USF Libraries
USF Digital Collections

Transport and survival of water quality indicator microorganisms in the ground water environment of Florida

MISSING IMAGE

Material Information

Title:
Transport and survival of water quality indicator microorganisms in the ground water environment of Florida implications for aquifer storage and waste disposal
Physical Description:
Book
Language:
English
Creator:
John, David E ( David Eric )
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
bacteria
viruses
inactivation
temperature
total dissolved solids
wetlands
septic systems
contamination
floridan aquifer
aquifer storage and recovery
Dissertations, Academic -- Marine Science -- Doctoral -- USF   ( lcsh )
Genre:
government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: Ground water resources are heavily used for drinking water supply and often as a receptacle for waste water. One concern is the possible contamination of wetland areas by ground water receiving septic system infiltration. To investigate this, two tracer studies were performed using the bacteriophage PRD-1 by seeding septic systems adjacent to wetlands with the phage and monitoring migration towards wetland areas. Transport velocities were evaluated based on appearance of tracer in sampling wells at various distances from the injection point. Velocities were estimated to be 0.25 m/d and 0.4 m/d at the two sites. Some retardation with respect to the conservative tracer SF6 was observed, with a factor of about 1.5. Due to dry conditions, the water table was well below surface, so transport of the virus into surface water was not observed. Survival of public-health-related microorganisms in ground water is also a concern. The effects of temperature and total dissolved solids (TDS) on survival of 5 groups of indicator organisms were evaluated in controlled experiments. TDS did not have significant effects on inactivation of these microbes up to 1000 mg/l, but there was indication of reduced inactivation of enterococci at TDS concentrations of 3000 mg/l. Increased temperature consistently resulted in more rapid inactivation. Survival in aquifer and reservoir water samples was also evaluated, and significant effects due to water type, temperature, and pasteurization treatment were observed. Inactivation was more rapid in surface water sources, and pasteurization enhanced survival. For enterococci and DNA coliphage, pasteurization effects were more pronounced in surface water. DNA coliphage and perhaps fecal coliform appeared to be the more-conservative indicator organisms for aquifer injection monitoring. Lastly, it was observed that inactivation rates were considerably slower in pore water of saturated limestone than in the bulk water column of similar water sources and conditions, particularly for enterococci and fecal coliform.
Thesis:
Thesis (Ph.D.)--University of South Florida, 2003.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by David E. John.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 322 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001441482
oclc - 53961836
notis - AJM5922
usfldc doi - E14-SFE0000155
usfldc handle - e14.155
System ID:
SFS0024851:00001


This item is only available as the following downloads:


Full Text

PAGE 1

Transport and Survival of Water Quality Indicator Mi croorganisms in the Ground Water Environment of Florida: Implications for Aquifer Storage and Waste Disposal by David E. John A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy College of Marine Science University of South Florida Major Professor: Joan B. Rose, Ph.D. John H. Paul, Ph.D. Debra E. Huffman, Ph.D. Charles P. Gerba, Ph.D. Ronald W. Harvey, Ph.D. Date of Approval: November 10, 2003 Keywords: bacteria, viruses, inactivation, temperature, total dissolved solids, wetlands, septic systems, contamination, floridan aquifer, aquifer storage and recovery Copyright 2003, David E. John

PAGE 2

Dedicated to my parents, without whose loving support and encouragement, none of my achievements would have been possible.

PAGE 3

ACKNOWLEDGEMENTS The author would like to thank the following people: committee members Dr. Rose, Dr. Paul, Dr. Huffman, Dr. Gerba, and Dr. Harvey, and my defense chair Dr John Lisle; Don Ellison and the Southwest Florida Water Management District; Peter Kwiatkowski and the South Florida Water Management District; the Florida Department of Health; Anne and Werner Von Rosenstiel for the Von Rosenstiel Endowed Fellowship fund; Don Shea of the St. Petersburg Downtown Partnership and the Robert M. Garrels Endowed Fellowship fund; the many lab mates who have helped both directly or indirectly with this research, namely Tracy Berg, Angela Gennaccaro, St ephaney Shehane, Jennifer Seiter, Mili Protic, Molly McLaughlin, Troy Scott, Tracie Jenkins, Walter Quin tero, Angie Coulliette, Julie Havens, Laura Cooney, and Michelle Woodall; numerous people from other agencies and groups who have helped, including Eric DeHaven, Jason Hurst, and Rob Degraff from the SWFWMD; Dave Demonstranti and Pete Dauenhauer of the SFWMD; Mark McNeal and Fran Bennett fro m CH2MHill; David Pyne from ASR Systems LLC; Harmon Harden from FSU; the Plaza Materials Corp.; th e staff of Albert Whitted WWTP; and last but not least all my friends and family for their support, love, and friendship.

PAGE 4

i TABLE OF CONTENTS List of Figures iv List of Tables vii Abstract ix Introduction 1 Chapter 1. Virus transport from septic tank system s near seasonally inundated areas through shallow aquifers 6 Introduction 6 Study Methods 11 Site Descriptions 11 Well Installation 12 Tracer preparation and injection 16 Sampling 16 Results 17 Duval County tracer study 17 Lake County tracer study 24 Discussion 31 Chapter 2: A review of factors affectin g microbial survival in ground water 37 Introduction 37 Studies on Viruses 39 Studies on Viruses and Bacteria 43 Studies on Bacteria and Cryptosporidium 45 Summary and Conclusions 47 Chapter 3: Survival of water quality indicator microorganisms in the ground water environment of Florida: Temperature and Total Dissolved Solids effects 58 Introduction 58 Study Methods 59 Organism Populations 59 Organism preparation for survival experiments 61 TDS-temperature trials 62 Data Analysis 64

PAGE 5

ii Results 66 Fecal coliform bacteria 66 Enterococci bacteria 70 RNA and DNA coliphage 73 PRD-1 bacteriophage 78 PBS controls and Variability 81 Discussion 87 Chapter 4: Survival of indicators in representative supply and receiving waters for aquifer storage and recovery in Florida 92 Introduction 92 Methods 95 Results 98 Fecal coliform 101 Enterococci bacteria 110 F+ RNA coliphage 116 DNA coliphage 124 PRD-1 bacteriophage 131 Variability and PBS Controls 137 Discussion 138 Chapter 5: Survival of Fecal Coliform, Enterococci, and F+ RNA Coliphages in Saturated Limestone 147 Introduction 147 Methods 150 Results 154 Discussion 167 Concluding Remarks 174 References 179 Appendices 186 Appendix 1: Observed data plots and fitted mode l curves for TDS-temperature experiments 187 A. Fecal coliform TDS-Temperatur e experimental data charts 187 B. Enterococci TDS-temperature experimental data charts 193 C. F+RNA coliphage TDS-temperature experimental data charts 199 D. DNA coliphage TDS-temperature experimental data charts 204 E. PRD-1 TDS-temperature trial experimental data charts 210 Appendix 2: Observed data plots and fitted model curves for aquifer and reservoir water sample studies with indicator bacteria populations 216 A. Fecal coliform Bradenton site natura l water experimental data charts 216 B. Fecal coliform Bradenton site raw water plot s and fitted first-order regression models 221 C. Fecal coliform West Palm Beach site na tural water experimental data charts 222 D. Fecal coliform West Palm Beach site raw water plots and fitted first-orde r regression models 227

PAGE 6

iii E. Enterococci Bradenton site natural water experimental data charts 228 F. Enterococci Bradenton site raw water plot s and fitted first-order regression models 233 G. Enterococci West Palm Beach natura l water experimental data charts 234 H. Enterococci West Palm Beach site raw water plots and fitted first-order regression models 239 Appendix 3: Observed data plots and fitted model curves for aquifer and reservoir water sample studies with bacteriophage 240 A. F+ RNA coliphage Bradenton site natural water experimental data charts 240 B. F+ RNA coliphage Bradenton site raw water pl ots and fitted first-order regression models 244 C. F+ RNA coliphage West Palm Beach natural wa ter and PBS control experi mental data charts 245 D. F+ RNA coliphage West Palm Beach site raw wa ter plots and fitted firstorder regression models 250 E. DNA coliphage Bradenton site natural water experimental data charts 251 F. DNA coliphage Bradenton site raw water plot s and fitted first-order regression models 256 G. DNA coliphage West Palm Beach site na tural water experimental data charts 257 H. DNA coliphage West Palm Beach site raw water plots and fitted first-orde r regression models 262 I. PRD-1 Bradenton site natural wa ter experimental data charts 263 J. PRD-1 Bradenton site raw water plots a nd fitted first-order regression models 268 K. PRD-1 West Palm Beach natural water experimental data charts 269 L. PRD-1 West Palm Beach site raw water plot s and fitted first-order regression models 274 Appendix 4: Results of statistical comparisons on days for predicted 2-log10 (99%) decline 275 A. Fecal coliform Temperature-TDS trials 275 B. Enterococci Temperature-TDS trials 277 C. F+ RNA coliphage Temperature-TDS trials 279 D. DNA coliphage Temperature-TDS trials 281 E. PRD-1 bacteriophage Temperature-TDS trials 283 F. Fecal coliform natural water trials 285 G. Enterococci natural water trials 290 H. F+ RNA coliphage natural water trials 295 I. DNA coliphage natural water trials 300 J. PRD-1 natural water trials, inactivation rate statistics 305 About the author End Page

PAGE 7

iv LIST OF FIGURES Figure 1. Site map of the Duval County (Maxville) site showing the locations of the drainfield, slotted sampling wells and multi-level sampling wells. 14 Figure 2. Site map of the Lake County site showing the locations of the drainfield, slotted sampling wells and multi-level sampling wells. 15 Figure 3. Total extent of PRD-1 tracer at Duval County, as detected in wells at any time, with initial detection given for contour shading. 20 Figure 4. Total extent of PRD-1 tracer at Lake County as detected in wells at any time, with day of initial detection given for contour shading. 28 Figure 5. Bacteria inactivation rate s, averaged by temperature range. 49 Figure 6. Virus inactivation rates averaged by temperature range. 49 Figure 7. Scatterplot of coliform bacteria inac tivation rates compiled from reviewed studies. 50 Figure 8. Scatterplot and regression of reviewed enterococci inactivation rates with respect to temperature. 51 Figure 9. Coliphage inactivation rates from reviewed studies with respect to temperature. 53 Figure 10. Poliovirus inactivation rates vs. temperature with exponential regression model. 53 Figure 11. Enterococci mean days for 2-log d ecline as a function of temperature and TDS. 73 Figure 12. F+ RNA coliphage mean days for 2-log decline as a function of temperature and TDS. 77 Figure 13. F+ RNA coliphage mean days for 2-log decline as a function of temperature and TDS, lower concentrations. 77 Figure 14. DNA coliphage mean days for 2-log decline as a function of temperature and TDS. 78 Figure 15. PRD-1 inactivation rates averaged by temperature and TDS. 80 Figure 16. Mean days for 2-log inactivation by experimental set for DNA coliphage and PRD-1. 83 Figure 17 A-D. Set differences for experimental trials and PBS controls. 85 Figure 18. Predicted days for 2-lo g decline for each organism, averaged across all TDS concentrations. 89 Figure 19. Mean days and confidence intervals for 2-log decline of fecal coliform in ASR water trials (surface and ground water combined). 105

PAGE 8

v Figure 20. Mean days and confidence intervals for 2-log decline of fecal in raw water at 22and 30C (surface and ground water combined). 106 Figure 21. Mean days with conf idence intervals for 2-log decline of fecal coliform in surface water: evaluation of temperature and pasteurization effects. 106 Figure 22. Mean days for 2-log decline of fecal coliform in raw surface water. 107 Figure 23. Days for 2-log decline of fecal coliform (m ean and confidence interval s) in raw and pasteurized ground water. 108 Figure 24. Mean days for 2-log decline of fecal coliform in raw ground water. 108 Figure 25. Chart of mean days for 2-log inactivati on for fecal coliform in raw aquifer and surface water at ambient temperatures for Florida. 109 Figure 26. Enterococci inactivation in ASR water samples. 112 Figure 27. Enterococci inactivation as days for 2-log decline in raw water at 22and 30C (mean days and confidence intervals). 113 Figure 28. Temperature and pasteurization effects on enterococci mean days for 2-log decline in ground water. 114 Figure 29. Chart of mean days for 2-log inactivati on for enterococci in raw aquifer and surface water at ambient temperatures for Florida. 115 Figure 30. F+ RNA coliphage days for 2-log declin e in raw surface and ground water, means are average by temperature and water type. 119 Figure 31. Mean days and confidence intervals for 2-log decline of F+ RNA coliphage in raw surface and ground water at 22 and 30 C. 119 Figure 32. Temperature and pasteurization effects on days for 2-log decline of F+ RNA coliphage in surface water, mean and conf idence intervals by temperature and pasteurization. 120 Figure 33. Temperature and pasteurization effects on F+ RNA coliphage inactivation in ground water, as mean days for 2-log decline. 121 Figure 34. Mean days and confidence intervals for 2-log decline of F+ RNA coliphage in raw ground water. 121 Figure 35. Chart of mean days for 2-log inactivation for F+ RNA coliphage in raw aquifer and surface water at ambient temperatures for Florida. 123 Figure 36. DNA coliphage inactivation in ASR water samples, showing effects of temperature and treatment on mean days for 2-log d ecline (with confidence intervals). 126 Figure 37. Days for 2-log decline for DNA colipha ge in raw surface and ground water, averaged by temperature and water type. 127 Figure 38. Mean days for 2-log decline of DNA coliphage in raw water at 22 and 30 C, averaged by temperature and water type (with confidence intervals). 127 Figure 39. DNA coliphage mean days for 2-log in activation in raw and past eurized surface water. 128

PAGE 9

vi Figure 40. DNA coliphage mean days for 2-log inactivation in raw surface water. 128 Figure 41. DNA coliphage mean days for 2-log inactivation in raw and pasteurized ground water. 129 Figure 42. DNA coliphage mean days for 2-log inactivation in raw ground water. 129 Figure 43. Chart of mean days for 2-log inactiva tion for DNA coliphage in ra w aquifer and surface water at ambient temperatur es for Florida. 130 Figure 44. PRD-1 first order inactivation rate constants in ASR aquifer and reservoir samples (log units per day), in combined raw and pasteurized water, averaged by temperature and pasteurization (combining both water types). 133 Figure 45. PRD-1 first order inactivation rates (log / d) in raw surface and ground water. 134 Figure 46. PRD-1 first-or der inactivation rates (log / d) in raw surface and ground water at 22 and 30 C, averaged by temperature and water type. 134 Figure 47. Chart of mean days for 2-log inactiv ation for PRD-1 in raw a quifer and surface water at ambient temperatures for Florida. 136 Figure 48. Fecal coliform survival in saturated limestone pore water ground water samples. 157 Figure 49. Fecal coliform surviv al in saturated limestone pore water surface water samples. 158 Figure 50. Enterococci survival in saturated limestone pore water ground water samples. 159 Figure 51. Enterococci survival in saturated limestone pore water surface water samples. 160 Figure 52. F+ RNA coliphage survival in saturated limestone pore water ground water samples. 161 Figure 53. F+ RNA coliphage survival in saturate d limestone pore water surface water samples. 162 Figure 54. Days predicted for 2-log inactivation of the three organism populations in pore water of saturated limestone. 164 Figure 55. Results of survival in PBS at 22 and 30 C for A. fecal coliform, B. enterococci, and C. F+ RNA coliphage. 166 Figure 56. Comparison of days pr edicted for 2-log inactivation of f ecal coliform in saturated limestone pore water and suspended in water column. 170 Figure 57. Comparison of days predicted for 2-log inactivation of enterococci in saturated limestone pore water and suspended in water column. 171 Figure 58. Comparison of days predicted for 2-log inactivation of F+ RNA coliphage in saturated limestone pore water and suspended in water column. 171

PAGE 10

vii LIST OF TABLES Table 1. Detection of PRD-1 Tracer at Duval County (Maxville). 18 Table 2. First appearance (peak conc.) of PRD-1 tracer, Duval County (Maxville). 19 Table 3. Tracer Movement Rates for Duval County. 21 Table 4. Comparison of tracer detection for PRD-1 and SF6 at Duval County. 24 Table 5. PRD-1 detection in drainfield mound wells at the Lake County (Groveland) site. 26 Table 6. PRD-1 detection in wells off the drainfield mound at Lake County (Groveland). 27 Table 7. PRD-1 tracer movement rates at Lake County. 29 Table 8. Tracer velocities calculated from the Lake County SF6 results. 30 Table 9. Bacteria inactivation rates from reviewed studies, compiled by temperature group. 48 Table 10. Virus inactivation rates from reviewed studies, by temperature range. 48 Table 11. Trends in inactivation rates possibly due to various factors in reviewed studies. 56 Table 12. Composition of mixed ion solution for TDS-temperature trials (Instant Ocean sea salt). 62 Table 13. Days predicted for 2-lo g decline of fecal coliform at vary ing TDS and temperature conditions from model curve equations. 68 Table 14. Days predicted for 2-log decline of entero cocci at varying TDS and temperature conditions from model curve equations. 71 Table 15. Predicted days for 2-log inactivation for F+ RNA coliphage. 75 Table 16. Predicted days for 2-log inactivation for DNA coliphage. 75 Table 17. PRD-1 first-order inactivation rate constants. 79 Table 18. Significance of TDS and te mperature as factors affecting survival of indicator organisms in water. 87 Table 19. Physiochemical measurements of raw water from the two sample sites. 99 Table 20. Microbial background measurements from raw water at the two sample sites. 100 Table 21. Background microbial concentrations after pasteurization of aquifer and reservoir water. 101

PAGE 11

viii Table 22. Predicted days for 2-log decline of fecal coliform concentrations in ASR ground water samples. 103 Table 23. Predicted days for 2-log decline of fecal coliform concentrations in ASR reservoir samples. 103 Table 24. Statistically significant factors affecting fecal coliform in activation in ASR water sources. 108 Table 25. First-order inactivation rates of fecal co liform in raw aquifer and surface water sources. 110 Table 26. Predicted days for 2-log decline of enteroco cci concentrations in ASR ground water samples. 111 Table 27. Predicted days for 2-log decline of entero cocci concentrations in ASR reservoir samples. 111 Table 28. Significant variables affecting ente rococci survival in ASR water sources. 114 Table 29. First-order inactivati on rates of enterococci in raw a quifer and surface water sources. 116 Table 30. Predicted days for 2-log decline of F+ RNA coliphage concentrations in ASR ground water samples. 117 Table 31. Predicted days for 2-log decline of F+ RNA coliphage concentrations in ASR surface water samples 117 Table 32. Significant variables affecting F+ RNA coliphage survival in ASR water sources. 122 Table 33. First-order inactivation rates of F+ RNA coliphage in raw aquifer and surface water sources. 123 Table 34. Predicted days for 2-log decline of DNA coliphage concentrations in ASR ground water samples. 125 Table 35. Predicted days for 2-log decline of DNA coliphage concentrations in ASR surface water samples. 125 Table 36. Significant variables affecting DNA co liphage survival in natural water sources. 129 Table 37. First-order inactivation rates of DNA co liphage in raw aquifer and surface water sources. 131 Table 38. First-order inactivation rates from PRD-1 in ASR ground water. 132 Table 39. First-order inactivati on rates in ASR surface water. 132 Table 40. Significant variables affecting PRD-1 survival in ASR water sources. 135 Table 41. First-order inactivation rates of PRD-1 in raw aquifer and surface water sources. 136 Table 42. Physiochemical and b ackground microbial characteristics of water sources for saturated limestone experiments. 155 Table 43. Pore water TDS, pH and turbidity. 155 Table 44. Days for 2-log inactivation of fecal colifor m, enterococci, and RNA coliphage in pore water of saturated limestone. 163

PAGE 12

ix Transport and Survival of Water Quality Indicator Mi croorganisms in the Ground Water Environment of Florida: Implications for Aquifer Storage and Waste Disposal David E. John ABSTRACT Ground water resources are heavily used for drinking water supply and often as a receptacle for waste water. One concern is the possible contaminati on of wetland areas by ground water receiving septic system infiltration. To investigate this, two tracer studies were performed using the bacteriophage PRD-1 by seeding septic systems adjacent to wetlands with the phage and monitoring migration towards wetland areas. Transport velocities were evaluated based on appearance of tracer in sampling wells at various distances from the injection point. Velocities were estimated to be 0.25 m/d and 0.4 m/d at the two sites. Some retardation with respect to the conservative tracer SF6 was observed, with a factor of about 1.5. Due to dry conditions, the water table was well below surface so transport of the vi rus into surface water was not observed. Survival of public-health-related microorganisms in ground water is also a concern. The effects of temperature and total dissolved solids (TDS) on survival of 5 groups of indicator organisms were evaluated in controlled experiments. TDS did not have significant effects on inactivation of these microbes up to 1000 mg/l, but there was indication of reduced inactivation of enterococci at TDS concentrations of 3000 mg/l. Increased temperature consistently resulted in more rapid inactivation. Survival in aquifer and reservoir water samples was also evaluated, and sign ificant effects due to water type, temperature, and pasteurization treatment were observed. Inactiva tion was more rapid in surface water sources, and pasteurization enhanced survival. For enterococci and DNA coliphage, pasteurization effects were more pronounced in surface water. DNA coliphage and perhaps fecal coliform appeared to be the moreconservative indicator organisms for aquifer injection m onitoring. Lastly, it was observed that inactivation rates were considerably slower in pore water of saturated limestone th an in the bulk water column of similar water sources and conditions, particul arly for enterococci and fecal coliform.

PAGE 13

1 INTRODUCTION Ground water resources are heavily used for domestic drinking water supplies in the United States and most of the world. Nationally, 40% of the U.S. domestic water supply originates from ground water. Furthermore, over 40 million people use ground water to supply their drinking water via domestic wells (Alley 1999). Of public water systems in the U.S., 92 % rely primarily on ground water for supply (Craun 2002). Worldwide, ground water represents a large majority of the drinking water supply in many nations, including Denmark, Portugal, Italy, Switzerland, Belgium, and the Netherlands, all of which derive more than 2/3 of their drinking water from ground water (Pedley and Howard 1997). Aquifers have, until the last few decades, been generally considered protected from potential sources of microbial or chemical contamination typically found in surface waters. Due to increasing population densities, agriculture, development and in dustrialization, and increased withdrawals from aquifers, however, the quality of ground water is in creasingly a concern. Numerous instances of ground water contamination and waterborne illness due to ingestion of ground water have been documented. Microbial contamination of ground water has been responsible for many disease outbreaks. In the U.S., at least 356 outbreaks of disease caused by contaminated ground water were documented between 1971 and 1994, representing 58% of all waterborne illness outbr eaks (Craun and Calderon 1997). Data for a more recent period (1991 1998) indicated that 74 outbreaks of waterborne illness occurred due to public water systems that used ground water, representing 68% of the waterborne disease outbreaks during that period (Craun 2002). This is likely an underestimation of overall incidence of illness due to frequent nondetection of outbreaks and a lack of reporting on s poradic and self-resolving illnesses. However, serious consequences can be the result, as estimated annual wate rborne disease deaths in the U.S. were reported in one review to be 900 1800 (Macler and Merkle 2000). While these high numbers are an estimate, documented illness due to drinking water contamination in the U.S. from 1991 1998 included 126 outbreaks and, excluding the large outbreak of cryptosporidiosis in Milwaukee in 1993, caused 8

PAGE 14

2 documented deaths. (Craun 2002). Since many individual, non-community, and community ground water wells and systems often are considered to be potable without treatment, disinfection has frequently not been required unless the aquifer is determined to be under direct influence of surface water. Regarding disease outbreaks due to both community and non-community ground water systems, inadequate disinfection or lack of disinfection was responsible for a significant proportion of the outbreaks (Craun and Calderon 1997). Specifically, inadequate or failed treatmen t in systems using ground water caused 58 illness outbreaks from 1991 1998, 31 of these outbreaks were due to untreated water, mostly in non-community water systems (Craun 2002). Beyond the important concern of waterborne disease due to consumption of ground water contaminated by surface sources, c ontaminated ground water may al so contribute to surface water microbial pollution. Several studi es employing virus tracers and/or chemical tracers have documented transport of wastewater from on-site sewage disposal systems (OSDS, septic tanks) to nearby surface water bodies such as canals, rivers, and marine environments (Paul 1995; Rose and Zhou 1995; Paul 1997; Dillon 1999; Paul 2000; Callahan 2001; Lipp 2001). Contamination of surface water via gr ound water flow can be more problematic in areas receivi ng high annual precipitation and that have a high water table. As discussed later, these conditions along with an of tentimes highly conductive hydrogeological setting are particularly evident in the state of Florida. Take n together, these factors pr esent a situation in which ground water contamination, particularly from OS DS, can have significant impacts on surface water microbial quality. A large number of different pathogenic or op portunist microorganisms can be responsible for ground water contamination. The pathogenic microo rganisms of concern include three major classes of microbes: viruses, bacteria, and protozoa. These or ganisms, as reviewed by Macler and Merkle, include waterborne viruses such as coxsackievirus, echovirus, rotavirus, norovirus, calicivirus, astrovirus, and hepatitis A and E. Bacteria of concern are chiefly pathogenic E. coli such as serotype 0157:H7, Salmonella and Shigella spp., Campylobacter jejuni and Aeromonas hydrophila among others. The main waterborne protozoa that may potentially be transmitted by ground water are Cryptosporidium parvum and Giardia lamblia (Macler and Merkle 2000). Recen t studies on the incidence of mi crobial contamination include an examination of wells in Wisconsin for enteric viral pathogens and indicators which detected viruses in 4

PAGE 15

3 out of 50 wells monitored four times over a year. Wells tested positive for rotavirus in three cases, and rotavirus, Norwalk-like virus and enteroviruses in the fourth positive well. However, there was not a correlation to the presence of F+ RNA bacteriophage or bacterial indicators, and contamination appeared to not be continuous since wells were not positive on consecutive samples (Borchardt 2003). A recent examination of waterborne disease in Finland determined that 13 of 14 waterborne illness outbreaks from 1998 1999 were caused by non-disinfected ground water. The cause in eight of these outbreaks was determined as Norwalk-like-virus (NLV, norovirus) and Campylobacter in three outbreaks (Miettinen 2001). Another European study reported on a community outbreak of illness due to Shigella sonnei attributed to well contaminat ion in Greece (Alamanos 2000). Cryptosporidium parvum has been implicated in a number of illness outbreaks from ground water as well. Over the period of 1984 1994, 4 out of 10 cryptosporidiosis outbreaks from U.S. drinking water systems were attributed to contamination of wells or wells influenced by surface water (Craun 1998). The regulatory framework governing ground wa ter use and impacts in the United States is provided mostly by the U.S. Safe Drinking Water Act (SDWA) and subsequent amendments, thus being administered by the U.S. Environmental Protection Agency (USEPA) but largely delegated to various state agencies for direct implementation. Regarding the use of ground water as a drinking water source, the SDWA establishes regulations and requirements only fo r public water systems, which are drinking water systems that have at least 15 connections or serve 25 people per day at least 60 days of the year. This includes community water systems, which serve the sa me people year round, and non-community systems which do not serve the same people year-round. No n-community systems are divided into transient (same people served but not year-round) and non-transient (serves different individuals for more than 6 months) water systems. Individual, privately owned wells are not regulated by the Safe Drinking Water Act or the USEPA (USEPA 1999). This regulatory framework sets standards for maximum contaminant levels (MCL) and treatment technology effectiveness for su rface water and ground wate r systems that are under the direct influence of surface water. Ground wa ter under the direct influe nce of surface water is determined by microscopic examination of samples from the aquifer and detection of particulates associated with surface water such as insect parts, pl ant debris, rotifers, and other materials. Thus, ground water under the direct influence of surface water is regulated the same as surface water under the Surface

PAGE 16

4 Water Treatment Rule and the Interim Enhanced Su rface Water Treatment Rule, including the use of sanitary surveys, source water protection practices, monitoring for indicators and pathogens such as Cryptosporidium and Giardia and requirements for filtration effec tiveness and disinfection standards. Recently enacted provisions under the USEPA’s Ground Water Rule set requirements for monitoring and treatment when necessary of ground water used for drinking water that is not considered under the direct influence of surface water. The Ground Water Rule a pplies to public water systems, not privately owned wells. These new requirements for protection of ground water used as drinking water include system sanitary surveys, hydrogeological sensitivity analyses for undisinfected systems, source water monitoring for undisinfected systems, and requirements for disinfection if f ecal contamination of the ground water source is indicated by monitoring. Furthermore, the Ground Water Rule sets disinfection requirements such that 4-log10 reductions of viruses should be achieved (USEPA 2002). The Safe Drinking Water Act along with stat e agencies also sets regulations and permit requirements for injec tion of water to aquifers an d the subsurface. The rules are set forth in Underground Injection Control (UIC) provisions, and classify such wells into 5 classes. Class I UIC wells inject hazardous and non-hazardous wastes to subsurface regions not considered to be underground sources of drinking water (USDW) with overlying confining layers. A USDW typically is considered to be an aquifer with total dissolved solids < 10,000 mg/L. Class II and III UIC wells are permitted for use by the oil, gas, and mining industries, while Class IV wells inject contaminated water to a USDW and are prohibited unless re-injecting treated water taken from a contaminated a quifer. Class V UIC wells are all other types of underground injection wells, and generally input water into or above a USDW. Aquifer storage and recovery (ASR) wells fall into this category, along w ith storm water drainage and aquifer recharge wells. Class V wells cannot adversely affect the quality of wa ter in the aquifer to which they inject (USEPA 2002). The subsurface environment of the Florida peninsula is dominated by abundant ground water, along with a small to non-existent vadose zone in many places. Three main aquifer systems are used for water supply in the state: the surficial aquifer, cons isting of the Biscayne aquifer and sand and gravel aquifer, the intermediate aquifer, and the deeper Floridan aquifer system, divided into the Upper Floridan aquifer, the middle confining unit, and the Lower Floridan aquifer. The Floridan aquifer system underlies

PAGE 17

5 the entire state of Florida, while ex tending northward to Georgia and even South Carolina. Florida receives a large amount of annual rainfall. The statewide average is 53 inches annually, and ranges from 69 inches in the panhandle to 40 inches in Key West. Part of this precipitation percolates to recharge aquifers, then discharging downward into deeper layers or laterally to surface water including springs, streams, lakes, wetlands, and marine waters. Thus, the surface wa ter and ground water environments are highly interconnected in many places, and th is interconnection can influence th e quality of both the ground water and surface water. Aquifers in Florida are composed of sedimentary rock forma tions, with the Floridan aquifer system being carbonate sedimentary rock such as dolomite and limestone of Tertiary age. Karst features are typical in many regions of Florida aquifers; karst geology is characterized by numerous solution channels, fissures, caves and often sinkholes This type of geology results in very highly conductive aquifers and possibly rapid lateral transport rates of ground water and potential contaminants. Also, much of the Floridan aquifer sy stem in south Florida is typified by water of high TDS, greater than 500 mg/L in the central part of the state and even higher, greater than 1000 mg/L, south of Lake Okeechobee (Berndt 1998). Along with a heightened state of awareness ab out potential ground water contamination has come interest within the regulatory, public health, and research communities to gain more information about the sources, transport, and fate of waterborne microorganisms in relation to aquifers and ground water. The research presented in this dissertation involves two obj ectives: (1) to describe virus transport associated with mounded septic tank systems near seasonally-in undated wetlands, and (2) to define fecal indicator microorganism survival in waters an d conditions representative of the Flor ida subsurface, particularly with emphasis towards survival of microorganisms in the context of aquifer injection or aquifer storage and recovery of surface water to the Upper Floridan aquifer.

PAGE 18

6 CHAPTER 1. VIRUS TRANSPORT FROM SEPTIC TANK SYSTEMS NEAR SEASONALLY INUNDATED AREAS THROUGH SHALLOW AQUIFERS Introduction Potentially harmful microbes may enter ground water via poor well construction, ground water recharge/infiltration from the surface, faulty septic tanks and/or sewer lines, land application of sewage sludge, and percolation of landfill leachate (Sobse y 1979; Pedley and Howard 1997). The fate of microorganisms in the subsurface depends on two basic processes, survival and transport/retention (Gerba and Bitton 1984). Study of the transport of microorganisms to and through ground water is an entire field onto itself. Considerable work has been done to define factors affecting microbial transport in ground water, generally with two motivating reasons: public health implications from contamination by potential pathogens, and transport of biodegrading bacteria to aquifer regions contaminated with chemical constituents. Transport studies often involve the use of columns to model movement through a soil matrix, or in-situ studies of microbial transport which employ monitoring wells to detect the organisms of interest, often a tracer organism, as they ar e transported with ground water across a study site. Column studies are useful for isolating and/or defining specific impacts controlling transport as they offer a controlled environment, while in-situ studies allow for evaluating the impact of other factors in the natural environment that are difficult or impossible to model with column studies. Such factors could include predation and antagonism by other organisms, alterations in adsorption and survival in response to natural geochemical constituents and pore size or transmissiv ity effects of the undisturbed aquifer material, and interrelation of these and other variables (Harvey 1997). Also, many physical parameters of water and contaminant transport, such as dispersion, have scale dependency, and thus in-situ studies more accurately model these parameters.

PAGE 19

7 Numerous factors have been identified which impact transport of bacteria and/or viruses in ground water. Beyond the bulk flow of water in an aquifer or soil (advection), physical and chemical parameters of the solid matrix, the ground water, and the organism s affect the degree to wh ich microbial particles are retained or transported and the relative rates at which they might move compared to the water itself. The primary mechanisms of retention in soil and aquifers are thought to be adsorption for viruses and sizedependent straining for bacterial and protozoan cells, although bacteria and to a lesser degree protozoa are also retained by adsorption (Gerba and Bitton 1984; Newby 2000). Electrostatic adsorption is one mechanism of retention. A major force governing adsorption is the electrostatic interaction between microbial particles and solid surfaces. This force is generally repulsive since microbes and soil surfaces generally have net negative charges. Two major determinants of surface charge on organisms are the isoelectric point of the cell/virion and pH of the wate r. By and large, microbial cells/particles have a negative surface charge in near-neutral water (Gerba 1984; Klein and Ziehr 1990; Krekeler 1991). But the overall charge on an organism is hi ghly variable. Isoelectric point (p I) is the pH at which the surface groups on a particle are neutralized via the bonding of an H+ ion at a negative site or loss of H+ from a positively charged site such as -OH2 + or phosphate groups (Gerba 1984 ). While positive and negative charges may remain on the surface, the net charge is ze ro. Thus, organisms in water of pH below their pI will be neutral to positively charged. The isoelectri c point of viruses varies among types and strains, and these variations are a major control on adsorption, such that adsorption is negatively correlated to isoelectric point (Dowd 1998). This effect also varies with soil and wa ter chemistry, such that water of neutral or higher pH facilitates adsorption, as does great er ionic strength (I) of the water and the abundance of trivalent or divalent cations such as aluminum and calcium on soil particles (Fontes 1991; Newby 2000). Adsorption will occur when the electrostatic repulsion is diminished enough for attractive forces to overcome it; attractive forces are a combination of van der Waals forces and hydrophobic interactions. Hydrophobicity of a microbial cell/particle also play s a role in its adhesion to surfaces. Generally, virus particles and bacteria have lipid side chains or lip id coats, as well as portio ns of surface proteins that will be hydrophobic. Usually, proteins fold in such a way that hydrophilic amino acid regions are exposed, but some hydrophobic regions will be exposed nonetheless and these may play a role in hydrophobic interactions (Gerba 1984; Newby 2000). Thus, organic content of the soil and ground water also affect

PAGE 20

8 adsorption. While higher organic content of soil ca n facilitate binding of more hydrophobic organisms (Newby 2000), greater dissolved organic matter can enhance transport by blocking hydrophobic binding sites or reduce adsorption of non-hydrophobic organisms such as the coliphage MS-2 (Powelson 1991). The water content of soil plays a large role in mi crobial transport. Comp arisons of retention of viruses have revealed that removal is greater through unsaturated vs. saturated ground water flow (Powelson 1990; Jin 2000) This is largely because of closer proximity of microbial particles to soil particle surfaces and possibly more rapid inactivati on under unsaturated conditions. Physical components of the soil or aquifer material such as grain size and other size-dependent exclusion factors such as cell size have a role in controlling transport via straining (G erba and Bitton 1984; Fontes 1991; Harvey 1997). Other hydrological factors are also important such as advective flow velocity and the heterogeneity of the aquifer system (Harvey 1997). Transport studies are often used to define modeling equation components, which can then be useful for predicting transport rates and distances through a particular aquifer based on the values of various parameters (Yates and Yates 1988; Sinton 2000). Septic tanks, or on-site sewage disposal systems, are the leading contributor of waste water to the sub-surface environment in the United St ates, and are believed to be respons ible for the majority of disease outbreaks relating to ground water contamination (Yates 1985). Tracer studies have been used to establish transport of virus particles from septic tanks to gro und water. A study on the bovine enterovirus BE-1 found it to have rapid, extensive movement through grou nd water in sandy soils with a transport rate of 35 m in 2 days (Scandura and Sobsey 199 7). Ground water contamination determined by tracer levels and fecal indicator constituents was more ex tensive at the site with the most co arse soil and shallowest aquifer. However, these studies were conducted in an area with a steep hydraulic gradient. Another recent field study found seeded MS-2 and X174 bacteriophage migr ated away from a septic tank at a rate of at least 12.9 m/d (over a 17.5 m distance) and concluded the 30.5 m setback distance allowed would not permit adequate natural disinfection to prot ect the ground water table (DeBorde 1998). Generally, optimal septic tank functioning is dependent upon a sufficient vertical distance between the tank and the water table. However, this may be more of a concern in areas lik e Florida with frequently very shallow water tables. Thus, areas in which there is a high density of septic tanks are at greater risk fo r contamination of shallow

PAGE 21

9 unconfined aquifers and drinking water sources need to be chosen carefully to avoid the risk of waterborne illness. Over the last several decades, much attention has been paid to issues regarding the transport of pathogens from septic tanks to potential drinking water sources and wells. However, in addition to ground water contamination, septic tank leachate has been found by a number of studies to impact surface water quality, via interaction between contaminated shallow a quifers and surface water. In particular, several reports have described the impact of septic tank effluent on canal and marine water quality in the Florida Keys. An initial published report on the presence of f ecal indicator bacteria in the surface water and near surface aquifer of Key Largo, Fl orida (Paul 1995) was followed by tracer studies which confirmed transport of bacteriophage from septic tanks to can als and near-shore marine water at Key Largo (Paul 1995) and at other locations in the Florida Keys (Pau l 1997; Paul 2000). Seeding septic tanks with the bacteriophage PRD-1 and HSIC-1 at Key Largo (Paul 1995) revealed transport to canals adjacent to test properties in approximately 11 hours and to marine water in 23 hours. Calculated rates of lateral transport based on time to detection at sample sites of various distances revealed transport rates from septic tank seeding that varied from 1.3 to 24.2 m/h, with an average of 13.5 m/h. Stud ies elsewhere in the Keys further delineated transport of virus from septic tanks to surface water on the order of several hours, with transport rates ranging from 2.5 35 m/h (Paul 1997). A similar study with a co nservative chemical tracer, sulfur hexafluoride gas (SF6) revealed detection in canals adjacent to test septic tanks in 8 hours (Dillon 1999). Such rapid transport to surface water in Key Largo sites was largely attributed to tidal pumping, while sites where tidal pumping was not as significant demonstrated slower transport rates of 0.12 2 m/h (Paul 1997). The presence of human enteroviruses in surface water of the Florida Keys was also demonstrated (Griffin 1999). Ther e are also indications that coral mucus may concentrate microorganisms of fecal origin, including human ente roviruses, leading to unk nown but possibly delete rious effects on these organisms and the associated stresse d marine ecosystem of coral reefs (Lipp 2002). Besides the Florida Keys, tracer studies have implicated septic tanks in contamination of surface water at estuarine (Rose and Zhou 1995; Lipp 2001) and freshwater sites in Florida, including demonstration of virus migration from a seeded septic tank to adjacent river waters (Callahan 2001).

PAGE 22

10 In 1990, 30% of Florida’s population used septic systems, discharging an estimated 450MG of wastewater per day (FDOH 1999). Current Florida Department of Health regulations require a 2 ft. (0.615m) separation from the high water table to the botto m of the septic tank drainfield to allow for proper treatment of the effluent. Since the water table is often close to the soil surf ace in Florida, drainfield mounds are a common solution to obtain this separation. However, the use of drainfield mounds allow for the installation of onsite sewage systems in areas previously considered too wet for traditional nonmounded systems. Due to locally shallow water tables, mounded systems are often installed near depressional wetlands, known as seas onally inundated areas (SIAs). SIAs are intermediate between terrestrial and aquatic environments both in their spatial location and in the amount of water to which they are accustomed. They are defined by the state of Florida as: “specific soil mapping units, of at least 0.025 acre that are classified in the Soil Legend of the applicable USDA Natural Resource Conservation Service (NRCS) Florida County soil survey as frequently flooded, ponded, depressional or slough, that are described in the detailed Soil Map Units of the applicable NRCS Florida County so il survey as very poorly drained; or that are classified in the Soil Legend of the NRCS County soil survey for Taylor County as commonly flooded.” (99-395 Laws of Florida) Since many of these depressional wetlands are characterized by very slight edge slopes, there is potential for larger areas to become inundated in their vicinity upon minor changes in water levels due to heavy episodic or seasonal rainfall. This storm flood zone around SIAs may present the potential for interaction of gr ound and surface waters and may threaten public health when ground waters containing human pathogens from septic systems emerge as surface water. To study this, tracer studies were undertaken to evaluate transport of viruses away from mounded drainfields towards an SIA at two sites in Florida to gauge virus and ground water transport rates and patterns in these model systems. This study was initiated by the Florida Department of Health, and involved three types of tracers: an inert gas, sulfur hexafluoride (SF6), two fluorescent dyes, fluorescein and rhodamine WT, and the bacteriophage PRD-1. This chapter reports on the transport study with the PRD1 tracer, with comparisons to transport of SF6 as a conservative tracer.

PAGE 23

11 PRD-1 has been used as a viral ground water tracer in a number of studies and serves as a surrogate for human pathogenic viruses (Bales 1995; Blanc and Nasser 1996; Ryan 1999). PRD-1 has also been successfully used as a ground wate r tracer in the Florida Keys as described above. It is an icosahedral lipid containing bacterioph age with a double-stranded DNA genome and a diameter of 62 nm (Olsen 1974). Several aspects of this organism make it useful as a virus transport model: its size and transport properties are similar compared to some human enteric viruses, detection methodology is relatively inexpensive and easy to perform, it is not commonly found as a natural inhabitant of environmental waters, it is harmless to humans, animals or plants, and it is rather persistent once introduced to ground water aquifers. Survival studies have suggested that PRD-1 may be a more appropriate model for survival of resistant pathogenic viruses such as hepatitis A virus in ground water (Blanc and Nasser 1996). Sulfur hexafluoride (SF6) is a water-soluble gas that is biologically and chemically inert, has a low background atmospheric concentration (10-15 mol/L), and can be detected at extremely low concentrations (Wanninkhof 1985). The potential of SF6 as a ground water tracer has also been reported, including its use in karst limestone and shallow, sandy aquifers (Wilson and Mack ay 1993; Dillon 1999; Corbett 2000). Study Methods Site Descriptions Florida Department of Health (FDOH) in cooperation with County health department staff selected five sites having septic tanks adjacent to SIAs for the complete study. Four of the sites had single family residential wastewater inputs to the septic ta nk while one site (Lake County) had inputs from a daycare center. All sites had three to seven feet of fine to medium sand over a clay-rich layer, with at least one foot of ground water (typically two to four feet) above the clay. Water in this shallow saturated zone above the clay lens typically flowed toward and discharged into the adjacent SIA. Two of these sites were used in a fall 2000 study for tracers tr ansport. A description of these two sites follows. Duval County (Maxville): The system site served a single-family home and was located 35 feet from an SIA mapped in the NRCS Duval County Soil Survey as #66 Surrency loamy fine sand,

PAGE 24

12 depressional, 0 to 2 percent slopes, and described in the detailed Soil Map Units as having a natural drainage setting of “very poorly drained”. Lake County (Groveland) : The system site was a child day care center and was located 30 feet from an SIA mapped in the NRCS Lake County Soil Survey as PmA, Placid and Myakka sands, 0 to 2 percent slopes, and described as “very poorly drained” and “in low marshy depressions”. Well Installation Each site was instrumented with both slotted and multi-level sampling wells (MLS). Slotted wells tend to provide integrated samples that are a mixture of different zones within the screened interval. The use of multi-level sampling wells allows for sampling of ground water at discrete depth intervals from a single bore hole. For installation of the slotted sampling wells, about 10 wells were employed at each site. A typical well pattern included wells located near the toe of the drainfield mound and wells that fanned out in a wider array approaching the SIA (see well location maps, Figure 1 and Figure 2). One well was located upgradient of the drain fi eld to provide backgroun d ground water quality at each site. Wells were installed typically to a depth of 3 to 7 feet using a hand auger. PVC well screen was used within the saturated zone with a solid casing attached which extended from the top of water table to the ground surface. After the well was placed at the appropriate depth, a sand pack with a bentonite plug was placed in the upper portion of the well bore. MLS wells were constructed using 1.9 cm OD PVC pipe as the housing to which 0.6 cm OD polypropylene tubing was attached. For this study, 3-7 polypropylene tubes were attached to the outside of a 1.5 to 3 m section of PVC pipe by plastic cable ties. Upon installation of the well, the PVC pipe was filled with material removed from the borehole and then capped. Sample depths were identified at the top of each piece of tubing. MLS sampling wells were in stalled using a hand auger with a 7.5-cm hollow barrel. To prevent the hole from back filling during construction, a 10-cm PVC casing (outer-casing) was inserted into the hole and moved downward as the hole was dug deeper. Once the desired depth was reached, the well was inserted into the hole, contained by the outer-cas ing. The outer-casing was then removed from the hole, allowing the aquifer materials to collapse around the sampler, isolating sampling points of the MLS at each level in the borehole. Add itional soil material, originally removed from the hole,

PAGE 25

13 was back filled to complete the well as necessary. Wells were typically cut flush to the ground and covered with a removable 15 cm plastic cover. Multi-level samplers (MLS) were installed to a depth of 1.5 to 3 meters. The MLS wells were arranged in three rows down gradient from the drain field injection point (Figure 1 and Figure 2). In this way flow rates at several intervals could have been calculated: (1) from the injection point to the first row on the mound, (2) from the first row to the second row at the toe of the mound, an d (3) from the second to the third row. Additional MLS wells were installed in the event that the tracer and wastewater plume traveled in an unexpected direction. There were 26 wells installed at Duval County site and 20 wells installed at the Lake County site.

PAGE 26

14 Figure 1. Site map of the Duval County (Maxville) si te showing the locations of the drainfield, slotted sampling wells and mult i-level sampling wells. Drain Field Top of Slope Toe of Slope Edge of Lawn s3 s5 s4 s6 s7 s8 s2 s1 1 2 9 5 ( n o t t o s c a l e ) 22 o 0,0 1,0 2,0 3,0 4,0 5,0 0,10,2 0,30,4 0,5 s93 2 0 30 o NorthM.T. Brown 10/26/99 (rev.11/25/99) (rev.12/29/99) 01020feetScales10Duval County South Maxville, FL m24 m22 m20 m19 m23 m21 m14 m13 m7 m6 m5 m4 m12 m1 m2 m3 m15 m16 m17 m18 m8 m9 m10 m11 m26 m25 LegendSlotted wells Multi-level sampling wells

PAGE 27

15 Figure 2. Site map of the Lake County site showin g the locations of the drainfield, slotted sampling wells and multi-level sampling wells. . s4 s3 s6 s8 s2 s1 s5 1 8 2 5 4 9 2 138 os9 W.L#1 W.L#2 Drain Field Top of Slope Toe of Slope North 01020feetScaleM.T. Brown 11/11/99 (rev.11/25/99) (rev.12/29/99)Wetland Edge 0,0 1,0 2,0 3,1 4,2 5,2 6,2 7,2 0,10,20,3 0,4 0,5 2,6 2,7 2,82,9 s7Lake CountyGroveland, FL s10 DP1 DP2 Legend Slotted wells Multi-level sampling wells m17 m16 m18 m8 m9 m10 m11 m12 m15 m13 m19 m20 m1 m2 m3 m4 m5 m6 m7 m14

PAGE 28

16 Tracer preparation and injection PRD-1 was propagated from an isolate obtained from University of North Carolina (Mark D. Sobsey) using the host Salmonella typhimurium strain LT2 (American Type Culture Collection ATCC #19585, Manassass VA). Phage were propagated in liquid culture 16 hours at 37 C in a shaking incubator. Host cell debris was removed by cent rifugation and filtration through 0.22 m filters. At each site, freshly prepared PRD-1 were diluted in approximately 160 lite rs onsite tap water (dechlorinated with 10% sodium thiosulfate) bubbled with 99.8% pure SF6 (Scott Specialty Gases) for at least 40 minutes prior to injection. Total quantities of PRD-1 seeded were 1.29 x 1014 PFU at the Duval County site and 4.2 x 1013 PFU at the Lake County site. Injection was performed by gravity feed over 2 hours into a pipe which was down stream of the septic tank and upstream of the drain field at each site. Sampling After tracer was injected into the septic system drain field, samples of ground water were taken from wells throughout the site to ascertain the exte nt and speed of virus transport with ground water through the subsurface. Pr ior to each injection, pre-seed samples were taken from all possible wells to ensure that no background PRD-1 were present. Afte r seeding, samples were collected from MLS and slot wells using 60-ml polypropylene syringes affixed with silicone tubing. Sampling syringes and tubes were sterilized by autoclaving prior to each sampling and changed for each well. Samples were transported on ice to the laboratory facility for processing within 16 hours. Samples were taken from the Duval County site at 0, 1, 2, 3, 7, 16, 21, 27, 35, 44, 56, 69, and 79 days post-seed and from the Lake County site at 1, 2, 4, 6, 12, 19, 26, 40, 54 and 65 days post-seed. Initial sampling events generally covered wells close to the injection point, and more wells were sampled in a wide r array at later dates, such that towards the end of each study nearly all wells with water were sampled. Analysis for the PRD-1 tracer was done with the host S. typhimurium LT2 (ATCC# 19585) using the double agar overlay method. Five replicates of 2 mL for each sample were done for a total of 10 ml sample analyzed. When necessary, samples were serially diluted in sterile 1 x PBS (phosphate buffered saline) to obtain a readable plate.

PAGE 29

17 Results Duval County tracer study A summary of samples that were positive for the presence of tracer at Duval County is given in Table 1, including concentrations of tracer detected an d the depth at which it was detected. In the case of slotted wells positive for tracer (DsS1, DsS4, and Ds S6), the depth indicates depth to water level below surface at that time. With the ex ception of the two sampling wells on the drainfield mound, (DsM1 and DsM2), PRD-1 was only detected in a given well on one sample event. Therefore, the first and only detection of tracer in each well was the peak c oncentration point. For wells on the mound, peak concentrations were observed on the day tracer wa s initially detected. Table 2 lists the day of initial detection for the tracer at Duval County. A rain event on Day 10 flooded much of the drainfield area and submerged several wellheads. Tracer was initially detected in several wells both on and off the drainfield mound on Day 21. This rainfall theref ore preceded initial tracer peaks by 11 days. Rainfall recorded at the nearby (20 miles) Jacksonville NAS was 2.41 inches over 48 hours. Smaller rain events were also recorded at the air station on Day 17 (0.18” ), Day 32 (0.57”), Day 37 (0.21”), and Day 67-68 (0.98” over 48 hrs.). Figure 3 portray estimates of the total horizontal exte nt of PRD-1 based on wells that were positive for the phage at anytime during each experiment.

PAGE 30

18 Table 1. Detection of PRD-1 Tr acer at Duval Co unty (Maxville). Date Days After Seeding Well Depth (m) A vg. Phage Conc. (pfu/ml) 10/10/00 21 DsM1 1.0 403.6 DsM1 1.4 177.3 DsM2 1.4 1.25 DsM6 0.6 34.75 DsM10 0.6 15 10/16/00 27 Note: all samples negative 10/24/00 35 DsM1 1.4 1.75 DsM1 1.8 2.75 DsM2 1.4 0 DsM2 1.8 0.25 DsM2 2.2 0 DsS1 0.47 3.25 DsS4 0.26 0.25 11/2/00 44 DsM9 1.0 25 DsM11 1.0 0.5 DsM16 0.6 2.5 DsM17 0.6 9 DsM18 1.0 0.5 DsS6 1.03 7 11/14/00 56 Note: all samples negative 11/27/00 69 DsM17 0.6 0.15 12/7/00 79 DsM15 0.6 0 DsM15 1.0 0.25

PAGE 31

19 Table 2. First appearance (peak conc.) of PRD-1 tracer, Duva l County (Maxville). Well Distance (m) Date Days after seed DsM1 3 10/10/0021 DsM2 1 10/10/0021 DsM6 5.7 10/10/0021 DsM10 9 10/10/0021 DsS1 5.4 10/24/0035 DsS4 9.7 10/24/0035 DsM9 10 11/2/00 44 DsM11 10 11/2/00 44 DsS6 12.9 11/2/00 44 DsM18 13.1 11/2/00 44 DsM17 14.1 11/2/00 44 DsM15 13.3 12/7/00 79

PAGE 32

20 Figure 3. Total extent of PRD-1 tr acer at Duval County, as detected in wells at any time, with initial detection given for contour shading. A travel velocity determination was made based on the first pulse of the tracer plume detected, in which it moved to wells DsM1, DsM2, DsM6, and DsM10 by Day 21. The average velocity for movement to DsM10 by this day was 0.429 md-1. This velocity included transport both within the drainfield mound and at the natural grade. It is important to make th e distinction between transport within the mound, which

PAGE 33

21 is likely composed of disturbed, more heterogeneous soil, to transport at the natural grade of the land outside the mound with relatively undi sturbed soil. But due to the pattern in which the tracer was detected, calculating transport velocity from the injection point to the most distant well in which peak tracer concentrations were detected is the best available op tion. There was an initial pulse of the tracer plume, and all wells sampled, including the mound wells, were negative on Day 27. Thus, it is possible that the detection of tracer at these later times was due to slow desorption of virus from the soil and subsequent transport to outlying wells in a more lateral pattern, or slower movement outside a preferential flow path. Movement rates were also calculated for detection of tracer in the more outlying wells. These rates were determined by using the farthest well in which the tr acer made its appearance on that sampling day. On T=35 days, the tracer was detected at well DsS4 for the first time (9.7 m), on T=44 days, the most distant positive well was DsM17 (14.1 m), and on T=79 days, tracer was detected in well DsM15 (13.3 m) for the first time. Tracer was thus eventually detected at di stances up to 14.1 m from the injection point. Rates calculated to these wells are given in Table 3 and the range of all calculated rates was 0.168 to 0.429 md-1. When values for this site were averaged, the mean travel velocity was 0.299 md-1. These averaged velocities include calculated transport after the initial observed pulse and may be that of a secondary pulse of desorbed virus. Table 3. Tracer Movement Rates for Duval County. Reference Point to Well… Distance (m)Days Movement Rate (m/d) Injection DsM10 9 21 0.429 DsS4 9.7 35 0.277 DsM17 14.1 44 0.320 DsM15 13.3 79 0.168 The conservative SF6 tracer was seeded concurrently with PR D-1, and results from the analyses of its transport performed by another research group ar e available in the published article on this study (Harden 2003). A day after injection, SF6 was observed in the mound wells, DsM1, DsM2, DsM3, and by the third day throughout the drain field. Background samples prior to this injection contained SF6,

PAGE 34

22 presumably from a previous tracer study the prior winter /spring. The rains prior to this injection could have mobilized a portion of the SF6 injection slug. Relatively high concentrations of SF6 were observed in well DsM2 four months after injection for the previous study, indicating SF6 can reside in septic mounds for long periods of time. SF6 was observed 3 days after injection in all wells sampled, including the slotted wells in the SIA (DsS5, DsS6, DsS7, DsS8, and DsS9), lending further support to this hypothesis. Rates could not be calculated for the SF6 observations during the first three days of the experiment due to uncertainty in the initial time reference. After three days, concentrations of SF6 concentrations generally decreased and later peaks were observed, possibly associ ated with a rain event on Days 9 and 10 of the experiment. Table 4 compares detections of these SF6 peaks to appearance of PRD-1 in respective wells. In comparing PRD-1 data to SF6 tracer data (Brown 2001 ; Harden 2003), the SF6 plume, indicated by concentration peaks of SF6, was initially detected at the same most-distant point of well DsM10 at the same time as PRD-1, Day 21. Thus, the velocity calculated for the initial plume movement in the direction towards the SIA was identical, 0.429md-1. After exact correlation of these two tracers initially, less agreement was observed in their de tection over subsequent sample events and in more distant wells. Spread of the SF6 tracer at Duval County covered a greater area near the mound than PRD-1, with peaks detected in wells DsM4, DsM5, DsM7, DsM8, and slo tted wells DsS4, S1 and S3. However, PRD-1 was detected, at low concentrations, in more outlying wells than SF6, and in a more widely dispersed fashion, as shown in Figure 3. From Table 4, for wells at the toe of the drainage mound, well DsS1 had a peak of SF6 on Day 27, but not until Day 35 for PRD-1. In third row wells, “peak” SF6 concentrations were not detected until Day 79 in well DsM9 and not at all in well DsM11, while PRD-1 was found in these wells by Day 44. Lower concentrations of SF6 were detected in these wells at earlier ti me points, but were below what would be considered a peak for determining tracer breakthrough (Brown 2001). Also, SF6 was detected in wells DsM8, DsS2 and DsS3 on Days 21 and 27, while PRD1 was never detected. PRD-1 was detected in the outer MLS wells (on Days 44 and 79), in lo w concentrations, while no associated SF6 peak was observed except in well DsS6, where SF6 detection was much sooner than PRD-1 in that well. It may be that SF6 concentrations were below detection limits in the ot her outer wells, as PRD-1 concentrations were very low. Tracer plume velocities that were determined based on detection of SF6 peaks averaged 0.36 0.16

PAGE 35

23 md-1. This is only slightly more rapid than the average velocity determined from PRD-1 detection of 0.299 md-1. However, many wells included in these calcu lations were different between those positive for SF6 and those positive for PRD-1. Direct comparisons we re made for several add itional wells in which both PRD-1 and SF6 were detected. These were well DsS1, with velocities in md-1 of 0.26 (SF6) vs. 0.15 (PRD1), well DsS4 with 0.36 (SF6) vs. 0.28 (PRD-1), and well DsS6 with 0.42 (SF6) vs. 0.26 (PRD-1). For well DsM9, due to the detection of multiple trace concentrations of SF6 after Day 3 (when all wells were positive) but before Day 79 when the “significant” peak occurred, there is more ambiguity surrounding the detection of tracer at this well. However, if veloc ities were compared based on reported breakthrough, they were 0.13 md-1 for SF6 and 0.23 md-1 for PRD-1, making it the only compared set of velocity values where PRD-1 was not slower than SF6.

PAGE 36

24 Table 4. Comparison of tracer detect ion for PRD-1 and SF6 at Duval County. Distance SF6 Peak PRD-1 Peak (m) (Days) (Days) Wells on mound DsM1 3 21 21 DsM2 1.5 21 21 DsM3 3 21 Wells at Toe of mound DsM4 8.5 27 DsM5 5.8 21 DsM6 5.7 21 21 DsS1 5.4 21 35 Third Row, Wells away from mound DsM8 12.7 21 DsM9 10 79 44 DsM10 9 21 21 DsM11 10 44 DsS2 10.3 27 DsS3 8.2 21 DsS4 9.7 27 35 Outer wells, on edge of SIA DsM15 13.3 79 DsM16 13.7 44 DsM17 14.1 44 DsM18 13.1 44 DsS6 11.4 27 44 Lake County tracer study At the Lake County site, PRD-1 tracer quickly migrated on the drain field mound and was first detected in well LM1, on T=1d. From there, tracer was detected at wells LM2 and LDP2 on the mound, and LM6 off the mound, on T=12d. By T=19d, it was also present at LM11. It remained detectable in these wells for the duration of the sampling efforts, up to T=65d. Unlike most years, conditions in fall 2000 were very dry. Florida was in one of its worst droughts in history and as a result the water table fell

PAGE 37

25 below many of the wells. Only a few samples were possi ble from slot wells as almost all were dry. Table 5 and Table 6 show detection of PRD-1 in the Lake County wellfield in wells on the drainfield mound (Table 5) and at the natural grade (Table 6). Initial detection on the mound wells was in LM1 by Day1, and off the mound in well LM6 on Day 12. Well DP2 is a drainpipe well that was similar to the slot wells. PRD-1 concentrations were much higher over the duration of the study here than at the Duval County site. In the first 20 days, PRD-1 concentrations were generally greater than or equal to 103 pfu/ml, and declined to concentrations less than 103 after 20 days. Figure 4 portrays the extent of PRD-1 tracer detection at this site. After Day 19, the tracer was not detected in any additional wells, but remained detectable in the shaded areas.

PAGE 38

26 Table 5. PRD-1 detection in drainfield mound wells at the Lake County (Groveland) site. DepthAvg. Phage Conc. (m)(pfu/ml) 10/5/001LM11.81858 10/6/002LM11.81350 LM12.210,775 LM12.638,250 10/8/004LM11.82550 LM12.26500 LM12.630,250 10/10/006LM11.85825 10/16/0012LM12.24925 LM22.23750 LM22.63700 LM23.01400 DP2NR559 10/23/0019LM11.81000 LM12.620 LM21.80 LM23.0800 DP2NR167 10/30/0026LM1alldry LM22.2450 LM23.0559 DP2NAdry 11/13/0040LM1all dry LM2 2.2727 LM22.61015 LM23.01100 11/27/0054LM22.6535 LM2 3.0 435 12/8/0065LM11.812 LM12.276 LM22.2115 LM22.6725 LM23.0784 DateDays After Seeding Well

PAGE 39

27 Table 6. PRD-1 detection in wells off the dr ainfield mound at Lake County (Groveland) DepthAvg. Phage Conc. (m)(pfu/ml) 10/16/0012LM61.40 LM62.6386 10/23/0019LM61.8403 LM62.6121 LM112.61080 10/30/0026LM61.40 LM62.6521 LM112.28000 LM112.6982 11/13/0040LM61.4300 LM62.8872 LM111.4305 LM112.8280 11/27/0054LM61.868 LM62.6723 LM112.6580 12/8/0065LM61.820 LM62.257 LM111.89 LM112.260 DateDays After Seeding Well

PAGE 40

28 Figure 4. Total extent of PRD-1 tracer at Lake Count y, as detected in wells at any time, with day of initial detection given for contour shading.

PAGE 41

29 The rates of tracer movement throughout and off the drainage mound were calculated. Table 7 summarizes these calculations. The key movement for determining flow rates was the migration of tracer from well LM6 to LM11, since this velocity measures transport due to more natural gradient flow conditions in the relatively undisturbe d soil, rather than conditions in th e mound soil. The travel velocity from well LM6 to LM11 was 0.314 md-1. Movements of the tracer from LM1 to LM6 and LM1 to LM11 were also considered, although th ese are transport velocities of the tracer in moving off the drainfield mound and represent somewhat forced-flow conditions due to the artificial elev ation gradient. These velocities were 0.590 md-1 from LM1 to LM6, and 0.467 md-1 from LM1 to LM11. Rates from LM2 to LM11 were not considered, since the tracer was already detected past LM2, in LM6, on its first appearance in LM2. Travel velocities on the drainfield mound ranged from 13.7 to 0.93 md-1. Although not shown on Table 7, the velocity from injection point to LM6 was 1.3 md-1 and to LM11 was 1.05 md-1. Table 7. PRD-1 tracer mov ement rates at Lake County. Reference Point to Well… Distance (m) Days Movement Rate (m/d) Injection LM1 13.7 1 13.7 LM2 13.2 12 1.1 LDP2 11.2 12 0.93 LM1 LM6 6.5 11 0.59 LM11 8.4 18 0.467 LM6 LM11 2.2 7 0.314 Samples taken for PRD-1 analyses were concurrently analyzed for concentrations of the conservative tracer SF6 by the Florida State University research group. The comprehensive SF6 tracer results are reported elsewhere (Brown 2001 ). A summary of the Lake County SF6 data are presented here for comparison. Multiple SF6 peaks were observed at wells on and off the drain mound. In the mound wells LM2 and LM3, and also wells LM4, LM5, LM6, and LM7 at the toe of the mound, peaks were observed 2 and 7 days after injection. Peaks occurred at LM1 on Days 7 and 19. On Day 7 there were also concentration peaks in wells away from the mound at LM9, LM10, LM11, LM12, LM13 and LS4. Well LM11 also had later peaks on Days 26 and 40, which was the only well not on or at the toe of the mound to

PAGE 42

30 have additional concentration peaks after the 7th day. Table 8 shows velocity of the SF6 tracer determined from migration to several key wells. Velocities were based on the greatest extent of the tracer plume on a given sample day, such that a well was not considered for velocity calculation if the tracer plume also first appeared beyond it at another more distant well on the same day. The first two values in Table 8 represent the first detection of tracer migration on Day 2, wher e it was found as far as wells LM6 and LM7. This velocity then represents migration of SF6 both throughout the drainage mound and off the slope. The remaining velocity values are for wells at the natural grade. While the average of velocities in moving off the mound was about 8 md-1, the average velocity for wells at grade was only 0.93 md-1. The rate of 0.48 md-1 calculated from the first peak concentrations found in both LM6 and LM11 was the only situation where presence corresponded to that of PRD-1. Although SF6 arrived at both these wells sooner than PRD1, the velocity between the two wells was similar for the two tracers with PRD-1 moving at 0.314 md-1. However, this is only 1/3 as fast as the average off-mound velocity of SF6 of 0.93 md-1. Table 8. Tracer velocities calculated from the Lake County SF6 results. These values are based on the farthest extent in a given direction of the SF6 tracer plume at detection of initial peak concentrations. InjectionLM616.028.00 LM716.128.05 LM4LS43.450.67 LM5LM1310.152.02 LM103.450.67 LM6LM112.250.48 LM7LM114.350.86 LM124.350.86 Average Velocity (m/d) Reference point to well…Distance (m) Days Transport patterns of the two tracers on and off th e drainfield mound were different other than the movement from LM6 to LM11. PRD-1 moved quickly to LM1 where SF6 was not detected until Day 7. The lateral expanse of plume movement (lateral to direction of the SIA) was more widespread for SF6 than that shown for PRD-1, extending more towards the east and west by Day 7. In contrast, phage was not detected off the mound until Day 12, at well LM6, and then in wells LM6 and LM 11 by Day 19.

PAGE 43

31 Therefore, although velocities of the two tracers we re similar when considering the movement from LM6 to LM11, the spread of the plume of SF6 was much greater spatially and more rapid than the PRD-1. Discussion As described in this chapter, the bacteriophage PRD-1 was used as a tracer of virus movement from mounded septic systems to the underlying ground water at two sites. Detection of the tracer in sampling wells to the surficial saturated soil at these sites confirmed migration of viruses away from the septic tanks’ drainage mound to the underlying shallow ground water and towards depressional wetland areas. No published studies were found that previously document viral tracer studies in this type of mounded system near seasonally-inundated areas. It is also important to recognize in the context of other ground water tracer studies that the present studies involved sites that were actively used septic systems, and thus provide a picture of realis tic situations rather than controlled-conditions. These involved the tracing of ground water flow due to the natural hydraulic gradient surrounding the septic drainfields under actual loading conditions provided by normal use of the facilities at the sites studied. Thus, we had no control over loading rates or the velocity and direction of flow of effluent from the septic systems. In-situ microbial transport studies through porous media are re ported relatively infrequently due to several general reasons: there is a much larger degr ee of complexity in the field environment than in laboratory studies to examine transport and attenuation behavior, there are issues of site selection and sampling well construction that complicate matters, site availability c oncerns, and issues with obt aining cooperation from willing landowners when private facilities are studied. Thus, this study is relatively rare in that two private facilities were studied in the course of normal use, an d the results are very important for evaluation of virus transport behavior under these natural, uncontrolled conditions. The sites used for these studies were typified by the use of mounded drainage areas for dispersion of septic tank effluent and to provide an additional distance of unsaturated soil above the water table to allow adequate attenuation of possibly harmful microorganisms in effluent prior to introduction into saturated zones where lateral transport and survival may be enhanced. The mounds are necessary because

PAGE 44

32 water tables at these and other similar sites are quite sh allow. Nonetheless, we have confirmed transport of virus from the septic drainage system to shallow ground water at the natural grade of the surrounding area. Vertical transport of virus from a model septic system in sandy soil of Florida similar to that found at these sites has been described previously under controlled loading rates at a lysimeter station. Under loading rate conditions of 0.063 md-1 and 0.032 md-1, PRD-1 peak breakthrough after transport through 0.6 m was observed at 2.2-3 days under high loading rate cond itions and 4 days under lower loading rate conditions (Nicosia 2001). That study suggested current Florida regulations requiring 0.6 m of unsaturated soil above the water table for septic drainage systems would not pr event significant concentrations of virus from septic effluent from reaching the saturated zone. In the pr esent study, we further demonstrate that under normal household loading conditions at the tw o sites, tracer viruses were also detected in the ground water in sampling wells away from the drainage mound. Both septic systems employed in this study were undergoing active, normal use; however, the nature of the locations differed in that the Duval County site was a single-family residence and the Lake County site was a child day-care center. The differ ent nature of the two sites likely influenced loading rates, such that the flow from the day care septic syst em would be expected to be greater. This may have influenced the pattern of transport of tracer at the two sites. At Duval County, tracer migration was not detected either on or off the drainage mound until Day 21, for PRD-1 and SF6. This migration also occurred as a pulse, moving the distance of 9 m to well DsM10 in the 5 days between Day 16 and Day 21, whereas no tracer had been detected in the 16 days prior to this. In contrast, at Lake County PRD-1 was detected at well LM1 within the drainfield mound by 1 day after seeding, and at LM6 off the mound by 12 days after seeding. SF6 was detected at numerous wells off the mound by Day 2. Also, PRD-1 was detected continuously at several wells throughout the study period. Thus, there was an apparently much more rapid, expansive (for SF6) spread of septic effluent at the Lake County site, while at Duval County movement and detection was more sporadic and at lo wer concentrations. The loading rates of the two different sites may be partly responsible for this diff erence. Also, weather conditions were such that the Duval County site received more rain events than La ke County. While rain events at a nearby weather station were recorded for Duval County on Days 10 (~ 2.5”), 17, 32, 37 and 67, no rain was recorded for Lake County at the nearby (14 mi.) Okahumpka weather station until Day 52 of the study, and this was the

PAGE 45

33 only rain event. The patterns of tracer movement an d detection at the two sites suggest that at the presumably higher-loading day care center, pore water flow and associated tracer movement could have been more a function of effluent loading from the septic system, while at the single family residence rainfall and percolation may have resulted in the moveme nt pulse. Also, the pulsed pattern of migration of PRD-1 at Duval County, and subsequent low-level detec tion in laterally more dispersive pattern, suggests the existence of a preferential flow path which the virus followed to well DsM10, perhaps due to flushing from rainfall on Days 10 and 17, with less rapid dispersion occurring throughout the well array at much reduced concentrations, possibly due to adsorption. Ob servations such as these highlight the importance of this study as an examination of transport conditions at the field scale and demonstrate that natural-gradient studies in normal-use situations su ch as these can provide useful info rmation about transport patterns of viruses. Although we did not employ controlled loading conditions, these results reveal interesting differences in transport patterns that may be due to the different types of normal us e. These studies involve mounded septic systems and such studies have not recei ved examination for virus impacts to ground water in published literature to date. The mere detection of viruses in significant concentrations in sampling wells at grade demonstrated the potential impact of moun ded septic systems in areas with high water tables, particularly in areas that this shallow ground water may drain to surface water such as an SIA. Due to the greater level of control that may be exerted over parameters regulating transport, most transport studies involve columns or other devices at the laboratory scale to investigate intricacies of microbial transport through satura ted media, including parameters such as adsorption and retention kinetics. Although several field viral tracer studies have been reported, they are relatively few. Septic tank tracer studies have been used previously to investig ate migration of viruses from such systems and to delineate their transport patterns in the underlying ground water. In cases where other septic tank tracer studies have been published, it is important to note that at the field scale, all sites are different, particularly with regards to basic hydrological parameters such as hydraulic conductivity and gradients, type of soil/rock, water chemistry, and bulk pore-water velocity and direction. Several tracer studies involving septic systems in coastal areas of Florida (Florida Keys) have been published, which describe transport of seeded viruses to the shallow ground water, canals, and near-shore surface water. Some transport velocities of seeded virus reported in these studies were much more rapid than observed in our studies, including

PAGE 46

34 velocities of 2.5 to 35 mh-1 (60 840 md-1) where tidal pumping was implicated and slower movement rates of 0.12 to 2 mh-1 (2.9 48 md-1) where tidal pumping was not thought to be significant (Paul 1997). Another septic tank tracer study in the Florid a Keys observed velocities of 1.7 to 57.5 mh-1 (41 1380 md-1) (Paul 2000). These studies demonstrate the impact of a much more fluid, dynamic hydrological environment on tracer studies, whereas rates observed for the tracer moving off the drainage mounds in our studies were on the order of 0.2 0.5 md-1 and tracer migration was measured on the scale of days and weeks rather than hours. However, some velocities found for movement within the drainage mound at Lake County were comparable to lower rates observed in the Keys were tidal pumping was not significant. Other viral tracer studies in more-inland areas ha ve also been reported. A septic system tracer study involving bovine enteroviruses characterized rapid transport of the virus from the septic system on the order of 35 m over 2 days, but the steep land slope and hydrologic gradient likely influenced this (Scandura and Sobsey 1997). A bacteriophage (MS2 and phi-X174) tracers study involving a high-school septic system found transport of a portio n of the seeded viruses to be 1 to 3 md-1, which was at least as fast as the conservative Brtracer (DeBorde 1998). In general, virus tr ansport velocities in field situations seem to be chiefly dictated by background hydrological conditions. Most field studies do not report significant retardation of viruses such as PRD-1 relative to conser vative tracers. Examples of this are found in natural gradient studies performed at field study sites that did not involve septic systems (Bales 1995; Pieper 1997; Deborde 1999), where no retardation of PRD-1 was observed. A forced-gradient study with PRD-1 and other viruses also observed no retardation (Woessner 2001). These studies all involved sandy soils and/or sand and gravel aquifers, not typified by fractured fl ow. In a viral tracer study (MS2 and PRD-1) involving a fractured clay-rich till (McKay 199 3), virus velocities were much greater than the bromide tracer, indicating micropore exclusion colloid transport behavior. Interestingly, in the studies reported here, also in sandy soils, observed detection of the two tracers did suggest some retarda tion of PRD-1 transport compared to SF6. For instance, at Duval County, initial de tection of the tracers as a pulsed movement between Day 16 and Day 21 (to as far as well DsM10) was the same for both tracers, possibly indicating a preferential flow path. But when comparing velocitie s for wells off the mound in which both tracers were detected after this “pulse”, possibly of a more late ral dispersion of tracers, the averaged retardation coefficient for PRD-1 was 1.5 (excluding transport to well DsM9 due to ambiguity surrounding SF6

PAGE 47

35 detection there). At Lake County, transport ve locities off the mound between wells LM6 and LM11 showed the same value of 1.5 for retardation. However, here transport of SF6 was in general much more rapid and disperse than PRD-1, with velocities leaving the mound to well LM6 showing a retardation factor for PRD-1 of 6.2 (8.0 md-1 vs. 1.3 md-1). Thus, in general, behavior of PRD-1 at these sites showed some degree of retardation, indicating equilibrium adsorption behavior. From the results presented here, making observations on behavior related to PRD-1 concentrations is difficult, such as kinetics of inactivation and removal by soil passage. Die-off of the virus was not specifically evaluated. Also, we did not evaluate concentrations of PRD-1 immediately upon leaving the drainage pipe area but rather at down-gradient wells, after the effluent had passed through soil for several meters. However, even though the objective of the stud y was velocity and extent of transport, rather than attenuation, declines of PRD-1 concentrations were ob served over time. This decline was more steady and gradual at Lake County, and presence of the tracer was observed over the entire duration after it first appeared. At Duval County, PRD-1 was already at low concentrations when first detected at Day 21, and sporadic detection of PRD-1 after the initial pulse pr ecluded observation of steady declines, but low levels were detected throughout the study period up to 79 days. The phage at Duval County may have passed through the wellfield more rapidly and completely due to possibly greater bulk water flow as a result of rain events and a preferentia l flow path. Nonetheless, our results indicate that PRD-1 is a stable virus even at higher temperatures observed in subtropical regions like Florida, and inactivation is slow enough to allow its use in tracer studies fo r periods of over 2 months. Weather conditions during the fall, 2000 portion of this study were unseasonably dry, and the area was still in a severe drought. As a result, rainfall during the previous months was significantly below average. Surface water was not cons istently present, and the slotted wells at the Lake County site were almost all dry. Thus, although migration rates of th e viral tracer were determin ed for the sh allow aquifer surrounding the drain field, the inci dence of migration of virus from th e shallow aquifer to surface water in seasonally inundated areas was not established. But these tracer stud ies did document movement of virus from septic systems through surficial ground water towards seasonally inundated areas. The migration rates and patterns observed were a function of bulk ground water flow, which indicated that movement towards the edge of the SIA in each site was the pr imary direction of flow. Therefore, even though

PAGE 48

36 conditions were unseasonably dry, tran sport observed in this study demonstrated the potential for migration of viruses from mounded septic systems towards nearby wetlands. It may be possible that given wetter weather and higher water tables, interaction of groun d water containing microorganisms of human fecal origin and surface water in seasona lly inundated areas would occur. The migration rates towards these SIAs may be used to estimate appropriate setback distances of septic tanks from SIAs when viral inactivation rates are incorporated in the calculation. Future studies in more normal climatic conditions would be beneficial to determine mo re accurately if septic system effluents near depressional wetlands may actually impact surface water quality. Future investigations involving actively-used septic systems to document impacts on ground water quality would also be helpful. Such studies can provide information about virus and septic tank effluent transport that cannot be obtained from forced-gradient studies and even studies in established field test sites where conditions are better controlled. While more diff icult to model and predict, transport behavior in response to “natural” temporal loading patterns and rates from septic systems would be very useful to document for comparison to controlled-condition studies. This type of normal use tracer study would be enhanced by accurate measurement of household input to septic systems (i.e. showers and flushing), along with good characterization of soil, background hydrology and engineer ed parameters (i.e. placement of drainpipes). Such a study would be much more difficult to implement than one at a field experimental study site, largely due to the requisite participation from private parties. Also, aggressive sampling efforts over long periods would enable better resolution of preferential tracks and pulsed flow patterns, but also increase the cost and difficulty of such studies. This study represents an important step in the employment of actively-used septic systems for tr acer studies on septic tank impacts to ground water quality, and is the first known report of tracer studies us ing mounded septic drainfields.

PAGE 49

37 CHAPTER 2: A REVIEW OF FACTORS AFFECTING MICROBIAL SURVIVAL IN GROUND WATER Introduction As discussed in Chapter 1, two basic factors chiefly control the fate of potentially harmful microorganisms once in ground water: transport/retention and survival. Considerable effort has been put forth in defining factors which regulate retention and adsorption, particularly of viruses. A summary of many of these findings was presented in Chapter 1. Survival of potential pathogens in ground water is also important to understanding the duration and spatial extent of ground water contamination impacts. This type of knowledge is also important for enabling info rmed decisions about activities that will affect ground water microbiological quality, such as septic tanks, agricu lture lots, surface application of biosolids, or use of recharge or injection wells, such as Class V UIC wells that might impact underground sources of drinking water. Here it is important to draw distinctions between behavior of microorganisms in the vadose zone and behavior in the ground water (saturated) zone. In general, the vadose zone is considered to offer protection of underlying ground water regions from surface activities that could input pathogens to aquifers. Percolation through the vadose zone has been shown to be a significant factor in removing possible contaminants before ground water reaches the saturated zone of an aquifer (Gerba and Bitton 1984). This occurs mostly due to slower transport and bulk flow of ground water, greater retention, and faster inactivation of bacteria, viruses, and protozoa in unsaturated soil. However, information on survival of microbes once in the saturated zone is particularly important in areas with shallow aquifers or in situations where possibly cont aminated surface water may come in direct contact with the aquifer. Also, areas with hi gh annual or seasonal rainfall may experience situations where more rapid transport of surface organisms to aquife rs occurs due to greater advective flow of water. Florida is one such region with high seasonal precip itation, limited vertical topography, and therefore in

PAGE 50

38 many places shallow water tables/small vadose zones. The karst geology of the Florida peninsula is also a contributor to more rapid transport of surface water to a quifers than most other areas of the continental U.S. Most studies on survival of public-health-relate d microorganisms in ground water have considered inactivation of viruses, as these organisms are often considered the most readily transported through the subsurface and most threatening to ground water supplies. But, given the karst geology of Florida, with associated solution channels and sometimes relatively high bulk porosity, larger organisms such as bacteria and intestinal parasite cysts and oocysts are of equal co ncern. Thus, it is important to examine the fate or survival of all groups of microorganisms in ground water, in the bulk liquid phase of the ground water environment. A review by Hurst (Hurst 1988) compiled data from other published reports on factors that influenced survival or inactivation rates of ente roviruses and rotaviruses in surface fresh waters. Quantitative data such as rates were not presented, a lthough a figure of temperature effects was given. However, the temperature effect for data summarized by Hurst varies and no average is given. The author believes that for these studies there are likely ot her factors beyond those analyzed accounting for differences in temperature effects. Factors which were determined to have a statistically significant effect on waterborne virus survival include: chloride concentration over the range of < 0.5 to 16.3 mg/l pH over the range of 6.0 to 7.8 total organic carbon from <1 to 17 mg/l hardness from 29 to 339 mg CaCO3 temperature from 4 37 C turbidity from <2.5 to 36 NTU. Sunlight also has a significant effect on survival but this is not a consideration in ground water survival, except to say there is a lack of it which would allow long er survival than viral particles in surface water. This introduction to survival of microorganisms in ground water seeks to summarize the current state of knowledge on many organisms of concern from a quantitative perspective. Since no standard exists for reporting results of studies on microbial inactivation, data have been reported by various authors in many different ways. The information presented herein summarizes methods and findings of studies on

PAGE 51

39 microbial survival in ground water and in some cas es surface water, and collates these findings into expressions of inactivation rates in terms of log10 decline in the viable or culturable organisms per day. In many cases, authors have reported inactivation data in th ese terms, in other cases the data as presented were converted to log/day declines based on times to achieve a given level of reduction or approximated from graphical data. In general, rates converted from grap hical data express an average rate resulting from the total decline in viable counts observed over the study period. Individual survival studies are grouped here based on the organisms that were evaluated. Lastly, analyses of data extracted from reviewed research are presented that combine inactivation rates for the organisms studied to express ranges and other summary statistics. Studies on Viruses Studies published by Yates and others in 1985 and 1990 have reported the effect of numerous parameters on virus survival in ground water. An an alysis of the effect of ch emical and physical factors, namely total dissolved solids (TDS), hardness, turbidity, pH, and nitrate concentrations, on virus survival in ground water was reported by Yates, et al. (Yates 1985). All ground water samples were analyzed in their natural state without treatment. The authors reported inactivation rates which ranged from 0.035 to 0.676 log/d for poliovirus, 0.051 to 0.628 log/d for echovirus, and 0.012 to 0.325 log/d for the coliphage MS-2. Multiple regression analyses of the data by the authors revealed that incubation temperature was the only factor significantly correlated to inactivation rate (P =0.05) of all viruses while calcium hardness was also correlated to the decay rate of MS -2, with increasing calcium concentra tions correlating to increased MS-2 inactivation. TDS, which ranged from 37 to 1,110 mg /l, was not found to significantly affect inactivation rates. A related paper by Yates and Gerba (Yates and Gerba 1985) incorporat ed a comparison of the impact of indigenous ground water bacteria on MS-2 survival in ground water. The addition of filter sterilized vs. raw ground water as a parameter did not change the analysis of significant factors for MS-2 survival from that described in the previous paper. Still, temperature and calcium hardness were the only

PAGE 52

40 factors significantly correlated to inactivation rates for MS-2. Approximated inactivation rates ranged from 0.028 to 0.167 log/d. To comprehensively compare the effect of indigenous bacteria on survival of introduced viruses in ground water, and evaluate the eff ect of numerous physical, chemical and microbiological factors on viral persistence, a third study by Yates and others (Yat es 1990) compared multiple samples with differing environmental variables in microcosm survival studies. Essentially, two lines of experimentation were reported. In one, a total of nine ground water samples from four separate states were incubated at the natural water temperature of the aquifer, which ranged from 12 to 23 C. Survival of seeded viruses was evaluated at each temperature while simultaneously comparing duplicates of each ground water sample after filtering through a 0.22 m filter to remove bacteria. In a ddition, 19 samples of well water from the Tucson, AZ basin were compared in light of numer ous parameters measured for each sample, including pH, turbidity, sulfate, nitrate, ammonia, magnesium hardness, iron, calcium hardness, total hardness, TDS, and heterotrophic bacteria. Statistical correlations were performed by the authors in order to establish if any factors could be significantly asso ciated to trends in inactivation ra tes. Temperature was again the only factor to consistently correlate to inactivation, with faster inactivation at higher temperatures. Subsequent studies using the same Tucson-area ground water samples to evaluate inactivation along with changes in bacterial population densities over the experimental time frame did reveal that MS-2 reduction was significantly correlated w ith an increase in bacteria l numbers. However, th e presence or absence of bacteria (raw vs. filtered) was not found to significantly affect decay rates of either MS-2 or poliovirus. This generalization applied when considering all wate r samples examined. But large variations in decay rates did exist between samples incubated at the same temperatures; for some inactivation was more rapid in unfiltered water while for others it was more rapid in filtered waters. In others still, no significant difference existed between filtered and unfiltered waters. The authors conclude that the lack of a consistent trend for the factors examined in all samples except for temperature may thus indicate interactions exist which could vary considerably between different water sources, and may make drawing generalizations for virus inactivation in ground water prohibitively difficult. Survival of hepatitis A virus (HAV), poliovirus, an d echovirus in ground water was evaluated with respect to the effect of temperature, aquifer substrate (soil), and presence of autochthonous microorganisms

PAGE 53

41 by Sobsey et al. (Sobsey 1986). Virus survival was evaluated in ground water alone or in ground water with one of several soil substrates suspended in it. Each trial was done at 5 and 25 C. In many cases, less than 2 log inactivation was observed, and no data for experiments at 5 C are reported as little inactivation was observed for the ground water or soil suspensions, regardless of th e presence of microbes. At 25 C, approximate inactivation rates ranged from < 0.024 to 0.095 for hepatitis A virus, 0.032 to 0.095 for poliovirus, and 0.032 to 0.071 for echovirus, all in log/d. Inferences drawn from these data by the authors are that HAV appeared to survive longer in soil su spension than echovirus 1 and perhaps poliovirus 1 at 25 C, regardless of soil type, while all viruses survived well at 5 C. It is also worth noting that among the sterile/non-sterile pair comparisons, the non-sterile replicate resulted in more rapid inactivation for 4 out of 6 HAV experiments, 1 out of 2 poliovirus experiments, and 2 out of 2 echovirus experiments, or 7 out of 10 pairs total showed more rapid inactivation in the non-s terile water or soil-water suspension. The authors also concluded that HAV was affected to a lesser extent than the other viruses by temperature and the presence of native microbes in the ground water and/or ground water-soil suspensions, thus poliovirus 1 and echovirus 1 are not effective indicators for pr edicting the survival of hepatitis A virus. A brief study by Yahya, et al. (Yahya 1993) evaluated inactiva tion of the bacteriophages MS-2 and PRD-1 in four different ground water samples, in cubated at the ambient temperature of the aquifer for each sample. Ground water samples came from Arizon a (3) and Canada. No parameters of ground water samples were reported and no mention was made of any tr eatment to the ground water such as filtration, so the water samples were assumed to be in a raw state. Inactivation rates estimated from this study ranged from no decline at 7 C to about 0.325 log/d at 23 C for MS-2 and no decline at 7 C to 0.12 log/d at 23 C for PRD-1. Conclusions derived from this study are that little difference in in activation rates between MS2 and PRD-1 was observed at lower temperatures, while elevated temperatures increased the inactivation rate of MS-2 much more than PRD-1. Alvarez, et al. (Alvarez 2000) evaluated inactivation of MS-2 and poliovirus in ground water samples that were either filtered (0.22 m) or used raw at 27 C. In general, inactivation did not follow first order kinetics, so the rates extrapolated from this study express average approximate rates of inactivation over the total duration of the experiment or until <1 pfu/ml was detected. Approximate inactivation rates for MS-2 in ground water were either 0.78 log/d or around 2.1-2.5 log/d depending on the

PAGE 54

42 virus seed preparation, with purifi ed MS-2 suspensions declining faster than unpurified cell lysates. Poliovirus suspensions were purified and ground water inactivation rates were approximately 1.4-1.7 log/d. For each virus, inactivation in the filtered gro und water subsample was slightly faster. The effect of hydrostatic pressure on poliovirus-1 survival was evaluated by Bitton et al. (Bitton 1983) using ground water and seawater. Ground water survival experiments were stored for 24 hours at 24 C (75 F) at initial pressures of atmospheric pressure (control), 500, 1000, 2000, 3000, and 4000 psi (range of 34-272 atm). Little effect was observed in the ground water samples as a result of pressure. Survival of poliovirus-1 ranged from 82.5% of the control at 3000 psi to 100% of the control at 4000 psi. A significant effect was observed in seawater samples stored at 2 C at 1000 psi for up to 24 hours. In that instance, only 15.6% of the control virus co ncentration was surviving at 24 hours. Jansons et al. evaluated several types of viruses for survival in dialysis tube devices while suspended in several bore holes containing ground water (Jansons 1989). The study site in this research was influenced by artificial recharge of the aquifer with wastewater effluent. Viruses evaluated were coxsackievirus B5, echovirus 6, 11, and 24, and poliovirus 1. Survival of viruses in 7 boreholes was evaluated, using a single virus type in each bore hole. Poliovirus 1 was used in three bore holes and the rest of the viruses in one each. The pl ume of recharge water created a gr adient of dissolved oxygen and temperature, such that bore holes more influenced by the effluent had a higher DO and lower temperature than the native aquifer water. This gradient allowed a comparison of poliovirus survival in response to the three different DO concentrations. It was found that inactivation was greater in the bore hole with higher mean DO concentrations, such that inactivation rates we re -0.09 log / d in a mean of 5.4 mg/l DO and -0.03 log / d in water with a mean DO of 0.2 mg/l. The te mperature was on average the same in these holes at 15.7 and 15.9 C. No other direct comparisons of the same organisms could be made in the ground water, although in sterile PBS the authors also found higher temperature to increase the inactivation rate of poliovirus 1. The authors also speculate that microbial activity in the poliovirus 1 dialysis tube at higher DO concentration could have led to more-rapid inactivation, due to the detection of high numbers of Pseudomonas maltophila in samples from this microcosm which were not detected in the other bore holes. Inactivation rates for other viruses in this study were -0.11, -0.10, and -0.05 for echovirus type 6, 11, and 24 respectively, -0.05 for coxsackievirus, and -0.07 for poliovirus 1 (all log / d) in a bore hole with mean

PAGE 55

43 DO concentration of 0.06 mg/l and 21.7 C. These rates were included in summary data analysis for reviewed studies later in this chapter. Studies on Viruses and Bacteria Keswick, et al. (Keswick 1982) evaluated survival of several indicator organisms and animal viruses in-situ in well water using polycarbonate membrane survival chambers which allowed exchange of water and dissolved compounds while retaining test organisms within the chamber and excluding autochthonous microbes. Organisms evaluated were coxsackievirus B3, poliovirus 1, echovirus 7, rotavirus SA-11, f2 bacteriophage, E. coli S. typhimurium and fecal streptococci (e nterococci). The water temperature over the 24-day duration of the experiment varied from 3 to 15 C. Inactivation rates were as follows, in log decline per day: coxsackievirus B3 0.19 poliovirus 1 0.21 fecal streptoc occi 0.23 E. coli 0.32 rotavirus SA-11 0.36 bacteriophage f2 0.39. Bitton et al. (Bitton 1983) evaluated survival of a number of indicator organisms and pathogens in a single Florida ground water source. The die-off of E. coli Streptococcus (Enterococcus) faecalis Salmonella typhimurium f2 bacteriophage, and poliovirus type 1 was evaluated in ground water microcosms. Seeded ground water flasks were incubated at 22 C for 15 days. In addition, a field study was performed in which samples were taken from 6 shallow monitoring wells tapping ground water underlying a cypress stand th at received primary septic tank effluent Septic discharge was halted due to excessively dry conditions, and sampling of the shallow wells was conducted with the cessation of discharge to evaluate the survival of fecal indicator bacteria that had in filtrated through th e cypress stand. In bench-scale laboratory studies, E. coli (0.16 log/d) and S. typhimurium (0.13 log/d) were much more rapidly inactivated in this ground water than were poliovirus 1 (0.046 log/d), but inactivation of S. faecalis

PAGE 56

44 (0.03 log/d) was approximately simila r to poliovirus 1. Also, inactivatio n of fecal (0.03 log/d) and total coliform (0.04 log/d) in the field study closely paralleled that of poliovirus 1 in the laboratory. The survival of numerous fecal indicator and pathogenic bacteria and coliphage as impacted by several factors in fresh and sea water was reported by Evison (Evison 1988). Although the present study is focused on ground water, some data from this paper were included since the authors quantitatively evaluate survival of a number of potential pathogens for which little survival data exists fr om ground water studies. The following organisms were used in microcosm batch studies: E. coli fecal streptococci, maroon fecal streptococci, Salmonella typhimurium strains 12, 12a and 110, Sal. anatum Shigella sonnei Sh. flexneri Yersinia enterocolitica Campylobacter fetus MS-2 bacteriophage, and f2 bacteriophage. Conditions evaluated for impacts on inactivation were temperat ure, salinity, nutrient amendment using sterilized sewage, and dark vs. light intensity. Only those studies performed under dark conditions with fresh water and no sewage amendment were included in this discussion. Rates for individual organisms from this study may be summarized as follows (converted to approximate log decline per day): for bacteria, E. coli rates ranged from 0.033 (2) to 0.35 (25), fecal streptococci fr om 0.04 (2) to 0.43 (20), S. typhimurium from 0.026 (2) to 0.14 (20), S. anatum from 0.025 (10) to 0.212 (2), Sh. Sonnei from 0.081 (15) to 0.42 (2), Y. enterocolitica from -0.22 (15) to 0.038 (5), and C. fetus from 0.16 (5) to 0.089 (25). The range of bacteriophage rates were, for MS-2, 0.02 (10) to 0.088 (20) and for f2 0.1 (5) to 1.6 (25). Regarding the response to temperatures, the author states that for most organisms a linear relationship existed between temperature and inactivation, indicating more rapid in activation at increasing temperatures. Growth was frequently observed for Y. enterocolitica notably at low salinity values at 15 C, and in fresh water at 15, 20, and 25 C. Overall, the author concluded that E. coli is an adequate indicator for the presence of culturable pathogens in freshwater, but the fecal st reptococci (enterococci) are a better indicator for seawater. Nasser and Oman (Nasser and Oman 1999) examined temperature effects on several organisms in ground water. Organisms evaluated were hepatitis A virus (HAV), male specific bacteriophage (F+ phage), E. coli and poliovirus 1. It is interesting that inactivation of E. coli was most rapid at 4 C. Otherwise, survival was negatively impacted by temperature based on comparative observation of figures in this study. Graphically-depicted inactivation rates of the four organisms were generally in the range of 0.01 to 0.05 log

PAGE 57

45 / d, although low poliovirus and hepatitis A virus inactivation rates of about 0.005 and 0.001 log / d respectively were reported at 4 10 C. Dowd and Pillai (Dowd and Pillai 1997) evaluated survival of two bacteria and two bacteriophage in ground water microcosms. Experiments were run with Salmonella typhimurium and a Klebsiella species, and MS-2 and PRD-1. Survival microcosms were incubated at 21 C for up to 32 days. Bacterial inactivation rates were about 0.094 log / d for Klebsiella and 0.52 log / d for S. typhimurium while bacteriophage inactivation was stated as approximately 0.8 log/d for both phage strains. Studies on Bacteria and Cryptosporidium Survival of several pathogenic and indicator/facultatively pathogenic bacteria in a single ground water source from Germany was evaluated by Filip, et al. (Filip 1988). Experiments were performed with microcosms using filter sterilized water samples held at 10 C, the ambient temperature of the source aquifer. Organisms analyzed for survival were as follows: Escherichia coli Salmonella typhimurium Pseudomonas aeruginosa Yersinia enterocolitica Staphylococcus aureus Streptococcus faecalis Bacillus cereus Bacillus megaterium and Clostridium perfringens Inactivation rates were approximately 0.03 for E. coli 0.012 for Str. faecalis 0.04 for Sa. typhimurium 0.2 for Sta. aureus 0.008 for Y. enterocolitica and 0.55 for B. megaterium From the authors’ original graphs, B. cereus declined about 3.6 log in 10 days, but showed no further decline to 100 days. C. perfringens showed little to no decline over 100 days, and P. aeruginosa increased in concentration to 11 days, and ther eafter declined by approximately 1 log to 100 days. Since these declines were not even approximat ely first-order, they were not included in summary inactivation rate data for this review. The impact of indigenous microbiota on survival of Aeromonas hydrophila in freshwater microcosms was evaluated by Kersters, et al. (Kersters 1996). Survival experiments were conducted at room temperature and experimental durations extended for up to 15 days. Inactivation rates ranged from 0.03 to 0.97 log / d. Significant differences were observed between sterilized samples and raw samples,

PAGE 58

46 with inactivation proceeding more rapi dly in the raw samples on average. In addition, the authors reported interaction of water source and ster ility was significant. There was a la rger increase in inactivation rates between sterile and raw water for surface water than for ground water. No studies which specifically evaluated survival of Cryptosporidium parvum oocysts in ground water were located; however, two articles wh ich employed surface water were reviewed. C. parvum inactivation in river water and in response to numerous environmen tal pressures including freezing, drying, and metal exposure was assessed by Robertson, et al. (Robertson 1992). To determine oocyst survival, samples were taken on various days up to 176 d, and the percentage of oocysts exhibiting PI exclusion and DAPI inclusion were determined. In order to interpre t these data for inclusion in this review, it was noted that the percentage of DAPI+ oocysts declined from 74.4% at day 0 to 10.0% at day 176. That equates to an N/N0 ratio of approximately 0.134, thus log N/N0 = -0.87. Thus for expressing this decline as a rate as elsewhere in this review, the inactivation rate is approximately 0.005 log / d. Medema et al. (Medema 1997) determined inactivation kinetics of several indicator species and C. parvum in river water that was either sterilized or raw at two temperatures (5 and 15 C). C. parvum inactivation was on the order of 0.01 to 0.02 log decline/ d, but within that range exhibited no difference between inactivation in sterilized or non-sterile water at 5 C and declined more rapidly in non-sterile water at 15 C. For E. coli and Ent. faecium inactivation was generally slower in autoclaved water than nonsterile water, with inactivation rates ranging from approximately 0.01 to 0.2 log / d. E. coli increased in titer in autoclaved water at 15 C then remained constant for the du ration of the experiment. Decay of Ent. faecium was slower at 15 C than in 5 C water. C. perfringens was 3-4 times more persistent in raw water than Cryptosporidium oocysts, but this trend was reversed in autoclaved water. Although inactivation of these organisms was evaluated in surface water, rather than a ground water source, the author’s inactivation rates were included in summary statistics since light as a parameter was not included. The effect of temperature and salinity on Cryptosporidium parvum infectivity was evaluated by Freire-Santos, et al. (Freire-Santos 1999) using mouse infectivity analyses. Conditions evaluated in a multi-variate experiment were storage times of 2, 21, and 40 days, temperatures of 4 11 and 18 C, and salinity of 0, 17 and 35 ppt. The effect of these f actors was modeled into an infection intensity function, which revealed maximum infection potential with oocysts stored in the lowest salinity and for the shortest

PAGE 59

47 time. Oocysts stored for the longest time and at the highest salinity produced the lowest infection density. These data are of limited use for the context of th is review since they do not express a measure of inactivation, only of effects on infection intensity as determined by evaluation of Cryptosporidium abundance in mouse intestinal tissue after inoculation. Therefore, a quantitative estimate of inactivation rates or a similar statistic was not possible. Summary and Conclusions To summarize inactivation rate data for these organisms, most rates as reported in or approximated from reviewed studies described above were compiled for each type of organism. From these compilations, some statistics were determined in order to gauge po ssible trends from pooled data. To facilitate data analyses, organisms were grouped into several categories: coliphage, poliovirus, echovirus, hepatitis A virus, PRD-1 bacterioph age, coliform bacteria (total an d fecal), enterococci/streptococci, Salmonella spp., Shigella spp., Clostridium perfringens Yersinia enterocolitica Aeromonas hydrophila and Cryptosporidium parvum All the rates compared were converted to log N/N0 inactivation values if necessary. Since most survival studies report significant effects due to temperature, rates were grouped for each organism into temperature ra nges. Table 9 and Table 10 cont ain temperature-grouped inactivation rates for each organism group, showing mean rates and standard deviations, medi an rates for temperature groups with 4 or more observations, and ranges. The temperature ranges were chosen to be 0 10 11 15 16 20 21 25, and 26 30 C, although for some these groupings were altered if data were sparse. Inactivation rate values as grouped in Table 9 and Table 10 for organism categories with at least 10 observations (along with Shigella spp. with n=6) were used to construct graphs showing the mean inactivation rates (log/day) for each organism by temperature group (Figure 5 and Figure 6).

PAGE 60

48 Table 9. Bacteria inactivation rates from rev iewed studies, compiled by temperature group. Organism Temperature Group (o C) Mean rate (log/d) Median rate (log/d) Std. Dev. (log/d) Range (log/d) n (values) 0 100.05140.040.0320.01 0.1027 11 150.0860.0750.078-0.008 0.2034 16 200.1180.1450.015 0.222 21 250.2010.1330.094 0.353 26 300.0351 Enterococci/0 100.080.0760.06280.012 0.219 11 150.0970.0710.0840.005 .2335 16 200.3950.04950.36 0.432 21 250.240.1830.029 0.363 Salmonella spp. 0 100.0580.0330.0630.025 0.2127 11 150.0940.0060.088 .102 16 200.1450.0050.14 0.152 21 250.2340.1450.1640.13 0.5174 Shigella spp. 0 100.2770.1050.17 0.423 11 150.0810.0811 20 250.190.050.14 0.242 0 100.005 0.006250 0.0123 11 150.0160.01560.005 0.0272 0 100.0230.0150.008 0.0383 15 25-0.1370.0802-0.163 Vibrio cholerae 110.31 Aeromonas hydrophila 220.3940.4350.3220.03 0.978 0 100.010.0100.01 0.014 11 150.01480.01450.007890.006 0.0244 Cryptosporidium parvum Clostridium perfringens Coliform bacteria Fecal streptococci Yersinia enterocolitica Table 10. Virus inactivation rates from reviewed studies, by temperature range. Organism Temperature Group (o C) Mean rate (log/d) Median rate (log/d) Std. Dev. (log/d) Range (log/d) n (values) Poliovirus0 100.00750.00350.005 0.012 11 150.08680.0720.05060.026 0.18519 16 200.0970.0810.0540.03 0.1859 21 250.2670.0830.2840.032 0.67610 26 301.030.8570.055 1.673 hepatitis A0 100.00550.00550.0060.001 0.012 20 250.05570.0360.04480.015 0.147 26 300.03750.003540.035 0.042 echovirus11 150.1070.0790.05790.051 0.1867 16 200.0980.0960.0410.05 0.1514 21 250.1690.0710.2080.057 0.6287 coxsackievirus3 150.191 190.051 rotavirus3 150.361 coliphage0 100.0290.020.02640 0.113 11 150.0970.060.09770.028 0.431 16 200.1430.0810.1890.02 0.639 21 250.4260.3240.3640.048 1.41612 26 301.2420.781.0350.022 2.55 PRD-10 100.0190.02690 0.0382 21 250.3240.4140.052 0.83

PAGE 61

49 Figure 5. Bacteria inactivation rates, average d by temperature range. Error bars represent standard deviation. 0 0.1 0.2 0.3 0.4 0.5 0 1011 1516 2021 2526 30Temperature groups (o C)Mean inactivation rates (log10/d) Coliform bacteria Enterococci Salmonella spp. Shigella spp. Figure 6. Virus inactivation rates averaged by temperature range. E rror bars are standard deviations. 0 0.5 1 1.5 20 1011 1516 2021 2526-30Temperature groups (o C)Mean inactivation rates (log10/d) Poliovirus Hepatitis A Echovirus Coliphage

PAGE 62

50 For coliform bacteria, including E. coli Klebsiella spp., and non-specific total and f ecal coliform results, inactivation rates from 7 studies were compar ed, in which the temperature ranged from 2 30 C, for a total of 20 rate values. The response of rate s in relation to temperature ranges for coliform bacteria showed that inactivation increased w ith temperature up to 25 C. From Table 9 and Figure 5, the slowest mean rate was observed at higher te mperatures, with the next slowest being at the lowest temperatures considered. However, only one valu e represents the inactivation rate at 26 30 C. A scatterplot and regression of inactivation rates fr om coliform bacteria studies was constructed (Figure 7). The linear regression shows a general upward trend for inactiv ation rates with temperature, but a very low r2 value of 0.185. This may be related to differences in experimental procedures and methods for rate calculation and reporting of data, compounded by few observations in the higher temperature ranges, or even an indication of growth of the bacteria at higher temperatures. If inactivation rates for colifor m bacteria are considered in light of whether the water was sterile (including filtered water) or non-sterile, without regards to temperature, mean rates are 0.088 log/day in sterile water vs. 0.078 log/day in non-sterile water. These averages also include inactivation rates determined from studies using sterile buffered saline. Figure 7. Scatterplot of coliform bacteria inac tivation rates compiled from reviewed studies. Equation is a linear model. y = 0.0045x + 0.0401 R2 = 0.18520 0.1 0.2 0.3 0.4 05101520253035Temperature (o C)Inactivation rate (log10/d)

PAGE 63

51 Five reviewed studies included evaluations of en terococci and/or fecal streptococci for a total of 21 observations; temperatures in these studies ranged from 2 to 25 C (Table 9). Some studies employed isolates such as Enterococcus faecalis, Streptococcus equines or Streptococcus bovis while most values were derived from studies evaluating non-specific enterococci. Wide ranges of values were observed, with three out of four temperature groups having ranges of greater than an order of magnitude. Averaged inactivation rates from these groups show that above 15 C, inactivation rates were greater on average. However, the mean rate declined fo r temperatures between 21 25 C co mpared to rates from 15 20 C, with greater variability at 21 25 C (Figure 5). Also there were fewer observations at higher temps than at 15 C and less. A regression of temperature vs. inactivation rates for enterococci (Figure 8) shows an overall positive in the data with greater rates of inactivation at higher temperatures, and an r2 value of 0.383, about twice that of the coliform regression. If experiments performed in sterile water (16 rate values) are compared to those in nonsterile water (5 values), mean inactivation rates are 0.161 log / day in sterile vs. 0.077 log / day in non-sterile experiments. Figure 8. Scatterplot and regressi on of reviewed enterococci inac tivation rates with respect to temperature. y = 0.0113x 0.0005 R2 = 0.38260 0.1 0.2 0.3 0.4 0.5051015202530 Temperature (o C)Inactivation rate (log10/d)

PAGE 64

52 Salmonella species bacteria were used in 4 studies reviewed here, with 15 observations being used to summarize inactivation rates. Temperatures for Salmonella experiments ranged from 2 to 25 C. Mean inactivation rates increase with in creasing temperature ranges. Little difference was observed between mean inactivation rates from sterile (0.370 log/day) and non-sterile (0.324 log/day) conditions. Other bacterial groups were evaluated in fewer reviewed studies, such as Shigella spp. (n=6 observations), Clostridium perfringens (n=5 observations), Yersinia enterocolitica (n=6 observations), and Aeromonas hydrophila (n=8 observations). Estimated inactivation rate s from these organisms are summarized in Table 9, subdivided into temperature ranges. The limited number of observations of these organisms makes drawing general conclusions difficult. Cryptosporidium parvum survival in water was evaluated by three studies reviewed here, but it was only possible to estimate quantitative inactivation rates from two of those. The mean inactivation rate in studies performed at 5 – 15 C was 0.0116 log/day, and the median value was 0.01 log/day. Estimated or reported inactivation rates ranged from 0.005 – 0.024 log/day. When considering virus survival studies, more data were available for poliovirus and coliphage than for any of the bacterial groups (Table 10). Some type of coliphage were evaluated in 10 studies reviewed here, for a total of 72 observations. Poliovirus were included in 8 studies for 43 observations (once again, not including sterile buffered saline or sterile de-ionized water conditions). The temperature ranges covered by data for both viruses was 4 30 C. From Table 10, a more consistent effect of temperature can be observed with the viruses than for bacteria. When looking at temperature groupings for poliovirus and coliphage, a consistent increase was observed in both the median and mean inactivation rates for each temperature category as temperature increased. Column gra phs of mean inactivation rates for each temperature group (Figure 6) also demonstrate this trend. Scatterp lots of the reported and estimated rates for coliphage and poliovirus against temperature also reveal generally great er inactivation rates at higher temperatures (Figure 9 and Figure 10). The optimal regression model for coliphage and poliovirus temperature effects was exponential, rather than a linear equation, based on maximum r2 values for each. The equations and r2 values are shown on Figure 9 and Figure 10. R2 values are both below 50%, and it visually appears that the curve does not fit greater inactivation rates above 25 C as well.

PAGE 65

53 Figure 9. Coliphage inactivation rates from rev iewed studies with respect to temperature. Exponential regression model result is shown. y = 0.0135e0.1275xR2 = 0.43980 0.5 1 1.5 2 2.5 3 05101520253035Temperature (o C)Inactivation rate (log10/d) Figure 10. Poliovirus inactivation rates vs. te mperature with exponential regression model. y = 0.0137e0.0988xR2 = 0.24130 0.5 1 1.5 2 2.5 3 05101520253035Temperature (o C)Inactivation rate (log10/d)

PAGE 66

54 Hepatitis A and echovirus mean inactivation rates also increase at higher temperatures, with mean inactivation being faster for hepatitis A above 20 C than below 10 C, while echovirus inactivation increases slightly between the three temperature brackets as shown in Table 10. Figure 6 also shows the greater inactivation rates for echovirus and hepatitis A at temperatures above 10 C. However, the increase is not as dramatic as for poliovirus and coliphage. One point, although, is that the total number of rate values represented by the means for echovirus an d hepatitis A are much fewer than for poliovirus 1 and the coliphage. Thus the significance of these inactivation rates being much slower than poliovirus or coliphage, which are more commonly used as indicators of virus presence, is uncertain for ground water. Clearly, a closer look should be taken at the persistence of hepatitis A and echovirus under these conditions to determine if indeed they are more persistent th an coliphage or enteroviruses such as polio 1. Inactivation rates compiled from these various studi es were separated into sterile vs. non-sterile water to determine if a consistent trend regarding th is was observed for viruses. For coliphage, mean inactivation rates were 0.278 log/day in sterile water (n =32) and 0.205 log/day in non-sterile water (n=46). Poliovirus means were 0.196 log/day in sterile water (n=19) and 0.166 log/day (n=31) in non-sterile water, while hepatitis A mean rates were 0.0228 log/day in st erile water (n=6) and 0.0455 log/day in non-sterile (n=8) water. For echovirus studies, only three observa tions were recorded for sterile water, but the mean rates were 0.058 log/day (n=3) and 0.136 log/day in non-sterile water (n=16). Given these comparisons, there does not appear to be an obvious effect when co nsidering sterile against non-sterile water for all rates reported or estimated for each virus group. Poliovirus and coliphage rates were slightly slower in nonsterile water, while the opposite was true of hepatitis A and echovirus. However, these comparisons were drawn regardless of temperature, and thus temperature differences between the two respective sets of data (sterile or raw) for each organism may obscure general trends related to background bacteria levels. Part of the purpose for this review was to analyze a large body of published data to elucidate possible trends in inactivation rates in response to environmental variables. For the bacteria, increasing temperature generally increases inactivation rates, alth ough at higher temperatur es this effect is not consistent. However, fewer data were available at these higher temperat ures. The regression analyses do indicate an overall increase in inactiv ation at higher temperatures for coliform and enterococci. For some viruses such as polioviruses and coliphage, temperature e ffects were very apparent. Likewise, the effect of

PAGE 67

55 indigenous microbiota (as in considering sterilized vs. raw water) does not appear to result in consistently faster inactivation rates. However, these experiment s were performed under different temperature regimes, thus it may be important to consider comparisons performed under the same temperature conditions to evaluate this possible effect more thoroughly. It is also important to consider that the inactivation rates used for these analyses were derived from many independently-performed experiments, some which directly calculated rates in these terms and others which did not. The data from these reports were originally presented in disparate ways, and the number of observations for a given condition may not be large enough to draw conclusions. In addition, many variables may come into play in bench-scale experiments, particularly the source and handling of organisms. For instance, some studies involved populations of bacteria or viruses derived from natural sources such as wastewater or animal feces, while others utilized pure strains maintained in laboratory conditions for many generations. Treatment of organisms prior to seeding survival experiments also varied, such as propagation and purification procedures. Evaluating the impact of these protocol variations among many studies is difficult. However, some individual studies evaluated the impact of various parameters within more or less controlled conditions and the findings of these types of studies may reveal more on possible trends. Table 11 summarizes trends regarding inactivation rates for various organisms that were observed in several studies reviewed here.

PAGE 68

56 Table 11. Trends in inactivation rates possibly due to various factors in reviewed studies. OrganismConditions contrastedReference MS-2 (coliphage)increasedw / increasing temp.Yates '85 MS-2 (coliphage)increasedw / increasing temp.Yates & Gerba '85 MS-2 (coliphage)increasedw / increasing temp.Yates '90 poliovirus 1increasedw / increasing temp.Yates '85 poliovirus 1increasedw / increasing temp.Yates & Gerba '85 poliovirus 1increasedw / increasing temp.Yates '90 poliovirus 1increasedat 25 vs 5 CSobsey '86 poliovirus 1increasedw / increasing temp.Nasser '99 echovirusincreased at 25 vs 5 CSobsey '86 hepatitis Aincreasedat 25 vs 5 CSobsey '86 hepatitis Aincreasedw / increasing temp.Nasser '99 Cryptosporidiumincreasedat 15 vs 5 CMedema '97 Enterococcus faeciumdecreasedat 15 vs 5 CMedema '97 poliovirus 1increasedat higher DO concentrationJansons MS-2 (coliphage)increasedw / increasing Ca hardnessYates '85 Cryptosporidiumincreasedin non-sterile vs sterileMedema '97 MS-2 (coliphage)decreasedin non-sterile vs sterileAlvarez '00 poliovirus 1decreasedin non-sterile vs sterileAlvarez '00 E. coliincreasedin non-sterile vs sterileMedema '97 poliovirus 1increasedin non-sterile vs sterileSobsey '86 echovirusincreasedin non-sterile vs sterileSobsey '86 hepatitis Aincreasedin non-sterile vs sterileSobsey '86 Aeromonas hydrophilaincreasedin non-sterile vs sterileKersters '96 Effect on inactivation rate As was observed by comparing inactivation rates compiled from many studies for the impact of temperature, several investigators observed that vi rus inactivation increases with increasing temperature within their respective studies (Yates and Gerba 19 85; Yates 1985; Sobsey 1986; Yates 1990; Nasser and Oman 1999); however, similar consistent trends for bacteria were not reported. Several studies also described an increase in inactivation rates in non-sterile vs. sterile water sources, but the opposite was also observed in some cases. Regarding the effect of salin ity or TDS, only one study directly evaluated TDS as a factor for viral inactivation (Yates 1985), and did not find significant differences over the range from 37 to 1100 mg/l. Unfortunately, a large proportion of reviewed studies did not include TDS as a reported parameter, which made analysis of TDS effects using data from many studies, as was done with temperature, difficult. One point of concern is the extrapolation of experimental results from bench-scale studies to insitu behavior of these many types of organisms. Of the studies reviewed here, only part of one evaluated in-situ decline of indicator organisms. Recalling the results from Bitton et al. inactivation rates of total

PAGE 69

57 coliform, fecal coliform, and fecal streptococci (e nterococci) were approx imately 0.02 0.03 log inactivation per day. In conclusion, studies following consistent experimental procedures need to be performed to hopefully reduce variability among investigators’ findings. Standard s for performing bench scale survival studies should include protocols for the propagation and preparation of seeded organisms, and should include controls such as ATCC strains of MS-2 and E. coli to preclude differences in the organisms themselves. In addition, more field studies are needed. While the introduction of potentially harmful microorganisms into the environment is generally opposed, innovative studies of ground water contamination by natural sources could prove helpful. If the proper safeguards could be ensured, controlled field studies involving seeded non-pathogenic microorganisms could prove even more beneficial if the results of such studies are expressed in quantitative terms and are published in pe er-reviewed literature to enable wide dissemination of this information.

PAGE 70

58 CHAPTER 3: SURVIVAL OF WATER QUALITY INDICATOR MICROORGANISMS IN THE GRO UND WATER ENVIRONMENT OF FLORIDA: TEMPERATURE AND TOTA L DISSOLVED SOLIDS EFFECTS Introduction One aspect of ground water micr oorganism survival that has recei ved little attention is the effect of variations in total dissolved solids on the inactivation of potential pathogens and their indicators. TDS of ground water in Florida can have a large range. For instance, in the Upper Floridan aquifer, TDS ranges from less than 100 to over 20,000 mg/l (Berndt 1998). The median for the Upper Floridan is about 300 mg/l, while the middle 50% of measurements are between about 200 to 700 mg/l. Dissolved solids concentrations generally increase in this aquifer towards the coasts and going south. The Upper Floridan in central coastal regions and south of Lake Okeechobee has TDS concentrations in excess of 1000 mg/l. Previous research has demonstrated that survival of the various microorganisms of concern to water quality is generally inversely related to increasing salinities in brackish and marine waters (i.e. 10 to above 30 ppt) (Fleisher 1991; Garcia-Lara 1991; Solic and Krstulovic 1992). In addition, temperatures in the deeper aquifers such as the Floridan aquife r system of central and south Florida are typically in the range of 20 30 C. Given the higher temperatures and salinity of ground water in many parts of the Floridan aquifer system, an evaluation of public-health related microbes under a range of conditions emulating these conditions is desirable. A greater proportion of published studies on survival of pathogens and indicators in ground water situations involve viruses, rather than the larger bacteria and protozoa. Thus, it is beneficial to continue to build the available body of knowledge on bacteria of conc ern. This is particularly important since bacterial indicators (i.e. total coliform bacteria) remain the stan dard for determining the quality of drinking water sources (USEPA 2002) and numerous bacterial illness ou tbreaks have been attributed to consumption of

PAGE 71

59 ground water (Craun 2002). Since little information exists on specific impacts of total dissolved solids for most indicator organisms, a direct comparison of this parameter would be beneficial. Also, many of the studies reviewed examine a particular strain of phage, enteric virus, or bacteria in their bench-scale work. While the use of specific strains is useful for compar isons to other research, these organisms are not the actual strains found in the environment. Thus, it is of interest to evaluate isolates from the environment, particularly of a composite population, to establish somewhat of a more realistic indicator assemblage. It may be that organisms which can be isolated from the environment have survived to the point of sampling and detection because they are more hardy variants or otherwise more fit than laboratory strains. The objective of research described in this chapter was to evaluate directly the effects of TDS and temperature on survival of several water-quality-indicator organism populations in bench-scale microcosm studies. These studies enabled an isolation of TDS and temperatur e effects, as well as interactive effects of the two parameters. The dissolved solids were provided by a mixed-ion salt which emulated seawater ionic composition and relative concentrations, and was mixed to concentrations ranging from 200 to 3000 mg/l. Temperatures evaluated were 5 C for a low-temperatur e control, and two higher temperatures of 22 and 30, which are more relevant to subtropical ground water environments such as the Floridan aquifer system. Study Methods Organism Populations Five types of microbes were employed for survival studies in TDS-temperature experiments. They were composite populations of fecal coliform and enterococci bacteria, composite populations of DNA coliphage (combined somatic and male-specific) and RNA coliphage (male-specific only), and the Salmonella bacteriophage PRD-1. Bacterial isolates were obtained from water samples of Bullfrog Creek, Hillsborough County and from storm water collected from Lake Jackson, Hillsborough County. Samples were initially assayed on mFC (APHA 1992) or mEI agar plates (US EPA Method 1600) using membrane filtration to select for fecal coliform or enterococci colonies respectively. Typical fecal colif orm (blue on mFC) or enterococci (gray with blue halo on mEI) colonies were picked an d streaked to isolate on non -selective tryptic soy agar

PAGE 72

60 (TSA) plates, which were incubated at 37 C for 24 hours. Isolates were then identified using the API identification system (Biomerieux, Hazelwood, MO) and 10 fecal coliform and 9 enterococci isolates were selected to comprise the populations for each bacterial type. The speci es as identified by API were, for fecal coliform, 8 Escherichia coli and 2 Klebsiella pneumoniae and for enterococci, 7 Enterococcus faecalis 1 Enterococcus faecium and 1 Enterococcus durans Isolates were propagated separately in tryptic soy broth (TSB) and frozen as a 50% mixture with dimethylsulfoxide for storage at –70 C. Male-specific (F+) RNA coliphage isolates were obtained from secondary wa stewater effluent at the Albert Whitted wastewater treatment facility in St. Petersburg. The effluent sample was taken by grab sampling directly from the secondary clarifier basin and aliquots of this sample were immediately plated by the double agar overlay method (described in EPA method 1602) on E. coli Famp (designation HS[pFamp]R, ATCC #700891) host cells to select for male-specific phage. Plaques were picked and isolated by additional plating and plaque selection using additional agar overlay procedures with Famp, then isolates were typed using the method described by Hsu (Hsu 1995) involving use of ribonuclease A (RNAse) to determine which were RNA phage. Ten is olates were chosen from these for the F+ RNA phage population, which were then propagated in TSB, filtered to remove host cells and debris, and stored as multiple aliquots of each isolate at -70 C. The DNA coliphage population was created from phage isolated from a water sample collected from Bullfrog Cr., Hillsborough County. Water sample aliquots were plated using the double agar overlay technique with E. coli strain C-3000 (ATCC #15597) which selects for both male-specific and somatic coliphage. Resulting plaques were then purified in the same way as for the RNA coliphage. An individual plaque from each isolate was again propagated in E. coli C-3000, filtered, and 10 isolates were stored as multiple aliquots at -70 C. These phage isolates we re also typed with RNAse, and determined to be all DNA phage (all grew in the presence of RNAse). Of the 10 isolates, 2 were male-specific and 8 were somatic phage, based on inability to grow on E. coli Famp cells. A pure stock of PRD-1 was obtained from the laboratory of Mark D. Sobsey (University of North Carolina, Chapel Hill, NC). This phage isolate was diluted and plated in an agar overlay with Salmonella typhimurium (designation LT2, ATCC #19585) to create plaques, then a single plaque was chosen to

PAGE 73

61 propagate in broth culture to further ensure purity. After filtration to remove host cells, aliquots of the suspension were stored frozen at -70 C. Organism preparation for survival experiments Prior to each experiment, each bacterial isolate wa s propagated separately from frozen cultures in 5 ml TSB, incubated overnight at 37 C. The following day, 0.1 ml of each isolate culture were combined into 50 ml of TSB, with fecal coliform and entero cocci being grown as sepa rate populations. These cultures were again incubated with shaking overnight at 37 C. Cells were washed in phosphate buffered saline solution (PBS) three times by successive centrifugation and re-suspension in PBS. Centrifugation parameters were 7 minutes at 2000 x g. The resulting viable titer after these procedures were consistent at about 8 x 108 cfu/ml for fecal coliform and 5 x 108 cfu/ml for enterococci. These rinsed cells were used for seeding into experimental vessels at the initiation of each trial. Coliphage isolates were grown in dividually from freezer stocks prior to each trial in cultures of respective bacterial strains ( E. coli Famp for F+ RNA phage and E. coli C-3000 for DNA phage), then purified by centrifugation and filtration through 0.22 m membrane filters. The titer of each isolate was determined independently, and for each trial since propagation results were not sufficiently consistent. Isolates for the respective populations of DNA or RNA coliphage were then combined in equal proportions (in terms of total pfu) into sterile reagent-grade wa ter for seeding into survival experiment vessels. Likewise, PRD-1 was grown from freezer stock and titere d for each trial. In a ll phage experiments, the DNA phage population and PRD-1 were combined together in each experimental bottle, and F+ RNA phage were tested in separate experiments. Also fecal coliform and enterococci populations were evaluated together in the same experimental vessels, since selective media was used to determine survival results. However, phage and bacteria were always evaluated separately. Thus, the experimental groups could be summarized as bacteria, DNA coliphage and PRD-1, and F+ RNA coliphage, with the members of each group being combined in re spective experimental bottles.

PAGE 74

62 TDS-temperature trials Solutions for evaluation of the effects of TDS and temperature on survival were created using Instant Ocean artificial sea salt (Aquarium Systems, Ment or, Ohio). The artificial seawater was utilized to provide a solution containing a mixture of ions as opposed to a solution composed only of sodium chloride or other single salt. Table 12 contains the composition of Instant Ocean solutions at 34 ppt (seawater salinity) as given by the manufacturer and the concen trations of these ions if the solution is at a concentration of 1000 mg/l (ppm) TDS. Table 12. Composition of mixed ion solution for TDS-temperature trials (Instant Ocean sea salt). Major components Chloride19251 Sodium 10757 Sulfate 2659 Magnesium 1317 Potassium 402 Calcium398 Carbonate/bicarb192 Strontium8.6 Boron5.6 Bromide65 Fluoride1 Iodide0.22 Lithium0.18 Trace components at 34 ppt Copper Trace (< 0.03)Arsenic Trace (< 0.0002) Iron Trace (< 0.03)Cadmium Trace (< 0.02) NickelTrace (< 0.04)Chromium Trace (< 0.0006) ZincTrace (< 0.02)AluminumTrace (< 0.04) ManganeseTrace (< 0.01)TinTrace MolybdenumTrace (< 0.0l)AntimonyTrace Cobalt Trace (< 0.05)RubidiumTrace Vanadium Trace (< 0.04)BariumTrace (< 0.05) Selenium Trace MercuryNone Lead Trace (< 0.005)NitrateNone PhosphateNone 1.912 0.029 0.006 0.005 11.706 5.647 0.253 0.165 316.382 78.206 38.735 11.824 Concentration at 34 ppt (mg/l) Concentration at 1000 ppm (mg/l) 566.206 Experiments were conducted in re-usable polypropylene bottles as reaction microcosms, either 250 ml or 100 ml in size. Prior to each trial run, each bottle was acid-washed w ith 10% HCl, rinsed with water and thoroughly washed with detergent, then triple-rinsed with tap water followed by de-ionized water. Bottles were air-dried then sterilized by autoclaving. Test solutions of the varying TDS concentrations were made by dissolving carefully weighed amounts of Instant Ocean in reagent-grade deionized water in glass bottles. E xperiments were performed using TDS concentrations of 200, 500, 1000, and 3000 mg/l. The pH was neutralized to between pH 6.5 – 7.5 using HCl or NaOH, and the actual conductivity/TDS was recorded. Solutions were then sterilized by autoclaving in the glass bottles. The

PAGE 75

63 bottles used had pressure-resistant closures, such that the caps were fastened tightly to prevent any concentration of TDS by evaporation. When pH was re-checked after autoclaving for initial representative solutions, it did not change due to autoclaving so in general pH was measured and adjusted prior to autoclaving only. In addition to the varying mixed ion concentration trials, phosphate buffered saline (PBS) microcosms were employed at each temperature to serve as control survival conditions for assessing overall variability between experimental sets. Thus fo r each experimental trial, a set of PBS bottles were also seeded with the same organism s as those evaluated in the respec tive experimental trials. PBS was made according to EPA Method 1623 an d sterilized by autoclaving, also in glass bottles. Immediately prior to each experimental trial, te st solutions were distributed to the sterilized bottles and solutions were placed at the respective temperature le vels. For the 5 C temperature, bot tles were placed in a foil-covered basket to exclude light and placed in a 5 C refrigerator. Bottles at 22 C were placed in a water bath at room temperature and covered to exclude light. The 30 C temperature was maintained in a water bath. The building climate control for indicat or experiments (at USF in St. Petersburg) was very stable such that the room temperature water bath remained constant at 22 C without adjustment. Each survival experiment was initiated by s eeding respective bottles of test solutions with organisms, using a quantity to achieve an approximate concentration of 2-3 x 105 cfu or pfu/ml. Enterococci and fecal coliform were s eeded to the same bottles, and were not used in the same bottle as bacteriophage. The time 0 sample was taken from each bottle immediately after seeding. Subsequent samples were taken on or about days 1, 2, 4, 7, 10, 14, 21, and 28. Survival of seeded bacteria and viruses was determined by evaluation of the concentration of culturable organi sms at each time point. Standard membrane filtration methods were em ployed for culturable counts for b acteria, using 0.45 m pore-size membrane filters to capture bacteria. Fecal colifor m were assayed using mFC agar, and enterococci were assayed using mE agar. Dilutions of water samples were performed using sterile PBS. Plates for fecal coliform were incubated at 44.5 for 24 hours, enterococci at 41.5 C for 48 hours. Bacteriophage counts were done using double agar overlay methods adapted from EPA method 1602 (Single Agar Layer Method, also describes double agar layer method). Phage were assayed for routine sample measurements using the same bacterial hosts as used for isolation of the population components. DNA coliphage were assayed using the E. coli C-3000 host, F+ RNA coliphage using the E. coli Famp host, and PRD-1 using Salmonella

PAGE 76

64 typhimurium LT2. Growth media for host was TSB, overlay agar was TSB with 1% agar, and bottom agar was TSA. Plates were incubated 24 hrs at 37 C. Positive controls for bacteria were E. coli for the fecal coliform, Enterococcus faecalis for enterococci, and MS-2 for the DNA and F+ RNA coliphage. Stock PRD-1 was used for positive controls on the PRD-1 assa y. Negative controls were used for bacteriophage hosts, to ensure host cells were not contaminated by phage. Negative controls for host involved plating 1 ml of host culture. Also, when performed simultaneously, host cultures of E. coli C-3000 and S. typhimurium LT2 were cross-checked with the opposing phage (MS-2 on S. typhimurium and PRD-1 on E. coli ) to ensure hosts were not cross-contaminated sinc e the two phage were combin ed in microcosms. For the bacteria, rinse buffer and dilution buffer (PBS) were checked to ensure that th ey were not contaminated with the target types of bacteria. Data Analysis The time series of data points comprised of log N/N0 ratios for individual data sets (data sets being each experimental trial for a given set of conditions) were analyzed i ndependently. The primary step of data analysis was fitting observed data to a regression equation. The purpose of fitting data to equations was to enable prediction of a number of days before 2 log (99%) declines in culturable/infective concentrations would be observed. The number of days for 2-log decline was further used as a statistic for comparative analysis of various fact ors on observed inactivation. To fit data for assessing inactivation behavior, observed ratios of surviving organisms were fit with an inactivation equation. The general equation used is shown here as Equation 1. Log N/N0 = k timem (1) This model incorporates an inactivation rate constant which describes the overall slope of the curve ( k ) and an exponent m which enables a better matching of inactivation kinetics. If m = 1, the curve is a straight line, indicating first-order kinetics, if m < 1, tailing is indicated, and if m > 1, a shouldering effect is indicated. The use of such a parameter was necessary since many of the data sets deviated from first-order kinetics, and a better fit to observed data was desired. All time values were in days. Observed data were fit to model equations by the use of least squares analysis to adjust the values of k and m to minimize the error sum of squares between observed log N/N0 and that predicted based on the model equation. Microsoft

PAGE 77

65 Excel was used to perform the iterative problem solving. In many cases, inactivation over the duration of a trial was too slight to allow the use of the additional exponent. The variability between time points overshadowed the general trend of decline and the mode l in Equation 1 was not able to fit observed data. Thus, in general, with data sets for which tota l decline was less than 1 log (90%), the exponent m was set to 1 and the model became first-order. In a few cases, kinetics over the course of the ex periment were such that an initial increase in viable counts was observed, followed by a decrease after some time. A model such as Equation 1 would not fit such a situation, and in thes e cases an alternative model was used in order to obtain a prediction of days before 2-log decline for further comparison. Equa tion 2 is a polynomial curve which when fit to data in these cases, has the shape of a shallow inverted parabola. Log N/N0 = k1 time2 + k2 time (2) However, this type of curve often extrapolated from observed data an assumed kinetic behavior which could not be observed (since actual time points only extended to approximately 28 days). The nature of this function is such that the slope increases with tim e resulting in the predicted increasing rate of change (dC/dt). If there was an observed decrease after an initia l increase in concentration, the curve fit this trend. However, if no decrease in concentra tion was observed up to the terminus of the trial run, this curve predicted an assumed a decrease at some future po int following kinetics that mirrored the initial increase and flattening of the observed data points trend. Th e use of this curve was only to enable a prediction for comparative analysis in situations where it was necessary, and it is just as reasonable to assume that observed kinetics fit to this type of function may be extrapolated in the same way as first-order kinetics were extrapolated to predicted inactivation pe riods beyond the experimental time frames. The predictive model with variab le component values fit for each data set was then used to estimate a number of days for 2-log decline with Microsoft Excel. Predicted days for 2-log decline were used as a statistic for analyses of variance (A NOVA) with Minitab rel.12. However, for PRD-1, inactivation over the 28-day experiments was slow to the point that frequently, very large numbers of days were predicted. Since these periods were well outside the experimental durations, and kinetics could be accurately modeled with first-order models in these par ticular cases, the statistic used for analysis was firstorder inactivation rate constants ( k in Eq. 1, with m at 1). ANOVA were used for each organism group to

PAGE 78

66 determine statistically significant factors accounting fo r variability in observed data. Factors compared for each organism were the TDS concentration and temper ature. Additional analyses were performed which examined only TDS concentrations equal to and less th an 1000 mg/l. These analyses were performed to statistically compare inactivation between various conditions and microbes. Results Fecal coliform bacteria Evaluation of fecal coliform su rvival in TDS-temperature micr ocosms indicated a significant effect of temperature on the predicted number of days until 2 log decline (p < 0.01). However, significant differences between TDS concentrations (for all temp erature levels combined) were not observed, nor was there a significant interaction of TD S concentration and temperature. Due to the large number of figures, plots of actual culturable counts fit to regression equations are shown in Appendix 1-A for visual reference. Examples of curves fit to Equation 2 can be seen in the data for fecal co liform bacteria in 3000 mg/l TDS at 22 and 30 C in Appendix 1-A. This type of m odel was used only in a very few cases with the TDStemperature trials (2) in order to estimate a value of days to 2-log decline for comparison to other conditions. Results from PBS cont rol microcosms for each organism are shown in Appendix 1 under the respective section for each organism. Model equations for each data set were used to predict the number of days for a 2-log decline in culturable concentrations. These model predictions were determined independently for each trial. Table 13 shows the predicted days for 2-log inactivation of fecal coliform at the various TDS concentrations and temperatures. In some cases, inactivation was so slight during the period of the experiment that the model for that data set predicted a very large number of days before a decline of 2 log. Since such long periods were well outside experimental durations, and it is r ealistic to assume a change in kinetics such that inactivation rates would increase after an extended period, any predicted period greater than 200 days is shown as “ > 200” and 200 was used for the value in statistical analyses of trends. This method is a compromise to capture the difference in order of magnitude between kinetics observed under various

PAGE 79

67 conditions while avoiding overly unrealistic values, some in excess of a year’s time based on 4 weeks of observed data. Likewise, the methods used could not accurately determine periods of time less than 1 day, so if a best-fit model equation predicted less than 1 day for 2-log inactivation, the value was rounded up to 1. Wide ranges of predicted values were determined for each temperature, the ranges were 19 over 200 days at 5 (both at 3000 mg/l), 4 to 126 days at 22, and 1 to 55 days at 30 (also at 3000 mg/l).

PAGE 80

68 Table 13. Days predicted for 2-log decline of fecal coliform at varying TDS and temperature conditions from model curve equations. TrialTDSTemperature12005101 219 376mean std. dev.65 42 12002234 24 319mean std. dev.19 15 12003046 23 39mean std. dev.19 23 15005143 271 386mean std. dev.100 38 15002226 218 327mean std. dev.24 5 15003015 26 38mean std. dev.10 5 110005> 200 2113 352mean std. dev.> 122 1100022126 231 320mean std. dev.59 58 110003020 214 34mean std. dev.13 8 13000519 2> 200mean std. dev.> 110 130002211 256mean std. dev.34 32 13000301 255mean std. dev.28 38Days to 2-log inactivation Using the predicted number of days until 2-lo g decline as a statistic, ANOVA were performed which resulted in the description of statistically significa nt variations in survival due to temperature but not

PAGE 81

69 TDS. ANOVA results for fecal coliform bacteria are shown in Appendix 4-A. Also, three outlying observations for fecal coliform inactiva tion were observed, including the tw o that were > 200. All outlying observations were at 5 C, one at 1000 mg/l, and both values at 3000 mg/l. To further identify the relationship of TDS and temperature on fecal coliform inactivation, a regression was performed. The results of this are shown in Appendix 4-A. Once ag ain, the effect of temperature was significant when averaged across all TDS concentr ations, but TDS was not a signifi cant predictor of fecal coliform inactivation time. The regression equation was determined to be: FC days = 106 3.35 Temp + 0.00688 TDS Thus temperature negatively affected days for 2-lo g decline, indicating an increase in activation. Regressions of this type allowed assigning significance levels to the respective predictors, such that in this case the constant and temperature variable were signif icant in the regression at the 99% level, and the overall regression equation allows a st atistically significant prediction of variability in inactivation times (p < 0.01). However, TDS was not a significant predictor. The r2 value of 40.6% also shows that the relationship is not very robust. This is may be due to large differences between inactivation in the respective trials at the same conditions in some cases. This regression analysis also highlighted 4 unusual observations as seen at the bottom of the table in Appendix 4-A. These were the three values identified as outliers from the ANOVA (200 days at 1000 mg/l and 5, 19 and 200 days at 3000 mg/l and 5), and the value of 126 days at 1000 mg/l and 22 C. Lower TDS values may be considered more importa nt when considering injection of contaminated surface water to aquifers, such as concentrations below 1000 mg/l as w ould be found in injected surface water. Thus, fecal coliform results were also analy zed while considering only TD S concentrations of 200, 500, and 1000 mg/l (Appendix 4 A). Significant differences among inactivation times were attributed to temperature (p < 0.01), but not to TDS variation. The mean days for predicted 2-log decline across all temperatures did increase with increasing TDS, from 34 days at 200 mg/l to 64 days at 1000 mg/l. However, variability between replicates did not allow the demonstration of statistical significance for this trend. Observed results from TDS-temperature trials fo r fecal coliform thus indicated that temperature significantly increased inactivation; however, variations in TDS concentrations did not produce statistically

PAGE 82

70 significant differences or trends, ev en when considering the interactio n of temperature and TDS. Also, large differences were frequently observed between rep licate trials testing the same conditions, as can be seen by looking at means and standard deviations in Table 13. These trials were different both in terms of different batches of the fecal coliform population and in the batches of prepared experimental solutions. It is likely that other factors related to the microcosm experiment were contributin g to variations in fecal coliform survival. The most egregiou s examples of poor agreement between trials were at 200 mg/l at 5 C, 1000 mg/l at 22 C, and in 3000 mg/l at all three temperatures. Two of these condition sets included values identified as outliers in the ANOVA or regression analysis. Enterococci bacteria Results for enterococci bacteria in TDS-temperature trials were determined and analyzed in the same manner as described for fecal coliform. Te mperature produced strongly significant effects on inactivation of enterococci (p < 0.01 ), while the effect of TDS was significant to the 90% level but not the 95% level (p < 0.1). The interaction of TDS and temperature was not significant. Individual plots of measured culturable concentration, ex pressed as the log of ratios at time points to starting concentrations, are included in Appendix 1-B. Predicted days for 2-log inactivation were determined from model equations fit to data from each trial and are shown in Table 14. Unlike fecal coliform bacteria, only Equation 1 was used to fit enterococci inactivation data, since in no case was an initial increase in concentration observed. Ranges of 2log inactivation days at each temperat ure were 25 to 114 days at 5, 3 to 59 days at 22, and 1 to 25 days at 30.

PAGE 83

71 Table 14. Days predicted for 2-log decline of en terococci at varying TDS and temperature conditions from model curve equations. TrialTDSTemperature1200535 2114 338mean std. dev.62 45 1200223 25 34mean std. dev.4 1 1200301 21 31mean std. dev.1 0 15005110 257 340mean std. dev.69 37 15002259 25 33mean std. dev.22 32 1500307 22 31mean std. dev.3 3 11000525 261 329mean std. dev.38 20 11000224 28 34mean std. dev.5 2 11000301 23 31mean std. dev.2 1 130005113 276mean std. dev.95 26 130002221 249mean std. dev.35 20 13000304 225mean std. dev.15 15Days to 2-log inactivation

PAGE 84

72 Appendix 4-B shows results of statistical analyses on predicted days for 2-log decline. Along with the description of significance due to the two parameters, two values were determined to be outliers, the value of 114 days in 200 mg/l and 5, and the value of 110 days in 500 mg/l and 5. To further examine the relationship of temperature and TDS to enterococci inactivation, a regression was performed as for fecal coliform. The regression equation determined for th e relationship of temperature and TDS to enterococci decline is: Ent Days = 65.7 2.44 Temp + 0.00825 TDS As indicated by regression p values, both temperature and TDS concentration are significant components of the prediction model at the 95% level, and with an r2 value of .587, the model fits data for predicted number of days for 2 log decline marginally better than the fecal coliform regression. Also, the same 2 observations were deemed unusual, being much larger than the regression model predicted. Besides these two condition sets, there were large differences between trials at 22 and 500 mg/l and 30 and 3000 mg/l. Enterococci data were also comp ared for only the three lower TDS concentrations. This analysis showed that unlike the case in which all four TDS co ncentrations are considered, TDS was not a significant factor for enterococci inactivation while temperatur e still was significant (Appendix 4-B). Figure 11 displays a graphic of means and conf idence intervals for enterococci at these TDS concentrations. As the means for each TDS level show, there wa s an inconsistent trend in terms of increasing TDS concentration. Thus, the increase in mean days for 2 log inactiva tion at the three temperatures at 3000 mg/l was great enough to offset the observed decline in the mean at 1000 mg/l and pr oduce a significant effect due to TDS.

PAGE 85

73 Figure 11. Enterococci mean days for 2-log d ecline as a function of temperature and TDS. Individual 95% CI TDS mg/L Mean -+---------+---------+---------+---------+ 200 22.4 (-------------*------------) 500 31.6 (------------*-------------) 1000 15.1 (-------------*------------) -+---------+---------+---------+---------+ 0.0 12.0 24.0 36.0 48.0 Individual 95% CI Temp C Mean ------+---------+---------+---------+----5 56.6 (------*-----) 22 10.6 (-----*------) 30 2.0 (------*-----) ------+---------+---------+---------+----0.0 25.0 50.0 75.0 The means are sharply lower for te mperatures of 22 and 30 than at 5. However, the difference between inactivation at 22 and 30 is not as great. Thus, as with fecal coliform, temperature is the stronger determinant of enterococci inactivation un der these controlled conditions. Still, replicate experiments in some cases had quite large differences in behavior, particularly trials at 500 mg/l and 22, and 3000 mg/l and 30. In fact, the predicted days for 2-log decline at 30 C in 3000 mg/l (Table 14) show that in one trial, the value was very close to those observed at 30 for the lower TDS concentrations, while the other trial showed a much longer period. Therefore, longer inactivation periods at higher TDS may be an artifact at this temperature at least. RNA and DNA coliphage For results of TDS-temperature st udies with the two types of colipha ge, graphs with observed data and model curves fit to these data are in Appendix 1-C and 1-D. Two trials were performed for all conditions with these organisms. For DNA coliphage each separate trial was conducted using separate batches of phage and ionic solutions. However, for the RNA coliphage, trials for 200 and 1000 mg/l were performed simultaneously as were both trials for 500 and 3000 mg/l, but the two sets were separate from each other. Predicted periods for 2-log decline are shown in Table 15 (F+ RNA coliphage) and Table 16 (DNA coliphage). Statistical analysis of these peri ods indicated both TDS concentration and temperature significantly affected RNA coliphage inactivation, and a significant interaction of the two factors also existed (p < 0.05). Thus the relative impact of TDS was greater at greater temperatures. For DNA

PAGE 86

74 coliphage, TDS was not a significant factor for vari ability of inactivation periods. Temperature was only significant at the 90% level, and not at the 95% leve l. In examining the mean values for each condition between the two types of phage, inactivation for RN A phage was generally more rapid than for DNA phage. This was apparent qualitatively at all condition se ts except one. Also, the variability between trials was less for RNA phage than for DNA phage, indica ting that unknown differences between batches of phage or in preparation of the test solutions could be resulting in noticeable variation of inactivation kinetics. Inactivation in one trial for DNA coliphage resulted in a predicted number of days to 2-log decline that was greater than 200 days.

PAGE 87

75 Table 15. Predicted days for 2-log inactivation for F+ RNA coliphage. TrialTDSTemperature1200546 244mean std. dev.45 1 12002232 242mean std. dev.37 7 12003013 214mean std. dev.14 1 1500582 239mean std. dev.61 30 15002252 251mean std. dev.52 1 15003021 227mean std. dev.24 4 11000564 250mean std. dev.57 10 110002241 232mean std. dev.37 6 110003011 213mean std. dev.12 1 130005155 2140mean std. dev.148 11 130002247 242mean std. dev.45 4 13000306 25mean std. dev.6 1Days to 2-log inactivation Table 16. Predicted days for 2-log inactivation for DNA coliphage. TrialTDSTemperature12005141 265mean std. dev.103 54 120022> 200 267mean std. dev.> 134 12003063 212mean std. dev.38 36 1500590 266mean std. dev.78 17 15002289 276mean std. dev.83 9 15003047 211mean std. dev.29 25 110005112 259mean std. dev.86 37 110002273 243mean std. dev.58 21 110003039 29mean std. dev.24 21 130005110 2167mean std. dev.139 40 130002256 2136mean std. dev.96 57 130003049 2123mean std. dev.86 52Days to 2-log inactivation

PAGE 88

76 The results of ANOVA for RNA coliphage are in Appendix 4-C. Besides the ANOVA results showing significant differences due to temperatur e and TDS, this can be shown quantitatively by a regression of the two factors. The regression equation for RNA coliphage is: RNA days = 78.7 + 0.0110 TDS 2.47 temp and all components of the regression equation we re significant to the 95% level. However, the r2 value was .579, which is still fairly low and indicates a number of points were not modeled well by this equation. These statistical analyses quantitatively describe a trend that temperature increases inactivation of RNA coliphage, and greater TDS reduces inactivation, th ereby resulting in greater days until 2-log decline. However, a graphic presentation of means for F+ RNA coliphage in Figure 12 shows that values averaged across the entire temperature spectrum are closely grouped at TDS concen trations up to 1000 mg/l, and increase at TDS concentrations of 3000 mg/l. Furtherm ore, in looking at the actual values estimated for 2log decline in Table 15, it is apparent that the source of differences between TDS concentrations is due to greater number of days (148) at 5 in 3000 mg/l TDS, and this large value is of course a result of extrapolation of observed inactivation kinetics, which were slow during the 28-day experiment. The means at 5 at the other TDS concentrations were 45, 61, an d 57 days. Averages by temperature and TDS in this table indicate that at ambient temperatures of the Flor idan aquifer system, there was little to no difference between the TDS concentrations, with inactivation being most rapid actually at 30 C at 3000 mg/l. This explains the statistically significant interaction of temperature and TDS that was described by a 2-way ANOVA. In looking further at these results, the mean s by TDS are quite similar at 22 C and are actually somewhat less at 3000 mg/l at 30 C than the other TDS concentrations. Due to the variation with temperature, and the fact that the TDS effect actually reverses at highe r temperatures, it may be that this apparent effect is more due to experimental variability or even the result of extrapolation of inactivation kinetics beyond the measured periods (28 days).

PAGE 89

77 Figure 12. F+ RNA coliphage mean days for 2-log decline as a function of temperature and TDS. Individual 95% CI temp C Mean --------+---------+---------+---------+--5 77.5 (---*---) 22 42.4 (---*---) 30 13.8 (---*---) --------+---------+---------+---------+--20.0 40.0 60.0 80.0 Individual 95% CI TDS mg/L Mean -----+---------+---------+---------+-----200 31.8 (-----*-----) 500 45.3 (-----*-----) 1000 35.2 (-----*------) 3000 65.8 (-----*-----) -----+---------+---------+---------+-----30.0 45.0 60.0 75.0 An ANOVA for only the three lower TDS concentr ations was also done (Appendix 4-C) and this analysis revealed that TDS concentr ation in this lower range was not a significant factor (to the 90% or 95% level), while temperature still was significant (p < 0.01). The means and confidence intervals from this comparison are shown here in Figure 13. Figure 13. F+ RNA coliphage mean days for 2-log decline as a function of temperature and TDS, lower concentrations. Individual 95% CI TDS mg/L Mean ---------+---------+---------+---------+-200 31.8 (----------*---------) 500 45.3 (---------*----------) 1000 35.2 (---------*----------) ---------+---------+---------+---------+-30.0 40.0 50.0 60.0 Individual 95% CI temp C Mean ------+---------+---------+---------+----5 54.2 (------*------) 22 41.7 (------*------) 30 16.5 (------*------) ------+---------+---------+---------+----15.0 30.0 45.0 60.0 Interestingly, when all TDS c oncentrations are considered, the mean 2-log inactivation days are fairly evenly spaced across the three-temperature re gime, unlike the grouping ob served for the means at only TDS 200-1000 mg/l, in which the means at 5 and 22 are more closely grouped. The comparison of

PAGE 90

78 means at each temperature shows the effect of removing 3000 mg/l experiments. Means do not change substantially at 22 and 30 between Figure 12 and Figure 13, but at 5 the effect of large values at this temperature in 3000 mg/l is apparent. Statistical analyses results for DNA coliphage are given in Appendix 4-D. An ANOVA for temperature and TDS up to 1000 mg/l revealed that TDS over this range was not a significant variable, while temperature was significant to the 90% level. Re gression results also indicated that temperature and not TDS (over the entire examined range) was a significant component of a model equation to fit the observed data. However, with an r2 value of 0.218, the regression was not an accurate model for predicting the number of days for 2-log decline. This is possibly due to the variability present between replicates at the same set of conditions. The regression equation is: DNA days = 106 + 0.0104 TDS-phage 1.98 temp. The diagram in Figure 14 shows mean days for 2-log decline of DNA coliphage in these experiments. It is apparent that larger differences were observed on average between 22 and 30 than between 5 and 30 C. Figure 14. DNA coliphage mean days for 2-log decline as a function of temperature and TDS. Individual 95% CI temp C Mean --------+---------+---------+---------+--5 101 (----------*---------) 22 96 (---------*----------) 30 44 (----------*---------) --------+---------+---------+---------+--35 70 105 140 Individual 95% CI TDS mg/L Mean -------+---------+---------+---------+---200 95 (-----------*-----------) 500 63 (-----------*-----------) 1000 56 (-----------*-----------) 3000 107 (------------*-----------) -------+---------+---------+---------+---35 70 105 140 PRD-1 bacteriophage Results for survival experiments with the Salmonella bacteriophage PRD-1 indicated that it was considerably hardier under the examined conditions. For this reason, and because many of the predicted number of days until 2-log inactivation were greater than 200, the statistic used for analysis of PRD-1 decline was a first-order inactivation rate. Beside s the extended periods for decline predicted from

PAGE 91

79 observed data values, most all of the PRD-1 data sets showed less than 1-log decline and were fit with a first-order variant of Equation 1. Thus, since this statistic was used, the remaining trials for which a nonlinear model was employed were replaced with a linear one. Table 17 contains inactivation rates in log N/N0 change per day from TDS-temperature trials. It is important to note that as these values are negative rate constants, the greater magnitude (more negative) numbers indicate a greater ra te of inactivation. Table 17. PRD-1 first-order in activation rate constants. TrialTDSTemperature12005-0.012 2-0.023mean std. dev.-.018 .008120022-0.020 2-0.036mean std. dev.-.028 .011120030-0.016 2-0.083mean std. dev.-.050 .04715005-0.022 2-0.010mean std. dev.-.016 .008150022-0.012 2-0.012mean std. dev.-.012 0150030-0.023 2-0.147mean std. dev.-.085 .088110005-0.022 2-0.022mean std. dev.-.022 01100022-0.014 2-0.047mean std. dev.-.031 .0231100030-0.011 2-0.229mean std. dev.-.120 .154130005-0.010 2-0.010mean std. dev.-.010 01300022-0.014 2-0.020mean std. dev.-.017 .0041300030-0.034 2-0.017mean std. dev.-.026 .012Linear inact. rate (log10/d)

PAGE 92

80 These inactivation rates were used to perform statistical analyses of the relative impacts of temperature and TDS concentration in the same fashion as for the other phage. Results of ANOVA using all TDS concentrations, regression results for the effect of the two predictors, and ANOVA results from only the three lower TDS concentrations are in Appe ndix 4-E. In short, neither ANOVA set nor the regression model indicated a statistically significant impact of TDS or temperature with the conditions evaluated. However, the mean inactivation rate for PRD-1 did increase at 30 C. This also follows the same pattern observed for mean inactivation at all TD S concentrations for the DNA coliphage with respect to temperature effects: means are more closely gro uped at 5 and 22 C, and inactivation is noticeably greater at 30 C. This pattern was also indicat ed with RNA coliphage when considering only TDS concentrations under 1000 mg/l. Figure 15 shows th e mean inactivation rate averaged across all four TDS concentrations at the three temperature levels. Figure 15. PRD-1 inactivation rates averaged by te mperature and TDS. All rates in log change per day. Individual 95% CI temp C log/d --+---------+---------+---------+--------5 -0.016 (-----------*-----------) 22 -0.022 (-----------*-----------) 30 -0.070 (-----------*-----------) --+---------+---------+---------+---------0.105 -0.070 -0.035 0.000 Individual 95% CI TDS-phage log/d -+---------+---------+---------+---------+ 200 -0.032 (-------------*-------------) 500 -0.038 (------------*-------------) 1000 -0.057 (-------------*------------) 3000 -0.017 (-------------*-------------) -+---------+---------+---------+---------+ -0.105 -0.070 -0.035 0.000 0.035 Although temperature was not statistically significant as a factor for PRD-1 inactivation rates, there was a trend such that more rapid inactivation wa s observed at the highest temperature (30 C). Mean inactivation rate constants from all TDS concentrations increased from -0.016 at 5 C and -0.022 at 22 C (very similar) to -0.070 at 30 C (Figure 15, all log/d). Regarding TDS, the most rapid rates on average

PAGE 93

81 were at 1000 mg/l, and declined at 3,000 mg/l. Ho wever, variability between trials may be somewhat confounding in this situation. For the second tria l at TDS concentrations of 200, 500, and 1000 mg/l (performed simultaneously), there wa s a considerable increase of inactivation rates from the first trial, mostly at 30 C. Trials at 3000 mg/l were performed at separate times than these three. Thus, additional factors relating to the experimental environment likely in fluenced inactivation rates at higher temperatures. If only the first trial at 200, 500, and 1000 mg/l and the two trials at 3000 mg/l are considered, inactivation rates at the three TDS and three temperatures have a mu ch more narrow range and no consistent trends with temperature and TDS (Table 17). PBS controls and Variability In cases where apparently large differences exis t between predicted days for 2-log inactivation, and some or all of those times are in excess of 28 days part of the difference may be due to extrapolation of slower initial kinetics. In other words, there may ha ve been a change in kinetics of decline after the experimental duration such that actual periods for 2-log decline were more similar, namely that they would be shorter. This may be more likely in situations at lower temperature (5), in which inactivation over the experimental durations was often slow and thus minor differences in rates and kinetics, when extrapolated to 2-log declines, resulted in peri ods that were quite different. Variation in initial kinetics of decline still cannot fully explain the sometimes large variability between replicates however, clearly in cases such as fecal coliform at 3000 mg /l and 30 C, other factors were at play (Table 13). Additional analyses of variance for variability of inactivation with the experimental set, regardless of TDS, showed that significant differences existed due to set (replicate) for fecal coliform, enterococci DNA coliphage, and PRD-1 (all p < 0.05). (F+ RNA coliphage data could not be analyzed in this way since surviv al in 500 mg/l TDS was not evaluated at all three temperatures at once; 30 C experiments were performed separately, with experiments at 3000 mg/l.). The results of these ANOVA are shown in Appendix 4 under the respective section for each organism. Sets for enterococci and fecal coliform were performed such that Set 1 evalua ted TDS concentrations of 500 and 3000 mg/l, Set 2 was 200 and 1000 mg/l, Sets 3 and 4 were 200, 500, and 1000 mg/l, and Set 5 was 3000 mg/l. For the DNA coliphage and PRD-1, Sets 1 and 2 evaluated concentrations of 200, 500, and 1000 mg/l, and Sets 3

PAGE 94

82 and 8 evaluated concentrations of 3000 mg/l. For RNA coliphage, Set 1 evaluated concentrations of 200 and 1000 mg/l at all temperatures and 500 mg/l at 5 and 22, while Set 2 evaluated 3000 mg/l and 500 mg/l at 30. The RNA coliphage trials for each of these conditions were performed in duplicate. The 2way ANOVA, which also included temperature as a factor, showed variability of mean inactivation rates/days by set varied with temperature (significant interaction existed) for enterococci and PRD-1. In addition, the significance of variability was greater due to set than due to temperature for PRD-1 and DNA coliphage, both of which had p values for significance of set differences below 0.05 but between 0.05 and 0.1 for temperature. Set differences for these two phages, which we re tested together in experimental microcosms, also showed similar patterns in variability. Figure 16 shows a column chart of the mean days for 2-log decline (PRD-1 first-order rate constants co nverted), averaged across the three temperatures and separated by experimental set. There is an apparent similarity in the relative trend between sets for the two phages. Most notably, inactivation in Set 2 was much more rapid on average. This may indicate that differences in experimental solutions or actual bo ttles used may be partly responsible. A similar comparison for the bacteria, fecal co liform and enterococci, rev ealed no similarity in the relative variation of mean values for sets (not shown).

PAGE 95

83 Figure 16. Mean days for 2-log inactivation by experimental set for DNA coliphage and PRD-1. Sets 1 and 2 were TDS concentrations of 200, 500, and 1000 mg/l, sets 3 and 8 were 3000 mg/l only. Each column represents the mean of values from all three temperatures. 98 45 72 118 30 103 127 1420 50 100 150 1238SetEst. days for 2-log decline DNA Coliphage PRD-1 To gain a rough picture of any covariance between set differences for experimental trials and set differences in PBS control solutions figures were constructed which ju xtaposition average days for 2-log inactivation of each organism by set (across all temperatures and TDS conc entrations tested) with averages across all temperatures for the PBS controls for each re spective set. These figure s are shown as Figure 17 A-D below. In each chart, set averages were arra nged in order of increasin g predicted survival in experimental trials, while PBS was shown corresponding to each respective experimental set. From these charts, there was little evidence for a general covariance of experimental trial set averages and those of PBS controls. However, for enterococci, there was a genera l parallel to variations in set averages between the two types of solutions for three of the four trials. Thus, for enterococci, differences between batches of cells may contribute to observable differences in su rvival potential even in the more isotonic, buffered solution. Also, for the DNA coliphage, inactivation was relatively slower in Set 8 for both the experimental trials and the PBS controls, but the much more rapid inactivation of experimental Set 2 trials observed for both PRD-1 and DNA coliphage was not re flected in the PBS control averages for either

PAGE 96

84 organism. The overall slower inactivation in both types of solution for DNA coliphage in Set 8 is noteworthy in light of set differences for 3000 mg/l trials, for which inactivation in the second trial was much less rapid at higher temperatures than in the first trial (Table 16). Recalling that Set 8 corresponded to the second trial for 3000 mg/l TDS only, the relatively less rapid inactivation in PBS on average for this trial as well indicates that the effect was more likely due to differences in the batch of coliphage. For the organisms with poor agreement of relative variability be tween experimental trials and PBS trials, it may be that batch differences were more significant in the low salinity, unbuffered mixed ion solution. Alternatively, it may be that differences in the mixe d ion solutions were responsible, but affected the two bacterial populations to different degrees in their case. From these charts, it is apparent that variability existed between sets for PBS as well, but the relative differences for each set did not correspond to similar relative differences in experimental sets on average, except for the enterococci cases.

PAGE 97

85 Figure 17 A-D. Set differences for experimental tr ials and PBS controls. Afecal coliform, Benterococci, CDNA coliphage, DPRD-1. A. Fecal coliform0 40 80 120 3425SetDays for 2-log10 decline experimental trials0 40 80 120Days for 2-log10 decline PBS controls Experimental trial averages PBS control averages B. Enterococci0 15 30 45 60 2435SetDays for 2-log10 decline experimental trials0 40 80 120Days for 2-log10 decline PBS controls Experimental trial averages PBS control averages

PAGE 98

86 C. DNA coliphage0 80 160 2318SetDays for 2-log10 decline experimental trials0 70 140Days for 2-log10 decline PBS controls Artificial seawater average PBS average D. PRD-1-0.1 -0.05 0 2318SetInactivation rate constant experimental trials (log d-1)-0.1 -0.05 0Inactivation rate constant PBS controls (log d-1) Experimental trial average PBS control average

PAGE 99

87 Discussion Modeling of observed viable concentrations of the five organism groups provided estimates of days to achieve 2-log (99%) reductions in concentrations or first-order inactivation rate constants (PRD-1). These periods or rates were used to compare variab ility between conditions with analyses of variance, which revealed factors that were determined to have a statistically significant aff ect on survival. Factors which were statistically significant to the 90% or 95% level are listed for each organism group in Table 18. Temperature in these trials ranged from 5 to 30 C, and survival was inversely related to temperature in cases where it was significant. TDS was a significant va riable for enterococci and F+ RNA, but only when considering the high TDS concentration of 3000 mg/l. Also, recall that for F+ RNA, the effect of TDS reversed with temperature, such that at 5 C, inac tivation was on average relatively slower at high TDS (3000 mg/l), and at 30 inactivation was slower at low TDS. Table 18. Significance of TDS and temperature as factors affecting survival of indicator organisms in water. OrganismTDS range (mg/L)Significant factors (95% unless noted) fecal coliform200 3000 temp 200 1000 temp enterococci200 3000 temp, TDS (90%) 200 1000 temp F+ RNA coliphage200 3000 temp, TDS, interact 200 1000 temp DNA coliphage200 3000 temp (90%) 200 1000 temp (90%) PRD-1200 3000 none 200 1000 none In this study, we have specifically addressed TDS such that the waters were artificially created and thus should have been standardized with regard to other parameters that would vary in ground water samples from the environment. Variability between expe rimental trials with identical temperature regimes

PAGE 100

88 was significant, such that other trends may have b een obscured. However, in the case of lower TDS concentrations, which were evaluated concurrently for the most part, TDS was still not significant as a factor affecting inactivation. This study is significant in the types of organisms that have been evaluated. Most studies investigating survival in ground water or model ground water situations focus on viruses due to their greater transport ability. We have included tw o bacterial indicator groups in our assessment. Also, organisms used except for PRD-1 were evaluated as composite populations of 9-10 isolates for each group. These organism populations were isolated from the environment and may present a better model for survival under environmental conditions that single strains. Certainly, fecal organisms isolated from the environment are composed of diverse populations, so the use of a composite may help to better approximate their survival. The di rect comparison of a range of TD S concentrations is an important contribution of this work, as no other study was found that evaluated this parameter on survival of such a range of organisms under the same set of conditions fo r comparative analysis. Also, the choice of a mixedion salt for evaluating TDS effects is important, since the dissolved solids composition in environmental waters is more similar to such a mixture than to a single-salt such as sodium chloride. The demonstration of significant decreases of su rvival potential under starvation conditions with increasing temperature is important in the context of survival in Florida ground waters. For instance, temperatures in the Floridan aquifer system range from about 20 to 30 C. Su rvival of fecal indicator organisms is likely to be reduced under these cond itions. However, the range of TDS concentrations observed in the Floridan aquifer system is not likely to influence these types of organisms with the possible exception of enterococci at TDS concentrations of 3000 mg/l or greater. Regarding trends in decline between types of or ganism, for the two populations of bacteria, fecal coliform had greater days for 2-log in activation than enterococci, more no tably at higher temperatures and in the lower TDS concentrations (Table 13 and Tabl e 14). This indicates perhaps a less pronounced temperature-based inactivation of fecal coliform than enterococci, especially in the less-saline water conditions. F+ RNA decline in these trials was gene rally more rapid on average than DNA coliphage, but they were fairly close in most cases. The difference did not appear dependent on temperature, but larger differences between the two phage populations were obse rved at 3000 mg/l TDS. A statistical analysis of average periods for 2-log decline was performed us ing data from fecal coliform, enterococci, RNA

PAGE 101

89 coliphage, and DNA coliphage (Appendix 4) as a 2way ANOVA comparing variability due to temperature and organism type. This analysis revealed significan t differences due to organism and temperature (p < 0.01), but no interactive effects. Figure 18 show s mean days for 2-log decline for each of the above organisms at the three temperatures. From this figure it is apparent that inactivation of enterococci was most rapid on average, RN A coliphage and fecal coliform were sim ilar, and DNA coliphage died off the slowest. These differences were thus statistic ally significant. An additional ANOVA comparing inactivation at only 22 and 30 reveal ed that the differences between organisms at higher temperature were significant as well. Figure 18. Predicted days for 2-log decline for each organism, averaged across all TDS concentrations. 63 15 4 98 34 16 78 42 14 101 96 440 50 100 150 200 52230Temperature (C)Days for 2-log10 decline (+ Std dev.) enterococci fecal coliform RNA coliphage DNA coliphage The results of this study suggest that, for the tested organisms, a direct relationship of TDS to inactivation was not generally apparent. Enterococci were the only organisms for which results may suggest an effect of TDS concentrations. Inactivati on was certainly less rapid at 3000 mg/l than the other TDS concentrations at 22 (Table 14) and was slower at 30 in one trial but not the other. That enterococci

PAGE 102

90 may be hardier under higher salinity conditions in our studies is in agreement with other researcher’s findings that such organisms are generally hardier an d thus more conservative indicators of fecal pollution in marine waters than in freshwater (Evison 1988). The TDS effect for F+ RNA coliphage reversed with temperature. Thus, there is no conclusive evidence from these experiments that TDS from 200 to 1000 mg/l, as an independent factor or wh en considered with temperature ranging from 5 to 30 C, consistently has a verifiable impact on survival of these waterquality-related microorganisms. These findings are similar to those of Yates et al. in that they did not find TDS significantly affected inactivation rates of 3 virus groups, when considering the innate TDS of various ground water samples ranging from 34 to 1100 mg/l (Yates 1985). Temperature effects were more pronounced, such that inactivation of all organisms increased with increasing temp eratures. This conforms with findings of many studies on virus survival in ground water (Keswick 1982; Yates and Gerba 1985; Yates 1985; Yahya 1993; Blanc and Nasser 1996; Nasser and Oman 1999; Alvarez 2000). Others have de monstrated reduced survival of enteric bacteria at higher temperatures as well (Evison 1988), while in some cases there has been indication that higher temperatures do not always result in faster inactivati on under environmental conditions for bacteria such as fecal coliform (Toranzo s 1991; Medema 1997). This may be rela ted to the re-growth potential of some bacteria, particularly in higher-nutrient waters. As discussed in Chapter 2, a compilation on survival data from many studies, consisting of both ground and surface waters, indicated a general increase of inactivation rates for colifor m bacteria (total and/or fecal) and en terococci/streptococci Analysis of temperature effects on survival of composite populations of these b acteria revealed a statisticallysignificant decrease of survival potential at higher temperatures in this study as well. Only PRD-1 did not indicate significantly more-rapid inac tivation at higher temperatures, lending credence to its reputation as a hardy phage under a range of environmental conditions. One aspect of these studies that could be improved for future work is the impact of calcium concentrations and/or hardness on survival. Yates et al. has found calcium hardness to negatively impact survival of MS-2 in ground water (Yates 1985). While we evaluated a mixture of ions for TDS concentration effects, the relative conc entration of calcium was similar to that of seawater. However, in a carbonate aquifer such as the Floridan aquifer system, calcium concentrations would be relatively higher. It would be beneficial in the future to specifically evaluate changes in hardness, for instance at a fixed TDS

PAGE 103

91 concentration, for impacts on survival of a composite population of coliphage and perhaps other enteric organisms. This could be accomplished by adding additional calcium chloride or calcium bicarbonate to the ionic mixture. Such a direct comparison wo uld provide a more accurate evaluation of calcium concentration effects directly, rather than indirect analysis by evaluation of several native ground water samples of differing hardness as was done previously. Also, a direct comparison of variable dissolved oxygen concentrations under non-aerated incubation conditions would be beneficial to evaluate survival in aquifers, where DO varies considerably and in places of the Floridan aquifer system or contaminated ground water can be quite low.

PAGE 104

92 CHAPTER 4: SURVIVAL OF INDICATO RS IN REPRESENTATIVE SUPPLY AND RECEIVING WATERS FOR AQUI FER STORAGE AND RECOVERY IN FLORIDA Introduction Ground water is a heavily used resource in th e state of Florida. For 1995, ground water represented about 60% of the total freshwater withdraw als in the state, for a total of about 4.3 billion gallons per day. Of that total, 1.8 billion gallons daily went to public supply and about 300 million gallons per day was withdrawn by domestic self-supply wells. Another 1.5 billion gallons per day was withdrawn via agricultural self-supply wells (Marella 1999). Water supply for Florida’s 14 million people is a continuing concern, and ground water makes up the vast majority93%, of drinking water supply in the state. Thus the quantity and quality of ground water is a major issue statewide, along with many other regions. Due to continued population growth, along with an interest in remediation of ecological and hydrological impacts caused by excessive withdrawal an d re-direction of water, new sources and practices for water supply are being sought. One technology of rising importance that involves injection of surface water to aquifers for storage and later recovery for use is termed Aquifer Storage and Recovery (ASR). In 1999, there were 1185 aquifer recharge and ASR wells docume nted by the EPA in the U.S. (USEPA 1999). In Fl orida alone there were 130 ASR wells, and up to 488 wells not differentiated between ASR and aquifer recharge wells (with most of these being aquifer recharge wells). ASR wells are different from aquifer recharge wells in that water is stored then recovered from the same well w ith an ASR well; aquifer recharge wells simply inject water to replenish the aquifer. More current data on ASR wells in Florida indicate that, according to the Florida DEP, there are at least 148 ASR wells in various stages of planning, construction, testing or operation in the state (data available at FDEP website). These are divide d into 94 treated drinking water wells, in the planning, construction, or testing/operational phases, 28 part ially treated surface water wells in

PAGE 105

93 the planning or construction phases, 14 reclaimed water and 11 raw ground water wells in planning, construction, or testing phases, and 1 raw surface water well that is constructed but idle. As a means of storage, ASR is beneficial in its cost efficiency and limited impact on the environment (as compared with reservoir construction), and can possibly enable the storage of very large volumes of water when compared to storage tanks. As such, ASR systems can be a valuable tool for water supply management officials to effectively manage supplies for drinking water, irriga tion, or ecosystem preservation and restoration. The Comprehensive Everglades Restoration Project is planned to involve 333 ASR wells to potentially store up to 1.7 billion gallons per day in the Upper Floridan aquifer (NRC 2001) but this project currently accounts for only a few test wells as detailed above, which are par tially treated surface water wells. The scale of this proposed use of ASR is unprecedented, and is an important option for restoration of stable water flows through the Everglades ecosystem. One environmental concern over the implementation of ASR systems however, is the impact of stored water on the quality of the existing water in aquifers. Although ASR water is generally pumped into receiving aquifers with lesser quality water, public co ncern has arisen over the migration of stored water and possible contamination of pristine aquifers if raw surface water is employed. Currently, the Florida Department of Environmental Protection, under the guidelines of the Safe Drinking Water Act’s Underground Injection Control regulations, requires that water injected into or above aquifers classified as Underground Sources of Drinking Water (USDW) via ASR wells meets all primary and secondary drinking water standards (Drew 2001). One major component of these standards is stipulated by the total coliform rule, which states that no total coliform bacteria be detected per 100 ml of water. Considerable interest among proponents of ASR technology exists in dete rmining the feasibility of relaxing pre-treatment requirements for stored water, assuming that natural attenuation of potentially harmful microorganisms that may be introduced occurs due to biological, physical, and chemical factors present in the subsurface environment. But the primary reas on for public concern ov er injection of untreated surface water is the presence of potentially pathogenic microorganisms in surface water and lack of information on their fate once in the aquifer. Of primary concern is the pote ntial transport of introduced pathogenic microbes to individual domestic wells which are not regulated by the Safe Drinking Water Act and frequently no not disinfect water prior to use. As a result, there is an urgent desire in the state of Florida for data to

PAGE 106

94 demonstrate the fate of allochthonou s organisms such as fecal indicators in water typical of what may be employed in ASR systems. The research described in this chapter was performed to investigate survival of five groups of water quality indicator organisms in surface reservoir and Upper Floridan aquifer water samples that are representative of those to be used for ASR. The objective was to quantitatively describe inactivation of these organisms in these water sources under temperature conditions that may be observed in the Upper Floridan aquifer, while also investigating the potential impact of background microbial levels. In addition, survival at higher temperatures in the range of 22 to 30 C, as found in the Floridan aquifer system, was compared to a low temperature of 5 C as may be seen in aquifers of colder regions. As described in the review of survival studies in ground water in Chapter 2, several investigators have examined the impact of background or autochthonous microbes in ground water on inactivation of seeded public-health-rel ated microorganisms. Observations with regard to the direct comparison of sterile vs. raw water sources for survival studies were somewh at varied in these studies. Some did not observe a statistically significant impact of filtered or autoclaved water, or saw conflicting trends. For example, in studies by Yates and others on virus survival in ground water (Yates and Gerba 1985; Yates 1990), the effect of filter sterilizing ground water samples prio r to seeding did not produce consistent effects on inactivation rates of the viruses (statistically significa nt differences were observed, but without consistent trends in terms of increasing or decreasing inactivatio n rates). In a study on viruses in ground water and soil suspensions, Sobsey et al. observed more rapid inactivation in ra w water vs. sterilized more often than not (7 out of 10 pairs, for viruses including hepatitis A, poliovirus, and echovirus) (Sobsey 1986). Thus for three sets, inactivation was more rapid in the pre-sterilized water. Also for viruses in ground water, Alvarez et al. found slightly faster inactivation of polio an d MS-2 in filter-sterilized water than in raw conditions (Alvarez 2000). Interestingly, in a comparis on of background microorganism effects on survival of seeded Aeromonas hydrophila in ground water and surface water, inactivation in sterile water was slower on average, but the effect wa s more dramatic in surface water than in ground water (K ersters 1996). Regarding possible temperature interactions with background microbi al effects, Medema et al noted that C. parvum viability declined more rapidly in sterile vs. raw river water at 15 C but not at 5 C. Similar trends were observed for E. coli and E. faecium in this water source (Medema 1997). For the investigations

PAGE 107

95 on survival in representative ASR so urce and receiving water samples desc ribed in this chapter, the impact of background microbial concentrations on survival was also investigated. It seems from the few findings described above, water type interactions may also o ccur and this possibility was examined in the current research by comparing pasteurized water samples to su rvival in raw water under identical conditions, while also contrasting aquifer and surface reservoir sources. Methods Two sample sites were utilized to provide wa ter sources for representative aquifer and surface reservoir waters that may be involved in ASR projects. The two sites were near the cities of Bradenton and West Palm Beach. The Bradenton site involved the B ill Evers reservoir which pr ovides raw water to the City of Bradenton drinking water treatment facility Aquifer water was drawn from the Avon Park formation of the Uppe r Floridan aquifer using the ROMP TR4-7 well. West Palm Beach surface reservoir source water was the Clear Lake reservoir, which supp lies raw water to the adjacent City of West Palm Beach drinking water treatment facility. Ground wa ter samples were taken from the Lake Lytal Park wellsite using the PBF-3 well which taps the Upper Florid an aquifer. Samples from all sites were collected in 1-L polypropylene bottles (pre-sterilized). Avon Park samples were collected after purging of 3 well volumes via a gasoline centrifugal pump, followed by pumping using an electric peristaltic pump to withdraw sample water. Lake Lytal Park well purging was performed by opening this artesian well 24 hours prior to sampling to allow development. Both reservoir site samples were taken from just below the surface. Temperature, co nductivity, and pH of the water were m easured on site when possible. Temperature data was not recorded on site for the first set of West Palm Beach samples. Water samples were transported or shipped on ice to the laboratory for use in survival studies. The water samples were divided and half of the water was subjected to heat treatment to reduce background microbial populations prior to seeding with test organisms. The heat pasteurization procedure involved raising the temperature of water samples while still in collection bottles to 70 C in a hot water bath, and holding the temperature at 70 C for 30 minutes. Th is step was performed to gain some comparative

PAGE 108

96 assessment of the impact of intact native microbial populations on survival of seeded non-native water quality indicator microorganisms. Some background microbial parameters were also determined from raw and pasteurized water samples. These parameters were fecal coliform bacter ia, enterococci bacteria, combined somatic and malespecific coliphage (DNA and/or RNA), heterotrophic pl ate count (HPC) bacteria, and in samples for PRD-1 survival study, any background PRD-1. Bacterial indicator concentrations were assayed with standard membrane filtration procedures, using mFC media for fecal coliform (APHA 1992), and mEI media for enterococci (as described in EPA method 1600). Coliphage were determined by assaying 10 ml in 2-ml aliquots by the double agar overlay procedure, using E. coli C-3000 host. HPC bacteria numbers were determined by the spread-plate method for raw water and membrane filtration for pasteurized samples, using R2A agar incubated at 22 C (room temperature) for 1-4 days. PRD-1, when measured, was performed using 10 ml of water in 2-ml aliquots as for coliphage, with host Salmonella typhimurium LT2 as described previously in Chapter 3. Test microorganism populations employed for thes e “natural water” experiments were the same as those used previously in TDS-temperature experiments described in Chapter 3. Briefly, the bacterial populations were composed of fecal coliform bacteria (8 E. coli isolates and 2 K. pneumoniae isolates) and enterococci (7 Enterococcus faecalis 1 E. faecium and 1 E. durans ). Coliphage populations were composed of DNA coliphage (10 isolates, 2 male-s pecific, 8 somatic) and F+ (male specific) RNA coliphage (10 isolates). Also, a single PRD-1 isolate was employed. Organisms were prepared prior to seeding in the same way as described for prior experiments (Chapter 3). For initiating survival trials, water samples were distributed to polypropylene test bottles which were prepared as described for TDStemperature trials. Water samples were set to the experimental temperatures of 5, 22 and 30 C for temperature equilibration for a short time, then test organisms were seeded. Bacteria and viruses were seeded to achieve a concentra tion of approximately 2-3 x 104 organisms per ml in each microcosm. Time 0 samples were taken immediately. Subsequent time-point samples were taken on or about days 2, 3, 5, 7, 10, 14, 21, and 28. Sample analyses for culturable /viable organisms at each time point were performed using the same methods as described for TDS-temperat ure trials, using membrane filtration with mFC agar for fecal coliform and mE for ente rococci, and double agar overlay assays for the phage with their

PAGE 109

97 respective hosts. Experiments were performed in sets with the first two water sample events (August and November, 2002) being used for replicate trials for bacteria (mixed together) and DNA coliphage and PRD-1 (mixed together). Water from the third sample event was used for simultaneous duplicates with F+ RNA coliphage. As was done for TDS-temperature trials described in Chapter 3, survival in phosphate buffered saline control microcosms was evaluated for each experimental set. These PBS bottles were set at the same temperature regime as the environmental water. PBS was made as described in Chapter 3 methods, and was sterile. Results of infective/culturable counts for each time point were analyzed in the same manner as described in Chapter 3. In short, plots of Log N/N0 ratios for each data set were fit to one of two curves. Equation 1 was used most often, and in cases where less than 1 log decline was observed, data were fit using a first-order variant. Alternatively, the po lynomial Equation 2 was employed in cases of fecal coliform and in one case DNA coliphage data sets wher e initial increases in counts were observed. In either case, observed data were fit to the equation using least-squares analysis in Microsoft Excel. Log10 N/N0 = k timem (1) Log N/N0 = k1 time2 + k2 time (2) After a model was fit to each data set, the equation with best-fit parameters was used to estimate the number of days for 2-log decline in the concentration of infective or culturable organisms. However, as in Chapter 3 for TDS-temperat ure experiments, PRD-1 inactivation wa s expressed in terms of first-order inactivation rates. Equation 1 was used to fit PRD-1 data, while m was fixed at 1 and thus the value of k was used as a statistic for comparison of relative effects of the various parameters. To gain some insight on parameters of these experiments which may play a role in controlling inactivation of contaminant microorganisms in thes e types of water and conditions, ANOVA tests were performed using predicted days for 2-log inactivation or inactivation rate (PRD-1) as a statistic. ANOVA were performed independently for each organism group (in other words, fecal coliform, enterococci, and each type of phage were considered independently). Comparisons performed included analyses of impacts of temperature and pasteurization in both water types combined, 3-way analyses of water type (ground water or surface water), temperature, and pasteurizatio n, 2-way analyses of water type and temperature in only raw water samples, and analyses of pasteuri zation and temperature effects in each water type

PAGE 110

98 considered separately. In addition, an examina tion of temperature effect s at only the two higher temperatures, indicative of conditions in the Upper Flor idan aquifer, along with water type in raw water was also performed. Results Representative water samples from two ASR sites in Florida were utilized for natural water source survival studies. Water from each s ite was characterized for each trial run in terms of several microbial and basic chemical parameters. In performing survival ex periments with these water samples, water from the first two sample events was used for the first and se cond replicates on fecal coliform, enterococci, DNA coliphage, and PRD-1. Water from the third sample event was used for both replicates with F+ RNA coliphage, performed simultaneously. All sample events occurred in the summer or fall. The results of basic characterization analyses for th e raw (unpasteurized) water samples from each of these sites are given in Table 19 (chemical parameters and temperatur e) and Table 20 (microbial parameters).

PAGE 111

99 Table 19. Physiochemical measurements of raw water from the two sample sites. Avon Park Well Bill Evers Reservoir Lake Lytal Park Well Clear Lake Reservoir T (oC) a 30322226 b30292327 c30272330 pH a7.16.67.07.0 b7.27.27.38.1 c7.16.97.58.4 TOC (mg/l)a2.13180.7313.2 b1.1416.1< 0.511.8 cNDNDNDND Conductivitya 2.98 mS/cm2384 S/cm28.15 mS/cm2417 S/cm2b 2.94 mS/cm2353 S/cm26.23 mS/cm2317 S/cm2c 3.08 mS/cm2224 S/cm28.16 mS/cm2450 S/cm2a15002004,000200 b15001753,000150 c15001124,000225 ND = analysis not done Approx TDS (mg/l) One notable observation from these measurements was the expected differences in conductivity and thereby TDS between the groun d water and surface water sources, such that the surface water TDS concentrations were much lower than either ground wa ter source’s value. Also, TDS in the Lake Lytal Park (West Palm Beach area) aquifer source was 3,000 4,000 mg/L, while that of the Avon Park aquifer source was only 1500 mg/L. Temper ature was consistently in the range of 22 to 30 C, with the two aquifers having very stable temperature measurements over the three sample events. As expected, surface water reservoirs were slightly more variable. The pH of all ground water samples was near neutral, while the pH of Bill Evers reservoir water was slightly be low neutral and that of Clear Lake reservoir was generally higher, being around 8.

PAGE 112

100 Table 20. Microbial background measurements from raw water at the two sample sites. Avon Park Well Bill Evers Reservoir Lake Lytal Park Well Clear Lake Reservoir HPCa 1.0 x 107 7.3 x 105 1.1 x 105 2.0 x 106(cfu /100 ml)b 2.0 x 1053.3 x 1068.7 x 1054.0 x 105c 6.1 x 1042.7 x 1061.1 x 105 1.7 x 105Fecal coliform a< 0.526< 0.5849 (cfu /100 ml)b< 175< 1408 c< 185< 1995 Enterococci a< 0.523< 0.5689 (cfu /100 ml) b 1552< 1977 c < 1197< 1230 Coliphage a< 10430< 1050 (pfu / 100 ml) b < 10< 10< 1050 c < 1024< 10< 10 PRD-1 a < 10< 10< 1010 (pfu / 100 ml) b < 10< 10< 10< 10 cNDNDNDND The data in Table 20 show background microbial concentrations of heterotrophic plate count bacteria (HPC), fecal coliform, enterococci, combined somatic and male-specific coliphage (DNA and RNA) and analysis for any background PRD-1. HPC analysis showed a range of over 2 orders of magnitude between the three Avon Park samples. HPC b acteria in the first Avon Park sample were heavily dominated by a single type of small, clear colony and it is unknown why these were present at the initial sampling and absent for subsequent ones. The othe r aquifer samples and surface water samples had HPC counts in the range of 105 to 106 CFU / 100 ml. Water quality indicat or organisms were found in the raw surface water samples and in one instance in a ground wa ter sample (enterococci in the second Avon Park sample). The presence of enterococci in the subsurface at this region seems unlikely; the colonies were confirmed as enterococci by esculin iron reduction, but it is not known if the presence of these organisms was due to contamination of the well hole, of the samp le, from the presence of enterococci-like organisms in the aquifer, or if the finding was in error. Wate r quality indicator bacteria concentrations were much higher in Clear Lake than Bill Evers reservoir, although coliphage did not show the same trend. Salmonella

PAGE 113

101 typhimurium phage such as PRD-1 were gene rally not present in these samples, and were not assayed for the F+ RNA coliphage survival samples, since PRD-1 were not used for these trials. Table 21 shows the microbial counts in natural water samples after treatment. Large reductions in HPC counts were achieved, and all water quality i ndicator organism counts were reduced to below detection limits. Table 21. Background microbial concentrations af ter pasteurization of aq uifer and reservoir water. Avon Park Well Bill Evers Reservoir Lake Lytal Park Well Clear Lake Reservoir HPC a 1880096< 1 (cfu /100 ml)b 239003677 c 50.576011.52400 Fecal coliform a < 0.5< 0.5< 1< 1 (cfu /100 ml)b< 1< 1< 1< 1 c< 1< 1< 1< 1 Enterococci a < 0.5< 0.5< 1< 1 (cfu /100 ml) b < 1< 1< 1< 1 c < 1< 1< 1< 1 Coliphage a < 10< 10< 10< 10 (pfu / 100 ml) b < 10< 10< 10< 10 c < 10< 10< 10< 10 PRD-1 a < 10< 10< 10< 10 (pfu / 100 ml)b< 10< 10< 10< 10 cNDNDNDND Fecal coliform Survival studies in natural wa ter sources with fecal coliform we re performed in duplicate with water collected at separate times an d organisms grown in separate batche s. Results of culturable fecal coliform counts at each time point were analyzed in the same way as for the TDS-temperature trials. The resulting statistic obtained for comparison was the nu mber of predicted days until 2-log decline and was based on model curve equations fit to observed data po ints. Charts of data points and curves fit to these data appear in Appendix 2A for Bradenton experiments and 2-C for West Palm Beach experiments. Table

PAGE 114

102 22 and Table 23 contain predicted days for 2-log d eclines under various conditions, for ground water and surface water respectively. It is im portant to consider at this point that these pe riods were determined solely as a comparative statistic to analyze the impacts of various parameters on relative inactivation of examined microbe populations. Since many of the conditions evaluated, such as pasteurized and 5 C samples, would not occur in actual Florida ground wa ter environments, such trial sets should not be considered indicative of actual survival in the field. The predicted days for 2-log inactivation were used for comparison of temperature, past eurization, and water type (reser voir or aquifer) effects.

PAGE 115

103 Table 22. Predicted days for 2-log decline of fecal coliform concentrations in ASR ground water samples. Trial Temp. oC 1Avon Park aq.547 2raw99mean std. dev.73 37 1Avon Park aq.2222 2raw17mean std. dev.20 4 1Avon Park aq.308 2raw12mean std. dev.10 3 1Avon Park aq.563 2pasteurized63mean std. dev.63 0 1Avon Park aq.22110 2pasteurized159mean std. dev.135 35 1Avon Park aq.3051 2pasteurized51mean std. dev.51 0 1Lake Lytal aq.514 2raw18mean std. dev.16 3 1Lake Lytal aq.2245 2raw35mean std. dev.40 7 1Lake Lytal aq.3011 2raw12mean std. dev.12 1 1Lake Lytal aq.5> 200 2pasteurized31mean std. dev.> 116 1Lake Lytal aq.2274 2pasteurized76mean std. dev.75 1 1Lake Lytal aq.3049 2pasteurized54mean std. dev.52 4Water source Days to 2-log inactivation Table 23. Predicted days for 2-log decline of fecal coliform concentrat ions in ASR reservoir samples. Trial Temp. oC 1Bill Evers res.524 2raw71mean std. dev.48 33 1Bill Evers res.224 2raw6mean std. dev.5 1 1Bill Evers res.302 2raw1mean std. dev.2 1 1Bill Evers res.5108 2pasteurized34mean std. dev.71 52 1Bill Evers res.2250 2pasteurized9mean std. dev.30 29 1Bill Evers res.3050 2pasteurized1mean std. dev.26 35 1Clear Lake res.525 2raw26mean std. dev.26 1 1Clear Lake res.226 2raw10mean std. dev.8 3 1Clear Lake res.304 2raw5mean std. dev.5 1 1Clear Lake res.5119 2pasteurized93mean std. dev.106 18 1Clear Lake res.2255 2pasteurized60mean std. dev.58 4 1Clear Lake res.3038 2pasteurized49mean std. dev.44 8Water source Days to 2-log inactivation More experiments resulted in initial growth of the fecal coliform popu lations than for TDStemperature effect trials (Chapter 3), but exclusively in th e pasteurized sub-samples, and these necessitated the use of the polynomial (parabolic) model (Equation 2). Also, as with TDS-te mperature experiments in Chapter 3, some data sets resulted in predicted days for 2-log inactivation in excess 200 days, and these were treated as described in the results for Chapter 3: periods for inactivation were capped at 200 days so as

PAGE 116

104 to capture the differences in magnitude of observed inactivation kinetics over the 28-day experiments but avoid excessively skewing mean values with very long durations based on experiments that were over an order of magnitude shorter. ANOVA were performed on these modeling results in order to compare the effects of various factors in natural water experiments. ANOVA tests were done using varying levels of categorization of the conditions evaluated, so as to clar ify differences between trends in su rface and ground water, or in raw water only. Appendix 4-F contains results of fecal coliform statistical tests fo r the aquifer and reservoir water studies. In a comparison of temperature and pasteurizati on effects in all natural water sources combined, variability due to both temperature and pasteurization were statistically significant, to the 95% and even 99% level. The graphic in Figure 19 presents the re lative effects of both factor s, indicating a decrease in the number of days and thus reduction of survival at higher temperatures. Also, the raw water resulted in much more rapid decline than treated water for average values across all temperature levels. There was not a statistically significant interacti on between temperature and pasteurizat ion, however, indicating that the impact of treatment was not greater at higher temp eratures, nor was the imp act of temperature more pronounced in raw water rather than treated. This an alysis design combined the results from both surface and ground water, only considering temperature and treatment. Inclusion of water type as a potential factor in a 3-way ANOVA was also performed (Appendix 4-F). The significant factors when water type, pasteurization, temperature, and their interactions were considered were all thr ee independent factors, but no interactions (p < 0.05). Thus, inactivation was si gnificantly different, greater in this case, in surface water than in ground water.

PAGE 117

105 Figure 19. Mean days and confiden ce intervals for 2-log decline of fe cal coliform in ASR water trials (surface and ground water combined). Individual 95% CI temp C Mean ------+---------+---------+---------+----5 64.7 (-------*--------) 22 46.1 (-------*-------) 30 24.9 (-------*--------) ------+---------+---------+---------+----20.0 40.0 60.0 80.0 Individual 95% CI treated Mean ------+---------+---------+---------+----no 21.8 (------*------) yes 68.6 (-----*------) ------+---------+---------+---------+----20.0 40.0 60.0 80.0 For evaluating actual behavior in real situations, the raw water trials are much more important. Thus, analysis of temperature and water type as importa nt factors in fecal coliform survival in only the raw water sources was also performed. In this comparison, only temperature was statistically significant (p < 0.01). The mean for numbe r of days across all temperatures was greater in ground water than surface water, as was found when considering both pasteurized and raw water sources. Means were 15 days for surface water and 28 days for ground water (average of predicted days from all temperatures). However, this difference was not statistically significant. If only the two higher temperatures of 22 and 30 were considered, as would be observed in the Floridan aquifer system, great er differences were observed due to water type. Temperature, water type, and the interaction were all statistically significant to the 95 % level. As Figure 20 shows, inactivation is quite a bit more rapid at 30 than 22 in raw water. Also, inactivation in surface water was more rapid than in ground water. The water type-temperature interaction indicated that the differences between 22 and 30 were relatively greater in ground water than in surface water. As may be seen in Table 22 and Table 23, inactivation of fecal co liform was fairly rapid at 22 in surface water, and thus the decrease in days for 2log inactivation between 22 and 30 was relatively more pronounced in ground water samples.

PAGE 118

106 Figure 20. Mean days and confid ence intervals for 2-log decline of fecal coliform in raw water at 22and 30C (surface and gr ound water combined). Individual 95% CI temp C Mean --------+---------+---------+---------+--22 18.1 (-------*--------) 30 6.9 (-------*--------) --------+---------+---------+---------+--6.0 12.0 18.0 24.0 Individual 95% CI type Mean -+---------+---------+---------+---------+ ground 20.3 (------*------) surface 4.7 (------*------) -+---------+---------+---------+---------+ 0.0 7.0 14.0 21.0 28.0 The intricacies of introduced fecal coliform survival in thes e water sources can be further described by considering the behavior in surface water and ground water separately. This is appropriate since analysis of variance on comb ined results determined a statisti cally significant difference between survival in these water types. Tables of the resu lts from these ANOVA are in Appendix 4-F as well. Figure 21 shows mean days from an examination of temperature and pa steurization effects in surface water sources only, combining the results from Clear Lake and Bill Evers reservoirs. In Figure 22 are the results for analysis on temperature in raw surface water only. Figure 21. Mean days with confid ence intervals for 2-log decline of fecal coliform in surface water: evaluation of temperature a nd pasteurization effects. Individual 95% CI temp C Mean ---------+---------+---------+---------+-5 62.5 (-------*--------) 22 25.0 (--------*-------) 30 18.8 (-------*--------) ---------+---------+---------+---------+-20.0 40.0 60.0 80.0 Individual 95% CI treatedMean ----------+---------+---------+---------+no 15.3 (------*------) yes 55.5 (------*------) ----------+---------+---------+---------+20.0 40.0 60.0 80.0

PAGE 119

107 Figure 22. Mean days for 2-log decline of fecal coliform in raw surface water. Individual 95% CIs For Mean Based on Pooled StDev Temp C N Mean StDev -------+---------+---------+--------5 4 36.50 23.01 (------*-------) 22 4 6.50 2.52 (------*-------) 30 4 3.00 1.83 (-------*------) -------+---------+---------+--------Pooled StDev = 13.41 0 20 40 When considering both treated and raw surface wa ter, pasteurization and temperature significantly affected survival, and in raw surface water, the temperature effect was still significant (p < 0.05 in all cases). With fecal coliform, in both the raw and comb ined surface water sources, th e grouping of days for 2-log inactivation is such that the mean number of da ys at 22 and 30 are more closely associated, while the number is considerably greater at 5 C. Graphics of analyses for ground water are shown in Figure 23 and Figure 24, while the tables of statistical values are in Appendix 4-F. While pasteu rization produced a statistically significant impact in ground water, as well, temperature was not significant to the 95% or 90% level in combined or raw ground water. However, it is evident that the means for nu mber of days to 2-log in activation show the typical temperature-related trend, with shorter periods for inac tivation at higher temperature. Also, it is important to consider Figure 20 in this inst ance. When only the two higher temp eratures are considered, temperature effects were statistically significant in raw water, with noticeable differe nces between the relative effects in surface vs. ground water. The mean days for 2-log inactivation at 22 an d 30 C were 30 vs. 10 days in ground water, while only 6 vs. 3 in surface water. When considering all tested temperatures, trials at 5 C in raw water showed greater variability, at least in Avon Park ground water and Bill Evers reservoir samples (Table 22 and Table 23). This served to incr ease the pooled standard devi ation and resulted in lack of statistical significance for temper ature in ground water with the th ree temperature regime. Table 24 summarizes the comp arison results for fecal coliform.

PAGE 120

108 Figure 23. Days for 2-log decline of fecal coliform (mean and confidence intervals) in raw and pasteurized ground water. Individual 95% CI temp C Mean ---+---------+---------+---------+-------5 78 (----------*-----------) 22 67 (----------*-----------) 30 31 (----------*----------) ---+---------+---------+---------+-------0 35 70 105 Individual 95% CI treated Mean --+---------+---------+---------+--------no 28 (---------*----------) yes 89 (----------*---------) --+---------+---------+---------+--------0 30 60 90 Figure 24. Mean days for 2-log decline of fecal coliform in raw ground water. Individual 95% CIs For Mean Based on Pooled StDev Temp C N Mean StDev -------+---------+---------+--------5 4 44.50 39.20 (----------*----------) 22 4 29.75 12.69 (----------*----------) 30 4 10.75 1.89 (---------*----------) -------+---------+---------+--------Pooled StDev = 23.81 0 25 50 Table 24. Statistically significant factors affecting fecal coliform inactivation in ASR water sources. Fecal Coliform Water typeTreatmentComparison parametersSignificant factors (95% unless noted) surface & groundraw and pasteurized2-way: temp, treatmnt temp, treatmnt surface & groundraw and pasteurized3-way: temp, treatmnt, type temp, treatmnt, type surface & groundraw only2-way: temp, type temp surface & ground raw only, 22o and 30o C 2-way: temp, type temp, type, interact. surface water raw and pasteurized2-way: temp, treatmnt temp, treatmnt surface water raw only1-way: temp temp groundwaterraw and pasteurized2-way: temp, treatmnt treatmnt groundwaterraw only1-way: temp not sign. (but clear trend)

PAGE 121

109 In a subtropical environment such as Florida, the two higher temperature levels of 22 and 30 are more of a concern in regards to survival and inactiv ation of contaminant microo rganisms. To graphically summarize inactivation trends in the four natural wate r sources (raw), Figure 25 displays the mean number of days for 2-log inactivation as listed in Table 22 and Table 23. Error bars represent the standard deviation of each mean value. It is important to recall that as the bars represent number of days, the smallest bars signify most rapid inactivation. Figure 25 graphically shows the trends described above, for raw water only, wherein it is apparent that inactivation is more rapid in surface water at both temperatures than in ground water, and for both types of water, in activation increases on averag e at 30 over 22 C. Figure 25. Chart of mean days for 2-log inactivation for fecal coliform in raw aquifer and surface water at ambient temperatures for Florida. Fecal coliform 0 10 20 30 40 50 2230 Temperature (oC)Days for 2-log10 (99%) decline Avon Park aq. Lake Lytal aq. Groundwater Bill Evers res. Clear Lake res. Surface water Due to the importance of actual inactivation rates under the raw water conditions as shown above, in contrast to the artificial situations created for co mparative analysis of effect s on survival, first-order inactivation rates were also determined for the combined data from each s ite. Charts of the raw water data

PAGE 122

110 points combined for each aquifer and reservoir source are given in Appendices 2-B and 2-D. Linear regression analysis of the combined points at all th ree temperatures are shown on these charts, and these rates and corresponding estimated days until 2-log and 3-log inactivation are shown here in Table 25. These rates are helpful for comparison to other studies, as first-order inactivation rates are often used to express results from studies such as these. Since da ta points from both trials in each water source were modeled for the linear regression together, the rate expre ssed here is essentially an average of data from the two replicates of each condition. Table 25. First-order inactivation rates of fecal coliform in raw aquifer an d surface water sources. Temperature (C)Avon Park aq.5-0.02195143 22-0.1012030 30-0.1661218 Lake Lytal aq.5-0.1411421 22-0.0653146 30-0.1491320 Bill Evers res.5-0.0523858 22-0.42057 30-1.01223 Clear Lake res.5-0.0663045 22-0.1731217 30-0.296710Linear inact. rate (log10/d) Est. days to 99% decline Est. days to 99.9% decline Water source Enterococci bacteria Enterococci survival studies in the aquifer an d surface water sources were performed alongside fecal coliform studies. The data obtained from enteroco cci experiments were analyzed in the same way as the fecal coliform, and results were compiled in the same types of tables and figu res. Figures of sample point counts and model curves appear in Appendices 2-E and 2-G. Table 26 and Table 27 display predicted days until 2-log decline for enterococci in gr ound water and surface water in both raw and pasteurized conditions.

PAGE 123

111 Table 26. Predicted days for 2-log decline of enterococci concentrations in ASR ground water samples. Trial Temp. oC 1Avon Park aq.560 2raw40mean std. dev.50 14 1Avon Park aq.2217 2raw5mean std. dev.11 8 1Avon Park aq.307 2raw2mean std. dev.5 4 1Avon Park aq.564 2pasteurized36mean std. dev.50 20 1Avon Park aq.2219 2pasteurized3mean std. dev.11 11 1Avon Park aq.308 2pasteurized2mean std. dev.5 4 1Lake Lytal aq.5175 2raw> 200mean std. dev.> 188 1Lake Lytal aq.2232 2raw35mean std. dev.34 2 1Lake Lytal aq.3013 2raw16mean std. dev.15 2 1Lake Lytal aq.5158 2pasteurized115mean std. dev.137 30 1Lake Lytal aq.2241 2pasteurized56mean std. dev.49 11 1Lake Lytal aq.3010 2pasteurized18mean std. dev.14 6Water source Days to 2-log inactivation Table 27. Predicted days for 2-log decline of enterococci concentrations in ASR reservoir samples. Trial Temp. oC 1Bill Evers res.529 2raw37mean std. dev.33 6 1Bill Evers res.224 2raw3mean std. dev.4 1 1Bill Evers res.302 2raw1mean std. dev.2 1 1Bill Evers res.5177 2pasteurized73mean std. dev.125 74 1Bill Evers res.2233 2pasteurized2mean std. dev.18 22 1Bill Evers res.304 2pasteurized1mean std. dev.3 2 1Clear Lake res.528 2raw29mean std. dev.29 1 1Clear Lake res.226 2raw7mean std. dev.7 1 1Clear Lake res.304 2raw4mean std. dev.4 0 1Clear Lake res.5148 2pasteurized125mean std. dev.137 16 1Clear Lake res.2238 2pasteurized22mean std. dev.30 11 1Clear Lake res.3012 2pasteurized8mean std. dev.10 3Water source Days to 2-log inactivation The same types of analyses of variance as we re performed for fecal co liform were performed for enterococci. Appendix 4-G displays the results of these analyses, first of a two-way test for water temperature and pasteurization treatment effects in all natural water sources, a three-way test for these two factors and water type, and then a two-way test for th e significance of temperature and water type in only

PAGE 124

112 raw waters. These test results reveal some significant relationships and interactions regarding enterococci inactivation. Temperature was strongly significan t (p < 0.01) in all these statistical tests. Figure 26. Enterococci inactivation in ASR water samples. Predic ted days for 2-log inactivation from both water types averaged by temperature and treatment. Individual 95% CI temp C Mean ----+---------+---------+---------+------5 93.4 (-----*-----) 22 20.2 (-----*-----) 30 7.0 (-----*------) ----+---------+---------+---------+------0.0 30.0 60.0 90.0 Individual 95% CI treated Mean -------+---------+---------+---------+---no 31.5 (-----------*------------) yes 48.9 (------------*-----------) -------+---------+---------+---------+---24.0 36.0 48.0 60.0 The diagrams of mean 2-log inactivation periods in Figure 26 indicate the temperature effect when considering combined averages from surface and ground water. Also, a trend in grouping of means by temperature is visible, similar to that seen in severa l fecal coliform tests, in wh ich the means at 22 and 30 were more closely grouped, and the average at 5 wa s considerably greater. When all water sources and treatments were considered in a two-way ANOVA with temperature and pasteurization as the factors (such as Figure 26 above), pasteurization was not quite significant at the 90% level, although it was nearly so (p = 0.108). When more factors and interactions were introduced to explain variability in a 3-way ANOVA, pasteurization treatment (averaged across all temperat ure levels) was significant to the 90% level. Inactivation was more rapid in raw water than in pasteu rized. In addition, the interactions of water type and treatment, and the interaction of water type, treat ment, and temperature were both significant to the 95% level. However, water type as an independent factor was not statistically significant. Pasteurization reduced inactivation in surface water to a much greater degree than in ground water. The temperature interaction with pasteurization effects and water type was interesting in that relative differences between raw and pasteurized conditions increased with increasin g temperature in surface wa ter, but not in ground water. This was determined from analysis of mean inactivation periods from each condition set as computed from values in Table 26 and Table 27.

PAGE 125

113 When considering raw water only, both temperature and water type were significant to the 95% level, and the interaction of the two factors was at the 90% level. Inactivation was more rapid in surface water, and this difference was relatively more pronounce d at 5 C than at the higher temperatures (hence the statistically-significant interaction of water type an d temperature). If only te mperatures of 22 and 30 were considered, in the raw state, temperature was only statistically significant to the 90% level. As shown in Figure 27, inactivation was more rapid at 30 than 22. Also, inactivation was still considerably more rapid in surface water at these temperatures than in ground water, and this trend was significant to the 95% level. The interaction of the two parameters was not statistically significant however. Figure 27. Enterococci inactivation as days for 2-log decline in raw water at 22and 30C (mean days and confidence intervals). Averag ed by temperature and water type. Individual 95% CI temp C Mean ----------+---------+---------+---------+22 13.6 (-----------*-----------) 30 6.1 (-----------*-----------) ----------+---------+---------+---------+5.0 10.0 15.0 20.0 Individual 95% CI type Mean ----+---------+---------+---------+------ground 15.9 (--------*---------) surface 3.9 (--------*---------) ----+---------+---------+---------+------0.0 6.0 12.0 18.0 Results of ANOVA performed independently on the two types of water are given in Appendix 4-G also. As with fecal coliform, both types were an alyzed independently, while also considering both combined raw and pasteurized and raw-only conditions. In surface water, temperature, treatment, and the interaction of the two factors were all strongly significant (p < 0.01), and as expected so was temperature when considering only raw surface wa ter (p < 0.01). However, in gr ound water, only temperature was significant (p < 0.01), while pasteurization did not produ ce a significant effect. As can be seen in Figure 28, the means are very close together between treated and raw water experiments, and there is almost total overlap of the confidence intervals. The difference between ground water and surface water in significance of treatment and the interaction of temperature an d treatment reveals the source of the interactive dependence on water type demonstrated when consid ering the 3-way ANOVA comp arison. It is evident

PAGE 126

114 that reduction of the native microbial populations enhanced enterococci survival in surface water more so than in the ground water sources. Table 28 summ arizes results of analys es for factors affecting enterococci survival. Figure 28. Temperature and pasteurization effects on enterococci mean days for 2-log decline in ground water. Individual 95% CI treated Mean --------+---------+---------+---------+--no 50.2 (---------------*----------------) yes 44.2 (---------------*----------------) --------+---------+---------+---------+--30.0 45.0 60.0 75.0 Individual 95% CI temp C Mean ------+---------+---------+---------+----5 106 (-------*------) 22 26 (-------*------) 30 10 (------*-------) ------+---------+---------+---------+----0 40 80 120 Table 28. Significant variables affecting en terococci survival in ASR water sources. Enterococci Water typeTreatmentComparison parametersSignificant factors (95% unless noted) surface & groundraw and pasteurized2-way: temp, treatmnt temp surface & groundraw and pasteurized3-way: temp, treatmnt, type temp, treatmnt (90%), type-tmnt interact, 3way interact surface & groundraw only2-way: temp, type temp, type, interatction (90%) surface & ground raw only, 22o and 30o C 2-way: temp, type temp (90%), type surface water raw and pasteurized2-way: temp, treatmnt temp, treatmnt, interaction surface water raw only1-way: temp temp groundwaterraw and pasteurized2-way: temp, treatmnt temp groundwaterraw only1-way: temp temp

PAGE 127

115 Figure 29 graphically displays the means of days for 2-log decline for enterococci as in Figure 25 for fecal coliform. The averages at both temperatures were greater in La ke Lytal Park water than in Avon Park water. This site-related difference was also the case between Clear Lake reservoir and Bill Evers reservoir. The column chart also demonstrates the more rapid inactivation in surface water compared to ground water at the respective temper atures, and the more rapid decline at 30 C than 22 C for each water type. Figure 29. Chart of mean days for 2-log inactiva tion for enterococci in raw aquifer and surface water at ambient temperatures for Florida. Enterococci 0 10 20 30 40 50 2230 Temperature (oC)Days for 2-log10 (99%) decline Avon Park aq. Lake Lytal aq. Groundwater Bill Evers res. Clear Lake res. Surface water Charts of combined data points from experiments in each water source, from which average firstorder inactivation rates were determined by linear re gression analysis, are shown in Appendices 2-F and 2H. These rates and corresponding estimated days until 2-log and 3-log inactivation are shown here in Table 29.

PAGE 128

116 Table 29. First-order in activation rates of enterococci in raw aquifer and surface water sources. Temperature (C)Avon Park aq.5-0.0603350 22-0.1621219 30-0.253812 Lake Lytal aq.5-0.011182273 22-0.0623248 30-0.1321523 Bill Evers res.5-0.0494161 22-0.37758 30-0.77434 Clear Lake res.5-0.0543756 22-0.265811 30-0.50146Water source Linear inact. rate (log10/d) Est. days to 99% decline Est. days to 99.9% decline F+ RNA coliphage Plots of log-transformed survival ratios for F+ RNA coliphage and resulting best-fit model curves for natural water trials are in Appendix 3-A and 3-C. Predicted days for 2-log decline are given in Table 30 and Table 31. Unlike the experiments for the other four types of organisms, duplicate trials for the RNA coliphage were performed simultaneously. Therefor e, the variability that would be associated with obtaining water samples at different times and from using different batches of cultivated test organisms was not present in the results. Thus, not unexpectedly, the overall variability between tr ials shown in the results given in these tables was generally much less than was present with the other sets of organisms. The same types of ANOVA comparisons as were done for the b acteria were performed for F+ RNA coliphage 2-log days and tables showing these are given in Appendix 4-H.

PAGE 129

117 Table 30. Predicted days for 2-log decline of F+ RNA coliphage concentrations in ASR ground water samples. Trial Temp. oC1Avon Park aq.55 2raw3mean std. dev.4 1 1Avon Park aq.221 2raw3mean std. dev.2 1 1Avon Park aq.301 2raw1mean std. dev.1 0 1Avon Park aq.528 2pasteurized46mean std. dev.37 13 1Avon Park aq.223 2pasteurized5mean std. dev.4 1 1Avon Park aq.301 2pasteurized1mean std. dev.1 0 1Lake Lytal aq.533 2raw34mean std. dev.34 1 1Lake Lytal aq.224 2raw4mean std. dev.4 0 1Lake Lytal aq.301 2raw1mean std. dev.1 0 1Lake Lytal aq.548 2pasteurized41mean std. dev.45 5 1Lake Lytal aq.224 2pasteurized5mean std. dev.5 1 1Lake Lytal aq.301 2pasteurized1mean std. dev.1 0Water source Days to 2-log inactivation Table 31. Predicted days for 2-log decline of F+ RNA coliphage concentrations in ASR surface water samples. Trial Temp. oC 1Bill Evers res.5118 2raw82mean std. dev.100 25 1Bill Evers res.227 2raw7mean std. dev.7 0 1Bill Evers res.303 2raw2mean std. dev.3 1 1Bill Evers res.560 2pasteurized51mean std. dev.56 6 1Bill Evers res.2211 2pasteurized12mean std. dev.12 1 1Bill Evers res.303 2pasteurized3mean std. dev.3 0 1Clear Lake res.521 2raw25mean std. dev.23 3 1Clear Lake res.222 2raw2mean std. dev.2 0 1Clear Lake res.301 2raw1mean std. dev.1 0 1Clear Lake res.522 2pasteurized24mean std. dev.23 1 1Clear Lake res.222 2pasteurized2mean std. dev.2 0 1Clear Lake res.301 2pasteurized1mean std. dev.1 0Water source Days to 2-log inactivation When both pasteurized and raw water of both types (surface and ground water) were considered, temperature was strongly significant in a two-way test, while treatment and the interaction of the two did not result in any significant effects. If water type wa s also considered in a three-way test, temperature was still determined to be significant (P < 0.05), while wa ter type was significant to the 90% level. Also, the three-way interaction of temperature, water type, an d treatment was significant to the 90% level. But differences due to water type and treatment were primar ily due to the differences at 5 C, and a consistent trend was not present. Pasteurization increased survival at 5 in ground water and decr eased survival at 5 in surface water. Thus, while pasteurization produced a statistically-significant re duction in in activation in

PAGE 130

118 ground water and did not in surface water, this trend was due to variability at lo w temperature and is of limited importance when considering our environment of concern. In raw water of both types, temperature was once again a very significant factor (p < 0.05), and water type and the interaction of water type and temperat ure were significant to the 90% level. In contrast to both types of bacteria (fecal coliform and enteroco cci), the relationship of water type to survival was reversed. As the diagram of means in Figure 30 shows, inactivation was more rapid for the RNA coliphage in raw ground water than in surface water. But wh ile survival was less on average in ground water compared to surface water in both raw and combined ra w and pasteurized analyses, this effect was due to differences at 5 more so than at the higher temperatur es of concern to the Florida subsurface. This may be inferred from examining mean 2-log inactivation days from Table 30 and Table 31. The pattern of temperature response was the same as that observed in many comparisons with the bacteria in natural water sources, with days for 2log decline at 22 and 30 closely grouped, and much longer times required at 5 C. Due to this grouping, there is question of whether differences of inactivation at these two higher temperatures are statistically different. An ANOVA for raw water data at only 22 and 30 C indicated that the difference between inactiv ation at the higher temperatures was statistically significant to the 95% level, but in reality there was little difference considering the time frame of the experiment and sample points. Mean days for 2-log decline were 3.75 at 22 and 1.4 at 30 C. The effect of water type at these temperatures was not significant. From this comparison, it is evident that water type was a statistically significant variable due to differ ences at 5 C, hence the temperature-water type interdependence when considering all temperatures. Figure 31 displays mean days for 2-log inactivation at these temperatures. Although there were differences between inactivation periods among these conditions, taken as a whole, these values were quite rapid compared to other organisms. All times were equal to or less than 1 week in raw water, regardless of wate r type or temperature (between 22 and 30).

PAGE 131

119 Figure 30. F+ RNA coliphage days for 2-log decline in raw surface and ground water, means are average by temperature and water type. Individual 95% CI temp C Mean -------+---------+---------+---------+---5 40.1 (-------*-------) 22 3.7 (-------*------) 30 1.4 (-------*------) -------+---------+---------+---------+---0.0 20.0 40.0 60.0 Individual 95% CI type-raw Mean -----+---------+---------+---------+-----ground 7.6 (------------*-----------) surface 22.6 (------------*-----------) -----+---------+---------+---------+-----0.0 10.0 20.0 30.0 Figure 31. Mean days and confiden ce intervals for 2-log decline of F+ RNA coliphage in raw surface and ground water at 22 and 30 C. Individual 95% CI temp C Mean ----------+---------+---------+---------+22 3.75 (----------*----------) 30 1.38 (---------*----------) ----------+---------+---------+---------+1.20 2.40 3.60 4.80 Individual 95% CI type raw Mean ---+---------+---------+---------+-------ground 2.00 (------------*------------) surface 3.12 (------------*------------) ---+---------+---------+---------+-------1.00 2.00 3.00 4.00 From data in Table 30 and Table 31, virtually no difference was observed between pasteurized and raw water inactivation at 22 and 30 C. This is wh y treatment was not a statistically significant variable independently in a 3-way ANOVA, but the three-way in teraction of water type, treatment, and temperature was determined to have a significant effect. Mean days for 2-log decline at the higher temperatures were 3 days and 4 days in raw and pasteurized ground water respectively at 22 C, and 1 day in both raw and pasteurized samples at 30 (calculation results not shown on tables). In surface water, values at 22 C were 5 days in raw and 7 in pasteurized, and 2 days in both at 30 C. It becomes apparent that at these temperatures, inactivation in the two water types is both similar and rapid, regardless of treatment, and there is a slight increase in survival at 22 over 30 C. Comparisons of factors affecting in activation were also broken down into the separate water types. The tables of these ANOVA tests are also in Append ix 4-H. In surface water, temperature remained strongly significant (p < 0.01), but pasteurization was not. As a graphic of the mean days for these

PAGE 132

120 coliphage shows (Figure 32), grouping with respect to temperature was similar to that seen in most cases for the bacteria in these water sources and for F+ RNA phage in raw water of both types. Also, the mean values for all temperatures compared by pasteurization treatment were cl ose together at 16 and 22 days, with considerable overlap of confidence intervals. Figure 32. Temperature and pasteurization effects on days for 2-log decline of F+ RNA coliphage in surface water, mean and confidence inter vals by temperature and pasteurization. Individual 95% CI temp C Mean -------+---------+---------+---------+---5 50.4 (-------*-------) 22 5.6 (-------*-------) 30 1.9 (-------*-------) -------+---------+---------+---------+---0.0 20.0 40.0 60.0 Individual 95% CI treatedMean ------+---------+---------+---------+----no 22.6 (---------------*---------------) yes 16.0 (---------------*---------------) ------+---------+---------+---------+----8.0 16.0 24.0 32.0 In ground water, both when combining treated an d raw, and in raw only, the temperature effect was the same and was significant, although only at th e 90% level in raw ground water. As Figure 33 and Figure 34 below show, there was a large difference be tween the means at 5 compared to the two higher temperatures, but the large standard deviation at 5 in raw ground water (Figure 34) reduced the significance level assigned to this effect. Also, pasteurization created a significant impact on inactivation in ground water (p < 0.05), and the effect was inte rrelated with temperature (interaction was significant with p < 0.05). Once again this interaction is due to a larger difference between pasteurized and raw water survival at 5 compared with the effect at environmenta l temperatures (2 2 and 30).

PAGE 133

121 Figure 33. Temperature and pasteurization effects on F+ RNA coliphage inactivation in ground water, as mean days for 2-log decline. Individual 95% CI temp C Mean -----+---------+---------+---------+-----5 29.8 (-----*-----) 22 3.6 (-----*----) 30 1.0 (-----*-----) -----+---------+---------+---------+-----0.0 10.0 20.0 30.0 Individual 95% CI treated Mean -----+---------+---------+---------+-----no 7.6 (--------*---------) yes 15.3 (---------*--------) -----+---------+---------+---------+-----5.0 10.0 15.0 20.0 Figure 34. Mean days and confidence intervals for 2-log decline of F+ RNA coliphage in raw ground water. Individual 95% CIs For Mean Based on Pooled StDev Temp C N Mean StDev ---------+---------+---------+------5 4 18.750 17.056 (---------*--------) 22 4 3.000 1.414 (--------*---------) 30 4 1.000 0.000 (--------*--------) ---------+---------+---------+------Pooled StDev = 9.881 0 12 24

PAGE 134

122 Table 32. Significant variables affecting F+ RNA coliphage survival in ASR water sources. F+ RNA coliphage Water typeTreatmentComparison parametersSignificant factors (95% unless noted) surface & groundraw and pasteurized2-way: temp, treatmnt temp surface & groundraw and pasteurized3-way: temp, treatmnt, type temp, type (90%), 3-way interact (90%) surface & groundraw only2-way: temp, type temp, type (90%), interact (90%) surface & ground raw only, 22o and 30o C 2-way: temp, type temp surface water raw and pasteurized2-way: temp, treatmnt temp surface water raw only1-way: temp temp groundwaterraw and pasteurized2-way: temp, treatmnt temp, treatmnt, interact groundwaterraw only1-way: temp temp (90%) The mean number of days predicted for 2-log d ecline at temperatures of 22 and 30 C in raw water sources are shown in Figure 35. The means displayed in this column chart underscore the more rapid inactivation observed with F+ RNA coliphage than for fecal coliform or enterococci overall, particularly in ground water. Data from duplicates for each cond ition were combined and linear regression rates were obtained, and these charts are displayed in Appe ndices 3-B and 3-D. The first-order rates and corresponding days predicted for 2and 3-log decline are shown in Table 33.

PAGE 135

123 Figure 35. Chart of mean days for 2-log inactivation for F+ RNA coliphage in raw aquifer and surface water at ambient temperatures for Florida. F+ RNA coliphage 0 10 20 30 40 50 2230 Temperature (oC)Days for 2-log10 (99%) decline Avon Park aq. Lake Lytal aq. Groundwater Bill Evers res. Clear Lake res. Surface water Table 33. First-order inactivation rates of F+ RNA coliphage in raw a quifer and surface water sources. Temperature (C)Avon Park aq.5-0.271711 22-0.50946 30-1.59412 Lake Lytal aq.5-0.0643147 22-0.44847 30-2.42711 Bill Evers res.5-0.0484263 22-0.249812 30-0.63335 Clear Lake res.5-0.0912233 22-0.93823 30-1.99912Water source Linear inact. rate (log10/d) Est. days to 99% decline Est. days to 99.9% decline

PAGE 136

124 DNA coliphage The predicted days for 2-log decline from DNA coliphage (combined male-specific and somatic) experiments in natural water sources are shown in Ta ble 34 and Table 35, while charts of the data and model curves are in Appendices 3-E and 3-G. In several cases, inactivation of DNA coliphage was quite a bit slower in natural water than F+ RNA coliphage and the two bacterial populations. In five cases, all in the pasteurized water sources, inactiv ation was so slow that the days predicted for 2-log decline were capped and the maximum for these analyses of variance of 200 days was used for comparison. Once again, pasteurized conditions are unnatural and used solely for comparison of the effect of reducing the native microbial population on survival of introduced organisms.

PAGE 137

125 Table 34. Predicted days for 2-log decline of DNA coliphage concentrations in ASR ground water samples. Trial Temp. oC 1Avon Park aq.591 2raw190mean std. dev.141 70 1Avon Park aq.2245 2raw28mean std. dev.37 12 1Avon Park aq.3016 2raw10mean std. dev.13 4 1Avon Park aq.5112 2pasteurized71mean std. dev.92 29 1Avon Park aq.2290 2pasteurized44mean std. dev.67 33 1Avon Park aq.3016 2pasteurized39mean std. dev.28 16 1Lake Lytal aq.562 2raw82mean std. dev.72 14 1Lake Lytal aq.2241 2raw24mean std. dev.33 12 1Lake Lytal aq.3014 2raw10mean std. dev.12 3 1Lake Lytal aq.5> 200 2pasteurized78mean std. dev.> 139 1Lake Lytal aq.22> 200 2pasteurized28mean std. dev.> 114 1Lake Lytal aq.3098 2pasteurized22mean std. dev.60 54Water source Days to 2-log inactivation Table 35. Predicted days for 2-log decline of DNA coliphage concentrations in ASR surface water samples. Trial Temp. oC 1Bill Evers res.530 2raw114mean std. dev.72 59 1Bill Evers res.2211 2raw19mean std. dev.15 6 1Bill Evers res.3011 2raw5mean std. dev.8 4 1Bill Evers res.5> 200 2pasteurized147mean std. dev.> 174 1Bill Evers res.22> 200 2pasteurized27mean std. dev.> 114 1Bill Evers res.30155 2pasteurized2mean std. dev.79 108 1Clear Lake res.582 2raw156mean std. dev.119 52 1Clear Lake res.2219 2raw28mean std. dev.24 6 1Clear Lake res.3013 2raw6mean std. dev.10 5 1Clear Lake res.5> 200 2pasteurized149mean std. dev.> 175 1Clear Lake res.22> 200 2pasteurized170mean std. dev.> 185 1Clear Lake res.30163 2pasteurized102mean std. dev.133 43Water source Days to 2-log inactivation Results of ANOVA tests on both surface and ground water sources are shown in Appendix 4-I. When considering both water types, strong significance may be attributed to the effects of temperature and pasteurization (p < 0.01), but no significant inter action between the two factors was apparent. Higher temperature decreased survival, and pa steurization increased survival on av erage. Inclusion of water type as a variable in a three-way ANOVA revealed th at water type was significant at the 90% level

PAGE 138

126 independently, and the interaction of water type and tr eatment was significant to the 95% level. As with the F+ RNA coliphage, inactivation of DNA coliphage was more rapid on average in ground water than in surface water when considering both pasteurized an d raw water across the three temperature regime. However, there were larger differe nces at each temperature between raw and pasteurized conditions in surface water than in ground water, pa rticularly in surface water at higher temperatures, thus the interaction of water type and treatment. But while temperature was significant as an independent factor, the interactive effect of temperature on the relative difference due to pasteurization between the two water types was not significant (no three-way interacti on as with enterococci). The di fferences due to temperature and treatment in a two-way comparison may be seen in Figure 36. Figure 36. DNA coliphage inactivation in ASR wa ter samples, showing e ffects of temperature and treatment on mean days for 2-log d ecline (with confidence intervals). Individual 95% CI temp C Mean ------+---------+---------+---------+----5 123 (------*-------) 22 73 (-------*------) 30 43 (------*-------) ------+---------+---------+---------+----35 70 105 140 Individual 95% CI treated Mean --+---------+---------+---------+--------no 46 (------*-------) yes 113 (------*------) --+---------+---------+---------+--------30 60 90 120 When considering only raw water samples, temperat ure was statistically-significant (P < 0.05), but water type and the interaction were not. Figure 37 di splays mean inactivation periods for raw water trials, wherein it is apparent that there was little differen ce between water types in raw water, but mean inactivation periods in surface water were slightly shor ter when all temperatures were combined. Thus, the differences due to water type that were statistically significant in a 3-way ANOVA with both pasteurized and raw water were due to differences in the pasteu rized water samples between water types. Also, the trend was reversed; inactivation in raw water was more rapid in surface water than in ground water. In Figure 37 the large separation of mean inactivation pe riods between 5 and the two temperatures found in the Floridan aquifer system is apparent. If only th ese two higher temperatures are compared, the mean 2-

PAGE 139

127 log inactivation period was still signi ficantly shorter at 30 vs. 22 C, and inactivation in surface water was significantly faster than in ground water (p < 0.05 for both). Figure 38 depicts these trends graphically. Figure 37. Days for 2-log decline for DNA coliphage in raw surface and ground water, averaged by temperature and water type. Individual 95% CI type Mean ---+---------+---------+---------+-------ground 51.1 (----------------*---------------) surface 41.2 (---------------*----------------) ---+---------+---------+---------+-------24.0 36.0 48.0 60.0 Individual 95% CI temp C Mean ----+---------+---------+---------+------5 101 (------*------) 22 27 (------*------) 30 11 (------*------) ----+---------+---------+---------+------0 35 70 105 Figure 38. Mean days for 2-log decline of DNA coliphage in raw water at 22 and 30 C, averaged by temperature and water type (with confidence intervals). Individual 95% CI temp C Mean ---+---------+---------+---------+-------22 26.9 (------*-------) 30 10.6 (------*------) ---+---------+---------+---------+-------7.0 14.0 21.0 28.0 Individual 95% CI water type Mean ---+---------+---------+---------+-------ground 23.5 (---------*---------) surface 14.0 (---------*---------) ---+---------+---------+---------+-------10.0 15.0 20.0 25.0 Results of statistics for comparisons of the tw o water types taken independently are also in Appendix 4-I. In surface water and ground water, in both comparisons (raw and pasteurized or raw only), temperature was significant to the 95% level. The in teraction of water type and treatment is explained further in these tests; for surface water sources, past eurization had a significant effect on inactivation (p< 0.05), but was not significant to the 90% level or greater in ground water (p = 0.12). Nonetheless survival

PAGE 140

128 still increased on average in pasteurized ground wate r. This difference was more notable at higher temperatures (Table 34). Mean days for 2-log in activation from these comparisons are shown here in Figure 39, Figure 40, Figure 41, and Figure 42. A summary of statistically-significant factors for DNA coliphage decline are summarized in Table 36. Figure 39. DNA coliphage mean days for 2-log inac tivation in raw and past eurized surface water. Individual 95% CI temp C Mean ------+---------+---------+---------+----5 135 (---------*--------) 22 84 (---------*---------) 30 57 (--------*---------) ------+---------+---------+---------+----40 80 120 160 Individual 95% CI treatedMean --------+---------+---------+---------+--no 41 (-------*-------) yes 143 (-------*-------) --------+---------+---------+---------+--40 80 120 160 Figure 40. DNA coliphage mean days for 2-log inactivation in raw surface water. Individual 95% CIs For Mean Based on Pooled StDev Temp C N Mean StDev ------+---------+---------+---------+ 5 4 95.50 53.15 (------*------) 22 4 19.25 6.95 (------*------) 30 4 8.75 3.86 (------*------) ------+---------+---------+---------+ Pooled StDev = 31.03 0 50 100 150

PAGE 141

129 Figure 41. DNA coliphage mean days for 2-log inactivation in raw and pasteurized ground water. Individual 95% CI temp C Mean ---+---------+---------+---------+-------5 111 (--------*--------) 22 62 (--------*--------) 30 28 (--------*--------) ---+---------+---------+---------+-------0 40 80 120 Individual 95% CI treated Mean --+---------+---------+---------+--------no 51 (----------*-----------) yes 83 (-----------*-----------) --+---------+---------+---------+--------25 50 75 100 Figure 42. DNA coliphage mean days for 2-log inactivation in raw ground water. Individual 95% CIs For Mean Based on Pooled StDev Temp C N Mean StDev ------+---------+---------+---------+ 5 4 106.25 57.13 (------*-------) 22 4 34.50 10.08 (-------*------) 30 4 12.50 3.00 (-------*------) ------+---------+---------+---------+ Pooled StDev = 33.54 0 50 100 150 Table 36. Significant variables affecting DNA coliphage survival in natural water sources. DNA coliphage Water typeTreatmentComparison parametersSignificant factors (95% unless noted) surface & groundraw and pasteurized2-way: temp, treatmnt temp, treatmnt surface & groundraw and pasteurized3-way: temp, treatmnt, type temp, treatmnt, type (90%), type-trmnt interact surface & groundraw only2-way: temp, type temp surface & ground raw only, 22o and 30o C 2-way: temp, type temp, type surface water raw and pasteurized2-way: temp, treatmnt temp, treatmnt surface water raw only1-way: temp temp groundwaterraw and pasteurized2-way: temp, treatmnt temp groundwaterraw only1-way: temp temp

PAGE 142

130 Figure 43 is a graphic presentation of the mean days predicted for 2-log decline of each water type, in the raw state at 22 and 30 C. This figure shows graphically the large di fference between survival of RNA coliphage (Figure 35) and DNA coliphage in aquifer and reservoir water sources at ambient temperatures found in Florida. Th e mean values at each temperature we re less for surface water than for ground water, as was seen with the indicator bact eria. As the two ANOVA tests on raw water results showed, water type was a statistically significant vari able when considering the two higher temperatures, but not when values at 5 C were included. It appe ars that the difference between the two water types was larger at 22 C. Charts of combined data points fo r raw water at all temperatures are given in Appendices 3-F and 3-H; these combined values of Log N/N0 for each experiment were used for linear regression, to obtain first order inactivation rates as for the other or ganisms. These inactivation rate constants are given in Table 37. Figure 43. Chart of mean days for 2-log inactivation for DNA coliphage in raw aquifer and surface water at ambient temperatures for Florida. DNA coliphage 0 10 20 30 40 50 2230 Temperature (oC)Days for 2-log10 (99%) decline Avon Park aq. Lake Lytal aq. Groundwater Bill Evers res. Clear Lake res. Surface water

PAGE 143

131 Table 37. First-order inactivation rates of DNA coliphage in raw aquifer and surface water sources. Temperature (C)Avon Park aq.5-0.0365683 22-0.0643147 30-0.1321523 Lake Lytal aq.5-0.0355786 22-0.0722842 30-0.1481420 Bill Evers res.5-0.0375481 22-0.1171726 30-0.1671218 Clear Lake res.5-0.017118176 22-0.0922233 30-0.161319Water source Linear inact. rate (log10/d) Est. days to 99% decline Est. days to 99.9% decline PRD-1 bacteriophage The bacteriophage PRD-1 is generally considered more resistant to environmental stresses than most enteric viruses and bacteria. However, it is useful as a tracer of vi rus movement through the subsurface for these reasons and may be a good predictor of survival for the more-resistant hepatitis A virus than F+ RNA coliphage (Blanc and Nasser 1996). Thus, it was evaluated in these trials to gauge its persistence under the conditions as may be found in aquifer injection scenarios in Florida. As was found for temperature-TDS trials in Chapter 3, inactivation of this organism was slower than the other indicator organisms on average. The methods employed to describe inactivation kinetics for other indicator organisms (the use of 2-log inactivation periods) very frequently gave estimated numbers of days in excess of 200, particularly at low temperatures and in pasteu rized samples. Since each e xperiment only lasted for 28 days, a different method was used to express results. Thus, the statistical comparisons of factors affecting survival for PRD-1 employed first-order inactivation rates. Plots of the PRD-1 data from natural water sources are displayed in Appendices 3-I and 3-K. These charts allow inspection of kinetics of decline and show that in most cases, a linear model is the most useful. Table 38 and Table 39 present the inactivation rate constants from these trials.

PAGE 144

132 Table 38. First-order inactivation rates from PRD-1 in ASR ground water. Trial Temp. oC 1Avon Park aq.5-0.001 2raw-0.017mean std. dev.-.009 .011 1Avon Park aq.22-0.011 2raw-0.022mean std. dev.-.017 .008 1Avon Park aq.30-0.010 2raw-0.019mean std. dev.-.015 .006 1Avon Park aq.50.004 2pasteurized-0.026mean std. dev.-.011 .021 1Avon Park aq.220.006 2pasteurized-0.016mean std. dev.-.005 .016 1Avon Park aq.30-0.003 2pasteurized-0.039mean std. dev.-.021 .025 1Lake Lytal aq.5-0.005 2raw-0.036mean std. dev.-.021 .022 1Lake Lytal aq.22-0.015 2raw-0.040mean std. dev.-.028 .018 1Lake Lytal aq.30-0.038 2raw-0.052mean std. dev.-.045 .010 1Lake Lytal aq.50.004 2pasteurized-0.018mean std. dev.-.007 .016 1Lake Lytal aq.22-0.020 2pasteurized-0.043mean std. dev.-.032 .016 1Lake Lytal aq.30-0.007 2pasteurized-0.032mean std. dev.-.020 .018Water source Linear inact. rate (log10/d) Table 39. First-order inactivation rates in ASR surface water. Trial Temp. oC 1Bill Evers res.5-0.072 2raw-0.023mean std. dev.-.048 .035 1Bill Evers res.22-0.139 2raw-0.068mean std. dev.-.104 .050 1Bill Evers res.30-0.124 2raw-0.220mean std. dev.-.172 .068 1Bill Evers res.5-0.005 2pasteurized0.0004mean std. dev.-.002 .004 1Bill Evers res.22-0.009 2pasteurized-0.052mean std. dev.-.031 .030 1Bill Evers res.30-0.004 2pasteurized-0.377mean std. dev.-.191 .264 1Clear Lake res.5-0.025 2raw-0.043mean std. dev.-.034 .013 1Clear Lake res.22-0.088 2raw-0.079mean std. dev.-.084 .006 1Clear Lake res.30-0.089 2raw-0.14mean std. dev.-.115 .036 1Clear Lake res.5-0.005 2pasteurized-0.015mean std. dev.-.010 .007 1Clear Lake res.22-0.004 2pasteurized-0.022mean std. dev.-.013 .013 1Clear Lake res.30-0.004 2pasteurized-0.024mean std. dev.-.014 .014Water source Linear inact. rate (log10/d)

PAGE 145

133 Aside from the different statistic used for co mparison of factors affecting PRD-1 survival, the analyses that were performed on the data for PRD-1 were the same as those for the other organisms. Thus, ANOVA result tables were created fo r these trials as well and are shown in Appendix 4-J. In a 2-way ANOVA considering temperature and tr eatment effects for all water source s, the increase of inactivation rates with temperature was statistically-significant to the 90% level. However, while inactivation rates decreased in pasteurized water over raw water, the effect was just beyond the range of statistical significance with p = 0.141 (Appendix 4-J). In a 3-way ANOVA that added differences due to water type, temperature and water type significantly affected inactivation rates (p < 0.05), while treatment effects were still not quite significant with p = 0.107. As shown in Figure 44, mean inactivation rates steadily increased with temperature, when both water types and both tr eatments were combined. Also, inactivation was more rapid in raw water than in pasteurized, as with all other organisms evaluated. Figure 44. PRD-1 first order inactivation rate constants in ASR aquifer and reservoir samples (log units per day), in combined raw and pasteurized water, averaged by temperature and pasteurization (combining both water types). Individual 95% CI temp C Mean ------+---------+---------+---------+----5 -0.018 (----------*----------) 22 -0.039 (----------*----------) 30 -0.074 (---------*----------) ------+---------+---------+---------+-----0.090 -0.060 -0.030 0.000 Individual 95% CI treated Mean --+---------+---------+---------+--------no -0.0573 (------------*-------------) yes -0.0296 (------------*------------) --+---------+---------+---------+---------0.0800 -0.0600 -0.0400 -0.0200 The diagram of mean rates in Figure 45 demonstr ates the effect of temperature and water type in raw water, such that inactivation was much more rapid in surface water than in ground water. Temperature, water type, and the interaction of thes e factors were all significant to the 95% level in unpasteurized water. As detailed below, increased inactivation due to temperature was significant in surface water but not in gro und water sources (raw). A two-way A NOVA for these parameters at only 22 and 30 revealed that at temperatur es common in the Floridan aquifer system, inactivation was significantly

PAGE 146

134 faster in surface water than in ground water (p < 0. 05), but while mean inactivation rates increased going from 22 to 30 the effect was not statistically significant for both water types averaged (p = 0.115, Appendix 4-J and Figure 46). Figure 45. PRD-1 first order inactivation rates (log / d) in raw surface and ground water. Individual 95% CI temp C Mean ----+---------+---------+---------+------5 -0.028 (--------*--------) 22 -0.058 (--------*--------) 30 -0.087 (-------*--------) ----+---------+---------+---------+-------0.100 -0.075 -0.050 -0.025 Individual 95% CI type Mean -------+---------+---------+---------+---ground -0.0222 (-----*-----) surface -0.0925 (-----*-----) -------+---------+---------+---------+----0.0900 -0.0600 -0.0300 0.0000 Figure 46. PRD-1 first-order inactivation rates (log / d) in raw surface and ground water at 22 and 30 C, averaged by temperature and water type. Individual 95% CI temp C Mean -------+---------+---------+---------+---22 -0.058 (------------*------------) 30 -0.087 (------------*------------) -------+---------+---------+---------+----0.100 -0.080 -0.060 -0.040 Individual 95% CI type Mean --+---------+---------+---------+--------ground -0.026 (-------*------) surface -0.118 (------*-------) --+---------+---------+---------+---------0.140 -0.105 -0.070 -0.035 ANOVA results for PRD-1 inactivation rates from comparisons of each of the water types taken independently are in Appendix 4-J as well. In su rface water, the increase of inactivation rates with temperature was significant to the 90% level when av eraging both raw and pasteu rized water, and to the

PAGE 147

135 95% level in raw water considered alone. Treatment in surface water resulted in a decreased average inactivation rate, by about , but the effect was not significant (p = 0.152). In ground water, neither treatment nor temperature resulted in statistically-signi ficant effects on PRD-1 in activation rates. There was a consistent increase of rate with temperature such that in raw ground water, the inactivation rate at 5 was about that at 30 C. However, variability in thes e cases resulted in fairly high p values, all in excess of 0.3 for these ANOVA. Table 40 summarizes factor s found to be significant for variability of PRD-1 inactivation rates. Table 40. Significant variables affectin g PRD-1 survival in ASR water sources. Water typeTreatmentComparison parametersSignificant factors (95% unless noted) surface & groundraw and pasteurized2-way: temp, treatmnt temp (90%) surface & groundraw and pasteurized3-way: temp, treatmnt, type temp, type surface & groundraw only2-way: temp, type temp, type, interaction surface & ground raw only, 22o and 30o C 2-way: temp, type type surface water raw and pasteurized2-way: temp, treatmnt temp (90%) surface water raw only1-way: temp temp groundwaterraw and pasteurized2-way: temp, treatmnt none groundwaterraw only1-way: temp none Although many of the conditions resulted in sl ow inactivation with predicted days for 2-log decline of over 200 days, like the other organisms PRD-1 inactivation was generally most rapid at the higher temperatures of 22 and 30 C in raw water, conditions which most closely emulate the environment of concern. Thus, for comparison, Figure 47 presents predicted 2-log days in a graphical format. All the values used for this graph were less than 200 days. However, since the number of days was much greater on average than that predicted for the other groups of organisms, the scale on Figure 47 is different than on similar figures for the other organisms. This is important to note when comparing this to the other graphs. Linear regression rates of combined data for the above conditions plus at 5 C are given in tabular format as was done for the other organisms previously, and this information is presented in Table 41.

PAGE 148

136 Figure 47. Chart of mean days for 2-log inactivation for PRD-1 in raw aquifer and surface water at ambient temperatures for Florida. PRD-1 bacteriophage 0 50 100 150 200 250 2230 Temperature (oC)Days for 2-log10 (99%) decline Avon Park aq. Lake Lytal aq. Groundwater Bill Evers res. Clear Lake res. Surface water Table 41. First-order inactiva tion rates of PRD-1 in raw aq uifer and surface water sources. Temperature (C)Avon Park aq.5-0.009222333 22-0.017118176 30-0.015133200 Lake Lytal aq.5-0.015133200 22-0.02774111 30-0.0454467 Bill Evers res.5-0.0474364 22-0.1031929 30-0.1531320 Clear Lake res.5-0.0345988 22-0.0842436 30-0.1151726Water source Linear inact. rate (log10/d) Est. days to 99% decline Est. days to 99.9% decline

PAGE 149

137 Variability and PBS Controls Experiments for survival in these reservoir an d aquifer water samples were performed in two separate sets for fecal coliform, en terococci, DNA coliphage and PRD-1. These four organism groups were evaluated in water samples collected at the same time for each trial set. For each, experiments in water from each of the two sites were performed using separa te batches of organisms so that the test organisms were as fresh as possible. Experiments for F+ RN A coliphage were performed in duplicate using water collected at one sample event, with the same batch used for water from each site. For the bacteria, even though each trial was performed using water collect ed at separate times months apart (August and November) and with separate batches of cultivated organisms, fairly good agreement was observed between replicates for each organism Instances of large inter-trial variability were observed for fecal coliform in Lake Lytal Park pasteurized water at 5 an d in Bill Evers reservoir water pasteurized at all three temperatures (Table 22 and Table 23). For enterococci, a large degree of variability was observed for Avon Park well water pasteurized at 22 and Bill Evers water pasteurized at 22 C (Table 26 and Table 27). As would be expected for F+ RNA coliphage, no cases of qualitatively large differences were observed between replicates (Table 30 and Table 31). Instances of larger difference between inactiv ation in the two trials for DNA coliphage were observed in Lake Lytal Park well water and Bill Evers re servoir water, both pasteurized at 22 and 30 C. Variability between sets in pasteurized Avon Park well water at 30 and perhaps 22 was also present. Similarly, inactivation of PRD-1 showed large differen ces between the two trials in all pasteurized water samples at 22 and 30, with inactivation being much mo re rapid in the second se t in each case. Although there is an apparent non-randomness to this difference, no known reason in terms of procedural differences was suspect. Much more rapid inactivation at higher temperature in the second trial with Bill Evers pasteurized water was observed for all four organism gr oups that were evaluated in this water collected at the same time. Thus, there may have been some factor in this surface water source at the second sampling even and not at the first which had a detrimental effect on all four enteric organism groups; this could have been a chemical constituent that was either heat stabile or was liberated as a result of pasteurization. Control microcosms with PBS were used for side-b y-side survival trials for each organism in each experimental set. Results of PBS bottle experiments are shown in Appendix 2 and 3 with the results of

PAGE 150

138 each organism for natural water trials. Inactivation data were analyzed in the same way as for experimental sets. ANOVA were performed on the results of PBS control microcosms to determine if statisticallysignificant differences existed between sets. Fo r each organism, a two-way ANOVA comparing effects due to set and temperature was performed, and a covari ance analysis for set differences and temperature as covariate was also performed. In all cases, no statisti cally significant differences were observed due to set for PBS control microcosms. Results of these analyses are shown in Appendix 4 under the section for each organism group. In regard to experimental trial variability observed for PRD-1 in pasteurized samples, PBS control results did not indicate that inactivation of the respective batches of PRD-1 used in the second set for each site experiments was significantly more rapid in PBS as was observed in natural water samples. Discussion Based on comparative analyses of observed inactivation behavior of these five microorganism groups, several overall trends were determined. Prim arily, predicted days for 2-log inactivation decreased steadily with increasing temperature; this was true in both combined raw and pasteurized comparisons and in raw water comparisons, as well as when considerin g only 22 and 30 C. In addition, inactivation was typically more rapid in surface water under raw conditio ns. Heat pasteurization to reduce native organism populations also had a noticeable effect of reducing in activation in most cases, and its effect was sometimes more significant in surface water than ground water (f or enterococci and DNA coliphage this was the case). Another point of interaction was the relatively greater increase of inactivation of fecal coliform and PRD-1 with increasing temperature in surf ace water over ground water. These observations invite speculation on what factors may be responsible for variability of survival in the microcosms. The trend of increasing inactivation at higher temperatures could be due to both reduced stability of biomolecular structure of the organisms themselves, and due to temperature effects on other antagonistic factors. The more rapi d inactivation in raw water as opposed to pasteurized conditions indicates that in most cases, the native mi crobial populations had an antagonistic effect on

PAGE 151

139 seeded bacteria and viruses. In raw water microc osms, inactivation was generally more rapid in surface water than in ground water. Thus, there were perhap s biological or chemical f actors present in the surface water that negatively impacted survival of the non-na tive microbes that had less of an effect in ground water. TDS may be one chemical factor in the gr ound water resulting in long er survival. But recalling results from Chapter 3, it was demonstrated that in direct comparisons for the effect of TDS and temperature on survival of these organism groups, enterococci was the only organism which may have been significantly influenced by TDS. Longer survival was observed in 3000 mg/l at 22 and in one trial but not the other at 30 than the other, lower concentrations where predicted 2-log inactivation times were fairly well grouped. The enhancement of surv ival due to TDS at this higher concentration was significant to the 90% level. It may be significant then, that in considering enterococci inactivation in these experiments, not only did enterococci survival longer on average in ground water than surface water, but between ground water sources survival was longer in the Lake Lytal Park well water, with a measured TDS of about 3,000 4,000 mg/l (Table 19), as compared to survival in Avon Park well water with a TDS of about 1500 mg/l. This longer survival may be seen from inactivation rates and predicted periods for decline in Table 29. Inactivation between surface water sources was similar for enterococci. Such a trend was not observed for the other organisms. Thus there is further indication of enhanced survival of enterococci due to increasing TDS going from fresh to slightly saline water. The significant enhancement of pasteurization eff ects with increasing temperatures was noted for enterococci in surface water, but not in ground water. This may then be due to greater activity of other microorganisms in the water that ne gatively impacted survival of ente rococci, such as phage, predatory protozoa, and other bacteria whose activity may be antagonistic to the seeded organisms. It is logical to presume that some antagonistic organisms may be mo re numerous in surface water, since ecologically the presence of enterococci would be common in surface water but uncommon in the deep aquifer. Thus specific interactions contributing to more rapid demise of enterococci would not be as common. For instance, phage capable of infectin g enterococci would be unlikely in the Floridan aquifer system, and larger bacterivorous protozoa would be less numerous in ground water due to filtration. It would follow that pasteurization would eliminate most or all of these types of antagonists, leading to enhanced survival with relatively greater differences in surface water, and at higher temperatures those organisms are more

PAGE 152

140 active at temperatures close to those of their natural environment. Along with enterococci, pasteurization significantly increased surv ival of DNA coliphage in surface water, but not in ground water. While no statistically-significant temperature interaction existe d, this shows that surface water microorganisms may also have a relatively larger impact on viruses such as coliphage, perhaps through consumption as dissolved organic matter. Background microorganism levels aff ected survival of fecal co liform bacteria, in both surface water and ground water. Al so, inactivation of PRD-1 was reduce d in pasteurized water of both types, and to a larger degree in su rface water, but the effects were not statistically significant. This may indicate more of an indirect effect on PRD-1, or that they are more resistant to whatever stresses are imposed by native microorganisms (bacteria most likely). Besides the absence of evolved ecological interactions with fecal indi cator bacteria and viruses at depths of the Floridan aquifer system, the aerated, surface-pressure environment imposed by microcosms may have impeded the overall activity of native ground water bacteria and thereby re duced their impact on seeded organi sms. In fact, for fecal coliform, initial increases of culturable concentr ations, an indication of re-growth, were observed for all pasteurized microcosms at 22 and 30 except for trial 2 with Bill Evers reservoir water. Therefore, native microorganisms in surface water and perhaps to a lesse r degree in ground water likely have a direct or indirect influence on the survival of allocthonous bacteria and viruses. The metabolic activity of native microorganisms ma y also be a factor in their impact on seeded organisms. Organic carbon content of surface water sources was much greater than of the ground water (Table 19). This may have also led to a larger degree of overall metabolic activity in the surface water microcosms, particularly at greater te mperatures. Another part of the environment that could lead to such a difference is the oxidation conditions of the water. The aquifer environment where ground water samples originated was typified by reducing conditions, judging by the sulfurous odor of the water. Naturally, surface water microorganisms originated from an aerated surface reservoir. Therefore, introduction of the ground water microorganisms to the oxidizing environment of the microcosms may have limited their metabolic activity, along with the much lower organic carbon concentrations. Nonetheless, the greater metabolic activity at higher temperatures in surface and perhaps a lesser degree in ground water could have been a reason why inactivation was more rapid in raw water, and why pasteurization generally reduced inactivation.

PAGE 153

141 Other reports have identified autochthonous microor ganism presence as a f actor in survival of seeded public-health-relat ed bacteria. Of most importance, Kersters, et al. made similar observations to those found here on the relative impact of sterilizing water samples prior to seeding; they observed that for survival of A. hydrophila pre-sterilization had a larger impact on survival in surface water than in ground water (Kersters 1996). Other research has shown enhanced survival of E. coli in 0.22 m-filtered stream water over unfiltered or even 1 m-filtered water (Janakiraman and Leff 1999). Sobsey et al. demonstrated enhanced survival in sterilized water over raw for seve ral enteric viruses (poliovirus 1, echovirus, hepatitis A) (Sobsey 1986). Perhaps of more importance, an older study on survival of poliovirus in water amended with sewage effluent found that with sewage effluent addition under non-sterile aerobic conditions, poliovirus decline more rapidly but under sterile aerob ic and anaerobic conditions and non-sterile anaerobic conditions the sewage had no effect (Hurst 1980). This potential effect of oxidation state of the water could play a role in aquifer injection scenarios in Florida and elsewhere, where aquifer dissolved O2 concentrations can be quite low. The potential impact of aeration and redox state of water is addressed more in the following chapter. Several investigators have examined survival of water-quality-related microorganisms in water previously, but the studies reported here are significant in a number of ways. We have evaluated a suite of microbes in a comprehensive fashion, and these organisms consisted of composite populations of bacteria and indicator phage rather than single isolates. Us e of composites provided a more-realistic picture of survival behavior of these organisms in environmental waters. Furthermore, by evaluating a number of different populations of indicators under simila r conditions, direct comparisons of the types of microorganism can be made. The two water types and temperature and pasteurization levels employed for these studies contribute to their comprehensiveness, in order to provide a broad spectrum of survival behavior and some information on parameters affecting survival in the environment of concern. While observations on the factors which impact su rvival of these indicator microbes are useful for understanding possible processes in the environment, the actual rates of decline and periods necessary to allow a certain concentration to die out are very important for gauging the risk of ground water contamination or injection of potentially harmful microorganisms. Naturally, most interest lies with bacteria and virus survival and inactivation in water conditions as would be found in the environments of

PAGE 154

142 concern. Figures in the Results section depict days predicted for 2-log decline in raw water sources for each organism. These figures provide a summary pict ure of survival in raw water at 22 and 30, temperatures that would be observed in subtropical aquifers like the Floridan aquifer system. Important trends are visible in these figures. For both of the bacterial groups, DNA coliphage, and PRD-1, inactivation was more rapid overall in surface water th an in ground water at thes e temperatures. While the same cannot be said for F+ RNA coliphage, averag ed periods for inactivation were similar and low between the two water types, such that the difference was nearly indistinguishable. At these temperatures, fecal coliform and entero cocci inactivation was similar in Figure 25 and Figure 29. From Table 25 and Table 29, estimated days for 2-log decline from first order inactivation rates were similar for these two bacteria at Floridan aquifer system temperatur es as well. For fecal coliform, 2log inactivation was predicted over periods on the order of 2 4 weeks in ground water and 1 2 weeks in surface water; enterococci predictions ranged slightly less at around 1 4 weeks in ground water and about 1 week in surface water sources. A comparison of th e two types of coliphage reveals that DNA coliphage were much hardier in the conditions evaluated, as Figure 35 and Figure 43 show. F+ RNA had the shortest periods for 2-log decline, thus the most rapid inactiva tion in these experiments. At these temperatures, 2log inactivation in both surface and ground water were estimated at less than 1 week for F+ RNA. DNA coliphage results, on the other hand, indicated 2-log inactivation over periods on the order of 2 4 weeks in both water types. As expected, PRD-1 was the most stable of the organisms evaluated. As can be seen in Figure 47, larger differences between the two ground water sites were observed than for the indicator organisms, but results in surface water were similar. Also, inactivation in su rface water was much more rapid on average than in ground water. Table 41 in the Results section shows predicted periods for inactivation of PRD-1, but even in these more harsh conditions inactivation in ground water was limited and predicted periods for 2-log decline were well outside the experimental duration. For instance, periods for 2-log decline in Avon Park well water were predicted at around 4 months at 22 or 30 based on observed inactivation rates over 28 days. Predicted durations were shorter in other water sources, and were within experimental durations for th e surface water at 2 -3 weeks. Based on these observations, comparisons of all the organisms for survival in raw ground and surface water may be made. Excluding PRD-1, which is rarely found in high quantities in contaminated

PAGE 155

143 water, the indicator with the longest survival times at 22 and 30 was DNA coliphage with mean predicted 2-log inactivation periods of 22 days in ground water and 16 days in surface water. The bacteria were slightly less in ground water and somewhat less in surf ace water, but were similar to each other. Days for 2-log decline averaged between 22 and 30 were 19 da ys in ground water and 6.5 days in surface water for fecal coliform, and 17 and 5 days respectively for ente rococci. Inactivation of the F+ RNA coliphage was thus most rapid under these conditions with average da ys for 2-log decline of 2.5 days in ground water and 3.5 days in surface water. These comparisons indi cate that under the conditions found in Florida’s subsurface, with regards to injecti on of possibly contaminated surface wate r, neither fecal coliform bacteria nor enterococci has an obvious greater margin of safety as an indicator. However, as discussed previously, for aquifer water of higher TDS enterococci may survive longer. In reality though, if these organisms were input to the subsurface with surface wa ter injection, the aquifer zone in which they would be found would likely have much lower TDS than the displaced high -TDS native aquifer water. Higher salinities could possibly be found at the periphery of surface-derived water underground. DNA coliphage appear to have a slight edge as a more-conservative general indicator in terms of their survival. The difference of genomic material between DNA and RNA coliphage may be one reason for their relatively longer survival. The RNA genome of the F+ RNA coliphage may be less stabile under these conditions. One point to consider also is that the DNA coliphage population was isolat ed from surface water, while the F+ RNA coliphage isolates came from secondary wastewater effluent. Thus, the DNA coliphage had more environmental exposure prior to collection in the water sample and may thus be hardier variants. No F+ RNA coliphage were found in the surface water source when the expe rimental populations were created. Therefore, differences among these isolate populations may be a factor for their relative su rvival, differences which may not translate to all such indicators. Given that th e survival of these organisms was fairly similar, but they could potentially have different behavior in the environment at large, monitoring of any aquifer injection of raw surface water to the subsurface shou ld employ fecal coliform, enterococci, and nonspecific coliphage such as those infecting E. coli C-3000. However, if techno logy or funds were limiting, fecal coliform and coliphage monitori ng would likely be sufficient, with the omission of enterococci. Since inactivation rates of similar organisms from a number of studies in environmental water were compiled for the literature review in Chapter 2, results of these studies may be compared to those

PAGE 156

144 ranges of values. Coliform bacteria in reviewed studies had inactivation rates ranging from about 0.01 to 0.1 log / d at temperatures of 0 10 , with a mean of 0.05; fecal coliform in raw water trials at 5 in the present study had inactivation rates of about 0.02 to 0.07 log / d, near the mean of reviewed studies, except for inactivation in Laky Lytal Park water at 5 which wa s more rapid at 0.14 log / d (Table 25). At higher temperature, reviewed rates ranged from 0.9 to 0.35 log / d (mean ~0.2) at 21-25 C, and rates from our study were at the higher end of this range at 0.1 in ground water and 0.3 log / d in surface water, both at 22. Enterococci inactivation rates observed in this st udy were also within the ra nges of reviewed rates, and near the means of respective temperature ranges; the mean of reviewed rates at 0 10 was 0.08 log/d, and most enterococci inactivation rates at 5 from this study were around 0.05 log / d, except in Lake Lytal Park water which was much lower at 0.01 log / d. At higher temperature, the mean from reviewed papers at 21 25 was 0.24 log / d, from our work at 22 C enterococci inactivation rates were about those of fecal coliform, 0.1 log / d in gr ound water and 0.3 log / d in surface water. For reference to some pathogenic bacteria, inactivation rates for Salmonella and Shigella spp. bacteria were near these for temperatures over 20 C. Coliphage inactivation rates from reviewed studies as means for studies at each temperature range were about 0.03 log/d at 0 10, 0.4 log/d at 21 25, and 0.4 log/d at 26 30. Almost all phage inactivation rates were from studies in ground water. The inactivation rates of DNA coliphage at 5 closely match the average from reviewed studies at about 0.02 to 0.03 log/d in both water types, while RNA coliphage rates at this temperature were faster at 0.07 5 to 0.1, but still within the range of observed studies of 0 0.1 log/d. At 22, RNA coliphage fell more in line with reviewed rates, with mean inactivation rates in both water types at 0.5 log/d, and DNA coliphage rates were much slower at about 0.07 to 0.1 log/d. At 30 temperatures, RNA coliphage inactivation was on the high side of the range from reviewed studies, at around 1 2 log/d with rates from reviewed studies ranging from 0.2 to 2.5 log/d. However, DNA coliphage inactivation was below the reviewed rates average and the lower bound of the range at about 0.15 log/d in both water types. PRD-1 inactivation rates observed from the work described in this chapter were near those found in reviewed studies at low temperature (0 10 C), such that rates in this study were about 0.01 log/d in ground water and 0.04 log/d in surface water, while the m ean of reviewed PRD-1 inactivation rates at low temperatures was 0.02 log/d. At temperatures around 21 25, rates from this study were

PAGE 157

145 below rates from reviewed studies, which ranged from 0.05 to 0.8 log/d (mean 0.32). Our results indicated PRD-1 inactivation rates at 22 were around 0.02 to 0.03 in ground water and about 0.1 log/d in surface water. Interestingly, although our PRD-1 inactivation rates in ground water at 22 were lower than rates from reviewed studies, they were a good approximat ion to hepatitis A virus inactivation rates from two studies which evaluated its survival in ground water and ground water/soil microcosms (Sobsey 1986; Nasser and Oman 1999). At temperatures of 20 25 C, these HAV rates ranged from 0.015 to 0.14, with an average of 0.05. This comparis on suggests that both PRD-1 and DNA coliphage survival from this study (at 22 C) could be indicative of hepatitis A survival. Others have found that PRD-1 survival may be a good model for that of hepatitis A in ground water (Blanc and Nasser 1996). Compiled poliovirus inactivation rates averaged somewhere in between RNA coliphage and DNA coliphage rates, at about 0.27 log/d at 21 25 C and about 1 log/d at 26 30 C. The objective of the experiments described in this chapter was to better define important aspects of microorganism survival in waters and conditions that might approximate what would be observed in raw-water ASR scenarios in the Floridan aquifer sy stem and Florida subsurface in general. A broadspectrum analysis of indicator mi croorganism survival was accomplishe d, with some interesting and important trends identified for several populations of in dicator bacteria and viruses. While this study has involved a broad look at a number of different organisms, and their survival in response to several general factors like temperature and pasteurization was evaluate d, more specific investigations might reveal more about mechanisms surrounding the trends observed here. For instance, a specific ev aluation of only one or two organism groups in raw water sources might eval uate the actual mechanism of less-rapid inactivation in pasteurized water. One technique that could be used is evaluation of the size fractionation of radiolabelled cells, to see if they are taken into a larger size fraction by protozoan predators perhaps. Alternatively, microscopic comparisons of stained ce lls may reveal information on their fate. Other parameters of the background biological activity might reveal more on mechanisms of inactivation as well, such as heterotrophic plate counts, and bacterial and viral direct counts during the course of a survival experiment to determine if these numbers change. These results have indicated that survival of enterococci, fecal colif orm, and DNA coliphage (male-specific and/or somatic) survival was fairly sim ilar to each other at temperatures resembling those of

PAGE 158

146 the Floridan aquifer system. This suggests that they ma y be effective as a suite of indicators for monitoring survival in larger-scale situations. In an actua l ASR system, surface water (that if untreated would potentially contain some microorganisms of health concern) would be injected into the Upper Floridan or other aquife r region, thereby displacing the existing ground water. This displacement would result in some mixing near the edges of the stored water zone, and other chemical and biological changes occurring in the stored water. Since these changes might impact survival of contaminant microbes, a larger-scale investigation of survival under a model injection scenario would be beneficial. Such a large-scale apparatus might include saturated porous media, initially saturated with ground water from the Upper Floridan aquifer, which would undergo a “storage cy cle” with seeded reservoir water where the surface water is injected to displace the ex isting ground water and incubated to observe storage effects on survival in a more-realistic environment. In such a case or with an actual field-scale injection experiment, monitoring of microbial and other parameters would be important. Survival of seeded microorganisms or naturally-occurring fecal indicators s hould be monitored both at the fringe of the storage zone and in its center to evaluate differences. Chemical parame ter measurements such as dissolved oxygen, H2S concentrations and pH, along with TDS changes, woul d be informative, particularly to correlate with observed survival characteristics of microorganisms. The experiments described here represent an important step in investigation of issues surrounding the important technology of aquifer storage and recovery in Florida. The appropriat e next step is now an “up-scaling” of this type of work to examine behavior in the real environment.

PAGE 159

147 CHAPTER 5: SURVIVAL OF FECAL COLIFORM, ENTEROCOCCI, AND F+ RNA COLIPHAGES IN SATURATED LIMESTONE Introduction Previous studies described in Chapters 3 and 4 have examined a number of parameters for their relative impact on surv ival of fecal indicator bacteria and viruse s, focusing on ground water environments and injection of surface water. Isolating the effects of total dissolved solids and temperature revealed that of the range of 5 to 30, temperature produced a st atistically-significant decreas e for survival of all the organisms evaluated except PRD-1. This was to be expected in light of previously reported work. However, TDS over the range from 200 3000 mg/L did not seem to have notable effects on survival variability for most organisms. There was some indication that enterococci may be stabilized by TDS of 3000 mg/L, but no significant impacts were observed for lower concentrations, below 1000 mg/L. Survival of F+ RNA coliphage was longer at the high TDS conc entration in lower temperatures, but this trend was reversed at higher temperatures and likely there is not a significant enhancement of survival overall due to TDS for these viruses. Regarding survival of the same groups of indicator microorganisms in environmental water sources, typical of those that would be employed for aquifer storage and recovery projects in Florida, increasing temperature was once again seen to negativel y impact survival. Pasteurization of water samples generally served to increase survival of seeded organisms, and in the case of enterococci and DNA coliphage, the effect of pasteurization was relativel y more pronounced in surf ace water sources than in ground water. The impact of pasteurization was also greater at higher temper atures for fecal coliform, enterococci in surface water, an d DNA coliphage such that relative differences between raw and pasteurized water were greater at higher temperatures. Inactivation was also more rapid in surface water on average, without regard to pasteu rization. F+ RNA coliphage inactiv ation at ambient temperatures for

PAGE 160

148 Florida aquifers was invariably rapid, with little obs erved difference due to the various parameters. These results indicated that possibly one of the main controls on survival of allocthono us fecal microorganisms is by native microbial populations perhaps via directly antagonistic interactions such as predation (by protozoa) and parasitism (by viruses), or via indirect interactions due to metabolism of the native microbial populations. One important point to note in interpreting the results obtained in this line of research so far is that all experiments were conducted under aerated conditions. Thus, the oxidative state of the water may not match that of the aquifer in question. Notably, water that was sampled from the Upper Floridan aquifer via both the Avon Park well and the Lake Lytal Park we ll had a strong hydrogen sulfide odor indicative of sulfate reduction under oxygen-limited conditions. Also, in ground water (saturated) zones of aquifers, there is no air-water interface as was present in the microcosms employed for previous experiments, and there is the presence of a mineral matrix providing a di fferent environment for attachment than found in the bulk water column. The effect on survival of attachment to a solid surface has been studied in several cases. Numerous reports have suggested that attachment offers protection to virus particles and reduces inactivation rates, as described in a review by Sc hijven and Hassanizadeh (Schijven and Hassanizadeh 2000). Specifically, Ryan et al. determined that inactivation of PRD-1 was 3 x more rapid in solution than when attached to solid su rfaces (Ryan 2002). Sakoda et al. also found that various E. coli phage were significantly more stable when adso rbed to various solids surfaces than when in suspension (Sakoda 1997), while Rossi and Aragno found that presence of colloidal clay particles protects the coliphage T7 from rapid inactivation (Rossi and Aragno 1999). Stream sediments are frequently found to harbor much greater nu mbers of fecal coliform bacteria and likely confer protectio n from inactivation in surface water (Buckl ey 1998; Crabill 1999) In tropical climates, fecal coliform bacteria may be able to reproduce in se diments, thus maintaining their concentrations at consistent levels and contributin g to their increased detection in overlying waters (Toranzos 1991; Roll and Fujioka 1997; Desmarais 2002). A study which specifically compared survival of fecal coliform and fecal streptoco cci in water with and without sedime nt found longer survival of these bacteria in water incubated with sediment (Sherer 1992). However, results of other studies have indicated that inactivation of viruses may be more rapid in wate r with solid particles pres ent, as the case for MS-2

PAGE 161

149 and PRD-1 in studies by Blanc and Nasser (Blanc and Na sser 1996), or a solid matrix had little effect on inactivation, as found in studies in anaerobic ground water with Enterococcus faecalis (Pavelic 1998). Regarding impacts from the oxidative state of ground water, few studies were located which evaluated these effects. However, Jansons et al found a small increase in inactivation of poliovirus in water with a greater dissolved O2 concentration, manifested as 2-lo g inactivation in 20 days in DO concentration of 5.4 mg/L vs. 50 days in ground water with DO concentration of 0.2 mg/L (Jansons 1989). However, a study by Banning et al. found that in anaerobic ground wa ter, an evaluation of survival of E. coli in sterilized and non-sterile water revealed that st erilizing dramatically enhanced survival of seeded bacteria (Banning 2002), suggesting that anaerobic ground water microorganisms produce antagonistic effects on introduced fecal indicators. This may be im portant in light of results presented in Chapter 4, in which a significant enhancem ent of survival was produced by pasteurizing water samples, and this effect was more pronounced in some cases in surface water than ground water. Furthermore, inactivation was typically more rapid in surface water than in ground wate r. It would be important to consider then, that an oxygenated environment is likely foreign and quite possibly harmful to the native ground water microorganisms found in regions of the Upper Floridan aquifer, where oxygen tensions are generally considerably lower and sulfate reducers are active. Th is may reduce the effect of native aquifer microbial communities on introduced fecal indicators. But while the activity of obligate anaerobes would be reduced/eliminated, aerobic metabolic activity woul d obviously be much greater overall in aerated microcosms than in an anaerobic aquifer. Naturally, overall metabolic rates and resulting biomass degradation are also more rapid in oxygenated vs. anaerobic or oxygenlimited water. Conversely, in an environment with limited or no aeration, dynami cs of native surface water and aquifer microbial communities could be different, along with their pote ntial impacts on introduced fecal microorganisms. Since studies with the representative ASR water samples demonstrated a general decrease of inactivation of seeded microbes with reduction of native microbial communities (pasteurization), some of these potential interactions may be different when in a less-aerated en vironment or when solid material is present to offer protection from predation or other antagonistic interactions. To investigate possible impacts of having a solid ma trix present and lack of aeration on survival in microcosm environments, another set of experiments was conducted. Survival in the pore water of

PAGE 162

150 saturated limestone tubes was evaluated with three of the microbial populations in each of the model ASR surface and aquifer water samples. This type of expe riment presented a more realistic environment which served to limit effects of aeration on the microcosms, while also providing a solid matrix typical of that found in the Upper Floridan aquifer. Ocala limestone was used to provide the solid media. This Eocene series limestone comprises all or part of the Upper Floridan in a large portion of peninsular Florida (Berndt 1998). Survival of fecal colifor m, enterococci, and F+ RNA coliphage was evaluated in these static saturated rock microcosms at 22 and 30 C, temperatures typical of the Upper Floridan aquifer. Methods The experiment to examine survival of fecal indicator organisms was performed using crushed limestone gravel saturated with water from 2 reservoi r sites and 2 nearby well sites that withdraw water representative of aquifer water used or planned for use as receiving zones for ASR systems. The two reservoirs served as raw water sources for drinking wate r plants that treat water for use in treated drinking water ASR wells and could in the future be involved in partially treated surface water or raw surface water ASR wells should they be permitted. These water so urces are the same as thos e used for representative ASR source and receiving water waters in studies on microorganism survival described in Chapter 4. Limestone used in this experiment was obtained as a gravel sample from the Plaza Materials Corp.’s Zephyr mine in Zephyrhills, Florida. Ocala limestone is the rock mined in this region. The gravel sample used was originally designed for use as road base or other general gravel uses. The limestone gravel was collected in a 5-gallon bucket which had previously been rinsed with deionized water. After collection, gravel was washed thoroughly, first by flushing the rock in the bucket with tap water using a hose placed at the bottom of th e bucket to wash material from the bottom-up. The bucket was agitated several times to enhance flushing of fine particles from the gravel and was left to flush overnight. The following day, deionized water was flushed through gravel for several hours to flush away any chlorine residual. After this time, the gravel wa s allowed to drain and dry pa rtially. Gravel was then crushed by gently tapping with a hand sledge; this served to produce pulverized limestone composed of

PAGE 163

151 mainly sand-sized particles, with a mixture of larger and smaller grain sizes. This pulverized limestone was placed in 4-L polypropylene c ontainers and thoroughly washed again in these containers with deionized water (non-sterile). Initially, rinse water wa s very turbid and the limestone sand was rinsed until the water was not visibly turbid, thus removing the larg e majority of small, clay-sized particles. After washing, limestone designated for e xperiments in each of the two sites’ water samples (City of Bradenton and City of West Palm Beach) was saturated in gro und water from the Avon Park (Bradenton) or Laky Lytal Park (West Palm Beach) wells that had previously been stored frozen at -70 C from the prior sampling event (F+ RNA coliphage water-column experiments). The limestone was saturated by pouring half of the pulverized volume into each ground water sample in 4-L containers, such that a small volume of free water remained above the level of sand. Saturated limestone was stored in this way at room temperature for 5 days, up until use in filling static microcosm tubes, to allow some conditioning of the material by each ground water, par ticularly to facilitate some establis hment of a microbial community that might adsorb to particles from any surviving micr obial populations in the thawed ground water. Organisms evaluated for survival in pore wate r of saturated limestone were fecal coliform, enterococci, and F+ RNA coliphage. Organism populations were prepared from the same frozen isolates as used for experiments described in Chapters 3 and 4. In short, these were 8 Escherichia coli and 2 Klebsiella pneumoniae isolates for fecal coliform bacteria, 7 Enterococcus faecalis 1 E. faecium and 1 E. durans isolates for enterococci bacteria, and 10 isolates of F+ RNA coliphage isolated and typed as described previously in Chapter 3. Organisms were grown and purified as described previously as well. Each phage isolate was titered, while each of the bacterial mixed popula tions were assumed to be at the consistent concentration obtained from propagation procedures of 8 x 108 cfu/ml for fecal coliform and 5 x 108 cfu/ml for enterococci. Test organisms were grown and titered (for phage) at USF in St. Petersburg and transported to Michigan State University in Ea st Lansing with limestone and water samples for experiments. Water samples were obtained from the same sites and in the same way as described previously for water column microcosms in Chapter 4. They were in short, ground water from the Upper Floridan aquifer obtained from the Avon Park well at the City of Bradenton ASR test site and the Lake Lytal Park well in the City of West Palm Beach. This well is not a test site for an ASR project, but produces water

PAGE 164

152 representative of the aquifer in the region that is or will be employed as receiving zone for ASR. Concurrently, surface water was sampled from the B ill Evers reservoir in Bradenton and Clear Lake reservoir in West Palm Beach. Water samples were taken on simultaneous days by the author at Bradenton and CH2MHill personnel (F. Bennett) at West Palm B each, and samples were transported on ice overnight to MSU for experiments. Field physical and chemical parameters were recorded at sampling, with the exception of the conductivity of West Palm Beach wate r that was recorded at MS U. Total organic carbon was not measured in these water samples. In the laboratory, background microbial parameters were determined for concentratio ns of fecal coliform bacteria, enterococci bacteria, combined somatic and malespecific coliphage on E. coli C-3000, and heterotrophic plate count bacteria. No samples were pasteurized. To initiate the experiment, water of each of the 4 sources was seeded in bulk (100 ml aliquots) with appropriate volumes of bacterial suspensions or mixed phage suspensions to give approximately 3x104 per ml. Water samples were divided such that bacterial cultures were combined in one group and the phage in the other set. Microcosms consisted of sterile 15-ml polypropylene tubes. These were created in multiples so that a tube could be sacrificed for each time point, this to allow sampling of the entire pore water volume. Water for each corresponding set was seed ed immediately prior to packing of tubes for that set. Each tube was first rinsed with approximately 5 ml of non-seeded water corresponding to the water source each was to receive. Tubes were filled with ap proximately 2.5 3 ml of seeded water, and then approximately 8 10 ml of limestone was added. Tubes were packed lightly by gently tapping tubes on the lab counter. In this way, limestone was completely submerged in water to ensure 100% saturation and even distribution of microorganisms. Seeded water samples were added to limestone that had been conditioned in corresponding aquifer water. Thus, Avon Park and Bill Evers reservoir water was added to limestone conditioned in Avon Park well water, and Clear Lake and Lake Lytal Park water was added to limestone conditioned in Lake Lytal Park well water. Tubes for T0 microbial concentrations were packed and immediately sampled before the other tubes for each set were packed. For sampling, the entire contents of a tube were emptied into a sterile 50-ml polypropylene tube, and the pore water was extracted using a sterile syringe and hypodermic needle. Pore water was then ejected into another 15-ml tube and refrigerated until processing. T0 samples were performed in duplicat e for each water source. Initiation of each set took approximately 30 minutes, and after all b acterial tubes were completed they were placed in

PAGE 165

153 water baths at appropriate temperatures. This was done before packing of coliphage tubes was initiated. Tubes were placed in water baths set at 22 and 30 C, and covered to exclude light. After all phage tubes were set, analysis of refrigerated T0 bacterial and phage samples was pe rformed. Thus, seeding of water and packing of tubes was performed as expeditiously as possible, and in sets so that T0 samples were as representative as possible of initial conditions. Analyses of viable microbial concentrations were performed using methods and media as described previously in Chapters 3 and 4. Pore water of a tube from each set was sampled on days 1, 2, 4, 6 or 7, and 9 for coliphage, and days 1, 2, 4, 7, and 14 for bacteria. In addition, phosphate buffered saline microcosms were employed to offer comparison of microorganisms’ survival to other experiments described for Chapter 4. For these, PBS was mixed acco rding to EPA method 1622 and sterilized in a glass bottle. PBS microcosms for this e xperiment were set in 50-ml polypropylene tubes rather than the aforementioned polypropylene bottles. PBS microcosms were actually initiated 1 day later than the T0 time point of limestone microcosm tubes. PBS microcosms were seeded with bacterial populations in one set and coliphage in another set, with temperatures corr esponding to limestone tubes at 22 and 30 C. PBS tubes were only water-column sets, limestone was not us ed; however, seed concentrations were the same at approximately 3x104 organisms per mL. Sample points for PBS were days 0, 1, 3, 6, and 8 for both organism groups. A single tube was used for each organism population/temperature combination and samples were withdrawn from bulk fluid as in previous experiments. Since the conditioned limestone had the potential to influence basic chemical parameters in the saturating water, both from effects due to the limest one and effects of residual ground water used for conditioning on surface water, pH and TDS of pore water was measured for each sample. Only one measurement was performed, since actua l volumes of pore water were relatively small. TDS was measured from representative bacterial tubes on Day 14, and pH was measured from bacterial tubes on Day 7. Measurements were taken after removal of suspended mineral particles present in pore water samples by centrifugation for 10 min at 2000 x g. Although crushed limestone had been rinsed thoroughly prior to conditioning in ground water, activity such as abrasion from transfers of rock, and perhaps physical softening due to water contact which enhanced erosion, resulted in considerable increases in turbidity of pore water removed from limestone. To quantify th is, representative samples of each water source were

PAGE 166

154 analyzed for turbidity on Day 7, along with source water samples that had not been used in microcosm tubes. To express data, a slight variation of data analysis methods described previously for other experiments was used. Kinetics of viable counts were different from bulk water column trials in that typically a sharp decline was observed in T1 samples from T0 values, followed by an occasionally sharp increase (for fecal coliform only) or flattening of kinetics on T2 samples and beyond. In some cases numbers (of fecal coliform) increased up to T2 before beginning to decline. Thus, to describe inactivation, log N/N0 survival ratios from day 1 or 2 and beyond were fit to linear regression tre ndlines. In this way, data for the second part of bi-phasic kinetics were used to describe inactivation, and the initial rapid decrease from T0 to T1, and any subsequent increase from T1 to T2 were not modeled by the regression. As in prior experiments, results from linear regressions we re used to estimate days for 2-log inactivation to be achieved for comparison. In some cases, 2-log decline was not observed in the 9 or 14 day experiments, and so time for 2 log decline was extrapolated based on observed kinetics. However, all periods were well within 200 days. Results Several background parameters were measured in water samples either at the time of sampling or in the laboratory, as appropriate. These descriptiv e characteristics of surface and ground water samples are shown in Table 42. Temperatures were in the range of experimental incubation temperatures for these and other environmental water survival studies (22 30 C). As was observed for prior sampling events for these water sites, TDS was greatest in the Lake Ly tal Park well water samples, and both surface water samples were much lower in TDS than the ground wate r samples. HPC bacteria were generally fewer in these samples than in prior sample events. This may be due to longer holding times for these samples, a result of transport of samples to East Lansing and other logistical considerations for preparing the experiments. Also, a low number of enterococci were detected in Lake Lytal Park well water, which would not be expected.

PAGE 167

155 Table 42. Physiochemical and back ground microbial characteristics of water sources for saturated limestone experiments. Avon Park Well Bill Evers Reservoir Lake Lytal Park Well Clear Lake Reservoir T (oC) 30302229 pH7.17.27.68.1 Conductivity 3.00 mS/cm2347 S/cm26.45 mS/cm2 323 S/cm215001743,000162 HPC 4.5 x 1035.7 x 1051.9 x 1043.5 x 105(cfu /100 ml) Fecal coliform < 170< 1870 (cfu /100 ml) Enterococci < 1154850 (cfu /100 ml) Coliphage < 10< 10< 10< 10 (pfu / 100 ml) Approx TDS (mg/l) Measurements of pore water pH and TDS are shown in Table 43. These results show large increases of surface water conductiv ities after incubation with limestone The TDS increase was due in part to residual ground water used for conditioning each respective limestone batch. Interestingly, conductivities of each ground water sample decreased from measured conductivity of raw bulk water samples. In the case of Avon Park ground water, this decrease was about 33%. The decrease of Lake Lytal Park well water conductivity was relatively less. Ch anges in pH were larger for samples from the Bradenton site, which were initially of lower pH. Th e pH of pore water was not apparently related to incubation temperature ov er 7 days, and all were in the range of 7.80 to 8.06. The increase of turbidity in pore water samples was approximately three orders of magnitude. Table 43. Pore water TDS, pH and turbidity. Avon ParkBill Evers Lake Lytal ParkClear Lake Conductivity 1919 S1215 S 5.90 mS2.29 mS pH 22o C sample 7.87.927.948.06 pH 30o C sample 7.888.017.837.97 pore water turbidity (ntu) 2150225017502400 source water turbidity (ntu) 0.53.81.52.0

PAGE 168

156 Survival results in pore water microcosms are shown in the following figures. Each figure portrays observed results fro m viability analyses of each organism, in terms of log N/N0 viable organisms per ml. For each set of water samples, time 0 results were pooled and averaged. The following results were from a selected tube for each condition set (wat er sample-temperature-organ ism group combination). Thus, data points on these charts represent the viab le concentration in the en tire extractable pore water volume from a discrete sample tube for each sample day. Duplicate tubes were assayed on each sample day. Linear regressions on the lo g-transformed survival ratios are s hown on each chart as well. Initial kinetics varied from previously described experiments in that for many, a sharp drop in infectious/culturable counts was observed between time 0 and day 1, and this was sometimes followed by a more linear decline or an increase from day 1 to day 2, after which kinetics were more stable and firstorder. Thus, linear regressions did not fit the origin (time 0) as a data point, and sometimes did not include day 1 if there was a large increase in counts from day 1 to day 2. However, all points for which data were obtained are shown on the charts. The equation for each regression is displayed on each chart, and since trendlines for 22 C conditions were always vertically above 30 C trendlines, the equations correspond to this placement (equations for 22 trendlines are always above 30 C trendlines). Results of fecal coliform survival are shown in Figure 48 for ground water samples and Figure 49 for surface water samples. It is apparent from these that once change in culturable counts stabilized by Day 1 or Day 2, a first-order model fit the data fairly well. In looking at these trendlines, it seems that there was little difference between inactivation at 22 and 30, although at 30 decline was slightly more rapid in all cases.

PAGE 169

157 Figure 48. Fecal coliform survival in saturate d limestone pore water ground water samples. Avon Park welly = -0.0398x 0.027 R2 = 0.7154 y = -0.0259x 0.2764 R2 = 0.4497 -4 -3 -2 -1 0 1 0246810121416time (d)Log N/N0 22 C data 30 C data 22 C regression 30 C regression Lake Lytal Park welly = -0.1396x 0.0392 R2 = 0.9131 y = -0.1053x 1.03 R2 = 0.9552 -4 -3 -2 -1 0 1 0246810121416time (d)Log N/N0 22 C data 30 C data 22 C regression 30 C regression

PAGE 170

158 Figure 49. Fecal coliform survival in saturate d limestone pore water surface water samples. Bill Evers reservoiry = -0.0907x + 0.1031 R2 = 0.9134 y = -0.1253x + 0.0987 R2 = 0.9634 -4 -3 -2 -1 0 1 0246810121416time (d)Log N/N0 22 C data 30 C data 22 C regression 30 C regression Clear Lake reservoiry = -0.0694x + 0.0147 R2 = 0.7349 y = -0.0934x + 0.0487 R2 = 0.9516 -4 -3 -2 -1 0 1 0246810121416time (d)Log N/N0 22 C data 30 C data 22 C regression 30 C regression Figure 50 shows ground water results for enterococci, while Figure 51 shows surface water results. The difference between 22 and 30 appears greater for ground water samples, while in surface water, inactivation is similar but slightly faster at 30 C.

PAGE 171

159 Figure 50. Enterococci survival in saturated limestone pore water ground water samples. Avon Park welly = -0.0541x 0.8407 R2 = 0.7976 y = -0.1378x 1.0742 R2 = 0.9389 -4 -3 -2 -1 0 1 0246810121416time (d)Log N/N0 22 C data 30 C data 22 C regression 30 C regression Lake Lytal Park welly = -0.0343x 0.8031 R2 = 0.774 y = -0.0863x 1.0834 R2 = 0.8766 -4 -3 -2 -1 0 1 0246810121416time (d)Log N/N0 22 C data 30 C data 22 C regression 30 C regression

PAGE 172

160 Figure 51. Enterococci survival in saturated limestone pore water surface water samples. Bill Evers reservoiry = -0.0231x 0.8414 R2 = 0.462 y = -0.0485x 0.8942 R2 = 0.8829 -4 -3 -2 -1 0 1 0246810121416time (d)Log N/N0 22 C data 30 C data 22 C regression 30 C regression Clear Lake reservoiry = -0.0722x 0.1988 R2 = 0.9135 y = -0.0609x 0.6499 R2 = 0.8706 -4 -3 -2 -1 0 1 0246810121416time (d)Log N/N0 22 C data 30 C data 22 C regression 30 C regression Results from F+ RNA coliphage microcosms are shown in Figure 52 for ground water and Figure 53 for surface water. There was visi bly a larger difference between 22 an d 30 C inactivation rates. Also,

PAGE 173

161 none of the coliphage conditions showed evidence of an increase of infectious counts on Day 2 from Day 1 numbers as was observed for fecal coliform in some cases. Figure 52. F+ RNA coliphage surv ival in saturated limestone po re water ground water samples. Avon Park welly = -0.2286x + 0.3002 R2 = 0.9641 y = -0.4321x 0.1001 R2 = 0.9496 -4 -3 -2 -1 0 1 0246810time (d)Log N/N0 22 C data 30 C data 22 C regression 30 C regression Lake Lytal Park welly = -0.2464x 0.3769 R2 = 0.9604 y = -0.4757x 0.6782 R2 = 0.9543 -4 -3 -2 -1 0 1 0246810time (d)Log N/N0 22 C data 30 C data 22 C regression 30 C regression

PAGE 174

162 Figure 53. F+ RNA coliphage surv ival in saturated limestone pore water surface water samples. Bill Evers reservoiry = -0.2818x 0.4367 R2 = 0.9883 y = -0.4875x 0.8411 R2 = 0.9904 -4 -3 -2 -1 0 1 0246810time (d)Log N/N0 22 C data 30 C data 22 C regression 30 C regression Clear Lake reservoiry = -0.297x 0.4388 R2 = 0.982 y = -0.5339x 0.8491 R2 = 0.98 -4 -3 -2 -1 0 1 0246810time (d)Log N/N0 22 C data 30 C data 22 C regression 30 C regression The regression equation for each condition set was solv ed for days (x-axis) to determine the number of days in each case that would be needed for 2-log inactivation, since th is statistic was used for

PAGE 175

163 comparison in previously described survival experiments in Chapters 3 and 4. These periods are shown for each organism and water sample-temperature combinatio n in Table 44. Also, th ese values are shown in graphic format in Figure 54. Table 44. Days for 2-log inactivation of fecal co liform, enterococci, and RNA coliphage in pore water of saturated limestone. Times based on linear regre ssion of culturable/infectious concentrations over time. Water ConditionsFecal coliformEnterococci F+ RNA coliphage Avon Park well, 22o C 502110 Avon Park well, 30o C 6774 Lake Lytal Park well, 22o C 14357 Lake Lytal Park well, 30o C 9113 Bill Evers reservoir, 22o C 23506 Bill Evers reservoir, 30o C 17232 Clear Lake reservoir, 22o C 29255 Clear Lake reservoir, 30o C 22222

PAGE 176

164 Figure 54. Days predicted for 2-log inactivation of the three organism popul ations in pore water of saturated limestone. Fecal coliform 0 20 40 60 80 2230 Temperature (oC)Days for 2-log10 (99%) decline Enterococci 0 20 40 60 80 2230 Temperature (oC)Days for 2-log10 (99%) decline F+ RNA coliphage 0 20 40 60 80 2230 Temperature (oC)Days for 2-log10 (99%) decline From these values, it is apparent that colipha ge inactivation was much more rapid under these conditions than the b acteria. Also, the relative rates and inactiv ation periods for the two bacteria were

PAGE 177

165 variable, such that fecal coliform su rvived longer than enterococci in some cases and vice versa in other cases. In all cases except one, 2-log inactivation was pr edicted to take longer at 22 than at 30. Large differences between survival in water from the two site s of each type (at each te mperature) were observed for fecal coliform in ground water, and to a lesser degree for enterococci in surface water, but only at 22 C. Otherwise, inactivation was si milar among water samples of each type. Differences for coliphage inactivation were most appare nt due to temperature, while site and wa ter type differences were small. For fecal coliform, the order of water samples producing mo st rapid inactivation to least rapid was Lake Lytal Park well water, the two surface water samples (close together), and Avon Park well water with the least rapid inactivation. Conversely, enterococci inaction was most rapid in Avon Park water, with no grouping due to water type at 22 C. At 30 C, inactivation in the two ground wa ter samples was more rapid than the two surface water samples. Results from survival in PBS solutions for each organism at 22 and 30 are shown in Figure 55. Inactivation in PBS for this experiment was much more rapid than had been observed on average for PBS control solutions for experiments described in Chapters 3 and 4. This was the case for all three organisms. For fecal coliform, 2-log inactivation in PBS, based on solution of the regression equations, occurred by 7 days at both 22 and 30. Previous averages for 2-log declines in PBS for water column studies with environmental waters were 85 days at 22 and 39 da ys at 30. These average predicted days for 2-log decline in PBS are shown in Appendix 4-F. The differences for enterococci were 10 days (pore-water control) vs. 45 days (previous average) at 22 and 4 days (pore-water control) vs. 19 days (previous average) at 30 (Appendix 4-G). For F+ RNA coliphage, days for 2-log decline in PBS were 5 and 6 days at 22 and 30 respectively, while averages for previous trials were 38 days at 22 and 14 days at 30 (Appendix 4-H).

PAGE 178

166 Figure 55. Results of survival in PBS at 22 and 30 C for A. fecal coliform, B. enterococci, and C. F+ RNA coliphage. Regression trendlines shown for each data set. A. Fecal coliform, PBS, 22o & 30o C, limestone pore-water trial-6 -5 -4 -3 -2 -1 0 1 0246810DaysLog N/N0 PBS, 22 C actual PBS, 22 C pred. PBS, 30 C actual PBS, 30 C pred. B. Enterococci, PBS, 22o & 30o C, limestone pore-water trial-6 -5 -4 -3 -2 -1 0 1 0246810DaysLog N/N0 PBS, 22 C actual PBS, 22 C pred. PBS, 30 C actual PBS, 30 C pred. C. F+ RNA coliphage, PBS, 22o & 30o C, limestone pore-water trial-6 -5 -4 -3 -2 -1 0 1 0246810DaysLog N/N0 PBS, 22 C actual PBS, 22 C pred. PBS, 30 C actual PBS, 30 C pred.

PAGE 179

167 Discussion The underlying purpose of this experiment was to identify and describe differences between survival of key microbes an d conditions when in pore wa ter of saturated limestone as compared to when in the water column, since water column studies were employed for more extensive experiments described in Chapter 4. A further change was the aerated state of water column microcosms, whereas the limestone studies described in this chapter were not disturbe d until each sample point an d had considerable solid volume within each that limited mixing, thus these micr ocosms had limited air exchange after being set. As compared to water column studies, the conditions in saturate d limestone tubes differed in several respects that affected basic parameters of pore water. Before water from each test source was filled with crushed limestone to simulate an aquifer, the limestone had been conditioned in ground water from each of the two sites (Avon Park well water from Br adenton and Lake Lytal Park well water from West Palm Beach). The purpose of this activity was to simulate conditions involved when injecting surface water into an aquifer, and to compare to static conditions such as continuous pr esence of the gr ound water. This conditioning ground water was dr ained before limestone was used in microcosm tubes. However, due to surface tension/capillary effect s, a proportion of co nditioning ground water could not be removed, particularly from around smaller grains. Therefore, in cases where surface water was employed, there was a partial mixing of the resident ground water and su rface water which was added. This mixing, and the presence of limestone itself, resulte d in changes to the pH and TDS of pore water in microcosm tubes. Table 42 and Table 43 reflect values of these parame ters in water sources and pore water measurements. Regarding TDS, there was an expected increase of conductivity for reservoir water after contact with limestone and residual ground water on the grains. Conductivity of Bill Evers reservoir water increased from about 350 S to about 1200 S, and Clear Lake reservoir water increased from 320 S to about 2.3 mS or 2300 S. This relatively greater change would be expected due to the gr eater TDS concentration in Lake Lytal Park ground water than in Avon Park water. However, th ere was also a decrease of measured conductivity in Avon Pa rk and Lake Lytal Park pore water from values in sampled wa ter. Thus, rather large increases of surface water conductivity were observed after exposure to the pre-conditioned limestone. This is mostly due to mixing with residual high-TDS ground water and possibly due to some

PAGE 180

168 dissolution from the calcium carbonate rock. The reduction of conductivity in ground water tubes was possibly due to measurement error due to small samp le volumes or ion exchange which affected the composition of TDS. Changes were also observed in pH of pore water. In general, a buffering effect was observed, to a consistent pH of near 8. Three water sources were at in itial pH of less than this, from 7.1 to 7.6. All were measured to be about 7.9 after saturation of the limest one. Also, Clear Lake water was already at pH 8.1, and this was reduced slightly to about pH 8.0. These microcosms we re very small scale representations of the process that may be involved in an ASR scenario. In reality, the largest im pacts on injected water chemistry would be in a “mixing” zone at the edge of the stored water zone, and de eper within this area of stored water, different ef fects would be observed. There would be less mixing with the displaced ground water and more interaction and dissol ution from the aquifer rock. Howeve r, in general th e water chemistry deep within the stored water zone would be more lik e that of the surface water source than at the mixing zone at the edge of the injected water displacement. In these experiments, consider able turbidity was released fro m the crushed limestone. This occurred after thorough rinsing initially after crushing that largely rem oved particles smaller than sand-size grains. Thus, continued abrasion and possibly dissolutio n during handling of the limestone “sand” released colloidal-size particles. These settled within microcosms such that after saturating the limestone, the pore water appeared clear by 1 day. However, once pore water was sampled the turbidity was resuspended and thus was present in samples when analyzed for micr obial concentrations. Th is likely had a role in distribution of both seeded and native microorganisms. The majority of both would likely have been adsorbed to solid particles, including the small clay-size particles that resulted in observ ed turbidity. Since much of this small particulate matte r was recovered in the sampling proce ss, it is likely that more of the attached microorganisms were recovered than would have been the case if small pa rticles were not present and the pore water was truly that of consolidated rock as in an aquifer. It has been reported that inactivation rates of viruses may vary between attach ed at free particles, being generally slower when attached (Schijven and Hassanizadeh 2000). Thus, it is possible that the result of small particles in these studies was a more-conservative view of microorganism su rvival, in that the surviv al of a large number of

PAGE 181

169 attached microorganisms was included and this may be longer than survival of free, unattached cells or virus particles. Another difference between pore water microcosms and water column experiments is the degree of aeration and its effect on dissolved oxygen (DO). Although DO was not measured, the pore water tubes were not disturbed during the course of incubation, si nce each tube represented a separate sample point, which was sampled destructively to obtain water. However, water column microcosms described in Chapter 4 were aerated at sampling points. Also, the solid media in each tube did not allow mixing or agitation of the pore water. Thus, it is reasonable to assume that after initial creation of microcosm tubes for saturated limestone experiments, gas exchange with the headspace of each tube was more limited and as a result, any aerobic bacter ial activity would consume oxygen and resu lt in more oxygen limited conditions than was present in bulk water column microcosms. Table 44 presents values for predicted days to achieve 2-log inactivation of fecal coliform, enterococci, and F+ RNA coliphage in pore water micr ocosms. From these data and the graphic depiction of them in Figure 54, some trends were evident. Overall, inactivation periods were on the order of 2 to 8 weeks, but all times beyond 14 days were outside the experimental duration and thus changes in kinetics would not be accounted for. Ente rococci population survival was sim ilar to that of fecal coliform on average. However, inactivation was more rapid in A von Park well water and not as rapid in Lake Lytal Park water, and a larger increase in inactivation due to higher temperatur e was observed in ground water. In the surface water microcosms, inactiv ation of enterococci was very sim ilar to that of fecal coliform at 30 C. Inactivation of F+ RNA coli phage was uniformly more rapid than the bacter ia, with little difference between the types of water and the two sources of each. Reduced survival due to higher temperature was by a factor of about to in terms of days for 2-log decline. These inactivation periods may be compared to those obtained from experiments from bulk water column microcosms to gain some insight on the eff ects of solid media and limited aeration/gas exchange. Days for 2-log inactivation from expe riments in Chapter 4 (as shown in tables of first-order inactivation rates from raw water sources) were combined with thos e obtained from this experiment to offer a direct comparison. Figure 56 shows this comparison for f ecal coliform. These inactiv ation periods indicate that in ground water the effect of a solid matrix was mixed: for Avon Pa rk samples, inactivation was more rapid

PAGE 182

170 when suspended in the bulk water column, but the oppo site was true in Lake Lytal Park water. Also, there was almost no difference between survival in Lake Ly tal Park water at 30 C in the water column and in pore water. In surface water sources, the difference was consistent with more rapid inactivation in the water column than pore water. Inte restingly, the two sets of column s for Bill Evers reservoir and Clear Lake reservoir are very similar in the relative effects of temperature and the comparison of pore water to water column results, being almost identical in appear ance except for scale. In activation was slightly less rapid on average in Clear Lake water than in Bill Ever s reservoir water, but othe rwise the two sets reflect the same pattern and approximate relative differences. Figure 56. Comparison of days predicted for 2-log inactivation of fecal coliform in saturated limestone pore water and suspended in water column. Fecal coliform 0 20 40 60 80 AP 22AP 30LP 22LP 30BE 22BE 30CL 22CL 30 Water source/temperature (oC)Days for 2-log10 (99%) decline Pore water Water column AP = Avon Park well LP = Lake Lytal Park well BE = Bill Evers res. CL = Clear Lake res. A comparison of entero cocci inactivation is shown in Figure 57. In ground water samples, little difference existed between pore water su rvival and survival in the water column, the differences and trends observed for fecal coliform survival were not observed. Inactivation of enterococci was much more rapid in Avon Park samples in pore water, but was less rapi d than fecal coliform in Lake Lytal Park water at 22 and was the same at 30 C. Like fecal coliform, while there was no clear and consistent effect of placing the seeded ground water in contact with rock material, the effect on inactivation in surface water was

PAGE 183

171 definite with the same trend as for fecal coliform. Survival was greatly enhanced in pore water compared to survival in the bulk water column. Figure 57. Comparison of days predicted for 2-log inactivation of enterococci in saturated limestone pore water and suspended in water column. Enterococci 0 20 40 60 80 AP 22AP 30LP 22LP 30BE 22BE 30CL 22CL 30 Water source/temperature (oC)Days for 2-log10 (99%) decline Pore water Water column AP = Avon Park well LP = Lake Lytal Park well BE = Bill Evers res. CL = Clear Lake res. Figure 58. Comparison of days predicted for 2-log inactivation of F+ RNA coliphage in saturated limestone pore water and suspended in water column. F+ RNA coliphage 0 5 10 15 20 AP 22AP 30LP 22LP 30BE 22BE 30CL 22CL 30 Water source/temperature (oC)Days for 2-log10 (99%) decline Pore Water Water column AP = Avon Park well LP = Lake Lytal Park well BE = Bill Evers res. CL = Clear Lake res.

PAGE 184

172 Figure 58 shows a similar comparison for the F+ RNA coliphage. However, the scale of this chart is that of the bacterial populations’ charts. Inactiva tion was more rapid in the water column than in pore water except in Bill Evers reservoir water, in which the two were fairly sim ilar at both temperatures. Inactivation was rapid in all pore wa ter conditions compared to the bacter ia, with periods for 2-log decline all under 10 days. Temperature is likely more of a co ntrolling factor for RNA phag e than for bacteria, such that differences between conditions in pore water and the water column are not as significant. The most striking trend from this experiment is the similarly large increase of survival capacity for fecal coliform and enterococc i in surface water when in pore water of saturated limestone. The effect of examining survival in ground water between these two environments was not as consistent, and thus less dramatic; little difference wa s observed at all for enterococci and the effect was inconsistent with the fecal coliform. For enterococci, this observation is similar to what wa s observed for inactivation in raw vs. pasteurized samples: pasteurization ha d a larger effect in surface water th an in ground water. The impact of solid media in this experiment, and the more pr onounced trend on average in surface water, perhaps indicates that aerobic activity of native microorganisms has a definite antagonistic impact on seeded fecal bacteria. However, there is some protection from these effects when in a solid matrix such as when attached to particles in saturated soil, aquifers, or se diments. Clearly, in studies on surface water sediments and fecal bacteria, it has been shown that sediment ha rbors larger numbers of these organisms, and often may allow their re-growth (Buckley 1998; Desmarais 2 002). It may be that protozoan predators have a large influence on survival of some fecal bacteria. Davies et al. demonstrated that i nhibition of protozoan predators greatly enhanced E. coli survival in sediments, and actually allowed for net increase in culturable counts (Davies 1995). Sherer et al. evaluated survival of fecal coliform and fecal streptococci in stream sediments of two grain size compositions at 8 C. Inac tivation rates they found in sediment were on the order of 0.01 to 0.03 log/d, equating to days for 2-log decline as we ha ve reported of 200 to 66. Given the much lower temperature used, this considerably longer survival would not be unexpected. Also, in their study they evaluated survival in the supernatan t (water column) and found inactivatio n to be 3 10 X more rapid. Blanc and Nasser evaluated die-off of 4 viruses including MS-2 (an F+ RNA coliphage) at 10 in soil saturated with ground water. They quantified 1.0 and 2.4 log inactivation in 20 days for MS-2, also much

PAGE 185

173 less rapid inactivation than found in this experiment likely due to the lower temperature (Blanc and Nasser 1996). Given the specific interest in limestone sa turated with surface water, as in an ASR injection scenario, our study is the first re port among those found that describes inactivation of water quality indicator microorganisms in such conditions. Particular ly, the temperatures of concern are not frequently used for ground water evaluations. Clearly, inactivation of F+ RNA coliphage under these temperature conditions is quite rapi d compared to that of bacter ia. Thus, their presence as an indicator would be a sign of recent sewage or other fecal contamination. Th e longer survival of enterococci and fecal coliform, particularly in the saturated rock studies employed here, would make them more conservative indicators for evaluating removal of enteric microorganisms due to inactivation in aquifer injection test studies.

PAGE 186

174 CONCLUDING REMARKS The work presented in this disse rtation pursued two main objectives : description of virus transport through shallow ground wa ter from a mounded septic system toward s a seasonally-inundated area, with emphasis on transport velocity and direction of spread, and an evaluation of survival behavior for a suite of fecal indicator organism po pulations in several situations. These studies were performed with a distinct regional emphasis, attempting to focus on conditions that might be observed in the subsurface of the Florida peninsula or other humid subtropical regions. Th e sites for virus transport studies were typified by a mounded septic system drainfield, something common to rural areas with high water tables, such that additional vertical distance between se ptic tank drain pipes and the satura ted zone is often required. This type of system would be unusual in most other areas of the continental U.S., bu t conditions typically found in Florida and other areas of the southeastern U.S. may require such systems. For survival studies, specific attention was paid to temperature c onditions and TDS concentrations as found in Florida ground waters, to offer descriptions of surv ival behavior found in th ese conditions, conditio ns for which relatively fewer data exist from other ground water survival studies. Also, survival in higher temperatures was contrasted to that in low temperatures found in aquifers of colder regions. In addition, survival studies were performed with composite organism populations isolated from local surface water or secondary waste water effluent, and represented a better approximation of organisms found in the local environments of concern than type strain isolates (i.e. MS-2). This regional emphasis to enteric microorganism surv ival was also applied to water samples typical of ASR projects in the state of Florida, from two sites. Lastly, the pore-water survival study was performed usin g a solid matrix of Ocala limest one, which composes much of the Floridan aquifer system. The pore-water study provided a miniature simulation of ASR injection, where surface water was used to saturate the limestone matrix and survival of enteric microorganisms in the pore water was compared to that in ground water and in water without solid substrate present.

PAGE 187

175 The viral transport study involving the Salmonella bacteriophage PRD-1 documented travel velocities at two sites near seasonally-inundated areas. This study was valuable in the context of ground water quality since septic systems an d other onsite waste water disposal systems are in wi de use throughout the state and world. Such systems have been shown to have significant impact s on ground water and even surface water quality, particularly in areas that are more vulnerable due to shallow ground water and a large degree of connectivity between ground water and surf ace water. Thus impacts of septic systems are frequently very site-specific, and regional or local focus can be important in studies evaluating these impacts. For instance, septic tank effects on mari ne water contamination in coastal areas have been documented in the Florida Keys (Paul 1995; Paul 1995; Paul 1997; Paul 2000). The studies described in Chapter 1 described transport of viruses from mounded septic systems to the underlying ground water at the natural grade of the immediately surrounding areas and towards nearby SIA’s. However, these surface water areas were dry at the time of the study due to a continuing drought, during what would have been the end of the rainy season. Thus, these studies did no t represent trans port of the viruses under worst-case situations for environmental impact s (actually the conditions probably represented best-case conditions for minimal environmental impacts). Still, the tracer studie s documented transport of vi rus to the ground water under each drainfield mound, and migration in a plum e towards the SIA at each site. Also, tracer viruses were detected in the ground water beneath the dry SI A at the Duval County site. Estimated average travel velocities of PRD-1 were about 0.3 m/d at the Duval Co unty site, and were similar at the Lake County site at 0.36 m/d. These are slightly less than the estima ted average ground water pore water velocity of about 0.5 m/d at each site, calculated from modeling of ob served heads at each site and estimated hydraulic conductivities of the sandy soils (Bro wn 2001). Since the drainfield at each site is approximately 10 m from the adjacent SIA, under wors e conditions transport of enteric organisms through the ground water from the mounded septic systems to beneath and possibl y into surface water of the SIA might occur in less than 33 days. Given the average inactivation rates compiled from reviewed studies on pathogenic virus inactivation in Chap ter 2 at temperatures above 20 C, this time would be adequate to result in significant reductions ( > 2 log10) of all but the most resistant viruses (such as hepatitis A) at temperatures that would be found during summer months due to inactivation alone. However, the ranges of estimated inactivation rates from all the potentially pathogenic viruses in reviewed studies included lower bounds at these

PAGE 188

176 temperatures that would allow less th an 2-log reduction by 33 days by in activation. In a ddition, adsorption would result in additional cons iderable reductions of vi rus concentrations. Still, these predictions are based on estimates of inactivation from a wide variety of ot her studies in ground water of different areas. It would be most beneficial to continue studies on vi ral transport from mounded septic systems with the hopes of evaluating transpor t behavior in conditions that are less favorable to virus removal and reduced migration. For instance, this type of study would be more informative for worst-case scenarios if it were performed during a time that the SIA was indeed i nundated and thus surface water was present, and interaction of the shallow ground water with water in the wetland might occur. Naturally, such in-situ studies are difficult to arrange, partic ularly when the cooperation of priv ate landowners is required. But given the indications of potential tr ansport behavior under dry conditions observed by our studies, further investigation to document potential input of virus from septic systems to surface water in this type of area would be valuable to gauge effectiveness of current practices for construction of mounded septic drainfields in protecting nearby waters from contamination. The other issue of regional concern investigated by this work was survival of several organisms in conditions and water samples indicative of those that may be encountered during the course of aquifer storage and recovery projects in Florida. ASR is continuing to grow as an important water supply management technology, and it is imperative that the science surrounding its use maintain pace with its implementation. Specifically, there is continued interest by involved parties in the potential natural attenuation of enteric microorganisms resident in stored surface water once in the subsurface, should untreated surface water be employed in ASR systems. Due to the scal e of planned ASR construction in Florida, investigations with a regional emphasis ar e critical. The Comprehensive Everglades Restoration Project plans to make use of hundreds of ASR wells storing/recovering up to over a billion gallons per day. While no current use of raw surface water for ASR is occu rring, the interest, and concern, over its potential use exists. The survival studies pe rformed as desc ribed in Chapters 3 through 5 focused on conditions that could possibly control dynamics of inactivation of se veral microorganism groups. Naturally, temperature was found to play a large role in relative inactivati on, both in controlled-TDS water experiments and in environmental water sources. Thus, comparisons to ground water microbiology studies done in colder

PAGE 189

177 regions should take temperature in to account, and ou r results further suggest th at inactivation of enteric microorganisms under environmental conditions found in subtropical regions is likely to be more rapid. But there was also indication of a la rge role played by native microbial populations in controlling survival of seeded enteric microbes. Pasteurization of water samples for survival studies described in Chapter 4 generally resulted in longer survival of seeded bacter ia and virus compared to raw water conditions. This effect was more pronounced in surf ace water in some cases, as for DNA coliphage and enterococci. If the communities of water microorganisms in surface water play a strong role in surviv al of potential pathogens, there could be implications for inj ection of surface water to the subsurface. Such a change of environment could negatively impact surface water microbial comm unities, via redox changes, pressure effects, and chemical effects such as mi xing with water high in H2S or other reduced species. Furthermore, if protozoan predators are responsible fo r decline of fecal bacteria, for in stance, their larger size may limit their distribution and activity in th e subsurface. Indeed, studies performed with saturated limestone indicated that survival of bacteria seeded in surface water was enhanced in pore water over that in the bulk water column. Along with effects on native microbial communities, othe r factors could influence dynamics of survival for enteric mi croorganism themselves such as ionic constituents, H2S or NH4 + ion, or metals that may become more solubilized in low-TDS surface water which disp laces ground water. One study indicated that MS-2 inactivation wa s enhanced by the presence of an air-water-solid interface, something that would be absent in the subsurface (Thompson and Yates 1999). The line of work described in this dissertation represents a novel examination of the potential dy namics of public-health-related microorganism survival in surface water injection scenarios, particularly in this region. The transport and survival experiments presen ted here represent what is in some ways a preliminary look at some importan t aspects of water quality related microorganisms in some interesting situations and conditions. The results described here ca n lead to additional examin ations for future work. The indications of important roles played by backgr ound microbial co mmunities for controlling survival of fecal microbes in ground water or in surface water injected to aquifers can lead to more concise, detailed research to confirm and describe thei r interactions in these water conditio ns. Of particular interest would be conditions surrounding the implementation of ASR injection in Florida, an d the fate of enteric organisms in representative surface water sources. Such studies could more-speci fically examine survival

PAGE 190

178 in surface water sources, partic ularly when in saturated limestone pore water, and as mixtures with native ground water to simulate the conditio ns in an ASR storage zone of the Upper Floridan aqui fer. The work described in Chapter 5 showed obvious impact of mixing with ground water that was saturating the rock prior to the study, as evidenced by TDS changes. It would be beneficial to look at larger-scale model environments to gauge the survival behavior of ente ric microbes more in the center of an ASR storage zone. In general, the microcosms employed for su rvival studies described here presented a highly miniaturized situation. Future work that employs more cooperation with hydrogeological and engineering interests would be very valuable. For example, surv ival studies using intact limestone cores or blocks saturated with seeded surface water might paint a more realistic picture of the environment in an ASR storage zone. Specific evaluation of what mechanisms l ead to enhanced survival of organisms in saturated solid media over the water column will also continue to be valuable. In the case of these experiments, aspects of the pore water microcosms that could have had an impact on survival included pH changes, reduction of aeration/oxygen tensions, reduced micr obial activity, and perhaps attachment to surfaces, particularly the colloidal particles which were largely recovered for surviving organism analyses. The ground water environment of Florida presents some uni que conditions, conditions which could have a profound imp act on microbial dynamics. In additi on, the multiple and growing number of ways which humans impact the ground water environm ent present continuing ch allenges for science to maintain pace in describing these impacts. Since gro und water is such a critical and, oftentimes poorlyunderstood, environment, con tinued work is always necessary. The re search presented here is an important contribution towards unde rstanding the effects of human activ ities on this precious resource.

PAGE 191

179 REFERENCES Alamanos, Y., Maipa, V., Levidiotou, S. and Gessouli, E. (2000). "A community waterborne outbreak of gastro-enteritis attributed to Shigella sonnei." Epidemio logy and Infection 125(3): 499-503. Alley, W. M., Reilly, T. E. and Fr anke, O. L. (1999). Su stainability of GroundWater Resources, U.S. Geological Survey Circular 1186 Denver, U.S. Geological Survey. Alvarez, M. E., Aguilar, M., Fountai n, A., Gonzalez, N., Rascon, O. and Saenz, D. (2000). "Inactivation of MS-2 phage and poliovirus in groundwater." Can J Microbiol 46(2): 159-165. APHA, AWWA and WEF (1992). Standard Methods for the Examination of Water and Wastewater Baltimore, MD. Bales, R. C., Li, S. M., Maguire, K. M., Yahya, M. T., Gerba, C. P. and Harvey, R. W. (1995). "Virus and Bacteria Transport in a Sandy Aquifer, Cape-Cod, Ma." Ground Water 33(4): 653-661. Banning, N., Toze, S. and Mee, B. J. (2002). "Escherichia coli survival in groundwater and effluent measured using a combination of propidium iodide and the green fluorescent protein." J Appl Microbiol 93(1): 69-76. Berndt, M. P., Oaksford, E. T., Mahon, G. L. and Schmidt, W. (1998). Groundwater. Water Resources Atlas of Florida Fernald, E. A. and Purdum E. D. Tallahassee, Institute of Science and Public Affairs, Florida State University: 38-63. Bitton, G., Farrah, S. R., Ruskin, R. H., Butner, J. and Chou, Y. J. (1983). "Survival of Pathogenic and Indicator Organisms in Groundwater." Ground Water 21(4): 405-410. Bitton, G., Pancorbo, O. C. and Farrah, S. R. (1983). "Effect of hydrostatic pressure on poliovirus survival in ground water." Ground Water 21(6): 756-758. Blanc, R. and Nasser, A. (1 996). "Effect of effluent quality and temper ature on the persiste nce of viruses in soil." Water Sci Technol 33(10-11): 237-242. Borchardt, M. A., Bertz, P. D., Spencer, S. K. and Battig elli, D. A. (2003). "Inciden ce of enteric viruses in groundwater from household wells in Wisconsin." Appl Environ Microbiol 69(2): 1172-1180. Brown, M. T., Annable, M. D., Delfino, J. J., Jawitz, J. W., Cohen, M., Hall, E., Harden, H. S., Chanton, J. P., Burnett, W., Rose, J. B., Paul, J. H., Griffin, D., Lipp, E. K. and John, D. E. (2001). Determination of an appropriate onsite sewage system setback distance to seasonally inundated areas. Tallahassee, FL, Florida Department of Health. Buckley, R., Clough, E., Warnken, W. and Wild, C. (1 998). "Coliform bacteria in streambed sediments in a subtropical rainforest conservation reserve." Water Res 32(6): 1852-1856.

PAGE 192

180 Callahan, M. R., Rose, J. B. and Paul, J. H. (2 001). Bacteriological and pathogenic water quality assessment of the upper reaches of the Chassahowitzka river. Citrus County, FL, Citrus County Department of Public Works Utility Division. Corbett, D. R., Dillon, K. and Burnett, W. (2000). "Tracing groundwater flow on a barrier island in the north-east Gulf of Mexico." Estu arine Coastal and Shelf Science 51(2): 227-242. Crabill, C., Donald, R. Snelling, J., Foust, R. and Southam, G. (1999). "The impact of sediment fecal coliform reservoirs on seasonal water qu ality in Oak Creek, Arizona." Water Res 33(9): 21632171. Craun, G. F. and Calderon, R. L. (1997). Microbial risks in groundwater systems: epidemiology of waterborne outbreaks. Under the Micros cope: Examining Microbes in Groundwater Denver, CO, American Water Works Research Foundation. Craun, G. F., Hubbs, S. A., Frost, F., Calderon, R. L. and Via, S. H. (1998). "Waterborne outbreaks of cryptosporidiosis." J ournal American Wate r Works Association 90(9): 81-91. Craun, G. F., Nwachuku, N., Calderon, R. L. and Craun, M. F. (2002). "Outbreaks in drinking-water systems, 1991-1998." Journal of Environmental Health 65(1): 16-23. Davies, C. M., Long, J. A. H., Donald, M. and Ashbolt, N. J. (1995) "Survival of Fecal Microorganisms in Marine and Fresh-Water Sediments." Appl Environ Microbiol 61(5): 1888-1896. Deborde, D. C., Woessner, W. W., Kiley, Q. T. and Ball, P. (1999). "Rapid transport of viruses in a floodplain aquifer." Water Res 33(10): 2229-2238. DeBorde, D. C., Woessner, W. W., Lauerman, B. and Ball, P. N. (1998). "Virus occurrence and transport in a school septic system and unconfined aquifer." Ground Water 36(5): 825-834. Desmarais, T. R., Solo-Gabriele, H. M. and Palmer, C. J. (2002). "I nfluence of soil on fecal indicator organisms in a tidally influenced subtro pical environment." Appl Environ Microbiol 68(3): 11651172. Dillon, K. S., Corbett, D. R., Chanton, J. P., Burnett, W. C. and Furbish, D. J. (1999). "The use of sulfur hexafluoride (SF6) as a tracer of septic ta nk effluent in the Florida Keys." J Hydrol 220(3-4): 129140. Dowd, S. E. and Pillai, S. D. (1997 ). "Survival and transport of select ed bacterial pathoge ns and indicator viruses under sandy aquifer cond itions." J Environ Sci Health Part A-Environ Sci Eng Toxic Hazard Subst Control 32(8): 2245-2258. Dowd, S. E., Pillai, S. D. Wang, S. Y. and Corapcio glu, M. Y. (1998). "Delin eating the specific influence of virus isoelectric point and size on virus adsorption and transport through sandy soils." Appl Environ Microbiol 64(2): 405-410. Drew, R. (2001). Aquifer Storage and Recovery – UIC Class V Wells: DEP’s Perspective. Orlando, FL. Evison, L. M. (1988). "Comparative Studies on the Survival of Indicator Organisms and Pathogens in Fresh and Sea-Water." Water Sci Technol 20(11-12): 309-315. FDOH (1999). Annual report, 1999. Tallahassee, FL, State of Florida, Department of Health. Filip, Z., Kaddumulindwa, D. and M ilde, G. (1988). "Sur vival of Some Pathog enic and Facultative Pathogenic Bacteria in Gro undwater." Water Sci Technol 20(3): 227-231.

PAGE 193

181 Fleisher, J. M. (1991). "A reanalysis of data suppor ting U. S. federal bacterio logical water quality criteria governing marine recreational waters." Resear ch Journal, Water Pollution Control Foundation 63(3): 259-265. Fontes, D. E., Mills, A. L ., Hornberger, G. M. and Herman, J. S. (1991). "Physical and Chemical Factors Influencing Transport of Microorganisms through Porous-Media." Appl Environ Microbiol 57(9): 2473-2481. Freire-Santos, F., Oteiza-Lopez, A. M., Vergara-Castib lanco, C. A. and Ares-Mazas M. E. (1999). "Effect of salinity, temperature and storage time on mouse experimental infection by Cryptosporidium parvum." Vet Parasitol 87(1): 1-7. Garcia-Lara, J., Menon, P., Servais, P. and Billen, G. (1991). "Mortality of Fecal Bacteria in Seawater." Appl Environ Microbiol 57(3): 885-888. Gerba, C. P. (1984). "Applied and Theoretical Aspects of Virus Adsorp tion to Surfaces." Advances in Applied Microbiology 30: 133-168. Gerba, C. P. and Bitton, G. (19 84). Microbial pollutants: their su rvival and transp ort pattern to groundwater. Groundwate r Pollution Microbiology Bitton, G. and Gerba, C. P. New York, NY, John Wiley and Sons: 65-88. Griffin, D. W., Gibson, C. J., Lipp, E. K., Riley, K., Paul, J. H. and Rose, J. B. (1999). "Detection of viral pathogens by reverse transcriptase PCR and of mi crobial indicators by standard methods in the canals of the Florida Keys." Appl Environ Microbiol 65(9): 4118-4125. Harden, H. S., Chanton, J. P., Rose, J. B., John, D. E. and Hooks, M. E. (2003). "Comparison of sulfur hexafluoride, fluorescein and rhodamine dyes an d the bacteriophage PRD1 in tracing subsurface flow." J Hydrol 277(1-2): 100-115. Harvey, R. W. (1997). In Situ and laboratory methods to study su bsurface microbial transport. Manual of Environmental Microbiology McInerney, M. J. and Hurst, C. J. Washington, D.C., ASM Press: 586-599. Hsu, F. C., Shieh, Y. S. C., Vanduin, J., Beekwilder, M. J. and Sobsey, M. D. (1995). "Genotyping MaleSpecific Rna Coliphages by Hybridization with O ligonucleotide Probes." Appl Environ Microbiol 61(11): 3960-3966. Hurst, C. J. (1988). "E ffect of Environmental Variables on En teric Virus Survival in Surface FreshWaters." Water Sci Technol 20(11-12): 473-476. Hurst, C. J., Gerba, C. P. and Cech, I. (1980). "E ffects of Environmental Variables and Soil Characteristics on Virus Survival in Soil." Appl Environ Microbiol 40(6): 1067-1079. Janakiraman, A. and Leff, L. G. (1999). "Comparison of survival of different species of bacteria in freshwater microcosms." J Freshw Ecol 14(2): 233-240. Jansons, J., Edmonds, L. W., Speight, B. and Bucens, M. R. (1989). "Survival of Viruses in Groundwater." Water Res 23(3): 301-306. Jin, Y., Chu, Y. J. and Li, Y. S. (2000). "Virus removal and transport in saturated and unsaturated sand columns." J Contam Hydrol 43(2): 111-128.

PAGE 194

182 Kersters, I., Huys, G., VanDuffel, H., Vancanneyt, M. Kersters, K. and Verstraete, W. (1996). "Survival potential of Aeromonas hydrophila in freshwaters and nutrient-poor waters in comparison with other bacteria." J Appl Bacteriol 80(3): 266-276. Keswick, B. H., Gerba, C. P., Secor, S. L. and Cech, I. (1982). "Survival of Enteric Viruses and Indicator Bacteria in Groundwater." J Environ Sci Hea lth Part A-Environ Sci Eng Toxic Hazard Subst Control 17(6): 903-912. Klein, J. and Ziehr, H. (1990). "Immobilization of Microbial-Cells by Adsorption." Journal of Biotechnology 16(1-2): 1-16. Krekeler, C., Ziehr, H. and Klein, J. (1991). "Influence of Physicoche mical Bacterial Surface-Properties on Adsorption to Inorganic Porous Supports." Applied Microbiology and Biotechnology 35(4): 484490. Lipp, E. K., Farrah, S. A. and Rose, J. B. (2001). "Assessment and impact of microbial fecal pollution and human enteric pathogens in a co astal community." Mar Pollut Bull 42(4): 286-293. Lipp, E. K., Jarrell, J. L., Griffin, D. W., Lukasik, J., Jacukiewicz, J. and Rose, J. B. (2002). "Preliminary evidence for human fecal contamination in cora ls of the Florida Keys USA." Mar Pollut Bull 44(7): 666-670. Macler, B. A. and Merkle, J. C. (2000). "Current knowledge on groundwater microbial pathogens and their control." Hydrogeology Journal 8(1): 29-40. Marella, R. L. (1999). Water withdrawals, use, discharg e, and trends in Florida, 1995. Tallahassee, Florida, United States Geological Survey. McKay, L. D., Cherry, J. A., Bales, R. C., Yahya, M. T. and Gerba, C. P. (1993). "A Field Example of Bacteriophage as Tracers of Fract ure Flow." Environ Sci Technol 27(6): 1075-1079. Medema, G. J., Bahar, M. and Schets, F. M. (1997). "Survival of Cryptosporidium parvum, Escherichia coli, faecal enterococci and Clostridium perfringens in river water: Influence of temperature and autochthonous microorganisms." Water Sci Technol 35(11-12): 249-252. Miettinen, I. T., Zacheus, O., von Bonsdorff, C. H. and Vartiainen, T. (2001). "Waterborne epidemics in Finland in 1998-1999." Water Sci Technol 43(12): 67-71. Nasser, A. M. and Oman, S. D. ( 1999). "Quantitative assessment of th e inactivation of pathogenic and indicator viruses in natural water sources." Water Res 33(7): 1748-1752. Newby, D. T., Pepper, I. L. and Maier, R. M. (2 000). Microbial Tr ansport. Environmental Microbiology Maier, R. M., Pepper, I. L. and Gerb a, C. P. San Diego, Academic Press: 147 175. Nicosia, L. A., Rose, J. B., Stark, L. and Stewart, M. T. (2001). "A field study of virus removal in septic tank drainfields." J Environ Qual 30(6): 1933-1939. NRC (2001). Aquifer Storage and Recovery in the Comprehensive Everglades Restoration Plan Washington, D.C., National Academy Press. Olsen, R. H., Siak, J. S. and Gray, R. H. (1974). "C haracteristics of Prd1, a Plasmid-Dependent Broad Host Range DNA Bacteriophage." J Virol 14(3): 689-699.

PAGE 195

183 Paul, J. H., McLaughlin, M. R., Griffin, D. W., Lipp, E. K., Stokes, R. and Rose, J. B. (2000). "Rapid movement of wastewater from on-site disposal systems into surface waters in the Lower Florida Keys." Estuaries 23(5): 662-668. Paul, J. H., Rose, J. B., Br own, J., Shinn, E. A., Miller, S. and Fa rrah, S. R. (1995). "Viral Tracer Studies Indicate Contamination of Marine Waters by Sewage Disposal Practices in Key-Largo, Florida." Appl Environ Microbiol 61(6): 2230-2234. Paul, J. H., Rose, J. B., Jiang, S., Kellogg, C. and Shinn, E. A. (1995). "Occurrence of Fecal Indicator Bacteria in Surface Waters and the Subsurface A quifer in Key-Largo, Fl orida." Appl Environ Microbiol 61(6): 2235-2241. Paul, J. H., Rose, J. B., Jiang, S. C., Zhou, X. T., Cochran, P., Kellogg, C., Kang, J. B., Griffin, D., Farrah, S. and Lukasik, J. (1997). "Evi dence for groundwater and surface marine water contamination by waste disposal wells in the Florida Keys." Water Res 31(6): 1448-1454. Pavelic, P., Ragusa, S. R., Flower, R. L., Rinck-Pfeiff er, S. M. and Dillon, P. J. (1998). "Diffusion chamber method for in situ measurement of pathogen inactivation in groundwater." Water Res 32(4): 11441150. Pedley, S. and Howard, G. (1997). "The public health implications of microbiological contamination of groundwater." Q J Eng Geol 30: 179-188. Pieper, A. P., Ryan, J. N., Harvey, R. W., Amy, G. L., Illangasekare, T. H. and Metge, D. W. (1997). "Transport and recovery of bacteriophage PRD1 in a sand and gravel aquifer: Effect of sewagederived organic matter." Environ Sci Technol 31(4): 1163-1170. Powelson, D. K., Simpson, J. R. and Gerba, C. P. (1990). "Virus Transport and Survival in Saturated and Unsaturated Flow through So il Columns." J Environ Qual 19(3): 396-401. Powelson, D. K., Simpson, J. R. and Gerba, C. P. (1991). "Effects of Organic-Matter on Virus Transport in Unsaturated Flow." Appl Environ Microbiol 57(8): 2192-2196. Robertson, L. J., Campbell, A. T. and Smith, H. V. (1992). "Survival of Cryptosporidium-Parvum Oocysts under Various Environmental Pressures." Appl Environ Microbiol 58(11): 3494-3500. Roll, B. M. and Fujioka, R. S. (199 7). "Sources of faecal in dicator bacteria in a br ackish, tropical stream and their impact on recreational water quality." Water Sci Technol 35(11-12): 179-186. Rose, J. B. and Zhou, X. T. (19 95). Phillippi Creek Water Quality Repo rt., Sarasota Bay National Estuary Program. Rossi, P. and Aragno, M. (1999). "Analysis of bacteriophage inactivation and its attenuation by adsorption onto colloidal particles by batch agitation techniques." Can J Microbiol 45(1): 9-17. Ryan, J. N., Elimelech, M., Ard, R. A., Harvey, R. W. and Johnson, P. R. (1999). "Bacteriophage PRD1 and silica colloid trans port and recovery in an iron oxide-coated sand aquifer." Environ Sci Technol 33(1): 63-73. Ryan, J. N., Harvey, R. W., Metge, D., Elimelech, M., Navigato, T. and Pieper, A. P. (2002). "Field and laboratory investigations of inac tivation of viruses (PRD1 and MS2) attached to iron oxide-coated qauartz sand." Environ Sci Technol 36(11): 2403-2413. Sakoda, A., Sakai, Y., Hayakawa, K. and Suzuki, M. (1997). "Adsorption of viruses in water environment onto solid surfaces." Water Sci Technol 35(7): 107-114.

PAGE 196

184 Scandura, J. E. and Sobsey, M. D. (1997). "Viral and bacterial contamination of groundwater from on-site sewage treatment systems." Water Sci Technol 35(11-12): 141-146. Schijven, J. F. and Hassanizadeh, S. M. (2000). "Removal of viruses by soil passage: Overview of modeling, processes, and paramete rs." Crit Rev Environ Sci Technol 30(1): 49-127. Sherer, B. M., Miner, J. R., Moore, J. A. and Buckh ouse, J. C. (1992). "Indicat or Bacterial Survival in Stream Sediments." J Environ Qual 21(4): 591-595. Sinton, L. W., Noonan, M. J., Finlay, R. K., Pang, L. and Close, M. E. (2000). "Transport and attenuation of bacteria and bacteriophages in an allu vial gravel aquifer." N Z J Mar Freshw Res 34(1): 175186. Sobsey, M. D. (1979). "Detection of Enteric Viruse s in Solid-Waste Landfill Leachates Response." American Journal of Public Health 69(4): 390-390. Sobsey, M. D., Shields, P. A., Hauchman, F. H., Hazard, R. L. and Caton, L. W. (1986). "Survival and Transport of Hepatitis a Virus in Soils, Groundwater and Wast e-Water." Water Sci Technol 18(10): 97-106. Solic, M. and Krstulovic, N. (1992) "Separate and Combined Effects of Solar-Radiation, Temperature, Salinity, and Ph on the Survival of Fecal -Coliforms in Seawat er." Mar Pollut Bull 24(8): 411-416. Thompson, S. S. and Yates, M. V. (1999). "Bacterioph age inactivation at the airwater-solid interface in dynamic batch systems." Appl Environ Microbiol 65(3): 1186-1190. Toranzos, G. A. (1991). "Current and possible alternate indicators of fecal contamin ation in tropical waters: a short review." Environmenta l Toxicology and Water Quality 6: 121-130. USEPA (1999). The Class V Underground Injection Control StudyVol. 21: Aquifer Recharge and Aquifer Storage and Recovery Wells. Washington, D. C., U.S. Environmental Protection Agency. USEPA (1999). Understanding the Safe Drinking Water Act. EPA 810-F-99-008 USEPA (2002). Proposed Ground Water Rule. EPA 815-F-00-003 USEPA (2002). Protecting Drinking Water Through Underground Injection Control; Drinking Water Pocket Guide #2. EPA 816-K-02-001 Wanninkhof, R., Ledwell, J. R. and Broecker, W. S. (1985). "Gas-Exchange Wind-Speed Relation Measured with Sulfur-Hexafluoride on a Lake." Science 227(4691): 1224-1226. Wilson, R. D. and Mackay, D. M. (1993). "The Use of Sulfur-Hexafluoride as a Conservative Tracer in Saturated Sandy Media." Ground Water 31(5): 719-724. Woessner, W. W., Ball, P. N., DeBorde, D. C. and Troy, T. L. (2001). "Viral transp ort in a sand and gravel aquifer under fiel d pumping conditions." Ground Water 39(6): 886-894. Yahya, M. T., Galsomies, L., Gerba, C. P. and Bales, R. C. (1993). "Survival of Bacteriophages Ms-2 and Prd-1 in Ground-Water." Water Sci Technol 27(3-4): 409-412. Yates, M. V. (1985). "Septic-Tank Density and Groundwater Contamination." Ground Water 23(5): 586591.

PAGE 197

185 Yates, M. V. and Gerba, C. P. (1 985). "Factors Controlling the Survival of Viruses in Groundwater." Water Sci Technol 17(4-5): 681-687. Yates, M. V., Gerba, C. P. and Kelley, L. M. (1985). "Virus Persistence in Groundwater." Appl Environ Microbiol 49(4): 778-781. Yates, M. V., Stetzenbach, L. D., Gerba, C. P. and Sinclair, N. A. (1990). "The Effect of Indigenous Bacteria on Virus Survival in GroundWater." J Environ Sci Health Part A-Environ Sci Eng Toxic Hazard Subst Control 25(1): 81-100. Yates, M. V. and Yates, S. R. (1988). "Virus Surv ival and Transport in Ground-Water." Water Sci Technol 20(11-12): 301-307.

PAGE 198

186 APPENDICES

PAGE 199

187 Appendix 1: Observed data plots and fitted model curves for TDS-temperature experiments A. Fecal coliform TDS-Temperature experimental data charts A. Fecal coliform ASW 200 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred. B. Fecal coliform ASW 200 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred. C. Fecal coliform ASW 200 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred.

PAGE 200

188 Appendix 1-A (continued) Fecal coliform ASW 500 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred. Fecal coliform ASW 500 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred. Fecal coliform ASW 500 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred.

PAGE 201

189 Appendix 1-A (continued) Fecal coliform ASW 1000 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred. Fecal coliform ASW 1000 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred. Fecal coliform ASW 1000 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred.

PAGE 202

190 Appendix 1-A (continued) Fecal coliform ASW 3000 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Fecal coliform ASW 3000 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Fecal coliform ASW 3000 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 203

191 Appendix 1-A (continued) Fecal coliform, PBS, 5o C, Sets 2, 3, & 4 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 2 Set 2 pred. Set 3 Set 3 pred. Set 4 Set 4 pred. Fecal coliform, PBS, 5o C, Sets 1 & 5 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 1 Set 1 pred. Set 5 Set 5 pred. Fecal coliform, PBS, 22o C, Sets 2, 3, & 4 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 2 Set 2 pred. Set 3 Set 3 pred. Set 4 Set 4 pred.

PAGE 204

192 Appendix 1-A (continued) Fecal coliform, PBS, 22o C, Set 5 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 5 Set 5 pred. Fecal coliform, PBS, 30o C, Sets 2, 3, & 4 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 2 Set 2 pred. Set 3 Set 3 pred. Set 4 Set 4 pred. Fecal coliform, PBS, 30o C, Set 5 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 5 Set 5 pred.

PAGE 205

193 B. Enterococci TDS-temperatur e experimental data charts Enterococci ASW 200 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 05101520253035DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred. Enterococci ASW 200 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred. Enterococci ASW 200 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred.

PAGE 206

194 Appendix 1-B (continued) Enterococci ASW 200 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred. Enterococci ASW 500 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred. Enterococci ASW 500 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 0510152025DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred.

PAGE 207

195 Appendix 1-B (continued) Enterococci ASW 1000 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred. Enterococci ASW 1000 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 0510152025DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred. Enterococci ASW 1000 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 0510152025DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred.

PAGE 208

196 Appendix 1-B (continued) Enterococci ASW 3000 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Enterococci ASW 3000 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Enterococci ASW 3000 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 0510152025DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 209

197 Appendix 1-B (continued) Enterococci, PBS, 5o C, Sets 2, 3, & 4 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 2 Set 2 pred. Set 3 Set 3 pred. Set 4 Set 4 pred. Enterococci, PBS, 5o C, Sets 1 & 5 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 1 Set 1 pred. Set 2 Set 2 pred. Enterococci, PBS, 22o C, Sets 2, 3, & 4 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 2 Set 2 pred. Set 3 Set 3 pred. Set 4 Set 4 pred.

PAGE 210

198 Appendix 1-B (continued) Enterococci, PBS, 22o C, Set 5 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 5 Set 5 pred. Enterococci, PBS, 30o C, Sets 2, 3, & 4 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 2 Set 2 pred. Set 3 Set 3 pred. Set 4 Set 4 pred. Enterococci, PBS, 30o C, Set 5 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 5 Set 5 pred.

PAGE 211

199 C. F+RNA coliphage TDS-temperat ure experimental data charts F+ RNA coliphage ASW 200 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage ASW 200 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage ASW 200 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 212

200 Appendix 1-C (continued) F+ RNA coliphage ASW 500 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage ASW 500 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage ASW 500 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Trial 3 Trial 3 pred.

PAGE 213

201 Appendix 1-C (continued) F+ RNA coliphage ASW 1000 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage ASW 1000 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage ASW 1000 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 214

202 Appendix 1-C (continued) F+ RNA coliphage ASW 3000 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage ASW 3000 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage ASW 3000 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 215

203 Appendix 1-C (continued) F+ RNA coliphage, PBS, 5o C, Sets 1A, 1B & 2 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 1A Set 1A pred. Set 1B Set 1B pred. Set 2 Set 2 pred. F+ RNA coliphage, PBS, 22o C, Sets 1A, 1B & 2 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 1A Set 1A pred. Set 1B Set 1B pred. Set 2 Set 2 pred. F+ RNA coliphage, PBS, 30o C, Sets 1A, 1B & 2 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 1A Set 1A pred. Set 1B Set 1B pred. Set 2 Set 2 pred.

PAGE 216

204 D. DNA coliphage TDS-temperat ure experimental data charts DNA coliphage ASW 200 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage ASW 200 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage ASW 200 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 217

205 Appendix 1-D (continued) DNA coliphage ASW 500 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage ASW 500 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage ASW 500 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 218

206 Appendix 1-D (continued) DNA coliphage ASW 1000 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage ASW 1000 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage ASW 1000 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 219

207 Appendix 1-D (continued) DNA coliphage ASW 3000 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage ASW 3000 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage ASW 3000 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 220

208 Appendix 1-D (continued) DNA coliphage, PBS, 5o C, Sets 1 & 2 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 1 Set 1 pred. Set 2 Set 2 pred. DNA coliphage, PBS, 5o C, Sets 3 & 8 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 3 Set 3 pred. Set 8 Set 8 pred. DNA coliphage, PBS, 22o C, Sets 1 & 2 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 1 Set 1 pred. Set 2 Set 2 pred.

PAGE 221

209 Appendix 1-D (continued) DNA coliphage, PBS, 22o C, Sets 3 & 8 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 3 Set 3 pred. Set 8 Set 8 pred. DNA coliphage, PBS, 30o C, Sets 1 & 2 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 1 Set 1 pred. Set 2 Set 2 pred. DNA coliphage, PBS, 30o C, Sets 3 & 8 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 3 Set 3 pred. Set 8 Set 8 pred.

PAGE 222

210 E. PRD-1 TDS-temperature trial experimental data charts PRD-1 ASW 200 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1 ASW 200 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1 ASW 200 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 223

211 Appendix 1-E (continued) PRD-1 ASW 500 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1 ASW 500 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1 ASW 500 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 224

212 Appendix 1-E (continued) PRD-1 ASW 1000 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1 ASW 1000 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1 ASW 1000 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 225

213 Appendix 1-E (continued) PRD-1 ASW 3000 mg/l, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1 ASW 3000 mg/l, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1 ASW 3000 mg/l, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 226

214 Appendix 1-E (continued) PRD-1, PBS, 5o C, Sets 1 & 2 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 1 Set 1 pred. Set 2 Set 2 pred. PRD-1, PBS, 5o C, Sets 3 & 8 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 3 Set 3 pred. Set 8 Set 8 pred. PRD-1, PBS, 22o C, Sets 1 & 2 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 1 Set 1 pred. Set 2 Set 2 pred.

PAGE 227

215 Appendix 1-E (continued) PRD-1, PBS, 22o C, Sets 3 & 8 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 3 Set 3 pred. Set 8 Set 8 pred. PRD-1, PBS, 30o C, Sets 1 & 2 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 1 Set 1 pred. Set 2 Set 2 pred. PRD-1, PBS, 30o C, Sets 3 & 8 (ASW trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 3 Set 3 pred. Set 8 Set 8 pred.

PAGE 228

216 Appendix 2: Observed data plots and fitte d model curves for aquifer and reservoir water sample studies with indicator bacteria populations A. Fecal coliform Bradenton site na tural water experimental data charts Fecal coliform, Avon Park aquifer, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Fecal coliform, Avon Park aquifer, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Fecal coliform, Avon Park aquifer, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 229

217 Appendix 2-A (continued) Fecal coliform, Avon Park aquifer, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Fecal coliform, Avon Park aquifer, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Fecal coliform, Avon Park aquifer, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 230

218 Appendix 2-A (continued) Fecal coliform, Bill Evers reservoir, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Fecal coliform, Bill Evers reservoir, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Fecal coliform, Bill Evers reservoir, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 231

219 Appendix 2-A (continued) Fecal coliform, Bill Evers reservoir, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Fecal coliform, Bill Evers reservoir, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Fecal coliform, Bill Evers reservoir, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 232

220 Appendix 2-A (continued) Fecal coliform, PBS, 5o C, Sets 6 & 8 (Bradenton water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 6 Set 6 pred. Set 8 Set 8 pred. Fecal coliform, PBS, 22o C, Sets 6 & 8 (Bradenton water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 6 Set 6 pred. Set 8 Set 8 pred. Fecal coliform, PBS, 30o C, Sets 6 & 8 (Bradenton water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 6 Set 6 pred. Set 8 Set 8 pred.

PAGE 233

221 B. Fecal coliform Bradenton site raw water plots and fitted first-order regression models Bill Evers reservoir, raw, combined fecal coliformy = -0.0521x R2 = 0.6043 y = -0.0521x R2 = 0.6043 y = -0.4195x R2 = 0.9247 y = -1.0115x R2 = 0.7497-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 Avon Park aquifer, raw, combined fecal coliformy = -0.021x R2 = 0.8093 y = -0.021x R2 = 0.8093 y = -0.101x R2 = 0.8706 y = -0.101x R2 = 0.8706 y = -0.166x R2 = 0.8078 y = -0.166x R2 = 0.8078-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 () 5 C 22 C 30 C 5 C model 22 C model 30 C model

PAGE 234

222 C. Fecal coliform West Palm Beach site natural water experimental data charts Fecal coliform, Lake Lytal aquifer, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Fecal coliform, Lake Lytal aquifer, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Fecal coliform, Lake Lytal aquifer, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 235

223 Appendix 2-C (continued) Fecal coliform, Lake Lytal aquifer, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred Trial 2 Trial 2 pred. Fecal coliform, Lake Lytal aquifer, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 2 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Fecal coliform, Lake Lytal aquifer, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 2 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 236

224 Appendix 2-C (continued) Fecal coliform, Clear Lake reservoir, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Fecal coliform, Clear Lake reservoir, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Fecal coliform, Clear Lake reservoir, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 237

225 Appendix 2-C (continued) Fecal coliform, Clear Lake reservoir, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Fecal coliform, Clear Lake reservoir, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 2 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Fecal coliform, Clear Lake reservoir, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 2 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 238

226 Appendix 2-C (continued) Fecal coliform, PBS, 5o C, Sets 7 & 9 (West Palm Bch. water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 7 Set 7 pred. Set 9 Set 9 pred. Fecal coliform, PBS, 22o C, Sets 7 & 9 (West Palm Bch. water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 7 Set 7 pred. Set 9 Set 9 pred. Fecal coliform, PBS, 30o C, Sets 7 & 9 (West Palm Bch. water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 7 Set 7 pred. Set 9 Set 9 pred.

PAGE 239

227 D. Fecal coliform West Palm Beach site ra w water plots and fitted first-order regression models Lake Lytal Park aquifer, raw, combined fecal coliformy = -0.1409x R2 = 0.8764 y = -0.065x R2 = 0.7784 y = -0.065x R2 = 0.7784 y = -0.149x R2 = 0.8056-4 -3 -2 -1 0 1 05101520253035time (d)Log N/N0 Clear Lake reservoir, raw, combined fecal coliformy = -0.0655x R2 = 0.6944 y = -0.0655x R2 = 0.6944 y = -0.1727x R2 = 0.5187 y = -0.296x R2 = 0.6327-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 () 5 C 22 C 30 C 5 C model 22 C model 30 C model

PAGE 240

228 E. Enterococci Bradenton site natu ral water experimental data charts Enterococci, Avon Park aquifer, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Enterococci, Avon Park aquifer, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Enterococci, Avon Park aquifer, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 241

229 Appendix 2-E (continued) Enterococci, Avon Park aquifer, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Enterococci, Avon Park aquifer, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Enterococci, Avon Park aquifer, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 242

230 Appendix 2-E (continued) Enterococci, Bill Evers reservoir, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Enterococci, Bill Evers reservoir, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Enterococci, Bill Evers reservoir, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 243

231 Appendix 2-E (continued) Enterococci, Bill Evers reservoir, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Enterococci, Bill Evers reservoir, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Enterococci, Bill Evers reservoir, pateurized, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 244

232 Appendix 2-E (continued) Enterococci, PBS, 5o C, Sets 6 & 8 (Bradenton water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 6 Set 6 pred. Set 8 Set 8 pred. Enterococci, PBS, 22o C, Sets 6 & 8 (Bradenton water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 6 Set 6 pred. Set 8 Set 8 pred. Enterococci, PBS, 30o C, Sets 6 & 8 (Bradenton water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 6 Set 6 pred. Set 8 Set 8 pred.

PAGE 245

233 F. Enterococci Bradenton site raw water plots and fitted first-order regression models Avon Park aquifer, raw, combined enterococciy = -0.0601x R2 = -0.3077 y = -0.1622x R2 = 0.5434 y = -0.2525x R2 = 0.3411-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 Bill Evers reservoir, raw, combined enterococciy = -0.0488x R2 = 0.6694 y = -0.0488x R2 = 0.6694 y = -0.3769x R2 = 0.7681 y = -0.7742x R2 = 0.5528-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 () 5 C 22 C 30 C 5 C model 22 C model 30 C model

PAGE 246

234 G. Enterococci West Palm Beach natural water experimental data charts Enterococci, Lake Lytal aquifer, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Enterococci, Lake Lytal aquifer, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Enterococci, Lake Lytal aquifer, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 247

235 Appendix 2-G (continued) Enterococci, Lake Lytal aquifer, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Enterococci, Lake Lytal aquifer, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Enterococci, Lake Lytal aquifer, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 248

236 Appendix 2-G (continued) Enterococci, Clear Lake reservoir, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Enterococci, Clear Lake reservoir, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Enterococci, Clear Lake reservoir, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 249

237 Appendix 2-G (continued) Enterococci, Clear Lake reservoir, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Enterococci, Clear Lake reservoir, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. Enterococci, Clear Lake reservoir, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 250

238 Appendix 2-G (continued) Enterococci, PBS, 5o C, Sets 7 & 9 (West Palm Bch. water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 7 Set 7 pred. Set 9 Set 9 pred. Enterococci, PBS, 22o C, Sets 7 & 9 (West Palm Bch. water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 7 Set 7 pred. Set 9 Set 9 pred. Enterococci, PBS, 30o C, Sets 7 & 9 (West Palm Bch. water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 7 Set 7 pred. Set 9 Set 9 pred.

PAGE 251

239 H. Enterococci West Palm Beach site raw water plots and fitted first-order regression models Lake Lytal Park aquifer, raw, combined enterococciy = -0.0105x R2 = 0.6072 y = -0.0105x R2 = 0.6072 y = -0.0622x R2 = 0.9206 y = -0.0622x R2 = 0.9206 y = -0.1321x R2 = 0.9229 y = -0.1321x R2 = 0.9229-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 Clear Lake reservoir, raw, combined enterococciy = -0.0542x R2 = 0.8286 y = -0.0542x R2 = 0.8286 y = -0.2649x R2 = 0.8771 y = -0.5008x R2 = 0.9651 y = -0.5008x R2 = 0.9651-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 () 5 C 22 C 30 C 5 C model 22 C model 30 C model

PAGE 252

240 Appendix 3: Observed data plots and fitte d model curves for aquifer and reservoir water sample studies with bacteriophage A. F+ RNA coliphage Bradenton site natural water experimental data charts F+ RNA coliphage, Avon Park aquifer, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage, Avon Park aquifer, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage, Avon Park aquifer, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 253

241 Appendix 3-A (continued) F+ RNA coliphage, Avon Park aquifer, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage, Avon Park aquifer, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage, Avon Park aquifer, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 254

242 Appendix 3-A (continued) F+ RNA coliphage, Bill Evers reservoir, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage, Bill Evers reservoir, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage, Bill Evers reservoir, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 255

243 Appendix 3-A (continued) F+ RNA coliphage, Bill Evers reservoir, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage, Bill Evers reservoir, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage, Bill Evers reservoir, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 256

244 B. F+ RNA coliphage Bradenton site raw wa ter plots and fitted first-order regression models Avon Park aquifer, raw, combined F+ RNA coliphagey = -0.271x R2 = 0.4032 y = -0.5094x R2 = 0.5498 y = -1.5936x R2 = 0.9364-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 Bill Evers reservoir, raw, combined F+ RNA coliphagey = -0.0481x R2 = 0.6208 y = -0.2491x R2 = 0.9561 y = -0.633x R2 = 0.9524-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 () 5 C 22 C 30 C 5 C model 22 C model 30 C model

PAGE 257

245 C. F+ RNA coliphage West Palm Beach na tural water and PBS control experimental data charts F+ RNA coliphage, Lake Lytal aquifer, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage, Lake Lytal aquifer, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage, Lake Lytal aquifer, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 258

246 Appendix 3-C (continued) F+ RNA coliphage, Lake Lytal aquifer, pateurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage, Lake Lytal aquifer, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage, Lake Lytal aquifer, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 259

247 Appendix 3-C (continued) F+ RNA coliphage, Clear Lake reservoir, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage, Clear Lake reservoir, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage, Clear Lake reservoir, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trail 2 pred.

PAGE 260

248 Appendix 3-C (continued) F+ RNA coliphage, Clear Lake reservoir, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage, Clear Lake reservoir, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. F+ RNA coliphage, Clear Lake reservoir, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 261

249 Appendix 3-C (continued) F+ RNA coliphage, PBS, 5o C, Sets 3 & 4 (natural water trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 3 Set 3 pred. Set 4 Set 4 pred. F+ RNA coliphage, PBS, 22o C, Sets 3 & 4 (natural water trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 3 Set 3 pred. Set 4 Set 4 pred. F+ RNA coliphage, PBS, 30o C, Sets 3 & 4 (natural water trials) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 3 Set 3 pred. Set 4 Set 4 pred.

PAGE 262

250 D. F+ RNA coliphage West Palm Beach si te raw water plots and fitted first-order regression models Lake Lytal Park aquifer, raw, combined F+ RNA coliphagey = -0.0644x R2 = 0.8955 y = -0.4478x R2 = 0.9869 y = -2.4272x R2 = 0.9976-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 Clear Lake reservoir, raw, combined F+ RNA coliphagey = -0.0909x R2 = 0.89 y = -0.9379x R2 = 0.9562 y = -1.9991x R2 = 0.9341-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 () 5 C 22 C 30 C 5 C model 22 C model 30 C model

PAGE 263

251 E. DNA coliphage Bradenton site natu ral water experimental data charts DNA coliphage, Avon Park aquifer, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage, Avon Park aquifer, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage, Avon Park aquifer, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 264

252 Appendix 3-E (continued) DNA coliphage, Avon Park aquifer, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage, Avon Park aquifer, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage, Avon Park aquifer, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 265

253 Appendix 3-E (continued) DNA coliphage, Bill Evers reservoir, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage, Bill Evers reservoir, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage, Bill Evers reservoir, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 266

254 Appendix 3-E (continued) DNA coliphage, Bill Evers reservoir, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage, Bill Evers reservoir, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage, Bill Evers reservoir, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 267

255 Appendix 3-E (continued) DNA coliphage, PBS, 5o C, Sets 4 & 6 (Bradenton water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 4 Set 4 pred. Set 6 Set 6 pred. DNA coliphage, PBS, 22o C, Sets 4 & 6 (Bradenton water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 4 Set 4 pred. Set 6 Set 6 pred. DNA coliphage, PBS, 30o C, Sets 4 & 6 (Bradenton water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 4 Set 4 pred. Set 6 Set 6 pred.

PAGE 268

256 F. DNA coliphage Bradenton site raw water plots and fitted first-order regression models Avon Park aquifer, raw, combined DNA coliphagey = -0.0362x R2 = 0.6337 y = -0.0644x R2 = 0.8206 y = -0.1321x R2 = 0.8073-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 Bill Evers reservoir, raw, combined DNA coliphagey = -0.0371x R2 = 0.5615 y = -0.1174x R2 = 0.6397 y = -0.1669x R2 = 0.4183-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 () 5 C 22 C 30 C 5 C model 22 C model 30 C model

PAGE 269

257 G. DNA coliphage West Palm Beach site natural water experimental data charts DNA coliphage, Lake Lytal aquifer, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trail 1 pred. Trial 2 Trial 2 pred. DNA coliphage, Lake Lytal aquifer, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage, Lake Lytal aquifer, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 270

258 Appendix 3-G (continued) DNA coliphage, Lake Lytal aquifer, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage, Lake Lytal aquifer, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage, Lake Lytal aquifer, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 271

259 Appendix 3-G (continued) DNA coliphage, Clear Lake reservoir, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage, Clear Lake reservoir, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage, Clear Lake reservoir, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 272

260 Appendix 3-G (continued) DNA coliphage, Clear Lake reservoir, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage, Clear Lake reservoir, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. DNA coliphage, Clear Lake reservoir, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 273

261 Appendix 3-G (continued) DNA coliphage, PBS, 5o C, Sets 5 & 7 (West Palm Bch. water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 5 Set 5 pred. Set 7 Set 7 pred. DNA coliphage, PBS, 22o C, Sets 5 & 7 (West Palm Bch. water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 5 Set 5 pred. Set 7 Set 7 pred. DNA coliphage, PBS, 30o C, Sets 5 & 7 (West Palm Bch. water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 5 Set 5 pred. Set 7 Set 7 pred.

PAGE 274

262 H. DNA coliphage West Palm Beach site ra w water plots and fitted first-order regression models Lake Lytal Park aquifer, raw, combined DNA coliphagey = -0.0349x R2 = 0.9038 y = -0.0724x R2 = 0.91 y = -0.1481x R2 = 0.8747-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 Clear Lake reservoir, raw, combined DNA coliphagey = -0.0167x R2 = 0.6665 y = -0.0917x R2 = 0.9021 y = -0.1596x R2 = 0.7391-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 () 5 C 22 C 30 C 5 C model 22 C model 30 C model

PAGE 275

263 I. PRD-1 Bradenton site natural water experimental data charts PRD-1, Avon Park aquifer, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1, Avon Park aquifer, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1, Avon Park aquifer, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 276

264 Appendix 3-I (continued) PRD-1, Avon Park aquifer, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1, Avon Park aquifer, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1, Avon Park aquifer, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 277

265 Appendix 3-I (continued) PRD-1, Bill Evers reservoir, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1, Bill Evers reservoir, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1, Bill Evers reservoir, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 278

266 Appendix 3-I (continued) PRD-1, Bill Evers reservoir, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1, Bill Evers reservoir, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1, Bill Evers reservoir, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 279

267 Appendix 3-I (continued) PRD-1, PBS, 5o C, Sets 4 & 6 (Bradenton water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 4 Set 4 pred. Set 6 Set 6 pred. PRD-1, PBS, 22o C, Sets 4 & 6 (Bradenton water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 4 Set 4 pred. Set 6 Set 6 pred. PRD-1, PBS, 30o C, Sets 4 & 6 (Bradenton water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 4 Set 4 pred. Set 6 Set 6 pred.

PAGE 280

268 J. PRD-1 Bradenton site raw water plots and fitted first-order regression models Avon Park aquifer, raw, combined PRD-1y = -0.009x R2 = -0.0068 y = -0.0167x R2 = 0.075 y = -0.0148x R2 = 0.1173-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 Bill Evers reservior, raw, combined PRD-1y = -0.0469x R2 = 0.6 y = -0.0469x R2 = 0.6 y = -0.1033x R2 = 0.5334 y = -0.1033x R2 = 0.5334 y = -0.1525x R2 = 0.7541 y = -0.1525x R2 = 0.7541-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 () 5 C 22 C 30 C 5 C model 22 C model 30 C model

PAGE 281

269 K. PRD-1 West Palm Beach natura l water experimental data charts PRD-1, Lake Lytal aquifer, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1, Lake Lytal aquifer, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1, Lake Lytal aquifer, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 282

270 Appendix 3-K (continued) PRD-1, Lake Lytal aquifer, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1, Lake Lytal aquifer, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1, Lake Lytal aquifer, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 283

271 Appendix 3-K (continued) PRD-1, Clear Lake reservoir, raw, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1, Clear Lake reservoir, raw, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1, Clear Lake reservoir, raw, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 284

272 Appendix 3-K (continued) PRD-1, Clear Lake reservoir, pasteurized, 5o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1, Clear Lake reservoir, pasteurized, 22o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred. PRD-1, Clear Lake reservoir, pasteurized, 30o C -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Trial 1 Trial 1 pred. Trial 2 Trial 2 pred.

PAGE 285

273 Appendix 3-K (continued) PRD-1, PBS, 5o C, Sets 5 & 7 (West Palm Bch. water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 5 Set 5 pred. Set 7 Set 7 pred. PRD-1, PBS, 22o C, Sets 5 & 7 (West Palm Bch. water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 5 Set 5 pred. Set 7 Set 7 pred. PRD-1, PBS, 30o C, Sets 5 & 7 (West Palm Bch. water) -6 -5 -4 -3 -2 -1 0 1 051015202530DaysLog N/N0 Set 5 Set 5 pred. Set 7 Set 7 pred.

PAGE 286

274 L. PRD-1 West Palm Beach site raw water pl ots and fitted first-or der regression models Lake Lytal Park aquifer, raw, combined PRD-1y = -0.0153x R2 = 0.04 y = -0.0269x R2 = 0.129 y = -0.0447x R2 = 0.5198-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 Clear Lake reservoir, raw, combined PRD-1y = -0.034x R2 = 0.7408 y = -0.0837x R2 = 0.9155 y = -0.1148x R2 = 0.7063-4 -3 -2 -1 0 1 051015202530time (d)Log N/N0 () 5 C 22 C 30 C 5 C model 22 C model 30 C model

PAGE 287

275 Appendix 4: Results of statistical comp arisons on days for predicted 2-log10 (99%) decline A. Fecal coliform Temperature-TDS trials Fecal coliform, 2-log ANOVA, all TDS concentrations, unbalanced design Factor Type Levels Values Temp-bac fixed 3 5 22 30 TDS-bact fixed 4 200 500 1000 3000 Analysis of Variance for FC days, using Adjusted SS for Tests Source DF Seq SS Adj SS Adj MS F P Temp-bac 2 40786 39868 19934 9.33 0.001 TDS-bact 3 4588 4588 1529 0.72 0.554 Temp-bac*TDS-bact 6 3913 3913 652 0.31 0.927 Error 21 44887 44887 2137 Total 32 94174 Unusual Observations for FC days Obs FC days Fit StDev Fit Residual St Resid 19 200.000 121.667 26.693 78.333 2.08R 28 19.000 109.500 32.692 -90.500 -2.77R 29 200.000 109.500 32.692 90.500 2.77R R denotes an observation with a large standardized residual.

PAGE 288

276 Appendix 4-A (continued) Regression analysis of fecal coliform inactivation periods at 4 TDS concentrations The regression equation is FC days = 106 3.35 Temp-bact + 0.00688 TDS-bact Predictor Coef StDev T P Constant 106.19 16.86 6.30 0.000 Temp-bac -3.3500 0.6980 -4.80 0.000 TDS-bact 0.006885 0.007388 0.93 0.359 S = 41.80 R-Sq = 44.3% R-Sq(adj) = 40.6% Analysis of Variance Source DF SS MS F P Regression 2 41760 20880 11.95 0.000 Residual Error 30 52414 1747 Total 32 94174 Source DF Seq SS Temp-bac 1 40243 TDS-bact 1 1517 Unusual Observations Obs Temp-bac FC days Fit StDev Fit Residual St Resid 19 5.0 200.00 96.32 12.18 103.68 2.59R 22 22.0 126.00 39.37 7.57 86.63 2.11R 28 5.0 19.00 110.09 19.10 -91.09 -2.45R 29 5.0 200.00 110.09 19.10 89.91 2.42R R denotes an observation with a large standardized residual Fecal coliform, 2-way ANOVA, TDS of 200, 500, & 1000 mg/L only Analysis of Variance for FC days Source DF SS MS F P TDS-bact 2 4173 2087 1.44 0.262 Temp-bac 2 32712 16356 11.31 0.001 Interaction 4 3688 922 0.64 0.642 Error 18 26036 1446 Total 26 66610 Individual 95% CI TDS-bact Mean -------+---------+---------+---------+---200 34.6 (------------*-------------) 500 44.4 (------------*-------------) 1000 64.4 (------------*-------------) -------+---------+---------+---------+---20.0 40.0 60.0 80.0 Individual 95% CI Temp-bac Mean ----+---------+---------+---------+------5 96 (------*-------) 22 34 (-------*------) 30 14 (-------*-------) ----+---------+---------+---------+------0 35 70 105

PAGE 289

277 B. Enterococci Temperature-TDS trials Enterococci, 2-log ANOVA, all TDS concentrations, unbalanced design Factor Type Levels Values Temp-bac fixed 3 5 22 30 TDS-bact fixed 4 200 500 1000 3000 Analysis of Variance for Ent Days, using Adjusted SS for Tests Source DF Seq SS Adj SS Adj MS F P Temp-bac 2 21873.2 22335.8 11167.9 21.71 0.000 TDS-bact 3 4280.7 4280.7 1426.9 2.77 0.067 Temp-bac*TDS-bact 6 1505.2 1505.2 250.9 0.49 0.810 Error 21 10805.0 10805.0 514.5 Total 32 38464.1 Unusual Observations for Ent Days Obs Ent Days Fit StDev Fit Residual St Resid 2 114.000 62.333 13.096 51.667 2.79R 10 110.000 69.000 13.096 41.000 2.21R R denotes an observation with a large standardized residual. Regression analysis results for enterococci at all TDS concentrations. The regression equation is Ent Days = 65.7 2.44 Temp-bact + 0.00825 TDS-bact Predictor Coef StDev T P Constant 65.662 8.987 7.31 0.000 Temp-bac -2.4428 0.3720 -6.57 0.000 TDS-bact 0.008252 0.003937 2.10 0.045 S = 22.27 R-Sq = 61.3% R-Sq(adj) = 58.7% Analysis of Variance Source DF SS MS F P Regression 2 23579 11789 23.76 0.000 Residual Error 30 14885 496 Total 32 38464 Source DF Seq SS Temp-bac 1 21399 TDS-bact 1 2180 Unusual Observations Obs Temp-bac Ent Days Fit StDev Fit Residual St Resid 2 5.0 114.00 55.10 7.23 58.90 2.80R 10 5.0 110.00 57.57 6.80 52.43 2.47R R denotes an observation with a large standardized residual

PAGE 290

278 Appendix 4-B (continued) Enterococci, 2-way ANOVA, TDS of 200, 500, & 1000 mg/L only Analysis of Variance for Ent days Source DF SS MS F P TDS-bact 2 1222 611 1.16 0.337 Temp-bac 2 15497 7748 14.67 0.000 Interaction 4 975 244 0.46 0.763 Error 18 9508 528 Total 26 27201 Individual 95% CI TDS mg/L Mean -+---------+---------+---------+---------+ 200 22.4 (-------------*------------) 500 31.6 (------------*-------------) 1000 15.1 (-------------*------------) -+---------+---------+---------+---------+ 0.0 12.0 24.0 36.0 48.0 Individual 95% CI Temp C Mean ------+---------+---------+---------+----5 56.6 (------*-----) 22 10.6 (-----*------) 30 2.0 (------*-----) ------+---------+---------+---------+----0.0 25.0 50.0 75.0

PAGE 291

279 C. F+ RNA coliphage Temperature-TDS trials F+ RNA coliphage, 2-way ANOVA, all TDS concentrations Analysis of Variance RNA days Source DF SS MS F P temp-pha 2 16313 8156 77.59 0.000 TDS-phag 3 4220 1407 13.38 0.000 Interaction 6 9766 1628 15.48 0.000 Error 12 1262 105 Total 23 31560 Individual 95% CI temp C Mean --------+---------+---------+---------+--5 77.5 (---*---) 22 42.4 (---*---) 30 13.8 (---*---) --------+---------+---------+---------+--20.0 40.0 60.0 80.0 Individual 95% CI TDS mg/L Mean -----+---------+---------+---------+-----200 31.8 (-----*-----) 500 45.3 (-----*-----) 1000 35.2 (-----*------) 3000 65.8 (-----*-----) -----+---------+---------+---------+-----30.0 45.0 60.0 75.0 Regression analysis results for F+ RNA coliphage at all TDS concentrations The regression equation is RNA days = 78.7 + 0.0110 TDS 2.47 temp Predictor Coef StDev T P Constant 78.65 11.49 6.84 0.000 TDS-phag 0.010981 0.004496 2.44 0.024 temp-pha -2.4743 0.4709 -5.25 0.000 S = 24.05 R-Sq = 61.5% R-Sq(adj) = 57.9% Analysis of Variance Source DF SS MS F P Regression 2 19416.3 9708.1 16.79 0.000 Residual Error 21 12143.7 578.3 Total 23 31560.0 Source DF Seq SS TDS-phag 1 3449.6 temp-pha 1 15966.7 Unusual Observations Obs TDS-phag RNA days Fit StDev Fit Residual St Resid 19 3000 155.00 99.22 11.61 55.78 2.65R R denotes an observation with a large standardized residual

PAGE 292

280 Appendix 4-C (continued) F+ RNA coliphage, 2-way ANOVA, TDS of 200, 500, & 1000 mg/L only Analysis of Variance for RNA days Source DF SS MS F P TDS phg 2 593 297 2.35 0.151 temp phg 2 4417 2208 17.50 0.001 Interaction 4 132 33 0.26 0.895 Error 9 1136 126 Total 17 6278 Individual 95% CI TDS mg/L Mean ---------+---------+---------+---------+-200 31.8 (----------*---------) 500 45.3 (---------*----------) 1000 35.2 (---------*----------) ---------+---------+---------+---------+-30.0 40.0 50.0 60.0 Individual 95% CI temp C Mean ------+---------+---------+---------+----5 54.2 (------*------) 22 41.7 (------*------) 30 16.5 (------*------) ------+---------+---------+---------+----15.0 30.0 45.0 60.0

PAGE 293

281 D. DNA coliphage Temperature-TDS trials DNA coliphage, 2-way ANOVA, all TDS concentrations Analysis of Variance for DNA days Source DF SS MS F P temp-pha 2 15859 7930 3.45 0.065 TDS-phag 3 10963 3654 1.59 0.243 Interaction 6 6509 1085 0.47 0.816 Error 12 27558 2297 Total 23 60889 Individual 95% CI temp C Mean --------+---------+---------+---------+--5 101 (----------*---------) 22 96 (---------*----------) 30 44 (----------*---------) --------+---------+---------+---------+--35 70 105 140 Individual 95% CI TDS mg/L Mean -------+---------+---------+---------+---200 95 (-----------*-----------) 500 63 (-----------*-----------) 1000 56 (-----------*-----------) 3000 107 (------------*-----------) -------+---------+---------+---------+---35 70 105 140 Regression analysis for DNA coliphage at all TDS concentrations. The regression equation is DNA days = 106 + 0.0104 TDS-phage 1.98 temp-phage Predictor Coef StDev T P Constant 105.75 22.75 4.65 0.000 TDS-phag 0.010372 0.008901 1.17 0.257 temp-pha -1.9793 0.9322 -2.12 0.046 S = 47.61 R-Sq = 21.8% R-Sq(adj) = 14.4% Analysis of Variance Source DF SS MS F P Regression 2 13295 6647 2.93 0.075 Residual Error 21 47595 2266 Total 23 60889 Source DF Seq SS TDS-phag 1 3077 temp-pha 1 10217 Unusual Observations Obs TDS-phag DNA days Fit StDev Fit Residual St Resid 3 200 225.00 64.28 13.33 160.72 3.52R R denotes an observation with a large standardized residual

PAGE 294

282 Appendix 4-D (continued) DNA coliphage, 2-way ANOVA, TDS of 200, 500, & 1000 mg/L only Analysis of Variance for DNA days Source DF SS MS F P TDS phg 2 5345 2673 1.20 0.344 temp phg 2 15509 7755 3.49 0.075 Interaction 4 3750 938 0.42 0.789 Error 9 19996 2222 Total 17 44601 Individual 95% CI TDS mg/L Mean -------+---------+---------+---------+---200 95 (-----------*------------) 500 63 (-----------*-----------) 1000 56 (-----------*-----------) -------+---------+---------+---------+---35 70 105 140 Individual 95% CI temp C Mean ----+---------+---------+---------+------5 89 (----------*----------) 22 95 (----------*----------) 30 30 (----------*---------) ----+---------+---------+---------+------0 40 80 120

PAGE 295

283 E. PRD-1 bacteriophage Temperature-TDS trials PRD-1, 2-way ANOVA, all TDS concentrations Analysis of Variance for PRD rate Source DF SS MS F P temp-pha 2 0.01393 0.00696 2.41 0.132 TDS-phag 3 0.00496 0.00165 0.57 0.644 Interaction 6 0.00591 0.00098 0.34 0.902 Error 12 0.03466 0.00289 Total 23 0.05945 Individual 95% CI temp C log/d --+---------+---------+---------+--------5 -0.016 (-----------*-----------) 22 -0.022 (-----------*-----------) 30 -0.070 (-----------*-----------) --+---------+---------+---------+---------0.105 -0.070 -0.035 0.000 Individual 95% CI TDS-phag Mean -+---------+---------+---------+---------+ 200 -0.032 (-------------*-------------) 500 -0.038 (------------*-------------) 1000 -0.057 (-------------*------------) 3000 -0.017 (-------------*-------------) -+---------+---------+---------+---------+ -0.105 -0.070 -0.035 0.000 0.035 Regression analysis for PRD-1 at all TDS concentrations The regression equation is PRD rate = 0.0092 0.00186 temp-phage +0.000007 TDS-phage Predictor Coef StDev T P Constant -0.00923 0.02307 -0.40 0.693 temp-pha -0.0018600 0.0009453 -1.97 0.062 TDS-phag 0.00000722 0.00000903 0.80 0.433 S = 0.04827 R-Sq = 17.7% R-Sq(adj) = 9.8% Analysis of Variance Source DF SS MS F P Regression 2 0.010515 0.005257 2.26 0.130 Residual Error 21 0.048937 0.002330 Total 23 0.059452 Source DF Seq SS temp-pha 1 0.009023 TDS-phag 1 0.001491 Unusual Observations Obs temp-pha PRD rate Fit StDev Fit Residual St Resid 18 30.0 -0.22900 -0.05781 0.01441 -0.17119 -3.72R R denotes an observation with a large standardized residual

PAGE 296

284 Appendix 4-E (continued) PRD-1, 2-way ANOVA, TDS of 200, 500, & 1000 mg/L only Analysis of Variance for PRD rate Source DF SS MS F P TDS phg 2 0.00219 0.00110 0.29 0.758 temp phg 2 0.01637 0.00819 2.14 0.174 Interaction 4 0.00322 0.00080 0.21 0.926 Error 9 0.03450 0.00383 Total 17 0.05629 Individual 95% CI TDS phg Mean ---+---------+---------+---------+-------200 -0.032 (---------------*---------------) 500 -0.038 (---------------*----------------) 1000 -0.057 (----------------*---------------) ---+---------+---------+---------+--------0.105 -0.070 -0.035 0.000 Individual 95% CI temp phg Mean ---------+---------+---------+---------+-5 -0.019 (----------*-----------) 22 -0.023 (----------*-----------) 30 -0.085 (----------*----------) ---------+---------+---------+---------+--0.100 -0.050 -0.000 0.050

PAGE 297

285 F. Fecal coliform natural water trials Fecal coliform; 2-way ANOVA, all natural water Analysis of Variance for 2-log days Source DF SS MS F P temp 2 12700 6350 5.87 0.006 treated 1 26274 26274 24.29 0.000 Interaction 2 815 408 0.38 0.688 Error 42 45430 1082 Total 47 85218 Individual 95% CI temp C Mean ------+---------+---------+---------+----5 64.7 (-------*--------) 22 46.1 (-------*-------) 30 24.9 (-------*--------) ------+---------+---------+---------+----20.0 40.0 60.0 80.0 Individual 95% CI treated Mean ------+---------+---------+---------+----no 21.8 (------*------) yes 68.6 (-----*------) ------+---------+---------+---------+----20.0 40.0 60.0 80.0 Fecal coliform; 3-way ANOVA, all natural water Factor Type Levels Values wate type fixed 2 ground surface treated fixed 2 no yes temp fixed 3 5 22 30 Analysis of Variance for 2-log days Source DF SS MS F P type 1 4622 4622 4.62 0.038 treated 1 26274 26274 26.25 0.000 temp 2 12700 6350 6.34 0.004 type*treated 1 527 527 0.53 0.473 type*temp 2 3195 1598 1.60 0.217 treated*temp 2 815 408 0.41 0.669 type*treated*temp 2 1051 525 0.52 0.596 Error 36 36036 1001 Total 47 85218

PAGE 298

286 Appendix 4-F (continued) Fecal coliform; 2-way ANOVA, raw surface and ground Analysis of Variance for days-raw Source DF SS MS F P temp-raw 2 4688 2344 6.28 0.009 type-raw 1 1014 1014 2.72 0.117 Interaction 2 315 158 0.42 0.662 Error 18 6721 373 Total 23 12737 Individual 95% CI temp-raw Mean -----+---------+---------+---------+-----5 40.5 (---------*---------) 22 18.1 (--------*---------) 30 6.9 (---------*--------) -----+---------+---------+---------+-----0.0 15.0 30.0 45.0 Individual 95% CI type-raw Mean -------+---------+---------+---------+---ground 28.3 (----------*-----------) surface 15.3 (----------*-----------) -------+---------+---------+---------+---10.0 20.0 30.0 40.0 Fecal coliform; 2-way ANOVA, raw surface and ground, higher temperatures only Analysis of Variance for days Source DF SS MS F P temp-hig 1 506.2 506.2 11.63 0.005 type-hi 1 961.0 961.0 22.07 0.001 Interaction 1 240.3 240.3 5.52 0.037 Error 12 522.5 43.5 Total 15 2230.0 Individual 95% CI temp C Mean --------+---------+---------+---------+--22 18.1 (-------*--------) 30 6.9 (-------*--------) --------+---------+---------+---------+--6.0 12.0 18.0 24.0 Individual 95% CI type Mean -+---------+---------+---------+---------+ ground 20.3 (------*------) surface 4.7 (------*------) -+---------+---------+---------+---------+ 0.0 7.0 14.0 21.0 28.0

PAGE 299

287 Appendix 4-F (continued) Fecal coliform; 2-way ANOVA, surface water Analysis of Variance for days-sw Source DF SS MS F P temp-sw 2 8958 4479 8.82 0.002 treated1 9680 9680 19.06 0.000 Interaction 2 450 225 0.44 0.649 Error 18 9141 508 Total 23 28230 Individual 95% CI temp C Mean ---------+---------+---------+---------+-5 62.5 (-------*--------) 22 25.0 (--------*-------) 30 18.8 (-------*--------) ---------+---------+---------+---------+-20.0 40.0 60.0 80.0 Individual 95% CI treatedMean ----------+---------+---------+---------+no 15.3 (------*------) yes 55.5 (------*------) ----------+---------+---------+---------+20.0 40.0 60.0 80.0 Fecal coliform; 1-way ANOVA, raw surface water Analysis of Variance for days sw Source DF SS MS F P temp sw 2 2713 1356 7.54 0.012 Error 9 1618 180 Total 11 4331 Individual 95% CIs For Mean Based on Pooled StDev Temp C N Mean StDev -------+---------+---------+--------5 4 36.50 23.01 (------*-------) 22 4 6.50 2.52 (------*-------) 30 4 3.00 1.83 (-------*------) -------+---------+---------+--------Pooled StDev = 13.41 0 20 40

PAGE 300

288 Appendix 4-F (continued) Fecal coliform; 2-way ANOVA, groundwater Analysis of Variance for days gw Source DF SS MS F P temp gw 2 9588 4794 1.68 0.215 treated 1 22022 22022 7.70 0.012 Interaction 2 1287 643 0.22 0.801 Error 18 51491 2861 Total 23 84388 Individual 95% CI temp C Mean ---+---------+---------+---------+-------5 78 (----------*-----------) 22 67 (----------*-----------) 30 31 (----------*----------) ---+---------+---------+---------+-------0 35 70 105 Individual 95% CI treated Mean --+---------+---------+---------+--------no 28 (---------*----------) yes 89 (----------*---------) --+---------+---------+---------+--------0 30 60 90 Fecal coliform; 1-way ANOVA, raw groundwater Analysis of Variance for days gw Source DF SS MS F P temp gw 2 2290 1145 2.02 0.189 Error 9 5102 567 Total 11 7393 Individual 95% CIs For Mean Based on Pooled StDev Temp C N Mean StDev -------+---------+---------+--------5 4 44.50 39.20 (----------*----------) 22 4 29.75 12.69 (----------*----------) 30 4 10.75 1.89 (---------*----------) -------+---------+---------+--------Pooled StDev = 23.81 0 25 50

PAGE 301

289 Appendix 4-F (continued) Fecal coliform; 2-way ANOVA for PBS control sets Analysis of Variance for FC days Source DF SS MS F P FC set 3 9163 3054 1.82 0.243 FC temp 2 5461 2730 1.63 0.272 Error 6 10043 1674 Total 11 24667 Individual 95% CI FC set Mean -----+---------+---------+---------+-----6 39 (-----------*----------) 7 102 (----------*-----------) 8 41 (----------*-----------) 9 36 (----------*-----------) -----+---------+---------+---------+-----0 50 100 150 Individual 95% CI FC temp Mean ----+---------+---------+---------+------5 40 (-------------*--------------) 22 85 (-------------*--------------) 30 39 (-------------*--------------) ----+---------+---------+---------+------0 35 70 105 Fecal coliform; covariance analysis for PBS sets, temperature as covariant Factor Type Levels Values FC set fixed 4 6 7 8 9 Analysis of Variance for FC days, using Adjusted SS for Tests Source DF Seq SS Adj SS Adj MS F P FC temp 1 206 206 206 0.09 0.768 FC set 3 9163 9163 3054 1.40 0.321 Error 7 15298 15298 2185 Total 11 24667 Term Coef StDev T P Constant 47.04 28.06 1.68 0.138 FC temp 0.397 1.295 0.31 0.768 Unusual Observations for FC days Obs FC days Fit StDev Fit Residual St Resid 5 200.000 103.525 27.268 96.475 2.54R R denotes an observation with a large standardized residual.

PAGE 302

290 G. Enterococci natural water trials Enterococci; 2-way ANOVA, all natural water Analysis of Variance for 2-log days Source DF SS MS F P temp 2 69285 34643 25.76 0.000 treated 1 3623 3623 2.69 0.108 Interaction 2 2629 1314 0.98 0.385 Error 42 56483 1345 Total 47 132019 Individual 95% CI temp C Mean ----+---------+---------+---------+------5 93.4 (-----*-----) 22 20.2 (-----*-----) 30 7.0 (-----*------) ----+---------+---------+---------+------0.0 30.0 60.0 90.0 Individual 95% CI treated Mean -------+---------+---------+---------+---no 31.5 (-----------*------------) yes 48.9 (------------*-----------) -------+---------+---------+---------+---24.0 36.0 48.0 60.0 Enterococci; 3-way ANOVA, all natural water Factor Type Levels Values temp fixed 3 5 22 30 treated fixed 2 no yes water ty fixed 2 ground surface Analysis of Variance for 2-log days Source DF SS MS F P temp 2 69285 34643 33.34 0.000 treated 1 3623 3623 3.49 0.070 water ty 1 2338 2338 2.25 0.142 temp*treated 2 2629 1314 1.27 0.294 temp*water ty 2 853 426 0.41 0.666 treated*water ty 1 6557 6557 6.31 0.017 temp*treated*water ty 2 9332 4666 4.49 0.018 Error 36 37403 1039 Total 47 132019

PAGE 303

291 Appendix 4-G (continued) Enterococci; 2-way ANOVA, raw surface and groundwater Analysis of Variance for days-r Source DF SS MS F P temp-r 2 22672 11336 10.11 0.001 type-r 1 8363 8363 7.46 0.014 Interaction 2 7812 3906 3.48 0.053 Error 18 20188 1122 Total 23 59034 Individual 95% CI temp C Mean -------+---------+---------+---------+---5 74.7 (-------*-------) 22 13.6 (--------*-------) 30 6.1 (-------*-------) -------+---------+---------+---------+---0.0 30.0 60.0 90.0 Individual 95% CI type Mean ----+---------+---------+---------+------ground 50.2 (---------*---------) surface 12.8 (---------*----------) ----+---------+---------+---------+------0.0 20.0 40.0 60.0 Enterococci; 2-way ANOVA, raw surface and ground, higher temperatures only Analysis of Variance for days Source DF SS MS F P temp-hig 1 225.0 225.0 3.77 0.076 type hig 1 576.0 576.0 9.65 0.009 Interaction 1 110.2 110.2 1.85 0.199 Error 12 716.5 59.7 Total 15 1627.8 Individual 95% CI temp C Mean ----------+---------+---------+---------+22 13.6 (-----------*-----------) 30 6.1 (-----------*-----------) ----------+---------+---------+---------+5.0 10.0 15.0 20.0 Individual 95% CI type Mean ----+---------+---------+---------+------ground 15.9 (--------*---------) surface 3.9 (--------*---------) ----+---------+---------+---------+------0.0 6.0 12.0 18.0

PAGE 304

292 Appendix 4-G (continued) Enterococci; 2-way ANOVA, surface water Analysis of Variance for days-sw Source DF SS MS F P temp-sw 2 27513 13756 36.91 0.000 treated1 9963 9963 26.74 0.000 Interaction 2 10764 5382 14.44 0.000 Error 18 6708 373 Total 23 54948 Individual 95% CI temp C Mean ----+---------+---------+---------+------5 80.7 (----*-----) 22 14.4 (-----*----) 30 4.5 (-----*-----) ----+---------+---------+---------+------0.0 25.0 50.0 75.0 Individual 95% CI treatedMean ----------+---------+---------+---------+no 12.8 (------*------) yes 53.6 (------*-------) ----------+---------+---------+---------+16.0 32.0 48.0 64.0 Enterococci; 1-way ANOVA, raw surface water Analysis of Variance for days sw Source DF SS MS F P temp sw 2 1936.17 968.08 125.36 0.000 Error 9 69.50 7.72 Total 11 2005.67 Individual 95% CIs For Mean Based on Pooled StDev Temp C N Mean StDev -+---------+---------+---------+----5 4 30.750 4.193 (--*--) 22 4 5.000 1.826 (--*--) 30 4 2.750 1.500 (--*--) -+---------+---------+---------+----Pooled StDev = 2.779 0 10 20 30

PAGE 305

293 Appendix 4-G (continued) Enterococci; 2-way ANOVA, groundwater Analysis of Variance for days gw Source DF SS MS F P treated 1 216 216 0.13 0.726 temp gw 2 42625 21313 12.50 0.000 Interaction 2 1197 598 0.35 0.709 Error 18 30695 1705 Total 23 74733 Individual 95% CI treated Mean --------+---------+---------+---------+--no 50.2 (---------------*----------------) yes 44.2 (---------------*----------------) --------+---------+---------+---------+--30.0 45.0 60.0 75.0 Individual 95% CI temp C Mean ------+---------+---------+---------+----5 106 (-------*------) 22 26 (-------*------) 30 10 (------*-------) ------+---------+---------+---------+----0 40 80 120 Enterococci; 1-way ANOVA, raw groundwater Analysis of Variance for days gw Source DF SS MS F P temp gw 2 28547 14274 6.39 0.019 Error 9 20119 2235 Total 11 48666 Individual 95% CIs For Mean Based on Pooled StDev Temp C N Mean StDev --------+---------+---------+-------5 4 118.75 80.45 (--------*--------) 22 4 22.25 13.94 (--------*--------) 30 4 9.50 6.24 (--------*-------) --------+---------+---------+-------Pooled StDev = 47.28 0 60 120

PAGE 306

294 Appendix 4-G (continued) Enterococci; 2-way ANOVA for PBS control sets Analysis of Variance for Ent days Source DF SS MS F P Ent set 3 1443 481 2.02 0.213 Ent temp 2 7992 3996 16.74 0.004 Error 6 1432 239 Total 11 10867 Individual 95% CI Ent set Mean -----+---------+---------+---------+-----6 55.0 (----------*---------) 7 44.7 (----------*----------) 8 32.7 (----------*----------) 9 61.7 (----------*----------) -----+---------+---------+---------+-----20.0 40.0 60.0 80.0 Individual 95% CI Ent temp Mean -+---------+---------+---------+---------+ 5 81.5 (-------*------) 22 45.5 (------*-------) 30 18.5 (------*-------) -+---------+---------+---------+---------+ 0.0 25.0 50.0 75.0 100.0 Covariance Analysis for PBS control sets, temperature as covariant Factor Type Levels Values Ent set fixed 4 6 7 8 9 Analysis of Variance for Ent days, using Adjusted SS for Tests Source DF Seq SS Adj SS Adj MS F P Ent temp 1 7872.4 7872.4 7872.4 35.52 0.001 Ent set 3 1443.0 1443.0 481.0 2.17 0.180 Error 7 1551.6 1551.6 221.7 Total 11 10867.0 Term Coef StDev T P Constant 95.184 8.935 10.65 0.000 Ent temp -2.4571 0.4123 -5.96 0.001

PAGE 307

295 H. F+ RNA coliphage natural water trials F+ RNA coliphage; 2-way ANOVA, all natural water Analysis of Variance for 2-log days Source DF SS MS F P temp 2 14709 7354 24.49 0.000 treated 1 4 4 0.01 0.908 Interaction 2 8 4 0.01 0.986 Error 42 12610 300 Total 47 27331 Individual 95% CI temp Mean -----+---------+---------+---------+-----5 40.1 (-----*-----) 22 4.6 (-----*-----) 30 1.4 (-----*-----) -----+---------+---------+---------+-----0.0 15.0 30.0 45.0 Individual 95% CI treated Mean -+---------+---------+---------+---------+ no 15.1 (-----------------*-----------------) yes 15.7 (-----------------*-----------------) -+---------+---------+---------+---------+ 8.0 12.0 16.0 20.0 24.0 F+ RNA coliphage; 3-way ANOVA, all natural water Factor Type Levels Values temp fixed 3 5 22 30 water ty fixed 2 ground surface treated fixed 2 no yes Analysis of Variance for 2-log da Source DF SS MS F P temp 2 14708.6 7354.3 29.65 0.000 water ty 1 736.3 736.3 2.97 0.093 treated 1 4.1 4.1 0.02 0.899 temp*water ty 2 984.3 492.1 1.98 0.152 temp*treated 2 8.3 4.1 0.02 0.983 water ty*treated 1 616.3 616.3 2.48 0.124 temp*water ty*treated 2 1342.8 671.4 2.71 0.080 Error 36 8930.5 248.1 Total 47 27331.2

PAGE 308

296 Appendix 4-H (continued) F+ RNA coliphage; 2-way ANOVA, raw surface and groundwater Analysis of Variance for days-raw Source DF SS MS F P temp-raw 2 7548 3774 9.07 0.002 type-raw 1 1350 1350 3.24 0.088 Interaction 2 2311 1155 2.78 0.089 Error 18 7491 416 Total 23 18700 Individual 95% CI temp C Mean -------+---------+---------+---------+---5 40.1 (-------*-------) 22 3.7 (-------*------) 30 1.4 (-------*------) -------+---------+---------+---------+---0.0 20.0 40.0 60.0 Individual 95% CI type-raw Mean -----+---------+---------+---------+-----ground 7.6 (------------*-----------) surface 22.6 (------------*-----------) -----+---------+---------+---------+-----0.0 10.0 20.0 30.0 F+ RNA coliphage; 2-way ANOVA, raw surface and ground water, higher temperatures only Analysis of Variance for days raw Source DF SS MS F P temp raw 1 22.56 22.56 8.02 0.015 type raw 1 5.06 5.06 1.80 0.205 Interaction 1 0.56 0.56 0.20 0.663 Error 12 33.75 2.81 Total 15 61.94 Individual 95% CI temp C Mean ----------+---------+---------+---------+22 3.75 (----------*----------) 30 1.38 (---------*----------) ----------+---------+---------+---------+1.20 2.40 3.60 4.80 Individual 95% CI type raw Mean ---+---------+---------+---------+-------ground 2.00 (------------*------------) surface 3.12 (------------*------------) ---+---------+---------+---------+-------1.00 2.00 3.00 4.00

PAGE 309

297 Appendix 4-H (continued) F+ RNA coliphage; 2-way ANOVA, surface water Analysis of Variance for days-sw Source DF SS MS F P temp-sw 2 11650 5825 13.43 0.000 treated1 260 260 0.60 0.449 Interaction 2 740 370 0.85 0.442 Error 18 7806 434 Total 23 20457 Individual 95% CI temp C Mean -------+---------+---------+---------+---5 50.4 (-------*-------) 22 5.6 (-------*-------) 30 1.9 (-------*-------) -------+---------+---------+---------+---0.0 20.0 40.0 60.0 Individual 95% CI treatedMean ------+---------+---------+---------+----no 22.6 (---------------*---------------) yes 16.0 (---------------*---------------) ------+---------+---------+---------+----8.0 16.0 24.0 32.0 F+ RNA coliphage; 1-way ANOVA, raw surface water Analysis of Variance for days sw Source DF SS MS F P temp sw 2 9102 4551 6.19 0.020 Error 9 6613 735 Total 11 15715 Individual 95% CIs For Mean Based on Pooled StDev Level N Mean StDev ---------+---------+---------+------5 4 61.50 46.85 (--------*-------) 22 4 4.50 2.89 (-------*--------) 30 4 1.75 0.96 (--------*-------) ---------+---------+---------+------Pooled StDev = 27.11 0 35 70

PAGE 310

298 Appendix 4-H (continued) F+ RNA coliphage; 2-way ANOVA, groundwater Analysis of Variance for days gw Source DF SS MS F P temp gw 2 4042.6 2021.3 32.36 0.000 treated 1 360.4 360.4 5.77 0.027 Interaction 2 610.7 305.4 4.89 0.020 Error 18 1124.3 62.5 Total 23 6138.0 Individual 95% CI temp C Mean -----+---------+---------+---------+-----5 29.8 (-----*-----) 22 3.6 (-----*----) 30 1.0 (-----*-----) -----+---------+---------+---------+-----0.0 10.0 20.0 30.0 Individual 95% CI treated Mean -----+---------+---------+---------+-----no 7.6 (--------*---------) yes 15.3 (---------*--------) -----+---------+---------+---------+-----5.0 10.0 15.0 20.0 F+ RNA coliphage; 1-way ANOVA, raw groundwater Analysis of Variance for days gw Source DF SS MS F P temp gw 2 756.2 378.1 3.87 0.061 Error 9 878.8 97.6 Total 11 1634.9 Individual 95% CIs For Mean Based on Pooled StDev Temp C N Mean StDev ---------+---------+---------+------5 4 18.750 17.056 (---------*--------) 22 4 3.000 1.414 (--------*---------) 30 4 1.000 0.000 (--------*--------) ---------+---------+---------+------Pooled StDev = 9.881 0 12 24

PAGE 311

299 Appendix 4-H (continued) F+ RNA coliphage; 2-way ANOVA, PBS control sets Analysis of Variance for RNA days Source DF SS MS F P RNA set 1 1096 1096 3.81 0.190 RNA temp 2 1826 913 3.17 0.240 Error 2 575 288 Total 5 3497 Individual 95% CI RNA set Mean -------+---------+---------+---------+---3 49.7 (-------------*-------------) 4 22.6 (--------------*-------------) -------+---------+---------+---------+---0.0 30.0 60.0 90.0 Individual 95% CI RNA temp Mean -+---------+---------+---------+---------+ 5 56 (--------------*--------------) 22 38 (--------------*--------------) 30 14 (--------------*--------------) -+---------+---------+---------+---------+ -35 0 35 70 105 F+ RNA coliphage; covariance analysis for PBS sets, temperature as covariant Factor Type Levels Values RNA set fixed 2 3 4 Analysis of Variance for RNA days, using Adjusted SS for Tests Source DF Seq SS Adj SS Adj MS F P RNA temp 1 1670.1 1670.1 1670.1 6.85 0.079 RNA set 1 1096.2 1096.2 1096.2 4.50 0.124 Error 3 731.0 731.0 243.7 Total 5 3497.3 Term Coef StDev T P Constant 66.56 13.25 5.02 0.015 RNA temp -1.6005 0.6113 -2.62 0.079

PAGE 312

300 I. DNA coliphage natural water trials DNA coliphage; 2-way ANOVA, all natural water Analysis of Variance for 2-log days Source DF SS MS F P temp 2 52285 26143 9.68 0.000 treated 1 53734 53734 19.89 0.000 Interaction 2 4902 2451 0.91 0.411 Error 42 113466 2702 Total 47 224388 Individual 95% CI temp C Mean ------+---------+---------+---------+----5 123 (------*-------) 22 73 (-------*------) 30 43 (------*-------) ------+---------+---------+---------+----35 70 105 140 Individual 95% CI treated Mean --+---------+---------+---------+--------no 46 (------*-------) yes 113 (------*------) --+---------+---------+---------+--------30 60 90 120 DNA coliphage; 3-way ANOVA, all natural water Factor Type Levels Values water ty fixed 2 ground surface treated fixed 2 no yes temp fixed 3 5 22 30 Analysis of Variance for 2-log days Source DF SS MS F P water ty 1 7450 7450 2.94 0.095 treated 1 53734 53734 21.19 0.000 temp 2 52285 26143 10.31 0.000 water ty*treated 1 14560 14560 5.74 0.022 water ty*temp 2 110 55 0.02 0.979 treated*temp 2 4902 2451 0.97 0.390 water ty*treated*temp 2 36 18 0.01 0.993 Error 36 91310 2536 Total 47 224388

PAGE 313

301 Appendix 4-I (continued) DNA coliphage; 2-way ANOVA, raw surface and groundwater Analysis of Variance for days-raw Source DF SS MS F P type-raw 1 590 590 0.57 0.462 temp-raw 2 37027 18513 17.74 0.000 Interaction 2 134 67 0.06 0.938 Error 18 18789 1044 Total 23 56541 Individual 95% CI type Mean ---+---------+---------+---------+-------ground 51.1 (----------------*---------------) surface 41.2 (---------------*----------------) ---+---------+---------+---------+-------24.0 36.0 48.0 60.0 Individual 95% CI temp C Mean ----+---------+---------+---------+------5 101 (------*------) 22 27 (------*------) 30 11 (------*------) ----+---------+---------+---------+------0 35 70 105 DNA coliphage; 2-way ANOVA, raw surface and ground water, higher temperatures only Analysis of Variance for days hig Source DF SS MS F P temp hig 1 1056.3 1056.3 24.30 0.000 type hig 1 361.0 361.0 8.31 0.014 Interaction 1 132.3 132.3 3.04 0.107 Error 12 521.5 43.5 Total 15 2071.0 Individual 95% CI temp C Mean ---+---------+---------+---------+-------22 26.9 (------*-------) 30 10.6 (------*------) ---+---------+---------+---------+-------7.0 14.0 21.0 28.0 Individual 95% CI water type Mean ---+---------+---------+---------+-------ground 23.5 (---------*---------) surface 14.0 (---------*---------) ---+---------+---------+---------+-------10.0 15.0 20.0 25.0

PAGE 314

302 Appendix 4-I (continued) DNA coliphage; 2-way ANOVA, surface water Analysis of Variance for days-sw Source DF SS MS F P temp-sw 2 24831 12416 4.62 0.024 treated1 62118 62118 23.11 0.000 Interaction 2 2727 1364 0.51 0.610 Error 18 48378 2688 Total 23 138055 Individual 95% CI temp C Mean ------+---------+---------+---------+----5 135 (---------*--------) 22 84 (---------*---------) 30 57 (--------*---------) ------+---------+---------+---------+----40 80 120 160 Individual 95% CI treatedMean --------+---------+---------+---------+--no 41 (-------*-------) yes 143 (-------*-------) --------+---------+---------+---------+--40 80 120 160 DNA coliphage; 1-way ANOVA, raw surface water Analysis of Variance for days surface water, raw Source DF SS MS F P temp sw 2 17933 8967 9.31 0.006 Error 9 8664 963 Total 11 26598 Individual 95% CIs For Mean Based on Pooled StDev Temp C N Mean StDev ------+---------+---------+---------+ 5 4 95.50 53.15 (------*------) 22 4 19.25 6.95 (------*------) 30 4 8.75 3.86 (------*------) ------+---------+---------+---------+ Pooled StDev = 31.03 0 50 100 150

PAGE 315

303 Appendix 4-I (continued) DNA coliphage; 2-way ANOVA, groundwater Analysis of Variance for days gw Source DF SS MS F P temp gw 2 27564 13782 5.78 0.012 treated 1 6176 6176 2.59 0.125 Interaction 2 2211 1106 0.46 0.636 Error 18 42931 2385 Total 23 78883 Individual 95% CI temp C Mean ---+---------+---------+---------+-------5 111 (--------*--------) 22 62 (--------*--------) 30 28 (--------*--------) ---+---------+---------+---------+-------0 40 80 120 Individual 95% CI treated Mean --+---------+---------+---------+--------no 51 (----------*-----------) yes 83 (-----------*-----------) --+---------+---------+---------+--------25 50 75 100 DNA coliphage; 1-way ANOVA, raw groundwater Analysis of Variance for days gw Source DF SS MS F P temp gw 2 19228 9614 8.55 0.008 Error 9 10125 1125 Total 11 29353 Individual 95% CIs For Mean Based on Pooled StDev Temp C N Mean StDev ------+---------+---------+---------+ 5 4 106.25 57.13 (------*-------) 22 4 34.50 10.08 (-------*------) 30 4 12.50 3.00 (-------*------) ------+---------+---------+---------+ Pooled StDev = 33.54 0 50 100 150

PAGE 316

304 Appendix 4-I (continued) DNA coliphage; 2-way ANOVA, PBS control sets Analysis of Variance for DNA days Source DF SS MS F P DNA set 3 523 174 0.52 0.681 DNA temp 2 5028 2514 7.57 0.023 Error 6 1994 332 Total 11 7545 Individual 95% CI DNA set Mean ---------+---------+---------+---------+-4 43.0 (---------------*---------------) 5 43.3 (---------------*---------------) 6 27.7 (---------------*---------------) 7 42.3 (---------------*----------------) ---------+---------+---------+---------+-16.0 32.0 48.0 64.0 Individual 95% CI DNA temp Mean ----+---------+---------+---------+------5 63.0 (--------*--------) 22 41.3 (--------*-------) 30 13.0 (--------*--------) ----+---------+---------+---------+------0.0 25.0 50.0 75.0 DNA coliphage; covariance analysis for PBS sets, temperature as covariant Factor Type Levels Values DNA set fixed 4 4 5 6 7 Analysis of Variance for DNA days, using Adjusted SS for Tests Source DF Seq SS Adj SS Adj MS F P DNA temp 1 4644.6 4644.6 4644.6 13.68 0.008 DNA set 3 522.9 522.9 174.3 0.51 0.686 Error 7 2377.4 2377.4 339.6 Total 11 7544.9 Term Coef StDev T P Constant 74.94 11.06 6.78 0.000 DNA temp -1.8873 0.5103 -3.70 0.008

PAGE 317

305 J. PRD-1 natural water trials, inactivation rate statistics PRD-1; 2-way ANOVA, all natural water Analysis of Variance for rate Source DF SS MS F P temp 2 0.02579 0.01289 3.15 0.053 treated 1 0.00922 0.00922 2.25 0.141 Interaction 2 0.00065 0.00033 0.08 0.923 Error 42 0.17217 0.00410 Total 47 0.20783 Individual 95% CI temp C Mean ------+---------+---------+---------+----5 -0.018 (----------*----------) 22 -0.039 (----------*----------) 30 -0.074 (---------*----------) ------+---------+---------+---------+-----0.090 -0.060 -0.030 0.000 Individual 95% CI treated Mean --+---------+---------+---------+--------no -0.0573 (------------*-------------) yes -0.0296 (------------*------------) --+---------+---------+---------+---------0.0800 -0.0600 -0.0400 -0.0200 PRD-1; 3-way ANOVA, all natural water Factor Type Levels Values temp fixed 3 5 22 30 water ty fixed 2 ground surface treated fixed 2 no yes Analysis of Variance for rate Source DF SS MS F P temp 2 0.025786 0.012893 3.83 0.031 water ty 1 0.028743 0.028743 8.53 0.006 treated 1 0.009224 0.009224 2.74 0.107 temp*water ty 2 0.015638 0.007819 2.32 0.113 temp*treated 2 0.000655 0.000327 0.10 0.908 water ty*treated 1 0.005491 0.005491 1.63 0.210 temp*water ty*treated 2 0.000957 0.000479 0.14 0.868 Error 36 0.121337 0.003370 Total 47 0.207831

PAGE 318

306 Appendix 4-J (continued) PRD-1; 2-way ANOVA, raw surface and groundwater Analysis of Variance for rate-r Source DF SS MS F P temp-r 2 0.013808 0.006904 7.76 0.004 type-r 1 0.029681 0.029681 33.34 0.000 Interaction 2 0.007660 0.003830 4.30 0.030 Error 18 0.016024 0.000890 Total 23 0.067173 Individual 95% CI temp C Mean ----+---------+---------+---------+------5 -0.028 (--------*--------) 22 -0.058 (--------*--------) 30 -0.087 (-------*--------) ----+---------+---------+---------+-------0.100 -0.075 -0.050 -0.025 Individual 95% CI type Mean -------+---------+---------+---------+---ground -0.0222 (-----*-----) surface -0.0925 (-----*-----) -------+---------+---------+---------+----0.0900 -0.0600 -0.0300 0.0000 PRD-1; 2-way ANOVA, raw surface and ground water, two higher temperatures only Analysis of Variance for rate Source DF SS MS F P temp hig 1 0.00331 0.00331 2.89 0.115 type hig 1 0.03423 0.03423 29.89 0.000 Interaction 1 0.00176 0.00176 1.54 0.238 Error 12 0.01374 0.00114 Total 15 0.05303 Individual 95% CI temp C Mean -------+---------+---------+---------+---22 -0.058 (------------*------------) 30 -0.087 (------------*------------) -------+---------+---------+---------+----0.100 -0.080 -0.060 -0.040 Individual 95% CI type Mean --+---------+---------+---------+--------ground -0.026 (-------*------) surface -0.118 (------*-------) --+---------+---------+---------+---------0.140 -0.105 -0.070 -0.035

PAGE 319

307 Appendix 4-J (continued) PRD-1; 2-way ANOVA, surface water Analysis of Variance for rate Source DF SS MS F P temp-sw 2 0.04072 0.02036 3.16 0.067 treated1 0.01447 0.01447 2.24 0.152 Interaction 2 0.00158 0.00079 0.12 0.886 Error 18 0.11615 0.00645 Total 23 0.17293 Individual 95% CI temp C Mean -+---------+---------+---------+---------+ 5 -0.023 (---------*---------) 22 -0.058 (---------*---------) 30 -0.123 (---------*--------) -+---------+---------+---------+---------+ -0.180 -0.120 -0.060 0.000 0.060 Individual 95% CI treated Mean -+---------+---------+---------+---------+ no -0.093 (-------------*------------) yes -0.043 (-------------*-------------) -+---------+---------+---------+---------+ -0.140 -0.105 -0.070 -0.035 0.000 PRD-1; 1-way ANOVA, raw surface water Analysis of Variance for rate sw Source DF SS MS F P temp sw 2 0.02102 0.01051 6.89 0.015 Error 9 0.01372 0.00152 Total 11 0.03474 Individual 95% CIs For Mean Based on Pooled StDev Temp C N Mean StDev --+---------+---------+---------+---5 4 -0.04075 0.02269 (------*-------) 22 4 -0.09350 0.03142 (------*-------) 30 4 -0.14325 0.05542 (------*------) --+---------+---------+---------+---Pooled StDev = 0.03904 -0.180 -0.120 -0.060 0.000

PAGE 320

308 Appendix 4-J (continued) PRD-1; 2-way ANOVA, groundwater Analysis of Variance for rate gw Source DF SS MS F P temp gw 2 0.000704 0.000352 1.22 0.318 treated 1 0.000241 0.000241 0.84 0.373 Interaction 2 0.000034 0.000017 0.06 0.943 Error 18 0.005183 0.000288 Total 23 0.006162 Individual 95% CI temp gw Mean --------+---------+---------+---------+--5 -0.0119 (-----------*------------) 22 -0.0201 (------------*-----------) 30 -0.0250 (------------*------------) --------+---------+---------+---------+---0.0300 -0.0200 -0.0100 0.0000 Individual 95% CI treated Mean -------+---------+---------+---------+---no -0.0222 (-------------*--------------) yes -0.0158 (-------------*--------------) -------+---------+---------+---------+----0.0280 -0.0210 -0.0140 -0.0070 PRD-1; 1-way ANOVA, raw groundwater Analysis of Variance for rate gw Source DF SS MS F P temp gw 2 0.000450 0.000225 0.88 0.448 Error 9 0.002304 0.000256 Total 11 0.002754 Individual 95% CIs For Mean Based on Pooled StDev Level N Mean StDev --+---------+---------+---------+---5 4 -0.01475 0.01571 (-----------*-----------) 22 4 -0.02200 0.01283 (-----------*-----------) 30 4 -0.02975 0.01887 (-----------*-----------) --+---------+---------+---------+---Pooled StDev = 0.01600 -0.045 -0.030 -0.015 0.000

PAGE 321

309 Appendix 4-J (continued) PRD-1; 2-way ANOVA, PBS control sets Analysis of Variance for PRD rate Source DF SS MS F P PRD set 3 0.001694 0.000565 1.69 0.268 PRD temp 2 0.005297 0.002649 7.92 0.021 Error 6 0.002007 0.000335 Total 11 0.008998 Individual 95% CI PRD set Mean ---+---------+---------+---------+-------4 -0.0352 (------------*------------) 5 -0.0086 (------------*------------) 6 -0.0220 (------------*------------) 7 -0.0387 (------------*------------) ---+---------+---------+---------+--------0.0600 -0.0400 -0.0200 0.0000 Individual 95% CI PRD temp Mean --+---------+---------+---------+--------5 -0.0142 (--------*--------) 22 -0.0084 (--------*--------) 30 -0.0556 (--------*--------) --+---------+---------+---------+---------0.0750 -0.0500 -0.0250 -0.0000 PRD-1; covariance analysis for PBD sets, temperature as covariant Factor Type Levels Values PRD set fixed 4 4 5 6 7 Analysis of Variance for PRD rate, using Adjusted SS for Tests Source DF Seq SS Adj SS Adj MS F P PRD temp 1 0.0023519 0.0023519 0.0023519 3.32 0.111 PRD set 3 0.0016938 0.0016938 0.0005646 0.80 0.533 Error 7 0.0049528 0.0049528 0.0007075 Total 11 0.0089985 Term Coef StDev T P Constant -0.00059 0.01596 -0.04 0.972 PRD temp -0.001343 0.000737 -1.82 0.111

PAGE 322

ABOUT THE AUTHOR David John was born and raised in Mesa, Arizona a nd graduated from Mountain View High School in 1991 with honors. He received a Bachelor of Scienc e degree from Northern Arizona University in Environmental Science with an emphasis in microbiology After two years workin g in industry and with the U.S. Forest Service on the Tont o National Forest of central Arizona, he attended graduate school at the University of Arizona. There we worked under Professor Charles Gerba and performed research on disinfection of microsporidia while earning a Master of Science degree in Soil, Water, and Environmental Science. Honors and awards he has received include the Robert M. Garrels fello wship, the Von Rosenstiel endowed fellowship, Soil, Water and Environmental Science departmental fellowship, Wesley Clark memorial scholarship, Gove Allen memorial scho larship, and NAU Student Employee of the Year.


xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001441482
003 fts
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 031203s2003 flua sbm s000|0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0000155
035
(OCoLC)53961836
9
AJM5922
b SE
SFE0000155
040
FHM
c FHM
090
GC11.2
1 100
John, David E.
q (David Eric)
0 245
Transport and survival of water quality indicator microorganisms in the ground water environment of Florida
h [electronic resource] :
implications for aquifer storage and waste disposal /
by David E. John.
260
[Tampa, Fla.] :
University of South Florida,
2003.
502
Thesis (Ph.D.)--University of South Florida, 2003.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 322 pages.
520
ABSTRACT: Ground water resources are heavily used for drinking water supply and often as a receptacle for waste water. One concern is the possible contamination of wetland areas by ground water receiving septic system infiltration. To investigate this, two tracer studies were performed using the bacteriophage PRD-1 by seeding septic systems adjacent to wetlands with the phage and monitoring migration towards wetland areas. Transport velocities were evaluated based on appearance of tracer in sampling wells at various distances from the injection point. Velocities were estimated to be 0.25 m/d and 0.4 m/d at the two sites. Some retardation with respect to the conservative tracer SF6 was observed, with a factor of about 1.5. Due to dry conditions, the water table was well below surface, so transport of the virus into surface water was not observed. Survival of public-health-related microorganisms in ground water is also a concern. The effects of temperature and total dissolved solids (TDS) on survival of 5 groups of indicator organisms were evaluated in controlled experiments. TDS did not have significant effects on inactivation of these microbes up to 1000 mg/l, but there was indication of reduced inactivation of enterococci at TDS concentrations of 3000 mg/l. Increased temperature consistently resulted in more rapid inactivation. Survival in aquifer and reservoir water samples was also evaluated, and significant effects due to water type, temperature, and pasteurization treatment were observed. Inactivation was more rapid in surface water sources, and pasteurization enhanced survival. For enterococci and DNA coliphage, pasteurization effects were more pronounced in surface water. DNA coliphage and perhaps fecal coliform appeared to be the more-conservative indicator organisms for aquifer injection monitoring. Lastly, it was observed that inactivation rates were considerably slower in pore water of saturated limestone than in the bulk water column of similar water sources and conditions, particularly for enterococci and fecal coliform.
590
Adviser: Rose, Joan B.
653
bacteria.
viruses.
inactivation.
temperature.
total dissolved solids.
wetlands.
septic systems.
contamination.
floridan aquifer.
aquifer storage and recovery.
690
Dissertations, Academic
z USF
x Marine Science
Doctoral.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.155