xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim
leader nam 22 Ka 4500
controlfield tag 007 cr-bnu---uuuuu
008 s2010 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0004799
Impact of tank material on water quality in household water storage systems in cochabamba, bolivia
h [electronic resource] /
by Cynthia Schafer.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains X pages.
Thesis (MSEV)--University of South Florida, 2010.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
ABSTRACT: The importance of water as a mechanism for the spread of disease is well recognized. This study conducted household surveys and measured several physical, chemical, and microbial water quality indicators in 37 elevated storage tanks constructed of different materials (polyethylene, fiberglass, cement) located in a peri-urban community near Cochabamba, Bolivia. Results show that although there is no significant difference in physical and chemical water quality between polyethylene, fiberglass and cement water storage tanks there is a difference in microbial contamination as measured by E. Coli counts (p = 0.082). Evidence points toward elevated water temperatures that increase along the distribution system (from 10.6C leaving the treatment plant) to within the black polyethylene storage tank (temperatures as high as 33.7C) as the most significant factor in promoting bacterial growth. Results indicate that cleaning frequency may also contribute to microbial water quality (p = 0.102).
Advisor: James Mihelcic, Ph.D.
Storage Tank Materials
x Civil & Environmental Eng.
t USF Electronic Theses and Dissertations.
Impact of Tank Material on Water Quality in Household Water Storage Systems in Cochabamba, Bolivia b y Cynthia Anne Schafer A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Environmental Engineering Department of Civil & Environmental Engineering College of Engineering University of South Florida Major Professor: James Mihelcic, Ph.D. Maya Trotz, Ph.D. Christian Wells, Ph.D. Date of Approval October 19, 2010 Keywords: Developing Country, Drinking Water, E. coli Peri Urban Storage Tank Materials Copyright 2010 Cynthia Anne Schafer
A CKNOWLEDGMENTS I would like to thank the National Science Foundation for support of this project through the Sustainable Water Management Research Experience in Bolivia Project (OISE 0966410). I would like to thank the following people for their help gathering data and laboratory work : Sara Connelly, Daragh Gibson, Gonzalo Mercado, Andre Riverez and Shirley Rodri guez. I would like to thank all the people of Tiquipaya that so kindly opened up their homes and their willingness to allow me to climb on their roofs in order to collect water samples I would also like to thank Linda Phillips and Dennis Magolan for their guidance before, during and after my trips to Bolivia. And finally, I would like to thank my advisor, Dr. James Mihelcic for his support and guidance throughout my studies and research.
i TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ .......... iii LIST OF FIGURES ................................ ................................ ................................ ........ v i ABSTRACT ................................ ................................ ................................ ......... v i ii INTRODUCTION ................................ ................................ ................................ ........... 1 Motivation and Hypotheses ................................ ................................ ................. 5 PREVIOUS RESEARCH ................................ ................................ ............................... 6 Waterborne Diseases ................................ ................................ ......................... 6 Distribution Systems ................................ ................................ ........................... 8 Health Issues of Stored Water ................................ ................................ ........... 10 Environmental Factors Affecting Stored Water Qualit y ................................ ....... 10 Water Storage Studies ................................ ................................ ....................... 12 STUDY LOCATION AND SYSTEM CHARACTERISTICS ................................ ...... 16 METHODS ................................ ................................ ................................ .................... 2 5 Background ................................ ................................ ................................ ...... 2 5 General Survey of Tank Type and Availability ................................ .................... 2 6 Sampling Procedures ................................ ................................ ......................... 2 6 Initial Water Quality Analysis ................................ ................................ .............. 2 8 In D epth Water Quality Analysis ................................ ................................ ........ 2 9 Temperature Study ................................ ................................ ............................ 3 2 Treatment Plant and Wells ................................ ................................ ................. 3 2 Statistical Analysis ................................ ................................ ............................. 3 3 Removal of Data ................................ ................................ ................................ 3 4 Potential Errors ................................ ................................ ................................ .. 3 4 RESULTS ................................ ................................ ................................ ..................... 3 6 Elevated Storage Tank Types ................................ ................................ ............ 3 6 Household Survey ................................ ................................ .............................. 3 8 Water Quality Initial Screening ................................ ................................ ........ 41 Water Quality In D epth Analysis ................................ ................................ ..... 5 7 Microbial Results ................................ ................................ ................................ 5 8 Temperature Study ................................ ................................ ............................ 6 3
ii Effect of Residence Time ................................ ................................ ................... 6 6 Treatment Plant and Wells ................................ ................................ ................. 6 8 DISCUSSION ................................ ................................ ................................ ............... 6 9 Community Perceptions ................................ ................................ ..................... 7 2 CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE RESEARCH ............... 7 4 REFERENCES ................................ ................................ ................................ ............. 7 7 APPENDICES ................................ ................................ ................................ .............. 82 Appendix A : IRB Approval L etter ................................ ................................ ........ 8 3 Appendix B : Study Information Sheet for Survey P articipants: Cochabamba, Bolivia ................................ ................................ ..... 8 5 Appendix C : Household Survey Q uestion naire ................................ .................. 8 7 Appendix D: Household Survey Questionnaire R esponses ................................ 9 5 Appendix E : Raw Data for Elevated Storage T anks in Tiquipaya Noreste (Bolivia) ................................ ................................ ......................... 9 7 Appendix F : Raw Data for Underground C isterns inTiquipaya Noreste (Bolivia) ................................ ................................ ...................... 10 1 Appendix G : Raw D ata for the Tiquipaya Noreste (Bolivia) W ater Distribution S ystem ................................ ................................ ...... 10 3 Appendix H : Raw D ata for the Tiquipaya Noreste (Bolivia) W ater Treatment P lant ................................ ................................ ........... 10 5 Appendix I : Results for MANOVA Comparing Water Quality Parameters for Samples Taken Directly from Storage Tanks with Those Tak en from T aps ................................ ................................ ......... 10 7 Appendix J : Results for MANOVA Comparing Water Quality P arameters for E ach T ank T ype ( Polyethylene, Fiberglass and Fiber C ement) ................................ ................................ ...................... 11 0 Appendix K : BART R esults ................................ ................................ .............. 1 1 5 Appendix L : BART Test Information S heets ................................ ..................... 1 1 6 ABOUT THE AUTHOR ................................ ................................ ..................... End Page
iii LIST OF TABLES Table 1: Burden of diarrheal disease by g lobal r egion, 2000 ................................ .......... 2 Table 2: Percent of positive test results for microbial contaminants from study in urban Trinidad. ................................ ................................ ................................ 12 Table 3: Sample survey questions concerning elevated storage tan k properties and h ousehold u se ................................ ................................ .......................... 2 8 Table 4: Distribution of samples taken directly from storage tanks and samples taken from taps by tank type ................................ ................................ ........... 2 9 Table 5 : Water q uality p arameters and a nalytical m ethods e mployed. ........................... 31 Table 6 : Detection limits of test kits used in laboratory analysis. ................................ .... 3 5 Table 7: Percentages of each tank type found within the Tiquipaya Noreste distribution system and of those included in the study ................................ ..... 3 6 Table 8 : Age d istribution of e levated storage t anks; 37 tanks sampled. ......................... 40 Table 9 : Frequency of rooftop water storage tank c leaning; 36 tanks sampled. ............. 41 Table 10 : Results for MANOVA comparing water quality parameters for samples taken directly from elevated storage tanks or from taps ................................ 4 2 Table 11 : Results for tests of between subject effects using MANOVA ......................... 4 3 Table 12: Assigned v alues for coded free chlorine data ................................ ................. 4 5 Table 1 3 : Overall physical and chemical water quality results for each water storage tank type in Tiquipaya Noreste (Bolivia) ................................ ............. 4 6 Table 1 4 : Results for r andomized b lock ANOVA of various water quality parameters versus tank age and cleaning schedule in Tiquipaya Noreste (Bolivia) ........... 50 Table 15: Results for one way ANOVAs comparing E. coli and total coliform counts for various cleaning sche dules ................................ ................................ ........ 5 3
iv Table 1 6 : Sample sizes for treatments for r andomized b lock ANOVA design. ............... 5 4 Ta ble 1 7 : Results for randomized b lock ANOVA of the effects of tank age and cleaning schedule on various water parameters within different tank types in Tiquipaya Noreste (Bolivia) ................................ ................................ 5 5 Table 18: One way ANOVAs for E. coli comparing storage tank types and treatments ................................ ................................ ................................ ..... 5 7 Table 1 9 : Percent of s amples that e xceed the B olivian w ater q uality s tandards for E. coli (0.0 CFU/mL) ................................ ................................ .................. 5 9 Table 20 : BART test results for three different microbial indicators reported as percent of positive tests recorded for each tank type. ................................ ...... 60 Table 21 : Maximum and minimum water t emperatures (C) recorded in elevated storage tanks in Tiquipaya Noreste (Bolivia) ................................ ................... 6 5 Table 2 2 : P values for two tail independent t tests comparing E. coli a nd total coliform counts within the distribution system, cisterns and elevated storage tanks in Tiquipaya Noreste (Bolivia) ................................ ................... 6 8 Table D1: Demographic information ................................ ................................ .............. 9 5 Table D2: Storage tank properties ................................ ................................ ................. 9 5 Table D3: Uses and practices o f storage tank ................................ ............................... 9 6 Table D4: Health e ffects of s tored w ater ................................ ................................ ........ 96 Table E 1: Elevated storage tank location and material and age characteristics ............. 9 7 Table E 2: Physical chemical water quality data for elevated storage tanks .................... 98 Table E 3: Total and free chlorine water quality data for elevated storage tanks ............. 99 Table E 4: Microbial water quality data for elevated storage tanks ................................ 10 0 Table F 1: Underground cistern characteristics ................................ ............................. 10 1 Table F 2: Physical chemical water quality data for underground cisterns .................... 10 1 Table F 3: Total and free chlorine water quality data for underground cisterns. ............ 10 2
v Table F 4: Microbial water quality data for underground cisterns ................................ .. 10 2 Table G 1: Tiquipaya Noreste water distribution system characteristics ........................ 10 3 Table G 2: Physical chemical water quality data for the Tiq u i paya Noreste water distribution system ................................ ................................ ....................... 10 3 Table G 3: Total and free chlorine water quality dat a for the Tiquipaya Noreste water distribution system ................................ ................................ ............. 10 3 Table G 4: Microbial water quality data for the Tiqu ipaya Noreste water distribution system ................................ ................................ ....................... 10 4 Table H 1: Tiquipaya Noreste water treatment plant c haracteristics ............................. 10 5 Table H 2: Physical chemical water quality data for the Tiquipaya Noreste water treatment plant ................................ ................................ ............................ 10 5 Table H 3: Total and free chl orine water quality data for the Tiquipaya Noreste water treatment plant ................................ ................................ ................... 10 6 Table H 4: Microbial water quality data for the Tiquipaya Noreste water treatment plant ................................ ................................ ................................ ........... 10 6 Table I 1 : Multivariate tests for water quality parameters for samples taken directly from storage tanks compared with those taken from taps ................ 10 7 Table I 2 : Tests of between subjects effects for water quality parameters for samples taken directly from storage tanks and those taken from taps ......... 10 7 Table J1 : Multivariate tests for water quality parameters for each tank type ................ 1 1 0 Table J2 : Tests of between subjects effects for water quality parameters for each tank type (polyethylene, fiberglass and fiber cement) .......................... 11 0 Table J3: Multiple comparisons using MANOVA and the TukeyHSD test statistic ....... 11 3 Table K 1 : Raw data for in depth microbial testing ................................ ........................ 1 1 5
vi LIST OF FIGURES Figure 1: Access to water and sanitation statistics and child mortality rates for Bolivia ................................ ................................ ................................ .............. 3 Figure 2: Causes of u nder 5 m ortality ................................ ................................ ............. 4 Figure 3: Elevated storage tank and cistern photos ................................ ....................... 1 8 Figure 4: Study location maps ................................ ................................ ...................... 1 9 Fig ure 5: Diagram of h ousehold w ater s ystem ty pical of Tiquipaya (Bolivia) .................. 20 Figure 6: Speciation plot of [HOCl]/[OCl ] ................................ ................................ ....... 22 Figure 7 : Tiquipaya Noreste (Bolivia) water treatment plant ................................ ........... 3 3 Figure 8 : Locations of all elevated storage tanks within study area. ............................... 3 7 Figure 9 : Five most commonly found elevated storage tanks observed in Tiquipaya Noreste community ................................ ................................ ......... 3 8 Figure 10 : Sample location map s in Tiquipaya Noreste community ............................... 3 9 Figure 1 1 : Results for conductivity, total dissolved solids, dissolved oxygen and pH for water storage tanks in Tiquipaya Noreste (Bolivia) ........................... 4 4 Figure 1 2 : Results for turbidity, free chlorine, total c oliforms and E. coli for water storage tanks in Tiquipaya Noreste (Bolivia) ................................ ................ 4 5 Figure 13: Results for conductivity, total dissolved solids, dissolved o xygen and pH by c leaning fre quency of elevated storage tanks in Tiquipaya Noreste (Bolivia) ................................ ................................ ........................... 4 7 Figure 1 4 : Results for turbidity, free chlorine, total c oliforms and E. coli by cleaning frequency of elevated storage t anks in Tiquipaya Noreste (Bolivia) ................................ ................................ ................................ ........ 4 8 Figure 1 5 : Results for conductivity, total dissolved solids, dissolved o xygen and pH by age of elevated storage tanks in Tiqupaya Noreste (Bolivia) .............. 4 9 Figure 1 6: Results for turbidity, free chlorine, total c oliforms and E. coli by age of e le vated storage tanks in Tiquipaya Noreste (Bolivia) ................................ .... 50
vii Figure 1 7 : Results of in depth analysis of iron, sulfate and nitrate levels in different storage tank types as well as within the distribution system in Tiquipaya Noreste (Bolivia) ................................ ................................ ........ 5 8 Figure 1 8 : Histogram of E. coli c ounts ................................ ................................ ........... 5 9 Figure 1 9 : Levels of heterotrophic aerobic and slime forming bacteria measured in distribution system and household cistern s and water storage tanks in Tiquipaya Noreste (Bolivia) ................................ ................................ ........... 6 2 Figure 20 : Levels of iron related bacteria measured in distribution system and household cistern s and water storage tanks in Tiquipaya Noreste (Bolivia) ................................ ................................ ................................ ......... 6 3 Figure 2 1 : Water temperature within three types of elevate d storage t anks in Tiquipaya Noreste (Bolivia) ................................ ................................ ............ 6 4 Figure 2 2: Difference between ambient air temperature and stored w ater t emperature in storage tanks in Tiquipaya Noreste (Bolivia) ......................... 6 5 Figure 2 3 : Water quality changes as water travels from the treatment plant through the system to household cisterns and storage tanks. ...................... 6 7
viii A BSTRACT The importance of water as a mechanism for the spread of disease is well recognized. This study conducted household surveys and measured several physical, chemical, and microbi al water quality indicators in 37 elevated storage tanks constructed of different materials (polyethylene, fiberglass, cement) located in a peri urban community near Cochabamba, Bolivia. Results show that although there is no significant difference in physical and chemical water quality between polyethylene, fiberglass and cement water storage tanks there is a difference in microbial conta mination as measured by E. Coli counts ( p = 0.082). Evidence points toward elevated water temperatures that increas e along the distribution system (from 10.6 C leaving the treatment plant ) to within the black polyethylene storage tank (temperatures as high as 33.7 C) as the most significant factor in promoting bacterial growth. Results indicate that cleaning frequency may also contribute to microbial water quality ( p = 0.102).
1 INTRODUCTION The importance of water as a mechanism for the spread of disease has long been recognized as seen by the large amount of peer reviewed articles concerning the relationship of health to water quality and sanitation ( e.g., Semenza et al 1998; Craun and Calderon, 2001; Egorov et al. 2002 ). In addition, international organizations such as the World Health Organization (WHO), the United Nations (UN) and the World Bank have given much attention to this subject. For example, a ccording to the WHO World Health Report (2004), approximately 3.2% of deaths and 4.2% of Disability Adjusted Life Years (DALYs) worldwide from diarrheal diseases are attributed to the consumption of contaminated water and lack of sanitation and hygiene practices. Th is corresponds to 88% of reported diarrheal diseases worldwide with over 99% of deaths occurring in developing countries, 90% of whom are children under the age of 5 (Nath et al. 2006). The UN reports that more than 2.2 million people, most of which resid e in developing countries, die each year due to diseases associated with poor water and sanitation. Table 1 provides global and regional data on disease burden from the year 2000 related to diarrheal diseases.
2 Table 1 : Burden of diarrheal disease by global r egion, 2000 Deaths and DALY Totals for 2000 Global Africa Americas South East Asia Europe East Mediterranean West Pacific Mortality due to Diarrheal Disease 3.2% 6.6% 0.9% 4.1% 0.2% 6.2% 1.2% DALYs due to Diarrheal Disease 4.2% 6.4% 1.6% 4.8% .5% 6.2% 2.5% Source: Nath et al. 2006 Often in developing countries with high morbidity and mortality numbers the health problem s are related to poor water quality, limited water availability, limited sanitation and/ or poor hygiene practices. Common interventions in these situations include: improving access to water, providing household treatment options, improving sanitation and hygiene education. The e ffect of improving access to water and sanitation services is most easily seen by looking at the under 5 mortality rates. For example, Bolivia has an under 5 mortality rate of 69 deaths per 1,000 live births while a s a region the Americas have an under 5 mortality rate of 25 deaths per 1,000 live births (WHO, 2006). Figure 1 shows how modest increases in access to water and sanitation services can help lower under age 5 mortality. Figure 1 shows that in 2002, 84% of the population in Bolivia had access to improved water sources and only 59% had access to sanitation services. In 19 90, when data for these two parameters began being recorded, under 5 mortality began decreasing at a greater rate. While this alone does not signify correlation, numerous studies have shown
3 that improving access to improved water and sanitation services ha ve shown that a correlation with reducing under 5 mortality rates exists ( e.g. Sobsey et al ., 2003 ). Figure 1 : Access to water and sanitation statistics and child mortality rates for Bolivia. a. Percent of Bolivian population with access to improved water s ource s; b. Percent of u rban Bolivian population with access to sanitation facilities; c. Under 5 m ortality per 1,000 b irths for Bolivia Source: Visualization from Gapminder World powered by Trendalyzer from www.gapminder.org Accessed online April 2010. Figure 2 shows the different causes of death for children under 5 years old. This figure shows that more than 10% of deaths for children under 5 are caused by diarrheal b. c. a.
4 diseases. Additionally, although more difficult to measure, early childhood diarrhea has shown to cause stunted growth and lower co gnitive function later in life (Berkman, 2002). Figure 2 : Causes of under 5 m ortality. Source: WHO, 2006. The issues discussed above can also be exacerbated by rapid population growth, espe cially in impoverished areas While the same organisms that make adults sick also make children sick, children are more susceptible to dying because their immune systems are not as well developed ; this effect is exacerbated when children a lao suffer from malnutrition ( Pelletier et al ., 1994 ).
5 Motivation and Hypothes e s The motivation for this study comes from the need for more research into water quality in modern water distribution systems and the causes of microbial contamination of water in household storage tanks. Numerous studies have been done focusing on physical, chemical and microbial water quality of household storage containers in situations where water is collected at a community source and then transported to the home ( e.g, Quick et al ., 199 9 Quick et al ., 2002 Cl asen and Bastable, 2003 Wright et al., 2004 ). There have also been studies performed that show how water quality degrades when supply is intermittent and as the residence time associated with distribution and storage increases ( Kerneis et al ., 1995, Toka jian and Hashwa, 200 3 ). However, few studies have been performed on elevated household storage tanks In addition, no peer reviewed articles were found by the author on field studies evaluating water quality of elevated household storage tanks commonly f ound in the developing world This study examines the effects of tank material, tank water temperature, and user behaviors on water quality in elevated household storage tanks in the city of Tiquipaya Bolivia The overall objective is to determine how the materials used to construct household water storage tanks and user operation/maintenance impact physical, chemical, and microbial quality of water in household storage tanks as well as document water quality as the water travels from the treatment plan t through the distribution system to the user This study will test three hypotheses : 1. Tank material impacts water quality within the household storage tank; 2. Tank material affects water temperature which impacts microbial water quality ; and 3. Tank factors such as cleaning frequency and age impact water quality within the household storage tank.
6 PREVIOUS RESEARCH Waterborne Diseases Access to safe water and sanitation facilities (e.g. latrines) as well as knowledge of proper hygiene practices can reduce the risk of illness and death from waterborne diseases, leading to improved health, poverty reduction, and socio economic development (CDC 2010 ). Water is an important vector for the transport of waterborne diseases which are generally caused by pathogenic microbes that can survive and often grow in water. Most waterborne diseases cause diarrheal illness and disproportionally affect children. Water can be contaminated by v arious pathways such as lack of hygiene, inad equate treatment or poorly maintained infrastructure. For example, an outbreak of t yphoid fever believed to be due to poor water quality in the distribution system in Dushanbe, Tajikistan between January 1996 through June 1997 led to 8,901 reported cases and 95 deaths (Mermin et al. 1999). Among a number of variables contributing to the spread of disease was a lack of residual chlorine in the distribution system (Mermin et al. 1999). The outbreak of ch olera that spread to 19 countries in Central and Sou th America in 1991 infected over 533,000 people and caused 4,700 deaths D rinking unboiled water was associated with becoming infected with V. cholerae (Swerdlow et al. 1992). A review published by Gundry et al (2004) found that samples of stored water c ontaminated with V. cholerae resulted in cholera cases and that treatment and improved storage interventions were successful at preventing cholera.
7 Numerous studies have found that the consumption of poor quality water is responsible for higher diarrheal inc idence (Semenza et al. 1998). However, unlike t yphoid f e ver and c holera which are each caused by a specific organism, numerous pathogens are responsible for causing diarrhea. As a result, low levels of indicator bacteria may correspond to high numbers of diarrhea cases and high levels of indicator bacteria may not always correspond to an increased number of cases of diarrhea (Gundry et al. 2004). This may be due to indicator bacteria not being a good measure of pathogens; this has been shown to be the case with thermotolerant coliforms (Gleeson and Gray, 1997; Hamer et al. 1998; Gundry et al. 2004). Additionally, diarrhea is a symptom of many illnesses, which makes the association with improved water quality and a reduction of diarrhea incidence diff icult to prove (Gundry et al. 2004). One cause of waterborne pathogens being present in water distribution systems is the failure to disinfect the water (Cardenas et al 1993; Rab et al. 1997; Craun et al 2002 ). The primary reason to maintain a disinfectant residual in a water supply is to guard against the re growth of pathogens and to neutralize pathogens that enter the system after treatment. Lack of a disinfectant residual in a system in which the water has undergone disinfection by chlorinat ion often indicates that contaminants are entering the system (Agard et al ., 2002 ). It has been shown that low concentration s of free chlorine, less than 0.2 mg/L, in potable water has led to substantially more coliform occurrences than water with higher f ree chlorine concentrations (LeChevallier et al ., 1996). A study done in Trinidad has shown a correlation between the loss of a residual chlorine concentration and an increased prevalence of total coliforms (from 0% to 80%) in water as it travels from the treatment plant to the user (Agard et al. 2002).
8 Distribution Systems In the U S recent focus on water quality issues has been on chemical contamination occurring within the distribution system. Evidence has been found indicating that the switch from chlorine to chloramine for disinfection increases corrosion of brass pipe which leads to elevated lead levels in the water (Edwards and Dudi, 2004) The presence of chlorine has also been implicated in higher rates of copper corrosion (Boulay and Edwards 2001). Another study has shown that maximum corrosion rates occur at 30 C which coincides with maximum bacterial growth (Arens et al., 1995). In developing countries, the focus has been on improving microbial water quality of drinking water supplies Although the presence of a water distribution system is often seen as a sign of improved water quality, it does not imply that the water is free of pathogens and therefore adequate for human consumptio n (Lee and Schwab, 2005). Often times, water leaving treatment systems or arriving at community taps is microbiologically safe, however contaminants may enter a distribution system after treatment or during household storage (Nath et al. 2006). In fact, in the United States alone approx imately 18% of waterborne disease outbreaks were linked to contaminants entering the distribution system after treatment (Craun and Calderon, 2001). Worldwide, contaminated water has been transported through distribution systems and has been implicated in the spread of outbreaks of typhoid fever, c holera and diarrheal diseases (Semenza et al. 1998; Egorov et al. 2002; Mermin et al. 1999; Swerdlow et al ., 1992 ). These pathogens have been found to be present in unimproved as well as improved water sources (Gundry et al ., 2004). There is also a growing body of evidence that distribution systems can cause a decrease in the quality of water which can lead to illness in consumers in developed countries
9 (e.g. LeChevallier et al ., 1996; Craun and Calderon, 200 1), emerging countries (e.g. Gayton et al ., 1997; Mermin et al. 1999; Basualdo et al. 2000; Egorov et al ., 2002) and developing countries ( e.g. Geldenhuys, 1995; Dany et al ., 2000; Agard et al ., 2002 ; Lee and Schwab, 2005 ). Compounding the issue is the c ommon practice in some communities of storing large volumes of water at the household level which enables contaminant organisms to grow and multiply. In many communities, treatment of water for drinking and cooking occurs within the home even when the wate r is piped to the household. In both rural and urban distribution systems f ecal contamination may enter a piped water supply due to deficiencies such a poor source quality, inadequate treatment or disinfection, and infiltration of contaminated water (e.g. sewage) (Sobsey et al. 2003). This is often due to poor infrastructure maintenance of the distribution system. Old and failing infrastructure leads to stoppages in service thereby requiring residents to store large quantities of water within the househol d in large storage tanks Such storage offers another route for contamination to enter the water before consumption (Nath et al. 2006). Another way that contaminants can enter the water distribution system is through the addition of untreated water into the distribution network (Ford, 1999; Craun and Calderon, 2001). This can be either intentional, for example where there is more than one source of water for a distribution system and n ot all sources are treated; or i t can be unintentional as is the case for leaky systems. The addition of untreated water may result in the presence of microbes, some possibly pathogenic, causing the consumer to become ill (Ford, 1999; Craun and Calderon, 2001). Contaminates can also enter water distribution systems by other pathways; studies have shown that failure to disinfect or to maintain a disinfectant residual (LeChavallier et al. 1996); long residence times (Tokijian and Hashwa, 2004 ); and changes in pressure within the network (LeChevallier et al .,
10 White Paper N o Date) can all lead to the presence of pathogens within a distribution system. Health Issues of Stored Water Microbial quality of potable water supplies is important not only in the developing world but also in developed countries. WHO (2006) guidelines state that water intended for human consumption should contain no microbiological agents that are pathogenic to humans. The WHO (2006) guidelines for Escherichia coli ( E. coli ) and thermotolerant coliforms are 0 colony forming units ( CFU) per 100 mL because even low levels of fecal contaminants may potentially cause illness. Sobsey (2006) concluded that world wide as well as in the US the greatest risk of waterborne disease is due to microbial contamination of potable water supplies. I n developing countries, it is estimated that the consumption of unsafe drinking water is responsible for 15 % to 20% of community diarrheal disease, with recent studies indicating that the percentages may even be higher (Sobsey et al ., 2003). In developed c ountries similar issues remain. B etween 15% and 30% of community diarrheal disease is a result of contaminated municipal drinking water despite the state of the art treatment technology employed (Payment et al ., 1991, 1997 from Sobsey 2003 ). Environmental Factors Affecting Stored Water Quality Temperature of the stored water is an important influence on the growth rate of bacteria that have survived treatment processes. Various field studies have shown that significant bacteria growth can occu r in water of 15C or higher ( Fransolet et al., 1985; Donlan and Pipes, 1988; Smith et al ., 1989; Donlan et al ., 1994 From LeChevallier et al. 1996). For example, Fransolet et al. (1985 ) showed that a temperature increase from 7.5C to 17.5C reduced th e lag phase of growth for Pseudomnas putida from 3 days to 10 hours
11 (From LeChevallier et al ., 1996). Another study found that coliform bacteria occurred more frequently a nd in higher concentrations at water temperatures greater than 15C (LeChevallier et al. 1996). Results from that study indicate that for a temperature increase from 5C to greater than 20C there was an 18 fold increase of coliform occurrence in free chlorinated systems ( p < 0.0001) (LeChevallier et al ., 1996). Turbidity in water is us ually caused by suspended matter such as clay, silt, organic and inorganic matter, plankton and other microorganisms and is a useful water quality indicator (LeChavallier et al. 1981). These particles can provide either nutrients for bacteria or other pat hogens, or they may protect microorganisms themselves from chlorination (LeChavallier et al. 1981). A study by LeChavallier et al. (1981) showed that coliforms in high turbidity wate r (13 NTU) were reduced by 80% f rom their original concentration after chlorination while coliforms in low turbidity water (1.5 NTU) were undetectable after chlorination. Their results also showed that g iven a constant chlorine dose a turbidity increase from 1 NTU to 10 NTU results in an e ightfold decrease in dis infection efficiency Residence time has major impact on water quality. Many studies have shown that water quality degrades as the water travels through the distribution system and in some cases is stored before use (e.g., Evison and Sunna, 2001; Tokajian and Hashwa, 200 3 ) A study of a water distribution system in urban Trinidad found that microbial water quality degraded significantly as the water traveled through the distribution system (see Table 2 ) even though the reservoir repeatedly tested negat ive for microbial contamination (Agard et al. 2002). The presence of E. coli suggests fecal contamination is occurring within the distribution system.
12 Table 2 : Percent of positive test results for microbial contaminan ts from study in urban Trinidad Drinking Water Samples from Households in Urban Trinidad (n = 104) Total Coliforms Thermotolerant Coliforms E. coli Treated Reservoir Water 0% 0% 0% Distribution System 46.9% 16% 33.3% Household 80.8% 53.8% 67.3% Source: Agard et al., 2002 Water Storage Studies Microbial re growth in potable water supplies is often a problem that is intensified by household water storage practices. A laboratory study found that f actors such as long retention times of 4 to 7 days low or no chlorine residual and temperatures above 15 C have all been shown to increase microbial re growth in commonly used 1000 L fiberglass, polyethylene and cast iron household storage tanks (Evison and Sunna, 2001). This study also found that water temperature inside the tank and tank age were the parameters mo st important for bacterial growth and were responsible for 77.7% of the heterotrophic plate count values measured for water stored for 4 days (Evison and Sunna, 2001). Additionally, the HPC counts between the water stored for 4 days and the water stored fo r 7 days were not significantly different which this author believes indicates that the bacteria in the tank had been shocked initially by the chlorination but had survived in the distribution system and were able to grow in the conditions provided by the storage tank and that an increase in bac terial growth may be observed for shorter residence times Furthermore, this study did not find significant variations in HPC counts or in physical and chemical parameters between the different tank types tested (po lyethylene, fiberglass and cast iron). However, it did find that the bacteria taxa within
13 the different tanks did differ, most likely due to differences in water temperature and light penetration (Evison and Sunna, 2001). A separate laboratory study look ing at the effects of cast iron and black polyethylene household storage tanks (1000 L capacity) found that the stored water deteriorated significantly ( p in both types of storage tanks, but did not find a significant difference in HPC counts between the two types of storage tanks (Tokajian and Hashwa, 200 3 ). HPC counts varied seasonally, with the highest levels being measured during the summer months (Tokajian and Hashwa, 200 3 ). Increased microbial gro wth in household storage tanks compared to source water may also be due to the design of household storage tanks I t is not possible to completely empty most tank s, and that allows for sediment buildup which can act as a growth medium for microbes in the incoming water (Tokajian and Hashwa 2004 ). This leads to persistence o f coliforms in the stored water. I ncreased storage time, water temperature and microbial quality of the incoming water are also significant factors that contribute to poor water qualit y (Tokajian and Hashwa 2004 ). One study found significant total coliform and E. coli growth in black polyethylene storage tanks in rural Bolivia, however, both total coliforms and E. coli were also detected at the source indicating the problem is occurr ing prior to point of use (Omisca, 2010). More common are studies on household storage containers used to retrieve water and store it inside the home. For example, a study in Malawi found that fecal coliform levels
14 increased in household storage containe rs after only 1 hour of storage Even when investigators chlorinated water in storage containers contamination was only eliminated for the first 4 hours after collection After 6 hours of storage there was considerable microbiological growth (Roberts et a l ., 2001). A study looking at post supply drinking water quality in rural Honduras (Trevett et al ., 2004) found that source water quality appeared to be a significant factor in determining household water quality and that storage factors such as covering the household storage tank, tank material and residence time did not make a significant difference on the stored water quality. There was also no correlation between storage container type and water quality, although this may be due to the relatively smal l sample size (43 storage containers). The source water in this study came from hand dug and bore hole community wells of varying water quality, but every source saw a deterioration of water quality between collection and consumption. Contamination was mea sured by the presence of thermotolerant coliforms found in the household storage containers. These containers were either made of plastic or clay and had either wide openings in which water was ladled or dipped out or narrow openings in which water was pou red. Residence time was determined simply by asking the female head of household the last time water was collected; no specific times were reported. Due to the small size of the water storage containers (~25 L) this author believes the residence ti mes to have been relatively short (< 1 day). This indicates that contamination was occurring between the point of supply and consumption and that the bacteria were able to grow within the household storage container Clasen et al. (2003) noted that inter vention studies that employ a 3 part intervention program involving 1) narrow mouth storage containers with spigots that prevent hands
15 from entering container; 2) point of use disinfection; and 3) community hygiene education have led to reductions in water borne disease incidence as can be seen by a 50% reduction in diarrhea incidence in Bangledash (Sobsey et al ., 2003), 44% and 50% Bolivia (Quick, 2002 and Sobsey et al. 2003, respectively ) and 62% in Uzbekistan (Semenza et al. 1998). Another in tervention study using a narrow neck clay container found that cholera carrier rates were 17.3% in the control group and 4.4% in the intervention group (Deb et al ., 1986). Th ese results agree with the results from Trevett et al. (2005) which found that the type of storage container and whether the container allowed contact of hands with the stored water were associated with increased diarrheal disease incidence.
16 STUDY LOCATION AND SYSTEM CHARACTERISTICS The department of Cochabamba is located in the central pa rt of Bolivia on the eastern edge of the Andes Mountains ( Figure 4 a). It is divided into 47 municipalities and has an area of more than 55,000 km 2 While a majority o f the residents speak Spanish, there are three additional languages spoken in the area, Quechua, Aymara and Guaran, the first two with a significant number of speakers. The capital of the department of Cochabamba is also called Cochabamba. It is the most populated city in the department. The department has 1,455,000 inhabitants with 51% of the population living in urban areas and 49% living in rural areas ( Insituto Nacional Estadistica de Boliva, 2009 ) This study takes place in the peri urban mun icipal ity of Tiquipaya ( Figure 4 b) which is located 11 km west of the city of Cochabamba. Due to its proximity to Cochabamba, Tiquipaya is quickly becoming an urban area, as is shown by a yearly population growth rate of over 13% (Insituto Nacional Estadistica de Boliva, 2009). The munic ipality of Tiquipaya is divided into 6 districts with D istricts 1, 2 and 3 are located in the mountains and are sparsely populated and Districts 4, 5 and 6 are located in the valley Districts 4, 5 and 6 are more densely populated and these districts are a lso where most agricultural activity in the region occurs (Butterworth et al. 2007). The valley area represents less than 10% of the total area but is where 71% of the population resides (Butterworth et al ., 2007).
17 Within Districts 4, 5 and 6 of Tiquipa ya there are about 40 neighborhoods each with thei r own water distribution system that provide s residents with household water. Approximately 50% of the water distribution systems in the region have been built within the last 15 years ( Mejoramiento del Si stema de Agua Potable y Ampliaci n de la Red de Alcanterillado Sanitario de la Comunidad Colcapirhua Tiquipaya, 2003). Water for these systems comes from groundwater and rivers; the region is underlain with two aquifers, one at about 45 meters and the othe r at about 80 meters depth ( Ing. Mario Severiche 2009). The shallower of the two aquifers is said to have been contaminated from nearby septic systems ( Ing. Mario Severiche 2009) Historically water availability was periodic and as a result, most househ olds have underground cisterns which store water before it is pumped to the water storage tanks located on the roofs of their homes in order to have a constant supply of water However many of the water distribution systems within the municipality have be en updated in recent years and now almost 60% of the systems provide service 24 hours a day (Mejoramiento del Sistema de Agua Potable y Ampliacin de la Red de Alcanterillado Sanitario de la Comunidad Colcapirhua Tiquipaya, 2003). The residents say that the water is of poor quality. Figure 3 a and 4 b show an elevated water storage tank and an underground cistern respectively.
18 Figure 3 : Elevated storage tank and cistern photos. a) Elevated water storage tank located on the roof of a home; b) Underground cistern located next to home near street. Th ere are over 80 ,000 inhabitants in Tiquipaya (Insituto Nac ional de Estadstica, 2009). Tiquipaya has an area of 320 km 2 (Bustamante et al. no date) and Districts 4, 5 and 6 are divided into about 40 neighborhoods. Most neighborhood s have their own water distribution system, most of which are operated by community organizations, or in the urban area, a larger association of multiple systems which is operated by the C omit de Agua Potable y Alcantarillado para Tiquipaya (COAPAT) The scope of this study is office in District 4 of Tiquipaya. See Figure 4 for study locat ion.
19 Figure 4 : Study location maps. a) Bolivia and its departments; b) Tiquipaya, study location shown in or ange. Each grid represents 1 by 1 km. The specific water distribution system under in vestigation has approximately 500 connections with about 50% of households using an elevated water storage tank ( Ing. Hector Escalera Estrad 2010). The treatment plant was constructed about 15 years ago while the distribution system itself was updated in 2007 2008 to use PVC pipe (Ing. Hector Escalera Estrad, 2010). The tanks are made of various materials such as fiber cement fiberglass and polyethylene In addition to the elevated household storage tank, almost every household also has a below ground cistern for additional water storage. Figure 5 shows that water from the distribution system feeds into the below ground cistern which is then pumped to the elevated storage tank before being used throughout the house. a. b.
20 Figure 5 : Diagram of h ousehol d water s ystem typical of Tiquipaya (Boli via) Water flows from distribution system to an underground cistern to an elevated storage tank. Water for the system comes mainly from the River Khora but is also supplemented by two wells. Water from River Khora is also shared with farmers in the area with approximately 1/6 th of the flow going towards irrigation (Butterworth et al. 2007). The river water is treated and then mixed with the well water for distribution. Treatment of the river water consists of a sedimentation basin and storage tank upstr eam of the main treatment plant. From there, the water is piped to the treatment plant. The water passes through a series of open tanks to encourage sedimentation of suspended solids; the water is then chlorinated and enters a closed storage tank before en tering the distribution system. Each day, 2 kg of chlorine in the form of NaOCl (assumed 100% purity) is mixed wit h 450 liters of water and then combined with water from the river over the course of the day with the goal of achieving an approximate concent ration of 0.6 mg/L Cl 2 ( Ing. Hector Escalera Estrad 2010). The desired chlorine residual is between 0. 6 and 0. 7 mg/L as it leaves the treatment plant and 0.2 to 0.3 mg/L when it arrives at homes or other connections ( Ing. Hector Escalera Estrad 2010 ) Re sidents generally
2 1 have water service 24 hours a day; however, service is occasionally interrupted for system cleaning and maintenance and for road and sewer construction. In order to determine if a sufficient amount of chlorine was being added to the ri ver water, the following calculations were made Based on this calculation, the amount of free chlorine in the treated water should be about 0.5 mg/L, which does not meet the treatment plant goal of 0.6 to 0.7 mg/L. Additionally chlorine is a strong oxidant and th ese calculation s d o not take into affect reactions of chlorine with reduced species in the water which would reduce the amount of chlorine available for disinfection In chlorine chemistry, there are three forms of chlo rine; total chlorine, free chlorine and combined chlorine. Total chlorine is the sum of free chlorine and combined chlorine, free chlorine is the chlorine available for disinfection in the form of HOCl and OCl and combined chlorine is chlorine that has r eacted with nitrogen containing compounds to form chloramines. Chloramines can still deactivate microbial contaminants, but the reaction mechanism is slower than with free chlorine. HOCl is a more powerful disinfectant than OCl ; concentrations of HOCl and OCl vary with pH.
22 In the case of the Tiquipaya Noreste water treatment plant, pH varies between 6.5 and 7.8 The associated relation of HOCl to OCl is shown by the following equations. Assuming the solution behaves ideally (i.e., = 1), a t 25 C (Benjami n, 2002) Rearranging, At a pH of 6.5, At a pH of 7.8, Figure 6 displays this information graphically. Figure 6 : Speciation plot of [HOCl]/[OCl ]. 0 10 20 30 40 50 60 70 80 90 100 0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 45.00 5.5 6.5 7.5 8.5 [HOCl]/[OCl ] pH [HOCl]/[OCl ] % [HOCl]
23 Due to the low cost, 10 Bs per 20 m 3 or 0.5 Bs per m 3 of water ($1.43 USD per 5,283 gallons or $0.27 per 1,000 gallons), water usage is quite high within the community. The engineer that oversees the water distribution system estimated water usage to be between 150 and 200 liters per person per day (~ 40 5 0 gallons per person per day). The low cost of water means that not very much money is collected ; improvements to the system can only be made with national government fund ing. Money collected from users is used to pu rchase chlorine and electricity for pump s. In Tiquipaya, the rainy season begins in December and ends in May; the rest of the year it is dry with occasional rainfall. Days are usually warm year round, 24 27 C and nights cool off to about 5 12 C (Weather Underground 2010 ). During the dry season, both the wells and the river water are used to provide water to the distribution system. The wells provide 6 10 L/s of water and the river supplies about 30 L/s but has the capacity to provide 40L/s. During the rainy season, the river water is t oo turbid for use and only the wells are used which causes water shortage problems (Ing. Hector Escalera Espad, 2010) The following calculations w ere made based on these numbers.
24 The assumptions used in these calculations were that there were 8 people per household (based on results from the household survey) and that the pumps for the well operate on a 24 hour/day basis. Based on these results, it appears that water demand is much lower than water availability, even in the case when only one source is used. The discrepancy may be due to a number of reasons such as inaccurate production rates of water from the wells or river, a greater number of connections than reported, or significant leakages in the system.
25 METHODS Background All data collection occurred during June, July and August of 2010 (winter months in Bolivia) in the co mmunity served by the Tiquipaya Noreste water distribution system. There are approximately 500 connections to the distribution system (Ing. Hector Escalera Estrad 2010 ). A pproximately 150 households in the study site had visible elevated water storage tan ks and 37 (25%) of those tanks are included in this study. Additionally, water samples were taken from 14 different underground cisterns, 7 locations within the distribution system, both wells and at 9 locations within the treatment plant. For the in depth microbial analysis, 11 tanks, 8 cisterns and 2 locations within the distributi on system were revisited for further analysis. Figure 10 in the R esults c hapter should be consulted for location information related to the various sampling points A ll households included in the study are provided water by this distribution system and have an elevated storage tank. The majority of sampling occurred between the hours of 8:00 am and 12:00 pm, however, on two occasions sampling was done between 3:00 pm a nd 6:00 pm in an attempt to obtain samples from households where homeowners were not present during the earlier sampling period. Measurements for the temperature study were taken every 30 minutes during daylight hours (7:00 am 7:00 pm)
26 General Survey of Tank Type and Availability Initially, the households, schools and businesses that are provided water by the Tiquipaya Noreste water distribution system were surveyed for the different types of water storage tanks present. The location and tank type of each tank was recorded. This was achieved by walking the streets of the community and noting the types of tanks present in homes, schools and other businesses and marking the locations with a Garmen eTrex H GPS (Olathe, Kansas) Tanks found in houses or buildings that appeared to be uninhabited were not counted. From this information the five most common tank types were selected for the study and were assigned numbers The tanks were then randomly selected by a random number generator and a list of tanks and their corresponding GPS locations was created Sampling Procedures An initial water quality screening of 37 elevated storage tanks and 14 underground cisterns was performed. Additionally, samples from various locations within the water distributio n system, both wells and treatment plant were taken. In addition to these initi al water quality measurements, a subset of 11 elevated storage tanks and 3 cisterns were chosen for in depth analysis (see next section). The households that were randomly sele cted were then visited in an attempt to obtain a water sample and administer a survey, however, often times the homeowner was not present and the sample was not obtained. In this situation, the next household on the list of households designated for furthe r analysis was visited. Due to numerous situations in which homeowners were not present almost every home, school or business with an elevated storage tank in the community served by the Tiquipaya Noreste water distribution system was visited in order to obtain a sufficient number of samples. In the case where water samples were obtained from schools, only survey questions pertaining to storage tank characteristics
27 and behaviors were used. See Figure 10 for a figure showing locations of the elevated storage tanks sampled in this study Interviewers obtained informed consent of study participants before conducting survey s or sampling (See Appendix A for the IRB Approv al form, Appendix B for Study Information Sheet and Appendix C for Study Questionnaire) All respondents were of 18 years of age or older. If someone under the age of 18 answered the door, investigators asked if an adult was present. In the event that an adult was not present, the household was visited at a later date when an adult was prese nt. If the homeowner/school director/business owner agreed to participate in the study a survey asking about use and behaviors related to the rooftop storage tank was administered. The survey was semi structured and questioned the user about water storag e tank age, cleaning and disin fection frequency and practices, see Table 3 for example question s The detailed survey (i.e., the Study Questionnaire) is provided in Appendix C
28 Table 3 : Sample survey questions concerning elevated storage tank properties and household u se. Water Storage Tank Properties and Access to Water What material is your tank made of? What is the age of the tank? How many days a week do you have access to piped running water? When you have access to piped running water, how long do you have access? Household Water Practices & Use Is the water Stored in the tank used for drinking water? What is the water from the storage tank used for? In general, how frequently do you clean your storage tank? What do you use to clean your storage tank? Initial Water Quality Analysis Physical/chemical parameters of the water in the rooftop storage tank were measured on site using a Hydro Lab Quanta Probe (Hach, Loveland, CO) The Hydro Lab Quanta Probe measures temperature, conductivity, total dissolved solids, dissolved oxygen, pH, and turbidity. In addition, water samples tot aling 350 mL were collected in two separate bottles for further analysis. A 100 mL plastic bottle containing sodium thiosulfate (as provided by Idexx Laboratories) was used to collect the water for analysis of coliforms and E. coli a nd a sterile 250 mL HDPE bottle was used to collec t water for free and total chlorine analysis. Sterile s ample bottles and all laboratory equipment were purchased and transported to Bolivia Initially, samples were tested for lead and copper H owever because detectable levels of lead or copper were not det ected in initial samples and PVC pipe is used for the distribution system, lead and copper testing was discontinued after an initial round of testing All samples were stored in a cooler at 4C and analyzed within 6 hours of collection at our field lab oratory
29 Whenever possible, physical/chemical parameters were measured and water samples were taken directly from the water storage tank. However, some homeowners were not comfortable allowing someone to climb on their roof in order to collect a water samp le directly from the storage tank. Of the 37 elevated storage tanks sampled, 20 (54%) of the samples were taken directly from the tank while 17 (46%) samples were taken from taps connected to the tank In the case where the sample was collected from a tap it was taken from the tap location closest to the tank. The tap was allowed to run for 30 seconds before the sample was collected. See Table 4 for information regardi ng the number of samples taken from tanks and taps for each tank type. Table 4 : Distribution of samples taken directly from storage tanks and samples taken from taps by tank type. Storage Tank Type Number of Samples Taken Directly from Storage Tanks Number of Samples Taken from Taps Polyethylene 11 ( 69%) 5 (31%) Fiberglass 5 (45 %) 6 (55 %) Fiber Cement 8 (80 %) 2 (20 %) Table 5 lists the parameters measured in both the initial and in depth water quality analysis studies. In order to measure physical parameters with the Quanta Hydrolab probe, a 4 liter glass jar was used to collect water from the tap and then the probe was placed in the jar and results were recorded. Data locations were noted whether the sample was collected direct ly from the tap or directly from the storage tank. In D epth Water Quality Analysis In addition to the initial water quality measurements, a subset of 11 elevated storage tanks, 4 cisterns and 2 locations along the distribution system were chosen for a mor e
30 in depth microbial analysis. Table 5 lists the parameters measured and the method used to measure them for both the initial study and the in depth analysis. Elevated storage tanks were chosen based on accessibility and willingness of homeowner to participate further. At this time of the study 3 samples from distribution system and the water leaving directly from treatment plant were also chosen for in depth analysis. Samples were collected in 100 mL plastic bottles containing sodium thiosulfate (as provided by Idexx Laboratories) for coliforms and E. coli analysis and a sterile 250 mL HDPE plastic bottle was used to collect water for free and total chlorine, iron, nitrate, sulfate, iron related bacteria, heterotrophic aerobic bacteria and slime form ing bacteria analysis. Samples were stored in a cooler at 4 C and analyzed within 6 hours of collection.
31 Table 5 : Water quality parameters and analytical methods e mployed. Parameter Method Screening Analysis In Depth Analysis Temperature Quanta Probe in situ measurement pH Quanta Probe in situ measurement Turbidity Quanta Probe in situ measurement Conductivity Quanta Probe in situ measurement Dissolved Oxygen Quanta Probe in situ measurement Total Dissolved Solids Quanta Probe in situ measurement Total Coliforms Idexx Laboratories Coli Lert Quanti Tray/2000 E. coli Idexx Laboratories Coli Lert Quanti Tray/2000 Total Chlorine Hach Test Kit: Smart Colorimeter II Chlorine Free Chlorine Hach Test Kit: Smart Colorimeter II Chlorine Iron Lamotte Smart Reagent System Nitrate Lamotte Smart Reagent System Sulfate Lamotte Smart Reagent System Copper Lamotte Smart Reagent System Lead Lamotte Smart Reagent System Alkalinity Hach Alkalinity Test Kit Iron Related Bacteria BART TM Test Kit Heterotrophic Aerobic Bacteria BART TM Test Kit Slime Forming Bacteria BART TM Test Kit For Total Coliforms and E. coli measurements the Coli Lert Quanti Tray system (IDEXX Laboratories, Westbrook, ME) was used which employs a Most Probable Number (MPN) method which is used to enumerate colony forming units (CFU) per 100 mL.
32 Temperature Study W ater temperature was measured inside three types of elev ated storage tanks for a period of 12 hours. A temperature probe ( TDSTestr11+, Oakton Instruments Vernon Hills, IL ) was placed within the tank and measurements were recorded every 30 minutes over a period of 12 hours covering the time of sunrise to sunse t (7:00am 7:00pm) Three tanks were included in th e temperature study. Both the fiber cement tank and the fiberglass tank were elevated and remained in direct sunlight throughout daylight hours. The polyethylene tank was located at ground level with a wall located on its west side. This meant that starting at about 2:30pm the tank was in the shade Since most storage tanks included in this study were located on rooftops, the storage tanks chosen for the temperatu re study are representative, since they too were exposed to sunlight through most of the day. Treatment Plant and Wells In addition to the water sampling previously mentioned, samples were taken from 8 locations within the treatment plant and at both well source s. Temperature, conductivity, total dissolved solids, dissolved oxygen, pH and turbidity measurements were measured using the Hydro Lab Quanta Probe. Total and free chlorine analysis in locations after disinfection wa s performed at the time of sampling as well as in the field laboratory. Additionally, source water, water after initial sedimentation water entering treatment plant (Item 1, Figure 7 b ), within the treatment plant (Items 2 and 3, Figure 7 b ), water before disinfection (Item 4, Figure 7 ) water after disinfection (Item 5, Figure 7 ) water from the s torage tank before distribution system, both source wells and 3 locations within the distribution system were analyzed for iron, nitrate, sulfate and alkalinity. The sample taken from the storage tank before the water enters the distribution system was also analyzed for iron related bacteria, heterotrophic aerobic
33 bacteria and slime forming bacteria. See Figure 7 b for treatment plant sampling locations. Figure 7 : Tiquipaya Noreste (Bolivia) water treatment plant. a) Photo s of Tiquipaya Noreste treatment plant; b) T reatment plant schematic and sampling locations Statistical Analysis Statistical analysis included a series of one way randomized block ANOVAs and gene ral linear MANOVAs as well as multiple regression analysis to determine if correlations and relationships between water quality parameters exist Two sample t tests were performed to analyze changes in water quality at different points in the system. S tati stical analysis was performed using Minitab 15 software ( LEAD Technologies, Inc. State College, PA) and SPSS PASW Statistics, v. 18.0 software (IBM, Somers, NY) a b .
34 Removal of Data Due to measurements of total coliforms and E. coli that were too high to count in one fiberglass tank and associated cistern that were not located within the water distribution system under study, these data were removed from the study for analysis. Additionally, it was found that for fiber cement tanks total chlorine measureme nts taken from taps were statistically different from measurements taken directly from fiber cement tank s. These data were also removed from the analysis. Potential Errors The potential for errors in sampling arises due to the inability of the researcher to view every elevated storage tank which may have resulted in underreporting of the numbers and types likely storage tanks Another potential source of error is related to the detection limits of the equipment. For example, 62 % of total chlorine and 75% of free chlorine measurements were reported at or below the lower detection limit (0.02 mg/L as Cl 2 ) The value from the instrument was coded into three categories as shown in Table 12 and displayed in Figure 12 The values from the instrument were used in the statistical analysis, but it is not known if these values are actual ly 0. This has the potential to skew the results indicating that chlorine is present in the water when indeed it is not. See Table 6 for the detection limits of all test kits used in this study.
35 Table 6 : Detection limits of tes t kits used in laboratory analysis. Parameter Detection Limit Total Chlorine 0.02 mg/L to 2.00 mg/L as Cl 2 Free Chlorine 0.02 mg/L to 2.00 mg/L as Cl 2 Iron 0.02 6.00 ppm Nitrate 0.02 3.00 ppm Sulfate 2 100 ppm Copper 0.02 6.00 ppm Lead 0.02 5.00 ppm Alkalinity 20 400 mg/L as CaCO 3 The t iming of sampling is another potential source of error. For example, i t wa s not known how recently the storage tank was filled from the municipal wate r supply prior to sampling. Agitation of settled part icles and microbes may occur during filling and this has been shown to produce significantly higher microbial counts in smaller water storage containers (Roberts et al. 2001).
36 RESULTS Elevated Storage Tank Types A general survey of the elevated storage tanks present in the Tiquipaya Noreste community found 145 elevated storage tanks of which 56 (38%) are polyethylene tanks, 50 (34%) are fiberglass tanks and 39 (27%) are fiber cement tanks in the area. Figure 8 shows the locations and tank type of all the elevated storage tanks found within the study area The tanks most commonly used are fiber cement black po lyethylene gray polyethylene round fiberglass and sideways fiberglass Figure 9 provides photographs of each specific tank type For purpose s of analyzing the results, th e tanks have been grouped into three categories: polyethylene, fiberglass and fiber cement Polyethylene is a commonly used plastic that is composed of long ethylene chains. Thin fibers of glass are used to form fiberglass. F iber cement is a composite material that is composed of sand, cement, and cellulose fibers. Table 7 : Percentages of each tank type found within the Tiquipaya Noreste distribution system and of those included in the study. Storage Ta nk Type % of Tank Type Found in Community % of Tank Type Sampled Polyethylene 38% 43% Fiberglass 34% 30% Fiber Cement 27% 27%
37 Figure 8 : Locations of all elevated storage tanks within study area.
38 Figure 9 : Five most commonly found elevated storage tanks observed in Tiquipaya Noreste community S tarting from the top left and moving clockwise: gray polyethylene sideways fiberglass, fiber cement black polyethylene and round polyethylene Household Survey Over the course of one week in June, 2010, a total of 35 surveys were administered 37 household water storage tanks were sampled (two households had two tanks), and 14 household cisterns and 7 points along the dis tribution system were sampled. Fourteen of the survey respondents were the female head of household and 21 respondents were the male head of household. A total of 10 fiber cement tanks, 11 fiberglass (6 round, 5 sideways), and 16 polyethylene (9 black, 5 gray, and 2 red) were sampled. Locations of the elevated storage tanks, cisterns and points along the distribution system that were sampled for general analysis are shown in Figure 10
39 Figure 10 : Sample location maps in Tiquipaya Noreste community a) Locations of elevated storage tanks included in general study; b) Locations of underground cisterns and samples taken from distribution system. a b
40 Table 8 shows the age distribution of the tanks by tank materials. 32 out of 3 6 or 89 % of the tanks sampled in this study are 10 years old or younger G enerally storage tanks are sold with a 20 year guarantee. Table 8 : Age d istribution of elevated s torage t anks ; 37 tanks sampled. Tank Age Tank Material 0 3 4 10 11 15 16 20 Unknown Totals Fiber cement 1 6 1 1 1 10 Fiberglass 4 5 2 0 0 11 Polyethylene 8 8 0 0 0 16 Totals 13 19 3 1 1 Table 9 shows the frequency in which study participants (n = 37 ) clean their rooftop tanks. When asked about storage tank cleaning methods 19 study participants said they used bleach, detergent or di sinfectant to clean their elevated storage tank. When asked about treating the water from the rooftop tank before use 23 study participants said they boil their water, 1 participant said s/he disinfects the water in the elevated storage tank, 8 participan ts said they did not treat the water (including the school) and 2 participants gave no answer because they are owners of apartment buildings in which residents may use various point of use treatment techniques 1 1 due to difficulties encountered in reaching storage tanks. Instead disinfection may be occurring at point of use within the household and that there was a miscommunication in either the survey question or in the household answering the survey question. In addition, o ne participant treats water for all uses while all others who responded stated they treat the water and only use the treated water for drinking or cooking. treated for all uses because treatment method was boiling water and it is unlikely that boiled water for activities such as bathing or washing was used. On ce again there was some miscommunication in either the survey question or in the household answering the survey question.
41 Table 9 : Frequenc y of rooftop water storage tank c leaning ; 36 tanks sampled. How Often Rooftop Tank is Cleaned Every 2 Years Annually Biannually Every 3 Months Monthly Never Other 2 11 3 4 8 5 3 Households with no regular cleaning schedule Thirty six respondents reported they had access to water 24 hours a day and 36 respondent s said that they had access to water 7 days a week from the distribution system (different study participant was the lone individual who did not have access 7 days a week ) Because all residents are connected to the same distribution system these responses mostly likely reflect occasional cuts in service for maintenance and are not characteristic of the system w hich generally provides water 24 hours a day 7 days a week. Water Q uality Initial S creening Before analyzing results for correlations between parameters or for differences in water quality versus tank types, tank properties and user behaviors a stati stical analysis was performed to see if differences exist between the samples taken directly from the elevated storage tanks and samples taken from household taps. In order to determine this, a series general linear MANOVA was performed Table 10 provides a summary of results and Appendix I can be consulted for more complete results. These results show that the results for eac h parameter d o not vary significantly between samples taken directly from the storage tanks themselves and samples taken from taps fed by storage tanks
42 Table 10 : Results for MANOVA comparing water quality parameters for samples taken directly from elevated storage t ank s or from tap s The results show that water samples taken from taps do not differ significantly ( sig. < 0.05) from samples taken directly from storage tanks Multivariate Tests b Effect Value F Hypothesis df Error df Sig. Tank or Tap Pillai's Trace .290 .982 a 10.000 24.000 .484 Wilks' Lambda .710 .982 a 10.000 24.000 .484 Hotelling's Trace .409 .982 a 10.000 24.000 .484 Roy's Largest Root .409 .982 a 10.000 24.000 .484 a. Exact statistic b. Design: Intercept + Tank or Tap Additionally, the data were analyzed to see if there were any differences between parameters for samples taken directly from storage tanks or from taps with various water quality parameters ( Table 11 ). See Appendix I for more detailed results.
43 Table 11 : Results for tests of between subject effects using MANOVA. The results show that no significant differences exist for any of the parameters between samples directly from tanks and those from taps. Source Dependent Variable Type III Sum of Squares df Mean Square F Sig. Corrected Model dimension1 Temperature .959 a 1 .959 .165 .688 Conductivity .006 b 1 .006 .785 .382 TDS .003 c 1 .003 .898 .350 DO .378 d 1 .378 1.422 .242 pH .000 e 1 .000 .004 .951 Turbidity 244.647 f 1 244.647 2.572 .118 Total Coliforms 416344.281 g 1 416344.281 2.307 .138 E. coli 92.740 h 1 92.740 .040 .844 Total Chlorine .000 i 1 .000 1.184 .284 Free Chlorine .000 j 1 .000 2.049 .162 Tank or Tap dimension1 Temperature .959 1 .959 .165 .688 Conductivity .006 1 .006 .785 .382 TDS 003 1 .003 .898 .350 DO .378 1 .378 1.422 .242 pH .0 00 1 .000 .004 .951 Turbidity 244.647 1 244.647 2.572 .118 Total Coliforms 416344.281 1 416344.281 2.307 .138 E. coli 92.740 1 92.740 .040 .844 Total Chlorine .000 1 .000 1.184 .284 Free Chlorine .000 1 .000 2.049 .162 Data collected from the initial scree ning indicates that the physical, chemical and initial microbial water quality parameters do not vary significantly between tank types, underground cisterns and within the water distribution system. There are no statistically
44 F = 1.081, p = .398), although pH differs between plastic and fiberglass tanks at p Figure 11 and Figure 12 depict the results graphically, for specific values see Table 13 and for detailed statistical analysis see Appendix J Figure 11 : Results for conductivity, total dissolved solids, dissolved oxygen and pH for water storage tanks in Tiquipaya Noreste (Bolivia) Except for a few outliers, results for conductivity, TDS and DO show no difference between tank type. For pH, there is a difference between polyethylene and fiberglass tanks and betwee n fiberglass and fiber cement tanks ( p = 0.001 and 0.043 respectively ) ; but there is no difference between polyethylene and fiber cement tanks ( p = 0.722 )
45 Figure 12 : Results for turbidity, free chlorine, total c oliforms and E. coli for water storage tanks in Tiquipaya Noreste (Bolivia). The outliers for turbidity, total coliforms and E. coli results shown in Figure 12 correspond to storage tanks that are cleaned 2 times a year or less. The results for free chlorine were coded due to a majority of the results were below the detection limit of the instrument. See Table 12 for coding. Table 12 : Assigned values for coded free chlorine data. Instrument Reading Assigned Value Below Detection Limit 0 0.02 0.03 mg/L 1 > 0.03 mg/L 2
46 Table 13 : Overall physical and chemical water quality r esults for each water storage tank type in Tiquipaya Noreste (Bolivia) The listed Bolivian standards apply only to the source water. Looking at the Bolivian standards provided in Table 13 turbidity occasionally exceeds the Bolivian standards while on average total coliforms and E. coli counts exceed the Bolivian standards. Total and free chlorine levels are lower than called for by the Bolivian standards, however, the standards are for water leaving treatment facilities and are not
47 generally used for water at th e household level. Average free and total chlorine levels are ne ar the detection limits of the instrument, actual values may be lower. The results were further analyzed by separating data by cleaning frequency ( which was recorded for each tank during the household survey ) For the physical and chemical water parameters, the results do not vary significantly between cleaning frequencies (see Figure 13 ) Figure 13 : Results for conductivity, t otal dissolved solids, d issolved o xygen and pH b y c leaning frequency of elevated storage t anks in Tiquipaya Noreste (Bolivia) While no significant relationship was seen between cleaning schedule and bacterial growth, Figure 14 shows both total coliform and E. coli levels are lower for tanks cleaned more than 3 times a year than for tanks that are never cleaned.
48 Figure 14 : Results for t urbidity, free chlorine, total c oliforms and E. coli by cleaning f requency of elevated storage t anks in Tiquipaya Noreste (Bolivia). Figure 15 and Figure 16 show the results for physical, chemical and microbial water quality parameters for elevated storage tanks grouped by age. These results indicate that storage tank age is not an important fact or and that cleaning frequency may have a larger impact on water quality. This may be due to the limited number of storage tanks over 10 ye ars old that were sampled (n = 4 ) or that 3 of the 4 sto rage tanks over 10 years old were reported as being cleaned m onthly
49 Figure 15 : Results for conductivity, total dissolved solids, dissolved o xygen and pH by a ge of elevated storage t anks in Tiquipaya Noreste (Bolivia). In addition chlorine measurements were measured near the detection limits of the instrument; it is possible that free chlorine levels are actually lower than reported.
50 Figure 16 : Results for turbidity, free chlorine, total c oliforms and E. coli by a ge of elevated storage t anks in Tiquipaya Noreste (Bolivia). Randomized block ANOVAs were used to analyze the effect of tank ages and cleaning schedules on all tanks. Tanks were grouped by age, (0 3 years and >4 years) and cleaning schedule (3 or more times per year, 1 2 times per year and less than 1 time per year). See Table 14 for results. Table 14 : Results for randomized b lock ANOVA of various water qual ity parameters versus tank age and cleaning schedule in Tiquipaya Noreste (Bolivia). E. coli (CFU/100 mL) p value Tank Age (years) Mean -+ --------+ --------+ --------+ ------0 3 20.6667 ( -----------* -----------) ( -----------* -----------) -+ --------+ --------+ --------+ ------10 20 30 40 0.396
51 Table 14 Continued Cleaning Schedule Mean ----+ --------+ --------+ --------+ ---1 8.1667 ( --------* ---------) 2 33.1667 ( --------* ---------) 3 34.3333 ( ---------* --------) ----+ --------+ --------+ --------+ ---0 15 30 45 0.102 Total Coliforms (CFU/100 mL) p value Tank Age (years) Mean -------+ --------+ --------+ --------+ 0 3 136.222 ( -----------* -----------) -----------* -----------) -------+ --------+ --------+ --------+ 0 250 500 750 0.328 Cleaning Schedule Mean + --------+ --------+ --------+ --------1 36.000 ( ---------* ---------) 2 202.833 ( ----------* ---------) 3 565.333 ( ---------* ---------) + --------+ --------+ --------+ --------350 0 350 700 0.269 Free Chlorine (mg/L) p value Tank Age (years) Mean -------+ --------+ --------+ --------+ 0 3 0.0136111 ( -----------* ----------) ----------* ----------) -------+ --------+ --------+ --------+ 0.0060 0.0120 0.0180 0.0240 0.390 Cleaning Schedule Mean + --------+ --------+ --------+ -------1 0.0158333 ( -----------* -----------) 2 0.0087500 ( -----------* ----------) 3 0.0083333 ( -----------* -----------) + --------+ --------+ --------+ -------0.0000 0.0070 0.0140 0.0210 0.527 Turbidity (NTU) p value Tank Age (years)Mean -------+ --------+ --------+ --------+ 0 3 4.82222 ( -------------* --------------) --------------* -------------) -------+ --------+ --------+ --------+ 4.20 4.80 5.40 6.00 0.828 Cleaning Schedule Mean -------+ --------+ --------+ --------+ 1 3.73333 ( -------* --------) 2 6.11667 ( --------* --------) 3 4.36667 ( -------* --------) -------+ --------+ --------+ --------+ 3.6 4.8 6.0 7.2 0.055
52 Table 14 shows that w hile none of the results are significant at the 95% confidence level, (p value < 0.05) tanks which are cleaned 3 or more times per year have less E. coli than tanks that are cleaned less frequently ( p = 0.102). Similarly, turbidity is lower in tanks that are reported to be cleaned 3 or more times per year compared to tanks that are reported to be cleaned 1 2 times per year ( p = 0.055) although the difference is less for tanks that are cleaned less than once per year. Tank age appears to have very little eff ect on water quality for all parameters. Since chlorine levels are near the detection limits (0.02 mg/L) of the equipment, it is difficult to make any specific conclusions about the effects of tank age and cleaning schedule on chlorine concentrations based on the results. Based on the results from Table 14 one way ANOVAs were performed to reveal differences between E. coli and total coliform counts for various cleani ng schedules. The results are shown in Table 15 These results show that there is a significant difference between E. coli and total coliform counts in storage tanks t hat are cleaned three or more times per year compared to storage tanks that are cleaned less than once per year ( p = 0.006 and 0.033, respectively). The results also indicate at difference exists between storage tanks that are cleaned three or more times p er year and storage tanks that are cleaned once or twice per year, however the difference is not significant at the 95% confidence interval ( p = 0.151).
53 Table 15 : Results for one way ANOVAs comparing E. coli and total coliform counts for various cleaning schedules. E. coli p value Cleaning Schedule N Mean StDev --------+ --------+ --------+ --------+ -----------* -----------) 1 2 15 40.33 66.79 ( --------* ---------) --------+ --------+ --------+ --------+ 0 25 50 75 0.151 Cleaning Schedule N Mean StDev + --------+ --------+ --------+ --------9.70 15.39 ( ----* ----) < 1 3 42.17 12.49 ( ---------* ----------) + --------+ --------+ --------+ --------0 16 32 48 .006 Total Coliforms p value Cleaning Schedule N Mean StDev ------+ --------+ --------+ --------+ -------* -----) > 1 3 882.4 1331.3 ( ------------* -----------) ------+ --------+ --------+ --------+ -0 500 1000 1500 0.033 A series of r andomized block ANOVAs were used to analyze the data for differences in water quality while taking into account differences in tank ages and cleaning schedules. The data were divided into the following 6 groups text) and analyzed by tank type ( polyethylene fiberglass and fiber cement ): 1. Tanks age 0 3 years; cleaned >3 times per year 2. Tanks age >4 years; cleaned > 3 times per year 3. Tanks age 0 3 years; cleaned 1 2 times per year 4. Tanks age >4 years; cleaned 1 2 times per year 5. Tanks age 0 3 years; cleaned less than once per year 6. Tanks age 0 4 years; cleaned less than once per year
54 Due to sampling limitations, no fiber cement tanks were sampled for treatment 3 and values were interpolated based on values for group 2 and 4 Table 16 provides the number of samples available for each treatment. Table 16 : S ample sizes for treatments for randomized b lock ANOVA design. Polyethylene Fiberglass Fiber cement Cleaning 1 3 1 1 Age: 0 3 Cleaning 1 1 1 3 Age: >4 Cleaning 2 4 1 0 Age: 0 3 Cleaning 2 5 3 2 Age: >4 Cleaning 3 2 1 1 Age: 0 3 Cleaning 3 1 3 1 Age: >4 Table 17 show s the results the r andomized b lock ANOVAs; tank types are analyzed to see if tank age or cleaning schedule affects various water quality parameters Although none of the results are stati stically signifi cant (p < 0.05 ), the results in Table 17 do provide some insight as to what relationships may exist and where further research should focus.
55 Table 17 : Results for randomized b lock ANOVA of the effects o f tank age and cleaning schedule on various water parameters within different tank types in Tiquipaya Noreste (Bolivia) E. coli p value Tank Type Mean --------+ --------+ --------+ --------+ Polyethylene 40.8333 ( -------* -------) Fiberglass 20.1667 ( -------* --------) Fiber cement 14.6667 ( -------* -------) --------+ --------+ --------+ --------+ 15 30 45 60 0.082 Treatment Mean ------+ --------+ --------+ --------+ -1 1.3333 ( ------* -----) 2 15.0000 ( ------* ------) 3 38.6667 ( -----* ------) 4 27.6667 ( ------* ------) 5 22.0000 ( ------* ------) 6 46.6667 ( ------* ------) ------+ --------+ --------+ --------+ -0 25 50 75 0.127 Total Coliforms p value Tank Type Mean -------+ --------+ --------+ --------+ P olyeth ylene 204.000 ( ------------* -----------) Fiberglass 159.000 ( -----------* ------------) Fiber cement 441.167 ( ------------* ----------) -------+ --------+ --------+ --------+ 0 300 600 900 0.647 Treatment Mean + --------+ --------+ --------+ --------1 5.00 ( --------* --------) 2 67.00 ( -------* --------) 3 287.67 ( --------* -------) 4 118.00 ( --------* --------) 5 116.00 ( --------* --------) 6 1014.67 ( -------* --------) + --------+ --------+ --------+ --------600 0 600 1200 0.300 Free Chlorine p value Tank Type Mean ----+ --------+ --------+ --------+ ---P olyethylene 0.0116667 ( ---------* --------) Fiberglass 0.0166667 ( ---------* --------) Fiber cement 0.0045833 ( ---------* --------) ----+ --------+ --------+ --------+ ---0.0000 0.0080 0.0160 0.0240 0.230
56 Table 17 Continued Treatment Mean ------+ --------+ --------+ --------+ -1 0.0233333 ( -------* --------) 2 0.0083333 ( --------* --------) 3 0.0041667 ( -------* --------) 4 0.0133333 ( --------* --------) 5 0.0133333 ( --------* --------) 6 0.0033333 ( --------* --------) ------+ --------+ --------+ --------+ -0.000 0.012 0.024 0.036 0.346 Turbidity p value Tank Type Mean + --------+ --------+ --------+ --------P olyethylene 4.48333 ( ----------* ----------) Fiberglass 5.56667 ( ----------* ----------) Fiber cement 4.16667 ( ----------* ----------) + --------+ --------+ --------+ --------3.0 4.0 5.0 6.0 0.331 Treatment Mean + --------+ --------+ --------+ --------1 3.80000 ( ------* ------) 2 3.66667 ( ------* -------) 3 6.70000 ( ------* -------) 4 5.53333 ( -------* ------) 5 3.96667 ( ------* ------) 6 4.76667 ( -------* ------) + --------+ --------+ --------+ --------2.0 4.0 6.0 8.0 0.238 The results show that at the 90% confidence level polyethylene tanks have higher E. coli values than fiberglass and fiber cement tanks ( p = 0.082) Treatment type also appears to have an effect on E. coli growth within the tank although not statistically significant, ( p = 0.127) showing that tanks aged 0 3 years that are cleaned 3 or more times a year (Treatment 1) have less E. coli compared to tanks that are 4 years old or ol der and cleaned less frequently. Based on the results from Table 17 one way ANOVAs were performed to show more specifically the differences in E. coli counts between storage tank types and treatments. Table 18 suggests that difference for E. coli counts between storage tank types exist, however the differences are not statistically significant at the 95% confidence interval. The results shown in Table 18 also indicate that treatments do effect E. coli counts,
57 although from these results it is not clear how great of an affect cleaning schedule or tank age have individually. Table 18 : One way ANOVAs for E. coli co mparing storage tank types and treatments. E. coli p value Tank Type Mean -+ --------+ --------+ --------+ ------Polyethylene 40.8333 ( ----------* ---------) Fiber Cement 14.6667 ( ---------* ----------) -+ --------+ --------+ --------+ ------0 16 32 48 0.098 Tank Type Mean -------+ --------+ --------+ --------+ Polyethylene 40.8333 ( ----------* ----------) Fiberglass 20.1667 ( ----------* -----------) --------+ --------+ --------+ --------+ 15 30 45 60 0.170 Treatment Mean StDev --------+ --------+ --------+ --------+ 1 1.33 2.31 ( -------* --------) 6 46.67 24.58 ( --------* -------) --------+ --------+ --------+ --------+ 0 30 60 90 0.034 Water Quality In D epth Analysis In depth analysis of water quality included measuring iron, sulfate and nitrate levels in 11 tanks, 4 cisterns and 2 locations within the distribution system. These chemical parameters did not vary significantly between the tank types (see F igure 17 ) Iron is however present in the distribution system in higher concentrations than what was found in the cisterns ( p = 0.042) and in tanks ( p = 0.115).
58 F igure 17 : Results of in depth analysis of iron sulfate and nitrate levels in different storage tank types as well as within the distribution system in Tiquipaya Noreste (Bolivia). Microbial Results E. coli results from samples taken from various tank types as well as the distribution system are presented in Figure 18 All samples analyzed from the distri bution system meet Bolivian standards ( 0 CFU/ mL ) except for two samples. One of these samples was taken from the point furthest from the treatment plant and the other was after water service had been cut off 2 and most likely does not accurately represent t rue water quality at this location. All samples obtained from household storage tank s (and all tank types ) 2 Service was cut off in a section of the distribution system during sampling one morning. This disconnection of service is not beli eved to have affected results because samples were taken from storage tanks and cisterns in other parts of the distribution system. Also, storage tanks and cisterns were at or near storage capacity at time of sampling indicating that the cut in service had not significantly impacted water supplies.
59 had measureable E. coli values above Bolivian standards (0 CFU/mL) Round Fiberglass storage tanks appear to have the most samples above Bolivian standards with over 70% of samples failing to meet water quality standards for E. coli ( Table 19) Figure 18 : Histogram of E. coli c ounts Includes initial and in depth water analysis from elevated storage tanks, cisterns and the water distribution system in Tiquipaya Noreste (Bolivia) Table 19 : Percent of samples that exceed the Bolivian water quality s tandards for E. coli (0.0 CFU/mL) Tank Type Distribution System (n = 7) Black Poly (n = 10) Gray Poly (n = 5) Round Fiberglass (n = 5) Sideways Fiberglass (n = 5) Fiber cement (n = 9) Cistern (n = 13) 33.3 % 57.1 % 71.4 % 28.6 % 54.5 % 42.9 % 22.2 %
60 In addition to testing for total coliforms and E. coli a subset of samples w as also tested for iron related bacteria, heterotrophic aerobic bacteria and slime forming bacteria ( Table 20 ) All samples taken from the distribu tion system, cisterns and storage tanks were positive for iron related bacteria suggesting widespread prevalence of these bacteria in the distribution system All cisterns tested positive for all three types of bacteria. A sample taken of effluent water from the treatment plant tested negative for all three types of bacteria Table 20 : BART test results for three different microbial indicators reported as percent of positive tests recorded for each tank type. Iron Related Bacteria Heterotrophic Aerobic Bacteria Slime Forming Bacteria Polyethylene (n = 5) 100% 40% 80% Fiberglass (n = 4) 100% 75% 75% Fiber cement (n = 2) 100% 0% 100 % Cistern (n = 4) 100% 100% 100% System (n = 2) 100% 0% 50% Treatment Plant (n = 1) 0% 0% 0% Figure 19 and Figure 20 show that there is n o observable spatial correlation found for the iron related bacteria, heterotrophic aerobic or slime forming bacteria ( p = 0.245, 0.847, and 0.934 respectively ). This indicates that while the distribution system may be responsible for transporting the bacteria to the household, the cisterns and ele vated storage tan ks are providing habitat for bacteria to grow th This idea is supported by the
61 lower prevalence of heterotrophic aerobic and slime forming bacteria found in the distribution system.
62 Figure 19 : Levels of heterotrophic aerobic and slime forming bacteria measured in distribution system and household cistern s and water storage tanks in Tiquipaya Noreste (Bolivia).
63 Figure 20 : Levels of iron related bacteria measured in distribution system and household cistern s and water storage tanks in Tiquipaya Noreste (Bolivia) Temperature Study Figure 21 and Figure 22 show the results from the temperature study. Temperatures were greatest and had the highest variability in the black polyethylene tank; temperatures were lowest and had the lowest variability in the fiberglass tank.
64 Temperatures in all three tanks were greater than 15 C indicating that significant bacteria growth is possible (LeChevallier et al ., 1996) Figure 21 : Water t emperature within three t ypes of elevated storage t ank s in Tiquipaya Noreste (Bolivia) Figure 21 shows that w ater temperature in the black polyethylene tank peaks earlier than the other two tanks due to shading of the black polyethylene tank around 14:30 while the other 2 tanks remained in direct sunlight until sunset. 14.00 19.00 24.00 29.00 34.00 39.00 6:00 7:00 8:00 9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 Temperature ( C) Time Round Fiberglass Fiber Cement Black Polyethylene
65 Table 21 : Maximum and minimum water t emperatures (C) recorded in elevated storage tanks in Tiquipaya Noreste (Bolivia) Fiberglass (n = 1) Fiber cement (n = 1) Black Polyethylene (n = 1) Maximum Water Temperature (C) 19.83 22.40 33.70 Minimum Water Temperature (C) 15.18 17.50 23.10 Difference (C) Between Max and Min Temperatures 4.65 4.90 10.60 Temperatures in the black polyethylene tank were greater than the ambient air temperature during the entire measurement period, shown by the positive values in Figure 22 Both the fiberglass and fiber cement tank had temperatures greater than the ambient air temperature in the morning, but had cooler water temperatures during the days as shown by th e negative values in Figure 22 Figure 22 : Difference between ambient air t emperature and stored water t emperature in storage tanks in Tiquipaya Noreste (Bolivia) 15 10 5 0 5 10 15 0:00 4:00 8:00 12:00 16:00 20:00 Water Temp Air Temp ( C) Time Fiberglass Fiber Cement Black Polyethylene
66 One i mplication of the warm water temperatures found in all elevated storage tanks, but especially in the black polyethylene tank is that there is the potential for increased bacterial growth. The climate in Cochabamba (11 km east of Tiquipaya) is moderate with average monthly temperatures between 13 C and 19 C (climate zone.com). The average temperature for August, when the temperature study took place, is 16 C. This implies that the results o f this temperature study are representative of year round water temperatures found inside the storage tanks. Effect of Residence Time Water s amples analyzed from treatment plant, locations within the distribution system, cistern and storage tanks show a loss of chlorine residual (almost immediately), an increase in total coliforms and E. coli and an increase in temperature as the water travels from the treatment plant to the household cisterns and storage tanks.
67 Figure 23 : Water quality changes as water travels from the treatment plant through the system to household cisterns and storage tanks. a) Temperature (C); b) Free chlorine (mg/L Cl 2 ); c) Total coliforms and E. coli (CFU/100 mL) As shown by p values less than 0.05 in Table 22 s ignificant differences in E. coli counts can be found between water from the distribution system and cisterns and between the distributio n system and storage tanks. For total coliforms, significant differences can be found between cisterns and storage tanks and between the distri bution system and storage tanks a b c
68 Table 22 : P values for two tail independent t tests c omparing E. coli and total coliform counts within the distribution system, cisterns and elevated storage tanks in Tiquipaya Noreste (Bolivia) Pair t test E. coli Total Coliforms System vs. Cistern 0.026 0.548 Cistern vs. Tank 0.964 0.020 System vs. Tank 0.049 0.024 Treatment Plant and Wells Analysis of water samples from the Tiquipaya Noreste water treatment plant show ed that treatment was sufficient to inactivate bacteria in the drinking water supply leaving the treatment plant Free chlorine was measured at 0.47 mg/L Cl 2 in the effluent water from the treatment plant Total coliforms were detected in Well 2 (534 CFU/100 mL) but were not detected in Well 1. Neither E. coli nor total coliforms were detected at locations in the dist ribution system near the respective wells. Well water is not chlorinated and low free chlorine levels were detected in the water at locations in the distribution system near the wells (0.04 mg/L Cl 2 for Well 1 and 0.05 mg/L Cl 2 for Well 2).
69 DISCUSSIO N This study found evidence of microbiological contamination of the potable water supply in Tiquipaya Noreste (Bolivia) that could potentially have negative health consequences for users. Based on previous studies p otential sources of the contamination include: 1) the addition of untreated well water, 2) leakages with in the distribution system, 3) inadequate treatment of source wa ter, 4) long residence times, 5) elevated water temperature and 6) low chlorine residual The addition of u ntreated well water creates an additional chlorine demand thereby lowering the amount of chlorine in the water that would otherwise provide protection against bacterial re growth. The existence of leakages in the distribution system were not detected durin g this study, however, extensive testing of the system was not done. Leakages could potentially allow contaminants to enter the distribution system. Water leaving the Tiquipaya Noreste treatment facility meets Bolivian water quality standards, therefore in adequate treatment is not believed to be responsible for the increased bacterial growth found between water in the distribution system and water in household cisterns and elevated storage tanks ( p = 0.026 and 0.049 for E. coli respectively). Multiple st udies have shown that increases in storage time lead to decreases in water quality (Evison and Sunna, 2001; Roberts et al., 2001; Agard et al., 2002; Tokajian and Hashwa, 2003). While this study did not directly measure residence time, by storing water at the household level residents are increasing water residence time prior to use This study also found that water temperature increases as the water travels from the treatment plant through the distribution system to the household cisterns and finally to
70 th e elevated storage tanks. This result is supported by studies that suggest that i n countries where access to water is unreliable the problem of microbial re growth is intensified by long water storage times (Evison and Sunna, 2001). This study also found t hat the water temperature inside the elevated storage tanks is above the threshold level of 15 C cited by other studies as causing increased microbial growth (Fransolet et al., 1985; Donlan and Pipes, 1988; Smith et al., 1989; Donlan et al., 1994 From Le Chevallier et al., 1996). Low to no chlorine residual detected in the water from this study may be allowing microbes to overcome the initial sho ck of chlorination and to grow. This obser ved increase in microbial growth also corresponds to an increase in water temperature as the water moves from the source to h ousehold water storage tanks. Long retention time s, low or no residual chlorine and high water temperatures within the household storage tank are found to increase the likelihood of m icrobial growth (Schoenen, 1990; Schoenen and Scholer, 1985; LeChevallier et al. 1981; Schoenan and Dott, 1977; Grabow et al. 1975). Previous studies have shown that storage tank materials do not contribute significantly to differences in microbial wat er quality of stored water (Evison and Sunna 2001; Tokajian and Haswa, 200 3 ). This study found, however that there may be a difference in microbial water quality between polyethylene storage tanks and fiberglass and fiber cement tanks ( p = 0.082). However physical and chemical water parameters were not found to differ significantly between the storage tank types. One possible cause for the difference in microbial water quality observed in different storage tank types may be water temperature inside the storage tanks. A longer duration study that measured water temperature in three representative storage tank types found that water temperatures inside black polyethylene tanks reach upwards of 34 C as
71 opposed to 20 C and 23 C in fiberglass and fiber cement tanks respectively Increased microbial growth has previously been documented in water with temperatures exceeding 15 C (Donlan and Pipes, 1988; Fransolet et al., 1989; Smith et al., 1989; Donlan et al., 1994 From LeChevallier et al., 1996). The temperatures found in three different storage tank types indicate the potential for increased bacterial growth which is a health concern because e ven low levels of bacterial growth have the potential to cause illness in users (WHO, 2006). This stud y also showed that water temperature and total coliforms and E. coli counts increase d as the water travels from the treatment plant through the distribution system to household cisterns and elevated storage tanks This result agrees with other studies that have shown increased microbial growth as residence time increases (Evison and Sunna, 2001; Roberts et al., 2001; Agard et al., 2002; Tokajian and Hashwa, 200 3 ). Storage tank cleaning frequency also appears to impact the microbial water quality of the s tored water. Although not statistically significant s torage tanks that are reported to be cleaned 3 or more times per year have less E. coli than tanks cleaned less frequ ently ( p = 0.102). Additionally, no correlation between storage tank age and E. coli or total coliform counts was found indicating that storage tank age does not significantly impact water quality. This study encountered storage tanks that were over 10 years old, but were cleaned monthly and as a result no coliforms were detected in the st ored water. According to a report released in 1996, 72.4% of water distribution systems in Bolivia practice disinfection (Espana et al. 1996). However, this study has found that the chlorine residual present in water that reaches the household to be at or below the
72 analytical detection level of 0.02 mg/L indicating that although chlorine is added to the water supply it is not added in sufficient quantities to provide users with protec tion against pathogens Since sampling is usually done immediately aft er treatment, the report may be misleading about the safety of potable water supplies in Bolivia. One study found significant growth of total coliforms in waters where the free chlorine concentrations were less than 0.2 mg/L (LeChevallier et al., 1996) I n the Tiquip aya Noreste distribution system free chlorine levels that are one tenth of that are commonly found in the system, cisterns and storage tanks A lack of free chlorine in the supply water may also be an indication that contaminants are entering the system after treatment For example, a study by Agard et al., (2002) found post treatment contamination to be the cause of microbial contamination of the drinking water supply. The addition of untreated well water being blended into the Tiquipaya Nore ste system may also be causing the decrease in chlorine residual into the system due to reactions of the chlorine with the additional microbes and other compounds introduced into the system. Studies have shown that the addition of untreated water into a di stribution system reduces chlorine residuals and increases the likelihood of illness in consumers ( Ford, 1999 ; Craun and Calderon, 2001 ). Community Perceptions During m ultiple instances during sample collection, the investigators were told by residents that the water provided by the system was contaminated by the time it reached their homes. While this may be the case during different parts of the year, the investigators did not find conclusive evidence to confirm these claims. Conta minants may be entering the distribution system or the cisterns and storage tanks may be seeding the influent water, either way it appears that the cisterns and storage tanks are providing habitat for bacterial growth Many community members also did not
73 a ppear to understand the connection between not cleaning their storage tank and reduced water quality.
74 CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE RESEARCH The objective s for this study was to look at physical, chemical, and microbial water quality inside household storage tanks commonly found in the developing world and to document water quality changes as the water travels from the source to the user Few studies have lo oked at microbial water quality in household elevated stora ge tanks in laboratory settings but this author was unable to find field studies concerning physical, chemical and microbial water quality in ele vated storage tanks Studies done in the US and othe r developed countries have looked at physical, chemical and microbial water quality but few studies measuring more than microbial water quality have been done in developing countries. The first hypothesis that this study investigates i s that tank mater ial impacts water quality of water inside household storage tanks. This study found that the E. coli was present in higher concentrations inside polyethylene storage tanks compared to fiberglass and fiber cement storage tanks ( p = 0.082). Physical and chem ical water quality parameters were not found to vary significantly between storage tank types. The second hypothesis is that the water temperature insi de storage tanks a ffects water quality. This study found that temperature was highest in black polyethy lene storage ta nks and that temperatures in each of the storage tank types investigated reached levels previously shown to induce increase d bacterial growth and that polyethylene tanks had higher E. coli counts ( p = 0.082).
75 The third hypothesis is that storage tank use factors also affect water quality. This study found that storage tanks cleaned 3 or more times per year had lower E. coli counts and turbidity than storage tanks cleaned less frequently ( p = 0.102 and 0.055, respectively). However, t ank age was not found to have a significant difference in water quality indicating that maintenance (i.e. cleaning) is more important to water quality. Additionally, this study provided evidence that as the water travels from the treatment plant through the distribution system to elevated storage tanks that water E. coli and total coliform counts increase ( p = 0.049 and 0.024, respectively) as does temperature. Current ly, g uidelines for water quality are for source water/water leaving treatment faciliti es and not at the point of consumption. Evidence presented in this study as well as by other researcher s has shown that the re is potential for contamination of water supplies during transport from t he source/treatment to occur in the distribution system an d during storage and that the potential for illness exists. Generally speaking, the risk for developing waterborne illness is relatively unknown since the water quality of consumed water is often unknown. Based on the results of this study, it is recomme nded that homeowners discontinue their use of cisterns and storage tanks. Water service is provided 24 hours a day every day of the week thereby negating the necessity for storage in this instance. For communities where service is intermittent and water st orage is necessary, it is recommended that elevated storage tank owners clean their tanks 3 or more times per year This study results also suggest that the age of the elevated storage tank is not as important as maintenance (cleaning) on water quality. Also, when cost is not an issue fiberglass and fiber cement storage tanks are preferred over polyethylene storage tanks because of
76 lower water temperature in the fiberglass and fiber cement tanks In instances where polyethylene storage tanks are used, the y should be sited in shady areas to mitigate increases in water temperature. Additionally, it is recommended that the well water is chlorinated in the Tiquipaya Noreste distribution system to increase chlorine residual in order to provide more protection o f users against waterborne diseases. Further research into the effects of tank material on water quality could look at water temperatures inside the elevated storage tanks to find more conclusive evidence linking increased microbial growth to temperature This study provides a snapshot of the water quality inside elevated storage tanks, but more research should b e done to investigate seasonal e ffects. More research into the chlorine residual levels in water distribution systems that use chlorine for dis infection since this study found that chlorine levels were not sufficient at preventing microbial growth. Although at least 72% of water distribution systems in Bolivia chlorinate their potable water supplies, chlorine residuals may be too low to prevent m icrobial growth resulting which could potentially lead to illness in users. Results from the bacteria study show that numerous bacteria are present in the water in the distribution system, cisterns and elevated storage tanks. Further research could att empt to identify more specifically what bacterial species are present and evaluate the potential health concerns.
77 REFERENCES Agard, L., Alexander, C., Green, S., Jackson, M., Patel, S., & Adesiyun, A. 2002 Microbial quality of water supply to an urb an community in Trinidad. Journal of Food Protection, Vol 65 (8 ) pp 1297 1303. Allen, M.J., Edberg, S.C., Reasoner, D.J. 2004. Heterotrophic plate count bacteria what is their significance in drinking water? International Journal of Food Microbiology Vo l 92 pp 265 274. Arens P., Tuschewitzki G., Wollman M., Follner H. and Jacobi H. 1995. Indicators for microbiologically induced corrosion of copper pipes in a cold water plumbing system. Zentralblatt fuer Hygiene und Umweltmedizin Vol 196 pp 444 454. Basualdo, J., Pezzani, B., De Luca, M., Cordoba, A. & Apezteguia, M. 2000 Screening the municipal water system of La Plata, Argentina, for human intestinal parasites. International Journal of Hygiene, Environment and Health Vol 203(2) pp 177 182. Benjamin, M. Water Chemistry Mc Graw Hill, Columbus, OH, 2002. Berkman, D. S., Lescano, A. G. and Gilman, R. H. 2002. Effects of stunting, diarrhoeal disease and parasitic infection during infancy on cognition in late childhood: a follow up study. The Lan cet Vol 359(9306) pp 564 571. Boulay N. and Edwards M. 2001. Role of temperature, chlorine and organic matter in copper corrosion by product release in soft water. Water Research Vol 35(3) pp 683 690. Brick, T., Primrose, B., Chandrasekhar, R., Ro y, S., Muliyil, J., & Kang, G. 2004 Water contamination in urban south India: household storage practices and their implications for water safety and enteric infections. International Journal of Hygiene and Environmental Health, 207 (5) pp 473 480. Cardenas, V., Saad, C., Varona, M. and Linero, M. 1993. Waterborne cholera in Riohacha, Columbia, 1992. Bulletin of PAHO Vol 27(4) pp 313 330. CDC C enters for Disease Control and Prevention, 2010. Global Water Sanitation and Hygiene 1600 Clifton Rd. Atlanta, GA 30333, USA Accessed Online: August 2010, http://www.cdc.gov/healthywater/global/index.html Clasen, T. F., & Bastable, A. 2003. Faecal contamination of drinking water during collection and h ousehold storage: the need to extend protection to the point of use. Journal of Water and Health Vol 1 (3) pp 109 115.
78 Climate Zone.com. 2004. Accessed Online: October 2010, http://www.climate zone.com/climate/bolivia/celsius/cochabamba.htm Craun, G. F. and Calderon, R. L. 2001. Waterborne disease outbreaks caused by distribution system deficiencies. J ournal of the Am erican Water W or ks Assoc iation, Vol 93(9) pp 64 75. Craun G. F., Nwachuku, N., Calderon, R. L. & Craun, M. F. 2002. Outbreaks in drinking water systems, 1991 1998. J ournal of Environ mental Health Vol 65(1) pp 16 23. Dany, V., Visvanathan, C. & Thanh, N. C. 2000 Evaluation of water supply systems in Phnom Penh City: a review of the present status and future prospects. International Journal of Water Resource Development Vol 16(4) pp 677 689. Deb, B. C., Sircar, B. K., Sengupta, P. G., De, S. P. Mondal, S. K., Gupta, D. N., Saha, N. C., Ghosh, S., Mitra, U. and Pal, S. C. 1986. Studies on interventions to prevent El Tor cholera transmission in urban slums. Bulletin of the World Health Organization, Vol 64 pp 127 131. Donlan, R. M., and W. O. Pipes. 1988. Selected drinking water characteristics and attached microbial population density. Journal of the American Water Works Association, Vol 80(11) pp 70 76. Donlan, R. M., W. O. Pipes, and T. L. Yohe. 1994. Biofilm formation on cast iron substrat a in water distribution systems. Water Res earch Vol 28 pp 1497 1503. Edwards M., and Dudi A. 2004. Role of chlorine and chloramines in corrosion of lead bearing plumbing materials. Journal of the American Water Works Association Vol 96(10) pp 64 82. Eg orov, A.; Ford, T.; Tereschenko, A., Drizhd, N.; Segedevich, I.; Fourman, V., 2002. Deterioration of drinking water quality in the distribution system and gastrointestinal morbidity in a Russian city. International Journal of Environmental Health Research, Vol 12 pp 221 223. Espana et. al. 1996. Reporte de evaluacion de los sistemas de agua potable y tratamiento de agua residual de Bolivia Universidad Tecnologia de Bolivia, La Paz. Evison, L.; Sunna, N., 2001. Microbial re growth in household water storage tanks. Journal of the American Water Works Association, Vol 93 pp 85 94. Ford, T. E. 1999. Microbiological safety of drinking water: United States and global perspectives. Environ mental and Health Perspect ive, Vol 107(Suppl. 1) pp 191 206. Fransolet, G., G. Villers, and W. J. Masschelein. 1985. Influence of temperature on bacterial development in waters. Ozone Sci ence, Vol 7 pp 205 227. Gapminder.org. Accessed Online April 2010. Visualization from Gapminder World powered by Trendalyzer from www.gapminder.org.
79 Gaytan, M., Castro, T., Bonilla, P., Lugo, A. & Vilaclara, G. 1997 Preliminary study of selected drinking water samples in Mexico City Revista Internacional de Contaminacin Ambiental, Vol 13(2) pp 73 78. Geldenhuys, J. 1995 Chlo ramination to preserve microbiological quality: experience at Rand Water. Water Supply, Vol 13(2) pp 31 3 316. Gleeson, C. and Gray, N. 1997. The Coliform Index and Waterborne Disease 1 st edn. E FN Spon Ltd. London, UK. Grabow, W. et al ., 1975. Behaviour in a river and of coliform bacteria with transferable or nontransferable drug resi stance. Water Res earch Vol 9(9) p p 777. Gundry, S.; Wright J.; Conroy, R., 2004. A systematic review of the health outcomes related to household water quality in developing countries. Journal of Water and Health Vol 2 (1) pp 1 13 Hamer, D.H., Simon. J., Thea, D.M and Keusch, G. 1998. Childhood Diarrhoea in Sub Saharan Africa Harvard University, Harvard. p p 1 32. Ing. Hector Escalera Estrad, Gerente General, COAPAT. Personal Communication, July 2010. Ing. Mario Severiche, Former Mayor of Colcapirhua, Bolivia. Personal Communication, December 2009. Kerneis A., Nakache F., Dequin A. and Feinberg M. 1995. The effects of water residence time on the biological quality in a distribution network. Water Research, Vol 29 pp 1719 1727. LeChev all ier, M.W., Evans, T.M. and Seidler, R.J. 1981. Effect of turbidity on chlorination efficiency and bacterial persistence in drinking water. Applied and Envir onmental Microbiol ogy Vol 42 (7) p 159. LeChevallier, M.W. Welch, N.; Smith, D., 1996. Full scale studies of factors related to Coliform re growth in drinking wter. Applied and Environmental Microbiology Vol 62 (7) pp 2201 2211. LeChevallier M.W., Gullick, R. W., Kari m, M. No Date. The potential for health risks from intrusion of contaminants into the distribution system from pressure transients. Distribution system white paper from the American Water Works Association Service Company. Lee, E.J; and Schwab, K. 2005. Deficiencies in drinking water distribution systems in developing countries. Journal of Water and Health Vol 3(2) pp 109 127 Mermin, J. H.; Villar, R.; Carpenter, J.; Roberts, L.; Samaridden, A.; Gasanova, L.; Lomakina, S.; bopp, C.; Hutwagner, L.; Mea d, P.; Ross, B.; and Mintz, E.D., 1999. A massive epidemic of multidrug resistant Typhoid Fever in Tajikistan associated with consumption of municipal water. The Journal of Infectious diseases Vol 179 pp 1416 1422.
80 Nath KJ, Bloomfield SF, Jones M (2006). Household water storage, handling and point of use treatment. A review commissioned by IFH; Accessed Online: June 2010 http://www.ifh homehygiene.org Omisca, E. 2010. University of South Florida Doctoral Candidate. Personal Communication. May 2010. Payment, P., Richardson, L., Siemiatycki, J., Dewar, M., Edwards, M. and Franco, E. 1991. A randomized trial to evaluate the risks of gastrointestinal disease due to consumption of drinking water meeting curren t microbiological standards. American Journal of Public Health Vol 81 pp 703 708. Payment, P., Siemiatycki, J., Richardson, L., Gilles, R., Franco, E. and Prevot, M. 1997. A prospective epidemiological study of gastrointestinal health effects due to the consumption of drinking water. International Journal of Environmental Health Research Vol 7 pp 5 31. Pelletier, D. L., Frongillo, E. A., Schroeder, D. G., and Habicht, J. P.1994 A Methodology for Estimating the Contribution of Malnutrition to Mortality in Developing in Developing Countries. Journal of Nutrition, Vol 124(10 Suppl) pp 2106S 2122S. Quick, R. E., Venczel, L. V., Mintz, E. D., Soleto, L., Aparicio, J., Gironaz, M., Hutwagner, L., Greene, K., Bopp, C., Maloney, K., Chavez, D., Sobsey, M. and Tauxe, R. V 1999. Diarrhoea prevention in Bolivia through point of use water treatment and safe storage: a promising new strategy. Epidemiology and Infection Vol 122 pp 83 90. Quick RE, Kimura A, Thevos A et al. 2002 Diarrhea prevention through household level water disinfection and safe storage in Zambia. American Journal of Tropical Medicine and Hygiene Vol 66 pp 584 589. Rab, M. A., Bile, M. K., Mubarik, M. M., Asghar, H., Sami, Z., Siddiqi, S., Dil, A. S., Barzgar, M. A., Chaudhry, M. A. and Burney, M. I. 1997 Water borne hepatitis E virus epidemic in Islamabad, Pakistan: a common source outbreak traced to the malfunction of a modern water treatment plant. Am erican J ournal of Trop ical Med icine and Hyg iene Vol 57(2) pp 151 157. Roberts, L., Chartier, Y., Chartier, O., Malen ga, G., Toole, M., & Rodka, H. 2001 Keeping clean water clean in a Malawi refugee camp: a randomize d intervention trial. Bulletin of the World Health Organization, Vol 79 (4) pp 280 287. Schoenen, D. and Dott, W. 1997. Microbial Growth on an expansion joint of a reservoir for drinking water. Zentralblatt Fur Bakeriology Hygiene I. Abt. Orig. B ., Vol 165 p 464. Schoenen, D., 1990. Influence of Materials on the Microbiological Colonization of Drinking Water. Proc. Federation of Eu r opean Microbiological Societies Sym., Cranfield Institute of Technology, Bedford, United Kingdom. Schoenen, D.; Sc holer, H., 1985. Drinking Water Materials: Field Observations and Methods of Investigation Ellis Horwood Ltd., Chichester, England.
81 Schoenen, D.; Dott, W., 1977. Microbial Growth on an Expansion Joint of a Reservoir for Drinking Water. Zentralbl att fur B akteriology Hyg iene I. Abt. Orig. B ., Vo l 165 p 464. Semnenza, J.C.; Roberts, L.; Henderson, A.; Bogan, J. and Rubin, C.H., 1998. Water distribution system and diarrheal disease transmission: a case study in Uzbekistan. Am erican J ournal of Trop ical Med i cine, Vol 59 (6) pp 941 946. Smith, D. B., A. F. Hess, and D. Opheim. 1989. Control of distribution system coliform re growth pp. 1009 1029. In Proceedings of the AWWA Water Quality Technology Conference. American Water Works Association Denver. Sobsey, M. D., Handzel, T., and Venczel, L. 2003. Chlorination and safe storage of household drinking water in developing countries to reduce waterborne disease. Water Science and Technology, Vol 47 (3) pp 221 228. Sobsey, M. D. 2006 Drinking water and he alth research: a look to the future in the United States and globally. Journal of Water and Health, Vol 4( Suppl 1 ) pp 17 21. Swerdlow D.L., Mintz, E.D. Rodriguez, M. et al. 1992. Waterborne transmission of epidemic cholera in Trujillo, Peru: Lessons for a continent at risk Lancet Vol 340 pp 28 33. Tokajian S. and Hashwa F. 2003 Water quality problems associated with intermittent water supply. Water Science and Technology Vol 47(3) pp 229 234. Tokajian, S., and Hashwa, F., 2004. Microbial Quality and Genotypic Speciation of Heterotrophic bacteria isolated from potable water stored in household tanks. Water Quality Research Journal of Canada Vol 39 (1) pp 64 73. Trevett, A.; Carter, R.; Tyrrel S., 2004. Water quality deterioratio n: a study of household drinking water quality in rural Honduras. International Journal of E nvironmental Health Research. Vol 14 (4) pp 275 283. Trevett, A. F., Carter, R. C., & Tyrrel, S. F. 2005. Mechanisms leading to post supply water quality deteriora tion in rural Honduran communities. International Journal of Hygiene and Environmental Health, Vol 208 (3) pp 153 161. United Nations, 2010. Fact Sheet on Water and Sanitation. http://www.un .org/waterforlifedecade/factsheet.html Accessed Online 30 August 2010. Weather Underground. 2010. Accessed Online: August 2010 http://www.wunderground.com (WHO) World Health Organization, Geneva, 2004. The World Health Report 2004. (WHO), World Health Organization. 2006. Microbial aspects. In Guidelines for drinking water quality (3 ed., Vol. 1). Geneva: World Health Organization. Wright, J.; Gundry, S.; and Conroy, R., 2004. Household drinking water in d eveloping countries: a systematic review of microbiological contamination between source and point of use. Tropical Medicine and International Health Vol 9 (1) pp 106 117.
82 A PPENDICES
83 Appendix A : IRB Approval Letter June 10, 2010 Cynthia Schafer Civil and Environmental Engineering RE: Expedited Approval for Initial Review IRB#: Pro00001177 Title: Impact of Tank Material and Residence Time on Water Quality in Household Water Storage Systems in Cochabamba, Bolivia Dear Cynt hia Schafer: On 6/10/2010 the Institutional Review Board (IRB) reviewed and APPROVED the above referenced protocol Please note that your approval for this study will expire on 6 10 2011. Approved Items: Protocol Document(s): Study Protocol.docx 0.01 Consent/Assent Document(s): Waiver of Informed Consent Documentation for the Verbal English and Spanish Information Sheet/Consents It was the determination of the IRB that your study qualified for expedited review which includes activities that (1) present no more than minimal risk to human subjects, and (2) involve only procedures listed in one or more of the categories outlined below. The IRB may review research through the expedited review procedure authorized by 45CFR46.110 and 21 CFR 56.110. The
84 Appendix A Continued research proposed in this study is categorized under the following expedited review category: (7) Research on individual or group characteristics or behavior (including, but not limited to, research on perception, cognition, motivation, identity, language, communication, cultural beliefs or practices, and social behavior) or research employing survey, interview, oral history, focus group, program evaluation, human factors evaluation, or quality assurance methodologies. Please note, the informed consent/assent documents are valid during the period indicated by the official, IRB Approval stamp located on the form. Valid consent must be documented on a copy of the most recently IRB approved consent form. Your study qualifies for a waiver of the requirements for the documentation of informed consent as outlined in the federal regulations at 45CFR46.116 (d) which s tates that an IRB may approve a consent procedure which does not include, or which alters, some or all of the elements of informed consent, or waive the requirements to obtain informed consent provided the IRB finds and documents that (1) the research invo lves no more than minimal risk to the subjects; (2) the waiver or alteration will not adversely affect the rights and welfare of the subjects; (3) the research could not practicably be carried out without the waiver or alteration; and (4) whenever appropri ate, the subjects will be provided with additional pertinent information after participation. As the principal investigator of this study, it is your responsibility to conduct this study in accordance with IRB policies and procedures and as approved by th e IRB. Any changes to the approved research must be submitted to the IRB for review and approval by an amendment. We appreciate your dedication to the ethical conduct of human subject research at the University of South Florida and your continued commitme nt to human research protections. If you have any questions regarding this matter, please call 813 974 9343. Sincerely, Krista Kutash, PhD, Chairperson USF Institutional Review Board Cc: Various Menzel, CCRP USF IRB Professional Staff
85 Appendix B : Study Information Sheet for Survey Participants: Cochabamba, Bolivia Interviewer :________________ Date:______________ Survey #: ______________ You are being asked to participate in a research study described below: STUDY TITLE: Impact of Tank Material and Residence Time on Water Quality in Household Water Storage Systems in Cochabamba, Bolivia PERSON IN CHARGE: Dr. James Mihelcic PHONE NUMBER: 1 813 974 9896 EMAIL: firstname.lastname@example.org LOCAL CONTACT: Nathan Reents PHONE NUMBER: 5 91 722 38444 EMAIL: email@example.com PURPOSE: The purpose of this research is to understand how water quality within household storage tanks is affected by tank type and individual practices relating to storage tank usage. RISKS, BENEFITS, AND ALT ERNATIVES: There are no known risks or benefits to participation in this study. You have the alternative to choose not to participate. Your participation is voluntary. You may withdraw at any time without penalty. CONFIDENTIALITY: We will not collect any i nformation about you that could be used to identify you. The information we will collect will be combined with information from other
86 Appendix B Continued sources to meet the research objectives. Results of the study may be published, but will not contain any personally identifiable information about you. All information that you provide will be stored in a secure location in which only the primary investigator ha s access to. CONSENT: Your consent to participate in this study was obtained verbally. If you decide at any time that you want your information to be excluded from this research study, please contact any of the people listed above and provide your survey n umber (in the top right corner of the information sheet) so that your information can be removed from the study. QUESTIONS OR COMPLAINTS: If you have any concerns, do not hesitate to call the numbers listed above. If you have questions about your rights, g eneral questions, complaints, or issues as a person taking part in this study, call the Division of Research Integrity and Compliance of the University of South Florida at 1 813 974 9343.
87 Appendix C: Household Survey Questionnaire Date: ID: Interviewer: Interviewee: Male Female Impact of Tank Material and Residence Time on Water Quality in Household Water Storage Systems in Cochabamba, Bolivia Community Survey Demographic Information 1. What is your age? a. 18 35 b. 36 50 c. 50 65 d. Over 65 2. How many persons live in your household? __________ 3. How many are adults aged 18 and above? ____________ 4. How many children aged 5 17? _______________ 5. How many children under 5 years? ____________
88 Appendix C Continued 6. What is the occupation of male head of household? ___________ 7. What is the occupation of female head of household? ___________ Water Storage Tank Properties and Access to Water 8. What material is your tank made of? a. Plastic b. Metal (aluminum, tin) c. Fiber fiber cement d. Ceramic e. fiberglass f. Other __________________________ 9. What is the age of the tank? a. 0 3 years b. 4 10 years c. 11 15 years d. 16 20 years e. Older than 20 years
89 Appendix C Continued 10. How many days a week do you have access to piped running water? a. 1 day per week b. 2 days per week c. 3 4 days per week d. 5 6 days per week e. everyday 11. When you have access to piped running water, how long do you have access? a. All day b. 12 hours a day c. 6 11 hours a day d. 2 5 hours a day e. 1 hour a day 12. How often is your tank filled? a. Every day b. Every 2 da ys c. Every 3 5 days d. Every 6 7 days e. Less than once a week
90 Appendix C Continued 13. How is the tank filled? a. Pumped by a pipe network directly connected to municipal system i. (Do you share a pump? Yes _____ No _____) b. By municipal system using gravity c. Using a hose connected to an outside tap (Do you share a tap? Yes ______ No ______) d. Other __________________________________ Household Water Practices & Use 14. Is the water stored in the tank used for drinking water? a. Yes b. No 15. If no, what is the source for drinking w ater? _________________________ 16. What is the water from the storage tank used for? (circle all that apply) a. Drinking b. Washing food/cooking c. Hand washing d. Bathing e. Brushing teeth f. Clothes washing g. Other _____________
91 Appendix C Continued 17. What methods do you use to treat your water before use? a. Storage tank disinfection (What kind of disinfection? ___________________________) b. Point of use disinfection (What kind of disinfection? ____________________________) c. Boiling d. Filter (what type? ____________) e. Other __________ f. None 18. Is water treated for all uses or only for drinking? a. Yes b. No c. Treated for drinking and _________________ 19. Does someone disinfect the water in the storage tank? (If answer is NO, skip to Question 23) a. Yes b. No
92 Appendix C Continued 20. If yes to disinfection, how frequently is the water disinfected? a. Daily b. Weekly c. Monthly d. Every 6 months e. Annually f. Rarely g. Other _____________ 21. If yes to disinfection, when was the last time of disinfection? a. Within the last two weeks b. Within the last month c. Within the last six months d. Within the last year 22. Who is the main person responsible for disinfecting the water in the storage tank? a. Male head of household b. Female head of household c. Child d. Other __________
93 Appendix C Continued 23. In general, how frequently do you clean your storage tank? (If answer is NEVER, skip to Question 27) a. Never b. Daily c. Weekly d. Monthly e. Every 6 months f. Annually g. Other ____________ 24. What do you use to clean your storage tank? ________________________ 25. When was the last time the tank was cleaned? a. Within the last two weeks b. Within the last month c. Within the last six months d. Within the last year e. Other _____________ 26. Who is the main person responsible for cleaning the water storage tank? a. Male head of household b. Female head of household c. Child d. Other__________________ ________
94 Appendix C Continued Health Effects 27. In the last 2 weeks have you or someone in the household experienced an illness resulting from drinking the water in your storage tank? a. Yes b. No 28. If yes to the illness, who was ill? a. Male head of household b. Female head of household c. Child (Under 5: yes ____ no ____) d. Other ________________________ 29. If yes to the illness, what symptoms were present (circle all that apply)? a. Diarrhea b. Stomach pains/cramps c. Fever d. Nausea e. Skin rash/infection f. Loss of appetite g. Other_______________________ Thank you for your time. The survey is now complete.
95 Appendix D: Household Survey Questionnaire Responses Table D1 : Demographic information Gender of Respondents Female Male 14 21 # of People in Household 1 5 6 10 > 10 13 15 6 Ages of People in Households > 18 Years 5 17 years Under 5 194 42 23 Table D2 : Storage tank properties Tank Material Polyethylene Fiberglass Fiber Cement 16 11 9 Tank Age 0 3 Years 4 10 Years 11 15 years 16 20 Years Unknow n 13 19 3 1 1 # of Days per Week with Access to Water < 2 Days 3 4 Days 5 6 Days 7 Days 0 0 1 35 # of Hours per Day with Access to Water < 6 Hours 6 11 Hours 12 23 Hours 24 Hours 0 1 0 35 Method Used to Fill Tank Gravity Pump 9 27
96 Appendix D Continued Table D3 : Uses and practices of storage tank. Water from Tank is Used for Drinking Yes No NA 28 5 2 Other Sources of Drinking Water Bottle d Water SODIS Directly from Syste m 11 1 2 Other Uses of Water from Storage Tank Cooking Washing Clothes Bathing Brushing Teeth 27 34 34 30 Method of Treating Water Boil Water Disinfect With Chlorine No Treatment NA 23 1 8 2 Frequency of Cleaning Storage Tank Every 2 Years Annually Biannual ly Every 3 Months Monthly Never Other 2 11 3 4 8 5 3 What is Used to Clean Storage Tank Disinfectant Detergent Broom Brush Rag 7 10 9 15 5 Person Responsible for Cleaning Storage Tank Male Head of Household Female Head of Household Other 18 1 12 Table D4: Health effects of stored w ater Illness Experience Within Last 2 Weeks Yes No NA 3 27 6 Symptoms Diarrhea Stomach Ache Fever Nausea Headache Chills 2 3 1 2 1 1
97 Appendix E : Raw Data for Elevated Storage Tanks in Tiquipaya Noreste (Bolivia) Table E 1: Elevated storage tank location and material and age characteristics.
98 Appendix E Continued Table E 2: Physical chemical water quality data for elevated storage tanks.
99 Appendix E Continued Table E 3: Total and free chlorine water quality data for elevated storage tanks D etection level is 0.02 mg/L
100 Appendix E Continued Table E 4: Microbial water quality data for elevated storage tanks.
101 Appendix F : Raw Data for Underground Cisterns in Tiquipaya Noreste (Bolivia) Table F 1: Underground cistern characteristics. Table F 2: Physical chemical water quality data for underground cisterns.
102 Appendix F Continued Table F 3: Total and free chlorine water quality data for underground cisterns. Detection level is 0.02 mg/L. Table F 4: Microbial water quality data for underground cisterns.
103 Appendix G : Raw Data for the Tiquipaya Noreste (Bolivia) Water Distribution System Table G 1: Tiquipaya Noreste water distribution system characteristics. Table G 2: Physical chemical water quality data for the Tiqui paya Noreste water distribution system. Table G 3: Total and free chlorine water quality data for the Tiquipaya Noreste water distribution system Detection level is 0.02 mg/L.
104 Appendix G Continued Table G 4: Microbial water quality data for the Tiquipaya Noreste water distribution system.
105 Appendix H : Raw Data for the Tiquipaya Noreste (Bolivia) Water Treatment Plant Table H 1 : Tiquipaya Noreste water treatment plant c haracteristics. Table H 2: Physical chemical water quality data for the Tiquipaya Noreste water t reat ment plant.
106 Appendix H Continued Table H 3: Total and free chlorine water quality data for the Tiquipaya Noreste water treatment plant. Detection level is 0.02 mg/L. Table H 4: Microbial water quality data for the Tiquipaya Noreste water treatment plant.
107 Appendix I : Results for MANOVA Comparing Water Quality P arameters for Samples Taken D irectly from S torage Tanks with Those Taken from T aps Table I 1 : Multivariate tests for water quality parameters for samples taken directly from storage tanks compared to those taken from taps Effect Value F Hypothesis df Error df Sig. Intercept Pillai's Trace .999 4229.953 a 10.000 24.000 .000 Wilks' Lambda .001 4229.953 a 10.000 24.000 .000 Hotelling's Trace 1762.480 4229.953 a 10.000 24.000 .000 Roy's Largest Root 1762.480 4229.953 a 10.000 24.000 .000 TankorTap Pillai's Trace .290 .982 a 10.000 24.000 .484 Wilks' Lambda .710 .982 a 10.000 24.000 .484 Hotelling's Trace .409 .982 a 10.000 24.000 .484 Roy's Largest Root .409 .982 a 10.000 24.000 .484 Table I 2 : Tests of between subjects effects for water quality parameters for samples taken directly from storage tanks and those taken from taps Source Dependent Variable Type III Sum of Squares df Mean Square F Sig. Corrected Model dimension1 Temperature .959 a 1 .959 .165 .688 Conductivity .006 b 1 .006 .785 .382 TDS .003 c 1 .003 .898 .350 DO .378 d 1 .378 1.422 .242 pH .000 e 1 .000 .004 .951 Turbidity 244.647 f 1 244.647 2.572 .118 Total Coliforms 416344.281 g 1 416344.281 2.307 .138 E. coli 92.740 h 1 92.740 .040 .844 Total Chlorine .000 i 1 .000 1.184 .284 Free Chlorine .000 j 1 .000 2.049 .162
108 Appendix I Continued Table I2 Continued Intercept dimension1 Temperature 9620.269 1 9620.269 1650.834 .000 Condu ctivity 1.119 1 1.119 152.208 .000 TDS .538 1 .538 143.819 .000 DO 76 7.924 1 767.924 2889.429 .000 pH 1561.052 1 1561.052 32619.80 6 .000 Turbi dity 1739.403 1 1739.403 18.288 .000 Total Coliforms 1494181.007 1 1494181.007 8.278 .007 E. co li 21652.959 1 21652.959 9.226 .005 Total Chlorine .010 1 .010 27.318 .000 Free Chlorine .006 1 .006 43.895 .000 Tank or Tap dimension1 Temperature .959 1 .959 .165 .688 Condu ctivity .006 1 .006 .785 .382 TDS .003 1 .003 .898 .350 DO .3 78 1 .378 1.422 .242 pH .000 1 .000 .004 .951 Turbi dity 244.647 1 244.647 2.572 .118 Total Coliforms 416344.281 1 416344.281 2.307 .138 E. coli 92.740 1 92.740 .040 .844 Total Chlorine .000 1 .000 1.184 .284 Free Chlorine .000 1 .000 2.049 .162 Error d imension1 Temperature 192.308 33 5.828 Conductivity .243 33 .007 TD S .123 33 .004 DO 8.770 33 .266 pH 1.579 33 .048 Tu rbidity 3138.721 33 95.113 Total Coliforms 5956374.322 33 180496.192 E. coli 77450.511 33 2346.985 To tal Chlorine .012 33 .000 Free Chlorine .005 33 .000
109 A ppendix I Continued Table I2 Continued Total dimension1 Temperature 10547.668 35 Co nductivity 1.403 35 TD S .680 35 DO 822.090 35 pH 1673.454 35 Turbidity 4904.010 35 To tal Coliforms 7567820.210 35 E. coli 101516.280 35 Total Chlorine .022 35 Free Chlorine .011 35 Corr ected Total dimension1 Temperature 193.268 34 Conductivity .248 34 TDS .127 34 DO 9.148 34 pH 1.579 34 Turbidity 3383.367 34 Total Coliforms 6372718.603 34 E. coli 77543.251 34 Total Chlorine .013 34 Free Chlorine .005 34 a. R Squared = .005 (Adjusted R Squared = .025) b. R Squared = .023 (Adjusted R Squared = .006) c. R Squared = .026 (Adjusted R Squared = .003) d. R Squared = .041 (Adjusted R Squared = .012) e. R Squared = .000 (Adjusted R Squared = .030) f. R Squared = .072 (Adjusted R Squared = .044) g. R Squared = .065 (Adjusted R Squared = .037) h. R Squared = .001 (Adjusted R Squared = .029) i. R Squared = .035 (Adjusted R Squared = .005) j. R Squared = .058 (Adjusted R Squared = .030)
110 Appendix J : Results for MANOVA Comparing Water Quality Parameters for Each Tank Type (Polyethylene, Fiberglass, and Fiber C ement) Table J 1 : Multivariate tests for water quality parameters for each tank type. Effect Value F Hypothesis df Error df Sig. Intercept Pillai's Trace 1.000 4764.810 a 10.000 23.000 .000 Wilks' Lambda .000 4764.810 a 10.000 23.000 .000 Hotelling's Trace 2071.656 4764.810 a 10.000 23.000 .000 Roy's Largest Root 2071.656 4764.810 a 10.000 23.000 .000 Tank Type Pillai's Trace .621 1.081 20.000 48.000 .398 Wilks' Lambda .469 1.058 a 20.000 46.000 .421 Hotelling's Trace .940 1.034 20.000 44.000 .446 Roy's Largest Root .641 1.538 b 10.000 24.000 .186 Table J 2 : Tests of between subjects effects for water quality parameters for each tank type (polyethylene, fiberglass and fiber cement) Source Dependent Variable Type III Sum of Squares df Mean Square F Sig. Corrected Model dimension1 Temperature 11.413 a 2 5.706 1.004 .378 Conductivity .003 b 2 .001 .170 .844 TDS .003 c 2 .002 .452 .641 DO .893 d 2 .446 1.731 .193 pH .328 e 2 .164 4.187 .024 Turbidity 177.354 f 2 88.677 .885 .423 Total Coliforms 147746.201 g 2 73873.100 .380 .687 E. coli 2189.369 h 2 1094.684 .465 .632 Total Chlorine .001 i 2 .001 1.451 .249 Free Chlorine 9.613E 5 j 2 4.806E 5 .310 .735
111 Appendix J Continued Table J2 Continued Intercept dimension1 Temperature 9573.897 1 9573.897 1684.6 65 .000 Conductivity 1.108 1 1.108 144.26 2 .000 TDS .513 1 .513 133.01 1 .000 DO 77 6.498 1 776.498 3009.9 01 .000 pH 1579.958 1 1579.958 40387. 919 .000 Turbi dity 1648.871 1 1648.871 16.458 .000 Total Coliforms 1266567.432 1 1266567.432 6.511 .016 E. co li 19446.248 1 19446.248 8.258 .007 Total Chlorine .009 1 .009 25.243 .000 Free Chlorine .006 1 .006 35.865 .000 Tank Type dimension1 Temperature 11.413 2 5.706 1.004 .378 Condu ctivity .003 2 .001 .170 .844 TDS .003 2 .002 .452 .641 DO .8 93 2 .446 1.731 .193 pH .328 2 .164 4.187 .024 Turbi dity 177.354 2 88.677 .885 .423 Total Coliforms 147746.201 2 73873.100 .380 .687 E. co li 2189.369 2 1094.684 .465 .632 Total Chlorine .001 2 .001 1.451 .249 Free Chlorine 9.613E 5 2 4.806E 5 .310 .735 Error d imension1 Temperature 181.855 32 5.683 Co nductivity .246 32 .008 TD S .123 32 .004 DO 8.255 32 .258 pH 1.252 32 .039 Tu rbidity 3206.013 32 100.188 Total Coliforms 6224972.402 32 194530.388 E. coli 75353.883 32 2354.809 To tal Chlorine .011 32 .000 Free Chlorine .005 32 .000
112 Appe ndix J Continued Table J2 Continued Total dimension1 Temperature 10547.668 35 Conductivity 1.403 35 TD S .680 35 DO 822.090 35 pH 1673.454 35 Tu rbidity 4904.010 35 To tal Coliforms 7567820.210 35 E. coli 101516.280 35 To tal Chlorine .022 35 Free Chlorine .011 35 Corr ected Total dimension1 Temperature 193.268 34 Conductivity .248 34 TDS .127 34 DO 9.148 34 pH 1.579 34 Turbidity 3383.367 34 Total Coliforms 6372718.603 34 E. coli 77543.251 34 Total Chlorine .013 34 Free Chlorine .005 34 a. R Squared = .059 (Adjusted R Squared = .000) b. R Squared = .011 (Adjusted R Squared = .051) c. R Squared = .027 (Adjusted R Squared = .033) d. R Squared = .098 (Adjusted R Squared = .041) e. R Squared = .207 (Adjusted R Squared = .158) f. R Squared = .052 (Adjusted R Squared = .007) g. R Squared = .023 (Adjusted R Squared = .038) h. R Squared = .028 (Adjusted R Squared = .033) i. R Squared = .083 (Adjusted R Squared = .026) j. R Squared = .019 (Adjusted R Squared = .042)
113 Appendix J Continued Table J3: Multiple comparisons using MANOVA and the Tukey HSD test statistic Dependent Variable Tank Type Tank Type Mean Difference Std. Error Sig. 95% Confidence Interval Lower Bound Upper Bound dimension1 Temperature dimension2 1 dimension3 2 1.196625 .9609798 .436 1.164859 3.558109 3 1.083958 .9932906 .526 1.356926 3.524842 2 dimension3 1 1.196625 .9609798 .436 3.558109 1.164859 3 .112667 1.0953253 .994 2.804288 2.578955 3 dimension3 1 1.083958 .9932906 .526 3.524842 1.356926 2 .112667 1.0953253 .994 2.578955 2.804288 Conductivity dimension2 1 dimension3 2 .019875 .0353261 .841 .106684 .066934 3 .012819 .0365138 .934 .102547 .076909 2 dimension3 1 .019875 .0353261 .841 .066934 .106684 3 .007056 .0402647 .983 .091890 .106001 3 dimension3 1 .012819 .0365138 .934 .076909 .102547 2 .007056 .0402647 .983 .106001 .091890 TDS dimension2 1 dimension3 2 .021 .0250 .676 .040 .083 3 .002 .0259 .996 .066 .061 2 dimension3 1 .021 .0250 .676 .083 .040 3 .023 .0285 .695 .093 .047 3 dimension3 1 .002 .0259 .996 .061 .066 2 .023 .0285 .695 .047 .093 DO dimension2 1 dimension3 2 .3314 .20475 .253 .8345 .1718 3 .3077 .21163 .326 .8278 .2124 2 dimension3 1 .3314 .20475 .253 .1718 .8345 3 .0237 .23337 .994 .5498 .5971 3 dimension3 1 .3077 .21163 .326 .2124 .8278 2 .0237 .23337 .994 .5971 .5498 pH dimension2 1 dimension3 2 .2299 .07973 .019 .4258 .0339 3 .1074 .08241 .404 .3099 .0951 2 dimension3 1 .2299 .07973 .019 .0339 .4258 3 .1224 .09088 .380 .1009 .3458 3 dimension3 1 .1074 .08241 .404 .0951 .3099 2 .1224 .09088 .380 .3458 .1009
114 Appendix J Continued Table J3 Continued Turbidity dimension2 1 dimension3 2 1.438 4.0349 .933 11.353 8.478 3 5.515 4.1706 .393 15.764 4.733 2 dimension3 1 1.438 4.0349 .933 8.478 11.353 3 4.078 4.5990 .652 15.379 7.224 3 dimension3 1 5.515 4.1706 .393 4.733 15.764 2 4.078 4.5990 .652 7.224 15.379 Total Coliforms dimension2 1 dimension3 2 1.508750 177.7953 1.00 438.41800 435.40050 3 149.2298 183.7733 .698 600.82922 302.36950 2 dimension3 1 1.508750 177.7953 1.00 435.40050 438.41800 3 147.7211 202.6512 .748 645.71052 350.26830 3 dimension3 1 149.2298 183.7733 .698 302.36950 600.82922 2 147.7211 202.6512 .748 350.26830 645.71052 E. coli dimension2 1 dimension3 2 7.838750 19.56160 .916 40.231386 55.908886 3 19.47430 20.21932 .605 30.212082 69.160693 2 dimension3 1 7.838750 19.56160 .916 55.908886 40.231386 3 11.63555 22.29632 .861 43.154812 66.425923 3 dimension3 1 19.47430 20.21932 .605 69.160693 30.212082 2 11.63555 22.29632 .861 66.425923 43.154812 Total Chlorine dimension2 1 dimension3 2 .009188 .0076405 .460 .027963 .009588 3 .005313 .0078974 .781 .014094 .024719 2 dimension3 1 .009188 .0076405 .460 .009588 .027963 3 .014500 .0087087 .234 .006900 .035900 3 dimension3 1 .005313 .0078974 .781 .024719 .014094 2 .014500 .0087087 .234 .035900 .006900 Free Chlorine dimension2 1 dimension3 2 .001563 .0050178 .948 .013893 .010768 3 .002882 .0051865 .844 .009863 .015627 2 dimension3 1 .001563 .0050178 .948 .010768 .013893 3 .004444 .0057193 .720 .009610 .018499 3 dimension3 1 .002882 .0051865 .844 .015627 .009863 2 .004444 .0057193 .720 .018499 .009610 Based on observed means.
115 Appendix K : BART Results Table K 1 : Raw data for in depth microbial testing.
116 Appendix L : BART Test Information Sheets
117 Appendix L Continued
118 Appendix L Continued
ABOUT THE AUTHOR Cynthia A. Schafer received her B.S. in Environmental Engineering from Michigan Technological University, where she was awarded the 2008 Nicole Roth Award for Leadership in Environmental Sustainability. At Michigan Tech, she served as a member of the Envir onmental Sustainability Committee and worked for the Carbon Academic Quality Improvement Program to inventory and offset carbon dioxide emissions on the Michigan Tech campus. She spent 2006 and 2007 as a participant in the U.S. l Change Education Program Summer Undergraduate Research Experience, working with faculty and graduate students to study the role of biological soil crusts in alpine ecosystems and carbon dioxide flux over Lake Superior. While at the University of South Florida, Cynthia was the 2009 2010 Sustainability Fellow. During her spare time, Cynthia volunteered with the Gulf Coast Jewish Family Services helping refugee children and their families adjust to life in the US.