Water quality trends in Tampa Bay, Florida

Citation
Water quality trends in Tampa Bay, Florida

Material Information

Title:
Water quality trends in Tampa Bay, Florida
Creator:
Bendis, Brian
Place of Publication:
Tampa, Florida
Publisher:
University of South Florida
Publication Date:
Language:
English
Physical Description:
viii, 84 leaves : ill. (some col.) ; 29 cm.

Subjects

Subjects / Keywords:
Water quality -- Measurement -- Florida -- Tampa Bay ( lcsh )
Dissertation, Academic -- Marine science -- Masters -- USF ( FTS )

Notes

General Note:
Thesis (M.S.)--University of South Florida, 1999. Includes bibliographical references (leaves 66-68).

Record Information

Source Institution:
University of South Florida
Holding Location:
Universtity of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
027008401 ( ALEPH )
44852813 ( OCLC )
F51-00143 ( USFLDC DOI )
f51.143 ( USFLDC Handle )

Postcard Information

Format:
Book

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

WATER QUALITY TRENDS IN TAMPA BAY FLORIDA by r BRIAN BEN DIS A thesis submitted in partial fulfillment of the requirement for the degree of Master of Science Department of Marine Science University of South Florida December 1999 Major Professor : Gabriel A. Vargo Ph D

PAGE 2

Graduate School Univers i ty of South Florida Tampa, Florida CERTIFICATE OF APPROVAL Master's Thesis This is to certify that the Master's Thesis of BRIAN BENDIS with a major in Marine Science has been approved by the Examining Committee on October 27, 1999 as satisfactory for the thesis requirement for the Master of Science degree Examining Committee : ifajofProfessor: Gabriel A. l>'}rgo, Ph.D. Member : Kent A. Fanning, P!j;d:

PAGE 3

DEDICATION I wish to thank my parents Gerald and Louise Bendis for their continual support through the years as I have pursued my goals Their encouraging words are always a source of strength. I dedicate this thesis to my wife, Katrina. Her enduring faith in my abilities and support of my dreams have kept me going through what was at times an arduous journey I have truly been blessed

PAGE 4

ACKNOWLEDGEMENTS The author wishes to acknowledge the Tampa Bay National Estuary Program for financial support at the beginning of this project. Holly Greening, TBNEP facilitated obtaining the river flow data and has been a supporter of this study since the start I wish to thank Andy Squires Pinellas County Department of Environmental Management, for his computer training and support. Richard Boler of the Hillsborough County Environmental Protection Commission provided the data, without which all points are mute. I also wish to thank all the friends at USF Marine Science and the Florida Marine Research Institute that have helped in their various capacities Bill Richardson in particular, has been a great off i cemate with whom many late nights and weekends were shared I especially wish to thank Drs Carm Tomas and Karen Steidinger my supervisors at FMRI. This would not have been completed without their support I also wish to thank my committee for their individual and collective advice throughout the entire graduate school process. I am especially grateful to Dr Gabe Vargo for his infinite patience

PAGE 5

TABLE OF CONTENTS List of Tables ...... .. ................ ....... .... ....... .. ...... ... ... .... ........ ..... .. ................................... .. ii List of Figures ......... . ....... ... ..... ..... ...... .. .. ...... .. .. ... ...... ..... ........ .... .... . . .. .......... .. ........ iii Abstract ... ....... ... ...... .......... ... .... ..... .... ..... ............. .. .................. .. ..... ... ..... . .. ............... vi Introduction ...................... ... ... ....... .. ........ ..... .. ........... ...................... ...................... .. .... 1 Summary of TBNEP Technical Publication #07-92 ........ ... ....... ............. ........... .... .... 7 Objectives .. .............. .. ......... ...... ... . .. . ..... ... ................ .... .. . ..... ......... .... ........ .. .......... 15 Methods .................. ......... .... ... .. . ........... ............ . ...... ......... ......... ............... ... .... .... ..... 16 Results ............. ................ ........... .. . ............ .. ........... .. ........................... .. .. .............. 19 Raw Data ...... ............... .. ..... ... ................. ..... .... .... ... .. ............ ..... .... .... .. ... .............. 19 Cluster Analysis ... .. .. ...... .... . ....................................................... ........ ... .. ...... ... ... . 20 Climatological Monthly Averages ........ ....... ......... ................. ....... .. .. ...... .. ... ................ 24 Annual Averages ...... ... ..................... . ..... . ...... ...... .. ....... .. .... ................................ 34 Mixing Diagrams . .................. ... .... ...... ..... ...... ........... ........ .. ...... .... ...... ... ... ... ..... ....... 43 Discussion . ................ .... .................... .. ...................... .............. ... ...................... .... 47 Cluster Analysis .... . .... .... ..... ...... ....... ......... .... ............ .. ................. . .... ......... . . ........ 47 Monthly Averages .................. .... ........ .. .. .... ... .... ... .............. ..... .... ........ .............. .... ... 49 Annual Averages .............. .. ... .. .... .......... .. ..... .. ... ....... .............. ......... ................... .. 52 Spatial Gradient. ..... ... ........ . ....................... ... ... ... ...................................... ...... .... 57 Conclus i ons and Recommendations . .... ........... .. .... ........................ ... .... .......... .. .. ....... 63 References ........... ... ....... ......... ... .................. ........ ........... .... .... . ............. .... ... .... . ... 66 Appendices ....... ... ..... .............. .... ..... ............................................... ....................... ... 69 Appendix 1 Descriptive Statistics ..... ..... ..... .. ....................... .... ....... .... .... ....... ... ......... 70 Appendix 2 CMA Box Plots. .. ...... ..... ...... .... ....... ... .... .. ... ... . .. .. ... ....... .... .... ...... .... ...... 73

PAGE 6

LIST OF TABLES Table 1 Percent variance accounted for by the first three factors from each PCA .... .. 19 Table 2. Des criptive statistics of l arge clusters of Tampa Bay water quality stations based o n CMAs ... .. . . . . .... .. .... . .. .. . . . .. . .. .. . ..... ... . ... . .. .... . ..... . .... .. . . 25 Table 3. Descr i ptive statistics of small clusters of Tam pa Bay water quality stat i ons based on CMAs .. ..... . ... ....... . ..... ..... ... .. .. ...... . .. . .... .... . .. . . ...... .... .. . . 26 T able 4. Correlation matrix from PCA of CMAs .... ....... .. .... . . . .... .... . .. . ... .. .. ......... . 32 Tabl e 5 Correlation matrix from PCA of annual averages ... .. . .. . .... ............. ......... 40 T ab l e 6. Climatological monthly averages of s mall clusters .... ......... .... ... ....... . . .. .... 70 ii

PAGE 7

LIST OF FIGURES Figure 1. HCEPC water monitoring stations in Tampa Bay . . ........ .............................. 2 F i gure 2. Time series of total phosphorus at station 16 from 1975-1994 . . .. ... . .... ...... 8 F i gu r e 3 Time series of total phospho r us at stat i on 6 from 1975-1994 .... . . ........... .. .. 9 Figure 4. Annual ave r ages of total phospho r us from select stations in Middle Tampa Bay .. ..... . .............. ....... ....... ... ...... ............... .. .... ..... . . ........ 9 Figure 5. Annual averages of total phosphorus from select stat i ons i n Hillsborough Bay .. ..... ......... ... ... . . .. . .. .... ... ......... . . .... ... .... .. ... ........ ... .. 1 0 Figure 6 Annual averages of total phosphorus from select stations in Low e r Tampa Bay ..... ...... . ..... . . .. ... . .... ............... .... .. ...... ...... ... ... ... ...... ....... 10 Figure 7. T i me series of chlorophyll a at station 16 from 1975 1994 ... .... . ...... .... .... 11 Figu r e 8 Time series of chlorophyll a at stat i on 6 from 1975-1994 ... . . .. ...... .. .. . . . 11 F i gure 9. Annual averages of chlorophyll a from select stations in Hillsbo r ough Bay . 12 F i gure 10 Annual averages of chlorophyll a from s elect stations in Middle Tampa Bay .. ...... ..... ..... .. .. ... .... ........... ....... . ............ ... ... ... ................ ....... 12 F i gure 11. Total annual riv e r flow for the 4 major ri vers flowing into Tampa Bay ....... . . .. from 197 4 -94 ... . ....... . .......... .. ... ...... ... ... . .. . ..... . .... ....... . . .... ... .... .......... 13 Figure 12 Cluster analys i s tree diagram ... .... .. . ... ........ . ... .... . ...... ....... .. .. . ..... ..... ... 21 Figu r e 13 Map of la rge clusters ............ .... ... .. ........ . .. .. ................ ... . ...... .... ..... .. . 22 Figure 14 Map of sm all clusters ......... ... . .. . .................. .. .. .. .. ... .. ... . ............ . ........... 23 Figure 15 Climatological monthly averages of wat er temperatur e by small cluster . .... 27 Figure 16 Climatological monthly averages of dissolved oxygen by small cluster .... . 27 F i gure 17 Average total monthly river flow for major r i vers flowing into Tampa Bay .... .. . from 1981-93 ... .... ..... ..... ............ .... ... .... ..................... . ..... ....... .. ..... ... 28 111

PAGE 8

Figure 18. Climatological monthly averages of salinity by small cluster ...... ................. 28 Figure 19 Climatological monthly averages of total nitrogen by small cluster .... .... ....... 29 Figure 20 Climatological monthly averages of total phosphorus by small cluster ......... 30 Figure 21 Climatolog ica l monthly averages of chlorophyll a averaged by small cluster ..... .... ... ......... ... ....... ... ... ................. ........... ... . ..... ... .......... ..... 31 Figure 22 Factor loadings plot from PCA of CMAs ...................................... ..... ... .. .... . 32 Figure 23. Scatter plot of factor scores from PCA of CMAs coded by month ...... .... .. .. 33 F i gure 24. Scatter plot of factor scores from PCA of CMAs coded by small cluster ....... 34 Figure 25. Annual averages of water temperature by small cluster ...... . . ..................... 35 Figure 26 Annual averages of dissolved oxygen by small cluster ........................ .. ...... 35 Figure 27 Annual averages of sal i nity by small cluster .... ......... ...... ...... .. ...... ... ....... ... 37 Figure 28. Annual averages of total nitrogen by small cluster ...... . ................ .............. 37 Figure 29. Annual averages of total phosphorus by small cluster ... . ... .. .. ... .. .... .. .......... 39 Figure 30. Annual averages of chlorophyll a by small cluster ...... ............................ .... 40 Figure 31. Factor loadings plot from PCA of annual averages ....... ...... .... .................. 41 Figure 32 Scatter plot of factor scores from PCA of annual averag es coded by year ... 42 Figure 33. Scatter plot of factor scores from PCA of annual averages coded by small cluster . .... ............. ... .... . ...... ... ............. .... .... .... ..... . . ... .. .. ..... ... . ...... ... 42 Figure 34 M ixing diagrams of chlorophyll a concentration though the shipping channel. ...... ....... .... ... .. ............... .. ..................................... ................ ....... 44 Figure 35. Mixing diagrams of total phosphorus concentration though the shipping channel ..... .................................... ... . ............ ...................... .. .... ......... ... 45 Figure 36. Mixing diagrams of total nitrogen concentration though the shi pping channel .. .......... ........... ............................................... ..... ........... .. .......... 46 Figure 37. Annual averages of inorganic Nand P and organic N from tributaries to Tampa Bay .................................................................................................. 56 Figure 38. Mixing diagram adapted from Day eta/ (1989) .................... .... ...... ............ 59 Figure 39. Box plot of CMAs of water temperature ............... .......... ......... ... ............... 73 Figure 40. Box plot of CMAs of dissolved oxygen .............. .............. .. .......................... 74 iv

PAGE 9

Figure 41 Box plot of CMAs of salinity .. ........... ......... ....... ............ ......... ... ...... ..... 75 Figure 42. Box plot of CMAs of total nitrogen ... ... ....... .... ........ ....... .. ...... ..................... 76 Figure 43 Box plot of CMAs of total phosphorus .... ... . .. . .... . .. ...... ..... .... . . .. ... .. ..... 77 Figure 44. Box plot of CMAs of chlorophyll a ...... .. ........... ............... ............ .......... ...... 78 Figure 45. Box plot of annual averages of water temperature ....... ........... .. .. ... .......... 79 F i gure 46. Box plot of annual averages of dissolved oxygen ... .................................... 80 Figure 47 Box plot of annual averages of salinity .................................................. ... ... 81 Figure 48. Box plot of annual averages of total nitrogen ........ .......... ........... .............. 82 Figure 49. Box plot of annual ave r ages of total phosphorus ....... ............. ............... .. 83 Figure 50. Box plot of annual averages of chlorophyll a ................... .. . ........ ...... . ...... 84 v

PAGE 10

WATER QUALITY TRENDS IN TAMPA BAY, FLORIDA by BRIAN BENDIS An Abstract Of a thesis submitted in partial fulfillment of the requirement for the degree o f Master of Science Department of Marine Science University of South Florida December 1999 Major Professor : Gabriel A Vargo, Ph.D V I

PAGE 11

Tampa Bay water quality data was obtained from the Hill s borough County Environmental Protection Commi ssio n (HCEPC). The HCEPC maintains a water quality moni toring program of 52 stations throughout Tampa Bay and measures over 50 water quality parameters on a monthly basis A complimentary network exists for the tributaries of Tampa Bay Results from an earlier study of the HCEPC database, National Tampa Bay Estuary Program technical publication #07 -92 showed a dramatic decline in total phosphorus from 1979 to 1981 and then a subsequent decline in chlorophyll a values in 1984. This study focuses on water quality data from 1981 to 1993 and includes the variables chlorophyll a, total nitrogen, total phosphorus, salinity, water temperature and dissolved oxygen. USGS river flow data was also analyzed Three subsets of the database were used: (1) monthly data from 1981-93, (2) climatological monthly averages and (3) annual averages Climatological monthly averages (CMAs) were calculated as averages of all January s, Februarys, Marches, etc. to obtain an average monthly cycle over the 13 year database Principal components analysis was performed on each the three data subsets. Factor scores from the PCA were then used as input to a cluster analysis. A clear seasonal pattern exists for ri ver flow, TN, TP and Chi a with maxima occurring in the rainy season. The seasonality is stronger in upper Tampa Bay where nut rient loadings are greater. Chlorophyll a, TN and TP are inversely related to salinity The PCA of the CMAs results in 87% of the var i ability captured by factors 1 and 2 The PCA also clearly discriminates between the wet and dry seasons The patterns are not as tight among the variables using annual averages as with the monthly averages PCA does not reveal any clear interannual differences However, significant correlations among the var iables are evident. Furthermore, spatial d ifferences VII

PAGE 12

are clear along the salinity/Chi a/nutrient gradient from the upper bay to the lower bay. The high river flow years of 1983 and 1987 were also El Nifio years and El Nifio years result i n above average rainfall for Florida Both chlorophyll a and TP values decline after 1984 and remain at lower levels than before 1984. Cluster and principal component analysis proved to be effective tools for reveal i ng patterns i n this large water quality data set. Tampa Bay water quality stations cluster in a dist i nct pattern from the head to the mouth of the bay along what may cons i dered a salinity/Chi a/nutrient gradient. This pattern moves from the nutrient enriched riverine sources to the cleaner Gulf of Mexico waters near the mouth The riverine influence is clear in the upper parts of the bay Meanwhile, the Gulf of Mexico i nfluence is evident in lower Tampa Bay The spatial gradient is driven by high nutrient inputs in the upper regions of the bay. High rainfall amounts intensify the gradient in late summer/early fall. The homogeneous clustering of the stations suggests that thi s gradient is maintained throughout the year, as well as from year to year ajor Professor: Gabriel Professor Department of argo, Ph. D. arine Science Date Approved : 1 l Vlll

PAGE 13

INTRODUCTION The Tampa Bay area is one of the fastest growing areas in the United States Florida is fourth in population and ten of the fifteen fastest growing counties around the Gulf of Mexico are in southwest Florida (Rote 1991 ) These demographics illustrate very well the amount of societal pressure on an estuary such as Tampa Bay. The sheer number of people who rely on or simply live near the bay both contribute to the pressure on the bay, and, alternately demand that environmental impacts to the estuary be minimized. Coastal pollution is one the country's largest threats to sustaining our lifestyle in term s of aes thetics and recreat ion, but also in supplying food to our growing population. The water quality of an estuary is critical to ma i ntaining i t as a viable resource for future generations. Tampa Bay is a large, subtropical estuary on Florida's west -cent ral coast. It has 1030 km2 of open water su rface area with a watershed at 5700 km2 (Johannsen 1991 ). Tampa Bay is divided into the following seven subbasins by Lewis and Whitman (1985) : Old Tampa Bay (OTB) Hillsborough Bay (HB), Lower Tampa Bay (L TB), Middle Tampa Bay (MTB) Boca Ciega Bay, Terra Ceia Bay, and Manatee River (Figure 1). The mean annual air temperature is 22.3 o c (Wooten 1985) and the m ean annual water temperature of Tampa Bay is 23.7 oc. The average depth of Tampa Bay is three meters and mixed tides dominate with an average tidal range of 0.7 meter s (Goodwin 1989) Annual rainfall is 137 em whereas evapotranspiratation is 76-122 em annually, with significant spatial variat ion (Flannery 1989). The subtropical climate of the area creates

PAGE 14

Figure 1. HCEPC water monitoring stations in Tampa Bay b ..... 0 00 N 0 0 0 00 N b lO 0 ,..._ N b '<:!" 0 ,..._ N b C") 0 ,..._ N b N 0 ,..._ N 82 20' II 5 0 5 10 15 20 Kilometers -----s I I -:-_,...._, i r .. ', ,.,, I ""::/ -I __ ---___ J "' CX> 0 q "' CX> 0 0 q "' -..J 0 q "' -..J 0 (...) q "' -..J 0 "' q 2

PAGE 15

a wet/dry season with 60% of the rainfall occurring from June to September (Zarbock 1991 ) As expected river flow follows the local meteorological conditions w i th peak river flows occurring at the end of the rainy season The major rivers flowing into Tampa Bay are the Hillsborough, Alafia, Little Manatee and Manatee (Figure 1 ) Collectively these rivers drain approximately 75 percent of the bay's watershed and account for 82 percent of the total flow, which is approximately 2011 cfs (Flannery 1989) Flannery (1989) also recognized 44 minor tributaries, including Alligator Creek, Bullfrog Creek, and Delaney Creek most of which are ungauged and less than 17.5 miles long. Furthermore, Hillsborough Bay receives 63 to 77 percent of the total freshwater flow into Tampa Bay {Flannery 1989). Industrial and urban centers are more prevalent in the northern watershed, whereas the southern rivers flow through more agricultural and rural lands As with most urban estuaries, many of the small tidal creeks have been severely impacted through channelization, bank hardening, urban runoff, industrial discharges, and flow alteration. Similarly, the Hillsborough River has been impo unded 11 miles from the mouth to create a reservoir for the City of Tampa's drinking supply. None of the rivers flowing into Tampa Bay are very large ; all originate within the state of Florida and have relatively small watersheds. When compared to other estuaries e.g San Francisco Bay, freshwater inflow is small relative to the tidal prism which results in a vertically well mixed estuary A distinguishing feature of the bay's topography is a shipping channel that runs along the longitudinal axis of the bay from Lower Tampa Bay through Middle Tampa Bay, where it forms a Y and bifurcates into Hillsborough Bay and Old Tampa Bay. The channel is maintained at 15 meters depth and extends the influence of the Gulf of Mexico farther up the bay (King Engineering 1992) The relatively shallow water depth of Tampa Bay, combined with the subtropical climate, provide suitable habitat for seagrasses. Seagrasses cover approximately 6000 hectares of the bay bottom an 3

PAGE 16

estimated 80 percent decline since the 1800's (Lewis eta/. 1982) This decl i ne has been attributed to poor water quality conditions (FWPCA 1969). Water quality problems are not new to Tampa Bay Noxious odors from proliferating macroalgae and decomposition due to discharge of untreated domestic waste were reported as early as 1928 (FWPCA 1969). During the 1960s high bacteria counts prompted officials to post s i gns to warn swimmers of unsafe conditions (Johansson 1991 ). In 1967 and 1968, the Federal Water Pollut i on Control Administration (FWPCA) performed a water quality study of Hillsborough Bay They identified the Hillsborough River and targeted chemical companies as being responsible for 63 percent of the total nitrogen (TN) and 94 percent of the total phosphorus (TP) entering Hillsborough Bay (FWPCA 1969) Recommendations by the FWPCA included reducing total nitrogen loads by 90 percent and that the Hookers Point wastewater treatment plant remove 90 percent of the carbonaceous material from its effluent. Total nitrogen loads from Hookers Point have since been reduced by 80 percent (Johannsen 1991) and the Alafia River and the coastal Hillsborough Bay drainage basin are now the largest sources of nitrogen and phosphorus to Tampa Bay (Zarbock eta/. 1994). Fugitive losses are those losses of bulk fertilizer that occur dur i ng transfer to storage areas or vessels These losses occur through stormwater runoff and dust, and account for almost half of the TN and TP loads to the coastal Hillsborough Bay drainage basin (Zarbock eta/. 1994) Primary wastewater treatment for effluent discharged into Hillsborough Bay from the City of Tampa s Hookers Point Wastewater Treatment Facility came on l ine in 1951 and advanced wastewater treatment (AWT} became required by law in 1972 AWT came on line in 1979 however the phosphate requ ir ement of 1 mg per liter was waived due to evidence of nitrogen limitation in Tampa Bay (Johannsen 1991). This facility is currently permitted at 70 MGD 4

PAGE 17

There is a large fertilizer industry east of Tampa Bay with ship loading terminals on Tampa Bay, specifically Hillsborough Bay In 1988, 80 percent of the U S production of phosphatic fertilizer and 50 percent of the n i trogenous fertilizer was shipped from Hillsborough Bay (Johannsen 1991 ). Companies such as Nitram, Inc and Cargill Fertilizer, Inc. have historically been major contributors of nutrients to the bay (Johannsen 1991). However Johannsen found that nitrogen loading from these sources decreased substantially from 1967 to 1990 Stormwater runoff and product losses from fertilizer related facilities are significant nonpoint sources of both nitrogen and phosphorus Zarbock eta/. (1994) estimated fugitive losses to contribute 7 percent and 15 percent of the total TN and TP loads respectively to Tampa Bay Nonpoint source and atmospheric deposition are the largest loading sources to Tampa Bay of both TN and TP. However, point sources are more significant in Hillsborough Bay than any other bay segment (Zarbock eta/. 1994) In fact, Hillsborough Bay receives 60 percent of the total TP load for all of Tampa Bay (Zarbock eta/. 1994) Johannsen (1991) has shown a decrease in chlorophyll a concentrations in Hillsborough Bay since the mid-1980s, mostly due to a decrease in the blue -green alga Schizothrix calcico/a This decrease in chlorophyll a has been correlated to reduced TN loadings because of AWT, especially Hookers Point, and better practices within the fertilizer industry due to stronger regulations Through state legislation, the Environmental Protection Commission Act (1969) established the Hillsborough County Environmental Protection Commission (HCEPC) (Boler 1992). In 1972, the HCEPC began a water quality monitoring program which entailed routine sampling of stations located throughout Tampa Bay as well as its tributaries In 1994, there were 52 bay sampling stations and 40 t ributary stations (Figure 1) with 93 parameters measured This extensive sampling regime provides data for not only monitoring the water quality in the bay but also baseline data to characterize 5

PAGE 18

Tampa Bay and to study the dynamic processes affecting the bay. The HCEPC managers along with state regulatory agencies also use the water quality monitoring program for permitting policies and enforcing regulations The HCEPC generates biennial reports summarizing the surface water quality data, as mandated by the Legislature. These reports are excellent summaries of the most recent conditions in the bay and are used extensively by the scientific community and water managers in the bay area. Especially useful is the general water quality index (WQI). This index is generated by weighting several parameters according to their influence on water quality : dissolved oxygen, chlorophyll a, total coliform biochemical oxygen demand, total phosphorus, total nitrogen and effective light penetration. The individual weighted scores are then summed for each station to calculate a final water quality index with a range of 1 to 100 points and 100 being the highest possible water quality. Trends indicate increasing WQis since 1981 throughout Tampa Bay with Hillsborough Bay showing the most improvement (Boler 1995) Tampa Bay was awarded a National Estuary Program (TBNEP) in 1990. One of the first projects administered by the TBNEP was to King Engineering Associates, Inc The purpose of the project was to describe the physical, chemical and biological water quality components of Tampa Bay. The HCEPC's data set (1974-1990) was the primary data source along with those obtained from The City of Tampa, Southwest Florida Water Management District and the USGS. Results of the project can be found in TBNEP technical publication# 07-92 "The Review and Synthesis of Historical Tampa Bay Water Quality Data What does all of this mean for water quality in Tampa Bay? One thing is certain. The conditions affecting Tampa Bay's water quality have not been static over the past 20 years with the advent of AWT, the increase in fertilizer shipping from the Port of Tampa, and the apparent improved water quality in Hillsborough Bay One of the most common 6

PAGE 19

eutrophication indicators is phytoplankton biomass This is a very good indicator because of a quick response to changes in inputs to the estuary such as nutrients and freshwater flow. Chlorophyll a is an indicator of phytoplankton b i omass and therefore often used to determine the water quality status or health of an estuary. Since all of the sources mentioned above affect chlorophyll a, chlorophyll a shall be the focus of this study. Specifically, the spatial and temporal patterns of chlorophyll a and other relevant parameters, as evident in the HCEPC data set were investigated. Summary of TBNEP Technical Publication #07-92 The following is a summary of the TBNEP report No 07 92 "Review and Synthesis of Historical Tampa Bay water quality. This summary is presented in order to give additional background information that will put my study in context and to give a longer history of water quality information The TBNEP report covers the years 1974 to 1990 whereas my study uses data from 1981 to 1993 One of the most dramatic results is the observed decline of total phosphorus in the bay during the period from 1975 to 1982 (Figures 2-6) This decline occurs baywide, even in Lower Tampa Bay at stations well removed from major sources Johannson (1991) attributes this decline to better regulation of point and non point discharges into Hillsborough Bay, especially in the fertilizer industry. Also at this time, advanced wastewater treatment came online at Hookers Point treatment plant and the use of phosphate free detergents became more widespread These practices were s i gnif i cant in reducing nutrient loading to the bay (Johannson 1991 ). Since 1982, TP values have remained relatively low and steady. Sim i lar to TP, chlorophyll a concentrations have declined in the bay (Figures 7 1 0) The decline in chlorophyll a occurred from 1982 to 1984, about four years after nutrient loadings declined It may have taken the bay time to equilibrate to the lower 7

PAGE 20

nitrogen levels (Johannson 1991 ). Maximum chlorophyll a levels occurred in 1979 and 1983 at some locations in the bay, coincident with peaks in freshwater input (Figure 11 ) The relationships between freshwater flow, salinity and chlorophyll a were examined by calculating correlation coefficients of the annual averages Weak correlations between flow and chlorophyll a were found baywide, whereas salinity and chlorophyll a had stronger inverse relationships in Middle Tampa Bay and Old Tampa Bay Also salinity and flow were strongly negatively correlated in Middle Tampa Bay, Old Tampa Bay and H i llsborough Bay Figure 2 Time series of total phosphorus at station 16 from 1975-1994 c-. 0, E (/) :J ..... 0 .s:::. a. (/) 0 .s:::. a.. <0 1 2 1 .0 8 6 .4 2 0.0 7501 7701 7902 8102 83 03 8503 8703 8903 9103 9303 7601 780 2 8002 8202 8403 8603 8803 9003 9203 9403 YRMO 8

PAGE 21

. Figure 3 Time series of total phosphorus at station 6 from 1975 1994 2.0 :::::::-0, E -C/) 1 5 :::J .... 0 .s:: a. C/) 0 .s:: 1 0 a.. co 5 7501 7701 7901 8101 8302 8502 8702 8902 9102 9302 7601 7801 8001 8201 8402 8602 8802 9002 9202 9402 YRMO Figure 4. Annual averages of total phosphorus from select stations in Middle Tampa Bay 1 .0.....------------------, 8 E -C/) :::J 6 .... 0 .s:: a. C/) /"' 0 .s:: .4 I a.. Station 1 6 .$ 19 2 28 82 75 77 79 81 83 85 87 89 91 93 76 78 80 82 84 86 88 90 92 94 Year 9

PAGE 22

Figure 5. Annual a v erages of total phosphorus from select stations in Hill sb orough Bay 3 .0r------------------------, I 2.5 \ I ', \ E 2.0 I --;;:; \ 2 I / 0 \ / -a 1 5 en 0 a... $ 1 0 5 0 0 .._....--.......-.....-.....--.--.---...---...---.---.---...---..---.---.,...-,.....--..--.--.........-l. 75 77 79 8 1 83 85 87 89 91 93 Station 7 8 44 80 76 7 8 80 82 84 86 88 90 92 94 Year Figure 6. Annual averages of total phosphorus from select stations in Lower Tampa Bay .6 5 -0, E .4 .__. en ::l '-0 ..c 3 a. en 0 ..c a... co .2 1 Station 75 77 79 81 83 85 87 89 91 93 76 78 80 82 84 86 88 90 92 94 Year 23 91 94 95 10

PAGE 23

Figure 7 Time series of chlorophyll a at station 16 from 1975-1994 30 -en ell >. 20 ..c. c. 0 ,_ 0 ..c. () 10 7501 7701 7902 8102 8303 8503 8703 8903 9103 9303 7601 7802 8002 8202 8403 8603 8803 9003 9203 9403 YRMO Figure 8. T ime series of chlorophyll a at station 6 from 1975-1994 160 140 120 -en 100 ell >. 80 ..c. c. 0 ,_ 0 60 ..c. () 40 20 7501 7701 7901 8101 8302 8502 8702 8902 9102 9302 7601 7801 8001 8201 8402 8602 8802 9002 9202 9402 YRMO 11

PAGE 24

Figure 9 Annual averages of chlorophyll a from select stations in Hillsborough Bay -0, :l. .._... Cl3 >-.c a. 0 0 :c () 60 ;I 50 II I I I \ I I I I I I I \ 40 I I \ I I \ I \ I I I 30 \I I I \ ' I \ 20 /'', I '-..--.1 I \ -Station 7 8 10 44 80 75 77 79 81 76 78 80 83 82 85 87 89 84 86 88 Year 91 90 93 92 94 Figure 10. Annual averages of chlorophyll a from select stations in Middle Tampa Bay co >-10 .c a. Station 0 ...... ..Q .c () /: 0._ __ 75 77 79 81 76 78 80 83 85 82 84 86 Year 87 89 88 90 91 93 92 94 16 19 28 82 12

PAGE 25

Figure 11. Total annual river flow for the 4 major rivers flowing into Tampa Bay from 197 4-94 --------------------------------------------.::: () :0 -------------------------------:::J ;: .Q u. -------------------------------------------Year A regression analysis of the annual averages for chlorophyll a, total nitrogen and total phosphorus was also done to examine potential relationships. The correlation coefficient between TN and TP was 0.47. Both TN and TP were positively correlated to chlorophyll a with relatively high Y-intercepts, suggesting that detrital material may be a significant fraction of TN and TP. Palmer and McClelland {1988) found that most of the total nitrogen was in the organic fraction. If the organic fraction of TN is mostly refractory then it is not directly utilizable by phytoplankton Examination of the records in the HCEPC data that contain both ortho-phosphate and total phosphorus reveals that 80 percent of the TP in those samples was ortho-phosphate DIN: DIP ratios average less than 1 for the entire bay, suggesting anN-limited system according to the Redfield ratio. This corroborates other evidence that Tampa Bay is nitrogen limited (Rodriguez 1991 ) 1 3

PAGE 26

Empirical orthogonal function (EOF) analysis was also performed on the climatological mean data for several water quality parameters EOF analysis provides a characterization of the variability patterns among parameters. This analysis showed that, in general, fluctuations of various parameters are spatially uniform in Tampa Bay with values rising and falling in unison across the bay through the seasons. EOF analysis of the 17 -year time series of chlorophyll a reveals that Hillsborough Bay and areas near river mouths are the most variable and the influence of the shipping channel is quite evident. Especially important is the horizontal salinity gradient that drives the buoyancy-driven (baroclinic) circulation in Tampa Bay (Weisberg 1991 ) Galperin eta/ {1991) found this mode of circulation to be more important than the tidally-driven (barotropic) circulation. It is this horizontal pressure gradient that sets up the gradients in the water quality parameters, i.e. nutrients and chlorophyll a. 14

PAGE 27

OBJECTIVES Based upon the results of NEP Techn i cal Publication #07-92, specifically the preliminary results of the chlorophyll patterns, the objectives of this study were the following : (1) To characterize the spatial and temporal trends in chlorophyll a and quantify the trend or lack of one (2) To determine any statistical relationships among key water quality parameters, i.e. chlorophyll a, salinity, phosphorus river flow (3) To examine any possible effects of the shipping channel in establishing gradients in chlorophyll a salinity, or particular nutrients (4) To determine if similar trends occur among any stations in Tampa Bay (is there clustering of stations with respect to various parameters) (5) To determine if patterns in Hillsborough Bay's chlorophyll a levels are reflected in other parts of the bay (6) To evaluate the HCEPC sampling program for possible alternatives to the stations sampled 15

PAGE 28

METHODS Tampa Bay water quality data was obtained from the Environmental Protect i on Commission of Hillsborough County as D-Base Ill (*.dbf) files The database included tributary and saltwater stations with over 50 measured parameters at 92 stations from 197 4 to 1994 (Figure 1 ) The HCEPC has maintained its water quality monitoring network since 1972 and is the most extensive database of its kind for Tampa Bay This study focuses exclusively on the 52 saltwater stations in Tampa Bay and incorporates the following parameters : chlorophyll a salinity (ppt), total nitrogen (TN, mg/1), total phosphorus (TP, mg/1), water temperature (0C) and dissolved oxygen (DO, mg/1). These variables were chosen because of their importance and utility in defining water quality as well as to maintain continuity with the Tampa Bay NEP Report No 07-92. Salinity, DO and water temperature are measured in the field at three depths -top middle and bottom The mid depth measured value was used since that is where samples are taken for the laboratory measured variables : chlorophyll a, TN, TP. The HCEPC samples the 52 saltwater stations on a monthly basis over a three week period during which roughly one-third of the stations are sampled on one day in each of three consecutive weeks (Boler eta/. 1991 ) Therefore caut i on should be exercised in data interpretation under this asynoptic sampling regime Another caveat is that mid-depth is not the same water depth at different stations especially when comparing a 1 0 m deep channel station to 2 m deep flats station The sampling details can be obtained from the HCEPC surface water quality reports (Boler 1992, 1995) 16

PAGE 29

Due to analytical uncertainties in the total nitrogen data prior to 1981 (K i ng Engineering, 1992) and incomplete data for 1994, data from 1981 to 1993 was used for the majority of the analyses in this study (clustering, PCA, descriptive statistics). TN is calculated by the HCEPC by adding the total Kjeldahl nitrogen (unfiltered) and the NO:/N02 (filtered) values TP analyses are performed on unfiltered samples Most statistical analyses were performed us i ng SPSS for Windows statistical software (SPSS Inc 1996). From this main database two derivative files were created one of the annual averages and one of the climatological monthly averages, using the aggregate file function in SPSS The climatological monthly averages (CMAs) are defined as the average for a specific month for all years for a particular variable at each stat i on In other words, the CMA for chlorophyll is calculated by averaging all the chlorophyll a values of all the Januarys, Februarys, Marches etc. so that an average monthly cycle is obtained. Annual averages are calculated for examining interannual variations. Line graphs, box plots and descriptive statistics were created using these data subsets. Box plots provide information about the variability within, as well as among, clusters. The box represents the interquartile ranges (50 % of the values) while the line across the box is the median. The whiskers extend from the box to the highest and lowest values excluding outliers Principle component analysis (PCA) was performed on the three derivations of the data: (1) raw data from 1981 to 1993, (2) the climatological monthly averages and (3) the annual averages The resultant factor scores were saved as variables in their respective databases. Using PCA results for all data from 1981 to 1993 the average and standard deviation of each of the first three factor scores was computed for each variable for each station This matrix of 52 stations X 6 variables was then used as input for a hierarchical cluster analysis The resultant dendrogram was then used to visually designate clusters of stations based on previous knowledge of Tampa Bay. Principal components analysis is often used to examine the inherent variability in a water quality 17

PAGE 30

dataset as well as to reduce the number of variables by expressing the variability in a $Ubset of independent variables (King Engineering, 1992; Fourqurean 1993; Christenson, 1998). These new variables retain the variability from the original dataset. Cluster analysis is used to objectively group stations into groups of similar water quality. Boyer eta/ (1992) applied a PCA-clustering combination to a multiparameter water quality data set to describe "zones of similar influence" in Florida Bay. This method has also been applied to the variability of phytoplankton blooms (Vezina eta/. 1995) and macrophyte and invertebrate abundance (Harlin eta/. 1996) A field identifying the cluster of each station was then created in each derivative database to facilitate data interpretation based on the cluster analysis The river flow data originates from the USGS gauging stations, station number in parentheses, and was obtained from Coastal Environmental, Inc through the Tampa Bay NEP Monthly flow data through 1994 (starting in various years) was totaled for the Hillsborough River near Tampa (02304500), the Alafia River at Lithia (02301500), the Little Manatee River near Wimauma (02300500), and the Manatee River near Myakka (02299950). The Hillsborough River station closest to Tampa Bay was chosen so that withdrawals for use in the reservoir would be accounted. Total annual river flow was calculated by summing the monthly flows from all the tributaries for each year. Average monthly flows were calculated by summing the flows from all the tributaries and then averaging these values for each month over the data record of interest (1981 to 1993). 1 8

PAGE 31

RESULTS Raw Data A 13 year (1981-1993) record of water qua l ity data from Tampa Bay was analyzed to assess the spat ial and temporal trends in total nitrogen total phosphorus chlorophyll a, salinity, water temperature and dissolved oxygen The location of all water quality stations is dep i cted in Figure 1 A principal components analysis, based on the s i x water quality variables at all 52 stations, was used to assess the correlation among variables The first three factors accounted for 77.8% of the variance (Table 1 ) The average and standard deviation of factors 1 2 and 3 for each station was then calculated to y i eld 6 new variables. Th is matrix of 6 x 52 was then used as input to a hierarchical cluster analysis Table 1 Percent var i ance accounted for by the first three factors from each PCA Separa t e analyses were run using all data po i nts (12382 records); climatological monthly averages for each station and variable ; and annual averages for each station and variable Factor % variance Cumulative % Raw Data 1981-93 PCA 1 42.1 42 1 2 24. 9 67 0 3 10. 8 77 8 Climatological Monthly Averages PCA 1 61.4 61.4 2 25.4 86.8 3 5 6 92 5 Annual Averages PCA 1 51. 1 51.1 2 20. 7 71.8 3 14. 2 86 0 1 9

PAGE 32

Cluster Analysis The cluster analys i s r eveals three large groups of stations each of which can be further separated i nto 2 smaller groups (Figures 13 and 14) Figure 12 i s the resultant tree diagram from this cluster analys i s The linkage distance is a normalized way of describing the similarity among stations A large linkage distance would simply group all the water quality stations into one cluster, whereas an extremely small linkage distance would create 52 clusters, one for each sta t ion. The objective is to choose a l inkage d i stance at which the result i ng clusters are reasonable based on wh a t is known about the studied system. If a line is drawn i n Figure 12 at a linkage distance of 1 the result i s three clusters (Figure 13) ; whereas a linkage distance of 0 5 yields 6 clusters (Figure 14) The stations cluster in quite cohesive groups with one exception in Old Tampa Bay, station 65. Station 65 is near a once active point source and is more related to stations in Hillsborough Bay than other stations in Old Tampa Bay. The three large clusters (Figure 13), defined as A, B and C closely follow the bay segments as defined by Lewis and Whitman (1985 Figure 1 ) Cluster A i nc l udes all stations in Lower and Middle Tampa Bay south of the Little Manatee River while cluster B i ncludes all stat i ons in O l d Tampa Bay and a group of locations in Middle Tampa Bay north of the Litt l e Manatee River Cluster C includes stations along the Hillsborough Bay Middle Tampa Bay boundary and all HB locations At a linkage distance of 1 stations 8 and 54 do not cluster with other stations and can be considered outliers. The smaller clusters are obtained by splitting each of the three large clusters into two clusters when a linkage distance of 0.5 is used This finer reso l ution of funct i onal stat ion groups will be used to discuss water quality trends throughout this paper Station 58 becomes an outlier when def i ning the small clusters (Fi gure 12) 20

PAGE 33

N Figure 12. Cluster analysis tree diagram Cluster analysis performed on average and standard deviat i on of factors 1,2 ,3 froni the pri nciple components analysis on water qual i ty data from 1981-1993 ID u c ro 0 ID en ro c 7 6 ... 5 ... 4 ... 3 ... 2 ... 1 ... 0 Tree Diagram for Cluster Analysis on Avg and S D of PCs 1, 2, 3 Unweighted pair-group average ; Squared Euclidean distances HCEPC Station

PAGE 34

Figure 13 Map of large c l usters. Results from the cluster analysis show i ng the course scale (large clusters) grouping of EPC Tampa Bay wate r quality stat i ons N s 0 10 20 Miles I Large Clusters e A 8 c 22

PAGE 35

14.
PAGE 36

Climatological Monthly Averages The climatological monthly averages (CMAs) are calculated by averaging all the monthly values of a particular variable over the data record for each station. For th i s database all the Januaries Februaries, Marches etc. from 1981-93 for each water quality variable at each station were averaged to give an average year or "average monthly picture" for individual stations. The CMAs can then be averaged by a spatial delineation i.e. cluster, bay segment etc. in order to obtain regional averages. For illustrative purposes only, the averages of the outlier stations are depicted as a group instead of individually. Even though this does not follow statistical protocol this was done in order to give some idea of the values of these outliers i n relation to the clustered stations Water temperature is quite uniform throughout the bay with highs of 30 C in July or August and lows of 16-1 yo C in January (Figure 15). Conversely, dissolved oxygen (Figure 16) values are highest in January for all clusters, ranging from 7.52 for 81 to 8.43 for C1 (Table 6) The lowest DOs occur in cluster 81 (4 97) during July The greatest difference among clusters also occurs in July with C1 at 7 .02 and 81 at 4.97. The highest overall average of 7 .12 occurs in cluster C1, which is of particular interest due to its location Even though river flow is not a variable in the water quality database, it was used to define wet/dry seasons and for identifying relationships to the water quality parameters. Average monthly river flow (Figure 17) shows a strong seasonal signal. Maximum river flow of 1750 cfs occurs in September which is the peak of the rainy season A secondary peak of 700 cfs occurs in March w ith low flow in late spring (May) and throughout the w i nter (Nov-Jan) Average monthly salinity values for most subclusters are inversely related to river flow with high values during winter and late spring and minima occurring in September (compare Figures 17 and 18) The average monthly salinitie s of clu s ter A (31. 1 ppt) are always higher than clusters 8 (25 .1) and C (24 8) (Table 2). Within cluster A, the salinity of 24

PAGE 37

subcluster A 1 is on the average, higher than A2 whereas the average salinity of 82 is higher than 81 (Table 3) Clusters C1, C2, and the two outliers have similar average salinities and are intermediate to clusters 81 and 82. Table 2. Descriptive statist i cs of l arge clusters of Tampa Bay water quality stations based on CMAs Std. Cluster Minimum Maximum Mean Error Chi a outliers 7 .67 35.97 17 .96 1 .264 A 1 .83 13.44 5 02 162 8 2 95 21. 54 9 .96 .220 c 7.48 33 .94 16 26 .483 Salinity outl iers 17 14 28.03 24.47 .355 A 25.71 36 12 31.11 166 8 17 .27 29.94 25 .11 .127 c 19 38 28.57 24 .83 .155 TN outliers .67 1 .38 .98 025 A .34 1 .22 .50 008 8 .44 1 10 .71 .008 c .53 1 .31 86 .012 TP outliers .42 1 .31 .66 .040 A .04 .38 .17 .006 8 25 .63 .38 .004 c 32 1.12 .56 .010 WTemp outliers 17 24 30.38 24.26 .773 A 15 .78 30.58 23 86 .377 8 16 .08 30.91 23. 77 281 c 16 .45 30. 82 24 .07 .385 DO out l iers 1 .94 8 .50 5.60 .283 A 5 23 8 20 6 .64 063 8 3 .89 8 18 6 70 055 c 4.46 8 .68 6 .66 077 25

PAGE 38

Table 3 Descriptive statistics of small clusters of Tampa Bay water quality stations based on CMAs Cluster Minimum Maximum Mean Std Error Chi a outliers 7 67 35 97 16 65 1.702 A1 1.83 9 .52 4 44 175 A2 2 04 13.44 6 07 284 81 3.69 21. 54 10 69 .441 82 2 95 20 15 9 64 .2 49 C1 9 16 32.32 17 92 1 090 C2 7.48 33 94 16 20 510 TN outliers 67 1.38 96 034 A1 .34 69 .47 007 A2 .42 1 22 56 016 81 .51 1 09 80 .015 82 .44 1 10 67 008 C1 56 1 19 86 025 C2 53 1 .31 87 014 TP outliers .42 1 .31 73 054 A1 04 30 13 006 A2 16 38 .25 007 81 28 63 .39 008 82 25 54 .38 004 C1 .43 1 12 63 026 C2 32 95 53 009 Salinity outliers 17 14 28 03 24 .51 .464 A1 28 29 36 12 32 10 167 A2 25 .71 33 85 29 33 .208 81 17 27 25 68 23 29 196 82 21. 74 29 94 25 .91 123 C1 21. 90 27.79 24 99 .250 C2 19 38 28 57 24 74 178 WTemp outliers 17 24 30 38 24 32 958 A1 15 86 30 58 23 88 474 A2 15 78 30 06 23.82 628 81 16 56 29 83 23 52 .501 82 16 08 30 .91 23.88 340 C1 17 20 30 82 24 .51 784 C2 16 45 30.45 23.94 .419 DO outliers 1 94 8 50 5 76 387 A1 5.33 8 16 6 68 077 A2 5 23 8 20 6 55 111 81 3.89 8 .01 6 33 .108 82 4 59 8 18 6 86 060 C1 5.43 8 .61 7.16 128 C2 3 32 8 68 6 37 092 26

PAGE 39

F i g ure 15 Climatolo g ica l month ly averages of water temperature by small cluster 40.0.-----------------......... C l uster 0 0 (1) 30.0 out l ier s ..... :J ...... ro A1 ..... (1) a. E A2 ..... (1) co 20. 0 81 8 2 C1 10. 0 C2 1 2 3 4 5 6 7 8 9 10 1 1 12 Month Fig u re 16 Climatological month l y averages of d issolved oxygen by small clus t er 9 0T------------------, 8 0 /. Cluster ::::::: C> E 7 0 outlier s c (1) A 1 C> 6 0 0 A2 "0 Q) > 0 5 0 en 81 0 8 2 4.0 C 1 3.0 C2 1 2 3 4 5 6 7 8 9 10 11 12 Month 2 7

PAGE 40

Figure 17 Average total monthly river flow for major rivers flowing into Tampa Bay from 1 981-93 2000 1800 1600 1400 2 1200 0 1000 u:: .... 800 a: 600 400 200 0 2 3 4 5 6 7 8 9 10 11 12 Month Figure 18. Climatological monthly averages of sa linity by small cluster / ....... Cluster / \ --....... ............. __ , ' / ..... ..... _ 30. 0 ....... outliers --. ____ --:;:::;-----c.. . ,', c.. A1 -c: A2 / (f) 20.0 81 82 C 1 10. 0 C2 1 2 3 4 5 6 7 8 9 10 11 12 Month 28

PAGE 41

Figure 19. Climatological monthly averages of total nitrogen by small cluster 1.2 Cluster -0> E 1.0 ou tli ers c A1 Q) 0> 8 0 "-A2 z l9 .6 ...... ', 81 0 I-, ..... --'-' ' / / \ 82 / / .4 -/ C 1 .2 C2 1 2 3 4 5 6 7 8 9 10 11 12 Month The CMAs of total n itrogen (Figure 19) display a seasonal pattern simi lar to r iver flow TN values peak in the late summer/early fall at the he igh t of the rainy season. Average monthly TN is highest in cluster C1 and C2 and lowest i n A 1 Moreover, maximum total nitrogen concentrations occur somewhat later (Oct.) in subclusters A 1 and A2 than in other areas. TN peaks occur in September for C1 (1.13), C2 (1.09), 81 (0 97), and 82 (0.81) and in October for A1 (0 58) and A2 (0 69) (Table 6). T h e ou t l i ers follow a similar seasonal pattern as cluster C1, only with s l ightly higher values. Total phosphorus (Figure 20) also exhib its a seasona l pattern ; however th i s pattern is less dramatic in clusters A and 8 than in C. The increase of TP in the clusters is gradua l through the summer unt i l the peak in early fall, except with the outliers wh i ch display sharper increases and decreases th rou gh the year Maximum total phosphorus l evels occur in August for C1 (0.78), September for C2 (0.64} 81 (0.48} 8 2 (0.47), a n d A2 ( 0 32) 29

PAGE 42

and in October for A 1 (0.20) (Table 6) Lowest TP values occur in February for the A and B clusters and i n December for the C clusters The spatial pattern of TP is similar to TN and salinity with a gradient in total phosphorus values f r om the mouth to the head of Tampa Bay. There is also a progression of maximum values over time during which max ima occur earlier at head of the bay than at the mouth of the bay. Figure 20. Climatological monthly averages of total phosphorus by small cluster ...-.. 0> E ........ (/) ::l .... 0 a. (/) 0 c.. nl 8 6 .4 2 ... -...... ,, ... -----......... --_ ___ ....... .... ---, _, ' 0 0 +-----.--......------.--......------.--......------.--......------.--......----l. Cluster o utl i ers A1 A 2 81 8 2 C1 C2 1 2 3 4 5 6 7 8 9 10 11 12 Month Chlorophyll a values (F i gure 21) follow a pattern similar to river flow and nutr i ents with high concentrations in summer / fall and low values i n winter The highest ch l orophyll concentrations are found in clusters C1 whi l e the l owest va l ues occur down t he bay in cluster A 1 The down bay excursion of the seasonal chlorophyll max i mum i s evident. Maximum values occur dur i ng July/August for upper bay locations (C 1, C2, outliers) shift 3 0

PAGE 43

toward August/September for B1 and B2, and then occur in September / October for clusters A 1 and A2, locations near the mouth of Tampa Bay Accordingly, both a temporal and a spatial d i splacement of chlorophyll maxima results in Tampa Bay. Figure 21. Climatological monthly averages of chlorophyll a averaged by small cluster -:::::::: 0> :::l. m >, .c 0.. 0 ...... 0 .c () 30.0 25.0 20.0 15 0 10. 0 5 0 I I I I J /' / / / / ...; --/ / I I I ,..._....._ I I I ' ....... ..---, ', ..,.,., ............... ' Cluster ou t li ers A1 A2 81 82 C1 0.0 C2 1 2 3 4 5 6 7 8 9 10 11 12 Month A PCA based on the climatological monthly averages resulted in the first two factors accounting for 86 8 % of the variance (Table 1 ) The correlation matrix (Table 4) shows that chlorophyll a is h i ghly corre lated with both TN and TP, and negatively correlated to sa l inity Salinity is also strongly negatively correlated to TN and TP. As observed in the monthly plots, water temperature and DO have a st r ong inverse relationship. 3 1

PAGE 44

Table 4. Correlation matrix from PCA o f CMAs Chi a Salinity TN TP Wtemp Chi a 1.00 Salinity -.65 1 .00 TN .85 76 1 00 TP .84 .71 .8 1 1 .00 Wtemp .46 -.07 .41 25 DO .3 7 .08 -.43 -.20 F ig ure 22. Fa cto r loadings plot f rom PCA of CMAs 1 Principal Component Analysis Factor Loadings Climato l ogica l month l y averages Factor 1 0 0 8 0.6 0.4 .2 0. 2 0.4 DO .._sallnl y 1.00 -.78 0.6 ... Y,.temp 0 8 I 0 8 0 6 0.4 A TP 0.2 ClJlL "' 0 g 0 0.2 0 4 .6 0 8 1 A plot of th e f acto r l oa ding s for the monthly averages (Figure 22} illustrates th ese relationships expressed in Table 4 TN TP an d ch l orophyll a are all l oaded strong l y positive by factor 1 and sa linity i s s trongly ne gat iv e along the same axis. Sa linity is a l so l oa d ed slig htly negativ e ly by fact o r 2 H owever, DO and water temperature are mostly 32

PAGE 45

loaded by f actor 2, inversely to one another. Using the same axis scaling as in Figure 22, one can also plot the factor scores of each station for each month to produce a scatter plot with as many points as records in the database. Figure 23 has the points coded by month and Figure 24 has the points coded by cluster. Ellipses on Figure 23 depict the stations' congruity by season. T his type o f coding reveals the strong seasonality among the months with sta t ions from all regions of the bay clustering according to wet or dry season. Furthermore, a gradient between the dry and wet seasons is evident with the transition months of May and October grouping between the climatic seasons. Coding the points by cluster (Figure 24) clearly reveals the transition from the head to the mouth of the bay that was evident i n the monthly p l ots T here seems to be a transition zone between clusters where symbols from adjacent clusters mix and yet there is a clear pattern in the grouping of symbols. Figure 24 also shows the similarity among various stations within regions of Tampa Bay. Furthermore, the cohesive grouping of stations within clusters indicates good similarity among stations within regions of the bay Figure 23. Scatter plot of factor scores from PCA of CMAs coded by month PCA factor scores plot 3r---------------------------------, 2 C\J .9 0 (.) C'C L.L -1 -2 3 DRY -2 1 0 F a ct o r 1 A ... ... WET T 2 3 4 M o nth A 1 2 11 1 0 T 9 ... 8 7 6 ., 5 4 3 2 1 33

PAGE 46

Figure 24. Scatter plot of factor scores from PCA of CMAs coded by small cluster 3r---------------------------------. 2 C\1 .... .8 0 u \\1 u... -1 -2 -3 Annual Averages -2 -1 0 0 0 0 Factor 1 0 0 0 0 0 2 0 0 0 3 4 Cluster C2 0C1 0 82 ll 81 0 A2 o A1 o outliers Annual water temperature (Figure 25) varies only about 3 8 degrees C from its lowest value in 1982 for 81 to its highest value i n 1989 for cluster C1. There seems to be a general increase in water temperature from 1982 to 1985/86 and then a sharp decrease in 1988, at which point all clusters are very similar. After 1988 clusters C1 and C2 remain at relatively warmer temperatures than previously whereas clusters A 1 and A2 decrease in water temperature steadily after 1989. Also after 1988 clusters 81 and 82 seem to vary around 24. 5 C, then decrease in 1992 and 1993 Dissolved oxygen (Figure 26) values do not seem to follow any type of pattern from 1981-1993 Clusters A 1 and A2 have relatively lower values for 1984-1989 and the n a slight increase till 1993 Clusters 81, 82, C2 follow a similar pattern from 1985-1989 Spatially, cluster C1 has the highest DO values and the outliers have the lowest. 34

PAGE 47

Figure 25. Annual averages of water temperatu r e by small cluster 25 0 0 -:J 24 (lj ...... Q.) a. E Q.) f23 22 C l uster outlie rs A1 A2 81 82 C1 C2 81 82 83 84 85 86 87 88 89 90 91 92 93 Year F igure 26 Ann ua l averages of dissolved oxygen by small cluster 8.00 7 .50 ;, I ....... I I Cluster \ I -<:::::: 7 .00 \ 0> E \ outliers c Q.) 6 .50 0> A1 / 0 "0 6.00 A2 Q.) > 0 81 C/) 5.50 0 82 5 .00 C1 4 .50 C2 81 82 83 84 85 86 87 88 89 90 91 92 93 Year 35

PAGE 48

Total annual river flow (Figure 11) is the sum of the monthly river flows for each year. High river flows occurred in 1979 1982, 1983, 1987, 1988 and 1994 and low river flows occurred in 1977, 1980, 1981, 1984, 1985, 1989 and 1990 Monthly values show that the high flow years of 1983 and 1987 were due to higher than norma l flows in the springtime (Feb-May), whereas high flows in 1982 were due to an extended wet season (June -October). The high flows in 1988 can be accounted for by very high flows in September November and December, whereas the increase in 1991 was mainly due to an extremely wet July The low flows during 1981 and 1985 were due to a shortened wet season (August to September), while on the contrary 1990 had low flows throughout the wet season These results indicate that variations in annual river flows can result from deviations in the length as well as the magnitude of the wet or dry seasons Annual salinities (Figure 27) are basically diametrical to the river flow data (Figure 11 ). Low salinity values occur during high flow years (1983, 1987) and high salinity values occur during low flow years (1981, 1985, 1990). Salinities range from 25-35 ppt with the highest and least variable values in clusters A 1 and A2 It is interesting to note that the salinities for clusters 81, 82, C1 and C2 are all quite similar and that this does not hold true for other parameters. The annual average box plot for salinity (Figure 47) cluster A1 has the highest within cluster variability and C1 has the lowest. Annually averaged total nitrogen values can vary considerably from year to year (Figure 28). Clusters A 1 and A2 have the lowest average TN, while clusters C1 and C2 have the highest TN values; a logical distribution since sources of h igher TN are near the head of the bay. Clusters A 1 and A2 have high TN values i n 1983, 1987 and 199 3 (about 0 6 mg /1) and low TN values in 1981, 1986 and 1991 ( < 0.4 mg /1) with cluster A2 having slightly higher values than A 1 The highest concentrations of TN in the 8 clusters are in 1984 and 1987, while the low TN occurs in 1981, 1986 1988 and 1990 Furthermore, 36

PAGE 49

Fig u re 2 7. Annu al ave r ages of sali nity b y small c lu s ter /--..__ / ..__ Cluster / ....... ....... ------/ ,'-... '' 30 -./ , '-.... outliers .--. ' a. a. A 1 c A2
PAGE 50

cluster 81 is always higher in total nitrogen than cluster 82. Clusters C1 and C2 have very similar TN values from 1981-1993 with h i gh values in 1983 and 1987, and then lower less var i able TN after 1987. The outlier stations average higher TN than any cluster and their TN values rise and fall in sync with the C clusters. To some degree, annual total nitrogen values vary with river flow. The high TN concentrations of 1983 and 1987 occur during years of high river flow. However, 1984 was a year of lower river flow than 1983 yet TN values were still high and even increased in the 8 clusters Also, 1988 was a year with elevated river flow but TN values decreased throughout the bay (Figure 28) The TN box plot (Figure 48) reveals that, for the most part, the A clusters are different than the C clusters and the 8 clusters have intermediate values From 1981-85, the differences among the clusters are more defined than later in the data set when the TN values of the C clusters decreased. Also relative to TN concentration the 82 cluster is more similar to the A2 cluster than the C clusters from 1990-93. Total phosphorus values have a clearer spatial resolution. Total phosphorus does not vary as much from year to year as total nitrogen or chlorophyll a especially in the A and 8 clusters (Figure 29) Also all clusters have decreasing total phosphorus after 1990. However the spatial differences in TP are obvious with the lowest TP values in A 1 and highest in C1. Furthermore, clusters 81 and 82 have very similar TP concentrations, suggesting that TP is not a differentiating factor between these clusters Except for a general decline of TP and chlorophyll after 1990 total phosphorus levels do not vary with river flow (Figure 11 ) TN (Figure 28) or chlorophyll (Figure 30). The peak river flow years of 1983 and 1988 show slight increases in TP. The TP box plot (Figure 49) shows that there is a clear separation among clusters with respect to TP. Clusters 81 and 82 are the most similar with the least spatial variability Also, cluster A2 is usually significantly higher than cluster A1. 38

PAGE 51

Throughout the bay chlorophyll a (Figure 30) values have similar trends with h i gher values prior to 1984, then a sharp decrease i n 1984, a slight increase in 1985 and finally relatively lower and steady values through 1993 As with the nutrients, a spatial gradient is evident in chlorophyll a from the head to the mouth of the bay with the highest concentrations in the C clusters intermediate values in the B clusters and lowest chlorophyll a i n the A clusters. Annual chlorophyll a values do not vary with river flow. Chlorophyll levels in 1983 are slightly elevated over 1982, but do not reflect the large increase of river flow. The decline of river flow in 1984 is accompanied by a decline in chlorophyll a as well. The chlorophyll a box plot (Figure 50) reveals the same trends as the annual average line graphs Figure 29. Annual averages of total phosphorus by small cluster 1 .00..--------------------., 80 Cluster E (/) ::I ..... 0 .s:::. 0. (/) 0 .s:::. a.. (ij 0 I-I 60 v .40 I '""-... I ', __ ___ , ....... ---.20 .. __ ..-----/ ___ ....,........ -......... -. 0 00 82 83 84 85 86 8 7 88 89 90 91 92 93 81 Year outliers A1 A2 81 82 C1 C2 3 9

PAGE 52

Figure 30. Annual averages of chlorophyll a by small cluster 30 Cluster -::::::: 0> :J.. outliers cu >-2 0 ..c A1 a. 0 ..... A 2 .2 ..c 0 81 10 82 0 81 82 83 84 85 86 8 7 88 8 9 9 0 91 92 93 Year A principle component analys i s on the annual averages resulted in the fi rst two factors accounting for 71. 8 % of the variance (Table 1 ) The correlat i on matri x for annual averages (Table 5) shows that chlorophyll a is negatively correlated to salinity and positively co rr elated to total nitrogen and total phosphorus Salin ity is a l so negat i vely correlated to both total nitrogen and total phosphorus Water temperature a n d disso l ved oxygen are not strongly correlated as they were with the monthly averages. Table 5 Correlat i on matrix from PCA of annual averages Chi a Salinity TN TP Wtem_Q Chi a 1 .00 Salinity -.57 1 .00 TN .67 78 1 00 TP .74 66 .68 1.00 Wtemp 02 15 -.13 03 1 00 DO 14 03 0 7 -.03 -.23 4 0

PAGE 53

Figure 31. Factor loadings plot from PCA of annual averages Annual mean data Principal Component Analysis Factor Loadings -1 -o.8 -o.s -o.4 -o.2 Factor 1 0 0.2 0.4 0.6 0.8 Wte no A TP ... lini .A. TN ... Chla DO r-0.8 0 6 0.4 0.2 N 0 g '" Ll.. -o.2 -0 4 -o.s -o. 8 -1 The factor loadings plot (Figure 31) for the annual averages is similar to the one for the monthly averages (Figure 22) Total nitrogen, total phosphorus and chlorophyll a are all loaded strongly positive by factor 1 and salinity is strongly negative along the same axis. Dissolved oxygen and water temperature are mostly loaded by factor 2, inversely to one another. A factor scores scatter plot of the annual averages coded by year does not reveal much of a distinguishable pattern (Figure 32}. The only year separated much from the mass of points in the middle of the plot is 1982. The annual averages factor scores plot coded by cluster (Figure 33} is very similar to the one for the monthly averages with the factor scores segregated according to cluster in a gradient from the upper to lower bay regions. 41

PAGE 54

Figure 32 Scatter p l o t of factor sco r es from PCA of annua l averages coded by year 4 .... 93 .... 92 3 .... 0 91 2 ... 90 0 89 ... 0 8 8 C\J 0 0 ...... 0 .. ... 8 7 t5 0 <0 LL.. .... 86 -1 ... 85 .,. ... 84 -2 t;. 0 83 -3 82 4 81 -3 -2 -1 0 2 3 4 Factor 1 Figure 33. S catte r p lot of factor scores from P C A of a n n u al averages coded by small cluster 4 0 0 0 3 0.6 0 Cob c O 0 0 0 2 0 i:fl .6 0 0 0 0 Cluster 0 0 0 C\J 0 C2 ...... .8 0 C1 (.) 0 <0 LL.. 0 0 -1 0 8 2 t 0 .6 81 -2 D O 0 .6 0 0 0 0 0 A2 ID d> .6 -3 0 A1 4 0 o utlier s -3 -2 -1 0 2 3 4 Factor 1 4 2

PAGE 55

Mixing Diagrams Another way to examine the down-bay gradient is by plotting the CMAs of certain variables for selected stations along a salinity gradient from the mouth to Hillsborough Bay (Figures 34-36). The selected stations were chosen on the basis of central location and depth, i.e. in the shipping channel, and are 94 95, 91, 23, 19, 16, 81, 80, 55 Plotting concentration versus salinity yields a mixing diagram with Hillsborough Bay at the low salinity end and Lower Tampa Bay at the high salinity end The chlorophyll a mixing diagrams (Figure 34) all have nearly the same general shape with two distinct regions For plot A, Jan to June, the slopes of the curves are greatest between 27 and 33 ppt, whereas the steepest slopes occur between 23 and 28 ppt from July to Dec in plot B. In both plots A and B the curves are much flatter at higher salinities. Furthermore, the curves are displaced along the salinity axis based on wet or dry season as the salinity in this region varies with the season. The dry months (May, June Dec) are shifted to the left (higher salinity) whereas the wet months (March, Aug, Sept Oct) are shifted to the right (lower salinity) The mixing diagrams for total phosphorus (Figure 35) follow more closely a 1:1 relationship than the chlorophyll a curves such that the TP curves are nearly straight lines S i milar to chlorophyll, the TP mixing curves are slightly separated temporally with .the dryer months shifted to higher salinities than the wet months The mixing diagrams for TN (Figure 36} do not have the linear shape of the TP curves nor the steep slopes and flat regions of the chlorophyll curves. 43

PAGE 56

Figure 34. Mixing diagrams of chlorophyll a concentration though the shipping channel. Plots based on the climatological monthly averages for locations in the channel from the mouth of Tampa Bay to Hillsborough Bay, (A) Jan-June, (B) July-Dec cu > ..c a.. 25 20 15 e 1o 0 ..c u 5 --Jan Feb -Mar ----Apr --.---May -June 36 35 34 St. 94 mouth 33 32 31 30 29 28 salinity 27 26 25 24 St. 55 HB 25.---------------------------------------------------, 20 ::::'15 0) ::l.. .__.. cu '10 a.. 0 .... 0 ..c u 5 --July -Aug -Sept ---Oct ---.-Nov Dec B 36 35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 St. 94 mouth St. 55 HB salinity 44

PAGE 57

Figure 35 Mixing diagrams of total phosphorus concentration though the shipping channel. Plots based on the climatological monthly averages for locations in the channel from the mouth of Tampa Bay to Hillsborough Bay, (A) Jan-June, (B) July-Dec. 1 1 .-------------------------------------------------, 1 0 -Jan 0.9 :::::--0.8 0> E 0 7 en :J 0 0.6 .c g. 0 5 0 .c 0. 0.4 ca .s 0 .3 0 2 0.1 Feb --Mar ----Apr ___.May June 0.0 -:::::: 0> E en :J 1.10 1.00 0 90 0.80 0 70 0 0 60 .c g. 0 50 0 .c 0. 0.40 .s 0 30 0 20 0.10 36 35 34 33 32 31 30 29 28 27 26 25 24 St. 94 mouth salinity St. 55 HB -..July -Aug -Sept ---Oct --..--Nov Dec B 0 00 .f.-.-rrrm..,..,rr-r.,.........,.-.-rr,...,...,...rrr"T'T'--rr-['T""",..,.-r,..,...,rrr-TT"'...-..,--rr-,.....,..,.,...,..,-,rrr.....-j 36 35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 St. 94 mouth salinity St. 55 HB 45

PAGE 58

Figure 36. Mixing diagrams of total nitrogen concentration though the shipping channel. Plots based on the climatological monthly averages for locations in the channel from the mouth of Tampa Bay to Hillsborough Bay, (A) Jan-June, (B) July-Dec :::::::: 0> E -c Q) 0> 0 ..... ..... c: m .... 0 .... -1 0 0 9 -Jan --Feb -Mar 0 8 ---Apr ----May --June 0 7 . 0 6 0.5 0.4 0 3 A 36 35 34 33 32 31 30 29 28 27 26 25 24 St. 94 mouth salinity St. 55 HB -July 0 90 --Aug -Sept 0 .80 .. --Oct ----Nov 0 .70 . ._______ o_ec__, -c Q) g> 0.60 .... ..... c: en o.5o .... 0 .... 0.40 0.30 0 .20 36 35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 St. 94 mouth salinity St. 55 HB 46

PAGE 59

DISCUSSION Cluster Analysis The cluster analysis shows how the stations group based on the six variables included in the analysis in both a spatial and temporal framework. Combining the cluster analysis with a PCA, line graphs and descriptive statistics allows us to infer relationships about processes governing locations or regions of Tampa Bay, e.g tidal mixing and circulation The clusters (Figures 12, 13 14) reflect the main sources of "new water" or what some may consider forcing functions. The location of cluster A 1 suggests that these stations are strongly influenced from the Gulf of Mexico water, whereas cluster A2 is a transition region that receives a mixture of inputs from the Gulf and from upper Middle Tampa Bay Residence t ime of water masses in this region of the bay are relatively low, maybe even less than a week (Mark Luther, personal communication) Station 24 stands out from other Lower Tampa Bay /A1 stations because of its location. Since the circulation within Tampa Bay moves the ebb tidal flow water out along the banks (Galperin 1991 ), water from the Port Manatee area may flow south and be trapped by the southern Sunshine Skyway bridge causeway. This flow pattern would influence the characteristics of station 24 differently than other L TB stations. Cluster B1 departs from other areas in Old Tampa Bay. This region receives flow from the small tidal creeks surrounding upper Old Tampa Bay, i.e Alligator Creek, Lake Tarpon outfall Rocky Creek, Sweetwater Creek and others, and has poor 47

PAGE 60

circulation north of the bridge causeways. Meanwhile, cluster B2 typifies most of Old Tampa Bay and upper M i ddle Tampa Bay The Middle Tampa Bay stat i ons in cluster B2 have intermediate values for the variables analyzed due to influence from HB, OTB and L TB. This area is essentially where water from the upper and lower bay mix to create the transition that was observed in the factor scores plots. Hillsborough Bay water can influence Old Tampa Bay when water is transported around the t i p of lnterbay peninsula into Old Tampa Bay by southwest winds and/or tidal mixing The pumping action created by the tides can slowly move water coming out of Hillsborough Bay into Old Tampa Bay (Mark Luther, personal communication). Station 84 is strongly influenced by the Little Manatee River However, the Little Manatee River is the least developed river residentially or commercially, of Tampa Bay and therefore station 84 clusters with the less river-impacted stations of cluster B2, instead of B1. Clusters in Hillsborough Bay along with one location in Old Tampa Bay, are characterized by elevated anthropogenic inputs. Cluster C1 stations {52, 71, 73) may be more influenced by industrial discharges than river input. Stations 52 and 71 are very close to industrial discharges in McKay Bay and effluent from Hookers Point Wastewater Treatment plant and are likely heavily influenced by these inputs Station 73 is near Delaney Creek, historically one of the most industrially polluted of Tampa Bay's tributaries (Flannery 1989) However, the influence of the Alafia river on station 73 is mitigated by a series of shallow spoil banks south of the river's channel. The outliers are either in McKay Bay (stations 54 and 58) or at the mouth of the Alafia river (station 8). High nutr i ent concentrations and low DO values may be the determining factors creating these single station clusters Cluster C2 stations are influenced more by river inputs Stations 44, 70, 6 and 7 are heavily influenced by the Hillsborough River whereas station 55 is influenced by the Alafia River Stations 11, 80 and 9 are transition stations from Hillsborough Bay to Middle Tampa Bay Furthermore, the industrial and strong 48

PAGE 61

riverine inputs affecting the C clusters are likely different en o ugh from the mos tl y urban and agricultural runoff inputs from the small tidal creeks characterizing the B clusters in OTB to help cause a distinction between these two regions of the bay Monthly Averages The seasonality evident in all parameters at all locations throughout the bay is a function of the strong dichotomy between the wet and dry seasons experienced in the Tampa Bay region In order for seasonality to exist in the variables there must be some seasonality of the inputs Areas of the bay that receive more seasona l inputs, i.e Hillsborough Bay and Old Tampa Bay exhibit stronger seasonal signals and are inherently more variable The PCA of the CMAs captures most of the variability and i s therefore a valid representation of the data. This also lends credence to describ i ng the relationships among variables and locat i ons using the factor loadings and scores plots. Therefore, clusters closer to inputs such as stormwater runoff and rivers differ not only i n magnitude from other clusters but also are more variable (Table 6) Also, the spatial gradient is larger when the inputs are la r ger, i.e. the differen c e between cluster s is largest during the rainy season and smallest during the dry sea s on (Figures 16, 18-21 ). The trends exhibited in the CMA line g r aphs are also apparent in the box plots (F i gures 39 44) which provide the added information of variabi l ity. The box plots clearly show that the upper bay clusters (C1, C2, B2) are more variable, spatially and temporally, than the lower bay clusters. Also variability within clu s ters (spatial) is largest duri ng the wet season as shown by the length of the boxes Stations affected the most by these inputs become outliers e g Stations 8 and 54. Station 8 is at the mouth of the Alafia River whereas station 54 is in M c Kay Bay (Figure 1) wh i ch h i stor i cally received heavy industrial runoff and f low from the Tampa 49

PAGE 62

Bypass Canal. Station 58, also in McKay Bay, is an outlier from the small clusters but becomes part of cluster C when the linkage distance is increased (Figure 12} Clusters C1 and C2 are most often at opposite ends of the y-scales of the variables analyzed relative to clusters A 1 and A2. The water temperature is an exception to this characterization (Figure 15). Water temperature is not strongly affected by what we would consider traditional inputs, e.g rivers and runoff, but is mostly controlled by solar input, which is presumed to affect the entire bay equally. In addition, the difference in DO between clusters decreases (Figure 16) during the dry season when water temperature is the main determinant and the influence of other factors is less. However, in the summer when factors such as production and respiration are operating at high rates, the differences in DO among many clusters increases, which suggests that factors other than water temperature can have significant impacts on the DO. Of particular interest is the increase in DO in July for clusters C1 and C2, which is coincident with an increase in chlorophyll a (Figure 21 ). This is the first large increase of chlorophyll a for the season and therefore the accompanying increase in primary production may not be offset by large respirat i on values The first algal bloom of the summer season, as indicated by the increase in chlorophyll, may not have died out yet. Therefore, decay of the associated organic matter and respiration are not consuming the 0 2 as fast as the phytoplankton and other primary producers are producing 02, thereby resulting in an increase in dissolved oxygen As the rainy season intensified, the increase was followed by a decrease in dissolved oxygen due to various factors such as increasing water temperature, increased respiration and decay of organic matter associated with the higher river flows and the higher algal biomass. The seasonality in chlorophyll a is strongly related to the seasonality of river flow and therefore salinity as well. The chlorophyll maxima for all clusters occur during the rainy season. The chlorophyll peaks progress to later in the rainy season moving from 50

PAGE 63

cluster C to cluster A (Figure 21 ) A high chlorophyll plume seems to move down the bay through time as if along a dilution line decreasing downbay as the influence of the Gulf of Mexico increases. The chlorophyll a maxima in the upper bay clusters occurs in July/August, whereas the lower bay chlorophyll a peaks occur in October TN and TP follow a similar pattern with higher values early in the rainy season in the upper bay clusters and later in lower bay clusters. The seasonality is also reflected in the factor loadings plot (Figure 22} and the factor scores plot (Figures 23 and 24). In Figure 22, the posit i ve loading of factor 1 on TN TP, and chlorophyll a and nega t ive loading on salinity results in h i gh nutrient, high chlorophyll a, low salinity data are d i splaced to the right. Conversely data with low nutrients, low chlorophyll a and high salinity are displaced to the left. The loading of factor 2 on salinity tilts this effect so a diagonal displacement occurs Also, opposite loadings of factors 2 and 1 on DO and water temperature produce an antagonistic diagonal effect that aligns the data along a diagonal from the southwest to northeast quadrant. Coding the factor scores by month (Figure 23) eluc i dates the seasonal effects Wet months are warmer, provide more inputs from runoff and rivers and therefore result in lower salinities and higher nutrients and chlorophyll a than in dry months This displacement is more pronounced for clusters C1 and C2 which are proximal to inputs (Figure 14). In Figure 24, the high salinity, low nutrient, low chlorophyll a stations of clusters A 1 and A2 are to the lower left. These stations are highly influenced by Gulf of Mexico waters. The low salinity, high nutrient, high chlorophyll a stations of clusters C1 and C2 are oriented to the upper right whereas the B cluster stations are intermediate to the others 51

PAGE 64

Annual Averages The CMAs are useful for determin i ng what an average year looks like whereas annual averages reveal how much a variable changes from year to year and if trends a r e apparent. Water temperatu r e varies from year to year with unclear spatial differences (Figure 24) There is a gene r al warming trend from 198 2 to 1986 then a sharp decrease to 1988, and then warmer values from 1989 to 1993 The upper and middle bay clusters are generally more variable than the Lower Tampa Bay clusters. Th i s is a reflection of the influence of Gulf water in Lower Tampa Bay. More energy is required to change the temperature of a larger deeper body of water such as the Gulf of Mexico than a shallow body of water like Tampa Bay The Band C clusters are not only shallower but also receive more freshwater influence, which is goi ng to be more var i able in temperature than the Gulf of Mexico The apparently lower dissolved oxygen values from 1984-1985 (Figure 26) do not seem to be rela ted to any parameters analyzed in this study Cluster C1 has the highest average dissolved oxygen possibly because of the freshwater influence i.e o x ygen is more soluble in freshwater than saltwater (Table 2, Parsons eta/. 1984) The outl i er stations also rece i ve freshwater ; however the influence of the increased chlorophyll a at these stations counteracts the freshwater dissolved oxygen effect due to increased phytoplankton resp i ration and the decay of organic matter Annual salinity patterns are very much influenced by river flow The salinity values and variablilites are determined by a particular station s location in reference to the saltwater source at the mouth and f r eshwater source at the head of the bay. Stations closer to the bay s head have lower and more variable salinities than those towards the mouth Clusters A 1 and A2 have higher and less variable salinity than B or C clusters (Figure 27, Table 6) The degree of change i n river flow is also man i fested in bayw i de 5 2

PAGE 65

salinity. The high variability in river flow from 1981-1987 result in pronounced changes in salinity, whereas the milder variations after 1988 did not result in nearly as much variation in the salinities The variance for salinity at all stations before 1988 is 14.36 ppt and after 1988 is 9 33 ppt. The occurrence of two El Nifios 1982/83 and in 1987 / 88 may explain the greater variability prior to 1988. El Nifio usually results in above average rainfall during the winter and spring for central Florida, whereas the counterpart La Nifia results in very dry springs. The springs of 1983 and 1987 had above normal river flows due to strong El Nifios, which essentially set up two wet seasons, spring and fall The association of total nitrogen with river flow is not as clear as the association of flow and salinity. Total nitrogen has more sources and sinks than salinity and therefore is expected to be more variable. When comparing TN, TP, and chlorophyll a levels, the TN and TP concentrations in 1983 are similar to the concentrations in 1987, but the corresponding chlorophyll peak in 1983 does not occur in 1987. Differences in the composition of TN in 1987 compared to 1983 (organic vs inorganic) may account for the differences in chlorophyll a levels. Total nitrogen in clusters B1 and B2 (Old Tampa Bay) does not decrease in 1984 as in the other clusters. Runoff is a relatively more important freshwater source for Old Tampa Bay when compared to Hillsborough Bay since there are no major rivers flowing into OTB (Zarbock eta/. 1994). If runoff to the tidal creeks does not decrease like river flow does, due to an activity such as lawn watering, the inputs of total nitrogen to that part of the bay may not decrease as much as in other parts. These kinds of differences in freshwater sources are important considerations when one attempts to link water quality to contaminant loadings. Loadings are more appropriately expressed as mass rather than concentration. For instance, river flow can increase but concentration may decrease. Loadings c ould increase in this scenario if the total mass of a nutrient, e.g. nitrogen, delivered to the 53

PAGE 66

estuary is greater due to the increased river flow irrespective of the concentration It is this type of complication that makes it difficult to decipher relationships between input variables such as TN and a response variable such as chlorophyll a when using only the ambient concentrations. The decrease in total nitrogen in 1986, even though river flow increased, could be a result of the timing of the low flows in 1985. The 1985 annual flow was reduced because of a shortened wet season of only August and September Therefore, there was not a high input of total nitrogen at the end of 1985 to maintain higher total nitrogen values at the beginning of 1986 through "seeding" of the nitrogen pool. Total nitrogen includes both the particulate and d i ssolved forms of both the inorganic and organic species of nitrogen Hence, low flows and lower total nitrogen values at the end of 1985 cannot maintain the high chlorophyll a values normally associated with the rainy season and thereby "seed" the nitrogen pool. Therefore, there is less total nitrogen available to regenerate or recycle into the upcoming dry season. Approximately 90 percent of the total nitrogen in Tampa Bay is organic nitrogen (Palmer and McClelland 1988). If most of the organic nitrogen i s refractory, then it is not directly utilizable by phytoplankton and therefore a nitrogen limited system can develop Contrary to TN, Palmer and McClelland (1988) state that most of the total phosphorus in Tampa Bay is in the inorganic fraction, orthophosphate. This indicates that the majority of phosphorus entering Tampa Bay from the rivers is available to the phytoplankton directly for uptake. From 1981 to 1993, total phosphorus remains relatively steady (Figure 29) The most dramatic changes are decreases in total phosphorus after 1988 in the C1 and C2 clusters and after 1990 for the A and B clusters This decrease is most likely due to the lower river flows from 1989-1993 compared to 1981-1988 The Alaf i a river i s a dominant source of TP to Tampa Bay (Zarbock et at. 1994). Because the rivers entering Tampa Bay are naturally high in phosphorus, there may be a threshold below which river flow must drop in order for 54

PAGE 67

significant changes to be noticed in the bay. In other words, high background leve l s require large variations in order for significant changes to be noticed in the amb i ent concentrations. TP has remained relatively steady since and therefore dramatic chlorophyll changes have not been observed. Chlorophyll a concentrations follow an expected pattern spatially but not temporally. The baywide decrease in chlorophyll a in 1984 (Figure 30) occurs during relatively high total nitrogen levels (Figure 28). Also, the 1987 total nitrogen peak attributed to El Niiio is not reflected in the chlorophyll a signal. One reason for this apparent lack of correlation may be the source of nitrogen. Figure 37 depicts the average N and P concentrations from the tributary stations in the HCEPC database from 1974 to 1994. It is clear that inorganic nitrogen values were lower in 1983 than in 1987. The low inorganic nitrogen concentration in 1984 corresponds to the low chlorophyll a of that same year. Phytoplankton growth in Tampa Bay is nitrogen limited (Rodriguez 1991) and therefore a decrease in the inorganic-N supply would account for less biomass and chlorophyll a. The nitrogen levels then may have rebounded to normal" concentrations and therefore the increase of nitrogen in 1987 was not high enough to elicit a large increase in chlorophyll a. The high nitrogen values in 1987 were most likely caused by two chemical spills of ammonium-nitrate that occurred at a fertilizer facility on Hillsborough Bay. Unfortunately, inorganic-N values prior to 1983, which would help in establishing average levels, are not available in this dataset. The spatial differences in chlorophyll a are typical estuarine distributions with areas closer to the head of the bay higher in chlorophyll a than those at the mouth of the bay. This spatial distribution is apparent in the factor loadings plot and the factor scores plot coded by cluster. 55

PAGE 68

Figure 37 Annual averages of inorgan i c N and P and organic N f rom tributarie s t o Tampa B a y Data from 40 tr i butary stations was averaged for each year from 1974 to 1994 2 5 r----------------, 2.0 J 0:::: ,, 0> II E I\ I I 1 5 1 I a. I I 0 I I z Q) I 0> 1 0 v Q) > ... <( . '\
PAGE 69

of individual stations Therefore, water temperature and DO are relatively inconsequent i al on an annual basis, compared to other variables in dete r mining the spatial differences among stations in the bay Coding the factor scores by year does not reveal any d i stinct year to year differences except for 1982 (Figure 32) 1982 was the coldest year (EI Nino again) in the data analyzed and since factor 2 loads positively on water temperature, a cold year would be highly negative on the factor 2 axis The main message from the scatter plots (Figures 32, 33) and annual time series plots (Figures 2530} is that the stations segregate spatially but the annual variability may not be large enough to d i sti n guish d i fferences from year to year This may also be a result of the smoothing effect of averaging and that those differences may be clearer using raw data. The correlation matrix generated by the PCA of annual averages reveals the same relationships as the monthly averages except for water temperature and DO The negative correlation between chlorophyll a and salinity, total nitrogen and total phosphorus i s weaker on annual basis than on a monthly basis The annual average line g r aphs (Figures 25-30) illustrate this weak relationship in a qualitative manner. The correlations are weaker on an annual basis due to the annual variability versus monthly variability The variability among seasons is well defined; the magnitudes may change but the general pattern remains the same. The strong seasonality repeats from year to year, however the year-to-year var i ations do not necessarily repeat. Spatial Gradient The temporal lag in the maximum CMAs from the head to the mouth of the bay is an interesting feature of the bay Maximum values of cluster A 1 lag the maximum values of clusters C1 and C2 by 2-3 months for chlorophyll a and 1 month for TN and TP (Figures 39-44, Table 6) The chlorophyll a peak in cluster C1 occurs before the peaks of TN, TP or river flow. This disparity suggests that the peaks i n TP and TN associated 57

PAGE 70

with the peak in river flow and minimum salinity are mostly due to river borne particulates, not autochthonous material. The variability apparent in the box plots makes it difficult to discern when the peak occurs, especially in cluster C1. Because the HCEPC samples the stations in a staggered format over 3 weeks, the time lag is actually longer than revealed with the CMAs. For instance, the TP peak in cluster A 1 occurs 7 weeks after the TP peak in C2 and 11 weeks after the TP peak in C1. The lag time associated with the peaks in water quality variables may be related to the flushing rate of the bay. Using the values in Table 6 to discern maximums, it would take approximately two months to move particulate matter from cluster C1 in Hillsborough Bay to cluster A 1 in Lower Tampa Bay This does not mean that the same particle of TP is conserved as it moves down the bay since chlorophyll a is part of TP. The P is being recycled through various biochemical pathways as a parcel of water moves down the bay. It seems as though the major algal bloom remains in HB for the summer. The baroclinic flow is then intensified by the increased river flow in August and September, which speeds up the transport of water out of HB. If this theory is true, then about 1.5-2 month later the bloom reaches L TB. Figure 38 is adapted from Day eta/. (1989) and is used to investigate the removal/addition of dissolved and particulate substances during estuarine mixing. A conservative constituent would yield a straight line since changes in concentration are strictly due to dilution However a convex or concave line would mean the estuary is acting as a source or sink, respectively (Day eta/. 1989). Figure 38 is the theoretical curve that forms the basis for constructing and interpreting the mixing diagrams for chlorophyll a, TP and TN (Figures 3436). The chlorophyll a mixing diagram shows that chlorophyll a is not a conservative property (Figure 34). An interesting feature of Figure 34 is a region where there is a large change in the X and Y values. The salinity range for this region of the curve is 58

PAGE 71

between 27 and 33 ppt for plot A and between 23 and 28 ppt for p lot B Geographically, this reg i on is between stations 16 and 81 in upper Middle Tampa Bay (Figure 1 ) On the less saline (up-bay) side of this region, the slopes of the curves are the steepest and therefore this is where the highest loss rates occur. Steeper slopes indicate a greater than a 1 : 1 change in chlorophyll a with salinity. Therefore accord i ng to Figure 38, this region can be viewed as a sink for chlorophyll a due to the shape of the plot and the decrease in chlorophyll a. Towards the mouth of the bay (higher salinity) from this reg i on, the curves tend to be flatter and the points more condensed relative to the Xaxis The change in the slopes to flatter curves suggests that the rates of the processes affecting chlorophyll a are lower in MTB and L TB than in HB From station 16 to the mouth of the bay, the flat shape of the curves also indicates that chlorophyll a concentration does not change very much along this transect through the channel. Figure 38 Mixing diagram adapted from Day eta/. (1989). This figure shows the shape of a curve of an example water quality parameter along a gradient, e g. salinity, in an estuary from the river (head) to the sea (mouth) 12 10 Source 8 < Conservat 10 6 0 >-4 Sink 2 0 0 2 4 6 8 10 12 Sea X Data River 59

PAGE 72

King Engineering (1992) produced contour plots that show a salty tongue" extending up the center of the bay The temporal displacement of the mixing curves suggests that the influence of the Gulf of Mexico reaches farther up the bay during the dry months and that the increased freshwater flow during the rainy season extends the riverine i nfluence more towards the mouth of the bay During the rainy season the stations' positions on the plot are displaced to the less saline side Essentially, the "salty tongue" of Gulf water is pushed farther down the bay by stronger baroclinic circulation This influx of Gulf water penetrates farther up the bay during dry months than in wet months This source of water has a more consistent character to it and is therefore less variable through the seasons. It also has lower concentrations of nutrients and chlorophyll a than the water coming down the bay from Hillsborough Bay and Old Tampa Bay The effect of this "salty tongue is to dilute the "Hillsborough Bay source water'' This dilution gives the mixing curves the flat, linear character on the saline side of the s i nk area The mixing diagram for total phosphorus (Figure 35) displays a similar temporal displacement to that of chlorophyll a However the dramatic change in slope of the curve does not happen with TP Although a slight curvature to the plots is evident due to the composition of TP (chlorophyll containing phytoplankton are included TP), total phosphorus displays a more conservative mixing curve The nearly conservative nature of TP may be due to either equal-rate processes along the salinity gradient, or more possibly dilution The high inputs of phosphorus to Tampa Bay tend to favor the dilution theory for the semi-conservative nature of TP. Furthermore, the fact that phytoplankton populations in Tampa Bay are nitrogen limited would argue that phosphorus is present in excess amounts. The phosphorus rema i n i ng after nitrogen is exhausted is then exported out of the bay, getting diluted by Gulf water as it travels down the bay The total nitrogen mixing diagram (Figure 36) does not have a clear zone of 60

PAGE 73

change as chlorophyll a does yet total nitrogen behaves non-conservatively, like chlorophyll. Except for May and June, the temporal separation evident in TP and chlorophyll a is not as clear in TN. Total nitrogen includes various forms of N so a conservative behavior should not be expected Consequently the variations of TN concentration a l ong the channel give a rather confusing picture of TN. Generally the shapes of the TN mixing curves imply the nature of a sink rather than a source The mixing diagrams for Tampa Bay do not provide a complete picture because they do not provide insight to the mechanism of removal or the nature of the product removed, part i culate or dissolved (Day eta/. 1989). In addition, these diagrams are not classic mixing curves since there is no freshwater end point. Another assumption of mixing diagrams is that the end points do not change over the time course of samp l ing along the dilution gradient. The end points of these mixing d i agrams (Figures 34-36) are sampled two weeks apart. Therefore, it is difficult to assume that a constant state was maintained at each end point. However, the points in the curve from station 81 northward are sampled on the same day and therefore the assumption is not completely violated. In spite of these caveats, the mixing diagrams contained herein do g i ve some insight to the behavior of constituents moving down the bay. Chlorophyll exhibits the most nonconservative behavior with most of the loss occurring between 23 and 31 ppt depending on the season. The technical report by Wade and Janicki (1994) contains a map (Figure 1, pg 27) which illustrates accumulation of fine grained sediments in th i s reg i on The sink character of this region may be due to physi c al reasons, not biolog i cal. As water flows down Hillsborough Bay into M i ddle Tampa Bay, the part i culate material carr i ed in by the rivers, as well as that generated in HB may be e ncounter i ng a circulation environment that encourages settlement to the bottom. On the other hand, total phosphorus exhib i ts a more conservative behavior because of the high inputs of TP to Tampa Bay Conversely, total nitrogen exhibit s a 6 1

PAGE 74

complex behavior that may exhibit the pattern of a sink or source depending on location in the bay or time of the year Since TN is a limiting factor of phytoplankton growth in Tampa Bay (Rodriguez 1991 ), the complex nonconservative nature of TN is expected the higher degree of variability Moving down the channel, removal rates of TN by flocculation, adsorption, and uptake may vary and sources of TN may be encountered as well through various biochemical pathways. 6 2

PAGE 75

CONCLUSIONS AND RECOMMENDATIONS Cluster and principal component analysis proved to be an effective combination of tools for revealing patterns in this large water quality data set. Tampa Bay water quality stations group in a dist i nct pattern from th e head to the mouth of the bay along what may considered a salinity/chlorophyll a/nutrient gradient. This pattern moves from the nutrient enriched river i ne sources to the lower biomass Gulf of Mexico waters near the mouth The riverine influence i s clearly evident in the upper parts of the bay (clusters C2 C1, B1 ) Meanwhile, the Gulf of Mexico influence is more pronounced i n lower Tampa Bay (clusters A 1, A2) The proximity of stations to inputs (rivers, industry etc ) determines not only the magnitude but also the variability of water quality variables in a particular area The cluster analysis provides a similarity index of the stations based on the variability captured by the PCA. A clear seasonal pattern exists for river flow, TN TP and chlorophyll a This seasonality i s driven by rainfall The season a lity is stronger in upper Tampa Bay where freshwater sources have a larger influence Contour p l ots (TBNEP Technical Rpt #0792) show that the Gulf influence is not as strong during the rainy season TN, TP, and chlorophyll a are all negatively correlated to salinity on a monthly basis. The segregation of the stations by month i n Figure 23 (wet vs dry) and the percent variance accounted for by factors 1 & 2 (Table 1) indicate that climatolog i cal monthly means are a useful way to simplify the data into a more manageable format. 6 3

PAGE 76

The temporal patterns are not as clear among the variables using annual means as with the monthly means, but the spatial patterns remain. As with the CMAs, TN, TP, and chlorophyll a are negatively correlated to salinity, however the relationships are weaker The high river flow years of 1983 and 1987 were El Nifio years, and El Nifio years result in above average rainfall for Florida The effect of El Nifio is most realized in the annual salinity and TN values. The decline in TP after 1984 may be a result of declines in inputs from various sources, including water treatment plants, the fertilizer industry, and non-point sources Removal of a significant amount of nitrogen from wastewater is the most likely reason for the decrease in chlorophyll a concentration. If we make the assumption that levels of TN TP and chlorophyll a are indicators of the health of an estuary, Tampa Bay is "healthier'' now that was prior to the early 1980s. One of the goals of this study was to evaluate the HCEPC sampling program on a cursory basis. The question has been raised as to whether too many or too few stations are being sampled by the HCEPC. The question of the timing of the sampling has also been raised and some have suggested a synoptic sampling relative to the tidal cycle While this would certainly be a worthwhile exercise, the logistics of this type of sampling regime on such a large system is burdensome to personnel and financial resources. In light of that reducing the number of stations may ease the burden and make it feasible to pursue a synoptic sampling effort, if only on an irregular basis. This study has shown that there is a certain degree of homogeneity within certain regions of Tampa Bay The tree diagram from the cluster analysis can be used to choose a subset of stations from the current program. Stations that cluster together at a small linkage distance may be represented equally well by a single station. The outliers stations (8, 54, 58) would still be sampled using this technique of choosing sampling stations Regions of high variability, like upper Old Tampa Bay and Hillsborough Bay, will require more samples in order to estimate water quality parameters precisely. 64

PAGE 77

Furthermore, stations near loading sources i.e. rivers and industry, will need to be continued in order to detect changes in the source or potential violations Based on these condit i ons the following stations are recommended in a possible new sampling program : 8 7 44, 80, 54, 52, 73, 58, 13, 32, 51, 38, 67 63, 84, 47, 64, 66, 61, 19, 24, 23, 25, 91, 93 This reduces the number of stations from 52 to 25. An examination of the trends and descriptive statist i cs of this reduced sampling program has shown that the results are not very different from the original 52 station prog r am Obviously, the personnel at the HCEPC that are currently sampling the bay are more familiar with the nuances of sampling Tampa Bay and therefore better able to refine this sampling effort. However, this is the type of exercise that can be done in order to make it possible to perform a synoptic sampling of Tampa Bay. 65

PAGE 78

REFERENCES Boler RN. 1992. Surface water quality 1990-1991, Hillsborough County, Florida Hillsborough County Environmental Protection Commission 209 p. Boler RN. 1995 Surface water quality 1992 1994, Hillsborough County, Florida H i llsborough County Environmental Protection Commission Boler RN Molloy RC, Lesnett EM 1991 Surface water quality monitoring by the Environmental Protection Commission of Hillsborough County. In : proceedings from BASIS II, Treat SF, Clark PA, ed i tors. available from TEXT, Tampa, FL 528 p Boyer JN, Fourqurean JW, Jones RD 1997. Spatial characterization of water quality in Florida Bay and Whitewater Bay by multivariate analyses : zones of similar influence Estuaries 20{4):743-758. Christenson T. 1998. Mesoscale spatial and temporal water quality trends in the Rookery Bay Estuary. MS Thesis University of South Florida, St. Peterburg FL. 157 p Day JW, Jr, Hall CAS, Kemp WM, Yanez-Arancibia A. 1989 Estua r ine Ecology New York : Wiley-lnterscience 558 p Flannery MS 1989. Tampa and Sarasota Bays' watersheds and tributaries In : Estevez ED editor NOAA Estuary-of -the Month Series No. 11. Washington D C. NOAA. 215 p. Fourqurean JW, Jones RD Zieman JC. 1993 Processes influencing water column nutrient characteristics and phosphorus limitat i on of phytoplankton biomass in Florida Bay, FL USA : inferences from spatial distribution Estuar i ne, Coastal and Shelf Science 36 : 295 314. FWPCA. 1969 Problems and management of water quality in Hillsborough Bay, Florida. Hillsborough Bay Technical Assistance Projec t Technical Programs, Southeast Reg i on Federal Water Pollut i on Control Administration 88 p Galperin B, Blumberg AF Weisberg RH. 1991 A time-dependent three-dimensional model of circulation in Tampa Bay In : proceedings from BASIS II, Treat SF, Clark PA, editors. available from TEXT, Tampa, FL. 528 p 66

PAGE 79

Goodwin CR. 1989. Circulation of Tampa and Sarasota Bays In: Estevez ED editor NOAA Estuary-of the-Month Ser i es No 11. Washington, D C NOAA. 215 p Harlin MM Rines HM R o se CE, Abel MB. 1996 Spatial heterogeneity of macrophytes and se l ected invertebrates i n Narragansett Bay (RI, U S A.) as revealed by princi p le component and cluster ana l ys i s Estuarine Coastal and Shelf Sc i ence 42 : 123-134. Johannsen JOR. 1991. Long-term trends of nitrogen load i ng, water quality and biological indicators in Hillsbo r ough Bay, Florida. In: p r oceedings from BASIS II, Treat SF, Clark PA ed i tor s. available from TEXT Tampa FL. 5 2 8 p King Engineering Associates, Inc. 1992 Review and synthesis of historical Tampa Bay water quality data P r epared f or Tampa Bay Nationa l Estuary Program Techn i ca l Publication #07 -92. Lewis RR Whitman RL, Jr. 1985 A new geographic description of the boundaries and subdiv i sions of Tampa Bay . In: p r oceedings from BASIS I Treat SF, Simon JL, Lewis Ill RR, Whitman Jr. RL, editors. Florida Sea G r ant College Tampa FL. Bellwether Press 663 p Lewis RRI Estevez ED 1988 The ecology of Tampa Bay Florida : an estuarine profile U S Fish and Wildli f e Serv Bio i Rep Report 85(7.18) 132 p Palmer SL, McClelland Sl. 1988 Tampa Bay water qua l ity assessment (2050) water quality study} Florida Department of Environmental Regulation 124 p Rodriguez P 1991. Nitrogen enrichment of Tampa Bay and the Little Manatee River phytoplankton populations MS Thesis University of South Flo rida, St. Peterburg FL. 92 p Rote JW 1991 Coastal nonpo i nt pollut i on control. In : proceed i ngs from BASIS II, Treat SF, Clark PA editors available f rom TEXT, Tampa FL. 5 2 8 p SPSS 1996 SPSS fo r Windows Ve r sion 7 0 : SPSS, Inc Vez i na AF Gratton Y V i net P. 1995. Mesoscale phys i cal-b i ological variability during a summer phytop l ankton bloom in t he Lower St. Lawrence estuary Estuarine Coastal and Shelf Science 41 : 393 -411. Weisberg RH, Williams RG. 1991 Init ial finding s on t he circulation of Tampa Bay In : proceedings from BASIS II, Treat SF, Clark PA editors available from TEXT, Tampa FL. 528 p Wooten GR. 1985. Meteoro l ogy of Tampa Bay In: proceedings from BASIS I, Treat SF, Simon JL, Lewis Ill RR, Whitman Jr. RL ed i tors Florida Sea Grant College Tampa, FL. Bellwether Press. 663 p Zarboc k HW. 1991. Past, present and future freshwater inflow to Tampa Bay-effects of a changing watershed In: pro c eedings from BASIS II, Treat SF, Clark PA, editors available from TEXT Tampa FL. 528 p 67

PAGE 80

Zarbock H, Janicki A, Wade 0 Heimbuch 0, Wilson H 1994. E s timates of total nitrogen, total phosphorus, and total suspended solids to Tampa Bay Florida. Prepared by Coastal Environmental Inc for Tampa Bay National Estuary Program. Technical Report #04-94. 68

PAGE 81

APPENDICES 69

PAGE 82

Appendix 1 Descriptive Statistics Table 6 Climatological monthly averages of small clusters Cluster Month Chi a Salinity TN TP WTefTip outliers 1 10.37 26. 38 79 .64 17.33 2 9 17 25. 13 83 .60 18 12 3 11. 59 25 13 .96 .60 20.44 4 13 .61 23.91 89 .62 22.62 5 13.60 25 54 .96 .91 26 .64 6 15.26 27.58 85 76 29 36 7 22. 46 24. 26 1 00 78 30 .19 8 26. 87 23. 17 1 .21 74 30.27 9 18 67 20.60 1.23 .86 28.68 10 18 54 22. 95 95 70 26. 08 11 21.45 24 .01 1 .04 .88 22.36 12 18 17 25.45 85 69 19 77 A1 1 2.87 32.19 .3 9 13 16.00 2 2 83 32 63 44 10 17 .71 3 3.47 33.13 .44 10 19 75 4 3.45 31.64 .46 10 22. 93 5 3.60 33.47 .53 .12 26.14 6 4 .61 35.00 .46 .12 29.29 7 4 .68 31.64 .43 14 30 15 8 5 .69 32 24 .48 15 29.95 9 6 88 30 .63 55 .18 28 86 10 6 87 30 02 .58 20 25.25 11 5 09 31.80 .44 .15 22. 04 12 3 28 30.78 .41 12 18.52 A2 1 3.59 29.56 .47 23 15 92 2 3.94 29.76 .49 .21 17 88 3 5 .43 29.78 48 .22 19. 90 4 4 80 28 78 56 .21 23.03 5 5.46 30 67 .63 .24 26 51 6 5 89 32 50 54 .22 29 .20 7 7.07 29 22 .51 .26 29.86 8 8.50 29.03 .60 26 29 .64 9 9 08 27.37 .67 32 28. 70 10 8.70 28.03 .69 .32 25 08 11 6 09 29.00 .65 .24 21.68 12 4 29 28.24 .47 .24 18 .43 81 1 8 08 23 .61 68 36 17.00 2 6.14 24. 94 62 30 17. 88 3 6 .01 24.85 64 30 19. 30 (continued on next page) DO 7.55 7.21 6.57 6 00 5.44 5 07 4 .66 4 03 3.90 5 30 6.22 7 13 7.89 7 58 7 .41 7.02 6 33 6 .01 5.73 5.86 5 88 6 .27 6 .91 7.32 7.92 7 .53 7.39 6 .81 6 09 5 .60 5.41 5.63 6 09 6 10 6 75 7 33 7 52 7 .43 7 09 70

PAGE 83

Appendix 1 Table 6 (Continued) Cluster Month Chi a Salinity TN TP WTemp DO 81 4 7 .38 24. 38 .72 .30 21.99 6.43 5 9.27 23.17 .88 .36 25.58 6 .08 6 11.71 25.00 .94 .41 28.40 5 31 7 13.48 24.47 .92 .44 29.27 4 .97 8 16.21 23. 37 .92 .46 29.64 5 .04 9 16.65 19 .66 .97 .48 28.16 5 .69 10 14 .55 21.08 .83 .45 25. 11 6 .25 11 11.22 21.53 .83 .43 21.59 6 .91 12 7 .64 2 3.36 .67 .33 18.28 7 .25 82 6 .21 26.36 .56 .35 17. 11 7.86 2 5 .25 27.17 51 .32 17.78 7.75 3 7 35 27.14 .59 33 19.47 7 .59 4 8 .74 26.63 .64 .3 3 22.3 5 7.06 5 8 89 25. 56 .75 .36 26. 0 3 6 .56 6 11 .23 27.84 .76 .39 28.84 5 85 7 12 .81 27.02 .72 .40 29.92 5 93 8 13.66 26.07 .76 .43 30. 08 5 .92 9 14 .20 23.52 .81 .47 28.60 6 33 10 12.59 23.74 .72 .45 25.48 6 .70 11 8 .87 24.35 .69 38 22.06 7 .20 12 5 90 25.52 .54 .32 18.80 7.53 C1 1 10.64 26.88 .67 .52 17.47 8.43 2 10.07 25.29 .73 .52 18 .17 8.11 3 14.13 25.21 87 .56 20.65 7 .72 4 16.65 24.42 87 .60 22.83 7 .36 5 14.58 25.42 .90 .73 26.96 6 .69 6 1 7.44 27.33 .75 .65 29.44 6 .70 7 28.02 24.46 .90 .77 30.45 7 .02 8 27.77 24. 11 1 .09 .78 30.45 6 .26 9 25.86 22.27 1 13 .76 29.11 5 91 10 20. 54 23.60 .89 .64 26.03 6.75 11 17.56 24.24 .88 .58 22.79 7.23 12 11.81 26.63 .66 .47 19.81 7 .72 C2 1 12 .10 26.67 .70 .47 17.17 7 .77 2 10 .25 25.47 .72 .44 17 .59 7 .60 3 11.68 25. 03 .88 .49 20.01 6.95 4 15.24 24.40 .85 51 22. 08 6 .65 5 12 56 25.34 .91 .52 26.23 5 .86 6 16 .54 27. 17 .87 .60 28.68 5.33 (continued on next page) 71

PAGE 84

Appendix 1 (Continued) Table 6 (Continued) Cluster Month Chi a Salinity TN TP WTemp DO C2 7 23 29 24. 50 88 .61 29 76 5 59 8 22. 46 24. 00 1 03 .61 29.73 5 13 9 23 08 21. 12 1 09 6 4 28.68 5 .69 10 18 07 22. 99 90 56 25.64 5 88 11 16 72 24. 10 89 54 22.22 6 75 12 12 .3 6 26.11 .71 42 19 49 7 .2 4 72

PAGE 85

-..J VJ ...-(.) 0 .._ Q) ,_ ::s -+-1 Q) c_ E Q) t-,_ Q) -+-1 co s 34 24 14 Figure 39. Box plot of CMAs of water temperature I I I I ijL ; e Q I 61-ftJ$ '11;.1 I "4 -9 "* I I l 1 2 3 4 5 6 7 8 9 10 11 12 Month Cluster O outliers () )>'"0 '"0 C:Oct> .A2 0 ::J xo. I -u x Ql\) -C]B1 en .82 .C1 0C2

PAGE 86

--.J """ ....::::::::: 0> E -c Q) 0> >< 0 "'C Q) > 0 CJ) CJ) -0 10 -8 6 4 2 0 Figure 40. Box plot of CMAs of dissolved oxygen I I I I I r n I il I I ---l l _ l ___ _ ---.-_____ i ---1 2 3 4 5 6 7 8 9 10 11 12 Month Cluster O outliers )> "'0 "'0 CD ::J L]A1 c.. x I 1\.) .A2 0 0 ::J !:::!". ::J 0B1 c CD c.. .82 .C1 0C2

PAGE 87

-.j Vl -0. a. -c (\l en 40 30 20 10 Figure 41. Box plot of CMAs of salinity I I I \ I I ' B r 8 r E 0 h 9 $ a Cluster O outliers )> "'0 "'0 (!) ::l Cillilll A 1 a. x 1\:) .A2 0 0 ::l !:::!". ::l [2]JB1 c I (!) a. .82 .C1 I I I ......._ 0 C2 1 2 3 4 5 6 7 8 9 10 11 12 Month

PAGE 88

0\ ::::::: 0> E .._ c Q) 0> e ...... z cu ...... 0 ...... 1.6 1 4 1 2 1 0 8 6 .4 2 Figure 42 Box plot of CMAs of total nitrogen I I I I n 8 e l D I 8 J t 8 1 I f ...1 ...! I -1 2 3 4 5 6 7 8 9 10 11 12 Month Cluster D outliers )> "0 "0 (1) ::J a. x 1\) -.A2 () 0 ::J -s c: (]81 (1) a. -.82 .C1 0C2

PAGE 89

Figure 43. Box plot of CMAs of total phosphorus 1.4 I 1.2 I Cluster -:::::: 1.0 0> E .._ (/) ::J 8 L... 0 .r:. a. (/) .6 0 .r:. a.. co .4 .... 0 1-. 2 0 0 I I \ t'l Q \ 1 J 8 1 1 I I r o1 I I j i D outliers )> "'0 "'0 CD :J a. x 1\) -.A2 0 0 :J : :J C]B1 c CD a. .82 .C1 0 C2 1 2 3 4 5 6 7 8 9 10 11 12 MONTH -.1 -.1

PAGE 90

.....,J 00 :::::::: C> ::1. _... (U .r:. a. 0 L.. 0 .r:. (.) 4 0 30 20 1 0 0 Figure 44 Box plot of CMAs of chlorophyll a I -Cluster 1 I D o ut l i ers I n )> -o -o (t) ::J [ZJA 1 a. n t I l -, r D 0 J J I u I f 8' 1 t x 1\) ......... .A2 0 0 ::J ::J c 0 B 1 (t) a. .82 .C1 0C2 1 2 3 4 5 6 7 8 9 1 0 11 1 2 Month

PAGE 91

-..J \0 -() 0 -:::s _. ro I... Q) a. E Q) I-I... Q) _. ro s 27 26 25 24 23 22 21 Figure 45. Box plot of annual averages of water temperature I I I Cluster []outliers )> "'0 "'0 (I) ::J [}TIA1 0.. x 1\) .A2 () 0 I I Q I ::J :; [JB1 c (I) 0.. -. 82 .C1 ' [Jc2 81 82 83 84 85 86 87 88 89 90 91 92 93 Year

PAGE 92

00 0 ::::::::: 0> E -c (I) 0> X 0 "0 (I) > 0 en en 0 10.0 9 0 8 0 7 0 6 0 5 0 4 0 3 0 Figure 46. Box plot of annual averages of dissolved oxygen -I I Cluster I 1 I i I f .f, I I I g IB' I D outliers )> "'0 "'0 (l) :::l a. x 1\) ........ .A2 () 0 :::l !::!: :::l C]B1 c: J (l) a. .82 I -C1 I i 0 C2 81 82 83 84 85 86 87 88 89 90 91 92 93 Year

PAGE 93

Figure 4 7. Box plot of annual averages of salinity 40 I I j 1 1 I Cluster )> "0 "0 (1) ::J a. x 1'\) ........ 0 0 20 ; I : I D 81 ::J !:t. ::J c (1) a. 82 .._.. I I I I I I I I I I I 1 c1 1o I I I I I I I I I I I I I 1 D c 2 I I I I I I I I I I I I I 81 82 83 84 85 86 87 88 89 90 9 1 92 93 Year 00 -

PAGE 94

00 tv Figure 48 Box plot of annual averages of total nitrogen 1 6 I I I I I I I I I I I I I 1.4 I I :a; 1 I I __ 1 0 L i j I 1 11 e e 8 I I rl I 9 ll I +-' z ro 6 .4 j r . : ( e' f j 2 Cluster O outliers E]A1 .A2 0B1 82 .C1 0.0 I I I I I I I I I I I D C2 I I I I I I I I I I I I I 81 82 83 84 85 86 87 88 89 90 91 92 93 Year )> "0 "0 co ::J a. x "' 0 0 ::J ::::!: ::J c: co a. -

PAGE 95

00 w ...-:::::: 0> E -en :::J '-0 .c a. en 0 .c a_ (\] 0 I-2 0 1 5 1.0 5 0 0 Figure 49. Box plot of annual averages of total phosphorus I Cluster --I I -I I I I I I 0 I .. " 1 ; t f i., ,.. ; I ijj ij I I D outliers )> "'0 "'0 ct> ::::l [2JA1 9: X 1\) .A2 () 0 ::::l = ::::l 0B1 c ct> 0. .82 .C1 0 C2 81 82 83 84 85 86 87 88 89 90 91 92 93 YEAR

PAGE 96

00 --.. ::::::::: C) ...._ ro .r::. a. 0 L.. 0 .r::. () 50 40 30 20 10 0 Figure 50. Box plot of annual averages chlorophyll a Cluster O out lier s .. )> "0 ' I I I I J II I I I [ I r I I l I i ' 9 .. 1 ... I J ' f f f r t 9 9' --"0 C1> ::l DA1 0.. x 1\) -.A2 0 0 ::l !::!'. :::1 0B1 c C1> 0.. ........ 82 .C1 0C2 81 82 83 84 85 86 87 88 89 90 9 1 92 93 Year


printinsert_linkshareget_appmore_horiz

Download Options

close
No images are available for this item.
Cite this item close

APA

Cras ut cursus ante, a fringilla nunc. Mauris lorem nunc, cursus sit amet enim ac, vehicula vestibulum mi. Mauris viverra nisl vel enim faucibus porta. Praesent sit amet ornare diam, non finibus nulla.

MLA

Cras efficitur magna et sapien varius, luctus ullamcorper dolor convallis. Orci varius natoque penatibus et magnis dis parturient montes, nascetur ridiculus mus. Fusce sit amet justo ut erat laoreet congue sed a ante.

CHICAGO

Phasellus ornare in augue eu imperdiet. Donec malesuada sapien ante, at vehicula orci tempor molestie. Proin vitae urna elit. Pellentesque vitae nisi et diam euismod malesuada aliquet non erat.

WIKIPEDIA

Nunc fringilla dolor ut dictum placerat. Proin ac neque rutrum, consectetur ligula id, laoreet ligula. Nulla lorem massa, consectetur vitae consequat in, lobortis at dolor. Nunc sed leo odio.