Abstract - Biological and Agricultural Engineering

Transcription

Abstract
BRIGHT, TIFFANY MARIE. An Examination of a Dune Infiltration System’s
Impact on Coastal Hydrology and Bacteria Removal (Under the direction of Dr.
William F. Hunt and Dr. Michael R. Burchell).
The Beaches Environmental Assessment and Coastal Health Act of 2000 (BEACH Act)
requires states to monitor bacteria levels in recreational coastal waters. Increased levels
of bacteria increase the potential for many illnesses to beach goers, so coastal towns are
forced to post advisories or close beaches after many rainfall events, which may
negatively impact tourism. Stormwater outfalls, common in many coastal towns, deliver
stormwater-borne bacteria and other pollutants into the ocean or estuaries.
The NC Department of Transportation and the Town of Kure Beach wanted to
reduce the amount of stormwater discharging on Kure Beach’s recreational beach. Two
stormwater Dune Infiltration Systems were designed to divert a portion of the flow into
the beach dunes. Sand filtration stormwater practices have historically been successful in
bacterial removal. The infiltration systems were constructed using commerciallyavailable open-bottomed infiltration chambers. Due to limited land area, the systems
were designed to infiltrate 1.3 cm storms, which comprised approximately 80% of the
rainfall events at the site. The watersheds of both sites (L and M) were small (1.9 ha and
3.2 ha, respectively) and of mixed urban and residential land use. Water table
measurements indicated a tidal influence, but approximately 2 m of sand was available
for infiltration in the vertical direction.
Data were collected from twenty-five storms during the months of March through
October 2006 to determine the Dune Infiltration System’s viability as a BMP. From
those 25 storms, Site L’s Dune Infiltration System captured total volume of 645 m3 of
stormwater runoff, allowing no runoff to bypass. Site M’s Dune Infiltration System
capacity was exceeded during 20 percent of the storms, capturing a total stormwater
runoff volume of 2313 m3 of the total 2412 m3. At both sites, the Dune Infiltration
Systems significantly (p < 0.01) reduced runoff volume and peak flow discharging
directly onto the beach. Routing the stormwater runoff through the dune’s soil and into
groundwater below did not cause noticeable increase fluctuations in the groundwater.
Bacteria concentrations in the stormwater runoff flowing into the Dune
Infiltration System ranged from 181 CFU/100 ml to 19400 CFU/100 ml with a median of
8600 CFU/100 ml for fecal coliform concentrations and from <10 CFU/100 ml to >2005
CFU/100 ml with a median of 1298 CFU/100 ml for enterococcus. The groundwater
bacteria concentrations were significantly (p<0.01) lower than those of the inflow,
ranging from <1 CFU/100 ml to 214 CFU/100 ml with a median of 1.5 CFU/100 ml for
fecal coliform concentrations and from <10 CFU/100 ml to 2005 CFU/100 ml with a
median of 10 CFU/100 ml for enterococcus concentrations. Groundwater bacteria levels
at Site L never exceeded North Carolina state’s standard; whereas, 23 percent of
groundwater samples from Site M did. The samples that exceeded the standards were
towards the end of the study and associated with relatively large runoff volumes.
A laboratory experiment was performed using Kure Beach’s soil to analyze the
effect bacteria had on infiltration and vice versa. Soil columns were treated with
bacteria- free stormwater or bacteria-spiked (Escherichia coli) stormwater. It was found
at a 95% confidence level that the infiltration rate of the Escherichia coli stormwater
treatment was statistical lower than the bacteria-free columns’ infiltration rate. A
correlation found was between the concentrations of bacteria found in the Escherichia
coli stormwater treatment columns’ effluent to the infiltration rate of these columns.
The Dune Infiltration System’s viability as a stormwater BMP requires continued
research. However, initial results are promising. The Dune Infiltration Systems did
reduce the amount and rate of stormwater directly discharging into the ocean, while
maintaining groundwater hydrology. Specifically, more research is needed to better
understand the Dune Infiltration System’s bacteria removal efficiency. As demonstrated
in the laboratory study, sediment accumulation caused diminished infiltration rates, but
correspondingly decreased the bacteria concentration in the groundwater. The reduction
in infiltrations rates may result in a decrease in the system’s retention volume, potentially
leading to more overflows. The relationship between infiltration rate to bacteria removal
needs to be evaluated when designing the Dune Infiltration System and developing a
maintenance schedule.
An Examination of a Dune Infiltration System’s
Impact on Coastal Hydrology and Bacteria
Removal
By
TIFFANY MARIE BRIGHT
A thesis submitted to the Graduate Faculty of
North Carolina State University
in partial fulfillment of the
requirements for the Degree of
Master of Science
Biological and Agricultural Engineering
Raleigh
2006
APPROVED BY:
________________________________
Dr. William F. Hunt III
Co-Chair of Advisory Committee
_________________________________
Dr. Michael R. Burchell II
Co-Chair of Advisory Committee
________________________________
Dr. Francis de los Reyes III
Member of Advisory Committee
i
Biography
One cold December day in Reykjavik, Iceland, Richard R. Bright II and Janet H. Bright
received one of the four greatest presents in there lives. Did they win the lottery? Did
they finally move from the cold dark country with the highest number of alcoholics?
No…even better, Tiffany Marie Bright greeted the world on the 27th in 1981, saying
“Hello World, It’s Me, T. Marie. Why is it so cold?” She was not supposed to arrive
until the mid January 1982. Having pre-existing knowledge of Chinese astrology, she
realized her personality follows that of a rooster and not of a dog (Tiffany wishes she had
prescient knowledge of graduate school, her future might have been different). The
Rooster is a flamboyant personality, feisty and outwardly confident. The Rooster is also
a trustworthy, hardworking individual. Roosters are happiest when they are surrounded
by others, at a party or just a social gathering. They even enjoy the spotlight and will
exhibit their charisma and wit in a minute (the author could not describe herself better).
The Bright family resided in Iceland for one year after Tiffany was born. Next
they moved to Virginia and Maryland, finally residing in a suburb of Philadelphia,
Chester Springs, Pennsylvania. It seems that each time the Bright’s moved the family
expanded. By the time Tiffany lived in PA, she was blessed with three younger sisters,
Ariel, Tamara, and Rachel.
Tiffany graced Downingtown Senior High School with her presence for four
years. Not only did she excel in the classroom, but also in the athletic department.
Tiffany played field hockey, swam, and ran track. Realizing that not all the world was so
cold, she continued her education at the University of Florida (UF) in Gainesville,
Florida. Well as the saying goes, if you’re not a gator then your gator bait.
ii
Tiffany immensely enjoyed her college years; palm trees on the campus,
knowledge everywhere, and great football. Tiffany started in the materials engineering
track, but soon realized she was not a material girl. She changed her focus to
Agricultural and Biological Engineering, specializing in water resource engineering.
Although Tiffany did not realize it at that time, but that decision was life changing.
Studying water resource engineering allowed her to work on various research projects,
Escondida, Chile, being her favorite one. But more importantly it allowed her to grow as
a person increasing her knowledge in the subject as well as meet interesting people who
will impact her as a person forever.
Upon graduating with honors from UF, Tiffany followed the recommendation of
one of her professors and attended North Carolina State University (NC State) in hopes of
graduation with a Masters of Science. Here Tiffany realized how little she knew and how
much knowledge is available. She focused on her research and classes, but made time for
socializing. Tiffany joined NC State’s water polo team and Engineers Without Borders
as well as started the T-Unit club.
As this thesis proves, Tiffany now has her M.S. in Biological and Agricultural
Engineering! Next Tiffany, along will the other member of the T-Unit, will take a well
deserved 2 month vacation in Europe. The rest of Tiffany’s life, as of now, remains for
the future to know and for her and her love ones to discover.
iii
Acknowledgments
The author would like to acknowledge many of the people who contributed to the success
of this research project. First off the author would like to thank Dr. Dukes for
recommending NC State. The author would like to thank her co-advisors, Dr. Bill Hunt
and Dr. Mike Burchell, for without their guidance, this research project would not have
become a reality. Many thanks for their encouragement and support whenever a
complication arose with the research. The author would like to thank Dr. Francis de los
Reyes for introducing her to the microorganism world and for all his guidance in the lab
study.
The author would also like to especially thank all those who worked directly on
the field research site. Special thanks to Sonny Becker and the Town of Kure Beach
Public Works Department for all there help and support. Also, the author would like to
thank Marc Horstman for his help correctly grabbing samples and downloading data
(good help is hard to find in Wilmington). The author would also like to thank Rachel
Huie, the Oxford Laboratory, JD Potts and the Shellfish Sanitation Laboratory for
analyzing samples.
The author would like to thank each member of the stormwater team. She would
like to thanks Jason Wright for the field help and presentation help. The author wants to
thank Ryan Smith for helping her with GIS and e-mailing her aerial photo links at least 5
times over the past year. The author wants to thank Jon Hathaway for keeping a smile on
her face when she had Sargent issues (what graduate student doesn’t have Sargent
issues?). The author would like to also specially thank Kelly Collins for all her support
as a co-graduate student and friend. She learned a lot from her inside and outside of the
iv
classroom. Also, she wants to thank Smarty Jones for all his knowledge on random
question she had. The author thinks every computer should have a Smarty Jones icon
where people can click and get answers, bet Windows Vista does not have that icon,
shame on you Bill Gates. And shame on you again Bill Gates, why is Word so hard to
format?
The author would also like to thank her fellow graduate students, Jackie Cotter,
Nick Lindow, Dale Hyatt, Bobby Boaz, and officemate Justin Spangler. Thanks for your
smiling faces (most the time) in Weaver. I enjoyed getting to know y’all and sharing my
time here will you. The author would like to especially thank Jenn Johnson and Jodi
Lindgren for all their help and support during graduate school. Tiffany is thankful that
they went to graduate school at NC State, having these two crazy women present during
graduate school allowed for fun times and life long friends.
Special thanks again to Bill for serving not only as the co-advisor, but also as a
mentor and a friend. The author congratulates him and his wife on the addition to their
family. The author sends thanks to all her friends and family for always offering support
through the many months of research and the final months of preparing the thesis.
Last but not least, the author would like to thank the other member of the T-Unit,
Tim Witwer. Without his continuous support and help, the author would have had
difficultly creating a thesis of this caliber. The author looks forward to continually
learning from him. Thanks again, son.
v
TABLE OF CONTENTS
LIST OF TABLES.......................................................................................................... viii
LIST OF FIGURES ............................................................................................................ x
1.0 INTRODUCTION ........................................................................................................ 1
2.0 LITERATURE REVIEW ............................................................................................. 4
2.1 GOVERNMENT’S ROLE IN COASTAL WATER QUALITY ............................................... 4
2.1.1 CLEAN WATER ACT 1972 ................................................................................... 4
2.1.1.1 BEACH Act 2000 ........................................................................................ 5
2.2 COASTAL MICROORGANISM CONTAMINATION .......................................................... 7
2.2.1 PATHOGENS IN POLLUTED WATERS ..................................................................... 7
2.2.1.2 Protozoa Coastal Pathogens .................................................................... 10
2.2.2 INDICATOR BACTERIA....................................................................................... 11
2.3 CAUSES OF COASTAL CONTAMINATION ................................................................... 13
2.3.1 SEWAGE ............................................................................................................ 13
2.3.2 STORMWATER RUNOFF ..................................................................................... 14
2.3.3 BOAT WASTE, WATER FOWL, AND OIL SPILLS ................................................. 16
2.4 NORTH CAROLINA BEACH MONITORING: POSTING CLOSURES AND ADVISORIES .... 17
2.5 ECONOMIC IMPACTS OF COASTAL CONTAMINATION ............................................... 18
2.6 BEST MANAGEMENT PRACTICES (BMPS)................................................................ 20
2.6.1 INTRODUCTION TO BMPS ................................................................................. 20
2.6.2 SAND FILTRATION BMP ................................................................................... 21
2.6.2.1 Introduction to Sand Filtration Systems as BMPs.................................... 21
2.6.2.2 Implementation of Sand Filtration Systems.............................................. 25
2.7 LABORATORY STUDIES ON BACTERIA REMOVAL FROM SAND COLUMNS ............... 29
3.0 HYPOTHESES AND OBJECTIVES ........................................................................ 33
4.0 Dune Infiltration System Field Study......................................................................... 37
4.1 SITE DESCRIPTION ................................................................................................... 37
4.1.1 LOCATION OF DIS............................................................................................. 37
4.2 DIS DESIGN CONSIDERATIONS ................................................................................ 38
4.2.1 DIS PRECONSTRUCTION MONITORING ............................................................. 38
4.2.2 PRECONSTRUCTION SAMPLING PROTOCOL ....................................................... 41
4.2.3 DIS PRECONSTRUCTION FIELD MEASUREMENTS .............................................. 41
4.2.3.1 Site Survey ................................................................................................ 41
4.2.3.2 Single Ring Infiltrometer Test................................................................... 43
4.3 DIS DESIGN ............................................................................................................. 45
4.3.1 HYDROLOGIC CALCULATIONS .......................................................................... 46
4.3.1.1 Rational and Natural Resources Conservation Service Method (NRCS)
Calculations .......................................................................................................... 46
4.3.1.2 Darcy’s equation ...................................................................................... 48
4.3.2 DIS DESIGN ...................................................................................................... 50
4.3.3 DIS INSTALLATION ........................................................................................... 51
4.3.4 DIS MONITORING ............................................................................................. 53
4.3.4.1 Monitoring Equipment.............................................................................. 53
4.3.4.2 Sampling Collection Protocol .................................................................. 58
4.4 DIS RESULTS AND DISCUSSION ............................................................................... 58
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
vi
4.4.1 PRECONSTRUCTION RESULTS AND DISCUSSION ................................................ 58
4.4.2 POST CONSTRUCTION HYDRAULIC DATA ......................................................... 61
4.4.2.1 Summary of Storm Events......................................................................... 61
4.4.2.2 Groundwater Results and Discussion....................................................... 64
4.4.2.3 Flow Mitigation Results and Discussion .................................................. 68
4.4.2.3.1 Site L Results and Discussion ........................................................... 68
4.4.2.3.2 Site M Results and Discussion .......................................................... 71
4.4.2.4 Design Discussion .................................................................................... 77
4.4.3 BACTERIA DATA RESULTS AND DISCUSSION ................................................... 81
4.4.3.1 Summary Results....................................................................................... 81
4.4.3.2 Statistical Analysis and Discussion .......................................................... 83
4.5 DIS SUMMARY ........................................................................................................ 89
5.0 Sand Column Infiltration and Bacteria Laboratory Study.......................................... 93
5.1 EXPERIMENTAL DESCRIPTION .................................................................................. 93
5.1.1 HYPOTHESES..................................................................................................... 94
5.1.2 EXPERIMENTAL VARIABLE CONTROL ............................................................... 94
5.1.3 EXPERIMENTAL MODEL .................................................................................... 98
5.2 EXPERIMENTAL METHOD......................................................................................... 99
5.2.1 VARIABLE CONTROL ........................................................................................ 99
5.2.2 EXPERIMENT PREPARATION .............................................................................. 99
5.2.3 EXPERIMENT PROCEDURE ............................................................................... 100
5.3 EXPERIMENTAL RESULT S AND DISCUSSION .......................................................... 104
5.3.1 E. COLI CULTURE ............................................................................................ 104
5.3.2 INFILTRATION RESULTS AND DISCUSSION ...................................................... 107
5.3.2.1 Pre-Trial Variable Control Results and Discussion............................... 107
5.3.2.2 Trial Infiltration Rates Results and Discussion...................................... 108
5.3.3 BACTERIAL RESULTS AND DISCUSSION .......................................................... 112
5.4 LABORATORY SUMMARY....................................................................................... 120
6.0 Conclusions .............................................................................................................. 122
6.1 Field Study............................................................................................................ 122
6.2 Laboratory Summary............................................................................................ 125
6.3 Overall Recommendations ................................................................................... 128
7.0 Future Research ........................................................................................................ 131
REFERENCES ............................................................................................................... 134
APPENDIX SECTION................................................................................................... 141
A.0 Appendix A-Field Study Storm Summary .............................................................. 142
A.1 Site L Total Summary.......................................................................................... 142
A.2 Site M Total Summary......................................................................................... 143
A.3 Individual Storm Summary.................................................................................. 144
B.0 Appendix B-Field Study Hydrology Statistics ........................................................ 169
B.1 Flow Mitigation-Volume ..................................................................................... 169
B.2 Flow Mitigation-Peak Flow Rate......................................................................... 172
B.3 Correlation Between Rainfall Intensity and Bypass Storms................................ 175
B.4 Correlation Between Peak Inflow Intensity and Bypass Storms ......................... 177
C.0 Appendix C-Field Study Bacteria Statistics ............................................................ 179
C.1 Inflow/Groundwater Fecal Coliform Concentration............................................ 179
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
vii
C.2 Inflow/Groundwater Enterococcus Concentration .............................................. 181
C.3 Groundwater Fecal Concentration Before and After DIS.................................... 183
D.0 Appendix D-Laboratory Infiltration Rate Curves.................................................... 185
E.0 Appendix E-Laboratory MPN Counts ..................................................................... 190
F.0 Appendix F-Laboratory Statistics ............................................................................ 191
F.1 Variation of CSW and T treatment’s to CDI Infiltration Rate ............................. 191
F.2 Variation of CSW and T treatment’s Infiltration Rate ......................................... 193
F.3 Variation of CSW and T treatment’s Bacteria Concentration.............................. 194
F.4 Correlation Between Infiltration Rate and Total Coliform Concentration........... 196
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
viii
LIST OF TABLES
Table 2-1. Pathogens and Swimming-Associated Illnesses (Dorfman 2004). ................. 9
Table 2-2. Major Pollution Sources Causing Beach Closings/Advisories in 2005
According to (Dorfman 2005). ................................................................................. 15
Table 2-3. Percent pollutant removal effectiveness for surface sand filters................... 23
Table 2-4. Typical Pollutant Removal Efficiency in Sand Filters (EPA 1999b). ........... 26
Table 2-5. Summary of Construction Cost Curves, Annual Maintenance Cost Curves
and Surface Area for five Stormwater BMPs in North Carolina, C = Cost in $, x =
Size of watershed in acre, SA = Surface Area in acre (Wossink and Hunt 2003).... 28
Table 4-1. Groundwater bacteria monitoring well specifications. .................................. 40
Table 4-2. Preconstruction groundwater fecal coliform levels. ...................................... 59
Table 4-3. Preconstruction stormwater runoff bacteria levels. ........................................ 59
Table 4-4. Site L Storm Characteristics. ........................................................................ 62
Table 4-5. Site M Storm Characteristics. ....................................................................... 63
Table 4-6. Site M summary result of bypassing storms. ................................................. 72
Table 4-7. Maximum stage in bypass storm in Site M’s chambers. .............................. 79
Table 4-8. Summary of Fecal Coliform levels for the 25 storms................................... 81
Table 4-9. Summary of Enterococcus levels for 22 storms. .......................................... 82
Table 5-1. Influent E. coli concentrations for each trial................................................ 107
Table 5-2. Average treatment infiltration rate per column........................................... 109
Table 5-3. E. coli concentration measured for T and CSW treatment for four trials using
Colilert™ (Standardized SM 9223) testing method. .............................................. 112
Table 5-4. Average total coliform concentration in T and CSW treatment per trial..... 114
Table 5-5. List of coliforms in the Enterobacteriacea family (Leclerc et al. 2001). ..... 119
Table A-1. Summary Table of the 25 Storm Events at Site L....................................... 142
Table A-2. Summary Table of the 25 Storm Events at Site M. .................................... 143
Table A-3. March 21, 2006 Storm Summary. ............................................................... 144
Table A-4. April 17, 2006 Storm Summary. ................................................................. 145
Table A-5. April 27, 2006 Storm Summary. ................................................................. 146
Table A-6. May7, 2006 Storm Summary. ..................................................................... 147
Table A-7. May 14, 2006 Storm Summary. .................................................................. 148
Table A-10. June 5, 2006 Storm Summary. .................................................................. 151
Table A-11. June 12, 2006 Storm Summary. ............................................................... 152
Table A-12. June 14, 2006 Storm Summary. ................................................................ 153
Table A-16. July 6, 2006 Storm Summary.................................................................... 157
Table A-17. July 16, 2006 Storm Summary.................................................................. 158
Table A-21. August 21, 2006 Storm Summary............................................................. 162
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
ix
Table A-22. August 23, 2006 Storm Summary............................................................. 163
Table A-23. September 1, 2006 (Tropical Storm Ernesto) Storm Summary. ............... 164
Table A-24. September 5, 2006 Storm Summary. ........................................................ 165
Table A-25. September 14, 2006 Storm Summary. ...................................................... 166
Table A-26. October 8, 2006 Storm Summary. ............................................................ 167
Table A-27. October 18, 2006 Storm Summary. .......................................................... 168
Table D-1. Trial 1 infiltration times for each column ................................................... 185
Table D-10. Trial 10 infiltration times for each column ............................................... 187
Table E-1. Number of Positive Tubes per Treatment ................................................... 190
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
x
LIST OF FIGURES
Figure 2-1. Plan view and profile of Austin sand filter (City of Boise Public Works
Professional Advisory Group 1998). ........................................................................ 22
Figure 2-2. Plan and profile view of Delaware sand filter (City of Boise Public Works
Professional Advisory Group 1998). ........................................................................ 24
Figure 2-3. Plan view and profile of underground sand filter (City of Boise Public
Works Professional Advisory Group 1998).............................................................. 25
Figure 4-1. Map of NC illustrating the location of New Hanover County and the Town
of Kure Beach. .......................................................................................................... 37
Figure 4-2. (a) Site L Watershed Area (b) Site M Watershed Area. ............................... 38
Figure 4-3. Installation of preconstruction groundwater well......................................... 39
Figure 4-4. Pre-construction hydrology monitoring wells. ............................................. 40
Figure 4-5. Survey of Site L and Site M in Kure Beach, NC (feet). ............................... 42
Figure 4-6. Actual Survey Elevations of Site L and M Kure Beach, NC (feet).............. 43
Figure 4-7. Site L’s cumulative infiltration versus time for three single ring infiltrometer
tests. .......................................................................................................................... 45
Figure 4-8. Sites L’s and M’s estimated inflow hydrograph for a 12.5 mm/hr (0.5 in/hr)
storm. ........................................................................................................................ 48
Figure 4-9. StormChambers™ schematic (courtesy Hydrologic Solutions Inc.)............ 49
Figure 4-10. Top view of DIS layout. ............................................................................. 51
Figure 4-11. Installation of a StormChamber™.............................................................. 52
Figure 4-12. Planting Sea oats in Site M’s dunes............................................................ 53
Figure 4-13. AutoCAD drawing of Site L and Site M vault (feet unless otherwise noted).
................................................................................................................................... 53
Figure 4-14. View of monitoring vault from manhole. ................................................... 55
Figure 4-15. Sargent pulley tape system. ........................................................................ 56
Figure 4-16. ISCO sampler in JOBOX. .......................................................................... 57
Figure 4-17. DIS system at Site L after installation. ....................................................... 57
Figure 4-18. Preconstruction groundwater elevations at Site L. ..................................... 60
Figure 4-19. Preconstruction groundwater elevations at Site M. .................................... 61
Figure 4-20. Rainfall intensity versus rainfall amount.................................................... 64
Figure 4-21. Site L groundwater fluctuations from July to October 2005 and 2006. .... 65
Figure 4-22. Site M groundwater fluctuations from July to October 2005 and 2006. .... 65
Figure 4-23. Wrightsville Beach tidal influences on groundwater elevations in Kure
Beach, NC. ................................................................................................................ 66
Figure 4-24. Site L and Site M fluctuations in groundwater since DIS implementation. 68
Figure 4-25. Volume of runoff captured Site L............................................................... 69
Figure 4-26. Site L peak inflow per storm. ..................................................................... 70
Figure 4-27. Site L Tropical Storm Ernesto and October 8, 2006 inflow hydrograph. .. 71
Figure 4-28. Volume of runoff captured versus overflow per storm at Site M............... 72
Figure 4-29. Site M inflow hydrograph, stage in vault and stage in StormChambers
during Tropical Storm Ernesto (8/31/06-9/01/06). ................................................... 73
Figure 4-30. Peak inflow rate for captured and bypass storm at Site M. ........................ 74
Figure 4-31. Peak inflow rate versus peak outflow rate per storm at Site M. ................. 74
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
xi
Figure 4-32. Site M comparison of June 14, 2006 and September 13, 2006 inflow
hydrographs............................................................................................................... 75
Figure 4-33. Peak rainfall intensity versus rainfall amount for captured and bypassed
storms for Site M. ..................................................................................................... 76
Figure 4-34. Variation in runoff volume for Site L and Site M. ..................................... 77
Figure 4-35. Variation in peak inflow rate for Site L and Site M. .................................. 77
Figures 4-36. (a) Site L semi-log fecal coliform concentration (b) Site L semi-log
enterococcus concentration (c) Site M semi-log fecal coliform concentration (d) Site
M semi-log enterococcus concentration during 2006............................................... 85
Figure 4-37. Semi-log of Site L’s groundwater enterococcus concentration and volume
of runoff per storm event. ......................................................................................... 86
Figure 4-38. Semi-log of Site M’s groundwater enterococcus concentration and volume
of runoff per storm event. ......................................................................................... 87
Figure 4-39. SAS output for Site L of fecal coliform groundwater concentration before
DIS (square symbol) and after (plus symbol). .......................................................... 88
Figure 4-40. SAS output for Site M of fecal coliform groundwater concentration before
DIS (square symbol) and after (plus symbol). .......................................................... 88
Figure 5-1. Kure Beach’s soil particle size distribution.................................................. 95
Figure 5-2. Initial column set-up, allowing 2 L of DI water to compact the column .... 97
Figure 5-3. Finishing construction the columns by adding stone to the columns. ........... 97
Figure 5-4. Final sand column design. ............................................................................ 98
Figure 5-5. (a) Extracting 6 ml of E. coli culture to inculcate sterilized stormwater in the
1.5 L beaker. (b) Sterilizing E. coli culture flask................................................... 101
Figure 5-6. (a) Inoculating sterilized stormwater with bacteria stormwater. (b)
Measuring 2 L of bacteria stormwater in 2 L cylinder. (c) Pouring treatment in the
sand column. ........................................................................................................... 102
Figure 5-7. Timing the water front movement to various table levels. ......................... 103
Figure 5-8. OD650 of Grown E. coli cultures over 13 day period................................ 105
Figure 5-9. Calculating generation time from OD curve. ............................................. 106
Figure 5-10. Average wetting front advancement rate curve for each treatment.......... 108
Figure 5-11. Graph of average treatment infiltration rate. ............................................ 110
Figure 5-12. Graph of variation of infiltration rate for treatment T and CSW.............. 111
Figure 5-13. Semi-log plot of total coliform concentration for each trial. .................... 115
Figure 5-14. Total coliform concentrations versus infiltration rates for treatments T and
CSW........................................................................................................................ 116
Figure 5-15. Semi-log plot of total coliform and E. coli concentration per trial........... 117
Figure A-1. Site L Inflow Hydrograph and Rainfall Amount for March 21, 2006 Storm.
................................................................................................................................. 144
Figure A-2. Site M Inflow Hydrograph and Rainfall Amount for March 21, 2006 Storm.
................................................................................................................................. 144
Figure A-3. Site L Inflow Hydrograph April 17, 2006 Storm, Rainfall Not Available. 145
Figure A-4. Site M Inflow Hydrograph April 17, 2006 Storm, Rainfall Not Available.
................................................................................................................................. 145
Figure A-5. Site L Inflow Hydrograph April 27, 2006 Storm, Rainfall Not Available. 146
................................................................................................................................. 146
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
xii
Figure A-7. Site L Inflow Hydrograph and Rainfall Amount for May 7, 2006 Storm. 147
Figure A-8. Site M Inflow Hydrograph and Rainfall Amount for May 7, 2006 Storm. 147
Figure A-9. Site L Inflow Hydrograph and Rainfall Amount for May 14, 2006 Storm.
................................................................................................................................. 148
Figure A-10. Site M Inflow Hydrograph and Rainfall Amount for May 14, 2006 Storm.
................................................................................................................................. 148
................................................................................................................................. 149
................................................................................................................................. 149
................................................................................................................................. 150
................................................................................................................................. 150
Figure A-15. Site L Inflow Hydrograph and Rainfall Amount for June 5, 2006 Storm.
................................................................................................................................. 151
Figure A-16. Site M Inflow Hydrograph and Rainfall Amount for June 5, 2006 Storm.
................................................................................................................................. 151
................................................................................................................................. 152
................................................................................................................................. 152
................................................................................................................................. 153
................................................................................................................................. 153
................................................................................................................................. 154
Figure A-22. Site M Inflow Hydrograph and Rainfall Amount for June 25, 2006 Storm
................................................................................................................................. 154
................................................................................................................................. 155
................................................................................................................................. 155
................................................................................................................................. 156
................................................................................................................................. 156
Figure A-27. Site L Inflow Hydrograph and Rainfall Amount for July 6, 2006 Storm.
................................................................................................................................. 157
Figure A-28. Site M Inflow Hydrograph and Rainfall Amount for July 6, 2006 Storm.\
................................................................................................................................. 157
................................................................................................................................. 158
Figure A-30. Site M Inflow Hydrograph and Rainfall Amount for July 16, 2006 Storm.
................................................................................................................................. 158
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
xiii
................................................................................................................................. 159
................................................................................................................................. 159
................................................................................................................................. 160
................................................................................................................................. 160
................................................................................................................................. 161
................................................................................................................................. 161
Figure A-37. Site L Inflow Hydrograph and Rainfall Amount for August 21, 2006
Storm....................................................................................................................... 162
Figure A-38. Site M Inflow Hydrograph and Rainfall Amount for August 21, 2006
Storm....................................................................................................................... 162
Storm....................................................................................................................... 163
Storm....................................................................................................................... 163
Figure A-41. Site L Inflow Hydrograph and Rainfall Amount for September 1, 2006
(Tropical Storm Ernesto) Storm.............................................................................. 164
Figure A-42. Site M Inflow Hydrograph and Rainfall Amount for September 1, 2006
(Tropical Storm Ernesto) Storm.............................................................................. 164
Storm....................................................................................................................... 165
Figure A-44. Site M Inflow Hydrograph and Rainfall Amount for September 5 2006
Storm....................................................................................................................... 165
Storm....................................................................................................................... 166
Storm....................................................................................................................... 166
Figure A-47. Site L Inflow Hydrograph and Rainfall Amount for October 8, 2006 Storm.
................................................................................................................................. 167
Figure A-48. Site M Inflow Hydrograph and Rainfall Amount for October 8, 2006
Storm....................................................................................................................... 167
Figure A-49. Site L Inflow Hydrograph and Rainfall Amount for October 18, 2006
Storm....................................................................................................................... 168
Storm....................................................................................................................... 168
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
1
1.0 INTRODUCTION
Coastal areas, which comprise only 17 percent of the land area in the United State are
host to over 50 percent of the total U.S. population. According to the Natural Resource
Defense Council (NRDC), the coastal population grew by 37 million people between
1970 and 2000, and by 2015 is projected to increase by another 21 million (Dorfman
2004). Urban development increases stormwater runoff, while limiting the amount of
available land that can be used to treat the stormwater. Stormwater runoff may contain
pollutants such as hydrocarbons, nutrients, metals, bacteria, pathogens, and sediment. To
control the amount of bacteria entering the ocean, the US Congress passed the Beach
Environment Assessment and Coastal Health (BEACH) Act, which required states to
monitor bacteria levels in recreational coastal waters and to post advisories of closures if
a state’s bacteria standards are exceeded. The bacteria standard was initially fecal
coliform, but many coastal states are now choosing enterococcus.
North Carolina spends approximately $550,000 annually to operate the water
quality monitoring program for its coastal recreational waters to protect the public safety
of residents and more than 6.5 million tourists that utilize the state’s beaches each year.
Although these beaches are being monitored, they are still threatened by pollution from
agricultural, septic system, and development runoff (Dorfman 2004).
Baker et al. (2005) conducted a study using data from two California sites,
Newport and Huntington Beach, to estimate the economic impact of illnesses associated
with polluted recreational waters. It was found that recreational swimming at these two
beaches cost the public $3 million per year in health related expenses. Another negative
2
effect of stormwater contamination is the loss of income associated with tourism.
According to the U.S Environmental Protection Agency (2000), $44 billion dollars were
spent on coastal tourism in that year. National Resource Defense Council’s 2004 Testing
the Waters Report indicates that there were 2,635 beach closures and advisories in the
U.S due to increased ocean bacterial levels associated with stormwater runoff. In 2004,
North Carolina had 555 beach closures or advisories, which equated to 20% of the U.S
closures (Potts 2005). As North Carolina’s coastal population continues to increase,
stormwater must be mitigated to reduce the potential of human exposure to bacteria and
other pathogens in order to prevent the more serious problems that have been
documented on the west coast.
The Town of Kure Beach is located in New Hanover County, south of
Wilmington, North Carolina. The North Carolina Department of Transportation
(NCDOT) and the Town of Kure Beach sought a technology to reduce the amount of
runoff entering Kure Beach’s recreational swimming areas. Stormwater outfalls,
common in many coastal towns, discharge stormwater and associated bacteria and
pollutants directly into the ocean. Greenberg (1956), Carlucci and Pramer (1959) and
Mitchell (1968) concluded that die-off of coliforms in marine waters is a fairly rapid
event that is controlled by a variety of factors, including toxicity due to high salt
concentrations, predation, competition by native microflora, heavy metals, and limited
nutrient supply. Typical die-off curves for Escherichia coli (E. coli) in seawater show an
initial lag phase followed by a mortality of up to 90% in 3 to 5 days (Gerba and McLeod
1975). Even though bacteria will eventually die-off, they pose an initially threat if
stormwater discharges into swimming areas. Thus, after rain events if the bacteria count
3
exceeds states’ standards, communities must temporarily post advisories or close
beaches.
In order to capture stormwater runoff, decreasing the public’s contact with
bacteria and, therefore, maintaining beach revenue, a Dune Infiltration System (DIS) was
implemented to demonstrate and research the potential of the system’s technology to treat
small to mid-sized rainfall events that frequently occur along the North Carolina Coast.
An additional laboratory study was conducted to estimate the effect of infiltration rates
on bacterial removal in this type of system to develop maintenance criteria for the DIS.
4
2.0 LITERATURE REVIEW
The first portion of this chapter describes government standards and regulations for
coastal water quality. The next segment elucidates causes, effects, and monitoring
practices of beach contaminations. Then Best Management Practices (BMPs) role in
managing coastal environments is described, with specific details on sand filtration BMP
design and implementation in coastal areas. The final section discusses bacteria removal
efficiency of sand filtration as tested in laboratory experiments.
2.1 GOVERNMENT’S ROLE IN COASTAL WATER QUALITY
2.1.1 CLEAN WATER ACT 1972
The United States Environmental Protection Agency (EPA) implemented the Federal
Water Pollution Control Act Amendments of 1972 to increase public awareness and
concern for controlling water pollution. As amended in 1977, the law became known as
the Clean Water Act (CWA) and established regulatory pollutant standards in the United
States. The CWA established enforceable water quality standards for contaminants in
surface waters and recognized the need to address the problems posed by non-point
source pollution. This also gave the EPA the authority to employ pollution control
programs (US EPA 2006b). The CWA is the United States’ basis for water quality
protection since it utilizes a variety of regulatory and non-regulatory provisions to reduce
point source pollution into waterways, finance municipal wastewater treatment facilities,
and manage polluted runoff (US EPA 2002).
5
Beginning in the late 1980s, efforts to address polluted runoff have increased
substantially. Evolution of CWA programs over the last decade has also included a shift
from a program-by-program, source-by-source, pollutant-by-pollutant approach to more
comprehensive watershed-based strategies. In this approach, equal resources are devoted
to both protecting healthy waters and restoring impaired ones. A full array of issues are
addressed, not just those subject to CWA regulatory authority (US EPA 2003). One
example is the safety of coastal waters for swimming and recreational activities. To
protect coastal recreation waters, the EPA has published scientifically justified limits for
a range of pollutants in coastal waters, known as the protective criteria for coastal waters.
Individual states are responsible for writing their own legal standards for pollutants and
adopting the protective criteria, pending EPA approval. The states can do this by: (1)
adopting the EPA’s recommended criteria, (2) modifying the EPA’s recommended
criteria to reflect site-specific conditions, or (3) adopting criteria that are as protective as
the EPA’s recommendation based on scientific methods (US EPA 2006a).
2.1.1.1 BEACH Act 2000
As of 2000, many states had not adopted the recommended federal bacteria criteria for
monitoring E. coli and/or enterococcus bacteria levels. In response, Congress passed the
Beach Environment Assessment and Coastal Health (BEACH) Act on October 10, 2000,
giving states until April 2004 to adopt protective bacteria criteria into their state
standards. For states that did not meet the deadline, Congress required EPA to issue
federal standards to ensure national protection (US EPA 2006a). The EPA's published
standards include a geometric mean value for multiple samples taken over 30 days and an
instantaneous single sample value. Based on these measures, local authorities should
6
issue beach closings or advisories if either standard is exceeded. Many states, only use
one measure, either the geometric mean or the single. North Carolina’s standard is
currently a single maximum of enterococci (Dorfman 2004). The EPA standards are
based on how often the beach is used. The categories and the standards for various
beaches are as follows:
•
Tier 1 - These beaches are used on a daily basis and must conform to a single
sample maximum 104 enterococci per 100 ml water or a geometric mean of 35
enterococci per 100 ml water.
•
Tier 2 - These beaches are used an average of four times per week and are
considered useable with less than a single sample maximum of 276 enterococci
per 100 ml water.
•
Tier 3 - These beaches are used an average of no more than twice a month, but
more intensively for special events such as triathlons. These beaches adhere to a
single sample maximum for these sites are 500 enterococci per 100 ml of water
(Potts 2005).
The BEACH Act funding contributes to research for protecting coastal
recreational waters. The EPA has developed an expedited laboratory test for enterococci.
This improved test produces results in 24 hours rather than the 48 hours currently
required for existing E. coli test methods. Also, the EPA works with other agencies at all
levels of government to develop and validate predictive models for gauging where and
when beach pollution is likely to occur. The goal is to assist public health officials to
determine when warnings may be necessary to alert beach goers of potential problems
during and immediately following a storm or other pollution event. Furthermore the
7
EPA-sponsored research improves the scientific efforts in support of local, state, and
tribal actions to protect public health at bathing beaches (US EPA 2006b). The BEACH
Act also directs the EPA to study issues associated with pathogens and public health and
to publish new or revised criteria based on that study by October, 2005, and every five
years thereafter. States must then adopt these new or revised criteria (Dorfman 2005).
2.2 COASTAL MICROORGANISM CONTAMINATION
2.2.1 PATHOGENS IN POLLUTED WATERS
While the BEACH Act established bacterial concentration standards to protect coastal
recreational waters, these standards may not be enough to ensure swimmer safety. A
recent study by Griffin et al. (2003) concluded, “…a majority of pathogens responsible
for outbreaks of human illnesses acquired from marine recreational exposure have not
been identified.” A study conducted in 2000 by the Department of Environmental
Analysis and Design, University of California, found human adenoviruses in 4 of 12
samples taken at the mouths of major rivers and creeks on beaches from Malibu, CA to
the United States border of Mexico in February and March 1999 (Jiang et al. 2001).
Researchers also tested for the bacterial indicators commonly used for beach water
monitoring including total coliform, fecal coliform, and enterococci, but found no
correlation of these indicators to adenoviruses. The study recommended that current
recreational water quality standards be modified to account for the presence of viruses
and that regular monitoring for human viruses be conducted on a regular basis (Griffin et
al. 2003).
8
Polluted waters may contain several different types of disease causing pathogens,
specifically bacteria, virus, and protozoa. According to the Natural Resources Defense
Council (NRDC 2004), 88% of beach closing and advisories in 2003 were due to
detected bacteria levels that exceed recreational coastal waters standards. Six percent
were from precautionary warnings due to rainfall known to carry pollution to swimming
water and 4% were in response sewage treatment plant failure and breaks in sewage
pipes, both of these causes polluting the water with bacteria and virus pathogens. The
last 2% were due to dredging problems and algal blooms.
2.2.1.1 Viral and Bacterial Coastal Pathogens
Table 2-1 list common coastal bacteria and viral water pathogens and the illness
associated with them. Research has shown that fecal-oral viral pathogens present various
health concerns. Ocean goers exposed to bacteria-enriched recreational waters have
symptoms ranging from asymptomatic to severe gastrointestinal, respiratory, and eye,
nose, ear, and skin infections. The two most common fecal-oral viral pathogens are
adenoviruses and Norwalk viruses. Adenoviruses are commonly found in wastewaterimpacted marine environments and can cause acute upper respiratory tract infections as
well as ocular and gastrointestinal infections. Norwalk-like viruses (small round
structured viruses [SRSV]) are a major cause of shellfish-associated disease and may be
the most significant cause of adult viral gastroenteritis (Griffin et al. 2003). Other
microbial bacteria diseases that can be contracted by swimmers include salmonellosis,
shigellosis, and infection caused by E. coli (Dorfman 2005).
9
In a 1995 large-scale epidemiological study, the Santa Monica Bay Restoration
Project investigated possible adverse health effects associated with swimming in ocean
waters contaminated by urban runoff. This study confirmed the increased risk of illness
associated with swimming in areas with high densities of indicator bacteria.
Table 2-1. Pathogens and Swimming-Associated Illnesses (Dorfman 2004).
Pathogenic Agent
Disease
Bacteria:
Campylobacter jejunii
Gastroenteritis
E. coli
Gastroenteritis
Salmonella typhi
Typhoid fever
Other salmonella species
Various enteric fevers, gastroenteritis, septicemia
Shigella dysenteriae and
other species
Bacterial dysentery
Vibrio cholera
Cholera
Yersinia spp.
Acute gastroenteritis (including diarrhea, abdominal pain)
Viruses:
Adenovirus
Respiratory and gastrointestinal infections
Coxsackievirus (some
strains)
Various, including severe respiratory diseases, fevers, rashes,
paralysis, aseptic meningitis, myocarditis
Echovirus
Various, similar to coxsackievirus
Hepatitis
Infectious hepatitis (liver malfunction); also may affect kidneys
and spleen
Norwalkvirus
Gastroenteritis
Poliovirus
Poliomyelitis
Reovirus
Respiratory infections, gastroenteritis
Rotavirus
Gastroenteritis
The Santa Monica Bay Restoration Project study began with initial interviews of 15,492
beachgoers who swam and immersed their heads, followed by interviews with 13,278 to
10
determine the occurrence of certain symptoms such as fever, chills, nausea, and diarrhea.
Water samples were taken and analyzed for total and fecal coliform, enterococcus, and E.
coli. Water samples were also collected at stormdrains and analyzed for enteric viruses.
The study found an increase in risk of illness (with symptoms including fever, chills, ear
discharge, and vomiting) associated with swimming near flowing stormdrain outlets in
Santa Monica Bay, compared with swimming more than 400 yards away. Swimmers near
stormdrains were found to have a 57 % greater incidence of fever than those swimming
farther away. This study also confirmed the increased risk of illness associated with
swimming in areas with high densities of indicator bacteria. Illnesses were reported more
often on days when the samples were positive for enteric viruses (Haile et al. 1996).
One of the primary concerns of public health officials is the relationship between
the presence of pathogens and the recreational risk to human health in polluted marine
environments. While a number of studies have attempted to address this issue, the
relationship is still poorly understood. A contributing factor to the slow progress in the
field has been the lack of methods sensitive enough to detect the broad range of both
bacterial and viral pathogens (Griffen et al. 2003).
2.2.1.2 Protozoa Coastal Pathogens
Phytoplankton, a type of protozoa, are coastal water microscopic organisms that form the
basis of the marine food web. Sixty-three of the thousands of species of phytoplankton
are known to be toxic to animals. High concentrations of phosphorus and nitrogen that
enter the ocean via sewage discharge or stormwater, artificially stimulate phytoplankton
population. The result is rampant multiplication with resultant blooms that can last for
days or months. Depending on the type of toxic organism, ocean swimmers exposed to
11
the toxic algae can experience illnesses ranging from respiratory problems and eye
irritation to neurotoxic poisoning that can cause short-term memory loss, dizziness,
muscular aches, peripheral tingling, vomiting, and abdominal pain (Bushaw-Newton and
Sellner 1999).
Although the most common consumer health impact of toxic blooms arises from
eating contaminated shellfish, there are numerous instances, which such blooms have
directly affected fishermen, swimmers and other recreational users of nearshore marine
waters. Toxic outbreaks of such organisms as Pfiesteria piscicida, which was first
discovered in North Carolina in 1991, have been found to be associated with fish kills
and with skin and neurological damage as well as memory loss (Trainer 2002). Red-tide
algal blooms of Gymnodinium brevii affected west coast beaches of Florida, which
resulted in many respiratory illnesses for ocean goers in the year 1996, 1999, and 2005
(Bushaw-Newton and Sellner 1999). Other outbreaks occurred in California in 2000,
Texas in 2004, and North Carolina in 1987 and 1988 (Tester et al. 1991).
2.2.2 INDICATOR BACTERIA
Research studies conducted during the past decades demonstrate a strong relationship
between the amount of indicator bacteria in coastal water and the incidence of
swimming-associated illnesses. Common indicator bacteria are total and fecal coliform,
enterococcus, and E. coli, the later two being the most common. E. coli is defined as
“gram-negative, facultative anaerobic, nonspore-forming bacillus commonly found in the
intestinal tracts of humans and other warm-blooded animals…Escherichia coli is
considered the primary indicator of recent fecal pollution” (Symons and Bradley 2001).
Enterococcus genus is defined by North Carolina Shellfish Sanitation and Recreational
12
Water Quality Section of the North Carolina Department of Environment and Natural
Resources (NCDENR) as “a gram-positive coccoid-shaped bacteria that is found in the
intestinal tracts of warm-blooded animals that include Enterococcus faecalis,
Enterococcus faecium, Enterococcus avium, and Enterococcus gallinarium” (Potts 2005).
E. coli is still being used in some states as indicator bacteria, but the EPA recommends
enterococci for the indicator bacteria for recreational coastal waters. Many studies
proved that as indicator bacteria, E. coli has a shorter persistence in the environment
when compared with enterococcus (Lleo et al. 2005 and Bordalo et al. 2002). A study in
Britain specifically proves that from a range of possible bacterial indicators only fecal
streptococcus (a subset of enterococcus) was an accurate indicator for gastroenteritis.
The authors suggest that, “fecal streptococci do seem to be a better indicator of water
quality than the traditional coliform counts. Bathing water standards should be revised
with these findings in mind” (Kay et al. 1994). Also, Haile et al. (1996) in their Santa
Monica Bay study found that neither fecal nor total coliform by itself is an accurate
indicator.
Even though indicator bacteria may not be directly harmful to humans, they are
relatively easy to test for and are typically found in the presence of more harmful
pathogens; however, the effectiveness of bacterial indicators as predictors of viral
contamination is questionable (Griffin et al. 2003).
Another problem with using microorganisms as indicators of fecal contamination is
the 24-hour lag time between sample collections and test results. In the meantime, ocean
goers may be exposed to contaminated water. Scientists are researching non-biological
indicators that may eventually replace or supply conventional indicators to provide
13
instantaneous results. This includes testing for caffeine concentration in sewage
contamination or using chemical fluorescence techniques to detect fecal contamination on
processed meat products (Buerge 2003).
2.3 CAUSES OF COASTAL CONTAMINATION
Recreational coastal water can be contaminated by polluted storm water runoff, sewer
line breaks, sewage spills and overflows, waste from domestic animals, marine mammals
and birds, poorly maintained septic systems, boat waste, and oil spills.
2.3.1 SEWAGE
According to Potts (2005) there are no ocean sewage outfalls or combined sewer
overflows in coastal North Carolina. Sewage treatment plants in North Carolina typically
discharge to rivers which in turn lead to the Atlantic. But sewage can still contaminate
coastal waters through combined sewer overflows, sanitary sewer overflow, sewage line
breaks, and sewage treatment plant malfunctions.
Sanitary sewers contaminate coastal waters when they overflow due to rain
overload or have a line break caused from old, inadequately maintained, or breached
sewage lines. The EPA has estimated between 23,000 and 75,000 sanitary sewer
overflows (SSO) occur annually, discharging a total of 3 billion to 10 billion gallons per
year. An estimated 1.8 to 3.5 million people contract gastroenteritis each year from
swimming in raw sewage (EPA 2004). In September, 1999, Hurricane Floyd caused
sewage to contaminate the mouth of the Neuse River, in North Carolina. According to
Paerl et al. (1999), low oxygen levels as well as high bacteria levels suffocated fish,
shellfish, and the smaller organisms on which they feed (The Associated Press 1999).
14
In January 2005, several beaches in Long Beach and Orange Counties, California,
were closed due to sewage contamination. A series of powerful rainstorms sent 2.4
million gallons of raw sewage from Long Beach into the Los Angles River and 4 million
gallons from Orange County into the Santa Ana River. Accompanying the sewage was
several hundred of thousands of gallons of bovine effluent. Consequently, enterococcus
levels rose 10 times higher than the CA’s standards (Chong and Wride 2005).
In addition to threatening humans, harmful bacteria negatively impact the ocean
ecosystem. In the Florida Keys and the Caribbean, fecal contamination from sewage is
thought to be a major source of disease in the surrounding coral, causing more then 90%
of Elkhorn coral to die over the past decade (National Marine Fisheries Service 2006).
According to a U.S. Virgin Island study, raw sewage discharged into the ocean killed
coral reefs at elevated rates. In the study, 30% of the coral exposed to raw sewage was
infected with two coral diseases that can kill a foot-long colony in a week; whereas, the
coral that was not exposed to sewage had infection rates of 3-4% (Probacso 2005).
2.3.2 STORMWATER RUNOFF
Stormwater runoff is also recognized as an important beach pollutant source, resulting in
elevated bacteria levels. Almost every coastal and Great Lakes state reported at least one
beach where stormwater drains onto or near bathing beaches. Stormwater is created
when rain or snowmelt travels on pervious and impervious areas, dissolving contaminants
and carrying them from their origin. Common contaminants include oil, grease, heavy
metals, pesticides, litter, fecal matter from pets and other urban animals, and pollutants
from vehicle exhaust. Even though separate storm sewer systems are designed to carry
15
only stormwater, human sewage can enter through leaks in adjacent sewage pipes or from
sewage pipes that are illegally hooked up to the stormdrains (Dorfman 2004).
As reported by the EPA (1998) about a quarter of our nation’s polluted estuaries
and lakes are fouled by urban stormwater. There are 17 stormdrains in North Carolina
that discharge directly into the ocean waters (Potts 2005). Urban stormwater was the
number one cause of known beach closings and advisories in 2004 and 2005 (Table 2-2).
Table 2-2. Major Pollution Sources Causing Beach Closings/Advisories in 2005
According to (Dorfman 2005).
Pollution Source
Number of Closings/Advisories
Elevated bacteria levels of unknown
14,602 days plus 69 extended and 39
origin
permanent events
Stormwater runoff
permanent events
Sewage spills and overflows
898 days plus 7 permanent events
Other (algal blooms, dredging, wildlife,
333 days plus 1 extended and 3
etc.)
permanent event
Rain or preemptive closing usually due
to stormwater or sewer overflows
permanent events
In California, it is common to have beach closings caused by rain. In January,
2005, Orange County experienced a series of intense rainstorms. This county had more
then 50 stormdrains, creeks, and rivers emptying into the ocean. Due to the high rain
amount, runoff elevated bacteria levels to 20 times higher then the state’s acceptable
limit. Stormwater helped deposit 10 tons of trash on Orange County beaches (Chong and
Wride 2005). According to Mehta (2002) on rainy days, 40% of the state's beaches
receive poor sanitary marks.
In March 1999, North Carolina health officials placed warning signs along 122 m
(400 ft) of beach to warn swimmers of polluted runoff discharging from the stormwater
pipes in Kure and Carolina Beach. New Hanover County health officials have recorded
16
high bacteria counts near the outfall pipe and attribute blame to septic tanks leaking into
the stormwater system along the oceanfront road (Feagans 1999).
Another advisory caused by polluted stormwater runoff was in Myrtle Beach,
South Carolina, when swim advisories were posted during the beginning of tourist season
2005 (Marshall and Ritch 2005). Also in Morehead City, North Carolina, shellfishing
was restricted for 10 days due to stormwater runoff pollution in June 2003. Prior to those
10 days, there had 58 closings that year for shellfishing water due to high bacteria levels
from stormwater runoff (Smith 2003). In 2004, Morehead City experienced three
swimming advisories due to high enterococcus counts from dog and bird waste in
stormwater runoff. Swimming advisories were posted at Radio Island, North Carolina,
for 28 days from mid-June to mid-July (Smith 2004).
2.3.3 BOAT WASTE, WATER FOWL, AND OIL SPILLS
Boat waste is an independent source of pathogenic bacteria. Sobsey et al. (2003)
conducted a 6-day study encompassing a holiday weekend, monitoring fecal coliform in
coastal waters. It was found that levels of fecal coliform increased correspondingly with
an increased number of boats. The highest measured fecal coliform levels exceeded
North Carolina’s limits and were noted near the boats.
In addition to human impacts, seasonal waterfowl migration can result in high
concentration of birds on and around beaches, and on suburban areas that drain to a
beach. The fecal matter from these animals can sometimes overload the beach’s
absorptive capacity for wastes. A study performed in Lake County, Ohio, connected
high bacteria counts in the waters to seagull droppings. Lake County officials used DNA
analysis to identify seagull droppings as the top source of E. coli bacteria in water
17
samples. Lake County had 178 beach closings in 2003, and seagulls were the cause of
most of them (Hawthorne 2004).
Oil forms globules that can float for days and wash onto beaches for weeks after a
spill. Oil enters coastal waters during tanker accidents, pipeline breaks, refinery
accidents, and stormwater runoff. A report from the Natural Research Council (2003)
stated, “nearly 85% of the 29 million gallons of petroleum that enter North American
ocean waters each year as a result of human activities comes from land-based runoff,
polluted rivers, airplanes, and small boats and jet skis . . .”. The amount of oil entered in
the ocean as runoff from trucks and cars is increasing in coastal areas where the
population is increasing and roads and parking lots are expanding.
2.4 NORTH CAROLINA BEACH MONITORING: POSTING CLOSURES AND
ADVISORIES
In June 1997 North Carolina Shellfish Sanitation and Recreational Water Quality Section
of NCDENR was delegated the responsibility of monitoring the ocean beaches, sounds,
bay and estuarine rivers. North Carolina monitors all 240 of the state’s coastal beaches.
Of the 240 beaches, there are 92 Tier 1 sites, 104 Tier 2 sites, and 44 Tier 3 sites.
Recreational beach water quality monitoring is performed on the ocean and sound-side
weekly from April 1st to September 30th and twice a month in October. Monitoring and
testing continues on a monthly basis from November through March (Potts 2005).
If a certain area along the coast has a problem with water quality, the Shellfish
Sanitation Branch will recommend people not swim within 61 m (200 ft) of a posted sign,
list the area on the local county’s website, and notify the local media and county health
department. The state health director and local health directors have the authority
18
necessary to close a beach if they deem it an imminent hazard to public health (Potts
2005).
Since imperfections exist in the monitoring system, there continue to be risks that
ocean goers can get sick. These risks can be reduced if the amount of bacteria entering
the ocean is reduced, such as BMP implementation.
2.5 ECONOMIC IMPACTS OF COASTAL CONTAMINATION
NRDC (2004) reported that in the year 2000 economic activity associated with the ocean
contributed more then $200 billion to the American economy. Approximately 85% of all
tourism revenues are received in coastal states. According to the EPA Liquid Assets 2000
report, a third of all Americans visit coastal areas each year, making a total of 910 million
trips while spending about $44 billion. In 1997 North Carolina received $2.9 billion from
coastal tourism that generated 44,800 jobs related to coastal tourism (EPA 2000). Ocean
pollution puts coastal tourist revenues at risk.
Beach closings and advisories also cause losses to those who planned to visit the
beach and swim in the water. Economists estimate that a typical swimming day is worth
$30.84 to each individual (Rabinovici et al. 2004). If the number of potential visitors to
the beach is high, a consumer surplus loss can be quite considerable. In 2003 and 2004
North Carolina had 874 and 55 beach days that were under a warning advisory or
closings, respectively. A study on consumer surplus of beach closure was done in Lake
Michigan. This study found that estimates of the economic loss of beach closings due to
pollution ranged from $ 7,935 to $ 37,030 per day (Rabinovici et al. 2004).
19
One of the most crowded surfing and swimming spots in Orange County,
California, Bolsa Chica State Beach, was closed in June 1996 due to elevated bacteria
counts from raw sewage seeping into the ocean from 44 breaks in 20-year old sewer line
that served the state beach’s restrooms. This event hurt California’s economy. Bolsa
Chica State Beach attracted 885,186 visitors in the fiscal year 1994-1995, nearly 600,000
during the summer months alone. Tourism generated about $1 million in revenues
annually from the $5 dollar parking fee alone. In 1995 there were 26,296 visitors the
same week as when the beach was closed in 1996 (Schoch 1996).
Beach pollution can also cause costly health reparations generated by swimming
in polluted water. Ocean goers can swim in unhealthy water if they ignore the advisory
signs or swim in polluted water before the 24 hours needed to detect and post unsafe
bacteria levels. Baker et al. (2005) conducted a study using data from two popular
Orange County, California, beaches, Newport and Huntington, to estimate the economic
impact of illnesses associated with polluted recreational waters. It was found that
swimming in these coastal waters cost the public $3 million per year in health related
expenses. This calculation is based on doctors’ fees to treat more than 74,000 incidents
of stomach illness, respiratory disease, and skin, eye, and ear infections caused by
exposure to polluted waters in a typical year and based on the lost income of typical
Orange County salaries.
Another economic aspect that is affected by polluted coastal water is the seafood
market. Every year, the Great Lakes, Gulf of Mexico, as well as the Pacific and Atlantic
coastal areas produce more than 10 billion pounds of fish and shellfish (US EPA 2000).
20
According to Clark and Stoner (2001), stormwater runoff costs the commercial fish and
shellfish industries between $17 million to $31 million annually.
2.6 BEST MANAGEMENT PRACTICES (BMPS)
2.6.1 INTRODUCTION TO BMPS
To help minimize the amount and improve the quality of stormwater runoff entering the
ocean, Best Management Practices (BMPs) can be implemented. EPA (1999) defines a
stormwater BMP as “a technical measure or structural control that is used for a given set
of conditions to manage the quantity and improve the quality of stormwater runoff in the
most cost effective manner.” Structural BMPs are engineered, constructed systems.
Non-structural BMPs are educational and pollution prevention practices designed to limit
the generation of stormwater runoff or reduce the amounts of pollutants contained in the
runoff. Structural and non-structural BMPs are used to minimize flooding, erosion, and
the amount of metals, nutrients, and bacteria (US EPA 1999a).
BMPs utilize the concepts of infiltration, filtration, detention, and retention.
Infiltration systems capture a volume of runoff allowing infiltration into the ground.
Filtration systems use combinations of granular filtration media such as sand, soil,
organic material, or carbon to remove constituents found in runoff. Detention systems
capture a volume of runoff and temporarily retain that volume for subsequent release, but
do not retain a significant permanent pool of water between runoff events. Retention
systems capture a volume of runoff and retain that volume until it is displaced by the next
runoff event, thus maintaining a significant permanent pool volume of water between
runoff events (US EPA 1999a). Common BMPs that employ one or more of these
21
concepts are wet ponds, wetlands, bioretention areas, sand filters, riparian buffers and
level spreaders, and reinforced grassy swales (Hunt 1999).
No single BMP can address all stormwater problems since each type has certain
limitations based on several factors including: drainage area served, available land space,
cost, pollutant removal efficiency, soil types, slopes, and depth of the groundwater table.
BMPs, effectively designed, increase pollutant removal and flow control (US EPA
1999a).
2.6.2 SAND FILTRATION BMP
2.6.2.1 Introduction to Sand Filtration Systems as BMPs
Sand filters are BMPs that have been borrowed from the treatment of wastewater and
drinking water. Sand filters consist of self-contained beds of sand that are either
underlain with drains or cells, and include baffles at inlets and outlets. Stormwater runoff
is filtered through the sand, removing contaminants via physical entrapment and sorption.
The type of media used and its grain size determines the pollutant particle size captured.
Coarser sands have larger pore space, allowing for high flow-through rates, but also
allowing commensurately larger particles to pass through. Fine sand has smaller pore
spaces with accordingly slower flow-through rates and filters out small total suspended
solids (TSS) particles (Urbonas 1999).
There are three commonly used sand filter systems: Austin sand filter, Delaware
sand filter, and the Washington, D.C., sand filter. The primary differences among these
designs are location (below or above ground), drainage area served, filter surface area,
land requirements, and quantity of runoff they treat (US EPA 1999a). In addition to the
22
three basic filtering systems, there are a number of variations and combinations of these
systems in use.
The Austin sand filter design evolved into two chambers: a sediment chamber or
pond followed by a surface sand filter with collector under drains in a gravel bed (Figure
2-1). First, the stormwater runoff enters the pretreatment sediment chamber, where
gravity removes coarse particles. The runoff then follows over a weir or through a riser
into the sand filter bed. Additional storage volume is provided above the filter bed to
increase the volume of water that can be temporarily ponded in the system prior to
filtration (City of Austin 1991).
Figure 2-1. Plan view and profile of Austin sand filter (City of Boise Public Works
Professional Advisory Group 1998).
Austin sand filters are designed for drainage areas less then 20.2 ha (50.0 ac).
They are designed to capture and treat the first 1.27 cm (0.50 in) of stormwater runoff
(US EPA 1999b). The two-basin configuration can limit premature clogging of the filter
23
bed due to excess sediment loading. The design concept of the Austin sand filter is most
like the Dune Infiltration System (DIS) implemented at Kure Beach, North Carolina.
Removal efficiencies for Austin sand filters are shown in Table 2-3.
Table 2-3. Percent pollutant removal effectiveness for surface sand filters.
Study
TSS TP TN NO3 Metals
City of Austin (1990)
75
59 44
-13
34-82
92
80 71
23
84-91
87
61 32
-79
60-81
Welborn & Veenhuis (1987)
78
27 27 -111
33-60
Source: (Strecker et al. 2001)
Shaver and Baldwin (1991) designed another type of sand filter in Delaware for
used around the perimeter of parking lots, called the Delaware sand filter. The Delaware
sand filter (Figure 2-2) consists of parallel sedimentation and sand filter trenches
connected by a series of level weir notches to assure sheet flow onto the filter.
Stormwater runoff enters the sediment chamber, then flows over the series of weirs into
the sand filter chamber. Additional storage volume is provided by water temporarily
ponding in both chambers. After being filtered, the stormwater is collected by a series of
gravity pipe under drains and flows into a clarifying well that is connected to a storm
drain system (Shaver and Baldwin 1991). This type of sand filtration captures and treats
2.54 cm (1.00 in) of stormwater (US EPA 1999b).
24
Figure 2-2. Plan and profile view of Delaware sand filter (City of Boise Public Works
Professional Advisory Group 1998).
The Underground Sand Filter was developed in Washington, D.C. in the late
1980’s. This filter is placed underground but contains the same components as the
Austin sand filter (Figure 2-3). The filter has three chambers. The first is a 0.9 m (3 ft)
deep chamber containing a permanent pool of water and functions as a sedimentation
chamber and an oil and grease trap. The second chamber is a 46-61 cm (18-24 in) sand
filter bed with a submerged opening. The second chamber contains an under drain
system as well as inspection and cleanout wells. The last chamber routes the flow to the
downstream receiving drainage systems. This type of filter is used for 0.4 ha (1 ac) or
less and can capture and treat 1.27 cm (0.5 in) of stormwater runoff (US EPA 1999b).
25
Figure 2-3. Plan view and profile of underground sand filter (City of Boise Public
Works Professional Advisory Group 1998).
2.6.2.2 Implementation of Sand Filtration Systems
Sand filters are primarily intended for water quality enhancement. They are preferred
over infiltration practices when contamination of the groundwater by suspended solids,
and fecal coliform are of concern. Sand filters can be highly effective stormwater BMPs
since they have high removal rates of sediment and fecal bacteria and require less land
then other BMPs. Typical pollutant removal efficiency in sand filters is shown in Table
2-4 (US EPA 1999b).
26
Table 2-4. Typical Pollutant Removal Efficiency in Sand Filters (EPA 1999b).
Nitrogen removal is complicated in sand filters. Sand filters are nitrate creators,
trapping organic nitrogen in an aerobic environment forming nitrate. Thus nitratenitrogen (NO3-N) levels increase through the use of sand filters. In Alexandria, Virginia,
two large Delaware style sand filters were monitored for six months to establish the
actual pollutant removal efficiency. It was found that sand filters were susceptible to
anaerobic conditions, which have a negative impact on total phosphorous removal but a
positive effect on total nitrogen removal. Also, placing a 33-cm (13-in) flooded gravel
filter beneath the sand filter may enhance nitrogen removal if sufficient organic carbon
was present. Forty-seven percent removal efficiency of total nitrogen (TN) and 72 %
removal efficiency of total phosphorous (TP) in aerobic state was found. Removal
efficiencies for total suspended solids (TSS) exceeded 80% (Strecker et al. 2001).
Barrett (2003) conducted a study on five Austin style sand filters that were
constructed by the Californian Department of Transportation (CALTRANS) in the Los
Angles and San Diego metropolitan areas in California. Performance analysis using a
27
linear-regression technique indicated that for sediment and almost all particle associated
constituents, effluent concentrations were independent of influent concentrations. The
constant effluent quality produced for the particular constituents indicated that the
calculation of a percent reduction is more indicative of the influent concentration rather
than the filter’s performance.
Along with efficient pollutant removal, sand filters are relatively easy to retrofit
into the available space. In Rehoboth, Delaware, Delaware style sand filters were
installed along streets in the city to reduce the bacteria levels in the stormwater runoff by
reducing the volume of pollutants in the runoff as well as collect litter and food waste
washed into stormdrains in the commercial resort area. These sand filters treated storms
under 2.54 cm (1.00 in) and decreased the amount of grease, oil, phosphorous,
hydrocarbons, lead, and nickel by 80% (Shaver 1994).
Sand filters have been used as a retrofit for water quality in existing drainage
basins in the nation’s capital. Dee (1997) used an Austin sand filter as a BMP in a 1.4 ha
(3.5 ac) existing watershed. Dee retrofitted a sand filter BMP into an existing basin,
which provided high pollutant removal efficiencies, as well as debris reduction.
Sand filters are typically designed for a drainage area ranging from 0.2-4.0 ha (0.5
-10 ac). Grisham (1995) presented the concept of engineers designing for the ‘first flush’
of a storm, describing the first flush as the runoff from the first 15 minutes of a storm
generally considered the first 1.27 cm (0.50 in) of stormwater runoff. This ‘first flush’
contains proportionately high levels of pollutants relative to rain thereafter. Thus, sand
filters are typically designed to capture 1.27-2.54 cm (0.50-1.00 in) of the storm.
28
One disadvantage of sand filters is the cost to construct and maintain. Wossink
and Hunt (2003) found sand filters to be relatively very expensive due to construction
materials and methods. Table 2-5 shows the predicted construction and maintenance of
sand filter compared to other BMPs.
Table 2-5. Summary of Construction Cost Curves, Annual Maintenance Cost Curves
and Surface Area for five Stormwater BMPs in North Carolina, C = Cost in $, x = Size of
watershed in acre, SA = Surface Area in acre (Wossink and Hunt 2003).
Maintenance is required for sand filters to work properly according to Grisham
(1995). When designing sand filters, permeability calculations should be based on the
assumption that the filter is 50 % clogged due to the expected clogging and improper
maintenance. Accumulated trash, paper, and debris should be removed from sand filters
every 6 months, or as necessary. Corrective maintenance of filtration chambers includes
removal and replacement of the top layers, 2.5-7.6 cm (1.0 -3.0 in) of sand (Hunt 1999).
Laboratory and field tests show that a filter media consisting of concrete sand
provides a good balance between flowrates and filtering efficiency (Urbonas 1999).
Initially the flowrates of the stormwater through the sand media are high, but as the
filtration of fine sediment accumulates on its surface, flowrates are reduced. Field tests
29
show that the effluent quality improves initially, but may degrade over time. This leads
to constituents leaching out from the filtrate and a need for maintenance.
In California it was found that rejuvenation of the filter bed was required at three
sites after 3 years of operation when solid loading was between 5 and 75 kg/m2 (1.0 to
15.4 lb/ft2) Barrett (2003). The study concluded that routine maintenance, including
periodic removal of the top layer of sand, will prolong operation.
The main impediment for adoption of this technology is the high construction
cost. However the small amount of land required for filter/basin configurations may
reduce the cost substantially. Thus, sand filters are a viable technology for stormwater
treatment where low concentrations of sediment and particle-associated, such as bacteria,
constituents are desired.
2.7 LABORATORY STUDIES ON BACTERIA REMOVAL FROM SAND COLUMNS
The majority of the laboratory studies examining bacteria removal through sand filtration
focus on wastewater systems’ removal of the bacteria via slow and fast sand filters as
well as bacteria removal in different types of filtration technologies. A study performed
by Gomez et al. (2006) showed that a pressure sand filter in a wastewater system
removed 36.88 ± 24.68% of fecal coliform and 34.1± 34.23 % of E. coli. Large
variations of bacteria removal are common in this type of system as well as in laboratory
studies involving sand column bacteria removal.
The adsorption of bacteria onto soil is affected by the physical and chemical
characteristics of the soil and water, the size and morphology of the bacterial cells, and
the water-flow characteristics in the soil. Abu-Ashour and Abu-Zreig (2005) studied the
30
effect of interstitial velocity on the adsorption of E. coli onto sandy soil. The results
showed that E. coli was retained in the sandy soil at lower interstitial velocities; whereas
the higher interstitial velocity resulted in higher shear forces which caused more
desorption of the E. coli cells from the surfaces of the soil particles.
Soil type is a physical characteristic that affects the adsorption of bacteria onto the
surrounding soil. Meschke and Sobsey (2003) conducted an experiment to directly
compare mobility and reduction of bacteria indicator, E. coli, in various soil systems.
They constructed 10 cm (4 in) deep soil columns, filled with either sand, organic muck or
clay and compared the mobility and reduction of E. coli in each. Rapid mobility with
limited reduction (1.6 log) was observed in the sand as well as the organic muck (2.8
log). No E. coli was shown to pass through the clay columns, allowing for a reduction
greater then 3.8 log.
Grain size and ionic strength of the soil are other physical and chemical
characteristics that affect bacteria adsorption onto the soil. Bolster et al. (2001)
examined the effect of grain size and ionic strength on a phenomenon known as blocking.
Blocking is when large bacteria loadings lead to high coverage of sediment surfaces
resulting in a decrease in deposition of bacteria. The presence of previously deposited
bacteria can result in decreased deposition rates. To test the effect of grain size, two
quartz sand columns were constructed, one composed of fine sand 0.42 to 0.50 mm in
diameter and the other with coarse sand 0.707 to 0.850 mm in diameter. Radio labeled
bacteria were introduced into each column. The effect of grain size on maximum
bacteria retention capacity was not found to be significant. However, an additional test
was performed on the columns, which altered the sand’s ionic strength. It was found that
31
decreasing ionic strength from 10-1 to 10-2 M KCL resulted in a decrease sticking
efficiency and maximum surface coverage. This could be an issue of concern for the
Dune Infiltration System, due to the low ionic strength associated with system’s sandy
soil.
Bolster et al. (2001) studied the effect of coating the sand with positively charged
aluminum and ferric hydroxides. It was found that the presence of Al- and Fe-coated
sand increased both deposition rates and maximum fractional surface coverage of
bacteria on sediment surfaces. Lukasik et al. (1999) studied the effective bacteria
removal of gravity flow raw sewage through sand columns coated with metallic
hydroxides to that of unmodified sand. Greater than a 4 log10 reduction in of E. coli was
achieved in the modified sand; whereas, less then a log10 was achieved in the unmodified
sand. The metal used for coating the sand could not be detected in the columns’ effluent,
indicating that the coatings were stable.
The modified sand seemed to better remove microorganisms due to increased
electrostatic interactions. It is interesting to note that by the end of the experiment there
was no significant difference in the ability of removing E. coli from the columns
containing treated sand compared to the columns containing unmodified sand. This was
attributed to a development of a microbial biofilm on the unmodified sand, which
decreased infiltration rate (Lukasik et al. 1999).
A similar effect on infiltration rate was observed in an experiment correlating the
bacteria production of extracellular polymers to saturated hydraulic conductivity (Ks).
Bacterial reductions of Ks in natural porous media have been traditionally associated with
development of anaerobic conditions and the production of large amounts of extracellular
32
polymers (Vandeviver and Baveye 1992). Vandeviver and Baveye (1992) tested a series
of percolation experiments to determine the extent to which aerobic bacteria were able to
clog permeameters filled with fine sand. It was found that strictly aerobic bacteria were
able to reduce Ks by up to four orders of magnitude. Initially, rapid reductions in Ks are
associated with the formation of a bacterial mat at the inlet boundary of the sand
columns. When the colonization of the inlet is prevented, clogging proceeds within the
bulk of the sand at a noticeably slower rate. Under oxygen- or glucose-limited growth
conditions, Ks decreases within the sand due to large aggregates of bacterial cells that
form local plugs within the pores. In all cases, the coverage of solid surface by the
bacteria cells was found to be sparse and heterogeneous.
33
3.0 HYPOTHESES AND OBJECTIVES
The literature review established the importance of keeping bacteria counts in coastal
waters below government recommended bacteria standards. If bacteria levels in coastal
and estuarine waters rise, the risk of illness to ocean goers and those who ingest shellfish
and other ocean life increase. Increased bacteria levels also decrease the coastal
communities’ economic viability by causing beach closures and advisory days, which
hurt local businesses. Stormwater runoff is the number one known cause of beach
closures and advisories.
As coastal communities develop, the amount of stormwater runoff increases,
while land availability decreases. Sand filters are an effective BMP for mitigating flow
and removing stormwater constituents where land is limited. Recently, coastal
communities installed successful sand filter BMPs. Sand filters have yet to be placed in
un-developable coastal land, the beach sand dune.
The overall goal of this research was to test a potential BMP, called a Dune
Infiltration System (DIS), which does not consume valuable coastal developable
property. The research will establish whether the DIS decreases the potential health
dangers associated with stormwater ocean outfalls for local coastal residences, tourists,
and coastal wildlife. Decreasing potential health dangers involve treating/removing the
bacteria transported in the stormwater ocean outfalls. To achieve the overall goal, a field
study and laboratory experiment was designed to answer several objectives.
34
The objectives of the field study were as follows:
1. Identify a range of fecal coliform and enterococcus concentrations in an urban
coastal community’s stormwater runoff.
2. Design a Dune Infiltration System that will capture all runoff associated with
a rainfall intensity of 12.5 mm/hr (0.5 in/hr) or less.
3. Determine if implementing a DIS decreased the amount and peak rate of
stormwater runoff directly discharged on the beach.
4. Determine if routing and discharging stormwater runoff in the dunes elevated
the level of the groundwater beneath the dunes.
5. Determine bacteria removal efficiency of the DIS by monitoring the inflowing
and outflowing bacteria concentrations for 25 storm events.
6. Determine if routing and discharging stormwater runoff into the dunes
increased the bacteria level in the groundwater beneath the dunes.
The hypotheses of this research were tested using a 95% level of confidence. The
field hypotheses were established to evaluate the overall goal of assessing the DIS as a
viable BMP. The field hypotheses are as follows:
First Hypothesis: The amount of stormwater runoff directly discharged onto the beach is
significantly less than the amount of stormwater captured in the DIS (α =0.05).
Second Hypothesis: The peak rate of stormwater runoff directly discharged onto the
beach is significantly less than the peak rates inflowing to the DIS (α = 0.05).
Third Hypothesis: The fecal coliform concentration in the inflowing stormwater runoff is
significantly greater than in the ground water samples (α = 0.05).
35
Fourth Hypothesis: The enterococcus concentration in the inflowing stormwater runoff is
significantly greater than in the ground water samples (α = 0.05).
Fifth Hypothesis: The fecal coliform concentration in the groundwater before DIS was
installed is significantly greater than to the fecal coliform concentration in the
groundwater after the DIS was installed (α = 0.05).
The laboratory study presented herein, three treatments of bacteria-free
stormwater (Control stormwater), bacteria stormwater (Test), and control deionized (DI)
water (Control DI) will be applied to sand columns. It is anticipated that the bacteria
treatment columns will clog faster then the other two controls, in response to aggregates
of bacterial cells forming plugs within the pores. This decrease in infiltration rate, in
return, should decrease the amount of bacteria filtering though the system, as predicted
by Lukasik et al. (1999).
The objectives of the laboratory study were as follows:
1. Determine the removal efficiency of E. coli by sand columns and if E. coli
removal efficiency is affected by sand clogging.
2. Determine the effect that the stormwater runoff contaminants have on the
infiltration rate in the sandy soil in order to determine a maintenance schedule
for the DIS.
From previous studies cited in the literature review, it is has been demonstrated
that slower influent velocity, along with a corresponding growth of microbial biofilm
increases the sand’s bacteria removal efficiency. To better understand the correlation of
stormwater inflow rate to the DIS bacteria removal efficiency, the following laboratory
hypotheses were tested:
36
First Hypothesis: The infiltration rates of the three Test columns are significantly less at
the end of the 60-day test period than infiltration rates of the Control DI water (α =0.05).
Second Hypothesis: The infiltration rates of the Control Stormwater columns are
significantly less than the infiltration rates of the Control DI water at the end of the 60day test period (α = 0.05).
Third Hypothesis: The infiltration rates of the Test columns are significantly less than the
infiltration rate of the Control Stormwater at the end of the 60-day test period (α =0.05).
Four Hypothesis: The initial E. coli concentrations in the Test columns’ effluent are
significantly greater than the initial Control Stormwater columns’ effluent (α =0.05).
Fifth Hypothesis: The E. coli concentrations in the Test columns’ effluent are
significantly less than the influent at the end of the 60-day test period (α = 0.05).
Sixth Hypothesis: The E. coli concentrations in the Test columns’ effluent are directly
correlated with the infiltration rates of the columns (α =0.05).
37
4.0 Dune Infiltration System Field Study
4.1 SITE DESCRIPTION
The Dune Infiltration System (DIS) demonstration project was implemented in the Town
of Kure Beach, North Carolina,
located at 34°00’11” East
latitude and 77°54’21” North
longitude according to the North
American Datum of 1983
(NAD83). The Town of Kure
Beach is located in New Hanover
County, as shown in Figure 4.1.
Kure Beach
Figure 4-1. Map of NC illustrating the location of
New Hanover County and the Town of Kure Beach.
4.1.1 LOCATION OF DIS
The Town of Kure Beach has 22 stormwater ocean outfalls. Two of these ocean outfalls
draining small watershed areas in Kure Beach were selected for the DIS demonstration
and research project. Site L (Figure 4-2 (a)), named after the street it borders, is a 1.8 ha
(4.5 acres) mixed urban and residential land use watershed, with a Rational equation
runoff coefficient of C = 0.8. Site M (Figure 4-2 (b)), also named after the street it
borders, is a 3.3 ha (8.1 acre) predominately dense residential land use watershed, with a
C = 0.7. The DIS were designed to be placed under the site’s dunes, described as a
Newhan Fine sand, composed of 99.4 % sand and 0.6% silt (NRCS, 2005). The North
Carolina Division of Coastal Management issued a Coastal Area Management Act
38
(CAMA) exemption to permit the work within the dune system. This not only facilitated
the logistics of the design but also substantially reduced the cost of the project, since
valuable ocean-front real estate was not purchased. Special care was taken not to disturb
the dunes during sea-turtle nesting season.
Outfall Pipe
Outfall
Pipe
(a)
(b)
Figure 4-2. (a) Site L Watershed Area (b) Site M Watershed Area.
4.2 DIS DESIGN CONSIDERATIONS
4.2.1 DIS PRECONSTRUCTION MONITORING
Prior to DIS installation, groundwater elevations and bacteria concentrations were
monitored at Sites L’s and M’s culvert to establish baseline levels.
Preconstruction groundwater levels were monitored to gauge groundwater depths,
daily tidal influences, and storm-induced fluctuations. Groundwater elevations were
measured using an encased INFINITY 1 continuous water table recorder (INFINITIES
F
USA, Inc., Daytona Beach, FL). The INFINITY’s casing was a 4.3 m (14 ft) long, 5.1
1
The use of trade names does not imply endorsement by North Carolina State University.
39
cm (2 in) diameter polyvinyl chloride (PVC) pipe. The pipe is capped at the bottom end,
with four 1 cm (3/8 in) diameter holes spaced around the circumference of the pipe. The
1 cm (3/8 in) diameter holes were drilled in 15 cm (6 in) increments, along the bottom 2.4
m (8 ft) of the casing. Two layers of drainage sock were placed over the holes to reduce
the amount of fine sediment entering the casing.
In June, 2005, a drill rig with a 30.5-cm (12-in) diameter auger, depicted in Figure
4-3, drilled a 3.8 m (12.5 ft) deep hole in the dunes at each site. The pre-constructed
casing was lowered into the holes and back filled with sand. A 5.1 cm (2.0 in) x 10.2 cm
(4 in) PVC coupling was glued to the top of the casing so that the data logger could be
affixed.
Figure 4-3. Installation of preconstruction groundwater well.
In June, 2005, six groundwater bacteria monitoring wells were installed at each
site. These wells were constructed using 5.1 cm (2 in) diameter PVC, of varying lengths,
as reported in Table 4-1.
40
Table 4-1. Groundwater bacteria monitoring well specifications.
Site L
Well
Name
L-4
L-6
L-8
L-10
L-12
L-14
Site M
Well
Name
M-4
M-6
M-8
M-10
M-12
M-14
Length of PVC
Pipe
1.68 m (5.5 ft)
1.68 m (7.5 ft)
1.68 m (9.5 ft)
1.68 m (11.5 ft)
1.68 m (13.5 ft)
1.68 m (15.5 ft)
Depth of hole
drilled
1.2 m (4 ft)
1.8 m (6 ft)
2.4 m (8 ft)
3.0 m (10 ft)
3.7 m (12 ft)
4.3 m (14 ft)
Range of Water Table
Measured
0.9-1.2 m (3-4 ft)
1.5-1.8 m (5-6 ft)
2.1-2.4 m (7-8 ft)
2.7-3.0 m (9-10 ft)
3.4-3.7 m (11-12 ft)
4.0-4.3 m (13-14 ft)
Well pipes were cut to the requisite length, and four, 1 cm (3/8-in) holes were drilled
around the circumference in 5.1 cm (2 in) vertical increments from the bottom 0.3 m (1
ft). The bottoms of the wells were capped and two layers of drainage sock were secured
over the holes. Using the same drill rig, six holes were drilled below the surface of each
site’s dunes at various depths indicated in Table 4-1. Figure 4-4, depicts the
preconstruction groundwater sampling wells in Site M’s dunes.
Figure 4-4. Pre-construction hydrology monitoring wells.
Precipitation measurements were necessary to correlate the rise in the water table
with the size of the storm. A Davis Rain Collector™ tipping bucket recorder with a 0.25
mm (0.01 in) capacity bucket (Davis Instruments, Hayward CA, Model 7852) along with
a HOBO® data logger (Onset Computer Corporation) were installed in site M’s dunes.
41
A backup manual gauge was installed near the tipping bucket. Due to a malfunction in
the original HOBO tipping bucket, a second one was later installed at Site L.
4.2.2 PRECONSTRUCTION SAMPLING PROTOCOL
Fecal coliform samples were intermittently collected during the months of July, 2005,
through September, 2005, within 24 hours of a storm event. Stormwater runoff samples
were collected from each site’s ocean outfall pipes. A total of six groundwater samples
were obtained from the following installed monitoring wells: L-10, L-12, L-14, M-12,
and M-14. Before collecting the sample, monitoring wells were purged 2 well volumes
using the well’s designated bailer. One 200 ml sample was collected from each well in a
sterile 250 ml bottle, with a sodium thiosulfate tablet in the bottle as preservative. The
samples were put on ice and transported to Oxford Laboratory, Inc., in Wilmington, NC,
for fecal coliform analysis (EPA method SM 922D).
Along with bacteria sampling, groundwater and rainfall data was obtained.
INFINITY data loggers were downloaded using a Hewlett-Packard™ calculator with
INFINITY software. The HOBO® data logger was downloaded using a HOBO® shuttle
data logger (Onset Computer Corporation). For each storm the amount of rain collected
in the manual rain gauge were recorded and emptied out. Both groundwater and rainfall
data were later uploaded into a Microsoft® Excel spreadsheet for analysis.
4.2.3 DIS PRECONSTRUCTION FIELD MEASUREMENTS
4.2.3.1 Site Survey
To accurately design DIS, Site L’s and Site M’s dune areas were surveyed. These areas
were surveyed with a Sokkia Total Station model SET5 30R. Survey points were
42
measured approximately every 3 m (10 ft) apart from the western side of the dune,
Atlantic Avenue, to the eastern side of the dune, approximately to the mean high tide
water line. These points were surveyed within the north and south boundaries of the
public beach access boardwalks. Particular detail was taken in surveying the location of
the groundwater and bacteria monitoring wells. Figure 4-5 is an AutoCAD drawing of
the relative elevations in Site L’s and M’s dune.
Figure 4-5. Survey of Site L and Site M in Kure Beach, NC (feet).
Surveying the North Carolina Geodetic Survey Stations ANT and THRID, located
near Site L’s and M’s dune, allowed the conversion of relative elevations, to become
actual elevations in the NAD83. The new survey was overlaid onto a 2002, aerial
43
photograph download from New Hanover County website (2006). Figure 4-6 is the
Autodesk Land Desktop® drawing of the NAD83 elevations of Site L and M.
Figure 4-6. Actual Survey Elevations of Site L and M Kure Beach, NC (feet).
4.2.3.2 Single Ring Infiltrometer Test
From the Nature Resource Conservation Service (NRCS) Soil Data Mart (2006), the site
soil was defined as Newhan Fine sand, having a profile with a uniform measured moist
bulk density between 1.60 g/cm3 (0.057 lbs/in3) to 1.75 g/cm3 (0.063 lbs/in3) and uniform
saturated hydraulic conductivity of 392 cm/ hr (154 in/hr) (NRCS 2006). Single ring
infiltrometer tests were preformed to verify the saturated hydraulic conductivity. The
single ring infiltrometer test measures infiltration rates, rather then saturated hydraulic
conductivity values. Infiltration rate is defined by Schwartz and Zhang (2003) as “the
process of downward water entry into soil. The rate of infiltration is usually sensitive to
44
near-surface conditions as well as the antecedent water conditions of the soil;” thus, it is a
characteristic of the soil’s surface. However, the hydraulic conductivity “relates specific
discharge to the hydraulic gradient” (Schwartz and Zhang 2003) making hydraulic
conductivity a parameter describing the ease of flow through a porous soil media.
Therefore, while infiltration rate is dependent on the hydraulic conductivity of the media
being infiltrated, hydraulic conductivity is not dependent on the infiltration rate, but
rather soil properties.
ASTM D 3385, the “Standard Test Method for Infiltration Rate in Field Soils
Using Double-Ring Infiltrometer” was the procedural basis for measuring surface
infiltration rates. This test measures infiltration rates for soils with a hydraulic
conductivity between 10-6 cm/s (3.9 x 10-7 in/s) and 10-2 cm/s (3.9 x 10-3 in/s). The
double ring infiltrometer test was modified to a single ring infiltrometer test, due to the
NRCS reporting the saturated hydraulic conductivity greater than the standard’s range.
The single-ring infiltrometer used consisted of 16 gauge galvanized steel rings
with inner diameter of 30.5 cm (12 in.). Six single infiltrometer tests were performed at
three separate locations at each site in the dunes after a 24-hour dry period. The locations
of the tests were at least 8.1 m (30 ft) apart on a level dune surface with no vegetation
located inside the ring. The ring was hammered 15 cm (6 in) into the ground with a
sledge and a wooden block. The test was done according to the ASTM 3385 standard,
with one exception. The test was designed to run until the water can no longer infiltrate.
This was deemed impossible due to the high infiltration rate of Newhan sand, so the test
was run until the water source ran dry.
45
Data were plotted as cumulative infiltration volumes versus infiltration time
(Figure 4-7). Since the infiltration rate is equivalent to the maximum-steady state or
average incremental infiltration velocity, the slope of the least squares line for each test
was determined to be equal to the surface infiltration rate of the tested surface. The
surface infiltration rate for each site was determined by averaging the results from the
three tests. The average surface infiltration rate was measured to be 329 cm/hr (130
in/hr) for Site L and 419 cm/hr (165 in/hr) for Site M. The overall average of the test was
372 cm/hr (147 in/hr), which was close to the saturated hydraulic conductivity, 392 cm/hr
Cummulative infilitration (cm)
(154 in/hr), measured by NRCS (2005).
3) y = 365.32x + 1.8381
R2 = 0.9999
140
1) y = 311.44x + 10.937
R2 = 0.9998
120
100
80
2) y = 309.82x + 8.5213
R2 = 0.9999
60
40
20
0
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
Time (hr)
Trial 1
Trial 2
Trial 3
Linear (Trial 3)
Linear (Trial 2)
Linear (Trial 1)
Figure 4-7. Site L’s cumulative infiltration versus time for three single ring
infiltrometer tests.
4.3 DIS DESIGN
Pre-construction monitoring from July, 2005, until January, 2006, indicated that water
table depths below the dune surface were, on average, 3.5 m (11.5 ft) at Site L and 4.0 m
46
(13.1 ft) for Site M. This provided sufficient depth to allow for vertical infiltration of
stormwater runoff.
4.3.1 HYDROLOGIC CALCULATIONS
4.3.1.1 Rational and Natural Resources Conservation Service Method (NRCS)
Calculations
The DIS was designed to capture the amount of stormwater runoff produced by a 12.5
mm (0.5 inch) per hour storm. Using previously measured watershed characteristics and
a given design storm, the system was designed based on the Rational Method as well as
the NRCS Method. The Rational Equation, (EQN 4-1) was used to calculate peak
discharge of each site’s drainage area (Schwab et al. 1993).
q = 0.0028* C * i * A
(4-1)
Where: q = peak discharge (m3/s),
C = Rational method runoff coefficient (0.8 Site L, 0.7 Site M),
i = rainfall intensity (mm/hr),
A = Watershed Drainage area (ha).
The peak discharge based on the design storm, area of the watershed, and C values,
needed to be diverted to the Dune Infiltration System was 0.05 m3/s (1.88 cfs) for Site L
and 0.07 m3/s (2.69 cfs) for Site M. The method of calculating the time of concentration,
Tc, was the Kirpich/Ramser (EQN 4-2) (Schwab et al. 1993).
t c = 0.0195* L0.77 * S −0.385
(4-2)
Where: tc = time of concentration (min)
L = hydraulic watershed length (m)
S = average hydraulic gradient (m/m)
S was calculated from the surveying data, equaling 0.02 m/m, which yielded an estimated
Tc value of 12 minutes for Site L and 16 minutes for Site M.
47
Another relevant empirical formula for determining the quantity of runoff was the
NRCS Equations used to calculate a Unit Hydrograph. This method estimated the time to
peak and the peak discharge, remembering that:
1) Weighted CN must be over 40.
2) The CN procedure is less accurate when runoff is less than 13
mm/hr (0.5 in/hr).
First, surface storage and runoff depth were calculated using equations 4-3 and 4-4
(Schwab et al. 1993)
⎛ 25400 ⎞
S =⎜
⎟ − 254
⎝ CN ⎠
( I − 0.2 S ) 2
Q* =
( I + 0.8S )
(4-3)
(4-4)
Where: S = maximum potential differences of rainfall and runoff (mm)
CN = curve number
I = storm rainfall (mm)
Q* = direct surface runoff depth (mm)
The CN used was a composite CN, using CN=98 for impervious area and CN= 50 for
pervious surfaces, based on the value assigned to Kure Beach’s soil. The composite CN
was estimated to be 88 for Site L and 69 for Site M. Next, total runoff volume (EQN 45) was calculated so that time of peak runoff could be calculated (EQN 4-6) (Malcom
1989).
Vol = 10* Q * A
Vol
Tp =
1.39* Q p
Where: Vol = volume of water under hydrograph (m3)
Q = direct surface runoff depth (mm)
A = watershed Drainage area (ha)
Tp = time to peak of the design hydrograph (sec)
Qp = peak discharge (m3/s)
(4-5)
(4-6)
48
Equation 4-7 was used to graph each site’s unit hydrograph as derived by H.R. Malcom
(1989).
0 ≤ t ≤ 1.25* Tp → Q =
Q p ⎛ 1 − cos(π t ) ⎞
⎜
⎟⎟
Tp
2 ⎜⎝
⎠
t > 1.25* Tp → 4.34* Q p * e
(4-7)
⎛ −1.3t ⎞
⎜
⎟
⎝ Tp ⎠
Outflow (m^3/s)
Where:
Q = Watershed Inflow (m3/s)
Tp = Time to peak of the design hydrograph (sec)
Qp = peak discharge (m3/s)
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
Site L
Site M
0
10
20
30
40
50
60
Time (min)
Figure 4-8. Sites L’s and M’s estimated inflow hydrograph for a 12.5 mm/hr (0.5 in/hr)
storm.
4.3.1.2 Darcy’s equation
The DIS was constructed using commercially-available open bottomed, high density
polyethylene (HDPE) infiltration chambers called StormChambers™ , produced by
HydroLogic Solutions, Incorporated, Occoquan, VA. The chambers were 1.1 m (3.5 ft)
high, 1.5 m (5.0 ft) wide, and 2.5 m (8.2 ft) long (Figure 4-9).
49
Figure 4-9. StormChambers™ schematic (courtesy Hydrologic Solutions Inc.).
The number of StormChambers™ necessary was calculated by combining the
hydrologic calculations previously mentioned, with Darcy’s equation (equation 4-8).
Q = AK
___
Δh
L
(4-8)
Where: Q = volumetric flowrate (m3/s or ft3/s),
A = flow area perpendicular to L (m2 or ft2),
K = hydraulic conductivity (m/s or ft/s),
L = flow path length (m or ft),
Δh = change in hydraulic head over path L (m or ft)
Darcy’s equation was used to conservatively estimate the number of chambers required to
accommodate stormwater infiltrate into the dunes. The use of this equation assumed only
vertical flow. Since this was a demonstration of an untested coastal BMP, ignoring
lateral flow during design provided a conservative design to help ensure dune protection.
The vertical hydraulic conductivity, K, used in Darcy’s equation was the single
ring infiltrometer test (described in Bean, 2005) result, which averaged to a value of 372
cm/min (0.003 ft/s). The flow path, L, was the depth of the sand to the water table,
which ranged from 2.0-2.6 m (6.6-8.5 ft). The area perpendicular to the flow, A, was
equal to the open area of the bottom of an individual chamber, 3.4 m2 (36.5 ft2). The
change in hydraulic head, h, ranged from 3.4 m (11ft), (the height of the chamber, 1.1 m
(3.5 ft), full of stormwater plus the average depth of sand) to approximately zero meters
50
at the water table. Using the maximum and minimum change in hydraulic head, the
volumetric flow rate, Q, at its maximum and minimum can be calculated based on the
number of chambers in the system. The combination of the hydrologic calculations and
Darcy’s equation yielded that twelve chambers total were needed at Site L and 22
chambers total at Site M to divert a rain event of intensity 12.5 mm (0.5 inch).
4.3.2 DIS DESIGN
The DIS utilized pre-existing stormwater piping system, by designing a divergent
monitoring vault to intercept the pipe at the west end of the dunes. The depth of burial
was not designed; rather it was a result of the original outfall conditions. Site L’s system
was deeper that Site M because the outfall was deeper at the connection point. The
invert elevations of the StormChambers™ allowed an average of 2.0 m (6.5 ft) for Site L
and 2.6 m (8.5 ft) for Site M of sand for the stormwater runoff to infiltrate through before
reaching the groundwater. This also allowed the 1.07 m (3.5 ft) StormChambers™ to be
buried 0.8 m (2.5 ft) at Site L and 0.5 m (1.5 ft) at Site M under the dunes, which added
protection to the DIS systems. The StormChambers™ inflow pipe inverts were 0.6 m (2
ft) higher than the StormChambers™ invert elevation to allow for a passive system. A 31
cm (12 in) diameter pipe, sloping less than 0.01 m/m, lead from the diversion vault to the
StormChambers™. As seen in Figure 4-10, stormwater from the outfall is diverted
within a buried concrete vault into a “T” intersection, allowing for two separate laterals
of StormChambers™ for flexibility if stormwater runoff debris clogged one of the
entrances or in the event routine maintenance was performed. Clean-out pipes were
designed and installed at the beginning and end of each StormChamber™ row, to
51
facilitate maintenance, which the Town of Kure Beach Department of Public Works
agreed to perform.
WT Observation
WQ Wells
WT Observation
WQ Wells
Overflow
Isco 6712
Portable Sampler ™
Maintenance/Access
point
Continuous Internal
Water Level Recorder
Weir
Sample
tube
Sediment Trap
Float-Pulley System
Inflow
Isco 730 Bubbler
Module™
Figure 4-10. Top view of DIS layout.
4.3.3 DIS INSTALLATION
The Town of Kure Beach Public Works Department, under the supervision of North
Carolina State University, installed the Dune Infiltration System in February, 2006. In
order to install the chambers in the dunes, a trench 2.7 m (9-ft) wide by 1.8 m (6-ft) deep
by 29 m (96 ft) long for Site L and 2.7 m (9-ft) wide by 1.5 m (5-ft) deep by 54 m (176
ft) long for Site M was constructed. Banks were stabilized with a geotextile fabric.
Then, 15.2 to 30.5 cm (6 to 12 inches) of 2.5 to 5.1 cm (1 to 2 inch) washed stone was
placed on top of the sand at the bottom of the trench to achieve uniform grade. Heavy
duty nylon netting was placed on top of the stones to secure them during future
maintenance (sediment removal) from the chambers. The chambers were placed on top
of the netting, and the trench and chambers were filled midway with washed stone
52
(Figure 4-11 (a)). The 31cm (12 in) pipe from the diversion vault was attached to the
start chambers and sealed.
(a)
(b)
Figure 4-11. Installation of a StormChamber™.
Access points for chamber maintenance were installed at the beginning and end of
the chambers. Water level monitoring points, 10 cm (4 in) in diameter, were also
installed at these locations within the chambers to accommodate an INFINITY™ water
level recorder (Figure 4-11 b).
Upon completion of installation, American beach grass (Ammophila
breviligulata) was replanted in the dunes in Site L and Site M in March, 2006 to help
initially stabilize the dunes. Fertilizer was spread, then sea oats (Uniola paniculata), a
nature NC plant, were planted during the plants’ optimal survival period, June, 2006, to
more effectively vegetate and stabilize the dunes. Figure 4-12 shows Kure Beach
volunteers planning sea oats.
53
Figure 4-12. Planting Sea oats in Site M’s dunes.
4.3.4 DIS MONITORING
4.3.4.1 Monitoring Equipment
The amount of stormwater diverted into the chambers was calculated from measurements
recorded in the monitoring vault. Figure 4-13 shows the two diversion vaults’ design
schematic.
Figure 4-13. AutoCAD drawing of Site L and Site M vault (feet unless otherwise noted).
54
An ISCO 730 Bubbler Module™ was attached to the bottom of the existing stormwater
outflow pipe. Manning’s equation (equation 4-9), with a maximum Manning’s n for
corrugated metal pipe, flowing full, near a manhole, of 0.024 was used to program the
ISCO Bubbler to calculate inflow rate of stormwater runoff into the vault.
2
R 3S
Q=
n
1
2
(4-9)
Where: Q= Discharge (m3/s)
R= Hydraulic Radius (m)
S = Friction Slope (m/m)
Since this was a demonstration of a new concept in coastal stormwater
management, the Dune Infiltration Systems were designed to capture only the amount of
stormwater runoff produced by an 1-hour storm with an intensity of 1.3 cm (0.5 inch).
For storms greater than 1.3 cm/hr (0.5 in/hr) design intensity, bypass was expected. The
overtop was designed to overflow a rectangular weir in the vault and discharge through
the original stormwater pipe onto the beach. To calculate the volume of overflow, 0.6 m
(2 ft) high, 1.2 m (4.0 ft) wide, and 0.8 m (2.5 ft) rectangular concrete weir without end
contractions was constructed as part of the vault (Figure 4-13). The weir was positioned
0.6 m (2 ft) from the connection with original outflow stormwater pipe. A 0.15 m (0.50
ft) metal plate was later drilled onto the concrete weir, giving the weir a total height of
0.76 m (2.5 ft). With the known type and height of the weir, the overflow rate was
calculated using the equation below (Grant & Dawson 2001).
Q = 6618*(1.2 H − 0.2 H ) * H 1.5
Where : Q = Discharge (m3/s)
H = Head Over Weir (m)
(4-10)
55
The head over the weir during a storm event was measured and recorded using a
Sargent® (SGT Engineering, Champaign, Illinois) float-pulley system which was
constructed and installed in the vault near the inflow stormwater pipe (Figure 4-14). The
Sargent float-pulley system consisted of a 5-turn pulley connecting a float with a
matching counterweight, which was attached to a 10-KΩ Newark™ potentiometer.
Figure 4-14. View of monitoring vault from manhole.
The potentiometer was attached to a two channel 12-bit Sargent data logger. The
system is powered by a 12-volt brick battery (Figure 4-15). A stilling well was
constructed with a wooden box with drilled 1.3 cm (0.5 in) holes and encased the Sargent
float-pulley system to damper the effect of turbulence on the float during large storm
events.
56
Figure 4-15. Sargent pulley tape system.
In instances when the system was overwhelmed with stormwater runoff, the
calculated overflow volume was subtracted from the total measured inflow, obtaining the
volume treated in the DIS. If no flow was recorded over the weir, then the flow into the
chambers was reasonably assumed to be equal to that measured from the inflow pipe.
The bacteria concentration entering the system was measured by water quality
grab samples captured during a storm. At Site L and Site M an ISCO 6712 Portable
Sampler™ was programmed using Manning’s equation to capture stormwater runoff at
flow weighted points along the inflow hydrograph. Site L’s ISCO was programmed to
capture a 200 ml sample for every 2.4 m3 (85 ft3) of stormwater runoff that entered the
vault. Site M’s ISCO was set to capture a 200 ml sample for every 3.9 m3 (137 ft3) of
stormwater that entered the vault. Samples were collect in the vault 15 cm (6 in) below
the pipe that leads to the StormChambers™. A 0.6 m (2.0 ft) diameter, 10 cm (4 in) tall
ring with an 8 cm (3 in) slit was inserted between the manhole and the manhole cover to
allow the ISCO bubbler and sampling tubing to exit the vault. The ring was caulked with
silicone to prevent the tubing from being cut by the ring each time the manhole cover was
57
removed. The ISCO samplers were each powered with a 12-volt battery that was
recharged by a 15 W Solarex Solar™ Panel. The samplers were individually stored in a
locked JOBOX from Ben Meadows Company (Figure 4-16).
Figure 4-16. ISCO sampler in JOBOX.
Groundwater wells to monitor bacteria were constructed and installed in
duplicate at each site. These were installed approximately 1 m (3.1 ft) down slope of the
Dune Infiltration System and 0.6 m (2-ft) and 1.2 m (4-ft) below the bottom of the
chambers to capture and store samples following rainfall events (Figure 4-17).
Cleanout Pipes
Monitoring
Vault
Groundwater
Wells
INFINITY
Rain Gauge
JOBOX with ISCO
Figure 4-17. DIS system at Site L after installation.
58
4.3.4.2 Sampling Collection Protocol
Samples were collected during the months of March, 2006, through October, 2006,
within 24 hours of a storm event. Two samples were collected for each site in order to
analyze for both fecal coliform and enterococcus. Fecal coliform samples were taken to
Oxford Laboratory, Inc., Wilmington, NC, as described in the preconstruction sampling
protocol. Enterococcus samples were collected in a 60 ml sterile bottle and taken to the
NCDENR Division of Shellfish Sanitation Laboratory located in Wrightsville Beach, NC,
who analyzed water samples using the IDEXX Laboratories, Inc., developed method,
Enterolert™ (ASTM method D6503-99).
Two sets of samples were collected from each site: inflow stormwater runoff and
groundwater. Stormwater runoff bacteria samples were collected from the ISCO
samplers at each site. The sample was collected during the storm in 1 liter bottles and
stored inside the sampler. If the storm event was large enough to fill more than a 1 liter
bottle, a composite sample was taken to be analyzed. A composite sample was formed
by taking the same volume of sample per liter bottle filled. Groundwater samples were
obtained from L-12 and M-12, as described in the preconstruction sampling protocol,
since the groundwater wells installed with the DIS remained dry.
4.4 DIS RESULTS AND DISCUSSION
4.4.1 PRECONSTRUCTION RESULTS AND DISCUSSION
From July, 2005, through September, 2005, bacteria samples were collected from the
groundwater wells. The groundwater samples were collected after five rain events; three
from groundwater wells at Site L: L-10, L-12, and L14 and two from Site M: M-12 and
59
M-14. Site M surface elevation was higher than Site L, the M-10 monitoring well was
dry; so no groundwater bacteria samples were collected.
Site L’s groundwater fecal coliform colony forming units (CFU) ranged from less
than 1 CFU/100 ml to 190 CFU/100 ml; whereas, Site M’s ranged from less than 1
CFU/100 ml to 200 CFU/100 ml. All groundwater fecal coliform levels remained equal
to or less than North Carolina’s standard of 200 CFU/100 ml. Table 4-2 shows the
groundwater bacteria concentrations for the five storms. Oxford Laboratory bottles were
not used on July 12, 2005, which may explain the increased bacteria counts.
Table 4-2. Preconstruction groundwater fecal coliform levels.
7/12/2005
7/24/2005
8/10/2005
8/24/2005
9/21/2005
L-10
CFU/100ml
115
<1
15
<1
<1
L-12
CFU/100ml
110
<1
23
<1
22
L-14
CFU/100ml
190
11
<1
<1
<1
M-12
CFU/100ml
200
1
12
<1
<1
M-14
CFU/100ml
54
66
11
29
29
Runoff from four storms was sampled directly from Site L’s and Site M’s ocean
outfall pipes. Site L’s stormwater fecal coliform levels ranged from 1,300 to 22,300
CFU/100 ml; where as Site M’s ranged from 1,820 to 6,000 CFU/100ml. All stormwater
inflow rates exceeded the state’s standard. Table 4-3 shows the stormwater runoff
bacteria concentrations.
Table 4-3. Preconstruction stormwater runoff bacteria levels.
7/12/2005
8/10/2005
8/23/2005
10/24/2005
Site L
CFU/100
ml
5320
7240
22300
1300
Site M
CFU/100
ml
6000
1820
3000
2200
60
Site L’s and Site M’s groundwater elevation ranged from 2.6 m (8.5 ft) and 3.6 m
(11.6 ft) below the dune surface (Figure 4-18 and Figure 4-19). Data is missing from Site
M’s groundwater elevation from September 22, 2005, until October 16, 2005, due to
equipment malfunction.
Site L’s three large groundwater elevation peaks (Site M’s two large peaks)
shown in Figure 4-18 (Figure 4-19) were fluctuations in the groundwater caused by
tropical systems. The largest peak in the groundwater occurred on September 14, 2005,
during Hurricane Ophelia. This storm’s rainfall total was approximately 432 mm (17 in),
which caused Site L’s groundwater to rise 1.5 m (4.9 ft) and 2 m (6.6 ft) rise in Site M’s
stormwater. Next was Tropical Storm Tammy on October 6, 2005, shown only in Site
L’s groundwater. Around 130 mm (5 in) of rain caused water table to rise 1.0 m (3.3 ft).
Then on October 23, 2005, remnants of Hurricane Wilma caused a rainfall total of about
8 cm (3 in) with a corresponding rise of 0.7 m (2.3 ft) in Site L’s groundwater.
5
4.5
Invert of Dune Infiltration System as designed= 4.45 m
4
Approximately 2 m depth of sand
EL (m)
3.5
3
2.5
2
1.5
1
Sensor Elevation =1.0 m
0.5
0
06/27/05
08/16/05
10/05/05 Date
11/24/05
Figure 4-18. Preconstruction groundwater elevations at Site L.
01/13/06
EL (m)
61
Invert of Dune Infiltration System as designed = 4.75 m
5
4.5
Approximately 2.5 m depth of sand
4
3.5
3
2.5
2
1.5
Sensor Elevation =1.6 m
1
0.5
0
06/27/05
08/16/05
10/05/05 Date
11/24/05
01/13/06
Figure 4-19. Preconstruction groundwater elevations at Site M.
4.4.2 POST CONSTRUCTION HYDRAULIC DATA
4.4.2.1 Summary of Storm Events
Twenty-five storm events were captured during the months of March through October,
2006 (Table 4-4 and Table 4-5). A storm event was defined as rainfall separated from
another by an inter-event dry period of at least 6 hours. Storm intensity was calculated
using the US EPA procedure for 2-yr-15 minute storms (U.S. EPA, 2002b).
The following hydraulic data was measured for each storm: rainfall amount,
rainfall duration, inflow rates, inflow durations, water level in the beginning and end of
the chambers, and the stage of the water in the monitoring vault. From this data the
following was calculated: peak rainfall intensity, peak inflow rate, total runoff volume,
volume runoff treated, and volume runoff overflow.
62
Table 4-4. Site L Storm Characteristics.
Storm
Date
Rainfall
Amount
Duration
Peak
Intensity
Peak
Flow
3/21/2006
4/16/2006
4/26/2006
5/7/2006
5/14/2006
5/15/2006
5/20/2006
6/5/2006
6/12/2006
6/14/2006
6/25/2006
6/26/2006
6/27/2006
7/6/2006
7/16/2006
7/23/2006
7/25/2006
7/30/2006
8/21/2006
8/22/2006
9/1/2006
9/6/2006
9/13/2006
10/8/2006
10/17/2006
(mm)
11.9
19.3
26.4
13.0
20.6
3.8
23.4
9.1
7.9
17.0
8.4
6.6
5.6
11.7
4.6
40.1
29.0
4.1
10.7
48.8
105.2
8.6
49.8
76.2
6.6
(hr)
10.3
N/A
N/A
2.0
3.3
0.3
19.1
5.9
11.6
4.0
2.4
3.5
6.1
4.8
1.8
24.3
23.3
8.2
0.7
6.4
21.8
13.1
10.8
15.3
18.9
(mm/hr)
2.79
N/A
N/A
33.53
30.48
14.30
10.20
41.15
5.08
39.62
73.15
15.25
11.12
27.94
18.30
50.80
43.69
1.27
19.30
88.90
22.86
12.19
6.10
88.90
4.32
(m3/s)
0.002
0.026
0.012
0.011
0.006
0.005
0.004
0.011
0.002
0.016
0.008
0.005
0.003
0.003
0.006
0.017
0.019
0.001
0.004
0.014
0.012
0.003
0.013
0.039
0.002
Runoff
Watershed
Depth*
(mm)
0.57
0.95
3.10
0.70
1.06
0.41
1.66
1.79
0.87
1.13
0.42
0.43
0.50
0.40
0.28
2.15
2.01
0.51
0.32
1.20
6.60
0.70
2.87
4.79
0.78
Total=
Total
Runoff
Volume
Captured
Total
Runoff
Volume
Bypass*
(m3)
10.3
17.3
56.4
12.8
19.3
7.4
30.2
32.7
15.8
20.5
7.6
7.8
9.1
7.2
5.1
39.2
36.5
9.4
5.9
21.9
120.2
12.8
52.2
87.2
14.2
659
(m3)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
*Indicates calculated values, the rest were directly measured
All this information can be found on a per storm basis in Appendix A. The
statistical output for the hydraulic data can be found in Appendix B. Table 4-5 and Table
4-6 summarizes each storm’s characteristics, although rainfall duration and peak intensity
are missing for the April storms due to tipping bucket malfunctions.
63
Table 4-5. Site M Storm Characteristics.
Storm
Date
3/21/2006
4/16/2006
4/26/2006
5/7/2006
5/14/2006
5/15/2006
5/20/2006
6/5/2006
6/12/2006
6/14/2006
6/25/2006
6/26/2006
6/27/2006
7/6/2006
7/16/2006
7/23/2006
7/25/2006
7/30/2006
8/21/2006
8/22/2006
9/1/2006
9/6/2006
9/13/2006
10/8/2006
10/17/2006
Rainfall
Amount
(mm)
11.9
19.3
26.4
13.0
20.6
3.8
23.4
9.1
7.9
17.0
8.4
6.6
5.6
11.7
4.6
40.1
29.0
4.1
10.7
48.8
105.2
8.6
49.8
76.2
6.6
Duration
(hr)
10.3
N/A
N/A
2.0
3.3
0.3
19.1
5.9
11.6
4.0
2.4
3.5
6.1
4.8
1.8
24.3
23.3
8.2
0.7
6.4
21.8
13.1
10.8
15.3
18.9
Peak
Intensity
(mm/hr)
2.79
N/A
N/A
33.53
30.48
14.30
10.20
41.15
5.08
39.62
73.15
15.25
11.12
27.94
18.30
50.80
43.69
1.27
19.30
88.90
22.86
12.19
6.10
88.90
4.32
Peak
Flow
(m3/s)
0.002
0.048
0.043
0.028
0.013
0.010
0.017
0.030
0.002
0.053
0.023
0.014
0.005
0.019
0.015
0.059
0.062
0.002
0.015
0.055
0.047
0.009
0.054
0.180
0.004
Runoff
Watershed
Depth*
(mm)
0.70
1.36
5.79
1.18
1.38
0.41
2.07
3.26
1.16
2.29
0.72
0.53
0.53
0.90
0.41
5.58
5.14
0.59
0.69
2.77
17.05
1.60
6.65
11.13
0.67
Total =
Total
Runoff
Volume
(m3)
22.8
44.3
189.1
38.5
45.2
13.3
67.5
106.6
37.7
72.7
23.6
17.4
17.3
29.4
13.5
175.8
163.0
19.2
22.6
87.9
556.7
52.3
217.0
280.2
21.9
2336
Total
Runoff
Volume
Bypass*
(m3)
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2.1
0.0
0.0
0.0
0.0
0.0
6.3
4.8
0.0
0.0
2.6
0.0
0.0
0.0
83.3
0.0
99
Seventeen of the 23 storms rainfall intensity exceeded the design intensity of 13
mm/hr (0.5in/hr), averaging 28.9 mm/hr (1.14 in/hr). Figure 4-20 is a graph of rainfall
intensity versus rainfall amount, showing the variety of storms captured. The intensity
ranged from 1.27 mm/hr (0.05 in/hr) on July 30, 2006, to 89 mm/hr (3.5 in/hr) on July 23,
2006, and October 8, 2006. The mean rainfall amount was 22.7 mm (0.89 in) and ranged
64
from 3.81 mm (0.15 in) occurring May 15, 2006, to 105.2 mm (4.14 in), occurring
Rainfall Intensity (mm/hr)
September 1, 2006, during Tropical Storm Ernesto.
100.00
80.00
60.00
40.00
20.00
0.00
0.00
20.00
40.00
60.00
80.00
100.00
120.00
Rainfall Amount (mm)
Under 12.5 mm/hr
Over 12.5 mm/hr
Figure 4-20. Rainfall intensity versus rainfall amount.
The majority of these storms are categorized as Type III storms, with relative
short durations of peak intensity occurring at the beginning of the storms. Type III
storms represent Gulf of Mexico and Atlantic coastal area where tropical storms bring
large 24-hour rainfall amounts (Schwab et al. 1993). The months of March through
October 2006, were of average rainfall relative to the last decade of rainfall events
measured from New Hanover County Airport ,Wilmington, North Carolina (State
Climate Office of North Carolina 2006).
4.4.2.2 Groundwater Results and Discussion
Figures 4-21 and 4-22 show the variation of groundwater for Site L elevations between
the months of July through October, both before and after the DIS was implemented in
2005 and 2006.
EL (m)
65
5
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
27-Jun 12-Jul
27-Jul 11-Aug 26-Aug 10-Sep 25-Sep 10-Oct 25-Oct
Date
2005
2006
Figure 4-21. Site L groundwater fluctuations from July to October 2005 and 2006.
The water table elevations for July through October, 2006, are similar to the water
table elevations for July through October, 2005. As previously discussed, in 2005 there
were 3 large storms, Hurricane Ophelia, Tropical Storm Tammy, and Hurricane Wilma.
In 2006 there was only one large storm event, Tropical Storm Ernesto.
5
4.5
4
El (m)
3.5
3
2.5
2
1.5
1
0.5
0
27-Jun
12-Jul
27-Jul
11-Aug
26-Aug
Date
10-Sep
25-Sep
2005
10-Oct
25-Oct
2006
Figure 4-22. Site M groundwater fluctuations from July to October 2005 and 2006.
66
The statistical analysis did not take into account rainfall variation in the two years. The
amount of rainfall affects the level of groundwater, since the rainfall amount established
the volume of water available to runoff or percolate into the groundwater.
The tide also influenced the water table elevation. Figure 4-23 shows the effect of
tidal fluctuations on the water table elevation of Site L and Site M. Tidal data were
obtained from a NOAA station located in Wrightsville Beach, located about 20 miles
north of Kure Beach (NOAA 2006). The datum for the tidal data was taken from the
mean lower low water (MLLW), which is defined as the average height of the lower low
3.5
3.5
3.0
3.0
2.5
2.5
2.0
2.0
1.5
1.5
1.0
1.0
0.5
0.5
0.0
0.0
-0.5
27-Jun
17-Jul
6-Aug
26-Aug
15-Sep
5-Oct
Tidal fluctuation (m)
Watertable Elevation
(m)
waters at a location over a 19-year period. (IHO 2001).
-0.5
25-Oct
Date
Date
(2006)
Site L
Site M
Tide Data
Figure 4-23. Wrightsville Beach tidal influences on groundwater elevations in Kure
Beach, NC.
Tide elevations varied from -0.36 m (-1.18 ft) to 2.11 m (6.92 ft), yielding a 2.5 m
(8.1 ft) difference. The water table elevation range at Site L was 1.92m (6.30 ft) to 3.98
m (13.1 ft) in 2005 and 1.73 m (5.68 ft) to 2.84 m (9.3 ft) in 2006. Site M’s water table
elevation range was 1.77 m (5.81 ft) to 3.88 m (12.7 ft) in 2005 and 1.71 m (5.61 ft) to
67
3.50 (11.5 ft) in 2006. Site L’s and Site M’s water table elevation ranges in 2005 and
2006 were less than the tidal fluctuations.
Figure 4-24 shows the groundwater fluctuations from February, 2006, until
October, 2006. The biggest groundwater fluctuations occurred during the two largest
recorded rain events. These were Tropical Storm Ernesto and the October 8, 2006, storm,
with rain totals of 105 mm (4.12 in) and 76 mm (3.0 in). At site M there was a 1.5 m (4.9
ft) rise during Tropical Storm Ernesto and a 1 m rise on October 8, 2006. For Site L
there was only a 0.5 m (3.3 ft) rise during Tropical Storm Ernesto and a 0.75 m (2.5 ft)
rise during the October 8, 2006 storm. This rise was a combination of stormwater
infiltrating into the groundwater as well as a tide change from high to low. For Tropical
Storm Ernesto, Site M rose 1 m (3.3 ft) more than Site L, indicating an increase in
groundwater due to the stormwater runoff infiltrating, since Site M produces more runoff
than Site L. Figure 4-23 shows the largest peak in the groundwater occurred on
September 14, 2005, during Hurricane Ophelia. This storm’s rainfall total was
approximately 43 cm (17 in), which caused Site L’s groundwater to rise 1.5 m (4.9 ft)
and a 2 m (6.6 ft) rise in Site M’s stormwater. Without the DIS, large storms increased
groundwater levels. When Hurricane Ophelia occurred, the tide was receding, so that
fluctuation is expected to have been from the storm. Storm sizes and intensities appear to
have a more pronounced impact on groundwater elevations than the incorporation of the
DIS.
68
5
4.5
Tropical Storm
Ernesto
4
EL (m)
3.5
3
2.5
2
1.5
1
0.5
0
01/28/06
03/19/06
05/08/06
06/27/06
Date
08/16/06
Site L
10/05/06
Site M
Figure 4-24. Site L and Site M fluctuations in groundwater since DIS implementation.
Routing large the amounts of stormwater runoff through the dunes has not had a
strong effect on water table elevations. The tidal fluctuation remained greater than the
variation of water table elevation at Site L and Site M. Thus, for a watershed less than
3.3 ha (8.1 acre) with groundwater elevation greater than 2.5 m (8.1 ft), a DIS designed to
capture storms with an intensity of 13 mm/hr (0.5 in/hr) or less should not hydraulically
overload the groundwater.
4.4.2.3 Flow Mitigation Results and Discussion
4.4.2.3.1 Site L Results and Discussion
The 25 storms analyzed generated 659 m3 (23, 272 ft3) of stormwater runoff from the L
watershed, ranging from 5.1 m3 (180 ft3) to 120 m3 (4,237 ft3), and averaging 26.4 m3
(932 ft3). No incidents of system overflow were measured. Therefore, as hypothesized
the volume of stormwater runoff captured in the DIS was significantly greater than the
69
volume of stormwater runoff that bypassed the DIS (p<0.01). Figure 4-25 depicts the
volume of stormwater runoff captured per storm.
Total Runoff Volume (m
3
)
120
100
80
60
40
`
20
17-Oct,
13-Sep
1-Sep
21-Aug
25-Jul
16-Jul
27-Jun
25-Jun
12-Jun
20-May
14-May
26-Apr
21-Mar
0
Storm Date (2006)
Figure 4-25. Volume of runoff captured Site L.
The largest runoff volume captured occurred on September 1, 2006, during
Tropical Storm Ernesto. The peak intensity of Ernesto at Kure Beach was 23 mm/hr
(0.89 in/hr), resulting in a peak runoff rate of 0.012 m3/s (0.424 cfs). This rate was
substantially less than the infiltration rate within the sand dunes. The water level rise in
the beginning chambers was 0.17 m (0.55 ft) out of the possible 1.01m (3.34 ft) of
storage height.
Since there was no bypass flow, the peak inflow rate of stormwater runoff
entering the DIS for Site L was significantly greater than the peak rate bypassing the DIS
(p<0.01). Peak flow into the system ranged from 0.0007 m3/s (0.0247 cfs) to 0.0391 m3/s
(1.380 cfs), with a mean of 0.0098 m3/s (0.3461 cfs). The maximum peak intensity
occurred during an October 8, 2006 storm event, which caused the stage in the
70
monitoring vault to rise within 4.2 mm (0.14 in) of the overflow weir. Figure 4-26 shows
the various peak flow inflow rates per storm.
0.045
Qp, Peak Inflowrate (m3/s)
0.040
0.035
0.030
0.025
0.020
0.015
0.010
0.005
17-Oct,
13-Sep
1-Sep
21-Aug
25-Jul
16-Jul
27-Jun
25-Jun
12-Jun
20-May
14-May
26-Apr
21-Mar
0.000
Storm Date (2006)
Figure 4-26. Site L peak inflow per storm.
Figure 4-27 shows the inflow hydrograph of both Tropical Storm Ernesto and the
October 8, 2006, storm. As noted in Figures 4-25 and 4-26, the runoff volume and peak
runoff rates for Tropical Storm Ernesto were 120 m3 (4237 ft3) and 0.012 m3/s (0.424
cfs), and 87.2 m3 (3079 ft3) and 0.0391 m3/s (1.380 cfs) for the October 8, 2006, storm.
During Tropical Storm Ernesto, the maximum stage in the vault was 0.50 m (1.64 ft). In
comparison, the October 8th storm maximum stage reached 0.72 m (2.36 ft), almost
overflowing the bypass weir. This may be attributed to the October 8, 2006 storm’s peak
inflow rate exceeding Tropical Storm Ernesto’s by more than a factor of three.
71
0.05
Weir Elevation = 0.76 m0.75
0.04
0.65
0.03
0.55
0.02
0.45
0.01
0.35
0
8/31/2006
0:00
8/31/2006
4:48
8/31/2006
9:36
8/31/2006
14:24
8/31/2006
19:12
9/1/2006
0:00
9/1/2006
4:48
Stage (m)
Inflow rate (m 3/s)
Tropical Storm Ernesto
0.25
9/1/2006
9:36
Date
Inflow
Stage in Vault
Weir Elevation =0.76 m
0.05
0.75
0.04
0.65
0.03
0.55
0.02
0.45
0.01
0.35
0
10/8/06
9:36
10/8/06
12:00
10/8/06
14:24
10/8/06
16:48
10/8/06
19:12
10/8/06
21:36
10/9/06
0:00
10/9/06
2:24
10/9/06
4:48
Stage (m)
Inflow rate (m3/s)
October 8, 2006
0.25
10/9/06
7:12
Date
Inflow
Stage in Vault
Figure 4-27. Site L Tropical Storm Ernesto and October 8, 2006 inflow hydrograph.
4.4.2.3.2 Site M Results and Discussion
The volume of stormwater runoff captured in the DIS at Site M was significantly greater
than the volume of stormwater runoff that bypassed by the DIS (p<0.001). Five of the 25
storms caused overflow of Site M’s DIS system, but 97% of the total measured inflow
volume was captured (Figure 4-28).
72
300
250
200
Runoff
Volume 150
(m3)
100
50
21 Ma
16- r
Ap
26- r
Ap r
7-M
a
14 - y
Ma
15 - y
Ma
20 - y
Ma
y
5-J
un
12 Jun
14 Jun
25 Jun
26 Jun
27 Jun
6-J
ul
16Jul
23Jul
25Jul
30 Ju
21 - l
Aug
22 Au g
1-S
ep
6-S
ep
13 Se p
8-O
17- ct
Oc
t,
0
Volume Captured
Volume Bypassed
Figure 4-28. Volume of runoff captured versus overflow per storm at Site M.
The volume of the 20 storms completely captured ranged from 13.3 m3 (470 ft3) to 557
m3 (19,670 ft3), averaging 77.8 m3 (2,747 ft3). For the 5 bypassing storms the total runoff
volume (including volume captured and volume passed) ranged from 74.8 m3 (2,642 ft3)
to 364 m3 (12,855 ft3), averaging 176 m3 (6,215 ft3). Table 4-6 summarizes the volume
of bypassed storm’s runoff that was either captured or bypassed, as well as the rainfall
amount and peak inflow rate.
Table 4-6. Site M summary result of bypassing storms.
Stormdate
6/14/06
7/23/06
7/25/06
8/22/06
10/8/06
Rainfall
Amount
Peak Intensity
mm (in)
mm/hr (in/hr)
17.02 (0.67) 27.94 (1.10)
40.13 (1.58) 88.90 (3.50)
29.96 (1.14) 27.94 (1.10)
34.5 (1.28)
52.07 (2.05)
76.20 (3.00) 88.90 (3.50)
Rainfall
Durantion
hr
4.02
24.32
23.27
6.38
15.33
Peak Inflow
Runoff Rate
3
m /s (cfs)
0.053 (1.87)
0.059 (2.09)
0.062 (2.20)
0.055 (1.95)
0.180 (6.35)
Stormwater
Entering vault *
3
3
m (ft )
75 (2640)
182 (6431)
168 (5927)
90 (3194)
363 (12836)
Overflow
Amount
3
3
m (ft )
2.1 (75)
6.3 (222)
4.8 (168)
2.6 (90)
83.3 (2942)
*Note: From outfall leading from Site M
The largest runoff volume was from September 1, 2006, Tropical Storm Ernesto,
shown in Figure 4-29. The runoff volume from this storm almost doubled the maximum
bypassing storm’s runoff volume, but the stage in the beginning chambers only rose to
73
0.39 m (1.26 ft) out of the possible 1.01m (3.34 ft) of storage height. This was due to the
relatively low peak inflow rate of Tropical Storm Ernesto, 0.047 m3/s (1.660 cfs). The
water in the monitoring vault rose to a stage of 0.71 m (2.33 ft), less than the 0.76 m (2.5
ft) necessary to create bypass. As shown in Figure 4-29, Tropical Storm Ernesto, lasted
almost 24 hours, but exhibited staged rainfall. This allowed the previous runoff to
percolate into the system before the next relatively high intensity part of the storm.
As hypothesized for Site M the peak inflow rate of stormwater runoff entering the
DIS was significantly greater than the peak flow rate bypassing the DIS (p<0.01). Peak
flow into the system ranged from 0.002 m3/s (0.071 cfs) to 0.114 m3/s (4.026 cfs),
0.8
0.05
Weir Elevation = 0.76 m
0.045
0.7
0.04
0.6
0.035
0.5
0.03
0.025
0.4
0.02
0.3
0.015
0.2
0.01
0.1
0.005
0
0
8/31/2006 8/31/2006 8/31/2006 8/31/2006 8/31/2006 9/1/2006 9/1/2006 9/1/2006 9/1/2006
0:00
4:48
9:36
14:24
19:12
0:00
4:48
9:36
14:24
Date
Inflow
Stage in Chamber
Stage In Vault
Figure 4-29. Site M inflow hydrograph, stage in vault and stage in StormChambers
during Tropical Storm Ernesto (8/31/06-9/01/06).
Peak inflow rates exceeding 0.051 m3/s (1.80 cfs) caused bypass (Figure 4-30).
The bypass flow rate ranged from 0.009 m3/s (0.328 cfs) to 0.156 m3/s (5.513 cfs),
averaging 0.045 m3/s (1.589 cfs), Figure 4-31.
Stage (m)
Flowrate (m 3/s)
averaging 0.030 m3/s (1.059 cfs).
Peak Inflow rate (m3/s)
74
0.2
0.18
0.16
0.14
0.12
0.1
0.08
0.06
0.04
0.02
0
2/2/2006
3/24/2006
5/13/2006
7/2/2006
8/21/2006
10/10/2006
11/29/2006
Date of Storm
Captured
Overflowed
Figure 4-30. Peak inflow rate for captured and bypass storm at Site M.
0.20
0.18
0.16
Peak 0.14
Flow 0.12
(m3/s) 0.10
0.08
0.06
0.04
0.02
21Ma
r
16Apr
26Apr
7-M
ay
14Ma
y
15Ma
y
20Ma
y
5-J
un
12Jun
14Jun
25Jun
26Jun
27Jun
6-J
ul
16Jul
23Jul
25Jul
30Ju
21 - l
Aug
22Aug
1-S
ep
6-S
ep
13Sep
8-O
c
17- t
Oct
,
0.00
Peak Inflow
Peak Overflow
Figure 4-31. Peak inflow rate versus peak outflow rate per storm at Site M.
One captured storm that exceeded the June 14, 2006, peak inflow rate, occurred
on September 13, 2006, characterized by a peak inflow rate equaling 0.054 m3/s (1.910
cfs). Figure 4-32 shows the inflow hydrograph and the stage in the vault for June 14
compared to September 13.
75
Figure 4-32. Site M comparison of June 14, 2006 and September 13, 2006 inflow
hydrographs.
On June 14 the vault’s stage reached a height of 0.77m (2.53 ft). On September 13 the
runoff height in the vault reached 0.73 m (2.40 ft), 0.03 m (0.10 ft), less than required to
over-top the overflow weir. The June 14 storm produced overflow for approximately
seven minutes; whereas, the September 13 storm did not. This result was attributable to
the more sustained peak inflow rate on June 14 storm relative to September 13. This is
demonstrated by a wider peak in June 14’s hydrograph than that of September 13’s
hydrograph. With this exception, all of the other storms with a peak runoff rate of less
76
than 0.053 m3/s (1.872 cfs) were completely routed through the DIS. Statistics used to
compare runoff volume of captured storms to bypass storms showed no statistical
significance at α=0.05.
The storm’s rainfall intensity can be used to predict if a storm will overflow the
system. Rainfall intensity is significantly (p<0.05) predictive of bypass at Site M. Figure
4-33, shows the rainfall intensity versus rainfall amount for captured storms. A rainfall
event with an intensity of 27.9 mm/hr (1.1 in/hr) and amount greater than 17 mm (0.67
in) caused bypass.
Rainfall Intesity (mm/hr)
90
80
70
60
50
Captured
40
Overflow
30
20
10
0
0
20
40
60
80
100
Rainfall Amount (mm)
Figure 4-33. Peak rainfall intensity versus rainfall amount for captured and bypassed
storms for Site M.
Figure 4-33, shows one captured storms with high rainfall intensity, but low
rainfall amount. This storm, circled in Figure 4-33, occurred on May 7, 2006. This May
storm was a 2 hour event with a rainfall amount of 13 mm (0.5 in) and peak intensity of
78.7 mm/hr (3.10 in/hr). There has been no rain for five days prior to this storm. Thus,
the dry watershed’s soil was able to infiltrate some of the rainfall allowing for maximum
77
peak flow rates of only 0.028m3/s (0.989 cfs). This peak rate was able to be captured by
the DIS, since the diversion pipe leading to the chambers and the infiltration rate of the
soil had a greater flow rate than the peak flow of the stormwater runoff.
4.4.2.4 Design Discussion
Figures 4-34 and 4-35 shows the difference in runoff volume and peak inflow rate for
storms at Site L and Site M.
100
80
Site M
60
Site L
40
20
0
21
-
M
16 ar
-A
p
7- r
M
a
14 y
-M
15 ay
-M
a
5- y
Ju
12 n
-J
u
14 n
-J
u
25 n
-J
u
26 n
-J
u
27 n
-J
un
6Ju
16 l
-J
u
30 l
-J
21 ul
-A
u
6- g
Se
17 p
-O
ct
,
Total Runoff Volume (m3/s)
120
Storm Date (2006)
Figure 4-34. Variation in runoff volume for Site L and Site M.
Qp, Peak Inflowrate (m3/s)
0.060
0.050
0.040
Site M
0.030
Site L
0.020
0.010
21
-
M
16 ar
-A
p
7- r
M
14 ay
-M
15 ay
-M
a
5- y
Ju
12 n
-J
u
14 n
-J
u
25 n
-J
u
26 n
-J
u
27 n
-J
un
6Ju
16 l
-J
u
30 l
-J
21 ul
-A
u
6- g
Se
17 p
-O
ct
,
0.000
Storm Date (2006)
Figure 4-35. Variation in peak inflow rate for Site L and Site M.
78
Difference in runoff volume and peak discharge was expected for Site L and Site
M. Site L is a 1.8 ha (4.5 acres) with CN = 88, while Site M is a 3.3 ha (8.1-acre) with
CN = 69. The NRCS method predicted a runoff volume of 15 m3 (530 ft3) for Site L and
31 m3 (1098ft3) for Site M for a 12.7 mm/hr (0.50 in/hr) size storm. The May 7, 2006,
storm produced 13 mm (0.51 in) of rain, which translated into 12 m3 (424 ft3) for Site L
and 39 m3 (1377 ft3) for Site M. This slight disparity between predicted and measured
values can be due to an incorrect calculation of CN for each site or inaccurate watershed
delineation. When back calculating CN for the monitored storms using the NRCS
method, it appears that these watersheds exhibit a different CN for storms less than 25.4
mm (1 in) than for storms over 25.4 mm (1 in). At Site L, for storms less than 25.4 mm
(1 in), the CN for was calculated as 92, but for storms greater than 25.4 mm (1 in) the CN
was calculated as 74, averaging a CN of 83. For Site M, a CN of 89 was backed
calculated for storms less than 25.4 mm (1 in) and a CN=59 for larger storms, with an
average CN=74. As storm sizes increases, CNs more accurately characterizes the
watershed (Schwab 1993). The DIS was designed to capture relatively small storms.
Thus, the average CN is most applicable when designing the DIS using the NRCS
method. Perhaps, the Rational Method should be used when designing the system with
smaller watersheds.
It also should be noted that a majority of Site L’s street curb storm drains were
partially clogged with sand. This could have caused Site L’s runoff to divert to Site M’s
storm drains. Also, various yards in Site L’s watershed exhibited small depressions that
allowed for surface storage. Thus, the measured runoff volume for Site L was somewhat
under predicted and Site M’s runoff was slightly over predicted than predicted. Also,
79
there was continuous flow in Site M’s stormwater outflow pipe, most likely indicating
shallow groundwater intrusion. These are possible explanations for why the calculated
average CN for Site L was less than the design CN and why the calculated average CN
for Site M was greater than the design CN.
Lastly, both the DIS for Site L and Site M were either both over-designed with
regards to the number of DIS chambers or under-designed with respect to the diameter of
the pipe leading to the chambers, or both. This conclusion was predicated upon the fact
that none of the storms filled any of the chambers to the potential storage height of 1.01m
(3.34 ft). The INFINITY water level meter located at the end of Site L’s chambers never
recorded a stage increase. The INFINITY located at the end of Site M’s chamber only
showed small stage difference (less that 13 cm (6 in)) in 3 of the 25 storms. Table 4-7
lists the maximum stage in the beginning chamber for each of the five bypassed storm for
Site M. The maximum stage in beginning chamber was recorded to reach a height of
0.69 m (2.25 ft).
Table 4-7. Maximum stage in bypass storm in Site M’s chambers.
Storm
date
6/14/06
7/23/06
7/25/06
8/22/06
10/8/06
Max Stage in
Chamber
m (ft)
0.69 (2.25)
0.29 (0.96)
0.38 (1.29)
0.24 (0.80)
0.57 (1.87)
Since this was the first implementation of the DIS and dune area was not limiting,
over-design of the system was a conservative decision. However, as research on the
systems progressed, it was evident that a more rigorous design procedure should be used.
Future DIS could be designed using Green-Ampt equation instead of Darcy’s equation
(EQN 4-11) to determine the infiltration capacity of the dune’s soil. In Darcy’s equation
80
the soil is assumed saturated. This is not the case under the DIS where unsaturated
conditions below the system prior to rainfall events. There is at least 2.0 m (6.5 ft) of
sand between the invert of the DIS the average groundwater elevation. The DIS’s soil
has a relatively high hydraulic conductivity. Completely saturating the soil in this system
is highly unlikely.
Green-Ampt is an approximation model utilizing Darcy’s Law. Water is assumed
to infiltrate into the soil as slug flow resulting in a sharply defined wetting front, which
represents the wetted and un-wetted zones. Green-Ampt presents an analysis of the flow
of water in a soil based on the assumption that the soil may be regarded as a bundle of
tiny capillary tubes, irregular in area, direction and shape. The infiltration capacity can
be predicted by the following equation (Skaggs et al. 1969):
⎧1 + B′( P + H ) ⎫
F = A⎨
⎬
f
⎩
⎭
(4-11)
Where: F = accumulative infiltration
f = infiltration capacity
H= head of water on the surface
P = matrix potential at wetting front
A,B’=constants dependent on soil type and conductivity
This equation assumes homogeneous deep soil with uniform initial water content and a
ponded surface, best describing the soil and situation associated with the DIS
(Manivannan and Sundar Raman 2002). If the DIS was designed using the Green-Ampt
equation, both systems would be half the size than the original design.
81
4.4.3 BACTERIA DATA RESULTS AND DISCUSSION
4.4.3.1 Summary Results
All 25 storm events captured during the months of March through October, 2006, all
were analyzed for fecal coliform concentrations and 22 were analyzed for enterococcus
concentrations. Fewer were measured for enterococcus counts due to NCDENR’s
laboratory schedule. Appendix C lists the statistical tests for field bacteria data.
Table 4-8 lists Site L’s and Site M’s fecal bacteria concentrations for each storm.
Table 4-8. Summary of Fecal Coliform levels for the 25 storms.
Site L
Site M
Stormwater
Site L
Stormwater
Runoff
Runoff
Groundwater
CFU/100 ml CFU/100 ml CFU/100 ml
3800*
<1
2280*
3/21/2006
2300*
<1
17200*
4/16/2006
181
<1
19400*
4/26/2006
2700*
1
3000*
5/7/2006
358*
<1
760*
5/14/2006
570*
<1
940*
5/15/2006
2000*
<1
5000*
5/20/2006
2900*
1
5100*
6/5/2006
5800*
4
4700*
6/12/2006
820*
<1
3100*
6/14/2006
TNTC*
1
TNTC*
6/25/2006
19000*
<1
15000*
6/26/2006
4100*
1
3300*
6/27/2006
10000*
1
9000*
7/6/2006
47662*
<1
6800*
7/16/2006
8200*
<1
TNTC*
7/23/2006
TNTC*
<1
TNTC*
7/25/2006
7100*
2
8000*
7/30/2006
TNTC*
2
TNTC*
8/21/2006
TNTC*
54
TNTC*
8/22/2006
TNTC*
4
TNTC*
9/1/2006
TNTC*
<1
TNTC*
9/6/2006
TNTC*
4
TNTC*
9/13/2006
4800*
1
16600*
10/8/2006
28300*
1
6500*
10/17/2006
*Exceeded North Carolina State Standard of 200 CFU/100 ml
TNTC= To Numerous To Count
Site M
Groundwater
CFU/100 ml
3
3
3
<1
8
8
2
2
1
1
<1
<1
<1
4
43
18
86
3
66
214*
TNTC*
4
18
<1
37
82
Table 4-9. Summary of Enterococcus levels for 22 storms.
Site L
Site M
Site L
Stormwater
Stormwater
Runoff
Groundwater
Runoff
CFU/100 ml CFU/100 ml CFU/100 ml
344*
<10
>2005*
4/16/2006
306*
<10
2005*
4/26/2006
334*
10
>2005*
5/7/2006
1652*
64
1445*
5/14/2006
945*
64
>2005*
5/15/2006
870*
<10
334*
5/20/2006
1013*
<10
504*
6/5/2006
>2005*
<10
504*
6/13/2006
2005*
<10
1184*
6/14/2006
>2005*
<10
>2005*
6/25/2006
>2005*
<10
1298*
6/26/2006
1013*
<10
478*
6/27/2006
453*
40
1298*
7/16/2006
2005*
<10
>2005*
7/23/2006
>2005*
10
>2005*
7/25/2006
10
31
<10*
7/30/2006
42
<10
271*
8/21/2006
738*
<10
1184*
8/22/2006
>2005*
31
>2005*
9/6/2006
1013*
42
>2005*
9/14/2006
1091*
10
>2005*
10/8/2006
>2005*
<10
>2005*
10/17/2006
*Exceeded North Carolina State Standard of 104 CFU/100 ml
TNTC= To Numerous To Count
Site M
Groundwater
CFU/100 ml
<10
<10
31
31
31
10
10
64
31
10
20
<10
10
429*
406*
<10
10
137*
2005*
150*
124*
20
It is noteworthy that a North Carolina Tier 1 coastal beach (such as Kure Beach) will
have to post an advisory if fecal coliform levels exceed 200 CFU/100 ml. Inflow fecal
coliform levels ranged from 181 CFU/100 ml to 47,662 CFU/100 ml with a median of
7,100 CFU/100 ml for Site L and ranged from 760 CFU/100 ml to 19,400 CFU/100 ml
with a median of 9,000 CFU/10 0ml for Site M. All stormwater runoff bacteria
concentrations exceeded the state’s standard for swimmable water except for the one
measured at Site L. The ground water bacteria levels ranged from <1 CFU/100 ml to 54
CFU/100 ml and with a median of 1 CFU/100 ml for Site L and ranged <1 CFU/100 ml
to 12,000 CFU/100 ml with a median of 3 CFU/100 ml for Site M. None of the Site L’s
83
groundwater samples exceeded the state’s standards, but two samples from Site M’s
groundwater did.
For statistical purposes, when the upper limit value was reached, (i.e. too
numerous to count (TNTC)), the maximum number measured by the analysis, 6,000, was
multiplied by two. Also when the lower limit was not reached, 1, the lowest test value
allowed was divided by two (Spooner 1991).
For Site L, stormwater runoff enterococcus levels ranged from <10 CFU/100 ml
to 4,010 CFU/100 ml with a median of 1,013 CFU/100 ml. For Site M, enterococcus
concentrations ranged from <10 CFU/100 ml to 4,010 CFU/100 ml with a median of
1,725 CFU/100 ml for Site M. One storm event from Site L and two events from Site M
did not exceed the state’s standard (104 CFU/100ml). The groundwater bacteria levels
ranged from 5 CFU/100 ml to 64 CFU/100 ml with a median of 5 CFU/100 ml at Site L
and ranged 5 CFU/100 ml to 2,005 CFU/100 ml with a median of 26 CFU/100ml, at Site
M. None of Site L’s groundwater samples exceeded the state’s enterococcus standard,
but six samples from Site M’s groundwater exceeded the state’s standard. As done with
the fecal coliform data, for statistical analysis, when the upper limit value was exceeded,
2,005, the maximum number allowed by the test was multiplied by two. Also, when the
lower limit was not reached, 10, the lowest test value allowed was divided by two
(Spooner 1991).
4.4.3.2 Statistical Analysis and Discussion
Statistical analysis indicated that the concentration of fecal coliform flowing into the
system was significantly greater than the concentration of fecal coliform in the
groundwater for both Site L and Site M (p<0.01). The same was also true for
84
enterococcus. Figures 4-36 (a), (b), (c), and (d) are semi-log graphs depicting the amount
of bacteria per storm in the stormwater runoff and groundwater.
At Site M, the two storms that exceeded the fecal coliform standard were August
22, 2006, and September 1, 2006. The August 22, 2006 event also caused groundwater
enterococcus levels (137 CFU/100 ml) to exceed the standards. This storm was a large
event, allowing 87.4 m3 (3,087 ft3) of stormwater to infiltrate into the dunes, with an
average concentration of TNTC (>6000 CFU/100 ml). Groundwater concentration
following Tropical Storm Ernesto had a groundwater concentration of 214 CFU/100 ml
and inflow concentration of TNTC. Enterococcus analysis could not be performed for
Tropical Storm Ernesto sample because the NCDENR Shellfish Sanitation lab was
closed.
It is interesting to note that the runoff from Tropical Strom Ernesto caused the largest
volume of stormwater routed into the dune, 557 m3 (19,670 ft3) with an average
concentration of TNTC (>6,000 CFU/100 ml), caused the largest rise in groundwater
fecal concentration.
The DIS at Site M, with a watershed approximately 2 times larger than Site L,
captured a total runoff of 2,237 m3 (78,999 ft3), 3.5 times that of the total runoff than Site
L, 659 m3 (23,272 ft3). In addition, Site M’s stormwater runoff had a median bacteria
concentration greater than Site L. Therefore, Site M was infiltrating more stormwater
runoff with higher bacteria concentrations than Site L. This may explain the increased
groundwater bacteria concentrations for Site M and not Site L during large summer storm
events.
85
Figure 4-40 b: Site L Semi-Log transform of Enterococcus
Concentration
100
10
1
0.1
21
-M
a
26 r
-A
p
14 r
-M
ay
20
-M
a
12 y
-J
un
25
-J
u
27 n
-J
un
16
-J
ul
25
-J
u
21 l
-A
ug
1Se
13 p
-S
e
17 p
-O
ct
,
Stormwater Runoff
NC's Standard
104 CFU/100ml
100
10
1
0.1
Groundwater
Stormwater Runoff
Groundwater
Date
Date
(a)
(b)
Figure 4-40 d: Site M Semi-Log transform of
Enterococcus Concentration
Figure 4-40c: Site M Semi-Log transform of Fecal
Coliform Concentration
100000
10000
10000
NC's Standard
200 CFU/100ml
1000
100
10
Enterococcus
concentration
(CFU/100ml)
1000
NC's Standard
104 CFU/100ml
100
10
1
Date
6Se
p
8O
ct
Groundwater
0.1
16
-J
ul
25
-J
ul
21
-A
ug
21
-M
a
26 r
-A
14 pr
-M
a
20 y
-M
ay
12
-J
u
25 n
-J
u
27 n
-J
un
16
-J
u
25 l
-J
u
21 l
-A
ug
1Se
13 p
-S
e
17 p
-O
ct
,
Stormwater Runoff
5Ju
n
14
-J
un
26
-J
un
0.1
pr
7M
ay
15
-M
ay
1
16
-A
Fecal concentration
(CFU/100ml)
8O
ct
NC's Standard
200 CFU/100ml
6S
ep
1000
1000
16
-A
pr
7M
a
15 y
-M
ay
5Ju
n
14
-J
un
26
-J
un
16
-J
ul
25
-J
ul
21
-A
ug
10000
10000
(CFU/100ml)
Fecal concentration
(CFU/100ml)
100000
Enterococcus Concentration
Figure 4-40a: Site L Semi-Log transform of Fecal
Coliform Concentration
Stormwater Runoff
Groundwater
Date
(d)
(c)
Figures 4-36. (a) Site L semi-log fecal coliform concentration (b) Site L semi-log enterococcus concentration (c) Site M semi-log
fecal coliform concentration (d) Site M semi-log enterococcus concentration during 2006.
86
Site L did not appear to experience bacteria loading and it should continue to be
under state standards. As previously stated, for both bacteria indicators Site L’s bacteria
concentration never exceeded state standards. Site L stayed below the limit during
summer months, when concentrations are measured to be the highest (Whitlock et al.
2002). The bacteria in the DIS system are expected to die off during North Carolina’s
drier months of October and December (Van Donsel et al. 1967). The concentration of
bacteria in the stormwater runoff entering the system should not be as high in the winter
as the summer due to the reduced temperature as well as the decrease amount of fecal
coliform sources.
Figures 4-37 and 4-38 show the enterococcus concentration and stormwater
State Standard =104 CFU/100 ml
100
80
60
40
20
Volume Infiltrating in DIS
17-Oct,
8-Oct
14-Sep
6-Sep
22-Aug
21-Aug
30-Jul
25-Jul
23-Jul
16-Jul
27-Jun
26-Jun
25-Jun
14-Jun
13-Jun
5-Jun
20-May
15-May
14-May
7-May
26-Apr
0
16-Apr
Volume of Runoff (m3) and
Enterococcus Concentration (CFU/100ml)
runoff volume per storm.
Enterococcus in Groundwater
Figure 4-37. Semi-log of Site L’s groundwater enterococcus concentration and volume
of runoff per storm event.
450
400
350
300
250
200
150
100
50
0
Volume Infiltrating in DIS
17-Oct,
8-Oct
14-Sep
6-Sep
22-Aug
21-Aug
30-Jul
25-Jul
23-Jul
16-Jul
27-Jun
26-Jun
25-Jun
14-Jun
13-Jun
5-Jun
20-May
15-May
14-May
7-May
26-Apr
State Standard =104 CFU/100 ml
16-Apr
3
Volume of Runoff (m ) and
Enterococcus Concentration (CFU/100ml)
87
Enterococcus in Groundwater
Figure 4-38. Semi-log of Site M’s groundwater enterococcus concentration and volume
of runoff per storm event.
The six storms that exceeded the enterococcus standard at Site M did not occur
until five months after the systems had been implemented. The first storm was on July
23, 2006, and was quickly followed by a July 25, 2006, storm. However, it is difficult to
determine if this is actually bacterial overloading from previous storms or a result of
several large storms occurring close together in the warmest months of the study.
After those two storms, Site M’s groundwater surpassed enterococcus state
standards on August 22, September 6, September 14, and October 8. All of these storms
infiltrated at least 52.3 m3 (1,847 ft3). For Site L, only the volume for the October 8 event
(87.4 m3 (3,087ft3)) exceeded the runoff volume of 52.3 m3 (1,847 ft3), which was
substantially less than 363 m3 (12,819ft3) of runoff at Site M, for the same event.
Despite Site M’s last storm’s groundwater fecal coliform concentration exceeding
the state standards, Site M’s and Site L’s groundwater fecal coliform bacteria
concentrations, after the implantation of the DIS, were significantly similar (p <0.05) to
the groundwater fecal coliform bacteria concentrations before the DIS. Figures 4-39 and
88
4-40 are SAS generated graphs that show log probability plot for groundwater bacteria
concentration before and after DIS installation at for Site L and Site M.
It should be noted that there were a limited number of groundwater bacteria
samples collected before the DIS was installed. It was difficult to determine if there is a
seasonal variation in the groundwater bacteria data. Whitlock et al. (2002) and Van
Donsel et al. (1967) have reported seasonal variations in survival of indicator bacteria.
Groundwater
Log (fecal
coliform)
Number of Day Since Beginning of Study
Figure 4-39. SAS output for Site L of fecal coliform groundwater concentration before
DIS (square symbol) and after (plus symbol).
Groundwater
Log (fecal
coliform)
Number of Day Since Beginning of Study
Figure 4-40. SAS output for Site M of fecal coliform groundwater concentration before
DIS (square symbol) and after (plus symbol).
89
Another consideration is the constituents found in stormwater runoff. Anderson
and Rounds (2003) reported E. coli concentrations, at a mixture of urban and agricultural
sites, to be statistically correlated with concentrations of suspended sediment, TP, and
NO3-N. Anderson and Rounds found that E. coli concentrations were not statistically
significantly correlated to temperature, but found the largest E. coli concentration amount
occurring during the warmest water temperature. Since inflow nutrient and sediment
levels are not known in this study, a comparison cannot be made, but it is important to
keep these correlations in mind when analyzing the data.
Even without analyzing stormwater constituents, the data indicated increased
bacteria loading in Site M. The fact that Site L infiltrated less stormwater runoff and
never exceeded the enterococcus state standard, and Site M only started to surpass the
standards near the conclusion of measurement, indicates the potential of increased
bacteria loading in Site M’s system. Bacteria colonies may have been stabilizing and
growing using organic matter deposited from the sediment in stormwater runoff. Gerba
and McLeod (1975) reported a longer survival of E. coli colonies in marine waters when
a greater content of organic matter was present.
4.5 DIS SUMMARY
The Dune Infiltration System captured all runoff associated with the designed rainfall
intensity of 12.5 mm/hr (0.5 in/hr) or less. Thus, the DIS achieved the goal of decreasing
the potential health hazard associated with stormwater ocean outfalls. The DIS
implemented at Site L never overflowed, capturing storms with rainfall intensities up to
90 mm/hr (3.5 in/hr). The DIS at Site M captured all storms with intensities up to 28
90
mm/hr (1.1 in/hr) and only overflowed 5 times. Both DIS systems captured a measured
total of 2,896 m3 (102,271 ft3) and bypassed 99 m3 (3,500 ft3), routing 96.6% of
measured inflowing stormwater runoff into the dunes.
One objective was to determine if routing and discharging stormwater runoff in
the dunes elevated the level of the groundwater beneath the dunes. Routing the 659 m3
(23,272 ft3) into Site L’s dune and 2237 m3 (78,999 ft3) into Site M’s dunes did not
substantially change the elevation of the water table. The largest storm-induced
fluctuation, 1.5 m (4.9 ft) occurred at Site M during Tropical Storm Ernesto. Preimplementation groundwater data shows a substantially greater groundwater elevation
increase during Hurricane Ophelia in 2005 before the DIS was installed. Maximum tidal
fluctuations caused groundwater to elevate 2.5 m (8.1 ft). Thus, there appears to be
limited groundwater mounding beneath the DIS system at Site L and Site M.
Another objective was to determine if routing and discharging stormwater runoff
into the dunes increased the bacteria level in the groundwater beneath the dunes. This
was tested by identifying a range of fecal coliform and enterococcus concentrations
stormwater runoff. Inflowing stormwater runoff had concentrations of fecal coliform
concentrations ranging from 181 CFU/100 ml to 19,400 CFU/100 ml with a median of
8,600 CFU/100ml. Inflow enterococcus concentrations ranged from <10 CFU/100 ml to
>2,005 CFU/100 ml with a median of 1,298 CFU/100ml. The groundwater
concentrations were significantly less (p< 0.001) than the inflow with fecal coliform
concentrations ranging from <1 CFU/100 ml to 214 CFU/100 ml with a median of 1.5
CFU/100ml. Groundwater enterococcus concentrations the range was from <10
CFU/100 ml to 2005 CFU/100 ml with a median of 10 CFU/100ml. The groundwater
91
fecal coliform concentrations at both sites were significantly (p<0.01) less than the
stormwater runoff inflow concentration.
The purpose of measuring the inflow and groundwater bacteria concentration was
to determine bacteria removal efficiency of DIS. North Carolina’s indicator bacteria
standards were exceeded only in Site M’s groundwater. Groundwater samples surpassed
the limit on 2 of the 25 events for fecal coliform and 6 of the 22 for enterococcus. These
incidents occurred five months after the system was implemented and during large storm
events. In addition, these samples were collected approximately 50 m (150 ft) from the
surf zone.
Thus far, both Dune Infiltration Systems have proven to decrease the likelihood
of beach closures, near where the old outfalls discharged, obtaining the overall goal of
decreasing the potential health dangers associated with stormwater ocean outfalls for
local coastal residences, tourists, and coastal wildlife.
In between storms, bacteria colonies could be growing, maintaining, or dying
depending upon the amount of useable organic matter available. If vast amounts of
stormwater runoff enter the system at high peak inflow rates, the bacteria colonies could
be transported by the infiltrating stormwater into the groundwater. The infiltration rate of
the soil was measured to be approximately 0.0009 m/s (0.003 ft/s), allowing bacteria to
flow through the slightly charged to neutral sand particles. Bolster et al. (2001) showed
when a large number of bacteria are introduced into subsurface, sediment surfaces near
the insertion point may have become saturated with bacteria. This blocking phenomenon
limits the concentration of deposited bacteria available for metabolic transformation of
92
contaminants and in turn limits biodegration rates. The presence of previously deposited
particles also affected deposition rates, allowing bacteria filter through the system.
93
5.0 Sand Column Infiltration and Bacteria Laboratory
Study
5.1 EXPERIMENTAL DESCRIPTION
As mentioned in Chapter 2, sand filters must be maintained for them to function properly.
In practice, insufficient maintenance and subsequent clogging are ubiquitous. When
designing sand filters, Grisham (1995) suggested that 50 percent of the measured
infiltration rate should be used. This design parameter is based on experiments that were
conducted to analyze the effect of runoff sediment on sand filters.
In wastewater research, the filter’s infiltration rate is a control, while the
efficiency of bacteria removal (E. coli and fecal coliforms) is variable (Urbonas 1999),
(Gomez 2006). The objective of this laboratory experiment was to combine stormwater
and wastewater research areas to better understand the effect of sediment clogging on
bacteria removal in DIS.
Nine sand columns were constructed with soil from the DIS at Site L. The
columns were divided into three experimental units (EU), and each EU was subject to
one of three discrete treatments: control with DI water (CDI), autoclaved stormwater
runoff (CSW), or bacteria spiked stormwater runoff (T). The experiment lasted 60 days
with treatments applied every third day. For each trial, infiltration rates of all columns
were measured, as well as total coliform (TC) counts for the CSW and T columns’
effluent using American Public Health Association Standard Method 9221. Every fourth
run, CSW and T columns’ effluent was analyzed for TC and E. coli at the NC State BAE
94
Environmental Analysis Lab using the IDEXX Laboratory, Inc. developed method of
Colilert™ (Standardized SM 9223).
5.1.1 HYPOTHESES
The first goal of the experiment was to statistically correlate T columns’ infiltration rates
with the T columns’ bacteria removal efficiency. The second was to analyze if the
addition of E. coli in the T columns statistically decreased the columns’ infiltration rate.
This was done by testing the following hypotheses (α=0.05):
1) The CDI sand columns’ day 60 infiltration rates are significantly
similar to day 1 infiltration rates.
2) The infiltration rate of the CSW and T columns are significantly less
than CDI at the end of the 60 days.
3) The infiltration rates of the T columns are significantly less than the
CSW columns’ infiltration rates at the end of the 60 days
4) The decrease in TC and E. coli concentration within the T columns’
effluent are correlated with infiltration rate.
5) The concentration of TC and E. coli within CSW sand columns’
effluent is significantly less than that of the T sand columns’.
5.1.2 EXPERIMENTAL VARIABLE CONTROL
This laboratory study was designed to emulate Site L’s DIS system, while controlling
certain key variables to generate statistically significant results. It was designed so that
one sand column represented one StormChamber unit in Site L’s DIS. Of particular
concern were the frequency of the trials and the amount of stormwater applied per trial.
The three day trial frequency was calculated from the preceding decade’s historical
rainfall data from New Hanover County Airport ,Wilmington, North Carolina (State
Climate Office of North Carolina 2006). The corresponding laboratory-scale amount of
95
influent needed for each column’s trial was calculated to be 2 liters per column. This
volume was calculated using the amount of stormwater runoff expected from Site L
during a 12.5 mm (0.5 in) event, as explained previously in Chapter 4.3.1. This number
was scaled down by applying the ratio of the volume of the cylinder to the volume of a
chamber in the DIS.
Soil was collected from Kure Beach Site L during installation implementation of
the DIS system in February 2006. A soil profile was obtained from a side of the trench
shown in Figure 4-12.
Columns were sized using a column to particle diameter ratio of 50, where the
particle diameter used was the d10, the diameter at which 90% of the soil’s particles have
a greater diameter than the d10. The d10, was determined using the ASTM D 422
“Standard Test Method for Particle-Size Analysis” (2002). The particle distribution
curve, shown in Figure 5-1, indicated that the d10 was 0.85 mm (.003 in). Thus, the inner
diameter of the column must exceed 4.25 cm (1.67 in). The columns used in this lab
were clear 0.38 cm (0.13 in) thick plastic tubing 1.8 m (6 ft) long with an outer diameter
of 5.08 cm (2 in) and inner diameter of 4.45 cm (1.75 in). The columns were thoroughly
cleaned and rinsed with DI water before use.
100%
Percent Finer
80%
60%
40%
20%
0%
0
0.1
0.2
0.3
0.4
0.5
Size (mm)
Figure 5-1. Kure Beach’s soil particle size distribution.
0.6
0.7
0.8
0.9
96
Before filling the columns with Site L’s soil, the soil was autoclaved for 30
minutes at 121◦C (250◦ F) and 1.4 atmosphere (20 psi). The sand was then oven dried at
105◦C (221◦ F) for 24 hours. The soil was sieved through on 4.75 mm (0.19 in) opening,
sieve number 4, to remove clay aggregates from the soil. The clay presence in the dune’s
otherwise sandy soil stemmed from a previously conducted beach nourishment project in
1997, during which soil from the ocean was pumped onto the beach.
From the NRCS, Soil Data Mart (2006), the soil density for Newhan Fine sand
ranges from 1.60 g/cm3 (0.057 lbs/in3) to 1.75 g/cm3 (0.063 lbs/in3) thus, 4.2 kg (9.2 lbs)
to 4.5 kg (10 lbs) of sand was needed to fill the columns. 4.3 ± 0.1 kg (9.6 ± 0.2 lbs) of
sand was weighed out and filled into each of the nine columns. A 13 cm (5 in) by 13 cm
(5 in) drainage sock square was wrapped over the end and taped to the bottom of each
column to prevent sand loss. Tape was placed 30.5 cm (12 in), 91.4 cm (36 in), and 152
cm (60 in) on the columns’ exterior, as measured from the top of the column for later
measurement purposed. Two liters (0.5 gal) of DI water were poured into each column,
to dispel air pockets. Next, 0.1 kg (0.3 lb) to 0.5 kg (1.1 lbs) of sand was added to the
columns so that 1.7 m (5.5-ft) of the column was filled with sand. Two more liters (0.5
gal) of DI water were poured into the sand filled column to compact the sand (Figure 52).
97
Figure 5-2. Initial column set-up, allowing 2 L of DI water to compact the column
A rock layer was added to the top of the columns to mimic the rock layer on the
full-scale DIS. Landscape pea pebbles were washed and sieved through a sieve opening
of 9.5 mm (0.38 in). The pebbles that sieved through were weighed to be 0.18 kg (0.5
lbs) ± 0.05 kg (0.1 lb) and funneled on top of the sand in each column (Figure 5-3). Two
liters (0.5 gal) of DI water were poured over the gravel and through the columns.
Additional pebbles, less than 0.1kg (0.2 lbs), were added in five of the columns. Upon
completion, there were nine, 1.8 m (6.0ft) columns filled with 1.7 m (5.5ft) of sterile
Newhan Fine sand, topped by 0.2 m (0.5 ft) of washed pebbles (Figure 5-3).
Figure 5-3. Finishing construction the columns by adding stone to the columns.
98
5.1.3 EXPERIMENTAL MODEL
This experiment was designed with three treatments; each treatment was comprised of
three replicates. Each column was randomly assigned a treatment. Standard PVC
reducer couplings (7.6 to 5.1 cm diameter) made of pliable rubber created a waterproof
linkage between the 5.1 cm (2.0 in) outer diameter sand columns and the 7.6 cm (3 in)
PVC reservoirs. The PVC reservoirs accommodated 2 liters (0.5 gal) of stormwater and
comprised the final, upper most part of the apparatus. Directly below each column a
correspondingly labeled 1 liter (0.3 gal) Nalgene™ bottle as place to capture the
respective column’s effluent (Figure 5-4).
Figure 5-4. Final sand column design.
Additional laboratory apparatus utilized in the experiment were as follows.
Three, 7.6 L (2.0 gal) buckets with bottom mounted spigots were labeled either: DI water,
99
E. coli stormwater or stormwater. These buckets were used to hold the different
treatments during a trial. Three, 2 L (0.5 gal) graduated cylinders with the same labeling
convention were used to pour the different treatments into each column. Lastly, a
calibrated Fischer Science™ brand timer was used to time all infiltration measurements.
5.2 EXPERIMENTAL METHOD
5.2.1 VARIABLE CONTROL
To ensure consistency between each column in terms of particle size distribution, density,
and composition and to establish a baseline infiltration rate per column, calibration tests
were performed. Two L (0.5 gal) of DI water were poured into each of the nine columns
and their infiltration rates measured according to the tape marks corresponding to
infiltrate volumes. Effective porosity of the soil was measured to be 33%, thus the tape
marks represented volumes of 0.18 L (0.05 gal), 0.53 L (0.14 gal), 0.88 L (0.23 gal), and
2 L (0.5 gal). Three days later a similar variable control trial was preformed.
5.2.2 EXPERIMENT PREPARATION
Due to the large amount of stormwater needed for the laboratory experiment, 240 L (63.4
gal) of actual stormwater runoff was collected from a watershed in Raleigh, NC, that was
predominately impervious, similar to Site L’s watershed. Stormwater runoff was
collected within the first hour of rainfall and later disposed if not used within three
weeks.
A pure E. coli strain culture was grown before the beginning of the experiment.
In a 200 ml (6.8 oz) Erlenmeyer flask, 100 ml (3.4 oz) of Lauryl Tryptose Broth (LSB)
was prepared occurring to the formula reported by American Public Health Association
100
(APHA) Standard Method 9221. This flask was then autoclaved for 15 minutes at 121◦C
(250◦ F) and 1.4 atmosphere (20 psi). Once the flask cooled, the sterile LSB broth was
relocated under a biological fume hood for inoculation. A standard wire loop was
inoculated with One Shot® E. coli strain INVαF´ purchased from Invitrogen Corporation.
The wire loop was then dipped into the sterile LSB broth inoculating the broth. The top
of the Erlenmeyer flask was then placed over a Bunsen burner to reduce the risk of
ambient microbial contamination. The E. coli-inoculated LSB was stored in a shake table
incubator at 37 ± 2°C.
The cells were used for a trial run after 24 hours of incubation. This liquid culture
was used for trials until 11 days after incubation, based on the work of Hicks et al.
(2005). On the 11th day, another 100 ml (3.4 oz) of LSB was prepared in either a 200 ml
(6.8 oz) Erlenmeyer flask or a 300 ml (10 oz) Erlenmeyer flask with a right-side arm.
The LSB flask was then autoclaved for 15 minutes 121◦C (250◦ F) and 1.4 atmosphere
(20 psi) and then cooled. The sterile LSB broth was placed under a biological fume hood
and inoculated with 1 ml (0.03 oz) of the previous, 11 day old, LSB and E. coli mixture.
Then, the E. coli inoculated LSB was stored in a shake table incubator at 37 ± 2°C. If the
LSB and E. coli mixture were prepared in the 300 ml Erlenmeyer flask with a right-side
arm, absorbance readings were taken using a Milton Roy Spectronic 21™
spectrophotometer at 650 nm every 24 hours for 14 days. The absorbance measurements
were recorded and plotted to ensure the cell density was between experiments.
5.2.3 EXPERIMENT PROCEDURE
The day before a trial, the stormwater runoff was sterilized. 12.2 L (3.2 gal) of
stormwater were autoclaved in 2 L (0.5 gal) polypropylene plastic (PP) bottles for at least
101
15 minutes at 121◦C (250◦ F) and 1.4 atmosphere (20 psi). The bottles were closed and
cooled overnight.
Approximately 2 hours prior to the trial, Most Probable Number (MPN) tubes
were made. Thirty 40 ml (1.4 oz) and eighty 15 ml (0.5 oz) PP tubes were filled with 20
ml (0.7 oz) and 10 ml (0.3 oz) of LSB broth, respectively. Bisulfate was added to later
detect presence of the coliform. The tubes were autoclaved for at least 15 minutes at
121◦C (250◦ F) and 1.4 atmosphere (20 psi) and allowed to cool inside the biological
fume hood.
To begin a trial, the three 7.6 L (2 gal) buckets were filled, one with DI water, one
with 6 liters of the sterile stormwater, and the last with the sterile 5.5 L (1.5 gal) of
stormwater. The remaining 700 ml (23.7 oz) of sterile stormwater were poured into a 1 L
(0.3 gal) beaker and placed under the biological fume hood. Pure E. coli culture was
removed from the incubator and used to inoculate with the 700 ml (23.7 oz) of
stormwater with 6.2 ml (0.2 oz) of the pure culture Figure 5-5 (a) and (b).
(a)
(b)
Figure 5-5. (a) Extracting 6 ml of E. coli culture to inculcate sterilized stormwater in
the 1.5 L beaker. (b) Sterilizing E. coli culture flask.
102
The bacteria-spiked stormwater were then poured into the 7.6 L (2 gal) bucket with 5.5 L
(1.5 gal) of sterile stormwater and stirred for 2 minutes Figure 5-6 (a).
It was found that 1 ml (0.03 oz) of pure culture to 1 L ( 0.3 gal) of DI water
yielded the desired concentration range of 2200-2800 MPN index/100 ml, which
approximated the average concentration of fecal coliform in stormwater runoff measured
at Site L from March to July 2006.
The 2 L (0.5 gal) graduated cylinder labeled “E. coli stormwater” was filled with
two liters of the spiked stormwater. The solution was then poured at an approximate rate
of 8 L/min (0.3 ft3/min) into one of the three T labeled columns, T1, T2, and T3, and the
timer was started. This procedure was offset by one minute per column for the two other
T columns (Figure 5-6 (b) and (c)).
a)
b)
c)
Figure 5-6. (a) Inoculating sterilized stormwater with bacteria stormwater. (b)
Measuring 2 L of bacteria stormwater in 2 L cylinder. (c) Pouring treatment in the sand
column.
The order of addition of the spiked stormwater on the column was randomly assigned
using a random number generator, RAND, in Microsoft Excel®. The remaining 200 ml
103
(6.8 oz) of spiked stormwater was poured into a 300 ml (10 oz) beaker and placed in the
biological hood for later MPN testing.
The time required for the treatment to infiltrate to 0.18 L (0.05 gal), 0.53 L (0.14
gal), 0.88 L (0.23 gal), and 2 L (0.5 gal) was recorded, Figure 5-7. Once the wetting
front on all the treatment’s columns passed the 0.88 L (0.23 gal) tape, trials for the
bacteria-free columns, CSW1, CSW2, and CSW3 began. The 2 L (0.5 gal) graduated
cylinder labeled “stormwater” was filled with 2 L (0.5 gal) of the stormwater and poured
into one of the three CSW columns. This procedure was followed in one minute
increments for the two other CSW columns, with the order of pouring randomly assigned.
Figure 5-7. Timing the water front movement to various table levels.
Infiltration time was recorded for each of the three columns for the stormwater to
infiltrate volumes of 0.18 L (0.05 gal), 0.53 L (0.14 gal), 0.88 L (0.23 gal), and 2 L (0.5
gal). Once all three CSW columns’ wetting fronts had passed the 0.88 L (0.23 gal) tape,
this basic procedure was initiated for the remaining control DI columns, CDI1, CDI2, and
CDI3.
104
After 1 L (0.3 gal) of effluent had drained from T1, T2, T3, CSW1, CSW2, and
CSW3, the Nalgene™ bottles were sealed and taken to the biological fume hood for
MPN testing. A set of 15 MPN tubes, 3 dilutions with 5 tubes per dilution were prepared
according to the APHA Standard Method 9221. Seven sets in total were produced, one
set for each effluent and one for the E. coli spiked stormwater influent. The MPN tubes
were incubated at 37 ± 2°C for 48 hours on a table shaker. The MPN tubes were checked
at 24 and 48 hours to see if there was growth, indicated by gas production or a color
change in the tubes. The number of positive tubes in each dilution was recorded for each
set and compared to the MPN tables.
After each trial run, the gallon buckets, 2 L (0.5 gal) graduated cylinders,
Nalgene™ effluent bottles, were soaked in soapy water for 15 minutes, sprayed with
isopropyl alcohol, and thoroughly rinsed with DI water. After allowing 48 hours of
incubation, MPN tubes were autoclaved at least 15 minutes at 121◦C (250◦ F) and 1.4
atmosphere (20 psi) and then soaked in soapy water for 15 minutes and rinsed thoroughly
with DI water.
5.3 EXPERIMENTAL RESULT S AND DISCUSSION
5.3.1 E. COLI CULTURE
To ensure the E. coli cultures grown throughout the experiment’s duration remained
consistent, turbidity measurements were taken at the beginning of the experiment, July
17, 2006, and near the end, September 7, 2006. Figure 5-8 shows the optical density at
650 nm (OD650) and Klett units over the twelve day trial.
1.6
1.4
800
700
1.2
1
600
500
0.8
0.6
0.4
400
300
200
0.2
0
100
0
350
0
50
100
150
200
250
September
July
300
Klett Units
OD 650nm
105
Time (hr)
Figure 5-8. OD650 of Grown E. coli cultures over 13 day period.
This graph shows a similarity of OD650 measurements for beginning and ending
cultures. This similarity was further analyzed by calculating the generation time (tgen)
and growth rate (k). The generation time was calculated using equations 5-1, 5-2 and 5-3
and is shown in Figure 5-9 (Madigan et al. 2003).
n = 3.32 [ log10 N − log10 N o ]
k=
n
t
t gen =
(5-1)
(5-2)
1
k
(5-3)
Where: n = number of divisions since time of inoculation
N = OD reading during exponential growing phase (Figure 5-7)
N0 = beginning OD reading during exponential growing phase
t = change in time between N and N0 (min)
k = growth rate (1/min)
tgen = time for a cell to travel through complete cell cycle (min)
The initial E. coli culture had a k value equal to 0.035 min-1 and a tgen equaling 29
min. The ending E. coli culture had similar values with k equaling 0.037 min-1 and tgen
equaling 27 min. Other laboratory research shows generation times between 26 through
106
30 minutes with various strains of E. coli in glucose broth with pH equal to 7 (Hicks et al.
1.6
1.4
1.2
1
0.8 N
0.6
No
0.4
0.2
0
0
800
700
600
500
400
300
200
100
0
t
5
10
15
Klett Units
OD 650nm
2005) (Plank and Harvey 1979).
20
Time (hr)
September
July
Figure 5-9. Calculating generation time from OD curve.
Measuring comparable absorbance curves, generation times, and growth rates for
the initial and ending E. coli cultures, indicate that the cultures’ growth at different times
were similar. To ensure the concentration of bacteria were as similar to each other as
possible, MPN tests were performed after each trial on unused inoculated stormwater.
The results are shown in Table 5-1, averaging 2594 MPN index/100 ml with a standard
deviation of 604 MPN index/100 ml
107
Table 5-1. Influent E. coli concentrations for each trial.
Trial
Day
6
9
12
15
18
21
24
27
30
36
39
42
45
48
51
54
Average
E. coli Concentration
(MPN index/100 ml)
2800
3500
2800
2200
3500
2200
2800
2800
3500
1700
3000
2800
2800
2200
1700
2200
2656
5.3.2 INFILTRATION RESULTS AND DISCUSSION
5.3.2.1 Pre-Trial Variable Control Results and Discussion
The infiltration rates for each column’s wetting front advancement curve and
experimental trial are found in Appendix D. Figure 5-10 shows the results of the
average wetting front advancement curves for each treatment. The average infiltration
rate, 155 cm/hr (61 in/hr) of the Control DI Water (CDI) columns was slightly lower than
the infiltration rate for the bacteria-inoculated stormwater columns, Test (T), and the
sterile stormwater columns, Control Stormwater (CSW). The average infiltration rate of
the T columns, 208 cm/hr (82 in/hr) cm/hr was similar to the average infiltration rate of
the CSW columns, 211cm/hr. (83 in/hr).
Wetting Front Advancement
Rate (cm/hr)
108
800
700
600
500
400
300
200
100
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Time (hr)
CDI
T
CSW
Figure 5-10. Average wetting front advancement rate curve for each treatment.
The variation between the average CDI infiltration rate and the average T and
CSW infiltration rate is acceptable, since the CDI columns were designed as a control for
the infiltration rate variable. It was more imperative that the average infiltration rate of
the T columns and CSW columns were equal in order to better evaluate the effect of
bacteria on infiltration rate. The average infiltration rates of the CDI columns were lower
than the T and CSW. During the trials, T and CSW columns gradually become clogged
with sediment from the stormwater, thus dramatically decreasing their infiltration rates.
Thus, the slight variation between the average infiltration rate for CDI relative to average
infiltration rates of T and CSW treatments, was not significant.
5.3.2.2 Trial Infiltration Rates Results and Discussion
The infiltration times for each column and trial are in listed in Appendix D, with
laboratory statistical output located in Appendix F. Table 5-2 lists the average infiltration
rate per treatment for the 20 trials, with trial day zero representing the calibration curve.
0.9
109
Table 5-2. Average treatment infiltration rate per column.
Trial
Number
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Trial
Day
0
3
6
9
12
15
18
21
24
27
30
33
36
39
42
45
48
51
54
57
60
CDI
(cm/hr)
154.74
157.84
164.02
168.73
163.92
164.55
163.71
165.78
165.00
164.68
164.98
173.81
174.14
173.20
170.90
163.39
163.93
165.26
165.42
165.29
163.93
T
(cm/hr)
207.79
177.43
140.87
136.69
78.28
75.06
77.82
67.76
62.98
58.31
57.50
58.99
48.89
41.43
17.56
15.36
6.75
2.52
1.36
0.64
0.49
CSW
(cm/hr)
211.57
237.46
230.28
231.78
181.70
158.11
157.06
110.70
112.40
110.00
109.84
75.79
66.35
59.08
28.19
27.87
22.55
8.77
3.65
2.40
2.41
This table is illustrated in Figure 5-11, which graphs the mean and standard errors for
infiltration rate per treatment of the 20 trials. The diagonally striped bars depicted the
average CDI infiltration rate per trial. The CDI trial averaged 166 cm/hr (65.4 in/hr) with
a standard deviation of 4.70cm/hr (1.85 in/hr). As seen in the table and the graph, at the
beginning of the experiment, average CDI infiltration rates were slightly higher than near
the end of the experiment. This may be attributed to the removal of finer sized particles
during previous the initial trial runs.
Infiltration Rate (cm/hr)
110
300
250
200
150
100
50
0
0
3
6
9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Trial Day
CD1
T
CSW
Figure 5-11. Graph of average treatment infiltration rate.
Columns T and CSW average infiltration rates decreased significantly (p<0.01)
throughout the course of the experiment. This was primarily due to the volume of
sediment in the stormwater runoff. The sediment accumulated on top of the sand, forcing
the water to pass through the smaller pore space in the sediment layer, before infiltrating
through the column. With each successive trial, the sediment layer thickness increased,
resulting in a corresponding decrease in infiltration rate. By trial day 30, the water
infiltrating through the columns carried the finer sized sediment particles down a few
centimeters, visibly creating a second semi-confinement layer.
The sediment clogged layers caused the infiltration rate to decrease to that of a
silty soil, 0.5 cm/hr (0.2 in/hr) to7.6 cm/hr (3 in/hr) and clay, less than 0.5 cm/hr (0.2
in/hr) (Boul et al. 2003). T columns infiltrated at the rate of a silt soil by the 16th
treatment and infiltrated at the rate of a clay soil during the 20th treatment. The CSW
columns infiltrated at the rate of a silt soil during the 18th treatment, as shown in Figure
5-12.
111
Infiltration Rate (cm/hr)
300
CSW1
CSW2
CSW3
T1
T2
T3
250
200
150
100
50
Silt Infiltration rate ( 2 cm/hr -7.6 cm/hr)
0
0
6
12
18
24
30
36
42
48
54
60
Trial Day
Figure 5-12. Graph of variation of infiltration rate for treatment T and CSW.
The infiltration rate of the T treatments were significantly different (p<0.05) from
the infiltration rate of the CDI columns from trial number 4 until 20. The infiltration rate
of the CSW treatments were significantly different (p<0.05) from the infiltration rate of
the CDI columns from trial number 10 until 20. The infiltration rates of T columns were
significantly different (p<0.1) from infiltration rate of CSW columns during trials
numbered 4 through 7 and then from trial number 16 through 20 (p<0.05). The
significant difference between the infiltration rates of T columns’ versus CSW columns’
infiltration rates from trials number 4 through 7 could be due to CSW 2’s large
infiltration rate values. A brief examination of individual column’s infiltration rates
follows.
As shown in Figure 5-12, the squares tend to increase between trial days 18 until
30. Significant differences between T columns’ infiltration rate and CSW columns’
infiltration rate from trials number 16 through 20 could be caused by bacteria blocking,
112
(Bolster et al. 2001), or by large aggregates of bacterial cells forming local plugs within
the pores, which was found to reduce the saturated conductivity up to four magnitudes
(Vandevivere and Baveye 1992).
5.3.3 BACTERIAL RESULTS AND DISCUSSION
Table 5-3 compares the measured E. coli concentrations of all three T columns and 2
randomly selected CSW columns for four trials. The means (standard deviations) for the
T columns for trials numbered 4, 9, 14, and 18 reported in CFU/100 ml were 2144 (690),
46.33 (15.50), 1.33 (0.57), and 0.67 (0.28), respectively. These values were substantially
larger than the mean (standard deviation) for the CSW columns for trial number 4, 9, 14,
and 18 , which where measured in CFU/100 ml as 1(0), 1 (0), 0.75 (0.35), and 0.5 (0),
respectively.
Table 5-3. E. coli concentration measured for T and CSW treatment for four trials using
Colilert™ (Standardized SM 9223) testing method.
Trial
Number
4
24
39
51
Trial
Day
12
27
42
51
T1
T2
2282
35
1
<1
2755
64
2
1
T3
CFU/100ml
1395
40
1
<1
CSW1
CSW2
CSW3
1
N/A
N/A
<1
1
1
<1
N/A
N/A
1
1
<1
It is interesting to note that the largest T column concentration for trial number 4,
2,282 CFU/100 ml, was slightly larger than the measured influent concentration of 2,200
MPN index/100 ml. This could be due to the standard error of measuring E. coli influent
concentration using MPN method. There is a 95% confidence interval of 1000 to 5,800
MPN index/100 ml that goes along with this measurement (APHA 1999).
Another explanation for the larger bacteria concentration in the effluent is bacteria
growth within the column combined with detachment of the bacteria from the soil. The
113
affect of the previous treatment and the quick infiltration rate of the soil can cause growth
and detachment. Since this high measurement occurs on the fourth trial, there were three
previous trials where bacteria had entered the system. Assuming some of the bacteria
from these trials bound to the soil or the deposited sediment, there were 12 days for the E.
coli cells to adapt to the new environment. Mandelstam (1958) determined that nongrowing populations of E. coli were able to synthesize new enzymes at a rate
approximately equal to protein breakdown, indicating that protein synthesis occurs at the
expense of utilization of endogenous material. In fact, laboratory cultures usually upset
community balance by causing quick enrichment of certain portions of the populations
(Roszak and Colwell 1987). In addition, every third day there was more stormwater
with bacteria added to the column allowing the bacteria population to increase with new
cells and new sediment. Van Donsel et al. (1967) reported that E. coli colonies in the
natural environment need 3-10 dry days to decrease the bacteria concentrations by 90%.
When the treatment was applied during these first four trials, the infiltration rate
was measured to be 396 cm/hr (156 cm/hr) in the top 30 cm (12 inches), where the
majority bacteria of the colony would be located, according to Alm et al. (2003). This
rate is fast enough that as the stormwater infiltrates through the sediment and top sand
particles, stormwater could remove bacteria cells from the established E. coli colony
attached to the sediment and sand particles. These bacteria ended up in the effluent,
which is why the concentration in the effluent was high.
To better understand the overall trend of bacterial removal in the system, total
coliform (TC) was analyzed. TC MPN concentrations were measured per trial for the
three T and CSW columns and can be found in Appendix E. Table 5-4 lists the average
114
TC MPN measurement per treatment for the 18 trials. This table is illustrated in Figure 513, which is a log graph of the mean total coliform concentration and standard error per
treatment throughout the 18 trials. Bacteria concentrations were not analyzed after the
18th trial due to the slow infiltration rates.
Table 5-4. Average total coliform concentration in T and CSW treatment per trial.
Trial
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Trial
Day
3
6
9
12
15
18
21
24
27
30
33
36
39
42
45
48
51
54
T
(MPN index/100 ml)
4
61
253
13670
13842
11833
2067
767
674
700
237
283
297
270
167
10
23
1
CSW
(MPN index/100 ml)
5
11
203
1000
1800
1867
410
260
337
153
107
29
15
11
11
9
14
1
The diagonal line columns in the graph represent the influent concentration of E.
coli per run. The concentration of TC in the T column’s effluent was significantly larger
(p<0.01), than the CSW column’s effluent, due to the addition of E. coli in T treatment
influent. There is a dramatic increase in TC from T column’s effluent as well as CSW
column’s effluent from the trial day 12 through trial day 21. There was also a
significant (p<0.05), decrease from T column’s effluent at trial day 24 and 48.
Log (T.C Concentration)
(CFU/100ml)
115
100000
10000
1000
100
10
1
0
3
6
9
12 15 18 21 24 27 30 33 36 39 42 45 48 51 54
Trial Day
T
CSW
Influent E.Coli
Figure 5-13. Semi-log plot of total coliform concentration for each trial.
There was a direct correlation between infiltration time and T column’s effluent
TC concentration at an α=0.05 level. Figure 5-14 is a graph of the average T and CSW
column’s effluent TC concentrations as well as the average T and CSW column’s
infiltration rates for each run. At trial days 21, 33, and 48 there are corresponding drops
in column’s effluent TC concentration as well as the columns’ infiltration rates.
T columns’ average effluent TC concentration pattern, represented by circle
points, follows a common drinking water treatment filtration phenomenon. The average
effluent TC concentrations in the beginning trials were low, then around the 5th trial (trial
day 15) the average effluent TC concentration increased seven fold. By the 7th trial (trial
day 21) the average effluent TC concentration was low again. The periods of increased
average effluent TC concentration is known as the ripening period. The ripening period
is associated with drinking water treatment granular filtration. As water passes through a
filter consisting of a packed bed of granular materials, microbes are removed as they
deposit on the filter medium. After a period of operation, the effluent quality deteriorates
116
to an unacceptable level. Passage of microbial pathogens during the ripening period can
250
16000
14000
12000
10000
8000
6000
4000
2000
0
200
150
100
50
Infilitration rate
(cm/hr)
Effluent TC
Concentration (MPN
index/100 ml)
be very high, which Figure 5-14 indicates (AWWA, 1999).
0
3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Trial Day
T TC Concentration
CSW TC Concentration
T Infiltration rate
CSW Infiltration Rate
Figure 5-14. Total coliform concentrations versus infiltration rates for treatments T and
CSW.
Since there was direct correlation between a decrease in T columns’ effluent TC
concentration and a decrease in T columns’ infiltration rate, a correlation between E. coli
concentration and infiltration rate can be inferred, although not enough data were
obtained for E. coli.
Figure 5-15 is a graph of the three T columns’ effluent E. coli concentrations and
the average of the T column’s effluent TC concentration over the duration of the
experiment, showing similar concentration patterns.
Semil-log (Effluent E. coli
Concentration) (CFU/100ml)
117
100000
10000
1000
100
10
1
0.1
0
10
20
30
40
50
Trial Day
T1 E. coli
T2 E.coli
T3 E.coli
T Average Total Coliform
Figure 5-15. Semi-log plot of total coliform and E. coli concentration per trial.
From Figure 5-15, the average TC effluent concentrations were notably higher
than E. coli concentrations in the columns’ effluent for three of the four days examined.
This is highlighted by the fact there was no bacteria added to CSW column, yet they
show an increase of TC concentration, although not of E. coli concentration. This is
because TC is a measure of all coliforms, and E. coli is a member of coliform family,
Enterobacteriacea. Table 5-5 lists the other coliforms in the Enterobacteriacea family
(Leclerc et al. 2001).
Other coliforms occur in nature, some in soil. Alm et al. (2003) concluded in a
freshwater beach fecal indicator study that the presence of fecal indicator bacteria in
beach sand suggests that pathogenic bacteria of intestinal origin may also be present in
the sand. In this laboratory experiment, the sand was initially autoclaved, so low TC
concentrations were measured in the effluent in the first couple of runs.
Since the sand as well as the stormwater for the CSW treatments were initially
autoclaved, the effluent TC concentration in the CSW should be non-existent. It is
60
118
possible that some bacteria in the sand survived the autoclave due to the large volume of
sand autoclaved. Another explanation for the TC concentration in the CSW treatments’
effluent could be due to contamination during the laboratory set-up or bacteria in the
clear plastic columns.
A Gerba and McLeod (1975) lab study reported that survival of coliforms and E.
coli was attributed to the greater content of organic matter present in the sediment. Each
time a trial was performed, sediment was added to the T and CSW columns. This caused
the TC concentrations in the T and CSW column’s effluent to initially increase. The T
columns’ effluent increased more dramatically since the treatment contained bacteria. As
previously discussed, the initially high infiltration rates allowed for removal of bacteria
from the sand particles into the effluent. But as infiltration rates decreased, the
concentration of E. coli and TC in the column’s effluent decreased in similar fashion,
allowing for the bacteria to remain within the column and not in the effluent.
Laboratory findings resulted in a better understand of the DIS function on a
microbial level. These findings helped explain why there was bacteria loading in Site
M’s groundwater. In addition to better understanding of the system, the laboratory
results impacted DIS design. The correlation between the infiltration rate and bacteria
removal efficiency found in the laboratory experiment established a range of infiltration
rates suitable for designing a DIS with optimal bacteria removal efficiency. Furthermore,
the laboratory study measured the effect of sediment on infiltration rate and leads to the
development of a proposed maintenance schedule for the DIS.
119
Table 5-5. List of coliforms in the Enterobacteriacea family (Leclerc et al. 2001).
120
5.4 LABORATORY SUMMARY
The objectives of the laboratory study were as follows:
1. Determine the removal efficiency of E. coli by sand columns and if E. coli
removal efficiency was affected by sand clogging.
2. Determine the effect that the stormwater runoff contaminants have on the
infiltration rate in the sandy soil in order to determine a maintenance schedule
for the DIS.
For objective one it was found that the column’s effluent E. coli concentration
was greater than the influent E. coli concentration during trial 4. This could be due to the
large confidence intervals associated with the MPN testing method or due to the
relatively high infiltration rate of the treatment flushing out E. coli cultures located on the
sediment and top soil layers. By the 9th trial, the bacteria stormwater columns’ effluent
had E. coli concentrations ranging from 35 CFU/100 ml to 64 CFU/100 ml, lower than
the state’s standards. By the 14th trial run, E. coli concentrations in the test columns’
effluent were 1 CFU/100 ml or less.
Total coliform measurements were used to estimate E. coli removal efficient in
the bacteria stormwater treatment columns. There was a direct correlation (p<0.05)
between infiltration rate and total coliform concentration in the columns’ effluent. The
correlation between infiltration rate and total coliform concentration in the columns’
effluent can be used to estimate correlation between E. coli concentration and infiltration
rate since E. coli is in the Enterobacteriacea family of total coliforms.
121
As for objective two of the laboratory study, it was shown that the average
infiltration rate of bacteria-free stormwater columns was reduced by 50% after 11
treatments. After 18 treatments, the average infiltration rate of these columns was
equivalent to that of silt soils. For the bacteria-spiked stormwater columns, the
infiltration rate was reduced by 50% after 4 treatments and was equivalent to the
infiltration rate of a silt soil after 16 treatments. There was a significant difference
(p<0.05) between the infiltration rate of bacteria-free and bacteria-spiked stormwater
treatments during trials 16 through 20. The significant difference of the various
treatments’ average infiltration rate during trials 16 through 20 could be caused by
bacteria aggregation between the soil’s pore spaces of the bacteria-spiked stormwater
columns.
The primary goal of the lab study was to better understand bacteria removal
efficiency in the actually DIS system. Measuring the effect of clogging on infiltration
and bacteria removal will help better design the DIS system and devise a maintenance
schedule for it. Clogging allows for more bacteria removal, but at the cost of decreased
infiltration rates. The correlation between infiltration and bacteria removal established in
the laboratory study can be used to establish maintenance for various sized DISs.
122
6.0 Conclusions
6.1 Field Study
The main objective of this study was to analyze the DIS as a BMP. If the system
worked as designed, stormwater runoff that would have normally been directly
discharged onto the beach would be routed into the dunes and infiltrate the underlying
sand. The objectives outlined in Chapter 3 were used as guidelines to analyze the DIS.
One objective was to determine if routing and discharging stormwater runoff in
the dunes elevated the level of the groundwater beneath the dunes. Routing the 659 m3
(23,272 ft3) into Site L’s dune and 2237 m3 (78,999 ft3) into Site M’s dunes did not
substantially change the elevation of the water table. The largest storm-induced
fluctuation, 1.5 m (4.9 ft) occurred at Site M during Tropical Storm Ernesto. Preimplementation groundwater data shows a greater groundwater elevation increase during
2005 Hurricane Ophelia. Maximum tidal fluctuations cause groundwater to elevate 2.5 m
(8.1 ft). Thus, there is no substantial groundwater mounding beneath the DIS system at
Site L and Site M.
Another objective was to determine if routing and discharging stormwater runoff
into the dunes increased the bacteria level in the groundwater beneath the dunes. This
was tested by identifying a range of fecal coliform and enterococcus concentrations in an
urban coastal community’s stormwater runoff. Inflowing stormwater runoff had
concentrations of fecal coliform concentrations ranging from 181 CFU/100 ml to 19400
CFU/100 ml with a median of 8600 CFU/100 ml and from <10 CFU/100 ml to >2005
CFU/100 ml with a median of 1298 CFU/100 ml for enterococcus. The groundwater
123
concentrations were significantly less (p< 0.001) than the inflow with fecal coliform
concentrations ranging from <1 CFU/100 ml to 214 CFU/100 ml with a median of 1.5
CFU/100ml. For enterococcus concentrations the range was from <10 CFU/100 ml to
2005 CFU/100 ml with a median of 10 CFU/100ml. The groundwater fecal coliform
concentrations at both sites were significantly (p<0.01) less than the stormwater runoff
inflow concentration.
The purpose of the measuring the inflow and groundwater bacteria concentration
was to determine bacteria removal efficiency of DIS. North Carolina’s indicator bacteria
standards were exceeded only in Site M’s groundwater. Groundwater samples surpassed
the limit on 2 of the 25 events for fecal coliform and 6 of the 22 for enterococcus. These
incidents occurred five months after the system was implemented and during large storm
events. This indicated bacteria loading at Site M, due to large volumes of highly
concentrated bacteria infiltrating into the groundwater at a relatively fast rate. When
sediment starts clogging the DIS, bacteria blocking should become apparent, decreasing
the infiltration rate, allowing bacteria to concentrate in the sediment and top 30 cm (12
in) of the soil rather than filter through into the groundwater. As the results show, it can
be concluded that both Dune Infiltration Systems reduces pathogenic bacteria
concentrations in the short term and pathogenic bacteria concentrations reduction should
continue as more sediments is captured in the chambers.
Lastly, the Dune Infiltration System captured all runoff associated with the
designed rainfall intensity of 12.5 mm/hr (0.5 in/hr) or less. Thus, the DIS achieved the
goal of decreasing the potential health hazard associated with stormwater ocean outfalls.
Both DIS systems captured a measured total of 2,896 m3 (102,271 ft3) and bypassed 99
124
m3 (3,500 ft3). The DIS was designed to route stormwater runoff from storms with an
intensity of 12.5 mm/hr (0.50 in/hr) or less into the dunes. The DIS implemented at Site
L never overflowed, capturing storms with intensity 90 mm/hr (3.5 in/hr) or less. The
DIS at Site M overflowed 5 times capturing all storms with intensities up to 28 mm/hr
(1.1 in/hr). The overflowing storms produced stormwater runoff inflow rates greater than
the flow capacity of the 0.3 m (1ft) diameter pipe leading to the StormChambers™.
During the five overflow events, the maximum stage in the beginning chambers
only rose to a height of 0.69 m (2.25 ft), thus the full storage potential of 1.01m (3.34 ft)
of the StormChambers was not utilized. Since the maximum stage in chambers was
never reached, the pipes leading to the chambers appeared to have been inadequately
sized. If the system were correctly designed to mitigate stormwater runoff resulting from
rainfall intensities of less than 12.5 m/hr (0.5 in/hr) by decreasing the diameter of the
inflow pipe and the amount of chambers, the volume of stormwater entering the
groundwater will decrease. This could allow bacteria concentration to remain under the
state’s acceptable limit, but cause more stormwater runoff to discharge directly onto the
ocean. Contrastingly, the inflow pipe leading to the chambers could be increased in
diameter to utilize of all the chambers. This could have potential captured all the runoff
at Site M. However, routing the extra stormwater runoff into the dunes could have
caused pathogenic bacteria loading in the groundwater and potential groundwater
mounding below the DIS.
The DIS has potential to an effective BMP at the remaining ocean outfalls in Kure
Beach and elsewhere. The DIS implementation at other beach’s ocean outfalls is a
possibility only if the groundwater is well below the surface. If more DIS’s were to be
125
implemented, the current design should be adjusted. When calculating infiltration
capacity of the system, the Green-Ampt equation is more appropriate than the Darcy
equation. Using the Green-Ampt equation when designing the DIS will decrease the
number of chambers needed to capture the design storm of12.5 mm/hr (0.5 in/hr). But
using Darcy’s equation will allow for more chambers, giving the design a built in factor
of safety.
Another design alterative would be to use various design storm intensities for
various sized watersheds. Site L captured storms with higher intensities than Site M, and
Site L’s groundwater bacteria concentration remained under the standards. Thus, smaller
sized watersheds like Site L could be designed for larger intensities, since relatively
smaller volumes of stormwater will be routed there. Relatively larger sized watersheds
like Site M may be designed using a less intense storm since that DIS will be capturing
relatively larger volumes of runoff. By adjusting design storm intensities, the appropriate
volume of stormwater runoff will be treated. The large watersheds will not cause
bacteria loading in groundwater beneath the DIS and the smaller watersheds could use
DIS to its full capacity.
6.2 Laboratory Summary
One objective of the laboratory study was to determine the effect that stormwater runoff
contaminants have on the infiltration rate in the sandy soil to determine a maintenance
schedule for the BMP sand filter. It was shown that the average infiltration rate of
bacteria-free stormwater columns was reduced by 50% after 11 treatments. After 18
treatments, the average infiltration rate of these columns was equivalent to that of silt
126
soils. For the bacteria-spiked stormwater columns, the infiltration rate was reduced by
50% after 4 treatments and was equivalent to the infiltration rate of a silt soil after 16
treatments. There was a significant difference (p<0.05) between the infiltration rate of
bacteria-free and bacteria-spiked stormwater treatments during trials 4 through 7 and then
from trials 16 through 20. The significant difference from trials 4 through 7 could be due
to an outlier infiltration rate in one bacteria free treatment column. The significant
difference of the various treatments’ average infiltration rate during trials 16 through 20
could be caused by bacteria aggregation between the soil’s pore spaces of the bacteriaspiked stormwater columns.
An additional objective was to determine the removal efficiency of E. coli by sand
columns and if E. coli removal efficiency is affected by sand clogging. By the 9th trial,
the bacteria stormwater columns’ effluent had E. coli concentrations ranging from 35
CFU/100 ml to 64 CFU/100 ml, lower than the state’s standards. By the 14th trial run, E.
coli concentrations in the test columns’ effluent were 1 CFU/100 ml or less.
Total coliform measurements were used to estimate E. coli removal efficient in
the bacteria stormwater treatment columns. There was a direct correlation (p<0.05)
between infiltration rate and total coliform concentration in the columns’ effluent. The
correlation between infiltration rate and total coliform concentration in the columns’
effluent can be used to estimate correlation between E. coli concentration and infiltration
rate since E. coli is in the Enterobacteriacea family of total coliforms.
A short term maintenance schedule for DIS’s was extrapolated from these
laboratory results. Two correlations from the laboratory experiment helped established a
maintenance schedule for the DIS. The correlation between trial number and infiltration
127
rate was extrapolated to determine the maintenance frequency, while the relationship
between infiltration rate and bacteria removal was used to determine at what time of year
maintenance should be preformed. It was found at the 12th trial, the infiltration rate was
reduced by 50%. Each column represented a chamber in Site L’s DIS. Thus, a
relationship was established (EQN 6-1).
Md = 12* NC * FRF
(6-1)
Where: Md = the number of days before maintenance,
NC = the number of chambers in the DIS,
FRF = frequency of rainfall.
The frequency of rainfall at Kure Beach, as mentioned in Chapter 5, was every 3
days. Thus, Site L’s DIS systems should be maintained every 1.1 years and Site M’s DIS
should be maintained every 2.1 years. (It should be noted that stormwater collected for
the lab experiment from Raleigh, NC, so there may be variation in sediment type and
amount found in Kure Beach’s stormwater runoff). DIS maintenance would require a
vacuum truck to attach to the clean-out pipes located at the beginning and end of each
chamber section (4 clean-out pipes in each of the DIS systems). Maintenance would also
include cleaning sediment and debris from the monitoring vault.
Field measurements coincide with the laboratory extrapolated maintenance
schedule. During the eight months of monitoring the system, there were no signs of
clogging. The systems never fully utilized all the chambers and retained the same storage
capacity throughout the experiment.
The timing of maintenance is important. The laboratory study established that the
infiltration rate of 63 cm/hr (25 in/hr) was needed to remove enough bacteria so that the
effluent’s bacteria concentration meets state standards. This occurred during trial 8. The
128
highest inflow bacteria concentrations in the field occurred during the months of July
through September. For these months, infiltration rates of 63 cm/hr (25 in/hr) or less
would optimize bacteria removal. Thus, DIS maintenance should be performed just
following the summer months. For Site L and Site M maintenance is recommended to be
performed between the months of November through January. This will allow for
slower infiltration rates during summer months (highest concentration of bacteria) to help
reduce the high concentration of bacteria from entering the groundwater below the
system.
The primary goal of the lab study was to better understand bacteria removal
efficiency in the actually DIS system. Measuring the effect of clogging on infiltration
and bacteria removal will help better design the DIS system and devise a maintenance
schedule for it. Clogging allows for more bacteria removal, but at the cost of decreased
infiltration rates. The correlation between infiltration and bacteria removal established in
the laboratory study can be used to establish maintenance for various sized DISs.
6.3 Overall Recommendations
The overall goal of the project was to evaluate the possibility of DIS as a potential BMP.
EPA (1999) defines BMP as “a technical measure or structural control that is used for a
given set of conditions to manage the quantity and improve the quality of stormwater
runoff in the most cost effective manner.” For the Kure Beach sites, the DIS was a viable
BMP. Both Dune Infiltration Systems did significantly (p<0.01) reduce the amount and
rate of stormwater directly discharging into the ocean, while not substantially altering
groundwater hydrology. Also, the groundwater fecal coliform concentrations at both
129
sites were significantly (p<0.01) less than the stormwater runoff inflow concentration.
But Site M’s groundwater bacteria concentrations exceeded the NC standards twice for
fecal coliform and six times for enterococcus. Thus, designing a system for a watershed
larger than Site M’s 3.3 ha (8.1 ac) watershed might further increase groundwater
pathogenic bacteria. Since Site M’s DIS captured storms with intensity of 2.54 cm/hr (1
in/hr) or less, a DIS system could be designed for a watershed area twice the size of Site
M with half the storm intensity.
The laboratory results showed greater reductions in effluent E. coli concentration
with slower infiltration rates. The experiment also demonstrated the quickness in which
the bacteria and bacteria-free stormwater treatment columns’ infiltration rate decreased.
For realistic DIS soil infiltration rate, the infiltration rate of media in the DIS should be
designed at half the measured infiltration rate.
In conclusion, the following recommendations are suggested:
1) Dune Infiltration System should be implemented at other ocean outfalls
that drain a watershed area of less than 6.6 ha (16 ac)
2) Dune Infiltration Systems should be designed with the following changes:
i. Design using Green-Ampt equation for infiltration.
ii. Design using 50% of measured infiltration rate.
iii. Multiply the numbers of chambers by safety factor.
iv. Design storm intensity should maximize number of storms
captured without overloading the system.
3) Dune Infiltration Systems must be properly maintained.
Thus, if Kure Beach has another watershed comparable to Site L’s watershed, a
DIS system should be implemented. A design storm intensity of 89 mm/hr (3.5 in/hr)
130
and an infiltration rate of 136 cm/hr (6.1 ft/hr) (50% of the average measured of 372
cm/hr (12.2 ft/hr)) would be used. When using Green-Ampt equation to estimate actual
soil infiltration capacity, it was calculated that five chambers were needed. Using a
safety factor of 1.5, a total of six chambers with a 0.3 m (1.0 ft) diameter inflow pipe
would be required to capture storm’s with intensities of 89 mm/hr (3.5 in/hr) or less. The
maintenance schedule for this system would be around every seven months, occurring at
end of March as well as the end of October.
131
7.0 Future Research
The main reason for installing DIS’s was to protect public health by reducing the public’s
exposure to harmful bacteria and other microorganisms. Transferring the problem
elsewhere is not a long term solution. If stormwater bacteria are simply being diverged
into groundwater simply delaying exposure to beach goers, than the beach goers might
still be at risk. Future DIS design and bacteria research is recommended before the DIS
can achieve widespread implementation.
First and foremost, monitoring should continue at both DISs. These systems were
monitored for 8 months. It is recommended that Sites L’s and M’s DIS system should be
monitored for at least full year, if not a complete second year, to better understand the
system. Further monitoring of these sites will establish if there is seasonal variation in
stormwater bacteria levels and if there are any trends not detected in earlier research.
Future research will also test and refine the recommended maintenance schedule.
Research is also needed for some of the new DIS design suggestions. It is
suggested that the DIS should be designed with a storm intensity that maximizes the
number of storms captured without overloading the system. Future research could
establish a correlation between a watershed’s discharge to design storm intensity. This
study showed various loading capacities at the two sites with different watershed areas,
but similar watershed characteristics. Assuming most coastal communities have similar
watersheds, the bacteria concentration associated with the stormwater runoff is expected
to be relatively similar. Thus, the amount of stormwater entering the system as well as
amount of media (sand) between the surface and the water table primarily controls
groundwater bacteria loading. Both of these parameters can be tested and used to
132
quantify the amount of stormwater that could enter the system without overloading the
groundwater with bacteria.
Bacteria loading was only quantified in the groundwater, not in the soil. The
laboratory study showed that there was a correlation between bacteria removal and
infiltration rate. The laboratory study did not examine the bacteria concentration in the
columns’ soil. The field study measured the amount of bacteria in the groundwater and
not within the soil. A study performed in Michigan at six freshwater beaches found fecal
indicator bacteria were more abundant in sand than in water. Compared to water,
enterococci counts in sand were 4–38 times higher and E. coli counts were 3–17 times
higher. The results of this study were consistent with work on freshwater beaches in
England, where fecal indicator counts were an order of magnitude greater in sand than in
the overlying water (Alm et al. 2003).
Future research should be performed analyzing the DIS soil’s ability to retain
bacteria. Although it might be difficult to obtain soil samples at various depths from
under the chamber, it is necessary that the soil samples come from the soil directly below
the StormChambers™. Kjeldgaard and Ranade (1966) reported that bacterial cells in the
environment vary in size and chemical composition. The variation is caused by
fluctuations in environmental conditions when cells are transitioning from the resting
state to the exponential growth phase. Cohen and Barner (1954) proposed that bacteria
survive in media containing a small number of specific nutrients, a condition that could
be extrapolated to aquatic environments, where metabolites leach away and do not
accumulate. In order to fully understand how the bacteria are acting within the DIS
system, the sample should come from the field.
133
The system was conservatively designed, but the results of this research indicated
that the number of DIS chambers could be limited. This will allow for the whole system
to be used and decrease the cost. In addition, when all the chambers are used, the stage
of water within the chambers will increase, causing a longer anaerobic period. This may
affect the life cycle of the microorganisms in the soil. In order to devise an effective
bacteria study, the DIS system studied needs to be designed as recommended to the
public, with the Green-Ampt equation.
Lastly, the DIS has only been tested in one type of soil, Newhan Fine sand,
composed of 99.4 % sand and 0.6% silt (NRCS, 2005). Coastal soils vary across the
nation. If Dune Infiltration Systems are to be implemented in coastal communities
around the nation, additional research is needed locally. Laboratory or field research
should analyze various coastal soils’ bacteria removal efficiencies and well as clogging
rates. This will result in an improved DIS design with an appropriate maintenance
schedule for various soil classifications.
134
REFERENCES
Abu-Ashour, J. and M. Abu-Zreig, (2005). "Effect of interstitial velocity on the
adsorption of bacteria onto soil." Adsorption Science Technology, 23(7), 535.
Alm, E. W. J. Burke, and A. Spain. (2003). "Fecal indicator bacteria are abundant in wet
sand at freshwater beaches." Water Research, 37(16), 3978.
American Public Health Association, Water Pollution Control Federation. (1999)
“Standard Methods for Examination of Water and Wastewater, 20th edition.” American
Public Health Association.
American Society for Testing and Materials (ASTM). (2003). "Standard Test Method for
Infiltration Rate of Soils in Field Using Double-Ring Infiltrometer." Rep. No. ASTM D
3385, West Conshohocken, Pa.
American Society for Testing and Materials (ASTM). (2002). “Standard Test Method for
Particle-Size Analysis” Rep. No ASTM D 422, West Conshohocken, Pa.
American Water Works Association (AWWA), Symons, James M., and Bradley, L., Jr.
(2001). The Drinking Water Dictionary. McGraw-Hill, Blacklick, Ohio.
American Water Works Association (AWWA). (1993). Water Quality & Treatment 5th
edition. McGraw-Hill Inc., Blacklick, Ohio.
Baker, D., B. Olsan, and J. Semenza. (2005). "Researcher Estimate Polluted O.C.
Beaches Cost Public $3.3 Million Annually." Science Daily.
Barrett, M. E. (2003). "Performance, Cost, and Maintenance Requirements of Austin
Sand Filters’." Journal of Water Resources Planning and Management, 129(3), 234-242.
Bean, E.Z. (2005). “A Field Study To Evaluate Permeable Pavement Surface Infiltration
Rates, Runoff Quantity, Runoff Quality, and Exfiltrate Qualtiy.” M.S thesis, North
Carolina State University, Biological and Agricultural Engineering Department.
Bolster, C. H., A.L Mills, G.M. Hornberger, and J.S Herman. (2001). "Effect of surface
coatings, grain size, and ionic strength on the maximum attainable coverage of bacteria
on sand surfaces." Journal of Contaminant Hydrology, 50(3), 287.
Bordalo, A. R. Onrassami, and C. Dechsakulwantana (2002). "Survival of faecal
indicator bacteria in tropical estuarine waters (Bangpakong River, Thailand)." Journal of
Applied Microbiology, 93(5), 864.
Brown, D. G. (2002). "Effects of porous media preparation on bacteria transport through
laboratory columns." Water Research, 36(1), 105.
135
Buerge, I. (2003). “Caffeine, an Anthropogenic Marker for Wastewater Contamination
of Surface Waters,” Environmental Science & Technology, 37, 691–700.
Buol, S.W., R.J. Southard, R.C. Graham, and P.A. McDaniel. (1997). Soil Genesis and
Classification. 5th edition. Iowa State press, State Avenue Ave, Iowa.
Bushaw-Newton, K.L. and K.G. Sellner, (1999). “Harmful Algal Blooms,” State of the
Coast Report, National Oceanic and Atmospheric Administration, Silver Spring, MD,
1999.
Carlucci, A.F. and D. Pramer. (1959). “Factors Affecting Survival of Bacteria in
Seawater.” Applied Microbiology, 7, 388-392.
Chen, J., S. Truesdail, F. Lu,G. Zhan, C. Belvin, B. Koopman, S.Farrah and D. Shah.
(1998). "Long-term evaluation of aluminum hydroxide-coated sand for removal of
bacteria from wastewater." Water Research, 32(7), 2171.
Chong, J., and Wride, N. (2005). "Southland’s Record Rainfall; Rain Overwhelms
Sewage Systems; Massive amount of runoff cause wastewater to overflow from pipes and
treatment plants, fouling Southland waterways and beaches." Los Angles Times, B 1.
City of Austin. (1990). “Removal Efficiencies of Stormwater Control Structures.”
Environmental Resources Management Division, Environmental and Conservation
Services Department, City of Austin, Austin, TX.
City of Boise Public Works Professional Advisory Group. (1998). "Boise Storm Water
Management Design Manual." http://www.cityofboise.org/public_works/services/
water/storm_water/manual/index.aspx?id=design_manual (July, 20, 2005).
H
Clark, G., and Stoner, N. (2001). "Stormwater Strategies: The Economic Advantage."
Stormwater, 2(1), 10-18.
Cohen, S.S. and H.D Barner. (1954). “Studies on Unbalanced growth in Escherichia
coli.” National Academy of Science, 40(1)179-185.
Dee, D. (1997). "Water Quality Retrofit of an Existing Drainage Basin Using a Sand
Filter Design." 24th Annual Water Resources Planning and Management Conference:
Aesthetics in the Constructed Environment, Civil Engineering, Texas, 205-210.
Dorfman, M. (2005). "Testing the Waters 2006: A Guide to Water Quality at Vacation
Beaches." Natural Resources Defense Council.
Dorfman, M. (2004). "Testing the Waters 2005: A Guide to Water Quality at Vacation
Beaches." Natural Resources Defense Council.
136
Entry, J.A. and N. Farmer. (2001). "Movement of coliform bacteria and nutrients in
ground water flowing through basalt and sand aquifers." Journal of Environmental
Quality, 30(5), 1533.
Feagans, B. (1999). "Hanby Pollution Hazard Frustrates Neighborhood." Morning Star,
Wilmington N.C, B 1-4.
Gerba, C.P. and J.S. McLeod. (1975). “Effect of Sediments on the Survival of
Escherichia Coli in Marine Waters.” Applied and Environmental Microbiology, 32(1)
114-120.
Gomez, M. (2006). "Urban wastewater disinfection by filtration technologies."
Desalination, 190(1), 16.
Grant, D. M., and B. D Dawson. (2001). Isco Open Channel Flow Measurement
Handbook. 5th edition. Isco, Inc., Lincoln, Nebraska.
Greenberg, A.E. (1956). “Survival of Enteric Organisms in Sea Water.” Public Health
Rep, 71,77-78.
Griffin D. W., K.A. Donaldson, J. H. Paul, and J. B. Rose. (2003). "Pathogenic human
viruses in coastal waters." Clinical Microbiology Reviews, 16(1), 129.
Grisham, D. (1995). "Designing for the ‘First Flush." Civil Engineering, 65(11), 67-69.
Haile, R. W., J. Alamillo, K. Barrett, R. Cressey, J. Dermond, C. Ervin, A. Glasser, N.
Harawa, P. Harmon, J. Harper, C. McGee, R. C. Millikan, M. Nides, and J. S. Witte
(1996). "An Epidemiological Study of Possible Adverse Health Effects of Swimming in
Santa Monica Bay." Santa Monica Bay Restoration Project, CA.
Hawthorne, M. (2004). "Seagull Droppings Top Source of Chicago-Area Beach
Bacteria." Knight Ridder Tribune Business News, 1.
Hunt, W. F., A.R. Jarrett, J.T. Smith, and L.J. Sharkey (2006). "Evaluation Bioretention
Hydrology and Nutrient Removal at three Field Sites in North Carolina.” Journal of
Irrigation and Drainage Engineering, 132(6)600-608.
Hunt, W. F. (1999). "Stormwater Structures Best Management Practices (BMPs)." Rep.
No. AG-588-1, North Carolina Cooperative Extension Service, Raleigh, NC.
International Hydrographic Bureau (IHO). (2001). International Hydrographic Bureau
(IHO) Dictionary S-32, 5th Edition, 3145.
Jiang, S., R. Noble, and W. Chu (2001). "Human adenoviruses and coliphages in urban
runoff-impacted coastal waters of Southern California." Applied and Environmental
Microbiology, 67(1), 179.
137
Kay, D., J. M. Fleisher, R.L. Salmon, F. Jones, M.D. Wyer, A.F. Godfree, Z. ZelenauchJacquotte, and R. Shore (1994). "Predicting Likelihood of Gastroenteritis from Sea
Bathing: Results from Random Exposure." The Lancet, 344, 905-911.
Korganonkar, K.S., and S.S Ranade. (1966). “Evaluation of Acridine Orange Fluorescence Test in Viabaility Study on Escherichia coli.” Canadian Journal of
Microbiology, 12, 186-190.
Lleo, M.M., B. Bonato, M.C. Tafi, C. Signoretto, C. Pruzzo and P. Canepari. (2005).
"Molecular vs culture methods for the detection of bacterial faecal indicators in
groundwater for human use." Letters in Applied Microbiology, 40(4), 289.
Lukasik, J., Y. Cheng, F. Lu, M. Tamplin, and S. R. Farrah. (1999). "Removal of
microorganisms from water by columns containing sand coated with ferric and aluminum
hydroxides." Water Research, 33(3), 769.
Madigan, M.T., J.M Martinko, and J. Parker. (2003). Brock Biology of Microorganisms.
10th edition. Pearson Education, Inc. Upper Saddle River, NJ.
Malcom, H. R. (1989). Elements of Urban Stormwater Design. North Carolina State
University, Raleigh, NC.
Mandelstam, J. (1958) . “Turnover of protein in growing and non-growing populations of
Escherichia coli.” Biochemistry Journal 69,110-119.
Manivannan ,S. and S. Sundar Raman (2002). “Green-Ampt Runoff Model: A Review.”
Indian J. Soil Cons, 31(2),105-113.
Marshall, K., and M. Ritch, (2005). "Group Joins to Curb Pollution." The Sun New
Myrtle Beach, SC.
Martin, M. J., B.E Logan, W.P. Johnson, D.G Jewett, and R.G Arnold. (1996). "Scaling
bacterial filtration rates in different sized porous media." Journal of Environmental
Engineering, 122(5), 407.
Mehta, S. (2002). "The Region; For Coast, When it Rains, It’s Poor; Pollution: Heal the
Bay’s annual Beach Report Card gives high marks during dry weather, but storms bring
out the worst in toxic runoff." Los Angles Times, B 7.
Meschke, J. S. (2003). "Comparative reduction of Norwalk virus, poliovirus type 1, F
RNA coliphage MS2 and Escherichia coli in miniature soil columns." Water Science and
Technology, 47(3), 85.
Mitchell, R. (1968). “ Factors affecting the Decline of Non-Marine Microorganism in
Seawater.” Water Resource, 2, 535-543.
138
National Marine Fisheries Service (NMFS). (2006) “Endangered and Threatened
Species: Final Listing Determinations for Elkhorn Coral and Staghorn Coral,” Federal
Register, 71(89), 26852.
National Oceanic Atmospheric Administration(NOAA). (2006). “Tides and Currents.”
http://tidesandcurrents.noaa.gov/( 1 November 2006).
H
National Research Council. (2003). Oil in the Sea III: Inputs, Fates, and Effects. National
Academy Press, Washington D.C..
Nature Resource Conservation Service (NRCS). (2006). "Soil Data Mart: NC 129-New
Hanover County, North Carolina." http://soildatamart.nrcs.usda. gov (1 October 2006).
H
North Carolina Shellfish Sanitation and Recreational Water Quality Section, Shellfish
Sanitation Section (2005). Shellfish Sanitation Section Marine Fisheries NCDENR..
http://www.deh.enr.state.nc.us/shellfish/ (27 June 2005).
H
Paerl, H.W., J.L. Pinckney, J.M. Fear, and B.L. Peierls . (1999). "Fish kills and bottomwater hypoxia in the Neuse River and Estuary: reply to Burkholder et al." Marine
Ecology Progress Series, 186, 307.
Potts, J. D. (2005). "North Carolina Beach Monitoring Project Quality Assurance Project
Plan." Division of Environmental Health: North Carolina Shellfish Sanitation &
Recreational Water Quality Section, Raleigh, NC.
Probacso, M. (2005). "Study: Raw sewage killing coral reefs." STLtoday.Com.
Rabinovici , S.J., R.L. Bernknopf, A.M. Wein, D.L. Coursey, and R.L. Whitman. (2004).
“Economic and Health Risk Trade-Offs of Swim Closures at a Lake Michigan Beach,” in
Environmental Science and Technology, 38 (10), 2742.
Rawlins, W. (2003). "Coastal Wetlands Reborn." The Raleigh News & Observer, March
9.
Roszak, D.B and R.R. Colwell. (1987). “Survival Strategies of Bacteria in the Natural
Environment.” Microbiological Reviews, 51(3), 365-379.
Schoch, D. (2000). "Board to Order Halt to Discharge on Beach." Los Angles Times, 3.
Schoch, D. (1996). "Bolsa Chica State Beach Closed by Sewage Leaks; Health: Officials
declare two-thirds of popular spot off-limits indefinitely after bacteria levels escalate."
Los Angles Times, 1.
Schwab, G. O., Fangmeier, D. D., Elliot, W. J., and Frevert, R. K. (1993). Soil and Water
Conservation Engineering. 4th ed. John Wiley & Sons, Inc., New York.
139
Schwartz, F. W., and Zhang, H. (2003). Fundamentals of Ground Water. John Wiley &
Sons Inc., New York.
Shapiro, N. (2003). "The Stranger Amongst Us: Urban runoff, the Forgotten Local Water
Resource." National Conference on Stormwater: Enhancing Programs at Local Levels,
U.S. Protection Agency, Ohio, 395-408.
Shaver, E. (1994). "Beach Community Adds Sand Filters to Storm Drain." Water
Environment & Technology, 5(5), 18-20.
Skaggs, R.W., L. E. Huggins, E. J. Monke, and G. R. Foster (1969). “Experimental
Evaluation of Infiltration Equations.” ASAE, 12 (6).
Smith, P. (2004). "Swimming Advisories posted for AB Water." The Daily News,
Jacksonville, NC.
Smith, P. (2003). "Stormwater Runoff in Coastal North Carolina Closes Shellfishing."
Daily News; Jacksonville NC.
Sobey, M., R. Perdue, M. Overton, and J. Fisher. (2003). "Factor influencing fecal
contamination in coastal marinas." Water Science & Technology, 47(3), 199-204.
Spooner, J. (1991). “Censored Data Values: Description and Effect of Censoring on
Statistical Trend Analyses (Part 2).” NWQEP Notes 48.
State Climate Office of North Carolina (2006). “North Carolina Climate Retrieval and
Observation Network Of the Southeast (NC CRONOS) database- Wilmington (Wso)
Airport, NC (319457).” http://www.nc-climate.ncsu.edu/ (March 3, 2006).
H
Strecker, E., L. Mayo, M. Quigley, and J. Howell. (2001). “Stormwater Best
Management Practices in an Ultra-Urban Setting: Selection and Monitoring” U.S
Department of Transportation Federal Highway Administration (FHWA) Report.
Tester, P.A., R.P. Stumpf, F.M. Vukovich, P.K. Folwer, and J.T. Turner. (1991) “An
expatriate red tide bloom: Transport, distribution, and persistence.” Limnology and
Oceanography, 36,1053–1061.
The Associated Press. (1999). "Damage Sever in Coastal Waters." Greensboro News
Record, B 5.
Trainer, V.L. (2002). “Unveiling an Ocean Phantom.” Nature, 413, 925–926.
Urbonas, B. (1999). "Design of a Sand Filter for Stormwater Quality Enhancement."
Water Environment Research, 71(1), 102-114.
140
US EPA. (2006a). "Bacterial Rule for Coastal and Great Lakes Recreational Water."
http://www.epa.gov/waterscience/beaches/bacteria-rule.htm (May, 11, 2005).
H
US EPA. (2006b). "Basic Information on the Beach Standards, Monitoring, and
Notification." http://www.epa.gov/waterscience/beaches/about.html (May 16, 2006).
H
US EPA. (2006c). "Laws and Regulations: Clean Water Act." http://www.epa.gov/
region5/water/cwa.htm (May 16, 2005).
H
US EPA. (2004). “Report to Congress: Impacts and Control of CSOs and SSOs.” April
24, 2004 EPA 833-R-04-001.
US EPA. (2003). "Introduction to the Clean Water Act." http://www.epa.gov
/watertrain/cwa/ (May 16, 2005).
H
US EPA. (2002a). "Federal Water Pollution Control Act." 2004(33 U.S.C), 101-607.
U.S. EPA. (2002b). Urban Stormwater BMP Performance Monitoring: A Guidance
Manual for Meeting the National Stormwater BMP Database Requirements. EPA-821-b01-001. Pg 128. Washington, D.C.: U.S. Environmental Protection Agency.
US EPA. (2000). "Liquid Assets 2000: America's Water Resources at a Turning Point."
Rep. No. EPA-840-B-00-001, United States Environmental Protection Agency,
Washington D.C.
US EPA. (1999a). "Preliminary Data Summary of Urban Storm Water Best Management
Practices." Rep. No. EPA-821-R-99-012, US EPA Office of Water, Washington, D.C.
US EPA. (1999b). "Storm Water Technology Fact Sheet Sand Filters." Rep. No. EPA832-F-99-007, US EPA Office of Water, Washington, D.C.
Vandevivere, P and P. Baveye (1992). "Saturated hydraulic conductivity reduction
caused by aerobic bacteria in sand columns." Soil Science Society of America Journal,
56(1), 1.
Van Donsel, D.J, E.E. Geldrich, and N.A Clark. (1967), “Seasonal Variations in
Survival of Indicator Bacteria in Soil and Their Contribution to Stormwater Pollution.”
Applied Microbiology, 15(6), 1362-1370.
Whitlock, J.E, D.T. Jones, and V.J. Harwood. (2002). “Identification of the Sources of
Fecal Coliforms in an Urban Watershed using Antibiotic Resistance Analysis.” Water
Research, 36(4), 4273-4282.
Wossink, A. and B. Hunt. (2003). “The Economics of Structural Stormwater BMPs in
North Carolina.” Water Resources Research Institute of North Carolina. WRRI-50260.
Weiss, T. H. (1995). "Effect of bacterial cell shape on transport of bacteria in porous
media." Environmental Science Technology, 29(7), 1737.
141
APPENDIX SECTION
142
A.0 Appendix A-Field Study Storm Summary
A.1 Site L Total Summary
Table A-1. Summary Table of the 25 Storm Events at Site L.
Storm
Date
3/21/2006
4/16/2006
4/26/2006
5/7/2006
5/14/2006
5/15/2006
5/20/2006
6/5/2006
6/12/2006
6/14/2006
6/25/2006
6/26/2006
6/27/2006
7/6/2006
7/16/2006
7/23/2006
7/25/2006
7/30/2006
8/21/2006
8/22/2006
9/1/2006
9/6/2006
9/13/2006
10/8/2006
10/17/2006
Rainfall
Amount
(mm)
11.9
19.3
26.4
13.0
20.6
3.8
23.4
9.1
7.9
17.0
8.4
6.6
5.6
11.7
4.6
40.1
29.0
4.1
10.7
48.8
105.2
8.6
49.8
76.2
6.6
Duration
(hr)
10.3
N/A
N/A
2.0
3.3
0.3
19.1
5.9
11.6
4.0
2.4
3.5
6.1
4.8
1.8
24.3
23.3
8.2
0.7
6.4
21.8
13.1
10.8
15.3
18.9
Peak
Intensity
(mm/hr)
2.79
N/A
N/A
33.53
30.48
14.30
10.20
41.15
5.08
39.62
73.15
15.25
11.12
27.94
18.30
50.80
43.69
1.27
19.30
88.90
22.86
12.19
6.10
88.90
4.32
Peak
Flow
(m3/s)
0.002
0.026
0.012
0.011
0.006
0.005
0.004
0.011
0.002
0.016
0.008
0.005
0.003
0.003
0.006
0.017
0.019
0.001
0.004
0.014
0.012
0.003
0.013
0.039
0.002
Runoff
Watershed
Depth*
(mm)
0.57
0.95
3.10
0.70
1.06
0.41
1.66
1.79
0.87
1.13
0.42
0.43
0.50
0.40
0.28
2.15
2.01
0.51
0.32
1.20
6.60
0.70
2.87
4.79
0.78
Total=
Total
Runoff
Volume
Captured
(m3)
10.3
17.3
56.4
12.8
19.3
7.4
30.2
32.7
15.8
20.5
7.6
7.8
9.1
7.2
5.1
39.2
36.5
9.4
5.9
21.9
120.2
12.8
52.2
87.2
14.2
659
Total
Runoff
Volume
Bypass*
(m3)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
143
A.2 Site M Total Summary
Table A-2. Summary Table of the 25 Storm Events at Site M.
Storm
Date
3/21/2006
4/16/2006
4/26/2006
5/7/2006
5/14/2006
5/15/2006
5/20/2006
6/5/2006
6/12/2006
6/14/2006
6/25/2006
6/26/2006
6/27/2006
7/6/2006
7/16/2006
7/23/2006
7/25/2006
7/30/2006
8/21/2006
8/22/2006
9/1/2006
9/6/2006
9/13/2006
10/8/2006
10/17/2006
Rainfall
Amount
(mm)
11.9
19.3
26.4
13.0
20.6
3.8
23.4
9.1
7.9
17.0
8.4
6.6
5.6
11.7
4.6
40.1
29.0
4.1
10.7
48.8
105.2
8.6
49.8
76.2
6.6
Duration
(hr)
10.3
N/A
N/A
2.0
3.3
0.3
19.1
5.9
11.6
4.0
2.4
3.5
6.1
4.8
1.8
24.3
23.3
8.2
0.7
6.4
21.8
13.1
10.8
15.3
18.9
Peak
Intensity
(mm/hr)
2.79
N/A
N/A
33.53
30.48
14.30
10.20
41.15
5.08
39.62
73.15
15.25
11.12
27.94
18.30
50.80
43.69
1.27
19.30
88.90
22.86
12.19
6.10
88.90
4.32
Peak
Flow
(m3/s)
0.002
0.048
0.043
0.028
0.013
0.010
0.017
0.030
0.002
0.053
0.023
0.014
0.005
0.019
0.015
0.059
0.062
0.002
0.015
0.055
0.047
0.009
0.054
0.180
0.004
Runoff
Watershed
Depth*
(mm)
0.70
1.36
5.79
1.18
1.38
0.41
2.07
3.26
1.16
2.29
0.72
0.53
0.53
0.90
0.41
5.58
5.14
0.59
0.69
2.77
17.05
1.60
6.65
11.13
0.67
Total =
Total
Runoff
Volume
(m3)
22.8
44.3
189.1
38.5
45.2
13.3
67.5
106.6
37.7
72.7
23.6
17.4
17.3
29.4
13.5
175.8
163.0
19.2
22.6
87.9
556.7
52.3
217.0
280.2
21.9
2336
Total
Runoff
Volume
Bypass*
(m3)
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2.1
0.0
0.0
0.0
0.0
0.0
6.3
4.8
0.0
0.0
2.6
0.0
0.0
0.0
83.3
0.0
99
144
A.3 Individual Storm Summary
Table A-3. March 21, 2006 Storm Summary.
Inflow (m 3/s)
L
M
Rainfall
Duration
(hr)
10.3
10.3
Peak
Intensity
(mm/hr)
2.794
2.794
Peak
Flow
(m3/s)
0.0022
0.0025
Total
Runoff
Volume
Bypass
(m3/s)
0
0
M vs. L
Peak Flow
Difference
(%)
11%
Site M vs.
L Runoff
Volume
Difference
(%)
55%
0.0025
0.5
0.002
0.4
0.0015
0.3
0.001
0.2
0.0005
0.1
0
0
3/21/2006 3/21/2006 3/21/2006 3/21/2006 3/21/2006 3/21/2006 3/21/2006
2:24
4:48
7:12
9:36
12:00
14:24
16:48
Rainfall Amount (cm)
Rainfall
Amount
(mm)
11.9
11.9
Total
Runoff
Volume
(m3/s)
10.3
22.9
Date
Inflow
Rainfall
0.003
0.6
0.0025
0.5
0.002
0.4
0.0015
0.3
0.001
0.2
0.0005
0.1
0
3/21/2006
2:24
3/21/2006
4:48
3/21/2006
7:12
3/21/2006
9:36
3/21/2006
12:00
3/21/2006
14:24
3
Inflow (m /s)
Figure A-1. Site L Inflow Hydrograph and Rainfall Amount for March 21, 2006 Storm.
0
3/21/2006
16:48
Date
Inflow
Rainfall
Figure A-2. Site M Inflow Hydrograph and Rainfall Amount for March 21, 2006 Storm.
145
Table A-4. April 17, 2006 Storm Summary.
L
M
Rainfall
Amount
(mm)
19.304
19.304
Rainfall
Duration
(hr)
N/A
N/A
Peak
Intensity
mm/hr
N/A
N/A
Peak
Flow
(m3/s)
0.026
0.048
Total
Runoff
Volume
(m3/s)
17.3
44.3
Total
Runoff
Volume
Bypass
(m3/s)
0.0
0.0
Site M vs.
L Peak
Flow
Difference
(%)
46%
Site M vs.
L Runoff
Volume
Difference
(%)
61%
0.03
Inflow (m 3 /s)
0.025
0.02
0.015
0.01
0.005
0
4/17/2006 4/17/2006 4/17/2006 4/17/2006 4/17/2006 4/17/2006 4/17/2006 4/17/2006
15:36
16:04
16:33
17:02
17:31
18:00
18:28
18:57
Date
Inflow
Figure A-3. Site L Inflow Hydrograph April 17, 2006 Storm, Rainfall Not Available.
0.06
Inflow (m 3 /s)
0.05
0.04
0.03
0.02
0.01
0
4/17/2006 4/17/2006 4/17/2006 4/17/2006 4/17/2006 4/17/2006 4/17/2006 4/17/2006
15:36
16:04
16:33
17:02
17:31
18:00
18:28
18:57
Date
Inflow
146
Table A-5. April 27, 2006 Storm Summary.
Inflow (m 3/s)
L
M
Rainfall
Amount
(mm)
26.4
26.4
Rainfall
Duration
(hr)
N/A
N/A
Peak
Intensity
(mm/hr)
N/A
N/A
Peak
Flow
(m3/s)
0.012
0.043
Total
Runoff
Volume
(m3/s)
56.4
189.1
Total
Runoff
Volume
Bypass
(m3/s)
0.0
0.0
Site M vs.
L Peak
Flow
Difference
(%)
71%
Site M vs.
L Runoff
Volume
Difference
(%)
70%
0.005
0.0045
0.004
0.0035
0.003
0.0025
0.002
0.0015
0.001
0.0005
0
4/27/2006 4:48
4/27/2006 9:36
4/27/2006 14:24
4/27/2006 19:12
4/28/2006 0:00
Date
Inlow
3
Inflow (m /s)
Figure A-5. Site L Inflow Hydrograph April 27, 2006 Storm, Rainfall Not Available.
0.02
0.018
0.016
0.014
0.012
0.01
0.008
0.006
0.004
0.002
0
4/27/200 4/27/200 4/27/200 4/27/200 4/27/200 4/27/200 4/27/200 4/27/200 4/27/200
6 2:24
6 4:48
6 7:12
6 9:36
6 12:00 6 14:24 6 16:48 6 19:12 6 21:36
Date
Inflow
147
Table A-6. May7, 2006 Storm Summary.
Rainfall
Duration
(hr)
2.0
2.0
Peak
Intensity
(mm/hr)
78.7
78.7
Peak
Flow
(m3/s)
0.011
0.028
Total
Runoff
Volume
Bypass
(m3/s)
0.0
0.0
Site M vs.
L Peak
Flow
Difference
(%)
62%
Site M vs.
L Runoff
Volume
Difference
(%)
67%
0.012
1.4
0.01
1.2
1
0.008
0.8
0.006
0.6
0.004
0.4
0.002
0.2
0
Inflow (m 3/s)
L
M
Rainfall
Amount
(mm)
13.0
13.0
Total
Runoff
Volume
(m3/s)
12.8
38.5
0
5/7/2006 5/7/2006 5/7/2006 5/7/2006 5/7/2006 5/7/2006 5/7/2006 5/8/2006 5/8/2006
19:12
19:55
20:38
21:21
22:04
22:48
23:31
0:14
0:57
Date
0.03
1.4
0.025
1.2
1
0.02
0.8
0.015
0.6
0.01
0.4
0.005
0
5/7/2006
19:12
0.2
0
5/7/2006
19:55
5/7/2006
20:38
5/7/2006
21:21
5/7/2006
22:04
5/7/2006
22:48
5/7/2006
23:31
5/8/2006
0:14
5/8/2006
0:57
Date
Inflow
Rainfall
Inflow (m3/s)
148
Table A-7. May 14, 2006 Storm Summary.
Rainfall
Duration
(hr)
3.3
3.3
Peak
Intensity
(mm/hr)
12.4
12.4
Peak
Flow
(m3/s)
0.006
0.013
Total
Runoff
Volume
Bypass
(m3/s)
0.0
0.0
Site M vs.
L Peak
Flow
Difference
(%)
56%
0.007
2.5
0.006
Inflow (m 3/s)
Site M vs.
L Runoff
Volume
Difference
(%)
57%
2
0.005
0.004
1.5
0.003
1
0.002
0.5
0.001
0
5/14/06
16:48
5/14/06
18:00
5/14/06
19:12
5/14/06
20:24
5/14/06
21:36
5/14/06
22:48
5/15/06
0:00
5/15/06
1:12
Rainfal Amount (cm)
L
M
Rainfall
Amount
(mm)
20.6
20.6
Total
Runoff
Volume
(m3/s)
19.3
45.2
0
5/15/06
2:24
Time
Inflow
Rainfall
Inflow (m 3/s)
0.012
2
0.01
0.008
1.5
0.006
1
0.004
0.5
0.002
0
5/14/06
16:48
2.5
0.014
0
5/14/06
18:00
5/14/06
19:12
5/14/06
20:24
5/14/06
21:36
5/14/06
22:48
5/15/06
0:00
5/15/06
1:12
5/15/06
2:24
Time
Inflow
Rainfall
149
Inflow (cfs)
L
M
Rainfall
Duration
(hr)
0.3
0.3
Peak
Intensity
(mm/hr)
14.3
14.3
Peak
Flow
(m3/s)
0.005
0.010
Total
Runoff
Volume
Bypass
(m3/s)
0.0
0.0
Site M vs.
L Peak
Flow
Difference
(%)
52%
0.005
0.0045
0.004
0.0035
0.003
0.0025
0.002
0.0015
0.001
0.0005
0
5/15/06 21:36 5/15/06 22:04 5/15/06 22:33 5/15/06 23:02 5/15/06 23:31
Site M vs.
L Runoff
Volume
Difference
(%)
45%
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
Rainfall
Amount
(mm)
3.8
3.8
Total
Runoff
Volume
(m3/s)
7.4
13.3
0.05
0
5/16/06 0:00
Time
Inflow
Rainfall
0.45
0.4
0.35
0.012
inflow (m3/s)
0.01
0.3
0.25
0.2
0.15
0.008
0.006
0.004
0.1
0.05
0
0.002
0
5/15/06 21:36 5/15/06 22:04 5/15/06 22:33 5/15/06 23:02 5/15/06 23:31
5/16/06 0:00
Date
Inflow
Rainfall
150
L
M
Rainfall
Duration
(hr)
19.1
19.1
Peak
Intensity
(mm/hr)
5.3
5.3
Peak
Flow
(m3/s)
0.004
0.017
Total
Runoff
Volume
Bypass
(m3/s)
0.0
0.0
Site M vs.
L Peak
Flow
Difference
(%)
77%
Site M vs.
L Runoff
Volume
Difference
(%)
55%
0.0045
2.5
Flowrate (m 3/s)
0.004
2
0.0035
0.003
1.5
0.0025
0.002
1
0.0015
0.001
0.5
0.0005
0
5/20/2006 4:48 5/20/2006 9:36
Rainfall
Amount
(mm)
23.4
23.4
Total
Runoff
Volume
(m3/s)
30.2
67.5
0
5/20/2006
14:24
5/20/2006
19:12
Tim e
Inflow
5/21/2006 0:00 5/21/2006 4:48
Rainfall
2.5
0.016
2
3
Flowrate (m /s)
0.014
0.012
1.5
0.01
0.008
1
0.006
0.004
0.5
0.018
0.002
0
5/20/2006
4:48
0
5/20/2006
9:36
5/20/2006
14:24
5/20/2006
19:12
Time
Inflow
5/21/2006
0:00
5/21/2006
4:48
Rainfall
151
Table A-10. June 5, 2006 Storm Summary.
Rainfall
Duration
(hr)
5.9
5.9
Peak
Intensity
mm/hr
41.1
41.1
Peak
Flow
(m3/s)
0.011
0.030
Total
Runoff
Volume
Bypass
(m3/s)
0.0
0.0
Site M vs.
L Peak
Flow
Difference
(%)
62%
Site M vs.
L Runoff
Volume
Difference
(%)
69%
0.014
3.5
0.012
3
0.01
2.5
0.008
2
0.006
1.5
0.004
1
0.002
0.5
0
6/5/2006
1:55
Inflow (m 3/s)
L
M
Rainfall
Amount
(mm)
31.2
31.2
Total
Runoff
Volume
(m3/s)
32.7
106.6
0
6/5/2006
3:21
6/5/2006
4:48
6/5/2006
6:14
6/5/2006
7:40
6/5/2006
9:07
6/5/2006
10:33
6/5/2006
12:00
Date
Inflow
Rainfall
3.5
Inflow (m 3/s)
0.03
3
0.025
2.5
0.02
2
0.015
1.5
0.01
1
0.005
0
6/5/2006
1:55
0.5
0
6/5/2006
3:21
6/5/2006
4:48
6/5/2006
6:14
6/5/2006
7:40
6/5/2006
9:07
6/5/2006
10:33
6/5/2006
12:00
Date
Inflow
Rainfall
0.035
152
3
Inflow (m /s)
L
M
Rainfall
Duration
(hr)
11.6
11.6
0.0018
0.0016
0.0014
0.0012
0.001
0.0008
0.0006
0.0004
0.0002
0
6/12/06
16:48
Peak
Intensity
(mm/hr)
1.5
1.5
6/12/06
19:12
Peak
Flow
(m3/s)
0.002
0.002
6/12/06
21:36
6/13/06
0:00
Total
Runoff
Volume
Bypass
(m3/s)
0.0
0.0
6/13/06
2:24
6/13/06
4:48
Site M vs.
L Peak
Flow
Difference
(%)
34%
6/13/06
7:12
Site M vs.
L Runoff
Volume
Difference
(%)
58%
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
6/13/06
9:36
Rainfall
Amount
(mm)
7.9
7.9
Total
Runoff
Volume
(m3/s)
15.8
37.7
Date
Inflow
Rainfall
0.9
0.8
0.0025
0.7
0.002
0.6
0.5
0.0015
0.4
0.001
0.3
0.2
0.0005
0.1
0
0
6/12/06 6/12/06 6/12/06 6/13/06 6/13/06 6/13/06 6/13/06 6/13/06
16:48
19:12
21:36
0:00
2:24
4:48
7:12
9:36
3
Inflow (m /s)
0.003
Date
Inflow
Rainfall
153
Inflow (m 3/s)
L
M
Rainfall
Duration
(hr)
4.0
4.0
Peak
Intensity
(mm/hr)
27.9
27.9
Peak
Flow
(m3/s)
0.016
0.053
Total
Runoff
Volume
Bypass
(m3/s)
0.0
2.1
Site M vs.
L Peak
Flow
Difference
(%)
70%
0.018
0.016
0.014
0.012
0.01
0.008
0.006
0.004
0.002
0
6/14/06
2:24
Site M vs.
L Runoff
Volume
Difference
(%)
73%
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
6/14/06
3:36
6/14/06
4:48
6/14/06
6:00
6/14/06
7:12
6/14/06
8:24
6/14/06
9:36
6/14/06
10:48
Rainfall
Amount
(mm)
17.0
17.0
Total
Runoff
Volume
(m3/s)
20.5
72.7
6/14/06
12:00
Date
Inflow
Rainfall
0.8
0.75
0.7
0.65
0.6
0.55
0.5
0.45
0.4
0.35
6/14/06
2:24
6/14/06
3:36
6/14/06
4:48
6/14/06
6:00
6/14/06
7:12
6/14/06
8:24
6/14/06
9:36
6/14/06
10:48
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
6/14/06
12:00
3
Inflow (m /s)
Date
Inflow
Rainfall Amount
154
Rainfall
Duration
(hr)
2.4
2.4
Peak
Intensity
mm/hr
73.2
73.2
Peak
Flow
(m3/s)
0.008
0.023
Total
Runoff
Volume
Bypass
(m3/s)
0.0
0.0
Site M vs.
L Peak
Flow
Difference
(%)
65%
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.009
0.008
0.007
0.006
0.005
0.004
0.003
0.002
0.001
0
6/25/06
12:00
Site M vs.
L Runoff
Volume
Difference
(%)
68%
6/25/06
12:43
6/25/06
13:26
6/25/06
14:09
6/25/06
14:52
6/25/06
15:36
Inflow (m 3/s)
L
M
Rainfall
Amount
(mm)
8.4
8.4
Total
Runoff
Volume
(m3/s)
7.6
23.6
6/25/06
16:19
Time
Inflow
Rainfall
Inflow (m3/s)
0.02
0.015
0.01
0.005
0
6/25/06 12:00
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
6/25/06 16:48
0.025
6/25/06 13:12
6/25/06 14:24
6/25/06 15:36
Time
Flow
Level
.
Figure A-22. Site M Inflow Hydrograph and Rainfall Amount for June 25, 2006 Storm
155
L
M
Rainfall
Duration
(hr)
3.5
3.5
Peak
Intensity
(mm/hr)
38.1
38.1
Peak
Flow
(m3/s)
0.005
0.014
Site M vs.
L Peak
Flow
Difference
(%)
61%
Site M vs.
L Runoff
Volume
Difference
(%)
55%
0.8
0.006
0.004
0.6
0.5
3
Inflow (m /s)
0.7
0.003
0.4
0.3
0.001
0.2
0.1
0.000
6/26/06
20:24
6/26/06
21:36
6/26/06
22:48
6/27/06
0:00
6/27/06
1:12
6/27/06
2:24
6/27/06
3:36
Rainfall
Amount
(mm)
6.6
6.6
Total
Runoff
Volume
Bypass
(m3/s)
0.0
0.0
Total
Runoff
Volume
(m3/s)
7.8
17.4
0
6/27/06
4:48
Time
Inflow
Rainfall
0.016
0.014
0.012
0.01
0.008
0.006
0.004
0.002
0
6/26/06
20:24
6/26/06
21:36
6/26/06
22:48
6/27/06
0:00
6/27/06
1:12
6/27/06
2:24
6/27/06
3:36
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
6/27/06
4:48
Inflow (m 3/s)
Time
Inflow
Rainfall
156
Inflow (m 3/s)
L
M
Rainfall
Duration
(hr)
6.1
6.1
Peak
Intensity
(mm/hr)
4.1
4.1
Peak
Flow
(m3/s)
0.003
0.005
Total
Runoff
Volume
Bypass
(m3/s)
0.0
0.0
Site M vs.
L Peak
Flow
Difference
(%)
51%
Site M vs.
L Runoff
Volume
Difference
(%)
48%
0.003
0.6
0.0025
0.5
0.002
0.4
0.0015
0.3
0.001
0.2
0.0005
0.1
0
6/27/06 8:24
Rainfall
Amount
(mm)
5.6
5.6
Total
Runoff
Volume
(m3/s)
9.1
17.3
0
6/27/06 10:48
6/27/06 13:12
6/27/06 15:36
6/27/06 18:00
Time
Inflow
Rainfall
0.006
0.6
0.005
0.5
0.004
0.4
0.003
0.3
0.002
0.2
0.001
0.1
0
6/27/06 8:24
Inflow (m3/s)
0
6/27/06 10:48
6/27/06 13:12
6/27/06 15:36
6/27/06 18:00
Time
Inflow
Rainfall
157
Table A-16. July 6, 2006 Storm Summary.
Flowrate (m 3/s)
L
M
Rainfall
Duration
(hr)
4.8
4.8
Peak
Intensity
(mm/hr)
27.9
27.9
Peak
Flow
(m3/s)
0.003
0.019
Site M vs.
L Peak
Flow
Difference
(%)
84%
Site M vs.
L Runoff
Volume
Difference
(%)
76%
0.0035
1.4
0.003
1.2
0.0025
1
0.002
0.8
0.0015
0.6
0.001
0.4
0.0005
0.2
0
7/6/2006
12:00
7/6/2006
13:12
7/6/2006
14:24
7/6/2006
15:36
7/6/2006
16:48
7/6/2006
18:00
7/6/2006
19:12
Rainfall
Amount
(mm)
11.7
11.7
Total
Runoff
Volume
Bypass
(m3/s)
0.0
0.0
Total
Runoff
Volume
(m3/s)
7.2
29.4
0
7/6/2006
20:24
Tim e
Inflow
Rainfall
0.025
1.2
Inflow (m 3/s)
0.02
1
0.015
0.8
0.01
0.6
0.4
0.005
0
7/6/2006
12:00
0.2
7/6/2006
13:12
7/6/2006
14:24
7/6/2006
15:36
7/6/2006
16:48
7/6/2006
18:00
7/6/2006
19:12
1.4
0
7/6/2006
20:24
Date
Inflow
Rainfall
Figure A-28. Site M Inflow Hydrograph and Rainfall Amount for July 6, 2006 Storm.\
158
Rainfall
Duration
Peak
Intensity
Peak
Flow
(mm)
4.6
4.6
(hr)
1.8
1.8
(mm/hr)
13.5
13.5
(m3/s)
0.006
0.015
(m3/s)
5.1
13.5
Total
Runoff
Volume
Bypass
Site M vs.
L Peak
Flow
Difference
Site M vs.
L Runoff
Volume
Difference
(m3/s)
0.0
0.0
(%)
64%
(%)
62%
0.006
0.6
0.005
0.5
0.004
0.4
0.003
0.3
0.002
0.2
0.001
0.1
0
7/16/2006
18:00
Flowrate (m 3/s)
L
M
Rainfall
Amount
Total
Runoff
Volume
0
7/16/2006
18:28
7/16/2006
18:57
7/16/2006
19:26
7/16/2006
19:55
7/16/2006
20:24
7/16/2006
20:52
7/16/2006
21:21
Tim e
Inflow
Rainfall
0.018
0.016
0.014
0.012
0.01
0.008
0.006
0.004
0.002
0
7/16/2006 7/16/2006 7/16/2006 7/16/2006 7/16/2006 7/16/2006 7/16/2006 7/16/2006
18:00
18:28
18:57
19:26
19:55
20:24
20:52
21:21
0.6
0.5
0.4
0.3
0.2
0.1
3
Flowrate (m /s)
0
Time
Inflow
Rainfall
159
Flowrate (m 3/s)
L
M
Rainfall
Duration
(hr)
24.3
24.3
0.02
0.018
0.016
0.014
0.012
0.01
0.008
0.006
0.004
0.002
0
7/23/2006
19:12
Peak
Intensity
(mm/hr)
88.9
88.9
Peak
Flow
(m3/s)
0.017
0.059
Total
Runoff
Volume
Bypass
(m3/s)
0.0
6.3
Site M vs.
L Peak
Flow
Difference
(%)
71%
Site M vs.
L Runoff
Volume
Difference
(%)
78%
4
3.5
3
2.5
2
1.5
1
0.5
7/23/2006
20:24
7/23/2006
21:36
7/23/2006
22:48
Date
Inflow
7/24/2006
0:00
7/24/2006
1:12
Rainfall
Amount
(mm)
40.1
40.1
Total
Runoff
Volume
(m3/s)
39.2
175.8
0
7/24/2006
2:24
Rainfall
Flowrate (m 3/s)
0.07
4
0.06
3
0.05
0.04
2
0.03
0.02
1
0.01
0
0
7/23/2006 7/23/2006 7/23/2006 7/23/2006 7/24/2006 7/24/2006 7/24/2006
19:12
20:24
21:36
22:48
0:00
1:12
2:24
Date
Inflow
Rainfall
160
L
M
Rainfall
Duration
(hr)
23.3
23.3
Peak
Intensity
(mm/hr)
27.9
27.9
Peak
Flow
(m3/s)
0.019
0.062
Site M vs.
L Peak
Flow
Difference
(%)
69%
Site M vs.
L Runoff
Volume
Difference
(%)
78%
Flowrate (m 3/s)
0.025
3.5
3
0.02
2.5
0.015
2
1.5
0.01
1
0.005
0
7/24/2006
19:12
0.5
Rainfall
Amount
(mm)
29.0
29.0
Total
Runoff
Volume
Bypass
(m3/s)
0.0
4.8
Total
Runoff
Volume
(m3/s)
36.5
163.0
0
7/25/2006
0:00
7/25/2006
4:48
7/25/2006
9:36
7/25/2006
14:24
7/25/2006
19:12
7/26/2006
0:00
7/26/2006
4:48
Date
Inflow
Rainfall
0.07
3.5
0.06
3
0.05
2.5
0.04
2
0.03
1.5
0.02
1
0.01
0.5
0
38922.8
Flowrate (m 3/s)
0
38923
38923.2
38923.4
38923.6
38923.8
38924
38924.2
Date
Inflow
Rainfall
161
L
M
Rainfall
Duration
(hr)
8.2
8.2
Peak
Intensity
(mm/hr)
1.3
1.3
Peak
Flow
(m3/s)
0.001
0.002
Total
Runoff
Volume
Bypass
(m3/s)
0.0
0.0
Site M vs.
L Peak
Flow
Difference
(%)
55%
0.0008
Flowrate (m 3/s)
0.0007
0.0006
0.0005
0.0004
0.0003
0.0002
0.0001
0
7/29/2006
21:36
7/30/2006
0:00
7/30/2006
7/30/2006
2:24
Date 4:48
Inflow
7/30/2006
7:12
7/30/2006
9:36
Site M vs.
L Runoff
Volume
Difference
(%)
51%
0.5
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
7/30/2006
12:00
Rainfall
Amount
(mm)
4.1
4.1
Total
Runoff
Volume
(m3/s)
9.4
19.2
Rainfall
0.002
0.5
0.4
0.0015
0.3
0.001
0.2
0.0005
0.1
0
0
7/29/2006 7/30/2006 7/30/2006 7/30/2006 7/30/2006 7/30/2006 7/30/2006
21:36
0:00
2:24
4:48
7:12
9:36
12:00
Flowrate (m 3/s)
Date
Inflow
Rainfall
162
Table A-21. August 21, 2006 Storm Summary.
L
M
Rainfall
Duration
(hr)
0.7
0.7
Peak
Intensity
(mm/hr)
15.6
15.6
Peak
Flow
(m3/s)
0.004
0.015
Total
Runoff
Volume
Bypass
(m3/s)
0.0
0.0
Site M vs.
L Peak
Flow
Difference
(%)
73%
Site M vs.
L Runoff
Volume
Difference
(%)
74%
Flowrate (m 3/s)
0.005
1.2
1
0.004
0.8
0.003
0.6
0.002
0.4
0.001
0.2
0
0
38950.7 38950.7 38950.7 38950.7 38950.7 38950.8 38950.8 38950.8
2
4
6 Date 8
2
4
Inflow
Rainfall
Amount
(mm)
10.7
10.7
Total
Runoff
Volume
(m3/s)
5.9
22.6
Rainfall
0.018
0.016
0.014
0.012
0.01
0.008
0.006
0.004
0.002
0
8/21/2006
16:48
1.2
1
0.8
0.6
0.4
0.2
8/21/2006
17:16
8/21/2006
17:45
8/21/2006
18:14
8/21/2006
18:43
8/21/2006
19:12
8/21/2006
19:40
0
8/21/2006
20:09
Date
Inflow
Rainfall
Storm.
3
Flowrate (m /s)
Storm.
163
Table A-22. August 23, 2006 Storm Summary.
Flowrate (m 3/s)
L
M
Rainfall
Duration
(hr)
6.4
6.4
Peak
Intensity
(mm/hr)
52.1
52.1
Peak
Flow
(m3/s)
0.014
0.055
Site M vs.
L Peak
Flow
Difference
(%)
74%
Site M vs.
L Runoff
Volume
Difference
(%)
76%
0.016
4
0.014
3.5
0.012
3
0.01
2.5
0.008
2
0.006
1.5
0.004
1
0.002
0.5
0
8/22/2006
9:36
8/22/2006
12:00
8/22/2006
14:24
8/22/2006
16:48
8/22/2006
19:12
8/22/2006
21:36
8/23/2006
0:00
Rainfall
Amount
(mm)
34.5
34.5
Total
Runoff
Volume
Bypass
(m3/s)
0.0
2.6
Total
Runoff
Volume
(m3/s)
21.9
87.9
0
8/23/2006
2:24
Date
Inflow
Rainfall
Storm.
0.06
3.5
3
Flowrate (m /s)
0.05
3
0.04
2.5
0.03
2
1.5
0.02
1
0.01
0
8/22/2006
9:36
0.5
8/22/2006
12:00
8/22/2006
14:24
8/22/2006
16:48
8/22/2006
19:12
8/22/2006
21:36
8/23/2006
0:00
0
8/23/2006
2:24
Date
Inflow
Rainfall
Storm.
4
164
Table A-23. September 1, 2006 (Tropical Storm Ernesto) Storm Summary.
Flowrate (m 3/s)
L
M
Rainfall
Duration
(hr)
21.8
21.8
Peak
Intensity
(mm/hr)
22.9
22.9
Peak
Flow
(m3/s)
0.012
0.047
Total
Runoff
Volume
Bypass
(m3/s)
0.0
0.0
Site M vs.
L Peak
Flow
Difference
(%)
74%
Site M vs.
L Runoff
Volume
Difference
(%)
78%
12
10
8
0.01
6
4
0.005
2
0
0
8/31/200 8/31/200 8/31/200 8/31/200 8/31/200 9/1/2006 9/1/2006 9/1/2006
6 0:00 6 4:48 6 9:36 6 14:24 6 19:12
0:00
4:48
9:36
0.015
Date
Inflow
Rainfall
12
0.05
10
0.04
8
0.03
6
0.02
4
0.01
2
0
0
8/31/20 8/31/20 8/31/20 8/31/20 8/31/20 9/1/200 9/1/200 9/1/200
6 0:00 6 4:48 6 9:36
06
06 0:00 06 4:48 06 9:36
06
19:12
14:24
Flowrate (m 3/s)
(Tropical Storm Ernesto) Storm.
Date
Inflow
Rainfall
(Tropical Storm Ernesto) Storm.
Rainfall
Amount
(mm)
105.2
105.2
Total
Runoff
Volume
(m3/s)
120.2
556.7
165
Table A-24. September 5, 2006 Storm Summary.
Rainfall
Duration
(hr)
13.1
13.1
Peak
Intensity
(mm/hr)
9.7
9.7
Peak
Flow
(m3/s)
0.003
0.009
Total
Runoff
Volume
Bypass
(m3/s)
0.0
0.0
Site M vs.
L Peak
Flow
Difference
(%)
68%
Site M vs.
L Runoff
Volume
Difference
(%)
76%
0.0035
1
0.003
0.8
0.0025
0.6
0.002
0.0015
0.4
0.001
0.2
0.0005
0
0
9/5/2006 9/5/2006 9/6/2006 9/6/2006 9/6/2006 9/6/2006 9/6/2006 9/6/2006
19:12
21:36
0:00
2:24
4:48
7:12
9:36
12:00
Flowrate (m 3/s)
L
M
Rainfall
Amount
(mm)
8.6
8.6
Total
Runoff
Volume
(m3/s)
12.8
52.3
Date
Inflow
Rainfall
0.01
1
0.008
0.8
0.006
0.6
0.004
0.4
0.002
0.2
0
0
9/5/2006 9/5/2006 9/6/2006 9/6/2006 9/6/2006 9/6/2006 9/6/2006 9/6/2006
19:12
21:36
0:00
2:24
4:48
7:12
9:36
12:00
Date
Inflow
Rainfall
Figure A-44. Site M Inflow Hydrograph and Rainfall Amount for September 5 2006
Storm.
Flowrate (m 3/s)
Storm.
166
Table A-25. September 14, 2006 Storm Summary.
Rainfall
Duration
(hr)
10.8
10.8
Peak
Intensity
(mm/hr)
3.6
3.6
Peak
Flow
(m3/s)
0.013
0.054
Total
Runoff
Volume
Bypass
(m3/s)
0.0
0.0
Site M vs.
L Peak
Flow
Difference
(%)
75%
Site M vs.
L Runoff
Volume
Difference
(%)
76%
0.016
5.5
Inflow rate (m 3/s)
0.014
4.5
0.012
3.5
0.01
0.008
2.5
0.006
1.5
0.004
0.5
0.002
0
9/14/2006 0:00
L
M
Rainfall
Amount
(mm)
49.8
49.8
Total
Runoff
Volume
(m3/s)
52.2
217.0
-0.5
9/14/2006 3:36
9/14/2006 7:12
9/14/2006 10:48
Date
Inflow
Rainfall
0.06
5.5
0.05
4.5
0.04
3.5
0.03
2.5
0.02
1.5
0.01
0.5
0
9/14/2006 0:00
Inflow rate (m 3/s)
Storm.
-0.5
9/14/2006 3:36
9/14/2006 7:12
9/14/2006 10:48
Date
Inflow
Rainfall
Storm.
167
Table A-26. October 8, 2006 Storm Summary.
Rainfall
Duration
(hr)
15.3
15.3
Peak
Intensity
(mm/hr)
88.9
88.9
Peak
Flow
(m3/s)
0.039
0.114
Total
Runoff
Volume
Bypass
(m3/s)
0.0
83.3
Site M vs.
L Peak
Flow
Difference
(%)
66%
Site M vs.
L Runoff
Volume
Difference
(%)
76%
0.045
0.04
0.035
0.03
0.025
0.02
0.015
0.01
0.005
0
10/8/06
12:00
9
8
7
6
5
4
3
2
1
0
10/8/06
14:24
10/8/06
16:48
10/8/06
19:12
10/8/06
21:36
10/9/06
0:00
10/9/06
2:24
10/9/06
4:48
Inflow rate (m 3/s)
L
M
Rainfall
Amount
(mm)
76.2
76.2
Total
Runoff
Volume
(m3/s)
87.2
280.2
10/9/06
7:12
Date
Inflow
Rainfall
Inflow rate (m 3/s)
0.14
9
8
7
6
5
4
3
2
1
0
0.12
0.1
0.08
0.06
0.04
0.02
0
10/8/06
12:00
10/8/06
14:24
10/8/06
16:48
10/8/06
19:12
10/8/06
21:36
10/9/06
0:00
10/9/06
2:24
10/9/06
4:48
10/9/06
7:12
Date
Inflow
Rainfall
Storm.
Storm.
168
Table A-27. October 18, 2006 Storm Summary.
Rainfall
Duration
(hr)
18.9
18.9
Peak
Intensity
(mm/hr)
4.3
4.3
Peak
Flow
(m3/s)
0.002
0.004
0.0025
Site M vs.
L Peak
Flow
Difference
(%)
41%
Site M vs.
L Runoff
Volume
Difference
(%)
35%
0.7
0.6
0.002
Inflow (m 3/s)
Total
Runoff
Volume
Bypass
(m3/s)
0.0
0.0
0.5
0.0015
0.4
0.001
0.3
0.2
0.0005
0.1
0
L
M
Rainfall
Amount
(mm)
6.6
6.6
Total
Runoff
Volume
(m3/s)
14.2
21.9
0
10/17/2006 10/17/2006 10/17/2006 10/17/2006 10/18/2006 10/18/2006 10/18/2006
7:12
12:00
16:48
21:36
2:24
7:12
12:00
Date
0.004
0.7
0.0035
0.6
Inflow (m 3/s)
0.003
0.5
0.0025
0.4
0.002
0.3
0.0015
0.001
`
0.2
0.1
0.0005
0
0
10/17/2006 10/17/2006 10/17/2006 10/17/2006 10/18/2006 10/18/2006 10/18/2006
7:12
12:00
16:48
21:36
2:24
7:12
12:00
Date
Storm.
Storm.
169
B.0 Appendix B-Field Study Hydrology Statistics
B.1 Flow Mitigation-Volume
¾ SAS Data Input
Storm
Date
03/21/06
04/16/06
04/26/06
05/07/06
05/14/06
05/15/06
05/20/06
06/05/06
06/12/06
06/14/06
06/25/06
06/26/06
06/27/06
07/06/06
07/16/06
07/23/06
07/25/06
07/30/06
08/21/06
08/22/06
09/01/06
09/06/06
09/13/06
10/08/06
10/17/06
Site L
captured
Site L
overflow
Site M
captured
Site M
overflow
(m3)
10.33
17.34
56.44
12.83
19.35
7.38
30.23
32.67
15.82
20.50
7.63
7.85
9.10
7.20
5.09
39.19
36.53
9.35
5.89
21.90
120.18
12.82
52.24
87.15
14.15
(m3)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
(m3)
22.943
44.291
189.149
38.520
45.215
13.326
67.547
106.611
37.745
72.667
23.557
17.430
17.329
29.408
13.496
175.794
163.047
19.188
22.586
87.850
556.712
52.328
217.021
280.194
21.891
(m3)
0
0
0
0
0
0
0
0
0
2.1
0
0
0
0
0
6.3
4.8
0
0
2.6
0
0
0
83.3
0
¾ SAS Analysis
The data in not normally distributed, thus a non-parametric analysis was performed.
Since the data being compared are numerically very different (0 verse 30-550), univariate
test was performed. The volume treated was statistically different from the volume
bypassed for both Site L and Site M (p<0.0001).
170
---------------------------------------------- site=L --------------------------------------------The UNIVARIATE Procedure
Variable: scordat
Moments
N
Mean
Std Deviation
Skewness
Uncorrected SS
Coeff Variation
25
26.3664
27.6456359
2.2070484
35722.4246
104.851765
Sum Weights
Sum Observations
Variance
Kurtosis
Corrected SS
Std Error Mean
25
659.16
764.281182
5.16639951
18342.7484
5.52912717
Basic Statistical Measures
Location
Mean
Median
Mode
Variability
26.36640
15.82000
.
Std Deviation
Variance
Range
Interquartile Range
27.64564
764.28118
115.09000
23.57000
Tests for Location: Mu0=0
Test
-Statistic-
-----p Value------
Student's t
Sign
Signed Rank
t
M
S
Pr > |t|
Pr >= |M|
Pr >= |S|
4.768637
12.5
162.5
Quantiles (Definition 5)
Quantile
100% Max
99%
95%
90%
75% Q3
50% Median
25% Q1
10%
5%
1%
0% Min
Estimate
120.18
120.18
87.15
56.44
32.67
15.82
9.10
7.20
5.89
5.09
5.09
<.0001
<.0001
<.0001
171
---------------------------------------------- site=M ----------------------------------The UNIVARIATE Procedure
Variable: scordat
Moments
N
Mean
Std Deviation
Skewness
Uncorrected SS
Coeff Variation
25
89.4698
116.782597
2.93612618
527437.326
130.527392
Sum Weights
Sum Observations
Variance
Kurtosis
Corrected SS
Std Error Mean
25
2236.745
13638.1749
10.5368611
327316.199
23.3565194
Location
Mean
Median
Mode
Variability
89.46980
44.29100
.
Std Deviation
Variance
Range
Interquartile Range
116.78260
13638
543.38600
84.02500
Test
-Statistic-
-----p Value------
Student's t
Sign
Signed Rank
t
M
S
Pr > |t|
Pr >= |M|
Pr >= |S|
3.830614
12.5
162.5
Quantile
100% Max
99%
95%
90%
75% Q3
50% Median
25% Q1
10%
5%
1%
0% Min
Estimate
556.712
556.712
217.021
196.894
106.611
44.291
22.586
17.329
13.496
13.326
13.326
0.0008
<.0001
<.0001
172
B.2 Flow Mitigation-Peak Flow Rate
¾ SAS Data Input
Storm
Date
03/21/06
04/16/06
04/26/06
05/07/06
05/14/06
05/15/06
05/20/06
06/05/06
06/12/06
06/14/06
06/25/06
06/26/06
06/27/06
07/06/06
07/16/06
07/23/06
07/25/06
07/30/06
08/21/06
08/22/06
09/01/06
09/06/06
09/13/06
10/08/06
10/17/06
Site L Peak
Flow
captured
Site L Peak
Flow
bypassed
Site M Peak
Flow
captured
Site M
Peak Flow
bypassed
(m3 /s)
0.0022
0.0256
0.0124
0.0107
0.0058
0.0046
0.0040
0.0113
0.0016
0.0158
0.0082
0.0053
0.0025
0.0030
0.0055
0.0174
0.0191
0.0007
0.0042
0.0142
0.0124
0.0030
0.0135
0.0391
0.0022
(m3/s)
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
(m3/s)
0.0025
0.0475
0.0432
0.0280
0.0130
0.0096
0.0170
0.0301
0.0025
0.0530
0.0234
0.0135
0.0051
0.0192
0.0155
0.0593
0.0622
0.0015
0.0153
0.0554
0.0472
0.0094
0.0535
0.1800
0.0037
(m3/s)
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0093
0.0000
0.0000
0.0000
0.0000
0.0000
0.0099
0.0346
0.0000
0.0000
0.0151
0.0000
0.0000
0.0000
0.1561
0.0000
¾ SAS Analysis
The same univariate test was performed as mentioned in Section B.2. The peak
rate of inflow was statistically different from the peak rate that bypassed in both Site
L and Site M (p<0.0001).
173
---------------------------------------------- site=L ----------------------------------The UNIVARIATE Procedure
Variable: scordat
Moments
N
Mean
Std Deviation
Skewness
Uncorrected SS
Coeff Variation
25
0.009772
0.00889445
1.73404163
0.00428597
91.0197652
Sum Weights
Sum Observations
Variance
Kurtosis
Corrected SS
Std Error Mean
25
0.2443
0.00007911
3.80413241
0.00189867
0.00177889
Location
Mean
Median
Mode
Variability
0.009772
0.005800
0.002200
Std Deviation
Variance
Range
Interquartile Range
0.00889
0.0000791
0.03840
0.01050
NOTE: The mode displayed is the smallest of 3 modes with a count of 2.
Test
-Statistic-
-----p Value------
Student's t
Sign
Signed Rank
t
M
S
Pr > |t|
Pr >= |M|
Pr >= |S|
5.493312
12.5
162.5
Quantile
100% Max
99%
95%
90%
75% Q3
50% Median
25% Q1
10%
5%
1%
0% Min
Estimate
0.0391
0.0391
0.0256
0.0191
0.0135
0.0058
0.0030
0.0022
0.0016
0.0007
0.0007
<.0001
<.0001
<.0001
174
---------------------------------------------- site=M ----------------------------------The UNIVARIATE Procedure
Variable: scordat
Moments
N
Mean
Std Deviation
Skewness
Uncorrected SS
Coeff Variation
25
0.023424
0.0167809
0.40159763
0.02047546
71.6397663
Sum Weights
Sum Observations
Variance
Kurtosis
Corrected SS
Std Error Mean
25
0.5856
0.0002816
-1.1738085
0.00675837
0.00335618
Location
Mean
Median
Mode
Variability
0.023424
0.019200
0.002500
Std Deviation
Variance
Range
Interquartile Range
0.01678
0.0002816
0.05200
0.03070
Test
-Statistic-
-----p Value------
Student's t
Sign
Signed Rank
t
M
S
Pr > |t|
Pr >= |M|
Pr >= |S|
6.979364
12.5
162.5
Quantile
100% Max
99%
95%
90%
75% Q3
50% Median
25% Q1
10%
5%
1%
0% Min
Estimate
0.0535
0.0535
0.0494
0.0475
0.0403
0.0192
0.0096
0.0025
0.0025
0.0015
0.0015
<.0001
<.0001
<.0001
175
B.3 Correlation Between Rainfall Intensity and Bypass Storms
¾ SAS Data Input
Success (0) or
Failure (1)
Success =
Captured
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
Rainfall
Amount
Rainfall
Intensity
(mm)
4.064
7.9
11.938
11.684
5.588
6.604
25.4
8.636
6.604
4.572
3.81
10.668
105.156
31.2
8.382
12.954
76.2
49.784
20.574
17.018
28.956
40.132
48.768
76.2
(mm/hr)
1.27
1.524
2.794
27.94
4.064
4.318
5.334
9.652
38.1
13.462
14.2875
15.61171
22.86
41.148
73.152
78.74
88.9
3.556
12.446
27.94
27.94
88.9
52.07
88.9
¾ SAS Analysis
A logistic test was used to determine if overflow could be predicted based on the
rainfall intensity and amount. A logistic test is a binary test that test for the
probability of success. It was found that there was no significant evidence (p>0.05)
of using a storm’s rainfall amount to predict the probability of bypass, but there was
significant evidence (p<0.05) of using a storm’s rainfall intensity to predict the
probability of bypass.
176
The LOGISTIC Procedure with Amount and Intensity
Analysis of Maximum Likelihood Estimates
Parameter
DF
Estimate
Standard
Error
Wald
Chi-Square
Pr > ChiSq
Intercept
intensity
amount
1
1
1
3.5409
-0.0415
-0.0260
1.2939
0.0218
0.0207
7.4893
3.6280
1.5824
0.0062
0.0568
0.2084
The LOGISTIC Procedure-Intensity Only
Parameter
DF
Estimate
Standard
Error
Wald
Chi-Square
Pr > ChiSq
Intercept
intensity
1
1
2.6865
-0.0343
0.9860
0.0174
7.4241
3.8908
0.0064
0.0486
177
B.4 Correlation Between Peak Inflow Intensity and Bypass
Storms
¾ SAS Data Input
Success
(0) or
Failure (1)
Success =
Captured
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
1
1
0
0
1
0
0
0
0
1
Peak
Runoff
rate
(m3/s)
0.00248
0.047544
0.043231
0.02797
0.013045
0.009595
0.017035
0.030053
0.00248
0.052976
0.023398
0.013514
0.005142
0.019232
0.015475
0.05927
0.06224
0.001538
0.015273
0.055416
0.047191
0.009436
0.053535
0.18
0.00371
¾ SAS Analysis
A logistic test was used to determine if overflow could be predicted peak inflow
rate (See B.4). It was found that there was no significant evidence (p>0.05) of using
a storm’s peak inflow rate to predict the probability of bypass.
178
The LOGISTIC Procedure
.
Parameter
DF
Estimate
Standard
Error
Wald
Chi-Square
Pr > ChiSq
Intercept
intensity
1
1
2.0794
9.4570
1.0607
184.7
3.8436
0.0026
0.0499
0.9592
179
C.0 Appendix C-Field Study Bacteria Statistics
C.1 Inflow/Groundwater Fecal Coliform Concentration
¾ SAS Data Input
3/21/06
4/16/06
4/26/06
5/7/06
5/14/06
5/15/06
5/20/06
6/5/06
6/12/06
6/14/06
6/25/06
6/26/06
6/27/06
7/6/06
7/16/06
7/23/06
7/25/06
7/30/06
8/21/06
8/22/06
9/1/06
9/6/06
9/13/06
10/8/06
10/17/06
Site L
Stormwater
Runoff
Site L
Groundwater
Site M
Stormwater
Runoff
Site M
Groundwater
CFU/100 ml
3800
2300
181
2700
358
570
2000
2900
5800
820
12000
19000
4100
10000
47662
8200
12000
7100
12000
12000
12000
12000
12000
4800
28300
CFU/100 ml
0.5
0.5
0.5
1
0.5
0.5
0.5
1
4
0.5
1
0.5
1
1
0.5
0.5
0.5
2
2
54
4
0.5
4
1
1
CFU/100 ml
2280
17200
19400
3000
760
940
5000
5100
4700
3100
12000
15000
3300
9000
6800
12000
12000
8000
12000
12000
12000
12000
12000
16600
6500
CFU/100 ml
3
3
3
0.5
8
8
2
2
1
1
0.5
0.5
0.5
4
43
18
86
3
66
214
12000
4
18
0.5
37
¾ SAS Analysis
Since the data was slightly skewed, the natural log of the bacteria concentrations
were taken. Proc Mixed was run in SAS, since the data was normalized and
dependant. A significant difference was found (p <0.001) between the runoff fecal
coliform concentration and the groundwater bacteria concentration for both sites.
180
--------------------------------------------- site=L -----------------------------------The Mixed Procedure
Covariance Parameter
Estimates
Cov Parm
Estimate
SP(POW)
Residual
0.9000
2.0941
Null Model Likelihood Ratio Test
DF
Chi-Square
Pr > ChiSq
1
0.00
1.0000
Type 3 Tests of Fixed Effects
Effect
Intercept
Num
DF
Den
DF
F Value
Pr > F
1
24
854.85
<.0001
--------------------------------------------- site=M -----------------------------------The Mixed Procedure
Estimates
Cov Parm
Estimate
SP(POW)
Residual
0.9000
5.2414
DF
Chi-Square
Pr > ChiSq
1
0.00
1.0000
Effect
Intercept
Num
DF
Den
DF
F Value
Pr > F
1
24
235.20
<.0001
181
C.2 Inflow/Groundwater Enterococcus Concentration
¾ SAS Data Input
4/16/2006
4/26/2006
5/7/2006
5/14/2006
5/15/2006
5/20/2006
6/5/2006
6/13/2006
6/14/2006
6/25/2006
6/26/2006
6/27/2006
7/16/2006
7/23/2006
7/25/2006
7/30/2006
8/21/2006
8/22/2006
9/6/2006
9/14/2006
10/8/2006
10/17/2006
Site L
Stormwater
Runoff
CFU/100 ml
344
306
334
1652
945
870
1013
4010
2005
4010
4010
1013
453
2005
4010
10
42
738
4010
1013
1091
4010
Site L
Groundwater
CFU/100 ml
5
5
10
64
64
5
5
5
5
5
5
5
40
5
10
31
5
5
31
42
10
5
Site M
Stormwater
Runoff
CFU/100 ml
4010
2005
4010
1445
4010
334
504
504
1184
4010
1298
478
1298
4010
4010
5
271
1184
4010
4010
4010
4010
Site M
Groundwater
CFU/100 ml
5
5
31
31
31
10
10
64
31
10
20
5
10
429
406
5
10
137
2005
150
124
20
¾ SAS Analysis
dependant. A significant difference was found (p <0.001) between the runoff
enterococcus concentration and the groundwater bacteria concentration for both sites.
182
---------------------------------------------- site=L ----------------------------------The Mixed Procedure
Estimates
Cov Parm
Estimate
SP(POW)
Residual
0.9000
3.6850
DF
Chi-Square
Pr > ChiSq
1
0.00
1.0000
Effect
Intercept
Num
DF
Den
DF
F Value
Pr > F
1
21
122.03
<.0001
---------------------------------------------- site=M ----------------------------------The Mixed Procedure
Estimates
Cov Parm
Estimate
SP(POW)
Residual
0.9000
2.8308
DF
Chi-Square
Pr > ChiSq
1
0.00
1.0000
Effect
Intercept
Num
DF
Den
DF
F Value
Pr > F
1
21
107.22
<.0001
183
C.3 Groundwater Fecal Concentration Before and After DIS
¾ SAS Data Input
Used data entered in C.1 along with table below
L-12
CFU/100ml
110
0.5
23
0.5
22
7/12/2005
7/24/2005
8/10/2005
8/24/2005
9/21/2005
M-12
CFU/100ml
200
1
12
1
0.5
¾ SAS Analysis
dependant. No significant difference was found at Site L or Site M (p >0.05)
between bacteria concentrations in the groundwater before and after the system was
implemented.
---------------------------------------------- site=L ----------------------------------The Mixed Procedure
Convergence criteria met.
Covariance Parameter Estimates
Cov Parm
Subject
SP(POW)
Residual
system
Estimate
0.3092
1.9191
DF
Chi-Square
Pr > ChiSq
1
0.58
0.4464
Effect
Num
DF
Den
DF
F Value
Pr > F
system
1
26.5
2.30
0.1410
184
---------------------------------------------- site=M ----------------------------------The Mixed Procedure
Convergence criteria met.
Covariance Parameter Estimates
Cov Parm
Subject
SP(POW)
Residual
system
Estimate
0.8618
5.3049
DF
Chi-Square
Pr > ChiSq
1
9.35
0.0022
Effect
Num
DF
Den
DF
F Value
Pr > F
system
1
9.62
0.05
0.8330
185
D.0 Appendix D-Laboratory Infiltration Rate Curves
Table D-1. Trial 1 infiltration times for each column
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.47
5.18
11.38
55.87
CDI2
0.40
4.60
9.90
45.03
CDI3
0.48
4.67
10.18
47.33
T1
0.47
4.95
10.65
47.40
Time
(min)
T2
0.37
4.45
9.93
43.75
T3
0.42
4.45
9.78
40.20
CSW1
0.38
4.02
10.25
32.87
CSW2
0.27
3.42
8.93
32.83
CSW3
0.33
3.73
10.82
32.02
T3
0.68
5.48
10.95
44.12
CSW1
0.42
4.23
8.97
36.50
CSW2
0.28
3.08
6.43
31.17
CSW3
0.33
3.90
7.70
33.50
T3
0.75
5.93
10.77
44.62
CSW1
0.55
4.97
10.47
40.02
CSW2
0.35
3.25
7.03
28.33
CSW3
0.25
4.02
10.67
33.75
T3
1.33
8.87
16.53
71.10
CSW1
1.25
8.68
16.33
68.00
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.43
5.10
11.17
53.05
CDI2
0.40
4.55
9.82
42.97
CDI3
0.38
4.45
9.75
46.50
T1
3.00
16.60
29.18
90.00
Time
(min)
T2
0.90
6.38
15.00
47.92
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.37
4.88
9.93
52.12
CDI2
0.37
4.50
9.78
42.73
CDI3
0.42
4.60
10.00
43.73
T1
3.13
34.35
39.50
70.00
Time
(min)
T2
1.45
8.53
16.18
61.25
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.40
5.77
11.75
53.37
CDI2
0.37
5.37
10.30
44.80
CDI3
0.40
5.37
10.52
44.38
T1
2.35
14.95
27.48
142.00
Time
(min)
T2
2.27
12.30
22.13
108.00
CSW2
0.35
4.40
7.35
29.73
CSW3
0.87
5.12
10.28
45.15
186
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.40
4.77
10.82
52.67
CDI2
0.37
4.57
9.75
43.47
CDI3
0.42
4.40
9.63
45.78
T1
3.50
19.80
26.52
137.00
Time
(min)
T2
3.40
18.57
24.25
118.00
T3
1.93
11.90
21.20
74.93
CSW1
3.48
11.70
16.02
69.00
CSW2
0.63
4.95
9.90
34.50
CSW3
0.83
6.57
12.78
56.00
T3
1.33
8.87
16.53
72.00
CSW1
3.48
11.70
16.02
71.00
CSW2
0.63
4.95
9.90
34.50
CSW3
0.83
6.57
12.78
56.00
T3
2.38
16.28
25.47
90.00
CSW1
3.40
15.60
27.12
98.00
CSW2
0.95
6.78
13.00
48.00
CSW3
1.83
10.57
18.62
84.00
T3
3.25
28.28
32.43
114.00
CSW1
2.02
17.22
23.12
90.00
CSW2
1.95
5.67
11.32
46.00
CSW3
2.37
5.57
10.72
93.00
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.40
5.77
11.53
53.37
CDI2
0.37
5.37
10.30
44.80
CDI3
0.40
5.37
10.52
44.55
T1
2.35
15.80
27.48
142.00
Time
(min)
T2
2.27
12.30
22.13
108.00
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.40
4.78
10.50
52.00
CDI2
0.38
4.43
9.35
44.38
CDI3
0.42
4.47
9.53
44.35
T1
3.93
20.85
35.93
177.00
Time
(min)
T2
5.08
21.97
28.23
105.00
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.40
4.65
10.73
52.00
CDI2
0.38
4.33
9.30
44.00
CDI3
0.42
4.83
9.60
45.35
T1
4.33
20.80
31.13
163.00
Time
(min)
T2
3.58
19.53
30.50
105.00
187
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.38
4.72
10.45
52.50
CDI2
0.37
4.40
9.42
43.75
CDI3
0.42
4.40
9.68
45.50
T1
2.75
18.83
32.83
172.00
Time
(min)
T2
1.82
19.00
27.75
118.00
T3
0.90
26.42
33.15
120.00
CSW1
3.12
18.17
24.55
103.00
CSW2
1.97
5.83
10.97
43.00
CSW3
1.83
6.17
12.70
103.00
T3
1.23
18.15
25.33
126.00
CSW1
1.55
18.48
26.23
115.00
CSW2
0.97
4.37
9.10
41.00
CSW3
2.15
17.20
24.18
105.00
T3
2.50
15.42
24.05
123.00
CSW1
3.92
19.65
26.77
117.00
CSW2
2.75
9.80
18.28
110.00
CSW3
2.60
15.32
20.08
85.00
CSW2
3.17
16.63
21.90
106.50
CSW3
2.27
12.30
17.42
92.63
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.38
4.93
10.97
53.13
CDI2
0.37
4.38
9.47
44.03
CDI3
0.42
4.55
9.87
44.50
T1
2.00
15.97
29.72
162.00
Time
(min)
T2
1.45
11.18
24.72
122.00
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.38
4.65
10.37
48.17
CDI2
0.37
4.42
9.45
42.00
CDI3
0.42
4.40
9.92
43.75
T1
3.22
20.48
31.93
163.00
Time
(min)
T2
2.63
11.58
20.58
116.00
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.38
4.28
9.33
47.75
CDI2
0.35
4.02
8.73
41.00
CDI3
0.37
4.22
9.20
45.00
T1
3.88
18.75
34.83
255.00
Time
(min)
T2
4.42
23.25
48.55
168.00
T3
2.27
13.25
22.00
110.00
CSW1
4.50
18.97
36.13
180.00
188
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.38
4.47
9.90
48.33
CDI2
0.35
4.00
8.72
40.92
CDI3
0.37
4.42
9.30
45.33
T1
8.00
15.75
35.18
250.00
Time
(min)
T2
9.42
19.25
30.55
185.00
T3
2.53
14.25
25.75
150.00
CSW1
4.50
25.67
33.32
156.00
CSW2
4.13
24.08
38.00
115.00
CSW3
2.80
17.92
29.90
128.00
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.38
4.67
10.30
49.67
CDI2
0.35
4.30
9.30
41.92
CDI3
0.37
4.55
9.73
44.83
T1
7.75
43.45
75.00
455.00
Time
(min)
T2
14.17
50.62
82.00
430.00
T3
8.07
35.53
58.20
437.00
CSW1
8.58
39.00
68.00
384.00
CSW2
2.67
17.08
31.00
200.00
CSW3
4.68
29.92
58.90
300.00
T3
4.67
22.75
38.75
254.00
CSW1
12.73
54.00
71.00
608.00
CSW2
9.00
22.87
30.00
200.00
CSW3
3.93
20.50
42.00
240.00
T3
7.67
42.50
75.00
565.00
CSW1
7.38
37.00
62.00
469.00
CSW2
5.38
39.83
73.00
190.00
CSW3
9.02
54.92
95.00
740.00
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.40
4.67
9.78
52.33
CDI2
0.35
4.42
9.67
43.92
CDI3
0.37
4.67
9.25
46.50
T1
20.00
107.00
150.00
1247.00
Time
(min)
T2
8.00
35.33
72.00
820.00
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.38
4.75
10.55
51.67
CDI2
0.35
4.17
9.62
44.60
CDI3
0.38
4.67
570.00
45.83
T1
24.17
139.00
230.00
7080.00
Time
(min)
T2
17.28
84.00
150.00
1410.00
189
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.38
4.80
10.58
51.50
CDI2
0.37
4.47
9.53
43.73
CDI3
0.38
4.60
9.93
45.82
T1
N/A
N/A
N/A
N/A
Time
(min)
T2
28.00
240.00
N/A
3522.00
T3
12.83
180.00
133.00
2715.00
CSW1
6.33
37.67
67.00
665.00
CSW2
5.00
31.25
63.00
744.00
CSW3
19.00
67.00
140.00
1805.00
CSW1
12.08
61.00
150.00
1417.00
CSW2
17.80
89.00
150.00
2880.00
CSW3
38.50
99.00
149.00
2750.00
CSW1
24.80
73.00
199.00
1938.00
CSW2
26.00
N/A
N/A
4578.00
CSW3
70.00
N/A
N/A
5070.00
CSW1
30.00
117.00
N/A
1912.00
CSW2
N/A
N/A
N/A
4758.00
CSW3
N/A
N/A
N/A
4950.00
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.38
4.80
10.40
51.17
CDI2
0.33
4.42
9.43
43.92
CDI3
0.40
4.67
10.18
45.75
T1
N/A
N/A
N/A
7170
Time
(min)
T2
N/A
N/A
N/A
17330
T3
42.00
180.00
N/A
3030.00
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.38
4.82
10.75
51.22
CDI2
0.35
4.30
9.60
43.17
CDI3
0.42
4.70
10.65
46.67
T1
N/A
N/A
N/A
N/A
Time
(min)
T2
N/A
N/A
N/A
N/A
T3
58.18
N/A
N/A
12100.00
Treatment Type
Infiltration
Volume (L)
0.08
0.47
0.78
1.12
CDI1
0.38
4.75
10.55
51.67
CDI2
0.35
4.17
9.62
44.60
CDI3
0.38
4.67
570.00
45.83
T1
78.00
204.00
N/A
15780.00
Time
(min)
T2
N/A
N/A
N/A
N/A
T3
N/A
N/A
N/A
N/A
190
E.0 Appendix E-Laboratory MPN Counts
Table E-1. Number of Positive Tubes per Treatment
Trial Number
T1
T2
T3
1
100
201
110
2
240
20
410
3
530
441
521
4
550*
542*
551*
5
555*
555*
555*
6
545*
554*
555*
7
533*
543*
533*
8
101*
211*
220*
9
511
522
512
10
531
512
530
11
220
511
212
12
434
321
440
13
501
511
221
14
542
552
512
15
301
321
211
16
220
221
231
17
421
411
420
18
0
100
0
*Made with 1 ml, 0.1 ml and 0.001 ml dilutions
CSW1
122
101
112
541*
350*
533*
520*
110*
501
330
311
430
320
231
231
231
302
100
CSW2
101
0
211
451*
350*
542*
431*
110*
440
230
220
421
321
120
223
211
321
100
CSW3
120
122
323
551*
552*
541*
502*
110*
512
321
211
440
320
401
122
220
311
0
E.coli
543*
544*
543*
542*
544*
542*
543*
543*
544*
532*
540*
551*
543*
543*
552*
541*
542*
543*
191
F.0 Appendix F-Laboratory Statistics
F.1 Variation of CSW and T treatment’s to CDI Infiltration Rate
¾ SAS Data Input
Day
0
3
6
9
12
15
18
21
24
27
30
33
36
39
42
45
48
51
54
57
60
CD1
(cm/hr)
135.0
138.4
145.8
148.4
144.9
146.8
144.9
148.7
148.7
147.3
145.5
160.6
162.0
160.0
155.7
147.8
149.7
150.2
151.1
151.0
149.7
CD2
(cm/hr)
170.5
171.7
180.0
181.0
172.6
177.9
172.6
174.2
175.8
176.8
175.6
184.1
188.6
189.0
184.5
176.1
173.4
176.8
176.1
179.2
173.4
CD3
(cm/hr)
158.7
163.4
166.3
176.8
174.2
168.9
173.6
174.4
170.5
170.0
173.8
176.8
171.9
170.6
172.5
166.3
168.7
168.8
169.0
165.7
168.7
T1
(cm/hr)
187.6
163.2
85.9
110.5
54.5
56.4
54.5
43.7
47.4
45.0
47.7
47.4
30.3
30.9
17.0
6.2
1.1
N/A
1.1
N/A
0.5
T2
(cm/hr)
209.9
176.8
161.4
126.3
71.6
65.5
71.6
73.7
73.7
65.5
63.4
66.7
46.0
41.8
18.0
9.4
5.5
2.2
0.4
N/A
N/A
T3
(cm/hr)
225.9
192.4
175.3
173.3
108.8
103.2
107.4
85.9
67.8
64.4
61.4
62.9
70.3
51.6
17.7
30.4
13.7
2.8
2.6
0.6
N/A
CSW1
(cm/hr)
192.8
235.3
211.9
193.3
113.7
112.1
108.9
78.9
85.9
75.1
67.2
66.1
43.0
49.6
20.1
12.7
16.5
11.6
5.5
4.0
4.0
CSW2
(cm/hr)
237.1
235.5
248.1
272.9
260.1
224.2
224.2
161.1
168.1
179.8
188.6
70.3
72.6
67.2
38.7
38.7
40.7
10.4
2.7
1.7
1.6
CSW3
(cm/hr)
204.9
241.5
230.8
229.1
171.3
138.1
138.1
92.1
83.2
75.1
73.7
91.0
83.5
60.4
25.8
32.2
10.5
4.3
2.8
1.5
1.6
¾ SAS Analysis
As done in the field study, Proc Mixed was run in SAS, since the data was
normalized and dependant. Slices were run to establish a test statistic for each run.
There is a significant difference (p<0.001) between all treatments. More specifically,
there is a significant difference (p<0.05) between all treatments from trial day 30 until
completion.
192
Differences of Least Squares Means
Effect
trt
trt
trt
trt/day
1
1
2
trt/day
2
3
3
Standard
Error
Estimate
2.0145
1.2626
-0.7518
0.1772
0.1760
0.1772
DF
5.92
5.83
5.92
t Value
11.37
7.18
-4.24
Tests of Effect Slices
Effect
day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
3
6
9
12
15
18
21
24
27
30
33
36
39
42
45
48
51
54
57
60
Num
DF
Den
DF
F Value
Pr > F
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
34.8
34.8
34.8
34.8
34.8
34.8
34.8
34.8
34.8
34.8
34.8
34.8
34.8
34.8
34.8
34.8
38.3
34.8
46.5
44.8
0.96
1.58
1.55
4.60
4.34
3.92
4.64
5.14
5.90
6.00
6.97
10.22
12.14
31.11
38.24
70.82
105.13
147.03
166.36
135.59
0.3910
0.2211
0.2260
0.0169
0.0208
0.0292
0.0163
0.0111
0.0062
0.0058
0.0029
0.0003
0.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
<.0001
Pr > |t|
<.0001
0.0004
0.0056
193
F.2 Variation of CSW and T treatment’s Infiltration Rate
¾ SAS Data Input
See F.1
¾ SAS Analysis
As done in F.1 , Proc Mixed was run in SAS, this time without the CDI treatment.
Slices were run to establish a test statistic for each run. There is a significant
difference (p<0.05) between CSW and T treatments. More specifically, there is a
significant difference (p<0.05) between treatments from trial day 12 until 18 and then
from trial day 45 until completion.
Effect
trt_day
trt
2
trt_day
3
Estimate
Error
-0.7519
Standard
DF
0.2070
3.93
t Value
Pr > |t|
-3.63
0.0228
Effect
day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
3
6
9
12
15
18
21
24
27
30
33
36
39
42
45
48
51
54
57
60
Num
DF
Den
DF
F Value
Pr > F
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
24.2
24.3
24.3
24.3
24.3
24.3
24.3
24.3
24.3
24.3
24.3
24.3
24.3
24.3
24.3
24.3
28.4
24.3
40.8
37.4
0.61
2.06
2.06
4.90
3.89
3.47
1.66
2.10
2.24
2.08
0.45
0.76
0.97
1.37
3.74
15.58
14.34
9.05
15.64
5.09
0.4422
0.1638
0.1637
0.0365
0.0602
0.0748
0.2097
0.1599
0.1476
0.1619
0.5070
0.3934
0.3349
0.2537
0.0648
0.0006
0.0007
0.0060
0.0003
0.0301
194
F.3 Variation of CSW and T treatment’s Bacteria Concentration
¾ SAS Data Input
Trial Number
T1
T2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
2
80
240
900
11199
3500
1700
500
488
700
90
340
300
236
110
9
26
0
7
34
220
1600
17329
16000
2800
900
920
600
500
170
500
517
270
9
21
2
T3
CSW1
CSW2
CSW3
4
9
140
1600
3500
2200
330
280
345
120
90
26
17
5.2
9
7
17
2
6
17
300
1600
4884
1700
400
220
365
170
90
34
14
16
11
9
14
0
(MPN index/100 ml)
4
70
300
1600
12997
16000
1700
900
613
800
120
340
90
58
120
12
22
0
6
6
170
500
3654
1700
500
280
300
170
140
27
14
12
12
12
11
1
¾ SAS Analysis
were taken. Using Proc Mixed, s significant difference (p <0.01) was found between
bacteria concentrations in the effluent of treatment CSW and T. More specifically,
there was a difference (p <0.05) from trail day 15 through 33, excluding trial 27 and
then from trial day 36 until 45.
195
Effect
trt
trt-day
2
trt _day
3
Estimate
Error
1.1985
Standard
DF
t Value
0.1404
4
8.53
Effect
day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
trt*day
3
6
9
12
15
18
21
24
27
30
33
36
39
42
45
48
51
54
Num
DF
Den
DF
F Value
Pr > F
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
58.3
58.3
58.3
58.3
58.3
58.3
58.3
58.3
58.3
58.3
58.3
58.3
58.3
58.3
58.3
58.3
58.3
67.9
0.73
23.42
0.52
0.28
11.26
20.15
19.03
8.20
3.25
17.21
2.00
37.09
56.78
64.69
52.75
0.05
1.90
0.11
0.3959
<.0001
0.4727
0.5959
0.0014
<.0001
<.0001
0.0058
0.0768
0.0001
0.1622
<.0001
<.0001
<.0001
<.0001
0.8205
0.1730
0.7438
Pr > |t|
0.0010
196
F.4 Correlation Between Infiltration Rate and Total Coliform
Concentration
¾ SAS Data Input
See F. 1 and F.3
¾ SAS Analysis
Since the data was slightly skewed, the natural log of the bacteria
concentrations were taken. To establish a correlation between effluent bacteria
concentration and infiltration rate of dependant variable, Proc Genmod was used.
This procedure uses maximum likelihood estimation to fit generalized linear
models. This method is a general statistical modeling tool which fits generalized
linear models to data. A significant correlation (p<0.01) found with between
infiltration rate, trial day to log of the effluent’s total coliform concentration.
Analysis Of Parameter Estimates
Parameter
Intercept
day
infil
DF
Estimate
1
1
1
13.8105
-0.1803
-0.0489
Standard
Error
1.6900
0.0370
0.0110
Wald 95% Confidence
Limits
10.4982
-0.2528
-0.0704
17.1227
-0.1079
-0.0274
ChiSquare
Pr > ChiSq
66.78
23.81
19.81
<.0001
<.000
<.0001

Abstract - Biological and Agricultural Engineering

Transcription

Similar documents

city place

Stormwater Pioneers Award Ceremony

Brochure - XP Solutions

Jim Gills, Lake County Engineer

Downspout Disconnection

PDF for printing tri-fold brochure

Hydromedia Brochure

Borgert Permeable Pavements Brochure

Conestoga College Cambridge Campus

Lockwood Folly River Restoration Plan