bacteria & viral indicator contamination of stormwater - a multi-watershed study

29
1 Prioritizing Stormwater Enforcement Efforts, a Multi-Watershed Study (Project No. 98-04/104) Final Report Tom Mahin, MADEP; Dave Gray, EPA Region 1; Ron Stoner, MADEP; Susan Gifford, MADEP; Oscar Pancorbo, MADEP * A* Executive Summary This project was a collaborative project between four watershed associations and the Massachusetts Department of Environmental Protection (MADEP). A total of 131 samples were collected during 4 storm events at 18 locations in 4 Massachusetts watersheds (Merrimack, Charles, Neponset and Ipswich basins. In addition, the Grantee (MADEP) conducted a comprehensive review of epidemiological studies completed since EPA’s 1986 recommendations relative to receiving water bacterial indicators. This project was proposed to assist in answering the following questions: What are “average” levels of bacterial indicators and viral indicators (coliphages) associated with discharges from municipal storm drains in Massachusetts? How do levels of enterococci compare to levels of E. coli in stormwater in Massachusetts? Does land use have a demonstrable impact on levels of bacterial indicators or viral indicators (coliphages)? Do coliphages correlate with traditional bacterial indicators? It can be argued that the value of coliphages as pathogen indicators is that they may not correlate with bacterial indicators, which are known to have limitations. Are the results from the epidemiological studies conducted since EPA’s 1986 receiving water bacterial standard recommendations consistent with the recommendations that the states use enterococci for marine waters and either E. coli or enterococci for freshwaters? Analysis of the laboratory results led to the following conclusions: Enterococci counts were much higher than E. coli levels. Rivers, ponds, lakes, etc. heavily impacted by stormwater in the watersheds studied may be more likely to be associated with a water quality standard violations depending on whether samples are analyzed for E. coli or enterococci. Bacterial indicators (E. coli, fecal coliform, enterococci) did correlate with each other

Upload: tom-mahin

Post on 21-Dec-2014

302 views

Category:

Health & Medicine


3 download

DESCRIPTION

 

TRANSCRIPT

Page 1: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

1

Prioritizing Stormwater Enforcement Efforts, a

Multi-Watershed Study (Project No. 98-04/104)

Final Report

Tom Mahin, MADEP; Dave Gray, EPA Region 1; Ron Stoner, MADEP;

Susan Gifford, MADEP; Oscar Pancorbo, MADEP

* A*

Executive Summary

This project was a collaborative project between four watershed associations and the

Massachusetts Department of Environmental Protection (MADEP). A total of 131

samples were collected during 4 storm events at 18 locations in 4 Massachusetts

watersheds (Merrimack, Charles, Neponset and Ipswich basins.

In addition, the Grantee (MADEP) conducted a comprehensive review of

epidemiological studies completed since EPA’s 1986 recommendations relative to

receiving water bacterial indicators.

This project was proposed to assist in answering the following questions:

What are “average” levels of bacterial indicators and viral indicators (coliphages)

associated with discharges from municipal storm drains in Massachusetts?

How do levels of enterococci compare to levels of E. coli in stormwater in

Massachusetts?

Does land use have a demonstrable impact on levels of bacterial indicators or

viral indicators (coliphages)?

Do coliphages correlate with traditional bacterial indicators? It can be argued that

the value of coliphages as pathogen indicators is that they may not correlate with

bacterial indicators, which are known to have limitations.

Are the results from the epidemiological studies conducted since EPA’s 1986

receiving water bacterial standard recommendations consistent with the

recommendations that the states use enterococci for marine waters and either

E. coli or enterococci for freshwaters?

Analysis of the laboratory results led to the following conclusions:

Enterococci counts were much higher than E. coli levels. Rivers, ponds, lakes,

etc. heavily impacted by stormwater in the watersheds studied may be more likely

to be associated with a water quality standard violations depending on whether

samples are analyzed for E. coli or enterococci.

Bacterial indicators (E. coli, fecal coliform, enterococci) did correlate with each

other

Page 2: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

2

The viral pathogen indicator used (male-specific coliphages) did not correlate

with the three bacterial indicators or with the water chemistry parameters.

Significantly higher levels of coliphages were found in certain locations (see

Appendices B and C) raising the question whether very high levels of coliphages

may be indicative of illicit sewage connections?

Bacterial indicator densities did not correlate with land use.

In addition, the review of epidemiological studies found that the results generally

were consistent with using enterococci for marine waters and either E. coli or

enterococci for fresh waters.

Page 3: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

3

Table of Contents

Executive Summary Page 1

Introduction Page 5

Field Methods Page 6

Laboratory Analyses Page 8

Use of Coliphages as an Alternative Pathogen Indicator Page 10

Results/Conclusions Page 12

References Page 23

Appendix A – Conference Paper Generated by Grant Page 25

Appendix B – Project Quality Assurance Program Plan (QAPP)

Page 4: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

4

Disclaimer/Acknowledgements

The project has been financed partially funded with federal Funds from the

Environmental Protection Agency (EPA) to the Massachusetts Department of

Environmental Protection (the Department) under Section 104(b)(3) of the Clean Water

Act. The contents do not necessarily reflect the views and policies of EPA or of the

MADEP, nor does the mention of trade names or commercial products constitute

endorsement or recommendation for use.

The MADEP thanks the following organizations and their staff for their support and

assistance during the project:

The Charles River Watershed Association

The Merrimack River Watershed Council

The Ipswich River Watershed Association

The Neponset River Watershed Association

In addition, thanks to Gary Gonyea for reviewing and commenting on the draft of the

Final Project Report.

Contact for Questions

For questions relative to this study, please contact the Project Officer for the study,

Tom Mahin of MADEP at [email protected]

Page 5: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

5

INTRODUCTION

This project was a collaborative project between 4 watershed advocacy groups and the

MADEP Northeast Regional Office and Wall Experiment Station. Stormwater samples

were collected during 4 storm events at 18 locations in freshwater portions of 4

northeastern Massachusetts watersheds: the Merrimack, Charles, Neponset and Ipswich.

The objective of the study was to compare a variety of pathogen indicators for their

potential in prioritizing stormwater remediation efforts. The project was funded by an

EPA 104(b)(3) grant (Project No. 98-04/104) with a state match of a portion of the funds.

Stormwater pollution has been identified as the leading cause of “pathogen” (bacterial

indicator) water quality standards violations in many watersheds. For example, there has

been a considerable amount of work done in the Lower Charles Basin to assess pollutant

loads and characterize water quality conditions. MWRA in its recently issued CSO

Facilities Plan has identified stormwater pollution to be the most prominent pollution

source in the Lower Charles Basin. This conclusion appears to be supported by the water

quality sampling done to date in the basin.

This project was proposed to assist in answering the following questions:

What are “average” densities of bacterial indicator and viral indicator (coliphages)

organisms associated with discharges from municipal separate storm drains in

Massachusetts?

How do fecal coliform, enterococci, and E. Coli observed in stormwater and

select instream locations compare in terms of densities?

Does land use have a demonstrable impact on observed densities of bacterial or

viral indicators (coliphages)?

Do coliphages correlate with traditional bacterial indicators? It can be argued that

the value of coliphages as pathogen indicators is that they may not correlate with

bacterial indicators, which are known to have limitations.

Field Methods

Sampling Locations & Methodology

All samples were collected, preserved and transported to MADEP’s Wall Experiment

Station in accordance with the project’s approved Quality Assurance Project Plan

(QAPP) dated November 1999 (attached as Appendix E). 117 aqueous samples were

collected from a total of 15 mainstem and tributary stormwater outfalls, plus three

culverted brook locations from Fall 1999 – Summer 2000 (including Winter). Sampling

locations are shown in Figure 1 and described in Appendix A. Outfalls ranged in size

from 8-inch diameter to 7’ x 12’ box culverts, servicing a variety of land uses, both with

and without the contribution of suspected illicit discharges.

Page 6: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

6

Quantitative precipitation predictions from the National Weather Service were used with

local commercial weather forecasting outlets to determine if a qualified precipitation

event was forthcoming. For the purposes of this study, a qualified rainfall event was

defined as a minimum 0.25-inch rainfall that generated sufficient stormwater volume and

duration to facilitate collection of all required samples at all stations. Qualified events

included those large enough in extent to have generated the minimum rainfall required

throughout all four watersheds (e.g. a frontal storm) and isolated precipitation events in

only some of the river basins.

Samples were collected after varied antecedent dry period conditions (1 to 10 days),

cumulative rainfall depths (<0.1 to 1.39-inches), and rainfall/runoff duration (first flush to

24-hours after start of precipitation). Though most were single grab samples, 15-minute

grab samples were collected at one station in each watershed during each sampling event

in an attempt to assess any temporal variability. In general, sample volume was collected

directly into 10-liter carboys equipped with a dispensing spigot. Where direct collection

was not possible due to flow angles or access, flow volume was collected and transferred

into carboys from sanitized 2-gallon buckets or from 1-liter sample bottles using a swing

sampler on a telescoping pole. Once full, the carboys were continuously agitated and

individual sample bottles were filled from the dispensing spigot.

Physical Observations & Field Measurements

In addition to collecting aqueous samples for laboratory analysis of physical, chemical,

and pathogen indicators, sampling crews collected field measurements of pH and

temperature, and noted physical observations regarding odor, color, clarity, floatables,

deposits/stains, and vegetation on standard forms, in field notebooks, and through

photography.

Page 7: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

7

Figure 1 - Sampling Locations

Page 8: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

8

Laboratory Analysis

Summary of Parameters Evaluated as Part of This Study:

Bacterial Indicators – Enterococcus, E. coli, Fecal Coliform & Clostridium

Perfringens Viral Indicators – Male-Specific Coliphages & Somatic Coliphages

Water Chemistry – Ammonia, Biochemical Oxygen Demand (BOD), Total

Suspended Solids (TSS), Anionic Surfactants (as MBAS) and Fluorescent

Whitening Agents, Fluoride, Specific Conductance, pH, Temperature, Chronic

Toxicity

Flow Related – Storm Duration, Precipitation, Storm Intensity, Antecedent Dry

Period

Land Use

Analytical Methods - Physicochemical Analyses

BOD by SM 5210B

TSS by SM 2540D

Ammonia-N by EPA 350.1 (Automated phenate colorimetry)

Fluoride by EPA 300.0 (Ion chromatography)

Anionic surfactants as MBAS by SM 5540C

Specific conductance by SM 2510B

Fluorescent whitening agents by HPLC (Fluorescence Detector)

Analytical Methods - Pathogen Indicators and Toxicity

Total Coliform SM9222B1

Fecal Coliform SM9222D1

E. coli SM9213D1

Enterococci SM9230C1

C. perfringens EPA-ICR membrane filtration method3

Male-specific coliphage Double-layer agar plaque assay3,4,5

Chronic Microtox Toxicity Test Azur Environmental6 1

American Public Health Association. 1995. Standard Methods for the Examination of Water and

Wastewater, 19th

edition. APHA, Washington, D.C.. 2

U.S. Environmental Protection Agency. 1983. Methods for the Chemical Analysis of Water and Wastes. EPA600/4-79-020. USEPA, Cincinnati, Ohio. 3 U.S. Environmental Protection Agency. 1996. ICR Microbial Laboratory Manual. EPA/600/R-95/178.

USEPA, Washington, D.C. 4 Grabow, W.O.K. and P. Coubrough. 1986. Practical direct plaque assay for coliphages in 100-mL

samples of drinking water. Applied and Environmental Microbiology. 52(3): 430-433.

Page 9: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

9

5 Sobsey, M.D., K.J. Schwab, and T.R. Handzel. 1990. A simple membrane filter method to concentrate and evaluate

male-specific RNA coliphages. J. Am. Water Works Assoc. 82(9): 52-59. 6 Azur Environmental. 1996. Microtox Chronic Toxicity Test. Azur Environmental, Carlsbad, CA.

Coliphage Analyses at Wall Experiment Station as Part of This Grant

Coliphages were concentrated from 1-L storm water samples by membrane filtration-

elution and assayed by the DAL method as follows:

Somatic coliphages – assayed on E. coli C or CN-13 host with phage fx174 as

positive control

Male-specific (F+) coliphages – assayed on E. coli Famp as host with phage

MS2 as positive control.

Coliphage Analysis at Wall Experiment Station

Page 10: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

10

Use of Coliphages as an Alternative Pathogen Indicator

Background on Coliphages

Coliphages are viruses that infect E. coli coliform bacteria and are

nonpathogenic to humans.

They are believed to be more similar to enteric viruses with respect to physical

characteristics, persistence in the environment and resistance to treatment

processes than are traditional indicator bacteria such as fecal coliforms.

They are relatively easy and inexpensive to analyze.

Coliphages have been reported to occur in high concentrations in sewage

treatment plant influent and reproduce in sanitary sewers under appropriate

conditions.

A variety of domestic and un-domesticated animals shed coliphages in their feces, but

usually at lower levels than found in human sewage (Calci et al, 1998).

A study was conducted during the summers of 2002 and 2003 in Madison, Wisconsin

(EMPACT, 2004) during which a total of 223 water samples were collected at the three

beach sites for determination of male-specific coliphages. Male-specific coliphages

were detected in 33 of these samples ranging in relatively low concentrations from <1 to

23 PFU/100mL.

A study conducted by the USGS during 2000 and 2002 (Bushon and Koltun, 2003),

found that coliphages didn’t correlate well with other microorganisms in Cuyahoga River

water and samples from a tributary wastewater treatment facility in Ohio (see table

below).

Source - (Bushon and Koltun, 2003)

Page 11: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

11

Investigations conducted by the Massachusetts Water Resources Authority from 1995-

2003 (Ballester et.al., 2004) documented no or poor correlation between coliphages

(male-specific and somatic) and other bacteria indicator organisms (fecal coliform,

enterococci, and E. Coli.) found in Boston Harbor, the Charles River, and the influent and

effluent of the Cottage Farm CSO facility and the Deer Island Treatment Plant.

Investigators concluded that coliphages could be used as relatively conservative tracers of

sewage in the region since they were observed to be more persistent through wastewater

treatment and in the environment than were the bacterial indicators examined.

Page 12: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

12

Results/Conclusions

Discussion of Results From This Study

For the traditional bacterial indicators, the highest concentrations were associated with

the summer and fall seasons. It should be noted though that actual loadings to

waterbodies are a factor of flow and concentrations, so overall loadings in the spring may

exceed loadings in the summer (on an overall season basis). The USGS Oregon study

(USGS Oregon, 2002) only looked at one bacterial indicator but found higher E. coli

levels in the summer than in the other 2 seasons studied (spring and winter).

Seasonal Comparison of Concentrations of Select Pathogen Indicators in This Study

Pathogen Indicator

(Geomeans)

Spring

Summer Fall

Winter

Bacterial Indicators

Enterococcus

(cfu/100 mL)

2,736 14,035 16,204 1,625

E. coli (cfu/100 mL) 350 1,906 1,584 312

Fecal Coliform

(cfu/100mL)

871 5,705 3,999 1,061

Viral Indicator

Male-Specific

Coliphages (pfu/L)

48 137 63 207

Because the summer season is most closely associated with recreational water uses that

are the basis of receiving water standards, a comparison of indicator levels by watershed

during summer storm events was conducted (see table below).

Comparison of Bacterial Indicator Levels by Watershed for Summer Storm Events

Pathogen Indicator

(Geomeans)

Merrimack

Basin

Charles Neponset Ipswich

Enterococci

(cfu/100 mL)

16,686 13,944 8,124 18,537

E. coli (cfu/100 mL) 4,778 NA 3,539 767

Fecal Coliform

(cfu/100mL)

13,897

1,999

5,090

2,882

Comparison of The Results From This Study With Other Stormwater Studies

Page 13: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

13

A number of studies have been done relative to bacterial levels in stormwater. The

Oregon USGS study noted below (USGS Oregon, 2002) looked at E. coli levels near

Portland, Oregon and the Charles River USGS Study (USGS Charles River, 2002) looked

at enterococci and fecal coliform levels. Neither of these two studies however looked at

both E. coli and enterococci levels.

The USGS conducted a study of E. coli levels from stormwater runoff in a creek

classified as 100% urban near Portland Oregon. The figure shows the E. coli levels for 3

storms during 1998-1999 (median values are the horizontal line on the box whiskers).

________________________________________________________________________

E. Coli Levels From Storm Runoff in an Urban Creek in Oregon (Source - USGS Oregon,

2002)

Page 14: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

14

The USGS conducted a study of the Charles River that included analyzing for fecal

coliform and enterococci but not E. coli in stormwater (USGS Charles River, 2002). As

part of the USGS Charles River Study they determined the “mean” enterococci

concentration from a number of different stormwater studies. These “other studies”

include stormwater data collected from 23 cities between 1978 and 2000 by many

different municipalities and agencies and. The “mean” value of 6,400 CFU/100 mL of

enterococci from mixed land use in the table below (the mean of the enterococci

concentrations from the selected other studies) is similar to the geometric mean (6,700

CFU/100 mL) and median (6,950 CFU/100 mL) of all samples analyzed for this study

(see page 17 of this report).

Summary of stormwater data from other studies (Source: USGS Charles River, 2002))

Page 15: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

15

In addition, the USGS Charles River study evaluated enterococci levels at a number of

locations in the Charles River basin, the results of which are shown below: The median

enterococci concentration for the wet weather samples analyzed as part of the Charles

River Study was 13,000 CFU/100 mL (Page 111 USGS Charles River Study, 2002).

Enterococci levels detected as part of the USGS Charles River study (Source - USGS

Charles River, 2002)

Page 16: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

16

Conclusion 1: The three bacterial indicators did correlate with each

other (see parameters in bold yellow/underlined in table below).

Pearson Correlation Coefficients

(P < 0.05) for Microbial Parameters in MA

Storm Water

0.460.610.59Som-phage

0.410.480.45MS-phage

C. perfringens

1.000.760.77Enterococci

0.761.000.87E. coli

0.770.871.00Fecal

Coliform

EnterococciE. coliFecal Coliform

Page 17: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

17

Conclusion 2: Enterococci counts were much higher than E. coli counts

in the same sample especially at lower E. coli densities

Page 18: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

18

Conclusion 3: The viral indicator male-specific coliphages

(“MS-phage”) did not correlate well with bacterial indicators. This

finding is not surprising given that the potential advantage of a viral indicator is that it

potentially more closely mimics characteristics of actual pathogens (higher survivability

in the environment, etc.) than do bacterial indicators. Therefore it does NOT mean that

male-specific coliphages are not a good pathogen indicator. It was beyond the scope of

this project to sample and analyze for a variety of waterborne pathogens. Such a

comparison of the correlation of coliphages with actual pathogens versus the correlation

of bacterial indicators versus actual pathogens in stormwater would be a worthwhile

future project.

Pearson Correlation Coefficients

(P < 0.05) for Microbial Parameters in MA

Storm Water

0.460.610.59Som-phage

0.410.480.45MS-phage

C. perfringens

1.000.760.77Enterococci

0.761.000.87E. coli

0.770.871.00Fecal

Coliform

EnterococciE. coliFecal Coliform

Correlation Coefficients for Coliphages (“MS-phage/Som-phage”) with Bacterial

Indicators

Regrsn. of E. coli vs. Male-Specific Coliphages in Storm Water

Correlation: r = 0.48 p < 0.05

E. coli Conc. (Log10

CFU/100 mL)

F+ C

oli

ph

ag

e C

on

c.

(Lo

g 10P

FU

/L)

0

1

2

3

4

5

6

0 1 2 3 4 5

Regression line with

95% confidence limits

Theoretical x = y line

Page 19: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

19

Conclusion 4: None of the bacterial or viral indicators correlated well

with pH, temperature, specific conductance or BOD.

Pearson Correlation Coefficients

(P < 0.05) for Microbial and Selected

Physicochemical Parameters in MA Storm Water

-0.210.55Som-phage

MS-phage

C. perfringens

0.27-0.300.46-0.19Enterococci

-0.400.39E. Coli

0.19-0.390.35Fecal coliforms

BODSpecific

Cond.

Water

Temp. pH

Correlation Coefficients for Bacterial Indicators and Coliphages

(“MS-phage/Som-phage”) with pH, Water Temp., Spec. Conductance and BOD

Page 20: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

20

Conclusion 5: Significantly higher levels of male-specific coliphages

found in samples from certain locations.

This raises the question of whether very high levels of male-specific coliphages are

indicative of illicit sewage connections in the watersheds studied. As shown in the figure

below, coliphage levels varied from less than 10 PFU/L to greater than 100,000 PFU/L in

this study.

Regrsn. of E. coli vs. Male-Specific Coliphages in Storm Water

Correlation: r = 0.48 p < 0.05

E. coli Conc. (Log10

CFU/100 mL)

F+ C

oli

ph

ag

e C

on

c.

(Lo

g10P

FU

/L)

0

1

2

3

4

5

6

0 1 2 3 4 5

Regression line with

95% confidence limits

Theoretical x = y line

Page 21: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

21

Conclusion 6: Land use did not significantly impact bacterial levels in

this study (bacterial indicator levels were generally elevated for all types of land use,

see graph below). It was unclear why there was not a correlation with land use. One

possible explanation is that sewage cross connections/illicit connections to storm drains

contributed to pathogen indicator levels regardless of land use type. Also animal scat

contributes to bacterial levels in stormwater regardless of predominant land use type. It

is possible that use of composite sampling might have reached a different conclusion

relative to correlation with land use.

FC

Outliers

E_COLI

Outliers

ENTCOCCI

Outliers

CP

Outliers

MS_PHAGE

Outliers

Extremes

SOM_PHAG

Outliers

Microbial Densities in MA Storm Water by Land Use

Median; Box: 25%, 75%; Whisker: Non-Outlier Min, Non-Outlier Max

LAND USELo

g1

0 B

ac

teria

l (C

FU

/10

0 m

L)

or C

oli

ph

ag

e (

PF

U/L

) C

on

c.

0

1

2

3

4

5

6

Residential

Ind-Comm Mix

Open Space-Rec

Res-Comm Mix

Page 22: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

22

Conclusion 7: The study was successful in identifying stormwater

outfalls with particularly high bacterial and viral indicator levels

that should be evaluated further for potential remediation.

Examples of such locations follow (these are not necessarily the most significant

levels detected but rather examples, the reader is referred to Appendix A Sample

Locations and Appendix B Detailed Study Results for more detailed information):

Sample

Location Site

# Fecal

Coliform (cfu/100 mL)

Enterococci (cfu/100 mL)

E. coli (cfu/100

mL)

Male-Specific

Coliphages (pfu/L)

Somatic Coliphages

(pfu/L)

Charles

CRWA- 1-3-1

3 53,000 96,000 NA 660,000 56,000

CRWA- 1-4-1

4 50,000 86,000 NA 660,000 14,000

Merri-mack

MRWC 2-1-1

1

280,000

1,000,000

110,000 1,700

1,700

MRWC 1-3-1

3

55,000

75,000

20,000

120,000

5,600

MRWC 4-3-1

3

26,000

72,000

8,000 200,000

9,600

MRWC 2-3-1

3

260,000

500,000

60,000 42,000

14,000

MRWC 1-4-1

4

91,000

23,000

20,000

180,000

6,800

Neponset

NepRWA-

1-5-1

5

20,000

5,000

20,000 100

43,000

NepRWA-

1-5-2

5

31,000

41,000

22,000 210

57,000

Ipswich

IRWA- 2-1-1

1

31,000

20,000

15,000 8

11,000

IRWA- 1-2-1

2

660

330,000

NA 8

65,000

Page 23: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

23

REFERENCES

(Ballester et.al., 2004) Ballester, N.A., Rex, A.C., and Coughlin, K.A. 2004. Study of

anthropogenic viruses in Boston Harbor, Charles River, Cottage Farm CSO Treatment Facility

and Deer Island Treatment Plant: 1995-2003. Boston: Massachusetts Water Resources Authority.

Report Enquad 2004-15.57 pp., at URL http://www.mwra.state.ma.us/harbor/enquad/pdf/2004-

15.pdf

(Bushon and Koltun, 2003) “Microbiological Water Quality in Relation to Water-

Contact Recreation, Cuyahoga River, Cuyahoga Valley National Park, Ohio, 2000 and

2002” Rebecca N. Bushon and G.F. Koltun USGS WRIR 03-4333

(Calci et al, 1998) “Occurrence of Male-Specific Bacteriophage in Feral and Domestic

Animal Wastes, Human Feces, and Human-Associated Wastewaters” Kevin R. Calci,

William Burkhardt III, William D. Watkins, and Scott R. Rippey, Applied and

Environmental Microbiology, December 1998, p. 5027-5029, Vol. 64, No. 12

(Cole et al, 2003) “Evaluation of F+ RNA and DNA Coliphages as Source-Specific

Indicators of Fecal Contamination in Surface Waters” Dana Cole, Sharon C. Long, and

Mark D. Sobsey, Appl Environ Microbiol. 2003 November; 69(11): 6507–6514.

(EPA, 1986) Ambient Water Quality Criteria for Bacteria – 1986, EPA440/5-84-002,

January 1986

(Gray and Mahin, 1999) Proceedings of EPA’s BEACH Conferences, Tampa, Florida

and San Diego, CA

(Metcalf & Eddy, 1979) Wastewater Engineering: Treatment Disposal Reuse, table 3-16

Page 103, Published by McGraw-Hill Company 1979

(USGS Oregon, 2002) Phosphorus and E. coli and Their Relation to Selected

Constituents During Storm Runoff Conditions in Fanno Creek, Oregon, 1989-99, USGS

Water Resources Investigations Report 02-4232

(USGS Charles River, 2002) Streamflow, Water Qualirt, and Contaminant Loadings in

the Lower Charles River Watershed Massachusetts, 1999-2000

Water-Resources Investigations Report 02-4137 02-4137

Page 24: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

24

APPENDIX A

Conference Paper Generated

by This Grant

Page 25: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

25

Bacterial Indicators and Epidemiological Studies at Beaches;

Implications for Stormwater Management (WEFTEC 2001 proceedings)

Tom Mahin, Chief of Municipal Services Section

Massachusetts Department of Environmental Protection, Northeast Regional Office

205a Lowell St., Wilmington, MA 01887

Phone: (978) 661-7696, Fax: (978) 661-7615

e-mail: [email protected]

Introduction

Based on epidemiological studies at beaches in the U.S., the USEPA has recommended

for a number of years that states use enterococci as the bacterial indicator for marine

waters and either enterococci or E. coli as the indicator for freshwaters (USEPA 1986).

The Massachusetts Department of Environmental Protection (DEP) recently completed a

comprehensive review and critical analysis of all the more recent (mostly non-EPA)

published epidemiological studies that were conducted subsequent to EPA’s original

recommendation. The goals of the review were as follows:

(1) To evaluate the more recent epidemiological studies to determine whether they

justified changing the DEP water quality standards for fresh and marine waters

(currently fecal coliform for both types of waters), and

(2) To analyze the potential implications for stormwater management given that

stormwater discharges are the main cause of exceedances of bacterial water quality

standards in Massachusetts.

Both the conclusions and the methodologies used in the studies were reviewed in detail.

Examples of some of the major epidemiological studies reviewed are noted below.

Page 26: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

26

Examples of Relevant Epidemiological Studies

During 1989-1992 during four consecutive summers, epidemiological studies were

carried out at marine beaches in England , the “UK beach studies” (Kay et al. 1994).

The UK beach studies differed from previous epidemiological studies in two

important ways. First, volunteers were randomly assigned as either bathers or non-

bathers. Secondly rather than relying of self-describing of symptoms, clinical

examinations were included as part of the study. The studies involved a total of 1216

participants. The studies found a dose-response relationship between fecal

streptococci (FS) and gastrointestinal (GI) illness. It should be noted that the

definition of fecal streptococci as used in these studies is very similar or the same as

enterococci as used in the U.S. An increase in GI illness rates was observed when FS

levels exceeded 32 colony-forming units (cfu) per 100 ml.

The studies also reported what was described as a “clear dose-response relationship”

between respiratory illness and fecal streptococci levels. The threshold level for

increased illness was 60 cfu/100 ml. While these studies only dealt with marine

waters and not fresh waters, the results appear consistent with the work done by EPA

that indicated that enterococci works well as an indicator of rates of GI illness in

marine waters whereas fecal coliform does not.

A major epidemiological study was conducted in Hong Kong in 1992 involving

25,000 beach-goers at coastal beaches (Kueh et al. 1995). Unfortunately fecal

streptococci/enterococci was not analyzed for. The study did find that “no direct

relationship between GI symptoms and E. coli or fecal coliforms could be identified

in this study”. The findings of the study appear consistent with USEPA’s position

that fecal coliform and E. coli are not effective at predicting GI illness in users of

marine waters.

An epidemiological study was conducted in 1995 of swimmers in the marine waters

of Santa Monica Bay (Haile et al. 1996). The study included 111,686 subjects.

Illness rates were compared for those swimming near stormwater outfalls versus

those swimming further away. Illness rates were also compared to various bacterial

indicators. Fecal coliform levels > 400/100 ml correlated only to skin rash and

E .coli correlated only with earache and nasal congestion. Enterococci levels

>106/100 ml were statistically correlated with “highly credible GI illness” and also

with “diarrhea with blood”.

Conclusions and Unresolved Issues

How much of a risk does wet weather stormwater/urban runoff pose to recreational

beach-goers? The Santa Monica study doesn’t appear to have answered this question

because the samples presumably included either mostly dry weather flow (given the

Page 27: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

27

climate in Southern California) or non-local origin flow. The dry weather flow

presumably could include significant amounts of illicit sewage connections. This

could have been responsible for significant percentage of the illness rates detected.

None of the epidemiological studies described above appear to have relied strictly on

traditional wet weather stormwater conditions as occur in non-arid areas of the U.S.

An epidemiological study was conducted by Yale University and EPA staff at a

pond used for swimming in Connecticut that received only runoff contaminated by

animal feces and not sewage (Calderon et al. 1991). The study that included 104

families did not detect a correlation between illness rates and levels of traditional

bacterial indicators but did find that bather density correlated with increased rates of

gastroenteritis in swimmers.

It is unclear what the source of contamination is in many of the studies reviewed.

EPA’s original epidemiological studies may have involved contamination resulting

mostly (or in significant part) from chlorinated effluents.

Since stormwater discharges are mostly unchlorinated, they may exhibit lower

pathogen to bacterial indicator levels than may have been present (but not analyzed

for) in many of the epidemiological studies if chlorinated effluents were the primary

source. Such a lower pathogen to indicator ratio, if confirmed, could have the

potential to overestimate the risk due to stormwater relative to previous EPA

studies.

Given the high levels of enterococci and other bacterial indicators that are

commonly detected in stormwater in urban areas around the country, evaluating the

true risk of stormwater becomes of critical importance. It should be noted however

that many stormwater drainage systems in urban areas (at least associated with the

aging infrastructure in the in the Northeast U.S.) contain significant amounts of

illicit sewage connections. Given this fact, a conservative approach would argue for

adopting the levels recommended by EPA at least until more progress is made in

reducing illicit connects and until future epidemiological studies (if conducted) can

provide better information relative to the specific risk from stormwater.

Can a single indicator adequately predict a range of illnesses in swimmers in

marine waters? USEPA recommends that only enterococci be used for marine

waters. The UK beach studies found that only increased levels of fecal coliform

organisms were predictive of ear ailments among bathers in the coastal waters studied

(Fleisher et al. 1996). In addition, the Santa Monica study found that E. coli was the

best predictor of earache after swimming (marine waters). Both of these more recent

studies seem to back up the argument that enterococci be used as the overall best

indicator for marine waters at least for gastroenteritis and respiratory illness.

However they also seem to point to the need for additional epidemiological studies to

clarify whether a single indicator is adequate to predict illness in swimmers using

marine beaches.

Page 28: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

28

References

Calderon, R. 1991. Health Effects of Swimmers and Nonpoint Sources of Contaminated

Waters. International Journal of Environmental Health Research 1, 21-31

USEPA 1986. Dufour, A., and R. Ballentine Ambient Water Quality Criteria for Bacteria

– 1986. EPA 440/5-84-002

Fleisher, J. M. et al. 1996. Marine waters contaminated with domestic sewage: nonenteric

illness associated with bather exposure in the UK. Am J Public Health 86: 1228-34

Haile, W., et al. 1996. An Epidemiological Study of Possible Adverse Health Effects of

Swimming in Santa Monica Bay. Final Report, May 6, 1996).

Kay, D. et al. 1994. Predicting likelihood of gastroenteritis from sea bathing; results

from randomized exposure. Lancet 344, 905-09

Kueh, C.S. et al, 1995. Epidemiological Study of Swimming-Associated Illnesses

Relating to Bathing-Beach Water Quality, Wat. Sci Tech.

Page 29: Bacteria & viral indicator contamination of stormwater - a multi-watershed study

29

Appendix B – Project Quality

Assurance Program Plan (QAPP)

Available upon request at MADEP, please contact Gary Gonyea at [email protected]