Appendices

Box A1. Major Elements of a Watershed-Based Strategy

• A watershed instead of political-boundary basis.
• Centralizing responsibility and authority for implementation with a municipal lead permittee working in partnership with other municipalities in the watershed as co-permittees—RC6 (Stormwater) C2, specifically C2(2) (inform and support implementation and adoption of NPDES permits).
• Embracing the full range of sources of aquatic ecosystem problems now usually under uncoordinated management and permitting; integration of all local water permits under the co-permittee system organized by watersheds—RC6 (Stormwater) C2, specifically C2(9) (implement NPDES industrial permits, WSDOT permits, DOE oversight).
• Extending full permit coverage, as appropriate, to any area in the watershed zoned or otherwise projected for development at an urban scale (e.g., more than one dwelling per acre)—RC6 (Stormwater) C2.
• Comprehensively covering all stages of urbanization: construction, new development, redevelopment, retrofit—RC1 (Land Protection) A2, specifically A2.2.8 (develop incentives to increase and improve redevelopment within UGSs); RC6 (Stormwater) C2, specifically C2(6) (retrofit stormwater systems).
• Adopting a minimum goal in every watershed to avoid any further loss or degradation of designated beneficial uses within the watershed’s component water bodies.
• Assessing water bodies that are not providing designated beneficial uses in order to set goals aimed at recovering these uses—RC1 (Land Protection) A1, specifically A1(3) (initiate and complete watershed assessments); RC2 (Flow Protection) A3.
• Defining careful, complete, and clear beneficial-use-attainment objectives to be achieved as the essential compliance endpoints.
• Concern with water quantity along with water quality—RC2 (Flow Protection) A3;
• Efficient, advanced scientific and technical watershed analysis to identify negative impact sources and set objectives and strategies—RC1 (Land Protection) A1, specifically A1(3) (initiate and complete watershed assessments); RC2 (Flow Protection) A3.
• Strategies to emphasize maximum isolation of receiving waters from impact sources; i.e. maximize application of low-impact development (LID) (retitled by the committee Aquatic Resources Conservation Design, ARCD) principles and methods—RC2 (Flow Protection) A3, specifically A3.3.2 (allow and promote rainwater harvesting) and A3 new strategies; RC6 (Stormwater) C2, specifically C2(3) (assist cities and counties in incorporating LID into all stormwater codes), C2(4) (develop and implement LID incentives), C2(6) (retrofit stormwater systems), and C2(8) (private stewardship and incentives for pollution prevention).
• Assigning municipalities more responsibility, along with more authority and funding, for the range of sources within their jurisdictions.
• Developing and appropriate allocating funding sources to enable municipalities to implement effectively—RC1 (Land Protection) A2, specifically A2(5) and A2(8) (both funding and technical assistance).
• A monitoring system composed of direct measures to assess compliance and progress toward achieving objectives and diagnosing reasons for the ability or failure to meet objectives, along with a research component to address information gaps—RC6 (Stormwater) C2, specifically C2(1) (establish regional coordinated monitoring program for stormwater under NPDES).
• Organizing consortia of agencies to design and conduct monitoring programs—RC6
• (Stormwater) C2, specifically C2(1) (establish regional coordinated monitoring program for stormwater under NPDES).
• An adaptive management framework to apply monitoring results and make early course corrections toward meeting goals and objectives, if necessary.
• A system of in lieu fees and trading credits to compensate for legitimate inability to meet requirements on-site by supporting equivalent effort elsewhere within the same watershed.

In addition to the Results Chain strategies denoted in the list, the NRC committee’s recommended program could serve as a framework to promote strategies RC1 (Land Protection) A1, specifically A1(1) (convene regional planning forum for coordinated vision), and RC2 (Flow Protection) A3, specifically A3.2 (reform state water laws). Implementation of other Results Chain strategies probably could also benefit, although perhaps less directly, from the recommendations in the NRC (2009) report.

Horner, May and Livingston (2003) put forward the following recommendations based on their data and the trends signified within them:

1. Systematically collect data on regionally representative stream benthic macroinvertebrate and fish communities. Extend the program’s coverage over the full range of urbanization. Use the data to develop regionally appropriate biological community indices.

2. Develop a geographic information system to organize and analyze watershed land use and land cover (LULC) data. Collect data on regionally appropriate LULC variables, particularly measures of impervious and forested cover in the watershed as a whole, at least two riparian bands extending to points relatively near and far from the stream, and in other local areas fairly close to the stream.

3. Base stream watershed management on specific objectives tied to desired biological outcomes.

4. If the objective is to retain an existing levels of stream function, very broadly preserve the extensive watershed and riparian natural vegetation and soil cover almost certainly present through mechanisms like outright purchase, conservation easements, transfer of development rights, etc.

5. If the objective is to prevent further degradation when partially developed areas urbanize more, maximize protection of existing natural vegetation and soil cover in areas closest to the stream, especially in the nearest riparian band. In the uplands, generally develop in locations already missing characteristic natural vegetation. As much as possible, preserve existing natural cover and limit conversion to impervious surfaces. The lower the level of existing development, the more important it is to protect existing natural vegetation and soil cover

6. In addition, fully serve newly developing and redeveloping areas with stormwater quantity and quality control best management practices (BMPs) sited, designed, and operated at state-of-the-art levels. Attempt to retrofit these BMPs in existing developments. The higher the level of existing development, the more important it is to control stormwater, since extensive land conversion results in the loss of natural vegetation and soil cover..

7. Where riparian areas have been degraded by encroachment, crossings, or loss of mature, natural vegetation, give high priority to restoring them to extensive, unbroken, well vegetated zones. This strategy could be the most effective, as well as the easiest, step toward improving degraded stream habitat and biology. Riparian areas are more likely to be free of structures than upland areas and more directly influence stream ecology. Also, riparian restoration fits well with other objectives, like flood protection and provision of wildlife corridors and open space.

Booth et al. (2001) interpreted their results to devise explicit strategies for protecting and restoring Puget Sound’s tributary streams, starting with a set of general strategies applying over the gradient of urbanization:

1. Recognize and preserve high-quality, low-development watershed areas.

2. Aggressively (and completely) rehabilitate streams where recovery of ecosystem elements and processes is possible. This condition is likely to be met only in low-development areas that happen to have relatively low to moderate levels of ecological health, because the agents of degradation are probably easier to identify and more amenable to correction.

3. Rehabilitate selected elements of mid-range urban watersheds, where complete recovery is not feasible but where well-selected efforts may yield direct improvement, particularly in areas of public ownership.

4. Improve the most degraded streams by first analyzing the acute cause(s) of degradation, but recognize that the restoration potential for populations of original in-stream biota is minimal.

5. In the most highly developed watersheds, education and/or community outreach is not just appropriate but crucial. Here, the level of public interest is likely to be highest, stream-side residents have greater direct individual influence over whether healthy stream conditions are maintained, and most of the riparian corridor is not under public ownership or control.

Booth et al. (2001) went on to offer specific recommendations for rehabilitation efforts:

1. Make direct, systematic, and comprehensive evaluation of stream conditions in areas of low to moderate development.

2. Recognize that the hydrologic consequences of urban development cannot be reversed without extensive redevelopment of urban areas. Likewise, the recovery of physical and biological conditions of streams is infeasible without hydrologic restoration over a large fraction of the watershed land area. This conflict can be resolved only if there are particular, ecologically relevant characteristics of stream flow patterns that can be managed in urban areas. Effective hydrologic mitigation will require approaches that can: (1) delay the timing of storm-flow discharges in relatively small storms, and (2) store significant volumes of rain for at least days or weeks. In the long run the goal should be to mimic the hydrologic responses across the hydrograph and not just truncate the high or low flow components. The rate of rise and decline of the hydrograph is just as important as the existence of peaks and lows. This approach almost certainly requires greater reliance on on-site storage to better emulate the hydrologic regime of undisturbed watersheds, either through dispersed infiltration, on-site detention, or forest preservation.

3. Where overall basin development is low to moderate, natural riparian corridors have significant potential to maintain or improve biological condition. Protecting high quality wetland and riparian areas that persist in less developed basins may also serve as a source of colonists (e.g., plants, invertebrates, fish) to other local streams that are subject to informed restoration efforts. At the same time, even small patches of urban land conversion in riparian areas can severely degrade local stream biology. As both a conservation and restoration strategy, protection and revegetation of riparian areas is critical for preventing severe stream degradation, but these measures alone are not adequate to maintain ecosystem function in streams draining highly urban basins.

Table B1 presents, in four categories, the elements of strategies drawn from the recommendations developed from the two large research projects on Puget Sound watersheds and streams (see above for fuller descriptions). It gives general notes regarding estimating the probable effectiveness and relative certainties associated with major strategies and references to sources of more information. Table B1 also relates the various strategies given here with those in the Results Chain memo. The strategies in Table B1 address multiple threats to the Puget Sound ecosystem, including stream channel hydromodification, salmon spawning and rearing habitat degradation, stream food web disruption, acute and chronic toxicity effects on aquatic organisms from metal and organic pollutants and increased pollutant loadings to all downstream waters, including Puget Sound.

Table B1. Strategies for Watershed Management to Protect and Restore Puget Sound’s Stream Tributaries

Category	Strategies	Notes	Associated Results Chain Strategies
Database development	Stream biological communities Land use and cover		RC1-A1(3); RC4-A1(3)
Establish objectives for an integrated approach	For streams with high biological integrity     For streams with reduced biological integrity	Appropriate objectives would be to retain existing WCI and/or forest cover and EIA balance. Appropriate objectives would be to retain existing WCI or recover a selected WCI and/or forest cover and EIA balance.	RC6-C2
Manage watersheds of streams with high biological integrity	Preserve existing watershed and riparian vegetation and soil cover through land use purchase, planning, and regulatory mechanisms.	Estimate effectiveness and relative certainty according to the data and methods presented by Horner, May, and Livingston (2003) and Booth, Harley, and Jackson (2002).	RC1-A1, A2, A3, A4, A4(6)
Manage watersheds of streams with reduced biological integrity	Maximize protection of existing vegetation and soil closest to the stream. Restore riparian areas to extensive, unbroken, well vegetated zones. Emphasize development in already disturbed locations. Serve newly developing and redeveloping areas with state-of-the-art stormwater quantity and quality control BMPs, especially low-impact development types. Retrofit existing development with these BMPs. Perform in-stream rehabilitation as appropriate to watershed conditions. Conduct watershed resident education.	Estimate effectiveness and relative certainty according to the data and methods presented by Horner, May, and Livingston (2003) and Booth, Harley, and Jackson (2002).     Refer to stormwater management segment of this chapter below for information on effectiveness and relative certainty.           Refer to stream restoration segment of this chapter below for information on effectiveness and relative certainty.	RC4-B1, B1(1), B1(3)             RC2-A3.3.2, A3 new strategies; RC6-C2(3), C2(4), C2(6), C2(8)             RC4-B1(3)

Water level fluctuation (WLF) was computed as the difference between crest stage and average base stage. Crest stage was determined with a crest-stage gauge, which records the maximum stage in a time period through the deposition level of cork dust on a plastic tube within a pipe housing. Average base stage was calculated as the mean of the stage at the beginning and end of the time period. WLF statistics were compute over extended time intervals involving a number of separate determinations. Table C1 depicts the relationship calculated by Chin (1996) and (Horner et al. 2001) between mean annual WLF and watershed TIA. Clearly, the two variables are not independent, as installation of impervious cover often accompanies removal of forest. Loss of watershed forest cover has been shown to be an important factor driving increases in WLF (Reinelt and Taylor 2001).

Table C1. Relationship Between Mean Annual Water Level Fluctuation (WLF) and Watershed Total Impervious Area (TIA) (after Chin 1996, Horner et al. 2001)

Mean Annual WLF Was:	If TIA Was:	Cases Where True:
< 20 cm	< 6%	100%
> 20 cm	> 21%	89%
> 30 cm	> 21%	50%
> 30 cm	> 40%	75%
> 50 cm	> 40%	50%

Box D1. Algal biomass control techniques from Cook et al. (2005).

• Nutrient diversion (removal or treatment of direct external inputs);
• Protection from diffuse nutrient sources (e.g., urban, agricultural, and forestry stormwater runoff);
• Dilution (to reduce nutrient concentrations) and flushing (to increase water exchange rate and consequent algal cell washout);
• Hypolimnetic (lower thermal layer) withdrawal (to discharge nutrient-rich water resulting from sediment release in the low-oxygen environment of thermal stratification);
• Phosphorus inactivation (precipitation by aluminum salt addition) and sediment oxidation (calcium nitrate injection to stimulate denitrification and oxidize organic matter);
• Biomanipulation (managing other trophic levels [zooplankton, fish] to control algae); and
• Copper sulfate (algicide) addition.

Macrophyte control mechanisms covered by Cooke et al. (2005) are:

• Restoring desirable plants to replace undesirable ones;
• Water level drawdown (to desiccate undesirables);
• Preventing invasion and physically removing undesirables;
• Sediment covers and surface shading
• Chemical controls; and
• Biological controls (insects, fish, other).

Three methods convey multiple benefits:

• Hypolimnetic aeration and oxygenation (to raise oxygen content and open habitat to cold-water fish; also to reduce sediment phosphorus release);
• Artificial circulation (use pumps, jets, or diffused air for the same purposes, plus move algal cells out of the lighted zone); and
• Sediment removal (for deepening, nutrient control, toxic substances removal, and/or rooted macrophyte control).

From the NRC report (2009, p405-406):

In water bodies that are not in attainment of designated uses, it is likely that the physical stresses and pollutants responsible for the loss of beneficial uses will have to be decreased, especially as human occupancy of watersheds increases. Reducing stresses, in turn, entails mitigative management actions at every life stage of urban development: (1) during construction when disturbing soils and introducing other contaminants associated with building; (2) after new developments on Greenfields are established and through all the years of their existence; (3) when any already developed property is redeveloped; and (4) through retrofitting static existing development. Most management heretofore has concentrated on the first two of those life stages.

 

The proposed approach recognizes three broad stages of urban development requiring different strategies: new development, redevelopment, and existing development. New development means building on land either never before covered with human structures or in prior agricultural or silvicultural use relatively lightly developed with structures and pavements (i.e., Greenfields development). Redevelopment refers to fully or partially rebuilding on a site already in urban land use; there are significant opportunities for bringing protective measures to these areas where none previously existed. The term existing development means built urban land not changing through redevelopment; retrofitting these areas will require that permittees operate creatively. What is meant by redevelopment requires some elaboration. Regulations already in force typically provide some threshold above which stormwater management requirements are specified for the redeveloped site.

 

All urban areas are redeveloped at some rate, generally slowly (e.g., roughly one or at most a few percent per annum) but still providing an opportunity to ameliorate aquatic resource problems over time. Extending stormwater requirements to redeveloping property also gradually “levels the playing field” with new developments subject to the requirements. … Some jurisdictions offer exemptions from stormwater management requirements to stimulate desired economic activities or realize social benefits. Such exemptions should be considered very carefully with respect to firm criteria designed to weigh the relative socioeconomic and environmental benefits, to prevent abuses, to gauge just how instrumental the exemption is to gaining the socioeconomic benefits, and to compensate through a trading mechanism as necessary to achieve set aquatic resource objectives.

 

It is important to mention that not only residential and commercial properties are redeveloped, but also streets and highways are periodically rebuilt. Highways have been documented to have stormwater runoff higher than other urban land uses in the concentrations and mass loadings of solids, metals, and some forms of nutrients (Burton and Pitt, 2002; Pitt et al., 2004; Shaver et al., 2007). Redevelopment of transportation corridors must be taken as an opportunity to install storm-water control measures (SCM) effective in reducing these pollutants.

 

Opportunities to apply SCMs are obviously greatest at the new development stage, somewhat less but still present in redevelopment, but most limited when land use is not changing (i.e., existing development). Still, it is extremely important to utilize all readily available opportunities and develop others in static urban areas, because compromised beneficial uses are function of the development in place, not what has yet to occur. Often, possibly even most of the time, to meet watershed objectives it will be necessary to retrofit a substantial amount of the existing development with SCMs. To further progress in this overlooked but crucial area, the Center for Watershed Protection issued a practical Urban Stormwater Retrofit Practices manual (Schueler et al., 2007).

From the NRC (2009) report:

All of the principles discussed above apply to industrial and construction sites as well: minimize the quantity of surface runoff and pollutants generated in the first place, or act to minimize what is exported off the site. Unfortunately, construction site stormwater now is managed all too often using sediment barriers (e.g., silt fences and gravel bags) and sedimentation ponds, none of which are very effective in preventing sediment transport. Much better procedures would involve improved construction site planning and management, backed up by effective erosion controls, preventing soil loss in the first place, which might be thought of as ARCD for the construction phase of development. Just as ARCD for the finished site would seek to avoid discharge volume and pollutant mass loading increase above pre-development levels, the goal of improved construction would be to avoid or severely limit the release of eroded sediments and other pollutants from the construction site.

Other industrial sites are faced with some additional challenges. First, industrial sites usually have less landscaping potentially available for land-based treatments. Their discharges are often more contaminated and carry greater risk to groundwater. On the other hand, industrial operations are amenable to a variety of source control options that can completely break the contact between pollutants and rainfall and runoff. Moving operations indoors or roofing outdoor material handling and processing areas can transform a high-risk situation to a no-risk one. It is recommended that industrial permits strongly emphasize source control (e.g., pollution prevention) as the first priority and the remaining ARCD measures as secondary options. Together these measures would attempt to avoid, or minimize to the extent possible, any discharge of stormwater that has contacted industrial sources.

It is likely that the remaining discharges that emanate from an industrial site will often require treatment and, if relatively highly contaminated, very efficient treatment to meet watershed objectives. Some industrial stormwater runoff carries pollutant concentrations that are orders of magnitude higher than now prevailing water quality standards. In these cases meeting watershed objectives may require providing active treatment, which refers to applying specifically engineered physicochemical mechanisms to reduce pollutant concentrations to reliably low levels (as opposed to the passive forms of treatment usually given stormwater, such as ponds, biofiltration, and sand filters). Examples now in the early stages of application to stormwater include chemical coagulation and precipitation, ion exchange, electrocoagulation, and filtration enhanced in various ways. These practices are undeniably more expensive than source controls and other ARCD options and traditional passive treatments. If they must be used at all, it is to the advantage of all parties that costs be lowered by decreasing contaminated waste stream throughput rates to the absolute minimum.

 

Pollutantb	Detention Pondsc	Wet Ponds	Treatment Wetlands	Biofiltersc	Media Filtersc	Hydrodyn. Devicesc
TSS	31 (16-46)	13 (7-19)	18 (9-26)	24 (15-33)	16 (10-22)	38 (21-54)
T N	2.72 (1.81-3.63)	1.43 (1.17-1.68)	1.15 (0.82-1.62)	0.78 (0.53-1.03)	0.76 (0.62-0.89)	2.01 (1.37-2.65)
T P	190 (120-270)	120 (90-160)	140 (40-240)	340 (260-410)	140 (110-160)	260 (120-480)
D P	120 (70-180)	80 (40-110)	170 (30-310)	440 (210-670)	90 (70-110)	90 (40-130)
T Cu	12.1 (5.4-18.8)	6.4 (4.7-8.0)	4.2 (0.6-7.8)	10.7 (7.7-13.7)	10.2 (8.2-12.3)	14.2 (8.3-20.0)
T Zn	60 (21-100)	29 (21-38)	31 (13-67)	40 (28-52)	38 (17-58)	80 (53-107)
T Pb	15.8 (4.7-26.9)	5.3 (1.6-9.0)	3.3 (2.3-4.2)	6.7 (2.8-10.6)	3.8 (1.1-6.4)	10.6 (4.3-16.9)
D Cu	7.4 (3.3-11.5)	4.3 (3.7-5.7)	No data	8.4 (5.7-11.5)	9.0 (7.3-10.7)	13.9 (4.4-23.4)
D Zn	26 (11-41)	33 (18-48)	No data	25 (19-32)	51 (29-73)	42 (10-75)
D Pb	2.1 (0.9-3.2)	2.5 (1.6-9.0)	0.9 (0.85-0.89)	2.0 (1.3-2.7)	1.2 (0.8-1.6)	3.3 (2.2-4.5)

REMOVAL OF FECAL COLIFORMS FROM STORMWATER RUNOFF:

A LITERATURE REVIEW

Report to City of Blaine

By

Richard R. Horner 230 NW 55th Street Seattle, Washington 98107

(206) 782-7400 rrhorner@msn.com

Tarelle Osborn Osborn Consulting, Inc. 10631 NE 16th Street Bellevue, Washington 98004

(425) 451-4009 tarelle@osbornconsulting.com

November 2005

INTRODUCTION

SCOPE OF REVIEW

Urban stormwater runoff is a widely recognized source of shellfish contamination by potential disease-causing organisms, which can lead to the closure of beds to harvest for human consumption. The literature search was intended to provide a current portrait of management options to reduce shellfish bed pathogen contamination problems associated with urban stormwater. More specifically, it concentrated on stormwater treatment methods that could be investigated further by the City of Blaine, Washington to protect shellfish harvest areas in adjacent marine waters.

The review encompassed exploring scientific and technical research databases provided by University of Washington Libraries, as well as the Internet using Google. Research databases accessed included Environmental Engineering Abstracts, Water Resources Abstracts, National Technical Information Service (NTIS), and U.S. Environmental Protection Agency (EPA) Publications Online.

The principal keyword used in the search was “fecal coliforms”, because of the prominence of this contamination indicator group in assessing shellfish bed status. The investigation did not use the broader categories “bacteria” and “pathogens” or specific microorganisms; but items reported in these terms were collected if they appeared to be relevant. “Stormwater”, “treatment”, “removal”, and “reduction” were used as secondary delimiters when necessary to narrow the inquiry to items of most direct interest.

BACKGROUND

Fecal coliforms represent a group of bacteria that have long been used as indicators of contamination by a whole host of potentially disease-causing microorganisms. Their popularity is mainly because: (1) they are relatively easy and inexpensive to measure; and (2) they have an association, and sometimes a demonstrated statistical correlation, with pathogenic organisms (Kadlec and Knight 1996).

The use of fecal coliforms (FCs) to indicate possible disease agents is not a perfect solution for several reasons. They originate from the intestinal tracts of all warm-blooded animals, and thus do not necessarily indicate human disease potential. Virulent pathogens, especially viruses, can be present even with relatively low FCs or absent with comparatively high values. Furthermore, FCs are very dynamic and responsive to a number of variables in the natural environment, such as temperature, growth substrate, and the kinetic energy of flow or currents. Nevertheless, no feasible replacement for routine monitoring has emerged, and FCs are the most common basis for regulating and managing aquatic resources. Because of this standard, FCs were taken as the basis for this literature review.

FCs fit within the broader group termed total coliforms, some of which have sources other than animal intestines (e.g., natural soils). The Escherichia coli, a subset of the fecal coliforms, are sometimes used as an alternative indicator, especially outside the United States. Other bacterial groups that have served this purpose include the enterococci and fecal streptococci. This review concentrates on FCs because of their general prominence as an index of pathogen contamination and their specific importance in water quality management within Drayton Harbor and the City of Blaine itself. In addition, the City of Blaine has acquired historical FC data within Drayton Harbor that can be used for monitoring the effectiveness of stormwater best management practices (BMPs) that are developed in the future. Data on other indicators are reported when they appear in the references consulted for information on FCs.

Evidencing their variability, FCs in urban stormwater runoff can range over a number of orders of magnitude. They most commonly fall into a range of about 10²-10⁴ colonies/100 mL (henceforth to be abbreviated as n/100 mL). However, values of <10 and as high as ~10⁶/100 mL are not uncommon (Schueler 1999a). Relatively high values are usually associated with a sewage release through an event like septic system failure, sanitary sewer overflow, or illicit connection (Pitt 1998). The mean in wide-ranging large data sets has been reported as approximately 15,000-20,000/100 mL (Pitt 1998, Schueler 1999a).

To protect shellfish harvesting in the State of Washington, Chapter 173-201A WAC requires that the geometric mean of FC readings in shellfish waters not exceed 14/100 mL, with no more than 10 percent of the measurements surpassing 43/100 mL. It is clear that to meet these criteria, typical concentrations in urban stormwater must be greatly decreased in almost any case, perhaps only excepting the most expansive and well flushed receiving waters. Reduction of mean concentration by 99 percent would still leave FCs at 150-200/100 mL, an order of magnitude higher than a 14/100 mL target. Therefore, FCs in typical urban stormwater must be reduced by source control, treatment, or both to levels more like 99.9 percent to assure protection of shellfish resources.

FCs generally fall in the range 10⁵-10⁷/100 mL in municipal wastewater effluents following both primary and secondary treatments but before disinfection. The distinction in concentrations between stormwater and wastewater is important, because the efficiency of reduction (percent removal) in a treatment system depends in part on the influent concentration; i.e., a higher efficiency in percentage terms is frequently registered with a “dirtier” than a “cleaner” influent. This phenomenon has been widely observed in stormwater treatment systems for various contaminants. What is also often seen is that the ultimate effluent quality produced by a treatment is comparable with varying influent concentrations and efficiencies. Therefore, effectiveness of a stormwater best management practice should be gauged in terms of both efficiency and consistently produced effluent quality.

Information is more scarce for FCs and other bacteriological measures than for other common stormwater contaminants, both for initial and treated runoff quality. This scarcity is due primarily to the relatively short holding time before microbiological analyses must start (6 hours) and the need to disinfect any surfaces that a sample contacts during monitoring. This period is shorter than typical full storm lengths, especially in the Pacific Northwest (the mean wet season length is 21 hours in Seattle). It would be very difficult to disinfect all of the tubing and surfaces in automatic sampling equipment. It is therefore virtually impossible to generate full storm composite samples for FC analysis. Monitoring must rely on a single grab sample, which is an unlikely representative of the overall event, or a series of burdensome grab samples taken throughout the storm and composited in relation to simultaneous flow measurements. The relative variability of FCs, and their consequent high statistical variance, also impedes obtaining data from which decisive conclusions can be drawn.

Broadly speaking, the bacterial content of stormwater runoff can be restricted by source controls, treatment BMPs, or both. Source controls are means of preventing contact between contaminants and rainfall or runoff. Hence, they are preventive practices; if there is complete lack of contact, they are 100 percent effective. Treatment BMPs are engineered devices intended to remove pollutants after they have already entered runoff. The principal types are constructed wetlands, ponds of various configurations, swales or surfaces that expose pollutants to vegetation and soil where pollutant removal mechanisms operate, and media filters. It is impossible according to inviolate physical laws to recapture all substances once released. Therefore, treatment BMPs having a surface discharge are never 100 percent effective in preventing delivery of pollutants to the receiving water.

The most common BMP investigated for bacteria reduction is some form of constructed wetland, with ponds being second in frequency. Both of these treatment systems have extended residence times, generally some days in length. The entering and exiting water streams are thus from different storms. Nevertheless, many studies compare influent and effluent quality without accounting for this fact. This failing is particularly evident in bacteria sampling because of the near impossibility of compositing samples from different points in time.

This literature review considers these data collection issues and interprets the utility of the results accordingly. Caution is applied when reporting results gained through incomplete sampling or from theoretical considerations with no or insufficient empirical demonstration.

EFFECTIVENESS OF STORMWATER BMPS IN FECAL COLIFORM REDUCTION

PRE-2000 EXPERIENCE

Schueler (1999b) summarized the experience in treating stormwater for FC reduction through the late 1990s. He covered sources, removal mechanisms, BMP treatment abilities, and recommendations for improving the quality of discharges to receiving waters from the pathogen standpoint. This review draws mostly from the last two topics. Schueler’s summary was based on 24 performance studies representing 10 stormwater ponds, nine sand filters, and five biofiltration swales. Most, but not all, focused on fecal coliforms, and grab sampling was the usual monitoring technique.

In Schueler’s database mean pond efficiency for FCs was 65 percent (range –5 to 98 percent). The corresponding figures for sand filters were mean 50 percent and a range of –68 to 97 percent). Swales generally discharged higher FC concentrations than entered (mean removal –58 percent). Pet wastes and in situ multiplication of bacteria were cited as the primary reason for poor swale performance. Schueler also reported effluent concentrations, with the means being 5144/100 mL for ponds, 5899/100 mL for sand filters, and 2506/100 mL for swales. It is apparent that influent concentrations were generally lower in the few swale studies than in the more numerous accounts for the other two BMPs.

The results indicate that ponds and sand filters can reduce stormwater bacterial contamination but not in a consistent and reliable manner. Effluent concentrations were still of the order 10³/100 mL, much higher than shellfish water quality criteria (~10¹/100 mL). It is true that dilution could lower the concentration sufficiently to meet criteria, but on the other hand it is also true that continuing large inputs of viable organisms would form a basis for sustaining a reproducing bacterial community in the receiving water.

Schueler concluded with a number of recommendations to improve performance. They included BMP structural modifications but highlighted source controls as means to prevent contaminant introduction in the first place. It would appear to be unlikely that effluents could be improved to the ~10¹/100 mL levels with structural fixes of these conventional BMPs alone. If this level is to be reached, some combination of highly effective source controls and advanced treatment BMPs will be needed.

The experience in using constructed wetlands to treat domestic wastewater can offer some insights applicable to stormwater. Kadlec and Knight (1996) covered all aspects of that topic following an intensive period of research on the subject. They summarized 21 studies in which fecal coliforms were measured before any disinfection. Reduction efficiencies ranged from <0 to 99.9 percent, 67 percent above 95 percent. However, the great majority of effluent concentrations were still of the order 10²/100 mL, including all but one case in the group having efficiency exceeding 95 percent. The authors concluded that outflow concentrations cannot be reduced to near zero without disinfection, if the wetland is open to wildlife. More specifically, they declared it technically infeasible to achieve FC consistently <500/100 mL in this situation.

DEVELOPMENTS SINCE 2000

Introduction

Since Schueler’s report some additional studies were performed on a variety of BMPs. Constructed wetlands were most commonly investigated in recent years, with the realization that chemicals exuded by plants could be bactericides. Other conventional stormwater BMPs receiving attention were ponds, media filters, vegetated filter strips and swales, and infiltration. There was limited reporting on stormwater disinfection by ultraviolet light.

This review covers each of the types, with the exception of infiltration. If suitable soils and hydrogeologic conditions allow infiltration, it can reduce pollutant inputs to surface waters by 100 percent. However, these conditions are unlikely to be prevalent in Blaine because of the predominance of glacial till soils.

Commercial enterprises have introduced a variety of proprietary BMPs to the market in recent years. The literature reports the success in FC reduction of three types: StormFilter, a media filter; StormTreat, a packaged wetland system, and the Stormceptor and Vortechnics devices, which employ hydrodynamic mechanisms for removing particles by centrifugal or centripetal force.

Constructed Wetlands

Australian researchers studied the bacteria reduction performance of a stormwater constructed wetland, as well as a wet pond (Davies and Bavor 2000; Bavor, Davies, and Sakadevan 2001). The wetland was elongated relative to its width (length:width ratio approximately 7:1) and was planted extensively with Phragmites australis. Discrete (presumably, grab) inflow and outflow samples were collected weekly. Mean removal efficiencies for FCs, enterococci, and heterotrophic bacteria were 79, 85, and 87 percent, respectively, with influent concentrations of the order 10²-10⁵ for the first two organism groups and 10⁶-10⁷ for heterotrophs. The lowest effluent FC concentration was 200/100 mL, well above the Washington shellfish criterion of ≤14/100 mL as a geometric mean.

Bavor, Davies, and Sakadevan (2001) reported on settling experiments, which demonstrated that bacteria were almost exclusively associated with particles less than 2 µm in size. Others (e.g., Dale 1974) noted this tendency of microorganisms to adsorb to particles, especially the finer ones. Wong, Breen, and Somes (1999) observed that bacteria are removed from stormwater principally through sedimentation. The very small particles transporting most of the bacterial load are difficult to settle, but filtering through vegetation assists settling. Once deposited, sediment-bound bacteria still can be resuspended back into the water column through disturbance by subsequent high storm flows (Crabill et al. 1999). Good vegetation cover could again assist performance by stabilizing sediments and reducing perturbation by flow (Davies and Bavor 2000).

Relative performance of a stormwater and a wastewater wetland was compared in Sweden (Stenstrom and Carlander 2001). The stormwater wetland had a sedimentation pond, shallow vegetated zone, and denitrification pond, with an overall water residence time of 3-5 days. The wastewater wetland had two parallel pond systems providing a 7-day residence time. The sampling procedures were not described. The wastewater wetland achieved very high removal efficiencies for E. coli, FC, and Clostridium (an anaerobic spore-forming bacterium) in both warmer and cooler seasons (E. coli—99.8% May, 97.5% November; FC—99.9% May and November; Clostridium—98.7% May, 95.9% November). The researchers observed a relationship between efficiencies of bacteria and particulate reductions, indicating again bacterial transport with the solids and removal through settling. The stormwater wetland reduced only total coliforms, and those bacteria only by one order of magnitude. However, entering concentrations were already relatively low for urban runoff at 10²-10³.

The Swedish research included sediment survival studies in the stormwater wetland. It took 24-27 days and 27-53 days for 90 percent die-off of E. coli and enterococci, respectively (and much longer for Clostridium and viruses). Thus, pathogens are vulnerable to remobilization by disturbances for a relatively long time.

California Department of Transportation (Caltrans, 2004) comprehensively studied the full range of conventional treatment BMPs, including a constructed wetland, at highway, maintenance station, and park-and-ride sites. Samples for FC analysis were collected as single grabs from the influent and effluent, and removal efficiencies were not computed. Influent concentrations at the constructed wetland, which was within a freeway right of way, ranged from 2 to 50000/100 mL, and at the outlet 2 to 7000/100 mL. The majority (65 percent) of the effluent samples had concentrations of the order 10¹/100 mL. In contrast, discharge concentrations at other BMPs included in the program were 10²-10³/100 mL in the majority of cases (see reports under the headings Ponds, Media Filters, and Vegetated Filter Strips and Swales below).

Two Alaska sedimentation basin-constructed wetlands systems receiving highway runoff were monitored during the fall season without description of the sampling scheme (Nyman et al. undated). Fecal coliforms were reduced to less than 10/100mL from already low (but unreported) numbers in the influent. A risk of using constructed wetlands or ponds for treatment of FC contamination is that the open water often attracts water fowl and wildlife, ultimately increasing contamination levels. A team of California researchers found this risk to be real in a constructed saltwater marsh near Huntington Beach. They found that Talbert Marsh regularly flushes millions of gallons of bird droppings into the Pacific Ocean. The research concluded that saltwater marches should be designed to discharge at a slower rate. A slower flow rate would likely prevent most contamination, since longer exposure to salt water and sunlight kills the bacteria (Grant et al. 2001). Any open water treatment facility should be carefully designed with this risk in mind.

StormTreat, a Modular, Manufactured Constructed Wetland

StormTreat is an in-ground modular device 2.9 meters (9.5 ft) in diameter consisting of several chambers manufactured and marketed by StormTreat Systems, Inc. A series of sedimentation chambers at the entrance are constructed to skim floatables (e.g., oils) as well as settle solids. The ultimate chamber is a vegetated wetland planted in gravel, where the water enters at the root zone. StormTreat is intended to treat the first 1.27 cm (0.5 inch) of runoff from relatively small storms or the first flush of larger events. Serving very large areas or attempting to treat larger flows requires a number of parallel units and a complex distribution arrangement. In many situations the standard StormTreat design basis would not comply with the 1992 Washington Department of Ecology designated water quality design storm, the 6-month, 24-hour rainfall event, which is equivalent to approximately 1.4 inch in Blaine. This storm would produce 0.5 inch or less of runoff only if the runoff coefficient were under 0.36. The 2005 Ecology Manual requires effective treatment for 91% of the runoff volume, which is actually less than providing treatment for the 6-month, 24-hour rainfall event. For simplicity, the cursory calculations completed for this memo were based on the 1992 requirements.

Sonstrom, Clausen, and Askew (2002) conducted a thorough study of a StormTreat system treating runoff from a roof and parking lot at a commercial site in Connecticut over a 2-year period. Two parallel units served 0.27 hectare (0.67 acre). This installation could treat only the first 0.46 cm (0.18 inch) of runoff. More tanks would have been necessary to meet the standard design basis, but the site owner would not make available the needed space and budget to do so. Excess runoff bypassed and was not monitored. Therefore, this study does not portray performance in the recommended configuration but does provide data on the device’s capabilities when individual units receive the design flow.

Grab sampling of the influent and effluent of the parallel units provided 16 samples for FC analysis. The hydraulic residence time was determined to average 9 days. Accordingly, effluent concentrations were compared to influent concentrations from the preceding week. This study thus made some attempt to compensate for the usual problem of inflows and outflows being from different water volumes.

Over the full Connecticut study the influent had median FC of 12000/100 mL and a mean of 590/100 mL. The effluent mean was < 1/100 mL. The researchers estimated cumulative loading reduction of FC at 99 percent. They attributed the high degree of retention to entrapment, filtration, and die-off.

This StormTreat system was thus shown to be capable of meeting water quality criteria for shellfish at discharge. It must be recalled, though, that it treated only a fraction of the runoff generated by the catchment. Assuming a runoff coefficient of 0.8 for the highly impervious site, it would have taken 13 units to meet the 1992 Washington Department of Ecology’s design criterion of treating runoff from 1.4 inch of rainfall.

Other reports of FC reduction in StormTreat systems range from 83 percent (Federal Highway Administration, undated) to 97 percent (StormTreat Systems, Inc, undated). The latter report from the manufacturer’s website incorporates data from several client studies verified by a certification program for proprietary BMPs operated by the state of Massachusetts.

Ponds

The Australian research on constructed wetlands reported above also included monitoring of a wet pond (Davies and Bavor 2000; Bavor, Davies, and Sakadevan 2001). A wet pond has a permanent or semi-permanent pool in which water has a relatively long residence time for reduction of small solids and dissolved substances, differing from a constructed wetland in having less or no submerged or emergent vegetation. The Australian pond had three cells, each approximately 2.5 meters (8.2 ft) in depth, with a fringe of Typha (cattails). This pond removed little or no bacteria (efficiencies of –2.5, 23, and 22 percent for FC, enterococci, and heterotrophic bacteria, respectively). It was in a watershed undergoing construction and had a significantly higher proportion of particles smaller than 5 µm than did the catchment feeding the wetland.

Mallin et al. (2002) grab sampled the inflow and outflow from three wet ponds receiving urban runoff over a 29-month period and measured FC concentrations. The geometric means declined from 488 to 70/100 mL and 97 to 43/100 mL in two ponds (efficiencies of 86 and 56 percent, respectively) but increased from 74 to 85/100 mL in a pond receiving golf course runoff. Therefore no pond effluent would meet the Washington shellfish criterion of ≤ 14/100 mL as a geometric mean.

A report from the Virgin Islands (Anonymous, undated) recounted comparative FC measurements at the inlet and outlet of a pond through a storm (presumably with grab sampling). Eight inflow samples varied from 18 to 810/100 mL. Mean removal efficiency was 76 percent, but the median was higher at 90 percent. The geometric mean of the effluent concentrations was 41/100 mL, again above the Washington criterion.

The Caltrans (2004) research included extended-detention ponds, which held runoff for up to 72 hours. This residence time is not nearly as long as in a constructed wetland or a wet pond but does offer some enhanced settling. Influent FC concentrations ranged from 110 to 28000/100 mL. Effluents exhibited concentrations ranging from 2 to 90000/100 mL, with the majority of values being of the order 10²-10³/100 mL.

Media Filters

The Caltrans (2004) study also encompassed sand filters and a StormFilter unit, which was at a maintenance station. Sand filters were of two types: the “Austin” design, in which flow enters a sedimentation chamber at a single point and then discharges via a perforated riser pipe onto sand; and the “Delaware” design, in which sheet flow enters a sedimentation chamber along a broad flow path and then passes over a weir to the sand chamber. A StormFilter has a bank of canisters containing a filtration medium, in this case perlite-zeolite. It is manufactured and marketed by Stormwater Management, Inc. (now Stormwater360).

Sand filter influent concentrations ranged from 23 to 200000/100 mL, with effluents covering the range 2 to 50000/100 mL. The majority of effluent concentrations were of the order 10²-10³/100 mL. Flows in the StormFilter ranged from 8 to 9000/100 mL. The effluent range was 2 to 3000, with 71 percent of the values of the order 10²-10³/100 mL.

Stormwater360 believes that subsurface constructed wetlands may be the most cost-effective treatment solution for FC reduction in stormwater. Stormwater360 is in the conceptual stage of a pilot project using the StormFilter in conjunction with subsurface wetlands. This eventual pilot will be in conjunction with Stephen Lyons, Ph.D., P.E., and/or Orange County Water District (Anaheim, CA).

Vegetated Filter Strips and Swales

Casteel et al. (2005) quantified bacterial indicators of fecal contamination in stormwater before and after diversion to a natural vegetated riparian buffer adjacent to a lake in the San Francisco. Lake concentrations of E. coli, enterococci, and total coliforms were about two to three orders of magnitude (99-99.9%) lower with treatment in the buffer than levels in stormwater, presumably based on grab sampling.

The Caltrans (2004) research covered both filter strips and swales. Filter strips are broad vegetated slopes receiving sheet flow, while swales are vegetated channels flowing at some depth. Filter strips experienced inflows having FCs from 30 to 90000/100 mL and discharged 17 to 9000/100 mL. The equivalent ranges for swales were 17 to > 200000/100 mL in the inflows and 17 to > 200000/100 mL in the effluents. The majority of effluent concentrations were of the order 10²-10³/100 mL for both BMP types.

Stormwater Disinfection

The city of Encinitas, CA studied ozonation and ultraviolet (UV) processes for disinfecting stormwater runoff to protect a swimming beach (Rasmus and Weldon 2005). A preliminary paper assessment rejected ozonation on a variety of logistical, cost, and performance grounds. Monitoring of the selected UV system for three months in the fall of 2002 showed the following reductions in geometric means of daily data: total coliforms—23437 to 2/100 mL, FC—1849 to 2/100 mL, and enterococci—1563 to 2/100 mL. Therefore, UV disinfection can reliably meet water quality criteria, although with considerable difficulty and expense to treat large stormwater volumes.

Hydrodynamic Devices

Neary and Boving (2004) reported on the performance of a Vortechs Stormwater Treatment System, a product of Vortechnics, Inc. (now Stormwater360). Flow enters the unit tangentially to a grit chamber, which promotes a swirling motion driving particles toward the center, where velocities are lowest and some settling occurs. Water then passes under a baffle to separate floatables. Flows above the design quantity bypass the unit. The authors did not describe the sampling procedure for FCs. Their removal ranged from 50% to 88% during three spring sampling events.

Other reports on Vortechs are less encouraging. The net removal was negative as reported in two studies by Clausen et al. (2002) and West et al. (2001).

Stormceptor is another commercial hydrodynamic device from the Stormceptor Group of Companies. Stormwater flows into an upper bypass chamber, where a weir and orifice assembly diverts flows less than the design rate into a lower treatment chamber. Velocity slows when water enters the treatment chamber. Here floatables rise and solids settle by gravity. From the treatment chamber, water is displaced up through a riser pipe into the bypass chamber on the downstream side of the weir for discharge. Clausen et al. (2002) and Waschbusch (1999) studied performance of Stormceptor units and found their net FC removal to be negative.

It was established above that FCs have a strong association with the smallest particles. These hydrodynamic devices have little capability of capturing relatively small particles and function well only in removing large solids like trash and the high end of the particle spectrum.

LOCAL PILOT PROJECTS AND RESEARCH

In 2003-04, the Port of Bellingham, Whatcom County Marine Resources Committee, Whatcom County, City of Blaine and the Drayton Harbor Shellfish Advisory Committee, through a cooperative effort, researched and developed stormwater treatment management practices to reduce bacterial pollution in Blaine Harbor, specifically near the Blaine Marina. The effort resulted in two pilot projects that were developed and implemented; installation of spiders on the breakwater to discourage seagull and pigeon roosting, and stormwater planters at the downspouts of the webhouse roof in the Blaine Marina (Landau Associates, Inc. 2004).

The stormwater planters were designed to "filter" the water for fecal coliform bacteria and other pollutants before the runoff drains into the marina waters. The rain water running off the roof of Webhouse 1 has very high concentrations of these bacteria, likely from the rain washing bird droppings left by the many gulls that regularly roost on the webhouse roof. Rainwater from the roof's downspouts is collected in the stormwater planter where it slowly filters through plant roots, soil and sand. The fecal coliform bacteria are captured in the soils where they break down and get absorbed by the plant roots. Filtered water empties into a storm drain that carries it to the marina (Landau Associates, Inc. 2004).

The stormwater planters were installed in the spring of 2004. Because of funding limitations the planters were not installed as originally specified. In fact, the total planter area was undersized by 85-97%. The projected removal rate for the planters was 99%. Because the system was so severely undersized, several of the monitored events overflowed the planters. Not counting this event, the removal rate was 50% (Hirsch Consulting Services 2004).

Considering that the planters were extremely undersized, the removal rates appear to be promising. The planter box pilot project experienced dead vegetation, possibly as a result of over-fertilization. A key recommendation included in the monitoring report of the stormwater planters suggested specifying plants that can tolerate high organic loading (Hirsch Consulting Services 2004). This recommendation should be considered with the construction of any treatment facility that includes vegetation as part of the treatment, such as wetlands and the StormTreat system. Using the stormwater planter technology on a much larger scale may be a feasible option within the City of Blaine.

SUMMARY AND CONCLUSIONS

Urban stormwater runoff is a widely recognized source of shellfish contamination by potential disease-causing organisms, which can lead to the closure of beds to harvest for human consumption. The fecal coliform group of bacteria is a convenient indicator of disease potential associated with a variety of microorganisms. FCs do have several disadvantages associated with their broad range of extra-human sources, lack of uniform association with pathogens, variability, and monitoring difficulties. Nevertheless, no better alternative has yet emerged, and FCs are used as the basis for assessing shellfish bed status.

The State of Washington sets as water quality criteria for shellfish waters a geometric mean of FC readings not to exceed 14/100 mL, with no more than 10 percent of the measurements surpassing 43/100 mL. Therefore, stormwater discharge targets should be of the order 10¹/100 mL, unless great dilution of the discharge can be assured.

Two general methods exist to prevent or reduce shellfish bed contamination by urban stormwater: pollution source controls and runoff treatment. Source controls separate the points of pollution origin from contact with rainfall or runoff; if the separation is complete, they are 100 percent effective in preventing contamination. Runoff treatments attempt to remove pollutants already in runoff; they can reduce but cannot entirely prevent contamination, unless all runoff infiltrates the soil and only emerges to surface water after full pathogen die-off.

This literature review investigated commonly used urban stormwater treatment techniques: constructed wetlands, ponds, media filters, vegetated filter strips and swales, and hydrodynamic devices. It also covered the small amount of information available on stormwater disinfection.

Excluding disinfection, constructed wetlands yielded the best performance in terms of fecal coliform reduction efficiency and effluent quality. All other options reviewed, except disinfection, generally produced effluents with FC concentrations two to three orders of magnitude higher than the presumed target of ~10¹/100 mL. Ultraviolet disinfection has been shown, as would be expected, to lower concentrations below detection. While this option could receive more consideration by the City of Blaine, it is likely to prove too logistically difficult and expensive for widespread application to protect shellfish beds.

Even with constructed wetlands, effluent FC concentrations were still generally an order of magnitude above the ~10¹/100 mL target. The major exception to this observation was the StormTreat system, a modular, manufactured constructed wetland on the commercial market, which reduced influent concentrations ranging 10²-10⁴/100 mL to a mean below detection.

Kadlec and Knight (1996), in evaluating results from municipal wastewater treatment in wetlands, offered an important clue regarding why the StormTreat system can out-perform large, more naturalistic constructed wetlands in FC reduction. They concluded that constructed wetland outflow concentrations cannot consistently be reduced to near zero, or even close, without disinfection, if the wetland is open to wildlife. This point was also illustrated in the research of Grant et al (2001) on the man-made Talbert Marsh, concluding that the additional seagull droppings were a direct source of FCs in the surf zone along Huntington Beach. The StormTreat units are not conducive to wildlife occupancy or access by domestic animals. The Caltrans (2004) experience with a constructed wetland in an urban freeway right of way adds evidence supporting this conclusion. This wetland was not easily accessible or attractive to wildlife and domestic animals. It exhibited the lowest effluent concentrations among the installations reviewed, although they were still considerably above the StormTreat levels.

The StormTreat system thus has potential for serious further consideration by the City of Blaine from the performance standpoint. However, treatment of the full State of Washington water quality design storm would require multiple units on all but the smallest sites, with the attendant issues of space, hydraulics, and cost.

More broadly, the City should investigate other ways to use constructed wetland technology while excluding animals that excrete fecal coliforms. There are models of wetland configurations from the municipal treatment experience that have not been investigated enough, if at all, for stormwater treatment, particularly the subsurface-flow class of constructed wetlands. These wetlands differ from the usual type used in stormwater management and reviewed here by having an artificial growth medium in a geometrically regular, constructed chamber with the water level at or below the medium surface, and often a surrounding fence. In other words, they are built like a wastewater treatment system, with little to attract animals. In contrast, the usual stormwater constructed wetlands have open water pools and emergent plant zones, a natural soil substrate, irregular shape, and open access. In other words, they are built somewhat like a natural water body and attract at least urban animals.

Coupled with further investigation of proprietary and non-proprietary constructed wetland designs, the City of Blaine should catalogue and assess every possible source control strategy that might be used to reduce initial FC concentrations in stormwater runoff to the minimum possible. Implementing the best feasible source controls would not replace the need for treatment but would add assurance to its success.

REFERENCES

Anonymous. Undated. Summary Report on Fecal Coliform Bacteria Removal Efficiency for Stormwater Runoff BMPs in the Virgin Islands. http://water.usgs.gov/wrri/02-03grants_new/prog-compl-reports/2003VI9B.pdf.

Bavor, H.J., C.M. Davies, and K. Sakadevan. 2001. Stormwater treatment: Do constructed wetlands yield improved pollutant management performance over a detention pond system? Water Science and Technology 44(11-12):565-570.

California Department of Transportation. 2004. BMP Retrofit Pilot Program Final Report (Appendix F), CTSW-RT-01-050. California Department of Transportation, Sacramento, CA.

Casteel, M.J., G. Bartow, S.R. Taylor and P. Sweetland. 2005. Removal of Bacterial Indicators of Fecal Contamination in Urban Stormwater Using a Natural Riparian Buffer. 10th International Conference on Urban Drainage, Copenhagen/Denmark, 21-26 August 2005. http://www.lmtf.org/FoLM/Plans/Water/VistaGrande/Casteeletal_10icud_paper.PDF.

Clausen, J.C., P. Belanger, S. Board, M. Dietz, D. Phillips, and R. Sonstrom. 2002. Final Report, Stormwater Treatment Devices, Section 319 Project. Connecticut Department of Environmental Protection, Hartford, CT.

Crabill, C., R. Donald, J. Snelling, R. Foust, and G. Southam. 1999. The impact of sediment fecal coliform reservoirs on seasonal water quality in Oak Creek, Arizona. Water Research 33:2163-2171.

Dale, N.G. 1974. Bacteria in intertidal sediments: Factors related to their distribution. Limnology and Oceanography 19:509-518.

Davies, C.M. and H.J. Bavor. 2000. The fate of stormwater-associated bacteria in constructed wetland and water pollution control pond systems. Journal of Applied Microbiology 89:349-360.

Federal Highway Administration. Undated. http://www.fhwa.dot.gov/environment/ultraurb/uubmp3p9.htm.

Grant, S.B., B.F. Sanders, A.B. Boehm, J.A. Redman, J.H. Kim, R.D. Mike, A.K. Chu, C.D. Gouldin, C.D. McGee, N.A. Gardiner, B.H. Jones, J. Sveikovsky, G.V. Leipzig, and A. Brown. 2001. Generation of enterococci bacteria in a coastal saltwater marsh and its impact on surf zone water quality. Environmental Science and Technology 35: 2407–2416.

Hirsch Consulting Services. 2004. Blaine Harbor Stormwater Treatment Pilot Monitoring. Prepared for the Port of Bellingham by Hirsch Consulting Services, Lummi Island, WA.

Kadlec, R.H. and R.L. Knight. 1996. Treatment Wetlands. Lewis Publishers, Boca Raton, FL.

Landau Associates, Inc. 2004. Specification for Stormwater Filter/Planter. Prepared for the Port of Bellingham by Landau Associates, Inc., Edmonds, WA.

Mallin, M.A., S.H. Ensign, T.L. Wheeler, and D.B. Mayes. 2002. Pollutant removal efficacy of three wet detention ponds. Journal of Environmental Quality 31(2):654-660.

Neary, K.A. and T.B. Boving. 2004. Stormwater Runoff Treatment with Vortechs® Structures. Geological Society of America Annual Meeting, Denver, CO, November 7–10, 2004. http://gsa.confex.com/gsa/2004AM/finalprogram/abstract_77164.htm.

Nyman, D., C. McCauley, D. Maddux, W. Schnabel, C. Woolard, and C. Adler. Undated. Evaluation of Stormwater Treatment by Constructed Wetlands: A Study of Two Sedimentation Basin/Constructed Wetlands at Soldotna, Alaska. www.awra.org/state/alaska/ameetings/2003am/abstracts/Wed-Nyman-wtlds.doc.

Pitt, R. 1998. Epidemiology and stormwater quality management. In Stormwater Quality Management, CRC/Lewis Publishers, New York, NY.

Rasmus, J. and K. Weldon. 2005. Moonlight Beach urban runoff treatment facility: Using ultraviolet disinfection to reduce bacteria counts. Stormwater, March 2005.

Schueler, T.R. 1999a. Microbes in urban watersheds: Concentrations, sources, and pathways. Watershed Protection Techniques 3(1):554-565.

Schueler, T.R. 1999b. Microbes in urban watersheds: Ways to kill ‘em. Watershed Protection Techniques 3(1):566-574.

Sonstrom, R.W., J.C. Clausen, and D.R. Askew. 2002. Treatment of parking lot stormwater using StormTreat system. Environmental Science and Technology 36:4441-4446.

Stenstrom, T.A. and A. Carlander. 2001. Occurrence and die-off of indicator organisms in the sediment in two constructed wetlands, Water Science and Technology 44(11-12):223-230.

StormTreat Systems, Inc. Undated. http://www.stormtreat.com/.

Waschbusch, R.J. 1999. Evaluation of the Effectiveness of an Urban Stormwater Treatment Unit in Madison, Wisconsin, 1996-1999. USGS Water Resources Investigation Report 99-4195. U.S. Geological Survey, Middleton, WI.

West, T.A., J.W. Sutherland, J.A. Bloomfield, and D.W. Lake. 2001. A Study of the Effectiveness of a Vortechs Stormwater Treatment System for Removal of Total Suspended Solids and Other Pollutants in the Marine Village Watershed, Village of Lake George, New York. New York Department of Environmental Conservation, Albany, NY.

Wong, T.F.H., P.F. Breen, and N.L.G. Somes. 1999. Ponds Vs. Wetlands—Performance Considerations in Stormwater Quality Management. Proceedings of the Comprehensive Stormwater and Aquatic Ecosystem Management First South Pacific Conference, 22-26 February 1999, Auckland, New Zealand, pp. 223-231.

Following is a compilation of all strategies identified in Section 3 for the protection and restoration of watersheds and tributaries. See Appendix 4H, Table H1 for a summary of references with more information, threats addressed, and relationships with PSP’s Results Chain strategies.

Key Strategy (4A): Develop a comprehensive watershed-based management system.

Key Strategy (4B): Manage stream watersheds using a data- and objective-based approach with appropriate specific strategies for streams depending on their levels of ecological condition.

Key Strategy (4C): Synthesis of guiding principles for stream restoration

• Protect well functioning streams and their habitats, where they exist.
• Consider what actions are necessary in the contributing watershed to achieve restoration goals and objectives. Either take these actions according the Strategy 4B or, if they cannot be performed, adjust goals and objectives to what is attainable or transfer restoration activity to a location where they can.
• Identify in-stream restoration options and apply the hierarchical strategy of Roni et al. (2002) to prioritize among them. That strategy emphasizes habitat reconnection as generally the most effective and certain of in-stream strategies, where prior disconnection is among the problems. The strategy then guides a user through consideration of riparian restoration and road improvements, with in-stream structural placements to follow or occur simultaneously with any of the other actions, as appropriate.

Key Strategy (4D): Protect, restore, and create wetlands according to the known preferences and tolerances of target biological communities, particularly geomorphic, hydrological, and hydroperiod requirements.

Key Strategy (4E): Protect and restore lakes applying the established specific strategies of algal biomass and macrophyte control.

Key Strategy (4F): As the principal basis of urban stormwater management, apply Aquatic Resources Conservation Design practices in a decentralized (i.e., close to the source), integrated fashion to new developments, redevelopments, and as retrofits in existing developments as necessary to meet established protection and restoration objectives. If a full, scientifically based analysis shows that it is indeed impossible to meet objectives with these practices, employ, first, in lieu fees or trading credits or, as a second priority option, conventional stormwater management practices according to the following key strategy:

Key Strategy (4G): Employ conventional stormwater management practices when the above options do not fully meet objectives. Increase the effectiveness of conventional vegetation- and soil-based practices whenever possible by using ARCD landscaping techniques. Apply enhanced filtration, ion exchange, or a treatment train involving both in industrial situations when source controls and ARCD measures are insufficient to meet objectives.

Key Strategy (4H): Address special stormwater problems as follows A. Promote source control under a broad ARCD program by assessing ubiquitous, bioaccumulative, and/or persistent pollutants that can only be controlled well by substituting with non-polluting products and enact bans on the use of products containing those pollutants. B. Improve construction site stormwater control by prioritizing, first, construction management practices that prevent erosion and other construction pollutant problems; second, practices that minimize erosion; and, last, sediment collection after erosion has occurred. C. To counteract dispersed sources of pathogens that compromise shellfish production and other beneficial uses, implement strong source controls and treat remaining sources with subsurface-flow constructed wetlands, assuming additional research and development verifies the promise of that technique.

Key Strategy (4I): Bolster incomplete combined sewer overflow reduction programs by using ARCD techniques identified for application in that setting to decrease stormwater flows.

Key Strategy (4J): If nitrogen discharge from a municipal treatment plant must be reduced below 1 mg total nitrogen/L to remove a threat to marine dissolved oxygen resources, apply reverse osmosis tertiary treatment with highly efficient filtration as a pretreatment. If analysis demonstrates that a lesser reduction will suffice, apply membrane bioreactor treatment. Key Strategy(4K): If discharges from on-site wastewater treatment systems are a serious threat to: (1) marine dissolved oxygen resources as a result of nitrogen; or (2) shellfish production or contact recreation as a result of pathogens, assess as possible solutions: (1) construct sewers and a municipal treatment plant, with advanced treatment for nitrogen if that is the threat, to replace problem on-site systems; or (2) apply advanced on-site treatment, tested and verified to reduce the problem sufficiently to remove the threat (note: at this point more testing is required for both on-site nitrogen removal systems and small-scale disinfection).

Key Strategy (4L): Upgrade the implementation of established agricultural best management practices, especially where agricultural runoff is: (1) a eutrophication threat as a result of nitrogen (N) and/or phosphorus (P); or (2) a threat to shellfish production or contact recreation as a result of pathogens. Manage nitrogen and phosphorus in concert by: (1) employing a phosphorus index to target management of critical P source areas, generally near receiving waters; and (2) applying N-based management to all other areas. Maintenance of riparian buffers advances both facets of the strategy by keeping agricultural activities out of the potentially most critical P production area and providing a sink for N to capture the majority of it before it can enter the water.

Key Strategy (4M): Upgrade the implementation of established forestry best management practices to protect stream water quality and hydrology in the vicinity of forestry activities and minimize the delivery of pollutants from those activities to downstream receiving waters, including Puget Sound.

Key Strategy	Report Reference for Details	Principal Guidance References	Applications	Threats Addressed	Results Chain Strategies Addressed
4A	Appendix 4A, Box A1	DeBarry (2004); Heathcote (2009); NRC (2009); for forestry issues Brooks et al. (2003)	Protection, restoration	All threats originating in Puget Sound watersheds, including those of urban, agricultural, forestry, and rural residential origin	See Appendix 4A, Box A1
4B	Appendix 4B, Table B1	Booth et al. (2001); Horner, May, and Livingston (2003)	Protection (of existing level of biological integrity), restoration (to improve biological integrity from a reduced level)	Stream channel hydromodification; salmon spawning and rearing habitat degradation; stream food web disruption; acute and chronic toxicity effects on aquatic organisms from metal and organic pollutants; increased pollutant loadings to all downstream waters, including Puget Sound	See Appendix 4B, Table B1
4C	Section 4-2 discussion under the heading Effectiveness and Relative Certainty of Stream Restoration	FHWA, (2007); Montgomery et al. (2003); NRCS.   (2007a); Roni et al. (2002); Saldi-Caromile et al. (2004); Stewart-Kloster et al. (2009); WDFW.  (2003)	Restoration	Restriction of anadromous fish passage; salmon spawning and rearing habitat degradation; stream food web disruption; if watershed restoration involved, threats under Strategy 4B also addressed	RC2 A3; RC4 B1, specifically B1(1), B1(3), and B1(4)
4D	Section 4-2, discussion under the heading Effectiveness and Relative Certainty of Wetlands Management Efforts	Azous and Horner (2001); Granger et al. (2005); Mitsch and Gosselink (2007); Mitsch et al. (2009); Sheldon et al. (2005);	Protection, restoration	Threats associated with their functions, not only to their internal ecosystems but also to waters and terrestrial environments associated with them	A broad range of strategies in this column, because of association of wetlands with other waters

Here we enumerate the major tasks foreseen by the authors as needed to bring the recommended strategies to full fruition. In some cases these tasks involve research in scientific, technical, or policy arenas; i.e., a systematic inquiry into a subject to discover facts or principles. In other cases the tasks would be more developmental, in the sense of bringing a known method or process to a more advanced or effective state. These research and development (R and D) needs are aligned with the distinct strategies identified in each chapter. Please see the relevant chapter for the citations repeated here.

Research and development needs for implementation of Overarching, Large-Scale Protection and Restoration Strategies

Most of the “Synthesizing Guidance for Puget Sound Protection and Restoration Strategies” relies on basic principles of ecology or well-established scientific findings in the Puget Sound region. Nevertheless, it would be highly valuable to determine the likely gross-scale impacts on key indicators for the Puget Sound ecosystem from different allocations of population growth across the region (i.e., as opposed to the county-by-county projections used for allocating population growth under the Growth Management Act). This could potentially take advantage of the watershed characterizations currently being completed by the Washington Departments of Ecology and Fish and Wildlife, applying them across WRIAs instead of strictly within WRIAs to determine at a regional scale the highest priority locations for protection and restoration and where new development would likely have the least impact. To the extent possible, this analysis should integrate anticipated impacts of climate change, which differ in their scope and severity across the region.

The field of ecological economics asserts that, instead of attempting to calculate the “correct” value of negative or positive environmental externalities, we should act on our knowledge that zero is incorrect. Accepting this challenge, the key research and development need is a feasibility assessment of candidate taxes or fees. The Puget Sound Partnership could choose candidates from potential taxes or fees identified in the Action Agenda and Chapter 4-1.

Fully implementing the identified protection and restoration strategies for watersheds and tributaries requires a mix of scientific, technical, and institutional research and development activities, as follows.

Key Strategy: Develop a comprehensive watershed-based management system.

• Develop a municipal co-permittee system to manage an integrated set of water-based permits, with a lead permittee working in partnership with other municipalities in the watershed as co-permittees.
• Establish state and municipal partnerships by watershed to set goals and objectives for protection and restoration, according to the principles outlined in Section 4-2.
• Establish a highly professional structure to perform the scientifically and technically based watershed analyses necessary to set and achieve goals and objectives.
• Set up the legal, regulatory, and financing mechanisms as necessary to assign authority and responsibility to municipal co-permittees for achieving goals and objectives and to ensure adequate funding for doing so.
• Determine the extent of institutional and financial barriers to retrofitting watersheds with stormwater and wastewater infrastructure necessary to meet goals and objectives and how they can be overcome.
• Develop an in lieu fee and credit trading system to make it possible for development project sponsors to compensate for legitimate inability to meet requirements on-site by supporting equivalent effort elsewhere within the same watershed.
• Incorporate recommended monitoring strategies into the monitoring program development efforts proceeding separately from the Puget Sound Science Update.

Key Strategy: Manage stream watersheds using a data- and objective-based approach with appropriate specific strategies for streams depending on their levels of ecological condition.

• Develop the watershed databases necessary to perform the recommended assessments.

Key Strategy: Restore streams according to a set of following principles given in Section 4-2.

• Adapt for urban application the hierarchical strategy for prioritizing restoration developed by Roni et al. (2002).

Key Strategy: Protect, restore, and create wetlands according to the known preferences and tolerances of target biological communities, particularly geomorphic, hydrological, and hydroperiod requirements.

• Determine the barriers that have impeded the application of knowledge about preferences and tolerances of target biological communities in wetland mitigation projects and act to remove them.

Key Strategy: Protect and restore lakes applying the established specific strategies of algal biomass and macrophyte control.

• No additional R and D required.

Key Strategy: As the principal basis of urban stormwater management, apply Aquatic Resources Conservation Design (ARCD) practices in a decentralized (i.e., close to the source), integrated fashion to new developments, redevelopments, and as retrofits in existing developments as necessary to meet established protection and restoration objectives. If a full, scientifically based analysis shows that it is indeed impossible to meet objectives with these practices, employ, first, in lieu fees or trading credits or, as a second priority option, conventional stormwater management practices according to next key strategy.

• Perform research to make objective determinations of the pavement widths actually needed for streets with various service levels and other paved areas.
• Determine how best to move the construction industry to act in such a way that soil disturbance is minimized during construction.
• Perform research to determine the best techniques for maximizing evapotranspiration (ET) from ARCD facilities, and the contribution ET can make in the Puget Sound region to reducing surface runoff from developed areas.
• Perform research to determine the best soil amendment techniques (composition and quantity) for maximizing soil storage, infiltration, and ET in ARCD facilities.
• Perform research on the various permeable pavement types to determine how best to extend their life both structurally and hydrologically.
• Perform research to determine the best vegetated-roof design techniques to maximize storage and ET, and the contribution green roofs can make in the Puget Sound region to reducing surface runoff from developed areas.
• Perform research to determine how much building with full ARCD application can be allowed, starting from different levels of existing development, and still prevent deterioration of biological integrity below existing levels in waters receiving storm runoff.

Key Strategy: Employ conventional stormwater management practices when the above options do not fully meet objectives. Increase the effectiveness of conventional vegetation- and soil-based practices whenever possible by using ARCD landscaping techniques. Apply enhanced filtration, ion exchange, or a treatment train involving both in industrial situations when source controls and ARCD measures are insufficient to meet objectives.

• Perform research to determine the benefits of applying ARCD landscaping principles and methods in vegetation- and soil-based conventional stormwater facilities.

Key Strategy: Address special stormwater problems as follows:

A. Promote source control under a broad ARCD program by assessing ubiquitous, bioaccumulative, and/or persistent pollutants that can only be controlled well by substituting with non-polluting products and enact bans on the use of products containing those pollutants.

• Catalogue ubiquitous, bioaccumulative, and persistent pollutants threatening the Puget Sound ecosystem, less threatening alternatives already available, and cases where development of such alternatives is needed to make substitutions.
• Develop legal, legislative, and regulatory structures for banning threatening chemicals in relation to alternative availability.

B. Improve construction site stormwater control by prioritizing, first, construction management practices that prevent erosion and other construction pollutant problems; second, practices that minimize erosion; and, last, sediment collection after erosion has occurred.

• No additional R and D needed.

C. To counteract dispersed sources of pathogens that compromise shellfish production and other beneficial uses, implement strong source controls and treat remaining sources with subsurface-flow constructed wetlands, assuming additional research and development verifies the promise of that technique.

• Test subsurface flow wetlands, designed to exclude wildlife, for pathogen reduction in stormwater runoff and develop design and maintenance specifications that provide maximum reduction.

Key Strategy: Bolster incomplete combined sewer overflow reduction programs by using ARCD techniques identified for application in that setting to decrease stormwater flows.

• No additional R and D required.

Key Strategy: If nitrogen discharge from a municipal treatment plant must be reduced below 1 mg total nitrogen/L to remove a threat to marine dissolved oxygen resources, apply reverse osmosis tertiary treatment with highly efficient filtration as a pretreatment. If analysis demonstrates that a lesser reduction will suffice, apply membrane bioreactor treatment.

• Perform research to determine the level of municipal wastewater nitrogen reduction required to protect marine dissolved oxygen resources in specific cases.
• If reverse osmosis is required for protection in at least some cases, perform research to determine if its cost can be reduced sufficiently to improve its cost-effectiveness substantially.

Key Strategy: If discharges from on-site wastewater treatment systems are a serious threat to: (1) marine dissolved oxygen resources as a result of nitrogen; or (2) shellfish production or contact recreation as a result of pathogens, assess as possible solutions: (1) construct sewers and a municipal treatment plant, with advanced treatment for nitrogen if that is the threat, to replace problem on-site systems; or (2) apply advanced on-site treatment, tested and verified to reduce the problem sufficiently to remove the threat (note: at this point more testing is required for both on-site nitrogen removal systems and small-scale disinfection).

• Thoroughly test promising on-site nitrogen removal technologies under Puget Sound conditions to determine if such a system can reduce nitrogen sufficiently to protect marine dissolved oxygen resources in specific cases where they are threatened by on-site treatment system discharges.
• Further develop small-scale disinfection technologies to improve their cost-effectiveness.

Key Strategy: Upgrade the implementation of established agricultural best management practices, especially where agricultural runoff is: (1) a eutrophication threat as a result of nitrogen (N) and/or phosphorus (P); or (2) a threat to shellfish production or contact recreation as a result of pathogens. Manage nitrogen and phosphorus in concert by: (1) employing a phosphorus index to target management of critical P source areas, generally near receiving waters; and (2) applying N-based management to all other areas. Maintenance of riparian buffers advances both facets of the strategy by keeping agricultural activities out of the potentially most critical P production area and providing a sink for N to capture the majority of it before it can enter the water.

• Develop the framework to institutionalize this strategy in watersheds subject to the negative impacts of eutrophication and, in general, to provide more directed guidance on the full range of contaminant issues to Puget Sound agricultural concerns.

Key Strategy: Upgrade the implementation of established forestry best management practices to protect stream water quality and hydrology in the vicinity of forestry activities and minimize the delivery of pollutants from those activities to downstream receiving waters, including Puget Sound.

• Reinvigorate the Timber Fish Wildlife process to implement this strategy in a strong partnership with the Puget Sound Partnership.

Research and development needs for implementation of Marine and Estuarine Protection and Restoration Strategies 1. Expand and improve our understanding of the sources, pathways, quantities, and fate of pollutants (nutrients, pathogens and toxics) in Puget Sound estuaries and marine waters. Determine how and where they are introduced into estuaries and Puget Sound waters.

2. Determine the effects of priority pollutants on aquatic species and human health. What are the ecological effects of “legacy toxics” such as PCBs and DDT?

3. Identify adaptive mechanisms at organism, population, and community levels that buffer (i.e., reduce vulnerability and promote recovery) the deleterious effects of pollutants.

4. Improve knowledge of times and places (“hotspots”) where water quality and sediment are impaired to the point that aquatic biota and/or humans are at risk.

5. What are the times of the year and associated conditions when estuary and marine ecosystems are most at risk?

6. What physical processes affect the distribution and potency of pollutants over time and space?

7. Identify the primary processes affecting the vulnerability and resiliency of PS to perturbation.

8. What effects will climate change have on these processes in the future?

9. Identify areas where the natural and human systems are not integrated, are particularly sensitive to perturbation, or are prone to dysfunction.

10. Eliminate gaps in knowledge and/or uncertainty by conducting research, including controlled, large-scale experiments, modeling and monitoring.

11. What strategies do we recommend to deal with unexpected developments, including catastrophic events?

12. Evaluate the relative effectiveness of current regulatory programs in protecting estuaries and marine areas and mitigating the impacts of human activities.

13. Evaluate the effects of increasing human-caused variation (frequency, amplitude, rates, etc.) in physical conditions (suspended sediment, salinity, etc.) on ecological processes and components.

14. What is the “lag time” between implementation of protection and restoration measures and the expected beneficial effects? What affects the time it takes for ecosystem response and recovery?

15. Develop a comprehensive “data gaps and uncertainties” matrix; update it regularly to ensure that resources are expended where most needed.

16. What are the cumulative effects of bulkheads, docks, piers, etc.?

Research and development needs for implementation of Fisheries and Wildlife Protection and Restoration Strategies

Much research has been conducted on fish and wildlife, particularly salmon, waterflow, and marine mammals. However, in terms of protection and restoration effectiveness, there are still a number of unknowns that need to be addressed. They generally fall into the following categories:

• Dynamic relationships between habitat changes, natural variation, and species’ population ecology
• Effects of direct human disturbance on species’ behavior (e.g., cetaceans, seabirds, waterfowl)
• Lethal and chronic sub-lethal effects of known and suspected pollutants, (e.g., copper, lead, nano-toxins, surfactants, personal care products, pharmaceuticals, etc.)
• Harvest management (salmon, waterfowl, and shellfish)
• Hatchery management (genetics, competition, mixed-stock fisheries, etc.)
• Effects of ambient light and noise on fish and wildlife behavior
• Quantification of illegal and undocumented harvest