Bridge Stormwater Runoff Analysis and Treatment Options (2014)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

19 Assessment Procedure This chapter provides a description of a simplified assess- ment procedure and a more complex assessment procedure if the DOT is required to assess the impact of the bridge crossing on the receiving water as a part of the 404/401 process. The purpose of using these procedures is to provide the DOT with numeric justification for their selected mitigation program. It is important to note that the assessment procedures in this chapter will not be useful for all DOTs in all situations. There may be cases in which treatment BMPs are mandated. Examples include requirements of the DOT’s stormwater permit; the bridge crosses a water body where the implemen- tation plan of the TMDL requires BMPs; and special situa- tions involving endangered species or high-quality waters. In these cases, there may be no level of assessment sufficient to demonstrate to the regulator that a BMP will not have some level of benefit. If that is the case, the user of this guide should skip forward to Chapters 4 and 5 to identify the most effective and cost efficient combinations of BMPs. 3.1 Overview of Assessment Approach Three fundamental cases are discussed in this section: rural watersheds, urban watersheds, and special situations. The practitioner will be able to classify the project watershed and complete the assessment following the steps provided in this section and as outlined in Chapter 1. The two assessment approaches investigate the likely impact on downstream water quality based on a mass balance approach. These assessments will help explain to regulators the decision of the DOT on the level of mitigation provided for a new, renovated, or existing bridge. Both the simple and complex assessments consist of per- forming a mass balance. A detailed discussion of the terms in a mass balance equation is provided in Appendix B. For each of the two assessments described, the mass balance has been simplified by assuming that the analysis is conducted imme- diately downstream of the bridge such that all constituents are considered conservative and processes steady state. An underlying premise in this analysis is that the quality of runoff from bridges is not meaningfully different than the runoff from any other impervious area subject to traffic loads (see Section 2.3). Although bridges may discharge directly to receiving waters from deck drains or scuppers on the bridge, the impact is not materially different than if an equivalent amount of impervious cover were constructed adjacent to the receiving water and runoff piped directly to the water body. 3.1.1 Rural Areas In a rural, largely undeveloped watershed, any impairments or degraded water quality would not be the result of an iso- lated bridge, but would be associated with either natural con- ditions or human activities such as Confined Animal Feeding Operations, agriculture, and logging, which would also be the source of the vast majority of stormwater flow in the receiv- ing water. As described previously, a number of studies have been conducted to evaluate the impact of bridge runoff on receiving waters and aquatic ecosystems and none of these has documented increased environmental impairment in the area immediately downstream of a bridge as compared to the upstream condition (URS 2010; Bartelt-Hunt 2012). Conse- quently, the impact of bridge runoff in primarily rural water- sheds is de minimis unless species and site-specific studies identify a unique situation. If the level of mitigation required under a statewide stormwater permit or the 404/401 certifi- cation process does not include water quality requirements, further assessment is not warranted. 3.1.2 Urban Areas The primary factor affecting receiving water health in urbanized areas is the volume and quality of runoff from impervious surfaces in the watershed. The bridge itself is C H A P T E R 3

20 one of many small impervious parcels contributing runoff and, consequently, it is logical that it be subject to the same regulations as other impervious area. That is, the de minimis assessment approach does not apply because of the cumula- tive impact of many small impervious parcels. Since DOTs are also subject to stormwater permit requirements in urban areas, the level of mitigation for bridge runoff should be simi- lar to that required for new impervious cover anywhere in the DOT permitted area. If implementation of structural BMPs to treat runoff from new impervious cover is considered nec- essary to comply with the MEP reduction in the discharge of pollutants as implemented in the applicable NPDES or other regulatory permit, then structural BMPs should be implemented either at the bridge crossing itself or offsite (preferred for performance and cost reasons). Since the level of mitigation required is specified in the stormwater permit, an assessment of water quality impacts is likely unnecessary in this case. 3.1.3 Special Situations There may be special situations in both urban and rural watersheds where implementation of stormwater treatment is requested by regulatory authorities. This can occur as part of the 404/401 process or where the water body has special environmental constraints, such as: • TMDL watersheds • ONRW • Domestic water supply/hazardous spill control • Presence of endangered species The primary purpose of the assessment approaches described in this chapter is to address regulatory concerns in these special situations. Before undertaking any of the assessment procedures described in the following sections, it is recommended that the DOT confer with the appropriate regulatory authority to determine which of the assessment procedures the agency will accept for determining the need for bridge deck runoff treatment. 3.1.4 Summary Bridges in urban areas should include a level of storm water treatment consistent with the local definition of MEP. For many jurisdictions, MEP may be satisfied using the practices described in Chapter 4. If treatment is desired or required, the treatment may take place either onsite or offsite (preferred for cost and performance reasons). The water quality impact of bridges located in rural areas is typically de minimis and no BMPs (beyond those selected as applicable from Chapter 4) are needed. Special situations require coordination with the appropriate regulatory authority to determine which assess- ment procedures would be accepted to indicate the need or lack thereof for mitigation. Therefore, only a small number of bridges would need to have an assessment performed. 3.2 Simple Assessment Procedure The simple assessment approach will be used in the case where a regulatory agency is requesting an analysis to assess the change in constituent loading as a result of a bridge cross- ing. This approach uses dilution calculations to estimate the increase in pollutant load resulting from discharges from the bridge deck to the receiving water. A worked example is pro- vided in Appendix B. To demonstrate that the water quality impact of any par- ticular bridge is de minimis, a mass balance should be per- formed. A mass balance consists of determining the percentage of load for any specific constituent of concern contributed by the bridge. EPA has also established policy that a de minimis discharge produces no more than a 10% decrease in water quality for any given water body, and the maximum aggre- gate decrease in water quality based on multiple de minimis findings is 20% for a water body (King 2006). Taking a more conservative approach, we will consider the contribution de minimis if the bridge contributes less than one percent of the load in the receiving water downstream of the bridge. If the load from the bridge is larger than 1%, then a more com- plex assessment should be used to determine whether the impact is sufficiently large to justify including either on-site or off-site treatment controls. ( )= + ×Load Increase Bridge Load Bridge LoadUpstream Load 100 where the Load Increase is the percentage of the load down- stream of the bridge contributed by the bridge itself, Bridge Load is the load conveyed by the bridge runoff, and Upstream Load is the load in the receiving water upstream of the bridge. Bridge Load Rainfall Runoff Coefficient Area of the Bridge Deck Concentration = × × × where the Rainfall is the average annual rainfall for the specific location, the runoff coefficient is assumed to be 1.0 (a conserva- tive value, since actual runoff coefficients tend to be somewhat less), and the Concentration is the average concentration of the constituent of concern (see Table 2-1). If the constitu- ent of concern is bacteria, and there is a need to account for the contribution of birds roosting in the bridge substructure, then a value of 5.0 × 108 MPN/d/nest for cliff swallows or 5.6 × 109 MPN/d/bird for pigeons can be used in the loading calculation (Sejkora 2011).

21 Upstream Load Annual discharge of the receiving water Average Stream Concentration = × where the average flow in the receiving water is determined by gauged data at the proposed site or calculated based on a com- parison of the upstream catchment area to the flow observed at a gauged location in the vicinity with a known catchment area or computed using one of the many synthetic methods available in public domain programs such as the hydrologic engineering center hydrologic modeling system (HEC-HMS). For a lake setting, if the bridge crosses a tributary arm of the lake, the most common physical configuration, then the same basic comparison of loads may be made in which the flow in the receiving water is that originating upstream from the bridge crossing. The only difference is the greater upstream surface area of the tributary arm than the stream crossing case, due to backwater from the lake, which affects only the precipitation load term. In some regulatory situations, there may be a concern about accumulation and build-up of pol- lutants within the main body of the lake, which should be addressed with a more complex assessment. If the bridge crosses a tributary arm of an estuary, then the relative load in the tributary versus bridge runoff are, again, compared in the same way as the stream crossing, using the drainage area of the tributary arm upstream from the bridge crossing, and the flow estimated from an upstream gauge extrapolated by drainage-area ratio to the bridge crossing, by transfer of record from a nearby gauge, or by application of the rational method using rainfall data. More complex situa- tions may require estimating the additional dilution afforded by tidal exchange. In a specific situation in the stream, lake, or coastal inlet setting, additional aspects of the receiving water may need to be explicitly addressed. These more complex situations may include the need to consider a mixing zone, or the involve- ment of kinetics of the pollutants of concern. It may prove adequate to use a somewhat more complicated, but still rela- tively straightforward order-of-magnitude analysis to quantify the bridge impacts for these situations. Methods are summa- rized for each watercourse type in the following section. 3.3 Complex Assessment Procedure The complex assessment procedure is suitable to evaluate the concentration of a specific pollutant found in bridge deck runoff in the receiving water once construction or rehabilita- tion of the bridge is complete. This procedure will be useful for special conditions, such as crossings of ONRW, domestic water reservoirs, or in the case of a TMDL watershed. NCHRP Report 474 will be referenced for other methods that can fur- ther support the complex assessment approach. The complex assessment approach will be described with the option of using the US Geological Survey (USGS) Stochastic Empirical Dilution Model (SELDM) to complete the assessment. There may be cases when regulatory agencies require a more rigorous analysis than the simple determination described above of whether the bridge impact in a rural watershed is de minimis. This might occur in situations such as Outstand- ing National Waters, presences of endangered species, 303(d) listed water bodies where no additional pollutant loading is considered acceptable, or where a TMDL requires pollutant reductions, or other site-specific situations. Consequently, the following sections describe how to calculate pollutant con- centrations downstream of the bridge to determine whether water quality standards are likely to be exceeded and treatment BMP implementation should occur. These calculations evalu- ate a worst-case scenario based on the impact of a design storm occurring during a period of low flow. They require substan- tially more data and a decision on the design storm size and critical stream discharge (e.g., 7Q10) as compared to the simple assessment. A detailed derivation of the equations presented in this section is provided in Appendix B. If the specific regulatory circumstance requires still more detail or sophistication, then the approaches summarized in NCHRP Report 474 and URS (2010) are suggested. 3.3.1 Stream Environment The geometric feature that characterizes the stream environ- ment is the large ratio of watercourse length (measured along the principal axis) to width (measured cross-channel, i.e., per- pendicular to the longitudinal axis). For a channel of constant cross section a, R = Q/xa, x denoting a distance downstream from the bridge and the volume of influence V = xa. In this case, R is the reciprocal of the time of travel. For regulatory purposes, the near-field solution is of interest, because most states typi- cally allow a zone of initial dilution (ZID), particularly for those constituents considered toxic. Within the ZID, exceedance of the stream standard is allowed. The ZID may be explicitly speci- fied as, in a stream, a distance downstream from the pollutant source. Even if not explicit, as is often the case with nonpoint sources, demonstration of a concentration within standards for a distance 100 to 1,000 yards downstream will suffice. Within such a short distance, the concentration down- stream of the bridge can be calculated as: 1 1 1 (1)c c D c D u b( )= − + where c = concentration in the water body downstream of the bridge cu = concentration in the water body upstream of the bridge cb = concentration in the bridge runoff D ≡ Q/Qb

22 Equation (1) is a suitable screening test. For storm runoff from even a small single-cell storm, D is on the order of 104. An analogous equation is offered as Method 1 in NCHRP Report 474, in which it is recommended that a “worst-case” condition of a high (95% exceedance suggested) bridge run- off flow be combined with an extreme low (5% exceedance suggested) flow in the receiving stream. This is truly worst case, in that it assumes that rain falls only on the bridge. While a critical low flow may be an appropriate choice for the ambi- ent flow Qa, the storm flow does not occur in isolation on the bridge, so the other component flows, given in Equation (5) of Appendix B: Simple and Complex Assessment Methods and Worked Example, must be considered. Even if one only includes the precipitation on the surface of the stream, there is still a factor of 104 dilution of bridge runoff. The far-field problem in the river is straightforward and needs only to address the length of stream to its mouth. For most streams, this will translate to a travel time of days to weeks, so the problem devolves to selecting suitable flows to be averaged over this time period. Evaporation rates are rarely high enough to warrant inclusion (except perhaps in the southwest), but it may be necessary to retain the first-order rate K in Equations (3) or (4) of Appendix B for reactive con- stituents. A simplified version is suggested as Method 7 for the stream environment in NCHRP Report 474. 3.3.2 Lake Environment The most common instance of a bridge crossing of a lake is a traverse of an arm, typically a stream channel now immersed by backwater from the lake. Typically, however, it is the impact on the main body of the lake that is the primary concern. The defining feature of a lake is the large ratio of storage to inflow, i.e., long residence time, so the lake is a cumulative watercourse. It is not the immediate response to runoff events but rather the accumulation of constituents in the lake that potentially affect the beneficial uses of the watercourse. The time scale of analysis is therefore much longer than that of the stream environment. This might be annual flows for large lakes with quasi-permanent temperature structure, but a seasonal analy- sis, notably the summer, is frequently more appropriate for impact analysis since it targets the season of greatest biological production. Equation (2) is directly applicable, where V is the volume of the entire lake, Qs is the mean seasonal inflow, and Qa is incorporated into Qs in the averaging process. 1 (2)c t c Q c Q R K V EA e c e a a s s R K EA V t o R K EA V t( )( ) ( )= + + − − +( ) ( )− + − − + − where Qa = ambient flow in the watercourse ca = ambient concentration in the watercourse Qs = storm runoff flow into the watercourse cs = storm runoff concentration E = evaporation rate in depth of water per unit time A = surface area of the volume of influence K = first-order decay coefficient in inverse time (i.e., per unit time) The separation of Qs into components is necessary in order to isolate the effect of bridge runoff, but now each component must be estimated based upon seasonal storm occurrences, and each component is determined for the entire volume of the lake, not just the segment upstream from the bridge. For de minimis argument, it may be sufficient to consider only the comparative magnitudes of Qb and Qp, the latter being easily estimated from seasonal precipitation data. One consequence of the long integration time of a lake is the increased importance of the first-order kinetic terms K and EA/V in Equation (2). Even the concentration of a conser- vative substance, such as salts or some metals, will be affected over the long term by evaporation. The uptake of reactive constituents like nitrogen species becomes substantial over the summer production season, since metals, over time may become less biologically available as sorption processes occur. A simplified version of Equation (2) is presented as Method 7 in NCHRP Report 474 for addressing the lake environment (see also Method 2). If the residence time in the lake is less than the analysis period (e.g., a season), then the steady-state version in Equation (3) may suffice for order-of-magnitude estimates. = ε (3)Q H A PT T T 3.3.3 Coastal Inlet Environment While the coastal inlet watercourse is arguably the most com- plex of those considered, its complexity in many respects sim- plifies the estimation problem. Like the lake watercourse, it is the impact on the total volume of the coastal inlet that is of pri- mary concern, so a mass budget over a seasonal or annual time scale is appropriate, and the complex geometry of the inlet is avoided. Unlike the lake, however, the coastal inlet enjoys addi- tional sources of dilution water due to its free connection to the sea and tidal exchanges. Some of these, such as internal circula- tions driven by density differences between ocean and coastal waters, and storm-driven exchanges can be site-specific and difficult to estimate. The tide, however, is a ubiquitous marine factor whose contribution to exchange can be estimated. The flow rate entering the inlet on the flood tide (the “tidal prism”) is estimated from the product tidal range HT and the surface area of the inlet A: = ε (4)Q H A PT T T Range is the difference between the heights of high tide and low tide, and statistics on range are readily available for

23 each of the NOAA tide gauges that dot the coastline of the United States. PT is the period of the tide (or, more precisely, the dominant period), and for present purposes PT takes on one of two values, 12.4 hours for semidiurnal tides, character- istic of most of the Atlantic and Pacific coasts, and 24.8 hours for most of the gulf coast. The coefficient e is a measure of the proportion of “new” water brought into the inlet on each tidal cycle, in contrast to inlet water carried out to sea but then returned to the inlet on the next tide. Equation (3) is applied to the inlet, for which Qa is neglected, and (5)Q Q Q Qs p T b= + + where Qs = mean seasonal inflow Qp = flow from precipitation on the water surface QT = volume entering the inlet on flood tide Qb = runoff from the bridge Like the lake, the volume of influence V is the entire volume of the inlet and the period of analysis is long-term, either sea- sonal or annual. The precipitation component Qp is that for the surface area of the inlet, and QT is computed from Equa- tion (4). Residence time in a coastal inlet is typically relatively short, so the steady-state solution Equation (3) may suffice for estimation purposes. As noted above, the estuary represents an extremely impor- tant special case of the coastal inlet, whose defining characteris- tic is a source of freshwater inflow, which is some combination of riverine inflow and runoff from the surrounding drainage area. For estimation purposes, the same procedure as for the lake may be applied, except now the tidal prism flow QT is added to the separation of flow components, though for a de minimis argument, it may be sufficient to consider only Qp. A common bridge crossing configuration in an estuary is a riverine or tributary arm of the estuary. In this case, the concern may be impacts on that reach of the estuary immedi- ately downstream from the bridge. This problem is addressed exactly like the stream environment applying Equation (2) or (3) in which Qa may be retained as an ambient inflow, per- haps a critical low flow. Depending upon local circumstances, a long time scale may be used in which inflows and rainfall are averaged over an extended period of time, or, if the near- field problem is a concern, specific characteristic storms may be used to estimate the relative contribution of the compo- nent flows. Flow from the upstream drainage area Qd might be included in this estimation, but, as above, it may be suf- ficient for the de minimis argument to consider only the pre- cipitation on the water surface Qp. In marginal cases, it may be desirable to include the tidal prism component explicitly, but generally this can be omitted with the observation that an estuary has additional sources of dilution besides inflows. So their neglect is, in effect, a worst-case approximation. In more complex modeling problems, such as for waste-load alloca- tions, these are represented by a large dispersion term in the mass budget. Appendix B provides a worked example using the complex assessment approach.