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.
26 Chapter 4. Hydrologic Guidance This chapter outlines the major components of the hydrologic guidance developed through this research project. First, the chapter describes how appropriate watershed boundaries are selected for application of the watershed approach to stormwater quantity mitigation and defines the concept of assessment points (APs), one of which will be the watershed outlet. Next, the chapter describes the hydrologic impacts of transportation infrastructure and the range of potential metrics for measuring the impacts and evaluating mitigation. Subsequently, the chapter discusses how these impacts can be modeled using existing watershed modeling tools. The final two sections describe approaches for screening level and more site-specific (detailed) evaluation of mitigation techniques. 4.1. Defining the Watershed and APs Within the decision framework summarized in Figure 3.1, the State DOT will, in conjunction with potential external partners/stakeholders, define the watershed and APs in step 4. The watershed will contain the transportation project or projects under consideration and will also include the area within which out-of-kind mitigation techniques may be applied to compensate for project hydrologic impacts. One AP will be at the watershed outlet. Others may be selected at other highway/stream crossings. The number and locations of APs will depend on where it is appropriate to analyze impacts and where the fulfillment of regulatory requirements is mandated. One of the benefits of watershed-based analyses is that it allows a single watershed-wide model application to provide impact and performance information for multiple APs, and thereby improves the cost-effectiveness of the application. Two general approaches to defining a watershed are considered: 1) relative and 2) fixed. A relative watershed definition is referenced to the project location including some or all of the watershed upstream of the project and, possibly, an additional drainage area to a point on the waterway downstream of the project. A fixed watershed definition uses pre-determined watershed delineations such as hydrologic unit code (HUC) watershed boundaries. The choice of watershed definition depends on several factors including the objectives of the watershed analysis, compatibility with regulatory requirements (e.g., required APs), and available tools for assessment. 4.1.1. Relative Watershed Definition The traditional approach is to define the watershed and AP at the project location and mitigate for hydrologic impacts evaluated at that location. To fully employ a watershed-based approach, a broader view of the watershed within which the project and corresponding mitigation are located is considered. This section describes options for defining the watershed outlet using a relative watershed approach including: ⢠At the project location (traditional approach). ⢠Downstream of the project location just upstream of the confluence with another waterway. ⢠Downstream of the project location and at the confluence with another waterway of the same stream order or greater.
27 ⢠Downstream of the project location a distance corresponding to fixed limit for an increase in drainage area, for example, finding a location downstream that increases the watershed area at the project location by no more than 50 percent. Consider a project site in Oklahoma where a roadway crosses a stream shown in Figure 4.1. The pin indicates the project location that could represent a culvert or other hydraulic structure. The contributing watershed is shown in yellow. In the historical approach, the AP is at the project site and any mitigation would be required on site. In a watershed approach with this definition of watershed, the AP would also be at the site, but mitigation could be applied anywhere within the watershed area if mitigation requirements were met at the AP. This definition of watershed limits all mitigation to be upstream of the project site, which may be a relatively small area. Alternatively, the watershed could be defined by specifying the watershed outlet downstream of the project site just above the junction with another stream of the same order as shown in Figure 4.2. The drainage area to this location is greater and includes more of the highway infrastructure within the watershed rather than simply the project site. This definition of watershed provides the opportunity for placing mitigation strategies downstream of the project site as well as in upstream locations. The main AP is downstream of the site. This definition of watershed also reveals opportunities for providing mitigation strategies for more than one project simultaneously if it is within the same watershed. With a list of potential mitigation techniques (with locations), that is, the portfolio of mitigation alternatives developed in step 6 of Figure 3.1, analysts can use spatial analysis tools to identify whether these locations lie within the defined watershed. If so, the mitigation alternatives can be assessed using the screening tools according to step 8 of Figure 3.1. If no locations are found, the downstream location can be moved to the next junction increasing the size of the defined watershed. Practical limits will be required on how far to carry this approach, such as limiting the extension downstream of the project site to a 50 percent increase in area as there will be some distance where a mitigation strategy, no matter how effective, is unlikely to have a meaningful relationship to the impacts of a particular project. The analyst must consider the stream configuration and orientation of the roadway system to select meaningful APs and their associated watershed boundaries. Small changes in the location of the downstream AP may result in large changes in watershed area. For example, consider the example shown in Figure 4.3. In this case the downstream AP is defined at a road crossing of a stream with the resulting watershed shown in yellow. If the analyst moves the AP downstream to just above the confluence with a larger stream, as shown in Figure 4.4, the watershed area increases only slightly. If the analyst continues further downstream below the confluence with the larger stream, as shown in Figure 4.5, the total watershed area draining to that potential AP is much larger than shown with the previous two illustrations. Which downstream AP is most appropriate depends on the objectives of the analysis and the regulatory framework within which the analysis is conducted. In addition, the AP must be located so that it captures the effects of the project impact and mitigation effects.
28 Figure 4.1. Watershed determined by project location (from United States Geological Survey (USGS) StreamStats). Figure 4.2. Watershed determined by the junction immediately downstream from project (from USGS StreamStats).
29 Figure 4.3. Example near a confluence with a larger waterway (from USGS StreamStats). Figure 4.4. Example near a confluence with a larger waterway moved to just upstream of the next junction (from USGS StreamStats).
30 Figure 4.5. Example near a confluence with a larger waterway moved to just downstream of the next junction (from USGS StreamStats). 4.1.2. Fixed Watershed Definitions Fixed watersheds represent the other general approach for defining a watershed and uses pre- determined watershed delineation tools such as HUC watershed boundaries. One of the standard sources of watershed delineations in the United States is the Watershed Boundary Dataset (WBD) from the USGS (https://www.usgs.gov/core-science-systems/ngp/national-hydrography/watershed- boundary-dataset). The hydrologic units (HU) in the WBD form a standardized system for organizing, collecting, managing, and reporting hydrologic information for the nation. The HUs in the WBD are arranged in a nested, hierarchical system with each HU in the system identified using a unique code as shown in Figure 4.6. The USGS developed HUC using a progressive two-digit system where each successively smaller areal unit is identified by adding two digits to the identifying code. The WBD contains eight levels of progressive HU identified by unique 2- to 16-digit codes. The dataset is complete for the United States to the 12-digit hydrologic unit. The 14- and 16-digit HU are optional and are not complete for the nation. There are approximately 2400 HUC8s for the United States with a median size of roughly 1400 square miles. At the HUC10 level there are roughly 19,000, with a median size of nearly 200 square miles, and at the HUC12 level there are approximately 100,000, with a median size of 35 square miles.
31 Figure 4.6. HUC hierarchy (source: USGS). Since the HUC boundaries are an established reference, their use provides some objectivity to identify the watershed within which a project is located with the HUC outlet serving as the downstream AP. For example, if the project is located within HUC 180902030303 â Upper Marble Canyon, as shown in Figure 4.6, mitigation could be considered anywhere within that HUC12 watershed. The watershed could also be defined to include upstream HUC12 areas, such as HUC 180902030301 in the example to provide additional opportunities for locating mitigation techniques. USGS and USEPA provide another standard hydrologic unit through the National Hydrography Dataset Plus (NHDPlus) catchments (https://www.epa.gov/waterdata/nhdplus-national- hydrography-dataset-plus). Combined with the National Elevation Dataset (NED), the NHDPlus provides watershed boundaries at the catchment level and drainage areas of each NHD stream segment. These catchments vary in size but on average tend to be around a square mile. Since these NHDPlus catchments are readily available, they could be used as the basis for defining a watershed area for analysis of project impacts and mitigation effectiveness.
32 4.1.3. Tools for Defining a Watershed Regardless of how the outlet of a watershed is determined, various geographic information system (GIS)-based tools are available to delineate a watershed using digital elevation data. Some examples include ArcHydro from ESRI, TauDEM from Utah State University [used in EPAâs Better Assessment Science Integrating Point and Nonpoint Sources (BASINS) software], and StreamStats. The USGS, with support by ESRI and partly based on ArcHydro tools, built the StreamStats web application to provide information about the watershed draining to user-selected locations on designated streams. The information includes computed watershed boundaries and drainage areas and many other characteristics including outlet elevation, various precipitation values (e.g., mean annual precipitation), composite soil permeability, 10-85 channel slope, and mean basin slope. In addition, a shape file of the watershed boundaries is generated so that analysts can use it as a boundary file to extract an elevation raster from the NED (embedded within the National Map). The StreamStats and NED are accessible using existing scripting technology and can be embedded into web-services (or even spreadsheet tools). The requisite scripting codes are supplied by the USGS at (https://github.com/USGS-R/nhdplusTools). Figure 4.7 depicts the status of the StreamStats toolkit. Most of the United States is fully implemented, partially implemented, or in the process of implementation; the exceptions are Texas, Florida, Nevada, Nebraska, and Michigan. In these states, the USGS has insufficient funding to implement the tool, however, the underlying data are archived at the USGS. In Alaska, StreamStats is available in very limited parts of the states. Figure 4.7. StreamStats coverage (source: USGS).
33 Other tools, such as EPA WATERS (https://www.epa.gov/waterdata/waters-watershed- assessment-tracking-environmental-results-system) use the NHDPlus catchments and find the containing catchment and all of the catchments upstream for a given location. 4.1.4. Recommendations for Defining a Watershed and APs Selection of APs and the associated watershed depends on the objectives of the State DOT and potential stakeholders, as well as the regulatory framework applicable to project impacts and potential mitigation techniques. At a minimum, the traditional approach to mitigation impacts on- site would define a single AP at the project site and consider the applicable watershed to be the drainage area to the project site. With this definition, the watershed approach to mitigation may be applied by considering the effectiveness of off-site mitigation techniques upstream of the project. When an expanded definition of the watershed is desirable and feasible, another AP can be chosen downstream of the project. A second AP downstream of the project will provide another location to consider the project impacts and mitigation effectiveness. It also allows location of mitigation techniques downstream of the project. The State DOT and potential stakeholders may choose the downstream AP based either on a relative or fixed definition of watersheds as previously discussed. Using relative definition, the AP may be downstream of the project but upstream of the confluence with a stream of equal or higher order. How far downstream depends on what is useful for evaluating project impacts and mitigation effectiveness. Extending so far downstream that project impacts and mitigation effectiveness are minimal is likely not to result in useful information or acceptance from regulatory agencies. Choosing a second AP that has a drainage area no more than twice the drainage area at the project site, that is, at the initial AP, may be a good rule of thumb. Using the fixed definition of a watershed the downstream (second) AP could be the outlet of a predefined watershed delineation such as a HUC12 drainage unit. The contributing watershed to this AP is the HUC12 unit itself plus any upstream HUC12 units draining to the unit containing the project. If desired, the State DOT could select the AP and the next downstream HUC12 further expanding the drainage area to the AP. As with a relative definition, the downstream AP should not be so far downstream that it does not provide useful information to the decision framework for evaluating off-site mitigation techniques. The same rule of thumb, that the drainage area at the downstream AP is no larger than twice the drainage area at the project site AP, is applicable regardless of the approach for defining watersheds. State DOTs and potential stakeholders may choose to define additional APs. Within the watershed defined by the downstream AP, there may be locations where project impacts are a source of concern or where mitigation co-benefits may be of interest. Additional APs may be required by regulatory agencies. While there is no theoretical limit to the number of APs, practical constraints of the assessment and modeling tools will limit the number of locations at which useful evaluation or decision-making information can be developed. For evaluation of multiple projects, larger watersheds than would be needed for a single project may be appropriate to include projects in different locations and to analyze a broader array of mitigation techniques. In this case, the number and location of APs will be determined based on the goals of the watershed analyses and the relevant regulatory constraints.
34 4.2. Hydrologic Impacts, Objectives, and Metrics Identifying the hydrologic objectives, that is, which hydrologic impacts are to be addressed through mitigation is a critical element of the decision framework for mitigating the hydrologic impacts of transportation facilities at a watershed scale. In step 2 and step 3 of the decision framework depicted in Figure 3.1, State DOTs and partners/shareholders determine the hydrologic objectives for the project. These objectives may relate to mitigation of impacts related to environmental/low flows, wetland and/or stream restoration hydrologic regimes, channel- forming flows, or flood flows. Hydrologic metrics to quantify project impacts and to evaluate the value of mitigation alternatives. Weinstein et al. (2017) addressed a single hydrologic metric: average annual runoff. Average annual runoff is useful in contexts including implementation of a TMDL but less so for quantifying stormwater runoff impacts from transportation projects. For hydrologic objectives related to environmental/low flows, possible metrics include a 7Q10 or a water quality storm (e.g., first 0.5 inch). For habitat evaluation, hydrologic metrics might be based on a particularly sensitive species while for wetlands, they might consider frequency of flooding and/or duration of saturation periods. For hydrologic objectives related to channel-forming discharges hydrologic metrics could include bankfull discharge or dominant discharge that represent a single flow value that can be used in design. Other objectives may require consideration of the pattern of flow throughout a specific period in which case a flow-duration curve is useful. Stream restoration objectives will require multiple metrics that include hydrologic measures. For example, the U.S. Army Corps of Engineers (USACE) developed a Stream Quantification Tool (SQT) tool with measures related to hydrology, hydraulics, geomorphology, physicochemical properties, and biology. The SQT hydrologic metrics include runoff, baseflow, and floodplain connectivity. Metrics to evaluate mitigation of flood (high) flow impacts commonly include a peak discharge for a given discharge quantile, such as a 100-year flow, but may also include total runoff volume or runoff duration for a given discharge. Most State DOTs have established design standards for designing highway structures including flows with return periods of 10-years, 25-years, 50-years, 100-years, and occasionally higher/longer periods for extreme events. Less common, but on the increase, are hydrologic standards for assessing water quality, sediment transport, ecosystem impacts, and channel-forming flows. Each state, and likely each project, will select the appropriate metrics because requirements vary from state to state and project to project necessitating that the decision framework be sufficiently flexible to provide for varied state requirements. The following sections describe possible hydrologic methods for a range of project objectives, an overview of State DOT hydraulic design standards, available tools for computing hydrologic metrics, and recommendations for typical hydrologic metrics for use in evaluating project impacts and mitigation effectiveness of stormwater impacts of transportation projects in a watershed context. 4.2.1. Possible Hydrologic Metrics Alteration of the hydrologic regime and the associated impacts on water quality, sediment, and ecosystems can be characterized by five time-variable attributes of streamflow summarized in Table 4.1 (Richter et al. 1996). These attributes directly affect geomorphology of streams as well
35 as sediment erosion, deposition, and transport, water quality, and ecosystems. Richter et al. (1996) further identified five groups of statistics that characterize relevant alterations of hydrologic regimes as they relate to ecosystems: Group 1: Magnitude of monthly water conditions: ⢠Mean value for each calendar month over record of analysis. Group 2: Magnitude and duration of annual extreme events: ⢠1-, 3-, 7-, 30-, and 90-day means for maxima and minima. Group 3: Timing of annual extreme water conditions: ⢠Julian date of each annual 1-day maximum and minimum. Group 4: Frequency and duration of high and low pulses: ⢠Flood frequencies derived from annual flood peaks and statistical analyses. ⢠Number of high and low pulses each year. ⢠Mean duration of high and low pulses within each year. Group 5: Rate and frequency of water condition changes: ⢠Mean of all positive differences between consecutive daily means. ⢠Mean of all negative differences between consecutive daily values. ⢠Number of rises and number of falls. Table 4.1. Attributes of streamflow for ecosystems (Richter et al. 1996). Attribute Description and Influences Magnitude The magnitude of the water condition at any given time is a measure of the availability or suitability of habitat and defines such habitat attributes as wetted area or habitat volume, or the position of a water table relative to wetland or riparian plant rooting zones. Timing The timing of water conditions can determine whether certain life-cycle requirements are met or can influence the degree of stress or mortality associated with extreme water conditions such as floods or droughts. Duration The duration of time over which a specific water condition exists may determine whether a particular life-cycle phase can be completed or the degree to which stressful effects such as inundation or desiccation can accumulate. Frequency The frequency of occurrence of specific water conditions such as droughts or floods may be tied to reproduction or mortality events for various species, thereby influencing population dynamics. Rate of Change The rate of change in water conditions may be tied to the stranding of certain organisms along the waterâs edge or in ponded depressions, or the ability of plant roots to maintain contact with phreatic water supplies.
36 Richter et al. (1996) use these 32 indicators as part of the IHA method to assemble a comprehensive profile for a water body that allows assessment of alteration of hydrologic regime by comparing pre- and post-change values. These indicators, possibly supplemented by additional âtraditionalâ statistics (e.g., variances, coefficients of variance, etc.) are useful within the IHA and as building blocks for developing other indices to measure sediment, water quality and ecosystem impacts/alterations. Changes in each of the streamflow attributes (magnitude, timing, frequency, duration, rate of change) not only produce relevant indicators for ecosystem alteration; they are also relevant factors (and potential indicators) of alterations to sediment and water quality regimes. Comprehensive indices for sediment, water quality, and ecosystem regimes can be created by combining the streamflow indicators with analogous data or model estimations for magnitude, timing, frequency, duration, rate of change of the sediment delivery and chemical loadings (point and nonpoint sources) and by representation of biochemical transformations within the water body. Metrics used in design and regulatory activities are representative of some aspect of magnitude, timing, duration, frequency, and rate of change. Common examples include: ⢠7Q10. The 7Q10 is the lowest 7 consecutive day low flow with an average recurrence frequency of once in every 10 years. It is used by many states and the federal government in setting discharge limits in National Pollutant Discharge Elimination System (NPDES) water quality permits. A permit will only be granted if the proposed pollutant discharge will not significantly impair the designated uses, such as drinking or swimming, when the streamflow falls to the 7Q10 level. In other words, NPDES permit holders are restricted from discharging pollutants that would cause pollutant concentrations in the receiving water to exceed permit limits, even at very low (i.e., 7Q10) streamflow levels. Although such a low streamflow value, roughly equivalent to a ten-year drought, is appropriately used in the context of limiting pollution discharges, the 7Q10 flow statistic is sometimes inappropriately claimed to represent an adequate streamflow for maintaining a healthy aquatic ecosystem, when in fact much higher streamflow levels are required. ⢠3Q2. The 3Q2 is the lowest 3 consecutive day low flow with an average recurrence frequency of once in every 2 years. It has been proposed as an alternative to the 7Q10 statistic and is used in one state; several other states use a 7Q2 statistic. ⢠30Q5. The 30Q5 is the lowest 30 consecutive day low flow with an average recurrence frequency of once in every 5 years. In the state of Idaho, it is proposed for use to establish a human health related regulatory level for non-carcinogens. ⢠nth percentile. Another alternative to the 7Q10 that is based on the flow-duration curve. It has been proposed by some states based on ease of calculation and communication with the public. ⢠0.5-inch runoff. The first 0.5-inch of runoff is often used as a capture amount for mitigating the âfirst flushâ of pollutants. ⢠Harmonic mean flow. The harmonic mean flow is a long-term mean flow value calculated by dividing the number of daily flows analyzed by the sum of the reciprocals of those daily flows. In the state of Idaho, it is proposed for use to establish a human health regulatory level for carcinogens.
37 ⢠4B3. The 4B3 is a biologically-based statistic that indicates an allowable exceedance of a chemical concentration for four consecutive days with an average recurrence frequency of once every 3 years. In the state of Idaho, it is proposed for use to measure chronic toxicity for aquatic life. (7Q10 is proposed as well.) ⢠1B3. The 1B3 is a biologically-based statistic that indicates an allowable exceedance of a chemical concentration for one day with an average recurrence frequency of once every 3 years. In the state of Idaho, it is proposed for use to measure acute toxicity for aquatic life. (1Q10 is proposed as well.) ⢠Index of Biological Integrity (IBI). An IBI (Karr 1991) can be used to evaluate the health of wetlands, streams, lakes, or estuaries. An IBI combines multiple indicators of biological condition (metrics) in a composite index value. The value can be compared to reference values and help managers assess the relative health of individual water bodies. Various IBIs have been developed and used. 4.2.2. State DOT Hydraulic Design Standards In a 2010 survey of highway hydraulic design practices of the 50 states (plus Puerto Rico), the FHWA cataloged design standards used for various types of stream crossing structures (FHWA 2012). Figure 4.8 is a sample of the survey results for design standards for interstate highways and the associated frequency of the design discharge return periods applied. Results are also provided in the survey report for other roadway types, such as major and minor arterials, local roadways, and temporary structures. Design standards varied widely, as might be expected given the wide range of climatic, hydrologic, topographic, and geomorphic conditions covered. However, there were some specific common practices noted: ⢠50-year and 100-year return periods for culvert and bridge hydraulics. ⢠100-year and 500-year return periods for bridge foundations and scour design. ⢠10-year to 50-year return periods for storm drains. 4.2.3. Available Software/Tools for Calculating Hydrologic Metrics There are a variety of computational tools and software for calculating various hydrologic metrics and model performance statistics included in BASINS, HSPEXP+, SARA, SWSTAT, and other tools. The BASINS Climate Assessment Tool (CAT) provides a means for assessing the effects of possible climate change on flow using the watershed models in BASINS (HSPF, SWAT, and SWMM). Appendix B provides additional information on many of these tools. The USGS SWSTAT program (Lumb et al. 1990) provided standardized methods (e.g., Log Pearson Type III) for performing a wide range of statistical analyses, including frequency, N-day generation, percentiles, flow duration, and hydrograph tables/curves. Modern interfaces have been developed for these analyses and the suite of components has been bundled into plugins that are available in EPAâs BASINS (USEPA 2019) and USGSâs Surface Water Toolbox (Kiang 2018). The Surface Water Toolbox also incorporates EPAâs DFLOW program, which focuses on biologically-based design flows. The plugin architecture of the analysis tools means they are readily available for incorporation into other interactive analysis frameworks. In addition, the
38 core computational code of the analysis components could be accessed in a batch mode if preset computations are needed with little to no user interaction. Figure 4.8. Peak discharge return period used for design (FHWA 2012). An additional set of statistics can be produced using HSPEXP+ (Mishra et al. 2017), an expert system program that supports HSPF model applications. These include metrics such as event peak discharge and volume which are not readily available in the tools mentioned above. The tool also computes a wide array of model performance statistics (percent error, NashâSutcliffe efficiency (NSE), root mean square error (RMSE), and RMSE-observations standard deviation ratio (RSR)), which can serve as a resource for assessing model value in simulation of the mitigation techniques being considered. As with the statistical plugins discussed above, HSPEXP+ computational components are readily available for incorporation into other analysis frameworks. It is important to understand that model performance statistics, which are used to compare observed and simulated values in hydrologic modeling, are different than the use of hydrologic metrics in this research which attempt to show the differences between a base (no project) condition and a proposed change on the watershed. Other tools for specialized purposes are also available. For example, NCHRP Report 853 (Bledsoe et al. 2017) provides guidance for addressing the complex issues of channel stability and stream restoration when highway and other transportation projects must address stream crossing designs within a watershed. It summarizes available guidance and provides a set of
39 decision support tools that are scientifically based and practical. Although Bledsoe, et al. (2017) provides a method for determining hydrologic metrics, its intent is for design of channel stability and stream restoration associated with stream crossing designs and not for evaluating stream restoration for hydrologic mitigation benefits. 4.2.4. Recommendations for Hydrologic Metrics Previous sections described metrics, standards, and tools for assessing the hydrologic impacts of transportation projects and evaluating the hydrologic performance of mitigation techniques. State DOTs and their partners/stakeholders may choose or be required to use any of these, or others not mentioned. For this research project, there is a list of hydrologic metrics that may be more commonly needed and will be used in developing the hydrologic screening tools and serve as the basis for detailed analyses. These hydrologic metrics are: 1. Change in event peak discharge (2-, 10-, 25-, 50-, and 100-yr). 2. Change in event volume (2-, 10-, 25-, 50-, and 100-yr). 3. Change in flow-duration curve (based on daily values). a. Change in 90% exceedance flow. b. Change in 10% exceedance flow. c. Change in the percentile corresponding to 0.5-inch runoff. 4. Change in low-flow statistics (e.g., 7Q10, 3Q2, 4Q3). This list is not intended to be exclusive of valid hydrologic measures that a State DOT might find useful for all projects. It merely represents those measures that would be more commonly useful. Each of the hydrologic metrics is obtainable from comprehensive watershed modeling approaches, but only a subset can likely be generated from screening analyses discussed in Section 4.4. Thus, the specific metrics available for evaluation will likely change with the level of analyses being performed for a specific project. 4.3. Simulation of Mitigation Techniques in Hydrologic Modeling The out-of-kind mitigation techniques addressed in this study must be credibly represented within the analytical tool set to estimate their hydrologic effects on the relevant hydrologic metrics within the watershed context. This section describes the general aspects of representing these techniques within watershed models. Each of the mitigation techniques can be represented, sometimes more explicitly than others, in a watershed modeling application. However, some mitigation techniques may not be readily represented in existing tools, especially those tools which were designed for a different purpose. Appendix C provides a summary of the ability of existing tools to effectively represent these techniques. An important consideration for modeling various landscape modification mitigation techniques, such as wetlands, forest, or uplands restoration is how to represent the separation of rainfall into the portion that is retained by the landscape (largely from infiltration), and the portion that runs off. HSPF separately considers pervious and impervious land segments requiring the modeler to
40 specify the pervious land parameters that control infiltration. Other watershed models use the Natural Resources Conservation Service (NRCS) CN approach for estimating runoff and infiltration. Although there is not a direct relationship between various methods, they can be compared by applying multiple models to a watershed and totaling the amount of runoff produced by each model in response to an identical rainstorm. It would also be possible to develop tables of parameters, such as curve numbers, that could be used for various landscape mitigation techniques under a specified range of conditions. The Maryland Department of the Environment, for example, developed a series of recommended effective curve numbers based on the depth of water retained by alternative green roof designs (MDE 2018). Described below are methods for representing each of the out-of-kind mitigation techniques in watershed models such as HSPF, SWMM, and SWAT. In some cases, modelers can represent the mitigation technique by describing the land cover type associated with the technique. In other cases, the modeler creatively adapts the model to represent techniques not explicitly addressed in the model, such as stream restoration. 4.3.1. Wetland Restoration/Creation Generally, this mitigation technique can be simulated as a land cover change in a watershed model. In HSPF, for example, wetlands could be simulated either as a pervious land category (PERLND) or a stream reach or reservoir (RCHRES). Most often wetlands are simulated as one of the PERLND land cover categories. The mitigation using wetlands is represented in the model by redefining some area from one land cover category to wetlands assuming the model already has a wetland category. Because of the way HSPF represents land covers lumped by subcatchment, the smaller the subcatchment the more meaningful the simulation results will be, since the spatial orientation of the wetland area within the subcatchment is not considered. For a more detailed modeling effort one might simulate wetlands as a PERLND with other PERLNDs draining to it, or as a RCHRES with its own hydraulic functions table (FTABLE). Wetland restoration and creation would be represented as a change in land cover. 4.3.2. Forest Restoration/Creation This mitigation technique is also represented as a land cover change, assuming the model already has an appropriate forest category. Further GIS processing might be needed if there are multiple forest categories for different forest types, soils, slopes, etc. Forest restoration and creation would be represented as a change in land cover. 4.3.3. Stream Stabilization/Restoration/Improvement In HSPF, hydraulic capacity changes to the stream channel are represented through changes to the FTABLEs, which specify the depth-surface area-volume-discharge relationship for a stream. In SWMM these changes can be represented more directly as changes to the channel parameterization. Bed and bank stabilization and riparian buffer restoration might affect water quality more than hydrology/hydraulics. However, these techniques might influence channel roughness and could slow the flow through the stream segment.
41 In-stream enhancement might include a return to a more natural channel, restoring stream sinuosity for example. This would be represented as a change in channel geometry. Floodplain reconnection could mean removal of levees allowing water to return to the natural floodplain or the addition of floodplain culverts outside the main channel. Other means of floodplain reconnection include reconfiguring channel cross-sections where entrenchment or encroachment has occurred in the channel so that the entrenchment is reduced and streamflow spreads to the floodplain more frequently. Floodplain reconnection in HSPF might mean a simple FTABLE change, in reaches with existing disconnected floodplains. 4.3.4. Uplands Restoration Uplands restoration includes techniques such as tree/grass planting, increased vegetation, and removal of impervious cover anywhere in the watershed with the goal of recreating more natural conditions and reducing the rate and volume of runoff coming off the restored area. All these techniques can be implemented through land cover changes in a watershed model. 4.3.5. Agricultural Practices Modification or Land Conversion Agricultural practice modification includes moving from conventional to low till plowing, land left fallow, removal of drain tiles, etc., which could have some hydrologic impact in retaining water. For example, practices that increase organic matter in agricultural soils, could help retain water and reduce runoff, peak flows, as well as mitigate the effects of drought. In HSPF such a technique would be represented by a land cover change including parameterization of the encouraged agricultural practice. Conversion of agricultural land to another land cover, e.g., uplands restoration as described in Section 4.3.4, is represented as a land cover change. 4.4. Screening Evaluation of the Hydrologic Impacts and Effectiveness of Mitigation Techniques In step 8 of the decision framework described in Section 3.1, the State DOT and external stakeholders/partners assess the portfolio of mitigation techniques using screening tools. This section describes the development and use of screening tools to determine potential hydrologic performance of various mitigation techniques (step 8A). Because screening tools, by definition, are more generalized to facilitate relatively easy application, they may not provide definitive information on which to base decisions. Each project and planning effort will exhibit unique characteristics. In some cases, actionable information may not be attained through screening alone and more detailed analysis as anticipated in step 11 of the decision framework may be required. Initial efforts to identify appropriate screening tools focused on adaptation and use of existing assessment tools that do not require watershed modeling expertise. Examples include the National Stormwater Calculator or regional tools such as the WWHM. Unfortunately, this effort did not result in the identification of existing tools appropriate for the goals of this project. Appendix C summarizes the tools evaluated for this purpose and, where appropriate, identifies modifications to the tools that could increase the utility of each tool. After concluding that existing screening tools would not effectively serve this study, the research team developed a new screening tool based on site-specific watershed model runs adapted for nationwide application. This section describes a matrix of detailed model watershed runs
42 performed to develop the screening tool followed by a description of the development of the screening tool and an explanation of its application. 4.4.1. Watershed Model Run Matrix To develop a hydrologic screening tool for evaluating mitigation techniques in a watershed context, the research team created a simulation matrix of watershed model runs incorporating a range of watershed conditions, transportation infrastructure locations, and mitigation techniques. This matrix guided implementation of detailed watershed modeling runs at multiple sites. The research team synthesized the watershed modeling results into a nationwide screening tool that is described later in this section. In this section, the simulation matrix parameters â location, mitigation techniques, mitigation technique size, and watershed context â are described. Locations (L): Several locations within the conterminous U.S. representative of a range of climate conditions (e.g., rainfall and temperatures) were evaluated. The work focused on locations with existing watershed models in Colorado, Minnesota, and Georgia. Mitigation techniques (T): Several mitigation techniques were investigated for representation in the selected watershed models. Section 4.3 summarizes how these techniques can generally be represented in watershed models. This mitigation techniques evaluated were: ⢠Wetland restoration/creation (represented as a change in land cover). ⢠Forest restoration/creation (represented as a change in land cover). ⢠Stream stabilization/restoration/improvement (represented by a single change in channel geometry). As described previously, this category of mitigation may include distinct subcategories such as bed and bank stabilization, riparian buffer restoration, in-stream enhancement, and floodplain reconnection. These subcategories are not individually evaluated as part of the screening tool development. ⢠Uplands restoration (represented as a change in land cover). Mitigation technique size (S): Each of the mitigation techniques may vary in size and may perform differently as the technique is scaled up. A range of possible mitigation technique sizes were investigated to define a range of possible results and to determine if results could be reliably extrapolated beyond modeled conditions. Watershed context (W): The position of the mitigation technique within the watershed (e.g., near an AP or in the upper or lower parts of the watershed), watershed slope, and watershed infiltration characteristics influence the effectiveness of the techniques in mitigating hydrologic effects of transportation projects. The research team performed model runs varying watershed position and infiltration conditions. The number of continuous simulations required to represent possible combinations of the four factors is the product of L x T x S x W where L represents the number of locations, T represents the number of mitigation techniques, S represents the number of sizes, and W represents the number of watershed location/infiltration combinations. The research team analyzed the simulation results for trends and patterns that could be used as part of a national screening tool.
43 4.4.2. Detailed Modeling Results for Screening Tool Development Development of the screening tool involved analysis of the results of approximately 1200 HSPF model runs at three locations (Colorado, Georgia, and Minnesota), each with two subwatersheds, and four mitigation options. Results were evaluated applying 10 hydrologic metrics. The work explored several possible tool types and formats resulting in a nationwide screening tool available in both a tabular form and as a spreadsheet. 4.4.2.1. General Methodology For each of three locations, the research team selected a subset of two to six subbasins from the larger existing watershed model as the focus of the modeling. These subsets were selected to: 1) avoid influences from large lakes or impoundments, 2) include a mix of simulated land cover types, and 3) include both pervious and impervious areas. To limit the overall watershed size at the most downstream AP, the research team selected headwater subbasins so that the effects of the transportation projects and mitigation techniques would be measurable in the hydrologic metrics. In most cases additional impervious area was added to the land cover distribution within each model so that each subbasin would contain potential impervious area that could be converted to a different land cover to represent application of a mitigation technique. Because mitigation scenarios in the modeling matrix involved conversion of up to 20 percent of a subbasin from impervious to another land cover category, it was necessary to start with subbasins with at least 20 percent imperviousness. The research team refers to these modified models as the baseline hypothetical scenario. All other model scenarios were developed from, and compared with, these baseline scenarios. 4.4.2.2. Watershed Model Locations Figure 4.9 displays the locations in Georgia, Colorado, and Minnesota of the established, calibrated HSPF watershed model applied for screening tool development. With additional time and resources, other locations could be added, but this distribution of project locations provides a good range of possible precipitation conditions and watershed responses. A summary of each watershed model and a short description of its geography follows. Cherry Creek, CO: Cherry Creek is a watershed in central Colorado, southeast of Denver. Watershed land cover consists mainly of grassland and scrubland, except for suburban areas near Denver. The area is characterized as the Colorado Piedmont subregion of the Great Plains, with a semiarid climate and average annual precipitation of 19.2 inches. Mean annual runoff for pervious and impervious surfaces in the basin is 0.71 inches and 12.5 inches, respectively. Figure 4.10 shows Piney Creek, the small tributary to Cherry Creek that is the focus of this study. The transportation project impact is indicated in subbasin 276 and APs are located at the outlets of subbasins 278 and 276. Table 4.2 summarizes the subbasin size and percent imperviousness for the hypothetical baseline conditions.
44 Figure 4.9. Geographic location of watershed models. Figure 4.10. Piney Creek (Cherry Creek), Colorado.
45 Table 4.2. Piney Creek, Colorado, hypothetical baseline watershed summary. Subbasin Area (acres) Area (square miles) Percent Impervious 272 4,451.7 7.0 7.8 274 1,722.7 2.7 11.1 276 3,458.7 5.4 19.2 277 1,692.4 2.6 16.5 278 2,711.7 4.2 27.4 Total 14,037.2 21.9 15.9 Upatoi Creek, GA: The Upatoi watershed is in western Georgia, near the City of Columbus and contains the Fort Benning army base. Except for the areas developed for military purposes, land cover within the watershed is mostly forested. The Upatoi Creek watershed is located just below the fall line where the rolling hills of the Piedmont meet the Coastal Plain. The climate is generally mild in winter and hot and humid in summer, with an average annual rainfall of 48.8 inches. Mean annual runoff for pervious and impervious surfaces in the basin is 17.9 inches and 40.5 inches, respectively. Figure 4.11 shows the Upper Pine Knot Creek, a small tributary to Upatoi Creek that is the focus of this study. The transportation project impact is indicated in subbasin 23 and APs are located at the outlets of subbasins 24 and 23. Table 4.3 summarizes the subbasin size and percent imperviousness for the hypothetical baseline conditions. Figure 4.11. Upper Pine Knot Creek (Upatoi Creek), GA.
46 Table 4.3. Upper Pine Knot Creek, Georgia, hypothetical baseline watershed summary. Subbasin Area (acres) Area (square miles) Percent Impervious 23 10,633.9 16.6 20.7 24 10,139.3 15.8 22.9 Total 20,773.2 32.4 21.8 Lower Minnesota, MN: The Lower Minnesota River is a watershed in southeastern Minnesota, near Minneapolis. The watershed is primarily grassland and forested, with significant urbanization near Minneapolis. The area is characterized by gently rolling plains, with many lakes formed from glaciation. The climate is characterized by temperature extremes and moderate precipitation of 30.8 inches per year. Mean annual runoff for pervious and impervious surfaces in the basin is 6.8 inches and 25.7 inches, respectively. Figure 4.12 shows the Bluff Creek watershed that is tributary to the Lower Minnesota River used in the simulations. The transportation project impact is indicated in subbasin 501 and APs are located at the outlets of subbasins 501 and 502. Table 4.4 summarizes the subbasin size and percent imperviousness for the hypothetical baseline conditions. Figure 4.12. Bluff Creek (Lower Minnesota River), MN.
47 Table 4.4. Bluff Creek, Minnesota, hypothetical baseline watershed summary. Subbasin Area (acres) Area (square miles) Percent Impervious 501 2,220.4 3.5 21.3 502 2,220.4 3.5 21.3 Total 4,440.8 7.0 21.3 4.4.2.3. Transportation Project Impacts and Mitigation Technique Effectiveness The research team assessed the hydrologic impacts of transportation projects by modeling hypothetical projects as a conversion of the predominant natural condition (forest or grassland) to an impervious (paved) land cover. To assess possible differences in effects on a range of hydrologic metrics, the research team used four scenarios representing impact areas representing 2, 5, 10, and 20 percent of the subbasin area. The research team also assessed the effectiveness of four mitigation techniques in compensating for the hypothetical transportation project impacts. Three of these â wetland restoration/creation, forest restoration/creation, and uplands restoration â were each represented by converting a portion of the baseline watershed to wetlands, forest, or grassland, respectively. As with the transportation impacts, four mitigation sizes were modeled representing 2, 5, 10, and 20 percent of the subbasin. For each mitigation technique and mitigation extent, two conversions were assessed. One assumed the baseline land cover represented an impervious land cover category. The second assumed the baseline land cover represented a medium runoff land cover category (grassland). In both cases, the land cover was converted to wetland, forest, or grassland depending on the mitigation technique being evaluated. In some cases, the converted land covers were an aggregate of non-contiguous areas of that land cover type. A further variation considered the location of the mitigation site relative to the project impact site. Two alternatives were considered: 1) upstream of the transportation project in the subbasin containing the transportation project and 2) downstream of the transportation project in a downstream subbasin. The fourth mitigation technique â stream restoration â could not be represented in the watershed models as a conversion of land cover. Stream restoration was represented by altering the representation of the stream channel to simulate a more winding (sinuous) and consequently longer stream. (Implementation of added sinuosity would require additional analysis of sediment transport effects of changing the channel grade.) Other modifications such as changing channel roughness or channel cross-section geometry are also valid approaches to representing stream restoration, but were evaluated in this research project. Four alternative channel lengths were modeled representing increased stream lengths of 2, 5, 10, and 20 percent. These mitigations were also placed in two different locations, one upstream of the transportation project in the subbasin containing the transportation project, and one in the next subbasin downstream of the transportation project.
48 4.4.2.4. Hydrologic Metrics For each model simulation, model output was assessed just below the assumed transportation project location and at the next HUC12 boundary downstream. At each location, statistics were computed for the following statistical terms using the hourly continuous simulation output: Based on hourly discharge: ⢠100-year peak discharge (cfs). ⢠50-year peak discharge (cfs). ⢠25-year peak discharge (cfs). ⢠10-year peak discharge (cfs). ⢠2-year peak discharge (cfs). Based on runoff event volumes: ⢠100-year event volume (ac-ft). ⢠50-year event volume (ac-ft). ⢠25-year event volume (ac-ft). ⢠10-year event volume (ac-ft). ⢠2-year event volume (ac-ft). The frequency statistics (peak discharges and event volumes) are based on the Log Pearson Type III distribution, as computed by USGS Surface Water Statistics software tools. For consistency across all computations and because each of the three modeling locations had flow time series of approximately 20 years, the baseline site-specific skew was employed for the baseline and transportation project and mitigation scenarios in the Log Pearson Type III distribution fitting. To compute event volumes from a time series of hourly discharge values, definition of an event was established. For this study, events are defined by looking for flows exceeding a certain flow rate threshold. This threshold is calculated by looking through the time series of baseline simulation discharges and finding the highest flow rate in each year; the smallest of these annual flow values multiplied by 0.1 is considered the minimum event and is considered the threshold. Then a list of events is computed using this threshold, where a unique event is defined as starting where the flow exceeds this threshold and ending when the flow returns below this threshold for at least a half-day. By inspection of the list of event volumes, the maximum event of each year of the simulation period is identified. The resulting series of annual maximum events is used (along with unit conversion from cfs-hr to ac-ft) in computing the volume statistics using the Log Pearson Type III as discussed. Multiple Simulations Results of 1200 model runs were compiled from three locations (Colorado, Georgia, and Minnesota), each consisting of two watersheds models, and four mitigation options for 10 hydrologic metrics.
49 In addition to peak and volume statistics, other hydrologic metrics were computed from the simulation model output. These metrics included the 10- and 90-percent exceedance flows from the flow-duration curve computed based on a simple ranking of the daily simulated streamflow values aggregated from the hourly model output. The research team also computed the X-percent exceedance flow that corresponded to 0.5 inches of runoff. However, none of these three lower flow metrics were able to consistently detect changes from the baseline to project impact and then from project impact to impact mitigation. With respect to determining the X-percent exceedance flow corresponding to 0.5 inches of runoff, this amount of runoff rarely occurred during the model simulation because very few rainfall events were sufficiently large to produce 0.5 inches of runoff over a day at the subbasin scale. 4.4.2.5. Modeling Limitations Evaluating the effectiveness of each mitigation technique depends on having a well-calibrated hydrologic model that produces a time series of flow from which peak flow quantiles can be reasonably estimated. For this effort, the research team used existing previously calibrated watershed models. Proper calibration and validation should be conducted when extending this methodology to other sites. In a few of the initial simulations, the research team found that some of the hydrologic metrics were not behaving in the direction and/or magnitude expected. Further investigation revealed that large variations in the sizes of adjacent subbasins may affect the results. To mitigate this, subbasin outlet locations were adjusted or new subbasins were selected so that subbasins upstream and downstream of the transportation project were comparably sized and that stream reaches upstream and downstream are similar in length and slope. The subbasin descriptions provided in Section 4.4.2.2 reflect adjusts made by the research team. Each model produced approximately 20 years of simulated hourly flows for each model simulation. Although this is sufficient for developing an annual maximum series and computing peak flow quantiles using the Log Pearson Type III distribution with regional skew values, statistics for larger events, such as the 100-year event, are estimated with less confidence than if a longer simulation period had been available. The same is true when estimating larger events from gauge records with 20 years of record. The modeling performed to develop the screening tool also revealed insights about potential limitations of watershed modeling for facilitating a watershed approach for out-of-kind landscape mitigation. Watersheds that are substantially larger than those modeled in this project may be so large that neither the hydrologic effects of the transportation project nor the effects of the mitigation will be measurable in the hydrologic metrics. That is, the hydrologic changes are too small compared to the overall hydrology within the watershed. Conversely, watershed sizes may be too small if they provide an insufficient area for the placement of out-of-kind landscape mitigation. In addition, while it is not required that the mitigation areas be contiguous or distributed evenly throughout the watershed, the effects of the mitigation have been evaluated using a lumped parameter approach. Therefore, the specific locations of the mitigation techniques within a subbasin are not represented in detail effectively meaning that the composite mitigation area is essentially distributed throughout the subbasin.
50 4.4.2.5.1. Mitigation from Pervious Land Covers The watershed modeling for this research revealed that under some circumstances, mitigation of the impacts of a transportation project could potentially be accomplished through conversion of certain pervious land covers (grasslands, scrub/shrub, urban parks and lawns, and agricultural lands/pasture) to forests or wetlands. The effectiveness of this landscape conversion is less effective than mitigation from conversion of impervious land covers (locations where the land to be converted is currently developed with an imperviousness greater than 95 percent). The modeling also found that conversions from pervious land cover to forests or wetlands only provided mitigation benefits where those lands are in areas with particular climate conditions. Further investigation determined that these climate conditions can be tied to locations where there is abundant precipitation and humidity (where yearly precipitation typically exceeds evaporation and evapotranspiration) which results in more prolific vegetation and vegetative decomposition creating organic soil conditions. The location of these climate and soils conditions can be identified by using average annual runoff as a representative surrogate for the climate and soil conditions. Comparing climate, precipitation, and runoff across the U.S. with the watershed modeling, the researchers concluded that mitigation conversion of existing pervious land cover will likely only produce benefits where the average annual runoff exceeds approximately 5 inches. Figure 4.13 illustrates average annual runoff for the conterminous U.S. (USFS 1981). Although the data supporting the figure are pre-1981, the general areas are representative of where 5 inches of runoff annually or less are likely to occur. This map does not cover Alaska or Hawaii and data for average annual runoff at specific transportation project sites is not readily available. However, average annual precipitation for most of the United States can be readily found from online sources. Figure 4.14 presents average annual precipitation for the conterminous U.S. Comparing this map to the average annual runoff map reveals that a five-inch average annual runoff contour in Figure 4.13 approximates a 24-inch average annual precipitation contour in Figure 4.14. Therefore, researchers concluded that for screening tool application, areas where the average annual precipitation is 24 inches or less are unlikely to provide significant hydrologic mitigation benefits from conversion of pervious land covers. Therefore, the screening tool should not be used for mitigation evaluation from pervious land covers in these areas. Detailed analysis is appropriate. It also follows, that conversion of a pervious land cover to a similar pervious land cover (something other than forest / wetlands) is not going to have significant hydrologic mitigation benefit, no matter the climate and soils conditions, so this information is also reflected in the screening tool. 4.4.2.5.2. Watershed Size and Project Impact Area Runoff conditions in a specific watershed vary greatly based upon percent imperviousness, stream and watershed slope, soils, vegetation, aspect, and other hydrologic and hydraulic factors. The intent of the screening tool is to provide a method to determine the potential for mitigation of the hydrologic impacts of a transportation project without extensive watershed evaluation and modeling. Screening tool development employed a range of conditions previously documented but do not include every possible watershed condition.
51 Figure 4.13. Average annual runoff. Figure 4.14. Average annual precipitation.
52 The screening tool was developed by testing a range of transportation project sizes ranging from 2 to 20 percent of the tributary drainage area at the AP and on watersheds smaller than 40 square miles. The ratios in this range of conditions were determined by interpolation. Applying this tool for cumulative project impacts less than 2 percent or greater than 20 percent (relative to the area at the AP) or for watersheds larger than 40 square miles assumes that the interpolated ratios can be extended outside of the range of computed values by extrapolation. This increases the risk that using the mitigation ratios developed for the tool outside these limits may be imprecise. In addition, when the AP is located well downstream of a transportation project, the adequacy of the mitigation effort (whether above or below the transportation project) will be affected by inflows from intervening streams and watersheds. Such conditions could limit the ability of the screening tool to provide a reliable mitigation ratio. Therefore, the researchers recommend the following screening tool application limits: ⢠Downstream APs should be no further downstream from the transportation project location such that the watershed area to the AP is no more than twice the watershed area just downstream of the transportation project. ⢠New imperviousness associated with the transportation project should preferably be between 2 and 20 percent (inclusive) of the watershed area to the AP. Application of the screening tool to highway impacts greater than 20 percent of the watershed area to the AP is more uncertain than application to highway impacts less than 2 percent of the watershed area. However, it is recommended that applications less than 2 or greater than 20 percent be assessed using detailed modeling analysis. ⢠Total watershed area to the AP should be less than or equal to 40 square miles. The screening tool may be used to estimate mitigation requirements outside of these suggested limits with the caveats described recognizing that the actual site conditions may require a larger (or smaller) mitigation ratio than what the screening tool indicates. In these instances, a detailed study may be warranted to determine site-specific mitigation ratios. 4.4.3. Hydrologic Screening Tool The hydrologic screening tool consists of a series of tables representing different watershed conditions, mitigation techniques, and hydrologic metrics. The tables contain mitigation ratios based on acreage for land cover conversion techniques (forest restoration, wetland restoration, and uplands restoration) and based on channel length for stream restoration techniques. Both are computed from the site-specific watershed model simulations described in the previous sections and provide the extent of mitigation needed per unit of transportation project impact. The tables provide estimates of the number of acres of mitigation (for land cover conversion Mitigation Ratios Mitigation ratios provide a consistent metric for comparing varying project and mitigation options. For land use conversion, the ratio is the number of acres of mitigation needed to compensate for one acre of highway impact. For stream restoration, the ratio is the increase of stream length (sinuosity) in feet to compensate for one acre of highway impact.
53 mitigation) or increased stream length (for stream restoration) required to compensate for one acre of highway development impact. Based on modeling diverse watersheds, the tool is intended to be applicable nationwide. The screening tool is most useful for planning and evaluation of the potential to use a watershed approach to mitigating the hydrologic impacts of transportation projects. However, more detailed modeling may be needed on a case-by-case basis to satisfy the State DOT, stakeholders, and collaborators that any proposed mitigation is successful in achieving the desired outcomes. 4.4.3.1. Determination of Mitigation Ratios The mitigation ratios were determined through a synthesis of the modeling results for each location and highway impact size interpolating between the 2, 5, 10, and 20 percent mitigation areas (for land cover mitigation) or 1.02, 1.05, 1.1, and 1.2 times the stream length (for stream restoration) to estimate the amount of mitigation required to return each output metric to its baseline level. These numbers were then converted to mitigation ratios for land cover mitigation (area of restoration per area of highway impact), and to units of feet per acre of new impervious highway impact for stream restoration. As shown in the example in Table 4.5, combined with a 2 percent transportation project impact, a 2 percent forest restoration mitigation, increases the 100-yr flow by 1.8 percent compared to the baseline. Similarly, a 5 percent forest restoration mitigation combined with a 2 percent transportation project impact, decreases the 100-year flow by 1.0 percent compared with the baseline of no project and no mitigation. Interpolating between these points (2,1.8) and (5,-1.0) the point at which the peak flow returns to the baseline is identified (3.9,0.0). In this case, a 3.9 percent mitigation is required to have zero change in the 100-yr hydrologic metric, thus mitigating the highway impact. In this example the highway impact size is 2 percent, therefore, the mitigation ratio for this example is 3.9/2 = 1.95, meaning 1.95 acres of forest restoration are required per acre of transportation project impact. Table 4.5. Example model results for percent change in the 100-year peak flow using forest mitigation of impervious areas downstream of a 2 percent transportation project site. Percent Forest Mitigation Percent Change in 100-yr Peak Flow 0 3.7 2 1.8 5 -1.0 10 -5.7 Table 4.6 summarizes a range of mitigation ratios computed for differing conditions representing the site-specific variability in watershed conditions, transportation project sizes, and geographic locations. The similarity in ratios across mitigation measures is the result, in part, of how these measures are represented in the watershed models as changes to land covers and streams. However, given all these variabilities, the high end of the range encompasses some of these uncertainties and is considered most useful for planning purposes. To provide the most
54 conservative estimate (and thus the highest confidence level), the highest mitigation ratio from among the geographic locations and highway impact sizes was identified and compiled into the final tables. The final tables reflect a balancing between a straightforward screening tool and defensible groupings of mitigation ratios. Table 4.6. Example mitigation ratios for the 100-year peak flow for forest restoration of impervious areas located downstream of the project site. Location % Highway Mitigation Ratio for 100yr Peak Flow CO 2 1.95 GA 2 1.00 MN 2 1.00 CO 5 1.92 GA 5 1.00 MN 5 1.02 CO 10 1.74 GA 10 1.00 In some cases, the mitigation ratios were highly variable, showing inconsistent relationships between highway impact size and required mitigation amount. In other cases, the mitigation ratio was found to be less than 1.0, which generally results from hydrograph attenuation during routing. This result was more common in the cases where the mitigation and APs are downstream of the transportation project. Mitigation ratios of less than one were not considered in the screening tool. In addition, for some transportation project sizes and mitigation techniques, the largest amount of mitigation simulated did not return the output metric to the baseline level. Rather than attempt extrapolation, these cases were also discarded from the analysis. Cases such as these have been noted in the final tables as not available (N/A) as no consistent pattern was detected. For these cases, a detailed analysis is recommended to better understand the effectiveness of mitigation techniques for the specific situation. The full set of tabular mitigation ratios from the analysis of modeling results is presented in Appendix D. All of the mitigation ratios for hydrologic screening are applicable for watersheds 40 square miles or less and where cumulative transportation project impacts are between 2 and 20 percent of the watershed area to the AP. Section 4.4.2.5 discusses these limits and use of the hydrologic screening tool outside of these limits.
55 4.4.3.2. Land Cover Conversion Mitigation Techniques For forest restoration/creation, wetland restoration/creation, and uplands restoration, the screening tool identifies the required mitigation ratio that specifies the number of acres of mitigation to compensate for one acre of highway (development) impact. This approach is similar in concept to wetland mitigation requirements. For example, a forest creation mitigation ratio of 1.5 under a certain set of conditions means that for every acre of highway impact, 1.5 acres of forest creation is needed for the specific hydrologic metric. Because the screening tool was developed based on modeling from diverse watersheds, the tool is applicable nationwide. The screening tool is most useful for planning and evaluation purposes and may be sufficient to meet the needs for implementing a watershed approach for mitigating hydrologic impacts of transportation projects. User input to the screening tool requires the following: ⢠Area of the watershed at the APs (AAP). The APs is the location where the hydrologic metric is evaluated. ⢠Area of the transportation project impact (AH). The highway impact is defined as the new impervious area compared with existing conditions. If the project is 100 percent impervious and the existing conditions are 100 percent pervious, the transportation project impact is the total project footprint. ⢠Area of the watershed between the transportation project and the APs (ADS). ⢠Proposed mitigation location (upstream or downstream of the transportation project). ⢠Existing land cover type of the proposed mitigation area. Table D.1 through Table D.6 (see Appendix D) summarize mitigation ratios for three landscape mitigation techniques: forest restoration, uplands (grassland) restoration, and wetland restoration. The mitigation ratio represents the area of mitigation required per area of transportation project impact. Table D.1 through Table D.3 apply to mitigation sites that are predominantly (at least 95 percent) impervious. Restoration of these types of land covers to forest, uplands, or wetlands are more effective than conversion of other types of land covers. Table D.4 through Table D.6 apply to mitigation sites that are predominantly pervious (no more than 5 percent impervious). These land covers may include grassland, pasture, scrub/shrub, and urban pervious lands. In addition, the tables for pervious land cover mitigation sites are only applicable for locations with at least 24 inches average annual precipitation. Section 4.4.2.5.1 provides the background for the use limitations for pervious land cover sites. Each table provides mitigation ratios for combinations of the following: ⢠AP location: outlet of the downstream subbasin or at the transportation project location. ⢠Mitigation location: downstream or upstream of the transportation project location. Mitigation Options Mitigation can be in the form of land use conversion (impervious or pervious) or a stream restoration/length increase. Mitigation can be upstream or downstream of the highway project as long as there is an Assessment Point (AP) further downstream.
56 ⢠Mitigation site: baseline land cover before mitigation represents an impervious or pervious land cover. For the screening tool: o Impervious land cover is defined as mitigation sites where the existing land is at least 95 percent impervious. o Pervious land cover is defined as mitigation sites where the existing land cover is no more than 5 percent impervious. For existing land cover between 5 and 95 percent at potential mitigation sites, use of the pervious designation will result in conservative screening results underestimating the potential mitigation benefits. Detailed modeling is recommended in these cases. 4.4.3.3. Stream Restoration Mitigation Techniques In addition to land cover conversion mitigation techniques, the research team simulated stream restoration as a mitigation technique. Stream restoration can take many forms including modifying channel cross-section geometry and roughness, reconnecting floodplains, and adding sinuosity. For development of the screening tool, the models were modified to simulate more winding (sinuous), longer streams than exist in the baseline. Figure 4.15 depicts the role of sinuosity in classifying stream types. The modifications addressed here are not intended to change stream classification but increase the sinuosity within a classification. Understanding the geomorphology of the stream is critical for successful mitigation. A stream may be naturally straight because of sediment load and/or gradient, for example. Just as important is knowing the trajectory of the watershed. Increasing sinuosity may not be appropriate if the mitigation reach will receive high sediment loads. Stream restoration is complicated by the need to fit in with sediment and flow transport compatible with upstream and downstream reaches. Figure 4.15. Stream classification by sinuosity (source: TxDOT 2019). Four mitigation scenarios were modeled to reflect a range of mitigation options. The options were to increase the baseline stream length by 2, 5, 10, and 20 percent. Two mitigation locations were also evaluated: one upstream of the transportation project in the same subbasin as the transportation project, and one in the next subbasin downstream of the transportation project. For the stream mitigation technique, user information to apply the screening tool will require the following: ⢠Area of the watershed AAP. The AP is the location where the hydrologic metric is evaluated. ⢠Area of the transportation project impact (AH).
57 ⢠Area of the watershed between the transportation project and the APs (ADS). This could be zero if they are at the same location up to some reasonable value. ⢠Proposed mitigation location (upstream or downstream of the transportation project). Table D.7 (see Appendix D) summarizes the estimated stream restoration length (feet) needed to offset every acre of new imperviousness from a transportation project. For each of the cases, a single mitigation amount is applicable to the 2-year to 100-year peak flows, but the amount varies for event volumes. As with the landscape mitigation techniques, both the peak flow and volume metrics are available. The State DOT uses the stream restoration mitigation table by considering the appropriate metrics, mitigation location, and APs. The appropriate table row is then identified from which to collect mitigation values for the metrics of interest. 4.5. Detailed Evaluation of the Hydrologic Impacts and Mitigation The decision framework described in Section 3.1 anticipates that evaluating transportation impacts and mitigation technique effectiveness with screening tools may not result in actionable results. In these cases, the State DOT may decide to move forward with more detailed site-specific evaluation as described in step 11 of the decision framework. Detailed tools to support step 11 will require watershed modeling expertise. In general, detailed evaluation will involve adapting existing watershed models, where available, or developing new watershed models. Adaptation of existing watershed models requires a more modest level of watershed modeling expertise. However, certain standards are needed to ensure the model is setup to gain the required insights. Developing a new watershed model for a particular transportation project or region for use in multiple projects is the most time and cost-intensive option but has the potential to provide the most useful site-specific information. High levels of watershed modeling experience, as well as extensive land cover, soils, weather, and other data are required. The following sections provide examples of watershed model adaptation and the process of developing new watershed models. 4.5.1. Adaptation of Existing Watershed Models Adaptation of existing watershed models developed by the State DOT or other agencies or organizations for their own purposes is a viable option. Two examples of model adaptation - Upatoi Creek in Georgia and Cherry Creek in Colorado - are discussed in the following sections. 4.5.1.1. Upatoi Creek Example As an example of a detailed analysis, the research team developed a hypothetical transportation project impact and evaluated mitigation techniques using an existing HSPF model of the Upatoi Creek in Fort Benning, GA. Figure 4.16 shows the full watershed model within the EPA BASINS system. Figure 4.17 focuses on a subset of the subbasins in the larger model with special attention to subbasin 31 where a hypothetical planned transportation project is shown traversing the subbasin. Using the detailed watershed model, the research team demonstrates assessment of the project impacts downstream at the AP where this tributary meets the main stem of Upatoi Creek at the outlet of subbasin 33.
58 Figure 4.16. HSPF integration with EPA BASINS. Figure 4.17. Hypothetical Georgia transportation project example using HSPF.
59 Starting with a baseline model, the modeler modifies the baseline model to represent the altered land cover with the increased impervious area from the new transportation project. Figure 4.18 illustrates the Windows interface to HSPF in BASINS (WinHSPF) that the modeler may use to make the needed land cover changes. Through this interface it is apparent that the proposed highway will be traversing land that is predominately forested within subbasin (RCHRES) 31. Figure 4.18. WinHSPF user interface. The modeler will evaluate the mitigation of the transportation project by a combination of reforestation and stream restoration. Reforestation can be modeled in HSPF by shifting land area from a developed category back into the forest category. For stream restoration, the input stream
60 length is extended and the HSPF âFTABLEâ associated with the stream reach is modified accordingly with a larger surface area and volume. The modeler then runs this new HSPF scenario with the project impact and the mitigation techniques in place. Since HSPF is a continuous simulation model, the modeler assesses the time series flow output at the outlet for three scenarios: 1) the baseline condition, 2) the proposed highway condition with no mitigation, and 3) the proposed highway with mitigation. Figure 4.19 displays the flow-duration plot from BASINS for the three scenarios. At the scale displayed, the three scenarios appear to have similar flow-duration curves. Figure 4.19. BASINS flow duration plot. Zooming in on the higher flows, as shown in Figure 4.20, it is apparent that the proposed highway increases the higher stream flows (red line) compared with the baseline. The figure also shows that, with the mitigation techniques in place, the higher stream flows (green) are reduced compared to the baseline (blue). The effects are also apparent in the short-term hydrograph shown in Figure 4.21. Many more hydrologic metrics are also available. For example, the 1-day flow with a 100-year recurrence interval is 2822 cfs in the baseline, 2939 cfs in the highway scenario, and 1633 cfs in the mitigated scenario. The previous section discusses at length the possible metrics that might be examined to determine if the mitigation objectives have been met, and any of the metrics can be calculated using the simulated streamflow time series. However, the proposed mitigation in this case successfully compensates for the highway hydrologic impact.
61 Figure 4.20. HSPF streamflow mitigation results for the higher flow rates. Figure 4.21. HSPF short-term hydrograph results.
62 4.5.1.2. Cherry Creek Example A second example of adapting an existing HSPF application is based on a model of the Cherry Creek watershed in Colorado as shown in Figure 4.22 within the EPA BASINS system. In this example, the modeler considers a hypothetical transportation project cutting across two of the model subbasins (206 and 202) as shown in Figure 4.23. These subbasins include two tributaries to Cherry Creek, known as Sulphur Gulch and Tallman Gulch. Figure 4.22. HSPF-BASINS Cherry Creek Colorado model interface. The modeler adjusts the baseline model to represent the altered land cover with the increased impervious area from the new transportation project. Using the data from the HSPF model the proposed highway will traverse land that is predominately in the âGrass/Shrub/Barren/Pastureâ land cover category. In this case, the modeler will assess the mitigation effectiveness related to the hydrologic effects of this transportation project by using a water reuse strategy. In this application, the water reuse consists of reducing the point source discharge directly into the stream and capturing peak
63 precipitation events through rain barrels and similar storage mechanisms for reuse. Water reuse can be modeled in HSPF in a simple hypothetical way by changing the multiplier on the point source discharge, increasing the PERLND/IMPLND retention storage terms, and shifting land area from a developed category back into the forest category. Because no screening tool for water reuse is available, water reuse is an example of where detailed modeling is needed to assess mitigation effectiveness. Figure 4.23. Hypothetical Cherry Creek Colorado highway example using HSPF. The modeler reruns the HSPF model with this mitigation technique and examines the time series output representing simulated flows at the outlet of subbasin 208 for three model scenarios: 1) the baseline condition, 2) the proposed highway condition without mitigation, and 3) the proposed highway condition with mitigation. Figure 4.24 summarizes the resulting flow-duration curves for all three scenarios. At the higher flows, the proposed highway increases the peak stream flows (red line). With the hypothetical mitigation technique in place the peak streamflow (green) is somewhat lower than the non-mitigated highway scenario, but not enough compared to the baseline (blue). The results show that this mitigation technique has a much more pronounced effect on the mid-range and low flows as shown in Figure 4.25. The effects are also visible in the time series plot shown in Figure 4.26 for a portion of the simulation time span. Based on this evaluation the mitigation technique successfully mitigates the effects of the proposed transportation project for the more frequent flows, but not for the higher flows.
64 Figure 4.24. Cherry Creek flow-duration curves. Figure 4.25. Close-up of the higher flows from the Cherry Creek flow-duration curves.
65 Figure 4.26. HSPF time series plot. 4.5.2. Development of New Site-Specific Watershed Models For step 11 of the decision framework, watershed models such as HSPF, SWMM, or SWAT could be used for a more detailed evaluation of out-of-kind mitigation techniques. These evaluations would be applied when screening methods are insufficient for a project. If existing watershed models are unavailable the State DOT project team may choose to develop a new model for the watersheds of interest. Development and application of a new watershed model is an extensive and watershed-specific process. If this analysis is required, the methods to model out-of-kind mitigation techniques in these models are described in Section 4.3. Once the method to represent the out-of-kind technique is selected the remaining model setup can follow the requirements of the selected model.
