Hydromodification Management Plan. Draft Report Chapters 4-6. Lower Silver - Thompson Creek Subwatershed. April 7, 2003

Transcription

1 Hydromodification Management Plan Draft Report Chapters 4-6 Lower Silver - Thompson Creek Subwatershed April 7, 2003 by GeoSyntec Consultants Walnut Creek, CA Balance Hydrologics, Inc. Berkeley, CA Philip Williams & Associates San Francisco, CA Raines, Melton, and Carella, Inc. Lafayette, CA 94549

2 Table of Contents 4 HYDROLOGIC MODELING Modeling Approach HEC-HMS Model Drainage Area Delineation Drainage Area Characteristics Excess Rainfall Hydrograph Generation Reach Routing Precipitation Model Verification Event Based Model Continuous Model Event-Based Model Results Continuous Model Results Stability Assessment Overview Channel Stability and Equilibrium Natural Erosion Processes Channel Stability and Thresholds Work Concepts Failure Mechanisms Methods i F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

3 5.3.1 Effective Work Index Defining Stable Conditions Summary of Development Scenarios Used in the Stability Analysis Determining Critical Values Hydraulic Computations Results (Pre-Urban and Existing Conditions) Percentage of Watershed Imperviousness Testing the Assessment Methods Ability to Predict Instability Erosion Potential (Ep) Selecting a Threshold of Adjustment Evaluation of Results Using Discrete Storm Events Comparing Effective Work Comparing Erosion Potential Correlations of Total Effective Work Predicting Erosion Potential Under Future Build-Out Conditions Effective Work Predicted Erosion Potential Conclusions Effective Work Erosion Potential (Ep) Threshold of Adjustment Hydromodification Control Standards and Measures Hydromodification Control Standards ii F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

4 6.1.1 Peak Flow Control vs. Duration Control Example Hydromodification Control Standards Hydromodification Control Standard Options Flow Duration Control Erosion Potential Control Flow Duration versus Erosion Potential Control Potential Effectiveness of Project Control Measures in the Lower Silver Thompson Creek Subwatershed Hydromodification Controls Measures Project Controls Drainage Area Controls In-Stream Controls Glossary References Appendix C Modeling Appendix D Stability Assessment iii F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

5 List of Figures All figures are provided at the end of this document. Figure 4-1: Figure 4-2: Figure 4-3: Figure 4-4: Figure 4-5: Figure 4-6: Figure 4-7: Figure 4-8: Figure 4-9: Figure 4-10: Figure 4-11: Figure 4-12: Figure 4-13: Figure 5-1: Figure 5-2: Figure 5-3: Figure 5-4: Figure 5-5: Drainage Area Delineations for the Lower Silver Thompson Creek Subwatershed Soil Classifications for the Lower Silver Thompson Creek Subwatershed Land Use for the Lower Silver Thompson Creek Subwatershed Mean Annual Precipitation (inches/year) Flood Frequency for Junction J-5 Using Event-based Model Results Flood Frequency for Junction J-7 Using Event-based Model Results Flood Frequency for Junction J-9 Using Event-based Model Results Flood Frequency for Junction J-10 Using Event-based Model Results Flood Frequency for Junction J-11 Using Event-based Model Results Flood Frequency for Junction J-12 Using Event-based Model Results Flood Frequency at Cross-section YB-7 Using Discrete Event-based Model Results Flood Frequency for Junction J-5 Using Continuous Model Results Flood Frequency for Junction J-12 Using Continuous Model Results Example of Erosion Observed at Cross-sections TC3-1 and TC5-4 Conceptual Example of Bank Sloughing and Retreat Histogram of Flows, Excess Shear Stress and Effective Work Curve For Segment TC5-7 Areas and Percent Imperviousness for Sub-basins Draining into Cross-sections in Segment TC5 Effective Work Curves for Stream Segments TC2, TC3 and TC5: Continuous Modeling Using Pre-Urban Flows with Bed Resistance iv F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

6 Figure 5-6: Figure 5-7: Figure 5-8: Figure 5-9: Figure 5-10: Figure 5-11: Figure 5-12: Figure 5-13: Figure 5-14: Figure 6-1: Effective Work Curves for Stream Segment TC5: Continuous Modeling Using Existing Flows with Bed Resistance Total Effective Work: Continuous Modeling Using Existing and Pre-Urban Flows with Bed Resistance Erosion Potential Chart: Continuous Modeling Using Existing Flows with Bed Resistance Probability of Erosion Instabilities Comparisons of Cumulative Effective Work Between Discrete and Continuous Modeling Using Existing Flows with Bed Resistance Total Effective Work Using Discrete Event Modeling Erosion Potential Chart for a 2-Year Event: Discrete Event Modeling Using Existing Flows Correlations Between Predicted Effective Work for Discrete and Continuous Modeling Correlation Between Continuous Modeling and Discrete Modeling for the Summation of a 2, 5, 10, 25 and 50-Year Event Flow Duration Control Concept v F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

7 List of Tables Table 4-1: Soil Moisture Accounting Parameters.5 Table 5-1: Range of Critical Shear Stress and Velocity for Measured Bed Material in Thompson Creek..18 Table 5-2: Selected Critical Values for Bank Material Found in Thompson Creek 18 Table 5-3: Comparison of Flows From Measured High Water Marks 21 Table 5-4: Comparison of Estimated HMP Channel Capacities to SCVWD Flood Control Estimates 22 Table 5-5: Comparison of Geomorphically Significant Flows and Discrete Events..25 Table 5-6: Percent Increase in the Erosion Potential Under Future Build-Out Conditions for Unstable Cross Sections 31 Table 5-7: Percent Increase in the Erosion Potential Under Future Build-Out Conditions for Stable Cross-Sections.31 vi F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

8 4 HYDROLOGIC MODELING This chapter describes the project team s hydrologic modeling of the Lower Silver Thompson Creek subwatershed. Modeling was conducted to simulate creek flows under existing, future and pre-urban conditions for use in the hydromodification assessment and planning process. 4.1 Modeling Approach The project team modeled creek flows under existing, build-out and pre-urbanization land use conditions using the U.S. Army Corps of Engineers Hydrologic Engineering Center - Hydrologic Modeling System (HEC-HMS) rainfall-runoff model in conjunction with the HEC Geospatial Hydrologic Modeling Extension (HEC-GeoHMS). The U.S. Army Corps of Engineers Hydrologic Engineering Center developed the HEC-HMS to supersede the HEC-1 Flood Hydrograph Package. Unlike HEC-1, HEC-HMS allows continuous hydrograph simulation over long periods of time in addition to event-based analysis. As described below, project team modelers ran the HEC-HMS in both continuous and event-based modes for this assessment. HEC-GeoHMS was used to analyze spatial data and generate inputs to HEC-HMS directly from GIS. The Soil Moisture Accounting (SMA) module in HEC-HMS allows for a continuous accounting of soil infiltration and other losses over a long time series of historical rainfall data. This is in contrast to event-based modeling, which assumes simplified losses for a single rainstorm event, normally generated from a synthetic design rainfall hyetograph. The continuous model is designed to model the dynamic effect of soil infiltration and other losses on storm runoff over the course of the long-term rainfall record. It can therefore be used to identify the hydromodification effects of development on small, frequent flows and to evaluate their impacts on stream stability. Event-based modeling is also useful because it provides a method for comparing hydrograph results under different land use conditions for statistically relevant design storms. In addition, event-based modeling is a convenient and commonly accepted approach for evaluating flood risk and design alternatives. However, the smallest design storm that can normally be evaluated using this approach is the two-year event, so it is less useful in evaluating the effects of hydromodification. 4.2 HEC-HMS Model The following sections describe the methods and data sources used to generate the inputs to the HEC-HMS model. The modeling approach generally followed methods and procedures for HEC-1 modeling outlined in the Santa Clara Valley Water District s Hydrology Procedures (SCVWD, 1998) as noted below. A previous hydrology study performed by Nolte Associates for SCVWD (Nolte, 2000) also provided background information for comparison, both of the estimated flow peaks, and application of SCVdWD procedures. 1 F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

9 4.2.1 Drainage Area Delineation Using GIS data, the project team defined approximately 150 catchments associated with individual storm drain outfalls and undeveloped areas in the Lower Silver - Thompson Creek subwatershed. The catchments were consolidated to reflect existing local drainage and land-use patterns, resulting in a total of 50 drainage areas comprising the subwatershed (see Figure 4-1 and Appendix C, Figure C-1)). To the extent possible, individual drainage areas were delineated so as not to include both urban and undeveloped areas, since many model parameters are derived from a drainage area s weighted average characteristics. Drainage areas were further consolidated for the pre-urbanization land use scenario (Appendix C, Figure C-2). Drainage area consolidation is summarized in Appendix C, Table C Drainage Area Characteristics Based on GIS land use and soils data, 15 different land uses and 20 NRCS soil types (in 3 different NRCS Hydrologic Soil Groups) were identified in the Lower Silver - Thompson Creek subwatershed (see Appendix C, Tables C-2 through C-4). The project team overlaid the drainage area delineations on those data to derive soil and land use characteristics used in modeling for each drainage area (see Figures 4-2 and 4-3). Existing hydrologic conditions for the Thompson Creek watershed were modeled using detailed soils and land use GIS data from SCVWD. The land use data were then modified to model hydrologic conditions for future and past (pre-urbanized) conditions, since GIS data were not available for these scenarios. For future conditions, the percentage of impervious land for each subwatershed under current conditions was increased based on future build-out percent impervious information from the current general plan. All other land uses for each subwatershed were then decreased in proportion to the increase in impervious area. The project team reviewed aerial photos from 1964 (County of Santa Clara, 1968) and USGS topographic maps of the region to characterize the pre-urbanized land use conditions in the Thompson Creek watershed. These sources provided a representation of the pre-urbanization distribution of urban, agricultural, and woodland/grassland areas for each subwatershed, which was then converted into model input parameters Excess Rainfall HEC-HMS uses estimates of the amount of water that infiltrates into the soil as well as other losses described below to calculate the excess precipitation that contributes to stormwater runoff. The event-based HEC-HMS modeling used the Soil Conservation Service (SCS) method to estimate losses, and the continuous simulation used the Soil Moisture Accounting (SMA) method (unique to HEC-HMS). 2 F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

10 SCS Method (Event-based Model) The SCS method uses a watershed runoff curve number (CN) to represent runoff potential for various soil types and land uses. To determine CN, the project team first assigned a soil infiltration potential to each identified soil type by using the NRCS Soil Hydrologic Group designation (A, B, C, or D with soil type A having the highest infiltration potential). Using guidance from Hydrology Procedures (SCVWD, 1998) and assuming an average antecedent moisture condition (AMC II), the team then assigned a CN value for each land use and soil group combination. Low CNs reflect a high infiltration potential (see Appendix C, Table C-4). In regions where the land use indicated impervious cover, the CN value was set at the maximum value (CN = 98) regardless of soil type. The project team then generated a CN for each drainage area using an area-weighted average of CN values present. The project team also used the SCS method to estimate initial abstraction losses as a function of curve number (SCVWD, 1998). These parameters are summarized by drainage area in Appendix C, Table C-5. Soil Moisture Accounting Method (Continuous Model) The SMA method provides a more complex method of evaluating rainfall runoff processes in a watershed. In this approach, actual measured rainfall over an extended time period is used as input. Losses are computed on a continuous basis, and include evaporation, surface storage, and infiltration. Parameters to compute these losses include climatic data, land use conditions, vegetation cover, and soils data. Table 4-1 summarizes the types of losses estimated by SMA, along with the data sources and methods used to develop the required input data. Further details are provided in Appendix C, Tables C-2 through C-4. Parameters are listed by drainage area in Appendix C, Table C-6. For each computational time step in the model, HEC-HMS calculates storage in each of the loss categories outlined in the table, which allows for a continuous accounting of losses and runoff over a long time series. For infiltration, the model initially assumes that water enters the soil at the maximum infiltration rate and percolates out of the soil at the maximum percolation rate. Once the soil layer becomes saturated, the infiltration rate is reduced to the percolation rate Hydrograph Generation HEC-HMS offers a variety of methods to transform the excess precipitation from any given storm into a runoff hydrograph for each model drainage area. SCVWD s Hydrology Procedures (1998) recommends and provides guidance for using Clark s synthetic unit hydrograph method in HEC-1. That guidance was used to estimate the two drainage area parameters required for Clark s method, time of concentration (T c ) and storage coefficient (R) (see Appendix C, Tables C-7 through C-11). The procedure described in Hydrology Procedures (SCVWD, 1998) requires further sub-division of urbanized drainage areas into pervious and impervious areas, and those areas that drain to F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003 3

11 storm drains. Runoff passing through storm drains is then routed using the Modified Puls routing method contained in the model using typical storage-discharge relationships for Santa Clara Valley storm drains (SCVWD, 1998). Under the SCVWD procedure, hydrographs from pervious and impervious urban areas within a drainage area are routed through the storm drain system, and then combined with the hydrograph for the portion of the drainage area (if any) that does not enter the storm drain system (see Appendix C, Table C-12 and Figure C-1). This routing and combination routine is internal to each drainage area, in contrast to reach routing between drainage areas (described below) Reach Routing HEC-HMS provides a variety of reach routing methods to translate the hydrograph from one drainage area downstream to a point where it can be combined with another drainage-area hydrograph. This assessment uses the Muskingum method, which uses basic channel (or culvert) dimensions and characteristics to estimate hydrograph translation and attenuation over the routing reach. For existing and future conditions, a combination of surveyed cross-sections, available storm drain data, and information from Nolte, 2000 was used to characterize channel dimensions and characteristics for reach routing. For the pre-urbanization scenario, project team modelers made reasonable assumptions regarding channel characteristics and flow paths, and verified these assumptions using the 1964 aerial photos. Reach routing parameters are summarized in Appendix C, Tables C-13 and C F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

12 Table 4-1 Soil Moisture Accounting Parameters Input Parameter Source Method Maximum Canopy Interception Storage Land use layer (GIS) from William Lettis & Assoc. (WLA) Vegetation-interception relationships from literature Mean Annual Precipitation (MAP) data from SCVWD 1. Interception storage assigned for each vegetation type present using published data from similar environments (% of MAP) 2. MAP assigned for each drainage basin 3. Interception converted to depth using precipitation depth for frequent design storm from SCVWD 1998 (1 ); depths verified vs. values from literature Maximum Surface Depression Storage Land use data (GIS) from WLA HEC-HMS Manual 1. Surface storage depression value assigned per land use (GIS) using HEC-HMS manual guidance 2. Drainage area surface storage determined as a weighted average of land use using GIS data Maximum Infiltration Rate Soil data from SCVWD database Soil parameters estimated from the literature (Rawls et al., 1982) 1. The initial maximum infiltration rate was estimated for each of the Hydrologic Soil Groups present (B, C, and D) using the Green-Ampt equation and soil parameter estimates from the literature. 2. All infiltration rates were then scaled up during model verification Maximum Percolation Rate Maximum Soil Profile Storage Tension (Unsaturated) Zone Storage Maximum Groundwater Layer Storage Soil data from SCVWD database Saturated hydraulic conductivity (K s ) for representative soils estimated from literature Soil data from SCVWD database Soil profile depth from Santa Clara County Soil Survey (NRCS) and effective porosity from Rawls et al, 1982 Soil data from SCVWD database HEC-HMS manual and Santa Clara County Soil Survey (NRCS) N/A 1. The Maximum Percolation Rate was set equal to saturated hydraulic conductivity (K s ) 1. Soil profile storage for each soil type taken from SCVWD data (in GIS) as a product of effective soil depth (root zone depth) and effective porosity. 2. Weighted average Maximum Soil Profile Storage determined for each drainage area using SCVWD data in GIS 1. Tension zone storage estimated as Available Water holding Capacity (A.W.C) from the County Soil Survey 1. Following guidance from the literature, groundwater storage and percolation rate were not considered in this assessment due to the lack of base flow in the creeks 5 F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

13 4.2.6 Precipitation The project team ran the HEC-HMS model using both event-based and continuous precipitation data as described below. Event-based Model In this approach, synthetic rainstorms were developed that were intended to simulate various magnitude flood events. Rainfall depths for the 2, 5, 10, 25, 50, and 100-year storms (24 hour duration) were estimated using the return period-duration-specific (TDS) regional equation provided in Hydrology Procedures (SCVWD, 1998). The TDS equation provides a rainfall depth for each storm based on a site s mean annual precipitation (MAP) and coefficients developed by SCVWD from long-term rainfall records (See Figure 4-4). The project team estimated rainfall depth for each design storm for each model drainage area based on the distribution of MAP. The project team modelers used GIS to divide MAP into nine categories, each covering a half-inch increment in MAP, and estimated rainfall depth for each category based on the mean MAP for that category. One of the nine categories (and associated rainfall depth) was assigned to each model drainage area based on average MAP for that drainage area (also determined by GIS). The balanced hyetograph routine in HEC-HMS generated storm hyetographs for each return period based on the rainfall depths described above. A storm hyetograph describes the time distribution of the rainfall intensity over the duration of the storm. A balanced hyetograph provides a distribution of the rainfall that maintains the same recurrence event for all time steps within the storm; thus, a balanced 100-year, 24-hour hyetograph will include 100-year rainfall rates for the 0.5-hour storm, the 1-hour storm, and the 3-hour storm imbedded within the rainfall distribution. This method is useful in simulating runoff in a multi-drainage area watershed model with drainage areas of different sizes, as in this study Continuous Model The continuous simulation was run using National Climatic Data Center (NCDC) continuous, hourly rainfall data from gage station ID , the City of San Jose gage. This gage corresponds to SCVWD station 6086, which was moved in 1986 to a nearby location known as Station 6131 (San Jose Airport). The data collection method was also changed at that time, from a 0.01-inch recording increment to a 0.10-inch increment. This change is reflected in the rainfall record, in that the smallest rainfall event recorded after 1986 is 0.10 inch. Using a process similar to that described for the event-based model, the project team scaled the continuous rainfall record for each drainage area based on drainage-area MAP. 6 F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

14 4.3 Model Verification A variety of parameters must be estimated in any hydrologic model. These estimated parameters affect the size and shape of the storm hydrograph predicted by the model compared to what may result from any individual actual storm. Whenever possible, modelers compare model results to recorded concurrent rainfall and flow data to calibrate the model by adjusting various parameters to reproduce the actual flow resulting from measured rainfall. While it is preferable to calibrate models to recorded flow data, uncalibrated models can still be very useful in assessing relative results for different land use scenarios and management options No historic stream gage data are available to calibrate the HEC-HMS model to actual flows in the Lower Silver - Thompson Creek system. However, the SCVWD stream gage on Thompson Creek at Quimby Road has been functional since January 2003, and additional flow data for Thompson Creek is currently being collected as part of a separate project. These sources should provide useful calibration data in the future. Until the model is calibrated, model results for continuous simulation should be used for comparative assessments only (such as comparing various land use and control alternative scenarios). In the absence of calibration data, the project team compared model results to other available peak flow estimates in order to verify the reasonableness of the model results. Peak flow estimates used in this comparison include those from SCVWD 1998 ( flood quantiles or regional regression equations), Nolte 2000, and FEMA 1986, as described below. In addition, we estimated peak flows from a December 2002 storm from high water marks observed in Thompson Creek. Rainfall records from that storm were then run through the model and the results compared to the estimated peak flows Event Based Model The project team compared peak flows generated by the event-based model to peak flows predicted by the regional regression equations provided in SCVWD The regional equations provide a method for estimating peak flows for a variety of return periods, based on contributing subwatershed area, MAP and other parameters (SCVWD, 1998). The equations are derived from analysis of long-term historical stream gage records for the region, and are generally assumed to represent undeveloped watershed conditions. Eventbased results were also compared to flow estimates provided in FEMA 1986 and Nolte Some model parameters were adjusted as a result of this verification process Continuous Model A continuous record of stream gage data is not available to calibrate the continuous simulation of the HEC-HMS model to actual flows in the Lower Silver - Thompson Creek system. The project team therefore used the results of the event-based model to verify that F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003 7

15 the continuous model produces reasonable results. The project team identified peak flows for the 2, 5, 10, 25, 50 and 100-year return periods from the continuous model results by performing a statistical flood frequency analysis of model results for the current land use condition. We then compared these peak flows to those estimated by the event-based model for the same return period and land use scenario, and adjusted the initial estimates of SMA model parameters to produce results that were in reasonable agreement with the event-based model. 4.4 Event-Based Model Results Figures 4-5 through 4-11 show event-based model results for past, existing and future land use conditions at each segment of Thompson Creek, and for the upper portion of Yerba Buena Creek (see Figure 4-1 for junction locations). Peak flows for a variety of return periods are plotted together for the three land use scenarios. This comparison demonstrates the effect of urbanization on peak flows across the spectrum of return periods, with the most dramatic impact being on the smaller, more frequent storms. The results of the regional regression equations are also plotted on Figures 4-6 and 4-10 for comparison. Although the model was not specifically calibrated to the regional regression equations, the peak flows generated by the HMS model show reasonable agreement with those estimated by the equations, especially recognizing the fact that the equations are assumed to most applicable to undeveloped watershed conditions. A summary of model results (peak flows and discharge volumes) is presented in Appendix C, Table C-15, for relevant locations on Thompson Creek and Yerba Buena Creek. In addition, model results at J3 and J12 for selected return periods are presented in Appendix C, Figures C-3 through C Continuous Model Results This section presents data from the continuous model for past, current and future land use conditions. Due to the large amount of data generated by the model for each land use scenario (50 years of flow estimates at half-hour intervals for multiple locations), summary results for Junctions 5 and 12 are presented in this section to demonstrate representative model results for the upper and lower portions of the Thompson Creek subwatershed, respectively. Figures 4-12 and 4-13 show the results of a flood-frequency analysis of continuous model results for the three land use conditions at the two representative locations. Figure 4-13 shows results for the lower portion of the watershed, reflecting aggregated hydromodification effects of development throughout the watershed. These results demonstrate the fact that the watershed is largely built-out, so that the projected hydromodification between the current 8 F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

16 and future conditions is much less than what has occurred from the 1960 s ( pre-urban condition) to the present. A comparison of 10-year peak flows at J12 (Figure 4-13) demonstrates an important aspect of hydromodification. As shown in Figure 4-13, the 10-year peak flow for the pre-urban condition (approximately 1500 cfs) now occurs approximately every 2.5 years under existing conditions, according to the continuous model results. This represents a dramatic increase in the frequency of flows of this magnitude, which has implications for stream stability as discussed in Chapter 5. The increase flow frequency is less dramatic between current land use conditions and future, but is still potentially significant. The model results for J5 (Figure 4-12) show a less dramatic difference between pre-urban and existing conditions, since the upper part of the watershed is not yet built out. However, a similar comparison peak flows show that the 10-year peak flow under pre-urban conditions now occurs approximately every six years under current conditions, according to the model results. 9 F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

17 5 Stability Assessment This chapter describes a method to evaluate how changes in stream flow affect channel erosion and stability. The method is applied to the Lower Silver - Thompson Creek subwatershed to evaluate its ability to predict unstable conditions and its effectiveness in distinguishing between levels of development. The stability assessment results are preliminary and will be refined as the assessment method is applied to the second and third sub-watershed. The stability assessment combines information from the modeling work and from the field geomorphic assessment to address erosion potential. The analysis provides estimates of the erosion potential of the stream segments under existing conditions, future built-out conditions, and built-out conditions with hydromodification controls. Previous work completed by the SCVURPPP to develop goals and objectives, a conceptual model, and a literature review were used in the development of this assessment method (GeoSyntec 2002). These work products and the proposed assessment methodology were reviewed by the expert panel. 5.1 Overview As explained in Chapter 1, increases in impervious surfaces caused by development and the associated changes in runoff clearly destabilize streams. In the Lower Silver - Thompson Creek subwatershed, development has created a large percentage of impervious surfaces that are interconnected via rain gutters, storm drains, and open channels discharging into the creeks. The increase in stream flow due to an increase in impervious surface and connectivity is the most significant cause of the observed channel erosion in Thompson Creek. The stability assessment methodology is based on the premise that a balance among flow energy, sediment loads, and channel resilience must be maintained in order for the stream network to remain stable. By applying this method and establishing management criteria, the intent is to maintain the natural stream sediment transport and erosion processes, not eliminate them. In watersheds that are already developed and the stream impacted, the intent is to reestablish this balance. The assessment method measures erosion potential by using an index measure representing the effective work done by flow energy in excess of the amount required to transport the available sediment load. To measure work done using continuous simulation, the method integrates the total time (duration) that flows exceed the critical shear stress for initial motion of bed material or erosion of bank material over a rainfall record of 50-years. Stream sections that are stable are used as a baseline against which to measure effective work and predict the erosion potential for existing and future conditions. For this assessment, 10 F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

18 stable sections are assumed to have existed prior to urbanization, which began in the late 1970 s. The 1968 historical photos show a densely vegetated riparian corridor and show no signs of canalizations or drainage ditches in this portion of Thompson Creek. Pre-urban cross sections were obtained from 1968 topographic maps with 2-foot contour intervals of Thompson and Yerba Buena Creeks. These cross sections show relatively gradual bank side slopes with no noticeable incision. A threshold of adjustment is considered that distinguishes between stable and unstable channel conditions. In natural stable stream systems, channels can tolerate a certain amount of variation in the stream flow and sediment characteristics without catastrophic failure through natural self-regulatory mechanisms. A disturbance of sufficient magnitude and duration that exceeds this innate ability to self-regulate is referred to as the threshold of adjustment. Once the imposed hydromodification changes exceed this threshold, channel incision and bank erosion begins (or other forms of adjustments occur depending on the stream system and local conditions). The approach involves the following: Computing the hydraulic parameters depth, velocity, and shear stress for the modeled stream flow rates (see section 5.3) Setting critical values for velocity and shear stress for initial motion of bed and for erosion of bank material (see section 5.3) Measuring the cumulative work done (see section 5.3) on the stream boundary based on model-predicted flows over a 50-year period of rainfall record (see Chapter 4) Quantifying potential stream boundary erodibility (see section 5.4) and using existing land use and field identified erosion conditions to test the effectiveness of the method Identifying a stability threshold that can be used to evaluate potential impacts from future development and for implementation of management strategies (see section 5.4) Measuring work done based on the methodology used for the prior step, but assuming future (build-out) land use (see section 5.4) Comparing the work done by flows in current stable channel reaches and under future land use conditions to assess the future channel erosion potential (see section 5.4) 11 F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

19 5.2 Channel Stability and Equilibrium Natural Erosion Processes Natural processes of weathering, landslides, debris flows, and hill slope erosion supply sediment to the valley floor and stream channels. This material is transported downstream, broken down into smaller material, deposited, scoured, and continually reworked. The hydraulic energy of stream flow imposes a shear force that cuts into the erodible soils and rock and helps shape the stream channel network. Over time, channels evolve to approximately stable equilibrium conditions that balance the imposed flow energy and sediment load with the channel boundary materials ability to resist erosion. The processes of runoff and sediment transport interact with the boundary materials establishing a cross sectional geometry, longitudinal slope and planform. As vegetation co-evolves with the channel, it too influences channel stability and shape Channel Stability and Thresholds A stable channel is defined as one that neither aggrades nor degrades, but instead maintains its average cross-section, planform, and profile features over time and within a range of variance. Several researchers have defined the equilibrium concept as one where the spectrum of discharges, slope, and channel geometry are adjusted to provide just the right velocity to transport the sediment load supplied to the system (e.g., Knighton 1998). Stream systems in equilibrium can tolerate a certain amount of variation in its flow and sediment loads through natural self-regulatory mechanisms. A stable channel can tolerate short-term disturbances without catastrophic failure. A disturbance of sufficient magnitude and duration that exceeds the ability to self-regulate is defined as the threshold of adjustment. This adjustment is observed as erosion, which may be observed in just a few storms or it may take years or decades to develop. Under such conditions, streams can migrate, change course, or incise into underlying materials, which would otherwise not be affecting the stream. Real stream systems may never truly be in perfect equilibrium, but over the time period of observation, stream systems tend to maintain consistent measurable characteristics within its variation, or if disturbed, streams tend to return to approximately their previous state Work Concepts In addition to the equilibrium concept, there also is a range of flows that are considered most important in defining channel form, adjustment and controlling the rate at which sediment is transported through the stream system (Leopold et al. 1964). Leopold et al. (1964) showed that a large percentage of the work done is performed by frequent flow events of moderate magnitude defined as geomorphically significant flows. Research has shown that urbanization significantly alters the frequency and duration of geomorphically significant flows (Hammer 1972, Hollis 1975). Bledsoe and Watson (2001) reported that the frequency of these flows increases by factors of 2.5 to 5 for watersheds with 18 percent impervious 12 F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

20 cover. MacRae et al. (1992, 1993) showed that the greatest increase in channel erosion results from increases in the small and moderate sized flow events. These small but frequent flow events have the energy to move sediment and erode stream bank material, and cumulatively have more influence over channel erosion and adjustment than infrequent larger storm events. Additionally, they can weaken the bank and accelerate the failure mechanism during larger storm events. As stated earlier, the stability assessment methodology is based on the premise that a balance among flow energy, sediment loads, and channel resilience must be maintained and that the flows to be managed are primarily the geomorphically significant flows. The intent of hydromodification management is to maintain a watershed s erosion processes and sediment transport characteristics, not eliminate them. In watersheds that are already developed and the stream impacted, the intent is to reestablish an approximate equilibrium between the current boundary conditions and sediment loads, and its new hydrologic regime Failure Mechanisms Channel erosion and adjustment can occur through a combination of several mechanisms, all of which are observed in the Thompson Creek subwatershed, although one mechanism may be more prevalent than others in any one segment: Channel incision and under-cutting of the bank toe due to shear erosion Slumping from over-steepened bank slopes or increased groundwater pore pressures following waning flood stages Loss of bank vegetation, reducing roughness and bank strength Water forced into the banks from obstructions such as boulders or large woody debris Shear erosion is the primary mechanism of erosion and is the primary mechanism considered in this assessment methodology. Channel incision and bank toe erosion is considered the beginning of channel instabilities, although other mechanisms may be observed as ultimate failure. Channel incision and shear erosion at the toe of stream banks increases the height of banks and oversteepens them, priming them for failure by slumping during larger flows. Figure 5-1 illustrates the observed channel adjustments due to erosion at two example crosssections TC3-1 and TC5-4. High flow events, rapid recession, and over saturation of soils can contribute to bank collapse. Creek banks can fail from excessive weight and internal groundwater pore pressure after they have been oversteepened, causing slumping of bank material into the stream channel. Figure 5-2 illustrates two types of bank slumping observed in the Thompson Creek subwatershed. 13 F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

21 Vegetation also influences channel form and its resistance to hydraulic shear. Dense vegetation adds roughness, slows flow velocity, reduces shear stresses on stream banks, and adds apparent soil cohesion through the root structure. Stream channel destabilization is often attributed to the removal of vegetation, especially if the pre-urban balance was established with vegetation present. Replacing vegetation while simultaneously reducing the shear stresses can contribute to channel stability. Local scour and erosion can be created by local obstructions that direct stream flows into the bank. This is common in many stream systems and is natural when the obstruction is not human induced (e.g., felled trees). 5.3 Methods Effective Work Index The stability assessment is based on measuring the magnitude of effective work done (W) by flows that exceed a specified critical value for the streambed or bank material at cross sections shown in Figure 3-2. W is a dimensionless index that represents the total work done on the channel boundary. This measure takes the following form: W = n ( τ τ ) t (1) 1 i c 1.5 where: W = index of total effective work done over the length of flow record. n = length of flow record τc = critical shear stress that initiates bed mobility or shear erosion τi = applied hydraulic shear stress, computed as ρgds t = duration of excess (in hours) d = depth of water S = longitudinal slope g = gravity constant ρ = density of water The time increment ( t) is determined by generating a histogram of the flows from the continuous simulation results, which are hourly data. For each flow range (Bin), the histogram provides the count or duration of time that flows are within the designated flow range. As an example, Figure 5-3 shows the flow histogram for cross section 5-7. There are 14 F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

22 143 hours of flow within the 5 to 10 cfs flow Bin. For the average flow within this range, the depth, velocity, and shear stress are computed. Equation 1 is solved for each flow Bin and multiplied by the duration ( t) to compute the incremental effective work done by that specific range of flows. The incremental effective work for each flow Bin represents the effective work curve shown in Figure 5-3. When equation 1 is summed over the 50-year flow record (histogram), the result (W) is an index measure of the total effective work done on the stream channel boundary. The total effective work index for stable stream channels in pre-urban conditions and undeveloped drainage areas (or drainage areas as close to undeveloped as possible) are compared to unstable channels in urbanized drainage areas. The comparison, expressed as a ratio, is defined as the Erosion Potential (Ep) (McRae, 1992, 1996). W unstable Ep = (2) W stable W unstable = work index for a stream section determined to be unstable (medium and high observed erosion, see Table 3-2) W stable = work index for a stream section determined to be stable (stable and low observed erosion, see Table 3-2) Defining Stable Conditions The stability assessment requires stable baseline conditions to compare to the existing and future conditions and predict erosion potential. Stable baseline conditions are being represented by pre-urban watershed conditions and modeling. Both the hydrologic modeling and the stream hydraulic modeling were completed using 1968 watershed and stream characteristics as provided by the District (SCVWD 1973). The 1968 historical photos show a densely vegetated riparian corridor throughout most of the study area and show no signs of canalizations or drainage ditches routing runoff to Thompson Creek, or within Thompson Creek. Pre-urban cross sections were obtained from 1968 topographic maps with 2-foot contour intervals of Thompson and Yerba Buena Creeks. These cross sections show relatively gradual bank side slopes with no noticeable incision, and resemble the stable cross section observed in the watershed Summary of Development Scenarios Used in the Stability Analysis This section outlines the characteristics of each development scenario considered in the stability analysis. 1. Pre-Urban Conditions 15 F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

23 a. Represented by the pre-urbanized watershed of 1968 with 1968 stream cross sections. b. It is assumed that the magnitude and frequency of shear stresses within the channel during the pre-urban period was associated with relatively stable streams, and serves as the default baseline. c. It is assumed that pre-urban bed and bank characteristics resembled that of currently healthy stable stream reaches observed today. d. The critical shear stress used to compute excess work is based on bed material, which is the least resistant boundary material observed. e. For currently healthy stream reaches, or slightly to moderately eroded, this pre-urban baseline becomes the condition to be preserved, or restored. 2. Existing Conditions a. Represented by existing watershed development and existing channel geometry and slope. b. Comparison between Pre-Urban and Existing shows the magnitude of hydromodification and helps determine which stream reaches, if any, can be restored to approximately pre-urban conditions. c. The excess work computed for existing healthy stable stream reaches should coincide with the pre-urban condition. The critical values are based on bed material for these reaches. d. For impacted stream reaches that have incised to hardpan or stiff clay layers, mobile bed material is no longer influencing channel geometry. In this case, the boundary material properties (i.e., its defined critical shear stress) of the weakest layer of unconsolidated sediments will be used when evaluating solutions for stabilization. 3. Future Conditions a. Represented by future watershed development and existing channel geometry and slope. b. Comparison between existing and future conditions will show which current healthy stable stream reaches could be affected or further affected by future development. c. Comparison between baseline (pre-urban) and future will show how much (and what types of) management might be required to: i. Preserve and maintain existing healthy stable reaches ii. Restore slightly to moderately incised and/or widening reaches iii. Stabilize and enhance extensively eroded reaches 16 F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

24 d. Existing boundary material properties as described in item 2 above will be used to evaluate the magnitude of controls required to reestablish stability. 4. Future Conditions with Control Measures a. Same as item 3 above b. Evaluates the effectiveness of control strategies, which may include: i. Source control of runoff ii. Appropriately reduced longitudinal slopes iii. Appropriately expanded channel geometries iv. Detention or other measures which may reduce peak flows and durations v. Restoration of woody vegetation of sufficient density and corridor width Determining Critical Values The ability of a stream bank to resist erosion depends on soil materials, stratigraphy, vegetation density, root strength, the degree of cohesion, bank height, and slope. Boundary material also influences vegetation assemblages which in turn provide resistance to bank erosion. Stream channels bounded by clays, compacted silts, and loess are often more resistant to erosion and respond more slowly to hydrologic changes than channels bound by loosely consolidated sands and gravels. Critical values of shear stress and velocity for the streambed and stream bank provide a measure of the stream s resistance to erosion. Critical values for bed material reflect the onset of sediment transport. Critical values for bank material reflect the onset of erosion of the bank, especially for weak stratigraphy layers. Streams (or boundary material) with larger critical values have more resilience to hydromodification. Chapter 3 described the physical properties of the observed streambed and bank materials, which are used to determine the critical values for the channel boundary. The bed was characterized by particle size distribution and depth. Stream banks were characterized by composition (percent clay, silt, sand and gravel). For the different bed material sizes, critical values of shear stress and velocity for bed mobility were estimated using two methods: Shield s equation and from permissible velocity tables published in ASCE Manual No. 77 (1992). Table 5-1 lists the range of estimated critical velocities and shear stresses using each method for comparison. The permissible velocity method can account for flows high in suspended sediment, such as in the lower segments of Thompson Creek where flows are typically highly turbid during storm events. For this first development of the stability assessment, the bed material is assumed to be 17 F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003

25 uniform throughout the subwatershed and an average critical shear stress and velocity were used; 0.14 lbs/sq-ft and 2.8 ft/sec, respectively. Table 5-1. Range of Critical Shear Stress and Velocity for Measured Bed Material in Thompson Creek Cross-section with Similar Bed Material D50 (mm) D84 (mm) Bed Slope Critical Bed Shear Stress (a) (lbs/ft^2) Critical Bed Shear Stress (b) (lbs/ft^2) Critical Velocity (a) (ft/s) Critical Velocity (b) (ft/s) TC TC1-6 TO TC TC2-5 TO TC TC2-9 TO TC TC TC TC5-1 TO TC TC TC TC TC YB YB Average Values a) Source: Computed using Shields Equation. Dimensionless parameter used = b) Source: ASCE Manual No. 77, page 334, Figure 9.5. The ASCE Manual No. 77 was used to estimate critical velocities and shear stress for the channel banks, including both the upper stiff clay layers and the lower weaker layers with sands and gravels. Table 5-2 lists the critical values selected for the different bank materials found in the Thompson Creek subwatershed. The composition of the bank soils was estimated qualitatively in the field. In the future, bank samples will be collected to better characterize bank material. Table 5-2. Selected Critical Values for Bank Material Found in Thompson Creek Critical Velocity Critical Shear Stress Material Type (V c ft/sec) (τ c lbs/ft^2) Riprap Hardpan F:\Sc44\Sc44-11\HMP Website\prgm_hmp_4-6_ doc 4/10/2003