Metals TMDLs for the Los Angeles River Watershed – Draft
Model Development for Simulation of Wet-Weather Metals Loading from the
Los Angeles River Watershed
May 2004
Prepared for:
USEPA Region 9
Los Angeles Regional Water Quality Control Board
Prepared by:
Tetra Tech, Inc.
1. Wet Weather Model
Wet weather sources of metals are generally associated with wash-off of loads accumulated on the land surface. During rainy periods, these metals loads are delivered to the waterbody through creeks and stormwater collection systems. Metals loads can be associated with sediment loadings, which can be linked to specific land use types that have higher relative accumulation rates of metals, higher relative loads of sediment from the land surface, or are more likely to deliver sediment and associated metals to waterbodies due to delivery through stormwater collection systems. To assess the link between sources of metals and the impaired waters, a modeling system may be utilized that simulates land-use based sources of sediment and associated metals loads and the hydrologic and hydraulic processes that affect delivery. Understanding and modeling of these processes provides the necessary decision support for TMDL development and allocation of loads to sources.
The U.S. Environmental Protection Agency’s (USEPA) Loading Simulation Program C++ (LSPC) was used to represent the hydrological and water quality conditions in the Los Angeles River watershed. LSPC is a component of the USEPA’s TMDL Modeling Toolbox, which has been developed through a joint effort between USEPA and Tetra Tech, Inc. It integrates a geographical information system (GIS), comprehensive data storage and management capabilities, a dynamic watershed model (a re-coded version of EPA’s Hydrological Simulation Program – FORTRAN [HSPF] [Bicknell et al., 2001]), and a data analysis/post-processing system into a convenient PC-based windows interface that dictates no software requirements. LSPC is capable of representing loading, both flow and water quality, from non-point and point sources and simulating in-stream processes. LSPC can simulate flow, sediment, metals, nutrients, pesticides, and other conventional pollutants, for pervious and impervious lands and waterbodies. LSPC was configured to simulate the Los Angeles River watershed as a series of hydrologically connected sub-watersheds.
2. Model Development
The watershed model represented the variability of non-point source contributions through dynamic representation of hydrology and land practices. The watershed model included all point and non-point source contributions. Key components of the watershed modeling included:
· Watershed segmentation
· Meteorological data
· Land use representation
· Soils
· Reach Characteristics
· Point Source Discharges
· Hydrology representation
· Pollutant representation
· Flow Data
2.1 Watershed Segmentation
In order to evaluate sources contributing to an impaired waterbody and to represent the spatial variability of these sources, the contributing drainage area was represented by a series of sub-watersheds. This subdivision was primarily based on the stream networks and topographic variability, and secondarily on the locations of flow and water quality monitoring stations, consistency of hydrologic factors, land use consistency, and existing watershed boundaries.
The subwatersheds for the Los Angeles River basin were delineated after dividing the watershed into two general components: headwaters and lower-elevation urban areas. The headwaters were generally more mountainous and have steeper slopes than the downstream portion of the watershed. In this mountainous region, Digital Elevation Models (DEMs) were utilized for delineating subwatersheds. Specifically, subwatershed boundaries were based upon slopes, ridges, and projected drainage patterns. Alternatively, in the downstream flatter areas of the watershed, maps illustrating the catchment network and drainage pipes were used to isolate sewersheds. The Los Angeles River watershed was ultimately delineated into 35 sub-watersheds for appropriate hydrologic connectivity and representation (Figure 1).
Figure 1. Subwatershed Delineation for the Los Angeles River Watershed
2.2 Meteorological Data
Meteorological data are a critical component of the watershed model. LSPC requires appropriate representation of precipitation and potential evapotranspiration. In general, hourly precipitation (or finer resolution) data are recommended for nonpoint source modeling. Therefore, only weather stations with hourly-recorded data were considered in the precipitation data selection process. Rainfall-runoff processes for each subwatershed were driven by precipitation data from the most representative station. These data provide necessary input to LSPC algorithms for hydrologic and water quality representation.
Precipitation data available from the National Climatic Data Center (NCDC) were reviewed based on geographic location, period of record, and missing data to determine the most appropriate meteorological stations. Ultimately, hourly rainfall data were obtained from 11 weather stations located in and around the Los Angeles River watershed for October 1988 through December 2001 (Table 1 and Figure 2).
Long-term hourly wind speed, cloud cover, temperature, and dew point data were available for the Los Angeles International Airport (WBAN #23174). These data were obtained from NCDC for the characterization of meteorology of the modeled watersheds. Using these data, hourly potential evapotranspiration was calculated.
Table 1. Precipitation and Meteorological Stations Used in the LSPC Watershed Model
Station # / Description / Elevation (ft) / Latitude / Longitude /CA1194 / BURBANK VALLEY PUMP PLA / 655 / 34.183 / -118.333
CA1682 / CHATSWORTH RESERVOIR / 910 / 34.225 / -118.618
CA3751 / HANSEN DAM / 1087 / 34.261 / -118.385
CA5085 / LONG BEACH AP / 31 / 33.812 / -118.146
CA5114 / LOS ANGELES WSO ARPT / 100 / 33.938 / -118.406
CA5115 / LOS ANGELES DOWNTOWN / 185 / 34.028 / -118.296
CA5637 / MILL CREEK SUMMIT R S / 4990 / 34.387 / -118.075
CA7762 / SAN FERNANDO PH 3 / 1250 / 34.317 / -118.500
CA7926 / SANTA FE DAM / 425 / 34.113 / -117.969
CA8092 / SEPULVEDA DAM / 680 / 34.166 / -118.473
CA9666 / WHITTIER NARROWS DAM / 200 / 34.020 / -118.086
Figure 2. Location of Precipitation and Meteorological Stations
2.3 Land Use Representation
The watershed model requires a basis for distributing hydrologic and pollutant loading parameters. This is necessary to appropriately represent hydrologic variability throughout the basin, which is influenced by land surface and subsurface characteristics. It is also necessary to represent variability in pollutant loading, which is highly correlated to land practices. The basis for this distribution was provided by land use coverage of the entire watershed.
Two sources of land use data were used in this modeling effort. The primary source of data was the County of Los Angeles Department of Public Works (LADPW) 1994 land use dataset that covers Los Angeles County. This dataset was supplemented with land use data from the 1993 USGS Multi-Resolution Land Characteristic (MRLC) dataset.
Although the multiple categories in the land use coverage provide much detail regarding spatial representation of land practices in the watershed, such resolution is unnecessary for watershed modeling if many of the categories share hydrologic or pollutant loading characteristics. Therefore, many land use categories were grouped into similar classifications, resulting in a subset of 7 categories for modeling. Selection of these land use categories was based on the availability of monitoring data and literature values that could be used to characterize individual land use contributions and critical metals-contributing practices associated with different land uses. For example, multiple urban categories were represented independently (e.g., residential, industrial, and commercial), whereas forest and other natural categories were grouped. Table 2 presents the land use distribution in each of the 35 subwatersheds.
LSPC algorithms require that land use categories be divided into separate pervious and impervious land units for modeling. This division was made for the appropriate land uses to represent impervious and pervious areas separately. The division was based on typical impervious percentages associated with different land use types defined by LADPW (DePoto et al., 1991).
Table 2. Land use Areas (square miles) of each Sub-Watershed
1 / 8.55 / 0.87 / 0.52 / 7.44 / 0 / 0 / 0.32 / 17.69
2 / 7.91 / 0.91 / 0.28 / 5.17 / 0.08 / 0.04 / 0.44 / 14.83
3 / 4.49 / 0.6 / 1.55 / 15.75 / 0.2 / 0 / 0 / 22.59
4 / 4.53 / 1.23 / 0.87 / 5.96 / 0.4 / 0.04 / 0.08 / 13.12
5 / 9.86 / 1.91 / 2.86 / 6.52 / 0 / 0 / 0.32 / 21.47
6 / 8.67 / 1.39 / 0.6 / 1.67 / 0.08 / 0 / 0 / 12.41
7 / 8.11 / 1.15 / 3.38 / 8.23 / 0.24 / 0.28 / 0.12 / 21.51
8 / 10.94 / 1.91 / 0.44 / 3.34 / 0.24 / 0.12 / 0.36 / 17.34
9 / 17.93 / 3.58 / 2.78 / 4.89 / 0.48 / 0.16 / 0.04 / 29.86
10 / 0.76 / 0 / 0 / 33 / 0.04 / 0.2 / 0 / 34
11 / 7.04 / 1.67 / 1.67 / 6.88 / 0.48 / 0 / 0.08 / 17.81
12 / 7.59 / 1.59 / 1.19 / 0.76 / 0.16 / 0 / 0 / 11.29
13 / 4.1 / 0.36 / 2.19 / 120.09 / 0.12 / 0.08 / 0 / 126.93
14 / 0.56 / 0.04 / 0.24 / 20.32 / 0.28 / 0 / 0 / 21.43
15 / 3.14 / 0.4 / 2.62 / 3.74 / 0.16 / 0 / 0 / 10.06
16 / 6.68 / 1.03 / 0.95 / 0.28 / 0 / 0 / 0 / 8.95
17 / 5.49 / 1.59 / 1.95 / 0.52 / 0 / 0 / 0 / 9.54
18 / 0.95 / 0.04 / 0 / 0.08 / 0 / 0 / 0 / 1.07
19 / 9.42 / 1.55 / 5.49 / 12.21 / 0.12 / 0 / 0.2 / 28.99
20 / 6.64 / 1.67 / 1.59 / 2.98 / 0.08 / 0.04 / 0.08 / 13.08
21 / 9.86 / 1.35 / 0.76 / 13.04 / 0 / 0 / 0.08 / 25.09
22 / 2.58 / 0.28 / 0.72 / 4.49 / 0 / 0 / 0 / 8.07
23 / 17.5 / 2.15 / 2.15 / 28.39 / 0.08 / 0 / 0.04 / 50.3
24 / 10.66 / 2.07 / 3.82 / 7.67 / 0.08 / 0 / 0.28 / 24.57
25 / 16.62 / 6.76 / 17.5 / 4.49 / 0.08 / 0 / 0.24 / 45.69
26 / 0 / 0.04 / 0.04 / 10.42 / 0 / 0 / 0 / 10.5
27 / 9.15 / 1.55 / 2.74 / 15.35 / 0.56 / 0.32 / 0.12 / 29.78
28 / 16.06 / 2.86 / 1.47 / 12.29 / 0.36 / 0 / 0 / 33.04
29 / 10.74 / 2.58 / 1.19 / 0.99 / 0 / 0 / 0.04 / 15.55
30 / 18.37 / 4.29 / 2.11 / 1.99 / 0.32 / 0.04 / 0.12 / 27.24
31 / 6.16 / 1.67 / 2.35 / 2.58 / 0.4 / 0.2 / 0 / 13.36
32 / 10.3 / 3.1 / 5.05 / 2.27 / 0.64 / 0 / 0.04 / 21.39
33 / 23.34 / 6.16 / 9.3 / 1.03 / 0.08 / 0.04 / 0.16 / 40.12
34 / 14.04 / 3.86 / 3.66 / 1.63 / 0.24 / 0 / 0.12 / 23.54
35 / 6.12 / 1.87 / 2.51 / 1.39 / 0.04 / 0.2 / 0.08 / 12.21
Percent of Total Area / 36.54% / 7.68% / 10.37% / 44.08% / 0.72% / 0.21% / 0.40%
2.4 Soils
Soil data for the Los Angeles River watershed were obtained from the State Soil Geographic Data Base (STATSGO). There are four main Hydrologic Soil Groups (Groups A, B, C and D). These groups, which are described below, range from soils with low runoff potential to soils with high runoff potential (USDA, 1986).
Group A Soils have low runoff potential and high infiltration rates even when wet. They consist chiefly of sand and gravel and are well drained to excessively-drained.
Group B Soils have moderate infiltration rates when wet and consist chiefly of soils that are moderately-deep to deep, moderately- to well-drained, and moderately course textures.
Group C Soils have low infiltration rates when wet and consist chiefly of soils having a layer that impedes downward movement of water with moderately-fine to fine texture.
Group D Soils have high runoff potential, very low infiltration rates and consist chiefly of clay soils. These soils also include urban areas.
The total area associated with each specific soil type was determined for all 35 subwatersheds. However, the dominant soil group ultimately represented each subwatershed in the model. Soil types within each subwatershed and the dominant soil group are presented in Table 3.
Table 3. Dominant Soil Group for each Subwatershed
Model Subwatershed / Dominant Soil Group / Model Subwatershed / Dominant Soil Group / Model Subwatershed / Dominant Soil Group1 / D / 13 / D / 25 / D
2 / D / 14 / C / 26 / C
3 / D / 15 / D / 27 / C
4 / D / 16 / D / 28 / D
5 / D / 17 / D / 29 / D
6 / D / 18 / D / 30 / D
7 / D / 19 / D / 31 / D
8 / D / 20 / D / 32 / D
9 / D / 21 / D / 33 / D
10 / D / 22 / D / 34 / D
11 / D / 23 / D / 35 / D
12 / D / 24 / D
2.5 Reach Characteristics
Each delineated subwatershed was represented with a single stream assumed to be completely mixed, one-dimensional segments with a trapezoidal cross-section. The National Hydrography Dataset (NHD) stream reach network for USGS hydrologic unit 18070105 was used to determine the representative stream reach for each subwatershed. Once the representative reach was identified, slopes were calculated based on DEM data and stream lengths measured from the original NHD stream coverage. In addition to stream slope and length, mean depths and channel widths are required to route flow and pollutants through the hydrologically connected subwatersheds. Mean stream depth and channel width were estimated from as-builts provided by the LADPW and were supplemented or verified through field reconnaissance. An estimated Manning’s roughness coefficient of 0.2 was also applied to each representative stream reach.