Hydrodynamic model of the San Francisco Bay and Delta, California

A two-dimensional hydrodynamic model of the San Francisco Bay and Delta was constructed using the Delft3D Flexible Mesh (DFM) modeling suite (www.deltares.nl/en/software/delft3d-flexible-mesh-suite/) to simulate water levels. Required model input files are provided to run the model for the time period from October 1, 2018, to April 30, 2019. This data release describes the construction and validation of the model application and provides input files suitable to run the model on Delft3D Flexible Mesh software version 2020.04. Model Description The San Francisco Bay and Delta Hydrodynamic Model (SFBDHM) utilizes the open-source Delft3D-Flexible Mesh software (www.deltares.nl/en/software/delft3d-flexible-mesh-suite/, 2020.04 release, SVN revision 601351) to compute Dynamic Still Water Levels (DSWLs) in San Francisco Bay and the Sacramento-San Joaquin Delta. DSWL captures the effects of meteorological and fluvial forcing on the coastal water levels; however, it excludes the impacts of wave setup and runup on the water level. The model covers the Delta up to the approximate upstream limit of tidal influence and extends seaward to the Pacific Ocean. It must be noted that the main purpose of the model was to simulate DSWL in open embayments of the San Francisco Bay. The model utilizes 1D elements used to represent tributaries and rivers flowing into the Bay and Delta. Model schematizations of the Delta (model grid and cross-section profiles) were derived from Delta Simulation Model II (DSM2, California Department of Water Resources, 2013). Topographic and bathymetric datasets from the USGS and California Department of Water Resources were applied across the San Francisco Bay and Delta Hydrodynamic model. In particular, the 2-meter resolution LEAN-corrected topography in the Bay (Buffington and others, 2016) and the seamless 10-meter resolution digital elevation model by Fregoso and others (2017) were applied. Data from the National Land Cover Database Land Cover (CONUS; Homer and others, 2020) were converted to roughness values using Manning's coefficients.The unstructured grid consists of more than 185,000 net nodes in the horizontal with a spatial resolution as fine as 100 meters. The 100-meter resolution model network is not fine enough to resolve smaller features such as narrow levees and dams. Therefore, an additional polyline has been included to account for constraining and rerouting effects of local levees and infrastructure. This file provides the location of each subgrid feature and, in combination with the latest topography, describes fine-scale elevations for the hydrodynamic simulations. The model is forced by astronomic tides and remote non-tidal residual (NTR) water levels at the offshore boundaries, fluvial discharges, and wind and atmospheric mean sea level pressure fields at the surface. The figure below depicts the model extent, location of offshore and discharge boundaries, levees, and NOAA tide stations used for model validation within the model domain. Figure 1: Map of model extent in the San Francisco Bay and Delta Offshore Boundaries The model's offshore boundary conditions in the Pacific Ocean are based on 67 measured tidal constituents at San Francisco with spatial variability derived from TPXO 8.0 (Egbert and Erofeeva, 2002). Tidal constituents were calibrated based on the difference between modeled and observed tidal constituents at the NOAA tide stations located throughout the bay. Remote NTR derived from measurements at the San Francisco NOAA tide station (#9414290) are applied uniformly across the ocean boundary. The tidal forcing files are included in the model package, as well as the NTR offshore boundary forcing files for the time period from Oct-2018 to Apr-2019. Discharge Boundaries Fluvial discharges from 16 USGS gauged rivers that flow into the Bay are included in the model (https://waterdata.usgs.gov/nwis/dv/?referred_module=sw). Six fluvial inflows to the Delta are based on Dayflow model outputs (https://data.ca.gov/dataset/dayflow). The discharge forcing files for the time period from Oct-2018 to Apr-2019 are included in the model package. The discharge stations incorporated in the SFBDHM are listed in table 1. Table 1. List of model discharge boundaries and their corresponding USGS discharge station identification numbers Name USGS Station Number Coyote Creek 11172175 Guadalupe River 11169025 Saratoga Creek 11169500 San Francisquito Creek 11164500 San Mateo Creek 11162753 Corte Madera Creek 11460000 San Rafael Creek 11459800 Novato Creek 11459500 Petaluma River 11459150 Sonoma Creek 11458500 Napa River 11458000 Wildcat Creek 11181400 San Lorenzo Creek #1 11181000 San Lorenzo Creek #2 11181008 Alameda Creek #1 11180500 Alameda Creek #2 11179000 Wind and Atmospheric Pressure Meteorological forcing conditions (wind and pressure) provided in this example dataset are based on ERA5 (Hersbach and others, 2020). ERA5 provides hourly estimates of a large number of atmospheric, land, and oceanic climate variables at 30-kilometer resolution. The ERA5 wind and pressure data for the time period from Oct-2018 to Apr-2019 is provided in the model package. Model Validation Measured and simulated water levels for the time period of Oct-2018 to Apr-2019 are provided in SFBD_his.zip. Measured water levels (referenced to the vertical datum NAVD88) were obtained from the NOAA tide stations (https://tidesandcurrents.noaa.gov/) within San Francisco Bay. Table 2 contains the measurement stations and model error statistics during the simulation period of Oct-2018 to Apr-2019 (NOAA Station ID, Station Name, Root Mean Square Error in cm, Mean Absolute Error in cm, Bias in cm). Root-mean-square errors (RMSE) are less than 10 cm for water levels at all observation sites (tide gauges). Model results are provided in the SFBD_his.nc file. Table 2. Statistics describing comparisons of observed and simulated water levels from the San Francisco Bay and Delta DFM model for the time period of Oct-2018 to Apr-2019 NOAA Station ID Station Name Root Mean Square Error (cm) Mean Absolute Error (cm) Bias (cm) 9414290 San Francisco 5.2 4.0 0.7 9414750 Alameda 6.6 5.0 2.4 9414863 Richmond 5.4 4.1 0.0 9414523 Redwood City 7.4 5.3 -2.0 9415144 Port Chicago 7.0 5.2 -3.0 9415102 Martinez 6.5 5.0 1.2