Modeling river flows and sediment dynamics for the Laguna de Santa Rosa watershed in Northern California

Abstract

The nonpoint source pollution model Soil and Water Assessment Tool (SWAT) was applied to understand management options that may improve water quality in the Laguna de Santa Rosa watershed in Sonoma County, California. Surface water quality in the Laguna watershed has been significantly impaired over recent years, as natural land cover has been urbanized or converted to agricultural uses. We first generated new maps of land cover and major land uses from satellite and airborne imagery for the watershed. The SWAT model output was checked against six streamflow gauges in the watershed. At the monthly time step, we found that the precalibrated model performed well at all gauges, with the coefficient of determination (r²) values ranging from 0.81 to 0.92. Calibration by modifications of groundwater extraction in the watershed resulted in notable increases to correlation values at all gauges, except at upstream locations on Santa Rosa Creek and Mark West Creek. Measured seasonal trends in sediment concentrations were tracked closely by the SWAT model predictions. Highest sediment loading rates were associated in the model results with pasture, rangeland, and vineyard cover areas. Model scenarios were tested for vegetation filter strips and improved ground cover conditions applied in subbasins, where soil erosion was shown to be elevated in previous simulations.