Modeling Sediment Deposition for Predicting Marsh Habitat Development

The South Bay Salt Pond Restoration Project (SBSPRP) is the largest tidal wetland restoration project on the west coast of the United States. The purpose of this pro ject was to use in-situ and remote sensing measurements to create a GIS model capable of predicting sediment depositi on in restored ponds in the Alviso Salt Pond Comple x. A sediment transport model, suspended sediment concentration maps, as well as laboratory analyses of in-situ sediment data were used to predict sediment deposit ion. Suspended sediment concentrations from our in-situ samples as well as the USGS’s continuous monitoring sites were correlated with Landsat TM 5, ASTER, and MODIS reflectance values using three statistical te chniques—an Artificial Neural Network (ANN), a linear regression, and a multivariate regression to map su spended sediment concentrations in the South Bay. Multivariate and ANN regressions using ASTER proved to be the most accurate correlation method, yielding R 2 values of 0.88 and 0.87 respectively. Sediment grain size data wer e collected from Pond A21 to determine particle set tling velocities, grain size distribution, bulk densities , and rates of deposition. These data coupled with tidal frequencies and suspended sediment maps were used in the Marsh Sedimentation (MARSED) model for predicting deposition rates for three years. Data from MODIS were used to track sediment transport pathways in the South Bay for further assessing future marsh development. Results from th is project were applied to the Regional Ocean Model ing System (ROMS) sediment transport module for understanding sediment dynamics in the South Bay. MARSED results for Pond A21 show an RMSD of 66.8mm (< 1 σ) between modeled and field observations and can th erefore be successfully used to model future wetland restorati on efforts.