Modeling Trace Element Concentrations in the San Francisco Bay Estuary from Remote Measurement of Suspended Solids

Investigator: 
Johanna Press
Advisor: 
Ron Smith
Start Date: 
June, 2014
Description: 

Suspended and dissolved trace elements are key determinants of water quality in estuarine and coastal waters. High concentrations of trace element pollutants in the San Francisco Bay estuary necessitate consistent and thorough monitoring to mitigate adverse effects on biological systems—such as toxicity to fish, waterfowl, and other marine organisms—and the contamination of water and food resources. Although existing monitoring programs collect annual in situ samples from fixed locations, models proposed by Benoit, Kudela, and Flegal (2010) enable calculation of the water column total concentration (WCT) and the water column dissolved concentration (WCD) of 14 trace elements in the San Francisco Bay from a more frequently sampled metric—suspended solids concentration (SSC). This study tests the application of these models with SSC calculated from remote sensing data, with the aim of validating a tool for continuous synoptic monitoring of trace elements in the San Francisco Bay.

Using HICO imagery, semi-analytical and empirical SSC algorithms were tested against a USGS dataset. A single-band method with statistically significant linear fit (p < 0.001) was chosen as the proxy for SSC values. The numerical models for WCT and the distribution ratio D were applied in MATLAB with terms to account for regional and seasonal effects, and results were used to calculate WCD. The modeled results were assessed against in situ data from the San Francisco Estuary Regional Monitoring Program. Quantile regression was used to evaluate model sensitivity to the distribution of regions, and outliers displaying regional aberrations were removed before robust regression was applied. Statistically significant and highly correlated results for WCT were found for 10 elements, with goodness of fit greater than or equal to that of the original models of seven elements. WCD was successfully modeled for six elements, with goodness of fit for each exceeding that of the original models. Concentrations of Arsenic, Iron, and Lead in the southern region of the Bay were found to exceed EPA water quality criteria for human health and aquatic life. The results of this study demonstrate the potential of monitoring programs using remote observation of trace element concentrations, and provide the foundation for future investigation of pollutant sources and pathways over time.