Statistical Methods Development and Sampling Design Optimization to Support Trends Analysis for Loads of Polychlorinated Biphenyls from the Guadalupe River in San Jose, California, USA

The Regional Monitoring Program for Water Quality in San Francisco Bay (RMP), with guidance from its Small Tributaries Loading Strategy (STLS) Team, has been conducting small tributary loading studies at several sites in the Bay Area since 2003. A current priority for the STLS is developing a strategy for measuring trends in pollutant loadings from small tributaries to the Bay. 

This technical report presents the work to develop a statistical model for trends in loads of polychlorinated biphenyls (PCBs), and estimate the power for proposed monitoring designs as a basis for detecting trends for the Guadalupe River watershed (San Jose, California). The statistical approach builds upon the turbidity surrogate methodology that has been employed in the STLS since 2003. A novel, two-stage statistical modeling approach was used to incorporate the significant turbidity-PCB relationships that exist, and evaluate climatic, seasonal, and inter-annual factors as additional potential drivers of PCB loads. The longest-running time series of tributary monitoring by the STLS on the Guadalupe River was selected as the case study for developing and testing the statistical approach. As a result of this effort, a multiple linear regression model for detecting trends in PCB loads in the Guadalupe River was developed.