While the EPA banned polychlorinated biphenyls (PCBs) in 1977, they continue to persist in San Francisco Bay (SF Bay), often at dangerously high concentrations due to their long half-life. However, in spite of their associated health and environmental risks, PCB monitoring within SF Bay is extremely limited, due in large part to the high costs, both in terms of labor and capital that are associated with it. In this study, a cost effective alternative to in-situ PCB sampling is presented by demonstrating the feasibility of PCB detection via remote sensing. This was done by first establishing relationships between in-situ measurements of sum of 40 PCB concentrations and total suspended sediment concentration (SSC) collected from 1998-2006 at 37 stations distributed throughout SF Bay. A correlation was discovered for all stations at (R2=0.32), which improved markedly upon partitioning stations into north bay, (R2=0.64), central bay (R2=0.80) and south bay (R2=0.52) regions. SSC was then compared from three USGS monitoring stations with temporally consistent Landsat 8 imagery. The resulting correlation between Landsat 8 (Rrs 654) and SSC measured at USGS stations (R2=0.50) was validated using an Airborne Visible/ Infrared Imaging Spectrometer (AVIRIS) image. The end product is a two-step empirical algorithm that can derive PCB from Landsat 8 imagery within SF Bay. This algorithm can generate spatial PCB concentration maps for SF Bay, which can in turn be utilized to increase ability to forecast PCB concentration. The observation that correlation between AVIRIS (Rrs 657) and SSC was stronger than that of Landsat 8 suggests that the accuracy of this algorithm could be enhanced with improved atmospheric correction.

Presentation Slides: https://www.scribd.com/document/321273834/Remotely-sensing-pollution-detection-and-monitoring-of-PCBs-in-the-San-Francisco-Bay#fullscreen&from_embed