As part of a historic remediation project, approximately 61 km 2 of salt evaporation ponds in the southern portion of San Francisco Bay, CA (USA) are scheduled for restoration to natural tidal marsh habitat over the next several decades. We have investigated the correlation of remotely sensed infrared spectral information with in situ field measurements and sampling, and evaluated the usefulness of a remote sensing approach to monitor salinity and population distributions of microbial communities in the hypersaline ponds. The Airborne Visible and Infrared Imaging Spectrometer (AVIRIS) instrument operated by the Jet Propulsion Laboratory has created a ten-year archive of spectral information concerning these ponds. We utilized spectral signatures of microbial populations that are sensitive to salinity, and trained a supervised classification algorithm to identify physical parameters from an AVIRIS scene based upon microbe spectra gathered in the field using a portable visible to near-infrared (VNIR) spectrometer. Our results indicate that automated analyses of hyperspectral observations are capable of detecting variations in microbial populations and discriminating corresponding salinity levels.

No abstract

We have investigated the correlation of remotely-sensed infrared spectral information with in situ field measurements and sampling to monitor ecosystem health, salinity, and population distributions of specific microbial communities occupying the salt ponds of southern San Francisco Bay. Approximately 61 km 2 of these salt evaporation ponds are scheduled to be restored to natural habitat over the next few decades as part of a historic remediation project. Significant cost reduction may be achieved through the use of a remote infrared monitoring approach in place of extensive sampling expeditions by field teams. Unique spectral signatures of microbial populations sensitive to salinity and other chemical concentrations provide the key to this method. The Airborne Visible and InfraRed Imaging Spectrometer (AVIRIS) instrument operated by the Jet Propulsion Laboratory (JPL) has created a ten-year archive of spectral information concerning these ponds which can be used to establish a baseline for comparison. We report here on in situ field sampling of microbial populations, including spectral measurements, and our results using these spectra to train a supervised classification algorithm (the US Geological Survey Tetracorder Algorithm) to identify microbial populations and physical parameters from an AVIRIS scene. Future measurements and data from satellite-based sensors may prove vital in monitoring the restoration process. High spectral resolution measurements from AVIRIS will be used to determine the efficacy of similar approaches using existing multispectral spaceborne sensors as well as to provide a reference for future, AVIRIS-class spacecraft as a surrogate for expensive ground surveys. Figure 1. The salt ponds of San Francisco Bay, as seen by AVIRIS. This image was generated by combining three channels from the 224-channel image cube. This produces an approximation of a true-color image. The colors of the ponds arise from microbial populations. Yellow rectangles denote pixels selected for spectral averages. BAY A4n A6n A7 A5 A8 A13 A15 A16 A17 A18 M3 M7 M6 M5 M4 A21 A20 A23 A19 A12 A14 A11 A9 A10 A3w A2w A2e B2 B1 A1 M2 M1 M8 A22 S N W E Remote

No abstract