Estimating phytoplankton biomass in coastal waters of Alaska using airborne remote sensing

Abstract

Empirical airborne remote-sensing relationships were examined to estimate chlorophyll a concentration in the first optical depth (chl FOD) of coastal waters of Afgonak/Kodiak Islands during July–August 2002. Band-ratio and spectral-curvature models were tested using satellite remote-sensing reflectance ( R rs( λ)) measurements. Additional shipboard and airborne R rs( λ) data were also analysed to evaluate consistency of proposed chl FOD– R rs( λ) relationships. Validation of chlorophyll algorithms was performed using data collected in the northern-part of the Gulf of Alaska and Bering Sea during 1996, 2002, and 2003 cruises. Likewise, oceanographic conditions during the surveys were typified to interpret variability of chl FOD fields. The SeaWiFS band-ratio algorithm OC2d was the most sensitive R rs combination ( R rs(509)/ R rs(553)) to detect chl FOD variability. Conversely, OC2a ( R rs(412)/ R rs(553)) had the lowest performance to derive chl FOD values. No valid statistical regressions were established for spectral–curvature relationships in the blue spectrum (< 500 nm). Fertile waters (> 5 mg m − 3 ) were preferentially located over shallow banks (∼50 m) and at the entrance of the bays. The approach used in this study to derive chl FOD values could be universal for Alaskan coastal waters. However, chl FOD– R rs( λ) relationships must be calibrated locally for a given season.