ESTIMATING THE BEAM ATTENUATION COEFFICIENT IN COASTAL WATERS FROM AVHRR IMAGERY

This paper presents an algorithm to estimate particle beam attenuation at 660 nm (c(p)660) in coastal areas using the red and near-infrared channels of the NOAA AVHRR satellite sensor. In situ reflectance spect ra and c(p)660 measurements were collected at 23 stations in Case I an d II waters during an April 1993 cruise in the northern Gulf of Mexico . The reflectance spectra were weighted by the spectral response of th e AVHRR sensor and integrated over the channel 1 waveband to estimate the atmospherically corrected signal recorded by the satellite. An emp irical relationship between integrated reflectance and c(p)660 values was derived with a linear correlation coefficient of 0.88. Because the AVHRR sensor requires a strong channel 1 signal, the algorithm is app licable in highly turbid areas (c(p)660 > 1.5 m(-1)) where scattering from suspended sediment strongly controls the shape and magnitude of t he red (550-650 nm) reflectance spectrum. The algorithm was tested on a data set collected 2 years later in different coastal waters in the northern Gulf of Mexico and satellite estimates of c(p)660 averaged wi thin 37% of measured values. Application of the algorithm provides dai ly images of nearshore regions at I km resolution for evaluating proce sses affecting ocean color distribution patterns (tides, winds, curren ts, river discharge). Further validation and refinement of the algorit hm are in progress to permit quantitative application in other coastal areas. Published by Elsevier Science Ltd.