ESTIMATING WATER QUALITY OF THE ARIAKE SEA IN JAPAN USING LANDSAT-TM DATA - Evaluation of SDD and SST-

Abstract

Using Lowtran 7's estimated Rayleigh scattered and aerosol scattered radiance, the radiance reflected at the sea surface (Lw(λ)) is derived from the measured radiance of Landsat-TM images taken over teh Ariake Sea. Then the Lw(λ) averaged from 4 x 4 windows of pixels contered at 33 sampling sites was regressed againts the observed Secchi disk depth (SDD) using linear regression algorithms. Result show that use of multi-date visible channels of Landsat-TM as the calibration data predicts more accurate and dependable SDD at the validation stage than use of single-date calibration data of Landsat-TM. This study confirms the feasibility of retrieving SDD (or turbidity/ suspended-sediments) from Landsat-TM data. Limited experiments in the modeling sea surface temperature (SST) show the potential of predicting SST from the thermal channel (6) og Landsate-TM data using  a simple liniear regression.