ASSESSMENT OF SEA SURFACE TEMPERATURE USING LANDSAT-TM DATA

摘要

The Thane creek, forming the East Coast of Island city of Mumbai, has been used as a convenient dumping site for disposal of treated as well as untreated effluents from industries and municipal corporations. The study area (approximately 100 sq. kms) in the Thane creek, has an unusual co-existence of highly industrialised urban setting and a thick stand of mangrove forest at the land-water interface. Eight field trips between January 1997 to February 1998 were carried out in the study area for collection of surface samples synchronised with the Landsat satellite passes. Sea Surface Temperature (SST) values were also recorded simultaneously. Linear regression model and Linear regression model with sun angle correction, were developed by co-relating the field SST data with Digital Numbers (DNs) from thermal IR band (10.4 μm to 12.5 μm, Thematic Mapper band 6 data) which were extracted from the digital remotely sensed data for the corresponding dates and locations. In case of linear regression model, part of the data from each field trip was used for calibration of the model and the remaining data were utilised for validation of the model developed in the present study. A reasonable "goodness of fit" was obtained based on chi-square test while comparing field observed and model predicted SST values (α=0.77 to 0.99). In case of linear regression with sun angle correction model, three field trips were utilised for calibration of model after sun angle correction. Data from remaining five field trips was utilised for validation of model after correcting for sun angle. "Goodness of fit" was assessed to be excellent based on high α values (0.99 to 1) obtained in chi-square test. Linear regression model with sun angle correction was found superior to the linear regression model developed in this study.