Abstract The potential utility of Landsat-5 Thematic Mapper (TM) thermal IR (Band 6) data for measuring and mapping surface water temperatures of the Great Lakes is analyzed. Two contiguous TM scenes were collected on 18 July 1984 over the Green Bay-North Central Lake Michigan and Milwaukee-South Central Lake Michigan regions, respectively. The TM thermal IR data were empirically calibrated using surface reference data. The regression model so developed, was then used to map surface temperature (°C) over the entire TM coverage area. Surface reference data points not used in the original calibration were used to test the predictive capability of the model over a wider geographic area. As part of the Landsat Image Data Quality Analysis effort (LIDQA), Bartolucci et al. (1985) have investigated the absolute calibration of TM thermal IR data. Application of these calibration equations to our data set verified their accuracy, resulting in a root mean square error of less than I(/dg)C.