Comparison of BTS and Landsat TM data from Jotunheimen, southern Norway

Abstract

Results from snow temperature measurements at the bottom of the winter snow cover (BTS) are compared with Landsat TM data in a preliminary analysis of the potential of using Landsat data in permafrost mapping in the area of Jotunheimen, southern Norway. Uncorrected apparent satellite temperatures (AST) from the thermal band were used to identify areas of high amplitude ground surface temperatures on the day of the satellite pass (15 August 1997) and the normalized difference vegetation index (NDVI) was applied to detect vegetation density. The BTS data are highly correlated with altitude. The analysis shows that AST makes a small, but significant, contribution towards explaining the variance of BTS at similar altitudes in a linear regression model. Use of the thermal data in the model seemed to remove the aspect dependency in the BTS data. Analysis of the residuals in a linear model including altitude and AST suggests that any improvement of the model must be based on parameters that vary systematically on a scale less than 200 m. NDVI is difficult to evaluate because of high correlation with altitude.