Comparison of MODIS land surface temperature and ground-based observed air temperature in complex topography

Abstract

Near-ground air temperature (T _a) and land surface temperature (T _s) are important parameters in studies related to variations in hydrology, biodiversity and climate change. However, complicated mountainous terrain tends to hinder observations in such areas. The scarce observations from mountainous areas can be augmented with data from a 1 km high spatial resolution data set. This data set is obtained from the land surface temperature element of the Moderate Resolution Imaging Spectroradiometer (MODIS) instruments installed on the Aqua and Terra Earth observation satellites from NASA. This study used regional monthly mean T _a data for Taiwan as a reference to assess the monthly mean T _s data set. The results showed that the two sets of data had correlation coefficients of 0.91–0.96, and the standard deviations of the differences between the two sets were 1.25–1.77°C. These results could serve as a reference for research related to climate and ecology. Further analysis indicated some possible sources of bias between T _s and T _a: (1) the significant influences caused by soil moisture between wet and dry seasons; (2) the difference between ground-based weather station elevation and 1 km grid-averaged elevation; and (3) interaction among the satellite view, solar zenith angle and terrain gradient. When the T _s product (V005) is used directly in ecological study and application, it is essential to have a clear knowledge of the bias and its possible causes.

Acknowledgements

We thank David Zelený for the comments on data analysis and two anonymous reviewers for very valuable comments on the previous versions of this article. This work was supported by the Experimental Forest, National Taiwan University, Taiwan (grants 98B01, 99B02 and 100B09). Ching-Feng Li was supported by the Czech Science Foundation (526/09/H025) and the Ministry of Education of Czech Republic (MSM 0021622416).