MODIS land surface temperature composite data and their relationships with climatic water budget factors in the central Great Plains
Authors: Park, S.; Feddema, J. J.; Egbert, S. L.
Source: International Journal of Remote Sensing, Volume 26, Number 6, 2005 , pp. 1127-1144(18)
Publisher: Taylor and Francis Ltd
Abstract:Daily land surface temperatures (Ts) derived from moderate resolution imaging spectroradiometer (MODIS) data were correlated with concurrent climatic water budget variables. Using a climatic water budget program, four daily water budget factors—percentage soil moisture (SM), actual/potential evapotranspiration ratio (AE/PE), moisture deficit (MD), and moisture deficit/potential evapotranspiration ratio (MD/PE)—were calculated at six weather stations across western and central Kansas. Correlation analysis showed that Ts deviations from air temperature had a significant relationship with the water budget factors. To do the analysis on a weekly basis, daily MODIS data were integrated into three different types of weekly composites, including maximum Ts, driest-day, and maximum Ts deviation (from maximum air temperature, or maxTa). Results showed that the maximum Ts deviation (Ts–maxTa) temperature composite had the highest correlation with the climatic water budget parameters. Time-integrated, or cumulative values and the moving average of the Ts deviation were meaningful measures of the relationship, but effective moving average periods varied spatially. Correction for different data acquisition times of MODIS thermal imagery improved the representativeness of signals for surface moisture conditions. The driest-day composite was most sensitive to time correction. After time correction, its relationship with soil moisture content improved by 11.1% on average, but the degree of correlation improvement varied spatially. Despite this improvement, the driest-day composite dataset did not have as strong a correlation with water budget factors as that of the maximum Ts deviation composite method.
Document Type: Research Article
Affiliations: Kansas Applied Remote Sensing Program, University of Kansas, Lawrence, Kansas 66047, USA
Publication date: March 1, 2005