Estimation of daily minimum land surface air temperature using MODIS data in southern Iran

Land surface air temperature (LSAT) is a key variable in agricultural, climatological, hydrological, and environmental studies. Many of their processes are affected by LSAT at about 5 cm from the ground surface (LSAT(5cm)). Most of the previous studies tried to find statistical models to estimate LSAT at 2 m height (LSAT(2m)) which is considered as a standardized height, and there is not enough study for LSAT(5cm) estimation models. Accurate measurements of LSAT(5cm) are generally acquired from meteorological stations, which are sparse in remote areas. Nonetheless, remote sensing data by providing rather extensive spatial coverage can complement the spatiotemporal shortcomings of meteorological stations. The main objective of this study was to find a statistical model from the previous day to accurately estimate spatial daily minimum LSAT(5cm), which is very important in agricultural frost, in Fars province in southern Iran. Land surface temperature (LST) data were obtained using the Moderate Resolution Imaging Spectroradiometer (MODIS) onboard Aqua and Terra satellites at daytime and nighttime periods with normalized difference vegetation index (NDVI) data. These data along with geometric temperature and elevation information were used in a stepwise linear model to estimate minimum LSAT(5cm) during 2003-2011. The results revealed that utilization of MODIS Aqua nighttime data of previous day provides the most applicable and accurate model. According to the validation results, the accuracy of the proposed model was suitable during 2012 (root mean square difference (RMSD) = 3.07 A degrees C, = 87 %). The model underestimated (overestimated) high (low) minimum LSAT(5cm). The accuracy of estimation in the winter time was found to be lower than the other seasons (RMSD = 3.55 A degrees C), and in summer and winter, the errors were larger than in the remaining seasons.