Estimation of daily maximum and minimum air temperature using MODIS land surface temperature products

摘要

Accurate estimation of spatially distributed air temperature (Ta) is useful for a wide range of disciplines such as hydrology, meteorology and ecology. Moderate Resolution Imaging Spectroradiometer (MODIS) land surface temperature (Ts) products hold the potential to get spatial estimates of land surface temperature at high temporal (daily) and spatial resolution (1. km) across the world. In this paper, the possibility of retrieving high-resolution Ta data from MODIS Ts data on both Terra and Aqua satellites was investigated over Xiangride River basin in the north Tibetan Plateau. The results of statistical analysis show that during the night MODIS Ts data on Terra satellite provides a good estimation of daily minimum air temperature with RMSE = 2.97 °C, MAE = 2.37 °C, and r = 0.94, while during the day the difference between Terra MODIS Ts data and daily maximum air temperature is large with RMSE = 7.45 °C, MAE = 6.21 °C, and r = 0.83. In order to improve the accuracy of the estimation of daily maximum air temperature, the temperature-vegetation index (TVX) method was used in this paper. By lowering the threshold for the negative correlation coefficient between NDVI (Normalized Difference Vegetation Index) and Ts to -0.1, the TVX method was applied to MODIS land products to estimate daily maximum air temperature successfully for the first time. The validation results show that the TVX method has improved the accuracy of daily maximum air temperature significantly with RMSE = 3.79 °C, MAE = 3.03 °C, and r = 0.83. Finally, the diurnal air temperature variation was simulated using a simple sine curve model for the period 21:00-9:00. UCT. The simulation results were in good agreement with observed values with RMSE = 3.43 °C, MAE = 2.54 °C, and r = 0.86. The results of this study show that daily maximum and minimum air temperature can be retrieved effectively from MODIS land surface products through the use of TVX method.