Comprehensive assessment of MODIS-derived near-surface air temperature using wide elevation-spanned measurements in China

This is the first comprehensive evaluation of the Moderate Resolution Imaging Spectroradiometer (MODIS)-derived near-surface air temperature, which has been used widely in a series of large-scale models varying in various disciplines ranging from climatology, hydrology to ecology. Four retrieval methods: the highest available pressure in the atmospheric profile product, interpolation by the adiabatic lapse rate, interpolation by the hypsometric equation and the combination with land surface temperature, were developed in the past, but only with validation in regional scale. All of these are evaluated in this paper against 2168 hourly meteorological recordings with an elevation span of over 5000 m in China. Results show that the method of the highest available pressure exhibits a serious underestimation, especially in areas of high elevation, such as the Tibetan Plateau. Interpolation by the hypsometric equation can only fix the underestimation to a very small extent, while interpolation by the adiabatic lapse rate can achieve a relatively good performance. In addition to the elevation influence, substantially variable estimates occur with the parabola-like distribution in low elevation areas, which implies the influence of cloud in Southern China. The combination of the underestimation from interpolation by the adiabatic lapse rate and overestimation in land surface temperature can eliminate the disturbance of both elevation and cloud, resulting in the best performance with r = 0.94, bias = -0.83°C and root-mean-square-error = 4.18°C.