Approaches to assessing the daily average ground surface soil heat flux on a regional scale over the Qinghai-Tibet Plateau

Abstract

The daily average ground surface soil heat flux (G0_daily) has been widely applied in land surface models. The accuracy of modeling G0_daily greatly impacts the estimation of the evapotranspiration and heat and moisture exchanges between the atmosphere and ground surface in frozen ground regions. We propose two methods for estimating G0_daily over the Qinghai-Tibet Plateau (QTP): the combination method (CM) and the sine method (SM). The instantaneous surface soil heat flux (G0_ins) is estimated at multiple times per day using the CM, and then, G0_daily is obtained. The SM simulates G0_daily by estimating G0_ins during the satellite transit based on the diurnal variation in the sinusoidal curves of G0_ins under clear-sky conditions. For G0_ins, the root mean square error (RMSE) and coefficient of determination (R2) of the CM are more than 10 W/m2 lower and 0.12 higher than those of the surface energy balance system (SEBS), respectively. For G0_daily, at the four sites (TGLMS, Naqu, Biru, and Nierong), the RMSE and the correlation coefficient (R) values of the CM are better than those of the SM by –0.86 W/m2, 9.04 W/m2, 4.33 W/m2, and 2.55 W/m2 and 0.19, 0.12, 0.10, and 0.08, respectively. The CM is applicable under all weather conditions and multi-period and multi-parameter observations throughout a day. Due to the principle of the SM, it can only simulate G0_daily when the satellite observes at least once each in the daytime and nighttime under clear-sky conditions. The SM overcomes the shortcomings of the CM, i.e., the requirements of complex input parameters and a large data volume. The advantage of the high spatial resolution of polar-orbiting satellite remotely sensed data (such as Moderate Resolution Imaging Spectroradiometer (MODIS) data) is given full play by the SM.