Estimation Of Sensible Heat Flux Based On SEBS Model And Its Application For Drought Monitoring

In this study, first used the Surface Energy Balance System (SEBS), remote sensing data of VIRR/FY-3A (Visible and InfraRed Radiometer) and MODIS/TERRA (MODerate-resolution Imaging Spectroradiometer), meteorological station data at Arou site to estimate a point-based sensible heat flux from May to September2010. Based on the LAS measurements at Arou site, sensible heat flux derived from VIRR and MODIS over alpine grass landscape were validated at satellite overpass time respectively, and sensitivity analysis of sensible heat were carried out. Second, the paper explores the utility of MERSI/FY-3A (MEdium Resolution Spectral Imager) in drought monitoring through a case study in Hunan Province of China in May,2011. Two indices including DSI (Drought Severity Index) and TVDI (Temperature Vegetation Dryness Index) are extracted to provide spatial information on drought. Correlation analysis is carried out between the two indices and soil moisture data, accumulated precipitation data collected from meteorological stations. MODIS/TERRA data are processed in the same way to provide a benchmark for MERSI performance evaluation.The main contents and conclusions were as follows:(1) Sensible heat flux estimation accuracy of VIRR is similar to that of MODIS, indicating VIRR data quality is as good as MODIS under normal conditions. RMSE (Root Mean Square Error) of VTRR_SEBS_H and MODIS_SEBS_H is45.2143w/m2(N=64) and57.3693w/m2(N=71) respectively.(2) Many common features are shared by VIRR and MODIS, which can be attributed to SEBS model performance. In May-June, H are over-estimated with big fluctuations and larger RMSE, but in July-September, H are under-estimated with fewer fluctuations and smaller RMSE. SEBS yields obvious over-estimation of sensible heat flux in June, probably because of high wind speed, large delta_T, and complicated thermodynamic state during transitional period when bare land transform to dense vegetation cover. More in-depth research is needed to address this phenomenon.(3) In terms of monthly means, sensible heat flux retrieved from MODIS is a little higher than VIRR due to about30min later in satellite overpass. And the high quality of VTRR/LST products have been proved. Evaporative fraction (EFr), which is more time-independent, shows higher agreement between MODIS and VIRR. (4) Sensitivity analysis shows potential temperature gradient (delta_T) plays a dominant role in determining magnitude and fluctuation of H. Variation of H calculation is very large when delta_T is high, that’s why H estimation during May-June has so large fluctuations. When delta_T is small, it has a more dominant role in the magnitude of H, so H estimation during July-September becomes smoother.(5) The dry grade distribution of drought monitoring results based on DSI and TVDI are consistent, but DSI showes advantage in complicated topography areas such as Dongting lake plain and Nanling mountains. It improves that drought monitoring based on DSI index is feasible.(6) The drought monitoring results derived from MERSI and MODIS shows the similar space distribution pattern, and high correlation between two indices and soil moisture data, accumulated precipitation data collected from meteorological stations have been founded. Although correlation coefficients between MODIS-derived indices and soil moisture are a little higher than MERSI-derived indices, MERSI is able to show more rich spatial details due to its high spatial resolution. The case study demonstrates that MERSI data quality is adequate for drought monitoring and it is worthwhile to apply MERSI data to more wide applications.