关键词: GARCE, GLDAS, MODIS数据, 蒸散发

Abstract: MODIS MOD16 evapotranspiration (ET) product is widely used in various applications in geosciences, and the product has been evaluated by many researchers in different regions over the world. However, when evaluating the ET product on catchment scales, previous studies were conducted on the annual scale under the assumption of no terrestrial water storage (TWS) change. But the annual TWS change is not negligible, and ignoring the TWS change may result in big biases when evaluating MOD16/ET product. In addition, the use of ET product on monthly scale is often required in many applications. In the present study, the monthly MODIS MOD16/ET product of Chinese river basins during 2002-2013 are compared with actual evapotranspiration (WBET) which is calculated based on the principle of water balance from ground observed precipitation data, discharge data, GRACE satellite observed terrestrial water storage (TWS) data and Global Land Data Assimilation System (GLDAS) simulated TWS data. The results show that: 1) on the annual scale, MOD16/ET is overestimated except for Songhua River Basin, and the difference between MOD16/ET and WBET increases from the Songhua River Basin in the north to the Yangtze River Basin in the south; 2) on the monthly scale, MOD16/ET tends to be overestimated compared to WBET in low value range (<20 mm/month) and underestimated in high value range (>20 mm/month), and the overestimation is prevalent across China in the period from November to March, but the consistency between MOD16/ET and WBET differs among the catchments in other months; 3) the consistency between monthly MOD16/ET and WBET is better in exorheic basins than in endorheic basin, better in the Songhua River Basin in northeastern China than in the Yangtze River Basin in the southern China.

Key words: evapotranspiration, GARCE, GLDAS, MODIS