Hydrological modelling and associated microwave emission of a semi-arid region in South-western Niger
Abstract
Studying surface–atmosphere feedback is often limited by the accuracy of the land surface observations (particularly soil moisture estimates) or the performance of the land surface models. To further our understanding of soil moisture effects on land–atmosphere fluxes, improvements in soil moisture mapping over large regions are necessary. The aim of this study was to obtain accurate soil moisture mapping over a 120 by 100 km2 area in West Africa using the considerable amount of measurements available from local and regional scales, recorded during the African monsoon multidisciplinary analysis (AMMA) experiment. The modelling strategy was based on the use of a land surface model (LSM), employed to provide high-resolution soil moisture mapping over the studied area. A microwave emission model was then used to simulate associated microwave brightness temperatures (TB) to compare with the Advanced microwave scanning radiometer (AMSR) at the same spatial (25 by 20 km2) and temporal resolution (daily). Discrepancies between observed and simulated TB were analysed and used to calibrate the LSM and the microwave emission models to match the specific hydrology and soil microwave behaviour of the studied area. Nevertheless, a positive bias of the near-surface soil moisture remained and the land surface model was still unable to reproduce the rapid dynamic of the near-surface soil moisture observed at the local and regional scales in this climatic context. To solve this problem, a secondary surface soil layer was added to match in situ soil moisture measurements as well as satellite microwave measurements. Additionally, the choice of the soil permittivity model was found to be of prior importance in order to perform suitable microwave brightness temperatures. Finally, a soil moisture retrieval algorithm based on AMSR and meteosat second generation (MSG) measurements was proposed in order to improve the quality of the soil moisture estimates over the studied area (the root mean square error decreases from 5.4 % vol. to 2.8 % vol).