Evaluation of simulated soil moisture in GLDAS using in-situ measurements over the Tibetan Plateau

Soil moisture is a key state variable of land surface which plays a significant role in many hydrological processes. The Global Land Data Assimilation System (GLDAS) can produce global, high-resolution and continuous soil moisture data sets which have been used in many applications. However, before using the soil moisture products, it is crucial to validate their accuracy and reliability. In this study, we evaluate the simulated soil moisture from four land surface models (LSM) in GLDAS against in-situ soil moisture measurements collected from the Maqu network located on the Tibetan Plateau at different soil depths. The results show that all the four LSMs are able to capture the temporal variation of observed soil moisture well in the Maqu network region. However, four LSMs all display biases when compared with the in-situ measurements. The biases are mainly caused by the high soil organic carbon contents in the Tibetan Plateau, and may also come from uncertainties in model structure, model parameters, forcing data, and measurement errors, indicating that great efforts are still needed to further improve the simulation skill of LSMs on the Tibetan Plateau.