Applicability evaluation of multiple sets of soil moisture data on the tibetan plateau

Based on the in situ data of the soil moisture-observation networks established at Maqu, Naqu, Ali, and Shiquanhe (Sq) on the Tibetan Plateau (TP), and using five evaluation indices [Pearson correlation coefficient (R), root mean square error (RMSE), mean deviation (bias), standard deviation ratio (SDV), and unbiased RMSE (ubRMSE)], the applicability of soil moisture datasets [COMBINED, ACTIVE, PASSIVE, ERA5, ERA5-Land (LAND), ERA-Interim (INTERIM), CLSM, and NOAH] was comprehensively evaluated. The results showed that, at the observation-network scale, ACTIVE exhibited the best applicability in Maqu (R = 0.704, ubRMSE = 0.040 m3/m3), COMBINED performed best in Naqu (R = 0.803, bias = 0.016 m3/m3), LAND displayed the best consistency with observations in Ali (R = 0.734, bias = -0.035 m3/m3), and ERA5 not only showed the best performance in Sq (R = 0.793, bias = -0.037 m3/m3) but also exhibited good results in the other three observation networks (R > 0.6). In a smaller-scale evaluation in Maqu, ACTIVE performed best, followed by ERA5. The COMBINED and PASSIVE products had serious gaps in Ali and Sq, and had the worst applicability in the western TP. In conclusion, considering the correlation results and temporal and spatial continuities, ERA5 is the most suitable soil moisture dataset for the TP.