土壤质地显著影响陆面水文模拟精度。利用5套土壤质地数据和内蒙古陆面同化系统，模拟了5组土壤水分数据。使用2016—2020年内蒙古63个国家气象站观测逐日0~10 cm土壤水分资料，评估了模拟土壤水分的适用性。结果表明：（1） 联合国粮农组织（FAO）和中国科学院（CAS）土壤质地与北京师范大学（BNU）、全球土壤数据集（GSDE）、第二次土壤调查（SNSS）土壤质地在空间上存在较大差异，SNSS、BNU和GSDE土壤质地较好地表征了内蒙古四大沙地和三大沙漠较高的砂土含量和较低的黏土含量。（2） 5组模拟实验和中国气象局陆面同化系统（CLDAS）土壤水分均能较好地再现内蒙古土壤水分自东北向西南逐步递减的空间分布，但均出现高估现象，其中，SNSS土壤水分与观测值的契合度最好。（3） 5组模拟实验和CLDAS土壤水分在内蒙古及其三个气候分区具有极显著的时间相关系数，FAO和CLDAS土壤水分的误差较大，SNSS土壤水分的误差和克林-古普塔系数均最优。综上，SNSS土壤质地适用性最好，SNSS模拟土壤水分的精度更优，各套模拟土壤水分在干旱区的适用性差于其在其他区域的适用性。

Soil texture is a key parameter affecting the accuracy of land surface hydrological simulation. Investigating the applicability of different soil textures to soil moisture simulation is an effective approach to improve the precision of regional land surface hydrological simulation and gridded drought and flood monitoring. On the basis of five latest domestic and international soil texture datasets and the Inner Mongolia Land Data Assimilation System (IMLDAS), five soil moisture simulation experiments corresponding to the five soil texture datasets were designed, and five soil moisture simulation datasets were implemented. Using daily surface soil (0-10 cm) moisture observation data from 63 national meteorological stations in the Inner Mongolia Autonomous Region from May to September from 2016 to 2020, the applicability of the five simulated soil moisture datasets was systematically evaluated. The results were as follows: (1) Soil texture datasets of the Food and Agriculture Organization of the United Nations (FAO) and Chinese Academy of Sciences (CAS) differed considerably from those of Beijing Normal University (BNU), Global Soil Dataset for use in Earth System Models (GSDE), and the Second National Soil Survey on China (SNSS) in space. The soil texture datasets of the SNSS, BNU, and GSDE better characterized the higher sand content and lower clay content in the four sandy lands and three desert areas of the Inner Mongolia Autonomous Region. (2) Compared with the observed soil moisture data, the spatial distribution characteristics of soil moisture from northeast to southwest in the Inner Mongolia Autonomous Region was well reproduced by the five soil moisture simulation experiments and the China Meteorological Administration Land Data Assimilation System (CLDAS) simulations of soil moisture, but there was an overestimation among them. Moreover, the trend of observed soil moisture with time was reflected well by all simulations. The soil moisture simulated from the SNSS experiment performed the best in the Inner Mongolia Autonomous Region and its three climatic zones, as it was the most consistent with the observed spatial distribution. (3) The five simulation datasets and CLDAS simulations of soil moisture in the Inner Mongolia Autonomous Region and its three climatic sub-regions had extremely significant temporal correlation with the observed soil moisture. The mean absolute error (MAE) and root mean square error (RMSE) of soil moisture simulated by the FAO and CLDAS were relatively large. Moreover, the MAE, RMSE and Kling-Gupta efficiency coefficient values of soil moisture simulated from the SNSS experiment were the best, followed by those corresponding to the BNU, CAS, and GSDE experiments, which were significantly better than those corresponding to the FAO and CLDAS soil moisture simulations. In conclusion, the SNSS simulation of soil moisture in the Inner Mongolia Autonomous Region and its three climatic sub-regions showed higher accuracy, and the applicability of SNSS soil texture was the best. Notably, the applicability of all simulated soil moisture datasets in the arid area was worse than that in other areas.