为探究WEPS模型的风蚀预测效果,在内蒙古乌兰布和沙漠油莎豆（Cyperus esculentus）种植区,对全采收（CK）、留4采6、留6采6油莎豆纯作和油莎豆-梭梭间作留茬4种不同采收模式地表进行风蚀监测,并利用2020年11月15日和12月26日2次实测风蚀数据对WEPS模型风蚀预测结果进行验证。结果表明：（1） 与全采收相比,3种留茬模式均能有效降低风力对种植区地表的侵蚀作用,地表防风固沙能力随留茬数量的增加而提高。（2） WEPS模型预测单宽风蚀量与实测结果有明显差异,模型值最大为实测值的10.16倍,最小为实测值的0.58倍;模型预测质量有不确定性,在植被覆盖度较高的地表预测效果较差。（3） 模型预测的单宽输沙量随实测值的增加呈幂函数增加,能够较准确地预测不同地表特征风蚀量变化趋势;在风蚀定量化估算上,WEPS模型还需根据实际风蚀环境对公式及参数进行修正。

The purpose of this study was to explore the wind erosion evaluation accuracy of the Wind Erosion Prediction System (WEPS). Wind erosion monitoring was carried out on the surface of four different harvest modes in the Ulan Buh Desert Cyperus esculentus planting area in Inner Mongolia, China as follows: full harvest, retain four harvest six, retain six harvest six, and retain intercrop (Cyperus esculentus and Haloxylon ammodendron). Wind erosion monitoring was carried out on the model surface, and the wind erosion prediction results of the WEPS model were verified using the wind erosion data, which was measured twice. The research results showed that compared with full harvesting, the three stubble modes effectively reduced the effect of wind erosion on the surface of the planting area. The intercropping stubble had the strongest wind-proof and sand-fixing ability among them, and the six remaining six harvested was the weakest. The surface wind-proof and sand-fixing ability was the weakest. With the increased number of remaining stubble, the structural characteristics of wind-sand flow gradually changed from a tortuous shape to the shape of “1”. This can provide a basis for the promotion of planting Cyperus esculentus in northern sandy areas to balance ecological and economic benefits. With the increased height, the sediment load of full harvest, with four remaining six harvested and six remaining six harvested, showed an exponential decreasing trend. The intercropping stubble sediment load showed a logarithmic decrease with the increase of height. The fitting degree decreased gradually, and the distribution law of sediment transport with height changed to a logarithmic direction. There was a certain difference between the WEPS prediction and the measured results. The maximum model value was 10.16 times the measured value, and the minimum was 0.58 times the measured value. The model prediction quality was uncertain, and the prediction effect was poor on the surface with high vegetation coverage. However, there was reasonable consistency as the unit-width sediment load predicted by the model increased as a power function with the increase of the measured value. Overall, the model is feasible in small-scale areas to predict the variation trend of wind erosion with different surface characteristics. However, in the quantitative prediction of wind erosion, the prediction value is less stable. It is necessary to strengthen the research on the model in multi-climate, multi-type surface vegetation coverage, multi-sequence prediction, and other factors related to wind erosion to establish a database that adjusts the model parameters and formulas according to the actual wind erosion environment to improve the universality of WEPS in the future.