以怒江源区那曲流域为例，基于4个试验场完全融化期（6—9月）的土壤体积含水量（0.15~0.51 cm3/cm3）和土壤基质势数据（0~200 kPa）实测数据，选择Van Genuchten（VG）、Brooks-Corey（BC）和Campbell 3个模型进行拟合，以均方根误差>ERMS和决定系数R2 为评价指标，分析3个模型对高海拔季节冻土区不同土层和不同土壤质地的适用性。结果表明：从整体上看，VG模型（平均R2 为0.992，平均ERMS为0.006cm3/cm3）的拟合效果优于BC模型（平均-R2 为0.972，平均ERMS为0.019cm3/cm3）和Campbell模型（平均R2为0.984，平均ERMS为0.014cm3/cm3）；但是在不同土层和不同土壤质地情况下模型的适用性有所区别，VG模型更适用于壤土和壤质砂土（平均R2为0.987，平均ERMS为0.008 cm3/cm3）以及土壤深层（10~35 cm土层，平均R2为0.990，平均ERMS为0.007cm3/cm3），Campbell模型更适用于砂质壤土（平均R2为0.985，平均ERMS为0.009 cm3/cm3）以及土壤表层（5 cm土层，平均R2为0.993，平均ERMS为0.006cm3/cm3），BC模型在不同条件下都不是最优模型；参数θr 取值大小会显著影响土壤水分特征曲线的形状。本研究可为深入研究高海拔季节冻土区的土壤水分运动特性以及中华水塔区的水源涵养作用提供支持。

Determining the soil water characteristic curve in the high altitude seasonal frozen zone is crucial to understand its soil water movement characteristics. However, today's research on modeling soil water characteristic curves in the high altitude seasonal frozen zones focuses more on the effects of permafrost, vegetation degradation, and other factors on the change of soil moisture content, and lacks the analysis of the applicability of soil water characteristic curves.The volumetric soil water content and soil matrix potential were measured by in-situ observation experiment, selected three models, Van Genuchten (VG), Brooks-Corey (BC), and Campbell for fitting, and analyzed the applicability of the soil water characteristic curve model in the Naqu watershed of the Nujiang River source area using the root-mean-square error and the coefficient of determination as the evaluation indexes. The results showed that: (1) The ERMS of the VG model was 66.6% and 56.9% lower than that of the BC and Campbell models, respectively, and the R2 of the VG model was higher than that of the BC and Campbell models (on average, 2.0% and 0.9% higher, respectively), and the VG model was considered sufficiently accurate as far as the R2 and the ERMS were concerned ( R2 > 0.97,ERMS ≤ 0.011cm3/cm3). While the BC model had an overall poor fitting effect, with an average R2 of 0.97 and an ERMS of 0.018cm3/cm3, the Campbell model had strong applicability in the soil surface layer, with an average R2 of 0.99 and an ERMS of 0.006cm3/cm3; (2) The parameter θr of the VG model plays a key role in the simulation performance of the VG model and its value. The parameter α and the parameter n of the VG model are stable parameters, which have less influence on the shape of the soil water characteristic curve. The parameter θr of the BC model is also a sensitive parameter, which has a significant influence on the shape of the soil water characteristic curve, while the parameter Φe of the BC model controls where the segmentation point of the BC model starts. The parameter Φe and the parameter λ of the Campbell model together control the shape of the soil water characteristic curve simulated by the Campbell model; (3) The error analysis of the model shows that the VG model has a better simulation performance under different soil texture and depth conditions, and the error of the measured 4-150 kPa soil matrix suction segment is limited to ±2.5%. The Campbell model has a strong fitting ability in the sandy loam area and the soil surface layer, while the BC model, due to its shape setting, often appears to be lower than the actual value in the middle section of the soil matrix potential and higher than the actual value in the high soil matrix potential section during the fitting process. In summary, the VG model has high applicability in the Naqu watershed of the Nujiang River source area, and it is recommended to use the VG model for fitting the soil water characteristic curve in the high altitude seasonal frozen zone . However, the model still has more obvious errors in the middle section of the soil matrix potential, and further work is still needed to prove its applicability to other alpine regions. Currently, there are still few studies on soil moisture profiles in the Tibetan Plateau, with scarce references and difficulties in field experiments. Therefore, in-situ observation experiments were chosen for data collection to obtain a long series of volumetric soil moisture content and soil matrix potential data, but soil moisture profiles are affected by wet and dry paths, and there are some errors in the soil moisture profiles. Therefore, it is necessary to analyze the paths of the collected data in the subsequent research and carry out validation experiments in the laboratory to improve and innovate the model to increase the simulation accuracy of the model, so as to make it more applicable in the Tibetan Plateau. This study can provide a basis for a deeper understanding of the soil moisture movement characteristics in the high altitude seasonal frozen zone and the water conservation role of the Chinese water tower zone.