中文摘要：

利用青藏高原东北部地区阿柔冻融观测站2013年5月至2014年11月观测资料,对通用陆面过程模式（CLM4.0）和动态陆面过程模式（DLM）青藏高原高寒地区土壤湿度模拟性能进行了评估。结果显示两种模式均能够较好的反映浅层（〈40 cm）土壤湿度动态变化,然而显著低估非冻结期土壤湿度;通过土壤有机质含量对土壤湿度模拟敏感性分析发现模式模拟土壤湿度偏干可能与模式中土壤有机质方案不足有关。在此基础上改进DLM模式土壤有机质和冻土液态水渗透方案,实验结果表明新参数化方案显著提高了高寒、高有机质含量地区模式土壤湿度模拟,平均偏差（BIAS）、均方根误差（RMSE）,均方差（MSE）和相关系数（R）分别达到0.032 m~3·m~（-3）,0.078 m~3·m~（-3）,0.010 m~3·m~（-3）和0.866。

英文摘要：

Land surface models（LSMs） provide mean for understanding soil moisture status and its responses and feedbacks to climate system. Firstly,we evaluated the Community Land Model version 4. 0（CLM4. 0） and Dynamic Land Model（DLM） modelled soil moisture at Arou station（ located in northeastern Qinghai-Tibetan Plateau） from May 2013 to November 2014. The results showed that both of the two models well reflected soil moisture dynamic changes of shallow layers（40 cm）,but significantly underestimated the soil moisture under unfroezn condition. Secondly,we analysed the sensitivity of soil moisture simulations to soil organic matter and found that the underestimates of soil moisture simulations were mostly resultd by the inappropriate soil organic matter and frozen soil impermeable fraction parameterizations. On the basis these results,we have improved the DLM model parameterization schemes of soil organic matter and hydraulic conductivity of soil within ice. The model test expriment showed that the new parameterizaion significantly improved soil moisture simulations in cold with high soil organic matter regions and the mean values of mean bias（ BIAS）,root mean square error（RMSE）,mean square error（MSE） and correlation coefficient（R） reached 0. 032 m~3·m~（-3）,0. 078 m~3·m~（-3）and 0. 01 m~3·m~（-3） and 0. 866,respectively.