中文摘要：

利用塔里木盆地南缘和田绿洲4眼观测井的2001年-2004年逐月实测平均地下水位埋深资料，基于帕森斯季节性指数理论，分别建立了季节性地下水位埋深动态变化预测模型，并进行了趋势预测检验．结果表明：各观测井地下水位帕森斯季节性指数模型拟合差≤0．10m的月数占总检验月数的比例为66．7％，拟合差≤0．20m且〉0．10m的月数占总检验月数的比例为29．2％，拟合差〉0．25的月数占总检验月数的比例为4％；各观测井预测误差≤0．10m的月数占总预报月数的比例为41．7％，预测误差≤0．20m且〉0．10m的月数占总预报月数的比例为37．5％，预测误差〉0．20m的平均为12％，说明帕森斯季节性指数模型可以对短期和田绿洲地下水位动态趋势进行预测。

英文摘要：

Using 48 month data on groundwater level in the period of 2001--2004 from 4 wells, the groundwater level prognostic equations for different wells were established to predict the short-term groundwater level in the Hetian oasis. The prediction shows that,in the period of 2001- 2004,the number of month with fitting error of the groundwater level prognostic equation less than 0. 05 m accounts for 38. 33 of total tested number of month,and that less than 0. 10 m accounts for 66. 7% of the total,but that exceeding 0. 1 m accounts only for 33.3%. The model of predicting groundwater level was tested by 12-month data from the 4 wells in 2003. The number of month with predicted error of groundwater level less than 0. 10 m accounts for 41.7% of the total predicted number of month,and that exceeding 0. 10 m and less than 0. 20 m accounts for 37. 5% of the total,but that exceeding 0. 20 m accounts for 12%. The model of predicting groundwater level with the Person＇s season＇s index theory is valid to predict groundwater level in Hetian oasis.