基于大气降水的华北地区土壤湿度预测模型  

The Soil Moisture Predictive Model Based on the Precipitation in North China

作　　者：王珊珊[1] 韩丽娟 崔恒建[1] 杨华[3] 

机构地区：[1]北京师范大学数学科学学院,北京100875 [2]国家气象中心,北京100081 [3]北京师范大学遥感科学国家重点实验室,环境遥感与数字城市北京市重点实验室,北京师范大学地理学与遥感科学学院,北京100875

出　　处：《应用气象学报》2011年第4期445-452,共8页Journal of Applied Meteorological Science

基　　金：公益性行业(气象)科研专项(GYHY201106027);十一五科技支撑项目(2006BAD04B04);国家自然科学基金项目(40901170)

摘　　要：基于引入随机变量的机理性模型方法,利用华北地区2000—2008年气象台站观测数据,以大气降水为随机变量,并考虑其延迟效应,利用回归方法建立了预测时效为1旬的土壤相对湿度预测模型。利用预测率和干旱等级预报精度两个评价指标,结合2009年土壤湿度实际观测数据,验证了预测模型预报率均在90%以上,绝大部分站点的干旱等级预报精度均在70%以上,得出该预测模型在华北地区应用的合理性,从而建立了一套客观、动态的土壤湿度预测方法,有利于及时掌握农田旱情程度和分布,主动采取防旱、抗旱应对措施。A better soil moisture predictive model for North China can enhance the accuracy of drought forecast, which will have great significance in predicting the extent and distribution of drought, playing an important role in taking positive and active measures for drought in time. It is well acknowledged that one of the key points and difficult points in research on dry farming is the change of soil moisture, which is influenced by various factors, such as rainfall, evaporation, the time of sunshine, the kind and texture of soil and so on. Of all the factors, the amount of rainfall is one of the most important factors, which contributes a lot to the change of the soil moisture, deserves more and more attentions to study their relationships. A random variable＇s mechanism-rational model is therefore introduced, and the meteorological data in North China during 2000 to 2008 is analyzed, taking the atmosphere precipitation as random variable as well as its delay effect, employing the regression method to set up the soil relative moisture predictive model, which is effective for ten days. 10-cm and 20-cm soil relative moisture predictive models for the 55 stations in the north of China are established, but only the fitting results of 10 stations are given, and their fitness accuracy （the Pearson correlation coefficient） is all above 60%. Meanwhile, the efficiency of this model is verified with soil moisture data of 2009 by means of the two indexes, indicating that all of the forecasting rates are above 90% and most of the drought prediction rates are above 70%, therefore a series of objective, dynamic soil moisture predictive method is established. The fitting graphs and the graphs of 95% confidence interval are also given for the 10 stations, from which the feasibility of this method can be verified. The method provides a new support for the prediction of the soil moisture in the field. The method can extend to every station in the country to set up local unique models. Though this ap proach is only tentative experi

关 键 词：土壤湿度 土壤蓄水饱和度 降水延迟效应 非线性回归方程 

分 类 号：S152.7[农业科学—土壤学]