指数平滑法在林地水分动态模拟和预测中的应用  

Application of Exponential Smoothing in Prediction and Dynamic Simulation of Soil Moisture in Forests

作　　者：崔哲伟[1,2] 毕华兴[1,2] 云雷[1,2] 田晓玲[1,2] 靳刚雷[1,2] 

机构地区：[1]北京林业大学水土保持学院水土保持与荒漠化防治教育部重点实验室,北京100083 [2]山西吉县森林生态系统国家野外科学观测研究站,山西吉县042200

出　　处：《水土保持通报》2011年第2期128-132,共5页Bulletin of Soil and Water Conservation

基　　金：国家自然科学基金项目"晋西黄土区森林植被对径流的影响及其尺度辨析"(30972419)

摘　　要：指数平滑法是广泛应用的基于时间序列进行预测的方法之一,但是目前该方法在土壤水分动态模拟与预测方面的应用较少。采用指数平滑法中的Simple模型,对晋西黄土区人工刺槐林地2009年6月1日至9月30日间每日土壤含水率和土壤蓄水量进行了模拟和预测。结果表明,该模型能够以一定精度对土壤含水率和土壤蓄水量进行模拟和预测;在0—150cm土层内的7个不同观测土层中,随着土层深度的增加,该模型的预测准确性增大。100—150cm土层预测值最大误差<5%,并且120—150cm土层的预测在7个观测土层中最为准确,其平均误差率仅为0.101 1%。对预测值与实测值进行比较后得出,降雨、蒸发蒸散、地表径流、植物根系吸收等影响因素可能对人工刺槐林地的土壤含水率和土壤蓄水量的影响深度可至120cm。Exponential smoothing based on time series is one of the methods widely used in predicting,but is rarely applied in the dynamic simulation and prediction of soil moisture.In this paper,the simple exponential smoothing was used to simulate and predict daily soil moisture and soil water storage capacity of an artificial locust forest in the loess region of Western Shanxi Province,between June 1st and September 30th,2009.The results show that soil water moisture and soil water storage capacity were simulated and predicted with a reasonable accuracy level.Observations of seven different layers（0—150 cm） indicated that the prediction accuracy of soil moisture increased with soil depth.The maximum error of the predictions within the layer of 100—150 cm was less than five percent,and the most accurate prediction（with an average error of 0.1 %） were found in the layer of 120—150 cm.Compared with the predicted and measured values,it was found that rainfall,evaporation and evapotranspiration,surface runoff,plant uptake and other factors may affect the artificial locust forest soil water moisture and soil water storage capacity up to the depth of 120 cm.

关 键 词：指数平滑 土壤水分 预测 人工刺槐林 

分 类 号：S152.7[农业科学—土壤学]