时间序列自回归模型在土壤水分预测中的应用研究

Application of Temporal Sequence Auto-Regression Model to Soil Moisture Forecast

在黄土丘陵半干旱区,多年生柠条林地土壤旱化严重,为了实现土壤水资源可持续利用,需要依据土壤水分状况,调控柠条生长与土壤水关系,而监测和预报土壤水分是防旱抗旱、调控柠条生长与土壤水关系的基础。为此选取黄土丘陵半干旱地区柠条林地2011年4月至2013年2月的土壤含水量时间序列作为研究对象,将各土层土壤含水量变异系数分为3个变化范围,并从3个范围内各选取一个代表土层,建立自回归模型,运用AIC准则及最大似然估计法求解模型中的参数,卡方检验的结果认为建立的模型较好。经过实例验证,实测值与预测值的相关误差均小于10%,说明时间序列自回归模型能够很好地预测黄土丘陵半干旱地区柠条林地的土壤含水量。

In the semi-arid loess hilly region,the arid trend of perennial Caragana korshinskii land is severe. In order to achieve the sustainable utilization of soil water resources,it needs to control the relationship of Caragana korshinskii growth and soil water according to the soil moisture condition. Monitoring and forecasting soil moisture are the fundamental to drought control and relationship control of Caragana korshinskii growth and soil water. The paper selected the soil moisture content sequence from April 2011 to February 2013 in the Caragana korshinskii land of semi-arid loess hilly region as an object of study,divided the variable coefficient of soil water content of each layer into three ranges and selected three representative layers from each range to establish auto-regression model. It solved the model parameters by using Akaike's Information Criterion and the maximum likelihood estimate method and the results of Chi-square test showed that the built model was good.The relative errors of measured value and predicted value were all smaller than 10% through case verification,showing the model could well predict the soil water content of Caragana korshinskii land in the semi-arid loess hilly region.