基于多变量时间序列（CAR）模型的洞庭湖区域地下水资源量预测

为了预测洞庭湖区域地下水资源量的变化趋势,制定应对地下水资源量较低年份的有效措施,通过分析洞庭湖区域河川天然径流量、长江三口入水量、城陵矶出水量与地下水资源量的变化趋势和相关性,运用多变量时间序列CAR模型,建立了洞庭湖区域地下水资源量的预测模型,验证了模型的预测效果;并运用该模型对洞庭湖区域地下水资源量的变化进行了预测。研究结果表明,洞庭湖区域河川天然径流量、长江三口入水量和城陵矶出水量的变化是引起地下水资源量变化的主要因素;本研究所建的模型模拟值与实测值相对误差的最大值为-7.47%,平均值为3.63%,表明该模型的预测精度高、效果好;通过对未来10年洞庭湖区域的地下水资源量进行预测可知,其平均值下降了0.49亿m3,最低值下降了0.66亿m3,表明该区域地下水资源量仍将持续降低;据不同方案下的地下水资源量预测结果可知,通过增加长江三口入水量,可以有效提高洞庭湖区域的地下水资源量。

Groundwater resources prediction based on multivariate time series CAR model in Dongting Lake area

In order to predict the change trend of groundwater resources in Dongting Lake area,effective measure is formulated to respond to the lower value year of groundwater resources.By analyzing the change trend and correlation between the nature river flow,Yangtze three inlet inflow water resources and Chenglingji outflow water with groundwater resources,by using the multivariate time series CAR model,the prediction model of groundwater resources was established,the prediction effect of model was validated.The results show that（1）The change of nature river flow,Yangtze three inlet inflow water resources and Chenglingji outflow water resources in Dongting Lake area are the main factors,which caused the change of groundwater resources.（2）The maximum relative error between simulated value and measured value of the model built in this article is-7.47%,the average value of relative error is 3.63%,it shows the prediction accuracy and effect of the model is good.（3）By predicting the groundwater resources in Dongting Lake area in the next 10 years,we can see that the average value of groundwater resources lower 0.049 billion m3,the lowest value of groundwater resources lower 0.066 billion m3 than the previous decade,it shows the groundwater resources in Dongting Lake area will continue to cut down.（4）According to the predicted result of groundwater resources under different programs,we can know that the groundwater resources in Dongting Lake area can be improved by increasing the Yangtze three inlet inflow water resources.