基于灰色马尔可夫链的兰村泉域地下水位预测

Prediction of Groundwater Depth the Based on Grey Markvo Model

将兰村泉域S1长观孔2001—2010年的年平均水位埋深作为特征因素序列,降水量和人工开采量作为相关因素序列,采用灰色系统理论将GM(1,N)模型应用于兰村泉域岩溶地下水位埋深预测,并应用马尔可夫模型对输出结果进行残差修正。结果表明:经过修正后的GM(1,3)模型的拟合精度达到97.41%,比没有经过残差修正的GM(1,3)模型高出9.62%,修正后的预测值更加贴近原始值,准确性提高。采用马尔可夫残差修正模型对2011—2013年兰村泉域水位埋深值进行预测,结果表明:2011年、2012年、2013年的地下水位预测值分别为33.24、32.01、31.12 m,地下水位缓慢回升。

Using the value of annual average underground water depth of S1 hole which in Lancun karst spring area,the values of rainfall and groundwater withdrawal as characteristic factor sequence and labor sequence respectively and combining with the grey system theory GM (1, N)model that forecasting underground water depth in Lancun karst spring area to simulate and modify the output residual error based on the Markov model. The results show that the fitting accuracy of the modified GM (1,3)model is 97. 41% that is 9. 62% higher than that of be-fore. And the predictive values which had been corrected were closer to the original date. The results of applying markov residual error cor-rection model for predicting the underground water depth (2011-2013)in Lancun karst spring area show that the predictive value of ground-water depth in the year of 2011,2012 and 2013 are 33. 24,32. 01 and 31. 12 m respectively,indicating that the groundwater table is rising slowly.