区域性地面沉降量预测的灰色与时间序列方法

Regional land subsidence forecast with GM and ARMA

由于特殊的地形、地质条件,海河流域因地下水超采造成的地面沉降现象较为严重.选用海河流域沧州地区的地面沉降历史情况作为特殊示例,运用灰色模型 (GM)和时间序列模型 (ARMA)进行建模、预测,最后尝试将两种模型混合,建立新的预测模型,根据实测沉降数据进行建模,预测未来地面沉降.此外,对三种模型进行比较、说明,来阐述混合模型优越性,以期为整个海河流域的地面沉降研究提供适合的方法.最后通过实例检测证明新的模型综合了原两种模型的特点,优势互补,达到了较好的效果,从而证明了其实用性与可信度.

The regional land subsidence induced by the over-extraction of groundwater is more serious within Haihe River Basin due to the special topographic and geological conditions. As a special case, the historical data of the land subsidence in Cangzhou area within Haihe River are to be used to develop the models for the forecast of the subsidence therein with both GM model and ARMA model, and then a new forecast model for predicting the future subsidence concerned is developed with the combination of both the models based on the actual measured data.Moreover, comparison and interpretation are made among the models to explain the advantage of the combined model, from which a suitable method for forecasting the ground subsidence in the whole basin is expected as well. With the integration of the features from both the original models, the new model is proved to be satisfactory through the examination made on the actual case with better practicality and reliability at last.