基于ARIMA-MC模型的滑坡位移预测

中国是一个滑坡灾害极为频繁的国家，三峡库区更是滑坡灾害的多发区和重灾区，GPS地表位移监测是滑坡稳定性监测的重要手段。以三峡库区树坪滑坡为例，先利用自回归移动平均模型（ARIMA）对树坪滑坡GPS监测点的累积位移数据进行时序拟合，之后利用马尔可夫链模型（MC）对拟合结果进行拟合优化，最后建立自回归移动平均-马尔可夫链模型，并将其用于对树坪滑坡位移的时序预测之中。拟合和预测结果表明，该模型能够有效提高滑坡位移预测精度并实现短期内的动态滚动预测。

Prediction of landslide displacement based on ARIMA-MC model

Landslides occur frequently in China, especially in the Three Gorges Reservoir, which is the landslide-prone area. Surface displacement monitoring with GPS is an important means for monitoring the landslide stability. Taking the Shuping landslide in Three Gorges Reservoir area as an example, firstly, ARIMA model is used to fit the accumulated displacement time series of Shuping landslide and then MC model is used to optimize the fitting. At last, ARIMA-MC model is built and used to forecast the displacement time series of Shuping landslide. Fitting and forecasting results show that the model can effectively improve the prediction accuracy and achieve rolling forecasts of landslide displacement in short term dynamically.