灰色GM(1,1)模型在汶川县泥石流活动趋势预测中的应用

泥石流的发生是一个动态变化的过程,灾变时间及活动强度具有明显的非线性特征,认识和掌握泥石流发生、发展规律,对未来泥石流活动趋势进行预测,是地质灾害防灾减灾与宏观规划防灾预案工作的基础。本文以四川省汶川县1960~2010年50a的泥石流灾情数据作为原始数据序列,建立灰色GM(1,1)灾变预测模型,采用2011~2019年9a泥石流灾情数据来验证模型预测精度。为了提高预测精度,采用傅里叶变换修正误差残值。经计算GM(1,1)模型的预测结果相对误差为3.582%,傅里叶变换误差修正后相对误差为0.073%,预测结果与实际发生年份一致,并预测未来一段时间汶川县将于2033年或2034年、2051年或2052年发生大规模群发性泥石流。

APPLICATION OF GREY GM(1,1)MODEL IN DEBRIS FLOW ACTIVITY TREND PREDICTION IN WENCHUAN COUNTY

The occurrence of debris flow is a dynamic process.The time and intensity of disasters have obvious non-linear characteristics.Understanding and mastering the law of occurrence and development of debris flow and forecasting the trend of future debris flow activities are the basis of geological disaster prevention and mitigation and macro-planning of disaster prevention and preparedness.Taking the debris flow disaster data from 1960to 2010in Wenchuan County of Sichuan Province as the original data sequence,the grey GM(1,1)disaster prediction model is established.The prediction accuracy of the model is verified by the debris flow disaster data from 2011to 2019.In order to improve the prediction accuracy,Fourier transform is used to correct the residual error.The relative error of GM(1,1)model is 3.582%and that of Fourier transform error is 0.073%.The predicted result is consistent with the actual year.It is predicted that large-scale mass debris flows will occur in Wenchuan County in the future in 2033or 2034,2051or 2052.