基于SARIMA时间序列模型的台风频次预测

Typhoon Frequency Prediction Based on SARIMA Time Series Model

台风是一种破坏力极大的灾害性天气,因此预测和预报台风,历来是气象工作的一项重要任务。对台风进行精准预测,并制定相应的预防和应急措施,是减轻台风造成灾难的重要手段。本文基于季节性差分自回归滑动平均模型(Seasonal Autoregressive Integrated Moving Average, SARIMA)研究台风频次的预测方法。该模型通过考虑时间序列的季节性和趋势性变化研究台风频次,旨在对未来台风频次提供准确预测方法。通过预处理台风发生频次相关数据,对时间序列进行平稳性检验以及白噪声测试,采用AIC遍历对模型定阶,计算模型的均方根误差(RMSE)和平均绝对误差(MAE)并绘制模型的残差分布和自相关图,分析比较后认为该模型的拟合效果较好。最后对2024年1月至2024年12月的台风频次进行预测,为提高自然灾害应对和相关政策制定提供了有力支持。

Typhoon is a highly destructive and catastrophic weather, so predicting and forecasting typhoons has always been an important task in meteorological work. Accurately predicting typhoons and de-veloping corresponding prevention and emergency measures is an important means of mitigating disasters caused by typhoons. This article is based on the Seasonal Autoregressive Integrated Mov-ing Average (SARIMA) model to study the prediction method of typhoon frequency. This model studies typhoon frequency by considering the seasonal and trend changes of time series, aiming to provide accurate prediction methods for future typhoon frequencies. By preprocessing the fre-quency related data of typhoons, conducting stationarity tests and white noise tests on the time se-ries, using AIC traversal to determine the order of the model, calculating the root mean square er-ror (RMSE) and mean absolute error (MAE) of the model, and drawing the residual distribution and autocorrelation diagram of the model, it is believed that the fitting effect of the model is good after analysis and comparison. Finally, the prediction of typhoon frequency from January 2024 to De-cember 2024 provides strong support for improving natural disaster response and related policy formulation.