基于时间序列的PM_(2.5)时空插值算法改进研究

Research on the Improvement of Spatio-temporal Interpolation Algorithm for PM_(2.5) Based on Time Series

该文以成都市及其周边市区分布的环境监测站点作为研究对象,将空气污染物指数作为数据源,进行时间序列平稳性处理,改进了时空插值算法的经典算法——约减法,从而提高了PM_(2.5)数据插值精度。同时还分析了PM_(2.5)的自相关特点,对时间序列数据进行分段处理,在自区间内使用函数模型法实现对PM_(2.5)数据的实时预测。实验表明,对时空插值算法改进后,在进行既往缺失数据修补及实时PM_(2.5)数据预测时,结果精度及可靠性高。

This paper takes environmental monitoring sites distributed in Chengdu and its surrounding urban areas as the research object,uses air pollutant index data as the data source to process the stability of time series,improves the classic algorithm of spatio-temporal interpolation algorithm——reduction algorithm,and improves PM_(2.5)data interpolation accuracy.The autocorrelation characteristics of PM_(2.5)are analyzed,time series data is processed in sec‐tions,and the function model method is used to realize real-time prediction of PM_(2.5)data in the self-interval.Ex‐periments show that after improving the spatiotemporal interpolation algorithm,the accuracy and reliability of the results are high when repairing historical missing data and predicting the real-time PM_(2.5)data.