Comparison of three prediction strategies within PM2.5 and PM10 monitoring networks

Research output: Journal Publications and Reviews (RGC: 21, 22, 62)21_Publication in refereed journalpeer-review

11 Scopus Citations
View graph of relations



Original languageEnglish
Pages (from-to)590-597
Journal / PublicationAtmospheric Pollution Research
Issue number3
Online published12 Dec 2019
Publication statusPublished - Mar 2020


An air pollution monitoring network is a primary tool for measuring, managing and assessing air quality in urban areas. But plenty of equipment in monitoring network inevitably result in financial costs. With this consideration, this study aims to explore some strategies for prediction instead of measurement. To do so, the relationships within PM2.5 and PM10 monitoring networks are first explored respectively. Then the relationship between PM2.5 with PM10 monitoring networks is also investigated. Based on these identified relationships, three prediction strategies are proposed and PM2.5 concentration is selected for predictions. The results verified that PM2.5 concentration can be well estimated under these strategies, in particular the local strategy based on the local pollutants and mixed strategy based on the local and surrounding pollutants. That means, without measurement, the pollution levels of some pollutants can be successfully determined through prediction. These findings provide a possibility to estimate the missing or unmonitored values in term of the available data at surrounding stations.

Research Area(s)

  • Air quality monitoring network, Cluster analysis, Generalised regression neural network, PM10, PM2.5