Premium
Spatial–temporal models for ambient hourly PM 10 in Vancouver
Author(s) -
Li Kathy H.,
Le Nhu D.,
Sun Li,
Zidek James V.
Publication year - 1999
Publication title -
environmetrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.68
H-Index - 58
eISSN - 1099-095X
pISSN - 1180-4009
DOI - 10.1002/(sici)1099-095x(199905/06)10:3<321::aid-env355>3.0.co;2-d
Subject(s) - interpolation (computer graphics) , scale (ratio) , spatial variability , spatial correlation , multivariate interpolation , spatial ecology , environmental science , temporal scales , sampling (signal processing) , field (mathematics) , computer science , meteorology , statistics , geography , cartography , mathematics , animation , ecology , computer graphics (images) , filter (signal processing) , computer vision , pure mathematics , bilinear interpolation , biology
In this article we develop and compare several space–time models for hourly ambient PM 10 collected in the Vancouver area. The models consist of a deterministic trend plus stochastic residuals. We find that the ambient PM 10 field over the area has an essentially constant temporal pattern across its monitoring sites; the spatial variation is relatively small. Our comparison leads us in the end to adopt a model with a common temporal correlation structure for all the sites. The residuals after eliminating serial correlation prove to have at most small‐scale spatial correlations, suggesting that dense monitoring networks would be necessary to make widespread spatial interpolation meaningful. Copyright © 1999 John Wiley & Sons, Ltd.