Premium
Spatial dependence in the tails of air pollutant distributions: alternatives to the spatial correlogram
Author(s) -
Arbia G.,
Copetti M.,
Lafratta G.
Publication year - 2009
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/env.934
Subject(s) - correlogram , bivariate analysis , univariate , spatial analysis , environmental science , extreme value theory , statistics , pollutant , econometrics , spatial dependence , air pollutants , spatial distribution , joint probability distribution , mathematics , meteorology , multivariate statistics , geography , air pollution , ecology , biology
The study of the pollutants needs a better understanding of their extreme behaviours which could potentially cause adverse health effects. When analysing spatial dependence of the pollutant, the dependogram proposed by Arbia and Lafratta is preferred to the traditional correlogram used in the spatial statistics literature because it captures nonlinear relationships in the tails of the joint distributions and helps in detecting a pattern of spatial regularities. In this paper, we present a new method to estimate the spatial dependogram that uses univariate and bivariate threshold models. The method is applied to a set of hourly NO 2 data collected by seven monitoring stations in the city of Rome (Italy) during the years 2000 and 2001. Copyright © 2008 John Wiley & Sons, Ltd.