Investigating the relationship among extreme climate indices by a varying-coefficient regression model

ChunHong Wang,Jiangshe Zhang,Xiaodong Yan
DOI: https://doi.org/10.1109/ICCASM.2010.5620821
2010-01-01
Abstract:Because the changing frequency of extreme climate events generally has profound impacts on our living environment, research of the related topic has received much attention in social sciences. Based on the daily data collected from 753 meteorological stations in China during 1961-2005, the present work studies the relationship among some extreme climate indices, that is, FEP(frequency of extreme precipitation), FWD(frequency of warm days), FWN(frequency of warm nights), FCD(frequency of cold days) and FCN(frequency of cold nights). Since a varying-coefficient regression (VCR) model is a useful extension of a linear regression model by allowing the coefficients to vary with a covariate such as time, it was used to examine the extent to which FWD, FWN, FCD and FCN affect FEP over the period 1961-2005 by the coefficient functions estimated by local linear fitting method. The results show that FWD and FWN exert a much stronger influence on FEP than FCD and FCN. Meanwhile, the varying trends of FWD and FWN on FEP are almost identical except for the fact that the relationship between FWD and FEP is negative while that between FWN and FEP is positive. © 2010 IEEE.
What problem does this paper attempt to address?