Open AccessA semi‐parametric regression approach to climatological quantile estimation for generating percentile‐based temperature extremes indices
Author(s) -
Yang Chi,
Xu Jing
Publication year - 2017
Publication title -
atmospheric science letters
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.951
H-Index - 45
ISSN - 1530-261X
DOI - 10.1002/asl.724
Subject(s) - percentile , quantile regression , quantile , parametric statistics , estimation , econometrics , regression , statistics , computer science , mathematics , economics , management
A semi‐parametric regression approach to quantile estimation for daily temperature data is proposed, in which both the biases and inhomogeneity are negligible, and is applied to the calculation of the six percentile‐based Expert Team on Climate Change Detection and Indices ( ETCCDI ) temperature extremes indices. Comparisons of the results with those from the CLIMDEX datasets show that the three warmth indices in the latter are probably biased such that their linear trends under the RCP4 .5 scenario seem to be overestimated. In order to avoid drawing misleading conclusions, it is necessary to re‐examine currently adopted algorithms and available datasets, and to develop new methods for generating the percentile‐based ETCCDI indices.