Premium
Structural and sampling uncertainty in observed UK daily precipitation extremes derived from an intercomparison of gridded data sets
Author(s) -
Simpson Ian R.,
McCarthy Mark P.
Publication year - 2019
Publication title -
international journal of climatology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.58
H-Index - 166
eISSN - 1097-0088
pISSN - 0899-8418
DOI - 10.1002/joc.5789
Subject(s) - interpolation (computer graphics) , sampling (signal processing) , data set , precipitation , environmental science , set (abstract data type) , grid , climatology , meteorology , computer science , data mining , geography , geology , filter (signal processing) , geodesy , animation , computer graphics (images) , computer vision , programming language , artificial intelligence
This paper compares multiple gridded data sets of daily UK precipitation to evaluate structural uncertainty in our reconstructions of historical rainfall. The data sets compared reflect two different sampling strategies and three different grid interpolation methods. In order to separate the influence of sampling and interpolation uncertainties, one of the data sets (produced by the Met Office) has been recreated using the sampling strategy of stations used in the European (E‐OBS) data set. The results confirm and build upon previous studies showing that relying on a relatively sparse but homogeneous network of stations limits the ability of the resulting data set to reliably estimate extreme rainfall at the daily timescale. It is shown that gridding methods that additionally make use of reference climatological data can avoid systematic bias in both the average and extreme events even when using a relatively sparse network of observations. This is an encouraging result in terms of our potential to reliably extend such data sets further back in time where the availability of digitized data is substantially lower than the more modern era.