Premium
REGEOTOP: New climatic data fields for East Asia based on localized relief information and geostatistical methods
Author(s) -
Thomas Axel,
Herzfeld Ute C.
Publication year - 2004
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.1058
Subject(s) - elevation (ballistics) , kriging , multivariate interpolation , climatology , interpolation (computer graphics) , digital elevation model , variogram , geostatistics , precipitation , environmental science , climate model , spatial variability , principal component analysis , evapotranspiration , climate change , meteorology , geography , statistics , geology , mathematics , remote sensing , computer science , frame (networking) , bilinear interpolation , telecommunications , ecology , oceanography , geometry , biology
Abstract Climate data fields represent essential tools for climate, biogeographical and agricultural research to run models and to provide observational data for the verification of global climate models (GCM). Climate data fields are generated through interpolation of observations taken at meteorological stations. Most current interpolation procedures try to describe the influence of topography on spatial climatic variations by relating them directly to absolute elevation or by introducing simple relief variables such as exposure. In both cases this may not properly describe spatial climatic variations, particularly not those of precipitation. This paper describes a regionalization procedure (REGEOTOP) that was applied to generate monthly spatial climatic data fields of temperature, precipitation and potential evapotranspiration (1951–90) for East Asia at 0.25° resolution from a new climate data base. REGEOTOP proceeds in several steps, combining statistical and geostatistical methods. First, basic relief types are determined by application of principal component analysis to moving windows in a digital elevation model (DEM) of the study area. Second, climatic variables are related to relief parameters by regression analysis with respect to basic relief types, location, and elevation. To account for large‐scale variation of climate variables, geostatistical variogram analysis (step three) and interpolation (step four) are applied to the regression residuals. Finally, maps of regression estimates plus kriged residuals are calculated, for a total of 1440 cases. The relief parameterization retains about 90% of the variance of a DEM in 10–18 principal components, depending on input parameters. The REGEOTOP method is computationally expensive, but results justify the effort. Owing to the thorough analysis in the REGEOTOP method and its application to the most comprehensive climate databases that exist outside China to date, the resultant maps provide a solid basis for GCM verification or hydrological and agro‐ecologic investigations and prognoses for East Asia. Copyright © 2004 Royal Meteorological Society