Computer assisted Koeppen climate classification: a case study for Brazil

摘要

Koeppen's climate classification, developed one century ago, is still used as a reference for large-scale climate mapping and useful for climatologic research (climate shift, climate model testing, and ecological research). This paper describes a software that was developed to allow non-assisted Koeppen climate classification based on a simple database (monthly mean temperatures and rainfall) as input. The software follows, as strictly as possible, the original concepts suggested by Koeppen, but does not include climates that cannot be defined by simple input parameters. Combined with GIS tools, the software was used to produce the first comprehensive climate map of Brazil using a large database. The map showed 21 Climate Equations in three climate zones (A = tropical moist; B = dry; C = moist with mild winter). This map was compared with an existing climate map of Brazil (FAO/SDRN). The two maps showed good agreement only in the less detailed information (climate zone and type). The complete Climate Equations were not comparable in relation to spatial distribution and class coincidence. Improvements were made for Koeppen's climate classification, avoiding computational or database errors, and it increased the details for climate classification of Brazil. Copyright (c) 2006 Royal Meteorological Society