Introduction

In 1999, the FAO Environment and Natural Resources Service published some global climate grids, a map showing the Koeppen climate classification next to various derived products such as the agroclimatic production potential according to Lieth (details). Regardless of several inconsistencies which users soon discovered, the maps turned out to be very popular and were widely copied, pasted into other web pages as well as linked to.

We are presenting below some updated maps based on improved recent global climate grids (click here for the source of the original data) and a more consistent methodology ([pdf] with an overview of the definition of Koeppen classes). Results of a recent recalculation of climatic net primary production can be found here. The image below shows average Koeppen climate for the period 1951-2000 based on the CRU and VASClimO datasets. Click here to download a larger version of the map. A very nice map can also be downloaded here from the Department of Natural Sciences at the University of Veterinary Medicine, Vienna.

The new maps are presented for two time periods, to allow users to assess changes that have taken place recently. For instance, the map below shows in red all the areas where the climate has changed (3^rd level of Koeppen classification) between the two reference periods from 1951 to 1975 and 1976 to 2000.

A larger version as well as a georeferenced variant of the image in IDA/WinDisp format can be downloaded by clicking here.

Methodology

The new globally gridded Koeppen climatologies are based on climatological observations rasterized to grids of 0.5�x0.5� degrees. They cover 4 different periods and are based on 3 different precipitation datasets. This leads to a total number of 12 rasterized Koeppen climatologies.

Mean annual cycles of temperature are always taken from the Climatic Research Unit (CRU, Mitchell and Jones, 2005). Precipitation is taken from 3 different sources:

1. Precipitation data set of CRU based on about 10,000 station records (Mitchel and Jones, 2005),
2. GPCC Full data, based on all time series fragments available to GPCC (>40,000 stations, Rudolf et al., 2005),
3. GPCC VASClimO data, based on 9,343 stations that provide at least 90% coverage within 1951-2000 (Beck, Grieser and Rudolf, 2005).

1. 1951 � 2000 (all),
2. 1961 � 1990 (norm),
3. 1951 � 1975 (early),
4. 1976 � 2000 (late).

All different datasets are converted to the same 0.5�x0.5� grid, which results from USGS gtopo30. Antarctica and Greenland are left out due to lack of data.

For the users convenience data are provided in several formats: comma separated values, IDA/Windisp geo-referenced grids, and GeoTIFF, which can be viewed either as images (TIFF) or georeferenced grids. For both IDA/WinDisp and GeoTIFF, the Koeppen classes, that are identified as strings of letters (e.f. E, or Af, or Cfa) have to be recoded as numbers. The following coding system (�byte representation�) was adopted:

The coding is such that A climates cover the range from 13 to 43, Cf ranges from 112 to 118 etc. Oceans and large freshwater bodies are coded as 254, while land areas that were assigned no climate values (for instance very small islands) are indicated with a value of 255.

Downloads as comma separated values (csv):

For each data source, the file contains a list of comma separated values (csv) ready for the use with Microsoft Excel. There are of 63,660 lines, each corresponding to a grid point over land. The five columns (data items on a line) are the gridpoint number, the longitude (in 0.01�), the latitude (in 0.01�), the fraction (in %) of the area of the 0.5�x0.5� pixel covered by land, and the Koeppen class using the standard 3-letter codes. The csv files are available zipped or unzipped.

Downloads as IDA (Windisp):

IDA stands for Image Display and Analysis format. These files can be easily displayed and rescaled with Windisp. Windisp also allows further processing of the data including the easy manipulation of colour scales.

We provide the Koeppen maps in their original 0.5�x0.5� resolution. Each pixel that covers a fraction of land is drawn as a whole pixel. We furthermore resampled these maps on a 5�x5� grid. The resulting fine-scale images still contain the classes on a .5�x.5� grid but the coastlines are represented more precisely.

Windisp can display the Koeppen classes up to the first, second or third letter. This allows the maps to be displayed in full (3 letters) or medium (2 letters) accuracy or to concentrate on the 5 major classes (1 letter). The colour scales come in two flavors, with and without an explicit byte representation. The resulting 6 colour scales are available here.

A list of the byte representations is here.

Downloads in GeoTIFF format:

In order to display the GeoTIFF files in colour with ESRI ArcView GIS, ESRI legend files are provided for Koeppen classification up to the first, second and third letter, respectively. A customised legend can be created with the help of the byte code used within the GeoTIFF files. This byte code is presented here. The GeoTIFF images can also be viewed using the FAO AgroMetShell software (AMS) that can be downloaded from the FAO-agromet FTP site. Note that the GeoTIFF files can be viewed with the colour tables listed under �Downloads as IDA (Windisp)� above.

Additional information

For further information on the area size covered by different Koeppen climate classes click here.

Download this file as pdf.

For any questions or to report problems please contact [email protected]

References

Beck, C., J. Grieser and B. Rudolf, 2005: A New Monthly Precipitation Climatology for the Global Land Areas for the Period 1951 to 2000. Klimastatusbericht 2004, 181-190, DWD. [pdf]

Mitchell, T., and P. Jones, 2005: An improved method of constructing a database of monthly climate observations and associated high-resolution grids. Int. J. Climatol., 25, 693-712. http://www.cru.uea.ac.uk/

Rudolf, B., C. Beck, J. Grieser, U. Schneider, 2005: Global Precipitation Analysis Products of the GPCC. Internet publication at http://gpcc.dwd.de/ [pdf]