Köppen climate classification

As a diagnostic tool to quantify climate variation and change

World map of Köppen climate classification

The purpose of this website is to share information about the Köppen climate classification, and provide data and high-resolution figures from the paper Chen and Chen, 2013: Using the Köppen classification to quantify climate variation and change: An example for 1901–2010 (PDF).

The Köppen climate classification was developed based on the empirical relationship between climate and vegetation. This type of climate classification scheme provides an efficient way to describe climatic conditions defined by multiple variables and their seasonalities with a single metric. Compared with a single variable approach, the Köppen classification can add a new dimension to the description of climate variation. Further, it is generally accepted that the climatic combinations identified with the Köppen classification are ecologically relevant. The classification has therefore been widely used to map geographic distribution of long term mean climate and associated ecosystem conditions. Over the recent years, there has also been an increasing interest in using the classification to identify changes in climate and potential changes in vegetation over time. These successful applications point to the potential of using the Köppen classification as a diagnostic tool to monitor changes in the climatic condition over various time scales.

This work used a global temperature and precipitation observation dataset to reveal variations and changes of climate over the period 1901–2010, demonstrating the power of the Köppen classification in describing not only climate change, but also climate variability on various temporal scales. It is concluded that the most significant change over 1901–2010 is a distinct areal increase of the dry climate (B) accompanied by a significant areal decrease of the polar climate (E) since the 1980s. The areas of spatially stable climate regions for interannual and interdecadal variations are also identified, which have practical and theoretical implications.

The plots below show world maps of the Köppen climate classification and time series of the area change for different types (click on the thumbnails to enlarge). Individual images can be saved by right clicking on the thumbnails and choosing "Save Link As", or you can download all images in one zip file.

Long-time average (1901–2010)

These maps show the Köppen classification for the long-term average climate (1901–2010). The data source is a global observational dataset by Kenji Matsuura and Cort J. Willmott, which combines data from several sources (including GHCN2) interpolated onto a 0.5 ° longitude × 0.5 ° latitude grid.

Stable and unstable regions

In the maps below, the Köppen classification was applied on temperature and precipitation averaged over shorter time scales, from interannual to decadal and 30 year. The 30 year averages were calculated with an overlap of 20 years between each sub-period, while the interannual and decadal averages did not have overlapping years. Black regions indicate areas where the major Köppen type has changed at least once during 1901–2010 for a given time scale. Thus, the black regions are likely to be sensitive to climate variations, while the colored regions identify spatially stable regions.

Change in areas

A time series for the global area of each Köppen type was obtained using the 30 year classifications. For each time series, the area anomalies were normalized by the mean area for the whole period to yield the relative area change.

Major types


The data and figures presented here can be downloaded below for scientific use. We would be happy to hear from you if you are interested in using the data. Just send an email to hans.chen@psu.edu and tell us about your work.

Description File size Downloads
All data and figures 9.8 MB Download
Data only 3.0 MB Download
Figures only 6.8 MB Download

Citation: Chen, D. and H. W. Chen, 2013: Using the Köppen classification to quantify climate variation and change: An example for 1901–2010. Environmental Development, 6, 69-79, 10.1016/j.envdev.2013.03.007.

How to read

The data are saved in tab delimited ASCII text files with CRLF line endings. The first line contains the header, followed by 85794 lines with centered grid box coordinates and the Köppen type. Each grid box has a size of 0.5° longitude × 0.5° latitude. For the interannual, decadal, and 30 year time scale there are multiple columns for the type, one column for each average period as indicated by the header.

Here is an example of how the data can be read in Python:

# Example of reading the koppen_1901-2010.tsv file in Python
import numpy as np
koppen = np.genfromtxt("koppen_1901-2010.tsv", dtype=None, names=True)
print("The Koppen type at {} latitude and {} longitude is {}".format(
        koppen['latitude'][2], koppen['longitude'][2],

And an example for MATLAB:

% Example of reading the koppen_1901-2010.tsv file in MATLAB
koppen = tdfread('koppen_1901-2010.tsv');
disp(['The Koppen type at ' num2str(koppen.latitude(3)) ' latitude and ' num2str(koppen.longitude(3)) ' longitude is ' koppen.p1901_2010(3,:)])

This work used the same criteria for the Köppen classification as Kottek et al. (2006) and the tables below are larely based on the tables in their paper. The scheme is described in this section.

The Köppen climate classification uses monthly temperature and precipitation for the twelve months, usually averaged over a long period of time. Climate types are represented by a two or three letter combination in which the first letter defines the major type. The major types can be further divided into subtypes based on the precipitation pattern (second letter, except for the E type) and the temperature (third letter).

Subtypes satisfy the criterion of their parent type(s). There can only be one climate type in a region and it is determined using the following hierarchy: ET, EF, BSh, BSk, BWh, BWk, Af, Am, As, Aw, Csa, Csb, Csc, Cwa, Cwb, Cwc, Cfa, Cfb, Cfc, Dsa, Dsb, Dsc, Dsd, Dwa, Dwb, Dwc, Dwd, Dfa, Dfb, Dfc, Dfd. So, for example, if the climate satisfies the criterion of both ET and BW, it is classified as ET.

For simplicity, the criterion for each Köppen type is described using variables, which are defined in the next subsection. Example code for implementing the scheme in MATLAB can be found in classKoppen.m from the MeteoLab toolbox.

Variables and subscripts

1 The threshold is defined as follows: if Pw ≥ ⅔ Pann then Pth = (2 Tann), else if Ps ≥ ⅔ Pann then Pth = 2 Tann + 28 °C, else Pth = 2 Tann + 14 °C.
2 For example, 4 Tmon < +10 °C means "at least 4 months with mean temperature below 10 °C."
3 The summer months are April through September (AMJJAS) on the Northern Hemisphere and the winter months are October through March (ONDJFM), and vice versa for the Southern Hemisphere.
Variable Description
T Air temperature
P Precipitation
min Minimum monthly value for the whole year
max Maximum monthly value for the whole year
ann Annual mean value
th Threshold value 1
mon Number of months satisfying the criterion 2
smin Minimum monthly value for the summer months 3
wmin Minimum monthly value for the winter months 3
smax Maximum monthly value for the summer months 3
wmax Maximum monthly value for the winter months 3
s Mean value for the summer months 3
w Mean value for the winter months 3

First and second letter

Type Description Criterion
A Tropical climates Tmin ≥ +18 °C
Af Tropical rain forest Pmin ≥ 60 mm
Am Tropical monsoon Pann ≥ 25(100 - Pmin) mm
As Tropical savannah with dry summer Pmin < 60 mm in summer
Aw Tropical savannah with dry winter Pmin < 60 mm in winter
B Dry climates Pann < 10 Pth
BW Desert (arid) Pann ≤ 5 Pth
BS Steppe (semi-arid) Pann > 5 Pth
C Mild temperate -3 °C < Tmin < +18 °C
Cs Mild temperate with dry summer Psmin < Pwmin, Pwmax > 3 Psmin, Psmin < 40 mm
Cw Mild temperate with dry winter Psmax > 10 Pwmin, Pwmin < Psmin
Cf Mild temperate, fully humid Not Cs or Cw
D Snow Tmin ≤ -3 °C
Ds Snow with dry summer Psmin < Pwmin, Pwmax > 3 Psmin, Psmin < 40 mm
Dw Snow with dry winter Psmax > 10 Pwmin, Pwmin < Psmin
Df Snow, fully humid Not Ds or Dw
E Polar Tmax < +10 °C
ET Tundra Tmax ≥ 0 °C
EF Frost Tmax < 0 °C

Third letter

Type Description Criterion
h Hot arid Tann ≥ +18 °C
k Cold arid Tann < +18 °C
a Hot summer Tmax ≥ +22 °C
b Warm summer Tmax < +22 °C, 4 Tmon ≥ +10 °C
c Cool summer Tmax < +22 °C, 4 Tmon < +10 °C, Tmin > -38 °C
d Cold summer Tmax < +22 °C, 4 Tmon < +10 °C, Tmin ≤ -38 °C

Feel free to contact us if you have any comments or questions.

  • Hans Chen
  • Department of Meteorology and Atmospheric Science
  • The Pennsylvania State University
  • 624 Walker Building
  • University Park, PA 16802
  • Email: hans.chen@psu.edu
  • Website: http://hanschen.org