Climate classification: a historical perspective

E. Linacre


The Greek origin of the word 'climate' relates to the Sun's declination. The seasonal variation of this declination depends only on latitude. Egyptian and Helenistic writers were aware of the profound effect of latitude on climate. The modern distinction between climatic regions is rooted primarily in the efforts of several German scientists during the last 2 centuries. In 1817 Alexander von Humboldt drew annual-mean temperatures on a world map. Wladimir Koeppen (1846-1940) refined this map and plotted seasonal temperature range in 1884, leading to his climate classification (Section 16.2 in the book). This classification followed that of plants (by Carolus Linnaeus in 1735) and that of clouds (by Luke Howard in 1802), being likewise hierarchical, with major categories subdivided, and then subcategories divided again, and so on. In fact Koeppen had initially studied botany at St Petersburg, later completing a Ph.D. at Heidelberg on the effect of temperature on plant growth. At the highest level his system is based on five sets of temperature limits. These were developed from his categorisation in 1884 of thermal zones suited to various kinds of vegetation. He refined his system several times, notably in 1918 and in 1936 (at the age of 90).

In 1924 Koeppen accompanied son-in-law, the meteorologist Alfred Wegener (1880-1930), in moving from Germany to Graz in Austria. Wegener had been ostracised by Germany's universities because of his theory of continental drift. In Graz, Koeppen deduced geological climates in support of the theory. Also, he became associated with Rudolf Geiger (1894-1981) and collaborated with him in producing the 1936 system of climate classification. Geiger was later responsible for further revisions. Geiger established the discipline of microclimatology as he collected a wealth of observations to understand 'the climate near the ground' (to quote the title of his book, translated into English in 1960), and its variations due to topography and land use.

Tor Bergeron (1891-1971), Swedish but educated in Germany, developed a different way of classifying climates in 1928, in terms of causes. His 'genetic' system categorises a place according to the frequencies with which it experiences various kinds of airmass. Each airmass is labeled in terms of the latitude at which its temperature had been determined previously, and the kind of surface there, either marine or continental. This procedure was applied by Helmut Landsberg (1906-1985) in central Europe and Pennsylvania in the late 1930's, but it is rarely used nowadays. The number of classes is too small and the system requires considerable information.

In Australia, Griffith Taylor (1880-1963) incurred political odium in the 1920's by vehemently insisting on the validity of the large fraction of the continent dubbed 'desert' (BW) or 'semi-desert' (BS) by Koeppen (Fig. 16.19), with consequent limits to the nation's growth. His geography textbook was banned by ambitious politicians in Western Australia. So he left the country in 1928 for Toronto and then Chicago.

In the USA, Warren Thornthwaite (1892-1963) developed a hierarchical classification in 1931, essentially in terms of the annual pattern of soil-moisture conditions. These were regarded as depending in a complicated manner on the monthly input as rain, and implicitly on the output as evaporation, indicated by temperature. Studies in New Zealand, for example, showed that Thornthwaite's classification made more sense than Koeppen's, except at low latitudes. Later, the connection to soil moisture was made more explicit in Thornthwaite's landmark 'approach toward a rational classification of climate' (to quote the title of his paper in 1948). However, it involved an empirical formula for estimating evaporation, which was superseded the same year by the physics-based formula of Howard Penman (1909-1984) in England.

All these classification systems differ from what has been done since, in solving practical problems. Now we use complex statistical procedures to group the climates of places and to define areas of 'similar' climates.