Agroforestry Systems 10: 3345, 1990. 33 © 1990 Kluwer Academic Publishers. Printed in the Netherlands. A climatic analysis method for expert systems assisting tree species introductions TREVOR H. BOOTH CSIRO Division of Forestry and Forest Products, PO Box 4008, Canberra, Australia ACT 2600 Key words: Species selection, species trials, agroforestry, similarity, Eucalyptus citriodora Abstract. There is a need for improved methods to suggest whether a particular tree species is worth testing at a new location. A method is proposed which compares the climatic similarity of a potential new location with conditions within a species natural distribution, and identifies if similar sites exist. When information is available, climatic comparisons can also be made with successful sites of introduction outside the natural range. A program is described which carries out these comparisons and Eucalyptus citriodora is analysed as an example. It is concluded that the method offers advantages over systems which describe climatic require- ments simply as ranges of suitable conditions. The importance of soil factors is recognised and ways in which these could be analysed along with climatic factors are noted. The integration of the similarity analysis into existing databases or its development as part of a complete expert system are discussed. Introduction Species of Australian origin, principally eucalypts, account for about half the area of tropical forest plantations [12]. Many lesser-known Australian species have potential for fuelwood and agroforestry uses [19]. For the last twenty-five years, the Australian Tree Seed Centre, part of CSIRO's Division of Forestry and Forest Products, has supplied seed to researchers in over 100 countries. In the 1986-87 period over 12 000 seedlots from a total of 571 species were distributed for trials [16]. Most of these samples are being used to determine which species or provenance (i.e. a species from an identified site) is best suited to a particular location. There is clearly a need for a system that can interpret the vast amount of information arising from these trials and recommend whether a species or provenance would be worth testing at other sites. Expert systems are computer programs designed to assist decision- making by simulating the problem-solving ability of human experts in specialized fields [18]. One of the first expert systems developed for a forestry application was designed to assist with the species selection problem. The