In agriculture, an essential question is "What can be grown in a specific site?" Farmers want to know what will grow well on their farm or in a particular field of their farm. Answering this question provides the basis for analyzing, then defining, the best possible business options in agriculture. For major commodity crops, information is readily available on the species and varieties likely to perform well for given conditions.

For the myriad of crops considered as of minor importance globally but, which are, nevertheless, significant to rural communities and local economies, reliable information on what can be successfully grown is not readily available. Local and expert knowledge tends to be restricted to those species or crops already grown in a given region or at least on the same continent, and access to literature is organized on a species-by-species basis.

Such geographical limitation on knowledge is of particular concern in light of the consistent dominance of a few, non-native species in commercial agriculture throughout the world. Such limitation is currently true of tropical fruits, and is likely to continue being so. Thus, the most promising new crops for local consumption or export are most likely to come from regions distant from the farmer or exporter.

The advent of computerized geographic information systems (GIS) opens the way to managing the large amounts of data required to match particular crops or species to the growing conditions of given sites. We have established a conceptual basis for developing two interlinked systems—CropIdent™ and Homologue™—which will enable users to determine what crops or varieties are likely to grow well in a given site. The procedure would be as follows:

• First, we input into the CropIdent™ system a combination of expert opinion and local knowledge on a particular species' requirements.
• This knowledge is then combined with the identification of a small number of specific sites where the crop or species is known to grow well.
• All this information is fed into Homologue™, which then identifies other similar areas ("homologs") where the crop is likely to grow successfully.

An important advantage of this approach is that no specific crop parameters related to individual factors, such as "critical temperature for pollination", are needed. The fact that a crop grows and performs well in a known and adequately described site is sufficient information.

The proposed system will be built on experience already gained at CIAT with the development of MarkSim™, which manages large climatic databases, and FloraMap™, which uses information on the known distribution of species to determine the probability of finding that species in another site. CIAT also has experience in linking expert knowledge with GIS in forage species in Central America. This experience will also be used to develop crop specific tables, which will form the basis for developing CropIdent™.

Once the system is developed, it will have multiple uses: farmers, farmer groups, technologists, planners, agronomists, entrepreneurs, and other agencies would be able to identify:

• Potential species or crops for the users' particular conditions
• The most promising sites or areas to visit to obtain specific technology
• Those sites where particular genotypes and crop management technologies are likely to function effectively
• Areas where current climates in given areas will be similar to future climates in other areas. Thus, users will be able to determine which existing crops and technologies will be suitable as site conditions respond to climatic change.