Research on many problems in ecology, geography, and forestry depends on being able to accurately identify (to species) and map a large number of individual trees. Doing this on the ground is straightforward, but consumes so much time and expense as to be prohibitive for areas larger than a hundred or so acres -- smaller than the scale on which most forests vary. Consequently, researchers have been hoping to be able to use aerial photography and/or satellite data to do this. However, identification methods have generally not been very accurate. Recently, a new method for identification from satellite or photographic data was developed; it differs from other methods in that each species is compared successively with all other species, rather than simultaneously attempting to match individuals to the descriptions of species. A preliminary study in the UK suggested that this method has much promise. However, it remains to be seen how useful this method is in more challenging settings where there are many species, including many that closely resemble one another. This project will constitute a much more severe trial of this new method, in that it will attempt to correctly identify and map individuals of 18 species (out of 200+) in a first-growth Mexican tropical forest. For each species there is at least one closely-related (and thus similar) species present. This study will thus provide critical data allowing other researchers to determine when this method is useful. Should this particular test be successful, it will also provide a remarkable dataset on the spatial distribution of trees in an intact tropical forest.
