The conjunctive use of principal component and cluster analysis, tools provided by ArcInfo in raster format, allowed us to carry out an objective climatic zonation. The procedure was applied at the Mojana and Darien regions in Colombia (South America). The regions were divided into cells of four and 10 hectares respectively. For each one of the cells the water balance was applied, taking into account the rainfall spatial variation as well as other climatic variables, hence the decade excess water was obtained. Principal component analysis was applied to the temporal (decade) variation of excess water obtained from a regional water balance. It was found that both regions can be adequately represented as superimposition of two leading maps (Eigen Vector) representing in both cases more than 90 percent of the variation. Natural groups for each region can thus be identified by cluster analysis of the two leading maps.
The zonation was applied to an agroclimatic study for the Mojana Region in order to develop an agricultural planning program. At the Darien Region, the zonation was applied for an environmental study of alternative roads for Darien Gap, which, due to its biotic vulnerability, is the only unbuilt portion of the Pan-American road.
Carlos Alberto González
Universidad Nacional de Colombia
Carrera 54A No. 76-66 Ap. 505
Santa Fé de Bogotá,