Characterization, Delineation and Visualization of Agro-Ecozones Using Multivariate Geographical Clustering

Agro-ecozoning is a delineation of landscape into relatively homogeneous regions of expected similar crop performance. Past classifications have been subjective, crop specific and did not take into account spatial correlation. A quantitative approach is proposed to unambiguously locate, characterise and visualise agro-ecozones and their boundaries which can be allied to different environmental conditions. In this study the environmental parameters, including climatic and soil characteristics, hypothesized to be generally relevant to many crops in Capitanata-Foggia (South Italy), were used. Cokriged environmental estimates at 500 m scale were used in a clustering algorithm based on non-parametric multivariate density estimation. A 3D map of density estimation and red-green-blue colour triplet were used for visualisation of agro-ecozones as a unique combination of environmental factors. The proposed approach produced the delineation of the study area in five compact classes in the space of environmental attributes that were also contiguous in geographic space. The resulting agro-ecozones may provide a framework for useful application in land use decision making.