| dcterms.abstract | This paper describes and discusses a multivariable methodology for environmental regionalization. The methodology include the following basic processes: 1) A discrimination of the universal space into unit environmental elements (UEE) with maximal homogeneity in all relevant environmental variables (at the level of perception considered); 2) A numerical and/or nominal characterization of each UEE according to values of basic and generated variables; 3) A reinterpretation and evaluation of the characterization variables, through multivariable statistical analysis and 4) A clustering of the UEE into environmentally homogeneous areas (EHA) in a hierarchy of regions, sub-regions, sub-sub-regions, etc. down to the level of the UEE. This approach is geared to environmental planning purposes, and emphasizes flexibility, repeatability, and adequacy for different planning levels, from macro to micro. Those are the main advantages of the methodology. The main disadvantages are the requirements for computer processing, the possible lack of relation between the EHA and field-visualizable units, the need for a careful selection of values and ranges of the characterization variables, and the dependency of the results upon the clustering techniques used, which requires good judgment by the user | en_US |
