Land Classification of Spain: A Flexible Sclae Multifactor Framework for Forestry Purposes

Abstract

A nation-wide Land Classification of Spain useful for renewable resources assessment, inventory and research, has been developed at INIA after several successful experiences carried out at regional level. The construction of this classification mainly aimed to provide a land model based on physical factors with undoubted forestry significance that are nonchanging at human life span. Because of the size of Spain (504.750 km2), its ecological diversity and available cartographic information, we had to design an original land classification method showing he following features: a) Divisive hierarchical multifactor framework in four progressive phases with increasing resolution power (25, 4, 1 and 0,25 km2 sample size) ad source map scale (1/400.000, 1/200.000, 1/50.000 and 1/10.000), and decreasing land coverage in accordance with land variability. b) Climatic, geological, and physiographic attributes actively used in land classes definition process, and biological, current land use, soil or remote sensing information passively used as land classes test. c) Twinspan multivariate analysis applied in order (1) to detect the main existing multifactor gradient (2) to arrange and classify the sample squares (3) to identify squares throughout the analysed land using a reduced number of indicator attributes. In the first phase, an ecoregionalization process took place, and eight main ecoregions were defined and delimited by means of 570 sample squares of 25-km2 size, characterized by 174 climatic, geological, and physiographic attributes. The resulting ecoregions showed highly significant correlation with phytosciological taxa distribution. The second phase has been developed separately for each ecoregion. After its completion, 252 Land Classes have been defined using a systematically selected sample of 5450 squares of 4 km2. Afterwards, 125000 squares have been identified and all the Spanish Land Classes were mapped and forestry characterized by means of an ARC-INFO Geographical Information System. The obtained framework can be used in different scale forest studies. At present, it has been successfully used in studies on native and nonnative forests (Pinus nigra plantations, Quercus suber woodlots, etc.). Important ecological and silvicultural features have appeared highly dependent on the Land Classification. As a consequence, land suitability maps for different forestry purposes have been designed.