| Abstract: | Abstract The coastal landscape in Santander, Spain, is analyzed in terms of landscape quality and fragility (or vulnerability). Quality is intended to indicate those areas of major landscape value for conservation purposes. Fragility is utilized to detect areas which could easily deteriorate if certain human activities are carried out. The authors’ study has three clearly differentiated phases. In the first, landscape units are defined and mapped. In the second, units are classified by computerized techniques and grouped according to significance. In the third phase, the true significance of the groups is interpreted. A mountainous, coastal zone in northern Spain was chosen for this study. Landscape units were determined by visual criteria with strong topographical bias. The area covered by each unit is that visible to an observer situated approximately in its center. Each unit is represented by certain biophysical and visual variables. Data were subjected to clustering analysis on the basis of using three ad casum defined “distances.”; The first distance gives unit classifications grouped according to similarity or proximity of their characteristic variables. In this case, inventory data only are used, no subjective value judgments were introduced in the process. Value judgments are made at the end on the resulting groupings of units. Two other distance classifications are determined by the similarity or proximity of the units making up each group according to the landscape quality and fragility values of each unit. In these latter cases, classification is influenced by value judgments introduced at the beginning of the process. These procedures are more subjective but give more congruent results. With each mathematical distance configuration, an analysis is made of the variables shown to be most representative and which have, therefore, shown greatest discriminatory power in selecting final groups. Finally, the different possibilities of the above techniques are discussed as well as their potential generalization to other fields demanding the treatment of qualitative landscape variables. |
