Comparing regional differentiation of land cover changes in natural and administrative regions of the Czech Republic using multivariate statistics

The detection and evaluations of land cover changes represent the major task for landscape transformation studies in post-communistic countries. The results of such evaluation are, however, highly influenced by spatial delimitation of monitored units as natural or administrative regions. Moreover, the objective quantitative assessment of land cover changes and their typologies in different types of regions can be hardly done by traditional map-interpretation approach. The aim of the present research was to evaluate the differences in results of land cover changes in the Czech Republic detected in natural (93 geomorphological units) and administrative (77 districts) regions using multivariate statistics. To analyse land cover (LC) changes we used STATISTICA 9 software. The application of principal component analysis (PCA), factor analysis (FA) and cluster analysis (CA) reveals the main overall trends in land cover changes in the Czech Republic. We applied PCA, CA and FA to land cover data from CORINE projects in 1990, 2000 and 2006. We analyzed LC changes in geomorphological units of Czechia as a whole. We made our calculation based on standardised data for land cover classes. The final number of variables (LC classes) used in the study was 11, drawing upon generalisation of only those land cover classes that are present in Czech landscape. For both sets of territorial units (i.e. natural and administrative), we calculated the Euclidean distance (full connections) between the cases (territorial units). The k-means method and hierarchical clustering were used for clustering. Based on these methods we set the typology of land cover changes in natural and administrative units. Finally, we assessed the differences between these typologies as regards statistical distribution of regions among the individual types. The factors influencing differences between these typologies are discussed, concluding in considerations on a role of spatial delimitation in land cover changes studies.