Mathematical method for measuring the similarities of Hungarian microregions
It would be useful to know the grade of similarity of microregions for landscape classification, landscape potential determination and landscape analysis. Geography measures landscape similarity on the basis of landscape elements – e.g. relief difference, water supply, soil productivity, etc. – and land use data or built-up grades. Unfortunately, comparison of the discrete, nominal or interval data of such indicators is very difficult.
Our method is aimed for the more objective determination of the so called taxonometric distance described above. Nine indicators have been selected: average relative relief, productive layer thickness, PaDI drought index, built-up area ratio, forest covered area ratio, vineyard ratio, NDVI value of vegetation, effective mesh size index of the landscape structure and hemeroby level. Based on these indicators a data matrix was created for all of the 230 microregions of Hungary. Then the so called difference co-efficient was determined with which the Euclidean distance of the microregions according to all of the indicators was calculated. For measuring differences the Multidimensional Scaling (MDS) method, Kruskal stress measurements were used and Agglomerative Hierarchical Clustering (AHC) was performed.
Applying MDS and AHC methods the differences and similarities of microregions in Hungary can be measured more accurately than before. Microregions can be identified that require greater attention in the course of for example landscape planning because, based on the results, they belong to different clusters and are different from the neighbouring microregions at the same landscape hierarchy level.