Hana STANKOVÁ, Lenka ŠOŠOVIČKOVÁ

Landscape change modelling and prediction using CORINE Land Cover data: case study from north-east Slovakia

Stanková, H., Šošovičková, L.: Landscape change modelling and prediction using CORINE Land Cover data: case study from north-east Slovakia. Kartografické listy, 2013, 21 (1).

Abstract: Analysis of land use (LU) and land cover (LC) changes belong to the most important tasks in the geographical applications. Recent inventories of operational models for LU and LC changes are numerous. The Land Change Modeler (LCM) for Ecological Sustainability is a software solution designed to address the pressing problem of accelerated land conversion. The aim of this study is to model and predict the changes in LC in the northeastern part of Slovakia using LCM with CORINE Land Cover data from years 1990, 2000 and 2006. In years 1990 – 2000 there were three main processes within the area of interest: plant succession, deforestation and intensification of agriculture. We predicted these processes for year 2006 and compared the results with actual LC from year 2006. Comparing the whole LC layers, accuracy was relatively high (Kappa index higher than 0,8) due to the large areas without change. Comparing only the areas of change, accuracy was very low (Kappa index 0,2 was the best result, obtained for the process of plant succession). The critical disadvantage of LCM approach is the change prediction based on the previous development only. The discrepancy between predicted and actual LC changes is mainly due to the changes in trends of land conversion, especially after Slovakia joined the EU in May 2004. The change of agriculture policy caused an increase of agriculture intensification and conversely a decrease of plant succession on the abandoned land. Besides these changes in trends, another difficulty represents the processes closely related to human decisions (i.e. deforestation). In this manner, we could predict ongoing processes, but the completely new changes caused by political, economical or other human decisions are unpredictable without very detailed additional information.

Keywords: land cover, change prediction, modelling, Land Change Modeler (LCM), CORINE Land Cover (CLC)