Faculty of Mathematics and Natural Sciences - Applied Geoinformation Science

Modeling cropland dynamics in Romania

Modeling cropland dynamics in Romania – a comparison of logistic regressions and neural networks

 

(German: Modellierung von landwirtschaftlicher Nutzung in Rumänien – ein Vergleich von logistischer Regression und neuronalen Netzen)

Cropland change is the dominating land use change process in Romania, the second largest new member state of the European Union. This project investigates cropland dynamics in Argeş County, Romania, in cooperation IAMO and in continuation of a project of the junior research group “Postsocialist Land Reforms and Land Use”. Land cover change is monitored and analyzed using hybrid classification of Landsat data. We study cropland change by using two land use modelling techniques. We assess the underlying causes of cropland change using spatially explicit logistic regressions. Maps of likely cropland change are derived by neural network analysis (Land Transformation Model). We identify hot spots of change as those areas that are under most eminent threat of change as a possible tool for spatial decision-making.

Principal Investigators:
Tobia Lakes (Humboldt-Universität zu Berlin)
Daniel Müller (Leibniz Institute of Agricultural Development in Central and Eastern Europe (IAMO))