GIS-gestützte Modellierung der räumlichen Verteilung der Vegetation im Tidebereich von Ästuaren unter den Bedingungen einer Klimaänderung mittels der Klassifikations- und Regressionsanalyse (CART) am Beispiel der Unterweservorländer

Abstract

The objective of this thesis was to develop a GIS-based model to predict changes of vegetation types due to climate change and sea level rise. The investigation area was the Weser estuary in northern Germany covering 50 km2. Classification tree analysis (CART) was used to test specific environmental variables affecting the land use and the dominant vegetation patterns of the tidal wetlands based on two different land use scenarios. The predictive accuracy was assessed using cross validation. The hit rate varied between 65 % for the land use types and 81 % for the vegetation types. The vegetation of the Weser estuary was modelled using explanatory topographic and hydrological variables. The CART model produced accurate and plausible predictions. The results illustrate the qualification and suitability of the modelling approach to predict changes of vegetation types for estuaries. The thesis is based upon the results of the BMBF project 'Klimaänderung und Unterweserregion' (KLIMU).