Spatial modelling of greenhouse gas compartments for improved full carbon accounting in boreal ecosystems

In the framework of the EU project SIBERIA-II (2002 - 2005), the International Institute for Applied Systems Analysis (IIASA) developed a regression-based semi-empirical model for diagnostic predictions of carbon stocks and fluxes for a 3 Mio km² study region in Central Siberia. This thesis deals with the development of geographical methods for the operational spatial delineation of an input database for this model, such as data fusion, spatial modelling and regionalisation of geocomplexes. A concept has been developed for the regionalisation of so called Greenhouse Gas Compart-ments (GGCs) – regions with the same characteristics regarding carbon pools and fluxes. Since soil and vegetation are the main terrestrial carbon pools and their distribution and type are the main parameters causing differences in carbon assimilation and emissions, GGCs are defined as regions with the same vegetation and soil types. Thus, this thesis deals with two applications of landscape ecology, a) predictive vegetation modelling using GIS and decision rules for the post classification of remote sensing products, and b) soil landscape modelling using GIS and digital elevation models for the refinement of an already existing soil map. Combining the resulting vegetation and soil map GGCs could be delineated and used as input for full carbon accounting.