Mohamed Jabloun1,2, Jørgen E. Olesen2, Marianne Zandersen3, Kari Petri Hyytiäinen4, Erik Smedberg5
1School of Biosciences, University of Nottingham, Loughborough, UK
2Dept. of Agroecology, Aarhus University, Tjele, Denmark
3Dept. of Environmental Science, Aarhus University, Roskilde, Denmark
4Dept. of Economics and Management, University of Helsinki, Helsinki, Finland
5Baltic Nest Institute Sweden, Baltic Sea Centre, Stockholm University, Stockholm, Sweden
The Baltic Sea has suffered from severe effects of eutrophication for many decades and achieving greater sustainability and ecological restoration is becoming urgent. Therefore, a scenario-based modeling framework is needed to support the analysis of possible impacts of land-use change, and to delineate potential mitigation strategies on nutrient loading under an uncertain future climate. Whilst a wide range of climate change scenarios have been available to the community for several years, the development of land-use change scenarios can be considered relatively recent and a small range of models are available. Thus, to effectively prepare for change, policymakers would need as a first step information about how socioeconomic scenarios and different future climate projections may influence future land use and land cover (LULC) in the Baltic Sea Basin.
In this study, we developed a scenario-based modeling framework to analyze potential future land-use change in the Baltic Sea Basin. Expert knowledge was used to develop quantitative demand for future LULC change for three shared socio-economic pathways i.e. the sustainability pathway (SSP1), the Middle of the Road (SSP2) and Fossil Fueled Development (SSP5). The changes related to urban/built up areas were determined separately and the projected future population in 2050 under the different SSPs was used to determine the urban expansion as compared to current situation. For the other LULC classes (i.e. forest, grassland, cropland, and bare/spare vegetation) a LULC model was developed using the Random Forest (RF) classification tree. Data on historical LULC in 2010 and selected biophysical and historical climate indices were used as drivers. The RF LULC model was used to generate the probability maps of each of the LULC classes for four different climate projections for the relatively near-future time frame period 2041-2060 under RCP8.5 scenario. These probability maps were used to allocate the different land use demand for the different SSP story lines and spatially explicit future land-use maps were produced for the different climate projections.