1 Department of Agroecology - Agricultural Systems and Sustainability, Department of Agroecology, Science and Technology, Aarhus University2 Department of Bioscience - Ecoinformatics and Biodiversity, Department of Bioscience, Science and Technology, Aarhus University3 Department of Bioscience - Biodiversity and Conservation, Department of Bioscience, Science and Technology, Aarhus University4 Department of Agroecology - Agricultural Systems and Sustainability, Department of Agroecology, Science and Technology, Aarhus University5 Department of Bioscience - Biodiversity and Conservation, Department of Bioscience, Science and Technology, Aarhus University6 Department of Bioscience - Ecoinformatics and Biodiversity, Department of Bioscience, Science and Technology, Aarhus University
Complex topography buffers forests against deforestation in mountainous regions. However, it is unknown if terrain also shapes forest distribution in lowlands where human impacts are likely to be less constrained by terrain. In such regions, if important at all, topographic effects will depend on cultural-historical factors and thus be human-driven (anthropogenic) rather than natural, except in regions where the general climate or extreme soils limit the occurrence of forests. We used spatial regression modeling to assess the extent to which topographic factors explain forest distribution (presence-absence at a 48×48 m resolution) in a lowland agricultural region (Denmark, 43,075 km2) at regional and landscape scales (whole study area and 10×10 km grid cells, respectively), how landscape-scale forest-topography relationships vary geographically, and which potential drivers (topographic heterogeneity, forest cover, clay content, coastal/inland location) determine this geographic heterogeneity. Given a moist temperate climate and non-extreme soils all landscapes in Denmark would naturally be largely forest covered, and any topographic relationships will be totally or primarily human-driven. At regional scale, topographic predictors explained only 5% of the distribution of forest. In contrast, the explanatory power of topography varied from 0%–61% at landscape scale, with clear geographic patterning. Explanatory power of topography at landscape scale was moderately dependent on the potential drivers, with topographic control being strongest in areas with high topographic heterogeneity and little forest cover. However, these conditioning effects were themselves geographically variable. Our findings show that topography by shaping human land-use can affect forest distribution even in flat, lowland regions, but especially via localized, geographically variable effects.
Journal of Geographical Sciences, 2014, Vol 24, Issue 1, p. 76-92