Spatial analysis and modelling of soil organic carbon using data mining and digital soil mapping along a climatic gradient in West Greenland

Soils are an important carbon sink, which stores up to 2500 Pg carbon worldwide. Thereof, arctic soils contain the major portion of soil organic carbon (SOC). Being a sensible ecosystem, the Arctic is sensitive to climate change. Hence, thawing of permafrost-affected soils to greater depth and for longer periods increases the release of CO2 to the atmosphere, which queries soils as important carbon pool. There are several studies on predicting SOC content of arctic soils focusing on different depths, scales and regions. Most surveys base on soil or landscape units representing averaged point information. Being limited respecting intense natural and spatial variation of soils, these approaches are susceptible to errors of estimation. The spatial heterogeneity of soil properties, especially soil carbon, results from environmental factors varying highly in space and scale. Although, dealing with diversified soil properties, pixel based approaches are less widespread at pedological studies of the Arctic. This study comprises an area-wide prediction approach of SOC at two different study areas in Western Greenland using digital soil mapping (DSM) and data mining (DM). The objectives are identifying major environmental factors controlling spatial variation of SOC in Western Greenland, creating high-resolution maps of relevant soil properties, and to estimate carbon stocks.