Organic carbon stored in soils of the northern circumpolar permafrost region is a potentially vulnerable component of the global carbon cycle. Recent studies indicate that soil organic carbon (SOC) stocks in the region are likely to be much larger than previous estimates. Yet, the quantity and decomposability of organic carbon contained in permafrost region soils remain highly uncertain, limiting our ability to predict the release of greenhouse gases due to permafrost thawing and the resulting carbon-climate feedbacks under future warming scenarios. We are using soil-forming factors and geospatial approaches to predict the spatial and vertical distributions of SOC stocks at regional scales in studies of climate change impacts on soil carbon in permafrost regions.
Earth system models (ESMs) have been promoted as a convincing way to predict future atmospheric greenhouse gas (GHG) levels and their impact on future warming. Uncertainty due to carbon cycle climate feedback accounts for a large proportion of the overall uncertainty in predictions of future GHG concentrations, anthropogenic and climatic impacts on soil carbon dynamics, and associated climate changes.
Though several recent studies have integrated improved representations of soil dynamics, current ESMs rely on an unrealistic representation of the land surface to infer spatial variability and also lack important soil-forming processes. Therefore, ESM predictions need to be compared with observation-based estimates to test and improve the ESM representations of the terrestrial surface.
Spatially explicit predictions of important soil and ecosystem properties at national and global scales could be essential for testing and improving the ESMs. However, fine-resolution predictions of soil and ecosystem properties that represent natural variability currently do not exist. We are applying our expertise in geospatial modeling, software application and development, data visualization, and web applications to improve the spatial heterogeneity in ESMs and create regional and global land-benchmarking products for data-model intercomparisons.