Predicting Spatial Distribution and Stability of Soil Carbon Based On Soil-Landscape Models, Terrain Attributes and Remote Sensing.

Terrestrial ecosystems can potentially capture and hold carbon at modest social costs. The Intergovernmental Panel on Climate Change (IPCC) Second Assessment Report estimates that by the year 2050 between 60 and 87Gt of C could be conserved or sequestered in forests and between 23 and 44Gt in agricultural soils.  The objective of this research was to identify areas within agricultural fields with the highest potential for storing carbon and predicting these areas at watershed scale.  Approximately 300 georeferenced soil samples were collected, via 10 point transects, across a soil catena covering the soils from the lowest to the highest elevation in the landscape and analyzed for total and mineralizable C.  Total C spatial variability followed the soil catena being higher for Mollisols between 4-6% compared to Alfisols between 1-3%.  Total C also indicated a significant relationship with Topographic Wetness Index (TWI) (R² = 0.5, p-value<0.01).   Mineralizable C also showed a significant positive relationship with TWI.  However, the strength of the relationship decreased with time varying from R² = 0.6 for the day three of incubation to R² = 0.42 (day14^th) to R²<0.01 (day 28^th), indicating the presence of labile C in the lowest landscape positions.  Cumulative mineralizable C decreased with time but remained higher for the Mollisols compared to Alfisols.  The high resolution aerial photography when overlaid with the georeferenced soil C results and the soil map showed a clear threshold boundary between darker and brighter areas of the image, coinciding with the boundary between Mollsols and Alfisols, respectively.  This research suggests that the combination of soil landscape models, terrain attributes, remote sensing, and georeferenced field data can be used to predict spatial distribution and stability of soil C and areas with the highest potential to store soil C.