The prairie soil carbon balance project – modelling and gis component: landscape-scale modelling of changes in soil organic carbon and extrapolation to regional scales

Non-Peer Reviewed

Carbon sequestration in soils has potential to reduce atmospheric CO2 levels and help Canada meet its commitments under the Kyoto Protocol. However, in order for soil carbon sinks to be recognized for purposes of the Protocol, a method for quantifying and verifying changes in soil carbon stocks at regional and national levels must be developed. The objective of the Prairie Soil Carbon Balance Project is to develop a scientifically defensible methodology for quantifying and verifying soil organic carbon (SOC) changes in response to various agricultural management practices. The objective of the modelling and GIS component of the project, being conducted at the University of Saskatchewan, is to combine CENTURY model predictions of changes in SOC levels under alternative management practices, with soil, land-use, and climate databases in order to provide Prairiewide predictions of changes in SOC storage. This modelling process requires “scaling-up” of model point-based output to larger areas. A key element to scaling-up is to define the spatial units over which average sets of driving variables are to be used as model input. The distribution of important driving variables for modelling changes in SOC, such as water inputs, plant residue production, soil redistribution, and initial SOC levels are largely controlled by topography at the landscape scale. Thus, the basic spatial unit to which model runs are applied should be topography-based landscape segments. An approach is being developed to apply landscape-scale SOC modelling to regional scale estimates. This approach involves using soil inventory data to define landscape units and using soil series information to define distinct modal landscape segments and their spatial extent within the larger unit. The model would then be run for each landscape segment, using a segment-specific set of driving variables. Model output would be scaled-up using the spatial extent of each landscape segment within the larger unit.