Improving carbon modeling in aspen-dominated forests: Lessons from CBM-CFS3 evaluation against long-term field data

Boreal forests cover a substantial portion of Canada's land base and play a central role in the national carbon budget through carbon storage and fluxes across multiple above- and belowground pools. The Carbon Budget Model of the Canadian Forest Sector (CBM-CFS3) simulates and projects forest carbon dynamics under varying conditions and management scenarios, and ongoing model evaluation and calibration are essential to maintain accuracy in carbon monitoring, accounting, and reporting. This study assesses CBM-CFS3 performance using independent ground plot data from 64 aspen-dominated stands that have been monitored since 2000 as part of the Climate Impacts on the Productivity and Health of Aspen (CIPHA) project. We compared modeled and observed C stocks in aboveground biomass, snags, forest floor, and mineral soil pools, and examined the influence of stand characteristics, climate variables, and soil chemical properties on model residuals. Model performance varied across carbon pools: CBM-CFS3 most accurately simulated aboveground biomass C stocks, performed moderately well for snags, but failed to capture spatial variability in forest floor and mineral soil C stocks. Our findings suggest that incorporating stand density into the yield curve model and stand age into volume-to-biomass conversion equations would enhance the accuracy of aboveground biomass C stock estimates in aspen-dominated stands. Representation of snag carbon dynamics could be improved by accounting for lagged mortality following drought events and by increasing snag fall rates. Further refinement of forest floor carbon dynamics will require integrating the influence of moisture on organic matter decomposition, while improving mineral soil carbon dynamics will require incorporating the effect of clay content on decomposition and reducing the temperature sensitivity of decay. Because these pools are interdependent, refinements in one pool will have cascading effects on downstream pools and overall carbon dynamics. The CBM-CFS3 remains a key tool for national carbon estimation, and the availability of high spatial and temporal resolution data, such as CIPHA, for model evaluation will further strengthen the model's predictive capability.