Quantifying multi-source carbon cycle model uncertainties: Sensitivity analysis, perturbed parameter ensemble, and uncertainty attribution

Results from terrestrial carbon cycle models have multiple sources of uncertainty, but the relative importance of these sources and how they combine has received little attention. This study investigates how various sources of uncertainty propagate, temporally and spatially, in CASA-D model. CASA-D simulates the impact of climatic forcing and disturbance legacies on forest carbon dynamics with forest inventory & analysis (FIA) data and satellite-derived maps. This study assesses uncertainty from driving data including climatic, fPAR and FIA inputs, and perturbed parameter ensemble. We quantify and report relative importance from each source in estimating forest carbon sink/source in southeast US.