Impact of uncertainty in attributing modeled North American terrestrial carbon fluxes to anthropogenic forcings

Abstract:

Although much progress has been made in the past decade in constraining the net North American terrestrial carbon flux, considerable uncertainty remains in the sink magnitude and trend. Terrestrial carbon cycle models are increasing in spatial resolution, complexity and predictive skill, allowing for increased process-level understanding and attribution of net carbon fluxes to specific causes. Here we examine the various sources of uncertainty, including driver uncertainty, model parameter uncertainty, and structural uncertainty, and the contribution of each type uncertainty to the net sink, and the attribution of this sink to anthropogenic causes: Increasing CO2 concentrations, nitrogen deposition, land use change, and changing climate. To examine driver and parameter uncertainty, model simulations are performed using the Community Land Model version 4.5 (CLM4.5) with literature-based parameter ranges and three different reanalysis meteorological forcing datasets. We also examine structural uncertainty thorough analysis of the Multiscale Terrestrial Model Intercomparison (MsTMIP). Identififying major sources of uncertainty can help to guide future observations, experiments, and model development activities.