Search

Changing climate conditions impact ecosystem dynamics and have local to global impacts on water and carbon cycles. Many processes in dynamic vegetation models (DVMs) are parameterized, and the unknown/unknowable parameter values introduce uncertainty that has rarely been quantified in projections of forced changes. In this study, we identify processes and parameters that introduce the largest uncertainties in the vegetation state simulated by the DVM Top‐down Representation of Interactive Foliage and Flora Including Dynamics (TRIFFID) coupled to a regional climate model. We adjust parameters simultaneously in an ensemble of equilibrium vegetation simulations and use statistical emulation to explore sensitivities to, and interactions among, parameters. We find that vegetation distribution is most sensitive to parameters related to carbon allocation and competition. Using a suite of statistical emulators, we identify regions of parameter space that reduce the error in modeled forest cover by 31±9%. We then generate large initial atmospheric condition ensembles with 10 improved DVM parameterizations under preindustrial, contemporary, and future climate conditions to assess uncertainty in the forced response due to parameterization. We find that while most parameterizations agree on the direction of future vegetation transitions in the western United States, the magnitude varies considerably: for example, in the northwest coast the expansion of broadleaf trees and corresponding decline of needleleaf trees ranges from 4 to 28% across 10 DVM parameterizations under projected future climate conditions. We demonstrate that model parameterization contributes to uncertainty in vegetation transition and carbon cycle feedback under nonstationary climate conditions, which has important implications for carbon stocks, ecosystem services, and climate feedback. Plain Language Summary: Changing climate conditions can impact the composition of an ecosystem. As plant types better suited for the new climate conditions begin to thrive, they can out‐compete the historical vegetation. This transformation can have impacts on the local water cycle and, if changes are large enough, on the global carbon cycle. Dynamic vegetation models (DVMs) are used to analyze the impacts of climate on vegetation and project future changes. DVMs are used across scales ranging from watersheds, informing management decisions, to the globe as an important component of the global carbon cycle. The mathematical representations of processes in DVMs are imperfect and introduce uncertainty in model projections. In this study we identify processes that introduce the largest uncertainties in the vegetation distribution modeled by a DVM. We then identify settings of model parameters that improve the modeled historical vegetation when compared to observations. While unique model parameterizations simulated the historical vegetation well, each parameterization simulated different future vegetation transitions under future climate conditions. Our study shows that model parameterization introduces uncertainty in the projected future changes in vegetation. Quantifying and appreciating this uncertainty are important when considering management decisions based on projected vegetation changes at the local to global scale. Key Points: Multiple unique parameterizations of a dynamic vegetation model met evaluation criteriaParameterizations simulated varying amounts of vegetation transitions under nonstationary climate contributing to uncertainty in the simulated carbon cycleSimulated vegetation distributions were most sensitive to parameters related to carbon allocation and competition [ABSTRACT FROM AUTHOR]

Understanding the unfolding challenges of climate change relies on climate models, many of which have large summer warm and dry biases over Northern Hemisphere continental midlatitudes. This work, with the example of the model used in the updated version of the weather@home distributed climate model framework, shows the potential for improving climate model simulations through a multiphased parameter refinement approach, particularly over the northwestern United States (NWUS). Each phase consists of (1) creating a perturbed parameter ensemble with the coupled global–regional atmospheric model, (2) building statistical emulators that estimate climate metrics as functions of parameter values, (3) and using the emulators to further refine the parameter space. The refinement process includes sensitivity analyses to identify the most influential parameters for various model output metrics; results are then used to cull parameters with little influence. Three phases of this iterative process are carried out before the results are considered to be satisfactory; that is, a handful of parameter sets are identified that meet acceptable bias reduction criteria. Results not only indicate that 74 % of the NWUS regional warm biases can be reduced by refining global atmospheric parameters that control convection and hydrometeor transport, as well as land surface parameters that affect plant photosynthesis, transpiration, and evaporation, but also suggest that this iterative approach to perturbed parameters has an important role to play in the evolution of physical parameterizations. [ABSTRACT FROM AUTHOR]