Ecological models help us understand how ecosystems function, predict responses to global change, and identify future research needs. However, widespread use of models is limited by the technical challenges of model–data synthesis and information management. To address these challenges, we present an ecoinformatic workflow, the Predictive Ecosystem Analyzer (PEcAn), which facilitates model analysis. Herein we describe the PEcAn modules that synthesize plant trait data to estimate model parameters, propagate parameter uncertainties through to model output, and evaluate the contribution of each parameter to model uncertainty. We illustrate a comprehensive approach to the estimation of parameter values, starting with a statement of prior knowledge that is refined by species‐level data using Bayesian meta‐analysis; this is the first use of a rigorous meta‐analysis to inform the parameters of a mechanistic ecosystem model. Parameter uncertainty is propagated using ensemble methods to estimate model uncertainty. Variance decomposition allows us to quantify the contribution of each parameter to model uncertainty; this information can be used to prioritize subsequent data collection. By streamlining the use of models and focusing efforts to identify and constrain the dominant sources of uncertainty in model output, the approach used by PEcAn can speed scientific progress. We demonstrate PEcAn's ability to incorporate data to reduce uncertainty in productivity of a perennial grass monoculture (Panicum virgatum L.) modeled by the Ecosystem Demography model. Prior estimates were specified for 15 model parameters, and species‐level data were available for seven of these. Meta‐analysis of species‐level data substantially reduced the contribution of three parameters (specific leaf area, maximum carboxylation rate, and stomatal slope) to overall model uncertainty. By contrast, root turnover rate, root respiration rate, and leaf width had little effect on model output; therefore trait data had little impact on model uncertainty. For fine‐root allocation, the decrease in parameter uncertainty was offset by an increase in model sensitivity. Remaining model uncertainty is driven by growth respiration, fine‐root allocation, leaf turnover rater, and specific leaf area. By establishing robust channels of feedback between data collection and ecosystem modeling, PEcAn provides a framework for more efficient and integrative science.

Terrestrial biosphere models are designed to synthesize our current understanding of how ecosystems function, test competing hypotheses of ecosystem function against observations, and predict responses to novel conditions such as those expected under climate change. Reducing uncertainties in such models can improve both basic scientific understanding and our predictive capacity, but rarely are ecosystem models employed in the design of field campaigns. We provide a synthesis of carbon cycle uncertainty analyses conducted using the Predictive Ecosystem Analyzer ecoinformatics workflow with the Ecosystem Demography model v2. This work is a synthesis of multiple projects, using Bayesian data assimilation techniques to incorporate field data and trait databases across temperate forests, grasslands, agriculture, short rotation forestry, boreal forests, and tundra. We report on a number of data needs that span a wide array of diverse biomes, such as the need for better constraint on growth respiration, mortality, stomatal conductance, and water uptake. We also identify data needs that are biome specific, such as photosynthetic quantum efficiency at high latitudes. We recommend that future data collection efforts balance the bias of past measurements toward aboveground processes in temperate biomes with the sensitivities of different processes as represented by ecosystem models.

Climate change is leading to a disproportionately large warming in the high northern latitudes, but the magnitude and sign of the future carbon balance of the Arctic are highly uncertain. Using 40 terrestrial biosphere models for Alaska, we 5 provide a baseline of terrestrial carbon cycle structural and parametric uncertainty, defined as the multi-model standard deviation () against the mean (x) for each quantity. Mean annual uncertainty (/x) was largest for net ecosystem exchange (NEE) (−0.01±0.19 kgCm−2 yr−1), then net primary production (NPP) (0.14± 0.33 kgCm−2 yr−1), autotrophic respiration (Ra) (0.09±0.20 kgCm−2 yr−1), gross pri10 mary production (GPP) (0.22±0.50 kgCm−2 yr−1), ecosystem respiration (Re) (0.23± 0.38 kgCm−2 yr−1), CH4 flux (2.52±4.02 gCH4m−2 yr−1), heterotrophic respiration (Rh) (0.14±0.20 kgCm−2 yr−1), and soil carbon (14.0±9.2 kgCm−2). The spatial patterns in regional carbon stocks and fluxes varied widely with some models showing NEE for Alaska as a strong carbon sink, others as a strong carbon source, while still others as 15 carbon neutral. Additionally, a feedback (i.e., sensitivity) analysis was conducted of 20th century NEE to CO2 fertilization () and climate ( ), which showed that uncertainty in was 2x larger than that of , with neither indicating that the Alaskan Arctic is shifting towards a certain net carbon sink or source. Finally, AmeriFlux data are used at two sites in the Alaskan Arctic to evaluate the regional patterns; observed seasonal NEE 20 was captured within multi-model uncertainty. This assessment of carbon cycle uncertainties may be used as a baseline for the improvement of experimental and modeling activities, as well as a reference for future trajectories in carbon cycling with climate change in the Alaskan Arctic

Abstract. Climate change is leading to a disproportionately large warming in the high northern latitudes, but the magnitude and sign of the future carbon balance of the Arctic are highly uncertain. Using 40 terrestrial biosphere models for Alaska, we provide a baseline of terrestrial carbon cycle structural and parametric uncertainty, defined as the multi-model standard deviation (σ) against the mean (x) for each quantity. Mean annual uncertainty (σ/x) was largest for net ecosystem exchange (NEE) (−0.01± 0.19 kg C m−2 yr−1), then net primary production (NPP) (0.14 ± 0.33 kg C m−2 yr−1), autotrophic respiration (Ra) (0.09 ± 0.20 kg C m−2 yr−1), gross primary production (GPP) (0.22 ± 0.50 kg C m−2 yr−1), ecosystem respiration (Re) (0.23 ± 0.38 kg C m−2 yr−1), CH4 flux (2.52 ± 4.02 g CH4 m−2 yr−1), heterotrophic respiration (Rh) (0.14 ± 0.20 kg C m−2 yr−1), and soil carbon (14.0± 9.2 kg C m−2). The spatial patterns in regional carbon stocks and fluxes varied widely with some models showing NEE for Alaska as a strong carbon sink, others as a strong carbon source, while still others as carbon neutral. Additionally, a feedback (i.e., sensitivity) analysis was conducted of 20th century NEE to CO2 fertilization (β) and climate (γ), which showed that uncertainty in γ was 2x larger than that of β, with neither indicating that the Alaskan Arctic is shifting towards a certain net carbon sink or source. Finally, AmeriFlux data are used at two sites in the Alaskan Arctic to evaluate the regional patterns; observed seasonal NEE was captured within multi-model uncertainty. This assessment of carbon cycle uncertainties may be used as a baseline for the improvement of experimental and modeling activities, as well as a reference for future trajectories in carbon cycling with climate change in the Alaskan Arctic.