Unprecedented warming of Arctic–boreal regions (ABR) has poorly understood consequences on carbon cycle processes. Uncertainties in annual methane (CH4) budgets partly arise because of limited data availability during winter. In this study, winter CH4 flux measurements were conducted using the snowpack diffusion gradient method over five ABR ecosystem types in Canada and Finland: closed–crown and open–crown coniferous boreal forest, boreal wetland and erect–shrub and prostrate–shrub tundra. Boreal forest uplands acted as net CH4 sinks, while the boreal wetland acted as net CH4 source during winter. We identified several wetland tundra CH4 emission hotspots and large spatial variability in boreal wetland CH4 emissions. In the boreal forest uplands, soil liquid water content was identified as an important environmental control of winter CH4 fluxes. Our results indicate non–negligible winter CH4 flux, which must be accounted for in annual carbon balance and terrestrial biosphere models over ABR.

Drought conditions caused by soil moisture stress and/or high vapour pressure deficit pose a challenge to many terrestrial ecosystem models (TEMs). The Canadian LAnd Surface Scheme Including biogeochemical Cycles (CLASSIC) employs an empirical approach to link soil moisture stress with stomatal conductance. Such soil moisture-based empirical approaches typically perform poorly during drought. Here, we implemented an explicit plant hydraulics parameterization, i.e., Stomatal Optimization based on Xylem hydraulics (SOX), in CLASSIC, thereby connecting the soil-plant-atmosphere continuum through plant hydraulic traits. Performance of the resulting CLASSIC$_{SOX}$ was evaluated against carbon and water fluxes measured with eddy covariance at eight boreal forest flux tower sites in North America. Compared to CLASSIC, CLASSIC$_{SOX}$ better simulated gross primary productivity (GPP) across all sites, i.e., coefficient of determination (R$^{2}$) increased (0.51 to 0.59), root mean square error (RMSE) and bias decreased (1.85 to 1.54 g C m$^{-2}$ d$^{-1}$) and (-0.99 to -0.58 g C m$^{-2}$ d$^{-1}$), respectively. Under drought conditions, identified using the Palmer drought severity index, GPP simulated with CLASSIC$_{SOX}$ improved compared to CLASSIC, i.e., R$^{2}$ increased (0.51 to 0.60), and RMSE and bias decreased (1.79 to 1.46 g C m$^{-2}$ d$^{-1}$) and (-0.97 to -0.53 g C m$^{-2}$ d$^{-1}$), respectively. In contrast, CLASSIC$_{SOX}$ simulated evapotranspiration worsened, i.e., R$^{2}$ decreased (0.61 to 0.42), RMSE increased (0.54 to 0.62 mm d$^{-1}$), and bias changed direction (0.09 to -0.09 mm d$^{-1}$). As evaporation is a highly parameterized process in CLASSIC, it likely needs to be re-parameterized to account for the SOX transpiration behaviour.

Process-based land surface models are important tools for estimating global wetland methane (CH4) emissions and projecting their behavior across space and time. So far there are no performance assessments of model responses to drivers at multiple time scales. In this study, we apply wavelet analysis to identify the dominant time scales contributing to model uncertainty in the frequency domain. We evaluate seven wetland models at 23 eddy covariance tower sites. Our study first characterizes site-level patterns of freshwater wetland CH4 fluxes (FCH4) at different time scales. A Monte Carlo approach has been developed to incorporate flux observation error to avoid misidentification of the time scales that dominate model error. Our results suggest that 1) significant model-observation disagreements are mainly at short- to intermediate time scales (< 15 days); 2) most of the models can capture the CH4 variability at long time scales (> 32 days) for the boreal and Arctic tundra wetland sites but have limited performance for temperate and tropical/subtropical sites; 3) model error approximates pink noise patterns, indicating that biases at short time scales (< 5 days) could contribute to persistent systematic biases on longer time scales; and 4) differences in error pattern are related to model structure (e.g. proxy of CH4 production). Our evaluation suggests the need to accurately replicate FCH4 variability in future wetland CH4 model developments.