Carbon-concentration and carbon-climate feedbacks in CMIP6 models, and their comparison to CMIP5 models

Abstract

<p><strong>Abstract.</strong> Results from the fully-, biogeochemically-, and radiatively-coupled simulations in which CO<sub>2</sub> increases at a rate of 1 % per year (1pctCO2) from its pre-industrial value are analyzed to quantify the magnitude of two feedback parameters which characterize the coupled carbon-climate system. These feedback parameters quantify the response of ocean and terrestrial carbon pools to changes in atmospheric CO<sub>2</sub> concentration and the resulting change in global climate. The results are based on eight comprehensive Earth system models from the fifth Coupled Model Intercomparison Project (CMIP5) and eleven models from the sixth CMIP (CMIP6). The comparison of model results from two CMIP phases shows that, for both land and ocean, the model mean values of the feedback parameters and their multi-model spread has not changed significantly across the two CMIP phases. The absolute values of feedback parameters are lower for land with models that include a representation of nitrogen cycle. The sensitivity of feedback parameters to the three different ways in which they may be calculated is shown and, consistent with existing studies, the most relevant definition is that calculated using results from the fully- and biogeochemically-coupled configurations. Based on these two simulations simplified expressions for the feedback parameters are obtained when the small temperature change in the biogeochemically-coupled simulation is ignored. Decomposition of the terms of these simplified expressions for the feedback parameters allows identification of the reasons for differing responses among ocean and land carbon cycle models.</p>

Ensemble projections elucidate effects of uncertainty in terrestrial nitrogen limitation on future carbon uptake

“…As many other studies (Arora et al., 2019), this study does not consider the limitation of the terrestrial C cycle by the phosphorus (P) cycle. A first field‐scale free‐air CO 2 enrichment experiment in an Australian Eucalypt forest with low P availability has shown that P limitation can severely constrain the response (Ellsworth et al., 2017).…”

“…These models generally suggest that N constraints attenuate the land C response to global change (Sokolov et al., 2008; Thornton et al., 2009; Wårlind, Smith, Hickler, & Arneth, 2014; Zaehle, Friedlingstein, & Friend, 2010; Zhang, Wang, Matear, Pitman, & Dai, 2014). This is also the case in the new set of ESMs in the Coupled Model Intercomparison Project Phase 6 (CMIP6; Arora et al., 2019), where about half of the models include a N cycle representation. However, adequate ecosystem‐scale characterization of terrestrial N cycle processes remains a challenge because of the complexity of N cycle processes, their spatial heterogeneity and the long‐term cumulative effect of comparatively small fluxes such as biological nitrogen fixation (BNF) on ecosystem N availability (Thomas, Brookshire, & Gerber, 2015).…”

“…Terrestrial biosphere models (TBMs) that estimate the future C‐uptake potential of land vegetation increasingly consider the dynamics of the global N cycle and its linkage to the C cycle (Arora et al., 2019; Zaehle & Dalmonech, 2011). These models generally suggest that N constraints attenuate the land C response to global change (Sokolov et al., 2008; Thornton et al., 2009; Wårlind, Smith, Hickler, & Arneth, 2014; Zaehle, Friedlingstein, & Friend, 2010; Zhang, Wang, Matear, Pitman, & Dai, 2014).…”

Beyond Modeling: A Roadmap to Community Cyberinfrastructure for Ecological Data-Model Integration

“…Assessment of model skills and representations are also beneficial when multiple models are considered to identify shortfalls or generalizability of different approaches (Schwalm et al, 2019), and to specify areas where mechanistic understanding is incomplete (Tuomi et al 2008). Model intercomparison projects (MIPs) have also been constructed to estimate the range of possible responses to future change or processes at large scales and to identify commonalities, divergence, and uncertainties across models (Friedlingstein et al 2006(Friedlingstein et al , 2014Palosuo et al 2012;Todd-Brown et al 2013;Huntzinger et al, 2017;Arora et al, 2019). While MIPs are scarce in the rest of ecology, they are relatively common for terrestrial ecosystem models (Warszawski et al, 2014;De Kauwe et al, 2014;Rollinson et al, 2017;Huntzinger et al, 2017;Müller et al, 2019.…”

“…Yet, having more data than ever before, we have not seen comparable progress in our capacity to forecast natural systems with our process-based models. For example, model projections out to the year 2100 do not agree on whether the land will be a carbon sink or source in response to climate change, and this has not changed despite years of apparent model improvement (Friedlingstein et al, 2006(Friedlingstein et al, , 2014Arora et al, 2019). The goal of this paper is to better characterize the bottlenecks that have obstructed the rates at which new information has been translated into knowledge and then integrated into models, and to lay out a roadmap for how to overcome these bottlenecks.…”