TY  - JOUR
A1  - Zechlau, Sabrina
A1  - Schlund, Manuel
A1  - Cox, Peter M.
A1  - Friedlingstein, Pierre
A1  - Eyring, Veronika
T1  - Do Emergent Constraints on Carbon Cycle Feedbacks Hold in CMIP6?
Y1  - 2022-12-05
VL  - 127
IS  - 12
JF  - Journal of Geophysical Research: Biogeosciences
DO  - 10.1029/2022JG006985
PB  - 
N2  - Emergent constraints on carbon cycle feedbacks in response to warming and increasing atmospheric CO<sub>2 </sub> concentration have previously been identified in Earth system models participating in the Coupled Model Intercomparison Project (CMIP) Phase 5. Here, we examine whether two of these emergent constraints also hold for CMIP6. The spread of the sensitivity of tropical land carbon uptake to tropical warming in an idealized simulation with a 1% per year increase of atmospheric CO<sub>2 </sub> shows only a slight decrease in CMIP6 (−52 ± 35 GtC/K) compared to CMIP5 (−49 ± 40 GtC/K). For both model generations, the observed interannual variability in the growth rate of atmospheric CO<sub>2 </sub> yields a consistent emergent constraint on the sensitivity of tropical land carbon uptake with a constrained range of −37 ± 14 GtC/K for the combined ensemble (i.e., a reduction of ∼30% in the best estimate and 60% in the uncertainty range relative to the multimodel mean of the combined ensemble). A further emergent constraint is based on a relationship between CO<sub>2 </sub> fertilization and the historical increase in the CO<sub>2 </sub> seasonal cycle amplitude in high latitudes. However, this emergent constraint is not evident in CMIP6. This is in part because the historical increase in the amplitude of the CO<sub>2 </sub> seasonal cycle is more accurately simulated in CMIP6, such that the models are all now close to the observational constraint.
N2  - Plain Language Summary: The statistical model of so‐called emergent constraints help to better understand the sensitivity of Earth system processes in a changing climate. Here, we analyze the robustness of two previously found emergent constraints on carbon cycle feedbacks, using models from the Coupled Model Intercomparison Project (CMIP) of Phases 5 and 6. First the decrease of carbon storage in the tropics due to increasing near‐surface air temperatures, which is found to be robust on the choise of model ensemble. Giving a constraint estimate of −52 ± 35 GtC/K for CMIP6 models, being within the range of uncertainty for the previously estimated result for CMIP5. Second, the increase of carbon storage in high latitudes due to CO<sub>2 </sub> fertilization effect, which is found to be not evident among CMIP6 models. This is in part because the historical increase in the amplitude of the CO<sub>2 </sub> seasonal cycle is more accurately simulated in CMIP6, such that the models are all now close to the observational constraint.
N2  - Key Points: An emergent constraint on the sensitivity of tropical land carbon to global warming, originally derived from Coupled Model Intercomparison Project Phase 5 (CMIP5), also holds for CMIP6. The combined CMIP5 + CMIP6 ensemble gives an emergent constraint on the sensitivity of tropical land carbon to global warming of −37 ± 14 GtC/K. An emergent constraint on the fertilization feedback due to rising CO<sub>2 </sub> levels, previously derived, is not evident in CMIP6.
UR  - http://resolver.sub.uni-goettingen.de/purl?gldocs-11858/11458
ER  -