Key Processes Underlying CO₂ Fertilization Effect In Land Surface Models And Parameter Uncertainty

Terrestrial ecosystems are one of the most important regulators for climate change,in that terrestrial ecosystems behave as an important sink of anthropogenic CO₂ emissions.Accurate estimates of terrestrial carbon?C?sink and its response to climate change are critical for projecting future climate change.Land surface models?LSMs?behave as indispensable tools for simulating land surface processes and C cycles.However,large uncertainty of modelled responses to climate change,especially to the increasing CO₂concentration in the atmosphere,still exist among different LSMs.This study focuses on identifying the key processes underlying the response of the terrestrial C cycle to increasing CO₂?i.e.,CO₂ fertilization effect,often denoted as??in a LSM.In addition,uncertainty in model parameters is another very important source of model uncertainty.We also investigated how parameters vary spatially in this study.In this study,model simulations were conducted using the Australian land surface model—CABLE.CO₂ fertilization effects at various hierarchical levels from leaf biochemical reaction to canopy gross primary production?GPP?,net primary production?NPP?,and ecosystem carbon storage were calculated for different C₃ plant functional types?PFTs?.Our results show that?values at leaf level vary little across different PFTs,but greatly diverge at canopy and ecosystem levels in all simulations.The major jump in the variation of ? values from leaf-to canopy-and ecosystem-levels results from divergence in the modelled leaf area index?LAI?within and among PFTs.Then?at the ecosystem level in the CABLE model was systematically decomposed into sensitivities of relevant C-cycle components:peak leaf-level photosynthesis(P_max),peak leaf area index(LAI_max),phenological index,C-use efficiency?CUE?,and C turnover time???.Results show that the sensitivity of LAI_max contributes most to the magnitude and variation of?.Results from other models also show a high correlation between?at the ecosystem level and LAI response to increasing CO₂.In this study,a synthesis of the responses of photosynthesis,LAI,NPP,biomass,and ecosystem C storage to elevated CO₂ in CO₂ enrichment experiments confirms that LAI is the most influential factor for ?.Model parameters need to be adjusted according to specific locations and climate change scenarios.By assimilating C flux data at 12 eddy-covariance towers in the conterminous USA into an ecosystem model,the variation of parameters across different sites was studied.Results show that only a small fraction of parameters can be fixed,and many parameters vary a lot across sites.Some of the variations are related to the acclimation of ecological processes to the environment and model deficiency.Therefore,considering the adjustment of parameters for modeling LAI under climate change is the key to advancing the prediction power of LSMs.Experimental attempts should also focus more on the change of plant traits under increasing CO₂ and nutrient limitations.