Global Multiscale Modeling of Coupled Carbon-Climate Processes

To investigate the effects of spatial scales on land-atmosphere interactions, we performed three 30-year simulation experiments with the Community Earth System Model (CESM). In the control experiment, each atmospheric grid column interacted with a single instance of the land model (CLM). We refer to this arrangement as Single Atmosphere / Single Land, or SASL. In a second experiment, all subgrid-scale atmospheric processes (e.g., radiation, clouds, precipitation, turbulence) in the Community Atmosphere Model (CAM) were replaced with a nonhydrostatic Cloud Resolving Model (CRM). Here we averaged the effect of subgrid-scale atmospheric processes that interacted with a single instance of CLM. We refer to this configuration as Multiple Atmosphere / Single Land (MASL). In the third experiment, land-atmosphere interactions are simulated on the 4-km CRM grid, with 32 separate instances of CLM in every CAM column. This configuration is called Multiple Atmosphere / Multiple Land, or MAML.

We found that although total precipitation was very similar among the experiments, precipitation intensity was much greater (and more realistic) in the MAML and especially the MASL runs. As a result, tropical and other moist forests experienced more runoff and less soil moisture storage than in SASL. This change in the fate of land-surface water changed the seasonal behavior of ecosystems and the development of drought stress. The distribution of solar radiation was also quite different in MAML and MASL compared to SASL. As a result of these interactions, global photosynthesis was reduced by about 20% in MAML relative to MASL and SASL.