Linking Biogeochemistry and Atmospheric Transport in the NCAR CSM

 NSF Small Grants for Exploratory Research 1999-2000

The Climate System Model (CSM) at the National Center for Atmospheric Research (NCAR) is a modular system of state-of-the-science components for the simulation of processes and interactions in the Earth system that determine our climate (Boville and Gent, 1998). CSM components have been used to simulate biogeochemical cycles on land (Bonan, 1995) and in the ocean (S. Doney, personal communication), and these simulations have been compared to appropriate local or regional data. I propose an exploratory study linking the biogeochemistry modules of the NCAR Climate System Model (CSM) with the atmospheric transport code in CCM3, and comparing the results to available observations. A first step will be to evaluate the passive chemical tracer transport of CCM3 by repeating an SF6 experiment recently completed by other major models in an international intercomparison (TransCom 2, Denning et al, 1999). Assuming this is successful, a full follow-on proposal will be submitted to develop improved biogeochemical linkages among the CSM components, and to test these flux simulations against atmospheric data.

Some simulations of atmospheric CO2 have been performed with the CCM. Erickson (1996) performed seasonal simulations with CCM2 using prescribed surface fluxes and found reasonable agreement with observations. CCM2 was also a participant in the first phase of the TransCom intercomparison, in which it was found to exhibit a strong interhemispheric gradient of CO2 due to seasonal exchange with vegetation (Law et al, 1996). This “rectifier effect” arises from seasonal covariance between CO2 fluxes and atmospheric transport (Denning et al, 1995, 1996). Some effort has also been made to test CSM-simulated CO2 fluxes against atmospheric CO2 data (Craig and Holmen, 1998; Craig et al, 1999), but these simulations lacked realism because the carbon budget was badly out of balance in many parts of the world. The size of carbon pools in soils and vegetation in LSM were prescribed a priori and were not allowed to adjust to simulated fluxes, leading to very strong regional sources and sinks and regional anomalies in the simulated atmospheric CO2 concentration field that were not consistent with observations.