GC41F-0660:
From Land Use to Land Cover: Restoring the Afforestation Signal in a Coupled Integrated Assessment - Earth System Model and the Implications for CMIP5 RCP Simulations

Thursday, 18 December 2014
Alan V Di Vittorio1, Louise P Chini2, Ben P Bond-Lamberty3, Jiafu Mao4, Xiaoying Shi4, John Truesdale5, Anthony Craig5, Katherine V Calvin3, Andrew D Jones1, William Collins1, James Edmonds3, George C Hurtt2, Peter E Thornton4 and Allison M Thomson3, (1)Lawrence Berkeley National Laboratory, Berkeley, CA, United States, (2)University of Maryland College Park, College Park, MD, United States, (3)Joint Global Change Research Institute, Pacific Northwest National Laboratory, College Park, MD, United States, (4)Oak Ridge National Laboratory, Oak Ridge, TN, United States, (5)Independent contractor with Lawrence Berkeley National Laboratory, Berkeley, CA, United States
Abstract:
Climate projections depend on scenarios of fossil fuel emissions and land use change, and the IPCC AR5 parallel process assumes consistent climate scenarios across Integrated Assessment and Earth System Models (IAMs and ESMs). The CMIP5 project used a novel “land use harmonization” based on the Global Land use Model (GLM) to provide ESMs with consistent 1500-2100 land use trajectories generated by historical data and four IAM projections. A direct coupling of the Global Change Assessment Model (GCAM), GLM, and the Community ESM (CESM) has allowed us to characterize and partially address a major gap in the CMIP5 land coupling design: the lack of a corresponding land cover harmonization. The CMIP5 CESM global afforestation is only 22% of GCAM’s 2005 to 2100 RCP4.5 afforestation. Likewise, only 17% of GCAM’s 2040 RCP4.5 afforestation, and zero pasture loss, were transmitted to CESM within the directly coupled model. This is a problem because afforestation was relied upon to achieve RCP4.5 climate stabilization. GLM modifications within the directly coupled model did not increase CESM afforestation. Modifying the CESM land use translator in addition to GLM, however, enabled CESM to simulate 66% of GCAM’s afforestation in 2040, and 94% of GCAM’s pasture loss as grassland and shrubland losses. This additional afforestation increases vegetation carbon gain by 19 PgC and decreases atmospheric CO2 gain by 8 ppmv from 2005 to 2040, implying different RCP4.5 climate scenarios between CMIP5 GCAM and CESM. Although the IAMs and ESMs were not expected to have exactly the same climate forcing, due in part to different terrestrial carbon cycles and atmospheric radiation algorithms, the ESMs were expected to project climates representative of the RCP scenarios. Similar land cover inconsistencies exist in other CMIP5 model results, primarily because land cover information is not shared between IAM and ESM models. High RCP4.5 afforestation might also contribute to inconsistencies as some ESMs might impose bioclimatic limits to potential forest area and have different rates of forest growth than projected by RCP4.5. Further work to harmonize land cover among models will be required to address this problem.