30-31 Jan 2017, Hamburg, MPI-M (Geomatikum Building, Bundesstr. 55, room 1729)
Conveners: Victor Brovkin and Lena Boysen (MPI, Germany)
The meeting of CRESCENDO terrestrial modelling groups was dedicated to the preparation of experiments designed to quantify land use effect on climate. 20 scientists from 7 European ESM groups discussed new aspects of the land use change (LUC) dataset prepared for the historical CMIP6 simulation. This new LUC forcing is much more complex than it was for CMIP5. How could models with different land cover structure interpret LUC scenarios in a consistent way? For example, this new forcing includes rangeland as a new LUC class. Some models assume that rangelands are natural vegetation and ignore the new class, while others treat rangelands as pastures, which are cleaned of forests, and suppress fires etc. Another divergence issue is on sub-grid scale transitions. Several models use so-called gross transitions accounting for crop rotations within grid cell, while others use simplified net transitions. Gross transitions are more difficult to implement; however, they better reflect changes in the land carbon budget. These and other details of the LUC dataset and their usage in ESMs were intensely discussed in a Q&A telecon session with George Hurtt of the University of Maryland, head of the group developing the land use forcing for CMIP6.
All presentations from the meeting are now available in our Intranet
Another topic of the meeting was on the soil carbon model evaluation under LUC. Usually, soil carbon decreases after transition from natural to agricultural land. However, the rate of changes is different among soils and land use classes. Axel Don, a guest scientist from the Thünen Institute, explained details of available metadata on soil carbon chronosequences (soil change in decadal-scale agricultural experiments). The groups agreed on selecting several evaluation sites in temperate and tropical regions.
The meeting ended with discussions on tier-1 LUMIP experiments, scheduled for as soon as the DECK runs are finished. In particular the idealized deforestation experiments were discussed, based on results of earlier idealized deforestation experiments and new insight into possible non-linearities in model response, as were the factorial land management experiments, for which a common platform with LUMIP to exchange details on the protocol was suggested.
Contact person for workshop details: Lena Boysen [email protected]