Global research alliance on agricultural greenhouse gases - benchmark and ensemble crop and grassland model estimates
Sandor, R and Ehrhardt, F and Basso, B and Bathia, A and Bellocchi, G and Brilli, L and Cardenas, L and De Antoni Migliorati, M and Doltra Bregon, J and Doris, C and Doro, L and Fitton, N and Giacomini, S and Grace, P and Grant, B and Harrison, M and Jones, S and Kirschbaum, M and Klumpp, K and Laville, P and Leonard, J and Liebig, M and Lieffering, M and Martin, R and McAuliffe, R and Meier, E and Merbold, L and Moore, A and Myrgiotis, V and Pattey, E and Qing, Z and Recous, S and Rolinski, S and Sharp, J and Silvia Massad, R and Smith, P and Smith, W and Snow, V and Soussana, J-F, Global research alliance on agricultural greenhouse gases - benchmark and ensemble crop and grassland model estimates, Modelling Grassland-Livestock Systems under Climate Change Conference 2016, 15-16 June 2016, Potsdam, Germany (2016) [Conference Extract]
The Soil Carbon and Nitrogen Cycling Cross-cutting Group (Soil CN group) of the Global Research Alliance on Agricultural Greenhouse gases (GRA) promotes a coordinated activity across multiple international projects (e.g. CN MIP and Models4Pastures of the FACCE-JPI https://www.faccejpi.com) to benchmark and compare simulation models that simulate GHG emissions from arable crop and grassland systems. Ten long-term experimental sites are studied covering a variety of climatic and geographic conditions worldwide (Australia, Brazil, Canada, France, India, New Zealand, Switzerland, United Kingdom and United States). Twenty-four process-based models of different complexity have contributed to the modelling exercise in different stages, each with access to gradually more detailed data to run and evaluate models of a multi-stage protocol. We present a comparison of model estimates of production (e.g. grain yield, gross primary production, above-ground net primary production, grassland grazing or defoliation) and vegetation (e.g. leaf area index) outputs, as well as GHG emissions (e.g. ecosystem respiration, nitrous oxide, enteric methane) from individual models to the multi-model ensemble. We found substantial discrepancies across different models, indicating considerable uncertainties regarding the simulation of crop and grassland processes. We show that uncertainties are considerably reduced after calibration with detailed production and phenology data. The multi-model approach also allowed for improved performance, according to relative root mean square error and relative bias performance metrics. Calibrated models provide a reliable basis for testing mitigation options at the studied sites.
climate change, modelling, model comparison, crop, grassland, model ensemble, greenhouse gas emissions, mitigation, adaptation