Fig. 2
Comparison of different estimates of FLUC for two of the focus countries in ESA-CCI RECCAP2, Germany and France (different rows). The left panels show annual time-series of FLUC simulated by two DGVMs (OCN [107] and ORCHIDEE-MICT [41]) based on two LULCC datasets: oneforced with LUH2 GCB2021 (blue lines) and HILDA + (yellow lines) for the period 1960–2020. These are compared tothe ensemble of bookkeeping models (black line for the mean and grey shades for the range of the models), the respective NGHGIs for each country (black line with triangle markers) and FAO (open circles). The right panels show mean decadal fluxes for individual models (OCN in filled bars, ORCHIDEE-MICT in hatched bars and BLUE-HILDA + in open bars) forced with the two LULCC datasets (blue colours for LUH2 GCB2021 and yellow for HILDA +). The markers show the corresponding values estimated by BK models (squares with vertical lines showing model spread), NGHGIs (triangles) and FAO (open circles). To estimate FLUC with DGVMs we followed the commonly used approach in Global Carbon Budgets [31, 70, 83]: we run two simulations forced with changing CO2 and climate, but one with fixed LULCC distribution (in this case in 1950) and another with changing LULCC fields. The difference between the two allows estimating the effect of LULCC on the simulated carbon fluxes. Information about the respective LULCC datasets can be found in the Section on Land cover EO datasets and more details about the forcing datasets and model simulations is provided in Additional file 1