Constraints from atmospheric CO2 and satellite-based vegetation activity observations on current land carbon cycle trends.

Terrestrial ecosystem models used for Earth system modelling show a significant divergence in future patterns of ecosystem processes, in particular carbon exchanges, despite a seemingly common behaviour for the contemporary period. An in-depth evaluation of these models is hence of high importance to achieve a better understanding of the reasons for this disagreement. Here, we develop an extension for existing benchmarking systems by making use of the complementary information contained in the observational records of atmospheric CO₂ and remotely-sensed vegetation activity to provide a firm set of diagnostics of ecosystem responses to climate variability in the last 30 yr at different temporal and spatial scales. The selection of observational characteristics (traits) specifically considers the robustness of information given the uncertainties in both data and evaluation analysis. In addition, we provide a baseline benchmark, a minimum test that the model under consideration has to pass, to provide a more objective, quantitative evaluation framework. The benchmarking strategy can be used for any land surface model, either driven by observed meteorology or coupled to a climate model. We apply this framework to evaluate the offline version of the MPI-Earth system model's land surface scheme JSBACH. We demonstrate that the complementary use of atmospheric CO₂ and satellite based vegetation activity data allows to pinpoint specific model failures that would not be possible by the sole use of atmospheric CO₂ observations. [ABSTRACT FROM AUTHOR]