Grain yield observations constrain cropland CO2 fluxes over 1 Europe

Carbon exchange over croplands plays an important role in the European carbon cycle over daily to seasonal time scales. Not only do crops occupy one eighth of the global land area, but their photosynthesis and respiration are large and aect CO2 mole fractions at nearly every atmospheric CO2 monitoring site. A better description of this crop carbon exchange in terrestrial biosphere models { most of which currently treat crops as unmanaged grasslands { could strongly improve their calculated uxes. Available longterm observations of crop yield, harvest, and cultivated area allow such improvements, when combined with the new crop-specic modeling framework we present here. In this framework we model gross carbon uxes of the major European crops on a 25 x 25 km grid and daily time-step, while aiming to make calculated seasonal grain yield agree with observations. For each crop species and region of Europe, this follows a two-step procedure. In the rst step, we calculate crop growth over the full growing season with the process-based WOrld FOod STudies (WOFOST) agricultural crop growth model, which results in a simulated crop yield. Simulated yields are optimized by minimizing their dierence to regional crop yield observations from the Statistical Oce of the European Union (EUROSTAT) by estimating one yearly regional scaling parameter for each crop species, the so-called \yield gap factor". In a second step, we run our WOFOST model for the full European 25km2 gridded domain using the optimized yield-gap factors for each crop species and region, to make wall-to-wall carbon exchange uxes that also take into account the local variations of weather, soil properties, and sowing calendar. We combine the resulting GPP and Raut uxes with a simple soil respiration expression to obtain the gridded total ecosystem respiration (TER) and net ecosystem exchange (NEE). We assess our model's ability to represent the seasonal GPP, TER and NEE uxes using 40 site-years of observations at 7 European FluxNet cropland sites and compare it with cropland carbon uxes produced by a typical terrestrial biosphere model used widely for European carbon cycle studies. We conclude that our new model framework provides a more detailed, realistic, and strongly observation-driven estimate of carbon exchange over European croplands. Its products will be made available to the scientic community through the ICOS Carbon Portal, and serve as a new cropland component in the CarbonTracker Europe inverse model ux estimates.