ORCHIDEE-CROP, a new process based Agro- Land Surface Model: model description and evaluation over Europe

The responses of crop plants to changing climate and CO2 could have large effects on food production, and impact carbon, water and energy fluxes, causing feedbacks to climate. To simulate the responses of temperate crops to changing climate and CO2, accounting for the specific phenology of crops mediated by management practice, we develop a process-oriented terrestrial biogeochemical model ORCHIDEE-CROP, which integrates a generic crop phenology and harvest module and a very simple parameterization of nitrogen fertilization, into the DGVM ORCHIDEEv196, in order to simulate biophysical and biogeochemical interactions in croplands, as well as plant productivity and harvested yield. The model is applicable for a range of temperate crops, but it is tested here for maize and winter wheat, with the phenological parameterizations of two European varieties. We evaluate this new model against eddy covariance and biometric measurements at 7 winter wheat and maize sites in Europe. The specific ecosystem variables used in the evaluation are Net Ecosystem Exchange (NEE), latent heat and sensible heat fluxes. Site additional measurements of leaf area index (LAI), aboveground biomass and yield are used as well. Evaluation results reveal that ORCHIDEE-CROP reproduces the observed timing of crop development stages and the amplitude of pertaining LAI changes in contrast to ORCHIDEEv196, that by default applies to crops the same phenology of grass. A near-halving of the root mean square error of LAI from 2.38±0.77 m2 m-2 to 1.08±0.34 m2 m-2 is obtained between ORCHIDEEv196 and ORCHIDEE-CROP across the 7 study sites. Improved crop phenology and carbon allocation lead to a general good match between modelled and observed aboveground biomass [with a normalized root mean squared error (NRMSE) of 11.0%-54.2%], crop yield, as well as of carbon and energy fluxes with NRMSE of ~9.0-20.1% and ~9.4-22.3% for NEE, and sensible and latent heat fluxes, respectively. The model data misfits for energy fluxes are within uncertainties of the measurements, which show an incomplete energy balance closure within the range 80.6-86.3%. The remaining discrepancies between modelled and observed LAI and other variables at specific sites are partly attributable to unrealistic representation of human management. In addition, ORCHIDEE-CROP is shown to have the ability to capture the spatial gradients of both biogeochemical (gross primary productivity and NEE) and biophysical (sensible and latent heat fluxes) variables across the sites in Europe, an important requirement for future spatially explicit simulations. Further improvement of the model with an explicit parameterization of nutrition dynamics and management, is expected to improve its predictive ability to simulate croplands in an Earth System Model