Calibration procedure for a potato crop growth model using information from across Europe

In the FertOrgaNic EU project, 3 years of field experiments with drip irrigation and fertigation were carried out at six different sites across Europe, involving seven different varieties of potato. The Daisy model, which simulates plant growth together with water and nitrogen dynamics, was used to simulate the field experiments. An initial potato parameterisation was generated from an independent dataset and was used for site-specific calibrations. At those sites where the same variety was used for all 3 years, the calibration of the initial potato model was based on the first 2 years using the last year for validation. From the site-specific parameterisations, shared traits were extracted into a common parameterisation. This resulted in a list of the parameters that six independent people found had the most impact on the simulations. This common parameterisation we argue is a valuable platform for adaptation of the Daisy model to new potato varieties or for the improvement of the existing parameter set. The procedure is then, as a starting point, to focus the calibration process on the recommended list of parameters to change. We demonstrate this approach by showing the procedure for recalibrating three varieties using all relevant data from the sites. We believe these new parameterisations to be more robust, because they indirectly were based on information from the six different sites. We claim that this procedure combines both local and specific modeller expertise in a way that results in more robust and general parameterisations than if the common parameterisation step had been skipped.