Just had a wheat fertiliser rate simulation on sandy loam of Missouri floodplain. 600 mm in-crop rainfall, drainage around 200 to 300 mm in-crop. 0 N rate observed yield 2317 kg/ha and simulated 477. Couldn't get near observed yield by adjusting soil characteristics until I found an old comment in the Silwat code that the solute flux and flow efficiencies had been changed from the original Ceres-maize 0.5 to 1.0. Changing the now externalised efficiencies in the soil.xml to 0.5 brought the simulated yield to 2344. It was similar with the other treatments on this soil and climate.
There needs to be a review of the solute flux and flow efficiency values with testing and validation against data to establish appropriate values for these calibration parameters.