Is CropSyst adequate for management-oriented simulation of growth and yield of processing tomato ?

Abstract: The model CropSyst has proven useful for management-oriented simulations of growth and yield of cereals and other field crops, but no scientific information is available with reference to processing tomato. The aim of this paper was to parameterise and validate the crop module of CropSyst for the simulation of potential fruit production in processing transplanted tomato (Lycopersicon esculentum Mill.). Parameterisation and calibration were performed by using field data from an experiment carried out in 1997 in Central Italy. The same set of parameters was validated against five independent experiments, carried out on the same location in 1998, 1999, 2000, 2001 and 2002. The simulation of aerial biomass was always very good, with RRMSE values ranging from 7.5 to 13.4% and modelling efficiencies (EI) always above 0.976. The simulation of LAI was very good during the first part of growing season (up to 40-50 days after transplanting), while the decreasing trend in the final part of growing cycle was not always reliably simulated. Indeed, RRMSE for LAI ranged from 13.5 to 26.8% and EI ranged from 0.849 to 0.966. The differences between simulated and observed final fruit yield were below 10%, except in one year (18% in 2001), confirming the practical value of this model, for management and legislative purposes. For research purposes, it is confirmed that the simulation of dry matter partitioning is a crucial issue in vegetable crops such as tomato, wherein the growth of sources and sinks coexists for a main part of crop cycle.