全球变化知识资源中心

Rice is one of the excellences of Italian agriculture, so that modern research techniques are mobilised to support farmers and local stakeholders in the main harvested areas in Northern Italy. Crop modelling is among the most prominent tools, as proved by the model-based services developed for yield forecasting and blast disease risk alert. Here we propose a methodology to obtain cultivar-specific parameters sets to improve the current crop modelling capabilities, using field observations collected on the main Italian rice cultivars. The definition of cultivar-specific parameters sets was based on the calibration of the crop model WOFOST_GT using long-term (2004-2015) experimental data collected in two sites (167 field trials). Our results proved a satisfactory accuracy in reproducing field observations of heading (average RRMSE = 5.9% +/- 1.3%) and maturity date (22.6% +/- 4.7%), canopy height (10% +/- 1.4%), and yield (11.1% +/- 4.1%) across cultivars and years. The analysis of the cultivar-specific parameters sets allowed highlighting the inter-relationships between model parameters, which revealed similarities with standard classifications of grain shape, phenological development and stem length. Our study is meant to deepen our modelling knowledge of the Italian rice varietal landscape, thus fostering future applications at different spatial and temporal resolutions.