A software component implementing a library of models for the simulation of pre-harvest rice grain quality.

Despite the availability of a variety of models to simulate crop growth and development, few operational approaches have been developed to assess pre-harvest quality of agricultural productions as a function of the conditions actually explored by the crop during the season. This represents a clear gap of knowledge researchers are trying to fill, in light of the evidences of a climate change-driven decline in the nutritional
properties of important food crops. Rice represents the staple food for half of the world’s population, and this explains the noticeable interest in rice grain quality, because of the direct implications on the economic value of productions, on their market destination, and on food security issues. This paper presents a framework-independent .NET software library, i.e., UNIMI.CropQuality, implementing models to simulate various aspects of rice quality: amylose, protein, lipids and starch content, viscosity profile, chalkiness, cracking and head rice yield. Alternate approaches for the simulation of the same quality property are included, to allow users to select the most suitable for specific modelling studies. A case study is also presented where the library was linked to the WARM rice model and used to simulate head rice yield and the percentage incidence of cracked and milky white kernels (severely chalky) for two rice varieties in the main European rice district. RRMSE ranged between 4.33% and 6.47% for head rice yield, 21.88% and 32.18% for cracking percentage, 35.92% and 55.01% for milky white chalkiness; modelling
efficiency were always positive. The component, developed according to the state-of-the-art of component-oriented software development, is released with a Software Development Kit containing help and code documentation files, as well as sample applications showing how to use the library with different crop simulators.