Crop Models

Crop models are simplified mathematical representations of crop systems that can be classified as mechanistic or empirical. The former uses the scientific knowledge of the system behavior to propose mathematical equations for the crop response to the environment. The latter are not constrained by scientific principles and use regression or curve fitting to observational data to predict system outputs. Crop models have been extensively applied in agricultural research to help interpret experimental findings, separate casualty from causality, synthesize research understanding, facilitate the development of preliminary hypotheses, and optimize crop performance across environments. Beyond research applications, crop models have been used in decision-making. Decision Support Systems are computer software tools that integrate external information (soil, weather, sensors) with crop models to assist farmers, stakeholders, and policymakers in making informed decisions. Crop models have undergone continuous development since their inception in the 1960s. Starting in the 2000s, the integration of machine learning techniques into crop modelling has shown success, particularly in addressing identification and classification challenges within the mechanization of harvest and postharvest operations, as well as crop protection. Nevertheless, machine learning should be used with caution for yield prediction.