An Integrated Forecast-Simulation System for Intelligent Preseason Farming Decision Support

Preseason farming decisions, including the choice between direct seeding and transplanting for crop planting, impact farm performance and resource efficiency, directly influencing agricultural productivity and farmers’ livelihood. Farmers, however, face a complex decision-making process when making these choices given the many scenarios and uncertainties they must navigate, including climate conditions, environmental variability, and expected yield. To help farmers navigate this complexity, this research presents an integrated decision support system that provides valuable foresight for making informed preseason decisions. It integrates machine-learning-based forecasting, crop evapotranspiration models, crop growth simulation, and uncertainty analysis. The system is demonstrated for preseason planning for sugar beet farming in northwestern Iran. The multivariate, multistation neural networks provide the most accurate forecasts, transplanting consistently improves crop yields, and irrigation emerges as a critical factor influencing yield variability. By projecting water requirements and farm performance across different scenarios and under uncertainty, the system empowers farmers with a comprehensive foresight and understanding of their options, enabling well-informed decision-making for the upcoming growing season.