Adaptation of the process-based CSM-CROPGRO model to simulate the growth and development of industrial hemp for seed and fiber production

Abstract

Industrial hemp (Cannabis sativa L.) is a re-emerging crop in the United States with unique agronomic challenges that require location-specific studies and guidance. Digital farming tools, such as crop growth models, can facilitate this process by enabling a better understanding of the farming system. Crop growth models predict the growth and development of crops over time using weather, soil, management, and physiological parameters as inputs. The goal of this study was to develop a new hemp model in the Cropping System Model (CSM)-CROPGRO module in the Decision Support System for Agrotechnology Transfer (DSSAT). Experimental data spanning two cultivars, both grown over two seasons and two sites in Florida, were used for model calibration and evaluation. Model adaptations were made in (1) tissue composition and assimilate partitioning, (2) cardinal temperatures for different growth and development processes, and (3) leaf photosynthesis and senescence. The results show a good simulation of aboveground biomass (d = 0.91, root mean square error [RMSE] = 482 kg ha⁻¹), stem weight (d = 0.83, RMSE = 430 kg ha⁻¹), and time to flowering (+4 to −5 days), capturing the differences among cultivars and planting dates. A seasonal analysis using the adapted model showed the impact of variable planting dates on hemp phenology, biomass, and grain production. Future work should include a more detailed observation and mechanistic simulation of self-thinning and evaluation with data representing different production environments and cultivars. The CROPGRO-Hemp model will provide a basis for growers, researchers, and other stakeholders to systematically analyze hemp production systems in Florida and internationally.