Modeling perennial sugarcane growth with the STICS soil-crop model in contrasted environments: Assessment of accuracy and robustness

Sugarcane holds significant economic importance due to its role in bioenergy and sugar production worldwide. However, as global demand increases, research is needed to identify the best management practices for sustainable yield improvement, and modeling approaches can be helpful. Our study aimed to parameterize and validate the STICS soil-crop model for sugarcane growth in an extensive range of environments. This model is not sugarcane specific but allows for the simulation of all the components of the carbon and nitrogen cycles and is useful for many research studies. The parameterization dataset included 20 crop cycles with Reunion Island varieties (mainly R570), and the validation dataset included 62 crop cycles with R570 variety and 158 crop cycles with 30 international varieties. The model was parameterized with belowground (root length, biomass, and perennial reserve) and aboveground variables, including leaf area index, intercepted radiation, and biomass components (stem, leaves). Validation was performed on aboveground variables. The model satisfactorily simulated all aboveground variables except for leaf area and mass, which were less accurately simulated. Despite being parameterized only on Reunion Island varieties, the model was robust enough to simulate some aboveground biomass at harvest for many international varieties. Additionally, the climatic domain of validity of the model was evaluated, and error on biomass at harvest increased for both high and low-yielded environments (high or low water availability and temperature), indicating that future improvements could be carried out in a complementary work. The STICS model can be used with good robustness to simulate sugarcane growth under similar varieties and pedoclimatic conditions, allowing exploration of management strategies that can boost productivity while maintaining the sustainability of cropping systems.