Using crop models, a decline factor, and a “multi-model” approach to estimate sugarcane yield compared to on-farm data

Sugarcane is an important crop in Brazilian agribusiness due to its diversified use. Crop forecast models are important tools for planning and making decisions regarding crop management. These models can be simple or complex, and choosing them will depend on the knowledge level of those using them. Thus, this study aimed to compare different methods for estimating sugarcane yield in three crop cycles. Data collection occurred in a sugarcane field in the municipality of Santo Antônio de Goiás, Brazil. The sugarcane variety evaluated was CTC-04. This variety was cultivated under dryland conditions, in cane plant, ratoon 1, and ratoon 2 cycles. Agrometeorological, biometric, and crop yield data were analyzed. Five crop models were used to estimate sugarcane yield: (i) FAO-Agroecological Zone (AEZ), (ii) agrometeorological-spectral (AEZs), (iii) Monteith (M), (iv) Scarpari (S), and (v) Martins and Landell (ML). Models AEZ, AEZs, M, and S showed average yield differences of about 15%, with the largest difference recorded by the ML model (39%). All models detected yield decline as a function of the number of harvests (kdec =  − 0.70). The multi-model approach reduced the differences between estimated and actual values, especially for the combinations “AEZ + AEZs” and “AEZ + AEZs + M.” The present findings contribute to the investigation of different models with the potential to estimate sugarcane productivity.