This study is focused on assessing the impacts of different regional climate model targeted simulations performed at convection-permitting resolution (CPRCM) in the AgS crop model yield simulations, evaluating to what extent climate model uncertainty impacts the modeled yield—considering the spatial and temporal variability of crop yield simulations over central-south Brazil. The ensemble of CPRCMs has been produced as part of a Flagship Pilot Study (FPS-SESA) framework, endorsed by the Coordinated Regional Climate Downscaling Experiment (CORDEX). The AgS simulated crop yield exhibited significant differences, in both space and time, among the simulations driven by the different CPRCMs as well as when compared with the simulations driven by observations. Rainfall showed the highest uncertainty in CPRCM simulations, particularly in its spatial variability, whereas modeled temperature and solar radiation were generally more accurate and exhibited smaller spatial and temporal differences. The results evidenced the need for multi-model simulations to account for different uncertainty, from different climate models and climate models parameterizations, in crop yield estimations. Inter-institutional collaboration and coordinated science are key aspects to address these end-to-end studies in South America, since there is no single institution able to produce such CPRCM-CropModels ensembles.

Researchers using crop models have been devising new roles for data and crop modeling based on the former’s increased availability and the new techniques developed for the latter. From the various available techniques, modeling may be tackled by data-driven methods or through a process-based approach. Process-based or mechanistic models may nonetheless take advantage of real-time observations through data assimilation. And while this approach has been widely used for field crops, this is not the case for crops grown in protected environments. We present a case study of data assimilation in a protected environment, capturing tomato growth data from different sources. We updated growth estimates of the Reduced State TOMGRO model, by assimilating observational data obtained through the continuous monitoring of plant mass and images captured by low-cost cameras, using the Unscented Kalman Filter and the Ensemble Kalman Filter. Since these techniques had not been used yet in the protected cultivation of tomatoes, it was necessary to develop the observation models as well, establishing the relationship between the observed variables and the ones estimated by the process-based model. The employed measurements, i.e., area of organs observed in pictures and plant-water mass, seemed suitable for tracking plant growth and for obtaining good approximations of the state variables estimated by the model. However, the quality of observations and of observation models was crucial for good performance of the assimilation techniques. As with other crops, it was not the case that assimilating one observation was useful for improving the value of others, including yield. We also observed that the assimilation performed better than calibrated models when there was a need to adjust the estimates to growth disturbances and that when filters lead to better yield estimates, continuous observations may not be required. There are then several steps and decisions that should be considered when bringing the idea from its application in field crops to protected environments and more studies are required to better determine the best approach.