Predicting crop yield at an early growth stage is a challenge, especially under stress conditions. Physiological traits such as photosynthesis and stomatal conductance are good yield predictors. Yet, they can be difficult to measure in a comparative way due to their dynamic nature and their interactions with ambient conditions. In recent years, many phenotyping platforms have become available, enabling high-throughput screening of many traits. We the PlantArray  gravimetric phenotyping system combined with soil, plant and atmospheric sensors to establish a framework of physiological quantitative traits (QPTs), creating high-resolution physiological-response profiles (PRPs) of young tomato plants under adequate-water conditions and in multiple drought-recovery scenarios and compared those profiles with yield parameters observed in a parallel field study. We used seven different genotypes of tomato grafted in reciprocal combinations to create a large variety of yield parameters. Among the more than 100 different QPTs and PRPs we collected, the best predictors of field yield were canopy stomatal conductance, transpiration rate (normalized to whole-plant weight) and cumulative transpiration, which was strongly positively correlated with the field-grown plants’ weights (R2: 0.94, P< 0.001), total yield (R2: 0.77, P< 0.05) and red-fruit weights (R2: 0.81, P< 0.05). Under drought stress, it became difficult to predict the yield based on the physiological traits, most likely due to the plant's strong defensive response, the pot effect and difficulties in replicating the stress treatments. Using a feedback irrigation algorithm, implemented separately for each plant, we overcame those challenges and created highly repeatable, standardized stress-treatment protocols, which enabled us to maintain near-field drought scenarios. This allowed us to develop a resilience-profile predictor with a high correlation to yield of drought-affected plants in the field. Our results suggest that QPTs and PRPs could be useful in pre-breeding and the development of high-yield tomato genotypes under dynamic environmental conditions.
