Genetic variation in the transpiration response to evaporative demand has been mostly studied in controlled environments, individual plants, and small genetic panels. Moving outdoors in field-like situations raises questions about what timeframe to measure transpiration, which metric to represent evaporative demand (vapor pressure deficit, VPD or reference evapotranspiration, ET_ref), which model to describe transpiration responses (linear slopes or segmented breakpoints), which plant spacing (individual plants, canopies), and which experimental setup for heritable genetic variation.

Transpiration responses to evaporative demand were phenotyped in 467 sorghum and pearl millet inbred lines in three contrasting platforms (high-tech, low-tech, and lysimetric systems) outdoors. Across all experiments, ET_ref consistently explained more variation in transpiration response than VPD. In both species, transpiration responses to increasing ET_ref were best described by linear relationships. Genotype-specific regression slopes were retained as trait estimates.

In pearl millet, transpiration response slopes and transpiration efficiency exhibited significant genetic variation with high heritability (0.61-0.73) across platforms. In sorghum, trait heritability depended on the platform, with higher heritability in the low-tech. Transpiration efficiency was positively associated with transpiration response slopes in both species, while relationships with root traits revealed species-specific patterns of water use and uptake.

Overall, ET_ref was a reliable descriptor of evaporative demand, and linear modelling captured transpiration responses to ET_ref well outdoors. The low-tech platform enabled reliable phenotyping of transpiration response to ET_ref of individual plants in large germplasm panels while both the high-tech and lysimeter platforms were reliable to measure transpiration response to ET_ref in canopies.