Non-water-stressed canopy temperature for crop water stress index estimation in soybean and maize: growth-stage-specific measurement timing and dynamic modeling
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Efficient irrigation management in water-limited cropping systems depends on timely identification of crop water stress, yet thermal methods remain limited by the difficulty of establishing a representative non-water-stressed canopy-temperature reference. This thesis evaluated two complementary, growth-stage-aware approaches for strengthening that reference: selecting appropriate measurement times for soybean non-water-stressed baselines and dynamically predicting non-water-stressed canopy temperature in maize. Canopy-temperature and meteorological observations were collected during the 2018 to 2020 growing seasons at North Platte, Nebraska. For soybean, candidate measurement periods were compared within a Bayesian hierarchical framework using the relationship between the resulting crop water stress index estimates and yield. For maize, an energy-balance-informed, data-based mechanistic model represented non-water-stressed canopy temperature as a first-order continuous-time system driven by air temperature, solar radiation, and vapor pressure deficit. The selected soybean periods shifted from mid-to-late afternoon during vegetative development to midday or early afternoon during reproductive development. Relative to these selected periods, the conventional period from 2:00 to 3:00 p.m. increased posterior mean root mean square error by 3.3 to 43.0 percent. Baseline slopes and intercepts also varied among growth stages. In maize, model performance was strongest at the milk and dough stages, where mean absolute error ranged from 0.093 to 0.342 degrees Celsius and the coefficient of determination reached 0.958; performance was weaker at the V8, V9, and blister stages. Collectively, the results show that the non-water-stressed thermal reference changes with crop development. Accounting for growth stage in both measurement timing and dynamic prediction can improve the consistency of crop water stress index estimation and strengthen its use in irrigation management.