Using Leaf Temperature to Improve Simulation of Heat and Drought Stresses in a Biophysical Model

Abstract

Despite evidence that leaf temperatures can differ by several degrees from the air, crop simulation models are generally parameterised with air temperatures. Leaf energy budget is a process-based approach that can be used to link climate and physiological processes of plants, but this approach has rarely been used in crop modelling studies. In this study, a controlled environment experiment was used to validate the use of the leaf energy budget approach to calculate leaf temperature for perennial pasture species, and a modelling approach was developed utilising leaf temperature instead of air temperature to achieve a better representation of heat stress impacts on pasture growth in a biophysical model. The controlled environment experiment assessed the impact of two combined seven-day heat (control = 25/15 °C, day/night, moderate = 30/20 °C, day/night, and severe = 35/25 °C, day/night) and drought stresses (with seven-day recovery period between stress periods) on perennial ryegrass (Lolium perenne L.), cocksfoot (Dactylis glomerata L.), tall fescue (Festuca arundinacea Schreb.) and chicory (Cichorium intybus L.). The leaf temperature of each species was modelled by using leaf energy budget equation and validated with measured data. All species showed limited homeothermy with the slope of 0.88 (P < 0.05) suggesting that pasture plants can buffer temperature variations in their growing environment. The DairyMod biophysical model was used to simulate photosynthesis during each treatment, using both air and leaf temperatures, and the patterns were compared with measured data using a response ratio (effect size compared to the well-watered control). The effect size of moderate heat and well-watered treatment was very similar to the measured values (~0.65) when simulated using T leaf, while T air overestimated the consecutive heat stress impacts (0.4 and 0). These results were used to test the heat stress recovery function (Tsum) of perennial ryegrass in DairyMod, finding that recovery after heat stress was well reproduced when parameterized with T sum = 20, while T sum = 50 simulated a long lag phase. Long term pasture growth rate simulations under irrigated conditions in south eastern Australia using leaf temperatures predicted 6-34% and 14-126% higher pasture growth rates, respectively at Ellinbank and Dookie, during late spring and summer months compared to the simulations using air temperatures. This study demonstrated that the simulation of consecutive heat and/or drought stress impacts on pasture production, using DairyMod, can be improved by using leaf temperatures instead of air temperature.