Which temperature to simulate foliar epidemics × crop architecture interactions ?

Air temperature measured by weather stations is commonly used in epidemiological models to forecast the effect of temperature on the development of foliar fungal pathogens. However, leaf temperature is the temperature actually perceived by such pathogens. The leaf temperature depends on the leaf energy budget (e.g. air temperature, radiation, wind, transpiration, etc.), which itself strongly depends on the crop architecture (e.g. leaf position, leaf angle, leaf area density). Consequently, differences between air and leaf temperatures vary spatially between canopies with contrasted architectures and between leaves within a given architecture, especially between leaf layers, as well as temporally throughout the course of the epidemic (seasonal variations). We already characterized the effect of leaf temperature on the latent period of the fungus Mycosphaerella graminicola infecting wheat leaves. In this simulation study, we aimed at estimating whether the use of either air or leaf temperature as input data influences the development of the infectious cycle of M. graminicola within contrasted wheat canopy architectures. Various weather conditions were generated using actual weather data. For each leaf layer, leaf temperature was calculated using the one-dimensional Soil-Vegetation-Atmosphere Transfer model CUPID for different canopy architectures. From the thermal performance curves of the latent period established in the aforementioned study, the pathochron, defined as the number of leaves emerging per latent period, was calculated, using either air temperature or leaf temperature as input data. At the leaf scale, the type of temperature used as input data modified the pathochron, which could generate various disease dynamics into the canopy. Our results highlighted the weather conditions for which it is necessary to take into account leaf temperature rather than air temperature to estimate accurately the development of M. graminicola. Our simulation method could be applied to other foliar fungal pathogens. In a further step, we will use future climatic scenarios to explore the impact of climate change on disease dynamics. In a longer term, the integration of these findings to more elaborated epidemiological models is expected to improve their forecasting accuracy