Evaluating the use of aerially captured spectral data for characterising drought-tolerance traits in sugarcane

Drought is a frequent phenomenon in sugarcane production, and future climate change could cause more frequent and severe dry spells. Drought tolerance is therefore highly desirable in sugarcane genotypes. Stomatal conductance (gs) is a key drought tolerance trait that regulates carbon fixation and crop water use, influencing canopy temperature (Tc) through latent heat absorption of evaporating water. Remote sensing of crop reflectance is a potential technique for estimating Tc and gs in sugarcane crops. This technique needs to be tested before it can be considered for use in sugarcane breeding. The aim of the study was to assess whether aerially captured spectral data can be used to infer genotypic differences in stomatal conductance (gs) in response to water stress. Two genotypes, NCo376 and N19, were grown at a rainshelter facility at Mount Edgecombe, South Africa, under well-watered (WW) and water deficit (WD) conditions. Following canopy establishment, water was withheld in the WD plots during two events (lasting four and eleven weeks respectively), with a recovery period of three weeks in between. Ground-based measurements included soil water content (capacitance probes), canopy cover (FIPAR, with AccuPAR LP-80 ceptometer), gs and leaf temperature (Tleaf, with CIRAS-3 photosynthesis system and Optris MS Plus, respectively). Spectral data in the visible, near-infrared (Sentera NDVI sensor) and thermal (FLIR Vue Pro) bands were obtained on six occasions using a DJI Phantom 4 drone. WD values of gs was expressed relative to WW values for each genotype (gs*). These were compared to differences in Tc (derived from aerially captured thermal data) between the WD and WW treatments for each genotype (?Tc). The study found a fair correlation between gs* and ?Tc (R 2 =0.40), although obvious outliers were present. Strong correlations were found between gs and Tc, and between Tleaf and Tc when WD treatments were moderately or severely stressed, but not when they were unstressed or mildly stressed. The study also found a good relationship between FIPAR and the normalized difference vegetation index (NDVI, R 2 =0.80). Results suggest that aerially captured thermal data may hold promise for estimating genotypic differences in gs in response to water stress. This, when combined with remotely sensed FIPAR, could be used to estimate canopy conductance which has the potential as a selection trait in sugarcane breeding. Further testing on multi-environment trials with more genotypes is needed to further develop the technique for practical use. Key words Stomatal conductance, canopy temperature, water stress, drought tolerance, soil water content, canopy cover