Site-specific non-water-stressed and water-stressed baselines to calculate CWSI to schedule irrigation for furrow-irrigated sugarcane: a new approach

Water stress in irrigated broadacre crops can be detected as an elevation in canopy temperature using canopy-temperature sensors. A commonly used index that reflects water stress is the crop water-stress index (CWSI), defined as a comparison of measured canopy-air temperature difference with a non-water-stressed baseline (lower baseline, (T c – T a) LB) and a water-stressed baseline (upper baseline, (T c – T a) UB). CWSI can also be used to calculate the fraction of soil-moisture depletion (fDEP) to determine irrigation depth required. Currently, the default values of the lower and upper baselines used in the CWSI calculation are defined from studies conducted in France. This assumption may lead to inaccuracies in determining the CWSI as the baselines may not be transferrable to other locations or crops. This paper presents an investigation of the impact of site-specific upper and lower baselines on the CWSI and fraction of depletion in the root zone (fDEP) for furrow-irrigated sugarcane in Ayr, Queensland, Australia during the irrigation season of 2017–2018. CWSI values were calculated from canopy temperatures, air temperatures and vapor-pressure deficits (VPD) throughout the season and daily maximum CWSI values were linked to the fraction of soil- moisture depletion (fDEP) through the water-stress coefficient (ks). Results suggested that baselines developed on-site performed better to schedule irrigation than did baselines developed off-site. The lower baseline (LB) equation for sugarcane was developed using canopy temperature measured from a fully irrigated field as: (T c – T a) LB = –1.8271*VPD – 4.7854, with a coefficient of determination of R 2 = 0.9511, as compared to published equations with coefficients of determination of R 2 = 0.80 and 0.67. The daily maximum CWSI values obtained by using the lower baseline reported in literature were higher than the site-specific lower baseline with a mean absolute error of 0.32, indicating the need of irrigation for all days between two irrigation events. The mean absolute error in predicting root-zone depletion (Dr) by using CWSI values with the lower baseline from the literature were higher with a mean absolute error of 7.92 mm and 10.89 mm. Updating on-site baselines after each irrigation event throughout the irrigation season instead of using a single baseline developed from one irrigation event improved results further. Key words Remote sensing, canopy temperature, CWSI, water-stress coefficient, soil-moisture depletion