Key points are not available for this paper at this time.
Taro, a neglected and underutilised crop, is identified as a promising future smart crop with significant potential to enhance food security while diversifying cropping systems. Despite its reported resilience to climate change impacts, taro’s physiological responses to water deficits, including increased leaf temperature and reduced stomatal conductance, can significantly reduce its productivity and yield. These responses vary with growth stages; limited canopy cover at emergence may expose soil to higher temperatures and intensify early stress, while peak transpiration during the vegetative stage increases crop sensitivity to water deficits. This study addresses the practical challenge of monitoring of taro crop water status throughout the growing season in data-scarce smallholder systems, where irrigation infrastructure is lacking and traditional field-based assessments are often infeasible. Advances in thermal remote sensing technologies and UAVs enable non-invasive, high-resolution monitoring of crop water status. This study presents a novel multi-temporal framework integrating UAV-acquired thermal and multispectral sensing with a deep neural network algorithm to estimate the stomatal conductance and leaf temperature of smallholder taro at emergence, vegetative and maturity stages. Findings showed that thermal data, together with derived thermal indices, are critical predictors for both stomatal conductance and leaf temperature across the growth stages. The vegetative stage exhibited the highest prediction accuracies for stomatal conductance (R2 of 0.96 and rRMSE of 12.86%) and leaf temperature (R2 of 0.95 and rRMSE of 1.11%). This study contributes an innovative, scalable solution for high-throughput water-stress monitoring in under-researched NUS crops, enabling informed decision-making for sustainable agriculture in resource-limited smallholder systems.
Ndlovu et al. (Thu,) studied this question.