Assessing neglected and underutilised taro crop water status using physiological indicators and UAV multi-modal thermal-multispectral data

Abstract

Abstract Purpose Taro (Colocasia esculenta (L)) , a neglected and underutilized crop species (NUS), holds great potential as a future smart crop that can thrive under climate variability and change, hence sustaining food security. While taro exhibits tolerance to drought conditions, variations in physiological attributes such as leaf temperature that rises under water stress and the associated stomatal closure that is initiated to conserve water, compromise crop productivity and overall yield. Therefore, monitoring taro crop physiological indicators of water status allows for the implementation of timely interventions and targeted adaption strategies to mitigate the effects of water deficit on taro crop productivity. Methods Unmanned Aerial Vehicles (UAV), integrated with high-resolution thermal sensors, provide valuable platform for generating near-real-time spatially explicit information suitable for assessing taro crop water status physiological indicators at farm scale. Hence, this study sought to evaluate the utility of UAV multi-modal thermal remote sensing and deep neural network techniques to estimate the equivalent water thickness, fuel moisture content, stomatal conductance, canopy temperature, and the chlorophyll content of smallholder taro crops. Results Findings showed that the multi-modal variable method achieves higher estimation accuracies in comparison to a single-modal technique, achieving R 2 values greater than 0.91 and rRSME values less than 14.15% of equivalent water thickness, fuel moisture content, stomatal conductance, canopy temperature, and chlorophyll content. Additionally, the results illustrated that the thermal wavebands and derived thermal indices are the most influential variables in estimating stomatal conductance and leaf temperature, yielding R 2 of 0.96 and 0.95, respectively. Conclusion These research findings underscore the applicability of UAV-acquired thermal remote sensing in providing rapid and robust spatially explicit information on smallholder taro crop water status for ensuring crop productivity and developing early warning systems of water stress. These findings serve as a stepping stone towards advancing agricultural monitoring frameworks and integrating NUS, such as taro, into traditional farming.

Plant phenotyping relevance

UAV熱・マルチスペクトルデータと深層学習により、タロイモの水分状態、生理形質、クロロフィルなどを推定し、精度も評価しているため、植物フェノタイピング手法の応用・技術評価が中心である。

Code and data availability

The paper's UAV thermal-multispectral taro phenotyping data and physiological measurements are explicitly not public; the authors state they are available only on request due to funder restrictions. No author code, models, or public repository URLs are provided.

No evidence-backed public reproduction asset is currently recorded.