外文摘要：Assessing plant water status accurately in both time and space is crucial for maintaining satisfactory crop yield and quality standards, especially in the face of a changing climate. Remote sensing technology offers a promising alternative to traditional in situ measurements for estimating stem water potential (psi stem). In this study, we carried out field measurements of psi stem in an irrigated olive orchard in southern Italy during the 2021 and 2022 seasons. Water status data were acquired at midday from 24 olive trees between June and October in both years. Reflectance data collected at the time of psi stem measurements were utilized to calculate vegetation indices (VIs). Employing machine learning techniques, various prediction models were developed by considering VIs and spectral bands as predictors. Before the analyses, both datasets were randomly split into training and testing datasets. Our findings reveal that the random forest model outperformed other models, providing a more accurate prediction of olive water status (R2 = 0.78). This is the first study in the literature integrating remote sensing and machine learning techniques for the prediction of olive water status in order to improve olive orchard irrigation management, offering a practical solution for estimating psi stem avoiding time-consuming and resource-intensive fieldwork.
外文关键词：Vegetation indices;Olive;satellite;modeling;Irrigation management;spectral bands
作者：Garofalo, Simone Pietro;Giannico, Vincenzo;Vivaldi, Gaetano Alessandro;Salcedo, Francisco Pedrero;Costanza, Leonardo;Ali, Salem Alhajj;Camposeo, Salvatore;Lopriore, Giuseppe
作者单位：Univ Bari Aldo Moro;CEBAS CSIC
期刊名称：AGRONOMY-BASEL
期刊影响因子：0.0
出版年份：2024
出版刊次：14（1）
原文传递申请：江苏省科技资源（工程技术文献）统筹服务平台