Detection of Potential Date Palm Stress in an Arid Environment Using Time-Series Sentinel-2 Imagery and Unsupervised Classification

Date palms are among the most important agricultural crops in the United Arab Emirates (UAE), where water scarcity, high temperatures, soil salinity, and irrigation challenges can adversely affect crop health and productivity. Timely identification of vegetation stress is essential for supporting sustainable agricultural management in arid environments. This study investigates the use of multi-temporal Sentinel-2 imagery and vegetation indices to detect potential stress in date palm plantations located at Al Waha Al Aamerah Farm in Al Ain, UAE.

Sentinel-2 Level-2A images acquired during the 2025 growing season were processed and analyzed. Five vegetation indices representing vegetation vigor and moisture conditions were calculated, including the Normalized Difference Vegetation Index (NDVI), Normalized Difference Red Edge Index (NDRE), Normalized Difference Moisture Index (NDMI), Moisture Stress Index (MSI), and Soil-Adjusted Vegetation Index (SAVI). For each index, mean, standard deviation, and temporal slope metrics were derived and combined into a 15-band feature stack. A vegetation mask based on NDVI mean values greater than 0.18 was applied to exclude bare soil and non-vegetated areas. Unsupervised K-means clustering was then used to classify vegetation condition into three categories: Potential Stress, Moderate Vigor, and High Vigor.

A total of 907 vegetation points were extracted and classified. The results indicated that 51.71% of the vegetation was categorized as Potential Stress, 17.31% as Moderate Vigor, and 30.98% as High Vigor. The generated stress map revealed clear spatial variability across the plantation and highlighted areas requiring further investigation. The study demonstrates that time-series Sentinel-2 imagery combined with vegetation indices and machine learning techniques provides an effective, cost-efficient approach for monitoring date palm health and supporting precision agriculture in arid regions.