High spatiotemporal resolution monitoring of crop water stress across the contiguous United States using Harmonized Landsat and Sentinel-2 data

Accurate and timely monitoring of crop water stress is essential for efficient agricultural water management, ultimately maintaining and improving crop productivity. While Landsat has been used for this purpose, its temporal resolution hampers timely detection of crop water stress. The recently released Harmonized Landsat and Sentinel-2 Version 2.0 dataset, which enables a higher-frequency time series of satellite observations (2–3 days, 30 m), offers a promising solution to this challenge. However, its potential for crop stress monitoring remained unexplored. In this study, we utilized 923 HLS satellite tiles to assess crop water stress across the contiguous United States (CONUS). Crop water stress was monitored by analyzing normalized difference moisture index (NDMI) time series through applying the Breaks For Additive Season and Trend Monitor (BFAST monitor) and random forest models. We used HLS data from 2016 to 2019 as the historical period, and data from 2020, a year marked by intense droughts, as the monitoring period. We used stratified random points interpreted from Standardized Precipitation Index based drought products to validate the crop water stress alerts. Our results show that HLS data enables near-real-time alerts of crop water stress with an overall accuracy of water stress of 74.0 % and kappa coefficient of 0.48. We mapped approximately 12.3 Mha of water-stressed crops across the CONUS from March to August 2020, identifying around 3.8 million crop water stress events. Among these events, nearly 41.8 % affected areas smaller than 0.5 ha. Major crop water stress events (≥ 5 ha) were the least frequent, making up 10.0 % of events, yet they dominated in terms of area, affecting 74.2 % of the total mapped extent. For temporal accuracy, the mean time lag of detected crop water stress across the CONUS using HLS data is approximately 9 days. Our detected crop water stress demonstrates the feasibility of HLS data for providing timely crop water stress monitoring at a national scale. This highlights the potential of HLS-based monitoring to inform precision irrigation and support sustainable agricultural water resource management. • Crop water stress was monitored across the contiguous United States using HLS data. • Approximately 12.3 Mha of water-stressed crops were mapped. • About 3.8 million crop water stress events were mapped using HLS data. • HLS data demonstrates potential for monitoring crop water stress at large scale. • The mean time lag of detected crop water stress was about 9 days.