Abstract
This study investigates the optimization of deficit irrigation scheduling in winter wheat (Triticum aestivum L.) within the semi-arid Kura-Aras Lowland of Azerbaijan using a combination of Sentinel-2 Normalized Difference Vegetation Index (NDVI) and ground-based canopy temperature measurements. Over two consecutive growing seasons (2021–2023), three irrigation treatments were evaluated: full irrigation (100% crop evapotranspiration, ETc), moderate deficit irrigation (75% ETc), and severe deficit irrigation (50% ETc), scheduled using a dynamic crop coefficient (Kc) derived from Sentinel-2 NDVI and Crop Water Stress Index (CWSI) calculated from canopy-to-air temperature differences. The results demonstrated that integrating remote sensing NDVI with real-time CWSI allowed for precise identification of water stress thresholds. The moderate deficit irrigation treatment (75% ETc) maintained yield levels comparable to full irrigation, achieving a non-significant yield reduction of only 6.4% while improving irrigation water use efficiency (IWUE) by 22.8%. In contrast, severe deficit irrigation significantly compromised grain yield and quality. The crop coefficient derived from Sentinel-2 (Kc-NDVI) showed a strong correlation (R² = 0.88) with ground-measured lysimeter values. This study demonstrates that combining satellite-derived vegetation indices with thermal canopy monitoring provides a robust, scalable framework for enhancing water productivity in water-scarce agricultural zones like the Kura-Aras basin.
