Transition toward increased maize water productivity in the Hetao Irrigation District in northwest China

The dynamic changes in regional crop water productivity (WP) have been neglected before, while it is critical for sustainable water resources management in arid irrigation districts. In this study, we applied a remote sensing-based model to simulate maize evapotranspiration (ET), yield and WP across the Hetao Irrigation District (HID) in northwest China from 2001 to 2024. The model outperformed the existing MODIS and PML_V2 products, improving R² (fitting slope) of ET and gross primary productivity estimations from 0.82 to 0.96 (0.37–0.99) and 0.64–0.92 (0.18–0.87), respectively; meanwhile it achieved acceptable yield estimates that strongly correlated with years of county-level statistics and field observations (R 2 =0.98, relative RMSE<20%). Subsequently, the temporal analysis revealed a pronounced turning point around 2017. Prior to this, ET, yield, and WP remained generally stable; thereafter, all three increased significantly, with respective annual rates of 9.30 mm, 332.71 kg/ha, and 0.03 kg/m 3 . The widespread adoption of water-saving irrigation technologies, together with enhanced soil water–salt regulation, played an important role in driving the increases in maize yield and WP. During the past two decades, the maize planting area expanded about 5.24 times, resulting in a significant increase in total seasonal ET (from 20.29 ×10 7 –117.98 ×10 7 m 3 ) and total yield (from 39.52 ×10 7 –216.72 ×10 7 kg). The spatial analysis revealed a pronounced heterogeneity of maize WP, ranging from 1.44 kg/m 3 in the WLBH sub-region to 1.71 kg/m 3 in the JFZ sub-region. Logarithmic regression decomposition demonstrated that yield was the dominant driver of WP, accounting for 90.59% for the entire HID. The responses of WP to ET exhibited significant spatial heterogeneity and nonlinearity such that beyond a threshold increasing ET no longer improved WP. This study provides a robust framework for quantifying the spatiotemporal dynamics of crop water use and productivity, offering scientific guidance for differentiated management of water resources in arid irrigated districts.