Quantifying water productivity and nitrogen uptake of maize under water and nitrogen stress in arid Northwest China

Water and nitrogen (N) are critical determinants of crop water productivity (WP) and N uptake (Nuptake), and are also key factors in crop modelling. Through a two-year field experiment with three irrigation levels (full irrigation (W1), moderate water stress (W0.75), and severe water stress (W0.5)), three N application rates (150, 75, and 0 kg N ha⁻¹, i.e., N150, N75, and N0) and three replications for each treatment in Northwest China, we quantified the effects of water and N stress on the growth, yield, and water and N use efficiency (WP and NUE) of hybrid seed maize (Zea mays L. cv. Tianruifeng). Furthermore, nonlinear dynamic water productivity (WP) models and water (ET)-based Nuptake models were developed. The results showed that water stress had a larger impact than N stress on the growth and yield of maize, with water stress reducing yield (Y), final aboveground biomass (B), and N uptake by 36.5%, 15.4%, and 25.7% on average, respectively, and N stress decreasing them by 18.7%, 12.3%, and 13.6%, respectively. Biomass-based water productivity (WPB-ET) and yield-based water productivity (WPY-ET) at the end of the growth period responded differently to water stress but similarly to N stress because of a higher decrease of harvest index (HI) under water stress than under N stress. Specifically, WPY-ET decreased significantly while WPB-ET kept relatively stable as water stress increased. The variation of the measured WP during the growth period presented a single peak trend. The Logistic, Sigmoid, and Linear models showed similar high accuracy for quantifying the relationship between evapotranspiration accumulation and biomass growth. However, the derived WP values from the nonlinear dynamic models were more in line with the single peak variation of the measured WP during the growth period, with the agreement index (d) improved from 0.21 to 0.65–0.75 in comparison to the Linear model. The developed ET-based N accumulation models can quantify N uptake through evapotranspiration. The results can contribute to improving the accuracy of crop modelling under drought stress conditions, and the design of irrigation and N management in arid and semi-arid regions.