Modified water-nitrogen productivity function based on response of water sensitive index to nitrogen for hybrid maize under drip fertigation

Abstract Accurate simulation of yield under different irrigation and fertilizer conditions is the basis of precise water and nitrogen management, which is the determining factor for agricultural sustainable development. To find out the responses of actual water consumption (ETa), dry mass (DM) and yield(Y) of drip-fertigated hybrid maize to different water and nitrogen levels, three-year continuous field experiments were conducted in Hexi Corridor, Northwest China from 2016 to 2018. Three irrigation levels CK, W1, W2 (100%, 75% and 50% of full irrigation) at four growth stages S1, S2, S3 and S4 (seedling, jointing, tasseling and grain filling stage) combined with seven nitrogen rates (50, 150, 200, 250, 300 kg/ha in 2016, 0, 50, 100, 150, 200, 250, 300 kg/ha in 2017 and 2018) were set. The results showed that in addition to S1W1 treatment, the yield of CK was higher than the yield under deficit irrigation of the same nitrogen level. When the nitrogen application was 150 kg/ha, the average yield of S1W1 increased by 6% compared with CK. When nitrogen application increased from 150 kg/ha to 300 kg/ha, the yield of CK, S1W2 and S3W2 decreased in different extent, by 18%, 22%, and 17% respectively. Based on the parabolic relationships between yield and nitrogen rate, we modified Jensen model by introducing a new water sensitive index which considered the response of crop water sensitivity to the nitrogen application and background mineral in soil. Comparing the modified Jensen model (W*N−Jensen model), to Jensen model and N-Jensen model, which considering the parabolic relationship between the yield and nitrogen application only, we found that W*N−Jensen model can better simulate the yield under different combinations of irrigation and nitrogen levels, with higher determination coefficient (R2), higher affinity index (d), lower normalized root mean square error (NRMSE), and lower Akaike information criterion (AIC). W*N−Jensen model has higher accuracy and consistency, and can be used to provide scientific basis for water and nitrogen management and yield forecasting in the field.