Moderation of nitrogen input and integration of legumes via intercropping enable sustainable intensification of wheat-maize double cropping in the North China Plain: A four-year rotation study

Wheat-maize (W-M) double cropping is the dominant land use system in the North China Plain (NCP). This system has high grain output but suffers from high fertilizer input and nitrogen (N) surplus. Meanwhile, the market demands more protein and oil crops, such as soybean or peanut. Here, we assess whether incorporation of legumes into W-M via intercropping with maize can contribute to lower annual N input while maintaining high annual outputs and diversifying products with high yield stability. We compared yield, yield stability, and profitability of six rotation systems: W-M30 (at the maize inter-plant distance of 30 cm), the density-increased W-M20, wheat-soybean (W-S), wheat-peanut (W-P), and wheat with an intercrop of maize (at the inter-plant distance of 20 cm) and soybean (W-MS) or peanut (W-MP). Four annual N input levels were compared: N0 (no N input), N1 (reduced N input), N2 (target practice), and N3 (current high input). Results over four years showed that replacing maize by maize/legume intercrops had a similar wheat yield as that of W-M while the N input was intermediate between W-M and W-legume rotations. Total actual/equivalent grain yields and gross margins of W-MS or W-MP were consistently intermediate between W-M and W-legume rotations. Intercropping enhanced the yield and temporal yield stability of maize per plant, with benefits for both yield stability of the intercropping system of maize season and the annual rotation system. Averaged over six rotation systems, increasing N supply increased the annual total actual/equivalent grain yield, gross margin and yield stability to a plateau starting at N1 or N2, without a further significant increase at N3. Specifically, the response of total equivalent yield or gross margin of each annual system to increasing N supply could be fitted by the linear-plateau model. Compared to the response curve of W-M, W-MS and W-S reached the plateau with a lower N input and higher yield and profit, while W-MP and W-P reached the plateau with lower N input but decreased the yield and profit. Compared to W-M with 240–360 kg N/ha/year, W-MS used 210–320 kg N/ha/year saving 11.1–12.5% fertilizer N, while maintaining or improving production by 9.1–13.0%, and improving profitability by 12.1–15.6% and temporal yield stability by 12.8–50.6%. W-MS diversifies products with lower N input and higher outputs, profitability, and temporal yield stability than conventional W-M, thus is highly recommended towards productive