Farming systems’ productivity and soil organic carbon stocks following fertilisers, no-tillage or legumes on a fertility-depleted soil in a semi-arid subtropical region

Abstract

Depleted soil nitrogen supplies in long-term continuously cultivated soil for cereal grain cropping have resulted in reduced cereal yields, low grain proteins and hence low economic returns. This has necessitated the development of alternative management practices to sustain crop yields, as well as to restore and maintain soil fertility. In the present study we examined the comparative performance of several management options over a 12-year period, including: a 4-year rotation of grass + legume pasture followed by wheat (GL–wheat); 2-year rotations of lucerne–wheat, annual medic–wheat and chickpea–wheat; and continuous conventional tillage (CT) or no-tillage (NT), without or with fertiliser N application (0, 25 and 75 kg N ha–1 for each crop). Average wheat grain yields were highest in the chickpea–wheat rotation, followed by the NT wheat with 75 kg N ha–1; the lowest grain yields were in the CT or NT wheat treatment without fertiliser N application. Crop water use and gross margin were strongly correlated. However, there was an increasing potential for the deep leaching of nitrate-N at 75 kg N ha–1 application, as well as from the GL pasture initiated in 1987, but not from that initiated in 1986, emphasising the effect of variability in growing seasons. Soil organic C stocks increased under the 4-year GL pasture in the 0–0.1 m depth only, then decreased steadily following the cropping phase. The rotation of 4-year GL pasture followed by wheat cropping for 4–6 years may maintain initial soil organic C stock, but a shorter cropping phase is required to increase soil organic C and N stocks and soil fertility in the long term. Partial economic analysis of the treatments suggested that restoring or maintaining soil N fertility, either through legume-based pastures, grain legume and/or N fertiliser, provides long-term positive economic return.