Lucerne removal before a cropping phase

Growing dryland crops after lucerne is known to be risky because of the lack of residual soil water. We investigated ways of reducing this risk by removing portions of a lucerne pasture, using either herbicides or cultivation, at monthly intervals between November and April, before sowing a wheat crop in May, followed by a canola crop in the following year. The experimental site was on a red-brown earth in southern New South Wales. Lucerne removal was incomplete when the wheat was sown, so all lucerne plants were removed from half of each plot with a post-emergence herbicide, to allow comparisons of intercropped wheat-lucerne and wheat monoculture. Measurements were made on crop growth, yield, grain quality, soil water, and soil mineral nitrogen (N) before and after both crops. On average, each additional month between lucerne removal and wheat sowing led to a yield increase of 8% and a grain protein increase of 0.3 percentage units. The main reason for the increases was additional soil mineral N, associated with a longer period of mineralisation. The soil water content at the time of wheat sowing was greater with early lucerne removal but the growing season rainfall did not limit yields, and there was more residual soil water at the time of wheat maturity where lucerne had been removed late and yields were lower. Method of lucerne removal did not significantly affect wheat yield, grain protein, soil water, or soil mineral N. The portions of the plots containing lucerne plants that survived the initial removal attempt produced similar wheat yields to the portions where lucerne had been totally removed, but grain protein was lower. The following growing season was drier, but despite less residual soil water where lucerne had been removed earlier in the previous year, the average canola yield was 2.5% greater for each additional month of fallow. The increase again appeared to be due to more residual mineral N. The seed oil concentration also decreased in response to later lucerne removal but seed protein increased. Where lucerne plants had been retained in the previous wheat crop, canola yield was lower than where they had been totally removed, apparently because of less soil water at sowing. Over the 2 years of the experiment, the net supply of mineral N was 374 kg N/ha, equivalent to an annual net mineralisation of 2% of the total soil N. The initial mineralisation rate was slow, suggesting that the soil may be deficient in mineral N soon after lucerne removal.