Depleting weed seed banks within non-crop phases for the benefit of subsequent crops

摘要

In 2003 a rotation trial was established at Avondale research station Western Australia to evaluate the effectiveness of introducing various one and three year non-crop phases into the rotation to deplete the seed bank of herbicide-resistant, annual ryegrass (Lolium rigidum Gaudin) populations. Treatments included grazed, three year pasture phases of subclover (Trifolium subterraneum L.) cv. Dalkeith, biserrula (Biserrula pelecinus L.) cv. Casbah, and lucerne (Medicago saliva L.); and ungrazed, one year phases of green manured field peas (Pisum sativum L.) or canola (Brassica napus L.), and brown manured French serradella (Ornithopus sativus Brot.) cv. Cadiz. Key measurements included seed bank size and weed density in the non-crop phase and subsequent crops. Initially (2003) the site had over 1000 ryegrass plants m~(-2) and approximately 350 dormant ryegrass seeds in mid winter. As expected, weed seed-set control was highly effective in green and brown manured treatments; these interventions contained the ryegrass seed bank initially, but one year of seed-set control was inadequate to return to sequential cropping, for the ryegrass seed bank increased rapidly to more than 10,000 dormant seeds m~(-2)within two years. By contrast, grazed, three year pasture phases combined with non-selective seed-set control significantly depleted the annual ryegrass seed bank. Overall grazed subclover was the most effective treatment with the annual ryegrass seed bank falling below 40 seeds m~(-2) and legume content rising from below 20%, in the establishment year, to greater than 90% in the third year of the non-crop phase. Collectively the results suggest that introducing a non-crop phase can provide an opportunity to decimate weed seed banks and minimise thedetrimental impact of weeds on subsequent crops.