DRY MATTER DECOMPOSITION OF COVER CROPS IN A NO-TILLAGE COTTON SYSTEM

ABSTRACT No-tillage cotton systems require soil coverage with cover crop residue for a longer time due to the late cycle of cotton. However, decomposition rates may vary between cover crops, and the adjustment of models to describe it is critical to no-tillage cotton management. Two non-linear regression models, exponential (EM) and Michaelis-Menten (MM), were adjusted to dry matter decomposition of cover crops in a cotton no-tillage system, in Brazil. Three field trials were performed in 2012 for the cover crops Urochloa ruziziensis (brachiaria), Pennisetum glaucum (pearl millet), and Cajanus cajan (pigeon pea). Samples of cover crop were collected at 20, 50, 70, 110, 140, and 170 days after sowing upland cotton to measure dry matter decomposition. MM showed better adjustment than EM for all cover crops. The estimations of half-life parameters were different between the cover crops, suggesting that each cover crop has its own rate of decomposition. For pearl millet, brachiaria, and pigeon pea, the half-life estimation by exponential model was over the MM in 9, 12, and 12 days.

摘要：鉴于棉花生育周期较长，免耕棉田需依托覆盖作物残体维持更长时间的土壤覆盖。然而不同覆盖作物的残体分解速率存在差异，因此构建适配的分解模型对免耕棉田管理至关重要。本研究针对巴西免耕棉田的覆盖作物干物质分解过程，对两类非线性回归模型——指数模型（Exponential Model, EM）与米氏动力学模型（Michaelis-Menten, MM）——开展了拟合校准。2012年，针对鲁兹臂形草（Urochloa ruziziensis，又称臂形草）、御谷（Pennisetum glaucum，又称珍珠粟）以及木豆（Cajanus cajan，又称鸽豆）三种覆盖作物，完成了3组田间试验。在陆地棉播种后的第20、50、70、110、140及170天采集覆盖作物残体样本，以测定其干物质分解量。结果显示，相较于指数模型，米氏动力学模型对所有覆盖作物的分解过程均呈现更优的拟合效果。不同覆盖作物的半衰期参数估计值存在显著差异，表明各覆盖作物的残体分解速率具有各自的特征。针对珍珠粟、臂形草与木豆，指数模型估算得到的半衰期分别比米氏动力学模型结果高出9天、12天与12天。