论文部分内容阅读
本文选取英国ADAS 1987—1989年的三种蔬菜(花茎甘蓝、大白菜、韭葱)氮肥试验为材料,分别选取总产量、干物重和市场产量为因变量,以施N量为自变量,对具有代表性的三种肥料效应函数:二次型Y=b_0+b_1X+b_2X~2(M1)、线性加指数Y=a+br~x+cX(M2)、倒线性Y=dbX/(8+bX)+cX(M3)进行了系统的比较研究。研究结果指出:用三种模型所配置的60个肥料效应方程均有55个达到了F_(0.1)以上显著水准,其中M2具有最大的平均回归方差贡献率(93%)和最小的平均残差方差(15.9)。进一步t检验指出:用M1所预报的最高平均产量显著高于M2和M3的预报结果。但M2所计算的平均最高施肥量显著低于M1和M3的计算结果。从模型的极值点考虑,M2较适合于模拟花茎计蓝、大白菜的肥料效应曲线;M3较适合于模拟花茎甘蓝、韭葱的肥料效应的曲线;M1的通用性较强。
In this paper, three kinds of vegetables (broccoli, Chinese cabbage, leek) nitrogen fertilizers from ADAS in 1987-1989 were selected as the materials. Total yield, dry matter weight and market yield were selected as dependent variables and N as independent variables Representative three kinds of fertilizer effect function: quadratic Y = b_0 + b_1X + b_2X ~ 2 (M1), the linear additive index Y = a + br ~ x + cX + bX) + cX (M3) for a systematic comparative study. The results show that 55 of the 60 fertilizer effect equations configured by the three models achieve a significant level above F 0.1, of which M2 has the largest mean regression contribution (93%) and the smallest mean residual Variance (15.9). Further t-test indicated that the highest average yield predicted by M1 was significantly higher than the predicted results of M2 and M3. However, the average maximum fertilization amount calculated by M2 was significantly lower than that calculated by M1 and M3. From the extremum point of the model, M2 is more suitable for simulating the fertilizer effect curve of blue and Chinese cabbage; M3 is more suitable for simulating the fertilizer effect curve of broccoli and leek; M1 is more versatile.