利用红边特征参数监测小麦叶片氮素积累状况

Monitoring nitrogen accumulation in wheat leaf with red edge characteristics parameters

以不同类型小麦品种在氮素差异梯度下连续3a田间试验为基础,在关键生育时期同步测定冠层光谱反射率、叶片干物质量及氮含量,探索建立小麦叶片氮素状况估算的新型红边参数及监测模型。结果表明,冠层微分光谱在红边区域内随氮素水平提高呈明显规律性变化,而原始光谱反射率的变化却较为复杂。与叶片氮积累量关系密切的常见红边参数间存在差异,其中,以GM2、SR705和FD742表现最突出,线性回归模型拟合精度(R2)分别为0.854、0.848和0.873,估计标准误差(SE)分别为1.136、1.160和1.059。基于红边双峰特征分析,构建新型红边双峰特征参数,其中,红边左偏峰面积LSDr_REPLE对叶片氮积累量方程拟合取得很好效果,决定系数和估计标准误差分别为0.869和1.080。经不同年际独立数据的检验表明,以GM2、SR705和FD742为变量,模型预测平均相对误差(RE)分别为17.6%、17.0%和14.9%,而红边左偏峰面积LSDr_REPLE模型预测误差控制得更好,平均相对误差RE为14.5%。以上表明,红边参数GM2、SR705和FD742可以对小麦叶片氮素状况进行有效监测,而红边左偏峰面积LSDr_REPLE模型预测更为准确可靠。

Three field experiments were conducted with different nitrogen application rates and wheat cultivars across three growing seasons, and time-course measurements were taken on canopy spectral reflectance, leaf dry weight and leaf nitrogen concentration under the various treatments. The primary objective of this study was to explore the optimum red edge characteristics parameters and quantitative models for estimating leaf nitrogen accumulation in wheat （Triticum aestivum L.）. The results showed that the first derivative of the reflectance spectra changes regularly with increasing N rates in red edge region, and canopy spectral reflectance changes complexly. The analyses on relationships between the vegetable indices reported to leaf N accumulation indicated that red edge spectral parameters related most significantly to leaf N accumulation, differed among red edge spectral parameters. An integrated linear regression equation of leaf N accumulation to GM2, SR705 and FD742 described the dynamic pattern of change in leaf N accumulation in wheat, giving the determination of coefficients （RE） as 0.854, 0.848 and 0.873, respectively, and the standard errors （SE） as 1.136, 1.160 and 1.059, respectively. The two peak spectral parameters in red edge region were constructed on analysis of red edge characteristics, and differential vegetation index of two peak in red edge region LSDr_REPLE was highly correlated with leafN accumulation with 0.868 ofR2 and 1.080 of SE. When independent data were fit to the derived equations, the average relative error （RE） values as 17.6%, 17.0%, 14.9% and 14.5% between measured and estimated N accumulation using spectral parameters GM2, SR705, FD742 and LSDr_REPLE, respectively, indicating a good fit and better in LSDrREPLE. The result indicated that those models could be used to reliably estimate the leaf N states in wheat, and especially LSDrREPLE of new extracted parameters could indicate further steadily dynamic changes in leafN accumulation.