单作套作大豆叶片氮素积累与光谱特征

种植模式和氮肥水平直接影响作物的生长和氮素的吸收,无损、即时监测大豆叶片氮素水平对大豆生产中的氮肥精确管理十分重要。本研究设置4个氮肥水平,分析单作套作下大豆在不同生育时期叶片氮素动态和光谱特征,明确对叶片氮素敏感的光谱特征参数,构建单作套作大豆通用的叶片氮素积累量估测模型。结果表明,随大豆生育时期的推进,单作套作种植模式下的大豆冠层叶片氮素积累量均呈现单峰变化趋势,最大值出现在N3处理下的结荚期,两种模式两年最大值平均分别为8.70 g m~(–2)和8.38 g m~(–2);不同生育时期和种植模式的大豆冠层原始反射光谱的变化规律与冠层叶片氮素变化规律均为先增加后降低,原始反射光谱在700~1000 nm波段的反射率以结荚期为拐点先增大后减小,最大反射率达到60%~70%左右;通过对单作套作大豆冠层光谱一阶导数变换,红边幅值呈现先增加后降低的趋势,同时红边位置随叶片氮积累量的增加和减小出现"红移"与"蓝移"现象。经波段自由组合和回归分析表明,以DSI(771、755)构建的线性(y=–1.249+3.209x,R~2=0.847)和乘幂(y=–1.470x~(1.676),R~2=0.872)模型能较精确地估测不同生育时期大豆冠层叶片氮素状况。

Remote Detection of Canopy Leaf Nitrogen Status in Soybean by Hyperspectral Data under Monoculture and Intercropping Systems

Non-destructive monitoring of soybean nitrogen status is important for precise N management in soybean production. In this study, the quantitative correlation between soybean leaf nitrogen status and canopy hyperspectral reflectance was investigated. Field experiments were conducted. With four nitrogen application rates for two years under monoculture and intercropping systems. The nitrogen accumulation of canopy leaves showed a single-peak changing trend in the process of soybean growth. The maximum value in monoculture and intercropping was 8.7 g m–2, 8.38 g m–2, respectively, at pod stage under N3 treatment. The raw hyperspectral reflectance and the leaf nitrogen accumulation had the same changing trend at different growth stages with different planting patterns. The peak value of the raw hyperspectral reflectance in the 700–1000 nm occurred at pod stage. In the first-order derivative spectrum, the red edge amplitude values increased first and then decreased. The position of the red edge changed as “Red shift” and “Blue shift” with the increase or decrease of leaf nitrogen accumulation. The results of the correlation analysis showed that the linear model and the power model by using the Difference Spectral Index (DSI: 771, 755) based on the best spectral band combination (BSBC) had the greatest accuracy to estimate the leaf nitrogen status of soybean.