欢迎访问林业科学,今天是

林业科学 ›› 2017, Vol. 53 ›› Issue (5): 16-22.doi: 10.11707/j.1001-7488.20170503

• 论文与研究报告 • 上一篇    下一篇

桉树叶片水分状态的近红外光谱检测

卢万鸿, 杨桂丽, 林彦, 王楚彪, 罗建中   

  1. 国家林业局桉树研究开发中心 湛江 524022
  • 收稿日期:2016-08-05 修回日期:2017-02-21 出版日期:2017-05-25 发布日期:2017-06-22
  • 基金资助:
    中央级公益性科研院所基本科研业务费专项资金项目(CAFYBB2017MA022);广东省科技计划项目(2012B020302005);“十三五”科技部国家重点研发计划(2016YFD0600503)。

Assessing Leaf Water Status of Eucalyptus Using NIRS

Lu Wanhong, Yang Guili, Lin Yan, Wang Chubiao, Luo Jianzhong   

  1. Eucalypt Research Centre of State Forestry Administration Zhanjiang 524022
  • Received:2016-08-05 Revised:2017-02-21 Online:2017-05-25 Published:2017-06-22

摘要: [目的] 探索近红外光谱(NIRS)分析技术在赤桉和细叶桉群体叶片水分状态相关生理指标快速检测中应用的可行性。[方法] 选取赤桉7个种源21个家系和细叶桉5个种源26个家系的幼苗,每家系选20株苗木植于等规格塑料盆中,对其按平均日蒸发量的100%,70%,50%,30%和0%5个等级进行控水。控水处理60天后,每家系每控水处理选择3株幼苗,测量其顶端第2对完全展开叶片的相对含水量(RWC)和水势(Ψw),用手持式近红外仪采集对应叶片的近红外光谱信息,持续测量18天。[结果] 利用近红外光谱技术检测赤桉和细叶桉叶片RWC和Ψw的结果显示,与赤桉和细叶桉水分性状关系最密切的近红外光谱区为1 860~1 960 nm。但经二阶导数处理后的近红外光谱则显示,样本叶片RWC和Ψw在近红外全光谱区间存在多个变异峰值。用于预测叶片Ψw建立的偏最小二乘法(PLS)模型的决定系数R2和均方根误差RMSE分别为0.92和0.25,预测叶片RWC的PLS模型的R2和RMSE分别为0.84和1.31。[结论] 对于桉树叶片水分状态生理性状的预测建模应选取近红外全谱段光谱信息,近红外光谱技术可为桉树群体的水分状态检测提供极大便利。

关键词: 桉树, 近红外光谱, 叶片相对含水量, 叶片水势, 近红外PLS模型

Abstract: [Objective] [Objective] The aim of this study was to monitor the leaf relative water content (RWC) and water potential (Ψw) in structured Eucalyptus camaldulensis and E. tereticornis populations by Near Infrared spectroscopy (NIRS). [Method]The samples were collected from the breeding population, which contained 21 families of 7 provenances for E. camaldulensis, and 26 families of 5 provenances for E. tereticornis. A total of 20 average-growth seedlings per family were chosen, divided into 5 groups randomly, and were watered every day by replenishing 100%, 70%, 50%, 30% and 0% of the average water loss of these seedlings from pots by evapotranspiration, respectively. In 60 days after water controls, three seedlings per water control and per family were chosen, of which the top second pair fully expanded leaves were used to be scanned with a portable near infrared spectrometer for getting the NIRS spectra. After the scanning, the leaf RWC and leaf Ψwwere measured. The measurements were conducted once a day for consecutive 18 days.[Result] The results showed that there was a close relationship between water status traits and the raw NIRS at 1 860-1 960 nm. However, after the transform of the 2nd derivate preprocess for raw NIR spectra, there were significant differences of leaf RWC and Ψw among all samples in the whole range of NIRs. The NIRs calibrated PLS(partial least squares) model for the prediction of leaf RWC and Ψw both showed a good fitting. The correlation coefficient between predicted and measured value (R2) were 0.92 and 0.84 for leaf Ψw and RWC, respectively. The average differences between predicted and measured values (RMSE) were 0.25 and 1.31 for leaf Ψw and RWC, respectively. [Conclusion] The coefficient of variation curve showed that the information of whole NIRS range should be used to calibrate the PLS model for the water status traits associated physiological characters in Eucalyptus. All the findings in this study highlight the advantages and perspectives of NIRS in monitoring leaf water status traits in structured Eucalyptus populations.

Key words: Eucalyptus, NIRS, leaf relative water content, leaf water potential, NIRS calibrated PLS model

中图分类号: