loading page

Predicting the Photosynthetic Capacity and Leaf Nitrogen of Woody Bioenergy Crops from Hyperspectral Reflectance Models
  • +2
  • Thu Ya Kyaw,
  • Heidi Renninger,
  • Courtney Siegert,
  • Padmanava Dash,
  • Krishna Poudel
Thu Ya Kyaw
Mississippi State University

Corresponding Author:tk758@msstate.edu

Author Profile
Heidi Renninger
Mississippi State University
Author Profile
Courtney Siegert
Mississippi State University
Author Profile
Padmanava Dash
Mississippi State University
Author Profile
Krishna Poudel
Mississippi State University
Author Profile

Abstract

Generating renewable bioenergy crops requires varietals that are suited to grow under varying environmental conditions necessitating the development and testing of a wide range of poplar (Populus) genotypes. Meanwhile, there is an increasing demand for refining the selection process of high-performing poplars. However, a cost-effective method is still needed to predict the productivity of various poplar genotypes. Photosynthetic capacity and leaf nitrogen are important growth-related physicochemical traits, but measuring them in the field and laboratory is expensive and time-consuming. Alternatively, remote sensing of hyperspectral leaf spectra may serve as a proxy to rapidly estimate these traits, which are associated with absorption, reflection, and transmission of solar radiation. To quantify photosynthetic traits, CO2 response curves were used to estimate Rubisco-limited carboxylation rate (Vcmax), maximum electron transport rate (Jmax), and triose phosphate utilization (TPU). From the same leaves measured for photosynthesis, leaf reflectance was measured with a handheld spectroradiometer. We measured a total of 105 leaf samples, including 6 taxa with 61 different poplar genotypes. For data analyses, Least Absolute Shrinkage and Selection Operator and Principal Component Analysis were used to determine the wavelengths that were the most useful for capturing the variability in the physicochemical data. Results showed that leaf reflectance at 758 nm and 936 nm were crucial wavelengths for predicting Vcmax (RMSPE = 31%) and Jmax (RMSPE = 32%), while 687 nm and 757 nm were important predictors for TPU (RMSPE = 31%), and 709 nm and 927 nm were important predictors for leaf nitrogen (RMSPE = 22%). The wavelengths near 687 nm and 760 nm are the oxygen absorption bands, and also overlap with the chlorophyll fluorescence emission of plants. Therefore, it is possible to apply hyperspectral reflectance models for rapid clonal screening and high-throughput field phenotyping of photosynthetic capacity parameters and leaf nitrogen of various poplar genotypes.