Spectral slope as an indicator of pasture quality

Rachel Lugassi, Alexandra Chudnovsky, Eliyahu Zaady, Levana Dvash, Naftaly Goldshleger

Research output: Contribution to journalArticlepeer-review

18 Scopus citations

Abstract

In this study, we develop a spectral method for assessment of pasture quality based only on the spectral information obtained with a small number of wavelengths. First, differences in spectral behavior were identified across the near infrared-shortwave infrared spectral range that were indicative of changes in chemical properties. Then, slopes across different spectral ranges were calculated and correlated with the changes in crude protein (CP), neutral detergent fiber (NDF) and metabolic energy concentration (MEC). Finally, partial least squares (PLS) regression analysis was applied to identify the optimal spectral ranges for accurate assessment of CP, NDF and MEC. Six spectral domains and a set of slope criteria for real-time evaluation of pasture quality were suggested. The evaluation of three level categories (low, medium, high) for these three parameters showed a success rate of: 73%-96% for CP, 72%-87% for NDF and 60%-85% for MEC. Moreover, only one spectral range, 1748-1764 nm, was needed to provide a good estimation of CP, NDF and MEC. Importantly, five of the six selected spectral regions were not affected by water absorbance. With some modifications, this rationale can be applied to further analyses of pasture quality from airborne sensors.

Original languageEnglish
Pages (from-to)256-274
Number of pages19
JournalRemote Sensing
Volume7
Issue number1
DOIs
StatePublished - 1 Jan 2015
Externally publishedYes

Keywords

  • Metabolic energy concentration (MEC)
  • Neutral detergent fiber (NDF)
  • Pasture quality
  • Protein
  • Reflectance spectroscopy
  • Spectral slope

ASJC Scopus subject areas

  • Earth and Planetary Sciences (all)

Fingerprint

Dive into the research topics of 'Spectral slope as an indicator of pasture quality'. Together they form a unique fingerprint.

Cite this