Spectral discrimination of increaser species as an indicator of rangeland degradation using field spectrometry

Abstract

Discriminating increaser species is important for mapping rangeland degradation. The main objectives of this paper were to (1) determine whether four increaser species could be discriminated from each other and (2) determine the key wavelengths that have high discriminatory power. Spectral data were taken at canopy level from Hyparrhenia hirta, Eragrostis curvula, Sporobolus africanus and Aristida diffuse from Okhombe rangeland, KwaZulu-Natal province, South Africa. A field portable analytical spectral device with spectral range from 350 nm to 2500 nm was used in the analysis. Three-tier hierarchical techniques of One-way ANOVA, stepwise discriminant function analysis and canonical function analysis were used. The results revealed that there were statistically significant differences in spectral reflectance among four species on 439 wavelengths. The most important eight wavelengths (665 nm, 729 nm, 848 nm, 895 nm, 1039 nm, 998 nm, 681 nm and 972 nm) that have been selected for spectral discrimination were largely located in the visible, red-edge and near-infrared regions of the spectrum. The three tiers of analysis yielded species discrimination with an overall accuracy of 83 percent and a KHAT value of 0.77. The use of the spectroscopic approach applied in this study indicated that the increaser species were spectrally different. Therefore, these results encouraged further investigation into the possibility of mapping increaser species as indicators of different levels of rangeland degradation.

Keywords:

• rangeland degradation
• vegetation species
• indicator
• field spectrometry

Acknowledgements

We would like to thank Amazizi Traditional Administrative Council, for the permission to conduct this research at Okhombe ward. My appreciation extends to El-fashir University (Sudan) and University of KwaZulu-Natal (South Africa) for giving one of us (KM) the opportunity to read for a PhD. Our gratitude further extends to Dr Elhadi Adam, Monique Salomon, and colleagues in CEAD for their assistance and valuable comments.