Abstract
The Klipriviersberg Nature Reserve has proportionally large number of invasive
plant species (Morné Britz, personal communication). Management of these species
currently focuses on conspicuous woody species and less attention is placed on
smaller plant species that are likely to threaten biodiversity. This can potentially result
in more costly and labour intensive management programmes if imminent
environmental threats are not timeously identified. The use of timely spatial distribution
maps aids in improving invasive plant management strategies. Invasive plant
distribution maps have been developed using traditional mapping methods; but these
are costly and time consuming. Remote sensing techniques on the other hand have
shown the potential in characterizing invasive plants species in different studies. This
study aimed to extend this potential by discriminating selected invasive plant species,
namely, Artemisia afra, Asparagus laricinus and Seriphium plumosum from adjacent
land cover types using continuum spectra of a field spectrometer data. In addition, the
study aimed to investigate the use of spectra simulated according to bands of SPOT
and Landsat images in an effort to explore the potential of extending field based
analysis to airborne or spaceborne remote sensing systems. Data were analysed at
individual, plot and group levels, respectively. Results showed A. afra and A. laricinus
to be best discriminated from adjacent land cover types using the near infrared (NIR)
region from analysis using both original and simulated spectra. None of the regions
that were assessed for S. plumosum, however, did show the potential of discriminating
the species from grass using both the original and simulated spectra. Successful
discrimination of A. afra and A. laricinus from adjacent land cover types using
simulated bands shows the potential of upscaling field based techniques, particularly
the NIR region, to spaceborne and airborne remote sensing technologies such as
SPOT and Landsat. Further studies are, however, recommended to improve the
reliability of the findings obtained in this study. Such studies would need to address
the shortcomings encountered in this study by (1) using more samples, (2)
categorising data analysis according to plant phonological stages to help determine
best timing for discrimination of the species, and (3) taking of spectral measurements
under ideal environmental conditions. Studies on biochemical composition of the
species are also encouraged to inform on reflectance behaviours of the species as
plant compounds or pigments influence electromagnetic reflectance differently.
M.Sc.