93
Background
Habitat Mapping with Remote Sensing
Remote sensing has enormous potential as a source of information on landscape patterns, habitats and dominant species (Bock et al. 2005). Many advantages of remote
sensing data for improving the efficiency for habitat mapping and monitoring have
already been discussed, and similar to the advantages mentioned in Alexandridis
et al. (2009) can be summarised as: High Resolution (HR) coverage of large areas at
low cost; observation at several non-visible wavelengths of the spectrum, and; more
consistent processing across a study area. Mapping of broad habitat types as generic
land cover classes is a common practice using remote sensing and is done on a very
coarse scale (Wulder et al. 2004). At a global scale, land cover mapping has been
accomplished by utilising the Moderate Resolution Imaging Spectroradiometer
(MODIS) satellite at 500 m resolution (Friedl et al. 2010), while country and regional
level land cover classifications have been accomplished using medium resolution
sensors. The two main types of satellite imagery at this resolution, which are more
commonly used in ecological remote sensing, are the Landsat images and the SPOT
system, neither of which, with resolutions of 30 m and 10 m respectively, are capable
of providing the quality needed for the whole range of habitats mapped. This is particularly true if also trying to detect components associated with habitat condition –
always a requirement for conservation management and reporting.
The first satellite-derived pixel-based land cover map of the UK was generated
using Landsat TM in 1990 as part of the United Kingdom Land Cover Map (UK
LCM) (Fuller et al. 1994). Another map was generated in 2000 using an objectbased approach (Fuller et al. 2002) but these maps have been used reluctantly by the
ecological community due to, mainly, low resolutions and inaccuracies, but also due
to the lack of understanding amongst users of the limitations of remote sensing. The
updating of the Phase 1 survey in Wales in 2010 (Lucas et al. 2011) also used EO
data in the form of SPOT-5, ASTER and IRS time-series as a repeat field survey was
deemed unlikely. However, many of the previous problems still existed even with
higher resolutions and better accuracies. On the other hand, this study created the
first national habitat map (as opposed to land cover) generated through the
implementation of EO data, and can potentially be adapted to allow continual monitoring of the extent and condition of habitats (Lucas et al. 2011).
The prospect of monitoring vegetation phenology from EO platforms is also a
key area of interest when discussing the use of EO data and habitat monitoring.
With the emergence of long time data records from sensors it is now possible to
observe variations in phenological parameters, such as length of the growing season. Visible changes in vegetation phenology may be important indicators of climatic change, as phenology responds to the effect of several physiological and
Mapping Coastal Habitats in Wales
Background
Habitat Mapping with Remote Sensing
Remote sensing has enormous potential as a source of information on landscape patterns, habitats and dominant species (Bock et al. 2005). Many advantages of remote
sensing data for improving the efficiency for habitat mapping and monitoring have
already been discussed, and similar to the advantages mentioned in Alexandridis
et al. (2009) can be summarised as: High Resolution (HR) coverage of large areas at
low cost; observation at several non-visible wavelengths of the spectrum, and; more
consistent processing across a study area. Mapping of broad habitat types as generic
land cover classes is a common practice using remote sensing and is done on a very
coarse scale (Wulder et al. 2004). At a global scale, land cover mapping has been
accomplished by utilising the Moderate Resolution Imaging Spectroradiometer
(MODIS) satellite at 500 m resolution (Friedl et al. 2010), while country and regional
level land cover classifications have been accomplished using medium resolution
sensors. The two main types of satellite imagery at this resolution, which are more
commonly used in ecological remote sensing, are the Landsat images and the SPOT
system, neither of which, with resolutions of 30 m and 10 m respectively, are capable
of providing the quality needed for the whole range of habitats mapped. This is particularly true if also trying to detect components associated with habitat condition –
always a requirement for conservation management and reporting.
The first satellite-derived pixel-based land cover map of the UK was generated
using Landsat TM in 1990 as part of the United Kingdom Land Cover Map (UK
LCM) (Fuller et al. 1994). Another map was generated in 2000 using an objectbased approach (Fuller et al. 2002) but these maps have been used reluctantly by the
ecological community due to, mainly, low resolutions and inaccuracies, but also due
to the lack of understanding amongst users of the limitations of remote sensing. The
updating of the Phase 1 survey in Wales in 2010 (Lucas et al. 2011) also used EO
data in the form of SPOT-5, ASTER and IRS time-series as a repeat field survey was
deemed unlikely. However, many of the previous problems still existed even with
higher resolutions and better accuracies. On the other hand, this study created the
first national habitat map (as opposed to land cover) generated through the
implementation of EO data, and can potentially be adapted to allow continual monitoring of the extent and condition of habitats (Lucas et al. 2011).
The prospect of monitoring vegetation phenology from EO platforms is also a
key area of interest when discussing the use of EO data and habitat monitoring.
With the emergence of long time data records from sensors it is now possible to
observe variations in phenological parameters, such as length of the growing season. Visible changes in vegetation phenology may be important indicators of climatic change, as phenology responds to the effect of several physiological and
Mapping Coastal Habitats in Wales