124
Principles Guiding Mapping at all Scales
Spectral and Textural
Ecologists can use phenotypic characteristics of individual species, including the
combination of species in different parts of the canopy, including the understorey, to
help identify plant communities and habitats. With optical remote sensing, this is
not the case, and the features identifiable are only those seen from above, therefore
all (or with a sparse canopy, most) of the return from the signal comprises the canopy species. With a dense canopy no component of the understorey will be visible
in the signal. Therefore, habitats can only be identified clearly from EO if they are
distinguished in botanical terms by their main canopy dominant species. This
knowledge of the dominant cover species of each habitat is considered in terms of
how it is manifest in the imagery in terms of:
• Vegetation productivity (estimated using the Normalised Difference Vegetation
Index (NDVI) (Tucker 1979);
• Vegetation wetness/dryness (which requires the Short Wave Infra-Red band
(SWIR) (Gao 1996);
• The amount of living and dead vegetation and proportion of non-vegetated areas
(Tucker 1979);
• The structure of the vegetation, including woodiness (Kerr and Ostrovsky 2003);
• Variation in the spectral characteristics of different communities at different
stages in the growing season (Cole et al. 2014).
These factors determine the type, scale and amount of imagery needed to map
the habitats.
The Crick Framework
The Crick Framework, was developed as part of the project (Medcalf et al. 2011,
2013); it brings together ecological and Earth Observation knowledge and serves to:
• Categorise habitats in terms of their ability to be mapped remotely; and,
• Provide detailed descriptions of the capacity of EO to support the identification
of habitats.
A wide range of interacting factors has been considered along with ecological
knowledge, to develop a generic classification system that proposes categories
(tiers) of habitat groups (Fig. 1). Habitats are described in terms of spectral characteristics and the spatial detail needed to map them. For instance, small scale or narrow habitats can only be mapped with spatially detailed image data (of a finer scale
than the object itself, so it is clearly identifiable in the imagery), and are therefore
placed in Tier 3b.
K. Medcalf et al.
Principles Guiding Mapping at all Scales
Spectral and Textural
Ecologists can use phenotypic characteristics of individual species, including the
combination of species in different parts of the canopy, including the understorey, to
help identify plant communities and habitats. With optical remote sensing, this is
not the case, and the features identifiable are only those seen from above, therefore
all (or with a sparse canopy, most) of the return from the signal comprises the canopy species. With a dense canopy no component of the understorey will be visible
in the signal. Therefore, habitats can only be identified clearly from EO if they are
distinguished in botanical terms by their main canopy dominant species. This
knowledge of the dominant cover species of each habitat is considered in terms of
how it is manifest in the imagery in terms of:
• Vegetation productivity (estimated using the Normalised Difference Vegetation
Index (NDVI) (Tucker 1979);
• Vegetation wetness/dryness (which requires the Short Wave Infra-Red band
(SWIR) (Gao 1996);
• The amount of living and dead vegetation and proportion of non-vegetated areas
(Tucker 1979);
• The structure of the vegetation, including woodiness (Kerr and Ostrovsky 2003);
• Variation in the spectral characteristics of different communities at different
stages in the growing season (Cole et al. 2014).
These factors determine the type, scale and amount of imagery needed to map
the habitats.
The Crick Framework
The Crick Framework, was developed as part of the project (Medcalf et al. 2011,
2013); it brings together ecological and Earth Observation knowledge and serves to:
• Categorise habitats in terms of their ability to be mapped remotely; and,
• Provide detailed descriptions of the capacity of EO to support the identification
of habitats.
A wide range of interacting factors has been considered along with ecological
knowledge, to develop a generic classification system that proposes categories
(tiers) of habitat groups (Fig. 1). Habitats are described in terms of spectral characteristics and the spatial detail needed to map them. For instance, small scale or narrow habitats can only be mapped with spatially detailed image data (of a finer scale
than the object itself, so it is clearly identifiable in the imagery), and are therefore
placed in Tier 3b.
K. Medcalf et al.