127
type of data also ensures that the outputs are of practical use at the local level and
enabled the maps produced at differing scales in Norfolk to be aligned and comparable. This allows key features (e.g., roads, rivers and field boundaries) within the
landscape to be consistently mapped at all scales.
Landscape Context: Confining the Search for Habitats
Once real world objects are defined, these can be classified using the rule-based
approach which is informed by ecological knowledge. Often there is insufficient
spectral information within the EO data to allow classification of habitats given the
diverse range of land cover and habitats types present in the extensive areas covered
by a satellite image. To overcome this, the component elements in the landscape are
separated using the ecological properties of the main habitat types. This ‘landscape
context’ of the habitat (Lucas 2011) is significant when mapping at the regional
scale because some habitats only occur in specific locations in the landscape. For
example, Floodplain and Coastal Grazing Marsh is only found within the floodplain
or next to brackish waters which are seasonally inundated. The location of these
habitats therefore has a direct geographic range. By first considering and examining
this ‘macro-scale’ the ecologist is able to define the ‘ecologically coherent’ distinct
landscapes where particular habitats are known to be located. The EO specialist can
then exploit this information together with knowledge of the spectral properties of
the habitats to produce the classification. This is shown in Fig. 3 (see Bell et al.
2015; Medcalf et al. 2011; Medcalf et al. 2013).
Placing Boundaries in Ecotones
Placing boundaries between habitats on maps is challenging whether mapping by
fieldwork or by remote sensing as the habitats are not always clearly delineated. In
some environments, a transition area occurs where habitats meet and integrate,
which is referred to as an ecotone. In the coastal areas of north Norfolk, for example,
saltmarsh grades into grazing marsh which in turn grades into natural grasslands as
the distance from the sea and height of the land increases. Within these habitats,
each of these ‘ecotones’ is represented in the imagery by a change in productivity
and wetness in the spring. Distinguishing one habitat from another therefore
depends on understanding the relationship between the ecotones and making a
decision about where a ‘hard mapped’ boundary would best be placed along the
continuum (Natural England 2011a, b, c; 2012).
Integrated Monitoring for Biodiversity Using Remote Sensing: From Local…
type of data also ensures that the outputs are of practical use at the local level and
enabled the maps produced at differing scales in Norfolk to be aligned and comparable. This allows key features (e.g., roads, rivers and field boundaries) within the
landscape to be consistently mapped at all scales.
Landscape Context: Confining the Search for Habitats
Once real world objects are defined, these can be classified using the rule-based
approach which is informed by ecological knowledge. Often there is insufficient
spectral information within the EO data to allow classification of habitats given the
diverse range of land cover and habitats types present in the extensive areas covered
by a satellite image. To overcome this, the component elements in the landscape are
separated using the ecological properties of the main habitat types. This ‘landscape
context’ of the habitat (Lucas 2011) is significant when mapping at the regional
scale because some habitats only occur in specific locations in the landscape. For
example, Floodplain and Coastal Grazing Marsh is only found within the floodplain
or next to brackish waters which are seasonally inundated. The location of these
habitats therefore has a direct geographic range. By first considering and examining
this ‘macro-scale’ the ecologist is able to define the ‘ecologically coherent’ distinct
landscapes where particular habitats are known to be located. The EO specialist can
then exploit this information together with knowledge of the spectral properties of
the habitats to produce the classification. This is shown in Fig. 3 (see Bell et al.
2015; Medcalf et al. 2011; Medcalf et al. 2013).
Placing Boundaries in Ecotones
Placing boundaries between habitats on maps is challenging whether mapping by
fieldwork or by remote sensing as the habitats are not always clearly delineated. In
some environments, a transition area occurs where habitats meet and integrate,
which is referred to as an ecotone. In the coastal areas of north Norfolk, for example,
saltmarsh grades into grazing marsh which in turn grades into natural grasslands as
the distance from the sea and height of the land increases. Within these habitats,
each of these ‘ecotones’ is represented in the imagery by a change in productivity
and wetness in the spring. Distinguishing one habitat from another therefore
depends on understanding the relationship between the ecotones and making a
decision about where a ‘hard mapped’ boundary would best be placed along the
continuum (Natural England 2011a, b, c; 2012).
Integrated Monitoring for Biodiversity Using Remote Sensing: From Local…