476
S.D. King and D.R. Green
map (Joao, 1998). Spatial resolution on the other hand refers to the smallest object that
is discernable in the data. For example, in remote sensing, a resolution of 30m means
that features and objects less than 30m across would not be seen as separate to their
surroundings. Scale and resolution are therefore related because scale sets the lower
limit to the size of an object that may be shown on a map (Joao, 1998). In terms of
information content, a large-scale map of 1:10,000 will usually contain more detailed
information than a 1:50,000 map covering the same geographical area. Likewise, an
aerial photograph with a spatial resolution of 1m will potentially contain more detailed
information than a subset of a satellite image covering the same area, with a spatial
resolution of 30m. In some cases, however, a small-scale map (with greater
generalisation) will arguably provide more information, especially when considering
large-scale processes (Goodchild and Quattrochi, 1997). This would also be true of a
larger resolution image, where some patterns become more apparent when a larger area
of the Earth's surface can be seen.
Scale and resolution consequently play a crucial role in determining which data
is used to provide the information for a given situation. For instance, most people, when
trying to find out a road route between two locations will turn to a road atlas, perhaps
with a scale of 1:125,000. However, if a specific street in a city had to be found, then a
street map, with a scale of perhaps 1:15,000 would be more appropriate. The need for
greater information content requires the user to move from a small-scale to a larger
scale map. In essence, the user has matched the issue (or information requirement) to
the utility of the map, the process being made more simple because maps are organised
hierarchically by scale, with most users having the knowledge that large-scale maps
offer greater detail (information) than small-scale maps.
The same principle may conceivably be applied to remotely sensed imagery for
environmental management, particularly where the information required from the
imagery has a spatial component. In theory, remotely sensed imagery could be arranged
using a hierarchy based on spatial resolution, rather than scale, with detailed
information being available on high-resolution imagery, and more generalised
information available on low-resolution imagery. Such a scheme does rely heavily upon
the principle that most of information required from remotely sensed imagery will be
the information that is visually obvious. While this by no means disregards the
importance of spectral information (outside the visible), the work of King (1998) has
argued that most non-specialist users of the data will find it easier to visually extract
information from the imagery, either by manual or on-screen, computer-based,
interpretation. Again, this is largely because remotely sensed imagery provides us with
an instantaneous recognition (through shapes, patterns and colours) of geographical
features and relationships within the natural environment (relating to the geographic
metaphor suggested by Couclelis, 1998). Even where spectral information is important,
the spatial resolution will still inevitably govern the information that can be extracted, as
spectral patterns will have a multi-spatial and therefore multi-scale component.
Matching Issue to Utility – Using Coastal Zone Management as an Example
The coastal environment is perhaps one of the most unique in the world. The
multiplicity of biological and landscape types which the interaction between land, water
S.D. King and D.R. Green
map (Joao, 1998). Spatial resolution on the other hand refers to the smallest object that
is discernable in the data. For example, in remote sensing, a resolution of 30m means
that features and objects less than 30m across would not be seen as separate to their
surroundings. Scale and resolution are therefore related because scale sets the lower
limit to the size of an object that may be shown on a map (Joao, 1998). In terms of
information content, a large-scale map of 1:10,000 will usually contain more detailed
information than a 1:50,000 map covering the same geographical area. Likewise, an
aerial photograph with a spatial resolution of 1m will potentially contain more detailed
information than a subset of a satellite image covering the same area, with a spatial
resolution of 30m. In some cases, however, a small-scale map (with greater
generalisation) will arguably provide more information, especially when considering
large-scale processes (Goodchild and Quattrochi, 1997). This would also be true of a
larger resolution image, where some patterns become more apparent when a larger area
of the Earth's surface can be seen.
Scale and resolution consequently play a crucial role in determining which data
is used to provide the information for a given situation. For instance, most people, when
trying to find out a road route between two locations will turn to a road atlas, perhaps
with a scale of 1:125,000. However, if a specific street in a city had to be found, then a
street map, with a scale of perhaps 1:15,000 would be more appropriate. The need for
greater information content requires the user to move from a small-scale to a larger
scale map. In essence, the user has matched the issue (or information requirement) to
the utility of the map, the process being made more simple because maps are organised
hierarchically by scale, with most users having the knowledge that large-scale maps
offer greater detail (information) than small-scale maps.
The same principle may conceivably be applied to remotely sensed imagery for
environmental management, particularly where the information required from the
imagery has a spatial component. In theory, remotely sensed imagery could be arranged
using a hierarchy based on spatial resolution, rather than scale, with detailed
information being available on high-resolution imagery, and more generalised
information available on low-resolution imagery. Such a scheme does rely heavily upon
the principle that most of information required from remotely sensed imagery will be
the information that is visually obvious. While this by no means disregards the
importance of spectral information (outside the visible), the work of King (1998) has
argued that most non-specialist users of the data will find it easier to visually extract
information from the imagery, either by manual or on-screen, computer-based,
interpretation. Again, this is largely because remotely sensed imagery provides us with
an instantaneous recognition (through shapes, patterns and colours) of geographical
features and relationships within the natural environment (relating to the geographic
metaphor suggested by Couclelis, 1998). Even where spectral information is important,
the spatial resolution will still inevitably govern the information that can be extracted, as
spectral patterns will have a multi-spatial and therefore multi-scale component.
Matching Issue to Utility – Using Coastal Zone Management as an Example
The coastal environment is perhaps one of the most unique in the world. The
multiplicity of biological and landscape types which the interaction between land, water