478
S.D. King and D.R. Green
than most other current sensors, is perceived as being more useful. Finally, there is a
widely held belief that because there are relatively few sensors developed specifically
for coastal zone monitoring, then the imagery available from other sensors will not
provide useful information (see for example Doody and Pamplin, 1998).
All of these problems can be relatively easily overcome with better education,
greater use and practice. For example, most remotely sensed imagery can be displayed
using colour combinations that more closely represent what the human eye sees, and
most imagery can be output as a hard copy. Now and in the future, more simple
techniques for delivering and manipulating digital data will become available that make
digital imagery easier to use. In terms of resolving the last two problems, it is necessary
to return to the amount of information required from an image, and this relates to the
issue to which the information held in the image is being applied. If, for instance, the
position and general quality of a section of sea wall was required, then aerial
photography would most likely provide the answer, as only the high spatial resolution of
aerial photography can really show such detail. However, if the total area of an estuary
or firth was needed, then a satellite image would be more appropriate to use, because far
more aerial photographs would be needed to cover such a large area, with no benefit in
information content. What this shows us is that people need to be made aware that all
remotely sensed imagery can have some use, as long as the correct imagery is applied in
the correct situation. Nevertheless, this requires a thorough understanding of the nature
of the problem being investigated, an understanding of the information required, and a
reference guide as to which imagery will provide that information. In essence, the issue
must be matched to the utility or characteristics of the sensor.
There are a whole range of issues, both natural and anthropogenic, which need
to be resolved at the coast. To match these issues to the imagery, one must understand
the information required to solve the issue and then the spatial, spectral and temporal
characteristics needed to provide the information. These characteristics can then be
compared with and matched to the spatial, spectral and temporal characteristics of the
different sensors. For example, a coastal issue might be the perceived increase in coastal
flooding due to saltmarsh erosion. The information required to investigate this issue
would be, where are the saltmarshes, are they eroding, which properties and land are
vulnerable? The objective would be to find imagery that could be used to identify
saltmarsh areas, analyse coastal change, and identify vulnerable property or land. To
identify the general pattern and location of saltmarsh areas in an estuary would usually
require the imagery to have a spatial resolution of at least 30m (Fig. 1) (this would
provide enough detail as well as area coverage). The spectral characteristics would not
be too significant in this case, as areas of saltmarsh can often be identified by pattern
and apparent texture. However, different band combinations can be used to highlight
certain vegetation and the vegetation/water boundary. A series of images over a period
of time would be needed to identify saltmarsh erosion. Having identified any areas of
possible erosion (Fig. 2), it would then be possible to look at these areas in more detail
using higher resolution imagery which could then be used to locate any land or property
in danger (Fig. 3). The information provided by such an analysis would then be
available for the coastal manager to formulate a management plan for this issue.
S.D. King and D.R. Green
than most other current sensors, is perceived as being more useful. Finally, there is a
widely held belief that because there are relatively few sensors developed specifically
for coastal zone monitoring, then the imagery available from other sensors will not
provide useful information (see for example Doody and Pamplin, 1998).
All of these problems can be relatively easily overcome with better education,
greater use and practice. For example, most remotely sensed imagery can be displayed
using colour combinations that more closely represent what the human eye sees, and
most imagery can be output as a hard copy. Now and in the future, more simple
techniques for delivering and manipulating digital data will become available that make
digital imagery easier to use. In terms of resolving the last two problems, it is necessary
to return to the amount of information required from an image, and this relates to the
issue to which the information held in the image is being applied. If, for instance, the
position and general quality of a section of sea wall was required, then aerial
photography would most likely provide the answer, as only the high spatial resolution of
aerial photography can really show such detail. However, if the total area of an estuary
or firth was needed, then a satellite image would be more appropriate to use, because far
more aerial photographs would be needed to cover such a large area, with no benefit in
information content. What this shows us is that people need to be made aware that all
remotely sensed imagery can have some use, as long as the correct imagery is applied in
the correct situation. Nevertheless, this requires a thorough understanding of the nature
of the problem being investigated, an understanding of the information required, and a
reference guide as to which imagery will provide that information. In essence, the issue
must be matched to the utility or characteristics of the sensor.
There are a whole range of issues, both natural and anthropogenic, which need
to be resolved at the coast. To match these issues to the imagery, one must understand
the information required to solve the issue and then the spatial, spectral and temporal
characteristics needed to provide the information. These characteristics can then be
compared with and matched to the spatial, spectral and temporal characteristics of the
different sensors. For example, a coastal issue might be the perceived increase in coastal
flooding due to saltmarsh erosion. The information required to investigate this issue
would be, where are the saltmarshes, are they eroding, which properties and land are
vulnerable? The objective would be to find imagery that could be used to identify
saltmarsh areas, analyse coastal change, and identify vulnerable property or land. To
identify the general pattern and location of saltmarsh areas in an estuary would usually
require the imagery to have a spatial resolution of at least 30m (Fig. 1) (this would
provide enough detail as well as area coverage). The spectral characteristics would not
be too significant in this case, as areas of saltmarsh can often be identified by pattern
and apparent texture. However, different band combinations can be used to highlight
certain vegetation and the vegetation/water boundary. A series of images over a period
of time would be needed to identify saltmarsh erosion. Having identified any areas of
possible erosion (Fig. 2), it would then be possible to look at these areas in more detail
using higher resolution imagery which could then be used to locate any land or property
in danger (Fig. 3). The information provided by such an analysis would then be
available for the coastal manager to formulate a management plan for this issue.