WORLD ATLAS OF CORAL REEFS
for reef mapping are the NASA Airborne Visible Infrared
Imaging Spectrometer (AVIRIS) and the Compact Airborne
Spectrographic Imager (CASI). The costs and the complex
processing requirements associated with these systems are
still the main constraints to their general use for coral reef
mapping. The minimum capital required just to obtain
the imagery is typically many tens, if not hundreds, of
thousands of dollars. At the close of the 1990s, however,
these represented the most accurate available systems for the
preparation of fine-scale coral reef maps.
Active sensors
Satellites and cameras are sometimes termed passive
sensors - they gather information from existing light as it is
reflected from objects on the ground. By contrast, a number
of active sensors are used in the marine environment which
direct their own source of light or sound towards their
subject and measure the reflection. Acoustic signals, or
sonar, are perhaps the best known. Sonar systems are
carried by ship-borne sensors. Using higher frequencies,
good spatial resolution can be achieved (1-4 meters), while
in shallow water they are unaffected by water turbidity or
depth. Sonar is used in most bathymetric mapping, and thus
is an important element of many reef maps, ahhough it does
not provide a direct measure of ecological features.
Light detection and ranging (LIDAR) is a lightbased form of remote sensing which involves emitting
pulses from an airborne laser and receiving energy which
has been reflected from both the water surface and submerged features. The time difference between the two
types of return provides a highly accurate (±15 centimeters)
measurement of depth. The result is an extremely high
resolution bathymetric chart. LIDAR is much less affected
by water clarity than normal optical sensors and in clear
conditions can operate to depths of about 50 meters. Like
sonar, LIDAR only maps the topography of the seabed,
not ecological features. Furthermore, because of the vast
amounts of data processing involved and the requirement
for specialist aircraft, it is yet to be used routinely in coral
reef mapping.
Ground-truthing
The process of producing a supervised classification of
an image is highly reliant on the correct interpretation.
Ultimately this is linked to a detailed sampling on the
ground, either directly by the cartographers, or using
existing information gathered by others in photographs,
maps or ecological surveys. Ground-truthing is an
expensive but critical element in preparing maps from
remote sensors. Natural variation in reefs between locations and over time is often significant and too much
extrapolation from previous work, or from work undertaken in other areas, can lead to substantial errors. At
the same time, ground-truthing can present considerable
opportunities for the refinement of individual maps,
greatly increasing the accuracy and allowing for the
discernment of particular features or habitats which might
not have been visible in other areas.
Remote sensing now dominates mapping in almost every
field, and is a critical component of reef mapping around
the world. Apart from specialist maps covering relatively
small areas, however, most existing maps which show reefs
are composite productions which may have been prepared
using satellite or aerial imagery in combination with
bathymetric data from sonar surveys and even with much
older data from early charts. A more widespread and rapid
updating of reef maps worldwide is limited by the high cost
of remote sensing and the detailed technical skills required
for image interpretation and map production.
Despite the power of remote sensing as a mapping
tool, there are also several practical constraints to its use.
Many of the world's reefs are located in the humid tropics
where cloud cover is frequent, which greatly restricts the
acquisition of images. The nature of mapping a submarine
feature also creates its own set of problems, not least of
which is the inability to map deeper reefs. Although the
depth limits vary it is rarely possible to map features more
than 20-30 meters below the ocean surface with conventional satellite imagery. Even above these depths, the water
column greatly affects the light returning to the sensor,
changing the spectial signatures of particular seabed
characteristics depending on depth and water clarity.
Although these effects can be partially compensated during
image processing they cannot be totally removed or
corrected. Moreover, the nature of the water column above
a reef, especially turbidity and depth, is highly variable.
Reef geometry, too, does not lend itself easily to being
mapped by remote sensing - few parts are flat and most
coral tends to be concentrated on steeply sloping edges.
One further weakness is that the remote sensing tools
used for seabed mapping are typically different from those
used to draw bathymetric maps. Many reef maps prepared
using remote sensing do not give detailed bathymetric
data, although these are clearly an important feature of
many reef maps.
The only alternative to remote sensing for mapping
coral reefs is the use of boat-based surveys to map surface
features such as reef crests, plotting bathymetry, or even
undertaking detailed sampling of the seabed. Unlike remote
sensing, which samples the entire seascape, errors arise in
this method because of the possibility of overlooking some
habitats between adjacent sampling points. Conversely, one
advantage of ground survey methods is that they allow for
the mapping of additional resources such as different
habitats or benthic species, which cannot be distinguished
by remote sensors because of similar reflectance patterns or
sparse distribution. Ground methods may also allow greater
for reef mapping are the NASA Airborne Visible Infrared
Imaging Spectrometer (AVIRIS) and the Compact Airborne
Spectrographic Imager (CASI). The costs and the complex
processing requirements associated with these systems are
still the main constraints to their general use for coral reef
mapping. The minimum capital required just to obtain
the imagery is typically many tens, if not hundreds, of
thousands of dollars. At the close of the 1990s, however,
these represented the most accurate available systems for the
preparation of fine-scale coral reef maps.
Active sensors
Satellites and cameras are sometimes termed passive
sensors - they gather information from existing light as it is
reflected from objects on the ground. By contrast, a number
of active sensors are used in the marine environment which
direct their own source of light or sound towards their
subject and measure the reflection. Acoustic signals, or
sonar, are perhaps the best known. Sonar systems are
carried by ship-borne sensors. Using higher frequencies,
good spatial resolution can be achieved (1-4 meters), while
in shallow water they are unaffected by water turbidity or
depth. Sonar is used in most bathymetric mapping, and thus
is an important element of many reef maps, ahhough it does
not provide a direct measure of ecological features.
Light detection and ranging (LIDAR) is a lightbased form of remote sensing which involves emitting
pulses from an airborne laser and receiving energy which
has been reflected from both the water surface and submerged features. The time difference between the two
types of return provides a highly accurate (±15 centimeters)
measurement of depth. The result is an extremely high
resolution bathymetric chart. LIDAR is much less affected
by water clarity than normal optical sensors and in clear
conditions can operate to depths of about 50 meters. Like
sonar, LIDAR only maps the topography of the seabed,
not ecological features. Furthermore, because of the vast
amounts of data processing involved and the requirement
for specialist aircraft, it is yet to be used routinely in coral
reef mapping.
Ground-truthing
The process of producing a supervised classification of
an image is highly reliant on the correct interpretation.
Ultimately this is linked to a detailed sampling on the
ground, either directly by the cartographers, or using
existing information gathered by others in photographs,
maps or ecological surveys. Ground-truthing is an
expensive but critical element in preparing maps from
remote sensors. Natural variation in reefs between locations and over time is often significant and too much
extrapolation from previous work, or from work undertaken in other areas, can lead to substantial errors. At
the same time, ground-truthing can present considerable
opportunities for the refinement of individual maps,
greatly increasing the accuracy and allowing for the
discernment of particular features or habitats which might
not have been visible in other areas.
Remote sensing now dominates mapping in almost every
field, and is a critical component of reef mapping around
the world. Apart from specialist maps covering relatively
small areas, however, most existing maps which show reefs
are composite productions which may have been prepared
using satellite or aerial imagery in combination with
bathymetric data from sonar surveys and even with much
older data from early charts. A more widespread and rapid
updating of reef maps worldwide is limited by the high cost
of remote sensing and the detailed technical skills required
for image interpretation and map production.
Despite the power of remote sensing as a mapping
tool, there are also several practical constraints to its use.
Many of the world's reefs are located in the humid tropics
where cloud cover is frequent, which greatly restricts the
acquisition of images. The nature of mapping a submarine
feature also creates its own set of problems, not least of
which is the inability to map deeper reefs. Although the
depth limits vary it is rarely possible to map features more
than 20-30 meters below the ocean surface with conventional satellite imagery. Even above these depths, the water
column greatly affects the light returning to the sensor,
changing the spectial signatures of particular seabed
characteristics depending on depth and water clarity.
Although these effects can be partially compensated during
image processing they cannot be totally removed or
corrected. Moreover, the nature of the water column above
a reef, especially turbidity and depth, is highly variable.
Reef geometry, too, does not lend itself easily to being
mapped by remote sensing - few parts are flat and most
coral tends to be concentrated on steeply sloping edges.
One further weakness is that the remote sensing tools
used for seabed mapping are typically different from those
used to draw bathymetric maps. Many reef maps prepared
using remote sensing do not give detailed bathymetric
data, although these are clearly an important feature of
many reef maps.
The only alternative to remote sensing for mapping
coral reefs is the use of boat-based surveys to map surface
features such as reef crests, plotting bathymetry, or even
undertaking detailed sampling of the seabed. Unlike remote
sensing, which samples the entire seascape, errors arise in
this method because of the possibility of overlooking some
habitats between adjacent sampling points. Conversely, one
advantage of ground survey methods is that they allow for
the mapping of additional resources such as different
habitats or benthic species, which cannot be distinguished
by remote sensors because of similar reflectance patterns or
sparse distribution. Ground methods may also allow greater