Extraction of bathymetry from synthetic aperture radar (SAR) is also a mature
strategy, competitively priced, and since the launch of TerraSAR-X, also of high
resolution. Marine DEM-derivation is, however, a complex undertaking and has
limitations. The caveat to retrieving water depth from SAR is that the image must
be acquired under favourable meteorological and hydrodynamic conditions;
moderate winds of 3-10 m/s and significant currents of about 0.5 m/s (Alpers and
Hennings 1984; Vogelzang et al. 1997). These criteria are necessary to ensure that
the waterbody is in motion over the seabed. This flow (typically tidally driven)
interacts with the bottom topography causing modulation of the surface current
velocity, which in turn give rise to local variations in surface wave patterns. An
over-flying SAR senses such variations through modulations in the backscattered
radar signal (Lyzenga 1991). As with the previously discussed optical derivation of
water depth, but unlike LiDAR, reference soundings are required to tune the SAR
extraction model. Suitable conditions for SAR imaging can be elusive and have to
date mitigated widespread use of the technology for resolving reef geomorphology. Nonetheless, in cases where favorable conditions are found, the potential of
SAR is heightened by its ability to image the sea-surface through cloud cover at
any time of the day or night. On land, and if the data have already been acquired,
extraction of topography from stereo-paired satellite images or aerial photographs
yields an accurate DEM at a cost many orders cheaper than a LiDAR survey. For
shallow water at least, this technique is transferable to the marine realm, but due to
refraction at the air–water interface, the calculations are complex and the methodology has so far not been applied extensively (Murase et al. 2008).
Given the lack of accuracy and necessity of considerable ground-control for
deriving DEMs from satellite imagery, be it SAR or visible-spectrum, LiDAR has
become the technology of choice for bathymetric mapping. It delivers a reliable,
fast, and accurate representation of a reefscape through geo-referenced 3-D point
clouds. Whether terrestrial or submarine, LiDAR-based measurements are commonly used for three families of information: (1) the production of bare-earth
models, (2) DEM generation, and (3) geomorphology studies. The latter two have
most relevance for coral reef research, but for completeness all three will be
covered.
Many applications, for example contouring, require a bare-earth digital terrain
model (DTM). Unfortunately, the raw data points captured by LiDAR do not
constitute a bare-earth DTM. Even though most LiDAR systems can measure ‘‘last
return’’ data points, in terrestrial environments these often measure ground
parameters like shrubbery, cars, buildings, and the canopy of dense foliage. For
some applications, raw LiDAR points must therefore be post-processed to remove
these returns as they are considered undesirable. In others cases, such as terrestrial
bare earth mapping, the last returns are valuable and desired by engineers, scientists, and researchers interested in buildings, infrastructure, and forest canopy,
among other land-cover mapping applications. In the marine realm, the problem of
surface clutter is less pronounced since submerged aquatic vegetation has less
vertical relief than terrestrial foliage. Even so, if the data are to be used for
5 LiDAR Overview
129
strategy, competitively priced, and since the launch of TerraSAR-X, also of high
resolution. Marine DEM-derivation is, however, a complex undertaking and has
limitations. The caveat to retrieving water depth from SAR is that the image must
be acquired under favourable meteorological and hydrodynamic conditions;
moderate winds of 3-10 m/s and significant currents of about 0.5 m/s (Alpers and
Hennings 1984; Vogelzang et al. 1997). These criteria are necessary to ensure that
the waterbody is in motion over the seabed. This flow (typically tidally driven)
interacts with the bottom topography causing modulation of the surface current
velocity, which in turn give rise to local variations in surface wave patterns. An
over-flying SAR senses such variations through modulations in the backscattered
radar signal (Lyzenga 1991). As with the previously discussed optical derivation of
water depth, but unlike LiDAR, reference soundings are required to tune the SAR
extraction model. Suitable conditions for SAR imaging can be elusive and have to
date mitigated widespread use of the technology for resolving reef geomorphology. Nonetheless, in cases where favorable conditions are found, the potential of
SAR is heightened by its ability to image the sea-surface through cloud cover at
any time of the day or night. On land, and if the data have already been acquired,
extraction of topography from stereo-paired satellite images or aerial photographs
yields an accurate DEM at a cost many orders cheaper than a LiDAR survey. For
shallow water at least, this technique is transferable to the marine realm, but due to
refraction at the air–water interface, the calculations are complex and the methodology has so far not been applied extensively (Murase et al. 2008).
Given the lack of accuracy and necessity of considerable ground-control for
deriving DEMs from satellite imagery, be it SAR or visible-spectrum, LiDAR has
become the technology of choice for bathymetric mapping. It delivers a reliable,
fast, and accurate representation of a reefscape through geo-referenced 3-D point
clouds. Whether terrestrial or submarine, LiDAR-based measurements are commonly used for three families of information: (1) the production of bare-earth
models, (2) DEM generation, and (3) geomorphology studies. The latter two have
most relevance for coral reef research, but for completeness all three will be
covered.
Many applications, for example contouring, require a bare-earth digital terrain
model (DTM). Unfortunately, the raw data points captured by LiDAR do not
constitute a bare-earth DTM. Even though most LiDAR systems can measure ‘‘last
return’’ data points, in terrestrial environments these often measure ground
parameters like shrubbery, cars, buildings, and the canopy of dense foliage. For
some applications, raw LiDAR points must therefore be post-processed to remove
these returns as they are considered undesirable. In others cases, such as terrestrial
bare earth mapping, the last returns are valuable and desired by engineers, scientists, and researchers interested in buildings, infrastructure, and forest canopy,
among other land-cover mapping applications. In the marine realm, the problem of
surface clutter is less pronounced since submerged aquatic vegetation has less
vertical relief than terrestrial foliage. Even so, if the data are to be used for
5 LiDAR Overview
129