Johnson-Sea-Link II submersible. The submersible’s acrylic sphere provides
observers with a field of view [180°, while an external pan and tilt video camera
can record images along the track. In addition, the submersible is equipped with a
manipulator arm adapted to grab, scoop, or suck up target samples such as rock
slabs, sediments and organisms. The submersible navigates with an USBL positioning system that records the submersible’s real-time position every 4 s (Reed
et al. 2006). Analysis of USBL tracking accuracy for a worst-case tracking scenario estimated a maximum statistical positioning error of 9.6 m at a depth of
500 m (Opderbecke 1997). This positioning error is most easily accounted for and
corrected in sites that contain variable topographic relief. During data analysis,
regular changes in relief allow the observer to better connect actual submersible
position with the high-resolution DEM. Overall, video ground-truthing is performed with greatest confidence on habitat patches that have radii larger than the
statistical positioning error of the submersible (i.e., \9.6 m). Despite these positioning limitations of the submersible, the combination of the video ground-truthing with the AUV data permits cold-water coral habitats to be characterized in
terms of terrain, benthic fauna, sedimentary features, and environmental parameters at unprecedented resolution.
10.3.2 Survey Design and Data Analysis
The workflow for the surveyed sites described in this chapter consists of reconnaissance mapping, AUV survey, submersible ground-truthing, data processing,
and spatial and quantitative analyses of the integrated dataset. Reconnaissance
mapping was performed with hull-mounted multi-beam sonar systems, EM120 and
EM1002, which operate at 12 and 95 kHz frequencies, respectively (Table 10.1).
There is an inherent trade-off between operating the EM120 versus the EM1002.
The EM120 system covers a larger area due to its wider swath and faster survey
speeds, but produces a DEM map of lower resolution than the EM1002 (Fig. 10.2;
Table 10.1). Given this trade-off, the EM120 sensor was employed to explore
areas in which coral colonies and/or mounds had not been documented previously;
whereas in areas known to contain coral colonies and mounds, the EM1002 was
utilized to conduct more detailed reconnaissance surveys. Based on the reconnaissance maps, two sites in the Straits of Florida were selected for the C-Surveyor-II AUV deployment (Fig. 10.3). The sites are located at the base of the
Miami Terrace and on the slope of Great Bahama Bank (GBB), ranging in water
depths from 580 to 870 m (Fig. 10.3). Together these sites comprise 75 km
2 of
high-resolution mapped area, which were subsequently ground-truthed by a total
of six submersible transects (*8 km long).
The multi-beam datasets from the hull-mounted multi-beam systems were
rendered to produce DEMs of each study area. The EM120 and EM1002 multibeam systems generated maps with bin-size resolution of 50 and 20 m, respectively (Table 10.1). In contrast, the AUV-mounted EM2000 system produced a
10 Deep Acoustic Applications
261
observers with a field of view [180°, while an external pan and tilt video camera
can record images along the track. In addition, the submersible is equipped with a
manipulator arm adapted to grab, scoop, or suck up target samples such as rock
slabs, sediments and organisms. The submersible navigates with an USBL positioning system that records the submersible’s real-time position every 4 s (Reed
et al. 2006). Analysis of USBL tracking accuracy for a worst-case tracking scenario estimated a maximum statistical positioning error of 9.6 m at a depth of
500 m (Opderbecke 1997). This positioning error is most easily accounted for and
corrected in sites that contain variable topographic relief. During data analysis,
regular changes in relief allow the observer to better connect actual submersible
position with the high-resolution DEM. Overall, video ground-truthing is performed with greatest confidence on habitat patches that have radii larger than the
statistical positioning error of the submersible (i.e., \9.6 m). Despite these positioning limitations of the submersible, the combination of the video ground-truthing with the AUV data permits cold-water coral habitats to be characterized in
terms of terrain, benthic fauna, sedimentary features, and environmental parameters at unprecedented resolution.
10.3.2 Survey Design and Data Analysis
The workflow for the surveyed sites described in this chapter consists of reconnaissance mapping, AUV survey, submersible ground-truthing, data processing,
and spatial and quantitative analyses of the integrated dataset. Reconnaissance
mapping was performed with hull-mounted multi-beam sonar systems, EM120 and
EM1002, which operate at 12 and 95 kHz frequencies, respectively (Table 10.1).
There is an inherent trade-off between operating the EM120 versus the EM1002.
The EM120 system covers a larger area due to its wider swath and faster survey
speeds, but produces a DEM map of lower resolution than the EM1002 (Fig. 10.2;
Table 10.1). Given this trade-off, the EM120 sensor was employed to explore
areas in which coral colonies and/or mounds had not been documented previously;
whereas in areas known to contain coral colonies and mounds, the EM1002 was
utilized to conduct more detailed reconnaissance surveys. Based on the reconnaissance maps, two sites in the Straits of Florida were selected for the C-Surveyor-II AUV deployment (Fig. 10.3). The sites are located at the base of the
Miami Terrace and on the slope of Great Bahama Bank (GBB), ranging in water
depths from 580 to 870 m (Fig. 10.3). Together these sites comprise 75 km
2 of
high-resolution mapped area, which were subsequently ground-truthed by a total
of six submersible transects (*8 km long).
The multi-beam datasets from the hull-mounted multi-beam systems were
rendered to produce DEMs of each study area. The EM120 and EM1002 multibeam systems generated maps with bin-size resolution of 50 and 20 m, respectively (Table 10.1). In contrast, the AUV-mounted EM2000 system produced a
10 Deep Acoustic Applications
261