Mid-Ocean Ridge Exploration with an Autonomous Underwater Vehicle

Home , Oceanography

or collective redistirbution other or collective or means this reposting, of machine, is bySociety, photocopy article only anypermitted of withSociety.Send portion theall approvalOceanographycorrespondence P Oceanography to: [email protected] The of or Th e This article has been published in published been This has article S p e c i a l i S S u e O n O c e a n E x p l o r at i on

B y D a n a R . y O e r g e r , A l b e r t M . B r a d l e y,

Michael Jaku b a , Oceanography M a u r i c e A . T i v e y,

C h r i s to p h e r R . G e r m a n , , VolumeSociety. 4, a quarterlyOceanography 20, Number C The journal of T i mo t h y M . S h a n k , a nd r O B e r t W. e M B l e y

Mid-Ocean Ridge exploration with an Society.Oceanography A opyright 2007 by The Autonomous Underwater

Vehicle P reserved. ll rights

Human-occupied submersibles, towed sors, and cameras in the rugged deep-sea preclude acoustic contact with the sup- vehicles, and tethered remotely operated terrain that has been the focus of numer- port vessel. Unlike towed assets, ABE ermission is granted to in teaching copy this and research. for use article R vehicles (ROVs) have traditionally been ous scientific expeditions (e.g., those to can travel close to the seafloor along used to study the deep seafloor. In recent mid-ocean ridges and ocean margin set- well-controlled tracklines, enabling high- years, however, autonomous underwater tings). The Autonomous Benthic Explorer resolution seafloor imaging through a vehicles (AUVs) have begun to replace (ABE) is an example of an AUV that has variety of modalities. ABE generally col- these other vehicles for mapping and been used for over 20 cruises sponsored lects co-registered bathymetric and mag- survey missions. AUVs complement the by the National Science Foundation netic data while determining its position capabilities of these pre-existing systems, (NSF), the National Oceanic and through a combination of long-baseline offering superior mapping capabilities, Atmospheric Administration (NOAA) acoustic transponders, a Doppler veloc- improved logistics, and better utilization Office of Ocean Exploration (OE), and ity log, and a fluxgate magnetic compass

of the surface support vessel by allowing international and private sources. This (Yoerger et al., 2007). Equally important, 1931, R O Box other tasks such as submersible opera- paper summarizes NOAA OE-sponsored ABE also measures water-column prop- tions, ROV work, CTD stations, or mul- cruises made to date using ABE. erties using dual conductivity/tempera- ockville, MD 20849-1931, U ockville, epublication, systemmatic reproduction, tibeam surveys to be performed while the ABE commonly operates with full ture probes, an optical backscatter sen- AUV does its work. AUVs are particularly autonomy; after launch, the vehicle sor, and a redox potential probe. well suited to systematic preplanned sur- completes its mission without human Since its initial trials in 1994, ABE has veys using sonars, in situ chemical sen- intervention, often at distances that completed 210 deep-ocean dives, cover- S A .

52 Oceanography Vol. 20, No. 4 ing over 3600 km of bottom tracks at an 2003). Alvin safety considerations pre- and extinct hydrothermal sites (Tivey average depth exceeding 2000 m (Yoerger cluded simultaneous operations, so ABE and Johnson, 2002). et al., 2007). The mid-ocean ridge has averaged about 7.5 hours of bottom Near 86°W, ABE made four dives at been the focus of most of these efforts, time over seven dives, covering a total of a nominal survey height of 40 m and particularly sites with hydrothermal 107 km of bottom tracks. During these a trackline spacing of 60 m to produce activity. Some cruises featured detailed dives, ABE collected high-resolution a bathymetric map of the rift valley. study of previously discovered sites, while bathymetric data using a 675-kHz scan- This survey included the Rosebud site other cruises focused on unexplored ning sonar with a nominal pixel resolu- discovered by scientists in Alvin on the areas. NSF funded initial ABE develop- tion of 2–5 m with concomitant record- same cruise. ABE also identified sev- ment and field deployments. Since 2002, ing of conductivity and temperature data. eral smaller vents, primarily through five cruises have been sponsored by the One key objective of our 2002 cruise temperature anomalies on the order of NOAA OE program, accounting for was to revisit the Rose Garden site that 50 millidegrees that were later investi- about one-third of ABE’s at-sea time. was first discovered in 1979 by scientists gated using Alvin. CTD casts and tow- During these cruises, we made many in Alvin and revisited in 1988 and 1990 yo runs showed the water column to be important additions to ABE’s technical (Hessler et al., 1988). This site’s lush well stratified near the seafloor, which Mid-Ocean Ridge capabilities. Those funded by NOAA OE faunal communities that featured dense made the small temperature anomalies include the addition of a multibeam tubeworm clusters made it an icon of caused by the vent plumes conspicuous. mapping sonar, installation of a Doppler hydrothermal vent research. We found, In each case, the combination of ABE’s exploration with an velocity log, and development of an however, that the Rose Garden site had bathymetric mapping and delineation anchoring system to allow ABE to “park” been paved over by lava flows since the of water-column anomalies enabled sci- itself on the seafloor at the end of a dive last visit by Alvin. No sign of the vent entists in Alvin to find the vent. These Autonomous Underwater or in the event of a serious fault. This site was found by either ABE or Alvin, exercises illustrated the value of an AUV- capability greatly improves our ability nor did we find any signs of previous based vent search, and showed that tem- to operate ABE unattended, freeing up dives such as markers or dive weights. perature anomalies encountered at our Vehicle the vessel for other work. It also greatly The absence of temperature anomalies in nominal survey height could be used to improves ABE’s ability to work simulta- ABE data helped confirm that our inabil- find vents, at least under favorable con- neously with ROVs.

Galápagos 2002 Expedition unlike towed assets, ABE can travel close The goals of our Galápagos expedition were to revisit the sites where to the seafloor along well-controlled hydrothermal vents were first discov- tracklines, enabling high-resolution seafloor ered in 1977 and 1979 in the vicinity of 0°48´N, 86°13´W and to study how imaging through a variety of modalities. those sites changed in the ensuing years. We also surveyed unexplored sites near 89°W. This cruise was supported by ity to detect hydrothermal vent features ditions. Figure 1 shows the assembled the NOAA OE program, NSF, and the from Alvin was not due to navigational bathymetric map, the locations of the Woods Hole Oceanographic Institution. error. These conclusions were also sup- measured temperature anomalies, and We operated ABE cooperatively with ported by ABE magnetic data, which the locations of the vent sites. DSV Alvin, with ABE performing sys- showed a reduced zone of magnetism at The second site we visited on the tematic mapping at night while the Alvin the paved-over Rose Garden site (Shank Galápagos Spreading Center, near 89°W, dives focused on detailed inspection and et al., 2003). Such reduced zones of mag- taught us more important lessons. Three sampling during the day (Shank et al., netism have been associated with active dives in this area enabled production

Oceanography December 2007 53 Figure 1. A bathymetric view of the floor of the Galápagos Spreading Center rift valley at 86°W obtained with a scanning sonar and gridded at 5 m. The red crosses indicate locations where temperature anomalies were detected byABE while surveying ~ 40 m above the seafloor, including the nascent Rosebud vent and the small ALR vent. The highly stratified water column made temperature anomalies as small as ~20 millidegrees conspicuous. No sign of hydrothermal activity was found near the Rose Garden site. ABE data showed a magnetic low at the reported Rose Garden location, a result consistent with previous hydrothermal activity.

of another bathymetric map, which While the ~ 2450-m deep 86°W site is 86°W, and we could find no correlation included the Calyfield vent site that located in the rift valley, the 89°W area between temperature anomalies seen in had been found using Alvin. Unlike the lies near a topographic high in less than the ABE data and the Calyfield site. 86°W site, the Calyfield site could not 1700 m of water on the flank of the rift. We concluded that our vent-searching be clearly distinguished through tem- CTD data showed that the near-bottom technique required additional sens- perature anomalies in the ABE data. hydrography was more complex than at ing, and fortunately our collaboration with Ko-ichi Nakamura of the Japanese Dana R. Yoerger ([email protected]) is Associate Scientist, Department of Applied National Institute of Advanced Industrial Ocean Physics and Engineering, Woods Hole Oceanographic Institution, Woods Hole, MA, Science and Technology provided USA. Albert M. Bradley is Principal Engineer (Retired), Department of Applied Ocean just such a sensor: the redox poten- Physics and Engineering, Woods Hole Oceanographic Institution, Woods Hole, MA, USA. tial (Eh) probe, which we had first run Michael Jakuba is Postdoctoral Investigator, Department of Applied Ocean Physics and on ABE in 2000. Engineering, Woods Hole Oceanographic Institution, Woods Hole, MA, USA. Maurice A. Tivey is Associate Scientist, Department of Geology and Geophysics, Woods Hole Explorer Ridge 2002 Oceanographic Institution, Woods Hole, MA, USA. Christopher R. German is Senior Expedition Scientist, Department of Geology and Geophysics, and Chief Scientist for Deep Submergence, We investigated the southern Explorer Woods Hole Oceanographic Institution, Woods Hole, MA, USA. Timothy M. Shank Ridge, Northeast Pacific Ocean, in order is Associate Scientist, Department of Biology, Woods Hole Oceanographic Institution, to locate the actively venting Magic Woods Hole, MA, USA. Robert W. Embley is Senior Research Scientist, Pacific Marine Mountain hydrothermal field that was Environmental Laboratory, National Oceanic and Atmospheric Administration (NOAA), originally discovered in 1985 and last Newport, OR, USA. visited by submersible in 1986. The site

54 Oceanography Vol. 20, No. 4 Figure 2. Bathymetry results from the summit of the Southern Explorer Ridge using a multibeam sonar on ABE. The multibeam data permitted quantitative assessment of the faulting of the axial summit graben floor and provided new insights into the formation of the Magic Mountain site, shown in the expanded view. Contradictory to previous reports, the vents were located outside the axial rift graben adjacent to the graben’s large eastern bounding fault. The plots also show the location of collocated redox potential signals and temperature anomalies that provided a reliable indication of hydrothermal activity even in complex terrain in the presence of significant currents.

was not fully characterized in the previ- vehicles, we could operate ABE with- soundings that allowed us to create a grid ous expeditions and its exact location out outside scheduling constraints. of 1-m pixels and in some places a 25-cm and tectonic setting were unknown. During the seven ABE dives, we con- pixel size. This resolution level collected This expedition featured a new capa- ducted simultaneous operations, such over an area of several kilometers pro- bility for ABE, the Simrad SM2000 mul- as CTD casts and hull-mounted multi- vided a unique opportunity to investigate tibeam sonar, which we demonstrated beam mapping, from the research vessel the faulting that is so prominent in this under NOAA OE sponsorship. The Thomas G. Thompson. During the course area (Deschamps et al., 2007). experimental setup used in 2002 was of the cruise, ABE covered 181 km of Prior to this study, most analyses of later integrated fully into ABE. Using bottom tracks over those seven dives. seafloor faulting used ship-based mul- the SM2000, we were able to obtain On the last two dives, the SM2000 failed tibeam bathymetry and individual pro- 128 beams over a 90° swath from survey due to a problem in the logging software. files made from deep-towed bathymetry, heights ranging from 50 to 150 m. This ABE continued to collect soundings from which have limited resolution, especially instrument represented more than an its scanning sonar that provided lower- for distinguishing between volcanic- order of magnitude improvement over resolution bathymetry. The SM2000 log- and fault-controlled topography and the scanning sonar used on ABE previ- ging problems were readily resolved dur- resolving processes such as fault link- ously. We also used the Nakamura redox ing the subsequent integration effort. age. ABE’s high-resolution multibeam potential (Eh) probe on ABE during this Figure 2 shows results from the north- data provided a much more complete cruise to aid in identifying localization of ern portion of the survey area, which was characterization of fault geometry over active vent sites. fully covered with the SM2000 multi- a sufficiently wide area to be statisti- Because we had no requirement to beam sonar. The multibeam sonar pro- cally relevant. Contrary to the prevailing coordinate operations with any other duced a “point cloud” of bathymetric paradigm of seafloor faulting based on

Oceanography December 2007 55 subaerial fault systems, the ABE results On a cruise immediately following the crest immediately south of the Chain demonstrated that the ratio of fault ABE cruise, our collaborators used the and Romanche Fracture Zones in par- length to fault height is not constant, bathymetry and water-column anomaly ticular, to test a hypothesis that these which highlights the importance of fault maps of the Magic Mountain site to fracture zones, assumed to be hydrother- linkage and fault growth within the rela- guide ROV dives using the Canadian mally barren, might act to separate two tively thin, brittle layer of oceanic crust ROPOS system (Embley, 2002). distinct biogeographic provinces of vent (Deschamps et al., 2007). Faults thus fauna from the northern and southern grow by coalescing rather than by propa- Southern Mid-Atlantic sections of the Mid-Atlantic Ridge. We gation and are probably limited by the Ridge, 2005 and 2006 already knew that the fauna of the north- depth of the brittle-ductile transition. Working together with the UK’s ern Mid-Atlantic Ridge were quite dis- Collocated anomalies of temperature Southampton Oceanography Centre tinct from those of the Pacific and Indian and redox potential in combination with (SOC) in 2005, we were able to locate the Oceans but, with no known vent sites bathymetry allowed us to pinpoint and first vent sites anywhere in the southern having been discovered in the Atlantic document the Magic Mountain site. To Atlantic Ocean. Based on our prelimi- Ocean south of the equator, hypotheses our surprise, we found the field outside nary investigations, the first conducted on what controlled these biological “dis- of the axial rift graben and adjacent to by ABE from a non-US research ves- connects” remained untestable as well as its large eastern bounding fault. The sel, we were able to pinpoint new vent untested. Thus, these cruises represented fault association is likely an important sites and photograph both the vents an important convergence of ideals aspect in the longevity of the vent field. and animals. Based on the resulting between NOAA’s OE Program and a key The large size and geometry of the fault data, we were able to guide a German component of the Census of Marine Life suggests that it probably extends to interdisciplinary team equipped with ChEss Program, which is dedicated to the brittle-ductile transition and likely their Center for Marine Environmental understanding vent-faunal biogeography and biodiversity. Our 2005 cruise highlighted the role of an AUV in a fundamental exploration Working together with the UK’s Southampton effort. The cruise was split into two legs Oceanography Centre (SOC) in 2005, we were aboard the UK’s Natural Environment Research Council RRS Charles Darwin. able to locate the first vent sites anywhere On the first leg of the cruise, we started in the southern Atlantic Ocean. work by mapping the seafloor along a previously unexamined section of the ridge crest using the ship’s EM120 multibeam system. This survey was fol- provides a robust and long-lived path- Sciences (MARUM) ROV Quest directly lowed by deep-tow surveys using the way for hydrothermal fluid to reach the to these vent sites just three weeks after Southampton Oceanography Centre’s actively venting site. we had first mapped the seafloor in this 30-kHz side-scan sonar, TOBI (Towed In addition to the Magic Mountain area. Impressed with the efficiency of Ocean Bottom Instrument), to image site, ABE located another likely vent site this novel approach, we were invited the underlying seafloor. TOBI was aug- about 3 km to the southeast of Magic back to the area aboard the German ship mented with a series of in situ MAPR Mountain. Multiple simultaneous redox- FS Meteor in June 2006, where we con- (miniature autonomous plume recorder) potential readings and temperature tinued to find new vents in three sepa- units to detect hydrothermal plumes in anomalies as high as 150 millidegrees rate areas, at 5°S, 9°33´S, and at 8°18´S. the overlying water column. The com- provided solid evidence of active vent- We investigated the southern Mid- bination of multibeam, side-scan, and ing. This site remains unexplored. Atlantic Ridge in general, and the ridge- water-column data from TOBI converged

56 Oceanography Vol. 20, No. 4 to identify two sites of active hydrothermal venting between 2° and 5°S (German et al., in press). With ABE aboard ship for the second leg of the cruise, we returned to 5°S where all the data pointed to a saddlelike area in the center of the ridge segment. At a depth of nearly 3000 m, the saddle comprised the shallowest point of the ridge axis in this area. As we prepared ABE for deployment, follow-up CTD investigations of the water column indicated the presence of at least three nonbuoyant plumes, implying the pos- sible presence of multiple active vent sites within a few kilometers. (Based on this work, S.A. Bennett and colleagues have prepared a paper on the distribution and stabilization of dissolved Fe in deep-sea Figure 3. ABE tracklines for dives at the 5°S site on the southern Mid-Atlantic Ridge in 2005 hydrothermal plumes.) Consistent with and 2006. The initial Phase 1 survey (in red) in 2005, which used very wide trackline spacing water-column observations, the TOBI and alternated depths, provided clues leading to the discovery of three hydrothermal sites, side-scan data showed a smooth, unfrac- including one diffuse field (Wideawake) and two black smoker sitesT ( urtle Pits and Red Lion). In 2006, a tighter Phase 1 dive confirmed the existence of a vent site to the east. The more tured area consistent with geologically detailed and closer up Phase 2 and Phase 3 dives that followed led to the discovery of a black fresh lava flows extending over an area of smoker site (Sister’s Peak) and two diffuse fields (Golden Valley and Foggy Corner). 18 km2 (German et al., in press). To track the nonbuoyant plume signals to their sources on the seafloor, we used ABE to execute a three-phase fluids while mapping the appropriate survey. Tracklines were spaced at 1 km strategy first demonstrated on an NSF region of the seafloor. Phase 2 surveys with alternating depths of 2750 and Ridge2000 cruise to the Lau Basin in are conducted at a height of 50 m above 2875 m—the depths of the two strongest 2004 (Yoerger et al., 2006; German et the seafloor with a trackline spacing of nonbuoyant plumes we had encountered al., 2007). The approach starts with a 30 m. In the best case, the vehicle will from CTD investigations. Fortunately, constant-depth survey within a dispers- be forced up when it passes through the two consecutive Phase 1 lines showed ing nonbuoyant hydrothermal plume. rising plume stem on at least two adja- solid redox potential signals, some of The appropriate depth to intercept the cent tracklines, in addition to registering which correlated with increased opti- nonbuoyant plume is first determined anomalies in the quantities measured in cal backscatter. Based on the position by CTD. Phase 1 survey lines are usually Phase 1. A favorable Phase 2 result usu- of the largest Phase 1 anomalies, we planned several hundred meters apart. ally allows a new source of venting to be planned a Phase 2 dive using our stan- Anomalies in temperature, optical back- located to within < 100 m. In the Phase 3 dard line spacing (30 m) and survey scatter, and redox potential from the survey, the vehicle takes photos from a height (50 m). This dive yielded solid Phase 1 survey are then used, together height of about 5 m to pinpoint the vent indications of vent sites beneath one of with current-meter data, to predict the site and to identify fauna. our Phase 1 tracklines, near the southern source area on the underlying seafloor. Our time at the 5°S site in 2005 was extent of our Phase 2 survey (later identi- Based on the Phase 1 results, we design very limited, so we employed a some- fied as the Turtle Pits black smoker site Phase 2 surveys to intercept rising plume what risky strategy for our initial Phase 1 and Wideawake diffuse venting field).

Oceanography December 2007 57 Remarkably, however, ABE also encoun- The following year, we returned to the 2005, our photographs showed that the tered very strong anomalies beneath 5°S site on a cruise in partnership with lavas in this area were fresh and glassy another of our Phase 1 tracklines. These the same MARUM Quest ROV team and and in some areas they had flowed over were apparent on the northernmost our scientific colleagues from Germany. the top of Wideawake vent fauna. They Phase 2 trackline (later identified as the For most of the cruise, ABE and Quest were not just geologically young but Red Lion black smoker site). In essence, alternated time in the water, although we genuinely recent, perhaps years or even our Phase 2 dive identified not one but also made some concurrent dives. This months old. Dating of the lavas is under- two different high-temperature hydro- collaboration proved to be very efficient way to find the exact timing of eruption thermal fields, spaced more than 1 km and productive; in several instances, while intercomparison of ABE photo- apart, at the southern and northern Quest dove directly onto targets a few graphs from 2005 and 2006 will allow extremes of the survey grid. A follow-up hours after they were discovered by us to examine time-series changes in Phase 3 dive at the southern site veri- ABE. For example, ABE arrived on deck the biological communities discovered fied the presence of active vents through late one Sunday morning and, within the year before. water-column measurements, but a 12 hours, Quest had dived to the seafloor Further south, we made two ABE flooded camera cable prevented any pho- and investigated three new vent sites dives in the Lilliput area (9°33´S) where tographs from being gathered. While the pinpointed during ABE’s just-completed Quest found two diffuse hydrother- camera was being repaired, we expanded dive. Although the vents were spread mal fields in 2005. A Phase 2 ABE dive the Phase 2 coverage to the north and out over an area of ~ 200 m in diam- yielded a bathymetric map of the site, east. This dive confirmed the presence eter, navigation fixes for ABE and Quest provided an improved geological context of a vent site to the north (the Red Lion were consistent to within 1–5 m on the for the 2005 Quest discoveries, and col- black smoker site) and provided prelimi- seafloor. This strategy greatly improved lected redox potential signals that pro- nary evidence of vent sites to the east the efficiency ofQuest operations and vided strong indications of several new (confirmed in the next year’s cruise). improved sample yield. vent sites. Our second dive in this area, Following repairs to ABE’s camera Our return to the 5°S site began with a Phase 3 dive, provided photos of four system, we ran two more Phase 3 sur- a more focused Phase 1 survey to detect newly discovered diffuse vent sites to add veys, one at Turtle Pits and Wideawake, any plumes that might have been missed to the two already known in this ridge and another at Red Lion. Diffuse-flow during the loose Phase 1 survey done the segment (Koschinsky, 2006). areas were found at Wideawake and previous year. The new survey provided ABE made a three-phase survey in black smokers were imaged at Turtle further evidence of venting between the Niebelungen area, 8°18´S, searching Pits and Red Lion. The monochrome the Turtle Pits and Red Lion sites and for vents that had been detected by our photos revealed a variety of animals, indicated another vent site to the west. German colleagues in 2005. Although including mussels, clams, and shrimp. A series of Phase 2 and Phase 3 dives the plume was extensive, complex cur- Approximately three weeks later, our pinned down the site between the two rents and rough terrain made the search German colleagues used our maps to previously discovered sites. The area was difficult. The vent source remained guide dives to Turtle Pits, Wideawake, called Comfortless Cove and contained undiscovered despite several Quest dives and Red Lion fields with the MARUM a black smoker site (Sister’s Peak) and dedicated to finding it in 2005. Following Quest ROV. They took close-up two diffuse-flow sites (Golden Valley our three-phase search method, we were images and collected geological, bio- and Foggy Corner). Quest visited these successful in locating a black smoker logical, and fluid samples (Haase et al., sites shortly after they were detected (der Drachenschlund, or The Dragon’s 2007) that were collaboratively pro- by ABE to make more detailed surveys Throat) on the third ABE dive. The vent, vided to our team to test hypotheses and take samples. Another Phase 3 dive a single black smoker, was located in a related to whether or not the southern revisited the Wideawake and Turtle Pits crater on a steep slope and then surveyed Mid-Atlantic Ridge represents a new fields first imaged in 2005; Figure 4 is by Quest a few hours after ABE returned vent faunal province. a photomosaic from Wideawake. In to the surface. The previous year, the

58 Oceanography Vol. 20, No. 4 Quest ROV had approached to within surveyed 161 kilometers over eight dives Figure 4. Photomosaic assembled from less than 200 m of the vent, but investi- and spent nearly 20 hours “sleeping” on five electronic still images taken from ABE at a height of about 5 m. Recorded gators had turned away because the geo- the seafloor while other activities were in the diffuse Wideawake field, the logic setting appeared unlikely to sustain completed. We conducted one joint dive images show a dense field of mussels (Bathymodiolus), clams (Calyptogena), high-temperature venting (Koschinsky, with the Quest6000, launching ABE after conid snails (likely Phymorhynchus sp.), 2006), and they did not have three-phase the ROV reached the seafloor. and limpets (Neolepetopsidae) among data such as that provided by ABE. ABE completed three kinds of survey relatively recently erupted lobate lava flows. Dating of the lavas is underway. We also made Phase 1 and Phase 2 lines while mapping the Brothers cal- ABE dives at 7°57´S, but no vents were dera. Most were focused on bathymetry discovered before our on-station at a survey height of 50 m, with simulta- time ended. neous water-column and magnetic-field mapping. Based on existing multibeam Kermadec Arc maps (EM300) and observations from Under NOAA OE sponsorship in 2007, ROVs and submersibles, we knew the we joined colleagues from the United caldera walls had a slope of about 45° on States, New Zealand, and Germany average, but included many steep faces to study Brothers volcano, an intra- that would not show up in the existing oceanic arc submarine volcano located multibeam maps. Tracklines were pro- 310 km northeast of New Zealand in the grammed to follow along-contour based Kermadec arc (de Ronde et al., 2005). on the existing bathymetry, with spacing Brothers is one of the most studied sub- of approximately 60 m to ensure overlap marine arc volcanos, but had never been in the steep terrain. ABE also made two mapped at the fine scale that can be intensive temperature surveys with very achieved with an AUV. closely spaced tracklines (10 m) over This effort combined many types of known vent sites and ran a few single activities, including test dives for the lines to collect magnetic field data. new University of Kiel Research Center Figure 5 shows the tracklines from all for Marine Geosciences Quest6000 dives in the upper panel and the assem- ROV, deep-water testing of the ROV bled bathymetry in the bottom panel. winch, CTD casts and tow-yos, and ves- The lower parts of the caldera map were sel multibeam mapping. Many of these filled using EM300 data (region out- activities were distant from the survey lined in black). The areas with many site and were conducted while ABE closely spaced lines over the smaller cone surveyed the seafloor out of acoustic and on the northern face indicate the range of the vessel. The anchoring sys- regions where the temperature surveys tem was critical for this cruise because were performed; the long isolated lines it allowed ABE to operate safely without show the locations of the magnetic- interfering with the other cruise activi- field data surveys. ties. Because ABE can anchor when its This bathymetric map provides what is batteries are depleted or in the event of probably the most detailed, overall view a critical fault, the vessel can leave the of a submarine arc volcano produced site with no requirement to return for at to date. The large smooth cone is prob- least 48 hours, which greatly facilitates ably the site of recent volcanic eruptions. scheduling the other tasks. Overall, ABE Temperature, redox potential, and optical

Oceanography December 2007 59 backscatter measurements confirmed the presence of hydrothermal activity from the small crater at the top of the cone. The smaller, more-weathered cones on the flanks of the large cone are probably older but also show hydrothermal activity that extends to the caldera’s floor. The rough, eroded topography on the caldera walls has likely been undermined by hydrothermal activity. ABE dives delin- eated the extent of the previously known vent sites, identified at least one new site based on both water-column and magnetic-field measurements, and located an extinct field through seafloor magnetics.

Conclusion This paper summarizes a series of oceanographic cruises using the Autonomous Benthic Explorer (ABE) and related technological developments supported by the NOAA Ocean Exploration Program between 2002 and 2007. These cruises made up about one-third of ABE’s sea time over this period, with other sea- going efforts sponsored by the National Science Foundation and by private and international sources. On these cruises, ABE made several different types of surveys and operated simultaneously with a number of other shipboard operations such as CTD work, vessel multibeam mapping, ROV dives, and tow-sled operations. Fine-scale, near-bottom bathymetric measurements made by ABE with a mul- Figure 5. The upper panel shows the tracklines used to map Brothers volcano in the Kermadec arc. The majority of the lines were chosen to provide bathymetric and near- tibeam sonar produce high-resolution bottom magnetic data while following along-contour based on previously collected maps that enable detailed analyses of EM300 multibeam bathymetry. The vehicle flew at a height of ~50 m above the seafloor active hydrothermal sites and permit and the nominal trackline spacing was 60 m to avoid data gaps in the steep terrain. The lower panel shows the resulting bathymetry gridded at 2 m, with the center section quantitative analyses of phenomena such (outlined in black) filled with EM300 data. as seafloor faulting. The high spatial resolution allows volcanic topography to be distinguished from fault-controlled topography and permits key processes

60 Oceanography Vol. 20, No. 4 such as fault linkage to be quantified critical to ABE’s success because they References with statistical certainty. Ocean-crust substantially improved the vehicle’s sci- de Ronde, C.E.J., M.D. Hannington, P. Stoffers, I.C. Wright, R.G. Ditchburn, A.G. Reyes, E.T. Baker, G.J. fault models based on ABE bathymetry entific mapping capabilities and its abil- Massoth, J.E. Lupton, S.L. Walker, and others. 2005. contradict earlier studies made from ity to work productively under a variety Evolution of a submarine magmatic-hydrothermal system: Brothers Volcano, Southern Kermadec Arc, lower-resolution, ship-based, multibeam of cruise constraints. New Zealand. Economic Geology 100:1,097–1,133. data and also show that mechanisms for Deschamps, A., M. Tivey, R.W. Embley, and W.W. Chadwick. 2007. Quantitative study of the defor- fault linkage and fault growth on the Acknowledgements mation at Southern Explorer Ridge using high- seafloor differ significantly from typical We are grateful to the crews of the resolution bathymetric data. Earth and Planetary subaerial systems. research vessels Atlantis, Thomas Science Letters 259(1–2):1–17. Embley, R.W. 2002. Rediscovery and exploration of The combination of bathymetry and G. Thompson, Charles Darwin, Meteor, Magic Mountain, Explorer Ridge, NE Pacific. Eos magnetic-field measurements from an and Sonne for their assistance in ABE Transactions, American Geophysical Union, 83(47), Fall Meeting Supplement, Abstract T11C-1264. AUV such as ABE provides estimates operations during these NOAA Office German, C.R., S.A. Bennett, C. Boulart, D.P. Connelly, of crustal magnetization, which in turn of Ocean Exploration and Research- A.J. Evans, B.J. Murton, L.M. Parson, R.D. Prien, E. Ramirez-Llodra, M. Jakuba, and others. In permits the age and thickness of lava sponsored cruises. Alan Duester, Andrew press. Hydrothermal activity on the southern flows to be determined. Combined maps Billings, and Rodney Catanach pro- Mid-Atlantic Ridge: Tectonically and volcanically controlled venting at 4–5°S. Earth and Planetary made by ABE have been used to identify vided critical engineering and logisti- Science Letters. volcanic features such as lava flow units, cal support for ABE operations. Ko-ichi German, C.R., D.R. Yoerger, M. Jakuba, T.M. delimit their fronts, and estimate their Nakamura of the National Institute Shank, C.H. Langmuir, K. Nakamura. 2007. Hydrothermal exploration with the Autonomous thicknesses. Additionally, as active and of Advanced Industrial Science and Benthic Explorer. Deep Sea Research Part I extinct hydrothermal sites may show Technology in Japan has been an invalu- doi:10.1016/j.dsr.2007.11.004. Haase, K.M., S. Petersen, A. Koschinsky, R. Seifert, low crustal magnetization, these maps able collaborator, and we thank him for C. Devey, N. Dubilier, S. Fretzdorff, D. Garbe- provide an important tool for determin- providing his expertise and his redox Schönberg, C.R. German, O. Giere, and others. 2007. Young volcanism and related hydrothermal ing the location and distribution of vent potential (Eh) probe on many ABE activity at 5°S on the slow-spreading southern sites in basalt-hosted systems. cruises. Other key collaborators include Mid-Atlantic Ridge. Geochemistry, Geophysics, and Geosystems 8:Q11002, doi:10.1029/2006GC001509. We also used our three-phase search Dan Fornari from the Woods Hole Hessler, R.R., W.M. Smithey, M.A. Boudrias, C.H. methodology to locate and survey hydro- Oceanographic Institution (WHOI), Keller, R.A. Lutz, and J.J. Childress. 1988. Temporal thermal vents based on clues provided Edward Baker and William Chadwick changes in megafauna at the Rose Garden hydrothermal vent, Galápagos Rift, eastern tropical by towed systems. In addition to locat- of PMEL/NOAA, Andrea Koschinsky Pacific. Deep Sea Research 35(10–11):1,681–1,709. ing the vent sites, we made fine-scale of the Jacobs University Bremen, Colin Koschinsky, A. 2006. Discovery of new hydrothermal vents on the southern Mid-Atlantic Ridge (4S–10S) bathymetric maps of the vent sites prior Devey and Klas Lackschewitz of IFM- during cruise M68/1. InterRidge News 15:9–15. to taking photographs, enabling detailed GEOMAR at the University of Kiel, and Shank, T., D. Fornari, D.R. Yoerger, S. Humphris, A. Bradley, S. Hammond, J. Lupton, D. Scheirer, R. study of the geology as well as the iden- Cornel de Ronde of GNS Science, New Collier, A.-L. Reysenbach, and others. 2003. Deep tification of vent fauna. Using these Zealand. The authors also thank the submergence synergy – Alvin and ABE explore the Galápagos Rift at 86°W. EOS, Transactions of the methods on cruises sponsored by the anonymous reviewers whose comments American Geophysical Union 84(41):425. NOAA Office of Ocean Exploration and improved this paper significantly. Judy Tivey, M.A., and H.P. Johnson. 2002. Crustal mag- Research, we found and documented the Fenwick of WHOI also made valuable netization reveals subsurface structure of Juan de Fuca Ridge hydrothermal vent fields. Geology first known South Atlantic vent sites. editorial contributions. In addition to 30(11):979–982. The NOAA Office of Ocean Explora- the NOAA Office of Ocean Exploration Yoerger, D.R., A. M. Bradley, M. Jakuba, C. German, T. Shank, M. Tivey, 2006. Autonomous and tion and Research support also provided and Research, the expeditions presented remotely operated vehicle technology for hydro- continuous improvement of the vehicle in this paper were also funded in part thermal vent discovery, exploration, and sampling. Oceanography 20(1):152–161. system. Technological upgrades and by the National Science Foundation, Yoerger, D.R., M. Jakuba, A.M. Bradley, and B. enhancements included the multibeam WHOI, the UK Natural Environment Bingham. 2007. Techniques for deep sea near bottom survey using an autonomous underwater vehi- sonar, Doppler velocity log, and ABE’s Research Council, and the Deutsche cle. The International Journal of Robotics Research anchoring system. These upgrades were Forschungsgemeinschaft. 26(1):41–54, doi:10.1177/0278364907073773.

Oceanography December 2007 61