Variation in Seafloor Communities Across the Western New England Seamounts and Adjacent Submarine Canyons: Implications for Conservation Morgan J. Kilgour1*, Peter J. Auster1,2, Dave Packer3, and Les Watling4 1Northeast Undersea Research, Technology & Education Center and Department of Marine Sciences, University of Connecticut at Avery
Point, 1080 Shennecossett Rd., Groton, CT 06340; 2Mystic Aquarium, Sea Research Foundation, 55 Coogan Blvd., Mystic, CT 06355; 3NOAA National Marine Fisheries Service, Northeast Fisheries Science Center, James J. Howard Marine Sciences Laboratory, 74 Magruder Road, Highlands, NJ 07732; 4University of Hawaii at Manoa, Department of Biology, 2538 McCarthy Mall, Edmondson 312, Honolulu, HI 96822
* Current address: Gulf of Mexico Fishery Management Council, 2203 N Lois Avenue, Suite 1100, Tampa, Florida 33607 USA
[email protected]
INTRODUCTION Conservation and sustainable use of biological diversity in the deep sea will require spatial management approaches to avoid significant impacts to vulnerable species and communities. Most problematic is the lack of geographically comprehensive time series surveys of seafloor fauna that are not of economic importance, especially in areas of precipitous topography. Data produced from archived underwater imagery as well as imagery collected to fill geographic gaps can potentially overcome the lack of systematic surveys for initial conservation planning and management needs. Here we use imagery collected using multiple underwater vehicles over approximately one decade in the western New England Seamounts (Bear, Physalia, Retriever, Mytilus) and over four decades (1978‐2013) in adjacent submarine canyons (Oceanographer, Lydonia, Gilbert, and an intercanyon area between Lydonia‐Powell Canyons) (Fig.1, Table 1) to assess whether seamounts in the US EEZ support a unique fauna that requires specific attention for conservation or can be considered as simple extensions of canyon faunas. METHODS • Six different underwater vehicles/platforms were used to acquire imagery: DSV Alvin, DSV Johnson‐Sea‐Link, towed camera TowCam, AUV Remus 6000, ROV Hercules, ROV Deep Discoverer (Fig. 2, Table 1) along with different media for imagery: 35 mm still film, digital still imagery, standard definition video, high‐definition video. • This resulted in variations in fields of view, image resolution, ability to ID organisms, etc. Nevertheless, a data set quantifying taxon presence, geologic habitat, and depth, based on each occurrence attributed to each major geographic feature was produced. • Biological observations were recorded for each occurrence of a species in non‐overlapping still and video images. Each observation was time and geo‐referenced based on available navigation data. Higher taxonomic level identifications of all invertebrates was used to prevent bias based on variable image resolution. • For each observation the dominant and secondary seafloor habitat was recorded based on a defined habitat classification scheme. • PRIMER (Version 6) was used for statistical analyses. For non‐metric Multi‐Dimensional Scaling (MDS) analyses, observations were concatenated based on location (canyon or seamount), habitat type within that location, and depth within that location. Analyses were also performed on coral data alone to test if corals had location specific patterns. Table 1.
Fig. 1. Bathymetric map of the study area. Each dot is a different dive location and resulting biological observation.
Remus 6000 TowCam
Feature Base Gilbert Canyon Gilbert/Lydonia Canyons Lydonia Lydonia‐Powell Intercanyon Oceanographer Canyon
Depth Range (m) 2700‐2800 2000‐2200 1100‐1300 400‐700 600‐2200
Bear Seamount
1300‐1900
Physalia Seamount Mytilus Seamount Retriever Seamount
1800‐2700 2500‐3300 1900‐3900
Alvin Deep Discoverer
Hercules
Johnson‐Sea‐Link Fig. 2. Underwater vehicles used for this study. Location
Vehicle and Year TowCam 2005 Alvin 2005 Deep Discover 2013 Deep Discover 2013 Alvin 1970s Johnson‐Sea‐Link 1980s Alvin 2001 Deep Discover 2013 Alvin 2003 Hercules 2004 Remus 2012 Deep Discover 2013 Hercules 2004
RESULTS • MDS analyses concatenated based on location revealed patterns of localized similarity but larger scale differences. Gilbert/Lydonia canyon and base of Gilbert and Lydonia canyon were similar (60% similarity Fig. 3). Lydonia‐Powell Intercanyon was dissimilar from all other locations. Both Oceanographer and Lydonia were similar, and all seamounts were similar (50% similarity) except Physalia (Fig. 3). When only coral data were analyzed, these patterns were similar with the exception of the seamounts (Fig. 4); Retriever and Physalia were different from other seamounts and from each other (Fig. 4). When the data were concatenated by habitat types and 100 m increment depth zones within each feature, there was clear separation between seamounts and canyons, but the stress level was high, suggesting more complexity to this pattern (Fig. 5). • Each feature exhibited a degree of uniqueness. Each seamount was unique from other seamounts and the intercanyon area was dissimilar from most other canyons and from all seamounts. Of the canyon type features, Oceanographer canyon was most similar to Bear and Retriever seamounts. Of the seamounts, Bear was most similar to canyons. The overarching pattern was a clear separation between the seamounts and canyons, and the intercanyon area was different from both the seamounts and canyons. • Noteworthy is that imagery from submersible dives in the same area of Oceanographer Canyon in 1978, 1980, 2001, 2005 and 2013 facilitated an examination of community structure across time and found high similarity in community composition within depth zones over time (Fig. 6, 7). Transform: Square root Resemblance: S17 Bray Curtis similarity
Fig. 3. MDS plot of square root transformed data by location using a Bray Curtis similarity matrix. Key to abbreviations: BEAR‐ Bear, PHY‐Physalia, MYT‐ Mytilus, RET‐ Retriever, OCE‐ Oceanographer, LYD‐ Lydonia, BGIL‐ Base of Gilbert, GILY‐ Gilbert/Lydonia, INT‐ Lydonia‐ Powell Intercanyon.
Location
1800-1900
2D Stress: 0.11
date 1978 1980 2001 2005 2013
600-700 600-700 500-600 300-400 900-1000
Similarity 20 40 60 80
1300-1400 800-1300 800-1300 1200-1300
Location
1978 1300-1400 2100-2200
Fig. 5. MDS plot of all locations higher order concatenated taxa based on location, depth, and substrate type. Abbreviations same as Fig. 3. Note left to right separation of seamount and canyon samples.
Fig. 6. MDS plot of Oceanographer Canyon. Time series. Data from five time periods from 1978‐ 2013 were used in analyses. Noteworthy is similarity of community composition by depth over time.
CONCLUSIONS Overall these results indicate seamounts and canyons harbor communities unique enough to warrant Fig. 4. MDS plot of all locations corals only square root consideration of inclusive conservation actions to ensure ecological representation within a network of transformed Bray Curtis similarity matrix. conservation areas. Additionally, the composition of these communities should be considered when assessing Abbreviations same as Fig. 3. Lydonia‐Powell management requirements as rare species may need special consideration. The lack of significant differences Intercanyon area was so different from the other in the community composition in a region of Oceanographer canyon over time also supports the continued features that it does not appear in this MDS plot. use of data from historic imagery in the absence of more modern data sets for use in conservation planning. ACKNOWLEDGMENTS The authors thank NOAA, the WAITT Institute, and the Natural Resources Defense Council for funding the various missions, as well as the captains, crews, and engineers of the various ships and sampling platforms that helped to collect such a long time series of data.
2001
2013
Fig. 7. Time series imagery of an area of deep‐sea coral habitat in Oceanographer Canyon.
