SS09 Studying the Ecology, Biodiversity, and Abundance of Aquatic Animals

Session Organizers:
David Bailey
Oceanlab, University of Aberdeen, UK
d.bailey@abdn.ac.uk

Nikki King
Oceanlab, University of Aberdeen, UK
n.king@abdn.ac.uk

Fishing using bait, one of the oldest techniques known to man, was adapted through the use of baited drop cameras to reveal activity of scavenging fauna at abyssal depths in the oceans. Such cameras provide unequivocal evidence of presence of animals at defined depths. Autonomous landers that place the camera on the sea floor for hours or days furthermore reveal the time series of events following arrival of bait. Early work used large or protected baits to produce a maximal persistent attractant effect. By adopting a minimal bait (ca. 500g) the entire cycle of interception, consumption and disappearance of fauna from the bait site could be observed. Using such a standard bait differences can be discerned in behaviour of different species and of the same species in different environments at different depths. From time of arrival of the first fishes estimates can be made of their local abundance depending on models of behaviour. An apparently simple technique has evolved into a sophisticated non-invasive method of investigating ecology of active fauna of the deep sea.

Little is known about how sounds emitted from deep-sea vehicles bias what types of animals are observed. Eye-in-the-Sea (EITS), an unobtrusive battery operated deep-sea observatory, was fitted with a hydrophone and used to record the acoustic signatures of three underwater vehicles, away from the influence of ship and surface noises. Acoustical recordings were made from MBARI’s ROVs, Tiburon and Ventana, and HBOI’s Johnson-Sea-Link submersible (JSL I). Sound files were stored as imbedded audio in MPEG-4 files. Each vehicle was used to conduct a series of maneuvers and tests around EITS at incremental distances from the hydrophone. Vehicle activity was annotated to the ROV’s or submersible’s video and time codes from EITS and deployment vehicles were synced. Noise producing components of the underwater vehicles identified by time code were isolated from extracted sound files and their frequencies and intensities were compared. Presence and absence of animal activity around EITS was also recorded and compared during vehicle-present and vehicle-free recording periods. Clear indications of aversive effects of vehicle presence on animal behavior were documented and will be discussed.

Dramatic changes have occurred in the deep-sea fish populations of the Pacific and Atlantic oceans in the last 20 years. We have previously described how abyssal fish abundances have increased three-fold at one Northeast Pacific location, while other authors have shown major declines in the bathyal fishes of the Northwest Atlantic. Oceanographically-driven changes in food availability probably caused the changes seen in the Pacific, while fishing pressure has driven declines in other stocks over the same time period. We analyzed a newly-collated dataset for the Porcupine Seabight of the North Atlantic, covering the periods 1979 to 1983 (110 trawls) and 1997 to 2002 (61 trawls and 80 baited camera deployments). This dataset included fishes from 700 to 4800 meters water depth and allowed changes in the fish communities to be compared between fished and unfished depths, and between scavenging and non-scavenging fishes. Changes in body size and abundance with depth were analyzed using general additive modeling and compared between years. This allowed the relative effects of fishing pressure, fisheries discarding, and oceanographic "regime shifts" to be established.

Macrobenthos of the deep, northern Gulf of Mexico (GoM) have been sampled with large box cores along multiple cross-depth transects extending from depths of 200 m out to 3700 m. Four major depth zones have been identified based on the faunal similarities (Beta diversity) between geographic sites, with the two intermediate-depth zones being divided horizontally down the middle of the basin. Each zone and sub-zone can be described by a characteristic animal density, biomass and biodiversity (Alpha diversity). The eastern sub-zones have higher stocks than the western sub-zones, presumably due to the horizontal productivity gradient. The alpha diversity displays an intermediate depth maximum with a hotspot enclosing the mid section of the Mississippi and De Soto Canyon. The submergence of faunal zones in the northeast and central GoM may relate to the meso-scale eddies and down-slope sediment movements. In the head of the Mississippi Canyon, the distinct faunal assemblage with high standing stocks and low alpha diversity may be associated with the accumulation of organic matter from river runoff and the adjacent shallow margin. Overall, the input of food resources appears to control the observed zonal patterns.

Submarine canyons can act as traps for organic detritus moving down slope or sinking from the overlying water column. Enhanced macrofaunal abundance and biodiversity have been found in canyon habitats as a result of these detritus concentrating mechanisms. Here we investigate the effects of canyon environments and depth on the scavenger communities of the Hawaiian slope. Standard, 10-kg bait parcels of marlin were placed on the seafloor at 350-1800m to quantify relative abundance and biodiversity of benthopelagic scavenging megafauna off of the islands of Nihoa, Oahu, Molokai, and Maro Reef. Video data taken from 30 submersible dives indicated higher biodiversity in canyon versus non-canyon habitats: of the 16 taxa found to feed on bait, 13 occupied canyons and 9 occupied slopes (non-canyon). Relative abundance of scavengers based on maximum number, first arrival time, and percent of total observation period present suggest that some species of scavengers prefer canyons while others prefer the open slope. Species composition and body sizes varied markedly with depth, suggesting that bathymetric influences are stronger than canyon effects.

Baited cameras are now a widely used tool for investigating the distribution and abundance of scavenging fauna in the deep-sea. Baited camera images are numerically dominated by one or two scavenging species allowing inferences to be drawn regarding overlying productivity and scavenger abundance. However, many minor species are observed and receive little attention in the published literature. These minor species generally arrive hours after the bait has landed on the seafloor and are not observed to feed directly on the bait; a strategy which appears to be energetically counter-productive in a food-poor environment. The contributing factors, such as body size, mobility and estimated abundances, influencing first arrival times of scavenging species at baited landers will be explored and linked to optimal foraging strategies. Results are of interest to those focused on ecological niche separation in scavenging species and selective fisheries such as longlining.

The decapod fauna of six World War II shipwrecks from 75 to 2000 m depths were examined in 2004 in the Gulf of Mexico. All ships were sunk within a two month period in 1942 and were used as surrogates for deep sea drilling structures. Five genera of crabs were counted along video transects on the wreck and adjacent debris field with the XL-11 ROV. Differences in abundance per square meter were compared between three treatments: on the wreck, near the wreck and away from the wreck (>300m) to determine if the presence of hard substrates affected crab distributions. Similarly, crab distributions were compared with substrate and depth to see if either influenced crab abundance. The most abundant crab species were observed only at four of the six shipwrecks, and two species were observed at a single site. Significant correlations with substrate occurred for some crab species, but not all. Our findings provide information on the role of drilling structures as artificial reefs in the deep sea.

Bioluminescent coastal lagoons are unique environments of great ecological importance. This study describes the population dynamics of P. bahamense, the organism responsible for most of the bioluminescence in these lagoons, and C. furca at Laguna Grande, and the relationships between these populations and various physical-chemical parameters. A seasonal population fluctuation pattern was observed in P. bahamense where higher densities were observed mainly from April to September and lower densities from October to February. Population fluctuations of C. furca were more erratic and non-seasonal. The mean population density throughout the study period was significantly higher in P. bahamense than in C. furca. Significant positive correlations were observed in both species between surface and bottom samples. The effects of heavy flooding during November 20 - 21, 2003; and the effects of heavy rains during the path of Hurricane Jeanne, during September 15 - 16, 2004, were reflected in the measurements of visibility, salinity and fluorescence. However, their effect on P. bahamense and C. furca populations are not clear. To our knowledge this study is the longest one carried out in a bioluminescent lagoon.

We deploy remotely operated vehicles equipped with high resolution video cameras enabling quantitative video transects (QVTs) to be obtained that provide ecology data at the scale of individual organisms. QVT sampling advances studies in animal diversity, distribution and abundance. Analyzing QVTs, however, is labor intensive and costly, limiting marine ecological research and application to aquatic management. In the current work, an automated program for detecting and classifying organisms processes video frames with a neuromorphic-selective attention algorithm, modeled after the human vision system. Candidate locations are identified and tracked to determine interesting detected events; these events are marked in the video frames and undergo further processing with an automated classifier to determine the abundance and distribution of representative species. We present comparison between professional annotations and automated detection of organisms in midwater and benthic transects. We present automated classification of organisms in benthic video footage. We present data on detecting animals in video from fixed observatory seafloor cameras.

Zooplankton depth distributions were mapped in the upper 100 m of the water column using a digital, autonomous Video Plankton Recorder (VPR), equipped with a Wetlabs ECO Puck fluorometer/turbidity sensor and a Seabird Fastcat CTD. Zooplankton net samples and phytoplankton samples were collected at the same locations. Stations were located in both Arctic and Atlantic waters, and covered different phases of spring bloom development. Variability and consistence in the vertical structure of Calanus populations was analyzed with respect to population structure, and chlorophyll fluorescence distribution, as well as water mass characterization and water column stability. The relative impact of zooplankton population composition and abundance, chlorophyll availability, and physical environment on vertical structure of the zooplankton was evaluated.

The aggregation response of fish populations following the addition of artificial structures to seafloor habitats has been well documented in shallow water reefs and at deeper structures such as oil extraction platforms. A long-term time-lapse camera was deployed at 4100m in the eastern North Pacific to examine how deep-sea fish populations were affected by an isolated artificial structure and whether fish surveys at this site were biased by aggregation behavior. Counts were taken of grenadiers, Coryphaenoides sp., observed per week as well as numbers of a potential prey species: Echinocrepis rostrata, an epibenthic echinoid. No significant positive correlation was found between the duration of deployment and average number of fish observed at the site. There was also no evidence of associative behavior around the time-lapse camera by E. rostrata. The lack of attraction to the time-lapse camera at this site could be due to the energy needs of deep-sea foraging predators which require them to move constantly, the absence of prey concentrations around the camera, or reduced need for physical shelter from predators in the abyssal environment.