Seasonal dynamics of megafauna on the deep West Antarctic Peninsula shelf in response to variable phytodetrital influx

[This corrects the article DOI: 10.1098/rsos.140294.].


Summary
The deep West Antarctic Peninsula (WAP) shelf is characterized by intense deposition of phytodetritus during spring/summer months, while very little food material reaches the seafloor during winter. The response of the shelf benthic megafauna to this highly variable food supply is still poorly understood. In order to characterize the deposition of phytodetritus and the megabenthic community response, we deployed a seafloor timelapse camera at approximately 590 m depth on the mid WAP shelf west of Anvers Island for 15 months. Seafloor photographs were taken at intervals of 12 or 24 h nearly continuously from 9 December 1999 (austral winter) to 20 March 2001 (summer) and analysed for phytodetritus deposition and megafaunal dynamics. Seafloor images indicated a marked seasonal arrival of greenish phytodetritus, with large interannual and seasonal variability in the coverage of depositing phytodetrital particles. The surface-deposit-feeding elasipod holothurians Protelpidia murrayi and Peniagone vignoni dominated the epibenthic megafauna throughout the year, frequently constituting more than 80% of the megafaunal abundance, attaining total densities of up to 2.4 individuals m −2 . Elasipod abundances were significantly higher in summer than winter. During summer periods of high phytodetrital flux, Pr. murrayi produced faecal casts at higher rates, indicating intensified population-level feeding activity. In March-June 2000, faecal casts lasted longest, suggesting lower horizontal bioturbation activity during autumn-winter. Our data 2014 The Authors. Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited.

Introduction
Oceanic and coastal Polar regions are characterized by extremes in primary productivity during an annual cycle [1][2][3]. Periods of high phytoplankton production may also be subject to high interannual variation related to longer term climate cycles and climate change [4,5]. Such variations in organic-matter production and phytodetritus flux may affect planktonic and benthic heterotrophic populations through changes in food availability, larval survival, recruitment rates, competition for food, feeding behaviour and other processes [6].
On the West Antarctic Peninsula (WAP) shelf, large phytoplankton blooms can occur in December-March, resulting in intense bouts of particulate organic-carbon (POC) deposition on the seafloor [7][8][9]. The intense spring-summer phytoplankton blooms sustain the Antarctic food web [7,10] and provide the main food source, in the form of sinking phytodetritus, to seafloor communities on the deep WAP shelf [9,11,12]. These periods of high summer production and phytodetritus flux are separated by seasons of extremely low winter (April-October) productivity, when sea-ice cover and low-light conditions prevail, allowing very little food to reach the benthos [9,13], raising questions on how benthic detritivores survive over winter months [12,14].
In the water column, where seasonal variability in food concentration is especially dramatic, grazers/detritivores exhibit varied life-history strategies. Some animals, e.g. the copepod Paralabidocera antarctica, store lipids to allow fasting over the winter [15,16]. Other animals, such as krill and copepods, can store lipids [6,16,17] but may also feed on ice algae and at the seafloor, or migrate to ice-free oceanic waters [18][19][20]. On the deep WAP shelf, benthic animals do not appear to store lipids [21,22], and yet can remain active year round [22][23][24], feeding on a benthic food bank of organic material stored in sediments [9,12,14,25,26].
While it has been postulated that deposit-feeding megabenthos play an important role in the flow of organic carbon through Antarctic shelf benthic food webs [28,37,38], little is known about the foraging rates of elasipod holothurians on the WAP shelf, and their response to seasonal deposition of fresh phytodetritus.
In this study, we use a time-lapse camera system deployed on the deep WAP shelf (approx. 590 m depth) to (i) monitor the arrival of phytodetritus and (ii) evaluate variations in megabenthic community structure and surface-deposit faecal production, which is used as a proxy for foraging rates, over a 15-month period at the WAP shelf floor. We find that elasipod holothurians are active throughout the year and increase in population densities through migration, modulating their feeding rates with variations in the export of particulate material from the euphotic zone. Given the high volume of sediment processed, these organisms are likely to be very important in organic-carbon cycling at the WAP shelf seafloor.   (figure 2) [26]. The camera was mounted on an aluminium tripod at an elevation of 152 cm above the seafloor and took pictures at a 45 • angle of ca 4 m 2 of seafloor. All deployments used Kodak 5279 Vision 500 T motion picture film (ASA 500). The camera malfunctioned between 19 January 2000 and its recovery on 9 March 2000, and between 25 September 2000 and 25 October 2000, so there were no photographs during these intervals. In addition, no photographs were collected during the ca 10-day periods between recovery and redeployment during mid-project cruises. At each deployment/recovery, at least one otter trawl sample was taken to capture local megafauna for identification and reproductive studies [23].

Material and methods
In the laboratory, images were digitalized using a Nikon Super Coolscan 5000ED 35 mm/APS (IX240) film scanner into tagged image file format (TIFF) for later analysis. The automatic colour compensation provided by the scanner was reversed to recover the original slide colours. The reversal process used the black frame as a reference. Each of the original 1008 × 672 pixel TIFF images was converted to four (RGBI) matrices, centred through a pattern-recognition Matlab program that identified the position of the central crosshatch in the Photosea 2000 image. An area of 800 × 620 pixels centred on the crosshatch was kept and outside areas cropped from all images. A Canadian perspective grid (50 × 50 cm squares) [39] was then constructed based on the camera elevation and angle relative to the seafloor and superimposed on each oblique image using Matlab. As the oblique grid cells represented squares in the seafloor, the edges of a larger square (1 × 1 m) formed by the four bottom centre cells were used for a perspective transformation. This transformation places the image viewpoint at a right angle from the seafloor (assumed to be planar) allowing distances and areas to be calculated. The quantitative part of the study focused on the epibenthic megafauna living on soft sediments. During the course of the study the camera never faced hard bottoms. Owing to different lighting conditions and to minimize perspective correction  errors, only the well-illuminated bottom area of the images (1.6 m 2 or 491 × 491 pixels) was used in the analysis, thus reducing errors in the counting and identification of species [40]. We could resolve and identify animals exceeding approximately 2 cm in minimum dimension. Epifaunal organisms were counted in each image (N = 452) and numerical density presented as individuals per square metre. The number of times a single individual appeared in consecutive images was recorded in order to evaluate the activity of benthic organisms (mostly holothurians). Individuals were visually identified by their size, position in the frame and bioturbation track in consecutive frames. Holothurian movement rates were calculated as the average linear distance of single individuals between consecutive frames (12 h). As these animals move irregularly, distances represent minimal movement rates and are likely to be underestimated, as we could not estimate the distance travelled (or direction of movement) from individuals that only appeared once in photographs.
Faecal-cast numerical densities and persistence times were calculated for the entire period for the holothurian Protelpidia murrayi, which produced the most conspicuous faecal casts (figure 2). Faecal casts appearing in images were marked and tracked in each successive photograph until they could no longer be resolved. For each cast, length (i.e. the uncoiled faecal cast) and width (i.e. the diameter of a faecal-cast section) were measured and the volume of sediment processed was calculated assuming the faecal casts to be cylindrical. Individual faecal-cast production was estimated for individuals that could be tracked in pictures. Population-level faecal production was calculated using the total volume of faeces produced per day in a given area. Faecal-cast parameters were compared using either one-way ANOVA or Kruskal-Wallis test after testing for the homogeneity of variances. Differences were considered significant at p < 0.05. When significant differences were found, we applied the post hoc tests of Tukey or Dunn depending on the assumptions of normality of data (Bartlett's test).
Sediment reworking time for individual Pr. murrayi was calculated assuming that animals feed on the top millimetre of the sediment surface based on observations of animals kept in the laboratory. Therefore, total reworking of the top millimetre of 1 m 2 of sediment implies 1000 cm 3 of sediment ingestion. The total reworking time of Pr. murrayi population at a given season was then calculated dividing 1000 cm 3 m −2 by the production of faecal casts per day. Therefore, this measurement represents the time of foraging required to rework a square metre of sediment. Seasonal ingestion rates were also tested by ANOVA or Kruskal-Wallis tests at a significance level of 0.05 and the same post hoc tests applied as explained above. Data on faecal casts and reworking times were compared with POC fluxes obtained by Smith et al. [9] during the same period.     (table 2). Other echinoderms included the sea stars Psilaster charcoti, Henricia spp., and an unidentified species, as well as the sea urchins Amphipneustes spp. and Ctenocidaris perrieri. Crinoids and ophiuroids were observed occasionally, as were molluscs, including the octopus Pareledone charcoti and the gastropod Harpovoluta charcoti. Giant nemerteans (Parborlasia corrugatus) and fish were also present in small numbers. A number of species present in pictures, including small penatulaceans and shrimps, were not counted because they were difficult to identify and the densities were very low.

Movement rates and feeding dynamics
Individual holothurian movement, based on the average linear distance travelled by single individuals between consecutive 12 h interval frames (N = 2 to 14 individuals), showed that Pr. murrayi individuals had similar seasonal movement rates (1.  [9]. We observed no influence from animal size on movement rates, but note that our sample size was small. The average faecal-cast volume for both holothurians varied from 1.8 to 6.  The right Y-axis represents scores for phytodetritus coverage of the seafloor as follows: no greenish, flocculent phytodetritus visible at the seafloor (score = 0); diffuse, greenish, flocculent material visible in some areas but bioturbation traces readily visible (score 1); much of the seafloor (50-90%) covered with greenish phytodetritus, with bioturbation traces partially filled (score = 2); more than 90% phytodetritus cover (score = 3) [9]. Score 4 represents dense, up to 2 cm thick phytodetritus carpets. Vertical lines are ±1 s.e.

Discussion
Our time-lapse photographic record shows evidence of strong interannual and seasonal variability in deposition of phytodetritus over the WAP shelf, associated with significant changes in megafaunal abundance. This high variability is consistent with previous studies documenting substantial seasonal and interannual variability in the flux of particulate organic matter (POM) and chlorophyll a to the WAP shelf sediments [7,9,10,12,41,42]. The time-lapse images revealed no phytodetrital coverage of sediments from the summer of 1999-2000 until the winter of 2000, which was followed by a rapid (i.e. days) increase in detritus accumulation in November 2000 [9]. The high interannual variability of phytodetrital accumulation may result from complex pelagic nutrient dynamics and variable ice retreat controlling phytoplankton blooms, variable grazing by zooplankton and lateral transport of surface phytoplanktonrich waters over the WAP shelf [43][44][45]. During summer periods of high primary productivity, POM fluxes are enhanced by the formation of larger, heavier marine snow aggregates and zooplankton faecal pellets [46,47]. During the spring-summer of 2001, thick carpets (approx. 2 cm) of phytodetritus were observed on the seafloor and appear to have provided an important food supply to the deposit-feeding megafauna [9,12,45].
Previous studies have indicated that temporal variability in the abundance and/or activity of macro-and megabenthos over a 15-month period on the deep WAP shelf was largely decoupled from phytodetritus accumulation and POC flux at the seafloor [24]. The weak coupling between seasonal POC flux and megafaunal and macrofaunal communities in the WAP shelf has been postulated to result from the presence of a sediment 'food bank' for deposit feeders [12,24,26,45]. The time-lapse images give further support for the food bank hypothesis, revealing that holothurians forage across the seafloor (i.e. inferred from movements), and consume and egest sediments, even during winter periods of very low food flux to the WAP shelf floor.
The community dominant Pr. murrayi fed continuously on surface sediments throughout the year, as indicated by the production of faecal casts during all our study intervals (figure 5). Similar continuous feeding on the labile organic matter available in surface sediments and on fresh phytodetritus deposited on the seafloor during the spring/summer season has been documented for infaunal echiurans and holothurians on the WAP shelf [12,23,26]. These finding indicate the presence of a sediment food bank on the WAP shelf fuelled by summer blooms, which may persist due to low bacterial organic-matter degradation rates in the extremely cold Antarctic shelf temperatures [9,12].
Although Pr. murrayi does feed year round, this abundant surface-deposit feeder does appear to modulate its sediment ingestions rates in response to increased food availability at the seafloor, as suggested by Sumida et al. [24]. This response is indicated by a nearly fourfold increase in the production of leading to diminished gut residence times presumably owing to the higher food quality [9,45,48,49]. Protelpidia murrayi, which is relatively unselective during particle ingestion [25], relies on selective digestion of organic matter; the presence of higher concentrations of labile phytodetrital material may allow it to pass sediment through the gut at higher rates, while maintaining adequate digestive yield [50,51].
At the Pr. murrayi population level, the production of faecal casts is seven times higher during summer of 2000-2001 than the previous summer, as a consequence of higher population densities probably resulting from immigration, larger body sizes and higher individual-level feeding rates. These variations in Pr. murrayi population densities and faecal-cast production rates yield substantial temporal variation in reworking times for the top millimetre of the sediment, with a turnover time approximately 287 days during the period of intense phytodetrital accumulation and approximately 2114 days when fluxes were lowest in winter months ( figure 7).
Faecal-cast duration suggests that the benthic community as a whole passes through cycles of low and high bioturbation rates, with faecal casts persisting approximately twice as long (approx. 317 h) during winter months than during summer periods. Megafaunal movement rates and foraging activities are likely to be the main cause of the disappearance of faecal coils in this low-flow environment [52]. We found no evidence of sediment resuspension in our images, which suggests that currents are not important in erasing megafaunal traces and faecal casts from the sediment surface. These cycles of bioturbation activity may also be important in the vertical mixing of POM across the sediment-water interface [53]. Such vertical transport will make food available to subsurface feeders and can ultimately lead to carbon burial. For example, the Antarctic sea urchin Abatus ingens reworks the top 2 cm of shallowwater sediments 2 to 17 times per year, influencing a variety of sediment processes including nutrient mixing and oxygen penetration to lower layers [54]. In our study area, faecal-cast duration seems likely to be controlled by a combination of 'retracking' by epifauna and diffusive sediment mixing by infauna (cf. Wheatcroft et al. [52]). The higher population-level sediment reworking rates for Pr. murrayi during the period of high phytodetritus availability suggests that 'retracking' by epibenthos contributed to the reduction in faecal-cast duration during November 2000-March 2001.
Because of high population densities and high feeding and faecal-cast production rates, Pr. murrayi may be geochemically important [55], modulating the distribution of labile organic matter at the sediment-water interface, and within the sediment column, on the deep WAP shelf. During periods of high phytodetrital flux, the Pr. murrayi population appears capable of processing the top 1 mm of sediment, and the recently deposited phytodetritus, on 200-300 day time scales, i.e. over time scales roughly similar to the spring-summer bloom period and the occurrence of phytodetritus at the seafloor [9]. As a consequence, much of the labile phytodetritus depositing on the WAP floor during the summer bloom may pass through a holothurian 'filter', becoming assimilated or redistributed as faecal casts by surface-deposit-feeding holothurians before becoming available to the broader deposit-feeding community. Thus, the dynamics of deposit-feeding megafauna, in particular Pr. murrayi, should be included in models of biogeochemical cycling and climate change on the WAP shelf. Because surfacedeposit-feeding elasipod holothurians similar to Pr. murrayi are abundant globally at bathyal depths and have been shown to actively consume labile phytodetritus [56], these holothurians are likely to modulate the distribution of labile organic matter on many continental margins.