J. Priddle

Temporal variation in Antarctic sea-ice: analysis of a long term fast-ice record from the South Orkney Islands

Detection of climate-induced change in marine ecosystems requires a knowledge of the underlying variability of the environment. This paper uses a range of datasets to investigate the interannual variability in Southern Ocean sea-ice dynamics. We present the first analysis of a series of fast-ice duration data from Signy Island, which we have cross-calibrated and combined with an earlier series from the same island group. The combined series covers the period from 1903 to the present day. The analyses indicate that there has been a long term decline in the duration of sea-ice at the South Orkney Islands in the north-west Weddell Sea. This change has not been a simple linear decline but appears to have been the result of a reduction in the duration of fast-ice during the 1940s and 1950s. There was a pronounced sub-decadal year cycle in fast-ice duration at the South Orkney Islands from the mid-1960s to 1990. In recent years this cyclicity has broken down and fast-ice duration has been greater than expected. Analyses of satellite data have shown that fast-ice duration at Signy Island reflects the larger-scale ice dynamics of the Weddell Sea. Investigation of the Weddell Sea ice dynamics in relation to circumpolar ice extents indicates that the position of anomalies in the maximum sea-ice extent field precesses around the Antarctic continent with a period of approximately 7–9 years. Analysis of atmospheric and oceanic connections with the sea ice variability show that there are signals in both regimes. This environmental variability has significant implications for ecosystem function and the detection of short-term and long-term ecological change.

Large-Scale Fluctuations in Distribution and Abundance of Krill — A Discussion of Possible Causes

Unusually low abundance of krill may last for periods of several months in the Scotia Sea near South Georgia and in Bransfield Strait. Two longer data sets on krill predators suggest that such events may occur two or three times in a decade, and that the situation normally returns to normal in the following season. It seems most unlikely that these events can be ascribed to features of krill biology. Simple models of recruitment failure or mortality cannot explain the observed changes, and alteration in small-scale distribution is not indicated by the available data. More probable mechanisms must involve large-scale changes in distribution of krill brought about by ocean-atmosphere processes. Whilst natural variation in mesoscale features has an appropriate spatial scale, the likely duration is too short. However, a breakdown of hydrographic structure in the surface water over a large area would drastically decrease the residence time of krill and it would take a longer time to reestablish high krill biomass. A prolonged period of southwards airflow over the Scotia Sea is identified as the likely driving force in this model. Such an airflow has been identified from atmospheric pressure distribution in the winters of 1983 and 1986, and was associated with southwards displacement of both warm surface water and of pack ice in the northern Weddell Sea.

ANTARCTIC AQUATIC ECOSYSTEMS AS HABITATS FOR PHYTOPLANKTON

1. The Southern Ocean is a large‐scale, relatively homogeneous upwelling ecosystem whose phytoplankton apparently grows suboptimally over much of its area. By contrast there is a wide variety of freshwater habitats in the Antarctic and in some of these phytoplankton growth efficiency is very high. The two habitats share similar temperature and irradiance regimes, but differ markedly in availability of inorganic nutrients, in grazing pressure and in the time‐ and space‐scales on which various physical processes act.

Scales of Interaction Between Antarctic Krill and the Environment

A conceptual framework to allow a constructive approach to the analysis of the Antarctic marine ecosystem taking into account the variability of the system was developed. The temporal and spatial scale relationships of biotic and abiotic processes were found to be important in the investigation of ecosystem structure. These processes were regarded as forming hierarchies over temporal and spatial scales. The hierarchies define which processes interact over common scales and can also be used to consider which interactions can occur. Different levels of krill aggregation change the possible interactions in the system, and the scales of operation of these aggregations have important consequences for the investigation of ecosystem structure. The framework allows specific interactions, such as grazing and predation, as well as more general aspects such as the formation of sea ice to be considered and allows an overview of ecosystem structure and function.

Seasonal and annual change in seawater temperature, salinity, nutrient and chlorophyll a distributions around South Georgia, South Atlantic

Data collected between 1926 and 1990, during the Discovery Investigations and fourteen subsequent cruises, have allowed the description of spatial and temporal variability of temperature, salinity, phosphate, silicate, nitrate and chlorophyll a in the surface waters around the subantarctic island of South Georgia. Measurements made in Antarctic Circumpolar Current water were compared with others made in Weddell Sea water, and profiles from shelf, shelf-slope and oceanic sites were considered separately. In summer, Weddell Sea surface water was significantly colder than that of the Antarctic Circumpolar Current (1.66 and 2.59°C, respectively), but no changes of temperature corresponded with bathymetry. There were no systematic differences between the salinity measurements made in Weddell Sea surface water and those in Antarctic Circumpolar Current water; however, oceanic waters were always more saline than those over the shelf-slope and shelf (33.91, 33.89 and 33.86, respectively). Silicate levels correlated well with seawater temperature, and Weddell Sea surface water concentrations were substantially higher than those of the Antarctic Circumpolar Current (51 and 25 mmol m−3, respectively, in winter; 29 and 13 mmol m−3 in summer). No such differences were found for phosphate or nitrate, and no systematic differences in any of the nutrient levels were attributable to bathymetry. Although summer chlorophyll a levels appeared to be highest in Antarctic Circumpolar Current water over the shelf and shelf-slope (2.7–3.0 mg m−3), no significant differences were attributable to water-mass or bathymetry. A clear seasonal pattern was evident, with the warmest seawater conditions, minimum nutrient concentrations and highest chlorophyll a levels found between December and March. Phosphate and nitrate were never exhausted: the lowest recorded phosphate concentrations were around 0.6 mmol m−3 and for nitrate 11 mmol m−3. However, low concentrations of silicate (∼ 1.0 mmol m−3) were evident during some summer surveys from the 1920s through to the present day. Average nutrient deficits calculated either between winter and summer mixed-layer concentrations, or between summer mixed-layer and Tmin values, produced similar estimates of carbon fixation for both phosphate and silicate, while nitrate appeared to underestimate carbon production. Phosphate and silicate deficits were considered to be satisfactory predictors of carbon production, which was about 30–40 g C m−2 year−1 in a mixed-layer depth of 50 m. Considerable interannual variability was found, with winter-like conditions prevailing until January on some occasions, and apparent year-to-year variability in the timing and magnitude of nutrient utilisation (especially silicate) by phytoplankton. A relationship was found between sea surface conditions around South Georgia in summer and the preceding winters fast-ice duration at the South Orkney Islands, which implied that some of this variability was attributable to large-scale change over the Scotia Sea as a whole, as opposed to local influences.

The absence of krill from the South Georgia zone, winter 1983

Abstract An extremely low abundance of the Antarctic krill, Euphausia superba, was observed in the Scotia Sea during the austral winter of 1983. The decline could not be attributed to any of the environmental factors investigated, including phytoplankton biomass which was higher than expected. One effect of the reduced krill population, observed during the 1983 to 1984 summer, was a much lower than average breeding success for bird and seal predators of krill on Bird Island, South Georgia. Historical evidence suggests that drastic reductions in krill abundance may not be uncommon, and will have profound consequences for the developing krill fishery.

Some Environmental Factors Influencing Phytoplankton in the Southern Ocean Around South Georgia

SummaryData on phytoplankton and zooplankton biomass, and physical and chemical variables, are combined with a published multivariate description of diatom species composition to interpret variation within an area around South Georgia surveyed during an austral summer. Large-scale species distributions could be equated to the different water masses which reflected the interaction of the Antarctic Circumpolar Current with the island and the Scotia Ridge. Small-scale factors were found to act at an interstation scale and imposed local variation on the biogeographic pattern. Nutrient depletion could be related to phytoplankton biomass but no single inorganic nutrient of those measured (NO3−N, PO4−P and silica) could be identified as important. The ratio Si:P appeared to be more important as an ecological factor. The impact of grazing by krill and other zooplankton could only be resolved as differences in phytoplankton biomass and phaeopigment content. Diatom species composition showed a relation to local krill abundance very different from that suggested by published studies, but could be explained as the effect of earlier grazing outside the study area. The effects of vertical mixing could not account for interstation differences as pycnocline depth was uniformly greater than euphotic depth, and vertical stability very low. Some comparison was made with data collected in 1926–31 by the Discovery Investigations. Significant differences in the distribution of certain taxa such as Chaetoceros criophilum and C. socialis were traced to major differences in hydrology.

Variations in phytoplanktonic nitrogen assimilation around South Georgia and in the Bransfield Strait (Southern Ocean)

Abstract Nitrogen assimilation was measured in two austral summers in the Scotia Sea around the island of South Georgia as well as the Bransfield Strait. Nitrate and ammonium assimilation was measured using 15 N techniques and the population was divided into two size classes, less than and greater than 20 μ m. Water column integrated nitrogen assimilation rates varied between 2.43 and 26.50 mmol N m −2 day −1 , the distribution being highly heterogeneous. The highest assimilation rate was found at a station near South Georgia, where the chlorophyll standing stock was elevated. A high assimilation rate was observed at a station in the Bransfield Strait and was associated with a localised, shallow mixed-layer feature. The less than 20 μ m size fraction contributed to a variable but frequently significant proportion of the total assimilation (14–78%). f ratios were generally low, signifying a high dependence of the population on ammonium as a nitrogen source. Small phytoplankton exhibited a statistically significant greater preference for ammonium than large species, and the total community f ratio was influenced strongly by the proportion of the less than 20 μ m fraction and ammonium concentration. It is suggested that nutrients may play a more important role in the ecology of the phytoplankton in the Southern Ocean than is usually supposed.

Retention of phytoplankton by an eddy

Abstract A patch of high phytoplankton biomass, approximately 10 9 m 2 in area, remained apparently stationary off King George Island, South Shetland Islands, over a period of at least 11 days. The patch was centred on an eddy at the apex of a tight meander formed by water passing round the eastern end of King George Island and being turned back immediately by the strong northeastern flow of water within the Bransfield Strait. Chlorophyll a biomass approached 1 g m 2 and was concentrated within the top 50–75 m of water. The community consisted mainly of diatoms, which were growing actively. Growth in situ might have been sufficient to generate the observed high biomass from the general concentration of phytoplankton observed locally but only if all production had been retained within the patch. However, elevated biomass downcurrent of the patch suggested that physical retention was only about 50% efficient. It is concluded that the patch was derived from a pulse of high biomass which had been transported into the area and partially retained by the horizontal recirculation of water within the eddy. Dominant diatom species, Odontella weissflogii, Proboscia ‘alata’, Chaetoceros curvisetum and Thalassiosira tumida , were also predominant in a phytoplankton maximum off Brabant Island, upcurrent of the study site. This may have provided the seed population for the patch off King George Island. Vertical migration may have prevented krill accumulation within the eddy and it is concluded that some shallow eddies may partially insulate phytoplankton communities from heavy grazing pressure.

Differences between swarms of Antarctic krill and some implications for sampling krill populations

Antarctic krill, Euphausia superba Dana, were sampled from 28 discrete swarms in a small area south west of Elphant Island, South Shetlands, over a period of 14 d (24 February-9 March 1985). Four biological characteristics of the krill (length, sex, moult, gut fullness) were examined in a study of variation between swarms. Analysis of these characteristics indicated extensive differences between swarms and no single characteristic, or combination of characteristics, emerged as consistently accounting for the observed heterogeneity. There was no relationship between the degree of heterogeneity and the physical or temporal proximity of swarms. The variability between even close swarms strongly suggests that swarms are the basic unit of organization of krill populations. This conclusion has important implications for sampling programmes aimed at estimating population parameters. Thus, in the population studied here, we calculate that it was necessary to sample 23 swarms to estimate mean length and 10 swarms to establish a reliable estimate for the proportion of females in the population.