No description available.

From the Antarctic Ice Sheet calves every year into the Southern Ocean, an average of 2000 km3 of icebergs. The meltwater is spread over a large area in the Southern Ocean but the large temporal variability in iceberg calving and the clustering of iceberg distribution means that meltwater injection can be locally very high. This study quantifies iceberg distribution, movement and melting using remote sensing observations and modelling. Icebergs were detected and tracked on Synthetic Aperture Radar images using a new computer-based iceberg detection method. The method allows an efficient and systematic processing of large volumes of SAR images, necessary to build a climatology of icebergs in the Southern Ocean. Tests were conducted using ground data from a field campaign and against manual image classification. The method was applied to several SAR image collections, namely the RADARS AT RAMP mosaic for the totality of coastal Antarctica, providing the first picture of iceberg distribution over such a large area. Giant icebergs (icebergs above 100 km2 in area) were shown to carry over half the total mass of the Antarctic iceberg population. Estimates of the spatial distribution of giant iceberg melting over the ocean were made using observed tracks and modelling the melting and spreading along its path. The modelling of basal melting was tested using ICESat laser altimetry to measure the reduction in the freeboard of three giant icebergs in the Ross. The distribution of meltwater for giant icebergs was combined with an existing simulation of meltwater distribution from smaller icebergs to produce the first map of total iceberg meltwater for the Southern Ocean. The iceberg contribution to the freshwater flux is shown to be relevant to both the Weddell Sea and the Southern Ocean south of the Polar Front.

Ice disturbance is frequently cited as having a major role in structuring benthic communities in shallow polar waters, and yet comprehensive field studies on the direct effects of ice disturbance are rare. This thesis aimed to describe the role of ice disturbance in determining benthic community structure at sites around Rothera Point, Adelaide Island (67° 34.5' S, 68° 07.0' W). The work comprised of two main components; a range of observational studies to describe the general effects of ice disturbance, and an experimental quantification of the disturbance pressure, which was then used to make novel links between disturbance and community structure. A photographic survey using depth transects (0-35 in depth) at three sites showed that benthic assemblages changed continuously along a bathymetric gradient. Assemblages were most patchy at shallow depths and the relative abundance of sessile forms increased with depth, which suggested that disturbance intensity was greatest in the shallows. The immediate effects of iceberg impacts were also investigated. Communities within newly formed iceberg scours were sampled and compared with those in undisturbed areas; scoured assemblages were 95% lower in mean macrofaunal abundance and 75.9% lower in species richness. The recovery of three scour assemblages was monitored for -30 months following the disturbance event. Scoured assemblagesb ecame increasingly similar to undisturbed assemblages over time and marked spatial variability was observed in both scoured and unscoured zones. Experimental markers were designed to detect iceberg impacts and were deployed as 24 grids at four depth increments and two study sites. Markers were surveyed regularly for two years to quantify the frequency of iceberg impacts, which varied significantly with depth, site, season and year. The intensity of disturbance was greatest in the shallows (0-5 m depth) and significantly reduced at 25 m depth. Variation between site and season could be largely explained by the duration of winter fast ice. Extensive sampling at each of the disturbance grids showed that disturbance intensity has a significant and wide-ranging influence on macrobenthic community structure.

The sea ice thickness distribution, represented by the probability density function (PDF), is critical to Earth's climate system and knowledge of the distribution in the Antarctic is limited. A novel methodology, using an acoustic Doppler current profiler, was developed to measure sea ice draft based on measurements taken from Autosub (an autonomous underwater vehicle). The PDF has been derived for three missions undertaken in the eastern Amundsen Sea during March 2003. The PDFs for all missions were found to have a single mode although there is evidence for variability in mean thickness along the ice front, as the most western mission has a lower mean draft. Geostatistical analyses of the data have allowed the derivation of PDFs to account for the spatial sampling. A factor on the thickness of ice measured is the presence of icebergs within a study region. This thesis reports on work carried out to investigate whether a correction to the sea ice thickness PDF can be made to account for icebergs and over what scale(s) this correction is valid. To answer this question data from Autosub and satellite images were used to investigate whether icebergs were randomly distributed both in open ocean and within sea ice. In conclusion, it was found that icebergs do cluster on the scale of typical Autosub missions (~few km). However, there are differences between icebergs in sea ice compared with open water. Therefore, icebergs should be accounted for in the sea ice thickness PDF.

Antarctic ice shelves play a key role in the global climate system, acting as important sites for the cooling of shelf waters, thereby facilitating deep and bottom water formation. Many of the processes that take place under large ice shelves can be observed more conveniently beneath smaller ice shelves such as Fimbul Ice Shelf, an ice shelf in the eastern Weddell Sea. Fimbul Ice Shelf and nearby ice shelves might also play a significant regional role: although no bottom water is produced in this area, it is thought that Fimbul Ice Shelf and nearby ice shelves precondition the shelf waters that ultimately are converted to Weddell Sea Deep Water (WSDW) in the southern Weddell Sea. Using the first data ever to be collected beneath an ice shelf from an autonomous underwater vehicle (AUV), as well as data from the vicinity of the ice shelf using traditional oceanographic methods, this thesis discusses the circulation and processes occurring beneath the ice shelf. This has been supplemented by using a coupled ice shelf/ocean model, POLAIR, to simulate the circulation, melting, and tides under Fimbul Ice Shelf and in the surrounding area. Data from the ice shelf cavity show relatively large variability in temperatures, but all within approx. 0.25 °C of the freezing point. Melt rates are found to be lower than some previous model studies, but in better agreement with observations and glaciological models. The base of the ice shelf was found to be rough in places, corresponding to `flow traces' visible in satellite imagery. This could have implications for mixing beneath the ice shelf, at least in these limited areas. The Autosub AUV was found to be a useful platform for measuring hydrography and ice shelf cavity geometry with spatial coverage and resolution not available from surface measurements.

La calotte polaire Antarctique conditionne un flux d’eau douce dans l’océan Austral par deux voies d’égale importance : une injection immédiate et localisée par fonte des plateformes glaciaires, et une injection « offshore » et différée par production (« vêlage ») puis fonte d’icebergs. On estime ainsi que les icebergs, en fondant, pourraient modifier les caractéristiques hydrologiques et biogéochimiques de la colonne d’eau. Les modèles numériques visant à estimer cet impact présentent des résultats contrastés. Ils sont limités dans leurs stratégies de représentations des icebergs, notamment parce que les connaissances sur la distribution spatiale et de taille des icebergs ou encore leurs mécanismes de perte de masse sont réduites. Une méthode récente exploitant des mesures par altimétrie satellitaire a permis la création d’une base de données cartographiant la distribution des icebergs Antarctiques avec une couverture spatiale et temporelle inédite. Notre analyse conjointe entre ces données et des champs de concentration en glace de mer met en lumière le transport d’eau douce injecté par les icebergs et son impact sur la banquise. On analyse également les liens entre icebergs de différentes tailles : les gros peuvent être vus comme des réservoirs de volume de glace, qu’ils diffusent dans tout l’océan en se fragmentant en petits icebergs de différentes tailles. On étudie alors l’évolution de deux icebergs géants, on propose une première paramétrisation du phénomène de fracturation et analyse la distribution de taille résultante. Ces résultats peuvent permettre une représentation plus réaliste du flux d'eau douce conditionné par les icebergs dans les modèles. / The Antarctic polar ice cap constrains a freshwater flaux into the Austral Ocean through two equally important pathways : a localized and immediate injection through the melting of ice-shelves bases, and a delayed offshore injection through the calving and subsequent melt of icebergs. Some studies reckon that melting icebergs have the capacity to alter the hydrological and biogeo-chemical characteristics of the water column. The numerical models trying to evaluate this impact have shown contrasting results. Yet, they might suffer from a poor representation of the icebergs, namely due to our limited knowledge on both the spatial and size distributions of the icebergs, or even the processes involved in their mass loss. A new method using satellite altimetry measurements has lead to the creation of a database mapping Antarctic icebergs distribution with an unprecedented spatial and temporal coverage. Our joint analysis between these data and sea ice concentration fields highlights a possible transport of the freshwater injected by an iceberg and its impacts on sea ice.We also analyze the links between icebergs of different sizes : the large ones can be seen as ice buffers that diffuse across the whole ocean when breaking into small fragments of various sizes. We finally study the evolution of two giant icebergs, suggest the first parametrization of the fragmentation process and analyze the subsequent size distribution of the fragments. These results can be valuable to account in a more realistic way the fresh water flux constrained by icebergs in models.