Current patterns and processes
Ice scour, the action of floating ice colliding with the seabed, is a major structuring force acting on polar benthic communities (Sahade et al. 1998, Peck et al. 1999, Gutt and Starmans 2001, Barnes and Brockington 2003, Gutt and Piepenburg 2003, Conlan and Kvitek 2005, Smale 2007, Laudien et al. 2007, Smale et al. 2007b). Floating ice can be both sea (seasonal sea ice) and land (ice bergs) derived and, once in the sea, is moved by the action of winds and oceanic currents. If the keel depth of the ice exceeds water depth, the contact between ice and substratum can cause considerable changes to the physical environment, including altered topography, sediment structure and current flow (Lien et al. 1989, Rearic et al. 1990, Woodworth-Lynas et al. 1991). Furthermore, ice scour occurs across a considerable geographic area, as icebergs have been recorded between 75 to 26° (but more typically ca 40°) in the southern hemisphere and between 80 to 44° in the north (Wadhams 2000). Peck et al. (1999) suggested that ice scouring is extensive as it “affects one-third of the world's coastlines”. Gutt and Starmans (2001) later calculated that “iceberg scouring is amongst the five most significant disturbances that any large ecosystem on earth experiences”.
The frequency of ice collisions with the seabed varies considerably in space between poles, regions, sites and depths, and in time from longer term Milankovitch cycles to years and seasons. Broad scale estimates of scour frequency have been proposed for both the Arctic and the Antarctic. Reimnitz et al. (1977) suggested that in some Arctic regions, shallow shelf sediments are completely reworked by ice in 50 yr, which was later supported by Gutt et al. (1996) who calculated that each square metre of their shallow Arctic study area (east Greenland Sea) was impacted once every 53 yr, on average. Ice scouring may occur to depths of up to 600 m in the Antarctic (Dowdeswell and Bamber 2007), which is likely to be considerably deeper compared with the Arctic, where icebergs are generally smaller. Calculations primarily from the Weddell Sea would suggest that each square metre of the entire Antarctic shelf is disturbed by ice on average once every 340 yr, although variability between regions is undoubtedly high (Gutt et al. 1996, Gutt and Starmans 2001).
These estimates of scouring frequency were based on observations of relict scour marks, rather than measures of actual disturbance. There have, however, been two attempts to estimate present-day scouring frequency, both conducted at Adelaide Island on the WAP. In the first study of its kind, Brown et al. (2004) deployed concrete markers in the shallows of two contrasting sites, and recorded how often these markers were damaged by iceberg impacts over a two year period. They observed significant differences in scouring frequencies between sites and found that some areas of seabed at ca 10 m depth were disturbed at least three times during the two-year study. Smale et al. (2007b) expanded this study to determine the effects of depth, site and season on the frequency of ice disturbance. Their study showed that the frequency of iceberg disturbance significantly decreased with depth in the shallow waters of Adelaide Island, and that disturbance varied between sites, seasons and years. Temporal variability in ice scouring frequency is largely driven by the formation and duration of seasonal fast ice. When the sea surface freezes during the winter months to form fast ice, icebergs are trapped within the ice and locked into position, which may restrict their movements and reduce the frequency of ice scouring (Smale et al. 2007b).
Ice scouring has a wide range of effects on marine benthos, which can be considered on three spatial scales: 1) the localised effect of a discrete iceberg grounding, in terms of both the immediate alteration to community structure and the subsequent recovery processes, 2) at the scale of habitat and how communities respond to continuous disturbance and 3) the influence of ice scouring at the regional scale of the WAP and Scotia Arc. The smallest scale, that of the actual collision between ice and sea bed, can range in area from less than a square metre in the shallows to scour marks of up to 350 m width and 15 km in length in deeper waters (Hotzel and Miller 1983). Iceberg impacts are highly energetic, even in the shallows (Smale et al. 2007b), and cause severe damage and high mortality to benthos (Fig. 2). For example, in shallow waters Peck et al. (1999) observed a 99.5% decrease in macrofaunal abundance in newly formed scours compared with unscoured zones in the South Orkney Islands whilst similarly, Smale et al. (2007a) recorded a 94.9% average decrease in abundance on the Antarctic Peninsula. Significantly lower abundance, biomass and richness values in scours compared with unscoured areas have also been recorded in deeper waters in the Weddell Sea (Gerdes et al. 2003), and in coastal waters of McMurdo Sound in the Ross Sea (Lenihan and Oliver 1995) and the Canadian Arctic (Conlan et al. 1998) (Fig. 2). However, iceberg impacts also resuspend benthic material, including food (Peck et al. 2005) and larvae (Dayton 1989).
Figure 2. Total faunal abundance and species richness of macrofauna (i.e. ≥0.5 mm size) found in scours compared with adjacent undisturbed zones. The Signy Island study sampled a single scour at 9 m depth, 1 d after formation (Peck et al. 1999). The Adelaide Island data are means (±SE) from 12 scours of known age (1–20 d old) across 3 sites at 10–17 m depth (Smale et al. 2007a). Data from the Weddell Sea are means from 4 scours, thought to be recent, occurring at 225–360 m depth. The species richness data are for polychaetes only (Gerdes et al. 2003). Data from McMurdo Sound are means of 3 scours of unknown age (but thought to be recent) from 3 sites, at 25–40 m depth (Lenihan and Oliver 1995). Finally, the study conducted in Canada sampled 4 scours, also thought to be recent, across 3 sites at 6–15 m depth (Conlan et al. 1998).
Download figure to PowerPoint
Following the retreat of the ice from the impacted site benthic recolonisation and succession processes commence, although it is currently unclear to what extent these processes are predictable. For example, at some shallow polar locations the pioneer assemblage and early patterns of development are relatively constant across scours (Conlan and Kvitek 2005, Smale et al. in press), whilst at deeper shelf sites pioneer species may vary considerably between disturbed areas (Gutt et al. 1996). Even so, a few general points regarding post-scour recovery can be made. Following disturbance, a diverse group of motile scavengers, usually dominated by amphipods, echinoderms and nemerteans, move into scoured zones to feed on the damaged benthos (Richardson and Hedgepeth 1977). Subsequently the scour depression begins to infill with sediment and biogenic matter such as algal detritus, whilst motile deposit feeders and other scavengers move through/ across the substrata and into the scour (Peck et al. 1999, Conlan and Kvitek 2005). The scoured area, now exposed to new recruits, is soon recolonised by early pioneers. For example, on hard substrata the bryozoan Fenestrulina rugula and spirorbid polychaete worms (Brown et al. 2004, Bowden et al. 2006) are often conspicuous members of early post-disturbance assemblages and in soft sediments the polychaetes Capitella sp., Leitoscoloploskerguelensis and Kefersteineni fauveli are of comparable importance (Lenihan and Oliver 1995). Finally, small fauna, such as bivalves and ostracods, may be advected into scour depressions by strong currents or water movements induced by storms or nearby iceberg groundings (Peck et al. 1999).
These recolonisation processes act at differing timescales and, as such, each stage of benthic recovery may be characterised by a different community. Therefore, even within a single site there may be hundreds of iceberg scours of differing ages, which have different physical properties and support, to some degree, a distinct assemblage. However, at intensely disturbed locations assemblages can be constantly held at early successional stages, as the period between disturbance events is too short to allow any prolonged assemblage development. This is particularly evident in benthic assemblages inhabiting the intertidal and very shallow subtidal zones near glaciers or ice cliffs (Pugh and Davenport 1997, Brown et al. 2004). Also at the scale of habitat, ice disturbance almost certainly drives the patterns of community change along a depth gradient in the shallow waters along the WAP and elsewhere. Most floating ice is small in size, as large icebergs and sea ice floes ultimately break up into small pieces of “brash” ice. These small fragments of ice ground out in shallow waters, thus the shallows experience greatest frequency of disturbance (Smale et al. 2007b) and generally support assemblages with low species richness and complexity, and a high abundance of pioneers motile scavengers or deposit feeders. Thus, the shallows represent highly disturbed habitats and support few species, as could be predicted by environmental stress models (see Scrosati and Heaven 2007 for recent discussion). Conversely, the frequency of the disturbance in deeper waters (ca 20 m depth) is significantly reduced (Smale et al. 2007b) and as a result assemblages are considerably more diverse, abundant and massive compared with those in shallower waters. This community change occurs continuously along a depth gradient in polar coastal waters and patterns can be remarkably consistent between contrasting sites (Gambi et al. 2000, Smale 2008). However, ice scouring promotes high patchiness across the depth-gradient (horizontal variability), which in some cases may be as important as variability along the depth gradient (vertical variability).
Although catastrophic at small scales, at the regional spatial scale ice disturbance may play a key role in promoting and maintaining species richness. This is because in undisturbed communities, that is those which have reached a state of near-equilibrium but are probably still in recovery from the last glacial maxima, a number of species are out-competed and displaced (Barnes 2002, Gutt and Piepenburg 2003). However, such species may thrive in early successional stages and as the Antarctic shelf (to ca 300 m depth) is effectively a patchy mosaic of co-existing successional stages following ice disturbance, beta diversity is elevated and maintained. Increased habitat heterogeneity and niche separation, as a direct consequence of ice scouring, may explain (in part) the high biodiversity of Antarctic shelf fauna. In addition, the co-existence of different basal species is likely to provide a range of different microhabitats, which in turn can support a range of epifaunal and infaunal species (i.e. facilitation, Bruno et al. 2003).
It has been assumed for many years that post-disturbance recovery in the Southern Ocean is slow compared with lower latitudes, primarily because of slower growth rates and community development in the Antarctic, but only recently has sufficient data been collected to test this longstanding paradigm. It seems that growth rates are in fact considerably slower compared with lower latitudes (Clarke et al. 2004b, Heilmayer et al. 2004, Barnes et al. 2007) and that, at least in some locations, the rate of community development on hard substrata is an order of magnitude slower in the Southern Ocean, at least at equivalent depths (Stanwell-Smith and Barnes 1997, Bowden et al. 2006). So, community recovery in iceberg scours may take a long time, and early and intermediate successional stages may persist for hundreds of years and in doing so, promote regional diversity. It should be noted, however, that the time required for recovery is determined by the nature of the “undisturbed” community adjacent to the disturbance event. For example, in the highly disturbed shallows scour assemblages regain similarity to those outside the scour within decades (Peck et al. 1999, Conlan and Kvitek 2005, Smale et al. in press). Conversely, in deeper waters that are infrequently disturbed, scours may take hundreds of years to recover as background assemblages are dominated by slow-growing sessile species such as sponges or bryozoans, or structured by dense sponge spicule mats (Dayton 1990, Gutt 2001). Benthic community structure can also be affected over longer timescales, as relict ice scour marks (with altered sediment characteristics and topographies) may persist for thousands of years (Duncan and Goff 2001).
Other forms of ice disturbance that structure Antarctic benthic communities are the action of anchor ice and the seasonal presence of an ice foot. Anchor ice, the formation of ice platelets on the sea bed or organisms, is an important disturbance pressure acting on communities at high latitudes and is particularly evident at East Antarctic sites and the Ross Sea (Dayton et al. 1969). However, there are very few reports of its formation from the WAP and Scotia Arc region, and it is almost certainly very rare here (probably because of oceanographic conditions and the relative lack of large floating ice shelves compared with elsewhere) and it therefore falls outside the scope of this paper. The winter ice foot, on the other hand, is a prominent feature of the region and has a profound, but somewhat spatially confined, effect on its shallow water ecology. The ice foot is a “narrow fringe of ice attached to the coast” (World Meteorological Organisation definition 1970) and is formed during the winter as seawater meets the cold rocky shore and freezes. This layer of ice extends seaward and covers the intertidal and shallow subtidal zone to several metres depth, and persists for more of the year than seasonal fast ice. The encasement of the biota by the ice foot restricts the exchange of gases and water, affects substratum temperature and limits colonisation by new settlers to just a few months of the year (Barnes 1995b, 1999, Waller et al. 2006). Furthermore, the ice foot itself exerts a powerful crushing force on the underlying surface (Scrosati and Heaven 2006). At Adelaide Island, for example, the ice foot persists for up to 9 months of the year and assemblages on the surface of the intertidal and shallow subtidal zones are very species poor and consist only of transient summer grazers and no sessile species. Along some coastlines, however, the protection offered by boulders and cobbles allows the development of more complex subsurface assemblages, which may include over-wintering sessile species and their associated predators (Waller et al. 2006).
Future scenarios and likely benthic response
Currently we are in an interglacial and as such expect that a degree of the glacier and ice shelf retreat should be an expected (normal) response following emergence from a glacial maximum. Thus to evaluate changes in ocean ice-loading, the major form of disturbance to the shelf, we must consider current rapid change in the historical context outlined above. However, following the recent rapid warming of the WAP and Scotia Arc region, glacier retreat and ice shelf collapse have almost certainly occurred at a greater rate than would be expected during natural cycles. To expand, 87% of tidewater glaciers on the Antarctic Peninsula have retreated since the 1940s (Cook et al. 2005), and the rate of ice flow from land to sea along these glaciers has increased by ca 12% since the early 1990s (Pritchard and Vaughan 2007). The vast majority of these retreating maritime glaciers have not yet retreated past their grounding lines and therefore calving ice will, at least over short timescales, generate floating icebergs. Furthermore, the past two decades have seen the retreat of more than ten ice shelves on the Antarctic Peninsula (Vaughan and Doake 1996, Vaughan et al. 2003). To the west of the Antarctic Peninsula, the collapses of the Wordie and Wilkins ice shelves, and the recent retreat of the George VI Sound ice shelf, have almost certainly resulted in an increased ice loading into nearshore waters, again over short timescales. Thus, intuitively it seems that the number of icebergs calving into coastal waters in the WAP and Scotia Arc regions has increased, and will continue to do so until glaciers and ice fronts retreat past their respective grounding lines. It should be noted, however, that long-term iceberg population data are very scarce, and there is currently some debate as to whether the populations of icebergs around the Antarctic Peninsula, and indeed the Southern Ocean as a whole, have increased (Bindschadler and Rignot 2001, Ballantyne and Long 2002, Long et al. 2002).
A second environmental factor, which is also changing rapidly in the region and is linked to ice scouring frequencies, is the duration and extent of sea ice. Sea ice coverage of the Southern Ocean as a whole is highly variable between years and regions, and the region to the west of the Antarctic Peninsula is the only sector to show a significant decrease in sea ice extent. Twenty years of sea ice and air temperature data have been analysed by the Long Term Ecological Research programme at Palmer Station (Smith and Stammerjohn 2001). The study defined an area of ca 2×105 km2 to the west of the Antarctic Peninsula and used satellite imagery to detect a significant decrease in sea ice coverage in the region over the past two decades, which was mainly due to a reduction in the duration of winter sea ice (Smith and Stammerjohn 2001). Other data sets from the Bellingshausen Sea confirm this recent decrease in the extent of sea ice (Jacobs and Comiso 1993, Zwally et al. 2002). In nearshore environments along the WAP, the frequency of collisions between icebergs and the seabed is significantly reduced in winter compared with summer (Smale et al. 2007b). This is because during the winter months icebergs are “locked-in” by the surrounding sea ice, which restricts their movements and hence their potential to cause disturbance (Fig. 3). Intuitively, if icebergs are surrounded by restrictive sea ice for a shorter period of the year, annual ice scouring frequencies in shallow waters will increase.
Figure 3. Left: an aerial photograph showing icebergs “locked-in” by fast sea ice during winter, which restricts their movements. White areas are winter sea ice, whilst dark areas are patches of open water. Image reproduced from Peck et al. (2005). Right: actual frequencies of iceberg scouring at two contrasting sites at Adelaide Island, with grey boxes indicating periods of fast ice formation. Adapted from Smale et al. (2007b).
Download figure to PowerPoint
Over the century, it seems very likely that the frequency of iceberg disturbances to shelf benthos along the WAP will increase, as a consequence of greater ice loading and decreased extent and duration of sea ice. At small spatial scales (i.e. sites), increased ice scouring intensity is likely to have important implications for benthic community structure. Recent studies from the Arctic suggest that ice scouring has a positive effect on benthic richness, as sites that are occasionally disturbed by ice support a greater diversity of species than those that are disturbed rarely or never (Conlan and Kvitek 2005, Laudien et al. 2007). However, icebergs are far less numerous and considerably smaller in the Arctic compared with the Antarctic (Wadhams 2000) and the vast majority of nearshore habitats in the WAP and Scotia Arc regions are disturbed by ice within ecological time frames (Barnes 1995a, Gutt et al. 1996, Sahade et al. 1998, Brown et al. 2004). Thus, any increase in disturbance frequency in the region will more likely force assemblages at moderately disturbed sites to become increasingly similar to those at intensely disturbed sites. So, how do nearshore benthic communities shaped by frequent ice scouring differ from those shaped by a moderate frequency of disturbance? The only studies that have linked empirical disturbance frequencies with parameters of community structure have been conducted at Adelaide Island. Brown et al. (2004) quantified ice disturbance frequency at two adjacent sites and made novel and important links between ice disturbance and benthic assemblages. They recorded twice as many iceberg impacts at one site, and the bryozoan communities at this site had half the number of species, two-thirds the space occupation and twice the mortality level compared with those at the less frequently disturbed site (Brown et al. 2004). Similarly, Smale (2007) linked empirical disturbance data with megafaunal, macrofaunal and lithophylic assemblage composition. The study showed that increased ice disturbance was significantly related to decreased species richness, biomass and space coverage. Furthermore, the relative abundance of sessile macrofauna and megafauna was negatively correlated with ice scouring intensity (Smale 2007). These sessile fauna, such as sponges and erect bryozoans, provide habitats for a wide range of epifaunal species on the Antarctic shelf (Gutt and Schickan 1998) and as such play a key ecological role in structuring these communities (i.e. facilitation, Bruno et al. 2003). In summary, the experimental (rather than observational) studies suggest that increased ice disturbance in nearshore habitats along the WAP would lead to reduced species richness, benthic biomass, space coverage of hard substrata, and abundance of ecologically important large sessile taxa.
Despite a relative paucity of experimental work concerning the effects of ice scouring, there have been a number of observational studies on the influence of ice scouring on benthic communities. It seems that ice scouring may select for certain fauna, such as nematodes (Lee et al. 2001), opportunistic and mobile polychaetes and amphipods (Richardson and Hedgepeth 1977, Lenihan and Oliver 1995), pioneering bryozoans (Gutt and Starmans 2001, Brown et al. 2004), fish species (Brenner et al. 2001) and perhaps for small individuals (Smale et al. 2007a). Furthermore, habitats that are frequently impacted by icebergs may support an elevated relative abundance of secondary consumers, as ice impacts damage benthos and promote feeding opportunities for scavenging species (Richardson and Hedgepeth 1977, Conlan et al. 1998). Intensely disturbed locations are also characterised by the persistence of early successional stages (McCook and Chapman 1993, Pugh and Davenport 1997), skewed population structures (Peck and Bullough 1993, Brown et al. 2004), high levels of sublethal damage (Barnes and Arnold 2001, Clark et al. 2007), and a paucity of long-lived, large, sessile species (Sahade et al. 1998, Smale 2008). In the northern hemisphere, Scrosati and Heaven (2007) recently showed that species richness on rocky shores was negatively related to ice scour stress. Thus, if ice scouring to nearshore benthos intensifies over ecological time scales, it seems likely that communities at some locations will become dominated by opportunistic pioneers and highly motile scavengers, whilst species richness and the abundance of large sessile fauna will decline.
The frequency of ice scouring is broadly related to depth (Gutt 2000, Smale et al. 2007b); most icebergs are small and ground in relatively shallow water (i.e. 0–100 m depth) whilst only a few are large enough to impact the seabed to depths of up to 600 m (Dowdeswell and Bamber 2007). Thus, benthic assemblages inhabiting the upper shelf zone are, broadly speaking, more influenced by disturbance than those in deeper water. However, in a historical context, much shelf fauna has been repeatedly bulldozed down to the deep continental margin by the highly dynamic ice sheet, which has cyclically advanced and retreated throughout the last 2 ma. During interglacial periods, the largely defaunated shallow shelf is recolonised by fauna inhabiting either the deep continental margins or isolated ice-free refuges on the shelf itself (Thatje et al. 2005). This is evidenced by the wide bathymetric ranges of many Antarctic taxa compared with elsewhere (Brey et al. 1996), and high degree of similarity between shelf and slope fauna recorded at some Antarctic locations (Galeron et al. 1992, Cattaneo-Vietti et al. 2000). For example, a recent study of cryptobenthos at Shag Rocks (Scotia Arc) found that most continental slope species were already known from the shelf and that 6 species from 1500 m have even been found in the intertidal zone (Barnes in press). Thus, if ice disturbance to the upper shelf zone intensifies, the most likely response of typical Antarctic species would be a bathymetric shift in distribution. Such a shift might take the form of effectively a contraction of species bathymetric ranges to the lower depths, and thus might also involve changes in the abundance of secondary consumers due to altering food availability. It could be the case that an increased frequency of ice scouring along the WAP would select for species with planktonic larvae, as these species have a higher dispersal capability than brooding species and can rapidly recolonise disturbed areas or newly available space (Poulin et al. 2002). However, predictions based on developmental modes alone should be made with caution, as the predicted dispersal potential (based on reproductive strategies and larval dispersal) does not always equal the realised dispersal ability for a given species.
Benthic macroalgae, although somewhat spatially restricted and lower in species richness and biomass in Antarctica compared with elsewhere, contribute significantly to the Antarctic coastal food web, both directly and as detritus (Heywood and Whitaker 1984, Wiencke 1996). Species richness decreases with increasing latitude in the Southern Ocean, probably as a result of disturbance intensity and temperature (Heywood and Whitaker 1984). Even within a single site, intense ice disturbance in the shallows may select for certain species, which are often fast growing, poor-competitors (Quartino et al. 2001). Thus, any change in the frequency of ice scouring is likely to affect both species richness and biomass values of macroalgae assemblages in nearshore Antarctic habitats. However, the is a severe paucity of investigations on macroalgae form the Antarctic Peninsula, so it is difficult to make any detailed predictions with an acceptable degree of confidence.
In contrast to the shelf system along the WAP generally, many of the intertidal and immediate subtidal habitats in the region are likely to experience a decrease in ice disturbance intensity over the next century. This may be particularly the case in the northern part of the Antarctic Peninsula and towards the latter half of the century. It is likely that the winter icefoot, which presently covers the hard substrata of the intertidal and shallow subtidal zone to depths of up to 5 m, will decrease in its extent and duration as a consequence of regional warming. The recent warming has been relatively more pronounced during the winter months compared with summer (Turner et al. 2005), and the persistence of coastal sea ice is likely to continue to decrease over the next few decades. The winter ice foot is, on small spatial scales, a dominant structuring force and holds benthic communities at very early successional stages (Barnes 1995b, Waller et al. 2006). Thus, any reduction in its extent or duration would decrease the severity of these habitats and perhaps allow the development of more complex assemblages, particularly at sheltered sites that are protected from summer ice scour. Long-standing paradigms in intertidal polar ecology have had to be revised recently, as new evidence suggests that some species can “over-winter” in the intertidal zone beneath the ice foot and, in some respects, these habitats are not as barren as traditionally perceived (Waller et al. 2006). It is possible that a decreasing ice foot presence will permit permanent (but simple) year-round communities to establish in the very shallow waters along the WAP. Conversely, the icefoot provides some thermal insulation to intertidal communities from freezing winter temperatures (Waller et al. 2006, Scrosati and Eckersley 2007) and may protect intertidal assemblages from wave action for much of the year, so a reduction in the extent and duration of the ice foot could also have some negative impacts on community development.
Predicting the intensity of ice disturbance over timescales greater than a few decades is, of course, very uncertain, but a few broad generalisations can be made nonetheless. Currently, ca 55% of the seaward margin of the Antarctic Ice Sheet is floating, but this proportion will be considerably reduced as the ice retreats in response to sustained and increased warming. Once maritime glaciers and ice fronts retreat onto land and past their grounding lines, calving ice will remain landlocked rather than being deposited into coastal waters where it can disturb marine benthos. Similarly, although the disintegration of an ice shelf can lead to rapid coastal ice loading by accelerating ice flows (Scambos et al. 2004, Rignot et al. 2005), this can only be sustained for a finite period (though this may be centuries along the margins of Ellsworth and Marie Byrd Land). So, whilst the major Antarctic ice streams will continue to transfer ice from land to sea over long time scales, the rate of ice deposition by small maritime glaciers and ice shelves will almost certainly reduce over time. The maximum depth at which ice scouring can occur is also likely to decrease due to the thinning of ice shelves and glaciers in the region, which has recently been observed and is set to continue (Rignot and Thomas 2002, Thomas et al. 2004). Intuitively, icebergs with smaller draughts will calve from thinner ice shelves and glaciers, with a decreased frequency of ice disturbance to shelf benthos a likely consequence. So, whilst the current maximum draught of an Antarctic tabular iceberg is ca 600 m (Dowdeswell and Bamber 2007), icebergs calving in the WAP region will have considerably smaller draughts if ice thinning persists. Marine benthos inhabiting the deeper waters of the shelf (i.e. >600 m depth) will be unaffected by ice scouring in the absence of icebergs with large draughts, which could have severe implications for habitat heterogeneity and regional biodiversity. Johst et al. (2006) recently modelled the Weddell Sea system and showed that any decrease in ice scouring intensity across the shelf as a whole would be more detrimental to benthic biodiversity than an increase, due to an unevenness in the longevity of late and early successional stages.