In the face of climate variability and change, science in Antarctica needs to address increasingly complex questions. Individual small studies are being replaced by multinational and multidisciplinary research programmes. The Latitudinal Gradient Project (LGP) is one such approach that combines a series of smaller studies under a single broad hypothesis to provide information that uses a gradient in latitude as a surrogate for environmental gradients, particularly climate. In this way latitudinal differences can be used to indicate climate change differences. The Key Questions for the LGP were developed via national workshops in Italy, New Zealand, and the USA and via two international workshops at SCAR conferences. Science and logistics are currently jointly shared by New Zealand, Italy and the USA, and cover marine and inland ecosystem studies along the Victoria Land coast from 72° to 78°S with plans for extensions to 85°S. The LGP forms part of the SCAR Programme Evolution and Biodiversity in Antarctica. This Special Issue summarizes some of the work in the first three years of the LGP (2002–2005), between McMurdo Sound and Cape Hallett, to form a basis for future comparative studies as the research shifts along the latitudinal span in the next decade.

The soils of the Seabee Hook area of Cape Hallett in northern Victoria Land, Antarctica, were mapped and characterized. Seabee Hook is a low-lying gravel spit of beach deposits built up by coastal currents carrying basalt material from nearby cliffs. Seabee Hook is the location of an Adélie penguin (Pygoscelis adeliae) colony which influences the soils with additions of guano, dead birds, eggshells and feathers. A soil-landscape model was developed and a soil association was identified between the soils formed on mounds (relict beach ridges) favoured by penguins for nests (Typic Haplorthel) and the soils in the areas between the mounds (Typic Haplorthel/Typic Aquorthel). Soils formed on the mounds inhabited by penguins contained guano in the upper 50 cm, overlying sub-rounded beach-deposited gravel and sand. Soils between mounds had a thin veneer (< 5 cm) of guano overlying basaltic gravelly sand similar to that in the lower parts of the mound soils. The soils had high concentrations of nitrogen, organic carbon, phosphorus, cadmium, zinc, copper, and increased electrical conductivity, within horizons influenced by penguin guano. Five buried penguin bones were collected from the base of soil profiles and radiocarbon dated. The dates indicate that Seabee Hook has been colonized by penguins for at least 1000 years.

Seabee Hook is a low lying gravel spit adjacent to Cape Hallett, northern Victoria Land, in the Ross Sea region of Antarctica and hosts an Adélie penguin (Pygoscelis adeliae) rookery. Dipwells were inserted to monitor changes in depth to, and volume of, groundwater and tracer tests were conducted to estimate aquifer hydraulic conductivity and groundwater velocity. During summer (November–February), meltwater forms a shallow, unconfined, aquifer perched on impermeable ice cemented soil. Groundwater extent and volume depends on the amount of snowfall as meltwater is primarily sourced from melting snow drifts. Groundwater velocity through the permeable gravel and sand was up to 7.8 m day−1, and hydraulic conductivities of 4.7 × 10−4 m s−1 to 3.7 × 10−5 m s−1 were measured. The presence of the penguin rookery, and the proximity of the sea, affects groundwater chemistry with elevated concentrations of salts (1205 mg L−1 sodium, 332 mg L−1 potassium) and nutrients (193 mg L−1 nitrate, 833 mg L−1 ammonia, 10 mg L−1 total phosphorus) compared with groundwater sourced away from the rookery, and with other terrestrial waters in Antarctica.

We have determined the flux of calcium, chloride and nitrate to the McMurdo Dry Valleys region by analysing snow pits for their chemical composition and their snow accumulation using multiple records spanning up to 48 years. The fluxes demonstrate patterns related to elevation and proximity to the ocean. In general, there is a strong relationship between the nitrate flux and snow accumulation, indicating that precipitation rates may have a great influence over the nitrogen concentrations in the soils of the valleys. Aeolian dust transport plays an important role in the deposition of some elements (e.g. Ca2+) into the McMurdo Dry Valleys' soils. Because of the antiquity of some of the soil surfaces in the McMurdo Dry Valleys regions, the accumulated atmospheric flux of salts to the soils has important ecological consequences. Although precipitation may be an important mechanism of salt deposition to the McMurdo Dry Valley surfaces, it is poorly understood because of difficulties in measurement and high losses from sublimation.

Stable oxygen analyses and snow accumulation rates from snow pits sampled in the McMurdo Dry Valleys have been used to reconstruct variations in summer temperature and moisture availability over the last four decades. The temperature data show a common interannual variability, with strong regional warmings occurring especially in 1984/85, 1995/96 and 1990/91 and profound coolings during 1977/78, 1983/84, 1988/89, 1993/94, and 1996/97. Annual snow accumulation shows a larger variance between sites, but the early 1970s, 1984, 1997, and to a lesser degree 1990/91 are characterized overall by wetter conditions, while the early and late 1980s show low snow accumulation values. Comparison of the reconstructed and measured summer temperatures with the Southern Oscillation Index (SOI) and the Antarctic Oscillation (AAO) yield statistically significant correlations, which improve when phase-relationships are considered. A distinct change in the phase relationship of the correlation is observed, with the SOI-AAO leading over the temperature records by one year before, and lagging by one year after 1988. These results suggest that over the last two decades summer temperatures are influenced by opposing El Niño Southern Oscillation and AAO forcings and support previous studies that identified a change in the Tropical-Antarctic teleconnection between the 1980s and 1990s.

The geochemistry and vertical stratification of shallow meltwater ponds at 78°S near Bratina Island (McMurdo Ice Shelf) have been determined for late winter (October) and summer (January) conditions as part of the Latitudinal Gradient Project. Of the five frozen ponds investigated in October, all were stratified with respect to conductivity, and three had highly saline basal brines beneath the ice at temperatures of −16 to −20°C. In the ice column, inclusions of saline fluid were observed in channels between ice crystals; the abundance increasing with depth and decreasing ice crystal size. In January, seven of the ten ponds investigated (including ponds sampled in October) retained conductivity stratification, whereas significant thermal stratification was observed in only three ponds (maximum ΔT = 5.5°C). Basal brines, ice and meltwaters were Na-Cl or Na-SO4 dominated. FREZCHEM52 modelling, supported by changes in ion ratios, indicated that the precipitation of mirabilite (Na2SO4.10H2O) and gypsum (CaSO4.2H2O) during progressive freezing is an important determinant in chemical evolution of the basal brine. High pH (8.8–11.2) and over-saturation with respect to dissolved oxygen (> 20 mg L−1) in summer, and the presence of sulphide ions in basal brines in winter, occurred in those ponds which experienced high biological productivity during the summer months.

Meltwater ponds in the Victoria Valley and in the Labyrinth at the head of the Wright Valley of Victoria Land were sampled in January (summer) and October (late winter) of 2004 to establish their geochemistry and stratification, and to compare this with that of coastal meltwater ponds at a similar latitude near Bratina Island. In summer, vertical profiles were measured in 14 ponds; 10 were thermally stratified (maximum ΔT = 11.5°C) and 12 demonstrated a conductivity increase (∼25x) in the lowest 10–20 cm of the water column. When 11 of these ponds were resampled in October, the ice columns were stratified with respect to conductivity and five ponds had highly saline (up to 148 mS cm−1), oxygenated basal brines present under the ice. Basal brines and summer melt waters were Na-Cl dominated, and Victoria Valley pond meltwaters were enriched in Ca relative to the Labyrinth ponds. Early gypsum precipitation directs the chemical evolution of residual brine during freezing. These ponds were enriched in NO3 relative to the coastal ponds at Bratina Island, due to dissolution of nitrate-bearing soil salts, and the reduced influence of marine aerosols and biological productivity on pond chemistry.

Data from six sites in Victoria Land (72–77°S) investigating co-variation in soil communities (microbial and invertebrate) with biogeochemical properties showthe influence of soil properties on habitat suitability varied among local landscapes as well as across climate gradients. Species richness of metazoan invertebrates (Nematoda, Tardigrada and Rotifera) was similar to previous descriptions in this region, though identification of three cryptic nematode species of Eudorylaimus through DNA analysis contributed to the understanding of controls over habitat preferences for individual species. Denaturing Gradient Gel Electrophoresis profiles revealed unexpectedly high diversity of bacteria. Distribution of distinct bacterial communities was associated with specific sites in northern and southern Victoria Land, as was the distribution of nematode and tardigrade species. Variation in soil metazoan communities was related to differences in soil organic matter, while bacterial diversity and community structure were not strongly correlated with any single soil property. There were no apparent correlations between metazoan and bacterial diversity, suggesting that controls over distribution and habitat suitability are different for bacterial and metazoan communities. Our results imply that top-down controls over bacterial diversity mediated by their metazoan consumers are not significant determinants of bacterial community structure and biomass in these ecosystems.

A network for monitoring change in the vegetation communities to assess the impact of future climate changes was established in Victoria Land (Continental Antarctica) in 2002 and 2003. The network is within the framework of the SCAR project RiSCC (Regional Sensitivity to Climate Change in Antarctic terrestrial and limnetic ecosystems), and in cooperation with LGP (Latitudinal Gradient Project) assessing how ecosystem biodiversity and structure may change with latitude and with climate change. The network is composed of 19 permanent plots at nine sites along a latitudinal gradient from Apostrophe Island (73°30′S, 167°50′E) to Granite Harbour (77°00′S, 162°26′E), with a main coastal transect and three subtransects inland from the coast. The sites are representative of the most frequent environmental conditions and the most widespread communities occurring in Continental Antarctica and in Victoria Land. The selected communities show different ecological requirements and have different potential sensitivities to climate change. The vegetation of each plot was described using the RiSCC research protocol, developed originally in the Maritime Antarctic and here adapted, for plot size, to Continental Antarctica. This paper characterizes the vegetation at the sites as the starting point of the long term monitoring.

This paper reports on the remapping of a carefully documented vegetation plot at Cape Hallett (72°19′S 170°16′E) to provide an assessment of the rates of vegetation change over decadal time scales. E.D. Rudolph, in 1962, mapped in detail the vegetation of a site approximately 28 m by 120 m at Cape Hallett, Victoria Land, Antarctica. This site was relocated and remapped in January 2004 and changes were assessed using GIS techniques. This appears to be the longest available time period for assessing vegetation change in Antarctica. The analysis indicated that considerable change had occurred in moss and algae distribution patterns and this seems to have been caused by increased water supply, particularly in wetter areas. There was also evidence of some change in lichen distribution. The extent of the change indicates that vegetation cover can be used for monitoring change in areas as extreme as the Ross Sea region. For this analysis to be successful it was important that the mapping techniques used were totally explicit and could easily be replicated. Fortunately, Rudolph had defined his cover classes and the site was also clearly marked. The application of GIS mapping techniques allows the mapping to be more explicitly defined and easily replicated.

Microscopy, DNA sequencing, and amplified fragment length polymorphism (AFLP) were used to examine variation within Nostoc commune from collections between 72 and 78°S in Victoria Land, Antarctica. Although there is considerable bias of collected material towards southern latitudes, and this material varies greatly in age (collected between 1984 and 2004), an important new phylogeographic pattern was found. DNA sequencing of the tRNAleu(UAA) region, used recently to define form species N. commune, revealed little variation between collections. AFLP analysis, however, split the collected material according to habitat (irrigated soil communities versus ponds), rather than latitude. These results suggest that environmental factors linked to latitude are not the greatest drivers of genetic variation in Victoria Land. These may operate at a lower level but would require intensive sampling within narrowly defined habitat types at a range of latitudes to uncover. We advocate extensive sampling across local environmental gradients based on water availability, comparative culturing, and development of sequence characterised amplified regions (SCARs) across a range of latitudes in future seasons of the Latitudinal Gradient Project.

Diverse microbial communities survive within the sea ice matrix and are integral to the energy base of the Southern Ocean. Here we describe initial findings of a four season survey (between 1999–2004) of community structure and biomass of microalgae within the sea ice and in the underlying water column at Cape Evans and Cape Hallett, in the Ross Sea, Antarctica as part of the Latitudinal Gradient Project. At Cape Evans, bottom-ice chlorophyll a levels ranged from 4.4 to 173 mg Chl a m−2. Dominant species were Nitzschia stellata, N. lecointei, and Entomoneis kjellmanii, while the proportion of Berkeleya adeliensis increased steadily during spring. Despite being obtained later in the season, the Cape Hallett data show considerably lower standing stocks of chlorophyll ranging from 0.11 to 36.8 mg Chl a m−2. This difference was attributed to a strong current, which may have ablated much of the bottom ice biomass and provided biomass to the water below. This loss of algae from the bottom of the ice may explain why the ice community contributed only 2% of the standing stock in the total water column. Dominant species at Cape Hallett were Nitzschia stellata, Fragilariopsis curta and Cylindrotheca closterium. The low biomass at Cape Hallett and the prevalence of smaller-celled diatoms in the bottom ice community indicate that the ice here is more typical of pack ice than fast ice. Further data will allow us to quantify and model the extent to which ice-driven dynamics control the structure and function of the sea ice ecosystem and to assess its resilience to changing sea ice conditions.

As part of the Latitudinal Gradient Project (LGP) the coastal meroplankton community is being studied along the coast of Victoria Land, in the Ross Sea, Antarctica. In this preliminary analysis the meroplankton community from Cape Hallett (72°S) is compared to that from two previously sampled sites in the south-western Ross Sea; at Cape Roberts and in McMurdo Sound (c. 77°S). Non-metric multi-dimensional scaling (MDS), a hierarchical cluster analysis and permutational MANOVA in combination show that the meroplankton composition differs significantly between the three sites. Although this preliminary analysis includes only one northern Ross Sea site, the results suggest that there are differences in meroplankton composition along the Victoria Land coast. Several larval types, including annelid trochophores, unidentified annelid larva, echinospira, barnacle nauplius, asteroid, echinoid, and ophiuroid larvae were absent or rare in the south-western Ross Sea but found abundantly at Cape Hallett. Detailed analysis of the meroplankton community at Cape Hallett and further examination of changes in the meroplankton community along the Victoria Land coast will continue in future years of the LGP.

The northern Victoria Land coastal marine environment was investigated during the late summer 2004, within the framework of the Latitudinal Gradient Project (LGP), to describe the physical, chemical and biological patterns of the water column and their relationship with the pelagic and benthic compartments, and to determine to what extent they change with latitude. A latitudinal gradient from Cape Adare to the Terra Nova Bay–Cape Russell area was determined on the basis of abiotic and trophic factors. Cape Adare had lower values of organic matter (particulate organic carbon < 150 μg l−1) available for the benthic communities, but this organic matter had good trophic quality. In Terra Nova Bay the particulate organic matter was quantitatively higher (organic carbon > 400 μg l−1), presumably reaching the bottom via faecal pellets, but was more detrital, although its nutritive value was still high (carbon protein content nearly 40%), as confirmed by the great quantity of phytopigments in the sediments (> 4.0 μg g−1). The benthic communities changed with latitude as well, partially reflecting the environmental and trophic gradient, but also showing a large within-area variability (except for the Cape Adare area), due to a complex array of variables that did not change with latitude.

In 2004 as part of the joint Latitudinal Gradient Project in the Ross Sea, cruises of RV Tangaroa and RV Italica along the Victoria Land coast and the Balleny Islands collected 142 species of mollusc (four Polyplacophora, 99 Gastropoda, 37 Bivalvia and two Scaphopoda). About 20% of these species represent new records for the Ross Sea quadrant. The species richness was found to be higher at 71–72°S (Cape Hallett–Cape Adare) and lower at 74–75°S (Terra Nova Bay–Cape Russell) using Shannon-Wiener's H', Pielou's J' and Simpson's λ indices, as well as by using species richness estimators (e.g. ICE, Chao 2, Jack 2). The Balleny Islands (65–67°S), though not exhaustively sampled, show diversity values comparable to those of Terra Nova Bay–Cape Russell. These islands, located in the main Antarctic Coastal Current, appear to represent a crossroads in the Southern Ocean, with mollusc and other invertebrate species present, that have previously been recorded only from the Weddell Sea. The higher diversity in the Cape Hallett–Cape Adare area is not easy to interpret, but could be a result of the intense iceberg scouring off the two capes, observed from seafloor mapping, a factor that is known to enhance species diversity at the regional scale. The existence of a decreasing trend in diversity towards higher latitudes along the Victoria Land coast cannot yet be shown, due to the large gaps in sampling coverage.

This paper is a contribution to the Latitudinal Gradient Project. It describes macro and epifaunal assemblages and habitats at three shallow water locations at the southern end of the western Ross Sea coast, and investigates relationships between faunal composition and environmental characteristics. Many variables (e.g. substrate type, sediment composition, depth, latitude, longitude) contributed to explaining the differences in community composition between locations, with latitude (a likely surrogate for broader scale factors, e.g. ice cover) one of the most important. The percentage explained by environmental characteristics was strongly scale dependent, decreasing with increasing scale of observation. As much as 66% and 75% of the variability in macrofaunal and epifaunal assemblages, respectively, was explained at the smallest scale (i.e. between transects within a location), compared to 9–18% and 11–32%, respectively, at the scale of the entire study. This relationship was also true for species richness and total abundance. This suggests that while small-scale habitat variability will not confound our ability to detect latitudinal gradients in future studies, adequately quantifying the environmental factors important in structuring these communities at larger (latitudinal) spatial scales will be important. Finally, large differences in habitat structure did not translate into large differences in the diversity of fauna, illustrating the difficulty of predicting faunal composition in the Ross Sea based on seafloor topography alone.

The endemic Antarctic scallop Adamussium colbecki (Mollusca: Bivalvia) represents a key species in the Ross Sea littoral benthic ecosystem, locally reaching very high densities. This species has an annual gametogenic cycle, with a summer spawning event, and a pelagic larval behaviour. This paper aims at describing population structure and genetic polymorphism (using AFLP) of the large populations inhabiting the Ross Sea (Terra Nova Bay and McMurdo Sound) in order to investigate possible genetic exchange between A. colbecki in these areas. In Terra Nova Bay, size-frequency distributions show, generally, the dominance of large individuals, although site related differences are found in the abundance of smaller size classes (less than 40 mm), suggesting that recruitment is not a regular event. All McMurdo sites are characterized by large individuals and, at least during this sampling period, recruitment is completely absent. Nuclear DNA analyses show that the largest differences are found at the largest scale (between McMurdo Sound and Terra Nova Bay), but all populations sampled, even at a smaller spatial scale, have a well-settled genetic structure, notwithstanding the pelagic larval strategy. The panmixia hypothesis has therefore to be rejected for this species.

Data from two new cruises is used to assess whether latitude plays any role in influencing broad-scale spatial patterns in echinoderm assemblage composition along the Victoria Land coast and the Balleny Islands as a contribution to the Latitudinal Gradient Project. Our results indicate that a latitudinal gradient is influencing assemblage structure in subtle and non-linear ways. The Balleny Islands system is different from the Victoria Land coast, probably because of a different biogeographic origin and current oceanographic conditions. Along the Victoria Land coast, latitude related differences arise when taking into account benthic biodiversity at different spatial scales. Alpha diversity increases from north to south, but beta diversity shows the opposite trend, although not linearly, suggesting the different importance of the iceberg disturbance along the northern Victoria Land coast.