http://journals.cambridge.org Downloaded: 16 Dec 2013 IP address: 117.20.56.50 Antarctic Science 22(6), 591–602 (2010) & Antarctic Science Ltd 2010 doi:10.1017/S0954102010000829 The environmental basis of ecosystem variability in Antarctica: research in the Latitudinal Gradient Project CLIVE HOWARD-WILLIAMS 1 , IAN HAWES 2 and SHULAMIT GORDON 3 1 NIWA, PO Box 8602, Christchurch, New Zealand 2 Aquatic Research Solutions Ltd, 35 Queen St., Cambridge, New Zealand 3 Antarctica New Zealand, Private Bag 4745, Christchurch, New Zealand c.howard-williams@niwa.co.nz Abstract: After a decade of research, New Zealand’s Latitudinal Gradient Project (LGP) now includes primary sites from Cape Hallett (728S) to the Darwin Glacier (808S), while additional observations extend the latitudinal transect from 848S to sub-Antarctic regions. The LGP has been structured around a hypothesis that, in a frigid continent, ice dynamics is the key ecosystem variable. For terrestrial environments, two aspects of ice dynamics appear to underlie much of the observed variability. Firstly, the aridity of the region makes the transition from ice to water a key ecological factor, and secondly, the legacy of ice dynamics dating as far back as the Pliocene is imprinted on biogeography. These factors operate at difference temporal and spatial scales and neither is monotonically related to latitude. Both are also complicated by meso-scale cross gradients of altitude and distance from the sea and micro-scale local variability. Whilst climate does vary on a broad-scale, differences within the northern and central parts of Victoria Land that the LGP has so far examined are insufficient to impose any overarching effect that can overwhelm these more local effects. The result is a multiple-scale patchwork of habitats and communities, more or less replicated across the transect, in which variability at any given latitude generally exceeds variability between latitudes. A lesser quantum of research has been directed at marine ecosystems, but here there is a similar picture of local variability dominating within the Ross Sea, with significant latitude-scale effects only emerging when transects are extended into maritime- and sub-Antarctic regions. It is implicit, but not specifically recognized in the LGP context, that a further confounding effect on the interpretation of ‘transect’ information is the multiple stressor concept that requires a simultaneous analysis of interacting (synergistic or antagonistic) factors and environmental responses. As the LGP continues to extend further south, climate is expected to become more extreme, and water availability may change sufficiently for loss of habitat and species diversity to occur. Here we discuss options for refining the LGP approach to optimize its potential for understanding variability, and the factors underpinning this, in the Ross Sea Sector. Received 31 August 2010, accepted 10 September 2010 Key words: biogeography, climate, marine ecosystems, terrestrial ecosystems, Victoria Land, water availability Introduction New Zealand’s Latitudinal Gradient Project (LGP) is based on a central hypothesis that ice-driven dynamics control the structure and function of marine, terrestrial and freshwater ecosystems along the latitudinal transect down the length of Victoria Land. The dynamics of ice encompasses a range of processes that include the extent and duration of snow, glaciers, sea ice and other ice covers. By utilizing a latitudinal transect, along which changes in the dynamics of ice and the balance points in ice-water transitions can be expected to gradually change, the LGP seeks to provide insight into how ecosystems are structured by comparing them along environmental gradients. The LGP concept also contains an underlying assumption that climate is related to latitude and, in a warming world (Steig et al. 2009), that latitude can also be used as a predictive surrogate for time. Terrestrial studies in the LGP have so far spanned a range of 14 degrees of latitude along the western side of Victoria Land from Cape Adare at (728S) to the La Gorce Mountains at (868S) (Howard-Williams et al. 2006), a gradient along which mean annual temperature varies considerably (Fig. 1). Marine studies extend from 788S in McMurdo Sound to ship-based studies as far north as 668S, from areas where sea ice is present year round to those where Southern Ocean conditions apply. In addition, in this issue complementary studies from other sectors of Antarctica provide information across similar or greater latitudinal ranges. There can be no doubt that, simply by providing access to previously data-poor locations, the LGP has broadened our understanding of ecosystems 591