Antarctic climate trends, variability and controls during the last century

1. Resume

The aim of this project is to produce an improved description of the natural variability of the Antarctic climate system, together with a better understanding of the processes driving variability and change. This will

(a) provide a baseline, against which present and future climate variations can be judged and an informed assessment made as to whether they lie within the range of previously-observed natural variation, or whether they indicate the influence of some new (possibly anthropogenic) forcing

(b) clarify feedbacks in the complex air-ice-ocean system and hence improve model predictions of the regional climate and impact assessment, in particular of the effect of global warming on ice sheet mass balance and sea level rise

(c) improve estimates of spatial variability in the region of ice core drilling sites so that the climate history derived from such ice cores can be set in context.

One of the foci of this project consists of a series of tests of the ability of a coupled atmosphere-ocean Global Circulation Model developed by the Hadley Centre to reproduce the variability of the Antarctic climate system over the last century. We expect that these tests, involving comparison with field data and the output from other more detailed models, will clarify our understanding of processes and provide further verification of GCM performance and an assessment of its ability to represent Antarctic climate variability. A particular aim is to establish whether the GCM output can provide a framework which gives confidence that the very sparse instrumental observations and ice core palaeoclimate records can be related to one another, so that a regional climate record can be developed for Antarctica covering the past hundred years.

2. Scientific Background

In Antarctica the only long records of climate come indirectly, from analysis of the chemical composition of ice cores. Ice cores collected in regions of low accumulation (such as Vostok) do not have sufficient resolution to reveal variations at the decadal time scale, but several cores collected along the spine of the Antarctic Peninsula, where the precipitation is much higher, give data which have been used to reconstruct climate variations over the last hundred years (Peel, 1992). Isotope ratios and major ion concentrations in the ice are used to deduce not only the local temperature and annual accumulation but also information on source areas for the precipitation, such as the polynya which appears in some years in the western Weddell Sea (Peel and Mulvaney, 1992; Jones et al., 1993). Valuable though these records are as a source of information on temporal variability, they are too few to define the regional climate, especially since they come from a mountainous area where spatial variability, especially of precipitation, is relatively high. An extension of coring to the east and south (Ellsworth Land, Berkner Island, Coats Land and Dronning Maud Land) is required, to provide broader spatial coverage, including less mountainous regions, where it will be much easier to deduce a regional climate from ice core data.

Instrumental climate records from Antarctica are only available for the last forty years or so and are limited to a small number of stations which are mainly located in coastal areas. Some more-widespread meteorological data have been provided by Automatic Weather Stations, and satellite observations have provided a vastly improved spatial coverage of climate-linked parameters, but only over the last 20 years. Thus these records, though a valuable source of information on spatial and interannual variability, are too short to define climate variability on the century scale.

Despite the limitations of this sparse and heterogeneous set of data, considerable progress has already been made in understanding the processes controlling climate in the Antarctic. Examination of the recent record (Raper et al. , 1984) shows that Antarctica is not a homogeneous region as far as climate variability and change are concerned. Most records from East Antarctic stations show small warming trends but these are not statistically significant because of the high degree of interannual variability in the records. However, records from the west coast of the Antarctic Peninsula clearly fall into a separate group. Although interannual variability here is even larger than in East Antarctica, long-term warming trends are so large that they appear to be statistically significant (Stark, 1994; King, 1994). The extreme interannual variability of the climate in this region has been largely attributed to complex atmosphere-ocean-sea ice interactions occurring in the Bellingshausen and Amundsen seas (Weatherly et al. , 1991; King, 1994) but it is not yet clear whether the processes responsible for driving interannual variability are also implicated in the longer term trends. Recent preliminary studies at BAS have pointed to the importance of the whole of the South Pacific sector of the Southern Ocean as a "centre of action" which drives climate variability in this region.

We shall continue to investigate the processes controlling variability by careful analysis and comparison of directly observed meteorological data, using the methods established in the first quinquennium. However, we believe that the key to further progress lies in the use of mathematical models of the atmosphere. We shall base our work on the existing output from various versions of the UK Met Office Unified Model. These include an atmospheric GCM with fixed seasonally varying sea-surface-temperatures (SSTs) and sea-ice; the same atmospheric model forced by observed SSTs and sea-ice extent; the atmospheric model linked to a sea-ice model and a mixed- layer ocean model and, finally, the atmospheric model linked to sea-ice and deep ocean models. A detailed validation of the Antarctic climate of the atmosphere-only model (Connolley and Cattle, 1994) was carried out during the first quinquennial review period. Runs of the atmospheric model, forced by observed SSTs and sea-ice extent, have been carried out by the Hadley Centre for much of the 20th century, and at high and standard resolution for the 10 year Atmospheric Models Intercomparison Project (AMIP) period. The atmospheric model, linked to sea-ice and deep ocean models, has been run nominally from 1860 to 2100, and a 1000 year run is also available. It is output from these latter, and future, such runs which we shall use to investigate variability on the century time scale.

The UM can also be used as a Numerical Weather Prediction (NWP) model for detailed case studies over periods when observed meteorological data are available for assimilation. The quality of antarctic atmospheric analyses from this and other NWP models is being assessed by the FROST (First Regional Observing Study of the Troposphere) project (Turner et al. 1995) which will be completed during this quinquennium. We have access to data from various NWP centres and are also able to make use of a sea-ice model developed at the Alfred-Wegener-Institut (AWI).

3. Research and Methodology

There are three main strands to the programme of extending the climate record of Antarctic variability and of testing the ability of the coupled ocean-atmosphere GCM to reproduce this which can in practice run in parallel: sensitivity tests, case studies and extension of the observed data sets.

(a) Sensitivity tests

Over the AMIP period the Hadley Centre atmosphere model has been run at high (NWP) and standard climate model resolution. We shall compare the spatial and temporal variability of climate parameters produced by these runs to assess their sensitivity to model grid size. This should give some indication of the ability of the coupled ocean-atmosphere GCM to reproduce spatial variability over Antarctica pending higher resolution runs of this model (which may well take place during the quinquennium). We shall then investigate the nature of the GCM variability on seasonal to century timescales.

(b) Case studies

This element of the project will involve the detailed study of processes controlling variability over relatively short time periods. Attention will be focused on the last 20 years, as during this period instrumental observations from Antarctic stations have been supplemented by reliable atmospheric analyses and satellite observations of the ocean, sea-ice and atmosphere. While it is important to understand climate variability over the whole of Antarctica, the extreme nature of observed climate variability in the West Peninsula/South Pacific Sector, together with the availability of reliable historical records from this region, makes the region a prime candidate for these studies.

For each case study a wide range of techniques will be applied to gather together the maximum amount of observed data. High resolution imagery from the ARIES satellite receiver at Rothera will be used to track synoptic and mesoscale weather systems over the British Antarctic Territory area. New algorithms developed during the PELICON project will be used to produce improved data sets of sea ice extent from passive microwave data. Surface winds over the ocean will be deduced from scatterometer data and sea surface temperature (SST) from satellite radiometers.

NWP analyses will first be checked against the observed data (not all of which will have been assimilated into the forecast) and then used to force the AWI sea ice model. Comparison of the sea ice extent given by the model and the observed sea ice cover will allow us to assess the extent to which sea ice variations are linked to atmospheric processes. The NWP analyses will also be used to determine whether there have been any shifts in position of major storm tracks. The deeper understanding of the links between ice, ocean and atmosphere which we seek to achieve by these case studies will then allow us to assess whether the current coupled ocean-atmosphere GCMs reasonably represent the short period variability of the Antarctic region.

(c) Extension of the observed data sets

This element of the project combines collection of cores from relatively unexplored sectors of Antarctica with research to improve the basis for calibrating ice-core constituents in terms of climate parameters (Mulvaney et al., 1992; 1993).

A network of intersecting shallow-coring traverses will be undertaken to investigate spatial and temporal variability over the last century. This work will form the BAS contribution to two international programmes, ITASE (International Trans-Antarctic Scientific Expedition) and EPICA (European Programme for Ice Coring in Antarctica). The traverses, on Berkner Island, inland from Halley and in Dronning Maud Land will provide data from an area where the topography and climate gradients are not as extreme as in the Antarctic Peninsula, so we expect to be able to achieve a sufficiently dense network to define the climate, and its spatial variability, for the region.

The question of how far ice core constituents reflect climate parameters is clearly of key importance. Ideally climate parameters should be measured at the ice core site and, as a result of the development of a relatively cheap, robust AWS in the last quinquennium, we are now able to do this more easily. Key parameters are air temperature and snow accumulation. We will deploy devices for continuously monitoring the height of the snow surface so that the seasonal representativeness of the ice-core record can be compared at different locations. Although we can only hope to collect a few years data in this quinquennium, these devices will, for the first time, allow us to look at accumulation variability on the sub-annual time scale. There is also the possibility of using instrumental records from manned stations elsewhere in the region to evaluate the calibration of the ice-core records. The difficulty here is that although the directly-observed data are clearly more trustworthy and extend for several decades, they may not be representative of the ice core site. Initial attempts to relate ice core accumulation to frequency of reported precipitation events in the Antarctic Peninsula have not shown a strong correlation; the station data are collected too far away from the existing ice core sites. We therefore intend to look to other areas of Antarctica for linked ice core/climate data sets. Ice cores under the influence of the Weddell Sea circulation will be the first target for study.

We will also compare ice-core climate data and manned station data indirectly, using the results of the high-resolution AMIP run used in the sensitivity tests (section 3(a)). If, over the AMIP period, both the station data and the ice core data are well-represented by the model output at the appropriate grid points, it will clearly increase our confidence in the ice core calibration. The model will also give an indication of the spatial scales over which cores are representative of climate variability.

Given that we can establish a good link between ice core and climate data we shall proceed to compare the simulation of the Antarctic climate over the last century from the coupled ocean-atmosphere GCM with the ice core records. These studies will thus provide an important database against which to compare and verify the variability of the GCM over Antarctica on these longer timescales, study of which will form a component of the research as noted in section 3(a) above.

4. Wider Implications

This project will enable a more complete description of present climate in Antarctica and a more reliable perspective on longer-term climate changes. These advances will allow early recognition of climate trends outside the range of natural climate variability, perhaps induced by anthropogenic forcing. Studies on the natural mechanisms driving climate of the past century will contribute to improved GCM predictions of future climate.

5. References

Connolley, W.M. and Cattle, H. 1994 The Antarctic climate of the UKMO unified model. Ant. Sci. 6 , 115-122.

Jones, P.D., R. Marsh, T.M.L. Wigley and D.A. Peel. 1993. Decadal timescale links between Antarctic Peninsula ice-core oxygen-18, deuterium and temperature. The Holocene 3 (1), 14-26.

King, J.C. 1994. Recent climate variability in the vicinity of the Antarctic Peninsula. Int. J. Climatol.

Mulvaney, R., G.F.J. Coulson and H.F.J. Corr. 1993. The fractionation of sea salt and acids during transport across an Antarctic ice shelf. Tellus 45B , 179-187.

Mulvaney, R., E.C. Pasteur, D.A. Peel, E.S. Saltzman and P.-Y. Whung. 1992. The ratio of MSA to non-sea-salt sulphate in Antarctic Peninsula ice cores. Tellus 44B , 295-303.

Peel, D.A. 1992. Ice core evidence from the Antarctic Peninsula region. In Climate since A.D. 1500. Eds. Bradley, R.S. and P.D. Jones. Routledge, London, 549-571.

Peel, D.A. and R. Mulvaney. 1992. Time-trends in the pattern of ocean-atmosphere exchange in an ice core from the Weddell Sea sector of Antarctica. Tellus 44B , 430-442.

Peel, D.A., R. Mulvaney and B.M. Davison. 1988. Stable-isotope/air-temperature relationships in ice cores from Dolleman Island and the Palmer Land Plateau, Antarctic Peninsula. Ann. Glaciol. 10 , 130-136.

Raper, S.C.B., T.M.L Wigley, P.R Mayes, P.D. Jones and M.J. Salinger, 1984. Variations in surface air temperatures: Part 3. The Antarctic, 1957. Mon. Wea. Rev., 112 , 1341.

Stark, P. 1994. Climatic warming in the central Antarctic Peninsula area. Weather , 49 , 215-220.

Weatherly, J.W., J.E. Walsh, H.J. Zwally .1991. Antarctic sea ice variations and seasonal air temperature relationships. J. Geophys. Res. 96 , 15119.