The Southern Hemisphere westerlies in the Australasian sector over the last glacial cycle: a synthesis
Introduction
The westerlies are one of the three major zonal circulations in each hemisphere. They are driven primarily by pole–equator temperature and pressure gradients and are very well developed in the Southern Hemisphere where the temperature contrast between the Antarctica and the Southern Ocean provides a strong driving force, while the lack of land between 40°S and 60°S allows unlimited fetch lengths. They drive one of the strongest ocean surface current systems on the planet (the west wind drift) and are the cause of the evocative Roaring Forties, Fearsome Fifties and Screaming Sixties monikers for this region. The westerlies act as both a buffer and a conductor between the Antarctic and the rest of the global climate system. The Past Global Changes (PAGES) Pole–Equator–Pole II project (PEP II) is well sited to intercept this circulation in a critical location for reconstructing changes in this circulation and evaluating its role in controlling global climate. This paper reviews our knowledge of modern and past behaviour of the Westerlies and highlights patterns of change, possible forcing mechanisms and gaps in our understanding.
The average circulation in the Southern Hemisphere is strongly zonally symmetric, reflecting the central location of the Antarctic continent over the South Pole and the relative lack of landmasses between 40°S and 60°S (Fig. 1). The near-surface wind maximum lies close to 50°S on average and exhibits a zonal wave number 1 signature, being most prominent across the Indian Ocean sector where the meridional pressure and temperature gradients are strongest, and least prominent across the Pacific where meridional pressure and temperature gradients are weakest on average.
Zonal wave number 1 is the most prominent departure from zonal symmetry in the mean circulation, with wave numbers 2 and 3 also making smaller contributions in middle latitudes. The standing wave components of the circulation are generally small in comparison to their Northern Hemisphere counterparts (Hurrell et al., 1998). Much of the poleward transport of energy in the Southern Hemisphere circulation is achieved by the transient eddies (synoptic-scale baroclinic storms).
The vertical profile of zonal mean westerly winds (Fig. 2) shows the annual mean position of the sub-tropical jet near 30°S at the tropopause and the mid-latitude tropopause-level jet near 50°S. The sub-tropical jet is strongest during winter, when the meridional temperature gradient is strongest (Fig. 3). In summer, the upper-level wind maximum moves poleward to lie almost directly above the surface wind maximum. In winter, the sub-tropical jet maximizes in the Australasian sector, with a sub-polar branch of the jet lying south of New Zealand and a relative minimum in upper-level winds over New Zealand. This “split” in the upper flow encourages the development of slow-moving (blocking) anti-cyclones in the New Zealand region and further east across the South Pacific (Hurrell et al., 1998).
The role of these jets is pivotal to the track of westerly cyclonic systems. The jets circle around the Antarctic in the upper troposphere at about a 300–500 HPa elevation. Surface weather systems track beneath and somewhat poleward of the jet. The jets contain planetary long waves. The jet wave number changes seasonally from typically 0–3 node waves in winter to higher, typically, 5–7 node (Rossby) wave numbers in summer (Sturman and Tapper, 1996). Low node numbers are associated with zonal flows and high wave numbers with meridional flows. In simple terms this means, that if the jet has a high zonal wave number, surface air tends to be advected to and from the Antarctic whereas under a zero node wave, westerly cyclonic systems circle the Antarctic without interacting strongly with it. In the Northern Hemisphere, the sub-tropical jet is tied topographically to the Tibetan Plateau and the Rocky Mountains and the position of the jet is largely fixed. The Andes are not a wide enough obstruction to pin the southern sub-tropical jet in the same way. Only the Australian landmass significantly interacts with the zonal flow. It does not permanently pin the jet but it preconditions ridging in the jet under some nodal waves in the Tasman Sea sector (e.g. wave no. 3—Sturman and Tapper, 1996). This controls the direction of approach of fronts onto New Zealand and a change in node number or position significantly changes the nature of westerly flow over New Zealand, in particular determining whether largely sub-tropical or sub-polar air is advected over the country.
On the synoptic time scale (1–10 days), most variability in the circulation is in the form of baroclinic waves, which are responsible for the majority of the meridional transport of heat and momentum at all times of year (Karoly et al., 1998). Synoptic-scale eddies are most active across the Indian Ocean sector, in the region of strongest surface westerly winds. They are manifested as travelling wave packets with horizontal scale of zonal wave numbers 4–6 and occur at all times of year throughout the Southern Hemisphere mid-latitudes.
At time scales of a month or longer, one of the most prominent patterns of variability in the circulation is the “High Latitude Mode” (HLM, Kidson, 1988), also known as the Antarctic Oscillation (AO—Thompson and Wallace (2000a), Thompson and Wallace (2000b)). It represents a near-zonally symmetric seesaw in atmospheric mass between high- and mid-latitudes, and is reflected in variations in the strength and extent of the sub-polar wind maximum (the polar vortex). Its manifestation in the mean sea-level pressure field is illustrated in Fig. 4, which was calculated from NCEP reanalyses over the 52-year period 1948–1999.
The HLM does not have a strongly preferred time scale, but is known to vary randomly from the positive polarity (strengthened polar vortex) shown in Fig. 4 to the negative polarity (weakened polar vortex) according to stochastic momentum forcing provided by synoptic-scale eddies. It typically stays in one polarity for several weeks at a time, before flipping in the course of a few days to the opposite polarity (Hartmann, 1995; Kidson and Watterson, 1999).
On the seasonal to inter-annual time scale, the HLM remains a prominent mode of variability, supplemented by wave patterns across the Pacific/South American sector (Mo and Higgins, 1998; Renwick and Revell, 1999). The so-called Pacific–South American (PSA) mode appears to be forced by anomalous tropical heating (convection) associated both with the El Niño Southern Oscillation (ENSO) cycle and with shorter-term intra-seasonal variability such as the Madden–Julian Oscillation (Kiladis and Mo, 1998; Mo and Higgins, 1998). It has a strong influence on blocking anti-cyclone activity across the southeast Pacific and influences patterns of rainfall in western South America (Rutllant and Fuenzalida, 1991).
At decadal and longer time scales, the HLM is again prominent, as is its Northern Hemisphere counterpart, the Arctic Oscillation. Both modes of variability in the zonal wind have been trending towards increasing positive values (stronger polar vortex) over the last several decades, which may be related to the global temperature signal, and/or to decreases in stratospheric ozone over both poles during the past two decades (Thompson and Wallace, 2000b; Kushner et al., 2001; Shindell et al., 2001; Thompson and Solomon, 2002). In broad terms, it appears that warming (expanding) the Hadley circulations in the tropics acts to increase the strength of the zonal mean circulation, leading to a strengthening of the polar vortex and consequent isolation of the polar regions. Conversely, cooling (shrinking) the tropical Hadley circulations acts to weaken the polar vortex, allowing more meridional flow and a stronger connection between polar and mid-latitude regions.
Rossby waves are also teleconnected to the rest of the Pacific climate system and there is strong evidence (Renwick and Revell, 1999) that El Niño events force the propagation of Rossby waves from the Australian region across the south Pacific to the south Atlantic. Such wave events encourage blocking in the southeast Pacific.
Reconstruction of past circulations is complex. Unlike ocean currents, zonal winds do not contain preservable diagnostic microfloras or faunas. The primary proxy for palaeo-wind studies is wind-blown sediment. While there are distinct sedimentary characteristics for wind-blown grains (e.g. Krinsley and Doornkamp, 1973) these are rarely diagnostic and even if a wind-blown origin can be demonstrated (usually on grain size and sorting criteria) provenance studies are required to determine the sediment source and hence the direction of transport. All other proxies depend on biological or physical responses to the secondary effects of wind flows. For example, under higher wind speeds, oceanic upwelling is enhanced and this may generate a bloom in plankton as nutrient flux increases. Blooms in diatoms or other taxa are, however, generated for a large number of reasons other than wind-enhanced upwelling and demonstrating causality is difficult.
The other important sources of palaeo-wind information are annual records including tree-rings, ice cores and annually laminated lake sediments. These records typically relate to some facet of the synoptic climatology, such as the effect of rainfall and/or temperature on plant growth, rather than directly to wind flow. It is often possible to infer changes in wind fields from the reconstructed climatology, however, and these records are in many ways the most critical, as they get down to temporal scales of resolution appropriate to true climatic reconstructions.
Assuming that an aeolian forcing can be demonstrated there remains the issue of what specific aspect of circulation is being reconstructed. Markgraf et al. (1992) summarized the three main types of response to climate change in the Southern Hemisphere westerlies. These are (1) changes in the intensity of the circulation, (2) changes in the main latitudinal track of the circulation and (3) changes in the position of blocking highs and wave functions on the westerly system, causing changes in the tracks of surface frontal systems.
The first two of these changes are self-explanatory, while the third relates primarily to changes in the wave numbers of the tropopause jets. In terms of palaeo-climate reconstructions, intensity and change of track signals are relatively straightforward to identify, at least qualitatively. By contrast, reconstructing changes in the angle of attack of surface fronts is very problematic. It cannot be ignored, however, as changes in the size of the Australian landmass on glacial–interglacial timeframes due to sea-level change, mean that the tropopause jets will not have maintained the same patterns through glacial cycles. In particular, ridging locations are almost certain to have changed.
For palaeo-wind intensity and track studies based on aeolian sediments there are two components, wind velocity information and sediment flux rates.
Aeolian grain size is a direct indicator of past wind velocities which can be related to wind speed by a 3rd or 4th power function (e.g. Bagnold, 1954). Accordingly a measure of the coarsest aeolian grain fraction is often used as a proxy for maximum wind velocities. This is useful for shorter, event-based studies. In the context of long-term distal records, such as aeolian transport from Australia to New Zealand (e.g. Hesse and McTainsh, 1999), changes in modal aeolian grain size are more useful as they represent changes in typical wind conditions rather than extreme weather events, which are noise in the long-term climate record.
Sediment flux rates measure both the emissivity of the source area, which is a measure of the availability of wind transportable particles and their erodibility, and the strength and persistence of the wind field operating over the sediments. Consequently, changes in flux are important proxies for palaeo-circulation studies but vegetation patterns are critical and the effects of aridity, burning and anthropogenic clearance need to be deconvolved from the records.
For changes in wind tracks the primary requirement is a network of well-correlated sites with quasi-continuous records. Marine-based studies are likely to be the most appropriate.
There are frequently significant problems associated with the attribution of a source area for the aeolian sediment. The traditional approach to this problem is the use of diagnostic components in the wind-blown fraction. Often these are biological. For instance, Casuarina pollen is used in New Zealand to recognize Australian sourced material. Most frequently the geochemistry of the sediments themselves are used. Clays (especially kaolinite) and quartz grains are the primary tracers and are normally used to track deposition of allogenic material in terranes free of those materials, such as the influx of quartz into basaltic terranes. Sediment size is also used. The influx of coarse quartz into deep-sea sites beyond the range of fluvially or wave reworked material is often seen as diagnostic of aeolian action.
Determining the timing of periods of enhanced westerly flow is problematic at least beyond the Last Glacial Maximum (LGM—ca. 20 ka). In the ocean, it is often possible to correlate sedimentary records to an orbitally tuned isotope record, but this is a source of a potentially significant autocorrelation. Since the curves are orbitally tuned to Milankovitch signals it means that de jure, ‘ages’ derived from the calibrations cannot be used to tie aeolian records to Milankovitch forcing, which is the most probable control on long duration changes in circulation patterns.
The more intermittent terrestrial records in Australasia are usually poor targets for orbital tuning but it is often still attempted (e.g. Carter and Lian, 2000). The advent of luminescence dating has improved the situation but even in northern New Zealand where marker tephra abound, poor age control is a major barrier to identifying periods of altered westerly circulation.
Numerous authors make statements about one or more aspects of westerly circulation history in the southern part of the PEP II transect. In this paper, we will focus on identifying records with strong numerical age control. We will discuss the type of proxy record being produced and provide an evaluation of the reliability of the proxy. In doing so, we challenge a number of preconceptions about past westerly behaviour.
Section snippets
Terrestrial dust records
It has long been established that Australian dust and biological material is transported across the Tasman during storms (e.g. Marshall, 1903). It is typically identified by its distinctive red colour. This represents oxide and sesqui-oxide coatings acquired by quartz and other minerals under sub-tropical aerobic weathering environments. Kinematic trajectory modelling of air masses demonstrates the large westerly component in flows over New Zealand, with 83% of low-lying air masses over the
Conceptual models for Southern Hemisphere westerly wind changes
Late Quaternary changes in the position and strength of the southern westerlies has been much debated. Wyrwoll et al. (2000) found evidence in their general circulation model simulations for the LGM to support earlier claims of a poleward displacement of the core of the westerlies, although the pattern was highly variable. It should be noted that this study used CLIMAP ocean temperatures which have been recently revised. On the other hand, Lamy et al (1998), Lamy et al (1999) have marine core
Critical issues and directions for future research
We identify three significant gaps in our knowledge. Firstly, we still have an inadequate knowledge of the long-term geographic and temporal patterns of change in the westerlies. Secondly, we have an inadequate handle on the climatological processes driving the patterns of change. Thirdly, we lack satisfactory modelling capabilities to test models of westerly change.
Conclusions
This review has demonstrated that there are a plethora of data that can be related to westerly circulation in the Australasian sector of the Southern Hemisphere. A number of patterns are visible in the data sets, on all time scales, and we can at least construct testable hypotheses about past circulation behaviour. Nevertheless, there are substantial gaps in our knowledge base and while studies are now at least addressing climatological issues, most of our data are not of high enough temporal
Acknowledgements
We thank numerous colleagues for access to in press and unpublished data. Rachel Reverley undertook the massive task of cross-checking all the references. Comments from the referees improved the paper considerably.
References (220)
- et al.
Climate of the last glaciation in New Zealand, based on aerosolic quartz influx in an andesitic terrain
Quaternary Research
(1992) - et al.
Holocene vegetation and paleoclimatic and paleomagnetic history from Lake Johnston, Tasmania
Quaternary Research
(2001) - et al.
Late Pleistocene glaciation of the Kosciuszko Massif, Snowy Mountains, Australia
Quaternary Research
(2001) - et al.
The timing of the Last Glacial Maximum in Australia
Quaternary Science Reviews
(2002) - et al.
Insolation values for the climate of the last 10 million years
Quaternary Science Reviews
(1991) - et al.
Luminescence chronology of loess–paleosol sequences from southern South Island, New Zealand
Quaternary Science Reviews
(2002) - et al.
Radiocarbon dating of playa-lake hydrologic changesexamples from northwestern China and central Australia
Palaeogeography, Palaeoclimatology, Palaeoecology
(1986) - et al.
Glacial to interglacial changes in non-carbonate and carbonate accumulation in the SW Pacific Ocean, New Zealand
Palaeogeography, Palaeoclimatology, Palaeoecology
(2000) Late Quaternary environmental changes in eastern and central Australia, and their climatic interpretation
Quaternary Science Reviews
(1991)- et al.
Late Pleistocene vegetation and climate history of Lake Selina, western Tasmania
Quaternary International
(1999)