We have examined aspects of the association between Antarctic sea‐ice and Southern Hemisphere extratropical cyclones for the period 1973–1991, on a seasonal time‐scale. It has been found that a sea‐ice–cyclone link is not apparent over the entire sea‐ice zone. However, we have found suggestions of interannual connections over certain regimes during particular seasons. Among the connections identified was an association between sea‐ice and spatial‐system density in the Amundsen Sea in summer. Also, winter cyclogenesis density in the Ross and Bellingshausen Seas appears strongly connected to ice in the same region. By analysing the results obtained through lagging and leading the sea‐ice time series by one season, we have concluded that the forcing of sea‐ice distribution by atmospheric stresses (either thermally or via wind manipulation) is a more dominant factor than anomalous sea‐ice conditions altering cyclone tracks or spatial density, a finding supported by the more widespread incidence of significant correlation when the sea‐ice time series lagged both system density and cyclogenesis by one season. Our results also show that cyclone density appeared associated more strongly with sea‐ice either ‘upstream’ or ‘downstream’, especially for cyclones in certain sections of the Weddell and Ross Seas during spring. On the other hand, winter cyclogenesis was most strongly linked to local sea‐ice.

The interannual variability of the horizontal lower-tropospheric moisture transport associated with the Australian summer monsoon has been analysed for the 1958-99 period. The 41-season climatology of moisture flux integrated between the surface and 450 hPa showed moderate levels of westerly transport in the month before Australian monsoon onset, associated with cross-equatorial flow in the Sulawesi Sea and west of Borneo. In the month after onset the westerly moisture transport strengthened dramatically in a zonal belt stretching from the Timor Sea to the Western Equatorial Pacific, constrained between the latitudes 5 and 15°S, and associated with a poleward shift in the Intertropical Convergence Zone and deepening of the monsoon trough. Vertical cross-sections showed this transport extending from the surface to the 500 hPa level. In the second and third months after onset the horizontal flow pattern remained similar, although flux magnitudes progressively decreased, and the influence of trade winds became more pronounced over northern Australia.Nine El Niño and six La Niña seasons were identified from the data set, and composite plots of the affected years revealed distinct, and in some cases surprising, alterations to the large-scale moisture transport in the tropical Australian-Indonesian region. During an El Niño it was shown that the month prior to onset, in which the moisture flux was weaker than average, yielded to a dramatically stronger than average flux during the following month, with a zone of westerly flux anomalies stretching across the north Australian coast and Arafura Sea. The period of enhanced moisture flux during an El Niño is relatively short-lived, with drier easterly anomalies asserting themselves during the following 2 months, suggesting a shorter than usual monsoon period in north Australia. In the La Niña composite, the initial month after onset shows a tendency to weaker horizontal moisture transport over the Northern Territory and Western Australia. The subsequent 2 months show positive anomalies in flux magnitude over these areas; the overall effect is to prolong the monsoon.Comparison of these results with past research has led us to suggest that the tendency for stronger (weaker) circulations to arise in the initial month of El Niño (La Niña) events is a result of mesoscale changes in soil moisture anomalies on land and offshore sea surface temperature (SST) anomalies, brought about by the large-scale alterations to SST and circulation patterns during the El Niño-Southern Oscillation. The soil moisture and SST anomalies initially act to enhance (suppress) the conditions necessary for deep convection in the El Niño (La Niña) cases via changes in land-sea thermal contrast and cloud cover.

Long‐term observations indicate that the Indian Ocean displays significant low‐frequency variability in mean sea‐level pressure, near‐surface wind, cloud and sea‐surface temperature (SST). A general circulation model is used to study the response of the atmosphere to an idealized SST anomaly pattern (warm in southern mid‐latitudes, cool in southern tropics) that captures the essence of observed multidecadal SST variability as well as that associated with ENSO in the South Indian Ocean. The major objectives are to investigate air–sea interaction mechanisms potentially associated with the variability and whether the atmospheric response to the SST is likely to lead to maintenance or damping of the original SST anomaly pattern, and on what time scale. Two types of experiment are performed to tackle these objectives. An ensemble of roughly 1‐year‐long integrations suggests that the seasonal‐scale response of the atmosphere to the imposed SST anomaly includes reduced genesis and density of cyclones in the mid‐ to higher latitudes, and an indication of a shift in their tracks relative to climatology. It is argued that these changes together with those to the near‐surface winds could be expected to lead to variations in surface fluxes that would tend to reinforce the original SST anomaly pattern on seasonal scales. A 21 year integration of the model with the SST anomaly pattern imposed throughout indicates that a low is generated near, and downstream of, the warm mid‐latitude anomaly. On decadal/multidecadal scales, the associated changes to the surface winds are argued as being likely to lead to changes in surface fluxes and in the strength of the South Indian subtropical gyre that would oppose the original anomaly. The current and previous model results together with the observations then support the idea that the observed multidecadal variability in atmospheric circulation and SST of the South Indian Ocean during the past century may have arisen through a combination of basin scale atmosphere–ocean interaction and a remotely forced component. © 1998 Royal Meteorological Society

No abstract