O bservations of global average surface air temperature (SAT) show an unequivocal warming over the twentieth century 1 , however the overall trend has been interrupted by periods of weak warming or even cooling (Fig. 1). For example, warming largely stalled from the 1940s to the 1970s. Between 1975 and 2000 the overall upward SAT trend resumed, but it was not uniform, with a decade of accelerated warming from about 1975-1985 (ref. 2), as well as periods of little warming 3 . Since around 2001 a marked hiatus in global surface warming has occurred, raising questions about its cause, its likely duration and the implications for global climate change.Decadal periods of minimal warming, or even cooling, interspersing decades of rapid warming, are not inconsistent with a long-term warming trend; indeed this characterizes the interplay between steadily increasing greenhouse gas forcing and internally generated climate variability. Factors other than internal variability, such as volcanoes and changes in solar radiation, can also drive cooler decades against the backdrop of ongoing warming. Indeed, hiatus decades are expected to punctuate future warming trends, even under scenarios of rapid global warming 4,5 . Mechanisms proposed to explain the most recent observed hiatus include increased ocean heat uptake 2,3,6,7 , the prolonged solar minimum 4 and changes in atmospheric water vapour 8 and aerosols 9,10 . The cool surface waters of the eastern Pacific have also been linked to the global temperature hiatus 11 and consensus is building that the subsurface ocean, with its vast capacity for heat storage, is playing a significant role through enhanced heat uptake 2,3,7,12,13 . It remains unclear, however, where the bulk of anomalous ocean heat uptake has occurred, with the Pacific 2,3 , Atlantic 13,14 and Southern [14][15][16][17] Oceans all potential candidates. One notable aspect of the two most recent extended hiatus periods (1940-1975 and 2001-present), in contrast to periods of global SAT warming (1910SAT warming ( -1940SAT warming ( and 1976SAT warming ( -2000, is that they correspond closely to periods when the Interdecadal Pacific Oscillation 18-20 (IPO) has been in a negative phase (Fig. 1a). The IPO manifests as a low-frequency El Niño-like pattern of climate variability, with a warm tropical Pacific and weakened trade winds during its positive phase, and a cool tropical Pacific and strengthened winds during its negative phase. Recent analyses of climate model simulations suggest that hiatus decades are linked to negative phases of the IPO (refs 2,3,11). Here we examine the most recent hiatus in this context, particularly in relation to altered ocean dynamics and enhanced ocean heat uptake, and assess implications for the coming decades.To examine the ongoing hiatus compared with a period of warming, we start by considering climatic trends over the past two decades, spanning the transition from a period of global surface warming in the 1990s to the post-2000 hiatus. During this time the Pacific trade winds ...

Extremes such as summer heat waves and winter warm spells have a significant impact on the climate of Australia, with many regions experiencing an increase in the frequency and duration of these events since the mid-twentieth century. With the availability of Coupled Model Intercomparison Project phase 5 (CMIP5) climate models, projected changes in heat waves and warm spells are investigated across Australia for two future emission scenarios. For the historical period encompassing the late twentieth century (1950–2005) an ensemble mean of 15 models is able to broadly capture the observed spatial distribution in the frequency and duration of summer heat waves, despite overestimating these metrics along coastal regions. The models achieve a better comparison to observations in their simulation of the temperature anomaly of the hottest heat waves. By the end of the twenty-first century, the model ensemble mean projects the largest increase in summer heat wave frequency and duration to occur across northern tropical regions, while projecting an increase of ~3°C in the maximum temperature of the hottest southern Australian heat waves. Model consensus suggests that future winter warm spells will increase in frequency and duration at a greater rate than summer heat waves, and that the hottest events will become increasingly hotter for both seasons by century’s end. Even when referenced to a warming mean state, increases in the temperature of the hottest events are projected for southern Australia. Results also suggest that following a strong mitigation pathway in the future is more effective in reducing the frequency and duration of heat waves and warm spells in the southern regions compared to the northern tropical regions.

Future changes in the stratospheric circulation could have an important impact on northern winter tropospheric climate change, given that sea level pressure (SLP) responds not only to tropospheric circulation variations but also to vertically coherent variations in troposphere-stratosphere circulation. Here we assess northern winter stratospheric change and its potential to influence surface climate change in the Coupled Model Intercomparison Project-Phase 5 (CMIP5) multimodel ensemble. In the stratosphere at high latitudes, an easterly change in zonally averaged zonal wind is found for the majority of the CMIP5 models, under the Representative Concentration Pathway 8.5 scenario. Comparable results are also found in the 1% CO 2 increase per year projections, indicating that the stratospheric easterly change is common feature in future climate projections. This stratospheric wind change, however, shows a significant spread among the models. By using linear regression, we quantify the impact of tropical upper troposphere warming, polar amplification, and the stratospheric wind change on SLP. We find that the intermodel spread in stratospheric wind change contributes substantially to the intermodel spread in Arctic SLP change. The role of the stratosphere in determining part of the spread in SLP change is supported by the fact that the SLP change lags the stratospheric zonally averaged wind change. Taken together, these findings provide further support for the importance of simulating the coupling between the stratosphere and the troposphere, to narrow the uncertainty in the future projection of tropospheric circulation changes.

Despite global warming, total Antarctic sea ice coverage increased over 1979–2013. However, the majority of Coupled Model Intercomparison Project phase 5 models simulate a decline. Mechanisms causing this discrepancy have so far remained elusive. Here we show that weaker trends in the intensification of the Southern Hemisphere westerly wind jet simulated by the models may contribute to this disparity. During austral summer, a strengthened jet leads to increased upwelling of cooler subsurface water and strengthened equatorward transport, conducive to increased sea ice. As the majority of models underestimate summer jet trends, this cooling process is underestimated compared with observations and is insufficient to offset warming in the models. Through the sea ice-albedo feedback, models produce a high-latitude surface ocean warming and sea ice decline, contrasting the observed net cooling and sea ice increase. A realistic simulation of observed wind changes may be crucial for reproducing the recent observed sea ice increase.

A strengthening of the Amundsen Sea low from 1979 to 2013 has been shown to largely explain the observed increase in Antarctic sea ice concentration in the eastern Ross Sea and decrease in the Bellingshausen Sea. Here it is shown that while these changes are not generally seen in freely running coupled climate model simulations, they are reproduced in simulations of two independent coupled climate models: one constrained by observed sea surface temperature anomalies in the tropical Pacific and the other by observed surface wind stress in the tropics. This analysis confirms previous results and strengthens the conclusion that the phase change in the interdecadal Pacific oscillation from positive to negative over 1979–2013 contributed to the observed strengthening of the Amundsen Sea low and the associated pattern of Antarctic sea ice change during this period. New support for this conclusion is provided by simulated trends in spatial patterns of sea ice concentrations that are similar to those observed. These results highlight the importance of accounting for teleconnections from low to high latitudes in both model simulations and observations of Antarctic sea ice variability and change.

The Australian decade-long ''Millennium Drought'' broke in the summer of 2010/11 and was considered the most severe drought since instrumental records began in the 1900s. A crucial question is whether climate change played a role in inducing the rainfall deficit. The climate modes in question include the Indian Ocean dipole (IOD), affecting southern Australia in winter and spring; the southern annular mode (SAM) with an opposing influence on southern Australia in winter to that in spring; and El Niño-Southern Oscillation, affecting northern and eastern Australia in most seasons and southeastern Australia in spring through its coherence with the IOD. Furthermore, the poleward edge of the Southern Hemisphere Hadley cell, which indicates the position of the subtropical dry zone, has possible implications for recent rainfall declines in autumn. Using observations and simulations from phase 5 of the Coupled Model Intercomparison Project (CMIP5), it is shown that the drought over southwest Western Australia is partly attributable to a long-term upward SAM trend, which contributed to half of the winter rainfall reduction in this region. For southeast Australia, models simulate weak trends in the pertinent climate modes. In particular, they severely underestimate the observed poleward expansion of the subtropical dry zone and associated impacts. Thus, although climate models generally suggest that Australia's Millennium Drought was mostly due to multidecadal variability, some late-twentieth-century changes in climate modes that influence regional rainfall are partially attributable to anthropogenic greenhouse warming.

33As part of a special issue on natural hazards, this paper reviews the current state of 34 scientific knowledge of Australian heatwaves. Over recent years, progress has been 35 made in understanding both the causes of and changes to heatwaves. Relationships 36 between atmospheric heatwaves and large-scale and synoptic variability have been 37 identified, with increasing trends in heatwave intensity, frequency and duration 38 projected to continue throughout the 21st century. However, more research is required 39 to further our understanding of the dynamical interactions of atmospheric heatwaves, 40 particularly with the land surface. Research into marine heatwaves is still in its 41 infancy, with little known about driving mechanisms, and observed and future 42 changes. In order to address these knowledge gaps, recommendations include: 43 focusing on a comprehensive assessment of atmospheric heatwave dynamics; 44 understanding links with droughts; working towards a unified measurement 45 framework; and investigating observed and future trends in marine heatwaves. Such 46 work requires comprehensive and long-term collaboration activities. However, 47 benefits will extend to the international community, thus addressing global grand 48 challenges surrounding these extreme events. 49

Following a multidecade increase, Antarctic sea ice declined drastically during austral spring 2016. Suggested causes of the sea ice decline include lingering effects of the 2015/2016 extreme El Niño and a tropical Indian Ocean teleconnection to high‐latitude atmospheric circulation. Here, we conduct pacemaker experiments using a full coupled climate model forced with observed tropical sea surface temperature to examine the impact of the Indian and Pacific Oceans on southern high latitudes during spring 2016. Our experiments suggest that a Rossby wave teleconnection from the tropical Indian Ocean contributed to the sea ice decline during spring 2016, with less influence from the Pacific Ocean. However, we find considerable spread in the magnitude of sea ice anomalies across ensemble members, suggesting that while an Indian Ocean teleconnection likely played a role, intrinsic atmospheric variability and high‐latitude ocean conditions may also have been important in driving the observed 2016 spring sea ice decline.