Seasonal climate summary for Australia and the southern hemisphere (summer 2018–19): extreme heat and flooding prominent

Keywords: coastal inundation, extreme heat, flooding, heatwave, neutral ENSO season, positive OLR anomalies, summer 2018–19.

The Australian summer of 2018–19 was marked by widespread above-average temperature and punctuated by extreme heat and intense precipitation events in some areas. Here we summarise the climatic setting, the key climate drivers and diagnostics used by the Bureau of Meteorology to characterise climate variability in the Australian region. Further, we investigate these patterns in the broader context of the southern hemisphere and summarise key meteorological parameters such as rainfall and temperature. Finally, we identify some key extreme and noteworthy weather and climate events of the austral summer of 2018–19. Unless otherwise stated, the main sources of information are analyses prepared by, or derived from datasets curated by, the Australian Bureau of Meteorology.

In the southern hemisphere, the months of December 2018, January 2019 and February 2019 were not characterised by the influence of any specific climate drivers, with the El Niño Southern Oscillation (ENSO), Indian Ocean Dipole (IOD) and Southern Annular Mode (SAM) all taking values within the typical ranges (e.g. one standard deviation) of their respective indices, as shown in sea-surface temperature (SST) patterns in Fig. 1.

Fig. 1.  Sea-surface temperature analyses by the Bureau of Meteorology for (a) December 2018, (b) January 2019 and (c) February 2019.

Fig. 2 shows representative values of the Southern Oscillation Index (SOI), ENSO 5VAR Index (5VAR) and Dipole Mode Index (DMI) in the context of preceding seasons. If the Troup (1965) SOI (Fig. 2a) has sustained negative values of SOI below –7 this is often indicative of El Niño episodes, whereas persistently positive values of SOI above +7 are typical of a La Niña episode. The 5VAR is a composite monthly ENSO index, calculated as the standardised amplitude of the first principal component of the monthly Darwin and Tahiti mean sea level pressure (MSLP) and monthly indices NINO3, NINO3.4 and NINO4 SSTs (Fig. 2b; Kuleshov et al. 2009). Values of the 5VAR that are more than one standard deviation are typically associated with El Niño for positive values, while negative 5VAR values of a similar magnitude are indicative of La Niña. Although both metrics have values close to the El Niño thresholds neither exceed it, meaning the summer of 2018–19 is classified as a neutral ENSO season.

Fig. 2.  Time series of (a) SOI, (b) 5VAR and (c) DMI, placing the values of these diagnostic indicators in a longer-term context. SOI is plotted with monthly values and accompanying five-month weighted averaging (6 for central month, 4 for months one offset from central month and 1 for months two offset from central month), with red horizontal line indicating the El Niño threshold and the blue horizontal line indicating the La Niña threshold. 5VAR is plotted with monthly values and accompanying three-month averages with red horizontal line indicating the El Niño threshold and the blue horizontal line indicating the La Niña threshold. DMI is plotted with weekly values and accompanying five-week moving mean with red horizontal line indicating the positive IOD threshold and the blue horizontal line indicating the negative IOD threshold. Plots provided by Bureau of Meteorology.

The IOD is said to be in a positive phase when values of the DMI (Fig. 2c) are greater than 0.4°C, neutral when the DMI is sustained between –0.4°C and 0.4°C and negative when DMI values are less than –0.4°C. The influence of the IOD on Australian climate is typically weak during December to April. This is due to the monsoon trough shifting south over the tropical Indian Ocean and changing the overall wind circulation, which in turn prevents an IOD ocean temperature pattern from being able to form. This was true of summer 2018–19, classified as an IOD neutral season.

The Madden–Julian Oscillation (MJO) is a tropical convective wave anomaly which develops in the Indian Ocean and propagates eastwards into the Pacific Ocean (Madden and Julian 1971, 1972, 1994). The MJO takes approximately 30 to 60 days to reach the western Pacific, with a frequency of six to twelve events per year (Donald et al. 2004). When the MJO is in an active phase, it is associated with areas of increased and decreased tropical convection, with effects on the southern hemisphere often weakening during early autumn, before transitioning to the northern hemisphere. The diagnostic used operationally by the Bureau of Meteorology to identify the phase of the MJO is the Real-time Multivariate MJO index (RMM, Wheeler and Hendon 2004). An MJO event in a specific phase occurs when the line moves outside the central circle and into the octant corresponding with that phase. The RMM diagrams for October–December 2018 and January–March 2019 are shown in Fig. 3. During December to February, MJO phases five and six typically correspond with wetter-than-average conditions across northern Australia, and the MJO was active in these phases during late-December, early-January and again in late-January.

Fig. 3.  Phase-space representation of the Real-time Multivariate MJO index (RMM, Wheeler and Hendon 2004). (a) Shows daily values with the months of October 2018 in red, November 2018 in green, and December 2018 in blue. (b) Shows daily values with the months of January 2019 in red, February 2019 in green, and March 2019 in blue.

Outgoing longwave radiation (OLR) in the equatorial Pacific Ocean can be used as an indicator of enhanced or suppressed tropical convection. Increased positive OLR anomalies typify a regime of reduced convective activity, a reduction in cloudiness and, usually, rainfall. Conversely, negative OLR anomalies indicate enhanced convection, increased cloudiness and chances of increased rainfall. During La Niña, decreased convection (increased OLR) can be seen near the Date Line, whereas increased cloudiness (decreased OLR) near the Date Line usually occurs during El Niño. Similarly, when Australia is under the influence of a negative IOD event, OLR anomalies are negative over the eastern Indian Ocean where increased convection occurs. During summer 2018–19 there were positive OLR anomalies over much of the land area in the southern hemisphere, including Australia (Fig. 4).

The MSLP pattern for summer 2018–19 is shown in Fig. 5a, computed using data from the 0000 UTC daily analyses of the Bureau of Meteorology’s Australian Community Climate and Earth System Simulator (ACCESS) model¹. MSLP anomalies are shown in Fig. 5b, relative to 1979–2000, climatology obtained from the National Center for Environmental Prediction (NCEP) II Reanalysis data (Kanamitsu et al. 2002). The MSLP anomaly field is not shown over areas of elevated topography (grey shading). In general, the MSLPs over Australia and South America were higher than usual and lower than usual near Antarctica, broadly consistent with a positive SAM. For more information on the SAM index from the Climate Prediction Center (NOAA), see http://www.cpc.ncep.noaa.gov/products/precip/CWlink/daily_ao_index/aao/aao.shtml. Fig. 6 shows December 2018 to February 2019 low-level (850 hPa) wind anomalies which indicates periods of strong westerly wind bursts in mid-December and mid-February. Winds are computed from ACCESS and anomalies with respect to the 22-year 1979–2000 NCEP climatology.

Fig. 4.  (a) Average OLR totals and (b) anomalies for summer 2018–19. Anomalies calculated with respect to a base period of 1979–2010. Interpolated OLR data provided by the NOAA/OAR/ESRL PSD, Boulder, Colorado, USA. Available at: http://www.bom.gov.au/climate/mjo/#tabs=Cloudiness.

Fig. 5.  (a) Southern hemisphere mean sea level pressure (MSLP) pattern and (b) anomalies for summer 2018–19. Anomalies calculated with respect to base period of 1979–2000.

Fig. 6.  Low-level (850 hPa) wind anomalies. Winds are computed from ACCESS and anomalies with respect to the 22-year 1979–2000 NCEP climatology. Plot provided by Bureau of Meteorology (M. Wheeler pers. comm.).

The climate drivers and their diagnostic indicators discussed above are not the sole influences on the observed meteorological and oceanographic conditions in the southern hemisphere over the summer 2018–19 period. Synoptic weather systems and changes already observed in the climate system due to anthropogenic greenhouse gas emissions (IPCC 2014; Bureau of Meteorology and CSIRO 2020) are two further notable contributors of the state of the atmosphere and oceans at various points in the season and across the season as a whole.

Temperatures were well above the long-term average over the southern hemisphere in summer 2018–19. Almost all land areas outside the Antarctic had seasonal mean temperatures near or above the 1981–2010 average, with the only significant cool SST anomalies being in the eastern Indian Ocean. It was the hottest austral summer on record in the NOAA² (Huang et al. 2020) dataset, second-highest in GISS³ (Lenssen et al. 2019) and third in the HadCRUT4⁴ dataset, using a simple average of the December, January and February ensemble medians (Morice et al. 2012). There were 27 tropical cyclones in the 2018–19 season (World Meteorological Organization 2020), the greatest number of tropical cyclones observed in a season since the 2008–09 season. The South Indian Ocean region had 13 hurricane-intensity cyclones, the equal-highest number ever recorded. Two tropical cyclones in the Australian region that exceeded Category 2 strength (and hence, were Severe Tropical Cyclones) were Riley off the Kimberley coast in January, and Oma in the South Pacific in February.

In Australia, December 2018 broke records for highest nationally averaged mean, maximum and minimum temperature and was also overall drier than average (Bureau of Meteorology 2019a). The mean maximum temperature across the month was 2.41°C above average, and every state except Tasmania recorded a mean maximum temperature between 2°C and 3°C above average. The Northern Territory recorded a 3.28°C anomaly, which was its highest on record. Mean temperature anomalies were highest or second highest in every state in December, except Queensland which ranked third highest. January was similarly hot, but in February top-10 mean temperatures were recorded only in Western Australia and the Northern Territory, despite the Australian average being the fourth highest on record. All other mainland states recorded above-average temperatures with Tasmania recording below-average temperature, driven by cooler daytime maximum temperatures. Further discussion on January’s record-breaking temperatures is included in Section 4.2.

During December, rainfall was below average in every state and territory, except Victoria. This dry tendency continued throughout the summer with all states and the Northern Territory being drier than usual in January and February, except Queensland and Tasmania which recorded above-average rainfall totals in February. Tasmania’s rainfall anomaly of –79% was its second lowest on record. This intensified existing rainfall deficiencies. A summary of monthly state averages is provided in Table 1. Monthly maxima and averages are available on the Bureau of Meteorology website⁵. Fig. 7 shows the rainfall totals and deciles nationally, and Fig. 8 shows maximum and minimum temperature anomalies and deciles.

Table 1.  Area averaged rainfall totals and maximum and minimum temperature anomalies and ranking (if in top 10 highest or lowest) for Australia and individual states/territories over summer 2018–19

Fig. 7.  Rainfall (a) totals and (b) deciles. Decile ranges based on grid-point data with respect to all available data 1900–2020.

Fig. 8.  Minimum temperature (a) anomalies and (b) deciles and maximum temperature (c) anomalies and (d) deciles. Decile ranges based on grid-point data with respect to all available data 1900–2020.

New Zealand experienced its third-warmest summer on record, with all regions recording anomalies greater than 0.5°C, with Hastings in the Hawke’s Bay region of the North Island recording an anomaly of 2.4°C, its third-highest value since records began in 1965. Numerous locations recorded their highest summer average daytime maximum temperature including at Appleby in the Tasman district of the northern South Island where data extend back to 1932. Rainfall across New Zealand was overall below-average especially on the South Island, including in drought affected areas in the Tasman district. On the west coast, Hokitika recorded its driest summer in over 150 years. Some areas of the North Island recorded above-average rainfall, but this was mostly attributed to single very wet days interspersing largely drier-than-usual conditions⁶.

South Africa recorded above-average temperatures and below-average rainfall across much of the region in December except for western parts of Northern Cape and Western Cape which had near-normal rainfall and typically cooler temperatures. Much of the eastern half of the country experienced maximum temperature anomalies of greater than 3°C. These above-average temperatures and below-average rainfalls remained throughout January for much of the region although the area of greater than 3°C anomalies contracted to northern and central parts of the country. February was milder, with approximately equal proportions of near-, below- and above-average maximum temperatures. Some areas were classified as ‘somewhat dry’ but for most of the country, precipitation was near- or above-normal⁷.

The SSTs were generally near or above average in the southern hemisphere during the summer of 2018–19 (Fig. 1). Notably, a large area of above-average temperatures persisted in the Tasman Sea, peaking in January with a large area of +4.0°C above the monthly average. This moderate-intensity marine heatwave (Hobday et al. 2016, using visualisation of Schlegel 2020) was sustained over much of the region over the summer period but peaked in January 2019 as a strong to severe heatwave in the region west of New Zealand. A more short-lived heatwave of similar intensity and spatial extent also occurred near 240°E, 30°S in December. The February Antarctic sea ice extent⁸ was 9.41 million km², the fifth-lowest recorded since 1979 (Fetterer et al. 2017), resulting from the second-warmest year on record (Blunden and Arndt 2020).

This section is not intended to be exhaustive but rather provide examples of high-impact or notable events and how these affected societies, economies and the environment. Notable Australian events that will be considered here are the extreme heat in January and the extreme rainfall and flooding in February. Further discussions on these are included in Section 4.1 (rainfall and flooding) and Section 4.2 (the heatwave), as is an example of a tide-driven coastal flooding event (Section 4.3).

South America also experienced a heatwave in late-January and early February, culminating in the southernmost observation of a temperature exceeding 30°C at 53.8°S in Rio Grande, Argentina, the warmest temperature ever recorded so far south (Blunden and Arndt 2020; World Meteorological Organization 2020). Peru experienced its most intense heatwave for 30 years and Paraguay recorded its hottest days on record, while countries closer to the equator experienced droughts and water shortages (Blunden and Arndt 2020). January 2019 also saw major flooding in northern Argentina, Uruguay and southern Brazil resulting in losses estimated at $US2.5 billion (World Meteorological Organization 2020).

Indian Ocean islands experienced both extreme wet (Seychelles had its second wettest year out of the last 47) and extreme dry (Réunion Island had its driest rainy season on record), while Southern Africa was also generally dry (Blunden and Arndt 2020).

New Zealand experienced several extreme weather events that required evacuations of people in affected areas (NIWA 2019). In February more than 1000 people were evacuated as a large fire burned for 20 days in the Tasman region. One home was destroyed, and 2300 hectares were burnt. On 2 December, heavy rain led to the evacuation of homes in Hamilton and over 100 campers were evacuated from Waihi Beach as high tides coincided with heavy rainfall on Christmas Eve. Tropical Cyclone Oma caused large agricultural impacts due to heavy rains and floods in the South Pacific, especially Vanuatu and New Caledonia (Blunden and Arndt 2020).

Five named tropical cyclones were reported in the Australian region between 1 December 2018 and 28 February 2019: Owen, Kenanga, Penny, Riley and Oma. In addition to these, an un-named tropical depression 13U that was embedded in the monsoon trough brought flooding rains to much of northern Queensland. This was arguably the tropical weather system associated with the most intense impacts for the season (Bureau of Meteorology 2019b), despite it never reaching tropical cyclone intensity in the pressure- and wind-based metrics used by the Bureau of Meteorology. Severe Tropical Cyclone Owen was the primary reason that north-eastern Queensland did not suffer the same rainfall deficiencies as the rest of Australia in December (Bureau of Meteorology 2019a). The flooding associated with the 13U low-pressure system was centred principally off north-western Queensland and the Townsville region, with Townsville recording 1052.8 mm of rain in the seven days to 4 February 2019, and 1259.8 mm in the ten days to 8 February. These totals were 166.6 and 334.3 mm higher than the previous seven- and ten-day accumulation records, set in January 1998 and January 1953 respectively. Further information on the climatological and meteorological aspects of the heavy rainfall leading to this flood event are contained in Bureau of Meteorology (2019c).

This record-breaking rain produced record-breaking flooding in Townsville and surrounds. The most significant impacts were thousands of properties affected in Townsville, Giru (on the Haughton River) isolated for more than a week, the closure of key highways for over a week, and the spilling of the Burdekin Falls Dam. There was also substantial flooding elsewhere in northern Queensland, with the Daintree River and the Daintree Village recording their highest flood height in 118 years. Major flood thresholds were exceeded at Birdsville and Diamantina Lakes (Diamantina River), Sellheim (Burdekin River), Giru (Haughton River), Townsville Aplin Weir (Ross River), Floraville (Leichhardt River), Richmond and Walkers Bend (Flinders River), Julia Creek (Julia Creek) and Daintree Village (Daintree River). Impacts consistent with this level were observed across northern Queensland – closure of major rail and road routes, flooding of buildings above the floor level and impacts to utility services. Further information and technical details on the flooding during January and February 2019 are provided in Bureau of Meteorology (2019d).

The summer of 2018–19 was the hottest on record for Australia, on average 0.86°C hotter than any previously recorded summer. The heatwave conditions (defined by considering three-day periods relative to climatological means and recent observations) were most extreme in January. January was 0.99°C hotter than any previously recorded month, and centred on NSW, where it was more than 2°C hotter than the previous record. Victoria, Tasmania, Queensland and the Northern Territory also recorded their hottest month on record. The heatwave was also exceptionally prolonged and persistent, with 28 days exceeding the monthly 99th percentile across the summer, 2.5 times more often than in any previous summer. The area of highest-on-record and other deciles is shown in Fig. 9. Further information, including climatological context and meteorological information is available in Bureau of Meteorology (2019e). This period of extended heat was a factor that contributed to long-lived and large bushfires in Tasmania during January, with an area of approximately 2.6% of the total area of Tasmania burnt as the result of several large fires, many in remote areas (Bureau of Meteorology 2019f).

Fig. 9.  January 2019 mean temperature deciles show large area of Australia as 10th decile or highest-on-record. Daily mean temperature is calculated as the simple average of the daily minimum and daily maximum temperature.

Although the impacts associated with this event may not have been as severe as other events this season, or compared to historical riverine flood events, this event is notable as an example of a tide-driven flood event, occurring with only limited effects of statistically-extreme weather or ocean events. Such events have been documented in the United States (e.g. Ray and Foster 2016), but not yet documented in detail in the Australian scientific literature – hence motivating its inclusion here.

The high tide for 20 February was predicted to be the second-highest tide of the year. At this time Ex-Tropical Cyclone Oma was offshore, with some probability that it may move towards Brisbane, which generated high interest on both conventional and social media platforms. This media interest resulted in this widespread coastal inundation in estuaries and tidal rivers along the east coast being widely reported. Impacts were reported from Ku Ring Gai Chase National Park north of Sydney to Cairns in northern Queensland between 18 and 22 February, with the most reports of flooding around Brisbane and the Gold Coast. The maximum observed hourly sea level during this event was 2.78 m on 20 February, 17 cm above the minimum flood threshold level defined by Hague et al. (2019). The top-of-hour tide predictions generated using R TideHarmonics (Stephenson 2016) indicate that this height was 7 cm higher than would have been expected during typical meteorological and oceanographic conditions (i.e. the non-tidal residual). A sea level of 2.78 m is not a particularly rare occurrence in Brisbane – it has occurred for 94 hours (over 79 days) in the 40-year period 1980–2019. Similarly, astronomical tides of at least 2.71 m have been predicted for 93 hours (over 84 days). Residuals of 7 cm or more are even more common, occurring approximately 17% of the time.

The key meteorological and oceanographic events in the southern hemisphere summer of 2018–19 have been summarised herein. This includes the first published accounts of a moderate-intensity marine heatwave in the Tasman Sea during January and a widely documented tide-only flooding event in Brisbane, Australia. Further, we provide the broader climatological context for these and other key events and weather impacts. ENSO and IOD, the key drivers of climatic variability on the seasonal scale for Australia, were in neutral conditions. However, extreme rainfall and temperatures were still observed across Australia, and more broadly, the southern hemisphere. Australia experienced its hottest summer on record and its hottest January on record which was 0.99°C hotter than any previously recorded month, with extreme heat centred on NSW and Victoria.

The author declares no conflicts of interest.