The Great Barrier Reef Experiences Coral Bleaching During El Nin˜O–Southern Oscillation Neutral Conditions

Home , Boulia Airport, Coral Sea, Oodnadatta Airport

CSIRO PUBLISHING Journal of Southern Hemisphere Earth Systems Science, 2019, 69, 310–330 Seasonal Climate Summary https://doi.org/10.1071/ES19006

Seasonal climate summary for the southern hemisphere (autumn 2017): the Great Barrier Reef experiences coral bleaching during El Nin˜o–Southern Oscillation neutral conditions

Grant A. Smith

Bureau of Meteorology, 700 Collins Street, Melbourne, Vic., Australia. Email: [email protected]

Abstract. Austral autumn 2017 was classified as neutral in terms of the El Ninõ–Southern Oscillation (ENSO), although tropical rainfall and sub-surface Pacific Ocean temperature anomalies were indicative of a weak La Ninã. Despite this, autumn 2017 was anomalously warm for most of Australia, consistent with the warming trend that has been observed for the last several decades due to global warming. The mean temperatures for Queensland, New South Wales, Victoria, Tasmania and South Australia were all amongst the top 10. The mean maximum temperature for all of Australia was seventh warmest on record, and amongst the top 10 for all states but Western Australia, with a region of warmest maximum temperature on record in western Queensland. The mean minimum temperature was also above average nationally, and amongst top 10 for Queensland, Victoria and Tasmania. In terms of rainfall, there were very mixed results, with wetter than average for the east coast, western Victoria and parts of Western Australia, and drier than average for western Tasmania, western Queensland, the southeastern portion of the Northern Territory and the far western portion of Western Australia. Dry conditions in Tasmania and southwest Western Australia were likely due to a positive Southern Annular Mode, and the broader west coast and central dry conditions were likely due to cooler eastern Indian Ocean sea-surface temperatures (SSTs) that limited the supply of moisture available to the atmosphere across the country. Other significant events during autumn 2017 were the coral bleaching in the Great Barrier Reef (GBR), cyclone Debbie and much lower than average Antarctic sea-ice extent. Coral bleaching in the GBR is usually associated on broad scales with strong El Ninõ events but is becoming more common in ENSO neutral years due to global warming. The southern GBR was saved from warm SST anomalies by severe tropical cyclone Debbie which caused ocean cooling in late March and flooding in Queensland and New South Wales. The Antarctic sea-ice extent was second lowest on record for autumn, with the March extent being lowest on record.

Received 7 November 2018, accepted 16 May 2019, published online 11 June 2020

1 Introduction 2 Pacific and Indian Basin climate indices Australautumnin2017followedayearofrecordhigh- 2.1 Southern Oscillation Index surface temperature in 2016 that was dominated by a strong The Troup Southern Oscillation Index (SOI) for the period El Ninõ at the beginning of the year and transitioned to weak La Ninã–like conditions by the end of 2016 and start of 2017 January 2013 to May 2017 is shown in Fig. 1 (Troup 1965), (National Oceanic and Atmospheric Administration 2017). In together with a 5-month weighted moving average to smooth Australia for 2016, temperatures were very warm (fourth out volatility (Wright 1989). Following positive SOI values for warmest on record) driven by both El Ninõ and anthropogenic the second half of 2016, the first 5 months of 2017 alternated climate change (Bureau of Meteorology and CSIRO 2017). between positive and negative. The SOI values for March, The Australian national average rainfall in 2016 was ,17% April and May were þ5.1, 6.3, þ0.5 respectively, giving a above average. neutral seasonal average of 0.2. The 5-month weighted This summary reviews the climate patterns for Austral moving average finished negative at 3.3 at the end of autumn. autumn 2017, with particular attention given to the Australasian The monthly mean sea-level pressure (MSLP) anomalies for and Pacific regions. The main sources of information for this both Darwin (provided by the Bureau of Meteorology) and report are analyses prepared by the Bureau of Meteorology using Tahiti (provided by the Me´teó France inter-regional direction data sourced from a range of centres and data sets. for French Polynesia) were near to neutral in the 3-month

Journal compilation Ó BoM 2019 Open Access CC BY-NC-ND www.publish.csiro.au/journals/es BoM climate summary: autumn 2017 Journal of Southern Hemisphere Earth Systems Science 311

20 3 15 2 10 5 1 0 0 –5

–10 –1 –15 –2 –20

–25 –3 2013 2014 2015 2016 2017 2013 2014 2015 2016 2017 SOI 5-month weighted average 5VAR 3-month weighted average

Fig. 1. Southern Oscillation Index (SOI), from January 2013 to May 2017, Fig. 2. 5VAR composite standardised monthly El Nin˜o–Southern Oscilla- together with a 5-month binomially weighted moving average. The Troup tion (ENSO) index from January 2013 to May 2017, together with a SOI is 10 times the standardised monthly anomaly of the difference in mean weighted 3-month moving average. The principal component analysis and sea-level pressure between Tahiti and Darwin. The calculation is based on a standardisation of this ENSO index is performed over the period 1950–99. 60-year climatology (1933–92).

La Ninã signal, and therefore the ENSO indicators pointing period (between 0.2 and þ0.8 hPa). The 5-month weighted towards El Ninõ had little impact on the Australian climate for moving average was trending towards negative territory for the autumn 2017. autumn period. 2.3 Tropical convection and rainfall 2.2 Composite monthly El Ninõ–Southern Oscillation Index Outgoing long-wave radiation (OLR) is a good proxy for trop- (5VAR) and multivariate El Ninõ–Southern Oscillation Index ical convection and rainfall. Decreased OLR usually indicates The 5VAR is a composite monthly El Ninõ–Southern Oscilla- increased convection and enhanced associated cloudiness and tion (ENSO) Index developed by the Bureau of Meteorology, rainfall, and increased OLR usually indicates decreased con- calculated as the standardised amplitude of the first principal vection. During El Ninõ, OLR is often decreased near the Date component of monthly Darwin and Tahiti MSLP (data obtained Line indicating increased convection (and rainfall) as the from http://www.bom.gov.au/climate/current/soihtm1.shtml, Walker Circulation weakens, and the warm pool and associated accessed 29 April 2020) and monthly NINO3, NINO3.4 and tropical rainfall is displaced to the east. The reverse is true NINO4 sea-surface temperatures (SSTs) from the National during La Ninã events. Centers for Environment Prediction (NCEP), obtained from Standardised monthly anomalies of OLR are computed for an 8 8 8 8 ftp://ftp.cpc.ncep.noaa.gov/wd52dg/data/indices/sstoi.indices, equatorial region from 5 Sto5N and 160 E to 160 Wby accessed 29 April 2020 (Kuleshov et al. 2009). Persistent pos- NOAA’s Climate Prediction Centre (obtained from http://www. itive or negative 5VAR values in excess of 1 standard deviation cpc.ncep.noaa.gov/data/indices/olr, accessed 29 April 2020). þ þ þ 2 are typically associated with El Ninõ or La Ninã events Monthly values for autumn were 1.1, 0.4, 0.1 W m for respectively. The monthly 5VAR values for the period January March, April and May respectively, indicating that convection 2013 to May 2017 are shown in Fig. 2, along with the 3-month was suppressed in that area during autumn. The overall autumn þ moving average. The autumn period of March/April/May 2017 mean was 0.5, with the countries in the region reporting exhibited mildly positive values of þ0.16, þ0.75 and þ0.72 drought conditions, for example, Kiribati Meteorological respectively, which were all below the 1 standard deviation Service (2017), and the northern Marshall Islands (UNOCHA threshold for El Ninõ events. ROAP 2017). The spatial patterns of seasonal OLR anomalies 8 8 The multivariate ENSO index (MEI), produced by the across the Asia–Pacific region between 40 S and 40 N for Physical Sciences Division of the Earth Systems Research autumn 2017 are shown in Fig. 3. Across Indonesia in locales Laboratory (formerly known as the US Climate Diagnostics of Sulawesi, Borneo and West Sumatra, there were reports of Centre), is a standardised anomaly derived from a number of widespread floods and landslides during the autumn period atmospheric and oceanic parameters calculated as a 2-month (Davies 2017a). Rainfall associated with cyclone Debbie can mean (Wolter and Timlin 1993, 1998). As for 5VAR, significant be seen crossing the Queensland border in Fig. 4. The OLR and positive anomalies are typically associated with El Ninõ, rainfall patterns in western Pacific and Indonesia are consistent whereas large negative anomalies indicate La Ninã. The 2017 with a weak La Ninã, which is in contrast to what the 5VAR and February/March (0.08), March/April (þ0.74) and April/May MEI indicators were suggesting for this period. (þ1.44) values were increasing over the autumn period, similar to the 5VAR, although did not reach significant magnitudes. 2.4 Indian Ocean Dipole Contrary to this, rainfall patterns (Section 2.3) and sub-surface The Indian Ocean Dipole (IOD) describes the pattern of SST Pacific warming (Section 4.2) were more indicative of a weak anomalies across the equatorial Indian Ocean. A positive phase 312 Journal of Southern Hemisphere Earth Systems Science G. A. Smith

OLR 2.5 × 2.5 NOAA REAL TIME-OLRAV (W/M–2) 20170301 0000 20170528 0000

55 45 35 25 15 5 –5 –15 –25 –35 –45 –55

Fig. 3. Outgoing long-wave radiation (OLR) anomalies for autumn 2017 (W m2). Base period is 1979–2000. The mapped region extends from 408Sto408N and 708E to 1808E. 0° Latitude

60°S120°E 30°S150°E 30°N180° 60°N 150°W 120°W 90°W 60°W Longitude March–May 2017

–1000 –800 –600 –400 –200 0 200 400 600 800 1000 Precipitation anomaly [mm/season]

Fig. 4. Precipitation anomalies for autumn 2017 (mm) for the Pacific region. Base period is 1979–95, from https:// iridl.ldeo.columbia.edu/maproom/Global/Precipitation/Seasonal.html, accessed 29 April 2020. BoM climate summary: autumn 2017 Journal of Southern Hemisphere Earth Systems Science 313

1.5 near the equator and is objectively monitored by the real-time 1.0 multivariate MJO (RMM) index as described in Wheeler and Hendon (2004). Impacts of the MJO can be felt in many regions 0.5 of the globe, including outside the tropics (Donald et al. 2004), and its frequency and strength vary from year to year (Wheeler 0 and Hendon 2004). During El Ninõ years, MJO convection anomalies tend to propagate further eastwards over the warmer –0.5 waters of the central and eastern Pacific Ocean, but the influence –1.0 of ENSO on the MJO in other locations is either weaker or not well understood (Hendon et al. 1999). –1.5 The phase-space diagram of the RMM for autumn 2017 is 2014 2015 2016 2017 shown in Fig. 6, and the evolution of tropical convection IOD 5-week weighted average anomalies along the equator with time is shown in Fig. 7.At the beginning of autumn 2017, the MJO was in an active phase in Fig. 5. Weekly Dipole Mode Index values from January 2014 to May the Indian Ocean (phase 3) but was soon indiscernible for most 2017, together with a 5-week binomially weighted moving average. Base- of March and April while over the maritime continent and line period is 1961–90. IOD, Indian Ocean Dipole. therefore not having an impact on Australian climate. The MJO strengthened back to an active phase at the end of April of the IOD is characterised by cooler than usual water near into the western hemisphere and Africa (phases 8 and 1) until Indonesia and warmer than usual water in the tropical western mid of May. Indian Ocean. This pattern is usually associated with decreased convection, and hence less rainfall in the eastern Indian Ocean and across southern Australia during winter and spring. The 4 Oceanic patterns opposite is true for a negative IOD phase. 4.1 Sea-surface temperatures IOD events can be represented by the Dipole Mode Index SST anomalies for autumn 2017 from Optimum Interpolation (DMI) (Saji et al. 1999). The DMI is the difference in SST SST (Reynolds et al. 2002) are shown in Fig. 8. The majority of anomalies between the Western Tropical Indian Ocean node 8 8 the tropical Pacific had warm anomalies greater than 0.5 C, with centred on the equator off the coast of Somalia (50–70 E and regions in the subtropics and eastern Pacific between 1 and 28C. 108S–108N) and the Southeastern Tropical Indian Ocean 8 8 The cooler anomalies in extratropical north and south Pacific (SETIO) node near Sumatra (90–110 E and 10–0 S). Sustained and warm anomalies elsewhere are consistent with the warm values of the DMI below 0.48C indicate a negative IOD event, þ 8 phase of the Pacific Decadal Oscillation that has mostly per- whereas sustained values above 0.4 C indicate a positive IOD sisted since 2014. Much of the eastern seaboard of Australia was event. An IOD event typically starts from May or June and lasts warmer than average, and conversely, cool anomalies persisted to around November. A positive IOD event is often associated in the southern half off the coast of Western Australia. The most with below-average rainfall in central and southeastern Austra- significant warm anomaly in the Australian region was along the lia in spring (Risbey et al. 2009) and is more likely to occur east coast of Tasmania extending into the Tasman Sea. alongside El Ninõ. The percentile rankings in terms of deciles are shown in Following a strong negative IOD event that dominated much Fig. 9. The darkest green indicates the lowest SST on record of 2016, the IOD was mostly weakly positive in autumn 2017. since 1981, and dark orange is the highest on record. There are The peak IOD weekly value for this period was þ0.388C þ 8 significant patches of highest on record within the western warm (Fig. 5), not quite breaching the threshold value of 0.4 C for pool, which typically boasts the highest SSTs in the world. a positive IOD event. The SETIO node was dominated by cool Regions of highest on record for autumn 2017 can be seen in the anomalies (Fig. 8), as was much of the eastern Indian Ocean. Pacific subtropics and central Indian Ocean, as well as in the far The western and central Indian Ocean saw mostly warm western Pacific Ocean. Warm waters off the coast of South anomalies. This temperature gradient likely contributed to the America have been attributed to causing disastrous flooding dry conditions in western and central Australia as it limited the rains in both Peru and Chile (Davies 2017b, 2017c). Around supply of moisture to the atmosphere across the country. Australia, there are emerging regions of highest on record around Tasmania, and conversely, small regions of lowest on 3 Madden–Julian Oscillation record off the southern Western Australia coast. The Madden–Julian Oscillation (MJO) can be characterised as a The NINO indices were positive for most of autumn 2017, burst of tropical cloud and rainfall which develops over the with NINO4 in March being the only exception having a slightly Indian Ocean and propagates eastwards over to the Pacific cool anomaly. The highest anomaly values were in the far Ocean, and sometimes around the entire global equator (Madden eastern Pacific in the NINO1þ2 region, although these exhib- and Julian 1971, 1972, 1994). The MJO typically takes 30–60 ited a weakening trend over the autumn period, as did NINO3. days to cross from the eastern Indian Ocean to the central Both NINO4 and NINO3.4 show warming trends across the Pacific, with six to twelve events per year. autumn period. The NINO3 and NINO3.4 indices are both The location of the convective phase of the MJO can be below the El Ninõ threshold of þ0.88C, although NINO1þ2 detected by looking for large-scale negative OLR anomalies values for the autumn period are the highest on record for the 314 Journal of Southern Hemisphere Earth Systems Science G. A. Smith

(RMM1, RMM2) for 1–March–2017 to 31–May–2017

Western 76Pacific

5 8 1

0 RMM2 Maritime Continent and Africa West. Hem. West.

–1 End

1 4

–2 Start

–3

Indian 2 3 Ocean

–3 –2 –1 0 123 RMM1

Fig. 6. Phase-space representation of the Madden–Julian Oscillation (MJO) index for autumn 2017. Daily values are shown with March in red, April in green and May in blue. The eight phases of the MJO and the corresponding (approximate) locations of the near-equatorial enhanced convective are labelled. index in a year that falls outside of a basin-wide El Ninõ. For autumn 2017 was classified as ENSO neutral. The event is March, only the very strong El Ninõs of 1983 and 1998 have attributed to two main factors: global warming and local weather scored a higher NINO1þ2 anomaly, with March 2017 even patterns. The positive warming trend in the Coral Sea is evident eclipsing the El Ninõ 2016 event. See Fig. 10 for locations of the across the SST anomalies in Fig. 11 and supported by IPCC NINO boxes. (2013) that states 93% of global warming has occurred in the ocean between 1971 and 2010. In terms of local weather, a 4.1.1 Coral bleaching relatively low number of summer storms occurred over the Reef The term coral bleaching refers to the whitening of corals, until late in the season, leading to increased surface heating and indicating that the coral has been stressed by a significant reduced mixing. The SST anomalies at their peak in mid-March change to its preferred environmental conditions which can lead (Fig. 12a) were subsequently cooled after the crossing of to mortality if adverse conditions persist. In March 2017, the Cyclone Debbie which made landfall near Airlie Beach (just Great Barrier Reef (GBR) experienced its second consecutive north of Mackay). The SST anomalies in the northern half of the mass coral bleaching event following the bleaching event of GBR, which was the primary region of bleaching impact, were 2016 (Hughes et al. 2017; Hughes and Kerry 2017). The largely unaffected by Cyclone Debbie. The final degree heating bleaching driver was elevated SSTs, with the Coral Sea day (DHD) metric that describes total thermal stress over the experiencing the second highest autumn SST anomaly since 2016/2017 austral summer period is shown in Fig. 12b. Areas of 1900 using the ERSSTv5 data set shown in Fig. 11 (Huang et al. highest stress are shown to be the central third of the GBR. 2017). The only higher anomaly for autumn was in the previous year (2016) which also caused bleaching in the GBR. 4.2 Equatorial Pacific sub-surface patterns Unlike the mass bleaching event in 2016 which was attrib- Cross-sectional monthly sub-surface temperature anomalies uted to both climate change and amplified by the 2015/2016 El along the equator, encompassing latitudes 28Sto28N, are shown Ninõ event (Great Barrier Reef Marine Park Authority 2017), in Fig. 13. The first 2 months of autumn 2017 had a persisting BoM climate summary: autumn 2017 Journal of Southern Hemisphere Earth Systems Science 315

OLR anomalies; daily – averaged; base period 1979–2010 1–March–2017 to 31–May–2017 March 1

April 1

May 1

0° 40°E 80°E 120°E 160°E 160°W 120°W 80°W 40°W 0° W m–2 15.S – 15.N CAWCR/Bureau of Meteorology –90 –70 –50 –30 –10 10 30 50 70

Fig. 7. Time–longitude section of daily-averaged outgoing long-wave radiation (OLR) anomalies, averaged for 158Sto158N, for the period March to May 2017. Anomalies are with respect to a base period of 1979–2010.

REYN SSTA MEAN 1 March to 31 May 2017

SSTA (°C)

4.0 3.0 2.0 1.0 0.5 0.0 –0.5 –1.0 –2.0 –3.0 –4.0

Fig. 8. Anomalies of global sea-surface temperature (SST) for austral autumn 2017 (8C). Base period is 1981–2010. 316 Journal of Southern Hemisphere Earth Systems Science G. A. Smith

REYN SST percentiles 1 March to 31 May 2017 Distribution based on gridded data

Highest on record 10 Very much above average 8–9 Above average 4–7 Average 2–3 Below average Very much 1 below average Lowest on record

Fig. 9. Deciles of global sea-surface temperature (SST) for austral autumn 2017 (8C).

30°N

0°N NINO 3.4 2 NINO 4 NINO 3 1

30°S 120°E 150°E 180° 150°W 120°W90°W 60°W

Fig. 10. Locations of NINO boxes in equatorial Pacific Ocean.

1.2 the month of May saw a significant weakening in this anomaly, resulting in largely neutral conditions throughout most of the 0.8 equatorial Pacific, with small patches of sub-surface warming in the west and a remaining small cool anomaly close to the surface 0.4 in the far east. The equatorial thermocline is a region below the surface where the temperature gradient between warm near-surface 0 water and cold deep-ocean waters is greatest. The 208C isotherm depth is generally located close to this equatorial thermocline. –0.4 Therefore, measurements of the 208C isotherm make a good SST anomaly (°C) proxy for the thermocline depth. Positive anomalies correspond –0.8 to a deeper than average thermocline and vice versa, where shifts in the depth of the 208C isotherm provide an indication of –1.2 subsequent temperature changes in SSTs; a deeper thermocline 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 results in less cold water available for upwelling, and therefore warming of surface temperatures. The time–longitude Fig. 11. Autumn sea-surface temperature (SST) anomalies in the Coral Sea (Hovmo¨ller) diagram for the 208C isotherm depth anomaly for the years 1900–2017 (ERSSTv5). Averaging region for the Coral Sea along the equator for January 2015 to May 2017, obtained from spans from latitude 48Sto268S, and longitude 1428E to 1748E. the TAO Project Office, is shown in Fig. 14. The 208C isotherm was elevated in western Pacific at the warm anomaly west of the Date Line, and a cold anomaly as low commencement of autumn 2017. By the end of autumn 2017, the as 38C stretching eastwards from the Date Line to 1008W that 208C isotherm was mostly in a neutral position in both the far looked similar to what occurs during a La Nin˜a event. However, eastern and western Pacific and slightly suppressed in the central BoM climate summary: autumn 2017 Journal of Southern Hemisphere Earth Systems Science 317

(a) IMOS 14-day Mosaic: SST Anomaly (b) IMOS 14-day Mosaic: DHD 19 March 2017 GBR region 30 March 2017 GBR region

12°S 12°S

16°S 16°S Cairns Cairns

Townsville Townsville 20°S 20°S

Mackay Mackay

Rockhampton Rockhampton Land Land 24°S 24°S No data No data Climatology: 2002–11 IMOs Climatology: 2002–11 IMOs 144°E 148°E 152°E 144°E 148°E 152°E

–4 –3 –2 –1 0 1 2 3 4 0 10 20 30 40 50 60 70 80 90 100 110 120 130140150 160 °C Days °C IDYOC070 Quality Level ≥ 3 IDYOC070 Quality Level ≥ 3 Created: 10-September-2017 06:08:41 © Bureau of Meteorology 2017 Created: 09-September-2017 18:21:23 © Bureau of Meteorology 2017

Fig. 12. Sea-surface temperature (SST) anomaly 14-day mosaics over the Coral Sea and in the Great Barrier Reef marine park (a) approximate peak of the SST anomaly in March 2017 (b) degree heating days since 1 December, calculated by summing positive 14-day mosaic anomalies. The mosaics are created by aggregating the satellite coverage over the previous 14-day period, giving preference to the most recent observations.

Pacific. By the end of autumn 2017, both sub-surface SST The MSLP anomalies are shown in Fig. 17, relative to the 1979– anomies and isotherm location were close to neutral in much 2000 climatology obtained from NCEP II Reanalysis data of the Pacific, ending the La Nin˜a-like pattern observed over the (Kanamitsu et al. 2002). summer/autumn period. The subtropical ridge formed a band of high pressure that was situated below 308S, extending across the Indian Ocean with a 4.3 Sea level maximum pressure of 1022.8 hPa. Anomalies show that the Sea-level anomalies for autumn 2017 are shown in Fig. 15 and high-pressure bands were close to climatology for autumn 2017 reference a datum of mean sea-surface height from 1993 to in the Indian Ocean and across southern Australia. The band of 2017. Both the southern Pacific Ocean and Indian Ocean had anomalously higher pressure in the Southern Ocean and lower mostly positive sea-level anomalies. Around Australia, anom- pressure over Antarctica is indicative of a weak positive South- alies were also positive, with the highest anomalies experienced ern Annular Mode (SAM) (see Section 5.3). Other high-pressure in the Gulf of Carpentaria and off the coast of New South Wales centres associated with the subtropical ridge were located over (NSW). Flooding associated with the remnants of cyclone South Africa (1020.5 hPa) and off the west coast of Chile Debbie end of March 2017 at Tweed River and Brunswick River (1022.2 hPa). The largest significant low pressure was located 8 (Maddox 2017) were exacerbated by high tides and anoma- at the circumpolar trough off Antarctica between 180 and 270 E, lously high sea levels along the east coast. with anomalous pressures of 12.5 hPa. Across the Australian continent, anomalies were generally weak and close to neutral. 5 Atmospheric patterns 5.2 Mid-tropospheric analysis 5.1 Surface analysis The 500-hPa geopotential height, an indicator of the steering of The MSLP pattern for autumn 2017 is shown in Fig. 16, com- surface synoptic systems across the southern hemisphere, is puted using data from the 0000 UTC daily analyses of the shown for autumn 2017 in Fig. 18, with associated anomalies Bureau of Meteorology’s Australian Communicate Climate and shown in Fig. 19. The anomaly patterns show average geopo- Earth System Simulator (ACCESS) model (Puri et al. 2013). tential height across mainland Australia, with positive 318 Journal of Southern Hemisphere Earth Systems Science G. A. Smith

Pacific Ocean eq Anomaly Δ = 0.5°C

February 17

March 17

April 17 0 m 4 50 m 3 100 m 2 150 m 1 200 m 0 250 m –1 300 m –2 350 m May –3 17 –4 400 m 100°E 120°E 140°E 160°E 180°E 160°W 140°W 120°W 100°W 80°W 60°W Analysis done May 29 22:25

Fig. 13. Four-month sequence from March to May 2017 of vertical sea sub-surface temperature anomalies at the equator for the Pacific Ocean. Red shading indicates positive (warm) anomalies and blue shading indicates negative (cool) anomalies. The contour interval is 0.58C, from http://www.bom.gov.au/oceanog- raphy/oceantemp/pastanal.shtml, accessed 29 April 2020.

anomalies across Tasmania extending to a high of þ69 gpm accessed 29 April 2020). SAM was mostly positive from 2015 to approximately over the Macquarie Island. These positive mid-2016 and was negative for six consecutive months from heights across south of Australia (along with positive MSLP October 2016 to March 2017. During April and May 2017, SAM anomalies) reduced the strength of prevailing westerly winds returned to the positive phase, which has been shown to reduce across southern Australia, contributing to a drier southwest rainfall in the far southwest corner of Western Australia in Western Australia and Tasmania during autumn. The most autumn (Hendon et al. 2007). significant anomalies in the southern hemisphere are the low of 121 gpm that coincides with the location of the lowest MSLP 6 Winds þ anomaly in Fig. 17, and a high of 103 gpm over the Antarctic Low-level (850 hPa) and upper level (200 hPa) wind anomalies Peninsula. are shown in Figs 21 and 22 respectively. Isotach contours are at 5ms1 intervals. In the southern central Pacific, there was an 5.3 Southern Annular Mode anomalous feature that exists in both low (high of 6.6 m s1) and The SAM is a mode of variability that is characterised by the upper level winds (high of 12.8 m s1) that is associated with the north-south oscillation of westerly winds that encircles Ant- very strong low-pressure system off the coast of Antarctica arctica. An index of SAM is the zonal pressure difference (shown in MSLP anomaly Fig. 17). In Australia, the anomalies between the latitudes 40 and 658S(Marshall 2003). During a shown in the low-level are enhanced easterlies across the north positive SAM phase, westerly winds contract closer to Antarc- and weakened westerlies from the Tasman Sea to the Southern tica, and a negative phase moves westerly winds towards the Ocean going over Tasmania. In the upper level, mainland equator. Figure 20 shows the SAM index as described in Australia was dominated by an anti-cyclonic system that Marshall (2003) from 2015 to the end of autumn 2017 (data enhanced the subtropical jet in the southern half and weakened it obtained from https://legacy.bas.ac.uk/met/gjma/sam.html, in the northern half. BoM climate summary: autumn 2017 Journal of Southern Hemisphere Earth Systems Science 319

20°C isotherm depth anomalies (m) 60°N 300 200 30°N J 100 F

0° 0 mm M –100 30°S A –200 M –300 J 0° 45°E90°E135°E180° 135°W 90°W 45°W J A Fig. 15. Sea-level anomalies are a combination of data from TOPEX/ S Poseidon, Jason-1, Jason-2 and Jason-3 satellites. Inverse barometer effect O not removed; seasonal signal not removed; global trend not removed; glacial N isostatic adjustment removed. Data obtained from CSIRO Sea Level, Waves D J and Coastal Extremes: http://www.cmar.csiro.au/sealevel/sl_data_cmar. F html, accessed 29 April 2020. M A M rain in Queensland on 18 and 19 May also contributed to high J seasonal totals between the Gulf Country and the tropical and J central coasts, with some locations setting monthly rainfall A records for May in part due to this event. S O Rainfall deciles indicate where rainfall amounts rank in N terms of the 1900 to 2017 record, which are allotted into 10 D percentile bins (Fig. 24). Above-average rainfall was also J F observed in western Victoria and adjacent parts of NSW border M country and southeastern South Australia; an area of central

2017A 2016 2015 South Australia; the coastal Top End in the Northern Territory; M and areas of Western Australia’s between the southern Kimber- 140°E 160°E 180° 160°W 140°W 120°W 100°W ley and southern interior, and along a broad strip extending from the eastern Pilbara coast to the western Eucla coast. –40 0 40 Rainfall for the season was below average for much of western and southwestern Queensland, extending across the Fig. 14. Time–longitude section of the monthly anomalous depth of the southern half of the Northern Territory; on each of the Yorke, 208C isotherm at the equator (28Sto28N) for March to May 2017. From the Eyre and Fleurieu peninsula in South Australia; across the west TAO Project Office: http://www.pmel.noaa.gov/tao/jsdisplay, accessed 29 of Western Australia from the far western Pilbara to the south April 2020. coast around Esperance. In Western Australia the South West Land Division recorded its 11th-driest autumn on record, and driest autumn since 2012, and was the driest autumn for at least 7 Australian region two decades at many locations. In Tasmania, western and 7.1 Rainfall southern regions were below average. A rainfall summary for Australia nationwide and each state and 7.2 Drought territory is listed in Table 1. Rainfall was slightly below average in autumn 2017 over Australia but varied regionally across the Rainfall deficiencies are used to indicate areas of drought and country significantly (Fig. 23). Victoria, NSW and Western are assessed by determining regions in the lowest 10% of records Australia were slightly above the mean, Queensland was slightly (serious deficiency), lowest 5% (severe deficiency) and lowest below the mean, and the remainder of the states and territories on record. Deficiencies for autumn 2017 are shown in Fig. 25. were well below the mean by an order of 20–30%. Central The most prominent region of rainfall deficiency was along the Australia experienced very much below-average rainfall, pri- west coast of Western Australia, spanning from Gascoyne, marily in the southern Northern Territory, as did in smaller through the Mid-West and into the Wheatbelt. Small regions regions in the far west of Queensland. of lowest on record were experienced near the coast from The coastline of Queensland had above-average rainfall from Carnarvon to Shark Bay. The Northern Territory also had a approximately Townsville all the way south along the Queens- region of serious-to-severe deficiency south of Tennant Creek land and NSW coast to the Victorian border. Western Victoria reaching across the borders into Queensland and South was mostly above average. Southern Western Australia had very Australia. Serious rainfall deficiencies also emerged in pockets much below-average rainfall in the west; however, the remain- of the southern South West Land Division in Western Australia der of the state was either average or above average. Severe and on the Eyre Peninsula in South Australia. In Tasmania, tropical cyclone Debbie brought heavy rainfall and floods to rainfall deficiencies covered the western highlands region. parts of eastern Queensland and northeastern NSW at the end of March. The two highest daily totals for Queensland and NSW 7.3 Temperature shown in Table 1 (Mount Jukes and Chillingham respectively) A subset of the full temperature network was used to calculate for autumn 2017 were associated with cyclone Debbie. Heavy the spatial averages and rankings for Australia, and all states 320 Journal of Southern Hemisphere Earth Systems Science G. A. Smith

ACCESS MSLP SUM 1 March to 31 May 2017

1035 1030 1025 1020 1015 1010 1005 1000 995 990 985 980 975 970 965

Fig. 16. Autumn 2017 mean sea-level pressure (MSLP) (hPa). The contour interval is 5 hPa. The MSLP anomaly field is not shown over areas of elevated topography (grey shading) as the extrapolation to sea level can result in unrealistic structures.

NCEP Reanal. MSLP Anomaly (hPa) 1 March to 31 May 2017

20.0 17.5 15.0 12.5 10.0 7.5 5.0 2.5 0.0 –2.5 –5.0 –7.5 –10.0 –12.5 –15.0 –17.5 –20.0

Fig. 17. Autumn 2017 mean sea-level pressure (MSLP) anomalies (hPa), from a 1979 to 2000 climatology. BoM climate summary: autumn 2017 Journal of Southern Hemisphere Earth Systems Science 321

Z0500 2.5×2.5 ACCESS OP. ANAL. (M) 20170301 0000 20170531 0000

6200 6100 6000 5900 5800 5700 5600 5500 5400 5300 5200 5100 5000 4900 4800

Fig. 18. Autumn 2017: 500-hPa mean geopotential height (gpm).

Z0500 2.5×2.5 ACCESS OP. ANAL.-NCEP2 (M) 20170301 0000 20170531 0000

240 210 180 150 120 90 60 30 0 –30 –60 –90 –120 –150 –180 –210 –240

Fig. 19. Autumn 2017: 500-hPa mean geopotential height anomalies (gpm), from a 1979 to 2000 climatology. 322 Journal of Southern Hemisphere Earth Systems Science G. A. Smith and territories shown in Table 2 (maximum temperature) and acorn-sat/ for details, accessed 29 April 2020) and extends Table 3 (minimum temperature). This data set is known as from 1910 to present. The mean maximum temperature ACORN-SAT (see http://www.bom.gov.au/climate/change/ for autumn 2017 was þ1.218C above average for Australia, ranking seventh warmest out of 108 autumn periods on record. All states except NSW and Western Australia ranked in the 6 5 top 10 autumns for maximum temperature. Regions in western 4 Queensland and into the Northern Territory were highest 3 on record. National mean minimum temperatures were 2 þ0.428C above average, although only Queensland, Victoria 1 andTasmaniawerewithinthetop10autumnsforminimum 0 temperature. –1 Maximum temperatures for autumn (Figs 26 and 27)were –2 above to very much above average for most of the Northern –3 Territory, along the eastern border of Western Australia south –4 2015 2016 2017 of the Kimberley, across all of South Australia, most of Queensland, inland NSW, all of Victoria and very much Fig. 20. Autumn 2017: 500-hPa mean geopotential height anomalies above average for all of Tasmania, and an area of Western (gpm), from a 1979 to 2000 climatology. Australia between the western Pilbara and Central Wheatbelt.

|V| access0850windsano 1 March to 31 May 2017

25 20 15 10 5

Fig. 21. Autumn 2017: 850-hPa vector wind anomalies (m s1).

|V| access0200windsano 1 March to 31 May 2017

25 20 15 10 5

Fig. 22. Autumn 2017: 200-hPa vector wind anomalies (m s1). BoM climate summary: autumn 2017 Journal of Southern Hemisphere Earth Systems Science 323

Table 1. Summary of the seasonal rainfall ranks and extremes on a national and state basis for autumn 2017 The ranking in the second last column goes from 1 (lowest) to 118 (highest) and is calculated over the years 1900–2017 inclusive. Qld, Queensland

Region Highest seasonal Lowest seasonal total Highest daily total (mm) Area-averaged Rank of area- Difference from total (mm) (mm) total (mm) averaged rainfall mean (%)

Australia 1891.0 at Mount Zero at several locations 635.0 at Mount Jukes (Qld) 113.3 69 6.0% Below mean Jukes (Qld) on 30 March (¼120.50) Queensland 1891.0 at Mount Zero at several locations 635.0 at Mount Jukes on 156.5 73 4.1% Below mean Jukes 30 March (¼163.14) New South 1383.5 at Zero at Glenbrook 507.0 at Chillingham on 154.8 93 8.4% above mean Wales Chillingham Bowling Club 31 March (¼142.74) Victoria 451.2 at Wyelangta 55.8 at Avoca 103.2 at Woomelang on 164.3 81 4.8% Above mean (Homebush) 21 April (¼156.81) Tasmania 640.8 at Mount Read 59.6 at Bridgewater 147.6 at Cornwall on 21 May 262 21 23.0% Below mean (Treatment Plant) (¼340.42) South 229.0 at Lucindale 3.6 at Mount Dare 98.8 at Lucindale Post Office 42.95 58 23.7% Below mean Australia Post Office on 21 March (¼56.32) Western 393.4 at Port Zero at several locations 193.0 at Yeeda on 13 March 94.7 66 5.0% Above mean Australia Hedland Airport (¼90.23) Northern 1015.7 at Lake Zero at several locations 384.0 At Point Fawcett on 102.5 49 26.5% Below mean Territory Evella 5 March (¼139.34)

Rainfall (mm)

1200 800 600 400 300 200 100 50 25 10 2 0 Australian rainfall analysis (mm) 1 March to 31 May 2017 Australian Bureau of Meteorology

http://www.bom.gov.au © Commonwealth of Australia 2018, Australian Bureau of Meteorology ID code: Analyser Issued: 30/05/2018

Fig. 23. Autumn 2017 rainfall totals (mm) for Australia.

There was a region of warmest on record for autumn in although they were cooler than average for an area spanning the western Queensland, reaching across the border into the southwestern Northern Territory and the adjacent eastern inte- Northern Territory. rior and southeastern Kimberley in Western Australia. There Mean minimum temperatures (Figs 28 and 29) were also was a small region of very much below-average minimum above to very much above average for most of eastern Australia, temperatures on the Western Australia coast west of the South the east of South Australia, and the northern half and far Australia border, and additional pockets of below-average southeast of the Northern Territory. Minimum temperatures minimum temperatures on northeast of the Wheatbelt and near were mostly near average for Western Australia during autumn, Karratha. 324 Journal of Southern Hemisphere Earth Systems Science G. A. Smith

Rainfall decile ranges

Highest on record 10 Very much above average 8–9 Above average

4–7 Average

2–3 Below average

1 Very much below average Lowest on record

Australian rainfall deciles 1 March to 31 May 2017 Distribution based on gridded data Australian Bureau of Meteorology http://www.bom.gov.au © Commonwealth of Australia 2018, Australian Bureau of Meteorology ID code: AWAP Issued: 10/05/2018

Fig. 24. Autumn 2017 rainfall deciles for Australia: decile ranges based on grid-point values over all autumns from 1900 to 2017.

Rainfall percentile ranking Serious deficiency Severe deficiency Lowest on record

Rainfall percentile (all avail. data) 1 March to 31 May 2017 Distribution based on gridded data Australian Bureau of Meteorology http://www.bom.gov.au © Commonwealth of Australia 2018, Australian Bureau of Meteorology ID code: Analyser Issued: 13/08/2018

Fig. 25. Rainfall deficiencies for autumn 2017.

8 Southern hemisphere ocean in the National Oceanic and Atmospheric Administration Temperatures in the southern hemisphere were the warmest on (NOAA) data set (Smith and Reynolds 2005), and third warmest record for land and ocean in autumn 2017 according to the in the United Kingdom Meteorological Office Hadley Centre/ National Aeronautics and Space Administration data set Climatic Research Unit, University of East Anglia (HadCRU) (Hansen et al. 2010), second warmest on record for land and data set (Morice et al. 2012). The differences in ranks between BoM climate summary: autumn 2017 Journal of Southern Hemisphere Earth Systems Science 325

Table 2. Summary of the seasonal maximum temperature ranks and extremes on a national and state basis for autumn 2017 The ranking in the last column goes from 1 (lowest) to 108 (highest) and is calculated over the years 1910–2017 inclusive. As the anomaly averages in the tables are only retained to two decimal places, tied rankings are possible. AWS, automatic weather station; NSW, New South Wales; Qld, Queensland; Vic., Victoria; WA, Western Australia

Region Highest seasonal Lowest seasonal mean Highest daily maximum tem- Lowest daily maximum Area-averaged Rank of mean maximum (8C) maximum (8C) perature (8C) temperature (8C) temperature area-averaged anomaly (8C) temperature anomaly

Australia 35.9 At Century 9.5 At Mount Hotham 44.6 At Mandora (WA) on 3.7 At Thredbo AWS 1.21 102 Mine (Qld) (Vic.) 5 March (NSW) on 31 May Queensland 35.9 At Century 21.2 At Applethorpe 44.0 At Urandangi Aerodrome 13.5 At Applethorpe on 1.54 102 Mine on 28 March 27 April New South 29.5 At Mungindi 9.5 At Thredbo AWS 44.0 At Wilcannia Aerodrome 3.7 At Thredbo AWS on 1.01 97 Wales Post Office AWS on 27 March 31 May Victoria 25.4 At Mildura 9.5 At Mount Hotham 38.0 At Walpeup Research on 3.3 At Mount Hotham 1.21 101 Airport 2 March on 31 May Tasmania 20.2 At Launceston 9.8 At Kunanyi (Mount 33.2 At Bushy Park (Bushy 0.6 At Kunanyi (Mount 0.95 101 (Ti Tree Bend) Wellington Pinnacle) Park Estates) on 14 March Wellington Pinnacle) on 30 May South 30.2 At Moomba 18.4 At Mount Lofty 44.5 At Ceduna Amo on 8.5 At Mount Lofty on 1.46 102 Australia Airport 26 March 28 May Western 35.7 At Mandora 20.3 At Shannon 44.6 At Mandora on 5 March 11.5 At Manjimup on 0.73 84 Australia 21 May Northern 35.1 At Mcarthur 30.3 At Yulara Airport 42.0 At Curtin Springs on 18.2 At Arltunga on 1.66 102 Territory River Mine Airport 26 March 31 May

Table 3. Summary of the seasonal minimum temperature ranks and extremes on a national and state basis for autumn 2017 The ranking in the last column goes from 1 (lowest) to 108 (highest) and is calculated over the years 1910–2017 inclusive. As the anomaly averages in the tables are only retained to two decimal places, tied rankings are possible. AWS, automatic weather station; NSW, New South Wales; SA, South Australia; WA, Western Australia

Region Highest seasonal mean Lowest seasonal Highest daily minimum Lowest daily minimum Area-averaged Rank of minimum (8C) mean minimum temperature (8C) temperature (8C) temperature area-averaged (8C) anomaly (8C) temperature anomaly

Australia 27.4 at Browse Island 1.3 Perisher Val- 30.2 at Oodnadatta Airport 8.5 at Thredbo AWS 0.42 88.5 (WA) ley AWS (NSW) (SA) on 27 March (NSW) on 3 May Queensland 25.9 at Coconut Island 10.3 at 29.8 at Boulia Airport on 2.7 at Warwick on 0.98 99 Applethorpe 29 March 31 May 29.8 at Windorah Airport on 25 March 29.8 Mornington Island Airport on 29 March New South 17.1 at South West Rocks 1.3 at Perisher 27.8 at Fowlers Gap AWS on 8.5 at Thredbo AWS on 0.87 95 Wales (Smoky Cape Valley AWS 27 March 3 May Lighthouse) Victoria 14.2 at Wilsons Promon- 3.3 at Mount 23.2 at Echuca Aerodrome on 6.3 at Mount Hotham on 0.95 99 tory Lighthouse Hotham 16 March 3 May Tasmania 13.4 at Hogan Island 2.5 at Liawenee 21.1 at Launceston (Ti Tree 4.6 at Liawenee on 4 May 0.72 100 Bend) on 16 March 4.6 at Liawenee on 21.1 at Flinders Island Airport 3 April on 16 March South 16.2 at Moomba Airport 8.8 at Keith 30.2 at Oodnadatta Airport on 2.8 at Yunta Airstrip on 0.36 77 Australia (Munkora) 27 March 31 May Western 27.4 at Browse Island 8.4 at Jarrahwood 30.0 at Legendre Island on 0.2 at Jarrahwood on May 59 Australia 7 March 1 May 30.0 at Barrow Island 0.2 at Norseman Aero on Airport on 7 March 30 May 0.2 at Salmon Gums Res. Stn. on 2 May Northern 26.0 at Mccluer Island 12.8 at Alice 30.0 at Centre Island on 2.0 at Arltunga on 11 May 0.3 80 Territory Springs Airport 30 March 326 Journal of Southern Hemisphere Earth Systems Science G. A. Smith

6°C 5°C 4°C 3°C 2°C 1°C 0°C –1°C –2°C –3°C –4°C –5°C –6°C

Maximum temperature anomaly (°C) 1 March to 31 May 2017 Australian Bureau of Meteorology

http://www.bom.gov.au © Commonwealth of Australia 2017, Australian Bureau of Meteorology ID code: AWAP Issued: 29/11/2017

Fig. 26. Autumn 2017 maximum temperature anomalies (8C) from analysis of ACORN-SAT data.

Temperature decile ranges

Highest on record 10 Very much above average 8–9 Above average

4–7 Average

2–3 Below average

1 Very much below average Lowest on record

Maximum temperature deciles 1 March to 31 May 2017 Distribution based on gridded data Australian Bureau of Meteorology http://www.bom.gov.au © Commonwealth of Australia 2017, Australian Bureau of Meteorology ID code: AWAP Issued: 29/11/2017

Fig. 27. Autumn 2017 maximum temperature deciles from analysis of ACORN-SAT data: decile ranges based on grid-point values for all autumns from 1910 to 2017. the three data sets reflect the different methods they use for Land temperatures in autumn 2017 were mostly warmer assessing temperatures over data-sparse land measurements and than average in the southern hemisphere, with the only signifi- are typically greater in the southern hemisphere than the cant exception being the southwest coast of Western Australia northern due to the lower data density. (Fig. 30). Both southern Africa and eastern Australia had BoM climate summary: autumn 2017 Journal of Southern Hemisphere Earth Systems Science 327

6°C 5°C 4°C 3°C 2°C 1°C 0°C –1°C –2°C –3°C –4°C –5°C –6°C

Minimum temperature anomaly (°C) 1 March to 31 May 2017 Australian Bureau of Meteorology

http://www.bom.gov.au © Commonwealth of Australia 2017, Australian Bureau of Meteorology ID code: AWAP Issued: 29/11/2017

Fig. 28. Autumn 2017 minimum temperature anomalies (8C) from analysis of ACORN-SAT data.

Temperature decile ranges

Highest on record 10 Very much above average 8–9 Above average

4–7 Average

2–3 Below average Very much 1 below average Lowest on record

Minimum temperature deciles 1 March to 31 May 2017 Distribution based on gridded data Australian Bureau of Meteorology http://www.bom.gov.au © Commonwealth of Australia 2017, Australian Bureau of Meteorology ID code: AWAP Issued: 29/11/2017

Fig. 29. Autumn 2017 minimum temperature deciles from analysis of ACORN-SAT data: decile ranges based on grid-point values for all autumns from 1910 to 2017. significant expanses of up to þ18C warmer than the 1981– average, with even higher temperatures in the Bay of Plenty 2010 average. South America was mostly up to þ0.58C andAuckland(NIWA 2017). warmer than average. New Zealand’s North Island experi- Seasonal precipitation percentages with reference to normal enced temperatures up to þ1.28C above the 1981–2010 from 1951 to 2000 are shown for autumn 2017 in Fig. 31. 328 Journal of Southern Hemisphere Earth Systems Science G. A. Smith

Land and ocean temperature departure from average March 2017–May 2017 (with respect to a 1918–2010 base period) Data Source: GHCN–M version 3.3.0 & ERSST version 4.0.0

–5–4 –3 –2 –1 0 1 2345 National Centers for Environmental Information Degrees celsius Please note: Grey areas represent missing data Wed Jun 14 07:31:54 EDT 2017 Map projection: Robinson

Fig. 30. Mean temperature anomalies (8C) from a 1981–2010 base period for March to May 2017 (source: National Centers for Environmental Information).

GPCC Monitoring product version 6 Gauge–Based Analysis 1.0 degree precipitation percentage of normals 1951/2000 for season (March, April and May) 2017 (grid based) 90°N

60°N

30°N

30°S

60°S

90°S 180° 120°W 60°W 0 60°E 120°E 180° (c) GPCC 2018/11/12

20 40 60 80 100 120 167 250 500

Fig. 31. Precipitation as a percentage of the 1951–2000 average for March to May 2017 (source: Global Precipitation Climatology Centre). BoM climate summary: autumn 2017 Journal of Southern Hemisphere Earth Systems Science 329

Table 4. Monthly NINO sea-surface temperature anomalies for autumn 2017, obtained from ftp://ftp.cpc.ncep.noaa.gov/ wd52dg/data/indices/sstoi.indices, accessed 29 April 2020

Month NINO1þ2 anomaly (8C) NINO3 anomaly (8C) NINO4 anomaly (8C) NINO3.4 anomaly (8C)

March þ1.89 þ0.57 0.06 þ0.13 April þ0.93 þ0.59 þ0.15 þ0.32 May þ0.78 þ0.51 þ0.29 þ0.46

18 ) 2 16

4 Sea ice extent (× 1Sea ice extent 000 000 km 2

0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Average +/–1 standard deviation 2016 2017

Fig. 32. Antarctic sea-ice extent. Average is calculated from 1981 to 2010. Data from National Snow and Ice Data Center (http://nsidc.org/data, accessed 29 April 2020).

Above-average seasonal precipitation was observed across the confirmed to be the case for the period leading up to and west coast of, and southern, South America, with severe flood- including the start of autumn 2017, when a negative SAM from ing reported in Argentina, Peru and Chile (Davies 2017b, 2017c; October 2016 to March 2017 led to the record low for March. World Meteorological Organization 2018). Flooding along the west coast of South America is typical in the late phase of Acknowledgements El Ninõ, although autumn 2017 was defined as ENSO neutral. The author would like to thank David Martin for providing valuable infor- SSTs off the Peruvian coast were þ1.898C warmer than average mation regarding the generation of OLR and MJO products. The author also in the NINO1þ2 region (Table 4), which is indicative of a thanks Blair Trewin and Matthew Wheeler for reviewing drafts of this ‘coastal El Ninõ’ (Takahashi and Martıńez 2019). Drier than manuscript and providing helpful comments. This research did not receive average conditions were notable across northeastern South any specific funding. America and central and Western Australia. New Zealand’s North Island experienced more than 150% of the normal autumn References rain, with several locations recording their wettest autumn on Bureau of Meteorology, and CSIRO (2017). State of the Climate 2016. record (NIWA 2017). (Bureau of Meteorology and CSIRO: Melbourne, Vic., Australia.) Davies, R. (2017a). ‘Indonesia – Homes Destroyed, 7 Dead after Floods and 8.1 Antarctic climate Landslides in Sulawesi, Borneo and Sumatra’. Available at http://flood- Sea-ice extent in the Antarctic was third lowest on record since list.com/asia/indonesia-floods-sulawesi-borneo-sumatra-may-2017 1978 for the autumn 2017 period. The sea-ice extent (Fig. 32)in [Verified 29 April 2020]. March 2017 was lowest on record for that month. The signifi- Davies, R. (2017b). ‘Chile – Deaths and Evacuations After Floods in cantly low sea-ice extent follows near average conditions in Atacama and Coquimbo’. Available at http://floodlist.com/america/ chile-floods-atacama-coquimbo-may-2017 [Verified 29 April 2020]. autumn 2016, and a period of record high sea-ice extent in the Davies, R. (2017c). ‘Peru – Thousands Affected by Floods and Landslides 4 years from 2011 to 2015, with record highs ending in spring Across the Country’. Available at http://floodlist.com/america/peru- 2015 (Martin 2016). The SAM index is a known commodity in lima-floods-mudslides-march-2017 [Verified 29 April 2020]. determining Antarctic sea-ice extent, with negative SAM phases Donald, A., Meinke, H., Power, B., Wheeler, M., and Ribbe, J. (2004). linked to lower sea-ice extents (Hall and Visbeck 2002). This is ‘Forecasting with the Madden–Julian Oscillation and the applications 330 Journal of Southern Hemisphere Earth Systems Science G. A. Smith

for risk management’. 4th International Crop Science Congress, Madden, R. A., and Julian, P. R. (1994). Observations of the 40–50-day Brisbane. tropical oscillation – a review. Mon. Wea. Rev. 122(5), 814–837. doi:10. Great Barrier Reef Marine Park Authority (2017). Final report: 2016 coral 1175/1520-0493(1994)122,0814:OOTDTO.2.0.CO;2 bleaching event on the Great Barrier Reef. (Great Barrier Reef Marine Maddox, S. (2017). MHL2574: NSW Ocean and River Entrance Tidal Park Authority: Townsville.) Levels Annual Summary 2016–2017, Manly Vale. Hall, A., and Visbeck, M. (2002). Synchronous variability in the Southern Marshall, G. J. (2003). Trends in the Southern Annular Mode from Hemisphere atmosphere, sea ice, and ocean resulting from the observations and reanalyses. J. Clim. 16(24), 4134–4143. doi:10.1175/ annular mode. J. Clim. 15(21), 3043–3057. doi:10.1175/1520- 1520-0442(2003)016,4134:TITSAM.2.0.CO;2 0442(2002)015,3043:SVITSH.2.0.CO;2 Martin, D. (2016). Seasonal climate summary southern hemisphere (spring Hansen, J., Ruedy, R., Sato, M., and Lo, K. (2010). Global surface 2015): El Ninõ nears its peak. J. South. Hemisph. Earth. Syst. Sci. 66(3), temperature change. Rev. Geophys. 48, RG4004. doi:10.1029/ 228–261. 2010RG000345 Morice, C. P., Kennedy, J. J., Rayner, N. A., and Jones, P. D. (2012). Hendon, H. H., Thompson, D. W. J., and Wheeler, M. C. (2007). Australian Quantifying uncertainties in global and regional temperature change rainfall and surface temperature variations associated with the Southern using an ensemble of observational estimates: the HadCRUT4 data set. J. Hemisphere annular mode. J. Clim. 20(11), 2452–2467. doi:10.1175/ Geophys. Res. Atmos. 117, D08101. doi:10.1029/2011JD017187 JCLI4134.1 National Oceanic and Atmospheric Administration (2017). State of the Hendon, H. H., Zhang, C., and Glick, J. D. (1999). Interannual variation of climate: global climate report for annual 2016. Bull. Am. Meteorol. Soc. the Madden-Julian Oscillation during austral summer. J. Clim. 12(8) 98(8), 1–298. PART 2, 2538–2550. doi:10.1175/1520-0442(1999)012,2538: NIWA (2017). Seasonal Climate Summary: Autumn 2017. (NIWA: Auckland.) IVOTMJ.2.0.CO;2 Puri, K., Dietachmayer, G., Steinle, P., Dix, M., Rikus, L., Logan, L., Huang, B., Thorne, P. W., Banzon, V. F., Boyer, T., Chepurin, G., Law- Naughton, M., Tingwell, C., Xiao, Y., Barras, V., Bermous, I., Bowen, rimore, J. H., Menne, M. J., Smith, T. M., Vose, R. S., and Zhang, H. M. R., Deschamps, L., Franklin, C., Fraser, J., Glowacki, T., Harris, B., Lee, (2017). Extended reconstructed Sea surface temperature, Version 5 J., Le, T., Roff, G., Sulaiman, A., Sims, H., Sun, X., Sun, Zhu, H., (ERSSTv5): upgrades, validations, and intercomparisons. J. Clim. Chattopadhyay, M., and Engel, C. (2013). Implementation of the initial 30(20), 8179–8205. doi:10.1175/JCLI-D-16-0836.1 ACCESS numerical weather prediction system. Aust. Meteorol. Ocea- Hughes, T. P., and Kerry, J. T. (2017). ‘Back-to-back bleaching has now hit nogr. J. 63(2), 265–284. doi:10.22499/2.6302.001 two-thirds of the Great Barrier Reef’, Conversat. Available at https:// Reynolds, R. W., Rayner, N. A., Smith, T. M., Stokes, D. C., and Wang, W. theconversation.com/back-to-back-bleaching-has-now-hit-two-thirds- (2002). An improved in situ and satellite SST analysis for climate. J. of-the-great-barrier-reef-76092 [Verified 29 April 2020]. Clim. 15(13), 1609–1625. doi:10.1175/1520-0442(2002)015,1609: Hughes, T. P., Kerry, J. T., A´ lvarez-Noriega, M., A´ lvarez-Romero, J. G., AIISAS.2.0.CO;2 Anderson, K. D., Baird, A. H., Babcock, R. C., Beger, M., Bellwood, Risbey, J. S., Pook, M. J., McIntosh, P. C., Wheeler, M. C., and Hendon, H. H. D. R., Berkelmans, R., Bridge, T. C., Butler, I. R., Byrne, M., Cantin, (2009). On the remote drivers of rainfall variability in Australia. Mon. N. E., Comeau, S., Connolly, S. R., Cumming, G. S., Dalton, S. J., Diaz- Wea. Rev. 137(10), 3233–3253. doi:10.1175/2009MWR2861.1 Pulido, G., Eakin, C. M., Figueira, W. F., Gilmour, J. P., Harrison, H. B., Saji, N. H., Goswami, B. N., Vinayachandran, P. N., and Yamagata, T. Heron, S. F., Hoey, A. S., Hobbs, J. A., Hoogenboom, M. O., Kennedy, (1999). A dipole mode in the tropical Indian ocean. Nature 401(6751), E. V., Kuo, C., Lough, J. M., Lowe, R. J., Liu, G., McCulloch, M. T., 360–363. doi:10.1038/43854 Malcolm, H. A., McWilliam, M. J., Pandolfi, J. M., Pears, R. J., Smith, T. M., and Reynolds, R. W. (2005). A global merged land-air-sea Pratchett, M. S., Schoepf, V., Simpson, T., Skirving, W. J., Sommer, surface temperature reconstruction based on historical observations B., Torda, G., Wachenfeld, D. R., Willis, B. L., and Wilson, S. K. (2017). (1880-1997). J. Clim. 18(12), 2021–2036. doi:10.1175/JCLI3362.1 Global warming and recurrent mass bleaching of corals. Nature Takahashi, K., and Martıńez, A. G. (2019). The very strong coastal El Ninõ 543(7645), 373–377. doi:10.1038/NATURE21707 in 1925 in the far-eastern Pacific. Clim. Dyn. 52, 7389–7415. doi:10. IPCC (2013). ‘Climate Change 2013 – The Physical Science Basis. Contri- 1007/S00382-017-3702-1 bution of Working Group I to the Fifth Assessment Report of the Troup, A. J. (1965). The Southern Oscillation. Q. J. R. Meteorol. Soc. Intergovernmental Panel on Climate Change’. (Eds T. F. Stocker, 91(390), 490–506. doi:10.1002/QJ.49709139009 D. Qin, G.-K. Plattner, M. Tignor, S. K. Allen, J. Boschung, A. Nauels, UNOCHA ROAP (2017). ‘The Country Preparedness Package: Strengthen- Y. Xia, V. Bex and P. M. Midgley) (Cambridge University Press: United ing local humanitarian response in Marshall Islands and beyond’. Kingdom and New York.) Available at https://ocharoap.exposure.co/the-country-preparedness- Kanamitsu, M., Ebisuzak, W., Woollen, J., Yang, S.-K., Hnilo, J. J., Fiorino, package [Verified 29 April 2020]. M., and Potter, G. L. (2002). NCEP-DEO AMIP-II Reanalysis (R-2). Wheeler, M. C., and Hendon, H. H. (2004). An all-season real-time Bull. Am. Meteorol. Soc. 83(11), 1631–1643. doi:10.1175/BAMS-83- multivariate MJO index: development of an index for monitoring and 11-1631 prediction. Mon. Wea. Rev. 132(8), 1917–1932. doi:10.1175/1520- Kiribati Meteorological Service (2017). Kiribati Drought Update, Tarawa. 0493(2004)132,1917:AARMMI.2.0.CO;2 Available at https://reliefweb.int/sites/reliefweb.int/files/resources/ Wolter, K., and Timlin, M. S. (1993). Monitoring ENSO in COADS with a kdu0317.pdf [Verified 29 April 2020]. seasonally adjusted principal component index. Proc. 17th Clim. Diag- Kuleshov, Y., Qi, L., Fawcett, R., and Jones, D. (2009). Improving nostics Work, pp. 52–57. preparedness to natural hazards: Tropical cyclone prediction for the Wolter, K., and Timlin, M. S. (1998). Measuring the strength of ENSO Southern Hemisphere. Adv. Geosci. 12(Ocean Science), 127–143. events: How does 1997/98 rank? Weather 53(9), 315–324. doi:10.1002/ Madden, R. A., and Julian, P. R. (1971). Detection of a 40–50 day oscillation J.1477-8696.1998.TB06408.X in the Zonal Wind in the Tropical Pacific. J. Atmos. Sci. 28(7025)–708. World Meteorological Organization (2018). WMO Statement on the State of doi:10.1175/1520-0469(1971)028,0702:DOADOI.2.0.CO;2 the Global Climate in 2017. (World Meteorological Organization: Madden, R. A., and Julian, P. R. (1972). Description of global-scale circulation Geneva.) cells in the tropics with a 40–50 day period. J. Atmos. Sci. 29(6), 1109–1123. Wright, P. B. (1989). Homogenized long-period Southern Oscillation doi:10.1175/1520-0469(1972)029,1109:DOGSCC.2.0.CO;2 indices. Int. J. Climatol. 9(1), 33–54. doi:10.1002/JOC.3370090104