1. Introduction

November 2021 was Australia’s coolest November this century and the wettest in over 120 years, resulting in flooding in parts of eastern Australia (Figs. 1b,d) (Bureau of Meteorology 2022). One year prior, in November 2020, Australia experienced its warmest November on record and well below average rainfall across much of the continent (Figs. 1a,c). Two Novembers, at opposite ends of the climate scale, may not seem remarkable given Australia’s highly variable climate (Nicholls et al. 1997); of interest is that both Novembers occurred on a background of La Niña conditions, which have historically been associated with a wet and cool Australia, particularly during austral spring months (McBride and Nicholls 1983; Risbey et al. 2009b; Chung and Power 2017; Tozer et al. 2023). Indeed, Australia’s wettest spring on record occurred during a strong La Niña in 2010 (Evans and Boyer-Souchet 2012; Hendon et al. 2014).

View Full Size

Rainfall deciles for (a) November 2020 and (b) November 2021. Mean surface temperature deciles for (c) November 2020 and (d) November 2021. Data are from the AGCD dataset. Deciles are calculated relative to the 1900–2021 and 1910–2021 periods for rainfall and surface temperature, respectively. Australian state boundaries and names are indicated in (c).

Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0112.1

• Download Figure
• Download figure as PowerPoint slide

Differences in Australia’s climate outcomes during La Niña events are not unexpected given that other sources of variability, related or unrelated to El Niño–Southern Oscillation (ENSO), may obscure or enhance the impacts of an individual La Niña event (e.g., Gallant and Karoly 2009). For example, a co-occurring positive Indian Ocean dipole (IOD) event, which is typically associated with drying in Australia, has been linked to the dry conditions that occurred in southeast Australia during the 2007 La Niña event (Gallant and Karoly 2009; Cai et al. 2011). Conversely, Lim and Hendon (2017) suggest that a strong negative IOD acted to enhance the weak La Niña of 2016, which resulted in very wet conditions in Australia in the latter half of 2016.

Several studies have assessed the key drivers, beyond La Niña, of the very wet conditions across Australia during 2010/11. For spring 2010, Hendon et al. (2014) suggest that in addition to the La Niña, a strong positive Southern Annular Mode (SAM) was a major contributor to the extreme rainfall experienced. For December 2010, Evans and Boyer-Souchet (2012) found that the very warm sea surface temperatures (SSTs) around northern Australia and in particular, northeast of Australia, contributed to the high rainfall anomalies experienced in Queensland. Evans and Boyer-Souchet (2012) note that these SSTs were warmer than had been seen in previous La Niña events. Whelan and Frederiksen (2017) explored the atmospheric dynamics behind the extreme rainfall in January 2011. They implicate blocking in the Tasman Sea and the Madden–Julian oscillation (MJO), in particular the time spent in MJO phase 6, in the heavy rainfall.

Lim et al. (2021) also found that the MJO contributed to extreme rainfall in Australia, this time the large rainfall deficits that occurred during the La Niña in November 2020. During this period, the MJO was active, at high amplitudes, over the Indian Ocean (predominantly in phases 8, 1, and 2), which acted to suppress rainfall over northern and eastern Australia. Lim et al. (2021) also suggest that the SSTs north of Australia were not as warm during spring 2020 as one would expect during a La Niña, which also contributed to the rainfall deficits.

Undertaking retrospective analysis of extreme events or periods, like those studies described above, is important for honing our understanding of the processes that drive climate extremes in Australia and for managing our expectations around the likely outcomes when a La Niña event, for example, is predicted in the future (Tozer et al. 2023). As such, we seek to understand the drivers of the wet and cool conditions in November 2021 and dry and warm in November 2020. In particular, we ask the following: Why were the climate outcomes of the two Novembers so different and extreme given that they both occurred on a background of La Niña conditions?

Our answer to this question is based on a hierarchy of assessment using composite patterns, individual monthly means, and daily snapshots. These different approaches provide us with different information about the two Novembers. We begin by identifying, via composite analysis, the large-scale atmospheric and oceanic conditions during La Niña and El Niño Novembers. This establishes our expectation for what happens, on average, during ENSO events in November. Composites are useful as they can help identify consistent patterns across a particular class of events; however, they may “average out” and obscure important interevent variability.

We next focus on this interevent variability by assessing monthly mean patterns of key atmospheric and oceanic diagnostics for November 2020 and 2021. In doing so, we consider the state of ENSO, IOD, SAM, and MJO during each November. As alluded to above, these form a set of drivers that are commonly invoked to explain climate outcomes in Australia. We find that the large-scale drivers provide little insight into why the climate outcomes in November 2020 and 2021 were so different. Next, we assess the marked difference in jet signatures across the continent in the two Novembers. That is, we explore the activity of the polar and subtropical jets in each November. The jets act as channels (or “waveguides”) for Rossby waves, supporting the development and steering of short-wave features including low pressure troughs and cutoff lows, and are therefore linked to rainfall variability (e.g., Hoskins and Ambrizzi 1993; Lee and Kim 2003; Archer and Caldeira 2008; Risbey et al. 2009a). In particular, we posit a hypothesis that the Australian continental temperature can play a role in setting the meridional temperature gradient over Australia, which in turn can strengthen or lessen the thermal wind and the subtropical jet and subsequently contribute to an increase or reduction in rainfall over the continent. We describe how this mechanism contributed to the climate outcomes in November 2020 and 2021.

Atmospheric features like the jets exhibit strong variability from day to day, and thus, the daily time scale is critical. Accordingly, we also examine relevant atmospheric diagnostics on the daily (synoptic) time scale. This assessment adds important nuance to our hypothesis described above.

2. Data

b. Atmospheric data

Atmospheric data are from the Japanese 55-yr Reanalysis (JRA-55; Kobayashi et al. 2015) dataset. Data are available from 1958 to the present at a resolution of 1.25° × 1.25°. Anomalies are calculated relative to the 1958–2021 period. Variables assessed are geopotential height, surface temperature, and wind anomalies. We also calculate thickness anomalies between 500 and 1000 hPa, thermal wind magnitude, and Eady growth rate. Positive and negative thickness anomalies can be used to diagnose warm and cold air, respectively, to aid in inferring advection.

Eady growth rate is an indicator of baroclinic instability and is defined as (Lindzen and Farrell 1980; Hoskins and Valdes 1990) follows:

${\sigma }_{\text{BI}}=0.31\frac{f}{N}\left|\frac{\partial \mathbf{v}}{\partial z}\right|,$(1)

where f is the Coriolis parameter, N is the Brunt–Väisällä frequency (Lee and Mak 1994), z is the vertical height, and v is the horizontal wind vector. We calculate Eady growth rate as an average over 150–450 hPa.

Regarding jet stream identification, we first note that atmospheric jets tend to be described as either subtropical or polar jets, typically distinguished based on their latitudinal location and forcing mechanisms (Harnik et al. 2016). Subtropical jets sit at the edge of the Hadley cell, on average around 25°–35°S in the Southern Hemisphere, and form due to angular momentum transport by the Hadley circulation, which in turn is driven by the meridional temperature gradient across the warm tropics and cool subtropics (Held and Hou 1980; Bals-Elsholz et al. 2001; Lee and Kim 2003). As such, subtropical jets are often described as “thermally driven jets.” We also note that the subtropical jet core tends to be confined to the upper troposphere (Nakamura and Shimpo 2004). Polar jets, also called “eddy-driven jets,” primarily form due to eddy momentum and heat fluxes and are located in the Ferrel cell, around 50°–60°S on average (Harnik et al. 2016). The polar jet typically has a deep structure through the atmosphere and can be evident in both upper levels and near the surface (Nakamura and Shimpo 2004). Although these two jets are associated with different forcing mechanisms, in the real world, these mechanisms operate simultaneously and are not independent (Lee and Kim 2003; Li and Wettstein 2012; Harnik et al. 2016). For example, eddy forcing influences the Hadley circulation and therefore the subtropical jet (Walker and Schneider 2006; Kuroda 2017; Mitchell et al. 2019). Conversely, thermal gradients also influence the polar jet (Lee and Kim 2003; Harnik et al. 2016). Attributing an observed jet to a specific forcing mechanism can therefore be difficult (Li and Wettstein 2012; Kuroda 2017).

To first order, in the context of the two Novembers, we are primarily interested in identifying jets over and around the Australian continent. We use zonal and thermal wind diagnostics, which are relevant to both the subtropical and polar jets. The thermal wind expresses the balance between meridional temperature gradients and vertical wind shear. The thermal wind integrated through the troposphere is maximum in the jet stream (e.g., Holton and Hakim 2013). The thermal wind can thus be used as a marker of the locations of the jet streams and reflect the role of meridional temperature gradients in the development of the jet. The thermal wind magnitude between 300 and 700 hPa is defined as

${\mathbf{v}}_T=\frac{1}{f}\mathbf{k}\times {\nabla }_p({\mathrm{\Phi }}_{300\text{hPa}}-{\mathrm{\Phi }}_{700\text{hPa}}),$(2)

where f is the Coriolis parameter, k is the unit vector perpendicular to Earth’s surface, ∇_p is the gradient on a constant pressure surface, and Φ is the geopotential at 300 and 700 hPa.

Figure 2 shows the utility of the zonal and thermal wind diagnostics to identify the jets. Figure 2a presents a vertical cross section taken at 140°E (cutting through eastern Australia; see Fig. 1b) of the mean jets in November 2021. The zonal wind (green contours) reveals two distinct jet signatures—a jet signature confined to the upper troposphere, sitting across the Australian continent at around 30°S, and a jet signature around 55°S, which extends from the upper to lower troposphere. The link between the meridional temperature gradient from the surface to the jets above is also clear (gray shading, indicative of maximum temperature gradient). We identify these jet signatures as the “subtropical jet” (STJ) and “polar jet” (PJ), respectively, based on their latitudinal locations and vertical structure.

View Full Size

Jet identification (a) vertical cross section at 140°E from surface (1000 hPa) to upper atmosphere (100 hPa) for November 2021 showing monthly mean zonal wind (green contours) and temperature. Zonal wind values greater or equal to 10 m s⁻¹ are shown, with contours spaced at every 5 m s⁻¹. Gray shading indicates where the maximum meridional temperature gradient exists. The temperature gradient is calculated using centered differencing. Gradients greater than 0.7°C per degree latitude are shown. (b) For 12 Nov 2021, daily temperature anomalies at 850 hPa (colored shading), zonal wind at 300 hPa (green contours), geopotential height anomalies at 500 hPa (black contours), and thermal wind magnitude between 300 and 700 hPa (orange contour). Zonal wind values greater or equal to 20 m s⁻¹ are shown, with contours spaced at every 10 m s⁻¹. Thermal wind magnitude values of 3 m s⁻¹ are contoured. Solid black contours indicate positive geopotential height anomalies (high pressure), and dashed black contours indicate negative height anomalies (low pressure). Vertical black line indicates cross-section location (140°E). (c) Vertical cross section at 140°E from surface (1000 hPa) to upper atmosphere (100 hPa) for 12 Nov 2021 showing zonal wind (green contours), temperature, and meridional wind (gray contours). Zonal wind values greater or equal to 15 m s⁻¹ are shown, with contours spaced at every 10 m s⁻¹. Negative meridional wind values are shown in dashed dark gray contours, and positive values are shown in solid light gray contours. Negative meridional wind values indicate poleward flow, and positive values indicate equatorward flow. Solid black contour indicates meridional wind values equal to 0 m s⁻¹. Gray shading indicates where the maximum meridional temperature gradient exists. Gradients greater than 1°C per degree latitude are shown. In (a) and (c), thick dark green notches indicate northern and southern latitudes of the Australian continent. The subtropical jet and polar jet are labeled as STJ and PJ, respectively.

Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0112.1

• Download Figure
• Download figure as PowerPoint slide

To show that our jet diagnostics and associated jet identification are applicable to the daily time scale, particularly in relation to the subtropical jet, we present zonal wind at 300 hPa and thermal wind magnitude between 300 and 700 hPa for a selected day in November 2021 in Fig. 2b and a vertical cross section showing zonal wind, meridional wind, temperature anomalies, and associated maximum temperature gradient for the same day in Fig. 2c. In Fig. 2b, there is a strong jet signature (i.e., maxima in the zonal and thermal winds) over the Australian continent. The link between the thermal and zonal winds is clear (i.e., coinciding contours of zonal and thermal winds). Figure 2c demonstrates the link between the meridional temperature gradient over the continent and jets aloft. That is, the strongest temperature gradient at the surface occurs around 20°S (over northern Australia) and aloft around 28°S (gray shading), where there is a maximum in the jet signature, labeled as the subtropical jet. The negative meridional wind values at the latitude of this jet are indicative of poleward flow, i.e., the Hadley circulation, which provides further support for identifying it as the subtropical jet. Thus, in general, we identify jets across the Australian continent as the subtropical jet and jets south of the continent as the polar jet. We do acknowledge, however, the complexity in disentangling jet forcing mechanisms and therefore jet identification, particularly at the daily time scale.

c. Climate driver index data

To identify the ENSO events for the composite analysis (described in section 3a), we use the Niño-3.4 index available online (https://www.cpc.ncep.noaa.gov/data/indices/ersst5.nino.mth.91-20.ascii), with anomalies calculated relative to the 1991–2020 period, from the Extended Reconstructed SST (ERSST) dataset (Huang et al. 2017). We use data spanning 1958–2021, in line with the availability of the JRA data.

For our comparison of the climate conditions in November 2020 and 2021 (section 3b), we derive monthly indices for ENSO (represented by the Niño-3.4 index), IOD [represented by the dipole mode index (DMI)], ENSO Modoki [represented by the ENSO Modoki index (EMI)], and SST variability around northern Australia using the Optimum Interpolation Sea Surface Temperature (OISST) analysis (Reynolds et al. 2007). OISST data are available from 1982 to the present with anomalies calculated relative to 1982–2021. The Niño-3.4 index is calculated by averaging SST anomalies over 5°S–5°N and 190°–240°E (black box in Fig. 3b). The DMI is calculated as the difference between SST anomalies averaged over the tropical western Indian Ocean (10°S–10°N, 50°–70°E) and eastern Indian Ocean (10°S–0°, 90°–110°E) (Saji et al. 1999) (red boxes in Fig. 3b). The EMI is calculated as

$\text{Region}\hspace{0.17em}\mathrm{A}-0.5\times (\text{Region}\hspace{0.17em}\mathrm{B}+\text{Region}\hspace{0.17em}\mathrm{C}),$(3)

where the regions represent averaged SST anomalies in A (165°E–140°W, 10°S–10°N), B (110°–70°W, 15°S–5°N), and C (125°–145°E, 10°S–20°N), respectively (Ashok et al. 2007) (blue dashed boxes in Fig. 3b). SST anomalies around northern Australia (NthAUS) are calculated over 18°–8°S and 110°–160°E (green box in Fig. 3b). This region broadly encompasses the northwest, northern, and northeast Australia boxes identified by Evans and Boyer-Souchet (2012). For the SAM, we use the Antarctic oscillation index, which is derived using the first empirical orthogonal function of geopotential height at 700 hPa and is available online (https://www.cpc.ncep.noaa.gov/products/precip/CWlink/daily_ao_index/aao/aao.shtml). Data are available from 1979 to the present. For the MJO, we use the real-time multivariate (RMM) MJO index (Wheeler and Hendon 2004) provided by Australia’s Bureau of Meteorology (http://www.bom.gov.au/climate/mjo/). The index includes phase and amplitude components. Data are available from 1974 to the present. In addition, we produce a monthly thermal wind index. To do this, we average thermal wind over 110°–155°E and 30°–25°S (black dashed box in Fig. 3c). This location was selected to broadly capture where the maximum of thermal wind magnitude is over the Australian continent during winter months, when the jet is approximately at its maximum in the region during the year [see, e.g., Fig. 1 in Bals-Elsholz et al. (2001)]. We also extract Australia-averaged November rainfall and mean temperature anomalies from online (http://www.bom.gov.au/climate/change/), which are also derived from AGCD. Anomalies are calculated relative to 1900–2021 for rainfall and 1910–2021 for temperature.

View Full Size

Composites of large-scale climate fields during November in (a),(c),(e) La Niña and (b),(d),(f) El Niño events. Fields are (a),(b) surface temperature anomalies, (c),(d) zonal wind at 300 hPa and thermal wind magnitude between 300 and 700 hPa, and (e),(f) rainfall anomalies over Australia and geopotential height anomalies at 500 hPa. Solid red contours indicate positive geopotential height anomalies (high pressure), and dashed blue contours indicate negative height anomalies (low pressure). Geopotential height contours are spaced at 20-m intervals. Zonal wind values greater or equal to 15 m s⁻¹ are shown in green contours, with contours spaced at 5 m s⁻¹ intervals. La Niña Novembers are 1973, 1975, 1988, 1998, 1999, 2007, and 2010, and El Niño Novembers are 1965, 1972, 1982, 1997, 2002, 2009, and 2015 (calculated using the ERSST dataset). Rainfall data are from AGCD with rainfall anomalies calculated relative to 1900–2021. Data for all other variables are from the JRA dataset, with anomalies calculated relative to 1958–2021. Boxes used to calculate SST-based indices are indicated in (b) black box for Niño-3.4, red boxes for IOD, blue dashed boxes for EMI, and green box for NthAUS SSTs. The box used to calculate the thermal wind index is indicated in (c).

Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0112.1

• Download Figure
• Download figure as PowerPoint slide

3. Methods

4. Climate expectations in November during La Niña and El Niño events

We begin by setting the general climate expectations in November during La Niña and El Niño events (Fig. 3). In the La Niña composites, the signature cold tongue in the equatorial Pacific is evident as are the associated warm anomalies around northern Australia (Fig. 3a). The opposite is the case in the El Niño composites (Fig. 3b).

In the Pacific sector, particularly for El Niño, there is a wave train–like structure, reminiscent of the Pacific South American pattern, which other authors have identified when exploring average Southern Hemisphere extratropical conditions during ENSO events (e.g., Ding et al. 2012; Wang et al. 2022). More broadly, the geopotential height anomaly pattern in La Niña (Fig. 3e) is similar to a positive SAM pattern, which has a tendency to occur during La Niña (Fogt and Marshall 2020).

In regards to the subtropical jet, past research has identified an asymmetric response of the jet to ENSO events (Wang et al. 2022). In El Niño, the central and eastern equatorial Pacific warms, relative to the South Pacific, intensifying the meridional temperature gradient and strengthening the subtropical jet in the Pacific region (Rind et al. 2001; Seager et al. 2003; Gallego et al. 2005; Gillett et al. 2021; Black et al. 2022; Wang et al. 2022). The composite picture in Fig. 3d follows this description, with the core of the jet evident in the Pacific. We note that the subtropical jet is weak over Australia during El Niño. In La Niña, the colder tropical Pacific SSTs weaken the meridional temperature gradient in the region, and consequently, there is little to no subtropical jet in the composite in the Pacific region. Instead, in La Niña, the warm SSTs around northern Australia and cool continental temperatures (Fig. 3a) enhance a meridional temperature gradient in the Australian region, as opposed to the Pacific, and thus, there is a subtropical jet maximum across the Australian continent, indicated by the zonal and thermal wind signatures over Australia (Fig. 3c).

Across Australia, surface temperatures are generally cool and rainfall anomalies are positive in La Niña (Figs. 3a,e) and the opposite is the case in El Niño (Figs. 3b,f), which aligns with high pressure over the continent during El Niño (Fig. 3f) and low pressure during La Niña events (Fig. 3e).

We note that key SST patterns and jet features presented in the La Niña and El Niño composites in Fig. 3 are found to be significant according to the Monte Carlo test described in section 3 (Fig. A1).

5. The state of large-scale climate drivers during the two Novembers

We now explore conditions in November 2020 and 2021, starting with the large-scale climate drivers. Tropical SSTs are an important indicator for the Australian region (e.g., Evans and Boyer-Souchet 2012). SST anomalies for the two Novembers are presented in Fig. 4. Figure 5 presents values for relevant driver indices across September–December months, with the red dashed lines indicating one standard deviation, to highlight more notable index values.

View Full Size

SST anomalies and surface temperature anomalies across Australia in (a) November 2020 and (b) November 2021. SST data are from the OISST dataset and Australian temperature anomalies from AGCD. OISST anomalies are calculated relative to the 1982–2021 period, and AGCD temperature anomalies are calculated relative to 1910–2021. Anomalies greater than 3° or less than −3° are capped at ±3. This capping only affects anomalies across Australia and does not change the interpretation of the figures.

Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0112.1

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Variability of relevant climate indices across September to December 2020 and 2021. (a) Niño-3.4 index, (b) EMI, (c) DMI, (d) NthAUS, and (e) SAM. November is highlighted in blue. Red dashed lines indicate one standard deviation, calculated across all months for each index individually. Indices are calculated using the OISST dataset, with anomalies calculated relative to 1982–2021. Index box locations are given in Fig. 3b.

Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0112.1

• Download Figure
• Download figure as PowerPoint slide

d. MJO

Past studies have shown that large areas of Australia tend to be dry in spring when the MJO amplitude is greater than 1, during phases 1, 2, 3, and 8 and wet in phases 4, 5, and 6 (Wheeler and Hendon 2004; Marshall et al. 2021; Cowan et al. 2023). Here, we explore the degree to which different MJO phases in the past have conditioned rainfall in November. To do this, we plot past Australia-averaged daily rainfall in all Novembers from 1974 to 2021 as a function of MJO amplitude and phase (Fig. 6), highlighting the evolution of the MJO in both November 2020 and 2021. In November 2021, the MJO spent 17 days of the month, at amplitudes greater than 1, in phases 3, 4, and 5. In November 2020, the MJO spent 16 days with amplitudes greater than 1 in phases 8, 1, and 2. This means that approximately half of November 2020 and 2021 were spent in a weak MJO state (amplitude less than 1). Also of note is that relative to other days in November since 1974, the amplitude of the MJO during both November 2020 and 2021 was low. That is, the highest amplitude reached in either November 2020 or 2021 was 1.7, yet amplitudes in November have exceeded 3 in the past. More generally, Fig. 6 shows that high and low rainfall days in November can occur in any MJO phase and that high rainfall days in November can occur when the MJO amplitude is less than 1 (i.e., when the MJO is classed as weak; Wheeler et al. 2009). These points suggest that MJO amplitude and phase are not necessarily a strong conditioner of November rainfall outcomes. We note that similar conclusions can be drawn when only eastern Australia (as opposed to all Australia) rainfall is assessed (not shown).

View Full Size

Australia-averaged daily rainfall for each November day from 1974 to 2021, plotted with respect to the phase and amplitude of the RMM1 and RMM2 MJO values. MJO phases are indicated in each corner. Red circles indicate days in November 2020, and yellow circles indicate days in November 2021. The greater the size and the darker the shading of the circles, the higher the rainfall amount. Daily rainfall data are from AGCD. MJO index is available at http://www.bom.gov.au/climate/mjo/.

Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0112.1

• Download Figure
• Download figure as PowerPoint slide

In general, assessment of large-scale drivers has shown that the states of ENSO, IOD, and SAM were similar in the two Novembers and thus do not provide obvious clues as to why the rainfall outcomes were so different. We do not discount the potential role of MJO, acknowledging that Lim et al. (2021) identified a key role for MJO in the November 2020 rainfall deficits. Our results do highlight, though, that MJO variability does not appear to have strongly conditioned November rainfall totals in the past. These outcomes suggest that there is more to the story of why the climate outcomes for Australia were vastly different in November 2020 and 2021.

6. Monthly mean jet activity during the two Novembers

We now identify the mean jet activity in the two Novembers, with a particular focus on the subtropical jet, first by exploring the general relationship between the subtropical jet (represented by the thermal wind) and November rainfall across Australia. In Fig. 7, we present November thermal wind index values relative to November all Australia-averaged rainfall anomalies, with all Australia-averaged mean temperature anomalies shaded. The strong relationship between continental temperature and the thermal wind index in November is apparent in that the colder/warmer the continent the more positive/negative the thermal wind index. Similarly, the higher/lower the thermal wind index, the higher/lower the November rainfall. That is, positive thermal wind anomalies, indicative of a stronger subtropical jet over Australia, are associated with higher continental rainfall and cooler temperatures. On the opposite end of the scale, warm continental temperatures are associated with negative thermal wind anomalies (which indicate a weak jet relative to mean) and low rainfall. Of note is that November 2020 and 2021 sit at opposite “ends” of Fig. 7, with a relatively low thermal wind index value and associated low rainfall and high temperature in November 2020 and high thermal wind, high rainfall, and low temperature in November 2021. It should be noted that we undertook the same analysis as described above using zonal wind at 300 hPa instead of the thermal wind and drew similar conclusions; i.e., there is a strong relationship between rainfall and temperature across Australia and the upper-level zonal wind for wind values averaged across the box presented in Fig. 3c.

View Full Size

November thermal wind index anomalies vs Australia-averaged rainfall anomalies. Thermal wind index is calculated between 300 and 700 hPa for the box identified in Fig. 3c using the JRA dataset, with anomalies calculated relative to 1958–2021. Australia-averaged November rainfall totals and mean temperature anomalies are available from http://www.bom.gov.au/climate/change/ and are derived from AGCD. Rainfall anomalies were calculated relative to the 1900–2021 period and temperature anomalies relative to 1910–2021. Temperature anomalies have been linearly detrended. November 2020 and 2021 are highlighted.

Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0112.1

• Download Figure
• Download figure as PowerPoint slide

We next identify the jet signatures across the Southern Hemisphere in November 2020 and 2021 using zonal and thermal wind diagnostics. For context, we present the climatological location of the subtropical and polar jets in November (Fig. 8a). The subtropical jet is strongest over Australian longitudes in winter months and weakens in summer months. The gradient between the tropics and subtropics in the Australian region is strongest during winter months when the Australian continent cools, setting up a strong temperature gradient relative to the adjacent warmer oceans to the north (Taljaard 1972; Bals-Elsholz et al. 2001; Pook et al. 2013; Ummenhofer et al. 2013). This gradient is linked to an enhancement of the thermal wind (and subtropical jet) over Australia. As the continent warms in the summer months, the temperature gradient is reduced, which results in reduced thermal wind and fewer instances of the subtropical jet over the continent on average (Pook et al. 2013). November represents a transition period, as shown in Fig. 8a, where there is some thermal wind signature over Australia and a moderate continental subtropical jet on average. The main subtropical jet branch, and associated baroclinic instability (indicated by the Eady growth rate), sits downstream of the continent in November.

View Full Size

Thermal wind magnitude between 300 and 700 hPa (shading), zonal wind at 300 hPa (contours), and Eady growth rate (s⁻¹) averaged across 150 and 450 hPa (gray dots) for (a) November climatology, calculated across the 1958–2021 period, (b) November 2020, and (c) November 2021. Eady growth rate has been filtered to show values ≥ 0.8 × 10⁻¹⁰ s⁻¹. Zonal wind values greater or equal to 15 m s⁻¹ are shown in green contours with contours spaced at 5 m s⁻¹ intervals. Data are from the JRA dataset.

Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0112.1

• Download Figure
• Download figure as PowerPoint slide

Underlying the climatological average are periods when a subtropical jet signature is present over the Australian continent in November and periods when it is not. That is, there is year-to-year variability in jet locations (Gallego et al. 2005). This is clearly evident in the jet structures in the two Novembers. In November 2020, the Australian continent is largely devoid of jet activity and associated baroclinic instability. There is a jet branch, which steers south of the continent over southwest Western Australia. This indicates that any short waves (e.g., low pressure troughs) will be steered toward the southwest corner of the continent and then poleward, which fits with the pattern of rainfall deciles in November 2020 (Fig. 1a) where southwest Western Australia experienced wet conditions, but the rest of Australia was dry. In stark contrast, there is a clear thermal wind signature, and associated subtropical jet and baroclinic instability, over the continent in November 2021 (Fig. 8c). The jet signature in November 2021 draws similarities with the jet signature in the La Niña composites (Fig. 3).

Consideration of monthly mean diagnostics suggests that in the case of November 2020, the warm continental temperatures (Fig. 1c) reduced the temperature contrast with the warm SSTs north of Australia and displaced the region of strongest meridional temperature gradient southward to the boundary between the warm continent and the cooler southern ocean (Figs. 4a and 5d). This meant that there was a much-reduced temperature gradient and thus weak thermal wind over the continent, with the jet largely poleward of the continent through the month (Fig. 8b). In November 2021, the cold continental temperatures coupled with warm surface temperatures over far northern Australia (Fig. 1d) and warm SSTs around northern Australia (Figs. 4b and 5d) contributed to the establishment of a strong thermal gradient and associated enhanced thermal wind and subtropical jet across the continent (Fig. 8c). A subtropical jet over the continent would provide support for the development and steering of short waves across the continent, contributing to the Australia-wide wet conditions that occurred in 2021 (Fig. 1c).

A key question is why the continental temperatures set up as they did in November 2020 and 2021. We focus here on the role of horizontal advection in the continental temperature anomalies in the two Novembers. It is important to note though that temperature tendencies are a combination of horizontal advection and adiabatic and diabatic sources (e.g., Röthlisberger and Papritz 2023). Future work aims to tease out the role of adiabatic and diabatic sources in the temperature anomalies experienced in November 2020 and 2021. In Fig. 9, we present monthly mean thickness across 500 and 1000 hPa, wind speed vectors at 700 hPa, and geopotential height anomalies at 500 hPa during both Novembers. First, we note the trough around 250°E that appears in both periods, which is characteristic of the positive SAM phase (Risbey et al. 2021). Other similarities are the high pressure nodes around 180° and 330°E. These patterns in the Pacific broadly fit with the composite La Niña geopotential height patterns in Fig. 3. There are key differences in the thickness and geopotential height anomalies around Australian longitudes during the two Novembers, however. In November 2020 (Fig. 9a), the Australian continent is dominated by a ridge, which corresponds to the southern shift of the jet stream (Fig. 8b). The strong positive thickness anomalies over Australia during November 2020 indicate warm lower-troposphere temperatures, with warm air advection from the tropics across the continent (highlighted by the wind vectors), in line with the warm continental surface temperatures. The tilted trough over southwest Western Australia and ridge over southeast Australia in November 2020 represent the dominant long-wave pattern in the westerlies over the Australian sector (Pook et al. 2013). In November 2021 (Fig. 9b), there is a trough located over central Australia contiguous with a midlatitude trough south of Australia, coupled with blocking highs upstream and downstream of Australia. The negative thickness anomalies extending from the midlatitude trough across Australia, and the wind vectors, indicate cold air advection from higher latitudes onto Australia, cooling the continent and the atmosphere aloft.

View Full Size

Monthly thickness anomalies between 500 and 1000 hPa (shading), wind vectors at 700 hPa (green arrows), and geopotential height anomalies at 500 hPa (contours) for (a) November 2020 and (b) November 2021. Wind vectors are subsampled to show wind with magnitude greater than the 60th percentile. Only vectors at every third latitude and sixth longitude are shown to aid interpretation. Negative geopotential height anomalies (blue dashed contours) indicate troughs, and positive anomalies (red solid contours) indicate ridges in the midtroposphere. Contours are spaced every 25 m. Data are from the JRA dataset with anomalies calculated relative to the 1958–2021 period.

Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0112.1

• Download Figure
• Download figure as PowerPoint slide

7. Daily jet activity during the two Novembers and driving factors

We now move to daily scale diagnostics to further explore the synoptic dynamics that influenced the different climate outcomes of the two Novembers. In Figs. 10 and 11, we present daily near-surface temperature anomalies, geopotential height anomalies at 500 hPa, and zonal wind at 300 hPa for November 2020 and 2021, respectively. From early November 2021, two jet signatures are evident—a subtropical jet moving in from west of Australian longitudes and a polar jet linked to a midlatitude blocking high and trough, southwest of the continent. The subtropical jet appears to intensify as the high and trough approach the continent (e.g., around 6 November 2021). The gradient of temperature over the continent steepens due to the cold air advection from the high and trough (contrasting with the warm SSTs to the north) as they move into Australian longitudes. The contiguous cold temperature anomalies from the midlatitudes across the Australian continent are evident across more than half the month, in line with the presence of the high–trough–high pattern in the midlatitudes, south of Australia. The persistence of this pattern is further evident from the Hövmoller plots in Fig. 12. The plots present the daily midtropospheric flow variability, averaged across 40°–60°S latitudes (i.e., indicative of Southern Hemisphere storm track latitudes). For November 2021 (Fig. 12b), there is a quasi-stationary three-wave pattern across the hemisphere, persisting for approximately the first 20 days of the month. After this time, the flow becomes more progressive. This reflects that for at least the first 20 days of November 2021 a midlatitude quasi-stationary block was present in the Indian Ocean and associated trough over Australian longitudes, advecting cold air onto the continent, contributing to the cold continental temperature anomalies presented in Fig. 1 and thus the steep temperature gradient across the continent.

View Full Size

Daily zonal wind at 300 hPa (green contours), geopotential height anomalies at 500 hPa (red/blue contours), and 850-hPa temperature anomalies (colored shading) for November 2020. Solid red contours indicate positive height anomalies (high pressure) and solid blue contours indicate negative height anomalies (low pressure), with contours spaced every 50 m. Zonal wind values greater or equal to 20 m s⁻¹ are shown in green contours, with contours spaced at 10 m s⁻¹ intervals. Dates are indicated in the top-right corner of each subplot. Data are from JRA, with anomalies calculated relative to the 1958–2021 period.

Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0112.1

• Download Figure
• Download figure as PowerPoint slide

View Full Size

As in Fig. 10, but for November 2021.

Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0112.1

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Hovmöller plots of the midtropospheric flow averaged between 40° and 60°S latitude for (a) November 2020 and (b) November 2021. Both the shading and contours represent geopotential height anomalies at 500 hPa. The red shading and solid contours indicate positive height anomalies, and blue shading and dashed contours represent negative height anomalies. The green dashed line indicates the west and east longitudes of Australia. Contours are spaced every 50 m. Data are from the JRA dataset with anomalies calculated relative to the 1958–2021 period.

Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0112.1

• Download Figure
• Download figure as PowerPoint slide

Figure 11 indicates that there is a jet signature, albeit at varying strength, on most days across November 2021, providing support for the development and steering of short waves across the continent, contributing to the Australia-wide wet conditions (Fig. 1b). This is highlighted, for example, around 6 November 2021 where a trough moves into Australian longitudes to the west. In line with the jet signature, the trough, which eventually cuts off from the midlatitude storm track, moves across the continent until around 16 November. The rain associated with this event is discussed in section 8. The subtropical jet is characterized by regions of upper-level divergence and promotes the development of cutoff low systems and hence rainfall. Once these short-wave features have developed, they will advect cold air from higher latitudes to low latitudes (cyclonic flow) and therefore reinforce the existing temperature gradient and accompanying jets. Therefore, in this case they provide a positive feedback for their own sustenance and further development (Holton and Hakim 2013).

For November 2020, Australia is generally warm from the beginning of the month and the jet is largely poleward of the continent for the majority of the month (Fig. 10). Around 8 November 2020, a cutoff low pressure system moves toward Australia from the west. The system shifts across southwest Western Australia and then south of the continent, in line with the orientation of the jet signature over the following days. Similarly, another low pressure system moves over southwest Western Australia around 13 November. The system largely stays in that region over the next couple of days before shifting south of Australia. In November 2020, short-wave features embedded in the Southern Hemisphere storm track are relatively progressive across the month, particularly around Australian longitudes as indicated by the relatively rapid progression of troughs and ridges on the Hövmoller for 2020 (Fig. 12a). However, as noted above, the storm track is poleward of the continent here and does not play a strong role in setting up the warm anomaly in this year. The warm anomaly is set up by the high pressure ridge over the continent evident in Fig. 9a, promoting warm advection, clear skies, and enhanced incoming solar heating.

We described in section 6, for the monthly mean time scale, that the subtropical jet is associated with a strong thermal gradient and enhanced thermal wind. To demonstrate the link between meridional temperature gradients over the continent and jets aloft on the daily time scale, we present temperature profiles for selected days in each November for an arbitrarily selected longitude, 140°E. To illustrate the dry and wet periods in 2020 and 2021, we select periods in each November when the temperature gradient over the continent is likely to be weak (November 2020) and strong (November 2021) based on the signature (or lack thereof) of the subtropical jet over the selected longitude using Figs. 10 and 11 as reference. For November 2020, these dates are 18 to 22 November (Fig. 13a), and for November 2021, we select 11–15 November (Fig. 13b). For the selected days in November 2020, the largest temperature gradient sits across the southern portion of the continent spanning the boundary between the warm continent and cooler Southern Ocean. At the surface, this maximum gradient is around 35°S and aloft around 50°S (gray shading), where a jet is evident (green contours). This jet signature can be seen in Fig. 10 on the selected days. That is, there is a jet signature over the selected longitude, sitting south of the continent. On the selected November 2021 days, the largest temperature gradient occurs over the Australian continent, spanning the warm ocean to the north and the relatively colder continent. At the surface, this maximum gradient is around 20°S and aloft around 30°S, where a clear subtropical jet signature is evident (Fig. 13b). Importantly, Fig. 13 further highlights the connection between surface temperature gradients, temperature gradients aloft, and the jet stream.

View Full Size

Daily vertical (full field) temperature profiles across 0°–60°S latitudes and 1000–100-hPa atmospheric levels for (a) 18–22 Nov 2020 and (b) 11–15 Nov 2021. Solid green contours indicate 300-hPa zonal wind filtered to show values greater than or equal to 15 m s⁻¹, with contours spaced at every 10 m s⁻¹. Gray shading indicates where the maximum temperature gradient exists. The temperature gradient is calculated using centered differencing. Gradients greater than 1°C per degree latitude are shown.

Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0112.1

• Download Figure
• Download figure as PowerPoint slide

8. Key rainfall events in the two Novembers and their synoptic features

The results discussed above suggest that the absence, in the case of November 2020, and presence, in the case of November 2021, of the subtropical jet over the continent contributed to the rainfall anomalies experienced across Australia in both periods. Here, we provide further evidence of the role the subtropical jet played in the rainfall anomalies in both Novembers. The stark contrast between the amount of rain received on the Australian continent in November 2020 and 2021 is illustrated in Fig. 14a, which shows the cumulative sum of Australia-averaged daily rainfall for the two Novembers. November 2020 (red curve) accumulates rainfall slowly without marked major events and sits near the bottom of past November rainfall totals (gray shaded area). The wet November 2021 (blue curve) is the wettest November on record as indicated by the blue curve sitting near the top of the gray envelope of past November day totals and exceeding the envelope by the end of November. There were three major periods in November 2021 when the rate of accumulation of rainfall was particularly high, noting that rainfall may still have occurred over parts of the continent outside of these periods. These periods are marked in Figs. 14a and 14b, which presents the daily rainfall anomalies (relative to the mean) for each day in November 2020 and 2021.

View Full Size

Australia-averaged daily rainfall in the two Novembers. (a) Cumulative sum of Australia-averaged daily rainfall for November 2020 (red) and November 2021 (blue). Gray line shows the median November rainfall. Gray shading indicates the 1st and 99th percentile range of daily November rainfall. (b) Australia-averaged daily rainfall anomalies for November 2020 (red bars) and November 2021 (blue filled bars) calculated relative to the daily November mean. Key periods for November 2021 are indicated: 1) 1–4 Nov 2021, 2) 7–13 Nov 2021, and 3) 21–27 Nov 2021. Daily rainfall data are from AGCD. Statistics calculated relative to 1900–2021.

Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0112.1

• Download Figure
• Download figure as PowerPoint slide

To examine these wet periods more closely, we generate Hövmoller plots for each November where the latitudes are confined to 25°–35°S to focus on what happens over the main rainfall regions on the Australian continent. These Hövmollers are shown in Fig. 15 with near-surface (1000 hPa) geopotential height anomalies contoured and 500-hPa geopotential height anomalies shaded. The Australian continent is spanned by vertical green lines. In November 2020 (Fig. 15a), the continent is dominated by surface and upper ridges, resulting in mostly dry conditions. For November 2021, the three surges in continental rainfall marked in Fig. 14 can be traced to the sequences of events in the Hövmoller in Fig. 15b. In each case, there is a surface trough established over the continent (dashed blue lines in the Australian region). A midtropospheric trough then enters the continental region from the west (shaded blue). The midtropospheric trough develops under the influence of the subtropical jet over the continent (Figs. 8c and 11), deepening the trough and resulting in the formation of cutoff lows over the continent. The second of these events (7–13 November) yielded an intense cutoff low (identified via synoptic typing and evident in Fig. 11) that brought widespread rain to South Australia and southeastern Australia as it progressed. This behavior contrasts with November 2020, where synoptic typing shows that fewer midtropospheric troughs moved into Western Australia, as the main storm track steered poleward in these longitudes in 2020, and those that did around 8 and 13 November 2020, moved into an environment unfavorable for development (limited upper-level jet support—Figs. 8b and 10) and were terminated relatively quickly without yielding significant rainfall on the continent.

View Full Size

Hovmöller plots of the midtropospheric flow averaged between 25° and 35°S latitude for (a) November 2020 and (b) November 2021. The shading represents geopotential height anomalies at 500 hPa, and the contours represent geopotential height anomalies at 1000 hPa (e.g., near surface level). The red shading and solid contours indicate positive height anomalies, and blue shading and dashed contours represent negative height anomalies. The green dashed line indicates the west and east longitudes of Australia. Contours are spaced every 20 m. Key periods for November 2021 are indicated with bold numbers: 1) 1–4 Nov 2021, 2) 7–13 Nov 2021, and 3) 21–27 Nov 2021. The “s” indicates event start, and the “e” indicates event end. Data are from the JRA dataset with anomalies calculated relative to 1958–2021.

Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0112.1

• Download Figure
• Download figure as PowerPoint slide

9. Conclusions

We assessed the synoptic and large-scale processes associated with the very different rainfall and temperature anomalies that occurred across Australia during the La Niña events of November 2020 and 2021. Specifically, we set out to answer: Why were the climate outcomes of the two Novembers so different given that they both occurred on a background of La Niña conditions?

We found that some of the more well-known climate drivers for Australia including ENSO, IOD, MJO, and SAM did not provide obvious clues as to why the climate outcomes in November 2020 and 2021 were so different. In this work, we assessed monthly mean and daily diagnostics and highlighted the roles of continental temperatures and the meridional temperature gradient over the Australian continent and their association with the subtropical jet in the dry and wet Novembers.

Our composite picture of La Niña shows the following:

• La Niña sets up warm SSTs around the northern edge of the Australian continent (in the Maritime Continent region).

• If all else equal, those warm SSTs in the north, coupled with a relatively cooler Australian continent, enhance the meridional temperature gradient over Australia and hence thermal wind and subtropical jet over the continent.

• The enhanced subtropical jet provides more favorable conditions for the development of any short-wave features (troughs) moving into the Australian region and facilitates their movement across the continent.

• On average, this process yields wetter conditions over Australia.

However, our study of the two Novembers highlighted the following:

• The meridional temperature gradient over Australia can be sensitive to the broadscale temperature of the Australian continent—since the gradient reflects the difference between warm SSTs to the north (over the Maritime Continent) and the temperatures of the Australian landmass and Southern Ocean to the south.

• In Novembers 2020 and 2021, we have the perfect case study of two La Niña years where

  • In 2020, the continent was warmer than average, weakening the contrast with the temperature over the Maritime Continent. The strongest meridional temperature gradient moved to the southern edge of the continent where the relatively warm continent met the cooler southern oceans. This shift southward in the region of maximum temperature gradient resulted in weak or no subtropical jet over the continent itself, little development of short-wave systems over the continent, and a dry month.

  • In 2021, a persistent trough, associated with quasi-stationary blocking highs to the west and east of the trough, in the midlatitude storm track south of Australia, resulted in persistent advection of cold air over the continent across much of the month, contributing to a much colder than average continent. The colder Australian continent enhanced the temperature contrast with the warm SST over the Maritime Continent. This placed the region of maximum meridional temperature gradient and thermal wind over Australia. That in turn led to a persistent subtropical jet over the continent during the month, which presented a favorable environment for the development of a series of troughs and cutoff lows moving in from the west, and record continental rainfall. These short-wave features, in turn, influence the meridional temperature gradient and thus jet signature.

In short, La Niña and the other drivers are only part of the equation. La Niña tilts the odds to wetter conditions for Australia by warming SSTs to the north of the continent. In any given month, however, broadscale temperature anomalies over the Australian continent, which can be influenced by variability in the midlatitude storm track south of Australia, may strengthen or weaken the meridional temperature gradient and subtropical jet in the region. The presence or absence of the subtropical jet over the continent then plays an important role in enhancing or weakening the impact of storms on the continent.

Here, we have described a common mechanism that contributed to the rainfall anomalies experienced across Australia in November 2020 and 2021. There are clearly other mechanisms also at play, including the factors contributing to the differences in the hemispheric-scale circulation in each November. Exploration of other chapters in the tale of the two Novembers will be the focus of further research.

Data availability statement.

The Japanese Reanalysis dataset can be accessed via https://jra.kishou.go.jp/JRA-55/index_en.html. The Australian Gridded Climate Dataset (AGCD) can be accessed via the Bureau of Meteorology (http://www.bom.gov.au/climate/austmaps/about-agcd-maps.shtml). Optimum Interpolation Sea Surface Temperature (OISST) data can be downloaded from https://www.ncei.noaa.gov/products/optimum-interpolation-sst.