Abstract

Subtropical maritime low pressure systems frequently impact Australia’s eastern seaboard. Closed circulation lows in the Tasman Sea region are termed East Coast Cyclones (ECC); they can evolve in a range of climatic environments and have proven most destructive during the late autumn–winter period. Using criteria based on pressure gradients, inferred wind field, and duration, an objectively determined database of ECC occurrences is established to explore large-scale influences on ECC evolution. Subclassification based on evolutionary trajectory reveals two dominant storm types during late autumn–winter: easterly trough lows (ETL) and southern secondary lows (SSL). Synoptic composites are used to investigate the climatological evolution of each storm type. ETL cyclogenesis occurs along the eastern seaboard at the confluence of warm moist subtropical easterlies and cool air over the continent that is advected from higher latitudes. SSL develop when a cold extratropical cyclone moves equatorward and interacts with warm moist conditions in the Tasman Sea. At seasonal time scales, a complex interplay of tropical and extratropical influences contributes to high-frequency storm seasons. ETL are more frequent during neutral or positive phases of the El Niño–Southern Oscillation, cool sea surface temperature anomalies (SSTAs) in the tropical Indian Ocean, and neutral to positive southern annular mode phases. SSL are more frequent during years with warm SSTAs in the eastern Indian Ocean, warm SSTAs in the western Pacific, and high-latitude blocking.

1. Introduction

Subtropical cyclones typically combine characteristics of tropical and extratropical cyclones (Evans and Guishard 2009) to produce one of the most complex and destructive maritime storm types observed in either hemisphere. Although large anticyclones typically dominate the subtropics, intense cyclonic systems can develop via several mechanisms: transition from the tropics, transition from the extratropics, or in situ cyclogenesis. Throughout their life cycle subtropical storms derive energy from both thermal and baroclinic processes (Hart 2003). Precise structure and evolution varies markedly between regions and seasons; for example, central Pacific Kona storms are most common in the autumn months and typically originate in the extratropics (Otkin and Martin 2004), northwest Atlantic subtropical cyclones frequently undergo tropical transition to become hurricanes (Guishard et al. 2009), and western Pacific subtropical storms affecting New Zealand during the warm season are often tropical cyclones undergoing extratropical transition (Sinclair 2002, 2004). Subtropical maritime cyclones also impact Australia’s eastern seaboard; collectively known as East Coast Cyclones (ECC), this family of storms occur throughout the year and are most common during late autumn and early winter (Hopkins and Holland 1997). Although ECC contribute positively to east Australian water resources (Risbey et al. 2009b) they have historically caused significant damage to coastal infrastructure and ecosystems through strong winds, heavy rainfall and flooding, storm waves, and elevated ocean levels.

Surprisingly little is known about the climatological origin and evolution of ECC events; a major research initiative—the Eastern Seaboard Climate Change Initiative on East Coast Lows (ESCCI-ECL)—is currently under way to address this knowledge gap. ECC have previously been investigated in the context of individual extreme events (Bridgman 1985; Holland et al. 1987; Garde et al. 2010; Qi et al. 2006; McInnes et al. 1992; Mills et al. 2010), heavy rain occurrences (Hopkins and Holland 1997), extreme wave producing systems (Allen and Callaghan 2001; Shand et al. 2011), and elevated ocean levels (Public Works Department 1985, 1986). While there has been some effort to improve the representation of ECCs in forecast models (McInnes et al. 1992; Katzfey and McInnes 1996; Leslie et al. 1987; McInnes et al. 2002), at lead times of more than 1–2 days there is often divergence, meaning that advanced predictions still rely heavily on forecaster experience. At seasonal or longer time scales climatological influences on ECC occurrence are poorly understood and there is almost no predictability.

The most significant ECC impacts have been observed to occur through cumulative damages, when clusters of storms occur in the same winter (Allen and Callaghan 2001). Although Hopkins and Holland (1997) allude to a relationship between high-frequency ECC seasons and transitions in the Southern Oscillation index (SOI), no serious attempt has been made to establish robust linkages between seasonal storm frequency and major drivers of Australian climate variability. A recent publication by Dowdy et al. (2013) uses upper atmospheric diagnostics to examine the risk of ECC events. However, the authors acknowledge that without an accurate event database, their conclusions are difficult to evaluate.

There are several preexisting east coast storm datasets (Holland et al. 1987; Hopkins and Holland 1997; Speer et al. 2011; Public Works Department 1985, 1986; Harper and Granger 2000; Qi et al. 2006). However, none of these have been kept up to date and all have been subjectively determined using different classification schemes. The first step toward understanding ECC evolution is the development of an objective cyclone database. Automated cyclone detection and tracking—as applied in this study—is relatively straightforward (e.g., Raible et al. 2008). The problem, however, is that considerable ambiguity surrounds the precise definition of an ECC. The Australian Bureau of Meteorology (BOM) definition encompasses any closed circulation cyclonic weather system in the east coast maritime environment (Speer et al. 2009). While this definition is sufficient to describe general ECC characteristics, it does not distinguish between storms with different structures or origins. In reality, ECC can develop in situ or enter the east coast region from any direction; to study their evolution a clearly defined classification criteria is required. Origin-based classifications have proven effective in a range of cyclone studies (e.g., Otkin and Martin 2004). In this study we develop a subclassification scheme that is based on storm system origin, and therefore appropriate to study event evolution.

Synoptic climatology techniques provide a powerful tool to investigate the mean structure and evolution of weather events; some recent examples include, the study of cutoff cyclones (Risbey et al. 2009a; Pook et al. 2012), cold events (Ashcroft et al. 2009), heavy precipitation events (Warner et al. 2012), and subtropical cyclones (Otkin and Martin 2004). Applying this approach to study ECC events we identify the underlying mechanisms responsible for storm genesis and resolve key evolutionary differences between storm subtypes. This in turn leads to improved understanding of the large-scale influences responsible for high-frequency storm seasons. Relationships between high-frequency ECC seasons and regional climate drivers may provide some predictability, and allow long-term projections based on climate change scenarios.

The paper is structured as follows: section 2 describes the climate datasets used and the methodology for developing an objective subclassified ECC event database. Section 3 presents the climatological evolution of winter ECC events and the longwave circulation patterns that precede them; results are also presented from the investigation into potential relationships between seasonal ECC frequency and ocean–atmosphere variability in the Indo-Pacific region. Section 4 discusses the association between event scale ECC genesis and low-frequency Indo-Pacific climate variability. We also explore the potential utility of regional climate indices for seasonal ECC forecasting. Conclusions are presented in section 5.

2. Data and methods

a. ECC event database

To construct a database of ECC occurrences a three-stage algorithm was developed to objectively identify, track, and classify ECC events from 1.5° European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis Interim (ERA-Interim) 6-hourly sea level pressure (SLP) data provided by ECMWF covering the period from 1979 to 2011. Cyclone detection is restricted to Australia’s east coast subtropical maritime region, latitudinally bounded between 20° and 40°S and longitudinally bounded between the Australian coastline and 162°E (Fig. 1): the tracking region is unrestricted. Consistent with the BOM definition of an ECC, the algorithm initially identifies all closed contour low pressure systems in the study area by measuring pressure gradients from each 1.5° grid cell to all eight surrounding cells; a closed low is declared when all pressure gradients are negative. This method detects far more cyclones that have been identified in previous subjective analysis such as Speer et al. (2009). To eliminate weaker or less significant closed circulation lows, only those closed lows with a mean pressure gradient of >1 hPa between the central grid cell and all eight surrounding grid cells are retained. Additionally, where two or more closed lows are identified within 3° (two grid cells), only the system with the deepest central pressure is retained. Individual closed lows identified at 6-hourly time steps are grouped into events if the center of the lows are located within 6° (latitude–longitude) at sequential time steps. This method still identifies significantly more events than were identified in Speer et al. (2009); although the objective of our study is not to replicate Speer et al. (2009), their dataset is the most recent record of New South Wales, Australia (NSW), coastal storms and is useful for comparison purposes. To restrict the dataset to only those systems considered to have impacted the eastern seaboard, three additional criteria are adopted to constrain duration, intensity, and inferred geostrophic wind field: 1) the system is required to have a maximum duration of >18 h; 2) the system is required to achieve, at some stage during its existence, a pressure gradient of at least 5 hPa (3°)−1 (latitude–longitude) (~250 km); and 3) the minimum pressure gradient in criterion 2 needs be oriented such that the inferred geostrophic wind field is directed toward the coast (west, southwest, south, and southeast directions). These criteria are not meant to imply that the excluded systems are not ECC events; however, analysis of smaller, weaker, shorter-duration, and more localized systems is beyond the scope of this study and potentially beyond the effective resolution of the ERA-Interim SLP data.

Fig. 1.

Mean storm tracks for each classified storm type: easterly trough lows (ETL), southern secondary lows (SSL), inland trough (IT), continental lows (CL), and extratropical cyclones (XTC). A total of 527 storm events (ETL-124, SSL-169, IT-75, CL-122, and XTC-37) were identified and tracked from ERA-Interim 6-hourly SLP data during all months (January–December) between 1979 and 2011. Classification is based on the trajectory of the storm system prior to day 0—defined as the first time step the system is detected as a closed circulation low pressure system in the Tasman Sea region (boxed region). Storm tracks (colored lines) for each category are determined by calculating the mean location of all events at each 6-hourly time step (denoted by solid markers along storm track) from day 0−48h to day 0+48h. Large hollow circles denote the mean location of each storm type at day 0. Arrows denote the mean direction of movement for all events in each storm category. Note that there are numerous events—especially within the ETL and SSL categories—that display anomalous (with respect to the mean) westward motion; preliminary results suggest that these events have a longer residence time in the Tasman Sea and further research—beyond the scope of this study—is recommended. Small solid markers denote the mean location of each of the 527 individual events during the 48 h prior to day 0 and illustrate considerable variability in the origin of ECC event in general and within each storm category. The thick brown line denotes peak elevation of the Great Dividing Range.

Fig. 1.

Mean storm tracks for each classified storm type: easterly trough lows (ETL), southern secondary lows (SSL), inland trough (IT), continental lows (CL), and extratropical cyclones (XTC). A total of 527 storm events (ETL-124, SSL-169, IT-75, CL-122, and XTC-37) were identified and tracked from ERA-Interim 6-hourly SLP data during all months (January–December) between 1979 and 2011. Classification is based on the trajectory of the storm system prior to day 0—defined as the first time step the system is detected as a closed circulation low pressure system in the Tasman Sea region (boxed region). Storm tracks (colored lines) for each category are determined by calculating the mean location of all events at each 6-hourly time step (denoted by solid markers along storm track) from day 0−48h to day 0+48h. Large hollow circles denote the mean location of each storm type at day 0. Arrows denote the mean direction of movement for all events in each storm category. Note that there are numerous events—especially within the ETL and SSL categories—that display anomalous (with respect to the mean) westward motion; preliminary results suggest that these events have a longer residence time in the Tasman Sea and further research—beyond the scope of this study—is recommended. Small solid markers denote the mean location of each of the 527 individual events during the 48 h prior to day 0 and illustrate considerable variability in the origin of ECC event in general and within each storm category. The thick brown line denotes peak elevation of the Great Dividing Range.

The ECC storm dataset is objectively subclassified based on the SLP origin of the storm system. This requires the objective determination of storm system tracks prior to developing a closed low in the Tasman Sea region. ERA-Interim 6-hourly SLP data are again used to track the evolution of each event for 48 h prior to developing a closed low in the Tasman Sea. The backtracking algorithm operates by locating the storm center at time-step 0—first detection of a closed low in the study region—and then progressively locating the nearest trough at −6-hourly time steps, giving preference to closed or semiclosed lows.

Prestorm trajectory data are used to objectively subclassify the ECC event dataset into five main ECC storm types previously defined in Public Works Department (1985, 1986). Factors evaluated by the classification algorithm are the following: storm-track latitude, whether the storm evolved mostly over the land or the sea, whether it tracked north or south, and its orientation to Australia’s main topographic divide (i.e., the Great Dividing Range). Consideration of the Great Dividing Range was found to be essential, as orographic uplift is believed to be a key forcing in the initial development of many ECC events (Holland et al. 1987; Leslie et al. 1987). Storm-type nomenclature and specific criteria are as follows:

  1. Easterly trough lows (ETL): Events that track mostly east of the Great Dividing Range and in a southerly direction.

  2. Southern secondary lows (SSL): Events that track mostly over the ocean and in a northerly direction.

  3. Inland troughs (IT): Events that evolve mostly over land, west of the Great Dividing Range and north of 30°S.

  4. Continental lows (CL): Events that evolve mostly over land, west of the Great Dividing Range and south of 30°S.

  5. Extratropical cyclones (XTC): Differentiation between canonical ETL events and those that evolve from tropical cyclones based on storm track trajectories is problematic, as both storm types evolve in a similar region and follow similar tracks. However, the BOM maintains a database of all tropical cyclone occurrences and tracks from 1900 to present (available online at www.bom.gov.au). Events that evolved from storm systems included in this database are classified as XTC.

Mean storm tracks for each of the five ECC subclasses shown in Fig. 1 are consistent with the classification criteria. Individual markers show the mean location of each identified event during the 48 h prior to developing a closed low in the Tasman Sea and illustrate considerable variability in the origins of ECC storm systems in general, and within each subgroup.

b. Data and methods to investigate daily evolution and longwave circulation

A synoptic typology approach is employed to investigate the mean event scale evolution of winter ECC from 2 days prior (day −2) to 1 day after (day +1) each event developing a closed low in the Tasman Sea region. The evolving relationships between four atmospheric variables are explored using sequential composites of ERA-Interim data: 1) SLP anomalies (SLPA: all anomalies are calculated with reference to the 1979–2010 long-term mean) illustrate storm system origin and mean surface evolution; 2) 300-hPa mean zonally winds illustrate inferred jet stream strength and orientation; 3) atmospheric thickness anomalies—difference between geopotential height (GPH) at 1000 and 500 hPa—illustrate the interaction of air masses with different thermal properties; and 4) determination of warm-core (barotropic) verse cold-cored (baroclinic) vertical structure at two tropospheric levels is used identify the processes—baroclinic versus barotropic—driving storm intensification. Hart (2003) developed a phase space method of classifying evolving subtropical cyclones based on thermal wind. The method exploits the fundamental difference in thermal wind structure between tropical and extratropical cyclones: geostrophic wind magnitude above the cyclone center increases (decreases) with altitude in baroclinic (barotropic) systems; summarized briefly here, the method is explained in detail in Hart (2003). Vertical-longitudinal cross sections of geopotential height at 12 levels between 1000 and 100 hPa are calculated for each time step (at 1200 UTC) of each event over an east–west longitudinal range of 25.5° (17 grid cells) centered on the storm system. Maximum geostrophic winds are calculated at each pressure level over a radius of two grid cells (3° longitude ~350 km). Geostrophic wind magnitudes are interpolated between levels to give 16 values over which to calculate the thermal winds. The sign and magnitude of the thermal winds V are calculated from the vertical gradient in geostrophic wind magnitude at two levels, using a linear regression fit between 900 and 600 hPa (Vl) and between 600 and 300 hPa (Vu). Positive (negative) values of V indicate a cold-core (warm core) cyclone within the layer. Thermal winds for all events are used to provide a mean classification for each storm type, at each time step, as either predominantly warm cored (barotropic) or predominantly cold cored (baroclinic) in the upper and lower troposphere (after Hart 2003): where >60% of events are positive (negative), the mean structure—for that storm type at that time step—is classified as cold cored (warm cored); otherwise, a neutral classification is given.

Longwave circulation can precondition the local environment and promote storm genesis. Geopotential height data at the 500-hPa level are commonly used to identify longwave circulation patterns and regions of blocking (e.g., Lamb 1959; Trenberth 1980; Kidson and Sinclair 1995). Composites of 1200 UTC ERA-Interim 500-hPa GPH anomalies at day −5 are used to identify the long-wave circulation features prior to event formation.

c. Method to explore seasonal drivers of high-frequency ECC winters

Atmospheric circulation patterns associated with high-frequency ECC seasons are identified by gridpoint correlation with seasonal mean ERA-Interim 500-hPa GPH anomaly data and comparison to major modes of atmospheric variability. The southern annular mode (SAM) and Pacific–South American mode (PSA) each describe about 20%–30% of seasonal atmospheric circulation variability in the Southern Hemisphere midlatitudes (e.g., Mo and Paegle 2001). SAM influences the latitude and amplitude of the longwave pattern with a SAM positive (negative) characterized by positive (negative) pressure anomalies in the midlatitudes and negative (positive) pressure anomalies in the high latitudes (e.g., Thompson and Wallace 2000; Gong and Wang 1999). The PSA modes are primarily driven by Rossby wave propagation from the tropics and are associated with a high wavenumber meridional longwave pattern in the midlatitudes (Karoly 1989; Mo and Ghil 1987; Mo and Higgins 1998). SAM and PSA time series are calculated—using the method of Visbeck and Hall (2004)—from the principal component time series of the leading EOFs of monthly mean ERA-Interim 850-hPa GPH anomaly data between 20° and 80°S. Following the conventions of Mo and Higgins (1998) and Visbeck and Hall (2004) SAM is defined as the principle component time series of EOF1, while PSA1 and PSA2 are the principle component time series of EOF2 and EOF3, respectively.

Indo-Pacific sea surface temperature anomalies (SSTA) influence atmospheric circulation in the Australian region and are commonly used as a basis for seasonal forecasting (Holton 1998; Langford and Hendon 2013). Regions of significant SSTA associated with high-frequency ECC seasons are identified by gridpoint correlation with seasonal Integrated Global Ocean Services System (IGOSS) Reynolds optimum interpolation (OI) version 2, 1° × 1° SSTA data (Reynolds et al. 2002) and comparison to major regional modes of coupled ocean–atmosphere variability. IGOSS SST data are used to calculate commonly used climate indices for the Indian Ocean region: west Indian Ocean (WIO), east Indian Ocean (EIO), and Indian Ocean dipole mode index (DMI) as defined by Saji et al. (1999) and the Indonesian Throughflow (ITF) index as defined by Cai and Cowan (2008). To represent the oceanic component of the El Niño–Southern Oscillation (ENSO) we calculate the Niño-3.4 index as defined by Trenberth (1997). To represent the atmospheric component of ENSO we use the SOI obtained from the Australian BOM.

Climatic drivers of high-frequency ECC seasons are investigated by linear correlation between indices and seasonal storm day frequency. Bearing in mind the analysis period (1979–2011) covers only 33 yr and relationships may be nonlinear, we also identify the phase of each index for individual high-frequency ECC seasons. A high-frequency season is defined as a season with storm day occurrence in the top 15% of seasons between 1979 and 2011. To facilitate intercomparison, all indices have been normalized to Z scores (anomaly/σ). Seasonal index phase based on mean Z scores are defined as follows: ±0–0.4σ = neutral, ±0.5–1σ = positive (negative), and greater than ±1σ = very positive (very negative).

3. Results

a. Seasonal variability in storm system origin

ECC events occur throughout the year with peak occurrences in May–June and again in September. Figure 2a shows a clear seasonal progression in storm type and origin; during summer most ECC events evolve from tropical cyclones—XTC. In autumn and early winter most events develop in situ along the mid-north coast—ETL. During winter and early spring most storms originate in the extratropics—SSL. In spring and early summer most events originate over the Australian continent—CL. The IT can occur at any time of year, typically making up about 10%–20% of storm days in all seasons.

Fig. 2.

(a) Percentage contribution of each storm type to total monthly ECC storm day occurrence. (b) Total number of storm days in each winter (May–August).

Fig. 2.

(a) Percentage contribution of each storm type to total monthly ECC storm day occurrence. (b) Total number of storm days in each winter (May–August).

Late autumn to winter is of specific climatological interest as ECC events have historically had the largest impact on eastern Australia during this period (Hopkins and Holland 1997). Event occurrence in the early winter is dominated by both ETL and SSL events, yet they have different geographic origins (Fig. 1). The remainder of this paper will therefore focus on the climatological evolution of ETL and SSL events occurring during the extended winter period from May to August.

b. Synoptic life cycle of winter ECC events

During the May–August winter period (1979–2011) there were 57 ETL and 67 SSL events detected. Figures 3 and 4 present synoptic composites of SLPA, 300-hPa zonal wind, thickness anomaly, and thermal wind for all events from day −2 to day +1. Day 0 is the first day that the system was detected as a closed surface low in the Tasman Sea. Mean storm tracks are also plotted and the mean locations of the storm system at each time step are indicated with a cross. Composites are produced independently of track information and therefore provide an objective assessment of the accuracy of the tracking algorithm.

Fig. 3.

Synoptic composites of ETL evolution at 1200 UTC for days −2 to +1 of each event. Composites are constructed by averaging data from all 57 ETL events. Cool (warm) colors indicate low (high) values. Bold green contours encircle regions of statistical significance exceeding 95%. The bold red line shows the mean storm track from day −2 to day +1. Crosses indicate the mean storm location at each time step. (a) SLP anomaly (hPa), (b) mean zonal wind at 300 hPa (m s−1), (c) thickness anomaly (500–1000-hPa geopotential height); (d) (left panel) east–west cross section of geopotential height over the developing cyclone, vertical lines indicate 3° radius around storm center, (right panel) geostrophic wind magnitude calculated at each level from the height difference within the 3° radius. At each time step, the mean storm structure is classified as warm cored (barotropic) or cold cored (baroclinic) at two levels, based on the observation that geostrophic wind magnitude increases (decreases) with altitude in cold (warm) cored systems [after Hart (2003)]. The thermal wind (V) is the vertical gradient in geostrophic wind magnitude calculated from the slope of the liner regression fit (opaque green lines) at two tropospheric levels: 900–600 hPa (Vl) and 600–300 hPa (Vu). Mean fit lines for Vu and Vl are plotted; however, classification is based on the individual Vu and Vl values for all events, where >60% of events are positive (negative), the mean structure is classified as cold cored (warm cored). Otherwise, Vu and Vl are approximately zero and the classification is neutral.

Fig. 3.

Synoptic composites of ETL evolution at 1200 UTC for days −2 to +1 of each event. Composites are constructed by averaging data from all 57 ETL events. Cool (warm) colors indicate low (high) values. Bold green contours encircle regions of statistical significance exceeding 95%. The bold red line shows the mean storm track from day −2 to day +1. Crosses indicate the mean storm location at each time step. (a) SLP anomaly (hPa), (b) mean zonal wind at 300 hPa (m s−1), (c) thickness anomaly (500–1000-hPa geopotential height); (d) (left panel) east–west cross section of geopotential height over the developing cyclone, vertical lines indicate 3° radius around storm center, (right panel) geostrophic wind magnitude calculated at each level from the height difference within the 3° radius. At each time step, the mean storm structure is classified as warm cored (barotropic) or cold cored (baroclinic) at two levels, based on the observation that geostrophic wind magnitude increases (decreases) with altitude in cold (warm) cored systems [after Hart (2003)]. The thermal wind (V) is the vertical gradient in geostrophic wind magnitude calculated from the slope of the liner regression fit (opaque green lines) at two tropospheric levels: 900–600 hPa (Vl) and 600–300 hPa (Vu). Mean fit lines for Vu and Vl are plotted; however, classification is based on the individual Vu and Vl values for all events, where >60% of events are positive (negative), the mean structure is classified as cold cored (warm cored). Otherwise, Vu and Vl are approximately zero and the classification is neutral.

Fig. 4.

As in Fig. 3, but for 69 SSL events.

Fig. 4.

As in Fig. 3, but for 69 SSL events.

1) ETL events

Figure 3a shows that two days prior to a typical ETL event a surface trough begins to develop along the eastern seaboard at ~18°S. A broad region of high pressure extends from south of the Australian continent to the northeast Tasman Sea. Trough intensification upstream, near Western Australia has been noted prior to many events and serves to strengthen the subtropical jet (STJ) and amplify the midlatitude ridge (Mills et al. 2010). Strengthened 300-hPa zonal wind at ~20° and ~60°S indicates a well-developed split jet (Fig. 3b). Thickness and thermal wind composites show a midtropospheric cold pool over the Australian mainland (Fig. 3c) to the west of the developing cyclone (Fig. 3d).

As the storm intensifies from day −1 to day +1 the SLP trough deepens along the east coast while the anticyclone intensifies and moves from the southwest to the southeast (Fig. 3a). The 300-hPa zonal wind composites (Fig. 3b) show a strengthening of the polar front jet (PFJ) and curvature of the STJ over the developing cyclone. Thickness composites (Fig. 3c) show a thermal gradient intensifying over the storm, with cool anomalies to the west and warm anomalies to the east. Thermal wind profiles (Fig. 3d) indicate an initial baroclinic structure, with a cold core in the upper troposphere to the west of the developing storm. By day 0 the storm begins to take on the characteristics of a warm core in the lower troposphere due to the inflow of warm moist air at the surface. By day +2 (not shown) the cool thickness anomaly over the Australian continent has mostly dissipated and the storm acquires a predominantly warm core structure in both the lower and upper troposphere.

ETLs tend to weaken as they move offshore; late autumn SST in the Tasman Sea are typically around 23°C and this is insufficient to maintain a barotropic storm (Dare and McBride 2011; Davis and Bosart 2003). Figure 3 illustrates the mean evolution of ETL events; there are many individual exceptions. For example, partial tropical transition has been observed as late as March (Qi et al. 2006; Garde et al. 2010). There are also numerous examples of ETL events making a full extratropical transition to the midlatitudes.

2) SSL events

Figure 4a shows that two days before a typical SSL event a well-developed low pressure system exists to the south of the Australian continent. A broad region of low SLPA over the Australian mainland results from the influence of northwest trough systems that often precede individual events. These systems extend from the tropical Indian Ocean through to southeast Australia and are commonly referred to as northwest cloud bands. They are most frequent during the winter months and often occur in conjunction with an extratropical cyclone (Tapp and Barrell 1984). The extratropical cyclone at ~40°S has positive SLPA upstream and downstream and appears to be part of a propagating wave train evident in both the SLPA and thickness anomaly fields (Fig. 4c). The 300-hPa zonal winds (Fig. 4b) indicate a strengthening of the STJ equatorward of the storm system; however, the PFJ is not as strongly developed during SSL events as it is during ETL events. Thickness composites (Fig. 4c) indicate a pool of cold air is associated with the surface cyclone. Thermal wind profiles (Fig. 4d) show an initial baroclinic structure with a well-defined cold pool in the upper troposphere.

As the storm intensifies from day −1 to day +1 the extratropical cyclone and upper-tropospheric cold pool move in tandem from the Southern Ocean into the Tasman Sea. The anticyclone centers merge poleward of the low and intensify. Figure 4d shows that once the system moves over the warmer waters of the Tasman Sea it begins to take on the characteristics of a subtropical storm and develop a partial warm core in the lower troposphere. As noted by Mills et al. (2010) this is more of a warm seclusion as opposed to the warm core typically associated with tropical cyclones. A warm (positive thickness) anomaly develops over the eastern Tasman Sea (Fig. 4c) as warm moist air is advected in the strengthened easterly flow. By day +1 the vertical profile has straightened, yet unlike ETL events, SSL events appear to retain the cold core in the upper troposphere. Figure 4 illustrates mean SSL evolution. The SSL category contains the most ECC storms and there may be multiple mechanisms for their formation; for example, hybrid systems with characteristics of both ETL and SSL events have been observed and warrant further investigation.

c. Hemispheric longwave circulation

Figures 5a and 5b show mean 500-hPa GPH anomalies at day −5 (1200 UTC) before all (57) ETL and (69) SSL events; they illustrate the longwave circulation pattern prior to event formation. Negative anomalies in the midlatitudes indicate the locations of longwave troughs. Positive anomalies in the midlatitudes indicate blocking anticyclones that tend to be persistent and frequently “block” or divert the normal passage of extratropical cyclones (e.g., Sinclair 1996); they are often associated with cold-air outbreaks due to anticyclonic advection around the eastern periphery (Mo et al. 1987; Ashcroft et al. 2009).

Fig. 5.

Mean 500-hPa geopotential height anomaly (m) at 1200 UTC on day −5 for (a) 57 ETL and (b) 69 SSL events. Cool (warm) colors indicate low (high) values. Bold green contours enclose regions of statistical significance exceeding 95%. Mean storm tracks for all identified ETL/SSL events are plotted in red.

Fig. 5.

Mean 500-hPa geopotential height anomaly (m) at 1200 UTC on day −5 for (a) 57 ETL and (b) 69 SSL events. Cool (warm) colors indicate low (high) values. Bold green contours enclose regions of statistical significance exceeding 95%. Mean storm tracks for all identified ETL/SSL events are plotted in red.

Prior to ETL formation (Fig. 5a), anticyclonic circulation south of Australia cools the continent by advecting cold air from higher latitudes. Anticyclonic circulation around the Tasman Sea ridge has the opposite effect by simultaneously advecting warm moist air from the subtropics. ETL genesis occurs in the baroclinic environment at the confluence of these two airstreams. The Tasman Sea ridge remains essentially stationary during event evolution. In contrast, the ridge southwest of Australia is mobile and propagates eastward, covering ~10° longitude day−1. The midlatitude longwave pattern amplifies in conjunction with a low pressure system in the subtropical Indian Ocean, suggesting Rossby wave propagation from the tropics: a hypothesis that will be investigated in the next section.

Prior to SSL formation (Fig. 5b), negative anomalies in the southeast Indian Ocean are consistent with an extratropical origin. Amplification of the longwave trough in this region is also conducive to the development of northwest cloud-band events (Tapp and Barrell 1984). Similar to ETL events, the midlatitude ridge–trough couplet (initially in the southern Indian Ocean) propagates eastward covering ~10° longitude day−1. High-latitude blocking south of the Tasman Sea is a quasi-stationary feature, it moves only a few degrees during the evolution of the storm. Persistent blocking in this region is linked to a meridional storm track and airmass exchange with the cold Antarctic continent (Pook and Gibson 1999; Massom et al. 2004).

d. Interannual variability in ECC occurrence

Interannual variability in the total number of winter ECC storm days, and within each storm type, is shown in Fig. 2b. High-frequency ECC seasons—defined as having a storm day frequency in the top 15%—occurred in 1984, 1989, 1990, 1998, and 2007. The highest-frequency ECC winter, 1998, was dominated by IT and CL events. The years 1984, 1990, and 2007 were mostly dominated by ETL events and 1989 was the highest SSL winter on record (1979–2011). High-frequency IT seasons can co-occur with ETL or SSL seasons; however, ETL and SSL occurrence are anticorrelated (r = −0.44, p = 0.02) and share no common high-frequency seasons. Therefore, interannual variability in total ECC days can be determined by changes in the frequency of ETL, SSL, CL, or IT events.

Total ECC storm day frequency is moderately correlated to the SOI (r = 0.57, p < 0.01) and Niño 3.4 (r = −0.41, p = 0.02), weakly correlated to the DMI (r = −0.3, p = 0.1), and uncorrelated to the SAM and PSA indices. Table 1 shows that most high-frequency seasons occurred with cool or neutral SSTA in Niño 3.4 and all had neutral or positive SOI. The DMI was negative for all years excluding 2007 and the ITF was positive or neutral for most years. There is no consistency in the sign of the SAM; however, PSA1 was neutral or negative for all years. Because ETL and SSL events evolve from different regions it is difficult to ascribe a mechanism to remote climate driver relationships in terms of total ECC frequency—therefore, we evaluate each storm type separately.

Table 1.

Index phase for the highest-frequency storm seasons (top 15%). All climate indices have been normalized to Z scores (anomaly/σ). Negative values (<0.5σ) are italicized and positive values (>0.5σ) are boldface.

Index phase for the highest-frequency storm seasons (top 15%). All climate indices have been normalized to Z scores (anomaly/σ). Negative values (<0.5σ) are italicized and positive values (>0.5σ) are boldface.
Index phase for the highest-frequency storm seasons (top 15%). All climate indices have been normalized to Z scores (anomaly/σ). Negative values (<0.5σ) are italicized and positive values (>0.5σ) are boldface.

1) ETL

Because ETL events are infrequent in late winter (August) we restrict the seasonal investigation to May–June–July. The six highest-frequency ETL seasons (top 15%) were 1984, 1985, 1990, 1996, 1999, and 2008. Seasonal ETL frequency is correlated to cool SST in the central Pacific (Fig. 6a) and low 500-hPa GPH over northern Australia (Fig. 6e), which is consistent with La Niña conditions. Although correlations with Niño 3.4 and the SOI are weak (p > 0.05) ⅚ highest-frequency ETL seasons occurred with cool or neutral SSTA in Niño 3.4 and all had positive or neutral SOI. An Indian Ocean influence is implied by statistically significant gridpoint correlations with both SST and 500-hPa GPH (Fig. 6). Although seasonal ETL frequency is uncorrelated to the DMI, it is moderately correlated to the WIO index (r = −0.41, p = 0.03). Table 1 shows that all six highest-frequency ETL seasons occurred in conjunction with cool or neutral SSTA in the WIO and EIO. Positive SLP and 500-hPa GPH correlations (Figs. 6c,e) across the Tasman Sea are consistent with blocking high pressure that develops during ETL intensification. No statistically significant atmospheric correlations are observed poleward of ~55°S and seasonal ETL frequency is uncorrelated to the SAM and PSA indices. However, SAM may still influence ETL formation, as all six high-frequency ETL seasons occurred with weakly positive or neutral SAM. These results suggest a preferred combination of regional climate drivers for high-frequency ETL seasons: neutral or La Niña–like conditions in the Pacific, cool SSTA in the tropical Indian Ocean, and neutral or positive SAM.

Fig. 6.

Gridpoint Pearson’s correlation (r) between seasonal (a),(c),(e) ETL (May–July) and (b),(d),(f) SSL (May–August) storm day frequency and (a),(b) IGOSS SSTA; (c),(d) ERA-Interim SLP; and (e),(f) ERA-Interim 500-hPa geopotential height. Cool (warm) colors indicate negative (positive) correlations (r values). Bold green contours enclose regions of statistical significance exceeding 95%. Mean storm tracks for all identified ETL/SSL events are plotted in red.

Fig. 6.

Gridpoint Pearson’s correlation (r) between seasonal (a),(c),(e) ETL (May–July) and (b),(d),(f) SSL (May–August) storm day frequency and (a),(b) IGOSS SSTA; (c),(d) ERA-Interim SLP; and (e),(f) ERA-Interim 500-hPa geopotential height. Cool (warm) colors indicate negative (positive) correlations (r values). Bold green contours enclose regions of statistical significance exceeding 95%. Mean storm tracks for all identified ETL/SSL events are plotted in red.

2) SSL

The five highest-frequency SSL winters (top 15%) were 1981, 1989, 1995, 2001, and 2011. Seasonal SSL frequency is correlated to warm SSTA in the EIO, Indonesia, and the western Pacific (Fig. 6b); a pattern that is consistent with a strengthened ITF. Seasonal frequency is moderately correlated to the ITF index (r = 0.42, p = 0.015) and uncorrelated to all other major indices. Table 1 shows that all five high-frequency SSL seasons had warm or neutral SSTA in the ITF and EIO. SSL frequency is correlated to low SLPA over most of Indonesia (Fig. 6d). The region of strongest correlations (~15°S, ~100°E) is the main genesis location for northwest cloud-band systems (Tapp and Barrell 1984). Negative SLP correlations over SE Australia identify the mean location of SSL events. Positive SLP and 500-hPa GPH correlations at ~60°S indicate high-latitude blocking (south of the Tasman Sea) is a persistent feature of high-frequency SSL seasons. Positive correlations at ~30°S, 170°W indicate a strengthened west Pacific subtropical ridge and enhanced easterlies at ~20°S. Warm SSTA in the western Pacific, combined with enhanced easterly winds drive warm-air advection into the southern Tasman Sea region. These results show that although SSL frequency is not well correlated to the major climate drivers (e.g., SOI and SAM), there are several key features associated with high-frequency seasons: warm SSTA and enhanced convection in the tropical east Indian Ocean, high-latitude blocking south of the Tasman Sea, and warm SSTA in the western Pacific.

4. Discussion

SLPA composites for day 0 of SSL and ETL events are visually similar; both storm types intensify in the subtropics and display a north–south-orientated trough–ridge couplet that is typical of most ECC events (Hopkins and Holland 1997). Major differences between the two storm types become evident during the days prior to developing a closed low in the Tasman Sea. ETL events develop within the subtropics, as a dip in the easterlies, along the coast at ~18°S. SSL events develop out of preexisting extratropical low pressure systems that move equatorward into the Tasman Sea. Both storm types have strongly baroclinic structures during their initial development; however, these arise out of fundamentally different processes. Likewise, at the seasonal time scale, event frequency is associated with climatic drivers in different regions.

a. ETL

Our results support previous research (e.g., Hopkins and Holland 1997; Holland et al. 1987) that the primary mechanism driving ETL genesis is the mesoscale interaction between warm moist subtropical airflow and cold dry continental air; convection is enhanced by orographic uplift over steep coastal topography. Therefore, large-scale conditions that promote ETL genesis (Fig. 7a) are enhanced subtropical easterlies, cooling over the continent, and a Walker circulation that is conducive to mesoscale convection over northeast Australia.

Fig. 7.

Summary diagram of identified relationships for (a) event-scale ETL evolution and (b) event-scale SSL evolution, where L indicates low pressure anomalies and H indicates high pressure anomalies; arrows indicate direction of circulation and air advection. Gray arrows show direction of longwave propagation. (c) Location of indices used to explain seasonal variability in ETL frequency: WIO SSTA, SOI, and SAM. (d) Location of indices used to explain seasonal variability in SSL frequency: ITF SSTA, southwest Pacific (SWP) SSTA, and East Antarctica (EA) SLP.

Fig. 7.

Summary diagram of identified relationships for (a) event-scale ETL evolution and (b) event-scale SSL evolution, where L indicates low pressure anomalies and H indicates high pressure anomalies; arrows indicate direction of circulation and air advection. Gray arrows show direction of longwave propagation. (c) Location of indices used to explain seasonal variability in ETL frequency: WIO SSTA, SOI, and SAM. (d) Location of indices used to explain seasonal variability in SSL frequency: ITF SSTA, southwest Pacific (SWP) SSTA, and East Antarctica (EA) SLP.

Subtropical easterly airflow is enhanced by anticyclonic circulation associated with blocking high pressure over the southern Tasman Sea and New Zealand; a synoptic situation that is more (less) frequent during SAM positive (negative) (Jiang et al. 2012).

Large-scale convection over northeast Australia is dominated by the Walker circulation and enhanced (suppressed) during La Niña (El Niño) conditions (Philander 1983). High-frequency ETL seasons do not occur during negative SOI years, yet there is no preference for strongly positive SOI years either. This is because ETL genesis is not dependent upon large-scale convection alone. They form in a mesoscale baroclinic environment with orographically enhanced convection. However, large-scale (Walker cell) subsidence associated with SOI negative (El Niño) conditions would actively suppress locally generated convection. Hence there are no high-frequency ETL years during SOI negative (El Niño).

Continental cooling prior to ETL events allows baroclinic cyclogenesis to occur equatorward of 20°S: latitudes usually reserved for tropical cyclones (e.g., Shapiro and Goldenberg 1998). Cold-air outbreaks over Australia are closely linked to midlatitude blocking (Ashcroft et al. 2009). The blocking pattern that drives cold-air advection prior to ETL events forms as part of a wave train emanating from the tropical Indian Ocean (Figs. 5a and 7a). Cai et al. (2011b) show that suppressed convection in the tropical Indian Ocean drives a Rossby wave train that amplifies low pressure in the southeast Indian Ocean and blocking south of the Australian continent. Cool Indian Ocean SSTA observed during all high-frequency ETL seasons may act to suppress convection and generate similar Rossby waves trains. Cai et al.’s (2011b) findings focused on the EIO wave train in JJA; further research will be needed to fully understand the dynamics of Rossby wave generation in the central and western Indian Ocean during late autumn and its influence on southern Australia. During winter, blocking south of Australia is also influenced by the SAM (Cai et al. 2011a). Our results show that all high-frequency ETL seasons have occurred during neutral or moderately positive SAM; suggesting that a negative or strongly positive SAM may have a destructive influence on the Indian Ocean wave train.

Low-frequency variability associated with tropical SSTA implies that Rossby wave propagation would tend to reestablish blocking in the same region throughout a season. This would have the effect of replenishing the cold pool over the Australian continent after each event. If this occurs in conjunction with an overall strengthening of the subtropical ridge in the Tasman Sea and neutral to positive SOI, multiple recurrent events would be expected during the same season.

b. SSL

Our results show that winter SSL events typically form when a cold-cored extratropical cyclone moves equatorward over the Tasman Sea, often in conjunction with a northwest cloud-band-type system (Fig. 7b). In contrast to ETL events, the cold air associated with SSL events is advected into the Tasman Sea in conjunction with the surface cyclone. Cold-air advection can be so pronounced that anomalous snowfalls are often observed at low elevations (Bridgman 1985). Highly baroclinic conditions develop when the cold extratropical cyclone interacts with warm moist air, associated with the northwest cloud band, over the relative warmth of the Tasman Sea. The barocline is further strengthened by the easterly inflow of warm moist subtropical air, associated with the developing anticyclone poleward and eastward of the storm. In contrast to ETL events, the midtropospheric cold pool associated with SSL storms typically does not dissipate during the event; this may have implications for reintensification in the Tasman Sea or over New Zealand. Therefore, the large-scale conditions that promote SSL cyclogenesis are those that increase air temperatures in the Tasman Sea region, increase the frequency of northwest cloud-band events, and drive a meridional extratropical storm track.

Warm SSTA in western Pacific, combined with a strengthening of the west Pacific subtropical ridge, enhance the advection of warm moist surface air into the Tasman Sea region. Northwest cloud-band frequency in winter is closely linked to warm SSTA and enhanced tropical convection in the ITF and EIO region (Cai and Cowan 2008; Tapp and Barrell 1984). Enhanced tropical convection in the ITF region also drives an equivalent barotropic Rossby wave train that strengthens blocking south of the Tasman Sea (Cai and Cowan 2008; Cai et al. 2011b). Much of the interannual variability in winter rainfall over southeast Australia is associated with warm SSTA in the ITF (Cai et al. 2009) and the occurrence of cutoff cyclones (Risbey et al. 2009b). Although enhanced convection in the ITF region drives both northwest cloud bands and blocking south of the Tasman, there still may be an element of chance involved in the synchronization of the northwest trough and extratropical cyclone.

A meridional extratropical storm track may be forced by high-latitude blocking south of the Tasman Sea (Pook and Gibson 1999). The extratropical origins of SSL combined with significant correlations at high latitudes suggest an Antarctic influence on SSL genesis: especially as SSL events are associated with the anomalous advection of very cold air. Cold-air outbreaks from the Antarctic continent can drive mesocyclogenesis and transport cold air to the subtropics (Bromwich 1991; Parish et al. 1994). Cold-air outbreaks occur primarily because of katabatic outflow that is topographically controlled (Carleton 1992). The Amery ice shelf is one of Antarctica’s primary outflow regions for katabatic winds and northward mass transport (Bains and Fraedrich 1989; Parish 1992; Parish and Bromwich 1998, 2007). Statistically significant SLPA (both positive and negative) are observed in this region up to 10 days prior to SSL formation (SLPA composites for day −10 not shown). It is feasible, but remains to be proven, that cold-air outflow from Antarctica may play a role in SSL genesis.

c. Summary of the relationship between major climate drivers and East Coast Cyclones

There are no simple linear relationships between seasonal ECC frequency and any of the major modes of Indo-Pacific climate variability. Total winter ECC frequency is moderately anticorrelated to the DMI, yet the Indian Ocean appears to influence each storm type differently. ETL frequency is related to cool SSTA in the WIO and EIO, while SSL frequency is correlated to warm SSTA in the EIO and ITF region.

Total winter ECC frequency is moderately correlated to La Niña; however, ENSO influences the formation of each storm type differently. ETLs form at latitudes where they can be influenced directly by ENSO via reduced or enhanced subsidence over northern Australia. SSLs form at higher latitudes where ENSO has little direct influence in the winter; remote ENSO influences are mostly confined to the PSA pattern with maximum amplitude in the South Pacific (Mo and Paegle 2001). The PSA modes may influence ECC formation via a hemispheric perturbation in the longwave circulation; however, our results—spanning only the past 33 years—are inconclusive. The tropical Indian Ocean is the main pathway for ENSO to influence southern Australia in winter (Cai et al. 2011b). Warm SSTA in the Indo-Pacific that promote SSL events may be indirectly related to ENSO, as neutral or trending La Niña conditions drive a stronger/warmer ITF (Cai and Cowan 2008).

As previously mentioned, all high-frequency ETL seasons have occurred with a neutral or moderately positive SAM. SAM positive is associated with a strengthened ridge in the Tasman Sea (Cai et al. 2011a; Jiang et al. 2012) and a well-developed split jet (Gallego et al. 2005; Bals-Elsholz et al. 2001); both conducive to ETL genesis. We also speculate that a negative or strongly positive SAM may override the influence of Rossby wave propagation from the tropical Indian Ocean. The phase of the SAM appears to have little direct impact on SSL frequency; a negative SAM would be expected to result in more extratropical cyclones moving into the subtropics, while a SAM positive is more conducive to a strengthened subtropical ridge in the western Pacific. Therefore, the SAM may not be the most appropriate index to describe SSL variability.

The Madden–Julian oscillation (MJO) has a strong influence on convection in the Indo-Pacific region (Madden and Julian 1994) and would be expected to play some role in ETL and SSL formation. A cursory investigation of the role of the MJO on ETL and SSL frequency revealed no significant relationships: a detailed investigation is beyond the scope of this paper.

d. Forecasting ECC events

Our results indicate that there is potential to improve ETL and SSL forecasts at the event and seasonal time scales. At lead times of up to 5 days, synoptic composites identify key regions of 500-hPa GPH anomalies that are statistically and mechanistically related to event formation (Figs. 7a,b). Indicators of a developing ETL event are positive anomalies (500-hPa GPH) southwest of Australia combined with positive anomalies over the Tasman Sea. Indicators of a developing SSL event are negative anomalies (500-hPa GPH) southwest of Australia combined with positive anomalies at ~60°S, 130°E. The next phase of this work will investigate whether these relationships can be used in a predictive capacity to supplement conventional numerical forecasting.

At the seasonal time scale, relationships between past event frequency and regional climate drivers can provide an indication of the likelihood of a high-frequency ETL or SSL season. The identified relationships are complex and interdependent; multiple components of the climate system must be evaluated. The three main regions with relevance for ETL seasonal event probability are shown in Fig. 7c: WIO—cool, SAM—neutral or positive, and SOI—neutral or positive. During the past 33 years this combination of drivers (in May–June–July) has occurred 7 times (Table 2a); six of these years were the highest frequency ETL seasons on record.

Table 2.

(a) Number of storm days for all years conforming to conditions associated with high-frequency of ETL events: WIO cool, SOI neutral or positive, and SAM neutral or positive. (b) Number of storm days for all years conforming to conditions associated with high-frequency of SSL events: ITF positive or neutral, southwest Pacific SSTA positive or neutral, and East Antarctic SLPA positive or neutral. See section 2d for definitions of positive, neutral, and negative.

(a) Number of storm days for all years conforming to conditions associated with high-frequency of ETL events: WIO cool, SOI neutral or positive, and SAM neutral or positive. (b) Number of storm days for all years conforming to conditions associated with high-frequency of SSL events: ITF positive or neutral, southwest Pacific SSTA positive or neutral, and East Antarctic SLPA positive or neutral. See section 2d for definitions of positive, neutral, and negative.
(a) Number of storm days for all years conforming to conditions associated with high-frequency of ETL events: WIO cool, SOI neutral or positive, and SAM neutral or positive. (b) Number of storm days for all years conforming to conditions associated with high-frequency of SSL events: ITF positive or neutral, southwest Pacific SSTA positive or neutral, and East Antarctic SLPA positive or neutral. See section 2d for definitions of positive, neutral, and negative.

Seasonal SSL frequency is not well correlated to any of the conventional climate indices (e.g., SOI and SAM). The three main regions with relevance for seasonal SSL event probability are shown in Fig. 7d: ITF—warm or neutral, southwest Pacific SSTA (35°–25°S, 175°E–175°W)—warm or neutral and East Antarctic SLPA (65°–50°S, 130°–155°E)—positive or neutral. Indices calculated from these regions provide some potential predictability for high-frequency SSL seasons. During the past 33 years this combination of drivers has occurred 7 times (Table 2b). All of these years had above average SSL frequency and four of the five highest-frequency SSL seasons are included in these years.

5. Conclusions

A significant result from this work is the development of an objectively determined database of subtropical cyclone occurrence in the Tasman Sea region. The database has been objectively subclassified using clearly defined criteria based on storm system origin. Events occur throughout the year, but are most frequent in late autumn and early winter. A clear seasonal preference is evident in storm system origin; during summer most events evolve from tropical depressions, while during winter most events evolve from extratropical cyclones. Our results show two distinctly different storm types dominate the early winter period. ETLs form in situ along the east coast while SSL evolve out of preexisting extratropical cyclones as they move into the Tasman Sea region. Baroclinic conditions that initiate intensification of both storm types evolve via different processes. Thermal gradients that initiate ETL surface development arise via the interaction of warm subtropical air and a preexisting cold pool over the Australian continent. Thermal gradients that drive SSL intensification develop when a cold pool moves equatorward over the relatively warm Tasman Sea in conjunction with a surface cyclone. The end results are storm systems in the Tasman Sea having similar characteristics, but different origins.

Seasonal-scale climatic variability that may promote increased ETL or SSL frequency has been investigated. Linear relationships with major climate drivers are weak. Specific combinations of remote driver influence are required to produce conditions conducive to each storm type. ETL events require cooling over the Australian continent, subtropical easterly airflow, and convection over northeast Australia. Midlatitude blocking enhances cooling over the Australian continent; this can be driven by Rossby wave trains generated via suppressed convection in the tropical Indian Ocean. Easterlies are strengthened when a subtropical ridge persists over the Tasman Sea. Mesoscale baroclinic convection over northeast Australia is suppressed during the negative phase of the Southern Oscillation; high-frequency ETL seasons do not occur under these conditions.

SSL events are more frequent during seasons when there is anomalous advection of warm moist air into the southern Tasman Sea combined with a more meridional storm track for cold extratropical cyclones. Warm SSTA and enhanced convection in the ITF region increase the probability of SSL events via two related processes: increased frequency of northwest cloud-band-type systems and the generation of an equivalent barotropic Rossby wave train that strengthens blocking south of the Tasman Sea. A strengthening of the west Pacific subtropical ridge and warm SSTA north of New Zealand also serve to increase mean air temperatures in the southern Tasman Sea. The Antarctic climate may also contribute to SSL events; cold extratropical cyclones can form because of cold-air outbreaks from the Antarctic continent, while high-latitude blocking along the East Antarctic coast drives a meridional storm track.

Subjective analyses of individual events show there are many mechanisms for ECC genesis, often combining elements of multiple storm types; the issue of hybrid systems warrants further investigation. The IT and CL storms also require further investigation. The highest storm frequency winter (1998) consisted mostly of IT and CL; suggesting that under different—undetermined—climatic configurations they may become a dominant storm type.

The subclassification of ECC storm types and identification of typical synoptic features associated with storm evolution may be of considerable use to forecasters attempting to anticipate storm development, especially when confronted with divergent model output. A seasonal time scales, index configurations that have previously resulted in high-frequency ETL and SSL winters show promise for improved predictability.

Acknowledgments

This research was funded in part by a research grant to Ian D. Goodwin from the New South Wales Office of Environment and Heritage (NSW OEH) and the New South Wales Environmental Trust. The paper is a contribution to the Eastern Seaboard Climate Change Initiative on East Coast Lows (ESCCI-ECL), and the research project: “Extending the extreme East Coast Low climatology over the past millennium.” Stuart A. Browning was supported by an MQRES grant from Macquarie University and a student grant from NSW OEH. We also thank the anonymous reviewers whose constructive suggestions improved this paper.

REFERENCES

REFERENCES
Allen
,
M.
, and
J.
Callaghan
,
2001
: Extreme wave conditions for the south Queensland coastal region. Environmental Protection Agency Tech. Rep. 32, 28 pp.
Ashcroft
,
L. C.
,
A. B.
Pezza
, and
I.
Simmonds
,
2009
:
Cold events over southern Australia: Synoptic climatology and hemispheric structure
.
J. Climate
,
22
,
6679
6698
.
Bains
,
P.
, and
K.
Fraedrich
,
1989
:
Topographic effects on the mean tropospheric flow patterns around Antarctica
.
J. Atmos. Sci.
,
46
,
3401
3415
.
Bals-Elsholz
,
T. M.
,
E. H.
Atallah
,
L. F.
Bosart
,
T. A.
Wasula
,
M. J.
Cempa
, and
A. R.
Lupo
,
2001
:
The wintertime Southern Hemisphere split jet: Structure, variability, and evolution
.
J. Climate
,
14
,
4191
4215
.
Bridgman
,
H.
,
1985
:
The Sygna storm at Newcastle—12 years later
.
Meteor. Aust.
,
3
,
10
16
.
Bromwich
,
D. H.
,
1991
:
Mesoscale cyclogenesis over the southwestern Ross Sea linked to strong katabatic winds
.
Mon. Wea. Rev.
,
119
,
1736
1753
.
Cai
,
W.
, and
T.
Cowan
,
2008
:
Dynamics of late autumn rainfall reduction over southeastern Australia
.
Geophys. Res. Lett.
,
35
,
L09708
,
doi:10.1029/2008GL033727
.
Cai
,
W.
,
A.
Sullivan
, and
T.
Cowan
,
2009
:
Rainfall teleconnections with Indo-Pacific variability in the WCRP CMIP3 models
.
J. Climate
,
22
,
5046
5071
.
Cai
,
W.
,
P.
van Rensch
, and
T.
Cowan
,
2011a
:
Influence of global-scale variability on the subtropical ridge over southeast Australia
.
J. Climate
,
24
,
6035
6053
.
Cai
,
W.
,
P.
van Rensch
,
T.
Cowan
, and
H. H.
Hendon
,
2011b
:
Teleconnection pathways of ENSO and the IOD and the mechanisms for impacts on Australian rainfall
.
J. Climate
,
24
,
3910
3923
.
Carleton
,
A.
,
1992
:
Synoptic interactions between Antarctica and lower latitudes
.
Aust. Meteor. Mag.
,
40
,
129
147
.
Dare
,
R. A.
, and
J. L.
McBride
,
2011
:
The threshold sea surface temperature condition for tropical cyclogenesis
.
J. Climate
,
24
,
4570
4576
.
Davis
,
C. A.
, and
L. F.
Bosart
,
2003
:
Baroclinically induced tropical cyclogenesis
.
Mon. Wea. Rev.
,
131
,
2730
2747
.
Dowdy
,
A. J.
,
G. A.
Mills
,
B.
Timbal
, and
Y.
Wang
,
2013
:
Changes in the risk of extratropical cyclones in eastern Australia
.
J. Climate
,
26
,
1403
1417
.
Evans
,
J. L.
, and
M. P.
Guishard
,
2009
:
Atlantic subtropical storms. Part I: Criteria and composite analysis
.
Mon. Wea. Rev.
,
137
,
2065
2080
.
Gallego
,
D.
,
P.
Ribera
,
R.
Garcia-Herrera
,
E.
Hernandez
, and
L.
Gimeno
,
2005
:
A new look for the Southern Hemisphere jet stream
.
Climate Dyn.
,
24
,
607
621
,
doi:10.1007/s00382-005-0006-7
.
Garde
,
L. A.
,
A. B.
Pezza
, and
J. A.
T. Bye
,
2010
:
Tropical transition of the 2001 Australian duck
.
Mon. Wea. Rev.
,
138
,
2038
2057
.
Gong
,
D.
, and
S.
Wang
,
1999
:
Definition of Antarctic Oscillation index
.
Geophys. Res. Lett.
,
26
,
459
462
.
Guishard
,
M. P.
,
J. L.
Evans
, and
R. E.
Hart
,
2009
:
Atlantic subtropical storms. Part II: Climatology
.
J. Climate
,
22
,
3574
3594
.
Harper
,
B.
, and
K.
Granger
,
2000
: East Coast low risks. Natural Hazards and the Risks They Pose to South-East Queensland, K. Granger and M. Hayne, Eds., Australian Geological Survey Organisation, 5.1–5.6.
Hart
,
R. E.
,
2003
:
A cyclone phase space derived from thermal wind and thermal asymmetry
.
Mon. Wea. Rev.
,
131
,
585
616
.
Holland
,
G. J.
,
A. H.
Lynch
, and
L. M.
Leslie
,
1987
:
Australian East-Coast cyclones. Part I: Synoptic overview and case study
.
Mon. Wea. Rev.
,
115
,
3024
3036
.
Holton
,
I.
,
1998
:
Prediction of growing season rainfall and crop yields in southern Australia
.
Aust. Meteor. Mag.
,
47
,
327
338
.
Hopkins
,
L. C.
, and
G. J.
Holland
,
1997
:
Australian heavy-rain days and associated East Coast cyclones: 1958–92
.
J. Climate
,
10
,
621
635
.
Jiang
,
N.
,
G.
Griffiths
, and
A.
Lorrey
,
2012
:
Influence of large-scale climate modes on daily synoptic weather types over New Zealand
.
Int. J. Climatol.
,
33
,
499
519
,
doi:10.1002/joc.3443
.
Karoly
,
D.
,
1989
:
Southern Hemisphere circulation features associated with El Niño–Southern Oscillation events
.
J. Climate
,
2
,
1239
1252
.
Katzfey
,
J.
, and
K.
McInnes
,
1996
:
GCM simulations of eastern Australian cutoff lows
.
J. Climate
,
9
,
2337
2355
.
Kidson
,
J. W.
, and
M. R.
Sinclair
,
1995
:
The influence of persistent anomalies on Southern Hemisphere storm tracks
.
J. Climate
,
8
,
1938
1950
.
Lamb
,
H. H.
,
1959
:
The southern westerlies: A preliminary survey; Main characteristics and apparent associations
.
Quart. J. Roy. Meteor. Soc.
,
85
,
1
23
,
doi:10.1002/qj.49708536302
.
Langford
,
S.
, and
H. H.
Hendon
,
2013
:
Improving reliability of coupled model forecasts of Australian seasonal rainfall
.
Mon. Wea. Rev.
,
141
,
728
741
.
Leslie
,
L.
,
G.
Holland
, and
A.
Lynch
,
1987
:
Australian East-Coast cyclones. Part II: Numerical modeling study
.
Mon. Wea. Rev.
,
115
,
3037
3054
.
Madden
,
R.
, and
P.
Julian
,
1994
:
Observations of the 40–50-day tropical oscillation—A review
.
Mon. Wea. Rev.
,
122
,
814
837
.
Massom
,
R. A.
,
M. J.
Pook
,
J. C.
Comiso
,
N.
Adams
,
J.
Turner
,
T.
Lachlan-Cope
, and
T. T.
Gibson
,
2004
:
Precipitation over the interior east Antarctic ice sheet related to midlatitude blocking-high activity
.
J. Climate
,
17
,
1914
1928
.
McInnes
,
K. L.
,
L.
Leslie
, and
J.
McBride
,
1992
:
Numerical simulation of cut-off lows on the Australian East Coast: Sensitivity to sea-surface temperature
.
Int. J. Climatol.
,
12
,
783
795
.
McInnes
,
K. L.
,
G. D.
Hubbert
,
D. J.
Abbs
, and
S. E.
Oliver
,
2002
:
A numerical modelling study of coastal flooding
.
Meteor. Atmos. Phys.
,
80
,
217
233
,
doi:10.1007/s007030200027
.
Mills
,
G. A.
,
R.
Webb
,
N. E.
Davidson
,
J.
Kepert
,
A.
Seed
, and
D.
Abbs
,
2010
: The Pasha Bulker East Coast low of 8 June 2007. CAWCR Tech. Rep. 023, 62 pp. [Available online at http://www.cawcr.gov.au/publications/technicalreports/CTR_023.pdf.]
Mo
,
K. C.
, and
M.
Ghil
,
1987
:
Statistics and dynamics of persistent anomalies
.
J. Atmos. Sci.
,
44
,
877
902
.
Mo
,
K. C.
, and
R.
Higgins
,
1998
:
The Pacific–South American modes and tropical convection during the Southern Hemisphere winter
.
Mon. Wea. Rev.
,
126
,
1581
1596
.
Mo
,
K. C.
, and
J. N.
Paegle
,
2001
:
The Pacific–South American modes and their downstream effects
.
Int. J. Climatol.
,
21
,
1211
1229
,
doi:10.1002/joc.685
.
Mo
,
K. C.
,
J.
Pfaendter
, and
E.
Kalnay
,
1987
:
A GCM study on the maintenance of the June 1982 blocking in the Southern Hemisphere
.
J. Atmos. Sci.
,
44
,
1123
1142
.
Otkin
,
J. A.
, and
J. E.
Martin
,
2004
:
A synoptic climatology of the subtropical kona storm
.
Mon. Wea. Rev.
,
132
,
1502
1517
.
Parish
,
T. R.
,
1992
:
On the interaction between Antarctic katabatic winds and tropospheric motions in the high southern latitudes
.
Aust. Meteor. Mag.
,
40
,
149
167
.
Parish
,
T. R.
, and
D. H.
Bromwich
,
1998
:
A case study of Antarctic katabatic wind interaction with large-scale forcing
.
Mon. Wea. Rev.
,
126
,
199
209
.
Parish
,
T. R.
, and
D. H.
Bromwich
,
2007
:
Reexamination of the near-surface airflow over the Antarctic continent and implications on atmospheric circulations at high southern latitudes
.
Mon. Wea. Rev.
,
135
,
1961
1973
.
Parish
,
T. R.
,
D. H.
Bromwich
, and
R. Y.
Tzeng
,
1994
:
On the role of the Antarctic continent in forcing large-scale circulations in the high southern latitudes
.
J. Atmos. Sci.
,
51
,
3566
3579
.
Philander
,
S.
,
1983
:
El Niño–Southern Oscillation phenomena
.
Nature
,
302
,
295
301
.
Pook
,
M. J.
, and
T.
Gibson
,
1999
:
Atmospheric blocking and storm tracks during SOP-1 of the FROST project
.
Aust. Meteor. Mag.
,
48
,
51
60
.
Pook
,
M. J.
,
J. S.
Risbey
, and
P. C.
Mcintosh
,
2012
:
The synoptic climatology of cool-season rainfall in the central wheatbelt of western Australia
.
Mon. Wea. Rev.
,
140
,
28
43
.
Public Works Department
,
1985
: Elevated coastal levels; Storms affecting N.S.W. coast 1880–1980. Weatherex Meteorological Services and Public Works Department Rep. 85041, 574 pp.
Public Works Department
,
1986
: Elevated coastal levels; Storms affecting N.S.W. coast 1980–1985. Weatherex Meteorological Services Rep. 86026, 5 pp.
Qi
,
L.
,
L.
Leslie
, and
M.
Speer
,
2006
:
Climatology of cyclones over the southwest Pacific: 1992–2001
.
Meteor. Atmos. Phys.
,
91
,
201
209
,
doi:10.1007/s00703-005-0149-4
.
Raible
,
C.
,
P.
Della-Marta
,
C.
Schwierz
,
H.
Wernli
, and
R.
Blender
,
2008
:
Northern Hemisphere extratropical cyclones: A comparison of detection and tracking methods and different reanalyses
.
Mon. Wea. Rev.
,
136
,
880
897
.
Reynolds
,
R. W.
,
N. A.
Rayner
,
T. M.
Smith
,
D. C.
Stokes
, and
W.
Wang
,
2002
:
An improved in situ and satellite SST analysis for climate
.
J. Climate
,
15
,
1609
1625
.
Risbey
,
J. S.
,
M. J.
Pook
,
P. C.
Mcintosh
,
C. C.
Ummenhofer
, and
G.
Meyers
,
2009a
:
Characteristics and variability of synoptic features associated with cool season rainfall in southeastern Australia
.
Int. J. Climatol.
,
29
,
1595
1613
,
doi:10.1002/joc.1775
.
Risbey
,
J. S.
,
M. J.
Pook
,
P. C.
McIntosh
,
M. C.
Wheeler
, and
H. H.
Hendon
,
2009b
:
On the remote drivers of rainfall variability in Australia
.
Mon. Wea. Rev.
,
137
,
3233
3253
.
Saji
,
N.
,
B.
Goswami
,
P.
Vinayachandran
, and
T.
Yamagata
,
1999
:
A dipole mode in the tropical Indian Ocean
.
Nature
,
401
,
360
363
.
Shand
,
T.
,
I. D.
Goodwin
,
M. A.
Mole
,
J. T.
Carley
,
S. A.
Browning
,
I. R.
Coghlan
,
M. D.
Harley
, and
L. W.
Peirson
,
2011
: NSW coastal inundation hazard study: Coastal storms and extreme wave events. Water Research Laboratory Tech. Rep. 2010/16, 45 pp.
Shapiro
,
L. J.
, and
S. B.
Goldenberg
,
1998
:
Atlantic sea surface temperatures and tropical cyclone formation
.
J. Climate
,
11
,
578
590
.
Sinclair
,
M. R.
,
1996
:
A climatology of anticyclones and blocking for the Southern Hemisphere
.
Mon. Wea. Rev.
,
124
,
245
264
.
Sinclair
,
M. R.
,
2002
:
Extratropical transition of southwest Pacific tropical cyclones. Part I: Climatology and mean structure changes
.
Mon. Wea. Rev.
,
130
,
590
609
.
Sinclair
,
M. R.
,
2004
:
Extratropical transition of southwest Pacific tropical cyclones. Part II: Midlatitude circulation characteristics
.
Mon. Wea. Rev.
,
132
,
2145
2168
.
Speer
,
M. S.
,
P.
Wiles
, and
A.
Pepler
,
2009
:
Low pressure systems off the New South Wales coast and associated hazardous weather: Establishment of a database
.
Aust. Meteor. Oceanogr. J.
,
58
,
29
39
.
Speer
,
M. S.
,
L. M.
Leslie
, and
A. O.
Fierro
,
2011
:
Australian East Coast rainfall decline related to large scale climate drivers
.
Climate Dyn.
,
36
,
1419
1429
,
doi:10.1007/s00382-009-0726-1
.
Tapp
,
R. G.
, and
S. L.
Barrell
,
1984
:
The north-west Australian cloud band: Climatology, characteristics and factors associated with development
.
Int. J. Climatol.
,
4
,
411
424
,
doi:10.1002/joc.3370040406
.
Thompson
,
D. W. J.
, and
J. M.
Wallace
,
2000
:
Annular modes in the extratropical circulation. Part I: Month-to-month variability
.
J. Climate
,
13
,
1000
1016
.
Trenberth
,
K. E.
,
1980
:
Planetary waves at 500 mb in the Southern Hemisphere
.
Mon. Wea. Rev.
,
108
,
1378
1389
.
Trenberth
,
K. E.
,
1997
:
The definition of El Niño
.
Bull. Amer. Meteor. Soc.
,
78
,
2771
2777
.
Visbeck
,
M.
, and
A.
Hall
,
2004
:
Reply
.
J. Climate
,
17
,
2255
2258
.
Warner
,
M. D.
,
C. F.
Mass
, and
E. P.
Salathé
Jr.
,
2012
:
Wintertime extreme precipitation events along the Pacific northwest coast: Climatology and synoptic evolution
.
Mon. Wea. Rev.
,
140
,
2021
2043
.