1. Introduction
a. Background
The tropical southwest Pacific Ocean encompasses an area of about 10 million square nautical miles from ~135°E longitude in Australia’s Gulf of Carpentaria eastward to 120°W longitude and from the equator to 25°S. It includes part of the Maritime Continent, eastern Australia, New Zealand, and many small island nations and territories (Fig. 1). In such a largely ocean-dominated region, tropical cyclones (TCs) are a primary cause of natural disasters and account for 76% of the reported disasters in the region from 1950 to 2004 (World Bank 2006). The ability to provide better understanding of TC behavior as well as seasonal climate outlooks prior to the TC season has the potential to benefit small island nations through increasing capacity to prepare for these extreme events, possibly saving lives and helping to mitigate impacts on civil infrastructure. The knowledge of how busy a particular season will be—and, more importantly, where in the basin TCs are most likely to make landfall—may be inferred from statistical relationships constructed from past data.
Simplified map of the southwest Pacific basin, showing major climate features and area of responsibility for RSMCs and tropical cyclone warning centers. The SPCZ position that indicates a region of maximum vertical velocity (omega), SST isotherms, and minimum MSLPs depicting the location of the subtropical high pressure belt are derived from averages for November–April for the 1970–2010 period based on the NOAA–CIRES twentieth-century reanalysis version 2. The longitudinal extent of overlapping Niño sea surface temperature anomaly areas are shown at the top of the diagram (with Niño-3 extending farther east than the border of the figure).
Citation: Journal of Climate 26, 1; 10.1175/JCLI-D-12-00077.1
Research into southwest Pacific basin TCs is very timely and relevant, especially since relatively little research on TCs in this region exists in comparison to other basins (Basher and Zheng 1995; Sinclair 2002; Kuleshov et al. 2008; Terry and Gienko 2010; Diamond et al. 2012). Over the past several years much attention has been focused on the relationship of climate change to the frequency and intensity of TCs in this region (Webster et al. 2005; Alley et al. 2007; Kuleshov et al. 2010; Nott 2011; Peduzzi et al. 2012). The Pacific basin is also one of the most important from a global climate standpoint as it is the setting for the El Niño–Southern Oscillation, which has climate impacts across the planet via significant teleconnections. The small island nations of the SW Pacific are directly influenced by ENSO, which modulates sea surface temperatures (SSTs) in the tropical and subtropical Pacific, and the genesis and primary activity region of TCs in the southwest Pacific (Vincent et al. 2011). ENSO variability also repositions the South Pacific convergence zone (SPCZ) (Widlansky et al. 2010), a major source of rainfall and an incubation region for TCs. Case studies demonstrate that short time-scale motion of the SPCZ helps to control trajectories of tropical cyclones and is a factor in extratropical transition (ETT) into the midlatitudes (Lorrey et al. 2012). Improved assessment of TC variability in relation to ENSO will help improve seasonal TC outlook and associated regional risk management.
b. Prior work and purpose of research
Previous work by Basher and Zheng (1995) used an archive of TCs from the New Zealand Meteorological Service for the 20-yr period from 1970 to 1989 to investigate tropical cyclone distribution relationships to the Southern Oscillation (SO). The study included statistical analyses between TCs and SST, and the Southern Oscillation index (SOI). It found significant and consistent relationships between TCs and SSTs in the SW Pacific basin west of 170°E and a closer connection between TCs and atmospheric circulation east of 170°E.
Sinclair (2002), again using the New Zealand Meteorological Service TC archive, found that about one-third of all storms traversed the 35°S parallel and further noted that southwest Pacific TCs encounter baroclinic westerlies earlier in their existence than in other basins and, as such, begin the EET between 25° and 35°S. Sinclair further found that during La Niña conditions average storm motion was slower and more meridional in nature and ETT was confined to the area west of 170°W. During El Niño conditions, ETT occurred much faster and along a much wider range from 160°E to 130°W.
Kuleshov et al. (2008) used the Australian Bureau of Meteorology TC database from 1970 to 2006 and the Multivariate ENSO Index (MEI) (Wolter 1987) to examine TC activity in the Southern Hemisphere. That work concluded that the focus of cyclogenesis shifts east of 170°E in El Niño years, whereas in La Niña years the area of cyclogenesis shifted west to the area between 140° to 170°E. This was consistent with the findings of both Basher and Zheng (1995) and Callaghan and Power (2011). Furthermore, Kuleshov et al. (2008) found a significant positive trend in the occurrence of major TCs during the period of their study (1982–2006).
Terry and Gienko (2010) analyzed the TC database from the Fiji Meteorological Service (longitudinal extent 160°E to 120°W) from 1970 to 2008 and identified cyclogenesis origins, minimum pressures, durations, and track parameters (azimuth and length) of 291 storms. They found considerable temporal variability of TC parameters and could not find any regular pattern of interannual TC behavior. They found strong and significant relationships among storm longevity, track length, and minimum sea level pressure (MSLP) and between seasonally averaged measures of latitude of cyclone origin and the strength of the SOI and the MEI. However, they found no overall long-term linear trends in the data with the exception of MSLP, which showed a uniformly decreasing trend across the basin.
The recent development of the South Pacific Enhanced Archive for Tropical Cyclones (SPEArTC) dataset (Diamond et al. 2012) provides an opportunity to formulate a more complete climatological study of TCs in the southwest Pacific basin that builds on the work of others (e.g., Sinclair 2002). The 19701 TC season is generally considered to be the first for reliable data owing to the availability of geostationary satellite data between 135°E to 120°W (Holland 1981). The extent of the SPEArTC region is different from the area of coverage for southwest Pacific TCs used by the World Meteorological Organization, defined by the forecast areas of responsibility for the Regional Specialized Meteorological Centers (RSMCs) in Nadi (Fiji) and Wellington (New Zealand). Both RSMCs extend from longitude 160°E to 120°W and have different latitudinal coverage that join at approximately 25°S latitude (Fig. 1; see http://www.wmo.int/pages/prog/www/tcp/Advisories-RSMCs.html). The work of Diamond et al. (2012) sourced historical information (historical synoptic charts containing TC tracks) to outline past TC activity in the SW Pacific basin. One important finding of that study illustrated that many past TCs that affected the southwest Pacific community formed to the west of the present RMSC areas of responsibility (over the Coral Sea and toward the Gulf of Carpentaria).
Prior to the development of SPEArTC, considerable work was done documenting TCs in the southwest Pacific basin (Lourenz 1981; Kerr 1976; Revell 1981; Thompson et al. 1992; Giovannelli 1952; Visher 1925; Maunder 1995; Kuleshov et al. 2008; Terry and Gienko 2010). However, prior work did not result in a consolidated and quality-controlled database of TCs in the region. SPEArTC integrates all previous efforts to identify TC tracks by 1) tying the various TC sources together; 2) digitizing new track information along with post-1969 satellite data, allowing for further statistical and spatial analysis; and 3) making the data more easily available in convenient formats and at a centralized, secure location in line with the work of the International Best Tracks for Climate Stewardship (IBTrACS) (Knapp et al. 2010). IBTrACS was not a reanalysis, but a collation of currently available best-track data from agencies worldwide. The major difference from SPEArTC, as documented in Diamond et al. (2012), was that it employed a season-by-season and storm-by-storm quality control process of track morphologies in order to develop as high-quality a dataset as possible. In the process of constructing SPEArTC, a number of corrections were made, including the addition of new tracks, deletion of erroneous tracks, correction of tracks, etc. For example, in the period from 1970–2010, a total of 34 storms that were in IBTrACS were deleted in the resulting SPEArTC dataset.2 The SPEArTC dataset also was not constrained by the WMO boundary of 160°E as is the case in IBTrACS. As such, SPEArTC reflects a more holistic, quality-controlled, and consolidated dataset that we believe augments this study in a positive manner not replicated before.
Seasonal patterns of TC frequency and distribution across the southwest Pacific for well-coupled, ocean-dominated, atmosphere-dominated, or neutral ENSO typologies are delineated in this study. The well-coupled ocean–atmosphere terminology used here is based on the oceanic and atmospheric times series that make up the coupled ENSO index (CEI, see section 2b). This approach demonstrates the usefulness of highlighting nuances (with important spatial signatures) that are associated with different types of ENSO events, indicates how TCs develop and evolve during the evolution of ENSO events, and illustrates the importance of ENSO variability on southwest Pacific TC behavior. The details of the spatial extent and frequency of TCs and how they undergo ETT is a new addition to SW Pacific climatology information that will help underpin future improvements on TC seasonal forecasts and potential risk for Pacific island countries.
2. Data and methods
a. SPEArTC
The new SPEArTC database is the most comprehensive dataset of TCs in the southwest Pacific basin at present (Diamond et al. 2012). It includes TC tracks external to the independent datasets retained by each RSMC so some differences to prior work are expected.
SPEArTC considers the entire SW Pacific basin as a region from 135°E to 120°W longitude and from the equator to 50°S (or as far south as a storm is tracked). This region encompasses the Cape York Peninsula, Queensland, and the southeastern areas of Australia, New Zealand, and all of the small islands strewn across the SW Pacific waters south of the equator. The SPEArTC domain also includes the SPCZ region, plus the Coral Sea, the Tasman Sea, and the equatorial Pacific Ocean from east of the Maritime Continent to the Niño-3.4 region east of eastern Kiribati (Fig. 1).
b. Coupled ENSO index
ENSO dominates seasonal to interannual climate variability in the Pacific region (Philander 1990) and comprises two dynamically linked components. The atmospheric component (the Southern Oscillation) (Walker and Bliss 1932; Trenberth and Caron 2000) involves variations in the strength of the trade winds measured by the Southern Oscillation index (SOI): the normalized mean sea level pressure difference between Papeetee, French Polynesia, and Darwin, Australia (Fig. 1). The oceanic component involves heat content variability in the equatorial oceanic mixed layer, typified by alternating warm and cool phases known as El Niño and La Niña (Philander 1990). The oceanic ENSO state is usually measured as equatorial SST anomalies in the Niño-3, Niño-4, and Niño-3.4 regions (Trenberth 1997); see Fig. 1 for details.
We have chosen to use the CEI, devised to identify synchronous oceanic (using the Niño-3.4 index) and atmospheric (using the SOI) states that are associated with the ENSO (Gergis and Fowler 2005). The Niño-3.4 region is close to the eastern portion of the region of high TC occurrence, so it might be expected to have more direct local impact than other commonly used ENSO indices, especially those based on Niño-1 or Niño-2 region SSTs. Our use of the CEI to highlight ENSO phase analogs is different from previous work in that it defines ENSO events by examining how coherent atmospheric and oceanic forcing are through an ENSO cycle. Use of this index was applied with the understanding that there are nuances for TC activity related to waxing and waning of major ENSO components and their interplay and that there are significant complexities to each event with progression through the TC season. The CEI integrates monthly SOI and Niño-3.4 values, and a style of event is assigned on whether the 11-month smoothed SOI was equal to or exceeded ±0.2 standard deviations3, and if the 5-month smoothed Niño-3.4 anomaly exceeded ±0.5°C. The CEI reveals several subdivisions of ENSO and at least seven distinct ENSO typologies (phases) including well-coupled events (herein called simply Niño or Niña), ocean-dominated events (3.4-Niño or 3.4-Niña), atmosphere-dominated events (SOI Niño or SOI Niña), and neutral ENSO conditions. [See Table A1, following Gergis and Fowler (2005), for CEI monthly assignments and supporting index data.] For each month, if both ocean and atmospheric indices exceeded the critical thresholds, it was defined as a well-coupled event (Niño or Niña); if only the atmosphere exceeded the threshold it was defined as an atmospheric (an SOI-type) event; and if only the ocean exceeded the threshold it was defined as an oceanic (a Niño-3.4-type) event.
Many previous studies have relied on examining different ENSO phases on a more basic level (e.g., warm event, cold event, neutral) to define ocean and atmospheric circulation anomalies in the Pacific (e.g., Trenberth 1997; Trenberth and Caron 2000; Camargo and Sobel 2005). However, it is well recognized by the climate forecasting community in the southwest Pacific (and elsewhere) that ENSO idiosyncrasies include more than three basic “flavors.” While previous work has done a very good job in defining some of the distinct climate anomalies (rainfall, SSTs, SPCZ position) that are important for the region during El Niño or La Niña events (Sinclair 2002), many more nuances in terms of ENSO typologies are known. A similar approach has recently been used to define warm season rainfall patterns in the southwest Pacific (Lorrey et al. 2012), which has shown some success in defining distinct patterns related to ocean–atmosphere circulation anomalies and improves on the more simple and more widely used classification of only three ENSO states.
c. Temporal categorization of the TC season
Three distinct TC “seasons” were examined for the whole SPEArTC dataset to identify ENSO phase analogs: (a) full austral warm season (November to April); (b) early season (November to January; summer); and (c) late season (February to April; autumn). The two halves of the TC season often see significant differences in terms of activity as a whole in the region and spatial disparities in TC action. For each of the seasons above, the majority of months in the season had to be ascribed as a single ENSO typology to be considered as an analog [e.g., for a six- (three-) month window, at least four (two) of the six (three) months had to fit one of the seven ENSO types]. (See the appendix for details of CEI-based analog seasons and Table A1 for details of tie-breaking and situations when conflicting or nonmajority ENSO conditions existed.) After analog seasons were identified, TC tracks for common types of events were amalgamated and spatial patterns of TC frequency, distribution, and anomalies were generated. The climatology information for the full, early, and late season is found in Fig. 2. One could argue that, since the majority of storms occur during the period from January to March, that period should have been used instead. However, given the application of our work toward eventually supporting seasonal outlooks of TCs in the region, we opted for the full, early, and late season approach described above.
Climatology contour plot indicating the spatial distribution and average number of tropical cyclones coming within 5° of each grid point during the (top) full (November–April), (middle) early (November–January), and (bottom) late (February–April) season in the southwest Pacific basin from 1970–2010 inclusive, based on the SPEArTC dataset.
Citation: Journal of Climate 26, 1; 10.1175/JCLI-D-12-00077.1
d. Supporting tools and data
Supporting reanalysis data were gathered from the National Oceanic and Atmospheric Administration (NOAA)/Cooperative Institute for Research in Environmental Sciences (CIRES) twentieth-century reanalysis (20CR) version 2 (Compo et al. 2006, 2011; Whitaker et al. 2004) to augment the interpretation of TC genesis, ETT, and seasonal spatial anomalies seen for the analogs identified in our CEI-based division of SPEArTC. Diagnostics calculated from the reanalyses [e.g., 500-hPa winds, low-level relative vorticity, etc., Chan and Gray (1982) and Gray (1981), are used to help understand TC genesis, tracks, and ETT behavior.
While the distribution of warm waters (≥26.5°C) is critical for dictating thermocline structure as well as initiating and maintaining cyclogenesis, an ENSO phase space plot (Fig. 6) was also used to highlight ETT traits (bearing and departure longitudes as storms exit the tropics) in the range from 25° to 35°S (Sinclair 2002), largely compiled using a Google Earth technique called Graphical Interpretation of Tracks (GrIT) documented in Diamond et al. (2012).
e. Constraints on analyses
The relatively short-term nature of the high-quality TC time series coupled with the changing of ENSO phases and variability of events means that only three or four analog seasons can be identified for several of the ENSO phases. Nevertheless, the compositing of the TC tracks for these analog seasons allowed at least a preliminary examination of the differences between different variations of ENSO. Sufficient TCs have occurred in the past 40+ years to afford a robust statistical analysis of some spatial traits (ETT trajectory, central low pressure, wind speed, and average longitudinal position during ETT) associated with some ENSO phases, which will help underpin seasonal TC forecast efforts.
As noted earlier, the TC season extends from November to April with the latter half of the season from February to April being the busier period with over 57% of the storms occurring4 during those months. It should also be noted that a number of seasonal ENSO phases are represented by only a small number of cases. Hence, there is increased uncertainty associated with some of the results, as discussed below.
3. Results
The analog seasons for each ENSO phase (Table 1) were used to compile the preponderance of events through the full season (November–April), early season (November–January), and late season (February–April), when each assigned ENSO phase had to account for at least 50% of months in a season. While a theoretical total of 21 ENSO phase cases were possible (full, early, and late season for Niño, Niña, 3.4-Niño, 3.4-Niña, SOI Niño, SOI Niña, and neutral phases), there were no cases of 3.4-Niño conditions occurring during the full and early seasons between 1970 and 2010. In addition, a number of ENSO phases had less than three analog years associated with them (Table 1) and are not reported here.
CEI ENSO phases and analog TC seasons from 1970 to 2010.
a. Tropical cyclone climatology
The average number of tropical storms (including all tropical depressions and tropical cyclones) over the 41-yr period from 1970–2010 is 12.9 per season; of those 2.5 were no greater than tropical depression strength, 6.3 attained either an Australian Bureau of Meteorology (BoM) category 1 or 2 TC status, and 4.1 attained major TC status (Australian BoM categories 3–5; http://www.bom.gov.au/cyclone/about/intensity.shtml). Of the number of major TCs during the climatology period, a total of 20 (0.5 per season) attained the most severe Australian category 5 status with maximum sustained winds of (using a 10-min average) greater than 106 kt (54.5 m s−1).
In general, the latter portion of the TC season is the most active in the SW Pacific (Diamond et al. 2012), and the TC swarm plots in Figs. 3a–c demonstrate this point. During 1970–2010 a total of 169 major TCs occurred out of a total of 532 storms (31.8%). However, it is interesting to note that there were 83 major TCs that developed out of a total of 298 storms (27.9%) during the period 1970–90, while from 1991 to 2010 there were 86 major TCs out of a total of 234 storms (36.7%). The increased proportion of major TCs over the past 20 years is statistically significant (p < 0.005, t test with 19 degrees of freedom). Furthermore, of the 21 category 5 storms that occurred in the region during that time, nearly 85% of them occurred during 1991–2010. The increased proportion of major TCs in a basin with fewer overall TCs is quite consistent with findings by Webster et al. (2005), as well as Maue (2011). For the entire analysis period from 1970 to 2010, the area of greatest TC occurrence is west of the International Date Line5 in the areas around Vanuatu, New Caledonia, and Fiji (Fig. 2 and Table 2). The loci of main activity shift westward from Vanuatu into the Coral Sea late in the season, and there is a more southerly and easterly extent of TC activity as the season progresses.
(a) Contour plot and supporting track data from SPEArTC showing the average spatial distribution of TCs for different ENSO phases during the full tropical cyclone season (November–April). Analogs are drawn from the 1970–2010 period (see Tables 1 and A1 for details on analog selection). Contours (graduated at an interval of 0.5 storms) are seasonal averages based on the available analogs for each ENSO phase. Note: no analog is identified for a N3.4-Niño style event. (b) As in (a) but for the early season (November–January). Note: no analog is identified for a 3.4-Niño style event; also, only one analog season existed for a SOI Niña–style event. (c) As in (a) but for the late season (February–April). Note: only one analog season existed for a 3.4-Niña– and SOI Niña–style event.
Citation: Journal of Climate 26, 1; 10.1175/JCLI-D-12-00077.1
Forty-one year TC frequencies/occurrences by country: All months 1970–2010.
b. TC tracks and anomalies associated with variation in ENSO coupling
Climatology contour plots of TC tracks and swarms (or groupings of TCs) were made for full, early, and late season (Figs. 3a–c) using SPEArTC data. Differences between the climatology (Fig. 2) and the spatial patterns for each of the ENSO phases and seasons (Figs. 3a–c) produced anomaly plots (Figs. 4a–c) used to describe the effects that different ENSO phases have on TC behavior and to identify the number of storms that came within 550 km (a 5° latitude radius) of a country or island group (see Table 2 for individual country TC frequency details).
(a) Anomaly climatological TC contour plot (from 1970 to 2010) superposed on the regional wind field anomaly (×10) showing regions of lower than normal and above normal TC activity in the southwest Pacific during different ENSO phases. Contour intervals (graduated at an interval of 0.2 storms) indicate seasonal averages and (blue: negative, red: positive). (b). As in (a) but for the early season (November–January). Note: no analog is identified for a 3.4-Niño style event and only one analog season existed for a SOI-Niña style event. (c) As in (a) but for the late season (February–April). No analog is identified for a 3.4-Niño style event, and only one analog season existed for a SOI Niña–style event (wind field data are not available from the reanalysis for this analog).
Citation: Journal of Climate 26, 1; 10.1175/JCLI-D-12-00077.1
The general patterns that can be discerned from TC tracks and swarm plots (Figs. 3a–c) for the different ENSO phases echo that of the climatology plots (Fig. 2) and reinforce the finding that the late season is more active than the early season as a whole and that activity is more geographically expansive during February–April. The anomaly plots for the ENSO phases (Figs. 4a–c) are more revealing than the TC track and swarm composites, and they depict a diametric opposition of the patterns related to well-coupled La Niña and El Niño events (CEI-based Niñas and Niños) with increased (decreased) TC numbers in the southwest [northeast (NE)] quadrant during Niñas, and the opposite pattern for Niños. Other characterizations of TC anomalies for the various ENSO phases (based on Figs. 4a–c) are as follows.
Niños: Greater than normal (reduced) activity in the NE (SW) quadrant of the SW Pacific, largely to the east of the date line during both the early and late season. Much greater overall activity during the late season, particularly in French Polynesia.
Niñas: Consistency of greater than normal (reduced) activity in the SW (NE) quadrant of the SW Pacific, largely to the west of the date line during both the early and late season. Near-equal scale of TC anomalies (in terms of TC numbers) relative to climatology for early and late season. Positive anomalies greatest south of New Caledonia and into the North Tasman Sea.
SOI Niños: Muted TC anomalies as a whole, but patches of elevated activity near or to the east of the date line. Reduced activity south of Vanuatu and New Caledonia into the North Tasman Sea.
SOI Niñas: Largely below normal activity for entire basin for the whole season, except near north Queensland and well to the east of New Zealand. Activity in the early (late) season is amplified in a swath aligned to the date line (or just to the east). Some elevated action in the western Tasman Sea during the late season.
3.4-Niños: Typified by late season negative anomalies (December onward), with reduction in TC numbers in the NE part of the SW Pacific and east of the date line north and east of Vanuatu and north of Fiji. Increases in activity are more prominent during the late season and are localized near western Gulf of Carpentaria, the central Tasman Sea, and over and to the east of New Zealand.
3.4-Niñas: Largely reduced anomalies across the SW Pacific, particularly to the east of the date line and over New Zealand and the central Tasman Sea. Early and late season activity is elevated in the Coral Sea region, and TC counts are much higher over the Gulf of Carpentaria and North Queensland during these events.
Neutral: Largest full season positive anomalies are over and to the west of the date line north of 25°S, with negative anomalies occurring in a swath from Vanuatu and New Caledonia and south-southwest to the east of New Zealand. Some positive anomalies occur for southern New Zealand, Queensland, and the Gulf of Carpentaria for the full season. Early season anomalies are strongest for New Zealand and northern Australia, while late season anomalies are strongest for the Coral Sea, Solomon Islands, Vanuatu, New Caledonia, Fiji, Niue, Tonga, and Samoa.
Table 3 details the number and intensity of storms by ENSO phase and portion of the season. TC categories were determined by using the 10-min average maximum sustained winds6 as found in the SPEArTC database; then those wind values were compared to the Australian TC intensity scale. It is clear that the seasons characterized as Niños contain the greatest number of major TCs (Australian categories 3–5).
TC intensity statistics by ENSO phase and by full, early, and late season mode for the period from 1970 to 2010. Storm intensities are based on the Australian TC Intensity Scale (http://www.bom.gov.au/cyclone/about/intensity.shtml). TD indicates tropical depression level [<34 kt (17.5 m s−1)], TC indicates tropical cyclone level [34–63 kt (17.5–32.4 m s−1; Australian categories 1–2)], and major TC level indicates ≥64 kt (32.9 m s−1; Australian categories 3–5).
c. TC genesis and ETT
The main development region or average position of TC genesis and average ETT information (based on the data in the Table A1) are shown in Fig. 5 along with SST (the 29°C isotherm) and the 500-hPa 4 m s−1 wind contour, while surface wind anomalies (×10) are shown in Figs. 4a–c.
Average areal extent of the 29°C isotherm for Niña, neutral, and Niño analog years: dashed 29°C isotherm lines indicate common coverage areas for the extent of CEI phase SSTs. Zonally oriented lines beginning at ~20°S along the Australian east coast indicate position of 4 m s−1 steering winds at the 500-hPa geopotential height. SST and steering wind data are from the NOAA–CIRES twentieth-century reanalysis. (Inset) Google Earth projection of average TC genesis locations and ETT longitudes at 25°S.
Citation: Journal of Climate 26, 1; 10.1175/JCLI-D-12-00077.1
The MDR data depict the average genesis point for TCs across the region (supporting contours for TC genesis as depicted in Table 2). The 41-yr average period from 1970 to 2010 indicates that the average genesis point is at 13.1°S, 168.0°E in the Vanuatu archipelago, one of the most active areas for TC development and related impacts in the SW Pacific basin (Figs. 1 and 2). The longitude of extratropical transition for storms as they travel from 25° to 35°S tends to be east of the date line for most ENSO phases except for well-coupled La Niña events.
Figures 5 and 6 reveal many general details about genesis and ETT of southwest Pacific TCs, but with some new insights about the spatial traits of both TC components that are linked to different ENSO phases. There is a clear demarcation between the average location of TC genesis for different styles of El Niño and La Niña and neutral conditions (encompassing full, early, and late seasons; see Fig. 5 inset). Niño TC genesis tends to be farther north and east in the basin (for the full, early, and late seasons), while for Niñas inception is situated farther south and west relative to neutral ENSO conditions (see Fig. 5). In addition, the warm waters from the western equatorial warm pool that reside between the equator and 15°S are more constricted during La Niña events, and extend much farther to the east during El Niño events relative to neutral conditions. The spatial differences in the latitudinal position of the 4 m s−1 steering winds are clearer east of the date line than west, and shows these winds are more northerly during El Niño than during La Niña. Other observations include the following.
Seasonal averages for longitude (5%, 25%, 50%, 75%, 95%) for ETT (at 25°S) with a rose diagram indicating the average trajectory over the last 10° in the subtropics (25°–35°S) for full, early, and late season TC activity in the southwest Pacific. Insufficient analog data indicate that only one analog season was identified (and averages were not calculated).
Citation: Journal of Climate 26, 1; 10.1175/JCLI-D-12-00077.1
Niños: Average ETT to the east of the date line and this ENSO phase exhibits the greatest areal constraint for average track location and the average transition region through the tropics into the subtropics. Interception of tracks by land is highly probable for Fiji and southern Tonga.
Niñas: Average ETT close to and to the west of the date line. Interception of tracks by land is most likely for Vanuatu.
SOI Niños: Average genesis points north of Vanuatu, more easterly in the early season and more westerly during the late season. Average ETT is within a zone slightly west of Niño-style events, with more easterly slant to trajectories as the TC season progresses. Interception of tracks by land is most likely for Fiji (early season) and central and northern Vanuatu (late season).
SOI Niñas: Genesis flanking northernmost Vanuatu, with tracks making average ETT to the east of the date line. Interception of tracks by land is most likely for Fiji, central and southern Tonga, and northern Vanuatu.
3.4-Niños: Largely a late-season phenomenon but can occur in December, with genesis south of the Solomon Islands and Papua New Guinea. ETT zone east of the date line. Interception of tracks by land is likely for Vanuatu.
3.4-Niñas: Genesis is usually in the north or central Coral Sea region (the farthest west for any ENSO phase), and tracks make ETT at or east of the date line, resulting in lengthy tropical–subtropical tracks. Interception of tracks by land is most likely for central and southern Vanuatu.
Neutral: Genesis on average is situated north of Vanuatu and south of the Solomon Islands, and ETT also covers a broad region south of Fiji eastward to due south of Niue. As a result, neutral tracks have a tendency to cover one of the largest areas on average compared to other ENSO phases, and interception by land is common for Vanuatu, Fiji, and Tonga.
In terms of genesis, an overall greater number of storms develop during seasons with ocean-dominated La Niña events (15.3 per season). However, the greatest number of major TCs (Australian categories 3–5) develop during seasons with ocean-dominated El Niño events (6.0 per season), with the second (4.9 per season) and third (4.1 per season) greatest number of major TCs occurring during well-coupled and atmospheric-dominated El Niño events, respectively.
The trajectories and ETT longitudes for full, early, and late season ENSO phases (Fig. 6) indicate an average ETT largely within ±10° longitude of the date line. El Niño phases (Niño, 3.4-Niño, and SOI Niño) have a more easterly orientation of tracks and the average ETT longitude is at or east of the date line. La Niña phases experience more widely varied ETT longitudes; however, Niña events are characterized by an ETT on average to the west of the date line. From 1970 to 2010, across all phases of the CEI, 54% of storms traversed the area from 25° to 35°S with 70% of those storms crossing at 35°S; this is consistent with the findings by Sinclair (2002).
d. Central pressures
The most accurate and reliable measure of TC intensity is the minimum sea level pressure either estimated from aircraft reconnaissance flight level or obtained via direct observation (Knaff and Zehr 2007). However, as pointed out by Sinclair (2002), the rather sparse observational network make in situ measurements of minimum central pressure rare, so pressure values are most often estimates from subjective interpretation of satellite imagery. Table 4 documents TC pressures for each of the CEI phases from the standpoint of mean central pressure of all storms in those phases, the average minimum central pressure for all storms in a phase, as well as the average central pressure for all storms in a phase as they approach ETT at 25°S.
Pressure characteristics per ENSO phase and by full, early, and late season mode for the period from 1970 to 2010, and associated anomalies. The mean Cp (hPa) is the mean central pressure of TCs, the min Cp (hPa) is the mean minimum central pressure of TCs, and the mean CpETT is the mean central pressure of TCs as they pass over and then south of 25°S. Negative anomaly values are shown in italics.
The greatest numbers of major TCs (during Niño phases), described in section 3b, match up with the greatest negative mean sea level pressure (MSLP) anomalies (e.g., lowest pressures). That trait persists even as storms approach ETT. Interestingly, early season neutral phases have the greatest negative departure from normal pressure at ETT (−6.5 hPa). The lowest minimum central pressures during Niño phases are about 1% lower (with stronger winds) than the climatological normal for the 1970–2010 period, while the Niña phases have MSLPs nearly equal to the climatological normal pressures. A t test analysis of the 1% lower pressures during Niño versus Niña phases is statistically significant (p > 0.025) and is one explanation for a greater frequency of major TCs during Niño phases. This difference may also be related to the general track locations during El Niño events being closer to the equator than in La Niña years, which suggests a reduction of baroclinicity (a more barotropic storm structure), which plays a significant role in maintaining storm strength. The central pressures at ETT have a greater standard deviation (±13 hPa) than those in the tropical phases (±6.7 hPa), and this is probably due to the greater variability in pressures as TCs reach ETT while they lose strength. For instance, Sinclair (1997) documented how central pressures of TCs in the basin can vary without corresponding changes in the circulation about the low, and this coupled with the more estimated pressure data at ETT help explain the greater variability in pressure data at ETT found in SPEArTC.
4. Discussion
Examining the 41-yr period of TC activity from 1970 to 2010 with the various phases of CEI as described by Gergis and Fowler (2005), we have analyzed the occurrence and geographic distribution of TCs via their temporal development within a season (full season, early season, and late season) within each of the seven CEI phases. The use of analogs to generate climatological information about TCs in the region is consistent with their use in generating seasonal climate forecasts (Mullan and Thompson 2006).
While this is not the first attempt to study the behavior of TCs in the southwest Pacific, we have undertaken new analyses with the most comprehensive set of TC data compiled to date. The spatial distribution of TCs categorized by ENSO phase is in line with earlier findings of Basher and Zheng (1995), who demonstrated that during El Niño events the positive anomalous TC activity is in the eastern portion of the basin east of the date line, while during La Niña events the opposite is true. The full season anomaly plots for CEI phases (where sufficient analog cases existed; Figs. 3a and 4a) not only confirm the findings from Basher and Zheng (1995) but illustrate subtle yet statistically significant differences between a fully coupled Niño and Niña, atmospheric- or oceanic-dominated phases, or neutral conditions.
During a fully coupled El Niño, there is a clear region of anomalously positive TC occurrence across a wide swath of the basin from 120°W northwestward to an area of greatest occurrence in the Solomon Islands region. The demarcation between positive and negative TC anomalies is much more clear-cut in that particular case than for the atmospheric-dominated El Niño (SOI Niño) where the positive anomaly regions are more scattered across the basin and less intense.
Some of the more severe storms in the SPEArTC dataset coincide with well-coupled El Niños (Table 3). The westerly steering winds at 500 hPa are typically farther north, and the genesis region is displaced north and east relative to normal during the analog seasons that contain these severe storms (Fig. 5). ETT is also displaced on average east of normal during Niños (Fig. 6), along with the region of warmest water emanating from the western Pacific warm pool being extended well to the east of the date line. The differences seen for the TC genesis region that are related to the different ENSO phases and the coincidental changes in the SPCZ mean location during these events [see Vincent et al. (2011) for general orientations of the SPCZ during opposite ENSO phases] illustrate how the convective zone and warm subequatorial waters combine to provide an incubation region for TCs in the southwest Pacific. The thermodynamic link between the convection zone location (defined by the configuration of the surface wind field) and the underlying SST structure is therefore critical for dictating changes in location of TC activity in the southwest Pacific. Expansion of warm equatorial waters eastward during the “warm” ENSO phase and buildup of warmer-than-normal waters in the eastern equatorial Pacific has the effect of spreading the requisite conditions for TC inception and activity to a wider than normal region flanking the equator. Not only are positive surface temperature anomalies more broadly distributed in the tropical part of the southwest Pacific during well-coupled El Niños, but the eastward expansion of warm waters serve to deepen the thermocline from west to east along the equator coincident with anomalous Sverdrup transport away from the equator. Our findings are consistent with the premise that deepened thermocline depth within the subtropics farther east than normal [see the ENSO sea surface temperature oscillation schematic of Jin (1997) and Meinen and McPhaden (2000)] is a likely candidate in helping to maintain TC activity east of the date line during well-coupled El Niños. Conversely, during La Niñas, the anomalous Sverdrup transport is toward the equator with warmer waters “piled up” in the western Pacific (underpinned by a deeper thermocline). The differences in east–west distribution of warm surface waters would help change the zonal SST gradient, change the zonal wind stress, and help shift the SPCZ (and also TC action) relative to normal. Examination of the TC genesis region and ETT locations for different ENSO phases (Fig. 5) and the slant of the zone encompassing TC action between the time of genesis and ETT reinforces this point.
Placing the comparison of SST spatial characteristics to the ENSO phase TC traits in the context of warm water volume changes at the equator (Meinen and McPhaden 2000) suggests that the auto-oscillatory feedback provided by the oceanic component of ENSO (volumetric exchanges of oceanic surface and mixed layer waters) could be a significant contributor to triggering anomalous TC action in the southwest Pacific. It is therefore reasonable to suggest that the widely varying spatiotemporal characteristics of each ENSO phase style (including equatorial and subtropical SST anomalies), superposed on the annual cycle, could lead to subtle but important stylistic differences in Sverdrup transport anomalies that occur at the equator. We have not tested that hypothesis in this study, but future work should test this linkage and it requires an analysis of the Sverdrup transport in the Niño SST regions paired with the climatological analysis presented in this study. From the standpoint of this hypothesis, it would also be highly likely that follow-on effects in zonal wind changes would affect all elements linked to TC formation (including SPCZ position), transport through the subtropics, and ETT. The development of antecedent conditions emplaced by the “recharge mode” of ENSO (Meinen and McPhaden 2000) may be useful for seasonal climate prediction of TC activity. Vincent et al. (2011) clearly illustrated the importance of the eastern edge of the western Pacific warm pool for dictating SPCZ orientation. A conclusion similar to that of Vincent et al. (2011) is reached by the conjecture provided above, supported by the major details of this study, in that equatorial warm water spatial traits are crucially linked to southwest Pacific TC dynamics. What is suggested here that is new is the stylistic differences of ENSO (including timing and rate of late season collapse) dictate a dynamical response of the regional atmosphere and ocean circulation in the southwest Pacific, which in turn dictates a style of seasonal TC activity.
While the approach of atmosphere–ocean coupling has not yet been investigated in great detail or applied broadly to TC research, it has obvious potential for application to seasonal climate and severe weather prediction. Tying into the recommendation for future work made above, the current monitoring of central and southwest Pacific ocean anomalies by the Tropical Atmosphere Ocean (TAO) and Argo arrays, and a host of past information (historical climate data) about the Niño oceanic zones gathered by ships (Wilkinson et al. 2011), along with new reanalysis products (see Compo et al. 2011), could expand our understanding of pre-1970 TC behavior. This would be done in conjunction with SPEArTC (Diamond et al. 2012), which extends continuously back to 1903 and discontinuously to the 1840/41 season. Continued monitoring from the present TAO and Argo buoy array coverage, and spatial additions to those arrays, are therefore a requisite for providing valuable data that could lead to insights into ocean–atmosphere dynamics relevant for TC prediction linked to ENSO. In addition, maintenance of the current arrays is required for further testing of forecast schemes based on climatological information presented in this study with the ultimate aim of preventing or reducing impacts of TCs in a highly vulnerable region. Finally, to assist studies of rainfall catches in TAO-based capacitance rain gauges (Morrissey et al. 2012), the ability to better link the occurrence of TCs to specific TAO rain events would be important; for instance, underestimates of TAO rain catch totals are subject to problems due to wind biases, which may be substantial, and need to be closely studied. In doing so, the wider climate research community will continue to benefit from improved TC climatologies as well as from the development of modern climate monitoring data gathered in situ and remotely [i.e., TAO and the Tropical Rainfall Measuring Mission (TRMM)] that is further developed using traditional approaches (indices, etc.) and extended reanalysis assimilations (Compo et al. 2006).
5. Summary and conclusions
Building on a new and comprehensive database of tropical cyclones (TCs) in the southwest Pacific, and coupling that with data from twentieth-century reanalysis and the coupled ENSO index (CEI), we have developed the most comprehensive climatology to date of TCs in the region from 1970 to 2010. The CEI has allowed us to better refine the already known relationship between TCs and ENSO by examining the varying degrees of coupling between the ocean and atmosphere during various ENSO phases: this is the first study to employ the CEI-based ENSO approach. We have confirmed correlations among TCs, SSTs, and the SOI to derive ENSO relationships and found significant and consistent linkages, highlighting the interplay of TCs and SSTs in the SW Pacific basin west of 170°E and a closer atmospheric connection to the east of 170°E. We believe that the use of the CEI is an advancement on previous work because it provides at least seven different ENSO typologies to examine. While the dynamical mechanisms for the observed changes are still to be determined, the findings of this work have allowed us to pose hypotheses that indicate sea–air flux changes as well as the importance of Sverdrup transport anomalies in the equatorial Pacific region that need to be tested in future studies. Furthermore, use of the CEI approach adds to the overall body of knowledge regarding TC behavior in the southwest Pacific that could be applied to similar studies for other tropical basins.
Subdividing the SPEArTC data in terms of the degree of ocean–atmosphere coupling is seen as a step forward from the simpler division using only one ENSO index for defining an event, which subsequently undercharacterizes the TC activity that could be associated with such an event. The analysis presented here captures the subtleties of ENSO events traits, how they propagate and collapse through the austral warm season, and how the ocean and atmospheric dynamics and changing base climate state contribute to the generation of extreme weather systems. This new information is relevant to the dynamical understanding of how TCs develop in the SW Pacific and evolve during ENSO events, and further illustrates the importance of ENSO variability on regional circulation and regional TC behavior. Furthermore, details about the spatial extent and frequency of TCs and how they undergo ETT were assessed, which is a new addition to SW Pacific climatology information that will help underpin future improvements on TC seasonal forecasts and potential risk for affected countries.
We summarize and highlight the main findings of the study in the following four main areas of the study: 1) overall climatology, 2) anomalies associated with various degrees of ENSO coupling (following Gergis and Fowler 2005), 3) TC genesis and extratropical transition (ETT), and 4) central pressures.
By extending the arbitrary WMO boundary of 160°E west into the Gulf of Carpentaria, we gain a more realistic picture of TC numbers and intensities for the region as a whole, as documented in Table 3. On average, from 1970 to 2010 the region experiences 13 TCs per season with 6.4 attaining major status. Consistent with previous studies, the latter half of this period not only has experienced a fewer total number of storms, but the proportion of major storms has been shown to be higher, at a statistically significant level.
The CEI-based ENSO phase plots of TCs in Figs. 3 and 4 reveal some consistent patterns that reinforce previously found relationships between ENSO and TCs, but do so in a more refined way. Well-coupled El Niño and La Niña events result in distinct and quite opposing patterns of TC activity. During well-coupled Niño events there is greater than normal activity in the northeast portion of the basin (largely east of the date line) with much greater activity during the late season, particularly as far east as French Polynesia (typically an area of low TC activity). Conversely, the greatest positive anomalies during a fully coupled Niña event lie to the south of New Caledonia and into the north Tasman Sea. During atmospheric-dominated (SOI) events, TC anomalies generally reflect lower activity for either Niño or Niña events, with patches of increased activity south of Vanuatu in SOI Niño events and amplified late season activity along the date line in SOI Niña events. During ocean-dominated (3.4) events, TC anomalies show increased late season activity in the very western and southern portions of the basin in the Gulf of Carpentaria and New Zealand during 3.4-Niño events, while 3.4-Niña phases show increased late season activity farther north in the Coral Sea region. Finally, neutral events demonstrate the largest full season anomalies to the west of the date line with negative anomalies occurring in the area from Vanuatu and New Caledonia south and southwest to the east of New Zealand.
The main development region for TCs in the region is centered on Vanuatu and is depicted in Fig. 5 as it relates to the 500-hPa steering winds and the 26.5° and 29°C isotherms. In terms of genesis, an overall greater number of storms develop during seasons with ocean-dominated La Niña events (15.3 per season). However, the greatest number of major TCs develop during seasons with ocean-dominated El Niño events (6.0 per season), with the second (4.9 per season) and third (4.1 per season) greatest number of major TCs occurring during well-coupled and atmospheric-dominated El Niño events, respectively. Figures 5 and 6 show many of the general details related to cyclogenesis and ETT with well-coupled Niño events (as indicated by the CEI) resulting in cyclogenesis farther north and east in the basin for all seasonal phases, with fully coupled Niña events essentially being opposite of that. The spatial difference in the latitudinal position of the upper-air steering winds is more clearly to the east of the date line during all three phases of El Niño. The trajectories and ETT longitudes for full, early, and late season ENSO phases (Fig. 6) indicate an average ETT within ±10° of the date line, which exhibits an easterly orientation during well-coupled El Niño events. Table 4 documents TC pressures for each of the CEI phases from the standpoint of mean central pressure of all storms in those phases, the average minimum central pressure for all storms in a phase, and the average central pressure for all storms in a phase as they approach ETT at 25°S.
The greatest number of major TCs (occurring during Niño phases) is documented in section 3b and matches up with the greatest negative MSLP anomalies (e.g., lowest pressures). That trait persists even as storms approach ETT. Interestingly, early season neutral phases have the greatest negative departure from normal pressure at ETT (−6.5 hPa). The lowest minimum central pressures during Niño phases are about 1% lower (with stronger winds) than the climatological normal for the 1970–2010 period, while the Niña phases have MSLPs nearly equal to the climatological normal pressures. This difference may also be related to the general track locations during El Niños being closer to the equator than in La Niña years, which suggests that a reduction of baroclinicity (a more barotropic storm structure) plays a significant role in maintaining storm strength.
By demonstrating subtle differences between a fully coupled Niño and Niña compared to atmospheric- or ocean-dominated phases, or neutral conditions, the importance of ENSO activity on regional TC behavior has been better refined, including the spatial extent and frequency of TCs and how they undergo ETT, all of which have enormous implications for producing higher-quality seasonal TC outlooks. As such we believe that the findings of this study will provide a baseline for better characterizing the behavior of TCs in the southwest Pacific, and assist in improving seasonal outlooks and TC impact information for the people of the region.
Acknowledgments
While the research for this paper was conducted by the lead author as part of a Ph.D. course of study at the University of Auckland, it should be noted that he is also affiliated with the NOAA National Climatic Data Center. Thanks are due to Dr. Anthony Fowler at the University of Auckland’s School of Environment for comments on an earlier draft of this study, as well as to the two anonymous reviewers of the paper for their very positive feedback and useful comments. Finally, thanks also go to Dr. Mark Morrissey at the University of Oklahoma and Mr. Derrick Snowden at NOAA’s Integrated Ocean Observing System program, for consultation on some of the statistical work included herein. Financial support for Drs. Lorrey and Renwick came from the New Zealand Ministry of Science and Innovation contract C01X0701 (formerly “Adaptation to Climate Variability and Change”), which currently funds the NIWA core science project “Climate Present and Past.”
APPENDIX
CEI Phases [based on Gergis and Fowler (2005)] Used to Ascribe Full, Early, and Late TC Analog Seasons for the Southwest Pacific
The CEI is an index used to jointly indicate ocean and atmosphere conditions associated with ENSO (Gergis and Fowler 2005). The primary components of the CEI are the Southern Oscillation index—the normalized pressure difference between Tahiti in French Polynesia and Darwin, Australia (Troup 1965)—and the Niño-3.4 (N3.4) index, which is a measure of central-western equatorial Pacific sea surface temperature anomalies. Each of these indices is smoothed [as explained in Gergis and Fowler (2005)], and minimum threshold values indicate whether the atmosphere (SOI) or ocean (N3.4) indicates if an ENSO phase is present.
When both indices indicate the same ENSO state, the ocean–atmosphere system is considered to be well coupled (either a Niño- or Niña-style event). When only the N3.4 index or SOI indicates ENSO is present, a N3.4 Niño/N3.4 Niña–style or SOI Niño/SOI Niña–style of event is suggested, respectively. For more details about the CEI, we refer the reader to Gergis and Fowler (2005). Table A1 contains detailed CEI value data from the 2010 TC season back to the 1970 TC season.
CEI values from the 2010 TC season back to the 1970 TC season.
REFERENCES
Alley, R. B., and Coauthors, 2007: Summary for policymakers. Climate Change 2007: The Physical Science Basis, S. Solomon et al., Eds., Cambridge University Press, 1–18.
Basher, R. E., and X. Zheng, 1995: Tropical cyclones in the southwest Pacific: Spatial patterns and relationships to Southern Oscillation and sea surface temperature. J. Climate, 8, 1249–1260.
Callaghan, J., and S. Power, 2011: Variability and decline in the number of severe tropical cyclones making landfall over eastern Australia since the late nineteenth century. Climate Dyn., 37, 647–662.
Camargo, S. J., and A. H. Sobel, 2005: Western North Pacific tropical cyclone intensity and ENSO. J. Climate, 18, 2996–3006.
Chan, J. C. L., and W. M. Gray, 1982: Tropical cyclone movement and surrounding flow relationships. Mon. Wea. Rev., 110, 1354–1374.
Compo, G. P., J. S. Whitaker, and P. D. Sardeshmukh, 2006: Feasibility of a 100-year reanalysis using only surface pressure data. Bull. Amer. Meteor. Soc., 87, 175–190.
Compo, G. P., and Coauthors, 2011: The Twentieth Century Reanalysis Project. Quart. J. Roy. Meteor. Soc., 137, 1–28.
Diamond, H. J., A. M. Lorrey, K. R. Knapp, and D. H. Levinson, 2012: Development of an enhanced tropical cyclone tracks database for the southwest Pacific from 1840 to 2010. Int. J. Climatol., 32, 2240–2250, doi:10.1002/joc.2412.
Gergis, J., and A. M. Fowler, 2005: Classification of synchronous oceanic and atmospheric El Niño–Southern Oscillation (ENSO) events for palaeoclimate reconstruction. Int. J. Climatol., 25, 1541–1565.
Giovannelli, P. J., 1952: Les Cyclones Tropicaux en Nouvelle-Caledonia, au Cours d’un Siecle (1852–1951). Service de la Meteorologie, Noumea, New Caledonia, 25 pp.
Gray, W. M., 1981: Recent Advances in Tropical Cyclone Research from Rawinsonde Composite Analysis. World Meteorological Organization, 407 pp.
Holland, G. J., 1981: On the quality of the Australian tropical cyclone data base. Aust. Meteor. Mag., 29, 169–181.
Jin, F.-F., 1997: An equatorial ocean recharge paradigm for ENSO. Part I: Conceptual model. J. Atmos. Sci., 54, 811–829.
Kerr, I. S., 1976: Tropical Storms and Hurricanes in the Southwest Pacific, November 1939 to April 1969. New Zealand Ministry of Transport, 114 pp.
Knaff, J. A., and R. M. Zehr, 2007: Reexamination of tropical cyclone wind–pressure relationships. Wea. Forecasting, 22, 71–88.
Knapp, K. R., M. C. Kruk, D. H. Levinson, H. J. Diamond, and C. J. Neumann, 2010: The International Best Track Archive for Climate Stewardship (IBTrACS): Unifying tropical cyclone best track data. Bull. Amer. Meteor. Soc., 91, 363–376.
Können, G. P., P. D. Jones, M. H. Kaltofen, and R. J. Allan, 1998: Pre-1866 extensions of the Southern Oscillation index using early Indonesian and Tahitian meteorological readings. J. Climate, 11, 2325–2339.
Kuleshov, Y., L. Qi, R. Fawcett, and D. Jones, 2008: On tropical cyclone activity in the Southern Hemisphere: Trends and the ENSO connection. Geophys. Res. Lett., 35, L14S08, doi:10.1029/2007GL032983.
Kuleshov, Y., R. Fawcett, L. Qi, B. Trewin, D. Jones, J. McBride, and H. Ramsay, 2010: Trends in tropical cyclones in the south Indian Ocean and the South Pacific Ocean. J. Geophys. Res., 115, D01101, doi:10.1029/2009JD012372.
Lorrey, A. M., G. A. Dalu, J. A. Renwick, H. J. Diamond, and M. Gaetani, 2012: Reconstructing the South Pacific convergence zone position during the presatellite era: A La Niña case study. Mon. Wea. Rev., 140, 3653–3668.
Lourenz, R. S., 1981: Tropical Cyclones in the Australian Region, July 1909 to June 1980. Australian Bureau of Meteorology, 94 pp.
Maue, R., 2011: Recent historically low global tropical cyclone activity. Geophys. Res. Lett.,38, L14803, doi:10.1029/2011GL047711.
Maunder, J., 1995: An historical overview regarding the intensity, tracks, and frequency of tropical cyclones in the South Pacific during the last 100 years, and an analysis of any changes in these factors. WMO/TD-692, Tropical Cyclone Programme Rep. TCP-37, 48 pp.
Meinen, C. S., and M. J. McPhaden, 2000: Observations of warm water volume changes in the equatorial Pacific and their relationship to El Niño and La Niña. J. Climate, 13, 3551–3559.
Morrissey, M. L., H. J. Diamond, M. J. McPhaden, H. P. Freitag, and J. S. Greene, 2012: An investigation of the consistency of TAO-TRITON buoy-mounted capacitance rain gauges. J. Atmos. Oceanic Technol., 29, 834–845.
Mullan, A. B., and C. S. Thompson, 2006: Analogue forecasting of New Zealand climate anomalies. Int. J. Climatol., 26, 485–504.
Nott, J., 2011: Tropical cyclones, global climate change and the role of Quaternary studies. J. Quat. Sci., 26, 468–473.
Peduzzi, P., B. Chatenoux, H. Dao, A. De Bono, C. Herold, J. Kossin, F. Mouton, and O. Nordbeck, 2012: Global trends in tropical cyclone risk. Nature Climate Change, 2, 289–294.
Philander, S. G. H., 1990: El Niño, La Niña, and the Southern Oscillation. Academic Press, 289 pp.
Revell, C. G., 1981: Tropical Cyclones in the Southwest Pacific, November 1969 to April 1979. New Zealand Ministry of Transport, 53 pp.
Sinclair, M. R., 1997: Objective identification of cyclones and their circulation intensity, and climatology. Wea. Forecasting, 12, 595–612.
Sinclair, M. R., 2002: Extratropical transition of southwest Pacific tropical cyclones. Part I: Climatology and mean structure changes. Mon. Wea. Rev., 130, 590–609.
Terry, J. P., and G. Gienko, 2010: Climatological aspects of South Pacific tropical cyclones, based on analysis of the RSMC-Nadi (Fiji) regional archive. Climate Res., 42, 223–233.
Thompson, C. S., S. Ready, and X. Zheng, 1992: Tropical Cyclones in the Southwest Pacific: November 1979 to May 1989. New Zealand Meteorological Service, 35 pp.
Trenberth, K. E., 1997: The definition of El Niño. Bull. Amer. Meteor. Soc., 78, 2771–2777.
Trenberth, K. E., and J. M. Caron, 2000: The Southern Oscillation revisited: Sea level pressures, surface temperatures, and precipitation. J. Climate, 13, 4358–4365.
Troup, A. J., 1965: The Southern Oscillation. Quart. J. Roy. Meteor. Soc., 91, 490–506.
Vincent, E. M., M. Lengaigne, C. E. Menkes, N. C. Jourdain, P. Marchesiello, and G. Madec, 2011: Interannual variability of the South Pacific convergence zone and implications for tropical cyclone genesis. Climate Dyn., 36, 1881–1896.
Visher, S., 1925: Tropical Cyclones of the Pacific. Bernice P. Bishop Museum, Bulletin 20, 163 pp.
Walker, G. T., and E. W. Bliss, 1932: World weather V. Mem. Roy. Meteor. Soc., 4, 53–84.
Webster, P. J., G. J. Holland, J. A. Curry, and H. R. Chang, 2005: Changes in tropical cyclone number, duration, and intensity in a warming environment. Science, 309, 1844–1846.
Whitaker, J. W., G. P. Compo, X. Wei, and T. M. Hamill, 2004: Reanalysis without radiosondes using ensemble data assimilation. Mon. Wea. Rev., 132, 1190–1200.
Widlansky, M. J., P. J. Webster, and C. D. Hoyos, 2010: On the location and orientation of the South Pacific convergence zone. Climate Dyn., 36, 561–578, doi:10.1007/s00382-010-0871-6.
Wilkinson, C., and Coauthors, 2011: Recovery of logbooks and international marine data: The RECLAIM Project. Int. J. Climatol.,31, 968–979.
Wolter, K., 1987: The Southern Oscillation in surface circulation and climate over the tropical Atlantic, eastern Pacific, and Indian Oceans as captured by cluster analysis. J. Climate Appl. Meteor., 26, 540–558.
World Bank, 2006: Not if, but when: Adapting to natural hazards in the Pacific Islands region—A policy note. The World Bank, East Asia and the Pacific Region, 46 pp.
In the Southern Hemisphere, the TC season straddles the change of calendar year; therefore, the 1970 season extends from November 1969 through April 1970, and the season is thus labeled as 1970 here. This will be the seasonal convention utilized throughout this paper.
Work is underway to eventually incorporate the work done in constructing the SPEArTC dataset back into IBTrACS.
SOI is reported in normalized units of standard deviation. (Können et al.1998).
The height of the season is from January to March in which 67% of the storms occur.
The date line has no physical significance as it merely traverses around the antimeridian at 180° longitude, but is still a useful and popularly used reference point for describing TC activity in the region.
Based on a 10-min averaged wind speed.
