Show Summary Details

Page of

date: 26 October 2020

# The Global Climatology of Tropical Cyclones

## Abstract and Keywords

Tropical cyclones, also known as hurricanes or typhoons, are one of the most violent weather phenomena on the planet, posing significant threats to those living near or along coastlines where tropical cyclone–related impacts are most pronounced. About 80 tropical cyclones form annually, a rate that has been remarkably steady over the period of reliable historical record. Roughly two thirds of these storms form in the Northern Hemisphere from about June to November, while the remaining third form in the Southern Hemisphere typically during the months of November to May. Our understanding of the global and regional spatial patterns, the year-to-year variability, and temporal trends of these storms has improved considerably since the advent of meteorological satellites in the 1960s because of advances in both remote-sensing technology and operational analysis procedures. The well-recognized spatial patterns of tropical cyclone formation and tracks were laid out in a series of seminal papers in the late 1960s and 1970s and remain an accurate sketch even to this day. Concerning the year-to-year variability of tropical cyclone frequency, the El Niño Southern Oscillation (ENSO) has by far the most dominant influence across multiple ocean basins, so much so that it is typically used as the main predictor for statistical forecasts of seasonal tropical cyclone activity. ENSO has a modulating influence on atmospheric circulation patterns, even in regions remote to the tropical Pacific, which, in turn, can act to enhance or inhibit tropical cyclone formation.

While the meteorological and climate community has come a long way in our understanding of the global and regional climatological features of tropical cyclones, as well as some aspects of the broader relationship between tropical cyclones and climate, we are still hindered by temporal inconsistencies within the historical record of storm data, particularly pertaining to tropical cyclone intensity. Despite recent efforts to homogenize the historical record using satellite-derived intensity data back to the early 1980s, the relatively short period makes it difficult to discern secular trends due to anthropogenic climate change from natural trends occurring on decadal to multidecadal time scales.

# Introduction

Tropical cyclones (TCs) are among the most powerful and destructive weather systems on earth. At peak intensity, surface winds around the inner core have been observed to exceed 90 m s-1 (333 km h-1; 207 mph) (Kimberlain, Blake, & Cangialosi, 2016). In addition to their violent surface winds, tropical cyclones bring with them additional hazards of torrential rain and powerful storm surges, which individually may pose even greater threats than the wind, especially to those living along the coast (e.g., Rappaport, 2000, 2013; Rappaport & Blanchard, 2015).

There is no one concise and agreed-upon definition of the term “tropical cyclone,” arguably because it has different meanings depending on geographic location. In the meteorological community, a tropical cyclone is the general term for a “cyclone that originates over the tropical oceans” (American Meteorological Society, 2016), or, more technically, a non-frontal synoptic scale low-pressure system over tropical and subtropical waters with organized convection (i.e., thunderstorm activity) and a closed cyclonic low-level wind circulation. Most people, however, know these systems by their regionally specific names. People living in the United States, for instance, refer to tropical cyclones as tropical depressions, tropical storms, or hurricanes, depending on their intensity, whereas people living in Asia use the term “typhoon” to describe a tropical cyclone that has reached hurricane intensity.

# Tropical Cyclone Nomenclature and “Best Track” Data

Knowledge of the global and regional characteristics of tropical cyclone activity, including where and when they form, their intensities, and tracks, requires careful documentation and planned coordination from government agencies around the world. In 1972, the World Meteorological Organization (WMO) established the Tropical Cyclone Project, a predecessor to the Tropical Cyclone Programme (TCP) currently in operation, in response to the disastrous 1970 Bangladesh cyclone. The TCP has established six regional specialized meteorological centers (RSMCs) and six tropical cyclone warning centers (TCWCs) (Fig. 1), which are responsible for providing real-time warnings to the public with information such as a tropical cyclone’s location, size, present and forecast movement, and intensity, as well as archiving best estimates of location and intensity into historical datasets known as “best track” data. The best track data are used widely in studies of the regional and global impacts of climate variability and change on tropical cyclones. The geographical subregions shown in Figure 1 are grouped typically into seven global tropical cyclone basins: (i) North Atlantic (NA, region 1), (ii) Eastern North Pacific (ENP, regions 2 and 3), (iii) Western North Pacific (WNP, region 4), (iv) North Indian (NI, region 5), (v) South Indian (SI, region 6), (vi) Australian (AUS, regions 7, 8, 9, 10, and 11), and (vii) the South Pacific (SP, regions 12 and 13). While it is possible for tropical cyclones to form in the South Atlantic, instances are exceptionally rare (a notable exception was Hurricane Catarina in March 2004; McTaggart-Cowan et al., 2006; Pezza & Simmonds, 2005), and consequently it is not recognized officially as a tropical cyclone basin. Finally, there are a number of other government agencies, such as the Joint Typhoon Warning Center (JTWC) in Hawaii, which do not fall under the organizational structure of the WMO but nevertheless provide valuable real-time advisories and warnings of tropical cyclones and contribute to the collection of best track data.

The International Best Track Archive for Climate Stewardship (IBTrACS; Knapp, Kruk, Levinson, Diamond, & Neumann, 2010) provides a merged global dataset of tropical cyclone frequency, location, and intensity, collected from all the international centers (both WMO and non-WMO affiliated) for public use. The IBTrACS-WMO best track data for the period 1985–2014 are used for the climatology presented herein, with a system being counted as a tropical cyclone if its maximum sustained surface wind (MSSW) speed reaches at least 34 knots (17.5 m s-1) (corresponding to gale-force winds on the Beaufort scale). Most of the WMO-sanctioned agencies use a 10-minute average MSSW speed, following WMO guidelines, but the U.S. National Hurricane Center (NHC) and the India Meteorological Department (IMD) use a 1-minute and 3-minute average, respectively, to calculate the MSSW speed. For the purpose of this review, the NHC and IMD data are converted to a 10-minute average by multiplying by a factor of 0.88—a commonly used although by no means optimal conversion factor whose origins remain somewhat unclear (Harper, Kepert, & Ginger, 2010). The interested reader is referred to Harper et al. (2010) for a detailed discussion on converting between various wind-averaging periods in tropical cyclone conditions.

Figure 1. Map showing the six Regional Specialized Meteorological Centers (RSMCs) and the six Tropical Cyclone Warning Centers (TCWCs) assigned by the World Meteorological Tropical Cyclone Programme.

Public confusion commonly arises from the many different naming conventions used to classify tropical cyclones depending on their intensity, not only between the global basins but also within a particular basin. In the North Atlantic and Eastern North Pacific basin, including the Gulf of Mexico and the Caribbean Sea, tropical cyclones are referred to as tropical depressions, tropical storms, or hurricanes, depending on intensity. Tropical depressions and tropical storms have maximum sustained surface winds of less than and greater than 17 m s-1, respectively, whereas a hurricane is more intense, with maximum sustained surface winds of at least 33 m s-1. In the Western North Pacific basin, tropical cyclones are classified as tropical depressions (<17 m s-1), tropical storms (>17 m s-1), severe tropical storms (>24 m s-1), and typhoons (≥33 m s-1). In the North Indian Ocean, the naming convention is different yet again, with the term cyclonic storm favored over tropical storm, and very severe cyclonic storm used instead of hurricane. In the Southwest Indian Ocean, terms such as moderate tropical storm (>17 m s-1), severe tropical storm (>24 m s-1), tropical cyclone (≥33 m s-1), intense tropical cyclone (>44 m s-1), and very intense tropical cyclone (>58 m s-1) are used. Finally, in the Australian and South Pacific regions, the term tropical cyclone itself is used to describe systems whose MSSW speeds exceed 17 m s-1, whereas a severe tropical cyclone refers to a tropical cyclone with a MSSW speed of at least 33 m s-1.

# Global and Hemispheric Aspects of Tropical Cyclone Formation and Frequency

Each year, around 80 tropical cyclones form around the globe (Fig. 2)—a number that has been remarkably steady since the start of reliable global best track data. The annual global rate of tropical cyclone formation continues to mystify climate scientists (Emanuel & Nolan, 2004); that there are about 80 tropical cyclones, and not 40 or 160, for instance, is a statistic without a satisfactory explanation, and there is no extant theory for what sets the global rate of formation (Walsh et al., 2015). During the 25-year period from 1990 to 2014, when full global data are available in IBTrACS-WMO, an average annual number of 79 tropical cyclones occurred (based on a 10-minute average MSSW speed exceeding 17 m s-1), with a standard deviation of about seven. The global average number of storms varies slightly depending on dataset used and the wind-averaging period (e.g., a 1-minute versus a 10-minute average MSSW speed), but studies typically report somewhere between 80 and 90 tropical cyclones per year (e.g., Frank & Young, 2007; Lander & Guard, 1998; Schreck, Knapp, & Kossin, 2014). The most active years globally in the period 1990–2014 were 1992, 2005, and 2013—when 90 tropical cyclones formed—while the most inactive years were 1999 (65 TCs) and 2010 (69 TCs; Maue, 2011), respectively (Fig. 2).

Figure 2. Time series of global, hemispheric, and regional annual tropical cyclones counts for the period 1985 to 2014. The North Indian basin has a shorter complete period of best track data, beginning in 1990. Note that a tropical cyclone year in the Southern Hemisphere is defined from July 1 to June 30 (e.g., 2004 spans from July 1, 2003, to June 30, 2004). See Schreck et al. (2014) for a comparison of annual storm counts from IBTrACS-WMO and NHC+JTWC over the period 1981–2010.

The Northern Hemisphere is host to a disproportionate number of tropical cyclones, experiencing 70% of the global total, compared to just 30% in the Southern Hemisphere. Almost one third (31%) of all storms originate and track over the warm waters of the Western North Pacific, with the Eastern North Pacific and North Atlantic basins housing 19% and 16%, respectively. The North Indian basin experiences significantly fewer tropical cyclones, accounting for only about 4% of the global total (Fig. 3), despite it being one of the most vulnerable regions of the world in terms of tropical cyclone impacts (Peduzzi et al., 2012).

Figure 3. (Top panel): Figure from Gray (1968) showing the “location points of first detection of disturbances which later became tropical storms.” (©American Meteorological Society. Used with permission.) (Bottom panel): The global distribution of tropical cyclone formation points for the period of most reliable global best track data, 1985–2014. The percentage of tropical cyclones occurring in each basin (relative to the global total) is also shown based on data from 1990 to 2014.

In the Southern Hemisphere, tropical cyclones form in a semi-continuous latitudinal zone, stretching from the east coast of Africa to the central South Pacific Ocean. This long expanse of formation points in the Southern Hemisphere (Fig. 3) makes the dividing lines between individual tropical cyclone basins less obvious compared to the Northern Hemisphere, where land acts often as a natural barrier. Nevertheless, these basin boundary lines serve to separate the areas of responsibility assigned by the WMO to various government agencies, as shown in Figure 1. Of the 30% of global tropical cyclones that develop in the Southern Hemisphere, on average, 11% occur in the South Indian basin, 12% in the Australian region, and 7% in the South Pacific basin (Fig. 3). The fraction of the global total of tropical cyclone activity that each basin represents is insensitive to choice of best track data, with IBTrACS-WMO and NHC+JTWC producing similar statistics (Schreck et al., 2014).

# Seasonality, Variability and Trends in Tropical Cyclone Counts

There was no discernable trend in the global number of tropical cyclones from 1985 to 2014 or from 1990 to 2014 (Fig. 2), despite some studies predicting a decrease in response to increased greenhouse gases (Knutson et al., 2010). The most active months for tropical cyclones globally are August and September, which together account for about one third (31%) of the total annual number. The month of May sees the fewest tropical cyclones, with an average of just three per year, or 4% of the annual global total.

Tropical cyclones in the Northern Hemisphere can form during any month of the year, although most storms (90%) occur from June to November, with a peak in August–September (44%). The quietest period in the Northern Hemisphere is from January to March, which coincides with the most active part of the cyclone season in the Southern Hemisphere. The years 1992 and 2013 were the most active in the Northern Hemisphere for the period 1990–2014, with 67 storms occurring in each of those years compared to the hemispheric average of 55, whereas only 44 storms formed in 1999 and 2010. There was no significant trend in the annual number of tropical cyclones in the Northern Hemisphere from 1985 to 2014 or from 1990 to 2014.

In the Southern Hemisphere, tropical cyclones form typically during the months of November to April (94% of the annual count), with the peak of activity occurring during January–March (66% of the annual count), although at least one storm has formed in every month except for August based on data from 1985 to 2014. The off-season months, from June to October, coincide with the most active period in the Northern Hemisphere. The year 1997 (i.e., July 1996 to June 1997) was the most active in the period 1985–2014, with 33 tropical cyclones forming (9 above the long-term average of 24), whereas less than half that number, 16, formed in 1991. There was a slight downward, albeit insignificant, trend in the total annual number of Southern Hemisphere tropical cyclones from 1985 to 2014—a decrease of roughly one storm per decade.

## North Atlantic Basin

The North Atlantic tropical cyclone season, known as “hurricane season ” in the United States, runs officially from June 1 to November 30, with 97% of storms occurring in that period. In rare instances storms have occurred outside the official season, such as Tropical Storm Ana in April 2003, or Tropical Storm Alberto in May 2012. The peak months in the North Atlantic are August and September (Fig. 4), which see an average of about seven tropical cyclones forming in the basin (i.e., 58% of the annual total count). In 2005, a record-breaking number of storms formed in a single season, totaling 27 tropical cyclones (based on a 10-minute average MSSW speed exceeding 17 m s-1), which surpassed the previous record set during the post-satellite era by a staggering nine tropical cyclones. Fifteen of the 27 storms became hurricanes, including three Category 5 systems (Katrina, Rita, and Wilma), breaking the preceding record for the number of hurricanes in one season (Beven et al., 2008). In contrast, only six tropical cyclones formed in 1986, about half the annual average of 12. There was a significant upward trend in the annual number of tropical cyclones in the North Atlantic from 1985 to 2014, at a rate of about 2.4 storms per decade (Fig. 2), as has been noted in other studies for similar time periods (e.g., Kossin, Camargo, & Sitkowski, 2010; Kossin, Olander, & Knapp, 2013).

Figure 4. Histograms of monthly frequency of tropical cyclone formation count for the globe, each hemisphere, and seven individual basins, based on data from 1985 to 2014. Note that the Southern Hemisphere annual cycle begins in July and ends in June. See Schreck et al. (2014) for a similar analysis based on data from both IBTrACS-WMO and NHC+JTWC.

## Eastern North Pacific Basin

The Eastern North Pacific hurricane season runs from May 15 to November 30, although the majority of storms (93%) form between June and October. The seasonal cycle of tropical cyclone activity is shifted slightly compared to its North Atlantic neighbor, beginning in May and peaking in August (Fig. 4), and it is extremely rare for storms to develop outside the main months of May to November, with just four systems forming out of season between 1985 and 2014. Notably, Tropical Storm Hali in March 1992 was the only known tropical cyclone to form east of the Date Line in the month of March, and earlier that year Hurricane Ekeka formed as a tropical cyclone in late January, becoming the first observed Hurricane in the Central Pacific during the month of January in the post-satellite era. The basin typically has around 15 tropical cyclones annually, but this number has ranged from as low as 8 in 1999 and 2010 to a record-breaking 26 in 1992. The basin boasts the highest density of tropical cyclone activity of anywhere in the world (Figs. 3, 5, 6). There was no discernable trend in the annual number of tropical cyclones from 1985 to 2014.

## Western North Pacific Basin

The Western North Pacific (WNP) basin is rather unique in terms of its tropical cyclone climatology, being the only basin in the world to have had at least one storm form in every calendar month (Fig. 4); however, most storms (95%) form between May and December. The seasonal peak in activity occurs in August and September (45%), with an average of 10 tropical cyclones each year, making it the most active basin for formation anywhere in the world. The basin experiences about the same number annually as all the regions of the Southern Hemisphere combined (~24 tropical cyclones). Even during the off-season months from January to April, tropical cyclones have been known to occur. Typhoon Mitag, for instance, developed in late of February 2002 near the Federal States of Micronesia, subsequently intensifying to become the first Super Typhoon on record for the month of March. The most active year in the period 1985–2014 was 1994, when a total of 32 tropical cyclones formed, including 19 typhoons and 6 super typhoons (Lander & Guard, 1998). In contrast, just 14 tropical cyclones formed during 2010, making it the least active tropical cyclone season on record for the Western North Pacific. There was a significant downward trend in the annual number of tropical cyclones from 1985 to 2014, at a rate of about 2.4 storms per decade, which interestingly is of the same magnitude but opposite sign to the observed increase in the North Atlantic basin over the same period.

## North Indian Basin

Unlike other basins that display unimodal seasonal cycles of tropical cyclone activity, the North Indian basin has two distinct peaks (Fig. 4; Evan & Camargo, 2011; Gray, 1968; Li, Yu, Li, Murty, & Tangang, 2013) associated with the seasonal migration of the South Asian monsoon. The first peak occurs in April to June, accounting for 35% of the average annual total, but most storms develop in the post-monsoon period from October to December (about 60% of the annual average count). The North Indian basin has the lowest frequency of storms relative to the annual global total (Fig. 3), with just 77 tropical cyclones forming in the 25-year period 1990–2014, or roughly 3 per year. The region was the last of the WMO-sanctioned agencies to include routine documentation of wind speed data as part of their best tracks (relatively complete best track data date back to 1990), and thus it is difficult to gauge long-term trends there (Kossin et al., 2013; Landsea et al., 2006; Schreck et al., 2014). The most active years in the basin were 1996, 1998, and 2010, all with five tropical cyclones (i.e., two above the long-term average), whereas only one storm formed in 2001, 2002, and 2011. There was almost no trend in the annual number of tropical cyclones in the period 1990 to 2014.

## South Indian Ocean Basin

The South Indian Ocean basin is the westernmost basin the Southern Hemisphere, extending from the African coast to 90°E. It is home to 11% of the global annual storm count (Fig. 3) and 35% of the Southern Hemisphere count. Storms generally form between November and April (92%), with almost two thirds occurring in January–March (Fig. 4). Although less common, tropical cyclones occasionally develop in the months flanking the main season—October (about one every 5 years) and May (about one every 10 years). Since 1985, the most active seasons were 1994 and 2003, when 13 storms formed. The quietest season, 1999, saw only three tropical cyclones develop. There has been a small positive, albeit insignificant, upward trend in the number of storms forming in the region from 1985 to 2014 (Fig. 2).

## Australian Region

The Australian tropical cyclone region extends from 90°E to 160°E, south of the equator, and is part of a continuum of tropical cyclone activity that extends from the coast of Africa to French Polynesia in the South Pacific (Fig. 3). The official tropical cyclone season is from November 1 to April 30, with 94% of storms forming during this period, but at least one tropical cyclone has formed in every month except for August. It is also not uncommon to have cyclones form in May (there were nine in the 30-year period 1985–2014), despite the last day of the official tropical cyclone season being April 30. The most active months are from January to March, when an average of six storms form in the region. The annual average count in the most recent 30-year period is somewhat lower than if the entire post-satellite period (i.e., 1970 onwards) is considered, owing to some very active years in the 1970s and early 1980s. Nevertheless, for the period 1985–2014 the average annual count was about 9 storms per year, while the observed count for any given year has ranged from as many as 14 systems in 1985, 1986, and 1999, to as few as just 5 in 1988 and 2007. The most active seasons on record were 1974 and 1984 (i.e., 1973–74 and 1983–84), with 19 tropical cyclones in the Australian region (Blair Trewin, personal communication). There was a downward but statistically insignificant trend in the annual number of tropical cyclones between 1985 and 2014.

## South Pacific Basin

The South Pacific basin, which is situated directly to the east of the Australian region (Fig. 3), extends from 160°E to 120°W. Its western boundary at 160°E serves to separate the Australian Bureau of Meteorology’s region of responsibility from that of the Meteorological Services of Fiji and New Zealand (Fig. 1). The South Pacific basin has the lowest rate of tropical cyclone formation in the Southern Hemisphere, accounting for 7% of the average global annual rate (Figs. 2, 3), or roughly six systems per year. The seasonality of storms is very similar to the Australian region, with 95% forming between November 1 and April 30. No tropical cyclones formed between the months of July to September during the period 1985 to 2014. The basin has the largest year-to-year (interannual) variability of tropical cyclone frequency of any basin in the world, with a standard deviation of more than half the annual average rate (μ‎ = 6, σ‎ = 3.5). The most active year was 1998, with a record-breaking 15 tropical cyclones forming in the basin, whereas only one system formed in both 1991 and 2002. A downward trend of about 0.8 storms per decade was present from 1985 to 2014, though this trend was not statistically significant.

# Interannual Variability of Tropical Cyclone Counts

The interannual variability of tropical cyclones, both globally and regionally, continues to be a major area of research in the tropical cyclone community. The annual global number of tropical cyclones has been relatively steady at about 80 or so (μ‎ = 78.3, σ‎ = 6.9) since the start of reliable best track records. The rather small standard deviation of the annual global storm count, compared to the mean, has led to the commonly held view that the annual global count is more stable than might be expected given the large interannual variability present in individual basins; however, Frank and Young (2007) found contradictory evidence of this, demonstrating that the global variability of storm numbers is in fact indistinguishable from that when each basin was examined independently of the others. The interannual variability of tropical cyclone counts on the scale of individual basins is driven primarily by natural modes of variability such as the El Niño Southern Oscillation (ENSO), the Atlantic Meridional Mode (AMM), and the Quasi-Biennial Oscillation (QBO). Of these modes, ENSO has by far the most dominant influence on tropical cyclone variability across multiple basins.

## Australian Region

The first study to establish the feasibility of seasonal tropical cyclone prediction was Nicholls (1979), “A Possible Method for Predicting Seasonal Tropical Cyclone Activity in the Australian Region.” This pioneering study revealed that the year-to-year fluctuations in the number of tropical cyclones in the Australian region were linked to pressure anomalies at Darwin during the preceding winters, which, in turn, were linked to the Southern Oscillation component of ENSO. The relationship was found to be negative, such that anomalously low pressure associated with La Niña events was a precursor to an active season, whereas anomalously high pressure (during El Niño events) was an indicator of below-average tropical cyclone activity. Nicholls (1984) confirmed his earlier results using a longer dataset in addition to two other large-scale predictors related to ENSO: (i) sea surface temperature (SST) in the eastern Equatorial Pacific and (ii) SST in the region immediately to the north of Australia where many tropical cyclones develop. The results further validated Nicholls’s earlier work, indicating that ENSO indices could be used to predict seasonal tropical cyclone activity in the Australian region months before the official November 1 onset. Since Nicholls’s founding work, many other studies have shown strong statistical links between ENSO and Australian tropical cyclone activity, confirming that El Niño (La Niña) years are associated with below (above)-average tropical cyclone counts (e.g., Chand et al., 2013; Evans & Allan, 1992; Goebbert & Leslie, 2010; Hastings, 1990; Liu & Chan, 2012; Ramsay, Camargo, & Kim, 2012; Ramsay, Leslie, Lamb, Richman, & Leplastrier, 2008; Ramsay, Richman, & Leslie, 2014; Solow & Nicholls, 1990; Werner & Holbrook, 2011). The decrease in tropical cyclone numbers during El Niño years in the Australian region has been linked to a geographical shift in activity, whereby an increase in storm frequency in the South Pacific, east of about 170°E, tends to offset the decrease in the Australian region (e.g., Basher & Zheng, 1995; Dowdy et al., 2012; Evans & Allan, 1992; Kuleshov, Qi, & Jones, 2008). Notable exceptions to the statistical relationship between ENSO and tropical cyclone frequency in the Australian region have occurred, for instance, during the La Niña events of 2010/2011 and 2011/2012 when average to below-average tropical cyclone numbers were observed. Moreover, the statistical relationship between ENSO and the annual number of tropical cyclones in the region has weakened notably since about the late 1990s (e.g., Dowdy, 2014; Ramsay, Richman, & Leslie, 2017), although the cause of this is currently unknown.

## South Indian Basin

There is no robust relationship between ENSO and tropical cyclone counts in the South Indian Ocean basin when the entire basin is considered, but many studies have noted shifts in storm activity within the basin between different phases of ENSO. During El Niño events, more storms tend to form in the region west of roughly 75°E, whereas a greater frequency of storms have been observed to occur to the east of 75°E during La Niña episodes (e.g., Ho, Kim, Jeong, Kim, & Chen, 2006; Kuleshov & de Hoedt, 2003; Kuleshov et al., 2008; Ramsay et al., 2012). Ho et al. (2006) attributed the increase in tropical cyclone activity in the western half of the basin during El Niño to cyclonic atmospheric circulation anomalies there, whereas in the eastern part of the basin, anticyclonic circulation anomalies resulted in suppressed tropical cyclone formation rates. Contrary to the studies noted above, Jury et al. (1993) found no statistically significant relationship between the Southern Oscillation Index (SOI) and tropical cyclone numbers in the region west of 75°E. However, a statistical relationship with the QBO was observed such that when the QBO was in its easterly phase there tended to be higher-than-average numbers of tropical cyclones.

## South Pacific Basin

The year-to-year variability of tropical cyclone counts in the South Pacific region is strongly related to ENSO. There is a well-known northeastward shift in tropical cyclone genesis during El Niño such that anomalously high storm counts are observed to the east of about 170°E, affecting the islands of Polynesia including Fiji, Samoa, and the Cook Islands (e.g., Basher & Zheng, 1995; Chand & Walsh, 2009; Diamond, Lorrey, & Renwick, 2013; Dowdy et al., 2012; Ramsay et al., 2012; Revell & Goulter, 1986; Vincent et al., 2011). At the same time, the Australian region (90°E–160°E) experiences relatively few tropical cyclones during El Niño. Several large-scale environmental factors have been implicated to explain the northeast-southwest shift in storm activity between El Niño and La Niña episodes in the South Pacific, including variability in sea surface temperature, vertical wind shear (Dowdy et al., 2012), and lower tropospheric relative vorticity (Camargo, Emanuel, & Sobel, 2007). The interested reader is referred to Dowdy et al. (2012) for further discussion on large-scale environmental influences in the region.

## North Indian Basin

Relatively few storms form in the North Indian Ocean compared to other Northern Hemisphere basins, with a seasonal average of three, partly due to the constraint posed by the Asian Continent extending into the latitudinal zone for tropical cyclone formation. The region has a unique seasonal cycle; when the Western North Pacific, Eastern North Pacific, and North Atlantic basins are close to their climatological peaks in storm activity during July–September, the North Indian Ocean basin experiences almost no tropical cyclones (only two formed during July–September in the 25-year period 1990–2014). The first peak in activity during May and June occurs shortly after the spring “predictability barrier” of ENSO, while the second seasonal peak in October to December is much closer to the canonical peak of ENSO. Both peaks are associated with the migration of the seasonal monsoon trough (e.g., Evan & Camargo, 2011; Gray, 1968; Li et al., 2013). Singh, Khan, and Rahman (2000) found modest, albeit significant, correlations between the SOI and tropical cyclone formation rates in the Bay of Bengal, with a greater frequency of events when the SOI was positive, but ENSO was found to have almost no impact on storm frequency in the Arabian Sea. Felton, Subrahmanyam, and Murty (2013) reported a statistically significant negative correlation between Niño 3.4 sea surface temperature and storm activity in the Bay of Bengal, which is consistent with the positive SOI relationship found by Singh et al. (2000).

## Western North Pacific Basin

ENSO has a large impact on the interannual variability of tropical cyclones in the Western North Pacific basin. Mirroring the South Pacific, there is shift toward the southeast (northwest) during El Niño (La Niña events). Chan (1985) was the first to point out this shift, finding that storm activity east of 150°E tended to be above normal during El Niño years; however, during the following year (after the El Niño event had lapsed) there was a basin-wide decrease in tropical cyclone activity. Several other studies have since confirmed the intra-basin shift in storm activity depending on ENSO phase (e.g., Chen, Weng, Yamazaki, & Kiehne, 1998; Chia & Ropelewski, 2002; Wang & Chan, 2002). The eastward migration of the main storm-formation region in El Niño years makes it possible for storms to track longer distances over the warm waters of the basin, and therefore they tend to be more intense than other years—particularly compared to La Niña years (e.g., Camargo & Sobel, 2005; Chan, 2007; Wang & Chan, 2002). Several atmospheric circulation anomalies have been proposed to explain the observed relationship between ENSO and tropical cyclone activity in the Western North Pacific, details of which can be found in review articles on the topic by Chan (2005) and Camargo, Sobel, Barnston, and Klotzbach (2010).

The QBO also has been suggested as a mode of variability that exerts some control on storm frequency in the region. Zhang, Zhang, and Wei (1994) found that when the QBO was in its westerly phase, enhanced tropical cyclone activity occurred. Chan (1995) confirmed the QBO influence, suggesting that increased tropical cyclone activity occurred when the lower-stratospheric winds were strengthening from the west. However, Camargo and Sobel (2010) found no significant correlations between QBO phase and the number of tropical cyclones in the basin for the period 1953–2008. They also questioned the physical arguments used to explain the modulation of storm frequency by the QBO, finally concluding that QBO exerts no significant influence on tropical cyclones in the Western North Pacific or elsewhere.

## Eastern North Pacific

The connection between large-scale climate drivers and the interannual variability of tropical cyclones in the Eastern North Pacific has remained somewhat elusive (e.g., Whitney & Hobgood, 1997), posing a challenge to seasonal forecasts in the region. Furthermore, the region has received relatively less attention in the scientific literature than its North Atlantic and Western North Pacific neighbors. Nevertheless, ENSO has been shown to exert some influence on the seasonal variability of storms in the region, in terms of both geographic location and intensity. Irwin and Davis (1999) found that during strong El Niño events, the formation region and associated storm tracks were displaced roughly 6° further west than the long-term mean longitude. This is consistent with the finding that El Niño supports a greater frequency of tropical cyclones in the central North Pacific (140°W to the Date Line), posing a greater threat to the Hawaiian Islands (Chu, 2004; Chu & Wang, 1997). Several other studies have confirmed this westward shift in genesis during El Niño years (e.g., Camargo, Robertson, Barnston, & Ghil, 2008; Wu & Chu, 2007). Recently, Jien, Gough, and Butler (2015) found a statistically significant difference in the frequency and intensity of storms in the western portion of the main development region (112°W to 140°W) depending on ENSO phase, with more frequent and more intense events occurring during El Niño years.

El Niño has been shown to have a direct influence on the intensity of storms in the Eastern North Pacific. Gray and Sheaffer (1991) found that intense tropical cyclones (i.e., wind speeds greater than 50 ms-1) were twice as likely to occur in El Niño years than La Niña years. In addition to the heat provided by the ocean surface, subsurface heating plays a critical role in tropical cyclone intensity and intensification by limiting the amount of cold upwelling induced by strong surface winds (e.g., Lin et al., 2013; Vincent, Emanuel, Lengaigne, Vialard, & Madec, 2014). While there exists a considerable lag between the canonical peak of ENSO during Boreal winter and peak storm season over the Eastern North Pacific (i.e., July-September), as well as a latitudinal mismatch between equatorial warming and the main storm development region, there is evidence to suggest that El Niño plays an important role in stimulating high-intensity storms by discharging subsurface ocean heat anomalies to the region some two to three seasons after the seasonal peak of ENSO (i.e., during Boreal winter) (Jin, Boucharel, & Lin, 2014).

## North Atlantic

Internannual variability of tropical cyclones in the North Atlantic basin has a strong dependence on ENSO. Gray (1984) was the first to show that an El Niño event resulted in reduced storm activity over the western Atlantic for the season immediately following the onset of the El Niño. The reduction of storm activity was attributed to enhanced upper-level westerly winds. Indeed, during an El Niño year, vertical wind shear tends to be above average over a large part of the tropical Atlantic resulting in suppressed tropical cyclone formation. On the other hand, La Niña years are associated with reduced vertical wind shear in the region, which increases the likelihood of tropical cyclone formation (e.g., Goldenberg & Shapiro, 1996; Knaff, 1997). ENSO influences other environmental factors too, including increased moist static stability during El Niño years owing to increased upper-tropospheric temperatures in the tropics (Tang & Neelin, 2004). Aside from the variability in seasonal storm counts, ENSO modulates several other tropical cyclone metrics in the North Atlantic, including the number of hurricane days, the number of intense hurricanes (Gray, Landsea, Mielke, & Berry, 1993; Landsea, Pielke, Mestas-Nuñez, & Knaff, 1999), and the accumulated cyclone energy (ACE) (see Camargo et al., 2010, for a review).

A significant statistical relationship existed between the QBO and Atlantic tropical cyclone activity for a period of the historical record. Gray (1984) originally proposed that basin-wide hurricane activity was 50–100% higher in the westerly phase of the QBO compared to its easterly phase. Several other studies have since implicated the QBO as a factor for modulating seasonal activity in the region, either for a particular season or a period of the historical record (see Camargo et al., 2010, and Camargo & Sobel, 2010, for a list of relevant studies). The relationship was statistically significant from about 1950 to 1983, but not after that time, raising questions about the physical links underlying the strong statistical relationship in the earlier period (Camargo & Sobel, 2010). The QBO was removed as one of the predictors for seasonal hurricane activity by the Tropical Meteorology Project Group at Colorado State University (CSU) in 2007.

# Large-Scale Climatological Influences on Tropical Cyclone Formation and Tracks

It is clear from Figures 3, 5, and 6 that tropical cyclones tend to form in preferred regions of the globe. For instance, they rarely form poleward of 30°S in the Southern Hemisphere and poleward of 40°N in the Northern Hemisphere, and there is a dearth of formation points in the eastern South Pacific and South Atlantic basins. It is also apparent that storms only develop at some minimum distance away from the Equator.

Figure 5. (Top panel): figure from Gray (1968, p. 671) showing the “designation of various tropical storm development regions and percentage of tropical storms occurring in each region relative to the global total. Numbers in parentheses are those of the average number of tropical storms occurring in each region per year. The 26.5°C isotherm for August in the Northern Hemisphere and January in the Southern Hemisphere is also shown.” (©American Meteorological Society. Used with permission.) (Bottom panel): Tropical cyclone origin points for the period 1985–2014 from IBTrACS-WMO overlaid on climatological sea surface temperature data for August in the Northern Hemisphere and January in the Southern Hemisphere. The 26.5°C isotherm is outlined in blue.

Figure 6. The total number of tropical cyclones forming per 2.5° × 2.5° latitude-longitude bin in the Northern Hemisphere (top panel) and Southern Hemisphere (bottom panel) for the period 1985–2014 based on data from IBTrACS-WMO. Gray streamlines illustrate the typical surface wind patterns for August in the Northern Hemisphere and January in the Southern Hemisphere, following Gray (1968). The thick black dashed lines mark the positions of the climatological intertropical converegence zones (ITCZ) and monsoon troughs.

Palmen (1948) was the first to observe that tropical cyclones require a minimum threshold of sea surface temperature for formation and subsequent intensification. He stated that “hurricanes can be formed only in the oceanic regions outside the vicinity of the Equator where the surface water has a temperature above 26–27°C,” which is now commonly referred to as the “26.5°C (80°F) threshold,” and is pervasive in the tropical cyclone literature as well as many meteorological textbooks. In 1968, William (“Bill”) Gray published his seminal paper, “Global View of the Origin of Tropical Disturbances and Storms” (Gray, 1968), which highlighted that the global mean climatology of tropical cyclones was linked to various large-scale features of the atmosphere and ocean. These features included the Equatorial Trough (and associated Intertropical Convergence Zone and Monsoon Trough) (Fig. 6), instability of the lower half of the troposphere, a minimum sea surface temperature threshold of 26.5°C, and the vertical shear of zonal wind between 850 hPa and 200 hPa. Once formed, tropical cyclones are “steered ” by predominant atmospheric flow patterns—notably by the easterly winds on the equatorward side of the subtropical highs (Fig. 6) such that they move generally from east to west during their early stages of life before drifting poleward and possibly “recurving ” with the midlatitude westerlies (Fig. 7). An exception to this track pattern occurs in the South Pacific basin, where storms more often than not move in a southeasterly direction (e.g., Chand et al., 2009; Knapp et al., 2010; Ramsay et al., 2012). Figure 3 compares Gray’s original 1968 tropical cyclone climatology with modern-day data. The general agreement between the global patterns of storm formation is fairly remarkable given that routine visible satellite pictures were not available until 1966. The satellite-detected storm positions in the best track data since about 1970 have filled in a few gaps (Fig. 3), particularly in the central Pacific region, but even so Gray’s original work is a largely accurate depiction of global tropical cyclone formation patterns. In terms of the relative frequency of storms in each basin, the numbers have changed somewhat since 1968, but Gray himself acknowledged that “with acquisition of more satellite and other conventional information, these percentages may have to be changed somewhat” (Gray, 1968, p. 671). In hindsight, we can see that Gray (1968) overestimated the frequencies in the North Indian and Western North Pacific basins, while the formation rates in the Eastern North Pacific and North Atlantic regions were somewhat underestimated. Similarly, in the Southern Hemisphere, analyses of modern best track data indicate that Gray (1968) underestimated tropical cyclone occurrence in the Australian/South Pacific region by about 50%, consistent with the findings of Schreck et al. (2014).

Figure 7. Global tropical cyclone tracks for the period 1990–2010, color-coded by intensity on the Saffir-Simpson Hurricane Wind Scale. Tracks are from the IBTrACS-ALL best track dataset (i.e., both WMO and non-WMO data), and SSHWS categories are based on a 1-minute average wind speed.

Image provided by the IBTrACS Team at NOAA/NCEI (Knap et al., 2010).

In 1979, Gray updated and refined his original 1968 paper using 20 years of storm data from 1958 to 1977 (Gray, 1979). The annual average global and hemispheric tropical cyclone counts have remained almost constant between the periods 1958–1977 and 1990–2014, despite advances in satellite-based tracking techniques in the latter period. The global total average number of storms was 79.1 in 1958–1977, compared to 78.5 in 1990–2014. In the Northern and Southern hemispheres, the statistics for the same two periods are 54.6 (1958–1977) and 54.9 (1990–2014), and 24.5 (1958–1977) and 23.7 (1990–2014), respectively.

## Seasonal Genesis Parameters

Arguably the most significant aspect of Gray (1979) was the introduction of six necessary, large-scale conditions for tropical cyclone genesis. According to Gray (1979, pp. 167–168), “it appears that seasonal tropical cyclone frequency can be directly related on a climatological or seasonal basis to a combination of six physical parameters which will henceforth be referred to as primary climatological genesis parameters. These parameters are:

1. 1) low-level relative vorticity (ζ‎r),

2. 2) Coriolis parameter (f),

3. 3) the inverse of the vertical shear, Sz, of the horizontal wind between the lower and upper troposphere (1/Sz),

4. 4) “ocean thermal energy”—sea temperature excess above 26°C to a depth of 60 m (E),

5. 5) vertical gradient of θ‎e between the surface and 500 mb (∂θ‎e/∂p),

6. 6) midde tropospheric relative humidity (RH).”

Gray hypothesized that tropical cyclone formation would be most favored where the product of the six parameters above are maximized, in terms of both geographic location and season. To that end, he posed a seasonal genesis parameter (s.g.p.), which combined the three “dynamic potential” parameters (1)–(3) with the three “thermodynamic potential” parameters (4)–(6). That is,

$Display mathematics$

The interested reader is referred to Gray (1979) for further details on the physical relevance of the parameters, as well as their empirically derived thresholds, magnitudes, and units.

In an effort to further understand the relationship between tropical cyclone formation and the global climate system, Emanuel and Nolan (2004) refined Gray’s seasonal genesis parameter and developed the first Genesis Potential Index (GPI) based on modern day reanalysis data. The Emanuel and Nolan GPI, hereafter E-GPI, is defined by the equation

$Display mathematics$

where η‎ is the absolute vorticity in s-1, Η‎ is the relative humidity at 700 hPa in percent, PI is the potential intensity in ms-1, and Vshear is the magnitude of the vector wind difference between 850 and 200 hPa, in ms-1. The major difference between Gray’s original seasonal genesis parameter and E-GPI is the utilization of potential intensity (PI) as the major thermodynamic control, rather than a threshold for sea surface temperature. Potential intensity, also referred to as maximum potential intensity (MPI), is a theoretical upper limit of the intensity of a tropical cyclone given sea surface temperature and environmental profiles of atmospheric temperature and humidity (Bister & Emanuel, 1998; Emanuel, 1986; Emanuel & Rotunno, 2011). The 26.5°C sea surface temperature threshold bounds the global tropical cyclone formation regions reasonably well in the current climate (Fig. 5), even on relatively short time scales (Dare & McBride, 2011), but theoretical and modeling studies suggest that this threshold will increase as the planet warms (e.g., Royer, Chauvin, Timbal, Araspin, & Grimal, 1998). PI, on the other hand, although often incorrectly viewed as a proxy for sea surface temperature, is a more robust thermodynamic parameter because it is sensitive to both local and global sea surface temperature changes (e.g., Ramsay & Sobel, 2011; Vecchi & Soden, 2007). The E-GPI has been shown to replicate some of the observed climatological characteristics of tropical cyclones, such as the seasonal cycles of individual basins (e.g., Camargo et al., 2007). Emanuel (2010) revised the E-GPI, replacing the relative humidity term with a variable associated with the saturation deficit of the mid-troposphere (χ‎). The saturation deficit in the mid-troposphere is expected to increase with global warming, assuming a constant relative humidity, which has implications for predicting how the rate of tropical cyclone formation will respond to such warming.

In 2011, Tippett, Camargo, and Sobel (2011) proposed an alternative tropical cyclone genesis potential index (hereafter T-GPI) based on a Poisson regression model, in which the monthly number of storms in a given region was regressed against several monthly-averaged climate variables. Four important predictors emerged from the regression: (i) absolute voriticity at 850-hPa, (ii) 600-hPa relative humidity, (iii) relative SST, and (iv) the vertical wind shear between 850 and 200 hPa. The “relative SST,” which is the difference between the local SST and the tropical-mean SST, serves as a proxy for potential intensity (e.g., Ramsay & Sobel, 2011; Swanson, 2008; Vecchi & Soden, 2007). In developing the T-GPI, Tippett et al. (2011) found that the absolute vorticity term is important for tropical cyclone genesis at monthly time scales only up to a certain threshold (4 × 10−5 s-1), beyond which other factors, such as thermodynamics or vertical shear, may be the primary rate-limiting influences.

The seasonal genesis indices described above are just a few of a number of tropical cyclone genesis indices developed for regional and global studies, inspired by Gray’s original seasonal genesis parameter (s.g.p; Gray, 1979). Other genesis indices include the Yearly Genesis Parameter (YGP; Royer et al., 1998), the Genesis Parameter (GP; DeMaria, Knaff, & Conell, 2001), and the Cyclone Genesis Index (CGI; Bruyere, Holland & Towler, 2012). The interested reader is referred to Menkes et al. (2012) for a review of seasonal genesis indices.

# Tropical Cyclone Intensity

Tropical cyclone intensity estimates are an important component of the historical best track record and have significant implications for our understanding of the response of tropical cyclones to climate variability and change. Operational forecasting agencies rely heavily on satellite surveillance to locate storms and infer their intensities, particularly for regions where in situ measurements (from aircraft for example) are not available—which is everywhere but the Atlantic currently. The gold standard for inferring tropical cyclone intensity from satellite data is a subjective pattern recognition technique developed by NOAA scientist Vern Dvorak (Dvorak, 1975), known as the “Dvorak technique.” Meteorologists apply the technique in real time, based on available satellite imagery, and at the end of each season the intensity estimates are reassessed before being incorporated into the best track data. Despite efforts to merge the regional best track data into a single global archive (Knapp et al., 2010), there exists inherent inhomogeneities in the data due to (i) different wind-averaging periods used between agencies, (ii) changes in observational technology and analysis procedures over time, and (iii) the subjective nature of the Dvorak technique itself.

## Satellite Observations of Tropical Cyclones and the Dvorak Technique

The first experimental satellite images of tropical cyclones came available in 1960 after the launch of the polar-orbiting weather satellite, TIROS-1. These images gave meteorologists the first snapshots of cloud patterns from above the earth’s surface. In 1961, Hurricane Esther, the fifth storm of the season in the North Atlantic, became the first tropical cyclone to be discovered by weather satellite. By 1966, the first global satellite monitoring system had begun, providing meteorologists with thousands of visible and infrared images of cloud formations, including tropical cyclones. A major advancement came in the 1970s with the introduction of geostationary satellites, which had the capability to continuously monitor the same area of the globe and provide tropical cyclone forecasters with an unprecedented regularity of cloud images. Three-hourly geostationary visible and infrared images became routinely available in 1978 (Harper, Stroud, McCormack, & West, 2008; Kossin et al., 2013). The 1980s and 1990s fostered even further development of satellite data, including sea surface temperature measurements and cloud drift winds (Velden & Hawkins, 2010).

As satellite technology was rapidly advancing through the 1970s and 1980s, NOAA scientist Vern Dvorak was at the same time working on a technique to estimate tropical cyclone intensities based on satellite-derived cloud patterns. The so-called Dvorak Technique was first summarized in a paper by Dvorak himself in 1975 (Dvorak, 1975) and since then has become one of the most famous and widely applied techniques in meteorology. It has undergone several modifications since its first inception, particularly during the 1980s (Dvorak, 1984) with the implementation of the Enhanced Infra-Red (EIR) Dvorak Technique and the 2000s (Velden et al., 2006) with the development of the Advanced Dvorak Technique (ADT); still, the core principles of the technique remain to this day. In practice, the operational meteorologist follows a set of cloud pattern recognition rules to determine a “T-number,” which is subsequently converted to a CI (Current Intensity) number, before finally being translated into a maximum sustained surface wind speed (see Velden et al., 2006, for an example of CI number to MSSW speed conversions in the Atlantic and Western North Pacific basins). As Velden et al. (2006) point out, a remarkable aspect of the technique is its absolute accuracy (50% of the MSSW speed estimates are within 5 knots of in situ reconnaissance aircraft measurements) combined with its internal consistency. The interested reader is referred to the review article by Velden et al. (2006) for a detailed history and discussion of the Dvorak Technique.

## Climatological Aspects of Tropical Cyclone Lifetime Maximum Intensity

The global distribution of tropical cyclone lifetime maximum intensity (LMI) locations for the period 1985–2014, based on IBTrACS-WMO data, is shown in Figure 8. The LMI is defined as the first point at which a tropical cyclone reaches its maximum MSSW speed over the course of its life. It is apparent from Figure 8 that the most intense storms, those of category 4 and 5 on the Saffir-Simpson hurricane wind scale (SSHWS), reach their LMI at latitudes further equatorward than storms of lesser intensity. In the Northern Hemisphere, the mean latitudes of LMI for Category 5 and Category 4 systems are 17.2°N and 17.7°N, respectively, compared to a mean latitude of 20.4°N when all systems are taken into account. In the Southern Hemisphere, the mean latitude of LMI for Category 5 storms is 13.8°S, compared to a latitude of 17.4°S for all storms. The mean latitude of LMI has been shown to vary from basin to basin (Kossin, Emanuel, & Vecchi, 2014), with storms in the North Atlantic basin reaching their LMI at considerably higher latitudes compared to other tropical cyclone basins. Globally, 42% of tropical cyclones remain at or below tropical storm intensity (30 kt < LMI < 56 kt) on the SSHWS, with the remaining 58% reaching hurricane intensity (i.e., LMI >= 56 kt) (the majority of those being Category 1 storms). Of the 1,373 systems that reached hurricane intensity between 1985 and 2014, 458 were Category 1, (19%), 269 were Category 2 (11%), 297 were Category 3 (13%), 308 were Category 4 (13%), and 41 were Category 5 (2%). The hemispheric distributions of LMI are broadly similar to the total global distribution (Fig. 9), but there exists considerable basin-to-basin variability. An anomaly that stands out in Figures 8 and 9 is the lack of high-intensity tropical cyclones in the Western North Pacific, particularly Category 5 storms. The region is home to the highest annual number of tropical cyclones in the world as well as to the Western Pacific warm pool (Fig. 5), so it stands to reason that it should support a greater number of extreme tropical cyclones than is apparent in Figures 8 and 9. The discrepancy arises primarily from interagency differences in the best track data. The Japanese Meteorological Agency (JMA), which provides data to the WMO version of IBTrACS, documents systematically low LMIs compared to the Joint Typhoon Warning Center (JTWC). For example, Typhoon Haiyan in November of 2013 (Fig. 10) reached a maximum intensity of 125 knots (10-minute average) according to the JMA (Table 1), whereas the JTWC estimated its peak intensity to be 170 knots (1-minute average; equivalent to 10-minute average 150 knots). Such wind speed discrepancies between the JMA and the JTWC are known to affect systems with LMIs exceeding roughly 96 knots (Knapp & Kruk, 2010; Schreck et al., 2014). On the other hand, best track LMIs in the Australian Bureau of Meteorology tend to be systematically higher than in the JTWC best track data.

Figure 8. The locations of lifetime maximum intensities (LMI) of tropical cyclones for the period 1985–2014. LMI is color-coded according to category on the Saffir-Simpson Hurricane Wind Scale.

Figure 9. The distribution of lifetime maximum intensities (LMI) of tropical cyclones for all regions (Global), Northern Hemisphere (NH), Southern Hemisphere (SH), and each of the individual basins. The shaded rectangles show the interquartile range of LMIs, with the corresponding median value indicated by the thick horizontal line. The extreme upper and lower horizontal lines show the minimum and maximum LMI for each region, respectively (excluding outliers, which are indicated by open circles). The gray horizontal lines indicate intensity thresholds according to the Saffir-Simpson Hurricane Wind Scale, noting that the thresholds have been scaled to be consistent with 10-minute average MSSW speed data.

A list of the top 10 most intense tropical cyclones for each basin between 1985 and 2014, based on IBTrACS-WMO data, is provided in Table 1. Note that several new LMI records have since been set in the Eastern North Pacific basin (Hurricane Patricia), the South Indian basin (Very Intense Tropical Cyclone Fantala), and the South Pacific basin (Severe Tropical Cyclone Pam).

Table 1. Top 10 Most Intense Tropical Cyclones During the Period 1985–2014 for Each Basina

Rank

NAb

ENP

WNP

NI

SI

AUS

SP

1

Wilma

Linda

Megi

Odisha Oct

Jane-Ima

Monica

Zoet

Oct 2005

Sep 1997

Oct 2010

1999 123

Apr 2002

Apr 2006

Dec 2002

141 (882)

141 (902)

125 (885)

(912)

126 (915)

135 (916)

125 (890)

2

Gilbert

Rick

Haiyan

BOB 01

Hellen

Orson

Ron*

Sep 1988

Oct 2009

Nov 2013

Apr 1991

Mar 2014

Apr 1989

Jan 1998

141 (888)

136 (906)

125 (895)

112 (918)

125 (915)

134 (905)

125 (900)

3

Rita

John

Flo

BOB 01*

Gafilo

Inigo

Susan*

Sep 2005

Aug 1994

Sep 1990

May 1990

Mar 2004

Apr 2003

Jan 1998

136 (895)

132 (929)

120 (890)

112 (920)

125 (895)

130 (900)

125 (900)

4

Mitch

Kenna

Dot*

Gonu*

Kirsty

Graham

Percy*

Oct 1998

Oct 2002

Oct 1985

Jun 2007

Mar 1985

Dec 1991

Mar 2005

136 (905)

128 (913)

120 (895)

112 (920)

122 (920)

126 (915)

125 (900)

5

Katrina

Ioke

Yuri*

Forrest

Hudah*

Ingrid

Hina

Aug 2005

Aug 2006

Nov 1991

Nov 1992

Apr 2000

Mar 2005

Mar 1985

132 (902)

123 (915)

120 (895)

102 (952)

120 (905)

125 (924)

120 (910)

6

Dean

Marie

Lola

ARB 01

Harry*

Gwenda

Erica*

Aug 2007

Aug 2014

May 1986

May 2001

Mar 2002

Apr 1999

Mar 2003

132 (905)

123 (918)

120 (910)

101 (932)

120 (905)

121 (900)

115 (915)

7

Andrew

Guillermo

Ruth

BOB 02*

Thelma

Heta*

Aug 1992

Aug 1997

Oct 1991

May 1994

Apr 2005

Dec 1998

Jan 2004

132 (922)

123 (919)

115 (895)

101 (940)

120 (905)

120 (920)

115 (915)

8

Felix

Gilma

Vongfong

Phailin*

Edzani

Vance*

Meena*

Sep 2007

Jul 1994

Oct 2014

Oct 2013

Jan 2010

Mar 1999

Feb 2005

132 (929)

123 (920)

115 (900)

101 (940)

120 (910)

115 (910)

115 (915)

9

Ivan

Celia*

Jangmi*

Sidr

Bruce

Fay*

Olaf*

Sep 2004

Jun 2010

Sep 2008

Nov 2007

Dec 2014

Mar 2004

Feb 2005

128 (910)

123 (921)

115 (905)

101 (944)

120 (920)

115 (910)

115 (915)

10

Isabel

Elida*

Nida*

Nilofar

Dina*

Alex

Beni*

Sep 2003

Jul 2002

Nov 2009

Oct 2014

Jan 2002

Mar 1990

Jan 2003

128 (915)

123 (921)

115 (905)

97 (950)

115 (910)

115 (927)

110 (920)

(a) Intensity is ranked first by 10-minute average wind speed and then by minimum central pressure.

A dagger symbol (†) indicates that the storm no longer holds the record for the most intense tropical cyclone in its basin (e.g., Hurricane Patricia in 2015 became the strongest on record in the ENP).

An asterisk (*) indicates that the maximum intensity was shared with at least one other storm in the best track data for the same basin.

(b) NA, North Atlantic; ENP, Eastern North Pacific; WNP, Western North Pacific; NI, North Indian; SI, South Indian; AUS, Australian region; SP, South Pacific.

Source: IBTrACS-WMO data https://www.ncdc.noaa.gov/ibtracs/index.php?name=wmo-data.

Figure 10. Typhoon Haiyan at peak intensity on November 7, 2013, with 10-minute maximum sustained winds of 230 km h-1 (145 mph).

# Current Roadblocks to Our Understanding

Knowledge of the global distribution of tropical cyclones, including regional variability and trends, has progressed significantly since the advent of satellites in the 1960s. Yet despite rapid advancements in observational technology, the pioneering studies of tropical cyclone climatology, such as Gray (1968), provide a remarkably accurate depiction of the global spatial patterns of tropical cyclone formation (Fig. 3).

A significant roadblock to our understanding of past variability and trends in tropical cyclone activity resides in temporal heterogeneities present within the historical best track data. Changes in observational technology and analysis procedures over time have compromised the quality of the global best track data, particularly prior to the early 1980s. These heterogeneities affect not only the historical intensities of storms, but also regional and global storm counts because tropical cyclone classification depends on the maximum sustained surface winds exceeding a certain threshold. Furthermore, prior to satellite data, the chance of a storm being missed over the open ocean was much greater because such detection required that a ship encounter the tropical cyclone directly (e.g., Vecchi & Knutson, 2011). Even within the period of most reliable data since the mid-1980s, temporal heterogeneities exist due to, for example, changes in satellite viewing angles (Kossin et al., 2013), as well as interagency differences in wind-averaging periods and the subjective nature of the Dvorak technique itself. In an attempt to resolve some of these issues, considerable efforts have been invested into reanalysing historical meteorological data with the benefit of hindsight (e.g., Hagen, Strahan-Sakoskie, & Luckett, 2012; Harper et al., 2008; Hennon, 2012; Landsea, 2007; Landsea, Vecchi, Bengtsson, & Knutson, 2010; Landsea et al., 2012, 2014), including the development of temporally consistent, satellite-based, global tropical cyclone intensity data (e.g., Kossin, Knapp, Vimont, Murnane, & Harper, 2007; Kossin et al., 2013), which give a more accurate depiction of global and regional trends—particularly in relation to potential climate change impacts. For example, there has been a significant poleward migration of the mean latitude at which tropical cyclones attain their LMI during the period 1982–2012 (Kossin et al., 2014), which is one of the more robust results linking tropical cyclone intensity to climate change and variability.

On a more fundamental note, a satisfactory answer to the question of what sets the annual global rate of tropical cyclone formation, roughly 80 per year, has thus far evaded climate scientists. Several empirical relationships have been derived to relate tropical cyclone formation to large-scale climate variables, such as genesis potential indices, but there is to date no established theory relating tropical cyclone formation rate to climate. To understand how regional and global tropical cyclone activity will change with a changing climate, it is paramount that we first understand what sets the rate of tropical cyclone formation in the current climate.

## References

American Meteorological Society. (2000) cited 2016. Tropical cyclone. Glossary of Meteorology. Retrieved from http://glossary.ametsoc.org/wiki/Tropical_cyclone.

Basher, R. E., & Zheng, X. (1995). Tropical cyclones in the southwest Pacific: Spatial patterns and relationships to Southern Oscillation and sea surface temperature. Journal of Climate, 8, 1249–1260.Find this resource:

Beven, J. L., II, Avila, L. A., Blake, E. S., Brown, D. P., Franklin, J. L., Knabb, R. D., et al. (2008). Atlantic hurricane season of 2005. Monthly Weather Review, 136, 1109–1173.Find this resource:

Bister, M., & Emanuel, K. A. (1998). Dissipative heating and hurricane intensity. Meteorology and Atmospheric Physics, 52, 233–240.Find this resource:

Bruyere, C. L., Holland, G. J., & Towler, E. (2012). Investigating the use of a genesis potential index for tropical cyclones in the North Atlantic basin. Journal of Climate, 25, 8611–8626.Find this resource:

Camargo, S. J., Emanuel, K. A., & Sobel, A. H. (2007). Use of a genesis potential index to diagnose ENSO effects on tropical cyclone genesis. Journal of Climate, 20, 4819–4834.Find this resource:

Camargo, S. J., Robertson, A. W., Barnston, A. G., & Ghil, M. (2008). Clustering of eastern North Pacific tropical cyclone tracks: ENSO and MJO effects. Geochemistry, Geophysics, Geosystems, 9, Q06V05.Find this resource:

Camargo, S. J., & Sobel, A. H. (2005). Western North Pacific tropical cyclone intensity and ENSO. Journal of Climate, 18, 2996–3006.Find this resource:

Camargo, S. J., Sobel, A. H., Barnston, A. G., & Klotzbach, P. J. (2010). The influence of natural climate variability on tropical cyclones, and seasonal forecasts of tropical cyclone activity. In J. C. L. Chan & J. D. Kepert (Eds.), Global perspectives on tropical cyclones: From science to mitigation (pp. 325–360). Singapore: World Scientific Publishing.Find this resource:

Camargo, S. J., & Sobel, A. H. (2010). Revisiting the influence of the quasibiennial oscillation on tropical cyclone activity. Journal of Climate, 23, 5810–5825.Find this resource:

Chan, J. C. L. (1985). Tropical cyclone activity in the northwest Pacific in relation to the El Niño/Southern Oscillation phenomenon. Monthly Weather Review, 113, 599–606.Find this resource:

Chan, J. C. L. (1995). Tropical cyclone activity in the western North Pacific in relation to the stratospheric quasi-biennial oscillation. Monthly Weather Review, 123, 2567–2571.Find this resource:

Chan, J. C. L. (2005). Interannual and interdecadal variations of tropical cyclone activity over the western North Pacific. Meteorology and Atmospheric Physics, 89, 143–152.Find this resource:

Chan, J. C. L. (2007). Interannual variations of intense typhoon activity. Tellus, 59A, 455–460.Find this resource:

Chand, S. S., Tory, K., McBride, J., Wheeler, M., Dare, R., & Walsh, K. J. E. (2013). The different impact of positive-neutral and negative-neutral ENSO regimes on Australian tropical cyclones. Journal of Climate, 26, 8008–8016.Find this resource:

Chand, S. S., & Walsh, K. J. E. (2009). Tropical cyclone activity in the Fiji region: Spatial patterns and relationship to large-scale circulation. Journal of Climate, 22, 3877–3893.Find this resource:

Chen, T.-C., Weng, S. P., Yamazaki, N., & Kiehne, S. (1998). Interannual variation in the tropical cyclone formation over the western North Pacific. Monthly Weather Review, 126, 1080–1090.Find this resource:

Chia, H. H., & Ropelewski, C. F. (2002). The interannual variability in the genesis location of tropical cyclones in the northwest Pacific. Journal of Climate, 15, 2934–2944.Find this resource:

Chu, P.-S. (2004). ENSO and tropical cyclone activity. In R. J. Murnane & K.-B. Liu (Eds.), Hurricanes and typhoons, past, present and future (pp. 297–332). New York: Columbia University Press.Find this resource:

Chu, P.-S., & Wang, J. (1997). Tropical cyclone occurrences in the vicinity of Hawaii: Are the differences between El Niño and non–El Niño years significant? Journal of Climate, 10, 2683–2689.Find this resource:

Dare, R. A., & McBride, J. L. (2011). The threshold sea surface temperature condition for tropical cyclogenesis. Journal of Climate, 24, 4570–4576.Find this resource:

DeMaria, M., Knaff, J. A., & Conell, B. H. (2001). A tropical cyclone genesis parameter for the tropical Atlantic. Weather and Forecasting, 16, 219–233.Find this resource:

Diamond, H. J., Lorrey, A. M., & Renwick, J. A. (2013). A Southwest Pacific tropical cyclone climatology and linkages to the El Niño–Southern oscillation. Journal of Climate, 26, 3–25.Find this resource:

Dowdy, A. J., Qi, L., Jones, D., Ramsay, H., Fawcett, R., & Kuleshov, Y. (2012). Tropical cyclone climatology of the South Pacific Ocean and its relationship to El Niño–Southern Oscillation. Journal of Climate, 25, 6108–6122.Find this resource:

Dowdy, A. J. (2014). Long-term changes in Australian tropical cyclone numbers. Atmospheric Science Letters, 15, 292–298.Find this resource:

Dvorak, V. F. (1975). Tropical cyclone intensity analysis and forecasting from satellite imagery. Monthly Weather Review, 103, 420–430.Find this resource:

Dvorak, V. F. (1984). Tropical cyclone intensity analysis using satellite data. NOAA Techical Report NESDIS 11, Silver Spring.Find this resource:

Emanuel, K. A. (1986). An air–sea interaction theory for tropical cyclones. Part I: Steady-state maintenance. Journal of the Atmospheric Sciences, 43, 585–605.Find this resource:

Emanuel, K. A. (2010). Tropical cyclone activity downscaled from NOAA-CIRES Reanalysis, 1908–1958. Journal of Advances in Modeling Earth Systems, 2, 1–12.Find this resource:

Emanuel, K. A., & Nolan, D. S. (2004). Tropical cyclone activity and the global climate system. 26th Conference on Hurricanes and Tropical Meteorology, Miami, FL, American Meteorological Society, 10A.2.Find this resource:

Emanuel, K. A., & Rotunno, R. (2011). Self-stratification of tropical cyclone outflow. Part I: Implications for storm structure. Journal of Atmospheric Science, 68, 2236–2249.Find this resource:

Evan, A. T., & Camargo, S. J. (2011). A climatology of Arabian Sea cyclonic storms. Journal of Climate, 24, 140–158.Find this resource:

Evans, J. L., & Allan, R. J. (1992). El Niño/Southern Oscillation modification to the structure of the monsoon and tropical cyclone activity in the Australasian region. International Journal of Climatololgy, 12, 611–623.Find this resource:

Felton, C. S., Subrahmanyam, B., & Murty, V. S. N. (2013). ENSO modulated cyclogenesis over the Bay of Bengal. Journal of Climate, 26, 9806–9818.Find this resource:

Frank, W. M., & Young, G. S. (2007). The interannual variability of tropical cyclones. Monthly Weather Review, 135, 3587–3598.Find this resource:

Goebbert K. H, & Leslie, L. M. (2010). Interannual variability of north- western Australian tropical cyclones. Journal of Climate, 23, 4538–4555.Find this resource:

Goldenberg, S. B., & Shapiro, L. J. (1996). Physical mechanisms for the association of El Niño and West African rainfall with Atlantic major hurricane activity. Journal of Climate, 9, 1169–1187.Find this resource:

Gray, W. M. (1968). Global view of the origin of tropical disturbances and storms. Monthly Weather Review, 96, 669–700.Find this resource:

Gray, W. M. (1979). Hurricanes: Their formation, structure and likely role in the general circulation. In D. B. Shaw (Ed.), Meteorology over the tropical oceans (pp. 155–218). Bracknell: Royal Meteorological Society.Find this resource:

Gray, W. M. (1984). Atlantic seasonal hurricane frequency. Part I: El Niño and 30-mb quasi-biennial oscillation influences. Monthly Weather Review, 112, 1649–1668,Find this resource:

Gray, W. M., Landsea, C. W., Mielke, P. W., Jr., & Berry, K. J. (1993). Predicting Atlantic basin seasonal tropical cyclone activity by 1 August. Weather and Forecasting, 8, 73–86.Find this resource:

Gray, W. M., & Sheaffer, J. D. (1991). El Niño and QBO influences on tropical cyclone activity. In M. H. Glantz, R. W. Katz, & N. Nicholls (Eds.), Teleconnections linking worldwide climate anomalies (pp. 257–284). New York: Cambridge University Press.Find this resource:

Hagen, A. B., Strahan-Sakoskie, D., & Luckett, C. (2012). A reanalysis of the 1944–53 Atlantic hurricane seasons—the first decade of aircraft reconnaissance. Journal of Climate, 25, 4441–4460,Find this resource:

Harper, B. A., Kepert, J. D., & Ginger, J. D. (2010). Guidelines for converting between various wind averaging periods in tropical cyclone conditions. World Meteorological Organization, Geneva.Find this resource:

Harper, B. A., Stroud, S. A., McCormack, M., & West, S. (2008). A review of historical tropical cyclone intensity in northwestern Australia and implications for climate change trend analysis. Australian Meteorological Magazine, 57, 121–141.Find this resource:

Hastings, P. A. (1990). Southern Oscillation influences on tropical cyclone activity in the Australian/southwest Pacific region. International Journal of Climatology, 10, 291–298.Find this resource:

Hennon, C. C. (2012). Citizen scientists analyzing tropical cyclone intensities. Eos, Transactions American Geophysical Union, 93, 385–387.Find this resource:

Ho, C-H., Kim, J.-H., Jeong, J.-H., Kim, H.-S., & Chen, D. (2006). Variation of tropical cyclone activity in the south Indian Ocean: El Niño–Southern Oscillation and Madden-Julian Oscillation effects. Journal of Geophysical Research, 111, D22101.Find this resource:

Irwin, R. P., & Davis, R. (1999). The relationship between the Southern Oscillation Index and tropical cyclone tracks in the eastern North Pacific. Geophysical Research Letters, 26, 2251–2254.Find this resource:

Jien, J. Y., Gough, W. A., & Butler, K. (2015). The influence of El Niño–Southern Oscillation on tropical cyclone activity in the eastern North Pacific basin. Journal of Climate, 28, 2459–2474.Find this resource:

Jin, F.-F., Boucharel, J., & Lin, I.-I. (2014). Eastern Pacific tropical cyclones intensified by El Niño delivery of subsurface ocean heat. Nature, 516, 82–85.Find this resource:

Jury, M. R. (1993). A preliminary study of climatological associations and characteristics of tropical cyclones in the southwest Indian Ocean. Meteorology and Atmospheric Physics, 51, 101–115.Find this resource:

Kimberlain, T. B., Blake, E. S., & Cangialosi, J. P. (2016). National Hurricane Center tropical cyclone report—Hurricane Patricia, October 20–24, 2015. National Oceanographic and Atmospheric Administration, Miami.Find this resource:

Knaff, J. A. (1997). Implications of summertime sea level pressure anomalies in the tropical Atlantic region. Journal of Climate, 10, 789–804.Find this resource:

Knapp, K. R., & Kruk, M. C. (2010). Quantifying interagency differences in tropical cyclone best-track wind speed estimates. Monthly Weather Review, 138, 1459–1473.Find this resource:

Knapp, K. R., Kruk, M. C., Levinson, D. H., Diamond, H. J., & Neumann, C. J. (2010). The International Best Track Archive for Climate Stewardship (IBTrACS): Unifying tropical cyclone best track data. Bulletin- American Meteorological Society, 91, 363–376.Find this resource:

Knutson, T. R., McBride, J. L., Chan, J., Emanuel, K., Holland, G., Landsea, C., & co-authors. (2010). Tropical cyclones and climate change. Nature Geoscience, 3, 157–163.Find this resource:

Kossin, J. P., Camargo, S. J., & Sitkowski, M. (2010). Climate modulation of North Atlantic hurricane tracks. Journal of Climate, 23, 3057–3076.Find this resource:

Kossin, J. P., Emanuel, K. A., & Vecchi, G. A. (2014). The poleward migration of the location of tropical cyclone maximum intensity. Nature, 509, 349–352.Find this resource:

Kossin, J. P., Knapp, K. R., Vimont, D. J., Murnane, R. J., & Harper, B. A. (2007). A globally consistent reanalysis of hurricane variability and trends. Geophysical Research Letters, 34, L04815.Find this resource:

Kossin, J. P., Olander, T. L., & Knapp, K. R. (2013). Trend analysis with a new global record of tropical cyclone intensity. Journal of Climate, 26, 9960–9976.Find this resource:

Kuleshov, Y., & de Hoedt, G. (2003). Tropical cyclone activity in the Southern Hemisphere. Bulletin of the American Meteorological Society, 16, 135–137.Find this resource:

Kuleshov, Y., Qi, L., Fawcett, R., & Jones, D. (2008). On tropical cyclone activity in the Southern Hemisphere: Trends and the ENSO connection. Geophysical Research Letters, 35, L14S08.Find this resource:

Lander, M. A., & Guard, C. P. (1998). A look at global tropical cyclone activity during 1995: Contrasting high Atlantic activity with low activity in other basins. Monthly Weather Review, 126, 1163–1173.Find this resource:

Landsea, C. W., Harper, B. A., Hoarau, K., & Knaff, J. A. (2006). Can we detect trends in extreme tropical cyclones? Science, 313, 452–454.Find this resource:

Landsea, C. W. (2007). Counting Atlantic tropical cyclones back to 1900. EOS, 88, 197 and 202.Find this resource:

Landsea, C. W., Feuer, S., Hagen, A., Glenn, D. A., Sims, J., Perez, R., et al. (2012). A reanalysis of the 1921–1930 Atlantic hurricane database. Journal of Climate, 25, 865–885.Find this resource:

Landsea, C. W., Hagen, A., Bredemeyer, W., Carrasco, C., Glenn, D. A., Santiago, A., et al. (2014). A reanalysis of the 1931–43 Atlantic Hurricane Database. Journal of Climate, 27, 6093–6118.Find this resource:

Landsea, C. W., Pielke, R. A., Jr., Mestas-Nuñez, A. M., & Knaff, J. A. (1999). Atlantic basin hurricanes: Indices of climatic changes. Climatic Change, 42, 89–129.Find this resource:

Landsea, C. W., Vecchi, G. A., Bengtsson, L., & Knutson, R. (2010). Impact of duration thresholds on Atlantic tropical cyclone counts. Journal of Climate, 23, 2508–2519.Find this resource:

Li, Z., Yu, W., Li, T., Murty, V. S. N., & Tangang, F. (2013). Bimodal character of cyclone climatology in the Bay of Bengal modulated by monsoon seasonal cycle. Journal of Climate, 26, 1033–1046.Find this resource:

Lin, I.-I., Black, P., Price, J. F., Yang, C. Y., Chen, S. S., Lien, C. C. & co-authors (2013). An ocean coupling potential intensity index for tropical cyclones. Geophysical Research Letters, 40, 1878–1882.Find this resource:

Liu, K. S., & Chan, J. C. L. (2012). Interannual variation of Southern Hemisphere tropical cyclone activity and seasonal forecast of tropical cyclone number in the Australian region. International Journal of Climatology, 32, 190–202.Find this resource:

Maue, R. N. (2011). Recent historically low global tropical cyclone activity. Geophysical Research Letters, 38, L14803.Find this resource:

McTaggart-Cowan, R., Bosart, L. F., Davis, C. A., Atallah, E. H., Gyakum, J. R., & Emanuel, K. A. (2006). Analysis of Hurricane Catarina (2004). Monthly Weather Review, 134, 3029–3053.Find this resource:

Menkes, C. E., Lengaigne, M., Marchesiello, P., Jourdain, N. C., Vincent, E. M., Lefevre, J., et al. (2012). Comparison of tropical cyclogenesis indices on seasonal to interannual timescales. Climate Dynamics, 38, 301–321.Find this resource:

Nicholls, N. (1979). A possible method for predicting seasonal tropical cyclone activity in the Australian region. Monthly Weather Review, 107, 1221–1224.Find this resource:

Nicholls, N. (1984). The Southern Oscillation, sea-surface temperature, and internannual fluctuations in Australian tropical cyclone activity. International Journal of Climatology, 4, 661–670.Find this resource:

Palmen, E. H. (1948). On the formation and structure of tropical cyclones. In Geophysica (Vol. 3, pp. 26–38). Helsinki: University of Helsinki.Find this resource:

Peduzzi, P., Chatenou, B., Dao, H., De Bono, A., Herold, C., Kossin, J., et al. (2012). Global trends in tropical cyclone risk. Nature Climate Change, 2, 289–294.Find this resource:

Pezza, A. B., & Simmonds, I. (2005). The first South Atlantic hurricane: Unprecedented blocking, low shear and climate change. Geophysical Research Letters, 32, L15712.Find this resource:

Ramsay, H. A., Camargo, S. J., & Kim, D. (2012). Cluster analysis of tropical cyclone tracks in the Southern Hemisphere. Climate Dynamics, 39, 897–917.Find this resource:

Ramsay, H. A., Leslie, L. M., Lamb, P. J., Richman, M. B., & Leplastrier, M. (2008). Interannual variability of tropical cyclones in the Australian region: Role of large-scale environment. Journal of Climate, 21, 1083–1103.Find this resource:

Ramsay, H. A., Richman, M. B., & Leslie, L. M. (2014). Seasonal tropical cyclone predictions using optimized combinations of ENSO regions: Application to the Coral Sea basin. Journal of Climate, 27, 8527–8542.Find this resource:

Ramsay, H. A., Richman, M. B., & Leslie, L. M. (2017). The modulating influence of Indian Ocean sea surface temperatures on Australian region seasonal tropical cyclone counts. Journal of Climate. Published online, 27 February 2017.Find this resource:

Ramsay, H. A., & Sobel, A. H. (2011). Effects of relative and absolute sea surface temperature on tropical cyclone potential intensity using a single-column model. Journal of Climate, 24, 183–193.Find this resource:

Rappaport, E. N. (2000). Loss of life in the United States associated with recent Atlantic tropical cyclones. Bulletin- American Meteorological Society, 81, 2065–2073.Find this resource:

Rappaport, E. N. (2013). Fatalities in the United States from Atlantic tropical cyclones: New data and interpretation. Bulletin- American Meteorological Society, 95, 341–346.Find this resource:

Rappaport, E. N., & Blanchard, B. W. (2015). Fatalities in the United States indirectly associated with Atlantic tropical cyclones. Bulletin- American Meteorological Society, 97, 1139–1148.Find this resource:

Revell C. G., & Goulter, S. W. (1986). South Pacific tropical cyclones and the Southern Oscillation. Monthly Weather Review, 114, 1138–1145.Find this resource:

Royer, J.-F., Chauvin, F., Timbal, B., Araspin, P., & Grimal, D. (1998). A GCM study of the impact of greenhouse gas increase on the frequency of occurrence of tropical cyclone. Climatic Change, 38, 307–343.Find this resource:

Schreck, C. J., III, Knapp, K. R., & Kossin, J. P. (2014). The impact of best track discrepancies on global tropical cyclone climatologies using IBTrACS. Monthly Weather Review, 142, 3881–3899.Find this resource:

Singh, O. P., Khan, T. M. A., & Rahman, M. S. (2000). Changes in the frequency of tropical cyclones over the North Indian Ocean. Meteorology and Atmospheric Physics, 75, 11–20.Find this resource:

Solow, A., & Nicholls, N. (1990). The relationship between the Southern Oscillation and tropical cyclone frequency in the Australian region. Journal of Climate, 3, 1097–1101.Find this resource:

Swanson, K. L. (2008). Nonlocality of Atlantic tropical cyclone intensities. Geochemistry, Geophysics, Geosystems, 9, Q04V01.Find this resource:

Tang, B. H., & Neelin, J. D. (2004). ENSO influence on Atlantic hurricanes via tropospheric warming. Geophysical Research Letters, 31, L24204.Find this resource:

Tippett, M. K., Camargo, S. J., & Sobel, A. H. (2011). A Poisson regression index for tropical cyclone genesis and the role of large-scale vorticity in genesis. Journal of Climate, 24, 2335–2357.Find this resource:

Vecchi, G., & Soden, B. (2007). Effect of remote sea surface temperature change on tropical cyclone potential intensity. Nature, 450, 1066–1071.Find this resource:

Vecchi, G. A., & Knutson, T. R. (2011). Estimating annual numbers of Atlantic hurricanes missing from the HURDAT database (1878–1965) using ship track density. Journal of Climate, 24, 1736–1746.Find this resource:

Velden, C. S., & Hawkins, J. (2010). Satellite observations of tropical cyclones. In J. Chan & J. Kepert (Eds.), Global perspectives of tropical cyclones (3d ed., pp. 201–226). World Scientific Publishing Company, Singapore.Find this resource:

Velden, C. S., Harper, B., Wells, F., Beven, J. L., Zehr, R., Oladner, T., & co-authors (2006). The Dvorak tropical cyclone intensity estimation technique: A satellite-based method that has endured for over 30 years. Bulletin- American Meteorological Society, 85, 353–385.Find this resource:

Vincent, E. M., Emanuel, K. A., Lengaigne, M., Vialard, J., & Madec, G. (2014). Influence of upper ocean stratification interannual variability on tropical cyclones. Journal of Advances in Modeling Earth Systems, 6, 680–699.Find this resource:

Vincent, E. M., Lengaigne, M., Menkes, C. E., Jourdain, N. C., Marchesiello, P., & Madec, G. (2011). Interannual variability of the South Pacific convergence zone and implications for tropical cyclone genesis. Climate Dynamics, 36, 1881–1896.Find this resource:

Walsh, K., Camargo, S. J., Vecchi, G. A., Daloz, A. S., Elsner, J., Emanuel, K., & co-authors (2015). Hurricanes and climate: The U.S. CLIVAR Working Group on Hurricanes. Bulletin- American Meteorological Society, 96, 997–1017.Find this resource:

Wang, B., & Chan, J. C. L. (2002). How strong ENSO events affect tropical storm activity over the western North Pacific. Journal of Climate, 15, 1643–1658.Find this resource:

Werner, A., & Holbrook, N. J. (2011). A Bayesian forecast model for Australian region tropical cyclone formation. Journal of Climate, 24, 6114–6131.Find this resource:

Whitney, L. D., & Hobgood, J. (1997). The relationship between sea surface temperatures and maximum intensities of tropical cyclones in the eastern North Pacific Ocean. Journal of Climate, 10, 2921–2930.Find this resource:

Wu, P., & Chu, P.-S. (2007). Characteristics of tropical cyclone activity over the eastern North Pacific: The extremely active 1992 and the inactive 1977. Tellus, Ser. A, 59, 444–454.Find this resource:

Zhang, G., Zhang, X., & Wei, F. (1994). A study on the variations of annual frequency of tropical cyclone in northwest Pacific during the last hundred years. Journal of Tropical Meteorology, 11, 315–323.Find this resource: