|A Dynamically Based Climatology of
Subtropical Cyclones in the North Atlantic Basin
The National Hurricane Center (NHC) online glossary defines a subtropical cyclone (STC) as a “non-frontal low-pressure system that has characteristics of both tropical and extratropical cyclones…. Unlike tropical cyclones, subtropical cyclones derive a significant portion of their energy from baroclinic sources, and are generally cold-core in the upper troposphere, often associated with an upper-level low or trough” (OFCM 2014).
The NHC STC definition emphasizes the hybrid nature of STCs and suggests that both baroclinic and diabatic energy sources contribute to STC formation.
The opportunity to investigate the roles of baroclinic and diabatic processes during the evolution of STCs motivates this study.
An objective STC identification technique, based on the changing contribution of baroclinic and diabatic processes during the evolution of individual cyclones, will be formulated and used to construct a dynamically based North Atlantic STC climatology.
The intraseasonal variability associated with the location and frequency of North Atlantic STC formation will be presented and compared to the intraseasonal variability associated with the location and frequency of North Atlantic tropical cyclones (TCs).
This study seeks to eliminate the remaining ambiguity in real time STC identification and provide the foundation for documenting the structure, motion, and evolution of the upper-tropospheric features linked to North Atlantic STC formation in subsequent studies.
The hybrid nature of STCs, emphasized in the NHC STC definition, makes them likely candidates to become TCs via the tropical transition (TT) process (Davis and Bosart 2003, 2004), during which cold-core oceanic cyclones transition into warm-core TCs in the presence of an upper-tropospheric disturbance.
Operational forecasters began documenting STCs transitioning into TCs in the presence of an upper-tropospheric disturbance in the early 1950’s (e.g., Moore and Davis 1951; Simpson 1952).
Moore and Davis (1951) and Simpson (1952) analyzed cold-core oceanic cyclones, initially collocated with regions of relatively cold upper-tropospheric air, that transitioned into warm-core TCs above relatively warm sea surface temperatures (SSTs).
The STC analyzed by Moore and Davis (1951) was unprecedented in the North Atlantic basin, transitioning into a category one hurricane (TC Able) ~200 km east of Cape Canaveral, FL, in May 1951.
TC Able would ultimately strengthen into a category three hurricane ~115 km southeast of Cape Hatteras, NC, before moving out to sea, making it the strongest TC to develop in the North Atlantic basin before the official start of the North Atlantic TC season (Norton 1952).
The tendency for STCs to form and rapidly undergo TT close to Bermuda [e.g., TC Grace (October 1991) (Pasch and Avila 1992), TC Karen (October 2001) (Stewart 2001; Guishard et al. 2007; Evans and Guishard 2009; Hulme and Martin 2009a,b)], the east coast of North America [e.g., TC Bertha (July 1990) (Gerrish 1990), the “Perfect Storm” (October 1991) (Pasch and Avila 1992; Cordeira and Bosart 2010, 2011), TC Michael (October 2000) (Franklin et al. 2001; Hulme and Martin 2009a)], and the west coast of Europe [e.g., TC Edouard (August 1990) (Case 1990), TC Vince (October 2005) (Franklin 2006), and TC Delta (November 2005) (Beven 2006)] can create potential challenges for operational forecasters and emergency managers.
A recent example of a transitioning STC, STC Beryl (May 2012), formed off the east coast of North Carolina, underwent TT, and made landfall as a 55 kt tropical storm near Jacksonville Beach, FL, in ~52 h (Beven 2012).
Figure 1 depicts STC Beryl at 1753 UTC 26 May 2012, ~340 km southeast of Wilmington, NC, when the cyclone had attained a minimum central pressure of 999 hPa and maximum sustained winds of 40 kts.
Along with its rapid formation and TT, TC Beryl is also remembered for being the strongest TC to make landfall in the continental United States before the official start of the North Atlantic TC season—producing 3–7 in rainfall totals across northeastern Florida/southeastern Georgia and storm surges 1–3 ft above normal tide levels along portions of the East Coast (Beven 2012).
Despite the existence of an STC definition in the NHC online glossary and numerous examples of STCs transitioning into noteworthy North Atlantic TCs, the real-time classification of oceanic cyclones as STCs remains fairly ambiguous (González-Alemán et al. 2015).
Guishard et al. (2009) sought to eliminate this ambiguity by developing a quantitative definition of STCs in terms of their location of formation, hybrid synoptic structure, and surface wind speed.
North Atlantic STCs forming during 1957–2002 were objectively identified by Guishard et al. (2009) by applying a consistent set of criteria to oceanic cyclones within the 1.25° European Centre for Medium-Range Forecasts (ECMWF) Re-analysis (ERA-40) dataset (Uppala et al. 2005).
In their study, North Atlantic STCs are required to form between 20°N and 40°N, exhibit hybrid structure within the Hart (2003) cyclone phase space (CPS) for ≥36 h, attain gale-force winds (17 m s−1) at 925 hPa while exhibiting hybrid structure, and be identified within 24 h of becoming purely cold or warm core.
The overlap between the North Atlantic STCs identified by Guishard et al. (2009) and North Atlantic STCs identified by NHC during 1957–2002 is considerable, but not perfect—likely due to the different techniques and criteria used in their identification (González-Alemán et al. 2015).
Of the 197 North Atlantic STCs identified by Guishard et al. (2009), 53 were cyclones named by NHC and included in the v03r01 edition of the International Best Track Archive for Climate Stewardship (IBTrACS) dataset (Knapp et al. 2010).
About half of the cyclones included in both datasets (27/53) were identified as STCs by NHC, with the 26 remaining cyclones identified as TCs (González-Alemán et al. 2015). Conversely, 65 of the 92 STCs identified by NHC were not identified by Guishard et al. (2009), likely due to the criteria imposed to identify STCs developing in situ.
The pioneering work of Guishard et al. (2009) excited subsequent research exploring STC formation in the South Atlantic basin.
Evans and Braun (2012) objectively identified South Atlantic STCs forming during 1957–2007 by applying a consistent set of criteria extremely similar to the criteria used by Guishard et al. (2009) to oceanic cyclones within the 1.25° ERA-40 dataset (Uppala et al. 2005).
A subsequent study by Gozzo et al. (2014) objectively identified South Atlantic STCs by applying a similar, but less restrictive, set of criteria used by Evans and Braun (2012) [e.g., relaxation of hybrid structure requirement within the Hart (2003) cyclone phase space, no gale-force wind requirement] to oceanic cyclones within the 1.5° Interim ECMWF Re-analysis (ERA-Interim) dataset (Dee et al. 2011) and 2.5° National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis dataset (Kalnay et al. 1996).
The less restrictive set of criteria used by Gozzo et al. (2014) resulted in a ~600% increase in the number of STCs identified annually in the South Atlantic basin.
The contrasting results of Evans and Braun (2012) and Gozzo et al. (2014), as well as those of Guishard et al. (2009) and Knapp et al. (2010), suggest that STCs may display a wide range of structural characteristics while still exhibiting the baroclinic and diabatic processes emphasized in the NHC STC definition.
The development of a new, dynamically based, STC identification technique may provide further insight into the structure and evolution of STCs and eliminate the remaining ambiguity in real-time STC identification.
Previous work by Davis (2010) developed a methodology for quantifying the relative contributions of baroclinic and diabatic processes during the evolution of individual cyclones in order to identify STCs within idealized numerical simulations.
This study expands upon the work of Davis (2010), investigating the roles of baroclinic and diabatic processes during the evolution of individual cyclones by calculating three potential vorticity (PV) metrics from the 0.5° NCEP Climate Forecast System Reanalysis (CFSR) dataset (Saha et al. 2010).
An objective STC identification technique that uses these three PV metrics will be formulated and applied to candidate STCs identified in the global climatology of baroclinically influenced tropical cyclogenesis events created by McTaggart-Cowan et al. (2013) in order to construct a dynamically based 1979–2010 North Atlantic STC climatology.
The remainder of this paper is organized as follows.
The data and methodology used to formulate an objective STC identification technique are described in section 2.
Section 3 contains a dynamically based 1979–2010 North Atlantic STC climatology and a comparison of the intraseasonal variability associated with the location and frequency of North Atlantic STCs with that of North Atlantic TCs.
A discussion of key findings and conclusions are contained in section 4.
2. Data and methodology
a. Candidate STCs
Baroclinically influenced tropical cyclogenesis cases identified in McTaggart-Cowan et al. (2013) that occurred over the North Atlantic from 1979 through 2010 were considered for potential STC identification (460 candidate STCs).
The period from 1979 through 2010 was chosen to coincide with the period covered by the 0.5° CFSR dataset (Saha et al. 2010).
North Atlantic cyclone tracks were obtained from the v03r03 edition of the International Best Track Archive for Climate Stewardship (IBTrACS) dataset (Knapp et al. 2010).
North Atlantic cyclone tracks were extended backward 36 h from their first IBTrACS position using a reverse steering flow calculation described in detail in McTaggart-Cowan et al. (2008).
McTaggart-Cowan et al. (2013) separated tropical cyclogenesis cases into one of five development pathways based on two external forcings in the near-TC environment prior to TC formation:
QG forcing for ascent, Q, determined by the average convergence of the 400–200-hPa Q vector within 6° of the cyclone center
lower-tropospheric baroclinicity, Th, determined by asymmetries in the 1000–700-hPa thickness field within 10° of the cyclone center
The five development pathways identified in McTaggart-Cowan et al. (2013) include: 1) Strong TT, 2) Weak TT, 3) Trough induced, 4) Low-level baroclinic, and 5) Nonbaroclinic events. The relative values of Q and Th associated with each development pathway, as well as a brief physical description of each development pathway, are given in Table 1.
To ensure consistency with the NHC STC definition (OFCM 2014), only baroclinically influenced tropical cyclogenesis cases occurring in the presence of an upper-tropospheric disturbance were considered for potential STC identification, restricting the development pathways considered to those with “high” values of QG forcing for ascent in Table 1: Strong TT, Weak TT, and Trough induced.
Of the 460 cyclones identified in McTaggart-Cowan et al. (2013) that occurred over the North Atlantic from 1979 through 2010, only 222 cyclones formed in the presence of an upper-tropospheric disturbance and were considered for potential STC identification.
A track map of the 222 cyclones considered for potential STC identification is shown in Fig. 2. (Use Fig. 2 to say more about impacts to North America and Europe)
b. Adapted Davis (2010) methodology
The Davis (2010) methodology for STC identification is based on the concept of Ertel PV and formulated in terms of two PV metrics that quantify the relative contributions of baroclinic processes and condensation heating during the evolution of individual cyclones.
The Davis (2010) methodology distinguishes between cyclone types within an idealized numerical simulation based on the relative contributions of baroclinic processes and condensation heating during the evolution of individual cyclones and can be thought of as similar to the Hart (2003) cyclone phase space diagrams utilized for STC identification by Guishard et al. (2009), Evans and Braun (2012), Gozzo et al. (2014), and González-Alemán et al. (2015).
The transition from identifying STCs within an idealized numerical simulation to the 0.5° CFSR dataset requires the adaptation of the original Davis (2010) methodology.
All PV metrics considered in the present study are calculated in a 6° box centered over the surface cyclone. The first PV metric in the Davis (2010) methodology, PV1, represents lower-tropospheric baroclinic processes in terms of the near-surface potential temperature anomaly:
PV1 , (1)
is the absolute vorticity, and is the potential temperature anomaly calculated at an individual grid point from an 11-day centered mean. The potential temperature anomaly variations across the 6° box, and are averaged between 925 hPa and 850 hPa prior to computing The horizontal scales, and , are the length of 6° of latitude and the longitudinal length of the box as a function of latitude, respectively. The horizontal scales, and , represent the lengths of the northern edge, center, and southern edge of the 6° box, respectively. The vertical scale, , is equal to 425 hPa to match the vertical integration of the lower-tropospheric PV anomaly (see below).
The second PV metric in the Davis (2010) methodology, PV2, represents midtropospheric latent heat release in terms of the lower-tropospheric PV anomaly:
where is the PV anomaly calculated at an individual grid point from an 11-day centered mean and is equal to 425 hPa.
For the purposes of this study, the authors introduce two additional metrics used to evaluate upper-tropospheric dynamical processes and upper-tropospheric structure (i.e., cold core vs. warm core) above the surface cyclone.
The first additional metric, PV3, represents upper-tropospheric dynamical processes in terms of the upper-tropospheric PV anomaly:
where is equal to 300 hPa.
The second additional metric, ζT, represents upper-tropospheric structure in terms of the upper-tropospheric thermal vorticity:
where horizontal scales, and , are the length of 12° of latitude, the longitudinal length of the box as a function of latitude, and the length of the center of the 12° box, respectively.
A 12° box was used to calculate ζT in order to capture the large-scale structure of the upper troposphere directly over and surrounding the cyclone center.
A schematic representation of the regions over which PV1, PV2, PV3, and ζT are calculated is shown in Fig. 3. The values of PV1, PV2, PV3, and ζT are smoothed using a 1–2–1 temporal filter prior to STC identification.
c. STC identification
In order to determine the time and location of STC formation within the subset of baroclinically influenced tropical cyclogenesis cases specified in section 2a, an objective identification technique for detecting STC formation was formulated, incorporating PV2, PV3, and ζT, and applied to the 0.5° CFSR dataset.
PV1, which represents lower-tropospheric baroclinic processes, was not included due to the range of lower-tropospheric thermal gradients associated with the subset of baroclinically influenced tropical cyclogenesis cases specified in section 2a (Table 1).
STC formation was identified the first time (t = t0) at which the following criteria were met:
1) There is a positive upper-tropospheric PV anomaly (representing an upper-tropospheric low or trough) and positive lower-tropospheric PV anomaly (e.g., representing a PV tower) over the cyclone center (i.e., PV3 > 0 and PV2 > 0 at t = t0)
2) The upper-tropospheric PV anomaly begins to be eroded by midtropospheric latent heat release [i.e., d(PV3)/dt < 0 at t = t0 + 6 h and t0 + 12 h]
3) The magnitude of the upper-tropospheric PV anomaly decreases faster than the magnitude of the lower-tropospheric PV anomaly [i.e., d(PV3)/dt < d(PV2)/dt at t = t0 + 6 h and t0 + 12 h]
4) The cyclone is cold core in the upper-troposphere (i.e., ζT > 0 at t = t0)
5) The previous four criteria are met < 3 days after the first IBTrACS position
An example of the application of the objective identification technique for detecting STC formation is shown in Fig. 4 for the case of STC Sean, which formed over the western North Atlantic in November 2011. STC Sean was selected as an illustrative case of STC formation over the western North Atlantic in the presence of an upper-tropospheric disturbance (Avila 2012).
Figure 4 reveals the changing PV structure of STC Sean during its evolution. The objective identification technique for detecting STC formation indicates that Sean became an STC at 1200 UTC 7 November 2011, at the time when the upper-tropospheric PV anomaly began to be eroded by midtropospheric latent heat release (Raymond 1992) and the upper troposphere became cold core.