An Investigation of Composite Dropsonde Profiles for Developing

542 JOURNAL OF THE ATMOSPHERIC SCIENCES VOLUME 70

An Investigation of Composite Dropsonde Proﬁles for Developing and Nondeveloping Tropical Waves during the 2010 PREDICT Field Campaign

WILLIAM A. KOMAROMI Rosenstiel School of Marine and Atmospheric Science, University of Miami, Miami, Florida

(Manuscript received 15 February 2012, in ﬁnal form 20 August 2012)

ABSTRACT

Composite dropsonde profiles are analyzed for developing and nondeveloping tropical waves in an attempt to improve the understanding of tropical cyclogenesis. These tropical waves were sampled by 25 reconnaissance missions during the 2010 Pre-Depression Investigation of Cloud-Systems in the Tropics (PREDICT) field campaign. Comparisons are made between mean profiles of temperature, mixing ratio, relative humidity, radial and tangential winds, relative vorticity, and virtual convective available potential energy (CAPE) for genesis and nongenesis cases. Genesis soundings are further analyzed in temporal progression to investigate whether significant changes in the thermodynamic or wind fields occur during the transition from tropical wave to tropical cyclone. Significant results include the development of positive temperature anomalies from 500 to 200 hPa 2 days prior to genesis in developing waves. This is not observed in the nongenesis mean. Progressive mesoscale moistening of the column is observed within 150 km of the center of circulation prior to genesis. The genesis composite is found to be significantly moister than the nongenesis composite at the middle levels, while comparatively drier at low levels, suggesting that dry air is more detrimental to genesis when located at the middle levels. Time-varying tangential wind profiles reveal an initial delay in intensification, followed by an increase in organization 24 h pregenesis. The vertical evolution of relative vorticity, in addition to a warm- over-cold thermal structure, is more consistent with a top-down than a bottom-up genesis mechanism. Last, CAPE values are much greater for nongenesis than genesis profiles, indicating that greater instability does not necessarily favor genesis.

1. Introduction surface temperatures greater than or equal to 268C (Palmen 1948), an unstable or conditionally unstable Predicting tropical cyclogenesis remains one of the environment, and relatively high moisture content from great forecasting challenges to today’s meteorological the surface through 5 km (Gray 1979). However, despite community (Emanuel 2005). Much of our limited un- these well-known criteria, the exact sequence of events derstanding can likely be attributed to our inability to culminating in tropical cyclogenesis remains unknown. differentiate the often subtle physical differences be- Two differing views of tropical cyclone formation are tween developing and nondeveloping tropical cyclones the top-down and the bottom-up hypotheses. Ritchie (TCs), and any such differences, when observed, have and Holland (1997) and Simpson et al. (1997) describe been insufﬁciently documented (Dunkerton et al. 2009). a top-down mechanism for genesis by which successive Among the well-known necessary dynamic conditions mergers of mesoscale convective systems (MCSs) in- for tropical cyclogenesis are background cyclonic vor- crease the size and/or strength of the midlevel vortex, ticity, 850–200-hPa tropospheric wind shear of less than 2 2 which induces a surface circulation through vertical 15 m s 1 and preferably below 10 m s 1, and a sufﬁ- penetration and vortex stretching. Similarly, Bister and ciently high Coriolis parameter (Gray 1968). Ther- Emanuel (1997) propose that a stratiform rain region modynamic prerequisites exist as well, including sea associated with an existing MCS acts to moisten and cool the mid- to lower levels. The level of peak cooling descends within the stratiform rain region, thereby Corresponding author address: William Komaromi, RSMAS, Di- vision of Meteorology and Physical Oceanography, 4600 Rickenbacker lowering the level of maximum potential vorticity (PV) Causeway, Miami, FL 33149. production, while moistening acts to limit the occur- E-mail: [email protected] rence of dry downdrafts. Along with the necessity of

DOI: 10.1175/JAS-D-12-052.1

Ó 2013 American Meteorological Society Unauthenticated | Downloaded 09/25/21 07:15 AM UTC FEBRUARY 2013 K O M A R O M I 543 a strengthening midlevel circulation, Nolan (2007) also disturbance within the pouch is repeatedly moistened by found humidification of the inner core due to moist deep moist convection within the critical layer while detrainment and precipitation from deep convective remaining somewhat protected from lateral intrusion of towers preceding genesis. However, Nolan (2007) does dry air and deformation by horizontal and vertical shear. not necessitate a top-down genesis process. Lastly, a re- This protovortex, collocated with the critical latitude, is cent study by Raymond et al. (2011) of five tropical cy- then able to keep pace with the parent wave until it has clogenesis events in the northwestern Pacific suggests strengthened into a self-maintaining entity. Hypotheti- that tropical cyclogenesis is facilitated by a preexisting cally, the marsupial paradigm could be used in conjunc- midlevel vortex. This midlevel vortex creates a cold core tion with either the top-down or bottom-up genesis at low levels, which alters deep convection as to facilitate hypotheses. Dunkerton et al. (2009) assume a bottom- spinup. up progression of genesis. A slightly differing sequence, known as bottom-up As already alluded to, much of the difficulty in iden- genesis, is proposed by Hendricks et al. (2004) and tifying the exact order of processes that occur during Montgomery et al. (2006), in which individual deep genesis, or whether top-down or bottom-up sequences moist convective updrafts or vortical hot towers (VHTs) both occur under different conditions, can be attributed develop within the tropical wave, amplify preexisting to a lack of in situ data prior to genesis. In an attempt to cyclonic vorticity, and gradually consolidate to form expand upon the limited dataset, several field campaigns a low-level center of circulation. Latent heat released have sampled tropical cyclones during and shortly after within these VHTs aids in the development of the mid- the genesis stage, including the Tropical Experiment in level warm core, and surface convergence and upper- Mexico (TEXMEX; Bister and Emanuel 1997; Raymond level divergence commence. Observational evidence et al. 1998), the Tropical Cloud Systems and Processes supporting a top-down mechanism for genesis is pre- (TCSP) experiment in 2005 (Halverson et al. 2007), the sented by Ritchie and Holland (1997) and Mapes and National Aeronautics and Space Administration (NASA) Houze (1995), while Houze et al. (2009) find evidence component of the African Monsoon Multidisciplinary that support the VHT argument, all for individual case Analyses (AMMA) project in 2006 (Zipser et al. 2009), studies. the Tropical Cyclone Structure experiment in 2008 Regardless of the exact order of processes by which (TCS-08; Elsberry and Harr 2008), as well as a handful genesis occurs, the dependence upon some initial MCS of observations from the Hurricane Rainband and In- or VHTs assumes sufficient tropospheric instability tensity Change Experiment (RAINEX) of 2005 (Houze to allow deep convection. Using in situ data, Molinari et al. 2006). Case studies using data from these experi- and Vollaro (2010) find that highly sheared, generally ments, such as Zipser et al. (2009), emphasize the diffi- weaker tropical cyclones tend to be associated with culty of achieving genesis in excessively dry air masses. higher convective available potential energy (CAPE) Ritchie and Holland (1997), Davis et al. (2008), Houze than their nonsheared, generally stronger counterparts. et al. (2009), and Braun et al. (2010) have shown that the Similarly, Braun (2010) found higher CAPE in envi- progressive strengthening of a midlevel vortex, a grad- ronments for weakening TCs compared to strengthening ual moistening of the column in a region of deep con- TCs in the days following genesis. In idealized numerical vection, and the development of a warm core are all simulations, Nolan et al. (2007) found that greater maxi- evident in observations of various tropical cyclones mum potential intensity (MPI) resulted in greater likeli- during and shortly following genesis. While these studies hood of genesis, while greater CAPE did not. Nonetheless, allude to the development of a warm core, the altitude of the question of whether genesis becomes increasingly the warm-core maxima and the timing of the devel- favored with increasing instability, or whether there is opment of the warm core are generally neglected. Ear- some threshold beyond which decreasing stability is lier observational studies such as La Seur and Hawkins detrimental to genesis, has not been conclusively an- (1963) and Hawkins and Rubsam (1968) have found swered via observational evidence. maximum warm anomalies at around 250 hPa in mature A recent endeavor to better understand tropical TCs, while Hawkins and Imbembo (1976) and Stern and cyclogenesis from a wave-relative framework is under- Nolan (2012) suggest that the primary warm core is lo- taken by Dunkerton et al. (2009). Known as the mar- cated from 500 to as low as 650 hPa. The level of max- supial paradigm, tropical depression formation from imum warm anomalies for pregenesis disturbances a predepression wave trough in the lower troposphere is remains to be determined. greatly favored within the critical-layer ‘‘pouch’’—a re- The most recent of the field campaigns involving gion of closed material contours wherein the parent genesis, known as Pre-Depression Investigation of wave’s phase speed equals the mean flow. A young Cloud-Systems in the Tropics (PREDICT), was an

Unauthenticated | Downloaded 09/25/21 07:15 AM UTC 544 JOURNAL OF THE ATMOSPHERIC SCIENCES VOLUME 70 expansive investigation of tropical cyclogenesis in the also included. Identification of the vertical level, timing, central Atlantic during the 2010 season (Montgomery and magnitude of warm core development, if at all dis- et al. 2012). One of the overarching goals of PREDICT cernible prior to genesis, will be sought. The data will be was to gather in situ observations necessary to examine evaluated to determine whether a progressive increase the marsupial theory of genesis. Ultimately, PREDICT in moisture to near saturation or a top-down versus has provided the most expansive in situ dataset comprising bottom-up transition of the mean vortex are observed. both developing and nondeveloping tropical waves prior An investigation of whether cases of genesis are asso- to genesis, complemented by simultaneous observations ciated with greater instability is presented. from the Intensity Forecasting Experiment (IFEX) and An overview of the PREDICT dropsonde dataset and Genesis and Rapid Intensification Processes (GRIP) the methodologies and analysis techniques used in this campaigns. During PREDICT, the National Science study appear in section 2. In section 3, comparisons of Foundation (NSF)/National Center for Atmospheric Re- anomalous values of a variety of metrics with respect to search (NCAR) Gulfstream-V (G-V) aircraft provided the PREDICT mean sounding for genesis and nongenesis once- and, occasionally, twice-daily sampling of tropical profiles are made. Conclusions will appear in section 4. waves for up to five consecutive days. Unlike in previous field campaigns, by incorporating the definition of the 2. Data and methods critical layer, it was possible to target a potential genesis region much smaller than an entire tropical wave. In total, During PREDICT, 547 dropsondes were deployed 25 flight missions were completed between 15 August and over the course of 25 aircraft missions investigating 30 September 2010 (plus 1 calibration flight). The strength tropical waves in the Caribbean and western Atlantic of this dataset comes not only from the unprecedented (Fig. 1). Five cases of genesis, three cases of nongenesis, quantity of developing and nondeveloping tropical waves and four TCs named during or prior to investigation (TC sampled, but also from the temporally evolving nature of stage) constitute the PREDICT dataset (Table 1). The data associated with distinct pouches. TC stage category overlaps with three of the five genesis Two recent papers, Smith and Montgomery (2011) cases: TSs Fiona, Matthew, and Nicole, in addition to an and Davis and Ahijevych (2012), examined PREDICT investigation of TS Gaston prior to weakening to a dropsonde data for Tropical Storm (TS) Matthew, remnant low. Cases are sorted by genesis or nongenesis Hurricane Karl, and ex-TS Gaston. Smith and Montgomery based upon whether or not the tropical wave under in- (2011) found lower values of equivalent potential vestigation eventually yields a tropical storm/depression temperature between the surface and 3 km of non- as declared by the National Hurricane Center (NHC). developing ex-Gaston than in developing pre-Karl The genesis category is further separated temporally and pre-Matthew. The authors found evidence that dry into missions that occur 0–24 h pregenesis, 24–48 h pre- air for the nondeveloping case was not necessarily asso- genesis, 48–72 h pregenesis, and .72 h pregenesis. Mis- ciated with stronger downdrafts but rather that the drier sions that begin during one such time period and end in midlevel air weakened the convective updrafts and another are assigned to the period during which the thereby prevented sufficient amplification of system rel- majority of dropsondes are deployed. As such, the 0–24-h ative vorticity necessary for development. Last, greater pregenesis category includes data from the 30 August CAPE and convective inhibition (CIN) were associated flight into pre-TS Fiona, the 14 September flight into with ex-Gaston than either genesis event. Davis and pre-Hurricane Karl, and the 27 September flight into Ahijevych (2012) found that a misalignment of the mid- pre-TS Nicole. The .72-h pregenesis category includes and low-level circulation centers, due to vertical shear, data from the 10 and 11 September flights into pre-TS made TS Gaston more susceptible to intrusion of dry Karl, the 20 September flight into pre-TS Matthew, and air. They found that Karl and Matthew developed in the 30 September flight into pre-Hurricane Otto. Trop- a moister environment, with mid- to upper-level moisture ical Storm Gaston is a special case. The first mission into increasing with time. An initial vertical misalignment of Gaston occurred at a time when the storm was already the vortex delayed genesis of Karl until the vortex could a named system, and therefore data from this mission subsequently realign. is assigned to the TC category. However, subsequent This study differs from previous studies in that mul- PREDICT flights occurred after Gaston was down- ticase composite vertical profiles of temperature, mois- graded by the NHC to a remnant cyclone. Since Gaston ture, and wind for genesis and nongenesis cases are was no longer a named system during these missions, compared. Additionally, the time evolution of the mean and was given a 70% chance to redevelop by NHC but genesis profile is examined. Parcel-based metrics such as failed to do so, missions into the remnants of Gaston are lifted condensation levels (LCLs), CAPE, and CIN are added to the nongenesis category.

Unauthenticated | Downloaded 09/25/21 07:15 AM UTC FEBRUARY 2013 K O M A R O M I 545

FIG. 1. Map of all dropsonde deployment locations during PREDICT and corresponding genesis categories, from 15 Aug through 30 Sep 2010.

Dropsonde data are composited into a single vertical kilogram. Convective available potential energy will be proﬁle for each group and genesis time category. Com- calculated with positing of temperature T, mixing ratio q, and relative ð ! z T 2 T humidity (RH) involves a simple separation of drop- EL y parcel y CAPE 5 g dz, (1) sondes by genesis or nongenesis categories, interpolating T zLFC y onto a common pressure grid in 5-hPa increments, and averaging. For stability calculations, the virtual tem- where Ty_parcel is the virtual temperature of a surface perature adjustment Ty 5 T(1 1 «q) will be applied, parcel lifted dry adiabatically below the level of free where « 5 0.608 when q is expressed in kilogram per convection (LFC) and moist adiabatically above. The

TABLE 1. Cases of study during PREDICT comprising the genesis, nongenesis, and TC-stage groups with corresponding dates of G-IV deployments. Timing of mission prior to genesis is included for cases in which genesis occurred. Latitude and longitude of the target center location of each drop pattern, zonal phase speed Up of the wave, and the number of dropsondes released are shown.

21 Group Case Date Lat (8N) Lon (8W) Up (m s ) No. dropsondes Nongenesis PGI-27L 17 Aug 14.5 69.3 27.9 23 Nongenesis PGI-27L 18 Aug 16.2 77.2 27.7 24 Nongenesis PGI-30L 21 Aug 20.3 52.9 28.9 16 Nongenesis PGI-30L 23 Aug 20.8 69.4 27.9 12 Genesis (0–24 h) pre-Fiona 30 Aug 14.9 46.8 29.0 30 TC stage TS Fiona 31 Aug 16.0 56.8 27.0 30 TC stage TS Fiona 1 Sep 19.1 62.6 25.1 21 TC stage TS Gaston 2 Sep 13.9 39.7 23.5 20 Nongenesis PGI-38L 3 Sep 15.2 42.6 23.6 22 Nongenesis PGI-38L 5 Sep 17.4 51.3 27.4 21 Nongenesis PGI-38L 6 Sep 15.8 56.9 26.4 23 Nongenesis PGI-38L 7 Sep 16.1 64.3 26.1 22 Genesis (.72 h) pre-Karl 10 Sep (1) 13.5 61.1 24.8 21 Genesis (.72 h) pre-Karl 10 Sep (2) 14.0 61.3 25.2 21 Genesis (.72 h) pre-Karl 11 Sep 15.5 67.8 26.4 22 Genesis (48–72 h) pre-Karl 12 Sep 15.0 72.0 26.0 22 Genesis (24–48 h) pre-Karl 13 Sep 16.8 76.9 26.7 20 Genesis (0–24 h) pre-Karl 14 Sep 18.4 83.6 26.4 21 Genesis (.72 h) pre-Matthew 20 Sep 11.2 58.6 26.4 21 Genesis (48–72 h) pre-Matthew 21 Sep 12.2 62.8 26.9 22 Genesis (24–48 h) pre-Matthew 22 Sep 13.3 70.7 27.1 18 TC stage TS Matthew 24 Sep 14.9 82.6 28.5 22 Genesis (0–24 h) pre-Nicole 27 Sep 19.6 86.9 2.5 23 TC stage TS Nicole 28 Sep 19.5 84.6 2.6 24 Genesis (.72 h) pre-Otto 30 Sep 15.8 58.1 211.3 26

Unauthenticated | Downloaded 09/25/21 07:15 AM UTC 546 JOURNAL OF THE ATMOSPHERIC SCIENCES VOLUME 70

2 equilibrium level (EL) for a virtual surface-based parcel, result in 0.4 and 2.2 m s 1 mean anomalies, respectively. 2 while calculation of CIN will be equivalent to (1) except Similarly, perturbations result in 0.4 and 1.5 m s 1 that integration will be from the lowest level of negative anomalies for Vrad. These results indicate that the wind buoyancy (Ty parcel , Ty) to the LFC (Doswell and metrics used in this study are not particularly sensitive to Rasmussen 1994). possible errors in the chosen center location if these

Vortex-relative tangential (Vtan) and radial (Vrad) errors are approximately 18 latitude or longitude. Errors components of wind are also calculated. Since the on the order of 58 may be more problematic, but are tropical waves are moving within some background flow, unlikely to occur. and winds from the dropsonde data are earth relative, it Vertical wind shear is calculated as the vector wind is necessary to remove the parent wave’s zonal phase difference between the 850–500- and 850–200-hPa pres- speed (Up in vortex-relative wind calculations). Zonal sure levels. Prior to computing the total wind shear vec- phase speed is calculated from a consensus of National tor, an average within each quadrant relative to the Centers for Environmental Prediction (NCEP) Global circulation center is performed in order to eliminate any Forecast System (GFS) and European Centre for Medium- false signal associated with asymmetries in the drop Range Weather Forecasts (ECMWF) analyses and pattern. Quadrant averages are then averaged together. forecasts of meridional wind y and RH (Table 1). The However, vertical asymmetries of the vortex itself are values used in this study were determined operationally not removed from the shear calculation. during PREDICT (available online at http://met.nps. In addition to the previously mentioned vertical mean edu/;mtmontgo/storms2010.html) using Hovmo¨ ller di- profiles, azimuthal averages of temperature, mixing ratio, agrams to track features in these fields. Since meridional relative humidity, tangential winds, and relative vorticity phase speed is generally much weaker than Up for are calculated in order to depict changes in radial struc- westward-moving tropical waves and was not calculated ture of the tropical wave leading up to genesis. Data are operationally during PREDICT, it is not included here. binned in annuli of 100-km radius centered on the nearest

As such, the mean Vtan proﬁle is then computed as the 100-km grid point, with the exception of winds at the sum of the cyclonic (positive) and anticyclonic (nega- center of circulation where Vtan is set to zero at r 5 0 km. tive) contribution of each dropsonde with Up removed, Relative vorticity is computed in radial coordinates as normalized by the total number of dropsondes. For Vrad, the component of wind from each dropsonde blowing 1 ›(rV ) z 5 tan (2) away from (toward) the center of circulation contributes r ›r positively (negatively).

Computation of Vtan and Vrad requires selection of a using a centered difference approximation, where dr 5 center of circulation, which is chosen to be the point at 100 km, beginning at r 5 50 km since 1/r is not defined at which mean 850–700-hPa Vtan is maximized for each r 5 0 km. flight. The methodology follows Marks et al. (1992) with Vertical profiles will be depicted as an anomaly with one exception: since the radius of maximum winds respect to the PREDICT mean profile. Using this (RMW) is poorly defined for many cases in this study, framework, it will be possible to investigate and quantify mean Vtan will be computed with respect to all drop- the differences between the vertical profiles of de- sonde locations, rather than only those within an annu- veloping and nondeveloping tropical systems. While lus around the RMW. Computation of Vtan is performed there are certainly large-scale synoptic, mesoscale, and in 1/108 iterations over a 1083108 latitude–longitude box convective-scale differences between the two scenarios centered on the flight pattern. Dropsondes are distrib- that cannot be captured in a mean dropsonde profile, a uted relatively even in space within 300 km about the number of significant results can nonetheless be drawn. center of circulation (Fig. 2). Farther out, there is Midlevel moisture can often vary immensely over a tendency for greater data coverage to the east and relatively small distances over the spatial extent of a southeast, with less coverage to the west. A sensitivity tropical wave because of a number of factors, including test was performed to examine the sensitivity of com- advection of moist or dry air, drying associated with puted Vtan and Vrad profiles to choice of center location. subsidence, convective moistening from detrainment All center locations are perturbed by 618 and 58 latitude and precipitation processes, or drying associated with and longitude, and Vtan and Vrad are calculated with dry downdrafts. Much of both the top-down and bottom- respect to each new possible choice of center. The up literature note a general trend of increasing convec- magnitude of the wind anomalies, or the difference be- tion near the center of the cyclone. However, averaging tween each perturbed state and the control, is averaged over the full areal extent of any one case might lead to over all cases. For Vtan, center perturbations of 18 and 58 a net cancellation of numerous moistening and drying

Unauthenticated | Downloaded 09/25/21 07:15 AM UTC FEBRUARY 2013 K O M A R O M I 547

FIG. 2. Plots of sounding locations relative to the center of circulation in polar (km, degrees) coordinates for each genesis category: (a) genesis, (b) nongenesis, and (c) TC stage. processes, masking mesoscale variability. Therefore, in windows centered temporally on the mean time of each addition to domainwide averaging, proﬁles of q and RH dropsonde mission to the nearest half hour, and then from dropsondes located within a 150-km radius of the further composited over all cases in each genesis cate- approximate center of circulation will be composited in gory. While accurate estimates of static center locations order to investigate localized moisture anomalies. This were determined from dropsonde data alone, model will only be performed for genesis cases, and of these, data was preferred when attempting to locate the time- only for cases in which the center is well deﬁned. This evolving center location. As such, satellite composites limited dataset includes 8–12 dropsondes for each genesis are centered geographically on the center of the pouch, time block (0–24 h pregenesis, 24–48 h, etc.) consisting of determined by the intersection of the disturbance critical data from the pre-Fiona, pre-Karl, pre-Matthew, pre- line with the axis of the wave trough, from a consensus of Nicole, and pre-Otto missions. GFS and ECMWF analyses interpolated linearly be- Last, Geostationary Operational Environmental Sat- tween analysis times. Similar to Davis and Ahijevych ellite (GOES) infrared data are investigated in order to (2012), convective activity is depicted as a fraction of the relate any dynamic and thermodynamic phenomena total time comprising each category within each 6-h observed in the dropsonde data to the convective period, using half-hourly data in the comoving frame of structure of the tropical waves. GOES cloud-top imag- reference, that a grid box 10 km on a side exhibits an IR ery in full 30-min resolution is composited over 6-h time temperature less than 2508C.

Unauthenticated | Downloaded 09/25/21 07:15 AM UTC 548 JOURNAL OF THE ATMOSPHERIC SCIENCES VOLUME 70

FIG. 3. Composite vertical proﬁles of anomalies relative to the PREDICT mean of (a) temperature, (b) mixing ratio, (c) relative humidity, and (d) tangential component of wind for genesis, nongenesis, and TC stage categories.

It should be emphasized that the results presented of the warming associated with the genesis profile is the herein are only valid for tropical systems in the western beginnings of the development of the warm core. Radial Atlantic basin during the 2010 hurricane season. Whether composites reveal the greatest warm anomalies occur- or not these results can be generalized is yet to be seen. ring within a 200-km radius of the center of circulation for genesis and TC samples, supporting this hypothesis (the development of which is shown in section 3b). Al- 3. Analysis of dropsonde data ternatively, the nongenesis mean profile is associated with cold anomalies of 20.18 to 20.98C from 600–200 hPa. It a. Genesis versus nongenesis should be noted that standard deviations of these data We begin our investigation by comparing T profiles of are large, and only cold anomalies for nongenesis above genesis and nongenesis cases with the PREDICT mean. 500 hPa are greater than one standard deviation from Nongenesis cases are associated with slight warm the PREDICT mean. anomalies of 0.18 to 0.28C below 600 hPa, while the Maximum anomalies of q, both positive and negative, genesis mean is associated with cold anomalies of 20.28 exist between 700 and 500 hPa for genesis, nongenesis, and to 20.38C (Fig. 3a). Above 600 hPa, T anomalies TC stage cases (Fig. 3b). TC and pregenesis profiles are 2 steadily decrease with height in the nongenesis profile, associated with moist q anomalies of 10.1 to 10.5gkg 1 while the genesis profile is instead associated with warm while nondeveloping systems are associated with dry q 2 anomalies. While of a much lower magnitude, 0.28 ver- anomalies of 20.1 to 21.0 g kg 1 in this layer. While sus 0.88C, the greatest positive T anomalies for genesis these values appear to be small, the nongenesis profile is 2 appear within the same layer as for the TC stage profile 25% drier than the PREDICT mean q of 3.24 g kg 1 with 2 between 400 and 200 hPa. It is possible that at least part a 20.81 g kg 1 anomaly at 500 hPa. Radial composites

Unauthenticated | Downloaded 09/25/21 07:15 AM UTC FEBRUARY 2013 K O M A R O M I 549

TABLE 2. PREDICT, genesis, nongenesis, and TC-stage mean mixing ratio values, standard deviations s, and anomaly vs the PREDICT mean for select levels from 1000 to 200 hPa.

2 2 Pressure (hPa) Sounding No. cases Mixing ratio (g kg 1) s (g kg 1) Anomaly vs PREDICT mean 200 PREDICT mean 12 0.02 0.01 — Genesis 5 0.02 0.00 0.00 Nongenesis 3 0.01 0.01 20.01 TC stage 4 0.02 0.01 0.00 500 PREDICT mean 12 3.24 0.75 — Genesis 5 3.41 0.38 0.17 Nongenesis 3 2.43 0.48 20.81 TC stage 4 3.63 0.89 0.39 700 PREDICT mean 12 7.86 0.81 — Genesis 5 8.01 0.56 0.15 Nongenesis 3 7.16 0.57 20.70 TC stage 4 8.20 1.03 0.34 850 PREDICT mean 12 12.78 0.65 — Genesis 5 12.56 0.78 20.22 Nongenesis 3 12.78 0.20 0.00 TC stage 4 13.05 0.73 0.27 1000 PREDICT mean 12 18.68 0.68 — Genesis 5 18.35 0.52 20.33 Nongenesis 3 19.18 0.40 0.50 TC stage 4 18.72 0.88 0.03

2 reveal dry anomalies as large as 21.8 g kg 1 at greater suggests that the developing tropical wave is vulnerable than 400-km radius for nongenesis, with nongenesis to the potential detrimental effects of entrainment on drier than genesis through the center of circulation, parcel buoyancy. Consistent with q results, RH is nota- suggesting an influence of the dry air on the core of the bly lower by 10%–20% from 700 to 300 hPa for non- tropical wave. Interestingly, profiles of q indicate that genesis than for genesis, indicating greater potential for the nongenesis mean value below 850 hPa is more than entrainment of drier air. 2 0.5 g kg 1 moister than the genesis mean value, sug- Values of CAPE reveal an interesting result, in that gesting that perhaps dry air at the midlevels was more not only are the pregenesis and TC profiles no more detrimental to genesis than drier air at the surface dur- unstable than the nongenesis sounding, but the nongenesis ing PREDICT. The PREDICT mean is greater than one profile is in fact much more unstable than the genesis 2 standard deviation moister than the nongenesis mean profile (Table 3). Mean CAPE anomalies are 1336 J kg 1 2 2 from 700 to 500 hPa, suggestive of the relative signifi- for nongenesis, 2171 J kg 1 for genesis, and 242 J kg 1 cance of this dry air (Table 2). for TC cases. All departures are relative to the PREDICT 2 While comparing q profiles allows for direct compar- mean CAPE of 2096 J kg 1 when integrated to the isons of the mass of water vapor in a column of atmo- height of the EL, or the highest available pressure level. 2 sphere, examination of RH (Fig. 3c) is necessary to Large CAPE values of 1900–2500 J kg 1, in conjunction identify near saturation of the lower and midlevels—a sig- with low LFCs of 940–920 hPa and very low CIN of 22 2 nificant criterion for genesis in Bister and Emanuel (1997) to 210 J kg 1 are not unexpected given that calcula- and Nolan (2007). The presence of low ambient RH also tions are with respect to particularly moist surface-based

TABLE 3. Instability data for different categories. Included are the LFC, EL, CAPE, CIN, standard deviations of CAPE, and CAPE anomaly vs the PREDICT mean.

2 2 2 Sounding No. cases LFC (hPa) EL (hPa) CIN (J kg 1) CAPE (J kg 1) CAPE s (J kg 1) Anomaly vs PREDICT mean PREDICT mean 12 928 199 28 2096 539 — Genesis 5 920 200 210 1925 298 2171 Non-genesis 3 940 196 27 2433 314 336 TC stage 4 932 202 27 2054 571 242 .72 h pregenesis 5 924 197 210 2076 117 151 48–72 h pregenesis 2 900 210 216 1550 453 2882 24–48 h pregenesis 2 935 193 28 1980 576 275 0–24 h pregenesis 3 919 204 27 1868 296 2208

Unauthenticated | Downloaded 09/25/21 07:15 AM UTC 550 JOURNAL OF THE ATMOSPHERIC SCIENCES VOLUME 70

FIG. 4. GOES data in 30-min resolution composited over 6-h time windows centered temporally on the mean time of each dropsonde mission to the nearest half hour, composited over multiple missions for each genesis category. The percentage of total time in each 10 3 10 km2 grid point remains below 2508C is depicted for (a) genesis, (b) nongenesis, and (c) TC stage categories (courtesy D. Ahijevych). parcels. In fact, many genesis and nongenesis drop- and Vollaro (2010), that highly sheared, generally weaker sondes exhibited no CIN at all. While it is true that the tropical cyclones tend to be associated with higher height at which the dropsonde is deployed may be below CAPE than their nonsheared, generally stronger coun- the EL in some cases, Ty parcel 2 Ty in (1) is close to zero terparts, as well as the findings of Braun (2010), that for most cases at this altitude, and therefore the amount CAPE generally tends to be higher in environments of of CAPE ‘‘missed’’ should be small. Overall, these re- weakening TCs compared to strengthening TCs. Satel- sults suggest that the availability of additional instability lite imagery suggests that lower CAPE values for gen- in an already otherwise unstable tropical environment esis cases may be due, at least in part, to greater does not increase the likelihood of tropical cyclogenesis, consumption of CAPE by more widespread deep con- which is consistent with Nolan et al. (2007). Smith and vection (Fig. 4). 21 Montgomery (2011) also find greater CAPE associated Average Vtan values are 2–3.5 m s for nongenesis 2 with ex-Gaston than with either genesis case they stud- and 3–5 m s 1 for genesis profiles from the surface ied. These results are also not inconsistent with Molinari through 500 hPa (Fig. 3d). While significant variability

Unauthenticated | Downloaded 09/25/21 07:15 AM UTC FEBRUARY 2013 K O M A R O M I 551

TABLE 4. Vertical wind shear values and standard deviations for 850–500 and 850–200 hPa for genesis, nongenesis, and TC stage categories. Wind shear data for .72 h pregenesis, 48–72 h pregenesis, 24–48 h pregenesis, and 0–24 h pregenesis subsets also included.

2 2 2 2 Group 850–500-hPa shear (m s 1) 850–500-hPa s (m s 1) 850–200-hPa shear (m s 1) 850–200-hPa s (m s 1) Genesis 2.70 0.22 6.97 0.40 Nongenesis 2.22 0.64 7.39 0.71 TC stage 2.35 0.35 6.92 0.52 .72 h pregenesis 2.82 0.14 6.93 0.31 48–72 h pregenesis 2.77 0.10 7.12 0.22 24–48 h pregenesis 2.74 0.04 7.11 0.21 0–24 h pregenesis 2.45 0.31 6.85 0.76

2 exists, with standard deviations greater than 2 m s 1, associated with more consistent deep convection over the developing tropical waves are generally associated a larger spatial area than nongenesis, which, in turn, with stronger circulations. As expected, these circula- would promote the development of a sustained midlevel tions are still signiﬁcantly weaker than those of TC stage vortex. Greater spatial coverage of deep convection is systems. For genesis, nongenesis, and TC cases, Vtan is also consistent with the building of upper-level warm cyclonic from the surface through 300 hPa and anticy- anomalies as a result of additional latent heating, as well clonic above. as humidiﬁcation of the inner core due to moist detrain-

Radial wind Vrad is also calculated (not shown). Max- ment and precipitation from deep convective towers 2 imum negative values of up to 21ms 1 for TC stage preceding genesis. 2 and 20.3 m s 1 for genesis and nongenesis occur between b. Time progression leading up to genesis the surface and 900 hPa, suggesting a shallow layer of inflow and boundary layer convergence. While these A major benefit of the pouch-tracking framework values appear to be small, they are averages over the developed for the PREDICT field campaign was an entire sampled region of the tropical wave, and do not ability to routinely identify and sample regions of poten- correspond to maxima at a particular radius. All three tial genesis daily beginning a few days prior to the de- profiles are associated with positive Vrad above 300 hPa, velopment of a tropical depression. This tracking and indicative of divergence. Time-averaged genesis and sampling technique yielded an impressive temporal nongenesis Vrad profiles are generally indistinguishable, genesis dataset, in which it is possible to subcategorize and Vrad does not appear to be of much value as a dis- genesis profiles by time leading up to genesis. In this criminating characteristic for genesis. However, as will section, we will continue to examine differences between be demonstrated in section 3b, the genesis Vrad profile various mean profiles and the PREDICT mean, but from evolves considerably with time. the perspective of a temporal progression. Midlevel, 850–500-hPa vertical wind shear is slightly Examination of radial profiles of T reveals widespread 2 greater for genesis than nongenesis with 2.70 m s 1 cold anomalies at large lead times for all radii gradually 2 shear plus or minus a standard deviation of 0.22 m s 1 as transitioning to warm anomalies at small lead times 2 compared to 2.22 6 0.64 m s 1 for nongenesis (Table 4). through much of the troposphere (Figs. 5a–d). The .72-h Alternatively, 850–200-hPa deep-layer shear is slightly pregenesis profile is associated with negative T anoma- more hostile for the nongenesis than the genesis cases, lies ranging from 20.18 to 21.08C, with local minima 2 2 with 7.39 6 0.71 m s 1 compared to 6.97 6 0.40 m s 1. between 800–600 and 400–200 hPa. Conversely, the However, wind shear did not appear to be the primary entire 0–24-h pregenesis profile is associated with deep factor in differentiating genesis from nongenesis cases warm anomalies ranging from 10.18C at 1000 hPa to as in 2010 as values of wind shear are statistically identical large as 12.08C at 300 hPa within 100 km of the center between the two cases. Given the high percentage of of circulation. The most obvious warming trend occurs dropsondes deployed within 400 km of the center of from 400 to 200 hPa, evident during every successive 24-h circulation and comparatively few outside of this region, time increment. However, cold anomalies persist from it is also possible that these data are not fully repre- 900 to 700 hPa through 24–48 h pregenesis, until finally sentative of the true environmental wind shear. warming rapidly 0–24 h pregenesis. Warm anomalies GOES composites clearly demonstrate persistently are observed from 500 to 200 hPa above cold anomalies colder cloud tops over a much larger area for genesis below 500 hPa from 24–48 and 48–72 h pregenesis, than nongenesis (Fig. 4). In fact, the area of cloud tops characteristic of the development of a low-level cold colder than 2508C more than 60% of the time is larger core prior to genesis as described by Bister and Emanuel for genesis than our sample of TCs. Genesis is clearly (1997) and Nolan (2007). The observed upper-level

Unauthenticated | Downloaded 09/25/21 07:15 AM UTC 552 JOURNAL OF THE ATMOSPHERIC SCIENCES VOLUME 70

FIG. 5. Radial profiles of temperature anomalies (8C) with respect to the PREDICT mean. Data are azimuthally averaged in annuli of 100-km radius for (a) .72 h pregenesis, (b) 48–72 h pregenesis, (c) 24–48 h pregenesis, and (d) 0–24 h pregenesis. warming can likely be attributed to a combination of latent with, even four or more days in advance. In a separate heat released by convection and warm-core development calculation, q and RH are composited for only a small via subsidence induced by this latent heating. The pre- subset of the total sample, only including 8–12 dropsondes dominant warm anomalies 0–24 h pregenesis are maxi- for each 24-h period within a 150-km radius of the center mized near 300 hPa, at a similar altitude to what was of circulation. Local averaging reveals a moistening trend 2 found in La Seur and Hawkins (1963) and Hawkins and (Fig. 6a), with q increasing from 10.5 to 10.7 g kg 1 at 2 Rubsam (1968). These results suggest that warm core .72 h pregenesis to 11.2 to 11.6 g kg 1 0–24 h pre- development commences prior to genesis and at the same genesis between 800 and 500 hPa, with the greatest in- altitude one would expect to find the warm core in crease in moisture occurring 24–48 h pregenesis. Radial a mature TC. There even exists some hint of a secondary cross sections of q azimuthally averaged in annuli reveal T maximum developing between 500 and 600 hPa within similar trends (not shown). This observation is consis- 200 km of the center of circulation, as suggested by tent with Nolan (2007) in that moist detrainment and Hawkins and Imbembo (1976) and Stern and Nolan precipitation from deep convective towers are acting (2012), although much weaker than the primary warm locally to humidify the center of the wave, although the core. greater low-level moistening observed here suggests a In contrast with temperature, neither q nor RH in- greater fraction of moist detrainment associated with creases on average with time. Mixing ratio differences cumulus congestus type clouds. The fact that this trend between the .72-h pregenesis and 0–24-h pregenesis does not appear in the full composite may simply be cases tend to be small, demonstrating low variability with reflective of the larger area whose profile is difficult to no evident large-scale humidification. However, the full modify given the small area occupied by convection. wave mean for cyclogenesis events during PREDICT is This result is also consistent with pregenesis moistening simply more moist than for nongenesis events to begin local to the inner 60 km above 2 km observed in the

Unauthenticated | Downloaded 09/25/21 07:15 AM UTC FEBRUARY 2013 K O M A R O M I 553

FIG. 6. Composite vertical profiles of anomalies relative to the PREDICT mean of (a) mixing ratio within 150 km of the center of circulation, (b) relative humidity within 150 km of the center of circulation, (c) radial component of wind at all radii, and (d) tangential component of wind at all radii for .72 h pregenesis, 48–72 h pregenesis, 24–48 h pregenesis, and 0–24 h pregenesis categories. numerical experiments of Bister and Emanuel (1997). A pregenesis as the boundary layer warms and moistens threshold of 60-km radius is not used here, as decreasing rapidly enough to offset continued warming at the up- the radius by 40 km reduces the number of dropsondes per levels, followed by a small decrease in CAPE again encompassed by about 75%, which leads to a very small 0–24 h pregenesis. sample size. While Bister and Emanuel (1997) and Nolan Midlevel, 850–500-hPa wind shear gradually subsides 2 2 (2007) suggest that a similar local moistening trend from 2.82 m s 1 .72 h pregenesis to 2.45 m s 1 0–24 h should be evident in the RH field, one was not observed pregenesis (Table 4), although both values are associ- in the limited dataset of this study (Fig. 6b). Instead, a ated with very low shear and are highly favorable for very small increase RH from 800 to 600 hPa is observed tropical cyclogenesis. Upper-level, 850–200-hPa wind 24–48 h pregenesis, followed by a decrease in RH 0–24 h shear does not depict a coherent signal, fluctuating be- 2 pregenesis. Overall, the added moisture content nearly tween 6.8 and 7.2 m s 1 between pregenesis time bins. exactly cancels with the increased saturation deficit as- However, as was the case for the comparison between sociated with warming temperatures such that the profile genesis and nongenesis shear values, these values are of RH remains relatively steady. statistically indistinguishable.

A time progression of CAPE reveals maximum in- Profiles of Vrad indicate that mean 300–200-hPa upper- stability .72 h pregenesis (Table 3), at a time when level outflow fluctuates significantly with time, but with upper-level temperatures are the coldest. A sudden no obvious trend (Fig. 6c). It would appear that level of 2 decrease in CAPE by approximately 500 J kg 1 occurs organization of divergent flow aloft is not critical to any 2 48–72 h pregenesis as upper levels warm. The mean particular stage of genesis. Radial winds of 11.5 m s 1 sounding becomes slightly more unstable again 24–48 h at 200 hPa are observed both .72 h pregenesis and at

Unauthenticated | Downloaded 09/25/21 07:15 AM UTC 554 JOURNAL OF THE ATMOSPHERIC SCIENCES VOLUME 70 the time of genesis, indicating that upper-level condi- evolution. To further investigate this issue, the time pro- tions were favorable for genesis several days before it gression of relative vorticity is computed in radial co- occurred, but the tropical waves required additional ordinates relative to the center of circulation (Figs. 7a–d) time to organize. At lower levels, positive values of Vrad from tangential wind data binned in annuli of 100-km 2 of 10.5 to 11ms 1 are observed below 850 hPa for radius. As expected, the greatest values of z are observed both the .72-h pregenesis and the 48–72-h pregenesis within r , 100 km near the center of circulation at all composites, indicating surface divergence. This suggests times. The initial altitude of the z maximum is not well that, while upper-level outflow was favorable for gene- defined .721 h pregenesis, with relatively weak local sis, lack of surface convergence may have hindered the maxima between the surface and 700 hPa (Fig. 7a). process. While not evident in these composites, Davis By 48–72 h pregenesis, vorticity has amplified consid- and Ahijevych (2012) suggested that an initial vertical erably through a deep layer from 925 to 500 hPa, with misalignment of vortex centers may have also delayed the greatest amplification between 900 and 600 hPa genesis of Karl. By 24–48 h pregenesis, radial outflow (Fig. 7b). Vorticity remains only marginally cyclonic at 2 2 reverses to 21ms 1 inflow, increasing to 22ms 1 0–24 h the surface, likely because of a combination of pre- pregenesis. While it is likely that some of the large-scale viously demonstrated low-level divergence occurring in surface divergence 48–72 h pregenesis is associated with a region of cold anomalies and surface friction. Relative convective or mesoscale downdrafts, there does not vorticity continues to strengthen at all levels below appear to be sufficient dry air near the core of the system 400 hPa within 24–48 and 0–24 h pregenesis, although at any time for dry air induced downdrafts to be a pro- what was previously a broad region of maximum z from hibitive factor for genesis (Fig. 6b). Thereafter, a second- 900–600 hPa has evolved to become a distinct maxi- ary circulation begins to develop within 48 h of genesis, mum slightly near 800 hPa (Figs. 7c,d). While we find no and surface convergence increases rapidly with time. evidence of vorticity descending, a clear intensification Tangential winds (Fig. 6d) progress from weaker to of midlevel vorticity prior to the development of a ro- stronger and more cyclonic with time above 600 hPa. bust surface circulation is more consistent with a Bister This result is consistent with the progressive building of and Emanuel (1997), Nolan (2007), and Raymond et al. a mid- to low-level vortex described by Nolan (2007) and (2011) genesis framework. Raymond et al. (2011). However, the region of maxi- The satellite presentation of the temporal progression mum Vtan exists between 850–550 hPa, depending upon leading up to genesis is not as intuitive as the satellite the timeframe examined, and there is no well-defined comparison between genesis and nongenesis. From peak. The altitude of the tangential wind maximum is of .72 h pregenesis to 48–72 h pregenesis, there is a sig- lower altitude than the warm core, consistent with La nificant expansion of the persistent cloud tops colder Seur and Hawkins (1963), Hawkins and Rubsam (1968), than 2508C, indicating greater coverage of deep convec- Hawkins and Imbembo (1976), and Stern and Nolan tion and presumably a more mature system approaching (2012). While it is difficult to determine an exact thresh- genesis (Figs. 8a,b). However, during subsequent 24–48- old, genesis appears imminent when systemwide deep- and 0–24-h pregenesis time increments, the coverage of 2 layer positive tangential wind anomalies of 6–7 m s 1 deep convection diminishes considerably, both spatially develop between 850 and 700 hPa, with much weaker and temporally (Figs. 8c,d). Despite the fact that signifi- values observed 24 or more hours earlier. Below 700 hPa, cant changes are ongoing within the pregenesis vortex, the strength of the mean tangential wind actually fluctu- including the development of a warm core, moistening ates from 72–24 h pregenesis, followed thereafter by an of the core, and amplification of system relative vorticity, abrupt increase 24 h pregenesis. This is consistent with the organization and persistence of temporally averaged Nolan (2007) in that the transition to the intensification 2508C cloud tops is no more indicative of genesis 0–24 h stage can be sharp, having less to do with a continuous pregenesis than it is .72 h pregenesis. This result strengthening of the mean wind and perhaps more to do demonstrates the obvious advantage of having available with moistening of the core. These results show a 24-h lag in situ data when assessing how close a tropical distur- between the greatest increase in moisture 24–48 h pre- bance is to developing into a tropical cyclone. genesis and a strengthening of the vortex 0–24 h pregenesis. It is particularly striking that this sudden 4. Conclusions increase in organization is evident in the mean profile of five sampled systems. Observations from the 2010 PREDICT field cam- While average tangential wind profiles depict a tem- paign, when analyzed from a composite mean frame- poral progression of the strengthening vortex, they do not work, offer discernible differences between developing clearly demonstrate a top-down or bottom-up genesis and nondeveloping tropical waves that may be

Unauthenticated | Downloaded 09/25/21 07:15 AM UTC FEBRUARY 2013 K O M A R O M I 555

21 FIG. 7. Relative vorticity (s ) computed in radial coordinates for (a) .72 h pregenesis, (b) 48–72 h pregenesis, (c) 24–48 h pregenesis, and (d) 0–24 h pregenesis. advantageous to the understanding and prediction of than the primary warm core found at higher altitude. In tropical cyclogenesis. Temperature, mixing ratio, rela- contrast with the genesis cases, negative T anomalies of tive humidity, radial and tangential components of wind, 20.58 to 21.08C exist from 500 to 200 hPa for non- relative vorticity, and CAPE are examined. developing systems. Temperature profiles reveal a progressive building of In terms of moisture, positive q anomalies of 10.1 2 warm anomalies from 500 to 200 hPa, relative to the to 10.5 g kg 1 from 800 to 300 hPa are observed in PREDICT mean, of 10.58 to 11.08C at 24–48 h pre- developing systems, even 72 or more hours pregenesis. genesis, increasing to 11.0 to 12.08C 0–24 h pregenesis Moisture does not increase significantly with time on the within 200 km of the center of circulation. While the spatial scale of the entire tropical wave. Meanwhile, existence of a warm core in mature TCs has been well nondeveloping systems are associated with significant established in previous literature, the magnitude and dry anomalies from 800 to 300 hPa. When only exam- timing of the warm-core development with respect to ining dropsondes located within 150 km from the center time of genesis has not. The observation of maximum of circulation, moist convective processes appear to in- warm anomalies just below tropopause-level pregenesis crease moisture as the tropical wave approaches genesis, suggests that warm-core development occurs at the as suggested by Bister and Emanuel (1997), Nolan (2007), same altitude as observed in mature TCs by La Seur and others. The maximum increase in moisture of 2 and Hawkins (1963) and Hawkins and Rubsam (1968). 1gkg 1 from 800 to 600 hPa occurs 24–48 h pregenesis. A local maximum in warm anomalies below 500 hPa This trend is likely washed out when all dropsondes are also suggests that formation of a secondary warm core included because of the large spatial area of averaging in is possible pregenesis, at a similar altitude as the the full composite, possibly coupled with some large- Hawkins and Imbembo (1976) secondary warm core. scale entrainment of dry air into the wave circulation. This is also consistent with the level of the Stern and Nonetheless, the full q composite still demonstrates Nolan (2012) warm core, although it is not the primary that time-evolving genesis profiles are all significantly warm core as they suggest. It should be noted that any moister than nondeveloping systems, even more than 72 h presence of a secondary warm core is much weaker prior to genesis. Nongenesis RH profiles are on the order

Unauthenticated | Downloaded 09/25/21 07:15 AM UTC 556 JOURNAL OF THE ATMOSPHERIC SCIENCES VOLUME 70

FIG. 8. GOES data, as in Fig. 4 but for (a) .72 h pregenesis, (b) 48–72 h pregenesis, (c) 24–48 h pregenesis, and (d) 0– 24 h pregenesis (courtesy D. Ahijevych). of 10%–20% drier than the PREDICT mean from 700 greatest increase in moisture by 24 h. Radial wind pro- to 500 hPa, suggesting a greater potential for dry air files suggest that many cases of genesis may have been entrainment into convective towers. Conversely, the delayed by low-level outflow. Alternatively, an initial nongenesis mean is actually moister than the genesis stage of low-level outflow may instead be a necessary mean from the surface through 850 hPa, possibly sug- first step prior to genesis, induced by cool sinking air in gesting that dry air at the midlevels is more detrimental the stratiform precipitation region as in Bister and to genesis than dry air at the low levels. Emanuel (1997). During the final 48 h before genesis, 2 Examination of the wind field reveals a progressive low-level inflow of 1–2 m s 1 develops and strengthens strengthening of the vortex above 600 hPa, with an ini- with time. tial delay in intensification from 850 to 700 hPa. Tan- Vorticity fields reveal a broad region of maximum z gential wind at these levels fluctuates between 3 and from 900 to 600 hPa at 48–72 and 24–48 h pregenesis. 2 5ms 1 from 72 through 24 h pregenesis, before jump- This feature appears simultaneously with low-level cold 2 ing suddenly to 6–7 m s 1 less than 24 h pregenesis. This anomalies 24–72 h pregenesis, as well as divergence sudden intensification of the vortex appears to lag the 48–72 h pregenesis. Thereafter, a distinct z maximum

Unauthenticated | Downloaded 09/25/21 07:15 AM UTC FEBRUARY 2013 K O M A R O M I 557 develops near 800 hPa 0–24 h pregenesis. This evolu- REFERENCES tion of the vortex is consistent with a process described Bister, M., and K. A. Emanuel, 1997: The genesis of Hurricane by Bister and Emanuel (1997), in which the level of Guillermo: TEXMEX analyses and a modeling study. Mon. maximum PV production descends as the level of peak Wea. Rev., 125, 2662–2682. cooling descends in the stratiform rain region. Ritchie Braun, S. A., 2010: Reevaluating the role of the Saharan air layer in and Holland (1997) and Simpson et al. (1997) propose Atlantic tropical cyclogenesis and evolution. Mon. Wea. Rev., 138, 2007–2037. asimilarmechanismbywhichamidlevelvortexin- ——, M. T. Montgomery, K. J. Mallen, and P. D. Reasor, 2010: duces a surface circulation through vertical penetration Simulation and interpretation of the genesis of Tropical Storm and vortex stretching. While vertically descending Gert (2005) as part of the NASA Tropical Cloud Systems and vorticity was not identified, the development of a ro- Processes Experiment. J. Atmos. Sci., 67, 999–1025. bust midlevel vortex prior to the intensification of Davis, C. A., and D. A. Ahijevych, 2012: Mesoscale structural a surface vortex was more consistent with these studies, evolution of three tropical weather systems observed during PREDICT. J. Atmos. Sci., 69, 1284–1305. along with Nolan (2007) and Raymond et al. (2011), ——, C. Snyder, and A. Didlake, 2008: A vortex-based perspective than with studies that suggest a bottom-up genesis of eastern Pacific tropical cyclone formation. Mon. Wea. Rev., mechanism. Differences between genesis and non- 136, 2461–2477. genesis Vtan and z also reveal that developing waves Doswell, C. A., and E. N. Rasmussen, 1994: The effect of neglecting are, on average, associated with a slightly stronger cir- the virtual temperature correction on CAPE calculations. Wea. Forecasting, 9, 625–629. culation than nondeveloping waves. Dunkerton, T. J., M. T. Montgomery, and Z. Wang, 2009: Tropical Genesis cases are associated with slightly greater cyclogenesis in a tropical wave critical layer: Easterly waves. 850–500-hPa midlevel wind shear than nongenesis ca- Atmos. Chem. Phys., 9, 5587–5646. 2 ses: 2.70 versus 2.22 m s 1. On the other hand, genesis Elsberry, R. L., and P. A. Harr, 2008: Tropical Cyclone Structure cases are associated with slightly lower 850–200-hPa (TCS08) field experiment science basis, observational plat- 2 deep-layer wind shear: 6.97 compared to 7.39 m s 1. forms, and strategy. Asia-Pac. J. Atmos. Sci., 44, 209–231. Emanuel, K. A., 2005: Divine Wind: The History and Science of Midlevel wind shear also gradually decreases with time Hurricanes. Oxford University Press, 285 pp. 21 from 2.82 to 2.45 m s during the time progression Gray, W. M., 1968: Global view of the origin of tropical distur- toward genesis. However, in both cases, differences in bances and storms. Mon. Wea. Rev., 96, 669–700. wind shear are statistically indistinguishable. ——, 1979: Hurricanes: Their formation, structure, and likely role Last, calculation of virtual CAPE indicates signifi- in the tropical circulation. Meteorology over Tropical Oceans, D. B. Shaw, Ed., Royal Meteorological Society, 155–218. cantly greater instability associated with nongenesis Halverson, J., and Coauthors, 2007: NASA’s Tropical Cloud Sys- profiles than with either pregenesis or TC-stage profiles. tems and Processes Experiment: Investigating tropical cyclo- 21 Results suggest that 2000 J kg of CAPE may be suf- genesis and hurricane intensity change. Bull. Amer. Meteor. ficient for tropical cyclogenesis, and additional in- Soc., 88, 867–882. stability does not aid in the genesis process. Hawkins, H. F., and D. T. Rubsam, 1968: Hurricane Hilda, 1964. II. While other recent studies have examined and Structure and budgets of the hurricane on October 1, 1964. Mon. Wea. Rev., 96, 617–636. compared individual cases sampled during PREDICT, ——, and S. M. Imbembo, 1976: The structure of a small, intense we have presented an alternative perspective in com- hurricane—Inez 1966. Mon. Wea. Rev., 104, 418–442. paring genesis to nongenesis cases via creating com- Hendricks, E. A., M. T. Montgomery, and C. A. Davis, 2004: The posite vertical profiles for all sampled tropical waves, as role of ‘‘vortical’’ hot towers in the formation of Tropical well as examining the day-to-day evolution of multi- Cyclone Diana (1984). J. Atmos. Sci., 61, 1209–1232. Houze, R. A., Jr., and Coauthors, 2006: The Hurricane Rainband case pregenesis composites. Further in-depth inves- and Intensity Change Experiment: Observations and modeling of tigation of tropical waves with new aircraft data and Hurricanes Katrina, Ophelia, and Rita. Bull. Amer. Meteor. Soc., corroboration with model analyses could potentially 87, 1503–1521. increase the robustness of these results. ——, W.-C. Lee, and M. M. Bell, 2009: Convective contribution to the genesis of Hurricane Ophelia (2005). Mon. Wea. Rev., 137, 2778–2800. La Seur, N. E., and H. F. Hawkins, 1963: An analysis of Hurricane Cleo (1958) based on data from research reconnaissance air- Acknowledgments. The author acknowledges fund- craft. Mon. Wea. Rev., 91, 694–709. ing from the National Science Foundation, Grant Mapes, B. E., and R. A. Houze Jr., 1995: Diabatic divergence ATM-0848753. The author is grateful to professors profiles in western Pacific mesoscale convective systems. Sharan Majumdar, Brian Mapes, and David Nolan for J. Atmos. Sci., 52, 1807–1828. their comments and advice throughout the study. The Marks, F. D., R. A. Houze, and J. F. Gamache, 1992: Dual-aircraft investigation of the inner core of Hurricane Norbert. Part I: author would also like to thank Chris Davis for his Kinematic structure. J. Atmos. Sci., 49, 919–942. insight and suggestions, as well as Dave Ahijevych for Molinari, J., and D. Vollaro, 2010: Distribution of helicity, CAPE, providing the GOES composites. and shear in tropical cyclones. J. Atmos. Sci., 67, 274–284.

Unauthenticated | Downloaded 09/25/21 07:15 AM UTC 558 JOURNAL OF THE ATMOSPHERIC SCIENCES VOLUME 70

Montgomery, M. T., M. E. Nicholls, T. A. Cram, and A. Saunders, ——, S. L. Sessions, and C. Lo´ pez-Carillo, 2011: Thermodynamics 2006: A ‘‘vortical’’ hot tower route to tropical cyclogenesis. of tropical cyclogenesis in the northwest Pacific. J. Geophys. J. Atmos. Sci., 63, 355–386. Res., 116, D18101, doi:10.1029/2011JD015624. ——, and Coauthors, 2012: The Pre-Depression Investigation of Ritchie, E. A., and G. J. Holland, 1997: Scale interactions during Cloud-Systems in the Tropics (PREDICT) experiment: Sci- the formation of Typhoon Irving. Mon. Wea. Rev., 125, 1377– entific basis, new analysis tools, and some first results. Bull. 1396. Amer. Meteor. Soc., 93, 153–172. Simpson, J., E. A. Ritchie, G. J. Holland, J. Halverson, and Nolan, D. S., 2007: What is the trigger for tropical cyclogenesis? S. Stewart, 1997: Mesoscale interactions in tropical cyclone Aust. Meteor. Mag., 56, 241–266. genesis. Mon. Wea. Rev., 125, 2643–2661. ——, E. D. Rappin, and K. A. Emanuel, 2007: Tropical cyclogen- Smith, R. K., and M. T. Montgomery, 2011: Observations of the esis sensitivity to environmental parameters in environments convective environment in developing and non-developing of radiative–convective equilibrium. Quart. J. Roy. Meteor. tropical disturbances. Quart. J. Roy. Meteor. Soc., 137, Soc., 133, 2085–2107. 1–20. Palmen, E. H., 1948: On the formation and structure of tropical Stern, D. P., and D. S. Nolan, 2012: On the height of the warm core cyclones. Geophysica, 3, 26–38. in tropical cyclones. J. Atmos. Sci., 69, 1657–1680. Raymond, D. D. J., C. Lo´ pez-Carillo, and L. Lo´ pez Cavazos, 1998: Zipser, E. J., and Coauthors, 2009: The Saharan air layer and the Case-studies of developing east Pacific easterly waves. Quart. fate of African easterly waves—NASA’s AMMA field study of J. Roy. Meteor. Soc., 124, 2005–2034. tropical cyclogenesis. Bull. Amer. Meteor. Soc., 90, 1137–1156.

Unauthenticated | Downloaded 09/25/21 07:15 AM UTC