CHAPTER 4
UPPER-LEVELS
Evans and Prater-Mayes (2004) examined the extratropical transition of Hurricane Irene (1999) and concluded that an upper-level jet streak and trough contributed to the cyclogenesis ahead of the storm and affected the post-transition intensification. This supposition will be explored with Kate (2003) and Fabian (2003). The interaction with the midlatitude trough is more fully explored in Chapter 5 with the aid of vector frontogenesis functions. Here, the upper-level features near jet are examined in the PV framework of Agusti-Panareda et al. (2004) as well as the cyclone-lifecycle paradigms. The ET of Irene (1999) is well documented by the authors mentioned earlier, but the extratropical phase is not. Thus, attention will be focused on the evolution of the upper-level pattern with special emphasis paid to the resulting cyclone/frontal structure.
4.1 Hurricane Irene (1999)
The potential vorticity (PV) framework above developed by Agusti-Panareda et al. (2004) discussed the extratropical transition of Hurricane Irene (1999). They concluded that indeed the upper-level outflow associated with the transformation of Irene enhanced divergent flow downstream and resulted in significant extratropical development. The analysis that follows will focus more upon the upper-level jet changes, evolution of PV and thermal anomalies near the tropopause, and interaction with the midlatitude trough. The effects on the baroclinic zone will be briefly mentioned since the cross-sections allow for easy inspection of its evolution. Additional attention will be paid to the extratropical stage of Irene and the subsequent mature stage, which is indicative of a Shapiro-Keyser (1990) warm core seclusion and LC2 baroclinic lifecycle. Isentropic PV maps on the 325 K surface and vertical cross-sections both zonally and meridionally oriented will best show the interaction between the transforming vortex and the midlatitude circulation.
4.11 Jet
The upper-level jet at 300 hPa (Fig. 4.1) is characterized as high zonal index with anticyclonic shear poleward of the strong subtropical high over the central Atlantic. Irene reaches the right entrance region of the jet at 19/00z (Fig. 4.1a) and remains in a favorable area for extratropical cyclogenesis with dynamically forced ascent. However, the strength of this jet is preconditioned by the approaching tropical cyclone, which is visible on north-south vertical cross-sections of PV and normal component winds (Fig. 4.2). This depiction nicely captures the interaction between baroclinic zone, upper-level PV anomalies, and upper-level jet all situated northward of Irene at about 50° N. At 17/12z (Fig. 4.2a), the moist convective tower of Irene is less that 1500 km away from the baroclinic zone and the associated positive PV anomaly of the trough. During the next 48 hours Fig. 4.2b,c), the upper-level jet remains strong with speeds > 70 ms-1 primarily in the zonal direction (Fig. 4.1a). As the tower approaches associated with the dipping isentropes indicative of a warm thermal anomaly and high PV air, its outflow induces a negative PV anomaly downstream or poleward shown as a concave-up bump along the tropopause at about 42° N at 18/12z (Fig. 4.2b) and 46° N at 19/00z (Fig. 4.2c). As explained in the preamble to this PV analysis, a meridionally oriented PV dipole steepens the tropopause and strengthens the jet. Irene’s lateral interaction with the trough’s positive PV anomaly occurs to the northwest and is partially represented at 19/00z (Fig. 4.3a).
There are two positive potential-vorticity (PV) anomalies: one at low-levels from latent-heat release due to heavy rainfall in moist warm-core ascent and the other at upper-levels associated with a stratospheric air streamer or fold above the warm core. Hoskins et al. (1985) described this “action at a distance” baroclinic development as dominated by the mutual interaction of upper-level PV anomaly and a surface temperature anomaly, which acts like a positive PV anomaly itself. This interaction is facilitated by the latent heat release, which reduces the static stability and thus increases the Rossby penetration depth (Browning et al. 1998).
As Irene couples with the midlatitude circulation at approximately 19/00z, a definitive upstream tilt with height develops indicative of the potential for baroclinic instability. With dry air intruding from the north into the southwestern quadrant of Irene, strong convection and latent heat release continues along the baroclinic zone associated with the warm frontogenetic region. As Irene approached, the positive PV anomaly associated with the trough along the baroclinic zone increases in magnitude and spatial scale closer to the surface. Diabatically forced convection continues in this region (45°-50° N Fig. 4.2c) due to the remarkable low-level jet (>30 ms-1) advection of warm, moist air (high θe content Fig. 4.3a) up and over the baroclinic zone. Irene amplifies this baroclinic zone on its own with the poleward advection of warm moist air and the analogous equatorward cold advection. As Irene superimposed its deformation pattern upon the baroclinic zone, warm and cold advection patterns were markedly accentuated. The warm conveyor belt (Carlson 1991) is an important player in the resurrection of the PV tower, this time more identifiable as an extratropical frontal process as the tropical convection is eroded. With the upstream tilt to the NW of the new tower at 19/12z (Fig. 4.3b), rapid cyclogenesis commences.
A time series of isentropic PV maps along the 325 K surface (Fig. 4.4) illustrates the process outlined above in terms of lateral interactions near the jet level. The 325 K isentrope, as seen in the latitudinal and zonal vertical cross-sections, intersects the jet streak at around 300 hPa, a fair representation of the midlatitude circulation (Bluestein 1992). From 19/00z to 19/12z, it is clear that a LC2 cyclonic wrap begins as Irene moves onto the cyclonic shear side of a developing upstream jet streak (Fig. 4.1c). High-PV air with stratospheric origin begins to fold around the circulation. This frontal fracture, indicative of a Shapiro-Keyser cyclone lifecycle, is commensurate with the initialization of the LC2 cyclonic wrap-up. The frontal fracture in this case is associated with the dramatic intensification of the upstream jet due to a marked PV gradient. This advection of cold, dry stratospheric (high PV) air associated with the high-PV tongue trails Irene from 19/12z (Fig. 4.1c) to 19/18z (Fig. 4.1d). In this case, Irene actually enters into an anticyclonic shear environment south of the jet streak ahead of the midlatitude trough. This is not a particularly favorable configuration for the consequent LC2 cyclonic wrap, but more of a LC1 trough thinning variety. Thus, it is hypothesized that the positive PV anomaly associated with the decaying tropical PV tower and the rapidly developing extratropical tower facilitate large-scale descent upstream in the trough region. The development of a frontal fracture allows high-PV from the upper-level trough to spill into the cyclone core. This in turn drives the intensification of the jet streak along the southern edge of the high-PV intrusion. Browning et al. (1998) noticed similar dry air intrusion with Lili (1995) due to a favorable trough interaction (Hanley et al. 2001). Frontal fracture will also be addressed in terms of its impact on vector frontogenesis and evolution of frontal features at midlevels and near the surface in Chapters 5 & 6.
4.12 Frontal Fracture and Cyclonic Roll-up
Matano and Sekioka (1971), Brand and Guard (1978), and Klein et al. (2000) among others have studied extratropical transition and identified dramatic dry air intrusions from upper-levels spilling into the cyclone core characteristic of LC2 lifecycles and frontal fractures. This dry intrusion (low-wet bulb θw air) is representative of descending stratospheric air to about 400 hPa. Again considering the vertical cross-sections, upstream high-PV air (blue contour > 5 PVU Fig. 4.3b) descends to about the 400-hPa level. The north-south viewpoint (Fig. 4.2e) fails to capture the intrusion since it is upstream but it does highlight the intensification of the upstream jet streak, which is becoming more zonally oriented at the base of the LC2 cyclonic wrap-up (Fig. 4.4d). It is interesting to note the timing of this intrusion. Browning et al. (1998) and Agusti-Panareda et al. (2004) both indicate that the dissolution of the tropical diabatic PV tower directly preceded the intrusion of high-PV air, which reinvigorated a new tower and commenced rapid cyclogenesis. The phasing with the midlatitude circulation in the framework of positive trough interaction as described by Hanley et al. (2001) seems to be the key to understanding exactly when the frontal fracture will occur and subsequent LC2 cyclonic wrap-up.
As mentioned before, the right entrance and left exit regions of a jet streak are favorable for cyclogenesis. At 19/06z, Irene is accelerated into a superposition region where the two regions overlap both upstream and downstream. However, as the positive PV tower approaches the upper-level positive anomaly downstream, the tropopause flattens and the PV gradient decreases (Fig. 4.2e,f). The tail end of the downstream jet weakens at 19/18z with the strongest isotachs rounding the base of another far downstream trough (Fig. 4.1d). The advection of the PV tower along the tongue of the 325 K PV sheet to the northeast quickly abates and Irene translates on a due west course. Where the upper-level PV gradient is greatest, the equatorward jet intensifies to > 70 ms-1 in a zonal direction.
The advantage of the vertical cross-sections in the north-south and east-west direction is apparent as the normal winds are easily analyzed. The poleward exit region of the jet is actually advected around the PV tongue at 20/00z (Fig. 4.4f) and seen on the east-west cross-section (Fig. 4.3c) as a poleward jet flow due to a steepened tropopause associated with the cyclonic wrap-up of high-PV air. [This weak poleward flow then cyclonic wraps around the northern periphery of the broad PV tower and helps facilitate the development of a broad back-door cold front as seen on the surface analysis (Fig.5.5).] The north-south cross-sections capture the intensification of the equatorward jet streak, which increases from 60 ms-1 at 19/18z (Fig. 4.2f) to well over 70 ms-1 at 20/00z (Fig. 4.2g). At the latter time, Irene is at its deepest central pressure of 944 hPa, well below its minimum pressure of 960-hPa as a category 2 hurricane off the North Carolina coast at 18/12z. The maximum sustained winds were 100 knots (NHC) at this time and fairly constrained to the center of circulation. It is apparent from the N-S cross-sections that strong winds in the extratropical stage are occurring beneath the upper-level jet, which is situated along bent-back warm front baroclinic zone (Chapter 6) at about 900 hPa with >50 ms-1 winds. This low-level jet is addressed later with QuikScat wind observations and satellite representations.
4.13 Warm Core Seclusion
The PV tower in both cross-section representations is no longer tilted at its mature stage (lowest sea level pressure) at 20/00z (Fig. 4.2g, 4.3c). Baroclinic instability wanes and is replaced by barotropic decay with the maximum upper-level stratospheric PV anomaly moves directly overhead of the surface low as seen on the 325 K isentropic surface at 20/00z and 20/12z (Fig. 4.4g,h). The PV magnitude begins to slowly decrease beneath as diabatic PV creation winds finally due to the effects of the dry-air intrusion. Browning et al. (1998) describe this cyclonic wrap-up that eventually encircles the cyclone core as an extensive three-dimensional PV sheet. At upper-levels (~300-hPa), the area of maximum PV is located in the stratosphere and extends in a curved shape past the cyclone. At lower-levels (~600 hPa), the PV sheet remains nominally in the troposphere and wraps inward around the cyclone center. The cyclonic wrap-up is able to occur due to the greater translational speed of the PV sheet as compared to the storm motion as seen with Lili by Browning et al. (1998). By 21/00z (Fig. 4.4i), the PV aloft decreases in magnitude and expands spatially and interacts with a downstream upper-level PV anomaly over the British Isles. LC2 lifecycles take a while to slowly wind down in terms of central pressure and winds. It is also interesting to point out as a counterexample of LC1-type trough-thinning or equatorward Rossby-wave breaking occurring over the southeast United States associated with a digging trough. The winds associated with the warm core seclusion and bent-back warm front responsible for its development will be addressed later.
4.2 Hurricane Kate (2003)
The upper-level jet at 07/00z was highly amplified with Kate approaching the favorable right entrance region (Fig. 4.5a). The low-zonal index is a product of a strong subtropical high located over the central Atlantic and an equatorward-digging trough over eastern Canada, which produced a high-vertical shear environment. This facilitated the rapid ET of Kate in less than 24-hours between 06/00z and 07/00z as seen from the cyclone phase diagnostics (Fig. 6.2). However, the lateral interaction with the upper-level trough did not occur until later at 07/12z with only limited PV anomaly contact. This delayed the eventual extratropical cyclogenesis that deepened Kate from its final tropical minimum SLP of 987-hPa at 07/00z to 968-hPa at 09/00z. It will be shown in the PV-upper-level framework that Kate’s extratropical cyclogenesis resulted from interactions with PV-anomalies aloft associated with a strong, meridionally oriented trough. Kate’s baroclinic lifecycle is more reminiscent of an LC1-type (THM) with also contains many of the same features of the Shapiro-Keyser (1990) conceptual model. The importance of the dry-air intrusion due to the stratospheric extrusion of high-PV air will be shown to have developed an upstream jet streak that aided in the extratropical development of Kate through the PV-anomaly interactions discussed by Hoskins et al. (1985).
As Kate approaches the baroclinic zone, the downstream jet streak is enhanced (Fig. 4.5a,b) due to the steepening of the tropopause by the tropical cyclone upper-level outflow outlined earlier with the Irene example. The overall convective tower associated with Kate is weak (likely poorly resolved due to the system’s small size) but is resurrected by the formation of a new extratropical tower (Fig. 4.6a-c). The strong core of the jet streak is still located to the north of tower at 08/00z (Fig. 4.6c) yet strong lateral interaction with stratospheric air with high-PV from the northwest begins to spill into the middle troposphere (Fig. 4.7b). The necessary westward tilt for cyclogenesis (baroclinic instability) continues throughout this frontal fracture process clearly shown by the tropopause fold descending to about 500-hPa at 08/06z (Fig. 4.7c). The upper-level positive PV anomaly became vertically aligned with the tropospheric tower and barotropic decay ensued at 09/00z (Fig. 4.7e). Kate entered the warm seclusion phase and weakened fairly quickly over the next two days.
On the 325 K isentropic surface, Kate does not interact with the reservoir of upstream high-PV until 08/06z (Fig. 4.8e) coinciding with the frontal fracture phase and dry air intrusion. It is at this time as well that the upwind jet-streak intensifies to over 60ms-1, with Kate located in the left-exit region beneath dynamically forced ascent. There is a weaker superposition of favorable ascent regions related to the upstream and downstream jets since the latter one weakens and moves eastward at the same time as Kate begins to feel the effects of the dry air intrusion high-PV advection. The cyclonic wrap-up indicative of dry air intrusion advects Kate along this sheet of high-PV while at the same time wrapping around the cyclone core (Fig. 4.8e,f) from 08/06z to 08/12z. However, characteristic of a LC1 baroclinic lifecycle (THM), cyclonic wrap up continues until the anticyclonic behavior of the poleward isentropic flow of PV develops. This trough thinning process destroys the potential temperature gradient and causes the system to weak quickly. The sequence from 08/12z to 09/12z shows a moderate trough thinning, yet is clearly distinct from the rapid cyclonic wrap-up associated with Irene (LC2 lifecycle). Consequently, the storm becomes more elongated along a NE-SW axis as the warm seclusion process proceeds. Even though LC1 baroclinic lifecycles favor a Norwegian type occlusion, the frontal fracture is characteristic of the Shapiro-Keyser lifecycle, which predict the formation of a bent-back warm front and warm core seclusion. Thus, the extratropical phase of Kate features properties of different lifecycles proving the inability to simply categorize storms according to a certain scheme. Nature provides much variety especially in terms of cyclone evolution.
4.3 Hurricane Fabian (2003)
After undergoing ET, the PV tower associated with Fabian remained vertically aligned and coherent until encountering a broad warm-front deformation zone that zonally deformed the system. From the sequence of IPV-325 K maps (Fig. 4.9), Fabian is involved in the formation of large-scale cut-off low over the northern Atlantic from 08/00z to 10/18z. This is another clear example of LC2 baroclinic lifecycle but in this case, Fabian does not interact directly with the midlatitude circulation or upper-level trough in terms of exchange of PV, for instance. Instead, the poleward propagation of Fabian contributes indirectly to the LC2 process along the eastern periphery of a large trough over the northwestern Atlantic. Fabian does not cross the jet axis but pushes it northward near 70°N at the apex of a broad thermal ridge associated with warm advection. Moderate upstream jet intensification is noticeable (Fig. 4.10), which in turn transports Fabian westward around the periphery of the broad PV blob (Fig. 4.9d,e) from 09/00z to 10/18z. Little upper-level PV anomaly interaction occurs except for a weak intrusion of PV at 09/00z (Fig. 4.11b) that helps maintain the PV tower (Fig. 4.12). The trough interaction is not favorable in the Hanley et al. (2001) sense, the PV tower does not tilt westward with height, and cyclogenesis does not initiate. Instead, slow weakening occurs until Fabian is deformed and absorbed by the cutoff low to the west. Another explanation for the weakening concerns Fabian’s inability to precondition an upstream jet streak with its upper-level outflow and associated negative PV anomaly. Moreover, the weak jet is already downstream, a position not favorable for baroclinic instability.
4.4 Trajectories Through the Warm Seclusions of Kate (2003) and Irene (1999)
Kuo et al. (1992) modeled the airflow and thermal structure of an occluded marine cyclone named the Ocean Ranger storm and present a clear conceptualization of the process as well as a few hypotheses about low-level frontal structure that will be addressed forthwith. The use of relative stream trajectories on isentropic surfaces allows for easy determination of vertical motion assuming adiabatic motions. Thus, the origination of various air streams, whether ascending or descending, influences the thermal structure of the system and the low-level frontal structure. Browning et al. (1998) examined the evolution of Lili as it reintensified after extratropical transition in terms of a potential vorticity framework. The goal here is to tie the two conceptualizations together to create a clear picture of the upper-level airflow and PV anomaly interactions with the low-level frontal structure during the warm seclusion process.
The trajectories labeled A, B, C, and D (Fig. 4.13) are analogous to the warm and cold conveyor-belts described by Carlson (1980) and Browning (1986) representing spreading of rising air in different directions around the storm. The storm is embedded entirely in warm advection since all four trajectories veer (turn clockwise) with increasing height. The dry intrusion discussed as high-PV, low humidity air, and high static stability of stratospheric origin subsides as it advances on the center from the west and southwest depicted by trajectories E and F. After circulating around to the east of the center, airstream E is shown to rise consistent with dynamic lifting ahead of the upper-level trough or due to frontogenetical forcing along the occluded front (Chapter 6). In both cases, the airstream’s history is one of dryness and subsidence, which explains the pronounced dry slot in the mid- to upper-level cloud pattern ahead of the occluded front. Likewise, the lack of a pronounced cold front structure with billowy convection is a product of the high static stability, dryness, and history of subsidence of the air encircling the storm. F remains along a fairly level path owing its subsidence and drying to upstream processes behind the upper-level trough. Kuo et al. (1992) note that the thermal gradient aloft is too weak to characterize as a front, but rather as a confluence of two airstreams of widely different origins (Carlson 1980).
4.41 Irene Seclusion Trajectories
The 310 K isentrope corresponds roughly with the dynamic tropopause (PV = 2) at 19/12z (Fig. 4.3b) from the zonal cross-section analyses of the frontal fracture stage. As mentioned previously, the height of the tropopause is decreased on the cold side of upper jet as expected, but a region of locally even lower tropopause exists just west of the cyclone (Browning and Roberts 1994). As Irene initially developed beneath the right entrance region of the upper-level jet streak, there was a considerable band of high cloud associated with slantwise ascent north of Irene and collocated with the jet streak. As seen, the intrusion of dry, high-PV air from upstream at < 350-hPa descended into the warm boundary layer at roughly 750-hPa (Fig. 4.14 Top) a dual jet streak structure formed (Fig. 4.1c). The slantwise ascent slowly weakened beneath the right entrance region of the downstream jet. However, a new jet strengthened upwind due to the moisture and temperature gradient associated with the high-PV stratospheric air and frontal fracture. The dynamical tropopause slowly approached the low-level PV maximum associated with moist processes and later reached a state of vertical alignment. It is the transverse ageostrophic circulation at the exit of an upper-level jet that transports warm, moist air from the warm sector in the boundary layer from right to left across the jet axis towards the cloud head and beneath the dry intrusion air (Browning and Roberts 1994). Convective potential Instability results and convective cloud formation in the cloud hook region is clear visible (Fig. 6.6). It is this same ageostrophic circulation that forces the dynamic descent and dry air intrusion or frontal fracture. A feedback occurs as an upstream jet is rapidly strengthened, with Irene now in the left-exit region of the jet favorable for cyclogenetic development. It is during this process that Irene deepened the most rapidly.
The 300 K isentropic surface intertwines Irene on a path from near the surface to upper-levels past the storm as seen in the PV N-S cross-section (Fig. 4.2f). While not exact, the trajectories in the Kuo et al. (1992) model associated with the low- to mid-level conveyor belts match observations at 19/18z (Fig. 4.15b). Ahead of Irene, a fanning out of the airstreams is observed with a southern branch characterized by gentle ascent and quick downstream movement (A in Fig. 4.13), a second with marked ascent and weak veering (B), and a third representing the cold conveyor belt wrapping cyclonically around the hook cloud and the region of dry intrusion into the center of the storm (D). This process of hook-cloud formation occurred as the upper-PV anomaly associated with the frontal fracture wrapped around the diabatically generated low-level PV maximum and is also indicative of LC2 type baroclinic waves with barotropic shear (THM). As this cyclonic wrap-up intensifies, the superposition of the cyclonic circulations associated with frontal fracture positive PV anomaly and the low-level positive PV anomaly spin-up the system. The vertical shear along the western and south side of Irene associated with the marked PV gradient tightens and effectively forces a cyclonic low-level jet around the low-level PV maximum (Fig. 4.14). The source of this baroclinicity and temperature gradient is associated with the strengthening bent-back front encircling the relative warm air around center compared to the environment immediately to the west. This jet will be shown to be associated with very strong winds, especially on the southern side where the storm’s motion adds to the system-relative circulation.
A similar trajectory analysis is shown for Kate (2003) (Fig. 4.16) on the 310 K (top) and 294 K (bottom) isentropic surfaces. The former corresponds roughly to the upper-level frontal fracture and associated conveyor belts while the latter more closely represents the low-level advection patterns often observed during the seclusion process. As an example, the seclusion seen on the 294 K lower-level surface plot (Fig. 4.16, bottom) is identified by a sharp thermal ridge and wind shift line separating areas of warm and cold advection at levels primarily below 700-hPa. The pinching action of air ahead of the baroclinic zone by the occluded front is similar to that described by Bjerknes and Solberg (1922) in their original paper on the life cycle of extratropical cyclones. In comparison, it is easily seen that Kate is becoming under the influence of anticyclonic shear associated with the upper-level jet, which is not only sending Kate to the northeast, but is elongating the storm in that direction. So, a weaker incarnation of cyclonic wrap-up is apparent preceding definite anticyclonic shear associated with the trough thinning and LC1-type waves (THM). Yet, the seclusion process occurs with Kate nonetheless but it is weaker and less pronounced than Irene. It is hypothesized that the cyclonic wrap up associated with Irene is much stronger than Kate owing to a stronger frontal fracture and intrusion of dry air associated with it. The weaker frontal fracture only marginally strengthens the upwind jet, through which secondary ageostrophic circulations forces a weaker vertical motion response. Kate does not remain in the relative favorable jet streak location but instead comes under the effects of anticyclonic shear. It weakens fairly rapidly afterwards in contrast with Irene, which remains a deep, cut-off low-pressure system for nearly a week.
Figure 4.13: From Kuo et al. (1992) schematic showing airflow through the modeled Ocean Ranger storm. Fronts are in open wave and occluded stages with the usual convention; fluffy cloud boundaries at mid- to upper levels; and relative trajectories, arrows (open where rising, shaded where sinking, hatched where level); delta scale for pressure level, lower right.
Figure 4.14: Conceptualization of warm seclusion of Lili from Browning et al. (1998). The thin solid line and stippled shading represent θw > 12°C. This feature is terminated at 500-hPa and capped by a perspective view lid to give a 3-dimensional impression. JU and JL respectively denote the upper- and lower-level jet axes with the bold line representing the dynamic tropopause PV = 2 isopleth. The wavy line corresponds to the model’s RH = 80% isopleth, which is interpreted by Browning et al. (1998) as the top of the actual region of cloud.
Figure 4.15: Relative airstream analysis of Irene (1999). Top: 310 K isentropic surface at 10/19 1200z. Bottom: 300 K isentropic surface. Both with wind vectors (ms-1) and pressure level (shaded, hPa). Storm motion at 36° u = 14.5 ms-1 and v = 10.7 ms-1.
Figure 4.16: Relative airstream analysis of Kate (2003). Top: 310 K isentropic surface at 10/08 0600z. Bottom: 294 K isentropic surface. Both with wind vectors (ms-1) and pressure level (shaded, hPa). Storm motion at 75° u = 11.2 ms-1 and v = 2.5 ms-1.