Every year, the North Atlantic Ocean sees an average of eleven tropical storms and hurricanes (Elsberry 1987). Of these 11 tropical cyclones (TCs), about 45% (Hart and Evans, 2000a) transition into extratropical cyclones (ECs). Climatologically, most extratropical transitions occur between 35°N and 45°N, although this latitude varies seasonally (Hart and Evans, 2000a); this region is near the populated United States East Coast as well as over major shipping routes. Despite this proximity to a large population and financial interests, the process of extratropical transition remains poorly forecasted, in terms of both intensity change and the time period of transition.

The key in the extratropical transition of TCs is the development of baroclinicity. Generally, this is achieved by inducing a temperature gradient and/or wind gradient. With respect to a TC, these gradients can be induced by interaction with one or more of several features, including mid-latitude cyclones and upper-level features such as jet streaks and troughs. Whatever the means of interaction, a TC undergoing extratropical transition experiences a period of transition; during this period, the structure of the storm changes through surface to upper levels from a "pure" TC into a storm of extratropical characteristics. These structure changes include expansion of the wind field, surface frontogenesis, a shift in the level and radius of maximum winds, and a transition from (dynamically) warm core to cold core. The extratropical transition of a TC is not well defined by a point in time because of the nature of the transition process. Rather, the transition is a process, characterized by an objective start and finish

SYNOPTIC OVERVIEW

Irene formed from a low-level cyclonic convective system southwest of Cuba into Tropical Depression 13, upon reanalysis, at 06Z 13 Oct 1999 (henceforth, all dates given for Irene's lifecycle will be from 1999), and was upgraded to Tropical Storm Irene at 12Z 13 Oct, just 6 hours later, at 18.5°N 83.4°W. Irene remained a TS while crossing Cuba, then gained hurricane strength at 06Z 15 Oct, just before brushing Key West, FL. Irene remained a minimal hurricane, with 65kt winds and central pressure around 985hPa, while crossing the southern tip of FL and upon emerging into the Atlantic Ocean. Once over the Gulf Stream, Irene paralleled the East Coast, where she slowly began to intensify. Figure 1 shows the track of Irene from 00Z 17 Oct onward. Early on 18 Oct, a reconnaissance flight unexpectedly found a minimum central pressure of 958hPa, winds at 902hPa of 129kt, and surface winds of 95kt. This peak in intensity officially came at 0756Z 18 Oct, when Irene was centered just off the coast of NC.

Irene also began her transition somewhere around this point. Irene weakened slightly, to 80kt winds and central pressure of 968hPa, and began to accelerate to the northeast under the influence of the synoptic environment. Within a period of about 24hr, Irene transitioned directly from a Category 1 hurricane to an extratropical cyclone (EC). The NHC declared Irene extratropical at 03Z 19 Oct (NHC99).

MY CURRENT WORK

In the process of studying the transition of Irene, I simply wanted to create a model of the storm in the PSU-NCAR MM5v.2 model that came closest to "truth". In the process of modeling Irene, however, I've encountered several stories within the story. The case has turned out to be extremely sensitive to the convective parameterization scheme (CPS) used in the model. With the Betts-Miller (BM) scheme, the storm recurves too early and dissipates without ever transitioning or interacting with the jet streak. The Kain-Fritsch (KF) scheme corrects this, overshooting the track just a bit too far to the north but otherwise improving the forecast. It seems that propagation is fairly consistent between the runs, so my next step is to look more carefully at the CPSs themselves to learn what might make such a difference in position.

Because transition occurs early on 18 Oct, I am initializing the model with enough spin-up time (even though a 00Z 18OCT BM SW run seems to have represented the storm all right). The 18Z 16OCT KF SW and 12Z 16OCT KF SW runs do not have a significantly different track at all, so my choice for the "best" run is based on the fact that 300 hPa jet streak in the 18Z 16OCT KF SW run is underestimated. Thus, I chose the 12Z 16OCT KF SW run as "best".

For each of 6 model runs--my analysis runs--I have plotted several factors for analyzing the runs. These include 300hPa winds with MSLP, 700hPa vertical velocity with 6-hour surface pressure tendency and 300hPa wind above 50 m/s, and 310K isentropic potential vorticity (note: some pressure surfaces are actually close approximations to the sigma levels used in MM5). Then, for the "best" run (as well as, to some extent, the good runs), I have also made a few extra plots, such as equivalent potential temperature at 850hPa and 300hPa and 345K isentropic potential vorticity. The plots are in the directories below:

12Z 16OCT KF SW -- "best"
12Z 16OCT BM SW -- bad
18Z 16OCT KF SW -- good
18Z 16OCT BM SW -- bad
00Z 17OCT KF SW -- good
00Z 17OCT BM SW -- bad

NOGAPS Analyses

Tracks of the 6 analytical cases plus truth cases
Tracks of the best cases (easier to see differences between the close tracks)
Tracks of cases initialized on 12Z and 18Z 16 OCT
Tracks of cases initialized on 00Z 17 OCT

REFERENCES

Avila, L.A., 1999 (NHC99): Preliminary report: Hurricane Irene, 13-19 October 1999. National Hurricane Center, available at http://www.nhc.noaa.gov/1999irene.html.

Elsberry, R.L. (Ed), 1987: A Global View of Tropical Cyclones. Office of Naval Research, Washington DC, 185pp.

Hart, R., and J.L. Evans, 2000a: A climatology of the extratropical transition of Atlantic tropical cyclones. J. Climate (Accepted March 2000, in press).

----- and -----, 2000b: Extratropical transition of Atlantic tropical cyclones: Defining the transition lifecycle using objective indicators of transition. Mon. Wea. Rev. (Submitted April 2000, in review).


Look for continual updates as I update my research!
Last update: 6 March 2001