916 WEATHER AND FORECASTING VOLUME 17 The Evolution of a Hurricane±Trough Interaction from a Satellite Perspective DEBORAH E. HANLEY Center for Ocean±Atmospheric Prediction Studies, The Florida State University, Tallahassee, Florida 8 October 2001 and 11 February 2002 ABSTRACT A case study is presented that describes the interaction of Hurricane Bertha (1996) with an extratropical upper- tropospheric trough shortly before the hurricane's landfall in North Carolina. This case study illustrates the evolution of the interaction as seen in water vapor satellite imagery. The similarity of the features in the satellite images to those in gridded analyses of potential vorticity in the upper troposphere suggests that some types of trough interactions may be identi®able from satellite imagery. 1. Introduction ies of tropical cyclone±trough interactions, Hanley (1999) noted that there was very strong similarity be- Although most of the previous studies of the inter- tween the PV structure of the upper-tropospheric trough actions between tropical cyclones and extratropical up- and signatures observed in satellite water vapor imagery per-tropospheric troughs have concentrated on the use [i.e., Geostationary Operational Environmental Satel- of diagnostic calculations from analyses, a few have lite-8 (GOES-8) 6.7-mm imagery]. The size and shape made use of satellite imagery (e.g., Velden 1987; Bosart of the PV maxima on selected isentropic surfaces was et al. 2000). Hanley et al. (2001, hereinafter HMK) iden- represented by corresponding dark regions in the water ti®ed two different types of trough interactions based vapor imagery with similar structures. Water vapor im- on the potential vorticity (PV) structure of the upper- age brightness temperature contours in the satellite im- tropospheric trough. The ®rst type, de®ned as super- agery were noted to be approximately parallel to PV position, occurs when an upper-tropospheric PV max- contours. This note illustrates the similarities of the PV imum approaches within 400 km of the tropical cyclone and water vapor signatures for the case of Hurricane center. The second type, de®ned as distant interaction, Bertha (1996), an example of a superposition-type trop- occurs when the PV maximum is within 1000±400 km ical cyclone±trough interaction. of the center. HMK found that 78% of all superposition Hurricane Bertha originated on 1 July 1996 from a cases intensi®ed and 61% of all distant-interaction cases tropical wave that moved off the coast of West Africa. resulted in deepening. Subtle differences in the PV The storm followed a smooth path around the Atlantic structure of the upper-tropospheric troughs involved in subtropical ridge while a weak midlevel trough persisted intensifying and weakening distant-interaction cases in the western North Atlantic. Bertha strengthened into suggest that PV signatures alone may not be suitable as a hurricane on 8 July as it moved across the Leeward forecast methods (HMK). The high percentage of in- and Virgin Islands. At that time, maximum winds of 39 tensifying superposition cases observed by HMK sug- ms21 were recorded. On 9 July, the track turned to the gests that identi®cation of such interactions by numer- northwest, and the maximum sustained winds were mea- ical guidance or by satellite imagery could improve the sured to be 51 m s21 at 0600 UTC. Bertha began to forecasting of intensity change during this type of in- follow a more north-northwest path on 10±11 July at a teraction. forward speed of about 4 m s21. Maximum winds were The relationship between water vapor and PV has gradually decreasing at this time, from 51 m s21 on 9 been shown by several investigators (e.g., Appenzeller July to 36 m s21 on 11 July. Prior to landfall near Wil- and Davies 1992; Appenzeller et al. 1996; Mans®eld mington, North Carolina, at 2000 UTC 12 July, Bertha 1997; Demirtas and Thorpe 1999). In several case stud- accelerated to a speed near 8 m s21. As an upper-level trough approached the hurricane, winds in Bertha abruptly increased to 46 m s21, which was the estimated Corresponding author address: Dr. Deborah E. Hanley, Center for 1-min wind speed at landfall. Details of Hurricane Ber- Ocean±Atmospheric Prediction Studies, The Florida State University, 227 Johnson Bldg., Tallahassee, FL 32306-2840. tha's track and intensity evolution were obtained from E-mail: [email protected] Pasch and Avila (1999). q 2002 American Meteorological Society Unauthenticated | Downloaded 10/07/21 07:31 AM UTC AUGUST 2002 NOTES AND CORRESPONDENCE 917 FIG. 1. National Hurricane Center advisory locations of Hurricane Bertha (1996) every 3 h from 0100 UTC 9 Jul to 0500 UTC 14 Jul superimposed on 3-day average AVHRR SST (8C) ending 1238 UTC 9 Jul. (Figure obtained from R. Sterner and S. Babin, Applied Physics Laboratory, The Johns Hopkins University.) The storm was much stronger at landfall than had 2. Data and analysis methods been originally forecast. Although warnings were posted a. Data prior to landfall, Bertha resulted in approximately $270 million in damage in North Carolina, and 12 deaths were Six-hourly uninitialized 1.1258 European Centre for reported that were due to the storm (Pasch and Avila Medium-Range Weather Forecasts (ECMWF) gridded 1999). The failure of of®cial forecasts to describe ad- analyses are obtained from spectral coef®cients archived equately the intensi®cation of Bertha highlights the chal- at the National Center for Atmospheric Research. Data lenge of forecasting tropical cyclone intensity. The rapid are available on 13 pressure levels for this case study reintensi®cation of Bertha just prior to landfall indicates and are interpolated to 20 levels. Molinari and Vollaro the need for more forecasting methods with regard to (1990) and Molinari et al. (1992, 1995) have discussed identifying trough interactions. the bene®ts of ECMWF analyses of the tropical cyclone Unauthenticated | Downloaded 10/07/21 07:31 AM UTC 918 WEATHER AND FORECASTING VOLUME 17 FIG. 2. Time series of sea level pressure every 6 h (UTC) for Hurricane Bertha from the best-track dataset. FIG. 3. Time series (UTC) of 200±850-hPa wind shear (m s21), calculated over a 500-km storm-centered radius. environment. Sea level pressures and storm center lo- cations are taken every 6 h from the National Hurricane Center ``best-track'' dataset (Jarvinen et al. 1984). peratures, SST changes alone are not suf®cient to ex- The satellite data used in this study are available at plain the ¯uctuations in intensity observed in Hurricane the archive at the University of Illinois and include Bertha. Prior to the strengthening observed in Fig. 2 GOES-8 visible imagery and GOES-8 6.7-mm water va- (0600±1800 UTC 12 July), Bertha was weakening por imagery. Sea surface temperature data from the Ad- steadily as the storm moved northward along the east vanced Very High Resolution Radiometer (AVHRR) coast of Florida. satellite instrument, composited over a 3-day period The vertical wind shear in the 850±200-hPa layer is ending 1238 UTC 9 July, are obtained courtesy of the calculated by averaging over a 500-km radius about the Space Oceanography Group at The Johns Hopkins Uni- storm center (Fig. 3). The averaging removes the azi- versity. High-density winds derived from cloud and wa- muthal mean component and retains the cross-storm ter vapor motions in GOES-8 satellites are supplied by component (Molinari 1993). As the trough and hurri- the Cooperative Institute for Meteorological Satellite cane approach one another (1200 UTC 12 July), the Studies at the University of Wisconsin (UWÐCIMSS). vertical shear increases to more than 15 m s21 from the Details of the processing involved in creating these wind northwest. After 1200 UTC 12 July, the shear begins ®elds can be found in Velden et al. (1997, 1998). to decrease as it shifts to a southwesterly direction. Al- though it may appear that weakening shear was the rea- son for the intensi®cation noted in Fig. 2, the central b. Diagnostic methods pressure began to fall 6 h prior to the drop in shear HMK used vertical wind shear, eddy momentum ¯ux magnitude. convergence (EFC), and Ertel PV to describe the evo- Increasing values of vertical shear on 11 July (Fig. lution of trough interactions in composite tropical cy- 3) result in a downshear tilt in the hurricane, as evi- clones. These diagnostic calculations will also be used denced by the presence of an exposed eye in Figs. 4a,b, in this study. Details of these calculations are presented which depict Bertha at 1700 and 2000 UTC 11 July, in section 2b of HMK. respectively. This disruption of the vertical alignment of the hurricane results in rapid weakening during this period (Fig. 2). Velden (1997) has proposed that the 3. Discussion weakening observed in Hurricane Bertha may be related AVHRR sea surface temperatures (SST; Fig. 1) in- to ``backdoor'' vertical wind shear. This wind shear sce- dicate that Bertha was moving over regions of SST nario arises when a short-wave trough and associated warmer than 268C throughout its entire track over water. jet downstream of the hurricane begin to amplify During the period of intensi®cation near landfall (0600± through baroclinic processes. The resulting jet entrance 1800 UTC 12 July; Fig. 2), SSTs along Bertha's track region is located just to the north of the main hurricane remained relatively constant (Fig. 1). Because intensi- out¯ow channel and results in air accelerating from the ®cation occurred over relatively constant water tem- hurricane into the jet entrance region. High-density mul- Unauthenticated | Downloaded 10/07/21 07:31 AM UTC AUGUST 2002 NOTES AND CORRESPONDENCE 919 FIG. 4. Visible satellite imagery for the following times: (a) 1700, (b) 2000, and (c) 2100 UTC 11 Jul and (d) 1200 UTC 12 Jul.