1. INTRODUCTION
Numerous severe weather and bow echo events have been studied and documented
in the mid-latitudes. Occurrences of bow echoes in a tropical environment,
specifically the Caribbean are not well documented. On July 8, 1996 Hurricane
Bertha moved west to northwest at 10 to 15 mph across the Caribbean. During the
morning hours a couple of outer rainbands moved across Puerto Rico. This paper
will demonstrate that the phenomena of bow echoes embedded in an outer rainband
of a hurricane was: 1) observed, and 2) occurred in an environment of high storm
relative helicity values, but marginal Convective Available Potential Energy (CAPE),
reminiscent of mid-latitude bow echo events.
Przybylinski et al (1993) found that cool season bow echo events possessed the
following features: 1) high precipitation (HP) storms frequently embedded within
large rapidly moving squall lines, 2) HP storms characterized by shallow, but still
strong rotating updraft centers with 50 dBZ cores only extended up to altitudes of
4-5 km.
Structures of rainbands in tropical cyclones studied by Jorgensen (1984) and
Korotky et al. (1995) were found to have weak updrafts and reflectivities decreasing
rapidly with height above the freezing level.
2. ENVIRONMENTAL CONDITIONS
The 1200 UTC sounding data for San Juan, Puerto Rico (SJU) were analyzed with the
SHARP program. Convective indices indicated only a marginally unstable air mass
with a CAPE of 681 Jkg-1. A previous study by McCaul (1991) on tornado
environments in tropical cyclones that affected the United States found mean CAPE
values of 299.8 Jkg-1 and Bulk Richardson Numbers (BRN) shear values of 8.4 ms-1
in the left front quadrant. Puerto Rico was in the left front quadrant of Hurricane
Bertha, but exhibited much greater values of CAPE. Computation of BRN yielded a
value of 12, which fell within the range of 10 to 40 found by Weisman and Klemp
(1982, 1984) to most likely favor supercell type storms.
The hodograph for 1200 UTC 8 July 1996 was analyzed using the SHARP program.
SHARP estimate of sr helicity (0-3 km, AGL) for the observed sounding was 218 m2s-2. A modified SHARP estimate of sr helicity was produced using the storm motion
of 066 degrees at 34 kts. This yielded a sr helicity of 244 m2s-2. This fell within the
range of 150-299 m2s-2 found by Davies-Jones et al. (1990) to favor weak tornadoes.
A right moving storm in the outer rainband produced a weak (F0) tornado which
caused damage as it passed over Mayaguez.
The VAD wind profile is shown in Fig. 1a-c. Examination of the wind profile shows
an increase in depth of the 40 kt isotach from 1108 to 1114 UTC, but a much greater
total depth (13,000 ft) increase from 1142 to 1206 UTC. Maximum estimated VAD
wind speeds also increased to 55 knots at 10,000 ft by 1200 UTC. The increase in
depth of the 40 kt isotach and maximum wind speeds to 55 kts nearly coincided with
the passage of the leading edge of the outer rainband at San Juan. The momentum
of the winds aloft was carried down to the surface. At 1213 UTC the visibility
dropped to 1 3/4 nautical miles in moderate rain and the wind shifted from 330
degrees at 16 knots to 040 degrees at 26 kts with gusts to 36 kts at San Juan's Luis
Munoz Marin International Airport. Later, the ASOS remark of PRESRR at 1234 UTC
indicates a mesohigh was likely located on the east side of the rainband.
Figure 1a-c. VAD Wind Profile 8 July 1996 a) 1102-1200 UTC b) 1206-1304 UTC c) 1304-1402 UTC
3. DISCUSSION
At 1200 UTC an outer rainband was over the east coast of Puerto Rico. The rainband
had bow echoes embedded in it similar to mid-latitude mesoscale convective
complexes. The band had sharp low-level reflectivity gradients along the leading
edge and maximum reflectivities 50-54 dBZ. Rotating heads and rear inflow notches
(RINs) were visible in a large broad bow. At least three rear inflow notches (RINs)
appeared in the reflectivity data over the ocean north of Puerto Rico in the rainband.
Rotating heads appeared to be present with the smaller scale bows and suggested
the locations of mesoanticyclones. VIL values were 12 kgm-2 or less in the rainband
and maximum echo tops 20,000-25,000 ft. Maximum reflectivities of 50-54 dBZ were
confined to 9,000 ft or less. Maximum echo top data suggested the convection
producing the bow echoes was low top in nature, generally less than 30,000 ft. A
trailing band of stratiform precipitation was observed east of the bow echo.
Between 1200 UTC and 1419 UTC, the outer rainband progressed past San Juan
toward western Puerto Rico. Figure 2 shows a bow echo embedded in a rainband of
Hurricane Bertha as it crosses the north coast of Puerto Rico on 8 July 1996. A well
defined bow echo with maximum reflectivities of 50-54 dBZ, a RIN, and subtle hook
echo are visible at 1310 UTC before weakening by 1414 UTC.
Figure 2. Four-panel Base Reflectivity time series at 0.5o elevation for the period 1304UTC (upper
left) - 1321UTC (lower right) on 8 July 96
Another portion of the initial rainband (north of the coast) discussed, began bowing
southeast from the Atlantic toward northwest Puerto Rico. By 1419 UTC an incipient
bow echo was over northwest Puerto Rico and had increased in maximum
reflectivity to 50-54 dBZ. VIL values were 14 kgm-2 or less. A RIN present east of the
developing bow echo at 1419 UTC expanded and progressed into the bow. From
1425 to 1431 UTC a subtle hook and, or mesocyclone was suggested by the
reflectivity data near Mayaguez. Unconfirmed reports of a tornado were received
from the Mayaguez area in the morning. Examination of base reflectivity data for
1425 UTC indicated a shallow core to the bow echo, 35-39 dBZ limited to below
22,000 ft. VIL values in the bow echo were 12 kgm-2 or less. Maximum echo tops were
only around 20,000-25,000 ft. Other storms in the rainband north of the coast
exhibited subtle bow echo features also.
Figure 3. Shows the SRM data corresponding to the low-level reflectivity data .
Examination of this data confirms the existence of features normally associated with
bow echoes, such as a RIN, and suggests mesocirculation features were also
present in the outer rainband. Further diagnosis of SRM data for the period 1304
UTC to 1414 UTC showed weak mesocyclonic and mesoanticyclonic areas of shear
associated with the bow echoes. Maximum outbound velocities of 10-21 kts and
maximum inbound velocities of 22-29 kts occurred between 1316 and 1321 UTC.
Figure 3. Four-panel Storm Relative Velocity map time series at 0.5o elevation for the period 1304
- 1321UTC on 8 July 1996
From 1356 to 1414 UTC, velocities and shears gradually weakened in the original
bow echo which had moved over western Puerto Rico. The acceleration of the
portion of the rainband north over the Atlantic Ocean continued southeast into
northwest Puerto Rico. By 1414 UTC, outbound velocities east of the new bow echo
had increased from 10-21 kts to 22-29 kts. Weak subtle areas of shear began to
develop on the north and south ends of the concave shaped echo.
Outbound velocities peaked at 22-29 kts over western Puerto Rico immediately east
of the evolving bow echo at both 1419 and 1431 UTC. Subtle areas of meso -anticyclonic shear on the north end and mesocyclonic shear on the south end of the
bow echo increased. Inbound velocities at both ends of the bow echo peaked at 22-29 kts. The radar beam was likely overshooting much of the low-level outflow air
with a beam height of 9,000-10,000 ft. By 1437 UTC, the bow echo had begun to
dissipate.
4. SUMMARY AND CONCLUSIONS
WSR-88D data indicate that bow echoes are more frequent in hurricane rainbands over the Caribbean than previously thought. Examination of the data revealed more than one occurrence of bow echoes in the outer rainbands of Hurricane Bertha.
Bow echoes observed in Hurricane Bertha's outer rainbands were more similar in
structure and environmental conditions to that found in cool season mid-latitude
bow echo events than to warm season mid-latitude events. Low echo tops and VILs,
and limited vertical extent of the high reflectivity cores provided strong evidence of
weak updrafts in the rainbands, which is similar to earlier findings of research on
tropical cyclones. These structural features are opposite of those found for warm
season bow echo events at mid-latitudes. Environmental parameters, marginal
instability and higher sr helicity, also conformed to the conditions found present in
cool season bow echo and tornadic hurricane events.
Finally, topography and wind direction appear to be key factors in the formation and
longevity of bow echoes in Puerto Rico. Mountainous terrain and wind direction on
the island likely affected storm type. Most days convection over the island appears
to be pulse type storms, but the bow echoes in the outer rainbands on 8 July 1996
had features more common to HP supercells.
5. ACKNOWLEDGEMENTS
The authors wish to thank Rachel Gross for reading the manuscript and suggesting
helpful changes.
6. REFERENCES
McAdie, J. C., P. Dodge, and S. H. Houston, 1995: Mesoscale features of Tropical
Storm Beryl (15-16 August 1994) as detected by the WSR-88D. Preprints, 21st
AMS Conf. on Hurricanes and tropical Meteorology. Miami. FL, Amer. Meteor.
Soc. 320-322.
Jorgensen, D.P., 1984: Mesoscale and convective-scale characteristics of mature
hurricanes. Part 1: General observations by research aircraft. J. Atmos. Sci.,
41, 1268-1285.
Korotky, J., J.R. Michael, and C.E. Williford, 1995: Tropical Storm Beryl: A WSR-88D Radar Overview. Preprints, 21st Conf. on Hurricanes and Tropical
Meteorology. Miami, Fl, Amer. Meteor. Soc.
Parrish, J. R., R. W. Burpee, F. D. Marks, JR., and R. Grebe, 1982: Rainfall patterns
observed by digitized radar during the landfall of Hurricane Frederic (1979).
Mon. Wea. Rev., 110, 1933-1944.
Powell, M. D., P. D. Dodge, and M. L. Black, 1991: The landfall of Hurricane Hugo
in the Carolinas: Surface wind distribution. Wea. Forecasting, 6, 379-399.
Przybylinski, R.W., T. J. Shea, D. L. Ferry, E. H. Goetsch, R. R. Czys, N. E.
Wescott, 1993: Doppler Radar Observations of High-Precipitation Supercells
over the Mid-Mississippi Valley Region. Preprints, 17th Conf. Severe Local
Storms, St. Louis, Amer. Meteor. Soc., 158-163.
* Corresponding author address: John E Wright, National Weather Service