a. Synoptic History
The NHC surface analysis showed a weak 1019 mb frontal low
midway between Bermuda and the South Carolina coast at 0600
UTC 5 July. Maximum sustained winds around the low were 10 to
20 knots based on available ship reports. Satellite imagery
indicated that the clouds associated with the low gradually
became isolated from the frontal cloud band over the next 24
to 36 hours. The satellite imagery also revealed that a low-
level cloud system center
became better defined just to the west of a small cluster of deep convection, and it is estimated that
the frontal low transformed into Tropical
Depression Two near 1800 UTC 6 July (Fig. 1
(65K GIF) and Table 1). Upper-level westerly shear was evident
from the small area of deep convection remaining displaced to the east of the low-
level center. Little overall movement was noted on 5 and 6 July.
The center of circulation became better defined by a
curved low- to mid-level cloud band, and post-analysis suggests
that the depression strengthened into
Barry near 0600 UTC 7 July. During the day, the storm began
moving toward the north-northeast near 10 knots. This allowed
the storm-relative shear to decrease and deep convection to
move cyclonically around the western semicircle of the
circulation. The deepest convection moved from just north
through west to south of the circulation center. The presence
of a negatively tilted mid- to upper-level trough just to the
southwest of Barry appears to have favored the temporary increase in convection.
The maximum sustained winds are estimated to have occurred
near 2100 UTC 7 July. At this time, an
Air Force Reserve unit aircraft
reported hurricane force winds at a flight-level of
1500 feet, but the minimum central pressure of 998 mb reported
by the plane does not appear to support sustained hurricane
strength. Central convection decreased dramatically after the
strong winds were reported, and it is assumed that the
aircraft winds were associated with a transitory mesoscale feature.
Satellite imagery revealed a cloud-free center within
relatively weak surrounding convection by 0000 UTC 8 July.
The next aircraft reconnaissance report indicated that the
minimum central pressure had changed little, but the maximum
flight-level winds had decreased about 40 knots from those
that were measured the previous day. By 1800 UTC 8 July, a
small area of deep convection had developed near the low-level
circulation center. The storm began accelerating toward the
north-northeast in advance of a large amplitude trough moving
eastward over the eastern United States. The central dense
overcast grew until near 1200 UTC on 9 July. Some of this
increase in convection may have been related to the passage of
Barry over a warm water eddy that bulged northward from the
Gulf Stream to near 42°N between 63-66°W.
Convection associated with Barry began to weaken as the
continued to accelerate toward the north-northeast
over cooler water. The maximum winds began to spread out away
from the cyclone center as Barry gradually lost tropical
characteristics, although upper-air soundings indicated that
still exhibited a warm core when it passed near
Sable Island. The center of the storm crossed the eastern tip
of the peninsula of Nova Scotia, near Hart Island, around 2130
UTC 9 July and then continued north-northeastward over Cape Breton Island. Barry became
extratropical near the western
coast of Newfoundland shortly after 0600 UTC 10 July. As a
weakening extratropical cyclone, it could be tracked to near
the southeast coast of Labrador before losing its identity.
b. Meteorological Statistics
Figures 2 and 3
(46K GIF) show "best track"
curves of minimum central pressure and maximum
one-minute surface wind speed, respectively, as a function of time. The
observations on which the curves are based are also plotted and consist of
using satellite imagery, aircraft reconnaissance data on 7 and 8 July, a few land reports,
as well as synoptic fixes from the Eastern
Canadian Hurricane Center in Halifax, Nova Scotia. The typical relationship
between pressure and wind does not appear well established
during Barry. This is likely due to the fact that the minimum
pressure occurred at northern latitudes as Barry was losing
tropical characteristics and due to the fact that the maximum
winds appear to have been related to a transitory mesoscale feature.
On 9 July, Hart Island, Nova Scotia, reported 990.8 mb at
2145 UTC and Fourchu Head, Nova Scotia, reported 990.6 mb at
2248 UTC. The minimum central pressure curve is anchored to
these surface reports and to the reports from aircraft reconnaissance on 7 and 8 July.
The maximum wind reported by aircraft was 86 knots
at a flight-level of 1500 feet at 2050 UTC 7 July. The latest
available satellite wind estimates at that time were 35 knots
from both the NHC and the NESDIS
Synoptic Analysis Branch
(SAB). The satellite wind estimate was 30 knots from the
Force Global Weather Central (AFGWC) a few hours before and
after the maximum winds reported by aircraft. Similar scatter occurred
between the satellite estimates and the aircraft measurements of maximum
winds on 8 July as well. At 1200 UTC, the NHC satellite analyst estimated
65 knots while the SAB
analyst estimated 45 knots. At 1438 UTC, the analyst at the
estimated 30 knots. At 1328 UTC, the aircraft reported maximum winds of
43 knots at a flight-level of 1500 feet.
Given the large amount of scatter in maximum wind information,
there is obviously considerable uncertainty in the best track
wind speed on Tropical Storm Barry.
c. Casualty and Damage Statistics
No reports of casualties or damage associated with Barry
have been received at the NHC.
d. Forecast and Warning Critique
Barry was a tropical storm for only 3 days, so meaningful
quantitative forecast evaluations are not possible. However,
for the small sample, official track forecast errors were
somewhat better than the average over the past 10 years.
These track forecast errors ranged from 60 n mi at 24 hours
(9 cases) to 111 n mi at 48 hours (5 cases) to 280 n mi (1
case) at 72 hours. The only operationally available track
prediction model that had somewhat lower track forecast errors
than the official forecast at 48 hours was the AVNI model. In
general, the official forecasts as well as the track guidance
were good in indicating a fairly consistent track that kept
Barry well away from the United States east coast. More than
two days in advance, the
tropical cyclone discussions began
correctly mentioning Nova Scotia as the most likely land to be affected by Barry.
The largest intensity forecast error was a 30 knot under-
forecast. This was a 12-hour forecast made just before the
rapid, and still unexplained, strengthening on 7 July.
Meteorological observations and best track data in the
vicinity of Nova Scotia were kindly provided by Mr. Allan
MacAfee from the Eastern Canadian Hurricane Center in Halifax, Nova Scotia.