A WSR-88D Approach to Waterspout Forecasting
LT(jg) Barry K. Choy, NOAA Corps and Scott M. Spratt
National Weather Service Office Melbourne, FL
Abstract
The WSR-88D is being installed at National Weather Service (NWS) forecast
and warning offices and many military installations across the county.
The added capabilities of the WSR-88D over conventional radar provides
the forecaster a multitude of products which allow a more complete interrogation
of small scale weather features. In Florida, waterspouts and weak tornadoes
account for much of the state's severe weather. They have been observed
to form under certain synoptic conditions, most often during the summer
and fall. Along the east-central Florida coast, waterspouts and weak tornadoes
are most frequent in a relatively small area near Cape Canaveral. Observing
and identifying small scale boundary interactions and the intensification
of convective cells in this region using WSR-88D products from the Melbourne
NWS office has proven useful in forecasting these situations.
This paper will begin by providing a brief overview of the waterspout formation
process. It also offers a forecast strategy developed for the east central
Florida coast using specific WSR-88D products to recognize precursor signatures
to waterspout and weak tornado formation. Once a high potential for waterspout
formation exists, a special statement can be issued to heighten public
awareness. An example of such a statement is provided. While the techniques
introduced here were designed for the east-central Florida coast, they
may be applicable at other coastal offices equipped with the WSR-88D.
1. Introduction
The east-central Florida coast is affected by several waterspouts and tornadoes
each year. While the majority of these events tend to be weak and short-lived;
at least when compared to "classic" mesocyclone tornadoes (Schmocker et
al. 1990), they still can be dangerous, and special attention must be given
to prevent injury or fatality. Only through careful analysis of local geography,
environmental conditions, and recognition of certain identifying features
will forecasters be able to effectively warn for such events.
In this paper we will specifically address waterspouts and coastal tornadoes
in the WSO Melbourne coastal warning area. We will review the process of
waterspout/weak tornado formation, presenting brief case studies of recent
events. WSR-88D products will be used to highlight signatures which are
possible precursors to development. Finally, we introduce a forecast strategy
for the Melbourne area, which may be applicable at other coastal WSR-88D
sites.
2. Waterspout/Landspout Formation
Processes
Numerous waterspout studies have been conducted ranging from climatology
of frequencies, strengths and locations, to detailed structural analyses
and formation processes (Golden 1974a, 1974b, 1977; Simpson et al. 1986).
Bluestein (1985) coined the term "landspout" to differentiate between the
"classic" tornado often associated with mesocyclones, and that of the non-mesocylone
induced tornado. Bluestein, and Brady and Szoke (1988) have documented
landspouts over the midwest and western High Plains and comment that these
weak, short-lived tornadoes develop over land in a remarkably similar way
to that of waterspouts. Following Bluestein's convention, we will refer
to waterspouts and weak tornadoes which affect the east central Florida
coast collectively as "spouts".
The principal processes of waterspout formation have been well documented
(Simpson et al. 1986). It is believed that low level vortices produced
along convergent boundaries become stretched in coincidence with updrafts
of developing cumulus or cumulonimbus as they become vertically collocated.
These surface based vortices have been observed to form along shear axes
when a sea breeze interacts with the synoptic surface flow, along outflow
boundaries, or in pertubated flow as a result of land/sea frictional differences
such as down-wind of islands (Hess and Spillane 1990). Low level instability
and lift is further enhanced when boundary interaction occurs, with spouts
often developing near this intersection (Golden and Sabones 1991).
Unlike strong tornadoes associated with "classic" supercell storms, spouts
form in the absence of a mesocyclone. Spouts are believed to spin-up from
the surface by the stretching of a low level vortex in contrast to spin-down
from mesocyclones in "classic" tornadogenesis (Wakimoto and Wilson 1989).
VORTEX (Verification of the Origins of Rotation in Tornadoes EXperiment)
will attempt to confirm these hypotheses in 1994-95 across the central
United States (Rasmussen 1993).
Contrary to the environmental conditions most often associated with stronger
tornado development, large vertical wind shear and baroclinicity are generally
absent in the environment conducive to spout formation. Generally weak
tropospheric winds in a quasi- barotropic environment with high humidity
(mixing ratios of 15-20 g kg-1), and warm surface temperatures (25-31 deg
C) are characteristics of the pre-spout environment (Hess and Spillane
1990).
Most spouts are rapidly developing and dissipating features of short duration.
Golden (1973) determined the average lifetime for waterspouts in the Florida
Keys to be 13 minutes. Spouts can be associated with rapidly developing
cumulus which often do not have indications of precipitation or lightning.
However, stronger and longer lasting spouts have been observed near or
within intense convective cells. In the Florida Keys, ninety-five percent
of all spouts have been observed forming along cumulus congestus cloud
lines, but spouts have occasionally been observed to form beneath individual
cumuli (Golden 1973).
3. East central Florida spouts
Since October 1989, WSO Melbourne has had the warning responsibility for
an 80 nautical mile (nm) (149 km) stretch of the east-central Florida coast
from the Volusia/Brevard County line south to the Indian River/St. Lucie
County line and out to 50 nm (93 km) from shore (Fig.
1). Within this area, a local maxima of spouts is observed near Cape
Canaveral. Almost seventy-five percent of the near coastal spouts reported
to WSO MLB come from the intracoastal lagoons or islands just west and
south of Cape Canaveral and the adjacent offshore waters.
A climatological study conducted by Golden (1977) also indicated a higher
frequency of waterspouts and weak coastal tornadoes near Cape Canaveral
as compared with the adjacent coastal regions immediately north and south
of the area. While there may be a small degree of observer bias in this
data due to population clustering, it is not believed to be significant.
Although the number of events in the Cape Canaveral area remain below that
observed throughout the Florida Keys and portions of the Gulf Coast, damage
occurs along the coast a few times per year on average.
Cape Canaveral is a discontinuity along the coast protruding several miles
east of the surrounding coastline (Fig.
2). Maps and nautical charts of this area reveal that in addition to
the unusual shape and orientation of the coastline, there are several intracoastal
islands and large warm shallow lagoons. These geographic features produce
an area of localized convergence, due to differential heating, which tends
to initiate convective cells under certain synoptic regimes (Zhong and
Takle 1993). This localized convergence, combined with numerous transient
boundaries propagating away from thunderstorm activity over the Gulf Stream
(Fig.
1) and/or the central Florida interior may account for increased spout
occurrence.
With the WSO MLB WSR-88D located approximately 20 nm (37 km) south of the
preferred area of spout development, real-time and post- storm analyses
of these events are now possible. Until now however, even with the WSR-88D,
spouts were not usually detected in real-time and continued to be largely
nowcast/now-warn events based on public observation of funnel clouds or
actively occurring waterspouts or tornadoes. Unfortunately, for such short-lived
events this wait-and-form technique usually proved futile as the event
was often over and damage, if any, had already occurred prior to the warning.
Schmocker et al. (1990) noted that the majority of spouts which have caused
damage in the past were classified as F0 or F1 (Fujita 1981) in intensity
and, in general, have only produced isolated cases of damage to coastal
residences or watercraft. This is fortunate, considering the ever increasing
number of watercraft and housing developments along the coast. The number
of registered boaters in Brevard and Indian River Counties alone totaled
35,418 in year from October 1992 to September 1993.
4. Examples of Spouts
Six cases of near coastal waterspouts and one case classified as an F0
tornado have recently been observed by the MLB WSR-88D. All of these cases
were located near Cape Canaveral, with five of them occurring over the
Indian River, Banana River, or Mosquito Lagoon
(Fig.
2). These spouts occurred during the Summer and early Fall months of
June through October, when surface heating is the predominant mechanism
for cell initiation. All occurred during daylight hours, between 1500 and
0100 UTC (1100 a.m. and 900 p.m. local time).
The WSR-88D creates a number of products each volume scan (every 5 or 6
minutes) from base data. These products are used in the abbreviated post-analyses
presented in the next section to highlight pre-waterspout signatures. For
more information about the WSR-88D and products refer to: Federal Meteorological
Handbook No. 11 1991b; Crum and Alberty 1993; Klazura and Imy 1993.
a. Merritt Island Landspout
The synoptic scale pattern of 24 June 1993 was dominated by a light easterly
surface flow with the Bermuda high located over the Carolinas. Winds were
generally 15 knots (8 m s-1) or less from the surface through the 500 millibar
(mb) level (Fig. 3).
The 2200 UTC 24 June sounding from Cape Canaveral indicated a very deep
moist layer with moderate instability
(Fig.
4), typical for this time of year.
The Melbourne WSR-88D indicated numerous thunderstorms offshore over the
Atlantic, well east of Cape Canaveral throughout the late afternoon. These
storms appear to have initiated a convergence line which later extended
from dissipating storms northwestward toward the coast. By early evening,
showers began to build along the line toward shore with the Composite Reflectivity
(CR) product indicating a narrow band of moderate reflectivities (45-49
dBZ) extending from offshore to the southern edge of Cocoa Beach (Fig.
A).
At 0021 UTC, a stronger echo (50-54 dBZ) covered a small area just south
of the Merritt Island Causeway on the east bank of the Banana River. By
0026 UTC, this echo had shifted west over the central portion of the river,
and by the next volume scan (0031 UTC), the northern two-thirds of Horti
Point was covered by a 50-54 dBZ echo
(Fig.
A). The first public report of a tornado was received around 0035 UTC.
At that time the CR radar product indicated small area of 55-59 dBZ just
west of Sykes Creek (Fig.
2).
The WSR-88D Echo Top (ET) and Vertically Integrated Liquid (VIL) products
likewise indicated rapid cell intensification up until the time of the
event (Fig. 5).
Both the base Velocity (V) product at 1.5 degree elevation and the Storm
Relative Motion (SRM) product at 0.5 degrees revealed evidence of weak
rotation for a few volume scans near the estimated time of the observed
tornado. The indication was however, revealed by only a few pixels within
a region of random inbound/outbound scatter. A post storm ground survey
revealed that several light aircraft, a few trees, and weak structures
were damaged along a short path on Merritt Island.
b. Cocoa Beach Waterspout
Weak tropospheric winds dominated the area as the Bermuda high lay over
south Florida on 23 July 1993 (Fig.
6). The 1700 UTC Cape Canaveral sounding indicated a deep layer of
moisture and light winds (<10 knots 5 ms-1) from the surface to 200
mb (Fig. 7). Strong
sea breeze convergence developed near the coast during late morning and
early afternoon as weak synoptic scale westerlies at the surface collided
with the onshore flow.
Strong thunderstorms developed over the Mosquito Lagoon between 1830 and
1900 UTC and extended along the coast from the eastern tip of Cape Canaveral
southwest to Port Canaveral. An outflow boundary propagating away from
dissipating storms northwest of this region forced an area of offshore
convection, and initiated cells over the coastal lagoons. While the outflow
boundary propagated toward the sea breeze front, cells along the coast
continued to intensify and build southward (Fig.
B). At 1930 UTC, a large waterspout was reported just offshore Port
Canaveral. By that time, WSR-88D radar products indicated the rapid intensification
trend of the spout- producing cell had ceased. VIL, ET, and CR products
indicated the cell maintained nearly steady-state conditions over the next
30 minutes (Fig.
8). Numerous reports indicated the persistence of the large spout as
it moved south to the Cocoa Beach coastal waters and dissipated just prior
to 2000 UTC. Although there were no indications of rotation with this cell
in any of the velocity products, a small area of winds of 26-35 knots (13-19
m s-1) were detected in the 0.5 degree 0.54 nm (1 km) V product. The waterspout
remained offshore during its lifetime and did not produce any damage to
coastal structures or vessels.
c. Port Canaveral Waterspout
Weak easterly surface flow with mid-level ridging across north Florida
characterized the synoptic pattern on 10 August 1993 (Fig.
9). A 500 mb trough off the northeast Florida coast advected slightly
cooler mid-level temperatures into the northern half of Florida, likely
providing additional instability to the moist tropical airmass already
in
place (Fig.
10).
At 1704 UTC, there was no indication of radar-detectable boundaries or
precipitation echoes across the lagoon area south and west of Cape Canaveral.
By 1718 UTC, a moderately strong (55-59 dBZ) CR echo had rapidly developed
just west of the Port Canaveral and continued to intensify through the
next several volume scans (Fig.
C). At 1730 UTC, reports indicated that a waterspout was associated
with this isolated cell. VIL, ET, and CR products confirmed the steady,
rapid intensification of the cell prior to spout formation
(Fig.
11).
Convection was occurring over the central Florida interior and well offshore.
Analyses of all available radar data indicated the cell remained isolated
throughout its lifetime and apparently did not initiate from, or interact
with, any boundaries. The cell also lacked radar-observed rotational characteristics
and enhanced low- level wind fields. The waterspout remained offshore during
its lifetime and did not cause any reported damage.
d. Titusville Waterspout
The synoptic pattern on 25 August 1993 consisted of a weak vertically stacked
low over the central Gulf of Mexico with a inverted trough extending northeastward
across central Florida. A light east to southeast flow was observed at
low and mid levels across the Florida peninsula (Fig.
12). The morning Cape Canaveral sounding revealed abundant moisture
below 600 mb and moderate instability (Fig.
13).
At 1439 UTC, radar indicated a small area of moderately strong (45- 49
dBZ) echoes 7 nm (13 km) east of Titusville. The cell increased to 50-54
dBZ by the next volume scan, and then began to dissipate by 1451 UTC. At
this same time, another cell developed just north of Titusville over the
Indian River. Both cells appear to have developed in response to a boundary
stretching from the Atlantic well offshore of Cape Canaveral, southwestward
across the intracoastal waters. This boundary may have been initiated by
a cluster of thunderstorms offshore of Cape Canaveral.
The new cell continued to intensify and displayed 50-54 dBZ reflectivities
by 1503 UTC (Fig.
D). The following scans indicated the cell contained a 55-59 dBZ core.
A waterspout was reported with this cell at 1515 UTC. VIL, ET and CR products
all indicated maximum values at or just prior to the time of waterspout
formation
(Fig.
14). While the spout was occurring, weak rotation was apparent within
a few pixels in the 0.54 nm (1 km) resolution V product at the 1.5 degree
elevation angle.
The strong cell drifted west with the spout persisting for approximately
5 minutes before moving onshore at Titusville and dissipating. Newspaper
reports indicated that minor wind damage occurred to a few coastal businesses
and several automobile windows were damaged.
e. Banana River Waterspout
Synoptic scale conditions during 27 September 1993 indicated the presence
of a deep layer ridge axis across the Florida straits and Bahamas (Fig.
15). A weakening frontal boundary had dropped into Alabama and Georgia
and was approaching north Florida. During the early afternoon, strong pre-frontal
convection commenced over the north-central and central Florida peninsula.
The 1500 UTC sounding from Cape Canaveral revealed a generally moist airmass
from the surface to 300 mb with a unidirectional south southwest wind flow
(Fig.
16). Winds steadily increased from 10 knots (5 m s-1) at the surface
to 35 knots (18 m s-1) at 300 mb. An easterly sea breeze of around 10 knots
became well established by mid-afternoon along the coast.
At 1913 UTC, radar indicated strong thunderstorms forming over the central
Florida interior and adjacent coastal waters, with thunderstorm activity
over the coast in the southern portion of the Banana River. By 1937 UTC,
the sea breeze boundary (denoted by "1" in Fig.
E) was detected in the CR product by a line of less than 20 dBZ echoes
extending from Vero Beach, north over Merritt Island, to just west of Port
Canaveral. Two additional boundaries (indicated as "2" and "3" in Fig.
E) were detected west of the sea breeze front and were moving southwest
toward the coast. At this time, boundary "2" which extended from Lake Kissimmee
to Cocoa, was only 5 nm (9 km) west of boundary "1". The 2002 UTC volume
scan indicated boundary "2" had begun to intersect the sea breeze front,
from just northeast of Cocoa to 2 nm (4 km) north of Horti Point.
Within 15 minutes, the first convective cell developed over Cocoa and by
2027 UTC a second cell was developing over Port Canaveral. By the next
volume scan at 2032 UTC, the area from Port Canaveral to Cocoa exploded
with strong convective cells which eventually formed into a solid line
and stretched from just offshore of Playalinda Beach to 6 nm (11 km) west
of Melbourne by 2038 UTC. The rapid intensification of the cell is shown
in Fig. 17
with the low- level R, CR, VIL, and ET products all increasing rapidly
prior to, and during the initial stages of spout development.
Public reports indicated that a waterspout developed over the Banana River
about one-half mile south of the 528 Causeway (CWY) at 2030 UTC. Numerous
waterspout reports were received during the next half-hour as the spout
traversed northward over the west basin of Port Canaveral and moved to
a position over the north Banana River before dissipating around 2110 UTC.
No reports of damage were received.
f. Playalinda Waterspout
Within 24 hours of the Banana River waterspout (previous case), another
spout event occurred. By 1200 UTC on 28 September 1993
(Fig.
18), surface winds backed to a northeasterly direction in response
to a cold front moving over the area. Winds above 850 mb remained west
to southwest but weakened from the previous day (Fig.
15). The 1500 UTC Cape Canaveral sounding continued to indicate that
the airmass over the area was extremely moist and unstable
(Fig.
19).
Mid-morning satellite imagery indicated a northeast-to-southwest orientated
cumulus cloud line with embedded convective cells over the Gulf Stream
from 30 nm (56 km) east of Daytona Beach to just north of Playalinda Beach.
Throughout the morning, weak radar echoes below 20 dBZ were largely masked
by an extensive area of Anomalous Propagation (AP) surrounding the radar
site. The AP was indicative of the highly stable and moist morning atmosphere
beneath the nocturnally induced radiation inversion.
By late morning, after filtering out low reflectivity values, a few pixels
of 25 dBZ CR echo were detected, extending from an area of offshore convective
activity to the coast 3 nm (6 km) north of Playalinda Beach (Fig.
F1). These echoes appeared to be associated with a weak low-level boundary.
At 1549 UTC, a scattered area of 20-25 dBZ was indicated from Titusville
to 15 nm (28 km) east of Playalinda Beach. By the next volume scan (1555
UTC), a small echo over the Mosquito Lagoon, 2 nm (4 km) north of Playalinda
Beach had rapidly increased to an area of 45-50 dbZ (arrow, Fig.
F2). A brief waterspout, as described by public reports, occurred with
this cell beginning around 1558 UTC. The short-lived cell quickly weakened
to 30 dBZ by 1601 UTC and completely dissipated by the following volume
scan (1607 UTC).
VIL, ET, and CR products all peaked prior to waterspout generation, and
showed a steady decreasing trend once the spout dissipated
(Fig.
20). There were no reports of damage with this spout.
g. North Cape Canaveral Waterspout
The synoptic conditions on October 22, 1993, were similar to the previous
two events. A weak front was positioned over north Florida with prefrontal
conditions dominating the area over central Florida and the adjacent coastal
waters (Fig. 21).
The Cape Canaveral sounding indicated the airmass over the area was extremely
moist and unstable at 1500 UTC (Fig.
22).
During the morning hours, showers and thunderstorms were occurring over
the Gulf Stream about 50 nm (93 km) east of Daytona Beach and Cape Canaveral
- a continuation from activity the night before. A persistent stationary
boundary, indicated by a line of cumulus congestus, was evident in the
visible satellite imagery and the lowest elevation R and CR products. This
boundary extended from the area of showers and thunderstorms offshore over
the Gulf Stream southwestward to Cape Canaveral.
Convection also began to develop over central Florida as surface temperatures
climbed into the mid 80's over much of the area. By 1542 UTC showers had
begun to develop inland along the southwestern edge of the boundary and
an initial cell formed offshore, just east of Cape Canaveral (Fig.
G1). By 1548 UTC, the CR product indicated 50-54 dBZ reflectivities
in this cell. A SKYWARN observer reported a waterspout offshore just east
of Cape Canaveral commencing around 1552 UTC. The spout was also reported
by observers at the Kennedy Space Center soon thereafter.
The volume scan at 1600 UTC (Fig.
G2) indicated that CR values continued in the 50-54 dBZ range, decreasing
to 45-49 dBZ by 1606 UTC. The spout producing cell maintained its intensity
briefly before rapidly dissipating, which was apparent in the tendencies
of the ET, VIL, and CR values (Fig.
23). This brief spout did not threaten the coast or cause damage to
vessels.
5. Event Synopsis
Six of the seven spouts occurred under weak tropospheric winds. The remaining
case, the Banana River spout, occurred under a slightly stronger flow,
with winds in the 10 to 20 knot range from the surface to 500 mb. Deep
moisture was present for all cases and surface flow along the coast was
east to southeast during five of the events. During the other two events,
the synoptic surface flow was south to southwest with a northeast to east
sea breeze flow along the coast. All spouts occurred with cell development
in the Cape Canaveral area as the cells were rapidly developing.
In four of the cases, distant convective activity appears to have initiated
low-level boundaries near or along which the spout- producing cells later
developed. It was apparent only in the Cocoa Beach and Banana River waterspouts
that two boundaries actually intersected. However, for the Merritt Island
landspout case there were several boundaries indicated by radar near the
area and these may have influenced spout formation. The time between initial
cell detection and the first report of a spout sighting was around 25 minutes
for each case.
After careful review of the WSR-88D archived products, weak rotation was
noticed near the 2500 ft (758 m) level for two of the seven cases. The
rotation was less than 30 knots (15 m s-1) storm relative, over a small
area, and was observed only for a few volume scans during each event. This
rotation was only apparent in the 0.54 nm (1 km) resolution 1.5 degree
V and SRM products. A small area of radial winds in the 26 to 35 knots
(13 to 18 m s-1) range was indicated in the 0.5 and 1.5 degree elevation
V and SRM products for the Cocoa Beach, Banana River, and Cape Canaveral
waterspouts, but no rotation was apparent.
The VIL, ET, and CR products all showed steady increasing trends prior
to spout formation, indicative of rapid cell development. The spouts were
first observed around the time of maximum VIL, ET, and CR values and dissipated
as parent storms weakened. Each of the observed cells increased to reflectivities
around 55 dBZ with moderate VIL values, and generally possessed low tops
as compared to convective cells over land. These values are not unusually
large when compared with other storms across east central Florida; rather,
it appears the trend in intensity is a significant factor.
Paxton and Shepherd (1993) proposed the development of WSR-88D algorithms
tailored to severe weather detection for central Florida. Digitized WSR-57
data (RADAP) used in their study suggested that positive VIL growth rates
of 20 kg m-2 or greater in a ten minute period, or a positive echo top
height increase of 15,000 feet (4550 m) or more within 10 minutes frequently
led to severe weather. Data from the present study confirms that a VIL
and ET growth algorithm would be helpful in delineating potential spout
events. Two of the seven spouts analyzed by the MLB WSR-88D exceeded the
Paxton and Shepherd thresholds. Both of these cases involved relatively
long-lived waterspouts. The remaining cases approached the VIL and ET severe
weather criteria.
Additional research using a larger data set is needed to further evaluate
the prospects for such an algorithm. Forecasters at the Melbourne and Tampa
Bay WSO's are planning to examine this topic in more detail during future
years.
6. Forecast strategy
A systematic approach to coastal spout forecasting is needed to increase
the forecasters awareness of development potential and to convey this information
to the public prior to occurrence.
1) The initial strategy is to determine whether the synoptic conditions
are favorable for spout development since they are contrary to the environment
for strong tornadoes. The latest sounding should indicate moisture through
a deep layer from the surface to around 500 mb and a weak tropospheric
wind field, favoring an onshore flow at low-levels.
2) Once the synoptic scale environment is determined to be in place,
detecting the presence and movement of boundaries and rapid cell intensification
becomes very useful in determining the potential for spout formation. The
WSR-88D "Combine Up" function is useful in highlighting boundaries in the
CR and lowest elevation R products. This function can be used alone, or
in conjunction with filtering the lowest one or two reflectivity levels
to better enhance boundaries.
3) Rapid cell development can be detected by an increasing trend in
the ET, VIL and CR products. Noticing this development over the preferred
spout area, especially along or near boundaries and intersections, can
be important when trying to predict the location and occurrence of spouts.
Although these radar products are highlighted, a number of other radar
products - for example, Reflectivity Cross Sections and base R for various
elevation angles - are also useful in detecting development in cells. Satellite
imagery and surface observations can also confirm radar characteristics.
4) Once conditions suggest a high potential of spout formation,the WSR-88D
Routine Product Set (RPS) list should be updated (or a special spout RPS
list should be implemented) to include some high resolution products. In
addition to developing a spout RPS list, user functions (macros) can be
created in advance to display high resolution four-panels of the lowest
elevation angle V and SRM products. The 0.13 nm (0.25 km) V and SRM products
at elevation angles which will bisect developing cells generally below
4000 feet should be most useful in tracking the location of spouts (or
more likely an area of strong winds) after initial development.
Rotational characteristics of spouts typically are about an order of magnitude
smaller than the highest resolution radar product available, therefore,
associated cyclonic signatures are usually unresolvable. Also, since rotation
is most prevalent at very low levels within the parent cell, it is unlikely
that velocity analyses beyond about 30 nm (56 km) of the radar site will
be helpful. However, the base Spectrum Width (SW) product may prove useful
to highlight or confirm spout presence through the indication of isolated
strong turbulence and should be included on the spout RPS list. While the
VIL and ET products seem to perform well as storm intensification indicators,
spout-producing cells often are not the strongest cells as indicated by
radar. Subtle increases of CR, ET and VIL values combined with areas of
localized convergence and approaching boundaries remain the key identifying
factors.
Response time during a spout is paramount. When a spout develops, radar
products from the lowest elevation scans must be interrogated closely.
If possible, a Volume Coverage Pattern (VCP) RESTART command should be
considered (at the Unit Control Position, UCP) after the 2.4 or 3.4 deg
scan to allow for immediate re-evaluation of the lowest level products.
For warning purposes, the NWS does not differentiate between spouts and
strong tornadoes, since even weak spouts have some potential for serious
injury, if not death. A tornado warning is required if a tornado is sighted,
indicated by radar, or if a waterspout threatens the coast. For non-threatening
waterspouts well offshore, a Special Marine Warning is normally sufficient
to cover the event. For coastal spouts, however, tornado warnings frequently
are issued after the initiation of the event once public reports are received.
Many times the spout has dissipated by the time the warning is heard by
the public, and therefore only adds confusion to the situation.
For the benefit of coastal residents and mariners, it would be useful to
heighten awareness ahead of a potential spout event, at least under certain
conditions. Suggested conditions are:
1) When certain synoptic patterns occur which are indicative, or
supportive of spout formation.
2) When a developing line of convection is approaching or forms over
the area and nearby boundaries are detected.
3) When rapid initiation of isolated convection is occurring over an
area which has been determined to be prone to spout formation.
To alert the public in advance to the possibility of spout formation, a
Marine Weather Statement and/or Special Weather Statement (MWS or SPS)
can be issued along the lines of the following example.
...CONDITIONS ARE FAVORABLE FOR WATERSPOUT FORMATION OVER
THE BANANA RIVER BETWEEN PORT CANAVERAL AND THE MERRITT ISLAND CAUSEWAY...
AT 100 PM...NATIONAL WEATHER SERVICE RADAR INDICATED A RAPIDLY
DEVELOPING LINE OF SHOWERS OVER THE BANANA RIVER 1 MILE WEST OF THE PORT
CANAVERAL LOCKS. THESE SHOWERS WERE BUILDING SOUTH TOWARD THE MERRITT ISLAND
CAUSEWAY ALONG THE SEABREEZE FRONT. SHOWERS DEVELOPING IN THIS AREA UNDER
SIMILAR CONDITIONS HAVE OFTEN PRODUCED FUNNEL CLOUDS OR WATERSPOUTS. MARINERS
AND COASTAL RESIDENTS ARE ADVISED TO MAINTAIN A CLOSE WATCH FOR FUNNEL
CLOUD AND WATERSPOUT FORMATION IN THIS AREA AND MONITOR THIS DEVELOPING
WEATHER SITUATION OVER THE NEXT FEW HOURS.
IF YOU SIGHT A FUNNEL CLOUD OR WATERSPOUT NEARBY...SEEK SAFE HARBOR
AND SHELTER IN A REINFORCED BUILDING...AND REPORT THE SIGHTING TO THE NATIONAL
WEATHER SERVICE...U.S. COAST GUARD...OR TO YOUR LOCAL LAW ENFORCEMENT OFFICIALS.
Purposes of this statement are to:
1) Heighten awareness of potentially severe weather, since effective
advanced warning of the actual condition (spout) is often not possible.
2) Reduce the time needed for response, should a warning become necessary.
3) Provide mariners, especially those is slow vessels, enough time to
avoid an area of potential development.
4) Encourage timely reports from the public, should an event occur.
Issuing such a statement does not eliminate the necessity for warning once
an event occurs, but does offer an opportunity to heighten awareness and
prepare the public for a possible event. It is also important to note that
there will be more potential situations than actual events, so great care
must be taken not to overuse such statements and risk reducing their effectiveness.
Although this statement has been tailored to WSO Melbourne, it is hoped
that as future WSR-88D sites become operational along the southern United
States coastline, additional offices will experiment with statements such
as this to forewarn the public of potentially hazardous spouts.
7. Conclusion
The WSR-88D now provides the forecaster with the added capability of detecting
radial wind velocities, in addition to reflectivity and other useful products.
There are still limitations to this new technology; and recognizing these
limitations will allow for better use of these products in forecasts and
warnings.
In the WSO Melbourne coastal warning area, most of the past waterspout
and coastal landspout formation has been isolated to a small area just
west and south of Cape Canaveral. The warm, shallow and relatively large
intracoastal lagoons and shape of the coast tend to localize convergence
in this area. This produces a favorable location for thunderstorm initiation.
On occasion, low level boundaries propagating through this area interact
with rapidly developing cells to produce spouts. Due to the weak radar
signatures and short-lived nature of such events, they have generally been
warned for only after being sighted and often enough even after they have
dissipated.
The spouts presented in this paper can be categorized into two groups -
the stronger and longer lasting ones which form along boundary intersections
indicative of the Cape Canaveral and Banana River cases. It is important
to note that strong (F2+) spouts have formed with mesocyclone producing
storms (Golden and Sabones 1991), but these are rare and are not addressed
in this paper.
While it is well known that mesocyclones associated with supercells can
be detected at great distances from a WSR-88D, any circulations associated
with spouts often remain undetected since rotational signatures occur at
lower levels, are much weaker, and are smaller in horizontal/vertical scale.
Detailed post-analyses of seven recent spout events within 30 nm (56 km)
of the Melbourne WSR-88D indicate only a hint of weak rotation in two of
the cases. This rotation was only noticed in the 0.54 nm (1 km) resolution
(best resolution on RPS list at the time) 1.5 degree elevation V and SRM
products, and only for a few volume scans during the actual event.
Velocity signatures were only indicated with a few pixels and were embedded
within an area of random inbound/outbound velocities, making real-time
detection very difficult, and warning nearly impossible if based on velocity
signatures alone. Even with the use of higher resolution velocity products,
detection of low level circulations may prove difficult, since rotational
velocity signatures of spouts typically are too small for the radar to
decipher.
Detecting boundaries however, is a more common and easier task for the
forecaster. Close monitoring of boundaries and anticipating the intersection
or interaction among boundaries and rapidly developing cells near coastal
regions is essential in identifying areas where spouts are likely to form.
In addition, rapid, but subtle,increases in cell ET, VIL and CR often occur
prior to the development of spouts. These two factors in combination with
knowledge of local climatology have allowed forecasters at Melbourne to
successfully provide information to the public about potential waterspout
formation.
In this paper we have presented seven case studies which identify synoptic
and mesoscale conditions favorable for coastal spout development. By combining
local spout climatology studies with the knowledge of coastal geography
and basic meteorological principles, the meteorologist using the WSR-88D
may be able to provide coastal residents and mariners with additional lead-time
concerning potentially dangerous spout events. To heighten public awareness
for the potential of spout formation, a Marine Weather Statement or Special
Weather Statement can be issued for a portion of the coastline. We have
provided an example of such a statement which has been used at Melbourne
and may be applicable at other locations.
Acknowledgements
We would like to thank the entire WSO Melbourne staff for their efforts
in collecting and archiving radar products. Special thanks to Don Baker
(formerly the WSO Science and Operations Officer) who suggested the need
for a local spout study and helped with the initial literature search and
to Delbert Matney for teaching us numerous data processing techniques.
MIC Bart Hagemeyer, Andy Nash, Dan Petersen, and the Southern Region (SR)
Scientific Service Division (SSD) provided many improvements to the text.
Sounding data was provided by forecasters at Base Weather Operations, Cape
Canaveral Air Force Station.
References
-
Bluestein, H.B., 1985: The formation of a "landspout" in
a "broken-line" squall line in Oklahoma. Preprints, 14th Conf. on Severe
Local Storms, Amer. Meteor. Soc., Indianapolis, 267-270.
-
Brady, R.H., and E.J. Szoke, 1988: The landspout - a common
type of northeast Colorado tornado. Preprints, 15th Conf. on Severe Local
Storms, Amer. Meteor. Soc., Baltimore, 312-315.
-
Crum T.D., and R.L. Alberty, 1993: The WSR-88D and the operational
support facility. Bull. Amer. Meteor. Soc., 74, 1669-1687.
-
Federal Meteorological Handbook No. 11, 1991b: Doppler radar
meteorological observations, Part C, WSR-88D products and algorithms. FCM-H11C-1991,
Interim Version One. Office of the Federal Coordinator for Meteorological
Services and Supporting Research, Rockville, Maryland, 210 pp.
-
Fujita, T., 1981: Tornadoes and downbursts in the context
of generalized planetary scales. J. Atmos. Sci., 38: 1511-1534.
-
Golden, J.H., 1973: Some statistical aspects of waterspout
formation. Weatherwise, 26, 108-117.
-
----, 1974a: The life cycle of Florida Keys waterspouts.
I. J. Appl. Meteor., 13, 676-692.
-
----, 1974b: Scale-interaction implications for the waterspout
life cycle. II. J. Appl. Meteor., 13, 693-709.
-
----, 1977: An assessment of waterspout frequencies along
the U.S. East and Gulf Coasts. J. Appl. Meteor., 16, 231-236.
-
----, and M.E. Sabones, 1991: Tornadic waterspout formation
near intersecting boundaries. Preprints, 25th Inter. Conf. on Radar Meteorology,
Amer. Meteor. Soc., Paris, France, 420-423.
-
Hess, G.D., and K.T. Spillane, 1990: Waterspouts in the gulf
of Carpentaria. Aust. Meteor. Mag., 38, 173-179.
-
Klazura, G.E., and D.A. Imy,1993: A description of the initial
set of analysis products available from NEXRAD WSR-88D System. Bull. Amer.
Meteor. Soc., 74, 1293-1311.
-
Paxton, C.H., and J.M. Shepherd, 1993: Radar diagnostic parameters
as indicators of severe weather in central Florida. NOAA Technical Memoranda
NWS SR-149, 12 pp.
-
Rasmussen, E., 1993: VORTEX the Verification of the Origin
of Rotation in Tornadoes EXperiment. Tornado experiment planning document
NSSL, 1-54.
-
Schmocker, G.K., D.W. Sharp and B.C. Hagemeyer, 1990: Three
initial climatological studies for WFO Melbourne, Florida: a first step
in the preparation for future operations. NOAA Tech. Memo. NWS SR-132.,
52 pp.
-
Simpson, J., M.C. McCumber and R.S. Penc, 1986: Observations
and mechanisms of GATE waterspouts. J. Atmos. Sci., 43, 753-782.
-
Wakimoto, R.M., and J.W. Wilson, 1989: Non-supercell tornadoes.
Mon. Wea. Rev., 117, 1113-1140.
-
Zhong, S., and T.S. Takle, 1993: The effects of large-scale
winds and sea-land-breeze circulations in an area of complex coastal heating.
J. Appl. Meteor., 32, 1181-1195.
Back to Spratt publication
page.