Extreme Winds and Rapid Degeneration of the May 2004 McMurdo Antarctica Storm: Analysis Using the AM

1
Extreme Winds and Rapid Degeneration of the May
2004 McMurdo Antarctica Storm Analysis Using the
AMPS Forecast Model, AWS Data, and MODIS Data
Daniel F. Steinhoff
Polar Meteorology Group, Byrd Polar Research
Center, The Ohio State University, Columbus, OH
INTRODUCTION The storm of 15 May 2004 was one
of the most severe wind events to strike McMurdo
Base Antarctica (Fig 1a, b) in recorded history.
Wind speed observations are sparse and suspect,
as many anemometers blew away when wind speeds
exceeded 100 mph. Black Island, just south of
McMurdo, recorded a peak wind gust of 144 mph at
2015 UTC. Other stations in the area recorded
similar conditions, up to a gust of 188 mph at
Arrival Heights. Buildings and other
infrastructure at the base were damaged, and
fortunately injuries were avoided because of the
early morning arrival of the storm. The ability
to accurately forecast severe weather events like
this is critical to the success and safety of
operations at McMurdo base. Forecasts are
provided for operations by the Antarctic
Mesoscale Prediction System (AMPS), which is a
joint effort between the National Center for
Atmospheric Research (NCAR) and the Byrd Polar
Research Center (BPRC) at The Ohio State
University. AMPS uses Polar MM5 the
Fifth-Generation Pennsylvania State University
(PSU) NCAR Mesoscale Model, which is adapted
for use in polar regions. Analysis will show
that AMPS provided a very credible forecast for
the 15 May 2004 storm. However a slight bias in
the time of storm onset near McMurdo, along with
an under-prediction of wind speeds, implies that
critical aspects of the storm are not properly
represented by AMPS. Along with Automatic
Weather Station (AWS) data and MODIS imagery,
AMPS is used to diagnose the reasons for
inaccurate prediction of timing and winds, and
also to better understand this complex storm that
dissipated quickly after striking McMurdo.
KATABATIC AND BARRIER WIND REGIMES As the
cyclone propagates westward onto the Ross Ice
Shelf, easterly flow associated with the cyclone
is prevalent over Siple Coast. Downslope flow
over Siple Coast, along with convergence through
glacier passes, forms a katabatic wind regime.
Katabatic winds form when radiatively-cooled
negatively buoyant air near the surface
accelerates downslope. This flow propagates
downslope onto the Ross Ice Shelf (Fig. 5).
Katabatic airstreams appear as warm signatures on
thermal infrared imagery (Bromwich 1989), however
in this case the katabatic airstream is colder
than ambient conditions. As the surface cyclone
propagates westward, winds become more
northeasterly off of Siple Coast, and flow into
the Transantarctic Mountains at the southern edge
of the Ross Ice Shelf. When winds approach a
vertical barrier like the Transantarctic
Mountains, the resulting flow is dependent upon
the wind speed, the height of the barrier, and
the stability of the layer in front of the
barrier. Whether the flow will traverse the
obstacle or flow around it can be determined from
the Froude Number
LOCALIZED EFFECTS NEAR McMURDO The warm thermal
anomaly that formed near Siple Coast early on the
15th is advected with the circulation of the
cyclone around the Ross Ice Shelf. AMPS shows
this warm thermal anomaly at 850 hPa just
southeast of Ross Island (Fig. 7). The actual
storm track derived from MODIS imagery differs
from the storm track that AMPS predicts. The
AMPS storm track is about 100 km northwest and
about an hour ahead of the actual track. With
the displacement of the storm track, the warm
thermal anomaly present near Ross Island at 850
hPa is displaced southeastward. Hence the air
will be colder at 850 hPa in the McMurdo area,
and with relatively warm northwesterly flow into
McMurdo, the region becomes destabilized. Soon
after, the accelerated cold barrier flow regime
reaches McMurdo (Fig. 8), and air converges into
the region to support the convection. AMPS
represents some of the convergence features, but
not others (Fig. 9). The overall southerly flow
into McMurdo by the barrier winds is resolved,
along with the southwesterly katabatic flow down
Koettlitz Glacier. However, due to the incorrect
location of the storm, flow over Minna Bluff will
be more southerly than the southwesterly flow
AMPS predicts (confirmed by observations at Minna
Bluff). This brings flow directly into the
McMurdo area. Also, possible funneling through
Herbie Alley, not resolved on AMPS, can enhance
convergence at McMurdo. The stagnation zone at
Windless Bight, represented in AMPS, is common
with southerly wind events, and causes flow to be
deflected outward (OConnor and Bromwich, 1988).
The deflection of flow originating east of White
Island into McMurdo is not represented in AMPS.
These additional convergence features will
enhance convection in the McMurdo
area. Turbulence from instability and the
downward momentum flux cause the actual surface
winds to be stronger than those forecast by AMPS.
This also explains the temperature difference
between AMPS and observations, as AMPS is about 5
C colder than stations in the McMurdo area.
Turbulence causes the surface air to be
well-mixed and thus warmer compared to AMPS,
where the strong surface inversion holds. Even
after the cold, katabatic/barrier wind regime
reaches McMurdo, turbulence from topographic
internal waves maintains a well-mixed lower
troposphere and prevents the cold surface air
from immediately choking off uplift.
Fr U0(gH(??/ ?))-1/2 Where, for the
Transantarctic mountain range near the
southwestern portion of the Ross Ice Shelf at
15/12Z, U0 17.5 m s-1 g 9.8 m s-2 H 2000
m ? ? 12 K ? 260 K Fr 0.58
The Froude Number is not traverse the range, and will instead flow
northward, based on the balance between the
Coriolis force and the local pressure gradient
force set up by the flow into the barrier.
Combined with the effect of the synoptic cyclonic
circulation, flow along the Transantarctic
Mountains, west of the cyclone, will propagate
northward faster than the cyclone itself. The
cold katabatic airstream becomes entrained into
the circulation of the cyclone and into the
barrier wind circulation (Fig. 6).
Herbie Alley
Fig. 5 Surface wind barbs (Full barb 5 m/s)
at 15/06Z Notice The confluence at the southern
tip of the Ross Ice Shelf (red box).
Fig. 1 Overview map of a) Ross Ice Shelf Region,
b) Inset of a) showing McMurdo/Ross Island area
(heights in meters).
SYNOPTIC OVERVIEW The cyclone initially
develops over the Amundsen Sea early on the 14
May. The surface cyclone intensifies as it is
co-located with a region of positive vorticity
advection at 500 hPa. Surface support is
obtained through deformation of the thermal field
over the same time period. The surface cyclone
propagates over a low-level baroclinic zone, and
deforms the thermal field into an S shaped
pattern (Fig. 2). Warm air is advected ahead of
the cyclone, promoting uplift and forward
propagation. The cyclone deepens by 12 hPa from
14/00Z until it makes landfall over Marie Byrd
Land after 14/12Z (Fig. 3). The cyclone
propagates westward along the northern coast of
Marie Byrd Land until it reaches the Ross Ice
Shelf near Roosevelt Island around 15/03Z. The
cyclone maintains its strength from continuation
of upper level support and a favorable low-level
thermal field. Wind speeds increase due to
vortex stretching as the cyclone descends onto
the ice shelf. Also during this time period,
warm air is advected southward from the
Bellingshausen Sea into West Antarctica. This
flow is funneled through glacier passes and a
warm thermal anomaly forms off of Siple Coast
early on the 15 May (Fig. 4).
Above Left Fig. 7 850 hPa Potential Temperature
at 15/18Z Notice the warm thermal anomaly
southeast of Ross Island. With the observed
storm track being displaced southeastward from
AMPS at this time, actual 850 hPa conditions will
be colder than AMPS, resulting in decreased
stability in the region. Left Fig. 8 Surface
Potential Temperature at 15/21Z Notice the cold
tongue extending northward to Ross Island as
a result of the barrier wind and katabatic wind
regimes. Above Fig 9 Schematic of conditions
in McMurdo region just after 15/18Z. The
southerly barrier/katabatic wind regime results
in convergence near McMurdo (enhanced by
topographical features). Ambient NW flow into
region prior to 18Z, combined with colder air at
850 hPa results in convection.
Fig. 6 Thermal Infrared Image from
15/1320Z Notice the warm signature over the
southeastern Ross Ice Shelf from the katabatic
flow. It appears as a warm signature because
the surface warms from turbulent mixing, breaking
the strong surface inversion. However, the
layer influenced by katabatic flow is colder than
the surroundings, even at the surface. This
cannot be determined from the satellite image due
to clouds obscuring the surface. AWS temperature
observations (not shown) in the katabatic flow
are colder than observations in the general
synoptic flow over the Ross Ice Shelf.
CYCLOLYSIS The southerly cold air flow
associated with the barrier winds cannot traverse
topography west of Minna Bluff, and a secondary
barrier wind regime forms, turning the flow
eastward. This barrier wind flows underneath
southerly winds above that traverse Minna Bluff.
The westerly airstream is highly ageostrophic and
flows into the center of the cyclone, choking off
uplift (Fig. 10). The cyclone then fills in
and cyclolysis occurs.
Fig. 10 Mean Sea Level Pressure and surface wind
barbs for 15/21Z. Notice the ageostrophic flow
west of the cyclone (red box), advecting higher
pressure into the cyclone center.
Above Fig. 2 Surface Potential Temperature at
14/12Z Notice the deformation of the thermal
field in the white box over the Amundsen Sea.
Above Right Fig. 3 Mean Sea Level Pressure at
14/12Z Notice the cyclone in the red box
approaching Marie Byrd Land. Right Fig. 4 850
hPa Potential Temperature at 15/00Z Notice the
warm thermal anomaly along Siple Coast in the
white box.
ACKNOWLEDGEMENTS Bromwich, D.H., 1989 Satellite
analysis of Antarctic katabatic wind behavior.
Bull. Amer. Meteor. Soc., 70, 738-49 OConnor,
W.P and D.H. Bromwich, 1988 Surface airflow
around Windless Bight, Ross Island, Antarctica.
Quart. J. Roy. Met. Soc., 114, 917-938. Observati
ons were obtained from the Antarctic
Meteorological Research Center at the University
of Wisconsin-Madison (Matthew Lazzara). Special
thanks to Dr. Dave Bromwich for his oversight and
support, and to Andy Monaghan for logistical
support throughout the duration of the project.