Structure and intensity changes in Hurricane Ophelia

Structure and intensity changes in Hurricane Ophelia (2005):
Coupled atmosphere-ocean modeling and RAINEX observations
Andrew W. Smith
March 2, 2016
1. Introduction
Intensity and intensity change remain one of the most challenging issues
in tropical cyclone (TC) prediction. Complex interactions between the
storm itself and the atmosphere-ocean environment facilitate changes in
intensity vis-à-vis storm structure, which are difficult to both observe and
model. Hurricane Ophelia (2005) demonstrates well the impact of this
storm-environment coupling and its consequences for TC strength and
organization. Ophelia was well observed during the Hurricane Rainband
and Intensity Change Experiment (RAINEX, see Houze et al. 2006) in
2005. Although previous studies have explored linkages between
convective structure evolution and intensity in TCs observed during
RAINEX (Houze et al. 2006, 2009; Judt and Chen, 2010), the role of
coupling between the TC and large-scale atmosphere-ocean environment
to structure-intensity dynamics has not been examined. Understanding
this is important to explaining how Hurricane Ophelia survived extreme
storm-induced ocean cooling, changed structurally during and following
the collapse of its inner eyewall, and to improving model prediction of TC
intensity.
To address the coupling impact, I investigate structure and intensity
changes in Hurricane Ophelia using a high-resolution coupled
atmosphere-ocean model and unprecedented NOAA and U.S. Navy
Electra Doppler (ELDORA) radar, and satellite observations gathered
during the RAINEX field campaign. Aiming to improve model hurricane
intensity prediction, RAINEX comprised a series of flight missions through
Hurricanes Katrina, Ophelia, and Rita to study hurricane eyewallrainband dynamics and interactions (e.g. eyewall replacement cycles,
see Judt and Chen, 2010).
Hurricane Ophelia originated from a weak tropical storm north of the
Bahamas mired in weak environmental steering flow. With limited forward
movement, Ophelia induced strong upper-ocean cooling beneath its
circulation that persisted near the storm’s inner core and was likely
responsible for extremely shallow vertical and asymmetric convection.
2. Coupled Model and Experiments
The University of Miami Unified Wave Interface-Coupled atmosphere- ocean Model (UWIN-CM) consists of component models Weather
Research and Forecasting (WRF) v3.6.1 for the atmosphere, and HYbrid
Coordinate Ocean Model (HYCOM) v2.2.98 for the ocean. UWIN-CM is
configured with 12-,4-, and 1.3 km nested grids and 44 vertical levels for
WRF and 4 km grids for HYCOM. Six-hourly 0.5° NCEP FNL fields and
the daily 1/12° global HYCOM analysis are used for initial and boundary
conditions. Two model couplings were used for these experiments:
uncoupled atmosphere (UA), which holds SST constant, and
atmosphere-ocean (AO), which allows SST to evolve.
3. Effects of Air-Sea Interaction on Storm Structure and Track in
Hurricane Ophelia
a. Track and intensity
Six-hourly storm tracks from UWIN-CM UA and AO simulations reflect
very different TC evolutions (Fig 1). Coupling to the ocean in AO permits
TC awareness and response to storm-induced cooling, and alters the
large-scale environment the TC feels. From Ophelia’s anticyclonic loop
phase onward, AO diverges noticeably – pressure and wind intensity are
overestimated in UA whereas AO responds to storm-induced cooling with
gradual weakening (Fig 1).
b. Storm-induced Upper Ocean Cooling
Prolonged wind stress forcing by TC winds at the ocean surface induces
upwelling of colder sub-surface waters; consequential impact on storm
intensity depends upon forward speed of the TC. Average forward speed
for Atlantic TCs over the period 1948-2003 is approximately 5 m/s (Chen
et al. 2006). Comparatively, due to weak tropospheric steering in the
large-scale environment, Ophelia’s average speed was much slower –
about 2 m/s (Fig 2). As a result of this slow movement, Ophelia induced
upwelling of colder ocean waters beneath its circulation, particularly while
executing its anticyclonic loop September 11-12. National Data Buoy
Center (NDBC) buoy and TRMM Microwave Imager Advanced
Microwave Scanning Radiometer (TMI AMSR-E) satellite SST
observations confirm Ophelia cooled SST by 3-5°C beginning late on
September 10, with the cooling pattern following the storm’s motion
through September 13 (Fig 3-4).
c. Inner eyewall collapse
Storm-induced cooling of sea-surface temperatures beneath TCs disrupts
boundary layer and air-sea interface surface heat flux inputs needed to
maintain deep, moist convection. Late on September 11, GOES-E visible
and infrared satellites captured a localized collapse of convection in
Ophelia’s inner eyewall also realized in UWIN-CM AO simulation (Fig 5).
During this time, Ophelia’s circulation was moving south becoming southsouthwest and was situated above sea-surface temperatures 2-3°C
cooler than the commonly accepted 26.5°C threshold for TC sustenance.
Satellite imagery shows the convection decays on the W/SW portion of
the eyewall likely as a result of this strong cooling and was accompanied
by a sharp drop in wind intensity and steady to rising minimum pressure
(Fig 1).
d. Symmetry & Vertical Structure
Symmetry of the eyewall and distribution of convection at and beyond the
radius of maximum wind (RMW) reflect not only a storm’s inertial stability
but also the environment the TC is interacting with. NOAA P-3 lower
fuselage (LF) radar composites from late on September 11-12 highlight
weakened convective returns in the W/SW quadrant of Ophelia’s eyewall.
Asymmetric eyewall structure observed by RAINEX was clearly seen in
UWIN-CM AO flight-level radar reflectivity during this time, with weaker
and less spatially consistent returns even 24h later when the storm was
no longer above the most intense cooling. Conversely, the UA vortex has
a smaller, more symmetric, and more intense eyewall ring (Fig 6).
Unusually shallow and asymmetric vertical structure was also observed
by NOAA and NRL P-3 ELDORA radar during this time. In the late
afternoon of Sept. 11, flight-level LF composite and vertical transects
show maximum echo-top heights of only 10 kilometers, with strong
convective returns reaching maximum heights of only 4-6 km (Fig 7).
East-west vertical transects of model radar reflectivity for AO and UA
simulations taken at the time of RAINEX observations late on Sept. 11
reveal there are two different storms at this point – a deep (nearly 13 km),
symmetric, outward sloping vortex in UA and a shallower (about 10 km),
asymmetric, and more ragged vortex in AO (Fig 8).
Evaluating vertical structures in the UWIN-CM simulations from a
statistical framework, contoured frequency by altitude diagrams (CFADs)
of radar reflectivity (dBZ) were constructed by binning radar returns by
altitude and normalized by values at each particular level. Given for three
times late on Sept. 9, 11, and 13, 2005, while maximum convective
returns are similar in UA and AO, presence of 20+ dBZ returns at high
altitudes and a pronounced low-level bright brand maximum indicate the
presence of ice and hence, more vigorous vertical motion. Return
frequency at low-mid levels is also skewed further toward convective
returns (35+ dBZ) in UA, even late on Sept. 13 where both storms are
distanced from the location of intense cooling felt earlier by AO (Fig 9).
5. Summary, Conclusions, and Future Work
Improving prediction of hurricane intensity requires better understanding
of how storm-environment interactions incite changes in TC convective
structure. Hurricane Ophelia was a well-observed storm from RAINEX
that demonstrates the importance of these interactions. Weak large-scale
environmental flow facilitated prohibitive SST cooling beneath Ophelia
and the collapse of inner eyewall convection captured only by AO
simulations. This cooling likely contributed to the inner-eyewall collapse
and extremely shallow and asymmetric vertical structures observed in
RAINEX.
Future work will focus on determining the mechanisms responsible for the
cooling in detail, added coupling to the UWIN-CM wave model,
understanding how Ophelia survived the storm-induced cooling and
inner-eyewall collapse via TC internal dynamics and structure changes,
and retrospective assessment of landfall impacts and storm surge
prediction of Hurricane Katrina from the perspective of changing eyewall
and rainband structures in the near-shore and landfall phase.
6. References
Houze, R. A., Jr., et al., 2006: The Hurricane Rainband and Intensity Change
Experiment: Observations and modeling of Hurricanes Katrina, Ophelia,
and Rita. Bull. Amer. Meteor. Soc., 87, 1503-1521.
Houze, R.A., Jr., et al., 2009: Convective contribution to the genesis of Hurricane
Ophelia (2005): Mon. Wea. Rev., 137, 2778-2800.
Judt, F. and S. S. Chen, 2010: Convectively generated potential vorticity in
rainbands and formation of the secondary eyewall in Hurricane Rita of
2005: J. Atmos. Sci., 67, 3581-3599.
Dudhia, J., 1993: A nonhydrostatic version of the Penn State-NCAR Mesoscale
Model. Validation tests and simulation of an Atlantic cyclone and cold
front. Mon. Wea. Rev., 121, 1493-1513.
Grell, G. et al., 1994: A description of the fifth-generation Penn State/NCAR
Mesoscale Model (MM5). NCAR Tech. Note NCAR/TN-398+STR,138
7. Figures
Fig 1. Storm tracks for Hurricane Ophelia from National Hurricane Center (NHC)
best track and UWIN-CM simulations for Sept. 9-14, 2005.
Fig 4. Sea surface temperature (°C) evolution with tracks from UWIN-CM AO
simulation (top) and TRMM TMI-AMSR-E satellite observations with tracks from
NHC (bottom) for Sept. 10-13 1200Z (left to right).
Fig 5. GOES-E 4km visible satellite observed (left), UWIN-CM UA (center), and
UWIN-CM AO (right) clouds from before and during the main portion of the
anticyclonic loop phase (far left) in Hurricane Ophelia from Sept. 11, 2005 (here).
Fig 6. Radar reflectivity and clouds from NOAA P-3 LF composite with GOES-E
4km visible satellite (left), UWIN-CM UA (center) and UWIN-CM AO (right)
simulations for periods during and after the anticyclonic loop phase, Sept. 11-12,
2005.
Fig 7. Shallow and asymmetrical horizontal and vertical structures observed by
RAINEX NOAA P-3 LF composite (left) and ELDORA 0.4 km radar (center, right)
with vertical transect showing shallow maximum echo-top heights in Ophelia late
Sept. 11, 2005 (right).
Fig 2. Forward speed (m/s) from 6-hourly NHC best track for Sept. 9-14, 2005 with
average Atlantic TC speed indicated.
Fig 8. UWIN-CM UA (left) and AO (right) flight-level radar reflectivity and east-west
vertical transect with maximum echo top heights and maximum altitude of strong
convective returns (40 dBZ, denoted by dashed contours) indicated for Sept. 11,
2005 at 1900Z, as in Fig 13.
Fig 3. Storm tracks, sea surface temperature (top right) and surface pressure
(bottom right) observations from NHC and UWIN-CM simulations with NDBC 41002
buoy location.
Fig 9. Contoured-frequency by altitude diagrams (CFADs) for UWIN-CM UA (top)
and UWIN-CM AO (bottom) simulations for Sept. 9 (left), 11 (center), and 13 (right),
2005 1900Z. Black dashed contours denote frequency of 20%.