Cyclone Phase Space:
One method to diagnose current
& forecast cyclone structure
or
“Terapia para o ciclone com uma crise de identidade”
“Gustav” 2002
“Catarina” 2004
Motivation
• Cyclones are not simply “tropical” or “extratropical”
• There is a great range of hybrid-type cyclones
• These are often the most challenging since we do not
have conceptual models for them
• Do these cyclones exist because of competing energy
sources?
• We will take a fresh look at the range of cyclone
structure and propose a method to classify them all
Test: Separate the 5 tropical cyclones from the 5 non-tropical.
Images
courtesy
NCDC
Some relevant questions…
• What makes a cyclone warm or cold-core?
• If all low pressure areas result from a column of air that
is on average warmer than its environment, how can
there be cold-core cyclones?
• What are the hydrostatic consequences of this
thermodynamic structure & the resulting profile of
cyclone “strength”?
• What about existence of mixed phase cyclones?
• Why do we care?
60 knots is 60 knots!
Some practical issues related to
structure
The benefits and drawbacks of relying on
climatology
Model interpretation: What type of development?
PMIN=1009hPa
PMIN=1001hPa
PMIN=1003hPa
PMIN=1005hPa
Why is the phase of a cyclone important?
• Predictability is a function of cyclone phase
• Model interpretation/trust is a function of phase
• It is often not at first apparent what the model is
forecasting, or the nature of cyclone development
• Peak intensity is a function of cyclone phase
Additional relevance: Predictability
Additional relevance: Predictability
Non-conventional cyclones: Examples
1938 New England Hurricane
?
940hPa
Pierce 1939
•
Began as intense tropical cyclone
•
Rapid transformation into an intense
hybrid cyclone over New England
(left)
•
Enormous damage ($5-10 billion
adjusted to 2008). 10% of trees downed
in New England. 600+ lives lost.
•
Basic theories do not explain a frontal
hurricane
Example of nonclassic structure
21 December 1994
22 December 1994
23 December 1994
24 December 1994
12
Non-conventional cyclones: Examples
Christmas 1994
Hybrid New England Storm
NCDC
•
Gulf of Mexico extratropical cyclone that
acquired partial tropical characteristics
•
A partial eye was observed when the
cyclone was just east of Long Island
•
Wind gusts of 50-100mph observed across
southern New England
•
Largest U.S. power outage (350,000) since
Andrew in 1992
•
Forecast 6hr earlier: chance of light rain,
winds of 5-15mph.
Non-conventional cyclones: Examples
Catarina (2004)
NCDC
•
Demonstrates that we
cannot rely purely on
the historical record to
diagnose and forecast
structure
•
What was Catarina?
•
We can say it “looks”
more like a hurricane
than a significant
number of north
Atlantic hurricanes!
Summary of relevance:
• Classification
• Better understanding of the current state
• Applying conceptual models
• The type/extent of expected impact/damage
• Quantifying potential for intensity change and uncertainty
– How can intensity change be forecast if there is great structural uncertainty?
• Amount of intrinsic (mis)trust of numerical model
forecasts
The end result:
We need a diagnosis of basic
cyclone structure that is more
flexible than only tropical or
extratropical
Goal:
A more flexible approach to cyclone characterization
To describe the basic structure of tropical,
extratropical, and hybrid cyclones simultaneously
using a cyclone phase space.
Phase Space
Parameter A
What parameters to use?
• Maximum wind? Minimum pressure?
– Too simple & does not discriminate
• Vorticity?
– What level?
• Potential vorticity?
– Separating vorticity and stability is important
• Q-vectors?
– Does not simplify the continuum
• Energetics?
– Ideal, but not practical in real-time
• Something more basic: thermal wind & asymmetry
Let us begin with a review the structure of the
text-book cyclone types
Classic warm-core cyclone: TC
Hurricane Bonnie (1998) Temperature Anomaly
-
12km
+
6km
1km
Image courtesy Mark DeMaria, CIRA/CSU
www.cira.colostate.edu/ramm/tropic/amsustrm.asp
Low pressure results
from column of air on
average warmer than
environment, with the
anomalous warmth in
the troposphere
Source:
Advanced Microwave
Sounder (AMSU)
Temperature Anomaly
Classic warm-core cyclone: TC
TC Height Field (m)
from hydrostatic
balance
Warm: expansion of
surfaces
Cold: contraction of
height surface
Classic warm-core cyclone: TC
Height anomaly from
zonal mean shaded
Height anomaly
increases with
altitude in
troposphere
Classic warm-core cyclone: TC
• Intensifies through: sustained convection, surface fluxes.
• Cyclone strength greatest near the top of the PBL
 Gradient wind balance in a convective environ.
-
+
Cold
Stratosphere
Z
Troposphere
W
a
r
m
L
Height anomaly
Classic cold-core cyclone: Extratropical
Cleveland Superbomb Temperature Anomaly
Low pressure results
from column of air
on average
warmer than
environment, with
the anomalous
warmth in the
stratosphere
L
2.5 NCAR/NCEP reanalysis
Classic cold-core cyclone: Extratropical
Height anomaly from
zonal mean shaded
Height anomaly
decreases with
altitude in
troposphere
Classic cold-core cyclone: Extratropical
• Intensifies through: baroclinic development, tropopause
lowering.
• Cyclone strength greatest near tropopause
 QG balance in a minimally convective environ
Cold
-
+
Warm
Stratosphere
Z
Troposphere
Cold
Warm
L
Height anomaly over sfc center
Hybrid (non-conventional) cyclone
What if an occluded extratropical cyclone moves over warm water?
Characteristics of tropical and extratropical cyclones.
Stratosphere
Z
+
Warmer
Troposphere
Colder
Warmer
L
Height anomaly over sfc center
Cyclone Parameters 1 and 2: Vertical structure
-VT: Thermal Wind [Warm vs. Cold Core]
Warm core
Hybrid
Cold Core
300mb
-
+
-
+
-
+
600mb
900mb
Height anomaly
Height anomaly
Height anomaly
Cyclone Parameter -VT: Thermal Wind
Warm-core example:
Hurricane Floyd 14 Sep 1999
Z
Z
Z
Z
Z
Z
Vertical profile of
ZMAX-ZMIN is proportional
to thermal wind (VT).
 ( Z MAX  Z MIN )
  | VT |
 ln p
Two layers of interest
300hPa
 ( Z MAX  Z MIN )
  | VTU |
 ln p
600hPa
Z
Z
Z
600hPa
 ( Z MAX  Z MIN )
  | VTL |
 ln p
900hPa
Cyclone Parameter -VT: Thermal Wind
Cold-core example:
Cleveland Superbomb 26 Jan 1978
300hPa
 ( Z MAX  Z MIN )
  | VTU |
 ln p
600hPa
600hPa
 ( Z MAX  Z MIN )
  | VTL |
 ln p
900hPa
Third dimension?
• We now have two parameters of the CPS that
discriminate the vertical structure of a cyclone:
warm vs. cold vs. hybrid core
• What about the horizontal structure?
• How do we separate the horizontal structure of the
various types of cyclones?
• Ultimately, a good measure is frontal nature
(baroclinic vs. barotropic structure)
Cyclone Parameter 3: Horizontal structure
B: Thermal Asymmetry
Symmetric
Hybrid
Asymmetric
Cyclone Parameter B: Thermal Asymmetry
• Defined using storm-relative 900-600hPa mean
thickness field (shaded) asymmetry within 500km
radius:
B  Z 600hPa  Z 900hPa  Z 600hPa  Z 900hPa
R
L
L
Cold
B >> 0:
Frontal
Warm
B0: Nonfrontal
Cyclone Parameter B: Thermal Asymmetry
Conventional Tropical cyclone: B  0
Forming
L
Mature
Decay
L
L
Conventional Extratropical cyclone: B varies
Developing
L
B >> 0
Mature
L
B>0
Occlusion
L
B0
Constructing Phase Space
Constructing 3-D phase space from cyclone
parameters: B, -VTL, -VTU
A trajectory within 3-D generally too
complex to visualize in an
operational setting
 Take two cross sections (slices) :
B
-VTU
-VTL
-VTL
Hurricane Mitch (1998)
Case of symmetric, warm-core development and decay
Classic tropical cyclone
Symmetric warm-core evolution: Hurricane Mitch (1998)
Slice 1: B Vs. -VTL
Symmetric warm-core evolution: Hurricane Mitch (1998)
Slice 1: B Vs. -VTL
Symmetric warm-core evolution: Hurricane Mitch (1998)
Slice 2: -VTL Vs. -VTU
Upward
warm core
development
maturity, and
decay.
With
landfall,
warm-core
weakens
more rapidly
in lower
troposphere
than upper.
Symmetric warm-core evolution: Hurricane Mitch (1998)
Slice 2: -VTL Vs. -VTU
Upward
warm core
development
maturity, and
decay.
With
landfall,
warm-core
weakens
more rapidly
in lower
troposphere
than upper.
December 1987 Extratropical Cyclone
Case of asymmetric, cold-core development and decay
Classic occlusion of an extratropical cyclone
Asymmetric cold-core evolution: Extratropical Cyclone
Slice 1: B Vs. -VTL
Asymmetric cold-core evolution: Extratropical Cyclone
Slice 2: -VTL Vs. –VTU
Hurricane Floyd (1999)
Multiple phase evolution:
Case of extratropical transition of a tropical cyclone
5
Warm-to-cold core transition:
Extratropical Transition of Hurricane
Floyd (1999): B Vs. -VTL
4
3
B
5
2
4
3
1
-VTL
2
1
Warm-to-cold core transition:
Extratropical Transition of Hurricane Floyd (1999)
B Vs. -VTL
B
Provides
for objective
indicators of
extratropical
transition
lifecycle.
Extratropical transition
ends when –VTL < 0
Extratropical transition
begins when B=10m
-VTL
Spaghetti Plot of 34 Cyclone Phase Trajectories based
upon Navy NOGAPS operational analyses
960hPa 970hPa 980hPa 990hPa 1000hPa 1010hPa
Asymmetric
Cold-Core
Asymmetric Warmcore
B
Symmetric Cold-core
Symmetric Warm Core
-VTL
Composite Mean ET Structural Evolution Summary
34-Cyclone Composite Mean Phase NOGAPS-analysis based
Trajectory with key milestones labeled
TE+24h
TE
TMID
TE+48h
TB
TE+72h
TB-24h
TB-72h
TB-48h
Variability About the Composite Mean
Boxes represent the calculated one standard deviation spread about the
34-cyclone consensus mean trajectory for each time
Considerable
variability
about mean
once transition
completed=>
posttropical
phase can take
many forms….
Floyd (1999): Non-intensifying
cold-core development
Hugo (1989): Explosive cold-core
development
Charley (1986): Schizophrenia
Cindy (1999): Absorption.
Dennis (1999): “ET-Interruptus”.
Keith (1988): Explosive warmseclusion development
Hurricane Olga (2001)
Multiple phase evolution:
Case of tropical transition of a cold-core cyclone
Cold-to-warm core transition:
Tropical Transition of Hurricane Olga (2001)
-VTU Vs. -VTL
Cold-to-warm core transition:
Tropical Transition of Hurricane Olga (2001)
-VTU Vs. -VTL
-VTU
Tropical transition
begins when –VTL > 0
(subtropical status)
-VTL
Tropical
transition
completes when
–VTU > 0
(tropical status)
Catarina: What was it in this context?
Summary of cyclone types within the phase space
Summary of cyclone types within the phase space
?Polar lows?
Phase space limitations
• Cyclone phase diagrams are dependent on the quality of the
analyses upon which they are based.
• Three dimensions (B, -VTL, -VTU) are not expected to explain all
aspects of cyclone development
• Other potential dimensions: static stability, long-wave pattern, jet
streak configuration, binary cyclone interaction, tropopause
height/folds, surface moisture availability, surface roughness...
• However, the chosen three parameters represent a large
percentage of the variance & explain the crucial structural
changes.
Important Caveats:
Often model analysis representation is poor
Important Caveats:
Warm Seclusion Dilemma
• At middle to higher latitudes, often winter-type cyclones will
develop tropical-type structure
– Warm core
– Eye
– Extreme intensity
• These cyclones often are “the worst of both worlds”
– Tropical
– Extratropical
• They represent the largest forecast problems for oceanic
forecasting
Forecast South Atlantic Warm Seclusion this week
Cold-Core Post-transition ET
Warm Seclusion
Post-transition ET
Surface wind field of a warm seclusion
40kt, 75kt, 90kt
Figure courtesy Ryan Maue, FSU
Vertical structure of a warm seclusion
Warm core to the west, cold core to the east!!
Figure courtesy Ryan Maue, FSU
Revisit: Why is the phase of a cyclone important?
• Predictability is a function of cyclone phase
• Model interpretation/trust is a function of phase
• It is often not at first apparent what the model is
forecasting, or the nature of cyclone development
• Peak intensity is a function of cyclone phase
Cyclone Phase Forecasting &
Real-time Web Site
Real-time Cyclone Phase Analysis & Forecasting
• Phase diagrams produced in real-time for
various operational and research models.
• Provides insight into cyclone evolution that
may not be apparent from conventional
analyses
• Web site: http://moe.met.fsu.edu/cyclonephase
• Also available a historical archive of CPS
diagrams for nearly 100 cyclones
Cyclone Phase Web Page Overview
• Model analyses and forecast-based phase
diagrams:
–
–
–
–
–
–
–
–
GFS
CMC
GFDL
HWRF
MM5 (FSU)
NAM
NOGAPS
UKMET
(0,6,12,18 UTC)
(0,12 UTC)
(0,6,12,18 UTC)
(0,6,12,18 UTC)
(0,12 UTC)
(0,6,12,18 UTC)
(0,12 UTC)
(0,6,12,18 UTC)
70
Cyclone Phase Web Page Overview
• Trajectory through phase space describes structural evolution
–
–
–
–
–
–
A = When cyclone was first detected
C = Current analysis time
Z = Cyclone dissipation time or end of model forecast data
AC = cyclone structural history
CZ = cyclone structural forecast
Date is labeled at 00Z along phase trajectory
• Color of trajectory gives cyclone intensity in MSLP
• Size of marker gives average radius of 925hPa gale-force wind
• Cyclone track & underlying SST provided in inset
• Phase diagram quadrants are shaded to give more rapid
interpretation
71
Example real-time cyclone availability for GFS
Example GFS forecast oceanic extratropical cyclone
Forecast phase analysis
Zoomed
Ensembling
Structural Predictability
Sensitivity to initial conditions & physics
• Often there is phase dependency on the type of
data assimilation or model physics
11 November 2003 GFDL vs GFS
AVN
Multiple model solutions:
Measure of structural forecast uncertainty
Multiple model solutions:
Measure of structural forecast uncertainty
Tropical Storm Noel: GFS ENS
78
Tropical Storm Noel: GFS ENS
79
GFS ANALYSIS
DIFFERENCE
GFS 5-day FORECAST
CPS provides a
method to verify
model cyclone
structure
forecasts
80
Closing thoughts
• They have been shown to be very helpful for:
– Understanding the current analysis of a cyclone
– Understanding the subtlety among many models
– Quantifying the timing of
• Extratropical to tropical transition
• Tropical to extratropical transition
• Genesis
– Used by NHC, CHC, JTWC, JMA [approx 100 citations]
• We must remember they cannot replace other tools
– If used, they should be in addition to other tools
– They are only as useful as the source data
• We always welcome new model output to the web page
Obrigado
Extra Slides
Three key subcomposites
• Fast [<=12hr] vs. Slow [>=48hr] Transitioning
• Post-ET Intensification (N=6) vs. Weakening (N=11)
• Post-ET Cold-Core (N=15) vs. Warm-Seclusion (N=6)
TB: Fast (left) vs. Slow (right) Transitioning
500mb
Height
&
Anom.
SST &
Anom.
TB: Post-ET Weakening (left) vs. Intensification (right)
500mb
Height
&
Anom.
SST &
Anom.
Strength
en (N=6)
TE: Post-ET Cold-core (left) vs. warm-seclusion (right)
320K
PV
320K
PV
Strength
en (N=6)