Title: DIAGNOSIS AND PREDICTION OF TROPICAL CYCLONE BEHAVIOUR IN THE AUSTRALIAN REGION AND THE SOUTH ATLANT
1DIAGNOSIS AND PREDICTION OF TROPICAL CYCLONE
BEHAVIOUR IN THE AUSTRALIAN REGION AND THE
SOUTH ATLANTIC SOME COMPARISONS BETWEEN THE
TROPICAL TRANSITION OF CATARINA (SA) AND DONALD
(AR)
NOEL DAVIDSON, ALEX PEZZA, LUKE GARDE AND
GRAHAM MILLS Centre for
Australian Weather and Climate Research
The University
of Melbourne, Earth Sciences
2CONTENT
1. Subtropical Cyclones Catarina (S. Atlantic)
and Donald (Aust. Reg.) Large-scale Environment
(LSE) Internal Structure and Evolution Climatol
ogical Aspects 2. TC-LAPS Initialization and
Prediction Components Large-scale Environment,
Vortex Specification, High-resolution Analysis,
Initialization, High-resolution Prediction,
Verification 3. Initialization and Prediction for
Catarina (preliminary results) The Vortex
Specification Dilemma!! Initial Conditions
Vortex Structure Prediction 4. Summary 5.
Possible Future Work?
3Track forecast quite skilful Intensity forecast
quite skilful to 48 hours, and then apparently
doesnt get the intensification (or does it?).
VMAX fluctuates between 55 and 70 knots
throughout, even though CP lies between
995-1000 hPa !!!!!??? W-E XSECT of VCMP at 48
hours Intense, mesoscale, warm inner-core
embedded in a synoptic-scale, cold-cored trough
(Structure OK??) Q1 Do we know the structure of
the inner-core at any time? (when,
depth, RMW?) Q2 Is the Dvorak Method applicable
here? Q3 When and how does the mesoscale
inner-core form? Q4 Why and how does the model
maintain this beautiful large-scale and
inner-core structure? (TT)
4Locations with Minimum Central Pressure
Points of Origin
Dare and Davidson, 2004 Tropical Cyclone
Characteristics in the Australian Region.
MWR. Forecast issues Track, Genesis, (gtgtgtgtgt
TT) Intensification/Decay, ET, Landfall
Final Decay Locations
Data for Initialization and Validation nearly
NON-EXISTENT!! Cannot validate structure and
quantify forecast skill much beyond track. So,
validate over the US..
5SUB-TROPICAL CYCLONES (MILLS)
- A cold-cored low, developing from the upper
troposphere - Can move westwards large distances
- Circularly symmetric at low levels
- Weak circulations near centre
- Strong winds on poleward side at large radii
- Surface low equatorward of sub-tropical ridge
- (Simpson, 1952)
- Often termed a Kona Low if it occurs off the
Hawaiian coast.
6 Subtropical Cyclone Donald
28/2/2001 2330UTC GMS 5 IR
ex-TC Abigail pre-Donald
TC Paula
High Latitude Influences Large Amplitude, Long
Waves at Upper Levels
7The DUCK TC Donald Not Named March 2001
8Donald produced large rain totals at Landfall
Catarina produced much less rainfall??
9Catarina and Donald Similarities Large-Scale
Environment
- Blocking
- Split Jet
- STJ to N
- Moisture from NE
CATARINA
DONALD
10Similarities Structure Phase Diagrams for
Catarina and Donald (Hart, 2003)
Medium-to-Deep Cold-Core transitioning to
SHALLOW, SYMMETRIC WARM CORE
11Similarities MOTION?
Tracks start (UTC) finish times
(UTC) ERA40 02/03/01 (1800) 08/03/01
(1800) ?t 6hrs NCEP 02/03/01 (1200)
08/03/01 (1800) ?t 6hrs NCEP2 03/03/01 (1200)
08/03/01 (1800) ?t 6hrs Meso LAPS 03/03/01
(2300) 08/03/01 (1100) ?t 12hrs LAPS
(Simulation Results) 04/03/01 (0000) 08/03/01
(0000) ?t 6hrs
12Similarities Structure Change
?
Donald Internal Structure Change Lagrangian
(System-following) Diagnostics from ERA40
Time-Height Series 00UTC 4 March 00UTC 9
March, 6-hourly Vorticity, Omega, Theta-E over a
300 km radius circle.
?
?e
Moisture Advection, Surface Fluxes Budget?
13?
Similarities Structure Change
Catarina Internal Structure Change Lagrangian
(System-following) Diagnostics from GASP
Analyses Time-Height Series 00UTC 23 March
00UTC 28 March , 6-hourly Vorticity, Omega,
Theta-E over a 300 km radius circle.
?
?e
gtHypothesis Inhibition, BL Moistening, Release
Inhibition, Active Convection, RI
14gt Hypothesis Initially, a weak low-level
cyclonic circulation is present. Period of
Inhibition to Convection by LSE occurs With
sustained Moisture Advection Surface Fluxes,
PBL moistens Period of Inhibition
passes Active, Deep Convection develops in a
much moister PBL Convection occurs
in a low wind shear, cyclonic environment
Intensification ensues
15SOUTH ATLANTIC INDEX (1979-2004)
Some Historical Perspectives
South Atlantic Index (SAI) normalized lagged
difference (12 days) between high latitude
blocking index and minimum midlatitude shear.
(Pezza et al 2008, IJC, under review).
TT
- Usually positive
- Slight positive trend
- High peak March 2004
Pezza and Simmonds, 2008 Annals of the NY Academy
of Science ANYAS, in press
16TRENDS OVER THE CATARINA BOX (ERA40)
Pezza et al., 2008 IJOC, under review
17TC-LAPS - Operational from 1999 Davidson and
Weber, 2000, MWR
18OPERATIONAL TC-LAPS
19VORTEX SPECIFICATION
- Construction of Symmetric Vortex
- INPUT size (ROCI), intensity (Pc), past
motion (0, -6, -12 hour locations) - and LSE (Large Scale Environment) from DA
- Surface Pressure Distribution Analytical
form - (Fujita, Holland, Chan and Williams), merge
with LSE using ROCI. - Assume thermal structure at storm centre is
moist adiabat from surface (Tc) - Pc, Tc gt Zc (geopotential height at centre
of storm) - Interpolate between Zc and Zlse using an
empirical structure function and - Storm Size to obtain Z
- Compute winds from Z using Gradient Wind
Equation - gt Symmetric Vortex
- NO SECONDARY CIRCULATION
- LACKS BALANCE between MASS and WIND FIELDS
20(No Transcript)
21GMS Assimilation and Initialisation
GMS CTT Data 2002020111 400hPa initialised omega
Initialised analysis and 48 hour forecast of
950hPa windfield
22INITIALIZATION OUTCOMES
- After initialization, if all goes to plan, the
initial condition contains a vortex with - An accurate analysis of the large-scale
environment - A primary circulation which is consistent with
estimated characteristics of the TC (location,
past motion, intensity). - Model-balanced primary and secondary
circulations. - A vertical motion field consistent with satellite
imagery, - A structure consistent with its environment.
- Unfortunately, everything rarely goes to plan!!
- Additionally, the circulation may not fit all
conventional obs - So..
- .. Design a new Vortex, constructed from
estimates of CP, VMAX, RMW, R34, ROCI - OR
- .. Use 4DVAR to construct vortex from
standard observations.
23(No Transcript)
24Good OBS NETWORK sophisticated
High-Resolution Modelling System well-defined
LS Environment consistent, well-defined Vortex
Structure gt Some skill in Intensity
Forecasts!! BUT BUSTS TOO!!
25PREDICTION EXPERIMENTS FOR CATARINA
- 1. System Options
- NCEP, GASP Initial Conditions
- Vortex Specification or No Bogus (which Vortex
Structure?) - Nudging Initialization or Cold Start
- Analysed or Forecast BCS
- Satellite-based Convective Heating Profiles (not
yet) - 2. Experiments (so far)
- GASP Initial Conditions, Cold Start
- GASP Initial Condition, Nudging
- Vortex Specification, Cold Start
- Vortex Specification, Nudging
- Revised Vortex Specification, Nudging
- Application of Satellite-based Convective Heating
Profiles
26Catarina Vortex Structure The Vortex
Specification Dilemma!!
Catarina Structure W-E XSECT of VCMP 24, 25, 26,
27 March Horiz and Vert Structure VMAX at outer
radii? Cold-cored till 26,27?
27Sensitivity to Initial Conditions
VS, FBCS
NCEP REANAL ABCS
GASP, ABCS
Base time 00UTC, 20040325 The Vortex
Specification Dilemma!!
28GASP ICs. Forecasts from base times 00UTC, 21,
22, 23, 24, 25 26 March
21
23
22
26
24
25
29W-E CROSS SECTIONS OF V COMPONENT FOR CATARINA
(00UTC 25 March 2004)
Full Vortex Specification (VS)
Global Analysis
VS in Lower Troposphere
30Sensitivity to Initialization. (base time 00utc,
25 March)
Basetime 00UTC, 20040325 CS, NG, Full VS, Half
VS .MORE WORK
31SUMMARY OF MAIN POINTS
- Catarina and Donald had similar structures and
evolution. - Both developed warm-cored structures from
cold-cored systems in which a large-scale
midlatitude upper trough and subtropical jet
enveloped the low-level circulation. Catarina
underwent Extratropical to Tropical Transition
(TT), Donald underwent partial TT. - Diagnostics suggests that TT may be associated
with environmental descent, boundary layer
moistening (advection and surface fluxes), and
eventually, as the inhibition to ascent passes,
development of deep, active convection, in a low
wind shear, cyclonic environment. - Track prediction for Catarina, after the loop,
is quite encouraging, but intensity prediction is
so far from TC-LAPS, only fair. (although
intensity and structure estimates are uncertain?) - Prediction of the pre-hurricane stage and TT
should not be conducted with existing vortex
specification methods. The imposed TC structures
are too different from the pre-hurricane
structures!! (Physical Initialization,
Krishnamurti et al., or 4DVAR may be the
preferred pathway). - Some forecasts may benefit from a revised form of
Vortex Specification for Subtropical Cyclones
but is this the best way?. (It may improve
understanding).
32POSSIBLE FUTURE WORK?
- Do we understand how the LSE influences the
internal structure? - 1. Modulation of ascent and environmental
vorticity. - 2. Lorenz Energy Conversions suggest
dramatic changes occurred 22/23
March when Storm Track looped. - (Veiga, Pezza, Simmonds, Silva
Dias, GRL, 2008?) - How well do we predict these (rare) phenomena?
(Intercomparisons between basins) - What do we need to predict these phenomena?
- Need to know more about the internal structure of
subtropical cyclones and how the structure
evolves. (RMW, Vertical Structure) - Would a revised form of Vortex Specification be
useful for prediction of subtropical cyclones?
(May improve understanding?) - Assimilation of additional data at higher
resolution? - Application of imposed convective heating
profiles during initialization to improve
intensity, and possibly track prediction? - Physical Initialization, 4DVAR?
33(No Transcript)
34SOUTH ATLANTIC HURRICANE CATARINA
SST minus 1000 hPa temp. 20-28 March 2004
Pezza and Simmonds, 2005
Pezza and Simmonds, 2008 Annals of the NY Academy
of Science, ANYAS, in press
lt SST used in NWP Expts. OK??
3505/3/2001 2330UTC GMS 5 IR
ex-TC Abigail Donald