DIAGNOSIS 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
35
Embed
DIAGNOSIS AND PREDICTION OF TROPICAL CYCLONE BEHAVIOUR IN THE AUSTRALIAN REGION AND THE SOUTH ATLANTIC : SOME COMPARISONS BETWEEN THE TROPICAL TRANSITION.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
DIAGNOSIS 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
1. Subtropical Cyclones Catarina (S. Atlantic) and Donald (Aust. Reg.)
LAPS (Simulation Results) :04/03/01 (0000) – 08/03/01 (0000)Δt = 6hrs
Similarities: MOTION?
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
Similarities: Structure Change
Moisture Advection, Surface Fluxes – Budget?
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
Similarities: Structure Change
=>Hypothesis: Inhibition, BL Moistening, Release Inhibition, Active Convection, RI
=> 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
SOUTH ATLANTIC INDEX (1979-2004)
Pezza and Simmonds, 2008Annals of the NY Academy of Science
ANYAS, in press
•Usually positive
•Slight positive trend
•High peak March 2004
TT
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).
Some Historical Perspectives
TRENDS OVER THE CATARINA BOX (ERA40)
Pezza et al., 2008IJOC, under review
TC-LAPS - Operational from 1999 Davidson and Weber, 2000, MWR
TC-LAPS has five basic components – all are critical : 1. Data assimilation (Bourke, Steinle, Kepert, Tingwell, …) to
establish the storm’s large-scale environment (LSE) and outer structure. (0.3750X29L,~550S-550N, 700E-1800E)
2. Vortex specification (Weber and Davidson; Holland vortex) to
construct the inner core circulation and asymmetries consistent with the estimated size, intensity and past motion, and then to re-locate the circulation to its observed position.
3. High-resolution (HR) analysis, (Steinle, Mills, Davidson) with
appropriate observational errors, length scales and quality control tolerances, to merge the intense vortex into the LSE. (0.150X29L, 270Lat-Long)
4. Initialisation with diabatic, dynamical nudging, (Davidson and
Puri) to balance the vortex using the model’s dynamics, and to re-define the vertical motion field to be consistent with the satellite imagery.
5. High-resolution prediction (Puri et al., 1998) with the generalised
LAPS forecast model, which contains high order numerics and sophisticated physical parameterisations.
VERIFICATION OF TC-LAPS : Australian Region 2004-2005
Forecast model initialized and verified against AR Operational Track Data t = 00
hrs t = 24
hrs t = 48
hrs t = 72
hrs Mean Direct Position Error (kms)
14.7 148.9 245.1 361.6
Mean Abs Central Pressure Error (hPa)
6.1 13.0 17.6 21.6
Number of forecasts 62 56 43 31
0- 100
100-200
200-300
300-400
400-500
500-600
600-700
No of 48 hour errors in 100km bins
7
9
13
11
2
0
1
Quite competitive for track and intensity by any measure. Major Errors : At 48 hours , ~ 1 BUSTS (> 500km).
Related to (a) Initial Vortex Structure, and/or (b) Analysis and Forecasts of the Large Scale Environment (DA).
Systematic Errors : Westward and Poleward Bias . Underestimation of Intensity
(Central Pressure). Unusual Behaviour around Coast.
DA builds analysed rotational flows to construct LSE and outer structure of storm. VS constructs Synthetic Vortex and merges it with LSE and outer structure from A NI creates ‘mesoscale balance’ and local, “estimated” divergent flows, consistent with cloud imagery, while preserving rotational flows. (Insert empirical convective heating profiles in regions of satellite-observed deep convection. Model tries to remove inserted heat sources with adiabatic cooling by ascent => ascent develops where deep convective clouds are observed)
GMS Assimilation and Initialisation
GMS CTT Data 2002020111
400hPa initialised omega
Initialised analysis and 48 hour forecast of 950hPa windfield
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.
F. Application of Satellite-based Convective Heating Profiles
PREDICTION EXPERIMENTS FOR CATARINA
Catarina: 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?
Base time 00UTC, 20040325: The Vortex Specification Dilemma!!
Sensitivity to Initial Conditions
NCEP REANAL ABCS GASP, ABCS VS, FBCS
GASP ICs. Forecasts from base times 00UTC, 21, 22, 23, 24, 25 26 March
21 22 23
24 25 26
W-E CROSS SECTIONS OF V COMPONENT FOR CATARINA
(00UTC 25 March 2004)
Global Analysis Full Vortex Specification (VS) VS in Lower Troposphere
Basetime 00UTC, 20040325: CS, NG, Full VS, “Half” VS
…….MORE WORK
Sensitivity to Initialization. (base time 00utc, 25 March)
1. Catarina and Donald had similar structures and evolution.
2. 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.
3. 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.
4. 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?)
5. 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).
6. Some forecasts may benefit from a revised form of Vortex Specification for Subtropical Cyclones – but is this the best way?. (It may improve understanding).
SUMMARY OF MAIN POINTS
• 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?
POSSIBLE FUTURE WORK?
SOUTH ATLANTIC HURRICANE ‘CATARINA’
Pezza and Simmonds, 2008Annals of the NY Academy of Science,