usingobservaonsandmodelsto be2er&understandandpredict ...€¦ · robert rogers aoml program...

Robert Rogers AOML Program Review 4-6 March 2013

Using observa-ons and models to be2er understand and predict

hurricanes

AOML Program Review 2

What is the role of convec0ve-‐scale processes in tropical cyclone intensity change?

•  TC intensity forecas/ng has shown less improvement than track forecas/ng •  Intensity change involves mul/scale processes

•  Observa/ons key component of a balanced approach toward advancing understanding and improving forecasts of TC intensity change •  HRD uniquely posi/oned to contribute to this effort through a combina/on of data

collec/on and analysis and numerical model experiments

•  IFEX: Mul/-‐year field campaign intended to improve TC intensity forecasts •  Partnership among NOAA (NHC, EMC, AOC) and other government, academic agencies (NASA GRIP, HS3; NSF PREDICT)

•  Goals: 1. collect observa/ons that span TC life cycle in a variety of environments for model ini/aliza/on and evalua/on

2. develop and refine measurement technologies that provide improved real-‐ /me monitoring of TC intensity, structure, and environment

3. improve understanding of physical processes important in intensity change for a TC at all stages of its life cycle

Overview

Rogers et al. (2013a)


Hurricane Isaac (2012)

Synergy of high resolu/on forecast and airborne observa/ons

P-‐3 and G-‐IV QC Doppler

observa-ons

Improving ini-al condi-on in storm

core region

Data assimila-on

Improving high resolu-on regional

forecast Wind intensity at 10 m

IFEX Goal 1: Observa-ons for model ini-aliza-on and evalua-on


“…THE 12Z HWRF RUN SHOWED CONSIDERABLY LESS INTENSIFICATION WITH KAREN COMPARED TO PREVIOUS RUNS AFTER ASSIMILATING DATA FROM THE FROM THE NOAA P-3 TAIL DOPPLER RADAR. THIS MARKS THE FIRST TIME DOPPLER RADAR DATA HAVE BEEN ASSIMILATIED INTO AN OPERATIONAL HURRICANE MODEL IN REAL TIME. ..”

NHC Forecast Discussion on October 4, 5 PM:


Posi/ve impact of Tail Doppler radar data on TS Karen intensity forecast

15

20

25

30

35

40

45

50

55

60

65

70

75

80

0 12 24 36 48 60 72 84 96 108 120

Intensity Forecast for Karen (2012) Valid 12 UTC 4 October 2013

10-m

pea

k w

ind

(kt)

Forecast hour

Best track HWRF without Doppler HWRF with Doppler


IFEX Goal 2: Techniques for real-‐-me TC monitoring

Tail Doppler radar from the high-‐al/tude NOAA G-‐IV aircraY

Flight track for the G-‐IV on 15 Sept 2013 in Hurricane Ingrid

Doppler-‐derived wind speed (shaded, m s-‐1) and vectors and dropsonde measurements at 1, 6, 12-‐km al/tude for Hurricane Ingrid (2013)

80 75

70

65

60

55

50

45 40

35 30

25

20 15

10

5

m/s

1 km 6 km 12 km

Doppler observa0ons will be transmi:ed in real 0me this season

25

24

23

22

98 97 96 95

21

94

25

24

23

22

98 97 96 95

21

94

25

24

23

22

98 97 96 95

21

94


IFEX Goal 3: Improved understanding of TC processes: Role of convec-ve-‐scale processes

What is the difference in the inner-‐core structure of intensifying and steady-‐state hurricanes?

Intensifying Steady-‐state

•  Analyses obtained from composites of mul/ple intensifying vs. steady-‐state hurricanes •  Convec/ve bursts defined as top 1% of ver/cal velocity distribu/on at 8 km al/tude (5.5 m/s) •  Intensifying cases have more bursts, more inside 2-‐km RMW compared with steady-‐state cases

Radial distribu/on of convec/ve bursts (%) and axisymmetric vor/city (shaded, x 10-‐4 s-‐1)

Rogers et al. (2013b)

x 10-‐4 s-‐1 x 10-‐4 s-‐1


Airborne Doppler observa/ons of the rapid intensifica/on (RI) of Hurricane Earl (2010) Storm-‐rela/ve wind speed (shaded, m s-‐1) at 2-‐km and

CB loca/ons (black dots) during RI

•  Most convec/ve bursts located inside 2-‐km RMW during this flight •  UpdraY core originates from PBL inside RMW, nearly ver/cal ascent •  Slope of updraY core departs significantly from angular momentum (M) surface •  Peak updraY inside local RMW throughout ascent, in locally high iner/al stability regime

20

19

18

64 63 62

55 50 45 40 35 30 25 20 15 10 5

m/s

Iner/al stability (shaded, 10-‐7 s-‐2) and tangen/al wind (contour, m s-‐1) on

downshear side during RI

1

2

3

4

5

6

7

8

9

-1 -0.5 -0.2 0.2 0.5 1 1.5 3 2 5 10

25 50 75 100 125 radius (km)

M surface

peak w

Ver/cal velocity (shaded, m s-‐1) and tangen/al wind (contour, m s-‐1) on

downshear side during RI

1

2

3

4

5

6

7

8

9

10

11

12

10

11

12

height (k

m)

25 50 75 100 125 radius (km)

-0.8 -0.4 0 0.4 0.8 1.2 1.6 2.4 2

M surface

peak w

m s-‐1 10-‐7 s-‐2

shear

RMW

50

100

150

200


What is the inner-‐core convec/ve structure of a rapidly-‐intensifying hurricane?


~0200

IFEX Goal 3: Improved understanding of TC processes: The Tropical Cyclone Diurnal Cycle

•  Manifests as a radially expanding diurnal pulse in the cloud field; •  Begins at sunset and reaches R=400-‐600 km by the following aYernoon;

Is this a fundamental TC process? What are its implica0ons on structure & intensity?


2005 Emily

GOES IR GOES IR 6-hr T Trend

2007 Felix

~02 LST

~02 LST

~08 LST

~08 LST

~14 LST

~14 LST

10-yr

TC Diurnal Clock

Hovmoller: Daily IR Temp Trends (10-yr

MH Composite)


How well does HWRF produce microphysics fields in deep convec/on?

radius (km)

Radar reflec/vity (shaded, dBZ) from ER-‐2 in Bonnie (1998)

2.8

4.2

5.6

7.0

1.4

0

2D-P

Iwc,

10-

5 kg/

kg

W, m

/s

10

5

0

-5

Flight-‐level ver/cal velocity (top, m/s) and ice concentra/on (bo_om red, x 10-‐5 kg/kg) from

midlevel NASA DC-‐8 aircrab

• Thompson scheme produces graupel at high (12 km) al/tudes, but may produce too much

• Ferrier does not produce any graupel at these heights, produces much less overall than Thompson

• Bonnie was an unusual storm; more research needed in a spectrum of cases

24

21

18

15

12

9

6

3

33

30

27 70

60

50

40

30

20

10

80

2

4

6

8

10

12

14

16

18 20

2

4

6

8

10

12

14

16

18 20

0 50 100 150 radius (km)

200 250 0 50 100 150 radius (km)

200 250

height of DC-8 height of DC-8

Thompson scheme Ferrier scheme

Axisymmetric graupel concentra/on (shaded, x 10-‐5 kg/kg) at 54 h for idealized HWRF runs using Thompson (leb) and opera/onal Ferrier (right) scheme

10-‐5 kg/kg 10-‐5 kg/kg

Black et al. (2003)


Evalua/on of ice microphysics in HWRF simula/ons


Summary • HRD is advancing the IFEX goals of improving TC intensity forecas/ng through a combina/on of observa/ons, modeling, and theory

•  assimila/on of airborne Doppler, new radar pladorms, research on convec/on and its role in TC intensity change

• HRD is uniquely posi/oned to combine these approaches

• Ongoing work will con/nue to develop and refine our observa/onal and modeling capabili/es, covering the spectrum of spa/al and temporal scales important in TC intensity change

•  new sampling strategies, model evalua/on

• These efforts advance NOAA’s mission of building a Weather-‐Ready Na/on

QUESTIONS?

AOML Program Review

Papers cited here Black, R. A., G. M. Heymsfield, and J. Hallett, 2003: Extra large particle images at 12 km in a

hurricane eyewall: Evidence of high-altitude supercooled water? Geophys. Res. Lett., 30, 2124, doi:10.1029/2003GL017864, 21.

Dunion, J.P., C.D. Thorncroft, and C.S. Velden, 2014: The tropical cyclone diurnal cycle of mature hurricanes. Mon. Wea. Rev. (in press)

Rogers, R.F., S.Aberson, A. Aksoy, B. Annane, M. Black, J. Cione, N. Dorst, J. Dunion, J. Gamache, S. Goldenberg, S. Gopalakrishnan, J. Kaplan, B. Klotz, S. Lorsolo, F. Marks, S. Murillo, M. Powell, P. Reasor, K. Sellwood, E. Uhlhorn, T. Vukicevic, J. Zhang, and X. Zhang, 2013a: NOAA'S Hurricane Intensity Forecasting Experiment: A Progress Report. Bull. Amer. Meteor. Soc., 94, 859–882.

Rogers, R.F., P. Reasor, and S. Lorsolo, 2013b: Airborne Doppler Observations of the Inner-core Structural Differences between Intensifying and Steady-State Tropical Cyclones. Mon. Wea. Rev., 141, 2970-2991.


Extra slides



Spa/o-‐temporal scales targeted by IFEX field experiments




Pre-‐ IFEX 1956-‐2004

IFEX 2005-‐2011

Pre-‐TD 4.3 9.9 TD 7.2 5.5 TS 26.8 37.1

Cat 1-‐2 31.6 24.8 Cat 3-‐5 30.0 22.7

Percentage (%) of on-‐sta/on aircraY flight hours




m/s

N

S

N S

m/s

height (k

m)

IFEX Goal 2: Techniques for real-‐-me TC monitoring

Real-‐/me analyses of TC inner-‐core structure from airborne radar

Doppler-‐derived wind speed (shaded, m s-‐1) and vectors and dropsonde measurements at 1-‐km al/tude for Hurricane Irene (2011)

South-‐north cross sec/on of Doppler-‐derived wind speed (shaded, m s-‐1) through

Hurricane Irene (2011)

•  Analyses are available within 1-‐2 h aYer aircraY lands

1

2

3

4

5

6

7

8

9

10

11

29

28

27

26

79 78 77 76

80 75 70 65 60

55 50 45 40

35 30 25

20 15

10 5

50

45

40

35

30

25

20

15

26.4 26.7 27 27.3 27.6 27.9 28.2 28.5 28.8 26.1

m/s


Exp. 1826 HWRF Exp. 1226 HWRF

IFEX Goal 3: Improved understanding of TC processes: Convec-ve-‐scale processes

•  AYer bifurca/on period, coverage of bursts for Exp. 1826 increases markedly •  Enhanced coverage occurs prior to significant increase in symmetric wind speed •  Transient, limited coverage of bursts for Exp. 1226

2-km axisymmetric wind maximum and coverage of convective bursts between r* = 0.5 and 1

HWRF simula/ons of the rapid intensifica/on of Hurricane Earl (2010) 12

UTC

8/2

9

18 U

TC 8

/29

00 U

TC 8

/30

06 U

TC 8

/30

12 U

TC 8

/30

18 U

TC8/

30

00 U

TC8/

31

12 U

TC 8

/29

18 U

TC 8

/29

00 U

TC 8

/30

06 U

TC 8

/30

12 U

TC 8

/30

18 U

TC8/

30

00 U

TC8/

31

66

22

20

68 Box 2: 1648 – 1733 UTC

Convec/ve burst module in Tropical Storm Gabrielle (2013)

Infrared satellite image and flight track showing 2nd of 3 box pa_erns flown around convec/ve burst

Convec-ve burst module: •  Collect Doppler and dropsonde data in vicinity of a convec/ve burst at high /me frequency (~30-‐45 minutes) •  Document structure and evolu/on of convec/ve-‐scale proper/es, e.g., sta/s/cs of reflec/vity, ver/cal velocity, vor/city, mass flux over convec/ve/mesoscale /me scales

•  Quan/fy impact of convec/ve-‐scale processes on parent system •  Evaluate and improve HWRF microphysics parameteriza/on

heig

ht (k

m)

ver-cal velocity (m s-‐1)

Contoured frequency by al/tude diagram (CFAD) of ver/cal velocity (shaded, %) for 2nd box pa_ern

40 30 15 5 2 1 0.5 0.1 0.01

1 2 3 4 5 6 7 8 9

10 11 12 13 14 15 16

%

-12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 14 16

heig

ht (k

m)

vor-city (x 10-‐3 s-‐1)

Contoured frequency by al/tude diagram (CFAD) of ver/cal vor/city (shaded, %) for 2nd box pa_ern

30 15 5 1 0.5

0.1 0.2

1 2 3 4 5 6 7 8 9

10 11 12 13 14 15 16

%

-4 -3 -2 0 1 2 3 4 -1



max

win

d (k

t)

HWRF 3-‐km simula/ons of the rapid intensifica/on of Hurricane Earl (2010)

Intensity traces for Hurricane Earl (2010)

•  Two HWRF runs: ini/alized at 12 UTC 26 August (Exp. 1226) and 18 UTC 26 August (Exp. 1826) •  Both runs capture early intensity evolu/on well •  Bifurca/on period at 00 UTC 8/30 – RI aborted in Exp. 1226 for 24 h •  More convec/ve bursts inside 2-‐km RMW in Exp. 1826 during bifurca/on period

00 U

TC 8

/27

00 U

TC 8

/28

00 U

TC 8

/29

00 U

TC 8

/30

00 U

TC 8

/31

00 U

TC 9

/1

140

120

100

80

60

40

20

0

Observations Exp. 1226 Exp. 1826

Wind speed at 2 km(shaded, m s-‐1) Streamlines at 2 km(black), 8 km(white) during bifurca/on period

Convec/ve burst loca/ons (red contours)

Exp. 1226 HWRF (no RI)

150

100

50 0

-50

-100

-150 150 100

50

0

-50 -100 -150 km

km

Exp. 1826 HWRF (RI)

150

100

50 0

-50

-100

-150 150 100

50

0

-50 -100 -150 km

50

45

40

35

30

25

20

15


Can high-‐resolu/on numerical models capture structural features of RI?

usingobservaonsandmodelsto be2er&understandandpredict ...€¦ · robert rogers aoml program...

Documents