The Re-analysis of Hurricane Andrew (1992)

1
The Re-analysis of Hurricane Andrew (1992)

• Chris Landsea
• NOAA/Hurricane Research Division
• Miami, Florida, USA
• NHC Best Track Change Committee
• 1 August 2002

Contributors Pete Black, Peter Dodge, Jason
Dunion, James Franklin, Brian Jarvinen, Tim
Olander, Mark Powell, Chris Velden Comments
William Bredemeyer, Steve Feuer, Paul Hebert, Sam
Houston, Charlie Neumann, Hugh Willoughby
2
The Re-analysis of Hurricane Andrew (1992)

• Why re-analyze the Atlantic hurricanes and
  Andrew?
• Summary of changes suggested
• Flight-level wind to surface extrapolation
• Feature tracking from the Miami radar
• Pressure-wind relationships
• Satellite intensity estimates
• Storm surge and SLOSH implications
• Estimates of intensity from damage surveys
• Discussion of uncertainties

3
(No Transcript)
4
Work of Jose Partagas Historical
Reconstruction from 1851-1910
5
(No Transcript)
6
WHAT IS THE INTENSITY OF A TROPICAL CYCLONE?

• Maximum sustained surface wind Maximum wind,
  averaged over 1 minute interval at an altitude of
  33 ft (10 m), associated with the circulation of
  the tropical cyclone at a given point in time.
• With very, very few exceptions, direct
  observations of the maximum sustained surface
  wind in a tropical cyclone are not available.

7
HOW DO WE ESTIMATE INTENSITY?

• Satellite imagery using the Dvorak technique.

• Aircraft reconnaissance flight-level winds

• GPS dropwindsondes

8
(No Transcript)
9
Revised Best Track
10
Best Track Winds
Best Track Pressures
11
Revised HURDAT File
12
Revised Landfall Data
13
(No Transcript)
14
In the early 1990s, reduction factors used by
NHC ranged from 75-90 of the flight-level wind.
Powell and Black (1990) concluded 63-73 for
700mb to surface reduction factor, but had few
eyewall high-wind cases.
15
EYEWALL SCHEMATIC
EYEWALL
EYE
AIRCRAFT TRACK
DROPSONDE TRAJECTORY
80
70
100
90
10000 ft
1-2 miles
16
AVERAGE OF 357 GPS DROPSONDE PROFILES IN THE
HURRICANE EYEWALL. ON AVERAGE, THE SURFACE WIND
IS 90 OF THE WIND AT 700 MB AIRCRAFT
RECONNAISSANCE LEVELS (75-80 AT LOWER ALTITUDES).
Franklin et al. (2002)
17
EYEWALL STRUCTURE CAN VARY SIGNIFICANTLY FROM
STORM TO STORM, OR EVEN DURING DIFFERENT STAGES
OF THE LIFE CYCLE OF A SINGLE STORM.
Franklin et al. (2002)
18
Implications for Andrew
IMPLICATIONS FOR ANDREW Aircraft measured wind of
162 kt at 10,000 ft at 410 am, 1 hr prior to
landfall. If Andrew were occurring today, NHC
would estimate a surface sustained wind of 145
kt (165 mph), based on taking 90 of the recon
wind speed. Franklin, Pers. Comm.
Franklin et al. (2001)
19
Step 1 Flight Level to a Mean Boundary Layer (0
to 500m)
MBL to 700mb Ratio
Distance as a Function of RMW
Dunion, Landsea, Houston (2002) Dunion and
Powell (2002)
20
Step 2 Mean Boundary Layer (0 to 500m) to the
Surface
Dunion, Landsea, Houston (2002) Dunion and
Powell (2002)
21
New HWIND Analysis For Hurricane Andrew 150
kt 93 of flight level
22
(No Transcript)
23
Some New Hurricane Andrew Data Radar Feature
Tracking
24
(No Transcript)
25
Comparison of Flight Level Data (in Red) To
Radar Feature Data (in Green)
26
Pressure- Wind Relationships Where Does Andrew
Fall?
Brown and Franklin (2002)
27
Large versus Small Hurricanes Implication for
Pressure-Wind Relationship
28
Hurricane Andrew - Satellite Dvorak Estimates 127
kt/935 mb
29
(No Transcript)
30
Objective Dvorak Technique
31
Hurricane Andrews Storm Surge And SLOSH Runs
32
Structural Damage Surveys of Hurricane Andrew
33
Two Main Structural Damage Swaths Naranja Lakes
and Cutler Ridge/Tamiami
Fujita (1992) Wakimoto and Black (1994)
34
Two Main Structural Damage Swaths Naranja Lakes
and Cutler Ridge/Tamiami
Fujita (1992) Wakimoto and Black (1994)
35
The Re-analysis of Hurricane Andrew (1992)

• Flight-level wind to surface extrapolation 145 kt
  (10)
• Feature tracking from the Miami radar 145 kt
  (15)
• Pressure-wind relationships 145 kt (20)
• Satellite intensity estimates 145 kt (20)
• Storm surge and SLOSH implications 145 kt (25)
• Structural damage survey estimates 135 kt (30)

36
CONCLUSIONS Andrews intensity at landfall
(and elsewhere) will never be known with
certainty. Andrews intensity at landfall is
VERY LIKELY to be in the range of 136 to 155 kt
(Category 5) for the maximum sustained surface
winds in South Florida. The single best
estimate of intensity at landfall is 145 kt.
It is quite UNLIKELY that Andrew was a 125 kt
(Category 4) as originally thought.
37
(No Transcript)
38
        Atlantic Hurricane Re-Analysis Project
             
http//www.aoml.noaa.gov/hrd/hurdat/index.html

• Documentation
• Data By Year and Storm
• Reference

Picture from "Florida's Hurricane History", by
Jay Barnes
39
(No Transcript)
40
GPS DROPWINDSONDE

• Developed in conjunction with the NOAA
  Gulfstream-IV jet aircraft. First systematic
  use for intensity was in 1998s Hurricane Bonnie.
• GPS dropsondes provide, for the first time,
  direct measurements of the winds at low levels in
  the hurricane eyewall.
• Dropsonde data reveal that the structure of the
  eyewall is very complex, and can vary
  tremendously from storm to storm.

41
(No Transcript)
42
(No Transcript)
43
Observational Platforms for Atlantic Hurricanes
Neumann et al. (1999)

44
Incorrect Intensity And Location At Landfall
45
Too Rapid During Last 6 Hours
46
Pressure-Wind Relationship
47
Atlantic Major Hurricanes
Bias-removed
Landsea (1993)
48
HOW CAN WE USE THE DROPSONDE DATA TO IMPROVE OUR
OPERATIONAL INTENSITY ESTIMATES?
DIRECT MEASURMENTS OF SURFACE WINDS
INTERPRETATION OF FLIGHT-LEVEL WINDS
49
(No Transcript)
50
(No Transcript)
51
But nobody lives at 10,000 ft. How can we use
flight-level data to estimate surface winds?
90 kt
95 kt
90 kt
105 kt
RECON FLIGHT-LEVEL WINDS HURRICANE GEORGES
9/20/98 20-23Z
52
ATLANTIC WIND/PRESSURE RELATIONSHIPS
Winds (knots) computed from regional P/W
relationships
1)For GLFMEX Wind(knots)10.627(1013-p)0.5640
Sample size 664 r0.991 2)For lt25N
Wind(knots)12.016(1013-p)0.5337 Sample size
1033 r0.994 3)For 25-35N Wind(knots)14.172(1
013-p)0.4778 Sample size 922 r0.996 4)For
35-45N Wind(knots)16.086(1013-p)0.4333
Sample size 492 r0.974 5)For Kraft
Wind(knots)14.000(1013-p)0.5000 Sample size
13 r ??
P(MB) GLFMEX lt25N 25-35N 35-45N
KRAFT P(MB) P(IN)
960 100 100 94 90
102 960 28.35
53
DROPSONDES WILL FUNCTION EVEN IN THE TURBULENT
CONDITIONS OF THE HURRICANE EYEWALL. 211 MPH
STRONGEST WIND EVER OBSERVED BY A DROPSONDE IN A
HURRICANE.
54
Dvorak Technique
Tropical cyclones have characteristic cloud
patterns that correspond to stages of development
and certain intensities.
55
(No Transcript)
56
(No Transcript)