1. The Dvorak Technique The Dvorak Technique Originally By: TSgt Johnson
(JTWC Satellite Operations)
Updated by Mr. Paul McCrone
(HQ AFWA/XOGM)

2. The Dvorak Technique Positioning Intensity Estimation Coding
Final T-number
Data-T
Pattern-T
Model-T
Current Intensity
Coding

3. Positioning Eye Exposed Low Level Circulation Center (LLCC)
Central Dense Overcast (CDO)
Embedded Center (EMB CTR)
Spiral Band Curvature (SBC)
Partially exposed LLCC
Cloud minimum wedge
Central Cold Cover (CCC)
Cirrus outflow
Circle method
Conservative feature
Animation
Extrapolation

4. Common Abbreviations Low Level Circulation Center LLCC
Low Level Cloud Lines LLCLS
Upper Level Circulation Center ULCC
Central Dense Overcast CDO
Embedded Center EMB CTR
Spiral Band Curvature SBC
Central Cold Cover CCC
Enhanced Infrared EIR
Convection CNVCTN

5. Eye Fixes Visual (VIS) Dark cloud free spot
Shadowy spot for cloud filled eye
Without LLCC showing within eye, fix on center of eye
Measure width
Evaluate shape

6. Eye Fixes Infrared (IR)
IR Warm spot
As a guideline look for 2 shades warmer on the BD curve (MUST use BD CURVE!)
Use warmest spot.
Eye boundary (eye wall)defined by the tightest temperature gradient
Measure width and shape independent of Visual.

7. Exposed LLCC Fixes Look for tightening cyclonic low clouds.
The center will be within the tightest circle of clouds.
Much easier to find with VIS imagery

8. CDO Fixes VIS only (The term “CDO” implies “VIS only” - no exceptions!
Look for low level cloud lines to extrapolate underneath convection
Look for overshooting tops - bias LLCC toward tallest tops!
Look for SBC pattern in texture of CDO

9. EMB CTR Look for a warm spot
Look toward the edge with the tightest temperature gradient
Don’t forget continuity with past positions

10. SBC Fixes VIS Draw Streamlines on the image.
Place each streamline so the curve lies as close as possible to the low level cloud lines (LLCLS) and convective bands.
Follow the streamlines to the center.

11. SBC Fixes VIS Draw Streamlines on the image.
Place each streamline so the curve lies as close as possible to the low level cloud lines (LLCLS) and convective bands.
Follow the streamlines to the center.

12. SBC Fixes IR Draw Streamlines on the image.
Place each streamline so the curve lies as close as possible to the low level cloud lines (LLCLS) and connective bands.
Follow the curves to the center
Same idea as VIS

13. Partially Exposed LLCC Fixes VIS
Defined as “Less than half of the LLCC exposed”
Includes cirrus obscuration
The center will be within the tightest circle of clouds.
Similar to the Fully exposed LLCC, just harder - less obvious.

14. Partially Exposed LLCC Fixes IR
Again…. “Less than half of the LLCC exposed”
Same idea as visual, except this can be even harder
LLCLS don’t show up well in GOES/GMS EIR (but they show up fairly well in DMSP Thermal imagery).

15. Cloud Minimum Wedge
The center will be halfway along a line drawn from the tip of the wedge straight across the comma head.

16. Cloud Minimum Wedge
The center will be halfway along a line drawn from the tip of the wedge straight across the comma head.
Dry Slot

17. Cloud Minimum Wedge
The center will be halfway along a line drawn from the tip of the wedge straight across the comma head.

18. Cloud Minimum Wedge
The center will be halfway along a line drawn from the tip of the wedge straight across the comma head.

19. Cloud Minimum Wedge
The center will be halfway along a line drawn from the tip of the wedge straight across the comma head.

20. Cloud Minimum Wedge
The center will be halfway along a line drawn from the tip of the wedge straight across the comma head.

21. CCC Fixes VIS Rare; to be used with VIS imagery ONLY
Typically, storm exhibits a milky cirrus top - few convective cells visible
Only use if there is no evidence of the CDO or curved lines visible through the cirrus.
Center placement based mostly on continuity.

22. CCC Fixes IR Very difficult to accurately locate center.
Center placement based mostly on continuity.

23. Cirrus Outflow Fixes Use only when low clouds are not visible.
Follow the cirrus outflow anticyclonically back to the center.
This locates the Upper Level Circulation Center (ULCC)
ONLY USE AS A LAST RESORT.

24. Circle Method Fixes To be used with weak or very broad systems.
Use all available curvature to find center.
Where most circles meet, indicates most likely LLCC.

25. Conservative Features Fixes
If you had good reason to believe the LLCC was located at a certain point in relationship with a persistent feature, the LLCC will remain in the same relative location for up to 12 hours (at the very furthest in time)

26. Animation Fixes Self explanatory
Use with all type of fixes when available.
Use animation alone only when still imagery gives no indication of LLCC.

27. Extrapolation Fixes
Extrapolate past few fixes to determine where current position should be located.
Limited by the accuracy of the past fixes
Don’t technically need a satellite image to do this
This is a mathematical “dead-reckoning” approach.
This method is not encouraged as an analysis method

28. Intensity Estimation Step 1 - Locate Cloud System Center (CSC)
Step 2 - Determine the Data Type (Data-T)
Step 3 - CCC
Step 4 - Past 24 hour trend
Step 5 - Model Expected T-number (MET)
Step 6 - Pattern T-number (PT)
Step 7 - T-number determination
Step 8 - Final-T
Step 9 - Current Intensity (CI)
Step 10 - Final-T/CI encoding

29. Step 2, Data-T Step 2A Curved Band Step 2B Shear Pattern
Step 2C Eye Pattern
Step 2D CDO
Step 2E Embedded Center

30. Step 2A, Curved Band Use a LOG10 spiral overlay.
The spiral should lie along the axis of the of the band, and roughly parallel the inside edge of the band.
Measure the arc length.

31. Step 2A, Curved Band Use a LOG10 spiral overlay.
The spiral should lie along the axis of the of the band, and roughly parallel the inside edge of the band.
Measure the arc length.

32. Step 2A, Curved Band Measuring the arc length:
Follow the convection, not cirrus blow-off
Easier to do with Visual than Enhanced IR.
You may have small breaks in convection and draw through

33. Step 2A, Curved Band Use a LOG10 spiral overlay.
The spiral should lie along the axis of the of the band, and roughly parallel the inside edge of the band.
Measure the arc length.
LOG10
Spiral

34. Step 2A, Curved Band Measuring the arc length: Can be very subjective
Inexperienced analysts tend to go too high (fooled by cirrus).
This storm is somewhere between 0.70 and 0.85.

35. Step 2A, Curved Band
Note: Southern
Hemisphere
Example

36. Step 2A, Curved Band
Note: Southern
Hemisphere
Example

37. Step 2A, Curved Band
Note: Southern
Hemisphere
Example

38. Step 2A, Curved Band Draw your banding line
from the outside , then in.

39. Step 2A, Curved Band 360 degrees around equals “1.00” wrap 0.70 0.80
0.60
0.50
Start from your
LLCC and
count “pie slices”
or log10 sectors.
They are counted in tenths.
360 degrees around equals “1.00” wrap
0.40
0.10
0.30
0.20

40. Step 2A, Curved Band A wrap of 0.80 would equal a Data T of T3.5 0.70
0.60
0.50
0.40
0.10
0.30
0.20

41. I could have added
an additional 0.05
for this portion of
wrap, giving a total
wrap of 0.85
- Judgement call.
0.70
0.80
0.60
0.50
0.40
0.10
0.30
0.20

42. Step 2B, Shear Pattern
Measure the distance from the LLCC to the nearest convection.
In IR, use the dark gray (DG) shade on BD Curve ONLY to identify convection.

43. Step 2B, Shear Pattern
Measure the distance from the LLCC to the nearest convection.
In EIR, use the dark gray (DG) on BD Curve ONLY shade to identify convection.
70nm = T1.5

44. Step 2C, Eye Pattern Terms CF Formula: CF = E# + Eye adj
E#: Eye Number
Eye adj: Eye adjustment
CF: Central Feature
BF: Banding Feature
CF Formula: CF = E# + Eye adj
Data-T Formula: DT = CF + BF

45. Step 2C (E# VIS) Determine if Eye is banding type or not.
Measure the embedded distance of the eye, or the average band width if eye is a banding type.
For small eyes, measure distance from center.
For large eyes (> 30 nm) measure from edge.
Apply Values to E# Table

46. CDO type eye

47. CDO type eye
CDO

48. CDO type eye CDO Measure narrowest distance (This is for a Large eye,
which defined as 30 nm or
more inclusive)

49. CDO type eye CDO Measure narrowest distance (This is for a SMALL eye,
which defined as 29 nm or
less)

50. Banding Eye

51. Banding Eye

52. Banding Eye

53. Banding Eye Use average distance around the banding eye
(Outline the band if you need to)

54. Banding Eye Use average distance around the banding eye
Don’t use the smallest or largest
Can be subjective

55. Step 2C (Eye adj VIS)
For poorly defined or ragged eyes subtract .5 if E# < 4.5, and subtract 1.0 if E# > 4.5.
Add .5 or 1.0 if the eye is well defined, circular, and embedded in a smooth, very dense canopy
Add .5 or whole 1.0, IF the MET > 6.0, & the CF is less than the MET.

56. Step 2C (BF VIS)
Dense, mostly overcast band that curves at least 1/4 distance around the central feature.
Rarely used if CF is greater than 4.5.
Compare to chart.

57. Step 2C (E# IR)
Find the shade of the coldest ring completely surrounding the eye. MUST USE BD CURVE!
Measure the distance from the edge of the coldest color band to the outside edge of the coldest surrounding ring.
For EIR, DO NOT USE center of eye ever (center is only used in VISUAL)!
Apply Values to E# Table

58. Step 2C (E# IR, Example)

59. Step 2C (E# IR, Example)

60. Step 2C (E# IR, Example) 6.0 In this Example, use White.
Cold Medium Gray is not wide enough
(18 nm!)
Eye # is
6.0
Confused?
43 nm
18 nm
9 nm

61. Step 2C (E# IR, Example)
Why is it that Cold Medium Gray (CMG) is not wide enough?
The color
band you
pick must
meet the
distance
criteria
in the table
(at the
bottom
left)!
43 nm
18 nm
9 nm

62. Step 2C (E# IR, Example) The color band you pick must meet the
43 nm of
distance on the white
band meets the 6.0
requirement
The color
band you
pick must
meet the
distance
criteria
in the table
(at the
bottom
left)!
43 nm
18 nm
9 nm

63. Abbreviations:
Each abbreviation on the EIR Eye adjustment chart comes from the
BD enhancement curve, which Dvorak developed and used for EIR analysis on Tropical Cyclones (TC).
Each color is directly correlated to a temperature range.
It is standard to use these shortened color terms rather that the temperature values themselves in TC Analysis.
WMG Warm Medium Gray
OW Off White
DG Dark Gray
MG Medium Gray
LG Light Gray
B Black
W White
CMG Cold Medium Gray
CDG Cold Dark Gray
Eye Temperature
Surrounding
Ring
Temperature

64. Temperatures Abbreviation Gray Shade Temperature Range ( C)
WMG Warm Medium Gray > +09
OW Off White to -30
DG Dark Gray to -41
MG Medium Gray to -53
LG Light Gray to -63
B Black to -69
W White to -75
CMG Cold Medium Gray to -80
CDG Cold Dark Gray < -81 o o o o o

65. Step 2C (Eye adj IR)
Determine the warmest shade in the eye and the shade of the coldest ring surrounding the eye.
Do not use for eyes 45nm diameter or larger.
Apply to chart.
For elongated eyes, subtract .5 if none subtracted by chart.
Eye Temperature
Surrounding
Ring
Temperature

66. Step 2C (E Adj IR, Example)
Here,
distance
doesn’t
matter.
Use the
Cold Dark
Gray (CDG)
for the
surrounding ring temp.
Use Black for the eye.
Of course,
the eye
adjust
diagram
only goes
to CMG,
so use
CMG
instead.
Cold Dark
Gray only
needs to
surround
for eye
adjustment.

67. Step 2C (E Adj IR, Example)
Eye Temp
CDG (Cold Dark Gray) only needs to surround for eye adjustment
(Distance doesn’t
matter).
Surround.
Temp
Ring
So, for a Black eye
and CMG surround
ring temperature,
the eye adjustment
is -0.5.

68. Step 2C (E Adj IR, Example)
So, the E# was 6.0,
based on a white
band, and the eye
adjustment was -0.5,
based on a Black eye
and CMG surround
ring temperature.

69. Step 2C (E Adj IR, Example)
Again,
E# = 6.0
Eye adj= -0.5
(don’t forget minus!)
Now use the formula
for the CENTRAL
FEATURE (CF):
CF Formula:
CF = E# + Eye adj
Here,
CF=6.0 +(-0.5)=5.5

70. Step 2C (BF IR)
Use only when the CF is 4 or more and the CF is lower than the MET without adding the BF.
Band must curve 1/4 distance around, be MG or colder, and have a DG or warmer wedge.
Measure Wedge.
Apply Chart.

71. Step 2C (E Adj IR, Example)
Remember:
Here,
CF=6.0 +(-0.5)=5.5
There is no banding
feature on this storm.
BF=0

72. Step 2C (BF IR, Example)
This storm
DOES have a
Banding
Feature…..

73. Step 2C (BF IR, Example)
This storm
DOES have a
Banding
Feature
Here:

74. Step 2C (BF IR, Example) This storm DOES have a Banding Feature Here:
(This would be
about BF=0.5)

75. Step 2C, Eye Pattern Run the following Formulas:
CF Formula: CF = E# + Eye adj
Data-T Formula: DT = CF + BF
When calculating with VIS, eyes with a diameter > 30nm limit DT to 6.0 for well defined eyes, and 5.0 for all others.

76. Step 2C (E Adj IR, Example)
Remember:
Here,
CF=5.5
BF=0
Data-T Formula:
DT = CF + BF
DT= =5.5

77. Step 2D, CDO Only used with VIS images.
Determine if CDO is well defined.
Measure CDO diameter and apply to chart.
Always Add any Banding Feature

78. Step 2E, EMB CTR
Only used with IR images and when Final-T from 12 hours previous was > 3.5.
Determine distance at which the center is embedded into the shades of the BD curve and apply to chart.
Add Banding Feature.

79. Step 3, CCC
When past Final-T < 3, use MET for 12 hours then hold same.
When past Final-T > 3.5, hold same.
Use as Final-T and go to Step 9

80. Step 4, 24-Hour Trend Compare current image to image 24 hours ago.
Determine if the cloud features in the current image look better defined, the same or worse.
If better, the trend is Developed (D)
If the same, the trend is Same (S)
If worse, the trend is Weakened (W)

81. Step 5, Model-T (MET)
For systems with a 24-Hour Trend of D or W, determine Pattern Evolution and apply appropriate number to Final-T from 24 hours ago.
- Slow (+ .5)
Normal (+ 1.0)
+ Rapid (+ 1.5)
ASSUMES you are routinely doing Dvorak intensity estimates - can’t do a “one-timer”!

82. Step 6, Pattern T - PT (IR) A B C
Select the pattern in the diagram that best matches your storm picture; 100% SUBJECTIVE
Pattern T number. A B C
Pattern
Type:

83. Step 6, PT (VIS)
Select the pattern in the diagram that best matches your storm picture; same as EIR PT
Pattern T number. A B C
Pattern
Type:

84. Final T Number 1. Use DT when it is clear-cut and representative.
2. Use PT only when DT is not clear-cut and representative, and PT is.
Often happens with Shear DT estimates (Shear DT often too high).
Remember that the Data T is often calculated, so it may be unrealistic.
3. Use MET only when PT and DT are not clear-cut and representative.
Thumb-rule: PT is NOT clear cut when you have difficulty distinguishing amongst three different PT pictograms along the horizontal.
Example: on Vis PT diagram, you honestly have difficulty telling the difference between an “A4.0”, “A5.0”, and “A6.0”

85. Final-T Number Constraints
Initial Classification must be 1.0 or 1.5 (period).
During the first 48 hours of development Final-T can not be lowered at night (but you can during day-light).
24 hours after initial 1.0, Final-T must be < 2.5.
Final-T limits:
< 3.5 =FT can change only 0.5 per 6 hours.
> 4.0 = FT can change only1.0 per 6 hours, 1.5 per 12 hours, 2.0 per 18 hours, or 2.5 per 24 hours.
Final-T must be from MET.

86. Current Intensity (CI)
Make the CI the same as Final-T during development, but hold higher during weakening.
Do not lower CI until at least 12 hours have past.
CI will not be more that 1.0 greater than Final-T.

87. Current Intensity (CI)
Why hold CI higher during weakening?
Storm surface vorticity (spin) is conserved for a period of time, even though the cloud features seem unimpressive.
CAREFUL! The storm will appear weak, but still maintain enough energy to have strong winds

88. Short Term Trend (STT) STT is the trend based on less than 18 hours.
Always use a STT when you have less than 18 hours of history.
In addition, you can use STT when the regular trend does not tell the whole story.
NOTE: THERE IS NO MODEL T number until the system is AT LEAST 18 hours old.

89. Coding DVORAK coding STT coding FT/CI/Trend/Period
Example- T3.0/4.0/W1.5 24HRS
STT coding
Trend/Period
Example - T2.0/2.0/STT: D0.5/15HRS

90. Coding
What about those situations where the LLCC cannot be determined, but Dvorak Constraints keep us from finalling the system?
Enter TX.X for Final T
Keep last CI until Dvorak Constraints allow you to eliminate the system
Should normally happen with very weak systems.
Example- TX.X/1.0/STT: S0.0/14HRS

91. Coding Exercise

92. Exercise Answers

93. Questions? . Contact: . Paul J. McCrone, GS-12, Chief Forecaster
HQ Air Force Weather Agency (AFWA)
Meteorological Satellite Applications
Office Code: XOGM
106 Peacekeeper Dr. STE 2N3
Offutt AFB, NE
and
Phones: (402)
DSN:
Fax: (402)

94. Step 2A, Curved Band
Note: Southern
Hemisphere
Example

95. CDO type eye CDO Measure narrowest distance (This is for a Large eye,
which defined as 30 nm or
more inclusive)

96. Step 2C (E# IR, Example) 6.0 In this Example, use White.
Cold Medium Gray is not wide enough
(18 nm!)
Eye # is
6.0
Confused?
43 nm
18 nm
9 nm