Dvorak Technique (modified)

• A statistical method for estimating the intensity of tropical cyclones (TCs) from interpretation of satellite imagery
  • quantifies TC intensity on a 1-8 scale (at 0.5 intervals) called T-Numbers, which are used in a variety of ways
    
    

    CI Number	1-min MSW				NHC/CPHC/JTWC
    	kt	mph	km/hr	m/s	MSLP (ATL/EPAC)	MSLP (NW Pacific)
    1.0	25	29	46	13
    1.5	25	29	46	13
    2.0	30	25	56	15	1009 mb	1000 mb
    2.5	35	40	65	18	1005 mb	997 mb
    3.0	45	52	83	23	1000 mb	991 mb
    3.5	55	63	102	28	994 mb	984 mb
    4.0	65	75	120	33	987 mb	976 mb
    4.5	77	89	143	40	979 mb	966 mb
    5.0	90	104	167	46	970 mb	954 mb
    5.5	102	117	189	52	960 mb	941 mb
    6.0	115	132	213	59	948 mb	927 mb
    6.5	127	146	235	65	935 mb	914 mb
    7.0	140	161	259	72	921 mb	898 mb
    7.5	155	178	287	80	906 mb	879 mb
    8.0	170	196	315	87	890 mb	858 mb

    
    
  • final output of the technique is the Current Intensity number (CI)
    • CI number is driven by the Final-T Number (FT)
      • FT is driven by the Data-T number (DT), the Model Expected-T number (MET), and the Pattern-T number (PT or PAT)
        • DT is often created from other sub-numbers

  Step 1: Locate the Cloud System Center (CSC 雲系中心)

  • usually defined by an eye, a low level circulation center or by other cloud features
  • Locate the overall pattern center
    • Typical cloud pattern evolution
      • curved band pattern (彎曲雲帶型)
        
        
        
        1. the CSC is halfway between the end of the curved band (A) & the end of the associated dry slot or cloud minimum wedge (B)
        2. the CSC is the focal point of curved cloud lines
      • central dense overcast (CDO 中心密集雲層區) pattern
        • a dense solid-looking mass of clouds covering the CSC, often lying within the curve of the cyclone's curved cloud band
      • shear pattern (切變型)
        
        
               "+" denotes the expected center position
        
        
  • Look for small scale features
  • Compare center location with forecast
  • Compare center with previous pattern center
  • Make final location adjustments
  • Looking for lowest possible center in terms of altitude (Surface center if possible)
    
    
  • Step 1A: A T1 classification can be given when ...
    1. a convective cluster has persisted ≥12 hrs
    2. the cluster has a CSC defined ≤ 2.5°(~275km) latitude (緯度) wide which has persisted ≥6 hr
    3. associated convection is dark gray (DG, <-31°C) or colder on the Dvorak BD enhancement curve over an area >1.5°(~165km) diameter <2° from the center
    • an existing CSC that does not meet the criteria of Step 1A can be tracked as a system "too weak to classify"- a location without an intensity estimate

  Step 2: Analysis of the Cyclone Cloud Pattern - DT

  • Tool
    • measuring curved bands
      • Log 10° spiral
        
        
    • IR imagery
      • BD enhancement
        
        
    
    
  • Step 2A: Measuring Curved Bands
    • When available, VIS are preferable to their IR counterparts
    • fit the spiral parallel to the inner edge of the band (VIS 可見光衛星雲圖) or to the coldest tops in the band (IR 紅外線衛星雲圖)
      • the spiral should lie along the axis of the of the band
    • measure only the arc length of primary band of the cyclone - other bands don't count
    • Add 0.5 to the DT if the curved band is White (W) or colder
    • endpoints of bands can be rather subjective
    • the center of the log 10° spiral is usually not the center of the cyclone
    • nature does not always produce bands with 10° crossing angles
      
      
  • Step 2B: Measuring Shear Patterns
    • When available, VIS are preferable to their IR counterparts
    • measure the distance (in degrees of latitude) from the low level center to the nearest convection, the edge of the dense overcast (VIS) or to the edge of the DG shade (IR)
    • the edge of the convection can be rather subjective
    • Shear patterns tend to be rather unstable, as the convection often shows strong pulses or bursts. Therefore the DT is often considered not to be clear cut
      
      
    
    
  • Step 2C: Measuring Eye Patterns
    • Find the eye number (E-number)
      • intensity estimate for an eye pattern based on surrounding cloud top temperatures (IR) or embedded distance in a CDO (VIS)
      • If 24 hr old FT ≤2, go to step 2A or step 3
        • IR
          • Find the coldest color on the BD enhancement that completely surrounds the eye with a thickness greater than the specified width (closed ring surrounding the eye)
            
            

            Surrounding BD color		OW	DG	MG	LG	B	W	CMG
            		OW
            Narrowest Width	°	≥0.3	≥0.3	≥0.4	≥0.4	≥0.5	≥0.5	≥0.5
            	km	≥33	≥33	≥44	≥44	≥55	≥55	≥55
            Eye Number (E#)		4.0	4.5	4.5	5.0	5.5	6.0	6.5

          
          
        • VIS
          • Visible eye
            • Measure the distance ('embedded distance') from the center of the eye to the edge of the CDO
              • embedded distance are measured from
                • the center of the eye for small eyes
                • the edge of the eye for large eyes (≥30 nm/56 km in diameter on VIS)
          • Banding eye (雲卷風眼)
            • ≥1 cloud band fully surrounding the eye
            • E-number determined by the average width of the band surrounding the eye
            
            

            Eye in CDO - Embedded Distance	°	~0.25	~0.5	~0.75	~1	>1
            	km	~28	~55	~83	~111	>111
            Banding Eye - Avg. Width of Band Around Eye	°	0.25	0.75	1.25
            	km	28	83	138
            Eye Number (E#)		3.0	4.0	5.0	6.0	7.0

      
      
    • Perform an Eye Adjustment
      • intensity estimate for an eye pattern based on eye temperature (IR) or eye size and clarity (VIS)
        • IR
          • based on the color on the warmest BD enhancement color in the eye
            
            

            			Eye Temperature
            			WMG	OW	DG	MG	LG	B	W
            			WMG	OW
            Surr. BD Color	OW	OW	0	-0.5
            	DG		0	0	-0.5
            	MG		0	0	-0.5	-0.5
            	LG		+0.5	0	0	-0.5	-0.5
            	B		+1.0	+0.5	0	0	-0.5	-0.5
            	W		+1.0	+0.5	+0.5	0	0	-1.0	-1.0
            	CMG		+1.0	+0.5	+0.5	0	0	-0.5	-1.0

          
          
        • VIS
          • Poorly defined or ragged eyes
            • subtract 0.5 for E≤4.5
            • subtract 1 for E≥5
          • Large eyes (≥30mm/ 56km)
            • limit T-no to T6 for round well-defined eyes
            • limit T-no to T5 for large ragged eyes
          • for MET ≥6
            • 0.5 or 1 may be added to DT for well-defined eye in smooth CDO when DT
        
        
    • Create the Central Feature (CF) number from the E-number and Eye Adjustment
      • E-Number + Eye Adjustment = CF Number
    • Add applicable banding features (BF) to the CF number to get the DT number
      • CF + BF = DT
      • BF
        • intensity estimate based on the banding surrounding central features of a cyclone
        • can be used with the eye, CDO, & embedded center patterns
        • IR
          • used only
            1. when CF/DT without banding < MET
            2. for cloud patterns of CF ≥4
            3. Band must be MG or colder while dry slot must be DG or warmer
            
            
          
          
        • VIS
          • can add multiple bands when applicable up to a total of 2.5 BF numbers
            
            
          
          
    • there are significant differences between what is measured in the VIS and IR eye patterns as shown on the flow chart
      • IR Eye Pattern is the most objective of all Dvorak measurements, but it cannot produce a DT=8.0 without adding banding
      • beware large satellite zenith/viewing angles and not being able to see to the bottom of the eye
      • for small eyes (generally <10 nmi wide), the satellite may not be able to measure the warmest temperature at the bottom of the eye, resulting in an underestimate of the intensity in both subjective & objective Dvorak techniques
      • beware sucker holes
    
    
  • Step 2D: Measuring Central Dense Overcast (CDO) Patterns
    • uses VIS
    • when CDO<0.75°wide, go to step 2A or step 3
    • measures the size of the CDO
      • edge of CDO is often subjective
      • for an elliptical CDO, the CDO size is the average of the sizes of the long and short axes of the ellipse
    • all banding rules are applicable
      
      

      CDO edge is:		Irregular		Well-defined
      Diameter	°	1.0-1.5	>1.5	0.75	1.25	1.75	≥2.25
      	km	110-170	>170	85	140	190	≥250
      Central Feature Number (CF)		2.0	3.0	2.0	3.0	4.0	5.0

    
    
  • Step 2E: Measuring Embedded Center (嵌匿藏中心) Patterns
    • uses IR
    • when 12 hr old FT <3.5, go to step 2A or step 3
    • measures how far the CSC is embedded into specified colors on the BD curve
      • embedded center pattern can only be used when the 12 hr old FT≥3.5 as otherwise it can produce unrealistically high intensity estimates
      • embedded center pattern is the most uncertain of all Dvorak measurements - where the classifier puts the CSC makes a big difference in the intensity estimate
      • when available and appropriate, use of VIS CDO is preferable to use of IR embedded center
    • all banding rules are applicable
      
      

      Surrounding BD Color		OW	DG	MG	LG	B	W or colder
      		OW					W or colder
      Embedded distance	°	≥0.4	≥0.4	≥0.5	≥0.5	≥0.6	≥0.6
      	km	≥44	≥44	≥55	≥55	≥66	≥66
      Central Feature Number (CF)		3.5	4.0	4.0	4.5	5.0	5.0

    
    
  • Step 2F: Central Cold Cover (CCC 中心冷雲蓋) Pattern
    • a large cold or dense overcast covering the CSC that lacks structure & obscures the cyclone center
    • when past T-no.≤T3, use as T-no. for coming 12 hours; then remains the same after 12 hours
    • when past T-no.>T3.5, use as final T-no.

  Step 3: Determine 24 hr Trend

  • Developing
    • Increased convection near CSC (larger or colder CDO)
    • Increased curved banding (primary band or bands around the CDO)
    • Eye forms, or becomes warmer, or more distinct
    • Exposed center closer to overcast
    • Increased curvature of low clouds near CSC
  • Weakening
    • Decreased convection near CSC (smaller or warmer CDO)
    • Decreased curved banding
    • Eye disappears, or becomes cooler, or less distinct
    • Exposed center further from overcast or covered center becomes exposed
    • Decreased curvature of low clouds near CSC
  • Steady
    • No noticeable 24 hr change
    • Both developing & weakening signs present (mixed signals)
    • CCC in a cyclone ≥T3.5 or CCC ≥12 hr in a weaker cyclone
    
    
  • the trend for the initial classification is always a normal D
  • ≥24 hrs of Dvorak classifications are needed to change the development trend
  • the first 18 hrs after the initial T1 are always a normal D
  • changes in the development rate for step 4 from D to D- or D+ (or W to W- or W+) need to use the technique rules, it cannot be changed arbitrarily

  Step 4: Determine Model Expected T-Number (MET)

  • the intensity estimate from the 24 hr old FT number & a determined intensity trend
    • Two consecutive previous Dvorak measurements of rapid or slow development/weakening are needed to establish rapid or slow 24 hr changes, or
      • look at the previous two FT values and compare them to the respective FT values from 24 hours prior
      • if the difference between both of these values >(<)1.0, then there is rapid (slow) development/weakening (add the +/- to the D or W)
      • this does not count the measurement making currently
    • one previous Dvorak measurement and signs of strong intensification or weakening (Step 8)
  • the MET for first classification is 1.0
    • for a Steady trend
      • MET = 24hr old FT
    • for a Developing or Weakening trend
      • normal strengthening or weakening
        • ±1 T-number/day
      • rapid changes
        • ±1.5 T-numbers/day
      • slow changes
        • ±0.5 T-numbers/day
        
        

        	Developing	Weakening
        Rapid	MET=24hr old FT +1.5	MET=24hr old FT -1.5
        Normal	MET=24hr old FT +1.0	MET=24hr old FT -1.0
        Slow	MET=24hr old FT +0.5	MET=24hr old FT -0.5

  Step 5: Determine Pattern T-Number (PT or PAT)

  • the intensity estimate from comparing the cyclone cloud pattern to predetermined patterns
  • if PT ≥1 T-numbers from the MET, recheck
  • IR (a - curved band, b - CDO, c- shear)
    
    
    
  • VIS (c- shear)
    

  Step 6: Determine Final T-Number (FT) by using DT, MET & PT

  • When cloud features are clear-cut
    • an unambiguous cloud pattern measurement
      • eg. an infrared eye measurement is often considered clear cut
    • measurements using multiple cloud pattern types that give the same DT
    • use DT from Step 2
  • When DT is not clear & when PT is different from MET
    • ambiguous or hard to measure/interpret cloud pattern measurements
      • eg. shear pattern measurements are often not clear cut.
    • measurements using multiple cloud pattern types that give different DTs
    • use PT from Step 5
  • For all other cases
    • use MET from step 4
  • Beware constraints (Step 6A)
    
    
  • Step 6A: Apply FT Constraints
    • initial classification must be T1 or T1.5
    • during first 24 hours of development, FT cannot be lowered at night
    • 24 hr after initial T1, FT must be ≤2.5
    • modified FT limits
    • FT must = MET ± 1

  Step 7: Determine Current Intensity Number (CI)

  • during the initial development & some stages of re-development, CI=FT
  • For weakening or redeveloping systems, hold the CI to the highest FT during the preceding 12 hr, but never >1.0 above the current FT
  • CI never < FT
    
    
  • Reasons
    • Strong systems
      • sometimes don't intensify as quickly as the cloud pattern suggests
    • Weak systems
      • sometimes lose all convection during the diurnal minimum
      • cloud patterns sometimes look unrealistically strong
    • Weakening systems
      • the decay of winds & pressures usually somewhat lags behind that of the cloud pattern

  Step 8: Forecast 24-hr Intensity (FI)

  • based on the current CI & satellite-observed signals in the cyclone cloud pattern & the environment
    
    

    Rules	Signs
    Rule A - Strong Unfavourable Signs in Cloud Pattern	persistent convective warming >12 hr CCC persisting >3 hr signs of shear or pattern elongation
    Rule B - Strong Unfavorable Signs in Environment	cyclone about to move into stratocumulus clouds cyclone about to move onto land signs of shear
    Rule C - Strong Favorable Signs in Cloud Pattern	2 successive observations of rapid development (24 hr change) 1 observation of rapid development & either a cold comma cloud pattern or multiple outflow channels forecast: If FT≤5.5, forecast rapid development (1.5 T-Numbers in 24 hr)
    Rule D - Weakened Cyclone Leaving Unfavorable Environment	cyclone leaving conditions of Rule B forecast rapid development to prior maximum intensity, followed by normal development
    Rule E - Cyclone Leaving Environment Where Development was Slowed	cyclone leaving conditions of Rule B forecast previous rate of development
    Rule F - Developing Cyclone Leaving Unfavorable Environment	cyclone leaving conditions of Rule B forecast increase of 1 T-Number per day in rate of development
    Rule G - Cyclone Peaking	northward moving cyclones expected to peak 4 days after first T1 westward moving cyclones expected to peak 6 days after first T1 all other cyclones expected to peak 5 days after first T1 forecast no change in intensity this rule is based mostly on climatology
    Rule L - Large Eyes	for eyes ≥30 nm/56 km wide with FT ≥6.0 limit FI to 6.0
    Rule P - Persistence	use when no strong signals are present forecast trend from past 24 hr to continue

  
  
  

  Reference Dvorak Technique