[an error occurred while processing this directive]

BOM Training Documentation

DVORAK TROPICAL CYCLONE INTENSITY ANALYSIS

1. INTRODUCTION
2. A MODEL OF TROPICAL CYCLONE DEVELOPMENT
3. DVORAK ANALYSIS - LOCATING THE SYSTEM CENTRE
Step 1: Locate the Tropical Cyclone Centre in Satellite Imagery
(i) Locate the overall pattern centre
(ii) Examine the cloud pattern for small scale features
(iii) Compare centre locations with forecast positions
(iv) Compare current centre location to previous
(v) Make final centre location adjustments
Examples of common centre locations in tropical cyclones
4. DVORAK ANALYSIS - ESTIMATING TROPICAL CYCLONE INTENSITY
Step 1a: Initial Development
Step 2: Determine the Pattern Type of the Cloud System
Step 2a: Curved Band Pattern
Step 2b: Shear Pattern
Step 2c: Eye Pattern
Step 2d: Central Dense Overcast Pattern
Step 2e: Embedded Centre Pattern
Step 3: Central Cold Cover Pattern
Step 4: Past 24 Hour Trend
Step 5: Model Expected T-number
Step 6: Pattern T-number
Step 7: Determining the Final T-number
Step 8: Rules Constraining the T-number
Step 9: The Current Intensity Number
  1. INTRODUCTION

    Satellite imagery provides the most universal source of data for locating and estimating the intensity of tropical cyclones. The Dvorak technique is a method of evaluating satellite signatures of tropical systems to quantitatively estimate their intensity. The technique takes into consideration the past trends in development in order to estimate the current and near-future intensity. Indications of continued development and/or weakening can be identified in the cloud features. Using these features in combination with the expected model development of the system and a series of rules, an intensity analysis and forecast can be made.

  2. A MODEL OF TROPICAL CYCLONE DEVELOPMENT

    This model describes the daily evolution of the cloud pattern associated with the typical development of a tropical cyclone. The descriptions include both cloud feature measurements and illustrations of the cloud patterns. Each day of evolution of the developing cloud pattern is defined as an increase in intensity of one Tropical (T) number.

    The primary pattern type described by the model is the curved band pattern illustrated in Figure 1(a). The typical curved band pattern evolution begins as the band takes shape on the first day of development. As the system intensifies the band forms or coils around the cloud system centre at the daily rate of one T-number per day. At the T2.5 stage of development, the curved band is shown to wrap half the distance around the storm centre (Day 2 of Figure 1(a)). The T4 stage of development is reached once the curved band has coiled completely around the centre. Further system intensification is marked by the appearance of an eye or increasing cold deep convection about the system centre. The model also allows for rapid and slow rates of intensity change when the pattern evolution appears faster or slower than typical. Rapid system intensification results in T-numbers increasing by 1.5 per day, while slow intensification typically sees rates of increase of about 0.5 T-numbers per day.

    The simple curved band patterns are only part of the model. The model also includes common distortions to the curved band pattern that result from a variety of atmospheric factors. Two additional pattern types are shown in Figures 1(b) and 1(c). The Central Dense Overcast (CDO) pattern (Figure 1(b)) is used when the central features of the pattern are covered by dense overcast clouds. The shear type patterns (Figure 1(c)) are used when vertical wind shear distorts the curved band pattern in the manner shown in the drawings. The modelled surface centre positions in Figure 1 are indicated by '+'.

    Maximum 10 minute average surface wind speeds associated with tropical systems of various T- numbers are as follows:-

    T-number Maximum 10 minute wind speeds
    T2 20 knots
    T2.5 30 knots
    T3.0 35 knots (tropical cyclone intensity Australian region)
    T3.5 50 knots (category 2 tropical cyclone Australian region)
    T4.5 65 knots (category 3 tropical cyclone Australian region)
    T5.5 90 knots (category 4 tropical cyclone Australian region
    T6.5 115 knots (category 5 tropical cyclone Australian region)


  3. DVORAK ANALYSIS - LOCATING THE SYSTEM CENTRE

  4. Step 1: Locate the tropical cyclone centre in satellite imagery.

    (i) Locate the overall pattern centre.

    The overall pattern centre is located to ensure that the focus is placed on the systems primary centre of vorticity. This is done by using either cloud lines or band curvature, a modelled centre placement or a combination of these techniques.

    Locating the cyclone centre using the band curvature technique is illustrated in Figure 2(a) and involves extrapolating or extending the curved lines or bands of the pattern to define the smallest common central area. Only the large scale lines and bands should used at this point. This technique for locating the cyclone centre is recommended for weak initial signatures and for all difficult cloud patterns.

    The modelled centre placement technique for locating the cyclone centre is shown in Figure 2(b). Draw a curved band axis through the centre of the most dense or coldest clouds of the pattern (shown as a dashed line in Figure 2(b)). Then draw a straight line from the deepest incursion of the cloud minimum wedge at "B" to the clockwise extremity of the curved band axis at "A". The overall cloud pattern centre is placed at the mid-point of this line. The length of the line AB is directly related to the confidence in the analysis. Note that the plus symbols in Figure 2(b) lie at the same point relative to the curved band axis in both patterns even though the pattern on the right shows high clouds building over the low level centre. This technique can be used with any pattern displaying a cloud minimum wedge, even those deformed by vertical wind shear.

    (ii) Examine the Cloud Pattern for Small Scale Features

    Small scale features of the satellite imagery can assist in locating the system centre. Such small scale features include the following:

    • An eye or indications of an eye. The Cloud System Centre (CSC) is located at the centre of the eye or at the centre of curvature of a partial eye wall when one of these features is observed. An eye may be identified by a warm spot in dense, cold (<9o C) overcast, or a relatively warm spot in a dense, cold (<9o C) overcast that has a cooler band curving at least halfway around it with a diameter of curvature of 1.5o latitude or less.
    • Low level cloud line curvature, cloud minimum areas (holes), or an area of merging cloud lines.
    • Middle or upper level cloud features such as line or band curvature, cold cloud areas, protruding cumulonimbus tops or holes in cloud cover. For large CDOs, the centre may be defined by an arc of overshooting cloud tops or an isolated cluster of convective tops.

    (iii) Compare centre locations with forecast positions

    If there are two possible centres, the one that is more consistent with the past track is the preferred position.

    (iv) Compare current centre location to previous.

    At times it is difficult to locate the system centre within the cloud pattern whereas, six hours prior for example, the system centre within the overall pattern may have been more obvious. By achieving a best fit of cloud features that have persisted over the past six hours the analyst can often arrive at a good estimate of the current centre position. Features such as the curved band axis and the cold canopy boundaries are often found to be persistent.

    (v) Make final centre adjustments

    At times satellite imagery grid mapping may be inaccurate which will result in erroneous cyclone centre positions. Check grid mapping by identifying obvious land features such as coastlines and ensuring that they coincide with maps accurately. Also, if a cyclone is a large distance from the satellite, parallax positions error may need to be taken into account.

    Examples of common centre locations in tropical cyclones

    Figure 3 illustrates six common cloud patterns associated with tropical cyclones and their typical centre locations.

  5. DVORAK ANALYSIS - ESTIMATING TROPICAL CYCLONE INTENSITY

  6. An estimate of a tropical cyclones intensity is determined mostly from objective measurement called the Data T or DT-number.

    Step 1a: Initial Development

    Disturbances showing signs of developing to tropical cyclone intensity are classified as T1.

    To be so classified, the disturbance must have the following properties:

    • The cloud system has persisted for more than12 hours.
    • A cloud system centre can be defined within an area having a diameter of 2.5o latitude or less, which has persisted for at least 6 hours.
    • The cloud system has an area of dense, cold (<-31o C) overcast greater than 1.5o latitude in diameter located less than 2o latitude from the centre. The overcast may also appear in cumulonimbus lines that curve around the centre.

    Step 2: Determining the Pattern Type of the Cloud System

    The manner in which the cloud system centre is defined determines which of the following pattern types are used for estimating the intensity of the system. When the cloud pattern being analysed does not resemble one of the following patterns, proceed to Step 3.

    The cloud pattern types are:

    • Curved Band Pattern
    • Shear Pattern
    • Eye Pattern
    • CDO Pattern (Visible (VIS) imagery) and Embedded Centre Pattern (Enhanced Infrared (EIR) imagery).

    A description of each pattern type follows.

    Below is a range of intensities (typical T-numbers) associated with the four primary pattern types.



    Step 2a: Curved Band Pattern

    This is the most common tropical cyclone cloud pattern observed in satellite imagery. The intensity estimate is derived by measuring the arc length of the curved band fitted to a 10o logarithmic spiral overlay and is essentially the same for both VIS and EIR imagery. The spiral overlay is fitted to the axis of the coldest overcast gray shade (most dense clouds) within the cloud band and should be roughly parallel to the overcast edge on the concave side of the band. The relationship between curved band lengths and the associated DT-number is shown in Figure 4. Figure 5 is an example of the curved band pattern in both VIS and EIR imagery. In this example the image has been remapped to Mercator projection to eliminate distortion due to the systems distance from the satellite (lower left frame in Figure 5). Such remapping is not necessary for systems in Papua New Guinea's are of responsibility as the satellite is positioned closely overhead. The dense overcast in the VIS image and the deep convection, dark grey and colder (colder than -31oC) in the EIR image both follow the 10o logarithmic spiral curve for more than 8 tenths, which corresponds to a DT-number of 3.5 as shown in Figure 4.

    Step 2b: Shear Pattern

    Shear patterns most commonly occur during the early development and weakening phases of the tropical cyclone life cycle and are only defined for systems less than hurricane intensity. They are identified by deep convection moving to one side the cyclone centre and developing a sharp edge. To analyse the pattern, the CSC must first be identified and then the distance between the CSC and the dense cold overcast measured. The relationship betwen these factors and the DT-number is shown in Figure 6. Typically the better the centre definition and the closer the centre is to the dense overcast, the higher the DT-number.

    In analysing the shear pattern, it is often difficult to distinguish the position of the low level centre in relation to the overcast. In such circumstances the following rules of thumb should be applied.

    • For tropical cyclone intensity (T2.5-3.5), the centre will be defined by parallel, circularly curved, low cloud lines with a diameter of 1.5o latitude or less near or under the edge of a dense, cold (<-31o C) overcast cloud mass.
    • For weaker systems (T1.5-2.5), the low cloud centre will either be poorly defined in spiral lines within 1.25o latitude of the cold overcast clouds, or circularly defined near a small, dense overcast (< 1.5o latitude diameter).

    In general, the shear pattern is more easily analysed in VIS imagery than it is in EIR imagery as low level cloud lines are more clearly defined.

    Step 2c: Eye Pattern

    The eye pattern is the most complex of the five pattern types. In determining whether a cloud system denotes the eye pattern, a history of pattern evolution must be known. The disturbance must have at least been classified a T2 or higher for 24 hours before an eye pattern analysis can be performed. If this criteria is satisfied the embedded distance of the eye in the cloud system must first be determined by measuring from the centre of the eye across the dense overcast to the nearest banding break or shadow. Using either the diagram in Figure 7 for VIS imagery or Figure 10 for EIR imagery, an Eye Number (E-number) can then be determined. The VIS imagery analysis method is based on the distance the centre is embedded underneath the cold dense overcast. The EIR method is based on the temperatures of the centre and surrounding deep convection. Note that the EIR analysis method requires the surrounding ring of convection to exceed minimum width requirements. Eye Adjustment Rules are then applied to the E-number to determine the Central Feature number (CF-number). These rules for VIS imagery are described in Figure 8 and take into account the definition and size of the eye. A maximum value of one number may be subtracted for a poorly defined, ragged eye. Similarly, for a system with a well defined eye, a maximum value of one number may be added. In addition to eye definition, eye adjustment may also depend on eye size and the intensity of the cyclone expected from extrapolation. The Eye Adjustment rules for EIR imagery are described in Figure 11. These rules are more objective than those for VIS imagery and are based on the system eye temperature and the temperature of the coldest ring of convection (defined by the coldest gray shade using the Dvorak enhancement) that completely surrounds the system centre. Minimum width requirements (as described in Figure 10 for determining the E-number) do not apply here. The resultant CF number (E-number + Eye Adjustment) may then be modified by adding a banding feature number (BF-number). The banding feature is a measure of the amount of wide or narrow banding that wraps around the central features of the cloud pattern. The more an overcast band feature curves around the central features, the greater the banding feature addition. See Figure 9 for banding feature patterns and their associated BF-number for VIS imagery analysis. Typically a BF-number is not added for an EIR eye pattern, except when the CF-number is less than the expected model intensity.

    E-numbers are also given for banding eyes. Banding type eyes are defined when the curved band wraps all the way around a cloud minimum at the cloud system centre. For this type of pattern, it is only the width of the band that determines the E-number as shown in Figure 7. The wider the band the greater the E-number.

    Since EIR imagery makes the analysis much simpler and more objective, EIR imagery should be used whenever possible for cyclones of hurricane intensity.

    Examples of Eye Pattern Analysis.

    Figure 12 is a VIS image of tropical cyclone Susan. A VIS eye pattern analysis is performed by measuring the embedded distance of the system centre within the dense overcast to the nearest banding break or shadow. The cloud pattern associated with tropical cyclone Susan shows the centre embedded a distance of 3/4o latitude. Using the table in Figure 7 results in an E-number of 5. Applying the Eye Adjustment Rules as shown in Figure 8, results in an eye adjustment value of zero and corresponding CF-number of 5. As the pattern of Susan in Figure 12 shows a narrow band that completely surrounds the eye, a BF-number of 1 (as determined from Figure 9) is added, resulting in a DT-number of 6.

    Figure 13 is an EIR image of cyclone Katrina. In this example the surrounding temperature is the black shade (B) which completely surrounds the eye and is at least 0.5o latitude in width. By referring to Figure 10, this corresponds to an E-number of 5.5 The colder white shade does not qualify as the ring of deep convection as it is too narrow on the eastern side. The eye adjustment rule for EIR analysis is then applied with a Medium Gray shade (MG) centre and white (W) surround. Using the table in Figure 11 the eye adjustment value is zero, resulting in a CF-number of 5.5.

    Step 2d: Central Dense Overcast (CDO) Pattern (VIS imagery):

    The CDO pattern is only used with VIS imagery. The CDO is a dense overcast mass of clouds that appears within the curve of the curved band axis and covers the cloud system centre. Modelled illustrations of the CDO patterns are shown in Figure 14. The size of the CDO and its boundary definition ie irregular or well-defined, provides the CF-number. As can be seen from Figure 14, the larger the CDO and the better the definition of its boundary, the higher the CF-number. The DT-number is derived by adding a BF-number to the CF-number according to the amount of banding observed around the CDO. The CDO pattern is used rarely since the curved band pattern analysis can often be used for this pattern type by drawing the curved band axis through the CDO.

    Step 2e: Embedded Centre Pattern (EIR imagery):

    The embedded centre pattern is used only with EIR imagery. This pattern type is used only when the tropical cyclone has had a previous history of a T3.5 or greater intensity and when the CSC clearly falls within the central overcast of the pattern.. The analysis of this pattern is similar to the eye pattern analysis except that no eye adjustment factor is added (see Figure 15). Firstly determine the coldest overcast in which the centre is embedded which meets the required minimum distance criteria. Using this data, read the value of the CF-number directly from the analysis diagram of Figure 15.

    Step 3: Central Cold Cover Analysis

    The Central Cold Cover (CCC) pattern consists of an approximately circular, cold or dense overcast covering the cyclone centre or comma head and obscuring the expected signs of pattern evolution. It differs from the CDO in that the CDO is a smaller cloud mass observed beneath the cold cirrus clouds that lies within the curve of the curved band pattern. The CCC pattern can occur at any stage of development and may last for several hours or for several days. Persistence of the CCC is typically associated with arrested development of the system. Care should be exercised under the following conditions:

    • Do not confuse a CCC pattern with a very cold comma pattern, usually indicated by a very cold comma tail and head, with smooth texture and an indication of a wedge; Curved cirrus lines or boundaries usually appear around the cold pattern and not around the CCC pattern.
    • Do not assume weakening in a CCC pattern when the comma tail begins to decrease in size, as is commonly observed; also the CCC often warms as the eye of the T4 pattern begins to develop.

    The following rules apply to CCC patterns:

    • When the past T-number is less than or equal to 3, maintain the model trend of development for 12 hours, then hold at the same intensity.
    • When the past T-number is greater than or equal to 3.5 hold the T-number the same.

    Step 4: Past 24 Hour Trend

    Cloud features in the current image are compared with those in the 24 hour old image to determine changes in intensity to the system over this time. Comparisons are made over 24 hours, even when images at shorter intervals are available, to avoid the strong diurnal effects often observed in tropical cloud patterns. From this comparison we decide whether the disturbance has developed (D), weakened (W) or remained the same (S) for the past 24 hours. Development is generally associated with increased organisation and better defined central features. The following are typical signs of development:

    • In the curved band pattern, the curved band coils further around the CSC.
    • In the CDO pattern, the CDO becomes larger or an increase in banding features is noted.
    • In the shear pattern, the CSC becomes more tightly defined in curved cloud lines or appears closer to the dense overcast.
    • In the eye pattern, banding features increase, the eye is more distinct (warmer), less ragged, and becomes further embedded in the cold dense overcast.
    • No significant non-diurnal warming of the cloud system is noted.

    The cyclone has weakened (W) when its cloud pattern indicates a persistent trend opposite to those listed above for development. Watch in particular for patterns that become sheared or exhibit warming (lowering) of cloud tops that is not associated with the diurnal cycle.

    The cyclone has become steady state (S) when:

    • A central cold cover appears in a T3.5 or greater tropical cyclone or has persisted for more than 12 hours in a weaker cyclone.
    • The CSC's relationship to the cold clouds has not changed significantly.
    • There are conflicting indications of both development and weakening.

    Step 5: The Model Expected T-number (MET)

    The MET is determined by using the 24 hour old T-number, the past 24 hour trend of deepening, weakening or remaining steady, and the rate the intensity has been changing in the past. When the growth rate has not been established in the case of new developments or reversals in trend, a past rate of change of one T-number per day is assumed.

    Generally the following intensity trends are observed:

    Current Intensity (CI)/Yesterday's T-Number
    0 1 (1.5) 2 3 4 5 (5.5) 6 (6.5) 7 8
    24-Hour Forecast CI Number/Today's MET
    D 1 1/1.5 (2.0) 3 4 5 6 (6.0) 6.5 (6.5) 7 8
    S 0 1 (1.5) 2 3 4 5 (5.5) 6 (6.0) 6.5 7
    W 0 1 (1.0) 1 2 3 4 (4.5) 5 (5.5) 6 7

    Note that the MET is not an analysed intensity of the storm but only a first guess estimate. Rapid or slow rates of change are established when two consecutive analyses showing rapid or slow pattern evolution are observed at 6 hour or more intervals, or when one observation of strong intensification or weakening is made.

    Step 6: The Pattern T-number (PT)

    The pattern T-number (PT) is equal to the MET plus or minus 0.5. The PT-number differs from the MET only when the pattern is obviously stronger or weaker than the MET number implies. For instance, if the MET is 3.5, you compare your pattern to the PT 3.5 pattern (see Figures 16 and 17 for patterns for EIR and VIS imagery respectively). If the pattern looks more like PT 4, 0.5 is added to the MET. Similarly if the pattern resembles PT 2.5 or weaker, 0.5 is subtracted from the MET.

    Step 7: Rules for determining the T-number.

    The rules for selecting the final T-number from the DT-number, the PT-number and the MET are outlined in Figure 18. These rules simply imply that the more vague or conflicting the evidence of intensity, the more the estimate should be biased towards the MET.

    Step 8: Rules which constrain the final T-number under specified circumstances.

    The final T-number selected in Step 7 is subject to the constraints outlined in Figure 19. These constraints are essential to the reliability and consistency of the intensity estimate. There is a strong tendency for analysts to ignore these rules when patterns appear much stronger or weaker than the rules allow. But years of comparing satellite observations with the central pressure measurements of reconnaissance aircraft show that cloud pattern changes do race ahead of pressure changes in some instances and that developing cyclones in their initial stages can lose most of their clouds at night.

    Step 9: The Current Intensity Number

    Each intensity analysis results in a Current Intensity (CI) number as well as a final T-number. The CI-number relates directly to the intensity of the tropical cyclone and is determined from the T-number by the rules outlined in Figure 20. When system redevelopment is indicated, the CI-number is not lowered to the T-number but remains the same until the T-number rises to the existing CI-number.