Exercise – Constructing a best track from multiple data sourcesExercise – Constructing a best

track from multiple data sources

NATIONAL HURRICANE CENTER

NATIONAL HURRICANE CENTER

JACK BEVENJACK BEVEN

WHERE AMERICA’S CLIMATE AND WEATHER SERVICES BEGINWHERE AMERICA’S CLIMATE AND WEATHER SERVICES BEGIN

ObjectivesObjectives

• To construct two best tracks (wind/pressure every 6 hours) for a tropical cyclone (from the Atlantic)

• The first uses only satellite data (subjective and objective Dvorak CI numbers/scatterometer) and surface obs

• The second includes a full array of Atlantic aircraft data• Vertical black lines on the graphs denote landfalls• Note that there are data sources that could be used but

were omitted (doppler radar/AMSU/others)

• To construct two best tracks (wind/pressure every 6 hours) for a tropical cyclone (from the Atlantic)

• The first uses only satellite data (subjective and objective Dvorak CI numbers/scatterometer) and surface obs

• The second includes a full array of Atlantic aircraft data• Vertical black lines on the graphs denote landfalls• Note that there are data sources that could be used but

were omitted (doppler radar/AMSU/others)

Notes on “conventional” dataNotes on “conventional” data

• Satellite intensity estimates marked TAFB and SAB are subjective Dvorak estimates from the NHC Tropical Analysis and Forecast Branch and the Satellite Analysis Branch.

• The Objective T-Numbers are from the CIMSS ADT.• All Dvorak estimates use the Atlantic Dvorak wind/

pressure calibration.• The scatterometer ob is from QuikScat.• Surface pressure obs plotted on the chart are not

always central pressures. Sometimes they are near-center pressures used to calibrate the pressure curve.

• Satellite intensity estimates marked TAFB and SAB are subjective Dvorak estimates from the NHC Tropical Analysis and Forecast Branch and the Satellite Analysis Branch.

• The Objective T-Numbers are from the CIMSS ADT.• All Dvorak estimates use the Atlantic Dvorak wind/

pressure calibration.• The scatterometer ob is from QuikScat.• Surface pressure obs plotted on the chart are not

always central pressures. Sometimes they are near-center pressures used to calibrate the pressure curve.

Dvorak Technique OutputDvorak Technique Output1-minute MSW CI

Number (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 35 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

Note: Other warning centers and basins use different pressures and wind averaging periodsNote: Other warning centers and basins use different pressures and wind averaging periods

Part 1 – No Aircraft Data

Pressure (hPa)

Wind (kt)

RECONNAISSANCE FLIGHT PATH

Aircraft “ALPHA” Pattern

GPS DropsondesGPS Dropsondes

GPS sondes are used in the eyewall, elsewhere in the storm, and on G-IV missionsGPS sondes are used in the eyewall, elsewhere in the storm, and on G-IV missions

0

2000

4000

6000

8000

10000

100 110 120 130 140 150 160 170

Heig

ht

(ft)

Wind Speed (kt)

Hurricane Mitch - EyewallGPS Dropsonde Wind Profile

2337 UTC 27 October 1998

* Maximum flight-level windfrom reconaissance aircraft

Hurricane Mitch - EyewallGPS Dropsonde Wind Profile

27 October 98

Hurricane Mitch - EyewallGPS Dropsonde Wind Profile

27 October 98

Winds increase downward from flight-level (10,000 ft) because the

hurricane is “warm-core”.

Friction decreases wind in the lowest 1500 ft of the eyewall.

Reducing Flight-Level Data to the Surface Based on GPS Dropsonde Data

WL150MBL

Wind averaging over the lowest layers of the sonde data

Stepped-Frequency Microwave Radiometer

• Relates microwave radiation from ocean to surface wind speed

• Can measure max surface winds in core of major hurricane

• Only provides data along line of flight• First data from C-130s in 2007; on

entire fleet in 2008

SFMR Issues

• Shoaling – breaking waves in areas of shallow water can artificially increase the SFMR retrieved wind speed

• Interaction of wind and wave field can introduce errors (~ 5 kt)

• Rain impacts not always properly accounted for (mainly < 50 kt).

• Calibration has been undergoing revision. Algorithms still under development, and forecaster don’t fully understand the error mechanisms.

Notes on aircraft central pressure data

Notes on aircraft central pressure data

• Aircraft central pressures are either measured from dropsondes in/near the center or extrapolated from measured conditions at the flight-level center.

• Dropsondes measure the surface pressure, but do not always hit the lowest pressure/lightest winds area at the surface.

• Despite these issues, aircraft-measured pressures have a good track record compared to central pressures from surface obs. The aircraft data is usually good to within 1-3 hPa/mb.

• Aircraft central pressures are either measured from dropsondes in/near the center or extrapolated from measured conditions at the flight-level center.

• Dropsondes measure the surface pressure, but do not always hit the lowest pressure/lightest winds area at the surface.

• Despite these issues, aircraft-measured pressures have a good track record compared to central pressures from surface obs. The aircraft data is usually good to within 1-3 hPa/mb.

Notes on aircraft wind dataNotes on aircraft wind data• Wind data has a lot of scatter, as the plots include

winds from all quadrants of the radius of maximum winds – not just the strongest quadrant.

• Aircraft surface winds are from the SFMR.• Flight-level winds have already had the appropriate

reduction to the surface included on the plots.• Flight-level>Surface winds relationships can change

depending on the organization of the cyclone.• The Dvorak Pressure>Wind data points use the Dvorak

wind-pressure curve to convert central pressures to winds. Use these with caution!

• The dropsonde surface winds are the unaveraged winds at the last report from the sonde. They may or may not be representative.

• Wind data has a lot of scatter, as the plots include winds from all quadrants of the radius of maximum winds – not just the strongest quadrant.

• Aircraft surface winds are from the SFMR.• Flight-level winds have already had the appropriate

reduction to the surface included on the plots.• Flight-level>Surface winds relationships can change

depending on the organization of the cyclone.• The Dvorak Pressure>Wind data points use the Dvorak

wind-pressure curve to convert central pressures to winds. Use these with caution!

• The dropsonde surface winds are the unaveraged winds at the last report from the sonde. They may or may not be representative.

Part 2 – With Aircraft Data

Pressure (hPa)

Wind (kt)

The “Solution” - Pressure

The “Solution” - Wind