1. Tropical cyclone activity
1.1 Tropical cyclones ending during the month
|
|
Basin name abbreviations:-
NWP : North-West Pacific (west of 180°E)
NEP : North-East Pacific (east of 180°E)
NAT : North Atlantic
NI : North Indian (west of Malay Peninsula)
SWI : South-West Indian (west of 90°E)
AUS : Australian (east of 90°E)
1-minute MSW: Maximum sustained wind (knots) averaged over 1
minute (JTWC Hawaii, NHC Miami, CPHC Hawaii)
10-minute MSW: Maximum sustained wind (knots) averaged over 10
minutes (JMA Japan, Southern Hemisphere RSMCs and TCWCs)
Note: Although 1-minute average winds should be stronger than 10-minute average, this may not be the case here since estimates are obtained from different warning centres.
1.2 Tropical cyclones still active at the end of the month
|
2. FORECAST ERROR STATISTICS
The Met Office global Numerical Weather Prediction model produces a 6-day forecast every 12 hours. The performance of the model in forecasting this month's tropical cyclones is assessed below. Mean error statistics for each storm that ended during the month are presented together with a final table of statistics for all storms ending this month. Verification is performed at 24 hour intervals up to forecast time T+120. Some mean error statistics for last year's storms are included for the purposes of a comparison. The present global model has a resolution of 0.83°x 0.55°x 38 levels. This is equivalent to a horizontal resolution of 93km x 62km at the equator.
Click here for details of the TC verification technique and an explanation of error statistics and their sign conventions.
2.1 Individual Tropical Cyclone Summaries
Name: FRANCES Identifier: 06L
Basin: NORTH ATLANTIC
Start date: 25.08.2004 End date: 07.09.2004
1-minute MSW: 125 knots 10-minute MSW: -
|
* DPE for all storms in the 2003 season in the NAT basin
A plot of the storm's 00Z observed track and forecasts.
A plot of the storm's 12Z observed track and forecasts.
Forecasts for Hurricane Frances were very good with all track errors below last season's average. Skill scores against CLIPER were positive. There was a slight fast and right-of-track bias. The fast bias was as a result of some longer range forecasts not predicting the slow down of Frances just before landfall over the USA. However, the location of landfall was very well predicted.
Name: SONGDA Identifier: 22W
Basin: NORTH-WEST PACIFIC
Start date: 27.08.2004 End date: 08.09.2004
1-minute MSW: 125 knots 10-minute MSW: 90 knots
|
* DPE for all storms in the 2003 season in the NWP basin
A plot of the storm's observed track and forecasts.
Forecast errors were greater than last season's average at T+48 and beyond. There was a slow and left-of-track bias in forecasts. This was a result of many forecasts failing to predict recurvature.
Name: GASTON Identifier: 07L
Basin: NORTH ATLANTIC
Start date: 27.08.2004 End date: 01.09.2004
1-minute MSW: 60 knots 10-minute MSW: -
Table of Mean Error Statistics
|
* DPE for all storms in the 2003 season in the NAT basin
A plot of the storm's observed track and forecasts.
Forecast errors for this storm were near to or below last season's average. Skill scores against CLIPER were very high.
Name: HOWARD Identifier: 11E
Basin: NORTH-EAST PACIFIC
Start date: 30.08.2004 End date: 05.09.2004
1-minute MSW: 120 knots 10-minute MSW: -
|
* DPE for all storms in the 2003 season in the NEP basin
A plot of the storm's observed track and forecasts.
Track forecast errors for this storm were well below last season's average. The model showed skill over CLIPER at T+48 and T+72.
Name: PHOEBE Identifier: 01S
Basin: AUSTRALIAN (SE Indian)
Start date: 02.09.2004 End date: 04.09.2004
1-minute MSW: 55 knots 10-minute MSW: 50 knots
|
* DPE for all storms in the 2003-4 season in the AUS basin
A plot of the storm's observed track and forecasts.
This first storm of the southern hemisphere season was short-lived and just one forecast was verified.
Name: IVAN Identifier: 09L
Basin: NORTH ATLANTIC
Start date: 02.09.2004 End date: 24.09.2004
1-minute MSW: 145 knots 10-minute MSW: -
|
* DPE for all storms in the 2003 season in the NAT basin
A plot of the storm's 00Z observed track and forecasts.
A plot of the storm's 12Z observed track and forecasts.
Track forecast errors at all lead times were below last season's average and skill scores against CLIPER were positive for Hurricane Ivan. There was a persistent right-of-track bias in forecasts throughout the hurricane's lifetime. At lead times of 96 hours and shorter just one forecast erroneously predicted landfall over Louisiana. All other forecasts correctly predicted landfall over Alabama where Ivan eventually came ashore. No forecasts were verified during the brief phase when the remnants of Ivan regenerated over the Gulf of Mexico.
Name: SARIKA Identifier: 23W
Basin: NORTH-WEST PACIFIC
Start date: 04.09.2004 End date: 07.09.2004
1-minute MSW: 60 knots 10-minute MSW: 55 knots
|
* DPE for all storms in the 2003 season in the NWP basin
A plot of the storm's observed track and forecasts.
Track forecast errors were below last season's average for this storm.
Name: ISIS Identifier: 12E
Basin: NORTH-EAST PACIFIC
Start date: 08.09.2004 End date: 17.09.2004
1-minute MSW: 65 knots 10-minute MSW: -
|
* DPE for all storms in the 2003 season in the NEP basin
A plot of the storm's observed track and forecasts.
Track forecast errors were near to or below last season's average. Skill scores were positive at T+24 and T+48. Early forecasts had a fast bias and failed to predict the slow down which occurred near the end of the hurricane's life.
Name: - Identifier: 10L
Basin: NORTH ATLANTIC
Start date: 09.09.2004 End date: 09.09.2004
1-minute MSW: 30 knots 10-minute MSW: -
There are no verification statistics or forecast plot for this tropical
cyclone since its 1-minute averaged winds did not reach tropical storm
strength.
A plot of the storm's observed track.
Name: JAVIER Identifier: 13E
Basin: NORTH-EAST PACIFIC
Start date: 10.09.2004 End date: 19.09.2004
1-minute MSW: 130 knots 10-minute MSW: -
|
* DPE for all storms in the 2003 season in the NEP basin
A plot of the storm's observed track and forecasts.
Track forecast errors for Hurricane Javier were well below last season's average, although skill scores were only just positive at T+48 and T+72. There was a slight right-of-track bias in forecasts.
Name: HAIMA Identifier: 24W
Basin: NORTH-WEST PACIFIC
Start date: 11.09.2004 End date: 13.09.2004
1-minute MSW: 30 knots 10-minute MSW: 40 knots
|
* DPE for all storms in the 2003 season in the NWP basin
A plot of the storm's observed track and forecasts.
Haima was a short-lived tropical storm and just one forecast was verified.
Name: JEANNE Identifier: 11L
Basin: NORTH ATLANTIC
Start date: 13.09.2004 End date: 27.09.2004
1-minute MSW: 100 knots 10-minute MSW: -
|
* DPE for all storms in the 2003 season in the NAT basin
A plot of the storm's 00Z observed track and forecasts.
A plot of the storm's 12Z observed track and forecasts.
Track forecast errors for Jeanne were below last season's average up to T+96, but T+120 errors were large. Skill scores against CLIPER were positive. Early forecasts whilst Jeanne was still a developing tropical storm had a right-of-track bias and so did not predict the track across the Dominican Republic and Haiti. After re-emerging over sea and developing into a hurricane forecasts initially did not pick up on the extent of the eastward component of motion. However, later forecasts predicted the slow anti-cyclonic loop and westward movement towards landfall in Florida very well overall.
Name: KARL Identifier: 12L
Basin: NORTH ATLANTIC
Start date: 16.09.2004 End date: 24.09.2004
1-minute MSW: 120 knots 10-minute MSW: -
|
* DPE for all storms in the 2003 season in the NAT basin
A plot of the storm's observed track and forecasts.
Track forecast errors were near to or below last season's average for this hurricane. Skill scores against CLIPER were high. There was a fast bias in longer range forecasts.
Name: MEARI Identifier: 25W
Basin: NORTH-WEST PACIFIC
Start date: 20.09.2004 End date: 30.09.2004
1-minute MSW: 120 knots 10-minute MSW: 90 knots
|
* DPE for all storms in the 2003 season in the NWP basin
A plot of the storm's observed track and forecasts.
As with Typhoon Songda, the model's early forecasts did not predict recurvature well resulting in forecast errors above last season's average at lead times of 72 hours and greater. However, later forecasts predicted recurvature well and skill scores against CLIPER were high.
2.2 Summary of all Tropical Cyclones
Mean error statistics for all tropical cyclones (of at least tropical
storm strength) which have ended this month are shown below.
Total number of depressions ending this month:- 14
Number of storms used in statistics below:- 13
|
* DPE for all Northern Hemisphere storms in 2003 season
Comments:
September was an active month for tropical cyclones as it usually the case and track forecast errors were below last season's average at all lead times. Skill scores against CLIPER were positive. Although individual tropical cyclones had track biases, the mean bias for all tropical cyclones was very small. The intensity tendency skill score this month was a fairly modest 16%.