Tropical cyclone activity and forecasts - May 2003

1. Tropical cyclone activity

1.1 Tropical cyclones ending during the month

NWP NEP NAT NI SWI AUS TOTAL Total number ending during the month 3 1 0 1 1 0 6

 

Name Identifier Basin Start Date End Date 1-minute MSW 10-minute MSW Manou 28S SWI 03 May 10 May 65 80 - 01B NI 10 May 19 May 65 - - 03W NWP 17 May 20 May 35 - Chan-hom 04W NWP 19 May 27 May 115 85 Andres 01E NEP 20 May 26 May 40 - Linfa 05W NWP 25 May 31 May 60 55

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

Name Identifier Basin Start Date Nangka 06W NWP 31 May

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 which 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 39 levels. This is equivalent to a horizontal resolution of 93km x 62km at the equator.

2.1 Individual Tropical Cyclone Summaries

Name: MANOU Identifier: 28S
Basin: SOUTH-WEST INDIAN
Start date: 03.05.2003 End date: 10.05.2003
1-minute MSW: 65 knots 10-minute MSW: 80 knots

Table of Mean Error Statistics   T+0 T+24 T+48 T+72 T+96 T+120 Possibly Verified 14 12 10 8 6 4 Detection Rate (%) 100 100 100 100 83 75 Mean DX 3 20 17 -85 -410 -908 Mean DY 46 98 14 -153 -263 -389 Mean AT 24 34 41 10 13 -349 Mean CT 32 64 -4 -140 -443 -821 Track skill (%) **** -26 26 66 **** **** Mean DPE 77 147 261 236 548 995 2001-2 DPE * 45 186 300 388 487 624 Intensity skill (%) **** 67 40 0 20 -100

A plot of the storm's observed track and forecasts.

Forecast errors were low up to T+72, but high thereafter. A few early forecasts had a right of track bias and took the storm across Madagascar, which explains the large longer range forecast errors. Other forecasts correctly predicted the stalling of the storm near to the coast of Madagascar.

 

Name: - Identifier: 01B
Basin: NORTH INDIAN (Bay of Bengal)
Start date: 10.05.2003 End date: 19.05.2003
1-minute MSW: 65 knots 10-minute MSW: -

Table of Mean Error Statistics   T+0 T+24 T+48 T+72 T+96 T+120 Possibly Verified 17 15 13 11 9 7 Detection Rate (%) 100 100 100 100 89 86 Mean DX -16 -27 -101 -27 -88 -260 Mean DY 12 154 242 265 386 202 Mean AT -35 -42 -25 -99 72 -61 Mean CT 15 -84 -250 -281 -355 99 Track skill (%) **** -17 5 11 **** **** Mean DPE 89 255 353 490 656 523 2001-2 DPE * 81 208 381 636 **** **** Intensity skill (%) **** -7 23 9 0 -33

A plot of the storm's observed track and forecasts.

01B exhibited a very erratic track which forecasts did not predict particularly successfully. At the time that the storm stalled in mid Bay of Bengal forecasts had predicted continued movement north-eastwards towards the coast. The storm did eventually accelerate towards the coast and forecast errors were not far from last season's average.

 

Name: - Identifier: 03W
Basin: NORTH-WEST PACIFIC
Start date: 17.05.2003 End date: 20.05.2003
1-minute MSW: 35 knots 10-minute MSW: -

Table of Mean Error Statistics   T+0 T+24 T+48 T+72 T+96 T+120 Possibly Verified 1 0 0 0 0 0 Detection Rate (%) 100 **** **** **** **** **** Mean DX -110 **** **** **** **** **** Mean DY -56 **** **** **** **** **** Mean AT -110 **** **** **** **** **** Mean CT -50 **** **** **** **** **** Mean DPE 122 **** **** **** **** **** 2002 DPE * 43 141 234 355 455 561

* DPE for all storms in the 2002 season in the NWP basin

A plot of the storm's observed track and forecasts.

03W only briefly reached tropical storm strength and no forecasts were verified.

 

Name: CHAN-HOM Identifier: 04W
Basin: NORTH-WEST PACIFIC
Start date: 19.05.2003 End date: 27.05.2003
1-minute MSW: 115 knots 10-minute MSW: 85 knots

Table of Mean Error Statistics   T+0 T+24 T+48 T+72 T+96 T+120 Possibly Verified 16 14 12 10 80 6 Detection Rate (%) 100 100 100 100 100 100 Mean DX -12 -100 -186 -347 -730 -1115 Mean DY -6 -3 11 -191 -706 -1088 Mean AT -13 -43 -56 -325 -978 -1538 Mean CT -2 -63 -187 -265 -195 -118 Track skill (%) **** 13 34 28 **** **** Mean DPE 48 138 246 543 1027 1572 2001-2 DPE * 43 141 234 355 455 561 Intensity skill (%) **** 57 33 0 75 33

* DPE for all storms in the 2002 season in the NWP basin

A plot of the storm's observed track and forecasts.

Forecasts for Chan-hom had a slight left-of-track bias and also failed to predict the acceleration of the storm later in its life. This resulted in very large track forecast errors at T+72 and beyond. However, the model showed skill over CLIPER up to T+72.

 

Name: ANDRES Identifier: 01E
Basin: NORTH-EAST PACIFIC
Start date: 20.05.2003 End date: 26.05.2003
1-minute MSW: 40 knots 10-minute MSW: -

Table of Mean Error Statistics   T+0 T+24 T+48 T+72 T+96 T+120 Possibly Verified 10 8 6 4 2 0 Detection Rate (%) 100 100 100 100 50 **** Mean DX 71 174 377 546 653 **** Mean DY 4 13 55 167 245 **** Mean AT -68 -171 -357 -496 -587 **** Mean CT 17 46 133 284 383 **** Track skill (%) **** -2 23 30 **** **** Mean DPE 72 186 387 575 702 **** 2001-2 DPE * 39 131 211 264 312 466 Intensity skill (%) **** 100 67 100 100 ****

* DPE for all storms in the 2002 season in the NEP basin

A plot of the storm's observed track and forecasts.

Andres was a typical straight-running north-east Pacific storm. Forecasts were slow and had a right-of-track bias, which resulted in track errors above last season's average.

 

Name: LINFA Identifier: 05W
Basin: NORTH-WEST PACIFIC
Start date: 25.05.2003 End date: 31.05.2003
1-minute MSW: 60 knots 10-minute MSW: 55 knots

Table of Mean Error Statistics   T+0 T+24 T+48 T+72 T+96 T+120 Possibly Verified 11 9 7 5 3 1 Detection Rate (%) 100 100 100 100 100 100 Mean DX -20 -63 -196 -453 -976 -826 Mean DY -2 -59 -253 -518 -990 -1090 Mean AT -16 -90 -337 -647 -1287 -1348 Mean CT -9 11 -7 48 -488 -209 Track skill (%) **** 39 15 25 **** **** Mean DPE 73 189 363 693 1390 1365 2001-2 DPE * 43 141 234 355 455 561 Intensity skill (%) **** 56 14 60 -100 -100

* DPE for all storms in the 2002 season in the NWP basin

A plot of the storm's observed track and forecasts.

Like Typhoon Chan-hom earlier in the month, Linfa accelerated north-eastwards later in its life and the model forecasts did not predict this well. Hence, there was a huge slow bias and forecast errors were very high beyond T+48. However, track forecasts did show skill against CLIPER.

 

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:- 6
Number of storms used in statistics below:- 6

Table of Mean Error Statistics   T+0 T+24 T+48 T+72 T+96 T+120 Possibly Verified 69 58 48 38 28 18 Detection Rate (%) 100 100 100 100 89 89 Mean DX -1 -13 -52 -119 -435 -737 Mean DY 10 52 41 -57 -264 -473 Mean AT -20 -52 -106 -249 -466 -749 Mean CT 10 -16 -100 -144 -308 -175 Track skill (%) **** 5 19 30 **** **** Mean DPE 72 185 321 486 843 1058 2001-2 DPE * 43 139 231 333 440 584 Intensity skill (%) **** 48 33 21 20 -25

* DPE for all Northern Hemisphere storms in 2002 season

A plot of the observed tracks of this month's storms in the North-West Pacific basin.

A plot of the observed tracks of this month's storms in the North-East Pacific basin.

A plot of the observed tracks of this month's storms in the North Indian basin.

A plot of the observed tracks of this month's storms in the South-West Indian basin.

Forecasts this month had a marked slow bias and negative cross-track bias (left-of-track in the northern hemisphere). Hence, track forecast errors were above last season's average. The model showed some skill over CLIPER. The overall intensity tendency skill score was 29%.