Observed and forecast tracks: northern hemisphere 2009

All tropical cyclones in the northern hemisphere season in 2009 are listed below including details of their start and end dates and maximum sustained wind (MSW). Qualitative comments are included on the performance of the Met Office global model in predicting the track of each tropical cyclone.

Click on the name of the tropical cyclone to see a chart of the observed track and forecast tracks from the Met Office global model. Verification statistics on each tropical cyclone are available upon request from tropical_cyclones@metoffice.gov.uk. At the end of the season a verification report on the performance of the Met Office global model will be published.

North-West Pacific
Name & IDStart & end dates 1-/10-minute MSWComments on Met Office global model forecasts
Kujira (01W)02-07 May115/80 knotsTrack forecast errors were near to last season's average for this storm and the model showed skill over CLIPER. There was a slow bias in forecasts, although the direction of track was well predicted.
Chan-hom (02W)03-11 May85/75 knotsThe direction of motion of this storm was well predicted, but there was a large slow bias resulting in track errors larger than last season's average. However, the model still showed good skill over CLIPER.
Linfa (03W)17-22 June75/60 knotsForecasts for Linfa were very good. The direction and speed of movement were well pedicted and track errors were very low whilst skill scores were very high.
Nangka (04W)22-26 June45/40 knotsNangka took a similar track to Fengshen in 2008 and the model showed the same bias. Early forecast tracks tended to have a right-of-track bias and later tracks a slight slow bias. Hence, track errors were a little above last season's average.
Soudelor (05W)09-12 July35/35 knotsSoudelor was only briefly a tropical storm and just one 24-hour forecast was verified.
06W13-14 July30/30 knots06W did not reach tropical storm status and no forecasts were verified.
Molave (07W)15-19 July65/65 knotsForecasts for Molave were mixed with some showing a leftwards and fast bias and some very accurate. Track forecast errors were above last season's average at 24 and 48 hours, but below this average at 72 hours. Skill scores against CLIPER were high.
Goni (08W)02-08 August45/35 knotsTrack forecast errors were generally below last season's average and the model gave some indication of the southward motion back out over the ocean near Hainan Island.
Morakot (09W)03-10 August85/80 knotsTrack forecast errors were large for this storm, particularly at longer lead times. This is as a result of failing to capture the northward turn of Morakot once it had passed over Taiwan.
Etau (10W)08-13 August35/50 knotsThe recurvature of Etau south of Japan was well signalled by the model, although forecasts had a slow bias. Hence track errors were large at some lead times.
Vamco (11W)17-25 August115/90 knotsThere was a slight left-of-track bias in forecasts, but in general the motion of the storm was well predicted, with track forecast errors mostly below last season's average.
Krovanh (12W)28 August - 01 September60/60 knotsTrack forecast errors were above last season's average primarily due to a fast bias in early forecasts. However, later forecasts predicted the recurvature close to the coast of Japan reasonably well.
Dujuan (13W)03-09 September55/50 knotsAs with Krovanh, early forecasts were worse than later ones. Prediction of the rate of acceleration during extr-tropical transition were erratic resulting in track forecast error larger than last season's average, particularly at longer lead times. However, skill scores againsy CLIPER up to 72-hours were good.
Mujigae (14W)09-11 September30/35 knotsMujigae only briefly attained tropical storm status (according to JMA only) and no forecasts were verified.
Choi-wan (15W)12-20 September140/100 knotsAn early forecast failed to predict recurvature, but apart from that forecasts were good resulting in track errors below last season's average at all lead times except 144 hours and high skill scores against CLIPER.
Koppu (16W)13-15 September70/65 knotsKoppu was relatively short lived with just one 48-hour forecast verified. Track forecast errors were close to last season's average at 24 hours lead time.
Ketsana (17W)25-30 September90/75 knotsPredictions of the typhoon's westward motion and landfall over Vietnam were well predicted and track forecast errors were below last season's average.
18W27-30 September35/30 knotsjust one 12-hour forecast was verified for 18W.
Parma (19W)27 September - 14 October130/105 knotsApart from a couple of early forecasts, track predictions for Parma were good. In particular, the slow movement near the Philippines for several days was well predicted as was the eventual progression across the South China Sea. Track forecast errors were generally below last season's average and skill scores against CLIPER were high.
Melor (20W)29 September - 09 October145/110 knotsTrack forecast errors were below last season's average and skill scores high. Some forecasts prematurely recurved Melor and some had a slow bias. There was a slight right-of-track bias at the point of landfall over Japan, although the magnitude of the error was quite small.
Nepartak (21W)08-13 October55/45 knotsThe north-eastward acceleration of Nepartak was not well predicted on the whole resulting in track forecast errors above last season'a average.
Lupit (22W)14-27 October135/95 knotsShort lead time track errors were fairly good and the model did fairly well at predicting the first turn northwards followed by the return to a westward track. However, the model did not pick up on the eventual recurvature, instead predicting a slow west-southwestward drift. This resulted in some very large errors at longer lead times. Incidentally, the Met Office ensemble prediction system (MOGREPS) did a much better job at predicting the recurvature of Lupit than did the deterministic model.
Mirinae (23W)26 October - 02 November90/75 knotsThere was a slight right-of-track bias in forecasts for this typhoon, but generally the model did well. Track errors were below last season's average by a significant margin at longer lead times and skill scores against CLIPER were high.
24W02-03 November30/- knotsThis system was brief lived and did not reach tropical storm status.
25W07-09 November45/30 knotsThis system briefly reached tropical storm status according to JTWC, but no forecasts were verified.
Nida (26W)22 November-03 December160/115 knotsMet Office global model forecasts were predicting a westward turn for Nida when many other models were predicted rapid recurvature. Although, the westward turn did not happen immediately, neither did recurvature and Nida did eventually turn west prior to dissipation. Hence, track forecast errors were near to last season's average at short lead times and only significantly above average at 84 hours and beyond. Skill scores against CLIPER were very high. There was a left-of-track bias in results as would be expected given the premature forecast of the westward turn.
27W23-24 November30/- knotsThis system was brief lived and did not reach tropical storm status.
28W05 December35/- knots28W only briefly attained tropical storms status (according to JTWC) and no forecasts were verified.

North-West Pacific observed tracks in 2009
North-West Pacific tropical cyclone names

North-East Pacific
Name & IDStart & end dates 1-minute MSWComments on Met Office global model forecasts
01E18-20 June30 knots01E was a tropical depression and no forecasts were verified.
Andres (02E)21-24 June65 knotsThe model showed a right-of-track and slow bias for Andres resulting in track errors above last season's average. However, Andres was only above tropical storm strength for two days, so only a few forecasts were verified.
Blanca (03E)06-09 July45 knotsJust two 24-hour forecasts were verified for this storm and track errors were small.
Carlos (04E)10-16 July90 knotsThe westward movement of this hurricane was well predicted, but the model had a fast bias which resulted in track errors above last season's average. However, skill scores against CLIPER were positive.
Dolores (05E)15-17 July45 knotsOnly one 24-hour forecast was verified for this storm. The model had a left-of-track bias.
Enrique (07E)03-07 August50 knotsThere was a slow bias in forecasts resulting in 48-hour track errors above last season's average.
Felicia (08E)04-11 August125 knotsTrack forecast errors were mostly low and skill scores high for this hurricane. There was a slight left-of-track bias in forecasts.
09E09-13 August30 knots09E was a tropical depression and no forecasts were verified.
Guillermo (10E)12-20 August110 knotsTrack forecast errors were generally low and skill scores against CLIPER high for this hurricane.
Hilda (11E)22-28 August55 knotsTrack forecast errors were near to or below last season's avarega for this storm. Skill scores against CLIPER were high.
Ignacio (12E)24-27 August45 knotsTrack forecast errors were near to last season's average for this storm.
Jimena (13E)29 August - 04 September135 knotsForecasts for Jimena had a left-of-track bias resulting in increasingly large track forecast errors with lead time. However, skill scores against CLIPER were good.
Kevin (14E)29 August - 01 September45 knotsKevin was only briefly a tropical storm and just one forecast was verified.
Linda (15E)07-12 September70 knotsForecasts had a slow and left-of-track bias resulting in errors above last season's average.
Marty (16E)16-19 September40 knotsJust two 24-hour forecasts were verified for Marty. Track forecast errors were low.
Nora (17E)23-25 September50 knotsJust one 24-hour forecast was verified for Nora. The track error was low.
Olaf (18E)01-04 October40 knotsOlaf was fairly short lived and just one 24-hour forecast was verified.
Patricia (19E)11-14 October50 knotsForecast tracks were a little fast for this storm, but few forecasts were verified since it was fairly short-lived.
Rick (20E)15-21 October155 knotsThe north-eastwards turn of Rick was not picked up by the model. Instead it a continued north-westwards track. Hence, track forecast errors were generally high and skill scores relatively low.
Central North Pacific
Name & IDStart & end dates 1-minute MSWComments on Met Office global model forecasts
Lana (06E)30 July-03 August55 knotsTrack forecast errors were near to last season's average for this storm.
Maka (01C)11-18 August45 knotsMaka was a tropical storm briefly early in its life then after a period of being a tropical depression again strengthened to become a tropical storm. Only a few forecasts were verified.
02C29-30 August30 knots02C was a tropical depression and no forecasts were verified.
Neki (03C)19-27 October105 knotsEarly forecasts for Neki were poor since they failed to predict the northwards turn and slow down of the hurricane. Hence, track forecast errors were much above average particularly at longer lead times. However, the model still managed to show considerable skill against CLIPER.

North-East and Central North Pacific observed tracks in 2009
North-East Pacific tropical cyclone names
Central North Pacific tropical cyclone names

North Atlantic
Name & IDStart & end dates 1-minute MSWComments on Met Office global model forecasts
01L28-29 May30 knotsNo forecasts were verified since this system did not attain tropical storm strength.
Ana (02L)11-17 August35 knotsAlthough Ana existed for several days only a brief period of this was as a tropical storm, so just one 24-hour forecast was verified. The track forecast error was small.
Bill (03L)15-23 August115 knotsThere was a left-of-track bias in forecasts for Bill resulting in track forecast errors above last season's average at 72-hours and beyond. However, skill scores against CLIPER were high.
Claudette (04L)16-17 August45 knotsClaudette was only briefly a tropical storm and no forecasts at 24 hours or beyond were verified.
Danny (05L)26-29 August50 knotsThere was a left-of-track bias in forecasts, although track forecast errors were near to last season's average and the 24-hour skill score was high.
Erika (06L)01-04 September50 knotsErika was a short-lived storm. The few forecasts verified had a fast bias.
Fred (07L)07-12 September105 knotsThe formation of Fred from an easterly wave in the eastern Atlantic was very well signalled by the model. Track forecast errors were also low and the deceleration and dissipation were well predicted.
08L25-26 September30 knotsNo forecasts were verified for this tropical depression.
Grace (09L)05-06 October60 knotsAlthough Grace existed for some time as a non-tropical feature, it was only briefly a tropical storm and only one 24-hour forecast was verified.
Henri (10L)06-08 October45 knotsHenri was yet another Atlantic storm which succumbed to shear in 2009 and thus was fairly short-lived.
Ida (11L)04-10 November90 knotsIda was generally well predicted with track forecast errors below last season's average - particularly at longer lead times. Skill scores against CLIPER were high. There was a slight slow bias resulting in forecasts predicting final landfall over the USA a little later than occurred.

North Atlantic observed tracks in 2009
North Atlantic tropical cyclone names

North Indian
(forecast track charts for Global and South Asia models shown)
Name & IDStart & end dates 1-/10-minute MSWComments on Met Office global model forecasts
Bijli (01B)15-17 April50/40 knotsThe global model was too weak and too slow on the northward track of Bijli. Hence track errors were large. However, the higher resolution South Asia model (tracks appended below global model chart) gave a much better track and intensity prediction.
Aila (02B)24-25 May65/55 knotsThe global model (top chart) did a little better than the South Asia Model (bottom chart) with regard to landfall location. However, the latter model gave a better intensity prediction.
03B05 September35/30 knotsNo forecasts were verified for this short-lived storm.
Phyan (04A)07-11 November50/35 knotsTrack forecasts were better from the South Asia Model (bottom chart) than the global model (top chart). The former correctly predicted landfall, whilst the latter tended to keep the storm just offshore.
Ward (05B)11-14 December50/45 knotsGlobal model track forecast errors for Ward were larger than last season's average due to a propensity to move the storm westwards too quickly. South Asia Model (bottom chart) errors were slightly smaller.

North Indian observed tracks in 2009
  North Indian tropical cyclone names

Note on estimated wind speeds

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, IMD New Delhi)
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.

Last updated: 17 December 2013