Observed and forecast tracks: northern hemisphere 2012

All tropical cyclones in the northern hemisphere season in 2012 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 dates1-/10-minute MSWComments on Met Office global model forecasts
01W17-18 February25/30 knots01W did not attain tropical storm status and no forecasts were verified.
Pakhar (02W)29 March-01 April65/40 knotsTrack forecast errors were low for this storm, although landfall was not predicted.
Sanvu (03W)21-27 May80/60 knotsShort lead time forecasts had small errors, but at longer lead times there was a slow bias resulting in errors larger than last season's average.
Mawar (04W)31 May-06 June100/75 knotsTrack forecast errors were below last season's average and skill scores against CLIPER were high.
Guchol (05W)11-19 June130/100 knotsTrack forecast errors were lower than last season's average at some lead times, but higher at others. Skill scores were high. Recurvature was successfully predicted, but some of the track detail was in error.
Talim (06W)17-21 June50/50 knotsThe direction of motion was well predicted, but the predicted speed of motion was a little low.
Doksuri (07W)26-30 June50/40 knotsThe track and landfall of Guchol was generally well predicted.
Khanun (08W)15-19 July50/50 knotsTrack forecast errors were near to or below last season's average for this storm.
Vicente (09W)20-24 July115/80 knotsTrack errors were low and skill scores high for this storm.
Saola (10W)28 July-03 August90/70 knots
Track forecast errors were below last season's average and skill scores high for this typhoon.
Damrey (11W)28 July-03 August75/70 knotsTrack forecast errors were mostly low and skills cores high for this typhoon. A few longer lead times forecasts were larger with a slow bias.
Haikui (12W)02-09 August65/65 knots
Track forecast errors were low and skill scores high for this typhoon.
Kirogi (13W)04-10 August45/50 knots
Apart from the first forecast for this storm, track forecast errors were very low and skill scores very high.
Kai-tak (14W)12-18 August70/65 knots
The model had a marked right-of-track bias early on for this typhoon resulting in some larger than average track forecast errors and low skill scores.
Tembin (15W)19-30 August115/90 knotsTrack forecast errors were low and skill scores high for this typhoon. The first landfall on Taiwan and the subsequent reversal in direction were well predicted. The only poor aspect of forecasts was a few runs which merged Tembin with Bolaven instead of keeping them as separate entities.
Bolaven (16W)20-29 August125/100 knotsSome early forecasts had a fast bias and some forecasts showed an erroneous interaction with Tembin resulting in a left-of-track bias. Thus track forecast errors were mostly larger than average.
Sanba (17W)10-17 September150/110 knotsShort lead time track forecast errors were low, but longer range errors high due to a slow bias. Skill scores against CLIPER were high.
Jelawat (18W)20 September-01 October140/110 knotsTrack forecast errors were near to or above last season's average. there was a right-of-track bias. Skill scores against CLIPER were high.
Ewiniar (19W)24-29 September55/50 knotsEwiniar was exceptionally well forecast by the model. Skill scores were very high and track forecast errors low.
Maliksi (20W)30 September-04 October45/50 knots
Forecasts had a left-of-track bias resulting in high track forecast errors.
Gaemi (21W)01-06 October55/50 knotsThe reversal in direction of Gaemi was well predicted, but landfall predictions were slow resulting in some large longer lead time errors.
Prapiroon (22W)07-19 October100/90 knotsTrack forecast errors were below last season's average and skill scores very high. The temporary slow down and reversal in direction was well predicted from three days in advance.
Maria (23W)14-19 October55/50 knotsThe sharp right turn of Maria was poorly predicted resulting in some large track forecast errors.
Son-tinh (24W)23-29 October110/85 knotsTrack forecasts for this typhoon were excellent in all respects. Errors were low, skill scores high and the landfall very well predicted.
25W14-15 November25/30 knots25W did not attain tropical storm status and no forecasts were verified.
Bopha (26W)25 November-09 December140/100 knotsThe formation of Bopha was well predicted by the model. The global model had a right-of-track bias, but still had track errors near to or below last season's average. A 4km resolution regional model run at the time (right-hand frame) gave better predictions of landfall location than the global model.
Wukong (27W)24-28 December35/40 knotsTrack forecast errors were above last season's average for this storm due to a fast bias.

North-West Pacific observed tracks in 2012
North-West Pacific tropical cyclone names

North-east Pacific
Name & IDStart & end dates1-minute MSWComments on Met Office global model forecasts
Aletta (01E)14-19 May45 knotsTrack forecast errors were a little larger than average due to the failure to predict the sharp turn north as the storm weakened.
Bud (02E)21-26 May100 knotsThe change in direction of this storm were well predicted although there was a slow bias resulting in track forecast errors above last season's average.
Carlotta (03E)14-17 June95 knotsThe prediction of landfall was a little slow for this storm.
Daniel (04E)04-12 July100 knotsTrack forecast errors were low and skill scores high for this storm.
Emilia (05E)07-15 July120 knotsDespite a slight southward bias, track forecast errors were low and skill scores high for this storm.
Fabio (06E)12-18 July95 knotsTrack forecast errors were near to last season's average. There was a slight left-of-track bias during recurvature and the model tended to weaken the storm too soon.
Gilma (07E)07-11 August70 knotsTrack forecast errors were low and skill scores high for this hurricane.
Hector (08E)11-17 August45 knotsTrack forecast errors were low for this storm, although the model failed to predict the sharp northward turn later in its lifetime.
Ileana (09E)27 August-02 September75 knotsTrack forecast errors were well below last season's average for this storm.
John (10E)02-04 September40 knotsJohn was only briefly a tropical storm and no forecasts were verified.
Kristy (11E)12-17 September50 knotsTrack forecast errors were near to last season's average. Forecasts had a slight fast bias.
Lane (12E)15-19 September75 knotsForecasts had a left-of-track bias, but errors were below last season's average.
Miriam (13E)22-28 September105 knotsTrack forecast errors were near to last season's average.
Norman (14E)28-29 September45 knotsNorman was only briefly a tropical storm and no forecasts were verified.
Olivia (15E)06-09 October50 knots
Track forecast errors were mostly below last season's average for this storm.
Paul (16E)13-17 October105 knotsTrack forecast errors were mixed for this storm. Skill scores were high at 24 hours onwards.
Rosa (17E)30 October-04 November45 knotsRosa took an unusual south-westwards track, but was well predicted by the model with low track forecast errors and very high skill scores.
Central North Pacific
Name & IDStart & end dates1-minute MSWComments on Met Office global model forecasts

North-East and Central North Pacific observed tracks in 2012

North-East Pacific tropical cyclone names

Central North Pacific tropical cyclone names
North Atlantic
Name & IDStart & end dates1-minute MSWComments on Met Office global model forecasts
Alberto (01L)19-22 May50 knotsAlberto was short lived, but well predicted.
Beryl (02L)26-30 May60 knotsThe model did very well in predicting the formation and the track and landfall of Beryl. Forecast errors were low and skill scores very high.
Chris (03L)19-22 June75 knotsChris was very well predicted by the model with low errors and high skill scores.
Debby (04L)23-27 June55 knotsShort range forecasts were good, but the acceleration across Florida was not well handled.
Ernesto (05L)01-10 August85 knotsTrack forecast errors were mostly low and skill scores high for this hurricane. Only a few long lead time errors were somewhat larger with an incorrect forecast track into the central Gulf of Mexico.
Florence (06L)04-06 August50 knots
Florence was short-lived and very few forecasts verified.
Gordon (08L)15-20 August95 knots
Short range track forecast errors were low and skill scores high. There was a slow bias in longer lead time forecasts resulting in higher errors.
Helene (07L)09-19 August40 knotsAfter being a tropical depression for most of its lifetime, Helene was only a tropical storm briefly before landfall and no forecasts were verified.
Isaac (09L)21-30 August70 knotsTrack forecast errors were mostly low and skill scores high for this hurricane. Some longer lead time forecasts had a right-of-track bias resulting in landfall further east than occurred. The slow down in forward motion at landfall was well predicted.
Joyce (10L)22-21 August35 knotsJoyce was only briefly a tropical storm and no forecasts were verified.
Kirk (11L)28 August-02 September90 knotsTrack forecast errors were low and skill scores high for this storm whose track was very well predicted.
Leslie (12L)30 August-11 September70 knotsTrack forecast errors were close to last season's average, although skill scores were very high. Some forecasts failed to predict the storm's slow down, but others predicted this movement very well.
Michael (13L)03-11 September100 knotsDespite a left-of-track bias, track forecast errors were generally below last season's average (particularly at longer lead times) and skill scores high for this storm.
Nadine (14L)11 September-04 October80 knotsNadine was one of the longest-lived Atlantic tropical storms on record. Overall, track forecast errors were near to last season's average and skill scores high for this unusual storm. There were periods when track errors were very large - e.g. early on when a left turn towards Newfoundland was predicted and mid-life when a track towards western Europe was predicted. However, these large errors were offset by some excellent forecasts as the storm made its slow clockwise loop in the eastern Atlantic. During this period forecast errors were exceptionally low.
Oscar (15L)03-05 October45 knots
Oscar was short lived, but forecasts were good.
Patty (16L)11-13 October40 knotsPatty was very short-lived and no forecasts were verified.
Rafael (17L)13-17 October80 knots
Track forecast errors were near to or slightly above last season's average. However, skill scores were high.
Sandy (18L)22-30 October100 knotsTrack forecast errors were lower than last season's average at all lead times except 168 hours. Skill scores were also very high. This is probably due to the unusual turn westwards towards landfall over the USA. This turn was predicted by all runs of the model, but the sharpness and speed of the turn were generally underpredicted. This resulted in longer lead time forecasts predicting landfall too late and too far north. It should be noted that Met Office ensemble predictions (not shown) gave a very good indication of the left turn and landfall over the USA up to a week ahead of the event. The majority of Met Office model predictions gave a good indication of the intensity and size of the storm as it approached US landfall.
Tony (19L)22-25 October45 knots
Track forecasts were a little slow for this short-lived storm.

North Atlantic observed tracks in 2012

North Atlantic tropical cyclone names
North Indian
(forecast track charts for Global and South Asia models shown)
Name & IDStart & end dates1-/10-minute MSWComments on Met Office global model forecasts
Murjan (01A)23-26 October35/35 knotsTrack forecasts were a little slow for this storm.
Nilam (02B)29-31 October55/45 knotsThere was a right-of-track bias for forecasts for this storm.
03B17-19 November35/30 knotsOnly JTWC designated 03B as a tropical storm and it was very short-lived.
04A23-24 December40/30 knots04A was poorly analysed by the global model and forecast errors were also large. The higher resolution South Asia Model (right-hand frame) fared little better.

North Indian observed tracks in 2012

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