Total verification of tracks done in 2019 in East Pacific

These tracks, or lack of those, have been earlier discussed in separate blog entries system-wise or time-wise:

Alvin
Barbara
Cosme
TD 4-E
Dalila
Erick
Flossie
Gil
Henriette
Ivo
Juliette
Akoni
Kiko
Lorena
Mario
Narda
Octave
Priscilla
Raymond
TD 21-E

ghost storms Jan-May
ghost storms Jun-Jul
ghost storms Aug-Oct
ghost storms Nov-Dec

In 2019, I analysed 721 GEFS runs.

I made 134 tracks of storms, that actually formed.
I made 212 tracks of ghost storms.
I missed 177 times cyclogenesis, which actually happened within 384 hours.

At the lead time of 1 day (24 hours), there were 5 hits, 21 false detections and 7 misses. 15 available position comparisons produce an average error of 125 nautical miles.

At the lead time of 2 days (48 hours), there were 14 hits, 29 false detections and 13 misses. 24 available position comparisons produce an average error of 133 nautical miles.

At the lead time of 3 days (72 hours), there were 23 hits, 28 false detections and 16 misses. 28 available position comparisons produce an average error of 123 nautical miles.

At the lead time of 4 days (96 hours), there were 27 hits, 36 false detections and 24 misses. 37 available position comparisons produce an average error of 137 nautical miles.

At the lead time of 5 days (120 hours), there were 26 hits, 36 false detections and 32 misses. 36 available position comparisons produce an average error of 184 nautical miles.

At the lead time of 6 days (144 hours), there were 27 hits, 44 false detections and 39 misses. 36 available position comparisons produce an average error of 238 nautical miles.

At the lead time of 7 days (168 hours), there were 26 hits, 43 false detections and 45 misses. 33 available position comparisons produce an average error of 285 nautical miles.

At the lead time of 8 days (192 hours), there were 28 hits, 56 false detections and 43 misses. 38 available position comparisons produce an average error of 359 nautical miles.

At the lead time of 9 days (216 hours), there were 32 hits, 56 false detections and 44 misses. 40 available position comparisons produce an average error of 444 nautical miles.

At the lead time of 10 days (240 hours), there were 29 hits, 56 false detections and 50 misses. 34 available position comparisons produce an average error of 523 nautical miles.

At the lead time of 11 days (264 hours), there were 27 hits, 85 false detections and 58 misses. 33 available position comparisons produce an average error of 592 nautical miles.

At the lead time of 12 days (288 hours), there were 24 hits, 101 false detections and 61 misses. 28 available position comparisons produce an average error of 722 nautical miles.

At the lead time of 13 days (312 hours), there were 19 hits, 129 false detections and 66 misses. 28 available position comparisons produce an average error of 685 nautical miles.

At the lead time of 14 days (336 hours), there were 16 hits, 139 false detections and 64 misses. 25 available position comparisons produce an average error of 684 nautical miles.

At the lead time of 15 days (360 hours), there were 16 hits, 159 false detections and 64 misses. 26 available position comparisons produce an average error of 678 nautical miles.

At the lead time of 16 days (384 hours), there were 19 hits, 169 false detections and 61 misses. 28 available position comparisons produce an average error of 632 nautical miles.

The table below shows the number of cases, where a track with a certain peak intensity ended up actually peaking with a certain intensity.

The graph below shows the average position errors of my tracks in 2019 in red circles, NHC average forecast errors 2014-2018 in black circles, and a quadratic fit to NHC errors as black line.

The graph below shows a 30-day average of hits (in black), false detections (in red) and misses (in blue), calculated from analysed GEFS runs and interpolated values for the runs, that were skipped.

The following is otherwise the same as the previous, but the average is weekly.