Verification of tracks made of Erin using GEFS mean fields

I made six tracks of Tropical Storm Erin. Additionally, within 384 hours of cyclogenesis, there were eight instances, when I missed cyclogenesis entirely. It is interesting, that there were more detections and often a higher peak intensity, than with Hurricane Dorian.

As of this analysis, best track positions are available August 21, 18Z thru August 29, 06Z, and as such, position comparisons are possible only for this timespan.

Since I give only SSHS categories in my tracks, I don't calculate intensity errors in units of speed, but in categories. For this purpose, I define tropical or subtropical storms as "Category 0", tropical or subtropical depressions as "Category -1" and non-existence of a tropical or subtropical cyclone as "Category -2".

In the six tracks, formation times ranged between August 26, 12Z and August 29, 12Z. Average was August 28, 06Z and median was August 28, 18Z.

In the 14 cases (6 tracks and 8 complete misses), peak intensity ranged between Category -2 and Category 4. Average was Category -0.36 and median was Category -2.

In the six tracks, dissipation times ranged between August 30, 18Z and September 8, 18Z. Average was September 5, 12Z and median was September 6, 09Z.  

At the lead time of 1 day (24 hours), there were 1 hit, 0 false detections and 0 misses. 1 available position comparison produces an average error of 14 nautical miles.

At the lead time of 2 days (48 hours), there were 1 hit, 0 false detections and 1 miss. 1 available position comparison produces an average error of 41 nautical miles.

At the lead time of 3 days (72 hours), there were 0 hits, 1 false detection and 1 miss. No available position comparisons.

At the lead time of 4 days (96 hours), there were 2 hits, 1 false detection and 0 misses. 2 available position comparisons produce an average error of 164 nautical miles.

At the lead time of 5 days (120 hours), there were 1 hit, 1 false detection and 1 miss. 1 available position comparison produces an average error of 161 nautical miles.

At the lead time of 6 days (144 hours), there were 1 hit, 1 false detection and 2 misses. 1 available position comparison produces an average error of 196 nautical miles.

At the lead time of 7 days (168 hours), there were 0 hits, 2 false detections and 2 misses. No available position comparisons.

At the lead time of 8 days (192 hours), there were 0 hits, 2 false detections and 2 misses. No available position comparisons. 

At the lead time of 9 days (216 hours), there were 1 hit, 3 false detections and 2 misses. 1 available position comparison produces an average error of 741 nautical miles.

At the lead time of 10 days (240 hours), there were 0 hits, 5 false detections and 1 miss. No available position comparisons. 

At the lead time of 11 days (264 hours), there were 0 hits, 5 false detections and 1 miss. No available position comparisons.  

At the lead time of 12 days (288 hours), there were 0 hits, 5 false detections and 2 misses. No available position comparisons.

At the lead time of 13 days (312 hours), there were 0 hits, 4 false detections and 2 misses. No available position comparisons.

At the lead time of 14 days (336 hours), there were 0 hits, 4 false detections and 2 misses. No available position comparisons.

At the lead time of 15 days (360 hours), there were 0 hits, 2 false detections and 3 misses. No available position comparisons.

At the lead time of 16 days (384 hours), there were 0 hits, 2 false detections and 3 misses. No available position comparisons.

 

 

Circle coloration scheme

 

 

Here are the tracks once again.

 

August 19, 12Z run:

 

 

August 20, 06Z run:

August 21, 06Z run:

August 23, 06Z run:

August 25, 06Z run:

August 26, 12Z run:

The complete misses were from the following runs:

August 11, 06Z

August 12, 06Z

August 13, 00Z

August 14, 00Z

August 15, 06Z

August 16, 06Z

August 18, 12Z

August 22, 12Z