Verification of tracks made of Narda using GEFS mean fields

November 14, 2019

I made eleven tracks of Tropical Storm Narda. Within 384 hours of cyclogenesis, there were no instances, when I missed cyclogenesis entirely. Therefore, overall detection of the system was perfect. However, the timing of its existence as a tropical cyclone was often pretty erroneous. Peak intensity was mostly overestimated.

As of this analysis, best track positions are available September 26, 12Z thru October 1, 12Z, 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 eleven tracks, formation times ranged between September 27, 18Z and October 1, 18Z. Average was September 29, 16Z and median was September 28, 18Z.

In the eleven cases (11 tracks and 0 complete misses), peak intensity ranged between Category 0 and Category 5. Average was Category 2.45 and median was Category 3.

In the eleven tracks, dissipation times ranged between September 29, 18Z and October 9, 06Z. Average was October 5, 02Z and median was October 4, 12Z.

At the lead time of 1 day (24 hours), there were 1 hit, 1 false detection and 0 misses. 2 available position comparisons produce an average error of 58 nautical miles.

At the lead time of 2 days (48 hours), there were 1 hit, 1 false detection and 1 miss. 2 available position comparisons produce an average error of 66 nautical miles.

At the lead time of 3 days (72 hours), there were 1 hit, 1 false detection and 2 misses. 2 available position comparisons produce an average error of 35 nautical miles.

At the lead time of 4 days (96 hours), there were 2 hits, 0 false detections and 1 miss. 2 available position comparisons produce an average error of 172 nautical miles.

At the lead time of 5 days (120 hours), there were 2 hits, 0 false detections and 0 misses. 2 available position comparisons produce an average error of 295 nautical miles.

At the lead time of 6 days (144 hours), there were 1 hit, 1 false detection and 1 miss. 1 available position comparison produces an average error of 349 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 0 hits, 3 false detections and 1 miss. No available position comparisons.

At the lead time of 10 days (240 hours), there were 0 hits, 3 false detections and 1 miss. No available position comparisons.

At the lead time of 11 days (264 hours), there were 0 hits, 2 false detections and 2 misses. No available position comparisons.

At the lead time of 12 days (288 hours), there were 1 hit, 1 false detection and 1 miss. 1 available position comparison produces an average error of 973 nautical miles.

At the lead time of 13 days (312 hours), there were 1 hit, 2 false detections and 1 miss. 1 available position comparison produces an average error of 774 nautical miles.

At the lead time of 14 days (336 hours), there were 0 hits, 4 false detections and 1 miss. 1 available position comparison produces an average error of 409 nautical miles.

At the lead time of 15 days (360 hours), there were 2 hits, 2 false detections and 0 misses. 2 available position comparisons produce an average error of 772 nautical miles.

At the lead time of 16 days (384 hours), there were 2 hits, 3 false detections and 0 misses. 2 available position comparisons produce an average error of 468 nautical miles.

Circle coloration scheme