Warren-
Thanks for your post. I'm attaching an image from
the MEG tropical recap presentation given a few weeks ago. The
top panel is track error, and the bottom panel is intensity bias -
both for the Atlantic basin covering the past 3 hurricane seasons.
The red is the ops GFS, the blue is the original FV3GFS
configuration, and the green is the final configuration. (A
science change was made in early August to address the significant
low intensity bias in tropical cyclones).
It's clear that both the initial and final FV3GFS
improve the track, relative to the ops GFS. The original FV3GFS
configuration clearly had a significant weak intensity bias. It's
mitigated in the final configuration (change shown by the purple
arrow), but beyond day 2, the FV3GFS still has a low intensity bias
relative to the GFS. So your hunch that the FV3GFS is underdoing
the intensity of Florence may be very much on the right track.
That said, even though the composite stats don't reflect it, the GFS
is prone to going crazy with the central pressure in the strongest
storms (i.e. the sub 900 pressures in 2017 Irma forecasts and
perhaps in some Florence runs).
-Geoff
