Thread by @NickPBassill

I've seen some understandable comparisons between #Teddy's future and Hurricane Sandy (2012), so let me discuss how they're a bit similar, and also why they're different.

Sandy was known for its significant G(E)FS vs EC track splits, where the latter correctly had a "left hook"

2) Conventional wisdom said the EC was right because of (a) higher resolution, and/or (b) better data assimilation. Many still believe these things about Sandy.

But, you can get "correct" Sandy tracks with GEFS initial conditions+90 km grid+EC physics: agupubs.onlinelibrary.wiley.com/doi/full/10.10…

3) I showed why in the follow-up paper - the GFS' cumulus parameterization (CP) was built such that it could only respond to what was occurring near the tropical ocean surface, and thus couldn't "feel" the approaching upper trough, which limited phasing.

agupubs.onlinelibrary.wiley.com/doi/full/10.10…

4) Simply upping the value of entrainment in the CP, which decreases CP activity, suddenly allows for EC-like tracks (red), even with a low-resolution WRF model with G(E)FS ICs and G(E)FS physics! Blue tracks are "normal" GFS entrainment, and green are higher entrainment tracks.

5) I'll skip the gory details, but this allowed for more grid-scale (aka microphysics) latent heat release, mostly between the upper trough and Sandy, which helped build an upper-level anti-cyclone, and locally increase the horizontal PV gradient, thus "attaching" Sandy & trough.

6) Could this be at play with differences in EC vs. G(E)FS vs. ? forecasts for #Teddy? Absolutely! Maybe? I don't know-models get updated over time, and biases change. However, one reason I'm a bit uncertain is that with Sandy, the "players" were well forecast up to bifurcation.

7/7) In #Teddy's case, the two pieces (Teddy & the upper trough) have definitively NOT been consistently forecast run to run, as is evident in this model trend loop of upper level PV + vorticity in the deterministic GFS, via @TropicalTidbits. Sandy's trough was also MUCH bigger.