Thread by @SkyPixWeather

1/9 Those rooting for long-track tornadoes & destruction yesterday probably were wondering what went "wrong" with the forecast. Those not as sadistic wondered what we missed w/overforecasting in a results-based sense. I have ideas from deep professional/scientific experience.

2/9 Dedicated scientific forecasters are like best NFL cornerbacks: shake off previous blown coverage, be better next play, but learn from it in film room. Data is our "film room". Atmosphere speaks the language of data. We ought to listen/read its language.

3/9 Put another way: every forecast is a test, & the atmosphere always has the answer key. Probabilistic forecasters aim to calibrate around zero line on reliability diagrams. Over/under-forecasts balance out over time, for consistently best forecasters.

4/9 High-amplitude "misses" are obviously undesirable, but good forecasts over long haul dampen their reliability impact (statistical reality) & keep them from undermining long-term credibility (perception). We aim for reality=perception=excellence, knowing perfection impossible.

5/9 One HUGE issue yesterday: warm-sector (WS) storm coverage. WS initiation requires just subtle lift on weak-cap days (e.g. 3 May 99, on which I've published papers) & usually doesn't occur away from well-defined boundaries in a strong EML/cap setup.

6/9 Humans + many models forecast weak capping for several hours in corridor well E of dryline, multiple WS supercells in that window, starting before dryline episode. Only 1 real non-dryline-area discrete WS supercell (Cashion) happened. Why? Atmosphere has spoken via data.

7/9 Consider soundings. Here's OUN at 18, 21, 00Z. Compare. Note warming inversion layer above 700 mb through aftn. That will suppress subtly forced WS convection over time. Note wind direction in stable layer & examine upstream soundings yourself.

8/9 Look at series of 700-mb charts yourself too. Though warmer above, warm nose was detectable there too. Origin: Mexican plateau, Coahuila SWwd. Not "mid/upper subsidence inversion" but EML advection. Mex. EML was slightly stronger/more north/sooner than expected over area.

9/9 That alone likely is enough to account for reduced WS dvlpmt in subtle forcing despite historically intense CAPE+shear parameter space. Another factor: lack of insolation===>weak bndry-layer lapse rates, weaker WS lift.

Good fodder for deeper/formal research! Go for it.