A tale of two ensembles

Thursday morning we awoke to a line of thunderstorms across central OK. Eastern CO had what appeared to be a cold air dam and to the east southeast appeared to be situated a deformation zone. Given the stronger deep layer shear, this small area of convection intensified an a southward moving supercell developed around 12 UTC. This cell and its previous incarnation had dropped a bunch of hail (drifts near Blanchard, and larger hail with the supercell).



On the synoptic scale there was a trough moving south-southeast with an attendant cold front located across northeastern MO, with a developing low level trough forming south of the dam emerging from the mountains eastward across the NM and TX panhandle.

This unique environment set up the whole day of convection across west and west central TX. The CAPS ensemble handed this event remarkably well. It even developed the supercell coming out of the line and moving south and then southeast. This anchoring forecast helps the CAPS ensemble point to this solution, while the NSSL-WRF, the anchoring member of the SSEO, failed to depict any convection at all in OK. Thus the tale of two ensembles: one with radar data assimilation 12 hours prior that has the OK convection and the other without radar data assimilation that doesnt.

However, the NSSL-WRF hot start forecast did have the convection in OK. So clearly we are wandering around chaos theory. Exactly what the differences were will be a subject for future research. But obviously, we are talking about small differences in the initial state of the atmosphere and its subsequent effects that can make or break a high resolution forecast.

The resulting forecasts hindered on the outflow boundary moving westward in the TX panhandle and down into central north TX and initiating convection in the afternoon. There were many features of this forecast that remained complicated. One was how quickly the initial supercells would grow upscale. If they did grow upscale would they develop southeastward along the outflow boundary or south/southwestward in an environment that may ot be conducive to wind gusts? In any case, how severe would this convective system be?

We had to choose and our team opted to try and split the difference in a manner of speaking. We would target the initial supercells as initiating early in the 1st period to produce severe weather but target the late periods for significant severe and then try to position our probabilities along the line of storms that developed in all ensembles. We chose to go with a low speed evolution. I am pretty sure we did well in this approach, though it wasnt very satisfying.

In the end the severe weather reports stayed with the supercells and eve though a linear structure emerged it was a result of the southwestward propagation of convective development. Thus the models are capable of giving us signals about the potential evolution of a system but we really need to ascertain, from the environment, what plausible and most likely outcomes will be.