Severe Weather 2/28/12

An active pattern is back, and we have severe rounds of severe wx to watch out for in the coming days. In this post, I will focus on the round today, which has been pegged by most in the weather circles as a forgettable precursor to a bigger event on Friday. Now I’m going to go out on a limb and say this has potential to be bad on its own. The usual disclaimers apply – i.e. my reasoning can be off sometimes, I’m not an expert, listen to official sources, etc. But I did want to share my two cents, even knowing I could be completely off.

Let’s say we had nothing but observational data and H5 and surface model progs available to us. The first thing I would notice is that we have an incredibly dynamic system on our hands: at H5 0Z tmrw we have SW winds of 90-100 kt, temps of <-15C north of I-40; at San Diego 0Z today we had hail. The moisture is there: 60F isodrosotherm to I-40, 65F isodrosotherm to I-20. Without any other data, such as CAPE/SRH/EHI/STP/even H7/H85/H2 data… this looks like a MDT/HIGH risk day.

But nobody is considering it one. Why?

What’s wrong is that the thermodynamics are not there supposedly. The moisture is there, the cold upper air temps are there, but the surface temperatures are cool. H5-sfc lapse rates SUCK. CAPE SUCKSx2. But wait. Let’s look at it more closely:

^A forecast sounding valid for 0Z tomorrow, for Hutchinson, KS.
Lapse rates are HELLA steep below H7 – nearly dry-adiabatic. Sure, we could use some better directional shear – and for that reason I don’t expect tornadic supercells to explode and multiply in srn KS – but the thermodynamic profile isn’t all that bad. The warm layer above H7 is the limiting factor. Erase that, CAPE values are okay, and regardless, 0-3 km CAPE values look pretty good from an initial look. We all know how bad models can be in regions of strong temperature advection, so it’s hard to promise that the warm layer will be modeled correctly at game time. I wouldn’t be surprised to see some tornadoes from low-topped supercells between extreme nrn OK and srn NE, but only initially in their storm cycles (the paltry directional shear will hurt storm modes over time) and close to the dryline where there will be solar insolation. Note that many models do attempt to break out precipitation in this area.

Such a threat would initially remain near and west of I-35. Because of the veered nature of the flow regime and the drought in the srn Plains, I think the dryline will mix a bit further east than the NAM indicates. Further east, stratus will rule the day, and models keep some type of warm front over central AR due to ongoing convection or some cold pool. Basically, north of the warm front is stable land. Furthermore, models keep a strong cap in the area through 0Z.

But after 0Z things change bigtime. Between 0Z and 6Z temperatures actually go up as the LLJ veers into AR and transports warm air into the area. Either (a) storms move in from the west, or (b) storms blow up along the LLJ axis as height falls overspread the region, but if these can become discrete in Scenario (a) or root to the surface in Scenario (b), I wouldn’t be surprised to see some sig tornadoes in wrn AR or ern OK tomorrow. Note that because of the capping and weaker forcing to the south the threat will remain north of the Red River.

(Compare 0Z and 6Z surface temps over AR below.)

This is the scenario I’m envisioning. Some low-topped supercells or a marginally severe QLCS develops along the dryline between 21Z-0Z, from extreme nrn OK to NE. These may produce a few isolated tornadoes and a few marginal severe reports. (5% tor, 15% hail, 5% wind) After around 0Z, the LLJ axis veers and intensifies, transporting higher theta-E air to the Ozarks and Ouachitas of ern OK and wrn AR. In this regime the enhanced lift from the LLJ will initiate storms starting around 02Z. Of course the storms will be elevated in the beginning, but as CINH is eroded the strongest cells in the new batch plus possible convection left over from the west could become tornadic around 04-05Z. (10% hatched tor, 15% hail, 30% wind) These storms will move east and slowly morph into a QLCS as surface flow veers. The tornado threat will decrease accordingly around 08-09Z.

Again not a HUGE threat day, but I think we might get some notable tors by the end of the day if the latter scenario verifies. And that may end up surprising some people who will overlook the setup tomorrow.

…Added post-script….
As a final note, and to tie this thing up, because this trough is so beautiful and dynamic, and because we have the low-level moisture that is so often missing in these setups, it might not hurt to be slightly more bullish than usual in this situation. There are instances that my intuitive feel from the most basic maps end up being more correct. And in this case, the intuitive look of this trough is MDT/HIGH risk.


So Bad…

I’ve really never seen the computer models this bad at 72 hrs. Note that the following maps are all different predictions from different computer models valid for the same exact time. The simulations were all run at 4:00 AM PST this morning:

1. CMC (from Canada)

2. UKMET (from Britain)

3. GFS/AVN (from the US)

4. ECMWF (from the EU)

5. WRF (from the US)

You don’t need to be an expert to tell anyone how different those maps look for just 3 days out. This lack of consensus is probably due to the fact that the relevant pieces of the puzzle (the energy off the coast of British Columbia that is to amplify over the central US, and the energy near Baja California) are both over data-sparse areas (not a lot of upper-air or surface observations in Mexico or over the ocean), so a lot of interpolation is used to estimate the orientations/positions/magnitudes/etc. of the relevant fields. Because this setup is so sensitive to initial conditions, small differences in these estimations are leading to significant differences in later outcome. (Read this.) In general however, yesterday’s computer models had better consensus and were, in general, closer with the two pieces of energy.

This post will serve to fill space.

Really, I feel like my mind has drawn a blank in the last semester. So sad 😦 (I’ve been thinking about some stuff, but nothing too deep.)

Anyway, I don’t feel like studying, I do feel kinda stressed though, and I feel a bit sick. So what else is there to do but blog post and *hope* your mind will start generating bloggable stuff in the next 30 minutes of your life.

Only six people have responded to my poll question about spherical coordinate notation, which is too bad. One thing that has intrigued me lately is that appearance of the greek letters \phi and \theta naturally naturally lend to angulations.

First, let’s look at \theta. I hope all readers (lol) can see how the outline of theta resembles a sphere, and how the line/curve that cuts through the middle resembles a circle parallel to the x-y plane (use your imagination, at least!). This circle is a curve of constant radius r and constant polar angle, and spans the space of all possible azimuth angles in [0, 2\pi].

Now, let's look at \phi. Likewise, the "circular" thing is the sphere, and the line that cuts through the sphere resembles a line of constant azimuth angle, with the polar angle varying across the line.

That's why I personally think the physics notation [radius, azimuth angle, polar angle] \rightarrow [ r, \phi, \theta ] makes more sense, because \phi looks like a visual representation of a line with constant azimuth angle, while \theta looks like a visual representation of a great circle with constant polar angle. But I suppose the math notation [radius, azimuth angle, polar angle] \rightarrow [ r, \theta, \phi ] makes sense too, since the circle that cuts through \theta occupies the space of azimuth angles, while the line that cuts through \phi occupies the space of polar angles. To each his/her own.


Next short thought: modeling atmospheric parameters. We'll denote an arbitrary parameter P that depends on all three spatial variables: the horizontal position \vec{\rho} \equiv (x,y) and z. Note this does NOT include integrated quantities like CAPE or SRH whose integration effectively reduce one degree of freedom (in my examples, the z direction) when we attempt to look at the atmosphere. It does, however, include lesser-known parameters such as the positive buoyant energy which goes by the difference between the virtual temperature of a parcel and that of the environment. Or, of course, the wind velocity field.

The problem with what we have now is that we are given maps of either constant z (the weather maps on a standard pressure level such as H5, H7, H85, etc) or constant \vec{\rho} (skew-t or hodograph plots). And neither of these maps tell the complete, because we are always losing one dimension when we look at them. Not saying they aren't useful---because they are incredibly so, but I feel at least that one of the greatest challenges of an interpreter of these maps is to keep track of the bigger picture---one that includes both \vec{\rho} and z.

The obvious remedy is to create three-dimensional weather maps, but the computational power needed for such a maps might still be a long ways away. Until then, this is a challenge for weather forecasters. We can look at different soundings for different locations, but the most impressive (or nonimpressive) soundings will stand out, which may lead us to forget the spatial variance of the parameters in those soundings. Likewise, we can look at maps from standard pressure levels, but if we treat them independently rather than see how they link together from the tropopause to the surface, we will often miss something. It's a challenge, for sure, and my opinion is that for amateurs, this is one of the greatest obstacles to overcome.

Super Bowl

First off, I am going to say I am really happy about my championship game predictions. They were tough picks and I was very close with the Ravens-Patriots game, and only one point off for the Giants-49ers game. (Okay, I will stop bragging now.)

I’ll keep this short… I think the Giants have one of the most balanced offenses/defenses we’ve seen thus far in the league. They’ve got an immense pass rush and have already beaten the Patriots in the regular season. Recent history has shown that the record doesn’t matter, but rather how balanced you are on both sides of the ball during the playoffs. From the get-go I had an intuitive feeling that the Giants would make it to the Super Bowl. Now that they’ve made it, I think they will win. The Patriots are built very similarly to the Packers, which the Giants trounced three weeks ago.

X-factors: whether the secondary of the Patriots shows up, whether Patriots TE Gronkowski will play.

Final score: Patriots 31 Giants 37