Iowa and Northern Missouri in the Convective Crosshairs

I had reconciled myself with the thought that this autumn’s storm season would be a no-show in the Great Lakes and Corn Belt. Leave it to the atmosphere to prove me wrong, but I’m not complaining. How’s this for a NAM forecast sounding for October 23 in Des Moines, Iowa (click on image to enlarge)?

Maybe not the sexiest hodograph, but I won’t kick it out of bed for eating crackers, and you’ve got to love that 1,811 J/kg SBCAPE. Storm-relative helicity could be better, but still, you got yer 0-2 km EHI of 2.1, yer -6.4 LI, 4 km VGP of .3, nice influx of moisture, good upper level support…what more do you want this time of year? Eggs in your beer?

Sixty-five miles southeast of Des Moines in Moravia, the sounding for 21Z looks even better with slightly bigger CAPE, -6.8 LI, a curvier hodograph, a southwesterly H5 cruising along at a stout 50 knots, BRN of 26, and better 0-6 km shear. Bulk shear is actually a bit of a concern–the NAM sounding gives a less optimistic view of it than does the map–but we’ll see how that plays out in future runs. I have a hard time believing that shear won’t be adequate.

I’m seriously contemplating going after this scenario. And it’s just a forerunner; another, stronger system looks to be moving through the Great Lakes in the Monday/Tuesday time frame, with backing surface winds pumping a nice plume of moisture into the region.

All in all, those of us up here in the great north woods may get one or two last whacks at some decent storm chasing and maybe even a few tornadoes before the snows fly. I’m keeping a close eye on this setup and crossing my fingers.

Forecast Model Simulations for 1965 Palm Sunday Tornadoes: Part 2

The drive down to the WFO at State College, Pennsylvania, was well worth my while (see my previous post). Operational forecaster and research meteorologist David Beachler was a pleasure to work with–personable, patient, and eager to help me understand the exhaustive forecast simulations he had produced on the 1965 Palm Sunday Tornadoes. Having pored over the data with David, gaining his insights on its strengths and weaknesses, I am now extremely excited about what I’ve got on my hands.

David’s modeling uses the WRF-ARW 40 km. The resolution is too coarse to offer the fine details that the SPC is capable of producing, but it gives an excellent overall feel of what forecasters and storm chasers might see in the models if the Palm Sunday synoptic setup were to unfold today instead of forty-five years ago in 1965.

no images were found

There’s no way I can begin to cover all the material, which in any case I need to sift through in order to put together a reasonably concise and meaningful scenario. But I can at least give you a sample of some of the stuff I’ve got to work with. Click on the following images to enlarge them.

First, here is a hand analysis of the kind that is accessible to anyone through NOAA’s historical daily weather maps archives. Besides the surface map for April 11, 1965, you also get the previous day’s surface map, 500 mb chart, and other info. It’s what you would have encountered when you turned to the weather page in the newspaper that morning.

What you would never have seen–because parameters such as CAPE, CIN, helicity, and so on didn’t exist back then, and because even if they had existed, the forecast models which could have depicted them were still years down the road–is this map showing SBCAPE and low-level shear.

The map is for 2200Z, or 6 p.m. EST–roughly the time at which tornadoes began moving through northern Indiana.

It gets even better. Here is a model sounding for KGRR, also at 2200Z, using WRF-ARW Bufkit data. The skew-T and hodograph depict the conditions that were shaping up to produce the F4 Alpine Avenue tornado that formed

around 6:50 p.m., as well as other tornadoes in west and southwest Michigan that day. The helicity is impressive–and look at those winds! Forty knots at 850 millibars is no mere puff of air.

What really excites me is that, using RAOB’s cross-section feature, I should be able to reconstruct a vertical profile of the atmosphere for the entire outbreak area. I’m not sure how deeply I want to go with that, but I have the capacity.

Bear in mind that I’m just showing a couple of representative glimpses derived from a 00Z, day-one model initiation. In fact, David provided me with a range of initiation times that allows me to get a good sense of how the maps might have progressed from several days prior to the actual tornado outbreak.

In practical terms, the maps and model sounding data I’ve got correlate to the NAM. They’re not the NAM, but for storm chasers who typically work with the GFS, ECMWF, GEM, NAM, and RUC, what you see here is probably closest to what you’d find using the North American Mesoscale Model.

That’s all for now. This has been a time-consuming post, and at 2:30 in the afternoon, I need to pull away from it so I can bathe and eat. I didn’t arrive home until 3:00 a.m., so it’s time for this road warrior to reset his time clock and get on with the rest of life.

Year of the Cap Bust

I guess I’m just a slow learner when it comes to technical stuff that involves linear thinking. Sooner or later, though, if I stick with something long enough, I usually emerge more knowledgeable for having done so. Nowhere has this been more true than in storm chasing, an activity which obviously depends heavily on figuring out if and where there will be decent storms to chase.

Seems like I’m constantly being confronted with some new aspect of the atmosphere that I haven’t factored into my forecasting, or that I haven’t factored in as effectively as I needed to. The upside of that, though, is that I wouldn’t even be aware of what I don’t know if I hadn’t learned enough to at least recognize my areas of ignorance. If my forecasts aren’t as expert as those of a trained meteorologist, they’re nonetheless a seven-league bound beyond when I was just beginning to grapple with all those arcane terms and acronyms of convective weather such as SBCAPE, CIN, 0-6km bulk shear, LIs, helicity, and lapse rates, and when the only thing I could do with a skew-T or a hodograph was shake my head in bewilderment.

This has been the year of discovering the 700 mb/12C limit. By “discovering,” I mean through empirical experience, and by “empirical experience,” I mean cap busts. Of course, I’ve endured plenty of cap busts in my development as a storm chaser; I just didn’t understand exactly what was going on, or why the high risk area I was sitting in was producing nothing more than smug blue skies rather than carnivorous supercells.

One memorable day in Iowa drove home the lesson perfectly. MLCIN was supposed to erode by later in the afternoon, and it got to a point where it was eroding, at least according the RUC. With SBCAPE at some ridiculous figure like 7,000 j/kg, I figured that at some point a convective tower would punch through the cap and go absolutely gonzo. Instead, the clouds kept firing up into the nicely sheared environment and then dying, firing and dying, firing and dying. The reason? A 700 mb temperature of around 14C, possibly considerably higher. Lesson learned: it doesn’t much matter what the models have to say about the CIN eroding when you’ve got mid-level temperatures like that.

I experienced another cap bust yesterday, though I can’t feel too bad about it since I had no compelling reason to head out in the first place, the conditions were so marginal. It was interesting to notice that in this situation, the circumstances were reversed: RUC showing my area under very breakable 700 mb temps of around 10C, but with MLCIN creating some concern. However, the CIN appeared to be eroding, and when an SPC mesoscale discussion spotlighted the area I was in, I started feeling happy about having made the drive down to west-central Illinois.

But the CIN started building back in, and by 00Z I found myself socked in under values around -300 j/kg. Not much a parcel of air can do with that, I guess, no matter how big the CAPE is. I turned around and headed home.

A paper by Bunker, Wetencamp, and Schild of the NWS in Rapid City, South Dakota, explores the ins and outs of the 700 mb/12C limit and concludes that it has a limited, conditional application. However, as my buddy Mike Kovalchick pointed out to me, the paper also reveals that only 5 percent of tornadoes within the study period formed when H7 temps exceeded 12C, and virtually no violent (EF4 and EF5) tornadoes occurred above that threshold.

So for practical use in storm chasing, the 12C limit appears to be a very useful rule of thumb. The issue for me then becomes a matter of refining my ability to know when cold air advection will lower the 700 mb temps. But that’s a subject for another blog. I’m tired of thinking. It’s time to go meet my buddy Dewey down in Plainwell and grab a brew at Arie’s. Ciao!