Wind shear decides what kind of storm you get

Same moisture, same instability, same lift, different wind profile, completely different day. Shear is the parameter that turns a forecast of storms into a forecast of a storm mode.

Two kinds of shear, two depths

Speed shear

Wind speed changing with height while the direction stays put. Tilts the updraft downshear, which is what gets the precipitation to fall out beside the inflow instead of into it.

Directional shear

Wind direction changing with height. This is what generates horizontal vorticity with a component along the inflow, and it is where supercell rotation comes from in lesson 6.

0 to 6 km shear

Deep-layer shear. The parameter that decides whether a storm can organize into a supercell at all.

0 to 1 km shear

Low-level shear. The parameter that has something to say about whether a supercell is tornadic, which deep-layer shear does not.

Bulk shear is a subtraction, not an addition

Bulk shear over a layer is the vector difference between the wind at the top and the wind at the bottom. Not the difference in speeds. This trips people up constantly, and it always trips them in the same direction: they underestimate it when the direction changes and overestimate it when it does not.

Surface 10 kt from 150° u = -5.0 kt v = +8.7 kt
6 km AGL 50 kt from 250° u = +47.0 kt v = +17.1 kt

Shear vector du = +52.0 kt dv = +8.4 kt
Magnitude sqrt(52.0² + 8.4²) = 52.7 kt

A worked example. Note that 50 minus 10 would have given 40 kt and 50 plus 10 would have given 60. The answer is neither, because the wind turned 100° between the two levels.

52.7 kt of deep-layer shear is comfortably supercell territory. The threshold, from both idealised modeling (Weisman & Klemp 1982) and observations, is roughly 15 to 20 m/s, or 30 to 40 kt, of 0 to 6 km shear as a necessary condition for supercells.

What deep-layer shear does not tell you

Thompson et al. 2003 pulled proximity soundings from the Rapid Update Cycle for a large sample of storms and found that 0 to 6 km shear separated supercells from non-supercells cleanly: there was no overlap between the 10th percentile for supercells, around 15 to 18 m/s, and the 90th percentile for non-supercells. As a mode discriminator it is about as good as a single parameter gets.

The same study found that deep-layer shear does not discriminate tornadic supercells from non-tornadic ones. That job belongs to 0 to 1 km shear and to low-level moisture, which is the lesson 1 point about LCL height arriving from the other direction.

Hodographs, at recognition level

A hodograph plots the tip of the wind vector at each height and joins them up. It is the wind profile with height taken out, which sounds like a loss until you see what it makes obvious. Two shapes are worth recognizing.

• Straight. Speed shear with little turning. A storm that splits produces a left-mover and a right-mover of roughly equal strength, and you get mirror-image storms diverging from where the first one was.
• Clockwise-curved, especially in the lowest kilometer or two. The split is no longer symmetric: the right-mover is favoured and the left-mover withers. This is the classic tornadic-supercell profile, and lesson 6 explains why the curvature does that.

A parameter that aged badly

The Bulk Richardson Number is CAPE divided by a measure of the shear, and it was proposed as a single number that would say whether an environment favoured supercells. In the Weisman and Klemp simulations, supercells appeared at BRN values of roughly 10 to 40. That range went into textbooks.

Then observed supercells were checked against proximity soundings and averaged BRN 57 to 78 (Edwards and Thompson, 20th Conference on Severe Local Storms). Not a contradiction of the modeling, and not a reason to throw the parameter away. The simulations were run on a small set of idealised profiles, and the real atmosphere produces a wider variety of them.

Which classification are we using

There are three schemes in circulation and they do not nest neatly. The Field Guide teaches five storm types. Doswell frames convection as a dual spectrum: one spectrum of individual cells running from ordinary to supercell, and a second spectrum of mesoscale systems, with no clean correspondence between them. SPC uses a radar-based convective mode scheme (Smith et al. 2012), which is what the modern climatologies are built on.

• Discrete right-moving supercell
• Cluster of cells containing a right-moving supercell
• QLCS
• Bow echo
• Linear or marginal supercell
• Non-supercell discrete cells and everything else

This course teaches the SPC scheme, for one practical reason: it is the vocabulary the statistics are collected in. When lesson 8 says a given share of tornadoes comes from QLCSs, that number exists because somebody classified thousands of radar images using these categories. A private taxonomy cannot be compared with anything.

Shear relative to the boundary

The single most useful mode-forecasting idea in this lesson is not the magnitude of the shear. It is its angle to whatever is doing the lifting.

• Shear roughly parallel to the boundary favours a linear mode. New updrafts form along the boundary and immediately share a cold pool with their neighbors, and the whole thing grows upscale into a line.
• Shear more normal to the boundary, that is, crossing it, favours discrete cells. Storms move off the boundary rather than along it, each keeps its own cold pool, and they stay separated long enough to organize individually.

Dial, Racy and Thompson 2010 quantified this: storm mode about three hours after initiation correlated with the component of the deep-layer shear normal to the initiating boundary, together with the boundary-relative mean wind, and the skill was best using the mean wind and shear angle over roughly 2 to 6 km. This is why two days with identical CAPE and identical shear magnitude can give you a photogenic isolated supercell and a 200-mile squall line.

Check yourself

The surface wind is 15 kt from the south and the 6 km wind is 45 kt from the west. Which is closest to the 0 to 6 km bulk shear, and what does it support?

1. About 47 kt, which supports supercells
2. About 30 kt, which is marginal for supercells
3. About 60 kt, which supports supercells
4. About 45 kt, because the surface wind is nearly perpendicular and cancels out

Sources

1. NWS Weather Spotter's Field Guide (YPA-201154, June 2011), section 2
2. Weisman & Klemp 1982, Mon. Wea. Rev. 110, 504–520: the dependence of numerically simulated convective storms on vertical wind shear and buoyancy
3. Thompson et al. 2003, Wea. Forecasting 18(6), 1243–1261: close proximity soundings within supercell environments from the Rapid Update Cycle
4. Smith et al. 2012, Wea. Forecasting 27(5), 1114–1135: convective modes for significant severe thunderstorms, part I
5. Dial, Racy & Thompson 2010, Wea. Forecasting 25(5), 1430–1446: short-term convective mode evolution along synoptic boundaries
6. Doswell 2001, Severe convective storms: an overview. Meteorological Monographs Vol. 28, No. 50, Amer. Meteor. Soc., 1–26
7. AMS Glossary of Meteorology, bulk Richardson number