Multicells: how storms move by building new storms

A cluster can move northeast while every cell in it moves east, or sit still for three hours while its cells race away downstream. Once you can separate the two motions, radar loops stop being confusing.

The commonest storm there is

The Field Guide calls the multicell cluster the most common storm type, and gives the warm-season pattern: new cells form on the west or southwest edge, old cells decay on the east or northeast edge, each individual cell lasts 20 to 30 minutes, and the cluster as a whole lasts an hour or more.

Read those two timings together. The cluster outlives its own cells by a factor of three or more. Nothing in the cluster is old. The cluster is old.

The mechanism is lesson 2's cold pool, used constructively instead of fatally. The gust front from the mature cells keeps spreading, and where it runs into warm, moist inflow it lifts that air to its LFC and a new updraft goes up. The old cell dies exactly as before. It just leaves a successor.

Two motions, not one

This is the idea the lesson exists for. What a radar loop shows you is the sum of two different things:

1. Advection. Each cell is carried along by the mean wind through the depth it occupies. This is cell motion, and it is roughly the same for every cell in the cluster.
2. Propagation. New cells form preferentially on one flank and old ones die on the other. This is the cluster growing in one direction and shrinking in the other, and it is a motion of the system that no individual cell is performing.

System motion is the vector sum. If new cells keep forming on the upwind flank at the same rate the cells are being carried downwind, the two cancel and the system does not move at all while every cell inside it is moving at 30 kt. That is the flash-flood case, and it is why "the storms are moving fast" is not by itself a reason to relax.

backbuilding

New cells forming repeatedly on the upwind flank, so the system extends itself upwind as fast as its cells depart downwind. The system stalls or moves backwards relative to the cells.

training

Cells moving one after another along the same line, like carriages passing a level crossing. Each cell is brief; the rain over any point underneath is not.

Corfidi vectors

Corfidi put the propagation argument into two vectors that can be read off a forecast sounding. The propagation vector is taken as the opposite of the low-level jet, because that is where the moist inflow is arriving from and therefore where the new cells will go up.

Upwind (backbuilding) propagation
V_prop = V_cloud_layer - V_LLJ

Downwind (forward-propagating) MCS motion
V_mcs = V_cloud_layer + ( V_cloud_layer - V_LLJ )

V_cloud_layer is the mean wind through the cloud-bearing layer; V_LLJ is the low-level jet. Corfidi 2003. Which of the two applies depends on whether the system is cold-pool driven, which is the subject of lesson 8.

Learn the shape of the argument rather than the arithmetic. The upwind vector says a system fed by a strong low-level jet will try to build back into that jet, and can therefore sit still or move against the mean wind. The downwind vector says a system whose cold pool is doing the lifting adds propagation to advection instead of subtracting it, and can move considerably faster than the wind at any level. Those are the two ends of the same idea and they produce opposite hazards: flooding at one end, a bow echo crossing three counties in an hour at the other.

Boundary collisions

Where two outflow boundaries meet, air has nowhere to go but up, and both boundaries are already convergent lines in their own right. Collisions are among the most reliable initiation points there are, and on a summer afternoon here the collision is often between a sea-breeze front moving inland and an outflow boundary from an earlier storm moving the other way.

This is a genuinely useful nowcasting habit: watch the boundaries on the low-level reflectivity, not just the storms. The fine lines are where the next hour happens.

MCS and MCC

MCS

Mesoscale convective system. A general term for a collection of storms behaving as one system, of which a multicell cluster, a squall line and a bow echo are all examples.

MCC

Mesoscale convective complex. A specific, satellite-defined subset of MCS with size, shape and duration criteria attached.

Maddox defined the MCC in 1980 from infrared satellite imagery: a cloud shield with an area of at least 100,000 km² colder than −32 °C, containing an interior area of at least 50,000 km² colder than −52 °C, meeting both size criteria for at least six hours, and with an eccentricity of at least 0.7 at maximum extent. The point of showing you those numbers is not to memorise them. It is that MCC is a defined term with a measurable test, and MCS is a description. Using them interchangeably is a small error that compounds.

Check yourself

On radar, individual cells are racing northeast at 35 kt, but the cluster as a whole has barely moved in ninety minutes and the rain totals underneath are climbing fast. What is happening?

1. New cells are forming on the upwind flank as fast as the old ones depart, so propagation is canceling advection. This is a flash-flood signature
2. The radar is mis-tracking the cells, since a cluster cannot move differently from the cells inside it
3. The cluster is about to bow forward, since a stationary system means the cold pool is winning
4. The cells are weakening, which is why the cluster has stopped advancing

Sources

1. NWS Weather Spotter's Field Guide (YPA-201154, June 2011), section 2, the multicell cluster storm
2. Corfidi 2003, Wea. Forecasting 18, 997–1017: cold pools and MCS propagation, forecasting the motion of downwind-developing MCSs
3. Maddox 1980, Bull. Amer. Meteor. Soc. 61, 1374–1387: mesoscale convective complexes
4. NWS JetStream, the thunderstorm and stability sections