Convective Rain

Convective rain falls from buoyancy-driven updrafts. The distinguishing features are all consequences of the updraft being strong and narrow: high intensity, short duration, and small spatial extent, with sharp gradients at the edges. Stratiform rain, produced by gentle widespread ascent, is the opposite on every count — lower intensity, longer duration, larger area, smoother.

The distinction is not merely descriptive. It is why the same daily total means different things in different regimes. Fifty millimetres falling over twelve hours from stratiform cloud largely infiltrates. The same fifty millimetres in ninety minutes from a convective cell exceeds infiltration capacity and runs off, which is the difference between a wet day and a flash flood. Daily accumulation, the most commonly available variable, discards exactly the information that determines the impact.

Why models struggle with it

Convective cells are typically a few kilometres across. A climate model grid cell is tens to hundreds. The cell is therefore invisible to the resolved dynamics and must be handled by a cumulus parameterisation, which produces a grid-cell-mean rain rate rather than a cell.

Two systematic errors follow, and both are well documented across generations of models:

• Rain that is too frequent and too light. The scheme rains a little in many cells rather than a lot in a few. Means look reasonable; the intensity distribution is badly wrong, with too few extremes and too few dry hours.
• The diurnal cycle peaks too early. Schemes tend to fire as soon as CAPE appears, giving a rainfall maximum near local noon instead of late afternoon or evening. This is a trigger problem, closely tied to how CIN is handled.

Convection-permitting models at 1-4 km improve both substantially, because convection initiates and organises through the resolved flow. The improvement in the diurnal cycle is one of the clearest results in regional climate modelling. Intensity distributions improve too, though a tendency toward cells that are too intense and too infrequent is common — the errors change character rather than disappear.

Where I have run into it

In the Galapagos work, the distinction matters for what a rain gauge record means. Heavy rainfall events in the archipelago are convective and localised, so a gauge either catches a cell or misses it entirely, and network density limits how much can be said about spatial structure. The garua season presents the opposite problem: persistent low-intensity moisture delivery that gauges also handle poorly, for different reasons — see occult precipitation.

See also: precipitation for the general treatment, CAPE and CIN for the energetics, and cumulus convection schemes for the parameterisation problem.