Room temperature stratification: the air at your ankles

Air arranges itself by density and then holds that arrangement. Warmer air carries less mass in the same volume, so it settles above cooler air and stays there. Left undisturbed, a room becomes a stack of bands instead of one body of air at one temperature.

Nothing inside a still room undoes that sorting. The denser air is already at the bottom, so no imbalance remains to drive a return flow. Stirring has to arrive from outside the arrangement, which in a flat means a fan, an open door, or somebody walking through.

Every warm object in the space feeds that stack. ASHRAE defines a stratified zone as one where air movement is driven entirely by buoyancy from convective heat sources. A person is such a source. So is a laptop, a router, a downlight, and a patch of floor that took the afternoon sun.

Each source launches a plume, and the plume changes as it climbs. ASHRAE's chapter on space air diffusion describes it directly: a thermal plume rising by natural convection above a heat source draws in surrounding air, so it grows in size and volume and slows down. Warm air leaves the source, collects room air on the way up, and delivers the mixture to the ceiling.

The band near the ceiling therefore holds the room's own heat, carried there by a process that runs whether the aircon is on or off. Switching a unit on adds a second flow working against the first. Where the room settles is the outcome of the two, and neither of them is under the remote's control.

A wall unit reads the air arriving at its own intake, high on the wall. In a room being cooled, that is the warmest band in the space. It also sits at the opposite end of the stack from everybody in the room.

Manufacturers write this down in their own service literature. Daikin's service manual for inverter wall-mounted splits defines room temperature as the temperature of the lower part of the room, then lists room thermistor temperature as a separate quantity. The manual states there is a difference between the two, depending on the type of indoor unit and the installation condition.

What the control does about it is the interesting half. The same manual says temperature control is performed using a target temperature appropriately adjusted for the indoor unit, together with what the thermistor detects. In other words, the board applies an assumed offset between the plane it can measure and the plane people occupy.

An assumed offset is a factory figure for that model, not a measurement of this flat. A room that layers more heavily than the assumption receives exactly the same correction as one that layers less. So when a display reports the room has reached its target, the honest reading of that claim is narrower than it looks.

ASHRAE Standard 55 fixes its own planes for judging comfort. Ankle level is 0.1 m above the floor. Head level is 1.1 m for a seated occupant and 1.7 m for a standing one. The tallest of the three still sits below the intake of a wall unit in an ordinary flat.

The gradient is a balance, and both sides of it move. ASHRAE states that the amount of stratification in a room is determined primarily by the balance between total room airflow and heat load. Push more air across the room and the bands blend together. Add heat, or spread it further up the height of the room, and they separate again.

Where the heat sits turns out to matter as much as the total quantity. Researchers at Aalto University and Halton measured air temperature profiles at ten heights in a test room built at two different ceiling heights. With low-level heat sources, most of the gradient sat inside the occupied zone regardless of room height. Room height became significant when the heat loads were high up.

That study examined displacement ventilation, a design that sets out to stratify a room on purpose, so its numbers do not transfer to a wall split. The part that carries over is about heat sources. Their height decides where in the room the layering lands, which is why two rooms with the same cooling load can feel very different.

Glazing acts on both sides of the balance at once. Sun through a window warms the floor and the furniture, and those surfaces then behave as low-level sources feeding the occupied zone. A tall window admits that heat across more of the room's height, so the stack gets fed at several levels instead of one.

Ceiling height sets the size of the reservoir. More volume above head height means more air the outlet has to turn over before the lower band improves, and that volume is fixed on the day the flat is built.

None of these five is a fault, which is what makes the conversation difficult. Each one is a property the flat already had before any unit was fitted. A machine can be asked to work against them. It cannot be asked to remove them.

A room can satisfy its sensor and still fail the people inside it. That is the entire cost of the gradient, and it never surfaces as a fault code, a pressure reading, or anything else a technician can put on an invoice.

Direction matters more than most people expect. The ISO 7730 model for dissatisfaction from vertical temperature difference applies when temperature increases upwards, and the standard notes that people are less sensitive when it decreases. ASHRAE Standard 55 puts it more bluntly. Stratification in the opposite direction is rare, is perceived more favourably by occupants, and is not addressed by the standard at all.

Cooling reliably produces the direction people mind. Cold air is denser, so it settles low while the warm band stays high, which leaves the head in warmer air than the ankles. Producing the reverse takes a warm floor, which is why the standards treat it as an unusual case.

The requirement was written for exactly this occupant. ASHRAE 55 applies its local discomfort limits to people with clothing insulation below 0.7 clo and metabolic rates below 1.3 met. It calls that pairing lightly clothed and near-sedentary. A household sitting in a flat on a warm evening fits it without trying.

No single figure defines how much layering is too much. ASHRAE 55 sets the limit as an air temperature gradient between head and ankle level, and that limit moves with whole-body thermal sensation. The same layering is judged differently depending on how warm or cool the person already feels. The limit itself is drawn where at most five percent of a room would be dissatisfied.

Mixing is what removes a gradient, and a split system has only one route to it. The supply must reach the lower band and fold it into the upper one. Whether it does is a property of the room and of the mounting, and it turns on why cold air runs along the ceiling.

An overhead system is designed to prevent layering, not to tolerate it. ASHRAE sorts room air distribution into mixed, fully stratified, and partially mixed types, and places overhead air distribution in the mixed group, defined as producing little or no thermal stratification within the space. A wall split in a Singapore flat belongs to that group by design.

So persistent bands in an ordinary flat say the mixing has stopped, not that layering is normal for this equipment. Usually it stopped because somebody turned it off. Quiet settings, sleep modes and draught-free modes all reduce air movement, and reduced air movement is the exact condition layers need to survive.

The standards treat moving air as changing the question. ASHRAE 55 exempts a space from its vertical gradient requirement where elevated air speed is being used to deliver comfort. Air crossing the occupants is a different route to the same result, and the gradient limit was written for still conditions.

In a tall space this stops being a servicing question and becomes a design one. ASHRAE treats stratification height, the level at which the upper mixed zone begins, as a design quantity in stratified and partially mixed systems. A wall split gives nobody a way to place it. Given enough volume above the occupied zone, one high-level outlet cannot turn the room over, and bands form regardless of what the machine does.

Say that out loud before money is spent on it. A chemical wash does not change the height of a room. Neither does more capacity, and neither does gas. Where a quote answers a seated-height comfort complaint with any of the three, ask what was measured at seated height beforehand. If nothing was, the quote is answering a question nobody asked.