High CO₂ in a flat: what the meter is really telling you

The meter is counting people, not pollution

Apart from one exception worth knowing about, nothing in a home produces carbon dioxide in any quantity except the people inside it. Not the furniture, not the paint, not the cleaning cupboard. The exception is unflued gas combustion: a gas hob adds CO₂ of its own while it burns, which is one more reason the building code asks a kitchen for a specific extraction provision for cooking fumes on top of the dwelling's general ventilation.

Set the hob aside and the measurement becomes simple. People emit CO₂ continuously and fairly predictably while they are in a room — not in bursts, the way a shower releases steam, and not in a plume, the way a frying pan releases everything else. With a source that steady, the concentration in a room stops depending on the source and comes to depend on the one remaining variable: how quickly the air is being replaced. The device sold to you as an air quality monitor is, in a flat, a renewal gauge.

That is a narrow job, and accepting the narrowness early saves a lot of misreading later. The sensor answers one question well — how much of the air you are breathing has already been through somebody's lungs — and everything else it appears to say is your inference, not its measurement.

There is still a good reason to own one, given that your nose reports on the same thing for nothing. The nose only works on arrival, and it adapts within minutes. The sensor never adapts. That is the whole difference, and it is why the person who has spent the night behind a closed door notices nothing at all while the person who opens that door at eight in the morning notices immediately.

A number with no slope in it cannot be read

Picture a reading taken at random: a high value on the display. Is the room filling up, hour by hour, with the family in it? Or is that the same value on its way down, ten minutes after somebody opened the balcony door? Identical digits, opposite situations, opposite responses. An instantaneous value has no direction in it, which is most of what you need.

What carries information is the shape over hours. Monitoring work in occupied apartments describes exactly this: concentrations that follow occupancy, climbing while people are in a room and falling away when the flat empties or is opened up. Three features of that shape are worth learning, and each answers something different.

• How steeply it climbs once a room is occupied and shut. This mixes two things at once — how many people there are per cubic metre of room, and how much renewal is arriving — so it is meaningful against the same room on another day and close to meaningless against a different room.
• Where it settles, if it settles at all. A curve that flattens out means renewal has drawn level with the occupants, if only just. A curve that never flattens means it has not.
• How fast it comes back down. The cleanest of the three, and worth a section of its own.

If you already have threshold numbers in your head from somewhere else, a word about them. Such figures are easy to find, but each one comes from a document written for a particular setting — a workplace, a classroom, guidance issued in another country — and it carries the conditions of that document with it. This page quotes none of them on purpose. What your own flat can give you, and no threshold can, is a comparison of itself against itself.

The one measurement that is not ambiguous

When everybody leaves the room, or when you open up, the source disappears. From that moment the curve is no longer mixing two things: it is measuring renewal and nothing else. That makes the fall the most informative segment you will ever record, and the easiest to run deliberately.

1. Read the outdoor value first, on the balcony or in the street, and let the meter settle for a few minutes. Every indoor figure is a rise above that, not an absolute.
2. Let a room reach a high, roughly steady value with people in it and everything shut the way you normally shut it.
3. Change one thing, and note the time. One window. Or one internal door.
4. Time how long the value takes to come back near your outdoor baseline, and write down which windows and which doors were open while it did.
5. Repeat on another day with a different combination.

The interpretation is usually direct. A slow fall with a window wide open generally means air is going in and out through the same opening, because there is no route for it to take. Opening a second opening on a different orientation — or simply opening an internal door — often shortens the fall more than opening the first window wider ever will. In a flat with openings on a single façade there is no cross ventilation in the sense general advice assumes, and the only path available runs through the inside of the dwelling.

Where the window gives onto an interior light well — patio de luces in Spanish, pati de llum in Catalan, the narrow shaft shared with the flats above and below — the same gesture renews less, because air in a shaft moves far less than air on a street front. If your flat has openings on two orientations, take the outdoor baseline at both. There is no reason for the shaft and the street to give you the same number, and knowing which is which decides what you open and when.

The room that produces the CO₂ has no way of removing it

DB-HS3 — Documento Básico HS 3, Calidad del aire interior, the section of the Spanish building code dealing with indoor air quality — sorts a dwelling into rooms where outdoor air is admitted, meaning bedrooms and living rooms, and rooms from which air is extracted, meaning bathrooms and kitchens, with transfer openings in between. The direction is fixed: admission room, circulation space, wet room, outside.

Put people on that map and the consequence is awkward. They sit, work and sleep at the admission end, the end where nothing extracts anything. Every breath they exhale has to cross the dwelling to reach an exit. A door with no gap beneath it and no grille in the leaf cuts that route in half, and a CO₂ log demonstrates it better than any argument: run the same room twice in comparable conditions and change only the door. The distance between the two traces is the value of the internal air path, measured in your own flat.

Some vocabulary, because you will meet it on paperwork and from tradespeople. Airflow rates are caudales. A mechanical system is VMC, ventilación mecánica controlada. A shared vertical duct climbing through a block is a shunt — what happens inside one affects every flat on the stack, which is why it is not something to open or connect into on a private initiative.

The clearest trace you are likely to record comes from a bedroom shut overnight: a long uninterrupted climb that stops only when somebody opens the door in the morning. It reads so cleanly because nothing changes for hours — same people, same door, same window. Published reviews of the bedroom ventilation literature connect overnight air renewal and CO₂ accumulation with sleep-related outcomes. What to do about a curve like that, given noise, security and a cold draught landing on the pillow, is a question with its own answer; the log only tells you whether you have that question.

Four things the sensor will never see

• Moisture. A bathroom two minutes after a shower can read low CO₂ with the air saturated. Condensation depends on how much moisture the household produces, how much renewal removes it, and how cold the coldest surfaces are — and none of those three appears in a CO₂ trace. If your device also shows relative humidity, that is a second reading of a second problem, not confirmation of the first.
• Particles. Frying, a candle, a cigarette, traffic coming in off the street. None of it moves the CO₂ line. Throwing the windows open onto a busy road can lower CO₂ and raise particles in the same gesture.
• What the materials give off. New furniture, paint, varnish, adhesives and cleaning products emit with nobody home. A flat shut up for a fortnight can read essentially outdoor CO₂ and still smell of solvent when you walk in.
• The quality of what you let in. The meter records that you renewed. It passes no judgement on what you renewed with.

One assumption is better killed with the meter than with words: that air conditioning brings fresh air into a room. A domestic split unit recirculates the air already in the room: only refrigerant travels between the indoor and outdoor sections, so the appliance does not, by itself, bring in any outdoor air. An air purifier filters what is already there and admits nothing either. Run either one for an hour with the meter beside it and watch the line refuse to move — one of the more valuable hours a cheap sensor will give you.

So a low reading means recently replaced air, not good air. And a high reading does not describe a specific harm; it describes renewal failing to keep pace with occupancy. It is also why some ventilation equipment modulates its airflow from a humidity signal rather than a CO₂ one — different problem, different sensor.

What the Spanish rules do with CO₂, and what your meter cannot prove

DB-HS3 sets the indoor air quality requirement for dwellings and fixes the route the air takes through them, and its air quality objective is expressed in terms of CO₂ concentration. This page does not reproduce those figures, and the reason is worth understanding rather than resenting. The code's criteria come with reference conditions and a method of verification that a desktop sensor standing on a shelf does not reproduce. They are a requirement placed on the design and on the building's ventilation system, not a test a household has to pass with a portable gadget. Anyone assessing an installation technically measures caudales — airflow rates at the openings — not the value on your bookcase.

RITE, Real Decreto 1027/2007 as amended, governs fixed thermal installations in buildings, ventilation among them, with the design, maintenance and control requirements that go with them. It is the framework in which regulating airflow according to demand, instead of running permanently at one rate, makes sense — and CO₂ is one of the signals that allows it, precisely because it tracks occupancy.

One thing to leave at the airport, if you have moved here from elsewhere: the ventilation rules of the country you came from. Guidance figures and duct requirements from another jurisdiction do not transfer to a flat in Catalonia, and quoting them at a Spanish technician will not advance the conversation. The document that governs this dwelling is DB-HS3.

The practical upshot is modest and freeing. The meter is for understanding your own home and comparing it against itself. It is not an instrument of verification and not an argument to wave at anybody.

Buying one, and putting it somewhere it tells the truth

The real value of an inexpensive meter is not its accuracy. It is that it turns something invisible into something a household can look at together, and that changes the conversation: agreeing to leave a bedroom door ajar is far easier once everybody has seen last night's curve.

Four things decide whether the one you buy will be any use.

• Which sensor is inside. Non-dispersive infrared, NDIR, measures CO₂. Other devices estimate it from volatile organic compounds and label the result eCO₂ or "CO₂ equivalent". That second number responds to cleaning spray, alcohol and food, and can climb steeply without a single person entering the room. If the specification does not say NDIR, assume you are looking at an estimate.
• Whether it logs. A device that only displays the current figure requires you to be standing in front of it, and what you want to see happens while you are asleep or out. Look for a graph, a history or an export, even if it lives on your phone.
• How it calibrates. Automatic baseline calibration assumes the unit periodically sees outdoor air. In a flat that is kept shut, that assumption can drift. Take it out to the balcony from time to time, and always read against your own outdoor measurement rather than trusting the absolute figure.
• Where it stands. Not on the sill of an open window, where it measures the street. Not an arm's length from your face, where it measures your breath. Not in a draught and not above a radiator. Put it at the height where people actually are, sitting or lying down, away from any extract grille, and give it a few minutes before believing the first values.

When the curves start repeating themselves

Sooner or later the readings stop surprising anybody. The shut room climbs. Opening up brings it down. The closed door holds it high. At that point the meter has said what it knows, because the question has changed underneath it: it is no longer how much renewal there is, it is where the route breaks.

That second question begins with looking. Does air have a way in, a way through from room to room, and a way out — a bathroom or kitchen extract that genuinely pulls rather than merely making a noise? A thin sheet of paper held against the grille can screen for suction at that moment, but it does not measure airflow or rule out duct and pressure problems. And there are flats where the route does not exist in the form the code assumes: a single façade, the only bedroom window opening onto a light well, or a refurbishment and much more airtight windows that removed the uncontrolled infiltration the dwelling had quietly been relying on. No quantity of measurement will create a route that was never built.

Where the sensor earns its keep again is the day something changes — a grille cut into a door, an undercut cleared, an extract repaired, new windows fitted. Then you have the thing that is almost never available when a home is altered: a recorded before to hold the after against. That is worth considerably more than whatever it happens to be displaying today.