Bather load: how occupancy data can support lower pool resource use

Indoor pools are in a class of their own. A dynamic simulation study of German facilities presented at BauSIM 2022 puts average annual consumption at roughly 3,400 kWh per m² of water area for sport-oriented pools and 5,900 kWh/m² for leisure-oriented ones — against a few hundred kWh per m² of floor area for a typical office building or sports hall.

Even exemplary buildings remain heavy consumers. Monitoring of the Lippe-Bad in Lünen, the first indoor pool built to Passive House standard, recorded about 1,189 kWh per m² of water area for heat and 718 kWh/m² for electricity — best-in-class figures that are still an order of magnitude above ordinary public buildings.

This is no longer only a budget question. Under Directive (EU) 2023/1791, the recast Energy Efficiency Directive, member states must ensure that the combined final energy consumption of public bodies falls by at least 1.9% every year against a 2021 baseline, with transposition due by 11 October 2025. For municipalities with energy-intensive aquatic facilities, pools can therefore be an important part of that reduction pathway.

German water treatment practice makes the dependency explicit. In DIN 19643, the rated load — *Nennbelastung* — is expressed as bathers per hour, and the required treatment volume flow follows directly from it: divide the rated load by a capacity factor that depends on the filtration process. A 25 × 16.7 m non-swimmer basin with fixed-bed filtration, for example, works out at a treatment flow of some 308 m³/h.

The same logic governs water consumption. Public pools must continuously add fresh water to dilute what bathers bring in, and German practice takes roughly 30 litres of fresh water per bather as the reference figure. Not per hour, not per cubic metre of basin: per person.

DIN 19643 links treatment-flow calculations to rated bather load, not pool volume alone. Without reliable current occupancy data, operators may rely on conservative design values or fixed schedules that do not reflect changing use.

The research literature calls it anthropogenic pollutant release, and Keuten and colleagues, publishing in *Water Research*, divided it into three parts: the initial release rinsed off the body in the first seconds of immersion, the continual release during swimming, and incidental release from urine and other excreta.

The quantities are not trivial. Reviews of the literature report urea inputs of roughly 0.56 to 2.46 grams per bather, together with 2–85 mg of creatinine, 30–60 mg of ammonium salts and 15–50 mg of amino acids. The German Federal Environment Agency (Umweltbundesamt) works with an assumption of 25 to 30 ml of urine per bather, and notes that skin rinse-off alone contributes on average about 0.16 g of urea per person.

Nor is the release rate fixed. Keuten's measurements show sweat production rising from 0.1–0.2 to 0.8 litres per m² of skin per hour as water temperature increases from 29 to 35 °C. Pollutant input is therefore not a simple headcount multiplier: it scales with how many people are in the water, how warm that water is, and how hard those people are working.

Several downstream demands — including chlorine, filtration and dilution — change with the load introduced by bathers.

Urea and the other nitrogen compounds react with free chlorine to form chloramines. The most volatile of them, trichloramine (NCl₃), off-gasses into the hall: it is the source of the characteristic “chlorine smell”, and a recognised respiratory irritant for swimmers and, above all, for staff who spend entire shifts poolside. The WHO guideline value for trichloramine in pool air is 0.5 mg/m³; a Belgian research group has proposed a stricter 0.3 mg/m³.

Here the research produces a finding with blunt operational consequences. Work on trichloramine formation and mass transfer, published in *Water Research*, found that air concentrations depend strongly on air circulation and on water agitation by bathers — and are not correlated with combined chlorine measured in the pool water, because ventilation is the predominant parameter.

In other words, a water probe alone cannot manage natatorium air quality. Occupancy and surface activity can provide additional context where camera position, visibility, data quality and the approved purpose make those indicators reliable.

Occupancy and activity also affect one of the building’s largest energy flows. As the U.S. Department of Energy puts it, evaporation is by far the largest source of energy loss from a pool: every pound of 80 °F water that evaporates removes roughly 1,048 Btu from the water body. In an indoor pool, that becomes a moisture load the air-handling plant must remove.

Evaporation is not constant. The ASHRAE Handbook — HVAC Applications chapter on natatoriums scales the evaporation rate by an activity factor, precisely because water agitation and splashing during use raise it above the still-water baseline; unoccupied hours are conventionally computed at a factor of 0.5. Field measurements published in *E3S Web of Conferences* in 2022 confirmed the mechanism and quantified it: the number of bathers had a large influence on the evaporation rate, with measured activity factors of 0.7–0.8 in occupied pools (peaking at 1.1) against 0.50–0.57 when empty.

Between an empty basin and a busy one, evaporative load can rise substantially. A ventilation system run at the same setting throughout the day may use more energy than necessary when the hall is empty and respond poorly when activity peaks.

The counterpart is equally well documented. Covering a pool when it is not in use is the single most effective way to cut heating costs, worth 50–70% according to the U.S. Department of Energy, and preventing up to 95% of evaporation. For indoor pools the agency notes a second benefit: a cover also reduces the need to ventilate and lets exhaust fans be shut down — savings that only materialise if you can be certain the basin is genuinely empty.

Many traditional pool instruments measure the effects of changing bather load rather than bather load itself.

Humidity may rise after evaporation increases. Redox or free-chlorine readings may shift after pollutants enter and react. Turnstiles count building entries, but not necessarily how many people are in each basin at a given moment.

These feedback signals remain essential. Reliable occupancy data can add a leading indicator, helping the plant anticipate demand instead of responding only after conditions change.

A pool-safety project defines the zones that need to be observed and the conditions to be tested. Where there is a separate, approved purpose, the same infrastructure can also produce aggregated indicators about occupancy and use of those areas without requiring facial recognition or persistent identification.

Aqutron Vision can use cameras only or add spatial sensors based on patented 3D technologies. Safety-critical processing is designed to run locally. Coverage, accuracy, update frequency and alert paths are established for the individual facility rather than assumed from a generic specification.

When data quality and governance allow it, occupancy trends, peak times and area use can be shared with S.P.A.C.E. — Sensing Platform for Automation, Control & Execution to compare plant behaviour with actual demand. That secondary use requires its own purpose, access rules, retention choices and acceptance criteria.

Shared infrastructure may improve the business case, but safety and energy performance must still be assessed separately. Any efficiency benefit should be measured against a site-specific baseline; it should not be assumed to pay for the safety system before the data has been verified.

With a trustworthy live occupancy figure, a series of ordinary building-services strategies suddenly become available to a pool:

These are established building-control strategies. In pools, their use depends on a reliable occupancy signal, appropriate safeguards and operating rules that protect water and air quality, hygiene and safety.

This needs stating plainly, because it is where such systems can go wrong. Energy optimisation must operate strictly inside safety and hygiene limits, never against them.

Water treatment standards define a floor, not a target to be negotiated downwards: circulation and disinfection may follow real load only above the regulatory minimum, and never below it. Air quality is the same — the WHO trichloramine guideline is a ceiling that no efficiency measure may push against, and reducing ventilation to save energy while chloramines accumulate is not efficiency, it is a hazard to staff.

The detection function must also be protected. The site study defines and tests how changes to lighting or air treatment interact with the conditions required for detection. Independent interlocks and fallback rules give safety constraints priority over efficiency commands.

Counting occupancy need not require facial recognition or persistent identification. Where a separate purpose is approved and the function is configured accordingly, local processing can provide the building system with aggregate counts or load estimates rather than identities or biometric templates.

Aggregation can limit the data used downstream, but it does not automatically make the source images or processing anonymous. Purpose, legal basis, impact assessment, retention and access remain decisions for the data controller in the context of the installed system.

A saving you did not baseline is a saving you cannot claim. Before commissioning any load-driven control strategy, sub-meter the circuits that matter — pool water heating, air handling, circulation pumps — put a meter on the make-up water line, and log chemical consumption. Establish a reference period, then measure against it.

For public bodies in particular, reporting requirements make a documented reference period and measurement method valuable. Traceable data can show where consumption changed and under which operating conditions.

Occupancy is one useful input to energy, water and chemical-demand analysis. Where this secondary function is authorized, configured and shown to produce reliable indicators at the site, the data can support efficiency decisions. Its economic value must still be measured against the baseline rather than assumed in advance.