Tap anywhere outside image to close
Shell insulation is the item most argued over and least examined. This page sets out where a heated pool actually loses its heat, what the research says against what manufacturers claim, and — if insulation is justified — the only method that works.
Pool insulation is a thermal barrier placed between the heated water body and everything colder that surrounds it. In practice this means two separate things that are routinely confused with one another:
Shell insulation — a rigid or sprayed layer on the outside of the concrete structure, between the shell and the surrounding soil. It addresses conduction: heat travelling out through the floor and walls into the ground.
Surface insulation — a cover over the water. It addresses evaporation, convection and radiation: heat leaving upward from the water surface into the air and the night sky.
They are not alternatives. They act on entirely different loss paths, and, as the next section shows, those paths are not remotely equal in size.
A thermal barrier only works if it physically stands in the path the heat is taking. For a shell, that path runs from concrete into soil — so the barrier belongs between the concrete and the soil, on the outside face. A thermos flask is insulated on the outside of the container, never on the inside. Insulation cannot block a route it is not standing in.
Before deciding whether to insulate a shell, it is worth knowing what proportion of the problem a shell represents.
| Loss Mechanism | Share of Surface Loss | What Stops It |
|---|---|---|
| Evaporation — every litre that leaves as vapour carries its latent heat with it | 68% | A cover. Nothing else comes close. |
| Re-radiation — the water surface radiating long-wave heat to the open sky, worst on clear nights | 17% | A cover. |
| Convection — moving air stripping the warm boundary layer off the surface | 15% | A cover, and a windbreak. |
| Conduction — heat passing out through the floor and walls into surrounding soil | not in this split | Shell insulation. Unaffected by a cover. |
The 68 / 17 / 15 split is the widely cited breakdown published by the Florida Solar Energy Center. Read it carefully: those three figures sum to 100% because they describe the losses occurring at the water surface. Conduction through the shell is not one of the three and is not separately quantified in that source — on the basis of which it is generally treated as small for a below-grade pool in a mild climate.
Two honest observations follow. First, the surface is overwhelmingly where the heat goes, and any specification that ignores it is arguing about the smaller half of the problem. Second, the split was derived for a temperate outdoor pool — it is not a universal constant, and the conduction share rises as the ground gets colder relative to the water. Both of those matter for what comes next.
The literature does not agree, and it is worth showing exactly who is saying what, and on what basis.
If evaporation, radiation and convection account for essentially all of a pool's heat loss, then conduction through the walls and floor of a below-grade pool is a minor term. On that reasoning, additional shell insulation frequently does not repay its cost.
Basis: independent published research on outdoor pool heat loss. Test conditions are stated. Derived largely from temperate-climate conditions.
Insulation suppliers — predominantly European and cold-climate — market XPS basin insulation as cutting heat loss by up to 80%, and a 50 mm PU spray-foam layer as reducing energy cost by 60–70%.
Basis: supplier marketing material. No test conditions are stated behind either figure. One of the same manufacturers qualifies its own claim by noting that wall insulation only delivers when paired with a good cover.
We do not repeat the 80% and 60–70% figures as fact anywhere in a specification, and neither should a consultant. A percentage with no stated test condition is a sales number, not an engineering one. It is quoted here only so that a reader who meets it in a supplier submittal recognises where it comes from.
Shell insulation matters considerably more in a genuinely cold region than in a mild one, for three reasons that have nothing to do with marketing:
For a heated pool in the plains, or a pool that is not heated at all, the research position broadly holds: the money is better spent elsewhere. For a heated pool in the Shimla–Kasauli–Mussoorie band or colder, shell insulation is defensible on its own terms.
But even where it is justified, shell insulation addresses the smaller half of the problem. A specification that insulates the shell and says nothing about a cover is incomplete, whatever it costs.
50 mm XPS board wrapped outside the shell — under the floor and up all four walls, before backfilling.
Under-raft boards carry structural load and must be specified by grade, not just by thickness. 300 kPa and 500 kPa are the commercial grades sold for under-slab pool use. A board called up only by thickness and thermal conductivity is under-specified for load, and the specification is incomplete as written.
Indicative product figures for reference: a commercial pool-grade XPS board at 80 mm gives a thermal conductivity of 0.035 W/mK and an R-value of 2.29 m²K/W.
| Material | Verdict | Notes |
|---|---|---|
| XPS board | The default | Closed-cell, load-bearing, moisture-stable. Specify grade as well as thickness. |
| Closed-cell PU / PUR spray foam | Equal, situational | Sprayed on the outside of the shell for a seamless, sealed layer. Best where boards don't fit — irregular shapes, curves, retrofits. Conductivity around 0.022 W/mK, better than XPS. One worked example takes a bare concrete wall from roughly 170 to 0.6 W/m²K with a 50 mm layer. |
| EPS | Avoid below grade | Loses insulating value once it absorbs moisture, and is too soft under a pool floor. |
| Foamglas | Specialist | Performs better where groundwater is heavy, but costs far more. |
| Polyiso / PIR | Wrong application | Belongs to the building envelope of an indoor pool hall, not the shell. Natatorium roofs typically need 150–200 mm for R-38 to R-49 continuous. |
Hill-station locations in the Shimla–Kasauli–Mussoorie band are cold, but not months below freezing. At that level 50 mm is the accepted balance of cost and performance for luxury heated pools.
75–100 mm belongs to Manali or Kashmir conditions. Below that, the extra thickness is cost without matching return. Specifying more does not make a specification more rigorous — only more expensive.
Any break at the raft-to-wall junction becomes a continuous escape path. A thermal envelope with a gap in it is not a thermal envelope — it is a detour.
Without a protection layer they are damaged by backfill before the pool is ever filled, and nobody will ever see it. This is the failure that never gets diagnosed.
The version that appears most often: insulation applied once the shell is already built, from the inside, finished with a resin coating. It reads as insulation on paper. It does almost nothing in the water.
The concrete shell still touches cold soil directly, so heat escapes on exactly the path it did before. The barrier is not standing in the route the heat takes.
Anything on the water side sits behind tiles in permanent damp. Damp insulation stops insulating — and there is no maintenance regime that can dry it out.
A few millimetres of coating cannot do the work of a 50 mm board. It is a waterproofing product being asked to perform as a thermal one, and the two are not interchangeable.
A soft layer between concrete and tile adhesive introduces a debonding risk for the mosaic later. The specification does not merely fail to help — it adds a defect path.
Net result: the client pays for insulation, and the heating bill does not change.
Look at where the layer sits in the sequence: below the raft and behind the walls, laid before the concrete is poured and before the soil goes back. Once the shell is cast and backfilled, that face is sealed away permanently.
There is no later stage at which shell insulation can be added properly. Not during finishing, not at handover, not during a refurbishment. It is one of the very few items on a pool that genuinely cannot be revisited — which is exactly why the after-the-fact versions of it exist, and exactly why they do not perform. Once the correct moment has passed, the only thing left to sell is something that looks like insulation.
So the decision belongs at design stage, alongside waterproofing and the structure — not to the fit-out conversation that happens months later. Penetrations for pipes and lights have to be coordinated through the layer while it is going down, and the whole envelope inspected before backfill, because after that nobody will ever see it again.
On an infinity or overflow pool, heated water also sits in the balancing tank at all times. An uninsulated tank quietly gives away the heat the shell insulation was installed to save. Tank walls and lid should be insulated to the same standard, at the same stage, for the same reason.
Indoor pool air runs a dew point around 21°C, with interior air typically 20–24°C. Any surface inside a wall or roof cavity colder than about 21°C will condense, and the structure rots over time.
Shielding the pool from wind reduces both evaporation and convection — together 83% of surface loss — at no ongoing cost, and without blocking the sunlight that helps warm the water. It is described in the source literature as usually highly cost-effective. It is also free at siting stage and expensive at every stage after, which is exactly when it tends to be raised.
Around 68% of a pool's heat loss is surface evaporation, and a further 32% is radiation and convection — also from the surface. The cover is the only measure that acts on all three at once, and it is the item most often absent from a specification that argues at length about shell insulation.
| Cover Type | Heat Retention | Water Saving | Lifespan & Trade-offs |
|---|---|---|---|
| Opaque foam cover | Best available — up to R-12, though most commercial products only reach R-1 to R-2 | Virtually eliminates evaporation | Heavy, expensive, blocks sunlight |
| Thermal blanket | Twice as effective at retaining heat as a bubble/solar blanket | Not separately quantified | 8–10 years |
| Solar / bubble blanket | Ineffective at retaining heat overnight | ~90% reduction in water loss | A few years |
| Polycarbonate automatic slat cover | More heat-resistant and durable than blankets | ~30% of annual water bill | Most durable, most expensive |
Some cold-climate building codes already treat this as non-optional: heated pools above 32°C are required to carry a cover rated at a minimum of R-12.
Thermal blanket or automatic polycarbonate slats. The largest single win, acting on the 68 / 17 / 15 losses simultaneously — and the step most often skipped.
A no-cost addition if decided early. Reduces evaporation and convection without blocking sun.
Structural and thermal at once — it handles frost and ground moisture as well as conduction. Justified in genuinely cold regions; harder to justify in mild ones.
On overflow and infinity pools. Without it, part of the heated volume sits permanently outside the insulated envelope.
Not optional where there is a roof. Insulation without it damages the building it sits in.
Correctly specified external insulation and an internally applied resin coating are not the same work, and cannot be compared line-for-line, however similar the two lines look on a page. Matching the flawed version simply to keep two quotations comparable means paying for something that leaves the heating bill exactly where it was.
Insulate the shell where the climate justifies it. Cover the water everywhere.