The honest comparison
I make the app, so take the framing with that in mind — but the conclusion here is not the one that sells apps.
| Estimated from weather data | Certified instrument | |
|---|---|---|
| Radiation | Modelled from cloud, humidity and sun position | Measured by a black globe |
| Your microclimate | Invisible to it | Measured where you stand |
| Compliance record | No | Yes, that is the point |
| Forecast | Yes — it can tell you about this afternoon | No, it reads now |
| Cost | Nothing | Hundreds upwards |
| Always with you | Yes | If someone remembered it |
Where the estimate genuinely fails
Not vaguely — specifically:
- Hot, dry, still conditions. The analytic method underestimates the natural wet bulb by up to 2°C there. That is the desert afternoon, and the estimator flags it when you are in it.
- Enclosed or reflective microclimates. A trench, a courtyard, a plant room, a stadium bowl, a black membrane roof, a spot beside a south-facing glass wall. A weather station three kilometres away knows nothing about any of them, and the differences are large.
- Indoors. Weather data is outdoor data. There is no honest indoor estimate from it, which is why this app makes no indoor claims and why the indoor heat page starts by saying so.
- Anywhere a number becomes evidence. The moment a value goes into a record, its provenance matters more than its accuracy.
Where the estimate is genuinely useful
- Deciding whether the day needs measuring at all. Most days are not close to the threshold, and knowing which ones are is most of the value.
- Planning tomorrow. No instrument forecasts.
- Awareness for people who will never own a meter. A runner, a parent, someone walking a dog, a small crew with no safety budget. The alternative is not an instrument; it is nothing.
- Sanity-checking a reading. If the meter says 21°C and the estimate says 30°C, something is wrong with one of them — often a globe in shade or a dry wick.
If you are buying one
Three checks that matter more than the brand:
- Does it have an actual globe? A device with no black sphere is computing an index, not measuring WBGT. Several inexpensive "heat stress" meters are in this category.
- Does it state conformance to the standard your rules cite? ISO 7243 generally; JIS Z 8504 if you are working under the Japanese workplace heat rules.
- Is the wick maintained? A dry or dirty wick is the most common cause of a wrong reading in the field, and it always reads high — which looks conservative and is not, because it destroys trust in the instrument the first time someone checks it against a thermometer.