What the table is doing

Heat stress is a balance. Work produces heat inside the body; the environment decides how fast that heat can leave. When production outruns removal, core temperature climbs. Rest is how the balance gets settled: it cuts production to near zero for part of each hour.

So the table has two axes. WBGT says how bad the environment is at removing heat, and workload says how much heat you are making. Where they cross, you get the share of each hour that can be work.

Read the answer as a ceiling, not a target. Nothing about 30 minutes on and 30 minutes off makes a worker safe if they arrived dehydrated, are on their second day in the heat, or are wearing a vapour-barrier suit.

The four things that make the real limit stricter

  1. Not acclimatised. NIOSH publishes a separate, lower set of limits for unacclimatised workers. In the first week, treat the conditions as several degrees worse than measured. The acclimatisation plan builder sets out the schedule.
  2. Clothing. Coveralls, and especially anything impermeable, add to the effective WBGT before you enter the table.
  3. Dehydration or illness. The tables assume a worker starting the hour in normal fluid balance.
  4. Direct sun with no wind. Already in WBGT if you measured it properly — but if you estimated it, that is exactly the case where an estimate reads low. The WBGT estimator flags it.

What this cannot be used for

This is a presentation of published guidance, and it is not a compliance determination. If you need a record that stands up to an inspection, you need a measurement from a certified instrument and an assessment made under whichever rules apply to you — the general duty clause and state plans in the US, the 2025 regulation in Japan, the risk-assessment duty in the UK and Australia. App or WBGT meter is the honest version of that comparison.

For the reasoning behind the schedule rather than the numbers, work/rest schedules in heat goes through workload categories, break design and what to do when the schedule says stop.