Heat Index Calculator
Result
Heat index (°C)
- Heat index (°F)
- 105.2 °F
The heat index calculator answers the question a thermometer cannot: how hot the air actually feels once relative humidity is counted in. It combines air temperature with relative humidity into a single apparent temperature — the number the US National Weather Service (NWS) publishes as the heat index. At 35 °C and 50% humidity the result is 40.7 °C, nearly six degrees above the air temperature, because moist air slows the evaporation of sweat, which is how the body sheds most of its heat. Dry air does the opposite: 40 °C at 10% humidity reads 36.7 °C, below the air temperature. The page runs the full NWS formula, including both of its corrections, and prints the Celsius and Fahrenheit values together.
Heat index bands and what each one means
| Heat index | What it means |
|---|---|
| Below 80 °F (26.7 °C) | The regression behind the heat index is fitted from 80 °F upward, so below that line the index is not defined in a useful sense — this page reports the simplified form and says so rather than leaving the badge blank. It is not a safe/unsafe statement: mild air with high humidity still costs you something, and 79 °F at 90% humidity is muggy enough to matter. When a hot day's forecast sits just under the line, the useful question is not whether the index applies but whether the work is heavy and the shade is available. |
| 80 to 89.9 °F (26.7 to 32.2 °C) | From here, prolonged exposure and physical activity start to tell. The body can still keep up, but it is working harder to do it: sweating is doing the whole job, and the fluid and salt it costs have to be replaced. Fatigue is the ordinary outcome of a long afternoon in this band. This is also where acclimatisation pays off most — someone in their second week of hot weather handles it noticeably better than someone in their first, on identical inputs. |
| 90 to 102.9 °F (32.2 to 39.4 °C) | Heat cramps and heat exhaustion become possible with prolonged exposure, and possible during activity. This is the band where the shading and hydration rules stop being advice and start being what prevents an incident: fluids before thirst, shade at the middle of the day, and no long stretches of hard work in the sun. Exhaustion shows up as heavy sweating with weakness, dizziness, nausea or a headache, and the response is to stop, cool down and rehydrate — heat exhaustion is the body saying it is losing, well before it actually loses. |
| 103 to 124.9 °F (39.4 to 51.7 °C) | Heat exhaustion is likely with continued exposure, and heat stroke becomes possible. The important word is possible, not certain: at this load it depends on duration, workload and the person, which is exactly why the band is treated as a stopping condition rather than a caution. Outdoor work, organised sport and long errands should move to the early morning or be called off. The distinction to hold on to is that heat exhaustion and heat stroke are different emergencies — the first person is sweating heavily and unwell, the second may have stopped sweating and be confused, and only the second is a call-emergency-services situation. |
| 125 °F (51.7 °C) and above | Heat stroke is expected rather than possible, and it is the band where this index stops describing discomfort and starts describing a medical emergency. At this apparent load the body's cooling capacity is exceeded faster than any behaviour can compensate for, and core temperature rises on its own. There is no workload small enough to make it reasonable and no amount of acclimatisation that makes it safe. The band is rare, which is the only encouraging thing about it — and it is reached at air temperatures that are entirely ordinary when the humidity is high enough, which is why the index exists at all. |
The four NWS bands plus the band below them that this page adds, all computed from the equation rather than copied from the chart. The boundaries are half-open and are read off the value the panel prints, which is rounded to one decimal, so the last value inside a band is 89.9 / 102.9 / 124.9 rather than a round number. Both scales are given because the lines are NWS's and therefore Fahrenheit — 103 °F is exactly 103 while its Celsius reading is a repeating decimal — and because the Celsius numbers are what most readers will act on. The first row answers the question the page gets most: why a mild, sticky day shows a badge that says the index does not apply, instead of showing nothing.
Heat index chart (°C)
| °C | °F | RH 40% | RH 50% | RH 60% | RH 70% | RH 80% | RH 90% |
|---|---|---|---|---|---|---|---|
| 26.7 | 80 | 26.4 | 27.1 | 27.7 | 28.3 | 29 | 30.2 |
| 29.4 | 85 | 29.1 | 30.3 | 31.8 | 33.7 | 36 | 38.8 |
| 32.2 | 90 | 32.6 | 34.8 | 37.6 | 41.1 | 45.2 | 49.9 |
| 35 | 95 | 37.2 | 40.7 | 45.1 | 50.3 | 56.5 | 63.7 |
| 37.8 | 100 | 42.9 | 48 | 54.2 | 61.5 | 70.1 | 79.8 |
| 40.6 | 105 | 49.7 | 56.6 | 64.9 | 74.7 | 85.8 | 98.4 |
| 43.3 | 110 | 57.6 | 66.6 | 77.4 | 89.7 | 103.8 | 119.4 |
The index at seven air temperatures and six humidities, in the layout NWS prints, with a Celsius column added because the main result here is Celsius. Reading down a column shows how fast the index grows with temperature — at 50% humidity it climbs from 27.1 °C at 26.7 °C of air to 66.6 °C at 43.3 °C of air. Reading across a row shows the humidity effect, which is large and uneven: at 32.2 °C of air, going from 40% to 90% humidity adds 17.3 °C — nearly as much as the 20.9 °C the index gains when the air temperature rises from 26.7 to 37.8 °C at a fixed 50%. Every cell is computed from the same equation as the answer panel, so the two cannot disagree. The 100% column is deliberately absent: at the top of the temperature range the equation leaves its data and returns values that are not readings of anything.
Formula
Step one: HI_simple = 0.5 · [ T + 61 + (T − 68) · 1.2 + RH · 0.094 ], evaluated first. Only if HI_simple ≥ 80 °F, step two — the regression: HI = −42.379 + 2.04901523·T + 10.14333127·RH − 0.22475541·T·RH − 0.00683783·T² − 0.05481717·RH² + 0.00122874·T²·RH + 0.00085282·T·RH² − 0.00000199·T²·RH², followed by two corrections: if RH < 13 and 80 ≤ T ≤ 112, subtract [(13 − RH)/4] · √[(17 − |T − 95|)/17]; if RH > 85 and 80 ≤ T ≤ 87, add [(RH − 85)/10] · [(87 − T)/5]. T and HI are in Fahrenheit, RH is a percentage.
- T
- Air temperature in Fahrenheit, because that is the scale the equation is written in. The field on this page takes Celsius first — most of the world reads a thermometer in Celsius, and making the default unit one the reader has to convert in their head is a needless tax — so the conversion to Fahrenheit happens once at the start and the answer is converted back at the end. The four band lines stay in Fahrenheit for the same reason: they are NWS's, and 103 °F is exactly 103 while its Celsius reading is 39.44.
- RH
- Relative humidity as a percentage, 0 to 100. It is the whole reason this page exists: at the same air temperature, 40% and 90% are different worlds to a body. Sweat cools by evaporating, and evaporation slows as the air around the skin gets closer to saturation, so humid heat costs more than dry heat at the same thermometer reading. Above 100% the number is not merely unusual but meaningless — the equation would happily return a heat index hotter than the air, which is why the input is capped and the computation refuses anything outside the range.
- HI_simple
- The first step: Steadman's simplified form, a weighted average of the air temperature and a humidity term. NWS's instruction is to compute this one first and only fall through to the regression when it reaches 80 °F. The order matters and is easy to get wrong — below that line the regression produces numbers that are too high, so a page that always ran the regression would be quietly wrong in exactly the mild range where nobody checks it. The simplified form is also what gets reported below 80: it is a real answer, not a placeholder.
- HI
- The Rothfusz equation, a nine-term regression fitted to Steadman's full apparent-temperature tables. It is what makes the index more than a rule of thumb: the squared and cross terms encode how the humidity effect grows with temperature, so 5 points of extra humidity cost more at 38 °C than at 27 °C. It is only ever evaluated at or above the 80 °F gate, and if the regression lands back below 80 the simplified form is reported instead — a gap the original NWS wording covers with the phrase about averaging the result with the temperature.
- corrections (two)
- Two adjustments added after the regression, each covering a corner the fit handles badly. Dry heat (RH under 13%, 80 to 112 °F) subtracts a term that grows as the air gets drier, because the regression overstates how bad dry heat feels; it is why 40 °C at 10% humidity can read below the air temperature. Humid heat (RH above 85%, 80 to 87 °F) adds a small term instead, but only in that narrow warm band — higher up the regression is already steep enough on its own. Both are easy to drop without noticing, because dropping one changes a number without moving the band it falls in.
- 80 °F
- The gate, and the lower edge of the index's validity rather than a round number someone liked. NWS's regression was fitted from 80 °F up; below it the index is not defined in any useful sense, and this page still answers — with the simplified form — but says so, which is what the first band in the reference table is for. 80 °F is 26.7 °C, so the gate sits at a perfectly ordinary summer day in Celsius, and a reader typing 26.6 versus 26.7 sees the badge change. That is the boundary being honest, not a rounding bug.
Use it on any day when the thermometer reading and how you feel disagree — the muggy afternoon that is somehow worse than a hotter, drier one, or the dry 40 °C that is survivable in the shade and lethal in the sun. It is the number to quote when planning work, sport or an outdoor event, because the bands are tied to what happens to people rather than to a temperature: the first one starts where the regression does, and the top one is where heat stroke stops being a possibility and starts being the expectation. Two things it deliberately does not do. It is a shade-and-light-wind number, so full sun can add up to 8 °C and a stiff breeze takes some back; and it describes a typical adult, so children, older people and anyone on medication that interferes with sweating run hotter at the same reading. The reference table below is the NWS chart in Celsius and Fahrenheit, computed here rather than copied, so a cell can be read across to the answer panel without either being out of date. For the cold side of the same idea, see the wind chill calculator — it is the mirror image, and it comes from the same agency's work on what weather does to a body.
Worked examples
First screen: 35 °C and 50% humidity
- Air at 35 °C is 95 °F
- Simplified form first: 0.5 · [95 + 61 + 27 · 1.2 + 50 · 0.094] = 96.55 °F, which clears the 80 °F gate
- So run the regression on 95 °F and 50%: 105.22 °F, with neither correction applying — the air is neither very dry nor very humid
- Nothing to add or subtract, so the index is 105.2 °F
- Celsius row: (105.2 − 32) · 5/9 = 40.7 °C
The two steps agree here, which is the point of the gate: well above 80 °F the simplified form is already close, and the regression is what turns close into right. Read the answer as a sentence — the air is 35 °C and your body is being asked to cope with 40.7 °C — and the value of the page appears: 5.7 °C of extra load, from humidity alone, at a temperature that reads as an ordinary hot day. This lands in the danger band, which is also why the first screen shows a badge rather than an empty space where one should be.
The dry-heat correction: 40 °C and 10%
- The air is 40 °C, which is 104 °F — well past the gate
- Simplified form: 0.5 · [104 + 61 + 36 · 1.2 + 10 · 0.094] = 104.57 °F, so the regression runs
- Regression on 104 °F and 10%: 98.58 °F
- Dry correction, because the humidity is under 13%: [(13 − 10)/4] · √[(17 − |104 − 95|)/17] = 0.75 · 0.686 = 0.51 °F
- Subtract it: 98.58 − 0.51 = 98.07 °F, which prints as 98.1 °F; the Celsius row is (98.1 − 32) · 5/9 = 36.7 °C — below the air temperature
The most counterintuitive answer this page gives: the air is 40 °C and the index says 36.7 °C, cooler than the thermometer. It is not a bug and it is not comfort — it is sweat doing its job. In air this dry, water evaporates as fast as the skin can supply it, so the body's cooling mechanism runs at full capacity and the perceived load drops. The correction exists because the raw regression does not know that: it was fitted mostly on humid conditions and overstates dry heat, and the NWS patch subtracts more the drier the air gets. The practical read is the one heat-safety advice repeats — in dry heat, water and shade are enough for a long time; in humid heat, they may not be.
The humid-heat correction: 28 °C and 90%
- The air is 28 °C, which is 82.4 °F — past the gate, and inside the narrow 80 to 87 °F window the humid correction covers
- Simplified form: 0.5 · [82.4 + 61 + 14.4 · 1.2 + 90 · 0.094] = 84.57 °F, so the regression runs
- Regression on 82.4 °F and 90%: 92.75 °F
- Humid correction, because the humidity is over 85%: [(90 − 85)/10] · [(87 − 82.4)/5] = 0.5 · 0.92 = 0.46 °F
- Add it: 92.75 + 0.46 = 93.21 °F, which prints as 93.2 °F; the Celsius row is (93.2 − 32) · 5/9 = 34.0 °C
This one is worth reading next to the previous example, because the two inputs are 12 °C apart in air temperature and the answers are 2.7 °C apart in the other direction. The air is 28 °C — mild by any standard, cool enough that a thermometer alone would say there is nothing to worry about — and it feels like 34 °C, which is the extreme caution band. That is the humidity effect at its purest: at 90% there is almost no room left in the air for sweat to evaporate into, so the body's main cooling channel is throttled while the air is barely warm. It is also why the correction only applies below 87 °F — above that the regression already accounts for humidity, and adding the term again would double-count it.
The heaviest band: 38 °C and 60%
- The air is 38 °C, which is 100.4 °F
- Simplified form: 0.5 · [100.4 + 61 + 32.4 · 1.2 + 60 · 0.094] = 102.96 °F, comfortably past the gate
- Regression on 100.4 °F and 60%: 130.93 °F — neither correction applies, 60% is the humidity the fit was built around
- The index is 130.9 °F, and the Celsius row is (130.9 − 32) · 5/9 = 55.0 °C
55 °C is not a temperature people experience in air; it is what the body is being asked to manage. The air is 38 °C — a serious heatwave day, not an impossible one — and 17 °C of apparent load arrives on top of it from humidity that is merely above average. This is the band NWS treats as the top of its scale, where heat stroke is expected rather than possible, and the number is worth taking literally: at this load, sweating cannot keep pace for long, and the failure mode is not discomfort but a rising core temperature. The advice for this band is short and is not really a calculation: get out of it.
Below the range: 26.6 °C and 50%
- The air is 26.6 °C, which is 79.88 °F
- Simplified form: 0.5 · [79.88 + 61 + 11.88 · 1.2 + 50 · 0.094] = 79.92 °F
- 79.92 °F is below the 80 °F gate, so the regression is never evaluated and the simplified form is the answer
- The index and the air temperature coincide at 79.9 °F → 26.6 °C, a Celsius row equal to the Celsius input
- One tenth of a degree higher, 26.7 °C is 80.06 °F; its simplified form reaches 80.12 °F, the regression runs and returns 80.9 °F, and the badge flips to the lightest warning
Below the gate the simplified form collapses toward the air temperature, and that is the mathematically honest outcome rather than a failure: in mild air, humidity barely changes how hot it feels, so the index has almost nothing to add. The page still prints a number and a band instead of leaving the space empty, which is a decision made for the reader — an absent badge looks like a broken page, while a badge that says the index is not defined this far down is information. The pair of inputs here, 26.6 and 26.7, is the boundary itself: NWS's 80 °F line in the Celsius a reader actually types, where 0.1 °C of thermometer moves the reading from below the range to the lightest warning.
Limitations
The chart stops at 90% humidity on purpose. At 110 °F and 100% the equation returns 278 °F — not an arithmetic error but a regression being extrapolated past the data it was fitted to, and printing a cell like that invites someone to read it as a usable number. The same caution applies at the top of the temperature field: 60 °C is inside the input range because a page should not crash on a number someone types, but the index is meaningless there. The index itself assumes shade and light wind. Full sun adds up to about 8 °C of radiant load that no humidity number captures, and a strong wind removes some heat that the index does not credit — which is why the index is not a substitute for the wet-bulb globe temperature used in occupational heat standards, where sun and workload are inputs rather than caveats. It also describes a typical, acclimatised adult. Children, older people, pregnant readers, outdoor workers in their first week, and anyone taking medication that suppresses sweating or sweating's effect will run hotter at the same reading, and no equation on this page knows any of that. Finally, the index is a guide for planning a day, not a medical instrument: if someone is confused, has stopped sweating, or has a rising temperature in the heat, the number here is irrelevant and the response is to call for help.
Frequently asked questions
- What is the heat index formula?
- Two formulas, in a fixed order. The first is Steadman's simplified form, HI = 0.5 · [T + 61 + (T − 68) · 1.2 + RH · 0.094], with T in Fahrenheit and RH as a percentage. If that comes out at or above 80 °F, the second runs: the Rothfusz regression, nine terms in T, RH, their squares and their cross products. Two corrections follow it — a subtraction when the air is very dry and an addition when it is very humid and only mildly warm. The order is not a style choice: below 80 °F the regression reads too high, so the simplified form is both the gate and the answer.
- How do I use this heat index calculator?
- Enter the air temperature in Celsius or Fahrenheit, and the relative humidity as a percentage. There is nothing to solve for and nothing to leave blank. The result panel prints the index in both scales — Celsius first, since that is how the temperature went in — and badges it with the band it falls in. The Fahrenheit value is the one to compare against the NWS chart, because the four band lines live on that scale. Below 80 °F you will still get a number and a band; the first band says the index is out of its defined range there rather than silently withholding an answer.
- Why does it feel hotter than the temperature says?
- Because your body cools itself by evaporating sweat, and humid air makes that harder. The air can only take up water until it is saturated, so at 90% humidity there is very little room left for the sweat leaving your skin, and the cooling that was supposed to happen mostly does not. The heat has nowhere to go, so it accumulates — which your body reports as feeling hotter than the air. The reverse is the dry case: at 10% humidity, sweat evaporates as fast as your skin can produce it, so 40 °C can feel cooler than the thermometer reads. The index is essentially a measurement of how well that one mechanism is working.
- What are the heat index bands, and where do the numbers come from?
- Four bands, all of them NWS's and all of them defined in Fahrenheit: caution from 80 °F, extreme caution from 90 °F, danger from 103 °F, and extreme danger from 125 °F. This page adds a fifth below 80 °F, because the regression is not valid there and a blank badge would look like a fault. The lines are not round numbers in Celsius — they work out to about 26.7, 32.2, 39.4 and 51.7 °C — which is why the table lists both scales and the judgement is made on the Fahrenheit value. The bands are about consequences rather than comfort: the first is where prolonged exposure starts to cost you, the last is where heat stroke becomes the expected outcome rather than a risk.
- Is the heat index the same as the wet-bulb temperature or the apparent temperature?
- No, and the differences matter. The wet-bulb globe temperature is what occupational heat stress standards use: it measures sun, wind and radiant heat as well as temperature and humidity, so it answers a question about a job site rather than about a person in the shade. Apparent temperature is an umbrella term — the heat index is one, the Australian Bureau of Meteorology's version is another, and they differ because they were fitted to different data and different assumptions about wind. The heat index is specifically a shade-and-light-wind number, which is why the same conditions in full sun can feel up to 8 °C worse than this page reports.
- How accurate is it, and when should I not trust it?
- It is a regression fitted to a physiological model, so treat it as a good estimate rather than a measurement: the band it lands in is more trustworthy than the decimal place. It goes wrong at the extremes — above about 110 °F with very high humidity the equation is extrapolated beyond its data, which is why this page's chart stops at 90% humidity rather than printing 278 °F for 110 °F and 100%. It also says nothing about you: acclimatisation, age, pregnancy, some medications and any condition that limits sweating all shift the real answer, and full sun or heavy work can push it further still. When the index is high, the number is the least interesting part of the decision — shade, water and stopping are what actually matter.
References
- What is the heat index? — the NWS page this calculator implements, with the two-step procedure, the two corrections, and the band chart in Fahrenheit — National Weather Service (Amarillo, TX)
- Heat index — the safety framing behind the band names, and what each one means for people working or exercising outdoors — National Weather Service
- Heat stress — the physiology behind the index: evaporative cooling, why humidity limits it, and who is at elevated risk — Centers for Disease Control and Prevention (NIOSH)