Heat Index Calculator: NOAA's Own Regression, Both Adjustments Included, and What It Still Cannot See

The heat index in °C and °F from air temperature and either relative humidity or dew point, computed with the full NOAA/NWS Rothfusz regression and both of the documented adjustments that most implementations leave out. Banded on the four published National Weather Service categories, with the India Meteorological Department’s experimental °C thresholds shown beside them rather than resolved, because the two schemes disagree by a wide margin. Three things this page will keep saying: the figure assumes shade and NOAA’s own statement is that full sunshine can add up to 15 °F; it is not WBGT, which is what sports and occupational bodies use for activity limits; and a heat index of 125 °F or more is something to act on now.

Last updated 6 Oct 2026

These estimate an exposure — from the weather, or from what you have taken in — using published models, not a measurement of you. Individual tolerance varies a great deal, so treat the figure as a guide to the conditions rather than a verdict on your own safety. Where a page flags a dangerous level, that is one to act on now rather than later.

Air temperature and humidity, on the published regression, with both adjustments

air temperature and relative humidity or dew point -> heat index in °C and °F, and which NWS band it falls in

Units for every temperature on this pageThis changes the unit of the two temperature fields below and nothing else. The arithmetic is always done in °F, because that is the unit Rothfusz fitted the regression in; the conversion runs at the exact factors 9/5 and 32 and the headline is then converted back. The regression is not re-fitted in °C and must not be — its nine coefficients are only valid against Fahrenheit.

Air temperature, measured in the shadeA dry-bulb air temperature taken out of the sun. The page refuses anything below 80 °F (26.7 °C), because NOAA states that the allowed temperatures are those shown in its Heat Index Table and the table starts at 80 °F — below that the heat index is not defined and this is not a wind-chill page. It also refuses above 122 °F (50 °C). If you only have a temperature from a phone weather app, that is a shade air temperature and is the right number to use here; a thermometer lying in the sun is not.

Which humidity figure you haveDew point is offered because in humid climates it is the number people actually quote, and because NOAA’s own heat index calculator takes temperature and dew point rather than relative humidity. The two are converted with the National Weather Service’s own vapour pressure formula, e = 6.11 × 10^{(7.5 T / (237.3 + T))} with T in °C and e in hPa, and relative humidity = 100 × e/e_s. The field you are not using is locked and shows the figure your other entry implies.

Relative humidity (%)0 to 100. Worth knowing that NOAA’s published Heat Index Table has no column below 40%, and that NCAR’s documentation of its own implementation of the NWS algorithm describes the default formulation as appropriate for relative humidities above 40% — which is precisely why a separate low-humidity adjustment exists below 13%. Both ends are computed here and both are flagged under the result.

Dew point temperatureIn the unit chosen at the top. A dew point above the air temperature is physically impossible — it would mean a relative humidity above 100% — and the page refuses it rather than printing a number. A dew point equal to the air temperature is 100% humidity, which is fog or heavy rain.

45.6°CExample

an air temperature of 36.0 °C in the shade with a relative humidity of 55 %

One simple formula, one nine-term regression, a rule for choosing between them, and two adjustments

Step 1, always: HI_simple = 0.5 × [T + 61.0 + (T − 68.0) × 1.2 + RH × 0.094] · Step 2: average it with the temperature, HI_avg = (HI_simple + T)/2 · Step 3: if HI_avg < 80 °F, that average IS the heat index and you stop · Step 4, otherwise the Rothfusz 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 · Step 5, low humidity: if RH < 13 and 80 ≤ T ≤ 112, SUBTRACT [(13 − RH)/4] × √([17 − |T − 95|]/17) · Step 6, high humidity: if RH > 85 and 80 ≤ T ≤ 87, ADD [(RH − 85)/10] × [(87 − T)/5] · Dew point to humidity, when you enter a dew point: RH = 100 × 10^{[7.5·T_d/(237.3 + T_d) − 7.5·T_c/(237.3 + T_c)]}, T in °C

T: air temperature in degrees Fahrenheit, in the shade. The regression’s nine coefficients are only valid against °F; a Celsius entry is converted at 9/5 and 32 before anything else happens, and the regression is never re-fitted.
RH: relative humidity as a percentage, 0 to 100 — a number, not a fraction. Feeding 0.55 instead of 55 is the commonest way this regression is got wrong, and it does not error: it returns a confident figure a long way out.
80 °F: the switch. NOAA’s wording is “In practice, the simple formula is computed first and the result averaged with the temperature. If this heat index value is 80 degrees F or higher, the full regression equation along with any adjustment as described above is applied.” The test is on the averaged value, not on the raw simple formula and not on the air temperature.
±1.3 °F: the published error of the regression, stated in Rothfusz’s Technical Attachment: “Because this equation is obtained by multiple regression analysis, the heat index value (HI) has an error of ±1.3 °F.” That is 0.72 °C, which is why this page does not pretend to a tenth of a degree of accuracy.
6.11, 7.5, 237.3: the constants in the National Weather Service’s own vapour pressure formula, used only for the dew point conversion. The 6.11 hPa cancels in the ratio, so the relative humidity depends on the two exponents alone.
137 °F: this page’s ceiling, and not an arbitrary one: it is the largest value printed anywhere in NOAA’s Heat Index Table. The blank cells in that table are exactly the combinations whose heat index exceeds it, so “above 137 °F” and “off the published table” are the same statement.

Worked example

an air temperature of 36.0 °C in the shade with a relative humidity of 55 %
Convert first, because the regression is in Fahrenheit. 36.0 × 9/5 + 32 = 96.8 °F. Everything from here is done in °F and converted back at the end.
The simple formula, which is always computed first. 0.5 × [96.8 + 61.0 + (96.8 − 68.0) × 1.2 + 55 × 0.094] = 0.5 × [96.8 + 61.0 + 34.56 + 5.17] = 0.5 × 197.53 = 98.765 °F.
Average it with the temperature and test against 80. (98.765 + 96.8)/2 = 97.78 °F, which is at or above 80, so the full regression replaces it. Had it come out below 80, that 97.78 would have been the answer and the regression would never have run.
The nine terms. With T = 96.8 and RH = 55: −42.379 + 198.3447 + 557.8832 − 1196.5978 − 64.0722 − 165.8219 + 633.2474 + 249.7228 − 56.4065 = 113.92 °F. Each term is large and they very nearly cancel — the sum is about one twentieth of the biggest term, which is why a single mistyped coefficient does not produce a slightly wrong answer, it produces nonsense.
Test both adjustments. Neither applies. The low-humidity adjustment needs RH below 13 % and 55 is not; the high-humidity adjustment needs RH above 85 % and a temperature of 80 to 87 °F, and neither holds. So the heat index stays 113.92 °F, which rounds to 114 °F — the whole-degree grid NOAA publishes on.
Back to Celsius, from the rounded Fahrenheit figure. (114 − 32) × 5/9 = 45.6 °C. Converting from the rounded °F figure rather than the raw one is deliberate: it makes the band boundaries on this page exactly 80, 90, 103 and 125 °F, so the category you are shown is the category NOAA's own published number would give. The cost is that the Celsius headline moves in steps of 5/9 °C, which is 0.56 °C — still well inside the regression's own ±1.3 °F.
And the band. 114 °F is between 103 and 124, so this is DANGER on the National Weather Service scale: heat cramps or heat exhaustion likely, heat stroke possible with prolonged exposure. In full sun, NOAA's up-to-15-°F statement puts the same conditions at 129 °F, which is Extreme Danger. 36 °C with 55 % humidity is an ordinary May afternoon across much of India.

The four National Weather Service heat index categories, with the published effect on the body

The four categories and the effect wording are the National Weather Service’s own, a United States Government work. Two things to notice. The °C column is a conversion made here, not a published scale — the NWS publishes in °F only. And the printed ranges have a one-degree seam: Danger is given as 103 to 124 and Extreme Danger as 125 or higher, so nothing is said about 124 to 125. This page bands on the whole-degree Fahrenheit figure, which closes the seam by putting 124 in Danger and 125 in Extreme Danger.
Category	Heat index (°F)	Heat index (°C)	Effect on the body, as the NWS publishes it
Caution	80 to 90	26.7 to 32.2	Fatigue possible with prolonged exposure and/or physical activity
Extreme Caution	90 to 103	32.2 to 39.4	Heat stroke, heat cramps, or heat exhaustion possible with prolonged exposure and/or physical activity
Danger	103 to 124	39.4 to 51.1	Heat cramps or heat exhaustion likely, and heat stroke possible with prolonged exposure and/or physical activity
Extreme Danger	125 or higher	51.7 or higher	Heat stroke highly likely

What the regression gives across the range that matters in India, in °C with °F in brackets

Computed by this page’s own implementation of the algorithm, with both adjustments active, and checked cell by cell against NOAA’s printed table where the two overlap. “Off the table” means the heat index exceeds 137 °F, the largest figure NOAA prints. The row worth staring at is 38 °C: at 55 % humidity that is a heat index of 125 °F, the NWS Extreme Danger threshold, from an air temperature that is unremarkable in May in Delhi or Kolkata. The 85 % and 100 % columns at 34 °C and above are monsoon conditions, and they run off the published scale entirely.
Air temp	30 % RH	40 % RH	55 % RH	70 % RH	85 % RH	100 % RH
28 °C / 82.4 °F	27.2 (81)	27.8 (82)	28.9 (84)	30.6 (87)	32.8 (91)	36.1 (97)
30 °C / 86.0 °F	28.9 (84)	29.4 (85)	31.7 (89)	35.0 (95)	38.9 (102)	44.4 (112)
32 °C / 89.6 °F	30.6 (87)	32.2 (90)	35.6 (96)	40.6 (105)	46.7 (116)	53.9 (129)
34 °C / 93.2 °F	33.3 (92)	35.6 (96)	40.0 (104)	46.7 (116)	55.0 (131)	off the table
36 °C / 96.8 °F	36.1 (97)	38.9 (102)	45.6 (114)	53.9 (129)	off the table	off the table
38 °C / 100.4 °F	39.4 (103)	43.3 (110)	51.7 (125)	off the table	off the table	off the table
40 °C / 104.0 °F	43.3 (110)	48.3 (119)	58.3 (137)	off the table	off the table	off the table
42 °C / 107.6 °F	47.2 (117)	53.9 (129)	off the table	off the table	off the table	off the table
44 °C / 111.2 °F	51.7 (125)	off the table	off the table	off the table	off the table	off the table

Two published heat index scales, side by side, not reconciled

The IMD figures are the four colour thresholds given in a July 2023 Ministry of Earth Sciences answer in the Rajya Sabha, which also states that the IMD heat index is derived “using the heat index equation similar to what is used by National Weather Service, National Oceanic and Atmospheric Administration (NOAA), USA”. So the two scales are computing very nearly the same number and then disagreeing about what to call it. The disagreement is not small: IMD’s highest category does not start until above 55 °C, while the NWS says heat stroke is highly likely from 51.7 °C. This page bands on the NWS scale because the NWS publishes an effect-on-the-body statement for each band and a colour is not one; the IMD category is reported in prose under the result so neither is hidden.
Heat index	National Weather Service (USA)	India Meteorological Department (experimental)
Below 26.7 °C (80 °F)	Below the lowest band; nothing published	Green (less than 35 °C)
26.7 to 32.2 °C	Caution	Green (less than 35 °C)
32.2 to 35 °C	Extreme Caution	Green (less than 35 °C)
35 to 36 °C	Extreme Caution	No category as published
36 to 39.4 °C	Extreme Caution	Yellow (36 to 45 °C)
39.4 to 45 °C	Danger	Yellow (36 to 45 °C)
45 to 46 °C	Danger	No category as published
46 to 51.7 °C	Danger	Orange (46 to 55 °C)
51.7 to 55 °C	Extreme Danger	Orange (46 to 55 °C)
Above 55 °C	Extreme Danger	Red (greater than 55 °C)

The two adjustments, their windows, and how large they get

Both are in NOAA’s published equation and both are omitted by most calculators. Each falls to zero at every edge of its own window, so neither introduces a step. The low-humidity one matters for pre-monsoon dry heat; the high-humidity one matters for a muggy morning, and is the one where NOAA’s printed table and NOAA’s printed equation disagree.
Adjustment	Window	Sign	Largest it gets	Where
Low humidity	RH below 13 % and T from 80 to 112 °F	Subtracted	3.25 °F (1.81 °C)	T = 95 °F, RH = 0 %
High humidity	RH above 85 % and T from 80 to 87 °F	Added	2.10 °F (1.17 °C)	T = 80 °F, RH = 100 %

What the number underneath the heat index actually assumes about you

From the National Weather Service Technical Attachment in which Rothfusz published the regression, which lists Steadman’s assumptions explicitly and notes that wind IS in the model — at 5 knots — contrary to the common belief that the heat index assumes still air. This is the list to read before deciding whether the number applies to you. If you weigh 90 kg, are wearing a cricket helmet, or are standing still in the sun on a windless afternoon, it does not describe your situation and no adjustment on this page fixes that.
Assumption in Steadman’s model	Value Rothfusz’s Technical Attachment lists
Ambient vapour pressure of the atmosphere	1.6 kPa, that is 16 hPa
Dimensions of a human	5 ft 7 in tall, 147 pounds — about 170 cm and 67 kg
Clothing cover	84 % coverage: long trousers and a short-sleeved shirt
Core temperature	98.6 °F, that is 37.0 °C
Activity	180 W per m² of skin area — the model person walking outdoors at 3.1 mph
Effective wind speed	5 knots, about 9.3 km/h
Sunshine	None. The chart is for shady locations; NOAA states full sunshine can add up to 15 °F

Why the heat index is two formulas with a switch, and what it is blind to

The heat index is not a measurement and it is not a temperature. It is an approximation to a 1979 physiological model, fitted in 1990, and it answers one question: at what dry air temperature would a particular model person feel the same as they do in the air you have described. Everything confusing about it follows from that sentence. There is a model person with a height, a weight, a shirt and a walking pace. There is an assumed wind. There is no sun.

The model is Steadman’s, published in the Journal of Applied Meteorology in 1979 as a temperature-humidity index built from human physiology and clothing science. Steadman produced a table. In July 1990 Lans Rothfusz, then at the National Weather Service, fitted a nine-term regression to that table so forecasters could compute a value instead of looking one up, and published it in a Technical Attachment with a stated error of ±1.3 °F. That regression is what every heat index on every phone and every news bulletin is, and this page implements it as NOAA documents it rather than as it is usually copied.

Three things in NOAA’s documented algorithm are routinely dropped, and all three are here. The first is that the regression is not used alone: a simpler Steadman-consistent formula is computed first, averaged with the air temperature, and only if that average reaches 80 °F does the regression run at all. Below that, the averaged simple value is the heat index — which means the two branches are different functions, and in very dry air they do not meet. At 0 % humidity the switch sits at 81.1 °F, where the simple branch gives 80 °F and the regression branch gives 77 °F: a 2.6 °F step for a tenth of a degree. The gap closes at about 39 % humidity. Nothing in NOAA’s documentation smooths it and this page does not invent a smoothing.

The second and third are the adjustments. If the relative humidity is below 13 % and the temperature is between 80 and 112 °F, up to 3.25 °F is subtracted. If the humidity is above 85 % and the temperature between 80 and 87 °F, up to 2.1 °F is added. The first is the one that matters for a pre-monsoon afternoon in Rajasthan; the second for a muggy coastal morning. A calculator that omits them overstates dry heat and understates muggy warmth, and most calculators omit them.

One honest inconsistency, found by checking rather than assuming. Reproducing all 135 printed cells of NOAA’s own Heat Index Table from the bare regression matches 134 of them exactly — the single exception is 90 °F at 45 % humidity, where the table prints 93 and the regression gives 92.49, which rounds to 92. But applying the high-humidity adjustment that NOAA’s equation page documents breaks eight further cells, all in the 80 to 86 °F rows at 90 % humidity and above: 80 °F at 100 % humidity is 87 °F in NOAA’s table and 89 °F with the adjustment applied. NOAA’s equation page and NOAA’s chart therefore disagree in that corner. This page follows the equation page, and says so under the result whenever the adjustment fires.

What the heat index is blind to, in rough order of how much it matters. Sun. The number is for shade, and NOAA’s own figure is that full sunshine can raise it by up to 15 °F. That is not a rounding error; it is the difference between Caution and Danger. Exertion. Steadman assumed a walk at 3.1 mph. Nothing on this page knows whether you are about to play cricket. Acclimatisation. NIOSH’s position is that heat tolerance develops over repeated exposure, and its recommended build-up for a worker new to heat is no more than 20 % exposure on day one with no more than 20 % added per day; even an experienced worker returning to a hot job is held to 50 % on day one, 60 %, 80 % and then 100 % on day four. A week or more away from the heat can cost a significant part of that adaptation, which is regained in two or three days. OSHA’s figures put the stakes plainly: almost half of heat-related worker deaths occur on a worker’s first day on the job or first day back after an extended absence, and over 70 % within the first week. The heat index bands know none of this; two people in the same 45 °C are not in the same danger. Wind, beyond 5 knots. It is in the model, but fixed.

And what you should use instead, for activity decisions: WBGT. Wet Bulb Globe Temperature is, in the National Weather Service’s own description, a measure of heat stress in direct sunlight that takes account of temperature, humidity, wind speed, sun angle and cloud cover; the heat index uses temperature and humidity only and is calculated for shade. WBGT is what athletic bodies use to set practice limits and what OSHA points to for workplace decisions. This page does not estimate it, and you should distrust any page that claims to from temperature and humidity alone: WBGT needs a black globe thermometer to capture the radiant load, and the radiant load is exactly the thing the heat index cannot see. A heat index plus NOAA’s 15 °F sunshine allowance is an upper bound on how wrong the shade figure might be, not a WBGT.

Where this matters most. India is the reason this page is written the way it is. The India Meteorological Department began issuing an experimental heat index in 2023, derived, in the Ministry of Earth Sciences’ own words, “using the heat index equation similar to what is used by National Weather Service, National Oceanic and Atmospheric Administration (NOAA), USA” — so the arithmetic is the same arithmetic, and the number on this page is comparable to the number IMD publishes. What differs is the labelling, and it differs a great deal: IMD’s red category begins above 55 °C where the NWS has already said, at 51.7 °C, that heat stroke is highly likely. Both scales are reported here. Neither is averaged into the other.

For what to do about the fluid side, the daily water intake page and the adult fluid requirement page are the ones built for it, and this page deliberately prints no fluid volume of its own: heat changes the answer by an amount no formula on this site can predict. For the exercise side, target heart rate zones and the VO2 max page are the companions. In the medical set, heat stroke’s downstream problems have their own pages — the rhabdomyolysis creatine kinase page, serum osmolality, the free water deficit page and urine output.

Frequently asked questions

Why does the page refuse to tell me the heat index at 24 °C?

Because there is not one. NOAA’s page on the calculation says the allowed values for temperature are those displayed in the NWS Heat Index Table, and that table begins at 80 °F (26.7 °C). Below that the regression is explicitly not appropriate, and the simple formula that NOAA substitutes was built to be consistent with Steadman’s warm-weather results, not to describe cool air. Extrapolated downward it returns numbers with no physiological meaning — and it is certainly not a wind chill, which is a different model with a different author. A calculator that answers anyway is telling you something it does not know.

The weather app says 41 and this page says 45.6. Which is right?

They are answering different questions. 41 is likely the air temperature; 45.6 is the heat index, the dry-air temperature that would feel the same. If your app is also showing a “feels like” figure and it still disagrees with this page, the usual reasons are these. It may be using the bare regression without the two adjustments. It may be using a different apparent-temperature model altogether — the Australian apparent temperature, Humidex and the UTCI are all in circulation and none of them is the NWS heat index. It may be applying a sun or wind correction of its own. Or it may be reading a slightly different humidity: at 36 °C a humidity error of five points is worth about 2.5 °C of heat index. The regression’s own published error is ±1.3 °F, so any two correct implementations can legitimately differ by a degree.

Is a heat index of 45 °C the same as a wet bulb globe temperature of 45 °C?

No, and they are not close. WBGT is a different scale with a different numerical range, and it is measured in direct sun with a black globe thermometer. The National Weather Service describes WBGT as accounting for temperature, humidity, wind speed, sun angle and cloud cover, while the heat index uses temperature and humidity and is calculated for shade. WBGT numbers are much lower than heat index numbers for the same conditions — occupational and athletic limits sit in the twenties and low thirties of °C — so comparing them directly will mislead you badly in both directions. This page does not estimate WBGT, because doing so from temperature and humidity alone means inventing the radiant load, which is the whole point of the globe thermometer.

Why does the page ask for a dew point at all?

Because it is often the number you actually have, and in humid climates it is the more useful of the two. NOAA’s own heat index calculator takes temperature and dew point rather than relative humidity. Relative humidity swings through the day just because the temperature does, so “70 % humidity” means one thing at 6 am and another at 2 pm; the dew point barely moves unless the air mass changes. A dew point of 24 °C is oppressive at any hour. The conversion used here is the National Weather Service’s own vapour pressure formula and nothing else.

How much does the sun add, really?

NOAA’s published statement is “exposure to full sunshine can increase heat index values by up to 15 °F”, which is 8.3 °C. That is a stated upper bound rather than a correction to apply, which is why this page reports it as a separate row labelled as such and does not fold it into the headline. The reason it is only an upper bound is that the actual radiant load depends on sun angle, cloud, what you are standing on and what you are wearing — and capturing that properly is what WBGT exists to do. The practical reading: if you are deciding whether to work, run or play in the open, the row with the 15 °F added is nearer your situation than the headline is.

Does being used to the heat change the bands?

It changes the risk a great deal and it changes the bands not at all, which is the honest and uncomfortable answer. NIOSH’s recommended build-up for someone new to working in heat is no more than 20 % exposure on the first day and no more than 20 % more each day after; even an experienced worker returning to a hot job is held to 50 % on day one, then 60 %, 80 % and 100 % on day four. Adaptation develops over roughly one to two weeks of repeated exposure, a week or more away can cost a significant part of it, and two or three days back in the heat regains it. OSHA reports that almost half of heat-related worker deaths happen on a worker’s first day on the job or first day back after an absence, and over 70 % in the first week. The heat index cannot see any of this. Two people standing in the same 45 °C heat index are not in the same danger, and the one who flew in yesterday is in much more of it.

Why is the Celsius figure moving in jumps of 0.6?

Because it is computed from the whole-degree Fahrenheit value, deliberately. The National Weather Service publishes its bands as 80, 90, 103 and 125 °F, and its table in whole degrees Fahrenheit. If the Celsius figure were rounded independently to one decimal place, a reading could print as 38.9 °C and sit in Extreme Caution while its own Fahrenheit value printed as 103 and sat in Danger — one tenth of a degree Celsius is finer than one degree Fahrenheit, so the two grids cannot be made to agree. Deriving Celsius from the rounded Fahrenheit figure makes the category you are shown exactly the category NOAA’s own published number gives, at the cost of 5/9 °C steps. The regression’s own error is ±1.3 °F, which is more than twice that step, so nothing real is lost.

I entered a dew point higher than the air temperature and it refused. Why?

Because that describes air holding more water vapour than it can hold, which is a relative humidity above 100 % and does not occur in the conditions this page is about. Nearly always it is a transposition — the two fields swapped — or a unit mismatch, a dew point in °F typed while the page is set to °C. A dew point exactly equal to the air temperature is 100 % humidity, which is fog, cloud or rain, and that the page will compute.

Can I use this for a child, an infant or an older relative?

You can read the conditions from it, and you should not read a personal risk from it. The number describes the air, not a person: Steadman’s model is a 67 kg adult walking at 3.1 mph in trousers and a short-sleeved shirt. Infants, people who are bed-bound or immobile, people on diuretics or with heart or kidney disease, and anyone who cannot get themselves out of the heat or reach a drink are at substantially more risk at the same heat index, and none of that is in the model. The practical use of a Danger or Extreme Danger reading is as a prompt to check on those people, which is exactly what the published advice says.

Is the heat index the same thing as Humidex, or as the “feels like” number?

No. Humidex is a Canadian index with a different formula and a different scale, and “feels like” is a marketing phrase that different providers compute differently — some use the NWS heat index in warm weather and a wind chill in cold weather, some use the Australian apparent temperature, some use something proprietary. The number on this page is specifically the NWS heat index from the Rothfusz regression, with NOAA’s two adjustments, and it will match NOAA’s own calculator except in the high-humidity corner described under the result.

Related calculators

References

National Weather Service, Weather Prediction Center. The Heat Index Equation. wpc.ncep.noaa.gov/html/heatindex_equation.shtml. The Rothfusz regression, the simple formula, the 80 °F averaging rule and both adjustments with their windows, as implemented on this page.
Rothfusz LP. The Heat Index “Equation” (or, More Than You Ever Wanted to Know About Heat Index). National Weather Service Technical Attachment SR 90-23 / Western Region Technical Attachment No. 90-24, 10 July 1990. Source of the nine coefficients, of the ±1.3 °F error figure, and of the list of Steadman’s assumptions reproduced on this page.
Steadman RG. The assessment of sultriness. Part I: a temperature-humidity index based on human physiology and clothing science. Journal of Applied Meteorology 1979;18:861–873. The underlying model; the regression is an approximation to its table.
National Weather Service. What is the heat index? weather.gov/ama/heatindex. The four published categories, their heat index ranges and the effect-on-the-body wording, and the worked example of 100 °F at 55 % humidity giving 124 °F.
National Weather Service. Heat Index and Heat Forecast Tools. weather.gov/safety/heat-index. The statement that exposure to full sunshine can increase heat index values by up to 15 °F, and the worked example of 96 °F at 65 % humidity giving 121 °F.
National Weather Service. Wet Bulb Globe Temperature. weather.gov/tsa/wbgt. The variable-by-variable comparison of WBGT and the heat index used on this page.
National Weather Service, Weather Prediction Center. Calculating the Heat Index. wpc.ncep.noaa.gov/heat_index/details_hi.html. The statement that the allowed values for temperature, at and above 80 °F, are those displayed in the NWS Heat Index Table.
National Weather Service Little Rock. Heat Index and Wind Chill Charts. weather.gov/lzk/charts.htm. The 135-cell Heat Index Table used to verify this page’s implementation cell by cell.
NCAR Command Language documentation, heat_index_nws. ncl.ucar.edu. An independent description of the NWS algorithm, used here to confirm that below the 80 °F switch the averaged simple value is itself the answer, and the source of the statement that the default formulation is appropriate for temperatures of 80 °F and above and relative humidities above 40 %.
National Weather Service El Paso. Vapor Pressure calculator documentation. weather.gov/media/epz/wxcalc/vaporPressure.pdf. The e = 6.11 × 10^{(7.5T/(237.3+T))} formula used for the dew point conversion.
National Institute for Occupational Safety and Health. Heat-Related Illnesses. cdc.gov/niosh/heat-stress/about/illnesses.html. The symptom and first-aid lists for heat stroke and heat exhaustion quoted on this page.
National Institute for Occupational Safety and Health. Acclimatization. cdc.gov/niosh/heat-stress/recommendations/acclimatization.html. The 20 %-per-day and 50/60/80/100 % build-up schedules and the loss and regain of acclimatisation.
Occupational Safety and Health Administration. Heat — Protecting New Workers. osha.gov/heat-exposure/protecting-new-workers. “Almost half of heat-related deaths occur on a worker’s first day on the job” and “over 70 percent of heat-related deaths occur during a worker’s first week” (citing Tustin 2018).
Ministry of Earth Sciences, Government of India. Rajya Sabha unstarred question answer, 27 July 2023. moes.gov.in. States that the IMD heat index is derived using a heat index equation similar to the NWS/NOAA one, and gives the four colour thresholds in °C quoted on this page.
United States Food and Drug Administration. Spilling the Beans: How Much Caffeine is Too Much? Cited here only for the cross-reference to the caffeine page, which uses its 400 mg a day figure.

CalcEngines health calculators are for education and for checking arithmetic that has already been decided elsewhere. They are not medical advice, they do not decide what to give, and they do not replace the judgement of a doctor, nurse, midwife or dietitian who knows the person in front of them. Every figure depends on the values you enter and on the assumptions stated on the page — check it against the prescription, the product label and your local policy before acting on it.