Boneyard Tools

How the heat index is calculated

A plain walkthrough of the NWS Rothfusz regression, the 80 F threshold, and the dry and humid adjustments that shape the feels like number.

From a simple estimate to the full regression

The National Weather Service does not jump straight to its long equation. It first evaluates a simple form that blends temperature and humidity, then averages that with the air temperature. If the average lands below 80 F, humidity is barely influencing comfort, so that simple value is reported as is. Only when the average reaches 80 F or higher does the calculator switch to the full Rothfusz regression, which is why readings on a mild spring day look almost identical to the plain thermometer.

The nine-term Rothfusz regression

The core formula is a polynomial fit that Lans Rothfusz published in 1990 to approximate Robert Steadman's detailed 1979 human comfort model. It sums nine terms built from the temperature, the humidity, their squares, and their cross products, each scaled by a fixed coefficient. That structure lets a single equation reproduce a lookup table that would otherwise need hundreds of tabulated values. Because it is a statistical fit rather than physics, it carries an error of roughly plus or minus 1.3 F across its valid range.

The dry and humid adjustments

Two correction terms fine tune the edges of the humidity scale. When relative humidity falls below 13 percent and the temperature sits between 80 and 112 F, an amount is subtracted because extremely dry air lets sweat evaporate faster and feel cooler. When humidity climbs above 85 percent and the temperature is between 80 and 87 F, a small amount is added, since saturated muggy air feels even more oppressive than the base equation predicts. Outside those windows no adjustment applies.

Where the number stops being reliable

The regression was tuned for the warm, humid conditions that produce dangerous heat, so it should be trusted mainly at 80 F and above. It also assumes a person in the shade, walking at a light pace, in a gentle wind. Direct sun, hard exertion, dark clothing, or poor hydration can all push the real strain well past the calculated figure, so treat the output as a planning guide rather than a precise physiological verdict.

Frequently asked questions

Who created the heat index formula?

Robert Steadman built the underlying human comfort model in 1979 from assumptions about body mass, clothing, and sweat rate. Lans Rothfusz of the NWS then fit the fast polynomial regression to Steadman's tables in 1990 so forecasters could compute it quickly, and that regression is what this tool uses.

Why is the accepted error about 1.3 degrees?

Because the equation is a curve fit to a table, it cannot match every cell exactly. The published root mean square error of the Rothfusz regression against Steadman's values is roughly plus or minus 1.3 F, which is small enough for forecasting but means you should not read the last decimal as gospel.