Vitamin D

How all of this is calculated

No figure on this site is an eyeball estimate. They all come out of a published model, and here is which one, with what it knows and what it does not.

Solar geometry

The position of the sun is computed with the standard declination and hour-angle equations: for a latitude, a longitude and a day of the year, the solar elevation at every instant. Sunrise, sunset, civil twilight and day length all come from that. It is astronomy, not statistics: it does not depend on the weather.

Clear-sky UV index

UV is estimated with the Madronich (2007) model, corrected for seasonal ozone after van Heuklon. Ozone matters because it is precisely what absorbs UVB when the sun is low, and that is the regime deciding whether a vitamin D winter exists. The city and month pages use that model under clear skies: clouds can only reduce those values. Your daily dashboard is different — there the UV index comes from Open-Meteo's real forecast, which already has cloud cover in it.

Vitamin D synthesis

From the UV, the minimal erythemal dose is calculated for the Fitzpatrick skin type, adjusted for age and for the area of skin exposed. The result is the minutes needed for a target dose. It is a population estimate, not a diagnosis.

The two thresholds

Synthesis needs a UV index of at least 3. How high the sun must climb to get there is not a constant: it depends on the ozone above you, which varies with latitude and season. The site computes that elevation for each place and each day, ranging from about 29° where the ozone layer is thinnest to about 42° where it is thickest. A fixed threshold was used before, and that simplification is precisely what caused the error corrected in July.

What this model does not know

It does not know whether there are clouds at the moment you look. It does not account for shade from buildings or mountains, nor for reflection off snow or water, which can increase exposure appreciably. It treats skin as a flat surface facing the sky. And it describes an average person of each skin type, not you: medication, pregnancy, advanced age and several illnesses change synthesis. It is not a substitute for a blood test or for the advice of a healthcare professional.

Model corrections

A model that has never changed is a model nobody has checked. These are its corrections.

9 July 2026

The previous UV estimator overestimated by 3 to 4 times when the sun was low, because it ignored ozone. The consequence was serious and published: the site claimed that Boston, New York, Madrid, Chicago and Toronto synthesise vitamin D twelve months a year, contradicting Webb, Kline and Holick (1988). It was replaced by Madronich with van Heuklon ozone, and 51 of the 73 cities stopped being listed as “all year round”.

References

Review

The medical content of this site has not yet been reviewed by a healthcare professional. When it is, their name and professional registration will appear here.