Mont Blanc is 4,810 metres. Or 4,805.59. Or 4,792. Kebnekaise is 2,111, or 2,096.8, or 2,088.4. None of these numbers is wrong. They just answer different questions.
We build reliefs of mountains, so we get this question more often than you might expect: why does the sign say one thing, Wikipedia another and your father's old hiking map a third? The answer is that "height above sea level" contains three assumptions, and every one of them can move. The sea. The method. And the summit itself.
Height above which sea?
The sea surface is not an even sphere. The Earth's mass is unevenly distributed, so gravity varies from place to place, and water settles accordingly. The imaginary surface where gravity is equally strong everywhere, extended in under the continents, is called the geoid. That is the zero level.
In Sweden the system is called RH 2000. It became the official height system in 2005 and is based on the country's third precision levelling, a project that ran from 1979 to 2003. The heights are so-called normal heights, and the system has an epoch: the year 2000. That a height system has a year attached sounds odd until you consider the next point.
The ground does not stand still
Scandinavia is still rising after the ice sheet. The uplift reaches about a centimetre a year at its fastest, in the northern Gulf of Bothnia. Over a couple of decades that adds up to several decimetres, which is more than the uncertainty in a modern height measurement. Every height value therefore has to be tied to a year, and that is why the specification states an epoch. The height is a description of where the ground was in a given year.
GPS gives the wrong height on purpose
A GNSS receiver, which is what sits in your phone and in survey equipment, does not calculate height above sea level. It calculates height above a mathematically smooth ellipsoid approximating the Earth. The difference between that ellipsoid and the geoid can be tens of metres.
Getting from one to the other requires a geoid model. The Swedish one is called SWEN17_RH2000 and, according to Lantmäteriet, has an estimated uncertainty of 8 to 10 millimetres across the mainland, Öland and Gotland. So that is not where the errors come from when a mountain changes height. The millimetres are under control. It is the metres that move.
The summit can be snow
Here is the main explanation. A mountain's highest point does not have to be rock.
The south peak of Kebnekaise carries a glacier. At Tarfala Research Station it is measured at the end of every summer, precisely in order to capture the year's low point. In September 2025 it stood at 2,088.4 metres, a metre and a half lower than the year before. The north peak next to it is solid rock and sits at 2,096.8 metres. Since 2019 the north peak has been Sweden's highest point.
Mont Blanc has the same problem on a larger scale. Since 2001 the summit has been surveyed every two years by surveyors in Haute-Savoie. 2021 gave 4,807.81 metres, 2023 gave 4,805.59 metres. The rock beneath the snow sits at roughly 4,792 metres and is, in addition, some tens of metres west of the snow-covered high point. What varies between surveys is, in practice, how much snow the wind has piled onto the summit.
So the round figure of 4,810 printed on almost every map is not a measurement. It is an established number.
And mountains that actually grow
Etna does the opposite. The volcano's highest point moves between craters depending on where the latest eruptive sequence has deposited its material. In August 2021 INGV established that the Southeast Crater had reached 3,357 metres and thereby overtaken the Northeast Crater, which had held the top spot for forty years. Three years later, in September 2024, the Voragine crater had built itself up to 3,403 metres and taken over.
Forty-six metres of difference in three years. That is not measurement uncertainty, that is the mountain.
How we handle it
When we build a relief, the shape comes from elevation data, not from a number. Every point in the terrain has its own height, and that data has a year of collection. An Etna relief is therefore a snapshot of the craters as they were when the survey was made, exactly as a photograph is a snapshot.
The number on the nameplate is something else. It carries the figure the place is known by, the one on the sign at the summit and in the guidebook. That number is also a snapshot, but of a decade rather than a year.
Both are true. They just answer different questions, exactly as at the start.
Read on
If you want to see what elevation data actually looks like before it becomes a physical surface, start with FRMD by You, where you choose an area on the map and see the terrain in preview. And if you would rather just have Sweden's highest mountain on the wall: the Kebnekaise piece shows the whole massif, both summits included.