Using Mean Dynamic Topography for elevation leveling on lands

Altitudes on lands are expressed with respect to a reference surface that has to realize as good as possible an equipotentiel of the gravity field; historically, this one was accessed through a temporal mean of sea level derived from observations at, at least, one point per country, following the resolution of the first General Assembly of the International Association of Geodesy, held in 1862 in Berlin. This is why you can "gain" or "loose" several centimeters or even tens of centimeters in altitude in one step by just crossing a border. In (mainland) France, the reference is the Marseille tide gauge over the period 1885-1896. From that point, using land surveyor techniques such as spirit leveling, the "zero" was propagated from place to place, reaching the other coasts of the country, some around 1000 km away - and not by the Mediterranean Sea like Marseille, but by the Atlantic Ocean, English Channel or North Sea. For neighboring countries the principle can be the same - a given point as reference (e.g. Newlyn in Great Britain, Alicante in Spain mainland, etc), propagated over their whole territory. For islands, with no land in between to propagate the measurements, a dedicated similar reference has to be chosen, for instance Ajaccio between 1912 and 1937 in Corsica.

This technique, while highly accurate at the local level, may be prone to systematic errors over long distances, leading to a zero surface not horizontal -- e.g. Dunkerque, on the North Sea, nearly 900 km north of Marseille is 25 cm higher with respect to that altitude zero than it should. The other obvious one is the historical discrepancies between neighboring countries, and the impossibility of having the same zero of altitude on an island, even when in the same country. Moreover, when defining the "level of the sea" as zero, the mean dynamic topography was not considered, since not known at first, then not measurable anyway. Nevertheless, scientific leveling surveys and intercomparison studies with neighboring countries, conducted by the IGN (National Geographic Institute), have initiated an assessment of the horizontality of the elevation system for mainland France.

For the past years, however, we have had mean dynamic topography data, computed from altimetry Mean Sea Surface, a geoid, drifting buoys and temperature/salinity profiles. It enables to remove mean dynamic topography from the "level of the sea", and thus reference the zero of altitudes to the geoid used in the mean dynamic topography computation, something which can be done on all coastal regions. Referencing the zero of altitude to a geoid using Mean Dynamic Topography can be a way of having an international -- at least for countries with a coast --, and also overseas uniform reference. It is also important in coastal sea level rise vulnerability assessment.