… and with many large and small local variations that depart from an oblate spheroid.
The Earth’s surface is a very dynamic system due to plate tectonics. And land masses are still recovering from the post-glacial rebound so that, for example, some coastlines around the UK are rising out of the sea whilst others are sinking into it.
The whole concept of MSL is a rough one. The UK regularly has a difference of 15m between high tide and low tide. The UK reference is based on that measured at Newlyn in Cornwall between 1915 and 1921. A hundred years on, it’s still the reference for “0m” [no one wants to rework and reissue all UK maps] but it’s no longer the MSL even for Cornwall.
Hence my point about heights of hills and mountains – they may be getting measured more precisely according to the accepted definition but they are still arbitrary. IMHO there should be a grace zone around the definitions of Munros, Marilyns, etc [a bit like traffic speed cameras have] so they are not demoted on as little as a 30cm shortfall.