Handling gaps in long-term met-ocean measurement data

I have ten years of wave and wind measurements from a Baltic platform with gaps of hours to several weeks. For fatigue analysis I need continuous series. Do you prefer reanalysis data to fill gaps, statistical imputation, or simply excluding incomplete periods? Any references on how much gap-filling is acceptable?

Finish signup to comment

Comments

@tomek_zielinski

36d

If you go statistical, check that the imputation preserves extremes – most methods smooth them out, and extremes are exactly what fatigue cares about.

@KrzysztofNowicki

37d

W oceanografii standardem jest uzupełnianie krótkich luk interpolacją, a długich danymi z reanalizy po kalibracji do pomiarów. Kluczowe: opisać to przejrzyście i pokazać wrażliwość wyniku.