Add some criteria to determine which observations should be marked as suspect when the manual observation is off by some amount.
E.g.
All observations after manual obs is off by more than > threshold (e.g. 5cm) until next manual obs?
Use linear interpolation between difference manual obs and time series, then apply threshold? Linear interpolation should have some limits (e.g. large gaps in series, ... ).
dbrakenhoff
commented
3 days ago
Add some criteria to determine which observations should be marked as suspect when the manual observation is off by some amount.
E.g.
Pick one and apply.