You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It happened quite often that raw data from loggers were lost or corrupted.
For these periods, transmissions are used instead.
But transmissions used to be daily in the winter and therefore daily values have been inserted as hourly values at many sites.
See an example at KAN_U, all these lines are daily averages found in the KAN_U_hour.csv file:
It also cause some problems because those daily values are filtered out when interpreted as hourly values: #243
So either:
daily values should only appear in the daily file and the corresponding period in the hourly file should be left blank
duration or time bound fields could be added to indicate that there are mixed time steps
I believe data users do not always go through the whole doc and use a "hourly" file as hourly values without looking for more details. So I'd be opting for option #1.
The text was updated successfully, but these errors were encountered:
It happened quite often that raw data from loggers were lost or corrupted.
For these periods, transmissions are used instead.
But transmissions used to be daily in the winter and therefore daily values have been inserted as hourly values at many sites.
See an example at KAN_U, all these lines are daily averages found in the
KAN_U_hour.csv
file:It also cause some problems because those daily values are filtered out when interpreted as hourly values:
#243
So either:
I believe data users do not always go through the whole doc and use a "hourly" file as hourly values without looking for more details. So I'd be opting for option #1.
The text was updated successfully, but these errors were encountered: