EMEP hourly (mean) output vs EEA hourly observations: How is hourly data collocated?

[avaldebe]

EMEP mean output reports "time at the middle of the period", so EMEP hourly (mean) output for hour HH to HH+1 have time HH:30 (UTC).

EEA hourly (mean) observations have Start and End time columns in CET. The pyaro reader handles the zone conversion and provides start_times and end_times from those columns.

After browsing the pyaerocom source code with @dulte we could not figure how the time collocation uses start_times and end_times to match the model's "time at the middle of the period"

@dulte please add more info from our discussion

[dulte]

@heikoklein , do you know how this is done in the bridge between pyaro and ungriddeddata?

[heikoklein]

Pyaro and UngriddedData(Structured) keep Start and End time:

pyaerocom/pyaerocom/ungriddeddata_structured.py

Line 858 in 223b648

"start_time": var_data.start_times,

When converting from UngriddedData to StationData (for comparison with observations) the end_time information gets lost, e.g.

pyaerocom/pyaerocom/ungriddeddata_structured.py

Line 356 in 223b648

dtime = subset["start_time"]

but the ts_type (hourly) is kept.

I'm not sure how the hourly comparison in the collocation then is done.