Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

EBAS reader: Handle time step sizes that differ over the provided time series #1228

Open
jgriesfeller opened this issue Jun 26, 2024 · 2 comments
Milestone

Comments

@jgriesfeller
Copy link
Member

jgriesfeller commented Jun 26, 2024

This concerns all types of EBAS readers...

From EMEP report 2022:

Precipitation:
Amberd looks VERY high

Remaining issue: Når jeg ser på Amberd, så er måleperiodene totalt irregulære. Det varierer mellom en halv dag og flere dager 
(feks 10 jan fra 10:20 til 19:10, og 4.3 kl 7 til 9.3 kl 7). Dette klarer ikke Pyaerocom å håndtere (ihvertfall ikke enda), 
så disse dataene må vi kaste ut.
  • The first discussion needs to be about how we want pyaerocom to handle this?
  • What to do if not all time steps fit?

How it is today (to my best knowledge):
Because the colocation can't handle time colocation with start and stop times. it has to assume a certain time step size and do the colocation on the middle points. We do accept small deviations from the time step sizes we have (details are here:

class TsType:
VALID = TS_TYPES
VALID_ITER = VALID[:-1]
FROM_PANDAS = PANDAS_FREQ_TO_TS_TYPE
TO_PANDAS = TS_TYPE_TO_PANDAS_FREQ
TO_NUMPY = TS_TYPE_TO_NUMPY_FREQ
TO_SI = TS_TYPE_TO_SI
TS_MAX_VALS = {
"minutely": 360, # up to 6hourly
"hourly": 168, # up to weekly
"daily": 180, # up to 6 monthly
"weekly": 104, # up to ~2yearly
"monthly": 120,
} # up to 10yearly
# "monthly": "days" below is because each month does not have the same number of days
# netcdf does time calculation for you given starting day and days past (CF convention)
TSTR_TO_CF = {"hourly": "hours", "daily": "days", "monthly": "days"}
TOL_SECS_PERCENT = 5

But the devil might like in the details of the reader implementations

@jgriesfeller jgriesfeller added this to the m2024-09 milestone Jun 26, 2024
@lewisblake
Copy link
Member

I would say we should ask the modelers how they want the comparison done and then we implement that rather than us coming up with a solution which they might not be satisfied with. If we are to include this data, what should we do with it? If no solution is given, then it is fair to throw out this station.

@jgriesfeller
Copy link
Member Author

@lewisblake agreed. The "probblem" right now is that the station is not ignored automatically. (Yes, I should have mentioned that above)
Pyaerocom's focus is to get as much data as possible.

@heikoklein heikoklein modified the milestones: m2024-09, m2024-11 Aug 20, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants