PSUM buffer (?) Snowpack not starting
- When a beginning date (-b arg) corresponding exactly to the starting date of the meteo data in the smet is provided, snowpack complains about missing precip
missing { precipitation } on ...
. This does not arise somehow with the res1exp example but with the data in attachment (which is at hourly resolution and not 30 minutes).
- When no beginning date is provided, snowpack will refuse to start if the meteo is at hourly resolution. Either is complains about precip, either about missing TA TSG RH sw_radiation precipitation precip_splitting depending if data starts at 00:00 or later.
For 1, I imagine this is because meteoio/snowpack wants to accumulate PSUM over HH-timestep:HH. However this is highly inconvenient and non intuitive for users (how to guess they have to start one time-step later than the starting date of the meteo input?). We can imagine that snowpack checks it before starting and eventually uses the first PSUM value as input. Moreover, MeteoSwiss PSUM at hourly resolution is always HH-1:HH (I don't know if this is an international standard), so anyway snowpack should use PSUM at timestep + 1 insteaad of - 1 (then the problem will arise at the end...).
For 2, either data starts at 00:00 and we fallback to 1, or when starting later than midnight it seems that when no -b is provided snowpack tries to start at midnight, which is also non intuitive at all.
Edit: IT seems that ProfileDate in the sno file is used when no -b is provided. Is this the case? If yes, again not intuitive ;)