Write output separately for each year

[133040028]

Hi,
I am using MetSim to generate input files for VIC 5 image. It requires forcing yearwise. I could successfully run the MetSim package for my study region. However, the output is generated for entire 17 years which is my simulation period (which is a huge file!). Is there any way to write the output yearwise? I could see 'time_grouper:: str' in the documentation. I am not sure how to use it.

[arbennett]

Hi! Sorry for the far delayed response, but yes you should be able to use the time_grouper configuration for this purpose. If you are using the YAML configuration style this would just appear like:

MetSim:
    # Other settings...
    time_grouper: '1AS'

The {x}AS syntax derives from older versions of Pandas groupby operations for yearly aggregation. Newer syntax such as doing something like time_grouper: '1year' may also work but this hasn't been tested.

[yzhou65jhu]

Hi! I'm trying to run metsim in bash and to use the time grouper in the configure file, but it doesn't seem to work. I just put "time_grouper = 1AS" under the 'MetSim' section, but metsim writes output for my entire period, which is 2 years.

[133040028]

Hi,
For me, "out_freq = 1Y" worked. Can you please check if it works for you.

[yzhou65jhu]

Thanks @133040028 . It works! That just saved me a lot of time!

[dwoodson-usbr]

@arbennett Following up on this question - is there any way to specify multiple input forcing .nc files for MetSim? E.g. for running MetSim on daily data spanning 50-years, with 1 .nc for each year, it seems like all 50 .nc files must be stacked into one .nc file, then run through MetSim, is there any way to reference all 50 .nc files separately (or the directory they live in) as forcing inputs?

[arbennett]

@dwoodson-usbr - you should be able to use the * notation in the input files. So if your data has a naming scheme like forcing_year.nc you should be able to give the forcing entry as forcing_*.nc in the configuration.

[dwoodson-usbr]

Makes sense, thanks @arbennett !