cesm-lens files and xarray · python-questions

Stream: python-questions

Topic: cesm-lens files and xarray

Allison Baker (Apr 30 2021 at 17:21):

I am reading in a monthly CAM CESM-LENS file that starts Jan 1920, but when I look at the at time attributes in xarray, it looks like the first month is February -- not January. (I also tried a CESM-LENS2 file and found the same thing. I thought the first monthly average in this file was for Jan. I am trying to do annual averages so wanted to use this info (also to weight by days in month). Can someone tell me what I am missing here?

import xarray as xr
filename = '/glade/collections/cdg/data/cesmLE/CESM-CAM5-BGC-LE/atm/proc/tseries/monthly/TS/b.e11.B20TRC5CNBDRD.f09_g16.011.cam.h0.TS.192001-200512.nc'
ds = xr.open_dataset(filename)
ts = ds['TS']
months = ts.time.dt.month
years = ts.time.dt.year
month_length = ts.time.dt.days_in_month

Then I get the following:
months.png

Anderson Banihirwe (Apr 30 2021 at 17:34):

@Allison Baker, this is a well known issue when working with CESM data...The issue can be fixed by

telling xarray not to decode the time
computing new time axis by averaging the time bounds
and decoding the time manually...

n [25]: filename = '/glade/collections/cdg/data/cesmLE/CESM-CAM5-BGC-LE/atm/proc/tseries/monthly/TS/b.e11.B20TRC5CNBDRD.f09_g16.011.cam.h0.TS.192001-200512.nc'

In [26]: ds = xr.open_dataset(filename, decode_times=False)

In [27]: attrs = ds.time.attrs.copy()

In [28]: ds = ds.assign_coords(time= ds.time_bnds.mean('nbnd'))

In [29]: ds.time.attrs = attrs

In [30]: ds = xr.decode_cf(ds)

Anderson Banihirwe (Apr 30 2021 at 17:34):

In [31]: ds
Out[31]:
<xarray.Dataset>
Dimensions:       (ilev: 31, lat: 192, lev: 30, lon: 288, nbnd: 2, slat: 191, slon: 288, time: 1032)
Coordinates:
  * ilev          (ilev) float64 2.255 5.032 10.16 18.56 ... 967.5 985.1 1e+03
  * lat           (lat) float64 -90.0 -89.06 -88.12 -87.17 ... 88.12 89.06 90.0
  * lev           (lev) float64 3.643 7.595 14.36 24.61 ... 957.5 976.3 992.6
  * lon           (lon) float64 0.0 1.25 2.5 3.75 ... 355.0 356.2 357.5 358.8
  * slat          (slat) float64 -89.53 -88.59 -87.64 ... 87.64 88.59 89.53
  * slon          (slon) float64 -0.625 0.625 1.875 3.125 ... 355.6 356.9 358.1
  * time          (time) object 1920-01-16 12:00:00 ... 2005-12-16 12:00:00

Anderson Banihirwe (Apr 30 2021 at 17:34):

In [32]: ds.time.dt.month
Out[32]:
<xarray.DataArray 'month' (time: 1032)>
array([ 1,  2,  3, ..., 10, 11, 12])
Coordinates:
  * time     (time) object 1920-01-16 12:00:00 ... 2005-12-16 12:00:00

Deepak Cherian (Apr 30 2021 at 17:35):

FAQ time!

Anderson Banihirwe (Apr 30 2021 at 17:37):

FAQ time!

Will submit a PR shortly

Allison Baker (Apr 30 2021 at 17:38):

Thanks a lot! I have mostly been working with daily data so hadn't seen this...

Max Grover (Apr 30 2021 at 17:38):

Thanks a lot! I have mostly been working with daily data so hadn't seen this...

Thanks for asking the question!

Katie Dagon (Apr 30 2021 at 18:09):

Yes this is very helpful. I'll add that if you want to do this for land output, there are a couple of small tweaks you have to make based on how things are named. For Line [28] in @Anderson Banihirwe 's code above, I (think) it would be the following:
ds = ds.assign_coords(time=ds.time_bounds.mean('hist_interval'))

Keith Lindsay (Apr 30 2021 at 18:23):

I have a function time_set_mid at
https://github.com/klindsay28/CESM2_coup_carb_cycle_JAMES/blob/22aa55e60471078922d96f07033adb326769e47a/src/utils.py#L69-L103
that resets time to the midpoints of time bounds. It works if decode_times was True or False, and works with different components of CESM that have different names for the time bounds variables. I also have a test suite for it at
https://github.com/klindsay28/CESM2_coup_carb_cycle_JAMES/blob/22aa55e60471078922d96f07033adb326769e47a/tests/test_utils.py#L49-L87

Anderson Banihirwe (Apr 30 2021 at 18:26):

@Allison Baker, @Katie Dagon's shortcut works

In [11]: ds = xr.open_dataset(filename)

In [12]: ds = ds.assign_coords(time=ds.time_bnds.mean('nbnd'))

You may want to preserve your original attributes because xarray seems to throw those away during the ds = ds.assign_coords(time=ds.time_bnds.mean('nbnd'))

In [13]: ds.time.attrs
Out[13]: {}

Last updated: Jan 30 2022 at 12:01 UTC