Cmorization

The pyremo package includes a cmorization module that should be useful to rewrite remo output for publication. The module is using the official climate model output rewriter (CMOR3 API) provided by PCMDI. The goal of the cmorization process is to provide climate model data output based on the Climate and Forecast (CF) standards to make it easily comparable to other model output.

The model output data not neccessarily altered (except maybe for converting units) but it is rather organized in a certain manner. For example, the hourly output of remo might be resampled to 3-hourly and daily instantaneous and mean values. Also the variables names and attributes are harmonized to provide standard names and units. Additionally, the different output variables are stored in a certain directory tree that is required for uploading it to an ESGF node.

Examples

We will show some simple examples here of how the pyremo cmorization module can be used. Let’s load an example remo dataset that contains monthly means for the year 2000

[1]:

import xarray as xr

import pyremo as pr

ds = pr.tutorial.open_dataset("remo_EUR-11_TEMP2_mon")
ds

The dataset is slightly larger than the official CORDEX domain according to archive specifications. Also the variable name is not CF standard and global attributes are missing. To cmorize this dataset, we use the pyremo cmor module. We also use CMIP6 cmor tables that are provided by py-cordex.

[2]:

from cordex.tables import cmip6_cmor_table, cordex_cmor_table

from pyremo.cmor import cmorize_variable

[3]:

filename = cmorize_variable(
    ds,
    "tas",
    cmor_table=cmip6_cmor_table("CMIP6_Amon"),
    dataset_table=cordex_cmor_table("CORDEX_remo_example"),
    CORDEX_domain="EUR-11",
)

/Users/buntemey/python/packages/pyremo/pyremo/cmor/remo_cmor.py:292: UserWarning: time units are set to default: days since 1949-12-01T00:00:00
  warn("time units are set to default: {}".format(u))
/Users/buntemey/python/packages/pyremo/pyremo/cmor/remo_cmor.py:456: UserWarning: adding time bounds
  warn('adding time bounds')

C Traceback:
In function: cmor_grid
! 

!!!!!!!!!!!!!!!!!!!!!!!!!
!
! Warning: it is recommended you pass the latitude bounds values when defining a grid
!
!!!!!!!!!!!!!!!!!!!!!!!!!


C Traceback:
In function: cmor_grid
! 

!!!!!!!!!!!!!!!!!!!!!!!!!
!
! Warning: it is recommended you pass the longitude bounds values when defining a grid
!
!!!!!!!!!!!!!!!!!!!!!!!!!


C Traceback:
In function: cmor_write_var_to_file
! called from: cmor_create_var_attributes
! 

!!!!!!!!!!!!!!!!!!!!!!!!!
!
! Warning: Invalid value(s) detected for variable 'longitude' (table: grids): 62223 values were lower than minimum valid value (0). Minimum encountered bad value (-44.594) was at (axis: index/value): rlat: 411/21.835 rlon: 0/-28.375
!
!!!!!!!!!!!!!!!!!!!!!!!!!


C Traceback:
In function: cmor_close_variable
! 

!!!!!!!!!!!!!!!!!!!!!!!!!
!
! Warning: while closing variable 2 (tas, table Amon)
! we noticed you wrote 0 time steps for the variable,
! but its time axis 0 (time) has 12 time steps
!
!!!!!!!!!!!!!!!!!!!!!!!!!

The warnings generated here mostly originate from the fact, that the original CMIP6 cmor tables have not yet been updated for use in regional climate model data. We will ignore them for now and have a look at the output

[4]:

filename

[4]:

'CORDEX/CMIP6/CORDEX/GERICS/REMO2-0-0/ECMWF/ECMWF-IFS-LR/historical/r1i1p1f1/Amon/tas/gn/v20220412/tas_Amon_REMO2-0-0_ECMWF-IFS-LR_historical_r1i1p1f1_gn_200001-200012.nc'

[5]:

output = xr.open_dataset(filename)
output

You can see that the dataset now has the right dimension sizes and names, added time bounds (since the dataset contains mean values) and also lots of global attributes to self describe the dataset.

[6]:

output.tas.plot(col="time", col_wrap=4)

[6]:

<xarray.plot.facetgrid.FacetGrid at 0x10e9455b0>