Examples of CORDEX-CMIP6 cmorization

Examples of CORDEX-CMIP6 cmorization#

Introduction#

This notebook demonstrates how climate model output can be reformatted and standardized using CMOR (Climate Model Output Rewriter) for the CORDEX-CMIP6 framework.

What is CMORization?

CMORization is the process of converting raw model output into standardized netCDF files
It ensures compliance with CF (Climate and Forecast) conventions
It automatically handles unit conversions, time resampling, and metadata standardization

In this notebook, we’ll show examples of how to cmorize different types of climate data (fixed fields, monthly, and sub-daily data).

!ls ../Tables

CORDEX-CMIP6_1hr.json	      CORDEX-CMIP6_formula_terms.json
CORDEX-CMIP6_3hr.json	      CORDEX-CMIP6_fx.json
CORDEX-CMIP6_6hr.json	      CORDEX-CMIP6_grids.json
CORDEX-CMIP6_CV.json	      CORDEX-CMIP6_mon.json
CORDEX-CMIP6_coordinate.json  CORDEX-CMIP6_remo_example.json
CORDEX-CMIP6_day.json

Controlled Vocabulary (CV)#

The CORDEX-CMIP6_CV.json file defines the controlled vocabulary and required global attributes that the CMOR3 library uses for rewriting and validating the model output. This ensures all CORDEX-CMIP6 datasets follow the same conventions and contain consistent metadata.

Below we show the required global attributes that must be present in every CORDEX-CMIP6 dataset:

import json
import pprint

from IPython.display import JSON

# Opening JSON file
with open("../Tables/CORDEX-CMIP6_CV.json") as json_file:
    CV = json.load(json_file)["CV"]

# JSON(CV['required_global_attributes'])
pprint.pprint(CV["required_global_attributes"])

['activity_id',
 'contact',
 'Conventions',
 'creation_date',
 'domain_id',
 'domain',
 'driving_experiment_id',
 'driving_experiment',
 'driving_institution_id',
 'driving_source_id',
 'driving_variant_label',
 'frequency',
 'grid',
 'institution',
 'institution_id',
 'license',
 'mip_era',
 'product',
 'project_id',
 'source',
 'source_id',
 'source_type',
 'tracking_id',
 'variable_id',
 'version_realization']

CMORization Examples#

We use the cordex.cmor module to create examples of CORDEX-CMIP6 compliant datasets. The following section defines three functions that demonstrate how to cmorize different types of climate data:

Fixed fields (fx): Time-independent data like topography (orog)
Monthly data (mon): Variables averaged over monthly timescales (e.g., temperature)
Sub-daily data: Hourly, 3-hourly, and 6-hourly data that can be automatically resampled

Each function shows how to:

Specify the input variable mapping
Define the CMOR table for the output format
Handle coordinate systems and domain information
Convert units as needed

Show code cell source

Hide code cell source

import os

import cordex as cx
import xarray as xr
from cordex import cmor as cmor

table_dir = "../Tables"

# Set the CMOR table prefix for CORDEX-CMIP6 standard
cmor.set_options(table_prefix="CORDEX-CMIP6")


def test_cmorizer_fx():
    """
    CMORize a fixed field (time-independent) variable.

    This example demonstrates how to cmorize topography (orog) data,
    which is a fixed field that doesn't change over time. We use a dummy
    topography field generated from the EUR-11 CORDEX domain.

    Returns
    -------
    str
        Path to the cmorized netCDF file
    """
    ds = cx.cordex_domain("EUR-11", dummy="topo")
    filename = cmor.cmorize_variable(
        ds,
        "orog",
        mapping_table={"orog": {"varname": "topo"}},
        cmor_table=os.path.join(table_dir, "CORDEX-CMIP6_fx.json"),
        dataset_table=os.path.join(table_dir, "CORDEX-CMIP6_remo_example.json"),
        grids_table=os.path.join(table_dir, "CORDEX-CMIP6_grids.json"),
        CORDEX_domain="EUR-11",
        time_units=None,
        allow_units_convert=True,
    )
    return filename


def test_cmorizer_mon():
    """
    CMORize a monthly aggregated variable.

    This example demonstrates how to cmorize monthly averaged temperature (tas).
    The input data is loaded from a tutorial dataset containing hourly 2m temperature.

    Returns
    -------
    str
        Path to the cmorized netCDF file
    """
    ds = cx.tutorial.open_dataset("remo_EUR-11_TEMP2_mon")
    filename = cmor.cmorize_variable(
        ds,
        "tas",
        mapping_table={"tas": {"varname": "TEMP2"}},
        cmor_table=os.path.join(table_dir, "CORDEX-CMIP6_mon.json"),
        dataset_table=os.path.join(table_dir, "CORDEX-CMIP6_remo_example.json"),
        grids_table=os.path.join(table_dir, "CORDEX-CMIP6_grids.json"),
        CORDEX_domain="EUR-11",
        time_units=None,
        allow_units_convert=True,
    )
    return filename


def test_cmorizer_subdaily(table):
    """
    CMORize sub-daily (1hr, 3hr, 6hr) variables.

    This example demonstrates how to cmorize sub-daily data. The input data
    is hourly temperature, which can be automatically resampled to other
    frequencies (3-hourly, 6-hourly, daily) based on the target CMOR table.

    Parameters
    ----------
    table : str
        Name of the time frequency table (e.g., '1hr', '3hr', '6hr', 'day')

    Returns
    -------
    str
        Path to the cmorized netCDF file
    """
    ds = cx.tutorial.open_dataset("remo_EUR-11_TEMP2_1hr")
    filename = cmor.cmorize_variable(
        ds,
        "tas",
        mapping_table={"tas": {"varname": "TEMP2"}},
        cmor_table=os.path.join(table_dir, f"CORDEX-CMIP6_{table}.json"),
        dataset_table=os.path.join(table_dir, "CORDEX-CMIP6_remo_example.json"),
        grids_table=os.path.join(table_dir, "CORDEX-CMIP6_grids.json"),
        CORDEX_domain="EUR-11",
        time_units=None,
        allow_units_convert=True,
        allow_resample=True,  # Enable automatic resampling to target frequency
    )
    return filename

Examples of CORDEX-CMIP6 cmorization

Contents

Examples of CORDEX-CMIP6 cmorization#

Introduction#

Controlled Vocabulary (CV)#

CMORization Examples#

Example 1: Fixed Fields (fx) - Topography#

Generated Filename#

Interactive Dataset Exploration with xarray#

Example 2: Monthly Aggregated Data (mon) - Temperature#

Example 3: Sub-daily Data (1hr, 3hr, 6hr, day) - Temperature#