Year zero handling

The CMIP6 historical concentration data files use a gregorian calendar which has a reference year of zero. There is no year zero in a gregorian calendar so this case cannot be handled by iris. As a result, we provide a simple wrapper to handle this edge case. Note, as we have to read in the entire data file, it can be slow.

# NBVAL_IGNORE_OUTPUT
import datetime

import iris
import iris.coord_categorisation
import iris.plot as iplt
from netcdf_scm.iris_cube_wrappers import CMIP6Input4MIPsCube

import matplotlib.pyplot as plt

plt.style.use("bmh")

# NBVAL_IGNORE_OUTPUT
cmip6_hist_concs = CMIP6Input4MIPsCube()
cmip6_hist_concs.load_data_from_identifiers(
    root_dir="../../../tests/test-data/cmip6input4mips",
    activity_id="input4MIPs",
    mip_era="CMIP6",
    target_mip="CMIP",
    institution_id="UoM",
    source_id="UoM-CMIP-1-2-0",
    realm="atmos",
    frequency="mon",
    variable_id="mole-fraction-of-carbon-dioxide-in-air",
    grid_label="gr1-GMNHSH",
    version="v20100304",
    dataset_category="GHGConcentrations",
    time_range="000001-201412",
    file_ext=".nc",
)

# NBVAL_IGNORE_OUTPUT
cmip6_hist_concs.cube

Mole (1.e-6)	time	sector
Shape	24168	3
Dimension coordinates
time	x	-
sector	-	x
Attributes
Conventions	CF-1.6
activity_id	input4MIPs
comment	Data provided are global and hemispheric area-weighted means. Zonal means...
contact	malte.meinshausen@unimelb.edu.au
creation_date	2016-08-30T18:22:16Z
dataset_category	GHGConcentrations
dataset_version_number	1.2.0
frequency	mon
further_info_url	http://climatecollege.unimelb.edu.au/cmip6
grid	global and hemispheric means - area-averages from the original latitudinal...
grid_label	gr1-GMNHSH
institution	Australian-German Climate & Energy College, The University of Melbourne...
institution_id	UoM
license	GHG concentrations produced by UoM are licensed under a Creative Commons...
mip_era	CMIP6
nominal_resolution	10000 km
product	assimilated observations
realm	atmos
references	Malte Meinshausen, Elisabeth Vogel, Alexander Nauels, Katja Lorbacher,...
source	UoM-CMIP-1-2-0: Historical GHG mole fractions from NOAA & AGAGE networks...
source_id	UoM-CMIP-1-2-0
table_id	input4MIPs
target_mip	CMIP
title	UoM-CMIP-1-2-0: historical GHG concentrations: global and hemispheric means...
tracking_id	hdl:21.14100/3ef0a11f-2ed2-4004-9234-4087c2d41cee
variable_id	mole_fraction_of_carbon_dioxide_in_air
Cell methods
mean	time
mean	area

We also make a plot to show how the underlying data looks.

# NBVAL_IGNORE_OUTPUT
yearmin = 1700
cube = cmip6_hist_concs.cube.extract(
    iris.Constraint(time=lambda t: t[0].year > yearmin)
)

fig = plt.figure(figsize=(16, 9))

for i in range(3):
    region = cube.extract(iris.Constraint(sector=i))
    for title in region.coord("sector").attributes["ids"].split(";"):
        if title.strip().startswith(str(i)):
            title = title.split(":")[1].strip()
            break

    if "Global" in title:
        plt.subplot(322)
    elif "Northern" in title:
        plt.subplot(324)
    elif "Southern" in title:
        plt.subplot(326)

    iplt.plot(region, lw=2.0)
    xlabel = "Time"
    plt.title(title)
    plt.xlabel(xlabel)
    plt.xlim(
        [datetime.date(1965, 1, 1), datetime.date(2015, 1, 1),]
    )

    if "Global" in title:
        plt.subplot(121)

        iris.coord_categorisation.add_year(region, "time", name="year")
        region_annual_mean = region.aggregated_by(["year"], iris.analysis.MEAN)

        iplt.plot(region_annual_mean, lw=2.0)

        var_name = region.var_name.replace("_", " ")
        var_name = var_name.replace("in", "\nin")
        plt.ylabel("{} ({})".format(var_name, region.units))
        plt.title(title + "-annual mean")
        plt.xlabel("Time")

plt.tight_layout();