Inspect data¶

In [1]:

import logging

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd

from ghg_forcing_for_cmip import download_ground_based, download_satellite, plotting
from ghg_forcing_for_cmip.validation import compute_discrepancy_collocated

# silent prefect run
logging.getLogger("prefect").setLevel(logging.ERROR)

# run pipeline to download data
for gas in ["co2", "ch4"]:
    download_ground_based.download_surface_data(gas=gas, remove_original_files=False)
    download_satellite.download_satellite_data(gas=gas, remove_original_files=False)

import logging import matplotlib.pyplot as plt import numpy as np import pandas as pd from ghg_forcing_for_cmip import download_ground_based, download_satellite, plotting from ghg_forcing_for_cmip.validation import compute_discrepancy_collocated # silent prefect run logging.getLogger("prefect").setLevel(logging.ERROR) # run pipeline to download data for gas in ["co2", "ch4"]: download_ground_based.download_surface_data(gas=gas, remove_original_files=False) download_satellite.download_satellite_data(gas=gas, remove_original_files=False)

/home/docs/checkouts/readthedocs.org/user_builds/ghg-forcing-for-cmip/envs/latest/lib/python3.11/site-packages/pydantic_settings/main.py:426: UserWarning: Config key `pyproject_toml_table_header` is set in model_config but will be ignored because no PyprojectTomlConfigSettingsSource source is configured. To use this config key, add a PyprojectTomlConfigSettingsSource source to the settings sources via the settings_customise_sources hook.
  self._settings_warn_unused_config_keys(sources, self.model_config)
/home/docs/checkouts/readthedocs.org/user_builds/ghg-forcing-for-cmip/envs/latest/lib/python3.11/site-packages/pydantic_settings/main.py:426: UserWarning: Config key `toml_file` is set in model_config but will be ignored because no TomlConfigSettingsSource source is configured. To use this config key, add a TomlConfigSettingsSource source to the settings sources via the settings_customise_sources hook.
  self._settings_warn_unused_config_keys(sources, self.model_config)

/home/docs/checkouts/readthedocs.org/user_builds/ghg-forcing-for-cmip/envs/latest/lib/python3.11/site-packages/pydantic/_internal/_generate_schema.py:2249: UnsupportedFieldAttributeWarning: The 'default' attribute with value 'UTC' was provided to the `Field()` function, which has no effect in the context it was used. 'default' is field-specific metadata, and can only be attached to a model field using `Annotated` metadata or by assignment. This may have happened because an `Annotated` type alias using the `type` statement was used, or if the `Field()` function was attached to a single member of a union type.
  warnings.warn(

/home/docs/checkouts/readthedocs.org/user_builds/ghg-forcing-for-cmip/envs/latest/lib/python3.11/site-packages/pydantic/_internal/_generate_schema.py:2249: UnsupportedFieldAttributeWarning: The 'default' attribute with value 'UTC' was provided to the `Field()` function, which has no effect in the context it was used. 'default' is field-specific metadata, and can only be attached to a model field using `Annotated` metadata or by assignment. This may have happened because an `Annotated` type alias using the `type` statement was used, or if the `Field()` function was attached to a single member of a union type.
  warnings.warn(

/home/docs/checkouts/readthedocs.org/user_builds/ghg-forcing-for-cmip/envs/latest/lib/python3.11/site-packages/pydantic_settings/main.py:426: UserWarning: Config key `pyproject_toml_table_header` is set in model_config but will be ignored because no PyprojectTomlConfigSettingsSource source is configured. To use this config key, add a PyprojectTomlConfigSettingsSource source to the settings sources via the settings_customise_sources hook.
  self._settings_warn_unused_config_keys(sources, self.model_config)
/home/docs/checkouts/readthedocs.org/user_builds/ghg-forcing-for-cmip/envs/latest/lib/python3.11/site-packages/pydantic_settings/main.py:426: UserWarning: Config key `toml_file` is set in model_config but will be ignored because no TomlConfigSettingsSource source is configured. To use this config key, add a TomlConfigSettingsSource source to the settings sources via the settings_customise_sources hook.
  self._settings_warn_unused_config_keys(sources, self.model_config)

In [2]:

d_gb_ch4 = pd.read_csv("data/downloads/ch4/ch4_gb_raw.csv")
d_gb_co2 = pd.read_csv("data/downloads/co2/co2_gb_raw.csv")
d_gb_ch4.head()

d_gb_ch4 = pd.read_csv("data/downloads/ch4/ch4_gb_raw.csv") d_gb_co2 = pd.read_csv("data/downloads/co2/co2_gb_raw.csv") d_gb_ch4.head()

Out[2]:

	year	month	latitude	longitude	altitude	value	std_dev	numb	site_code	network	...	unit	version	instrument	lat	lon	bnd	lat_bnd	lon_bnd	time_fractional	time
0	1983	1	-75.604202	-26.601500	13.0	1560.179199	2.809700	216	HBA	noaa	...	ppb	2025-08-15T13:48:06.496334	noaa	-77.5	-27.5	0	-75	-25	1983.083333	1983-01-15 12:00:00+00:00
1	1983	1	-75.604202	-26.601500	13.0	1560.179199	2.809700	216	HBA	noaa	...	ppb	2025-08-15T13:48:06.496334	noaa	-77.5	-27.5	1	-80	-30	1983.083333	1983-01-15 12:00:00+00:00
2	1983	1	-64.774200	-64.052696	13.0	1557.342407	8.102857	180	PSA	noaa	...	ppb	2025-08-15T13:48:06.496334	noaa	-62.5	-62.5	0	-60	-60	1983.083333	1983-01-15 12:00:00+00:00
3	1983	1	-64.774200	-64.052696	13.0	1557.342407	8.102857	180	PSA	noaa	...	ppb	2025-08-15T13:48:06.496334	noaa	-62.5	-62.5	1	-65	-65	1983.083333	1983-01-15 12:00:00+00:00
4	1983	1	-37.950001	77.529999	153.0	1558.405273	10.514277	144	AMS	noaa	...	ppb	2025-08-15T13:48:06.496334	noaa	-37.5	77.5	0	-35	75	1983.083333	1983-01-15 12:00:00+00:00

5 rows × 23 columns

In [3]:

d_eo_ch4 = pd.read_csv("data/downloads/ch4/ch4_eo_raw.csv")
d_eo_co2 = pd.read_csv("data/downloads/co2/co2_eo_raw.csv")
d_eo_ch4.head()

d_eo_ch4 = pd.read_csv("data/downloads/ch4/ch4_eo_raw.csv") d_eo_co2 = pd.read_csv("data/downloads/co2/co2_eo_raw.csv") d_eo_ch4.head()

Out[3]:

	time	year	month	lat_bnd	lon_bnd	bnd	lat	lon	value	std_dev	numb	gas	unit	pre	column_averaging_kernel	vmr_profile_apriori
0	2003-01-16 12:00:00+00:00	2003	1	-85	-120	0	-82.5	-117.5	1611.184302	47.167322	17	ch4	ppb	0.95	1.188702	1714.261884
1	2003-01-16 12:00:00+00:00	2003	1	-85	-120	0	-82.5	-117.5	1611.184302	47.167322	17	ch4	ppb	0.85	1.156435	1713.341703
2	2003-01-16 12:00:00+00:00	2003	1	-85	-120	0	-82.5	-117.5	1611.184302	47.167322	17	ch4	ppb	0.75	1.115912	1712.569428
3	2003-01-16 12:00:00+00:00	2003	1	-85	-120	0	-82.5	-117.5	1611.184302	47.167322	17	ch4	ppb	0.65	1.073409	1712.066933
4	2003-01-16 12:00:00+00:00	2003	1	-85	-120	0	-82.5	-117.5	1611.184302	47.167322	17	ch4	ppb	0.55	1.017909	1711.855703

Coverage¶

Ground-based data¶

In [4]:

_, axs = plt.subplots(1, 2, constrained_layout=True, figsize=(11, 6))
plotting.plot_map(
    d_gb_ch4,
    # Don't need the +, just put the strings in brackets and
    # they're automatically joined
    (
        f"Observations sites ({d_gb_ch4.year.min()}-{d_gb_ch4.year.max()})"
        "\n $CH_4$ data"
    ),
    axs[0],
)
plotting.plot_map(
    d_gb_co2,
    (
        f"Observations sites ({d_gb_co2.year.min()}-{d_gb_co2.year.max()})"
        "\n $CO_2$ data"
    ),
    axs[1],
)

_, axs = plt.subplots(1, 2, constrained_layout=True, figsize=(11, 6)) plotting.plot_map( d_gb_ch4, # Don't need the +, just put the strings in brackets and # they're automatically joined ( f"Observations sites ({d_gb_ch4.year.min()}-{d_gb_ch4.year.max()})" "\n $CH_4$ data" ), axs[0], ) plotting.plot_map( d_gb_co2, ( f"Observations sites ({d_gb_co2.year.min()}-{d_gb_co2.year.max()})" "\n $CO_2$ data" ), axs[1], )

Downloading file 'ne_110m_land.zip' from 'https://naciscdn.org/naturalearth/110m/physical/ne_110m_land.zip' to '/home/docs/.cache/geodatasets'.

Out[4]:

<Axes: title={'center': 'Observations sites (1968-2024)\n $CO_2$ data'}>

Satellite data¶

Methane¶

In [5]:

_, axs = plt.subplots(2, 3, constrained_layout=True, figsize=(9, 3))

# Zip plus `.flatten` lets you avoid this enumerate thinking
# and double loop
for ax, year in zip(axs.flatten(), [2003, 2008, 2013, 2018, 2020, 2022]):
    plotting.plot_map(
        d_eo_ch4[d_eo_ch4.year == year],
        f"OBS4MIPs ({year}), $CH_4$ data",
        ax,
        "lon",
        "lat",
        ".",
        5,
    )

_, axs = plt.subplots(2, 3, constrained_layout=True, figsize=(9, 3)) # Zip plus `.flatten` lets you avoid this enumerate thinking # and double loop for ax, year in zip(axs.flatten(), [2003, 2008, 2013, 2018, 2020, 2022]): plotting.plot_map( d_eo_ch4[d_eo_ch4.year == year], f"OBS4MIPs ({year}), $CH_4$ data", ax, "lon", "lat", ".", 5, )

Carbon Dioxide¶

In [6]:

_, axs = plt.subplots(2, 3, constrained_layout=True, figsize=(9, 3))
# Zip plus `.flatten` lets you avoid this enumerate thinking
# and double loop
for ax, year in zip(axs.flatten(), [2003, 2008, 2013, 2018, 2020, 2022]):
    plotting.plot_map(
        d_eo_co2[d_eo_co2.year == year],
        f"OBS4MIPs ({year}), $CO_2$ data",
        ax,
        "lon",
        "lat",
        ".",
        5,
    )

_, axs = plt.subplots(2, 3, constrained_layout=True, figsize=(9, 3)) # Zip plus `.flatten` lets you avoid this enumerate thinking # and double loop for ax, year in zip(axs.flatten(), [2003, 2008, 2013, 2018, 2020, 2022]): plotting.plot_map( d_eo_co2[d_eo_co2.year == year], f"OBS4MIPs ({year}), $CO_2$ data", ax, "lon", "lat", ".", 5, )

Comparison of raw ground-based data with satellite product¶

In [7]:

# select collocated sites
d_gb_ch4_sel = d_gb_ch4[
    ["year", "month", "lat", "lon", "value", "site_code"]
].drop_duplicates()
d_eo_ch4_sel = d_eo_ch4[
    ["year", "month", "lat", "lon", "value", "pre", "vmr_profile_apriori"]
].drop_duplicates()
d_gb_co2_sel = d_gb_co2[
    ["year", "month", "lat", "lon", "value", "site_code"]
].drop_duplicates()
d_eo_co2_sel = d_eo_co2[
    ["year", "month", "lat", "lon", "value", "pre", "vmr_profile_apriori"]
].drop_duplicates()

d_colloc_ch4 = d_gb_ch4_sel.merge(
    d_eo_ch4_sel,
    on=["year", "month", "lat", "lon"],
    how="inner",
    suffixes=["_gb", "_eo"],
)
d_colloc_co2 = d_gb_co2_sel.merge(
    d_eo_co2_sel,
    on=["year", "month", "lat", "lon"],
    how="inner",
    suffixes=["_gb", "_eo"],
)

d_colloc_ch4.to_csv("data/downloads/ch4/ch4_collocated.csv")
d_colloc_co2.to_csv("data/downloads/co2/co2_collocated.csv")
d_colloc_ch4.head()

# select collocated sites d_gb_ch4_sel = d_gb_ch4[ ["year", "month", "lat", "lon", "value", "site_code"] ].drop_duplicates() d_eo_ch4_sel = d_eo_ch4[ ["year", "month", "lat", "lon", "value", "pre", "vmr_profile_apriori"] ].drop_duplicates() d_gb_co2_sel = d_gb_co2[ ["year", "month", "lat", "lon", "value", "site_code"] ].drop_duplicates() d_eo_co2_sel = d_eo_co2[ ["year", "month", "lat", "lon", "value", "pre", "vmr_profile_apriori"] ].drop_duplicates() d_colloc_ch4 = d_gb_ch4_sel.merge( d_eo_ch4_sel, on=["year", "month", "lat", "lon"], how="inner", suffixes=["_gb", "_eo"], ) d_colloc_co2 = d_gb_co2_sel.merge( d_eo_co2_sel, on=["year", "month", "lat", "lon"], how="inner", suffixes=["_gb", "_eo"], ) d_colloc_ch4.to_csv("data/downloads/ch4/ch4_collocated.csv") d_colloc_co2.to_csv("data/downloads/co2/co2_collocated.csv") d_colloc_ch4.head()

Out[7]:

	year	month	lat	lon	value_gb	site_code	value_eo	pre	vmr_profile_apriori
0	2003	1	-52.5	-67.5	1722.262451	USH	1631.436021	0.95	1725.226070
1	2003	1	-52.5	-67.5	1722.262451	USH	1631.436021	0.85	1720.670639
2	2003	1	-52.5	-67.5	1722.262451	USH	1631.436021	0.75	1717.441933
3	2003	1	-52.5	-67.5	1722.262451	USH	1631.436021	0.65	1715.137159
4	2003	1	-52.5	-67.5	1722.262451	USH	1631.436021	0.55	1714.451287

In [8]:

fig, axs = plt.subplots(1, 2, figsize=(11, 3), constrained_layout=True)
plotting.plot_monthly_average(d_colloc_ch4, "$CH_4$", axs[0])
plotting.plot_monthly_average(d_colloc_co2, "$CO_2$", axs[1])
fig.suptitle("Annual, global average of GHG concentration")

fig, axs = plt.subplots(1, 2, figsize=(11, 3), constrained_layout=True) plotting.plot_monthly_average(d_colloc_ch4, "$CH_4$", axs[0]) plotting.plot_monthly_average(d_colloc_co2, "$CO_2$", axs[1]) fig.suptitle("Annual, global average of GHG concentration")

Out[8]:

Text(0.5, 0.98, 'Annual, global average of GHG concentration')

In [9]:

plotting.plot_average_hemisphere(d_colloc_ch4, "$CH_4$")

plotting.plot_average_hemisphere(d_colloc_ch4, "$CH_4$")

In [10]:

plotting.plot_average_hemisphere(d_colloc_co2, "$CO_2$")

plotting.plot_average_hemisphere(d_colloc_co2, "$CO_2$")

In [11]:

d_col_sorted_co2 = (
    d_colloc_co2.groupby(["lat", "lon", "site_code"])
    .agg({"value_gb": "count"})
    .reset_index()
    .sort_values(by="value_gb", ascending=False)
)
d_col_sorted_ch4 = (
    d_colloc_ch4.groupby(["lat", "lon", "site_code"])
    .agg({"value_gb": "count"})
    .reset_index()
    .sort_values(by="value_gb", ascending=False)
)
d_col_sorted_ch4.rename(columns={"value_gb": "count"}).head(10)

d_col_sorted_co2 = ( d_colloc_co2.groupby(["lat", "lon", "site_code"]) .agg({"value_gb": "count"}) .reset_index() .sort_values(by="value_gb", ascending=False) ) d_col_sorted_ch4 = ( d_colloc_ch4.groupby(["lat", "lon", "site_code"]) .agg({"value_gb": "count"}) .reset_index() .sort_values(by="value_gb", ascending=False) ) d_col_sorted_ch4.rename(columns={"value_gb": "count"}).head(10)

Out[11]:

	lat	lon	site_code	count
148	17.5	-157.5	MLO	3640
60	-12.5	-172.5	SMO	2960
209	37.5	102.5	WLG	2500
206	37.5	-97.5	SGP	2400
205	37.5	-112.5	UTA	2280
215	42.5	-107.5	NWR	2250
173	22.5	7.5	ASK	2220
220	42.5	112.5	UUM	2060
211	37.5	127.5	TAP	1940
225	47.5	17.5	HUN	1870

In [12]:

max_lat = -50.0
d_col_sorted_co2[d_col_sorted_co2["lat"] <= max_lat].rename(
    columns={"value_gb": "count"}
).head(2)

max_lat = -50.0 d_col_sorted_co2[d_col_sorted_co2["lat"] <= max_lat].rename( columns={"value_gb": "count"} ).head(2)

Out[12]:

	lat	lon	site_code	count
5	-52.5	-67.5	USH	460
2	-57.5	-62.5	DRP	210

In [13]:

min_lat = 50.0
d_col_sorted_co2[d_col_sorted_co2["lat"] >= max_lat].rename(
    columns={"value_gb": "count"}
).head(2)

min_lat = 50.0 d_col_sorted_co2[d_col_sorted_co2["lat"] >= max_lat].rename( columns={"value_gb": "count"} ).head(2)

Out[13]:

	lat	lon	site_code	count
43	-12.5	-172.5	SMO	3070
103	17.5	-157.5	MLO	2690

In [14]:

sites_selected_co2 = [
    "SMO",
    "MLO",
    "USH",
    "OXK",
]
sites_selected_ch4 = [
    "SMO",
    "MLO",
    "USH",
    "BRW",
]

sel_co2 = d_col_sorted_co2[d_col_sorted_co2.site_code.isin(sites_selected_co2)]
sel_ch4 = d_col_sorted_ch4[d_col_sorted_ch4.site_code.isin(sites_selected_ch4)]

_, axs = plt.subplots(1, 2, constrained_layout=True, figsize=(8, 5))

ax1 = plotting.plot_map(
    sel_co2,
    "Selected sites - CO2",
    lon_value="lon",
    lat_value="lat",
    axs=axs[0],
)
ax2 = plotting.plot_map(
    sel_ch4,
    "Selected sites - CH4",
    lon_value="lon",
    lat_value="lat",
    axs=axs[1],
)

for d, ax in zip([sel_co2, sel_ch4], [ax1, ax2]):
    for lat, lon, label in zip(d.lat, d.lon, d.site_code):
        ax.text(
            lon + 15,
            lat + 5,
            label,
            fontsize=8,
            fontweight="bold",
            ha="right",
            va="bottom",
            color="black",
        )

sites_selected_co2 = [ "SMO", "MLO", "USH", "OXK", ] sites_selected_ch4 = [ "SMO", "MLO", "USH", "BRW", ] sel_co2 = d_col_sorted_co2[d_col_sorted_co2.site_code.isin(sites_selected_co2)] sel_ch4 = d_col_sorted_ch4[d_col_sorted_ch4.site_code.isin(sites_selected_ch4)] _, axs = plt.subplots(1, 2, constrained_layout=True, figsize=(8, 5)) ax1 = plotting.plot_map( sel_co2, "Selected sites - CO2", lon_value="lon", lat_value="lat", axs=axs[0], ) ax2 = plotting.plot_map( sel_ch4, "Selected sites - CH4", lon_value="lon", lat_value="lat", axs=axs[1], ) for d, ax in zip([sel_co2, sel_ch4], [ax1, ax2]): for lat, lon, label in zip(d.lat, d.lon, d.site_code): ax.text( lon + 15, lat + 5, label, fontsize=8, fontweight="bold", ha="right", va="bottom", color="black", )

In [15]:

d_col_sel_co2 = d_colloc_co2[
    d_colloc_co2.site_code.isin(sites_selected_co2)
].reset_index(drop=True)
d_col_sel_ch4 = d_colloc_ch4[
    d_colloc_ch4.site_code.isin(sites_selected_ch4)
].reset_index(drop=True)

d_col_sel_co2.to_csv("data/downloads/co2/co2_collocated_sites.csv")
d_col_sel_ch4.to_csv("data/downloads/ch4/ch4_collocated_sites.csv")

d_col_sel_co2 = d_colloc_co2[ d_colloc_co2.site_code.isin(sites_selected_co2) ].reset_index(drop=True) d_col_sel_ch4 = d_colloc_ch4[ d_colloc_ch4.site_code.isin(sites_selected_ch4) ].reset_index(drop=True) d_col_sel_co2.to_csv("data/downloads/co2/co2_collocated_sites.csv") d_col_sel_ch4.to_csv("data/downloads/ch4/ch4_collocated_sites.csv")

In [16]:

fig, axs = plt.subplots(2, 4, figsize=(11, 4), constrained_layout=True)

# Same zip trick as above
for site, ax in zip(sites_selected_co2, axs[0, :]):
    d = d_col_sel_co2[d_col_sel_co2.site_code == site].copy()
    rmse_co2 = compute_discrepancy_collocated(d, "co2", "rmse")
    ax = plotting.plot_monthly_average(d, "$CO_2$", ax)  # noqa: PLW2901
    ax.set_title(
        # As above re ability to use brackets instead of +
        # (although this is really very optional
        # and probably more a style choice than there being a right answer)
        site
        + f", RMSE: {rmse_co2['rmse_co2'].values[0]:.2f},"
        + f"\nBias: {rmse_co2['bias_co2'].values[0]:.2f}, "
        + f"SD: {np.sqrt(rmse_co2['var_co2'].values[0]):.2f}",
        fontsize="small",
    )
    ax.legend(fontsize="x-small", handlelength=0.4, frameon=False)

# zip could be used here too
for i, site in enumerate(sites_selected_ch4):
    d = d_col_sel_ch4[d_col_sel_ch4.site_code == site].copy()
    rmse_ch4 = compute_discrepancy_collocated(d, "ch4", "rmse")
    axs[1, i] = plotting.plot_monthly_average(d, "$CH_4$", axs[1, i])
    axs[1, i].set_title(
        site
        + f", RMSE: {rmse_ch4['rmse_ch4'].values[0]:.2f},"
        + f"\nBias: {rmse_ch4['bias_ch4'].values[0]:.2f}, "
        + f"SD: {np.sqrt(rmse_ch4['var_ch4'].values[0]):.2f}",
        fontsize="small",
    )
    axs[1, i].legend(fontsize="x-small", handlelength=0.4, frameon=False)

fig, axs = plt.subplots(2, 4, figsize=(11, 4), constrained_layout=True) # Same zip trick as above for site, ax in zip(sites_selected_co2, axs[0, :]): d = d_col_sel_co2[d_col_sel_co2.site_code == site].copy() rmse_co2 = compute_discrepancy_collocated(d, "co2", "rmse") ax = plotting.plot_monthly_average(d, "$CO_2$", ax) # noqa: PLW2901 ax.set_title( # As above re ability to use brackets instead of + # (although this is really very optional # and probably more a style choice than there being a right answer) site + f", RMSE: {rmse_co2['rmse_co2'].values[0]:.2f}," + f"\nBias: {rmse_co2['bias_co2'].values[0]:.2f}, " + f"SD: {np.sqrt(rmse_co2['var_co2'].values[0]):.2f}", fontsize="small", ) ax.legend(fontsize="x-small", handlelength=0.4, frameon=False) # zip could be used here too for i, site in enumerate(sites_selected_ch4): d = d_col_sel_ch4[d_col_sel_ch4.site_code == site].copy() rmse_ch4 = compute_discrepancy_collocated(d, "ch4", "rmse") axs[1, i] = plotting.plot_monthly_average(d, "$CH_4$", axs[1, i]) axs[1, i].set_title( site + f", RMSE: {rmse_ch4['rmse_ch4'].values[0]:.2f}," + f"\nBias: {rmse_ch4['bias_ch4'].values[0]:.2f}, " + f"SD: {np.sqrt(rmse_ch4['var_ch4'].values[0]):.2f}", fontsize="small", ) axs[1, i].legend(fontsize="x-small", handlelength=0.4, frameon=False)

In [17]:

plotting.plot_collocated_rmse(d_colloc_co2, d_colloc_ch4, "rmse")

plotting.plot_collocated_rmse(d_colloc_co2, d_colloc_ch4, "rmse")

Out[17]:

array([<Axes: title={'center': 'Carbon Dioxide (avg. RMSE:3.47)'}, xlabel='RMSE', ylabel='CO2 in ppm'>,
       <Axes: title={'center': 'Methane (avg. RMSE:77.63)'}, xlabel='RMSE', ylabel='CH4 in ppb'>],
      dtype=object)