Working with CMEMS datasets

You can use dreamcoat to download, process and plot daily surface ocean reanalysis / forecast data from the CMEMS datasets

GLOBAL_ANALYSIS_FORECAST_PHY_001_024 and
GLOBAL_ANALYSIS_FORECAST_BIO_001_028.

These two datasets are collected for the same space and time ranges and merged into a single xarray Dataset for further processing and plotting. The biogeochemical dataset comes at a lower resolution than the physical one so the former is nearest-neighbour interpolated onto the latter's latitude-longitude grid.

New files are only downloaded if they don't already exist locally and only the specified space and time ranges are downloaded.

Download a dataset

username and password should be your login details for the CMEMS data portal.
Don't commit these to a public repo!
They can optionally be added to the .dreamcoat directory as cmems_username.dat and cmems_password.dat.

Physical

Physical data from GLOBAL_ANALYSIS_FORECAST_PHY_001_024:

import dreamcoat as dc

dc.cmems.download_surphys(
    filename=None,
    filepath="",
    date_min=None,
    date_max=None,
    latitude_min=-90,
    latitude_max=90,
    longitude_min=-180,
    longitude_max=180,
    username=None,
    password=None,
    convert_nc=True,
    delete_nc=True,
    get_current_vertical=True,
)

Biogeochemical

Biogeochemical data from GLOBAL_ANALYSIS_FORECAST_BIO_001_028:

dc.cmems.download_surbio(
    filename=None,
    filepath="",
    date_min=None,
    date_max=None,
    latitude_min=-90,
    latitude_max=90,
    longitude_min=-180,
    longitude_max=180,
    username=None,
    password=None,
    convert_nc=True,
    delete_nc=True,
)

Open a dataset

Once both datasets above have been downloaded, to open the files use

surface = dc.cmems.open_surface(
    filepath="",
    date_min=None,
    date_max=None,
    latitude_min=-90,
    latitude_max=90,
    longitude_min=-180,
    longitude_max=180,
    username=None,
    password=None,
    with_current_vertical=False,
)

The filepath specifies where downloaded files can be found.
Dates should be given in %Y-%m-%d format. If left blank, today is used for both.

Plotting functions

Surface maps

At a single time point

surface_map draws a map of a single time point from the CMEMS dataset. You need to extract the single time slice yourself before running the plotting function, as shown below.

surface_time_slice = surface.isel(time=0)  # for example
surface_time_slice = surface.mean('time')  # alternative example

fig, ax = dc.plot.surface_map(
    surface_time_slice,
    fvar,  # string specifying which variable to plot
    ax=None,
    color_zoom_factor=0.9,
    dpi=150,
    figsize=[6.4, 4.8],
    land_visible=True,
    longitude_fmt=":03.0f",
    latitude_fmt=":02.0f",
    map_extent=None,
    map_projection=ccrs.Mercator(),
    quiver_alpha=None,
    quiver_coarsen=10,
    quiver_color="k",
    quiver_visible=False,
    save_extra="",
    save_figure=False,
    save_path="",
    contrast=None,
    ship_distance=None,
    ship_fade_concentric=True,
    ship_lon_lat=None,
    title="",
    vmin=None,
    vmax=None,
)

Loop through all days

To draw maps for every day of data in the dataset (each on a separate figure), use surface_map_daily:

dc.plot.surface_map_daily(data, fvar, **kwargs)

The above automatically sets a consistent vmin and vmax across all the figures for each variable. The kwargs are mostly the same as for surface_map above.

Time series at a point

Show a single variable

surface_timeseries draws a time series with PCHIP interpolation between the points. You first need to make the slice yourself so that time is the only remaining dimension in the dataset, as shown below.

surface_point_slice = surface.sel(
    longitude=10.0, latitude=-38.4, method="nearest"
)  # for example

fig, ax = dc.plot.surface_timeseries(
    surface_point_slice,
    fvar,
    ax=None,
    dpi=150,
    figsize=[6.4, 4.8],
    draw_line=True,
    draw_points=True,
    interpolate_pchip=True,
    show_offset_text=True,
)

Grids with multiple variables

You can draw a grid of all variables of a particular type in a single figure:

# Physical variables
fig, axs = dc.plot.surface_timeseries_grid_phys(
    surface_point_slice, dpi=150, figsize=[9.6, 7.2]
)

# Carbonate system variables
fig, axs = dc.plot.surface_timeseries_grid_co2(
    surface_point_slice, dpi=150, figsize=[9.6, 4.8]
)

# Biological variables
fig, axs = dc.plot.surface_timeseries_grid_bio(
    surface_point_slice, dpi=150, figsize=[9.6, 4.8]
)

# Nutrients
fig, axs = dc.plot.surface_timeseries_grid_nuts(
    surface_point_slice, dpi=150, figsize=[9.6, 4.8]
)

Or all the grids together (but there's currently no way to automatically save the figures that come out):

dc.plot.surface_timeseries_grids(surface_point_slice, dpi=150)

Polar current plot

To draw a time series of current in polar coordinates:

fig, ax = dc.plot.surface_currents(
    surface_point_slice, dpi=150, figsize=[6.4, 4.8]
)