MONET Accessors¶

MONET adds georeferencing and analysis tools to xarray and pandas data structures through accessor mechanisms. These tools can be accessed via a special .monet attribute, available for xarray.DataArray, xarray.Dataset, and pandas.DataFrame objects after importing monet.

Initializing the Accessor¶

All you have to do is import monet and xarray.

import monet
import xarray as xr

# Example opening a dataset (requires monetio or a standard NetCDF)
# ds = xr.open_dataset('model_output.nc')

# After import, the .monet accessor is available
# ds.monet.quick_map()

Interpolation Methods¶

The MONET accessor provides several useful interpolation routines including:

Finding the nearest point to a given latitude and longitude
Interpolating to a constant latitude or longitude
Vertical interpolation (stratification)
Remapping entire 2D DataArray or Dataset using xregrid.

Find Nearest Point¶

To find the nearest latitude/longitude point, use the nearest_latlon method.

# Find the closest grid point to lat=20.5, lon=-157.4
ds.monet.nearest_latlon(lat=20.5, lon=-157.4)

If you wanted to only find the nearest location for a single variable, you can use the accessor on the DataArray:

ds.O3.monet.nearest_latlon(lat=20.5, lon=-157.4)

Comparison and Statistics¶

The compare method allows for quick comparison between two DataArray objects. It automatically aligns the data and can compute various statistics or a simple difference.

# Compute difference (da1 - da2) and plot automatically
ds.O3.monet.compare(obs.O3, stat='diff', plot=True)

# Compute RMSE and return a DataArray (lazily if using Dask)
rmse_da = ds.O3.monet.compare(obs.O3, stat='RMSE', plot=False)

Supported statistics include metrics from monet-stats (e.g., RMSE, MB, MAE, NMB, IOA).

Faceted Time Map¶

Create a grid of map plots for each time slice of a variable.

# Create a 4-column facet grid of Ozone maps
fig, axes = ds.O3.monet.quick_facet_time_map(ncols=4)

Pairing Data¶

The .pair() method provides a high-level interface for matching model data with observations. It is available on all MONET accessors (DataArray, Dataset, and DataFrame).

# Pair model Dataset with observation DataFrame
paired = ds.monet.pair(obs_df)

# Pair model DataArray with observation Dataset (e.g. for trajectory)
paired_traj = ds.O3.monet.pair(obs_ds, interp_time=True)

Convention-Aware Coordinate Detection¶

MONET is designed to work with various data conventions without requiring destructive renaming of dimensions or coordinates. The accessor automatically detects latitude and longitude based on common naming patterns, CF standard names, units, and UGRID mesh topologies.

Supported Naming Patterns¶

MONET searches for a wide range of common names including: - Latitude: latitude, lat, y, XLAT, grid_yt, nav_lat, etc. - Longitude: longitude, lon, x, XLONG, grid_xt, nav_lon, etc.

CF Convention Support¶

MONET also checks for: - standard_name attributes like latitude or grid_latitude. - units attributes such as degrees_north or degrees_east.

UGRID Convention Support¶

For unstructured grids, MONET detects the mesh_topology variable (using cf_role="mesh_topology") and identifies coordinates using UGRID attributes: - node_coordinates - face_coordinates - edge_coordinates

Accessing Coordinates¶

You can access the detected latitude and longitude objects through the .monet.lat and .monet.lon properties:

# Access detected latitude coordinate/variable
lat_da = ds.monet.lat

# Access detected longitude coordinate/variable
lon_da = ds.monet.lon

These properties work for both Xarray (Dataset/DataArray) and Pandas (DataFrame) objects.

Utility Methods¶

Standardize¶

The standardize method ensures that coordinates have standard CF attributes and that longitudes are wrapped to the [-180, 180) range. It also updates the dataset's history.

ds_std = ds.monet.standardize()

Wrap Longitudes¶

Explicitly wrap longitudes to [-180, 180) without modifying other attributes.

ds_wrapped = ds.monet.wrap_longitudes()

Tidy¶

Applies both longitude wrapping and sorting by longitude.

ds_tidied = ds.monet.tidy()

Comparison and Analysis¶

Vertical Interpolation¶

The interpolate_vertical method provides vertical interpolation using pytspack tension splines, following the same API as pytspack.interpolate_vertical.

# Interpolate O3 to specific altitude levels
# 'altitude' is the name of the vertical dimension in the dataset
interped = ds.O3.monet.interpolate_vertical([100, 500, 1000], level_dim='altitude')

For Dataset objects, interpolate_vertical will interpolate all variables that have the specified vertical dimension.

# Interpolate entire dataset to pressure levels
ds_interp = ds.monet.interpolate_vertical([850, 700, 500], level_dim='pres')

Note

stratify() is still available as a deprecated backward-compatibility alias.

Land and Ocean Masking¶

MONET includes built-in support for land/ocean masking based on Natural Earth data.

# Get a boolean mask (True for land)
is_land = ds.monet.is_land()

# Return the dataset masked by ocean (preserving only ocean points)
ds_ocean = ds.monet.is_ocean(return_xarray=True)