regridcart Regridding lat/lon to local Cartesian coordinates

This package takes care of regridding data defined with latitude/longitude coordinates onto a local Cartesian grid of a fixed resolution. To use it you simply define the domain you want to regrid onto and then call the resample method.

Usage

import regridcart as rc
import xaray as xr

da_src = xr.open_dataarray("...")

target_domain = rc.LocalCartesianDomain(
    central_latitude=lat0,
    central_longitude=lon0,
    l_meridional=1000.0e3,
    l_zonal=3000.0e3,
)

dx = 1.0e3 # new resoluion 1km
da_regridded = rc.resample(target_domain, da=da_src, dx=dx)

The provided data-array is assumed to have latitude/longitude coordinates defined by one of the following:

1. lat and lon coordinates along which the data is aligned, i.e. lat and lon are given as 1D arrays
2. lat and lon are given as auxilliary variables so that the data isn't aligned along the lat/lon directions, but rather the lat and lon of every datapoint is given
3. the data-array has projection information defined in a CF-compliant manner using the grid_mapping attribute (http://cfconventions.org/Data/cf-conventions/cf-conventions-1.7/build/ch05s06.html)
4. the data-array was loaded from a raster-file using rioxarray.open_rasterio so that the projection information is available via da.rio.crs

The package also implements cropping (rc.crop_field_to_domain), plotting domain outline (domain.plot_outline) and can also with data already on a Cartesian grid with rc.CartesianDomain. See notebooks/examples.ipynb for detailed examples.

Installation

regridcart can be installed with pip from pypi, but it relies on cartopy and xesmf which in turn rely on proj and emsf, these can most easily be installed with conda:

conda install xarray cartopy xesmf -c conda-forge
pip install regridcart