Introduction to Pyresample

https://pyresample.readthedocs.io/en/latest/

This package seems a bit more speciallized, and does not have as tight of integration with xarray like xESMF and Verde do. If working with satellite or swath data, this is not one to miss! This package integrates with Satpy https://satpy.readthedocs.io/en/stable/ .

(more) Integration with xarray

This is on the to-do list: https://pyresample.readthedocs.io/en/latest/roadmap.html?highlight=xarray#xarray-and-geoxarray

Prerequisites

Knowing your way around xarray, numpy is beneficial. This is not deisgned to be an introduction to any of those packages. Would do this notebook after doing the xESMF one!

Imports

import pandas as pd
import numpy as np
import xarray as xr

import matplotlib.pyplot as plt

from appdirs import *

import dask.array as da

import pyresample
from pyresample import image, geometry
from pyresample.bilinear import NumpyBilinearResampler

from xarray import DataArray
from pyresample.bilinear import XArrayBilinearResampler

import os

%load_ext watermark
%watermark --iversions

pandas    : 2.2.2
dask      : 2024.5.2
numpy     : 1.26.4
matplotlib: 3.8.4
sys       : 3.10.14 | packaged by conda-forge | (main, Mar 20 2024, 12:45:18) [GCC 12.3.0]
pyresample: 1.28.3
xarray    : 2024.5.0

Resampling of gridded data using pyresample

Link to this turtorial is here: https://pyresample.readthedocs.io/en/latest/swath.html#pyresample-bilinear

We will be deconstructing it a bit to get into the details, but all of the code is from the above link.

target_def = geometry.AreaDefinition('areaD',
                                     'Europe (3km, HRV, VTC)',
                                     'areaD',
                                     {'a': '6378144.0', 'b': '6356759.0',
                                      'lat_0': '50.00', 'lat_ts': '50.00',
                                      'lon_0': '8.00', 'proj': 'stere'},
                                     800, 800,
                                     [-1370912.72, -909968.64,
                                      1029087.28, 1490031.36])

Unlike using xESMF, this does not depend or work with xarray:

print('target def type', type(target_def))

target def type <class 'pyresample.geometry.AreaDefinition'>

data = DataArray(da.from_array(np.fromfunction(lambda y, x: y*x, (500, 100))), dims=('y', 'x'))
type(data)

xarray.core.dataarray.DataArray

lons = da.from_array(np.fromfunction(lambda y, x: 3 + x * 0.1, (500, 100)))
lats = da.from_array(np.fromfunction(lambda y, x: 75 - y * 0.1, (500, 100)))
source_def = geometry.SwathDefinition(lons=lons, lats=lats)

resampler = XArrayBilinearResampler(source_def, target_def, 30e3)
result = resampler.resample(data)
type(result)

/home/runner/miniconda3/envs/gridding-cookbook-dev/lib/python3.10/site-packages/pyproj/crs/crs.py:1293: UserWarning: You will likely lose important projection information when converting to a PROJ string from another format. See: https://proj.org/faq.html#what-is-the-best-format-for-describing-coordinate-reference-systems
  proj = self._crs.to_proj4(version=version)

xarray.core.dataarray.DataArray

Can export to xarray

result.to_dataset()

<xarray.Dataset> Size: 5MB
Dimensions:                                   (x: 800, y: 800)
Coordinates:
  * x                                         (x) float64 6kB -1.369e+06 ... ...
  * y                                         (y) float64 6kB 1.489e+06 ... -...
Data variables:
    reshape-c1dd57ad960628aa19ec19ddee8ffa66  (y, x) float64 5MB dask.array<chunksize=(800, 800), meta=np.ndarray>

data.to_dataset()

<xarray.Dataset> Size: 400kB
Dimensions:                                 (y: 500, x: 100)
Dimensions without coordinates: y, x
Data variables:
    array-76c109148983aaf79e69a6d8292615f3  (y, x) float64 400kB dask.array<chunksize=(500, 100), meta=np.ndarray>

Summary

Pyresample is a speciallist program, with strong functionality with satpy. Would reccomend if swath/sat image data is part of your normal workflow. For others, the requirement of the data being CF compliant and API is a hurdle.