[Bug]: NaN handling in regrid2

[lee1043]

What happened?

When NaN value exists in the original field, interpolation using regrid2 replaces NaN values to zero, then interpolates, which introduces error in the resulted field. unmapped_to_nan=True that was discussed in #528 for xesmf is not available for regrid2, wondering if that could be added.

What did you expect to happen? Are there are possible answers you came across?

Original field has NaN values over the ocean, and when interpolate, it should look like one in the middle (example used xesmf). However, regrid2 injects zeros to NaN values before interpolation, then interpolates, resulting zero over land and gradation values along coastlines.

Minimal example code to reproduce the above plot is attached below.

Minimal Complete Verifiable Example (MVCE)

import matplotlib.pyplot as plt
import xcdat as xc

# Generate a sample data field that has value only over ocean
input_data = "ts_Amon_ACCESS1-0_historical_r1i1p1_185001-200512.nc"
input_data2 = "sftlf_fx_ACCESS1-0_amip_r0i0p0.nc"

ds = xc.open_dataset(input_data)
ds2 = xc.open_dataset(input_data2)
sftlf = ds2["sftlf"]
ds["sst"] = ds["ts"].where(sftlf==0)

# Make a target grid to interpolate to
target_grid = xc.create_uniform_grid(-90, 90, 5, 2.5, 357.5, 5)

# regrid
ds_regridded_xesmf = ds.regridder.horizontal("sst", target_grid, tool="xesmf", method="bilinear")
ds_regridded_regrid2 = ds.regridder.horizontal("sst", target_grid, tool="regrid2") # unmapped_to_nan=True is not available for regrid2

fig, axs = plt.subplots(1, 3, figsize=(15, 4))

ds["sst"].isel(time=0).plot(ax=axs[0])
ds_regridded_xesmf["sst"].isel(time=0).plot(ax=axs[1])
ds_regridded_regrid2["sst"].isel(time=0).plot(ax=axs[2])

axs[0].set_title("Original")
axs[1].set_title("Regridded: xesmf")
axs[2].set_title("Regridded: regrid2")

fig.tight_layout()

Relevant log output

No response

Anything else we need to know?

Sample input files used above:

• ts_Amon_ACCESS1-0_historical_r1i1p1_185001-200512.nc
• sftlf_fx_ACCESS1-0_amip_r0i0p0.nc

Environment

xcdat 0.6.1

xr.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.10.10 | packaged by conda-forge | (main, Mar 24 2023, 20:17:34) [Clang 14.0.6 ]
python-bits: 64
OS: Darwin
OS-release: 22.6.0
machine: x86_64
processor: i386
byteorder: little
LC_ALL: None
LANG: None
LOCALE: (None, 'UTF-8')
libhdf5: 1.14.2
libnetcdf: 4.9.2

xarray: 2023.11.0
pandas: 2.1.3
numpy: 1.23.5
scipy: 1.11.4
netCDF4: 1.6.5
pydap: None
h5netcdf: None
h5py: None
Nio: None
zarr: None
cftime: 1.6.3
nc_time_axis: 1.4.1
iris: None
bottleneck: None
dask: 2023.12.0
distributed: 2023.12.0
matplotlib: 3.7.1
cartopy: 0.22.0
seaborn: 0.12.2
numbagg: None
fsspec: 2023.12.0
cupy: None
pint: None
sparse: 0.14.0
flox: None
numpy_groupies: None
setuptools: 67.7.2
pip: 23.1.2
conda: None
pytest: None
mypy: None
IPython: 8.18.1
sphinx: None

[lee1043]

Hi @jasonb5, is there any chance this could be fixed in #533?

[jasonb5]

@lee1043 This will be resolved in another PR after #533, it needs to be fixed for all the regridding tools.

[lee1043]

Close this issue as it was resolved by #533.

Result of the minimal example code above after updating the xcdat to the latest main branch.

[tomvothecoder]

Close this issue as it was resolved by #533.

Result of the minimal example code above after updating the xcdat to the latest main branch.

Thank you for confirming!