Remap OISST to tx2_3

%matplotlib inline

import xarray as xr
import xesmf
import numpy as np
import datetime
import matplotlib.pyplot as plt

today = datetime.date.today().strftime("%y%m%d")
print(today)

260306

fname = '../mesh/tx2_3v3_grid.nc'
ds_out = xr.open_dataset(fname).rename({'tlon': 'lon','tlat': 'lat', 'qlon': 'lon_b',
                                        'qlat': 'lat_b', 'nx' : 'xh', 'ny' : 'yh',
                                        'depth' : 'z_l', 'tmask':'mask'})
ds_out

<xarray.Dataset> Size: 42MB
Dimensions:  (yh: 480, xh: 540, nxp: 541, nyp: 481)
Dimensions without coordinates: yh, xh, nxp, nyp
Data variables: (12/20)
    lon      (yh, xh) float64 2MB ...
    lat      (yh, xh) float64 2MB ...
    ulon     (yh, nxp) float64 2MB ...
    ulat     (yh, nxp) float64 2MB ...
    vlon     (nyp, xh) float64 2MB ...
    vlat     (nyp, xh) float64 2MB ...
    ...       ...
    tarea    (yh, xh) float64 2MB ...
    mask     (yh, xh) float64 2MB ...
    angle    (yh, xh) float64 2MB ...
    z_l      (yh, xh) float64 2MB ...
    ar       (yh, xh) float64 2MB ...
    egs      (yh, xh) float64 2MB ...
Attributes:
    Description:  CESM MOM6 2/3 degree grid
    Author:       Frank, Fred, Gustavo (gmarques@ucar.edu)
    Created:      2026-03-05T14:49:28.971877
    type:         Glogal 2/3 degree grid file

xarray.Dataset

• Dimensions:
  • yh: 480
  • xh: 540
  • nxp: 541
  • nyp: 481
• Data variables: (20)
  • lon
    (yh, xh)
    float64
    ...

    long_name :

    array of t-grid longitudes

    units :

    degrees_east

    [259200 values with dtype=float64]

  • lat
    (yh, xh)
    float64
    ...

    long_name :

    array of t-grid latitudes

    units :

    degrees_north

    [259200 values with dtype=float64]

  • ulon
    (yh, nxp)
    float64
    ...

    long_name :

    array of u-grid longitudes

    units :

    degrees_east

    [259680 values with dtype=float64]

  • ulat
    (yh, nxp)
    float64
    ...

    long_name :

    array of u-grid latitudes

    units :

    degrees_north

    [259680 values with dtype=float64]

  • vlon
    (nyp, xh)
    float64
    ...

    long_name :

    array of v-grid longitudes

    units :

    degrees_east

    [259740 values with dtype=float64]

  • vlat
    (nyp, xh)
    float64
    ...

    long_name :

    array of v-grid latitudes

    units :

    degrees_north

    [259740 values with dtype=float64]

  • lon_b
    (nyp, nxp)
    float64
    ...

    long_name :

    array of q-grid longitudes

    units :

    degrees_east

    [260221 values with dtype=float64]

  • lat_b
    (nyp, nxp)
    float64
    ...

    long_name :

    array of q-grid latitudes

    units :

    degrees_north

    [260221 values with dtype=float64]

  • dxt
    (yh, xh)
    float64
    ...

    long_name :

    x-distance between u-points, centered at t

    units :

    meters

    [259200 values with dtype=float64]

  • dyt
    (yh, xh)
    float64
    ...

    long_name :

    y-distance between v-points, centered at t

    units :

    meters

    [259200 values with dtype=float64]

  • dxCv
    (yh, xh)
    float64
    ...

    long_name :

    x-distance between q-points, centered at v

    units :

    meters

    [259200 values with dtype=float64]

  • dyCu
    (yh, xh)
    float64
    ...

    long_name :

    y-distance between q-points, centered at u

    units :

    meters

    [259200 values with dtype=float64]

  • dxCu
    (yh, xh)
    float64
    ...

    long_name :

    x-distance between t-points, centered at u

    units :

    meters

    [259200 values with dtype=float64]

  • dyCv
    (yh, xh)
    float64
    ...

    long_name :

    y-distance between t-points, centered at v

    units :

    meters

    [259200 values with dtype=float64]

  • tarea
    (yh, xh)
    float64
    ...

    long_name :

    area of t-cells

    units :

    meters^2

    [259200 values with dtype=float64]

  • mask
    (yh, xh)
    float64
    ...

    long_name :

    ocean fraction at t-cell centers

    units :

    none

    [259200 values with dtype=float64]

  • angle
    (yh, xh)
    float64
    ...

    long_name :

    angle grid makes with latitude line

    units :

    degrees

    [259200 values with dtype=float64]

  • z_l
    (yh, xh)
    float64
    ...

    long_name :

    depth at h points

    units :

    meters

    [259200 values with dtype=float64]

  • ar
    (yh, xh)
    float64
    ...

    long_name :

    grid aspect ratio (dyt/dxt)

    units :

    none

    [259200 values with dtype=float64]

  • egs
    (yh, xh)
    float64
    ...

    long_name :

    grid effective grid spacing

    units :

    degrees

    [259200 values with dtype=float64]

• Attributes: (4)

  Description :

  CESM MOM6 2/3 degree grid

  Author :

  Frank, Fred, Gustavo (gmarques@ucar.edu)

  Created :

  2026-03-05T14:49:28.971877

  type :

  Glogal 2/3 degree grid file

infile = '/glade/campaign/cgd/oce/datasets/obs/OI.SST.v2/monthly/sst.mean.????.??.nc'
ds_in = xr.open_mfdataset(infile)
ds_in

<xarray.Dataset> Size: 2GB
Dimensions:  (time: 444, lat: 720, lon: 1440)
Coordinates:
  * time     (time) datetime64[ns] 4kB 1982-01-16 ... 2018-12-16
  * lat      (lat) float32 3kB -89.88 -89.62 -89.38 -89.12 ... 89.38 89.62 89.88
  * lon      (lon) float32 6kB 0.125 0.375 0.625 0.875 ... 359.4 359.6 359.9
Data variables:
    sst      (time, lat, lon) float32 2GB dask.array<chunksize=(1, 720, 1440), meta=np.ndarray>
Attributes:
    Conventions:    CF-1.5
    title:          NOAA High-resolution Blended Analysis: Daily Values using...
    institution:    NOAA/NCDC
    source:         NOAA/NCDC  ftp://eclipse.ncdc.noaa.gov/pub/OI-daily-v2/
    comment:        Reynolds, et al., 2007: Daily High-Resolution-Blended Ana...
    history:        Mon Dec 23 22:15:20 2019: ncra -F -d time,1,31 sst.day.me...
    references:     https://www.esrl.noaa.gov/psd/data/gridded/data.noaa.oiss...
    dataset_title:  NOAA Daily Optimum Interpolation Sea Surface Temperature
    NCO:            netCDF Operators version 4.8.1 (Homepage = http://nco.sf....

xarray.Dataset

• Dimensions:
  • time: 444
  • lat: 720
  • lon: 1440
• Coordinates: (3)
  • time
    (time)
    datetime64[ns]
    1982-01-16 ... 2018-12-16

    long_name :

    Time

    delta_t :

    0000-00-01 00:00:00

    avg_period :

    0000-00-01 00:00:00

    axis :

    T

    actual_range :

    [66474. 66838.]

    cell_methods :

    time: mean

    array(['1982-01-16T00:00:00.000000000', '1982-02-14T12:00:00.000000000',
           '1982-03-16T00:00:00.000000000', ..., '2018-10-16T00:00:00.000000000',
           '2018-11-15T12:00:00.000000000', '2018-12-16T00:00:00.000000000'],
          shape=(444,), dtype='datetime64[ns]')

  • lat
    (lat)
    float32
    -89.88 -89.62 ... 89.62 89.88

    long_name :

    Latitude

    standard_name :

    latitude

    units :

    degrees_north

    actual_range :

    [-89.875 89.875]

    axis :

    Y

    array([-89.875, -89.625, -89.375, ...,  89.375,  89.625,  89.875],
          shape=(720,), dtype=float32)

  • lon
    (lon)
    float32
    0.125 0.375 0.625 ... 359.6 359.9

    long_name :

    Longitude

    standard_name :

    longitude

    units :

    degrees_east

    actual_range :

    [1.25000e-01 3.59875e+02]

    axis :

    X

    array([1.25000e-01, 3.75000e-01, 6.25000e-01, ..., 3.59375e+02, 3.59625e+02,
           3.59875e+02], shape=(1440,), dtype=float32)

• Data variables: (1)
  • sst
    (time, lat, lon)
    float32
    dask.array<chunksize=(1, 720, 1440), meta=np.ndarray>

    long_name :

    Daily Sea Surface Temperature

    units :

    degC

    valid_range :

    [-3. 45.]

    precision :

    2.0

    dataset :

    NOAA High-resolution Blended Analysis

    var_desc :

    Sea Surface Temperature

    level_desc :

    Surface

    statistic :

    Mean

    parent_stat :

    Individual Observations

    actual_range :

    [-1.8 34.48]

    cell_methods :

    time: mean

    Array Chunk Bytes 1.71 GiB 3.96 MiB Shape (444, 720, 1440) (1, 720, 1440) Dask graph 444 chunks in 889 graph layers Data type float32 numpy.ndarray

• Attributes: (9)

  Conventions :

  CF-1.5

  title :

  NOAA High-resolution Blended Analysis: Daily Values using AVHRR only

  institution :

  NOAA/NCDC

  source :

  NOAA/NCDC ftp://eclipse.ncdc.noaa.gov/pub/OI-daily-v2/

  comment :

  Reynolds, et al., 2007: Daily High-Resolution-Blended Analyses for Sea Surface Temperature. J. Climate, 20, 5473-5496. Climatology is based on 1971-2000 OI.v2 SST, Satellite data: Navy NOAA17 NOAA18 AVHRR, Ice data: NCEP ice.

  history :

  Mon Dec 23 22:15:20 2019: ncra -F -d time,1,31 sst.day.mean.1982.nc -o monthly/sst.mean.1982.01.nc Thu Aug 24 13:35:47 2017: ncatted -O -a References,global,d,, sst.day.mean.1982.v2.nc Version 1.0

  references :

  https://www.esrl.noaa.gov/psd/data/gridded/data.noaa.oisst.v2.highres.html

  dataset_title :

  NOAA Daily Optimum Interpolation Sea Surface Temperature

  NCO :

  netCDF Operators version 4.8.1 (Homepage = http://nco.sf.net, Code = http://github.com/nco/nco)

def regrid_tracer(fld, ds_in, ds_out, method='bilinear'):

    regrid = xesmf.Regridder(
        ds_in,
        ds_out,
        method=method,
        periodic=True,
    )
    fld_out = regrid(ds_in[fld])#[0,0,:]#.where(ds_in.lat>ds_out.geolat.min())
    return fld_out

Potential temperature

sst = regrid_tracer('sst', ds_in, ds_out).rename('sst')
sst

<xarray.DataArray 'sst' (time: 444, yh: 480, xh: 540)> Size: 460MB
dask.array<astype, shape=(444, 480, 540), dtype=float32, chunksize=(1, 480, 540), chunktype=numpy.ndarray>
Coordinates:
  * time     (time) datetime64[ns] 4kB 1982-01-16 ... 2018-12-16
Dimensions without coordinates: yh, xh
Attributes:
    regrid_method:  bilinear

xarray.DataArray

'sst'

• time: 444
• yh: 480
• xh: 540

• dask.array<chunksize=(1, 480, 540), meta=np.ndarray>

  Array Chunk Bytes 439.01 MiB 0.99 MiB Shape (444, 480, 540) (1, 480, 540) Dask graph 444 chunks in 895 graph layers Data type float32 numpy.ndarray

• Coordinates: (1)
  • time
    (time)
    datetime64[ns]
    1982-01-16 ... 2018-12-16

    array(['1982-01-16T00:00:00.000000000', '1982-02-14T12:00:00.000000000',
           '1982-03-16T00:00:00.000000000', ..., '2018-10-16T00:00:00.000000000',
           '2018-11-15T12:00:00.000000000', '2018-12-16T00:00:00.000000000'],
          shape=(444,), dtype='datetime64[ns]')

• Attributes: (1)

  regrid_method :

  bilinear

# add coordinates
sst = sst.assign_coords(geolat=ds_out.lat, geolon=ds_out.lon)
sst

<xarray.DataArray 'sst' (time: 444, yh: 480, xh: 540)> Size: 460MB
dask.array<astype, shape=(444, 480, 540), dtype=float32, chunksize=(1, 480, 540), chunktype=numpy.ndarray>
Coordinates:
  * time     (time) datetime64[ns] 4kB 1982-01-16 ... 2018-12-16
    geolat   (yh, xh) float64 2MB -81.56 -81.56 -81.56 ... 50.27 50.11 49.99
    geolon   (yh, xh) float64 2MB -286.7 -286.0 -285.3 ... 72.97 72.98 73.0
Dimensions without coordinates: yh, xh
Attributes:
    regrid_method:  bilinear

xarray.DataArray

'sst'

• time: 444
• yh: 480
• xh: 540

• dask.array<chunksize=(1, 480, 540), meta=np.ndarray>

  Array Chunk Bytes 439.01 MiB 0.99 MiB Shape (444, 480, 540) (1, 480, 540) Dask graph 444 chunks in 895 graph layers Data type float32 numpy.ndarray

• Coordinates: (3)
  • time
    (time)
    datetime64[ns]
    1982-01-16 ... 2018-12-16

    array(['1982-01-16T00:00:00.000000000', '1982-02-14T12:00:00.000000000',
           '1982-03-16T00:00:00.000000000', ..., '2018-10-16T00:00:00.000000000',
           '2018-11-15T12:00:00.000000000', '2018-12-16T00:00:00.000000000'],
          shape=(444,), dtype='datetime64[ns]')

  • geolat
    (yh, xh)
    float64
    -81.56 -81.56 ... 50.11 49.99

    long_name :

    array of t-grid latitudes

    units :

    degrees_north

    array([[-81.558125, -81.558125, -81.558125, ..., -81.558125, -81.558125,
            -81.558125],
           [-81.459689, -81.459689, -81.459689, ..., -81.459689, -81.459689,
            -81.459689],
           [-81.360114, -81.360114, -81.360114, ..., -81.360114, -81.360114,
            -81.360114],
           ...,
           [ 49.981551,  50.10076 ,  50.266038, ...,  50.266038,  50.10076 ,
             49.981551],
           [ 49.988104,  50.105166,  50.270311, ...,  50.270311,  50.105166,
             49.988104],
           [ 49.993627,  50.107298,  50.272397, ...,  50.272397,  50.107298,
             49.993627]], shape=(480, 540))

  • geolon
    (yh, xh)
    float64
    -286.7 -286.0 -285.3 ... 72.98 73.0

    long_name :

    array of t-grid longitudes

    units :

    degrees_east

    array([[-286.666667, -286.      , -285.333333, ...,   71.333333,   72.      ,
              72.666667],
           [-286.666667, -286.      , -285.333333, ...,   71.333333,   72.      ,
              72.666667],
           [-286.666667, -286.      , -285.333333, ...,   71.333333,   72.      ,
              72.666667],
           ...,
           [-286.970638, -286.909025, -286.86341 , ...,   72.86341 ,   72.909025,
              72.970638],
           [-286.987224, -286.947478, -286.919655, ...,   72.919655,   72.947478,
              72.987224],
           [-286.999825, -286.983349, -286.973852, ...,   72.973852,   72.983349,
              72.999825]], shape=(480, 540))

• Attributes: (1)

  regrid_method :

  bilinear

Visual inspection

Make sure original and remapped plots look similar.

# visual inspection. Make sure original and remapped plots look similar
for t in range(12):
  fig, axes = plt.subplots(nrows=1, ncols=2, figsize=(18,4))
  sst[t,:,:].plot.pcolormesh(x='geolon',y='geolat',ax=axes[0], vmin=-2,vmax=30)
  ds_in['sst'][t,:,:].plot.pcolormesh(ax=axes[1], vmin=-2,vmax=30)
  axes[0].set_title('Remapped')
  axes[1].set_title('Original grid')
  plt.suptitle('Month ='+ str(t+1))

/glade/u/apps/opt/miniforge/envs/npl-2026a/lib/python3.13/site-packages/dask/config.py:786: FutureWarning: Dask configuration key 'allowed-failures' has been deprecated; please use 'distributed.scheduler.allowed-failures' instead
  warnings.warn(

# Global attrs
sst.attrs['title'] = 'Monthly OISSTv2 remapped to tx2_3v3'
sst.attrs['references'] = 'https://www.esrl.noaa.gov/psd/data/gridded/data.noaa.oisst.v2.highres.html'
sst.attrs['author'] = 'Gustavo Marques (gmarques@ucar.edu)'
sst.attrs['date'] = today
sst.attrs['infile'] = infile
sst.attrs['url'] = 'https://github.com/NCAR/tx2_3/oisst'
# save
fname = 'oisstv2_to_tx2_3v3_{}.nc'.format(today)
sst.to_netcdf(fname)