Hydrology

In the following sections, we use Python to demonstrate how to gather data for a specific watershed in the HYDAT database

Environment setup

import xarray as xr
import intake
import hvplot.xarray
import dask

Accessing the data

We are now ready to access our catalog which uses Intake to organize all our datasets.

catalog_url = 'https://raw.githubusercontent.com/hydrocloudservices/catalogs/main/catalogs/main.yaml'
cat=intake.open_catalog(catalog_url)
cat

main:
  args:
    path: https://raw.githubusercontent.com/hydrocloudservices/catalogs/main/catalogs/main.yaml
  description: Master Data Catalog
  driver: intake.catalog.local.YAMLFileCatalog
  metadata: {}

Let’s open ERA5-Land reanalysis (reference dataset) from which we will gather data at our watershed of interest

ds_era5l = cat.atmosphere.era5_land_reanalysis.to_dask()
ds_era5l

<xarray.Dataset>
Dimensions:    (latitude: 701, longitude: 1171, time: 636240)
Coordinates:
  * latitude   (latitude) float64 85.0 84.9 84.8 84.7 ... 15.3 15.2 15.1 15.0
  * longitude  (longitude) float64 -167.0 -166.9 -166.8 ... -50.2 -50.1 -50.0
  * time       (time) datetime64[ns] 1950-01-01 ... 2022-07-31T23:00:00
Data variables:
    sd         (time, latitude, longitude) float32 dask.array<chunksize=(8760, 7, 7), meta=np.ndarray>
    t2m        (time, latitude, longitude) float32 dask.array<chunksize=(8760, 7, 7), meta=np.ndarray>
    tp         (time, latitude, longitude) float32 dask.array<chunksize=(8760, 7, 7), meta=np.ndarray>

xarray.Dataset

• Dimensions:
  • latitude: 701
  • longitude: 1171
  • time: 636240
• Coordinates: (3)
  • latitude
    (latitude)
    float64
    85.0 84.9 84.8 ... 15.2 15.1 15.0

    long_name :

    latitude

    standard_name :

    latitude

    stored_direction :

    decreasing

    units :

    degrees_north

    array([85. , 84.9, 84.8, ..., 15.2, 15.1, 15. ])

  • longitude
    (longitude)
    float64
    -167.0 -166.9 ... -50.1 -50.0

    long_name :

    longitude

    standard_name :

    longitude

    units :

    degrees_east

    array([-167. , -166.9, -166.8, ...,  -50.2,  -50.1,  -50. ])

  • time
    (time)
    datetime64[ns]
    1950-01-01 ... 2022-07-31T23:00:00

    array(['1950-01-01T00:00:00.000000000', '1950-01-01T01:00:00.000000000',
           '1950-01-01T02:00:00.000000000', ..., '2022-07-31T21:00:00.000000000',
           '2022-07-31T22:00:00.000000000', '2022-07-31T23:00:00.000000000'],
          dtype='datetime64[ns]')

• Data variables: (3)
  • sd
    (time, latitude, longitude)
    float32
    dask.array<chunksize=(8760, 7, 7), meta=np.ndarray>

    GRIB_NV :

    0

    GRIB_Nx :

    1171

    GRIB_Ny :

    701

    GRIB_cfName :

    lwe_thickness_of_surface_snow_amount

    GRIB_cfVarName :

    sd

    GRIB_dataType :

    an

    GRIB_gridDefinitionDescription :

    Latitude/Longitude Grid

    GRIB_gridType :

    regular_ll

    GRIB_iDirectionIncrementInDegrees :

    0.1

    GRIB_iScansNegatively :

    0

    GRIB_jDirectionIncrementInDegrees :

    0.1

    GRIB_jPointsAreConsecutive :

    0

    GRIB_jScansPositively :

    0

    GRIB_latitudeOfFirstGridPointInDegrees :

    85.0

    GRIB_latitudeOfLastGridPointInDegrees :

    15.0

    GRIB_longitudeOfFirstGridPointInDegrees :

    -167.0

    GRIB_longitudeOfLastGridPointInDegrees :

    -50.0

    GRIB_missingValue :

    9999

    GRIB_name :

    Snow depth

    GRIB_numberOfPoints :

    820871

    GRIB_paramId :

    141

    GRIB_shortName :

    sd

    GRIB_stepType :

    instant

    GRIB_stepUnits :

    1

    GRIB_totalNumber :

    0

    GRIB_typeOfLevel :

    surface

    GRIB_units :

    m of water equivalent

    coordinates :

    number time step surface latitude longitude valid_time

    long_name :

    Snow depth

    standard_name :

    lwe_thickness_of_surface_snow_amount

    units :

    m of water equivalent

    Array Chunk Bytes 1.90 TiB 1.64 MiB Shape (636240, 701, 1171) (8760, 7, 7) Dask graph 1238664 chunks in 2 graph layers Data type float32 numpy.ndarray

  • t2m
    (time, latitude, longitude)
    float32
    dask.array<chunksize=(8760, 7, 7), meta=np.ndarray>

    GRIB_NV :

    0

    GRIB_Nx :

    1171

    GRIB_Ny :

    701

    GRIB_cfName :

    unknown

    GRIB_cfVarName :

    t2m

    GRIB_dataType :

    fc

    GRIB_gridDefinitionDescription :

    Latitude/Longitude Grid

    GRIB_gridType :

    regular_ll

    GRIB_iDirectionIncrementInDegrees :

    0.1

    GRIB_iScansNegatively :

    0

    GRIB_jDirectionIncrementInDegrees :

    0.1

    GRIB_jPointsAreConsecutive :

    0

    GRIB_jScansPositively :

    0

    GRIB_latitudeOfFirstGridPointInDegrees :

    85.0

    GRIB_latitudeOfLastGridPointInDegrees :

    15.0

    GRIB_longitudeOfFirstGridPointInDegrees :

    -167.0

    GRIB_longitudeOfLastGridPointInDegrees :

    -50.0

    GRIB_missingValue :

    9999

    GRIB_name :

    2 metre temperature

    GRIB_numberOfPoints :

    820871

    GRIB_paramId :

    167

    GRIB_shortName :

    2t

    GRIB_stepType :

    instant

    GRIB_stepUnits :

    1

    GRIB_totalNumber :

    0

    GRIB_typeOfLevel :

    surface

    GRIB_units :

    K

    coordinates :

    number time step surface latitude longitude valid_time

    long_name :

    2 metre temperature

    standard_name :

    unknown

    units :

    K

    Array Chunk Bytes 1.90 TiB 1.64 MiB Shape (636240, 701, 1171) (8760, 7, 7) Dask graph 1238664 chunks in 2 graph layers Data type float32 numpy.ndarray

  • tp
    (time, latitude, longitude)
    float32
    dask.array<chunksize=(8760, 7, 7), meta=np.ndarray>

    GRIB_NV :

    0

    GRIB_Nx :

    1171

    GRIB_Ny :

    701

    GRIB_cfName :

    unknown

    GRIB_cfVarName :

    tp

    GRIB_dataType :

    fc

    GRIB_gridDefinitionDescription :

    Latitude/Longitude Grid

    GRIB_gridType :

    regular_ll

    GRIB_iDirectionIncrementInDegrees :

    0.1

    GRIB_iScansNegatively :

    0

    GRIB_jDirectionIncrementInDegrees :

    0.1

    GRIB_jPointsAreConsecutive :

    0

    GRIB_jScansPositively :

    0

    GRIB_latitudeOfFirstGridPointInDegrees :

    85.0

    GRIB_latitudeOfLastGridPointInDegrees :

    15.0

    GRIB_longitudeOfFirstGridPointInDegrees :

    -167.0

    GRIB_longitudeOfLastGridPointInDegrees :

    -50.0

    GRIB_missingValue :

    9999

    GRIB_name :

    Total precipitation

    GRIB_numberOfPoints :

    820871

    GRIB_paramId :

    228

    GRIB_shortName :

    tp

    GRIB_stepType :

    accum

    GRIB_stepUnits :

    1

    GRIB_totalNumber :

    0

    GRIB_typeOfLevel :

    surface

    GRIB_units :

    m

    coordinates :

    number time step surface latitude longitude valid_time

    long_name :

    Total precipitation

    standard_name :

    unknown

    units :

    m

    Array Chunk Bytes 1.90 TiB 1.64 MiB Shape (636240, 701, 1171) (8760, 7, 7) Dask graph 1238664 chunks in 2 graph layers Data type float32 numpy.ndarray

• Indexes: (3)
  • latitude
    PandasIndex

    PandasIndex(Float64Index([              85.0,               84.9,  84.80000000000001,
                   84.70000000000002,  84.60000000000002,  84.50000000000003,
                   84.40000000000003,  84.30000000000004,  84.20000000000005,
                   84.10000000000005,
                  ...
                  15.900000000000512, 15.800000000000512, 15.700000000000513,
                  15.600000000000513, 15.500000000000513, 15.400000000000514,
                  15.300000000000514, 15.200000000000514, 15.100000000000515,
                  15.000000000000515],
                 dtype='float64', name='latitude', length=701))

  • longitude
    PandasIndex

    PandasIndex(Float64Index([             -167.0,              -166.9,              -166.8,
                  -166.70000000000002, -166.60000000000002, -166.50000000000003,
                  -166.40000000000003, -166.30000000000004, -166.20000000000005,
                  -166.10000000000005,
                  ...
                   -50.90000000000567,  -50.80000000000567, -50.700000000005666,
                  -50.600000000005664,  -50.50000000000566,  -50.40000000000566,
                   -50.30000000000566,  -50.20000000000566,  -50.10000000000566,
                  -50.000000000005656],
                 dtype='float64', name='longitude', length=1171))

  • time
    PandasIndex

    PandasIndex(DatetimeIndex(['1950-01-01 00:00:00', '1950-01-01 01:00:00',
                   '1950-01-01 02:00:00', '1950-01-01 03:00:00',
                   '1950-01-01 04:00:00', '1950-01-01 05:00:00',
                   '1950-01-01 06:00:00', '1950-01-01 07:00:00',
                   '1950-01-01 08:00:00', '1950-01-01 09:00:00',
                   ...
                   '2022-07-31 14:00:00', '2022-07-31 15:00:00',
                   '2022-07-31 16:00:00', '2022-07-31 17:00:00',
                   '2022-07-31 18:00:00', '2022-07-31 19:00:00',
                   '2022-07-31 20:00:00', '2022-07-31 21:00:00',
                   '2022-07-31 22:00:00', '2022-07-31 23:00:00'],
                  dtype='datetime64[ns]', name='time', length=636240, freq=None))

• Attributes: (0)

Here, we open HYDAT dataset as a xarray Dataset :

ds_hydat = cat.hydrology.hydat.to_dask()
ds_hydat

<xarray.Dataset>
Dimensions:        (data_type: 2, id: 7881, spatial_agg: 2, timestep: 1,
                    time_agg: 1, latitude: 2800, longitude: 4680, time: 59413)
Coordinates: (12/15)
  * data_type      (data_type) <U5 'flow' 'level'
    drainage_area  (id) float64 dask.array<chunksize=(10,), meta=np.ndarray>
    end_date       (id, data_type, spatial_agg, timestep, time_agg) object dask.array<chunksize=(7881, 2, 2, 1, 1), meta=np.ndarray>
  * id             (id) <U7 '01AA002' '01AD001' ... '11AF004' '11AF005'
  * latitude       (latitude) float64 85.0 84.97 84.95 ... 15.07 15.05 15.02
  * longitude      (longitude) float64 -167.0 -167.0 -166.9 ... -50.05 -50.02
    ...             ...
    source         (id) object dask.array<chunksize=(7881,), meta=np.ndarray>
  * spatial_agg    (spatial_agg) object 'point' 'watershed'
    start_date     (id, data_type, spatial_agg, timestep, time_agg) object dask.array<chunksize=(7881, 2, 2, 1, 1), meta=np.ndarray>
  * time           (time) datetime64[ns] 1860-01-01 1860-01-02 ... 2022-08-31
  * time_agg       (time_agg) <U4 'mean'
  * timestep       (timestep) <U3 'day'
Data variables:
    mask           (id, latitude, longitude) float64 dask.array<chunksize=(1, 500, 500), meta=np.ndarray>
    value          (id, time, data_type, spatial_agg, timestep, time_agg) float64 dask.array<chunksize=(10, 59413, 1, 1, 1, 1), meta=np.ndarray>

xarray.Dataset

• Dimensions:
  • data_type: 2
  • id: 7881
  • spatial_agg: 2
  • timestep: 1
  • time_agg: 1
  • latitude: 2800
  • longitude: 4680
  • time: 59413
• Coordinates: (15)
  • data_type
    (data_type)
    <U5
    'flow' 'level'

    array(['flow', 'level'], dtype='<U5')

  • drainage_area
    (id)
    float64
    dask.array<chunksize=(10,), meta=np.ndarray>

    Array Chunk Bytes 61.57 kiB 80 B Shape (7881,) (10,) Dask graph 789 chunks in 2 graph layers Data type float64 numpy.ndarray

  • end_date
    (id, data_type, spatial_agg, timestep, time_agg)
    object
    dask.array<chunksize=(7881, 2, 2, 1, 1), meta=np.ndarray>

    Array Chunk Bytes 246.28 kiB 246.28 kiB Shape (7881, 2, 2, 1, 1) (7881, 2, 2, 1, 1) Dask graph 1 chunks in 2 graph layers Data type object numpy.ndarray

  • id
    (id)
    <U7
    '01AA002' '01AD001' ... '11AF005'

    array(['01AA002', '01AD001', '01AD002', ..., '11AF002', '11AF004', '11AF005'],
          dtype='<U7')

  • latitude
    (latitude)
    float64
    85.0 84.97 84.95 ... 15.05 15.02

    axis :

    Y

    long_name :

    latitude

    standard_name :

    latitude

    units :

    degrees_north

    array([85.   , 84.975, 84.95 , ..., 15.075, 15.05 , 15.025])

  • longitude
    (longitude)
    float64
    -167.0 -167.0 ... -50.05 -50.02

    axis :

    X

    long_name :

    longitude

    standard_name :

    longitude

    units :

    degrees_east

    array([-167.   , -166.975, -166.95 , ...,  -50.075,  -50.05 ,  -50.025])

  • name
    (id)
    object
    dask.array<chunksize=(7881,), meta=np.ndarray>

    Array Chunk Bytes 61.57 kiB 61.57 kiB Shape (7881,) (7881,) Dask graph 1 chunks in 2 graph layers Data type object numpy.ndarray

  • province
    (id)
    object
    dask.array<chunksize=(7881,), meta=np.ndarray>

    Array Chunk Bytes 61.57 kiB 61.57 kiB Shape (7881,) (7881,) Dask graph 1 chunks in 2 graph layers Data type object numpy.ndarray

  • regulated
    (id)
    float64
    dask.array<chunksize=(10,), meta=np.ndarray>

    Array Chunk Bytes 61.57 kiB 80 B Shape (7881,) (10,) Dask graph 789 chunks in 2 graph layers Data type float64 numpy.ndarray

  • source
    (id)
    object
    dask.array<chunksize=(7881,), meta=np.ndarray>

    Array Chunk Bytes 61.57 kiB 61.57 kiB Shape (7881,) (7881,) Dask graph 1 chunks in 2 graph layers Data type object numpy.ndarray

  • spatial_agg
    (spatial_agg)
    object
    'point' 'watershed'

    array(['point', 'watershed'], dtype=object)

  • start_date
    (id, data_type, spatial_agg, timestep, time_agg)
    object
    dask.array<chunksize=(7881, 2, 2, 1, 1), meta=np.ndarray>

    Array Chunk Bytes 246.28 kiB 246.28 kiB Shape (7881, 2, 2, 1, 1) (7881, 2, 2, 1, 1) Dask graph 1 chunks in 2 graph layers Data type object numpy.ndarray

  • time
    (time)
    datetime64[ns]
    1860-01-01 ... 2022-08-31

    array(['1860-01-01T00:00:00.000000000', '1860-01-02T00:00:00.000000000',
           '1860-01-03T00:00:00.000000000', ..., '2022-08-29T00:00:00.000000000',
           '2022-08-30T00:00:00.000000000', '2022-08-31T00:00:00.000000000'],
          dtype='datetime64[ns]')

  • time_agg
    (time_agg)
    <U4
    'mean'

    array(['mean'], dtype='<U4')

  • timestep
    (timestep)
    <U3
    'day'

    array(['day'], dtype='<U3')

• Data variables: (2)
  • mask
    (id, latitude, longitude)
    float64
    dask.array<chunksize=(1, 500, 500), meta=np.ndarray>

    grid_mapping :

    spatial_ref

    Array Chunk Bytes 769.44 GiB 1.91 MiB Shape (7881, 2800, 4680) (1, 500, 500) Dask graph 472860 chunks in 2 graph layers Data type float64 numpy.ndarray

  • value
    (id, time, data_type, spatial_agg, timestep, time_agg)
    float64
    dask.array<chunksize=(10, 59413, 1, 1, 1, 1), meta=np.ndarray>

    Array Chunk Bytes 13.95 GiB 4.53 MiB Shape (7881, 59413, 2, 2, 1, 1) (10, 59413, 1, 1, 1, 1) Dask graph 3156 chunks in 2 graph layers Data type float64 numpy.ndarray

• Indexes: (8)
  • data_type
    PandasIndex

    PandasIndex(Index(['flow', 'level'], dtype='object', name='data_type'))

  • id
    PandasIndex

    PandasIndex(Index(['01AA002', '01AD001', '01AD002', '01AD003', '01AD004', '01AD005',
           '01AD008', '01AD009', '01AD012', '01AD013',
           ...
           '11AE012', '11AE013', '11AE014', '11AE015', '11AE016', '11AE018',
           '11AF001', '11AF002', '11AF004', '11AF005'],
          dtype='object', name='id', length=7881))

  • latitude
    PandasIndex

    PandasIndex(Float64Index([              85.0,             84.975,  84.94999999999999,
                   84.92499999999998,  84.89999999999998,  84.87499999999997,
                   84.84999999999997,  84.82499999999996,  84.79999999999995,
                   84.77499999999995,
                  ...
                   15.24999999998414, 15.224999999984135,  15.19999999998413,
                  15.174999999984124, 15.149999999984118, 15.124999999984112,
                  15.099999999984107, 15.074999999984101, 15.049999999984095,
                   15.02499999998409],
                 dtype='float64', name='latitude', length=2800))

  • longitude
    PandasIndex

    PandasIndex(Float64Index([             -167.0,            -166.975,             -166.95,
                  -166.92499999999998, -166.89999999999998, -166.87499999999997,
                  -166.84999999999997, -166.82499999999996, -166.79999999999995,
                  -166.77499999999995,
                  ...
                  -50.249999999973454,  -50.22499999997345,  -50.19999999997344,
                   -50.17499999997344,  -50.14999999997343, -50.124999999973426,
                   -50.09999999997342, -50.074999999973414,  -50.04999999997341,
                    -50.0249999999734],
                 dtype='float64', name='longitude', length=4680))

  • spatial_agg
    PandasIndex

    PandasIndex(Index(['point', 'watershed'], dtype='object', name='spatial_agg'))

  • time
    PandasIndex

    PandasIndex(DatetimeIndex(['1860-01-01', '1860-01-02', '1860-01-03', '1860-01-04',
                   '1860-01-05', '1860-01-06', '1860-01-07', '1860-01-08',
                   '1860-01-09', '1860-01-10',
                   ...
                   '2022-08-22', '2022-08-23', '2022-08-24', '2022-08-25',
                   '2022-08-26', '2022-08-27', '2022-08-28', '2022-08-29',
                   '2022-08-30', '2022-08-31'],
                  dtype='datetime64[ns]', name='time', length=59413, freq=None))

  • time_agg
    PandasIndex

    PandasIndex(Index(['mean'], dtype='object', name='time_agg'))

  • timestep
    PandasIndex

    PandasIndex(Index(['day'], dtype='object', name='timestep'))

• Attributes: (0)

In the HYDAT dataset, we provide the mask variable which represents rasterized watershed deleniation at a 0.025 deg spatial resolution for most ids (stations) where streaflow is recorded. For instance, we can vizualize the mask for the 01AA002 station :

%%time

station_id = '02PL005'
da = ds_hydat.sel(id=station_id)
mask = da.mask

mask \
.where(mask==1, drop=True) \
.hvplot(geo=True, tiles='ESRI', width=750, height=500, colorbar=False, alpha=0.8)

CPU times: user 11.6 s, sys: 526 ms, total: 12.2 s
Wall time: 29.3 s

We then interpolate our mask to fit the reference dataset’s spatial resolution :

%%time
mask = mask.interp_like(ds_era5l)
mask = mask.where(mask==1, drop=True)

mask \
.hvplot(geo=True, tiles='ESRI', width=750, height=500, colorbar=False, alpha=0.8)

CPU times: user 9.69 s, sys: 307 ms, total: 9.99 s
Wall time: 23.8 s

We can now apply the mask to the reference dataset and average the data on the mask. Combining this data with hydrometric data such as streamflows provides inputs for interesting analysis such as hydrological studies.

%%time
with dask.config.set(**{'array.slicing.split_large_chunks': False}):
    ds = ds_era5l.where(mask==1, drop=True)
ds

CPU times: user 13.1 s, sys: 264 ms, total: 13.4 s
Wall time: 31.2 s

<xarray.Dataset>
Dimensions:        (time: 636240, latitude: 3, longitude: 4)
Coordinates:
  * latitude       (latitude) float64 46.2 46.1 46.0
  * longitude      (longitude) float64 -71.6 -71.5 -71.4 -71.3
  * time           (time) datetime64[ns] 1950-01-01 ... 2022-07-31T23:00:00
    drainage_area  float64 dask.array<chunksize=(), meta=np.ndarray>
    id             <U7 '02PL005'
    name           object dask.array<chunksize=(), meta=np.ndarray>
    province       object dask.array<chunksize=(), meta=np.ndarray>
    regulated      float64 dask.array<chunksize=(), meta=np.ndarray>
    source         object dask.array<chunksize=(), meta=np.ndarray>
Data variables:
    sd             (time, latitude, longitude) float32 dask.array<chunksize=(8760, 3, 4), meta=np.ndarray>
    t2m            (time, latitude, longitude) float32 dask.array<chunksize=(8760, 3, 4), meta=np.ndarray>
    tp             (time, latitude, longitude) float32 dask.array<chunksize=(8760, 3, 4), meta=np.ndarray>

xarray.Dataset

• Dimensions:
  • time: 636240
  • latitude: 3
  • longitude: 4
• Coordinates: (9)
  • latitude
    (latitude)
    float64
    46.2 46.1 46.0

    long_name :

    latitude

    standard_name :

    latitude

    stored_direction :

    decreasing

    units :

    degrees_north

    array([46.2, 46.1, 46. ])

  • longitude
    (longitude)
    float64
    -71.6 -71.5 -71.4 -71.3

    long_name :

    longitude

    standard_name :

    longitude

    units :

    degrees_east

    array([-71.6, -71.5, -71.4, -71.3])

  • time
    (time)
    datetime64[ns]
    1950-01-01 ... 2022-07-31T23:00:00

    array(['1950-01-01T00:00:00.000000000', '1950-01-01T01:00:00.000000000',
           '1950-01-01T02:00:00.000000000', ..., '2022-07-31T21:00:00.000000000',
           '2022-07-31T22:00:00.000000000', '2022-07-31T23:00:00.000000000'],
          dtype='datetime64[ns]')

  • drainage_area
    ()
    float64
    dask.array<chunksize=(), meta=np.ndarray>

    Array Chunk Bytes 8 B 8 B Shape () () Dask graph 1 chunks in 3 graph layers Data type float64 numpy.ndarray

  • id
    ()
    <U7
    '02PL005'

    array('02PL005', dtype='<U7')

  • name
    ()
    object
    dask.array<chunksize=(), meta=np.ndarray>

    Array Chunk Bytes 8 B 8 B Shape () () Dask graph 1 chunks in 3 graph layers Data type object numpy.ndarray

  • province
    ()
    object
    dask.array<chunksize=(), meta=np.ndarray>

    Array Chunk Bytes 8 B 8 B Shape () () Dask graph 1 chunks in 3 graph layers Data type object numpy.ndarray

  • regulated
    ()
    float64
    dask.array<chunksize=(), meta=np.ndarray>

    Array Chunk Bytes 8 B 8 B Shape () () Dask graph 1 chunks in 3 graph layers Data type float64 numpy.ndarray

  • source
    ()
    object
    dask.array<chunksize=(), meta=np.ndarray>

    Array Chunk Bytes 8 B 8 B Shape () () Dask graph 1 chunks in 3 graph layers Data type object numpy.ndarray

• Data variables: (3)
  • sd
    (time, latitude, longitude)
    float32
    dask.array<chunksize=(8760, 3, 4), meta=np.ndarray>

    GRIB_NV :

    0

    GRIB_Nx :

    1171

    GRIB_Ny :

    701

    GRIB_cfName :

    lwe_thickness_of_surface_snow_amount

    GRIB_cfVarName :

    sd

    GRIB_dataType :

    an

    GRIB_gridDefinitionDescription :

    Latitude/Longitude Grid

    GRIB_gridType :

    regular_ll

    GRIB_iDirectionIncrementInDegrees :

    0.1

    GRIB_iScansNegatively :

    0

    GRIB_jDirectionIncrementInDegrees :

    0.1

    GRIB_jPointsAreConsecutive :

    0

    GRIB_jScansPositively :

    0

    GRIB_latitudeOfFirstGridPointInDegrees :

    85.0

    GRIB_latitudeOfLastGridPointInDegrees :

    15.0

    GRIB_longitudeOfFirstGridPointInDegrees :

    -167.0

    GRIB_longitudeOfLastGridPointInDegrees :

    -50.0

    GRIB_missingValue :

    9999

    GRIB_name :

    Snow depth

    GRIB_numberOfPoints :

    820871

    GRIB_paramId :

    141

    GRIB_shortName :

    sd

    GRIB_stepType :

    instant

    GRIB_stepUnits :

    1

    GRIB_totalNumber :

    0

    GRIB_typeOfLevel :

    surface

    GRIB_units :

    m of water equivalent

    coordinates :

    number time step surface latitude longitude valid_time

    long_name :

    Snow depth

    standard_name :

    lwe_thickness_of_surface_snow_amount

    units :

    m of water equivalent

    Array Chunk Bytes 29.12 MiB 410.62 kiB Shape (636240, 3, 4) (8760, 3, 4) Dask graph 73 chunks in 30 graph layers Data type float32 numpy.ndarray

  • t2m
    (time, latitude, longitude)
    float32
    dask.array<chunksize=(8760, 3, 4), meta=np.ndarray>

    GRIB_NV :

    0

    GRIB_Nx :

    1171

    GRIB_Ny :

    701

    GRIB_cfName :

    unknown

    GRIB_cfVarName :

    t2m

    GRIB_dataType :

    fc

    GRIB_gridDefinitionDescription :

    Latitude/Longitude Grid

    GRIB_gridType :

    regular_ll

    GRIB_iDirectionIncrementInDegrees :

    0.1

    GRIB_iScansNegatively :

    0

    GRIB_jDirectionIncrementInDegrees :

    0.1

    GRIB_jPointsAreConsecutive :

    0

    GRIB_jScansPositively :

    0

    GRIB_latitudeOfFirstGridPointInDegrees :

    85.0

    GRIB_latitudeOfLastGridPointInDegrees :

    15.0

    GRIB_longitudeOfFirstGridPointInDegrees :

    -167.0

    GRIB_longitudeOfLastGridPointInDegrees :

    -50.0

    GRIB_missingValue :

    9999

    GRIB_name :

    2 metre temperature

    GRIB_numberOfPoints :

    820871

    GRIB_paramId :

    167

    GRIB_shortName :

    2t

    GRIB_stepType :

    instant

    GRIB_stepUnits :

    1

    GRIB_totalNumber :

    0

    GRIB_typeOfLevel :

    surface

    GRIB_units :

    K

    coordinates :

    number time step surface latitude longitude valid_time

    long_name :

    2 metre temperature

    standard_name :

    unknown

    units :

    K

    Array Chunk Bytes 29.12 MiB 410.62 kiB Shape (636240, 3, 4) (8760, 3, 4) Dask graph 73 chunks in 30 graph layers Data type float32 numpy.ndarray

  • tp
    (time, latitude, longitude)
    float32
    dask.array<chunksize=(8760, 3, 4), meta=np.ndarray>

    GRIB_NV :

    0

    GRIB_Nx :

    1171

    GRIB_Ny :

    701

    GRIB_cfName :

    unknown

    GRIB_cfVarName :

    tp

    GRIB_dataType :

    fc

    GRIB_gridDefinitionDescription :

    Latitude/Longitude Grid

    GRIB_gridType :

    regular_ll

    GRIB_iDirectionIncrementInDegrees :

    0.1

    GRIB_iScansNegatively :

    0

    GRIB_jDirectionIncrementInDegrees :

    0.1

    GRIB_jPointsAreConsecutive :

    0

    GRIB_jScansPositively :

    0

    GRIB_latitudeOfFirstGridPointInDegrees :

    85.0

    GRIB_latitudeOfLastGridPointInDegrees :

    15.0

    GRIB_longitudeOfFirstGridPointInDegrees :

    -167.0

    GRIB_longitudeOfLastGridPointInDegrees :

    -50.0

    GRIB_missingValue :

    9999

    GRIB_name :

    Total precipitation

    GRIB_numberOfPoints :

    820871

    GRIB_paramId :

    228

    GRIB_shortName :

    tp

    GRIB_stepType :

    accum

    GRIB_stepUnits :

    1

    GRIB_totalNumber :

    0

    GRIB_typeOfLevel :

    surface

    GRIB_units :

    m

    coordinates :

    number time step surface latitude longitude valid_time

    long_name :

    Total precipitation

    standard_name :

    unknown

    units :

    m

    Array Chunk Bytes 29.12 MiB 410.62 kiB Shape (636240, 3, 4) (8760, 3, 4) Dask graph 73 chunks in 30 graph layers Data type float32 numpy.ndarray

• Indexes: (3)
  • latitude
    PandasIndex

    PandasIndex(Float64Index([46.20000000000094, 46.10000000000094, 46.00000000000094], dtype='float64', name='latitude'))

  • longitude
    PandasIndex

    PandasIndex(Float64Index([-71.60000000000542, -71.50000000000543, -71.40000000000543,
                  -71.30000000000544],
                 dtype='float64', name='longitude'))

  • time
    PandasIndex

    PandasIndex(DatetimeIndex(['1950-01-01 00:00:00', '1950-01-01 01:00:00',
                   '1950-01-01 02:00:00', '1950-01-01 03:00:00',
                   '1950-01-01 04:00:00', '1950-01-01 05:00:00',
                   '1950-01-01 06:00:00', '1950-01-01 07:00:00',
                   '1950-01-01 08:00:00', '1950-01-01 09:00:00',
                   ...
                   '2022-07-31 14:00:00', '2022-07-31 15:00:00',
                   '2022-07-31 16:00:00', '2022-07-31 17:00:00',
                   '2022-07-31 18:00:00', '2022-07-31 19:00:00',
                   '2022-07-31 20:00:00', '2022-07-31 21:00:00',
                   '2022-07-31 22:00:00', '2022-07-31 23:00:00'],
                  dtype='datetime64[ns]', name='time', length=636240, freq=None))

• Attributes: (0)

%%time
(ds.t2m \
 .mean(['latitude','longitude']) \
 .hvplot(grid=True, color='green') + \
ds.tp \
 .mean(['latitude','longitude']) \
 .hvplot(grid=True, color='blue') + \
ds.sd \
 .mean(['latitude','longitude']) \
 .hvplot(grid=True, color='turquoise') + \
da.value.sel(data_type='flow') \
 .dropna('spatial_agg', how='all') \
 .squeeze() \
 .dropna('time') \
 .rename('flow') \
 .hvplot(grid=True, color='red')) \
.cols(1) \
.opts(shared_axes=True)

CPU times: user 8.78 s, sys: 747 ms, total: 9.53 s
Wall time: 44.1 s