cfgrib

本节介绍如何使用 cfgrib 加载 GRIB2 要素场，并使用 cedarkit-maps 绘图。

安装

使用 conda 安装 cfgrib

conda install -c conda-forge cfgrib

准备

导入需要的包

import xarray as xr
import pandas as pd
import cfgrib

设置绘图的数据参数，使用 CMA-MESO 2024 年 11 月 14 日 00 时次 024 时效数据。

system_name = "CMA-MESO"
data_type = "cma_meso_3km/grib2/orig"
start_time = pd.to_datetime("2024-11-14 00:00:00")
forecast_time = pd.to_timedelta("24h")

加载数据

设置 GRIB2 数据文件路径

file_path = '/g3/COMMONDATA/OPER/CEMC/MESO_3KM/Prod-grib/2024111400/ORIG/rmf.hgra.2024111400024.grb2'
file_path

'/g3/COMMONDATA/OPER/CEMC/MESO_3KM/Prod-grib/2024111400/ORIG/rmf.hgra.2024111400024.grb2'

注：可以使用 reki 库查找本地文件路径

Show code cell source Hide code cell source

from reki.data_finder import find_local_file

file_path_use_reki = find_local_file(
    data_type,
    start_time=start_time,
    forecast_time=forecast_time,
)
file_path_use_reki

Show code cell output Hide code cell output

PosixPath('/g3/COMMONDATA/OPER/CEMC/MESO_3KM/Prod-grib/2024111400/ORIG/rmf.hgra.2024111400024.grb2')

使用 cfgrib 加载 2 米温度要素场。

说明：

1. index="" 表示不生成 cfgrib 的索引文件

2. filter_by_keys 用于设置筛选要素场的条件，这里选择 shortName 为 2t 的场

ds = xr.open_dataset(
    file_path,
    engine="cfgrib",
    backend_kwargs={
        "filter_by_keys": {
            "shortName": "2t"
        },
        "indexpath": "",
    }
)
t_2m_field = ds["t2m"] - 273.15
t_2m_field

<xarray.DataArray 't2m' (latitude: 1671, longitude: 2501)> Size: 17MB
array([[-19.797974 , -19.870972 , -19.942978 , ..., -20.505981 ,
        -19.986984 , -19.959976 ],
       [ -6.6259766,  -6.7809753,  -6.383972 , ..., -21.558975 ,
        -21.516983 , -20.273972 ],
       [ -6.434967 ,  -6.696991 ,  -5.9329834, ..., -21.054977 ,
        -21.480972 , -20.025986 ],
       ...,
       [ 27.062012 ,  27.073029 ,  27.068024 , ...,  27.54303  ,
         27.550018 ,  27.659027 ],
       [ 27.096008 ,  27.051025 ,  27.068024 , ...,  27.57602  ,
         27.572021 ,  27.66101  ],
       [ 27.044037 ,  27.058014 ,  27.051025 , ...,  27.549011 ,
         27.572021 ,  27.66101  ]], dtype=float32)
Coordinates:
    time               datetime64[ns] 8B ...
    step               timedelta64[ns] 8B ...
    heightAboveGround  float64 8B ...
  * latitude           (latitude) float64 13kB 60.1 60.07 60.04 ... 10.03 10.0
  * longitude          (longitude) float64 20kB 70.0 70.03 70.06 ... 145.0 145.0
    valid_time         datetime64[ns] 8B ...

xarray.DataArray

't2m'

• latitude: 1671
• longitude: 2501

• -19.8 -19.87 -19.94 -20.02 -20.1 ... 27.54 27.58 27.55 27.57 27.66

  array([[-19.797974 , -19.870972 , -19.942978 , ..., -20.505981 ,
          -19.986984 , -19.959976 ],
         [ -6.6259766,  -6.7809753,  -6.383972 , ..., -21.558975 ,
          -21.516983 , -20.273972 ],
         [ -6.434967 ,  -6.696991 ,  -5.9329834, ..., -21.054977 ,
          -21.480972 , -20.025986 ],
         ...,
         [ 27.062012 ,  27.073029 ,  27.068024 , ...,  27.54303  ,
           27.550018 ,  27.659027 ],
         [ 27.096008 ,  27.051025 ,  27.068024 , ...,  27.57602  ,
           27.572021 ,  27.66101  ],
         [ 27.044037 ,  27.058014 ,  27.051025 , ...,  27.549011 ,
           27.572021 ,  27.66101  ]], dtype=float32)

• Coordinates: (6)
  • time
    ()
    datetime64[ns]
    ...

    long_name :

    initial time of forecast

    standard_name :

    forecast_reference_time

    [1 values with dtype=datetime64[ns]]

  • step
    ()
    timedelta64[ns]
    ...

    long_name :

    time since forecast_reference_time

    standard_name :

    forecast_period

    [1 values with dtype=timedelta64[ns]]

  • heightAboveGround
    ()
    float64
    ...

    long_name :

    height above the surface

    units :

    m

    positive :

    up

    standard_name :

    height

    [1 values with dtype=float64]

  • latitude
    (latitude)
    float64
    60.1 60.07 60.04 ... 10.03 10.0

    units :

    degrees_north

    standard_name :

    latitude

    long_name :

    latitude

    stored_direction :

    decreasing

    array([60.099998, 60.069998, 60.039998, ..., 10.059998, 10.029998, 10.      ])

  • longitude
    (longitude)
    float64
    70.0 70.03 70.06 ... 145.0 145.0

    units :

    degrees_east

    standard_name :

    longitude

    long_name :

    longitude

    array([ 70.  ,  70.03,  70.06, ..., 144.94, 144.97, 145.  ])

  • valid_time
    ()
    datetime64[ns]
    ...

    standard_name :

    time

    long_name :

    time

    [1 values with dtype=datetime64[ns]]

• Indexes: (2)
  • latitude
    PandasIndex

    PandasIndex(Index([         60.099998,          60.069998,          60.039998,
           60.009997999999996, 59.979997999999995, 59.949997999999994,
            59.91999799999999,  59.88999799999999,  59.85999799999999,
            59.82999799999999,
           ...
           10.269997999998452, 10.239997999998453, 10.209997999998453,
           10.179997999998454, 10.149997999998455, 10.119997999998455,
           10.089997999998456, 10.059997999998457, 10.029997999998457,
                         10.0],
          dtype='float64', name='latitude', length=1671))

  • longitude
    PandasIndex

    PandasIndex(Index([              70.0,              70.03,              70.06,
                        70.09,              70.12,              70.15,
                        70.18,  70.21000000000001,  70.24000000000001,
            70.27000000000001,
           ...
           144.73000000000283, 144.76000000000283, 144.79000000000283,
           144.82000000000284, 144.85000000000284, 144.88000000000284,
           144.91000000000284, 144.94000000000284, 144.97000000000284,
                        145.0],
          dtype='float64', name='longitude', length=2501))

• Attributes: (0)

绘图

使用 cedarkit-maps 绘制 2 米温度填充图

Show code cell source Hide code cell source

from cedarkit.maps.style import ContourStyle
from cedarkit.maps.chart import Panel
from cedarkit.maps.domains import EastAsiaMapTemplate
from cedarkit.maps.colormap import get_ncl_colormap

t_2m_level = [-24, -20, -16, -12, -8, -4, 0, 4, 8, 12, 16, 20, 24, 28, 32]
color_index = [2, 12, 22, 32, 42, 52, 62, 72, 82, 92, 102, 112, 122, 132, 142, 152]
t_2m_color_map = get_ncl_colormap("BlAqGrYeOrReVi200", index=color_index)
t_2m_style = ContourStyle(
    colors=t_2m_color_map,
    levels=t_2m_level,
    fill=True,
)
domain = EastAsiaMapTemplate()
panel = Panel(domain=domain)
panel.plot(t_2m_field, style=t_2m_style)
domain.set_title(
    panel=panel,
    graph_name="2m Temperature (C)",
    system_name=system_name,
    start_time=start_time,
    forecast_time=forecast_time,
)
domain.add_colorbar(panel=panel, style=t_2m_style)
panel.show()