python - DataArray.mean 不保留坐标

Question

DataArray.mean 不保留取决于应用平均值的维度的坐标。

注意:XLAT 和 XLONG 与时间无关；然而，一些 netcdf 文件具有沿这两个方向的时间坐标。

我有这个 netcdf 文件 wrfout_d03.nc，我用以下命令打开文件:

ds = xr.open_dataset('/Users/jacob/Desktop/wrfpy/wrfout_d03_may.nc')

然后给出一个 DataSet 对象:

<xarray.Dataset>
Dimensions:                (Time: 193, bio_emissions_dimension_stag: 41, bottom_top: 50, bottom_top_stag: 51, klevs_for_dvel: 1, seed_dim_stag: 12, soil_layers_stag: 4, south_north: 115, south_north_stag: 116, west_east: 115, west_east_stag: 116)
Coordinates:
    XLAT                   (Time, south_north, west_east) float32 ...
    XLONG                  (Time, south_north, west_east) float32 ...
    XTIME                  (Time) datetime64[ns] ...
    XLAT_U                 (Time, south_north, west_east_stag) float32 ...
    XLONG_U                (Time, south_north, west_east_stag) float32 ...
    XLAT_V                 (Time, south_north_stag, west_east) float32 ...
    XLONG_V                (Time, south_north_stag, west_east) float32 ...
Dimensions without coordinates: Time, bio_emissions_dimension_stag, bottom_top, bottom_top_stag, klevs_for_dvel, seed_dim_stag, soil_layers_stag, south_north, south_north_stag, west_east, west_east_stag
Data variables:
datavars...

然后我使用以下代码访问 PM2_5_DRY 变量:

pm25 = ds.PM2_5_DRY

生成的对象 pm25 具有以下尺寸和坐标:

<xarray.DataArray 'PM2_5_DRY' (Time: 193, bottom_top: 50, south_north: 115, west_east: 115)>
[127621250 values with dtype=float32]
Coordinates:
    XLAT     (Time, south_north, west_east) float32 ...
    XLONG    (Time, south_north, west_east) float32 ...
    XTIME    (Time) datetime64[ns] ...
Dimensions without coordinates: Time, bottom_top, south_north, west_east
Attributes:
    FieldType:    104
    MemoryOrder:  XYZ
    description:  pm2.5 aerosol dry mass
    units:        ug m^-3
    stagger: 

然后我操纵 pm25 对象并通过以下方式获得时间维度的平均值:

pm25_mean = pm25.mean(dim='Time', keep_attrs = True)

生成的对象是一个 DataArray，但没有坐标 XLAT 或 XLON。

<xarray.DataArray 'PM2_5_DRY' (bottom_top: 50, south_north: 115, west_east: 115)>
array([[[14.73083   , 14.756626  , 14.796355  , ..., 20.325712  ,
         20.855696  , 21.381271  ],
        [14.651459  , 14.34477   , 14.371858  , ..., 18.00389   ,
         18.4109    , 21.337002  ],
        [14.59026   , 14.257076  , 14.293012  , ..., 17.391146  ,
         18.217058  , 20.882664  ],
        ...,
        [27.356459  , 27.21468   , 27.757051  , ...,  8.084272  ,
          8.010168  ,  7.989942  ],
        [27.185486  , 27.02623   , 27.776043  , ...,  7.944748  ,
          7.8795266 ,  7.8552976 ],
        [26.926008  , 27.724253  , 28.427626  , ...,  7.8269224 ,
          7.773637  ,  7.741844  ]],

Dimensions without coordinates: bottom_top, south_north, west_east
Attributes:
    FieldType:    104
    MemoryOrder:  XYZ
    description:  pm2.5 aerosol dry mass
    units:        ug m^-3
    stagger:

检查以下代码:

pm25_mean.coords

Coordinates:
*empty*

我试着查看 xarray 中 mean 函数的文档；但是，我无法找到任何选项来将坐标从前一个对象复制到新对象。

关于如何进行此操作的任何提示？我在想我需要从文件中访问这些坐标，然后再次组合它们。但我不确定如何完成这个过程。

还有，跟这个有关系吗？

XLAT     (Time, south_north, west_east) float32

XLAT 是一个多维坐标，它也取决于时间。由于我在时间维度上获得了平均值，pm25 的维度数量从 4 个减少到 3 个。这是否也以某种方式影响了坐标？

我需要最终对象具有坐标 XLAT 和 XLONG，因为我会将其可视化。

感谢您的帮助!

Answer 1

您只需将坐标变量从 3 维转换为 2 维。

d = xr.open_dataset('.../pm25_sample.nc')
d['XLAT'] = d.XLAT.mean(dim = 'Time')
d['XLONG'] = d.XLONG.mean(dim = 'Time')

d['PM2_5_DRY'].mean(dim = 'Time')

     ...,
    [ 0.03839084,  0.03837739,  0.03835952, ...,  0.03414929,
      0.03412561,  0.03410038],
    [ 0.03837854,  0.03836632,  0.03834687, ...,  0.03414606,
      0.0341224 ,  0.03409675],
    [ 0.03836945,  0.03835024,  0.03833132, ...,  0.03414177,
      0.03411727,  0.03409337]]], dtype=float32)
Coordinates:
    XLAT     (south_north, west_east) float32 14.086891 14.086907 ... 15.111996
    XLONG    (south_north, west_east) float32 120.49799 120.50717 ... 121.55791
Dimensions without coordinates: bottom_top, south_north, west_east