xarray - 使用 groupby 对一年的气候每小时 netCDF 数据中的每一天进行分组

Question

我有一个地理范围超过一年的每小时 netCDF 气候数据，例如从 2017-01-01T00:00:00 到 2017-12-31T23:00:00.

<xarray.Dataset>
Dimensions:    (latitude: 106, longitude: 193, time: 8760)
Coordinates:
  * latitude   (latitude) float32 -39.2 -39.149525 ... -33.950478 -33.9
  * longitude  (longitude) float32 140.8 140.84792 140.89584 ... 149.95209 150.0
  * time       (time) datetime64[ns] 2017-01-01 ... 2017-12-31T23:00:00
Data variables:
    T_SFC      (time, latitude, longitude) float32 dask.array<shape=(8760, 106, 193), chunksize=(744, 106, 193)>
Attributes:
    creationTime:        1525708833
    creationTimeString:  Mon May  7 09:00:32 PDT 2018
    Conventions:         COARDS

正如它所说的那样，数据具有三个坐标（纬度、经度和时间）和一个变量，即每小时温度。

我的代码：

import xarray as xr
mds_temp_path = '../Archive/*/IDV71000_VIC_T_SFC.nc'    # netCDF
mds_temp = xr.open_mfdataset(mds_temp_path)    # open netCDF and read into a dataset object

print(mds_temp.groupby('time.dayofyear').mean('time'))

我得到的：

<xarray.Dataset>
Dimensions:    (dayofyear: 365, latitude: 106, longitude: 193)
Coordinates:
  * latitude   (latitude) float32 -39.2 -39.149525 ... -33.950478 -33.9
  * longitude  (longitude) float32 140.8 140.84792 140.89584 ... 149.95209 150.0
  * dayofyear  (dayofyear) int64 1 2 3 4 5 6 7 8 ... 359 360 361 362 363 364 365
Data variables:
    T_SFC   (dayofyear, latitude, longitude) float64 dask.array<shape=(365, 106, 193), chunksize=(1, 106, 193)>

我想要实现的目标是能够获得每天的平均温度值，例如生成的数据集中的时间坐标为“2017-01-01”、“2017-01-02”、“2017-01-03”、……、“2017-12-31”，而不是 1 , 2, 3, ……, 365.

Answer 1

您应该使用 resample method instead of groupby:

mds_temp.resample(time='1D').mean()

这些概念在文档的时间序列数据部分有更完整的描述：http://xarray.pydata.org/en/stable/time-series.html#resampling-and-grouped-operations

Answer 2

使用@jhamman 的回答很有用，但如果您的测量之间缺少一天，方法 resample(...) 将填补它，请参见此示例：

# create a dataset with data on days 1 & 3
t = ['2000-01-01T00:00:00.000000000', '2000-01-01T01:00:00.000000000', '2000-01-01T02:00:00.000000000', '2000-01-03T00:00:00.000000000', '2000-01-03T01:00:00.000000000', '2000-01-03T02:00:00.000000000']
t = pd.to_datetime(t)
ds = xr.Dataset({"foo": ("time", np.arange(len(t))), "time": t})

# reduce to days (but with interpolation on day 2!)
ds = ds.resample(time='1D').mean()
print(ds.time)

<xarray.DataArray 'time' (time: 3)>
array(['2000-01-01T00:00:00.000000000', '2000-01-02T00:00:00.000000000',
       '2000-01-03T00:00:00.000000000'], dtype='datetime64[ns]')
Coordinates:
  * time     (time) datetime64[ns] 2000-01-01 2000-01-02 2000-01-03

在我的例子中，我不想要这种行为，缺失的天数在按天聚合后必须保持缺失，所以我使用这种方法：

# set all dates to have time at 00h so multiple measurements in a day have the same label
ds.coords['time'] = ds.time.dt.floor('1D')

# group by 'date' using an average (mean)
ds = ds.groupby('time').mean()

<xarray.Dataset>
Dimensions:  (time: 2)
Coordinates:
  * time     (time) datetime64[ns] 2000-01-01 2000-01-03
Data variables:
    foo      (time) float64 1.0 4.0

print(ds)

# set all dates to have time at 00h so multiple measurements in a day have the same label
ds.coords['time'] = ds.time.dt.floor('1D')

# group by 'date' using an average (mean)
ds = ds.groupby('time').mean()

print(ds)
<xarray.Dataset>
Dimensions:  (time: 2)
Coordinates:
  * time     (time) datetime64[ns] 2000-01-01 2000-01-03
Data variables:
    foo      (time) float64 1.0 4.0

希望对大家有所帮助:)

xarray - 使用 groupby 对一年的气候每小时 netCDF 数据中的每一天进行分组

xarray - Use groupby to group by every day over a year's climatological hourly netCDF data

python

netcdf

pandas

python-xarray