使用xarray求月平均值

Question

我正在处理GLDAS. 时期的数据重新分析1 year。这些文件是.nc4. 我可以打开文件，但我不能groupby给出1 month。我不想手动或通过 for，但我发现确实xarray如此groupby。我的脚本是：

\n\n

In[16]:import xarray as xr\n\nIn[17]:gldas = xr.open_mfdataset('./GLDAS_2010/*.nc4', chunks=None, concat_dim='time', preprocess=None, engine='netcdf4', lock=None,compat='minimal',coords='minimal',data_vars='minimal')\n\nIn[18]: gldas\nOut[18]: \n<xarray.Dataset>\nDimensions:                (bnds: 2, lat: 40, lon: 48, time: 365)\nCoordinates:\n  * lat                    (lat) float32 -34.875 -34.625 -34.375 -34.125 ...\n  * lon                    (lon) float32 -59.875 -59.625 -59.375 -59.125 ...\n  * time                   (time) float64 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 ...\nDimensions without coordinates: bnds\nData variables:\n    Albedo_inst            (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    AvgSurfT_inst          (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    CanopInt_inst          (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    ECanop_tavg            (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    ESoil_tavg             (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    Evap_tavg              (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    LWdown_f_tavg          (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    Lwnet_tavg             (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    PotEvap_tavg           (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    Psurf_f_inst           (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    Qair_f_inst            (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    Qg_tavg                (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    Qh_tavg                (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    Qle_tavg               (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    Qs_acc                 (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    Qsb_acc                (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    Qsm_acc                (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    Rainf_f_tavg           (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    Rainf_tavg             (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    RootMoist_inst         (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    SWE_inst               (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    SWdown_f_tavg          (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    SnowDepth_inst         (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    Snowf_tavg             (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    SoilMoi0_10cm_inst     (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    SoilMoi100_200cm_inst  (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    SoilMoi40_100cm_inst   (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    SoilTMP0_10cm_inst     (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    SoilTMP100_200cm_inst  (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    SoilTMP10_40cm_inst    (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    SoilTMP40_100cm_inst   (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    Swnet_tavg             (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    Tair_f_inst            (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    Tveg_tavg              (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    Wind_f_inst            (time, lat, lon) float32 dask.array<shape=(365, 40, 48), chunksize=(1, 40, 48)>\n    time_bnds              (time, bnds) float64 dask.array<shape=(365, 2), chunksize=(1, 2)>\nAttributes:\n    CDI:                    Climate Data Interface version 1.6.9 (http://mpim...\n    Conventions:            CF-1.4\n    history:                Mon Feb  5 20:07:25 2018: /usr/bin/ncks -O -L 1 /...\n    source:                 Noah_v3.3\n    institution:            NASA GSFC\n    missing_value:          -9999.0\n    tavg definision::       past 3-hour average\n    acc definision::        past 3-hour accumulation\n    inst definision::       instantaneous\n    title:                  GLDAS2.1 LIS land surface model output\n    references:             Rodell_etal_BAMS_2004, Kumar_etal_EMS_2006, Peter...\n    conventions:            CF-1.6\n    comment:                website: http://ldas.gsfc.nasa.gov/gldas, http://...\n    MAP_PROJECTION:         EQUIDISTANT CYLINDRICAL\n    SOUTH_WEST_CORNER_LAT:  -59.875\n    SOUTH_WEST_CORNER_LON:  -179.875\n    DX:                     0.25\n    DY:                     0.25\n    CDO:                    Climate Data Operators version 1.6.9 (http://mpim...\n    NCO:                    20180205\n

\n\n

当我尝试时：

\n\n

In[19]:pp.set_index('time').groupby(freq='1M').mean('time')\n

\n\n

给出这个错误：

\n\n

TypeError: groupby() got an unexpected keyword argument 'freq'\n

\n\n

好吧，知道有这个论点，我将进行下一次尝试，即：

\n\n

In[20]:import pandas as pd\nIn[21]:pp.groupby(['Albedo_inst',pd.Grouper(key='\xe2\x80\x8c\xe2\x80\x8btime', freq='M')]).mean()\n

\n\n

错误是：

\n\n

TypeError: `group` must be an xarray.DataArray or the name of an xarray variable or dimension\n

\n\n

最后一次尝试没有频率：

\n\n

In[22]:pp.groupby('Albedo_inst').mean('time')\n

\n\n

和错误：

\n\n

ValueError: Level values must be unique: [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0] on level 0\n

\n\n

我尝试：

\n\n

gldas.resample('1MS', dim='time', how='mean')\n

\n\n

错误是：

\n\n

TypeError: Only valid with DatetimeIndex, TimedeltaIndex or PeriodIndex, but got an instance of 'Float64Index'\n

\n\n

我尝试：

\n\n

gldas.groupby('time.month').mean('time')\n

\n\n

错误是：

\n\n

AttributeError: 'IndexVariable' object has no attribute 'month'\n

\n\n

有人会帮助我吗？多谢你们！

\n\n

编辑：

\n\n

In [24]: pp['time']\nOut[24]: \n<xarray.DataArray 'time' (time: 365)>\narray([0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n       0., 0., 0., 0., 0., 0., 0., 0.])\nCoordinates:\n  * time     (time) float64 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 ...\nAttributes:\n    standard_name: time\n    bounds: time_bnds\n    units: day as %Y%m%d.%f\n    calendar: proleptic_gregorian\n

\n\n

编辑：

\n\n

如果我打开一个文件：

\n\n

In[108]:gldas\nOut[108]: \n<xarray.Dataset>\nDimensions:                (bnds: 2, lat: 40, lon: 48, time: 1)\nCoordinates:\n  * lat                    (lat) float32 -34.875 -34.625 -34.375 -34.125 ...\n  * lon                    (lon) float32 -59.875 -59.625 -59.375 -59.125 ...\n  * time                   (time) float64 0.0\n  * bnds                   (bnds) int64 0 1\nData variables:\n    Albedo_inst            (time, lat, lon) float64 15.66 16.0 16.23 15.98 ...\n    AvgSurfT_inst          (time, lat, lon) float64 302.8 302.3 302.3 302.4 ...\n    CanopInt_inst          (time, lat, lon) float64 0.001 0.0 0.0 0.0 0.0 ...\n    ECanop_tavg            (time, lat, lon) float64 2.48 0.89 0.16 0.0 0.02 ...\n    ESoil_tavg             (time, lat, lon) float64 46.39 46.58 49.18 55.22 ...\n    Evap_tavg              (time, lat, lon) float64 0.0001668 0.0001683 ...\n    LWdown_f_tavg          (time, lat, lon) float64 369.5 368.3 366.8 365.6 ...\n    Lwnet_tavg             (time, lat, lon) float64 -88.05 -87.24 -87.68 ...\n    PotEvap_tavg           (time, lat, lon) float64 622.3 610.4 601.1 599.6 ...\n    Psurf_f_inst           (time, lat, lon) float64 1.004e+05 1.006e+05 ...\n    Qair_f_inst            (time, lat, lon) float64 0.01117 0.01144 0.01172 ...\n    Qg_tavg                (time, lat, lon) float64 54.2 53.65 55.13 58.52 ...\n    Qh_tavg                (time, lat, lon) float64 110.4 101.9 104.5 114.7 ...\n    Qle_tavg               (time, lat, lon) float64 417.2 420.8 416.1 405.6 ...\n    Qs_acc                 (time, lat, lon) float64 0.0 0.0 0.0 0.0 0.0 0.0 ...\n    Qsb_acc                (time, lat, lon) float64 0.00253 0.00335 0.00528 ...\n    Qsm_acc                (time, lat, lon) float64 0.0 0.0 0.0 0.0 0.0 0.0 ...\n    Rainf_f_tavg           (time, lat, lon) float64 1.4e-06 5e-07 1e-07 0.0 ...\n    Rainf_tavg             (time, lat, lon) float64 1.4e-06 5e-07 1e-07 0.0 ...\n    RootMoist_inst         (time, lat, lon) float64 255.3 259.6 263.9 266.1 ...\n    SWE_inst               (time, lat, lon) float64 0.0 0.0 0.0 0.0 0.0 0.0 ...\n    SWdown_f_tavg          (time, lat, lon) float64 793.9 789.7 791.5 794.6 ...\n    SnowDepth_inst         (time, lat, lon) float64 0.0 0.0 0.0 0.0 0.0 0.0 ...\n    Snowf_tavg             (time, lat, lon) float64 0.0 0.0 0.0 0.0 0.0 0.0 ...\n    SoilMoi0_10cm_inst     (time, lat, lon) float64 24.86 25.17 25.47 25.51 ...\n    SoilMoi100_200cm_inst  (time, lat, lon) float64 254.4 258.8 266.1 274.1 ...\n    SoilMoi40_100cm_inst   (time, lat, lon) float64 153.7 156.7 159.8 161.6 ...\n    SoilTMP0_10cm_inst     (time, lat, lon) float64 293.8 293.5 293.4 293.7 ...\n    SoilTMP100_200cm_inst  (time, lat, lon) float64 291.3 291.0 290.9 290.9 ...\n    SoilTMP10_40cm_inst    (time, lat, lon) float64 292.4 292.2 292.1 292.2 ...\n    SoilTMP40_100cm_inst   (time, lat, lon) float64 292.2 292.0 291.8 291.9 ...\n    Swnet_tavg             (time, lat, lon) float64 669.5 663.3 663.0 667.7 ...\n    Tair_f_inst            (time, lat, lon) float64 300.2 300.0 299.8 299.5 ...\n    Tveg_tavg              (time, lat, lon) float64 368.3 373.3 366.8 350.4 ...\n    Wind_f_inst            (time, lat, lon) float64 3.704 3.704 3.704 3.704 ...\n    time_bnds              (time, bnds) float64 0.0 0.0\nAttributes:\n    CDI: Climate Data Interface version 1.6.9 (http://mpimet.mpg.de/cdi)\n    Conventions: CF-1.4\n    history: Mon Feb  5 20:07:25 2018: /usr/bin/ncks -O -L 1 /tmpdata/regridder/services_88661/cdoGLDAS_NOAH025_3H.A20100101.1500.021.SUB.nc4 /tmpdata/regridder/services_88661/deflatecdoGLDAS_NOAH025_3H.A20100101.1500.021.SUB.nc4\nMon Feb 05 20:07:21 2018: cdo -s -L -f nc4 -sellonlatbox,-60.0,-48.0,-35.0,-25.0 -selname,Albedo_inst,AvgSurfT_inst,CanopInt_inst,ECanop_tavg,ESoil_tavg,Evap_tavg,LWdown_f_tavg,Lwnet_tavg,PotEvap_tavg,Psurf_f_inst,Qair_f_inst,Qg_tavg,Qh_tavg,Qle_tavg,Qs_acc,Qsb_acc,Qsm_acc,Rainf_...\n    source: Noah_v3.3\n    institution: NASA GSFC\n    missing_value: -9999.0\n    tavg definision:: past 3-hour average\n    acc definision:: past 3-hour accumulation\n    inst definision:: instantaneous\n    title: GLDAS2.1 LIS land surface model output\n    references: Rodell_etal_BAMS_2004, Kumar_etal_EMS_2006, Peters-Lidard_etal_ISSE_2007\n    conventions: CF-1.6\n    comment: website: http://ldas.gsfc.nasa.gov/gldas, http://lis.gsfc.nasa.gov/\n    MAP_PROJECTION: EQUIDISTANT CYLINDRICAL\n    SOUTH_WEST_CORNER_LAT: -59.875\n    SOUTH_WEST_CORNER_LON: -179.875\n    DX: 0.25\n    DY: 0.25\n    CDO: Climate Data Operators version 1.6.9 (http://mpimet.mpg.de/cdo)\n    NCO: 20180205/usr/local/lib/python3.6/dist-packages/spyder/widgets/variableexplorer/utils.py:414: FutureWarning: 'summary' is deprecated and will be removed in a future version.\n  display = value.summary()\n

\n\n

结论：\n我明白了。如果有人需要帮助，请给我留言。

\n

Answer 1

如果您没有有效的时间变量（上面示例中的时间始终为 0...），您必须修复该问题。但一旦完成，我认为您正在寻找的行是：

gldas['Albedo_inst'].resample(time="1MS").mean(dim="time")