从python中的每日netcdf文件计算每月平均值

Question

Hello I have a netcdf file with daily data. Shape of the file is (5844, 89, 89) ie 16 years data. I tried to get monthly average from daily data. I am looking for simillar to resample function in pandas dataframe. Is there anyways to do that in python. As I know it is very easy to calculate by using cdo and nco but I am looking in python.

Sample code that I used to read netcdf file is:

import netCDF4
from netCDF4 import Dataset
fh = Dataset(ncfile, mode='r')
time = fh.variables['time'][:]
lon = fh.variables['longitude'][:]
lat = fh.variables['latitude'][:]
data = fh.variables['t2m'][:]
data.shape

Answer 1

@ jhamman Thank you for suggesting xarray.resample . It is simpler than I thought and the answer to my question is:

import xarray as xr
ds = xr.open_dataset(nc_file)
monthly_data = ds.resample(freq = 'm', dim = 'time', how = 'mean')

Answer 2

新版本的xarray，用法简单多了，如下

monthly_data=ds.resample(time='m').mean()

Answer 3

If you are working in Linux or macOS this can be easily done using nctoolkit, which uses CDO as a backend. (Installation instructions here ).

If you want to get the monthly mean, you just need the following:

import nctoolkit as nc
data = nc.open_data(ncfile)
data.tmean(["year", "month"])

This can be plotted:

data.plot()

If you wanted to convert it to a pandas dataframe:

df = data.to_dataframe()

Answer 4

May I also suggest the following solution:

data = fh.variables['t2m'][:] # The t2m variable is pulled out of the file as 'data'
stat1 = data.mean() # mean
stat2 = data.std() # standard dev