Time Slice Python Xarray Dataarray

Question

I need to know if I'm time-slicing the xarray data from Jan 1991 through Dec 2021 that I have. The coordinates look like this:

Coordinates:
 * time       (time) datetime64[ns] 1991-01-31 1991-02-28 ... 2021-12-31
number     int32 0
step       timedelta64[ns] 00:00:00
surface    float64 0.0
 * latitude   (latitude) float64 58.0 57.75 57.5 57.25 ... 23.5 23.25 23.0
 * longitude  (longitude) float64 -130.0 -129.8 -129.5 ... -63.5 -63.25 -63.0

The line of code that I'm using to slice through the dataarray (resultm) looks like this -

month_curr = resultm.sel(time=slice('2021-12','2021-12')).groupby('time.month').mean(dim='time')

And, my objective is to slice or extract all the December 2021 data - which should be a monthly value. It looks like the data might be in daily form but the download type is 'monthly_averaged_reanalysis' ERA5 data.

Thank you,

Answer 1

You can select the relevant data using the datetime accessor .dt where you need to combine both dt.month and dt.year using numpy.logical_and to generate a boolean index which corresponds to the required indices.

For your example, to generate a monthly mean of Dec 2021 you could do:

import numpy as np

month_curr = resultm.sel(
    time=np.logical_and(
        resultm.time.dt.year == 2021, resultm.time.dt.month == 12
    )
)
month_curr = month_curr.mean("time")

Here's a toy example (using the year 2013):

import xarray as xr
import numpy as np

x = xr.tutorial.load_dataset("air_temperature")

xs = x.sel(
    time=np.logical_and(x.time.dt.month == 12, x.time.dt.year == 2013)).mean("time")