Concatenate IVT from MERRA2 reanalysis

[laisgfernandes]

Hello all,

I am trying to concatenate MERRA2 files using iris, but I am having problems. Here is my code:

#####################################
"""Concatenate IVT from the MERRA2
reanalysis."""

import iris
import numpy as np
import matplotlib as mpl
from iris.util import unify_time_units
from iris.util import equalise_attributes

def clean_callback(cube,field,filename):
    cube.coord('time').attributes = {}

#read meridional IVT file.
IVT_file = '/Volumes/climate_lab2/MERRA2/Daily_and_Subdaily/IVT_daily/MERRA2.tavg1_2d_int_Nx.20180101.SUB.nc'

IVT = iris.load_cube(IVT_file, 'northward_flux_of_atmospheric_water_vapor', callback=clean_callback)

dd = np.arange('2018-01-02', '2019-01-01', dtype='datetime64')

dd = np.datetime_as_string(dd)

hyphenlist = dd
newlist = []

for x in hyphenlist:
    newlist.append(x.replace('-', ''))

dd = newlist

for i in dd:
    print(i)
    #read meridional IVT file.
    IVT_file = '/Volumes/climate_lab2/MERRA2/Daily_and_Subdaily/IVT_daily/MERRA2.tavg1_2d_int_Nx.'+str(i)+'.SUB.nc'
    IVT_n = iris.load_cube(IVT_file, 'northward_flux_of_atmospheric_water_vapor', callback=clean_callback)

    cube = iris.cube.CubeList([IVT,IVT_n])
    iris.util.unify_time_units(cube)
    equalise_attributes(cube)
    IVT = cube.concatenate_cube()

iris.save(IVT,'/Users/lais3/Desktop/IVT_conc/IVT_merid/MERRA2.tavg1_2d_int_Nx.2018.nc')
#######################################

Python is returning:
ConcatenateError: failed to concatenate into a single cube.
Dimension coordinates metadata differ: time != time

I believe the issue is about the dtype of the time coordinate. After type:

print(IVT.coord('time'))

Python shows me that the dtype for time is int32. For older files, it was float64, and I was being able to concatenate.

Please, I need to learn how I change the dtype from int32 to float64. I couldn't find an explanation in the iris user guide.

Many thanks for your assistance.
Cheers,
Laís

[trexfeathers]

Hi @laisgfernandes. The values in a DimCoord (DimCoord.points) are a standard NumPy ndarray, so the correct way of changing type is to use ndarray.astype() on the DimCoord's points, e.g.:

import numpy as np

my_coord = IVT.coord("time")
my_coord.points = my_coord.points.astype(np.float64)

[laisgfernandes]

Hi @trexfeathers,

Thanks for your reply! I know I had to use astype for NumPy. I forgot to use .points to apply Numpy functions. Thank you very much for remembering me. I could concatenate the data after that.

Cheers,
Laís