1

I am working with some atmospheric model outputs available in NetCDF format. I want to calculate Temperature Advection by 850 hPa winds. I tried using the metpy function https://unidata.github.io/MetPy/latest/api/generated/metpy.calc.advection.html#metpy.calc.advection for the same. I am attaching a part of my code. Please help to resolve this error. All required libraries are updated with their latest versions.

import xarray as xr import numpy as np import matplotlib.pyplot as plt from metpy.units import units import cartopy.crs as ccrs import cartopy.feature as cfeature import metpy.calc as mpcalc

temp = xr.open_dataset('ts.nc') u10 = xr.open_dataset('u10.nc') v10 = xr.open_dataset('v10.nc')

lat = temp['lat'] lon = temp['lon'] lonn, latt = np.meshgrid(lon,lat)

dx, dy = mpcalc.lat_lon_grid_deltas(lon, lat, )

time = np.array(temp.time[0:9])

T1 = temp['ts'][0:9,:,:]

u10_1 = u10['u10'][0:9,:,:]
v10_1 = v10['v10'][0:9,:,:]

T11 = xr.DataArray(data = T1.values, dims = ['time', 'lat', 'lon'], coords=dict(lat=lat, lon = lon, time = time)) u10_11 = xr.DataArray(data = u10_1.values, dims = ['time', 'lat', 'lon'], coords=dict(lat=lat, lon = lon, time = time)) v10_11 = xr.DataArray(data = v10_1.values, dims = ['time', 'lat', 'lon'], coords=dict(lat=lat, lon = lon, time = time))

advec = mpcalc.advection(T1, [u10_1, v10_1], (dx, dy))

[code][1]

Error: AttributeError: crs attribute is not available.

Please let me know if any further references are required.

Somoshree Chatterjee
  • 21
  • 2

2 Answers2

1

Mentioned error (AttributeError: crs attribute is not available) speaks itself the problem in the code.

This error is because of the absence of cartopy projection attribute in the data coordinates. You need to manually add the projection in the data.

T1=T1.metpy.assign_crs(grid_mapping_name='latitude_longitude', earth_radius=6371229.0)

You can pick any one of the projections from this list.

Dharman
  • 30,962
  • 25
  • 85
  • 135
Ankur Kumar
  • 11
  • 1
0

If there are CF projection metadata present in your dataset, MetPy may be able to automatically identify the proper CRS, by calling the parse_cf() method:

v10 = xr.open_dataset('v10.nc')
v10 = v10.metpy.parse_cf()
DopplerShift
  • 5,472
  • 1
  • 21
  • 20