Wind_data

Wind_data is a Python library to download data from the Climate Data Store, but currently only supports very few of the available datasets. For now, it is oriented to download data from the 'reanalysis-era5-single-levels' and 'reanalysis-era5-land' datasets and process the velocity at 10m.

Set up

As the data are downloaded through the Climate Data Store API, you need to follow a few steps:

• create an account here, and log in.
• setup the API key on your computer: see here (just follow the steps under Install the CDS API key)
• install the lib on your computer: As for most python package, use pip. In a terminal run pip3 install --upgrade https://github.com/Cgadal/Wind_data/tarball/master. If you also have git installed on your computer, you may prefer to use: pip3 install git+https://github.com/Cgadal/Wind_data.git.

General documentation can be found on the ECMWWF website, for ERA5 and ERA5Land.

Note that I am quite unsure on how this is going to work on other O.S. than Linus distributions.

Quick example

Script

import Wind_data.Era_5 as Era5
import numpy as np

########### CReating wind data object
Test = Era5.Wind_data('test_place', type = 'reanalysis-era5-single-levels')

### Specifying request
variables = ['10m_u_component_of_wind','10m_v_component_of_wind']  ## variables
year = [str(i) for i in np.arange(2010, 2011 + 1)]   ## years: from 2012 to 2018
month = ['{:02d}'.format(i) for i in np.arange(1, 12 + 1)] ## months: all of them
day = ['{:02d}'.format(i) for i in np.arange(1, 31 + 1)] ## days: all of them
time =  ['{:02d}'.format(i) +':00' for i in np.arange(0, 23 + 1)] ## hours: all of them
area = [-24, 15, -24, 15] ## geographical subset

variable_dic = {'product_type':'reanalysis',
                'format': 'netcdf',
                'variable': variables,
                'year': year,
                'month': month,
                'day': day,
                # 'grid': 0.1 # note that you can specify the grid span you want, and the CDS will interpolate for you. Native grid is 0.25 for ERA5 and 0.1 for ERA5Land.
                'time': time,
                'area': area}

### launching request
Test.Getting_wind_data(variable_dic)

##### Creating a KML file to show spatial points on Google Earth
Test.Update_coordinates()
Test.Write_coordinates()
Test.Create_KMZ()
##

Optional keywords

The functions have plenty of optional keywords, I will write a documentation as soon as possible. In the mean time, here is a list of usefull keywords in Test.Getting_wind_data() to play with.

• save_to_npy = True: if you want to convert netcdf files to .npy files. True is only supported for 10m velocities. Default is True.
• on_grid = True: if you want to have the spacial points given moved to correspond to the original grid of the downloaded dataset. Default it True.
• file = 'info.txt': Name of the file some informations will be written to.
• Nsplit = 1: Number of sub-requests you request will be split to. Theroretically, it is automatically calculated by the function. If an error ocurr, try to increase it. Default is 1 but replaced if found to be too low.

Differences between datasets

The different datasets have differences in the variable_dic passed to Test.Getting_wind_data(). In particular, for the 'reanalysis-era5-land' dataset, you need to remove the entry 'product_type':'reanalysis'. More details can be found on the original documentation of the CDSAPI.

Treatment of the wind data

For now, the library also integrates a few functions to calculate sand fluxes from the 10 velocity, depending on several parameters. What comes next are typical commands that may follow the Quick example presented above.

Script

############### This part is only needed if you run what follows in a different script

# import Wind_data.Era_5 as Era5
# import numpy as np
#
# ########### CReating wind data object
# Test = Era5.Wind_data('test_place', type = 'reanalysis-era5-single-levels') ## create the python object
# Test.load_spec()                                                            ## Load spec previously saved in 'info.txt' (or another file you might have chosen)

############## Otherwise just append this to the first script 'Quick example'

## Load the netcdf files (whose names are stored in Test.file_names.)
Test.Load_netcdf(Test.file_names, save_to_npy = False)

## Convert velocities to polar coordinates
Test.Cartesian_to_polar()

## Calculate sand fluxes.
# Optional keywords are the grain diameter [m] (grain_size = 180*10**-6), hydro. roughness [m] (z_0 = 1e-3), height of wind data [m] (z = 10), air density (rhoair = 1.293), grain density (rhosed = 2.55e3).
Test.Calculate_fluxes()

## Plot wind roses and flux roses
Test.Write_wind_rose('wind_roses')
Test.Write_flux_rose('flux_roses')