Skip to content
Python library for dual-Doppler uncertainty assessment
Python TeX
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.


Yet Another Dual-Doppler Uncertainty Model

GitHub release (latest by date) DOI Binder License

A Python package for simple and fast uncertainty assessment of dual-Doppler retrievals of wind speed and wind direction.

Table of Contents


YADDUM is focused on delivering a simple yet effective dual-Doppler uncertainty model. This package is based on the dual-Doppler uncertainty model developed by Nikola Vasiljevic and Michael Courtney. The model is based on the law of propagation of uncertainties. A full mathematical description of the model is enclosed in the YADDUM docstring. YADDUM is applicable for wind lidars and radars.


Getting Started

These instructions will get you a copy of YADDUM up and running on your local machine. Alternatively, you can run YADDUM without installing it using binder.


YADDUM is written in Python 3, therefore you will need to have Python 3 installed on your local machine. For the simplicity of YADDUM installation having Anaconda or Miniconda is advisable.


There are two ways how you can install YADDUM, either using conda or downloading the source code and manually installing the package.

In case you have Anaconda or Miniconda installed on your local machine you need to execute the following line in your (conda) terminal.

conda create -n UNCERTAINTY -c nikola_v yaddum

This command will create new conda environment UNCERTAINTY and install YADDUM package with all the dependencies. In this way your existing conda environments will be left intact.

Alternatively you can download the source code as a zip file, unzip it in a desired folder, in terminal navigate to the folder and run the following command:

python install

This command will download all the dependencies and install them first before installing YADDUM package. Beware that, unless you activate your preferred virtualenv, this command will overwrite any existing packaged. Therefore, if you are not confortable to setup virtualenv opt for the installation using conda.

Running the test

An example of Python script which uses YADDUM package is provided in test folder. Simply download the script and run the following command:


The end result should be a plot that looks like this:



YADDUM contains five classes (see image below) which are: Atmosphere, Measurements, Instruments, Lidars and Uncertainty. However, users only interacts with the class Uncertainty as this class inherits the properties of the remaining four classes.


The worflow with YADDUM is relatively simple and essentially consists of the following steps:

  1. Atmosphere parametrization using the method add_atmosphere(atmosphere_id, model, model_parameters)
  2. Localization and description of lidars using the method `add_lidar(instrument_id, position, category, **kwargs)``
  3. Localization of measurement points using the method add_measurements(measurements_id, category, **kwargs)
  4. Calculation of the measurement uncertainty using the method calculate_uncertainty(instrument_ids, measurements_id, atmosphere_id, model)

The methods add_atmosphere(), add_measurements() and add_lidar() create three Python dictionaries atmosphere, measurements and instruments, while calling the method calculate_uncertainty() produces two xarray datasets, namely wind_field and uncertainty.

Let's use the following example to show how to interact with YADDUM package. First, we import YADDUM package together with several additional packages and create a YADDUM object:

import yaddum as yaddum
import numpy as np  
import matplotlib.pyplot as plt
import xarray as xr

lidar_uncertainty = yaddum.Uncertainty()

Following the previous workflow we will parametrize atmosphere:


lidar_uncertainty.add_atmosphere('pl_1', 'power_law', model_pars)

The above commands will add power law model to the object (specifically dictionary lidar_uncertanty.atmosphere) together with the set of parameters. Currently YADDUM only supports this atmospheric model.

Next, we will add lidars to our object:

uncertainty_pars = {'u_estimation':0.1,

lidar_pos_1 = np.array([0,0,0])
lidar_pos_2 = np.array([1000,1000,0])

lidar_uncertainty.add_lidar('koshava', lidar_pos_1, **uncertainty_pars)
lidar_uncertainty.add_lidar('sterenn', lidar_pos_2, **uncertainty_pars)

With the code above we have added two lidars 'koshava' and 'sterenn' together with their positions and their intrinsic uncertainties which reflect their ability to:

  • Estimate radial velocity from the backscatter signal (u_estimation)
  • Resolve range (i.e., distance) from which the backscatter signal is coming from (u_range)
  • Point laser beams towards measurement points (u_azimuth and u_elevation)

Similarly to the atmospheric dictionary users have access to the lidar information by executing lidar_uncertainty.instruments command.

In the following step we will add measurement points to our object. We can either add an arbitrary array of points or create 2D horizontal mesh of points:

points = np.array([[500,-500,100]])
lidar_uncertainty.add_measurements('pts', 'points', positions = points)

lidar_uncertainty.add_measurements('mesh', 'horizontal_mesh', 
                                   resolution = 10, 
                                   mesh_center = np.array([0,0,100]), 
                                   extent = 5000)

In the first case we have added single measurement point with coordinates of (500, -500, 100) in Northing, Easting and Height, while in the second case we have provided position of the mesh center (0,0,100), set the mesh resolution (10) and mesh extent in Easting and Northing (5000). The unit for each of these values is meters. Both category of measurement points (i.e., single point and mesh points) now exist in our object and they are distinguishable by their ids ('pts' and 'mesh'). They are stored in the measurement dictionary:


The last step is to call the method calculate_uncertainty() and specify ids of lidars, measurement point, atmospheric and uncertainty model to be used for the calculations:

lidar_uncertainty.calculate_uncertainty(['koshava', 'sterenn'], 'mesh', 'pl_1', 

This last step will create the two xarray datasets contaning the results of the uncertainty calculation. xarray is a very powerful package and comes with a range of functionalities, thus it is worth spending time and learning how to operate with it. For example, xarray dataset can be viewed graphically and explore numerically:

lidar_uncertainty.uncertainty.azimuth_gain.sel(instrument_id = 'koshava').plot()


In case of confusion on what each method does and how to parametrize it simply check docstring methods. In our example to access the help for one of the methods we used you need to type:


Beware that YADDUM conforms to cf convention for naming and definition of atmospheric parameters. Check the Jupyter notebook providing an additional example on how to use YADDUM.

Built Using


How to cite

This package is persisted using Zenodo making it citable and preserved. Simply click on the banner bellow to find your preferred citation options (e.g., BibTex):



If you want to take an active part in the further development of YADDUM make a pull request or post an issue in this repository.


Michael Courtney for the support in developing dual-Doppler uncertainty model which is the basis of this package.

You can’t perform that action at this time.