Cloud-scattering code
CloudScat is a Montecarlo code that predicts the signal observed by a satellite due to an optical source located inside or above a thundercloud. It considers both Rayleigh (molecular) scattering and Mie scattering with cloud droplets.
This code is used in the following article, published in Geoscientific Model Development:
The code is distributed as a julia package. To install it, after you have downloaded and installed julia, start julia and go into the package manager by pressing ]. Then type
add http://github.com/aluque/CloudScat.jlYou can exit the package manager by pressing backspace at the beginning of the prompt.
The samples folder contains example computations. It is recommended that you
start by copying these files into another folder and use them as templates for
your simulations. After you have installed the CloudScat package you can find
the location of the samples folder by
> using CloudScat
> CloudScat.SAMPLES_PATHThe file sample.jl contains a simple example. You can run it
with
> julia --color=yes sample.jlor, if you prefer to do everything from the julia prompt,
julia> include("sample.jl"); run()If the simulation finishes correctly it produces a file called sample.h5 that contains the output data from the simulation. To view this output you can use one of the scripts contained in the util folder of the repo. You can find the location of this folder in your installation by
> using CloudScat
> CloudScat.UTIL_PATHFor example to plot the simulated photometer data of observer one uses
> python plot_photo.py input_file.h5 --observer=1To see the image recorded by the same observer use
> python plot_image.py input_file.h5 --observer=1CloudScat can manage very complex cloud geometries, defined as combinations and
linear transformations of a set of elementary shapes. Full details are given
in the example file cloud_geometry.jl. If you want to define new geometrical
shapes beyond those provided by default, look at src/geometry.jl. A geometry
type must define methods to compute (quickly if possible) intersections between
the defined shape and a straight line and to test whether a point is inside the
figure. For an example on how to do this look at src/rings.jl, where the cloud
top has the shape of concentric rings.
If your machine contains several cores you can benefit from using several threads for your simulation. This can be accomplished by setting the environment variable
JULIA_NUM_THREADS to the number of threads. For example, for 4 threads invoke
julia as e.g.
> JULIA_NUM_THREADS=4 julia --color=yes sample.jl