Unpack the "ModelOutput" and "Python" archives to the same directory and follow the instructions below for using the Martian dust storm tracking algorithm.

###########################
### ModelOutput folder ###
###########################

This folder contains model output data of the Mars PCM6 with slightly different configurations of the "callphys.def" file (see explanations of the variables in "callphys.def" at https://lmdz-forge.lmd.jussieu.fr/mediawiki/Planets/index.php/Dust_Cycle_in_Mars_PCM6):

- "Cinj025" folder contains model runs with default callphys.def (coeff_injection = 0.25)
All other folders contain model runs with coeff_injection = 0.75:
- "Default" (no further changes)
- "InjectionTF_10-14" (injection of dust only from 10am to 2pm localtime; ti_injection = 10. , tf_injection = 14.)
- "NoRds" (no rocket dust storm scheme; rdstorm = .false., topflows = .false., naerkind = 2)

All output data was reduced to 2pm localtime, using post-processing tools coming along with the model.

#####################
### Python folder ###
#####################

This folder contains the Python scripts for the automatic dust storm tracking algorithm. This is how to use them and how the folder structure works:

1) Edit the "input_vars.py" file with the path information of your folder structure and the parameters you wish to use (the variables dist_offset1, dist_offset and fdist are only relevant to step 3). You do not need to edit the other files.
2) Run the "storm_tracker.py" script (no need for editing). This file converts the CDOD maps (model input and output in the "ModelOutput" folder) data to a Mollweide projection, reads the information for initialization of each dust storm sequence from the "Seq_lsts" folder, calulates the sequence-specific CDOD threshold and does the image processing (contouring of dust storms). Results are stored in the "ProcessedData/Automatic_attribution/" folder (contour data, .csv files). The "Raw" folder contains the raw CDOD maps and the "Contours" folder the CDOD maps with contoured storms.
3) Run "seq_attrib.py" (no need for editing) for the attribution, of each individual storm contoured in step 2, to a sequence. The results (attribution in "*_seq.csv" files, summarized dust storm track information in "*_cc.csv" files) are stored in the "ProcessedData/Automatic_attribution/" folder along with the results from step 2. The "Seq_attrib" folder contains the CDOD maps with contours and sequence attribution information (Yes/No marking of each contour along with storm ID from the corresponding _seq.csv file).
4) You may use the data in the "*_cc.csv" files from steps 1 to 3 in order to plot the trajectories with your own plotting routine. These trajectories can be compared with the "*_manual_cc.csv" files from the manual attribution method (stored in "ProcessedData/Manual_attribution/"). 
5) If you want to create you own manual attribution, edit the "Seq" column in the "*_seq.csv" files in "ProcessedData/Manual_attribution/" (0 = No, 1 = Yes) and then run the "postproc_manual.py" script. This script will overwrite the existing "*_manual_cc.csv" files.

After step 2, you may also edit the "input_vars_testab.py" file and then run "seq_attrib_testab.py" to scan the parameter space for a and b (the parameters of the automatic attribution algorithm). The results are stored in the "ProcessedData_testab/" folder. 

Other files in this folder:
- "sol2ls.txt" is used by the .py scripts and containing information for the conversion of Sol time vector to Ls time vector (and vice-versa).
- "geography_fn.py" is imported by the .py scripts and contains functions for Mollweide projection, circular mean calculation of longitudes, and haversine calculations.