MANTRA Postprocessing scripts

This repository contains scripts to turn raw results from the "MountAiN glacier Transient snowline Retrieval Algorithm" into analysis-ready data.

In case you have a Slurm-based HPC cluster at hand, you may use the Bash shell files to run scripts in parallel.

Requirements

Data

First, you will need to process some MANTRA TSLA data. An small example file is ./data/MANTRA/.

Many of the scripts require a CSV file of Randolph Glacier Inventory (RGI) data to work. An example for High Mountain Asia is provided in ./data/RGI. For other regions, you may download the according shapefile, open it in your preferred GIS environment, and export the attribute table to CSV.

Software

• Python 3 (a current version of Anaconda is recommended).
• Pandas and Numpy (will be included in Anaconda).

Filter and clean raw GEE TSLA result data

Step 1: Merge GEE files

Append all CSV files but make sure the header line is included only once:

head -n 1 TSL-results-0001.csv > _TARGET_PATH_/TSL-results-merged.csv && tail -n+2 -q *.csv >> _TARGET_PATH_/TSL-results-merged.csv

Step 2: Basic filtering and cleaning

Make sure input_file and output_path are set correctly in tidy-raw-gee-tsl-csvs.py. Then run it ...

python tidy-raw-gee-tsl-csvs.py

Step 3: Remove winter maxima

The script to remove winter maxima is designed to be run in parallel on a HPC cluster using the workload manager Slurm. For convenience, you may want assign the path where the HDF5 file is located to a bash variable, here named H5_PATH.

cd _PATH_TO_HDF5_FILE_
H5_PATH=$( pwd )

~/launch_rwm.sh $H5_PATH/TSLA-HMA-2020-07-filtered.h5 $H5_PATH/glaciers-HMA-full.list 500

Step 4: Merge HDF5 files

Step 3 will result in a series of HDF5 files that need to be merged, subsequently. In the directory containing the files, first create a file containing a list of all file names that should be merged, using e.g.

ls *-noWinterMax.h5 > h5-files.list

And merge the HDF5 files by running:

python merge-multiple-hdfs.py <input_dir> <output_file> <file_list>

Optional scripts

Append CSVs to existing HDF5

Edit and run append_GEE_obs.py. This is supposed to be used when you process additional MANTRA data and what to append it to an existing data set. For instance, you may want to add a new year to an existing set of TSLA timeseries.

Append a preprocessed HDF5 to existing HDF5 file

Edit and run append_preproc_TSLA_data.py.

Drop duplicates

Edit and run drop_duplicates.py. This is convenience script to ensure, well, that there are no duplicates in your dataset.

Grid preprocessing

Edit and run grid_preprocessing.py. This will support to create a gridded dataset of TSLA anomalies, e.g. for EOF analyses.

Obtain TSLA maxima per year and glacier

python maxima-per-glacier-and-year.py <input_file> <glacier_list_file> <lower_limit> <upper_limit> <output_path>

Split big file to individual glaciers

Helper tool to convert a full TSL file into individual files per glacier.

python plot-glacier-timeseries.py <tsl_file> <output_dir> <output_type>

With tsl_file being a TSL result file in HDF format, output_dir a valid directory to which the output will be written, and output_type either csv or hdf5.

Citing MANTRA

If you publish work based on MANTRA, please cite as:

David Loibl (2022): MountAiN glacier Transient snowline Retrieval Algorithm (MANTRA) v0.8.2, doi: 10.5281/zenodo.7133644

Acknoledgements

MANTRA was developed within the research project "TopoClimatic Forcing and non-linear dynamics in the climate change adaption of glaciers in High Asia" (TopoClif). TopoCliF and David Loibl's work within the project were funded by DFG under the ID LO 2285/1-1.

The development of filters identification of adequate thresholds was supported by Inge Grünberg. Inge's contribution to this project was supported by Geo.X, the research network for Geosciences in Berlin and Potsdam.