Reconstructing past glacier mass change is of interest for different applications, e.g. for quantifying their contribution to sea-level change. One approach is to use a glacier model, forced by reconstructions of climate, to estimate past glacier states. However, glaciers respond to climate variability and change with time lags between a few years and many centuries, and the backwards reconstruction is impeded by the non-linear interaction between glacier geometry and mass balance.
This repository host a new method, based on OGGM that estimates past glacier states. The method is published in "The Cryosphere".
Our method consists of 3 steps and the workflow is shown in the following figure:
- generation of glacier states,
- identification of glacier candidates, and
- their evaluation based on the misfit between the modelled and the observed surface elevation in the year of the observation.
The current version was tested on glaciers located in the Alps and we reconstruct their state in 1850 by using synthetic experiments. Thus the reconstructed glacier states did not represent the reality and need to be handled with caution. A further developement that will also allow the application to real-world cases is under developement and a new manuscript will the submitted soon.
In most cases, the resulting reconstruction is non-unique, as multiple initial states converge towards the observed state in the year of observation.
You can install this custom package with:
pip install git+https://github.com/OGGM/reconstruction.git
However, what you'd probably like to do is to fork this repository and use it as a template for your own project. You can install it locally with:
pip install -e .
How to use this method is shown in this example. The usage of our method takes some time (hundreds of OGGM simulations are required to reconstruct the state of ONE year), and we recommend to use a super-computer environment for the reconstruction of multiple glaciers.
- View the source code on GitHub.
- Report bugs or share your ideas on the issue tracker.
- Improve the model by submitting a pull request.
- Or you can always send us an e-mail the good old way.