###### tags: `oggm`

# OGGM Online tutorial notes

16.04.2020, 14H
[Fabien Maussion](https://fabienmaussion.info)

**Agenda**

[TOC]

## Introduction

Large scale glacier modelling: what is that?

If you missed it, IARPC recorded a talk from Regine Hock and myself. You can watch it on [youtube](https://www.youtube.com/watch?v=ttJMxcwXUjw).

Discussion points for today:
- where is OGGM today in the landscape of (global) glacier models?
- Originated from [Marzeion et al 2010](https://www.the-cryosphere.net/6/1295/2012/tc-6-1295-2012.html), but is completely independant. Biggest legacy is the MB model calibration and the fact that Ben is a strong driver of OGGM development.
- very similar models: [GloGEM](https://www.frontiersin.org/articles/10.3389/feart.2015.00054/full) extended by [GloGEM-Flow](https://www.the-cryosphere.net/13/1125/2019/) and [PyGEM](https://www.frontiersin.org/articles/10.3389/feart.2019.00331).
- other global models: mostly scaling relationships, more or less obscure
- regional models (very similar in construction ut often tuned for a region or not able to handle as much data)
- see GlacierMip and latest paper by Marzeion and colleagues for an overview (out soon)
- OGGM (and the models listed above) are of a **very different kind** than the traditionnal ice sheet and ice flow models (e.g. ElmerICE, PISM, etc.). In OGGM, a very large part of the codebase is data and preprocessing workflow.
- in comparison to competitors:
- OGGM is far more advanced in terms of model structure and documentation
- most complete ice dynamics and geometry handling
- simpler MB model
- less publications (by far) than the main competitor GloGEM (but also less "magical", for good and for bad)

## Tools

We will use Jupyter notebooks. You can learn about them [here](https://edu.oggm.org/en/latest/notebooks_howto.html).

No need to install OGGM on your laptop! We are going to follow the tutorials online, by using the [OGGM-Edu](https://edu.oggm.org) platform. You can follow along, or just watch me going through them with you.

Optional: if you have OGGM installed and know about these things, you can also download the notebooks from the [repository](https://github.com/OGGM/oggm-edu-notebooks).

## Model workflow and model physics

Tutorial: **getting_started.ipynb**

We will go through the tutorial together, and jump between the tutorials and the documentation for explanations.

If time permits:
- inversion.ipynb
- massbalance-calibration.ipynb
- working_with_rgi.ipynb

## Installation and HPC set-up

- typical problems
- windows is almost never working properly
- on HPC:
- consider using singularity
- OGGM Docker: https://hub.docker.com/u/oggm
- pre-download your data

## Typical pitfalls and difficulties

- the problem of bed topography:
- currently no out of the box solution for all users (this is work in progress)
- MB model and bed topography have to be consistent to avoid spin-up issues. Discuss example with PyGEM.
- glacier directories and overall model complexity
- experience shows that people with limited python / computer knowledge struggle with the model
- the persistence on disk workflow is increasing complexity of certain workflows like sensitivity experiments. There are mechanisms for that, but it requires some thinking
- the pitfall of "thinking small" (guilty as well):
- large scale models will never be able to model each glacier individually with accuracy. The flowline and MB models will fail miserably in many cases. The strength of "pioneers" like Ben and Matthias Huss is their ability to abstract from that and focus on the big picture
- paleo simulations:
- how far can we simulate past glacier evolution?

**Important: keep in touch with us**

## Publications in preparation

At least the ones I know of:
- Ben Pelto: ice thickness in Alaska
- Matthias Dusch: calibration and simulation of glacier lengths for ~40 glaciers in the Alps 1850-present-future
- Julia Eis: global ice volue reconstruction 1900-today (challenging)
- Bea Recinos: inversion with calving in Greenland (uncertain if will be submitted)
- Fabien Maussion: global ice volume calibration (see pitfalls above)
- Li Fei: ice thickness in Tian Shan
- Nicolas Champollion: global projections with CMIP6

## Future model developments and changes

- **Important**:
- revamp of climate data handling for ERA5 and ERA5-land (horizon: weeks)
- ice thickness inversion calibration tools (horizon: months)
- OGGM Shop:
- GlaThiDa
- CMIP5 and CMIP6
- ITS LIVE
- Geodetic MB from where available
- Calving: colaboration with Lizz Ultee
- Mass-balance: PyGEM - OGGM compatibility

## "Adding value": how to publish with OGGM

Open discussion

## Keep in touch

- follow the changes on [Github](https://github.com/OGGM/oggm)
- join us on Slack