Developers Workshop

August 9, 2024

Volcanology Infrastructure for Computational

Tools and

Resources

Einat Lev

einatlev@ldeo.columbia.edu

Agenda – Day 1

Agenda – Day 2

Agenda – Day 3

What is VICTOR?

A cloud-based platform for volcanologists to find, use, and integrate numerical models and other computational tools

Goals

“VICTOR will help volcanologists at universities and observatories to find, use, compare, and link models and observations by providing a central access portal and standardized workflows based on a JupyterHub environment”

The VICTOR team

Columbia University

Tufts University

University of South Florida

Software engineer

Sam Krasnoff

Training and Education Coordinator (TBD)

VICTOR Advisory committee

U of Hong Kong

NASA/JPL

Idaho State University

INGV

U of Oregon

High-level design

Learn more: victor.ldeo.columbia.edu

Parallel efforts

Learn more: victor.ldeo.columbia.edu

Currently Available Models and Tools

1. IMEX-Lava Lava2d
2. pyFLOWGO
3. MrLavaLoba
4. MOLASSES
5. VolcFlow
6. VENUSS

New tools and models added all the time!

Developers Workshop next week!

Added-value Capability: Model Benchmarking

Learn more: victor.ldeo.columbia.edu

Set up common inputs, e.g.:

• Vent location
• DEM file
• Effusion duration / flux / volume

Visualize results

New Capability: remote-sensing workflows

1. Interactively define AOI

2. Granule Search using NASA OPERA

3. Interactive Time Series of Eruption Data

Upcoming capabilities – Inversion workflows

Future capabilities – Model Coupling + UQ

Conduit ascent model

Plume model

Magma chamber model

Ground deformation model

Ash model

VICTOR.ldeo.columbia.edu

Advisory Committee

Development and Management team

Workshops and Courses

victor.ldeo.columbia.edu

victor.ldeo.columbia.edu

Events

Learn more: victor.ldeo.columbia.edu

Cascades24 workshop, last week

Cities on Volcanoes 12, Guatemala

Documentation

victor.readthedocs.io

​

Learn more: victor.ldeo.columbia.edu

Education

Learn more: victor.ldeo.columbia.edu

@volcanocyber

Education

Learn more: victor.ldeo.columbia.edu

​

​

​

​

Teaching modules

​

​

​

​

​

​

​

​

​

​

​

​

​

​

Lava flows

Papers

Slides

Activities

Registration

• Sign-up link on the website
• Google registration form
• Currently requires a GitHub username for authentication

Learn more: victor.ldeo.columbia.edu

Scan to Sign up

A Technical View of the Lab

• Running Ubuntu 22.04
• Fully featured C/C++, Fortran compilers, as well as Javascript, Python, R interpreters
• A variety of python packages for machine learning, image recognition, and remote storage access (S3, Google Cloud, Azure)
• The majority of packages are installed through a conda environment, allowing for great flexibility
• Hosted through AWS EC2 instances managed by Kubernetes control panel

Connect to the VICTOR JupyterHub

1) From main site, or via victor.2i2c.cloud

Connect to the VICTOR JupyterHub

1) From main site, or via victor.2i2c.cloud

​

2) Choose a server type

Starting VICTOR

3) Wait for the server image to install (may take a few minutes)

The Launch Pad

A quick install script

Open a Terminal window, and type “victor setup”

​

Then choose a model to install

Pre-installed workflows

Learn more: victor.ldeo.columbia.edu

Pre-installed workflows

Learn more: victor.ldeo.columbia.edu

Virtual Desktop

To run applications with GUI, e.g. QGIS

Learn more: victor.ldeo.columbia.edu

Virtual Desktop

Learn more: victor.ldeo.columbia.edu

Agenda – Day 1