Category: Uncategorized

Here we are! We’ll be there! At the PATA days, obviously!

We invited Dr. Anika Braun to lead a hands-on workshop on the usage of Artificial Intelligence (AI) in Earth Sciences, and this is what we plan for you!

The Workshopwill be held during the PATA Days 2022 in France.
Tuesday, September 27 2022; 18:30 – 19:30 CET
Wednesday, September 28 2022; 18:30 – 19:30 CET

If you want to partecipate to the workshop (under the PATA days subscription), please register here. https://forms.gle/QZvh2HDLxi4Vu3aS8

Only 20 sits available!

“Everything You Always Wanted to Know About Artificial Intelligence* (*But Were Afraid to Ask)) – a hands-on workshop on AI in Earthquake Geology”

Artificial Intelligence has long found its way into remote sensing and earth sciences, with applications e.g. in feature detection, spatial prediction, and hazard modelling. With the increasing availability of data, e.g. from satellites and drones, we also have an increasing need to efficiently exploit this data for valuable information. In the past few years, there has also been an incredible development in the availability of (open) tools and educational materials, enabling us to easily tap this data treasure. In order to generate plausible, reliable, and reproducible information for earth sciences however, it is important that we critically address points such as input data quality, model type, and results.

In this short course, you will learn some basics of Artificial Intelligence for earth science applications, so you know where to start if you want to implement this technique for your own project. In a simple hands-on example, we will see how you can critically explore your input data and prepare it for modelling. We will train some models, and then explore tools to evaluate the modelling outcomes. You will also be sensitized for the most important pit-falls and sources of uncertainty.

Participants need a computer (Windows or Mac) with an up-to-date operation system and some free disc space. No programming skills required, but some basic understanding of statistics would be good.

Day 1

Theory:
Machine learning concepts
General workflow
Data preparation and exploration

Hands-on:
(Data export from GIS (SAGA GIS?))
Data preparation (IBM SPSS Modeler)

Day 2

Theory:
Model types
Evaluation Tools

Hands-on:
Modelling
Model Evaluation
(Data re-import in GIS (SAGA, QGIS))

Workshop materials

Installation Guide and licenses

Sample dataset and project

“Cookbook” with most important application scenarios, data requirements, available tools, further reading or videos (we could ask the participants in advance what kind of machine learning problem they are interested in to adapt this to their needs)

In spring 2021 (14-16^th April, 2021) the kick-off meeting of the EDITH project was held online, gathering over 80 participants from all over the globe, ranging from PhD Students, Early Career Researchers and Senior Scientists.

For a full description of the kickoff meeting plan and program, see the linked pages. A sum up of the meeting can also be read on Quaternary Perspectives, that covered our initiative in its Issue n. 30.

Keynote talks were given by eight invited speakers from across a range of career stages, institutions and backgrounds. Their talks introduced classical and innovative approaches to earthquake geology across four main themes, geodesy, paleoseismology, archaeoseismology and tectonic geomorphology.  You can find the abstracts and recorded presentations on Youtube.

In the second part, we invited the participants to a collective brainstorming divided into thematic breakout rooms which focused on the following topics:

1. Epistemic uncertainty;
2. deciphering the seismic cycle;
3. short vs long term time scales;
4. geologic data for probabilistic seismic hazard assessment
5. groundbreaking cooperations.

The discussions (each 60 minutes) were stimulating, and we came out of these sessions with a spectrum of ideas, including suggestions for Hot Topics, that were identified to be challenges that can be tackled by the EDITH community due to their interdisciplinary approach.

Therefore, we plan to keep the interest focused on Hot Topics identified by our community through:

-Workshops for ECS to share methodological background on upcoming and challenging methods related to the Hot Topics

-Thematic working groups aimed at discussing most recent advances in earthquake geology and proposing future collaborations

– Paper parties, discussing groundbreaking articles in earthquake geology or potentially linked research areas offering  promising new interdisciplinary approaches.

The Key points from the breakout groups we use to define further action are:

Uncertainties

How to deal with increasing uncertainties that arise from the availability of high resolution topographic data and how to define the best signal to noise ratio, were the key topics discussed. Maybe Machine Learning (ML) methods can help us to find solutions. Therefore, we plan a training course on ML techniques to share available knowledge to tackle these challenges. Another issue that was touched was the methodology of trenching. Lots of uncertainties arise from this methodology since it is only point data which is often not reproducible.

Deciphering the seismic cycle

We identified a major knowledge gap in unraveling the seismic cycle, especially on slow moving intraplate faults. Can state of the art techniques like InSAR or LIDAR help us in identifying faults which are in the final stages of the seismic cycle? This is a topic that was identified as “Hot Topic” and we will dedicate a workshop on it to further advance collaboration and ideas. Furthermore, we discussed possibilities on how to identify creep in paleoseismological trenches. This would be a key for more complete hazard assessment.

Short vs. long term time scales

How to approach the discrepancies between geodetic and geological strain rates, was a major point of discussion. Furthermore, how distributed faulting in complex fault systems influences strain rate measurements is an important related topic to be discussed. Healing processes need to be taken into account as major factor in affecting fault behavior on different time scales. Therefore, collaboration with scientists form the rock mechanics community is desirable.

Geologic data for probabilistic seismic hazard assessment

Discussing strain maps from InSAR time series analysis we think they could be a major contribution, if we consider transients in strain. In fact, we noticed, that we are lacking experts from the field of PSHA and definitively need to initiate joint discussions with them.

Groundbreaking cooperations

Continuation of surface faults to depth is a widely needed input into our models, as well as the distribution of microseismicity. Therefore, probably not new or groundbreaking, but for sure extremely desirable is the collaboration with geophysicists. We need to promote our initiatives to our colleagues from this field. Application of machine learning techniques and Deep GIS is the hot topic, but we need to make sure that methodological developments are sound and integrating ground truthing. How this can be done will be a topic in the planned training course.

We think that the following actions are good choices to further perpetuate our initiative and stay in touch to discuss and learn about our exciting science:

• Community building: We plan to organize an informal meeting (a few hours long) during the next PATA days (France, 26-30 Sept 2022) and INQUA meeting (Rome, 16-20 July 2023). This will be our first opportunity to meet in person, to know each other, and discuss future networking.
• Online activities: We plan one or two scientific, online activities, for each incoming year:
  • Short training courses on specific techniques (e.g., Using machine learning in the Earth sciences).
  • Literature review on very recent advances of specific topics. This review will be followed by a round table discussion (e.g., slow-moving intraplate faults and how to face the related challenges).

Thank you again for the great Input and hope to see you soon! 😉

Notwithstanding the impact of the COVID outbreak, that changed all our lives, TERPRO’s EDITH Project (From Earthquake DeformatIon To SHA) is moving ahead and we can report some important NEWS.
We decided to go online with the kickoff meeting, that was unfortunately postponed by almost one year. The meeting will be held in the days of 14-16 April 2021 via a videocall platform (probably ZOOM).
We have planned the kickoff meeting as a discussion roundtable, with several keynote lectures, by invited speakers, and discussion rooms organized per field of expertise.

Our invited speakers, listed by field of expertise are:
GEODESY: Sabrina Metzger (Postdam Univerisity); Yariv Hamiel (Geological Survey of Israel)
PALEOSEISMOLOGY: James McCalpin (GeoHaz Consulting); Tamarah King (Oxford University)
ARCHEOSEISMOLOGY: Julian Lozos (CSU Northridge); Paolo Forlin (Durham University)
TECTONIC GEOMORPHOLOGY: Joanna Faure Walker (UCL Institute for Risk and Disaster Reduction); Ramon Arrowsmith (Arizona State University).

In the next weeks we’ll have some details on the title of the invited lectures. For the moment, please save the date!

We have received final confirmation from the INQUA commission that EDITH will be funded and we are excited to have the opportunity to bring together a group of like-minded scientists to tackle some of the big questions in earthquake science.
For the next three years, we will have annual meetings, as well as other activities, to bring together a range of scientists, across disciplines, career stages and countries.
For our first kick-off meeting, we want to take an interactive approach and provide plenty of opportunities for informal interactions and discussions. We are planning a kick-off meeting in the middle of April 2021 . Given the current coronavirus situation, we still do not know the location but it will be probably held in Europe; more details will come as the current situation evolves.

---

EDITH Kickoff meeting SCHEDULE (preliminary)

1st day – afternoon

Welcome and Introduction to EDITH

Presentation of the three main topics of EDITH (Into the seismic cycle, Shaping the Earth, Toward Seismic Hazard Assessement)

Short oral presentations of participating scientists (competence,  studies and envisaged input to EDITH)

2nd day

Roundtable discussion accompanied by keynote introductory talks.

Session proposal for the Barcamp meeting

3rd day – morning

Barcamp meeting

Debriefing and Final outlook

What topics are of interest to EDITH? We’ve made a mind map of the committee’s initial thoughts, but we are sure that there are many more links to be explored and established.

---