Interview with Arnau Lagresa, creator of the database (BDD) for the GarMultiRisk project

The database (BDD) of the GarMultiRisk project compiles all available information about various natural hazard events that have affected La Garrotxa region over the past 123 years.

To learn more about this database, we spoke with Arnau Lagresa, a key contributor to the GarMultiRisk project and the primary creator of the BDD. Arnau holds a degree in History and Geography from the University of Girona and is an expert in Geographic Information Systems (GIS).

- Arnau, first of all, what is your connection with the Centre for Territorial Sustainability (CST), the coordinating entity of the GarMultiRisk project?

I started collaborating with CST during my first GIS course in my Geography degree. Since then, I’ve completed various assignments for the organization. I frequently use tools like QGIS, a software that allows for creating, analyzing, and visualizing maps and working with geographic and spatial data. When the GarMultiRisk project began in 2023, CST reached out to me again to see if I wanted to participate, and I agreed. In parallel, I also work as an environmental technician for the Alt Empordà Institution for the Defense and Study of Nature (IAEDEN).

- As part of the GarMultiRisk project, your main task has been collecting data for the BDD. How new was this role for you?

This was my first time creating a database, and I found the process particularly novel when it came to understanding the geological dynamics of the various natural hazards we studied. I had no prior knowledge of these dynamics, so I had to undertake significant research. This was a motivating challenge and gave me a broader understanding of natural processes in La Garrotxa. Additionally, the project allowed me to consolidate my knowledge: I applied GIS skills using QGIS and employed historical research methodologies to gather data from old records, such as local newspapers.

- Before diving deeper, what is the purpose of this database?

The goal is to train an algorithm capable of identifying interrelations between different natural hazards. This allows us to detect patterns and, more importantly, determine whether one natural hazard or process can trigger or increase the likelihood of another. This concept, called "multi-hazard," is the innovative focus of our project.

We use a probabilistic approach: we collect data from past events (via the BDD) to create future forecasts. For this project, the analysis period was defined based on the availability of historical data for La Garrotxa, which spans about 123 years. Once all the data is processed, we will produce final outputs such as interactive viewers, hazard maps, and future scenarios, which will be integrated into a Natural Risk Reduction and Management Plan for La Garrotxa. This plan will be delivered to policymakers and emergency personnel in the region, providing them with a comprehensive understanding of multi-hazard risks in La Garrotxa to support effective long-term planning.

- La Garrotxa also has volcanoes, but there haven’t been eruptions in the past 123 years. How do you handle such natural hazards that fall outside the studied timeframe?

As you said, there hasn’t been an eruption in the past 123 years, so this data isn’t included in the BDD. The most recent eruption is estimated to have occurred about 8,400 years ago (Revelles et al., 2023). However, since volcanism is a potential future hazard in La Garrotxa, it will be considered in the project’s final outputs, such as scientific articles and reports.

- Can you describe what the BDD looks like today to help us better visualize it?

The BDD is in the form of an Excel file, with each tab representing a different type of natural hazard. We’ve compiled information on six natural hazards: earthquakes, floods, land subsidence/sinkholes, landslides, rockfalls, and wildfires. Each hazard has specific data fields. Some are common across all hazards, such as location and date, while others are tailored to the specific dynamics of each hazard.

- Can you give us a concrete example?

Of course! Take floods, for instance. If you open the flood tab in the Excel file, you’ll see rows and columns. Each row corresponds to a flood event that occurred at a specific time and location within the 123-year study period. The columns represent fields or variables important for characterizing floods, such as start and end dates, location, affected area, impacted municipalities, travel distance, volume, topography, watershed, and more. When new data on a flood event becomes available, we add a row and fill in the relevant variables based on the information at hand.

- Gathering all this data must have been challenging. What were the main difficulties?

Even now, some columns or variables are missing data. One of our key goals was to design the ideal multi-hazard database, including all necessary fields to make each hazard record as complete as possible. However, we often encountered incomplete data when collecting information in the field. This means that, later on, when training the algorithm, we’ll need to make assumptions or define ranges for variables with less data.

Additionally, as a pilot project, GarMultiRisk offers an excellent opportunity to experiment and learn. The knowledge gained here will help us refine our methodology and techniques for future research.

- Was it harder to find information on certain hazards? Or was there anything surprising about the data?

Wildfires were challenging to document because there aren’t many in La Garrotxa, and when they occur, they tend to affect small areas or have anthropogenic rather than natural causes. Also, historical newspapers contained very little information about them. On the other hand, I was surprised by the number of landslides we documented, which is the hazard for which we have the most data.

For earthquakes and floods, there were already well-established and comprehensive databases available.

- You’ve mentioned geological dynamics and algorithm development. How has collaboration worked within your team, and what kinds of expertise does it include?

Our team is multidisciplinary, which is one of the aspects I find most motivating about the project. Scientific knowledge is provided primarily by Dr. Joan Martí Molist and Dr. Marta López Saavedra from the Natural Risks Assessment and Management Service (NRAMS) at IDAEA-CSIC. The IT component is handled by Marc Martínez Sepúlveda, also from NRAMS. Local scientific knowledge is supported by Dr. Llorenç Planagumà, Mireia Jiménez (CST members), and Alba Ocaña, an environmental technician with the Garrotxa Environmental and Public Health Consortium, a project collaborator.

- What steps have you taken so far in creating the BDD?

First, Marta and I identified information sources: where to search for data. We created an Excel sheet for each natural hazard, listing where to find the information—web links, book references, articles, historical newspapers, and so on.

We noticed two types of data: inventoried and non-inventoried. Inventoried data, such as official digital records, could be downloaded with a click and integrated almost immediately. Non-inventoried data took more time, requiring searches through scientific articles, popular science books, and historical press archives.

In January 2024, we held citizen science workshops to gather oral testimonies from local residents about natural hazard experiences. Participants included weather enthusiasts and Forest Defense Group workers who shared valuable information on fire and flood dynamics in La Garrotxa.

Throughout this process, I regularly updated the BDD and met with the team to discuss questions and make decisions based on the scientific and technical insights each member brought.

- What are the next steps?

To start training the algorithm, Marc Martínez (software developer) needs all our data in a standardized format that machines can read easily. Meanwhile, the scientific team must develop a robust conceptual model using an “event tree” methodology. This ensures the algorithm explores hazard interrelations based on the physical and geological laws governing these processes.

The project concludes in May 2025, when we’ll deliver a range of products, including static maps, online interactive viewers, and reports, all integrated into a management plan.

- How do you feel about the work done so far?

I believe we now have the most comprehensive natural hazard database for La Garrotxa, which is a significant achievement. The next step is to merge geology and computer science to make the database functional.

- If you could start over, what would you do differently?

I would have defined the necessary variables for each natural hazard more precisely from the beginning. Having to revisit information sources repeatedly to gather additional data has been time-consuming.

- Lastly, what are your overall thoughts on the project and your participation?

I’m highly motivated by this project. While geology and natural hazards aren’t my main expertise, I’m glad to contribute so concretely. This project has also opened a new professional chapter for me, as I’ll be joining NRAMS in January 2025. I’m excited to build on what we’ve achieved and to keep moving forward as a team.

- That’s great, Arnau! Best of luck in this new chapter, and Happy New Year 2025!

Thank you, same to you!

References:

Revelles, J., Martí Molist, J., Burjachs, F., et al. (2023). Socio-ecological impact of monogenetic volcanism in the La Garrotxa Volcanic Field (NE Iberia). Scientific Reports, 13, Article 8168. https://doi.org/10.1038/s41598-023-35072-0.

The GarMultiRisk Project is supported by the Fundación Biodiversidad of the Ministry for the Ecological Transition and Demographic Challenge, through the Call for Grants for the implementation of projects contributing to the National Climate Change Adaptation Plan (2021-2030).

Financial support by:

With the collaboration of: