Newsletter 6 - April 2021

At the beginning of the URBASIS project, we did not imagine (certainly like many other projects of this type) that we would have to re-invent the framework and organization of the project because of Covid-19. One of the objectives mentioned by the MSCA program of the European Union is to promote networking between partners in order to place the ESRs in the best conditions to accomplish their scientific work. Thus, they are exposed to various scientific environments and can develop their skills and expertise thanks to the multiple events, like those we had programmed to promote interactions.

And the Covid-19 went through this...

Of course, we have been able to move the training activities online, to organize monthly meetings by WP to exchange between partners on the basis of the work presented by the ESRs, we have organized the general assemblies, the ESR's advisory boards by zoom, allowing the students to progress in their doctoral work. The progress of the scientific work is not questioned, the dissemination activities, the scientific publications and the writing of deliverables are confirming (according to the scientific committee of the project) the excellent on-time scientific perspectives of the project.

But still!

Nothing beats face-to-face meetings, during which the question/answer sessions are  smoother, wide-ranging discussion during coffee and lunch breaks may be profound, spontaneous, impossible to predict, without the imperious respect of the timing, and the social evenings animated by exciting scientific discussions allow to break the ice between senior researchers and ESRs, the ESRs themselves being able to meet more simply...

Let's be optimistic!

The URBASIS project is about halfway through and we can imagine that this situation will not last forever. Let's hope that the second part of the ESRs' projects will be full of science and face-to-face events that will allow them to finalize their project in the best conditions.  Yes, let's be optimistic!
 

  Edito by Philippe Guéguen, coordinator of the ITN URBASIS project.

Incoming scientific deliverables

Deliverable D1.3 - Regionally adjusted stochastic ground-motion models, associated variabilities,  correlations and epistemic uncertainties
Both empirical models and stochastic models represent recent and promising developments in the field of engineering seismology. Empirical models are developed from a cluster of ground motion recordings from several earthquakes recorded at multiple stations in different regions to ensure that the developed models are robust and reliable for the prediction. One of the main alternative ways to derive empirical models from sparse data regions is to utilise stochastic modelling to simulate ground motion data. However, for Probabilistic Seismic Hazard Analysis (PSHA), there is great importance for a regionally adjusted ground motion model, i.e. to update the parameters of the model rather than developing a new model with sparse data in different regions. Here, we have tried to develop an algorithm that can be utilised for empirical and stochastic models to adjust them regionally for more precise predictions using the European Engineering Strong Motion (ESM) dataset. This procedure is implemented on one of the most commonly used European empirical models -  Bindi et al. (2014, B014)  and reduced the sigma and area metric (AM) of B014 by 15% and 30%, respectively, after the calibration process. In addition, we have tried to analyse the parameters of stochastic modelling using SMSIM and execute a similar algorithm to develop a regionally adjusted stochastic model.
 
The change of AM before and after the calibration process of Bindi et al. (2014, B014) to the data with RegionID SZ49. The orange curve represents the observed data and the blue curve is the prediction. a) before calibration b) after calibration
Deliverable D4.5 - Use of sparse information from earthquakes in urban areas
While we are marching into a Big Data world (for instance the exponential increase in seismic recordings), it seems inevitable and even more necessary to deal with bad data since the source data, in practice, are often limited, sparse, and even corrupted. Automatic procedures that can identify and even handle those bad data are of great interests to both earthquake engineers and seismologists. Deep Learning methods, known to demonstrate great performance in an end-to-end fashion, can provide prospects for us to explore the data and construct robust models. Specifically, attentions are paid to reflect the degree of indeterminacy from the sparse information, and to account for the uncertainties of the deep learning models. 
Deliverable D4.3 - Introducing time-dependancy in the probabilistic seismic hazard assessment
The seismic hazard level at a site or a region can be quantified following a deterministic or probabilistic approach, being the latter the most common one in several applications in both sides, the academia and the industry. However, this methodology is based on the fact that all events are independent of each other, and it is well recognized that earthquake activity varies with time, especially for aftershock sequences and induced seismicity applications where the spatial and temporal distribution of events depends on the characteristics of the triggering mainshock or injection rate. This deliverable addresses the problem for both types of seismicity focusing on a case study of a induced seismicity in UK where we explore several methods and techniques such as the Epidemic-Type Aftershock Sequence (ETAS) Model and the Seismogenic Index Model to take into account these seismicity variations in time and include them into the seismic hazard assessment.

Our URBASIS item

In January of this year, the URBASIS management team together with the ESRs have decided to create an promotional item in effigy of the URBASIS project. A contest was put in place to collect as many ideas as possible. The item had to respect certain rules such as being useful, eco-friendly and unisexe. 
After a brainstorming session, each participant proposed their ideas and imageries and then a vote was made. So today we are happy to present our new URBASIS item!

How about our ESRs?

Subash Ghimire - ESR4.1
“This year, I had many opportunities to know more about myself, to develop patience, resilience and many professional and life skills.”

"Wow!! It has already been 18 months, time flies! During this period, I have been exploring the application of machine learning techniques in earthquake building damage predictions, testing a machine learning-based building damage prediction method by considering a few key features of buildings, macroseismic intensity, and the corresponding building damage information observed due to earthquakes. I am comparing the performance of different machine learning techniques and their sensitivity with earthquake magnitude, region, damage grade definition. So far, I am excited to share that the machine learning technique provides a reliable estimate of the spatially distributed building damage!
This year, I had many opportunities to know more about myself, to develop patience, resilience and many professional and life skills. URBASIS is giving me a great platform to learn and grow day by day, thanks to my ESRs friends, supervisors, and all helping hands involved in this project. So far, I am ecstatic about my PhD and also excited to enjoy the wonderful things that will come after this global pandemic."

Chiara Amendola - ESR3.5

“I would better realize one day, that the opportunity I’m having now, was the best I could have ever experienced.”


"In 1978 Lucio Dalla sang: "Dear friend, since you left, there’s big news, the old year is over now, but something is still wrong here". It's not 1978 anymore and neither 2020 but still there's something wrong. We are facing a global education crisis, however, I found myself lucky being able to continue my research from home. Miles of distance fade as online platforms consolidate. The luckiest of us realize how true the "one face one race" expression is and 15 guys who one year ago were complete strangers, start to call themself friends, because as friends do, they support each other for good and for bad. Awards, conference and journal papers should measure our growth, our work's developments during this year, I would rather say that such aspects are just a series of lucky consequences of being part of an amazing project such as URBASIS. I would better realize one day, that the opportunity I’m having now, was the best I could have ever experienced. Plenty of researchers are believing in my work and are giving help and precious advice, and I can only keep going on and doing my best in my projects and my expectation. Human bonds which are now surfacing are the true value of the URBASIS project."

Srihari Sangaraju - ESR3.4 

“While smart working made the communication faster and kept me away from other distractions [...], some perks of student life were taken away.”

"In the first year of my PhD, and the life in the pandemic, I have faced a bit of everything. It was a whole new personal experience as an ex-pat in a new country, with restrictions changing every couple of weeks. While smart working made the communication faster and kept me away from other distractions especially near the deadlines, but some perks of student life were taken away.  It was also strange attending virtual conferences, while they save time and expenses, making new connections with researchers is quite difficult.

We always had an option in the past, to postpone the physical activities like workshops and secondments, but with the emergency continuing to yet another year, we may have to think about organizing them remotely, which I feel is not the same experience for ESR's. However, we had more internal meetings, where I got acquainted with the spectrum of contemporary problems in urban seismology on which my colleagues are working on. We have also picked out the areas where we could potentially collaborate. With this positive note, I am looking forward to having a safe and productive year ahead."

Our futur events

  • WP2 #2 meeting – Induced Seismicity – Apr. 2021
  • WP3 #2 meting – Urban seismology, engineering seismology and earthquake engineering – May 2021
  • WP4 #2 meeting – Vulnerability and risk assessment: improving the interface between engineering seismology, earthquake engineering and users – Jun. 2021
Come and join us !
More info