Newsletter 5 - November 2020

The second year of the URBASIS project came to an end under special conditions due to the COVID crisis. The ambition of the ITN URBASIS project is to provide the recruited ESRs with the opportunity to exchange and complete their skills in their research field. The program of training activities and the exchanges planned directly with the researchers involved in the network during the workshops and scientific meetings were designed to provide them with the best conditions to develop their personal development plan. 
The situation has modified our initial plans. But cheer up! We are convinced that the upcoming program will allow them to develop their project in the best possible conditions. As everywhere in Europe, we have organized ourselves and are moving forward in order to allow them to be exposed to an (online) scientific environment being as profitable and adapted to them as possible. The scientific activity is progressing, and the deliverables of the project are numerous and of high quality, as detailed in this newsletter. 
In this situation, such impredictible, our ability to adapt will have to be mobilized and I am certain that, in the end, the projects of each ESR will contribute to advancing our knowledge of seismic risk in urban areas.

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

The URBASIS project scientific progression

Deliverable D1.1 - Epistemic Uncertainties of large earthquakes stress-drop models
 
In order to develop ground motion models for Europe and evaluate the regional variability of source scaling, a waveform database has been constructed. The software stream2segment (Zaccarelli et al., 2018) has been used to extract waveform data and station metadata from EIDA (European Integrated Data Archive). The ISC (International Seismological Centre) event-catalog has been considered to guide the segment extraction for earthquakes occurred between 1991 and 2020 in central Europe, with magnitude above 2.5 and depth shallower than 60 km. The data set contains about half million downloaded segments. We first run the European earthquakes as an example and obtain the scaling of stress drops from the decomposed source function for the deliverable. The preliminary process and result shows that the regionalization is important to get more realistic effect which is our next step.                                                                          
Fig1. Our strategy of developing ground motion models for Europe and evaluate the regional variability of source scaling.
Deliverable D1.5 - Spatial variability of earthquake ground motion: insights from 3D numerical simulations 
 
One of the key issues in seismic risk assessment of large urban areas is the modeling of the spatial correlation of ground motion intensity measures. The majority of spatial correlation models have been developed using large databases including recordings from dense networks in different regions worldwide. However, recent studies have highlighted that spatial correlation is strongly region-dependent so that analyses derived on large datasets may not be appropriate. On the other hand, 3D physics-based numerical simulations (PBS) of seismic ground motion allow for detailed, region-specific studies on the spatial correlation of ground motion and on its dependence on the seismic source, directivity in near-fault conditions, path and site effects. This deliverable shows the main results of a systematic analysis of the spatial correlation (semi-variograms) in spectral accelerations from broadband physics-based simulated ground motions (SPEED code, coupled with an Artificial Neural Network technique for the high frequency part) for different regions (Po Plain, L’Aquila, Marsica, Sulmona, Norcia, in Italy; Istanbul, in Turkey; Thessaloniki, in Greece) and severe earthquake scenarios (M=6-7+) in near-source conditions. The dependence on physical factors, such as ground motion directionality, magnitude, source directivity and source-to-site path, is discussed.                                                                                                                                                                                            
Deliverable D2.1 - Model relating physical and system properties to induced seismicity
 
One of the key issues in seismic risk assessment of large urban areas is the modeling of the spatial correlation of ground motion intensity measures. The majority of spatial correlation models have been developed using large databases including recordings from dense networks in different regions worldwide. However, recent studies have highlighted that spatial correlation is strongly region-dependent so that analyses derived on large datasets may not be appropriate.
On the other hand, 3D physics-based numerical simulations (PBS) of seismic ground motion allow for detailed, region-specific studies on the spatial correlation of ground motion and on its dependence on the seismic source, directivity in near-fault conditions, path and site effects. This deliverable shows the main results of a systematic analysis of the spatial correlation (semi-variograms) in spectral accelerations from broadband physics-based simulated ground motions (SPEED code, coupled with an Artificial Neural Network technique for the high frequency part) for different regions (Po Plain, L’Aquila, Marsica, Sulmona, Norcia, in Italy; Istanbul, in Turkey; Thessaloniki, in Greece) and severe earthquake scenarios (M=6-7+) in near-source conditions. The dependence on physical factors, such as ground motion directionality, magnitude, source directivity and source-to-site path, is discussed.                                                                                                                          
Deliverable D2.4 - Toward New-Field Ground-Motion Prediction Equations for Induced Seismicity
 

Induced seismicity is currently drawing public attention as a potentially significant hazard. Several studies have been conducted to develop ground-motion prediction equations (GMPEs) for induced seismicity, however, many of them still rely on the assumption that induced events have similar source and attenuation parameters to those of natural earthquakes. We use Preston New Road (PNR) dataset recorded between 2018-2019 with local magnitudes ML < 3 at distances less than 40 km to develop GMPEs that are tuned to the key magnitude-distance range for induced seismicity applications. A preliminary observations focused on: (1) study of attenuation parameters using spectral fitting methods utilizing three different part of seismograms (S-wave, coda wave, and S-coda wave windows) and coda envelope decay methods; (2) estimations of residual site-specific exponential decay ; and (3) observation of site condition by calculating Vs30 using several bedrock depth assumptions and fundamental frequency obtained from horizontal-to-vertical spectral ratio (HVSR) discussed in this deliverable. In addition, an updated ML- MW  relationship model for PNR dataset is presented. These findings can subsequently be implemented for developing suitable GMPEs for induced earthquakes. Through a better understanding of ground motions and their controlling factors, future studies will hopefully can draw more robust conclusions on the behaviour of ground motions from these events and reduce the uncertainty associated with GMPEs for induced events.                                                                                                                                                            More information

Deliverable D3.6 - State-of-the-art fragility curves of typical RC and masonry structures considering SFSI and aging effects
 
The Deliverable D3.6 presents a comprehensive review of the up-to-date literature regarding SFSI and aging effects on vulnerability computation. The main finding is that the studies carried so far, prove that the conventional way of calculating building fragility considering fixed-base and uncorroded structures may lead to a potential underestimation of the seismic risk
Soil-foundation-structure interaction and local site effects are generally shown to be more pronounced in the case of soft soil formations and high-rise structures, causing considerable modification to the free-field motion, as well as to the dynamic response of the structure. In this light neglecting SFSI and especially site effects may lead to inaccurate fragility and loss estimates, which constitute fundamental components in the risk assessment. Moreover, aging effects on structures might increase their vulnerability and contribute to significant loss of their capacity via a slow, progressive and irreversible process caused by the deterioration of the material properties.
Even though the results of such studies provided the scientific community with valuable knowledge at site-specific vulnerability assessment, the reliability of risk analysis at urban scale is assessed with certain limitations. Further research is necessary for the development of generalized fragility functions applicable to different reinforced concrete and masonry buildings, which would take into account SFSI and site effects for a great variety of soil conditions.
The encouragement in the adoption of reliable SFSI and time-variant models will promote a more accurate assessment of the seismic safety of existing buildings. This aspect is particularly significant in urban risk assessment in order to identify the most appropriate short- and long-term earthquake mitigation policies.
Deliverable D4.1 - Building damage prediction equation for induced and natural seismic ground-motion parameters
 
Performance-based earthquake engineering is a probabilistic decision-making framework aimed to mitigate seismic risk, based on a comprehensive scientific foundation. In this framework, the ground motion intensity measures (IM) is linked to the threshold damage parameters (EDP) to measure the expected damage of the structures. Two natures of IM are defined: (1) an efficient IM defined as providing the smallest variability in EDP given IM relation, and (2) sufficient IM defined as providing EDP conditionally independent of earthquake magnitude and distance. Most of the studies use numerical methods to model the building response for given IM. Experimental data from the full-scale observations are much more representative of the complex physical process than even the most sophisticated laboratory or numerical experiments. Integrating these data into our modes helps to identify the sources of epistemic uncertainty. The objective of this study is to use experimental data collected from the buildings to explore the sources of epistemic uncertainties in EDP given IM relationship (i.e. σEDP|IM) and to test the IM efficiency and sufficiency. For this, we developed a database by collecting real strong motion values recorded at the top and at the bottom floors of the buildings from the US, Japan, and Romania.  The database contains 8,520 strong motion recordings that correspond to 118 buildings and 2,737 events. Several ordinate and spectral IM are considered as ground motion IMs and the normalized relative roof displacement of the building is considered as EDP. The relationship between EDP given IM is analyzed in order to identify the associated sources of uncertainties (i.e. σEDP|IM). Region-to-region, building-to-building, and within-building uncertainties associated with earthquake magnitude-distance and aging is explored. The efficiency and sufficiency of each IM from a large set of building and earthquake motion data are tested for different criteria characterizing the seismic source (magnitude and source-to-site distance) and considering several building classes, and a specific single-building analysis including aging due to cumulative earthquake damage over time is developed by utilizing the available experimental data and considering the most efficient IMs.                                                                                                                                                                                                       
Fig1. Summary of the variation of σEDP|IM values as a function of the IMs concerned, considering different components of the uncertainties in prediction models. The x-axis is the IMs considered (PGA/PGV/PGD: peak ground acceleration/velocity/displacement, AI: arias intensity, DP: destructive potential, CAV: cumulative absolute velocity, SA1/SV1/SD1: spectral acceleration/velocity/displacement at pre-seismic fundamental period, and SA2/SV2/SD2: spectral acceleration/velocity/displacement at co-seismic fundamental period). The y-axis is the σEDP|IM considering all dataset (black circle), all Japanese dataset (grey circle), Japanese steel reinforced concrete (SRC) building dataset (grey square), long term monitored  specific SRC building (ANX) from Japanese dataset (grey triangle), ANX dataset applying magnitude (M) and epicentral distance (R) criteria (MR2: R =120 ± 60% and M= 4.5 ± 0.5) (black asterisk), ANX dataset observed between (T3: 2011/03/01 to 2011/09/30) including magnitude-distance criteria MR2 (grey diamond).

What decisions for 2021?

In order to define COVID-19 impacts on the URBASIS schedule and to remodel the project in compliance with the participants needs, the Supervisory Board met and took the following decisions:
  • Workshop 3, expected for March 2021 will be organized online by ETHZ 
  • There will be no more face-to-face training activities until Sept. 2021
  • Each month a WP meeting will be organized to enable supervisors and ESRs to gather and discuss their topics. 

Progress of WP1 - Low probability / High consequences LPHC earthquakes - after a year of research

The aim of WP1 is to better understand the challenges that low probability/high consequence earthquakes pose to the scientific community and to the society. The research promoted within WP1 during the first year of PhD developed along different directions, involving both empirical analysis and ground motion simulations. Of particular interest for the ground shaking prediction in low seismic regions is the possibility to transfer models calibrated elsewhere.

Therefore, with ESR1.1 (Jaleena Sunny), a new metric to compare the observed and predicted data has been proposed, tested using also synthetic data, and applied to rank calibrated ground motion models given other sets of seismic data. The ranking of models, originally developed by the ESR1.1, is currently under peer-review, and will be presented at the next AGU Fall Meeting.

Another key issue in seismic risk assessment of large urban areas is the modeling of the spatial correlation of ground motion intensity measures. Within this context, the ESR1.3 (Jiayue Lin) has focused on the systematic analysis of the spatial correlation of broadband physics-based simulated ground motions, generated through the spectral element code SPEED coupled with an Artificial Neural Network technique for the high-frequency part, for different regions and fault rupture scenarios.

Finally, with ESR1.2 (Ming-Hsuan Yen), the physical factors driving the scaling of the period of large velocity ground-motion pulses with magnitude, and the pulses within-event variability, have been characterized using earthquakes occurred in Japan and Taiwan. The results show that the slip asperity characteristics and not only the magnitude and size of the earthquakes are critical for the occurrence of the strong-velocity pulses. The manuscript presenting the results of the research has been finalized and it is ready for the peer-review process.

How do ESRs feel after one year of research?

Andres Felipe Hernandez Estrada - ESR4.2
 
“[...] a first year full of personal and professional experiences [...]”

"During the last year, so many things have changed in my daily life and I would say that most of them were positive; moving to a new country is always exciting and even more when you are part of an interesting and challenging project like URBASIS. Unfortunately, it would be impossible to remember the first year without the world pandemic, because it changed us in so many different ways as well as how we interact with each other. From the academic point of view, although from home we could continue with our research activities, the uncertainty about the health situation around us, and the modifications in training activities, seminars, and conferences affected us in one way or another. However, after all of this, I could say that human beings are resilient and we adjust ourselves to adversities to overcome them. Personally, I found different ways to be productive as well as to relax during my free time, so a first year full of personal and professional experiences and with the best expectations for the years to come."

Astha Poudel - ESR4.5

“First-year was about getting acquaintance with the field of my research, trying to understand the different approaches and past work in the field of systemic vulnerability.”

"Realizing the immense importance of your research work and its impact on society gives you a strong motivation to persistently go ahead. Strong interest that has been developed about the subject is the most prominent achievement of my first year in this project.
I am currently working at the Aristotle University of Thessaloniki, Greece. My PhD is related to the systemic vulnerability and risk analysis on the city scale. Moving to a new place brought different opportunities, challenges and completely new experiences. First-year was about getting acquaintance with the field of my research, trying to understand the different approaches and past work in the field of systemic vulnerability. Being a part of the URBASIS project also gave me the opportunity to have direct networking with the experts in the field of earthquake engineering. Since it was a journey initiated along with other fourteen early-stage researchers like myself, I found new friends with whom I could share similar circumstances."

Yu Chen - ESR 4.4

“It was a struggle to be forced at home and try to find a balance between work and life.”
 

"It's hard to believe that it has been almost a year since our first General Assembly in Potsdam, the starting point.
We were thrilled then by the fantastic training opportunities given by URBASIS for boosting communication and collaboration between ESRs.
However, the COVID19 just surprisingly came and hit the Europe continent, bringing paradigm-shift changes to our life. 'Postponement', 'cancellation' and 'virtual' seemed to have made frequent appearances in our emails during that 5-month-long lockdown in the UK. It was a struggle to be forced at home and try to find a balance between work and life. But on the bright side, the solitary life might also serve as an opportunity for some purposeful people to 'nerdly' make achievements. There's some kind of irony that a whole year has passed and we spent half of it at home. Now the second wave is coming, and sadly Liverpool happens to be the region with the highest risk that has been forced into a Tier 3 lockdown. Hope everybody can hang in there. The show must go on.