Rémy Bossu

Rémy Bossu

orcid

Research Focus & Metrics

Academic Career Length

13Years (2014-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
26 October 2016 in Seismological Research Letters
Verified
Journal Article
Closed access
PDF file
RBRémy Bossu
Rémy Bossu
MLMatthieu Landès
Matthieu Landès
FRFréderic Roussel
Fréderic Roussel
RSRobert Steed
Robert Steed
GMGilles Mazet‐Roux
Gilles Mazet‐Roux
SHSusan Hough
Susan Hough
The collection of earthquake testimonies (i.e., qualitative descriptions of felt shaking) is essential for macroseismic studies (i.e., studies gathering information on how strongly an earthquake was felt in different places), and when done rapidly and systematically, improves situational awareness and in turn can contribute to efficient emergency response. In this study, we present advances made in the collection of testimonies following earthquakes around the world using a thumbnail‐based questionnaire implemented on the European‐Mediterranean Seismological Centre (EMSC) smartphone app and its website compatible for mobile devices. In both instances, the questionnaire consists of a selection of thumbnails, each representing an intensity level of the European Macroseismic Scale 1998. We find that testimonies are collected faster, and in larger numbers, by way of thumbnail‐based questionnaires than by more traditional online questionnaires. Responses were received from all seismically active regions of our planet, suggesting that thumbnails overcome language barriers. We also observed that the app is not sufficient on its own, because the websites are the main source of testimonies when an earthquake strikes a region for the first time in a while; it is only for subsequent shocks that the app is widely used. Notably though, the speed of the collection of testimonies increases significantly when the app is used. We find that automated EMSC intensities as assigned by user‐specified thumbnails are, on average, well correlated with “Did You Feel It?” (DYFI) responses and with the three independently and manually derived macroseismic datasets, but there is a tendency for EMSC to be biased low with respect to DYFI at moderate and large intensities. We address this by proposing a simple adjustment that will be verified in future earthquakes.

28 October 2015 in Seismological Research Letters
Verified
Journal Article
Closed access
PDF file
RBRémy Bossu
Rémy Bossu
MLMaud Laurin
Maud Laurin
GMGilles Mazet-Roux
Gilles Mazet-Roux
FRFrédéric Roussel
Frédéric Roussel
RSRobert Steed
Robert Steed
This study investigates how people affected by the 2015 Ghorka earthquake sequence in Nepal have adopted and used the LastQuake smartphone application and associated Twitter robot, which offer rapid earthquake information focusing on felt and damaging earthquakes. It is shown that mobile devices played a key role in Nepal by providing access to earthquake information and that the LastQuake smartphone application was rapidly identified and adopted through viral spread after the mainshock. Thumbnail-based questionnaires, which have replaced online macroseismic questionnaires on the smartphone application and on the website for mobile devices, ease language barriers and have proven to be very efficient for collecting testimonies within a few tens of minutes of a felt-earthquake occurrence.

5 April 2019 in Science Advances
Verified
Journal Article
Open access
PDF file
RSRobert J. Steed
Robert J. Steed
AFAmaya Fuenzalida
Amaya Fuenzalida
RBRémy Bossu
Rémy Bossu
IBIstván Bondár
István Bondár
AHAndres Heinloo
Andres Heinloo
ASAngelo Strollo
Angelo Strollo
Detection of earthquakes by seismic networks has been accelerated on a global scale using crowdsourced data.

29 September 2021 in Seismological Research Letters
Verified
Journal Article
Closed access
PDF file
RBRémy Bossu
Rémy Bossu
FFFrancesco Finazzi
Francesco Finazzi
RSRobert Steed
Robert Steed
LFLaure Fallou
Laure Fallou
IBIstván Bondár
István Bondár
Public earthquake early warning systems have the potential to reduce individual risk by warning people of approaching tremors, but their development has been hampered by costly infrastructure. Furthermore, both users’ understanding of such a service and their reactions to actual warnings have been the topic of only a few surveys. The smartphone app of the Earthquake Network initiative utilizes users’ smartphones as motion detectors and provides the first example of a purely smartphone-based earthquake early warning system, without the need for dedicated seismic station infrastructure and operating in multiple countries. We demonstrate that this system has issued early warnings in multiple countries, including for damaging shaking levels, and hence that this offers an alternative to conventional early warning systems in the foreseeable future. We also show that although warnings are understood and appreciated by users, notably to get psychologically prepared, only a fraction take protective actions such as “drop, cover, and hold.”

16 August 2016 in Natural Hazards
Verified
Journal Article
Open access
PDF file
SHSusan E. Hough
Susan E. Hough
SMStacey S. Martin
Stacey S. Martin
Vineet K GahalautVineet K Gahalaut
Vineet K Gahalaut
AJAnand Joshi
Anand Joshi
MLM. Landes
M. Landes
RBRémy Bossu
Rémy Bossu
We use 21 strong motion recordings from Nepal and India for the 25 April 2015 moment magnitude (MW) 7.8 Gorkha, Nepal, earthquake together with the extensive macroseismic intensity data set presented by Martin et al. (Seism Res Lett 87:957–962, 2015) to analyse the distribution of ground motions at near-field and regional distances. We show that the data are consistent with the instrumental peak ground acceleration (PGA) versus macroseismic intensity relationship developed by Worden et al. (Bull Seism Soc Am 102:204–221, 2012), and use this relationship to estimate peak ground acceleration from intensities (PGAEMS). For nearest-fault distances (RRUP < 200 km), PGAEMS is consistent with the Atkinson and Boore (Bull Seism Soc Am 93:1703–1729, 2003) subduction zone ground motion prediction equation (GMPE). At greater distances (RRUP > 200 km), instrumental PGA values are consistent with this GMPE, while PGAEMS is systematically higher. We suggest the latter reflects a duration effect whereby effects of weak shaking are enhanced by long-duration and/or long-period ground motions from a large event at regional distances. We use PGAEMS values within 200 km to investigate the variability of high-frequency ground motions using the Atkinson and Boore (Bull Seism Soc Am 93:1703–1729, 2003) GMPE as a baseline. Across the near-field region, PGAEMS is higher by a factor of 2.0–2.5 towards the northern, down-dip edge of the rupture compared to the near-field region nearer to the southern, up-dip edge of the rupture. Inferred deamplification in the deepest part of the Kathmandu valley supports the conclusion that former lake-bed sediments experienced a pervasive nonlinear response during the mainshock (Dixit et al. in Seismol Res Lett 86(6):1533–1539, 2015; Rajaure et al. in Tectonophysics, 2016. Ground motions were significantly amplified in the southern Gangetic basin, but were relatively low in the northern basin. The overall distribution of ground motions and damage during the Gorkha earthquake thus reflects a combination of complex source, path, and site effects. We also present a macroseismic intensity data set and analysis of ground motions for the MW7.3 Dolakha aftershock on 12 May 2015, which we compare to the Gorkha mainshock and conclude was likely a high stress-drop event.

24 May 2023 in Neural Computing and Applications
Verified
Journal Article
Open access
PDF file
FOFerda Ofli
Ferda Ofli
Muhammad ImranMuhammad Imran
Muhammad Imran
UQUmair Qazi
Umair Qazi
JRJulien Roch
Julien Roch
Catherine PenningtonCatherine Pennington
Catherine Pennington
RBRémy Bossu
Rémy Bossu
Lack of global data inventories obstructs scientific modeling of and response to landslide hazards which are oftentimes deadly and costly. To remedy this limitation, new approaches suggest solutions based on citizen science that requires active participation. In contrast, as a non-traditional data source, social media has been increasingly used in many disaster response and management studies in recent years. Inspired by this trend, we propose to capitalize on social media data to mine landslide-related information automatically with the help of artificial intelligence techniques. Specifically, we develop a state-of-the-art computer vision model to detect landslides in social media image streams in real-time. To that end, we first create a large landslide image dataset labeled by experts with a data-centric perspective, and then, conduct extensive model training experiments. The experimental results indicate that the proposed model can be deployed in an online fashion to support global landslide susceptibility maps and emergency response.