**Abstract**

From 2017 till 2020 a low cost seismic sensor network was built in the southern Vienna Basin, Lower Austria, as a part of ongoing educational and citizen science projects. The purpose of the project is to inform society about the seismic activity in this area and to include authorities and interested citizens into data acquisition and exploitation. Near real time (NRT) seismic data are made accessible online. Seismic events are detected and archived automatically. The visualization of these events online facilitates instantaneously estimates of the extent of the shaking area and potential damage. Peak ground velocities (PGV) are related to macroseismic intensities (EMS-98) derived from reports about ground motion felt in the vicinity of the network stations. Observed amplitudes and travel times are modeled by simple, but effective relations. Traditional and innovative localization methods based on travel times and amplitudes are applied and analyzed with respect to data quality and localization accuracy. All results are accessible online and the computer code is open and applicable, e.g. for educational purposes.

**Keywords:** public seismic network, NRT ground motion watching, peak ground velocity, macroseismic intensity, earthquake localization

## **1. Introduction**

Instrumental seismology started worldwide at the beginning of the 20th century. Data acquired by the continuously improved seismometers built and still build the basis for our present-day knowledge about seismic waves, the structure of the Earth's interior, the origin of earthquakes and their impact on infrastructures and humans. However, earthquake phenomena have been fascinating and even threatening mankind from time immemorial. Systematic documentation and classification was based on observations and reports of educated persons, officers, chroniclers, clerics, and presumably only a small fraction of scientists and specialists. To summarize, seismological research before 1900 was only possible with the contribution of the public.

Nowadays, the evaluation of reports about felt ground motion and damage caused by earthquakes is treated by a seismological subdiscipline. Historic macroseismic intensity scales (e.g., Rossi - Forel, Mercalli, Cancani, Medvedev Sponheuer – Karnik) have been refined (e.g. EMS-98) and correlations with

instrumentally recorded ground motions have been established. Near real-time (NRT) preparation of so-called instrumental intensity maps is a scientific task to support mitigation in case of an earthquake. However, the reports of citizens on their perceptions of ground motions during earthquakes and damage is still an essential scientific input. We concentrate on the following tasks to promote the interest of Citizen Science in seismological data acquisition and analysis:

Citizens are frequently prepared to report their perceptions about ground motion. However, they also want to immediately know, if a ground motion was caused by an earthquake, a blast in a nearby quarry or only by a very local source such as traffic or construction works. The public is interested in whether damage to buildings occurred or health and safety were at stake. We intend to answer these questions in NRT and intuitively interpretable information via the internet, based on the data provided by a public low-cost seismic network.

The stations of this low-cost sensor network are installed in private homes and industrial buildings, schools and offices. These locations are representative of places where people observe ground motion and report it. At best we get reports from citizens about felt ground motion directly from station locations. We take advantage of these circumstances to establish a very close correlation of instrumental data and intensity classifications.

Students of polytechnics were and are still involved in the production of lowcost sensors, coding of digitizers, and developing special tools for data visualization. We intend to maintain these cooperations, but also to demonstrate that seismological data analysis must not be a black box for students of polytechnics or grammar schools, alumni, and interested citizen. We will show that accurate hypocenter localization is possible with data from the low-cost sensor network. We try to achieve this goal with easily understandable algorithms.

**2.2 Sensor and network**

*marks the extent of the map.*

**Figure 1.**

throughout this chapter.

**39**

The development of our low-cost seismic sensor started within the scope of the

*Topography, geology (VBTF … Vienna Basin transfer fault) and felt earthquakes since 1200 (solid circles); the dashed polygon delimits the low-cost seismic sensor network. Insert: European seismic Hazard map, rectangle*

*Seismological Data Acquisition and Analysis within the Scope of Citizen Science*

*DOI: http://dx.doi.org/10.5772/intechopen.95273*

The essential MSS components are two orthogonally oriented, horizontal geophones, two 16bit Analogue to Digital Converters (ADC), and a single board computer (SBC), specifically a Raspberry Pi (**Figure 2a**). The 4.5 Hz natural frequency and the 0.7 damping coefficient of the geophones and first order 12.5 Hz RC low-pass filters determine the frequency response of the MSS (**Figure 2b**). The whole assembly is protected by a robust casing. The SBC controls signal processing and provides internet connectivity. Depending on the programmable pre-amplification gain of the ADC the sensitivity ranges from 0.28 μm/s/count to 2.24 μm/s/count. Accurate time information is provided by Network Time Protocol (NTP). Seismic data is formatted to MSEED (100 Hz sample rate) and sent every 10 s to the MSS-Server. The MSS deployment started in 2017. Up until October 2020 a total of 48 MSS were installed in the southern Vienna Basin and the surrounding area in the province Lower Austria. The selection and deployment of the MSS stations received much support from the federal warning center, local authorities, schools, one

national educational project "Schools & Quakes"in 2015 [4]. One goal was the design and assembly of seismic low-cost sensors from scratch until final operation by students of polytechnic schools. This activity was inspired by the Quake-Catcher Network [5], where low-cost MEMS accelerometers either integrated into computers or in external units are used to form a world wide seismic network. However, we could not reach the desired sensitivity on the basis of low-cost MEMS (Micro-Electrical–Mechanical-Sensor) accelerometers. Following the Raspberry Shake seismograph [6] we changed to classical geophones to transform ground motions into electrical signals. Our low-cost sensor is dedicated to collecting quantitative ground motion data of felt local earthquakes. Therefore, we call it "MacroSeismic Sensor" or MSS in order to emphasize its purpose. The term MSS will be used for our sensor

## **2. Public seismic sensor network**

#### **2.1 Area**

We chose the southern Vienna Basin and its surroundings for the installation of a low-cost seismic sensor network. This area belongs to the zone of relative high seismic hazard in Austria and is densely populated and industrialized. Therefore we reasonably presume that ground motion caused by earthquakes or other sources is an interesting issue for officials and citizens.

The Vienna basin is a representative example of a pull apart basin well-explored and documented in geological literature (e.g. [1]). The basin was created by lateral extrusion of the most eastern part of the Eastern Alps from the compressional zone in the west to the extensional Pannonian Basin in the east during Miocene [2]. The basin reaches a maximum depth of 6 km. It is surrounded by Austroalpine Crystalline, the Northern Calcareous Alps, and Flysch. Shallow quaternary sediments not outlined in the schematic geological map (**Figure 1**) may significantly influence the seismic response. The Vienna basin transfer fault (VBTF) corresponds to the southern strike-slip boundary of the pull apart basin. It is still active and constitutes the main tectonic process responsible for the seismicity in this area.

Since 1200 AD about 460 earthquakes have been documented as felt and classified according to the European macroseismic scale EMS-98 in or near the southern Vienna Basin [3]. The highest epicentral intensities have been evaluated for the Schwadorf (8th October 1927, I0 = VIII) and the Seebenstein (16th April 1972, I0 = VII-VIII) earthquakes.

*Seismological Data Acquisition and Analysis within the Scope of Citizen Science DOI: http://dx.doi.org/10.5772/intechopen.95273*

#### **Figure 1.**

instrumentally recorded ground motions have been established. Near real-time (NRT) preparation of so-called instrumental intensity maps is a scientific task to support mitigation in case of an earthquake. However, the reports of citizens on their perceptions of ground motions during earthquakes and damage is still an essential scientific input. We concentrate on the following tasks to promote the interest of Citizen Science in seismological data acquisition and analysis:

Citizens are frequently prepared to report their perceptions about ground motion. However, they also want to immediately know, if a ground motion was caused by an earthquake, a blast in a nearby quarry or only by a very local source such as traffic or construction works. The public is interested in whether damage to buildings occurred or health and safety were at stake. We intend to answer these questions in NRT and intuitively interpretable information via the internet, based

The stations of this low-cost sensor network are installed in private homes and industrial buildings, schools and offices. These locations are representative of places where people observe ground motion and report it. At best we get reports from citizens about felt ground motion directly from station locations. We take advantage of these circumstances to establish a very close correlation of instrumental data

Students of polytechnics were and are still involved in the production of lowcost sensors, coding of digitizers, and developing special tools for data visualization. We intend to maintain these cooperations, but also to demonstrate that seismological data analysis must not be a black box for students of polytechnics or grammar schools, alumni, and interested citizen. We will show that accurate hypocenter localization is possible with data from the low-cost sensor network. We try to

We chose the southern Vienna Basin and its surroundings for the installation of a

The Vienna basin is a representative example of a pull apart basin well-explored and documented in geological literature (e.g. [1]). The basin was created by lateral extrusion of the most eastern part of the Eastern Alps from the compressional zone in the west to the extensional Pannonian Basin in the east during Miocene [2]. The basin reaches a maximum depth of 6 km. It is surrounded by Austroalpine Crystalline, the Northern Calcareous Alps, and Flysch. Shallow quaternary sediments not outlined in the schematic geological map (**Figure 1**) may significantly influence the seismic response. The Vienna basin transfer fault (VBTF) corresponds to the southern strike-slip boundary of the pull apart basin. It is still active and constitutes

Since 1200 AD about 460 earthquakes have been documented as felt and classified according to the European macroseismic scale EMS-98 in or near the southern Vienna Basin [3]. The highest epicentral intensities have been evaluated for the Schwadorf (8th October 1927, I0 = VIII) and the Seebenstein (16th April 1972,

low-cost seismic sensor network. This area belongs to the zone of relative high seismic hazard in Austria and is densely populated and industrialized. Therefore we reasonably presume that ground motion caused by earthquakes or other sources is

the main tectonic process responsible for the seismicity in this area.

on the data provided by a public low-cost seismic network.

achieve this goal with easily understandable algorithms.

and intensity classifications.

*Earthquakes - From Tectonics to Buildings*

**2. Public seismic sensor network**

an interesting issue for officials and citizens.

I0 = VII-VIII) earthquakes.

**38**

**2.1 Area**

*Topography, geology (VBTF … Vienna Basin transfer fault) and felt earthquakes since 1200 (solid circles); the dashed polygon delimits the low-cost seismic sensor network. Insert: European seismic Hazard map, rectangle marks the extent of the map.*

#### **2.2 Sensor and network**

The development of our low-cost seismic sensor started within the scope of the national educational project "Schools & Quakes"in 2015 [4]. One goal was the design and assembly of seismic low-cost sensors from scratch until final operation by students of polytechnic schools. This activity was inspired by the Quake-Catcher Network [5], where low-cost MEMS accelerometers either integrated into computers or in external units are used to form a world wide seismic network. However, we could not reach the desired sensitivity on the basis of low-cost MEMS (Micro-Electrical–Mechanical-Sensor) accelerometers. Following the Raspberry Shake seismograph [6] we changed to classical geophones to transform ground motions into electrical signals. Our low-cost sensor is dedicated to collecting quantitative ground motion data of felt local earthquakes. Therefore, we call it "MacroSeismic Sensor" or MSS in order to emphasize its purpose. The term MSS will be used for our sensor throughout this chapter.

The essential MSS components are two orthogonally oriented, horizontal geophones, two 16bit Analogue to Digital Converters (ADC), and a single board computer (SBC), specifically a Raspberry Pi (**Figure 2a**). The 4.5 Hz natural frequency and the 0.7 damping coefficient of the geophones and first order 12.5 Hz RC low-pass filters determine the frequency response of the MSS (**Figure 2b**). The whole assembly is protected by a robust casing. The SBC controls signal processing and provides internet connectivity. Depending on the programmable pre-amplification gain of the ADC the sensitivity ranges from 0.28 μm/s/count to 2.24 μm/s/count. Accurate time information is provided by Network Time Protocol (NTP). Seismic data is formatted to MSEED (100 Hz sample rate) and sent every 10 s to the MSS-Server.

The MSS deployment started in 2017. Up until October 2020 a total of 48 MSS were installed in the southern Vienna Basin and the surrounding area in the province Lower Austria. The selection and deployment of the MSS stations received much support from the federal warning center, local authorities, schools, one

**Figure 2.**

*MSS – MacroSeismic sensor; (a) geophones, ADC and SBC mounted on base plate; (b) frequency response between f = 1 Hz and fNyquist = 50 Hz, (c) student at polytechnic Wiener Neustadt assembling MSS; (d) MSS mounted in office of district exchange Bruck an der Leitha together with contact persons.*

**3. Event detection, visualization, and archiving**

PGV (**Figure 5a**),

**Figure 3.**

*transparent symbols PGV of last minute.*

*DOI: http://dx.doi.org/10.5772/intechopen.95273*

instead of PGV (**Figure 5c, d**).

**41**

The PGV map as shown in **Figure 3** is transient. Significant seismic events should be detected and saved in order to keep this information and to make it available for more detailed analysis. The definition of a seismic event and proper trigger criteria should take the data quality into consideration. MSS stations are intentionally mounted in buildings where people potentially experience ground motions and report their observations. These places are frequently noisy and even high PGVs may be recorded due to nearby activities (e.g. traffic, construction work, washing machine, etc.). The main objective of a detection algorithm is to distinguish high amplitude noise at individual stations and regional events like earthquakes or quarry blasts. We perform a "Delaunay" triangulation [7] of the MSS station network and examine the triples of PGV values belonging to the different triangles. Once the minimum PGV value within one triple exceeds a preselected threshold the recorded PGV at all MSS-stations are classified as a seismic event. The duration of the seismic event is expanded by the triggering of other Delaunay triangles and prolongated by a listening time window. This time window takes care of the propagation of the maximum amplitude seismic waves over the network area. **Figure 4** shows the temporal sequence of the trigger status for an entire seismic event. As soon as a seismic event ends, the seismic data of the respective time window is archived. We offer two options for the visualization of the whole seismic event:

*(a) Recordings (stacked horizontal components) of the time interval 2019-06-14 12: 33:30–12:34:30; (b) corresponding NRT PGV-map; opaque symbols show PGV of last second (gray background in a),*

*Seismological Data Acquisition and Analysis within the Scope of Citizen Science*

• Coloring the Voronoi regions [8] of each MSS station according to the event

• contouring the PGVs at the MSS-stations by the Kriging method (**Figure 5b**).

The PGV values observed during an earthquake are strongly affected by specific geological and technical peculiarities at the individual MSS stations. In the Section 5.1 we introduce station amplification factors "SA" to improves the fit of PGV to a power law amplitude - distance relation. The application of SA significantly improve the spatial correlation of PGV. Contours become much smoother and better delineate the areas of felt ground motions and maximum shaking. Therefore, we also offer the aforementioned data display alternatively based on PGV/SA

Both visualizations are available in NRT after the seismic event.

quarry operator, and private citizens. The MSS were mounted by a single plug to a vertical, preferably a retaining wall in solidly constructed buildings, mainly at basement, ground floor, or first floor level.

### **2.3 Near real-time ground motion watching**

A third essential component of the MSS network, beside the MSS stations and the MSS-Server, is the MSS-homepage that provides data visualizations and access to numeric data (https://www.macroseismicsensor.at/). The MSS-network is meant to inform communities, governmental administration, civil protection organizations and last but not least citizens about the felt or presumed seismic activity in the southern Vienna Basin. Ground motions take 10–30 seconds to travel over the whole area of the MSS-net from the epicenter. Immediately people are curious to know the source of the vibration. The authorities contacted should be able to answer, at least preliminarily, these questions on the basis of the information and visualization provided on the internet. In the case of a stronger earthquake (intensities > = V), the staff of the civil protection organization should know if panic could arise or if there was damage to buildings and sensitive infrastructure. Therefore, visualization of the essential seismic data should be swiftly available and understandable.

We mainly try to meet these demands by using a map of the MSS data. We determine Peak Ground Velocity (PGV) as the maximum resultant horizontal ground velocity over a time interval of 1 sec and visualize it at each MSS station with symbols. The map is updated every 10 seconds. **Figure 3** shows seismic recordings and the essential components of the PGV-map for a 60 s time window covering the ML = 2.5 earthquake on 14th June 2019.

*Seismological Data Acquisition and Analysis within the Scope of Citizen Science DOI: http://dx.doi.org/10.5772/intechopen.95273*

#### **Figure 3.**

quarry operator, and private citizens. The MSS were mounted by a single plug to a vertical, preferably a retaining wall in solidly constructed buildings, mainly at

*MSS – MacroSeismic sensor; (a) geophones, ADC and SBC mounted on base plate; (b) frequency response between f = 1 Hz and fNyquist = 50 Hz, (c) student at polytechnic Wiener Neustadt assembling MSS; (d) MSS*

*mounted in office of district exchange Bruck an der Leitha together with contact persons.*

A third essential component of the MSS network, beside the MSS stations and the MSS-Server, is the MSS-homepage that provides data visualizations and access to numeric data (https://www.macroseismicsensor.at/). The MSS-network is meant to inform communities, governmental administration, civil protection organizations and last but not least citizens about the felt or presumed seismic activity in the southern Vienna Basin. Ground motions take 10–30 seconds to travel over the whole area of the MSS-net from the epicenter. Immediately people are curious to know the source of the vibration. The authorities contacted should be able to answer, at least preliminarily, these questions on the basis of the information and visualization provided on the internet. In the case of a stronger earthquake (intensities > = V), the staff of the civil protection organization should know if panic could arise or if there was damage to buildings and sensitive infrastructure. Therefore, visualization of the essential seismic data should be swiftly available and

We mainly try to meet these demands by using a map of the MSS data. We determine Peak Ground Velocity (PGV) as the maximum resultant horizontal ground velocity over a time interval of 1 sec and visualize it at each MSS station with symbols. The map is updated every 10 seconds. **Figure 3** shows seismic recordings and the essential components of the PGV-map for a 60 s time window covering the

basement, ground floor, or first floor level.

*Earthquakes - From Tectonics to Buildings*

ML = 2.5 earthquake on 14th June 2019.

understandable.

**40**

**Figure 2.**

**2.3 Near real-time ground motion watching**

*(a) Recordings (stacked horizontal components) of the time interval 2019-06-14 12: 33:30–12:34:30; (b) corresponding NRT PGV-map; opaque symbols show PGV of last second (gray background in a), transparent symbols PGV of last minute.*
