A comparison of GlobalQuake preliminary earthquake location data with INGV final results
Earthquake in Calabria on the 20th April 2024: provisional estimates and localization times, a comparison between Globalquake and INGV data
⚠️ This article cannot begin without stating that we are not criticizing the efforts of the Italian National Institute of Geophysics and Volcanology (INGV).
We highly regard their commitment to seismic research and recognize how important their input has been towards the geophysics and to all of the scientific community.
Building upon the valuable groundwork laid by INGV this article explores the ongoing advancements in seismic monitoring technology exemplified by GlobalQuake.
Context
On April 20th, 2024, at around 04:05 (UTC+2), an earthquake with a magnitude of 3.5 ML occurred in southern Calabria, in the province of Reggio Calabria.
On the following day, INGV followed up with a blog post, detailing their hypocentral location system and magnitude calculation methods.
The article, written by Alessandro Amato head of the INGV-CAT, states that INGV's system was able to identify the earthquake and in two minutes a preliminar estimate was published as soon as possible to social media, in particular on X (ex Twitter).
What is the issue?
Early Earthquake Warning (EEW) systems are now supported by various countries,
which acknowledged their significance in reducing risks associated with earthquakes.
It is therefore necessary for institutions such as INGV to quickly embrace advanced softwares
tailored not only for recording and analyzing earthquakes but also for potential future integration into EEW systems.
These programs provide an opportunity for better earthquake alerts that can save numerous lives and infrastructure.
For this reason, the integration of cutting-edge technologies into seismological frameworks is not just desirable but necessary for effective emergency planning and disaster response
in a world where seismic activities pose significant threats. Furthermore, it is not enough to just Tweet about the preliminar data to “warn” because if anything,
earthquakes require more strong and dependable warning systems.
Recognizing the limitations of current earthquake detection systems used by INGV, we developed an innovative, but still experimental, software which includes quicker processing times allowing for few vital seconds for earthquake alerts: GlobalQuake.
What is GlobalQuake?
GlobalQuake is an innovative, experimental, Java based software that allows for near real-time monitoring of global earthquakes. It enables users to track seismic activity worldwide by using seismic station data from publicly available networks through FDSNWS services and real-time Seedlink servers.
The program measures the arrival times of p-waves and s-waves at seismic stations so as to detect earthquakes.
With station coordinates, obtained via the FDSNWS, these data are supplied to a hypocenter location algorithm hence pinpointing earthquake epicenters.
This algorithm was invented and developed by Jakub Španger in the summer of 2023, with support from others including myself,
and will be analysed in the thesis of Jakub's for his Computer Science Bachelor, at Masaryk University, Brno, Czechia.
(PDF of the Thesis)
As soon as enough data has been gathered, GlobalQuake is able to accurately identify earthquake epicenters (shown by ‘X’ marks) and the propagation of P and S waves shown on an interactive three-dimensional map of the world. GlobalQuake’s algorithms provided, also, instant access to crucial information related to key earthquake parameters such as magnitude, depth, and maximum intensity among others.
The speed of GlobalQuake
According to INGV, at 04:05:18 (UTC+2) on April 20, 2024, an earthquake occurred in the Calabria Region.
GlobalQuake at 04:05:27, after just 9 seconds of the origin, identified that an event had occurred in the Calabria Region,
and the estimate placed the magnitude at 3.4 with a low quality rating (D) in the Calabria Region.
Over the next 4 seconds, GlobalQuake refined its estimates, lowering the magnitude to 3.2 but improving the quality rating to average (from D to C, and from C to B).
Finally, after 13 seconds, GlobalQuake settled on a magnitude of 3.6 with the best-quality rating based on data from only 13 stations (S).
(GlobalQuake uses an accuracy ratings where D=Bad, C=Low, B=Average, A=Good, S=BEST)
GlobalQuake confirmed the earthquake and continuously refined its data over 15 revisions as more stations provided information. The final revision (15th), considered the most accurate due to using the most data, estimated a magnitude of 3.1 at a depth of 10.7 kilometers (with an uncertainty of 1.8 kilometers) located at coordinates 38.3330, 16.0607 .
Interestingly, this final assessment by GlobalQuake occurred just two minutes after the earthquake's origin, probably even before the first INGV tweet had been published.
(Article in italian on how INGV uses Twitter for preliminar information)
It is important to note that INGV's two-minute timeframe for releasing preliminary data is not the first solution generated by their software. As an INGV representative explained, INGV decided to prioritizes data accuracy over speed to ensure the reliability of the information they release.
GlobalQuake Revisions
Here's a table of all the revisions:
| Revision | Latitude | Longitude | Depth | Magnitude | Station used | Quality | Calculation Time | Speed |
|---|---|---|---|---|---|---|---|---|
| 1 | 38,2835 | 16,1018 | 10,7±0,0km | 3,4 | 4 | D | 04:05:27 | 00:00:09 |
| 2 | 38,2882 | 16,1096 | 10,7±0,0km | 3,2 | 6 | C | 04:05:28 | 00:00:10 |
| 3 | 38,3103 | 16,0882 | 12,7±1,9km | 3,2 | 8 | B | 04:05:30 | 00:00:12 |
| 4 | 38,3070 | 16,0920 | 12,7±0,1km | 3,6 | 13 | S | 04:05:31 | 00:00:13 |
| 5 | 38,3154 | 16,0852 | 18,6±0,6km | 3,5 | 21 | S | 04:05:32 | 00:00:14 |
| 6 | 38,3155 | 16,0853 | 18,6±1,8km | 3,5 | 26 | S | 04:05:34 | 00:00:16 |
| 7 | 38,3154 | 16,0852 | 21,0±0,6km | 3,5 | 33 | S | 04:05:36 | 00:00:18 |
| 8 | 38,3213 | 16,0690 | 19,5±0,0km | 3,5 | 47 | S | 04:05:41 | 00:00:23 |
| 9 | - | - | 18,6±??? | 3,5 | - | - | 04:05:45 | 00:00:27 |
| 10 | 38,3181 | 16,0728 | 18,1±0,1km | 3,4 | 69 | S | 04:05:50 | 00:00:32 |
| 11 | 38,3264 | 16,0661 | 14,2±0,2km | 3,4 | 86 | S | 04:05:55 | 00:00:37 |
| 12 | 38,3264 | 16,0661 | 15,1±1,8km | 3,3 | 104 | S | 04:06:04 | 00:00:46 |
| 13 | 38,3311 | 16,0738 | 12,7±0,1km | 3,3 | 125 | S | 04:06:26 | 00:01:08 |
| 14 | 38,3358 | 16,0473 | 12,7±0,1km | 3,3 | 151 | S | 04:06:41 | 00:01:23 |
| 15 | 38,3330 | 16,0607 | 10,7±1,8km | 3,1 | 181 | S | 04:07:25 | 00:02:07 |
There has been a data gap for revision number 9 due to a temporary interruption on the GlobalQuake Italia YouTube Live Stream. While the recording showed the revision process, some information, specifically the location data, for revision 9 was obscured by another event in Greece.
How much off was the data?
There's a slight discrepancy between the final locations determined by GlobalQuake and INGV. GlobalQuake's location is 1.44 kilometers away from INGV's manual assessment.
However, it's important to consider the margin of error. GlobalQuake indicated a probable location with a 2.30 kilometer radius, and both INGV's and GlobalQuake's results fall within this designated area.
The difference in magnitude is also minimal, with only a 0.4 variation between the two assessments.
However, the difference in magnitude needs further consideration. While the variation between the two assessments is 0.4, an INGV representative pointed out that in terms
of energy release, a magnitude difference of 0.4 can be significant. For example, a magnitude 3.1 earthquake releases roughly four times less energy than a magnitude 3.5 earthquake.
Therefore, even a seemingly small difference in magnitude can have a substantial impact.
Similarly, the depth variations are relatively small, with a maximum difference of 4.3 kilometers.
Here's the graph of the calculations done by GlobalQuake:
Conclusion
This analysis highlights the promise of innovative solutions like GlobalQuake for earthquake early warning systems. By embracing new technologies, INGV has the opportunity to become a leader in developing and implementing more efficient systems, ultimately safeguarding lives throughout Italy in the event of an earthquake.
It's important to note that GlobalQuake is still under development, and like any new technology, it may occasionally experience errors. For instance, there was a recent incident where it mistakenly identified an earthquake in France with a magnitude of 10.0.
As highlighted by an INGV representative, the organization has historically adhered to a policy of withholding automatic solutions to prevent the dissemination of false alarms or inaccurate information. This cautious approach stems from a deep understanding of the delicate nature of their role and the potential consequences of releasing unreliable data. This demonstrate INGV's dedication to striking a balance between embracing technological advancements and ensuring the utmost reliability of the information they provide to the public.
Data
The networks used by GlobalQuakes are IT, IY, IX and some minor network such as GE, GR, VD and others.
Replay
Here's the link of the replay of the GlobalQuake calculation:
INGV Data
INGV, if you are reading this, can you give us the access to your networks so that we can improve our software? Thanks.
Final Discaimer
This is my inaugural attempt at writing an article of this sort. Forgive me in case I made any errors or omissions. be assured that I am dedicated to learning and improving. I would like to thank you in advance for your patience and understanding as I undertake to improve on them in the coming days.
Thank you!