This post is based on small research done inside a Kaggle notebook, using publicly available data: https://www.kaggle.com/code/iananich/analyse-carpathian-earthquakes-with-isc?scriptVersionId=203729120

Data was downloaded from ISC-EHB Bulletin using a simple script inside that notebook. It’s worth noting that while ISC provides multiple magnitude estimations for each earthquake, for the sake of simplicity, only the first one was used (which is ISC’s preferred one).

Plots are built with Plotly — a handy library that produces interactive plots that are great for data exploration.

—

Plot 1

Plot 1 presents a map of the Carpathian mountains, with a beautiful background layer provided by USGS.

Individual hexagons are colored according to total number of earthquakes recorded in a period starting from 1970 up to 2023. Regions with under 50 events are discarded (otherwise this image would be almost completely yellow — there’s barely any calm place). Most earthquakes were recorded in dark purple spots — there are three of them: in North-West, South-East, and South.

The red square represents the limits of the search rectangle — earthquakes outside of it are outside of the scope of this research.

Below there’s a histogram, with each bar’s height set to the number of earthquakes in each year.

Plot 2

Year 2023 stands out. If you are curious why the year 2024 is not discussed, the answer is simply — as of right now there’s very little data even from November 2023, let alone the following months.

—

Zooming in

Plot 3

This point on Google Maps: https://maps.app.goo.gl/4D4fmTk6iVyEZzgt8

It’s pretty close to Târgu Jiu

For plot 3 and onward, a smaller region is selected. This new map uses much smaller hexagons and helps localize earthquakes.

Below in plot 4 individual bars only count events per month.

Plot 4

There wasn’t much happening before 2023. Also notice that the earliest events are from 1990, unlike on the previous histogram from Plot 2

Let’s look at the depth and magnitude of these events:

Plot 5

Plot 5 shows an interesting phenomenon where depth distribution didn’t change much, but color (which is set according to magnitude) shifted to weaker magnitudes. That puts a concern that perhaps processes didn’t change, while estimation of strength — did. This is further emphasized by the fact that on plot 2 we saw a decline of increase in earthquakes per year in about the same year.

Change in magnitude estimation is more evident from the following plot:

Plot 6

Plot 6 also highlights unusually high magnitudes from 2023.

—

Year 2023

Note: plots 7, 8, 9 are just the same 4, 5, 6 but zoomed in on time scale — interactivity of Plotly plots allows you to adjust the viewport and take a screenshot.

Plot 7

Zoomed-in histogram nicely displays that most events occurred in February and March of 2023.

Plot 8

Note that most earthquakes happened between depths of 10 and 20 km. The median depth for events in this area in the whole year 2023 is at 16.5 km, q1 and q3 at 14.8 and 17.7 km respectively.

Also quite interesting that in the beginning of activity variation in depth was much larger.

Plot 9

That earthquake in February 2023 had a magnitude of 5.6, there’s more details about it on VolcanoDiscovery’s page: https://allquakes.com/earthquakes/quake-info/7391116/mag5quake-Feb-14-2023-Romania.html

It’s important that in a close proximity of that earthquake (in Târgu Jiu, at distance 8km) its max shaking intensity was reported as VI “Strong” on the Modified Mercalli Intensity Scale, meaning “Felt by all, any frightened. Some heavy furniture moved; few instances of fallen plaster. Damage slight”.

—

Would be interesting to get an opinion from those who have more expertise in the field: what it could have been? What to expect next?

But also, I want to encourage everyone with basic coding skills to do their own research — data is out there…