Laboratory Earthquake Prediction

Vishal Kumar
Department of Mathematics.
Birla Institute of Technology Mesra.

Abstract
Forecasting earthquakes is one of the most important problems in Earth science because of their
devastating consequences. Current scientific studies related to earthquake forecasting focus on three
key points: when the event will occur, where it will occur, and how large it will be.
In this project, I will address when the earthquake will take place. Specifically, I will predict the time
remaining before laboratory earthquakes occur from real-time seismic data.
If this research project is solved and the physics are ultimately shown to scale from the laboratory to
the field, researchers will have the potential to improve earthquake hazard assessments that could save
lives and billions of rupees in infrastructure.

1.Introduction
An earthquake is the shaking of the surface of the Earth, resulting from the sudden release of energy
in the Earth’s lithosphere that creates seismic waves. Earthquakes can range in size from those that are
so weak that they cannot be felt to those violent enough to toss people around and destroy whole
cities. The seismicity, or seismic activity of an area is the frequency, type and size of earthquakes
experienced over a period. The word tremor is also used for non-earthquake seismic rumbling.
At the Earth's surface, earthquakes manifest themselves by shaking and displacing or disrupting the
ground. When the epicentre of a large earthquake is located offshore, the seabed may be displaced
sufficiently to cause a tsunami. Earthquakes can also trigger landslides, and occasionally volcanic
activity.
In its most general sense, the word earthquake is used to describe any seismic event—whether natural
or caused by humans—that generates seismic waves. Earthquakes are caused mostly by rupture of
geological faults, but also by other events such as volcanic activity, landslides, mine blasts,
and nuclear tests. An earthquake's point of initial rupture is called its focus or hypocentre. The
epicentre is the point at ground level directly above the hypocentre.

Earthquake Prediction is a branch of the science of seismology concerned with the specification of the
time, location, and magnitude of future earthquakes within stated limits. Many methods have been
developed for predicting the time and place in which earthquakes will occur. Despite considerable
research efforts by seismologists, scientifically reproducible predictions cannot yet be made to a
specific day or month.

Forecasting.
While forecasting is usually considered to be a type of prediction, earthquake forecasting is often
differentiated from earthquake prediction. Earthquake forecasting is concerned with the probabilistic
assessment of general earthquake hazard, including the frequency and magnitude of damaging
earthquake in a given area over years or decades. For well-understood faults, the probability that a
segment may rupture during the next few decades can be estimated.
 Earthquakes (M6.0+) since 1900 through 2017

2.Data description
The goal of this project is to use seismic signals to predict the timing of laboratory earthquakes. The
data comes from a well-known experimental set-up used to study earthquake physics.
The acoustic_data input signal is used to predict the time remaining before the next laboratory
earthquake (time_to_failure).
The training data is a single, continuous segment of experimental data. The test data consists of a
folder containing many small segments. The data within each test file is continuous, but the test files
do not represent a continuous segment of the experiment; thus, the predictions cannot be assumed to
follow the same regular pattern seen in the training file.
For each seg_id in the test folder, you should predict a single time_to_failure corresponding to the
time between the last row of the segment and the next laboratory earthquake.

File descriptions

 train.csv - A single, continuous training segment of experimental data.

 test - A folder containing many small segments of test data.
 sample_sumbission.csv - A sample submission file in the correct format.

Data fields

 acoustic_data - the seismic signal [int16]

 time_to_failure - the time (in seconds) until the next laboratory earthquake [float64]
 seg_id - the test segment ids for which predictions should be made (one prediction per
segment)
New feature exploration

Aggregated features

Let's visualize the new features distributions. The graphs below show the distplot (histograms and
density plots) for all the new features, for train (green) and test (blue) data.
.
.

.
Scaled features
Let's scale now the aggregated features and show again the resulting graphs. We are fitting the scaler
with both train and test data. We use red from train and magenta for test data.
Aggregated features and time to failure.
Let's also show aggregated features and time to failure on the same graph .