TremorViz

The working goal of this project is to create a series of interactive data visualizations and/or dashboards that can be used to explore seismic data. These data can have X, Y, Z, and time information, so visualizing all of those dimensions can be challenging. Our hope is that by using interactive visualization tools, the user will be able to add or subtract complexity and move between scales.

Collaborators on this project

Team members: Ariane Ducellier (project lead), Marcos Llobera, Jacob Deppen, Erik Fredrickson

Data science lead: TBD

The problem

The aim of the project is to create a set of tools to visualize physical phenomena in 4D (latitude, longitude, depth, and time). The specific problem arises when we want to visualize at the same time two or more physical phenomena that are spatially and/or temporally correlated. What can we do in order to better see possible correlation patterns? Particularly, how do we deal with the case when the types of information that we have about the time and localization are very different from one physical phenomenon to another?

Application Example

Episodic Tremor and Slip (ETS) is a phenomenon that takes place mostly in subduction zones, when a geodetically detected slow slip event occurs concurrently with tectonic tremor. Tremor is a long (several seconds to many minutes), low amplitude seismic signal, with emergent onsets, and an absence of clear impulsive phases. At least a portion of the tremor is made of small low-frequency earthquakes (LFEs). Slow slip events can be detected by observing GPS time series. Tremor and LFEs can be detected, and their source can be located, using recordings of seismic stations. There is a spatial and temporal correlation between slow slip, tremor, and LFEs.

Sample data

Tremor

We can have two types of datasets:

1. For each five-minute-long time window when tremor is recorded, we have:
   • Latitude
   • Longitude
   • Approximate depth (depth is not very well known)

See Figure 1 for an example of visualization of tremor data. An example of this type of tremor dataset can be downloaded here.

Figure 1: Map of the locations of the tremors recorded in the Olympic Peninsula between May 9th and May 23rd 2018. Each dot represents the location of the source of the tremor during a five-minute-long time window. The color of the dot represents the time when the tremor was recorded (Image from [here](https://pnsn.org/tremor)).

OR

2. For each tremor episode, we have:
   • Latitude
   • Longitude
   • Approximate depth
   • Beginning time
   • Duration

An example of this type of tremor dataset can be downloaded here.

Low Frequency Earthquakes

They are divided into families. For each family, we have:

• Latitude
• Longitude
• Depth
• Timing of each earthquake
• (Sometimes magnitude of each earthquake)

See Figure 2 for an example of visualization of LFE data. A dataset of LFEs on the San Andreas Fault can be downloaded here.

Figure 2: Number of LFEs recorded per day as a function of time for a given family of LFEs located on the San Andreas Fault.

Slow slip

We can have both GPS data, and slip maps established using the GPS data.

GPS data

For each GPS station, we have the latitude and the longitude of the station, and the three components of the displacement in function of time (one value per day and per component).

See Figure 3 for an example of visualization of GPS data.

Figure 3: Longitudinal displacement measured at the GPS station PGC5 (southern Vancouver Island) as a function of time. The grey bars represent the timing of ETS events.

An example of GPS dataset can be downloaded here.

Maps

For each ETS episode, we can get the total amount of slip at the plate boundary from numerical modeling, but we have no indication about the time evolution of the slip.

See Figure 4 for an example of visualization of slow slip map.

Figure 4: Map of the total amount of slip observed during 16 ETS events between 1998 and 2008 (from Schmidt, D.A., and H. Gao (2010), Source parameters and time‐dependent slip distributions of slow slip events on the Cascadia subduction zone from 1998 to 2008, J. Geophys. Res., 115, B00A18, doi:10.1029/2008JB006045).

Specific Questions

Improve Figure 1.

Existing methods

Python visualization libraries bokeh or altair.

Proposed methods/tools

We are going to try altair.

Remaining problems to solve

• Putting a background map of western Washington
• Adding the depth contour lines of the plate boundary
• Finding a way to set the limits (in latitude, longitude) of the plot