Visualizing EarthScope Science: Motion

First Place:

 

1st with 1289 votes: Kasra Hosseini - University of Oxford

"Fly-through of a seismic tomography model of Earth"

Seismic tomography is the pre-eminent tool for imaging the Earth’s interior. Since the advent of this method in the mid-1970s, the internal structure of Earth has been vastly sampled and imaged at a variety of scales, and the resulting models have served as the primary means to investigate the processes driving our planet.

Over 2.5 million cross-correlation travel times and 3.3 million high frequency picked arrival times were used to create the model in this video. USArray, the seismological component of EarthScope, was one of the main data sources used to dramatically increase the resolution of this model beneath North America and the surrounding area. Moreover, by using the core-diffracted seismic waves recorded by the USArray stations and other global seismic networks, we were able to sharpen the imaged structures in the lower third of the mantle (~2000-2800 km depth).


Second Place:

 

2nd with 838 votes: Chengping Chai - Pennsylvania State University

"Receiver-function wavefield in the contiguous United States"

A major part of our data is from the EarthScope USArray. Our project is funded by the U.S. National Science Foundation (grants EAR-1053484 and EAR-1053363) and the U.S. DOE (grant LDRD-20120047ER).


Third Place:

 

3rd with 181 votes: Scott Burdick - University of Maryland

Like a CAT Scan, seismic tomography lets us slice through the Earth without cutting into it.  We find slices of “seismic velocity,” or the speed with which waves from earthquakes travel through the Earth. The seismic velocity is affected by many properties of the rocks making up the Earth’s mantle like mineral composition, melt, and fluid content, but the much of the variation in velocity is caused by temperature. Waves travel faster through rocks that are colder and stiffer.  For this reason, images of seismic velocity usually use blue where waves travel faster and red where they travel slower. Seafloor is subducted at convergent plate boundaries when one tectonic plate is pushed beneath another, giving rise to natural hazards like earthquakes and volcanoes. The subducted seafloor is cold and strong, so it shows up in tomography as faster than average.  The circle of convergent boundaries surrounding the Pacific Ocean is known as The Ring of Fire. The model in the video is created using over 13 million earthquake recordings made at seismometers from the USArray, the seismological component of EarthScope, and global seismic networks.


Other Submissions:

 

Kyle Murray - Cornell University

A model of the change in elevation of the earth’s surface from late 2002 to 2010 in a ~ 100x100 km area in the southern San Joaquin Valley of California. Large subsidence signals with rates reaching over 25 cm/yr over this time span are a result of extraction of hydrocarbons in the south and groundwater in the north. Periodic variations each year are a result of seasonal recharge of water to aquifers in wetter months - slightly offsetting the groundwater extraction. These seasonal variations are not seen in areas where hydrocarbons are being extracted such as the Lost Hills Oil Field in the south. Subsidence rates have increased in recent years. Data is from Earthscope’s archive of synthetic aperture radar (SAR) data from the Envisat satellite.


Kimberly Schmid - NOAA’s National Geodetic Survey

This video animation illustrates the apparent motion of select stations in the Continuously Operating Reference Systems (CORS) network, across a prototype NATRF 2022 reference frame. The plotted points are derived from weekly time series data from each station.


Xiaotao Yang - University of Massachusetts Amherst

EarthScope OIINK Plane-Wave Migration image volume using teleseismic P-wave receiver functions. Animation illustrating three-dimensional geometry of Moho surface estimated from EarthScope OIINK (Ozark, Illinois, Indiana, Kentucky) data.

This animation uses two independent views of the 3D scene. The upper view has a moving camera and the lower view has a fixed viewpoint from above to produce an approximate map projection throughout the full animation. White curves illustrate state boundary and coastline data in three-dimensions. At the beginning of the animation both views show topography from ETOPO1 (https://www.ngdc.noaa.gov/mgg/global/global.html last accessed February 7, 2017) colored by elevation.

The locations of EarthScope USArray Transportable Array (TA) seismic stations are illustrated with magenta spheres and the locations of OIINK stations are illustrated with smaller, green spheres. Within the first 1 s of the animation the topography vanishes to reveal underlying surfaces. In the upper view, we display the Moho surface estimated from the regional Central United States (CUS) data as a white surface with illumination creating shading. A grid of small yellow spheres is used to illustrate the higher resolution surface estimated from OIINK data. The lower frame displays the Great Unconformity surface described by Marshak et al. [in press] color coded by elevation. The same lighting as the upper frame also highlights steeper slopes. The seismic station spheres disappear later in the animation and the Moho surface from CUS data disappears from the upper view while the camera swings to a view from the south and slightly above Earth’s surface 7 s into the animation.  

During this time, the Great Unconformity surface in the lower view fades to translucent to provide a view of the underlying Moho estimated from the OIINK data. The remainder of the animation slices through the 3D OIINK image volume. The lower view provides a map perspective of where the section in the upper view is located. In the upper view, we clip the Moho surface with a plane parallel to the image slice plane but offset 20 km southward. This makes the Moho surface appear like a dashed line in the animation. We use red to blue color map to display the image volume with red showing positive amplitudes and blue showing negatives.


Estelle Chaussard - University of Buffalo 

Surface and deep junction of the Hayward and Calaveras Faults in the East San Francisco Bay Area. The InSAR mean velocity and gradient maps reveal the faults surface traces (black lines), while repeating earthquakes (color dots) highlight the geometry of the fault junction at depth.


Kristine Larson - University of Colorado

"Measuring snow in the western U.S. using GPS data from the EarthScope Plate Boundary Observatory"

Top: Weekly photograph for PBO site P360 in Island Park, Idaho; Bottom: Accumulated snow at P360 calculated for the same time period using reflected GPS signals from the PBO H2O group. This technique is now used to measure snow accumulation every day at ~180 PBO sites.
More information about PBO H2O can be found at http://xenon.colorado.edu/portal