On July 25, 1946 the United States detonated the first underwater nuclear weapon in history - code name "Baker" - at the Bikini Atoll. The explosion generated a gas bubble that pushed against the water, generating a supersonic shock wave which crushed the hulls of nearby target ships as it spread out. Seismic waves of this test were observed at seismograph stations around the globe and it was realized that these waves could be used to detect and potentially characterize a nuclear explosion.
Fig.1. Photography of the underwater "Baker" nuclear explosion of July 25, 1946 showing the white sphere of water and vapour formed by the second shock wave of the explosion (image in public domain).
The U.S. performed also the first fully underground explosion - code name "Ranier" - that was detected by about 50 seismic stations; however, it was confused in part with a "normal" earthquake.
With the ban of nuclear weapon (well, sort of...) testing in the year 1958 it became necessary to install an effective worldwide monitoring system. Three years later the set-up of the WorldWide Standardized Seismographic Network (WWSSN) began and in 1966 almost 112 stations were working in the monitoring project "Vela". Vela provided a large quantity of supplementary seismic data used to answer three questions: Where is the seismic event located? What is the source type (artificial or natural) of the event? How large is the event?
“It appears increasingly doubtful that an atomic-weapons test of significant dimension can be concealed either underground or in outer space. A five-kiloton nuclear explosion in an underground salt cavern near Carlsbad, N.M., in December was clearly recorded by seismographs as far away as Tokyo, New York, Uppsala in Sweden and Sodankyla in Finland. The seismograph records included tracings of the ‘first motion,’ considered critical in distinguishing between earthquakes and underground explosions.”
The signature of a natural earthquake shows a distinct pattern: a seismometer will first detect the Primary and Secondary Waves, followed by the more destructive Surface or Rayleigh Waves.
Seismic P Waves are compressional waves, similar to sound waves in the air. Secondary or Shear (S) Waves are transverse waves, like those that propagate along a rope. A sudden explosion generates a "sphere" of compressional waves travelling in all directions. In contrast an earthquake is caused by the sliding of rocks along a fracture and it will generate shear waves concentrated in a certain direction. Therefore an explosion will show a strong and sudden signal of P-waves, with a similar signal recorded by all the seismometers collocated around the explosion. An earthquake will show a more complex pattern, depending of the position of the seismometer, characterized by strong S-Waves and R-Waves.
Also an underground explosion does not generate very strong surface waves as a natural earthquake does.
Fig.2. Schematic seismogram with Primary (P; compressional waves), Secondary (S; shear waves), and Rayleigh (R; surface waves) phases for an artificial blast and a natural earthquake.
As every atomic explosion will generate a unique pattern, distinct from natural earthquakes, seismology is a reliable tool to control the ban of nuclear test and to supervise countries that still test atomic weapons.
The information recovered from seismograms of nuclear blasts can be applied in forensic seismology also to study detonations of common explosives. Most spectacular cases in the last years comprise the reconstruction of the Oklahoma City bombing in 1995 (see this abstract by HOLZER at the AGU meeting in 2002) and the investigation in the explosion on the Russian submarine "Kursk" in 2000 (see KOPER et al. 2001; the blog "About.com Geology" hosts many other examples).
Seismic waves can be generated not only by shear movements along faults or by the expansion of plasma (nuclear device) or gas (conventional device) during an explosion, but also by the impact of objects with the ground.
Seismic signals were already used to identify the location of rock-falls and recent research suggests that the signals can help to characterize the dynamics and volume of a landslide, Dave Petley discusses the significance and use of seismograms in various posts published on his "Landslide blog".
The analysis of seismic waves provided also insights on what happened September 11, 2001 in New York. Seismograph stations around the city recorded the signals generated by the aircraft impacts and the subsequent collapse of the two towers of the World Trade Center (the Lamont-Doherty Cooperative Seismographic Network provides a rich collection of datasets of the seismic activity around N.Y.). The collapse of the south tower generated a signal with a magnitude of 2.1 and the collapse of the north tower, whit a signal of magnitude 2.3, was recorded by 13 stations ranging in distance from 34 to 428km.
Also these seismograms show a distinct pattern if compared to the pattern caused by a natural earthquake. There are no P or S Waves, but the impacts of the buildings on the ground generated a sudden peak of short-period Rayleigh Waves.
Fig.3. Seismic recordings at the seismograph station Palisades (N.Y.) for events at World Trade Center on September 11, distance of station from Ground Zero ~ 34km. Note that impact 1 and collapse 2 relate to the north tower, and impact 2 and collapse 1 apply to the south tower. Expanded views of the first impact and first collapse shown in red. Figure from KIM et al. 2001, published here according to the Usage Permissions granted by AGU & authors.
The seismograms show also that the impact and explosion of the two airplanes generated a relative small amount of seismic energy. This confirms the observation that the collapses of the two towers were not a direct result of the impacts, but caused by the weakening of the supporting structures of the buildings due the subsequent fires.
Most energy of the collapses was dispersed into the deformation of the buildings and the formation of rubble and dust, only a small portion of potential energy was converted into seismic waves. The generated 2.1 and 2.3 earthquakes were too weak to destabilize nearby buildings, most damage was done by the kinetic energy of the debris and the displaced air.
Also the collision of the cruise ship "Costa Concordia" on January 13, 2012 was recorded by the seismograph station "Monte Argentario", situated on the Italian mainland. From the eyewitness testimony and the Automatic System of the ship the time of collision with a submerged rock was estimated at 20:45 (UTC). This time is confirmed by a sudden peak in the seismogram at 20:45:10 (the seismograph station is distant 18km from the site of the collision, the seismic waves needed almost 3-4 seconds to travel this distance). The seismogram shows also after the impact the "noise" generated by the hull of the ship grinding along the rocky substrate.
Fig.4. Seismogram recorded at the station "Monte Argentario" (Italy) showing the seismic waves generated by the impact of the "Costa Concordia" on January 13, 2012 20:45 (UTC). An accurate analysis of "The seismic wake of "Costa Concordia" (23.01.2012) can even specify the speed of the ship at the moment of the collision.
Figure used with permission and taken from the post "The earthquake of the Costa Concordia" by Italian seismologist Marco Mucciarelli, published on January 21, 2012 on his blog "terremoti, sismologia ed altre sciocchezze".
ANDERSON, D.N.; RANDALL, G.E.; WHITAKER, R.W.; ARROWSMITH, S.J.; ARROWSMITH, M.D.; FAGAN, D.K.; TAYLOR, S.R.; SELBY, N.D.; SCHULT, F.R.; KRAFT, G.D. & WALTER, W.R. (2010): Seismic event identification. WIREs Computational Statistics Vol.2, July/August: 414-432
KIM, W.-Y.; SYKES, L.R.; ARMITAGE, J.H.; XIE, J.K.; JACOB, K.H.; RICHARDS, P.G.; WEST, M.; WALDHAUSER, F.; ARMBRUSTER, J.; SEEBER, L.; DU, W.X. & LERNER-LAM, A. (2001): Seismic Waves Generated by Aircraft Impacts and Building Collapses at World Trade Center, New York City. EOS Vol.82 (47)
KOPER, K.D.; WALLACE, T.C.; TAYOLR, S.R. & HARTSE, H.E. (2001): Forensic seismology and the sinking of the Kursk. EOS, Vol.82 (4): 37