The terrorist bombing of the Alfred P. Murrah Federal Building in Oklahoma City on April 19, 1995, generated seismic waves that were recorded on two permanent seismographs about 7 and 26 km away from the bombing. The seismogram recorded at 26 km shows two low-frequency wave trains, discrete sets of oscillatory signals, that begin about 10 s apart. Public release of this record prompted speculation that each wave train was caused by a different energy source. On May 23, 1995, the U.S. Geological Survey monitored the demolition of the bomb-ravaged Federal Building with portable seismographs (Figure 1). Two wave trains were picked up again. The recordings indicate that the wave trains during both the bombing and demolition represent seismic waves traveling at different velocities. We conclude that the two wave trains recorded during the bombing are consistent with a single impulsive energy source.
Militia groups were among the early proponents that the two wave trains were caused by separate explosions and hinted that the "government orchestrated the tragedy" at the Federal Building [Kovaleski, 1995] . However, the scientific community was also uncertain about the significance of the two wave trains. The Oklahoma Geological Survey noted in an April 26 press release that "The location and source of the second surface wave recording was unknown. Detailed investigations at the building site may offer an explanation of the cause and origin of the second event." The AAPG Explorer  reported that the "first event was caused by energy from the explosion and the second from the fall of the building." This uncertainty partially arose from timing problems; the precise time of the bombing was unknown and the seismogram from the permanent seismograph 7 km from the bombing only had relative time.
To help understand the April 19 seismograms, the controlled demolition of the building on May 23, was recorded with an array of four portable digital seismographs (Figure 2). Analysis of these data indicates that the two wave trains are distinct seismic phases that traveled in a near-surface zone over which seismic velocity increases significantly with depth. The first phase appears to be a packet of scattered seismic shear waves, and the second is the fundamental-mode Rayleigh wave. The two wave trains recorded on April 19 are consistent with a single explosion, not evidence of two explosions 10 s apart. The investigation also permits an accurate estimation of the time of detonation between 9:01:52 and 9:01:57 a.m. local time (LT).
Two permanent seismographs recorded the bombing. FNO, which is northeast of Norman near Franklin, is 26.02 km south of the Federal Building (Figure 2). It is operated by the Oklahoma Geological Survey. Although the FNO seismogram fueled much of the controversy, the bombing also was recorded by another seismograph, OMN, which is operated by the Omniplex Science Museum for public display purposes (Figure 2). This seismograph is 6.86 km northeast of the Federal Building. Both seismographs consist of conventional vertical-component seismometers that record ground velocity at frequencies ranging from about 1 15 Hz.
The FNO seismogram is shown in Figures 3a and 4a. Although public attention focused on the two low-frequency wave trains (see wave trains 2 and 3 in Figure 4a), they were preceded by a high-frequency wave train (see wave train 1 in Figure 4a). Little attention was paid to the high-frequency wave train because the seismograph operator originally attributed it to local traffic. We can show that wave train 1 arrived after the bomb detonated and most likely consists of P and S waves generated by the bombing.
The OMN seismogram of the bombing (Figure 3a) is on scale for about 2 s and then increases and goes off scale twice for 10-s and 5-s intervals. The off-scale intervals begin about 16 s apart. We believe that the beginning of the seismic signal marks the arrival of the P wave, the first off-scale interval is the S wave, and the second off-scale interval is the air wave from the blast, which was heard clearly at Omniplex Science Museum. Due to a power failure the previous night, the OMN seismograph failed to accurately record the time of the bombing.
On May 23, 1995, the remainder of the Federal Building was razed (Figure 1). The building collapsed inward as approximately 68 kg of dynamite were ignited in the structure. The demolition was monitored by portable 6-channel digital seismographs [Borcherdt et al., 1985] placed next to the two permanent seismographs as well as 56 m from the building and at a site 12.04 km southeast of the building near Moore, which is referred to as MOR (Figure 2). Three-component velocity sensors were used at each site, and a three-component accelerometer was deployed at the Federal Building. Recorders were synchronized to an external clock. The deployment, instrumentation, and data are described by Dietel  .
The vertical component of ground velocity recorded by the portable seismographs during the demolition is shown in Figure 3b; digital seismograms were filtered with a low-pass 2-Hz filter because of high-frequency ambient noise, particularly at the three distant sites. The duration of the seismic signal increases as seismic waves propagate outward from the Federal Building, as is common to all seismograms as the numbers and lengths of ray paths increase. The seismograms also indicate that the seismic signal from the demolition changes from a single wave train to two wave trains as the seismic energy propagates outward to FNO (see wave trains 2 and 3 in Figure 3b). The velocities of wave trains 2 and 3, as inferred from their travel time between MOR and FNO, are about 1.7 and 1.0 km/s, respectively.
The permanent seismographs, FNO and OMN, also recorded the demolition (Figure 3c). The FNO seismogram again includes two wave trains about 10 s apart that are comparable in amplitude and frequency to the April 19 low-frequency wave trains (see wave trains 2 and 3 in Figure 4a). The OMN demolition seismogram went off scale only once, for 7 s beginning about 2 s after the arrival of seismic energy from the demolition.
Even without identifying the phases of the two low-frequency wave trains recorded about 10 s apart at FNO after the bombing and demolition, both their similarity (Figure 4a) and development as seismic energy propagates outward from the Federal Building (Figure 3b) indicate that they are associated with the propagation of seismic waves. They cannot be caused by separate seismic energy sources because the entire demolition lasted only 8 s.
In addition, even though the demolition lasted 8 s, analysis of the unfiltered seismogram recorded at the Federal Building indicates that most of the seismic energy generated by the demolition was released about 5 s after it started. Comparison of timed video coverage of the demolition with the seismogram indicates that explosives were completely detonated about 2.5 s into the demolition and did not generate as much seismic energy as the collapse of the building.
Because each type of seismic wave has a distinctive particle motion and velocity at which it propagates, the digital recordings can be used to help identify the wave types in the wave trains. The particle motion of the second low-frequency wave at FNO (wave train 3 in Figure 3b) is vertically polarized and retrograde elliptical, which is characteristic of seismic waves known as Rayleigh waves. Wave train 3 also shows some dispersion from longer to shorter periods, which is characteristic of Rayleigh waves. On the basis of the particle motion and the 1.0 km/s phase velocity, we interpret this phase to be a fundamental-mode Rayleigh wave. Particle motion of the first low-frequency wave train at FNO (wave train 2 in Figure 3b) has a less coherent pattern but has some vertical elliptical motion suggestive of Rayleigh waves. This pattern and the 1.7 km/s phase velocity suggest this wave train may be in whole or part a higher-mode Rayleigh wave. Alternatively, the less coherent pattern of particle motion indicates that this phase more likely consists of S waves that are scattered in a near-surface zone.
To confirm that this interpretation is consistent with what is known of the local seismic velocity structure, synthetic seismograms were computed for the May 23 demolition (Figure 4b) using the seismogram recorded near the Federal Building as the source-time function. A one-dimensional velocity structure was constructed based on downhole velocity surveys in local oil wells. Seismograms were computed using a numerical code described by Wang and Herrmann  and Herrmann and Wang  . To match the synthetic with the observed FNO seismogram required modification of the S wave velocity in the upper 350 m. In the final model, the S wave velocity in the shallow zone increases from about 0.3 1.3 km/s with depth. The inferred S wave structure is consistent with regional studies by Lee and McMechan  . To improve the match between amplitudes of the waveforms, we assumed an intrinsic attenuation, Q, of 50.
The time of the bombing on April 19, which is not accurately known, can be estimated from the time it took for the two low-frequency wave trains (wave trains 2 and 3) to travel from the Federal Building to FNO after the May 23 demolition. This is because the bombing- and demolition-generated wave trains appear to be the same wave phases (Figure 4a). On May 23, we measured a travel time of 30 s from the beginning of the demolition to the arrival of wave train 3. Applying this travel time to the April 19 FNO seismogram, which had accurate timing, indicates the bomb detonated at 9:01:52 a.m. LT.
The time 9:01:52 a.m. is probably the earliest that the bomb could have detonated. The exact time of the bombing is uncertain because of ambiguities about which part of the potential seismic sources on April 19 and May 23 generated the wave trains. As noted previously, most of the seismic energy from the demolition was generated about 5 s after the demolition started. If this pulse generated the low-frequency wave trains, then the May 23 travel time for wave train 3 is 25 s. The source of seismic energy on April 19 is ambiguous because no seismograph was next to the Federal Building, and both the bomb and building collapse are potential seismic sources. It seems likely that the bomb was the primary seismic source, but the building collapse cannot be excluded as a partial source of seismic energy. The approximately 1820-kg bomb released about 28 times as much energy as the explosives used in the demolition and was probably better coupled with the ground. In addition, energetic body waves were present in the April 19 seismogram at FNO (wave train 1 in Figure 3a) that were not present on May 23.
On the preceding basis, we conclude that the bomb was detonated sometime between 9:01:52 and 9:01:57 a.m. LT. These times imply average velocities of 2.7 and 5.5 km/s, respectively, for wave train 1 at FNO. We prefer 9:01:57 a.m. as the time of the bombing because the 5.5 km/s velocity is consistent with a deeply refracted P wave. A generalized velocity model (R. L. Brown, personal communication, 1995) indicates that the tops of the Arbuckle limestone, which has a P wave velocity of 6.1 km/s, and crystalline bedrock, which has a P wave velocity of 6.2 km/s, are at depths of 2.7 and 3.6 km, respectively.
Finally, we offer the following interpretation of the seismic phases in the OMN seismograms of April 19 and May 23. If one assumes that the second off-scale wave train at OMN on April 19 is the air wave generated by the bombing explosion, which was clearly heard at OMN, one can estimate when the explosion occurred relative to the arrival of the air wave by dividing the 6.86-km distance from the bomb site to OMN by the speed of sound in air. Using a speed of 340 m/s, which is appropriate for the 365-m elevation of Oklahoma City, yields a travel time of 20.1 s. It is also consistent with a 20.3-s travel time for a large-amplitude, high-frequency, wave train on the unfiltered OMN digital recording on May 23 [Dietel, 1995] , which probably was caused by the explosives detonated for the demolition. The air-wave travel time implies that the explosion occurred about 2.5 s before seismic energy started arriving at OMN and 4.5 s before the first off-scale wave train. These travel times yield average velocities of 2.7 and 1.5 km/s, respectively, which are appropriate P and S wave velocities for refracted waves at a 6.86 km distance.
Thus we infer that the beginning of the OMN seismic signal associated with the April 19 bombing is a P wave and the first off-scale wave train is an S wave, both of which were generated by the explosion. Furthermore, we conclude that the first seismic arrival and the off-scale wave train at OMN from the demolition on May 23 are the same P and S wave phases, respectively, identified at the beginning of the April 19 seismogram because they also arrive 2 s apart.
We thank the following people for their contributions to the research and preparation of this article: R. L. Brown, G. Christenson, T. Miller, F. Brooks, E. Dingel, S. R. Walter, D. P. Hill, J. E. Vidale, and T. C. Hanks.
AAPG Explorer, Bomb blast waves recorded, OC tragedy shook the earth, Am. Assoc. Pet. Geol. Explorer, 16(6), 24, 1995.
Borcherdt, R. D., J. B. Fletcher, E. G. Jensen, G. L. Maxwell, J. R. Van Schaack, R. E. Warrick, E. Cranswick, M. J. S. Johnston, and R. McClearn, A general earthquake observation system (GEOS), Bull. Seismol. Soc. Am., 75, 1783, 1985.
Dietel, C. M., Digital seismic recordings of the May 23, 1995, demolition of the Alfred P. Murrah Federal Building, Oklahoma City, Oklahoma, U.S. Geol. Surv. Open File Rep. 95-594, 14 pp., 1995.
Herrmann, R. B., and C. Y. Wang, A comparison of synthetic seismograms, Bull. Seismol. Soc. Am., 75, 41, 1985.
Kovaleski, S. F., We're not wackos, San Jose Mercury News, May 2, 1995, Sec. B, p. 7, 1995.
Lee, H.-Z, and G. A. McMechan, Estimation of shear-velocity profiles by imaging of surface waves, Geophys. J. R. Astron. Soc., 90, 649, 1987.
Wang, C. Y., and R. B. Herrmann, A numerical study of P-, SV-, and SH-wave generation in a plane layered medium, Bull. Seismol. Soc. Am., 70, 1015, 1980.
Back to Science for Everyone