The New Madrid seismic zone (NMSZ) in the central United States is the most seismically active intraplate region in North America. It includes two SW-NE-trending zones of right-lateral strike-slip faulting on subvertical faults and a zone of thrust faulting on 30 SW-dipping plane at a left step-over between the strike-slip fault zones (Figure 30.1). Three widely felt magnitude 7-8 earthquakes occurred in the 250 km zone in the the winter of 1811-1812, and the central thrust zone was sufficiently displaced during the 7 February 1812 event to create a waterfall on the Mississippi river.
The NMSZ is located within a failed rift that was active about 600 million years ago, followed by period of magmatic reactivation and igneous intrusion of mafic plutons 80 to 60 million ago. These episodes of activity introduced heterogeneities into the crust that may act as stress concentrators for the late Holocene seismicity, possibly initiated by the most recent deglaciation event (Grollimund and Zoback, 2001). Paleoseismic evidence indicates that 1811-1812 sized events have occurred throughout the late Holocene, most recently around 1450 and 900 A.D., but the small cumulative fault offsets inferred from seismic reflection data suggest that the current high level of seismic activity initiated recently, and there is scant paleoseismic evidence for more than 4 episodes prior to the historic events.
Low rates of strain and the lack of apparent active surface tectonics suggest that the central and eastern U.S. are within the stable interior of the North America plate. Geodetic studies of broadscale deformation within this region generally find that relative station velocities are consistent ( mm/yr) with a rigid plate with strain rates not significantly differing from zero. High strain rates (100 nanostrain/yr) were reported within a network spanning the southern NMSZ based on a 1991 GPS survey and triangulation data collected in the 1950's (Liu et al., 1992). We present new strain rate estimates within this southern NMSZ network from GPS surveys conducted in 1993 and 1997 that show the recent deformation rates are not significantly greater than zero.
The primary network consists of triangulation benchmarks installed in the mid-1950's that were reoccupied using GPS in 1991 by NGS. Enough of the benchmarks had survived to allow 40-year averaged shear-strain rates to be estimated. This network was reoccupied by GPS in 1993 and 1997 and stations were added to better span the southern seismic zone and the western rift boundary (Figure 30.1).
We processed the data from the 3 GPS surveys using GAMIT/GLOBK software using many of the same techniques used to process the BARD observations (Murray and Segall, 2001). Definition of a self-consistent reference frame is complicated by major changes to the global fiducial network that occurred between the 1991 and 1993 surveys. To better define the velocity frame throughout this period, we included SOPAC global solutions obtained for 45 days, at 64-day intervals, between March 1991 and December 1998. These solutions provide sufficient observations to estimate the positions and velocities of the fiducial stations included in our NMSZ analysis during each interval between significant changes of the station equipment and reference monuments. We defined a North America frame by minimizing the horizontal velocities of 15 stations, which had an rms deviation of 3.5 mm in position and 0.5 mm/yr in velocity.
The velocities of 32 sites in the NMSZ region are well determined with respect to the North America frame (Figure 30.2). The average horizontal velocity of the 32 stations relative to North America is mm/yr at NE, which significantly differs from zero (all quoted uncertainties are 95% confidence). The apparent average northward motion is due primarily to stations located in the interior of the network, whereas the average motion of the outlying stations (BLUF, HOPE, GP47, BROA, GP17) is mm/yr, NE, consistent with their being on stable North America.
We estimated strain rates using 1950's triangulation data to compare with the Liu et al. (1992) results. Because triangulation data are relatively insensitive to distances and the scale of the network, we estimated engineering shear strain rates, and , from which the maximum shear-strain rate and direction of maximum contraction can be derived. Liu et al. (1992) found that shear-strain rates in a 22-station network (Figure 30.1) were significantly greater than zero, particularly in the western half of the network that spans the rift boundary, with nanoradian/yr. Using the same stations, but including GPS data from all 3 years, we find the estimated shear strain rates in all cases do not significantly differ from zero. For example, using all or just the western stations is or nanoradian/yr, respectively. Therefore, in contrast to the 1992 study, we find no evidence for high strain rates in the southern New Madrid seismic zone.
The average motion of the interior stations relative to the outlying stations is mm/yr, NE, with the most central stations tending to have velocities slightly elevated above this level. The spatial coherence of this pattern suggests that some deformation--albeit marginally significant--may be present in the region. Projecting the average motion onto the NE seismic trend yields 0.7 parallel and -0.6 mm/yr perpendicular components, which is opposite of that predicted by simple elastic strain dislocation models that assume zero far-field deformation and backslip on the faults defined by seismic and paleoseismic studies. We are currently investigating alternative explanations for this tantalizing signal, such as from a gravitational instability of the underlying rift pillow (Pollitz et al., 2002), or from the relaxation of a weakened lower crustal zone proposed by Kenner and Segall (2000). This latter model, which predicts low rates of strain consistent with our geodetic observations, shows that low strain rates do not necessarily preclude the possibility of repeating large intraplate earthquakes, and that the seismic hazards in the NMSZ are still likely to be high.
We appreciate previous support for this project by the NSF. We thank Stacy Kerkela, Stephanie Prejean, and the staff and students of CERI in Memphis, Tennessee for assistance conducting the 1997 survey.
Grollimund, B., and M. D. Zoback, Did deglaciation trigger intraplate seismicity in the New Madrid seismic zone?, Geology, 29, 175-178, 2001.
Kenner, S. J., and P. Segall, A mechanical model for intraplate earthquakes: Application to the New Madrid seismic zone, Science, 289, 2329-2332, 2000.
Liu, L. B., M. D. Zoback, and P. Segall, Rapid intraplate strain accumulation in the New Madrid seismic zone, Science, 257, 1666-1669, 1992.
Murray, M. H., and P. Segall, Modeling broadscale deformation in northern California and Nevada from plate motions and elastic strain accumulation, Geophys. Res. Lett., 28, 4315-4318, 2001.
Pollitz, F. F., L. Kellogg, and R. Bürgmann, Sinking mafic body in a reactivated lower crust: A mechanism for stress concentration in the New Madrid Seismic Zone, Bull. Seismol. Soc. Am., 91, 1882-1897, 2002.
Berkeley Seismological Laboratory
215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760
Questions or comments? Send e-mail: firstname.lastname@example.org
© 2004, The Regents of the University of California