The Eastern Tennessee Seismic Zone (ETSZ) is an intraplate continental region located in the eastern part of North America and constitutes, after the New Madrid Seismic Zone, the second most active region of the continent east of the Rocky Mountains. It consists of a NE-trending, 300 km long by 100 km wide, belt of diffuse seismicity and is characterized by relatively low magnitude earthquakes occurring at mid-crustal depths. As a consequence, the rupture never propagates up to the ground surface and no obvious relationship seems to exist between the earthquake distribution and the faults known from geological mapping.
Our goal is thus to use indirect, geophysical observations to constrain the structure of the upper to mid-crust, with particular focus on the Precambrian basement, in order to identify any existing relationship between that structure and the seismic activity of the region. The geophysical available observations used in our work consist of seismological data obtained from a local network as well as of potential field data (magnetic and gravity data). This work is a collaborative study involving CERI (Memphis, TN) and NCCU (Durham, NC). Seismological data inversion (earthquake location and tomography) was achieved at CERI, while potential field data analysis was performed at NCCU.
Earthquake location and seismic tomography
Arrival time data from a local network of 433 stations were used to locate 1164 earthquakes in the ETSZ and to invert for the body-wave velocity field in that region. The hypocenters are mostly distributed between 5 and 20 km depth, with a maximum of the distribution located around 11km, which may reflect the depth of the brittle-ductile transition. As previously observed, the epicenters are distributed along a NNE-trending band and the epicenters located SE of the New York-Alabama Lineament (NYAL), a major linear magnetic feature of the region, appear more densely distributed than the ones located NW of that lineament (figure 1).
P- and S-wave velocity fields were inverted for by means of an iterative linearized procedure. The forward model was calculated from a Podvin-Lecomte finite-difference algorithm on a 12km x 12km x 4km grid. The results are shown for each 4km-thick slice for Vp, Vs, Vp/Vs and the seismic parameter Ф (figure 2). On the Vp map, the tomographic inversion shows the existence, at depths greater than 8 km, of a high velocity body to the West and of a low velocity body to the East, between which the earthquakes tend to occur. The NYAL, along which the earthquakes align, seems furthermore to correspond to a high Vp and Vs velocity gradient, with high velocities SE of that lineament and low velocities on its northwestern side. But the most striking result is the apparent affinity of earthquakes with low Vp/Vs as well as low Ф values. These results support the hypothesis according to which the earthquake distribution in that region is likely to reflect the geological structure at depth.
Fig. 1 Comparison of earthquake distribution and potential field maps in the Eastern Tennessee Seismic Zone.
Euler deconvolution and magnetic basement depth estimation
Euler deconvolution is a quantitative interpretation method for magnetic anomalies. It allows to invert for the position and depth of the magnetic sources by solving Euler’s homogeneity equation :
where T is the measured total field, x0, y0, z0 the spatial coordinates of the source, x, y, z the coordinates of the observation, B a regional value of the total magnetic field and N is a structural index (0-3) which reflects the geometry of the source. This method was applied to the total magnetic intensity field in the region of the ETSZ (figure 3).
Fig. 3 Total magnetic intensity field anomaly map (top) and magnetic anomaly source map (bottom) in the ETSZ. Red triangles represent earthquake epicenters located in this study.
On figure 4, we present NW-SE cross-sections in the 3D fields obtained from the tomographic inversion. We also plot the earthquakes located in this work as well as the location of the magnetic sources obtained from Euler deconvolution for several structural indices. The location of the earthquakes appears to coincide with an inflexion of the basement on Vp and Vs cross-sections. Such an inflexion can also be observed on the topography of the basement inferred from the magnetic data.
Fig. 4 NW-SE Cross-sections in Vp, Vs, Vp/Vs and Ф fields, and distribution of earthquakes (white dots) and magnetic sources as inferred from Euler deconvolution for different structural indices ( black dots).
Conclusions and perspectives
Earthquake distribution in the ETSZ appears to be significantly correlated with seismic velocity variations inferred from body-wave tomography and with observed potential field anomalies, thus suggesting some control of the seismicity by geological contrasts at depth. We applied Euler deconvolution technique to total magnetic field intensity data. The preliminary results provide magnetic basement depth estimates which can be compared to the variations of the velocity field inferred from tomographic inversion. These results, as well as information compiled from surface geology, available seismic profiles and well-log data, will ultimately be integrated into a geospatially referenced database and will be used to constrain structural models of the region. Compilation of such data is in progress.