HR: 0800h
AN: U21B-0814 [Abstracts]
TI: Ground-Penatrating Radar Investigations Across the Sawmill Branch Fault Near Charleston, South
Carolina
AU: * Dura-Gomez, I
EM: idura@geol.sc.edu
AF: Department of Geological Sciences, University of South Carolina, 701 Sumter Street
, Columbia, SC 29208
United States
AU: Addison, A
EM: aaddison@geol.sc.edu
AF: Department of Geological Sciences, University of South Carolina, 701 Sumter Street
, Columbia, SC 29208
United States
AU: Knapp, C C
EM: camelia@geol.sc.edu
AF: Department of Geological Sciences, University of South Carolina, 701 Sumter Street
, Columbia, SC 29208
United States
AU: Talwani, P
EM: pradeep@sc.edu
AF: Department of Geological Sciences, University of South Carolina, 701 Sumter Street
, Columbia, SC 29208
United States
AU: Chapman, A
EM: achapman@scprt.com
AF: Colonial Dorchester State Historic Site, 300 State Park RD, Summerville, SC 29485
United States
AB:
During the 1886 Charleston earthquake, two parallel tabby walls of Fort Dorchester broke left-laterally, and a strike of
~N25§W was inferred for the causative Sawmill Branch fault. To better define this fault, which does not have any surface
expression, we planned to cut trenches across it. However, as Fort Dorchester is a protected archeological site, we were
required to locate the fault accurately away from the fort, before permission could be obtained to cut short trenches. The
present GPR investigations were planned as a preliminary step to determine locations for trenching.
A pulseEKKO 100 GPR was used to collect data along eight profiles (varying in length from 10 m to 30 m) that were run across
the projected strike of the fault, and one 50 m long profile that was run parallel to it. The locations of the profiles were
obtained using a total station. To capture the signature of the fault, sixteen common-offset (COS) lines were acquired by
using different antennas (50, 100 and 200 MHz) and stacking 64 times to increase the signal-to-noise ratio. The location of
trees and stumps were recorded. In addition, two common-midpoint (CMP) tests were carried out, and gave an average velocity
of about 0.097 m/ns. Processing included the subtraction of the low frequency "wow" on the trace (dewow), automatic gain
control (AGC) and the application of bandpass filters. The signals using the 50 MHz, 100 MHz and 200 MHz antennas were found
to penetrate up to about 30 meters, 20 meters and 12 meters respectively. Vertically offset reflectors and disruptions of
the electrical signal were used to infer the location of the fault(s). Comparisons of the locations of these disruptions on
various lines were used to infer the presence of a N30§W fault zone We plan to confirm these locations by cutting shallow
trenches.
DE: 0910 Data processing
DE: 0994 Instruments and techniques
DE: 8107 Continental neotectonics (8002)
SC: Union [U]
MN: Fall Meeting 2005