NS21B-3881:
3D GPR Modeling and Imaging of Burials: Mueschke Historic Cemetery, Houston, Texas

Tuesday, 16 December 2014
Azizuddin Abdul Aziz, Robert Stewart and Susan L Green, University of Houston, Houston, TX, United States
Abstract:
3D ground-penetrating radar (GPR) surveys consisting of 6 grids were conducted from October 2013 until April 2014 to delineate burials at the historic Mueschke Cemetery in Houston, Texas. The surveys were primarily to assist historians and archeologists from the Mueschke Cemetery Association and Lone Star College in locating some 13 postulated unmarked burials. Antenna with three frequencies were used: 100 MHz, 250 MHz and 1000 MHz. Most surveys were conducted with the 250 MHz Sensors and Software NOGGIN System which has a maximum penetration of about 3 m. Three methods were used to estimate the soil velocity for time-to-depth conversion: Common mid-point (CMP) surveys, time-to-known depth matching, and hyperbola fitting. All three methods gave an average velocity of 0.06 m/ns in the upper 2 m of the soil. The time-to-known depth method was accomplished by digging a trench (1.5 m deep by 1.5 m wide by 3 m long) about 10 m from the cemetery entrance using a back-hoe. Rebar was hammered horizontally into the trench wall at 0.25 m increments from depths of 1.5 m to 0.25 m. The excavation also allowed us to observe soil strata which transitions from loam, to silty clay to mostly clay with increasing depth. We used finite-difference, time-domain computer modeling to synthesize the response of two types of burials: vaulted or concrete enclosed (post-1940) and non-vaulted or casket only (pre-1940). Modeling results indicate that vaulted burials have a flattened apex signature while non–vaulted burials have signatures that are more hyperbolic. The data were processed using gain, dewow, background removal, filtering, and migration. Survey data over known burials show distinct diffractions and a rectangular shape–correspondent to the computer modeling results. Burials before 1940 have weaker diffractions which complicates their detection. Tree roots, clay patches, and rocks can also present anomalies that must be carefully investigated. Nonetheless, several strong burial-type anomalies have been identified. High amplitude areas are especially evident in depth slices from the 3D data volumes and indicate the locations of these suspected burials. GPR surveys and images have been successfully used at the historic Mueschke Cemetery to template known burials and provide leads on the locations of a number of unmarked burials.