SIGNAL POLARITY TRENDS IN GEORADAR IMAGES OF MAMMAL BURROW NETWORKS IN A HETEROGENEOUS MOUNTAIN SLOPE REGOLITH

This study presents a simple and effective approach for characterizing shallow subsurface anomalies in ground-penetrating radar (GPR) images based on a dense grid of high-resolution (750 MHz) surveys of pocket gopher burrow networks along a soil-covered mountain slope, Oracle Ridge, southern Arizona. Within the upper 0.5-1.2 m of unsaturated sandy soil, hyperbolic anomalies with polarity inversions (relative to the ground wave) were interpreted as open burrows (air-filled or partially filled with vegetation), whereas normal polarities represented large tree roots or shallow rock clasts. Background grayscale depth profiles (mostly black and white end members) were established adjacent to hyperbolic anomalies, with residual values of the two traces indicating the direction and degree of overlap in polarity shifts. Typically, the first partial negative offset in grayscale (polarity reversal relative to ground reflection) represents a subsurface extension of an unfilled burrow. The depth range of this signature varies between 6-12 cm in profile segments adjacent to burrow entrances, with tunnel termini extending below 20 cm. Truncations of subsurface reflections were used to identify the extent of soil disturbance (e.g., inclined tunnels) and to calculate apparent dips (8-10 ̊). In addition, grayscale profiles allow determination of regolith thickness due to a strong reflection from bedrock surface and low-amplitude (mid-scale) response below this contact. Shallow excavations and burrow casts were used to groundtruth a number of shallow anomalies and demonstrate good correlation between observed and imaged subsurface burrow extensions. Where geophysical transects traverse spoil mounds, a characteristic concave reflection on raw radargrams allows determination of cross-sectional areas and serves as a guide for surface normalization. Whereas a number of post-processing techniques exist for diffraction analysis in GPR images, our approach indicates that grayscale profiles of digital and scanned analog records allow rapid characterization and quantification of key attributes in bioturbated heterogeneous substrates.