IMPLICATIONS OF GROUND PENETRATING RADAR SURVEYS IN ERODED VALLEY FILL AT PROVIDENCE CANYON STATE PARK

Ground Penetrating Radar (GPR) data collected with a 100 MHz Pulse Ekko Pro system along the axis and across an aggrading valley-floor characterize the thickness and internal architecture of fill derived from several large (»60m deep) actively eroding gullies at Providence Canyon State Park in southwest Georgia. Sediments liberated by incision and headward erosion of host Cretaceous shallow marine (Ripley Fm.), intertidal sands (Providence Fm.), and Tertiary iron-rich residuum (Clayton Fm.) have accumulated as reworked alluvial (braided stream) and colluvial deposits that exceed 6 m in thickness. Vibracore and valley-side exposures provide abundant control for radar interpretations.

Long- and cross-valley GPR transects together with detailed 3D surveys at up- and down-valley sites reveal 4 distinct radar facies. Facies 1 and 2 are associated with in situ Cretaceous Providence Fm. (at up-valley locations only) and Ripley Fm. deposits respectively. Discontinuous, low amplitude, and hummocky reflectors (F1) and subparallel (F2) reflectors are highly attenuated, surround, and are truncated by channel fill. Radar facies within this fill consist of intermediate amplitude reflectors that comprise contorted to channelized complex fill (F3) associated with modern braided stream sediments. These deposits overly high amplitude parallel reflectors (F4) which vibracore suggest are finer-grained flood plain deposits that pre-date land-clearing thought to be responsible for initiating the canyons. 3D radar visualizations indicate that valley fill is more readily defined at down-valley locations due to the contrast between sandy braided stream fill and underlying clay-rich Ripley Fm.

Radar data indicate that the volume of valley fill within 1 km of eroding headwalls increases down-valley as valley sides become less confined, braided streams merge, and channel fill depth and width increases. Nonetheless, the total volume of fill is clearly far less than eroded space in the dissected canyons. This together with flood plain deposits beneath braided stream fill suggests: (1) that land clearing in the 1800’s accelerated but did not initiate canyon formation, and (2) that more sediment has been transported down-valley than is stored locally near the eroding canyons.