Interpreting Carolina Bays as Glacier Ice impacts

Antonio Zamora

Abstract
The origin of the Carolina Bays remains a mystery more than eighty years after their discovery. This paper explores the hypothesis that the Carolina Bays were formed as the result of impacts of glacial ice ejected when a meteorite struck the Laurentide ice sheet in the Great Lakes region. A new impact model describes the elliptical Carolina Bays as conic sections representing the intersection of the tilted conical craters formed by the oblique entry of the glacier chunks with the level surface of the Earth. Melting of the ice impactors created pools that became shallower through viscous relaxation of the substrata and deposits of terrestrial sediments. The impact model can be used to predict the location where stones carried by the glacial ice impactors may have been deposited. The high density of the impacts indicates a possible local extinction event, but this paper does not speculate on the time frame when this would have occurred. Photographs of the experimental tests of the model are presented.

Figure 1. Carolina Bays

History
The Carolina Bays are approximately 500,000 elliptical, shallow lakes, wetlands and depressions, up to 10 km long which are found east of the Rocky Mountains and are concentrated mainly along the Atlantic seaboard. They are also known as Delmarva Bays, Maryland Basins or Nebraska Rainwater basins. The major axes of the ellipses generally point toward the Great Lakes. The bays were discovered in the 1930's from the first aerial photographs of the Atlantic coast. The rims of the bays are usually only one meter higher than the centers, making the bays virtually undetectable from the ground. The use of LiDAR, a laser imaging system, has been used to discover a great number of bays by emphasizing the small differences in elevation of the surface features (Figure 1). A web site for viewing Carolina Bay LiDAR images is available from Davias[1].

Bay Formation Theories
Many theories have been formulated to explain the origin of the Carolina Bays. The theories can be classified into those that propose forces within the Earth, such as wind or water currents, and those that propose impacts from an astronomical event, like a meteor shower. The proposals about impacts from meteorites or comets (Melton 1933, Firestone 2007; Kennett 2009a,b) have been opposed due to the lack of physical evidence resulting from a hyperspeed impact such as shatter cones, crystals displaying shock metamorphism, meteorite fragments or the occurrence of siderophile elements. This type of evidence is deemed essential to distinguish extraterrestrial impact structures from land features created by volcanism, erosion or other terrestrial processes (French and Koeberl, 2010).

Since the bays do not contain meteorite fragments or mineral crystals with the planar deformation features characteristic of hypervelocity impacts, astronomers and geologists have rejected the theory that the Carolina Bays were the result of meteorite or comet impacts (Pinter, et al. 2011).

The theories of bay creation by terrestrial processes include substrate dissolution, wind, ice, marine waves and currents that reduce the volume of karst-like depressions which are later modified by wind or ice-push processes (May and Warne, 1999). The bays have also been characterized as thermokarst or thaw lakes that are circular or elliptical in shape and are often aligned with the prevailing wind (Melosh 2011).

In general, the terrestrial process models do not apply well to the Carolina Bays because they do not provide a mechanism for the formation of elliptical bays with raised rims whose axes are aligned to radiate from a common point.

Thermokarst characteristics
Figure 2 and Figure 3 illustrate thermokarst lakes in Alaska and Russia. These lakes form when underground ice melts in a region underlain by permafrost and the ground collapses like a sinkhole. The cavity fills with water forming a lake, but unlike the Carolina Bays, the lakes are not all perfect ellipses, they have no overlaps, they have no preferential rim thickening, and they do not have raised rims. The shape and alignment of the thaw lakes is determined by the contours of the land, and the alignment is generally in the direction in which water drains toward the sea.

Click images to enlarge them

Figure 2.

Figure 3.
Thermokarst (Russia, 70.347, 71.771)

Carolina Bay characteristics
This section examines various physical characteristics of the Carolina Bays to try to determine mechanisms by which their features could have originated. Many of the features are best explained as a result of impacts, and it is for this reason that scientists persist in trying to develop an impact model that satisfies both astronomers and geologists.

Figure 4.
Tatum, SC (34.6441, -79.5868)

Figure 5.
Bowmore, NC (34.9196, -79.3206)

Figure 4 illustrates the great density of bays in the South Carolina coast. This image is completely covered with bays except in the stream basins where they have disappeared through erosion. The structural preservation of the Carolina Bays for at least a dozen millennia may be partly due to the flat porous landscape which allows rain water to quickly filter underground thereby preventing erosion from lateral water flow. It is remarkable that all the bays are elliptical and aligned with their major axes from the northwest to the southeast radiating from the Great Lakes region. Another salient feature is that many of the ovals overlap, which implies a chronology of stratigraphic formation. Figure 5 shows in more detail the overlaps in the bays. Some bays are completely contained within others, while other bays only partially overlap. One important detail is the thickened rim on the southeast side of the bays. This is a characteristic of oblique impacts which tend to push surface material in the direction of the impact. It is also important to notice that large ovals may have small ovals within them, and that large ovals cover smaller ovals. This can be explained by noting that small bays created by small impacts would have been obliterated when overlaid by a larger impact.

Carolina Bays occur in unconsolidated ground close to the water table along the Atlantic seaboard. The bays are also found in Nebraska and Kansas in what once were the shores of the Western Interior Seaway of North America. The bays in the Midwestern states differ from the ones in the east coast in that their major axes are aligned from the northeast to the southwest (Figure 5a). The fact that the Midwestern bays also radiate from the Great Lakes region provides support for an impact origin of the bays. Fewer Carolina Bays can be observed in the Midwestern states because only the larger bays have endured the erosion by water and the accumulation of layers of wind-blown dust and silt.

Figure 5a. Bays in Nebraska are degraded by erosion and deposition (40.566, -98.123)

Figure 6.
Carolina Bays (34.715, -79.470)

Figure 7.
Carolina Bays (34.791, -79.244)

Figure 8.
Carolina Bays, (34.773, -79.294)

Figure 9.
Carolina Bays (34.842, -79.218)

Figures 6 through 9 illustrate the variety of bay configurations that preserve the basic elliptical shape with northwest/southeast alignment and the thickened southeast rim. In some cases, erosion or farming machinery have destroyed these features. Here is a summary of the main morphologic characteristics of the Carolina Bays (Eyton 1975):

1. The width-to-length ratio averages 0.58 for both, the Carolina Bays and the Nebraska rainwater basins.
2. The Carolina Bays are ellipses although some lack bilateral symmetry along either the major or minor axis. The southeast portion of many bays is more pointed than the northwest end and the northeast side bulges slightly more than the southwest side. Known major axis dimensions vary from approximately 60 meters to 11 kilometers.
3. The Carolina Bays display a northwest-southeast orientation. Deviations from this orientation appear to be systematic by latitude (Prouty, 1952).
4. The bays are shallow depressions below the general topographic surface with a maximum depth of about 15 meters. Large bays tend to be deeper than small bays, but the deepest portion of any bay is offset to the southeast from the bay center.
5. Many bays have elevated sandy rims with maximum development to the southeast. Rim heights vary from 0 to 7 meters.
6. Carolina Bays frequently overlap other bays without destroying the morphology of either depression. One or more small bays can be completely contained in a larger bay.
7. The stratigraphy beneath the bays is not distorted (Preston and Brown, 1964; Thom, 1970).
8. Bays occur only in unconsolidated sediments. Bays in South Carolina are found on relict marine barrier beaches associated with Pleistocene sea level fluctuations, in dune fields, on stream terraces and sandy portions of backbarrier flats (Thom, 1970). No bays occur on modern river flood plains and beaches.
9. Carolina Bays appear to be equally preserved on terraces of different ages and formational processes.
10. Bays are either filled or partly filled with silt of organic and inorganic origin. Ghosts of semi-obliterated Carolina Bays appear to represent former bays which were filled by terrestrial sediments and organic materials.

The Glacier Ice Impact Hypothesis
Firestone (2009) and Davias (2010) have suggested the possibility that the Carolina Bays were created from impacts of ejecta from an asteroid or comet impact on the Laurentide ice sheet in the Great Lakes area. The Great Lakes area is a plausible point for an extraterrestrial impact because many of the straight lines drawn along the major axes of the bays in Nebraska, Kansas, Georgia, North Carolina, South Carolina and Virginia intersect around Michigan. The impactors that formed the bays have been suggested to be splashes of mud, ice-and-water slurry or just pieces of ice. However, to produce depressions like those of the Carolina Bays, the impacting body must resist deformation in order to penetrate the target surface. Viscous ejecta such as slurries or mud merely form a layer on top of a firmer target surface without creating a depression (Gault and Greeley, 1978). The premise for considering ice-boulder ejecta from an extraterrestrial impact on the Laurentide ice sheet is based on experiments with NASA's Ames Vertical Gun which have demonstrated that a high speed impact on an ice sheet shatters the ice sheet ejecting large quantities of ice pieces, rather than just melting or vaporizing the ice (Schultz 2009). The physics of excavation is nearly the same for small and large impacts (Melosh 1989, p.14), so the fracturing of an ice sheet and the ejection of ice pieces should scale up to an impact by a meteorite on a glacier.

The estimates for the formation of the Carolina Bays vary from 7,000 to 200,000 years ago (Brooks et al., 2010). The glacier ice impact hypothesis requires the extraterrestrial impact to have occurred at any time during the Pleistocene Epoch between 11,700 to 260,0000 years ago when North America had a thick cover of glacial ice.

 Impact on an ice sheet ejects pieces of ice (Schultz 2009)

This paper examines the hypothesis that the Carolina Bays were made by impacts of glacier ice boulders ejected at ballistic speeds from the Laurentide ice sheet by a meteorite impact. The examination will try to determine if the characteristics of the bays could be reasonably explained by impacts from ice boulders with substantially slower speeds than meteorites, and whether experimental tests provide support for the hypothesis. The hypothesis also considers that the impacts were made on viscous surfaces since the bays occur only on unconsolidated sediments and they are now shallow depressions instead of deep craters. No attempt will be made to establish the timeline of the meteorite impact, or to associate it with any mass extinction event.

Trajectory and Speed of Ejecta
If the hypothesis that at least one meteorite impacted the Laurentide ice sheet is true, and chunks of ice were ejected, the ejecta would have had to travel 1000 kilometers from central Michigan to North Carolina (Figure 10). We can use ballistic equations to calculate the speeds and angles necessary to launch a projectile this far. At an angle of 45°, which would cover the greatest distance at any given speed, the velocity v is given by the formula v2 = d×g, where g is the gravitational acceleration with a value of 9.81 m/s2 near the Earth's surface. The speed necessary to launch an object a distance d of 1000 kilometers at an angle of 45° is 3,130 meters per second, disregarding air resistance and the effect of the Earth's rotation. This is approximately 3 km/sec which is a relatively slow speed compared to the typical speeds of 17 km/sec for asteroid impacts or 50 km/sec for comet impacts.

Ejecta traveling a distance of 1,400 kilometers from the Great Lakes region to southern Georgia would need an initial speed of approximately 3.7 km/sec, which is still very slow compared to the speeds of asteroids or comets.

Figure 10. Trajectories of ejected glacier ice

The trajectories of the ejected material would have reached heights of 150 to 390 kilometers above the surface of the Earth, depending on the initial speeds and launch angles. Above 35 kilometers from the Earth's surface, the atmospheric pressure is below the triple point of water and water cannot exist in the liquid state regardless of the temperature. Liquid water in any ejected mud or water slurry would have boiled vigorously creating clouds of ice crystals and water vapor. The ice crystals would have blocked the light of the Sun, and could have contributed to a cooling interval or ice age. The glacier ice boulders would have traveled without significant aerodynamic drag above the Earth's atmosphere. During reentry, the leading surfaces of the ice boulders would have vaporized and acted as ablative heat shields. Aerodynamic stresses would have broken up some boulders to produce crater fields with many impacts. The breakup of the projectiles and mid-air collisions would have changed some of the trajectories and caused variations in the orientation of the resulting craters.

Mass of Impactors
If we consider that ice boulders were launched from the Laurentide ice sheet at a speed of 3 km/sec, we can estimate the mass of an ice boulder that would create a crater with a diameter of one kilometer, which is a common size of a Carolina Bay. The University of Arizona calculator for computing projectile size from crater diameter (Melosh 1999) allows the input of the crater diameter and various initial conditions to determine the diameter of the impacting object. Once the diameter of the impactor is known, it is possible to calculate its mass. The calculator also estimates the energy of the impact in Joules and megatons.

Input Parameters
Final Crater Diameter: 1 kilometer
Projectile Density: 917 kg/m3 (ice)
Impact Conditions:
Impact Velocity: 3 km/sec
Impact Angle: 45 degrees
Target Descriptors:
Target Density: 1500 kg/m3 (porous rock)
Acceleration of Gravity: 9.8 m/sec2
Target Type: loose sand

Results
The three scaling laws yield the following projectile diameters: (note that diameters assume a spherical projectile)
Yield Scaling: 1.79 × 102 meters
Pi Scaling (Preferred method!): 0.18 × 103 meters
Gault Scaling: 2.61 × 102 meters
Crater Formation Time: 6.56 seconds

Using the Pi Scaling method this impactor would have struck the target with an energy of 1.27 × 1016 Joules (3.03 MegaTons).

According to the calculation, a spherical ice boulder with a diameter of 180 meters traveling at a speed of 3 km/sec would create a crater of 1 kilometer when impacting at an angle of 45°. Using the formula for the volume of a sphere, the ice boulder would have a volume of 3 million cubic meters and would weigh 2.8 million metric tons. The ice boulder would be about the size of Yankee Stadium in New York City.

We can speculate that if there are 500,000 Carolina Bays (Firestone 2009) and each one was formed by an energy of 1.27 × 1016 Joules, then the total energy of the impacts was 6.35 × 1021 Joules. This total energy provides a rough lower limit of the kinetic energy transferred to the ejecta by an extraterrestrial impact. Additional energy would have been converted to heat, seismic energy and fracturing of the target and projectile.

Using the same number of bays, we can calculate that approximately 1.5 × 1012 cubic meters of ice were ejected by the impact on the Laurentide ice sheet. This amount of ice is also a minimum since much of the ice may have fallen on ground too firm for the formation of bays. If the ice sheet had a thickness of 1 kilometer at the point where the meteorite struck, the circular area containing this volume of ice would have had a diameter of 44 kilometers. The area would be comparable to the Chippewa Basin of Lake Michigan (highlighted in Figure 11). The Chippewa Basin was suggested as a possible extraterrestrial impact site by Firestone et al. (2006, p.267) because a seismic profile of the basin showed terrace faulting typical of an impact where large slabs of rock had cracked and slid downward. However, a USGS seismic stratigraphy analysis of Lake Michigan by Foster and Colman (1991) found an unconformity, but no evidence of faulting. The Michigan Basin that contains the Mississippian Aquifer is much older than the Carolina Bays and the thickening of the underlying Devonian, Silurian and Cambrian strata is not typical of an impact structure. Thus, the exact location of the extraterrestrial impact remains unknown.

Figure 11. Mississippian Aquifer in Michigan, USGS[1]

We can also calculate that an ejecta blanket containing 1.5 × 1012 cubic meters of ice distributed in a semicircle with a radius of 1400 km, would have covered approximately half of the contiguous United States, including the Midwest and the Atlantic coast, with pieces of ice to a depth of about half a meter.

Impact Model
This paper proposes an impact model that correlates well with the Carolina Bay characteristics. An oblique impact on a viscous surface creates a tilted conical cavity which at the intersection with the level surface of the Earth is a conic section that appears as an elliptical crater (Figure 12). An oblique impact by an ice boulder with a velocity of 3 to 4 km/sec would create a shallow oval crater. The craters would be deeper at the terminal end due to the inclined incidence of the impact. The overturned flaps from the impact would create raised rims around the crater, particularly at the terminal end of the ellipse (Figure 13). When the ice impactor melted, the melt water would create a pool in the depression. Because the ice was of terrestrial origin, the impact site would have no meteorite fragments or siderophile elements from the impactor. At most, it might be possible to find stones that were originally embedded in the glacier ice. Also, the slow-speed ballistic impacts on soft yielding ground would not have been able to generate the pressures of 2 to 5 GPa necessary to form shatter cones or other types of shock metamorphism. The elliptical craters created by the impacts would eventually fill by silting and viscous relaxation of the substrate.

Figure 12. Conic Sections
The eccentricity of an ellipse
depends on the angle of impact;
its size varies according to the
volume of material displaced.

Figure 13. Vertical cross section of an oblique impact

The interpretation of the Carolina Bays as conic sections makes it possible to predict where to find stones that might have been contained in the chunks of glacier ejected from the Laurentide ice sheet. The impact angle θ may be estimated from the ratio of the minor axis (W) to the major axis (L) using equations from forensic science.

sin(θ) = W / L

If the impactor comes to rest at the terminal end of the major axis of the ellipse, the depth D at which stones from the glacier might be found can be calculated as:

D = tan(θ) × L / 2

For a Carolina Bay with a minor axis of 520 meters and a major axis of 920 meters these equations calculate an impact angle of 34.4° and 315 meters as the depth where some of the stones carried by the glacier chunk might have ended up. Some stones might be found at shallower depths due to the buoyancy of ice. The larger bays may be more likely to contain some of these glacier stones.

Studies of the bays have not found evidence of disturbed strata, suggesting that no deep craters were created during their formation. However, unlike solid surfaces where the substrata are deformed by impacts (Melosh, 1989 p.79), the layers in unconsolidated viscous surfaces are penetrated and parted by the impacting body; the original stratigraphy is later partially reassembled when the depth of the cavity is reduced by viscous relaxation. An impact of an ice projectile on a viscous surface in combination with the restoring effect of viscous relaxation would leave little evidence of disturbance.

Why conical craters?
It is reasonable to ask why the craters created by the impacts of glacier ice have a conical shape. This is determined by cratering mechanics and the type of shock wave created by the impactor. Impacts on solid surfaces go through stages of compression, excavation and modification (Melosh, 1989). During the compression stage, the projectile is destroyed and most of its kinetic energy is transferred to the point of impact. In the excavation stage, a hemispherical shock wave expands from the point of impact creating a bowl-shaped crater. At the hyperspeeds of extraterrestrial impacts, the transfer of kinetic energy during the compression stage is virtually instantaneous so that even moderately oblique impacts create hemispherical shock waves and circular craters. Bottke et al. (2000) found that asteroid or comet impact angles of less than 12° from horizontal were required for the formation of elliptical-shaped craters on Mars, Venus, and the Moon. The modification stage is controlled mainly by gravitational forces that determine the final configuration of the crater.

By contrast, a collision with a viscous medium at a slower speed does not destroy the projectile. The excavation phase starts as the impactor continues to travel through the medium creating a conical or paraboloid
Paraboloid
shock wave until it is stopped by friction. A viscous surface with low elasticity will retain the conical shape, and it will gradually be modified by gravity through viscous relaxation. Figure 14 illustrates the shock wave produced by a bullet traveling at Mach 1.5 in air. Figure 15 shows the conical cavity made by a bullet on modeling clay.

Figure 14. A bullet at Mach 1.5
creates a conical shock wave

Figure 15. A bullet impact creates
a conical cavity on modeling clay

Eccentricity and time of emplacement
The size (major axis) of a bay is a function of the energy of the impact, but the eccentricity of the ellipse depends only on the angle of impact (Figure 12). For this reason, large and small bays have similar shapes, but the larger bays are made by larger projectiles.

The pattern of overlaps of the bays indicate the sequence in which the impacts occurred. In addition, using the formulas given above, the ratio of the minor to the major axes of the ellipses can be used to deduce the relative time of emplacement for bays which do not overlap.

Figure 15a. Bays with different eccentricities
(34.841, -79.221)

Figure 15b. Trajectories at different launch angles

Figure 15a illustrates Bay A which is more elongated than Bay B. Bay A has axes of 637×1166 meters with a ratio of 0.546 which corresponds to an impact angle of 33.114°.  Bay B has axes of 543×942 meters with a ratio of 0.576 which corresponds to an impact angle of 35.198°.  Trajectories of projectiles originating from the same point and traveling the same distance at different launch angles are illustrated in Figure 15b. Assuming that the launch angles of the ice projectiles are the same as the impact angles calculated from the bays, the ballistic equations indicate that for projectiles traveling a distance of 1120 km, the projectile corresponding to A with the smaller launch angle and lower trajectory had a speed of 3.465 km/sec and a flight time of 386 seconds. The projectile for B with the higher trajectory had a speed of 3.415 km/sec and a flight time of 401 seconds. Thus, the projectile for A would have reached its destination 15 seconds before the projectile for B.

Overturned Flaps
Gamble et al. (1977) suggested that the raised rims around the Carolina Bays might be secondary rims formed over time from eolian sand deposits which buried a primary rim at the edge of an initial depression. However, raised rims are a common characteristic of impact craters. The raised rims around an impact cavity are formed during the excavation phase as the shock wave of the impacting body penetrates the ground, displaces material laterally and ejects material above the surface. The overturned flaps are created during the remodeling phase when the ejected material falls back to the surface under the influence of gravity (Maxwell, 1977). On a viscous target, the ejected material acts like a breaking wave in which the base of the wave meets resistance from the surface while the crest continues its forward motion. The final overturned flaps usually display inverted stratigraphy. Figures 16a through 16c illustrate the development of the overturned flaps. Figure 17 indicates the locations of possible overturned flaps in several Carolina Bays.

Figure 16a. Excavation phase

Figure 16b. Remodeling phase

Figure 16c. Final overturned flap

Figure 17. Possible overturned flaps
in several Carolina Bays (34.857, -79.183)

Experimental Tests
Cratering mechanics are often studied using the CTH hydrocode software (Sandia[1]), but the physical principles of this impact model can be tested easily with a simple experimental setup. Figures 18 through 22 illustrate impact experiments conducted on a mixture of clay and sand that was mixed with enough water to have the consistency of mortar. Small ice spheres were used as projectiles propelled with a slingshot.

Click image to enlarge it

Figure 18. Oblique impact by ice creates a tilted conical cavity on clay-sand mixture.
The ice impactor is visible at the bottom of the crater.
Notice that the raised rim is actually the overturned flap from the impact.

Click image to enlarge it

Figure 19. A second oblique impact illustrates how the
rim material is pushed in the direction of the trajectory.
The ice of the first impact is already partially melted
and viscous relaxation has reduced the depth of the crater.

Click image to enlarge it

Figure 20. Melting of the ice impactors and viscous relaxation of the substrate
produce shallow oval pools with slightly raised rims.
The rim is thicker on the terminal side of the ellipse.

Projectiles which impact at a very shallow angle will ricochet, rather than penetrate the target. The craters left by these oblique impacts are also elliptical (Figure 21), but they differ from the conical craters made by projectiles that penetrate the surface (Figure 23).

Click image to enlarge it

Figure 21. An impact at a very inclined angle created a shallow oval crater
and caused the ice impactor to bounce or skip like a stone on water.
The top surface of the target was lightly covered with colored sand to enhance analysis.

Click image to enlarge it

Figure 22. One day after the impact, the crater contains water
that has filtered through the porous medium.
Craters formed by a bouncing impactor are not deeper at the terminal end.

Mechanism for the formation of overlapping bays
The following images illustrate the transformation of adjacent conical impact craters into overlapping depressions analogous to the Carolina Bays. The surface of the clay-sand target was thinly covered with colored sand to enhance the analysis of the impact images. Figures 23 and 24 show the conical craters of two adjacent large impacts created by ice ball projectiles. Figure 25 is a side view of the two impacts showing a thin layer of colored sand in the overturned flap that marks the boundary of the inverted stratigraphy. Figures 26 through 28 show the gradual reduction of crater depth due to viscous relaxation that results in elliptical water-filled depressions with raised rims. The overlap of the final depressions can be used to deduce the chronology of the impacts in the experimental tests. The same should be true for the Carolina Bays.

Click images to enlarge them

Figure 23. First conical crater

Figure 24. Second conical crater

Figure 25. Side view of the two impacts
showing boundary of inverted stratigraphy
in the overturned flap

Figure 26. Top view of craters

Figure 27. Reduction in crater depth
by viscous relaxation

Figure 28. Top view of final configuration

Detailed view of Figure 25 showing the boundary of the inverted stratigraphy.
A cross section of the overturned flap in the final configuration (Figure 28)
would reveal this boundary.

Mechanism for Soil Fluidization
The impact model for the Carolina Bays proposes impacts of ice boulders on a viscous surface. Several factors could have contributed to the formation of a viscous target surface at the time when the bays were formed.

The Carolina Bays occur only in unconsolidated sediments in low-lying areas where the water table is within a few meters from the surface (Eimers et al. 2001). Saturated, cohesionless soils like these may liquefy and flow like liquids when subjected to monotonic, cyclic, or shock loading (Sladen et al. 1985). The soil can undergo large deformations typical of fluids when the shear resistance of the soil becomes less than the static, driving shear stress (Martin et al. 1975). Geologic flow processes caused by seismic vibration have also been called "acoustic fluidization" (Melosh 1979) and "dynamic fluidization" (Richards et al. 1990).

A ground impact by a meteorite in the Great Lakes Region would have generated seismic shock waves that propagated at 2 to 8 kilometers per second, depending on the characteristics of the surface. P-waves traveling at 5 km/sec would have reached the Eastern seaboard approximately 3.5 to 4.5 minutes after the extraterrestrial impact. The shock waves of a large impact would have been able to liquefy saturated soil, as commonly happens during earthquakes. The ice boulders arriving 6 to 9 minutes after being ejected from the ice sheet would have hit a ground already fluidized by the seismic shocks of the extraterrestrial impact. In addition, the tremors produced by the multitude of impacts of huge ice boulders would also have contributed to the fluidization of the surface and speeded up viscous relaxation.

Viscous Relaxation
Hyperspeed impacts create different types of craters depending on the structure and properties of the target surface. Terrestrial craters with diameters of less than 4 km have a simple bowl shape, while larger craters typically have a central uplift from deeper strata or a multi-ring structure (French 1998). The uplift is created by the elastic properties of a solid surface which deforms and rebounds when impacted. By contrast, the Carolina Bays, even those with a major axis exceeding 10 km, have a uniform type of configuration due the way that impact cavities are formed on unconsolidated, saturated soils by projectiles at slower ballistic speeds.

A projectile at ballistic speed penetrates a viscous surface by parting the existing layers of the target during the excavation phase of the impact. All the energy of the projectile is spent moving the non-elastic material along its path. Viscous relaxation during the modification phase, which is driven by gravity, subdues topographic features, such as craters, and partially reconstitutes the stratigraphy by the flow of the material parted during the excavation. The leveling process gradually reduces the depth of the cavity, and ground vibrations may accelerate the rate of viscous relaxation by reducing the shear resistance of the soil.

A cavity in a viscous surface is filled by flow of the material surrounding the deepest part of the cavity. Pressure increases with depth and this creates a velocity gradient that promotes faster centripetal lateral flow at the bottom of a cavity in a uniformly viscous medium. The increase of pressure p with depth is given by the equation $$p = \rho g h$$, where ρ is the density, g is the gravitational constant and h is the depth. The flow of the medium can be modeled using the Navier-Stokes equations for incompressible flow, where v is the velocity, μ is the dynamic viscosity and f represents body forces such as gravity.

$$\rho \left({{\partial v} \over \partial t} + v \cdot \nabla v \right) = -\nabla p + \mu \nabla^2 v + f$$

The cavity cannot be filled by upwelling of substrata because the pressure from the weight of the terrain around the cavity which causes the remodeling cannot be fully transmitted to the substrata as long as material is being diverted by lateral laminar flow to fill the cavity. The flow into the cavity stops when the pressure is insufficient to overcome the frictional forces of the medium. Viscous relaxation basically reverses the sequence in which the cavity was created. Figures 29a and 29b illustrate that the material that fills any level is the same material that was parted by the shock wave of the impact. Photographs of an experiment to verify this process are shown in Figures 30 through 33.

Figure 29a. Centripetal laminar flow
characterizes viscous relaxation of cavities

Figure 29b. Strata are reconstituted
as cavity depth decreases

Stratigraphic Restoration Experiment
Figure 30 shows the crater made by the oblique impact of an ice ball on a sand-clay target that had been prepared with an underlying red layer approximately two centimeters below the surface. The penetration of the impactor through the red layer dragged along some of the red material. Figures 31 through 33 show the gradual reduction in depth as the viscous material adjacent to the deepest part flows to fill the cavity. The red layer remains at the same level and disappears from view when the impact structure reaches its final configuration because the cavity is filled by centripetal lateral flow of material. If the cavity had been filled by upwelling of material from lower strata, the red layer would have been raised and exposed to the surface.

Figure 30.

Figure 31.

Figure 32.

Figure 33.
Stratigraphic Restoration

Conclusion
The terrestrial models for thawed permafrost and other eolian and lacustrine processes lack a mechanism for explaining the regular elliptical shapes with raised rims of the Carolina Bays and their alignment toward a common radial point which are features more characteristic of impacts. The wide range of ages obtained for the Carolina Bays (e.g., Brooks et al., 2010) has been used to suggest that there were multiple periods of Carolina Bay formation by eolian processes and to refute the notion that a single impact event could have been responsible for the formation of the bays. However, if the bays were formed from glacier ice impacts and the stratigraphy was reconstituted by viscous relaxation after the ice melted, the material of the bays would have the same physical characteristics as before the impacts, and the age of the material where the bays occur would not be indicative of the time when the bays were formed.

The new model of slow-velocity impacts from ice ejecta resulting from a meteorite impact on the Laurentide ice sheet explains many of the characteristics of the Carolina Bays, including the lack of shock metamorphism and meteorite fragments. The model postulates that oblique impacts of glacier ice on the viscous surfaces of the coastline created slanted conical craters that became shallow elliptical pools through viscous relaxation of the substrata. The overturned flaps from the impacts became the raised rims of the bays. The model can be used to calculate the physical forces that may have been at play in the creation of the bays and provide some clues about the meteorite impact that created the ice boulders.

The glacier ice impact hypothesis can be confirmed by finding evidence of an extraterrestrial impact in the Laurentide ice sheet. Conversely, proving that the Carolina Bays were created by impacts may be used as evidence that there was an extraterrestrial impact on the glacier ice sheet. The use of conic sections as a mathematical model for the impacts will facilitate the analysis of the bays and provide guidance for geological field studies by predicting where to find stones carried within the glacier-ice impactors. Overturned flaps with inverted stratigraphy in the raised rims at the terminal end of the ellipses are also characteristic of impacts and they could generally be identified from relatively shallow core samples or trenches. Finding these overturned flaps and glacier stones may settle the controversy about the origin of the Carolina Bays and provide information about the site where the extraterrestrial impact occurred. Thus far, the strongest support for an extraterrestrial impact at the Younger Dryas boundary has been provided by analysis of Greenland Ice Sheet cores which showed a large platinum anomaly (Petaev, et al. 2013).

If the Carolina Bays were created by the proposed impacts, any fauna living in the central and eastern United States at that time would have seen a blinding flash followed by seismic shocks approximately four minutes later. It would have been almost impossible to avoid being hit by the barrage of enormous ice boulders seven minutes after the flash. Any survivors would have found their habitat destroyed. A local extinction event could have easily occurred under these circumstances, even without considering fires and other effects associated with the meteorite impact.

Acknowledgement: I would like to thank Dr. David Rajmon for his thoughtful and constructive comments on the initial version of this document.

References:
• Bottke, William F.; Love, Stanley G.; Tytell, David; Glotch, Timothy,
Interpreting the Elliptical Crater Populations on Mars, Venus, and the Moon, Icarus 145, 108–121 (2000) doi:10.1006/icar.1999.6323
• Brooks, M. J.; B. E. Taylor; and A. H. Ivester, 2010,
Carolina bays: time capsules of culture and climate change. Southeastern Archaeology. vol. 29, pp. 146–163
• Davias, M. [1], Carolina Bay LiDAR Imagery Viewer
• Davias, M., 2010, Correlating an Impact Structure with the Carolina Bays, GSA Denver Annual Meeting (31 October - 3 November 2010), Paper No. 116-13
• Eimers J.L.; Terziotti, S.; Giorgino, M.; (2001),
Estimated Depth to Water, North Carolina, Open File Report 01-487
http://nc.water.usgs.gov/reports/ofr01487/
• Eyton, J.R; Judith I. Parkhurst A Re-Evaluation Of The Extraterrestrial Origin Of The Carolina Bays, 1975
http://abob.libs.uga.edu/bobk/cbayint.html
• Firestone, R; West, A; Warwick-Smith, Simon; 2006,
The Cycle of Cosmic Catastrophes: How a Stone-Age Comet Changed the Course of World Culture
• Firestone, R.B., West, A., Kennett, J.P., Becker, L., Bunch, T.E., Revay, Z.S., Schultz, P.H., Belgya, T., Kennett, D.J., Erlandson, J.M., Dickenson, O.J., Goodyear, A.C., Harris, R.S., Howard, G.A., Kloosterman, J.B., Lechler, P., Mayewski, P.A., Montgomery, J., Poreda, R., Darrah, T., Que Hee, S.S., Smith, A.R., Stich, A., Topping, W., Wittke, J.H., Wolbach, W.S., 2007. Evidence for an extraterrestrial impact 12,900 years ago that contributed to the megafaunal extinctions and the Younger Dryas cooling. Proceedings of the National Academy of Sciences 104, 16016–16021.
http://www.pnas.org/content/104/41/16016.long
• Firestone, R. B., The Case for the Younger Dryas Extraterrestrial Impact Event: Mammoth, Megafauna, and Clovis Extinction, 12,900 Years Ago. Journal of Cosmology, 2009, Vol 2, pages 256-285.
[http://journalofcosmology.com/Extinction105.html ]
• Foster, D. S., and S. M. Colman, 1991,
Preliminary interpretation of the high-resolution seismic stratigraphy beneath Lake Michigan. Open-File Report no. 91-21. United States Geological Survey, Woods Hole, Massachusetts.
• French B. M. (1998) Traces of Catastrophe: A Handbook of Shock-Metamorphic Effects in Terrestrial Meteorite Impact Structures. LPI Contribution No. 954, Lunar and Planetary Institute, Houston. pp 23-28.
• French, Bevan M.; Christian Koeberl, "The convincing identification of terrestrial meteorite impact structures: What works, what doesn't, and why", Earth-Science Reviews, 98 (2010) 123–170.
• Gamble, E.E., Daniels, R.B., and Wheeler, W.H. (1977).
Primary and Secondary Rims of Carolina Bays. Southeastern Geology, 18, 199–212
• Gault, Donald E. and Ronald Greeley,
Exploratory Experiments of Impact Craters Formed in Viscous-Liquid Targets: Analogs for Martian Rampart Craters? Icarus 34, 486-495 (1978)
• Kennett, D.J., Kennett, J.P., West, A., Mercer, C., Que Hee, S.S., Bement, L., Bunch, T.E., Sellers, M., Wolbach, W.S., 2009a. Nanodiamonds in the Younger Dryas boundary sediment layer. Science 323, 94.
http://dx.doi.org/10.1126/science.1162819
• Kennett, D.J., Kennett, J.P., West, A., West, G.J., Bunch, T.E., Culleton, B.J., Erlandson, J.M., Que Heeg, S.S., Johnson, J.R., Mercer, C., Shen, F., Sellers, M., Stafford Jr., T.W., Stich, A., Weaver, J.C., Wittke, J.H., Wolbach, W.S., 2009b. Shock-synthesized hexagonal diamonds in Younger Dryas boundary sediments. PNAS 106, 12623–12628.
• Martin, G. R., Finn, W. D. L., and Seed, H. B. (1975).
"Fundamentals of liquefaction under cyclic loading." J. Geotech. Engrg. Div, ASCE, Vol. 101, No. GT5, May, pp. 423-438.
• Maxwell, D.E., 1977, Simple Z model of cratering, ejection, and the overturned flap,
Impact and Explosion Cratering, Pergamon Press (New York), p. 1003-1008.
• May, James H., and Andrew G. Warne, Hydrogeologic and geochemical factors required for the development of Carolina Bays along the Atlantic and Gulf of Mexico, coastal plain, USA, Environmental & Engineering Geoscience August 1999 v. 5 no. 3 p. 261-270
• Melosh, H. J., (1979) Acoustic fluidization: A new geologic process?
Journal of Geophysical Research, Vol. 84, Issue B13, pages 7513–7520, 10 December 1979 [http://dx.doi.org/10.1029/JB084iB13p07513 ]
• Melosh, H.J., 1989, "Impact Cratering: A Geologic Process"
Oxford University Press
• Melosh, H.J; Beyer, R.A., 1999, Computing Projectile Size from Crater Diameter
http://www.lpl.arizona.edu/tekton/crater_p.html
• Melosh, H.J., 2011, Planetary Surface Processes, Cambridge University Press, p. 462.
• Melton F.A., Schriever W. (1933) Geology 41: 52-66.
• Petaev, Michail I.; Shichun Huang, Stein B. Jacobsen, Alan Zindler, Large Pt anomaly in the Greenland ice core points to a cataclysm at the onset of Younger Dryas, PNAS July 22, 2013, doi: 10.1073/pnas.1303924110
• Pinter, Nicholas; Andrew C. Scott; Tyrone L. Daulton; Andrew Podoll; Christian Koeberl; R. Scott Anderson; Scott E. Ishman; The Younger Dryas impact hypothesis: A requiem, Earth-Science Reviews, Volume 106, Issues 3–4, June 2011, Pages 247–264
[http://dx.doi.org/10.1016/j.earscirev.2011.02.005 ]
• Preston, C. D., and Brown, C. Q., 1964. Geologic Section along a Carolina Bay, Sumter County, S. C., Southeastern Geology, vol. 6, pp. 21-29.
• Prouty, W. F., 1952. Carolina Bays and their Origin, Bulletin, Geological Society of America, vol. 63, pp. 167-224.
• Richards, R., Jr., Elms, D., and Budhu, M. (1990).
"Dynamic Fluidization of Soils." J. Geotech. Engrg., 116(5), 740–759.
• Sandia[1] - Sandia National Laboratories CTH Shock Physics
http://www.sandia.gov/CTH/index.html
• Schultz, P.H. (2009) in "Last Extinction: Megabeasts' Sudden Death", NOVA, written and produced by Doug Hamilton, Public Broadcasting System.
• Sladen, J. A., D'Hollander, R. D., and Krahn, J. (1985).
"The liquefaction of sands, a collapse surface approach." Canadian Geotech. J., Vol. 22, pp. 564-578.
• Thom, B. G., 1970. Carolina Bays in Horry and Marion Counties, South Carolina, Bulletin, Geological Society of America, vol. 81, pp. 783-814.
• Trajectory of a projectile, Wikipedia
http://en.wikipedia.org/wiki/Trajectory_of_a_projectile
• USGS[1] U.S. Geological Survey, Ground Water Atlas Of The United States,
Iowa, Michigan, Minnesota, Wisconsin, HA 730-J, Fig. 61,
http://pubs.usgs.gov/ha/ha730/ch_j/J-text5.html