A paper published in July 2021 used computer simulations to examine the effect of ice sheet thickness on the formation of the Hiawatha crater in Greenland. This presentation compares the results of the simulation with actual experimental impacts.
Transcript:
The Hiawatha Crater in Greenland revisited. On November 14, 2018, a paper by 22 scientists in the journal Science Advances announced the discovery of a crater in the northern part of Greenland that was buried under the Hiawatha glacier. The paper received wide coverage in the media because scientists speculated that the extraterrestrial impact that made the crater could be related to the extinction event in North America at the onset of the Younger Dryas cooling event 12,900 years ago. This video reports on a paper published in July 2021 that used computer modeling to determine the effect of ice sheet thickness on the formation of the Hiawatha crater.
The abstract of the paper by Elizabeth Silber and five co-authors says: The discovery of a large putative impact crater buried beneath Hiawatha Glacier along the margin of the northwestern Greenland Ice Sheet has reinvigorated interest into the nature of large impacts into thick ice masses. This circular structure is relatively shallow and exhibits a small central uplift, whereas a peak-ring morphology is expected. This discrepancy may be due to long-term and ongoing subglacial erosion but may also be explained by a relatively recent impact through the Greenland Ice Sheet, which is expected to alter the final crater morphology.
The abstract continues: Here we model crater formation using hydrocode simulations, varying pre-impact ice thickness and impactor composition over crystalline target rock. We find that an ice-sheet thickness of 1.5 or 2 km results in a crater morphology that is consistent with the present morphology of this structure. Further, an ice sheet that thick substantially inhibits ejection of rocky material, which might explain the absence of rocky ejecta in most existing Greenland deep ice cores if the impact occurred during the late Pleistocene. From the present morphology of the putative Hiawatha impact crater alone, we cannot distinguish between an older crater formed by a pre-Pleistocene impact into ice-free bedrock or a younger, Pleistocene impact into locally thick ice, but based on our modeling we conclude that latter scenario is possible.
The topic of an extraterrestrial impact on an ice sheet is of great interest for me. I have written three books and one peer-reviewed publication proposing that an extraterrestrial impact on the Laurentide Ice Sheet ejected ice boulders whose secondary impacts created inclined conical cavities that transformed into shallow elliptical bays by viscous relaxation.
I have even performed experiments demonstrating that oblique impacts by ice projectiles on a viscous surface can produce inclined conical cavities that look elliptical when viewed from above.
Figure 2 of Dr. Silber's paper has a time series of a modeled impact into a 1.5-km-thick ice sheet. Material is colored according to material type. Blue represents the ice sheet, light brown represents a granitic crust target material, and dark brown represents the iron impactor. The axis origin marks the point of impact. The inset on the right illustrates a cross-section of the impact cavity.
I was particularly disappointed to see the numerical simulation treat the ice sheet as a plastic material that can be deformed. This is not what really happens in nature.
Professor Peter Schultz from Brown University has conducted many tests of high speed impacts using the AMES high-speed gun of the National Aeronautics and Space Administration.
This is one of his tests. The computer model illustrates a layer of ice undergoing a plastic deformation almost as if it were liquid water, whereas the experimental impact shows that the ice sheet breaks apart from the shock of the impact and the pieces of ice are then ejected far from the impact zone. The computer model is wrong because it does not produce results that correspond to the real physical event.
The incorrect computer model also leads to the wrong conclusions, such as the statement that if the ice sheet was 1.5 to 2 kilometers thick at the time the putative Hiawatha crater formed, the impact by an iron asteroid would have melted 106 to 164 cubic kilometers of ice, comparable to the amount of water in Lake Tahoe. This assumes that the ice remains in the vicinity of the impact in order to be melted, but the physical experiment showed us that chunks of the ice layer are ejected far away from the impact zone and they don't stay around to have a chance to be melted by the heat generated by the impact.
The authors relied on the iSALE simulation code for their work, so they just trusted the results of the program. The iSALE shock physics code is described as a multi-material, multi-rheology shock physics code that is a well-established, world-class tool for studying impacts, and that it has been used in pioneering studies of the formation of large impact craters on the Earth and the influence of target property variations on crater formation as well as the influence of a water layer on crater formation. The deficiency of the code seems to be in the simulation of the ice layer.
A paper by Arakawa and two co-authors has information that could improve the computer simulation. The abstract says: We conducted impact experiments with water ice at an impact velocity of 3.6 km/s and observed shock wave and fracture propagation in it by means of ultra-high speed photography. We observed that a region in which HEL (Hugoniot elastic limit) followed the elastic precursor wave, expanded with a velocity of 3-2.5 km/s until the pressure fell below 240 MPa. Below that pressure, a damage region appeared 0.8-3/microseconds after the passage of precursor wave. In this region, dynamic shear strength of water ice was estimated to be 21 MPa. Below 80 MPa, the several radial cracks proceeded toward the rear surface and broke the sample before the tensile fracture caused by reflection waves from an antipodal point became visible. Therefore, the main mechanism to cause the largest fragment is the radial crack growth rather than a spallation at the rear.
The important findings in this paper are, that following an impact, the shock wave propagates at a velocity of 3 to 2.5 kilometers per second through the ice, and a damage region of radial cracks breaks the sample 0.8 to 3 microseconds after passage of the precursor wave. The image at 5.5 microseconds labels the precursor wave front as A, and the shear damage zone expansion is labeled C. The reflection waves from the side are labeled B. The mechanism of fracturing reported by Arakawa is exactly what we see in the experimental impacts by Professor Schultz.
Pay close attention. The projectile penetrates the ice sheet and sends a shock wave that breaks it up. The intense heat of the impact creates a steam plume explosion that ejects the ice pieces in ballistic trajectories.
This is the contact and compression stage of the impact. The book entitled Impact Cratering by Professor Jay Melosh says: As the projectile continues its plunge into the surface, the shock fronts spread and propagate into both projectile and target. Shock pressures developed during the early stages of most hypervelocity impacts generally reach hundreds of gigapascals and far exceed the yield strength of both projectile and target. Both materials may either melt or vaporize upon unloading from such pressures. High-speed jets of strongly shocked material squirt out from the interface between the projectile and target. Most of the projectile's initial kinetic energy is transferred to the target. The underlying rocks are compressed, heated, and accelerated to high speed.
When the excavation phase starts, we can already see that the ice sheet is fracturing from the shock wave of the impact. The expansion of the vapor plume helps to propel the ice pieces ejected by the impact. The innermost ejecta are launched first and travel the fastest following the steepest trajectories. Ejecta originating farther from the center are launched later and move more slowly, falling closer to the impact point.
This video points out the discrepancy between the current iSALE model of a hyperspeed impact on an ice sheet compared to what actually happens in physical experiments. An impact on an ice sheet does not melt the ice to create big pools of water. Instead, the impact on an ice sheet fractures the ice and the pieces of ice are ejected in ballistic trajectories originating from the impact point. Ice is a bad conductor of heat and the kinetic energy of an impact is more likely to be transferred as kinetic energy to the ice fragments produced by the impact than converted to heat for a phase transition of the ice from solid to liquid and from a liquid to a gas. All these transfers of energy are possible, and the computer model should partition the energy to produce results that conform to the experimental observations.
A good computer model should be able to simulate an impact on the Laurentide Ice Sheet to produce the ice boulders necessary to create the Carolina Bays from the secondary impacts.
