To study the effect of shallow excavation we performed numerical simulations of an impact against a three-layered target. We used the same numerical model as in Ormo et al. (2002), but increased the spatial resolution of the numerical grid (80 cells across water column) and explicitly included a sedimentary layer (10 cells thick). Based on the results of previous simulations we considered the impact of a 300-m-radius stony asteroid into 800-m-deep water at a velocity of 20 km/s. The crystalline basement was covered by a 100-m-thick sediment layer with low strength (107 Pa cohesion and zero internal friction). The strength model used for the sediments is rather arbitrary. We do not know the real strength and simply consider a limiting case of low strength media. The effect of spallation resulting from the interaction of the shock wave with solid-water boundary could cause an intrusion of sediment fragments into water. This could result in mixing of the water and sediments and formation of a mud-like medium with very low strength. We used the Tillotson equation of state (Tillotson 1962) for wet tuff (with a density of 2 g/cm3) to describe the thermodynamical properties of the sediments and the ANEOS equation of state (Thompson and Lauson 1972) for granite to describe thermodynamical properties of the impactor and crystalline basement. We consider an axisymmetric 2D problem. The nonuniform computational grid consists of 250^250 cells in vertical and horizontal directions. Near the sea floor we used rather high spatial resolution (space step equals 10 m, that is 10 cells across the sedimentary layer). The impactor was considered to be an ellipsoidally shaped granite body flattened in the vertical direction (with a diameter that is two times its height). Its mass equals the mass of a 300-m-radius spherical impactor. For land impacts this flattened shape provides a transient crater that is similar to the transient crater formed after an oblique impact (Boris Ivanov, personal communication).

The results of the numerical simulations are shown in Fig. 1. At the beginning the ejecta curtain consists of sediment fragments only. Later, a more powerful basement ejecta curtain form. We can see a considerable difference between the behavior of water and that of sediments, which is due to the higher density and strength (water has no strength) of sediments. The rather minor thickness of the sediment layer also plays some role in the different behavior of the sedimentary layer. In analogy to water we could say that we consider an impact into "shallow sediments". As a result there is no separate sediment cavity (similar to separate water cavity). However, moving outwards, the basement ejecta curtain pushes off low strength sediments (like a bulldozer). As a result an area that is cleared of sediments forms around the cavity. Similarly the strengthless water is pushed outwards after impact into shallow water (Ormo et al. 2002).

The width of the area cleared from sediments and covered by basement ejecta reaches a width of 1-1.5 km outside the basement crater rim. This roughly agrees with the field observations. Moreover, the area cleared of Ordovician strata (the upper part of the sedimentary layer) is slightly larger. The area cleared from sediments can also increase due to water erosion (during water cavity growth), which is not taken into account in these simulations. Dense, shock compressed water moves outside at a velocity of about 100 m/s and can strongly erode a sedimentary surface disturbed by shock wave and spallation. The water density does not change significantly (maximum 1.5-2 times), but pressure increases by orders of magnitude.

The erosion also can explain the wide range of available estimates of the outer crater diameter (from 12 to 24 km). The flow velocity continuously decreases with distance from the central crater, and the erosion also decreases. The erosion of the sea floor is determined by a powerful turbulent flow in the boundary layer near the sea floor (Dalwigk and Ormo 2001). When the surface is eroded by powerful gas flows, the eddy diffusion coefficient and the flux of the blown-off material are proportional to the squared velocity of the flow (Adushkin and Nemchinov 1994). This corresponds to the scenario in which a particle is broken from the surface by force, which arises as a result of the gas flow around the particle and is proportional to the squared velocity of the stream. Similar ideas are used for water flows in river engineering (Paintal 1971). If we assume that the quadratic dependence is correct also for impact-induced water-stream erosion, the thickness A of the soil washed away by water at distance R from the crater center can be estimated as A =k\[u(R)]2dt, where u is the flow velocity given from the numerical simulations described above and k is an unknown constant.

Figure 2 shows the dependence of the integral ¡u2dt on the radial coordinate R both for outward and inward (resurge) water flows. We can see that in the brim area (4 km<R<5.5 km) erosive power of the outward directed flow is comparable with that of the resurge (within a factor of two). Outside the brim most erosion of the sea floor is produced by the resurge. Note that the expression used for eroded depth is only a very

Fig. 2. Integral Wdt (erosive power) versus distance R from the crater center. Black line indicates outward flow, the thick gray line resurge flow.

Fig. 2. Integral Wdt (erosive power) versus distance R from the crater center. Black line indicates outward flow, the thick gray line resurge flow.

crude approximation. In particular, the value of k can depend on the soil density, the surface roughness, the size of the particles washed away, and so forth. Furthermore, there is a minimum velocity threshold for the stream at which soil particles of a definite size break away from the surface.

Was this article helpful?

## Post a comment