Predicting the Consequences of iron-NEO Collisions with Earth: The Impact Effects Knowledge-base

In recent years, interest in predicting the consequences of the encounter of a NEO with Earth has increased. Most studies focus on rocky or cometary objects [1-3]. However, in the most frequent size range of the smallest objects, the effects of iron objects are more severe than for equal-sized rocky objects. Indeed, most small craters on Earth have been formed by iron objects, which penetrate the atmosphere more efficiently than stony asteroids. Intact meteoroids form single craters (e.g. Kamil), fragmented meteoroids form crater/meteorite strewn fields (e.g. Morasko), and larger asteroids fragment and form a single crater (e.g., Barringer). The atmospheric entry and cratering generate shock waves, which can cause damage on the ground. In this project, we combine different approaches in a knowledge-base to predict the outcome of such events with sufficient accuracy in a fast way. Such a tool is highly desirable for ESA's Planetary Defence Office to release impact warnings and inform emergency response agencies. 

We describe the behaviour of iron objects of few metres in size up to ~50 m in diameter. We assume that their strength follows a typical Weibull power law. In order to simulate the atmospheric entry of an iron object, we follow different approaches: 1) small (i.e. ~1 m sized) meteoroids that are strong enough to avoid fragmentation in the atmosphere are subject to ablation and deceleration, 2) slightly larger asteroids (i.e. few metres) are described by a separate fragment model [e.g. 4,5], and 3) the largest asteroids are described by the pancake model [e.g. 6].

The effect of atmospheric shock waves (i.e. overpressure and wind speeds on the ground) are calculated based on fits to nuclear explosion data [1,7], which yield circular-symmetric results. For cases that are not accurately described by simple parameterisation, we use the shock physics code SOVA [8] to simulate the evolution of shock waves based on pre-calculated energy release curves. Tracer gauges measure the pressure and velocity distribution over time in atmosphere and store the maximum values.

Distinction between individual fragments and a homogenous cloud of fragments: The fragmentation of an iron meteoroid/asteroid and the altitude at which such an event occurs depends on the initial parameters (velocity, entry angle, size). Large objects typically have a lower strength than smaller ones and fragment earlier in the atmosphere. Once fragmented, the individual pieces separate from each other with a lateral dispersion velocity, which depends on the absolute velocity of the object [4]. To decide which parameterisation to use, we compare the total cross- sectional area of all fragments with their separation distance (similar approach has been used in [9]). As long as the cross-sectional area of all fragments is larger than the corresponding size of the expanding fragment cloud (estimated from the dispersion velocity), the pancake model can be applied. This condition is expressed by the following equation (assuming equal-sized fragments):

with the initial size R, separation velocity constant C, air density ρ_air, asteroid density ρ_asteroid, Weibull exponent α_W, atmospheric scale height H, number of fragments N and entry angle ϑ. The altitude of fragmentation (and, thus, the atmospheric density at fragmentation) depends on the initial parameters, which is why the equation cannot be solved analytically. The dependence is shown in Figure 1. A meteoroid of the size of the Morasko event (~3 m radius, [5]) falls into the regime where the separate fragment model is applied. An asteroid like the one that formed the Barringer crater (~25 m radius) can be modelled with the pancake model. We also apply this approach to fragments with exponential size-frequency distribution, and the results are essentially the same.

Figure 1: Expansion radii (colour lines) versus initial radius of a meteoroid (X-axis and grey line). Iron density is 7800 kg/m³, the reference strength for a 1kg sample is 440 MPa, α_W=0.2. If the coloured lines are below the grey line, the pancake model is valid (separation distance is smaller than the maximum radius of the compact body). If the coloured lines are above the grey line, the separate fragment model is applicable (fragments are well-separated when the second fragmentation occurs). Colours represent different number of fragments: black – 10; blue – 100; green – 1000; and red – 10⁶.

Overpressure: We show results for two size classes of objects, which represent the size of the Kamil and the Barringer cratering events (Figure 2). For the Kamil event (initial radius of ~ 1.5 m), the outcome depends on modelling parameters (e.g. strength) and can vary between the formation of a single crater or a strewn field. However, overpressure distributions show little variation with and without a fragmentation event. For larger objects, the strength is less important. Overpressures are caused mostly by the crater formation and to some part by the atmopsheric entry.

Figure 2: Overpressure caused by the entry of two different sized objects. Top: 1.5m radius, 20km/s, 45° entry angle, different strength parameters, Bottom: 23m radius, 18km/s, 30°- 45° entry angle (green, and red & blue, respectively). The red line represents a ten times weaker reference strength. Results based on nuclear explosion data [7] are shown in black.

Shock physics codes give most reliable results for the prediction of overpressures. However, for a range of input parameters, the usage of parameterisation in combination with nuclear explosion data produces reliable estimates much faster (cf. Figure 2). In this project, we use a combination of both approaches. For small iron objects, overpressures are low apart from a small range near the crater, and damage is localised within a few crater radii.

Acknowledgements

We acknowledge the funding by ESA SSA-NEO, contract code P3-NEO-VIII.

References

[1]Collins G.S., Melosh H.J. and Marcus R.A. (2005)M&PS,40(6),817-840. [2]Wheeler et al. (2018)Icarus,315,79-91. [3] Artemieva N. and Shuvalov V. (2019)M&PS,54(3),592-608. [4] Artemieva N. and Shuvalov V. (2001)JGR,106,3297-3309.  [5] Bronikowska et al. (2017)M&PS,52(8),1704-1721. [6] Chyba C.F., Thomas P.J. and Zahnle K.J. (1993)Nature,361,40-44. [7] Collins et al.(2017)M&PS,52(8),1542-1560. [8] Shuvalov V.V. (1999)ShockWaves,9,381-390. [9] Svetsov V.V., Nemtchinov I.V., and Teterev A.S. (1995)Icarus,116,131-153.