Content uploaded by Hugh G. Lewis
Author content
All content in this area was uploaded by Hugh G. Lewis on Feb 01, 2016
Content may be subject to copyright.
Near Earth Objects, our Celestial Neighbors: Opportunity and Risk
Proceedings IAU Sy mpo si um N o. 236, 2006
A. Milani, G.B. Valsecchi and D. Vokrouhlick´y, eds.
c
2007 International Astronomical Union
doi:10.1017/S1743921307003614
Near Earth Object impact simulation tool
for supporting the NEO mitigation decision
making process
Nick J. Bailey, Graham G. Swinerd,
Andrew D. Morley and Hugh G. Lewis
Department Aerospace Engineering, School ofEngineering Sciences,
University of Southampton, Southampton, UK
Abstract. This paper describes the development of a computer simulation tool, NEOSim, capa-
ble of modelling small NEO impacts and their effect on the global population. The development
of the tool draws upon existing models for the atmospheric passage and impact processes. Sim-
ulation of the land and ocean impact effects, combined with a population density model, leads
to casualty estimation at both a regional and global level. Casualty predictions are based upon
the intensity of each impact effect on the local population density, with consideration given to
the population inside or outside local infrastructure. Two case studies are presented. The first
evaluates the potential threat to the UK, and highlights coastal locations as being at greatest
risk. Locations around Cornwall demonstrate an increase in casualties above the local average.
The second case study concerns the potential impact of asteroid (99942) Apophis in 2036. Prop-
agation of the possible orbits along the line of variance leads to an extensive path of risk on
the Earth. Deflection of the asteroid, by a variety of means, will move the projected impact
site along this path. Results generated by NEOSim for the path indicate that South American
countries such as Colombia and Venezuela are at a greatest risk with estimated casualty figures
in excess of 10 million. Applications of this software to the NEO threat are discussed, along with
the next stage of NEO impact simulation development.
Keywords. Impact risk
1. Introduction
Near Earth Objects have been the subject of continued study for many years with
NASA’s Spaceguard Survey providing the most comprehensive search and activity to
date. Objects greater than 1 km in diameter are the focus of these surveys, while small
sub-kilometre bodies remain largely uncatalogued. This NEO natural hazard was brought
to public attention by the recent discovery of (99942) Apophis, a relatively small 320 m
diameter asteroid, as it passed by Earth in December 2004. Following this discovery,
the Earth witnessed the largest natural disaster on record when a 9.0 earthquake off
the coast of Sumatra generated a tsunami wave which inundated coastlines around the
Indian Ocean and parts of Africa. The calculated death toll for this event is 229,866
(United Nations 2006 and was a relatively ‘small’ tsunami compared to ocean impact
generated tsunamis. Both these events demonstrate the significant threat posed by the
sub-kilometre NEO impact hazard.
The work outlined in this paper began in the middle of 2004 and has focused on de-
veloping a computer simulation tool capable of modelling both land and ocean impact
scenarios for sub-kilometre bodies. The program is called NEOSim and generates a ca-
sualty estimate by assessing the interaction of the impact generated effects with human
populations. Two case studies have been investigated which deal with the local and global
477
478 N. J. Bailey et al.
threat. The first scenario looks at the threat posed to a small region by objects landing
in the vicinity. The UK was chosen as the test case, but this methodology can be applied
to any location on the globe to assess the local risk. The second scenario investigates
potential impact events of the asteroid Apophis along the predicted line of risk for the
potential Earth encounter in 2036. This line stretches from Kazakhstan, across the north
Pacific and Central America to the Cape Verde Islands in the Atlantic. Analysis shows
the potential consequences for a space mission that seeks to mitigate the impact threat
by altering the object’s orbit.
2. NEOSim Methodology
A study of the literature revealed that there was a deficiency in tools that enabled the
study of NEO impacts, including both land and ocean impacts. Three distinct phases
of an impact were identified to be incorporated into the NEOSim program: atmospheric
entry; impact energy transmission; and casualty prediction. These are briefly discussed
here.
2.1. Atmospheric entry
NEOs impact the Earth at hypersonic velocities (typically in excess of 12 km/s) and,
during atmospheric entry, generate complex hypersonic flows around the object. The
dominant feature is the high stagnation pressure at the leading edge of the object and the
expansion of the bow shock around the body. Behind the shock front the high pressure
generates high temperatures and these combined effects lead to a mass loss through
ablation of the surface material. If the stagnation pressure difference exceeds the internal
strength of the object, which is determined by the physical composition of the NEO,
the object will rupture leading to fragmentation. NEOSim incorporates three models for
studying the effect of fragmentation.
•Single object model for robust, high strength objects
•Catastrophic fragmentation for weak objects
•Progressive fragmentation for pre-fractured objects
The single object model assumes the NEO remains intact throughout the atmospheric
passage with mass loss only through ablation while the catastrophic fragmentation model,
developed by Chyba et al. (1993) and Lyne et al. (1996) commonly known as the ’Pan-
cake’ model, deals with very low strength objects that disintegrate in the atmosphere.
This model is thought to best represent the Tunguska event of 1908. the progressive frag-
mentation model was developed by Boroviˇcka et al. (1998), Baldwin & Sheaffer (1971)
and Foschini (1998) to model objects that are pre-fragmented in orbit with relatively
high internal strength. These objects break apart in the atmosphere but don’t entirely
disintegrate resulting in multiple ground impacts. A pre-defined break-up altitude is
implemented according to Klinkrad et al. (2004), set at 30 km.
2.2. Surface Impact Effects
During impact the object’s kinetic energy excavates a crater before being transmitted
through a number of mechanisms identified by Collins et al. (2004). For land impact
these mechanism include ejecta distribution, a seismic shock wave, a surface blast wave
and thermal radiation generated by an expanding fireball. These impact generated effects
can be likened to those associated with nuclear detonations. However, little data from
nuclear testing is publicly available. Instead, the models by Glasstone & Dolan. (1977)
and Collins et al. (2005) provide a good approximation and have been implemented in
the NEOSim software.
Tool for supporting the NEO mitigation decision making process 479
Ocean impacts account for approximately two thirds of all NEO impact events due
to the proportion of Earth covered by water. Ocean impacts are characterised by the
excavation of a transient cavity (or crater) through the deposition of the object’s kinetic
energy at the impact site. This cavity is naturally unstable and immediately in-fills from
the surrounding ocean. This in-filling water oscillates vertically generating the tsunami
wave train. NEOSim implements the models by Chesley & Ward (2003) to calculate the
tsunami’s shoaling characteristics (wave run-up and run-in distances) on surrounding
shorelines.
A special case exists where the transient cavity depth is greater than the ocean depth.
In this ‘bottoming out’ scenario both a tsunami and some land impact generated effects
are assumed to affect the surrounding region.
2.3. Casualty Prediction
Human population density data was obtained from NASA Visible Earth†. Although the
data set is global, the low resolution limits the accuracy of the casualty estimations. This
data is utilised by a dedicated casualty prediction algorithm that has been developed to
assess each impact scenario. Following the simulation of an impact, the population in the
area affected is used to calculate the casualty figure. The number of casualties at each
point is depended on the local population and the severity of each effect. Increasing the
distance from the impact site is the main factor in reducing the severity of each land
impact effect. Several casualty prediction variables can be manipulated by the user for
fine tuning of the scenario, including how many casualties are generated by the increasing
effect intensity. Consideration is given to the percentage of people inside or outside local
infrastructure, as this will offer some protection, assuming the buildings are not destroyed.
The sum of casualties generated by each effect is output as a casualty distribution map.
For ocean impacts, casualty prediction is implemented in a similar manner, with casu-
alties generated along the inundated littoral. The effectiveness of the wave in producing
casualties is based upon the run-up height of the wave at shoaling and the run-in distance.
The longer travel time of a tsunami provides populations with more time to evacuate,
provided that there is a warning system. This casualty-reducing factor is incorporated,
based on the wave travel time.
2.4. Data Output
NEOSim’s primary output is the total casualty figure, which is provided to the user via
a pop-up window. The user is then presented with the array of outputs provided by
NEOSim. These outputs overlay data concerning each impact generated effect (both for
land and ocean impacts) onto a world map. Furthermore, the casualty data is provided as
an overlay onto the casualty density map with shading from dark to bright to denote low
to high casualty densities. Figure 1 provides an example of the NEOSim outputs with a
test impact into the North Sea (denoted by the small cross). The figure is a composition,
with the left side showing the impact effects, in this example a tsunami, and on the right
side the casualty density map. Verification of the absolute casualty figure for any impact
is difficult due to the lack of first hand impact experience and the number of factors
involved in casualty generation. Thus, NEOSim focuses on the relative casualty density
over the area affected to provide information on the locations at greatest risk and to
determine which populations should be evacuated in the event of an impact.
†NASA Visible Earth, Population Density Map, [Online] http://visibleearth.nasa.gov/
view rec.php?=116 [15 March 2005].
480 N. J. Bailey et al.
Figure 1. Figure shows the output provided by NEOSim. The left side shows the impact
generated effects distribution, here the tsunami wave. The right provides the casualty density
map with light regions representing more casualties.
Table 1. Characteristics of the test NEO used in the Case Study One simulation.
Parameter Initial Value
Radius 150 m
Impactor Velocity 25 km/s
Altitude 106 km
Density 2600 kg/m3
Yield Strength 10000000 N/m2
Heat of Ablation 8000000 J/kg
3. Case Study 1 – United Kingdom NEO Impact Risk
3.1. Methodology
In order to study the effect of NEO impacts around a region on the Earth rather than
simply a single impact, a multi-run application was built into NEOSim. This multi-run
tool allows the user to select a region defined by a latitude and longitude bin over a
particular location of interest. This region is then divided into a number of cells (defined
by the user) into which a single object is impacted. The casualty figure output from
each cell impact is used to shade the output map. Mapping the results in this manner
avoids the problem of validation of any specific casualty figure by providing the relative
risk of an impact into each cell. For this case study the UK was chosen as the target
with the latitude bin from 40◦to 70◦north and longitude bin from 30◦west to 15◦east.
Resolution of the grid cells (the number of cells that fit within in the latitude longitude
bins) is restricted by the processing time required for the simulation. Increasing the
resolution and therefore number of cells improves the quality of the data but requires
increasingly long runtime. For the present case study a 400 cell grid was chosen as a
compromise between runtime and quality. The characteristics of the incident NEO are
defined in Table 1.
3.2. Resul ts
Figure 2 presents the results for the UK case study. The results show that the UK
population is at greatest risk from objects impacting in the Atlantic Ocean to the west
and south west. The peak amplitude of the tsunami is dependent on the depth of water at
the point of impact with deep transient craters resulting in large amplitude oscillation as
the cavity in-fills. Impacts into shallow oceans bottom out preventing the large oscillations
of water, generating only small tsunami waves. This explains the low expected casualties
Tool for supporting the NEO mitigation decision making process 481
Figure 2. Shading denotes the number of casualties by the impact in the centre of each cell
with dark shading representing more casualties.
from impacts into the North Sea where the water is only 30 m deep at maximum removing
the potential to generate large waves. This provides some protection to communities on
the eastern seaboard against tsunami inundation.
Populations highlighted as being at greatest risk are those along the westward coast-
line of Ireland and the south-western regions of the UK. In particular Cornwall and
Devon, with a historically high dependence on maritime fishing industries, are at very
high risk from such deep ocean impacts. NEOSim’s tsunami model does not include
wave diffraction which will reduce the effectiveness of the tsunami’s propagation along
the English Channel. Furthermore, the Bristol Channel, which experiences the second
greatest tidal range in the world along with a large bore phenomenon, is likely to amplify
the tsunami’s amplitude as it travels up the constricting passage. Thus low-lying regions
along the Severn Estuary are likely to experience catastrophic inundation from any deep
Atlantic Ocean impact to the south west. Central coastlines of England and Wales, and
in particular the major port of Liverpool, are shown at a reduced risk. This is explained
by the protection afforded by Ireland which receives the majority of the inundation from
impact tsunamis.
482 N. J. Bailey et al.
3.3. Conclusions
The threat to the UK from small, Earth-impacting asteroids is predominantly from
Atlantic Ocean impacts to the west and south west. Atlantic impacts to the south west
of the UK represent the greatest hazard with populations along the south west coastlines
of the UK, including Cornwall, Devon and south Wales at greatest risk. A number of
major cities are situated along the Bristol Channel including Cardiff, Newport, Bristol
and Gloucester. These large cities will be at an increased risk due to the funnelling effect
of the channel. It is expected that communities along the South coast of England will
also be at increased risk due to the funnelling of the tsunami waves as they move up the
Channel. Impacts in the shallow waters of the North Sea and Irish Sea present a much re-
duced casualty potential due to the small tsunami generated. These regions help to lower
the risk to the UK by increasing the Earth’s area where impacts can occur with little
consequences. Land impacts in the UK will be very rare due to the small total land area
(approximately 0.1% of the Earth’s surface area). The consequences, while potentially
severe even for these small NEOs, tend to be localised. The typically high population
density of England in particular increases the risk of many casualties resulting from a
land impact. NEOSim estimates casualty figures for England land impacts to be from 3
to 8 million, which is comparable to London’s population of 7.5 million†. The cell that
was created over London actually shows a larger casualty signal.
While this study focuses solely on the threat to the UK, in reality it is impossible to
ignore the shared threat to other countries in the region. Particularly important in this
study is the consequence for Ireland. When considering simply the UK, Ireland acts as a
barrier to block the majority of tsunami waves from reaching the shores of England and
Wales, effectively lowering the risk. However, it would be impossible not to view such
impacts as threatening when, living in as we do in the global community, the consequences
for neighbouring countries would be significant.
The high threat to the UK from Atlantic Ocean impacts calls for increased research
into mitigating the threat. Furthermore, considering the very small risk from tectonic
and volcanic activity and even catastrophic weather (such as hurricanes), the NEO risk
is significant, potentially the UK’s single greatest natural hazard (apart from, perhaps,
the long-term effects of global warming).
4. Case study 2 – Apophis Path of Risk
The probability of impact for the asteroid 99942 Apophis has been reduced to 1 in
30,000 (equivalent probability of 3.3e−5) and thus no longer presents a significant risk.
However, the example of Apophis has been used to demonstrate the potential of studying
an object’s path of impact risk. The speed of the simulation enables studies of newly
discovered risk paths to be performed quickly to provide an assessment of the potential
consequences of the impact.
4.1. Methodology
In order to study the effect of the potential impact of Apophis, the line of risk was
required in latitude/longitude coordinates. Work by the B612 Foundation [footnote 1:
website URL] has generated a path of risk for the predicted impact in 2036 using orbit
intersections. The developed path is shown in Figure 3, which was processed to generate
the latitude and longitude coordinates for every point along the path. A fixed longitude
division was used to produce equal cell widths. NEOSim’s multi-run tool was extended
†Wikipeda London [Online] http://en.wikipedia.org/wiki/London [15 August 2006].
Tool for supporting the NEO mitigation decision making process 483
Figure 3. The curved line is the calculated path of risk for the potential impact of asteroid
99942 Apophis in 2036. This line is determined using orbit integrations.
Table 2. Assumed physical characteristics of the asteroid Apophis used as the test impactor.
Data from http://neo.jpl.nasa.gov/risk/a99942.html.
Parameter Initial value
Radius 320 m
Impactor Velocity 12.65 km/s
Object Density 2600 kg/m3
Type Monolithic
to model a series of impact cells along the path. A single 1-sigma cell was added above
and below the path to represent the uncertainty in the lateral direction.
The estimated characteristics of Apophis, given in Table 2, were used to define the
impacting object in NEOSim. This object was impacted into the centre of each cell along
the path. The casualty figure for each impact was used to shade each cell as in the first
case study.
4.2. Resul ts
The western and eastern limbs of the modeled risk path are shown in Figures 4 and
5 respectively. For publishing reasons the shading has been converted to a grey scale
with black representing most casualties and white least. The western limb of the path
stretches from the Pacific Ocean, through Kamchatka and into Central Russia. The
region of greatest casualty generation is found at the path’s most westerly point with
impacts into Kazakhstan. The second most significant region of risk is associated with
impacts into the Sea of Okhotsk, north of Japan. Here impacts generate a tsunami that
inundates the populated coastal region of Hokkaido, Japan, generating many casualties.
Land impacts over northern parts of Russia appear to produce relatively few casualties.
The eastern limb of the path show ocean impacts into the Pacific far from land gen-
erating relatively few casualties compared to those close to populated land masses such
around Central America mainland or the Hawaiian archipelago. Impacts around the coast
of Central America (up to a range of 1200 kilometres, [750 miles]) and in the Gulf of
Mexico represent the greatest threat to human life, indicated by that dark shading in
484 N. J. Bailey et al.
Figure 4. This figure presents the western limb results of the Apophis risk path. The path
stretches from Kazakhstan, through northern Russia and into the Pacific Ocean. Cells shaded
dark represent higher casualty figures generated from the impact into the centre of the cell.
Figure 5. This figure presents the eastern limb of the Apophis path of risk stretching from the
Pacific Ocean across Central America to the Cape Verde Islands in the Atlantic Ocean. Dark
shaded cells are those which generated most casualties from the impact into the centre of that
cell.
Figure 5. These casualties are generated along the populated coastal regions of Central
America.
Land impacts in this region across Central America show some interesting features.
In general these impacts generate fewer casualties than the surrounding ocean impacts.
However, particular impacts in Venezuela produce casualty estimates comparable to these
oceanic events. This variability is due to the location dependence of land impacts. For
example a direct impact into Caracas will produce many more casualties than an im-
pact only a few kilometres outside the city. Overall the land impacts in Columbia and
Venezuela generate more casualties than those in the Central American countries of
Nicaragua, Costa Rica and Panama.
The most easterly potential impact sites approach the Cape Verde Islands off the
western coast of Africa. Impacts at this end of the path generate relatively high ca-
sualty figures. The generated tsunami affects the islands themselves as well as coastal
communities of Mauritania, Senegal and Gambia.
4.3. Conclusions
NEOSim demonstrates that, at both ends of the risk path, there exists a region. Deflection
of the asteroid impact site in either direction will increase the threat to either one of these
two regions. Thus an assessment would need to be made to determine which of these
populations would most feasibly accept the increasing risk. In either case the mitigation
Tool for supporting the NEO mitigation decision making process 485
mission launched needs to be fully capable of deflecting the asteroid’s orbit to prevent the
object ‘just hitting’ and thus catastrophically affecting either of these two populations.
Of interest are the relatively low casualty estimates for Pacific and Atlantic Ocean im-
pacts far out to sea. These impacts, while appearing to have little consequence, still gen-
erate many casualties and are a significant hazard. However, relative to impacts around
Central America, these ocean events represent a significantly lower threat. Impacts close
to Hawaii are highlighted as generating more casualties. Warning time is a factor that will
dramatically reduce the number of casualties resulting from an ocean impact. Accurate
prior knowledge of the impact site combined with the long travel time of the tsunami
wave (of the order of hours) will aid in evacuation and reduce the total casualty figure.
This will help prevent scenes such as those witnessed during the Sumatran earthquake
induced tsunami of 2004.
5. The Application of the NEOSim tool to aid the Decision Making
Process
5.1. TheUKRiskCaseStudy
Studying a specific region on the Earth allows for an assessment of what form the risk
to the region or a particular country in that region takes. The NEOSim output indicates
whether the country’s population is at greater risk from ocean or land impacts and
which impacts will generate most casualties. In the case of the UK, the threat is greatest
from ocean impacts and the generated tsunami waves. Many factors are involved in
determining this threat, particularly the amount of coastline exposed to deep ocean, the
number of coastal populations and the local topography and bathymetry.
The UK is protected from many possible Atlantic Ocean impacts by the presence
of Ireland. However, it is impossible to segregate the study to one particular country
as, while the majority of Wales is protected by the presence of Ireland, the western
communities of Ireland will be severely affected. Therefore it is impossible to study the
consequences for a whole region as many countries will be affected by any impact.
The threat mitigation decision making process for the UK concerns populations are
protected from the NEO impact hazard. NEOSim demonstrates that the greatest threat is
faced by tsunami inundation from Atlantic Ocean impacts. Thus, to mitigate this threat,
research is required into methods for protecting the major cities (indicated previously)
from this inundation. One feasible method would be through a tsunami warning system
with evacuation strategy put in place for each major city. Thus, in the event of an impact,
the most densely populated regions at risk places could be evacuated to reduce the overall
loss of life.
5.2. The Apophis Risk Case Study
Despite the recent orbit refinement of the asteroid Apophis essentially removing its im-
pact threat for 2036, the experience highlights the potential for Earth to be hit unex-
pectedly by a small NEO. Thus there is a real need for rapid assessment of the im-
pact consequences and determination of the populations at greatest risk. Examining the
Apophis path of risk demonstrates the dramatic consequences for an impact. The study
highlights regions around Central America as being at greatest risk from the asteroid.
Therefore any mitigation attempt would most likely focus at moving the asteroid’s orbit
away from this region. However, any mitigation manoeuvre will always have the effect
of increasing the risk to one region over another. In this example, large populations at
risk are highlighted at each end of the path. Thus any mitigation mission will move the
asteroid towards one of these regions.
486 N. J. Bailey et al.
Decisions regarding the relative importance of one community over another are difficult
to make, but need to be made before any mitigation scenario is attempted in terms of
modifications of a NEO’s orbit, with resultant change in impact site. Such discussion
would be required at an international level as the consequences are truly multi-national.
Application of the NEOSim tool enables easy assessment of various scenarios quickly and
cheaply to inform this discussion as well as highlighting which countries will be affected.
5.3. Future Work
Work has been ongoing to develop an advanced impact simulator called NEOimpactor
which improves on NEOSim in three major areas:
•The inclusion of an infrastructure damage model to determine the economic cost of
an impact event,
•The advancement of the tsunami model from a ray tracing method to a neural
network approach to cater for diffraction around coastlines, and
•The development of a database software architecture to enable manipulation of data
layers to power new and novel investigations. One application of this database model is
to provide feedback about the effects suffered by only one pre-selected country.
Early results have been compared to the outputs of NEOSim for equivalent impact
events and also compared to real world scenarios including the Sumatra Indian Ocean
tsunami.
Acknowledgements
We would like to thank Professor R. Crowther, the principal sponsor of the project,
for his useful input and advice through out.
References
United Nations Office of the Special Envoy for Tsunami Recovery, [Online] available from:
http://www.tsunamispecialenvoy.org/country/humantoll.asp [19 July 2006]
Chyba, C.F., Thomas, P.J. & Zahnle, K.J. 1993, Nature 361, 40
Lyne, J.E., Tauber, M. & Fought, R. 1996, J. Geophys. Res. 101, 23207
Boroviˇcka, J., Popova, O.P., Nemtchinov, I.V., Spurn´y, P. & Ceplecha, Z. 1998, Astron. Astro-
phys. 334, 713
Baldwin, B. & Sheaffer, Y. 1971, J. Geophys. Res. 76, 4653
Foschini, L. 1998, Astron. Astrophys. 337, 5
Klinkrad, H., Fritsche, B. & Lips, T. 2004, International Astronautical Congress 2004, vol. IAC-
04-IAA.5.12.2.07
Collins, G.S., Melosh, H.J. & Marcus, R.A. 2004, Lunar and Planetary Laboratory, [Online]
www.lpl.arizona.edu/ImpactEffects [28 January 2005]
Glasstone S. & Dolan P.J. 1977, The Effects of Nuclear Weapons, 3rd Edition, United States
Department of Defence
Collins, G.S., Melosh, H.J. & Marcus, R.A. 2005, Meteoritics & Planetary Science 40, 817
Chesley, S.R. & Ward, S.N. 2003, AAS/Division for Planetary Sciences Meeting Abstracts, vol.
35
Ward, S.N. & Asphaug, E. 2000, Icarus 145, 64
National Aeronautics and Space Administration 2003, [Online] http://visibleearth.nasa.gov/
view rec.php?id=116 [15 March 2005]
