Constraining The Cratering Chronology of Vesta

Download as pdf or txt
Download as pdf or txt
You are on page 1of 44

a

r
X
i
v
:
1
4
0
7
.
3
3
0
3
v
1


[
a
s
t
r
o
-
p
h
.
E
P
]


1
1

J
u
l

2
0
1
4
Constraining the Cratering Chronology of Vesta
David P. OBrien
Planetary Science Institute, 1700 E. Ft. Lowell, Suite 106, Tucson, AZ 85719
obrien@psi.edu
Simone Marchi
NASA Lunar Science Institute, Southwest Research Institute, Boulder, CO
Alessandro Morbidelli
Observatoire de la Cote dAzur, CNRS, Nice, France
William F. Bottke
NASA Lunar Science Institute, Southwest Research Institute, Boulder, CO
Paul M. Schenk
Lunar and Planetary Institute, Houston, TX
Christopher T. Russell
Institute of Geophysics and Planetary Physics, University of California, Los Angeles, CA
Carol A. Raymond
Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA
Accepted to Planetary and Space Science on May 1, 2014
1 Table and 7 Figures at end of Paper
Abstract
Vesta has a complex cratering history, with ancient terrains as well as
recent large impacts that have led to regional resurfacing. Crater counts
can help constrain the relative ages of dierent units on Vestas surface,
but converting those crater counts to absolute ages requires a chronology
function. We present a cratering chronology based on the best current models
for the dynamical evolution of asteroid belt, and calibrate it to Vesta using
the record of large craters on its surface. While uncertainties remain, our
chronology function is broadly consistent with an ancient surface of Vesta as
well as other constraints such as the bombardment history of the rest of the
inner Solar System and the Ar-Ar age distribution of howardite, eucrite and
diogenite (HED) meteorites from Vesta.
Keywords:
Vesta, Asteroids, Impact Cratering, Impact Chronology
1. Introduction
Even before NASAs Dawn Mission arrived at Vesta, it was known to have
an interesting cratering history. Hubble Space Telescope images revealed a
large central-peak crater near its south pole estimated to be 460 km in diam-
eter and 13 km deep, comparable in diameter to Vesta itself (Thomas et al.,
1997). Vesta had long been suggested as the parent body of the HED
(howardite, eucrite and diogenite) meteorites on the basis of spectral and
geochemical evidence (McCord et al., 1970; Consolmagno and Drake, 1977).
Its location in the inner asteroid belt between the 3:1 mean-motion resonance
with Jupiter and the
6
secular resonance with Saturn is favorable for deliv-
ering material to near-Earth space (Wisdom, 1985; Migliorini et al., 1997),
and there is a dynamical family of related bodies (often called vestoids)
in the vicinity of Vesta, stretching toward those resonances (Binzel and Xu,
1993). The detection of the large south pole crater quite literally provided
the smoking gun evidence linking Vesta to the vestoids and HEDs. Studies
of the HED meteorites showed that Vesta formed and dierentiated early in
Solar system history (see McSween et al., 2011, for a review), such that its
surface should provide a record of some of the earliest times in Solar System
history. Thus, there were many reasons for the selection of Vesta as one of
the targets of the Dawn Mission (Russell and Raymond, 2011).
More detailed observations once Dawn arrived at Vesta revealed that
there were in fact two overlapping large impact basins in the southern hemi-
sphere, the larger and younger one known from the HST imaging now named
Rheasilvia, and the older one named Veneneia (Schenk et al., 2012). Vesta
showed a strong dichotomy between north and south, with the north being
heavily cratered and the south showing relatively few craters (Marchi et al.,
2012c). Initial crater counts of the Rheasilvia and Veneneia basins placed
their ages at approximately 1 Ga for Rheasilvia, and at least 2 Ga for
2
Veneneia, with the uncertainty in the latter being due to the fact that it
was somewhat disrupted during the formation of Rheasilvia (Marchi et al.,
2012c; Schenk et al., 2012). The formation of these two large basins erased
nearly all pre-existing craters in the southern hemisphere, but apparently left
the northern hemisphere relatively undisturbed.
Rheasilvia has sharply-dened features such as ridged terrain on the
crater oor and a prominent rim scarp (Schenk et al., 2012), suggestive of
a relatively young age. Additional evidence for the relatively recent for-
mation of Rheasilvia comes from the Vesta asteroid family, the vestoids,
that are dynamically related to Vesta and were likely ejected in a large cra-
tering event. The size distribution of these bodies is quite steep compared
to the background population (eg. Cellino et al., 1991; Tanga et al., 1999;
Nesvorn y et al., 2008), and Marzari et al. (1996, 1999) showed that the fam-
ily would have to be less than about 1 Ga old, otherwise its size distribution
would collisionally grind down and not remain as steep as observed. Finally,
the Gamma Ray and Neutron Detector (GRaND) on Dawn detected hydro-
gen on the surface of Vesta, which is interpreted to be exogenic, and found
a much lower abundance of hydrogen within the perimeter of the Rheasilvia
basin compared to the rest of the surface, suggesting that Rheasilvia formed
and reset the surface relatively recently (Prettyman et al., 2012).
Broadly, the sum of this evidence points to an ancient surface that has
been modied over time by craters, including several large and relatively re-
cent ones, such that Vestas surface units span a wide range of ages. While
relative crater densities can be used to place these dierent units in a strati-
graphic sequence, one would always like to be able to determine absolute
ages. In practice, however, this is not always a straightforward process.
Marchi et al. (2012c) and Schenk et al. (2012) used estimates of the impact
rate in the current main belt to estimate the ages of Rheasilvia and Veneneia,
which can work reasonably well for surfaces dating back to about 3 Ga or
so, as the impact rate in the main belt was likely fairly constant over that
time period (see OBrien and Sykes, 2011, for a recent review of main-belt
dynamical and collisional history). Prior to that, however, the impact rate
was likely higher, such that there is no longer a linear relationship between
age and crater density (this relation is generally termed a chronology func-
tion or chronology curve). Furthermore, as we discuss in Sec. 2, the degree
to which the impact rate has changed over time has been a matter of debate.
Here we present a chronology based on the current understanding of main-
belt dynamical history, which we then constrain using the record of the largest
3
impacts on Vesta. We rst give a brief review of the current understanding of
the impact history of the inner Solar System in Sec. 2. In Sec. 3 we present
a mathematical description of the lunar chronology as well as the lunar-
like chronology proposed for Vesta by Schmedemann et al. (2014), and in
Sec. 4 we give a derivation of our model chronology for Vesta based on main-
belt dynamics. We present these curves in a normalized form, such that
they can be used with dierent estimates of the crater production function.
Our estimate of the crater production function is then discussed in Sec. 5. In
Sec. 6 we calibrate the model chronology curve to Vesta, using measurements
of the large crater population on its surface as a constraint. We summarize
the results and discuss the implications of this work in Sec. 7.
2. Background on the Impact History of the Inner Solar System
The lunar cratering record is the most well-studied in the Solar System,
and benets from the availability of radiometrically dated samples from some
surfaces that have been studied with crater counts. Despite this ground
truth, however, there remains some ambiguity. Radiometric dating of sam-
ples of lunar impact melts (eg. Papanastassiou and Wasserburg, 1971a,b;
Wasserburg and Papanastassiou, 1971; Turner et al., 1973) showed an unex-
pected clustering of ages around 3.9 Ga and absence of earlier ages, which ran
counter to the prevailing idea that the impacts on the moon and terrestrial
planets were due the dynamically-decaying remnants of planet formation.
Tera et al. (1974) coined the term terminal Lunar cataclysm for this spike
in impact activity, occurring 600 Ma after the formation of the Moon (it
is also widely referred to as the Late Heavy Bombardment, or LHB). Later
studies of melts in lunar meteorites (eg. Cohen et al., 2000) showed a similar
clustering of impact ages around 3.9 Ga, although the distribution was more
broad than that inferred from the Apollo samples.
Hartmann (1975) suggested that the apparent lack of impact melt ages
prior to 3.9 Ga may be due to what he termed the stone-wall, in which
there was a smoothly declining cratering rate, but prior to 3.9 Ga it was
so intense that no surface rocks were able to escape resetting. While this
cannot be strictly true, given that many older lunar rocks exist that did
not experience any impact melting, more complicated processes might have
occurred in which, for example, there is a selection eect towards older impact
melt rocks being more deeply buried and younger ones more likely to be
found on the surface (Hartmann, 2003). Further modeling is necessary to
4
fully understand regolith evolution and the types of selection eects it can
introduce.
Combining crater counts with known radiometric ages can shed some light
on the problem, although it is dependent to some degree on how well we can
measure the crater production populations and how well we actually know
the age of the surfaces. For example, the age of the Nectaris basin, a key
marker in the lunar chronology system, is placed at 4.1 Ga by some groups
and 3.9 Ga by others, depending on how the distribution of ages in Apollo
16 samples is interpreted. Assuming that the age of Nectaris is represented
by the 4.1 Ga age signature in the Apollo 16 samples, Neukum and Wilhelms
(1982), (also Neukum and Ivanov, 1994; Neukum et al., 2001) proposed that
the impact rate was relatively constant back to 3.5 Ga, and exponentially
increased back to the formation of the moon at 4.5 Ga (however, as that
analysis is only based on samples dating back to 4.1 Ga, it does not provide
a solid constraint on what happened before 4.1 Ga). On the other hand, if the
age of Nectaris is assumed to be represented by the 3.9 Ga signature in the
Apollo 16 samples, the exponential decay would be much steeper and cannot
be extrapolated back to 4.5 Ga without the moon accreting more than a
lunar mass of material (Ryder, 1990; St oer and Ryder, 2001; Ryder, 2002).
This argues that the decay could not have been monotonic, and some sort of
impact cataclysm must have occurred.
The latest work suggests that none of the available samples actually origi-
nate from Nectaris, making direct dating of that basin impossible (Norman et al.,
2010). Norman et al. nd that the 4.1-4.2 Ga ages seen in the Apollo 16
samples may actually be from the Serenitatis basin, with Serenitatis ejecta
thrown to the Apollo 16 site by the Imbrium impact. The Serenitatis basin
may be only slightly older than Nectaris (eg. Fassett et al., 2012), however,
so the work of Norman et al. is consistent with a 4.1 Ga age for Nectaris.
While its fair to say that the evidence for or against the declining ux
and impact cataclysm models from cratering records and dated lunar sam-
ples is still not conclusive (see, eg. Hartmann et al., 2000; Chapman et al.,
2007, for a discussion), other lines of evidence can provide additional con-
straints. For example, Morbidelli et al. (2012) discuss the constraints from
highly siderophile element abundances in lunar rocks (eg. Walker et al., 2004;
Day et al., 2007, 2010). They use those constraints to infer a total impacting
mass of 3.5 10
19
kg on the Moon since its formation, whereas the to-
tal impacting mass implied by extending the Neukum and Wilhelms (1982),
Neukum and Ivanov (1994) and Neukum et al. (2001) bombardment history
5
back to 4.5 Ga is roughly four times larger. Bottke et al. (2007) modeled the
dynamical and collisional evolution of possible long-lived dynamical reser-
voirs in the inner Solar System to determine if a decaying population of plan-
etesimals could remain massive enough and decay slowly enough to match
the number of basins that formed on the moon between 3.8 and 4.1 Ga.
They found that, especially when collisional grinding was considered, there
was no population capable of surviving long enough to explain the forma-
tion of those basins. Both of these studies point to the need for a delayed
bombardment, regardless of whether the actual age of Nectaris is 3.9 or 4.1
Ga.
Various mechanisms have been proposed to explain how a delayed in-
crease in the impact ux could occur, such as the late formation of Neptune
(Levison et al., 2001), the destabilization of a fth terrestrial planet that
originally existed between Mars and the asteroid belt (Chambers, 2007), or
the breakup of a Vesta-sized body in the Mars-crossing population (

Cuk,
2012). The leading scenario is generally called the Nice Model as all of the
authors were working in Nice, France when it was developed (Tsiganis et al.,
2005; Gomes et al., 2005; Morbidelli et al., 2005). In this model, Jupiter and
Saturn formed interior to a mutual mean-motion resonance (MMR, initially
proposed to be the 2:1 resonance but more likely even closer together in-
side their 3:2 resonance). The scattering of remnant planetesimals in a disk
beyond the giant planets caused Jupiter and Saturn to slowly diverge until
they crossed the resonance, triggering an instability that among other things
destabilized the asteroid belt, cleared out most of the remnant planetesi-
mals, and caused the giant planets to rapidly migrate to their current orbital
conguration.
In the original simulations, Gomes et al. (2005) found that the migration
of Jupiter and Saturn across their mutual 2:1 MMR would cause resonances
to sweep through the asteroid belt, sending those asteroids onto potentially
Earth-crossing orbits and depleting the mass of the belt by a factor of 10-
20. The impact rate on the Moon would spike fairly sharply, consistent with
the lunar cataclysm scenario suggested by Tera et al. (1974). However, more
detailed work with a wider range of simulations found that the best match
to a range of Solar System constraints was obtained when Jupiter has an
encounter with one of the ice giants (Uranus or Neptune) immediately fol-
lowing the resonance crossing, causing a much more rapid migration than
would be obtained simply by planetesimal scattering (Brasser et al., 2009;
Morbidelli et al., 2009, 2010). This was termed the Jumping Jupiter sce-
6
nario. Minton and Malhotra (2009) also found that a very short timescale
of Jupiters migration was necessary in order to match constraints from the
asteroid belt, consistent with that scenario.
Morbidelli et al. (2010) showed that unlike in the original simulations
of Gomes et al. (2005) where the asteroid belt was depleted by a factor of
10-20, the depletion factor in the jumping Jupiter case was only about 2.
Minton and Malhotra (2010) showed that there may be an additional factor
of 2 depletion due to the longer term chaotic diusion of unstable asteroids
following Jupiters migration, but that still means that the mass of the as-
teroid belt prior to Jupiters migration was only about 4x its current mass,
and could not provide sucient impactors to explain all of the large basins
on the Moon, especially the oldest ones.
An extension of this scenario was proposed by Bottke et al. (2012), who
suggested that the primordial asteroid belt extended inwards to roughly 1.8
AU (compared to roughly 2.1 AU now). This innermost part of the primor-
dial asteroid belt was termed the E-Belt, and would have been stable prior
to the giant planet migration in the Nice Model, since the
6
secular reso-
nance that currently denes the inner edge of the asteroid belt would have
been located beyond the asteroid belt when Jupiter and Saturn were closer
together, interior to their mutual 3:2 MMR. When the 3:2 MMR was crossed
and the Nice Model instability occurred, the
6
would have swept through
the belt to its current location, making the E-Belt unstable. Bodies from
that population would be sent on Earth/Moon-crossing orbits and have a
relatively high collision probability with the Earth and Moon, compared to
bodies derived from further out in the asteroid belt. Furthermore, the im-
pacting population would experience a relatively slow decay with time, rather
than the sharp spike in impacts produced in the early Nice Model simula-
tions. Bottke et al. found that the impacts from E-Belt bodies would create
roughly 10 lunar basins between 3.7 and 4.1 Ga, broadly consistent with the
number of basins in the Neukum and Wilhelms (1982), Neukum and Ivanov
(1994) and Neukum et al. (2001) chronology since 4.1 Ga (with Nectaris
likely being the rst or one of the rst basins formed by the destabilization
of the E-Belt).
The E-Belt impacts would not be enough to explain all of the large basins
on the moon, however, as several times as many basins formed prior to Nec-
taris than after it. Morbidelli et al. (2012) showed that the best t to the
lunar cratering record is from a combination of a declining impact ux due
to leftover planetesimals from terrestrial planet accretion, which would have
7
formed the pre-Nectarian basins, combined with an increased ux starting
around 4.1 Ga due to Jupiters migration and the destabilization of the
asteroid belt and E-Belt that would have formed the more recent basins.
Marchi et al. (2012b) nd that the crater size-frequency distribution (SFD)
on and near the Nectaris basin is dierent from that on pre-Nectarian ter-
rains, suggesting a transition in the impactor population around that time.
While there could be multiple interpretations of the nature of the transition,
they show that it is consistent with a change in velocity of the impactors
as the more highly-excited E-Belt population begins to dominate over the
primordial impacting population around the time of Nectaris formation.
It is important to note that while there have been many historical dis-
agreements about the interpretation of the lunar cratering record, this hy-
brid model of Morbidelli et al. (2012) is essentially consistent with the lunar
chronology proposed by Neukum and Wilhelms (1982), Neukum and Ivanov
(1994) and Neukum et al. (2001) back to 4.1 Ga (ie. back to the earliest
time for which we have samples that can potentially be associated with a
specic areas of the lunar surface). The primary uncertainties lie in what
happened before that time. Such a hybrid scenario was alluded to earlier by
Hartmann et al. (2000), although never explicitly modeled.
2.1. Implications for the Asteroid Belt
It is common practice to scale the crater production rate from one planet
to another in the inner Solar System using estimates of the orbital distri-
bution of impactors (namely the Near-Earth Asteroids, NEAs) and scaling
laws for crater production (eg. Ivanov, 2001). While this approach is not
without its diculties and uncertainties, it is generally reasonable given that
the dierent planets are all targets being hit by a single source population,
the NEAs. The same approach can not necessarily be applied to scale the
early lunar crater production rate to the asteroid belt, however. The dynam-
ical history that delivers the impactors from the main belt to the terrestrial
planet region may imply a much dierent collisional history for bodies in
the asteroid belt compared to the moon and terrestrial planets. A simple
example is that if the asteroid belt was suddenly reduced in mass by a factor
of 2 at 4 Ga, with the asteroids being delivered to the terrestrial planet
region, the impact rate in the asteroid belt would merely drop by a factor of
2, while bodies in the terrestrial planet region would experience a signicant
spike in their impact rate.
8
Another important factor is that the NEA size distribution may dier
from that of the main belt, because the NEAs are derived from the main belt
in part by the action of size-dependent forces, namely the Yarkovsky eect
(see, eg. Bottke et al., 2006, for a thorough review). In fact, Strom et al.
(2005) and Marchi et al. (2009) nd that there are two dierent crater SFDs
on the Moon, one for older highland terrains, and a steeper SFD for younger
maria terrains. The implication is that impactors hitting the older terrains
have a SFD that closely matches the main belt, and were derived from it
by a size-independent process (such as the eects of Nice-Model resonance
sweeping on the belt), while the younger surfaces are cratered primarily by
NEAs, which have a steeper SFD than the main belt due to the action of the
Yarkovsky eect. Thus, the impactor SFD inferred from small craters on the
moon, which are primarily counted on younger surfaces, is not necessarily the
same as the impactor SFD in the asteroid belt itself. In addition, NEAs are
delivered primarily through resonances in the inner asteroid belt, but Vesta
can be hit by central-belt and many outer-belt bodies as well, and all of those
regions have somewhat dierent size distributions (eg. Jedicke and Metcalfe,
1998; Masiero et al., 2011).
Instead, we must look at the dynamical history of the asteroid belt itself
and use that to determine the impact rate vs. time. As described in Sec. 2.1,
the asteroid belt since 4.1 Ga likely experienced a factor of 4 depletion
due to the combined eects of the resonance sweeping during the Nice Model
instability (Morbidelli et al., 2010) and the subsequent decay of unstable
asteroids (Minton and Malhotra, 2010). The E-Belt, while it dominated the
impacts on the terrestrial planets, would have only been a relatively small
fraction of the mass in the primordial belt. Prior to 4.1 Ga, rather than
increasing back in time, the mass of the asteroid belt may have remained
relatively constant at 4 times the current mass going back several hundred
million years.
At the earliest times, immediately following the formation of the Solar
System, the asteroid belt may have had signicantly more mass than it cur-
rently does (Wetherill, 1992; Petit et al., 2001; OBrien et al., 2007), which
would have been depleted over the rst 100 Ma and led to an increased
impact rate in the belt during that time. Another possible contributor to
the early impact rate in the asteroid belt could be leftover scattered planetes-
imals from the terrestrial planet region, although the eect of such bodies on
the asteroid belt has not been fully quantied. Regardless of the source of
these earliest impactors, it is likely that there was a somewhat larger impact
9
rate in the asteroid belt immediately following the formation of the Solar
System, which would have decayed to 4 times the current rate and stayed
at that level until 4.1 Ga, then decayed to its current rate following the
destabilization of the asteroid belt and E-Belt. This depletion at 4.1 Ga
would correspond to the beginning of the Late Heavy Bombardment on the
moon.
There are two additional sources of impactors that we do not consider
here. The giant planet migration and resulting depletion of the asteroid belt
and E-belt around 4.1 is driven by the scattering a massive primordial
trans-Neptunian disk of planetesimals, some of which will cross the inner
Solar System (eg Gomes et al., 2005; Levison et al., 2009). However, there
is not strong evidence for a signicant cometary component to the impact
ux in the inner Solar System. The SFD of ancient craters surfaces on the
moon, for example, is consistent with being derived entirely from main belt
impactors (Strom et al., 2005; Marchi et al., 2009, 2012b). Because highly
siderophile elements (HSEs, like the platinum-group elements) in the lunar
crust would have been heavily depleted during its dierentiation, current
abundances of those elements are likely due to subsequent impacts. Abun-
dances of HSEs in the lunar crust show no signature of primitive, carbona-
ceous chondritic material like CI or CM chondrite (taken to be good proxies
for cometary material), suggesting that comets did not play a major role in
the lunar bombardment (eg. Kring and Cohen, 2002; Galenas et al., 2011).
Similar results are obtained from studies of actual projectile fragments in re-
golith breccias from the Apollo 16 site (Joy et al., 2012). The lack of evidence
for cometary impactors could be due to disintegration of comets once they
reach into the inner Solar System (eg. Sekanina, 1984), or perhaps the primor-
dial trans-Neptunian disk was somewhat less massive (eg. Nesvorn y et al.,
2013) than initially envisioned in Gomes et al. (2005). Broz et al. (2013)
model the formation and evolution of asteroid families under the inuence of
an LHB-era cometary bombardment, and nd that comets could potentially
lead to a signicant impact ux, although to explain the number of observed
families it is likely that 80% of comets are disrupted due to close approaches
to the Sun before they are able to impact the asteroids. Hence, we ignore for
now the possibility of cometary impactors on Vesta, although this should be
revisited in the future pending better constraints on the impact ux.
It has also been proposed that the formation of Jupiter may have scat-
tered planetesimals through the asteroid belt region and caused a Jovian
Early Bombardment (Turrini et al., 2011; Turrini, 2014). The actual bom-
10
bardment rate in that model can be quite large, potentially intense enough
to erode Vestas surface, but varies signicantly based on chosen initial con-
ditions. All of the impacts in that model would have occurred very early, and
Vesta may have taken at least several million years to dierentiate and form
its crust (see McSween et al., 2011, for a review), so it is not clear that those
impacts would actually be recorded on its surface as seen today. Hence we do
not include it in the model presented here, but note that scattered planetes-
imals during Jupiters formation may be an additional impactor population
to include in future analysis, provided it can be better constrained.
3. The Lunar and Lunar-Like Chronologies
The cumulative number of craters larger than 1 km in diameter produced
on the moon per km
2
has been estimated as
N
l
(T) = a
l
[exp (
l
T) 1] + b
l
T (1)
where T is the time measured backwards from the present in Ga (Neukum et al.,
2001). From Neukum et al. (2001), the coecients are a
l
= 5.4410
14
km
2
,
b
l
= 8.38 10
4
km
2
Ga
1
,
l
= 6.93 Ga
1
(
l
= 1/
l
= 0.144 Ga).
Marchi et al. (2009) nd slightly dierent, but still similar values of a
l
=
1.2310
15
km
2
, b
l
= 1.3010
3
km
2
Ga
1
,
l
= 7.85 Ga
1
(
l
= 1/
l
=
0.127 Ga). This chronology is based on crater counts of areas for which
radiometrically-dated samples are available (although as discussed in Sec. 2
there is some debate over the actual source regions of some of the samples),
and is not constrained prior to 4.1 Ga even though it is often plotted ex-
tending back to 4.5 Ga. Recent model-based chronologies for the moon (eg.
Morbidelli et al., 2012) are in general agreement with this chronology from
4.1 Ga to the present time.
Taking the derivative of Eq. 1 gives the dierential production rate of
craters larger than 1 km diameter per km
2
per Ga
dN
l
dT
= a
l

l
exp (
l
T) + b
l
(2)
Eq. 2 can be normalized to give a value of 1 at present time (T = 0), which
for the generally-satised case where a
l

l
b
l
is
dN

l
dT
=
a
l

l
b
l
exp (
l
T) + 1 (3)
11
The asterisk denotes the normalized form. The normalized form of the cu-
mulative expression (Eq. 1) can be found either by integrating Eq. 3 or by
dividing Eq. 1 through by b
l
N

l
(T) =
a
l
b
l
[exp (
l
T) 1] + T (4)
Note that the resulting expression has units of time. The intuitive inter-
pretation is that if the cratering rate varies in time and a given number of
craters has accumulated in time T, N

(T) is the amount of time it would


have taken to accumulate the same number of craters if the impact rate were
xed at the present value. For an impact rate that increases at earlier times,
it will always be the case that N

(T) T. Normalized forms such as this


will be useful in the subsequent sections, as they can be multiplied by the
crater production function at the current time to give the total cumulative
crater production for any value of T.
Schmedemann et al. (2014) assume that the lunar curve can be directly
scaled to Vesta. If f is the current rate of formation of craters 1 km and
larger on Vesta (per km
2
per Ga) and r is the ratio of the formation rate on
Vesta to the rate on the Moon, then r = f/b
l
and Eq. 1 is modied to give
a lunar-like curve
N
ll
(T) = r N
l
(T) = a
l
r [exp (
l
T) 1] + f T (5)
Schmedemann et al. (2014) nd that for Vesta, f = 0.01979 km
2
Ga
1
and
r = 23.62. Taking the derivative of Eq. 5 (or alternatively multiplying Eq. 2
by r) gives the dierential production rate of craters 1 km and larger per
km
2
per Ga on Vesta
dN
ll
dT
= r
dN
l
dT
= a
l

l
r exp (
l
T) + f (6)
Eq. 6 can be normalized to give a value of 1 at present time (T = 0). This
yields the same result as the normalized lunar curve (Eq. 3), as expected
since they are just scaled versions of each another
dN

ll
dT
=
a
l

l
b
l
exp (
l
T) + 1 (7)
Similarly, the normalized cumulative curve is
N

ll
(T) =
a
l
b
l
[exp (
l
T) 1] + T (8)
12
As discussed in Sec. 2.1, while it may be reasonable to scale the impact
rate and chronology function between bodies in the inner Solar System, scal-
ing it to the asteroid belt does not have the same physical basis, and in fact
it may imply a particular history for the asteroid belt that may not be plau-
sible. The simplest physical interpretation of taking the chronology function
inferred for the moon and simply scaling it to match the current impact rate
in the asteroid belt is that the asteroids are the primary impactors on other
asteroids (which is most likely true), but also that the impact rate within the
belt going back in time, and hence its total mass, directly tracks the curve
given by Eq. 6.
4. A Model Chronology Curve Based on Main-Belt Dynamics
Here we present a chronology curve for the asteroid belt based on the
recent dynamical results described in Sec. 2, assuming three main processes:
(1) A primordial depletion from an initial impacting mass M
o
(in units of
current asteroid belt mass) with timescale
pd
, which could be due to the
depletion in mass of the primordial asteroid belt itself, perhaps by embedded
planetary embryos (Wetherill, 1992; Petit et al., 2001; OBrien et al., 2007),
or alternatively by the decay of scattered leftover planetesimals from the ter-
restrial planet region; (2) Rapid loss of mass by a factor f
LHB
2 starting at
time T
LHB
with decay time constant
LHB
, triggered by the sweeping of res-
onances through the belt during a rapid phase of planetesimal-driven giant
planet migration (ie. the Nice Model, Gomes et al., 2005; Morbidelli et al.,
2010); and (3) Loss of a factor f
cd
2 by post-LHB chaotic diusion
(Minton and Malhotra, 2010). We use the subscript LHB here for the time
of the instability, as it coincides with the Late Heavy Bombardment on the
terrestrial planets.
One key parameter is the time of destabilization of the asteroid belt by
the Nice Model resonance sweeping event T
LHB
, which is often quoted as
being around 3.9 Ga based on early dating of lunar samples (eg. Tera et al.,
1974), but more recent work (eg. Bottke et al., 2012; Morbidelli et al., 2012;
Marchi et al., 2013) suggests is more likely 4.1 Ga. We nominally assume
that
LHB
=
l
(ie. that the decay follows the same prole as the lunar curve),
although this does not necessarily have to be the case. Two other important
parameters are the original mass of impactors in or otherwise aecting the
asteroid belt M
o
(relative to the current mass), and the timescale of the decay
of the primordial impact rate
pd
. We nominally use a value of
pd
= 25 Ma
13
based on the results of Bottke et al. (2005), who nd that the primordial
impact ux in the asteroid belt dropped to 2% of the initial value within
100 Ma of its formation, roughly consistent with an exponential decay
timescale of 25 Ma. M
o
is variously estimated to be on the order of hundreds
to thousands (see, eg., Weidenschilling, 1977; Wetherill, 1992; Petit et al.,
2001; Bottke et al., 2005; OBrien et al., 2007). In Sec. 6 we constrain that
value, using the record of large impact basins on the surface of Vesta.
4.1. Derivation
Minton and Malhotra (2010) nd that there has likely been a decay in
the number of asteroids in the main belt due to chaotic diusion following
the LHB resonance sweeping event, which is when the current dynamical
structure of the asteroid belt is assumed to have been established. The pa-
rameterization of the decay assumed here is based on Minton and Malhotra,
Table 1 and Eq. 4, although we rescale it to give relative number in the belt
(n
cd
(T), equal to 1 at the present time), rather than fraction remaining, and
we parameterize it in Ga rather than years.
n
cd
(T) = C
cd
((T
LHB
T)/(1 Ga) + 0.001)
0.0834
(9)
where
C
cd
= (T
LHB
/(1 Ga) + 0.001)
0.0834
(10)
The loss factor due to chaotic diusion since T = T
LHB
is then
f
cd
= C
cd
(0.001)
0.0834
2 (11)
Note that this is probably a lower limit to the loss that may occur, as we
do not explicitly include the loss that may be due to collisional grinding.
Durda et al. (1998) nd that there could be as much as a factor of 3 loss due
to collisional grinding. Bottke et al. (2005) found that there would be mod-
est depletion due to collisional grinding. While they did not quantify the
amount, it would likely be less than found by Durda et al., since Durda et al.
assume a particularly weak strength law for asteroids.
To get the normalized dierential cratering curves, we make two simpli-
fying assumptions: 1) The impact rate in the asteroid belt is directly propor-
tional to the total mass of impactors; and 2) The impact velocity is constant
with time. These together imply that the production rate of a crater of a
14
given size is directly proportional to the mass of impactors of a given size.
The rst assumption is not strictly true given that the mass of the asteroid
belt is dominated by the largest bodies and while the loss of one of them may
signicantly change the belt mass, it would have little eect on the popula-
tion of small impactors. Thus, when we talk about the mass of impactors, we
refer to an idealized distribution that smooths over stochastic variations at
the large size end. Marchi et al. (2013) nd that while the E-belt impactors
would have a higher impact velocity than main-belt asteroids, their num-
bers are relatively small compared to main-belt impactors, in part justifying
our second assumption. It is possible that scattered planetesimals from the
terrestrial planet zone, however, would have a dierent impact velocity and
collision probability with the asteroids than other main-belt impactors. As
the relative contributions of the dierent impacting populations at dierent
times are not well-constrained, we keep with assumptions 1 and 2 in the
derivations that follow.
We can get a normalized dierential cratering curve for T less than T
LHB
by combining the chaotic diusion term (Eq. 9) with an exponential decay
term for the depletion of the belt following the instability at T
LHB
dN

m<
dT
= C
LHB
exp (
LHB
T)+C
cd
((T
LHB
T)/(1 Ga)+0.001)
0.0834
(12)
where
LHB
may be similar to the exponential decay constant
l
for the
Neukum et al. (2001) curve, although not necessarily. The coecient C
LHB
of the LHB decay term is obtained by setting C
LHB
exp (
LHB
T) = (f
LHB
f
cd

f
cd
) at T = T
LHB
, which sets the belt mass at T = T
LHB
to be f
LHB
f
cd
times the present mass
C
LHB
= (f
LHB
f
cd
f
cd
)/ exp (
LHB
T
LHB
) (13)
For T greater than T
LHB
, the normalized dierential curve is based on the
normalized belt mass at T = T
LHB
(which is f
LHB
f
cd
) combined with a
primordial depletion term to account for the decrease in bombardment rate
immediately following the formation of the Solar System
dN

m>
dT
= f
LHB
f
cd
+ C
pd
exp ((T T
o
)/
pd
) (14)
where
15
C
pd
= (M
o
f
LHB
f
cd
) (15)
and T
o
is the time at which the bombardment of Vestas surface begins to
be recorded. The rst solids in the Solar System formed at 4.567 to 4.568
Ga (Bouvier and Wadhwa, 2010; Connelly et al., 2012) and the excitation of
potential impacting bodies likely did not begin until the gas disk dissipated
several Ma after that (Haisch et al., 2001; Kita et al., 2005). Similarly, it
may have taken at least several Ma for Vesta to dierentiate and form its
crust (see McSween et al., 2011, for a review). Hence, we nominally set T
o
= 4.56 Ga.
While the exponential form of the primordial decay term in Eq. 14 is
straightforward and involves the fewest free parameters, an alternative pa-
rameterization using a stretched exponential function (eg. Dobrovolskis et al.,
2007) that that includes an additional free parameter to give a longer tail
to the decay is described in Appendix A. This may prove useful in future
work that attempts a more detailed tting of the model presented here to
the results of numerical simulations of possible impactor populations.
To obtain the cumulative chronology curves, we integrate Eqns. 12 and
14. For T less than T
LHB
, the integral of Eq. 12 gives
N

m<
(T) =

T
0
dN

m<
dT

dT

=
C
LHB

LHB
[exp (
LHB
T) 1]
C
cd
0.9166 Ga
1
((T
LHB
T)/(1 Ga) + 0.001)
0.9166
+ C
1
(16)
where
C
1
=
C
cd
0.9166 Ga
1
(T
LHB
/(1 Ga) + 0.001)
0.9166
(17)
For T greater than T
LHB
, the integral of Eqns. 12 and 14 gives
N

m>
(T) =

T
LHB
0
dN

m<
dT

dT

T
T
LHB
dN

m>
dT

dT

= C
2
+ f
LHB
f
cd
(T T
LHB
) +
C
pd

pd
exp ((T T
o
)/
pd
) C
3
(18)
16
where
C
2
=
C
LHB

LHB
[exp (
LHB
T
LHB
) 1]
C
cd
0.9166 Ga
1
(0.001)
0.9166
+ C
1
(19)
and
C
3
= C
pd

pd
exp ((T
LHB
T
o
)/
pd
) (20)
Note that C
2
is equal to Eq. 16 evaluated at T = T
LHB
.
5. Crater Production Functions
The crater production function, which we denote here as F(D), gives the
number of craters of a given size D or larger per unit area per unit time
(here we use units of km
2
Ga
1
). N

(T) from Eqns. 16 and 18 (for the


model chronology) or Eq. 8 (for the lunar-like chronology) can be used to
scale the crater production function to give the cumulative crater production
for any time T. The cumulative number of craters of a given diameter or
larger produced per unit area since time T, where T = 0 is the present time,
is found by
N(D, T) = N

(T) F(D) (21)


For F(D) on Vesta we use a model production function F
m
(D) that is derived
using the model main belt size distribution from Bottke et al. (2005), which
is constrained by the observed main-belt size distribution at large sizes and a
range of other constraints such as the cosmic-ray exposure ages of meteorites
and the number of asteroid families. Crater scaling laws (Holsapple and Housen,
2007) and estimates of the main-belt impact rate are used to convert this
main-belt size distribution into a crater production function for Vesta, as out-
lined in Marchi et al. (2010, 2012a,c); Marchi et al. (2014). Schmedemann et al.
(2014) take a somewhat dierent approach by scaling the estimated lunar
crater production function (from Neukum et al., 2001) to Vesta, accounting
for the dierences in impact velocity and the relative numbers of possible
impactors. While there are some dierences between these two production
functions, they both give similar production rates of 1 km-scale craters. We
will show in Sec. 6 that the primary reason for dierent age estimates of older
terrains on Vesta lies in the dierences between the chronology functions, not
the assumed crater production function.
17
6. Applying the Model Chronology to Vesta
From the crater catalog of Marchi et al. (2012c), with recent updates to
include the north polar region, there are ve craters roughly 200 km diameter
and larger and nine craters roughly 100 km diameter and larger. Other
unpublished crater catalogs have been compiled by members of the Dawn
team, with the number of craters 100 km in diameter and larger ranging
from 6 to 11, and it is possible that numerous craters of that size were erased
over Vestas history, especially by the formation of the large Rheasilvia and
Veneneia basins. Since the crater SFD of all but the two largest craters may
have been inuenced to some degree by erasure process and therefore may
not accurately reect the production function, we only use these two largest
craters to constrain the parameters in the expressions for the chronology
curve, namely M
o
.
Figure 1 shows the result of tting the model production function F
m
(D)
to the two largest craters on Vesta (note that while craters down to 100 km
diameter are shown, they are not included in the t). We solve for the N

value such that N

F
m
(D) best matches the two large craters, giving an N

value of 27.4 Ga. In addition, we perform the same calculation assuming that
the expected value of craters of that size over Vestas lifetime is either 1 or 4
(approximately a 1-sigma range). This gives N

values of 13.7 and 54.8 Ga,


respectively. Figure 1 shows that, based on the model production function
assumed here, it is possible that numerous 200 km diameter craters could
have formed and been erased over Vestas history, and it is highly likely that
many 100 km diameter craters have been erased.
As described in Sec. 4, many of the parameters in Eqns. 16 and 18 can
be estimated from modeling or theory, and we summarize those values in
Table 1. The eects of varying those parameters from their nominal values
will be described later in this section. To estimate M
o
, the original mass of
impactors relative to the current mass of the main belt, we can solve Eqn. 18,
the cumulative model chronology function, for M
o
by setting Eq. 18 equal to
the N

values determined by the crater count ts in Fig. 1 at T = T


o
. This
gives M
o
= 836 for the model production function, with a 1-sigma range of
288 to 1932. These M
o
values assume a timescale for primordial depletion

pd
of 25 Ma, and could be larger if
pd
is smaller. The current mass of the
main belt is estimated to be roughly 0.0006 Earth masses (Krasinsky et al.,
2002), so these M
o
values correspond to a possible range of 0.17 to 1.16
Earth masses for the primordial impacting mass, consistent with estimates
18
of amount of mass in the primordial asteroid belt (eg., Weidenschilling, 1977;
Wetherill, 1992; Petit et al., 2001; Bottke et al., 2005; OBrien et al., 2007).
Figure 2 shows a comparison of the normalized dierential production rate
of craters dN

/dT in the lunar-like chronology (Eq. 7 from Sec. 3) and the


model chronology (Eqns. 12 and 14 from Sec. 4), for the nominal parameter
values given above. Also shown is a constant linear production rate curve.
Note that the lunar-like curve prior to 4.1 Ga is an extrapolation, since the
lunar chronology on which it is based is only constrained back to 4.1 Ga.
While the lunar-like curve increases rapidly prior to 3 Ga, the model curve
rst increases to 4 times its current value prior to 4.1 Ga (the time of
destabilization of the asteroid belt and E-Belt), then increases again going
back to T
o
, due to the larger primordial mass of the asteroid belt, and/or
scattered planetesimals from the terrestrial planet region that can strike the
asteroids. For any given time prior to 3 Ga, the lunar-like curve implies a
much larger impact rate in the asteroid belt than the model curve.
The integrals of the dierential curves from Fig. 2 give the normalized
chronology functions N

(T). Figure 3 shows the a comparison of the nor-


malized chronology functions N

(T) in the lunar-like chronology (Eq. 8 from


Sec. 3) and the model chronology (Eqns. 16 and 18 from Sec. 4), for the
nominal parameter values given above, along with a linear chronology curve.
Note that while the model dierential curve in Fig. 2 shows sharp increases
near T = 4 and 4.5 Ga, the corresponding increases in the cumulative curve
are more gradual, because the dierential curves are integrated from T = 0
back to the time in question rather than just over a small interval where the
changes occur. The second plot in Fig. 3 shows an approximately 1-sigma
range based on uncertainties in the estimate of the initial impacting mass
M
o
.
Figure 4 shows the chronology curves for Vesta in terms of absolute num-
bers of craters larger than 1 km, using Eq. 21 along with the normalized
N

(T) curves from Fig. 3 and the crater production function F


m
(D). The
lunar-like curve of Schmedemann et al. (2014) (our Eq. 5) is also shown. The
two curves are roughly the same for T < 3.5 Ga, since the production rate
of 1 km craters is roughly the same in the lunar-like and model production
functions. However, they diverge signicantly prior to 3.5 Ga, because of the
divergence of the chronology functions (Fig. 3). For any given time prior to
3.5 Ga, the lunar-like curve implies that a given surface of that age would
have a higher crater density than is implied by the model chronology (or
alternatively, the lunar-like chronology gives a younger age than the model
19
chronology for a surface with a given crater density).
For the lunar-like curve in Fig. 4, crater densities as high as predicted
around 4.1 Ga would not actually be possible, as the surface would become
saturated and the actual observed crater numbers would lie below the produc-
tion curve (eg. Gault, 1970). For the model chronology curves, it is possible
that the levels achieved near T = T
o
would be close to the empirical satura-
tion level as well, and that could potentially aect the conversion of crater
density to absolute age. This is discussed in further detail in Marchi et al.
(2012c).
It is illustrative to compare our model chronology to the lunar chronol-
ogy. Figure 5 shows the crater production rates and cumulative chronology
curves in terms of 1 km diameter craters for both the moon and Vesta, where
the lunar curves are given by Eqns. 1 and 2 and the Vesta curves are given
by Eqns. 12, 14, 16 and 18, scaled appropriately by the production rate
of 1 km diameter craters from the Vesta model production function F
m
(D)
(approx. 0.0190 km
2
Ga
1
). Note in particular that while the impact rate
curve for Vesta lies above the lunar curve by a factor of 20 for T < 3
Ga, it is about a factor of 10 lower around 4 Ga, the nominal time of
the LHB. The reason these curves can follow substantially dierent proles
(rather than simply being scaled versions of one another, was briey alluded
to in Sec. 2.1. Around 4.1 Ga in our model chronology, the asteroid belt is
depleted by a factor of 4 through a combination of resonance sweeping and
subsequent chaotic diusion of bodies out of the belt. While this only leads
to a decrease in impact rate of factor of 4 within the asteroid belt (and
hence on Vesta), the several asteroid belts worth of mass that are rapidly
ejected from the asteroid belt lead to a huge increase in the impact rate on
the moon and terrestrial planets. For T < 3 Ga, after this inux of material
has decayed, impacts on the moon occur at a much lower rate, set by the
rate at which bodies slowly leak out of the asteroid belt to become NEAs.
While we do not explicitly calculate the lunar impact rate from the dynam-
ical assumptions of our model chronology, we note that Bottke et al. (2012)
and Morbidelli et al. (2012) have shown that the impact rate on the moon
implied by the dynamical scenario on which we build our model chronology
is generally consistent with that of Neukum et al. (2001) over the last 4.1
Ga, which is what we plot for the lunar curves in Fig. 5.
Figure 6 shows how the model chronology curve is aected by varying
several of the key parameters in Eqns. 16 and 18 from their nominal val-
ues given in Table 1. f
LHB
is varied from 1 to 10, relative to its nominal
20
value of 2. With f
LHB
= 1, there is no depletion and hence no change in
the impact rate at t
LHB
, the impact rate only changes during the early pri-
mordial depletion phase. f
LHB
= 10 is consistent with the early Nice Model
simulations (Gomes et al., 2005), although subsequent models have revised
that value downward to the current estimate of 2 (Morbidelli et al., 2010).
The primordial depletion timescale
pd
is varied by a factor of 2 around its
nominal value of 25 Ma (ie. a range of 12.5 Ma to 50 Ma). In both cases, the
range of variation in the initial N

(T) value at T
o
= 4.56 Ga is comparable
to the range of variation due to the 1-sigma range in M
o
estimates, although
the shape of the curves may be signicantly aected (particularly in the case
of high f
LHB
).
Figure 7 shows our model chronology function applied to actual crater
counts on Vesta. Two regions are chosen, the oor of Rheasilvia, and the
highly-cratered terrains (HCTs) in the northern hemisphere identied by
Marchi et al. (2012c) and subsequently revised. The model production func-
tion F
m
(D) is t to the crater counts for these two regions, such that N

F
m
(D)
best matches the observed crater size-frequency distributions. The resulting
N

value can be related to the surface age with the model chronology func-
tion in Fig. 3 (where the model chronology function is from Eqns. 16 and 18,
with the values in Table 1 and M
o
= 836).
We nd an age of Rheasilvia of 1 Ga, consistent with the earlier es-
timates by Marchi et al. (2012c). Assuming that the craters on the HCT
regions are primary and tting to the small end (D < 8 km) of the crater
SFD, we nd an age of 4.3 Ga, consistent with them being amongst the
oldest terrains on Vesta. It is likely that the HCTs are saturated or close to
saturation, so the true age could be closer to 4.5 Ga. Because the larger
craters on the HCTs lie below the production curve while smaller craters
are saturated, it has been suggested that the impacting population on the
HCT regions was primarily a steep secondary population from the Rheasil-
via impact. However, more detailed analysis of those regions nds that few
craters in that area show morphological signs of being secondaries (which
would have formed at less than Vestas escape velocity of 350 m/s). The
most likely explanation is that the HCTs are not really a specic geologi-
cal unit, but are regions chosen specically because they have the highest
crater density, and the model production function in the 3-20 km range of
craters on the HCTs has a slope somewhat shallower than -2. With a slope
shallower than -2, erasure by large craters dominates and the crater popu-
lation at smaller sizes mirrors the slope of the production population, with
21
the overall crater density dependent on stochastic erasure by large craters
(eg. Chapman and McKinnon, 1986; Melosh, 1989; Richardson, 2009). In
that case, the regions chosen as the HCTs, with the highest density of small
craters, will be the ones that avoided experiencing many larger impacts,
hence the low crater counts for D > 8 km.
7. Discussion
We present a cratering chronology for Vesta based on the best current
understanding of the dynamical evolution of asteroid belt and the impact
history of the inner Solar System, and we calibrate it to Vesta using the
record of large craters on its surface. In this chronology, Vesta would have
experienced two main stages: 1) An early declining impact ux due to the
dynamical depletion of the asteroid belt and/or scattered planetesimals from
the terrestrial planet region; and 2) A secondary decrease in impact ux
around 4 Ga due to the sweeping of resonances through the asteroid belt
driven by giant planet migration, as well as the chaotic diusion of unstable
asteroids out of the main belt. While there are necessarily some simplica-
tions and uncertainties in the model, our chronology is currently the best
educated guess we have as to Vestas impact history, and is broadly consis-
tent with estimates of the initial amount of mass present in the asteroid belt
and with Vesta having an ancient cratered surface.
The meteorite record provides additional evidence for the collisional and
dynamical history of the asteroid belt, and of Vesta. The HED meteorites
from Vesta, as well as the H chondrites, show a broad peak in impact-reset
ages from 3 to 4.2 Ga and a relative lack of ages between 4.2 and 4.5
Ga (Bogard, 1995; Bogard and Garrison, 2003; Swindle et al., 2009; Bogard,
2011). Marchi et al. (2013) show that this is likely a consequence of the
destabilization of the E-Belt, located at the inner edge of the primordial
asteroid belt. While those bodies would not lead to a signicantly increased
impact rate in the asteroid belt, they would lead to an increase in high-
velocity impacts, which would be capable of resetting Ar-Ar ages on asteroid
surfaces. Those surfaces would be sampled by later impacts that would
eject material that could eventually evolve onto Earth-crossing orbits. In the
case of Vesta, Rheasilvia is the likely candidate, as it appears to be the last
major impact, and could have sampled material from numerous pre-existing
impact basins with a range of ages. Our chronology, which includes the
destabilization of the E-Belt and asteroid belt at 4.1 Ga, with a plateau
22
in the impact rate prior to that leading back to the primordial spike in the
impact rate, is consistent with this picture.
We nd an age of 1 Ga for Rheasilvia using our model chronology, con-
sistent with previous estimates (Marchi et al., 2012c). This is consistent with
several other lines of evidence as well. Rheasilvia has a quite fresh appear-
ance relative to all other ancient basins. The size distribution of the Vesta
family members, or vestoids, is quite steep compared to the background
population. Marzari et al. (1996, 1999) showed that its size distribution
would collisionally grind down if the family were older than 1 Ga, sug-
gesting its relatively recent formation. A young age is also consistent with
the much lower abundance of Hydrogen within Rheasilvia Basin compared
to the rest of the surface, as found by the Gamma Ray and Neutron Detector
(GRaND) on Dawn (Prettyman et al., 2012) and interpreted as exogenic in
origin. Nesvorn y et al. (2008) estimate that the Vesta family would have to
be 1-2 Ga old in order to explain the presence of fugitive vestoids that
lie outside the bounds of the main family, although if radiation forces like
the YORP eect that can lead to a non-random distribution of spin axes are
taken into account, the age is closer to 1 Ga (D. Nesvorny, pers. comm.). It
may seem strange that the largest crater on Vesta would be so young, where
we dene young here as forming in the last 3 Ga, during which time the
impact rate is roughly the same as the current rate. We nd that while it
may be a low-probability event, it is still plausible given the impact history
implied by the model chronology. Assuming that one Rheasilvia-sized basin
formed over the lifetime of Vesta, the probability that it formed in the last
3 Ga is roughly estimated as N

(T = 3 Ga)/N

(T = T
o
) = 0.134, or 13.4%,
assuming the nominal model chronology parameters.
As described in Sec. 2.1, we do not account for several possible ad-
ditional sources of impactors, including cometary inux during the LHB
(Gomes et al., 2005; Levison et al., 2009), or scattered planetesimals dur-
ing the formation of Jupiter (Turrini, 2014), and in 4.1 we assume that the
impact rate on Vesta is directly proportional to the total impacting mass and
that the impact velocity is constant. We also chose a value of
pd
for the pri-
mordial decay of the impact ux of 25 Ma, based on models of the depletion
of the asteroid belt through perturbations by primordial planetary embryos.
If the depletion occurred through a dierent process (eg. the Grand Tack of
Walsh et al. (2011)), the depletion would have been faster, and the resulting
primordial impacting mass M
o
in our model would have to be correspond-
ingly higher. While all of these are reasonable assumptions to make, given
23
the uncertainties involved, it does mean that the chronology presented here
should not be taken as the nal word for calculating ages on Vestas sur-
face, and should certainly not be used to calculate 3-signicant-gure ages
as is common in the crater counting literature. Rather, we provide a general
framework for Vestan chronology that can be rened as further constraints
may become available, and which can be used to provide reasonable age esti-
mates in the context of our current understanding of Solar System evolution.
Because of the uncertainties involved, we stress that in presenting crater
counts for Vesta, the age should not be the only piece of information given.
Rather, since any absolute age calculation is model-dependent, a quantita-
tive measure of the crater density (such as the number of craters larger than
a certain size) should be given, as well as any other relevant information
such as the assumed impacting population, so that the reader can do their
own comparison between the results of dierent groups who may be making
dierent assumptions.
We contrast our approach with that of Schmedemann et al. (2014), who
apply the lunar chronology curve of Neukum et al. (2001) to Vesta by scaling
it to Vestas current impact rate in the main belt. As discussed in Secs. 2.1
and 3, while it may be reasonable to scale the curve between dierent bodies
in the inner Solar System since they are all impacted by a common popula-
tion, the NEAs, it is not necessarily true that one can simply scale it to the
asteroid belt. The dynamical history that delivers the impactors from the
main belt to the NEA population may imply signicantly dierent collisional
histories for bodies in the asteroid belt compared to the moon and terres-
trial planets. Fig. 5 illustrates this clearly, as the Lunar curve shown there
is broadly consistent (at least back to 4.1 Ga) with the dynamical scenario
(eg. Bottke et al., 2012; Morbidelli et al., 2012) that is the basis of our model
chronology curve for Vesta, yet the two chronology curves dier signicantly
from one another.
Taking the chronology function inferred for the moon and scaling it to
match the current impact rate in the asteroid belt, as done by Schmedemann et al.
(2014), is not based on any specic physical scenario, but the simplest phys-
ical interpretation of such a scaling is that the primary impactors in the
asteroid belt are other asteroids (which is likely true), and that the impact
rate in the belt going back in time, and hence its total mass, directly tracks
the curve given by Eq. 6 (shown in Fig. 2). While lunar chronology itself
is not constrained prior to 4.1 Ga, the scaling used to give the lunar-like
chronology for Vesta may imply an impact rate that continues to follow the
24
dashed curve in Fig. 2 prior to 4.1 Ga, possibly giving an unreasonably large
primordial impact rate and the production of signicantly more large basins
that the two seen on Vesta today. While there are potentially other scenarios
that could avoid this issue, we caution that the full implications and physical
interpretation of using a lunar-like chronology in the asteroid belt should be
more thoroughly explored if it is going to be employed in future work.
Dawn will arrive at Ceres in 2015, giving opportunity to test our model
chronology and perhaps rene it with further constraints, such as the record
of large impact basins on Ceres. However, Ceres is likely a very dierent
body than Vesta and may pose several diculties for the interpretation of its
cratering record. Given that Ceres may contain a signicant fraction of water
ice (eg. McCord et al., 2011), the scaling law for crater formation will likely
dier from that of Vesta, although that can be at least partially accounted
for. More problematic is that, as discussed by Bland (2013), signicant water
ice could mean that many craters, especially those in the warmer equatorial
regions, could viscously relax on short timescales, leaving little or no record
of their existence.
Acknowledgments
D. P. OBrien is supported by grant NNX10AR21G from NASAs Dawn at
Vesta Participating Scientist Program. The Dawn mission to asteroid Vesta
and dwarf planet Ceres is managed by JPL, for NASAs Science Mission
Directorate, Washington, DC. UCLA is responsible for overall Dawn mission
science. We thank David Minton and an anonymous referee for their helpful
reviews and comments.
Appendix A. Alternative Parameterization of Primordial Deple-
tion
An alternative parameterization of the primordial decay uses a stretched
exponential function (eg. Dobrovolskis et al., 2007) with a decay timescale

pd
modied by an exponent
pd
less that 1 that gives a longer tail to the
decay. While we do not use this form here, we include it for completeness,
as it may provide useful for tting the model presented here to the results of
numerical simulations of the primordial decay of the impactor population.
Using this new form, Eq. 14 for the normalized dierential curve for T
greater than T
LHB
becomes
25
dN

m>
dT
= f
LHB
f
cd
+ C
pd
exp

T
o
T

pd

pd

(A.1)
For T greater than T
LHB
, the integral of Eqns. 12 and A.1 gives the new
normalized cumulative curve
N

m>
(T) =

T
LHB
0
dN

m<
dT

dT

T
T
LHB
dN

m>
dT

dT

= C
2
+ f
LHB
f
cd
(T T
LHB
) +
C
pd

pd

pd

pd
,

T
o
T

pd

pd

C
3
(A.2)
where is the upper incomplete Gamma function, dened as
(s, x) =


s
x
s1
e
x
dx (A.3)
C
1
and C
2
remain the same as in Eqns. 17 and 19, and Eq. 20 becomes
C
3
=
C
pd

pd

pd

pd
,

T
o
T
LHB

pd

pd

(A.4)
26
References
Binzel, R. P. and Xu, S. (1993). Chips o of asteroid 4 Vesta - Evidence for
the parent body of basaltic achondrite meteorites. Science, 260, 186191.
Bland, M. T. (2013). Predicted crater morphologies on Ceres: Probing in-
ternal structure and evolution. Icarus, 226, 510521.
Bogard, D. (1995). Impact ages of meteorites: A synthesis. Meteoritics, 30,
244268.
Bogard, D. D. (2011). K-Ar ages of meteorites: Clues to parent-body thermal
histories. Chemie der Erde / Geochemistry, 71, 207226.
Bogard, D. D. and Garrison, D. H. (2003). 39Ar-40Ar ages of eucrites and
thermal history of asteroid 4 Vesta. Meteoritics and Planetary Science,
38, 669710.
Bottke, W. F., Durda, D. D., Nesvorn y, D., Jedicke, R., Morbidelli, A.,
Vokrouhlick y, D., and Levison, H. F. (2005). Linking the collisional history
of the main asteroid belt to its dynamical excitation and depletion. Icarus,
179, 6394.
Bottke, W. F., Levison, H. F., Nesvorn y, D., and Dones, L. (2007). Can
planetesimals left over from terrestrial planet formation produce the lunar
Late Heavy Bombardment? Icarus, 190, 203223.
Bottke, W. F., Vokrouhlick y, D., Minton, D., Nesvorn y, D., Morbidelli, A.,
Brasser, R., Simonson, B., and Levison, H. F. (2012). An Archaean heavy
bombardment from a destabilized extension of the asteroid belt. Nature,
485, 7881.
Bottke, Jr., W. F., Vokrouhlick y, D., Rubincam, D. P., and Nesvorn y, D.
(2006). The Yarkovsky and Yorp Eects: Implications for Asteroid Dy-
namics. Annual Review of Earth and Planetary Sciences, 34, 157191.
Bouvier, A. and Wadhwa, M. (2010). The age of the Solar System redened
by the oldest Pb-Pb age of a meteoritic inclusion. Nature Geoscience, 3,
637641.
27
Brasser, R., Morbidelli, A., Gomes, R., Tsiganis, K., and Levison, H. F.
(2009). Constructing the secular architecture of the solar system II: the
terrestrial planets. A&A, 507, 10531065.
Broz, M., Morbidelli, A., Bottke, W. F., Rozehnal, J., Vokrouhlick y, D., and
Nesvorn y, D. (2013). Constraining the cometary ux through the asteroid
belt during the late heavy bombardment. A&A, 551, A117.
Cellino, A., Zappala, V., and Farinella, P. (1991). The size distribution of
main-belt asteroids from IRAS data. MNRAS, 253, 561574.
Chambers, J. E. (2007). On the stability of a planet between Mars and
the asteroid belt: Implications for the Planet V hypothesis. Icarus, 189,
386400.
Chapman, C. R. and McKinnon, W. B. (1986). Cratering of planetary satel-
lites. In J. A. Burns and M. S. Matthews, editors, Satellites, pages 492580.
University of Arizona Press, Tucson, AZ.
Chapman, C. R., Cohen, B. A., and Grinspoon, D. H. (2007). What are the
real constraints on the existence and magnitude of the late heavy bom-
bardment? Icarus, 189, 233245.
Cohen, B. A., Swindle, T. D., and Kring, D. A. (2000). Support for the Lunar
Cataclysm Hypothesis from Lunar Meteorite Impact Melt Ages. Science,
290, 17541756.
Connelly, J. N., Bizzarro, M., Krot, A. N., Nordlund,

A., Wielandt, D., and
Ivanova, M. A. (2012). The Absolute Chronology and Thermal Processing
of Solids in the Solar Protoplanetary Disk. Science, 338, 651655.
Consolmagno, G. J. and Drake, M. J. (1977). Composition and evolu-
tion of the eucrite parent body - Evidence from rare earth elements.
Geochim. Cosmochim. Acta, 41, 12711282.

Cuk, M. (2012). Chronology and sources of lunar impact bombardment.


Icarus, 218, 6979.
Day, J. M. D., Pearson, D. G., and Taylor, L. A. (2007). Highly Siderophile
Element Constraints on Accretion and Dierentiation of the Earth-Moon
System. Science, 315, 217219.
28
Day, J. M. D., Walker, R. J., James, O. B., and Puchtel, I. S. (2010). Osmium
isotope and highly siderophile element systematics of the lunar crust. Earth
and Planetary Science Letters, 289, 595605.
Dobrovolskis, A. R., Alvarellos, J. L., and Lissauer, J. J. (2007). Lifetimes of
small bodies in planetocentric (or heliocentric) orbits. Icarus, 188, 481
505.
Durda, D. D., Greenberg, R., and Jedicke, R. (1998). Collisional Models
and Scaling Laws: A New Interpretation of the Shape of the Main-Belt
Asteroid Size Distribution. Icarus, 135, 431440.
Fassett, C. I., Head, J. W., Kadish, S. J., Mazarico, E., Neumann, G. A.,
Smith, D. E., and Zuber, M. T. (2012). Lunar impact basins: Stratigraphy,
sequence and ages from superposed impact crater populations measured
from Lunar Orbiter Laser Altimeter (LOLA) data. Journal of Geophysical
Research (Planets), 117, E00H06.
Galenas, M. G., Gerasimenko, I., James, O. B., Puchtel, I. S., and Walker,
R. J. (2011). Continued Study of Highly Siderophile Element Characteris-
tics of Apollo 17 Impact Melt Breccias. In Lunar and Planetary Institute
Science Conference Abstracts, volume 42 of Lunar and Planetary Institute
Science Conference Abstracts, page 1413.
Gault, D. E. (1970). Saturation and equilibrium conditions for impact cra-
tering on the lunar surface: Criteria and implications. Radio Science, 5,
273291.
Gomes, R., Levison, H. F., Tsiganis, K., and Morbidelli, A. (2005). Origin of
the cataclysmic Late Heavy Bombardment period of the terrestrial planets.
Nature, 435, 466469.
Haisch, Jr., K. E., Lada, E. A., and Lada, C. J. (2001). Disk Frequencies
and Lifetimes in Young Clusters. ApJ, 553, L153L156.
Hartmann, W. K. (1975). Lunar cataclysm - A misconception. Icarus, 24,
181187.
Hartmann, W. K. (2003). Megaregolith evolution and cratering cataclysm
modelsLunar cataclysm as a misconception (28 years later). Meteoritics
and Planetary Science, 38, 579593.
29
Hartmann, W. K., Ryder, G., Dones, L., and Grinspoon, D. (2000). The
Time-Dependent Intense Bombardment of the Primordial Earth/Moon
System. In R. M. Canup, K. Righter, and et al., editors, Origin of the
Earth and Moon, pages 493512. University of Arizona Press, Tucson, AZ.
Holsapple, K. A. and Housen, K. R. (2007). A crater and its ejecta: An
interpretation of Deep Impact. Icarus, 191, 586597.
Ivanov, B. A. (2001). Mars/Moon Cratering Rate Ratio Estimates.
Space Sci. Rev., 96, 87104.
Jedicke, R. and Metcalfe, T. S. (1998). The Orbital and Absolute Magnitude
Distributions of Main Belt Asteroids. Icarus, 131, 245260.
Joy, K. H., Zolensky, M. E., Nagashima, K., Huss, G. R., Ross, D. K.,
McKay, D. S., and Kring, D. A. (2012). Direct Detection of Projectile
Relics from the End of the Lunar Basin-Forming Epoch. Science, 336,
14261429.
Kita, N. T., Huss, G. R., Tachibana, S., Amelin, Y., Nyquist, L. E., and
Hutcheon, I. D. (2005). Constraints on the Origin of Chondrules and CAIs
from Short-lived and Long-Lived Radionuclides. In A. N. Krot, E. R. D.
Scott, and B. Reipurth, editors, Chondrites and the Protoplanetary Disk,
volume 341 of Astronomical Society of the Pacic Conference Series, pages
558587.
Krasinsky, G. A., Pitjeva, E. V., Vasilyev, M. V., and Yagudina, E. I. (2002).
Hidden Mass in the Asteroid Belt. Icarus, 158, 98105.
Kring, D. A. and Cohen, B. A. (2002). Cataclysmic bombardment through-
out the inner solar system 3.9-4.0 Ga. Journal of Geophysical Research
(Planets), 107, 5009.
Levison, H. F., Dones, L., Chapman, C. R., Stern, S. A., Duncan, M. J., and
Zahnle, K. (2001). Could the Lunar Late Heavy Bombardment Have
Been Triggered bythe Formation of Uranus and Neptune? Icarus, 151,
286306.
Levison, H. F., Bottke, W. F., Gounelle, M., Morbidelli, A., Nesvorn y, D.,
and Tsiganis, K. (2009). Contamination of the asteroid belt by primordial
trans-Neptunian objects. Nature, 460, 364366.
30
Marchi, S., Mottola, S., Cremonese, G., Massironi, M., and Martellato, E.
(2009). A New Chronology for the Moon and Mercury. AJ, 137, 4936
4948.
Marchi, S., Barbieri, C., K uppers, M., Marzari, F., Davidsson, B., Keller,
H. U., Besse, S., Lamy, P., Mottola, S., Massironi, M., and Cremonese, G.
(2010). The cratering history of asteroid (2867) Steins. Planet. Space Sci.,
58, 11161123.
Marchi, S., Massironi, M., Vincent, J.-B., Morbidelli, A., Mottola, S.,
Marzari, F., K uppers, M., Besse, S., Thomas, N., Barbieri, C., Naletto,
G., and Sierks, H. (2012a). The cratering history of asteroid (21) Lutetia.
Planet. Space Sci., 66, 8795.
Marchi, S., Bottke, W. F., Kring, D. A., and Morbidelli, A. (2012b). The
onset of the lunar cataclysm as recorded in its ancient crater populations.
Earth and Planetary Science Letters, 325, 2738.
Marchi, S., McSween, H. Y., OBrien, D. P., Schenk, P., De Sanctis, M. C.,
Gaskell, R., Jaumann, R., Mottola, S., Preusker, F., Raymond, C. A.,
Roatsch, T., and Russell, C. T. (2012c). The Violent Collisional History
of Asteroid 4 Vesta. Science, 336, 690694.
Marchi, S., Bottke, W. F., Cohen, B. A., W unnemann, K., Kring, D. A.,
McSween, H. Y., de Sanctis, M. C., OBrien, D. P., Schenk, P., Raymond,
C. A., and Russell, C. T. (2013). High-velocity collisions from the lunar
cataclysm recorded in asteroidal meteorites. Nature Geoscience, 6, 303
307.
Marchi, S., Bottke, W., OBrien, D., Schenk, P., Mottola, S., Sanctis, M. D.,
Kring, D., Williams, D., Raymond, C., and Russell, C. (2014). Small crater
populations on Vesta. Planetary and Space Science, in press.
Marzari, F., Cellino, A., Davis, D. R., Farinella, P., Zappala, V., and Vanzani,
V. (1996). Origin and evolution of the Vesta asteroid family. A&A, 316,
248262.
Marzari, F., Farinella, P., and Davis, D. R. (1999). Origin, Aging, and Death
of Asteroid Families. Icarus, 142, 6377.
31
Masiero, J. R., Mainzer, A. K., Grav, T., Bauer, J. M., Cutri, R. M., Dailey,
J., Eisenhardt, P. R. M., McMillan, R. S., Spahr, T. B., Skrutskie, M. F.,
Tholen, D., Walker, R. G., Wright, E. L., DeBaun, E., Elsbury, D., Gau-
tier, IV, T., Gomillion, S., and Wilkins, A. (2011). Main Belt Asteroids
with WISE/NEOWISE. I. Preliminary Albedos and Diameters. ApJ, 741,
68.
McCord, T. B., Adams, J. B., and Johnson, T. V. (1970). Asteroid Vesta:
Spectral Reectivity and Compositional Implications. Science, 168, 1445
1447.
McCord, T. B., Castillo-Rogez, J., and Rivkin, A. (2011). Ceres: Its
Origin, Evolution and Structure and Dawns Potential Contribution.
Space Sci. Rev., 163, 6376.
McSween, H. Y., Mittlefehldt, D. W., Beck, A. W., Mayne, R. G., and
McCoy, T. J. (2011). HED Meteorites and Their Relationship to the
Geology of Vesta and the Dawn Mission. Space Sci. Rev., 163, 141174.
Melosh, H. J. (1989). Impact cratering: A geologic process. Oxford University
Press, New York.
Migliorini, F., Morbidelli, A., Zappala, V., Gladman, B. J., Bailey, M. E.,
and Cellino, A. (1997). Vesta fragments from v6 and 3:1 resonances: Impli-
cations for V-type NEAs and HED meteorites. Meteoritics and Planetary
Science, 32, 903916.
Minton, D. A. and Malhotra, R. (2009). A record of planet migration in the
main asteroid belt. Nature, 457, 11091111.
Minton, D. A. and Malhotra, R. (2010). Dynamical erosion of the asteroid
belt and implications for large impacts in the inner Solar System. Icarus,
207, 744757.
Morbidelli, A., Levison, H. F., Tsiganis, K., and Gomes, R. (2005). Chaotic
capture of Jupiters Trojan asteroids in the early Solar System. Nature,
435, 462465.
Morbidelli, A., Brasser, R., Tsiganis, K., Gomes, R., and Levison, H. F.
(2009). Constructing the secular architecture of the solar system. I. The
giant planets. A&A, 507, 10411052.
32
Morbidelli, A., Brasser, R., Gomes, R., Levison, H. F., and Tsiganis, K.
(2010). Evidence from the Asteroid Belt for a Violent Past Evolution of
Jupiters Orbit. AJ, 140, 13911401.
Morbidelli, A., Marchi, S., Bottke, W. F., and Kring, D. A. (2012). A
sawtooth-like timeline for the rst billion years of lunar bombardment.
Earth and Planetary Science Letters, 355, 144151.
Nesvorn y, D., Roig, F., Gladman, B., Lazzaro, D., Carruba, V., and Mothe-
Diniz, T. (2008). Fugitives from the Vesta family. Icarus, 193, 8595.
Nesvorn y, D., Vokrouhlick y, D., and Morbidelli, A. (2013). Capture of Tro-
jans by Jumping Jupiter. ApJ, 768, 45.
Neukum, G. and Ivanov, B. A. (1994). Crater Size Distributions and Im-
pact Probabilities on Earth from Lunar, Terrestrial-planet, and Asteroid
Cratering Data. In T. Gehrels, M. S. Matthews, and A. M. Schumann,
editors, Hazards Due to Comets and Asteroids, pages 359416. University
of Arizona Press, Tucson, AZ.
Neukum, G. and Wilhelms, D. E. (1982). Ancient Lunar Impact Record. In
Lunar and Planetary Institute Science Conference Abstracts, volume 13 of
Lunar and Planetary Inst. Technical Report, pages 590591.
Neukum, G., Ivanov, B. A., and Hartmann, W. K. (2001). Cratering Records
in the Inner Solar System in Relation to the Lunar Reference System.
Space Sci. Rev., 96, 5586.
Norman, M. D., Duncan, R. A., and Huard, J. J. (2010). Imbrium provenance
for the Apollo 16 Descartes terrain: Argon ages and geochemistry of lunar
breccias 67016 and 67455. Geochim. Cosmochim. Acta, 74, 763783.
OBrien, D. P. and Sykes, M. V. (2011). The Origin and Evolution of the
Asteroid BeltImplications for Vesta and Ceres. Space Sci. Rev., 163,
4161.
OBrien, D. P., Morbidelli, A., and Bottke, W. F. (2007). The primordial
excitation and clearing of the asteroid beltRevisited. Icarus, 191, 434
452.
33
Papanastassiou, D. A. and Wasserburg, G. J. (1971a). Lunar chronology
and evolution from Rb-Sr studies of Apollo 11 and 12 samples. Earth and
Planetary Science Letters, 11, 3762.
Papanastassiou, D. A. and Wasserburg, G. J. (1971b). Rb-Sr ages of igneous
rocks from the Apollo 14 mission and the age of the Fra Mauro formation.
Earth and Planetary Science Letters, 12, 3648.
Petit, J., Morbidelli, A., and Chambers, J. (2001). The Primordial Excitation
and Clearing of the Asteroid Belt. Icarus, 153, 338347.
Prettyman, T. H., Mittlefehldt, D. W., Yamashita, N., Lawrence, D. J.,
Beck, A. W., Feldman, W. C., McCoy, T. J., McSween, H. Y., Toplis,
M. J., Titus, T. N., Tricarico, P., Reedy, R. C., Hendricks, J. S., Forni,
O., Le Corre, L., Li, J.-Y., Mizzon, H., Reddy, V., Raymond, C. A., and
Russell, C. T. (2012). Elemental Mapping by Dawn Reveals Exogenic H
in Vestas Regolith. Science, 338, 242246.
Richardson, J. E. (2009). Cratering saturation and equilibrium: A new model
looks at an old problem. Icarus, 204, 697715.
Russell, C. T. and Raymond, C. A. (2011). The Dawn Mission to Vesta and
Ceres. Space Sci. Rev., 163, 323.
Ryder, G. (1990). Lunar samples, lunar accretion and the early bombardment
of the moon. EOS Transactions, 71, 313.
Ryder, G. (2002). Mass ux in the ancient Earth-Moon system and benign
implications for the origin of life on Earth. Journal of Geophysical Research
(Planets), 107, 5022.
Schenk, P., OBrien, D. P., Marchi, S., Gaskell, R., Preusker, F., Roatsch, T.,
Jaumann, R., Buczkowski, D., McCord, T., McSween, H. Y., Williams, D.,
Yingst, A., Raymond, C., and Russell, C. (2012). The Geologically Recent
Giant Impact Basins at Vestas South Pole. Science, 336, 694697.
Schmedemann, N., Kneissl, T., Ivanov, B., Michael, G., Wagner, R.,
Neukum, G., Ruesch, O., Hiesinger, H., Krohn, K., Roatsch, T., Preusker,
F., Sierks, H., Jaumann, R., Reddy, V., Nathues, A., Walter, S., Neese-
mann, A., Raymond, C., and Russell, C. (2014). The cratering record,
34
chronology and surface ages of (4) Vesta in comparison to smaller aster-
oids and the ages of HED meteorites. Planetary and Space Science, in
press.
Sekanina, Z. (1984). Disappearance and disintegration of comets. Icarus, 58,
81100.
St oer, D. and Ryder, G. (2001). Stratigraphy and Isotope Ages of Lu-
nar Geologic Units: Chronological Standard for the Inner Solar System.
Space Sci. Rev., 96, 954.
Strom, R. G., Malhotra, R., Ito, T., Yoshida, F., and Kring, D. A. (2005).
The Origin of Planetary Impactors in the Inner Solar System. Science,
309, 18471850.
Swindle, T. D., Isachsen, C. E., Weirich, J. R., and Kring, D. A. (2009). 40Ar-
39Ar ages of H-chondrite impact melt breccias. Meteoritics and Planetary
Science, 44, 747762.
Tanga, P., Cellino, A., Michel, P., Zappal`a, V., Paolicchi, P., and dellOro,
A. (1999). On the Size Distribution of Asteroid Families: The Role of
Geometry. Icarus, 141, 6578.
Tera, F., Papanastassiou, D. A., and Wasserburg, G. J. (1974). Isotopic
evidence for a terminal lunar cataclysm. Earth and Planetary Science
Letters, 22, 121.
Thomas, P. C., Binzel, R. P., Gaey, M. J., Storrs, A. D., Wells, E. N.,
and Zellner, B. H. (1997). Impact excavation on asteroid 4 Vesta: Hubble
Space Telescope results. Science, 277, 14921495.
Tsiganis, K., Gomes, R., Morbidelli, A., and Levison, H. F. (2005). Origin of
the orbital architecture of the giant planets of the Solar System. Nature,
435, 459461.
Turner, G., Cadogan, P. H., and Yonge, C. J. (1973). Apollo 17 Age Deter-
minations. Nature, 242, 513515.
Turrini, D. (2014). The primordial collisional history of Vesta: Crater satu-
ration, surface evolution and survival of the basaltic crust. Planetary and
Space Science, in press.
35
Turrini, D., Magni, G., and Coradini, A. (2011). Probing the history of Solar
system through the cratering records on Vesta and Ceres. MNRAS, 413,
24392466.
Walker, R. J., Horan, M. F., Shearer, C. K., and Papike, J. J. (2004). Low
abundances of highly siderophile elements in the lunar mantle: evidence
for prolonged late accretion. Earth and Planetary Science Letters, 224,
399413.
Walsh, K. J., Morbidelli, A., Raymond, S. N., OBrien, D. P., and Man-
dell, A. M. (2011). A low mass for Mars from Jupiters early gas-driven
migration. Nature, 475, 206209.
Wasserburg, G. J. and Papanastassiou, D. A. (1971). Age of an Apollo
15 mare basalt: Lunar crust and mantle evolution. Earth and Planetary
Science Letters, 13, 97104.
Weidenschilling, S. J. (1977). The distribution of mass in the planetary
system and solar nebula. Ap&SS, 51, 153158.
Wetherill, G. W. (1992). An alternative model for the formation of the
asteroids. Icarus, 100, 307325.
Wisdom, J. (1985). Meteorites may follow a chaotic route to earth. Nature,
315, 731733.
36
T
o
4.56 Ga
T
LHB
4.1 Ga

LHB
6.93 Ga
1
f
LHB
2

pd
0.025 Ga
Table 1: Nominal values of key parameters in Eqns. 16 and 18
37
0.1
1
10
100
1000
10000
10 100
N
(
>
D
)

(
W
h
o
l
e

S
u
r
f
a
c
e
)
Crater Diameter (km)
Model Production Function
Model Prod. Func. (N*=27.4 Ga)
Approx. 1-sigma Range
Measured Large Craters
Figure 1: The model crater production function F
m
(D) for Vesta plotted along with the
large crater counts of Marchi et al. (2012c). Error bars for the crater counts are

N. The
solid curve is t to the two largest craters on Vesta (note that while craters down to 100
km diameter are shown, they are not included in the t). The dashed curves show an
approximately 1-sigma range, where an expected value of 1 or 4 large craters is assumed,
instead of just two.
38
0.1
1
10
100
1000
10000
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
d
N
*
/
d
T
Time (Ga)
Normalized Impact Rate
Linear Chronology Curve
Lunar Curve Scaled to Vesta
Model Chronology Curve
Figure 2: The normalized dierential production rate of craters dN

/dT in the model


chronology (Eqns. 12 and 14 from Sec. 4), for the nominal parameter values given in
Sec. 6. Also shown is the lunar-like chronology (Eq. 7 from Sec. 3) and a constant linear
production rate curve. The dashed part of the lunar-like curve prior to 4.1 Ga is an
extrapolation, since the lunar chronology on which it is based is only constrained back to
4.1 Ga.
39
0.01
0.1
1
10
100
1000
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
N
*
(
T
)

(
G
a
)
Time (Ga)
Normalized Cumulative Crater Production
Linear Chronology Curve
Lunar Curve Scaled to Vesta
Model Chronology Curve
1
10
100
1000
3 3.2 3.4 3.6 3.8 4 4.2 4.4
N
*
(
T
)

(
G
a
)
Time (Ga)
Normalized Cumulative Crater Production
Linear Chronology Curve
Lunar Curve Scaled to Vesta
1-sigma M
o
Range
Model Chronology Curve
Figure 3: The normalized chronology function N

(T) in the model chronology (Eqns. 16


and 18 from Sec. 4), for the nominal parameter values given in Sec. 6. Also shown is
the lunar-like chronology (Eq. 8 from Sec. 3) and a linear chronology curve. These are
the integrals of the dierential curves in Fig. 2. The dashed part of the lunar-like curve
prior to 4.1 Ga is an extrapolation, since the lunar chronology on which it is based is only
constrained back to 4.1 Ga. The bottom gure is the same as the top, but focuses on
times greater than 3 Ga, when the curves begin to signicantly diverge, and also shows
an approximately 1-sigma range based on uncertainties in the initial impacting mass M
o
.
0.001
0.01
0.1
1
10
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
N
>
1

k
m

p
e
r

k
m
2
Time (Ga)
Cumulative Crater Production
Linear Chronology Curve
Lunar Curve Scaled to Vesta
Model Chronology Curve
0.01
0.1
1
10
3 3.2 3.4 3.6 3.8 4 4.2 4.4
N
>
1

k
m

p
e
r

k
m
2
Time (Ga)
Cumulative Crater Production
Linear Chronology Curve
Lunar Curve Scaled to Vesta
1-sigma M
o
Range
Model Chronology Curve
Figure 4: Model chronology curve for the density of 1 km and larger craters on Vesta, using
the model chronology function from Eqns. 16 and 18 (with the nominal parameter values
given in Sec. 6) and the model production function F
m
(D). Also shown is the lunar-like
chronology curve (Eq. 5) and a linear chronology curve. The dashed part of the lunar-like
curve prior to 4.1 Ga is an extrapolation, since the lunar chronology on which it is based
is only constrained back to 4.1 Ga. The bottom gure shows the same curves as the top,
but focuses on times greater than 3 Ga, when the curves begin to signicantly diverge, and
also shows an approximately 1-sigma range based on uncertainties in the initial impacting
mass M
o
.
0.0001
0.001
0.01
0.1
1
10
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
N
>
1

k
m

p
e
r

k
m
2

p
e
r

G
a
Time (Ga)
Impact Rate
Lunar Curve
Vesta Model Chronology Curve
0.0001
0.001
0.01
0.1
1
10
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
N
>
1

k
m

p
e
r

k
m
2
Time (Ga)
Cumulative Crater Production
Lunar Curve
Vesta Model Chronology Curve
Figure 5: Comparison of the model chronology with the lunar chronology. The top gure
shows the formation rate of 1 km and larger craters for Vesta and the moon, while the bot-
tom gure shows the cumulative crater production (ie. the chronology curve). The dashed
part of the lunar curve prior to 4.1 Ga is an extrapolation, since the lunar chronology is
only constrained back to 4.1 Ga.
42
1
10
100
3 3.2 3.4 3.6 3.8 4 4.2 4.4
N
*
(
T
)

(
G
a
)
Time (Ga)
Normalized Cumulative Crater Production
Nominal Curve with 1-sigma M
o
Range
Model Chronology Curve with f
LHB
= 10
Model Chronology Curve with f
LHB
= 1
1
10
100
3 3.2 3.4 3.6 3.8 4 4.2 4.4
N
*
(
T
)

(
G
a
)
Time (Ga)
Normalized Cumulative Crater Production
Nominal Curve with 1-sigma M
o
Range
Model Chronology Curve with
pd
= 50 Ma
Model Chronology Curve with
pd
= 12.5 Ma
Figure 6: Similar to Fig. 3, but showing the eects of varying several of the parameters
from the nominal values given in Table 1. The top gure shows the eect of varying f
LHB
from 1 to 10, relative to its nominal value of 2. The case of f
LHB
= 1 corresponds to
no depletion and hence no change in the impact rate at t
LHB
, the only change in the
impact rate occurs during the early primordial depletion phase. f
LHB
= 10 is consistent
with the early Nice Model simulations (Gomes et al., 2005), before subsequent modeling
revised that value downwards. The bottom gure shows the eect of varying the primordial
depletion timescale
pd
from 12.5 Ma to 50 Ma, relative to the nominal value of 25 Ma.
For comparison, the grey curves in both plots show the nominal chronology curve with an
approximately 1-sigma range based on uncertainties in the initial impacting mass M
o
.
1e-06
1e-05
0.0001
0.001
0.01
0.1
1 10
C
u
m
u
l
a
t
i
v
e

N
u
m
b
e
r

o
f

C
r
a
t
e
r
s

p
e
r

k
m
2
Crater Diameter (km)
Fit to Measured Crater Counts
Heavily Cratered Terrain (HCT)
HCT MPF Best Fit (4.3 Ga)
Rheasilvia Floor
Rheasilvia MPF Best Fit (1 Ga)
Figure 7: Crater counts for two regions on Vesta, with age estimates using the model
chronology. HCT refers to the highly-cratered terrain identied by Marchi et al. (2012c).
44

You might also like