- Email: [email protected]
Post-impact thermal structure and cooling timescales of Occator crater on asteroid 1 Ceres
Accepted Manuscript
Post-Impact Thermal Structure and Cooling Timescales of Occator Crater on Asteroid 1 Ceres Timothy J. Bowling , Fred J. Ciesla , Thomas M. Davison , JenniferE.C. Scully , Julie C. Castillo-Rogez , Simone Marchi , Brandon C. Johnson PII: DOI: Reference:
S0019-1035(18)30018-6 https://doi.org/10.1016/j.icarus.2018.08.028 YICAR 13003
To appear in:
Icarus
Received date: Revised date: Accepted date:
12 January 2018 25 August 2018 29 August 2018
Please cite this article as: Timothy J. Bowling , Fred J. Ciesla , Thomas M. Davison , JenniferE.C. Scully , Julie C. Castillo-Rogez , Simone Marchi , Brandon C. Johnson , Post-Impact Thermal Structure and Cooling Timescales of Occator Crater on Asteroid 1 Ceres, Icarus (2018), doi: https://doi.org/10.1016/j.icarus.2018.08.028
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
M
AN US
CR IP T
Highlights • Occator crater on dwarf planet 1 Ceres hosts high albedo faculae that may have been deposited by a hydrothermal system • Impact induced heating during the formation Occator produced a hydrothermal system with volumes sufficient to form the faculae • This hydrothermal system lasted between 0.4 and 4 Myr, implying that the faculae are largely contemporaneous with the formation of the crater.
1
ACCEPTED MANUSCRIPT
Post-Impact Thermal Structure and Cooling Timescales of Occator Crater on Asteroid 1 Ceres. Timothy. J. Bowling (Southwest Research Institute, 1050 Walnut Street, Boulder, CO, 80304; corresponding author [email protected]) Fred. J. Ciesla (Department of Geophysical Sciences, University of Chicago, 5734 S. Ellis Avenue, Chicago, IL, 60637)
CR IP T
Thomas. M. Davison (Department of Earth Science and Engineering, Imperial College London, London SW7 2AZ, UK)
Jennifer. E. C. Scully (Jet Propulsion Laboratory, M/S 183-301, 4800 Oak Grove Drive, Pasadena, CA, 91109)
Julie. C. Castillo-Rogez (Jet Propulsion Laboratory, M/S 79-24, 4800 Oak Grove Drive, Pasadena,
AN US
CA, 91109)
Simone Marchi (Southwest Research Institute, 1050 Walnut Street, Boulder, CO, 80304) Brandon C. Johnson (Department of Earth, Environmental and Planetary Sciences, Brown
M
University, 324 Brook St., Box 1846, Providence, RI 02912)
ED
1) Abstract
Occator crater is perhaps the most distinct surface feature observed by NASA’s Dawn
PT
spacecraft on the Cerean surface. Contained within the crater are the highest albedo features on the planet, Cerealia Facula and Vinalia Faculae, and relatively smooth lobate flow deposits.
CE
We present hydrocode simulations of the formation of Occator crater, varying the water to rock ratio of our pre-impact Cerean surface. We find that at water to rock mass ratios up to 0.3,
AC
sufficient volumes of Occator’s post-impact subsurface would be above the melting point of water to allow for the deposition of Faculae like deposits via impact-heat driven hydrothermal effusion of brines. This reservoir of hydrothermally viable material beneath the crater is composed of a mixture of impactor material and material uplifted from 10’s of kilometers beneath the pre-impact surface, potentially sampling a deep subsurface volatile reservoir. Using a conductive cooling model, we estimate that the lifetime of hydrothermal activity within
2
ACCEPTED MANUSCRIPT
such a system, depending on choice of material constants, is between 0.4 and 4 Myr. Our results suggest that impact heating from the Occator forming impact provides a viable mechanism for the creation of observed faculae, with the proviso that the faculae formed within a relatively short time window after the crater itself formed.
CR IP T
2) Introduction
With an orbital semi-major axis of 2.8 AU and a mean radius of 473 km, dwarf planet 1 Ceres is the most massive object in the asteroid belt and lies at the cusp between the rocky inner and icy outer solar system. Currently the observational target of NASA’s Dawn spacecraft (Russell et
AN US
al., 2015), Ceres appears to be an ice rich rocky body, considerably more volatile rich than the smaller, denser asteroid 1 Vesta, and considerably rockier than many Jovian and Saturnian moons of similar size. Because of its transitional composition and location between the dry inner and icy outer regions of the solar system, quantifying the amount and character of
M
subsurface volatiles within Ceres can provide an important basis with which to understand
ED
planetesimal growth and the dynamic history of the Solar System.
While the presence of water within Ceres’s subsurface was suspected long before the arrival of Dawn, with several pre-encounter detections of water outgassing from the dwarf planet
PT
(A’Hearn and Feldman, 1992; Kuppers et al, 2014), constraints on the mass fraction of volatiles within the body could only be loosely inferred from the geodetic constraints (Thomas et al.,
CE
2005). Dawn has revealed Ceres to be surprisingly complex, with a surface composed of salts, ammoniated clays, magnesium rich serpentines, and organic rich regions (De Sanctis et al.,
AC
2015; Ammanito et al., 2016; De Sanctis, 2016; De Sanctis et al., 2017) hinting at a complicated evolutionary history in which subsurface fluid circulation played a major role. Beyond direct spectral detections of ice in freshly exposed regions (Combe et al., 2016; Platz et al., 2016) and of near surface hydrogen at the poles (Prettyman et al., 2017), the amount of water in Ceres’ subsurface today must be largely inferred from studies of the body’s surface morphology and geodetic state. Internal density (Park et al., 2016) and topographic relaxation (Fu et al., 2017)
3
ACCEPTED MANUSCRIPT
models imply that while some subsurface water must be present, Ceres’ outer layers are largely dominated by rock, salts, or other mechanically strong, relatively low density species such as methane clathrates. The viscous relaxation state of impact craters on the body implies an upper limit of 30% water by in the dwarf planet’s lithosphere (Bland et al., 2016). Landslide morphologies within craters on Ceres imply a lower limit, at least regionally, of 10% water in
CR IP T
the near surface (Schmidt et al., 2017).
The 92 km diameter crater ‘Occator’ in Ceres’ norther hemisphere is one of the most distinctive features on the dwarf planet. With steep, terraced crater walls and a largely flat floor, the relatively young (Natheus et al. 2015; Heisinger et al., 2016) crater has a morphology consistent
AN US
with craters of similar size observed on icy satellites (Schenk et al., 2016). Perhaps the most distinctive feature within Occator, however, are numerous bright regions on the crater floor, named ‘faculae’. With relatively high albedo, Occator’s faculae were discernable in pre-Dawn encounter telescopic observations of the dwarf planet (Thomas et al., 2005; Li et al., 2006).
ED
carbonates (De Sanctis et al., 2016).
M
Remote sensing observations strongly suggest the faculae are largely composed of salts and
The formation mechanism of Occator’s faculae is still unknown, and they may be precipitate deposits from past hydrothermal or cryovolcanic activity within the crater. Because they are
PT
localized within the floor of a significant, fresh crater, it is easy to draw a connection between impact induced heating and faculae formation. Less distinct faculae-like deposits are found
CE
elsewhere on Ceres’ surface, and mainly seem to be correlated with relatively large (>30 km), fresh impact craters (Stein et al., 2016). However, given the abundance of non-impact related
AC
cryovolcanic features on Ceres (e.g. Ahuna Mons; Reusch et al., 2016) and the possibility that the faculae post-date crater formation by many millions of years (Natheus et al., 2017), it is not certain if, or how, the driving mechanism behind faculae formation is related to the impact itself.
4
ACCEPTED MANUSCRIPT
Here, we use numerical models to simulate the formation of Occator crater in order to quantify the magnitude and distribution of post-impact temperature within the crater’s subsurface, providing constraints on the distribution and intensity of expected hydrothermal activity immediately following formation. We then model the subsequent conductive cooling of the post-impact temperature structure in order to estimate a timescale by which impact induced
CR IP T
heat within the crater’s subsurface would be available to directly drive faculae forming processes.
3) Description of Impact Modeling
AN US
In order to understand the post-impact temperature structure of Occator, we simulate the formation of the crater using the iSALE-2D ‘Dellen’ shock physics code. iSALE-2D is an extension of the SALE algorithm (Amsden et al., 1980), originally developed to model shock processes, processes that dominate the formation of impact craters on planetary surfaces. iSALE-2D
M
extends this work to include visco-elastic-plastic constitutive models, various empirically grounded equations of state for geologic media, and the capability to handle multiple materials
ED
within a single simulation (Melosh et al., 1992; Ivanov et al., 1997). The code also includes sophisticated strength models required to describe the flow of geologic materials under high stresses (Collins et al., 2004), and a numerically fast method of describing both porous
PT
compaction (Wunnemann et al., 2006; Collins et al., 2011) and dilatant bulking (Collins, 2014).
CE
Our simulations are run on a two dimensional, cylindrically symmetric, Eulerian (fixed cell) computational mesh (Table 1). The two-dimensional geometry of our simulations, a
AC
requirement for reduced computation time, limits our suite of simulations to only vertical impacts at 20 cells per projectile radius. Nonetheless, 2D simulations still provide valuable insights into the initial post-impact temperature state of Occator and possible formation mechanisms for the faculae. The initial target within our simulations is composed of a mixture of serpentine and water with a surface temperature of 150 K, a surface gravity of 0.27 m s-2, and a lithospheric temperature gradient consistent 0.5 K km-1. Our impactors are composed of
5
ACCEPTED MANUSCRIPT
dunite, have radii for a given simulation determined via complex crater scaling laws described in detail in Johnson et al. (2016b), and colliding with the target surface at 4.8 km s-1 (Marchi et al., 2016). Initial near surface target densities are directly dependent on the equation of state and are both temperature and pressure dependent.
CR IP T
The crater modification process, in which the initial transient cavity collapses under the force of gravity, plays a large role in determining the final distribution of strongly heated material within our Occator simulations. This process is governed by several separate sub-grid models within iSALE that describe how materials flow when subjected to various stresses. The four most important such models describe 1) the elastic strength of the target (Collins et al., 2004), or
AN US
how shear stress limits depend on pressure; 2) a damage model (Ivanov et al., 1997), which describes how the elastic strength of a material changes when undergoing increasing strain induced fragmentation; 3) a thermal weakening model (Ohnaka, 1995), which describes how elastic strength is reduced as a material’s temperature approaches its melting point; and 4) an
M
acoustic fluidization model (Bray et al., 2014), which describes a transient weakening mechanism by which material strength is reduced when subjected to acoustic energy deposited
ED
by a passing shock. A common technique in impact crater simulation studies is to use empirically determined inputs for models 1-3 based on assumptions about target rheology, then tune the acoustic fluidization model (4) to reproduce observed topography of a given
PT
crater set (e.g. Bray et al., 2014). However, while the relaxation state of craters (Bland et al., 2016) and topographic spectrum (Fu et al., 2017) on the Cerean surface suggest that the
CE
strength of the lithosphere is dominated by non-H2O components, crater morphologies and key metrics such as simple-to-complex crater transition diameters are consistent with an ice-
AC
dominated target (Schenk et al., 2016). As such, it is difficult to justify the use of solely ‘ice-like’ or ‘rock-like’ input parameters for models 1-3 (e.g. cohesion, coefficient of internal friction, melt temperature).
For this work, in order to reduce the number of simulations required, we employ an alternate approach where we assume that the Cerean surface has ‘rock-like’ rheologic properties for long
6
ACCEPTED MANUSCRIPT
term, slow- to no-strain rate modification processes, and ‘ice-like’ rheologic properties for highstrain rate and transient weakening mechanisms (Table 2). Specifically, we use ‘rock-like’ properties to describe the elastic strength and damage accumulation of materials (models 1-2), based on empirical fits to dunite (Davison et al., 2010; Potter et al., 2015; Johnson et al., 2016b) in the absence of relevant data on hydrated silicates, and ‘ice-like’ properties to describe
CR IP T
temperature and acoustic energy (Bray et al., 2014) dependent transient weakening
mechanisms (models 3-4). The physical justification of this assumption is as such: under low stress or strain rate conditions, an interconnected mixture of ice and rock will behave in a way consistent with its dominant species. However, at high strain rates, high temperatures, or high acoustic energies, an ice-rock mixture will behave in a manner consistent with its weakest
AN US
component. Specifically, in an intimate, low cohesion mixture with a relatively large ice/rock ratio, elastic strength will be greatly reduced when the melting point of water is reached and the mixture becomes a slurry. When subjected to large amounts of scattered acoustic energy, which principally reduces material strength in tension, the weakest component in an
M
interconnected intimate mixture will dominate the degree of fluidization. In addition to models described above, we employ an ice specific, temperature and strain rate dependent viscous
ED
term within our simulations (Johnson et al., 2016a).
Until the arrival of the Dawn spacecraft at Ceres, most detailed observations of craters on solar
PT
system bodies were constrained to either mostly ‘rocky’ or mostly ‘icy’ bodies. As such, there has not been a concerted effort to implement a rheology consistent with a high-rock/high-ice
CE
fraction target into an impact hydrocode, although considerable work has been done examining the strength of ice-rock mixtures in laboratory impacts (e.g. Arakawa and Tomizuka, 2004). The
AC
rheologic assumptions described above are qualitatively consistent with, but do not embrace the full complexity of, studies of the rheology of ice-rock mixtures. In reality, the strength of ice-rock mixtures depends intricately on strain rate, ice/rock ratio, temperature, and porosity (Durham et al, 1992; Yasui and Arakawa, 2008; Yasui and Arakawa, 2010). While we believe our rheology implementation is sufficient for this work, it is clear that this important subfield of impact crater modeling is worthy of further attention and effort.
7
ACCEPTED MANUSCRIPT
The process of impact crater formation is heavily dependent on shock propagation and attenuation, conditions under which materials are subjected to extremely high pressures and temperatures (Melosh, 1989). The amount of intra- and post-shock heating produced during a given impact is strongly dependent on the choice of equation of state (EOS), which relates how
CR IP T
pressure (P) varies as a function of material density () and temperature (T). Because Ceres’ surface is likely a mixture of various species including hydrated silicates (De Sanctis et al., 2016) and water ice (Bland et al., 2016; Schmidt et al., 2017), we follow the method of Pierazzo et al. (2005) to build equations of state for mixtures of serpentine (Brookshaw, 1998) and H2O (Turtle and Pierazzo, 2001). The thermodynamics of individual species are described using the ANEOS
AN US
semi-analytical equation of state package. The mixture technique of Pierazzo et al. (2005) assumes that both component species of a two-phase mixture are always in mechanical (P) and thermal (T) equilibrium, or that (
)
(
) where the subscripts refer to the
specific densities of serpentine and water. The thermal equilibrium assumption implies that the
M
target is composed of an intimate mixture of both species with grain sizes small enough that timescale of thermal equilibration is small compared to the shock compression and rarefaction approximately ⁄
ED
timescale. The shock compression timescale in the subsurface can be estimated as second, where D is the impactor diameter and U is the impact velocity.
The thermal equilibration timescale is approximately
⁄ where d is the grain size and is the
PT
thermal diffusivity (~10-6 m2 s-1). As such, the necessary thermal equilibrium assumption used
CE
in these models is only strictly valid for d<<1 mm. The grain sizes of silicate and volatile species within the Cerean subsurface are currently unknown.
AC
The method of Pierazzo et al. (2005) cannot describe mixture thermodynamics under low density (gaseous), low temperature conditions above the vaporization temperature of water and below the vaporization temperature of serpentine. Physically, this state would correspond to a suspension of liquid or solid serpentine particles within a cloud of water vapor, and would need to be addressed with a different type of numerical model oriented towards multiphase flows. When this regime is reached, we assume that both material pressure and temperature
8
ACCEPTED MANUSCRIPT
are set solely by the H2O component of the mixture, but that the density remains a linear admixture of both species. In addition, for the sake of numerical stability, we delete material from computational cells that reach a density below 100 kg m3. Because very little vaporization occurs at the impact velocities used in our simulations, this deletion method only results in a
CR IP T
loss of <2 impactor masses.
4) Results
In our simulation with a target composed of 80% serpentine and 20% water by mass, the
impact initially opens up a transient crater 24 km deep [Figure 1]. Hot impactor and target
AN US
material is spread along the floor and walls of the transient cavity. After 75 seconds, the
transient crater begins to collapse under the force of Ceres’ gravity. As the relatively unheated crater walls are subsumed by the collapse flow the hottest material, including the majority of material heated above the melting point of water (273 K) and a nominal brine eutectic of 240 K.
M
Much of this material is concentrated within a large central uplift that reaches approximately 15 km above the crater floor, as can be seen in Figure 1 panel C. Because this uplift is
ED
gravitationally unstable, it subsequently collapses, distributing heated material across the final crater floor, which is relatively flat. At the end of crater formation, much of the crater floor is covered by a mixture of target and projectile material that is above the melting point of water.
PT
This layer is deepest at the center of the crater where a ‘plug’ of hot material extends to 15 km depth. This feature may be exaggerated somewhat by the fact that axisymmetric impact
CE
models tend to overestimate central uplift heights and other effects at the axis of symmetry (Elbeshausen, 2009). The hottest material within the center of the crater is 648 K, and
AC
corresponds to delivered impactor material. 150 km3 of target material is heated above 373 K, 1031 km3 above 273 K, and 4118 km3 above 230 K is deposited in the near surface beneath the crater floor.
The choice of ice/rock ratio used to produce our mixed material equation of state affects the magnitude and distribution of subsurface temperature following crater formation as well as the
9
ACCEPTED MANUSCRIPT
depth of the hot central plug. Because we have only relatively loose constraints on the volatile content of Ceres, with nominal values being based on interpretation of geomorphic features and the inferred density and composition of the dwarf planet, we ran several simulations with varying amounts of ice content in the subsurface. Figure 2 shows a comparison of final temperature states between simulations with target ice/rock ratios of 0.1, 0.2, and 0.3 by mass.
CR IP T
The craters that form in more ice-rich targets are slightly deeper and host a larger volume of material above the melting point of water in its subsurface. A trend can be observed in which targets with higher ice contents produce larger volumes of material heated above the melting point of water.
AN US
Because the crater floor is largely composed of material from the collapse of the gravitationally unstable central uplift, heated material within the crater floor is composed of a mixture of hot impactor material, relatively hot target material with near surface provenance, and relatively cool material subsumed from considerable depth [Figure 3]. In the nominal simulation with an
M
ice/rock ratio of 0.2, material with provenance depths >20 km is deposited in the near surface beneath the crater floor, largely in a region near the base of the crater wall. Much of this
ED
material is above a nominal eutectic for a brine, and could be expected to host a hydrothermal system with surface commutativity. As such, if there is a volatile rich reservoir at depth within Ceres’ crust, it could be uplifted, heated, and exposed in regions of Occator’s crater floor,
PT
where fractures produced by the impact could facilitate the transport of fluids to the surface.
CE
5) Discussion
AC
In each of our impact simulations, a considerable volume of material is heated above the melting point of water and deposited in the near surface beneath the floor of the crater. Post impact, this region should be composed of hot, fractured rock containing a component of liquid brine. Hydrothermally driven effusion could cause some of this brine to be driven out on the surface of the crater floor. Hydrothermal modeling would be required to quantify volume of this effusion, but in general the amount extruded would be dependent on the volume and
10
ACCEPTED MANUSCRIPT
distribution of heated material in the subsurface, the efficacy by which fluids could separate from their post-impact location and migrate, and the pre-impact ice/rock ratio. Because Ceres’ surface is a near vacuum, liquid and solid brines extruded onto the surface would rapidly sublimate, leaving behind precipitate deposits like the observed faculae (Schenck et al, this
CR IP T
issue).
If brine extrusion was most rapid in the initial phase of post-impact crater modification, the earliest evaporate deposits within the crater would be associated with very short lived flows and ponds of liquid water on the surface, especially in the center of the crater where the subsurface is hottest. These liquid deposits could then be overlain by a second phase of
AN US
precipitate deposition in which ballistically driven salt rich ice particles would be ejected from fractures in the crater floor (Zolotov, 2017). If both of these deposition process occurred, evidence of the different depositional periods may manifest as albedo or spectral layering observable within regions of the faculae that have been extensively fractured or excavated by
M
small impacts.
ED
The most prominent facula within Occator is located at the very center of the crater within a central pit. Because this region of our models correlated with a deep plug of hot material heated above the melting point of water, it is reasonable to expect that the magnitude of
PT
extrusion and sublimation would be most extreme in this area. Additionally, gravitational settling of solid particles in a hot region with locally high ice/rock ratio may have formed a lake
CE
in the center of the crater immediately following impact. If such a lake existed, it would rapidly boil at the surface until the latent heat of sublimation produced a protective ice cover. In this
AC
state, subsurface flows from outside the crater center could recharge the temporary lake, concentrating the amount of solutes in the region and the volume of the associated precipitate deposit. Eventually, all liquid and solid H2O would sublimate or evaporate, potentially leaving behind a localized depression with a precipitate deposit at the bottom.
11
ACCEPTED MANUSCRIPT
If the extrusion and sublimation of impact heated brines onto the floor of Occator are indeed the formation mechanism which produced the observed faculae within the crater, a maximum timescale of deposition can be inferred based on the conductive cooling timescale of the postimpact thermal structure. Once the subsurface has entirely cooled beneath the eutectic of relevant brines, no more hydrothermal circulation or surface effusion could occur, and the
CR IP T
deposition of precipitate deposits on the crater floor would cease. The conductive cooling timescale provides an upper bound, as additional effects such as hydrothermal circulation and evaporative effects would only increase the rate at which the crater’s subsurface cools.
To estimate the thermal lifetime of liquids within the subsurface of Occator following its
AN US
formation, we use the cooling model of Davison et al. (2012). This model solved the conductive heat equation in two dimensions to quantify the post-impact thermal evolution of the crater. Within the model, a constant temperature of 150 K is imposed along the crater surface boundary, the vertical surface boundary corresponds to the geothermal gradient assumed in
M
our pre-impact iSALE target, and no geothermal heat flux is imposed at the base of the mesh. Material densities and temperatures are mapped directly from iSALE, and uniform values for
ED
heat capacity (1550 J kg-1 K-1) and thermal conductivity (2.7 W m-1 K-1) consistent with ice-rock mixtures (Barnhart et al., 2010) are used across the entire grid. The resolution of this model is the same as that of the iSALE models on which their initial conditions are based. Water/rock
PT
interactions and latent heat of crystallization are not accounted for in this model. Figure 4 shows the post-impact temperature evolution of one of our Occator simulations as a function
CE
of time. Initially, a hot plug of material in the center of the crater as well as the crater floor are above the melting point of water. Within 100 kyr, the floor of the crater has cooled off well
AC
below temperatures at which liquid brines could exist. Only a region within the initially hot central plug is still above the melting point of water. Within 400 kyr, the entire subsurface has cooled to the point that no more hydrothermal activity would be expected.
Several important caveats to the application of our cooling model to Occator need to be discussed. Firstly, if certain species with very low thermal conductivities, such as methane
12
ACCEPTED MANUSCRIPT
clathrates, dominate the volatile budget of Ceres’s subsurface, the cooling timescale could be considerably extended. Because the temperature rate of change within our simulations is linearly dependent on the thermal conductivity, we can estimate a liberal upper limit cooling timescale for Occator of about 4 Myr (corresponding to a very low thermal conductivity of 0.3 W m-1 K-1). Secondly, our cooling model does not include latent heat of fusion, which may
CR IP T
prolong the timescale of cooling. In our nominal model a total of 2.1 x 1021 J of heat is conducted from the crater subsurface as it cools. Including latent heat of fusion and
vaporization, assuming the total volume with melted volatiles is 4118 km3 and the total volume with vaporized volatiles is 150 km3, would add an additional 4.7 x 1020 J to the system. While this is not an inconsiderable amount, we do not expect that the inclusion of latent heat in our
AN US
model would drastically change our cooling timescale estimates. Thirdly, we use a relatively low thermal gradient in our models, in which the nominal eutectic of a brine in the pre-impact target is reached at approximately 50 km depth, we have not considered the influence of heat flux into the bottom of our cooling model, nor the influence of radiogenic heating within the
M
target, both of which are intimately tied to the geothermal gradient. Recent work describing Ceres’ topography may require a higher thermal gradient than used in our models to explain a
ED
low viscosity layer at depth (Fu et al., 2017). Qualitatively, the inclusion of radiogenic heating, basal heat flux, and/or a higher pre-impact thermal gradient within our cooling model would increase the cooling timescale, as the total volume of material heated above the nominal brine
PT
eutectic would be larger. However In a hotter pre-impact target the dynamics of crater formation and collapse would be altered, with more hot material being uplifted from depth,
CE
and it is difficult to decouple heat induced changes to the crater collapse/modification process from heat induced changes to the post-impact cooling process without a large suite of models
AC
covering many free parameters. Finally, and perhaps most importantly, our conductive cooling model cannot account for advective heat transfer via fluid flow within an active hydrothermal system immediately after crater formation. Qualitatively, the inclusion of convection should decrease the cooling timescale, as has been shown for other solar system craters (Abramov et al., 2004). To first order, the ratio of the advective cooling timescale to the conductive cooling timescale is proportional to the permeability of the system (Barnhart et al., 2010). Because the
13
ACCEPTED MANUSCRIPT
permeabilities of geologic media cover a huge range (>10 orders of magnitude) and the permeability of Occator’s subsurface is currently unknown, it is not possible to quantify how much a hydrothermal system could reduce the cooling timescale of Occator. However the effect could be considerable, and as such the cooling timescales reported in this paper should be considered as likely upper limits. Further modeling of Occator’s thermal evolution, including
CR IP T
a larger range of pre-impact geothermal gradients, subsurface compositions, the inclusion of latent heat of fusion, and the implementation of a hydrothermal model as an alternative to our conductive cooling model will be the subject of future work.
The age of Occator’s formation has been estimated based on small crater densities of deposits
AN US
on the crater floor. A recent work concludes that the age of Occator’s largest facula, located in the center of the crater, is approximately 30 Myr younger than the crater itself (Natheus, 2017). If this is the case, extrusion of brines directly driven by impact heat could not possibly be responsible for the formation of this particular deposit, as all fluids within the subsurface would
M
have cooled well beyond the freezing point on a timescale 1-2 orders of magnitude faster than observed. However, re-analysis of the cratering ages of the faculae and the rest of Occator
ED
suggest that there is little to no discernable difference in ages between the two structures, in which case impact induced heating provides a viable mechanism for the deposition of
PT
hydrothermally related deposits (Neeseman et al., this issue).
Depending on the distribution of volatiles in the subsurface before impact, the uplift of deep
CE
seated water rich material into the region directly beneath Occator’s center could provide a reservoir for later stage, non-impact heat driven cryovolcanism. In this study, we assume a
AC
fairly low pre-impact geothermal gradient for our Cerean target. However, if a higher thermal gradient is assumed, than the material that underlies the center of Occator crater may have had pre-impact temperatures at or above the melting point of liquid water. If this was the case, the cooling time of the crater may have been extended.
14
ACCEPTED MANUSCRIPT
6) Conclusions
Our simulations of the formation of Occator crater on dwarf planet 1 Ceres suggest that the impact should produce a considerable amount of target material heated above the melting point of water. The region of heated material is largely concentrated in a central plug at the
CR IP T
center of the crater, although a substantial amount of hot material is also spread across the crater floor by the collapse of the central uplift during crater modification. The volume of
heated water within the subsurface of the crater is dependent on the pre-impact ice/rock ratio of the target, with larger ice/rock ratios leading to larger volumes of hydrothermally viable post-impact target. As such, larger amounts of hydrothermal fluids would be available for
AN US
faculae formation in targets with higher ice/rock ratios. Assuming the faculae within Occator were deposited by impact heat driven hydrothermal processes, higher ice/rock ratios in the pre-impact target would likely lead to larger volume faculae deposits.
M
The volume of material within the crater’s subsurface that should contain post-impact liquid water is considerably larger than the faculae volumes described by Scully et al. (2017, this
ED
issue). We suggest that this hydrothermal reservoir, if extruded onto the surface of the crater floor either as a liquid or a solid and subsequently sublimated, can provide a compelling mechanism for facula formation. However, the lifetime of such a system would be limited to, at
PT
most, a few million years post impact. In either case, we suggest that volatile rich material uplifted from a pre-impact reservoir at considerable depth and deposited beneath the crater
CE
center may provide a region from which the material that formed the facula could be sourced.
AC
7) Acknowledgements
We would like to thank the developers of iSALE-2D, including Gareth Collins, Kai Wunnemann, Dirk Elbeshausen, Tom Davison, Boris Ivanov, and Jay Melosh. We would like to thank the member of NASA’s Dawn mission for making the underlying observations that form the basis for this work. This work was funded in part by NASA’s Postdoctoral Program.
15
ACCEPTED MANUSCRIPT
8) References A’Hearn, M. F. and P. D. Feldman (1992) Water Vaporization of Ceres, Icarus, 98, 54. Abramov, O. and Kring, D. A. (2004) Numerical modeling of an impact-induced hydrothermal system at the Sudbury crater, Journal of Geophysical Research, 109, E10007, doi:10.1029/2003JE002213.
CR IP T
Amsden, A. A., H. M. Ruppel, and C. W. Hirt (1980), SALE: A simplified ALE computer program for fluid flow at speeds, Los Alamos National Laboratories Report LA-8095. Arakawa, M. and D. Tomizuka (2004) Ice-silicate fractionation among icy bodies due to the difference of impact strength between ice and ice-silicate mixture. Icarus, 170, 193-201.
AN US
Barnhart, C. J., Nimmo, F., and B. J. Travis (2010) Martian post-impact hydrothermal systems incorporating freezing. Icarus, 208, 101-117. Bland, M.T., Raymond, C.A., Schenk, P.M., et al., 2016. Composition and structure of the shallow subsurface of Ceres revealed by crater morphology. Nat. Geosci. 9, 538–542. doi:10.1038/ngeo2743.
M
Bray, V. J., G. S. Collins, J. V. Morgan, H. J. Melosh, and P. M. Schenk (2014), Hydrocode simulation of Ganymede and Europa cratering trends— How thick is Europa’s crust? Icarus, 231, 394–406, doi:10.1016/j.icarus.2013.12.009.
ED
Brookshaw, L. 1998, An equation of state for serpentine, Tech. Rep. Working Paper Series SCMC-9813, Faculty of Sciences, University of Southern Queensland 2.1, 3.5
PT
Collins, G. S., H. J. Melosh, and B. A. Ivanov (2004), Modeling damage and deformation in impact simulations, Met. And Planet. Sci., 39, 217–231.
CE
Collins, G. S., H. J. Melosh, and K. Wünnemann (2010), Improvements to the e-a porous compaction model for simulating impacts into high-porosity solar system objects, Int. J. Impact Eng., 38, 434–493.
AC
Collins, GS (2014), Numerical simulations of impact crater formation with dilatancy, J. Geophys. Res. Planets, 119, 2600–2619 Collins, G. S., Elbeshausen, D., Davison, T. M.; Wünnemann, K.; Ivanov, B.; Melosh, H. J. (2016): iSALE-Dellen manual. figshare. Combe, J.-Ph., McCord, T.B., Tosi, F., et al., 2016. Detection of local H2O exposed at the surface of Ceres. Science 353, aaf3010. doi:10.1126/science.aaf3010.
16
ACCEPTED MANUSCRIPT
Davison, T. M., G. S. Collins, and F. J. Ciesla (2010), Numerical modelling of heating in porous planetesimal collisions, Icarus, 208(1), 468–481, doi:10.1016/j.icarus.2010.01.034. Davison, T.M., Ciesla, F.J. and Collins, G.S. (2012) Post-impact thermal evolution of porous planetesimals. Geochimica et Cosmochimica Acta, Volume 95, pp. 252–269(18).
CR IP T
De Sanctis, M.C., Ammannito, E., Raponi, A., et al., 2015. Ammoniated phyllosilicates with a likely outer Solar System origin on (1) Ceres. Nature 528, 241–244. doi:10.1038/nature16172. De Sanctis, M.C., Raponi, A., Ammannito, E., et al., 2016. Bright carbonate deposits as evidence of aqueous alteration on (1) Ceres. Nature 536, 54–57. doi:10.1038/ nature18290. De Sanctis, M.C., Ammannito, E., McSween, H.Y., et al., 2017. Localized aliphatic organic material on the surface of Ceres. Science 355, 719–722. doi:10.1126/ science.aaj2305.
AN US
Durham, W. B., S. H. Kirby, and L. A. Stern (1992), Effects of dispersed particulates on the rheology of water ice at planetary conditions, J. Geophys. Res., 97, 20,883–20,897, doi:10.1029/92JE02326.
Fu, R. R. et al. (2017) The interior structure of Ceres as revealed by surface topography. Earth and Planetary Science Letters, 476, 153-164.
M
Hayne, P.O., Aharonson, O., 2015. Thermal stability of ice on Ceres with rough topography. J. Geophys. Res. Planets 120, 1567–1584. doi:10.1002/2015JE004887.
ED
Hiesinger, H., Marchi, S., Schmedemann, N., et al., 2016. Cratering on Ceres: impli- cations for its crust and evolution. Science 353, aaf4759. doi:10.1126/science. aaf4759.
CE
PT
Ivanov, B. A., D. Deniem, and G. Neukum (1997), Implementation of dynamic strength models into 2-D hydrocodes: Applications for atmospheric breakup and impact cratering, Int. J. Impact Eng., 20, 411–430.
AC
Ivanov, B.A., Pierazzo, E., 2011. Impact cratering in H2O-bearing targets on Mars: thermal field under craters as starting conditions for hydrothermal activity. Me- teorit. Planet. Sci. 46, 601– 619. doi:10.1111/j.1945-5100.2011.01177.x. Johnson, B. C. et al. (2016a) Formation of the Sputnik Planum basin and the thickness of Pluto’s subsurface ocean, GRL, 10,068. Johnson, B. C. et al. (2016b) Spherule layers, crater scaling laws, and the population of ancient terrestrial impactors, Icarus, 271, 350-359. Kuppers, M. et al. (2014) Localized sources of water vapour on the dwarf planet (1) Ceres, Nature, 505, 525. 17
ACCEPTED MANUSCRIPT
Li, J.-Y., McFadden, L.A., Parker, J.Wm., et al., 2006. Photometric analysis of 1 Ceres and surface mapping from HST observations. Icarus 182, 143–160. doi:10.1016/j. icarus.2005.12.012. Melosh, H. J. (1989), Impact Cratering: A Geologic Process, Oxford Univ. Press, New York.
CR IP T
Melosh, H. J., E. V. Ryan, and E. Asphaug (1992), Dynamic fragmentation in impacts: Hydrocode simulation of laboratory impacts, J. Geophys. Res., 97, 14,735–14,795. Nathues, A.Hoffmann, M., Schaefer, et al., 2015. Sublimation in bright spots on (1) Ceres. Nature 528, 237–240. doi:10.1038/nature15754.
Nathues, A., Platz, T., Thangjam, G., 2017. Evolution of Occator Crater on (1) Ceres. Astronom. J. 153, 112. doi:10.3847/1538-3881/153/3/112, (12pp).
AN US
Neesemann, A., et al. The various ages of Occator crater, Ceres: Results of a c comprehensive synthesis approach. Icarus, this issue.
Ohnaka, M. (1995), A shear failure strength law of rock in the brittle-plastic transition regime, Geophys. Res. Lett., 22, 25–28.
M
Park, R.S., Konopliv, A.S., Bills, B.G., et al., 2016. A partially differentiated interior for (1) Ceres deduced from its gravity field and shape. Nature 537. doi:10.1038/ nature18955.
ED
Pierazzo, E. et al. (2005) Starting conditions for hydrothermal systems underneath Martian craters: Hydrocode modeling, GSA Special Paper 384.
PT
Platz, T., Nathues, A., Schorghofer, N., et al., 2016. Surface water-ice deposits in the northern shadowed regions of Ceres. Nat. Astron. 1. doi:10.1038/ s41550-016-0007, id. 0007.
AC
CE
Potter, R. W. K., D. A. Kring, and G. S. Collins (2015), Scaling of basin-sized impacts and the influence of target temperature, in Large Meteorite Impacts and Planetary Evolution V: Geological Society of America Special Paper, edited by Osinski, G. R., and Kring, D. A., vol. 518, pp. 99–113, Geol. Soc. of Am., Boulder, Colo., doi:10.1130/2015.2518(06). Prettyman, T.H., Yamashita, N., Toplis, M.J., et al., 2017. Extensive water ice within Ceres’ aqueously altered regolith: evidence from nuclear spectroscopy. Science 355, 55–59. doi:10.1126/science.aah6765. Ruesch, O.; Platz, T.; Schenk, P.; McFadden, L. A.; Castillo-Rogez, J. C.; Quick, L. C.; Byrne, S.; Preusker, F.; OBrien, D. P.; Schmedemann, N.; Williams, D. A.; Li, J.- Y.; Bland, M. T.; Hiesinger, H.; Kneissl, T.; Neesemann, A.; Schaefer, M.; Pasckert, J. H.; Schmidt, B. E.; Buczkowski, D. L.; Sykes, M. V.; Nathues, A.; Roatsch, T.; Hoffmann, M.; Raymond, C. A.; Russell, C. T. (2016). Cryovolcanism on Ceres, Science. 353 (6303): aaf4286–aaf4286. doi:10.1126/science.aaf4286. 18
ACCEPTED MANUSCRIPT
Russell, C. T. et al. (2016) Dawn arrives at Ceres: Exploration of a small, volatile-rich world, Science, 353, 6303 Schenk, P. et al., (2016) Impact crater on the small planets Ceres and Vesta: Central Pits and the Origin of Bright Spots. LPSC 47, 2697.
CR IP T
Schenk, P. M., et al. Geomorphology and formation of Occator crater and hydrothermal origins of its bright spots: interplanetary comparisons. Icarus, this issue.
Schmidt, B. E. et al., (2017) Geomorphological evidence for ground ice on dwarf planet Ceres, Nature Geoscience, 10, 338.
AN US
Stein, N., Ehlmann, B., Ammannito, E., et al., 2017. Characteristics, formation, and evolution of faculae (bright spots) on Ceres. Lunar Planet Sci. 48 Abstract 2547. Thomas, P. C. et al. (2005) Differentiation of the asteroid Ceres as revealed by its shape, 437, 224. Turtle, E. P., and E. Pierazzo (2001), Thickness of a Europan ice shell from impact crater simulations, Science, 294(5545), 1326–1328, doi:10.1126/science.1062492.
ED
M
Wünnemann, K., Collins, G., and Melosh, H. (2006). A strain-based porosity model for use in hydrocode simulations of impacts and implications for transient crater growth in porous targets. Icarus, 180:514--527.
PT
Yasui, M. and M Arakawa (2008) Experimental study on the rate depended strength of ice-silica mixture with silica volume fractions up to 0.63. GRL, 35, L12206.
CE
Yasui, M. and M. Arakawa (2010) Rate-dependent strength of porous ice-silica mixtures and its implications for the shape of small to middle-sized icy sattelites. Icarus, 210, 956-967.
AC
Zolotov, M. Y. (2017) Aqueous origins of bright salt deposists on Ceres, Icarus, 296, 289.
9) Acknowledgements
We gratefully acknowledge the developers of iSALE-2D, including Gareth Collins, Kai Wünnemann, Dirk Elbeshausen, Boris Ivanov and Jay Melosh. We also acknowledge Brandon Johnson and Gareth Collins for their suggestions and insight during the development of this work, and two anonymous reviewers for their thoughtful and helpful comments. Finally, we acknowledge NASA’s Dawn mission and science teams for their efforts providing the observations which underlie this work. 19
ACCEPTED MANUSCRIPT
10) Figures
B
AN US
CR IP T
A
D
AC
CE
PT
ED
M
C
Figure 1 Figure showing the formation timeline of an Occator-like crater in a target composed of 20% water ice and 80% serpentine by mass. (top left) Immediately following impact, the target is heated by the passage of the impact induced shock wave. (bottom left) The hottest material, including the impactor, is largely spread out across the crater floor. (top right) The collapse of the transient crater concentrates most of the highly heated material in a gravitationally unstable central uplift which subsequently collapses. (bottom right) The hottest material in the crater forms a plug directly beneath the center of Occator, with the denser
20
ACCEPTED MANUSCRIPT
impactor (red material) slowly settling to the bottom of the plug. Grey lines show tracer particles of equal pre-impact provenance depth. D
CR IP T
A
C
AN US
E
M
B
ED
A
AC
CE
PT
F
Figure 2 Figure showing the post-impact thermal fields of 3 Occator-like craters with target ice mass ratios of 10% (A), 20% (B), and 30% (C). In the lowest case (10%), most of the hot material forms a relatively shallow plug, or is spread out along the crater floor. As the amount of material heated above the melting point of ice is increased, the volume of the hot plug increases, as does its depth. Increasing the ice-rock ratio also leads to more distinct crater walls of steeper slope. Grey lines show tracer particles of equal pre-impact provenance depth.
21
ACCEPTED MANUSCRIPT
Panels D-F show the corresponding post-impact location of impactor material (red) and target material (grey). It can been clearly seen that the hottest material within each crater’s subsurface is impactor.
AN US
CR IP T
A
PT
ED
M
B
AC
CE
C
Figure 3 Figure showing the provenance depth of uplifted material beneath Ceres for 10% (A), 20% (B), and 30% (C) target ice mass ratios. Black lines correspond to isotherms of 273 and 230
22
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
M
AN US
CR IP T
K (the nominal melting point of water and of a brine eutectic). Each dot represents a massless Lagrangian tracer particle, which follow the flow of material as the simulation progresses, colored according to pre-impact provenance depth. In each simulation, the central hot plug is composed of a mixture of impactor and target excavated from considerable depth. Had a deep subsurface volatile reservoir existed before Occator formed, it would have been subsequently heated, uplifted, and deposited beneath the central floor of the crater.
23
ED
M
AN US
CR IP T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 4 Figure showing the conductive cooling of an Occator like structure (nominally similar to the 10% water by mass case). Within 100 kyr (B) after crater formation, hot material on the crater floor has completely cooled below the melting point of water and/or a brine eutectic. Hot material can remain above these temperature limits beneath the center of the crater for approximately 300-400 kyr (D-E), after which all possibility of impact induced hydrothermal activity will cease.
24
Grid Resolution 100 m Number of Horizontal Cells in Grid 1200 Number of Vertical Cells in Grid 800 Impactor Diameter 5.6-6.3 km Impactor Velocity 4.8 km s-1 Target Ice/Rock Ratio (by mass) 0.1-0.3 Target Suface Gravity 0.27 m s-2 Target Surface Temperature 150 K Pre-impact Target Temperature Gradient 0.5 K km-1 Table 1: Global Parameters Used in iSALE Simulations
Target ANEOS Serpentine/H2O 273 K 1318-1955 J kg-1 K-1 1.2 0.26-0.27 0.6 1.4
3.5 GPa 10 kPa 50 MPa 187-197 s 1.9-2.3x106 m2 s-1
AC
CE
PT
ED
M
AN US
Material Parameters Impactor Equation of State ANEOS Dunite Melting Temperature 1436 K Specific Heat Capacity 1000 J kg-1 K-1 Thermal Softening Parameter 2.0 Poisson Ratio 0.25 Coefficient of Internal 0.6 Friction (damaged) Coefficient of Internal 1.4 Friction (undamaged) Strength at Infinite Pressure 3.5 GPa Cohesion (damaged) 10 kPa Cohesion (undamaged) 50 MPa Acoustic Fluidization Time Decay Acoustic Fluidization Kinematic Viscosity Table 2: Material Parameters Used in iSALE Simulations
CR IP T
ACCEPTED MANUSCRIPT
25
Post-Impact Thermal Structure and Cooling Timescales of Occator Crater on Asteroid 1 Ceres Timothy J. Bowling , Fred J. Ciesla , Thomas M. Davison , JenniferE.C. Scully , Julie C. Castillo-Rogez , Simone Marchi , Brandon C. Johnson PII: DOI: Reference:
S0019-1035(18)30018-6 https://doi.org/10.1016/j.icarus.2018.08.028 YICAR 13003
To appear in:
Icarus
Received date: Revised date: Accepted date:
12 January 2018 25 August 2018 29 August 2018
Please cite this article as: Timothy J. Bowling , Fred J. Ciesla , Thomas M. Davison , JenniferE.C. Scully , Julie C. Castillo-Rogez , Simone Marchi , Brandon C. Johnson , Post-Impact Thermal Structure and Cooling Timescales of Occator Crater on Asteroid 1 Ceres, Icarus (2018), doi: https://doi.org/10.1016/j.icarus.2018.08.028
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
M
AN US
CR IP T
Highlights • Occator crater on dwarf planet 1 Ceres hosts high albedo faculae that may have been deposited by a hydrothermal system • Impact induced heating during the formation Occator produced a hydrothermal system with volumes sufficient to form the faculae • This hydrothermal system lasted between 0.4 and 4 Myr, implying that the faculae are largely contemporaneous with the formation of the crater.
1
ACCEPTED MANUSCRIPT
Post-Impact Thermal Structure and Cooling Timescales of Occator Crater on Asteroid 1 Ceres. Timothy. J. Bowling (Southwest Research Institute, 1050 Walnut Street, Boulder, CO, 80304; corresponding author [email protected]) Fred. J. Ciesla (Department of Geophysical Sciences, University of Chicago, 5734 S. Ellis Avenue, Chicago, IL, 60637)
CR IP T
Thomas. M. Davison (Department of Earth Science and Engineering, Imperial College London, London SW7 2AZ, UK)
Jennifer. E. C. Scully (Jet Propulsion Laboratory, M/S 183-301, 4800 Oak Grove Drive, Pasadena, CA, 91109)
Julie. C. Castillo-Rogez (Jet Propulsion Laboratory, M/S 79-24, 4800 Oak Grove Drive, Pasadena,
AN US
CA, 91109)
Simone Marchi (Southwest Research Institute, 1050 Walnut Street, Boulder, CO, 80304) Brandon C. Johnson (Department of Earth, Environmental and Planetary Sciences, Brown
M
University, 324 Brook St., Box 1846, Providence, RI 02912)
ED
1) Abstract
Occator crater is perhaps the most distinct surface feature observed by NASA’s Dawn
PT
spacecraft on the Cerean surface. Contained within the crater are the highest albedo features on the planet, Cerealia Facula and Vinalia Faculae, and relatively smooth lobate flow deposits.
CE
We present hydrocode simulations of the formation of Occator crater, varying the water to rock ratio of our pre-impact Cerean surface. We find that at water to rock mass ratios up to 0.3,
AC
sufficient volumes of Occator’s post-impact subsurface would be above the melting point of water to allow for the deposition of Faculae like deposits via impact-heat driven hydrothermal effusion of brines. This reservoir of hydrothermally viable material beneath the crater is composed of a mixture of impactor material and material uplifted from 10’s of kilometers beneath the pre-impact surface, potentially sampling a deep subsurface volatile reservoir. Using a conductive cooling model, we estimate that the lifetime of hydrothermal activity within
2
ACCEPTED MANUSCRIPT
such a system, depending on choice of material constants, is between 0.4 and 4 Myr. Our results suggest that impact heating from the Occator forming impact provides a viable mechanism for the creation of observed faculae, with the proviso that the faculae formed within a relatively short time window after the crater itself formed.
CR IP T
2) Introduction
With an orbital semi-major axis of 2.8 AU and a mean radius of 473 km, dwarf planet 1 Ceres is the most massive object in the asteroid belt and lies at the cusp between the rocky inner and icy outer solar system. Currently the observational target of NASA’s Dawn spacecraft (Russell et
AN US
al., 2015), Ceres appears to be an ice rich rocky body, considerably more volatile rich than the smaller, denser asteroid 1 Vesta, and considerably rockier than many Jovian and Saturnian moons of similar size. Because of its transitional composition and location between the dry inner and icy outer regions of the solar system, quantifying the amount and character of
M
subsurface volatiles within Ceres can provide an important basis with which to understand
ED
planetesimal growth and the dynamic history of the Solar System.
While the presence of water within Ceres’s subsurface was suspected long before the arrival of Dawn, with several pre-encounter detections of water outgassing from the dwarf planet
PT
(A’Hearn and Feldman, 1992; Kuppers et al, 2014), constraints on the mass fraction of volatiles within the body could only be loosely inferred from the geodetic constraints (Thomas et al.,
CE
2005). Dawn has revealed Ceres to be surprisingly complex, with a surface composed of salts, ammoniated clays, magnesium rich serpentines, and organic rich regions (De Sanctis et al.,
AC
2015; Ammanito et al., 2016; De Sanctis, 2016; De Sanctis et al., 2017) hinting at a complicated evolutionary history in which subsurface fluid circulation played a major role. Beyond direct spectral detections of ice in freshly exposed regions (Combe et al., 2016; Platz et al., 2016) and of near surface hydrogen at the poles (Prettyman et al., 2017), the amount of water in Ceres’ subsurface today must be largely inferred from studies of the body’s surface morphology and geodetic state. Internal density (Park et al., 2016) and topographic relaxation (Fu et al., 2017)
3
ACCEPTED MANUSCRIPT
models imply that while some subsurface water must be present, Ceres’ outer layers are largely dominated by rock, salts, or other mechanically strong, relatively low density species such as methane clathrates. The viscous relaxation state of impact craters on the body implies an upper limit of 30% water by in the dwarf planet’s lithosphere (Bland et al., 2016). Landslide morphologies within craters on Ceres imply a lower limit, at least regionally, of 10% water in
CR IP T
the near surface (Schmidt et al., 2017).
The 92 km diameter crater ‘Occator’ in Ceres’ norther hemisphere is one of the most distinctive features on the dwarf planet. With steep, terraced crater walls and a largely flat floor, the relatively young (Natheus et al. 2015; Heisinger et al., 2016) crater has a morphology consistent
AN US
with craters of similar size observed on icy satellites (Schenk et al., 2016). Perhaps the most distinctive feature within Occator, however, are numerous bright regions on the crater floor, named ‘faculae’. With relatively high albedo, Occator’s faculae were discernable in pre-Dawn encounter telescopic observations of the dwarf planet (Thomas et al., 2005; Li et al., 2006).
ED
carbonates (De Sanctis et al., 2016).
M
Remote sensing observations strongly suggest the faculae are largely composed of salts and
The formation mechanism of Occator’s faculae is still unknown, and they may be precipitate deposits from past hydrothermal or cryovolcanic activity within the crater. Because they are
PT
localized within the floor of a significant, fresh crater, it is easy to draw a connection between impact induced heating and faculae formation. Less distinct faculae-like deposits are found
CE
elsewhere on Ceres’ surface, and mainly seem to be correlated with relatively large (>30 km), fresh impact craters (Stein et al., 2016). However, given the abundance of non-impact related
AC
cryovolcanic features on Ceres (e.g. Ahuna Mons; Reusch et al., 2016) and the possibility that the faculae post-date crater formation by many millions of years (Natheus et al., 2017), it is not certain if, or how, the driving mechanism behind faculae formation is related to the impact itself.
4
ACCEPTED MANUSCRIPT
Here, we use numerical models to simulate the formation of Occator crater in order to quantify the magnitude and distribution of post-impact temperature within the crater’s subsurface, providing constraints on the distribution and intensity of expected hydrothermal activity immediately following formation. We then model the subsequent conductive cooling of the post-impact temperature structure in order to estimate a timescale by which impact induced
CR IP T
heat within the crater’s subsurface would be available to directly drive faculae forming processes.
3) Description of Impact Modeling
AN US
In order to understand the post-impact temperature structure of Occator, we simulate the formation of the crater using the iSALE-2D ‘Dellen’ shock physics code. iSALE-2D is an extension of the SALE algorithm (Amsden et al., 1980), originally developed to model shock processes, processes that dominate the formation of impact craters on planetary surfaces. iSALE-2D
M
extends this work to include visco-elastic-plastic constitutive models, various empirically grounded equations of state for geologic media, and the capability to handle multiple materials
ED
within a single simulation (Melosh et al., 1992; Ivanov et al., 1997). The code also includes sophisticated strength models required to describe the flow of geologic materials under high stresses (Collins et al., 2004), and a numerically fast method of describing both porous
PT
compaction (Wunnemann et al., 2006; Collins et al., 2011) and dilatant bulking (Collins, 2014).
CE
Our simulations are run on a two dimensional, cylindrically symmetric, Eulerian (fixed cell) computational mesh (Table 1). The two-dimensional geometry of our simulations, a
AC
requirement for reduced computation time, limits our suite of simulations to only vertical impacts at 20 cells per projectile radius. Nonetheless, 2D simulations still provide valuable insights into the initial post-impact temperature state of Occator and possible formation mechanisms for the faculae. The initial target within our simulations is composed of a mixture of serpentine and water with a surface temperature of 150 K, a surface gravity of 0.27 m s-2, and a lithospheric temperature gradient consistent 0.5 K km-1. Our impactors are composed of
5
ACCEPTED MANUSCRIPT
dunite, have radii for a given simulation determined via complex crater scaling laws described in detail in Johnson et al. (2016b), and colliding with the target surface at 4.8 km s-1 (Marchi et al., 2016). Initial near surface target densities are directly dependent on the equation of state and are both temperature and pressure dependent.
CR IP T
The crater modification process, in which the initial transient cavity collapses under the force of gravity, plays a large role in determining the final distribution of strongly heated material within our Occator simulations. This process is governed by several separate sub-grid models within iSALE that describe how materials flow when subjected to various stresses. The four most important such models describe 1) the elastic strength of the target (Collins et al., 2004), or
AN US
how shear stress limits depend on pressure; 2) a damage model (Ivanov et al., 1997), which describes how the elastic strength of a material changes when undergoing increasing strain induced fragmentation; 3) a thermal weakening model (Ohnaka, 1995), which describes how elastic strength is reduced as a material’s temperature approaches its melting point; and 4) an
M
acoustic fluidization model (Bray et al., 2014), which describes a transient weakening mechanism by which material strength is reduced when subjected to acoustic energy deposited
ED
by a passing shock. A common technique in impact crater simulation studies is to use empirically determined inputs for models 1-3 based on assumptions about target rheology, then tune the acoustic fluidization model (4) to reproduce observed topography of a given
PT
crater set (e.g. Bray et al., 2014). However, while the relaxation state of craters (Bland et al., 2016) and topographic spectrum (Fu et al., 2017) on the Cerean surface suggest that the
CE
strength of the lithosphere is dominated by non-H2O components, crater morphologies and key metrics such as simple-to-complex crater transition diameters are consistent with an ice-
AC
dominated target (Schenk et al., 2016). As such, it is difficult to justify the use of solely ‘ice-like’ or ‘rock-like’ input parameters for models 1-3 (e.g. cohesion, coefficient of internal friction, melt temperature).
For this work, in order to reduce the number of simulations required, we employ an alternate approach where we assume that the Cerean surface has ‘rock-like’ rheologic properties for long
6
ACCEPTED MANUSCRIPT
term, slow- to no-strain rate modification processes, and ‘ice-like’ rheologic properties for highstrain rate and transient weakening mechanisms (Table 2). Specifically, we use ‘rock-like’ properties to describe the elastic strength and damage accumulation of materials (models 1-2), based on empirical fits to dunite (Davison et al., 2010; Potter et al., 2015; Johnson et al., 2016b) in the absence of relevant data on hydrated silicates, and ‘ice-like’ properties to describe
CR IP T
temperature and acoustic energy (Bray et al., 2014) dependent transient weakening
mechanisms (models 3-4). The physical justification of this assumption is as such: under low stress or strain rate conditions, an interconnected mixture of ice and rock will behave in a way consistent with its dominant species. However, at high strain rates, high temperatures, or high acoustic energies, an ice-rock mixture will behave in a manner consistent with its weakest
AN US
component. Specifically, in an intimate, low cohesion mixture with a relatively large ice/rock ratio, elastic strength will be greatly reduced when the melting point of water is reached and the mixture becomes a slurry. When subjected to large amounts of scattered acoustic energy, which principally reduces material strength in tension, the weakest component in an
M
interconnected intimate mixture will dominate the degree of fluidization. In addition to models described above, we employ an ice specific, temperature and strain rate dependent viscous
ED
term within our simulations (Johnson et al., 2016a).
Until the arrival of the Dawn spacecraft at Ceres, most detailed observations of craters on solar
PT
system bodies were constrained to either mostly ‘rocky’ or mostly ‘icy’ bodies. As such, there has not been a concerted effort to implement a rheology consistent with a high-rock/high-ice
CE
fraction target into an impact hydrocode, although considerable work has been done examining the strength of ice-rock mixtures in laboratory impacts (e.g. Arakawa and Tomizuka, 2004). The
AC
rheologic assumptions described above are qualitatively consistent with, but do not embrace the full complexity of, studies of the rheology of ice-rock mixtures. In reality, the strength of ice-rock mixtures depends intricately on strain rate, ice/rock ratio, temperature, and porosity (Durham et al, 1992; Yasui and Arakawa, 2008; Yasui and Arakawa, 2010). While we believe our rheology implementation is sufficient for this work, it is clear that this important subfield of impact crater modeling is worthy of further attention and effort.
7
ACCEPTED MANUSCRIPT
The process of impact crater formation is heavily dependent on shock propagation and attenuation, conditions under which materials are subjected to extremely high pressures and temperatures (Melosh, 1989). The amount of intra- and post-shock heating produced during a given impact is strongly dependent on the choice of equation of state (EOS), which relates how
CR IP T
pressure (P) varies as a function of material density () and temperature (T). Because Ceres’ surface is likely a mixture of various species including hydrated silicates (De Sanctis et al., 2016) and water ice (Bland et al., 2016; Schmidt et al., 2017), we follow the method of Pierazzo et al. (2005) to build equations of state for mixtures of serpentine (Brookshaw, 1998) and H2O (Turtle and Pierazzo, 2001). The thermodynamics of individual species are described using the ANEOS
AN US
semi-analytical equation of state package. The mixture technique of Pierazzo et al. (2005) assumes that both component species of a two-phase mixture are always in mechanical (P) and thermal (T) equilibrium, or that (
)
(
) where the subscripts refer to the
specific densities of serpentine and water. The thermal equilibrium assumption implies that the
M
target is composed of an intimate mixture of both species with grain sizes small enough that timescale of thermal equilibration is small compared to the shock compression and rarefaction approximately ⁄
ED
timescale. The shock compression timescale in the subsurface can be estimated as second, where D is the impactor diameter and U is the impact velocity.
The thermal equilibration timescale is approximately
⁄ where d is the grain size and is the
PT
thermal diffusivity (~10-6 m2 s-1). As such, the necessary thermal equilibrium assumption used
CE
in these models is only strictly valid for d<<1 mm. The grain sizes of silicate and volatile species within the Cerean subsurface are currently unknown.
AC
The method of Pierazzo et al. (2005) cannot describe mixture thermodynamics under low density (gaseous), low temperature conditions above the vaporization temperature of water and below the vaporization temperature of serpentine. Physically, this state would correspond to a suspension of liquid or solid serpentine particles within a cloud of water vapor, and would need to be addressed with a different type of numerical model oriented towards multiphase flows. When this regime is reached, we assume that both material pressure and temperature
8
ACCEPTED MANUSCRIPT
are set solely by the H2O component of the mixture, but that the density remains a linear admixture of both species. In addition, for the sake of numerical stability, we delete material from computational cells that reach a density below 100 kg m3. Because very little vaporization occurs at the impact velocities used in our simulations, this deletion method only results in a
CR IP T
loss of <2 impactor masses.
4) Results
In our simulation with a target composed of 80% serpentine and 20% water by mass, the
impact initially opens up a transient crater 24 km deep [Figure 1]. Hot impactor and target
AN US
material is spread along the floor and walls of the transient cavity. After 75 seconds, the
transient crater begins to collapse under the force of Ceres’ gravity. As the relatively unheated crater walls are subsumed by the collapse flow the hottest material, including the majority of material heated above the melting point of water (273 K) and a nominal brine eutectic of 240 K.
M
Much of this material is concentrated within a large central uplift that reaches approximately 15 km above the crater floor, as can be seen in Figure 1 panel C. Because this uplift is
ED
gravitationally unstable, it subsequently collapses, distributing heated material across the final crater floor, which is relatively flat. At the end of crater formation, much of the crater floor is covered by a mixture of target and projectile material that is above the melting point of water.
PT
This layer is deepest at the center of the crater where a ‘plug’ of hot material extends to 15 km depth. This feature may be exaggerated somewhat by the fact that axisymmetric impact
CE
models tend to overestimate central uplift heights and other effects at the axis of symmetry (Elbeshausen, 2009). The hottest material within the center of the crater is 648 K, and
AC
corresponds to delivered impactor material. 150 km3 of target material is heated above 373 K, 1031 km3 above 273 K, and 4118 km3 above 230 K is deposited in the near surface beneath the crater floor.
The choice of ice/rock ratio used to produce our mixed material equation of state affects the magnitude and distribution of subsurface temperature following crater formation as well as the
9
ACCEPTED MANUSCRIPT
depth of the hot central plug. Because we have only relatively loose constraints on the volatile content of Ceres, with nominal values being based on interpretation of geomorphic features and the inferred density and composition of the dwarf planet, we ran several simulations with varying amounts of ice content in the subsurface. Figure 2 shows a comparison of final temperature states between simulations with target ice/rock ratios of 0.1, 0.2, and 0.3 by mass.
CR IP T
The craters that form in more ice-rich targets are slightly deeper and host a larger volume of material above the melting point of water in its subsurface. A trend can be observed in which targets with higher ice contents produce larger volumes of material heated above the melting point of water.
AN US
Because the crater floor is largely composed of material from the collapse of the gravitationally unstable central uplift, heated material within the crater floor is composed of a mixture of hot impactor material, relatively hot target material with near surface provenance, and relatively cool material subsumed from considerable depth [Figure 3]. In the nominal simulation with an
M
ice/rock ratio of 0.2, material with provenance depths >20 km is deposited in the near surface beneath the crater floor, largely in a region near the base of the crater wall. Much of this
ED
material is above a nominal eutectic for a brine, and could be expected to host a hydrothermal system with surface commutativity. As such, if there is a volatile rich reservoir at depth within Ceres’ crust, it could be uplifted, heated, and exposed in regions of Occator’s crater floor,
PT
where fractures produced by the impact could facilitate the transport of fluids to the surface.
CE
5) Discussion
AC
In each of our impact simulations, a considerable volume of material is heated above the melting point of water and deposited in the near surface beneath the floor of the crater. Post impact, this region should be composed of hot, fractured rock containing a component of liquid brine. Hydrothermally driven effusion could cause some of this brine to be driven out on the surface of the crater floor. Hydrothermal modeling would be required to quantify volume of this effusion, but in general the amount extruded would be dependent on the volume and
10
ACCEPTED MANUSCRIPT
distribution of heated material in the subsurface, the efficacy by which fluids could separate from their post-impact location and migrate, and the pre-impact ice/rock ratio. Because Ceres’ surface is a near vacuum, liquid and solid brines extruded onto the surface would rapidly sublimate, leaving behind precipitate deposits like the observed faculae (Schenck et al, this
CR IP T
issue).
If brine extrusion was most rapid in the initial phase of post-impact crater modification, the earliest evaporate deposits within the crater would be associated with very short lived flows and ponds of liquid water on the surface, especially in the center of the crater where the subsurface is hottest. These liquid deposits could then be overlain by a second phase of
AN US
precipitate deposition in which ballistically driven salt rich ice particles would be ejected from fractures in the crater floor (Zolotov, 2017). If both of these deposition process occurred, evidence of the different depositional periods may manifest as albedo or spectral layering observable within regions of the faculae that have been extensively fractured or excavated by
M
small impacts.
ED
The most prominent facula within Occator is located at the very center of the crater within a central pit. Because this region of our models correlated with a deep plug of hot material heated above the melting point of water, it is reasonable to expect that the magnitude of
PT
extrusion and sublimation would be most extreme in this area. Additionally, gravitational settling of solid particles in a hot region with locally high ice/rock ratio may have formed a lake
CE
in the center of the crater immediately following impact. If such a lake existed, it would rapidly boil at the surface until the latent heat of sublimation produced a protective ice cover. In this
AC
state, subsurface flows from outside the crater center could recharge the temporary lake, concentrating the amount of solutes in the region and the volume of the associated precipitate deposit. Eventually, all liquid and solid H2O would sublimate or evaporate, potentially leaving behind a localized depression with a precipitate deposit at the bottom.
11
ACCEPTED MANUSCRIPT
If the extrusion and sublimation of impact heated brines onto the floor of Occator are indeed the formation mechanism which produced the observed faculae within the crater, a maximum timescale of deposition can be inferred based on the conductive cooling timescale of the postimpact thermal structure. Once the subsurface has entirely cooled beneath the eutectic of relevant brines, no more hydrothermal circulation or surface effusion could occur, and the
CR IP T
deposition of precipitate deposits on the crater floor would cease. The conductive cooling timescale provides an upper bound, as additional effects such as hydrothermal circulation and evaporative effects would only increase the rate at which the crater’s subsurface cools.
To estimate the thermal lifetime of liquids within the subsurface of Occator following its
AN US
formation, we use the cooling model of Davison et al. (2012). This model solved the conductive heat equation in two dimensions to quantify the post-impact thermal evolution of the crater. Within the model, a constant temperature of 150 K is imposed along the crater surface boundary, the vertical surface boundary corresponds to the geothermal gradient assumed in
M
our pre-impact iSALE target, and no geothermal heat flux is imposed at the base of the mesh. Material densities and temperatures are mapped directly from iSALE, and uniform values for
ED
heat capacity (1550 J kg-1 K-1) and thermal conductivity (2.7 W m-1 K-1) consistent with ice-rock mixtures (Barnhart et al., 2010) are used across the entire grid. The resolution of this model is the same as that of the iSALE models on which their initial conditions are based. Water/rock
PT
interactions and latent heat of crystallization are not accounted for in this model. Figure 4 shows the post-impact temperature evolution of one of our Occator simulations as a function
CE
of time. Initially, a hot plug of material in the center of the crater as well as the crater floor are above the melting point of water. Within 100 kyr, the floor of the crater has cooled off well
AC
below temperatures at which liquid brines could exist. Only a region within the initially hot central plug is still above the melting point of water. Within 400 kyr, the entire subsurface has cooled to the point that no more hydrothermal activity would be expected.
Several important caveats to the application of our cooling model to Occator need to be discussed. Firstly, if certain species with very low thermal conductivities, such as methane
12
ACCEPTED MANUSCRIPT
clathrates, dominate the volatile budget of Ceres’s subsurface, the cooling timescale could be considerably extended. Because the temperature rate of change within our simulations is linearly dependent on the thermal conductivity, we can estimate a liberal upper limit cooling timescale for Occator of about 4 Myr (corresponding to a very low thermal conductivity of 0.3 W m-1 K-1). Secondly, our cooling model does not include latent heat of fusion, which may
CR IP T
prolong the timescale of cooling. In our nominal model a total of 2.1 x 1021 J of heat is conducted from the crater subsurface as it cools. Including latent heat of fusion and
vaporization, assuming the total volume with melted volatiles is 4118 km3 and the total volume with vaporized volatiles is 150 km3, would add an additional 4.7 x 1020 J to the system. While this is not an inconsiderable amount, we do not expect that the inclusion of latent heat in our
AN US
model would drastically change our cooling timescale estimates. Thirdly, we use a relatively low thermal gradient in our models, in which the nominal eutectic of a brine in the pre-impact target is reached at approximately 50 km depth, we have not considered the influence of heat flux into the bottom of our cooling model, nor the influence of radiogenic heating within the
M
target, both of which are intimately tied to the geothermal gradient. Recent work describing Ceres’ topography may require a higher thermal gradient than used in our models to explain a
ED
low viscosity layer at depth (Fu et al., 2017). Qualitatively, the inclusion of radiogenic heating, basal heat flux, and/or a higher pre-impact thermal gradient within our cooling model would increase the cooling timescale, as the total volume of material heated above the nominal brine
PT
eutectic would be larger. However In a hotter pre-impact target the dynamics of crater formation and collapse would be altered, with more hot material being uplifted from depth,
CE
and it is difficult to decouple heat induced changes to the crater collapse/modification process from heat induced changes to the post-impact cooling process without a large suite of models
AC
covering many free parameters. Finally, and perhaps most importantly, our conductive cooling model cannot account for advective heat transfer via fluid flow within an active hydrothermal system immediately after crater formation. Qualitatively, the inclusion of convection should decrease the cooling timescale, as has been shown for other solar system craters (Abramov et al., 2004). To first order, the ratio of the advective cooling timescale to the conductive cooling timescale is proportional to the permeability of the system (Barnhart et al., 2010). Because the
13
ACCEPTED MANUSCRIPT
permeabilities of geologic media cover a huge range (>10 orders of magnitude) and the permeability of Occator’s subsurface is currently unknown, it is not possible to quantify how much a hydrothermal system could reduce the cooling timescale of Occator. However the effect could be considerable, and as such the cooling timescales reported in this paper should be considered as likely upper limits. Further modeling of Occator’s thermal evolution, including
CR IP T
a larger range of pre-impact geothermal gradients, subsurface compositions, the inclusion of latent heat of fusion, and the implementation of a hydrothermal model as an alternative to our conductive cooling model will be the subject of future work.
The age of Occator’s formation has been estimated based on small crater densities of deposits
AN US
on the crater floor. A recent work concludes that the age of Occator’s largest facula, located in the center of the crater, is approximately 30 Myr younger than the crater itself (Natheus, 2017). If this is the case, extrusion of brines directly driven by impact heat could not possibly be responsible for the formation of this particular deposit, as all fluids within the subsurface would
M
have cooled well beyond the freezing point on a timescale 1-2 orders of magnitude faster than observed. However, re-analysis of the cratering ages of the faculae and the rest of Occator
ED
suggest that there is little to no discernable difference in ages between the two structures, in which case impact induced heating provides a viable mechanism for the deposition of
PT
hydrothermally related deposits (Neeseman et al., this issue).
Depending on the distribution of volatiles in the subsurface before impact, the uplift of deep
CE
seated water rich material into the region directly beneath Occator’s center could provide a reservoir for later stage, non-impact heat driven cryovolcanism. In this study, we assume a
AC
fairly low pre-impact geothermal gradient for our Cerean target. However, if a higher thermal gradient is assumed, than the material that underlies the center of Occator crater may have had pre-impact temperatures at or above the melting point of liquid water. If this was the case, the cooling time of the crater may have been extended.
14
ACCEPTED MANUSCRIPT
6) Conclusions
Our simulations of the formation of Occator crater on dwarf planet 1 Ceres suggest that the impact should produce a considerable amount of target material heated above the melting point of water. The region of heated material is largely concentrated in a central plug at the
CR IP T
center of the crater, although a substantial amount of hot material is also spread across the crater floor by the collapse of the central uplift during crater modification. The volume of
heated water within the subsurface of the crater is dependent on the pre-impact ice/rock ratio of the target, with larger ice/rock ratios leading to larger volumes of hydrothermally viable post-impact target. As such, larger amounts of hydrothermal fluids would be available for
AN US
faculae formation in targets with higher ice/rock ratios. Assuming the faculae within Occator were deposited by impact heat driven hydrothermal processes, higher ice/rock ratios in the pre-impact target would likely lead to larger volume faculae deposits.
M
The volume of material within the crater’s subsurface that should contain post-impact liquid water is considerably larger than the faculae volumes described by Scully et al. (2017, this
ED
issue). We suggest that this hydrothermal reservoir, if extruded onto the surface of the crater floor either as a liquid or a solid and subsequently sublimated, can provide a compelling mechanism for facula formation. However, the lifetime of such a system would be limited to, at
PT
most, a few million years post impact. In either case, we suggest that volatile rich material uplifted from a pre-impact reservoir at considerable depth and deposited beneath the crater
CE
center may provide a region from which the material that formed the facula could be sourced.
AC
7) Acknowledgements
We would like to thank the developers of iSALE-2D, including Gareth Collins, Kai Wunnemann, Dirk Elbeshausen, Tom Davison, Boris Ivanov, and Jay Melosh. We would like to thank the member of NASA’s Dawn mission for making the underlying observations that form the basis for this work. This work was funded in part by NASA’s Postdoctoral Program.
15
ACCEPTED MANUSCRIPT
8) References A’Hearn, M. F. and P. D. Feldman (1992) Water Vaporization of Ceres, Icarus, 98, 54. Abramov, O. and Kring, D. A. (2004) Numerical modeling of an impact-induced hydrothermal system at the Sudbury crater, Journal of Geophysical Research, 109, E10007, doi:10.1029/2003JE002213.
CR IP T
Amsden, A. A., H. M. Ruppel, and C. W. Hirt (1980), SALE: A simplified ALE computer program for fluid flow at speeds, Los Alamos National Laboratories Report LA-8095. Arakawa, M. and D. Tomizuka (2004) Ice-silicate fractionation among icy bodies due to the difference of impact strength between ice and ice-silicate mixture. Icarus, 170, 193-201.
AN US
Barnhart, C. J., Nimmo, F., and B. J. Travis (2010) Martian post-impact hydrothermal systems incorporating freezing. Icarus, 208, 101-117. Bland, M.T., Raymond, C.A., Schenk, P.M., et al., 2016. Composition and structure of the shallow subsurface of Ceres revealed by crater morphology. Nat. Geosci. 9, 538–542. doi:10.1038/ngeo2743.
M
Bray, V. J., G. S. Collins, J. V. Morgan, H. J. Melosh, and P. M. Schenk (2014), Hydrocode simulation of Ganymede and Europa cratering trends— How thick is Europa’s crust? Icarus, 231, 394–406, doi:10.1016/j.icarus.2013.12.009.
ED
Brookshaw, L. 1998, An equation of state for serpentine, Tech. Rep. Working Paper Series SCMC-9813, Faculty of Sciences, University of Southern Queensland 2.1, 3.5
PT
Collins, G. S., H. J. Melosh, and B. A. Ivanov (2004), Modeling damage and deformation in impact simulations, Met. And Planet. Sci., 39, 217–231.
CE
Collins, G. S., H. J. Melosh, and K. Wünnemann (2010), Improvements to the e-a porous compaction model for simulating impacts into high-porosity solar system objects, Int. J. Impact Eng., 38, 434–493.
AC
Collins, GS (2014), Numerical simulations of impact crater formation with dilatancy, J. Geophys. Res. Planets, 119, 2600–2619 Collins, G. S., Elbeshausen, D., Davison, T. M.; Wünnemann, K.; Ivanov, B.; Melosh, H. J. (2016): iSALE-Dellen manual. figshare. Combe, J.-Ph., McCord, T.B., Tosi, F., et al., 2016. Detection of local H2O exposed at the surface of Ceres. Science 353, aaf3010. doi:10.1126/science.aaf3010.
16
ACCEPTED MANUSCRIPT
Davison, T. M., G. S. Collins, and F. J. Ciesla (2010), Numerical modelling of heating in porous planetesimal collisions, Icarus, 208(1), 468–481, doi:10.1016/j.icarus.2010.01.034. Davison, T.M., Ciesla, F.J. and Collins, G.S. (2012) Post-impact thermal evolution of porous planetesimals. Geochimica et Cosmochimica Acta, Volume 95, pp. 252–269(18).
CR IP T
De Sanctis, M.C., Ammannito, E., Raponi, A., et al., 2015. Ammoniated phyllosilicates with a likely outer Solar System origin on (1) Ceres. Nature 528, 241–244. doi:10.1038/nature16172. De Sanctis, M.C., Raponi, A., Ammannito, E., et al., 2016. Bright carbonate deposits as evidence of aqueous alteration on (1) Ceres. Nature 536, 54–57. doi:10.1038/ nature18290. De Sanctis, M.C., Ammannito, E., McSween, H.Y., et al., 2017. Localized aliphatic organic material on the surface of Ceres. Science 355, 719–722. doi:10.1126/ science.aaj2305.
AN US
Durham, W. B., S. H. Kirby, and L. A. Stern (1992), Effects of dispersed particulates on the rheology of water ice at planetary conditions, J. Geophys. Res., 97, 20,883–20,897, doi:10.1029/92JE02326.
Fu, R. R. et al. (2017) The interior structure of Ceres as revealed by surface topography. Earth and Planetary Science Letters, 476, 153-164.
M
Hayne, P.O., Aharonson, O., 2015. Thermal stability of ice on Ceres with rough topography. J. Geophys. Res. Planets 120, 1567–1584. doi:10.1002/2015JE004887.
ED
Hiesinger, H., Marchi, S., Schmedemann, N., et al., 2016. Cratering on Ceres: impli- cations for its crust and evolution. Science 353, aaf4759. doi:10.1126/science. aaf4759.
CE
PT
Ivanov, B. A., D. Deniem, and G. Neukum (1997), Implementation of dynamic strength models into 2-D hydrocodes: Applications for atmospheric breakup and impact cratering, Int. J. Impact Eng., 20, 411–430.
AC
Ivanov, B.A., Pierazzo, E., 2011. Impact cratering in H2O-bearing targets on Mars: thermal field under craters as starting conditions for hydrothermal activity. Me- teorit. Planet. Sci. 46, 601– 619. doi:10.1111/j.1945-5100.2011.01177.x. Johnson, B. C. et al. (2016a) Formation of the Sputnik Planum basin and the thickness of Pluto’s subsurface ocean, GRL, 10,068. Johnson, B. C. et al. (2016b) Spherule layers, crater scaling laws, and the population of ancient terrestrial impactors, Icarus, 271, 350-359. Kuppers, M. et al. (2014) Localized sources of water vapour on the dwarf planet (1) Ceres, Nature, 505, 525. 17
ACCEPTED MANUSCRIPT
Li, J.-Y., McFadden, L.A., Parker, J.Wm., et al., 2006. Photometric analysis of 1 Ceres and surface mapping from HST observations. Icarus 182, 143–160. doi:10.1016/j. icarus.2005.12.012. Melosh, H. J. (1989), Impact Cratering: A Geologic Process, Oxford Univ. Press, New York.
CR IP T
Melosh, H. J., E. V. Ryan, and E. Asphaug (1992), Dynamic fragmentation in impacts: Hydrocode simulation of laboratory impacts, J. Geophys. Res., 97, 14,735–14,795. Nathues, A.Hoffmann, M., Schaefer, et al., 2015. Sublimation in bright spots on (1) Ceres. Nature 528, 237–240. doi:10.1038/nature15754.
Nathues, A., Platz, T., Thangjam, G., 2017. Evolution of Occator Crater on (1) Ceres. Astronom. J. 153, 112. doi:10.3847/1538-3881/153/3/112, (12pp).
AN US
Neesemann, A., et al. The various ages of Occator crater, Ceres: Results of a c comprehensive synthesis approach. Icarus, this issue.
Ohnaka, M. (1995), A shear failure strength law of rock in the brittle-plastic transition regime, Geophys. Res. Lett., 22, 25–28.
M
Park, R.S., Konopliv, A.S., Bills, B.G., et al., 2016. A partially differentiated interior for (1) Ceres deduced from its gravity field and shape. Nature 537. doi:10.1038/ nature18955.
ED
Pierazzo, E. et al. (2005) Starting conditions for hydrothermal systems underneath Martian craters: Hydrocode modeling, GSA Special Paper 384.
PT
Platz, T., Nathues, A., Schorghofer, N., et al., 2016. Surface water-ice deposits in the northern shadowed regions of Ceres. Nat. Astron. 1. doi:10.1038/ s41550-016-0007, id. 0007.
AC
CE
Potter, R. W. K., D. A. Kring, and G. S. Collins (2015), Scaling of basin-sized impacts and the influence of target temperature, in Large Meteorite Impacts and Planetary Evolution V: Geological Society of America Special Paper, edited by Osinski, G. R., and Kring, D. A., vol. 518, pp. 99–113, Geol. Soc. of Am., Boulder, Colo., doi:10.1130/2015.2518(06). Prettyman, T.H., Yamashita, N., Toplis, M.J., et al., 2017. Extensive water ice within Ceres’ aqueously altered regolith: evidence from nuclear spectroscopy. Science 355, 55–59. doi:10.1126/science.aah6765. Ruesch, O.; Platz, T.; Schenk, P.; McFadden, L. A.; Castillo-Rogez, J. C.; Quick, L. C.; Byrne, S.; Preusker, F.; OBrien, D. P.; Schmedemann, N.; Williams, D. A.; Li, J.- Y.; Bland, M. T.; Hiesinger, H.; Kneissl, T.; Neesemann, A.; Schaefer, M.; Pasckert, J. H.; Schmidt, B. E.; Buczkowski, D. L.; Sykes, M. V.; Nathues, A.; Roatsch, T.; Hoffmann, M.; Raymond, C. A.; Russell, C. T. (2016). Cryovolcanism on Ceres, Science. 353 (6303): aaf4286–aaf4286. doi:10.1126/science.aaf4286. 18
ACCEPTED MANUSCRIPT
Russell, C. T. et al. (2016) Dawn arrives at Ceres: Exploration of a small, volatile-rich world, Science, 353, 6303 Schenk, P. et al., (2016) Impact crater on the small planets Ceres and Vesta: Central Pits and the Origin of Bright Spots. LPSC 47, 2697.
CR IP T
Schenk, P. M., et al. Geomorphology and formation of Occator crater and hydrothermal origins of its bright spots: interplanetary comparisons. Icarus, this issue.
Schmidt, B. E. et al., (2017) Geomorphological evidence for ground ice on dwarf planet Ceres, Nature Geoscience, 10, 338.
AN US
Stein, N., Ehlmann, B., Ammannito, E., et al., 2017. Characteristics, formation, and evolution of faculae (bright spots) on Ceres. Lunar Planet Sci. 48 Abstract 2547. Thomas, P. C. et al. (2005) Differentiation of the asteroid Ceres as revealed by its shape, 437, 224. Turtle, E. P., and E. Pierazzo (2001), Thickness of a Europan ice shell from impact crater simulations, Science, 294(5545), 1326–1328, doi:10.1126/science.1062492.
ED
M
Wünnemann, K., Collins, G., and Melosh, H. (2006). A strain-based porosity model for use in hydrocode simulations of impacts and implications for transient crater growth in porous targets. Icarus, 180:514--527.
PT
Yasui, M. and M Arakawa (2008) Experimental study on the rate depended strength of ice-silica mixture with silica volume fractions up to 0.63. GRL, 35, L12206.
CE
Yasui, M. and M. Arakawa (2010) Rate-dependent strength of porous ice-silica mixtures and its implications for the shape of small to middle-sized icy sattelites. Icarus, 210, 956-967.
AC
Zolotov, M. Y. (2017) Aqueous origins of bright salt deposists on Ceres, Icarus, 296, 289.
9) Acknowledgements
We gratefully acknowledge the developers of iSALE-2D, including Gareth Collins, Kai Wünnemann, Dirk Elbeshausen, Boris Ivanov and Jay Melosh. We also acknowledge Brandon Johnson and Gareth Collins for their suggestions and insight during the development of this work, and two anonymous reviewers for their thoughtful and helpful comments. Finally, we acknowledge NASA’s Dawn mission and science teams for their efforts providing the observations which underlie this work. 19
ACCEPTED MANUSCRIPT
10) Figures
B
AN US
CR IP T
A
D
AC
CE
PT
ED
M
C
Figure 1 Figure showing the formation timeline of an Occator-like crater in a target composed of 20% water ice and 80% serpentine by mass. (top left) Immediately following impact, the target is heated by the passage of the impact induced shock wave. (bottom left) The hottest material, including the impactor, is largely spread out across the crater floor. (top right) The collapse of the transient crater concentrates most of the highly heated material in a gravitationally unstable central uplift which subsequently collapses. (bottom right) The hottest material in the crater forms a plug directly beneath the center of Occator, with the denser
20
ACCEPTED MANUSCRIPT
impactor (red material) slowly settling to the bottom of the plug. Grey lines show tracer particles of equal pre-impact provenance depth. D
CR IP T
A
C
AN US
E
M
B
ED
A
AC
CE
PT
F
Figure 2 Figure showing the post-impact thermal fields of 3 Occator-like craters with target ice mass ratios of 10% (A), 20% (B), and 30% (C). In the lowest case (10%), most of the hot material forms a relatively shallow plug, or is spread out along the crater floor. As the amount of material heated above the melting point of ice is increased, the volume of the hot plug increases, as does its depth. Increasing the ice-rock ratio also leads to more distinct crater walls of steeper slope. Grey lines show tracer particles of equal pre-impact provenance depth.
21
ACCEPTED MANUSCRIPT
Panels D-F show the corresponding post-impact location of impactor material (red) and target material (grey). It can been clearly seen that the hottest material within each crater’s subsurface is impactor.
AN US
CR IP T
A
PT
ED
M
B
AC
CE
C
Figure 3 Figure showing the provenance depth of uplifted material beneath Ceres for 10% (A), 20% (B), and 30% (C) target ice mass ratios. Black lines correspond to isotherms of 273 and 230
22
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
M
AN US
CR IP T
K (the nominal melting point of water and of a brine eutectic). Each dot represents a massless Lagrangian tracer particle, which follow the flow of material as the simulation progresses, colored according to pre-impact provenance depth. In each simulation, the central hot plug is composed of a mixture of impactor and target excavated from considerable depth. Had a deep subsurface volatile reservoir existed before Occator formed, it would have been subsequently heated, uplifted, and deposited beneath the central floor of the crater.
23
ED
M
AN US
CR IP T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 4 Figure showing the conductive cooling of an Occator like structure (nominally similar to the 10% water by mass case). Within 100 kyr (B) after crater formation, hot material on the crater floor has completely cooled below the melting point of water and/or a brine eutectic. Hot material can remain above these temperature limits beneath the center of the crater for approximately 300-400 kyr (D-E), after which all possibility of impact induced hydrothermal activity will cease.
24
Grid Resolution 100 m Number of Horizontal Cells in Grid 1200 Number of Vertical Cells in Grid 800 Impactor Diameter 5.6-6.3 km Impactor Velocity 4.8 km s-1 Target Ice/Rock Ratio (by mass) 0.1-0.3 Target Suface Gravity 0.27 m s-2 Target Surface Temperature 150 K Pre-impact Target Temperature Gradient 0.5 K km-1 Table 1: Global Parameters Used in iSALE Simulations
Target ANEOS Serpentine/H2O 273 K 1318-1955 J kg-1 K-1 1.2 0.26-0.27 0.6 1.4
3.5 GPa 10 kPa 50 MPa 187-197 s 1.9-2.3x106 m2 s-1
AC
CE
PT
ED
M
AN US
Material Parameters Impactor Equation of State ANEOS Dunite Melting Temperature 1436 K Specific Heat Capacity 1000 J kg-1 K-1 Thermal Softening Parameter 2.0 Poisson Ratio 0.25 Coefficient of Internal 0.6 Friction (damaged) Coefficient of Internal 1.4 Friction (undamaged) Strength at Infinite Pressure 3.5 GPa Cohesion (damaged) 10 kPa Cohesion (undamaged) 50 MPa Acoustic Fluidization Time Decay Acoustic Fluidization Kinematic Viscosity Table 2: Material Parameters Used in iSALE Simulations
CR IP T
ACCEPTED MANUSCRIPT
25