Depletion of Heat Producing Elements in the …...Depletion of Heat Producing Elements in the Martian Mantle Lujendra Ojha1, Saman Karimi2, Kevin W. Lewis2, Suzanne E. Smrekar3, and

Depletion of Heat Producing Elementsin the Martian MantleLujendra Ojha1 , Saman Karimi2 , Kevin W. Lewis2 , Suzanne E. Smrekar3 ,and Matt Siegler4

1Department of Earth and Planetary Sciences, Rutgers, The State University of New Jersey, Piscataway, NJ, USA,2Department of Earth and Planetary Sciences, Johns Hopkins University, Baltimore, MD, USA, 3Jet PropulsionLaboratory, California Institute of Technology, Pasadena, CA, USA, 4Planetary Science Institute, Tucson, AZ, USA

Abstract Heat is primarily generated in planetary interiors by the decay of long‐lived heat producingelements (HPE). Planetary heat flow estimates can thus provide critical insights into the thermal state ofa planet and the bulk distribution of the HPE. The lack of appreciable lithospheric deflection in the northpolar region of Mars by the weight of the polar ice cap is suggestive of low heat flow. Here we model thedeflection of the Martian lithosphere and show that the present‐day mantle heat flow cannot exceed7 mW m−2 in the north polar region of Mars. Our mantle heat flow estimate is notably lower than the heatflow expected from a chondritic mantle suggesting the Martian mantle to be depleted in HPE. If our result isglobally representative, lower levels of heat generation in the planet's mantle may have inhibitedwidespread late‐stage volcanism on Mars.

Plain Language Summary The geophysical evolution of a planet is governed by the rate of heatproduction within the planet and the mechanism and efficiency through which that heat is lost. After theaccretion and differentiation of a planet, the major source of internal heat generation is through thedecay of long‐lived radiogenic elements such as uranium, thorium, and potassium. We constrain thepresent‐day thermal structure in the north polar region of Mars by modeling observed geophysical flexurebeneath the north polar cap. Our results suggest that most of the heat producing elements on Mars may besequestered in the crust and that the mantle is depleted in these elements. If globally representative, ourresults have important implications for the thermal and geological evolution of Mars.

1. Introduction

Planetary evolution is governed in large part by the thermal history of its interior. The thermal evolution of aplanet depends on the rates of heat production and heat loss. After accretion and differentiation, the majorsource of heat production in the interior of terrestrial planets is the decay of long‐lived heat producing ele-ments (HPE) with half‐lives of order a billion years or greater (e.g., 238U, 235U, 232Th, and 40K). If the heatgenerated within the mantle of a planet is cooled from above, then, the resulting thermal convection candrive mantle convection, provided that the viscosity of the mantle is sufficiently low. Mantle convectioninfluences crustal tectonics and volcanism, thus constraining the thermal state of a planet's mantle is impor-tant for understanding planetary thermal evolution.

The thermal state of the Earth's interior has been constrained by more than twenty thousand surface heatflow measurements, and large variations in heat flow with age, tectonic setting, and composition are foundwithin the continental and the oceanic crust (Jaupart & Mareschal, 2007). The average oceanic and conti-nental heat flow are 101 and 65 mW m−2, respectively (Jaupart & Mareschal, 2007). The average reducedheat flux component, which represents the heat flowing from the lower crust andmantle of the Earth, is esti-mated to lie between about 10 and 34 mW m−2 (McLennan & Taylor, 1996). The only body other than theEarth for which we have in situ heat flow measurements is the Moon. At Apollo 15 and 17 landing sites,the heat flow was measured to be 21 + 3 and 15 + 2 mW m−2, respectively (Langseth et al., 1976).

While our understanding of the Martian surface has significantly improved in the last two decades, ourknowledge of the interior of Mars and its evolution remains relatively poor. To date, we do not have anyin situ measurement of the Martian heat flow, and most estimates of the surface heat flow are inferred fromeither cooling models, using parameterized convection schemes and compositional estimates

©2019. American Geophysical Union.All Rights Reserved.

RESEARCH LETTER10.1029/2019GL085234

Key Points:• We constrain the Martian mantle

heat flow by modeling thelithospheric deflection from theweight of the north polar cap onMars

• We find that the mantle heat flowcannot exceed 7 mW m−2 on thenorth pole of Mars

• The estimated low mantle heat flowsuggests a strong fractionation ofheat producing elements into theMartian crust

Supporting Information:• Supporting Information S1

Correspondence to:L. Ojha,[email protected]

Citation:Ojha, L., Karimi, S., Lewis, K. W.,Smrekar, S. E., & Siegler, M. (2019).Depletion of Heat Producing Elementsin the Martian Mantle. GeophysicalResearch Letters, 46. https://doi.org/10.1029/2019GL085234

Received 3 SEP 2019Accepted 4 NOV 2019Accepted article online 9 NOV 2019

OJHA ET AL. 1

(e.g., Stevenson et al., 1983; Hauck, 2002), fully dynamical convection schemes (e.g., Kiefer & Li, 2009; Plesaet al., 2015, 2018), or lithospheric thickness estimates based on models of observed topographic loading (e.g.,McGovern, 2004; Phillips et al., 2008). Overall, these thermal models of Mars are in general agreement withplanetary cooling models that suggest that heat flow has steadily declined throughout the last four billionyears of Martian history.

An unresolved question regarding the modern Martian thermal state concerns the low heat flow valuesinferred from lithospheric loading models of the north polar cap of Mars (Phillips et al., 2008), whichrequires the subsurface temperatures to be much colder than expected from a mantle with chondritic rela-tive proportions of HPE. In their seminal paper, Phillips et al. (2008) compared the heat flow implied by thelack of flexure in the north polar region to the heat flow modeled from a standard chondritic parameterizedthermal model of Hauck and Phillips (2002). The nominal thermal model from Hauck and Phillips (2002)considered a bulk distribution coefficient D of 0.1 and estimated the mantle heat flux to be ~17 mW m−2.The mantle heat flow implied by the lack of flexure in the north polar region is substantially lower thanthe estimate provided by Hauck and Phillips (2002), which led Phillips et al. (2008) to conclude that thelower estimated heat flowmust be due to a “subchondritic”Mars. Alternative hypotheses. such as large‐scalespatial heterogeneity in mantle heat flow (Grott & Breuer, 2009; Kiefer & Li, 2009; Plesa et al., 2016), and astronger fractionation of the HPE into the Martian crust than assumed by Hauck and Phillips (2002) havesince then been proposed (e.g., Plesa et al., 2018) to explain the lack of lithospheric flexure in Martian northpolar region. Either of these scenarios have significant implications for Martian thermal evolution.

The Heat Flow and Physical Properties Probe onboard the Interior Exploration of Mars using SeismicInvestigations, Geodesy, and Heat Transport (InSight) mission is expected to provide the first direct, near‐surface heat flow measurement of Mars (Spohn et al., 2018). As illustrated in Figure 1, the surface heat flowexpected to be measured by InSight will include contributions from the crust, the lithospheric mantle, andthe heat flow from the lower mantle into the base of the lithosphere. On Mars, our ability to separate heatflow contributions from the crust and mantle requires knowledge of the crustal thickness as well as the dis-tribution of HPE in both the crust and mantle. While the Seismic Experiment for Interior Structure

Figure 1. Model setup used for the calculation of the mantle heat flow. (a) The surface heat flow (qs) will includecontributions from the heat generated within the crust (qc), upper mantle (qm), and the heat flow from the lowermantle into the base of the lithosphere (qb). Tc and Tm are the crustal and lithospheric thicknesses, respectively. (b)Schematic diagram illustrating the geometry, boundary conditions, and the geodynamical context for the finite elementsimulation (axisymmetric and not to scale). τ: shear stress; σn: normal stress; u: velocity. The stress from the weight ofPlanum Boreum and Gemina Lingula is approximated as a distributed load using a cosine fit to the observed load (seeFigure S1).

10.1029/2019GL085234Geophysical Research Letters

OJHA ET AL. 2

instrument onboard InSight is expected to constrain the crustal thickness at the InSight landing site(Panning et al., 2017), the distribution of the HPE with depth in the Martian crust is unknown (e.g., Hahnet al., 2011; McLennan, 2001; Taylor et al., 2007) and cannot be constrained reliably with any instrumentson board InSight. Therefore, the estimate of the mantle heat flow from surface heat flow measurements willhave a large range of uncertainty. Here we place an upper limit on the local mantle heat flow in the northpolar region of Mars by modeling the flexural response of the Martian lithosphere to the stress imposedby the weight of the north polar cap using state‐of‐the‐art finite element techniques. This constraint onman-tle heat flow at the north polar region of Mars provides a much‐needed, albeit regional, insight into thepresent‐day thermal state of the Martian mantle.

2. Methods

The heat flow measured at the surface (qs) of a single plate planet such as Mars is the net sum of heat origi-nating within the crust (qc), the lithospheric mantle (qm), and the heat provided by the lower mantle to thebase of the lithosphere (qb; Figure 1). An indirect estimate of qb may be made by measuring the response ofthe planetary lithosphere to the stress imposed by large topographic loads on the surface and estimating anumber of thermal and mechanical properties of the lithosphere. Depending on the scope of the problem,the term lithosphere can have several different meanings. We adopt the thermal definition of a lithosphere,which is the uppermost layer of a planet in which thermal energy is transferred primarily by conduction. Incontrast, heat is transferred entirely by convection in the underlying mantle (Figure 1). A thick lithospherewill resist bending when a topographic load is emplaced on its surface. Conversely, a thin lithosphere will bemore susceptible to bending when a topographic load is imposed on its surface. The thermal lithosphere isdistinguished from the elastic lithosphere (e.g., Phillips et al., 2008) which refers to that fraction of the ther-mal lithosphere that is sufficiently cold and rigid so that elastic stresses are not relaxed on time scales of order109 years. In the lower, hotter part of the thermal lithosphere, stresses relax by solid‐state creep processes.

In this study, we focus on the north polar region of Mars because radar instruments have been able to pro-vide tight constraint on the thickness of the polar caps (Nerozzi & Holt, 2019) which combined with the bulkdensity estimates of the north polar cap inferred from gravity data (Ojha et al., 2019) allow us to estimate thenormal stress imposed by the polar caps on the underlying lithosphere (Text S1; Figure 2). Geophysical flex-ure models constructed with these constraints can then be compared with the observed lithospheric flexure(e.g., Phillips et al., 2008; Selvans et al., 2010) to constrain theMartian heat flow. The north polar cap of Marsis composed of the ice‐rich north polar layered deposits (NPLD) and an underlying unit called the basal unitthat is richer in lithics. The largest section of the north polar cap is the topographic dome of Planum Boreum(PB), which is composed of both NPLD and basal unit. A smaller lobe of the cap called Gemina Lingula (GL)is composed entirely of the NPLD and separated from the main dome of PB by a large canyon called ChasmaBoreale (Figure 2). We treat PB and GL as axisymmetric distributed loads emplaced on top of the Martiancrust (Text S1; Figure S1). We input the 2‐D stress profiles from PB and GL into finite element models to con-strain the heat flow required to produce the magnitude of lithospheric deflection observed in radar data.Here we use the commercially available Marc‐Mentat finite element package (http://www.mscsoftware.com) to model the flexure of the lithosphere under various thermal and mechanical conditions. The finiteelement modeling consists of three parts: (i) building a finite element mesh, (ii) running a thermal simula-tion, and (iii) running a mechanical simulation. All three aspects of the model are run in Marc‐Mentat finiteelement package.

(i) Finite Element Mesh: We model lithospheric deflection in 2‐D using a two‐layer axisymmetric planarfinite element mesh—a single crustal layer underlain by a mantle. Given the relatively small wave-length of our topographic loads (<10°), we use a planar mesh instead of a spherical mesh. Sphericalmesh is only important when modeling the effects of topographic loads with large wavelength forwhich planetary curvature may play a role (Karimi & Dombard, 2016). We fit the observed stress pro-files of PB and GL with cosine curves and use the fitted stress profiles as an axisymmetric distributedload in our finite element method models (Figure 1; Figure S1). The depth of the mesh (~350 km) issimilar to the maximum thickness of the Martian stagnant lid (Spohn et al., 2001; Hauck, 2002) andis sufficiently wide (3,000 km) that the far edge boundaries do not affect the potential lithospheric flex-ure. The crustal thickness is set to the values determined by previous gravity studies (~35 km; e.g.,

10.1029/2019GL085234Geophysical Research Letters

OJHA ET AL. 3

Neumann et al., 2004). The resolution of the mesh decreases with depth such that the crust has a higherresolution than the mantle. Overall, the total number of elements in the mesh is on the order of 104.

(ii) Thermal Simulation: The viscosity structure within the lithosphere is ultimately controlled by the ther-mal structure, thus determining the temperature profile is a critical section of our models. We run asteady‐state thermal simulation in which we set the boundary conditions including the surface tem-peratures and varying mantle heat flow values and set the heat flow from the sides of the mesh to zero(Figure 1; Text S2). The mean annual surface temperature at the north polar region is estimated to be~165 K which we use as a boundary condition for the surface temperature. The north polar cap itself isnot included in the thermal model, and the surface temperature boundary condition applies to the topof the Martian crust. Following Zuber 2001, the density values of the crust and mantle are assumed tobe 2,900 kgm−3 and 3,500 kgm−3, respectively. We have tested the sensitivity of these values and foundthat reasonable variations in density values do not significantly affect the results of our simulations. Weset the mantle and the crustal thermal conductivity to 4 and 2.5 W m−1 K−1 (Schumacher & Breuer,2006), respectively. We verify the accuracy of the thermal simulation by comparing the results to ana-lytical solutions (Text S3; Figure S2).

(iii) Mechanical Simulation: The final output from the thermal simulation is used as an input to themechanical simulation (Text S4). We model the mechanical deformation of the lithosphere for a max-imum of 10 million years after the emplacement of the north polar cap. Over this geologically short

Figure 2. (a) Radar based thickness map showing themajor geographic units of the north polar ice cap. (b) Lithostatic stress exerted by the polar caps on the under-lying crust. The red and white dotted lines show the locations of the topographic profile of Planum Boreum (PB) and Gemina Lingula (GL) used for the flexuremodeling. (c) Lithospheric deflections expected from the weight of the GL for various qb. (d) Same as (c), but for PB. (e) Lithospheric deflection expected from theweight of GL for various qb and qc. (f) Lithospheric deflection expected from the weight of the GL for various qb assuming an anhydrous rheology. All otherparameters (e.g., crustal thickness, density, and thermal conductivity) were identical to the hydrous flexure model show in (c; see Table S1 for a summary of theparameters).

10.1029/2019GL085234Geophysical Research Letters

OJHA ET AL. 4

time frame, the temperature structure of the lithosphere is not significantly affected by the secular cool-ing of the planet or variations in surface temperature (Text S2). The crust and mantle are assumed to becompositionally uniform with temperature‐dependent viscoelastic rheology. The rheology of eachmodel element within the mesh is allowed to evolve with local thermal and mechanical conditions(i.e., no a priori constraints on the rheology; Text S4). We keep the minimum viscosity of the materialsto 1021 Pa·s in the simulation; we have tested smaller minimum viscosity values but found it to have anegligible impact on the final result (Text S5).

3. Results and Discussion

Despite the large stress imparted by the weight of the north polar cap on the lithosphere (Figure 2),Selvans et al. (2010) found small (<200 m) deflection of the lithosphere under PB. At GL, the NPLDrests directly on top of the late Hesperian aged (3.7–3.0 Ga) Vastitas Borealis Interior Unit (Brotherset al., 2015; Phillips et al., 2008), where radar observations likewise show very little downwarping ofthe surface indicative of a small (<100 m) deflection of the lithosphere (Phillips et al., 2008; Putziget al., 2009).

We use 2‐D profiles from our derived stress maps and model the expected lithospheric deflection using thetwo‐layer axisymmetric planar finite element mesh. We first consider a scenario in which we set qc and qm tozero and only vary qb. This is a non‐physical scenario as both the crust and the lithospheric mantle can belarge sources of planetary heat. However, this scenario enables us to place an upper limit on qb. Usinghydrous rheology and assumptions described above for the Martian crust and mantle (see Table S1 for asummary of the assumed parameters), our models show that qb higher than 7 mWm−2 leads to lithosphericdeflection larger than 100 m at GL and 200 m at PB (Figure 2c and 2d).

There are several different parameters involved in our models, the effect of which we comprehensivelyassess in the following sections. We will primarily focus on GL as our case study unless otherwise noted,as the radar constraint is most robust there (Phillips et al., 2008). During the early differentiation process,it has been proposed that more than half of the incompatible HPE are partitioned into the Martian crust(Hahn et al., 2011; McLennan, 2001; Taylor et al., 2007). Thus, qc could be a significant fraction of theMartian surface heat flow. Previously, Hahn et al. (2011) and Parro et al. (2017) estimated qc on Mars bycombining gamma ray spectrometer elemental abundance data with crustal thickness models. Theyassumed a vertically homogenous crust with no change in the concentration of HPE with depth and esti-mated the maximum qc to be ~13 mW m−2 on present‐day Mars. Thus, we set the mantle heat flow to ourupper limit of 7 mW m−2 and set qc to 13 mW m−2 assuming constant heat production with depth. For athick (>350 km) stagnant lid such as expected in present‐day Mars (e.g., Grott & Breuer, 2009; Spohnet al., 2001), the heat produced in a thin crust (~35 km), as expected in the north polar region from gravitydata (e.g., Neumann et al., 2004), has a minor influence on the magnitude of the lithospheric deflection(Figure 2e). If qc is higher than 13 mW m−2 in the north polar region of Mars, then an even lower amountof qb is required to limit the lithospheric deflection to under 100 m.

The Martian crust is thought to be essentially basaltic in composition, the thermal conductivity (kc) of whichgenerally lies between 2.5 to 4 W m−1 K−1, depending on temperature and porosity (Clauser & Huenges,2013). We ran models with various values of thermal conductivity, but it did not change our results notably(Figure S3). If an extremely low value of kc is used (1 W m−1 K−1), assuming a porous Martian crust (e.g.,Goossens et al., 2017), then the magnitude of the lithospheric deflection increases instead. Thus, a reason-able variation in the thermal conductivity of the Martian crust does not notably change our upper limiton the background heat flow at the north polar region of Mars.

The rheology of the crust and the mantle can have a large effect on the lithospheric deflection. The magni-tude of lithospheric deflection is reduced in the case of anhydrous rheology in comparison to hydrous rheol-ogy (Figure 2f; Figure S3). The inclusion of water and other volatiles leads to more ductile rheology and amore easily deformable lithosphere. If the Martian interior was less hydrated than previously thought, itcould contribute to the limited amount of lithospheric flexure observed in the north polar region (e.g.,Figure S3). However, geochemical evidence suggests that the Martian mantle is water‐rich (e.g., Giestinget al., 2015; McSween et al., 2001; Usui et al., 2015). Regardless, even if the Martian interior was

10.1029/2019GL085234Geophysical Research Letters

OJHA ET AL. 5

completely anhydrous, the lithospheric deflection by the weight of GL should still exceed the radar con-straint if the heat flow is higher than 7 mW m−2 (Figure 2f).

The thermal time constant of a 350‐km lithosphere is ~109 years which is orders of magnitude larger than thetime scale considered in this work (Text S2). Thus, variation in surface temperature has a negligible effect onthe magnitude of the lithospheric deflection (Figure S4). The magnitude of the deflection may be small if notenough time has passed since the emplacement of the surface loads for the lithosphere to reach quasi‐equilibrium. Our flexure models show that the majority of the deflection happens in less than 100,000 years(less than one obliquity cycle) after emplacement of the load (Figure S4). Substantial deflection should bepresent even if the ice caps were deposited as recently as ~10,000 years ago (Figure S4). The exact deposi-tional age of the north polar cap has been difficult to estimate, but it is unlikely to be younger than 10,000years (Landis et al., 2016; Tanaka et al., 2008). The lack of evidence for a young age of the polar deposits thusallows us to rule out an incomplete flexural response as the cause of the inferred low mantle heat flow atthe poles.

The heat flow estimates from this work would be underestimated if radar observations have underestimatedthe full magnitude of the downwarping of the surface in the north polar region of Mars. This is however unli-kely; to exceed the 100‐m surface downwarping beneath GL, the integrated permittivity of the NPLD wouldhave to be less than 2.6, which is highly unlikely for the ice‐dust mixtures of the NPLD and substantial lithicfraction that makes up the basal unit (Nerozzi & Holt, 2019; Ojha et al., 2019; Phillips et al., 2008). On theother hand, while lithic components can have dielectric constants greater than 3, the time‐to‐depth conver-sion for permittivity values greater than 3 would lead to a convex basal surface (Phillips et al., 2008). We aretherefore confident that reasonable errors in interpretation of the radar data would not significantly affectour constraints for the lithospheric flexure models. This reinforces the value of 7 mW m−2 as a likely upperlimit on the present‐day mantle heat flow at the north pole of Mars.

Our heat flow estimate can be compared with previous thermal models of Mars to ascertain if the Martianmantle is depleted in HPE. A variety of geochemical models have been proposed for Mars (e.g., Lodders &Fegley, 1997; Wänke et al., 1994); however, remote sensing observation from the gamma ray spectrometer(Boynton et al., 2004) suggests that the model by Wänke and Dreibus (1994) is most consistent with theobservation of K and Th on the Martian shallow subsurface (hereafter referred to as the WD model). TheWD model suggests the Martian interior to be depleted in K but has a chondritic relative abundance ofrefractory lithophile elements (U and Th). HPE are incompatible lithophile elements; thus, they are knownto be strongly fractionated in crust. If ~40% of the bulk HPE from the WD model is assumed to be fractio-nated in the crust, then the resulting mantle heat flow at 350 km exceeds 14 mW m−2 (Plesa et al., 2018).If the mantle heat flow in the north polar region of Mars is as high as 14 mW m−2, then the lithosphericdeflection should exceed 400 m at PB and 200 m at GL (Figure 2), both of which would be resolvable in radardata. Thus, our upper limit of 7 mW m−2 on the present‐day mantle heat flow at the north pole of Marsimplies that either the bulk abundance of the heat producing elements on Mars is lower than expected fromthe chondritic WD model (e.g., Phillips et al., 2008), that a larger proportion of the bulk HPE has been frac-tionated into the crust (e.g., Plesa et al., 2018), that there are large‐scale spatial heterogeneity in mantle heatflow (e.g., Grott & Breuer, 2009; Kiefer & Li, 2009; Plesa et al., 2016) or any combination thereof.

Analysis of most Martian meteorites suggests the primitive Martian interior to have HPE relative abun-dances in the same proportion as chondritic values (see Kiefer & Li, 2009 for a summary). Thus, it is unlikelythat the estimated low mantle heat flow is due to Mars that is substantially less chondritic than the WDmodel. In contrast, analyses of Martian meteorites provide unambiguous evidence for a strong fractionationof the HPE in the Martian crust. The apparent depletion of Al in Martian meteorites suggests that Mars dif-ferentiated early with a melting event that involved depths of at least several hundred kilometers. Duringthis early differentiation process, it has been proposed that more than half of the HPE would be partitionedinto the crust (McLennan, 2001; Taylor et al., 2007). A strong partitioning of the HPE elements into the crustlowers heat production in the Martian mantle (Figure 3), while the resulting higher crustal heat flow willonly have a minor effect on the lithospheric strength (Figure 2e). Kiefer and Li (2009) assumed 50–70% ofthe HPE to be fractionated in the crust and showed that the style of mantle convection (upwelling vs down-welling) controls the observed lateral variation in the lithospheric thickness on present‐day Mars. In regionsof possible mantle downwelling, such as the north polar region of Mars, Kiefer and Li (2009) estimated the

10.1029/2019GL085234Geophysical Research Letters

OJHA ET AL. 6

mantle heat flow to be ~9 mW m−2. Similarly, Plesa et al. (2018) showed that only a limited set of models,those that have a strong fractionation of HPE in the Martian crust, can satisfy all available geophysical,petrological, and geological constraints for Mars. The mantle heat flow constraint from our model isremarkably close to the value obtained by Plesa et al. (2018) for the north pole of ~8 mW m−2 in, whichover 67% HPE inventory is assumed to be concentrated in the Martian crust (Plesa et al., 2018). Thus, ourpreferred hypothesis to explain the inferred low mantle heat flow is that there is a strong fractionation(60–70%) of the HPE in the Martian crust and that the Martian mantle is depleted in the HPE.

The heat flowmeasurement by Heat Flow and Physical Properties Probe on the InSight lander may allow usto test if mantle heat flow varies significantly on Mars. Numerical thermal evolution models that assumechondritic relative proportion of HPE in the Martian interior predict the surface heat flow at the InSightlanding region to range between 17 and 21 mW m−2 (Plesa et al., 2016). If heat production on Mars is only70–80% that of chondritic levels as suggested by Phillips et al. (2008), then the surface heat flow will be pro-portionally lower than 17–21 mWm−2. However, if the lowmantle heat flow is due to strong fractionation ofthe HPE from the primitive mantle into the crust, then the surface heat flow will be close to the range pre-dicted by chondritic geochemical models of Martian thermal evolution.

4. Conclusion

Our results place a tentative upper limit on the mantle heat flow of Mars by modeling the response of theMartian lithosphere to stress imposed by the weight of the north polar ice cap. We find that the mantle heatflow likely does not exceed 7 mW m−2 in the northern polar region of Mars. If the tentative upper limit onthe mantle heat flow from our work is globally representative of Mars, then the strong fractionation betweenthe crust and mantle on Mars may have precluded the mantle from undergoing late‐stage widespread melt-ing (Figure 3), significantly affecting the geological history of Mars (McLennan, 2001; Hauck, 2002).Depletion of the HPE in the Martian mantle along with a gradually thickening lithosphere may have ledto the cessation of widespread late‐stage volcanism on Mars.

ReferencesBoynton, W. V., Feldman, W. C., Mitrofanov, I. G., Evans, L. G., Reedy, R. C., Squyres, S. W., et al. (2004). The Mars Odyssey gamma‐ray

spectrometer instrument suite. Space Science Reviews, 110, 37–83. https://doi.org/10.1023/B:SPAC.0000021007.76126.15Brothers, T. C., Holt, J. W., & Spiga, A. (2015). Planum Boreum basal unit topography, Mars: Irregularities and insights from SHARAD. J.

Geophys. Res. E Planets, 120, 1357–1375. https://doi.org/10.1002/2015JE004830Clauser, C., & Huenges, E. (2013). Thermal conductivity of rocks and minerals. Rock Physics Phase Relations.. https://doi.org/10.1029/

RF003p0105Giesting, P. A., Schwenzer, S. P., Filiberto, J., Starkey, N. A., Franchi, I. A., Treiman, A. H., et al. (2015). Igneous and shock processes

affecting chassignite amphibole evaluated using chlorine/water partitioning and hydrogen isotopes. Meteoritics and Planetary Science,50(3), 433–460. https://doi.org/10.1111/maps.12430

Figure 3. The temperature profile of the Martian lithosphere with depth showing the effect of crustal fractionation of heatproducing elements. The models assume that the surface heat flow on Mars does not exceed 21 mW m−2.

10.1029/2019GL085234Geophysical Research Letters

OJHA ET AL. 7

AcknowledgmentsWe would like to thank Dr. StevenHauck II for his assistance with thecomputation. The radar data used inthis work are a published database thatmay be obtained from Nerozzi and Holt(2019) (https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2019GL082114). No other data aregenerated as part of this work. Reviewsfrom Paul Morgan and an anonymousreviewer significantly helped improvethe paper.

Goossens, S., Sabaka, T. J., Genova, A., Mazarico, E., Nicholas, J. B., & Neumann, G. A. (2017). Evidence for a low bulk crustal density forMars from gravity and topography. Geophysical Research Letters, 44, 7686–7694. https://doi.org/10.1002/2017GL074172

Grott, M., & Breuer, D. (2009). Implications of large elastic thicknesses for the composition and current thermal state of Mars. Icarus,201(2), 540–548. https://doi.org/10.1016/j.icarus.2009.01.020

Hahn, B. C., McLennan, S. M., & Klein, E. C. (2011). Martian surface heat production and crustal heat flow fromMars Odyssey gamma‐rayspectrometry. Geophysical Research Letters, 38(14), L14203. https://doi.org/10.1029/2011GL047435

Hauck, S. A., & Phillips, R. J. (2002). Thermal and crustal evolution of Mars. Journal of Geophysical Research, 107, 6–1. https://doi.org/10.1029/2001JE001801

Jaupart, C., & Mareschal, J.‐C. (2007). Heat flow and thermal structure of the lithosphere. Treatise on geophysics., 6, 217–252. https://doi.org/10.1016/B978‐0‐444‐53802‐4.00114‐7

Karimi, S., & Dombard, A. J. (2016). On the possibility of viscoelastic deformation of the large south polar craters and true polar wander onthe asteroid Vesta. Journal of Geophysical Research: Planets, 121, 1786–1797. https://doi.org/10.1002/2016JE005064

Kiefer, W. S., & Li, Q. (2009). Mantle convection controls the observed lateral variations in lithospheric thickness on present‐day Mars.Geophysical Research Letters, L18,203. https://doi.org/10.1029/2009GL039827

Landis, M. E., Byrne, S., Daubar, I. J., Herkenhoff, K. E., & Dundas, C. M. (2016). A revised surface age for the north polar layered depositsof Mars. Geophysical Research Letters, 43, 3060–3068.

Langseth, Keihm, Peters, K., Langseth, M. G., Keihm, S. J., Peters, K., 1976. Revised lunar heat‐flow values. Proc Lunar Sci Conf 7th.https://doi.org/10.1007/s13398‐014‐0173‐7.2

Lodders, K., & Fegley, B. (1997). An oxygen isotope model for the composition of Mars. Icarus, 126, 373–394. https://doi.org/10.1006/icar.1996.5653

McGovern, P. J. (2004). Correction to “Localized gravity/topography admittance and correlation spectra on Mars: Implications for regionaland global evolution.”. Journal of Geophysical Research, 109, E07007. https://doi.org/10.1029/2004JE002286

McLennan, S. M. (2001). Crustal heat production and the thermal evolution of Mars. Geophysical Research Letters, 28, 4019–4022. https://doi.org/10.1029/2001GL013743

McLennan, S. M., & Taylor, S. R. (1996). Heat flow and the chemical composition of continental crust. Journal of Geology, 104, 369–377.McSween, H. Y., Grove, T. L., Lentz, R. C., Dann, J. C., Holzheid, A. H., Riciputi, L. R., & Ryan, J. G. (2001). Geochemical evidence for

magmatic water within Mars from pyroxenes in the Shergotty meteorite. Nature, 409(6819), 487–490. https://doi.org/10.1038/35054011Nerozzi, S., & Holt, J. W. (2019). Buried ice and sand caps at the north pole of Mars: Revealing a record of climate change in the cavi unit

with SHARAD. Geophysical Research Letters, 46(13), 7278–7286. https://doi.org/10.1029/2019GL082114Neumann, G. A., Zuber, M. T., Wieczorek, M. A., McGovern, P. J., Lemoine, F. G., & Smith, D. E. (2004). Crustal structure of Mars from

gravity and topography. Journal of Geophysical Research, E Planets, 109. https://doi.org/10.1029/2004JE002262Ojha, L., Nerozzi, S., & Lewis, K. (2019). Compositional constraints on the north polar cap of Mars from gravity and topography.

Geophysical Research Letters, 46(15), 8671–8679. https://doi.org/10.1029/2019GL082294Panning, M. P., Lognonné, P., Bruce Banerdt, W., Garcia, R., Golombek, M., Kedar, S., et al. (2017). Planned products of the Mars structure

service for the InSight mission to Mars. Space Science Reviews, 211(1‐4), 611–650. https://doi.org/10.1007/s11214‐016‐0317‐5Parro, L., Jiménez‐Díaz, A., Mansilla, F., & Ruiz, J. (2017). Present‐day heat flowmodel of Mars. Scientific Reports, 7, 45629, https://doi.org/

10.1038/srep45629Phillips, R. J., Zuber, M. T., Smrekar, S. E., Mellon, M. T., Head, J. W., Tanaka, K. L., et al. (2008). Mars North polar deposits: Stratigraphy,

age, and geodynamical response. Science, 320(5880), 1182–1185. https://doi.org/10.1126/science.1157546Plesa, A. C., Grott, M., Tosi, N., Breuer, D., Spohn, T., & Wieczorek, M. A. (2016). How large are present‐day heat flux variations across the

surface of Mars? Journal of Geophysical Research: Planets, 121, 2386–2403. https://doi.org/10.1002/2016JE005126Plesa, A. C., Padovan, S., Tosi, N., Breuer, D., Grott, M., Wieczorek, M. A., et al. (2018). The thermal state and interior structure of Mars.

Geophysical Research Letters, 45(22), 12,198–12,209. https://doi.org/10.1029/2018GL080728Plesa, A. C., Tosi, N., Grott, M., & Breuer, D. (2015). Thermal evolution and Urey ratio of Mars. Journal of Geophysical Research: Planets,

120, 995–1010. https://doi.org/10.1002/2014JE004748Putzig, N. E., Phillips, R. J., Campbell, B. A., Holt, J. W., Plaut, J. J., Carter, L. M., et al. (2009). Subsurface structure of Planum Boreum from

Mars Reconnaissance Orbiter Shallow Radar soundings. Icarus, 204, 443–457. https://doi.org/10.1016/j.icarus.2009.07.034Schumacher, S., & Breuer, D. (2006). Influence of a variable thermal conductivity on the thermochemical evolution of Mars. Journal of

Geophysical Research, E Planets, 111(E2). https://doi.org/10.1029/2005JE002429Selvans, M. M., Plaut, J. J., Aharonson, O., & Safaeinili, A. (2010). Internal structure of Planum Boreum, from Mars advanced radar for

subsurface and ionospheric sounding data. Journal of Geophysical Research, E Planets, 115(E9). https://doi.org/10.1029/2009JE003537Spohn, T., Acuña, M.H., Breuer, D., Golombek, M., Greeley, R., Halliday, A., et al., 2001. Geophysical constraints on the evolution of Mars,

in: Space Science Reviews. https://doi.org/10.1023/A:1011949306989Spohn, T., Grott, M., Smrekar, S. E., Knollenberg, J., Hudson, T. L., Krause, C., et al. (2018). The Heat Flow and Physical Properties Package

(HP3) for the InSight mission. Space Science Reviews, 214(5), 1–33. https://doi.org/10.1007/s11214‐018‐0531‐4Stevenson, D. J., Spohn, T., & Schubert, G. (1983). Magnetism and thermal evolution of the terrestrial planets. Icarus, 54(3), 466–489.

https://doi.org/10.1016/0019‐1035(83)90241‐5Tanaka, K. L., Rodriguez, J. A. P., Skinner, J. A., Bourke, M. C., Fortezzo, C. M., Herkenhoff, K. E., et al. (2008). North polar region of Mars:

Advances in stratigraphy, structure, and erosional modification. Icarus, 196, 318–358. https://doi.org/10.1016/j.icarus.2008.01.021Taylor, G. J., Boynton, W. V., Brückner, J., Wänke, H., Dreibus, G., Kerry, K. E., et al. (2007). Bulk composition and early differentiation of

Mars. Journal of Geophysical Research, E Planets, 112(E3). https://doi.org/10.1029/2005JE002645Usui, T., Alexander, C. M. O. D., Wang, J., Simon, J. I., & Jones, J. H. (2015). Meteoritic evidence for a previously unrecognized hydrogen

reservoir on Mars. Earth and Planetary Science Letters, 410, 140–151. https://doi.org/10.1016/j.epsl.2014.11.022Wänke, H., & Dreibus, G. (1994). Chemistry and accretion history of Mars. Philosophical Transactions of the Royal Society London A, 359,

285–293. https://doi.org/10.1098/rsta.1994.0132

10.1029/2019GL085234Geophysical Research Letters

OJHA ET AL. 8