Figures and Tables

top

Introduction

The conversion of lherzolite and harzburgite to an orthopyroxene-poor or -free, clinopyroxene-rich rock classified as wehrlite – a process hereafter referred to as wehrlitisation – requires interaction with silica-undersaturated (ultra)basic melts (e.g., Wallace and Green, 1988

Wallace, M.E., Green, D.H. (1988) An experimental determination of primary carbonatite magma composition. Nature 335, 343–346.

; Yaxley et al., 1998

Yaxley, G.M., Green, D.H., Kamenetsky, V. (1998) Carbonatite metasomatism in the southeastern Australian lithosphere. Journal of Petrology 39, 1917–1930.

). Such melts encompass carbonatites, carbonated silicate melts (e.g., kimberlite) or CO₂-bearing silicate melts (e.g., melilitites and nephelinites), which can form by near-solidus melting of peridotite (e.g., Gudfinnsson and Presnall, 2005

Gudfinnsson, G.H., Presnall, D.C. (2005) Continuous gradations among primary carbonatitic, kimberlitic, melilititic, basaltic, picritic, and komatiitic melts in equilibrium with garnet lherzolite at 3-8GPa. Journal of Petrology 46, 1645–1659.

). Given the strong incompatibility of CO₂ in peridotite, low-volume melts are typically carbonated even if the source is not specifically C-rich, as long as the mantle source lies above the depth of redox melting (Hirschmann, 2010

Hirschmann, M.M. (2010) Partial melt in the oceanic low velocity zone. Physics of the Earth and Planetary Interiors 179, 60–71.

). In extensional continental settings, small-volume melts generated in the deep lithospheric or convecting mantle traverse ∼100 to 250 km of subcontinental lithospheric mantle (SCLM) and crust, with which they are initially out of chemical equilibrium, causing extensive reactions to occur (e.g., McKenzie, 1989

McKenzie, D. (1989) Some remarks on the movement of small melt fractions in the mantle. Earth and Planetary Science Letters 95, 53–72.

). Wehrlitisation can result from such reactions and involves the liberation of CO₂ vapour (e.g., Wallace and Green, 1988

Wallace, M.E., Green, D.H. (1988) An experimental determination of primary carbonatite magma composition. Nature 335, 343–346.

; Yaxley et al., 1998

Yaxley, G.M., Green, D.H., Kamenetsky, V. (1998) Carbonatite metasomatism in the southeastern Australian lithosphere. Journal of Petrology 39, 1917–1930.

). This, in turn, contributes to diffuse continental degassing, especially in rift settings where lithosphere thinning has occurred (Brune et al., 2017

Brune, S., Williams, S.E., Muller, R.D. (2017) Potential links between continental rifting, CO₂ degassing and climate change through time. Nature Geoscience 10, 941–946.

; Foley and Fischer, 2017

Foley, S.F., Fischer, T.P. (2017) An essential role for continental rifts and lithosphere in the deep carbon cycle. Nature Geoscience 10, 897–902.

). It is noticeable that wehrlites are frequently reported in the literature for basalt-borne xenolith suites associated with rifts, faults and basins. These structures are pathways for CO₂-rich fluids (Tamburello et al., 2018

Tamburello, G., Pondrelli, S., Chiodini, G., Rouwet, D. (2018) Global-scale control of extensional tectonics on CO₂ earth degassing. Nature Communications 9, 4608.

). However, a link between the release of CO₂ to the exosphere during diffuse, non-volcanic degassing and a specific petrological mechanism remains unexplored, and the passage of carbon through the lithosphere is itself poorly documented. Wehrlite-bearing xenolith suites have been entrained in magmas of various ages. Using literature data, we show that wehrlites, as both products and monitors of the passage of CO₂-bearing melts, can reveal the otherwise hidden CO₂ flux through the shallow SCLM, and its eventual tectonic degassing both in currently and formerly active rifts.

top

Depths, Hallmarks and Agents of Wehrlitisation

Published data show that basalt-borne xenolith suites from the spinel facies SCLM (∼40–100 km), mostly associated with off-cratonic lithosphere or cratonic lithosphere in various states of disruption and decratonisation, contain significant proportions of wehrlite (Table S-1). The data compilation encompasses garnet-free xenoliths from various on- and off-cratonic rift systems and basins (Supplementary Information). Indeed, the decarbonation reaction, which causes wehrlitisation, has been experimentally demonstrated to occur at relatively shallow depths corresponding to ∼1.5 to 2.0 GPa (Wallace and Green, 1988

Wallace, M.E., Green, D.H. (1988) An experimental determination of primary carbonatite magma composition. Nature 335, 343–346.

, and references therein). This is because the carbonated peridotite solidus features a “ledge” in pressure-temperature space so that on upward movement of carbonatite melts, they must freeze and react to form wehrlites (Wallace and Green 1988

Wallace, M.E., Green, D.H. (1988) An experimental determination of primary carbonatite magma composition. Nature 335, 343–346.

). There is pervasive evidence for the interaction of SiO₂-undersaturated, CO₂-bearing melts also with the deep lithosphere, and carbonatitic high-density fluids are observed in diamonds (Fig. 1, Table S-2). Nevertheless, the proportion of wehrlitic garnet both in xenoliths and as inclusions in diamond is minute (∼1 %) compared to the lherzolite or harzburgite paragenesis (Fig. S-1). Thus, wehrlitisation is not an important process in the garnet- and diamond-stable part of the SCLM (>∼60 and ∼120 km, respectively, depending on geotherm), which typically records lower oxygen fugacities than the shallow SCLM (Supplementary Information). Interaction of carbonated melts with the deep lithosphere leads to graphite/diamond precipitation instead, through a process called redox freezing (Rohrbach and Schmidt, 2011

Rohrbach, A., Schmidt, M.W. (2011) Redox freezing and melting in the Earth’s deep mantle resulting from carbon-iron redox coupling. Nature 472, 209–212.

).

Full size image

Wehrlitisation leads to high clinopyroxene modes at the expense of orthopyroxene relative to other peridotites in the same xenolith suite. Transitional rock types affected by the same process, but at lower melt-rock ratios, also occur, resulting in clinopyroxene-rich lherzolites and orthopyroxene-poor harzburgites with clinopyroxene/orthopyroxene ratios >1 (e.g., Yaxley and Green, 1996

Yaxley, G.M., Green, D.H. (1996) Experimental reconstruction of sodic dolomitic carbonatite melts from metasomatised lithosphere. Contributions to Mineralogy and Petrology 124, 359–369.

; Lin et al., 2020

Lin, A.-B., Zheng, J.-P., Aulbach, S., Xiong, Q., Pan, S.-K., Gerdes, A. (2020) Causes and consequences of wehrlitization beneath a trans-lithospheric fault: Evidence from Mesozoic basalt-borne wehrlite xenoliths from the Tan-Lu fault belt, North China Craton. Journal of Geophysical Research: Solid Earth 124, e2019JB019084.

). Depending on the style of melt interaction (porous flow vs. fractures), olivine-rich, pyroxene-poor to -free dunites may ultimately form (e.g., Shaw et al., 2018

Shaw, C.S.J., Lebert, B.S., Woodland, A.B. (2018) Thermodynamic Modelling of Mantle–Melt Interaction Evidenced by Veined Wehrlite Xenoliths from the Rockeskyllerkopf Volcanic Complex, West Eifel Volcanic Field, Germany. Journal of Petrology 59, 59–86.

). The metasomatic agents inferred for the various wehrlite-bearing xenolith suites range from carbonatite to carbonated silicate melts to CO₂-bearing silicate melts (Supplementary Information), which reflect increasing melt volumes and dilution with silicate components (e.g., Gudfinnsson and Presnall, 2005

Gudfinnsson, G.H., Presnall, D.C. (2005) Continuous gradations among primary carbonatitic, kimberlitic, melilititic, basaltic, picritic, and komatiitic melts in equilibrium with garnet lherzolite at 3-8GPa. Journal of Petrology 46, 1645–1659.

). Based on experimentally produced melts, there is a relationship between melt fraction and CO₂ (Fig. 1a) and SiO₂ content (Fig. 1b), producing an inverse correlation between CO₂ and SiO₂ (Fig. 1c) and between SiO₂ and CaO (Fig. 1d). The higher melt volumes involved in the generation of SiO₂-rich melts may therefore compensate for their lower CO₂ contents in terms of their ability to convert a given amount of orthopyroxene to clinopyroxene. The link between silica-undersaturated carbonated melts, wehrlites and decarbonation at low pressure is strengthened by direct observations of CO₂-rich fluid inclusions, carbonates, carbonate-bearing glass veins and melt pockets in wehrlites, with entrapment pressures of 0.8 to 1.7 GPa (e.g., Yaxley et al., 1998

Yaxley, G.M., Green, D.H., Kamenetsky, V. (1998) Carbonatite metasomatism in the southeastern Australian lithosphere. Journal of Petrology 39, 1917–1930.

; Loges et al., 2019

Loges, A., Schultze, D., Klügel, A., Lucassen, F. (2019) Phonolitic melt production by carbonatite Mantle metasomatism: evidence from Eger Graben xenoliths. Contributions to Mineralogy and Petrology 174, 93.

). It is further supported by experimental studies which show that wehrlite forms in equilibrium with carbonated melts (Yaxley and Green, 1996

Yaxley, G.M., Green, D.H. (1996) Experimental reconstruction of sodic dolomitic carbonatite melts from metasomatised lithosphere. Contributions to Mineralogy and Petrology 124, 359–369.

; Gervasoni et al., 2017

Gervasoni, F., Klemme, S., Rohrbach, A., Grützner, T., Berndt, J. (2017) Experimental constraints on mantle metasomatism caused by silicate and carbonate melts. Lithos 282–283, 173–186.

).

The effects of wehrlitisation on major element contents vary (Fig. S-2), reflecting the spectrum of SiO₂-undersaturated carbonated melts. In some suites (e.g., Eifel, North Atlantic Craton in Greenland), clinopyroxene in wehrlite is dominated by elevated CaO/Al₂O₃ (Figs. 2a, S-2), in others by elevated FeO (Fig. 2b). The effects on the trace element budget are also heterogeneous (Fig. S-3), depending not only on the identity of the metasomatic agent and type of melt-rock reaction, but also on lithosphere thickness, as the melt traverses and equilibrates with garnet-bearing peridotite in thick lithospheres (Supplementary Information).

Full size image

top

Modelling CO₂ Loss Via Wehrlitisation

The amount of CO₂ liberated from a volume of wehrlite-bearing peridotite due to interaction with CO₂-bearing, silica-undersaturated melt is estimated based on the decarbonation reaction

(e.g., Yaxley and Green, 1996

Yaxley, G.M., Green, D.H. (1996) Experimental reconstruction of sodic dolomitic carbonatite melts from metasomatised lithosphere. Contributions to Mineralogy and Petrology 124, 359–369.

). Dolomite is assumed to be dissolved in the metasomatic melt, and wehrlites with high clinopyroxene modes formed from lherzolites and harzburgites with lower clinopyroxene modes (Wallace and Green, 1988

Wallace, M.E., Green, D.H. (1988) An experimental determination of primary carbonatite magma composition. Nature 335, 343–346.

; Yaxley et al., 1998

Yaxley, G.M., Green, D.H., Kamenetsky, V. (1998) Carbonatite metasomatism in the southeastern Australian lithosphere. Journal of Petrology 39, 1917–1930.

). The proportion of a rock mass affected by wehrlitisation is estimated by counting Fe-rich “reaction” dunites and orthopyroxene-poor harzburgites and lherzolites with wehrlites (hereafter “wehrlite-group peridotites”), compared to “other peridotites” comprising harzburgites and lherzolites with clinopyroxene/orthopyroxene ratios <1. Reaction dunites, with low pyroxene modes, form during open-system processes (e.g., Shaw et al., 2018

Shaw, C.S.J., Lebert, B.S., Woodland, A.B. (2018) Thermodynamic Modelling of Mantle–Melt Interaction Evidenced by Veined Wehrlite Xenoliths from the Rockeskyllerkopf Volcanic Complex, West Eifel Volcanic Field, Germany. Journal of Petrology 59, 59–86.

), whereas here an equilibrium process is modelled. Therefore, reaction dunites and olivine-rich harzburgites with high clinopyroxene/orthopyroxene ratios are not considered in the calculation of the median clinopyroxene abundance or composition in the wehrlite-group peridotites (Table S-3).

The calculations assume closed-system reactions, resulting in minimum estimates for the volume of silica-undersaturated melt passing through the shallow lithosphere (Yaxley et al., 1998

Yaxley, G.M., Green, D.H., Kamenetsky, V. (1998) Carbonatite metasomatism in the southeastern Australian lithosphere. Journal of Petrology 39, 1917–1930.

). Because Equation 1 is based on diopside, whereas natural clinopyroxenes are not pure diopside, the difference in median clinopyroxene modes between wehrlite-group and other peridotites is weighted by the median diopside component in clinopyroxene (ΔDi). The mass of enstatite required to produce ΔDi is calculated based on the decarbonation reaction, with two moles of CO₂ liberated per mole of diopside formed (Eq. 1). The mass of CO₂ that passed through the shallow lithosphere is estimated by weighting the mass of CO₂ liberated during wehrlitisation by the proportion of wehrlite-group peridotites in each sample suite (Table S-3). The variation of the wehrlite-group proportion in multiple xenolith suites per area is taken as the uncertainty. The continental lithospheric area affected by wehrlitisation is estimated using area estimates from the literature or information on the total length of associated rifts, such as the Eastern Rift in the East African Rift and the European Cenozoic Rift System (Supplementary Information). Assuming a density of 3350 kg m⁻³ and a conservative 10 km interval of wehrlitisation, the total mass of liberated CO₂ is estimated. Finally, disequilibrium textures indicate that wehrlitisation occurred close in time to entrainment in the host magma, and that no major fluid- or melt-rock interaction has occurred since. Thus, wehrlitisation and CO₂ degassing are taken to be related to periods of active extension, which facilitates the formation and mobility of small-volume melts (Supplementary Information).

top

Results and Discussion

In the xenolith suites considered, 6–15 % clinopyroxene was added due to wehrlitisation (Table S-3), corresponding to ΔDi of 3.5 to 12 % and requiring conversion of 6.4 to 22 kg of enstatite (Table 1). Weighted by the proportion of wehrlite-group peridotites in the spinel facies rock column, this amounts to 0.2 ± 0.1 kg (Hoggar Swell) to 2.4 ± 1.5 kg (Eastern Rift) CO₂ per 100 kg of wehrlitised peridotite (Table 1). The continental area affected by wehrlitisation ranges from 110 × 10³ km² (European Cenozoic Rift System) to 4500 × 10³ km² (Tan Lu Fault Belt). Taking the proportion (and its variability gauged as 1σ) of wehrlite-group peridotites entrained with basalts in each area as representative, this yields total masses of released CO₂ from 24 (±15) × 10³ Gt to 2100 (±1700) × 10³ Gt. For estimated timespans of activity from 10 to 40 Myr, this amounts to CO₂ fluxes of 1.4 ± 0.1 Mt yr⁻¹ (Pannonian Basin) to 70 ± 58 Mt yr⁻¹ (Tan Lu Fault Belt) (Table 1). These estimates indicate significant mantle contributions to the total tectonic and volcanic CO₂ flux at the time of active rifting. For comparison, for conservative flux densities, 40 Mt CO₂ yr⁻¹ is estimated for combined present-day active rifts (Brune et al., 2017

Brune, S., Williams, S.E., Muller, R.D. (2017) Potential links between continental rifting, CO₂ degassing and climate change through time. Nature Geoscience 10, 941–946.

). Further, amplitudes >80 Mt CO₂ yr⁻¹ are estimated for the “Mesozoic CO₂ high”, which was associated with a total rift length of 50,000 km (Brune et al., 2017

Brune, S., Williams, S.E., Muller, R.D. (2017) Potential links between continental rifting, CO₂ degassing and climate change through time. Nature Geoscience 10, 941–946.

). This suggests that the Tan Lu Fault Belt, which was most active in the Early Cretaceous, was a major contributor to the contemporaneous greenhouse climate. For the Eastern Rift, we obtain ∼6.5 ± 4.1 Mt CO₂ yr⁻¹ compared to 6–18 Mt CO₂ yr⁻¹ attributed to magmatic intrusions into the crust based on surficial CO₂ flux measurements (Hunt et al., 2017

Hunt, J.A., Zafu, A., Mather, T.A., Pyle, D.M., Barry, P.H. (2017) Spatially variable CO₂ degassing in the main Ethiopian rift: implications for magma storage, volatile transport, and rift-related emissions. Geochemistry Geophysics Geosystems 18, 3714–3737.

). The similar order of magnitude for estimated mantle contributions to CO₂ degassing in the Eastern Rift suggests that wehrlites are well suited to monitor the present and past passage of CO₂ through the shallow lithosphere, which ultimately degassed to the atmosphere. All CO₂ mass estimates are minima because open-system processes (e.g., dunitisation) cannot be quantified using our method. Further, a proportion of clinopyroxene in lherzolites, which were attributed to “other peridotites”, may have resulted from wehrlitisation instead. It is also possible that the affected lithospheric depth interval is >10 km (e.g., in the North Atlantic Craton in Greenland it is 20 km; Supplementary Information). Moreover, the width of the affected lithosphere adjacent to rifts may be broader than assumed here.

Locality	Agent a	Area/duration b	ΔDi c	Enstatite d	Dolomite e	Perid CO2 (1σ) f	CO2 flux (1σ) g
Unit		1000 km2/Myr	%	kg	kg	kg	Mt yr−1
Eastern Rift	2	600/40	9.3	20	9.3	2.4 (1.5)	6.5 (4.1)
Aldan Shield	3	200/10	7.1	13	6.1	0.6 (0.4)	4.1 (2.4)
TLFB type 1	1		8.4	16	7.1	0.3 (0.3)	16 (13)
TLFB type 2	2		12	22	10.2	0.3 (0.2)	13 (11)
TLFB type 3	3		9.5	18	8.0	0.8 (0.7)	41 (34)
TLFB combined	1,2,3	4500/30				1.4 (1.2)	70 (58)
Middle Atlas	2	400/40	11	20	9.4	1.5 (0.4)	5.1 (1.2)
Hoggar Swell	3	785/40	3.5	6.4	2.9	0.2 (0.1)	1.5 (0.8)
WEVF	2	110/40	10	19	8.8	1.9 (1.2)	1.7 (1.1)
Pannonian basin	3	133/20	8.7	16	7.4	0.6 (0.1)	1.4 (0.1)

TLFB Tan Lu Fault Belt, WEVF West Eifel Volcanic Field
^a Type of metasomatic agent involved in wehrlitisation: 1 carbonatite (40–48 wt. % CO₂), 2 carbonated silicate melt (10–30 wt. % CO₂), 3 silica-undersaturated silicate melt (1–5 wt. % CO₂)
^b Estimated area affected by wehrlitisation and duration of metasomatism/degassing
^c Difference between median clinopyroxene abundances in wehrlite-group (reaction dunites not counted) and other peridotite xenoliths, weighted by median diopside component (Table S-3)
^d Mass of enstatite (per 100 kg of peridotite converted to wehrlite) required for conversion to mass of diopside corresponding to ΔDi
^e Mass of dolomite in the liquid (per 100 kg of peridotite converted to wehrlite) corresponding to 1/4 of the moles of enstatite as per the decarbonation reaction: 4 MgSiO₃ + CaMg(CO₃)₂ = 2 Mg₂SiO₄ + CaMgSi₂O₆ + 2 CO₂
^f Mass of CO₂ liberated from 100 kg wehrlite-bearing peridotite using proportion of wehrlite-group peridotites and its variability in Table S-3
^g Megatonnes CO₂ degassed per year for the estimated area and duration (comment b)

Modern degassing of mantle-derived CO₂-rich fluids and gases is correlated to active fault systems and extensional tectonic regimes (Brune et al., 2017

Brune, S., Williams, S.E., Muller, R.D. (2017) Potential links between continental rifting, CO₂ degassing and climate change through time. Nature Geoscience 10, 941–946.

; Tamburello et al., 2018

Tamburello, G., Pondrelli, S., Chiodini, G., Rouwet, D. (2018) Global-scale control of extensional tectonics on CO₂ earth degassing. Nature Communications 9, 4608.

). A link between wehrlitisation and extensional settings is also evident in all cases studied here. Tamburello et al. (2018)

Tamburello, G., Pondrelli, S., Chiodini, G., Rouwet, D. (2018) Global-scale control of extensional tectonics on CO₂ earth degassing. Nature Communications 9, 4608.

find that current degassing is more prevalent in central Western Europe and the western United States than in cratonic areas. This probably reflects that extension leads to lithosphere thinning, as occurred in the Wyoming Craton and in eastern North China Craton, which hosts part of the Tan Lu Fault Belt. In these settings, oxidised melts collect carbon previously stored in the SCLM (Foley and Fischer, 2017

Foley, S.F., Fischer, T.P. (2017) An essential role for continental rifts and lithosphere in the deep carbon cycle. Nature Geoscience 10, 897–902.

), followed by decarbonation as the melts encounter the solidus ledge of carbonated peridotite (Wallace and Green, 1988

Wallace, M.E., Green, D.H. (1988) An experimental determination of primary carbonatite magma composition. Nature 335, 343–346.

). In this case, not only the lithosphere associated with rifts and faults that are recognisable at the surface should be regarded as potential sites of wehrlitisation and CO₂ release, but also lithosphere affected by unsuccessful rifting and thinning, such as the North Atlantic Craton in Greenland, as well as cratonic regions recognised as partially or wholly decratonised (Aulbach, 2019

Aulbach, S. (2019) Cratonic Lithosphere Discontinuities: Dynamics of Small-Volume Melting, Meta-Cratonisation and a Possible Role for Brines. In: Yuan, H., Romanowicz, B. (Eds.) Lithospheric Discontinuities. American Geophysical Union, John Wiley & Sons, Washington DC, 177–204.

). Moreover, deep lithosphere loss causes lithospheric heating and decompression, as evidenced by microstructural and compositional evidence for garnet breakdown (Supplementary Information). This might not only exhume diamondiferous lithosphere to the shallower mantle where it is oxidised, but also causes crustal metamorphism, which is an important contributor to atmospheric CO₂ (Kerrick, 2001

Kerrick, D.M. (2001) Present and past nonanthropogenic CO₂ degassing from the solid Earth. Reviews of Geophysics 39, 565–585.

).

top

Conclusions and Outlook

Wehrlites typically constitute ∼20 % of mantle xenolith suites in extensional settings, where continental lithospheres are thinned, facilitating the generation and percolation of small-volume carbonated melts along rifts, faults or in basins. The decarbonation reaction can be applied to wehrlites to estimate the minimum amount of CO₂ that passed through the shallow (∼60–100 km) lithosphere. Based on wehrlite-bearing xenolith suites, we calculate CO₂ liberation of 24 ± 15 thousand to 2.1 ± 1.7 million Gt CO₂, with estimated CO₂ fluxes of 1.4 ± 0.1 Mt yr⁻¹ to 70 ±58 Mt yr⁻¹. Ultimate diffuse degassing of this CO₂ is expected to significantly affect climate. Importantly, wehrlitisation may occur wherever continental lithosphere is reactivated, also in lithospheric provinces where prominent rifts are absent and carbonated melts have not been emplaced at the surface. One such example is the basaltic volcanic province of southeastern Australia, where not only the link between wehrlitisation and carbonatite was first established, but also gas fields rich in mantle-derived CO₂ have been connected to the very volcanism that brought the wehrlite xenoliths to the surface (Supplementary Information).

top

Acknowledgements

This work and collaboration were stimulated by an invitation to SA and GMY to present at the Deep Carbon Observatory’s Deep Carbon 2019: Launching the Next Decade of Deep Carbon Science meeting in Washington DC (USA), and by an Alexander von Humboldt Fellowship to GMY, which we gratefully acknowledge. It was written while SA was funded through German Research Foundation fellowship AU356/11. We thank Shantanu Keshav, an anonymous referee as well as the editor, Ambre Luguet, for their very thorough reviews and incisive comments.

Editor: Ambre Luguet

top

References

top

Supplementary Information

The Supplementary Information includes:

• Oxygen Fugacity Control on Redox Freezing vs. Wehrlitisation during Interaction with Carbonated Melts
• Compositional Effects of Wehrlitisation and the Role of Lithosphere Thickness
• Selection Criteria, and Assignation to Wehrlite-group Peridotites
• Wehrlite-bearing Xenolith Suites and their Tectonic Setting
• CO₂ Liberation during Wehrlitisation: Modelling and Rationale
• Figures S-1 to S-3
• Tables S-1 to S-3
• Supplementary Information References

Download the Tables S-1 and S-2 (Excel).

Download the Supplementary Information (PDF).