Enhanced metamorphic CO2 release on the Proterozoic Earth

Authors: E. M. Stewart, Donald E. Penman

Abstract

Geologic sources of atmospheric CO2 were critical to maintaining Earth’s habitability during the early evolution and expansion of life. It is often assumed that higher rates of volcanic outgassing provided the greenhouse gases necessary to warm the planet when solar luminosity was low; however, we demonstrate that CO2 release due to metamorphic processes may have been substantially elevated from one to two billion years ago. Carbon cycle models suggest that enhanced metamorphism alone (or in combination with differences in marine weathering processes) can explain observations of a warm planet in this mid-Proterozoic Era.

Metamorphic CO2 degassing is a significant flux in the geologic carbon cycle with the potential to trigger global environmental change and drive the long-term evolution of the deep Earth, yet characterization of this important carbon flux prior to the Cambrian period is almost nonexistent. Nevertheless, there is reason to suspect that metamorphic CO2 devolatilization was substantially higher on the earlier Earth. For one, changes in the nature of sedimentary carbonate burial have altered the distribution of carbonate in sediments which are available for metamorphism (e.g., refs. 1 and 2). In addition, hotter crustal thermal regimes prior to the Cambrian period (e.g., refs. 3 and 4) could also enhance metamorphic CO2 loss.

On the modern Earth, metamorphic CO2 release is on the order of 1 to 5 Tmol CO2~ y^−1^, accounting for a significant fraction of solid Earth degassing. Though there are large uncertainties (from 5 to 80%), median estimates rank metamorphic degassing as 40% of total solid Earth CO2 release today (5–14). A larger Precambrian metamorphic carbon flux would have profound implications for the whole Earth system—contributing to higher atmospheric CO2~ concentrations and warmer climates—unless dampened or balanced by adjustment of other carbon fluxes. Thus, it is crucial to understand the factors which controlled metamorphic decarbonation on the early Earth.

Metamorphic decarbonation reactions usually occur via the reaction of silicate and carbonate minerals as often represented by the archetypical [1]CaCO3(calcite)+SiO2(quartz)→CaSiO3(wollastonite)+CO2(fluid),

where the minerals calcite and quartz react to form a new mineral, wollastonite, and release CO2 (15, 16). While this specific reaction is relatively rare on Earth, it represents two important factors shared by almost all metamorphic decarbonation 1) both carbonate minerals (e.g., calcite) and silicate minerals (e.g., quartz) must be present for the reaction to go forward and 2) the reaction is driven forward at higher temperatures, thus hotter rocks can release more CO2.

In fact, the ratio of carbonate to silicate material in a rock is a strong predictor of the degree of carbon loss it will experience. A rock with a high ratio of carbonate to silicate will exhaust all silicate reactant with much of the initial carbon still trapped in solid form. Conversely, a rock with a low carbonate to silicate ratio will be limited by carbonate availability and has the potential to degas a much higher percentage of its CO2 (see ref. 17).

Dolomitic rocks can also release more CO2 than metamorphism of analogous calcite-bearing sediments. This is due to the higher number of product silicate minerals which may form. Reaction 2 shows the reaction of dolomite and quartz to release CO2 and create calcite and forsterite as additional products.[2]2CaMg(CO3)2 dolomite+SiO2 quartz→2CaCO3 calcite+Mg2SiO4 forsterite+2CO2 fluid.

In addition to forsterite, reaction of dolomite and quartz may form talc, tremolite, diopside, or clinohumite (16). Calcite and quartz, on the other hand, can only react to form the mineral wollastonite at crustal P-T conditions. Many Mg-bearing silicates are stable at lower temperatures than wollastonite (e.g., talc at ~300 °C versus wollastonite at 700 °C; see ref. 16). At high temperatures, stable Mg-silicates exhibit lower cation-to-silicon ratios than wollastonite (e.g., 1 in forsterite versus 1 in wollastonite). Thus, dolomitic rocks undergo more efficient decarbonation than calcitic rocks if they are either a) silica limited and/or b) experienced peak temperatures below ~700 °C.

Given the above, this work focuses on how two broad variables have changed over billions of first, the nature of protolith sediments available to be metamorphosed, and second, the pressure-temperature (P-T) conditions of metamorphism itself. Here, we show that these intertwined variables may combine to dramatically increase metamorphic CO2 degassing and atmospheric CO2 concentrations on the earlier Earth, making metamorphism a critical component of our planet’s early habitability.

Exploring Metamorphic Decarbonation Efficiency.

Thermodynamic modeling of rock metamorphism can be used to explore the influence that protolith bulk composition and P-T conditions exert on decarbonation efficiency. As an example, the degree of decarbonation is calculated for a range of bulk one endmember is a representative siliceous sediment with no carbonate material (Marianas clay composition from ref. 18) and the other endmember is a pure calcite limestone. In Fig. 1A, carbon loss is calculated at an arbitrary moderate temperature and pressure of 600 °C and 0.55 GPa. At these conditions, rocks starting with <12 weight percent calcite are able to undergo 100% decarbonation, while more calcite-rich rocks retain more of their carbon in solid form. In Fig. 1B, carbon loss for a 90% clay–10% calcite protolith is calculated along two P–T a high T/P path from 310 °C and 0.2 GPa up to 600 °C and 0.55 GPa typical of Proterozoic metamorphism and a lower T/P path from 310 °C and 0.4 GPa up to 600 °C and 0.84 GPa, representing Phanerozoic Barrovian conditions (e.g., refs. 19 and 20). As expected, the rock metamorphosed along the higher T/P path releases more CO2~ at each temperature step. Finally, we consider decarbonation of calcite versus dolomite in Fig. 1C. Here, a carbonate + quartz system degasses CO2 at lower temperatures and achieves a higher total degree of decarbonation if the carbonate mineral is dolomite. Taken together, these data suggest that higher T/P metamorphism of a predominately dolomitic and mixed carbonate-silicate sedimentary sequence would release more CO2 than low T/P metamorphism of a less-mixed sequence containing an identical mass of carbon.

Global Changes in Sedimentation and Metamorphism.

Sedimentary carbonate is dominantly (by mass) deposited in relatively pure limestones and dolostones with little to no silicate material mixed in (e.g., ref. 21). As compared to the systems in Eqs. 1 and 2, these pure carbonate rocks do not have the necessary chemical components to release CO2 via decarbonation reactions at normal metamorphic conditions. While aqueous transport of, e.g., silica, into pure carbonates during metamorphism may drive decarbonation, this metasomatism is likely to affect only a small fraction of the total rock volume [~1 to 20 %; (22)].

However, the evolution of marine calcifiers around the Cambrian period had a profound effect on the nature and location of carbonate deposition (e.g., refs. 2 and 23), as documented by a rich literature on Precambrian carbonate sedimentation styles and structures (e.g., refs. 1 and 24–28). In particular, the Precambrian is characterized by a higher abundance of abiotic carbonate precipitate textures (e.g., ref. 1) and stromatolites (e.g., ref. 29). With respect to carbonate-silicate mixing, Grotzinger & James (ref. 1) note that Mesoproterozoic carbonate platforms are “muddier,” with more silt and clay mixed in. The ratio of dolomite to calcite in the sedimentary record has also been thoroughly explored, with dramatically greater dolomite abundance recognized in the Archean and Proterozoic Eons, (30–33).

A large new dataset offers the opportunity to explore these observations more quantitatively. Cantine et al. (ref. 34) compiled data on carbonate-bearing sediments from 3.8 billion to 490 Mya in a database that includes observation of more than 16,000 Precambrian rocks, spanning more than 45,000 m of measured stratigraphy. Most relevant to decarbonation, each entry is flagged if it contains dolomite, “significant carbonate component in a siliciclastic rock,” or “significant siliciclastic component in a carbonate rock.” Fig. 2 A and B show the distribution of rock types for each geologic era calculated from the dataset of ref. 34. A gradual increase in mixed carbonate-silicate rocks peaks in the Mesoproterozoic era, where more than 25% of the recorded section consists of these mixed rock types. Importantly, this database approach is consistent with the more qualitative observations previously made by ref. 1. In addition, the database reproduces observations of dolomitic carbonates dominating the rock record prior to the Cambrian. Thus, while there is certainly some preservation bias in these data—ref. 34 provides a more detailed discussion of possible biases and caveats in the database—we consider these two very substantial differences to represent a real change in carbonate sedimentation styles over time, consistent with the field observations made by previous authors (e.g., refs. 1 and 24).

With respect to temperature and pressure in the crust, existing analysis has already demonstrated higher metamorphic T/P conditions prior to the Cambrian (e.g., refs. 3, 4, and 35). This general cooling trend started about two billion years ago and is accompanied by spreading of P-T conditions into a bimodal distribution of hotter and colder environments. Brown & Johnson (ref. 35) report a database of 546 localities with metamorphic pressure and temperature spanning the last ~3.7 billion years of Earth history. Additional analysis by ref. 4 highlights the bifurcation trend and presents implications for the initiation of plate tectonics. These data are shown in Fig. 2C.

Estimating the Precambrian Metamorphic Carbon Flux.

In order to estimate metamorphic carbon fluxes on the Precambrian Earth, thermodynamic modeling of metamorphism is applied to representative sedimentary protoliths and P–T conditions derived from ref. 34 and ref. 35, respectively. Due to the age range available in these datasets, we consider time intervals between 490 and 3500 Ma. A 10,000—iteration Monte Carlo approach is used to account for uncertainty in a variety of terms including protolith characteristics, P-T-Xco2 conditions, and tectonic geometry (i.e., crustal thickness and area of metamorphism). We generate random time-bins for each iteration to avoid bias associated with binning; then, for each time interval, we calculate an average metamorphic flux as described below (Materials and Methods and SI Appendix). We calculate total solid-Earth degassing rates for each of our time bins by adding the calculated metamorphic flux to an assumed constant volcanic flux of 4 Tmol CO2 y^−1^. In addition, the carbon isotope composition of the solid Earth degassing flux can be estimated as a weighted average of a volcanic mantle source (δ^13^CVPDB = −5) and a heavy fractionated metamorphic source. We recognize the simplifications inherent in these calculations; however, given the consistent approach, the results can be expected to show robust patterns and relative changes between time intervals even if the absolute values carry large uncertainties.

Model inputs and results of thermodynamic modeling are shown in Fig. 3. Here and elsewhere we focus on the average model output, while 25th and 75th percentile contours are also shown. Rocks around the Archean–Proterozoic boundary at 2500 Ma undergo the lowest degree of metamorphic carbon loss, releasing an average of 2.5 Tmol CO2~ per year. Mean estimated metamorphic decarbonation reaches a maximum of 5.5 Tmol CO2~ per year at 1400 Ma before falling back to 3.5 Tmol y^−1^ in the Cambrian. When added to a constant volcanic flux, this corresponds to total solid Earth outgassing increasing from 6.5 Tmol CO2 per year on the earliest Earth to 9.5 Tmol CO2 per year in the Proterozoic (i.e., an increase of ~45%). The δ^13^C of the flux similarly rises from −2.4‰ to −1‰.

Modeling the Response of Atmospheric pCO2.

We take two approaches to estimate the potential effects of an elevated Proterozoic metamorphic degassing flux on the exogenic carbon cycle (particularly atmospheric pCO2 levels). The first approach considers the effects of elevated degassing on the carbon mass balance of the exogenic carbon solid-earth degassing is balanced by the chemical weathering of silicate minerals and subsequent marine carbonate burial (36–38), and the steady-state point is set by the pCO2 required to generate enough silicate weathering to balance a given degassing flux (39). The second technique uses a numerical model of the Precambrian carbon cycle [PreCOSCIOUS, (40) modified to include CO2 recycling by enhanced reverse weathering. Reverse weathering can be represented by the reaction[3]3Ca2+(aq)+2H4SiO4(silicic acid)+6HCO3−(aq)→Ca3Si2O5(OH)4(clay)+CO2(fluid)+5H2O(fluid),

where alkalinity is consumed by clay formation and CO2 is released. Isson & Planavsky (ref. 41) suggest that this is a potentially important feature of the Precambrian carbon cycle due to elevated seawater-dissolved silica concentrations in the absence of biosilicifying organisms (e.g., ref. 42), but the extent to which reverse weathering was elevated during the Precambrian is highly uncertain, and may not have been significantly higher than modern rates (43). We therefore present modeling results with and without enhanced reverse weathering as two possible endmember scenarios.

Results

Phase equilibria modeling predicts a gradual rise in the metamorphic carbon flux after ~2500 Ma. Given comparable pressure–temperature conditions in the Archean and Proterozoic, this apparent rise in the metamorphic flux can primarily be attributed to changing sedimentary protolith compositions. That is, sediments deposited after ~2500 Ma were more susceptible to carbon loss when heated. The fall in metamorphic degassing after ~1400 Ma is a result of both lower T/P metamorphic conditions and dilution of dolomitic and mixed carbonate-silicate sediments with the deposition of increasingly pure limestone. We can take the Cambrian interval (541 Ma to our minimum age of 490 Ma) as representing relatively “modern” conditions. By comparison, we estimate an average total volcanic + metamorphic degassing flux near ~130% of the modern at its Mesoproterozoic peak.

How might a higher Precambrian metamorphic degassing flux affect Earth’s carbon cycle and atmospheric pCO2 levels? We first consider that question in the context of the mass balance requirement for the exogenic carbon On long timescales, the sources of C to the atmosphere (including solid earth degassing) must balance the sinks (sedimentary carbon burial as carbonate minerals and organic carbon; ref. 44). The maintenance of this balance on geologic timescales is credited to a dynamic silicate weathering flux, which removes CO2 from the atmosphere and encourages carbonate mineral precipitation and burial in sediments. The rate of silicate mineral weathering is thought to accelerate under warm (high pCO2) conditions, thus removing more CO2 from the atmosphere and returning it to the solid earth as sedimentary carbonate. Under elevated solid Earth degassing, a new exogenic carbon cycle mass balance will be set at the atmospheric pCO2 required to increase silicate weathering to balance the new degassing flux, which, in turn, depends on the strength of the silicate weathering feedback (e.g., ref. 39). If a mathematical form of the pCO2-silicate weathering flux relationship is assumed, then the pCO2 required to balance any degassing flux can be solved analytically. A power law formulation is typically used to represent the sensitivity of silicate weathering (FSi) to changing pCO2 (e.g., refs. 39, 45, 46, and 47):[4]FSi=Fvc0*(RCO2)nSi,

where Fvc0 represents the initial degassing flux, RCO2 represents the ratio of atmospheric CO2 to a reference, initial pCO2 (for example, preindustrial pCO2 of 280 ppm), and nSi represents the strength of the silicate weathering feedback. This relationship can be rearranged to determine the RCO2 needed to balance a higher degassing flux (i.e., FSi = a new Fvc):[5]RCO2=(Fvc/Fvc0)1/nSi.

This relationship describes how equilibrium atmospheric pCO2 responds to changing background solid Earth degassing rate, but it is very sensitive to the silicate weathering feedback strength (nSi). Because the strength of the silicate weathering feedback is uncertain and almost certainly varies significantly over the billions of years of geologic time spanned by our metamorphic degassing reconstructions (e.g., refs. 39 and 47), we consider a wide range of possible nSi values (from 0.2 to 0.6) in calculations of equilibrium atmospheric pCO2 resulting from changing solid-earth degassing (Fig. 4).

The Fig. 4A illustrates the relationship between changing solid Earth degassing rate and equilibrium pCO2 (Eq. 5) across a range of silicate weathering feedback strength (nSi). Our reconstructed Mesoproterozoic degassing flux (a sustained solid Earth degassing rate 30% higher than modern) would result in an equilibrium pCO2 between 1.5 and 3.5 times higher than modern (median values), demonstrating the significant impact that enhanced metamorphic degassing would have on the Precambrian carbon cycle and climate.

The above framework omits a recently proposed (41) mechanism to support elevated pCO2 during the Precambrian: accelerated marine clay authigenesis (“reverse weathering”) under the high seawater dissolved silica concentrations inferred for Precambrian seas in the absence of biosilicifying organisms. If marine reverse weathering was indeed much more ubiquitous during the Precambrian, this would certainly affect the relationship between elevated solid-earth degassing, silicate weathering, and atmospheric pCO2. In order to investigate the effects of enhanced metamorphic CO2 release on a possible Proterozoic carbon cycle buffered by ubiquitous reverse weathering, we performed a suite of degassing rate experiments using a numerical model of the Precambrian carbon and silica cycles [PreCOSCIOUS, (40) modified to include reverse weathering (e.g., refs. 50 and 51). The sensitivity of reverse weathering rates to Precambrian seawater/pore water dissolved silica concentrations, seawater pH, temperature, and other variables is highly Isson and Planavsky (ref. 41) explored potential reverse weathering rates spanning orders of magnitude, concluding that elevated Precambrian dissolved silica concentrations accelerated reverse weathering rates enough to significantly affect the global carbon cycle and atmospheric pCO2 (a conclusion which has been questioned by ref. 43). For the present purposes, we parameterized reverse weathering as a function of seawater dissolved Si concentration, with a sensitivity scaled to give equilibrium atmospheric pCO2 levels similar to those proposed by ref. 41. From that stable equilibrium (which already features atmospheric pCO2 ~16× higher than preindustrial), the model was forced by systematically changing the total solid Earth degassing rate and then running the model for tens of millions of years until a new stable equilibrium was reached. As above, these experiments were performed under several choices of nSi (the sensitivity of the silicate weathering feedback). Because uncertainties in reverse weathering kinetics do not preclude a Precambrian carbon cycle without significantly elevated reverse weathering (43), we compare these experiments with scenarios that omit elevated reverse weathering as two possible endmembers.

The Fig. 4B shows the equilibrium atmospheric pCO2 resulting from different combinations of solid-earth degassing rate and silicate weathering feedback strength in PreCOSCIOUS simulations that include accelerated reverse weathering. pCO2 resulting from the same combinations of degassing rate and silicate weathering feedback strength without reverse weathering are also plotted for comparison. The absolute magnitude of atmospheric pCO2 levels in simulations that include accelerated reverse weathering are of course an order of magnitude higher than modern levels, but the effect of changing solid Earth degassing rate is similar. A lower nSi (weaker silicate weathering feedback) leads to more pronounced changes in atmospheric pCO2. Simulations with degassing rates 30% higher than modern (representing the increased metamorphic CO2 production we calculate for the Mesoproterozoic) feature equilibrium pCO2 levels between 1.5 and 4.1 times higher (depending on the value of nSi chosen) than those simulations run with a modern degassing rate.

The elevated rates of Proterozoic decarbonation predicted by our thermodynamic modeling would increase not just the rate of total-solid earth degassing, but its δ^13^C as well. Total solid-earth degassing is a mixture of relatively high-δ^13^C metamorphic CO2 (3.5‰) and lower-δ^13^C (−5‰) mantle-derived volcanic CO2~. An increase in the metamorphic component by 1.7× (as in our average model at 1400 Ma) would drive an increase in the δ^13^C of degassed CO2~ from −2.4‰ to −0.9‰. We use carbon-cycle flux and isotopic mass balance calculations (45, 52, 53) to calculate what effects elevated degassing CO2 δ^13^C would have on the exogenic carbon cycle and sedimentary δ^13^C records. Results of these calculations are shown in Fig. 5B.

The combined effects of elevated degassing rates and δ^13^C on sedimentary δ^13^C are quite small because they each independently drive opposing increasing the δ^13^C of solid-earth degassing serves to increase the δ^13^C of the exogenic carbon cycle, while increasing the rate of solid-earth degassing tends to decrease the δ^13^C of the exogenic carbon cycle. When both are increased simultaneously, as implied by our thermodynamic modeling of Proterozoic decarbonation, the effects nearly cancel We calculate sedimentary δ^13^C values for the entire Precambrian and Cambrian within 0.1‰ of one another.

Discussion

It has long been suggested that Earth’s atmosphere needed higher concentrations of greenhouse gasses to maintain habitability given substantially lower solar luminosity early in our history [see the “faint young sun” problem, (54, 55); reviews in refs. 56–58]. It is, therefore, common to propose that volcanic carbon degassing may have been elevated to account for higher equilibrium pCO2 (e.g., refs. 59 and 60) despite little direct evidence for enhanced volcanic CO2 release on the early Earth. Additional proposed explanations include less efficient CO2 draw down via silicate weathering (e.g., ref. 41). We would by no means suggest that these processes were constant over billions of years. However, we propose that enhanced metamorphic CO2 degassing offers an alternative and complementary explanation for high atmospheric pCO2, especially in the Mesoproterozoic.

Notably, our reconstructed interval of elevated metamorphic CO2 release corresponds to the longest recorded period in our planet’s post-Archean history devoid of global glaciation (Snowball Earth) events (e.g., refs. 61 and 62; Fig. 6). During the subsequent Neoproterozoic, the Sturtian (715 to 660 Ma, ref. 63) and Marinoan (635 Ma, ref. 61) “Snowball Earth” glaciations of the Cryogenian are the most extensive glaciations in Earth’s history. The global-scale glaciation during these events has been attributed to a runaway ice-albedo positive feedback following the crossing of a mid-latitude polar ice cap growth threshold during a long-term cooling trend (e.g., refs. 64 and 65). The fall in metamorphic CO2 degassing flux reconstructed by our thermodynamic modeling after ~1200 Ma may have contributed to this long-term cooling trend or at least cooled the background climate state of the Neoproterozoic such that secular cooling trends were more likely to cross the ice-albedo runaway threshold, triggering global glaciation.

In short, modeling suggests that metamorphic CO2 release could have dominated global carbon cycling on the Proterozoic Earth, with profound implications for the climate state, carbon isotope mass balance, and planetary habitability. Some substantial uncertainties remain—in particular, differences in the volumetric extent of metamorphism and reaction rates could modulate metamorphic CO2 degassing independent of the changing decarbonation efficiency we explore here. We note that mountain-building may also be associated with enhanced silicate weathering and corresponding draw-down of atmospheric CO2 (e.g., refs. 66 and 67). However, even on the modern Earth, mountain building appears to be a net source of CO2 (e.g., refs. 7 and 8) emitting, perhaps, an order of magnitude more CO2 than is removed (68). Thus, we consider this process a likely net source of CO2 on the earlier Earth, especially in light of evidence for enhanced metamorphic degassing.

Exactly how much enhanced degassing might have increased atmospheric pCO2 depends on the strength of the silicate weathering feedback (uncertain today, even more so for the Precambrian) and whether or not reverse weathering was supercharged to the extent that ref. 41 propose. Nevertheless, it is crucial that metamorphism be explored as a potentially major carbon source in consideration of the carbon cycle in deep time. Future work should include new field- and laboratory-based estimates of metamorphic CO2 degassing. The inclusion of carbonate metamorphism in numerical models will also help define the role it played in the success of life on the early Earth.

Materials and Methods

Thermodynamic Modeling of Metamorphic Decarbonation.

All calculations are performed in the software package Theriak-Domino (69) with the database of ref. 70 and compatible activity models with a water-bearing fluid present in excess. Details are given in SI Appendix.

For each time interval, we calculate an average sedimentary section using rock volumes from ref. 34 and we assume all carbonate rocks of equal or older age are available for metamorphism. Dolostones and limestones “with significant silicate component” are represented by adding Marianas clay to a pure carbonate rock, while significant carbonate component in a siliciclastic rock is approximated by adding either dolomite or calcite to a pure Marianas clay. We also consider metasomatic silica addition to pure carbonate rocks in a simple system of CaO-(±MgO)-SiO2-CO2-H2O. For each time period, the carbon loss resulting from metamorphism is calculated for each protolith type at a metamorphic condition consistent with the corresponding time interval from ref. 35.

Finally, the fractional mass loss of carbon is converted to a flux according to the following [6]F=∫z1z2Cco2·ρ·ΔCO2·t-1·A·dz,

where CCO2 is the initial mass fraction of CO2 for a given rock type, ρ is its density (fixed at 2,700 kg m^−3^), ΔCO2 is the fraction of carbon lost, t is the duration of metamorphism (fixed at 10 My), A is the area of metamorphism, and z is the depth of each rock type (such that Δz = thickness); z is estimated in the Monte Carlo simulation using a generated crustal thickness and volume fraction of carbonate-bearing rock throughout the crustal stack multiplied by the proportions of each carbonate-bearing subtype according to data from ref. 34 (Fig. 3D). See similar assumptions for flux calculations in refs. 5, 9, 71, and 72.

The δ^13^CVPDB of total degassed CO2 from each protolith along an open-system Rayleigh fractionation curve can be calculated according to[7]δ13Cdegassed=δ13CoFα-1+1000Fα-F·F-1-1,

where δ^13^Co is the isotopic composition of the original solid, F is the fraction of carbon remaining in the rock, and α is a temperature-dependent fractionation factor. The δ^13^Co values are taken as +0.8‰ (the average for all Precambrian carbonates from ref. 73). There is no fractionation associated with degassing in carbonated siliciclastics since they undergo 100% carbon loss (i.e., quantitative removal), but the siliceous limestones and dolostones release heavily fractionated CO2 (>4.0‰). The average metamorphic outgassing flux ranges from 1.1 to 2.5‰.

Supplementary Material