|Home | About | Journals | Submit | Contact Us | Français|
Porphyry deposits, our main source of copper and of significant amounts of Mo, Re and Au, form at convergent margins in association with intermediate-felsic magmas. Although it is accepted that copper is transported and precipitated by fluids released by these magmas, the magmatic processes leading to the formation of economic deposits remain elusive. Here we perform Monte Carlo petrological and geochemical modelling to quantitatively link crustal magmatic processes and the geochemical signatures of magmas (i.e., Sr/Y) to the formation of porphyry Cu deposits of different sizes. Our analysis shows that economic deposits (particularly the largest ones) may only form in association with magma accumulated in the lower-middle crust (P>~0.5GPa) during ≥2–3Ma, and subsequently transferred to and degassed in the upper crust over periods of up to ~2.0Ma. Magma accumulation and evolution at shallower depths (<~0.4GPa) dramatically reduces the potential of magmatic systems to produce economic deposits. Our modelling also predicts the association of the largest porphyry deposits with a specific Sr/Y interval (~100±50) of the associated magmatic rocks, which is virtually identical to the range measured in giant porphyry copper deposits.
Porphyry deposits, suppliers of ~75% of the world’s copper and of significant proportions of Mo, Au, Re, are formed in subduction-related magmatic arcs1 and, to some extent, in post-subduction collisional zones2, where they are associated with magmas of intermediate to felsic compositions (mostly andesitic to dacitic) having calc-alkaline to high-K calc-alkaline affinity1. It is widely accepted that copper in porphyry-type deposits is transported and precipitated by fluids extracted from these magmas3,4,5. Therefore, the amount of copper that is ultimately precipitated in porphyry deposits is a function of the amount of magmatic fluids available and of the copper concentration in those fluids, assuming that precipitation efficiencies do not change significantly from one magmatic system to another. In turn, the mass of available fluids and their copper6 concentrations are directly related to the amount of magma and its volatile and copper content, because the transfer of copper from magma to fluid is modulated by a specific interval of fluid-melt partition coefficient values.
Geochronologic studies show that the total period of magmatic-hydrothermal activity ranges from a few tens of thousands of years up to almost 2Ma in porphyry-type deposits of different sizes7,8,9,10,11,12 (Supplementary Information 1, Table S1.1). Such temporal constraints provide broad limits to the duration of magmatic-hydrothermal activity leading to the formation of these deposits. Additionally, although not fully understood, the relationship between magmas with high Sr/Y values (~100±50) and porphyry formation has empirically proved positive in several major porphyry Cu deposits13,14,15 and has been increasingly used as a fertility indicator of magmatic systems during exploration.
All the above observations must be accounted for to develop quantitative models describing the genesis of porphyry-type deposits. Annen et al.16 have developed a thermal model to quantify the generation and evolution of intermediate to felsic magmas, i.e., those typically associated with porphyry deposits, which is valid for a broad range of physico-chemical conditions (P, T, H2O contents, magma fluxes) deemed appropriate for subduction-related magmatic systems. This thermal modeling shows that, for a given magma flux, the size of crustal magmatic systems, which ultimately controls the maximum amount of Cu that they can deliver, is controlled by the depth (pressure, P) and duration of magma accumulation (time, t).
Here, using a Monte Carlo approach, we combine the model of ref. 16 with models of H2O solubility in silicate melts17, petrological and geochemical modelling, as well as mass balance calculations, to quantify crustal depths and timescales of formation of magmatic systems best suited to produce economic porphyry Cu deposits of different sizes within the time intervals constrained by radiometric dating (Table S1.1 in Supplementary Information 1).
We focus on Andean-type porphyry Cu deposits associated with syn-subduction arc magmatism. This magmatism has a rather restricted range of Cu concentrations18, thus, simple mass balance considerations set a lower limit to the minimum amount of felsic-intermediate melt required to provide the Cu precipitated in a porphyry deposit (see also refs 6 and 19). This implies that the larger the amount of Cu deposited, the larger must be the volume of magma that delivered the mineralising fluids. In order to calculate volumes of intermediate to felsic melts we have used the thermal models developed in refs 16 and 20. Specifically, the petrologic model of ref. 16 is used to quantify the volumes of intermediate to felsic magmas generated through time by concomitant fractional crystallization of parent hydrous basalts and partial melting of host rocks (amphibolites at lower crustal levels and graywackes at upper crustal ones) at different crustal levels. We considered a typical long-term arc magma flux of 0.0009km3/a into magmatic systems located at depths ranging from 5 to 30km (corresponding to pressures between 0.15 and 0.9GPa: Fig. 1). The model results presented in ref. 20 are used to quantify volumes of intermediate to felsic melts generated at shallower depths (6–10km corresponding to 0.18–0.3GPa) by a higher magma flux of 0.016km3/a (Fig. 1). These two end-member situations encompass the conditions and rates of magma transfer in the arc crust reasonably well.
Following the model of ref. 16 we simulate injection of mantle-derived basaltic melt as circular sills of 50m thickness and 7.5km radius every 10ka at depths ranging between 5 and 30km over timescales of up to several Ma. The injected basaltic melt cools and crystallizes at a rate that is inversely proportional to the depth of emplacement (assuming a geothermal gradient of 20°C/km in the model). Progressive magma injection leads to a temperature increase of the surrounding rocks until a residual melt (from fractionation of the injected basalt) starts to accumulate. Eventually, the temperature of the magmatic system may reach values appropriate for partial melting of the surrounding rocks and a crustal melt starts to accumulate and mix with the residual melt (from basalt fractionation), forming a hybrid melt.
The most important results of the modelling are: 1) the mass of residual melt increases with the rate and duration of magma injection; 2) injected basaltic melts cool increasingly rapidly at shallower levels in the Earth’s crust and partial melting of host rocks becomes increasingly difficult and less efficient, because the temperature of the host rocks steadily decreases with decreasing depth according to the geothermal gradient (20°C/km). Therefore, melt productivity (the amount of hybrid melt accumulated after a certain time since the onset of the injection process with respect to the amount of total injected basaltic melt: Supplementary Information 2) decreases with decreasing depth; 3) over time, hybrid melts have compositions increasingly closer to that of the recharging basaltic magma, which is the opposite of what happens during the fractionation of a single magma batch21.
The model of ref. 20 is parameterised to evaluate melt productivity in shallow magmatic systems under short-lived (up to 200ka), high magma fluxes (melt injection rate of 50mm/a equivalent to 0.016km3/a for a disk-shaped pluton with 10km radius). The modelling approach is the same as the scenario for mantle-derived basaltic melts, with the exception that the injected magma is dacitic in composition and no partial melting of the host rocks occurs.
We calculated the H2O concentration of hybrid melts for both scenarios assuming that H2O in the parent basaltic melt varies between 2 and 4wt.% and in the host crustal rocks between 0.2 and 1wt.% (Table 1). While we considered a fluid phase consisting solely of H2O, we took into account the variation of H2O solubility in silicate melts with pressure and chemical composition17 (Supplementary Information 2).
The amount of fluid needed to transport and precipitate Cu depends on the concentration of Cu in such a fluid. To calculate the concentration of Cu in the fluid phase, we first considered a range of Cu contents of magmas typical for thick volcanic arcs17 (Supplementary Information 2), and then we performed Monte Carlo simulations for a range of fluid-melt partition coefficients for Cu between 2 and 100 (e.g., ref. 22). We assume that there is no recycling of pre-existing sulphide cumulates, or, if there is through partial melting and incorporation into the hybrid melt, this is computed in the Cu budget of arc magmas with normal Cu concentrations18. The Cu concentrations in the exsolvable fluid are on the order of a few hundreds of ppm (up to 700ppm). This compares well with the Cu concentrations measured in intermediate density fluid inclusions from porphyry systems, also taking into account the high diffusivity of Cu in quartz, which might be responsible for a post-entrapment enrichment of measured Cu in intermediate density S-bearing fluid inclusions23.
We have used an empirical relationship between melt productivity and SiO2 to determine the SiO2 content of the hybrid melt (see Supplementary Information 2 for further details). To calculate the Sr/Y ratios of the hybrid melts we have used available experimental petrology data24,25 on hydrous basaltic parental melts undergoing fractional crystallizion at pressures ranging from <0.1 to 1.2GPa. The mineralogy as well as the proportions of minerals, and thus the bulk partition coefficients of Sr and Y, change both with pressure and residual melt fractions. For the fractionating minerals of the above experiments (olivine, spinel, plagioclase, orthopyroxene, clinopyroxene, amphibole, and garnet) we pooled all the available mineral/melt partition coefficients for Sr and Y (GERM database: https://earthref.org/KDD/) subdividing them according to melt composition (equivalent to residual melt fractions in a process of fractional crystallization). The median values of the partition coefficients for both elements and for every mineral display systematic changes with the composition (or residual fraction) of the melt (Supplementary Information 2 and Dataset 1). We fit the partition coefficients as a function of residual melt composition and model their contribution to the evolution of Sr/Y in the residual melt as a function of pressure and amount of melt (Supplementary Information 2 and Dataset 1).
In the Monte Carlo simulations initial H2O and Cu content of the melt, fluid-melt partition coefficients for Cu, duration of sill injection at depth (t), and depth of magma accumulation (P) were contemporaneously and randomly varied within specified ranges for the two selected magma fluxes (Table 1; Supplementary Information 3, Dataset 2). We performed more than 100,000 simulations.
High magma fluxes (0.016km3/a) at shallow crustal depths (6–10km) may result in the assembly of large, transient (≤~200ka), intermediate-felsic magmatic reservoirs20,26. Our results show that, under these conditions, magmas are H2O-saturated throughout the period of magma accumulation with more than 30vol.% of excess fluid in >90% of the simulations, which is sufficient for a continuous extraction of fluids27, potentially Cu-bearing (Fig. 2a). Assuming that these extracted fluids generate mineralization28, the median value of Cu accumulation after the maximum lifetime (~200ka) permissible for such a sustained flux20,26 would be <9Mt (<4.5Mt Cu for 50% efficiency: Fig. 2b). This could be a viable mechanism to form at best deposits with <~5Mt Cu (50% efficiency), but not larger ones (Fig. 2c). Interestingly, the Cu exsolution rates and Sr/Y values calculated for these short-lived magmatic systems are comparable to Cu emission rates and Sr/Y compositions of persistently degassing volcanoes (Etna29, Masaya30, White Island31, Stromboli32; Fig. 2d).
Mafic magma injection under average arc magma flux (0.0009km3/a) at shallow crustal levels (P<~0.4GPa) prevents the accumulation of significant amounts of hybrid melt and dissolved Cu-bearing fluids (Cu<1Mt even after 3Ma: Fig. 3a) because of the large amount of heat lost to the relatively cool host rock at shallow crustal levels16. The rate of potential Cu-release by these magmatic systems is significantly lower than the rates at which Cu is precipitated in porphyry systems (all Cu endowments of porphyry deposits are much higher for a given time than our model results in Fig. 3b). This indicates that shallow level accumulation and fractionation of mafic magmas under average arc fluxes is not a viable mechanism to generate giant porphyry deposits in the timeframes measured in natural deposits.
Basaltic melt injection at pressures >0.4GPa (>~13km) and with an average arc magma flux (0.0009km3/a) results in a higher hybrid melt productivity because of the higher temperature of the surrounding rocks16 (see km3 of hybrid melt in Fig. 3a). Additionally, the H2O content of the hybrid melt is higher (Fig. 3a) because of the strong dependence of fluid solubility on pressure17. Therefore, basaltic melt injection at pressures >0.4GPa over times >3Ma, leads to the accumulation of hybrid melt containing sufficient H2O to deliver up to 10Mt Cu (Fig. 3a). At accumulation pressures >0.5GPa and accumulation times >3.5Ma, the hybrid melt contains enough H2O to deliver >30 and up to 240Mt Cu (Fig. 3a).
Our simulations indicate that magmatic systems capable of forming porphyry deposits with >10Mt Cu must develop at mid- to lower crustal depths (>~17km, P>0.5GPa) and require accumulation times >2.5Ma with an average arc magma flux (Fig. 3a,c,d). The modelling results are in agreement with the evidence that large porphyry systems develop at the end of magmatic cycles lasting several Ma33,34,35,36 (2–5Ma; Supplementary Information 1) with average magma fluxes37.
Our results suggest that magmas containing the largest amounts of H2O and Cu (>30Mt) are systematically water-undersaturated because they formed at high pressures (>0.5GPa; Fig. 4a). Therefore, fluid- and Cu-bearing magmas accumulated in the lower to middle crust may rise towards shallower levels carrying the bulk of their water and Cu, before starting to release them upon reaching fluid saturation pressures. Calculations show that if these magmas rose adiabatically, the majority of the simulated fertile magmas would become H2O-saturated at pressures between 0.25 and 0.55GPa (Fig. 4b). This pressure interval corresponds to depths of ~8–18km, which are in the same range of the main magmatic reservoirs inferred for several porphyry-type deposits from geophysical, geochemical and mineralogical data1,6,19,38.
Thus, the formation of large accumulations of hybrid melts at mid- to lower crustal levels is an essential step in the formation of the porphyry system which sets the upper size limit of the potential deposit. However, it is the next step (the transfer of this large accumulation of hybrid melt to shallower levels from where fluids can be exsolved) that determines the formation and the ultimate size of the deposit (see also ref. 28). This is suggested by data on Cu endowment and overall duration of the ore deposition period for different deposits (Supplementary Information 1, Table S1.1). Here, the overall duration of the ore deposition period was bracketed using only Re-Os dating of molybdenite from different hydrothermal pulses and/or U-Pb zircon dating of syn- and late/post-mineral porphyries (Table S1.1). Although the measurements of both Cu endowment and ore deposition period are subject to large uncertainties related to methodological approaches, uncertainties in cross-calibration of dating techniques and sampling bias, a broadly positive correlation exists between Cu endowment and the overall duration of the ore deposition period (Fig. 2c). Therefore, it is reasonable to consider that Cu in most porphyry deposits is precipitated at a long-term average rate of a few tens (~40) of tons of Cu per year. This does not mean that Cu is continuously precipitated at this rate. Rather, as pointed out by several geochronological studies9,10,39, this suggests that larger deposits are formed by a higher number of short-lived (few tens of ka) pulses distributed over a longer timescale than in smaller deposits, with the long-term average rate of Cu deposition being similar in both cases. The overall longer lifetime of magmatic-hydrothermal activity in the larger deposits could be due either to a larger availability of hybrid melt in the deep reservoir of the larger deposits or to earlier tectonic interruption of the hybrid melt transfer from depth for the smaller deposits. The latter seem to be characterized by durations of precursor magmatic activity as long as those in the largest deposits (Supplementary Information 1). Hence, it is possible that tectonics play an important role in modulating the transfer of magma from the deep to the shallow reservoir.
Our modelling results quantitatively account for the association between supergiant porphyry Cu deposits (>10Mt of Cu at 50% efficiency) and magmas with Sr/Y values between ~50 and ~15012 (Figs 5 and and6).6). In fact, in our model this range of natural Sr/Y values associated with major porphyry copper deposits corresponds to timescales (t) and crustal levels (P) of magma accumulation that are the most favourable for the development of the largest and H2O-richest magma volumes (Figs 3a and 5a,b). The availability of such a large amount of H2O-rich magma (with its corresponding “diagnostic” Sr/Y values) allows a long-lived transfer to shallower levels from where Cu can be released to form mineralization.
Our model suggests, in agreement with natural porphyry data, that for both higher (>150) and lower (<50) Sr/Y values, the probability to form deposits with >10Mt Cu is extremely low (Fig. 6), because total amounts of magmas and associated H2O formed at the corresponding P-t conditions are not sufficiently large (Figs 3a and 5a,b). Such limited H2O and magma quantities are not sufficient to sustain the longer-lived ore deposition period required to form supergiant deposits (Fig. 5c). In contrast, for magmatic systems with Sr/Y between 50 and 150 the probabilities to deliver 30–50 and 70–90Mt Cu (50% efficiency) are ~9 and 4% of the simulations, respectively (Figs 5c and and6).6). The probability to form a deposit with >90Mt Cu (El Teniente-type) drops to about 0.2% (Figs 5c and and6).6). Additionally, the largest number of simulations (66.5%) results in the production of porphyry deposits with <10Mt Cu, which is also observed in natural deposits (Fig. 6). The remarkable fit between model predictions and natural porphyry data (Figs 5c and and6)6) suggests that our model (based on measurable parameters like the overall duration of ore deposition period and average Sr/Y values of associated magmatic rocks; Fig. 5c) could be used for preliminary evaluation of mineral resources during early stages of mineral exploration.
Our results suggest that the formation of arc magmatic systems associated with porphyry Cu deposits occurs in two steps: 1) long-lasting (>~2.5Ma) injection of hydrous basalts in the mid- to lower crust (>~17km) leading to the formation of large amounts (>800km3) of andesitic magma (SiO2=57–64wt.%) with a specific interval of Sr/Y ratios (50–150) and high H2O concentrations (85% of the simulations of the most fertile magmas, with exsolvable Cu>30Mt, are between 5.5–13wt.% H2O; Supplementary Information 4, Figure S4.1) under an average magmatic arc flux (e.g., 0.0009km3/a); 2) subsequent transfer of this magma to mid-/upper crustal levels (~8–18km), from where magma may provide copper-bearing fluids, during recurring episodes of mineralization lasting up to ~2Ma. Whereas the first step (formation of large accumulation of hybrid melts at mid- to lower crustal levels) sets the upper size limit of the potential deposit, it is the next step (duration of the transfer of this large accumulation of hybrid melt to shallower levels from where fluids can be exsolved) that seems to determine the ultimate size of the deposit. The occurrence of these two steps also provides an explanation for the observed transition of tectonic regimes associated with the formation of giant porphyry systems in magmatic arcs, sometimes accompanied by a magmatic lull, from a mainly compressional stage (favouring multi-Ma accumulation of magma at depth) to a near neutral stress or slightly extensional stage38,40 (promoting the migration of magma towards shallower depths and fluid exsolution within the timescales of porphyry deposit formation). The termination of the magma transfer during this second step could be due either to exhaustion of the mid-crustal magma reservoir or, more likely, to external tectonic processes. Magmatic systems that do not grow enough in size at depth cannot sustain the long-lived magmatic-hydrothermal activity that is recorded by the largest porphyry systems.
Overall, while not discounting the potential role of specialized magmas18,41, our model suggests that typical arc processes and melt compositions are able to explain the wide range of Cu endowments observed in porphyry deposits.
Our model uses a Monte Carlo approach to simulate (>100000 simulations) the amounts of hybrid melt produced in the crust (melt productivity=amount of hybrid melt accumulated/amount of total intruded basaltic melt), their water contents (both in solution and exsolvable), the Cu contents in the exsolvable water and the SiO2 as well as Sr/Y compositions of the hybrid melts produced within the crust.
We quantify melt productivity at different crustal depths under (i) an average arc magma flux [5mm/a of basaltic melt injection rate through a circular section of 7500m of radius, equivalent to 0.0009km3/a] and (ii) an episodically high magma flux [50mm/a injection rate of dacitic melt through a circular section with 10000m radius, equivalent to 0.016km3/a], using the models of refs 16 and 20.
For the average arc magma flux (0.0009km3/a) we parameterized melt productivity for a time interval between 0 and 5Ma and for pressures between 0.15 and 0.9GPa (corresponding to crustal depths of ~5 to ~30km)16.
We allowed pressure (P) and time (t) of maturation of the magmatic systems to vary randomly within the above mentioned fixed limits (Table 1) to obtain, for any random value of P and t, the corresponding value of melt productivity using a Monte Carlo method.
We have parameterized the curves of melt productivity from ref. 16 for both residual (Mresidual) and crustal melt (Mcrustal) fractions (Supplementary Information 2, Figure S2.1), which are expressed as polynomial functions of pressure (P) of the type
where M is the residual or crustal melt fraction, P is the pressure at which injection and accumulation of residual/crustal melt is occurring and x, y and z are variables that depend on the incubation time through best fit polynomial equations of the type
where OT is the time since the onset of the injection of basaltic magma and a’/”/”’, b’/”/”’, c’/”/”’, d’/”/”’, e’/”/”’, f’/”/”’, g’/”/”’ are constant values different for each one of the x, y and z variables.
For the high magma flux (0.016km3/a) scenario we parameterized melt productivity for a time interval up to 0.2Ma (a maximum estimate for a continuously high magma flux20) and for depths of magma emplacement shifting steadily through time from 6 to 10km due to continuous injection of the sills underneath the previous ones20 (Supplementary Information 2, Figure S2.2). Volume of melt is calculated from volume of mobile magma assuming that the melt fraction in the mobile magma ranges randomly between 60 and 80% of the magma volume20.
Melt productivity as determined above was coupled with H2O concentrations in the hybrid melt assuming a geologically sound random range of initial H2O contents in the mantle-derived basalt (2–4wt.%) and in the amphibolitic (lower) to greywacke (upper) crust (0.5–1.0wt.%) and assuming a completely incompatible behaviour of H2O during the hybrid melt accumulation process. We then used VolatileCalc17 to calculate water solubility in melts according to pressure of accumulation and hybrid melt composition (Supplementary Information 2, Figure S2.3). This way we could determine the H2O contents of the melts and the degree of H2O over- or under-saturation in the hybrid melts produced at different crustal levels (P) and after different durations of injection. This allowed us to determine how much H2O was associated with any specific hybrid melt produced after any injection time, at any crustal depth and under different magma fluxes. At any specified pressure, H2O solubility is linked to melt fraction (M) by a best-fit polynomial equation (Supplementary Information 2, Figure S2.4) of the type
where h’/”/”’, k’/”/”’, i’/”/”’ are constant values specific to each one of the r, s, t variables. Combining these equations allows us to reproduce for any random P and M the solubility of H2O to be used in the Monte Carlo simulations.
The amount of fluid (and consequently of melt) needed to transport and precipitate Cu as observed in natural deposits is dependent on the concentration of Cu in such a fluid. The latter depends on the Cu concentration in the melt and on the value of the fluid-melt partition coefficient, which determines how much Cu goes into the fluid once it separates from the melt. For Cu concentrations in the hybrid melt we use the composition-(SiO2−) dependent Cu concentrations of continental arc magmas of ref. 18. The SiO2-Cu relationship is best fitted by a second order equation (y=ax2+bx+c, where y=Cu (ppm) and x=SiO2) that expresses the covariation between median Cu and SiO2 values from thick arc magmas (>30km) (ref. 18) (Supplementary Information 2, Figure S2.6). Since there is some scatter in the SiO2-Cu relationship we consider all possible values within the upper and lower boundaries of this scatter (also defined by second order polynomial equations in Figure S2.6) and implement them in the Monte Carlo modelling. For Cu KDs we use a conservative approach according to which we attribute a random variation of Cu fluid-melt KDs between 2 and 100 for any type of hybrid melt produced.
In order to link the melt productivity of the model of ref. 16 to a chemical composition (SiO2) we use the relationship below (Supplementary Information 2, Table S2.1) between melt fraction and SiO2 based on the mid-values of SiO2 for the fields of basalt, basaltic andesite, andesite, dacite and rhyolite of the TAS diagram42 and the mid-values of the melt fraction and the corresponding composition attributed by ref. 16 (e.g., Figure 8 of ref. 16).
In a bivariate plot, the two variables above are linked through the equation
Finally we calculate Sr/Y values of melt fractions based on: (i) statistical treatment of available mineral-melt KD values of Sr and Y (GERM database: https://earthref.org/KDD/) for various minerals and their correlation with changing melt composition; (ii) experimentally-determined proportions of mineral phases during crystallization of hydrous basalts at various pressures; (iii) implementation of the changing Sr and Y KDs with melt composition and changing mineral phases with pressure to obtain Sr/Y values of hybrid melts as a continuous function of pressure and amounts of hybrid melt (see Supplementary Information 2 and Supplementary Dataset 1 for more details).
The overall duration of the ore deposition period in our simulations was obtained considering the average Cu flux rate in natural porphyry Cu deposits which is provided by the broad linear correlation (Supplementary Information 2, Figure S2.7) between Cu endowment and overall duration of the ore deposition periods for selected porphyry Cu deposits (i.e., those with the best constrained radiometric dating available: see Table S1.1 in Supplementary Information 1). Although measurements of both these parameters may be subject to large uncertainties related to methodological approaches, uncertainties in cross-calibration of dating techniques and sampling bias, the positive correlation between them (Overall duration of ore deposition=0.0202*Cu (Mt), R2=0.74, p<0.00001) provides a reasonable first order approximation of the average rate at which Cu is precipitated in porphyry Cu deposits of variable sizes (~40tons of Cu per year). This rate is used to quantify the time that is needed to flux the Cu contained in the exsolvable fluid associated with the variable amounts of hybrid melts produced by our simulations, using the equation
assuming that 50% of the exsolved Cu is actually deposited.
The molar volume of H2O at different pressures (0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8GPa) and at a temperature of 950°C (typical for andesitic melts) was calculated using the online fugacity calculator at “ https://www.esci.umn.edu/people/researchers/withe012/fugacity.htm” and parameterizing the variation of the molar volume of H2O with pressure by the following best fit equation
where P=pressure (in GPa). Using temperatures of ±100°C has a minimal effect on the molar volume of H2O.
How to cite this article: Chiaradia, M. and Caricchi, L. Stochastic modelling of deep magmatic controls on porphyry copper deposit endowment. Sci. Rep. 7, 44523; doi: 10.1038/srep44523 (2017).
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
We thank Celeste Mercer (USGS), Simon Tapster (Leicester University) and an anonymous reviewer for their constructive comments that helped to improve significantly our work. MC was financially supported by the Swiss National Science Foundation (Project no. 200020_137663). LC received funding for this project from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No. 677493).
The authors declare no competing financial interests.
Author Contributions M.C. designed the study. M.C. and L.C. developed the methods and wrote the manuscript.