Water plays an important role in geological processes. Providing constraints on what may influence the distribution of aqueous fluids is thus crucial to understanding how water impacts Earth’s geodynamics. Here we demonstrate that ductile flow exerts a dynamic control on water-rich fluid circulation in mantle shear zones. Based on amphibole distribution and using dislocation slip-systems as a proxy for syn-tectonic water content in olivine, we highlight fluid accumulation around fine-grained layers dominated by grain-size-sensitive creep. This fluid aggregation correlates with dislocation creep-accommodated strain that localizes in water-rich layers. We also give evidence of cracking induced by fluid pressure where the highest amount of water is expected. These results emphasize long-term fluid pumping attributed to creep cavitation and associated phase nucleation during grain size reduction. Considering the ubiquitous process of grain size reduction during strain localization, our findings shed light on multiple fluid reservoirs in the crust and mantle.
Water is one of the main characteristic components of the Earth. It significantly influences geological processes such as partial melting, ore genesis and the production of earthquakes. Deformation experiments have also shown that water may weaken single crystals and mineral aggregates during ductile flow1,2,3,4, with key implications for plate tectonics. For instance, water/hydrolytic weakening of quartz in the lower crust and of olivine in the uppermost mantle may change the mode of lithospheric deformation from localized (subduction zone, narrow rift) to distributed (ridge plateau, wide rift), or vice versa5,6,7. Although the physical process of this phenomenon remains to be resolved, intra-grain water is thought to enhance the glide or climb of line defects (dislocations) during plastic flow, resulting in mechanical weakening8,9.
Conversely, rock deformation may have an impact on water distribution. While brittle faults are accepted as controlling the circulation of fluids at shallow depth (< 10 km depth), natural observations—mainly of mid-crustal rocks—suggest that ductile shear zones also drive fluids at greater depth10. In the brittle regime, the presence of faults controls the migration of fluids by providing structural conduits or through seismic pumping11. But at deeper levels and higher temperatures, these fault-related processes are unlikely owing to plastic flow. At present, two possible mechanisms have been proposed to account for water draining in ductile shear zones, including a ‘passive’ convergence due to an increase of porosity where grain size is being reduced12, or ‘dynamic’ pumping through opening of micro-cavities13. Fluid migration is indeed particularly observed where mineral aggregates dominantly deform by grain-size-sensitive (GSS) creep, commonly including a component of grain boundary sliding (GBS)13,14,15. When GBS cannot be fully compensated by diffusive or plastic processes, some micro-cavities open and close in a continuous process because of local dilatancies, giving rise to creep cavitation16,17,18. This might result in fluid pumping—including water—towards the shear zone centre. Nevertheless, only a few vestiges of syn-tectonic water draining have been reported so far, and exclusively for crustal rocks13,19, allowing doubts to persist about the validity of this latter mechanism.
In this paper, we describe systematic changes in dislocation slip-system and related olivine fabric, combined with enrichments in amphibole, that document syn-tectonic water draining in natural mantle shear zones (Ronda massif, Spain). While water accumulates around ultramylonites that deform by GSS creep, fluid aggregation is coeval with multi-scale strain partitioning between water-poor and water-rich layers during olivine dislocation creep. Providing evidence of long-term water pumping during shear strain localization in upper mantle peridotites, our findings strongly suggest that water converges through the opening of micro-cavities coupled with phase nucleation processes.
The Ronda mylonitic complexes
Located in the internal domain of the Betic Cordillera in southern Spain, the Ronda massif encompasses several kilometre-scale lenses of sub-continental ultramafic bodies (Fig. 1a). During continental extension and subsequent thrusting onto continental crust20,21, these mantle peridotites were affected by several stages of deformation21,22 including the development of ductile shear zones in the southwestern massif (Fig. 1a)23. In this region, the outcrop-scale features consist of alternating protomylonitic and mylonitic layers within so-called mylonitic complexes. These complexes range from a few centimetres to more than 10 m in thickness and extend up to several hundred meters along strike (Fig. 1b,c). They developed from a protolith of porphyroclastic, spinel-bearing harzburgite with a foliation oriented 110°/50°NE with top-to-the-SW shear sense criteria (Fig. 1b). The transition from the protolith to a mylonitic complex is progressive and characterized by a thickening of mylonitic and protomylonitic layers (see Supplementary Fig. 1).
In thin section, the mylonitic peridotites contain several ultramylonitic bands of very fine grains that always wrap around pyroxene porphyroclasts (Fig. 1d). Indeed, the protolith and protomylonites have a mean grain size >1 mm for all phases, but the grain size substantially decreases to ∼150 μm close to mylonitic layers, ∼30–50 μm in mylonitic layers (that is, in between ultramylonites) and <15 μm in ultramylonitic layers, except for a few porphyroclasts of pyroxene (see the Methods section for grain size calculation). At the centre of mylonitic complexes, the density and size of ultramylonitic bands substantially increase, and mylonitic layers contain coarse spinel grains (chromite) with many mode-I cracks sealed by olivine and amphibole (tremolite) (Fig. 2). The occurrence of cracks resulting from tensile stress and sealed by a hydrous phase strongly suggests that the cracking process was the result of high fluid pressure24. Post-tectonic serpentine is also present in the protolith and protomylonites, but is less abundant in mylonites and is quasi-absent in ultramylonites.
Through high-resolution element maps collected by electron microprobe, we further demonstrate enrichment in secondary phases and the presence of amphiboles (pargasite±tremolite) in fine-grained ultramylonites (Fig. 3a,b). Most peridotites of the shear zones (protolith, protomylonites and mylonites) are indeed composed of olivine (85%) with a small amount of orthopyroxene (13%)±clinopyroxene (1%)±spinel (1%), whereas the ultramylonites contain far more orthopyroxene (∼30%) and ∼4.5% of calcium-bearing phases. These latter include either clinopyroxene (diopside) or amphiboles with local intergrowths (Fig. 3b). The amount of amphibole generally increases with reducing grain size, but we also document higher content of amphibole in some mylonitic areas located along ultramylonitic layers (see Supplementary Fig. 2). In ultramylonites, all phases are strongly mixed with a low-porosity, with backscatter electron (BSE) images revealing only a few pores of ∼1 μm size (Fig. 3c). Very little amphibole is present in the protolith25.
Based on mineral compositions in the mylonites and ultramylonites and using available geothermobarometers (see Supplementary Tables 1–6), we appraise (1) a pressure reducing from ∼1.2 GPa to ∼0.5 GPa, and (2) a coeval decrease in temperature from >1,000 °C to below 600 °C with reducing grain size (Fig. 3d). These conditions are consistent with previous estimates that documented decompression and cooling from ∼1 GPa/1,000 °C to ∼0.5 GPa/<700 °C in the study area22,25,26. Using Perple_X calculations to construct a P–T diagram of phase equilibrium (pseudosection), we further give an estimate of the ultramylonites water content. Focusing on a very fine grain size area, we estimated the mineral modal proportions through point counting using element map. This gives 65.9% for olivine, 28.8% for orthopyroxene, 3.0% for amphiboles, 1.5% for clinopyroxene and 0.8% for spinel, which was subsequently converted into oxide percentages. Based on our P–T estimates, as well as calculated isopleths for amphibole and clinopyroxene on the pseudosection, we need ∼600 p.p.m. H2O to reach 3% amphibole (2.25% pargasite+0.75% tremolite) and <2% clinopyroxene in the ultramylonites (Fig. 3e). As 3% amphibole is close to the maximum that we found in ultramylonites, a lower water content may be expected. However, the absence of plagioclase necessarily implies a minimum of 400 p.p.m. H2O (see Supplementary Fig. 3). For further information about microstructures, phase compositions and P–T estimates, the readers may refer to Hidas et al.25.
Olivine fabric and lattice misorientations
Through electron backscatter diffraction (EBSD) analyses, we documented lattice preferred orientation (LPO) and related features for olivine. All LPOs are characterized by olivine  axis maxima at low angles to the shear direction (X) and  axes close to the pole to the shear plane (Z), which is typical of an E-type fabric (Fig. 4a)27. This feature differs from the distal protolith where an A-type fabric has been documented22,25. Moreover, in some layers, the  and  axes are more dispersed and distributed as girdles normal to the shear direction. Together with a strong  axis maximum that is close to the shear direction, these features are typical of an axial- fabric or D-type fabric28,29, giving rise to a combined type of olivine LPO between E- and D-type fabrics.
To better characterize this combined LPO, here referred to as D/E-type fabric (Fig. 4a), we used the eigenvalues of each pole figure to calculate the BA index30,31. This index indicates a degree of combination between axial- fabric (BA=1) and axial- fabric (BA=0). A point maximum LPO such as an E-type fabric will have an intermediate value, typically around 0.5. Across the Ronda shear zones, the BA index indicates pure E-type fabric at the edge of the protolith along mylonitic complexes (BA=0.48), and at the edge of the mylonitic layers adjacent to ultramylonites (BA=0.49) (Fig. 4b). Elsewhere, olivine developed a D/E-type fabric with significant components of D-type fabric, including in the protomylonites (BA=0.67), the core of the mylonites (BA=0.73) and the core of the protolith layers located between mylonitic complexes (BA=0.81) (Fig. 4a,b). We also found D/E-type fabric in the ultramylonites (BA=0.70), but this value remains uncertain because of their very weak and dispersed LPOs (Fig. 4a).
In relation to this, the strength of olivine fabrics was investigated using the texture32 (J) and misorientation33 (M) indices. These both give the degree of mineral alignment from random (J=1; M=0) to single crystal-like LPO (J=∞; M=1). Across the Ronda shear zones, olivine fabrics range from moderately weak LPOs in the protolith (J=3.87; M=0.14) to very strong LPOs at the edge of mylonitic areas adjacent to ultramylonites (J=16.80; M=0.46). They also systematically strengthen from centre to edge within each layer (Fig. 4b). Indeed, the J and M indices never exceed 5 and 0.2, respectively, in the centres of the protolith layers between mylonitic complexes, the protomylonites and the mylonites. In contrast, the fabric strength largely exceeds J=10 and M=0.3 at the edge of these layers (Fig. 4a). This distribution (using M index) correlates with the distribution of shear strain (γ) across the Ronda shear zones, deduced from the angle between the long shape axis of olivine grains and the shear plane (Fig. 4b)34. It is also important to note that despite high shear strain in ultramylonites (γ>5), the olivine LPO remains extremely weak (J=1.64; M=0.01).
EBSD maps were used to document intra- and inter-granular structural features across the shear zones. These measurements identify olivine sub-grain boundaries in mylonites, with a sub-grain size (∼20 μm) that far exceeds olivine grain size in the ultramylonites (Fig. 5a). Sub-grain boundaries are also prevalent in both the protolith and protomylonitic bands, in contrast to ultramylonitic layers where only a few sub-grains are observed. The parameters used for the definition of grain boundaries and sub-grain boundaries are available in the methods section.
Of further interest, EBSD maps combined with inverse pole figures (IPF) emphasize different LPOs for olivine as a function of the distance from ultramylonite layers. In Fig. 5a, we display the distribution of olivine crystallographic axes with respect to the Y direction, that is, the deformation axis normal to the shear direction and parallel to the shear plane. As sketched in Fig. 5b, this map highlights different areas that include: (1) ultramylonitic layers with a very weak LPO (J=2.54; M=0.04) and a mean grain size of ∼12 μm (B1 area); (2) mylonitic connected bands adjacent to the ultramylonites with a strong LPO (J=16.80; M=0.46) and a grain size of ∼34 μm (B2 area) and (3) mylonitic areas with a moderate LPO (J=4.79; M=0.14) and a grain size of ∼33 μm (B3 area). The colour coding in Fig. 5a indicates that olivine -axes are parallel to Y in the B2 bands, consistent with the dominant E-type fabric in these areas. This feature is supported by a BA index of 0.49 (Fig. 5b and Supplementary Fig. 4). In contrast, olivine -axes are mostly aligned with Y in B3 areas, giving rise to a different olivine fabric. With a BA index of 0.74 (Fig. 5b), the corresponding LPO is characterized by a combined pattern between E- and D-type fabrics (see Supplementary Fig. 4). Olivine grains in ultramylonites do not show a noticeable LPO, but their LPOs also display a combined pattern of D/E-type fabric with BA=0.60 (Fig. 5b and Supplementary Fig. 4). Moreover, detailed maps in ultramylonite layers highlight tiny grains of orthopyroxene (∼2 μm) located at quadruple junctions of grain boundaries. These quadruple junctions only occur in very fine-grained ultramylonites (Fig. 5c).
For each area (B1, B2 and B3) and the whole map of Fig. 5a, we show the distribution of olivine misorientation axes and related dislocation slip-systems at sub-grain boundaries (Fig. 6). As shown experimentally35, most sub-grain boundaries are related to tilt boundaries instead of twist boundaries. Plotting misorientation axes on IFPs provides a key to identifying the rotation axis of the main dislocation slip-system36,37 (Fig. 6a). In Fig. 6b, IPF are displayed for the whole map (in black) and specific areas from B1 to B3, as referred to in Fig. 5b. For the whole map and the B1 (blue) and B2 (purple) areas, the distribution of misorientation axes indicates rotation around the  axis, consistent with dominant dislocation slip on the (001) slip-system. On the other hand, misorientation axes are highly concentrated around the  axis at intermediate distances from the ultramylonites (red B3 area). This rotation axis is related to activation of the (010) slip-system in olivine35. These results reveal, therefore, a change in dislocation slip-system from (001) to (010) in mylonitic layers with increasing distance from ultramylonites.
During mantle flow, olivine aggregates deform by several competing mechanisms that include grain-size-insensitive creep, typically dislocation creep, and GSS creep, such as diffusion creep. GSS creep is also referred to as superplastic flow when GBS significantly contributes to the deformation15,38. During grain size reduction, the dominant deformation mechanism commonly switches from dislocation creep to GSS creep, causing a substantial drop in strength3,4. Furthermore, intense grain size reduction is often coeval with phase mixing. This provides conditions to keep grain size small by inhibiting grain growth through grain boundary pinning, enhancing GSS creep39.
In the Ronda shear zones, weak olivine LPO and phase mixing indicate that deformation occurred dominantly by GSS creep in ultramylonites39,40, and most likely by superplastic flow as suggested by quadruple junctions. These junction types arise when grain boundaries align with the shear plane due to significant GBS41. The crystallization of new orthopyroxene grains at quadruple junctions also suggests that phase nucleation occurred during ductile flow. Based on recent experiments42, it is very likely that strain localization in ultramylonites results from this phase nucleation process and associated grain size reduction. In contrast, strong olivine LPOs and numerous sub-grains indicate that dislocation creep is the dominant deformation mechanism for the rest of the Ronda shear zones (protolith, protomylonites and mylonites). Using the size of sub-grains and an olivine piezometer43, we estimate a stress between 100 and 200 MPa during deformation.
When olivine mainly deforms by dislocation creep, several dislocation slip-systems are involved to accommodate plastic strain, but generally, one dominant slip-system governs the development and geometry of the olivine LPO. So far, five olivine fabrics and associated slip-systems have been described (resulting in A-, B-, C-, D- and E-type fabrics). Their preferential activation depends on the deformation conditions including stress, pressure and/or water content27,44,45. While E-type fabric is generally recognized as resulting from activation of the (100) slip-system during deformation of olivine containing a moderate amount of water27 (Fig. 7a), the (010) slip-system is preferentially activated during deformation of dry or poorly hydrated olivine, giving rise to an A-type fabric44 (Fig. 7b). As for the D-type fabric, it has only been observed for dry or poorly hydrated olivine28,46 and recent experiments have described it as an incipient A-type fabric resulting from the activation of the same water-poor slip-system29 (Fig. 7c). All of these fabrics can develop at pressures lower than 3 GPa and below a stress of 300 MPa (refs 27, 46).
The development of E-type fabric in Ronda indicates that water was present during strain localization. This is confirmed by the presence of amphibole and our estimate of ∼600 p.p.m. H2O in the ultramylonites (Fig. 3e). However, the occurrence of combined D/E-type fabric shows that water content has changed during deformation, either from dry to wet conditions (that is, from D- to E-type) or vice versa (that is, from E- to D-type). The occurrence of millimetre-scale water gradients preserved in the mineral fabric (Fig. 5) does not support progressive hydration through water infiltration. Indeed, even in shear zones where very high deformation rates are expected (10−12–10−10 s−1), many years of deformation are needed to generate a mineral fabric, particularly with high fabric strength29,30. At a millimetre scale, it is a matter of days to diffuse water through an olivine aggregate, even through the crystal lattice47. This rate of water diffusion excludes in situ measurements of water content in our samples (see the methods section). The systematic occurrence of E-type fabric, more or less combined with D-type fabric, also suggests that water was present everywhere in the study area before strain localization, contrary to the distal protolith where a dry A-type fabric has been described22,25. The most likely scenario is, therefore, that the D-type fabric overprinted the E-type fabric during strain localization.
The distribution of olivine fabrics implies that water progressively converged towards the ultramylonites during deformation. This gave rise to water accumulation at the edge of the mylonites (E-type fabric), and overprinting of E-type fabric by D-type fabric where olivine was progressively dried (Fig. 8a). Assuming that water mainly diffuses along grain boundaries—grain boundary diffusion is indeed one order of magnitude faster than volume diffusion in olivine48—this overprinting is attributed to a switch in the dislocation slip-system where water is being exsolved from olivine, that is, where interstitial water is no longer available (Fig. 8a). We further attribute the D/E-type fabric in ultramylonites to grain size reduction and subsequent water distribution through increasing grain boundary abundance. While water content increases per volume unit, as shown by the distribution of amphibole, it decreases per unit of grain boundary (Fig. 8a). Water convergence is also expected at a larger scale considering the higher density of ultramylonites at the centre of mylonitic complexes. This results in partially dried olivine (D/E-type fabric) in between mylonitic complexes, as well as strain partitioning between water-poor and water-rich layers because of water/hydrolytic weakening (Fig. 8b). The sole exception concerns the ultramylonitic layers, where the expected strain hardening due to olivine drying is far outweighed by grain size reduction and phase mixing, both of which promote weakening by leading to dominant GSS creep49. The high amount of water at the centre of complexes further accounts for the occurrence of mode-I cracks in this area, particularly in mylonites where we expect to have the highest amount of water at grain boundaries (Fig. 8b).
Altogether, these features show that water draining in ductile shear zones does not only occur in the crust, but also in the upper mantle, and necessarily on a long-term basis in order to be preserved into the olivine fabric. Our documentation of water accumulation around ultramylonites further excludes grain size reduction as a unique ‘passive’ process accounting for water draining. They show instead that water converged as a result of ‘dynamic’ pumping during ductile flow. Although we have no direct evidence of strain-induced cavities, the convergence of water towards layers dominated by superplastic flow supports creep cavitation as the source responsible for water-rich fluid pumping (Fig. 8a). GBS-induced quadruple junctions are also typical sites for cavitation41. In addition, the presence of nucleated phases at quadruple junctions, as well as a low porosity, suggests that opening cavities are filled by new precipitating phases, which may enhance long-term water pumping. While creep cavitation is a transient process that first opens cavities, by pumping fluid, and then closes cavities by expelling fluid too13, the filling of opening cavities by new solid phases precludes fluid expulsion. Lateral fluid-assisted diffusion of mineral components is likewise expected to compensate for the local deficit of fluid solute induced by phase nucleation25,50. This accounts for enrichment in secondary phases within ultramylonites, particularly for orthopyroxene, which only relies on increasing silica content.
To conclude, we here ascribe dynamic long-term fluid pumping during the deformation process of GBS and related creep cavitation, both arising from grain size reduction through phase nucleation. As grain size reduction is a ubiquitous feature of mantle and crustal shear zones39,40,51,52, this process is expected to occur wherever strain highly localizes, such as plate boundaries and at the base of the lithosphere-asthenosphere boundary. Our study therefore confirms a ductile process that could provide important fluid reservoirs, impacting the water distribution and related features of both the crust and mantle.
Scanning electron microscope
BSE images were acquired on polished (using 1 μm diamond paste) and carbon coated (20 nm) thin sections using a TESCAN scanning electron microscope (SEM) at ISTO/BRGM (University of Orléans, France). The SEM was set at an accelerating voltage of 20 kV, a probe current of ∼6 nA and a working distance of 10 mm. We also used the TESCAN SEM at ISTO/BRGM with an accelerating voltage of 15 kV to perform energy-dispersive X-ray spectroscopy (EDS) on cracked spinel.
Major-element compositions and element maps were acquired using a CAMECA SX-five electron microprobe at ISTO/BRGM (University of Orléans, France). The analyses were performed on carbon-coated (20 nm), polished (using 1 μm diamond paste) thin sections using an accelerating voltage of 15 kV, a probe current of 10 nA and a beam diameter of 1 μm. For point analyses, we applied a counting time of 10 and 5 s for peak and background signals, respectively. Calibration was carried out against natural and synthetic minerals, including albite, MnTiO3, Fe2O3, MgO, Al2O3, andradite, orthose, Cr2O3, NiO and vanadinite.
The pseudosections were calculated using Perple_X (ref. 53) with the thermodynamic database of Holland and Powell54. It was built in the NCFMASH (Na2O-CaO-FeO-MgO-Al2O3-SiO2-H2O) chemical system. The weight percent of oxides was calculated by combining the volume percent of mineral modes, which were determined by point counting on elements maps, and the composition of mineral phases determined at the microprobe. No phase was excluded, but we did not consider chromium in spinel for simplicity. The solution models are from Holland and Powell54,55 and Holland et al.56, except for antigorite57, amphibole58 and plagioclase59.
Electron backscatter diffraction
The EBSD data were collected using an SEM either coupled to an EDAX Pegasus system at ISTO/BRGM, or using an e−FlashHR detector at Bruker Nano Analytics GmbH (Berlin, Germany). Crystallographic information was also combined with EDX (Energy-Dispersive X-ray) spectroscopy to identify phases. The EBSD measurements were performed at a working distance of 15 or 20 mm with an accelerating voltage of 20 kV and a probe current of ∼6 nA on polished thin sections (diamond paste of 0.25 μm followed by colloidal silica). In order to avoid indexing errors, we constructed pole figures using one measurement per grain collected manually. The iso-contours and grey shadings on pole figures were plotted using the Unicef careware software package written by David Mainprice (www.gm.univ-montp2.fr/PERSO/mainprice/W_data/CareWare_Unicef_Programs). We further used the open-source matlab-based MTEX toolbox to construct IPF of misorientation axes and to calculate the texture (J) and misorientation (M) indices30,60. To plot pole figures and calculate the J index, we used a Gaussian half-width angle of 10°.
For acquisition data processing and analysis of EBSD maps, we used the OIM Analysis 7 software (EDAX) and/or HKL Tango channel 5 system (Oxford Instruments). The indexing hit rates were higher than 80% for all maps, including cavities. We used a minimum of six indexed bands and a maximum band mismatch of 2°. The outlier pixels and zero solutions were removed up to a maximum pixel count of 10. Grain detection was performed using correlated misorientation angles higher than 10° for grain boundaries and between 2 and 10° for sub-grain boundaries. Each grain contains a minimum of five pixels considering a hexagonal grid. The average grain size corresponds to the mean equivalent diameter calculated using the grain ellipsoid fit. EBSD maps were used only to calculate the grain size in the protomylonites (edge), mylonites and ultramylonites. In the protolith and centre of protomylonites, we applied the length-intercept method on BSE images because of their larger grain sizes.
In situ water content
A major problem with in situ measurements of water content in natural rocks is how far the measured value reflects the water content at lithosphere depths. It has been demonstrated that for orogenic peridotites where the exhumation is slow compared to xenoliths, there are limits in the use of Fourier transform infrared spectroscopy—or other analytical methods—in the determination of olivine water content61,62. Water from crustal fluids may be indeed introduced or may precipitate as fluid inclusions and hydrated phases in olivine during exhumation63. Note that the Ronda peridotites have been lying on the crust of the internal Betics since more than 20 Ma (ref. 21). Furthermore, lab experiments have demonstrated that water diffusion is fast in olivine (10−10 m2 s−1) relative to geological strain rates47. If water was originally present in olivine grains, it could have been easily diffused out during the peridotite exhumation stage64. The presence of post-tectonic serpentine in the studied peridotites may also have strongly modified olivine water content after deformation. It is therefore very likely that in situ measurements in olivine of the Ronda peridotites would not reflect the water content of the minerals prior to exhumation, and particularly during their deformation. In contrast, lab experiments have shown that olivine LPOs are sensitive to water content27,44. As they directly arise from rock deformation, they provide a more robust method—yet qualitative—to document syn-tectonic water content in upper mantle rocks61.
The authors declare that most of the data supporting the findings of this study are available within the paper (and its Supplementary Information files). The source files of EBSD data are available from the corresponding author on reasonable request.
How to cite this article: Précigout, J. et al. Water pumping in mantle shear zones. Nat. Commun. 8, 15736 doi: 10.1038/ncomms15736 (2017).
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Kronenberg, A. K. & Tullis, J. Flow strengths of quartz aggregates: grain size and pressure effects due to hydrolytic weakening. J. Geophys. Res. 89, 4281–4297 (1984).
Mackwell, S. J., Kohlstedt, D. L. & Paterson, M. S. The role of water in the deformation of olivine single crystal. J. Geophys. Res. 90, 11,319–11,333 (1985).
Karato, S.-I., Paterson, M. S. & FitzGerald, J. D. Rheology of synthetic olivine aggregates: influence of grain size and water. J. Geophys. Res. 91, 8151–8176 (1986).
Hirth, G. & Kohlstedt, D. L. Rheology of the upper mantle and the mantle wedge: a view from the experimentalists. Geophys. monogr. 138, 83–105 (2003).
Jackson, J. Strength of the continental lithosphere: time to abandon the jelly sandwich? GSA today 12, 4–10 (2002).
Burov, E. B. & Watts, A. B. The long-term strength of continenal lithosphere: ‘Jelly sandwich’ or ‘crème brûlée’. GSA Today 16, 4–10 (2006).
Gueydan, F., Morency, C. & Brun, J.-P. Continental rifting as a function of lithosphere mantle strength. Tectonophysics 460, 83–93 (2008).
Griggs, D. T. Hydrolytic weakening of quartz and other silicates. Geophys. J. R. Astro. Soc. 14, 19–31 (1967).
Kohlstedt, D. L. The role of water in high-temperature rock deformation. Rev. Min. Geochem. 62, 377–396 (2006).
McCaig, A. M. Deep fluid circulation in fault zones. Geology 16, 867–870 (1988).
Sibson, R. H., Moore, J., McM., & Rankin, A. H. Seismic pumping—a hydrothermal fluid transport mechanism. J. Geol. Soc. Lond. 131, 653–659 (1975).
Wark, D. A. & Watson, E. B. Effect of grain size on the distribution and transport of deep-seated fluids and melts. Geophys. Res. Lett. 27, 2029–2032 (2000).
Fusseis, F., Regenauer-Lieb, K., Liu, J., Hough, R. M. & De Carlo, F. Creep cavitation can establish a dynamic granular fluid pump in ductile shear zones. Nature 459, 974–977 (2009).
Chokshi, A. H. Cavity nucleation and growth in superplasticity. Mater. Sci. Eng. A 410–411, 95–99 (2005).
Oliot, E., Goncalves, P., Schulmann, K., Marquer, D. & Lexa, O. Mid-crustal shear zone formation in granitic rocks: constraints from quantitative textural and crystallographic preferred orientations analyses. Tectonophysics 612–613, 63–80 (2014).
Kassner, M. E. & Hayes, T. A. Creep cavitation in metals. Int. J. Plast. 19, 1715–1748 (2003).
Dimanov, A., Rybacky, E., Wirth, R. & Dresen, G. Creep and strain-dependent microstructures of synthetic anorthite-diopside aggregates. J. Struct. Geol. 29, 1049–1069 (2007).
Rybacki, E., Wirth, R. & Dresen, G. High-strain creep of feldspar rocks: Implications for cavitation and ductile failure in the lower crust. Geophys. Res. Lett. 35, L04304 (2008).
Menegon, L., Fusseis, F., Stünitz, H. & Xiao, X. Creep cavitation bands control porosity and fluid flow in lower crustal shear zones. Geology 43, 227–230 (2015).
Garrido, C. J. et al. Garnet lherzolite and garnet-spinel mylonite in the Ronda peridotite: vestiges of oligocene backarc mantle lithospheric extension in the western Mediterranean. Geology 39, 927–930 (2011).
Précigout, J., Gueydan, F., Garrido, C. J., Cogné, N. & Booth-Rea, G. Deformation and exhumation of the Ronda peridotite (Spain). Tectonics 32, 1011–1025 (2013).
Van der Wal, D. & Vissers, R. L. M. Structural petrology of the Ronda peridotite, SW spain: deformation history. J. Petrol. 37, 23–43 (1996).
Hidas, K. et al. Backarc basin inversion and subcontinental mantle emplacement in the crust: kilometre-scale folding and shearing at the base of the proto-Alborán, lithospheric mantle (Betic Cordillera, southern Spain). J. Geol. Soc. Lond. 170, 47–55 (2012).
Hubbert, M. K. & Willis, D. G. Mechanics of hydraulic fracturing. AAPG special volumes A075, 239–257 (1972).
Hidas, K. et al. Fluid-assisted strain localization in the shallow subcontinental lithospheric mantle. Lithos 262, 636–650 (2016).
Hidas, K. et al. Strain localization in pyroxenite by reaction-enhanced softening in the shallow subcontinental lithospheric mantle. J. Petrol. 54, 1997–2031 (2013).
Katayama, I., Jung, H. & Karato, S.-I. New type of olivine fabric from deformation experiments at modest water content and low stress. Geology 32, 1045–1048 (2004).
Bystricky, M., Kunze, K., Burlini, L. & Burg, J.-P. High shear strain of olivine aggregates: rheological and seismic consequences. Science 290, 1564–1567 (2000).
Hansen, L. N., Zhao, Y.-H., Zimmerman, M. E. & Kohlstedt, D. L. Protracted fabric evolution in olivine: implications for the relationship among strain, crystallographic, and seismic anisotropy. Earth Planet. Sci. Lett. 387, 157–158 (2014).
Mainprice, D., Bachmann, F., Hielscher, R. & Schaeben, H. in Rock Deformation from Field, Experiments and Theory: A Volume in Honour of Ernie Rutter Vol. 409, 251–271 (eds Faulkner, D. R., Mariani, E. & Mecklenburg, J.) (Geological Society, 2014).
Chatzaras, V. et al. Axial-type olivine crystallographic preferred orientations: the effect of strain geometry in mantle texture. J. Geophys. Res. Solid Earth 121, 1–28 (2016).
Bunge, H. Texture Analysis in Materials Science: Mathematical Models 593, Butterworths (1982).
Skemer, P., Katayama, I., Jiang, Z. & Karato, S.-I. The misorientation index: development of a new method for calculating the strength of lattice-preferred orientation. Tectonophysics 411, 157–167 (2005).
Ramsay, J. G. Shear zone geometry: a review. J. Struct. Geol. 2, 83–99 (1980).
Hansen, L. N., Zimmerman, M. E. & Kohlstedt, D. L. The influence of microstructure on deformation of olivine in the grain-boundary sliding regime. J. Geophys. Res. 117, B09201 (2012).
Mehl, L., Hacker, B. R., Hirth, G. & Kelemen, P. B. Arc-parallel flow within the mantle wedge: evidence from the accreted Talkeetna arc, south central Alaska. J. Geophys. Res. 108, 2375 (2003).
Précigout, J. & Hirth, G. B-type olivine fabric induced by grain boundary sliding. Earth Planet. Sci. Lett. 395, 231–240 (2014).
Langdon, T. G. Grain boundary sliding revisited: development in sliding over four decades. J. Mater. Sci. 41, 597–609 (2006).
Warren, J. M. & Hirth, G. Grain size sensitive deformation mechanisms in naturally deformed peridotites. Earth Planet. Sci. Lett. 248, 438–450 (2006).
Drury, M. R., Ave’Lallemant, H. G., Pennock, G. M. & Palasse, L. N. Crystal preferred orientation in peridotite ultramylonites deformed by grain size sensitive creep, Etang de Lers, Pyrenees, France. J. Struct. Geol. 33, 1776–1789 (2011).
Ree, J.-H. Grain boundary sliding and development of grain boundary openings in experimentally deformed octachloropropane. J. Struct. Geol. 16, 403–418 (1994).
Précigout, J. & Stünitz, H. Evidence of phase nucleation during olivine diffusion creep: a new perspective for mantle strain localisation. Earth Planet. Sci. Lett. 455, 94–105 (2016).
Van der Wal, D., Chopra, P., Drury, M. & Fitz Gerald, J. Relationships between dynamically recrystallized grain size and deformation conditions in experimentally deformed olivine rocks. Geophys. Res. Lett. 20, 1479–1482 (1993).
Jung, H. & Karato, S.-I. Water-induced fabric transitions in olivine. Science 293, 1460–1463 (2001).
Ohuchi, T., Kawazoe, T., Nishihara, Y., Nishiyama, N. & Irifune, T. High pressure and temperature fabric transitions in olivine and variations in upper mantle seismic anisotropy. Earth Planet. Sci. Lett. 304, 55–63 (2011).
Demouchy, S., Tommasi, A., Barou, F., Mainprice, D. & Cordier, P. Deformation of olivine in torsion under hydrous conditions. Phy. Earth Planet. Int. 202–203, 56–70 (2012).
Demouchy, S. & Mackwell, S. Water diffusion in synthetic iron-free olivine. Phys. Chem. Minerals 30, 486–494 (2003).
Demouchy, S. Diffusion of hydrogen in olivine grain boundaries and implications for the survival of water-rich zones in the Earth’s mantle. Earth Planet. Sci. Lett. 295, 305–313 (2010).
Fei, H., Wiedenbeck, M., Yamazaki, D. & Katsura, T. Small effect of water on upper-mantle rheology based on silicon self-diffusion coefficients. Nature 498, 213–216 (2013).
Linckens, J., Herwegh, M., Müntener, O. & Mercolli, I. Evolution of a polymineralic mantle shear zone and the role of second phases in the localisation of deformation. J. Geophys. Res. 116, B06210 (2011).
Drury, M. R., Vissers, R. L. M., Van der Wal, D. & Hoogerduijn Strating, E. H. Shear localisation in upper mantle peridotites. Phys. App. Geophys. 137, 439–460 (1991).
Rutter, E. H. & Brodie, K. H. The role of tectonic grain size reduction in the rheological stratification of the lithosphere. Geologische Rundschau 77, 295–308 (1988).
Connolly, J. A. D. The geodynamic equation of state: what and how. Geochem. Geophys. Geosyst. 10, Q10014 (2009).
Holland, T. & Powell, R. An internally consistent thermodynamic data set for phases of petrological interest. J. Metam. Geol. 16, 309–343 (1998).
Holland, T. & Powell, R. Thermodynamics of order-disorder in minerals. 2. Symmetric formalism applied to solid solutions. Am. Mineral. 81, 1425–1437 (1996).
Holland, T., Baker, J. & Powell, R. Mixing properties and activity-composition relationships of chlorites in the system MgO-FeO-Al2O3-SiO2-H2O. Eur. J. Mineral. 10, 395–406 (1998).
Padrón-Navarta, J. A. et al. Tschermak’s substitution in antigorite and consequences for phase relations and water liberation in high-grade serpentinites. Lithos 178, 186–196 (2013).
Dale, J., Powell, R., White, R. W., Elmer, F. L. & Holland, T. J. B. A thermodynamic model for Ca-Na clinoamphiboles in Na2O-CaO-FeO-MgO-Al2O3-SiO2-H2O for petrological calculations. J. Metam. Geol. 23, 771–791 (2005).
Newton, R. C., Charlu, T. V. & Kleppa, O. J. Thermochemistry of the high structural state plagioclases. Geochem. Cosmochim. Acta 44, 933–941 (1980).
Bachmann, F., Hielscher, R. & Schaeben, H. Texture analysis with MTEX—free and open software toolbox. Solid State Phenomena 160, 63–68 (2010).
Palasse, L. N. Microstructural evolution and seismic anisotropy of upper mantle rocks in rift zones. Geologica Ultraiectina. Vol. 300 (PhD thesis, Utrecht Univ. Repository, 2008).
Warren, J. M. & Hauri, E. H. Pyroxenes as tracers of mantle water variations. J. Geophys. Res. Solid Earth 119, 1851–1881 (2014).
Kostenko, O., Jamtveit, B., Austrheim, H., Pollok, K. & Putnis, C. The mechanism of fluid infiltration in peridotites at Almklovdalen, western Norway. Geofluids 2, 203–215 (2002).
Peslier, A. H. & Luhr, J. F. Hydrogen loss from olivines in mantle xenoliths from Simcoe (USA) and Mexico: mafic alkali magma ascent rates and water budget of the sub-continental lithosphere. Earth Planet. Sci. Lett. 242, 302–319 (2006).
Brey, G. P. & Köhler, T. Geothermobarometry in four-phase Lherzolites II. New thermobarometers, and practical assessment of existing thermobarometers. J. Petrol. 31, 1353–1378 (1990).
Witt-Eickschen, G. & Seck, H. A. Solubility of Ca and Al in orthopyroxene from spinel peridotite: an improved version of an empirical geothermometer. Contrib. Min. Petrol. 106, 431–439 (1991).
Jianping, L., Kornprobst, J., Vielzeuf, D. & Fabriès, J. An improved experimental calibration of the olivine-spinel geothermometer. Chinese J. Geochem. 14, 68–77 (1995).
Putirka, K. D. Thermometers and barometers for volcanic systems. Rev. Min. Geochem. 69, 61–120 (2008).
We thank Carlos J. Garrido for his help in acquiring field data. We also thank Holger Stünitz, Nicole Le Breton, Alexandra R. L. Kushnir and Jessica M. Warren for their help and comments, as well as Florian Fusseis for his fruitful reviews. This project has been funded by the ETH Zürich (ETH fellowship application FEL01 09-3), ERC RHEOLITH (Grant n° 290864), Labex Voltaire (ANR-10-LABX-100-01) and ANR DSP-Tibet (ANR-13-BS06-0012).
The authors declare no competing financial interests.
Supplementary Figures, Supplementary Tables and Supplementary References (PDF 9314 kb)
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
About this article
Cite this article
Précigout, J., Prigent, C., Palasse, L. et al. Water pumping in mantle shear zones. Nat Commun 8, 15736 (2017). https://doi.org/10.1038/ncomms15736
This article is cited by
Porosity induced by dislocation dynamics in quartz-rich shear bands of granitic rocks
Scientific Reports (2022)
Evaluation of grain boundaries as percolation pathways in quartz-rich continental crust using Atomic Force Microscopy
Scientific Reports (2021)
Oceanic transform fault seismicity and slip mode influenced by seawater infiltration
Nature Geoscience (2021)
Excess water storage induced by viscous strain localization during high-pressure shear experiment
Scientific Reports (2019)
Water-rich sublithospheric melt channel in the equatorial Atlantic Ocean
Nature Geoscience (2018)
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.