solution. In other words, they are not the product of provides a record of the original size of the grain. The local replacement of existing detrital grains by central idea behind the pds method is that principal recrystallziation. (4) There are many descriptions in directions with S<1. SMt deformation has reduced the literature of a"fine-grained matrix "in low-grade the average dimension of the detrital grains by a factor immature sandstones. The matrix problem"is at the equivalent to the principal stretch. In contrast, the centre of the old debate about the distinction of average initial dimension of a detrital grain should be greywacke from arkose(see Dickinson, 1970, for a preserved in the X direction because the grains lack review). More recently, some of our colleagues have any significant internal deformation and because the suggested that the matrix of the rock may account for original grain surface is mantled by fibre overgrowth the volume loss that we have measured. Deformed Therefore, a contractive principal stretch can be lithic grains can appear like matrix, but the outlines of determined by finding the average grain dimension such grains are usually easily recognized in plane light arallel to a contractive principal direction and (pseudomatrix of Dickinson, 1970). Our experience is dividing it by the average grain dimension parallel to that the rest of the"matrix " is fibre overgrowth. The X Dimensions are measured in a two-dimensional overgrowths appear as a fine-grained matrix"when thin section, so a correction is needed to get the viewed in sections oblique to the X direction. XY and appropriate three-dimensional result(see Feehan and XZ sections are needed to see the overgrowth texture Brandon, 1999 for details) with XZ sections providing the best view. Diagnostic Our measurements were made using a petrographic features include the elongated habit of the fibre microscope with a camera lucida tube and digitising minerals(commonly quartz and mica), the consistent tablet. Measurements are precise to better than #3 um orientation of the fibres across the section and the The dimension of each grain is represented by its generally uniform composition of the overgrowths. (5) caliper dimensions(or projected dimensions)in the We found no petrographic evidence for syntaxial principal directions lying in the section. The caliper overgrowths, but cathodoluminescence work is needed dimensions of the grains are not affected in any to fully test this conclusion significant way by grain rotations associated with compaction. For instance, PDS measurements on 4.2. Methods undeformed sandstones gave undeformed results(i.e Our study employs the pds(projected Dimension S-1)(Ring and Brandon, 1999 Strain), Mode, and Fiber methods for measuring The Mode method is used to determine the strains and internal rotations in sandstones deformed extensional strain recorded by the fibre overgrowths by the SMT mechanism. a brief summary is provided The modal percentage of fibres in a rock is directly here. Further details can be found in Feehan and related to the absolute extensional stretch in the rock Brandon(1999)and ring and brandon (1999) Fibre modes are most easily measured in the XZ Relevant computer programs are available at section. For unidirectional fibres, S,=(1-m where www.geologyyaleedu/-brandon m is the modal fraction of fibre Traditional methods, such as the r/ method, are Given absolute strains, the volume stretch Sr not suitable because the grains did not deform as final volume/initial volume) is equal to the product of passive markers but rather by truncation and the principal stretches(S,. Sy.S,). Because our precipitation along grain boundaries. In our discussion methods focus entirely on the loss of mass from grains here, the principal stretches are designated as Sr> Sy> and the amount of mass locally precipitated, our S, where S= final length/initial length estimates of Sp, only represent the mass-transfe Measurements were made using XZ and Xr thin omponent of the volume strain. Other sources of sections. Samples in this study have unidirectional volume strain include changes in porosity and mineral fibres, which means that Sx>I and Sy and s,< 1 density. Porosity is thought to have been small at the Thus, S, and S, were determined by the PDs method start of smt deformation and is thus ignored and Sy by the Mode method Changes in mineral density are insignificant at the low The Pds method is used to metamorphic grade in our study shortening produced by dissolution of grain The geometric relationship between the fibre boundaries. The method exploits the fact that for SMt overgrowths and the trace of cleavage was used to deformation, the dimensions of the detrital quartz and estimate the internal rotation associated with Smt feldspar grains remain unchanged in the X direction deformation. This method is described in Ring and (i.e, deformation is intergranular, not intragranular) Brandon(1999). The basic idea is that the fibre Therefore, the grain diameter in the X direction overgrowths track the incremental extension direction 6
6 solution. In other words, they are not the product of local replacement of existing detrital grains by recrystallziation. (4) There are many descriptions in the literature of a “fine-grained matrix” in low-grade immature sandstones. The “matrix problem” is at the centre of the old debate about the distinction of greywacke from arkose (see Dickinson, 1970, for a review). More recently, some of our colleagues have suggested that the matrix of the rock may account for the volume loss that we have measured. Deformed lithic grains can appear like matrix, but the outlines of such grains are usually easily recognized in plane light (pseudomatrix of Dickinson, 1970). Our experience is that the rest of the “matrix” is fibre overgrowth. The overgrowths appear as a “fine-grained matrix” when viewed in sections oblique to the X direction. XY and XZ sections are needed to see the overgrowth texture, with XZ sections providing the best view. Diagnostic features include the elongated habit of the fibre minerals (commonly quartz and mica), the consistent orientation of the fibres across the section, and the generally uniform composition of the overgrowths. (5) We found no petrographic evidence for syntaxial overgrowths, but cathodoluminscence work is needed to fully test this conclusion. 4.2. Methods Our study employs the PDS (Projected Dimension Strain), Mode, and Fiber methods for measuring strains and internal rotations in sandstones deformed by the SMT mechanism. A brief summary is provided here. Further details can be found in Feehan and Brandon (1999) and Ring and Brandon (1999). Relevant computer programs are available at www.geology.yale.edu/~brandon. Traditional methods, such as the Rf /φ method, are not suitable because the grains did not deform as passive markers but rather by truncation and precipitation along grain boundaries. In our discussion here, the principal stretches are designated as SX ≥ SY ≥ SZ, where S = final length/initial length. Measurements were made using XZ and XY thin sections. Samples in this study have unidirectional fibres, which means that SX > 1 and SY and SZ < 1. Thus, SY and SZ were determined by the PDS method, and SX by the Mode method. The PDS method is used to measure the average shortening produced by dissolution of grain boundaries. The method exploits the fact that for SMT deformation, the dimensions of the detrital quartz and feldspar grains remain unchanged in the X direction (i.e., deformation is intergranular, not intragranular). Therefore, the grain diameter in the X direction provides a record of the original size of the grain. The central idea behind the PDS method is that principal directions with S < 1, SMT deformation has reduced the average dimension of the detrital grains by a factor equivalent to the principal stretch. In contrast, the average initial dimension of a detrital grain should be preserved in the X direction because the grains lack any significant internal deformation and because the original grain surface is mantled by fibre overgrowths. Therefore, a contractive principal stretch can be determined by finding the average grain dimension parallel to a contractive principal direction and dividing it by the average grain dimension parallel to X. Dimensions are measured in a two–dimensional thin section, so a correction is needed to get the appropriate three–dimensional result (see Feehan and Brandon, 1999 for details). Our measurements were made using a petrographic microscope with a camera lucida tube and digitising tablet. Measurements are precise to better than ±3 µm. The dimension of each grain is represented by its caliper dimensions (or projected dimensions) in the principal directions lying in the section. The caliper dimensions of the grains are not affected in any significant way by grain rotations associated with compaction. For instance, PDS measurements on undeformed sandstones gave undeformed results (i.e. S ~ 1) (Ring and Brandon, 1999). The Mode method is used to determine the extensional strain recorded by the fibre overgrowths. The modal percentage of fibres in a rock is directly related to the absolute extensional stretch in the rock. Fibre modes are most easily measured in the XZ section. For unidirectional fibres, SX = (1 - m) -1, where m is the modal fraction of fibre. Given absolute strains, the volume stretch SV (= final volume/initial volume) is equal to the product of the principal stretches (SX ⋅ SY ⋅ SZ). Because our methods focus entirely on the loss of mass from grains and the amount of mass locally precipitated, our estimates of SV only represent the mass-transfer component of the volume strain. Other sources of volume strain include changes in porosity and mineral density. Porosity is thought to have been small at the start of SMT deformation and is thus ignored. Changes in mineral density are insignificant at the low metamorphic grade in our study area. The geometric relationship between the fibre overgrowths and the trace of cleavage was used to estimate the internal rotation associated with SMT deformation. This method is described in Ring and Brandon (1999). The basic idea is that the fibre overgrowths track the incremental extension direction
during the deformation, whereas the cleavage using the Hencky method where only the stretch approximates the XY plane of the finite deformation tensor is needed(see Appendix B of Brandon, 1995) Internal rotation was estimated using the FIBer program to model the shape of about 30-50 fibres 3.3. Results The internal rotation axis is assumed to parallel o digitised in the XZ section(Ring and Brandon, 1999) bles i and 2 list our deformation measurements for the Verrucano and melser sandstones from the Tables I and 2 report internal rotation and average Helvetic nappes above the glarus thrust, and for the kinematic numbers for Smt deformation Internal Taveyannaz and north helvetic flysch sandstones rotation is represented by a right-handed rotation axis from the Infrahelvetic complex below the thrust defined by a trend and plunge, and a rotation angle, Sample locations are shown in Fig. 2. Note that most Q2j. Wn and Wm are the average kinematic vorticity of our samples from the Helvetic nappes an e trom numbers and Am the average kinematic dilatancy more than 1 km above the thrust plane The stereograms(Fig. 5)show that the z directions number(Means et al., 1980; Passchier, 1991; Means are clustered around a steeply plunging maximum, and 1994; Ring and Brandon, 1999). Definitions and other the X and y directions are scattered in a weakly details are given in Ring and Brandon(1999). A brief defined subhorizontal girdle. The average for Z(Table review is provided here. The m subscript for the 2)defines the average flattening plane, which has a kinematic numbers indicates a path-averaged value strike of 30 and a dip of 10 to the SE. The average X assuming a steady three-dimensional deformation. If direction is close to horizontal, although its trend SMt deformation were unsteady, then Wm would changes from 200 above the glarus thrust to 160% have no direct relationship to the time history of the below the thrust. The Nadai plot(Fig. 6)shows a instantaneous kinematic vorticity number Wk. The scatter of both prolate and oblate strain symmetries simple geometry of the overgrowths in our samples The strain type(Fig. 7)is generally weakly suggests that Smt deformation was fairly steady, at constrictional, as indicated by S, <1 (Table 1). The principal stretches indicate that the constrictional least in its orientation. An asterisk indicates that the aspect of the strain is primarily the result of shortening kinematic number is based on the deviatoric stretching in the y and Z directions and not extension in X rate rather than the absolute stretching rate. Thus, a Above the glarus thrust. local measurements from coaxial deformation is indicated by =0 and a non- the Verrucano and melser sandstones have X strains coaxial simple-shear deformation by Wm=l ranging from +5 to +43%and Z strains, from to regardless of the amount of volume strain. A 52%. The tensor average indicates absolute principal stretches of 1. 07.0.89 and 068. Thus at the regional describes the average ratio of the volume -strain rate lative to the deviatoric stretching rate(Passche scale, SMT deformation was constrictional (1>S 1991; Ring and Brandon, 1999). The deformation is S,)and approximately plane strain(Sr= 1). This isochoric if A=0. dilatant if A>0 and surprising result, with S, =1, stems from the variable orientations of X and y in the flattening plane, which compactive if Am <0. For example, a deformation means that local extensional strains are averaged out at involving uniaxial shortening and an equal loss of the regional scale volume, would have A=-l because the rates of Below the Glarus thrust, individual samples show haller strain magnitudes in X and y. For instance. X olume strain and deviatoric strain would be equal but strains range from +3 to +12%. In contrast, the tensor opposite in sign average indicates absolute principal stretches of 0.99 Table 2 reports tensor averages for our deformation 0.88 and 0.73, which is nearly identical to the tensor measurements As discussed in Brandon(1995) average above the thrust Thus the low strain deformation data must be averaged in tensor form to magnitudes in X and y are only manifested at the local ensure that the magnitudes and directions of the cale. At the regional scale, sandstones above and I stretches and rotations are correctly below the fault record a similar smt deformation associated. If the rotational component of the deformation is small. then one can average the stretch involving constrictional plane strain The absolute strain data indicate pronounced mass- tensor and the internal rotation tensor separately, without introducing significant errors(Brandon, loss volume strains ranging from-9 to-54% in the 1995). In this study, tensor averages were calculated sandstones above the glarus thrust. and -8 to -49% for sandstones below the thrust. The average is the same for both 36%. At the outcrop
7 during the deformation, whereas the cleavage approximates the XY plane of the finite deformation. Internal rotation was estimated using the FIBER program to model the shape of about 30-50 fibres digitised in the XZ section (Ring and Brandon, 1999). The internal rotation axis is assumed to parallel Y. Tables 1 and 2 report internal rotation and average kinematic numbers for SMT deformation. Internal rotation is represented by a right-handed rotation axis, defined by a trend and plunge, and a rotation angle, Ωi. Wm and * Wm are the average kinematic vorticity numbers and * Am the average kinematic dilatancy number (Means et al., 1980; Passchier, 1991; Means, 1994; Ring and Brandon, 1999). Definitions and other details are given in Ring and Brandon (1999). A brief review is provided here. The m subscript for the kinematic numbers indicates a path-averaged value assuming a steady three-dimensional deformation. If SMT deformation were unsteady, then * Wm would have no direct relationship to the time history of the instantaneous kinematic vorticity number * Wk . The simple geometry of the overgrowths in our samples suggests that SMT deformation was fairly steady, at least in its orientation. An asterisk indicates that the kinematic number is based on the deviatoric stretching rate rather than the absolute stretching rate. Thus, a coaxial deformation is indicated by * Wm = 0 and a noncoaxial simple-shear deformation by * Wm = 1, regardless of the amount of volume strain. * Am describes the average ratio of the volume-strain rate relative to the deviatoric stretching rate (Passchier, 1991; Ring and Brandon, 1999). The deformation is isochoric if * Am = 0, dilatant if * Am > 0, and compactive if * Am < 0. For example, a deformation involving uniaxial shortening and an equal loss of volume, would have * Am = –1 because the rates of volume strain and deviatoric strain would be equal but opposite in sign. Table 2 reports tensor averages for our deformation measurements. As discussed in Brandon (1995), deformation data must be averaged in tensor form to ensure that the magnitudes and directions of the principal stretches and rotations are correctly associated. If the rotational component of the deformation is small, then one can average the stretch tensor and the internal rotation tensor separately, without introducing significant errors (Brandon, 1995). In this study, tensor averages were calculated using the Hencky method where only the stretch tensor is needed (see Appendix B of Brandon, 1995). 3.3. Results Tables 1 and 2 list our deformation measurements for the Verrucano and Melser sandstones from the Helvetic nappes above the Glarus thrust, and for the Taveyannaz and North Helvetic flysch sandstones from the Infrahelvetic complex below the thrust. Sample locations are shown in Fig. 2. Note that most of our samples from the Helvetic nappes are from more than 1 km above the thrust plane. The stereograms (Fig. 5) show that the Z directions are clustered around a steeply plunging maximum, and the X and Y directions are scattered in a weakly defined subhorizontal girdle. The average for Z (Table 2) defines the average flattening plane, which has a strike of 30° and a dip of 10° to the SE. The average X direction is close to horizontal, although its trend changes from 200° above the Glarus thrust to 160° below the thrust. The Nadai plot (Fig. 6) shows a scatter of both prolate and oblate strain symmetries. The strain type (Fig. 7) is generally weakly constrictional, as indicated by SY < 1 (Table 1). The principal stretches indicate that the constrictional aspect of the strain is primarily the result of shortening in the Y and Z directions, and not extension in X. Above the Glarus thrust, local measurements from the Verrucano and Melser sandstones have X strains ranging from +5 to +43% and Z strains, from −22 to −52%. The tensor average indicates absolute principal stretches of 1.07, 0.89 and 0.68. Thus, at the regional scale, SMT deformation was constrictional (1 > SY > SZ) and approximately plane strain (SX ≈ 1). This surprising result, with SX ≈ 1, stems from the variable orientations of X and Y in the flattening plane, which means that local extensional strains are averaged out at the regional scale. Below the Glarus thrust, individual samples show smaller strain magnitudes in X and Y. For instance, X strains range from +3 to +12%. In contrast, the tensor average indicates absolute principal stretches of 0.99, 0.88 and 0.73, which is nearly identical to the tensor average above the thrust. Thus, the low strain magnitudes in X and Y are only manifested at the local scale. At the regional scale, sandstones above and below the fault record a similar SMT deformation involving constrictional plane strain. The absolute strain data indicate pronounced massloss volume strains ranging from −9 to −54% in the sandstones above the Glarus thrust, and −8 to −49% for sandstones below the thrust. The average is the same for both groups, −36%. At the outcrop scale
there is no evidence of where this missing mass went areas adjacent to the Glarus thrust, but the details of The volume fraction of veins in outcrops is generally their measurements and the localities are not no greater than a few percent. Thus, we conclude that discussed. Neither study provides any information SMt deformation was influenced by a large flux of about principal directions. By themselves, strain-ratio fluid that was able to dissolve the sandstones and to data have limited utility, but they do provide transport the dissolved load over a scale larger than information about the symmetry and magnitude of the our study area. deviatoric component of the strain. Thus, comparisons Figure 8 shows that volume strain and deviatoric with our data are limited to the Nadai plot( Fig. 6) strain are uncorrelated, which implies that these strains all data sets show a similar clustering along the are controlled by different processes. Based on our prolate/oblate boundary, so they share the same strain previous work, we have found that a strong correlation symmetry. In contrast, they have very different strain between deviatoric strain and volume strain only magnitudes In Figure 6, the strain ratio R(=S/s) occurs where fibre overgrowths are small or absent provides a useful measure of deviatoric strain (e.g. Feehan and Brandon, 1999). The reason is that magnitude(note that R is independent of volume volume and deviatoric strains are solely a function of strain). Our measurements have Rz ranging from 1.4 shortening strains in Y and Z. Deviatoric strains to 2.3, whereas Siddans'(1979)measurements range become uncorrelated with volume strain when there from 1.5 to 9.5. However the highest strain are variations in extensional strain(as indicated by magnitudes for Siddans' data are those from localities variations in the modal abundance of fibre near the Glarus thrust(grey triangles in Fig. 6b). The overgrowth). Uncorrelated volume strain and remaining localities(black triangles in Fig. 6b)show a deviatoric strain are observed in the eastern belt of closer correspondence to our results. The"averages the franciscan Complex of California (Ring and given in Milnes and Pfiffner(1977) have an Rx of-4 Brandon, 1999), as well as in our Glarus study here for rocks below the Glarus thrust(open triangle in Fig The variations in fibre overgrowth means that some of 6b)and -13 for rocks directly above the thrust(open the dissolved grain mass is re-precipitated locally in circle in Fig. 6b) the rock. The rock must extend in at least one These results suggest that differences between direction to be able to accommodate the locally these studies is due, at least in part, to difference in precipitated mass the heterogeneity influenced, at least in part, by the s verage of a heterog strain field wi Only 7 out of the 18 samples have sufficient tensional strain to determine the rotational proximity to the glarus thrust. a denser and more component of the deformation, Q2i and wm(table 1) uniform coverage would be needed to test this Small extensional strains means short fibres As the interpretation. Grain-size effects might also be fibres get shorter, so does the resolution of the important. Our study focused exclusively on medium- incremental extension path. S, must be greater than grain sandstones, whereas Siddans'(1979) study was 1. 10 to get reliable estimates of Q2i and Wm(ring restricted to Verrucano mudstones where the reduction spots are found. Milnes and Pfiffner(1977) and Brandon, 1999). Our measurements indicate did not report what they sampled, but we suspect that minor non-coaxiality during smt deformation. One they also focused on the Verrucano mudstone sample has Wm=0. 29, but the rest are less than 0.18 because of the availability of reduction sp a9. The rotational component of the deformation may be elatively small, but the internal rotation axes show 4. Discussion consistent orientation and shear sense(Fig 9). The 4.1. Mass loss average rotation axis (Table 2, asterisk in Fig 9)is A surprising result of our study is the large mass loss, horizontal and indicates a general top- north sense of about 36%. in sandstones above and below the glarus shear. which is similar to the shear -sense direction thrust The amount of dissolved mass is large, and determined for the glarus thrust( schmid, 1975 there is no obvious repository for this dissolved mass Milnes and Pfiffner, 1980; Lihou, 1996) in the region around the Glarus thrust. The volume strain can be considered as a form of internal erosion 3.4. Previous regional strain work within the Alpine wedge. SMT deformation included Siddans (1979) presents principal strain ratios both closed and open exchange, involving local measured from reduction spots in mudstones at 12 precipitation of fibre overgrowths and wholesale loss localities in the Verrucano formation milnes and of mass from the rock. The open-system behaviour Pfiffner(1977)report some"average" strain ratios for was probably driven by dissolution and bulk removal
8 there is no evidence of where this missing mass went. The volume fraction of veins in outcrops is generally no greater than a few percent. Thus, we conclude that SMT deformation was influenced by a large flux of fluid that was able to dissolve the sandstones and to transport the dissolved load over a scale larger than our study area. Figure 8 shows that volume strain and deviatoric strain are uncorrelated, which implies that these strains are controlled by different processes. Based on our previous work, we have found that a strong correlation between deviatoric strain and volume strain only occurs where fibre overgrowths are small or absent (e.g. Feehan and Brandon, 1999). The reason is that volume and deviatoric strains are solely a function of shortening strains in Y and Z. Deviatoric strains become uncorrelated with volume strain when there are variations in extensional strain (as indicated by variations in the modal abundance of fibre overgrowth). Uncorrelated volume strain and deviatoric strain are observed in the Eastern Belt of the Franciscan Complex of California (Ring and Brandon, 1999), as well as in our Glarus study here. The variations in fibre overgrowth means that some of the dissolved grain mass is re-precipitated locally in the rock. The rock must extend in at least one direction to be able to accommodate the locally precipitated mass. Only 7 out of the 18 samples have sufficient extensional strain to determine the rotational component of the deformation, Ωi and * Wm (Table 1). Small extensional strains means short fibres. As the fibres get shorter, so does the resolution of the incremental extension path. SX must be greater than ~1.10 to get reliable estimates of Ωi and * Wm (Ring and Brandon, 1999). Our measurements indicate minor non-coaxiality during SMT deformation. One sample has * Wm = 0.29, but the rest are less than 0.18. The rotational component of the deformation may be relatively small, but the internal rotation axes show a consistent orientation and shear sense (Fig. 9). The average rotation axis (Table 2, asterisk in Fig. 9) is horizontal and indicates a general top-north sense of shear, which is similar to the shear-sense direction determined for the Glarus thrust (Schmid, 1975; Milnes and Pfiffner, 1980; Lihou, 1996). 3.4. Previous regional strain work Siddans (1979) presents principal strain ratios measured from reduction spots in mudstones at 12 localities in the Verrucano formation. Milnes and Pfiffner (1977) report some “average” strain ratios for areas adjacent to the Glarus thrust, but the details of their measurements and the localities are not discussed. Neither study provides any information about principal directions. By themselves, strain-ratio data have limited utility, but they do provide information about the symmetry and magnitude of the deviatoric component of the strain. Thus, comparisons with our data are limited to the Nadai plot (Fig. 6). All data sets show a similar clustering along the prolate/oblate boundary, so they share the same strain symmetry. In contrast, they have very different strain magnitudes. In Figure 6, the strain ratio RXZ (= SX/SZ) provides a useful measure of deviatoric strain magnitude (note that RXZ is independent of volume strain). Our measurements have RXZ ranging from 1.4 to 2.3, whereas Siddans’ (1979) measurements range from 1.5 to 9.5. However, the highest strain magnitudes for Siddans’ data are those from localities near the Glarus thrust (grey triangles in Fig. 6b). The remaining localities (black triangles in Fig. 6b) show a closer correspondence to our results. The “averages” given in Milnes and Pfiffner (1977) have an RXZ of ~4 for rocks below the Glarus thrust (open triangle in Fig. 6b) and ~13 for rocks directly above the thrust (open circle in Fig. 6b). These results suggest that differences between these studies is due, at least in part, to difference in sample coverage of a heterogeneous strain field, with the heterogeneity influenced, at least in part, by the proximity to the Glarus thrust. A denser and more uniform coverage would be needed to test this interpretation. Grain-size effects might also be important. Our study focused exclusively on mediumgrain sandstones, whereas Siddans’ (1979) study was restricted to Verrucano mudstones, where the reduction spots are found. Milnes and Pfiffner (1977) did not report what they sampled, but we suspect that they also focused on the Verrucano mudstones because of the availability of reduction spots. 4. Discussion 4.1. Mass loss A surprising result of our study is the large mass loss, about 36%, in sandstones above and below the Glarus thrust. The amount of dissolved mass is large, and there is no obvious repository for this dissolved mass in the region around the Glarus thrust. The volume strain can be considered as a form of internal erosion within the Alpine wedge. SMT deformation included both closed and open exchange, involving local precipitation of fibre overgrowths and wholesale loss of mass from the rock. The open-system behaviour was probably driven by dissolution and bulk removal
low temperatures and low permeabilities thought to typify this setting Burkhard et al. (1992)argues for advection of externally derived fluids, associated with large fluid- 4.2. Regional tectonic evolution to-rock ratios, during motion of the Helvetic thrust There is evidence in the Helvetic Alps that Carbonate-bearing veins in the taveyannaz sandstone deformation propagated northward with time, starting which is otherwise free of carbonate. also indicate in the structurally highest and most internal units to mass transfer over distances greater than the outcrop the south( Sardona and Blattengrat nappes )and scale moving outward to more external and structurally Our results here are similar to those from other deep units to the north. Lihou(1996)related cleavage studies we have done on sandstones in convergent formation in the Sardona and Blattengrat nappes to the wedge settings. These include the late cretaceous San Late Eocene D, Pizol phase and reported a-160- Juan-Cascade nappes(Feehan and Brandon, 1999), the trending extension direction. We concur with Milnes Late Cenozoic Olympic subduction complex( Brandon and Pfiffner(1977)that the Glarus thrust formed and Kang, 1995), the mesozoic franciscan subduction during the subsequent D, Calanda phase and complex(ring and Brandon, 1999; Bolhar and ring 2001), and the Permian-Cretaceous Torlesse o. commodated early transport of the Helvetic nappes nto the Infrahelvetic complex. Cleavage in the subduction complex in the South island of New Verrucano is thought to have formed early during the Zealand(Maxelon et al., 1998). All examples show Calanda phase. The average X direction for the significant mass loss in association with an Verrucano is -200 at that time(Siddans, 1979, this approximately plane-strain coaxial deformation at the study ) We tie cleavage formation in the taveyannaz regional scale sandstone and the north Helvetic Flysch to the late Our study highlights an interesting contrast in the Calanda phase and the D, Ruchi phase(Schmid character of mass transfer with increasing grade. The 1975). The average X direction is -160%at that time mass loss that we have documented in the low-grade and identical to that reported by Schmid(1975)for the sandstones tends to be pervasive throughout the rock D Ruchi phase. Therefore, we suggest that the In contrast, mass transfer in greenschist-and Taveyannaz sandstone and associated North Helvetic mphibolite-facies rocks tends to be more localized Flysch, which lie beneath the Glarus thrust,were occurring in association with veins and ductile shear deformed after the verrucano and melser sandstone zones(e.g. Selverstone et al., 1991; Newman and which lies above the thrust Mitra, 1993; Ague, 1994; Bailey et al, 1994; O Hara Overall. the data suggest that the downward 1994;Ring,1999 propagation of deformation during the Oligocene There has been recurring evidence that smt Calanda and Ruchi phases was associated with a deformation is commonly associated with large mass loss volume strains(e.g, Wright and Platt, 1982 However, the problem with interpreting this changing curving x direction that changed from -200 to -1 Wright and Henderson, 1992). This result, however, X direction is that our strain work indicates that little has been difficult to reconcile with the fact that silicate extension was associated with x above and below the minerals have very low solubilities in typic Glarus thrust. In other words, the regional-scale metamorphic fluids. Thus, large mass-loss volume constrictional plane-strain deformation associated with strains seem to require large fluid-to-rock ratios One D, and D, was produced by shortening in Y and z way out of this dilemma, proposed by Etheridge et al directions. We also note the interpretation of X as a (1983), is that fluid flow in low-grade settings direction of tectonic transport is only valid for a driven by thermal convection, which would allow re simple-shear deformation, which is not supported by circulation of a small volume of fluid. This our strain data. The observation that the smt strains interpretation is supported by theoretical models are, in general, only slightly non-coaxial also discussed by Wood and Hewitt(1982, 1984)and Criss challenges the interpretation that the ductile and Hoffmeister(1991), which show that there is no deformation in the nappe is related to shear coupling threshold for the onset of thermal convection when on the glarus thrust. In contrast, we argue that the isotherms are inclined. Convection will always occur, generally coaxial deformation away from the glarus although the flow velocities may become quite small if thrust indicates that the thrust was very weak. In fact, the thermal gradients are small. We speculate that experimental work by Hsu(1969)and Schmid(1975 porous-medium convective flow may be an under showed that the lochseitenkalk is very weak appreciated process in convergent wedges, despite the Likewise, very low shear coupling on the subduction
9 of the more soluble components of the rock, due to flow of a solvent fluid phase on a regional scale. Burkhard et al. (1992) argues for advection of externally derived fluids, associated with large fluidto-rock ratios, during motion of the Helvetic thrust. Carbonate-bearing veins in the Taveyannaz sandstone, which is otherwise free of carbonate, also indicate mass transfer over distances greater than the outcrop scale. Our results here are similar to those from other studies we have done on sandstones in convergent wedge settings. These include the Late Cretaceous San Juan-Cascade nappes (Feehan and Brandon, 1999), the Late Cenozoic Olympic subduction complex (Brandon and Kang, 1995), the Mesozoic Franciscan subduction complex (Ring and Brandon, 1999; Bolhar and Ring, 2001), and the Permian-Cretaceous Torlesse subduction complex in the South Island of New Zealand (Maxelon et al., 1998). All examples show significant mass loss in association with an approximately plane-strain coaxial deformation at the regional scale. Our study highlights an interesting contrast in the character of mass transfer with increasing grade. The mass loss that we have documented in the low-grade sandstones tends to be pervasive throughout the rock. In contrast, mass transfer in greenschist- and amphibolite-facies rocks tends to be more localized, occurring in association with veins and ductile shear zones (e.g. Selverstone et al., 1991; Newman and Mitra, 1993; Ague, 1994; Bailey et al., 1994; O’Hara, 1994; Ring, 1999). There has been recurring evidence that SMT deformation is commonly associated with large massloss volume strains (e.g., Wright and Platt, 1982; Wright and Henderson, 1992). This result, however, has been difficult to reconcile with the fact that silicate minerals have very low solubilities in typical metamorphic fluids. Thus, large mass-loss volume strains seem to require large fluid-to-rock ratios. One way out of this dilemma, proposed by Etheridge et al. (1983), is that fluid flow in low-grade settings is driven by thermal convection, which would allow recirculation of a small volume of fluid. This interpretation is supported by theoretical models discussed by Wood and Hewitt (1982, 1984) and Criss and Hoffmeister (1991), which show that there is no threshold for the onset of thermal convection when isotherms are inclined. Convection will always occur, although the flow velocities may become quite small if the thermal gradients are small. We speculate that porous-medium convective flow may be an underappreciated process in convergent wedges, despite the low temperatures and low permeabilities thought to typify this setting. 4.2. Regional tectonic evolution There is evidence in the Helvetic Alps that deformation propagated northward with time, starting in the structurally highest and most internal units to the south (Sardona and Blattengrat nappes) and moving outward to more external and structurally deep units to the north. Lihou (1996) related cleavage formation in the Sardona and Blattengrat nappes to the Late Eocene D1 Pizol phase and reported a ~160°- trending extension direction. We concur with Milnes and Pfiffner (1977) that the Glarus thrust formed during the subsequent D2 Calanda phase and accommodated early transport of the Helvetic nappes onto the Infrahelvetic complex. Cleavage in the Verrucano is thought to have formed early during the Calanda phase. The average X direction for the Verrucano is ~200° at that time (Siddans, 1979; this study). We tie cleavage formation in the Taveyannaz sandstone and the North Helvetic Flysch to the late Calanda phase and the D3 Ruchi phase (Schmid, 1975). The average X direction is ~160° at that time and identical to that reported by Schmid (1975) for the D3 Ruchi phase. Therefore, we suggest that the Taveyannaz sandstone and associated North Helvetic Flysch, which lie beneath the Glarus thrust, were deformed after the Verrucano and Melser sandstone, which lies above the thrust. Overall, the data suggest that the downward propagation of deformation during the Oligocene Calanda and Ruchi phases was associated with a curving X direction that changed from ~200° to ~160°. However, the problem with interpreting this changing X direction is that our strain work indicates that little extension was associated with X above and below the Glarus thrust. In other words, the regional-scale constrictional plane-strain deformation associated with D2 and D3 was produced by shortening in Y and Z directions. We also note the interpretation of X as a direction of tectonic transport is only valid for a simple-shear deformation, which is not supported by our strain data. The observation that the SMT strains are, in general, only slightly non-coaxial also challenges the interpretation that the ductile deformation in the nappe is related to shear coupling on the Glarus thrust. In contrast, we argue that the generally coaxial deformation away from the Glarus thrust indicates that the thrust was very weak. In fact, experimental work by Hsü (1969) and Schmid (1975) showed that the Lochseitenkalk is very weak. Likewise, very low shear coupling on the subduction