Click for next page ( 97


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 96
Hydrothennal Systems Associated with Regional Metamorphism and ~, . . 6 ~;rusta~ Anatexis: Pyrenees, France . STEPHEN M. WICKHAM University of Chicago HUGH P. TAYLOR, JR. California Institute of Technology INTRODUCTION Our understanding of the transport of fluids through the deeper parts of the crust, and in particular through rocks ~ . . , undergoing prograde regional metamorphism, is at an embryonic stage. This stems largely from the difficulty of making direct observations of such processes, in contrast to the situation in active shallow hydrothermal systems (which are accessible to study in boreholes). Stable iso- tope (~8O/160, 13C/~2C, and D/H) studies of metamorphic and igneous rocks and minerals may, however, be used to place constraints on the passage of H2O- and CO2-rich fluids through the crust. Oxygen, hydrogen, and carbon are major constituents of both rocks and typical crustal fluids, and these three elements show systematic differ- ences in isotopic composition in the different terrestrial reservoirs. For example, the mantle has a ALSO value of about +6, distinctly different from the range of values shown by most detrital sedimentary rocks (+9 to +18) and much lower than the values in sedimentary carbonates (+20 to +30~. All of these "normal" rock INTO values are distinctly higher than the values in seawater (EGO = 0) or meteoric water (BOO =-25 to 0) (Taylor, 1974, 1977~. Conse- quently, if aqueous fluids derived from the Earth's surface interact with igneous or metamorphic rocks at elevated 96 Examples from the temperatures where the equilibrium isotopic fractionations are small, this results in a change in the rock oxygen (and hydrogen) isotopic composition toward that of the water. Stable isotope analyses of such rocks can help to identify the source of infiltrating fluids, and, in the case of oxygen, which is the dominant constituent of H2O, CO2, and virtu- ally all crustal rocks, these data can be used to make material-balance calculations constraining the quantity of fluid involved. In a number of areas it has been recognized that whole- rock INTO values tend to decrease with increasing meta- morphic grade (e.g., Garlick~and Epstein, 1965; S-hieh and Taylor, 1969; Shieh and Schwarcz, 1974~. For instance, in Idaho the variation is from about +15 in low-grade shales to +11 in sillimanite-grade rocks formed during Creta- ceous metamorphism (Fleck and Criss, 1985~. There is, however, at present no clear consensus of opinion regard- ing the origin of this isotopic shift, its timing, or its rela- tionship to fluid transport during metamorphism. For example, the variation of 6~80 from typical sedimentary values to values closer to typical mantle values could involve minimal amounts of fluid and arise simply by exchange and homogenization of oxygen between high- ~8O metasediments and low-~8O mantle-derived igneous lithologies. Such an effect might be expected at increasing

OCR for page 96
HYDROTHERMAL SYSTEMS ASSOCIATED WITH REGIONAL METAMORPHISM AND CRUSTAL ANA TEXIS depth within the crust, where mantle-derived material becomes progressively more common and sedimentary material more rare. Such exchange conceivably could take place at a relatively low fluid-rock ratio with the fluid merely acting as the agent of isotopic exchange. On the other hand, the shift in 6~80 to lower values at higher metamorphic grade could reflect exchange of oxy- gen between higher-grade metamorphic rocks and low-~8O fluids at high temperatures, with the fluid itself forming the isotonically light oxygen reservoir. In this case large quantities of fluid would be required to bring about bulk changes in INTO by the amounts that are commonly ob- served (3 to 4 per mill. If the latter process is dominant, then permeabilities must be appreciable and fluid-rich conditions are likely to be common during metamorphism (e.g., Ferry, 1986), with fluid pressure at times equaling or locally exceeding lithostatic pressure. If the former inter- pretation is more common, a free fluid phase may only be intermittently and/or locally present. The Hercynian metamorphic terrane exposed in the Pyrenees offers a good opportunity to differentiate be- tween these two alternative scenarios. Here, a characteris- tic shift to lower-~8O values with increasing metamorphic grade is clearly observed, and good exposure, clear-cut geological relationships and access to a wide range of structural levels allow us to place tight constraints on both the nature and scale of fluid-rock interaction during meta- morphism (Wickham end Taylor, 1985, 1987; Wickham, 1987a; Bickle et al., 1988~. France [RO/S SE/G~RS '-~ ~XSS/F ~7 St. Girons ,~,/ If =~n~: 'art Spa i n Mesozoic and Tertiory sediments ~3 Low-grode Paleozoic High-grode mice schists, migmalites, anotectic reroutes ~3 Orthogneisses Basal gneisses ~3 Lote grenodior~tes SOKM , , 97 METAMORPHISM IN THE PYRENEES Hercynian Basement of the Pyrenees The Pyrenees are a roughly linear chain of mountains between France and Spain. They were uplifted during the lower Tertiary in response to convergence between Iberia and Western Europe. Uplift has exposed an extensive pre- Mesozoic basement terrane (the Hercynian basement) comprising Paleozoic and Precambrian sediments, metasediments, and gneisses as well as a variety of grani- toid plutons (see Figure 6.1~. All of these lithologies were metamorphosed or intruded at about 310 to 340 Ma during the Late Carboniferous, Hercynian orogeny (Zwart, 1979; Autran et al., 1980; Bard et al., 1980; Bickle et al., 1988~. Subsequent (e.g., Tertiary) deformation and recrystalliza- tion of the Hercynian metamorphic and granitic rocks were relatively minor. The Hercynian basement ranges from virtually unmeta- morphosed, fossiliferous Devonian and Carboniferous rocks, some of which were actually being deposited while metamorphism was occurring at depth, through Lower Paleozoic mica schists, marbles, and migmatites, finally to amphibolite and granulite-facies "basal gneisses" (Zwart, 1979~. The mineral assemblages in the gneisses formed at pressures of 4 to 7 kbar (Vielzeuf, 1984), compatible with their being the basement upon which the Paleozoic sedi- mentary sequence was deposited (Vitrac-Michard and Allegre, 1975~. The metamorphic sequences are charac FIGURE 6.1 Hercynian basement outcrop in the Pyrenees, showing the various lo- calities discussed in the text.

OCR for page 96
98 terized by very abrupt transitions from the low- to the high-grade regions and by extensive partial melting of Lower Paleozoic politic metasediments at temperatures of about 700C and depths of only about 10 to 12 km. This is recorded by abundant politic migmatites and peraluminous granitoids associated with the high-grade mica schists. Subdivision of the Hercynian Crust into Three Structural Levels Taken together the wide range of structural levels ex- posed in the different tectonic blocks of the Hercynian basement of the Pyrenees can be interpreted in terms of a composite section through the upper 25 km of the conti- nental crust. Because the oxygen isotope systematics are very distinctive at different structural levels within this section, it is convenient to divide it into three tectonostra- tigraphic zones: (1) a moderately deformed, fossiliferous sedimentary sequence that was deposited from the Upper Ordovician to the Upper Carboniferous; (2) a metamor- phic and migmatitic sequence developed within Cambro- Ordovician (mainly politic) metasediments, with equili- bration pressures of 2 to 4 kbar, that were locally exten- sively melted and intruded by peraluminous leucogranites and biotite-cordierite granites; and (3) a region of amphi- bolite- and granulite-facies orthogneisses and paragneisses equilibrated at 4 to 7 kbar. Although a continuous section through all of these crustal levels is not exposed in the Pyrenees, rocks from all three levels outcrop within several relatively restricted geographic areas (e.g., in the Agly and St. Barthelemy massifs). Figure 6.2 shows a section through the St. Barthelemy Massif, indicating the relationship between Zones 1, 2, and 3 in that region. FIGURE 6.2 Schematic ~ue-scale section through the St. Barthelemy Massif redrawn from Passchier (1984, Figure 3~. The line of section is approximately north to south (from low to high grade) and is about 15 km long. The Zone 3 "basal gneisses" have uncertain structural relationship to the other rocks in the massif because they are bounded by a mylonitic shear zone that approximately separates Zone 2 and Zone 3. STEPHEN M. WICKHAM AND HUGH P. TAYLOR, JR. Metamorphism and Anatexis The very abrupt, albeit progressive and gradational, transition from low to high grade in the metamorphic sequences in the Pyrenees implies the existence of very steep thermal gradients during metamorphism. Typically, a progression in the Zone 2 politic rocks (which comprise over 90 percent of the Lower Paleozoic metasediments) from chlorite-sericite phyllites through andalusite an :l sil- limanite schists to migmatites and biotite-cordierite grani- toids occurs over horizontal distances of only 3 to 5 km (e.g., in the Trois Seigneurs Massif, Figure 6.3~. Thin (10 to 50 m) carbonate-rich beds are interlayered within the Cambro-Ordovician pelites; these have developed various calc-silicate mineralogies consistent with the metamor- phic mineral assemblages in the adjacent pelites. The granitic rocks were mostly derived by partial melt- ing and homogenization of the high-grade politic metasedi- ments, and in general the granitic melts have not migrated very far from their region of generation (Wickham, 1987a). Migmatite textures and compositions imply that pelites melted to at least 50 to 60 percent by volume, and the observed rapid increase to these high degrees of melting at relatively low temperatures of 700 to 750C strongly implies water-rich conditions with aH2O buffered exter- nally to high values (see Wickham, 1987a). This is also suggested by common pegmatitic textures, widespread muscovitization and tourmalinization, and by calc-silicate mineral assemblages in the marble layers (they commonly contain clinozoisite and sometimes also grossular; see Ferry, 1983~. In contrast to the Zone 2 rocks, much smaller degrees of melting occurred within the amphibolite- and granulite- facies "basal gneisses" (Zone 3~. These rocks are litho- logically distinct from Zone 2 in that they contain far less ZONE 3 - ~ZONE 2 ~ -'a ZONE 1 (4-7kbar) ~(2-4 fiber) , (<2kbar) i; ! (2-4 kbar) mylonite zne~l~minmatites ~mien Hitch ~choler ASH_ 1 omphibolite and granulite ~ . . racier gne~sses

OCR for page 96
HYDROTHERMAL SYSTEMS ASSOCIATED WITH REGIONAL METAMORPHISM AND CRUSTAL ANA TEXIS SW I ,* 2000 E 1000 Carbonate //, \ ~ it'' +'%,~~E,~ Florid I * ~x ~ / I ~ Chlorite sericite phyllit" Carbonate E23 Andalusite schist Id l Andalusitc + sillimanite schist _~ Biotite - sillimanite schist Betide - sillimarute gneiss E;2l Biotite granite - quartz diorite {i:3 Leucogranite migmatite cone petite and larger amounts of granitic orthogneiss, quartzo- feldspathic biotite gneiss, and amphibolite. Zone 3 is locally migmatitic but contains virtually no large (>100 to 1000 m) bodies of peraluminous granite. Vielzeuf (1984) made a detailed study of the petrology of these rocks, which are commonly orthopyroxene bearing and clearly equilibrated at greater depth (4 to 7 kbar) and under much "dryer" peak metamorphic conditions than those in Zone 2. Although no H2O activity data are available from this work, some of the granulites probably equilibrated at aH2O > 0.5. This does not in itself preclude the possibility that more water-rich conditions prevailed at an earlier stage during the metamorphism. However, as described below, the stable isotope systematics within the basal gneiss Zone 3 rocks are significantly different from those in the mica schist-migmatite sequences (Zone 2~. This almost cer- tainly reflects contrasting fluid-rock interactions during metamorphism at these different levels within the Her- cynian crust. OXYGEN ISOTOPE SYSTEMATICS Zone 1 The Paleozoic rocks from Zone 1 comprise fossilifer- ous shales and carbonates that have been deformed but only very weakly metamorphosed. Rocks are typically fine "rained and may contain chlorite or fine-grained white mica but no higher-grade metamorphic minerals. Because they probably never experienced temperatures in excess of 400C during the Hercynian metamorphism, it is hardly surprising that these rocks in the main preserve typical 99 FIGURE 6.3 Schematic true-scale section through the Trois Seigneurs metamorphic sequence, indicating the progressive gra- dation from low-grade phyllites through high-grade mica schists and migmatites to the granitoids that were derived mainly through anatexis of the politic metasedi- ments. Note the proximity of the low- and high-grade rocks. sedimentary INTO values. However, recent 87Sr/86Sr data (Bickle et al., 1988) indicate that even these low-grade phyllites and shales were extensively modified by fluid infiltration during prograde metamorphism. 87Sr/86Sr ratios were homogenized (and in some cases lowered) to values of ~0.715 at 310 Ma, a process that is observed in even the lowest-grade Zone 1 shales, which could not have experi- enced temperatures any higher than 250 to 300C (see Table 6.1~. i80/~60 ratios may have been modified at the same time, but we have no way of detecting this because we cannot determine the depositional INTO values of these sediments (as we can with 87Sr/86Sr). The Sr data allow us to infer that significant isotopic modifications resulting from fluid infiltration commenced at the earliest stages of prograde metamorphism and continued up to peak meta- morphic temperatures (see Bickle et al., 1988, and below). The ~8o/~6o data are illustrated in Figure 6.4, where a compilation of INTO values for Zone 1 (calcite for the carbonate rocks and whole rock for shale samples) is pre- sented from four different parts of the Pyrenees. In the Trois Seigneurs and Agly massifs these rocks are Ordovi- cian and Silurian shales, whereas in the St. Barthelemy and Arize massifs they are Ordovician to Carboniferous shales and limestones. Calcite from the limestones ranges from +19 to +24, consistent with the typical range of isotopic values shown by many Paleozoic sedimentary carbonates (Baertschi, 1957; Veizer and Hoers, 1976~. Shales range from +13 to +16, with most values between +14 and +15. These kinds of values are also shown by all other Zone 1 Ordovician to Carboniferous sedimentary rocks that we have analyzed from other parts of the Pyre- nees.

OCR for page 96
100 STEPHEN M. WICKHAM AND HUGH P. TAYLOR, JR. TABLE 6.1 Oxygen and Strontium Isotope Data (average values) at Various Outcrop Localities of Pelitic Schist, Phyllite, and Shale in the Trois Seigneurs Massif, Arranged into Four Groups as a Function of Grade of Metamorphism Original Sedimentary Intermediate Grade High grade Parameter Rock (model values Low Grade (shales (vicinity of the (andalusite Studied 450 Ma)a and phyllites) biotite isograd) sillimanite zones) Mean 6~80 value Range of 6~80 Mean 87Sr/86Sr value Range of 87Sr/36Sr Average Sr content Average Rb content Average Rb/Sr ratio ? 0.7202 + 0.0053 (5) 0.7133 to 0.7326 89 ppm (5) 138 ppm (5) 1.55 +14.5 + 0.7 (8) +13.3 to +16.0 0.7130 + 0.0022 (5) 0.7089 to 0.7166 89 ppm (5) 138 ppm (5) 1.55 +12.1 + 0.9 (5) +10.5 to +13.3 0.7153 (1) 102 ppm (1) 98 ppm (l) 0.96 +11.5 + 0.6 (16) +10.6b to +12.7 0.7151 +0.0016 (6) 0.7129 to 0.7178 98 ppm (7) 160 ppm (7) 1.62 Note: Data are from Bickle et al. (1988) and Wickham and Taylor (1985). The tabulated 87Sr/86Sr, Sr, and Rb data represent the mean values for the various localities, using a single average value for each of these outcrop localities where multiple samples were studied. The + indicates average deviation from the mean value, with the number of analyzed localities given in parentheses. Al187Sr/86Sr values are calculated at 310 Ma except for the original sedimentary rock (model) values, which are calculated for 450 Ma. aThe 6~80 values are unknown, but the model strontium isotope values at 450 Ma can be calculated, as discussed in the text. bOne sample with an anomalously low-0 value of +8.8 is not included (Wickham and Taylor, 1985). Zone 2 In contrast to the characteristic sedimentary INTO values shown by the shales and carbonates of Zone 1, the Zone 2 mica schists, migmatites, and peraluminous granitoids have lower and much more uniform GINO values, mostly in the range +11 to +12 (see Table 6.1 and Figure 6.4~. This is particularly clear in the Trois Seigneurs Massif, where the Zone 1 to Zone 2 transition is well exposed and where the Zone 2 carbonates have internally homogeneous GINO values that are essentially identical to the INTO of the adjacent mica schists (Wickham and Taylor, 1985~. A detailed isotopic profile through one of the Trois Seigneurs meta- carbonate layers is shown in Figure 6.5. This unit lies between the "sillimanite in" isograd and the migmatite zone and is about 15 m thick, sandwiched between psam- mitic rocks with INTO values of about +12. The calcite has a fairly uniform GINO of between +13 and +14 throughout the layer, regardless of the calcite content of the sample or the distance of the sample from the margin of the layer. This indicates an extremely high degree of oxygen iso- topic equilibration between this layer and the surrounding metasediments, in contrast to the situation prior to meta- morphism when the carbonate layers would have had much higher GINO values than the surrounding rocks (as is still the case in Zone 1~. This homogeneity of carbonate and petite oxygen isotopic compositions within the Trois Seigneurs Zone 2 rocks is typical of the oxygen isotope systematics at this structural level throughout the Pyre- nees. Preexisting sedimentary heterogeneities in GINO have everywhere been smoothed out by pervasive oxygen iso topic exchange over wide regions. The Hercynian age of this process is proved by Rb-Sr isochrons obtained from suites of mica-schist samples (Bickle et al., 1988) from Zone 2, indicating homogenization of 87Sr/86Sr at values of ~0.715 at 310 Ma. Interestingly, in the case of Sr the 87Sr/ 86Sr ratios are not smoothed out everywhere; this is proba- bly because the giant reservoir of strontium in the carbon- ates (containing ~2000 ppm Sr) is much more resistant to change than the petite reservoir (~200 ppm Sr) and thus tends to retain its original sedimentary value of~0.708 (Bickle et al., 1988~. In addition to the whole-rock GINO values being much more homogeneous in Zone 2 than in Zone 1, there is an obvious, pronounced lowering of 6~80 going from Zone 1 to Zone 2 (Figure 6.4; Table 6.1~. This shift reflects a change in the bulk oxygen isotopic composition of the Paleozoic sedimentary pile of 3 to 4 per mil, from about +15 to about +11.5 (similar to that observed in some other regional metamorphic sequences (e.g., Garlick and Ep- stein, 1967; Rye et al., 1974; Fleck and Criss, 1985). This shift is particularly well characterized in the Hercynian of the Pyrenees, where it takes place fairly abruptly over a stratigraphic distance of 1 or 2 km within the metamorphic sequence. It is clear that the Zone 1 and Zone 2 politic rocks both belong to the same Paleozoic sedimentary succession and that they originally had similar bulk isotopic and chemical compositions. This implies that the observed shift in GINO values does not represent any lithological difference be- tween Zone 1 and Zone 2. During metamorphism these rocks were partially dehydrated and decarbonated, con

OCR for page 96
HYDROTHERMAL SYSTEMS ASSOCIATED WITH REGIONAL METAMORPHISM AND CRUSTAL ANA TEXIS versed to high-grade mica schists and marbles, were iso- topically homogenized, and overall had their bulk 6~80 values lowered by about 3 to 4 per milt The Zone 2 metamorphic rocks clearly must have exchanged oxygen with a large reservoir that had a 6~80 value significantly lighter than +1 1. Zone 3 Exposures of Zone 3 rocks in the Pyrenees are limited and principally occur in three regions: the Agly, St. Barthe- lemy, and Castillon massifs (see Figure 6.1~. Addition- ally, a small area near the village of Lapege in the Trois Seigneurs Massif exposes rocks that probably correlate with the basal gneisses elsewhere because they are litho- logically distinct from any other rocks in the Trois Seigneurs area (which clearly belong to Zone 2; see Wickham, 1987a). Zone 3 rocks comprise a lithologically heterogeneous sequence of granitic gneisses, quartzo-feldspathic biotite gneisses, amphibolites, carbonates, and relatively rare +24 _ +2` +~E ~0 ZONE 1 I ZONE 2 l O carbonates pastes by\ _ ~St. Barthelemy .~ ~ ~--e'0: +10 _ +8 Trois Saigneurs PALEOZOIC SEDIMENTS METAMORPHOSED PALEOZOIC (SHALES AND CARBONATES) (MICA SCHISTS AND MARBLES) FIGURE 6.4 Compilation of data for Zones 1 and 2 from the Agly, Arize, St. Barthelemy, and Trois Seigneurs massifs for whole-rock petite samples (solid circles) and calcite from car- bonate-bearing samples (open diamonds). The data are plotted as a function of metamorphic grade with their position on the diagram corresponding roughly to their relative distance from the Zone 1 to Zone 2 boundary in the field. All of the Zone 2 carbonate samples are from Trois Seigneurs, and the two data points with error bars represent the average values (and the 6~80 range) for the two detailed profiles in Figures 6 and 7 of Wick- ham and Taylor (1987~. There is a pronounced shift in SILO to lower values going from Zone 1 to Zone 2, and the Zone 2 values are much more homogeneous than those in Zone 1 (also see Table 6.1~. 101 TROI S SEIGNEURS / ~0(1%) ~ CARBONATE 12 _ - - ~v 10 - 8 He ~ 6 In 0~ (61%) 0 ~0 (31 %) o~hO(72%) 0~0 (77%) / +10 +12 +14 +16 +18 8180 o FIGURE 6.5 Oxygen isotopic profile through a carbonate unit in the Trois Seigneurs Massif (sample SI13, Wickham and Taylor, 1985) that crops out between the "sillimanite in" and "andalusite out" isograds (see Figure 6.3~. Note that the INTO of calcite is fairly uniform throughout the layer, regardless of either the dis- tance from the margin or the calcite content of the sample. The INTO values within the carbonate layer are similar to the psam- mites to either side, and they are much lower than the original sedimentary values, which must have been at least as high as +22 to +25. kinzigites (politic gneisses). Oxygen-isotope data from these regions are summarized in Figures 6.6 and 6.7 (see also Wickham and Taylor, 1987~. These data are not as clear cut as those obtained from Zones 1 and 2, mainly because we do not know for certain what the original premetamorphic INTO values of the Zone 3 lithologies were; we simply do not have access to the unmetamorphosed equivalents of these rocks (as we do in Zone 2~. The data in Figures 6.6 and 6.7 show that major lower- ing of the 6~80 values in carbonates occurred both at St. Barthelemy and at Castillon and that this was also accom- panied by some isotopic homogenization. However, the data from Agly and Lapege are rather different. Here the INTO values of the carbonates span a wide range from +13 to +22, including both values typical for sedimentary car- bonates (as in Zone 1) and values characteristic of the homogenized and ~8O-shifted marbles of Zone 2. In order to investigate these systematics further, we made detailed isotopic profiles across some of the individ- ual carbonate units in the Zone 3 basal gneisses. One such profile from the Agly Massif is shown in Figure 6.8. The

OCR for page 96
1kJ~ -1 +22 +2O +18 +16 ~ 180 +14 +8 +~ +24 ZONE 1 | ZONE 2 | ZONE 3 _~\ _~ 1~! \ ~ ~ it, ~1 /~' , 1 ~\A xx1 +2 ; w ~ G {,, ~ ~0~ /~ Ad/ v / FIGURE 6.6 Summary of the /0 data for metacarbonate lithologies from the Pyrenees (excluding the Anze data shown in Figure 6.4~. The calculated whole-rock carbonate 6~80 values (i.e., including the coexisting silicate minerals) are plotted as a function of metamorphic grade, with the relative width of each lithological unit along the abscissa being proportional to the total areal extent of each Ethology in the Castillon (Cast.), St. Barthel- emy (St. B.), Trois Seigneurs (T.S.), and Agly massifs. Within a given lithological subdivision, the data points are plotted propor- tional to the actual geographic distance of the sample locality from the boundary of this subdivision in the field. Silicates were not analyzed for two Castillon samples and one low-grade St. Barthelemy sample (marked with plus signs); these 6~80 values represent calcite rather than whole rock. In the Trois Seigneurs field two of the data points with error bars represent average values (and the 6~80 range) for the two profiles shown in Figures 6 and 7 of Wickham and Taylor (1987~. The field of whole-rock 6~80 values of pelites, granites, and gncisses from all regions (Figure 6.7) is shown for companson. Note the preservation of relatively high (sedimentary) 6~80 values in Zone 1, the ~80- depleted and homogeneous isotopic compositions of both the carbonates and the silicate lithologies in Zone 2, and the wide range of 6~80 values in all rock types in Zone 3. calcites in these carbonate layers have essentially retained their original sedimentary 6~80 values, indicating a lack of exchange with the adjacent biotite gneisses (which have 6~80 values of about +121. Steep isotopic gradients of as much as 10 per mil over 50 cm are preserved at the mar- gins of the carbonates. These systematics occur despite the fact that the Agly units are mostly substantially thinner than those at Trois Seigneurs and were metamorphosed at higher temperatures (both of which factors might be ex STEPHEN M. WICKHAM AND HUGH P. TAYLOR, JR. pected to favor isotopic homogenization). Similar sys- tematics are observed in the Lapege samples. Again, steep isotopic gradients, internal isotopic heterogeneity, and lack of homogenization with the adjacent lithologies character- ize the carbonate unit sampled here (see Wickham and Taylor, 1987; Figure 6.8~. This clearly implies a different style of fluid-rock interaction in the Agly and Lapege Zone 3 rocks, as compared with those in Zone 2 in the same areas. In this respect it is important to note that there is no obvious contrast in the present-day permeabilities of Zone 2 and Zone 3 rocks, although major contrasts may have existed between these lithologies at the time of active Hercynian metamorphism. At St. Barthelemy and Castillon, the oxygen isotope systematics in the Zone 3 basal gneisses have much more in common with the higher-level Zone 2 rocks. Metacar- bonates from these regions have 6~80 values similar to those for the Zone 2 rocks from Trois Seigneurs, and the steep isotopic gradients that are a characteristic feature at +22 8'~C . ZONE I Zn~F ~I ZONF ~ , At, +8 _ +6 _ +429 I I ~ ~1 ) ~~v + ~O FIGURE 6.7 Summary of the whole-rock /0 data for pe- lites, granites, and gneisses plotted as a function of metamorphic grade (as described in the caption for Figure 6.6~. The data points lying above the heavy black bar on the abscissa represent values for mafic rocks within the basal gneisses. The field for the whole-rock carbonate samples plotted in Figure 6.6 is also repeated in this figure for reference. Note the strong lowering of 6'80 in all lithologies (including the carbonates) going from Zone 1 to Zone 2. The homogeneous 6'80 values in Zone 2 reflect the infiltration of large volumes of aqueous fluid into all lithologies at high temperatures. The 6'80 values are more heterogeneous in Zone 3, indicating that the Zone 3 rocks did not experience such a massive water influx.

OCR for page 96
HYDROTHERMAL SYSTEMS ASSOCIATED WITH REGIONAL METAMORPHISM AND CRUSTAL ANA TEXIS AG LY 12 10 8 2 _ _ x Gneiss whole rocl' O Silicate residue O Colcile ~ Calculoted carbonate shore rock /~///X~//I OT ~ T E ///// o DO (70%) - CARBONATE ~ - A~5%} //////////. ~ ~(91 To) CARBONATE O 0( O(84~) L~//~//~/////~/~/: +16 tt8 +20 +22 81So Agly and Lapege have not yet been identified. Some isotopic heterogeneity does occur between the rare mafic rocks in these areas and the more abundant biotite gneisses, but it is not as marked as that observed between the car bonates and gneisses at Agly. In general, isotopic heterogeneity within the basal gneisses is more extreme than is observed anywhere in Zone 2, suggesting that the Zone 3 rocks throughout the Pyrenees were not subjected to the same isotopic homog enization process that occurred at higher structural levels. This increase in isotopic heterogeneity with increasing structural depth in the Hercynian crust strongly implies that fluid movement during metamorphism became less important at deeper structural levels. Interpretation of Zone I and Zone 2 Oxygen Isotope Systematics There are essentially two plausible low-~8O reservoirs available to account for the lowering of INTO in going from Zone 1 to Zone 2. One is mantle-derived lower crustal rocks, which would probably have had 6'8O values be tween +6 and +10. The other is a large volume of low-'8O aqueous fluid. Geological evidence in the Pyrenees favors the latter reservoir because suitable low-'8O rocks are notably absent from the wide range of structural levels exposed. Basal gneisses in Zone 3 contain small amounts of mafic rock with 6'8O values of +6 to +8, but are domi 103 FIGURE 6.8 Oxygen isotopic profile through two thin carbonate layers within granulite-facies quartzo-feldspathic gneiss in the Agly Massif (see Figure 2 of Wick- ham and Taylor, 1987, for locality). The weight percent calcite in each carbonate sample is shown. Despite the small di- mensions of the carbonate layers (<3 m thick), most of the calcites have essen- tially retained their original sedimentary INTO values, indicating a lack of /0 exchange with the adjacent gneisses. The original INTO values of these gneisses are not known, but they could have been nearly identical to their present metamorphic values (see text). nantly composed of metasedimentary biotite gneisses with 6'8O values of +11 to +12. Late granodiorite plutons have values typically in the range +8 to +11, but these were intruded after Hercynian metamorphism and were not available to take part in the earlier isotopic homogeniza- tion processes. The synmetamorphic peraluminous grani- toids were themselves mainly derived from politic metasedi- mentary material (Wickham, 1987a) and have average 6'8O values of +11 to +12. A suitable low-'8O rock reservoir is therefore lacking at any exposed level of the Hercynian crust. Furthermore, the isotopic data from the Zone 3 rocks at Lapege and Agly imply an increase in the degree of isotopic heterogeneity with increasing structural depth. This is opposite to what would be expected if the Paleo- zoic metasediments were being isotonically homogenized by some large-scale process involving interactions with low-'8O materials from the mantle or the lower crust. If the low-'8O reservoir were an aqueous fluid of some type, then simple material-balance calculations could be used to place constraints on the quantity of fluid involved. Clearly, this depends on the original 6'8O value of the fluid and the temperature of the isotopic exchange as well as the magnitude of the isotopic shift observed in the rocks. This is illustrated in Figure 6.9, where water-rock ratio is plotted against d'8O, the initial 6'8O value of infil- trating fluid (see Taylor, 1977, and Wickham and Taylor, 1985, for details), assuming a shift in 6'8O of 3.5 per mil between Zone 1 and Zone 2 for the bulk terrane. Exchange

OCR for page 96
104 is assumed to occur at a sufficiently high temperature for the final INTO values of rock and fluid to be approximately equal. Clearly the lighter the fluid, the smaller is the quantity required to generate the observed change in INTO. If the water were magmatic or metamorphic in ongin, it is unlikely to have had a INTO value of less than +6. From Figure 6.9 it is clear that at this value of 6~8Oi, the mate- r~al-balance, water-rock ratio is about 0.6, equivalent to about 35 percent by weight of the entire isotonically al- tered rock mass. Such a large quantity of water would need to be derived from plutons many times bigger than the Zone 2 homogenized terrane, and all of the water released would need to be channeled through this region. Metamorphic water from dehydrating minerals would also have to be derived from a very much larger region and effectively focused through the isotonically homogenized region. Either scenario seems implausible and in the latter case implies that large surrounding regions remain isotopi- cally unaffected. We have yet to identify such regions and consider that a magmatic or metamorphic source for the bulk of the water is unlikely, simply because it seems quantitatively inadequate to explain the Zone 1 to Zone 2 isotopic shifts. The only other plausible water sources are (1) connate formation waters that were trapped in the Paleozoic sedi- ments during deposition and subsequently mobilized dur- ing metamorphism or (2) water derived directly from the surface of the Earth. In either case the amounts of water available are far larger and are thus volumetrically easily capable of producing the observed isotopic shifts at all reasonable values of 6~8Oi. Formation waters might typi- cally have INTO values in the range +2 to +8 (Clayton et al., 1966) and could be drawn in from large regions out- side the isotonically homogenized zones. Surface waters represented by seawater (~80 ~ 0) and meteoric water (~80 > 0) for ~ effectively limitless source, but they would have to move from the surface to depths of 10 km or more in order to flow through the Zone 2 rocks during metamorphism. In summary, the oxygen isotope data for the Zone 1 and Zone 2 rocks imply that pervasive isotopic exchange and homogenization were accompanied by a towering Of 6~80 by several per mil in Zone 2 rocks during metamorphism. The oxygen data do not unambiguously identify the source of this fluid, but simple material-balance constraints imply that huge amounts of H2O are required and that Paleozoic formation waters or surface waters were the most likely source. Interpretation of Zone 3 Oxygen Isotope Data The isotopic data from Zone 3 are more difficult to interpret than the data from Zones 1 and 2, because the STEPHEN M. WICKHAM AND HUGH P. TAYLOR, JR. .4 1 1 1 1 -i 1 .80 _ 1.d 1.0 _ o.z CLOSED / SYSTEM ~ / / ~ OPEN / / SYSTEM , 1 ~ ~ , , , -4 -2 o +2 +4 +6 +8 +10 8180 infiltrating water In In _ 70 ~ Y OCR for page 96
HYDROTHERMAL SYSTEMS ASSOCIATED WITH REGIONAL METAMORPHISM AND CRUSTAL ANA TEXIS gneisses, contrary to the Zone 2 mica schists. The isotopic composition of the gneisses may simply reflect original sedimentary values that have remained essentially un- changed during metamorphism. Mafic rocks occur sporadically throughout Zone 3 and preserve INTO values of typically +6 to +8, not far from their presumed original mantle values. This is true even at Castillon and St. Barthelemy, where the metacarbonates have HO values similar to those of the biotite gneisses. Steep isotopic gradients are preserved at the margins of these mafic layers. Thus, although the nature of the iso- topic heterogeneity is different at St. Barthelemy and Castillon in that it does not occur between metacarbonate and adjacent units, it is still more pronounced than any- where we have sampled in Zone 2. Although we cannot accurately calculate water-rock ratios in Zone 3 because we do not know the original HO values of the rocks prior to Hercynian metamorphism, it seems certain that infiltra- tion has been much less extreme at this deeper level in the crust. This is especially true at Agly and Lapege, where we see extreme oxygen isotope heterogeneity preserved in and adjacent to the metacarbonates. The ]80/~60 data strongly support the idea that during metamorphism water was derived from the overlying Pa- leozoic supracrustal sedimentary pile or from the surface, because in this case a clear maximum depth of penetration would be expected, possibly due to the expected decrease in permeability with increasing structural depth in the crust. If the water were derived from a deep plutonic source or if the homogenization reflected bulk HO ex- change with the deep crust during metamorphism, there should certainly not be an increase in the degree of iso- topic heterogeneity at deeper structural levels, as is ob- served in the Pyrenees. Although the Castillon and St. Barthelemy data require some infiltration of the carbonate lithologies by H2O- (and CO2-) bearing fluids during metamorphism, the Zone 3 rocks certainly have not expe- rienced the pervasive flushing at material-balance water- rock ratios in excess of 0.5 that is required by the Zone 2 data at Trois Seigneurs. HYDROGEN ISOTOPE SYSTEMATICS Material-balance calculations and other considerations described above imply that the large volumes of aqueous fluid that infiltrated the Zone 2 rocks of the Pyrenees probably were originally derived from the surface of the Earth. To test this hypothesis further, we made D/H analy- ses on samples of muscovite from the metamorphic and granitic rocks of Zone 2 at Trois Seigneurs and elsewhere in the Pyrenees. The results are summarized in Figure 6.10, where these data are plotted as a function of meta- morphic grade. 105 Most igneous and metamorphic rocks have a relatively restricted range of SD values, typically in the range of-50 to -85 per mil (e.g., Taylor, 1974~. This range is distinct from the isotopic composition of present-day seawater (6D ~ 0) and many meteoric waters, particularly those found at high latitudes and high altitudes (6D = -100 to - 00), so that any interaction at high temperatures be- tween such surface waters and "normal" igneous and metamorphic rocks should result in a modification of rock ED values toward the characteristic meteoric or marine values. Because there is proportionally so much more hydrogen in water than in rocks, rock ED values can be changed significantly even at very low water-rock ratios involving only tiny amounts of infiltrating fluid. There- fore, under favorable conditions, D/H measurements can be very sensitive monitors of the involvement of surface waters with geological systems. The muscovite data of Figure 6.10 strongly support the notion that marine fluids interacted with the Zone 2 O Trois Seigneurs Muscovite _ in _ _ _ -40 _ ED -so -60 -7n _ 1 Canigou Aston Borouss:_~ Bosost Lys C. MOST IGNEOUS AND METAMORPHIC MlJC~nVITF ~- CARBONIFEROUS ~ . , OCEAN WATER __ ~ ~ , ~ ~_CALCULATED WATER, 450C -80 _ -- I'- _ _90 ~ ~/ `~' FIGURE 6.10 ED for muscovite from a variety of Zone 2 lithologies within the metamorphic sequence (leucogranites, quartz-muscovite veins, and boudin-neck fillings) plotted as a function of metamorphic grade. Trois Seigneurs samples are shown as open circles. The other data (shown as crosses) cover a wide area within the Pyrenees, ranging from the Bosost region in the west to the Canigou Massif, about 180 km to the east (see Figure 6.1~. The data are very uniform throughout this wide geographic range and have exceptionally high SD values, dis- tinctly higher than the typical range shown by most igneous and metamorphic muscovite. The calculated field of waters with which these muscovites would have been in equilibrium is shown and corresponds closely to the likely ED value of Carboniferous ocean water.

OCR for page 96
106 metasediments at high temperatures. All of the ED values are between - 0 and -25, which is distinctly heavier than the typical range shown by most muscovites from meta- morphic and igneous rocks. The values from the Trois Seigneurs Massif (see also Wickham and Tavlor 19RS- Figure 6.11) are particularly heavy (five values lie be- tween -32 and -25 per mil) and uniform over a wide range of structural levels and a variety of lithologies. Interest- ingly, the sample from the deepest structural level has the lightest (most normal) ED value and may possibly reflect the waning effects of surface water infiltration at deeper structural levels (similar to the downward increase in oxygen isotope inhomogeneity observed in Zone 3 at Agly and Trois Seigneurs). Figure 6.10 also shows a calculated field of waters that would be in equilibrium with these muscovite samples at plausible metamorphic temperatures (450C). This field mostly lies between 0 and -10 per mil and overlaps with the hypothetical field of Late Paleozoic (Hercynian) ocean water. In the past, ocean water may have fluctuated be- tween the present SD value (0 per mill and perhaps -15 per mil when there were no ice caps. Because no major late Carboniferous glaciation has been recognized, sea- water was probably closer to this latter value during the Hercynian. The D/H data of Figure 6.10 are thus thor- oughly consistent with the proposal that the Trois Seigneurs fluids were ultimately derived from late Paleozoic sea- water. Note that by the time this fluid infiltrated Zone 2, it still would have retained its original ED of about-5 to -10, but it almost certainly would not have retained its original INTO value of about 0, because it would have undergone extensive exchange with Paleozoic sedimen- tary rocks enroute. Thus, in this particular case the D/H data are much more informative about the original source of the H2O than are the ~8O/~60 data, although they of course say almost nothing about the amounts of H2O in- volved. High ED values have also been observed in Hercynian rocks elsewhere in Europe, for example, in Cornwall (Sheppard, 1977), in the Alps (Frey et al., 1976; Negga et al., 1986) and in the Iberian pyrite belt (Munha et al., 1986~. These data have generally been interpreted to represent the involvement of connate waters of some type, although not in terms of direct infiltration by marine sur- face waters. The values obtained from the Pyrenees indi- cate equilibration of Zone 2 metasediments with marine fluids over a wide region, though they do not tell us if the water was derived from connate fluids within the Paleo- zoic supracrustal metasediments or whether it had to be derived directly from the surface (see below). Similarly, the D/H data do not in themselves constrain the timing of the infiltration process. However, bearing in mind the homogeneity of ED and 6~80 both between and within STEPHEN M. WICKHAM AND HUGH P. TAYLOR, JR. individual terranes, it seems likely that both were homoge- nized during metamorphism more or less simultaneously by the same fluid infiltration event. STRONTIUM ISOTOPE SYSTEMATICS To constrain further the timing and scale of the hy- drothermal process that have affected the metasediments at Trois Seigneurs and elsewhere in the Pyrenees, Rb-Sr isotope analyses were made on a number of samples from the metamorphic sequence at Trois Seigneurs (Bickle et al., 1988~. These data are illustrated in Figure 6.1 1, where the initial 87Sr/86Sr ratios 310 Ma are plotted against PESO. The data from both the low- and high-grade metasedi- ments lie on crude Hercynian-age isochrons, indicating that, unlike the ~8o/~6o ratios, 87Sr/86Sr was homogenized in both the amphibolite facies mica-schists and in their low-grade protoliths, the Lower Paleozoic shales and phyllites, approximately 310 Ma. Because this 87Sr/86Sr homogenization is evident in even the lowest-grade shales (although not so extreme as in the higher-grade samples), it implies that hydrothermal activity commenced at the very lowest grades of regional metamorphism (<200C?) and continued up to temperatures in excess of 600C (see Bickle et al., 1988, for detailed discussion). The Trois Seigneurs politic schists occupy a much more homogeneous (87Sr/86Sr)3~0 field (0.713 to 0.718) than they would be expected to occupy if they had evolved as closed systems with their present-day 87Rb/86Sr ratios from Cam- brian sediments. The model values that they would have had 310 Ma are also plotted on Figure 6.11; these data imply that not only have these values been substantially homogenized, but also that some of the samples have undergone a bulk lowering of 87Sr/86Sr. This lowering must have occurred in response to mixing with relatively unradiogenic strontium (87Sr/86Sr < 0.715) during prograde Hercynian metamorphism (see Table 6.1~. A possible source for the low 87Sr/86Sr might be the metacarbonate layers that are present throughout the Lower Paleozoic sequences and that typically have fixed 87Sr/86Sr ratios of 0.708. However, detailed profiles measured through the metacarbonates at Trois Seigneurs (Bickle et al., unpublished data) show that there has been incomplete homogenization of Sr within the central parts of the car- bonate layers. An alternative Sr source is the hydrothermal fluids themselves. For example, the marine formation waters (brines) that may have circulated though these rocks may have contained several hundred parts per million Sr, and, judging by analyses of such waters elsewhere in the world, this Sr would have had a 87Sr/86Sr substantially lower than 0.715 (0.708?~. The hydrothermal exchange processes that affected the

OCR for page 96
HYDROTHERMAL SYSTEMS ASSOCIATED WITH REGIONAL METAMORPHISM AND CRUSTAL ANA TEXIS .725: .: O .715 - - .710 _ 70! 6 ~to.734 :~ UGLY Mar Acres I I ~ `-~NEISSES M1e 2-A/ ~A~/r ~---'t0 .708~ u2.aqA~astre us a~a~so-Na~/re ~: 1'"~\'. .' \ 6 +7 +a +9 +10 Al +12 +13 +14 +15 +~6 ~ 180 FIGURE 6.11 Plot of 87Sr/86Sr at 310 Ma versus 6~80 (after Bickle et al., 1988), comparing the Trois Seigneurs (T-S) granite and petite data with data from the Maladeta platonic complex in the central Pyrenees (M1, M2, and M3; Vitrac-Michard et al., 1980) and with data on some other massifs in the Pyrenees that contain large exposures of Zone 3 granulite-facies basal gneisses (Agly) and of orthogneiss (Canigou and Aston). Also shown are fields for the measured values of Trois Seigneurs shales and phyllites 310 Ma together with the calculated range of model values 310 Ma for these same rocks, assuming that they were originally deposited in the Ordovician with an initial 87Sr/86Sr = 0.707 (see Table 6.1~. The Trois Seigneurs politic schists are an obvious high-~8O, high-87Sr end-member for the biotite granite site, and they have much more homogeneous 87Sr/86Sr values of the shales and phyllites have been changed during prograde Hercynian metamorphism. Fields for the upper mantle, for the main part of the Peninsular Ranges batholith (PRB) from Taylor and Silver (1978) and for the Hercynian granites from Brittany and southwest England (Sheppard, 1986) are also shown. De- spite their having major metasedimentary components in their sources, these other Hercynian granitic rocks also have relatively low initial 87Sr/86Sr ratios, suggesting that the source rocks for these magmas may also have been affected by the same sort of hydrothermal 87Sr/86Sr homogenization process that appears to have affected the Trois Seigneurs pelites. regional metamorphism of the Pyrenees have profoundly modified the isotopic composition of both strontium and oxygen and hydrogen throughout huge volumes of the crust. In this respect it would be extremely misleading to take the INTO values and the calculated model 87Sr/86Sr values of low-grade Cambro-Ordovician shales from the Pyrenees and use them directly as representative protolith end-members for any of the synmetamorphic granitic litho 107 fogies magmas produced in this region, despite the fact that modified material of this type undoubtedly formed a major component of some of the Hercynian granitic mag- mas. The shales and phyllites underwent such profound isotopic modification by hydrothermal processes during prograde metamorphism that in any magma-mixing pro- cess this end-member would have very different oxygen and strontium isotopic compositions than if the original sediments had remained perfectly closed systems subse- quent to deposition. IMPLICATIONS FOR FLUID TRANSPORT DURING METAMORPHISM Timing of Infiltration The massive infiltration of aqueous fluid into the Zone 2 rocks implied by the ]80/~60 data from the Pyrenees is certainly related in some way to the Hercynian metamor- phism. First, the isotopic homogenization is observed only in the (Zone 2) Paleozoic sediments that have been meta- morphosed to medium or high grade. Second, this type of large-scale fluid circulation requires some sort of heat engine to drive it, and the intense, localized thermal anoma- lies generated during metamorphism provide the obvious candidate. Third, the Rb-Sr data (Bickle et al., 1988) confirm that there was a major (87Sr/86Sr) homogenization event in these rocks and that it occurred ~310 Ma during the Hercynian. Defining the timing of infiltration relative to the Her- cynian metamorphic peak is less straightforward. How- ever, minerals in the granites and mica schists preserve equilibrium 6'8O fractionations (e.g., between coexisting quartz, feldspar, and muscovite in leucogranites; see Wickham and Taylor, 19851. This- is contrary to the situ- ation observed in hydrothermal circulation systems asso- ciated with cooling plutons (e.g., Criss and Taylor, 1983; Taylor, Chapter 5, this volume) where large disequilib- rium effects are commonly observed between quartz and coexisting feldspar. The absence of such effects in the Pyrenean rocks rules out any appreciable postmetamor- phic infiltration of the rocks (i.e., following crystallization of the granites). It is also impossible for all of the infiltration to have occurred contemporarily with the Hercynian metamorphic peak. This is because the Hercynian isograd pattern clearly indicates that very steep thermal gradients (80 to 100C/ km; see Wickham, 1987a) existed at this time, and convec- tive circulation of pore fluids would tend to flatten such gradients. Furthermore, a large proportion of the isotopi- cally homogenized rocks were partially melted and/or undergoing ductile deformation at the metamorphic peak,

OCR for page 96
108 and this would have reduced their effective permeability, making them resistant to fluid infiltration. The most plausible scenario is one in which large-scale fluid infiltration of the Paleozoic metasediments occurred over a protracted interval of time during prograde heating of the terrane. This process ceased (or was displaced to higher structural levels) by the time peak metamorphic temperatures were reached. Conceivably, the onset of partial melting in the politic lithologies could have sealed them (and any deeper structural levels) to further infiltra- tion, though it remains plausible that some of the water remained available to promote the large-scale melting effects observed in the Zone 2 rocks (see below). This is further supported by strontium isotope data (Bickle et al., 1988) indicating that the 87Sr/86Sr ratios were partially homogenized in even the lowest-grade rocks, suggesting that isotopic homogenization commenced at the very ear- liest stages of regional metamorphism. Scale of Infiltration The Hercynian metamorphic terranes in the Pyrenees appear to have been pervasively infiltrated by aqueous fluid over depth ranges of at least 4 km (see Wickham and Taylor, 1985; Figure 6.4~. The horizontal extent of the process is much more difficult to estimate due to the frag- mentary nature of Hercynian basement exposures. Alpine deformation has broken up the Hercynian crust into a number of discrete tectonic blocks and slices, so it is impossible to know the original regional distribution of the Hercynian isograds with any certainty. However, in the four North Pyrenean Massifs for which data are re- ported in Figure 6.4, the isotopic homogenization is ob- served along strike for distances of at least 5 to 10 km. Today the high-grade regions occupy relatively restricted zones commonly cored by bodies of granite or gneiss and separated by geographically more widespread low-~rade Paleozoic rocks (Zwart, 1979~. It is therefore quite likely that the Hercynian metamorphic and anatectic effects were not continuous regionally at the same level in the crust and that the high-grade terranes formed a pattern of metamor- phic hot spots. Models for metamorphism in the Pyrenees have explained such a pattern in terms of diapiric doming of granitoids (Soula, 1982) or localized extensional tec- tonism and mafic intrusion (Wickham and Oxburgh, 1987~. The origin of the infiltrating marine fluid (whether derived from connate formation waters or directly from the surface) is critically related to the regional Hercynian thermal structure. Material-balance constraints (Figure 6.9) imply that for reasonable BIRO values of connate for- mation waters a water-rock ratio of about 0.5 would be necessary to account for the isotopic effects, equivalent to STEPHEN M. WICKHAM AND HUGH P. TAYLOR, JR. about 25 percent of the isotonically homogenized rock mass (or about 60 percent by volume). Since the maxi- mum reasonable porosity at this depth is perhaps about 2 percent, this implies that large reservoir regions adjacent to the infiltrated terrane would be required to supply this water if it were exclusively connate in origin. Hence, each homogenized region (Zone 2) should be surrounded by a much more extensive terrane that is isotonically unaf- fected (Zone 1~. We have yet to identify such reservoir regions in the Pyrenees and therefore consider it more likely that most of the fluid was derived directly from the surface. The maximum penetration depth of the fluid is not well constrained. Peak metamorphic mineral assemblages in the Zone 2 rocks suggest pressures of 3 to 4 kbar (10 to 12 km dentin) within the Trois Seigneurs migmatite zone (see Wickham, 1987a). Model calculations of Wickham and Oxburgh (1987) suggest that prograde heating of the metasediments probably took only 1 to 2 m.y., so that although fluid infiltration apparently preceded the attain- ment of maximum metamorphic temperatures, these pres- sure estimates give us our best indication of the depth of the rocks at the time of infiltration. The petrological and stable isotope data are therefore consistent with the pene- tration of surface-derived fluid to depths of at least 10 to 12 km below the surface during Hercynian metamorphism. The data from the Lapege and Agly Zone 3 rocks (basal gneisses) suggest less isotopic homogenization at deeper structural levels (at Agly the mineral assemblages imply Hercynian equilibration pressures of about 5 kbar; Viel- zeuf, 1984~. However, Figures 6.6 and 6.7 indicate that at St. Barthelemy and Castillon the systematics, particularly in the metacarbonates, are more homogeneous for 6~80. In these latter two localities (which equilibrated at pressures of 4 to 6 kbar), it is possible that some homogenization promoted by fluid infiltration occurred at an early stage in the metamorphic history, although without D/H data we cannot say if this water was marine in origin and therefore surface derived (as we can with the Zone 2 rocks). How- ever, if these terranes are also found to exhibit exception- ally heavy ED values, it would imply penetration of sur- face waters into some of the deeper-level rocks as well. Relation of Fluid Transport to Metamorphism and Anatexis The stable isotope data presented above allow calcula- tion of a plausible material-balance water-rock ratio that could generate the Zone 1 to Zone 2 shift in 6~80. How- ever, consideration of the petrology of migmatite zones such as those exposed in the Trois Seigneurs Massif (Wickham, 1987a) also places constraints on water bud

OCR for page 96
HYDROTHERMAL SYSTEMS ASSOCIATED WITH REGIONAL METAMORPHISM AND CRUSTAL ANA TEXIS gets during metamorphism and melting. Here there is a rapid increase from 0 percent to >40 percent partial melt- ing of politic metasediment over only 300 m of the meta- morphic section (representing a temperature increase of no more than 30C). Such a sharp increase implies that aH2O is buffered at fixed values close to unity so that melts can be progressively saturated with water as they are gener- ated. Inasmuch as the leucogranitic melt generated from the pelites probably had a water content close to 8 wt.% (see Wickham, 1987b), 50 percent melting would require the partially melted rocks to contain at least 4 wt.% H2O (disregarding any water that might be contained in hy- drous minerals such as biotite). This is two or three times more than the intrinsic water content of high-grade petite, particularly since much of this water will be held in resid- ual biotite during melting. Hence, an external water source is also required to account for petite anatexis, even though the size of the source is substantially smaller than that demanded by the oxygen isotope data. Although we can- not at this stage prove that the high-6D marine fluids that flushed the Zone 2 rocks were also responsible for the large-scale melting effects, circumstantial evidence points strongly to a link between the two, implying that the same fluid was available to enter the melting zone during petite anatexis. The relationship between the tectonic setting for Her- cynian high-thermal gradient metamorphism in the Pyre- nees and the large-scale fluid infiltration is still uncertain. However, based on a variety of lines of evidence, includ- ing the contemporaneity of marine surface sedimentation with metamorphism and anatexis at depth, the absence of high-pressure metamorphic rocks, the high thermal gradi- ents, and the evidence for deep groundwater circulation, Wickham and Oxburgh (1985, 1986, 1987) proposed that metamorphism occurred in a rift environment and was prompted by intrusion of high-temperature mafic magma into the lower crust. In this scenario localized rift zones (pull aparts?) provided the setting for a focused thermal flux, and deeply penetrating surface water pervasively infiltrated the Paleozoic high-grade metamorphic rocks and promoted melting. Inasmuch as all active continental rift zones have asso- ciated groundwater hydrothermal systems that in many cases penetrate to great but largely unknown depths (see Taylor, Chapter 5, this volume), a rift setting provides the ideal combination of high temperatures at shallow depth with the type of fluid-flow system that we infer from the isotope data. In the Pyrenees, where a thick supracrustal pile of Paleozoic sediment (dominated by politic material in its lower part) existed prior to metamorphism, rifting would have provided the ideal thermal conditions (i.e., temperatures of 700C at 10 to 12 km) to cause large-scale melting of the deeper parts of this sedimentary pile. 109 SUMMARY AND CONCLUSIONS The Hercynian basement in the Pyrenees is exposed over a wide range of structural levels, and these rocks record the effects of an intense Late Carboniferous meta- morphic-anatectic event. l80/~60 ratios vary in a system- atic fashion across this composite crustal section. At high structural levels the low-grade, moderately deformed Pa- leozoic sediments (Zone 1) have typical sedimentary INTO values (+20 to +25 in carbonates, +14 to +16 in shales). The metamorphosed, strongly deformed equivalents of these rocks (Zone 2) have uniform INTO values of about +11.5 in all lithologies. The average INTO value of Zone 2 politic rocks is about 3 per mil lower than in Zone 1, implying that there has been a bulk change in the oxygen isotopic composition of Zone 2, clearly a result of exchange with some type of low-~8O oxygen reservoir. Because low-~8O rocks are rare within the Hercynian crust of the Pyrenees at any structural level exposed, it is most likely that this low-~8O reservoir was aqueous fluid, which pervasively infiltrated the high-grade rocks, lowering and homogenizing INTO values. Mass-balance calculations imply that the quantity of water required to account for the isotopic shift is so large that it can only have been derived from connate formation water or from the Earth's surface. This is sup- ported by exceptionally heavy ED values shown by mus- covite from Zone 2 throughout the Pyrenees; such high-6D values are best interpreted as indicating equilibration with marine fluids at high temperatures, because seawater is the major hich-D fluid reservoir in the Earth's hvdrosDhere . .. . . ~ ~ ~ ----rim Oxygen-isotope systematics from the deep-level am- phibolite and granulite facies gneisses of Zone 3 are in general more heterogeneous than those in Zone 2, although some regions are clearly more isotonically heterogeneous than others. The Zone 3 data cannot be interpreted in such a clear-cut way as for Zones 1 and 2, but- they certainly preclude pervasive infiltration of the Zone 3 rocks by large volumes of aqueous fluid. This tends to support the idea that the Zone 2 infiltrating fluids were derived from above rather than below, since in this case a maximum penetra- tion depth would be expected. At Trois Seigneurs, lower Paleozoic shales and phyl- lites have 87Sr/86Sr values of 0.707 to 0.717 (at 310 Ma), but model values at 310 Ma of 0.709 to 0.736, based on an assumed depositional age of 450 Ma and an initial 87Sr/86Sr = 0.707. On a regional scale these 87Sr/86Sr ratios were homogenized to about 0.713 to 0.717 in both high- and low-grade politic schists during the 340- to 310-Ma meta- morphic events (see Bickle et al., 1988~. Much of this 87Sr/86Sr exchange occurred at very low grades (below the biotite isograd), but significant changes also accompanied the INTO lowering of the phyllites (+13 to +16) during their transformation to andalusite- and sillimanite-grade schists

OCR for page 96
110 (HO = +11 to +12~; all of these effects are attributed to pervasive interactions with hydrothermal fluids. A scenario for fluid flow in the Pyrenees is depicted in Figure 6.12; this shows a surface water flux penetrating to depths of 10 to 12 km in cool areas, adjacent to a localized thermal anomaly such as would have generated the meta- morphic sequences in the Pyrenees. This water migrates through the supracrustal sediments in the heated area dur- ing prograde metamorphism, but movement largely ceases once melting is initiated. Such melting may set up lateral gradients in aH2O that help to draw the deeply penetrating water into the melting zone. This water, however, is largely restricted from entering the Zone 3 gneisses that probably formed the basement to the Paleozoic supracrus- tal metasediments. This may reflect an intrinsic permea- bility contrast between the basal gneisses and the supracrus- tal sediments or, alternatively, the zone of anatexis may have formed a barrier to any deeper penetration of H2O. Note that we are not suggesting in Figure 6.12 that there is any multiple cycling of H2O through the metasedi- mentary sequence. It is clear from all kinds of considera- tions that the water that is drawn in laterally toward the thermal anomaly makes only one pass through the terrane before being heated and expelled upward. Even in rela- tively small hydrothermal convective systems around ig- neous intrusions, calculations show that there is only time for essentially one pass (see Norton and Taylor, 1979~. Our understanding of fluid transport mechanisms in rocks undergoing high-grade metamorphism is still insuf- ficient to say whether the large-scale fluid infiltration documented in Zone 2 in the Pyrenees is likely to be FIGURE 6.12 An approximately SO-km- wide section through the Hercynian crust of the Pyrenees showing listric normal faults and a schematic rift setting for low- pressure regional metamorphism. Melt- ing occurs at the base of the Paleozoic metasedimentary pile in response to heat- ing by mafic intrusions, which are em- placed predominantly within the lower crust (dot-pattern bodies in center of dia- gram). Extensive melting at still deeper levels in the crust generates a group of larger, late-stage granodiorite magma bodies (cross-pattern bodies in Zone 1 to the sides of the diagram, fed from "floored" source regions in the lower crust). The flux of surface-denved metamorphic pore fluids (heavy arrows) is very high in the Lower Paleozoic metasedi- mentary sequences (Zone 2), but either negligible or at least much lower in the underlying basal gneisses (Zone 3), which retain heterogeneous 6'80 values (gneiss = +11, carbonate = +17 STEPHEN M. WICKHAM AND HUGH P. TAYLOR, JR. typical in other metamorphic terranes (for discussion see Ferry, 1986; Wood and Walther, 1986~. However, inas- much as the degree of infiltration appears to be signifi- cantly higher in Zone 2 than in Zone 3, this area may provide an opportunity to identify the most important factors controlling effective permeability during metamorphism. These factors obviously must be linked to the characteris- tics of the rocks at the time of metamorphism, rather than to their present-day physical properties such as permeabil- ity. For example, devolatilization during prograde meta- morphism may have enhanced permeability in Zone 2 but not in Zone 3, where the rocks had previously been dehy- drated and where they already had a high-grade mineral- ogy by Hercynian times. If we are correct in our interpre- tation that the infiltrating fluid was derived from the sur- face or at least from connate fluids within the Paleozoic supracrustal pile, then the Zone 2 to Zone 3 contrast may be depth related, and we may simply be observing the lower, weaker parts of a very deep (hydrostatic?) hydro- thermal circulation system. Some authors (e.g., Wood and Walther, 1986) have suggested that such hydrothermal systems should die out at depths of 3 to 6 km because in some boreholes (e.g., the U.S. Gulf Coast) pore fluid pressures become lithostatic within this depth range. In rift-zone situations like Iceland this is clearly not true, and it is likely that the meteoric hydrothermal convective systems there extended down to at least 10 to 15 km depth (see Taylor, Chapter 5, this volume). However, we are as yet unable to predict where this hydrostatic to lithostatic pressure change will occur in other parts of the crust with different states of stress, volconian ~ ~, ~ . surface I | \~\~3~ woter '~/oto~3j/ ~Zone 1 ~2 ~ km ~12km ' ~ncte- /|/ f ~zol 2 lower limit, I- O f-_ _ water flux - _~ carbolic ~~0- gne i ss ~i\~\~ f111~ ~ __ ~ alto homogeneous Tower crust //////i///// ~ month to +23)9 in contrast to the homogenized Zone 2 values (= +12 to +14 in all lithologies). The granulite-facies gneisses thus seem to lie between the two major anatectic levels that affected the Hercynian crust of the Pyrenees.

OCR for page 96
HYDROTHERMAL SYSTEMS ASSOCIATED WITH REGIONAL METAMORPHISM AND CRUSTAL ANA TEXIS ~, ~ tectonic regimes, upper crustal composition, and thermal structure. Deep boreholes in continental settings with deep hydrothermal systems (e.g., the Salton Trough, Cali- fornia, and the Taupo Rift Zone, New Zealand) may in the future resolve this important question regarding the maxi- mum penetration depth of groundwater convective sys- tems. The data from the Pyrenees suggest that here at least the surface water reached depths of 10 to 12 km, resulting in profound isotopic modification of the rocks. At depth this presumably early-stage hydrostatic regime (or at least one where PH2O < PTOTA} ~ must have been transformed to a lithostatic regime with time, probably as wholesale melting began to occur and the style of rock deformation changed from a br~ttle-fracture mode to a ductile mode. We believe that, at least during the early stages of this transition, the H2O required to promote petite anatexis was still the same water as was responsible for isotopic ho- mogenization (and that this water was originally derived from a connate or marine source). Thus, at some point in the time-temperature history of this Deane (possibly at the onset of anatexis in politic lithologies), aqueous fluids that had been moving under a hydrostatic (Darcy's Law) flow regime became (locally?) involved in the lithostatic regime characteristic of a crustal melting process. At present our understanding of these complex three-phase fluid-rock-melt systems is insuffi- cient to predict the nature of this transition in any detail. However, it is possible that aqueous fluids may have still been able to migrate through fractures in refractory layers (e.g., calcsilicates or quartzites) embedded in the volumet- rically more extensive, partially molten politic layers, thereby providing a water supply to the anatectic zones and maintaining water-rich conditions throughout the melting event. The lithostatic-hydrostatic boundary proba- bly migrated upward and outward from the zone of en- hanced heat flow, at least up until the time that the thermal anomaly began to decay. Understanding the details of this complex lithostatic-hydrostatic transition and the effects of intercalated brittle and ductile layers in terms of space, time, temperature, and permeability will be difficult. It will necessarily involve a wide variety of geological, geochemical, and geophysical efforts and approaches. ~, ACKNOWLEDGMENTS Financial support for this work was provided by NSF grant no. EAR-8313106. We are grateful to M. J. Bickle, H. J. Chapman, and Ron Oxburgh for their collaboration and help on these problems. Discussions with Denis Norton and Steve Sparks were also very helpful. This is contr~bu- tion no. 4442, Division of Geological and Planetary Sci- ences, California Institute of Technology. 111 REFERENCES Autran, A., B. Barriere, B. Bonin, J. Didier, P. Pluck, S. Four- cade, P. Giraud, J. Jonin, J. Lameyre, J-B. Orsini, and G. Vivier (1980~. Les granitoides de France, in Evolutions Geolo- giques de la France, Bulletin de Recherche Ge'ologique et Mine'ralogique, Me'moire 107, Colloque C7, 51-97. Baertschi, V. P. (1957~. Messung und Deutung relativer Haufigkeitsvariationen van 'SO and l3C in karbonatgesteinen und mineralien, Schweizerische Mineralogische und Petro- graphische Mitteilungen 37, 73-151. Bard, J-P., B. Briand, J-M. Cantagrel, G. Guitard, J-R. Kienast, J. Kornrobst, B. Lasnier, C. LeCorre, and D. Santallier (1980~. Le metamorphisme en France, in Evolutions Geologiques de la France, Bulletin de Recherche Ge'ologique et Mine'ralo- gique, Me'moire 107, Colloque C7, 161-189. Bickle, M. J., S. M. Wickham, H. J. Chapman, and H. P. Taylor, Jr. (1988~. A strontium, neodynium, and oxygen isotope study of the hydrothermal metamorphism and crustal anatexis in the Trois Seigneurs Massif, Pyrenees, France, Contributions to Mineralogy and Petrology 100, 399-417. Clayton, R. N., I. Friedman, D. L. Graf, T. K. Mayeda, W. F. Meerts, and N. F. Shimp (1966~. The origin of saline forma- tion waters: I. Isotopic composition, Journal of Geophysical Research 71, 3869-3882. Criss, R. E., and H. P. Taylor, Jr. (1983~. An TO and D/H study of Tertiary hydrothermal systems in the southern half of the Idaho batholith, Geological Society of America Bulletin 94, 640-663. Ferry, J. M. (1983~. Regional metamorphism of the Vassalboro formation, South-Central Maine, U.S.A.: A case study of the role of fluid in metamorphic petrogenesis, Journal of the Geological Society of London 140, 551-76. Ferry, J. M. (1986~. Infiltration of aqueous fluid and high fluid :rock ratios during greenschist facies metamorphism : A reply, Journal of Petrology 27, 695-714. Fleck, R. J., and R. E. Criss (1985~. Strontium and oxygen isotopic variations in Mesozoic and Tertiary plutons of central Idaho, Contributions to Mineralogy and Petrology 90, 291- 308. Frey, M., J. C. Hunziker, J. R. O'Neil, and H. W. Schwander (1976~. Equilibrium-disequilibrium relations in the Monte Rosa granite, Western Alps: Petrological, Rb-Sr and stable isotope data, Contributions to Mineralogy and Petrology 55, 147-179. Garlick, G. D., and S. Epstein (1965~. Oxygen isotope ratios in coexisting minerals of regionally metamorphosed rocks, Geo- chimica et Cosmochimica Acta 31, 181-214. Munha, J., F. J. A. S. Barriga, and R. Kerrich (1986~. High-'8O ore-forming fluids in volcanic-hosted base metal massive sulphide deposits: Geologic, TO and D/H evidence from the Iberian pyrite belt; Crandon, Wisconsin; and Blue Hill, Maine, Economic Geology 81, 530-552. Nebula. H. S. S. M. F. Sheppard, J. M. Rosenbaum, and M Cuney (1986~. Late Hercynian U-vein mineralization in the Alps: Fluid inclusion and C, 0, H isotopic evidence for mixing

OCR for page 96
112 between two externally derived fluids, Contributions to Min- eralogy and Petrology 93, 179-186. Norton, D., and H. P. Taylor, Jr. (1979~. Quantitative simulation of the hydrothermal systems of crystallizing magmas on the basis of transport theory and oxygen isotope data, Journal of Petrology 20, 421-486. Passchier, C. W. (1984~. Mylonite-dominated footwall geometry in a shear zone, central Pyrenees, Geological Magazine 121, 429-436. Rye, R. O., R. D. Schuiling, D. M. Rye, and J. B. H. Jansen (1976~. Carbon, hydrogen and oxygen isotope studies of the regional metamorphic complex at Naxos, Greece, Geochimica et Cosmochimica Acta 40, 1031-1049. Sheppard, S. M. F. (1977~. The Cornubian batholith, SW Eng- land: D/H and TO studies of kaolinite and other alteration minerals, Journal of the Geological Society of London 133, 573-591. Sheppard, S. M. F. (1986~. Igneous rocks: III. Isotopic case studies of magmatism in Africa, Eurasia, and oceanic islands, in Stable Isotopes in High Temperature Geological Processes, J. W. Valley, H. P. Taylor, Jr., and J. R. O'Neil, eds., Reviews in Mineralogy, vol. 16, Mineralogical Society of America, Washington, D.C., pp. 319-371. Shieh, Y. N., and H. P. Schwarcz (1974~. Oxygen isotope stud- ies of granite and migmatite, Grenville Province of Ontario, Canada, Geochimica et Cosmochimica Acta 38, 21-45. Shieh, Y. N., and H. P. Taylor, Jr. (1969~. Oxygen and hydrogen isotope studies of contact metamorphism in the Santa Rosa range, Nevada, and other areas, Contributions to Mineralogy and Petrology 20, 306-356. Soula, J-C. (1982~. Characteristics and mode of emplacement of gneiss domes and plutonic domes in the central-eastern Pyrenees, Journal of Structural Geology 4, 313-342. Taylor, H. P., Jr. (1974~. The application of oxygen and hydro- gen isotope studies to problems of hydrothermal alteration and ore deposition, Economic Geology 69, 843-883. Taylor, H. P., Jr. (1977~. Water/rock interactions and the origin of H2O in granitic batholiths, Journal of the Geological Soci- ety of London 133, 509-558. Taylor, H. P., Jr., and L. T. Silver (1978~. Oxygen isotope relationships in plutonic igneous rocks of the Peninsular Ranges batholith, southern and Baja California, in Short Papers of 4th International Conference on Geochronology, Cosmochronol- ogy, and Isotope Geology, R. E. Zartman, ea., U.S. Geological Survey Open-File Report 78-701, pp. 423-426. Veizer, J., and J. Hoefs (1976~. The nature of TO and '3C/'2C secular trends in sedimentary carbonate rocks, Geochimica et Cosmochimica Acta 47, 697-706. STEPHEN M. WICKHAM AND HUGH P. TAYLOR, JR. Vielzeuf, D. (1984~. Relations de phases dans le facies granulite et implications geodynamiques. L'exemple des granulites des Pyrenees, Annales Scientifiques Universite' Clermont I1 79, 189 pp. Vitrac-Michard, A., and C. J. Allegre (1975~. A study of the formation and history of a piece of continental crust by 87Sr/ 86Sr method: The case of the French oriental Pyrenees, Contri- butions to Mineralogy and Petrology 50, 257-285. Vitrac-Michard, A., F. Albarede, C. Dupuis, and H. P. Taylor, Jr. (1980~. The genesis of Variscan (Hercynian) plutonic rocks: Inferences from Sr, Rb, and O studies of the Maladeta igneous complex, Central Pyrenees, Spain, Contributions to Mineral- ogy and Petrology 72, 57-72. Wickham, S. M. (1987a). Crustal anatexis and granite petro- genesis during low pressure regional metamorphism; the Trois Seigneurs Massif, Pyrenees, France, Journal of Petrology 28, 127-169. Wickham, S. M. (1987b). The segregation and emplacement of granitic magmas, Journal of the Geological Society of London 144, 281-297. Wickham, S. M., and E. R. Oxburgh (19851. Continental rifts as a setting for regional metamorphism, Nature 318, 330-333. Wickham, S. M., and E. R. Oxburgh (1986~. A rifted tectonic setting for Hercynian high thermal gradient metamorphism in the Pyrenees, in The Geological Evolution of the Pyrenees, E. Banda and S. M. Wickham, eds., Tectonophysics 129, 53-69. Wickham, S. M., and E. R. Oxburgh (1987~. Low pressure regional metamorphism in the Pyrenees and its implications for the thermal evolution of rifted continental crust, Philo- sophical Transactions of the Royal Society of London A321, 219-243. Wickham, S. M., and H. P. Taylor, Jr. (1985~. Stable isotopic evidence for large-scale seawater infiltration in a regional metamorphic terrane: The Trois Seigneurs Massif, Pyrenees, France, Contributions to Mineralogy and Petrology 91, 122- 137. Wickham, S. M., and H. P. Taylor, Jr. (1987~. Stable isotope constraints on the origin and depth of penetration of hydro- thermal fluids associated with Hercynian regional metamor- phism and crustal anatexis in the Pyrenees, Contributions to Mineralogy and Petrology 95, 255-268. Wood, B. J., and J. V. Walther (1986~. Fluid flow and its implication for fluid-rock ratios, in Fluid-Rock Interactions During Metamorphism, J. V. Walther and B. J. Wood, eds., Spnnger-Verlag, New York, pp. 89-108. Zwart, H. J. (1979~. The geology of the Central Pyrenees, Leidse Geol. Med. 50.1, 1-74.