<%BANNER%>

Analysis of porosity evolution during low temperature metamorphism of basaltic lavas and implications for fluid flow

University of Florida Institutional Repository

PAGE 1

ANALYSIS OF POROSITY EVOLUTION DURING LOW TEMPERATURE METAMORPHISM OF BASALTIC LAVA S AND IMPLICATIONS FOR FLUID FLOW By JANE E. GUSTAVSON A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2006

PAGE 2

Copyright 2006 by Jane E. Gustavson

PAGE 3

iii ACKNOWLEDGMENTS There are many people in my life who have b een influential in helping me to obtain my Master of Science. Of those, there are a few that require special mention. My father, James M. Gustavson, did all that he knew to pr ovide the best for his children. If it were not for his selflessness and honest love, I w ould not be the person I am today. Allyn Spear, though we have a mottled history, was the one who stood beside me and gave me the courage to pursue a higher level of educa tion when I was filled with so much self doubt. Dr. Philip S. Neuhoff was always there with a patient yet guiding hand to see me through my research. He has become more than just my advisor; he is also a good friend. And finally, PJ Moore, who has become the co mpanion I have searched for. His promise to walk beside me gives me the strength to en ter the real world with the confidence I need to succeed.

PAGE 4

iv TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii LIST OF TABLES.............................................................................................................vi LIST OF FIGURES.........................................................................................................viii ABSTRACT....................................................................................................................... xi CHAPTER 1 INTRODUCTION........................................................................................................1 Porosity and Permeability in Vesicular Lavas..............................................................3 Low Grade Alteration of Basaltic Lavas......................................................................6 Coupling Between Porosity Evol ution and Chemical Reactions.................................8 Role of the Present study..............................................................................................9 2 LOW-GRADE ALTERATION OF THE NORTH SHORE VOLCANIC GROUP, MINNESOTA.............................................................................................................10 Introduction.................................................................................................................10 Geologic Background.................................................................................................12 Methods......................................................................................................................15 Results........................................................................................................................ .17 Primary Basalt Composition and Mineralogy.....................................................17 Regional Alteration Mineralogy..........................................................................22 Laumontite zone lavas..................................................................................26 Stilbite-heulandite zone lavas.......................................................................27 Thomsonite-mesolite zone lavas..................................................................27 Analcime zone lavas.....................................................................................28 Alteration Mineral Chemistry..............................................................................29 Discussion...................................................................................................................31 Regional Variation of Metamorphic Grade.........................................................31 Conditions of Alteration......................................................................................38 Structural Interpretations.....................................................................................42

PAGE 5

v 3 PETROGRAPHIC AND DIGITAL ANAL YSIS OF POROSITY EVOLUTION DURING ALTERATION..........................................................................................45 Introduction.................................................................................................................45 Methods......................................................................................................................47 Field Techniques.................................................................................................47 Optical and Digital Techniques...........................................................................48 Observations...............................................................................................................49 NSVG Field Sites 3, 4 and 6...............................................................................49 Reaction Progress................................................................................................55 Partial infilling at low-grades.......................................................................55 Mafic phyllosilicate to zeolite infilling........................................................58 Multiple stage zeo lite infilling.....................................................................63 Discussion...................................................................................................................63 Nature of Porosity and Permeability in Basalts...................................................63 Controls on Reaction Progress............................................................................68 4 MODELING OF MINERAL PARAGENESES........................................................70 Introduction.................................................................................................................70 Methods......................................................................................................................71 Results........................................................................................................................ .75 Mineral Parageneses............................................................................................75 Volume Changes During Alteration....................................................................83 5 DISCUSSION.............................................................................................................86 Reaction Progress in Low-Grade Metabasalts............................................................86 Dependence of Reaction Progress on Pore Size.........................................................88 6 CONCLUSIONS........................................................................................................96 LIST OF REFERENCES...................................................................................................99 BIOGRAPHICAL SKETCH...........................................................................................110

PAGE 6

vi LIST OF TABLES Table page 2.1 Field study locations in the North Shore Volcanic Group.......................................16 2.2 Whole-rock chemical com positions (wt %) of samples...........................................18 2.3 Representative compositions of plagioclase............................................................20 2.4 Representative compositions of pyroxenes..............................................................21 2.5 Representative compositions of mafic phyllosilicates.............................................30 2.6 Representative compositions of thomsonites...........................................................32 2.7 Representative compositions of me solite, analcime, and laumontite.......................34 3.1 Statistical analysis of vesicle size (d iameter) as a function of reaction progress from outcrop scale measur ements at Site 6..............................................................51 3.2 Statistical analysis of ve sicle size (area) as a func tion of reaction progress from measured vesicles of a low-grade metaba salt from East Greenland (thin section 421505).....................................................................................................................57 3.3 Statistical analysis of vesicle size as a function of reaction progress for vesicles filled with chlorite and/or laumontite at the thin section scale (sample NS04-14) from Site 4................................................................................................................59 3.4 Statistical analysis of vesicle areas (mm2) and clay rim thicknesses (mm) for sample 94-80 from eastern Iceland..........................................................................63 4.1 Bulk rock composition for olivine tholeiite basalt (wt %).......................................72 4.2 Modal abundances and compositions of primary mineral phases in olivine tholeiite.....................................................................................................................7 2 4.3 Calculated anhydrous bulk composition of basaltic andesite before and after reaction.....................................................................................................................74 4.4 Reactant phase composition, abundance, and relative dissolution rate in basaltic andesite used as input for reaction path model.........................................................75

PAGE 7

vii 4.5 Mole, volume and mass amounts in una ltered and altered andesite for the computer (modeled) and mathema tical (calculated) models....................................81 4.6 Phase compositions and abundances at end of reaction path for basaltic andesite..82

PAGE 8

viii LIST OF FIGURES Figure page 1.1 Cross sectional view showing distributi on of primary porosity (vesicles, scoria, breccia) typical of thick aa lava flows........................................................................4 1.2 Schematic diagrams depicting the effect s of pore size, shape and connectivity on the permeability of vesicular basalts..........................................................................4 2.1 Generalized geological map of northeaste rn Minnesota showing the distribution of the NSVG lavas and associated Ke weenawan intrusives (after Miller et al. 2001).........................................................................................................................1 1 2.2 Flow boundary between two vesicu lar basalt flows at Site 12................................14 2.3 Total alkalis-silica diagram showing compositional ranges of NSVG basalts sampled in this study (black circles)........................................................................19 2.4 Field photo from Site 3 (Table 2.1) s howing an extensively altered pahoehoe flow. Visible within the flow are bleached haloes around vesicles connected by thin anastomosing bleached areas............................................................................23 2.5 Secondary alteration in NSVG lavas. All photomicrographs were taken through partially crossed polars.............................................................................................24 2.6 Compositions of mafic phyllosilicates fo rmed during regional metamorphism of the NSVG as a function of the number of non-interlayer cations (Si + Al + Mg + Fe) versus the interlayer ch arge (2 Ca + Na + K)....................................................31 2.7 Plot showing the compositional varia tion of analyzed thomsonites in NSVG basalts (black circles)...............................................................................................33 2.8 Generalized map of northeastern Minneso ta showing the distribution of NSVG lavas (gray) and interpretations of me tamorphic grade based on this study and the work of Schmidt (1993) and Schmidt and Robinson (1997)..............................35 2.9 Stratigraphy of the NSVG (Vervoort and Green, 1997; Miller et al. 2002) correlated along the fold axis dividing th e lavas into the southwest and northeast limb........................................................................................................................... 36

PAGE 9

ix 2.10 Schematic diagram showing the distri bution of minerals and mineral zones typically developed during very low gr ade metamorphism of large igneous provinces (after Walker, 1960; Neuhoff et al. 1997, 2000)....................................39 3.1 Outcrop photograph of vesicular basalt at Site 6 ( Table 2.1 ) illustrating typical variation in pore size, shape and spacing.................................................................50 3.2 Images of reaction zones around and betw een macroscopic pore space in NSVG lavas.......................................................................................................................... 52 3.3 Image is of sample NS04-14, showing the laumontite zone alteration found at Site 4......................................................................................................................... 53 3.4 Various levels of secondary alteration in the matrix of NSVG basalts sampled during the field season. All photomicr ographs were taken through crossed polars........................................................................................................................5 4 3.5 Example of digital analysis of vesicle f illings in a zeolite fa cies vesicular lava from East Greenland.................................................................................................56 3.6 Frequency of vesicle area s versus the amount of secondary alteration filling the vesicle for a sample of chabazite-tho msonite zone alteration from East Greenland (cf. Neuhoff et al. 1997)........................................................................58 3.7 Plot shows the thickness in mm of chlori te rims (black circles) and the total percent of the vesicle filled with chlorite (open circles) with respect to total vesicle area in mm2..................................................................................................60 3.8 Plot shows the percentage of mineral inf illing with respect to total vesicle area....61 3.9 Photomicrograph showing a thin section of a mesolite-scolecite zone lava from eastern Iceland..........................................................................................................62 3.10 Data compiled from digital analysis of mesolite-scolecite zone lava from eastern Iceland (refer to Figure 3.3 )....................................................................................64 4.1 Mineralogic composition of lavas as a function of reaction progress. Panels A, B, and C depict i ndividual models...........................................................................76 4.2 Instantaneous change in total mineral vol ume as a function of reaction progress. Panels A, B, and C correspond to models shown in Figure 4.1 ..............................84 5.1 Schematic representation of a spherical pore of radius rpore surrounded by a reaction aureole of thickness xauroele.........................................................................89 5.2 Graph depicting the percen tage of volume occupied by the pore in a given poreaureole local reaction re gion as a function of rpore and xaureole using pore sizes and aureole thicknesses typical of vesicular mafic lavas................................................90

PAGE 10

x 5.3 Schematic representation of a spherical pore showing secondar y infilling typical of the petrographically observed alteration in the basaltic ande site of the NSVG..92 5.4 Plot depicting the aureole thickness (x) required to completely fill pore space within a vesicle of a given pore radius (rpore) with alteration minerals, shown for total infilling and for clay or laumontite infilling....................................................93 5.5 Percent chlorite infilling as a function of the axial ratio for an ellipsoidal vesicle elongated in one direction assuming a cons tant volume and a constant chlorite rim thickness............................................................................................................94

PAGE 11

xi Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science ANALYSIS OF POROSITY EVOLUTION DURING LOW TEMPERATURE METAMORPHISM OF BASALTIC LAVA S AND IMPLICATIONS FOR FLUID FLOW By Jane E. Gustavson August 2006 Chair: Philip S. Neuhoff Major Department: Geological Sciences Basaltic lavas are the most abundant rock type in the crust and globally form important reservoirs for the migration and st orage of crustal fluids. As in any porous medium, the porosity and permeability of basalt lavas control the movement and extent of chemical interaction of these fluids and thus the availability and quality of obtainable resources (e.g., groundwater and petroleum). The highly-reactiv e nature of basaltic lavas near the surface leads to intimate coupling be tween hydrologic properties and extent of chemical alteration, which controls the a quifer/reservoir quality and chemical fluxes during alteration. Our ability to quantitatively understa nd these processes is limited by rigorous models of how chemical reactions and porosity/permeability modifications are coupled in these systems. In the present study, development of seconda ry mineral parageneses is used as a monitor of fluid access to pores and to obtain a quantitative descrip tion and interpretation

PAGE 12

xii of the evolution of pore space through time. This work has shown that fluid movement through vesicular basalts occurs along a pe rmeable network of microfractures and connected vesicles, which are depicted by the extremely weathered aureoles of alteration surrounding these pathways and the vesicles they connect. This research also shows that the extent of reaction progress, as evidenced by the temporal sequence of mineral infillings within pore space, is heavily influenced by the vesicle size and shape. Larger vesicles exhi bit greater degrees of reaction progress while smaller ones are closed off earlier in the reac tion progress. Irregular ly shaped vesicles may close off earlier in the reaction progress due to constriction caused by the shortest diameter. This dependence of reaction progre ss on pore size is more specifically based on the volume percent occupied by the pore with in a closed pore-aureole system. Each vesicle independently exchanges chemical co mponents with the surrounding lava matrix (the alteration aureole), implying that th e reaction progress exhibited within a given vesicle is based on the proportion of the vesicle to aureole size. For a given vesicle size, the pore space evolution due to secondary mi neralization will be less with a smaller alteration aureole, due to the low level of fluid rock interaction evidenced by this smaller alteration aureole. Inversely, a larger alteration aureole im plies a higher level of fluidrock interaction, which may create enough secondary minerali zation to progress to later stages of alteration or to completely fill the pore space.

PAGE 13

1 CHAPTER 1 INTRODUCTION Fluid movement is inherently coupled with geochemical reactions during earth processes. Mass transport of chemical species during fluid flow has received considerable attention as a means of explaining the chemi cal evolution of geochemical systems such as hydrothermal ore deposits, petroleum reservoirs, groundwater aquifers, and metamorphic systems (e.g., reviews and re ferences therein of Norton and Knight, 1977; Cathles, 1981, 1990; Etheridge et al., 1983; Garven and Freeze, 1984a,b; Norton, 1984; Wood and Hewett, 1984; Garve n, 1985, 1989, 1995; Broedehoeft and Norton, 1990; Steefel and Lasaga, 1992; Garven et al., 1993; Garven and Raffensperger, 1994; Raffensperger and Garven, 1995a,b; Lichtner et al., 1996; Person et al., 1996; and many others). Dissolution and precipitation of so lid phases during reactive transport processes leads to changes in permeability and porosity of rocks and sediments. Permeability influences the extent and distribution of chemical alteration during hydrothermal and metamorphic processes by controlling the rate s and magnitudes of supply and removal of chemical species (e.g., Lasaga, 1986, 1989; Or toleva et al., 1987; Norton, 1988; Dewers and Ortoleva, 1990; Steefel and Van Ca ppellen, 1990; Steefel and Lasaga, 1992, 1994; Manning et al., 1993; Lasaga and Rye, 1993; Ro se, 1995; Bolten et al., 1999). Porosity controls the amount of aqueous solution that is in contact with the porous media, and can also have considerable effects on permeability (Xu and Pruess, 2001). The interrelation between porosity, permeability, and chemical reaction is especially profound in basaltic lavas and othe r volcanic rocks. Th e effect of mineral

PAGE 14

2 paragenesis on porosity in low-grade meta basalts has been noted in many studies (Schmidt, 1990, 1993; Bevins, Rowbotham, and Robinson, 1991; Manning and Bird, 1991, 1995; Robert and Goff, 1993; Schmidt and Robinson, 1997; Neuhoff et al. 1999). Basaltic lavas are the most abundant rock t ype in the earth’s cr ust (Wedepohl, 1969), and can form extensive aquifers for the stor age and migration of groundwater (e.g., Kononov, 1978; Ingebritson and Scholl, 1993; Rose et al. 1996; Manga, 1997, 1999), geothermal fluids (e.g., Arnorsson, 1995a, b) and petr oleum (e.g., Pendkar and Kumar, 1999; Christiansen, 1994; Iijima, 2001). Once basalti c lavas are emplaced at the surface of the earth, the primary magmatic phases formed dur ing cooling (i.e., oliv ine, anorthite-rich plagioclase, pyroxenes, oxides and glass) are metastable. The highly-reactive nature of basaltic lavas near earth’s surface leads to extensive dissolution and hydrolysis of the primary magmatic phases that releases chemical constituents of the lavas into the aqueous phase. There they are transported by di ffusion and advection through the rock (e.g., Steefel and Lasaga, 1992) and/ or re-precipitated in seconda ry mineral phases such as clays, zeolites, and silica minerals. These phases have relatively large molar volumes and typically occlude pore space as they require additional room to gr ow beyond the space generated through dissolution or hyd rolysis of the primary basalt ic minerals. This causes portions of the rock to be closed off from the fluid flow pathway, which has a significant effect on the porosity and fluid movement through the rock. This thesis explores the relationship be tween porosity evolution during secondary mineral paragenesis and fluid flow through vesi cular basalts. The following sections of this chapter provide a review of the nature of porosity and permeability in vesicular lavas, low-grade alteration of basaltic lavas, and the relationships between chemical reaction,

PAGE 15

3 porosity, and permeability in terms of flui d movement through these lavas. This information is used to pose several questions about the coupling of chemical alteration and porosity evolution in lavas that are addressed by the work below. Porosity and Permeability in Vesicular Lavas Due to parameters such as temperature a nd pressure experience d by the lava flow during eruption, total porosity can vary signifi cantly within a single st ratigraphic flow as can factors such as the size, spacing and geometry of pore spaces (e.g., Larsen et al. 2004; Burgisser and Gardner, 2004). Vesicular ba salt flows can generally be divided into three sections based on the density of vesicl es (Figure 1.1): 1) the bottom, which is high in porosity with elongated pores due to the stress of overburden during flow; 2) the center, which tends to be ma ssive and lacks pores; and 3) the top, which has the highest porosity and more spherical pores due to less overburden pressu re during deposition (Neuhoff et al. 1999). As a consequence, porosity can vary dram atically from essentially 0% in the massive flow centers up to 85% in scoraceous zones (e.g., Cashman and Mangan, 1994). The macroscopic porosity depicted in Figur e 1.1 tends to dominate total porosity, as porosities determined by point counting or othe r digital analyses tend to agree well with those obtained by laboratory techniques (e.g., Saar and Manga, 1999, Al-Harthi et al., 1999).

PAGE 16

4 Figure 1.1: Cross sectional view showing di stribution of primary porosity (vesicles, scoria, breccia) typical of thick aa lava flows. Note heterogeneous but regular distribution of porosity through the flow. Figure 1.2 illustrates three di fferent possible pore geomet ries within the vesicular zones of basalt. In Figure 1.2A, the highly ve sicular lava contains coalesced pore spaces, leading to high permeability that is controlled by the apertures of the pore intersections (Saar and Manga, 1999); this situation is typical of scoraceous zones within flow tops. At lower porosities, where vesicle density is too low for widespread coalescence, variations in vesicle geometry can have a profound effect on the nature of permeability. Figure 1.2: Schematic diagrams depicting the effects of pore size, shape and connectivity on the permeability of vesicular basalts. The lava depicted in (A) has highest porosity, with coalesced vesicles that lead to connected fluid flow paths. The lavas shown in (B) and (C) have the sa me porosity, but different pore size and mean pore-pore distances that will affect the ease and direc tion of fluid flow through the rock.

PAGE 17

5 For instance, the lavas depicted in parts B and C have the same porosity, but different pore sizes. The distribution of pore space into a greater number of smaller pores in part C relative to that in part B leads to significan tly reduced pore-pore dist ances, which is likely to impact the ease of fluid movement between pores. These varia tions in porosity and pore geometry depicted in Figure 1.2 can exist even within a single ve sicular zone of the basalt (Gardner et al. 1996). Vesicular volcanic rocks e xhibit a wide range of permeabilities, ranging from about 10-7 to 10-16 cm2 (e.g., Klug and Cashman, 1996; Saar and Manga, 1999; Mueller et al. 2005). Zones within basaltic lavas exhibi ting macroscopic porosity generally have relatively high permeabilities, often in the range of 10-7 to 10-11 cm2. This is comparable to clean unconsolidated sand or karstic li mestone. Matrix permeabilities in non-porous lavas can be considerably lower, often as low as 10-16 cm2 (e.g., Saar and Manga, 1999; Sruoga et al. 2004; Mueller et al. 2004). Recent research demonstrates that ve sicular basalts do not follow porositypermeability relations typically applied to granular rocks (Saar and Manga, 1999). Previous studies of the relationship between porosity and permeability in granular rocks have applied percolation theory (e .g., Carman, 1956; Dullien, 1992; Bosl et al. 1998) which predicts a percolation threshold, or the minimum porosity at which a connected pathway for fluid movement through the sample exists, of 30%. Following this theory, when the porosity of a sample drops below th is critical value there will be a dramatic decrease in permeability. This approach lead s to widely applied power law dependencies of permeability on porosity such as the Kozeny-Carman equation (e.g., Carman, 1956; Dullien, 1992; Le Gallo et al. 1998; Balashov and Yardley, 1998; Park and Ortoleva,

PAGE 18

6 2003; Freedman et al. 2003). Experimental studies of lava degassing, transport modeling, and textural studies of pumices and other high-vesicular materials predict that interconnected pathways resulting from bubble co alescence of vesicles generally occur at porosities over 30 % (e.g., Klug and Cashman, 1996; Blower, 2001; Mueller et al. 2004; Burgisser and Gardner, 2005). However, it is clear that vesicular volcanic rocks exhibit relatively high, but variable, permeabilities (~10-8 cm2) down to porosities as low as a few percent, in contrast to the behavior of granular rocks an d the predictions of percolation theory (e.g., S aar and Manga, 1999; Mueller et al. 2004). Numerous models have been proposed to explain this behavior largely relying on te xtural explanations (e.g., Saar and Manga, 1999; Blower, 2001). Recen t work has suggested that the elevated permeabilities in volcanic rocks are associat ed with microfractures formed between vesicles during degassing, leadi ng to connected fluid pathways that are not quantitatively important components of total porosity (Mueller et al. 2004). Low Grade Alteration of Basaltic Lavas When basalt is emplaced at the earth’s surface it is metastable. The olivine, pyroxene, anorthite-rich plagioclase, oxide a nd glass phases in basalts form under higher temperature conditions and are easily w eathered at surface conditions. Chemical alteration of basalts begins to occur almo st immediately after emplacement (e.g., Hearn et al. 1981; Jakobsson and Moore, 1986), and the sp ecific reactions that take place between basaltic lavas and groundwat ers are strongly dependent on temperature and pressure conditions as well as the composition of the lavas and fluids (e.g., Kristmannsdttir and Tmasson, 1978; Neuhoff et al. 1999, 2000). In general, the sequence of reactions experienced by lavas of this nature can be summarized as 1) r eaction of lavas with oxygenated aqueous solutions forming iron oxyhydroxides, celadonite, amorphous silica,

PAGE 19

7 amorphous aluminosilicates, and clay minerals during weathering; 2) continued reaction of the lavas to form clay minerals (smectite chlorite) and later zeolites and other calcium aluminosilicates during burial metamorphism; and 3) localized, late stage hydrothermal alteration in fracture system s around intrusions (Neuhoff et al. 1999). These various stages of mineral formation can be dis tinguished through care ful use of geologic relationships and relative dating techniques (Neuhoff et al. 1999). The sequence of chemical reactions that alter basaltic lavas after emplacement typically leads to the development of depth-controlled zones, which reflect increasing temperature with depth. Clay minerals trend from dioctahedral smectite to trioctahedral smectite to mixed layer chlorite-smectite to chlorite with increasing depth/temperature (e.g., Schiffman and Friedliefsson, 1991; Neuhoff et al. 1999, 2006); these minerals are observed as replacements of glass, olivin e, pyroxenes, fine-grained groundmass, sometimes plagioclase, and are also found to rim vesicle walls. Zeolites occur as replacements of plagioclase and occasi onally groundmass, and more commonly as vesicle infillings after forma tion of mafic phyllosilicates. Prominent depth-controlled distribution of individual zeolite species within vesicles is often noted, allowing delinea tion of mineral zones defined by the occurrence of one or more index zeolite species. These zones are well described in many tholeiitic lava flows throughout the world, and as many as five separate zones exist in regionally metamorphosed basaltic lavas and Icela ndic geothermal systems (Walker, 1951, 1960a, b; Sukheswala et al. 1974; Kristmannsdottir and Tomasson, 1978; Jorgensen, 1984; Murata et al. 1987; Schmidt, 1990, 1993; Neuhoff et al. 1997, 1999; Christiansen et al. 1999). They are relatively uniform in th ickness over large distances and cross-cut

PAGE 20

8 individual lava flows. The thickness, orie ntation, and geologic relations between these mineral zones and lava stratigraphy provide critical information for assessing preerosional lava thicknesses as well as crustal deformation (e.g., Neuhoff et al. 1997, 2000). Coupling Between Porosity Evolution and Chemical Reactions As in any porous medium, the hydrologic pr operties of basalt lavas control the movement and extent of fluid -rock interaction, which affect s the formation of secondary minerals and thus the availa bility and quality of obtaina ble resources (e.g., groundwater and petroleum). In principle, porosity change during chemical alteration can be determined from the change in mineral vol ume during reaction due to mineral dissolution and precipitation. The extent of porosity de struction is a function of both the relative volumes of primary and secondary phases, and of their relative solubi lities in the aqueous phase (Putnis, 2002). Provided a suitable determinative relationship between porosity and permeability is known (such as th e Kozeny-Carman equation; Carman, 1956), changes in mineral volume during chemical reaction from petrographic or modeling results can be used to estimate changes in permeability (e.g., Steefel and Lasaga, 1994; Panda and Lake, 1995; Balashov and Yardley, 1998; Le Gallo et al. 1998; Saripalli et al. 2001, 2005; Park and Ortoleva, 2003; Xu and Pruess, 2001; Freedman et al. 2003). In low-grade metabasalts, the destruc tion of porosity is evident through the progressive infilling of vesicles by phases such as mafic phyllosilicates and zeolites. This is a direct consequence of the open crystal st ructures of mafic phyllosilicates and zeolites, which leads to a net increase in mineral volume (Neuhoff et al. 1999, 2000). These pore fillings will alter the size, geometry and connectivity of the pore space and can effectively close off portions of the rock to further alteration (Neuhoff et al. 1999).

PAGE 21

9 Because, as noted above, permeability of vesicu lar lavas is likely strongly dependent on the geometry of the pore space, these proce sses may lead to variations in permeability with reaction progress that are not readily predicted by simple power-law porositypermeability relationships. What is clear is that progressive alteration of lavas leads to distinct reductions in permeability, for instan ce in oceanic lavas (Fisher, 1998), petroleum reservoirs (Sruoga et al. 2004; Luo et al. 2005), and geothermal systems (Stimac et al. 2004). Role of the Present study The ability to quantify and further unde rstand the time dependent changes in the properties of porous media is sought after by geologists, hydrologists and oil companies among many others (Aharonov et al. 1997). The evolution of porosity and permeability during reactive transport has received surp risingly little attention (e.g., Steefel and Lasaga, 1994; Le Gallo, 1998; Saripalli et al. 2001; Freedman et al. 2003), and our ability to quantitatively understand these proc esses is limited due to a lack of rigorous models of how chemical reactions and poros ity/permeability modifications are coupled in these systems. This is particularly true fo r vesicular lavas for reasons noted above. The present study uses secondary mineral paragene ses in vesicular basalts to address three aspects of this problem: 1. What causes the high permeabilities of vesicular volcanic rocks, even at low porosities? 2. Does the extent of reaction progress (as evidenced by the tem poral sequence of mineral infillings of pore space) vary w ith aspects of pore geometry such as spacing, size, and shape? 3. What is the relationship between pore geometry and porosity evolution during mineral paragenesis?

PAGE 22

10 CHAPTER 2 LOW-GRADE ALTERATION OF THE NORTH SHORE VOLCANIC GROUP, MINNESOTA Introduction The distribution of individual zeolites, filling vesicles in low-grade metabasalts of large igneous provinces, is typically charac terized by several depth-controlled zones defined by the occurrence of one or more index zeolite species (e.g., Walker, 1960a,b; Sukheswala et al. 1974; Murata et al. 1987; Schmidt, 1993; Neuhoff et al. 1997, 1999). The thicknesses, orientations, and geologic relations between these mineral zones and lava stratigraphy provide critical information for assessing geothermal gradients, crustal deformation, and pre-erosiona l lava thicknesses (e.g., Neuhoff et al. 1997, 2000). This style of alteration has typically been obse rved in Mesozoic or younger, relatively undeformed lavas due to the fact that low gr ade terrains are often lost to erosion or overprinted by subsequent metamorphism in older provinces. One exception is the 1.1 Ga North Shore Volcanic Group (NSVG), which is regionally deformed, yet still contains well-developed zeolite facies metamor phic zonation (Schmidt, 1993; Schmidt and Robinson, 1997; Figure 2.1). The overall structure of the NSVG is a plungi ng synclinal fold with a hinge axis in the Tofte-Lutsen area (e.g., Miller et al. 2002) dividing the region into two sections; the southwest limb and the northeast limb. Previous studies (e.g. Schmidt, 1993, Schmidt and Robinson, 1997) established a progression of metamorphic mineral assemblages in the southwest limb that range from lower zeo lite facies in the uppermost parts of the

PAGE 23

11 Figure 2.1: Generalized geological map of northeastern Minnesota showing the distribution of the NSVG lavas and asso ciated Keweenawan intrusives (after Miller et al. 2001). Numbered locations (red circles) refer to sampling sites from the present study (cf. Table 2.1). stratigraphy through greenschist facies rocks in the lowermost lavas near Duluth. There appears to be little previous work on the metamorphism of the northeast limb as assessing the metamorphic history of this part of the province is complicated by the variable lava chemistry, which covers the whole range be tween rhyolites and oliv ine normative basalts (BVSP, 1981; Brannon, 1984; Schmidt, 1990; Miller et al. 2002; Boerboom, 2004). The present study investigates the distribution of metamo rphic grade in the NSVG, with a concentration on the alteration present in the northeast limb. Field, petrographic, and analytical studies are combined to asse ss the variation in metamorphic grade across

PAGE 24

12 the syncline. This data is then used to interpret the c onditions and timing of alteration relative to the structural development of this province. Geologic Background The most voluminous accumulations of basa ltic lava on continental crust are flood basalts, which typically outcr op as thick sections of subhor izontal, subaerial tholeiitic flows (Coffin and Eldholm, 1994; Winter, 2001), such as the basalts of the Keweenawan province associated with the Midcontinent Rift The Midcontinent rift system (MRS) is a major feature covering over 2,000 km of central North America (Van Schmus and Hinze, 1985; Cannon et al. 1989, Cannon, 1992), extending northeast from Kansas into central Lake Superior and then southeastward thr ough Michigan into Ohio (Green, 1982; Hinze et al. 1997; Ojakangas and Dickas, 2002). The MRS is a product of extension occurring between 1108 and 1094 Ma, which was aborted pr ior to complete continental rifting due to convergence from the Grenville province allochthon to the eas t (Cannon, 1994; Allen et al. 1997). Magmatism associated with this extension is believed to originate from a large asthenospheric plume centered below wh at is now Lake Superior (Vervoort and Green, 1997). The volcanic pile emplaced during magmatism is more than 8 km thick (Ojakangas and Matsch, 1982; Schmidt and Robinson, 1997) with an estimated volume of 1.3 x 106 km3 (Hutchinson et al. 1990; Allen et al. 1992). This province is one of the oldest and best preserved large igneous provinces (BV SP, 1981; Boerboom, 2004) with the majority composed of either intrusi ons or subaerial volca nics later buried by Phanerozoic sedimentary strata (Ojakangas a nd Dickas, 2002). Surface exposures of the MRS are known only in the Lake Superior re gion, where a 30 km thick sequence of volcanic and sedimentary rocks composes the Keweenawan Supergroup (e.g., Ojakangas and Dickas, 2002).

PAGE 25

13 The NSVG, a section of the Keweenawan province located in northern Minnesota, covers the extent of the Minne sota shoreline along Lake Supe rior and can be divided into two sub-basins of gently dipping strata (V ervoort and Green, 1997). These lavas consist mainly of olivine tholeiites with a con tinuous range of compositions from strongly olivine-phyric tholeiites to rhyolites (BVSP, 1981; Schmidt, 1990; Boerboom, 2004). Magmatism throughout most continental flood basa lts is dominantly basaltic with only about 1% of the flows having a felsic co mposition; within the NSVG, 10% of the southwest limb and 25% of the northeast limb is composed of felsic flows (Vervoort and Green, 1997). It still is unclear if the range in composition is due to large-scale crustal melts or differentiates of more primitive magma (Vervoort and Green, 1997). The structure of the MRS is dominated by ex tension followed by compression, forming horst and graben structures followed by reverse and, in some cases, tr ansform faulting (Van Schmus and Hinze, 1985; Vervoort and Gree n, 1997). Geophysical and geological data show that the Lake Superior region is a fault-bounded-asym metric basin, but the exact structure of the area encompassing the NSVG is still unknown (Van Schmus and Hinze, 1985; Hinze et al. 1997). The flows of the NSVG are mainly be tween 1 and 50 m thick (Schmidt and Robinson, 1997), with an average thickness between 5 and 25 m (Schmidt, 1990). In areas where lava flows vary in compos ition, the morphology of individual flows is characteristic for each major compositional type (Saemundsson, 1979). In the NSVG, the volumetrically dominant olivine tholeiites occur as pahoehoe flows, while other compositions, such as the basaltic ande sites occur as aa flows (Schmidt, 1990; Boerboom, 2004). Figure 1.1 shows a cross s ectional diagram depi cting the general

PAGE 26

14 morphology of a typical aa flow. The upper vesi cular zone can take up as much as half of the flow thickness in thinner flows and th e flow bottoms are rich in pipe amygdules that often coalesce into inverted Y-shapes (Schmidt, 1990). Pahoehoe flows typically are vesicular throughout the entire thickness. In areas where multiple pahoehoe flows occur in vertical succession, flow boundaries are easil y depicted due to the presence of pipe amygdules at the base and classic ropey pa hoehoe flow-top characte ristics at the top (Figure 2.2). Figure 2.2: Flow boundary between two vesicu lar basalt flows at Site 12. Field book (12x19 cm) shown for scale. White curve superimposed on image highlights the flow boundary. Lower arrow points to pahoehoe structures indicative of the flow top of the lower flow, and th e upper arrow points to the filled pipe amygdules found at the base of the upper aa flow.

PAGE 27

15 Methods Field work was conducted along the North s hore of Lake Superi or in Minnesota during July 15-26, 2004. Field observations we re made at 12 sites (Table 2.1, Figure 2.1). Several sites (3, 4, 6, 11, 12) were se lected for detailed mapping at the outcrop scale using standard geologic methods. Photo-mosaics were acquired with a high resolution digital camera for later analysis. Samples for geochemical and digital analysis were collected from outcrops that were re presentative of the variation in porosity, alteration, mineralogy, and pore types encountered. A suite of samples representative of th e lithologies and mineral parageneses found in the NSVG were selected and used for in depth chemical and digital analysis of the various flow chemistries and alteration historie s. In preparation for bulk rock analysis, twelve samples were powdered and sent to the Geosciences Laboratories of Ontario where major element contents were dete rmined by wavelength dispersive X-ray fluorescence (XRF) with a detection limit of 0.01 wt %. Thin sections also were prepared from each sample for detailed petrographic observations of primary and secondary minerals. Petrographic analysis was perfor med on each thin section to quantify primary porosity and determine mineral parageneses. Mineral chemistry was determined by electron probe microanalysis (EPMA) us ing an automated JEOL 733A electron microprobe operated at 15 kV accelerating poten tial and 15 nA beam current. Calibration was conducted using natural geologic standard s. Beam width for analysis of hydrous minerals (i.e., clays and zeolites) was 10 to 30 m to minimize alkali migration (this instrument is outfitted with wide detector sl its that allow for large spot size analyses). Raw counts were collected for 20 s (approximate ly 60 s total beam contact at each point) and converted to oxide weight percents using the CITZAF correc tion procedure after

PAGE 28

16Table 2.1: Field study locations in the North Shore Volcanic Group. Site Site Description Samples Alteration Mineralogy Zeolite Zone 1 Knife River Wayside N46 56.849', W91 47.505' 01 Prehnite-pumpellyite facies alteration, laumontite float Prehnite-pumpellyite 2 Cut Face Creek outwash at Wayside N47 43.854', W90 26.383' 02 Laumontite fault breccia and stilbitebearing amygdaloidal lava float Laumontite 3 Private shoreline southeast of Thomsonite Beach N47 43.358', W90 27.076' 04-10 Thomsonite + mesolite alteration Thomsonite-Mesolite 4 Cut Face Creek N47 43.887', W90 26.517' 11-12, 14, 3031 Laumontite + chlorite alteration; (also laumontite cemented breccia 2 miles upstream) Laumontite 5 Streambed to Butterwort Cliffs N47 43.266', W90 28.089' 16 Laumontite-cemented fault breccia float Thomsonite-Mesolite 6 Butterwort Cliffs (Cascade River State Park) N47 43.212', W90 28.033' 13B, 15, 17-19 Thomsonite + mesolite + very minor analcime +/calcite alteration Thomsonite-Mesolite 7 Shoreline southeast of Judge Magney State Park N47 48.716', W90 03.999' 03 Silica + clay + laumontite alteration Laumontite 8 Tofte Park N47 34.258', W90 50.389' 20 Hydrothermally a ltered lava with heulandite + scolecite + stilbite + laumontite + thomsonite Thomsonite-Mesolite 9 Lutsen Grandview Park N47 39.568', W90 38.349' 21 Thomsonite + mesolite + analcime Analcime 10 Roadcut along Hwy 61 N47 35.536’, W90 48.071' 22 Andesite (?) with laumontite Thomsonite-Mesolite 11 Gooseberry Falls State Park N47 08.286', W91 27.520' 23-29 Chlorite + calcite + stilbite + Laumontite alteration Laumontite 12 Temperance River State Park N47 33.076', W90 52.560' 32-38 Stilbite + heulandite + quartz + chlorite + thomsonite alteration Stilbite-heulandite

PAGE 29

17 accounting for unanalyzed oxygen fo llowing the methods of Tingle et al. (1996). The analytical conditions and correction procedures employed have previously been shown to provide analyses of zeolites and clays that ar e within error of compositions determined by other methods (Tingle et al. 1996). Results Field observations indicate th at low-grade secondary mineral assemblages are wellpreserved through most of the NSVG. Alterati on mineralogy is easily observed in the field as vesicle infillings in the high porosity zones. Vesicles typically show complete infilling with various zeolite minerals, and, in some cases, are initially lined with clay or silica rims. The lava matrix, in both the hi gh and low porosity zones, appears to have also undergone alteration. This matrix al teration shows some zeo litization, but is dominated by clay and iron oxide alteration minerals. The secondary minerals present in these lavas are similar to t hose previously observed in t holeiitic lava sequences (e.g., Walker, 1960; Neuhoff, 1997, 2000), ranging from lower zeolit e facies through greenschist facies alteration. Primary Basalt Composition and Mineralogy Sampled lavas show compositional variations consistent with pr evious studies of this province (e.g., Schmidt, 1993; Schmidt and Robinson, 1997). Whole rock compositions for selected samples are show n in Table 2.2 and the total alkali-silica diagram of Figure 2.3. Most of the lavas were olivine-normative basalts, with a few from the northeast limb exhibiting more evolved ba saltic andesite and dacitic compositions.

PAGE 30

18Table 2.2: Whole-rock chemical compositions (wt %) of samples. Sample NS04-03 NS04-05 NS04-10 NS04-14B NS04-15 NS04-19 NS04-21 NS04-25 NS04-29 NS04-31 NS04-37 NS04-38 Sitea 8 3 347710 121141313 SiO2 66.87 44.96 45.0555.5145.446.8245.28 43.2948.6646.6246.5944.92 TiO2 1.28 1.14 1.082.171.21.241.08 1.310.741.151.121.07 Al2O3 9.42 16.8 15.512.1515.6715.8816.35 15.3116.0416.861615.73 Fe2O3 b 12.6 9.98 9.6813.4510.0210.569.08 10.976.8610.269.89.53 MnO 0.1 0.13 0.150.240.160.170.14 0.160.170.160.160.14 MgO 1.5 7.48 9.263.588.59.418.4 7.446.877.987.898.31 CaO 2.06 7.9 7.073.488.426.177.92 9.586.829.689.058.83 Na2O 3.59 3.66 3.33.832.584.032.84 2.232.542.272.121.92 K2O 0.49 0.14 0.071.640.730.220.8 0.270.250.280.450.41 P2O5 0.39 0.12 0.110.440.120.120.12 0.140.050.120.130.13 LOI 2.48 8.16 9.033.947.245.928.11 8.6910.554.566.988.86 Total 100.78 100.47 100.3100.43100.04100.54100.12 99.3999.5599.94100.2999.85 aSite descriptions given in Table 2.1, and locations shown on Figure 2.1. bTotal Fe reported as Fe2O3.

PAGE 31

19 Figure 2.3: Total alkalis-silica diagram show ing compositional ranges of NSVG basalts sampled in this study (black circles). Some of the collected samples are more evolved basaltic andesites and dacites, while the majority of samples are olivine-phyric basalts and picrobasalts. Matrix alteration of the primary magmatic phases is observed to varying degrees in all of the analyzed samples. There was no glas s, mesostasis or olivin e observed in any of the samples, most likely due to alteration. Feldspar and pyroxene compositions within the lavas varied throughout each sample, and these variations showed no significant correlation to compositional variations of the lavas. The feldspar compositions (Table 2.3) range from Ab23An72Or5 to Ab99.5An0.3Or0.2. The near-end member albite compositions observed in some plagioclase ph enocrysts suggests th at they have been subject to secondary albitization. Clinopyr oxenes are present, with representative compositions shown in Table 2.4; the absence of orthopyroxenes could also be attributed to the alteration of each sample.

PAGE 32

20Table 2.3: Representative compositions of plagioclase. Sample NS04-03 NS04-29B NS04-05 NS04-19 NS04-10A NS04-25 NS04-38 NS04-31A Oxide weight % SiO2 68.46 69.28 61.7949.7345.7353.8453.2048.62 TiO2 0.04 0.03 0.050.040.020.110.100.10 Al2O3 19.55 19.94 19.0117.9125.1527.7429.6431.79 FeOa 0.05 0.09 0.767.010.330.870.620.75 MnO 0.00 0.01 0.020.160.000.020.000.03 Cr2O3 0.04 0.02 0.040.010.030.050.050.03 MgO 0.00 0.13 2.3611.530.270.360.240.18 CaO 0.06 0.58 1.261.568.1410.9313.0214.66 Na2O 12.28 11.99 9.374.135.895.044.133.12 K2O 0.03 0.11 0.471.430.040.320.120.12 Total 100.50 102.18 95.1493.5285.5999.26101.1299.41 Anhydrous formula unitsb Si 2.98 2.97 2.872.412.422.462.392.24 Ti 0.00 0.00 0.000.000.000.000.000.00 Al 1.00 1.01 1.041.051.571.491.571.73 Fe 0.00 0.00 0.030.330.010.030.020.03 Mn 0.00 0.00 0.000.010.000.000.000.00 Cr 0.00 0.00 0.000.000.000.000.000.00 Mg 0.00 0.01 0.170.970.020.020.020.01 Ca 0.00 0.03 0.060.090.460.540.630.73 Na 1.04 1.00 0.840.420.600.450.360.28 K 0.00 0.01 0.030.010.000.020.010.01 Mole % Ab 99.57 96.80 89.5380.7056.5044.6736.2027.60 An 0.27 2.60 6.8017.2543.2353.4763.1071.65 Or 0.17 0.60 3.682.050.271.870.700.75 aTotal Fe reported as FeO. bBased on 8 oxygen charge equivalents.

PAGE 33

21Table 2.4: Representative compositions of pyroxenes. Sample NS04-10A NS04-10A NS04-14B NS04-15C NS04-15C NS04-37 NS04-37 Oxide weight % SiO2 49.16 49.7650.2751.5352.6651.1350.03 TiO2 1.54 1.930.771.320.671.661.63 Al2O3 4.70 3.731.311.902.504.453.08 FeOa 10.11 10.0721.1013.476.8310.2812.68 MnO 0.22 0.220.500.330.170.200.20 Cr2O3 0.28 0.140.050.020.600.260.22 MgO 13.40 14.1610.3815.6716.4212.9612.17 CaO 18.85 19.7615.8216.0920.8320.7920.18 Na2O 0.52 0.350.220.330.300.320.32 K2O 0.02 0.010.010.030.050.000.01 Total 98.79 100.14100.43100.69101.04102.05100.53 Anhydrous formula unitsb Si 3.72 3.723.903.843.843.753.77Ti 0.09 0.110.040.070.040.090.09Al 0.42 0.330.120.170.220.380.27Fe 0.64 0.631.370.840.420.630.80Mn 0.01 0.010.030.020.010.010.01Cr 0.02 0.010.000.000.030.010.01Mg 1.51 1.581.201.741.791.421.37Ca 1.53 1.581.311.291.631.631.63Na 0.08 0.050.030.050.040.050.05K 0.00 0.000.000.000.000.000.00 Mole % Wo 41.60 41.7733.8733.2342.5344.4042.93 En 41.03 41.6330.9045.0746.6038.5036.03 Fs 17.37 16.6335.2321.7310.8717.1321.03 aTotal Fe reported as FeO. bBased on 12 oxygen charge equivalents.

PAGE 34

22 Regional Alteration Mineralogy Although flow morphologies and textures are well-preserved throughout the study area, all investigated lava fl ows exhibit some degree of secondary mineral alteration. In contrast to typical alterati on patterns in younger very low-gr ade metabasalts, alteration is present even in areas away from zones of hi gh primary porosity and is clearly observed in hand specimen. Within the matrix, the primary magmatic phases are usually replaced by secondary alteration minerals such as Fe(Ti) oxides, mafic phyllosi licates, zeolites, and occasionally silica minerals and calcite. The alteration mineralogy observed within vesicles is dominated by zeolite minerals, with occasional linings of silica and mafic phyllosilicates and overp rinting by calcite. The extent of chemical alteration within th e matrix is typically greater within zones of high primary porosity (e.g., vesicular flow tops and bottoms) than in massive flow centers. In some samples there is a strong presence of sub-mm sized spots of bright red Fe(III) oxides (probably hematite), and in most cases oxide alteration occurs further away from amygdules than the mafic phyllosilicates and zeolites. Mafi c phyllosilicates often are pervasive around vesicles as replacements of mesostasis, glass, and olivine, and also are commonly found lining pore walls. Plagio clase is variably replaced in vesicular zones by either near-end member albite or zeolites, and zeolite alteration within the matrix dominantly occurs in close proximity to amygdules also filled with zeolites. Often the only phenocryst phase that escapes altera tion in these zones is clinopyroxene. The overall extent of this alteration appears to be a function of distance from primary porosity, and in some samples (Figure 2.4) light -colored reaction aure oles are observed as bleached haloes (probably due to oxidation reactions) around and between primary pores within the rock.

PAGE 35

23 Figure 2.4: Field photo from Site 3 (Table 2.1) showing an extensively altered pahoehoe flow. Visible within the flow are bl eached haloes around vesicles connected by thin anastomosing bleached areas. Pencil shown for scale. Primary vesicles, especially in zones of relatively hi gh porosity at the tops and bottoms of flows, are generally completely f illed with zeolites and occasional linings of silica and mafic phyllosilicate minerals. Some vesicles undergo complete infilling with only one zeolite mineral, as shown by the thoms onite filled vesicle in Figure 2.5A. Many other vesicles show multiple st ages of zeolite alteration and contain two or more zeolite minerals. Figure 2.5B shows a vesicle fille d with thomsonite and mesolite, and the vesicle in Figure 2.5C contains analcime, thom sonite and mesolite. Some vesicle walls are rimmed by secondary minerals prior to zeolit e infilling. Silica mine rals are the first to form where present, as shown in Figure 2.5D This amygdule exhibits early lining of silica, followed by a thin clay layer before co mplete infilling of laumontite. Most often the vesicle walls are lined only with mafic phyl losilicates prior to zeolite infilling, as shown in Figure 2.5E where chlorite rims precede laumontite infilling.

PAGE 36

24 Figure 2.5: Secondary alteration in NSVG la vas. All photomicrographs were taken through partially crossed polars. A) Photomicrograph showing thomsonite alteration in sample NS04-31, collected directly above the sediment pile roadcut from Site 4 (Table 2.1). B) Photomicrograph of sample NS04-15 showing the thomsonite-mesolite alteration typical of Sites 3 and 6. C) Scanned image of thin section of sample NS04-15 showing analcimethomsonite (thom.)-mesolite alteration. D) Photomicrograph of sample NS0403 showing laumontite alteration typical of sample sites 4 and 7 with clearlydefined early silica and clay (rims on vesi cles. E) Photomicrograph of sample NS04-14 showing laumontite alteration with initial formation of chlorite rims. F) Photomicrograph of sample NS04-10A highlighting calcite crystallization in thomsonite filled amygdule.

PAGE 37

25 Figure 2.5: (continued) Quartz and calcite are widespread thr oughout the study area, though not necessarily pervasive. Local, late-stage alteration that cross-cuts and, in some cases overprints these assemblages, is observed around veins and faults Veins in these lo calities often contain quartz and calcite, along with laumontite, stilb ite, and/or heulandite. Silica occasionally is present as initial pore linings, but most of ten is present within the matrix, while calcite dominantly is found as post-zeolite inf illings of amygdules. Figure 2.5D shows an amygdule exhibiting early linings along the pore wall of quartz and interlayered chlorite/smectite, followed by complete zeoli te infilling of residual pore space. The progression of mineralization within this depict ed pore is typical of pores exhibiting silica linings. Figure 2.5F shows an amygdule filled with thomsonite that has experienced late calcite overprinting. The amygdules within a given sample that have experienced this late calcite overprinting appear to be randomly distributed. Four distinct zeolite asse mblages are observed within the lavas of the northeast limb of the study area. Mineral zones containi ng these assemblages are identified based on the presence of the index minerals analcime, thomsonite-mesolite, stilbite-heulandite, and laumontite within vesicles. In all cas es, these zones are present in contiguous

PAGE 38

26 sections of the NSVG, and, except for the st ilbite-heulandite zone are present in both limbs. Most of the isograds bounding thes e zones are not well-exposed, and their locations are approximate. Alteration higher in metamorphic grade than zeolite facies is present only in the very s outhern portion of the field area where alteration reached prehnite-pumpellyite and greenschist facies, and has already been discussed in detail by Schmidt (1993) and Schmidt and Robinson (1997). The mineralogy of the four distinct zeolite zones is described below: Laumontite zone lavas Laumontite zone alteration is the highest zeolite grade alteration observed in the NSVG lavas, and is characterized by the presen ce of quartz, interlayered chlorite/smectite (c/s), and laumontite. The alteration zone was observed in two locations along the shoreline, from just south of Grand Marais to Grand Portage and from Little Marais to just south of Two Harbors. Albitized plag ioclase laths and altered mesostasis dominate the matrix, and there are very minor am ounts of unaltered pyroxene phenocrysts. Alteration affects most of the matrix and consists of c/s, quartz and iron oxides. Quartz rims within amygdules are occasiona lly present where they precede rims of C/S (Figure 2.5D). Occasionally, C/S is visi ble in hand sample, and appears in thin section as a green fibrous mineral within vesicl es. Smaller vesicles within these lavas are completely filled with c/s, while larger vesicles are rimmed with C/S and later filled with laumontite. Laumontite appears white in hand sample, and in thin section appears either blocky or fibrous depending on th e orientation of the thin se ction to the fibrous mineral growth habit. Overprinting of laumontite w ith calcite is occasionally observed in the vesicles. The calcite exhibits twinning and, in most cases rhombohedral crystal habit.

PAGE 39

27 Stilbite-heulandite zone lavas Stilbite-heulandite zone alteration is rec ognized in the NSVG lavas by the presence of amygdules filled with stilbit e and/or heulandite. Unlike the other zeolite zones present along the NSVG, this zone was not found in the northeast limb. During this study it was only observed once along the southwest limb between the towns Little Marais and Tofte (and was observed by Schmidt ( 1993) in several local ities in this area). Iron oxides, chlorite, albitized plagiocl ase laths and minor amounts of unaltered pyroxene phenocrysts dominate the lava matrix. Trioctahedral smectite is pres ent either as rims along vesicle walls or as complete infilling of vesicles. Smaller vesicles are co mpletely filled with trioctahedral smectite, while in larger vesicles it is present as rims with the remaining open space filled with calcite. The calcite exhibits twinning and in some instances rhombohedral cleavage. It is possible that calcite occu rs as overprinting of the original zeolites that filled the vesicles in these analyzed samples, as stilbite and heul andite were observed as vesicle infillings in the field. Thomsonite-mesolite zone lavas Lavas exhibiting amygdules filled with t homsonite, mesolite and C/S interlayered clays comprise this alteration zone observed in the NSVG. This zone of alteration occurred to the northeast and to the sout hwest of Lutsen, bounding the analcime zone discussed below. Pyroxene phenocrysts, pl agioclase laths, and iron oxides dominate the lava matrix surrounding the vesicles. The pyroxene phe nocrysts often are unaltered, whereas plagioclase is typically replaced by zeol ites similar to those observed in vesicles. Amygdule minerals in hand samples are occasio nally green in color, but are dominantly

PAGE 40

28 either pink or white with dark-pink color ba nds. In thin section, the minerals appear colorless to white, with brown staining some times present around the rim of the vesicle. There are two dominant crystallization ha bits of thomsonite and mesolite in the vesicles of this zone. In vesicles where th ere is a prevalent crystallization sequence of early thomsonite pore rims followed by late mesolite filling of remaining pore space, the thomsonite occurs as a massive growth of many small fibrous bundles rimming the vesicle wall, or in some rare occurrences comp letely filling the vesicle. Mesolite appears either massive with no crystal habit, or fi brous and fills any remaining space after the thomsonite rims in vesicles. The second domina nt crystallization hab it originates at some apparently random nucleation point along the vesicle wall and grow s radially outward until the vesicle is completely filled (Figur e 2.5B). The growth sequence is usually continuous from thomsonite originating at th e nucleation site transi tioning into mesolite with no visible boundary between the two; in amygdules where there is a clearly defined boundary between the thomsonite and mesolite one, if not both minerals, exhibit a fibrous growth habit. In all vesicles obser ved in this zone, the relative amounts of these two minerals vary inconsistently. Also, clay s are not present as ri ms along the vesicle wall, but they do occasionally exist as interg rowths of interlayered chlorite-smectites within the zeolite minerals, and in a ma jority of the amygdules, there is a late overprinting of calcite, with crystals exhibiting twinni ng and rhombohedral cleavage (Figure 2.5F). Analcime zone lavas This alteration zone in the NSVG lavas is characterized by the presence of thomsonite and analcime within vesicles. Th is is the lowest grade of alteration observed in the field area, and is found only in the vicinity of Lutsen. Unaltered pyroxene

PAGE 41

29 phenocrysts, albitized plagioclase laths, and iron oxides, with occasional zones of thomsonite alteration, dominate the lava matr ix. No visible clay or silica rims were observed lining vesicle walls. The amygdule mine rals appear pink to colorless in hand sample and clear in thin section. There is usually a sequence of mineralization from thomsonite rimming the vesicle to analcime growth in the residual open space. The thomsonite crystallization habit ranges from a massive growth of many small fibrous bundles rimming the pore, to one or two larg e fibrous bundles individually filling up to half of the vesicle, while the analcime exhi bits a blocky habit. The relative amount of each mineral varies between vesicles, and in some instances the crystallization sequence is ambiguous. Alteration Mineral Chemistry Representative compositions of analyzed mafic phyllosilicates are listed in Table 2.5 and are plotted in Figure 2.6 in terms of the sum of Si + Al + Mg + Fe and the interlayer charge (2 Ca + Na + K). The bulk of the samples have compositions intermediate between trioctahedral smectites and chlorites, suggesting that they are interlayered chlorite/smect ite phases. These samples are found in the thomsonitemesolite and laumontite alteration zones. Samples representative of the stilbiteheulandite alteration zone are dominantly trioctahedral smectites, although some compositions are mixtures of dioctahedral and trioctahedral smectites. One last group exhibits compositions with (2Ca + Na + K) greater than 1.5 per 28 O equivalents and lie along an extension of the trioctahedral smectite-chlorite mixing trend. These samples are mainly from the thomsonite-mesolite alterati on zone, and may be al tered celadonite (c.f. Neuhoff et al., 1999).

PAGE 42

30 Table 2.5: Representative compos itions of mafic phyllosilicates. Sample NS04-03 NS04-19 NS04-25 NS04-37 NS04-37 NS04-37 NS04-38 Oxide weight % SiO2 36.40 33.3947.995.6249.4557.85 50.27 TiO2 0.47 0.000.0419.820.010.03 0.02 Al2O3 12.61 15.939.212.338.0823.47 8.05 FeOa 10.69 12.362.4858.960.790.55 0.69 MnO 0.25 0.460.150.320.180.07 0.19 Cr2O3 0.01 0.010.040.040.020.05 0.04 MgO 18.24 24.1924.302.5128.046.58 28.33 CaO 1.90 1.213.901.113.453.68 3.26 Na2O 0.15 0.040.050.570.040.05 0.05 K2O 0.08 0.010.050.060.010.12 0.01 Total 80.80 87.5988.2291.3490.0792.45 90.92 Anhydrous formula unitsb Si 7.58 6.538.611.568.689.44 8.72 Ti 0.07 0.000.014.170.000.00 0.00 Al 3.10 3.671.960.761.674.51 1.65 Fe 1.86 2.020.3813.770.120.07 0.10 Mn 0.04 0.080.020.080.030.01 0.03 Cr 0.00 0.000.010.010.000.01 0.00 Mg 5.67 7.066.641.047.331.60 7.33 Ca 0.42 0.250.760.330.650.64 0.61 Na 0.06 0.010.020.310.010.01 0.02 K 0.02 0.000.010.020.000.03 0.00 SAMFc 18.21 19.2917.5917.1317.8015.62 17.80 2Ca+Na+K 0.93 0.521.550.991.321.33 1.23 aTotal Fe reported as FeO. bBased on 28 oxygen charge equivalents. cSum of Si + Al + Mg + Fe in formula unit. Zeolite compositions generally agree we ll with previous reported values. Representative thomsonite compositions are li sted in Table 2.6. Thomsonites span the entire compositional range found in this mineral (between Ca1.5Na1.5Al4.5Si5.5O20 n H2O and Ca2NaAl5Si5O20 n H2O; Ross et al. 1992; Neuhoff and Ruhl, 2006; Figure 2.7) with some samples exhibiting compositions less Ca-Al -rich than previously observed in this

PAGE 43

31 Figure 2.6: Compositions of ma fic phyllosilicates formed during regional metamorphism of the NSVG as a function of the numbe r of non-interlayer cations (Si + Al + Mg + Fe) versus the interlayer charge (2 Ca + Na + K). Sample compositions were normalized to 28 O charge equivale nts for comparison. Gray areas show the positions of ideal endmember dioctahedral and trioctahedral smectites and chlorite. The variation in zeolite zone alteration with rela tion to composition is depicted by the black circles (thomsonite-mesolite), blue squares (stilbiteheulandite), or red triangles (laumontite). mineral. Mesolite and laumontite (Table 2.7) are essentially stoichiometric. Analcime compositions have Si/Al ratios close to 2.0 (T able 2.7), similar to other occurrences in low grade metabasalts (e.g., Passaglia and Sheppard, 2001; Neuhoff et al. 2006). Discussion Regional variation of metamorphic grade One of the difficulties encountered in asse ssing metamorphic grade in continental flood basalts is that the lavas are frequen tly overprinted by smal l-scale hydrothermal systems associated with local intrusions and fa ults. Indeed, a number of sites visited in

PAGE 44

32 Table 2.6: Representative compositions of thomsonites. Sample NS04-05 NS04-10A NS04-15C NS04-21 NS04-31A Oxide weight % SiO2 35.7537.0939.2537.02 43.35 TiO2 0.060.010.020.04 0.00 Al2O3 30.0730.6630.6930.84 28.01 FeOa 0.070.210.180.08 0.00 MnO 0.000.000.020.01 0.01 Cr2O3 0.000.020.040.02 0.00 MgO 0.220.470.000.21 0.00 CaO 12.9613.2212.5313.03 10.48 Na2O 4.003.784.444.13 5.44 K2O 0.010.020.040.03 0.01 Total 83.1485.4887.1985.41 87.31 Anhydrous formula unitsb Si 5.005.045.205.03 5.68 Ti 0.010.000.000.01 0.00 Al 4.964.914.794.94 4.32 Fe 0.010.020.020.01 0.00 Mn 0.000.000.000.00 0.00 Cr 0.000.000.010.00 0.00 Mg 0.050.100.000.04 0.00 Ca 1.941.921.781.90 1.47 Na 1.080.991.141.09 1.38 K 0.000.000.010.01 0.00 Si/Al 1.011.031.091.02 1.31 Ca/(Ca+Na+K) 0.640.660.610.63 0.52 Na/(Ca+Na+K) 0.360.340.390.36 0.48 aTotal Fe reported as FeO. bBased on 20 oxygen charge equivalents. this study (e.g., Sites 3, 4, 8, and 10) cont ain well developed laumontite + quartz alteration associated with small zones of brittle deformation that are likely small fault zones. At Site 8 (Tofte Park), the aa and pahoehoe flows expos ed along the coastline exhibit a severely over-constrained mineral assemblage of thoms onite + scolecite + stilbite + heulandite + laumontite + quartz. This assemblage is well-developed in other large igneous provinces, for instan ce in eastern Iceland (Neuhoff et al. 1999), where it is always associated with local hydrothermal alte ration. Phase rule anal ysis indicates that it is over-constrained. The mi nerals comprising this a ssemblage probably did not

PAGE 45

33 crystallize together, but rather sequentially in response to changes in temperature, pressure, and/or fluid compositi on. The flows at Site 8 are cr osscut by zeolite-filled veins that were likely major fluid conduits during hydr othermal alteration of this outcrop. Only in relatively isolated vesicular zones in th e center of the aa flows is the likely regional metamorphic assemblage present where the la vas were able to es cape later alteration. Therefore, observations of metamorphism in the lava flows of the NSVG are based on sample sites where regional metamor phic grade was not overprinted by local hydrothermal alteration. Figure 2.7: Plot showing the compositional va riation of analyzed thomsonites in NSVG basalts (black circles). Thomsonite solid solution (dashed line) between Ca1.5Na1.5Al4.5Si5.5O20 n H2O and Ca2NaAl5Si5O20 n H2O endmember compositions (blue squares) also shown (Rose et al. 1992; Neuhoff and Ruhl, 2006).

PAGE 46

34Table 2.7: Representative compositions of mesolite, analcime, and laumontite. Sample NS04-05 NS04-10A NS04-15A NS04-15C NS04-21 NS04-14B NS04-03 Mineral Mesolite Mesolite Mesolite Analcime Analcime Laumontite Laumontite Oxide weight % SiO2 43.79 44.4346.4253.5855.5651.5350.64 TiO2 0.03 0.010.030.050.070.010.01 Al2O3 25.48 25.5627.0922.9422.3622.2222.03 FeOa 0.00 0.010.100.000.000.090.04 MnO 0.00 0.000.000.000.020.030.00 Cr2O3 0.00 0.040.040.040.080.030.03 MgO 0.00 0.000.000.000.000.000.00 CaO 8.47 8.569.920.370.3512.0611.88 Na2O 5.99 5.835.5613.5613.220.070.01 K2O 0.08 0.050.030.090.010.090.28 Total 83.83 84.4889.1890.6391.6786.1284.93 Anhydrous formula units Si 2.97 2.982.961.992.033.983.96 Ti 0.00 0.000.000.000.000.000.00 Al 2.03 2.022.031.010.962.022.03 Fe 0.00 0.000.000.000.000.010.00 Mn 0.00 0.000.000.000.000.000.00 Cr 0.00 0.000.000.000.000.000.00 Mg 0.00 0.000.000.000.000.000.00 Ca 0.61 0.620.680.010.011.001.00 Na 0.79 0.760.690.980.940.010.00 K 0.01 0.000.000.000.000.010.03 Ob 10 1010661212 Si/Al 1.46 1.471.451.982.111.971.95 Ca/(Ca+Na+K) 0.44 0.450.500.010.010.980.97 Na/(Ca+Na+K) 0.56 0.550.500.980.980.010.00 aTotal Fe reported as FeO. bNumber of framework oxygens in formula unit.

PAGE 47

35 Figures 2.8 and 2.9 synthesize the observat ions of regional metamorphic grade from this study and those of Schmidt (1993) in terms of their spatial and stratigraphic distributions, respectively. Many of the lo calities studied by Sc hmidt (1993) were Figure 2.8: Generalized map of northeastern Minnesota showing the distribution of NSVG lavas (gray), the locations listed in Table 2.1, and interpretations of metamorphic grade based on this study and the work of Schmidt (1993) and Schmidt and Robinson (1997). Metamo rphic mineral zones correspond to standard metamorphic facies desi gnations (prehnite-pumpellyite and greenschist). Inferred boundaries be tween mineral zones are shown by the dashed lines.

PAGE 48

36 Figure 2.9: Stratigraphy of the NS VG (Vervoort and Green, 1997; Miller et al. 2002) correlated along the fold axis dividing the lavas into the southwest and northeast limb. Numbers in stratigraphic columns refer to sampling sites in this study (Figure 2.1, Table 2.1). Mine ral zones based on interpretations from this study and Schmidt (1993) and Schmidt and Robinson (1997). revisited in this study, and in some cases reinterpreted in light of the complex metamorphic history noted above. Along the southern limb, the transition from the laumontite zone to the higher grade prehnite-pumpellyite zone is placed at Site 1 where primary pore spaces in adjoining lava flows indicate a change in mineral assemblage. The laumontite zone continues up through th e stratigraphy for about 2700 meters until observations of stilbite and heulandite at Si te 12 suggest a transition to the lower-grade the stilbite-heulandite zone. Th e onset of this alteration coin cides with the basal contact of the Schroeder basalts along th e southern limb, which shows a stratigraphic thickness of ~900 meters assuming the hinge ax is as the upper extent of the section. The lower half of the Schroeder basalts contains the stilbite-heulandite zone, while the upper half of this

PAGE 49

37 section shows thomsonite-mesolite zone altera tion. This transition implies a correlation between shallower stratigraphic dept h and lower grades of alteration. The lowest grade of altera tion noted during this study o ccurs at Site 9 (Lutsen Grandview Park) and corresponds to the lowerm ost analcime zone. This outcrop consists of a series of vesicular pahoehoe flows cont aining abundant thomsonite alteration. The exact extent of the analcime mineral zone is unclear due to limited exposures, though what is exposed shows no signi ficant stratigraphic thickness. Typically in large igneous provinces there is a well-developed chabazi te-thomsonite zone overlying the analcime zone (e.g., Walker, 1960; Kristma nnsdttir and Tmasson, 1978; Larsen et al. 1989; Neuhoff et al. 1997, 2000, 2006) that is frequently as thick as the analcime, mesolitescolecite (likely of equivalent grade to the thomsonite-scole cite zone observed in the NSVG), and stilbite-heulandite zones combined. The absence of this zone, together with the limited extent of the analcime zone and the thicknesses of the thomsonite-mesolite an stilbite-heulandite zones (Figure 2.9), imp lies that approximately 1000-2000 meters of the volcanic stratigraphy may have been lost to erosion in the NSVG. The variation in metamorphic grade in th e northeast limb of the NSVG appears to mirror that in the southwest limb, where meta morphic grade increases with stratigraphic depth. The Lutsen basalts exposed along th e northeast limb of the NSVG are thinner (~450 meters) and contain well-developed thom sonite-mesolite zone alteration down to the lower stratigraphic boundary with the Cut Fa ce Creek sediments. This sediment pile is ~20 meters thick, and the andesites di rectly below them show laumontite zone alteration. Limited accessibili ty precluded assessmen t of metamorphic grade northeast of locality 7, and thus the total thickness of the laumontite zone in this limb is

PAGE 50

38 unconstrained. The most striking difference between the two limbs is the absence of stilbite-heulandite zone alteration in the northeast limb, discussed below. Conditions of Alteration Observed mineral assemblages of the NSVG are comparable to similar alteration of the lavas in East Greenland, eastern Icela nd, and the Icelandic geothermal systems (Walker, 1960; Kristmannsdttir and Tmasson, 1978; Larsen et al. 1989; Neuhoff et al. 1997, 2000). The distribution of zeolite minera ls with depth/temperature in basaltic terrains of Iceland and East Greenland are summarized in Figure 2.10 (Neuhoff et al. 1997; 1999; 2000). The minerals observed to de velop within each zeolite zone are shown with respect to temperature constraints of th e mineral distributions typically observed in tholeiitic lava flows. Each mineral zone forms during alteration, tr ansitioning to the next with increased temperature/depth. It has been observed that thick sequences of lavas with varying chemistries exposed to the same alteration conditions will ofte n exhibit substantially different mineral parageneses (e.g., Walker, 1960a; Robert, 2001; Neuhoff et al. 2006), causing a need for observation of rock type along with alteration minerals present. The Schroeder-Lutsen basalts of the NSVG are lower in silica than those of the Greenland and Iceland provinces showing similar mineralogy, which may stabilize thomsonite relative to more silica-rich minerals like mesolite and scolecite (Neuhoff et al. 2006). Therefore, the zone in Figure 2.10 labeled as the mesolite-scolecite zone is in terpreted to be equivalent in grade to the thomsonite-mesolite zone.

PAGE 51

39 Figure 2.10: Schematic diagram showing the di stribution of minerals and mineral zones typically developed during very low gr ade metamorphism of large igneous provinces (after Walker, 1960; Neuhoff et al. 1997, 2000). Frequency of mineral occurrence with depth is denoted by solid (ubiquitous) and dashed (occasionally present) lines. Mineral list is non-exclusive. Temperatures listed on left side of diagram are ba sed on observations by Kristmannsdttir and Tmasson (1978) of the distributi on of minerals as a function of temperature in Icelandic geothermal systems. Figure 2.9 shows the estimated stratig raphic thicknesses given by Miller et al. (2002) combined with my observations of re gional metamorphic grade boundaries with stratigraphic depth for the NSVG lavas. Co mparisons of these observations to the observations in Figure 2.10 can be used to in terpret the thermal and structural history experienced after emplacement (Neuhoff et al. 2000). Observations show zeolite to

PAGE 52

40 greenschist facies alteration along the expos ed stratigraphy of the NSVG (Schmidt and Robinson, 1997; Neuhoff et al. 2000), where alteration grade increases with stratigraphic depth along the exposure. Due to the synclin al structure of the region, the youngest lavas exposing the lowest grade of alteration are f ound near the hinge axis. This fold axis divides the region into two limbs, and stratig raphic depth as well as alteration grade increases to the southwest and northeast. Analcime zone alteration, the lowest gr ade of alteration observed within the NSVG basalts, occurs near the hinge axis separati ng the Schroeder basalts in the southwest limb from the Lutsen basalts in the northeast lim b. There is no significant thickness to this zone, though observations of analcime alterati on in similar provinces predict relatively consistent stratigraphic thicknesses of each a lteration zones. The limited thickness of the analcime zone, as well as the absence of a lteration minerals exhibi ting temperatures and pressures lower than analcime in grade, is likely due to erosion of once exposed stratigraphically younger lavas. Stratigraphically below the analcime zone is the thomsonite-mesolite zone of alteration, exhibiting a consistent thickness in each limb of approximately 500 m. The lower thomsonite-mesolite boundary in the northe ast limb is in direct contact with the higher grade laumontite zone, while in the so uthwest limb there is a 400 m thick stilbiteheulandite zone of alteration separating the thomsonite-mesolite zone from the laumontite zone. The thickness of the laumontite zone in the northeast limb is at least 1500 m, though the exact extent is unknown due to inadeq uate exposures further to the northeast. Assuming no structural complicatio ns, it can be expected to be as thick as the laumontite zone in the southwest limb, which is 2700 m thick and bounded below by the prehnite-

PAGE 53

41 pumpellyite zone. The prehnite-pumpellyite zone is the highest grade of alteration observed in the NSVG basalts, with the onset of alteration occurring at a stratigraphic depth of 3,860 m below the fold axis. The th ickness of this zone is unknown due to a lack of observations in outcrops further to the south. The boundary constraints defined by the mappe d zeolite zones can be correlated to the thermal boundaries corresponding to mine ral isograds in Icelandic geothermal systems (Figure 2.10) to estimate average ge othermal gradients during alteration. The well-constrained isograd locations bounding the thomsonite-mesolite and stilbiteheulandite zones in the southwest limb gi ve a combined thickness of 900 meters, and indicate a temperature range of ~40 C. This implies a geot hermal gradient of the NSVG region during alteration of approximately 42 C/km, which compares well to the estimated geothermal gradients of 55 C/km observed in LIPS formed during continental rifting in eastern Iceland (Neuhoff et al. 1999) and 40 C/km in East Greenland (Neuhoff et al. 2000). This information also can be used to cal culate overburden pressure and heat flow experienced during the time of alteration. The estimated stratigraphic thickness of the basalts experiencing zeolitiz ation during the time of a lteration is ~5500 m after combining the known stratigraphy with the estima ted loss due to erosi on. Pressures thus increased from a minimum of 1 bar (assuming atmospheric pressure) to maximum pressures ranging from ~550 to ~1600 bars, corresponding to hydrostatic and lithostatic pressures, respectively. Regional heat flow over the area during alte ration was about 1.32.2 heat flow units, given a thermal conductiv ity for water-saturated basalt of 1.5-2.0 W/mK (Oxburgh and Argell, 1982) and the imp lied geothermal gradient for the NSVG of

PAGE 54

42 42 C/km. These values are comparable to heat flow in presentday continental rift systems (Lysak, 1992). Structural Interpretations The overall structure of the NSVG is a plungi ng synclinal fold with a hinge axis in the area between Tofte and Lutsen (e.g. Miller et al. 2002). This fold has been interpreted as a subsidence feature formed during eruption of the lava pile due to evacuation of the underlying magma chamber a nd increased density of the growing lava pile (e.g., Ojakangas and Matsch, 1982), or alte rnatively as a late st ructural modification of the area due to strong influences from th e nearby Grenville orogen (Van Schmus and Hinze, 1985). The observation that metamorphic grade in creases with stratigraphic depth, in both limbs of the province, has considerable impli cations for the structural development of the NSVG. Based on my field observations comb ined with those of Schmidt (1990), the distribution of secondary mine ral paragenesis represented in the NSVG suggests that the metamorphic zones are folded along with the lavas. Regional metamorphism in large igneous provinces typically is initiated soon after volcanism and is of short duration (e.g., < 1 Ma; Neuhoff et al. 1997). Thus, if the folding of the NSVG was syn-volcanic, one would expect the mineral zones to cross-cu t the stratigraphy, whereas the isograds would be folded along with the lavas if deformati on were due to post-volcanic processes. Combining this information with the obser vation that the regional metamorphism is typically rapid and of short dur ation after complete eruption of the lava pile in large igneous provinces (e.g. Neuhoff et al. 1997, 1999, 2000), it appears that folding of the NSVG is a post-volcanic feature and was not a mechanism of crus tal accommodation of the immense lava pile developed in this province.

PAGE 55

43 The stratigraphic distributions of zeolite zones in th e NSVG (Figure 2.9) are not symmetrical between limbs, as indicated by the absence of the stilbite -heulandite zone in the northwest limb of the province. Previous observations of zeolite zones in other continental flood basalt provinces have shown that these zones are relatively continuous and uniform in thickness (Walker, 1960b; Neuhoff et al. 2000). Correlation of the overlying thomsonite-mesolite zone between the northeast and s outhwest limb show continuous and uniform thickness, implying that this should be the case in the continental flood basalts of the NSVG. The stilbite-heu landite zone, however, is mapped in the southwest limb with a thickness of almost 450m but is not present in the northeast limb. The expected position of the stilbite-heulandite zone in the northeast limb corresponds to the location of the Cutface Creek sediments, which are only 30-40 m thick and thus not extensive enough to contain the w hole range of stratigraphic depth expected for this zone. Two possible explanations might account for the lack of the stilb ite-heulandite zone in the northeast limb. The first is based on the fact that there is a marked change in lava composition between the thomsonite-mesolite zone occurrences at site 3 (and immediately overlying the Cutface Creek sediment s) and the laumontite zone lavas at site 4, with the latter exhibiting more silicic ba saltic andesite compositions (e.g., Walker, 1960a; Kristmannsdttir and Tmasson, 1978; Murata et al. 1987; Hearn et al. 1989; Robert, 2001; Neuhoff et al. 2006). Potentially this cha nge in lava composition may result in laumontite occurring at lower grades This should, however, stabilize the more Si-rich minerals stilbite and heulandite re lative to laumontite, not the other way around. A more plausible hypothesis is that there is a loss of stratigraphy in the region between sites 3 and 4. High-angle reverse faults are observed in the central and western regions of

PAGE 56

44 this area due to compression, causing juxtaposit ion of older rocks ne xt to younger rocks (e.g., White, 1966; Van Schmus and Hinze, 1985). Reverse faulting may have also occurred in the eastern region of the NSVG, a nd could explain the abse nce of the stilbiteheulandite zeolite zone in the northeast limb. The potential displacement of this alteration zone therefore occurr ed as relatively late faulti ng of the NSVG basalts, which agrees with the suggested post-volcanic compression.

PAGE 57

45 CHAPTER 3 PETROGRAPHIC AND DIGITAL ANALYS IS OF POROSITY EVOLUTION DURING ALTERATION Introduction Permeability in basaltic lavas has received considerable attention recently due to increased focus on their hydrologic properties an d the role of degassi ng of volatiles in the dynamics of eruptive events (e.g., Fisher, 1998; Saar and Manga, 1999; Blower, 2001a, b; Srouga, 2004; Stimac, 2004). From this work, it is becoming increasi ngly clear that the geometry and connectivity of the pore netw ork in vesicular lava s is fundamentally different than in more commonly considered porous media such as sediments and fractured rocks, and more attention is bei ng placed on the relationship between porosity and permeability. For lavas with porosities >60%, there appears to be no significant relationship between porosity and perm eability; when porosities are below 60%, however, experimental results from different studies show varying porosity-permeability relationships (Eichelberger et al. 1986; Klug and Cashma n, 1996; Saar and Manga, 1999). In vesicular basalts the permeability can be as high as 10-7 cm2, which is a value comparable to clean unconsolidated sand or karstic limestone. In basalts where the vesicles are small and closely spaced, there wi ll be an increased probability of connected fluid pathways developing between pores that will likely lead to these higher permeabilities (e.g., Saar and Manga, 1999). Pe rcolation theory (Lee, 1990; Sahimi, 1994; Grimmett, 1999) applied to basalts in experimental studies shows that the

PAGE 58

46 percolation threshold, the cri tical porosity below which perc olating clusters of bubbles should not exist, is around 30%, implying that little to no permeability can be expected below critical porosity. In th ese low porosity basalts, the vesi cles become more isolated. If connectivity between pores is dependant on fluid flow through the matrix, permeability can be as low as 10-18 cm2 (Johnson, 1980; Saar and Manga 1999). Multiple studies have shown, however, that vesicular lavas retain relatively high permeabilities (>10-8 cm2) down to porosities as low as a few percent (Feng et al. 1987; Sahimi, 1994, 1995; Saar and Manga, 1999; Blower, 2001b). This does not follow the predicted model for granular rocks, and implies that there must be pathways forming within the matrix allowing fluid to flow at a faster rate. Differences in porosity within basaltic la va flows leads to considerable flow-scale variations responsible for th e development of secondary mi neral paragenesis (Schmidt, 1990; Schmidt and Robinson, 1997). Under low-grade alteration, the chemical components necessary to precipitate the ve sicle-filling minerals are derived from a spatially restricted region (a lteration halos) around the ve sicle and the fluid pathways over which diffusion occurs. This suggests that the presence of mine ral infillings can be a reliable indicator of pores being part of a permeable network through the lava, as fluid flow is inherently coupled to changes in porosity and permeability brought about by mineralogic alteration. This ch apter explores using mineralogi c alteration to trace fluid pathways through the lavas and the tempor al evolution of pore space through time. These observations will prove important not only for further understanding of groundwater and petroleum reservoirs in ba salts, but also developing new methods to monitor the controls and affects of fluid fl ow and alteration within vesicular rocks.

PAGE 59

47 Methods Field Techniques Field work was conducted along the north s hore of Lake Superior in Minnesota during July 15-26, 2004. Of the sampling site s and methods discussed in Chapter 2, measurements were performed at sites 3, 4 and 6 (Table 2.1, Figure 2.1) of total initial porosity, pore size distributions, and the amount and type of secondary alteration within primary pore space and within the matrix. Measurements of total poro sity were made both on the thin section scale through digital analysis as well as on the outcrop s cale in the field. All porosity in analyzed samples is essentially zero, due to the comp lete infilling of all vesicles by secondary alteration, so primary porosity is analyzed base d on the initial void spac e in vesicles prior to alteration. In situ field measurements of primary macr oscopic porosity (defined in this paper as the percentage of a given volume of rock occupied by pore space) were made in the field following the methods of Manning a nd Bird (1991). This is done by placing a transparent mylar sheet embosse d with a precision drafted gr id (1073 nodes) directly on the outcrop and counting the number of nodes on th e grid that overlay pores. Porosity was taken to be the percentage of node s overlying pores (Manning and Bird, 1991). Measurements were made to assess the pore size distributions within sampled lava flows. Overall pore size distributions were assessed by measuring the maximum and minimum diameter of all pores within a re presentative area. Notes were made on the progress of reaction of each measured pore by assessing the level of secondary alteration observed with respect to pore size. Observ ations also were made of the amount of alteration visible in the matrix around and between pores to assess the characteristics of

PAGE 60

48 fluid flow through the rock as well as the amount of lava that has undergone alteration during low-grade metamorphism. Optical and Digital Techniques Numerous techniques have been applied to determine the distri bution and character of macroscopic pore space in lavas, includi ng impregnation with pl astics and dissolution of the lava to reveal casts of pore structure (e.g., Sahagian et al. 1989), three-dimensional imaging by X-ray computed tomography (e.g., Song et al. 2001; Sahagian et al. 2002), and digital analysis of two dimensiona l sections (e.g., Toramaru, 1990; Klug and Cashman, 1996; Sahagian and Pr ousevitch, 1998; Al-Harthi et al. 1999; Saar and Manga, 1999). In this study, I adopt the last approach, because digital analysis of two dimensional sections will permit direct correl ation of pore structure to the chemical and mineralogical observations. Three thin sections from different re gions are examined; a chabazite zone metabasalt from East Greenland, a mesolite-s colecite zone metabasalt from eastern Iceland, and a laumontite zone metabasalt fr om the NSVG (Site 4). These samples are chosen to represent progressive stages of zeo lite alteration with respect to the range of pore evolution experienced during alteration. Thin sections ar e prepared for each sample, and digital images are prepared from the orig inal thin section. This is done either by capturing digital images through a high powered petrographic micros cope, or by scanning the thin section using a backli t flatbed scanner to create a high resolution image. These images were imported into the image pr ocessing software Adobe Photoshop, which was used to exaggerate the contrast between the features of intere st and the background groundmass, remove residual background noise created during scanning, and convert the image to grayscale. The images were then imported into either Image J or ArcView

PAGE 61

49 GIS 3.2a (ArcView); both programs were used to measure and stat istically analyze the porosity and secondary mineral phases in the images. When imported into ArcView the images were converted into shape files by creating a vector data format, which represents features as polygons defined by coordinate pairs, allowing for da ta analysis to be performed on the image. For porosity meas urements, these methods were limited to pores visible at the thin section scale. This excludes macro-pores larger than the size of a standard thin section (27 46 mm) or micro-pores smaller than the pixilation ability of the digital image processing software (which varies between programs). Observations NSVG Field Sites 3, 4 and 6 In situ field measurements of total orig inal macroscopic porosity were made at Sites 4 and 6 (Table 2.1) with in the NSVG, where outcrop e xposures were adequate for sampling methods. Porosity measurements in the field were limited to macroscopic observations, which confined analysis to the high porosity tops and bottoms of each lava flow. Two measurements at Site 4 show primary porosities of 19.7% and 21.7%, and one measurement at Site 6 shows a primary poros ity of 15%. This highlights the localized variation of porosity experienced within thes e high porosity zones of each lava flow. Figure 3.1 shows a close-up of an outcrop exposure at Site 6 that is also representative of porosity at Site 4. It can be observed from this figur e that vesicle shape, size and spacing vary consider ably, and the numbered regions highlight examples of this random distribution. Circle 1 highlights an area of little to no porosity surrounded by vesicular areas. The average vesicle was oval to round in shape, though many vesicles exhibited little to no sphericity appearing mo re elongated as highlighted by the vesicle (4x12 mm) in circle 2. Vesicl e size distributions within these measured zones were

PAGE 62

50 highly variable, with average vesicle radii ranging from 46 to 1 mm. Table 3.1 gives the statistical analysis of vesicle size measur ements performed on the outcrop of all amygdules, as well as those showing only t homsonite alteration or thomsonite and Figure 3.1: Outcrop photograph of vesicular basalt at Site 6 (Table 2.1) illustrating typical variation in pore size, shape and spacing. Ruler (15 cm long) shown on the left for scale. Vesicles are filled with white thomsonite-mesolite aggregates surrounded by the gray basalt Numbered circles highlight various porosity characteristics of basalt; high lighted areas show 1) little to no porosity, 2) elongated vesicle, and 3) pores of various sizes in close proximity to each other. mesolite alteration (discussed further below). Circle 3 in Figure 3.1 contains vesicles within close proximity of each other that ra nge in size from ~1 to 10 mm. In each of these highlighted regions and in the areas su rrounding them, there are no vesicles that are 2 3 1

PAGE 63

51 visibly connected. The lack of interconnection between vesicles was observed at each sampling site. Table 3.1: Statistical analysis of vesicle size (diameter) as a function of reaction progress from outcrop scale measurements at Site 6. All vesiclesThomsonite onlyThomsonite + Mesolite Statistical Dataa n 185125160 Size Range 1 – 461 – 1010 – 46 Mean 9.482.6223.76 Mode 1.001.0022.00 St. Dev.b 11.102.258.17 Size Distributionc 25% 17.001.001.00 50% 22.502.004.00 75% 28.503.0017.0 100% 46.0010.0046.00 aAll diameter values are given in mm. bStandard Deviation. cPercentage of pores given value. All vesicles were completely occluded due to secondary alteration, causing them to appear white or pink in outcr op. Many of the sites contained vesicles that appeared partially or completely unfilled, and the pres ence of such vesicles increased along river bed and shoreline exposures. Cut sample s exposing fresh surfaces, however, showed complete infilling of all vesicles, thus these open vesicles can be at tributed to weathering of alteration minerals. Sites 3 and 6 containe d vesicles filled with either thomsonite or thomsonite and mesolite (Table 3.1). Vesicles at Site 4 were filled w ith either chlorite or chlorite and laumontite with occasional calcite overprinting of the laumontite. Observations of collected NSVG samples show that matrix alteration varies within a given sample from areas with little altera tion to areas showing complete alteration of primary magmatic phases. These areas of high alteration appear to exist as reaction aureoles around primary pore sp aces, and around what appear to be microfractures acting as pathways between vesicles through the ma trix. Figure 3.2A (also shown in Figure 2.4)

PAGE 64

52 depicts the visible aureoles at the outcrop sc ale (Site 3), appeari ng as light-pink areas around and between amygdules. Figure 3.2B show s a schematic representation of these observed aureoles. The aureol es around vesicles are genera lly thicker than those around the possible microfractures between vesicl es. In many cases the veins appear as connecting pathways between vesicles. Figure 3.2: Images of reaction zones around and between macroscopic pore space in NSVG lavas. A) Field photo from Site 3 (Table 2.1) show ing an extensively altered pahoehoe flow with bleached ha loes around vesicles connected by thin anastomosing bleached areas. Pencil for scale. B) Schematic depiction of matrix alteration shown in (A), with broad alteration halos (white) around vesicles (light gray) connected by thin strands of alteration. These aureoles were first observed as a weathering phenomenon at the outcrop scale, and further analysis of cut hand sample s and thin sections (F igure 3.3) show that they are present throughout the entire rock. They are clear ly visible as lighter areas around and between vesicles, though at hi gher magnifications they have a dark appearance compared to the less altered matrix surrounding them. Petrographic observations at the thin secti on scale show that alteration appears to occur in stages that are independent of bot h lava composition and grade of alteration. Figure 3.4 depicts three stages of alteration observed within the matrix of analyzed

PAGE 65

53 Figure 3.3: Image is of sample NS04-14, show ing the laumontite zone alteration found at Site 4. Image was captured using a re flected light binocul ar microscope. Chlorite is the first mineral to fill th e pore, as seen by the dark rim around the pore wall. Residual open space is fill ed by laumontite, the white mineral in the center of each pore. Alteration aureoles are seen here as light colored rims around and between pores. NSVG samples, with minor alteration in 3.4A to extensive alteration in 3.4C. Figure 3.4A shows a sample containing the highest pe rcentage of unaltered matrix (lighter matrix) where the areas of alte ration (darker band in matrix) appear as networking veins. In samples exhibiting more alteration (Figur e 3.4B), the areas of unaltered matrix are smaller and are surrounded by thicker netw orking veins. Other samples show no networking of alteration (Figure 3.4C), and instead the entire matrix has undergone alteration.

PAGE 66

54 Figure 3.4: Various levels of s econdary alteration in the matr ix of NSVG basalts sampled during the field season. All photomicr ographs were taken through crossed polars. A) Alteration (dar ker areas) occurs in a ne twork pattern through the matrix (sample NS04-31), surroundi ng areas of unaltered matrix and connecting zeolitized amygdules. Fr esh feldspars and pyroxenes can be observed in the unaltered matrix (lighter areas). B) Thicker network pattern of matrix alteration (sample NS04-15) surrounding small pa tches of unaltered matrix where fresh feldspars and pyroxenes can still be observed. C) Matrix has undergone complete alteration (sampl e NS04-14), with visible feldspar laths but no residual pyroxene. The dominant minerals observed within thes e alteration aureoles are iron oxides, clay minerals, and zeolites. There is a dominant expression of iron oxides occurring within the altered matrix further away fr om pore spaces, while zeolite mineralization increases closer to primary pore spaces. Sa mples exhibiting more in tense alteration show the complete disappearance of primary pyroxene and alteration of the plagioclase laths.

PAGE 67

55 Reaction Progress Low-grade metabasalt samples exhibiting zeoli te zone alteration were selected to show varying levels of secondary mineral formation through increasing alteration from chabazite to mesolite-scolecite to laumontite zeolite zones. Samples from Greenland, Iceland, and Sites 4 and 6 from the NSVG we re analyzed for aspects of pore geometry such as spacing, size and shape and were obser ved with respect to th e extent of reaction progress experienced during al teration. The progress of r eaction is evident through the progressive infilling of vesicl es. At very low grades, the vesicles experience partial infilling, and as grad e increases, so does the amount of secondary mineral precipitation within the vesicle. Vesicles are progressively filled with clay minerals, clay and zeolite minerals, or multiple generations of zeolite minerals. Partial infilling at low-grades A low-grade metabasalt from East Green land exhibiting chabazite zone alteration (Neuhoff et al. 1997) was used for analysis of very low grades of alteration, as alteration lower in grade than the analcime zeolite zone are not preserved with in the NSVG lavas. A thin section (421505) represen ting chabazite zone alteratio n is shown in Figure 3.5A. This thin section contains ve sicles exhibiting three levels of alteration, those that are variably empty, completely filled with chabazi te, or partially filled with chabazite, and analysis is performed to assess vesicle properties in relation to these alteration characteristics. Figure 3.5B shows the thin section converted into a shape-file for area analysis in ArcView 3.2a. The region highlig hted in both images by the dashed line depicts an area with very little secondary mineral infilling. Th e lack of mineralization is independent of vesicle size and shape, and inst ead may be an effect of limited fluid flow through this region during alteration.

PAGE 68

56 Figure 3.5: Example of digital anal ysis of vesicle fillings in a zeolite facies vesicular lava from East Greenland. The original thin section (A) contains vesicles (278 individual objects, comprising 13.2% of th e sample) that are partly filled with chabazite. The region highlighted within the dashed line (in A and B) shows almost no secondary mineralization; see text for discussion. Part (B) shows separation of the three degrees of vesicl e-scale alteration; vesicles unaffected (unfilled) by alteration (light brown), completely-filled vesicles (dark brown), and partially-filled vesicles (light and dark brown).

PAGE 69

57 Area measurements were conducted on 284 vesicles, and each vesicle was then categorized with respect to the level of alte ration with values shown in Table 3.2. Total original porosity of the sample (calculated from the area of vesicle space divided by the total thin section area) is 13.2%, and the zeolite infillings occupy 70.4% of the original pore space (corresponding to a 9.4% increase in solid volume during alteration). Figure 3.6 shows the frequency of vesicles as a f unction of area and the level of alteration observed within the pore. Visual inspection of Figure 3.6 suggests that, in this sample, chabazite infilling is most prominent in larger vesicles, and that larger chabazite-filled vesicles are more likely to retain porosity af ter zeolite formation. The large spread in data shows that while the level of alteration is affected by vesicle size (average area of vesicles retaining open porosity is larger th an those completely filled), this may not be the only controlling factor. Table 3.2: Statistical analysis of vesicle size (area) as a function of reaction progress from measured vesicles of a low-grade me tabasalt from East Greenland (thin section 421505). Vesicles are either completely filled (filled), completely open (unfilled), or partially filled with residual pore space (partially filled). All vesiclesFilledUnfilledPartially filled Statistical Dataa n 2786818921 Size Range 0.001 – 2.5020.001 – 1.3950.001 – 0.5430.053 – 2.502 Mean 0.0940.0820.0350.661 Mode 0.0020.0020.002-St. Dev.b 0.2390.0900.0720.586 Size Distributionc 25% 0.0040.0110.0030.280 50% 0.0180.0510.0100.401 75% 0.0860.1260.0330.907 100% 2.5020.3950.5432.502 aAll area values are given in mm2. bStandard Deviation. cPercentage of pores given value.

PAGE 70

58 0 10 20 30 40 50 60 70 0.0010.0050.010.050.10.515 vesicle area (mm2)frequency unfilled vesicles partially filled vesicles filled vesicles 0.034 0.082 0.661 Figure 3.6: Frequency of vesicl e areas versus the amount of secondary alteration filling the vesicle for a sample of chabazite-t homsonite zone alteration from East Greenland (cf. Neuhoff et al. 1997). Vesicles are unf illed, partially filled or completely filled with chabazite. Average vesicle area for each level of alteration is labeled. Mafic phyllosilicate to zeolite infilling Observation of lavas exhibiting zeolite al teration higher in gr ade than chabazitethomsonite zone found no residual pore space in vesicles observed in thin section. The relationship between reaction progress and alte ration minerals then becomes related to the progressive stages of mineral precipitat ion within vesicles. Low-grade metabasalts from Site 4 in the NSVG and from eastern Iceland exhibit early pr ecipitation of mafic phyllosilicates followed by la ter stage zeol ite infilling. Samples collected during the field season at the Cutface Creek location (Site 4, see Table 2.1) within the NSVG exhibit laumontite z one alteration. Microprobe analysis of a sample in thin section shows each vesicle ini tially filled with chlor ite rims (Figure 3.3),

PAGE 71

59 with any residual pore space completely fill ed with laumontite during later stages of alteration. Visual inspection of vesicles (Figure 3.3) also shows a consistency in thickness of chlorite rims irrespective of vesicl e size and shape. All vesicles within this thin section were measured for total area a nd for the amount (in area) filled by laumontite and chlorite, and values are represented in Table 3.3. Chlorite rim thicknesses in this sample also were measured (Table 3.3) and are statistically consistent throughout the thin section (0.52 +/0.12 mm). Figure 3.7 plots this consistency in thickness with respect to vesicle area, as well as the percentage of the vesicle filled with chlorite, and shows that the decrease in chlorite infilling as vesicle size increases is due to the consistency in chlorite rim thicknesses. Figure 3.8 again shows the percentage of chlo rite infilling with respect to area, and also shows the percentage of laumontite infilling with respect to area. Due to the consistency in clay rim thicknesses, the percentage of each vesicle filled with chlorite decreases as the area of the pore increases, leaving more residual space in larger pores for later infilling of laumontite. Table 3.3: Statistical analysis of vesicle size as a function of reaction progress for vesicles filled with chlorite and/or laumontite at the thin section scale (sample NS04-14) from Site 4. Ar ea data is given in mm2 for the entire vesicle and for the area of the vesicle filled by laumontite and chlorite. Chlorite rim thickness measurements are given in mm. Vesicle AreaLaumontite AreaChlorite Area Chlorite thickness Statistical Data n 232323 255 Size Range 1.39 – 23.300.00 – 16.841.29 – 14.21 0.29 – 0.90 Mean 10.335.015.32 0.52Mode --0.00-0.55 St. Dev.a 6.434.763.16 0.12 Size Distributionc 25% 3.761.362.41 0.41 50% 5.171.713.46 0.67 75% 14.127.286.84 0.50 100% 16.4315.141.29 0.51 aStandard Deviation. bPercentage of pores given value.

PAGE 72

60 Figure 3.7: Plot shows the thickness in mm of ch lorite rims (black circles) and the total percent of the vesicle filled with chlorite (open circles) with respect to total vesicle area in mm2. As vesicle area increases, the chlorite rims maintain a consistent thickness of ~0.52 mm, while the percentage of vesicle occlusion by chlorite steadily decreases. Further analysis of this consistency in thickness of mafic phy llosilicate rims was performed on a previously collected sample from eastern Iceland. Mesolite-scolecite zone alteration in a sample from eastern Icel and contains vesicles completely filled by interlayered chlorite-smectite (c/s) and scolecite. Each pore was initially filled with clay rims during early stages of alteration, leav ing residual pore space in larger pores that becomes completely filled with scolecite duri ng later stages of alte ration (Figure 3.9). Figure 3.9 highlights three vesicles, each exhibi ting a different size and shape. Each vesicle is initially filled with rims of C/ S that are of a consistent thickness between vesicles. The smallest vesicle is almost co mpletely occluded by the C/S while the other

PAGE 73

61 Figure 3.8: Plot shows the percen tage of mineral infilling wi th respect to total vesicle area. Smallest vesicles are completely f illed with chlorite. As the vesicle size increases, the percentage of the vesicle filled with chlorite decreases, and the residual space is filled with laumontite. two retain pore space for scolecite infilling. Measured clay rim thicknesses are given in Table 3.4 and are shown to be consistent th roughout the thin sec tion (0.091 +/0.02 mm) irrespective of pore size, although a slight variation in th ickness is observed in non-ovoid pores (Figure 3.9) due to geometric effects of taking a two-dimensional slice of a threedimensional object.

PAGE 74

62 Figure 3.9: Photomicrograph showing a thin se ction of a mesolite-scolecite zone lava from eastern Iceland. Each pore is filled initially with a ch lorite-smectite clay rim; larger pores are subsequently filled with scolecite. Clay rim thicknesses are consistent throughout the section (0.10 +/0.02 mm) irrespective of pore size, highlighted by the two circled pore s and the larger pore between them. Pores smaller than 0.1mm2 are completely filled with clay. Digital analysis of pore sizes compared to level of infilling is performed using ArcView and Image J software. The vesicles ar e analyzed as either filled with both C/S and stilbite or completely occluded by c/s. This information is plotted in Figure 3.10, and shows that pores with an area of 0.1 mm2 or less are completely filled with clay, while larger pores are filled with both clay and zeolites due to the consistency in C/S thicknesses.

PAGE 75

63 Table 3.4: Statistical analys is of vesicle areas (mm2) and clay rim thicknesses (mm) for sample 94-80 from eastern Iceland. Vesicle AreaClay thickness Statistical Data n 229344 Size Range 0.001 – 2.1910.038 – 0.163 Mean 0.0790.091 Mode 0.0030.092 St. Dev.a 0.2450.022 Size Distributionb 25% 0.0040.104 50% 0.0130.088 75% 0.0430.046 100% 2.1910.068 aStandard Deviation. bPercentage of pores given value. Multiple stage zeolite infilling Initial rims of mafic phyllo silicates were not observed in every analyzed sample, and instead analysis of reaction progress was based on successive levels of zeolite infilling. Observations at th e outcrop scale of Site 6 show that all vesicles exhibit thomsonite mineralization while only larger por es progress to crystallization of mesolite. A representative analysis of pore size distributions within the lavas (Table 3.1) shows that pores exhibiting only thomsonite alteration have an average radius of 2.62 mm, while those progressing to mesolite growth have an average radius of 23.76 mm. Discussion Nature of Porosity and Permeability in Basalts Previous laboratory studies of unaltered lavas suggest that fluid movement through the rock is dependant on the geometric and spat ial aspects of the vesicle space such as the aspect ratio, size, and spac ing, thus implying a dependenc y of fluid movement on the geometry and connectivity of vesicles with in the rock. Typical porosity within high vesicular zones of basaltic la vas, however, is highly variable, as shown by the correlation

PAGE 76

64 0 10 20 30 40 50 60 70 80 90 0.0050.010.050.10.51 vesicle area (mm^2)frequency all pores, n=172 clay filled, n=110 Figure 3.10: Data compiled from digital analys is of mesolite-scolecite zone lava from eastern Iceland (refer to Fi gure 3.3). Analysis performed at the thin section scale using ArcView and ImageJ software. Plot shows the area variation of all vesicles within the thin section, co mpared to a random sampling of those filled only with clay. Vesicles exceeding an area of 0.1 mm2 are filled with both clay and zeolites, while those smaller than 0.1 mm2 are filled completely with clay. of these observations from basaltic lava sample s between different regions. Observations at the outcrop scale of basalts from the NSVG (Figure 3.1) show local areas of little to no porosity surrounded by areas of hi gher porosity that display mu ch variation in the vesicle size and shape. Analysis of samples in th in section from the NSVG, East Greenland and eastern Iceland (e.g., Figure 3.9) show this o ccurring even at the mm scale, and it is apparent at this scale that all analyzed samples also e xhibited random vesicle-vesicle spacing with little to no in terconnection of vesicles. In spite of the high variabili ty in vesicle size, shape a nd spacing, all vesicles are partially to completely occluded due to sec ondary mineralization, im plying that there is some other factor controlling fluid flow th rough the rock and gaining access to pores.

PAGE 77

65 The observed alteration aureol es (Figures 3.2, 3.3 and 3.4) suggest that fluid is transported through the rock along a permeable network of microfractures and connected vesicles. This implies that vesicles will e xperience secondary alteration regardless of their size shape and spacing as long as they ar e a part of the fluid flow pathways. This also suggests that the overall permeability of vesicular basalts will remain high even in areas where porosity is low and vesicles are wi dely distributed. In these areas there will be pathways of high permeability surrounding areas of little to no permeability. The general highlighted region in Figure 3.5, for ex ample, reflects a portion of the sample that was not permeable enough to allow for diffusion of the chemical components necessary to precipitate the secondar y mineralization. This may reflect localized lowered permeability due to the absence of the fluid flow pathways necessary for these pores to be part of the permeable network through the lava. Observation of alteration within the ma trix shows these alteration aureoles occurring around and between vesicles. Th e primary magmatic phases are highly weathered within these aureoles signifying that these are preserving th e pathways of fluid movement through the rock during alterati on. This suggests that the chemical components necessary to precipitate the ve sicle-filling minerals are derived from a spatially restricted region (a lteration aureoles) around the ve sicle and the fluid pathways over which diffusion occurs. Almost as in a hydrologic system with streams flowing in and out of a lake, where the streams represent the pathways between ve sicles and the lake represents the vesicle, the residence time of fl uid within the lake is expected to be higher than within the stream due to its larger st orage capacity. This means that the alteration

PAGE 78

66 occurring in and around vesicles should be greater than those surrounding the pathways between vesicles, which is seen in Figure 3.2. The exact nature and initial formation of these pathways are still unclear, though there are many possible explanations. These mi crofractures are interp reted to be primary features created during formation and cooli ng of the lava, and not secondary openings due to external stress. Kowallis et al. (1982) observed similar microfractures in comparable basalts collected from the IRDP (Iceland Research Drilling Project) borehole in Reydarfjrdur, Iceland. Their observations show that the observed microfractures formed during cooling and have since under gone secondary alterati on and mineralization due to massive amounts of fluid movement through the basalt. Termination of these fractures appears blunt, rather than the ta pered termination typical of newly formed cracks, and the presence of de licate secondary mineral grow th within these features would not have survived the stress induced during a secondary cracking process (Kowallis et al. 1982). They also describe fractur es coincident with grain boundaries that are often produced by thermal stressi ng, as well as non-coincident grain boundary fractures produced by local strain variati ons. Though these fractures increase the accessibility of the basalt matrix to alteration, they do not appear to coincide with those observed in this study as they would not neces sarily bridge vesicles and have a major affect on permeability. Processes experienced by a given lava fl ow during formation ca n control the total amount of porosity and can ex ert a fundamental control on th e nature and connectivity of primary pore space. Lava viscosity, lava volat ile content, atmospheric pressure, and the degree of crystallinity of the lava, among other factors, all affect both the total

PAGE 79

67 vesicularity of a lava flow and the vesicle size distri bution (e.g., Cashman and Mangan, 1994; Sahagian and Maus, 1994; Prousevi tch and Sahagian, 1996, 1998; Sahagian et al. 2002). Srouga et al. (2004) suggests that autobreccia tion processes during cooling of rhyolitic lavas can also enhance the primary por osity within the lava. They suggest that even though these pore spaces are small they may form a connected network, which has potential for increasing the ra te of fluid flow through the rock. This potential network, however, is not similar to the network obser ved in the NSVG basalts as it would not necessarily connect vesicles formed due to trapped gas bubbles within the rock during cooling. Another possible explanation of the connected network between vesicles observed in this study is the creation of microcracks due to the el ongation of bubbles during lava flow and degassing prior to coo ling (Saar and Manga, 1999; Mueller et al. 2005). The strong presence of spherical vesicles in th e high permeability zones of the cooled lava, however, suggests that these vesicles retain gas after cooling which must later escape through the solid rock. This escaping gas, however, will cr eate a connected network of pathways between vesicles and through the matrix, which can then be reused as a pathway for fluid flow during later events of low-grade alterati on. Fluid moving through this network will have a longer residence time in the larger pore spaces, allowing for more of the surrounding matrix to undergo intera ction and alteration with the fluid. This will cause larger vesicles to be surrounded by thicker alteration aureoles than the smaller vesicles or thinner connecti ng pathways, creating alteration aureoles similar to those observed in the NSVG basalts.

PAGE 80

68 Controls on Reaction Progress The level of reaction progress experienced by a vesicle can be related to the amount and type of secondary mineralization obser ved within the vesicle, and is strongly dependant on the vesicle size and shape. In ge neral, larger vesicles will retain residual pore space in lower stages of alteration, as observed in the chabazite grade zeolite alteration present in the Greenland sample. Vesicles exposed to higher grades of zeolitization, however, retain no residual por e space and instead the presence of later stage mineralization is controlled by vesicle size. There is a general trend of secondary mineralization thr oughout the reaction progress from clay mineral precipitation to progressive stages of zeolite mineral precipitation. Observations show that larger vesicles will progress to later stages of alteration while smaller ones will become clos ed off earlier in the reaction progress. Also, irregularly shaped vesicles may close o ff earlier in the reaction progress even with large areas due to constriction caused by the smallest diameter. In vesicles where clay rims are present, these rims will generally form first and grow to a consistent thickness within a sample irrespective of vesicle size. Thus, smaller vesicles will become completely occluded by clay minerals, while larger ones will show later stage zeolitization. Co mparison of clay rim thicknesses between the measured samples discussed above shows that while the thicknesses within a given sample are consistent, variation will occur between lava fl ows. This shows no relation to vesicle size and instead must be due instead to differences in mineralogy of the rock and of the clay rimming the pores, which is explored in Chapter 4. All vesicles within basaltic lavas that are connected to the fluid flow pathways can be analyzed and used to interpret the reac tion progress experienced by the basalt during

PAGE 81

69 secondary alteration. Interpreta tions must not be made, however, wit hout close analysis of the vesicle size and shape characteristics as smaller vesicles and/or those with constrictions due to irregula rity in shape may not show the full extent of alteration experienced by the lava. Assessments of the la va chemistry and alteration grade are also necessary, as this will cause a variation in secondary mineral precipitation which will affect interpretations of reacti on progress experienced by the lava.

PAGE 82

70 CHAPTER 4 MODELING OF MINERAL PARAGENESES Introduction The field, petrographic, and digital observ ations summarized in preceding chapters illustrate several key factors involved in th e coupled interactions between pore space evolution and mineral paragenesi s during low-grade alteration of vesicular lavas. First, there is a definite crystallization sequence of secondary minerals that typically progresses from amorphous silica to mafic phyllosilica tes (chlorites and di -octahedral to trioctahedral smectites) to zeolites, followed in some instances by late zeolites or calcite. Second, these minerals typically occlude pr imary porosity, with the degree of reaction progress being sensitive to vesicle size. Fo r instance, measured thicknesses of mafic phyllosilicate rims are consistent at the th in section scale irre spective of the size, geometry, and connectivity of vesicles. This same phenomenon is manifested by the vesicle-size dependence of occurrence of late stage zeolites and preservation of residual pore space. Third, a notable feature observed in these rock s is the presence of reaction aureoles around primary pore spaces and netw orking microfractures. Alteration aureoles are clearly visible as lighter areas surrounding the vesicles and microfractures (Figure 3.3). These aureoles potentia lly represent fluid pathways through the matrix during alteration, suggesting that the chemical com ponents necessary to precipitate the vesiclefilling minerals are derived from a spatially restricted region (alteration aureoles) around the vesicle and the fluid pathways over which diffusion occurs.

PAGE 83

71 This chapter describes the results of i rreversible reaction path modeling conducted in order to interpret these features of lowgrade metabasalts. These calculations are used to 1) explain the genesis of the sequence of secondary minerals formed during alteration; and 2) interpret the e volution of porosity duri ng alteration. The result s are then used in concert with my field, petrogr aphic, and digital observati ons to develop a geometric model explaining the relationship between por e size, mineral authigenesis, and porosity evolution during alteration. Methods Mineral paragenesis and porosity evol ution during low gr ade alteration are modeled using the irreversible reaction path code EQ3/6 (Wolery, 1979; Wolery et al. 1990). The program computes the compositiona l evolution in systems of water-rock interaction during reactive mass transfer assuming local and partial equilibrium between an aqueous fluid and coexisting mineral and gas phases (Wolery, 1979). Basic inputs to the code are the composition of the star ting fluid, temperature (with pressure corresponding to liquid-vapor equilibrium for water), and the composition and abundances of phases that will be reacted with the fluid. In all models presented below, the initial fluid was taken to be 1 kg of dilute aqueous fluid u ndersaturated with respect to all minerals that interacts with a lava c ontaining 20% initial porosity (or a bulk rock volume of 4 liters). In the first set of models, a simplified ba salt composition was created as an input for the program (Table 4.1 and 4.2) that is re presentative of the mineral modes, mineral compositions, and the bulk rock composition of olivine tholeiites in the North Shore Volcanic Group (after Brannon, 1984). The mode l simulates titration of 4 liters of lava into 1 kg of a dilute aqueous solution (initially no saturated ph ases) in a closed system at

PAGE 84

72 Table 4.1: Bulk rock composition fo r olivine tholeiite basalt (wt %) Oxide Calculated Sampled*Average*Range* SiO2 47.75 47.2648.0747.6 – 48.5 Al2O3 18.01 16.8616.8016.5 – 17.4 CaO 13.67 11.0310.9310.6 – 11.2 Na2O 2.26 2.512.482.38 – 2.64 K2O 0.12 0.310.280.13 – 0.44 MgO 6.44 6.746.986.25 – 7.56 FeO 9.67 11.9812.2011.4 – 12.6 TiO2 1.83 1.671.711.43 – 1.93 *Values from Brannon, 1984 Table 4.2: Modal abundances and compositions of primary mineral phases in olivine tholeiite Initial Phase End-member X1 Moles2 Volume %Mass Rate3 Feldspar 23.6059.306425.79 1 Anorthite 65.00 Albite 33.80 Orthoclase 1.20 Pyroxene 27.4024.103059.99 10 Wollastonite 42.00 Enstatite 43.50 Ferrosilite 14.50 Olivine 4.3011.00708.31 100 Forsterite 63.90 Fayalite 36.10 Ilmenite 2.502.0037931 1 Magnetite 1.701.90393.62 1 Glass 2.601.70182.21 1000 1Molar percentages are represented by X. 2Moles are for 4 liters of basalt. 3Relative reaction rates represent dissolu tion rates of mineral phases. 70C, Psat. This bulk composition and temperature condition is typical of the thomsonitemesolite zone alteration found in the NSVG. Two isochemical versions of this model were run that differed only in the relative rate s of dissolution of components into the fluid. In the first, the lava was titrated into the fluid stoichiometri cally. In the second, the same bulk composition was distributed among separate primary phases in their relative modal proportions and allowed to re act at relative rates that approximate differences in bulk dissolution rates (Table 4.2; Gislason and Eugster, 1987; Palandri and

PAGE 85

73 Kharaka, 2004). Thermodynamic data for th is model was taken from the “combined” database provided with EQ3/6. Precipitation of epidote solid solutions, grandite garnet solid solutions, prehnite, and tremolite was suppressed in the model as these species are not expected to form under these conditions bu t are predicted to be stable using the data in the combined database. The second model uses a calculated phase asse mblage that is repr esentative of the mineral modes, mineral compositions and bulk rock compositions of a sampled basaltic andesite (sample NS04-14 from sampling site 4; see Table 4.3) from the NSVG. This sample is representative of upper zeolite f acies conditions, although it is somewhat more silicic than most of the NSVG. Because of the extensive matrix alteration of the sample that precluded optical determ ination of primary mineral mo des, the relative abundances of primary minerals were calculated with the computer program KWare Magma (Wohletz, 1999) by assuming a CIPW normative mi neralogy of the unaltered basalt. The resulting mineralogy (Table 4.4) compares favorably with previous observations of andesites in the NSVG (Brannon, 1984; Schm idt, 1990). Representative mineral formulas were taken from Chapter 2 along with reported values by Brannon (1984) and Schmidt (1990). In this model, the temper ature was set at 120C (typical of laumontite zone alteration; cf. Chapter 3), and the primary phases were titrated in their relative modal proportions using relative rate consta nts that approximate differences in bulk dissolution rates (Table 4.4; Gislason and E ugster, 1987; Palandri and Kharaka, 2004). Due to the simpler alteration assemblage enc ountered in this sample, the calculations were performed with a database based on SUPCRT92 (Johnson et al. 1992) augmented with data for ilmenite from Stefansson ( 2001) and hydroxyapatite and daphnite from the

PAGE 86

74Table 4.3: Calculated anhydrous bulk composition of basaltic andesite before and after reaction. Ilmenite HydroxylApatite ClinoPyroxene OrthoPyroxene Albite Anorthite K-Feldspar Quartz Hematite Chlorite Laumontite Whole Rock Unaltered Andesite SiO2 0.00 0.00 1.5114.7723.124.986.52 6.860.000.000.0057.67 TiO2 2.26 0.00 0.000.000.000.000.00 0.000.000.000.002.26 Al2O3 0.00 0.00 0.000.006.544.231.84 0.000.000.000.0012.62 FeOa 2.03 0.00 0.6011.360.000.000.00 0.000.000.000.0013.98 MgO 0.00 0.00 0.173.540.000.000.00 0.000.000.000.003.72 CaO 0.00 0.56 0.740.000.002.330.00 0.000.000.000.003.62 Na2O 0.00 0.00 0.000.003.980.000.00 0.000.000.000.003.98 K2O 0.00 0.00 0.000.000.000.001.70 0.000.000.000.001.70 P2O5 0.00 0.42 0.000.000.000.000.00 0.000.000.000.000.45 Total 4.29 0.98 3.0229.6633.6411.5410.06 6.860.000.000.00100.00 Altered Andesite SiO2 0.00 0.00 5.970.0023.150.006.52 13.960.006.201.8757.67 TiO2 2.26 0.00 0.000.000.000.000.00 0.000.000.000.002.26 Al2O3 0.00 0.00 0.000.006.550.001.84 0.000.003.510.7912.69 FeOa 2.03 0.00 1.790.000.000.000.00 0.002.907.420.0014.13 MgO 0.00 0.00 1.000.000.000.000.00 0.000.002.770.003.78 CaO 0.00 0.59 2.790.000.000.000.00 0.000.000.000.443.82 Na2O 0.00 0.00 0.000.003.980.000.00 0.000.000.000.003.98 K2O 0.00 0.00 0.000.000.000.001.70 0.000.000.000.001.70 P2O5 0.00 0.45 0.000.000.000.000.00 0.000.000.000.000.45 Total 4.28 1.05 11.550.0033.680.0010.07 13.962.9019.903.10100.48 aTotal Fe reported as FeO

PAGE 87

75 combined database. In addition, ideal chlo rite, plagioclase, and orthopyroxene solid solutions were defined. Andradite and gro ssular garnet, prehnite, annite, tremolite, epidote, talc, analcime, and magnetite are all phases that are suppresse d in the input file of the model so that they will not form as part of the alteration sequence. Table 4.4: Reactant phase composition, abundance, and relative dissolution rate in basaltic andesite used as i nput for reaction path model. Initial Phase End-member X1 Moles2 Volume %Mass Rate3 Feldspar 24.4962.396569.50 10 Anorthite 20.12 Albite 62.33 Orthoclase 17.55 Orthopyroxene 29.2323.633527.30 100 Enstatite 35.7 Ferrosilite 64.3 Clinopyroxene 2.992.71353.22 100 Diopside 100.00 Hedenbergite 0.00 Quartz 13.587.70815.86 1 Ilmenite 3.362.66510.21 100 Apatite-OH 0.230.91114.72 100,000 1Molar percentages are represented by X. 2Moles are for 4 liters of basalt. 3Relative reaction rates represent dissolu tion rates of mineral phases. Results Mineral parageneses Figures 4.1A and 4.1B depicts the mineralogic evolution of an oliv ine tholeiite as a function of reaction progress ( ) assuming either congruent dissolution of the bulk lava (4.1A) or dissolution controlled by the relativ e dissolution rates of the primary minerals (4.1B). Note that the plots in Figure 4.1 re present the relative abunda nces of minerals in terms of absolute volume. A value of log of -10 corresponds to the beginning of reaction; a value of 0 repres ents completion of reaction.

PAGE 88

76 A0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 -10.0-8.3-7.3-6.4-5.7-5.2-5.0-4.6-4.2-1.0-0.7-0.5-0.4-0.3-0.2-0.10.0log volume (cc) Homogenous Glass Others Oxides Clays Mesolite Scolecite Figure 4.1: Mineralogic composition of la vas as a function of reaction progress for models describing A) homogeneous dissolution of an olivine tholeiite at 70 C; B) incongruent dissolution of an olivine tholeiite controlled by relative dissolution rates of primary minerals at 70 C; and C) incongr uent dissolution of a basaltic andesite controlled by the relative dissolution rates of primary minerals at 120 C.

PAGE 89

77 B0 500 1000 1500 2000 2500 3000 3500 4000 4500 -10.0-9.0-8.4-7.2-6.1-5.7-4.8-3.0-2.3-1.7-1.3-1.3-1.0-0.7-0.5-0.4-0.3-0.2-0.10.0log volume (cc) Magnetite Olivine Basalt Glass Pyroxene Ilmenite Scolecite Mesolite Chlorite Titanite Serpentine Feldspar Figure 4.1: (continued)

PAGE 90

78 C0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 -7.0-6.0-5.0-4.0-3.0-2.0-1.5-1.4-1.2-1.1-1.0log volume (cc) LaumontiteChloriteHematiteQuartz K-Feldspar Plagioclase Ortho-Pyroxene Clino-PyroxeneApatite-OHIlmeniteSmectite Figure 4.1: (continued)

PAGE 91

79 For the congruent dissolution model in Figur e 4.1A, the onset of secondary mineral precipitation occurs simultaneously for all phases. The zeolites mesolite and scolecite comprise the largest fraction of the alteration assemblage. Clay minerals predicted by the model consist mostly of chlor ite with minor precip itation of biotite and saponite. Oxides have the least affect on volume increase with precipitation mostly occurring as titanite, though magnetite, rutile, hematite, and gibbsite also form. The remaining minerals are grouped as others. Diopside dominates the vo lume contribution of these other minerals during the run, with only minor am ounts of clinozoisite. At the final stage of the model the homogenous glass has dissolved completely a nd all volume in the system consists of secondary minerals. Figure 4.1B depicts the mineralogic evolut ion of an olivine tholeiite consisting initially of glass, olivine, pyroxene, feld spar, ilmenite and magnetite in their modal proportions. Each primary phase dissolves to various levels of co mpletion at different stages of the reaction controlled by their relativ e reaction rate. Contra ry to the results of the congruent dissolution model, the onset of secondary mineral precipitation varies between phases. The complete dissoluti on of glass and olivine, along with the compositional change of pyroxene from Wo0.42En0.435Fs0.145 to diopside, donates iron and magnesium needed to form the chlorite and serpentine clay minerals relatively early in the reaction. The complete dissolution of feldspar dona tes the calcium and sodium needed to form the zeolites mesolite and scol ecite. All potassium dissolved from the feldspar remains in the aqueous phase. Il menite completely dissolves, donating the titanium for precipitation of titanite. Magnetite is the only primary magmatic phase that remains stable th roughout the reaction.

PAGE 92

80 Although both of the above models predict roughly the same mineral assemblage, they differ in their predictions of the timing of mineral formation. Both models predict a mineral assemblage dominated by the zeolites mesolite and sc olecite along with pyroxene and mafic phyllosilicate, consistent with what is observed in olivine tholeiites within the thomsonite-mesolite zone of the NSVG. However, the model depicted in Figure 4.1B more closely approximates the mineral paragene ses observed in zeolite facies metabasalts than that shown in Figure 4.1A in terms of the relative abundance of secondary mineral phases and their relative sequence of preci pitation. Early formation of mafic phyllosilicate phases relative to zeolites that is observed in the relative timing of mineral precipitates within vesicles is a consequence of the different ial reaction rates of primary mineral phases in the incongruent dissolution model. Comparison of the results for olivine thol eiite dissolution eith er congruently or incongruently (Figures 4.1A and 4.1B) illust rates the importance of relative dissolution rates in controlling reaction pa ths. Alteration of the oliv ine tholeiite as a homogenous glass (4.1A) shows the onset of secondary mi neral precipitation to occur simultaneously for all phases, and allows the large volume increase due to zeolitization to dominate throughout precipitation. These results do not correlate to the petr ographic observations of altered basalts from the NSVG, indicating that the observed paragenesis of early mafic phyllosilicates followed by zeolites can not be explained by stoichiome tric reaction of the basalt, but rather is a consequence of di fferential dissolution rates for primary basalt phases as shown in Figure 4.1B, and in the inco ngruent dissolution of basaltic andesite in Figure 4.1C. Consequently, modeling of the mi neral paragenesis observed in the basaltic andesite employed only an inc ongruent dissolution approach.

PAGE 93

81 Figure 4.1C depicts the mineralogic evol ution of a basaltic andesite during interaction with dilute meteor ic solutions at 120 C. I nput and output values for the moles, volume, and mass of mineral phases ar e compared to calculated values and shown in Table 4.5. The final phase assemblage listed in Table 4.5 consists of hematite, chlorite, laumontite, K-feldspar, albite, clinopyroxene, quartz, hydroxyapatite, and ilmenite, which is essentially the assembla ge observed in sample NS04-14. The output Table 4.5: Mole, volume and mass amounts in unaltered and altered andesite for the computer (modeled) and mathem atical (calculated) models. Moles* Volume % Mass (g) Unaltered Calculated Modeled Calculated Modeled Calculated Modeled Quartz 13.58 13.587.707.70815.85 815.86 K-Feldspar 4.30 4.3011.7011.691196.42 1196.40 Albite 15.26 15.2638.2438.304000.77 4002.50 Anorthite 4.93 4.9312.4312.411372.44 1370.60 Ilmenite 3.36 3.362.662.66510.21 510.21 Apatite-OH 0.23 0.230.910.91116.55 114.72 Clinopyroxene 2.99 2.992.712.71353.22 353.22 Orthopyroxene 29.23 29.2323.6423.633527.43 3527.30 Total 73.88 73.88100.00100.0011892.89 11890.81 Altered Calculated Modeled Calculated Modeled Calculated Modeled Quartz 27.50 29.3014.3915.481652.24 1760.20 K-Feldspar 4.28 4.3010.7510.881192.05 1196.40 Albite 15.20 15.2635.1535.563986.04 4002.20 Ilmenite 3.34 3.362.442.48507.07 510.21 Apatite-OH 0.25 0.230.920.85123.72 114.72 Clinopyroxene 5.90 5.099.147.841366.98 1112.52 Hematite 4.29 3.973.002.80342.73 634.10 Chlorite 4.07 3.5319.8117.342355.68 2319.75 Laumontite 0.92 1.404.406.77366.38 659.30 Total 65.76 66.43100.00100.0011892.89 12309.40 *Values are for 4 liters of Basalt generated by the model shows mass remaining re latively constant, with an increase of only 3.5%. Mass increase largely reflects hydrat ion of the lava. The bulk composition of the lava is relatively unchange d other than this gain of H2O, as shown by the comparison of the initial and final bulk compositions s hown in Table 4.3. Early, complete dissolution

PAGE 94

82 of orthopyroxene and anorthite contributes the iron, magnesium and calcium needed to create the hematite, chlorite and laumontite obser ved in the sample. Iron released early in the reaction progress re-precipitates as he matite until enough magnesium and aluminum are present in the aqueous phase to reach chlorite saturation. Chlorite is the modally most important secondary mineral until laumontite saturation is reached, which is the last mineral to form. Table 4.6 lists the compositi ons of mineral solids solutions present at the end of the model. The anorthite compone nt in plagioclase is lost during dissolution, and all remaining plagioclase is completely albitized. Table 4.6: Phase compositions and abundances at end of reaction path for basaltic andesite given in their respective en d-members and molar proportions (where solid solutions exist), moles, volume %, and mass. Final Phase End-member X1 Moles2 Volume %Mass Feldspar 19.5646.441196.40 Anorthite 0.00 Albite 78.02 Orthoclase 21.98 Clinopyroxene 5.097.841112.52 Diopside 93.07 Hedenbergite 6.93 Chlorite 3.5317.342319.75 Clinochlore 35.23 Daphnite 64.77 Laumontite 1.406.77659.30 Hematite 3.972.80634.10 Quartz 29.3015.481760.20 Ilmenite 3.362.48510.21 Apatite-OH 0.230.85114.72 1Molar percentages are represented by X. 2Moles are for 4 liters of basalt. Orthopyroxene compositions are essentially cons tant while this phase is present in the model. Clinopyroxene increases in modal abundance with reaction progress, and also becomes steadily more Fe-rich. Chlorite com positions are initially clinochlore-rich, but become more daphnite rich with increasing reaction progress.

PAGE 95

83 The bulk solid composition is in excellent agreement with th at calculated by assuming that the lava completely altered to the observed mineral assemblage. The lower half of Table 4.5 shows the modal abundances of secondary minerals determined through a normative calculation analogous to that of the CIPW norm, but with the observed secondary mineral assemblage in place of pr imary igneous phases. These calculations indicate the creation of chlorite, laumontite, hematite, and an increase in clinopyroxene and quartz through complete destruction of the orthopyroxene and the anorthite. All other primary phases including the ilmenite, hydr oxyl-apatite, albite, and k-feldspar show no change in their compositions or amounts present throughout alteration. Volume Changes During Alteration Volume change observed during the reacti on of each modeled run can be observed both as the total volume change and as the instantaneous volu me change at each step of the reaction. The total volume change of the system with respect to reaction progress is presented in Figure 4.1. In each run, the mass is held constant throughout the reaction to maintain a constant initial primary porosity of ~20%; therefore, any volume increase observed in each run is a result of the forma tion of secondary minerals with higher molar volumes than initial primary magmatic phases Reaction of the oliv ine tholeiite shows a volume increase of 14% at the end of reacti on (Figures 4.1A and 4.1B ), while reaction of the basaltic andesite in (Figure 4.1C ) shows an 8% increase in volume. Instantaneous volume change, i. e., the net change in amount of products relative to reactants at each step, is calculated for each step of the reaction and depicted in Figure 4.2. Negative values along the y-axis of this plot correspond to a volume decrease in the system, which can also be explained as por osity increase. Conversely, positive values

PAGE 96

84 correspond to a volume increase which relates to porosity destruction. In each modeled reaction, there is an initial volume decrease in the system creating a subtle increase in porosity before there is a more significan t volume increase causing a destruction or occlusion of primary porosity. In the congr uent dissolution model for olivine tholeiite, the positive volume change occurs at the onset of mineral precipitati on, and is related to the formation of the whole phase assemblage. For the incongruent dissolution model, the instantaneous volume change becomes positive early due to the formation of serpentine and chlorite and before formation of zeo lites. The need for space during early phyllosilicate formation may account for the prominent clay rims observed in many samples. A similar trend is observed for reaction of the basaltic andesite. A-100 -80 -60 -40 -20 0 20 40 -10.0-8.3-7.3-6.4-5.7-5.2-5.0-4.6-4.2-1.0-0.7-0.5-0.4-0.3-0.2-0.10.0log % mineral volume change Figure 4.2: Instantaneous change in total mineral volume as a function of reaction progress. Panels A, B, and C correspond to models shown in Figure 4.1.

PAGE 97

85 B-100 -80 -60 -40 -20 0 20 40 -10.0-9.0-8.4-7.2-6.1-5.7-4.8-3.0-2.3-1.7-1.3-1.3-1.0-0.7-0.5-0.4-0.3-0.2-0.10.0log % mineral volume change C-100 -80 -60 -40 -20 0 20 40 -7.8-6.5-6.3-5.9-4.6-2.0-1.5-1.4-1.2-1.1-1.0log % mineral volume change Figure 4.2: (continued)

PAGE 98

86 CHAPTER 5 DISCUSSION Reaction Progress in Low-Grade Metabasalts Petrographic observations of low-grade metaba salts (cf. Chapter 3) show a definite progression from mafic phyllosilicat es to zeolites to late stage zeolites. In some cases, residual pore space is left over af ter alteration. Alteration with in the matrix is dominated by the presence of Fe(Ti) oxides, and mafic phyllosilicates along with variable amounts of silica, feldspars (albite and microcline at higher grades), and zeo lites. Clinopyroxenes tend to survive alteration, while other mafi c phases (olivine, orthopyroxene, and glass) tend to be completely altered, often pse udomorphically. The reaction path modeling described in Chapter 4 indicates that this is a consequence of inc ongruent dissolution of the lavas. The sequence of secondary mine ralization is controlled by the differential dissolution rates for primary ba saltic phases as well as the volume change throughout the reaction due to the variation in molar volumes of secondary minerals. In the olivine tholeiite (Figure 4.1B), the first modeled mine rals to form are titanite, serpentines and chlorite brought on by early dissolution of o livine and basaltic glass, which have the highest dissolution rates (T able 4.2). Figure 4.2B show s an instantaneous volume increase of 17% related to the high molar volu mes of the serpentines and chlorites. This positive volume change forces precipitation to progress from the matrix to the vesicle in search of available pore space, forming observed clay rims lining vesicle walls. The onset of feldspar and pyroxene dissolu tion as the alteration progresses leads to zeolite precipitation. There is still precipitation of the early forming minerals, but the

PAGE 99

87 system becomes dominated by these zeolites due to their large molar volumes compared to both the primary basaltic phases and the mafic phyllosilicates. They create a final instantaneous volume increase of 14% (Figure 4.2B), causing occlusion of any residual pore space within vesicles. The model predic ts mesolite to form first with late stage formation of scolecite, which is a comm on sequence observed in samples exhibiting mesolite-scolecite zone alteration (pers. com.; Neuhoff, P.S., 2006). A similar progression of secondary mineraliz ation is seen in the modeled alteration of basaltic andesite (Figure 4.1C), where he matite, smectite and chlorite are the first minerals to form with later progression to th e zeolite laumontite. The early dissolution of orthopyroxene and anorthite due to their high dissolution rates (Tab le 4.4) leads to the initial precipitation of Fe(Ti) oxides and mafic phyllosilicates. The small molar volume of hematite creates no volume increase duri ng precipitation, allowing precipitation to occur in situ without progressing to open vesi cle space. As the levels of magnesium and aluminum dissolved into the aqueous so lution increase (Table 4.3), clay mineral precipitation dominates the system. The pos itive instantaneous volume change reaching 14.41% at this stage in the reaction (Figure 4.2C) is then due to the high molar volumes of the chlorite and smectite, causing preci pitation to migrate into open pore space and form the observed chlorite rims lining vesicle wa lls in these samples (Chapter 3). At this stage in the reaction, the mode l also shows an increase in the amount of quartz in the system, which is observed petrographically as sporadic appearances of silica within the matrix or occasional silica rims observed lining pores. As the reaction progresses, the aqueous phase acquires enough calcium to become saturated with respect to the zeolite laumontite. This shift from clay minerals to zeolites

PAGE 100

88 appears to be sudden due to the observed cons istency in clay rim thicknesses. In other samples of metabasalts where no clay rims line vesicle walls, this shift may have occurred before significant volume change due to clay precipitati on, whereas very thick clay rims would suggest a large increase in vo lume prior to zeolite formation. Once this shift occurs, one or more generations of zeolites will fill any remaining pore space in the basalt, which is usually within or immedi ately surrounding vesicles. In the modeled basaltic andesite, the precipitation of laumontite causes an instantaneous volume increase of 8.84% (Figure 4.2C) and fills remaining por e space in vesicles. Smaller vesicles which have been closed dur ing earlier stages of al teration will not experience zeolitization, and the relative amount of zeolite crystallization in a given vesicle will increase with vesicle size due to th e consistency in clay rim thicknesses. Dependence of Reaction Progress on Pore Size Observations described in preceding chapters suggest that the alteration observed in and around vesicles in mafic lavas occurs thro ugh fluid-rock interact ion at the scale of individual pores. Most notably, the prominent vesicl e size dependence on reaction progress noted in Chapter 3 indicates that chemical components are not homogenized across the lava during alterati on, but rather that each vesi cle independently exchanges chemical components with the surrounding la va matrix. This is most prominently exhibited by the constancy of mafic phyllosilicate rim thic knesses. In addition, the variations in zeolite minera logy observed between pores in low-grade metabasalt samples observed in previous studies (e.g., Neuhoff et al. 1999; 2006) suggest that local interactions between fluids within vesicles and the surr ounding matrix are important. This is not to say that there is not ch emical communication between pores, as the relatively high permeabilities of vesicular lavas noted above requires that advective-

PAGE 101

89 dispersive transport leads to solute tran sport between vesicles. Nonetheless, the prominence of vesicle-scale chemical inte raction leads to a number of important conclusions about the coupling between poros ity and chemical alteration in vesicular lavas. The dependence of reaction progress on pore size (e.g., whether zeolites will be developed in a given pore) can be explaine d using these results of the EQ3/6 models through the use of geometric models. A syst em is considered where each pore and the lava immediately surrounding it (i.e., the re action aureole) act as a closed (batch dissolution) system. Figure 5.1 shows a schematic representation of a spherical vesicle of radius rpore surrounded by a reaction aureole of thickness xaureole, and the radius of the system (rsystem) is equivalent to rpore plus xaureole. I can then use this system to interpret the % volume change needed to fill in a pore space given set values of pore size and aureole size. Figure 5.1: Schematic representation of a spherical pore of radius rpore surrounded by a reaction aureole of thickness xauroele. The radius of the system (rsystem) is equivalent to rpore plus xaureole. x aureole r porer system

PAGE 102

90 The plot in Figure 5.2 depicts the percentage of the volume of the system shown in Figure 5.1 occupied by the vesicle in a give n pore-aureole local reaction region as a function of rpore and x using pore sizes and aureole thickness typical of petrographic observations. For a vesicle of a constant size, the necessary aureole thickness required to fill the pore decreases as the % volume change during reaction increases. Inversely, if I increase the aureole thickness (f or instance, if the rate of diffusion into the matrix is higher), then less volume increase is required to fill the pore. No te that as aureole thickness increases, the percen t of volume occupied by the pore in the system becomes relatively insensitive to pore si ze. Thus, factors influencing the size of the aureole (i.e., Figure 5.2: Graph depicting the percentage of volume occupied by the pore in a given pore-aureole local reaction region as a function of rpore and xaureole using pore sizes and aureole thicknesses typi cal of vesicular mafic lavas.

PAGE 103

91 the extent of aqueous diffusion between th e vesicle and the matrix) such as time, temperature, and lava texture should greatly affect the exte nt of mineral paragenesis and degree of porosity modification. In a system representative of typical alteration obse rved in vesicular basalts (Chapter 3), average vesicle radii range from 1-5mm and average total original porosities range from 10-20%. This constrains the aure ole thicknesses required to completely close off porosity through secondary mineralizati on to ~0.5-8 mm, and this observation is consistent with the thickness of the alterati on aureoles described in Chapter 3. If the increase in mineral volume during reaction ex ceeds the volume of th e pore, then reaction will stop. For example, in the above EQ 3/6 model of the olivine tholeiite, the instantaneous volume increase reaches 15% (Figure 4.2B) during mafic phyllosilicate precipitation. If the alteration aureole at th is point in the reacti on progress has a thickness of 1 mm, then all pores with radii of ~1 mm will be filled, whereas la rge pores still retain space for further mineral development (e.g., zeolite formation). The instantaneous volume increase during later zeolite precip itation reaches up to 20%, thus having the ability to fill larger pore sp ace and effectively decreasing porosity in the system to zero. I can then relate this geometric model to petrographic observations of pore size variations and alteration char acteristics to predict the am ount of basalt that must be altered to create these features Petrographic observations of alteration characteristics in the basaltic andesite samples from the NSVG shows initial clay rims lining vesicle walls with a consistent thickness of 0.52 +/0.12 mm, with the residual pore space filled by laumontite. These features are schematica lly represented in Figure 5.3, where the pore radius equals the clay rim thickness plus the ra dius of the laumontite infilling of residual

PAGE 104

92 pore space. The total thickness of the altera tion aureole can then be divided into the thickness required to create the consistent clay rim (xchlorite) plus the additional thickness required to fill residual pore space with laumontite (xlaum). These required thicknesses will be sensitive to the vesicle size, as shown in Figure 5.4 where aureole thickness is plotted with respect to pore radius. Figure 5.3: Schematic representation of a spherical pore showing secondary infilling typical of the petrog raphically observed alteration in the basaltic andesite of the NSVG. The vesicle with radius rpore is surrounded by a reaction aureole with thickness xtotal. Modeled alteration of this sample show s that the instanta neous volume change during mafic phyllosilicate precipitation is 14.41%, while the instantaneous volume change during precipitation of laumontite is only 8.84%. This suggests that for a given pore radius, the required amount of basalt altera tion to create the chlorite rims will be less than the amount required to create the laumontite infilling. Using Figure 4.6 and assuming a pore radius of 5mm shows that the aureole thickness required to create the chlorite rim is ~2 mm while the entire thickn ess has to be ~5.5 mm to fill the remaining pore space with laumontite. Xchlorite rpore Xtotal laumontite chlorite

PAGE 105

93 The spherical model in Figure 5.1 represen ts one end-member in a continuum of vesicle shapes that are present in lavas. Ve sicles shapes in many flows vary from nearly spherical to highly elongate, and this w ill have a profound imp act on the degree of reaction progress observed in a given vesicle. For instance, Figure 5.5 shows the percent Figure 5.4: Plot depicting the aureole thickne ss (x) required to completely fill pore space within a vesicle of a given pore radius (rpore) with alteration minerals, shown for total infilling and for cl ay or laumontite infilling. of chlorite infilling vesicles with volum es equivalent to spheres with radii (rpore) of 0.25, 0.5, and 1 cm assuming a constant chlorite rim thickness of 0.5 mm as a function of the ratio of the length of the long axis over th at of the shorter axes (assuming elongation in only one direction). For all pores, the percentage of chlorite infilling increases exponentially with increasing ellipticity (axial ratio). As pore size decreases, the effect of

PAGE 106

94 elongation one the degree of clay infilling incr eases dramatically. Thus, vesicle shape as well as vesicle size can exert a profound infl uence on the degree of reaction progress. 0 10 20 30 40 50 60 70 80 90 100 0100200300Axial RatioChlorite Rim (% ) rpore = 0.5 rpore = 0.25 rpore = 1 Figure 5.5: Percent chlorite infil ling as a function of the axial ratio (ratio of the length of the long axis over that of the two shor ter axes) for an ellipsoidal vesicle elongated in one direction assuming a constant volume equivalent to a spherical vesicle with radii of 0.25, 0.5, a nd 1 cm and a constant chlorite rim thickness of 0.5 mm. This system is sensitive to many variables, and any small change will cause the numerical values given above to change. For example, the fluid composition used in these models is a dilute aqueous solution wh ich may not be representative of groundwater compositions present during the tim e of alteration. More chem ical constituents available in the initial fluid would allow for more s econdary mineral precipi tation, thus creating a larger positive volume change and changing the exact values rela ting the pore radius, aureole thickness and volume % required to o cclude porosity. Also, the temperature and pressure conditions for each modeled r un are determined based on known conditions

PAGE 107

95 required to form the observed zeolite altera tion; these minerals do occur over a small range of conditions, however, and the exact conditions of the mode l may not coincide exactly with conditions experien ced during alteration. This can also be said for other set parameters in the model where assumptions are made in order to best represent the expected conditions, as the model sensitivity to slight variations in these parameters may alter the amount of secondary minera lization experienced during alteration.

PAGE 108

96 CHAPTER 6 CONCLUSIONS Secondary mineral parageneses were used to delineate time-dependent changes in the nature of fluid storage capac ity and transmission in vesicula r lavas. The review of the present state of understanding of the hydrologic properties of ve sicular lavas in Chapter 1 led to several questions that formed th e motivation for this study. The detailed observations are summarized below with respect to these questions: 1. What causes the high permeabilities of vesicular volcanic rocks, even at low porosities? Fluid movement through vesicular basalt s occurs along a permeable network of microfractures and connected vesicles. The microfractures are usually not visible, but their locations are depicted by the alterati on aureoles that surr ound these pathways and the vesicles they connect. The basalt is highl y weathered within these aureoles signifying that these are preserving the pathways of fluid movement through the rock during alteration. This implies that vesicles will experience secondary alteration regardless of their size, shape and spacing as long as they are connected to the fluid flow pathways. This can explain the high permeability observe d in vesicular basalts with low porosity and little to no visible interc onnection between vesicles, as well as observed patches of completely unaltered vesicles within altered lavas. There are a number of proce sses that create various type s of microfractures within vesicular basalts. Not all of these microfractures, howev er, create a network of pathways through the matrix that connect vesicles as observed in the basalts of the NSVG. The

PAGE 109

97 strong presence of spherical vesicles in th e high permeability zones of the cooled lava, however, suggests that these vesicles retain gas after cooling which must later escape through the solid rock. This escaping gas w ill create a connected ne twork of pathways between vesicles and through the matrix, wh ich can then be reused as fluid flow pathways during later events of low-grade alteration. 2. Does the extent of reaction progress (as evidenced by the tem poral sequence of mineral infillings of pore space) vary w ith aspects of pore geometry such as spacing, size, and shape? AND 3. What is the relationship between pore geometry and porosity evolution during mineral paragenesis? Though the vesicle-vesicle spacing shows no control on the extent of reaction progress experienced by a given vesicle, th ere is a heavy influence imposed by the vesicle size and shape. Observations show that mineral assemblages within larger vesicles exhibit greater degrees of reacti on progress while smaller ones will become closed off earlier in the reaction progress. In situations where vesicl es are ellipsoidal or irregularly shaped, they may close off earlie r in the reaction progr ess even with large areas due to constriction cause d by the shortest diameter. This dependence of reaction progress on pore size is not purely based on the size of the pore, but on the volume % that this por e occupies within a closed pore-aureole system. It has been shown that each vesicle independently exchanges chemical components with the surrounding lava matrix, wh ich implies that the progress of reaction exhibited within a given vesicle is based on the ratio of the vesicle to aureole size. A vesicle that may be small compared to the average size within a given sample has the ability to show low levels of reaction progress if the alterati on around this vesicle is also

PAGE 110

98 very small. Inversely, a smaller alteration aureole around a larger vesicle implies less fluid-rock interaction, which may be seen in lower grades of alte ration or younger lavas, as this may not create enough secondary mine ralization to progress to later stages of alteration or to completely fill the pore space.

PAGE 111

99 LIST OF REFERENCES Aharonov, E., Spiegelman, M., and Keleme n, P., 1997, Three-dimensional flow and reaction in porous media: Implications for the Earth’s mantle and sedimentary basins: Journal of Geophysic al Research, v. 102, p. 14,821-14,834. Allen, D.J., Hinze, W.J., Dickas, A.B., Mudrey, M.G. Jr, 1997, Integrated geophysical modeling of the North American Midcontin ent Rift System: New interpretations for western Lake Superior, north-weste rn Wisconsin, and eastern Minnesota, i n Ojakangas, R.W., Dickas, A.B., and Gr een, J.C., eds., Middle Proterozoic to Cambrian Rifting, Central North America: Boulder, Colorado, Geological Society of America Special Paper 312. Allen, D., Hinze, W., and Cannon, W., 1992, Drainage, topographic, and gravity anomalies in the Lake Superior region : Evidence for a 1100 Ma mantle plume: Geophysical Research Letters, v. 19, p. 2119-2122. Al-Harthi, A.A., Al-Amri, R.M., and Sheh ata, W.M., 1999, The porosity and engineering properties of vesicular basalt in Saudi Arabia: Engineering Geology, v. 54, p. 313320. Arnorsson, S., 1995a, Geothermal systems in Ic eland: Structure and conceptual models. 1. High temperature areas. Geothermics, v. 24, p. 561-602. Arnorsson, S., 1995b, Geothermal systems in Icel and: Structure and conceptual models. 2. Low temperature Areas: Geothermics, v. 24, p. 603-629. Balashov, V.N., and Yardley, B.W.D., 1998, M odeling metamorphic fluid flow with reaction-compaction-permeability feedbacks: American Journal of Science, v. 298, p. 441-470. Basaltic Volcanism Study Project, 1981, Basalt ic volcanism on the terrestrial planets: Pergamon Press, New York, 1289 p. Bevins, R.E., Robinson, D., and Rowbotham G., 1991, Compositional variations in mafic phyllosilicates from regional low-grade meta basites and application of the chlorite geothermometer: Journal of Metamorphic Geology, v. 9, p. 711-721. Blower, J.D., 2001a, Factors controlling poros ity-permeability relationships in magma: Bulletin of Volcanology, v. 63, p. 497-503.

PAGE 112

100 Blower, J. D., 2001b, A three-dimensional ne twork model of permeability in vesicular material: Computers and Geosciences, v. 27, p. 115-119. Boerboom, T., Miller, J., Green, J., 2004, Ge ologic highlights of new mapping in the southwestern sequence of the North Shor e Volcanic Group and in the Beaver Bay Complex, in Severson, M. J., and Heinz, J., ed., Institute on Lake Superior geology: Duluth, Part 2 – Field Trip Guidebook, p. 46-85. Bolton, E.W., Lasaga, A.C., and Rye, D.M., 1999, Long-term flow/chemistry feedback in a porous medium with heterogeneous perm eability: kinetic control of dissolution and precipitation: American J ournal of Science, v. 299, p. 1-68. Bosl, W.J., Dvorkin, J., Nur, A., 1998, A st udy of porosity and permeability using a lattice Boltzmann simulation: Geophysic al Research Letters, v. 25, no. 9, p. 14751478. Brannon, J.C., 1984, Geochemistry of successive lava flows of the Keweenawan North Shore Volcanic Group; Doctoral Thesis, Washington University, St. Louis, Mo. Broedehoeft, J.D., and Norton, D.L., 1990, Ma ss and energy transport in deforming Earth’s crust, in The Role of Fluids in Crustal Processes, Washington, D.C.: National Academy, p. 27-41. Burgisser, A., Gardner, J.E., 2004, Expe rimental constraint s on degassing and permeability in volcanic conduit flow: Bull. Volcanol., v. 67, p. 42-56. Cannon, W.F., 1994, Closing of the midcontinent ri ft – a far-field effect of Grenvillian compression: Geology, v. 22, p. 155-158. Cannon, W.F., 1992, The Midcontinent rift in th e Lake Superior region with emphasis on its geodynamic evolution: Tectonophysics, v. 213, p. 41-48. Cannon, W.F., Cannon, W.C., Green, A.G., Hutchi nson, D.R., Lee, M.W., Milkereit, B., Behrendt, J.C., Halls, H.C., Green, J.C., Dickas, A.B., Morey, G.B., Sutcliffe, R., Spencer, C., 1989, The Midcontinent rift be neath Lake Superior from GLIMPCE seismic reflection profiling: Tectonics, v. 8, p. 305-332. Carman, P.C., 1956, Flow of Gases Through Po rous Media: Academic, San Diego, 182 p. Cashman, K.V., and Mangan, M.T., 1994, Physical aspects of magma tic degassing II. Constraints on vesiculation processes from textural studies of eruptive products, in Carroll, M.R., and Holloway, J.R. (eds) Volatiles in magmas: Mineralogical Society of America Reviews in Mineralogy 30, p. 447–478. Cathles, L.M., 1981, Fluid flow and hydrot hermal ore deposits: Economic Geology, v. 75, p. 424-457.

PAGE 113

101 Cathles, L.M., 1990, Scales and effects of flui d flow in the upper crust: Science, v. 248, p. 323-328. Christiansen, F.G., 1994, Seeps and other bitumen showings: A review of origin, nomenclature and occurrences in Gree nland: Open File Series Grnlands Geologiske Undersgelse, 94/7, 41 p. Christiansen, F. G., Boesen, A., Bojesen-Koef oed, J., Dallhoff, F., Dam, G., Neuhoff, P. S., Pedersen, A. K., Pedersen G. K., St annius, L. S., and Zinck-Joergensen, K., 1999, Petroleum geological activities onshor e West Greenland in 1997: Geology of Greenland Survey Bulletin, v. 180, p. 10-17. Coffin, M.F., Eldholm, O., 1994, Large Igneous Provinces: Crustal stru cture, dimensions, and external consequences: Re views of Geophysics, v. 32, p. 1-36. Dewers, T., and Ortoleva, P., 1990, Geoche mical self-organization III: Mechanochemical mode of metamorphic differentia tion: American Journal of Science, v. 290, p. 473-521. Dullien, F.A.L., 1992, Porous media; fluid transp ort and pore structure: Academic Press, San Diego, CA, United States, 574p. Eichelberger, J. C., Carrigan, C. R., West rich, H. R., and Price, R. H., 1986, Nonexplosive silicic volcanis m: Nature, v. 323, p. 598-602. Etheridge, M.A., Wall, V.J., and Vernon, R. H., 1983, The role of the fluid phase during regional metamorphism and deformation: Journal of Metamorphic Geology, v. 1, p. 205-226. Feng, S., Halperin, B.I., Sen, P.N., 1987, Transp ort properties of con tinuum systems near the percolation threshold: P hysical Review B, v. 35, p. 197-214. Fisher, A.T., 1998, Permeability within basalti c oceanic crust: Reviews of Geophysics, v. 36, p. 143-182. Freedman, V.L., Saripalli, K.P., and Meyer, P.D., 2003, Influence of mineral precipitation and dissolution on hydrologic prope rties of porous media in static and dynamic systems: Applied Geochemistry, v. 18, p. 589-606. Gardner, J.E., Thomas, R.M.E., Jaupart, C., Tait, S., 1996, Fragmentation of magma during Plinian volcanic eruptions : Bull. Volcanol., v. 58, p. 144-162. Garven, 1989, The role of regional fluid flow in the genesis of the Pine Point deposit. reply: Economic Geology, v. 81, p. 1015-1020. Garven, G., 1985, The role of regional fluid flow in the genesis of the Pine Point deposit: Economic Geology, v. 81, p. 307-324.

PAGE 114

102 Garven, G., 1995, Continental-scale groundwater flow and geologic processes: Annual Review of Earth and Planetary Sciences, v. 24, p. 89-117. Garven, G., and Freeze, R.A., 1984a, Theoretical analysis of the role of groundwater flow in the genesis of a stratabound ore depos it: 1. Mathematical and numerical model: American journal of Science, v. 284, p. 1085-1124. Garven, G., and Freeze, R.A., 1984b, Theoretical analysis of the role of groundwater flow in the genesis of a stratabound ore deposit: 2. Quantitative results: AmericanJournal of Science, v. 284, p. 1125-1174. Garven, G., and Raffensperger, J.P., 1994, Hydrogeology and geochemistry of ore genesis in sedimentary basins, in Barnes, H.L., ed., The Geochemistry of Hydrothermal Ore Depos its, ch. 4, in press. Garven, G., Ge, S., Person, M.A., and Sverje nsky, D.A., 1993, Genesis of stratabound ore deposits in the Midcontinent basins of North America. 1. The role of regional groundwater flow: American Jour nal of Science, v. 293, p. 497-568 Gislason, S.R., and Eugster, H.P., 1987, Mete oric water-basalt interactions; I, A laboratory study: Geochimica et Cosmochimica Acta, v. 51, no.10, p. 2827-2840. Green, J.C., 1982, Geology of the Keweenawan extrusive rocks: Geological Society of America Memoir 156. Grimmett, G., 1999, Percolation, 2nd edition: Berlin Heider lberg New York, Springer. Hearn, P.P, Jr., Steinkampf, W.C., Horton, D.G., Solomon, G.C., White, L.D., and Evans, J.R., 1989, Oxygen-isotope composition of ground water and secondary minerals in Columbia Plateau basalts; implications for the paleohydrology of the Pasco Basin: Geology, v. 17, p. 606-610. Hinze, W.J., David, J.A., Lawrence, W.B ., Mariano, J., 1997, The Midcontinent Rift System: A major Proterozoic continental rift, i n Ojakangas, R.W., Dickas, A.B., and Green, J.C., eds., Middle Proterozoi c to Cambrian Rifting, Central North America: Boulder, Colorado, Geological Society of America Special Paper 312. Iijima, A., 2001, Zeolites in petr oleum and natural gass reservoirs, in Bish, D.L., and Ming, D.W., eds., Natural Zeolites: Occurr ence, Properties, Applications: Reviews in Mineralogy and Geochemistry, Minera logical Society of America and the Geochemical Society, Washington, D.C., p. 347-402. Ingebritsen, S. E., and M. A. Scho ll, 1993, The hydrology of Kilauea volcano. Geothermics, v. 22, p. 255-270. Jakobsson, S.P., and Moore, J.G., 1986, Hydrothe rmal minerals and alteration rates at Surtsey volcano, Iceland: Geological So ciety of America Bulletin, v. 97, p. 648659.

PAGE 115

103 Johnson, J.W., Oelkers, E.H., and Helges on, H., 1992, CSUPCRT92; a software package for calculating the standard molal therm odynamic properties of minerals, gases, aqueous species, and reactions from 1 to 5000 bar and 0 to 1000 degrees C: Computers & Geosciences, vol.18, no.7, pp.899-947. Johnson, D. M., 1980, Crack distribution in th e upper oceanic crust and its effects upon seismic velocity, seismic structure, formation permeability, and fluid circulation: Initial Reports, Deep Sea Dr illing Project, v. 51-53, p. 1479-1490. Jorgensen, O., 1984, Zeolite zones in the ba saltic lavas of th e Faeroe Islands, in The deep drilling project 1980-1981 in the Faeroe Is lands: Annales Societatia Scientiarum Faroensis, Supplementum, v. 9, p. 71-91. Klug, C., and Cashman, K. V., 1996, Permeabilit y development in vesiculating magmas: implications for fragmentation: Bulletin of Volcanology, v. 58, p. 87-100. Kononov, 1978, Hydrogeology of Iceland: Inte rnational Geology Review, v. 21, no. 4, p. 385-398. Kowallis, B.J., Roeloffs, E.A., and Wang, H.F ., 1982, Microcrack studies of basalts from the Iceland Research Drill ing Project: Journal of Ge ophysical Research, v. 87, no. B8, p. 6650-6656. Kristmannsdttir, H., and Tmasson, J., 1978, Zeolite zones in geothermal areas in Iceland, in Sand, L. B., and Mumpton, F. A., ed., Natural Zeolites: Oxford, Pergamon Press Ltd., p. 277-284. Larsen, J.F., Denis, M.H., Gardner, J.E ., 2004, Experimental study of bubble coalescence in rhyolitic and phonolitic melts: Geochemica et Cosmochimica Acta, v. 68, no. 2, p. 333-344. Lasaga, A.C., 1986, Metamorphic reaction rate laws and the development of isograds: Mineralogical Magazine, v. 50, p. 359-373. Lasaga, A.C., 1989, Fluid flow and chemical r eaction kinetics in metamorphic systems: a new simple model: Earth and Planet ary Science Letters, v. 94, p. 417-424. Lasaga, A.C., and Rye, D.M., 1993, Fluid fl ow and chemical reaction kinetics in metamorphic systems: American Journal of Science, v. 293, p. 361-404. Lee, S. B., 1990, Universality of continuum percolation: Physics Reviews, v. 42, p. 48774880. Le Gallo, Y., Bildstein, O., and Brosse, E., 1998, Coupled reaction-flow modeling of diagenetic changes in reservoir permeability, porosity, and mineral compositions: Journal of Hydrology, v. 209, p. 366-388.

PAGE 116

104 Le Maitre, R.W., Bateman, P., Dudek, A., Keller, J., Lameyre Le Bas, M.J., Sabine, P.A., Schmid, R., Sorensen, H., Streckeisen, A., Woolley, A.R., and Zanettin, B., 1989, A Classification of Igneous Rocks and Gl ossary of Terms: Oxford, Blackwell. Lichtner, P.C., Steefel, C.I., and Oelkers, E.H., eds., 1996, Reactive transport in porous Media: Mineralogical Society of Amer ican Reviews in Mineralogy v. 34, 438 p. Luo, J., Morad, S., Liang, Z., and Zhu, Y., 2005, Controls on the quality of Archean metamorphic and Jurassic volcanic reservoir rocks from the Xinglongtai buried hill, western depression of Liaohe basin, Chin a: American Association of Petroleum Geologists Bulletin, v. 89, p. 1319-1346. Lysak, S.V., 1992, Heat flow variations in continental rifts: Tectonophysics, v. 208, p. 309–323. Manning, C.E., and Bird, D.K., 1991, Porosity ev olution and fluid flow in the basalts of the Skaergaard magma-hydrothermal system, East Greenland: American Journal of Science, v. 291, no. 3, p. 201-257. Manning, C.E., and Bird, D.K., 1995, Porosit y, permeability, and basalt metamorphism: Special Paper – Geological Soci ety of America, v. 296, p. 123-140. Manning, C.E., Ingebritsen, S.E., and Bird, D.K., 1993, Missing mineral zones in contact metamorphosed basalts. American Journal of Science, v. 293, p. 894-938. Manga, M., 1997, A model for discharge in spri ng-dominated streams and implications for the transmissivity and recharge of quaternary volcanics in the Oregon Cascades: Water Resources Research, v. 33, p. 1813.1822. Manga, M., 1999, On the timescales characte rizing groundwater discharge at springs: Journal of Hydrology, v. 219, p. 56-69. Miller, J.D., Jr., Green, J.C., Severson, M.J ., Chandler, V.W., and Peterson, D.M., 2001, Geologic map of the Duluth Complex and related rocks, northeastern Minnesota: Minnesota Geological Survey Misc ellaneous Map M-119, scale 1:200000. Miller, J.D., Jr., Green, J.C., Severson, M. J., Chandler, V.W., Hauck, S.A., Peterson, D.M., and Wahl, T.E., 2002, Geology and mine ral potential of the Duluth Complex and related rocks of northeastern Minneso ta: Minnesota Geological Survey Report of Investigations 58, 207 p. Mueller, S., Melnik, O., Spieler, O., Sc heu, B., Dingwell, D.B., 2004, Permeability and degassing of dome lavas undergoing rapi d decompression: An experimental determination: Bull. Volcanol., v. 60, p. 526-538. Murata, K. J., Formoso, M. L. L., and Ro isenberg, A., 1987, Distribution of zeolites in lavas of southeastern Parana Basin, state of Rio Grande do Sul, Brazil: Journal of Geology, v. 95, p. 455-467.

PAGE 117

105 Neuhoff, P.S., Rogers, K.L., Stannius, L.S ., Bird, D.K., Pedersen, A.K., 2006, Regional very low-grade metamorphism of basaltic lavas, Disko-Nuussuaq Region, West Greenland: Lithos, in press Neuhoff, P. S., Fridriksson, T., Bird, D. K., 2000, Zeolite paragenesis in the North Atlantic Igneous Province: implications for geotectonics and groundwater quality of basaltic crust: Internationa l Geology Review, v. 42, p. 15-44. Neuhoff, P. S., Fridriksson, T., Arnorsson, S., 1999, Porosity evolution and mineral paragenesis during low-grade metamorphi sm of basaltic lavas at Teigarhorn, Eastern Iceland: American Jour nal of Science, v. 299, p. 467-501. Neuhoff, P.S., Watt, W.S., Bird, D.K., and Pedersen, A.K., 1997, Timing and structural relations of regional zeolite zones in ba salts of the East Greenland continental margin: Geology, v. 25, p. 803-806. Norton, D., 1988, Metasomatism and permeability: American Journal of Science, v. 288, p. 604-618. Norton, D.L., 1984, Theory of hydrothermal systems: Annual Review of Earth and Planetary Sciences, v. 12, p. 155-177. Norton, D., and Knight, J., 1977, Transport pheno mena in hydrothermal systems: cooling plutons: American Journal of Science, v. 277, p. 937-981. Ojakangas, R.W., and Dickas, A.B., 2002, The 1.1-Ga Midcontinent Rift System, central North America: sedimentology of two d eep boreholes, Lake Superior region: Sedimentary Geology, v. 147, p. 13-36. Ojakangas, R.W., and Matsch, C.L., 1982, Minnesota’s Geology: Minneapolis, the University of Minnesota Press, 255 p. Ortoleva, P., Chadam, J., Merino, E., and Sen, A., 1987, Geochemical self-organization II: The reactive-infiltration instability: American Jour nal of Science, v. 287, p. 1008-1040. Oxburgh, E.R., and Argell, S.G., 1982, Therma l conductivity and temperature structure of the Reydarfjrdur borehole: Journa l of Geophysical Research, v. 87, p. 6423– 6428. Palandri, J.L., and Kharaka, Y.K., 2004, A compilation of rate parameters of watermineral interaction kinetics for applica tion to geochemical modeling: Open-File Report U. S. Geological Su rvey, Report: OF 2004-1068, 64 p. Panda, M. N. and Lake, L. W., 1995, A physi cal model of cementation and its effects on single phase permeability: American Association of Petroleum Geologists Bulletin, v. 79, p. 431-434.

PAGE 118

106 Park, A.J., and Ortoleva, P.J., 2003, WRIS.TEQ : Multimineralic wate r-rock interaction, mass transfer and textural dynamics simu lator: Computers and Geosciences, v. 29, p. 277-290. Pendkar, N., and Kumar, A., 1999, Delineation of reservoir section in Deccan Trap basement, example from Padra Field, Cambay Basin: Bulletin of the Oil and Natural Gas Corporation Limited, v. 36, p. 83-88. Person, M., Raffensperger, J.P., Ge, S., a nd Garven, G., 1996, Basin-scale hydrogeologic modeling: Reviews of Geophysics, v. 34, p. 61-87. Putnis, A., 2002, Mineral replacement reactions; from macroscopic observations to microscopic mechanisms: Mineralo gical Magazine, v. 66, no. 5, p 689-708. Raffensperger, J.P., and Garven, G., 1995a, Th e formation of unconformity-type uranium ore deposits. 1. Coupled groundwater flow and heat transport modeling: American Journal of Science, v. 295, p. 581-636. Raffensperger, J.P., and Garven, G., 1995b, Th e formation of unconformity-type uranium ore deposits. 2. Coupled hydrochemical mode ling: American Journal of Science, v. 295, p. 639-696. Robert, C., 2001, Hydrothermal alteration proces ses of the Tertiary lavas of Northern Ireland: Mineralogical Magazine, v. 65, p. 543-554. Robert, C., and Goff, B., 1993, Zeolitizati on of basalts in subaqueous freshwater settings: Field observations and experime ntal study: Geochimica et Cosmochimica Acta, v. 57, p. 3597-3612. Rose, N.M., 1995, Geochemical consequences of fluid flow in porous basaltic crust containing permeability contrasts: Geochimica et Cosmochimica Acta, v. 59, p. 4381-4392. Rose, T.P., Davisson, M.L., Criss, R.E., 1996, Isotope hydrology of voluminous cold springs in fractured rock from an activ e volcanic region, northeastern California: Journal of Hydrology, v. 179, p. 207.236. Saar, M.O., 1998, The relationship between perm eability, porosity and microstructure in vesicular basalts [Mas ter’s Thesis]: University of Oregon, p. 1-101. Saar, M. O., Manga, M., 1999, Permeability – poro sity relationship in vesicular basalts: Geophysical Research Letters, v. 26, no. 1, p 111-114. Sahagian, D.L., and Proussevitch, A.A., 1998, 3D particle size distributions from 2D observations: Stereology from natural app lications: Journal of Volcanology and Geothermal Research, v. 84, p. 173-196.

PAGE 119

107 Sahagian, D.L., Anderson, A.T., and Ward, B ., 1989, Bubble coalescence in basalt flows: Comparison of a numerical model with na tural examples: Bulletin of Volcanology, v. 52, p. 49-56. Sahagian, D.L., Proussevitch, A.A., and Carl son, W.D., 2002, Anal ysis of vesicular basalts and lava emplacement processes for application as a paleobarometer/paleoaltimeter: Th e Journal of Geology, v. 110, p. 671-685. Saemundsson, K., 1979, Outline of the geol ogy of Iceland: Joekull, no. 29, p.7-28. Sahimi, M., 1994, Applications of percolati on theory: Taylor and Francis, London, p. 1300. Sahimi, M., 1995, Flow and Transport in Po rous Media and Fractured Rocks: VHC Verlagsgesellschaft mb H, Weinheim, p. 1-496. Saripalli, K.P., Meyer, P.D., Parker, K.E., and Lindberg, M.J., 2005, Effect of chemical reactions on the hydrologic properties of fractured a nd rubbelized glass media: Applied Geochemistry, v. 20, p. 1677-1686. Saripalli, K.P., Meyer, P.D., Bacon, D. H., and Freedman, V.L., 2001, Changes in hydrologic properties of aquifer media due to chemical reactions: a review: Critical Reviews in Environmental Scie nce and Technology, v. 31, p. 311-349. Schiffman, P., and Fridleifsson, G."., 1991, The smectite-chlorite transition in drillhole NJ-15, Nesjavellir geothermal field, Icel and: XRD, BSE and electron microprobe investigations: Journal of Me tamorphic Geology, v. 9, p. 679-696. Schmidt, S.Th., 1993, Regional and local patter ns of low-grade metamorphism in the North Shore Volcanic Group, Minnesota, USA: Journal of Metamorphic Geology, v. 11, p. 401-414. Schmidt, S.T., 1990, Alteration under conditions of burial metamorphism in the North Shore Volcanic Group, Minnesota – Mine ralogical and geochemical zonation [Ph.D. thesis]: Heidelberg, Heidelberg er Geowissenschaftliche Abhandlungen, 309 p. Schmidt, S.T., Robinson, D., 1997, Metamor phic grade and porosity and permeability controls on mafic phyllosilicat e distributions in a regiona l zeolite to greenschist facies transition of the North Shore Volcanic Group, Minnesota: GSA Bulletin, v. 109, no. 6, p. 683-697. Song, S.-R., Jones, K.W., Lindquist, W.B., Do wn, B.A., and Sahagian, D.L., 2001, Synchrotron X-ray computed microtomogr aphy: studies on vesiculated basaltic rocks: Bulletin of Volcanology, v. 63, p. 252-263.

PAGE 120

108 Sruoga, P., Rubinstein, N., and Hinterwimmer, G., 2004, Po rosity and permeability in volcanic rocks: a case study on the Serie Tobfera, South Patagonia, Argentina: Journal of Volcanology and Geot hermal Research, v. 132, p. 31-43. Steefel, C.I., and Lasaga, A.C., 1992, Putting transport into water-rock interaction models: Geology, v. 20, p. 680-684. Steefel, C.I., and Lasaga, A.C., 1994, A coupled model for transport of multiple chemical species and kinetic precipitati on/dissolution reactions with applications to reactive flow in single phase hydrothermal systems: American Journal of Science, v. 294, p. 529-592. Steefel, C.I., and Van Cappellen, P., 1990, A ne w kinetic approach to modeling waterrock interaction: the ro le of nucleation, precursor s, and Ostwald ripening: Geochimica et Cosmochimica Acta, v. 54, p. 2657-2677. Stimac, J.A., Powell, T.S., and Golla, G.U., 2004, Porosity and permeability of the Tiwi geothermal field, Philippines based on continuous and spot core measurements: Geothermics, v. 33, p. 87-107. Sukheswala, R. N., Avasia, R. K., and Gangopadhyay, M., 1974, Zeolites and associated secondary minerals in the Deccan Traps of western India: Mineralogical Magazine, v. 39, p. 658-671. Tingle, T.N., Neuhoff, P.S., Ostergren, J.D., Jones, R.E., and Donovan, J.J., 1996, The effect of “missing” (unanalyzed) oxy gen on quantitative electron microprobe microanalysis of hydrous silicate and oxi de minerals: Geological Society of America, Abstracts with Programs, v. 28, p. A-212. Toramaru, A., 1990, Measurement of bubble size distribution in vesiculated rocks with implications for quantitative estimati on of eruptive processes: Geothermal Research, v. 43, p. 71-90. Van Schmus, W.R., and Hinze, W.J., 1985, The Midcontinent Rift System: Annual Reviews of Earth and Planetary Sciences, v. 13, p. 345-383. Vervoort, J.D., and Green, J.C., 1997, Origin of evolved magmas in the Midcontinent rift system, northeast Minnesota: Nd-isotope evidence for melting of Archean crust: Canadian Journal of Earth Science, v. 34, p. 521-535. Walker, G. P. L., 1951, The amygdale minerals in the Tertiary lavas of Ireland. I. The distribution of chabazite habits and zeol ites in the Garron plateau area, County Antrim: Mineralogy Magazine, v. 29, p. 773-791. Walker, G. P. L., 1960a, The amygdale minerals in the Tertiary la vas of Ireland. III. Regional distribution: Mineral ogical Magazine, v. 32, p. 503-527.

PAGE 121

109 Walker, G. P. L., 1960b, Zeolite zones and dike distribution in relation to the structure of the basalts of eastern Iceland: Journal of Geology, v. 68, p. 515-528. Wedepohl, K.H., 1969, Composition and abunda nce of common igneous rocks: New York, Springer, p. 227-249. White, W.S., 1966, Tectonics of the Keweenawan basin, western Lake Superior region: U.S. Geological Survey Prof essional Paper 524-E, p. E1-E23. Winter, J.D., 2001, An Introduction to Igne ous and Metamorphic Petrology: Upper Saddle River, Prentice-Hall Inc., 697 p. Wohletz, K.H., 1999, Magma, Los Alamos Na tional Laboratory computer code LA-CC 99-28, Los Alamos, NM. Wolery, T. J., 1979, Calculation of chemical equilibria between a queous solutions and minerals: The EQ3/6 software package: Lawrence Livermore National Laboratory Report, UCRL-52658. Wolery, T. J., Jackson, K. J., Bourcier, W. L ., Bruton, C. J., Viani, B. E., Knauss, K. B., and Delany, J. M., 1990, Current status of the EQ3/6 software package for geochemical modeling, in Melchior, D. C., and Bassett, R. L., ed., Chemical Modeling in Aqueous Systems II.: Am erican Chemical Society, p. 104-116. Wood, J.R., and Hewett, T.A., 1984, Reservoir diagenesis and convective fluid flow, in McDonald, D.A., and Surdam, R.C., ed s., Clastic Diagenesis: American Association of Petroleum Ge ologists Memoir, v. 37, p. 3-13. Xu, T., and Pruess, K., 2001, Modeling multi phase non-isothermal fluid flow and reactive geochemical transport in vari ably saturated fractured rocks: 1. Methodology: American Journal of Science, v. 301, p. 16-33.

PAGE 122

110 BIOGRAPHICAL SKETCH Jane E. Gustavson was born in Milwaukee, Wisconsin, in April of 1980 to James M. Gustavson. She graduated from Wauwat osa East High School in the Spring of 1998, and started college immediately in the Fall to work towards a degree in geology. After a semester at Winona State University and a year at the University of Wisconsin – Milwaukee, she spent her last two and a half years at the Univer sity of Wisconsin – Oshkosh, receiving her Bachelor of Science in 2002. In 2003, she began her graduate work in the Department of Geological Scien ces at the University of Florida under the guidance of Dr. Philip Neuhoff. Following graduation, Jane plans to pursuit a career in geology that focuses on practical ap plications of her thesis work.


Permanent Link: http://ufdc.ufl.edu/UFE0015883/00001

Material Information

Title: Analysis of porosity evolution during low temperature metamorphism of basaltic lavas and implications for fluid flow
Physical Description: Mixed Material
Language: English
Creator: Gustavson, Jane E. ( Dissertant )
Neuhoff, Philip S. ( Thesis advisor )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2006
Copyright Date: 2006

Subjects

Subjects / Keywords: Geological Sciences thesis, M.S
Dissertations, Academic -- UF -- Geological Sciences
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
theses   ( marcgt )

Notes

Abstract: Basaltic lavas are the most abundant rock type in the crust and globally form important reservoirs for the migration and storage of crustal fluids. As in any porous medium, the porosity and permeability of basalt lavas control the movement and extent of chemical interaction of these fluids and thus the availability and quality of obtainable resources (e.g., groundwater and petroleum). The highly-reactive nature of basaltic lavas near the surface leads to intimate coupling between hydrologic properties and extent of chemical alteration, which controls the aquifer/reservoir quality and chemical fluxes during alteration. Our ability to quantitatively understand these processes is limited by rigorous models of how chemical reactions and porosity/permeability modifications are coupled in these systems. In the present study, development of secondary mineral parageneses is used as a monitor of fluid access to pores and to obtain a quantitative description and interpretation of the evolution of pore space through time. This work has shown that fluid movement through vesicular basalts occurs along a permeable network of microfractures and connected vesicles, which are depicted by the extremely weathered aureoles of alteration surrounding these pathways and the vesicles they connect. This research also shows that the extent of reaction progress, as evidenced by the temporal sequence of mineral infillings within pore space, is heavily influenced by the vesicle size and shape. Larger vesicles exhibit greater degrees of reaction progress while smaller ones are closed off earlier in the reaction progress. Irregularly shaped vesicles may close off earlier in the reaction progress due to constriction caused by the shortest diameter. This dependence of reaction progress on pore size is more specifically based on the volume percent occupied by the pore within a closed pore-aureole system. Each vesicle independently exchanges chemical components with the surrounding lava matrix (the alteration aureole), implying that the reaction progress exhibited within a given vesicle is based on the proportion of the vesicle to aureole size. For a given vesicle size, the pore space evolution due to secondary mineralization will be less with a smaller alteration aureole, due to the low level of fluid rock interaction evidenced by this smaller alteration aureole. Inversely, a larger alteration aureole implies a higher level of fluid-rock interaction, which may create enough secondary mineralization to progress to later stages of alteration or to completely fill the pore space.
Subject: alteration, analcime, basalt, flow, fluid, heulandite, laumontite, lava, metamorphism, modeling, NSVG, paragenesis, porosity, progress, reaction, stilbite, thermodynamic, thomsonite, vesicular, zeolite
General Note: Title from title page of source document.
General Note: Document formatted into pages; contains 122 pages.
General Note: Includes vita.
Thesis: Thesis (M.S.)--University of Florida, 2006.
Bibliography: Includes bibliographical references.
General Note: Text (Electronic thesis) in PDF format.

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: aleph - 003614694
System ID: UFE0015883:00001

Permanent Link: http://ufdc.ufl.edu/UFE0015883/00001

Material Information

Title: Analysis of porosity evolution during low temperature metamorphism of basaltic lavas and implications for fluid flow
Physical Description: Mixed Material
Language: English
Creator: Gustavson, Jane E. ( Dissertant )
Neuhoff, Philip S. ( Thesis advisor )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2006
Copyright Date: 2006

Subjects

Subjects / Keywords: Geological Sciences thesis, M.S
Dissertations, Academic -- UF -- Geological Sciences
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
theses   ( marcgt )

Notes

Abstract: Basaltic lavas are the most abundant rock type in the crust and globally form important reservoirs for the migration and storage of crustal fluids. As in any porous medium, the porosity and permeability of basalt lavas control the movement and extent of chemical interaction of these fluids and thus the availability and quality of obtainable resources (e.g., groundwater and petroleum). The highly-reactive nature of basaltic lavas near the surface leads to intimate coupling between hydrologic properties and extent of chemical alteration, which controls the aquifer/reservoir quality and chemical fluxes during alteration. Our ability to quantitatively understand these processes is limited by rigorous models of how chemical reactions and porosity/permeability modifications are coupled in these systems. In the present study, development of secondary mineral parageneses is used as a monitor of fluid access to pores and to obtain a quantitative description and interpretation of the evolution of pore space through time. This work has shown that fluid movement through vesicular basalts occurs along a permeable network of microfractures and connected vesicles, which are depicted by the extremely weathered aureoles of alteration surrounding these pathways and the vesicles they connect. This research also shows that the extent of reaction progress, as evidenced by the temporal sequence of mineral infillings within pore space, is heavily influenced by the vesicle size and shape. Larger vesicles exhibit greater degrees of reaction progress while smaller ones are closed off earlier in the reaction progress. Irregularly shaped vesicles may close off earlier in the reaction progress due to constriction caused by the shortest diameter. This dependence of reaction progress on pore size is more specifically based on the volume percent occupied by the pore within a closed pore-aureole system. Each vesicle independently exchanges chemical components with the surrounding lava matrix (the alteration aureole), implying that the reaction progress exhibited within a given vesicle is based on the proportion of the vesicle to aureole size. For a given vesicle size, the pore space evolution due to secondary mineralization will be less with a smaller alteration aureole, due to the low level of fluid rock interaction evidenced by this smaller alteration aureole. Inversely, a larger alteration aureole implies a higher level of fluid-rock interaction, which may create enough secondary mineralization to progress to later stages of alteration or to completely fill the pore space.
Subject: alteration, analcime, basalt, flow, fluid, heulandite, laumontite, lava, metamorphism, modeling, NSVG, paragenesis, porosity, progress, reaction, stilbite, thermodynamic, thomsonite, vesicular, zeolite
General Note: Title from title page of source document.
General Note: Document formatted into pages; contains 122 pages.
General Note: Includes vita.
Thesis: Thesis (M.S.)--University of Florida, 2006.
Bibliography: Includes bibliographical references.
General Note: Text (Electronic thesis) in PDF format.

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: aleph - 003614694
System ID: UFE0015883:00001


This item has the following downloads:


Full Text












ANALYSIS OF POROSITY EVOLUTION DURING LOW TEMPERATURE
METAMORPHISM OF BASALTIC LAVAS AND IMPLICATIONS FOR FLUID
FLOW















By

JANE E. GUSTAVSON


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2006


































Copyright 2006

by

Jane E. Gustavson















ACKNOWLEDGMENTS

There are many people in my life who have been influential in helping me to obtain

my Master of Science. Of those, there are a few that require special mention. My father,

James M. Gustavson, did all that he knew to provide the best for his children. If it were

not for his selflessness and honest love, I would not be the person I am today. Allyn

Spear, though we have a mottled history, was the one who stood beside me and gave me

the courage to pursue a higher level of education when I was filled with so much self

doubt. Dr. Philip S. Neuhoff was always there with a patient yet guiding hand to see me

through my research. He has become more than just my advisor; he is also a good friend.

And finally, PJ Moore, who has become the companion I have searched for. His promise

to walk beside me gives me the strength to enter the real world with the confidence I need

to succeed.
















TABLE OF CONTENTS



A C K N O W L E D G M E N T S ......... .................................................................................... iii

LIST OF TABLES ............. ....... .... ........ .. ........ .......... .......... .. vi

LIST OF FIGURES ............. ................... ............ .......... ................ viii

ABSTRACT .............. ......................................... xi

CHAPTER

1 IN TR OD U CTION ............................................... .. ......................... ..

Porosity and Perm ability in Vesicular Lavas...........................................................3
Low Grade Alteration of Basaltic Lavas ..................................................................6
Coupling Between Porosity Evolution and Chemical Reactions .................................8
R ole of the Present study .............................................. .. ..... .. ........ .. ..

2 LOW-GRADE ALTERATION OF THE NORTH SHORE VOLCANIC GROUP,
MINNE SOTA ........................ ........................ 10

Introduction ......... ...... ............................................................................10
Geologic Background ................................ ........ ........ ...... .. ............ 12
M eth o d s .............................................................................. 15
R results ...................... ..... .... ... ...... ....................................17
Primary Basalt Composition and Mineralogy ......... .....................................17
Regional Alteration M ineralogy ............... ......... ......................................22
L aum ontite zone lavas...................... .. .. .......... ....................... ...........26
Stilbite-heulandite zone lavas ........... .................. ...............................27
Thom sonite-m esolite zone lavas ....................................... ............... 27
A nalcim e zone lavas............................. .................. ........................ 28
A lteration M ineral C hem istry .............................................................................29
D discussion ....................................... ............................ .......... 31
Regional Variation of Metamorphic Grade ......................................................31
C conditions of A lteration ............................................................... ............... 38
Structural Interpretations ......................................................... ............. 42










3 PETROGRAPHIC AND DIGITAL ANALYSIS OF POROSITY EVOLUTION
D U R IN G A L TE R A T IO N ............................................................... .....................45

In tro d u ctio n ........................................................................................................... 4 5
M e th o d s ..............................................................................4 7
F field T ech n iq u e s ........................................................................................... 4 7
O ptical and D igital Techniques .................................................................... 48
Observations ......................................................................49
N SV G Field Sites 3, 4 and 6 ....................................................... 49
Reaction Progress ........................ ............ ............................ 55
Partial infilling at low-grades ......... ... ......... ..... ...... .. ............... 55
Mafic phyllosilicate to zeolite infilling ..................................... .......... 58
Multiple stage zeolite infilling ...........................................63
D discussion ................. ........... .. .........................................................63
Nature of Porosity and Permeability in Basalts................................................63
Controls on R action Progress ........................................ ........................68

4 MODELING OF MINERAL PARAGENESES ................................................70

In tro d u ctio n ........................................................................................................... 7 0
M e th o d s ..............................................................................7 1
R e su lts .........................................................................................................7 5
M ineral P arageneses ...................................................... .............. ... ............... 75
Volume Changes During Alteration .................................. ............ 83

5 D ISCU SSION ................ ........ ............................................... ..... 86

Reaction Progress in Low-Grade Metabasalts............... ...................... .......86
Dependence of Reaction Progress on Pore Size ......................... ....................... 88

6 CON CLU SION S .................................. .. .......... .. .............96

LIST O F R EFEREN CE S ............... ................................................ ......... ...... 99

BIOGRAPH ICAL SKETCH ............... ............. ............................... ............... 110

















v
















LIST OF TABLES


Table p

2.1 Field study locations in the North Shore Volcanic Group. .....................................16

2.2 Whole-rock chemical compositions (wt %) of samples ................ ..................18

2.3 Representative compositions of plagioclase. ............. .............................................20

2.4 Representative compositions of pyroxenes. ......................................................21

2.5 Representative compositions of mafic phyllosilicates. .........................................30

2.6 Representative compositions of thomsonites. ......................................................32

2.7 Representative compositions of mesolite, analcime, and laumontite....................34

3.1 Statistical analysis of vesicle size (diameter) as a function of reaction progress
from outcrop scale measurements at Site 6. ................................... ............... 51

3.2 Statistical analysis of vesicle size (area) as a function of reaction progress from
measured vesicles of a low-grade metabasalt from East Greenland (thin section
4 2 1 5 0 5 ) ................................................ ................................................ ...5 7

3.3 Statistical analysis of vesicle size as a function of reaction progress for vesicles
filled with chlorite and/or laumontite at the thin section scale (sample NS04-14)
fro m S ite 4 ........................................................................ 5 9

3.4 Statistical analysis of vesicle areas (mm2) and clay rim thicknesses (mm) for
sample 94-80 from eastern Iceland. .............................................. ............... 63

4.1 Bulk rock composition for olivine tholeiite basalt (wt %) .................. ..................72

4.2 Modal abundances and compositions of primary mineral phases in olivine
th o le iite ....................................................... ................ 7 2

4.3 Calculated anhydrous bulk composition of basaltic andesite before and after
reaction ..............................................................................74

4.4 Reactant phase composition, abundance, and relative dissolution rate in basaltic
andesite used as input for reaction path model............... ..................... ................75









4.5 Mole, volume and mass amounts in unaltered and altered andesite for the
computer (modeled) and mathematical (calculated) models...............................81

4.6 Phase compositions and abundances at end of reaction path for basaltic andesite..82















LIST OF FIGURES


Figure page

1.1 Cross sectional view showing distribution of primary porosity vesicless, scoria,
breccia) typical of thick aa lava flows.............. .............................................. 4

1.2 Schematic diagrams depicting the effects of pore size, shape and connectivity on
the permeability of vesicular basalts. ....................................................................... 4

2.1 Generalized geological map of northeastern Minnesota showing the distribution
of the NSVG lavas and associated Keweenawan intrusives (after Miller et al.,
2 00 1) ......... ........................................................ ....................... 11

2.2 Flow boundary between two vesicular basalt flows at Site 12. ............................14

2.3 Total alkalis-silica diagram showing compositional ranges of NSVG basalts
sampled in this study (black circles). ............................................ ............... 19

2.4 Field photo from Site 3 (Table 2.1) showing an extensively altered pahoehoe
flow. Visible within the flow are bleached haloes around vesicles connected by
thin anastom osing bleached areas. ........................................ ....................... 23

2.5 Secondary alteration in NSVG lavas. All photomicrographs were taken through
partially crossed polars ..................... ....... ..................................... ............... 24

2.6 Compositions of mafic phyllosilicates formed during regional metamorphism of
the NSVG as a function of the number of non-interlayer cations (Si + Al + Mg +
Fe) versus the interlayer charge (2 Ca + Na + K). ................................................31

2.7 Plot showing the compositional variation of analyzed thomsonites in NSVG
basalts (black circles). .............................. ..... .............. ...... ...... ...... 33

2.8 Generalized map of northeastern Minnesota showing the distribution of NSVG
lavas (gray) and interpretations of metamorphic grade based on this study and
the work of Schmidt (1993) and Schmidt and Robinson (1997).............................35

2.9 Stratigraphy of the NSVG (Vervoort and Green, 1997; Miller et al., 2002)
correlated along the fold axis dividing the lavas into the southwest and northeast
lim b ............................................................................. 3 6









2.10 Schematic diagram showing the distribution of minerals and mineral zones
typically developed during very low grade metamorphism of large igneous
provinces (after Walker, 1960; Neuhoff et al., 1997, 2000). ..................................39

3.1 Outcrop photograph of vesicular basalt at Site 6 (Table 2.1) illustrating typical
variation in pore size, shape and spacing. ..................................... ............... 50

3.2 Images of reaction zones around and between macroscopic pore space in NSVG
lav as ...............................................................................................52

3.3 Image is of sample NS04-14, showing the laumontite zone alteration found at
S ite 4 ............. ......... .. .. ......... .. .. ............................................... . 5 3

3.4 Various levels of secondary alteration in the matrix of NSVG basalts sampled
during the field season. All photomicrographs were taken through crossed
p o lars. ............................................................ ................ 5 4

3.5 Example of digital analysis of vesicle fillings in a zeolite facies vesicular lava
from E ast G reenland ..... ................................................................ .......... 56

3.6 Frequency of vesicle areas versus the amount of secondary alteration filling the
vesicle for a sample of chabazite-thomsonite zone alteration from East
Greenland (cf. Neuhoff et al., 1997). ............ ..................... ............... 58

3.7 Plot shows the thickness in mm of chlorite rims (black circles) and the total
percent of the vesicle filled with chlorite (open circles) with respect to total
vesicle area in m m 2. .................... .................... ..................... .. ......60

3.8 Plot shows the percentage of mineral infilling with respect to total vesicle area. ...61

3.9 Photomicrograph showing a thin section of a mesolite-scolecite zone lava from
eastern Icelan d ..................................................... ................ 6 2

3.10 Data compiled from digital analysis of mesolite-scolecite zone lava from eastern
Iceland (refer to Figure 3.3). .............................................................................. 64

4.1 Mineralogic composition of lavas as a function of reaction progress. Panels A,
B, and C depict individual m odels. ........................................ ....... ............... 76

4.2 Instantaneous change in total mineral volume as a function of reaction progress.
Panels A, B, and C correspond to models shown in Figure 4.1. ........................... 84

5.1 Schematic representation of a spherical pore of radius rpore surrounded by a
reaction aureole of thickness xauroele. ............. ................................. ...............89

5.2 Graph depicting the percentage of volume occupied by the pore in a given pore-
aureole local reaction region as a function of rpore and xaureole using pore sizes and
aureole thicknesses typical of vesicular mafic lavas ........................ ............90









5.3 Schematic representation of a spherical pore showing secondary infilling typical
of the petrographically observed alteration in the basaltic andesite of the NSVG. .92

5.4 Plot depicting the aureole thickness (x) required to completely fill pore space
within a vesicle of a given pore radius (rpore) with alteration minerals, shown for
total infilling and for clay or laumontite infilling. ................................................93

5.5 Percent chlorite infilling as a function of the axial ratio for an ellipsoidal vesicle
elongated in one direction assuming a constant volume and a constant chlorite
rim thickness. ...................................................... ................. 94















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

ANALYSIS OF POROSITY EVOLUTION DURING LOW TEMPERATURE
METAMORPHISM OF BASALTIC LAVAS AND IMPLICATIONS FOR FLUID
FLOW

By

Jane E. Gustavson

August 2006

Chair: Philip S. Neuhoff
Major Department: Geological Sciences

Basaltic lavas are the most abundant rock type in the crust and globally form

important reservoirs for the migration and storage of crustal fluids. As in any porous

medium, the porosity and permeability of basalt lavas control the movement and extent of

chemical interaction of these fluids and thus the availability and quality of obtainable

resources (e.g., groundwater and petroleum). The highly-reactive nature of basaltic lavas

near the surface leads to intimate coupling between hydrologic properties and extent of

chemical alteration, which controls the aquifer/reservoir quality and chemical fluxes

during alteration. Our ability to quantitatively understand these processes is limited by

rigorous models of how chemical reactions and porosity/permeability modifications are

coupled in these systems.

In the present study, development of secondary mineral parageneses is used as a

monitor of fluid access to pores and to obtain a quantitative description and interpretation









of the evolution of pore space through time. This work has shown that fluid movement

through vesicular basalts occurs along a permeable network of microfractures and

connected vesicles, which are depicted by the extremely weathered aureoles of alteration

surrounding these pathways and the vesicles they connect.

This research also shows that the extent of reaction progress, as evidenced by the

temporal sequence of mineral infillings within pore space, is heavily influenced by the

vesicle size and shape. Larger vesicles exhibit greater degrees of reaction progress while

smaller ones are closed off earlier in the reaction progress. Irregularly shaped vesicles

may close off earlier in the reaction progress due to constriction caused by the shortest

diameter. This dependence of reaction progress on pore size is more specifically based

on the volume percent occupied by the pore within a closed pore-aureole system. Each

vesicle independently exchanges chemical components with the surrounding lava matrix

(the alteration aureole), implying that the reaction progress exhibited within a given

vesicle is based on the proportion of the vesicle to aureole size. For a given vesicle size,

the pore space evolution due to secondary mineralization will be less with a smaller

alteration aureole, due to the low level of fluid rock interaction evidenced by this smaller

alteration aureole. Inversely, a larger alteration aureole implies a higher level of fluid-

rock interaction, which may create enough secondary mineralization to progress to later

stages of alteration or to completely fill the pore space.














CHAPTER 1
INTRODUCTION

Fluid movement is inherently coupled with geochemical reactions during earth

processes. Mass transport of chemical species during fluid flow has received

considerable attention as a means of explaining the chemical evolution of geochemical

systems such as hydrothermal ore deposits, petroleum reservoirs, groundwater aquifers,

and metamorphic systems (e.g., reviews and references therein of Norton and Knight,

1977; Cathles, 1981, 1990; Etheridge et al., 1983; Garven and Freeze, 1984a,b; Norton,

1984; Wood and Hewett, 1984; Garven, 1985, 1989, 1995; Broedehoeft and Norton,

1990; Steefel and Lasaga, 1992; Garven et al., 1993; Garven and Raffensperger, 1994;

Raffensperger and Garven, 1995a,b; Lichtner et al., 1996; Person et al., 1996; and many

others). Dissolution and precipitation of solid phases during reactive transport processes

leads to changes in permeability and porosity of rocks and sediments. Permeability

influences the extent and distribution of chemical alteration during hydrothermal and

metamorphic processes by controlling the rates and magnitudes of supply and removal of

chemical species (e.g., Lasaga, 1986, 1989; Ortoleva et al., 1987; Norton, 1988; Dewers

and Ortoleva, 1990; Steefel and Van Cappellen, 1990; Steefel and Lasaga, 1992, 1994;

Manning et al., 1993; Lasaga and Rye, 1993; Rose, 1995; Bolten et al., 1999). Porosity

controls the amount of aqueous solution that is in contact with the porous media, and can

also have considerable effects on permeability (Xu and Pruess, 2001).

The interrelation between porosity, permeability, and chemical reaction is

especially profound in basaltic lavas and other volcanic rocks. The effect of mineral









paragenesis on porosity in low-grade metabasalts has been noted in many studies

(Schmidt, 1990, 1993; Bevins, Rowbotham, and Robinson, 1991; Manning and Bird,

1991, 1995; Robert and Goffe, 1993; Schmidt and Robinson, 1997; Neuhoffet al., 1999).

Basaltic lavas are the most abundant rock type in the earth's crust (Wedepohl, 1969), and

can form extensive aquifers for the storage and migration of groundwater (e.g., Kononov,

1978; Ingebritson and Scholl, 1993; Rose et al., 1996; Manga, 1997, 1999), geothermal

fluids (e.g., Arnorsson, 1995a, b) and petroleum (e.g., Pendkar and Kumar, 1999;

Christiansen, 1994; Iijima, 2001). Once basaltic lavas are emplaced at the surface of the

earth, the primary magmatic phases formed during cooling (i.e., olivine, anorthite-rich

plagioclase, pyroxenes, oxides and glass) are metastable. The highly-reactive nature of

basaltic lavas near earth's surface leads to extensive dissolution and hydrolysis of the

primary magmatic phases that releases chemical constituents of the lavas into the aqueous

phase. There they are transported by diffusion and advection through the rock (e.g.,

Steefel and Lasaga, 1992) and/or re-precipitated in secondary mineral phases such as

clays, zeolites, and silica minerals. These phases have relatively large molar volumes and

typically occlude pore space as they require additional room to grow beyond the space

generated through dissolution or hydrolysis of the primary basaltic minerals. This causes

portions of the rock to be closed off from the fluid flow pathway, which has a significant

effect on the porosity and fluid movement through the rock.

This thesis explores the relationship between porosity evolution during secondary

mineral paragenesis and fluid flow through vesicular basalts. The following sections of

this chapter provide a review of the nature of porosity and permeability in vesicular lavas,

low-grade alteration of basaltic lavas, and the relationships between chemical reaction,









porosity, and permeability in terms of fluid movement through these lavas. This

information is used to pose several questions about the coupling of chemical alteration

and porosity evolution in lavas that are addressed by the work below.

Porosity and Permeability in Vesicular Lavas

Due to parameters such as temperature and pressure experienced by the lava flow

during eruption, total porosity can vary significantly within a single stratigraphic flow as

can factors such as the size, spacing and geometry of pore spaces (e.g., Larsen et al.,

2004; Burgisser and Gardner, 2004). Vesicular basalt flows can generally be divided into

three sections based on the density of vesicles (Figure 1.1): 1) the bottom, which is high

in porosity with elongated pores due to the stress of overburden during flow; 2) the

center, which tends to be massive and lacks pores; and 3) the top, which has the highest

porosity and more spherical pores due to less overburden pressure during deposition

(Neuhoff et al., 1999).

As a consequence, porosity can vary dramatically from essentially 0% in the

massive flow centers up to 85% in scoraceous zones (e.g., Cashman and Mangan, 1994).

The macroscopic porosity depicted in Figure 1.1 tends to dominate total porosity, as

porosities determined by point counting or other digital analyses tend to agree well with

those obtained by laboratory techniques (e.g., Saar and Manga, 1999, Al-Harthi et al.,

1999).









lava flow top


vesicular flow top (+/- scoria, breccia)


internal vesicular zone

isolated vesicles massive flow center
vesicular flow base and breccia


lava flow bottom

Figure 1.1: Cross sectional view showing distribution of primary porosity vesicless,
scoria, breccia) typical of thick aa lava flows. Note heterogeneous but regular
distribution of porosity through the flow.

Figure 1.2 illustrates three different possible pore geometries within the vesicular

zones of basalt. In Figure 1.2A, the highly vesicular lava contains coalesced pore spaces,

leading to high permeability that is controlled by the apertures of the pore intersections

(Saar and Manga, 1999); this situation is typical of scoraceous zones within flow tops.

At lower porosities, where vesicle density is too low for widespread coalescence,

variations in vesicle geometry can have a profound effect on the nature of permeability.


S0 0 0 -




S
**
A B C. ***


Figure 1.2: Schematic diagrams depicting the effects of pore size, shape and connectivity
on the permeability of vesicular basalts. The lava depicted in (A) has highest
porosity, with coalesced vesicles that lead to connected fluid flow paths. The
lavas shown in (B) and (C) have the same porosity, but different pore size and
mean pore-pore distances that will affect the ease and direction of fluid flow
through the rock.









For instance, the lavas depicted in parts B and C have the same porosity, but different

pore sizes. The distribution of pore space into a greater number of smaller pores in part C

relative to that in part B leads to significantly reduced pore-pore distances, which is likely

to impact the ease of fluid movement between pores. These variations in porosity and

pore geometry depicted in Figure 1.2 can exist even within a single vesicular zone of the

basalt (Gardner et al., 1996).

Vesicular volcanic rocks exhibit a wide range of permeabilities, ranging from about

10-7 to 10-16 cm2 (e.g., Klug and Cashman, 1996; Saar and Manga, 1999; Mueller et al.,

2005). Zones within basaltic lavas exhibiting macroscopic porosity generally have

relatively high permeabilities, often in the range of 10-7 to 10-11 cm2. This is comparable

to clean unconsolidated sand or karstic limestone. Matrix permeabilities in non-porous

lavas can be considerably lower, often as low as 10-16 cm2 (e.g., Saar and Manga, 1999;

Sruoga et al., 2004; Mueller et al., 2004).

Recent research demonstrates that vesicular basalts do not follow porosity-

permeability relations typically applied to granular rocks (Saar and Manga, 1999).

Previous studies of the relationship between porosity and permeability in granular rocks

have applied percolation theory (e.g., Carman, 1956; Dullien, 1992; Bosl et al., 1998)

which predicts a percolation threshold, or the minimum porosity at which a connected

pathway for fluid movement through the sample exists, of 30%. Following this theory,

when the porosity of a sample drops below this critical value there will be a dramatic

decrease in permeability. This approach leads to widely applied power law dependencies

of permeability on porosity such as the Kozeny-Carman equation (e.g., Carman, 1956;

Dullien, 1992; Le Gallo et al., 1998; Balashov and Yardley, 1998; Park and Ortoleva,









2003; Freedman et al., 2003). Experimental studies of lava degassing, transport

modeling, and textural studies of pumices and other high-vesicular materials predict that

interconnected pathways resulting from bubble coalescence of vesicles generally occur at

porosities over 30 % (e.g., Klug and Cashman, 1996; Blower, 2001; Mueller et al., 2004;

Burgisser and Gardner, 2005). However, it is clear that vesicular volcanic rocks exhibit

relatively high, but variable, permeabilities (-10-8 cm2) down to porosities as low as a

few percent, in contrast to the behavior of granular rocks and the predictions of

percolation theory (e.g., Saar and Manga, 1999; Mueller et al., 2004). Numerous models

have been proposed to explain this behavior, largely relying on textural explanations

(e.g., Saar and Manga, 1999; Blower, 2001). Recent work has suggested that the elevated

permeabilities in volcanic rocks are associated with microfractures formed between

vesicles during degassing, leading to connected fluid pathways that are not quantitatively

important components of total porosity (Mueller et al., 2004).

Low Grade Alteration of Basaltic Lavas

When basalt is emplaced at the earth's surface it is metastable. The olivine,

pyroxene, anorthite-rich plagioclase, oxide and glass phases in basalts form under higher

temperature conditions and are easily weathered at surface conditions. Chemical

alteration of basalts begins to occur almost immediately after emplacement (e.g., Hearn et

al., 1981; Jakobsson and Moore, 1986), and the specific reactions that take place between

basaltic lavas and groundwaters are strongly dependent on temperature and pressure

conditions as well as the composition of the lavas and fluids (e.g., Kristmannsd6ttir and

T6masson, 1978; Neuhoff et al., 1999, 2000). In general, the sequence of reactions

experienced by lavas of this nature can be summarized as 1) reaction of lavas with

oxygenated aqueous solutions forming iron oxyhydroxides, celadonite, amorphous silica,









amorphous aluminosilicates, and clay minerals during weathering; 2) continued reaction

of the lavas to form clay minerals (smectite, chlorite) and later zeolites and other calcium

aluminosilicates during burial metamorphism; and 3) localized, late stage hydrothermal

alteration in fracture systems around intrusions (Neuhoff et al., 1999). These various

stages of mineral formation can be distinguished through careful use of geologic

relationships and relative dating techniques (Neuhoff et al., 1999).

The sequence of chemical reactions that alter basaltic lavas after emplacement

typically leads to the development of depth-controlled zones, which reflect increasing

temperature with depth. Clay minerals trend from dioctahedral smectite to trioctahedral

smectite to mixed layer chlorite-smectite to chlorite with increasing depth/temperature

(e.g., Schiffman and Friedliefsson, 1991; Neuhoff et al., 1999, 2006); these minerals are

observed as replacements of glass, olivine, pyroxenes, fine-grained groundmass,

sometimes plagioclase, and are also found to rim vesicle walls. Zeolites occur as

replacements of plagioclase and occasionally groundmass, and more commonly as

vesicle infillings after formation of mafic phyllosilicates.

Prominent depth-controlled distribution of individual zeolite species within vesicles

is often noted, allowing delineation of mineral zones defined by the occurrence of one or

more index zeolite species. These zones are well described in many tholeiitic lava flows

throughout the world, and as many as five separate zones exist in regionally

metamorphosed basaltic lavas and Icelandic geothermal systems (Walker, 1951, 1960a,

b; Sukheswala etal., 1974; Kristmannsdottir and Tomasson, 1978; Jorgensen, 1984;

Murata etal., 1987; Schmidt, 1990, 1993; Neuhoff et al., 1997, 1999; Christiansen etal.,

1999). They are relatively uniform in thickness over large distances and cross-cut









individual lava flows. The thickness, orientation, and geologic relations between these

mineral zones and lava stratigraphy provide critical information for assessing pre-

erosional lava thicknesses as well as crustal deformation (e.g., Neuhoff et al., 1997,

2000).

Coupling Between Porosity Evolution and Chemical Reactions

As in any porous medium, the hydrologic properties of basalt lavas control the

movement and extent of fluid-rock interaction, which affects the formation of secondary

minerals and thus the availability and quality of obtainable resources (e.g., groundwater

and petroleum). In principle, porosity change during chemical alteration can be

determined from the change in mineral volume during reaction due to mineral dissolution

and precipitation. The extent of porosity destruction is a function of both the relative

volumes of primary and secondary phases, and of their relative solubilities in the aqueous

phase (Putnis, 2002). Provided a suitable determinative relationship between porosity

and permeability is known (such as the Kozeny-Carman equation; Carman, 1956),

changes in mineral volume during chemical reaction from petrographic or modeling

results can be used to estimate changes in permeability (e.g., Steefel and Lasaga, 1994;

Panda and Lake, 1995; Balashov and Yardley, 1998; Le Gallo et al., 1998; Saripalli et

al., 2001, 2005; Park and Ortoleva, 2003; Xu and Pruess, 2001; Freedman et al., 2003).

In low-grade metabasalts, the destruction of porosity is evident through the

progressive infilling of vesicles by phases such as mafic phyllosilicates and zeolites. This

is a direct consequence of the open crystal structures of mafic phyllosilicates and zeolites,

which leads to a net increase in mineral volume (Neuhoff et al., 1999, 2000). These pore

fillings will alter the size, geometry and connectivity of the pore space and can

effectively close off portions of the rock to further alteration (Neuhoff et al., 1999).









Because, as noted above, permeability of vesicular lavas is likely strongly dependent on

the geometry of the pore space, these processes may lead to variations in permeability

with reaction progress that are not readily predicted by simple power-law porosity-

permeability relationships. What is clear is that progressive alteration of lavas leads to

distinct reductions in permeability, for instance in oceanic lavas (Fisher, 1998), petroleum

reservoirs (Sruoga et al., 2004; Luo et al., 2005), and geothermal systems (Stimac et al.,

2004).

Role of the Present study

The ability to quantify and further understand the time dependent changes in the

properties of porous media is sought after by geologists, hydrologists and oil companies

among many others (Aharonov et al., 1997). The evolution of porosity and permeability

during reactive transport has received surprisingly little attention (e.g., Steefel and

Lasaga, 1994; Le Gallo, 1998; Saripalli et al., 2001; Freedman et al., 2003), and our

ability to quantitatively understand these processes is limited due to a lack of rigorous

models of how chemical reactions and porosity/permeability modifications are coupled in

these systems. This is particularly true for vesicular lavas for reasons noted above. The

present study uses secondary mineral parageneses in vesicular basalts to address three

aspects of this problem:

1. What causes the high permeabilities of vesicular volcanic rocks, even at low
porosities?

2. Does the extent of reaction progress (as evidenced by the temporal sequence of
mineral infillings of pore space) vary with aspects of pore geometry such as
spacing, size, and shape?

3. What is the relationship between pore geometry and porosity evolution during
mineral paragenesis?














CHAPTER 2
LOW-GRADE ALTERATION OF THE NORTH SHORE VOLCANIC GROUP,
MINNESOTA

Introduction

The distribution of individual zeolites, filling vesicles in low-grade metabasalts of

large igneous provinces, is typically characterized by several depth-controlled zones

defined by the occurrence of one or more index zeolite species (e.g., Walker, 1960a,b;

Sukheswala etal., 1974; Murata etal., 1987; Schmidt, 1993; Neuhoffet al., 1997, 1999).

The thicknesses, orientations, and geologic relations between these mineral zones and

lava stratigraphy provide critical information for assessing geothermal gradients, crustal

deformation, and pre-erosional lava thicknesses (e.g., Neuhoff et al., 1997, 2000). This

style of alteration has typically been observed in Mesozoic or younger, relatively un-

deformed lavas due to the fact that low grade terrains are often lost to erosion or

overprinted by subsequent metamorphism in older provinces. One exception is the 1.1

Ga North Shore Volcanic Group (NSVG), which is regionally deformed, yet still contains

well-developed zeolite facies metamorphic zonation (Schmidt, 1993; Schmidt and

Robinson, 1997; Figure 2.1).

The overall structure of the NSVG is a plunging synclinal fold with a hinge axis in

the Tofte-Lutsen area (e.g., Miller et al., 2002) dividing the region into two sections; the

southwest limb and the northeast limb. Previous studies (e.g. Schmidt, 1993, Schmidt

and Robinson, 1997) established a progression of metamorphic mineral assemblages in

the southwest limb that range from lower zeolite facies in the uppermost parts of the






11

















Geo ogy
N e Beaver Bay Complex
from | Duluth Complex
a11 o rLate Archean Rocks
1e Miscellaneous Intrusions
Th p n NSVG
.- c-t 1 Paleoproterozoic Rocks
andi"t ,:" [_ aRift Sediments
--I














10 0 10 20 Miles



Figure 2.1: Generalized geological map of northeastern Minnesota showing the

distribution of the NSVG lavas and associated Keweenawan intrusives (after
Miller et al., 200). Numbered locations (red circles) refer to sampling sites
from the present study (cf. Table 2.1).

stratigraphy through greenschist faces rocks in the lowermost lavas near Duluth. There

appears to be little previous work on the metamorphism of the northeast limb as assessing

the metamorphic history of this part of the province is complicated by the variable lava

chemistry, which covers the whole range betassociated Keween rhyolites and olivine normative basalts

(BVSP, 1981; Brannon, 2001984; Schmidt, 1990; Miller etd circles)., 2002; Boerboom, 2004).

The present study investigates the distribution of metamorphic grade in the NSVG,


with a concentration on the alteration present in the northeast limb. Field, petrographic,

and analytical studies are combined to assess the variation in metamorphic grade across









the syncline. This data is then used to interpret the conditions and timing of alteration

relative to the structural development of this province.

Geologic Background

The most voluminous accumulations of basaltic lava on continental crust are flood

basalts, which typically outcrop as thick sections of subhorizontal, subaerial tholeiitic

flows (Coffin and Eldholm, 1994; Winter, 2001), such as the basalts of the Keweenawan

province associated with the Midcontinent Rift. The Midcontinent rift system (MRS) is a

major feature covering over 2,000 km of central North America (Van Schmus and Hinze,

1985; Cannon et al., 1989, Cannon, 1992), extending northeast from Kansas into central

Lake Superior and then southeastward through Michigan into Ohio (Green, 1982; Hinze

et al., 1997; Ojakangas and Dickas, 2002). The MRS is a product of extension occurring

between 1108 and 1094 Ma, which was aborted prior to complete continental rifting due

to convergence from the Grenville province allochthon to the east (Cannon, 1994; Allen

et al., 1997). Magmatism associated with this extension is believed to originate from a

large asthenospheric plume centered below what is now Lake Superior (Vervoort and

Green, 1997). The volcanic pile emplaced during magmatism is more than 8 km thick

(Ojakangas and Matsch, 1982; Schmidt and Robinson, 1997) with an estimated volume

of 1.3 x 106 km3 (Hutchinson et al., 1990; Allen et al., 1992). This province is one of the

oldest and best preserved large igneous provinces (BVSP, 1981; Boerboom, 2004) with

the majority composed of either intrusions or subaerial volcanics later buried by

Phanerozoic sedimentary strata (Ojakangas and Dickas, 2002). Surface exposures of the

MRS are known only in the Lake Superior region, where a 30 km thick sequence of

volcanic and sedimentary rocks composes the Keweenawan Supergroup (e.g., Ojakangas

and Dickas, 2002).









The NSVG, a section of the Keweenawan province located in northern Minnesota,

covers the extent of the Minnesota shoreline along Lake Superior and can be divided into

two sub-basins of gently dipping strata (Vervoort and Green, 1997). These lavas consist

mainly of olivine tholeiites with a continuous range of compositions from strongly

olivine-phyric tholeiites to rhyolites (BVSP, 1981; Schmidt, 1990; Boerboom, 2004).

Magmatism throughout most continental flood basalts is dominantly basaltic with only

about 1% of the flows having a felsic composition; within the NSVG, 10% of the

southwest limb and 25% of the northeast limb is composed of felsic flows (Vervoort and

Green, 1997). It still is unclear if the range in composition is due to large-scale crustal

melts or differentiates of more primitive magma (Vervoort and Green, 1997). The

structure of the MRS is dominated by extension followed by compression, forming horst

and graben structures followed by reverse and, in some cases, transform faulting (Van

Schmus and Hinze, 1985; Vervoort and Green, 1997). Geophysical and geological data

show that the Lake Superior region is a fault-bounded-asymmetric basin, but the exact

structure of the area encompassing the NSVG is still unknown (Van Schmus and Hinze,

1985; Hinze et al., 1997).

The flows of the NSVG are mainly between 1 and 50 m thick (Schmidt and

Robinson, 1997), with an average thickness between 5 and 25 m (Schmidt, 1990). In

areas where lava flows vary in composition, the morphology of individual flows is

characteristic for each major compositional type (Saemundsson, 1979). In the NSVG, the

volumetrically dominant olivine tholeiites occur as pahoehoe flows, while other

compositions, such as the basaltic andesites occur as aa flows (Schmidt, 1990;

Boerboom, 2004). Figure 1.1 shows a cross sectional diagram depicting the general









morphology of a typical aa flow. The upper vesicular zone can take up as much as half

of the flow thickness in thinner flows and the flow bottoms are rich in pipe amygdules

that often coalesce into inverted Y-shapes (Schmidt, 1990). Pahoehoe flows typically are

vesicular throughout the entire thickness. In areas where multiple pahoehoe flows occur

in vertical succession, flow boundaries are easily depicted due to the presence of pipe

amygdules at the base and classic ropey pahoehoe flow-top characteristics at the top

(Figure 2.2).


Figure 2.2: Flow boundary between two vesicular basalt flows at Site 12. Field book
(12x19 cm) shown for scale. White curve superimposed on image highlights
the flow boundary. Lower arrow points to pahoehoe structures indicative of
the flow top of the lower flow, and the upper arrow points to the filled pipe
amygdules found at the base of the upper aa flow.









Methods

Field work was conducted along the North shore of Lake Superior in Minnesota

during July 15-26, 2004. Field observations were made at 12 sites (Table 2.1, Figure

2.1). Several sites (3, 4, 6, 11, 12) were selected for detailed mapping at the outcrop

scale using standard geologic methods. Photo-mosaics were acquired with a high

resolution digital camera for later analysis. Samples for geochemical and digital analysis

were collected from outcrops that were representative of the variation in porosity,

alteration, mineralogy, and pore types encountered.

A suite of samples representative of the lithologies and mineral parageneses found

in the NSVG were selected and used for in depth chemical and digital analysis of the

various flow chemistries and alteration histories. In preparation for bulk rock analysis,

twelve samples were powdered and sent to the Geosciences Laboratories of Ontario

where major element contents were determined by wavelength dispersive X-ray

fluorescence (XRF) with a detection limit of 0.01 wt%. Thin sections also were prepared

from each sample for detailed petrographic observations of primary and secondary

minerals. Petrographic analysis was performed on each thin section to quantify primary

porosity and determine mineral parageneses. Mineral chemistry was determined by

electron probe microanalysis (EPMA) using an automated JEOL 733A electron

microprobe operated at 15 kV accelerating potential and 15 nA beam current. Calibration

was conducted using natural geologic standards. Beam width for analysis of hydrous

minerals (i.e., clays and zeolites) was 10 to 30 [im to minimize alkali migration (this

instrument is outfitted with wide detector slits that allow for large spot size analyses).

Raw counts were collected for 20 s (approximately 60 s total beam contact at each point)

and converted to oxide weight percent using the CITZAF correction procedure after












Table 2.1: Field study locations in the North Shore Volcanic Group.
Site Site Description Samples Alteration Mineralogy Zeolite Zone
1 Knife River Wayside 01 Prehnite-pumpellyite facies alteration, Prehnite-pumpellyite
N460 56.849', W910 47.505' laumontite float
2 Cut Face Creek outwash at Wayside 02 Laumontite fault breccia and stilbite- Laumontite
N470 43.854', W900 26.383' bearing amygdaloidal lava float
3 Private shoreline southeast of 04-10 Thomsonite + mesolite alteration Thomsonite-Mesolite
Thomsonite Beach
N470 43.358', W900 27.076'
4 Cut Face Creek 11-12, Laumontite + chlorite alteration; (also Laumontite
N470 43.887', W900 26.517' 14, 30- laumontite cemented breccia 2 miles
31 upstream)
5 Streambed to Butterwort Cliffs 16 Laumontite-cemented fault breccia Thomsonite-Mesolite
N470 43.266', W900 28.089' float
6 Butterwort Cliffs (Cascade River 13B, 15, Thomsonite + mesolite + very minor Thomsonite-Mesolite
State Park) 17-19 analcime +/- calcite alteration
N470 43.212', W900 28.033'
7 Shoreline southeast of Judge 03 Silica + clay + laumontite alteration Laumontite
Magney State Park
N470 48.716', W900 03.999'
8 Tofte Park 20 Hydrothermally altered lava with Thomsonite-Mesolite
N470 34.258', W900 50.389' heulandite + scolecite + stilbite +
laumontite + thomsonite
9 Lutsen Grandview Park 21 Thomsonite + mesolite + analcime Analcime
N470 39.568', W900 38.349'
10 Roadcut along Hwy 61 22 Andesite (?) with laumontite Thomsonite-Mesolite
N470 35.536', W900 48.071'
11 Gooseberry Falls State Park 23-29 Chlorite + calcite + stilbite + Laumontite
N470 08.286', W910 27.520' Laumontite alteration
12 Temperance River State Park 32-38 Stilbite + heulandite + quartz + chlorite Stilbite-heulandite
N470 33.076', W900 52.560' + thomsonite alteration









accounting for unanalyzed oxygen following the methods of Tingle et al. (1996). The

analytical conditions and correction procedures employed have previously been shown to

provide analyses of zeolites and clays that are within error of compositions determined by

other methods (Tingle et al., 1996).

Results

Field observations indicate that low-grade secondary mineral assemblages are well-

preserved through most of the NSVG. Alteration mineralogy is easily observed in the

field as vesicle infillings in the high porosity zones. Vesicles typically show complete

infilling with various zeolite minerals, and, in some cases, are initially lined with clay or

silica rims. The lava matrix, in both the high and low porosity zones, appears to have

also undergone alteration. This matrix alteration shows some zeolitization, but is

dominated by clay and iron oxide alteration minerals. The secondary minerals present in

these lavas are similar to those previously observed in tholeiitic lava sequences (e.g.,

Walker, 1960; Neuhoff, 1997, 2000), ranging from lower zeolite facies through

greenschist facies alteration.

Primary Basalt Composition and Mineralogy

Sampled lavas show compositional variations consistent with previous studies of

this province (e.g., Schmidt, 1993; Schmidt and Robinson, 1997). Whole rock

compositions for selected samples are shown in Table 2.2 and the total alkali-silica

diagram of Figure 2.3. Most of the lavas were olivine-normative basalts, with a few from

the northeast limb exhibiting more evolved basaltic andesite and dacitic compositions.













Table 2.2: Whole-rock chemical compositions (wt %) of samples.
Sample NS04-03 NS04-05 NS04-In NS04-14B NS04-15 NS04-19 NS04-21 NS04-25 NS-4-2" NS04-31 NS04-37 NS04-38
Sitea 8 3 3 4 7 7 10 12 11 4 13 13


SiO2
TiO2
A1203
Fe203b
MnO
MgO
CaO
Na20
K20
P205
LOI
Total


66.87
1.28
9.42
12.6
0.1
1.5
2.06
3.59
0.49
0.39
2.48
100.78


44.96
1.14
16.8
9.98
0.13
7.48
7.9
3.66
0.14
0.12
8.16
100.47


45.05
1.08
15.5
9.68
0.15
9.26
7.07
3.3
0.07
0.11
9.03
100.3


55.51
2.17
12.15
13.45
0.24
3.58
3.48
3.83
1.64
0.44
3.94
100.43


45.4
1.2
15.67
10.02
0.16
8.5
8.42
2.58
0.73
0.12
7.24
100.04


46.82
1.24
15.88
10.56
0.17
9.41
6.17
4.03
0.22
0.12
5.92
100.54


45.28
1.08
16.35
9.08
0.14
8.4
7.92
2.84
0.8
0.12
8.11
100.12


43.29
1.31
15.31
10.97
0.16
7.44
9.58
2.23
0.27
0.14
8.69
99.39


48.66
0.74
16.04
6.86
0.17
6.87
6.82
2.54
0.25
0.05
10.55
99.55


46.62
1.15
16.86
10.26
0.16
7.98
9.68
2.27
0.28
0.12
4.56
99.94


46.59
1.12
16
9.8
0.16
7.89
9.05
2.12
0.45
0.13
6.98
100.29


44.92
1.07
15.73
9.53
0.14
8.31
8.83
1.92
0.41
0.13
8.86
99.85


"Site descriptions given in Table 2.1, and locations shown on Figure 2.1. bTotal Fe reported as Fe203.












16


14 -
Phonolite

12 -
12 Tephriphonolite
Foidite
Trachydacite
0
SPhonotephrtee

+ /
+ trachyandesite

Z
O 6 Tephrite


4 Basaltic Andesite
Basalt andesite

2
Picrobasall


35 40 45 50 55 60 65 70 75
SiO2 (wt%)

Figure 2.3: Total alkalis-silica diagram showing compositional ranges of NSVG basalts
sampled in this study (black circles). Some of the collected samples are more
evolved basaltic andesites and dacites, while the majority of samples are
olivine-phyric basalts and picrobasalts.

Matrix alteration of the primary magmatic phases is observed to varying degrees in

all of the analyzed samples. There was no glass, mesostasis or olivine observed in any of

the samples, most likely due to alteration. Feldspar and pyroxene compositions within

the lavas varied throughout each sample, and these variations showed no significant

correlation to compositional variations of the lavas. The feldspar compositions (Table

2.3) range from Ab23An720r5 to Ab99.5An0.30r0.2. The near-end member albite

compositions observed in some plagioclase phenocrysts suggests that they have been

subject to secondary albitization. Clinopyroxenes are present, with representative

compositions shown in Table 2.4; the absence of orthopyroxenes could also be attributed

to the alteration of each sample.












Table 2.3: Representative compositions of plagioclase.
Sample NS04-03 NS04-29B NS04-05 NS04-19 NS04-10A NS04-25 NS04-38 NS04-31A
Oxide weight %
Si02 68.46 69.28 61.79 49.73 45.73 53.84 53.20 48.62
TiO2 0.04 0.03 0.05 0.04 0.02 0.11 0.10 0.10
A1203 19.55 19.94 19.01 17.91 25.15 27.74 29.64 31.79
FeO" 0.05 0.09 0.76 7.01 0.33 0.87 0.62 0.75
MnO 0.00 0.01 0.02 0.16 0.00 0.02 0.00 0.03
Cr203 0.04 0.02 0.04 0.01 0.03 0.05 0.05 0.03
MgO 0.00 0.13 2.36 11.53 0.27 0.36 0.24 0.18
CaO 0.06 0.58 1.26 1.56 8.14 10.93 13.02 14.66
Na20 12.28 11.99 9.37 4.13 5.89 5.04 4.13 3.12
K20 0.03 0.11 0.47 1.43 0.04 0.32 0.12 0.12
Total 100.50 102.18 95.14 93.52 85.59 99.26 101.12 99.41
Anhydrous formula units
Si 2.98 2.97 2.87 2.41 2.42 2.46 2.39 2.24
Ti 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Al 1.00 1.01 1.04 1.05 1.57 1.49 1.57 1.73
Fe 0.00 0.00 0.03 0.33 0.01 0.03 0.02 0.03
Mn 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00
Cr 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Mg 0.00 0.01 0.17 0.97 0.02 0.02 0.02 0.01
Ca 0.00 0.03 0.06 0.09 0.46 0.54 0.63 0.73
Na 1.04 1.00 0.84 0.42 0.60 0.45 0.36 0.28
K 0.00 0.01 0.03 0.01 0.00 0.02 0.01 0.01
Mole %
Ab 99.57 96.80 89.53 80.70 56.50 44.67 36.20 27.60
An 0.27 2.60 6.80 17.25 43.23 53.47 63.10 71.65
Or 0.17 0.60 3.68 2.05 0.27 1.87 0.70 0.75
aTotal Fe reported as FeO. bBased on 8 oxygen charge equivalents.












Table 2.4: Representative compositions of pyroxenes.
Sample NS04-10A NSi-4-i1-1. NS04-14B NS04-15C NS04-15C NS04-37 NS04-37
Oxide weight %
Si02 49.16 49.76 50.27 51.53 52.66 51.13 50.03
TiO2 1.54 1.93 0.77 1.32 0.67 1.66 1.63
A1203 4.70 3.73 1.31 1.90 2.50 4.45 3.08
FeO" 10.11 10.07 21.10 13.47 6.83 10.28 12.68
MnO 0.22 0.22 0.50 0.33 0.17 0.20 0.20
Cr203 0.28 0.14 0.05 0.02 0.60 0.26 0.22
MgO 13.40 14.16 10.38 15.67 16.42 12.96 12.17
CaO 18.85 19.76 15.82 16.09 20.83 20.79 20.18
Na20 0.52 0.35 0.22 0.33 0.30 0.32 0.32
K20 0.02 0.01 0.01 0.03 0.05 0.00 0.01
Total 98.79 100.14 100.43 100.69 101.04 102.05 100.53

Anhydrous formula units
Si 3.72 3.72 3.90 3.84 3.84 3.75 3.77
Ti 0.09 0.11 0.04 0.07 0.04 0.09 0.09
Al 0.42 0.33 0.12 0.17 0.22 0.38 0.27
Fe 0.64 0.63 1.37 0.84 0.42 0.63 0.80
Mn 0.01 0.01 0.03 0.02 0.01 0.01 0.01
Cr 0.02 0.01 0.00 0.00 0.03 0.01 0.01
Mg 1.51 1.58 1.20 1.74 1.79 1.42 1.37
Ca 1.53 1.58 1.31 1.29 1.63 1.63 1.63
Na 0.08 0.05 0.03 0.05 0.04 0.05 0.05
K 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Mole %
Wo 41.60 41.77 33.87 33.23 42.53 44.40 42.93
En 41.03 41.63 30.90 45.07 46.60 38.50 36.03
Fs 17.37 16.63 35.23 21.73 10.87 17.13 21.03


"Based on 12 oxygen charge equivalents.


"Total Fe reported as FeO.









Regional Alteration Mineralogy

Although flow morphologies and textures are well-preserved throughout the study

area, all investigated lava flows exhibit some degree of secondary mineral alteration. In

contrast to typical alteration patterns in younger very low-grade metabasalts, alteration is

present even in areas away from zones of high primary porosity and is clearly observed in

hand specimen. Within the matrix, the primary magmatic phases are usually replaced by

secondary alteration minerals such as Fe(Ti) oxides, mafic phyllosilicates, zeolites, and

occasionally silica minerals and calcite. The alteration mineralogy observed within

vesicles is dominated by zeolite minerals, with occasional linings of silica and mafic

phyllosilicates and overprinting by calcite.

The extent of chemical alteration within the matrix is typically greater within zones

of high primary porosity (e.g., vesicular flow tops and bottoms) than in massive flow

centers. In some samples there is a strong presence of sub-mm sized spots of bright red

Fe(III) oxides (probably hematite), and in most cases oxide alteration occurs further away

from amygdules than the mafic phyllosilicates and zeolites. Mafic phyllosilicates often

are pervasive around vesicles as replacements of mesostasis, glass, and olivine, and also

are commonly found lining pore walls. Plagioclase is variably replaced in vesicular

zones by either near-end member albite or zeolites, and zeolite alteration within the

matrix dominantly occurs in close proximity to amygdules also filled with zeolites. Often

the only phenocryst phase that escapes alteration in these zones is clinopyroxene. The

overall extent of this alteration appears to be a function of distance from primary

porosity, and in some samples (Figure 2.4) light-colored reaction aureoles are observed as

bleached haloes (probably due to oxidation reactions) around and between primary pores

within the rock.




























Figure 2.4: Field photo from Site 3 (Table 2.1) showing an extensively altered pahoehoe
flow. Visible within the flow are bleached haloes around vesicles connected
by thin anastomosing bleached areas. Pencil shown for scale.

Primary vesicles, especially in zones of relatively high porosity at the tops and

bottoms of flows, are generally completely filled with zeolites and occasional linings of

silica and mafic phyllosilicate minerals. Some vesicles undergo complete infilling with

only one zeolite mineral, as shown by the thomsonite filled vesicle in Figure 2.5A. Many

other vesicles show multiple stages of zeolite alteration and contain two or more zeolite

minerals. Figure 2.5B shows a vesicle filled with thomsonite and mesolite, and the

vesicle in Figure 2.5C contains analcime, thomsonite and mesolite. Some vesicle walls

are rimmed by secondary minerals prior to zeolite infilling. Silica minerals are the first to

form where present, as shown in Figure 2.5D. This amygdule exhibits early lining of

silica, followed by a thin clay layer before complete infilling of laumontite. Most often

the vesicle walls are lined only with mafic phyllosilicates prior to zeolite infilling, as

shown in Figure 2.5E where chlorite rims precede laumontite infilling.









































Figure 2.5: Secondary alteration in NSVG lavas. All photomicrographs were taken
through partially crossed polars. A) Photomicrograph showing thomsonite
alteration in sample NS4-31, collected directly above the sediment pile





roadcut from Site 4 (Table 2.1). B) Photomicrograph of sample NS04-15
showing the thomsonite-mesolite alteration typical of Sites 3 and 6. C)
Scanned image of thin section of sample NS04-15 showing analcime-
.. .







thomsFigure 2.5 Seconite dairy alteration in NSVG lavas. All photomicrographs were taken4-
through partially crossed polars. A) Photomicrograph showing thomsonite



03 showing laumontite alteration typical of sample si04-3, collected directly above the sediment pilearly-
roadcutdefined early silica and clay (rims on vehicles. E) Photomicrograph of sample 04-
4-14 showing the thomsonite-mesolite alteration with initial of Sites 3 and 6. C) rims.
Scanned image of thin section of sample NS04-15 showing analcime-
thomsonite (thom.)-mesolite alteration. D) Photomicrograph of sample NS04-
03 showing laumontite alteration typical of sample sites 4 and 7 with clearly-
defined early silica and clay (rims on vesicles. E) Photomicrograph of sample
NS04-14 showing laumontite alteration with initial formation of chlorite rims.
F) Photomicrograph of sample NS04-10A highlighting calcite crystallization
in thomsonite filled amygdule.
























Figure 2.5: (continued)

Quartz and calcite are widespread throughout the study area, though not necessarily

pervasive. Local, late-stage alteration that cross-cuts and, in some cases overprints these

assemblages, is observed around veins and faults. Veins in these localities often contain

quartz and calcite, along with laumontite, stilbite, and/or heulandite. Silica occasionally

is present as initial pore linings, but most often is present within the matrix, while calcite

dominantly is found as post-zeolite infillings of amygdules. Figure 2.5D shows an

amygdule exhibiting early linings along the pore wall of quartz and interlayered

chlorite/smectite, followed by complete zeolite infilling of residual pore space. The

progression of mineralization within this depicted pore is typical of pores exhibiting silica

linings. Figure 2.5F shows an amygdule filled with thomsonite that has experienced late

calcite overprinting. The amygdules within a given sample that have experienced this

late calcite overprinting appear to be randomly distributed.

Four distinct zeolite assemblages are observed within the lavas of the northeast

limb of the study area. Mineral zones containing these assemblages are identified based

on the presence of the index minerals analcime, thomsonite-mesolite, stilbite-heulandite,

and laumontite within vesicles. In all cases, these zones are present in contiguous









sections of the NSVG, and, except for the stilbite-heulandite zone, are present in both

limbs. Most of the isograds bounding these zones are not well-exposed, and their

locations are approximate. Alteration higher in metamorphic grade than zeolite facies is

present only in the very southern portion of the field area where alteration reached

prehnite-pumpellyite and greenschist facies, and has already been discussed in detail by

Schmidt (1993) and Schmidt and Robinson (1997).

The mineralogy of the four distinct zeolite zones is described below:

Laumontite zone lavas

Laumontite zone alteration is the highest zeolite grade alteration observed in the

NSVG lavas, and is characterized by the presence of quartz, interlayered chlorite/smectite

(c/s), and laumontite. The alteration zone was observed in two locations along the

shoreline, from just south of Grand Marais to Grand Portage and from Little Marais to

just south of Two Harbors. Albitized plagioclase laths and altered mesostasis dominate

the matrix, and there are very minor amounts of unaltered pyroxene phenocrysts.

Alteration affects most of the matrix and consists of c/s, quartz and iron oxides.

Quartz rims within amygdules are occasionally present where they precede rims of

C/S (Figure 2.5D). Occasionally, C/S is visible in hand sample, and appears in thin

section as a green fibrous mineral within vesicles. Smaller vesicles within these lavas are

completely filled with c/s, while larger vesicles are rimmed with C/S and later filled with

laumontite. Laumontite appears white in hand sample, and in thin section appears either

blocky or fibrous depending on the orientation of the thin section to the fibrous mineral

growth habit. Overprinting of laumontite with calcite is occasionally observed in the

vesicles. The calcite exhibits twinning and, in most cases, rhombohedral crystal habit.









Stilbite-heulandite zone lavas

Stilbite-heulandite zone alteration is recognized in the NSVG lavas by the presence

of amygdules filled with stilbite and/or heulandite. Unlike the other zeolite zones present

along the NSVG, this zone was not found in the northeast limb. During this study it was

only observed once along the southwest limb between the towns Little Marais and Tofte

(and was observed by Schmidt (1993) in several localities in this area). Iron oxides,

chlorite, albitized plagioclase laths and minor amounts of unaltered pyroxene phenocrysts

dominate the lava matrix.

Trioctahedral smectite is present either as rims along vesicle walls or as complete

infilling of vesicles. Smaller vesicles are completely filled with trioctahedral smectite,

while in larger vesicles it is present as rims with the remaining open space filled with

calcite. The calcite exhibits twinning and in some instances rhombohedral cleavage. It is

possible that calcite occurs as overprinting of the original zeolites that filled the vesicles

in these analyzed samples, as stilbite and heulandite were observed as vesicle infillings in

the field.

Thomsonite-mesolite zone lavas

Lavas exhibiting amygdules filled with thomsonite, mesolite and C/S interlayered

clays comprise this alteration zone observed in the NSVG. This zone of alteration

occurred to the northeast and to the southwest of Lutsen, bounding the analcime zone

discussed below. Pyroxene phenocrysts, plagioclase laths, and iron oxides dominate the

lava matrix surrounding the vesicles. The pyroxene phenocrysts often are unaltered,

whereas plagioclase is typically replaced by zeolites similar to those observed in vesicles.

Amygdule minerals in hand samples are occasionally green in color, but are dominantly









either pink or white with dark-pink color bands. In thin section, the minerals appear

colorless to white, with brown staining sometimes present around the rim of the vesicle.

There are two dominant crystallization habits of thomsonite and mesolite in the

vesicles of this zone. In vesicles where there is a prevalent crystallization sequence of

early thomsonite pore rims followed by late mesolite filling of remaining pore space, the

thomsonite occurs as a massive growth of many small fibrous bundles rimming the

vesicle wall, or in some rare occurrences completely filling the vesicle. Mesolite appears

either massive with no crystal habit, or fibrous and fills any remaining space after the

thomsonite rims in vesicles. The second dominant crystallization habit originates at some

apparently random nucleation point along the vesicle wall and grows radially outward

until the vesicle is completely filled (Figure 2.5B). The growth sequence is usually

continuous from thomsonite originating at the nucleation site transitioning into mesolite

with no visible boundary between the two; in amygdules where there is a clearly defined

boundary between the thomsonite and mesolite, one, if not both minerals, exhibit a

fibrous growth habit. In all vesicles observed in this zone, the relative amounts of these

two minerals vary inconsistently. Also, clays are not present as rims along the vesicle

wall, but they do occasionally exist as intergrowths of interlayered chlorite-smectites

within the zeolite minerals, and in a majority of the amygdules, there is a late

overprinting of calcite, with crystals exhibiting twinning and rhombohedral cleavage

(Figure 2.5F).

Analcime zone lavas

This alteration zone in the NSVG lavas is characterized by the presence of

thomsonite and analcime within vesicles. This is the lowest grade of alteration observed

in the field area, and is found only in the vicinity of Lutsen. Unaltered pyroxene









phenocrysts, albitized plagioclase laths, and iron oxides, with occasional zones of

thomsonite alteration, dominate the lava matrix. No visible clay or silica rims were

observed lining vesicle walls. The amygdule minerals appear pink to colorless in hand

sample and clear in thin section. There is usually a sequence of mineralization from

thomsonite rimming the vesicle to analcime growth in the residual open space. The

thomsonite crystallization habit ranges from a massive growth of many small fibrous

bundles rimming the pore, to one or two large fibrous bundles individually filling up to

half of the vesicle, while the analcime exhibits a blocky habit. The relative amount of

each mineral varies between vesicles, and in some instances the crystallization sequence

is ambiguous.

Alteration Mineral Chemistry

Representative compositions of analyzed mafic phyllosilicates are listed in Table

2.5 and are plotted in Figure 2.6 in terms of the sum of Si + Al + Mg + Fe and the

interlayer charge (2 Ca + Na + K). The bulk of the samples have compositions

intermediate between trioctahedral smectites and chlorites, suggesting that they are

interlayered chlorite/smectite phases. These samples are found in the thomsonite-

mesolite and laumontite alteration zones. Samples representative of the stilbite-

heulandite alteration zone are dominantly trioctahedral smectites, although some

compositions are mixtures of dioctahedral and trioctahedral smectites. One last group

exhibits compositions with (2Ca + Na + K) greater than 1.5 per 28 O equivalents and lie

along an extension of the trioctahedral smectite-chlorite mixing trend. These samples are

mainly from the thomsonite-mesolite alteration zone, and may be altered celadonite (c.f.

Neuhoff et al., 1999).










Table 2.5: Representative compositions of mafic phyllosilicates.
Sample NS04-03 NS04-19 NS04-25 NS04-37 NS04-37 NS04-37 NS04-38
Oxide weight %


Si02
TiO2
A1203
FeO"
MnO
Cr203
MgO
CaO
Na20
K20
Total


Si
Ti
Al
Fe
Mn
Cr
Mg
Ca
Na
K
SAMFc
2Ca+Na+K


36.40
0.47
12.61
10.69
0.25
0.01
18.24
1.90
0.15
0.08
80.80


7.58
0.07
3.10
1.86
0.04
0.00
5.67
0.42
0.06
0.02
18.21
0.93


"Total Fe reported as FeO.
Mg + Fe in formula unit.


33.39
0.00
15.93
12.36
0.46
0.01
24.19
1.21
0.04
0.01
87.59


47.99
0.04
9.21
2.48
0.15
0.04
24.30
3.90
0.05
0.05
88.22


5.62
19.82
2.33
58.96
0.32
0.04
2.51
1.11
0.57
0.06
91.34


Anhydrous formula units


6.53
0.00
3.67
2.02
0.08
0.00
7.06
0.25
0.01
0.00
19.29
0.52


8.61
0.01
1.96
0.38
0.02
0.01
6.64
0.76
0.02
0.01
17.59
1.55


1.56
4.17
0.76
13.77
0.08
0.01
1.04
0.33
0.31
0.02
17.13
0.99


49.45
0.01
8.08
0.79
0.18
0.02
28.04
3.45
0.04
0.01
90.07


8.68
0.00
1.67
0.12
0.03
0.00
7.33
0.65
0.01
0.00
17.80
1.32


Based on 28 oxygen charge equivalents.


57.85
0.03
23.47
0.55
0.07
0.05
6.58
3.68
0.05
0.12
92.45


50.27
0.02
8.05
0.69
0.19
0.04
28.33
3.26
0.05
0.01
90.92


Zeolite compositions generally agree well with previous reported values.

Representative thomsonite compositions are listed in Table 2.6. Thomsonites span the

entire compositional range found in this mineral (between Cai.5Nai.5Al4.Si.5i.02o0nH20

and Ca2NaAl5Si5O20-nH20; Ross et al., 1992; Neuhoff and Ruhl, 2006; Figure 2.7) with

some samples exhibiting compositions less Ca-Al-rich than previously observed in this


9.44 8.72
0.00 0.00
4.51 1.65
0.07 0.10
0.01 0.03
0.01 0.00
1.60 7.33
0.64 0.61
0.01 0.02
0.03 0.00
15.62 17.80
1.33 1.23
CSum of Si + Al +











2.5
Thomsonite-Mesolite
Stilbite-Heulandite
A Laumontite
2.0


cr
(D dioctahedral A
0 1.5 smectite
CO trioctahedral
NC smectite
CL A
e 1.0
+ *
Z A
+ 0
m 0.5
N A

chlorite
0.0
15 16 17 18 19 20

Si + Al + Mg + Fe per 28 O equivalents

Figure 2.6: Compositions of mafic phyllosilicates formed during regional metamorphism
of the NSVG as a function of the number of non-interlayer cations (Si + Al +
Mg + Fe) versus the interlayer charge (2 Ca + Na + K). Sample compositions
were normalized to 28 O charge equivalents for comparison. Gray areas show
the positions of ideal endmember dioctahedral and trioctahedral smectites and
chlorite. The variation in zeolite zone alteration with relation to composition
is depicted by the black circles (thomsonite-mesolite), blue squares (stilbite-
heulandite), or red triangles (laumontite).

mineral. Mesolite and laumontite (Table 2.7) are essentially stoichiometric. Analcime

compositions have Si/Al ratios close to 2.0 (Table 2.7), similar to other occurrences in

low grade metabasalts (e.g., Passaglia and Sheppard, 2001; Neuhoff et al., 2006).

Discussion

Regional variation of metamorphic grade

One of the difficulties encountered in assessing metamorphic grade in continental

flood basalts is that the lavas are frequently overprinted by small-scale hydrothermal

systems associated with local intrusions and faults. Indeed, a number of sites visited in









Table 2.6: Representative compositions of thomsonites.
Sample NS04-05 NSI4-1I.A NS04-15C NS04-21 NS04-31A
Oxide weight %
Si02 35.75 37.09 39.25 37.02 43.35
TiO2 0.06 0.01 0.02 0.04 0.00
A1203 30.07 30.66 30.69 30.84 28.01
FeO" 0.07 0.21 0.18 0.08 0.00
MnO 0.00 0.00 0.02 0.01 0.01
Cr203 0.00 0.02 0.04 0.02 0.00
MgO 0.22 0.47 0.00 0.21 0.00
CaO 12.96 13.22 12.53 13.03 10.48
Na20 4.00 3.78 4.44 4.13 5.44
K20 0.01 0.02 0.04 0.03 0.01
Total 83.14 85.48 87.19 85.41 87.31
Anhydrous formula units
Si 5.00 5.04 5.20 5.03 5.68
Ti 0.01 0.00 0.00 0.01 0.00
Al 4.96 4.91 4.79 4.94 4.32
Fe 0.01 0.02 0.02 0.01 0.00
Mn 0.00 0.00 0.00 0.00 0.00
Cr 0.00 0.00 0.01 0.00 0.00
Mg 0.05 0.10 0.00 0.04 0.00
Ca 1.94 1.92 1.78 1.90 1.47
Na 1.08 0.99 1.14 1.09 1.38
K 0.00 0.00 0.01 0.01 0.00
Si/Al 1.01 1.03 1.09 1.02 1.31
Ca/(Ca+Na+K) 0.64 0.66 0.61 0.63 0.52
Na/(Ca+Na+K) 0.36 0.34 0.39 0.36 0.48
"Total Fe reported as FeO. "Based on 20 oxygen charge equivalents.

this study (e.g., Sites 3, 4, 8, and 10) contain well developed laumontite + quartz

alteration associated with small zones of brittle deformation that are likely small fault

zones. At Site 8 (Tofte Park), the aa and pahoehoe flows exposed along the coastline

exhibit a severely over-constrained mineral assemblage of thomsonite + scolecite +

stilbite + heulandite + laumontite + quartz. This assemblage is well-developed in other

large igneous provinces, for instance in eastern Iceland (Neuhoff et al., 1999), where it is

always associated with local hydrothermal alteration. Phase rule analysis indicates that it

is over-constrained. The minerals comprising this assemblage probably did not










crystallize together, but rather sequentially in response to changes in temperature,

pressure, and/or fluid composition. The flows at Site 8 are crosscut by zeolite-filled veins

that were likely major fluid conduits during hydrothermal alteration of this outcrop. Only

in relatively isolated vesicular zones in the center of the aa flows is the likely regional

metamorphic assemblage present where the lavas were able to escape later alteration.

Therefore, observations of metamorphism in the lava flows of the NSVG are based on

sample sites where regional metamorphic grade was not overprinted by local

hydrothermal alteration.



5.2


5.0




.0
14 15 168 1.7 1.8 1.9 2.0 2.1


o

M $
I. 4.4


4.2 -


4.0 .- -i
1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1

Ca per 20 framework O

Figure 2.7: Plot showing the compositional variation of analyzed thomsonites in NSVG
basalts (black circles). Thomsonite solid solution (dashed line) between
Cal.sNa. 5A14.5Si5.5020-nH20 and Ca2NaAl5Si5O20-nH20 endmember
compositions (blue squares) also shown (Rose et al., 1992; Neuhoff and Ruhl,
2006).












Table 2.7: Representative compositions of mesolite, analcime, and laumontite.
Sample NS04-05 NS04-10A NS04-15A NS04-15C NS04-21 NS04-14B NS04-03
Mineral Mesolite Mesolite Mesolite Analcime Analcime Laumontite Laumontite
Oxide weight %
Si02 43.79 44.43 46.42 53.58 55.56 51.53 50.64
TiO2 0.03 0.01 0.03 0.05 0.07 0.01 0.01
A1203 25.48 25.56 27.09 22.94 22.36 22.22 22.03
FeO" 0.00 0.01 0.10 0.00 0.00 0.09 0.04
MnO 0.00 0.00 0.00 0.00 0.02 0.03 0.00
Cr203 0.00 0.04 0.04 0.04 0.08 0.03 0.03
MgO 0.00 0.00 0.00 0.00 0.00 0.00 0.00
CaO 8.47 8.56 9.92 0.37 0.35 12.06 11.88
Na20 5.99 5.83 5.56 13.56 13.22 0.07 0.01
K20 0.08 0.05 0.03 0.09 0.01 0.09 0.28
Total 83.83 84.48 89.18 90.63 91.67 86.12 84.93
Anhydrous formula units
Si 2.97 2.98 2.96 1.99 2.03 3.98 3.96
Ti 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Al 2.03 2.02 2.03 1.01 0.96 2.02 2.03
Fe 0.00 0.00 0.00 0.00 0.00 0.01 0.00
Mn 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Cr 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Mg 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Ca 0.61 0.62 0.68 0.01 0.01 1.00 1.00
Na 0.79 0.76 0.69 0.98 0.94 0.01 0.00
K 0.01 0.00 0.00 0.00 0.00 0.01 0.03
Ob 10 10 10 6 6 12 12
Si/Al 1.46 1.47 1.45 1.98 2.11 1.97 1.95
Ca/(Ca+Na+K) 0.44 0.45 0.50 0.01 0.01 0.98 0.97
Na/(Ca+Na+K) 0.56 0.55 0.50 0.98 0.98 0.01 0.00
"Total Fe reported as FeO. "Number of framework oxygens in formula unit.









Figures 2.8 and 2.9 synthesize the observations of regional metamorphic grade

from this study and those of Schmidt (1993) in terms of their spatial and stratigraphic

distributions, respectively. Many of the localities studied by Schmidt (1993) were


Figure 2.8: Generalized map of northeastern Minnesota showing the distribution of
NSVG lavas (gray), the locations listed in Table 2.1, and interpretations of
metamorphic grade based on this study and the work of Schmidt (1993) and
Schmidt and Robinson (1997). Metamorphic mineral zones correspond to
standard metamorphic facies designations (prehnite-pumpellyite and
greenschist). Inferred boundaries between mineral zones are shown by the
dashed lines.


- ii


HGrand Portage
Pfovland


Grand Marais ,
41


<16


\ Approximate mineral zone
\ boundary at coast
o Sampling Location (Table 1)


* Town


* J










Mineral Zones
0-
10 Analcme 9 -Lutsen Basalts
d Thomson ite-Mesolite 5 3 ICut Face CreekSandstone
500 Schroeder Basalts- ood Harbor Bayndesis
S. Stilbite-Heulandit Good Harbor Bay Andesites
Stilbite-Heulan dite BreakwaterBasalt
Mn 12 '-Breakwater Basalt
S10- -Grand Marais Felsites
CU 1000-
Bell Harbor Lavas- -Crofiville Basalts
S1500- Palisade Head Rhyolite- -Devil's Track Rhyolite
a I- -Maple Hill Rhyolite
a- Baptism River Lavas- -Red Cliff Basalts
2000-
a Laumontite -Kimball Creek Rhyolite
C 2500-
Gooseberry River Basalts- 11 -Marr Island Lavas
S3000- --Naniboujou Basalts
o Two Harbors Basalts- -Devil's Kettle Rhyolite
3500-
Larsmont Basalts
rule River Lavas
4000f Sucker River Basalts- Prehnite-Pumpellyite
Chicago Bay Lavas

Southwest Limb Northeast Limb

Figure 2.9: Stratigraphy of the NSVG (Vervoort and Green, 1997; Miller et al., 2002)
correlated along the fold axis dividing the lavas into the southwest and
northeast limb. Numbers in stratigraphic columns refer to sampling sites in
this study (Figure 2.1, Table 2.1). Mineral zones based on interpretations
from this study and Schmidt (1993) and Schmidt and Robinson (1997).

revisited in this study, and in some cases reinterpreted in light of the complex

metamorphic history noted above. Along the southern limb, the transition from the

laumontite zone to the higher grade prehnite-pumpellyite zone is placed at Site 1 where

primary pore spaces in adjoining lava flows indicate a change in mineral assemblage.

The laumontite zone continues up through the stratigraphy for about 2700 meters until

observations of stilbite and heulandite at Site 12 suggest a transition to the lower-grade

the stilbite-heulandite zone. The onset of this alteration coincides with the basal contact

of the Schroeder basalts along the southern limb, which shows a stratigraphic thickness of

-900 meters assuming the hinge axis as the upper extent of the section. The lower half of

the Schroeder basalts contains the stilbite-heulandite zone, while the upper half of this









section shows thomsonite-mesolite zone alteration. This transition implies a correlation

between shallower stratigraphic depth and lower grades of alteration.

The lowest grade of alteration noted during this study occurs at Site 9 (Lutsen

Grandview Park) and corresponds to the lowermost analcime zone. This outcrop consists

of a series of vesicular pahoehoe flows containing abundant thomsonite alteration. The

exact extent of the analcime mineral zone is unclear due to limited exposures, though

what is exposed shows no significant stratigraphic thickness. Typically in large igneous

provinces there is a well-developed chabazite-thomsonite zone overlying the analcime

zone (e.g., Walker, 1960; Kristmannsd6ttir and T6masson, 1978; Larsen et al., 1989;

Neuhoff et al., 1997, 2000, 2006) that is frequently as thick as the analcime, mesolite-

scolecite (likely of equivalent grade to the thomsonite-scolecite zone observed in the

NSVG), and stilbite-heulandite zones combined. The absence of this zone, together with

the limited extent of the analcime zone and the thicknesses of the thomsonite-mesolite an

stilbite-heulandite zones (Figure 2.9), implies that approximately 1000-2000 meters of

the volcanic stratigraphy may have been lost to erosion in the NSVG.

The variation in metamorphic grade in the northeast limb of the NSVG appears to

mirror that in the southwest limb, where metamorphic grade increases with stratigraphic

depth. The Lutsen basalts exposed along the northeast limb of the NSVG are thinner

(-450 meters) and contain well-developed thomsonite-mesolite zone alteration down to

the lower stratigraphic boundary with the Cut Face Creek sediments. This sediment pile

is -20 meters thick, and the andesites directly below them show laumontite zone

alteration. Limited accessibility precluded assessment of metamorphic grade northeast of

locality 7, and thus the total thickness of the laumontite zone in this limb is









unconstrained. The most striking difference between the two limbs is the absence of

stilbite-heulandite zone alteration in the northeast limb, discussed below.

Conditions of Alteration

Observed mineral assemblages of the NSVG are comparable to similar alteration of

the lavas in East Greenland, eastern Iceland, and the Icelandic geothermal systems

(Walker, 1960; Kristmannsd6ttir and T6masson, 1978; Larsen et al., 1989; Neuhoff et al.,

1997, 2000). The distribution of zeolite minerals with depth/temperature in basaltic

terrains of Iceland and East Greenland are summarized in Figure 2.10 (Neuhoff et al.,

1997; 1999; 2000). The minerals observed to develop within each zeolite zone are shown

with respect to temperature constraints of the mineral distributions typically observed in

tholeiitic lava flows. Each mineral zone forms during alteration, transitioning to the next

with increased temperature/depth.

It has been observed that thick sequences of lavas with varying chemistries exposed

to the same alteration conditions will often exhibit substantially different mineral

parageneses (e.g., Walker, 1960a; Robert, 2001; Neuhoff et al., 2006), causing a need for

observation of rock type along with alteration minerals present. The Schroeder-Lutsen

basalts of the NSVG are lower in silica than those of the Greenland and Iceland provinces

showing similar mineralogy, which may stabilize thomsonite relative to more silica-rich

minerals like mesolite and scolecite (Neuhoff et al., 2006). Therefore, the zone in Figure

2.10 labeled as the mesolite-scolecite zone is interpreted to be equivalent in grade to the

thomsonite-mesolite zone.














u,,-
'E
0E^a
0~ 0_ ^K J


0 a



E
3i C

Up"


V)
X
0 C
0
L 3
u-u0


U,

Ca


4C
v,53
3 TT
0 0


En
Q.u


C)



E
EL MINERAL ZONES


0-
no
zeolites
25-
Si i chabazite

LU 50- I I I thomsonite

SI i analoime

< 75 -I I mesolite +
Sscolecite

100 heulandte + stilbitE

HI laumontite
125-


1' 5prehnite-bearing
150-
15-facies


200-


Figure 2.10: Schematic diagram showing the distribution of minerals and mineral zones
typically developed during very low grade metamorphism of large igneous
provinces (after Walker, 1960; Neuhoff et al., 1997, 2000). Frequency of
mineral occurrence with depth is denoted by solid (ubiquitous) and dashed
(occasionally present) lines. Mineral list is non-exclusive. Temperatures
listed on left side of diagram are based on observations by Kristmannsd6ttir
and T6masson (1978) of the distribution of minerals as a function of
temperature in Icelandic geothermal systems.

Figure 2.9 shows the estimated stratigraphic thicknesses given by Miller et al.

(2002) combined with my observations of regional metamorphic grade boundaries with

stratigraphic depth for the NSVG lavas. Comparisons of these observations to the

observations in Figure 2.10 can be used to interpret the thermal and structural history

experienced after emplacement (Neuhoff et al., 2000). Observations show zeolite to









greenschist facies alteration along the exposed stratigraphy of the NSVG (Schmidt and

Robinson, 1997; Neuhoff et al., 2000), where alteration grade increases with stratigraphic

depth along the exposure. Due to the synclinal structure of the region, the youngest lavas

exposing the lowest grade of alteration are found near the hinge axis. This fold axis

divides the region into two limbs, and stratigraphic depth as well as alteration grade

increases to the southwest and northeast.

Analcime zone alteration, the lowest grade of alteration observed within the NSVG

basalts, occurs near the hinge axis separating the Schroeder basalts in the southwest limb

from the Lutsen basalts in the northeast limb. There is no significant thickness to this

zone, though observations of analcime alteration in similar provinces predict relatively

consistent stratigraphic thicknesses of each alteration zones. The limited thickness of the

analcime zone, as well as the absence of alteration minerals exhibiting temperatures and

pressures lower than analcime in grade, is likely due to erosion of once exposed

stratigraphically younger lavas.

Stratigraphically below the analcime zone is the thomsonite-mesolite zone of

alteration, exhibiting a consistent thickness in each limb of approximately 500 m. The

lower thomsonite-mesolite boundary in the northeast limb is in direct contact with the

higher grade laumontite zone, while in the southwest limb there is a 400 m thick stilbite-

heulandite zone of alteration separating the thomsonite-mesolite zone from the laumontite

zone. The thickness of the laumontite zone in the northeast limb is at least 1500 m,

though the exact extent is unknown due to inadequate exposures further to the northeast.

Assuming no structural complications, it can be expected to be as thick as the laumontite

zone in the southwest limb, which is 2700 m thick and bounded below by the prehnite-









pumpellyite zone. The prehnite-pumpellyite zone is the highest grade of alteration

observed in the NSVG basalts, with the onset of alteration occurring at a stratigraphic

depth of 3,860 m below the fold axis. The thickness of this zone is unknown due to a

lack of observations in outcrops further to the south.

The boundary constraints defined by the mapped zeolite zones can be correlated to

the thermal boundaries corresponding to mineral isograds in Icelandic geothermal

systems (Figure 2.10) to estimate average geothermal gradients during alteration. The

well-constrained isograd locations bounding the thomsonite-mesolite and stilbite-

heulandite zones in the southwest limb give a combined thickness of 900 meters, and

indicate a temperature range of -40 C. This implies a geothermal gradient of the NSVG

region during alteration of approximately 42 C/km, which compares well to the

estimated geothermal gradients of 55 C/km observed in LIPS formed during continental

rifting in eastern Iceland (Neuhoff et al., 1999) and 40 C/km in East Greenland (Neuhoff

et al., 2000).

This information also can be used to calculate overburden pressure and heat flow

experienced during the time of alteration. The estimated stratigraphic thickness of the

basalts experiencing zeolitization during the time of alteration is -5500 m after

combining the known stratigraphy with the estimated loss due to erosion. Pressures thus

increased from a minimum of 1 bar (assuming atmospheric pressure) to maximum

pressures ranging from -550 to -1600 bars, corresponding to hydrostatic and lithostatic

pressures, respectively. Regional heat flow over the area during alteration was about 1.3-

2.2 heat flow units, given a thermal conductivity for water-saturated basalt of 1.5-2.0

W/mK (Oxburgh and Argell, 1982) and the implied geothermal gradient for the NSVG of









420 C/km. These values are comparable to heat flow in present-day continental rift

systems (Lysak, 1992).

Structural Interpretations

The overall structure of the NSVG is a plunging synclinal fold with a hinge axis in

the area between Tofte and Lutsen (e.g. Miller et al., 2002). This fold has been

interpreted as a subsidence feature formed during eruption of the lava pile due to

evacuation of the underlying magma chamber and increased density of the growing lava

pile (e.g., Ojakangas and Matsch, 1982), or alternatively as a late structural modification

of the area due to strong influences from the nearby Grenville orogen (Van Schmus and

Hinze, 1985).

The observation that metamorphic grade increases with stratigraphic depth, in both

limbs of the province, has considerable implications for the structural development of the

NSVG. Based on my field observations combined with those of Schmidt (1990), the

distribution of secondary mineral paragenesis represented in the NSVG suggests that the

metamorphic zones are folded along with the lavas. Regional metamorphism in large

igneous provinces typically is initiated soon after volcanism and is of short duration (e.g.,

< 1 Ma; Neuhoff et al., 1997). Thus, if the folding of the NSVG was syn-volcanic, one

would expect the mineral zones to cross-cut the stratigraphy, whereas the isograds would

be folded along with the lavas if deformation were due to post-volcanic processes.

Combining this information with the observation that the regional metamorphism is

typically rapid and of short duration after complete eruption of the lava pile in large

igneous provinces (e.g. Neuhoff et al., 1997, 1999, 2000), it appears that folding of the

NSVG is a post-volcanic feature and was not a mechanism of crustal accommodation of

the immense lava pile developed in this province.









The stratigraphic distributions of zeolite zones in the NSVG (Figure 2.9) are not

symmetrical between limbs, as indicated by the absence of the stilbite-heulandite zone in

the northwest limb of the province. Previous observations of zeolite zones in other

continental flood basalt provinces have shown that these zones are relatively continuous

and uniform in thickness (Walker, 1960b; Neuhoff et al., 2000). Correlation of the

overlying thomsonite-mesolite zone between the northeast and southwest limb show

continuous and uniform thickness, implying that this should be the case in the continental

flood basalts of the NSVG. The stilbite-heulandite zone, however, is mapped in the

southwest limb with a thickness of almost 450m but is not present in the northeast limb.

The expected position of the stilbite-heulandite zone in the northeast limb corresponds to

the location of the Cutface Creek sediments, which are only 30-40 m thick and thus not

extensive enough to contain the whole range of stratigraphic depth expected for this zone.

Two possible explanations might account for the lack of the stilbite-heulandite zone

in the northeast limb. The first is based on the fact that there is a marked change in lava

composition between the thomsonite-mesolite zone occurrences at site 3 (and

immediately overlying the Cutface Creek sediments) and the laumontite zone lavas at site

4, with the latter exhibiting more silicic basaltic andesite compositions (e.g., Walker,

1960a; Kristmannsd6ttir and T6masson, 1978; Murata et al., 1987; Hearn et al., 1989;

Robert, 2001; Neuhoff et al., 2006). Potentially this change in lava composition may

result in laumontite occurring at lower grades. This should, however, stabilize the more

Si-rich minerals stilbite and heulandite relative to laumontite, not the other way around.

A more plausible hypothesis is that there is a loss of stratigraphy in the region between

sites 3 and 4. High-angle reverse faults are observed in the central and western regions of






44


this area due to compression, causing juxtaposition of older rocks next to younger rocks

(e.g., White, 1966; Van Schmus and Hinze, 1985). Reverse faulting may have also

occurred in the eastern region of the NSVG, and could explain the absence of the stilbite-

heulandite zeolite zone in the northeast limb. The potential displacement of this

alteration zone therefore occurred as relatively late faulting of the NSVG basalts, which

agrees with the suggested post-volcanic compression.














CHAPTER 3
PETROGRAPHIC AND DIGITAL ANALYSIS OF POROSITY EVOLUTION
DURING ALTERATION

Introduction

Permeability in basaltic lavas has received considerable attention recently due to

increased focus on their hydrologic properties and the role of degassing of volatiles in the

dynamics of eruptive events (e.g., Fisher, 1998; Saar and Manga, 1999; Blower, 2001a, b;

Srouga, 2004; Stimac, 2004). From this work, it is becoming increasingly clear that the

geometry and connectivity of the pore network in vesicular lavas is fundamentally

different than in more commonly considered porous media such as sediments and

fractured rocks, and more attention is being placed on the relationship between porosity

and permeability. For lavas with porosities >60%, there appears to be no significant

relationship between porosity and permeability; when porosities are below 60%,

however, experimental results from different studies show varying porosity-permeability

relationships (Eichelberger et al., 1986; Klug and Cashman, 1996; Saar and Manga,

1999).

In vesicular basalts the permeability can be as high as 10-7 cm2, which is a value

comparable to clean unconsolidated sand or karstic limestone. In basalts where the

vesicles are small and closely spaced, there will be an increased probability of connected

fluid pathways developing between pores that will likely lead to these higher

permeabilities (e.g., Saar and Manga, 1999). Percolation theory (Lee, 1990; Sahimi,

1994; Grimmett, 1999) applied to basalts in experimental studies shows that the









percolation threshold, the critical porosity below which percolating clusters of bubbles

should not exist, is around 30%, implying that little to no permeability can be expected

below critical porosity. In these low porosity basalts, the vesicles become more isolated.

If connectivity between pores is dependant on fluid flow through the matrix, permeability

can be as low as 10-18 cm2 (Johnson, 1980; Saar and Manga, 1999). Multiple studies

have shown, however, that vesicular lavas retain relatively high permeabilities (>10-8

cm2) down to porosities as low as a few percent (Feng et al., 1987; Sahimi, 1994, 1995;

Saar and Manga, 1999; Blower, 2001b). This does not follow the predicted model for

granular rocks, and implies that there must be pathways forming within the matrix

allowing fluid to flow at a faster rate.

Differences in porosity within basaltic lava flows leads to considerable flow-scale

variations responsible for the development of secondary mineral paragenesis (Schmidt,

1990; Schmidt and Robinson, 1997). Under low-grade alteration, the chemical

components necessary to precipitate the vesicle-filling minerals are derived from a

spatially restricted region (alteration halos) around the vesicle and the fluid pathways

over which diffusion occurs. This suggests that the presence of mineral infillings can be

a reliable indicator of pores being part of a permeable network through the lava, as fluid

flow is inherently coupled to changes in porosity and permeability brought about by

mineralogic alteration. This chapter explores using mineralogic alteration to trace fluid

pathways through the lavas and the temporal evolution of pore space through time.

These observations will prove important not only for further understanding of

groundwater and petroleum reservoirs in basalts, but also developing new methods to

monitor the controls and affects of fluid flow and alteration within vesicular rocks.









Methods

Field Techniques

Field work was conducted along the north shore of Lake Superior in Minnesota

during July 15-26, 2004. Of the sampling sites and methods discussed in Chapter 2,

measurements were performed at sites 3, 4 and 6 (Table 2.1, Figure 2.1) of total initial

porosity, pore size distributions, and the amount and type of secondary alteration within

primary pore space and within the matrix.

Measurements of total porosity were made both on the thin section scale through

digital analysis as well as on the outcrop scale in the field. All porosity in analyzed

samples is essentially zero, due to the complete infilling of all vesicles by secondary

alteration, so primary porosity is analyzed based on the initial void space in vesicles prior

to alteration. In situ field measurements of primary macroscopic porosity (defined in this

paper as the percentage of a given volume of rock occupied by pore space) were made in

the field following the methods of Manning and Bird (1991). This is done by placing a

transparent mylar sheet embossed with a precision drafted grid (1073 nodes) directly on

the outcrop and counting the number of nodes on the grid that overlay pores. Porosity

was taken to be the percentage of nodes overlying pores (Manning and Bird, 1991).

Measurements were made to assess the pore size distributions within sampled lava

flows. Overall pore size distributions were assessed by measuring the maximum and

minimum diameter of all pores within a representative area. Notes were made on the

progress of reaction of each measured pore by assessing the level of secondary alteration

observed with respect to pore size. Observations also were made of the amount of

alteration visible in the matrix around and between pores to assess the characteristics of









fluid flow through the rock as well as the amount of lava that has undergone alteration

during low-grade metamorphism.

Optical and Digital Techniques

Numerous techniques have been applied to determine the distribution and character

of macroscopic pore space in lavas, including impregnation with plastics and dissolution

of the lava to reveal casts of pore structure (e.g., Sahagian et al., 1989), three-dimensional

imaging by X-ray computed tomography (e.g., Song et al., 2001; Sahagian et al., 2002),

and digital analysis of two dimensional sections (e.g., Toramaru, 1990; Klug and

Cashman, 1996; Sahagian and Prousevitch, 1998; Al-Harthi et al., 1999; Saar and

Manga, 1999). In this study, I adopt the last approach, because digital analysis of two

dimensional sections will permit direct correlation of pore structure to the chemical and

mineralogical observations.

Three thin sections from different regions are examined; a chabazite zone

metabasalt from East Greenland, a mesolite-scolecite zone metabasalt from eastern

Iceland, and a laumontite zone metabasalt from the NSVG (Site 4). These samples are

chosen to represent progressive stages of zeolite alteration with respect to the range of

pore evolution experienced during alteration. Thin sections are prepared for each sample,

and digital images are prepared from the original thin section. This is done either by

capturing digital images through a high powered petrographic microscope, or by scanning

the thin section using a backlit flatbed scanner to create a high resolution image. These

images were imported into the image processing software Adobe Photoshop, which

was used to exaggerate the contrast between the features of interest and the background

groundmass, remove residual background noise created during scanning, and convert the

image to grayscale. The images were then imported into either Image J or ArcView









GIS 3.2a (ArcView); both programs were used to measure and statistically analyze the

porosity and secondary mineral phases in the images. When imported into ArcView the

images were converted into shape files by creating a vector data format, which represents

features as polygons defined by coordinate pairs, allowing for data analysis to be

performed on the image. For porosity measurements, these methods were limited to

pores visible at the thin section scale. This excludes macro-pores larger than the size of a

standard thin section (27 x 46 mm) or micro-pores smaller than the pixilation ability of

the digital image processing software (which varies between programs).

Observations

NSVG Field Sites 3, 4 and 6

In situ field measurements of total original macroscopic porosity were made at

Sites 4 and 6 (Table 2.1) within the NSVG, where outcrop exposures were adequate for

sampling methods. Porosity measurements in the field were limited to macroscopic

observations, which confined analysis to the high porosity tops and bottoms of each lava

flow. Two measurements at Site 4 show primary porosities of 19.7% and 21.7%, and one

measurement at Site 6 shows a primary porosity of 15%. This highlights the localized

variation of porosity experienced within these high porosity zones of each lava flow.

Figure 3.1 shows a close-up of an outcrop exposure at Site 6 that is also

representative of porosity at Site 4. It can be observed from this figure that vesicle shape,

size and spacing vary considerably, and the numbered regions highlight examples of this

random distribution. Circle 1 highlights an area of little to no porosity surrounded by

vesicular areas. The average vesicle was oval to round in shape, though many vesicles

exhibited little to no sphericity appearing more elongated as highlighted by the vesicle

(4x12 mm) in circle 2. Vesicle size distributions within these measured zones were









highly variable, with average vesicle radii ranging from 46 to 1 mm. Table 3.1 gives the

statistical analysis of vesicle size measurements performed on the outcrop of all

amygdules, as well as those showing only thomsonite alteration or thomsonite and


Figure 3.1: Outcrop photograph of vesicular basalt at Site 6 (Table 2.1) illustrating
typical variation in pore size, shape and spacing. Ruler (15 cm long) shown
on the left for scale. Vesicles are filled with white thomsonite-mesolite
aggregates surrounded by the gray basalt. Numbered circles highlight various
porosity characteristics of basalt; highlighted areas show 1) little to no
porosity, 2) elongated vesicle, and 3) pores of various sizes in close proximity
to each other.

mesolite alteration (discussed further below). Circle 3 in Figure 3.1 contains vesicles

within close proximity of each other that range in size from ~1 to 10 mm. In each of

these highlighted regions and in the areas surrounding them, there are no vesicles that are









visibly connected. The lack of interconnection between vesicles was observed at each

sampling site.

Table 3.1: Statistical analysis of vesicle size (diameter) as a function of reaction progress
from outcrop scale measurements at Site 6.
All vesicles Thomsonite only Thomsonite + Mesolite
Statistical Dataa
n 185 125 160
Size Range 1 46 1 10 10 46
Mean 9.48 2.62 23.76
Mode 1.00 1.00 22.00
St. Dev.b 11.10 2.25 8.17
Size Distributionc
25% 17.00 1.00 1.00
50% 22.50 2.00 4.00
75% 28.50 3.00 17.0
100% 46.00 10.00 46.00
aAll diameter values are given in mm. bStandard Deviation. TPercentage of pores < given
value.

All vesicles were completely occluded due to secondary alteration, causing them to

appear white or pink in outcrop. Many of the sites contained vesicles that appeared

partially or completely unfilled, and the presence of such vesicles increased along river

bed and shoreline exposures. Cut samples exposing fresh surfaces, however, showed

complete infilling of all vesicles, thus these open vesicles can be attributed to weathering

of alteration minerals. Sites 3 and 6 contained vesicles filled with either thomsonite or

thomsonite and mesolite (Table 3.1). Vesicles at Site 4 were filled with either chlorite or

chlorite and laumontite with occasional calcite overprinting of the laumontite.

Observations of collected NSVG samples show that matrix alteration varies within

a given sample from areas with little alteration to areas showing complete alteration of

primary magmatic phases. These areas of high alteration appear to exist as reaction

aureoles around primary pore spaces, and around what appear to be microfractures acting

as pathways between vesicles through the matrix. Figure 3.2A (also shown in Figure 2.4)









depicts the visible aureoles at the outcrop scale (Site 3), appearing as light-pink areas

around and between amygdules. Figure 3.2B shows a schematic representation of these

observed aureoles. The aureoles around vesicles are generally thicker than those around

the possible microfractures between vesicles. In many cases the veins appear as

connecting pathways between vesicles.














Figure 3.2: Images of reaction zones around and between macroscopic pore space in
NSVG lavas. A) Field photo from Site 3 (Table 2.1) showing an extensively
altered pahoehoe flow with bleached haloes around vesicles connected by thin
anastomosing bleached areas. Pencil for scale. B) Schematic depiction of
matrix alteration shown in (A), with broad alteration halos (white) around
vesicles (light gray) connected by thin strands of alteration.

These aureoles were first observed as a weathering phenomenon at the outcrop

scale, and further analysis of cut hand samples and thin sections (Figure 3.3) show that

they are present throughout the entire rock. They are clearly visible as lighter areas

around and between vesicles, though at higher magnifications they have a dark

appearance compared to the less altered matrix surrounding them.

Petrographic observations at the thin section scale show that alteration appears to

occur in stages that are independent of both lava composition and grade of alteration.

Figure 3.4 depicts three stages of alteration observed within the matrix of analyzed



































Figure 3.3: Image is of sample NS04-14, showing the laumontite zone alteration found at
Site 4. Image was captured using a reflected light binocular microscope.
Chlorite is the first mineral to fill the pore, as seen by the dark rim around the
pore wall. Residual open space is filled by laumontite, the white mineral in
the center of each pore. Alteration aureoles are seen here as light colored rims
around and between pores.

NSVG samples, with minor alteration in 3.4A to extensive alteration in 3.4C. Figure

3.4A shows a sample containing the highest percentage of unaltered matrix (lighter

matrix) where the areas of alteration (darker band in matrix) appear as networking veins.

In samples exhibiting more alteration (Figure 3.4B), the areas of unaltered matrix are

smaller and are surrounded by thicker networking veins. Other samples show no

networking of alteration (Figure 3.4C), and instead the entire matrix has undergone

alteration.







































Figure 3.4: Various levels of secondary alteration in the matrix ofNSVG basalts sampled
during the field season. All photomicrographs were taken through crossed
polars. A) Alteration (darker areas) occurs in a network pattern through the
matrix (sample NS04-31), surrounding areas of unaltered matrix and
connecting zeolitized amygdules. Fresh feldspars and pyroxenes can be
observed in the unaltered matrix (lighter areas). B) Thicker network pattern
of matrix alteration (sample NS04-15) surrounding small patches of unaltered
matrix where fresh feldspars and pyroxenes can still be observed. C) Matrix
has undergone complete alteration (sample NS04-14), with visible feldspar
laths but no residual pyroxene.

The dominant minerals observed within these alteration aureoles are iron oxides,

clay minerals, and zeolites. There is a dominant expression of iron oxides occurring

within the altered matrix further away from pore spaces, while zeolite mineralization

increases closer to primary pore spaces. Samples exhibiting more intense alteration show

the complete disappearance of primary pyroxene and alteration of the plagioclase laths.









Reaction Progress

Low-grade metabasalt samples exhibiting zeolite zone alteration were selected to

show varying levels of secondary mineral formation through increasing alteration from

chabazite to mesolite-scolecite to laumontite zeolite zones. Samples from Greenland,

Iceland, and Sites 4 and 6 from the NSVG were analyzed for aspects of pore geometry

such as spacing, size and shape and were observed with respect to the extent of reaction

progress experienced during alteration. The progress of reaction is evident through the

progressive infilling of vesicles. At very low grades, the vesicles experience partial

infilling, and as grade increases, so does the amount of secondary mineral precipitation

within the vesicle. Vesicles are progressively filled with clay minerals, clay and zeolite

minerals, or multiple generations of zeolite minerals.

Partial infilling at low-grades

A low-grade metabasalt from East Greenland exhibiting chabazite zone alteration

(Neuhoff et al., 1997) was used for analysis of very low grades of alteration, as alteration

lower in grade than the analcime zeolite zone are not preserved within the NSVG lavas.

A thin section (421505) representing chabazite zone alteration is shown in Figure 3.5A.

This thin section contains vesicles exhibiting three levels of alteration, those that are

variably empty, completely filled with chabazite, or partially filled with chabazite, and

analysis is performed to assess vesicle properties in relation to these alteration

characteristics. Figure 3.5B shows the thin section converted into a shape-file for area

analysis in ArcView 3.2a. The region highlighted in both images by the dashed line

depicts an area with very little secondary mineral infilling. The lack of mineralization is

independent of vesicle size and shape, and instead may be an effect of limited fluid flow

through this region during alteration.

















































Figure 3.5: Example of digital analysis of vesicle fillings in a zeolite facies vesicular lava
from East Greenland. The original thin section (A) contains vesicles (278
individual objects, comprising 13.2% of the sample) that are partly filled with
chabazite. The region highlighted within the dashed line (in A and B) shows
almost no secondary mineralization; see text for discussion. Part (B) shows
separation of the three degrees of vesicle-scale alteration; vesicles unaffected
(unfilled) by alteration (light brown), completely-filled vesicles (dark brown),
and partially-filled vesicles (light and dark brown).









Area measurements were conducted on 284 vesicles, and each vesicle was then

categorized with respect to the level of alteration with values shown in Table 3.2. Total

original porosity of the sample (calculated from the area of vesicle space divided by the

total thin section area) is 13.2%, and the zeolite infillings occupy 70.4% of the original

pore space (corresponding to a 9.4% increase in solid volume during alteration). Figure

3.6 shows the frequency of vesicles as a function of area and the level of alteration

observed within the pore. Visual inspection of Figure 3.6 suggests that, in this sample,

chabazite infilling is most prominent in larger vesicles, and that larger chabazite-filled

vesicles are more likely to retain porosity after zeolite formation. The large spread in

data shows that while the level of alteration is affected by vesicle size (average area of

vesicles retaining open porosity is larger than those completely filled), this may not be

the only controlling factor.

Table 3.2: Statistical analysis of vesicle size (area) as a function of reaction progress from
measured vesicles of a low-grade metabasalt from East Greenland (thin
section 421505). Vesicles are either completely filled (filled), completely
open (unfilled), or partially filled with residual pore space (partially filled).
All vesicles Filled Unfilled Partially filled
Statistical Dataa
n 278 68 189 21
Size Range 0.001 2.502 0.001 1.395 0.001 0.543 0.053 2.502
Mean 0.094 0.082 0.035 0.661
Mode 0.002 0.002 0.002
St. Dev.b 0.239 0.090 0.072 0.586
Size Distributionc
25% 0.004 0.011 0.003 0.280
50% 0.018 0.051 0.010 0.401
75% 0.086 0.126 0.033 0.907
100% 2.502 0.395 0.543 2.502
aAll area values are given in mm2. Standard Deviation. cPercentage of pores < given
value.










70
m unfilled vesicles
60 m partially filled vesicles
O filled vesicles
50

S40 o
|| 0.082 |_
r30
4-
20

10


0.001 0.005 0.01 0.05 0.1 0.5 1 5
vesicle area (mm2)


Figure 3.6: Frequency of vesicle areas versus the amount of secondary alteration filling
the vesicle for a sample of chabazite-thomsonite zone alteration from East
Greenland (cf Neuhoff et al., 1997). Vesicles are unfilled, partially filled or
completely filled with chabazite. Average vesicle area for each level of
alteration is labeled.

Mafic phyllosilicate to zeolite infilling

Observation of lavas exhibiting zeolite alteration higher in grade than chabazite-

thomsonite zone found no residual pore space in vesicles observed in thin section. The

relationship between reaction progress and alteration minerals then becomes related to

the progressive stages of mineral precipitation within vesicles. Low-grade metabasalts

from Site 4 in the NSVG and from eastern Iceland exhibit early precipitation of mafic

phyllosilicates followed by later stage zeolite infilling.

Samples collected during the field season at the Cutface Creek location (Site 4, see

Table 2.1) within the NSVG exhibit laumontite zone alteration. Microprobe analysis of a

sample in thin section shows each vesicle initially filled with chlorite rims (Figure 3.3),









with any residual pore space completely filled with laumontite during later stages of

alteration. Visual inspection of vesicles (Figure 3.3) also shows a consistency in

thickness of chlorite rims irrespective of vesicle size and shape. All vesicles within this

thin section were measured for total area and for the amount (in area) filled by laumontite

and chlorite, and values are represented in Table 3.3. Chlorite rim thicknesses in this

sample also were measured (Table 3.3) and are statistically consistent throughout the thin

section (0.52 +/- 0.12 mm). Figure 3.7 plots this consistency in thickness with respect to

vesicle area, as well as the percentage of the vesicle filled with chlorite, and shows that

the decrease in chlorite infilling as vesicle size increases is due to the consistency in

chlorite rim thicknesses. Figure 3.8 again shows the percentage of chlorite infilling with

respect to area, and also shows the percentage of laumontite infilling with respect to area.

Due to the consistency in clay rim thicknesses, the percentage of each vesicle filled with

chlorite decreases as the area of the pore increases, leaving more residual space in larger

pores for later infilling of laumontite.

Table 3.3: Statistical analysis of vesicle size as a function of reaction progress for
vesicles filled with chlorite and/or laumontite at the thin section scale (sample
NS04-14) from Site 4. Area data is given in mm2 for the entire vesicle and for
the area of the vesicle filled by laumontite and chlorite. Chlorite rim thickness
measurements are given in mm.
Vesicle Area Laumontite Area Chlorite Area Chlorite thickness
Statistical Data
n 23 23 23 255
Size Range 1.39 23.30 0.00 16.84 1.29 14.21 0.29 0.90
Mean 10.33 5.01 5.32 0.52
Mode -- 0.00 -- 0.55
St. Dev.a 6.43 4.76 3.16 0.12
Size Distributionc
25% 3.76 1.36 2.41 0.41
50% 5.17 1.71 3.46 0.67
75% 14.12 7.28 6.84 0.50
100% 16.43 15.14 1.29 0.51
"Standard Deviation. bPercentage of pores < given value.












100 DO chlorite thickness (mm)
00 0 chlorite infilling (%)

80- 0 0


0 0 0
60 -o o
0 0 0
o o o


40 O Oo

20
20 -

0
0 0* 6* 0 0@ ** 0O0 0

0 5 10 15 20 25

vesicle area (mm2)

Figure 3.7: Plot shows the thickness in mm of chlorite rims (black circles) and the total
percent of the vesicle filled with chlorite (open circles) with respect to total
2
vesicle area in mm2. As vesicle area increases, the chlorite rims maintain a
consistent thickness of -0.52 mm, while the percentage of vesicle occlusion
by chlorite steadily decreases.

Further analysis of this consistency in thickness of mafic phyllosilicate rims was

performed on a previously collected sample from eastern Iceland. Mesolite-scolecite

zone alteration in a sample from eastern Iceland contains vesicles completely filled by

interlayered chlorite-smectite (c/s) and scolecite. Each pore was initially filled with clay

rims during early stages of alteration, leaving residual pore space in larger pores that

becomes completely filled with scolecite during later stages of alteration (Figure 3.9).

Figure 3.9 highlights three vesicles, each exhibiting a different size and shape. Each

vesicle is initially filled with rims of C/S that are of a consistent thickness between

vesicles. The smallest vesicle is almost completely occluded by the C/S while the other






61





100- o 0 chlorite
O0 A laumontite


A
80- o


60 0 A
5A 00 A 0
SA A 0 A
40 -
E A4


20 A


0 Au,

0 5 10 15 20 25

total vesicle area (mm2)

Figure 3.8: Plot shows the percentage of mineral infilling with respect to total vesicle
area. Smallest vesicles are completely filled with chlorite. As the vesicle size
increases, the percentage of the vesicle filled with chlorite decreases, and the
residual space is filled with laumontite.

two retain pore space for scolecite infilling. Measured clay rim thicknesses are given in

Table 3.4 and are shown to be consistent throughout the thin section (0.091 +/- 0.02 mm)

irrespective of pore size, although a slight variation in thickness is observed in non-ovoid

pores (Figure 3.9) due to geometric effects of taking a two-dimensional slice of a three-

dimensional object.






































Figure 3.9: Photomicrograph showing a thin section of a mesolite-scolecite zone lava
from eastern Iceland. Each pore is filled initially with a chlorite-smectite clay
rim; larger pores are subsequently filled with scolecite. Clay rim thicknesses
are consistent throughout the section (0.10 +/- 0.02 mm) irrespective of pore
size, highlighted by the two circled pores and the larger pore between them.
Pores smaller than 0.1mm2 are completely filled with clay.

Digital analysis of pore sizes compared to level of infilling is performed using

ArcView and Image J software. The vesicles are analyzed as either filled with both C/S

and stilbite or completely occluded by c/s. This information is plotted in Figure 3.10, and

shows that pores with an area of 0.1 mm2 or less are completely filled with clay, while

larger pores are filled with both clay and zeolites due to the consistency in C/S

thicknesses.









Table 3.4: Statistical analysis of vesicle areas (mm2) and clay rim thicknesses (mm) for
sample 94-80 from eastern Iceland.
Vesicle Area Clay thickness
Statistical Data
n 229 344
Size Range 0.001 2.191 0.038 0.163
Mean 0.079 0.091
Mode 0.003 0.092
St. Dev." 0.245 0.022
Size Distributionb
25% 0.004 0.104
50% 0.013 0.088
75% 0.043 0.046
100% 2.191 0.068
"Standard Deviation. bPercentage of pores < given value.

Multiple stage zeolite infilling

Initial rims of mafic phyllosilicates were not observed in every analyzed sample,

and instead analysis of reaction progress was based on successive levels of zeolite

infilling. Observations at the outcrop scale of Site 6 show that all vesicles exhibit

thomsonite mineralization while only larger pores progress to crystallization of mesolite.

A representative analysis of pore size distributions within the lavas (Table 3.1) shows that

pores exhibiting only thomsonite alteration have an average radius of 2.62 mm, while

those progressing to mesolite growth have an average radius of 23.76 mm.

Discussion

Nature of Porosity and Permeability in Basalts

Previous laboratory studies of unaltered lavas suggest that fluid movement through

the rock is dependant on the geometric and spatial aspects of the vesicle space such as the

aspect ratio, size, and spacing, thus implying a dependency of fluid movement on the

geometry and connectivity of vesicles within the rock. Typical porosity within high

vesicular zones of basaltic lavas, however, is highly variable, as shown by the correlation










90
80 m all pores, n=172
m clay filled, n=110
70
60
5 50
0 40
30 -
20


0
0.005 0.01 0.05 0.1 0.5 1

vesicle area (mm^2)


Figure 3.10: Data compiled from digital analysis of mesolite-scolecite zone lava from
eastern Iceland (refer to Figure 3.3). Analysis performed at the thin section
scale using ArcView and ImageJ software. Plot shows the area variation of all
vesicles within the thin section, compared to a random sampling of those
filled only with clay. Vesicles exceeding an area of 0.1 mm2 are filled with
both clay and zeolites, while those smaller than 0.1 mm2 are filled completely
with clay.

of these observations from basaltic lava samples between different regions. Observations

at the outcrop scale of basalts from the NSVG (Figure 3.1) show local areas of little to no

porosity surrounded by areas of higher porosity that display much variation in the vesicle

size and shape. Analysis of samples in thin section from the NSVG, East Greenland and

eastern Iceland (e.g., Figure 3.9) show this occurring even at the mm scale, and it is

apparent at this scale that all analyzed samples also exhibited random vesicle-vesicle

spacing with little to no interconnection of vesicles.

In spite of the high variability in vesicle size, shape and spacing, all vesicles are

partially to completely occluded due to secondary mineralization, implying that there is

some other factor controlling fluid flow through the rock and gaining access to pores.









The observed alteration aureoles (Figures 3.2, 3.3 and 3.4) suggest that fluid is

transported through the rock along a permeable network of microfractures and connected

vesicles. This implies that vesicles will experience secondary alteration regardless of

their size shape and spacing as long as they are a part of the fluid flow pathways. This

also suggests that the overall permeability of vesicular basalts will remain high even in

areas where porosity is low and vesicles are widely distributed. In these areas there will

be pathways of high permeability surrounding areas of little to no permeability. The

general highlighted region in Figure 3.5, for example, reflects a portion of the sample that

was not permeable enough to allow for diffusion of the chemical components necessary

to precipitate the secondary mineralization. This may reflect localized lowered

permeability due to the absence of the fluid flow pathways necessary for these pores to be

part of the permeable network through the lava.

Observation of alteration within the matrix shows these alteration aureoles

occurring around and between vesicles. The primary magmatic phases are highly

weathered within these aureoles signifying that these are preserving the pathways of fluid

movement through the rock during alteration. This suggests that the chemical

components necessary to precipitate the vesicle-filling minerals are derived from a

spatially restricted region (alteration aureoles) around the vesicle and the fluid pathways

over which diffusion occurs. Almost as in a hydrologic system with streams flowing in

and out of a lake, where the streams represent the pathways between vesicles and the lake

represents the vesicle, the residence time of fluid within the lake is expected to be higher

than within the stream due to its larger storage capacity. This means that the alteration









occurring in and around vesicles should be greater than those surrounding the pathways

between vesicles, which is seen in Figure 3.2.

The exact nature and initial formation of these pathways are still unclear, though

there are many possible explanations. These microfractures are interpreted to be primary

features created during formation and cooling of the lava, and not secondary openings

due to external stress. Kowallis et al. (1982) observed similar microfractures in

comparable basalts collected from the IRDP (Iceland Research Drilling Project) borehole

in Reydarfjordur, Iceland. Their observations show that the observed microfractures

formed during cooling and have since undergone secondary alteration and mineralization

due to massive amounts of fluid movement through the basalt. Termination of these

fractures appears blunt, rather than the tapered termination typical of newly formed

cracks, and the presence of delicate secondary mineral growth within these features

would not have survived the stress induced during a secondary cracking process

(Kowallis et al., 1982). They also describe fractures coincident with grain boundaries

that are often produced by thermal stressing, as well as non-coincident grain boundary

fractures produced by local strain variations. Though these fractures increase the

accessibility of the basalt matrix to alteration, they do not appear to coincide with those

observed in this study as they would not necessarily bridge vesicles and have a major

affect on permeability.

Processes experienced by a given lava flow during formation can control the total

amount of porosity and can exert a fundamental control on the nature and connectivity of

primary pore space. Lava viscosity, lava volatile content, atmospheric pressure, and the

degree of crystallinity of the lava, among other factors, all affect both the total









vesicularity of a lava flow and the vesicle size distribution (e.g., Cashman and Mangan,

1994; Sahagian and Maus, 1994; Prousevitch and Sahagian, 1996, 1998; Sahagian etal.,

2002). Srouga et al. (2004) suggests that autobrecciation processes during cooling of

rhyolitic lavas can also enhance the primary porosity within the lava. They suggest that

even though these pore spaces are small they may form a connected network, which has

potential for increasing the rate of fluid flow through the rock. This potential network,

however, is not similar to the network observed in the NSVG basalts as it would not

necessarily connect vesicles formed due to trapped gas bubbles within the rock during

cooling.

Another possible explanation of the connected network between vesicles observed

in this study is the creation of microcracks due to the elongation of bubbles during lava

flow and degassing prior to cooling (Saar and Manga, 1999; Mueller et al., 2005). The

strong presence of spherical vesicles in the high permeability zones of the cooled lava,

however, suggests that these vesicles retain gas after cooling which must later escape

through the solid rock. This escaping gas, however, will create a connected network of

pathways between vesicles and through the matrix, which can then be reused as a

pathway for fluid flow during later events of low-grade alteration. Fluid moving through

this network will have a longer residence time in the larger pore spaces, allowing for

more of the surrounding matrix to undergo interaction and alteration with the fluid. This

will cause larger vesicles to be surrounded by thicker alteration aureoles than the smaller

vesicles or thinner connecting pathways, creating alteration aureoles similar to those

observed in the NSVG basalts.









Controls on Reaction Progress

The level of reaction progress experienced by a vesicle can be related to the amount

and type of secondary mineralization observed within the vesicle, and is strongly

dependant on the vesicle size and shape. In general, larger vesicles will retain residual

pore space in lower stages of alteration, as observed in the chabazite grade zeolite

alteration present in the Greenland sample. Vesicles exposed to higher grades of

zeolitization, however, retain no residual pore space and instead the presence of later

stage mineralization is controlled by vesicle size.

There is a general trend of secondary mineralization throughout the reaction

progress from clay mineral precipitation to progressive stages of zeolite mineral

precipitation. Observations show that larger vesicles will progress to later stages of

alteration while smaller ones will become closed off earlier in the reaction progress.

Also, irregularly shaped vesicles may close off earlier in the reaction progress even with

large areas due to constriction caused by the smallest diameter.

In vesicles where clay rims are present, these rims will generally form first and

grow to a consistent thickness within a sample irrespective of vesicle size. Thus, smaller

vesicles will become completely occluded by clay minerals, while larger ones will show

later stage zeolitization. Comparison of clay rim thicknesses between the measured

samples discussed above shows that while the thicknesses within a given sample are

consistent, variation will occur between lava flows. This shows no relation to vesicle size

and instead must be due instead to differences in mineralogy of the rock and of the clay

rimming the pores, which is explored in Chapter 4.

All vesicles within basaltic lavas that are connected to the fluid flow pathways can

be analyzed and used to interpret the reaction progress experienced by the basalt during






69


secondary alteration. Interpretations must not be made, however, without close analysis

of the vesicle size and shape characteristics, as smaller vesicles and/or those with

constrictions due to irregularity in shape may not show the full extent of alteration

experienced by the lava. Assessments of the lava chemistry and alteration grade are also

necessary, as this will cause a variation in secondary mineral precipitation which will

affect interpretations of reaction progress experienced by the lava.














CHAPTER 4
MODELING OF MINERAL PARAGENESES

Introduction

The field, petrographic, and digital observations summarized in preceding chapters

illustrate several key factors involved in the coupled interactions between pore space

evolution and mineral paragenesis during low-grade alteration of vesicular lavas. First,

there is a definite crystallization sequence of secondary minerals that typically progresses

from amorphous silica to mafic phyllosilicates (chlorites and di-octahedral to tri-

octahedral smectites) to zeolites, followed in some instances by late zeolites or calcite.

Second, these minerals typically occlude primary porosity, with the degree of reaction

progress being sensitive to vesicle size. For instance, measured thicknesses of mafic

phyllosilicate rims are consistent at the thin section scale irrespective of the size,

geometry, and connectivity of vesicles. This same phenomenon is manifested by the

vesicle-size dependence of occurrence of late stage zeolites and preservation of residual

pore space. Third, a notable feature observed in these rocks is the presence of reaction

aureoles around primary pore spaces and networking microfractures. Alteration aureoles

are clearly visible as lighter areas surrounding the vesicles and microfractures (Figure

3.3). These aureoles potentially represent fluid pathways through the matrix during

alteration, suggesting that the chemical components necessary to precipitate the vesicle-

filling minerals are derived from a spatially restricted region (alteration aureoles) around

the vesicle and the fluid pathways over which diffusion occurs.









This chapter describes the results of irreversible reaction path modeling conducted

in order to interpret these features of low-grade metabasalts. These calculations are used

to 1) explain the genesis of the sequence of secondary minerals formed during alteration;

and 2) interpret the evolution of porosity during alteration. The results are then used in

concert with my field, petrographic, and digital observations to develop a geometric

model explaining the relationship between pore size, mineral authigenesis, and porosity

evolution during alteration.

Methods

Mineral paragenesis and porosity evolution during low grade alteration are

modeled using the irreversible reaction path code EQ3/6 (Wolery, 1979; Wolery et al.,

1990). The program computes the compositional evolution in systems of water-rock

interaction during reactive mass transfer assuming local and partial equilibrium between

an aqueous fluid and coexisting mineral and gas phases (Wolery, 1979). Basic inputs to

the code are the composition of the starting fluid, temperature (with pressure

corresponding to liquid-vapor equilibrium for water), and the composition and

abundances of phases that will be reacted with the fluid. In all models presented below,

the initial fluid was taken to be 1 kg of dilute aqueous fluid undersaturated with respect to

all minerals that interacts with a lava containing 20% initial porosity (or a bulk rock

volume of 4 liters).

In the first set of models, a simplified basalt composition was created as an input

for the program (Table 4.1 and 4.2) that is representative of the mineral modes, mineral

compositions, and the bulk rock composition of olivine tholeiites in the North Shore

Volcanic Group (after Brannon, 1984). The model simulates titration of 4 liters of lava

into 1 kg of a dilute aqueous solution (initially no saturated phases) in a closed system at









Table 4.1: Bulk rock composition for olivine tholeiite basalt (wt %)
Oxide Calculated Sampled* Average* Range*
SiO2 47.75 47.26 48.07 47.6 48.5
A1203 18.01 16.86 16.80 16.5- 17.4
CaO 13.67 11.03 10.93 10.6-11.2
Na20 2.26 2.51 2.48 2.38 2.64
K20 0.12 0.31 0.28 0.13 0.44
MgO 6.44 6.74 6.98 6.25 7.56
FeO 9.67 11.98 12.20 11.4 12.6
TiO2 1.83 1.67 1.71 1.43- 1.93


*Values from Brannon, 1984

Table 4.2: Modal abundances
tholeiite


and compositions of primary mineral phases in olivine


Initial Phase End-member X1 Moles2 Volume % Mass Rate3
Feldspar ________23.60 59.30 6425.79 1
Anorthite 65.00
iAlbite ]33.80
Orthoclase 1.20

Wollastonite 42.00
Enstatite 43.50
Ferrosilite 14.50
Olivine 4.30 11.00 708.31 100
Forsterite 63.90 ____ .
Fayalite 36.10 ______
Ilmenite 2.50 2.00 37931 1
Magnetite__ 1.70 1.90 393.62 1
Glass 2.60 1.70 182.21 1000
Molar percentages are represented by X. Moles are for 4 liters of basalt. Relative
reaction rates represent dissolution rates of mineral phases.

700C, Psat. This bulk composition and temperature condition is typical of the thomsonite-

mesolite zone alteration found in the NSVG. Two isochemical versions of this model

were run that differed only in the relative rates of dissolution of components into the

fluid. In the first, the lava was titrated into the fluid stoichiometrically. In the second,

the same bulk composition was distributed among separate primary phases in their

relative modal proportions and allowed to react at relative rates that approximate

differences in bulk dissolution rates (Table 4.2; Gislason and Eugster, 1987; Palandri and









Kharaka, 2004). Thermodynamic data for this model was taken from the "combined"

database provided with EQ3/6. Precipitation of epidote solid solutions, granite garnet

solid solutions, prehnite, and tremolite was suppressed in the model as these species are

not expected to form under these conditions but are predicted to be stable using the data

in the combined database.

The second model uses a calculated phase assemblage that is representative of the

mineral modes, mineral compositions and bulk rock compositions of a sampled basaltic

andesite (sample NS04-14 from sampling site 4; see Table 4.3) from the NSVG. This

sample is representative of upper zeolite facies conditions, although it is somewhat more

silicic than most of the NSVG. Because of the extensive matrix alteration of the sample

that precluded optical determination of primary mineral modes, the relative abundances

of primary minerals were calculated with the computer program KWare Magma

(Wohletz, 1999) by assuming a CIPW normative mineralogy of the unaltered basalt. The

resulting mineralogy (Table 4.4) compares favorably with previous observations of

andesites in the NSVG (Brannon, 1984; Schmidt, 1990). Representative mineral

formulas were taken from Chapter 2 along with reported values by Brannon (1984) and

Schmidt (1990). In this model, the temperature was set at 1200C (typical oflaumontite

zone alteration; cf Chapter 3), and the primary phases were titrated in their relative

modal proportions using relative rate constants that approximate differences in bulk

dissolution rates (Table 4.4; Gislason and Eugster, 1987; Palandri and Kharaka, 2004).

Due to the simpler alteration assemblage encountered in this sample, the calculations

were performed with a database based on SUPCRT92 (Johnson et al., 1992) augmented

with data for ilmenite from Stefansson (2001) and hydroxyapatite and daphnite from the













Table 4.3: Calculated anhydrous bulk composition of basaltic andesite before and after reaction.
Hydroxyl- Clino- Ortho- Whole
Ilmenite Apatite Pyroxene Pyroxene Albite Anorthite K-Feldspar Quartz Hematite Chlorite Laumontite Rock


Unaltered Andesite


SiO2
TiO2
A1203
FeO"
MgO
CaO
Na20
K20
P205
Total


0.00
2.26
0.00
2.03
0.00
0.00
0.00
0.00
0.00
4.29


0.00
0.00
0.00
0.00
0.00
0.56
0.00
0.00
0.42
0.98


1.51
0.00
0.00
0.60
0.17
0.74
0.00
0.00
0.00
3.02


14.77
0.00
0.00
11.36
3.54
0.00
0.00
0.00
0.00
29.66


23.12
0.00
6.54
0.00
0.00
0.00
3.98
0.00
0.00
33.64


4.98
0.00
4.23
0.00
0.00
2.33
0.00
0.00
0.00
11.54


6.52
0.00
1.84
0.00
0.00
0.00
0.00
1.70
0.00
10.06


6.86
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
6.86


0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00


0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00


0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00


57.67
2.26
12.62
13.98
3.72
3.62
3.98
1.70
0.45
100.00


AlteredAndesite


0.00 23.15


0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00


0.00
6.55
0.00
0.00
0.00
3.98
0.00
0.00
33.68


0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00


6.52 13.96 0.00 6.20


0.00
1.84
0.00
0.00
0.00
0.00
1.70
0.00
10.07


0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
13.96


0.00
0.00
2.90
0.00
0.00
0.00
0.00
0.00
2.90


0.00
3.51
7.42
2.77
0.00
0.00
0.00
0.00
19.90


1.87 57.67


0.00
0.79
0.00
0.00
0.44
0.00
0.00
0.00
3.10


2.26
12.69
14.13
3.78
3.82
3.98
1.70
0.45
100.48


"Total Fe reported as FeO


SiO2
TiO2
A1203
FeO"
MgO
CaO
Na20
K20
P205
Total


0.00
2.26
0.00
2.03
0.00
0.00
0.00
0.00
0.00
4.28


0.00
0.00
0.00
0.00
0.00
0.59
0.00
0.00
0.45
1.05


5.97
0.00
0.00
1.79
1.00
2.79
0.00
0.00
0.00
11.55









combined database. In addition, ideal chlorite, plagioclase, and orthopyroxene solid

solutions were defined. Andradite and grossular garnet, prehnite, annite, tremolite,

epidote, talc, analcime, and magnetite are all phases that are suppressed in the input file

of the model so that they will not form as part of the alteration sequence.

Table 4.4: Reactant phase composition, abundance, and relative dissolution rate in
basaltic andesite used as input for reaction path model.
Initial Phase End-member X1 Moles2 Volume % Mass Rate3
Feldspar ___24.49 62.39 6569.50 10
Anorthite 20.12
Albite 62.33
Orthoclase 17.55
Orthopyroxene O1 ]29.23 23.63 3527.30 100
Enstatite 35.7
Ferrosilite 64.3
Clinopyroxene 2.99 2.71 353.22 100
Diopside 100.00 .
Hedenbergite 0.00
Quartz ___ 13.58 7.70 815.86 1
Ilmenite ___3.36 2.66 510.21 100
Apatite-OH 0.23 0.91 114.72 100,000
1Molar percentages are represented by X. 2Moles are for 4 liters of basalt. 3Relative
reaction rates represent dissolution rates of mineral phases.

Results

Mineral parageneses

Figures 4.1A and 4.1B depicts the mineralogic evolution of an olivine tholeiite as a

function of reaction progress () assuming either congruent dissolution of the bulk lava

(4.1A) or dissolution controlled by the relative dissolution rates of the primary minerals

(4.1B). Note that the plots in Figure 4.1 represent the relative abundances of minerals in

terms of absolute volume. A value of log of-10 corresponds to the beginning of

reaction; a value of 0 represents completion of reaction.


















Scolecite


4500

4000

3500

3000

2500

2000

1500

1000

500

0


-10.0 -8.3 -7.3 -6.4 -5.7 -5.2 -5.0


-4.6 -4.2 -1.0 -0.7 -0.5 -0.4 -0.3 -0.2 -0.1


log 4
Figure 4.1: Mineralogic composition of lavas as a function of reaction progress for models describing A) homogeneous dissolution of
an olivine tholeiite at 70 C; B) incongruent dissolution of an olivine tholeiite controlled by relative dissolution rates of
primary minerals at 70 C; and C) incongruent dissolution of a basaltic andesite controlled by the relative dissolution rates
of primary minerals at 120 oC.


A 500ooo


HoInmo'enoLs Glass














B 4500

4000
000 "Scolecite

3500
Serpentine:
3000


U 2500
(D Feldspar

6 2000


1500


1000


500


0
-10.0 -9.0 -8.4 -7.2 -6.1 -5.7 -4.8 -3.0 -2.3 -1.7 -1.3 -1.3 -1.0 -0.7 -0.5 -0.4 -0.3 -0.2 -0.1 0.0
log

Figure 4.1: (continued)










































-6.0 -5.0 -4.0 -3.0 -2.0 -1.5 -1.4 -1.2 -1.1
log 4


Figure 4.1: (continued)









For the congruent dissolution model in Figure 4.1A, the onset of secondary mineral

precipitation occurs simultaneously for all phases. The zeolites mesolite and scolecite

comprise the largest fraction of the alteration assemblage. Clay minerals predicted by the

model consist mostly of chlorite with minor precipitation of biotite and saponite. Oxides

have the least affect on volume increase with precipitation mostly occurring as titanite,

though magnetite, rutile, hematite, and gibbsite also form. The remaining minerals are

grouped as others. Diopside dominates the volume contribution of these other minerals

during the run, with only minor amounts of clinozoisite. At the final stage of the model

the homogenous glass has dissolved completely and all volume in the system consists of

secondary minerals.

Figure 4.1B depicts the mineralogic evolution of an olivine tholeiite consisting

initially of glass, olivine, pyroxene, feldspar, ilmenite and magnetite in their modal

proportions. Each primary phase dissolves to various levels of completion at different

stages of the reaction controlled by their relative reaction rate. Contrary to the results of

the congruent dissolution model, the onset of secondary mineral precipitation varies

between phases. The complete dissolution of glass and olivine, along with the

compositional change of pyroxene from Wo0.42En0.435Fs0.145 to diopside, donates iron and

magnesium needed to form the chlorite and serpentine clay minerals relatively early in

the reaction. The complete dissolution of feldspar donates the calcium and sodium

needed to form the zeolites mesolite and scolecite. All potassium dissolved from the

feldspar remains in the aqueous phase. Ilmenite completely dissolves, donating the

titanium for precipitation of titanite. Magnetite is the only primary magmatic phase that

remains stable throughout the reaction.









Although both of the above models predict roughly the same mineral assemblage,

they differ in their predictions of the timing of mineral formation. Both models predict a

mineral assemblage dominated by the zeolites mesolite and scolecite along with pyroxene

and mafic phyllosilicate, consistent with what is observed in olivine tholeiites within the

thomsonite-mesolite zone of the NSVG. However, the model depicted in Figure 4.1B

more closely approximates the mineral parageneses observed in zeolite facies metabasalts

than that shown in Figure 4.1A in terms of the relative abundance of secondary mineral

phases and their relative sequence of precipitation. Early formation of mafic

phyllosilicate phases relative to zeolites that is observed in the relative timing of mineral

precipitates within vesicles is a consequence of the differential reaction rates of primary

mineral phases in the incongruent dissolution model.

Comparison of the results for olivine tholeiite dissolution either congruently or

incongruently (Figures 4.1A and 4.1B) illustrates the importance of relative dissolution

rates in controlling reaction paths. Alteration of the olivine tholeiite as a homogenous

glass (4.1A) shows the onset of secondary mineral precipitation to occur simultaneously

for all phases, and allows the large volume increase due to zeolitization to dominate

throughout precipitation. These results do not correlate to the petrographic observations

of altered basalts from the NSVG, indicating that the observed paragenesis of early mafic

phyllosilicates followed by zeolites can not be explained by stoichiometric reaction of the

basalt, but rather is a consequence of differential dissolution rates for primary basalt

phases as shown in Figure 4.1B, and in the incongruent dissolution of basaltic andesite in

Figure 4.1C. Consequently, modeling of the mineral paragenesis observed in the basaltic

andesite employed only an incongruent dissolution approach.









Figure 4.1C depicts the mineralogic evolution of a basaltic andesite during

interaction with dilute meteoric solutions at 120 oC. Input and output values for the

moles, volume, and mass of mineral phases are compared to calculated values and shown

in Table 4.5. The final phase assemblage listed in Table 4.5 consists of hematite,

chlorite, laumontite, K-feldspar, albite, clinopyroxene, quartz, hydroxyapatite, and

ilmenite, which is essentially the assemblage observed in sample NS04-14. The output

Table 4.5: Mole, volume and mass amounts in unaltered and altered andesite for the
computer (modeled) and mathematical (calculated) models.
Moles* Volume % Mass (g)
Unaltered Calculated Modeled Calculated Modeled Calculated Modeled


Quartz
K-Feldspar
Albite
Anorthite
Ilmenite
Apatite-OH
Clinopyroxene
Orthopyroxene
Total

Altered
Quartz
K-Feldspar
Albite
Ilmenite
Apatite-OH
Clinopyroxene
Hematite
Chlorite
Laumontite
Total


13.58
4.30
15.26
4.93
3.36
0.23
2.99
29.23
73.88

Calculated
27.50
4.28
15.20
3.34
0.25
5.90
4.29
4.07
0.92
65.76


13.58
4.30
15.26
4.93
3.36
0.23
2.99
29.23
73.88

Modeled
29.30
4.30
15.26
3.36
0.23
5.09
3.97
3.53
1.40
66.43


7.70
11.70
38.24
12.43
2.66
0.91
2.71
23.64
100.00

Calculated
14.39
10.75
35.15
2.44
0.92
9.14
3.00
19.81
4.40
100.00


7.70 815.85
11.69 1196.42
38.30 4000.77
12.41 1372.44
2.66 510.21
0.91 116.55
2.71 353.22
23.63 3527.43
100.00 11892.89

Modeled Calculated
15.48 1652.24
10.88 1192.05
35.56 3986.04
2.48 507.07
0.85 123.72
7.84 1366.98
2.80 342.73
17.34 2355.68
6.77 366.38
100.00 11892.89


815.86
1196.40
4002.50
1370.60
510.21
114.72
353.22
3527.30
11890.81

Modeled
1760.20
1196.40
4002.20
510.21
114.72
1112.52
634.10
2319.75
659.30
12309.40


*Values are for 4 liters of Basalt

generated by the model shows mass remaining relatively constant, with an increase of

only 3.5%. Mass increase largely reflects hydration of the lava. The bulk composition of

the lava is relatively unchanged other than this gain of H20, as shown by the comparison

of the initial and final bulk compositions shown in Table 4.3. Early, complete dissolution









of orthopyroxene and anorthite contributes the iron, magnesium and calcium needed to

create the hematite, chlorite and laumontite observed in the sample. Iron released early in

the reaction progress re-precipitates as hematite until enough magnesium and aluminum

are present in the aqueous phase to reach chlorite saturation. Chlorite is the modally most

important secondary mineral until laumontite saturation is reached, which is the last

mineral to form. Table 4.6 lists the compositions of mineral solids solutions present at

the end of the model. The anorthite component in plagioclase is lost during dissolution,

and all remaining plagioclase is completely albitized.

Table 4.6: Phase compositions and abundances at end of reaction path for basaltic
andesite given in their respective end-members and molar proportions (where
solid solutions exist), moles, volume %, and mass.
Final Phase End-member X Moles Volume % Mass
Feldspar 19.56 46.44 1196.40
Anorthite 0.00
Albite 78.02
Orthoclase 21.98
Clinopyroxene 5.09 7.84 1112.52
Diopside 93.07
Hedenbergite 6.93
Chlorite 3.53 17.34 2319.75
Clinochlore 35.23
Daphnite 64.77
Laumontite 1.40 6.77 659.30
Hematite 3.97 2.80 634.10
Quartz 29.30 15.48 1760.20
Ilmenite 3.36 2.48 510.21
Apatite-OH 0.23 0.85 114.72
1Molar percentages are represented by X. 2Moles are for 4 liters of basalt.

Orthopyroxene compositions are essentially constant while this phase is present in the

model. Clinopyroxene increases in modal abundance with reaction progress, and also

becomes steadily more Fe-rich. Chlorite compositions are initially clinochlore-rich, but

become more daphnite rich with increasing reaction progress.









The bulk solid composition is in excellent agreement with that calculated by

assuming that the lava completely altered to the observed mineral assemblage. The lower

half of Table 4.5 shows the modal abundances of secondary minerals determined through

a normative calculation analogous to that of the CIPW norm, but with the observed

secondary mineral assemblage in place of primary igneous phases. These calculations

indicate the creation of chlorite, laumontite, hematite, and an increase in clinopyroxene

and quartz through complete destruction of the orthopyroxene and the anorthite. All

other primary phases including the ilmenite, hydroxyl-apatite, albite, and k-feldspar show

no change in their compositions or amounts present throughout alteration.

Volume Changes During Alteration

Volume change observed during the reaction of each modeled run can be observed

both as the total volume change and as the instantaneous volume change at each step of

the reaction. The total volume change of the system with respect to reaction progress is

presented in Figure 4.1. In each run, the mass is held constant throughout the reaction to

maintain a constant initial primary porosity of -20%; therefore, any volume increase

observed in each run is a result of the formation of secondary minerals with higher molar

volumes than initial primary magmatic phases. Reaction of the olivine tholeiite shows a

volume increase of 14% at the end of reaction (Figures 4.1A and 4.1B), while reaction of

the basaltic andesite in (Figure 4.1C) shows an 8% increase in volume.

Instantaneous volume change, i.e., the net change in amount of products relative to

reactants at each step, is calculated for each step of the reaction and depicted in Figure

4.2. Negative values along the y-axis of this plot correspond to a volume decrease in the

system, which can also be explained as porosity increase. Conversely, positive values






84


correspond to a volume increase which relates to porosity destruction. In each modeled

reaction, there is an initial volume decrease in the system creating a subtle increase in

porosity before there is a more significant volume increase causing a destruction or

occlusion of primary porosity. In the congruent dissolution model for olivine tholeiite,

the positive volume change occurs at the onset of mineral precipitation, and is related to

the formation of the whole phase assemblage. For the incongruent dissolution model, the

instantaneous volume change becomes positive early due to the formation of serpentine

and chlorite and before formation of zeolites. The need for space during early

phyllosilicate formation may account for the prominent clay rims observed in many

samples. A similar trend is observed for reaction of the basaltic andesite.



40

20

C
0


E -20
0
-40

c -60


-80


-100
-10.0-8.3 -7.3 -6.4 -5.7 -5.2 -5.0 -4.6 -4.2 -1.0 -0.7 -0.5 -0.4 -0.3 -0.2 -0.1 0.0
log
Figure 4.2: Instantaneous change in total mineral volume as a function of reaction
progress. Panels A, B, and C correspond to models shown in Figure 4.1.










B 40

20

0-


E -20

S-40

S -60

-80


-100
-10.09.0-8.4-7.2-6.1-5.7-4.8-3.0-2.3-1.7-1.3-1.3-1.0-0.7-0.5-0.4-0.3-0.2-0.1 0.0
log s

C 40-

S20

u 0

E -20
0
> -40

-60
E
-80

-100
-7.8 -6.5 -6.3 -5.9 -4.6 -2.0 -1.5 -1.4 -1.2 -1.1 -1.0

log 5
Figure 4.2: (continued)














CHAPTER 5
DISCUSSION

Reaction Progress in Low-Grade Metabasalts

Petrographic observations of low-grade metabasalts (cf. Chapter 3) show a definite

progression from mafic phyllosilicates to zeolites to late stage zeolites. In some cases,

residual pore space is left over after alteration. Alteration within the matrix is dominated

by the presence ofFe(Ti) oxides, and mafic phyllosilicates along with variable amounts

of silica, feldspars albitee and microcline at higher grades), and zeolites. Clinopyroxenes

tend to survive alteration, while other mafic phases (olivine, orthopyroxene, and glass)

tend to be completely altered, often pseudomorphically. The reaction path modeling

described in Chapter 4 indicates that this is a consequence of incongruent dissolution of

the lavas. The sequence of secondary mineralization is controlled by the differential

dissolution rates for primary basaltic phases as well as the volume change throughout the

reaction due to the variation in molar volumes of secondary minerals. In the olivine

tholeiite (Figure 4.1B), the first modeled minerals to form are titanite, serpentines and

chlorite brought on by early dissolution of olivine and basaltic glass, which have the

highest dissolution rates (Table 4.2). Figure 4.2B shows an instantaneous volume

increase of 17% related to the high molar volumes of the serpentines and chlorites. This

positive volume change forces precipitation to progress from the matrix to the vesicle in

search of available pore space, forming observed clay rims lining vesicle walls.

The onset of feldspar and pyroxene dissolution as the alteration progresses leads to

zeolite precipitation. There is still precipitation of the early forming minerals, but the









system becomes dominated by these zeolites due to their large molar volumes compared

to both the primary basaltic phases and the mafic phyllosilicates. They create a final

instantaneous volume increase of 14% (Figure 4.2B), causing occlusion of any residual

pore space within vesicles. The model predicts mesolite to form first with late stage

formation of scolecite, which is a common sequence observed in samples exhibiting

mesolite-scolecite zone alteration (pers. com.; Neuhoff, P.S., 2006).

A similar progression of secondary mineralization is seen in the modeled alteration

of basaltic andesite (Figure 4.1C), where hematite, smectite and chlorite are the first

minerals to form with later progression to the zeolite laumontite. The early dissolution of

orthopyroxene and anorthite due to their high dissolution rates (Table 4.4) leads to the

initial precipitation ofFe(Ti) oxides and mafic phyllosilicates. The small molar volume

of hematite creates no volume increase during precipitation, allowing precipitation to

occur in situ without progressing to open vesicle space. As the levels of magnesium and

aluminum dissolved into the aqueous solution increase (Table 4.3), clay mineral

precipitation dominates the system. The positive instantaneous volume change reaching

14.41% at this stage in the reaction (Figure 4.2C) is then due to the high molar volumes

of the chlorite and smectite, causing precipitation to migrate into open pore space and

form the observed chlorite rims lining vesicle walls in these samples (Chapter 3). At this

stage in the reaction, the model also shows an increase in the amount of quartz in the

system, which is observed petrographically as sporadic appearances of silica within the

matrix or occasional silica rims observed lining pores.

As the reaction progresses, the aqueous phase acquires enough calcium to become

saturated with respect to the zeolite laumontite. This shift from clay minerals to zeolites









appears to be sudden due to the observed consistency in clay rim thicknesses. In other

samples of metabasalts where no clay rims line vesicle walls, this shift may have

occurred before significant volume change due to clay precipitation, whereas very thick

clay rims would suggest a large increase in volume prior to zeolite formation. Once this

shift occurs, one or more generations of zeolites will fill any remaining pore space in the

basalt, which is usually within or immediately surrounding vesicles. In the modeled

basaltic andesite, the precipitation of laumontite causes an instantaneous volume increase

of 8.84% (Figure 4.2C) and fills remaining pore space in vesicles. Smaller vesicles

which have been closed during earlier stages of alteration will not experience

zeolitization, and the relative amount of zeolite crystallization in a given vesicle will

increase with vesicle size due to the consistency in clay rim thicknesses.

Dependence of Reaction Progress on Pore Size

Observations described in preceding chapters suggest that the alteration observed in

and around vesicles in mafic lavas occurs through fluid-rock interaction at the scale of

individual pores. Most notably, the prominent vesicle size dependence on reaction

progress noted in Chapter 3 indicates that chemical components are not homogenized

across the lava during alteration, but rather that each vesicle independently exchanges

chemical components with the surrounding lava matrix. This is most prominently

exhibited by the constancy of mafic phyllosilicate rim thicknesses. In addition, the

variations in zeolite mineralogy observed between pores in low-grade metabasalt samples

observed in previous studies (e.g., Neuhoff et al., 1999; 2006) suggest that local

interactions between fluids within vesicles and the surrounding matrix are important.

This is not to say that there is not chemical communication between pores, as the

relatively high permeabilities of vesicular lavas noted above requires that advective-