| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Jones & Stokes, 2600 V Street, Sacramento, CA 95818 USA
b Soil & Water Science Program, Dep. of Environ. Sci., Univ. of California, Riverside, CA 92521-0424 USA
graham{at}citrus.ucr.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and methodsResultsDiscussionSummaryREFERENCES |
|---|
Abbreviations: COLE, coefficient of linear extensibility • Fed, citrate-bicarbonate-dithionite extractable Fe • OC, organic C
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionSummaryREFERENCES |
|---|
20% of California's land area, concentrated primarily in the Sierra Nevada and the southern California Peninsular Mountain Ranges (Norris and Webb, 1990). Where unaffected by Pleistocene glaciation, granitic bedrock is often weathered to depths of several meters, while overlying soils are comparatively thin (Wahrhaftig, 1965; Graham et al., 1997). Because of its widespread occurrence, not only in California, but throughout much of the world (Gerrard, 1994), weathered granitic bedrock is an important soil parent material. As such, its properties and the processes by which it transforms to soil are of interest from a pedological standpoint. Because weathered bedrock often dominates granitic regoliths, an understanding of its functions in natural systems is of interest relative to ecology, hydrology, environmental quality, and engineering as well. The roles of weathered granitic bedrock in soil genesis, water movement, and natural ecosystems have been studied worldwide (Isherwood and Street, 1976; Koppi and Williams, 1980; Megahan and Clayton, 1986; Evans and Bothner, 1993; Graham et al., 1994). Several studies have focused on granitic bedrock in the western USA. Nettleton et al. (1968, 1970) studied selected chemical, physical, mineralogical, and morphological properties of weathered granitic bedrock and overlying soils on a toposequence in the southern California Peninsular Ranges and used their data to determine processes of bedrock weathering and soil formation. In the Idaho batholith, Clayton and Arnold (1972), Clayton et al. (1979), and Clayton (1993) characterized and discussed various properties and processes in granitic bedrocks during different stages of weathering and developed a practical classification scheme. Weathered granitic bedrock in the San Jacinto Mountains in southern California has been shown to possess abundant lithogenic and pedogenic porosity, and to conduct water relatively rapidly in comparison with overlying soils (Johnson-Maynard et al., 1994; Graham et al., 1997). The ecological significance of weathered granitic bedrock in southern California and the southern Sierra Nevada has been demonstrated by studies showing that it serves as a rooting medium and as a source of plant available water for forest and chaparral vegetation (Arkley, 1981; Jones and Graham, 1993; Anderson et al., 1995; Sternberg et al., 1996; Hubbert, 1999).
The granitic bedrocks described in the previously cited studies, as with most granitic bedrocks, possess joint fractures produced by the release of geostatic and tectonic forces over time (Ollier, 1969; Twidale, 1971). Joint fractures facilitate the ingress of weathering agents such as water and plant roots, influence patterns of weathering and clay illuviation, and affect landform evolution (Ollier, 1965; Twidale, 1971; Clayton and Arnold, 1972; Graham et al., 1994; Ollier and Pain, 1996); however, the influence of fractures on specific pedogenic processes and their role in the transformation of granitic bedrock to soil have not been thoroughly studied.
The objectives of this study were (i) to characterize various chemical, physical, morphologic, and micromorphologic properties of a weathered granitic bedrock in southern California, with an emphasis on properties near joint fractures, and (ii) to use the data collected during this process to interpret dominant pedogenic processes and the overall role of joint fractures in the transformation of the fractured granitic bedrock to soil.
|
Materials and methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionSummaryREFERENCES |
|---|
8 km south of the town of Banning. The San Jacinto mountains are part of the larger southern California Peninsular Range complex (Norris and Webb, 1990), and are composed mostly of plutonic igneous rocks with lithologies ranging from mafic tonalite to K-feldspar-poor granodiorite. Compositional variation is minimal and ranges from 50 to 55% plagioclase (An30-40), 20 to 30% quartz, 1 to 8% K-feldspar, 10 to 15% biotite, and 0 to 5% hornblende, with mineral grain size usually <10 mm (Hill, 1988). The area is characterized by a Mediterranean climate, with precipitation occurring mostly as rain between the months of November and April. The closest weather stations are located in Idyllwild (20 km southeast, elevation 1640 m), which receives 645 mm of precipitation annually, and in Beaumont (12 km northwest, elevation 792), which receives 432 mm of precipitation annually (National Oceanic and Atmospheric Administration, 1992). Soils in the area consist of moderately deep to shallow Entisols, Alfisols, and Mollisols on gently rolling to steep hillslopes (Knecht, 1971; Graham et al., 1997). Vegetation consists of chaparral species dominated by chamise (Adenostoma fasciculatum Hook. & Arn.), Eastwood's manzanita (Arctostaphylos glandulosa Eastw.), and cupleaf ceanothus [Ceanothus greggii A. Gray var. perplexans (Trel.) Jep.]. The site investigated is on a nearly level ridge top. Soils are loamy, mixed, active, mesic, shallow Typic Haploxeralfs underlaid by >1 m of weathered granitic bedrock. The regolith materials are moderately acid, with pH values ranging from 4.95 in the uppermost soil horizon to 5.76 in the deepest weathered bedrock horizon studied (1:1 soil/water) (Frazier, 1997). As reported by Frazier (1997), the primary mineral composition of the bedrock at a depth of 1.25 m is 40% plagioclase feldspar, 21% quartz, 20% mica (mostly biotite), 13% K-feldspar, and 5% hornblende. The fine sand (100–250 µm) and medium silt (5–20 µm) fractions of the weathered bedrock horizons above a depth of 1.25 m contain quartz, plagioclase, hornblende, and biotite, and the secondary minerals kaolin, vermiculite, and regularly and randomly interstratified micavermiculite. The clay fractions (<2 µm) of the weathered bedrock and overlying soil horizons contain no primary minerals, and secondary minerals are identical to those contained in the coarser size fractions, with the addition of trace quantities of goethite.
Less mafic tonalites (<1% hornblende, 10% biotite) occur in the general study area, but produce soil and weathered bedrock morphologies that differ from those described in this study. In particular, much less illuvial clay is generated (Graham et al., 1997).
Field Methods
A trench 1.25 m deep, 2.25 m long, and 0.75 m wide was excavated using an electric chipping hammer and hand implements. The weathered bedrock at the study site and throughout the general area was extensively fractured. The widest and longest fractures exposed on the trench walls were vertically oriented joints — hereafter referred to as joint fractures — which were usually continuous from the soil–weathered bedrock interface to the lowermost exposed bedrock (1.25 m).
Alteration of the original bedrock material has produced a vertical sequence of weathered bedrock horizons and a series of morphologic zones emanating laterally from the fractures. Based on these observations, each weathered bedrock horizon was divided into four zones for description and sampling: the matrix, fracture rind, fracture coating, and fracture fill (Fig. 1)
. For the purposes of sampling, the matrix was considered to be the material located >2 cm from the fracture sidewalls. The fracture rind consisted of more intensely altered bedrock material that extended from the fracture sidewall to 
2 cm into the matrix. The fracture coating was a thin (<0.1 cm) layer of material covering the surface of the fracture rind. The fill material was contained within the fracture itself, and in most cases, did not completely fill the fractures. An additional zone, the matrix-rind zone, was not identified in the field, but was discovered and described during micromorphological examination (Fig. 1). This zone had a variable, undetermined width and consisted of matrix material immediately adjacent to the fracture rind. It was less altered than material in the fracture rind, but more altered than matrix material located farther from joint fractures.
|
10 to 15 cm in diam., were collected from each horizon in the region of the bedrock fractures using the sampling scheme shown in Fig. 1. Each sample contained the complete lateral morphologic zonation of matrix through fracture fill. Additional bulk and intact samples were collected from the weathered bedrock matrix at positions 10 to 20 cm from the fracture sidewalls (Fig. 1), and used for all analyses of the matrix materials. Bulk and, when possible, intact samples were also collected from the overlying soil horizons.
Laboratory Methods
Materials from the four different fracture morphologic zones were separated from the intact sample blocks using a metal spatula. Soil and weathered bedrock samples were air dried, crushed, and sieved to remove coarse fragments (>2 mm). Unless otherwise noted, all subsequent analyses were performed on the fine earth fraction (<2 mm). Particle–size distribution was determined by the pipette method after organic matter removal using H2O2 (Gee and Bauder, 1986). Total C was measured with a Carlo Erba NA1500 C/N/S analyzer (Fisons Instruments, Dearborn, MI) by measuring CO2 evolved during dry combustion at a temperature of 1029°C (Nelson and Sommers, 1982). Total C was assumed equal to organic C (OC) because carbonates were not detected in the soil or weathered bedrock. Bulk density was measured on intact samples using the paraffin-coated clod method (Blake and Hartge, 1986b). Particle density was measured using an Accupyc 1330 gas pycnometer (Micromeritics Instr. Corp., Norcross, GA) (Blake and Hartge, 1986a). Total porosity was calculated from bulk density and particle density. Samples were extracted with citrate-bicarbonate-dithionite to determine quantities of pedogenic Fe (Jackson et al., 1986). Soil and weathered bedrock color were measured with a Minolta CR-200 Chroma Meter (Minolta Corp., Ramsey, NJ). Coefficient of linear extensibility (COLE) values were determined for the weathered bedrock matrix from the bulk density of Saran-coated clods desorbed to -10 KPa and oven dryness (Method 4D; Soil Survey Staff, 1984). Data from the particle size, bulk density, OC, and Fed procedures were analyzed statistically using two-factor ANOVA, with the factors being horizon (depth) and morphologic zone. Multiple comparisons were made using Tukey's least significant difference procedure (Data Desk, 1995).
Thin sections (25 by 45 mm) were prepared from intact samples of the matrix and fracture-rind zones from each of the four weathered bedrock horizons. Thin sections were viewed with a petrographic microscope and described according to the terminology of Bullock et al. (1985). When referring to illuvial clay features, the terminology of Brewer (1976) was also used. Point counting procedures were used to estimate the volume percentages of channels and planar voids with diameters 
0.1 mm (i.e., macropores; Vepraskas et al., 1991), and void coatings and infillings. Estimates were based on a total of 300 counts per morphologic zone (spread equally across three slides).
|
Results |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionSummaryREFERENCES |
|---|
5 cm from fracture sidewalls.
|
0.4 cm in the Cr3 horizon (Table 1). These measurements were made on the trench face after it had been exposed to temperatures of 20°C for 10 h. The width of the fracture rind also decreased with increasing depth, from 0.6 to 2.2 cm in the BCrt horizon to 0.5 cm in the Cr3 horizon. At most depths, the rind was distinguished from the adjacent matrix largely by a contrast in color (Table 1). The color contrast, and thus the boundary between the two zones, was abrupt at all depths, but was much less pronounced in the BCrt horizon than in the Cr horizons. At all depths, roots contained within the fracture void space were concentrated along the fracture sidewalls. Few of these roots extended laterally into the adjacent fracture-rind and matrix zones. Fracture-coating materials in the BCrt, Cr1, and Cr2 horizons closely resembled pressure faces commonly found on the ped surfaces of clayey subsoils. Many of the sand-sized mica particles in these coating materials exhibited parallel-referred orientation patterns relative to the surfaces of the fracture rind, giving the rind surface a shiny appearance. Coating materials in the Cr3 horizon, and sometimes those in the Cr2 horizon, had a dull, powdery appearance, and more closely resembled illuvial silt films than pressure faces.
Fracture-fill materials were found in the void space of all four fractures studied. They were usually structureless with a loose or soft consistency, but occasionally had a compact platy structure oriented parallel to the vertical axes of the joint fractures. In most cases, the fill materials adhered somewhat to the fracture sidewalls.
Qualitative Micromorphology
Weathered Bedrock Matrix
The fabric of the matrix in all four weathered bedrock horizons was rock controlled, in that the distribution and orientation of mineral grains and void spaces was largely controlled by structural and textural properties of the bedrock (Stolt et al., 1991). Chemical mineral weathering increased upward in the profile. Dissolution pitting of feldspar minerals occurred as linear alterations along twin planes and as complex alteration patterns spread randomly across grain surfaces (Fig. 2a)
. The dissolution pits contained secondary weathering products, including Fe oxides, which imparted a dirty appearance on grain surfaces (Fig. 2a). Parallel linear alteration of biotite grains was apparent at all depths studied (Fig. 2a). More advanced stages of biotite weathering, as indicated by yellower grain colors and decreased birefringence and pleochroism, become more apparent upward in the profile. Hornblende alteration was weak and was apparent only from accretions of Fe oxides along cleavage planes and in pores contiguous with grain boundaries. Quartz grains appeared chemically unaltered, but exhibited abundant intragranular microcracks throughout the profile (Fig. 2a).
|
Coatings and infillings of mostly clay-sized weathering products, including Fe oxides, were present on the surfaces of channels and planar voids at all depths studied (void ferriargillans of Brewer, 1976) (Fig. 2a and 2b). They exhibited banded extinction patterns under cross polarizers, which indicated weak-to-strong parallel-referred orientation patterns relative to void surfaces and suggested an illuvial origin. Void coatings and infillings in the Cr horizons usually had a hue of 10YR, while those in the BCrt horizon had a hue of 7.5YR or redder.
Fragmented, sand-sized mineral grains occurred regularly throughout the matrix. The abundance of fragmented mineral grains and the degree of mineral grain fragmentation generally increased upward in the profile. Fragmentation occurred mainly along planes of weakness corresponding to infilled planar voids, an example of which is shown in Fig. 2a. The group of quartz grains in the center of the photomicrograph had a clustered distribution pattern in which each grain in the cluster had boundaries that appeared to have once accommodated adjacent grains. Grain fragmentation patterns such as these typically occur in areas of the matrix where the rock-controlled fabric has been locally disrupted.
Matrix-Rind Zone
When viewed in thin section, the region of the matrix within 1 to 3 cm of the rind had rock-controlled fabric, but had features that contrasted slightly with those in the remainder of the matrix. Because of its position, it was termed the matrix-rind zone. It was evident only under magnification, and was not sampled for analysis of chemical and physical properties.
The degree of chemical mineral weathering in the matrix-rind zone was greater than in the rest of the matrix, especially with regard to the degree of feldspar pitting. The most distinguishing characteristics of the matrix-rind zone, however, were that the abundance of fragmented mineral grains and the degree of mineral grain fragmentation were greater than in the adjacent matrix. As in the matrix, fragmented mineral grains were separated mainly along the axes of infilled planar voids and microcracks (Fig. 2c).
Fracture-Rind Zone
In the Cr3 horizon, the rind fabric was typically rock controlled and therefore very similar to the fabric of the adjacent matrix-rind zone; however, fragmented mineral grains were more abundant and the degree of grain fragmentation was more advanced in the rind than in the matrix-rind zone.
In the Cr2 and Cr1 horizons, the rind had gefuric and single-spaced porphyric c/f-related distribution patterns and contained more fine materials than in the Cr3 horizon (Fig. 2c). It lacked a definite microstructure, but the parallel-referred orientation pattern of the platy mica grains and elongate quartz and feldspar grains relative to the joint fracture sidewalls gave the rind in the Cr2 and Cr1 horizons a somewhat laminated appearance (Fig. 2c). Platy mica particles with "bow-like" or contorted configurations (Fig. 2c and 2d) were common throughout the rind in the Cr2 and Cr1 horizons, as were granostriated clay and mica pseudomorphs at the boundaries of sand-sized quartz and feldspar grains. The boundary between the rind and matrix-rind zone in both Cr2 and Cr1 horizons was sharp and prominent (at 20x magnification; Bullock et al., 1985), and was usually <1 mm wide. The boundary was sharpest where it ran along quartz grains, and was more diffuse where it abutted highly weathered feldspar, mica, and hornblende grains (Fig. 2c).
In the BCrt horizon, the rind exhibited two types of fabric. In the first, which was more common, the c/f-related distribution pattern and the referred orientation pattern of platy and elongate mineral grains were similar or identical to those described above for the rind in the Cr2 and Cr1 horizons (Fig. 2e). In the second, the c/f-related distribution pattern was also gefuric or single-spaced porphyric, but crack and fissure microstructures were apparent and the parallel-referred orientation patterns of sand-sized mica, quartz, and feldspar grains, which was strongly expressed in the first type of fabric, was weak or nonexistent (Fig. 2f). In some cases, the crack and fissure microstructures appeared together with weakly developed subangular blocky microstructure. In samples with this second type of fabric, the boundary between the rind and matrix rind was usually faint and diffuse (at 20x magnification).
Quantitative Micromorphology
Macroporosity
In the matrix, macroporosity increased upward in the profile (Fig. 3a)
. Planar voids composed most of the macropore space in the matrix of the Cr horizons, whereas root channels composed most of the macropore space in the matrix of the BCrt horizon. In the matrix-rind zone, macroporosity was lower than that in the matrix in the BCrt, Cr1, and Cr2 horizons, and was slightly greater than that in the matrix in the Cr3 horizon (Fig. 3a and 3b). In the fracture rind, macroporosity was about the same as, or slightly higher, than in the matrix rind, and consisted mostly of root channels (Fig. 3b and 3c).
|
Physical and Chemical Properties
Particle-Size Distribution
In the matrix, total clay content was highest in the BCrt horizon and decreased with increasing depth to a minimum in the Cr3 horizon (Table 2)
. In all four weathered bedrock horizons, total clay content in the fracture rind was higher than that in the matrix. In the BCrt and Cr1 horizons, total clay contents were highest in the fracture rind, whereas in the Cr2 and Cr3 horizons, they were highest in the fracture coating. Fine clay contents showed vertical and lateral trends identical to those for total clay (Table 2).
|
|
|
|
|
|
Discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionSummaryREFERENCES |
|---|
Clay Illuviation and Macroporosity
The presence of illuvial void coatings and infillings throughout the matrix indicates that clay translocation is not limited to fractures. Water flow through the matrix also redistributes clay in suspension. Depth functions of total and fine clay (Table 2) and void ferriargillans (Fig. 3d) indicate that the BCrt is the main illuvial horizon. Comparatively lower total and fine clay contents of the A and AB horizons show them to be eluvial and a source of clay for the BCrt horizon. The predominance of rock-controlled structure in the matrix suggests that the BCrt horizon attains its argillic character before it undergoes significant structural development. Thus, illuvial processes involved in the formation of the BCrt are facilitated largely by pedogenic porosity generated through weathering and root pedoturbation, and not by structural, interpedal porosity, as is the case in many soils.
In the matrix, quantities of void infillings and planar macrovoids are inversely related as a function of depth (Fig. 3a and 3d). These data suggest that planar macrovoids in the rock fabric become progressively clogged with illuvial clay over time, rendering them less effective for water flow, a phenomenon also observed in saprolites of the southeastern USA (Vepraskas et al., 1991; Williams et al., 1994; Vepraskas et al., 1996). Consequently, root channels become increasingly important as conduits for clay translocation in the upper regions of the profile as the BCrt horizon develops. Root channels, which compose >90% of the macrovoid volume in the BCrt matrix (Fig. 3a), also accumulate illuvial clay, but root growth creates new channels or reopens old ones (Fig. 2b), so they facilitate clay translocation when other types of voids have been clogged.
The color and elemental composition of void argillans have been used to determine the origins of illuvial clay in granitic bedrock (Nettleton et al., 1968; Chartres and Walker, 1988; Graham and Franco-Vìzcaino, 1992). Chartres and Walker (1988) found that the elemental composition of illuvial clay contained in deep weathered bedrock horizons differed from that of the soil clay and illuvial clay in shallower bedrock horizons, suggesting that the deeper clay deposits had been translocated from sources other than the soil. Similar conclusions were made in this study based on color rather than elemental composition. Ferriargillans in the BCrt matrix have redder hues than ferriargillans in the matrices of underlying Cr horizons, suggesting that clay translocation occurred from different sources in the regolith. Most of the illuvial clay in the BCrt matrix is probably derived from the overlying soil, while illuvial clay in the underlying Cr horizons appears to originate largely from localized translocation of clay generated from in situ weathering of biotite, hornblende, and plagioclase feldspars. These minerals are likely sources of illuvial clay during early stages of pedogenesis, as they are relatively easily weathered and abundant in granitic rock. The colors indicate a lower Fe oxide content and reflect the lesser degree of weathering in the Cr horizons.
In all four weathered bedrock horizons, clay contents are higher and the ratio of fine clay/total clay is higher in the fracture rind than in the adjacent matrix (Table 2), suggesting that the rind is relatively enriched with illuvial clay. As with weathering, illuvial enrichment of the rind appears to be facilitated by the movement of water down joint fractures. The most probable source of illuvial clay in the rind is the overlying soil, which generally has lower clay contents and lower fine/total clay ratios than the rind (Table 2).
Mineral Grain Fragmentation and Localized Fabric Disruption
Micromorphologic observations and gravel/(sand + silt) ratios (Table 3) indicate that the abundance of fragmented mineral grains and the degree of mineral grain fragmentation increase upward in the profile. The increase in grain fragmentation coincides with an increase in the abundance and thickness of void ferriargillans (Fig. 3d). Mineral grains separated by planar voids that are infilled with illuvial clay often exhibit clustered-related distribution patterns with boundaries that accommodate adjacent grains in the clusters (Fig. 2a). These features are usually associated with areas of locally disrupted rock-controlled fabric, and become more abundant upward in the profile. Thus, we speculate that shrink–swell activity of illuviated clay within planar voids and microcracks plays a major role in grain fragmentation and localized disruption of rock-controlled fabric. Rock-controlled fabric predominates throughout most of the matrix, however, indicating that shrink–swell activity generally has not been sufficient to induce the development of soil structure, except in parts of the BCrt matrix where weak secondary blocky structure exists. Similar observations and conclusions as to the role of illuvial clay in physical weathering of granitic bedrock have been made in southern California (Nettleton et al., 1968; Graham et al., 1994) and in southeastern Australia (Dixon and Young, 1981; Chartres and Walker, 1988).
Illuvial clay content and grain fragmentation trends similar to those that occur with depth also occur laterally from the rind to the matrix. In the Cr2 and Cr3 horizons, void ferriargillans (Fig. 3d and 3e) and fragmented mineral grains are more abundant, and the degree of grain fragmentation and localized fabric disruption are more advanced, in the matrix-rind zone than in the matrix. We interpret that the higher illuvial clay contents are responsible for the greater degree of grain fragmentation and fabric disruption in the matrix-rind zone. In the BCrt and Cr1 horizons, fragmented mineral grains and the degree of mineral grain fragmentation follow the same trends as described above, but illuvial clay contents in the matrix rind and matrix of these horizons are similar (Fig. 3d and 3e). These data and observations suggest that roots play a greater role in fabric disruption in these horizons than in the underlying Cr2 and Cr3 horizons where roots are generally restricted to the fracture rind and the fracture void space (Table 1).
Fabric Reorganization
Except for the Cr3 horizon, the rock-controlled fabric that predominates in the weathered bedrock matrix and matrix-rind zones does not occur within the fracture rind. The common and distinguishing features of the rind fabric in the Cr2, Cr1, and BCrt horizons, such as the parallel-referred orientation pattern of platy and elongate mineral grains (Fig. 2c and 2e), contorted mica pseudomorphs (Fig. 2c and 2d), and the granostriation of clay and mica pseudomorphs around quartz and feldspar grains, suggest that the fracture rind in these horizons has been laterally compressed. We hypothesize that swelling of pedogenic clay in the weathered bedrock causes joint fractures to close upon wetting, and that the resulting compressive and shear stresses are responsible for forming the fabric exhibited in the rind in the BCrt, Cr1, and Cr2 horizons. Although the mineral structures of vermiculite and kaolinite, the two main clay minerals present in the weathered bedrock (Frazier, 1997), expand only very slightly upon wetting, shrink–swell is largely the result of the loss and gain of water in the microporous network formed by linked phyllosilicate crystallites, rather than interlayer expansion (Wilding and Tessier, 1988). Measured COLE values indicate that there is sufficient clay present in the weathered bedrock matrix to cause as much as a 3% increase in volume because of swelling (Table 4). Additionally, the aperture of joint fractures in a nearby (0.5 m) weathered granitic bedrock decreased considerably, to the point of total closure at some depths, when wetted with an aqueous tracer suspension (Frazier, 1997). Apparently, the rock-controlled fabric in the matrix contains enough clay to expand upon wetting, but is sufficiently rigid to compress the rind material rather than deform when fracture sidewalls converge.
Bulk density trends provide further evidence that lateral compression of the rind has taken place (Table 4). Studies have shown that the bulk density of granitic rock decreases as it weathers (Nettleton et al., 1970; Clayton et al., 1979; Chartres and Walker, 1988; Graham et al., 1997). We found this to be true for the bedrock matrix, in which bulk density decreases upward from the Cr3 to the BCrt horizon. Because weathering also increases towards joint fractures, a similar decrease in bulk density would be expected from the matrix to the fracture rind. While this is the case in the Cr2 and Cr3 horizons, the bulk density of the fracture rind in the BCrt and Cr1 horizons is slightly greater than that in the adjacent matrix. Thus, we conclude that shrink–swell processes have compacted the rind materials in these most weathered, clay-enriched horizons.
Compaction may explain the relative paucity of void ferriargillans in the Cr1 and Cr2 rinds (Fig. 3f), despite the fact that they have higher clay contents and higher fine/total clay ratios than the adjacent matrix (Table 2). Apparently void ferriargillans are being disrupted faster than they are created in these horizons. The opposite appears to be true in the BCrt rind, where void ferriargillans are more abundant than in the adjacent matrix (Fig. 3d and 3f). Apparently, the rate of ferriargillan formation in this horizon exceeds the rate of disruption.
Lateral compression may also explain why roots rarely extend from joint fractures laterally into the rind of the BCrt, Cr1, and Cr2 horizons. Root penetration could be limited by the dense, compacted nature and vertically trending fabric of the rind in these horizons. Based on the abundance of vertically trending root channels in the rind (Fig. 2e), however, it appears that root growth downward through the rind is not restricted.
In the Cr3 horizon, the rind fabric does not differ significantly from that in the adjacent matrix-rind zone. Additionally, macroporosity, total porosity, and the volume of void ferriargillans are greater in the rind than in the adjacent matrix-rind zone (Fig. 3 and Table 4). These data and observations suggest that the Cr3 rind has not undergone lateral compression and compaction, and that the matrix-rind zone develops along the margins of joint fractures prior to fabric reorganization.
In weathered bedrock horizons where laterally compressed rind fabrics exist (i.e., BCrt, Cr1, and Cr2 horizons), there appears to be very little mixing of the material in the rind and matrix rind, as the boundary between the two zones is usually sharp and well defined (at 20x magnification; Bullock et al., 1985). Therefore, we deduced that thickening of the rind fabric occurs via two mechanisms. First, more resistant grains at the matrix rind–rind boundary (Fig. 2c) become dislodged and incorporated into the rind fabric by shrink–swell generated stresses when fracture sidewalls converge. Localized shrink–swell activity of illuvial clay within intra- and intergranular planar voids (Fig. 2c) may loosen grains from the rock-controlled matrix rind fabric, allowing them to be more easily dislodged and incorporated into the rind fabric when fracture sidewalls converge. Second, grains weakened by chemical weathering at the matrix rind/rind boundary may be broken and incorporated into the rind by similar stresses. Fragmentation and subsequent incorporation of grains into the rind via these mechanisms is probably responsible for the lower gravel/(sand + silt) ratios in the rind relative to the matrix (Table 3).
Structural Development
In the BCrt horizon, some fracture-rind samples have crack or fissure microstructure, usually together with incipient subangular blocky microstructure (Fig. 2f). The microstructural units are bounded by what appear to be desiccation cracks, attesting to the important role of shrink–swell processes in the development of subsoil microstructure. We speculate that formation of the laterally compressed, vertically trending fabric that occurs in most samples of the BCrt rind (Fig. 2e) precedes the development of crack, fissure, and subangular blocky microstructures on the basis that the rock-controlled fabric that initially exists along joint fracture margins would need to be disrupted before substantial microstructural development could begin. The matrix rind–rind boundary is less pronounced (at 20x magnification) where the BCrt rind has crack and subangular blocky microstructure, suggesting that fabric reorganization and microstructural development proceeds inward from the rind, following the weathering front and water infiltration pathways. Weak secondary soil structure develops concurrently at the soil–weathered bedrock interface (Table 1) and proceeds downward following the vertical weathering front and water infiltration pathways.
Fracture Coating
The ratio of fine clay/total clay is higher in the fracture coating than in overlying soil horizons (Table 2), suggesting that the fracture coatings are illuvial. In the BCrt and Cr1 horizons, both fine and total clay contents are lower in the fracture coating than in the rind (Table 2). The opposite relationship exists in the Cr2 and Cr3 horizons. Additionally, the fine/total clay ratio of the fracture coating in the BCrt horizon is lower than that for fracture coatings in deeper horizons. These data suggest that clay has been eluviated from the rind surface in the BCrt and Cr1 horizons and deposited as a coating on rind surfaces in the underlying Cr2 and Cr3 horizons. The illuvial nature of the rind surface in the lower profile is especially apparent in the Cr3 horizon where the clay content in the fracture coating is more than double that in the matrix (Table 2).
The smoothed fracture coatings in the upper part of the profile appear to result from the same compressive forces that reorganized the fabric in the rind. Thoma et al. (1992) found that mineralized coatings on the surfaces of fractures in weathered tuff had a significant effect on the migration of water between fractures and the surrounding matrix. In this study, the presence of smoothed coatings, or pressure faces, in the upper part of the profile, as well as the vertically trending fabric and low lateral pore continuity of the rind zones, may inhibit the flux of water into the rind and surrounding matrix in an analogous manner. As a result, water may move relatively rapidly and frequently across the BCrt rind surfaces, eluviating clay in the process.
Fracture Fill
The data collected during this study do not conclusively indicate the origin of the fracture-fill materials. Fill materials in the BCrt and Cr1 horizons have particle–size distributions similar to those for overlying soils (Table 2), suggesting that they may have sloughed into joint fractures from above. Conversely, fill materials in the Cr2 and Cr3 horizon have particle–size distributions that more closely resemble those in the adjacent matrix, suggesting that they may be derived from the sloughing of fracture sidewalls. The vertical platy structure of some fill materials is consistent with the lateral compression hypothesis invoked to explain fabric reorganization in the rind.
|
Summary |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionSummaryREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
Received for publication March 22, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionSummaryREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. Vargas and M. F. Allen Dynamics of Fine Root, Fungal Rhizomorphs, and Soil Respiration in a Mixed Temperate Forest: Integrating Sensors and Observations Vadose Zone J., August 13, 2008; 7(3): 1055 - 1064. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. E. Allen Physical and Chemical Characteristics of Soils Forming on Boulder Tops, Karkevagge, Sweden Soil Sci. Soc. Am. J., January 1, 2005; 69(1): 148 - 158. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. J. Kendrick and R. C. Graham Pedogenic Silica Accumulation in Chronosequence Soils, Southern California Soil Sci. Soc. Am. J., July 1, 2004; 68(4): 1295 - 1303. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Assessing Changes in Elemental Mass as a Result of Chemical Weathering of Granodiorite in a Mediterranean (Hot Summer) Climate Journal of Sedimentary Research, May 1, 2003; 73(3): 434 - 443.
|
|
|
|
|
|
C. S. Frazier, C. S. Frazier, R. C. Graham, P. J. Shouse, M. V. Yates, and M. A. Anderson A Field Study of Water Flow and Virus Transport in Weathered Granitic Bedrock Vadose Zone J., August 1, 2002; 1(1): 113 - 124. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. R. Hubbert, R. C. Graham, and M. A. Anderson Soil and Weathered Bedrock: Components of a Jeffrey Pine Plantation Substrate Soil Sci. Soc. Am. J., July 1, 2001; 65(4): 1255 - 1262. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||
Error bars are not visible on scale shown. 
n = 1, and therefore standard deviation could not be calculated