Vertical Magnetic Structure of Ocean Crust Determined From Near-bottom Magnetic Field Measurements

Maurice A. Tivey

Dept. of Geology and Geophysics, Woods Hole Oceanographic Institution, Woods Hole, MA

Abstract. Concensus on the source of marine magnetic anomaly stripes has remained one of the most elusive aspects of the Vine and Matthews seafloor spreading model which has been of such fundamental importance to global plate tectonics. The first attempt to measure the vertical magnetic structure of oceanic crust exposed at a scarp face using near-bottom magnetic measurements is reported along with the theoretical basis for the reduction and analysis of these measurements. Magnetometers mounted to deeptow and submersible platforms were used to determine the horizontal and vertical variation of crustal magnetization of the upper 2 km of young oceanic crust exposed by a steep submarine escarpment at the Blanco fracture zone in the northeast Pacific Ocean. Results show a large amplitude magnetic anomaly is consistently found at the extrusive to intrusive transition indicating that the extrusives are strongly magnetized compared to the weakly magnetized intrusive crust. Profiles also show the systematic presence of more than one polarity with depth indicating the possibility of dipping polarity boundaries within the extrusives. Forward models find that a dipping polarity boundary model correctly predicts the amplitude and shape of the overlying sea surface anomaly while a block model with vertical polarity boundaries overestimates the amplitude. These in situ results are used to directly assess the contribution made by the extrusive section to the overlying magnetic anomaly signal, in essence, to define the magnetic source layer. It is concluded that the extrusive lavas constitute the main magnetic source for magnetic anomalies in oceanic crust less than 2 m.y. in age.

Introduction

The spatial variation in marine magnetic anomalies has been used with great success to unravel the history of the ocean basins and to understand the process of crustal accretion at the midocean ridges, but knowledge of the vertical dimension of crustal magnetization has remained the subject of vigorous debate. Controversy concerning the source of the magnetic anomalies has persisted since the first observations [Vine and Matthews, 1963], and while a number of models exist for the nature of the vertical magnetic structure of oceanic crust, few direct measurements are available (see Harrison, 1987, for a review).This paper presents the results of a detailed near-bottom magnetic survey of a steep scarp that exposes a near vertical cross-section of oceanic crust, which includes the extrusive lavas and the upper part of the intrusive dike section. The primary objective of the magnetic survey was to investigate the in situ vertical magnetic structure of oceanic crust and the horizontal variation in this structure, in order to understand the magnitude and nature of the contribution of the upper crust to marine magnetic anomalies. In situ magnetic field measurements not only allow the vertical magnetic structure to be directly related to the overlying magnetic anomalies, but also help to elucidate the processes of oceanic crustal accretion and evolution.Below, I briefly review the various approaches that have been used in the past to constrain the source region of oceanic crustal magnetization and its variation with depth. This is followed by an outline of the vertically-oriented magnetic survey method and analysis approach used in this study. Blanco scarp submersible magnetic data are then presented and discussed in terms of their implications upon the nature of the magnetic source layer and the structure and formation of oceanic crust.

The Magnetic Source Layer

Constraints on the vertical magnetic structure of oceanic crust have been inferred from a diverse range of observations. Marine magnetic anomalies reflect the intensity and geometry of their source, and a number of important models have been proposed based on these observations. For example, models with dipping reversal boundaries and multiple layers have been proposed based on observations of anomalous magnetic skewness and transition widths from marine and satellite magnetic anomaly data [e.g., Cox et al., 1972; Blakely, 1976; Cande and Kent, 1976; Blakely and Lynn, 1977; Kidd, 1977; Schouten and Denham, 1979; Macdonald et al., 1983; Arkani-Hamed, 1988, 1989]. Tectonic rotation of crustal units has also been suggested as a reason for the anomalous shape and skewness of some anomalies [Cande and Kent, 1976; Cande, 1978]. The nonuniqueness of potential field methods, however, means that any number of possible models can give rise to the observed anomalies, and independent data are needed to constrain the source of marine magnetic anomalies.

Rock samples collected by dredge and submersible provide a direct measurement of oceanic crustal magnetization (see Talwani et al., 1979; Harrison, 1987 for reviews) but limitations due to crustal exposure, sample orientation and sampling bias strongly influence the results. Extrusive basalts form the majority of seafloor rock samples and are typically strongly magnetized, supporting the contention that they comprise the primary source of magnetic anomalies. Extrusive lavas also decrease in magnetization with age due to low-temperature oxidation [Irving, 1970; Johnson and Atwater, 1977; Bleil and Peterson, 1983; Johnson and Pariso, 1993], which suggests that the extrusive contribution to the magnetic signal becomes less important with age [Cande and Kent, 1976; Blakely, 1983; Raymond and LaBrecque, 1987]. Deep crustal rocks sampled from the seafloor are usually highly altered and may not be representative of crust at depth, although some intrusive rocks can carry significant magnetization [e.g. Fox and Opdyke, 1973; Kent et al., 1978; Davis, 1981].Drilling provides the most direct method of measuring vertical magnetic structure, but results have been inconclusive due to a lack of significant penetration into oceanic crust (see Talwani et al., 1979 and Harrison, 1987 for reviews). Drilling results confirm dredge-based studies and find that highly magnetic extrusives undergo low-temperature alteration with age, reducing their ability to produce magnetic anomalies [Johnson and Pariso, 1993]. Evidence for tectonic rotations has also been observed within the extrusive layer [e.g. Hall and Robinson, 1979; Verosub and Moores, 1981]. Deep Sea Drilling Project Hole 504B in the equatorial Pacific is the deepest continuous hole to date, penetrating almost to the base of the dike section at about 2000 m depth [Dick et al., 1992; Pariso & Johnson, 1991]. Based on Hole 504B, Smith and Banerjee [1986] argued that the dike section is an important contributor to the marine anomaly signal, but acknowledged that alteration, which may vary spatially and temporally, is an important factor in controlling the magnitude of this contribution. Results from deeper crustal sections also find coherently magnetized units that could contribute to the magnetic signal [e.g., Pariso and Johnson, 1993a,b]. Drilling constrains only the vertical dimension, whereas magnetic anomalies arise from both vertical and horizontal variations in magnetic structure, so that care must be taken in interpreting the drill hole results [see Denham and Schouten, 1979].Ophiolite studies also provide important insight into the magnetic structure of oceanic crust [e.g., Vine and Moores, 1972; Levi et al., 1978; Luyendyk et al., 1982; Banerjee, 1984; Swift and Johnson, 1984, Hall et al., 1987, 1989; Pariso and Johnson, 1989] although some ophiolites have geochemical affinities to arc environments and may not be directly applicable to crust created at midocean ridges [e.g., Pearce et al., 1984]. Most ophiolites have gone through some metamorphism and alteration, so that their magnetic signature may not truly reflect in situ oceanic crustal signals. Criteria to assess the validity of magnetic measurements from ophiolites have therefore been suggested [Levi et al., 1978]. Some studies found that the dikes were not a significant source region because of hydrothermal alteration [Hall et al., 1987; Luyendyk et al., 1982], while other ophiolite studies argued that gabbros could provide a significant contribution [Luyendyk and Day, 1982].Clearly, a consensus on the source of the magnetic stripes remains elusive, and while important constraints have been provided through drilling, rock samples and ophiolite studies, these approaches are limited by the indirect relationship of these measurements to the observed magnetic anomalies. Drilling and rock studies address the type and magnitude of magnetization, the timing of remanence acquisition and the suitability of various rock units as a magnetic source, but it is the direction of magnetization and the geometry and distribution of magnetization that have the most significant control on the magnetic anomaly stripes. When polarity boundaries are vertical, crustal magnetization intensity controls the magnitude and shape of the observed magnetic anomalies. If polarity boundaries are near horizontal, the rock could be highly magnetic, but would not contribute much to the overall magnetic anomaly signal except to introduce a phase shift in the anomalies. In the extreme case of flat lying boundaries, no anomaly would be produced even though the rocks could be highly magnetic. Thus, the vertical distribution of crustal magnetization, and the spatial variation in this vertical structure, are both important parts of the magnetic signal. Conventional magnetic field measurements address the spatial variation in magnetization in an indirect way but cannot resolve any variation with depth due to the inherent nonuniqueness of the magnetic method. Drilling can address the vertical variations in magnetic structure, but it is not practical to carry out sufficient drillholes to obtain a relevant spatial variation in the vertical structure.

Vertical Magnetic Profiling

The basic premise of the vertical magnetic profiling (VMP) approach is to measure the magnetic field of crust exposed in near-vertical exposures of oceanic crust (Figure 1), such as large submarine escarpments that are most often found at oceanic transforms and fracture zones. Large submarine escarpments from their geometry provide a window into the vertical structure and stratigraphy of oceanic crust and provide significant opportunities for mapping both vertical and horizontal variations in oceanic crustal structure [e.g. Karson et al., 1992]. Scarps are complicated features, but for the sake of clarity, we assume the simple geometry of a single fault in order to demonstrate the vertical profiling approach. The validity and implications of this idealized assumption will then be discussed.

Conventional analysis of marine magnetic field data collected over the seafloor assume the magnetic source region is composed of crustal prisms of finite thickness and of infinite length perpendicular to profile [e.g. Vacquier, 1962; Schouten and McCamy, 1972; Parker, 1973]. This layered model is a good first approximation of oceanic crustal structure, but the nonuniqueness of the magnetic inverse method means that either the source layer thickness or magnetization can be solved for, but not both. Typically, a constant thickness source layer is assumed to obtain an vertically averaged crustal magnetization for the layer [Parker and Huestis, 1974]. While this approach has been extremely useful, neither the thickness of the layer or the computed magnetization is truly representative of the actual magnetic source layer. In the vertical magnetic case, the geometry of layered oceanic crust can be used to advantage. Qualitatively, magnetic measurements adjacent to a vertical scarp face (Figure 2a) can be viewed in a new, rotated reference frame, where the scarp face is rotated into a horizontal position (Figure 2b). The original ambient magnetic field and magnetization vectors are also rotated within this new reference frame. In this new orientation, a reasonable and realistic approximation can be made that oceanic crust is composed of a series of thin tabular bodies, perpendicular to the profile and extending to infinite depth assuming that the magnetic anomalies are 2-dimensional and strike along the scarp perpendicular to profile (Figure 2b). The zero level of the magnetization contrast can be fixed by using the contrast of non-magnetic seawater with the top of the scarp. The direction of magnetization remains unknown and a geocentric dipole direction must be assumed, as in conventional analysis. As with all inverse techniques, however, solutions are nonunique so that any number of magnetization solutions may exist. For the more realistic case of a non-vertical scarp face (Figure 2c), the slope is rotated into the horizontal by an amount equal to the slope angle (Figure 2d). In this case, instead of vertically oriented tabular bodies, source bodies dip with an angle equivalent to the slope in the new coordinate system (Figure 2). This dipping geometry introduces a phase shift into the anomaly.The Fourier transform F of the magnetic field b(x) on a level plane a distance z above the origin for a dipping dike-like body of vertical thickness t is given by Pedersen [1978](1)where J(k) is the Fourier transform of the magnetization, k is wave number, ( is the dip angle of the dike from horizontal and T(k) is given by:(2)and are unit vectors in the direction of the ambient magnetic field and magnetization, respectively. Assuming the dike extends to infinite depth, the second exponential in equation (1) goes to zero and with a slight rearrangement the following expression is obtained(3)As the dip of the dike approaches vertical, the complex expression for the dipping dike goes to unity and the familiar expression for an infinite halfspace is obtained. For the scarp case, the directional components perpendicular to the strike of the scarp of the field inclination and magnetization inclination are rotated by an amount equal to the slope angle. Similarly, by rotating the slope into the horizontal and the originally horizontal layers become dipping layers with an angle equal to the slope angle (see Figure 2).Inversion for crustal magnetization can also be attempted by rearranging equation (3) in terms of the observed magnetic field. Note that both the forward and inverse models assume a level topography and observation plane. The inversion assumes that all the layers dip with the same angle and that the layers extend to infinity with constant magnetization within the layers. As with any inverse technique, the results are nonunique and represent just one of many possible solutions.To demonstrate the VMP approach, equation (3) is used to calculate the magnetic field in a forward model of a synthetic traverse up a steep scarp face (Figure 3). The model assumes a south-facing scarp, oriented magnetic east-west, with a 40° slope. A highly magnetized upper layer (10 A/m), equivalent to the extrusive section of ocean crust, is modeled overlying a non-magnetic lower layer representing dikes. A reversely magnetized model was chosen to compare with actual survey data presented later. The magnetic field observed at an altitude of 10 m from the scarp face is calculated using the new geometry (see Figure 3). The model shows a large amplitude asymmetrical anomaly with a magnetic low at the scarp top, indicating reversed magnetization and a magnetic high at the extrusive to dike transition (Figure 3). The depth that the VMP technique "samples into the wall" is wavelength dependent, with longer wavelengths representing greater penetration of the wall, and shorter wavelength features representing shallow and topographic signals.As in all near-bottom techniques, the short length of the profile limits the longest wavelengths that can be reliably sampled. At the other end of the spectrum, topography and thin, slipped blocks, for example, would produce short wavelength anomalies.

Advantages and Disadvantages of Vertical Magnetic Profiling

Structural Problems - VMP depends upon obtaining magnetic measurements on a scarp face that exposes a cross-section of oceanic crust. In the example (Figure 3), a single normal fault geometry was assumed, but nature is more complicated. Fracture zones and transforms are the primary environments where oceanic crust is exposed to any significant extent, although other environments, such as at the median valley walls of slow spreading ridges, may expose crustal sections [Mutter and Karson, 1992]. While fracture zones and transforms exhibit large relief, only a small proportion are known to expose any significant vertical section of the crust. Many studies have shown that transforms do not expose significant cross-sections of oceanic crust, either because of dip-slip faulting or mass-wasting covering up the outcrop [e.g., Francheteau et al., 1976; CAYTROUGH, 1979; OTTER, 1984; Tamayo Sci. Team, 1984]. More recent observations and models of crustal evolution at ridge transform intersections (RTIs) have added to the tectonic complexity. For example, elevated topography and deep crustal rocks are typically found at the inside corner of RTI's, compared to low relief topography and predominantly extrusive rocks on the outside corner of RTI's. Recent models [Karson and Dick, 1983; Fox and Gallo, 1984; Severinghaus and Macdonald, 1988; Mutter and Karson 1992; Tucholke and Lin, 1994] propose that the extrusive carapace of crust is removed or stripped off by low angle faulting at the inside corner of an RTI, exposing deep crustal rocks that ultimately form massifs along the walls of transforms [e.g., Auzende et al., 1994; Dick et al., 1992]. Seismic studies also find that oceanic crust thins towards fracture zones [e.g. Detrick and Purdy, 1980; Detrick et al., 1982, Mutter et al., 1984], which raises the question of whether fracture zone crust is representative of "normal" ocean crust.

Tectonic complexity clearly has an important bearing on the nature of the magnetic signal produced by the crust, and only by careful site selection and preliminary mapping can these effects be minimized in a successful application of vertical magnetic profiling. Those areas where magnetic anomalies are truncated cleanly at the transform or fracture zone would make the best candidates for VMP surveys. Despite the problems associated with fracture zones and transforms, significant vertical exposures of oceanic crust on the seafloor do exist, such as at King's Trough [Le Pichon and Sibuet, 1971; Kidd et al., 1982], Hess Deep [Lonsdale, 1988; Francheteau et al., 1990], and the Blanco Trough [McManus, 1967; Embley and Wilson, 1992].Mass wasting is also an important process that is responsible for shaping transform and fracture zone walls and significant amounts of talus are often found associated with such features. Talus was exploited as a source of samples in early fracture zone studies, but this debris frustrates attempts at geological mapping and outcrop sampling. From the point of view of magnetic surveys, talus should appear as relatively non-magnetic bodies due to a random arrangement of blocks, which will cancel out any remanent magnetization. Thus, magnetic surveys may be able to "see through" the talus cover to the rock below. Magnetic anomalies may be produced by induced magnetization in the talus cover, but this is likely to be important only for deeper crustal rocks or highly altered basalts. Talus also increases the distance from the true source, resulting in a loss of magnetic anomaly resolution.Crustal Alteration - Alteration of crust exposed in transform or fracture zone walls could attenuate the magnetic signal of crust exposed at the fracture zone relative to crust away from the fracture zone. This may not be significant for the extrusive section which undergoes pervasive low-temperature alteration, but it could be important for deeper crust. Weakening magnetic anomalies towards the fracture zone would be one indication that alteration has affected the crustal magnetization signal.

The Blanco Scarp

The Blanco fracture zone (BFZ) offers a good opportunity to study the vertical magnetic structure of oceanic crust and to demonstrate the VMP technique (Figure 4). Lineated anomalies formed at the Juan de Fuca Ridge are cleanly truncated by the BFZ [Tivey, 1994], and there are areas of very steep topography, notably in the Blanco trough region with up to 2 km of relief [Embley and Wilson, 1992]. Blanco scarp, which marks the north wall of Blanco trough, is oriented ENE-WSW, perpendicular to the geomagnetic declination and magnetic anomalies. Dredging and sampling have recovered extrusive lavas and intrusive dike rocks, but little or no gabbroic rocks [Melson, 1969], which suggests that the upper portion of crust is exposed and has not been removed tectonically.

Geologic and Magnetic Setting

Blanco fracture zone [McManus, 1967; Embley et al., 1987] is a 360 km long, right-lateral transform fault zone in the Northeast Pacific Ocean that separates the Juan de Fuca Ridge (JDF) to the north from the Gorda Ridge to the south (Figure 4). The western BFZ is marked by a deep depression, the Blanco trough [Melson, 1969], which has a steep northern wall, the Blanco scarp, and a more gradual southern wall that rises up to form Parks plateau (Figure 4). Parks plateau appears to be an elevated portion of seafloor with a series of amorphous ridges, equivalent in depth to the bathymetry on the Juan de Fuca plate, north of the BFZ. Parks plateau is truncated abruptly at its southern edge, where it abuts the Tufts abyssal plain (Figure 4). The southern slope of Parks plateau has been interpreted as the old trace of the Blanco transform fault, which has only recently (about 0.5 My ago) jumped northward to the northern side of the Blanco trough region to form the very steep Blanco scarp [Embley and Wilson, 1992]. Thus, the Blanco escarpment may represent a relatively "fresh" exposure of JDF crust.The main study region of this paper focuses on the Blanco scarp, the steep, 2.4 km high northern wall of the Blanco trough, which cleanly truncates ridge-parallel, abyssal hill fabric of ocean crust created at the southern JDF over the last 1.5 My (Figure 4). Major faults can be seen to intersect the scarp on Gloria and SeaMARC II sidescan coverage of the area [EEZ SCAN 84, 1986] and from SeaBeam bathymetry analysis [Dauteuil, 1995]. Based on sea surface magnetic data (Figure 5), the Blanco scarp study area is located 40 km east of the axis, within the reversed polarity Matuyama epoch, between the Jaramillo anomaly (0.99 Ma) and anomaly 2 (1.77 Ma).

Blanco Scarp Deeptow Magnetic Survey

A preliminary near-bottom magnetic survey of Blanco scarp was obtained as part of the 1987 side-looking sonar survey of the scarp [Delaney et al., 1987; Palmer et al., 1987] using a proton magnetometer sensor towed behind using the Scripps Institution of Oceanography (SIO) Deep Tow [Spiess and Tyce, 1973]. Seven magnetic profiles (12 to 15 km long) were obtained along constant depth contours parallel to the scarp face, including one profile along the top of the scarp (Figure 6). The measured total magnetic field was merged with the transponder navigation, altimetry, and vehicle depth. The 3 upper profiles were located within the extrusive lavas while the lower 4 lines traversed the intrusive dike section (Figure 7). By inspection, the most striking result of the Deep Tow magnetic field data (Figure 7) is the progressive reduction in the amplitude of the magnetic signal with depth down the scarp face. The 4 lower profiles are between 1 and 2.5 km below the top of the scarp and show very little variation, while the 3 upper profiles show high amplitude magnetic field variations. This difference between the upper and lower set of profiles suggests that a magnetic contrast exists between these two units.

Blanco Scarp Submersible Magnetic Surveys

A more detailed magnetic study of the Blanco scarp was carried out as part of a 23 dive program using the submersible Nautile [BLANCONAUTE, 1991; Juteau et al., 1995a,b] (see Figures 4 and 8). Magnetic field data relevant to this study were collected on 7 dives located within the main Blanco scarp study area (Figure 8). Two magnetic dive traverses were carried out in younger crust near the Brunhes-Matuyama reversal at 130° 12'W (Figure 9). Magnetic field data were obtained using the Alvin 3-axis fluxgate magnetometer mounted to the front sample basket of Nautile. Data were recorded inside the submersible on a laptop computer and later combined with depth, heading and altitude, which was simultaneously recorded by the Nautile data system. Submersible navigation used a ship-based acoustic transponder net with an estimated XY accuracy of ±10 m. The measured 3-axis magnetic field components were vector summed for total field, because no independent orientation data were available for vector analysis. Magnetic field measurements were corrected for the permanent and induced magnetic fields of the submersible by minimizing the total magnetic field variations [Press et al., 1986] measured during spins of the submersible on the descent and ascent of each dive. Resultant noise levels were generally below 100 nT in amplitude and insignificant compared to the geophysical signals of several thousands of nT. Finally, the International Geomagnetic Reference Field (IGRF) for 1991 was removed from the magnetic data [IAGA, 1987]. Magnetic measurements were generally made at a constant altitude of 3 to 5 m above the seafloor with flat topography and an altitude of 5 m was assumed for the inversions. Small scale topographic variations were ignored, but these are of short-wavelength and do not affect the main magnetic signal.

Magnetic Results

Figure 10 shows a compilation of the vertical magnetic field profiles collected up the Blanco scarp face in the study area. Profiles BN03 and BN04 overlap the same crust and can be combined to form a complete vertical traverse of the scarp face (Figure 11). The most striking observation on this composite profile (Figure 11) is the prominent asymmetrical trend of the magnetic anomaly ranging from a minimum at the scarp top (2205 m) to a strong (12,000 nT) maximum at about 3100 m depth. This result is entirely consistent with the modeled magnetic anomaly for a reversely magnetized layer (Figure 3). Below 3100 m, the observed field rapidly decreases and remains low to about 3400 m depth. Below 3400 m, the magnetic anomaly becomes relatively low in amplitude, in marked contrast to the large amplitude and highly variable magnetic field response of the crust shallower than 3400 m. The magnetic transition from 3100 to 3400 m depth on profile BN03/04 (Figure 11) also coincides with the observed transition region from the extrusive pillow lavas to the intrusive dike section [Juteau et al., 1995a]. The crust, above 3100 meters depth, corresponds to the extrusive lava section, where pillow lavas were found in outcrop to the top of the scarp, whereas predominantly diabase was observed and sampled below 3400 m depth [Naidoo et al., 1992]. The lithologic transition between dike and extrusive pillow lavas is masked by talus and was not directly observable on any of the dives in the study area [Juteau et al., 1995a]. Talus cover forces the sensor to be further from the magnetic source and this would tend to subdue anomalies. Nevertheless, the magnetic signal records a very sharp transition (less than 100 m in vertical extent), which suggests the talus cover does not significantly mask the magnetic signal.A forward model with a single strongly magnetized layer (-12 A/m), approx. 900 m thick, overlying a less magnetized section (1 A/m) fits the long wavelength and asymmetrical shape of the observed anomaly (Figure 11). This model assumes reversed polarity crust, so that the lowest magnetic field value corresponds to the top of the scarp and the highest magnetic field corresponds to the boundary between the highly magnetic upper layer and less magnetic lower layer (Figure 11). The observed field shows greater variability above 3100 m in comparison to the crust below 3100, which is consistent with the Deep Tow results. The ALVIN magnetometer was turned off just above the top of the scarp, as the submersible returned to the sea surface, so that the predicted decrease in magnetic field amplitude above the top of the scarp was not recorded. Other profiles do show the decrease in magnetic field amplitude above the scarp (Figure 10).Profile BN03/04 was inverted for magnetization assuming that the magnetic source region extends to infinite depth using the method described earlier. The scarp face is oriented along an azimuth of 110°, perpendicular to the ridge axis and local magnetic meridian (I=60° D=20°). Magnetic field data were bordered out to twice their length to minimize any edge effects from the Fourier transforms. The magnetization inversion is shown in Figure 12. Bandpass filtering was done to minimize the effect of small scale topographic variations with a short wavelength cutoff of 25 m and a long wavelength of 1.5 to 2 km, depending on the length of the profile. The magnetization inversion solution was shifted by a small DC amount to ensure zero magnetization of seawater above the scarp. Figure 12 shows that the crust above 3100 m depth is highly magnetized in a reversed polarity sense and highly variable in intensity from 0 to 26 A/m. The mean magnetization for this 900 m thick upper crustal section is approx. 10 A/m. Several peaks and troughs can be identified in the magnetic field (Figure 11) and magnetization (Figure 12) over this upper crustal section, suggesting an alternating strongly and weakly magnetized section. In addition to this variability, there is a well defined zone of positively and strongly magnetized crust located between 3100 and 3300 m depth, at the base of the extrusive section. This zone and its possible origin is examined in more detail in a later section. Finally, in contrast to the highly magnetic extrusives, the dike section, below 3400 m depth, is weakly magnetized, averaging approx. 1 A/m and a positive polarity, although this is not well constrained. The submersible magnetic results illustrated by profile BN03/04 (Figure 11) are representative of the entire Blanco scarp study region. The clearly identifiable magnetic anomaly maximum at 3100 m in profile BN03/04 (Figure 11) occurs at slightly different depths in the other profiles (Figure 10) but remains a systematic feature of all the vertical magnetic profiles obtained from the study area. This anomaly corresponds to the dike-extrusive contact and can be correlated laterally between the profiles over a distance of 10 km, to give an estimate of the thickness of the magnetic layer and its spatial variation. The long-wavelength, asymmetrical magnetic anomaly trend can be seen in all the profiles and indicates the presence of a coherently magnetized layer. Some profiles, BN02 and BN05, only closely traverse part of the scarp and then show very smooth magnetic fields (Figure 10) where the submersible left the scarp face (> 10 m), ascended vertically, and progressively increased the distance from the scarp face. The results are robust and repeatable down to a 10-meter scale in dives BN05 and BN06 (see Figure 10). In the study area, the top of the scarp is usually marked by a magnetic low, even when the submersible left the bottom, which indicates the crust is reversely magnetized, consistent with the location and age of the study area.A summary of the computed crustal magnetization profiles for the Blanco scarp study area are shown in Figure 13. These profiles show the extrusive section has highly variable magnetization with alternating zones of apparently strongly and weakly magnetized crust. The average magnetization of the extrusive crustal section is approx. 10 A/m when these high and low magnetizations are taken into account. The extrusive section is also reversely magnetized, in keeping with the predicted age of this crustal section, although several profiles also show a zone of apparent positive magnetization at the base of the extrusive layer. The thickness of the strongly magnetized layer also varies from about 800 m in the west to about 1100 m in the east. All the profiles show that the dike section is weakly magnetized with an average of approx. 1 A/m. These weak magnetization values also mean that the polarity of the dike section is difficult to assess. Two magnetic profiles were obtained in younger crust near the Brunhes-Matuyama boundary (see Figure 9 for location) and show the same basic vertical magnetic structure seen at the Blanco scarp study area but with important differences (Figure 14). These profiles are slightly more complicated for a number of reasons. The scarp is less steep is this area and is predominantly covered in talus, making it difficult to correlate the magnetic signal with lithology. Dive BN19 was also conducted in a zigzag pattern in an attempt to map the Brunhes/Matuyama polarity boundary (Figure 9). While this strategy was not immediately successful, the asymmetrical shape of the anomaly is opposite in sense to the profiles observed in the Blanco scarp study area and consistent with positively magnetized crust, with a positive magnetic anomaly at the scarp top (Figure 14). Inversion shows that the upper 500 m of crust is strongly and positively magnetized (Figure 15). Dive BN18, located 3 km east of BN19 within the reversed polarity Matuyama epoch (Figure 8), traversed straight up the vertical scarp face and shows a positive anomaly at the top of the scarp, also indicating normal polarity (Figure 14). A strong magnetic anomaly low is found at about 2700 m depth, however, followed by a weak magnetic high at 2800 m. This anomaly morphology indicates a polarity reversal within the upper part of the crust. Inversion for magnetization shows the upper 200 m of crust to be clearly positively magnetized and the lower crust to be reversely magnetized (Figure 15). The proximity of BN18 to the Brunhes epoch suggests that the upper part of the extrusive crust is probably of Brunhes age, while the lower reversed polarity extrusive crust is Matuyama age. Below 2900 m, the BN18 magnetization signal is weak, similar to the response of the dike section seen in the Blanco scarp study area. These magnetic observations are consistent with outcrop lithology observations and samples of pillow basalts appearing at ~2900 m depth and continuing to the top of the scarp at 2500 m. This suggests that the extrusive layer is only 400 m thick, slightly less than the 500 m of BN19 and significantly different from the 800 to 1000 m thickness measured at the Blanco scarp study site. Variability in the thickness of the extrusive layer is common on the Juan de Fuca Ridge based on magnetic models [Tivey and Johnson, 1993; Tivey, 1994] and seismic results [McDonald et al., 1994; White and Clowes, 1990]. The mean positive magnetization of BN18 and BN19, averaged over the upper crustal section, is approx. 14 A/m, which is slightly higher than those found at the older Blanco scarp study site.Finally, submersible dive BN12 traversed across the top of the scarp in a conventional seafloor survey mode (Figure 16). The BN12 dive track, shown in Figure 4, covers the same region traversed by the SIO Deep Tow and shows a similar magnetic field to the uppermost profile of the deep-tow survey (Figure 7). The anomalous magnetic high observed in the deep-tow survey at 130°W is also observed in the submersible profile (Figure 16). To remove the effect of topography, the magnetic field was inverted for crustal magnetization assuming a constant thickness layer [Parker and Huestis, 1974]. Inversion results show that a strong magnetization high, implying that the anomaly is not caused by topography (Figure 16). This positive magnetization could be due to off-axis normal polarity volcanism or, because of its age, it may be the Cobb event (1.12 Ma) [Mankinen et al., 1978; Hsu et al., 1990].

Discussion

Crustal Magnetization Structure

Vertically oriented magnetic field profiles of the Blanco scarp reveal a systematic anomaly pattern consistent with their age and predicted polarity. Large amplitude magnetic anomalies are found at the extrusive to dike contact, indicating a strong magnetic contrast and implying that the extrusive crust is much more magnetic than the underlying intrusive dike section (Figure 13). The transition from highly magnetized crust to less magnetized crust occurs over a very narrow depth range of < 100 m. This sharp transition in magnetic properties is also found in Hole 504B at the base of the extrusive lavas and at the dike-transition zone [Alt et al., 1993]. Inversions for crustal magnetization of vertically oriented profiles show that the 1.5 Ma old extrusives have highly variable magnetization ranging from near zero to 26 A/m (Figure 13). When averaged over the full extrusive section thickness, the extrusives at the Blanco scarp site have a mean magnetization of ~10 A/m, compared with ~1 A/m for the dike section. Younger crustal profiles (BN18, BN19) show slightly higher mean magnetization (~14 A/m), as would be expected for younger crust (Figure 15). These results are consistent with previous paleomagnetic measurements of rock samples taken from the Blanco fracture zone and southern Juan de Fuca area [e.g. Johnson and Salem, 1994; Tivey, 1994].The highly variable magnetization within the extrusive section appears to group into discrete and coherent zones of high or low magnetization, averaging between 50 and 150 m thick (Figure 13). With the exception of the possible normal polarity unit near the base of the extrusives, the majority of reversed polarity extrusive magnetization varies considerably in intensity but does not become opposite in polarity, suggesting that these variations are most likely due to magnetization intensity rather than polarity reversals. Downhole logging results from Hole 504B show a similar pattern with considerable magnetic field variability in the extrusives in comparison to the dike section and transition zone [Alt et al., 1993]. The large variability in magnetization could be a topographically-induced signal arising from variations in observation distance from the scarp face, or it could be due to alternating units of variably magnetized lavas either through alteration, faulting or volcanic emplacement history such as sills. Submersible observations [Juteau et al., 1995a] show no clear relationship between geology and the short wavelength magnetization, and correlations only exist in a very general way. While geological controls on magnetization are certainly possible they are not well-constrained by these data, and another mechanism may in fact be responsible for a majority of this fine-scale signal. It has been suggested that magnetization could be fractal in three-dimensions [Pilkington & Todoeschuck, 1993] so that small random variations in source magnetization could "leak" through the inversion filters to produce apparent anomalies of longer wavelength. This view is supported by high frequency variations of the observed anomalies compared to the band-passed inversion solutions. Furthermore, a lack of correlation between adjacent Blanco profiles (Figure 13) of short-wavelength magnetization peaks also indicates magnetization intensity varies randomly on a spatial scale at these wavelengths. If this large variability in computed crustal magnetization is truly representative of the crustal magnetization then it points out the pitfalls of sample-based crustal magnetization estimates when only a limited number of samples are available.In contrast to the extrusives, the Blanco dikes are only weakly magnetized and show little variability in intensity (Figure 13). This behavior is similar to results of downhole logging at Hole 504B, which found magnetic field variability markedly reduced in the dike section and virtually zero in the transition zone [Alt et al., 1993]. No overall trend in magnetization intensity with depth is observed as has been suggested from Hole 504B results [Pariso and Johnson, 1991]. For the Blanco data, such long-wavelength trends are difficult to assess with Fourier-based analysis because of the short length of the profiles (< 2 km) and the bandpass filtering.The magnetization structure of the upper crust exposed at Blanco scarp can be divided into two main units. The upper magnetic unit corresponds to the extrusive basalts and is highly magnetized with a high degree of variability in intensity. This variability results in an average magnetization of ~10 A/m for the Blanco scarp extrusives. The lower unit corresponds to the intrusive dike sequence and is only weakly magnetized to ~1 A/m. These results have important consequences for the source of the marine magnetic anomaly stripes as discussed below.

Magnetization Polarity Structure

The most exciting aspect of vertical magnetic profiling is the ability to map the spatial variation in magnetic structure, which ultimately determines the magnetic anomaly signal measured at the sea surface. For the Blanco scarp site, all profiles show reversed polarity, consistent with their Matuyama age (Figure 13). In contrast, profiles BN18 and BN19, which are close to the Brunhes-Matuyama boundary, show positive magnetization in the upper crustal section with BN18 clearly showing a zone of reversed magnetization midway down the extrusive section (Figure 15). The Blanco scarp site profiles also show a zone of apparent normal polarity crust at the base of the extrusive lava section. The interpretation of this positive zone is important for understanding the overall vertical magnetic structure of the crust.The apparent positively magnetized zone is correlatable between several of the profiles in the eastern half of the Blanco survey site (Figure 13). The positive zone can be correlated between profiles over a distance of several kilometers, progressively thinning towards the west, from about 200-250 m thick in the east until it is completely absent in the westernmost profile, BN13 (Figure 13). The westernmost profiles BN18 and BN19 show no similar zone at the dike to intrusive transition. The inversion amplitude of the positive horizon (about 5 to 10 A/m) is almost equivalent to the overlying reversed polarity crust (Figure 13). Several possible scenarios could account for this zone.Alteration zone - The positive zone could be caused by alteration at the transition zone similar to the extrusive-dike transition at Hole 504B, which shows very low magnetization primarily due to intense alteration at this horizon [Pariso & Johnson, 1991]. A nonmagnetic zone would produce an apparent positive magnetization zone relative to the overlying reversed polarity extrusives or underlying reversed dikes. The Blanco inversions, however, show this zone is truly positive rather than being near zero. The vertical magnetization DC baseline is fixed by the magnetization contrast with seawater at the top of the scarp and to make the positive zone zero would require a large DC offset, which would make the upper crust highly magnetic (e.g. a mean of 22 A/m). Such a high magnetization is unlikely for this age of crust so that the positive zone is unlikely to be a DC shift problem in the magnetization inversion estimate. Alteration could produce a zone of induced magnetization at the dike to extrusive transition, however, and this would enhance an apparently positive magnetization. Juteau et al., [1995a,b] find that the extrusive volcanics are only moderately altered to a brownstone facies, as expected for their age. The deeper crustal sections appear to be altered to greenschist facies with veining and fracturing common in the dike section. In a study of Blanco dredge samples, Johnson and Salem [1994] found remanent magnetization dominated induced magnetization for both extrusive and dike rocks indicating that alteration had not destroyed any primary magnetization. No concentrated, discrete zone of alteration was observed at the dike to extrusive transition however [Juteau et al., 1995a,b]. Thus, while alteration may be locally important it is unlikely to be the cause of the observed systematic pattern in magnetization. Tectonic or volcanic unit- Another possibility is that the positive zone corresponds to a discrete volcanic unit, such as a sill, that has been emplaced after the extrusive crust was formed. If this was the case, the sill must be younger than the surrounding crust with positive magnetization, which implies a Jaramillo or Brunhes age. This sill would have been emplaced up to 12 km from the Jaramillo zone or 19 km from Brunhes. While little is really known about the age structure of the crust, and some studies find that volcanism can extend off-axis [e.g. Macdonald et al., 1989], it is considered unlikely that crustal emplacement has taken place at these distances from the JDF axis. The possibility of a sill cannot be ruled out, although no single correlatable geological unit was mapped that corresponds to this positive zone [Juteau et al., 1995a,b].A further possibility is that the positive zone could be caused by a tectonic shear zone at the dike to extrusive transition, perhaps associated with axis-parallel normal faulting. This mechanism would suggest local control and indeed a large fault with associated landslides is located near BN10 (Figure 8). A tectonic shear zone at the extrusive-dike level would likely reduce the magnetization in this zone but would not change the polarity of the magnetization except under extreme degrees of rotation. While submersible observations may show some signs of tectonic disruption they do not support extreme rotation at this horizon [Juteau et al., 1995].Normal polarity - While the above possibilities are all plausible to some degree the simplest explanation is a normal polarity zone. The positive horizon at the base of the extrusives is likely to be the oldest part of the extrusive crust and could represent normal polarity anomaly 2 (1.77 Ma). It is generally thought that magnetic boundaries dip within the extrusive section towards the spreading axis, become near vertical in the dikes and dip away from the axis in the deeper gabbro section [e.g., Cox et al., 1972; Cande and Kent, 1976; Kidd, 1977; Schouten and Denham, 1979; Macdonald et al., 1980] (Figure 17). Profiles BN18 and BN19 (Figure 15) can be interpreted as being of Brunhes age with BN18 recording a polarity transition to reversed Matuyama-aged. There may be a reversed tail at the base of the BN19 extrusives, but the short profile and complicated track precludes being definitive about this.Using the dipping boundary model of Macdonald et al., [1980], the transition width of sea surface magnetic anomalies is approximately equal to the width of the transition zone in the extrusive layer. This distance is also approximately equal to the half-width of the zone of accretion at the neovolcanic zone. For the southern Juan de Fuca Ridge, sea surface anomalies show an average transition width of ~5 km [Tivey, 1994], which is approximately equivalent to the present day axial neovolcanic zone half-width. How do these widths compare with those from the vertical profiles? Profile BN18 is approximately 3.5 km from BN19 predicting a full transition width of 7 km, slightly larger than that determined from sea surface anomaly and present day morphology. In older crust, the Blanco scarp site profiles are ~7 km away from the surface expression of anomaly 2, predicting a full transition width of 14 km; i.e. much greater than the 5 km estimate from sea surface anomaly transitions and neovolcanic zone width.These discrepancies can be reconciled using the Schouten and Denham [1979] model, which employs a gaussian distribution of dike and extrusive emplacement centered on the spreading axis. This model predicts a sigmoidal magnetization polarity distribution with thin gaussian tails of opposite polarity within a given crustal section that can extend for a considerable distance. The gaussian distribution of emplacement at the spreading center is stretched as it evolves through time because, in this model, plate velocity increases to full plate speed over finite distance, and crustal stretching also occurs over a finite width [Denham and Schouten, 1979]. The trailing edge of the distribution (i.e. that closest to the spreading center) is stretched the most producing a long tail of opposite polarity within a given crustal section. Assuming a spreading rate of 30 km/Ma and one standard deviation ( of 1.25 km for both emplacement and stretching (i.e. 4( is 5 km), gaussian tails could extend 7 km from the surface expression of the anomaly (Figure 17). The stretching would include crustal rotation along ridge parallel faults. The model readily explains the Blanco Scarp data, which show a thin normal polarity zone at the base of the extrusives, approx. 7 km from the sea surface expression of anomaly 2. The two vertical profiles at the Brunhes-Matuyama can also be explained in terms of dipping magnetic boundaries. These results suggest that dipping polarity boundaries do exist in oceanic crust, which ultimately has important consequences for the source of magnetic anomalies.

The Source of the Marine Magnetic Anomaly Stripes

The final goal of this study was to address the question concerning the source of the magnetic anomaly stripes. The VMP method provides an estimate of the absolute magnetization intensity of the crust and also the vertical extent of the extrusive magnetic layer, both critical parameters for estimating the contribution of the upper crust to the sea surface magnetic anomaly stripes. The Blanco results show that the upper extrusives are highly magnetic in comparison to the intrusive dike section and the transition occurs over a relatively narrow depth range. This rapid transition from magnetic crust to less magnetic crust provides support for the commonly held idea that the extrusive crust approximates a magnetic source layer. Two models for the extrusive magnetic source layer are now considered. The first model is a standard block model, as is typically assumed in many studies, using magnetization intensity and layer thickness values derived from this study. This model is compared to the dipping boundary model, which probably more realistically emulates reversal boundaries within the extrusive portion of the magnetic source layer. The dipping boundary model uses the stochastic emplacement model of Schouten & Denham [1979]. To compare to these models, a representative sea surface magnetic profile has been chosen that traverses along the Blanco scarp edge, very near the main study site at anomaly 2 (see tracklines on Figure 5).For the block model, the source layer is assumed to be 0.8 km thick with a mean magnetization of 10 A/m (Figure 18). These parameters are based on the Blanco Scarp study site and thus are most appropriate for anomaly 2 aged crust rather than for younger crust. Two variations of the block model were calculated. The first approach is to model a vertically oriented block as is typically assumed in simple models. The second approach was to investigate the role of tectonic rotation upon the signal by rotating a vertical block about an axis parallel to the spreading axis [Verosub and Moores, 1981]. The model of a vertical block (A), generates an anomaly that is almost twice the observed amplitude of anomaly 2 and does not fit the observed horizontal gradients (Figure 18). This model has ignored the effects of multiple polarity with depth within the extrusives, which would reduce the anomaly amplitude. In the rotated block model (Figure 18 Model B), rotation has a large influence upon the resultant anomaly because of the reorientation of the magnetization vector, which results in anomalous phase shifts of anomalies as seen by Cande [1978]. For the Juan de Fuca region, such anomalous phase shifts are not seen in the young anomaly sequence, and so simple block rotation is probably not an important factor here. This is demonstrated by the forward model (B), which assumed a 30° outward rotation from the axis (Figure 18). The resultant field is quite markedly phase shifted and the amplitude is still too large compared to the observed anomaly.In the final model (C), we used the dipping reversal boundary model of Schouten and Denham [1979], which describes dike emplacement, lava extrusion and crustal stretching as gaussian distributions. A standard deviation of emplacement and stretching of 1.25 km was assumed for a spreading rate of 30 km/m.y. (Figure 17). The dipping polarity boundaries define curved sigmoidal polygons of crust, which are used in the Won and Bevis [1987] analytic method to calculate the magnetic field anomalies of the source bodies (see Figure 18). As in the block model, a crustal magnetization of 10 A/m and a layer thickness of 0.8 km is assumed. Figure 18 shows that the dipping source body configuration predicts the correct amplitude of anomaly 2 and closely matches the horizontal gradients. The model anomaly amplitude is less because only the steeper parts of the polarity boundary produces an anomaly. Those parts that approach horizontal produce essentially no external magnetic field. From their configuration (Figure 18), the steeper parts of these sigmoidal bodies account for only 60% of the layer thickness. In such a case, for a given layer thickness, a magnetic inversion would significantly underestimate the magnitude of the crustal magnetization. Crustal stretching implies faulting and possibly crustal rotation of lava sequences, but no rotation of vectors were considered in this study. Rotation of lavas as a consequence of emplacement, burial and the spreading process is an important factor in modeling the magnetic response of the lavas. The lower lava units would be rotated towards the spreading axis, opposite to the block-tilting rotation of Verosub and Moores [1981] as modeled above (personal communication, Hans Schouten). Evaluation of these effects is beyond the scope of this paper but this is an important aspect of crustal magnetization and worth future consideration.In summary, the block models overestimate the amplitude of the observed anomaly 2, while the gaussian emplacement model correctly predicts the observed amplitude. Figure 19 shows the Blanconaute vertical magnetic profiles interpreted in terms of dipping boundaries. This diagram shows the possible spatial variation in crustal magnetization both with depth and in the horizontal dimensions although a large data gap exists at the Jaramillo. The sea surface magnetic field generated by such a configuration is also shown in comparison with the observed magnetic field. This calculated field shows that the extrusives contribute a significant part of the magnetic anomaly signal in young crust, supporting the contention of the extrusives as the major source.

Summary and Conclusions

The success of the Blanco survey and the consistency of the results underscores the utility of the vertical magnetic profiling approach. Vertical magnetic field measurements integrate the magnetic properties of the crust to better represent the macroscopic observations of the magnetic anomaly stripes, in comparison to extrapolations made from measurements made on individual rock samples for example. Drillholes provide important constraints on vertical magnetic structure, but the spatial variation in the vertical structure remains unknown, unless a series of offset holes are drilled, a prohibitively expensive undertaking. Thus, while magnetic intensity is important, it is ultimately the geometry of the magnetic boundaries that gives rise to the magnetic stripes, something that rock sample measurements and single drillholes cannot address. VMP of steep scarp faces provides a realistic alternative to drilling by being able to examine both the vertical magnetic structure of the exposed crust and the spatial variation of that vertical structure. This is especially important for determining the shape of a reversal boundary. Finally, because VMP is carried out in situ, the measurements can be directly related to overlying magnetic anomalies. A number of conclusions can be drawn from the Blanco magnetic survey results:1/ Vertical magnetic profiles indicate that the extrusive crust is strongly but variably magnetized, averaging 10 A/m for ~2 m.y. aged crust in contrast to the weakly magnetized intrusive dike section, which averages approx. 1 A/m. Changes in magnetic anomaly character appear to systematically coincide with the lithologic boundaries, so that by inference, the magnetic layer variations also reflect the depth of the intrusive dike to extrusive lava transition, which lends credence to the common assumption that the magnetic source layer is primarily defined by the extrusives of layer 2A. The highly magnetized extrusive layer varies between 800 and 1000 m in thickness across the main Blanco scarp study area and is thinner by a factor of two in younger crust. Such crustal thickness variations are consistent with observed variability of layer 2A on the Juan de Fuca Ridge [Tivey and Johnson, 1993; Tivey, 1994].2/ The vertical magnetic profiles also record polarity reversals at depth within the extrusive section, which can be correlated laterally for several kilometers. The occurrence and spatial distribution of these polarity boundaries is consistent with the dipping polarity boundaries predicted by crustal accretion models [e.g., Schouten and Denham, 1979]. This dipping geometry has important consequences for determining the relative contribution of the extrusive lavas to the magnetic anomaly stripe signal. Only the steep portions of the magnetic polarity boundary that are found in the central part of the extrusive layer will contribute substantially to the magnetic stripe anomaly. The more shallow dipping parts of the magnetic boundary will produce a longer wavelength and low amplitude anomaly and will not contribute significantly to the observed magnetic anomaly stripes.3/ From the measurements made at the Blanco scarp study site, the extrusives form the main source region for the overlying magnetic stripe anomalies. Block models based on magnetization amplitudes estimated from VMP profiles overestimate the magnitude of magnetic anomalies but dipping polarity boundary models accurately predict the amplitude and shape of the observed magnetic stripe anomaly. The dike section does not contribute significantly to the observed anomaly. This suggests that along with the timing of remanence acquisition, magnetization intensity and direction, the shape of the polarity boundary is an important parameter in determining the source of the marine magnetic anomaly signal.

Acknowledgments I thank Thierry Juteau for the opportunity to participate in the BLANCONAUTE cruise. J. Delaney provided the BLANCOTROUGH deeptow magnetic data and was responsible for the original inception of the Blanco scarp studies. I thank Barrie Walden of Woods Hole Oceanographic Institution, for use of the Alvin magnetometer and the Nautile crew for their help installing the sensor and its operation during the dives. I thank the rest of the BLANCONAUTE scientific staff for their patience in collecting the magnetic data. Hans Schouten provided many stimulating discussions regarding this work. I also thank Kim Klitgord and Rick Blakely for their in-depth reviews which helped improve the manuscript. WHOI provided the initial funds to join the Blanconaute cruise and to collect the magnetic data. The majority of the work was supported by NSF grant OCE-9115298.

References

Alt, J.C., H. Kinoshita, L.B. Stokking and Shipboard Scientific Party, Proc. Ocean Drilling Program,. Init. Repts., Leg 148, 5-352,1993.

Arkani-Hamed, J., Remanent magnetization of the oceanic upper mantle, Geophys. Res. Lett., 15, 48-51, 1988.

Arkani-Hamed, J. , Thermoviscous remanent magnetization of oceanic lithosphere inferred from its thermal evolution, J. Geophys. Res., 94, 17,421-17,436, 1989.

Auzende, J-M, M. Cannat, P. Gente, J.P. Henriet, T. Juteau, J. Karson, Y. Lagabrielle, C. Mevel and M.A. Tivey, Observations of sections of oceanic crust and mantle cropping out on the southern wall of Kane FZ (N. Atlantic), Terra Research, 6, 143-148, 1994.

Banerjee, S.K., The magnetic layer of the oceanic crust. How thick is it?, Tectonophysics, 105, 15-27, 1984.

Blakely, R.J., An age-dependent, two-layer model for marine magnetic anomalies, in The Geophysics of the Pacific Ocean Basin and its Margins, geophys. Monogr. Ser. 19, 227-24, AGU, Washington, DC, 1976.

Blakely, R.J., and W.S. Lynn, Reversal transition widths and fast-spreading centers, Earth and Planet. Sci. Lett., 33, 321-330, 1977.

Blakely, R.J., Statistical averaging of marine magnetic anomalies and the aging of oceanic crust, J. Geophys. Res., 88, 2289-2296, 1983.

BLANCONAUTE Shipboard Party, The western Blanco transform: Preliminary insights into the oceanic crustal architecture and transform zone processes, EOS Trans. AGU, 72, 44, 519, 1991.Bleil, U., and N. Peterson, Variations in magnetization intensity and low-temperature titanomagnetite oxidation of ocean floor basalts, Nature, 301, 384-388, 1983.

Cande, S.C. and D.V. Kent, Constraints imposed by the shape of marine magnetic anomalies on the magnetic source, J. Geophys. Res., 44, 547-566, 1976.Cande, S.C., Anomalous behaviour of the paleomagnetic field inferred from the skewness of anomalies 33 and 34, Earth Planet. Sci. Lett., 40, 275-286, 1978.

CAYTROUGH, Geological and geophysical investigation of the Mid-Cayman Rise spreading center: initial results and observations. In Talwani, M., Harrison, C.G.A., and Hayes, D.E., (Eds) Deep Drilling results in the Atlantic Ocean: Ocean Crust. Maurice Ewing Ser., 2, 66-93, 1979.

Cox, A., R.J. Blakely, and J.D. Phillips, A two-layer model for marine magnetic anomalies, EOS Trans. AGU, 53, 974, 1972.Dauteuil, O., Fault pattern from Seabeam processing: the western part of the Blanco Fracture Zone (NE Pacific), Mar. Geophys. Res., 17, 17-35, 1995.

Davis, K.E., Magnetite rods in plagioclase as a primary carrier of stable NRM in oceanic floor gabbros, Earth Planet. Sci. lett., 55, 190-198, 1981.

Delaney, J.R., F.N. Spiess, W.E. Colony, J.L. Karsten, and P. Nehlig, A complete Deep-Tow swath map of the south-facing wall of the Blanco Trough, Juan de Fuca Region, EOS Trans. AGU, 68, 1509, 1987.

Denham C.R., and H. Schouten, On the likelihood of mixed polarity in oceanic basement drill cores, in DSDP Results in the Atlantic Ocean: Ocean Crust Maurice Ewing Ser., eds M. Talwani, C.G.A. Harrison, D.E. Hayes, 2, 151-159, AGU, Washington, DC, 1979.

Detrick, R.S. and G.M. Purdy, Crustal structure of the Kane Fracture Zone from seismic refraction studies, J. Geophys. Res., 85, 3759-3777, 1980.

Detrick, R.S., M.M. Cormier, R. Prince and D.W. Forsyth, Seismic constraints on the crustal structure within the Vema Fracture Zone, J. Geophys. Res., 87, 10,599-10,612, 1982.

Dick, H.J.B., and Shipboard Scientific Party, Proc. Ocean Drilling Program,. Init. Repts. Leg 140, 37-200, 1992.EEZ SCAN 84 Scientific Staff, Atlas of the Exclusive Economic Zone, Western Coterminus United States: U.S. Geological Survey Misc. Investigations Series, I-1792, 152p, 1986.

Embley, R.W. and D. Wilson, Morphology of the Blanco Transform fault zone- NE Pacific: Implications for its tectonic evolution, Mar. Geophys. Res., 14, 25-45, 1992.

Fox, P.J. and N.D. Opdyke, Geology of oceanic crust: magnetic properties of oceanic rocks, J. Geophys. Res., 78, 5139-5154, 1973.Fox, P.J. and D.G. Gallo, A tectonic model for ridge transform-ridge plate boundaries: Implications for the structure of oceanic lithosphere, Tectonophys., 104, 205-242, 1984.

Francheteau, J., P. Choukroune, R. Hekinian, X. Le Pichon and H.D. Needham, Oceanic fractures do not provide deep sections of the crust, Can. Jour. Earth Sci., 13, 1223-1235, 1976.

Francheteau, J., R. Armijo, J.L. Cheminee, R. Hekinian, P. Lonsdale and N. Blum, 1 Ma East Pacific Rise oceanic crust and uppermost mantle exposed by rifting in Hess Deep (equatorial Pacific Ocean), Earth Planet. Sci. Lett., 101, 281-295, 1990.

Hall, J.M., and P.T. Robinson, Deep crustal drilling in the north Atlantic Ocean, Science, 204, 573-586, 1979.

Hall, J.M., B.E. Fisher, C.C. Walls, S.L. Hall, H.P. Johnson, A.R. Bakor, V. Agrawal, M. Persaud and A.M. Sumaiang, Vertical distribution and alteration of dikes in a profile through the Troodos ophiolite, Nature, 326, 780-782, 1987.

Hall, J.M., B. Fisher, C. Walls, T. Ward, L. Hall, and H.P. Johnson, Magnetic oxide petrography and alteration in the Cyprus Crustal Project Drillhole CY-4. Section through the lower sheeted complex and upper plutonic complex of the Troodos Ophiolite, Cyprus, in: Cyprus Crustal Study Project: Initial Report, Hole CY-4 edited by I. L. Gibson, J. Malpas P. T. Robinson and C. Xenophontos, Geological Survey of Canada, Ottawa, Ont. 1989.

Harrison, C.G.A., Marine magnetic anomalies - The origin of the stripes, Ann. Rev. Earth Planet. Sci., 15, 505, 1987.Hsu, V., D.L. Merrill and H. Shibuya, Paleomagnetic transition records of the Cobb Mountain event from sediments of the Celebes and Sulu seas, Geophys. Res. Lett., 17, 2069-2072, 1990.

International Association of Geomagnetism and Aeronomy (IAGA), Division I, Working Group 1, International geomagnetic reference field revision 1987, IAGA News, 26, 87-92, 1987.

Irving, E., The mid-Atlantic ridge at 45°N, XIV. Oxidation and magnetic properties of basalt; review and discussion, Can. J. Earth Sci., 7, 1528, 1970.

Johnson, H.P., and Atwater, T., Magnetic study of basalts from the Mid-Atlantic ridge, lat. 37°N, Geol. Soc. Am. Bull., 88, 637, 1977.Johnson, H.P. and J.E. Pariso, Variations in oceanic crustal magnetization: Systematic changes in the last 160 million years, J. Geophys. Res., 98, 435-446, 1993.

Johnson, H.P., and B.L. Salem, Magnetic properties of dikes from the oceanic upper crustal section, J. Geophys. Res., 99, 21,733-21,740, 1994.Juteau, T., D. Bideau, O Dauteuil, G. Manac'h, D.D. Naidoo, P. Nehlig, H. Ondreas, M.A. Tivey, K.X. Whipple, and J.R. Delaney, A submersible study in the western Blanco Fracture Zone, N.E. Pacific: Lithostratigraphy, magnetic structure, and magmatic and tectonic evolution during the last 1.6 Ma, Mar. Geophys. Res., 17, 399-430, 1995a.

Juteau, T., D. Bideau, O Dauteuil, G. Manac'h, D.D. Naidoo, P. Nehlig, H. Ondreas, M.A. Tivey, K.X. Whipple, and J.R. Delaney, De la propagation a l'accretion oceanique: etude par submersible du mur nord de la Fosse Ouest Blanco (Zone de Fracture Blanco, Pacifique NE), Bull. Soc. geol. France, 66, 105-121, 1995b.Karson, J.A. and H.J.B. Dick, Tectonics of ridge-transform intersections at the Kane Fracture Zone, Mar. Geophys. Res., 6, 51-98, 1983.

Karson J.A., S.D. Hurst, and P. Lonsdale, Tectonic rotations of dikes in fast spread oceanic crust exposed near Hess Deep, Geology, 20, 685-688, 1992.

Kent, D.V., B.M. Honnorez, N.D. Opdyke and P.J. Fox, Magnetic properties of dredged oceanic gabbros and the source of marine magnetic anomalies, Geophys. J. R. astro. Soc., 55, 513-537, 1978.

Kidd, R.B., R.C. Searle, A.T. Ramsay, H. Pritchard, and J. Mitchell, The geology and formation of King's Trough, northeast Atlantic ocean, Mar. Geol., 48, 1-30, 1982.Kidd, R.G.W., The nature and shape of the sources of marine magnetic anomalies, Earth Planet. Sci. Lett., 33, 310-320, 1977.

Le Pichon, X. and J.C. Sibuet, Western extension of boundary between European and Iberian plates during the Pyrenean orogeny, Earth Planet. Sci. Lett., 12, 83-88, 1971.Levi, S., S.K. Banerjee, S. Beske-Diehl, and B. Moskowitz, Limitations of ophiolite complexes as models for the magnetic layer of the oceanic lithosphere, Geophys. Res. Lett., 5, 473-476, 1978.

Lonsdale, P.Structural pattern of the Galapagos microplate and evolution of the Galapagos triple junction, J. Geophys. Res., 93, 13,551-13,574, 1988.

Luyendyk, B.P., B.R. Laws, R. Day, and T.B. Collinson, Paleomagnetism of the Samail Ophiolite, Oman. 1. The sheeted Dike complex at Ibra, J. Geophys. Res., 87, 10,883-10,902, 1982.

Luyendyk, B.P. and R. Day, Paleomagnetism of the Samail Ophiolite, Oman. 2. The Wadi Kadir gabbro section, J. Geophys. Res., 87, 10,903-10,917, 1982.

Macdonald, K.C., S.P. Miller, S.P. Heustis, and F.N. Spiess, Three-dimensional modelling of a magnetic reversal boundary from inversion of deep-tow measurements, J. Geophys. Res., 85, 3670-3680, 1980.

Macdonald, K.C., S.P. Miller, B.P. Luyendyk, T.M. Atwater, and L. Shure, Investigation of a Vine Matthews magnetic lineation from a submersible: The source and character of marine magnetic anomalies, J. Geophys. Res., 88, 3403-3418, 1983.Macdonald, K.C., R.M. Haymon, and A. Shor, A 220 km2 recently erupted lava field on the East Pacific Rise near lat 8°S, Geology, 17, 212-216, 1989.

Mankinen, E.A., J.M. Donnelly, and C.S. Gromme, Geomagnetic polarity event recorded at 1.1 m.y. B.P. on Cobb Mountain, Clear Lake volcanic field, California, Geology, 6, 653-656, 1978.

McDonald, M.A., S.C. Webb, J.A. Hildebrand, B.D. Cornuelle, and C.G. Fox, Seismic structure and anisotropy of the Juan de Fuca Ridge at 45°N, J. Geophys. Res., 99, 4857-4888, 1994.

McManus, D.A., Blanco Fracture Zone, Northeast Pacific Ocean, Marine Geology, 3, 429-455, 1967.Melson, W.G., Preliminary results of a geophysical study of portions of the Juan de Fuca Ridge and Blanco Fracture Zone, ESSA Tech. Memo. CandGSTM, 6, 33pp, 1969.

Mutter, J.C., R.S. Detrick and North Atlantic Transect group, Multichannel seismic evidence for anomalously thin crust at Blake Spur fracture zone, Geology, 12, 534-537, 1984.

Mutter, J.C. and J.A. Karson, Structural Processes at slow-spreading ridges, Science, 257, 627-634, 1992.

Naidoo, D.D., J.R. Delaney, T. Juteau, A scarp map profile of upper oceanic crust at the west Blanco transform (WBT): Insights into magmatic accretion, EOS trans. AGU, 73, 502, 1992.

OTTER (J.A. Karson, P.J. Fox, H. Sloan, K.T. Crane, W.S.F. Kidd, E. Bonatti, J.B. Stroup, D.J. Fornari, D. Elthon, P. Hamlyn, J.F. Casey, D.G. Gallo, D. Needham, and R. Sartori), The geology of the Oceanographer transform: the ridge-transform intersection, Mar. Geophys. Res., 6, 109-141, 1984.

Palmer, J., J.R. Delaney and F.N. Spiess, Recent Deep-Tow observations at the Blanco Fracture Zone-Juan de Fuca Ridge intersection, EOS Trans. AGU, 68, 1492, 1987.

Pariso, J.E. and H.P. Johnson, Magnetic properties of an analog of the lower oceanic crust: Magnetic logging and paleomagnetic measurements from drillhole CY-4 in the Troodos ophiolite, in: Cyprus Crustal Study Project: Initial Report, Hole CY-4 edited by I.L. Gibson, J. Malpas P.T. Robinson and C. Xenophontos, pp 278-293, Geological Survey of Canada, Ottawa, Ont. 1989.

Pariso, J.E. and H.P. Johnson, Alteration processes at Deep Sea Drilling Project/Ocean Drilling Program Hole 504B at the Costa Rica Rift: Implications for magnetization of oceanic crust, J. Geophys. Res., 96, 11,703-11,722, 1991.

Pariso, J.E. and H.P. Johnson, Do lower crustal rocks record reversals of the Earth's magnetic field? Magnetic petrology of oceanic gabbros at Ocean Drilling Program Hole 735B, J. Geophys. Res., 98, 16,013-16,032, 1993a.

Pariso, J.E. and H.P. Johnson, Do Layer 3 rocks make a significant contribution to marine magnetic anomalies? In situ magnetization of gabbros at Ocean Drilling Program Hole 735B, J. Geophys. Res., 98, 16,033-16,052, 1993b.

Parker, R.L., The rapid calculation of potential anomalies, Geophys. J. R. astr. Soc., 31, 447-455, 1973.Parker, R.L., and S.P. Huestis, The inversion of magnetic anomalies in the presence of topography, J. Geophys. Res., 79, 1587-1593, 1974.

Pearce, J.A., S.J. Lippard and S. Roberts, Characteristics and tectonic significance of suprasubduction zone (s.s.z) ophiolites, in Marginal Basin Geology, edited by B.P. Kokellar and M. F. Howell, Geol. Soc. Spec. Publ. London, 16, 77-94, 1984.

Pedersen, L.B., A statistical analysis of potential fields using a vertical cylinder and a dike, Geophysics, 43, 943-953, 1978.

Pilkington, M., and J.P. Todoeschuck, Fractal magnetization of continental crust, Geophys. Res. Lett., 20, 627-630, 1993.

Press, W.H., B.P. Flannery, S.A.Teukolsky and W.T. Vetterling, Numerical recipes: The art of scientific computing, Cambridge University Press, Cambridge, UK, pp 818, 1986.

Raymond, C. A. and J. L. LaBrecque, Magnetization of the oceanic crust: TRM or CRM?, J. Geophys. Res., 92, 8077-8088, 1987.

Schouten, H. and K. McCamy, Filtering marine magnetic anomalies, J. Geophys. Res., 77, 7089-7099, 1972.Schouten, H., and C.R. Denham, Modeling the oceanic magnetic source layer, in DSDP Results in the Atlantic Ocean: Ocean Crust Maurice Ewing Ser., eds M. Talwani, C.G.A. Harrison, D.E. Hayes, 2, 151-159, AGU, Washington, DC, 1979.

Severinghaus, J.P., and K.C. Macdonald, High inside corners at ridge transform intersections, Mar. Geophys. Res., 9, 353-367, 1988.

Smith, G.M., and Banerjee, S.K., Magnetic structure of the upper kilometer of the marine crust at Deep Sea Drilling Project Hole 504B, eastern Pacific Ocean, J. Geophys. Res., 91, 10337-10354, 1986.

Spiess, F.N., and Tyce, R.C., Marine Physical Laboratory deep-tow instrumentation system, S10 Ref., 73, 1973.

Swift, B.A., and H.P. Johnson, Magnetic properties of the Bay of Islands ophiolite suite and implications for the magnetization of oceanic crust, J. Geophys. Res., 89, 3291-3308, 1984.

Talwani, M., C.G.A. Harrison and D.E. Hayes, DSDP Results in the Atlantic Ocean: Ocean Crust Maurice Ewing Ser., AGU, Washington, DC, 1979.

Tamayo Tectonic Team (D. Gallo, W. Kidd, P. Fox, J. Karson, K. Macdonald, K. Crane, P. Choukroune, M. Seguret, R. Moody, and K. Kastens), Tectonics at the intersection of the East Pacific Rise with the Tamayo transform fault, Mar. Geophys. Res., 6, 159-185, 1984.

Tivey, M.A., Fine-scale magnetic anomaly field over the southern Juan de Fuca Ridge: Axial magnetization low and implications for crustal structure, J. Geophys. Res., 99, 4833-4855, 1994.

Tivey, M.A., and H.P. Johnson, Variations in oceanic crustal structure and the implications for the fine-scale magnetic anomaly signal, Geophys. Res. Lett., 20,17 1879-1882, 1993.

Tucholke, B.E. and J. Lin, A geological model for the structure of ridge segment in slow spreading ocean crust, J. Geophys. Res., 99, 11,937-11,958, 1994.

Vacquier, V., A machine method for computing the magnetization of a uniformly magnetized body from its shape and a magnetic survey, Benedum Earth Magnetism Symposium, Univ. of Pittsburgh Press, Pittsburgh, Pa., 123-127, 1962.

Verosub, K., and E. Moores, Tectonic rotations in extensional regimes and their paleomagnetic consequences for oceanic basalts, J. Geophys. Res., 86, 6335-6349, 1981.

Vine, F.J., and D.H. Matthews, Magnetic anomalies over oceanic ridges, Nature, 199, 947-949, 1963.Vine, F.J., and E.M. Moores, Model for the gross structure, petrology, and magnetic properties of oceanic crust, Geol. Soc. Am. Mem., 182, 195-205, 1972.

White, D.J., and R.M. Clowes, Shallow crustal structure beneath the Juan de Fuca Ridge from 2-D seismic refraction tomography, Geophys. J. Int., 100, 349-367, 1990.

Won, I.J. and M.G. Bevis, Computing the gravitational and magnetic anomalies due to a polygon: Algorithms and Fortran subroutines, Geophysics, 52, 232-238, 1987.