Abstract

Twenty-three Mesozoic “Chrons” (specific time intervals) from M0 to M40, including several in the Jurassic Magnetic Quiet Zone (“Jurassic Quiet Zone”), as well as Cenozoic Chron C34, are identified and mapped between the Atlantis and Fifteen-Twenty fracture zones on the North American plate, and between the Atlantis and Kane fracture zones on the African plate. Asymmetric seafloor spreading is indicated by the distances spanned over Chron intervals for the western and eastern flanks of the Central Atlantic ocean basin: C34 to the Mid-Atlantic Ridge (84 Ma to 0 Ma), M0 to C34 (120.6 Ma to 84 Ma), and M25 to M0 (154 Ma to 120.6 Ma). Chron M40 (167.5 Ma) is mapped ∼65 km outboard of the S1 magnetic anomaly over the African flank, and its conjugate, the Blake Spur Magnetic Anomaly (“Blake Spur Anomaly”) over the North American flank. Another pair of conjugate anomalies, the S3 magnetic anomaly over the African flank, and the East Coast Magnetic Anomaly (“East Coast Anomaly”) over the North American flank, are respectively located ∼30 km and 180 km inboard of the S1-Blake Spur Magnetic Anomaly pair. Therefore, the ridge jump to the east between “Blake Spur” and “East Coast” anomalies at ∼170 Ma theorized by Vogt and others in 1971 is supported by this study. Between the Atlantis and Kane fracture zones, the width of the African Jurassic Magnetic Quiet Zone is ∼70 km greater (22%) than the North American Jurassic Magnetic Quiet Zone. Correlatable anomalies exist over the African plate, suggesting a second ridge jump, to the west. Modeling results indicate that this jump occurred between 164 Ma and 159 Ma (Chrons M38 and M32). The ridge jumps can be related to plate interactions as North America separated from Gondwana. However, we note that the second ridge jump occurred approximately at the time suggested for the onset of seafloor spreading in the Gulf of Mexico.

INTRODUCTION

The overall kinematic history of the Central Atlantic Ocean between 10° N and 40° N is well understood (Fig. 1) (Klitgord and Schouten, 1986; Muller and Roest, 1992; Muller and Smith, 1993; Muller et al., 1999; Vogt, 1986; Withjack et al., 1998). Following Middle to Late Triassic rifting between North America and Africa, seafloor spreading began at ca. 185 Ma (Withjack et al., 1998). Seafloor spreading anomalies corresponding to intervals Mid-Atlantic Ridge to C34 and M0 to M25 are well documented (Klitgord and Schouten, 1986; Muller and Roest, 1992, Muller et al., 1999). Magnetic data can be divided into five major anomaly provinces that form symmetric bands along the eastern and western sides of the Central Atlantic: (1) Cenozoic Chron C34 to the Mid-Atlantic Ridge; (2) the Cretaceous Magnetic Quiet Zone; (3) Mesozoic Chrons (M-Series) M25 to M0; (4) the Jurassic Magnetic Quiet Zone consisting of Mesozoic Chrons M41 to M26; and (5) a zone of low-amplitude anomalies between the East Coast Magnetic Anomaly (“East Coast”) and Blake Spur Magnetic Anomaly (“Blake Spur”) over the North American flank of the ocean basin is called the Inner Magnetic Quiet Zone (Rona et al., 1970; Vogt, 1986; Vogt et al., 1971). S1 and S3, similar to the “East Coast” and “Blake Spur” anomalies, are located over the African flank inboard of Chron M40 (Roeser et al., 2002).

The age and relative rotation history between the North American and African plates have been determined by combining fracture zones identified from Geosat and Seasat altimetry data with magnetic isochrons (Klitgord and Schouten, 1986; Muller and Roest, 1992; Muller et al., 1999). We use similar methods and integrate gridded magnetic data with a recently compiled, extensive profile-based magnetic data set to: (1) identify and map eighteen M-Series Chrons (M25 to M0) in the Central Atlantic Ocean; (2) identify and map five additional M-Series Chrons within the Jurassic Magnetic Quiet Zone (M40 to M28); (3) interpret the intersections between Chrons C34, M0, and M25 with the Fifteen-Twenty, Kane, and Atlantis fracture zones to calculate finite difference Euler poles needed to reconstruct the relative plate motion between North America and Africa from 154 Ma to present (Chron M25 to Mid-Atlantic Ridge); (4) estimate asymmetric spreading rates from Chron C34 to the Mid-Atlantic Ridge, Chrons M0 to C34, and Chrons M25 to M0; (5) document a previously proposed ridge jump in the North American plate that occurred at ca. 170 Ma (Vogt et al., 1971; Vogt, 1973, 1986); and (6) propose the existence of another ridge jump between 164 Ma and 159 Ma (Chrons M38 and M32).

GEOPHYSICAL DATA

The open-file geophysical data used in this study consisted of two magnetic anomaly grids, a gravity grid, and three magnetic anomaly profile data sets. Two open-file magnetic anomaly grids partially cover the Central Atlantic Ocean (Fig. 2) (Hinze et al., 1988; Verhoef et al., 1996). Magnetic anomaly profile data sources are: the National Oceanic and Atmospheric Administration/National Geophysical Data Center GEOphysical DAta System (GEODAS) database, the Kroonvlag project data from the Geological Survey of Canada (Collette et al. 1984), and forty-two lines digitized from Vogt et al. (1971)(Fig. 3).

The satellite-derived free-air gravity anomaly grid over the Central Atlantic Ocean is a two arc-minute grid of fairly uniform coverage acquired during the Geosat Geodetic Mission and the ERS-1 Geodetic Phase along closely spaced satellite tracks (Fig. 4) (Sandwell and Smith, 1997).

METHODS

Fracture Zones and Flowlines

Flowlines describe the relative motion between two accreting plates over time. Fracture zones provide approximations of these flow-lines by recording changes in seafloor spreading direction over time (Collette and Roest, 1992; Klitgord and Schouten, 1986). Fracture zones can be identified by their characteristic signatures in magnetic and gravity data. Magnetic lineations related to geomagnetic polarity reversals are often offset along fracture zones, and free-air gravity anomalies over fracture zones are typically distinct, 25–50 km wide, curvilinear minima. Therefore, we interpret fracture zones from maps of gravity and magnetic anomalies by tracing gravity minima through the transform offsets (Figs. 1, 5, and 6). From south to north, the Fifteen-Twenty, Kane, and Atlantis fracture zones span most of the Central Atlantic, or ∼2000 km, and they extend west and east close to the coasts of North America and Africa, or almost 6000 km (Fig. 1).

Geomagnetic Isochrons

Distinct magnetic anomalies are produced by rocks accreted along the Mid-Atlantic Ridge during specific time intervals (“Chrons”) that are associated with geomagnetic polarity reversals. Magnetic anomalies produced by sea-floor spreading are identified in a two-step process: (1) by simultaneous interpretation of gridded and profile magnetic anomalies to correlate significant anomaly trends, and (2) by comparison of selected anomaly profiles with synthetic anomaly profiles that are calculated from two-dimensional magnetic models based on the geomagnetic polarity reversal time scales of Channell et al. (1995) and Sager et al. (1998). The magnetic data coverage over the North American flank of the Central Atlantic ocean basin is relatively dense, including extensive ship-track coverage as well as gridded data sets (Figs. 2 and 3). The data coverage over the African flank has low ship-track data density and lacks gridded data south of 30° N. Therefore, we interpret Mesozoic Chrons from the Bahamas to just northeast of the Atlantis fracture zone on the North American flank, but only along a 200-km-wide corridor between the Atlantis and Kane fracture zones on the African flank.

We correlate magnetic anomalies in groups, or sequences of anomalies, for line-to-line coherency. Anomalies produced by geomagnetic polarity reversals should be consistent and roughly parallel to seafloor spreading centers, while anomalies produced by basement relief are typically shorter and less organized. Initially (step 1), prominent anomalies coinciding approximately with the globally mapped magnetic isochrons by Muller et al. (1997) were identified on maps: C34, M0, M4, M10N, M16, M21, and M25. Next, additional anomalies were identified between this first set of prominent anomalies by approximating distances based on time intervals defined by geomagnetic polarity reversals and anomaly character (Channell et al., 1995; Sager et al., 1998). Landward of M25 in the Jurassic Magnetic Quiet Zone, anomalies were also identified by approximating distances based on time intervals and by comparing anomaly shapes with Chrons identified by Roeser et al. (2002) and Sager et al. (1998): M28, M29, M32, M38, and M40.

Excellent examples of Jurassic Magnetic Quiet Zone anomalies, which can be used for comparison with the Central Atlantic “Jurassic Quiet Zone” anomalies, are found over the western Pacific Ocean and off the northwest shelf of Australia. Anomalies over the Jurassic Magnetic Quiet Zone in the Pigafetta Basin, just east of the Northern Mariana Basin, are produced by basement at depths of 6–6.5 km. Two deep-tow profiles were acquired over an ∼8 m. y. sequence (167.5 Ma to 157.5 Ma) containing eighty-eight M41 to pre-M29 Chrons (Sager et al., 1998). After upward continuing profiles to sea level, only 44% of the Chrons were retained (Sager et al., 1998). We use similar upward continued profiles to compare and identify anomaly patterns over the Central Atlantic.

Twenty-three magnetic anomaly profiles were selected between Atlantis and Kane fracture zones over the western flank of the ocean basin, and thirteen profiles were selected between Atlantis and Kane fracture zones over the eastern flank. These were correlated with synthetic anomalies generated from a two-dimensional model based on the geomagnetic polarity reversal time scale (step 2). Anomaly profiles were projected to straight-line segments, approximately parallel to the flow lines and displayed at the same horizontal and vertical scales. Correlated anomalies were then identified as Chrons according to their similarity with the synthetic profiles 01(Table 1). Line-to-line correlations were connected with solid lines, which were, in turn, connected by dotted lines to maintain a spatial reference frame over fracture zones or areas where profile data did not exist (Figs. 5–8). Two-dimensional magnetic models were generated using 2-km-thick blocks representing constant polarity intervals of the geomagnetic polarity time scale with 2-km-wide Gaussian polarity transitions between the Chrons. Magnetic anomalies produced by these models are compared to magnetic anomaly data profiles. Basement depths in the “Jurassic Quiet Zone” off North America are ∼7–9 km (Uchupi et al., 1984a, 1984b); therefore, the basement depth of our two-dimensional, magnetic anomaly model is 8 km for North America. We use 8-km basement depth for Africa as well.

Finite-Difference Poles

Stage poles were calculated for the North American and African flanks: Chron C34 to Mid-Atlantic Ridge, Chron M0 to C34, and Chron M25 to M0; and total reconstruction poles were calculated for C34 (84 Ma), M0 (120.6 Ma), and M25 (154 Ma) (02Tables 2 and 033). A method similar to that described by Engebretson et al. (1984) was used to find the best-fit Euler pole for pairs of control points defined by the intersections between interpreted geomagnetic isochrons (C34, M0, and M25) and fracture zones (Atlantis, Kane, and Fifteen-Twenty). Each pair of control points is assumed to have been originally coincident so that a plate rotation exists that will restore them to a single point. A computer program minimized mismatch between one set of control points and the rotated set of control points by least-squares approximation (Bird, 2004). The program required an input geographic seed location (latitude and longitude) for the center of a scan matrix and a scan increment (in degrees); it built the scan matrix and then searched for the best-fit Euler pole. Once the best-fit pole was located, its coordinates were input into the program as the next seed location using a smaller scan increment. Convergence to a solution yielding a 90% confidence region was typically achieved in three or four iterations (Bird, 2004).

ANALYSIS

Oceanic Crust and Fracture Zones

Gridded magnetic data and published seismic refraction data indicate that oceanic crust exists beneath the interpreted Jurassic Magnetic Quiet Zone Chrons, or M40 to M28 (Figs. 5–7, 04Table 4) (Ewing and Ewing, 1959; Houtz, 1980; Katz and Ewing, 1956; Sheridan et al., 1966). Crustal boundaries mapped by Uchupi et al. (1984a, 1984b) indicate that oceanic crust extends inboard almost as far as the East Coast Magnetic Anomaly (Fig. 5). Magnetic anomalies produced by geomagnetic polarity reversals coinciding with Chrons M25 through M0, and subtle, broad linear anomalies, similar to M40 to M28 between “East Coast” and “Blake Spur” anomalies, also support this interpretation of oceanic crust (Fig. 6).

The primary difficulty associated with mapping fracture zones using satellite-derived free-air gravity data over long distances, such as from the Mid-Atlantic Ridge to the continental shelves of North America and Africa, is their variable expression. Their expression can be affected by several factors including seafloor spreading rate, length of transform segment, magma supply, and sedimentation, which can result in fracture zones that seem to disappear and reappear along strike. Muller and Roest (1992) estimate the average error for identifying fracture-zone locations from satellite-derived gravity data to be 5 km.

Using satellite-derived gravity alone over the North American plate, the interpreted landward projections of the Atlantis, Kane and Fifteen-Twenty fracture zones extend to Chron 21, 75 km inboard of Chron 25, and to Chron 25, respectively (Figs. 1 and 5). The Atlantis fracture zone trace can be followed farther landward in the North American plate to over 100 km inboard of Chron M40 using the magnetic anomaly grids. M-Series Chrons are offset sinistral almost 80 km along the inboard projection of this fracture zone (Fig. 6). The landward projections of the Kane and Fifteen-Twenty fracture zones on the African plate extend to Chron M16 and Chron M25, with dashed lines indicating less confidence (Fig. 1). The volcanic Great Meteor Seamount and Saharan Seamount complexes mask parts of the Atlantis fracture zone on the African plate, where its inferred location between Chrons C34 and M25 is represented by a dashed line (Fig. 1).

Geomagnetic Isochrons

Geomagnetic Chrons M0 to M40 have been interpreted and mapped outboard of the North American and African continental slopes (Figs. 5–8). Due to their relatively high amplitudes, M25 to M0 Series anomalies are readily identified; however, “Jurassic Quiet Zone” anomalies (M28 to M40) are more difficult to interpret because: (1) anomaly amplitudes are much lower; (2) they are located over the deepest parts of the basin near the continental margins; and (3) geomagnetic polarity reversals occurred at a rapid rate during this time. All M-Series anomaly amplitudes over the eastern side of the Central Atlantic Ocean are lower than those over the western side.

The M25 to M0 province offshore North America is a region characterized by distinctive packages of anomalies (Fig. 9). M0 has a prominent positive peak just landward of a sharp anomaly minimum; then a broad low is followed by a sequence of three peaks associated with the M1, M3, and M4 Chrons. M4 typically has the greatest amplitude of these three anomalies, similar to M0. Landward of M4, following an interval lacking coherent anomaly trends, are Chrons M10N, M12A, and M14. M10N typically coincides with a broad high-amplitude anomaly, which sometimes is expressed as two peaks and it is identified at the crest of the western flank of this anomaly. M12A and M14 are slightly lower in amplitude than M10, but coincide with distinctive peaks that, together with M10N, can be mapped with consistency from the Atlantis to the Fifteen-Twenty fracture zones. The package of anomalies bounded by M16 to M21 includes seven Chrons: M16, M17, M18, M19, M20n-1, M20, and M21. Chrons M16, M20, and M21 are the highest amplitude anomalies overall in the package, but are not well defined over the entire interval from the Atlantis to Fifteen-Twenty fracture zones. However, these Chrons are consistent with respect to their relative spatial position. Between M16 and M20, Chrons M17, M18, M19, and M20n-1 are just as consistent with respect to their position, although Anomaly M17 is often lower in amplitude. Between the Atlantis and Kane fracture zones, the next three prominent anomalies are Chrons M22, M23, and M24. M25 is more difficult to interpret. Its general character is a broad anomaly, with the Chron coinciding with highest amplitude at the crest of its eastern flank; however, often the western flank of the anomaly is much lower in amplitude and even absent at times.

M25 to M0 anomalies on the African side of the Central Atlantic do not identically mirror those on the North American side (Fig. 9). However, distinctive anomalies are correlated and interpreted such that the same Chrons are identified on both sides of the ocean basin. Similar to M0 anomalies over the North American plate, M0 anomalies over the African plate are characterized by a sharp minimum to the west followed by a relatively broad, high-amplitude peak to the east. Chron M4 is similar in shape to Chron M0. M1 and M3 anomalies are associated with anomaly peaks, and their spatial relation between M0 and M4 is consistent. Chrons M10N, M12A, and M14 are evenly spaced peaks; however, M12A is identified as the eastern peak of a pair of low-amplitude anomalies superimposed on a broad-anomaly high between M10N and M14. Chron M16 is identified as the western peak of a pair of anomalies. Chrons M17–M21 over the African plate mirror the same package of anomalies over the North American plate by an evenly spaced sequence of anomaly highs. Similar to the North American plate, Chron M17 is often a low-amplitude peak near the landward side of Chron 16. Unlike the North American plate, Chron M21 over the African plate is consistently higher in amplitude, and more prominent, than M20. This might be the result of our comparatively limited area of investigation on the African side of the basin. Chrons M22, M23, and M24 are the next landward set of peaks on the African plate, and M22 and M23 are easily mapped. However, M24 can only be mapped south of Profile D (Fig. 9). The low-amplitude Chron M25 is transitional to the Jurassic Magnetic Quiet Zone, and it is mapped by correlating a broad, subtle, anomaly high. Also, anomaly M25, which is typically a low-amplitude anomaly, cannot be identified on profile B and only tentatively on Profile D because the Tropic Seamount produces high-amplitude (>400 nT) anomalies that interfere with Chrons along these profiles (Figs. 8 and 9).

Correlated “Jurassic Quiet Zone” anomalies over the North American plate include M28, M29, M32, and M40 (Fig. 10). M28 and M29 are similar in amplitude to M25, and, while their shapes vary somewhat from north to south, the distinct pair of relatively prominent anomalies is mapped with confidence in several locations. Chron M32 is also sometimes difficult to identify, but like M28 and M29, a distinct peak that is slightly higher in amplitude than the surrounding anomalies, similar to modeled anomalies, occurs in several locations. M40 is a persistent anomaly high that is mapped ∼50–75 km outboard of the Blake Spur Magnetic Anomaly.

Mirroring the North American plate, M28 and M29 are mapped as a pair of peaks similar to and just landward of M25. M32, over the African plate, is a distinct peak that is slightly higher in amplitude than the surrounding anomalies, similar to M32 over the North American plate. The characteristic package of anomalies between Chrons M38 and M39 are superimposed on a broad, asymmetric high with a steep flank over M38 (Sager et al., 1998). M40 is a relatively high-amplitude anomaly that is mapped ∼65 km outboard of S1.

Prominent, near-shore, magnetic anomalies, extending hundreds of kilometers parallel to the coasts of North America and Africa, have been identified: the “East Coast Anomaly” and “Blake Spur Anomaly” are ∼180 km apart over North America (Figs. 1 and 5–7) (Klitgord and Schouten, 1986; Vogt, 1986), and the S1 and S3 anomalies over Africa are ∼50 km apart (Figs. 1 and 8) (Roeser et al., 2002; Roeser, 1982; Verhoef et al., 1991). The amplitudes, trends, and lengths of “East Coast,” “Blake Spur,” S1, and S3 are similar to M-Series anomalies (M0 to M25). Vogt et al. (1971) and Vogt (1973; 1986) suggested that the Blake Spur Magnetic Anomaly is the result of an eastward jump of the spreading center away from the East Coast Magnetic Anomaly prior to 170 Ma.

The East Coast Magnetic Anomaly bends eastward at ∼40° N, which appears to also be the northernmost extent of the Blake Spur Magnetic Anomaly. To the north of this eastward bend, Roeser et al. (2002) suggested that S1 and “East Coast” are conjugate anomalies that coincide with the oceanic-continental crustal boundary. However, Roest (1987) suggested that to the south, S1 and “Blake Spur” are conjugate anomalies. We interpret that East Coast Magnetic Anomaly-S3 and Blake Spur Magnetic Anomaly-S1 are conjugate anomalies and that East Coast Magnetic Anomaly and S3 coincide with the ocean-continent boundary.

Ridge Jumps and Asymmetric Spreading

Ninety-percent confidence regions of calculated Euler poles, are shown in Figure 11. Total reconstruction poles reported by Muller and Roest (1992) lie within the 90% confidence regions. Stage poles for the western and eastern sides of the Central Atlantic Ocean basin appear to show significant asymmetry 03(Table 3). However, the asymmetry suggested by stage-pole rotation angles can be misleading due to the proximity of a given pole with respect to a given plate location. That is, a small change in this distance can produce a relatively large change in the rotation angle when the plate being rotated is very close to, or even contains, the rotation pole. For example, M25 to M0 spreading rates offshore Morocco are 10 mm/a for M21 to M0, 16 mm/a for M25 to M21, and 10 mm/a for Jurassic Magnetic Quiet Zone (Roeser et al., 2002). This slow rate for the “Jurassic Quiet Zone” is reasonable when considering its close proximity to the reconstruction pole that describes the plate motion over this time (“M0-M25 SPE”; Fig. 11). Therefore, it is important to measure distances between Chrons over specific time intervals to determine relative spreading rates 05(Table 5).

Asymmetric spreading is characterized by excess seafloor accretion on one side of a spreading center and is typically less than 5%; however, the percentage of asymmetric spreading of the East Pacific Rise is over 10% (Muller et al., 1998). We interpret the widths of anomaly provinces C34 to Mid-Atlantic Ridge and M0 to C34 to indicate asymmetric spreading 05(Table 5). The spreading rate for the high-amplitude M25 through M0 Chrons in the North American plate (14.4 mm/a) is ∼10.5% faster than the 12.9 mm/a spreading rate for the African plate, which indicates asymmetric spreading in this interval as well.

The spreading rate for the North American Jurassic Magnetic Quiet Zone reported by Klitgord and Schouten (1986) of 19 mm/a is close to the “Jurassic Quiet Zone” rate calculated in this analysis (19.2 mm/a); however, it differs significantly from our African “Jurassic zone” rate of 24.6 mm/a, which is ∼22% greater. Figure 10 shows correlated anomalies for the low-amplitude M40 to M28 sequence compared to two-dimensional synthetic models using two different sets of spreading velocities. This large difference in spreading rates suggests that a ridge jump may have occurred in this time interval. Therefore, the magnetic data over the African side were inspected for repeated anomalies. A sequence of anomalies near M38 appears to repeat on the few profiles we analyzed and could account for an additional ∼70 km of oceanic crust that seems to be absent from the North American side. The seafloor spreading models were modified by removing these Chrons from the North American model and inserting them into the African model such that correlated “Jurassic Quiet Zone” anomalies over Africa include M28, M29, M32, M38 (abandoned from the North American side), M38 (African side), and M40. The combined width of the conjugate Jurassic Magnetic Quiet Zones is 570 km, corresponding to a total symmetric spreading rate of 43.8 mm/a. We then use a half-spreading rate of 21.9 mm/a for each side of this low-amplitude sequence. Comparing “Jurassic Quiet Zone” anomalies with those calculated from our ridge-jump models suggests that a sequence of anomalies approximately between M38 and M32 could be produced by a sliver of oceanic lithosphere that was abandoned by a ridge jump.

The proximity of the African Jurassic Magnetic Quiet Zone stage pole, and the fanning of North American M25 to M0 and “Jurassic Quiet Zone” anomalies to the south, indicate that this distance increases southward to 10° N. Furthermore, Roeser et al. (2002) reported that interpreted anomalies over the Seine Abyssal Plain, offshore Morocco, approach S1 northward at a 10° angle, which also reflects the close proximity to the rotation pole for this interval. We suggest that the difference in “Jurassic Quiet Zone” widths indicates that a ridge jump occurred between 164 and 159 Ma (Chrons M38 to M32), abandoning ∼35 km of North American lithosphere on the African side of the Central Atlantic ocean basin, but that its extent and timing cannot be determined with precision.

Our identification of M40 along North America suggests that the high-amplitude western and eastern basin bounding anomalies, East Coast Magnetic Anomaly-S3 and Blake Spur Magnetic Anomaly-S1, are conjugate pairs. The asymmetric spreading required to create oceanic crust from “East Coast” to “Blake Spur” (180 km), and S3 to S1 (50 km), greatly exceeds the maximum asymmetry reported by Muller et al. (1998). We conclude that a ridge jump occurred within the Inner Magnetic Quiet Zone, substantiating the ridge jump hypothesized by Vogt et al. (1971) and Vogt (1973; 1986). This jump probably occurred prior to 167 Ma (M40). Roeser et al. (2002) identified Chrons M41 to M25 off the coast of Morocco, reported that Seaward Dipping Reflectors (SDR) coincide with S1, and assigned an age of 170 Ma for that anomaly. Assuming a constant spreading rate of 21.9 mm/a over the 65-km distance from “Blake Spur” to M40 and S1 to M40, the time interval would be 2.97 m.y. and indicates that this early ridge jump occurred ca.170 Ma as postulated by Vogt for the western flank, and it is consistent with the S1 age for the eastern flank (Roeser et al., 2002).

Plate Reconstructions

Figures 12, 13, and 14 show North American and African plate reconstructions for 84 Ma, 120.6 Ma, and 154 Ma, coinciding with Chrons C34, M0, and M25. Plates are rotated with respect to the trace of the Mid-Atlantic Ridge using the calculated stage poles 03(Table 3). Gravity anomalies over fracture zones for C34 and M0 reconstructions (Figs. 12 and 13) are continuous across the Mid-Atlantic Ridge, adding confidence to the plate reconstruction. Inspection of the M25 reconstruction (Fig. 14) illustrates the differences in distance between “Jurassic Quiet Zone” provinces as well as the relative distance spanned by the Inner Magnetic Quiet Zone section during this time.

DISCUSSION

Blake Plateau

Seafloor spreading in the Central Atlantic is marked by the onset of post-rift sediment deposition in early Middle Jurassic, or ca. 185 Ma (Withjack et al., 1998). Dunbar and Sawyer (1989) suggested that the preexisting structural grain of the continental crust controlled the amount of continental extension prior to breakup. They concluded seafloor spreading begins first along segments that follow the structural grain, that extension prior to breakup is symmetrical, and that the amount of this extension is two to three times less than the amount of extension that crosses the preexisting structural grain. They also concluded that the total range of continental extension in the Central Atlantic is 200 km to over 600 km, and that this variability is largely controlled by the preexisting weaknesses within the structural grain and the ultimate orientation of the continental break. The largest extension along North America in the Central Atlantic is over 600 km in the region of the Blake Plateau, south of the Blake Spur fracture zone (Dunbar and Sawyer, 1989), which offsets Chrons M25 to M0 by 30–50 km (Fig. 5).

The continental boundary in the Blake Plateau suggested by Dunbar and Sawyer (1989) extends to Chron M32 mapped in this study. Inboard of this margin, we map the southernmost 280 km of Chron M40. Although Jurassic Magnetic Quiet Zone anomalies are difficult to correlate, our interpretation is consistent with the trend of the Blake Spur Magnetic Anomaly and the mapped limit of oceanic crust (Uchupi et al., 1984b), which are located ∼200 km landward of Dunbar and Sawyer's (1989) proposed southeastward extent of continental crust beneath the Blake Plateau. Seismic refraction data in this part of the Blake Plateau include lines interpreted by Ewing and Ewing (1959), Katz and Ewing (1956), and Sheridan et al. (1966). 04Table 4 summarizes seismic refraction data in and near the Blake Plateau. Except for the crust beneath Line 102, crustal thicknesses range from 5.9 km to 6.7 km. Line 102 lies over the northwest projection of the “Blake Spur Fracture Zone,” and the crustal thickness along this line is consistent with the crustal thickness of other fracture zones in the Atlantic Ocean, which is ∼2–3 km (Fox and Gallo, 1984).

The Blake Plateau overlies a broad basin that is ∼5–6 km deep overall, but exceeds 8 km just north of 30° N, 78° W (Crosby et al., 1984). Seismic refraction data from several profiles were combined and interpreted over an 800-km-long, east-west cross section, from the west coast of Florida at 30° N eastward to ∼76° W at 29° N (Fig. 15) (Sheridan et al., 1966). Inboard of seismic refraction Line 102 they interpreted a 2-km-thick, 5.3–5.5 km/s layer overlying basement to be volcanics, and a ridge (5.7–6.1 km/s) that rises ∼2.5 km along the outer edge of the Blake Plateau to be basement. The deepest horizon detected between these volcanics and the basement ridge, at 3–4 km, is defined by velocities ranging between 5.5 and 6.0 km/s. They suggested that it might be equivalent to Paleozoic basement rocks of Florida.

Dietz (1973) proposed that the Bahamas overlie the Early Jurassic mantle plume that produced the extensive tholeiitic intrusions and flows of the Central Atlantic Magmatic Province (CAMP). Marzoli et al. (1999) reported that the total aerial extent affected by the CAMP Plume was greater than 7 million square kilometers on North America, Africa, and South America. They also reported that this volcanism occurred over just a few million years and that the peak activity was 200 Ma, or ∼15 m.y. before seafloor spreading commenced in the Central Atlantic Ocean. However, if, as Dietz (1973) suggests, the Bahamas overlie seamounts produced by the CAMP Plume, then volcanic activity continued as seafloor was accreted in the Central Atlantic Ocean. Therefore, the presence of thick volcanics on the oceanic basement of the Blake Plateau seems reasonable. Although a comprehensive understanding of the history and makeup of the Blake Plateau is beyond the scope of this paper, we suggest that it is not possible to rule out the presence of oceanic crust beneath the Plateau. Therefore, based on similarities of anomalies along nine profiles, we tentatively extend Chron M40 over the Blake Plateau with dashed lines.

Refraction data support the existence of oceanic crust extending landward at least to the Blake Spur Magnetic Anomaly (Ewing and Ewing, 1959; Houtz, 1980; Katz and Ewing, 1956) 04(Table 4). Moho depths are not measured beneath the deep Carolina Trough and Blake Plateau. However interpretations of reflection data summarized by Withjack et al. (1998) indicate that a zone of Seaward-Dipping Reflectors (SDR) coincides with the East Coast Magnetic Anomaly from its southern end at ∼30° N along the North American continental shelf to ∼43° N, just south of Nova Scotia.

Chron Identification and Ridge Jumps

Unlike magnetic anomalies over M25 to M0 Chrons, “Jurassic Quiet Zone” anomalies are characterized by extremely rapid reversal rates and low amplitudes. Vogt et al. (1970) suggested that these low amplitudes could be the result of a period of rapid polar wander or that they formed at the magnetic equator. Sager et al. (1998) reported that M27 to M30 have been verified by magnetic stratigraphy, and that while magne-tostratigraphic data corresponding to M38 and older are lacking, short wavelength anomalies are similar to other paleointensity variations corresponding to periods of 300–150 k.y. They noted, however, that magnetization strength data are poorly constrained and they concluded that the Jurassic geomagnetic field behavior was unusual. If the modeled polarity reversals exist, then the geomagnetic polarity reversal rate is extraordinarily high: ∼12 per m.y. or 20% higher than the period between M25 and M26, which would then be the second highest reversal rate (Sager et al., 1998). They suggested that the anomalies could be produced by paleointensity fluctuations instead of geomagnetic polarity reversals.

Both ridge jumps described in this study are consistent in dimension and duration with other ridge jumps observed around the world. The relocation of seafloor spreading centers, or ridge jumps, has been documented along the Mid-Atlantic Ridge near the Ascension Fracture Zone (Brozena, 1986), seven locations west of the East Pacific Rise, including two ridge jumps currently under way on the East Pacific Rise (Luhr et al., 1985; Morton and Ballard, 1986; Mammerickx et al., 1988; Mammerickx and Sandwell, 1986), south of the Chilean Ridge (Mammerickx et al., 1988), and three locations in the north Pacific (Mammerickx et al., 1988). Between Mexico and the Galapagos Ridge, Luhr et al. (1985) document several ridge jumps over the past 12 Ma with jumps ranging from ∼500 km to 1100 km.

CONCLUSIONS

Combining extensive magnetic data sets with satellite-derived, free-air gravity data has allowed us to map several M-Series Chrons in detail, estimate asymmetric spreading rates, and interpret two ridge jumps in previously poorly constrained swaths of the Central Atlantic Ocean. We calculate new Euler poles: stage poles for the North American and Africa sides of the basin, and total reconstruction poles for Chrons C34 (84 Ma), M0 (120.6 Ma) and M25 (154 Ma). Measured distances on the North American and African flanks between the Mid-Atlantic Ridge and Chron C34, Chrons C34 and M0, and Chrons M0 and M25 reveal asymmetric spreading 05(Table 5). From 154 Ma to the present, we calculate three time intervals of overall asymmetric spreading: 10.5% to the west from Chrons M25 to M0 (154 Ma to 120.6 Ma), 3.5% to the east during the Cretaceous Magnetic Quiet Zone (120.6 Ma to 84 Ma), and 10% to the west since the end of the Cretaceous Magnetic Quiet Zone (84 Ma to present).

The mapped southernmost extent of Chron M40 on the North American flank of the ocean basin, extending into the Blake Plateau, suggests that the crust beneath the plateau is oceanic and that the ocean-continent crustal boundary lies along trend with the ocean boundary interpreted by Uchupi et al. (1984a; 1984b). The second vertical derivative of total-intensity magnetic anomalies reveals subtle anomalies over the Inner Magnetic Quiet Zone that are sub-parallel to the “East Coast” and “Blake Spur” (Behrendt and Grim, 1985), further supporting the existence of oceanic crust between these two anomalies.

Newly mapped M-Series Chrons in the Jurassic Magnetic Quiet Zone, and their positions with respect to the prominent Blake Spur Magnetic Anomaly off the continental shelf of North America and S1 anomaly off the continental shelf of Africa, support the identification of two ridge jumps: (1) a previously theorized eastward jump at ca. 170 Ma (Vogt et al., 1971; Vogt, 1973; 1986), and (2) a westward jump between M32 and M38 (159 Ma and 164 Ma). These ridge jumps, especially the latter, could correspond with the opening of the Gulf of Mexico.

*Present address: Bird Geophysical, 16903 Clan Macintosh, Houston, Texas 77084, USA; dale@birdgeo.com

We would like to thank Walter Roest at the Geological Survey of Canada for sending a copy of the Kroonvlag project, which is a ship-track magnetic data set (also known as the “Collette Data”).