Site 990 is located 28 nmi east of the East Greenland coast, within the southern drilling transect EG63. It was one of three drill sites planned to complete the stratigraphic sampling of the earliest volcanism along this margin (Figs. 2, 4). The site was located at the position of previous ODP Site 915 in order to more deeply penetrate the lava succession to test the hypothesis that Iceland-type oceanic crustal accretion and steady-state production of Iceland-type tholeiites were initiated within this stratigraphic interval. Another important objective at the site was to sample material suitable for precise radiometric and magnetostratigraphic age determinations in order to assess the magnitude of a suspected hiatus in volcanic activity, located between the Middle and Upper series lavas at Site 917.
Because the sedimentary section had been cored at Site 915 during Leg 152, Site 990 was washed to a depth of 182.0 mbsf and rotary cored below that level. Sediments were recovered in the interval 182.0Ð202.3 mbsf and subdivided into two lithologic units. According to ODP convention, these two units are termed lithologic Unit I and lithologic Unit II, even though data from Site 915 indicate that the material above 182 mbsf probably forms two additional lithologic units. As a result, we correlate lithologic Units I and II at Site 990 with lithologic Unit III at Site 915. The ages of both Site 990 units are unknown, but the ages of the overlying sediment and underlying basalt at Site 915 suggest an early Eocene age.
Lithologic Unit I is a calcite-cemented mixed-cobble conglomerate, dominated by clasts of altered basalt, gabbro, and dolerite; quartzite and siliciclastic siltstone form the remainder of the cobbles. The cobbles are generally rounded to well rounded and range in size from 4 to >12 cm in diameter. The matrix is a poorly sorted silty sand, with angular grains, sand-sized mudstone intraclasts, and calcite cement.
Lithologic Unit II directly overlies basalt and is a clayey volcaniclastic breccia, dominated by basaltic debris. Clasts in the breccia are predominantly angular and composed of dark yellowish brown, altered basaltic material. The matrix of the breccia is dominated by clay and iron oxides, probably derived from the alteration of basaltic material, and minor well-rounded silt to fine sand-sized quartz grains. The presence of poorly developed flow indicators, the repeated vertical changes between clast-supported and matrix-supported fabric, and the absence of macroscopic pedogenic features indicate that Unit II was deposited by a moving fluid, such as a matrix-rich debris flow, after a limited distance/energy of transport. The large size and the rounding of the clasts in lithologic Unit I suggest that this unit was deposited in a high-energy environment, possibly a high-gradient stream; a shallow, wave-influenced marine setting; or a fan delta.
Additional sedimentary material, apparently untransported, was recognized as red, brecciated to clayey material on the tops of flow units within the basalt basement. This material has been described as part of the igneous sequence, but reflects in situ alteration and soil development.
Thirteen flow units were recognized in the core recovered from the interval 212Ð325 mbsf on the basis of changes in phenocryst assemblage or the presence of weathered and/or vesicular flow tops. Lava flows fall into one of three types: aa, pahoehoe, and transitional. Pahoehoe flows dominate in the lower part of the drilled sequence, whereas aa flows are ubiquitous in the upper portion. The top of the volcanic section at this site (and theÊpreviously drilled Site 915) is deeply weathered and oxidized, indicating that eruption occurred subaerially with some time gap between successive flow units.
Flow units cored at Site 990 range from aphyric to highly olivine or plagioclase-olivine-clinopyroxene phyric basalt. The olivine content decreases upward in the section, whereas both grain size and flow thickness increase upward. There is a subtle but systematic compositional variation in trace-element contents from the base to the top of the sequence (i.e., decreasing Cr and Ni and increasing V, Nb, Zr, and Y). In general, the lavas are moderately evolved, with low trace-element abundances, and geochemically similar to the one unit recovered from Site 915 and all units at Site 918 (72 km to the east; Fig. 5). No lavas similar to the Upper Series units cored at Site 917 (3 km to the west) were found, indicating that the transition from the breakup-related series to the Iceland-type tholeiitic series that dominates the oceanic SDRS is abrupt and occurs over a stratigraphic interval of <100 m.
The basaltic rocks recovered at Site 990 exhibit numerous planar, primary magmatic features that consist of vesicular layers, elongated patches of filled vesicles, and widely developed diffuse, thin flow bands. Many of these magmatic features occur in an almost horizontal attitude. The only evidence of deformation consists of a relatively dense network of veins and, to a lesser extent, of fractures, some of which show the development of slickensides. The veins, usually 1Ð2 mm thick, are commonly lined and filled with green clay. Other minerals less commonly seen in veins include zeolite minerals, carbonates, native copper, and gypsum.
Paleomagnetic data for the Site 990 basalts reveal a magnetic reversal between the upper two normally magnetized flows and the lower 11 reversely magnetized flows. Integration with Leg 152 results suggests that the normal polarity interval may be correlated with either magnetochron C25n or C26n and the underlying reverse polarity interval with C25r or C26r. The discovery of normal polarity intervals, together with results from future radiometric dating, offers the promise of a precise chronology for East Greenland margin volcanism.
Measurements of index properties from half-round cores and discrete minicores correlate with the flow structure. Specifically,P-wave velocities and bulk densities vary from 2 km/s and 2.3 g/cm3 in altered, vesicular flow tops to 6 km/s and 3.0 g/cm3 within the central, more compact portions. From the high rates of recovery and detailed sampling, it is apparent that the often reported differences between velocity and density measurements on discrete core samples and estimates derived from seismic reflection or downhole logging result from preferential recovery of the compact, central flow material. Magnetic susceptibilities range from 100 * 10-5 to 5000 * 10-5 SI. Thermal conductivity values are similar to those measured at Hole 989B, namely ~2 W/m/K.