Journal of Conference Abstracts

A Petrological Traverse Along the Mid-Atlantic Ridge Across the Azores Hot Spot

C. H. Langmuir Lamont-Doherty Earth Observatory, Palisades, NY 10964

langmuir@ldgo.columbia.edu

J. Reynolds Lamont-Doherty Earth Observatory, Palisades, NY 10964

H. Bougault IFREMER BP70 29263 Plouzane Brest, France

T. Plank Lamont-Doherty Earth Observatory, Palisades, NY 10964

L. Dosso IFREMER BP70 29263 Plouzane Brest, France

D. Desonie Lamont-Doherty Earth Observatory, Palisades, NY 10964

E. Gier Lamont-Doherty Earth Observatory, Palisades, NY 10964

Y. Niu Lamont-Doherty Earth Observatory, Palisades, NY 10964

During the Fazar expedition, 213 sampling stations were occupied along the neovolcanic zone of the mid-Atlantic Ridge from 33° 10'N, south of the Hayes transform fault , to 40° 30'N, just south of the Kurchatov transform. This range of latitudes traverses the bathymetric and geochemical anomaly associated with the Azores hot spot, which crosses the ridge at 39° 00'N. Previously, the long wavelength gradient in petrology and geochemistry was known for this region from the work of Schilling (Schilling, 1975; Schilling et al. 1983) who had one or two dredges from most segments. Very small scale variabilty, primarily across the strike of the rift valley, had been investigated in the FAMOUS and AMAR areas. No systematic sampling of the neovolcanic zone along complete second order segments had been carried out in this region. The recovered samples allow the investigation of two problems on two different scales.

(1) Geophysical models of ridge segmentation and mantle upwelling have simple and concrete predictions for basalt chemical compositions. For example, active upwelling in the centers of segments should lead to larger extents of melting and lower mean pressures of melting compared to passive upwelling beneath thicker lithosphere at segment margins. Are there systematic changes in basalt chemistry along ridge segments, and what constraints do they place on mantle flow and melting models on the segment scale?

(2) Regional depth varies from 1200 to 3000 meters for this portion of the MAR. The Azores region is known to be anomalous in a global context in terms of its depth-Na₈-Fe₈ relationships. Are there chemical systematics with depth in this region, and what do they reveal about mantle temperature, mantle heterogeneity or crustal processes?

To explore systematic variations in chemical composition that correlate with geographic position within a segment, all data from a given segment were plotted to determine the apparent slopes of liquid lines of descent. The slopes change as the hot spot influence increases, due primarily to the effects of water on the liquid line of descent. After correction for fractionation, most segments do not show changes in chemical composition that correlate regularly with geographical position within the segment . This conclusion is not simple, however, because of two effects that complicate interpretation of the data. First, as is well known, samples from near offsets tend to be more fractionated, and therefore it is difficult to compare samples of the same MgO content. Second, mantle hetergoeneity is a very strong influence in this region, even down to small scales. The occurrence of basalts of very different levels of incompatible element enrichment in the same segment makes it difficult to use the basalts to unambiguously investigate melting processes. However, if the signal were as large as would be predicted from the geophysical models, it should be very pronounced. Instead, in this region, the petrological signature suggests that basalts from the centers of segments are not derived by significantly greater extents of melting than those from the edges of segments. In some cases, gradients in major element chemistry seem to supersede segment boundaries.

The data suggest a large amount of variability can occur at the smallest scales within the MAR rift valley. Rock cores, which probably sample an area of less than 100m², not infrequently brought up two distinct glass compositions, and in at least one case, three compositions ranging from enriched to depleted. These results document a very fine scale of flow variability on the MAR floor. Individual melts of very different chemical compositions can be erupted in very close proximity. Therefore the "local variability" on the MAR is not a regular segment scale phenomenon-- it occurs on much smaller scales within the rift valley.

The one regularity on a segment scale is the tendency for enriched basalts to occur at segment centers (along with less enriched basalts). As the Azores is approached, the gradient in incompatible element ratios is not steadily progressive Instead, highly enriched lavas begin to appear at segment centers, until ultimately only such lavas are present very close to the hot spot. Preferential occurrence of enriched MORB at slow-spreading segment centers seems to be a fairly common phenomenon. Perhaps there is vertical zonation in the mantle beneath these segments, and the centers of segments preferentially tap deeper levels.

The regional chemical picture, looking at average chemistry along strike, has much cleaner results. One unexpected results is the excellent correlation between average depth and average MgO content. As the ridge shallows, MgO gets lower, reaching a mean value of only 6.2% MgO adjacent to the Azores. There are smooth average gradients in fractionation -corrected chemical composition going over the Azores hot spot, with a pronounced secondary peak centered at 35°N. This secondary peak is what causes the very asymmetrical profile of chemical compositions around the Azores. Once this peak is removed, the chemical variations are symmetrical, and the Azores influence is much narrower.

The Azores hot spot produces distinctive major element compositions that are not consistent with the global correlations of depth with basalt chemistry. Once the 35°N anomaly is removed, there are the opposite depth-chemistry relationships. around the Azores. The shallowest basalts have high Na₂O and low FeO. This anomalous chemical signal is very regular with latitude. It can be quantified by subtracting the major element signal that would be expected from the axial depth if the chemistry were normal. There are two important aspects to this distribution. First, the low Fe and high Na clearly indicate an important component of major element heterogeneity. The simplest model is one of a depleted mantle that has been enriched in Na and incompatible elements. Second, there is an exceptionally depleted zone symmetrically around the hot spot center. Such a depleted zone is present around other hot spots as well. This zone can be modeled as a consequence of 3-D mantle movement associated with the hot spot accompanied by vertical melt movement. The clear conclusions from this region are the importance of mantle heterogeneity as well as mantle temperature at both long and short length scales, and the lack of segment scale variability that would be consistent with simple geophysical models of segment scale active upwelling. The large variability in basalts at the smallest sampling scale suggests one important aspect of magma generation along the MAR is the production oh highly heterogeneous, small flows. In combination with dike propagation from the centers of segments, this could lead to highly non-systematic spatial variations within a segment.