Journal of Conference Abstracts

Influence of Spreading Rate, Magma Flux and Tectonic Setting on Basalt Crystallization and Assimilation Beneath Mid-Ocean Ridges

Peter J. Michael Dept. of Geosciences, University of Tulsa, 600 S. College Ave., Tulsa, OK 74104, USA

GEOS_PJM@centum.utulsa.edu

Winton Cornell Dept. of Geosciences, University of Tulsa, 600 S. College Ave., Tulsa, OK 74104, USA

Introduction

We have compiled analyses of Chlorine and major elements for 400 mid-ocean ridge basalts (MORB) (250 new analyses) globally to examine the influence of spreading rate, tectonics and magma flux on assimilation and crystallization beneath the Mid-Ocean Ridge (MOR). Cl and Cl/K in MORB are good indicators of assimilation of hydrothermally influenced material (Michael and Schilling, 1989) because Cl is low in the mantle and high in hydrothermal systems. Major element modelling is used to describe the extent and depth of crystallization (Grove et al., 1993; Danyushevsky et al., in press). The northern MAR provides critical observations in our study because it displays large variations in axial depth, axial profile, crustal thickness and extent of melting. Much of the ridge is deep and has a well-developed axial valley. MORB along these segments have formed by low or moderate extents of melting (Klein and Langmuir, 1987). In contrast, Reykjanes and Kolbeinsey Ridges are shallow and lack a deep median valley. Their axial profiles are similar to those observed at fast and super fast spreading ridges. They have thicker crust and their MORB have formed by large extents of melting, suggesting that magma flux plays a role in ridge characteristics (Phipps Morgan et al., 1994).

Behavior of Cl and Cl/K

Based on ionic charges and radii, Cl/K should remain nearly constant during mantle melting and crystallization, and should be higher for enriched sources (Schilling et al., 1980). We observe that N-MORB from most slow spreading ridges such as MAR near Kane and 15°N and 31-46°S, Mid-Cayman Spreading Center (MCSC), Southwest Indian Ridge (SWIR) and Australian-Antarctic Discordance (AAD) have very low Cl/K (<0.02), while E-MORB from these ridges have Cl/K up to about 0.08 (Fig. 1). We propose that this is a mantle trend, and reflects no assimilation of hydrothermally altered material. The higher Cl and Cl/K of many MORB from the other suites (up to 1.0) are related to high-level assimilation (Michael and Schilling, 1989). Depleted, low-Cl MORB are most sensitive to increases of their Cl and Cl/K by contamination.

Fig. 1: K/Ti vs. Cl/K for MORB glasses. Depleted N-MORB have low K/Ti, enriched E-MORB have high K/Ti. Open symbols are slow-spreading ridges with MORB that have crystallized only at high pressure. Filled symbols are slow-spreading ridges with MORB that have crystallized at lower pressures. Other symbols are medium- to fast-spreading ridges. Curve is the proposed mantle trend (see text). Glasses lying above the mantle trend have assimilated Cl-rich material.

Crystallization pressure

Crystallization pressures can be compared qualitatively for moderately evolved basalt liquids by determining the pressure at which clinopyroxene, plagioclase and olivine would be saturated at roughly the same temperature (Danyushevsky et al., in press). A similar approach can be used to compare the liquid lines of descent (LLD) for different suites of MORB (Grove et al., 1993), but this gives slightly different results.

Observations and interpretations

The most consistent correlation we observe is between elevated Cl/K and suites that contain MORB that have crystallized at low pressure. This includes most MORB from super fast and fast spreading ridges (East Pacific Rise (EPR) 28°-35°S, EPR 9°-12°N) and many MORB from ridges with medium spreading rates (Galapagos 85°W and 95°W, Explorer Ridge). Notably, certain MORB from Reykjanes and Kolbeinsey Ridges also appear to have equilibrated at low pressure. These suites also contain MORB with Cl/K that is greater than the proposed mantle values (Fig. 1). Like the fast-spreading ridges, Reykjanes Ridge has erupted Fe-Ti basalts. In contrast, MORB from most slow-spreading segments (e.g., MAR near Kane, 15°N and 31-46°S; AAD, SWIR, MCSC) are invariably low in Cl and Cl/K. Their major element chemistry suggests that they have crystallized at higher pressures and have not crystallized substantially at low pressure. Within suites whose MORB display a range of crystallization pressures, there is not a 1:1 correlation between samples with high Cl/K and low crystallization pressure. Cl/K and crystallization pressure of individual glasses are not related to location within any segment.

In detail, Cl/K is highest (up to 1.0) in evolved MORB from the super fast-spreading southern EPR and the propagating Galapagos Ridge at 85°W. On these ridges, Cl/K ratios are typically greater than mantle values and increase regularly with decreasing Mg# (Fig. 2), suggesting that magmas lose considerable heat and continuously assimilate their surroundings while crystallizing in shallow, long-lived magma chambers. Additionally, the Cl content of the crust might be higher along these ridges from greater hydrothermal activity. Cl/K is slightly lower and not as well correlated with Mg# in MORB from dying ridges and from fast- and medium-spreading ridges like the EPR at 9°N. On Reykjanes and Kolbeinsey Ridges, Cl/K is high but is not correlated with Mg#, suggesting that magmas may crystallize and assimilate at shallow levels, but not in long-lived chambers.

Fig. 2: Mg# = 100 x (Mg/[Mg+Fe²⁺]): a measure of the extent of crystallization. Cl/K indicates assimilation. Symbols as in Fig. 1. The line at Cl/K=0.08 indicates the maximum value for the mantle. Data for slow-spreading high-pressure LLD ridges all fall below this line and are not shown for clarity.

A suite's average crystallization pressure is correlated with extent of melting (Fig. 3) and crustal thickness (not shown), suggesting that the depth at which a MORB crystallizes is related to magma flux and crustal thickness, which is consistent with geophysical evidence and thermal and mechanical models (Phipps Morgan et al., 1994).

Fig. 3: Average crystallization pressure for each suite, calculated using the model of Danyushevsky et al. (in press). Only MORB having <7.7% MgO were included in the average, since more primitive MORB may never have been saturated with clinopyroxene. Average extent of melting is estimated from Na_8.0 (Langmuir et al., 1993) using all MORB for which we have data. Major element data are adjusted to a common standard. Symbols as in Fig. 1.

References

Danyushevsky, L. et al., Contrib. Mineral. Petrol. (in press)

Grove, T.L., Kinzler, R.J. & Bryan, W.B., in Mantle Flow and Melt Generation at Mid-Ocean Ridges, eds. Phipps Morgan et al., 281-310 (1993).

Klein, E.M. & Langmuir, C.H., J. Geophys. Res. 92, 8089-8115 (1987).

Langmuir, C.H., Klein, E.M. & Plank, T., in Mantle Flow and Melt Generation at Mid-Ocean Ridges, eds. Phipps Morgan et al., 180-279 (1993).

Michael, P.J. & Schilling, J.-G., Geochim. et Cosmochim. Acta 53, 3131-3143 (1989).

Phipps Morgan, J., Harding, A., Orcutt, J., Kent, G. & Chen, Y.J., in: Magmatic Systems, ed. M. Ryan, 139-178 (1994).

Schilling, J.-G. Bergeron, M.B. & Evans, R., Phil. Trans. Roy. Soc. London 297, 147-178 (1980).