Structure of the Ethiopian lithosphere: Xenolith evidence in the Main Ethiopian Rift

Geochimica et Cosmochimica Acta, Vol. 69, No. 15, pp. 3889–3910, 2005Copyright © 2005 Elsevier Ltd

Printed in the USA. All rights reserved

doi:10.1016/j.gca.2005.03.043

Structure of the Ethiopian lithosphere: Xenolith evidence in the Main Ethiopian Rift

TYRONE O. ROONEY,1,* TANYA FURMAN,1 GEZAHEGN YIRGU,2 and DEREJE AYALEW2

1Department of Geosciences, The Pennsylvania State University, University Park, PA 16802, USA2Department of Geology and Geophysics, Addis Ababa University, P.O. Box 1176, Addis Ababa, Ethiopia

(Received August 23, 2004; accepted in revised form March 24, 2005)

Abstract—The lithospheric and sublithospheric processes associated with the transition from continental tooceanic magmatism during continental rifting are poorly understood, but may be investigated in the centralMain Ethiopian Rift (MER) using Quaternary xenolith-bearing basalts. Explosive eruptions in the Debre Zeyit(Bishoftu) and Butajira regions, offset 20 km to the west of the contemporaneous main rift axis, host Al-augite,norite and lherzolite xenoliths, xenocrysts and megacrysts. Al-augite xenoliths and megacrysts derived frompressures up to 10 kb are the dominant inclusion in these recent basalts, which were generated as small degreepartial melts of fertile peridotite between 15 and 25 kb. Neither the xenoliths nor the host basalts exhibit signsof carbonatitic or hydrous (amphibole � phlogopite) metasomatism, suggesting that infiltration of silicatemelts resulting in pervasive Al-augite dyking/veining dominates the regional lithospheric mantle. Recentgeophysical evidence has indicated that such veining/dyking is pervasive and segmented, supporting theconnection of these Al-augite dykes/veins to the formation of a proto ridge axis. Al-augite xenoliths andmegacrysts have been reported in other continental rift settings, suggesting that silicate melt metasomatismresulting in Al-augite dykes/veins is a fundamental processes attendant to continental rift

0016-7037/05 $30.00 � .00

development. Copyright © 2005 Elsevier Ltd

1. INTRODUCTION

As continental break-up progresses to maturity, intraplatemagmatism dominated by lithospheric contributions gives wayto ocean floor magmatism driven by athenospheric melting.This transition between continental rifting and sea-floor spread-ing is poorly constrained. The Main Ethiopian Rift (MER) liesin a unique transitional zone between the continental rifting ofEast Africa and the seafloor spreading of Northern Afar and theRed Sea. Xenoliths and mafic lavas from this transitional zonereveal direct insight into materials and processes associatedwith the shift from continental to oceanic magmatism.

In Ethiopia, the continental–oceanic transition is manifest bya shift from broad-based to focused strain, which is clearly seenin the Quaternary division of the central MER into discretevolcanic segments described previously as narrow en-echelonfault belts (Ebinger and Casey, 2001). Tectonic and magmaticactivity in this region is attributed to a deep-seated mantleplume that impacts the lithosphere beneath Afar (e.g., Schillinget al., 1992; Coulie et al., 2003; Courtillot et al., 2003). To thesouth, compositionally distinct volcanism in the Turkana areaand throughout the Kenya dome is derived from either a largeheterogenous mantle plume that encompasses Afar (Furman etal., 2004) or from a separate small plume (George et al., 1998).The MER thus provides an opportunity to study the continen-tal–oceanic transition and the process of plume-driven conti-nental break-up.

As the process of rifting thins the continental crust, it mustalso thin the subcontinental lithospheric mantle. Elsewhere inthe East African Rift system (EARS), the compositionallyheterogeneous sub-continental lithospheric mantle (SCLM) hasbeen an important source for rift magmatism (Macdonald,

* Author to whom correspondence should be addressed ([email protected]).

3889

1994; Furman, 1995; Rogers et al., 1998; Furman and Graham,1999; Späth et al., 2001). This volumetrically significant par-ticipation of the lithospheric mantle in magma genesis indicatesthat it may melt preferentially. Enrichment of this regionthrough metasomatism is capable of producing a source that ispredisposed to melting because of the growth of hydrousphases (e.g., amphibole and phlogopite) as well as anhydrousminerals (e.g., apatite) with lower P-T stability than the sur-rounding peridotite (Dawson and Smith, 1988; Rudnick et al.,1993). The heterogeneous nature of the subcontinental lithos-pheric mantle (evident on a scale �10 km) make it difficult tocharacterize from eruptive lava chemistry alone. Fragments ofthe lithospheric mantle in the form of mantle xenoliths found atrare localities within the East African Rift provide direct com-positional information (Dawson and Smith, 1988; Rudnick etal., 1993; Dawson, 2002). Such xenoliths also constrain themetasomatic history and the P-T conditions of melting whichare important parameters associated with the rifting process.

We describe two new xenolith suites and associated hostlavas from the Debre Zeyit (also known as Bishoftu) andButajira regions of central Ethiopia (Fig. 1). These new xeno-lith suites and host lavas provide more than simply a windowinto the composition of the sub-MER lithospheric mantle; theyare evidence of the regional processes associated with destruc-tion of lithospheric mantle during the continental–oceanicmagmatic transition. Prior to this study, the three mantle xeno-lith localities associated with flood basalt emplacement andincipient rifting were located in the Lake Tana region (Roger etal., 1997; Conticelli et al., 1999; Roger et al., 1999; Ayalew etal., 2005), the Wollega region (Conticelli et al., 1999) and theSidamo area (Bedini et al., 1997; Bedini and Bodinier, 1999;Conticelli et al., 1999; Lorand et al., 2003). The Lake Tana andWollega xenoliths, both on the Ethiopian Plateau, were em-placed some 300 km west of the main rift axis and represent
fragments of the lithospheric mantle (lherzolites, dunites, har-

and in

3890 T. O. Rooney, T. Furman, G. Yirgu, and D. Ayalew

zburgites and hornblende-harzburgites) that were entrained bytheir host lavas (Conticelli et al., 1999). The Sidamo suite waserupted along the rift axis �500 km south of the central MER;hence well outside the proposed oceanic–continental transitionzone. These southern-rift xenoliths (lherzolites, dunites, har-zburgites and websterites) are similar to those found on thenorthern plateau and also represent fragments of the litho-spheric mantle (Conticelli et al., 1999). The Debre Zeyit andButajira suites, in contrast, provide important local constraintson intensive parameters associated with the rifting process,including the pressure and temperature of magma generation,the local geothermal gradient and the interaction betweenplume heat source and the SCLM.

2. STUDY LOCATION

This study concentrates on the western marginal graben of

Fig. 1. Map of the study area showing the approximat(shaded) from the central MER (unpatterned). Dashed linecircle � silicic volcanic complex of Ziqualla; solid linesgray, italic text) and Quaternary basaltic flows and cindMazzarini et al. (1999). Stars � samples; boldface italic �inferred faults (Mohr, 1961), respectively. (B) Butajira(Benvenuti et al., 2002). Sample localities as in (A). Faultsshown (Benvenuti et al., 2002).

the central MER known as the Silti-Debre Zeyit Fault Zone

(WoldeGabriel, 1988). The region is marked by recent basalticactivity at Debre Zeyit and Butajira (Fig. 1), separated by thesilicic centre of Ziqualla (1.28 to 0.85 Ma; Morton et al., 1979).Quaternary volcanism in the central sector of the MER isrestricted to this fault zone and its eastern equivalent, theWonjii Fault Belt (Mohr, 1967; WoldeGabriel, 1988) indicat-ing two parallel contemporaneous zones of activity. The faultzone is intersected between 9o05N and 8o20N by the cross-rifttrending Yerer-Tullu Wellel volcanotectonic lineament (Abebeet al., 1998). The currently active north–south structures withinthe MER combined with east–west extension along this faultzone results in a regional stress field conducive to volcanicactivity and is expressed as a local extensional axis along thewestern rift margin (Abebe et al., 1998; Mazzarini et al., 1999).The well-preserved Quaternary silicic volcanic centre of Zi-qualla and the less well-preserved silicic centre of Bede Gebabe

on of the rift wall (dotted line) separating the highlandsternary volcanic zones of Debre Zeyit and Butajira; gray

r roads. (A) Debre Zeyit region with silicic centres (lights (dark gray, unnamed). All flows and volcanoes after

ith/megacryst locality; solid and dotted lines � faults andwith Quaternary Butajira volcanics shown in dark grayferred faults (solid and dashed lines, respectively) are also

e locati� Qua

� majoer conexenol

region

(0.36 Ma; Mazzarini et al., 1999) predate the youngest phase of

3891Xenoliths in the Ethiopian continental-oceanic transition

basaltic volcanism (Gasparon et al., 1993), expressed as cindercones, maars and basaltic flows that are at times parasitic on theslopes of the silicic volcanoes. This most recent phase ofvolcanism has been dated at 0.13 Ma in Butajira (WoldeGabrielet al., 1990) and 0.11 Ma in Debre Zeyit (Chernet et al., 1998).Recent eruptive features in the Debre Zeyit region exhibit analignment that passes into known faults, while the volcanicclusters in the Debre Zeyit and Butajira regions are rooted onextensional fractures (Mohr, 1961; Korme et al., 1997).

The Debre Zeyit area has been investigated in two earlierstudies (Gasparon et al., 1993; Mazzarini et al., 1999), butxenoliths have not been reported previously. The centralMER has recently been investigated using numerous geo-physical techniques under the auspices of EAGLE (EthiopiaAfar Geoscentific Lithosperic Experiment). Seismic workconsisted of a 400 km wide-angle reflection/refraction pro-file along the rift axis and a series of fan profiles for 3Dcoverage combined with a wide-angled reflection/refractionexperiment that provided a detailed crustal and upper mantlevelocity model along a 400 km transect across the NorthernEthiopian rift. A network of broadband and short periodinstruments was used to reveal 3D variations in crust andmantle structure via tomographic (shallow and deep) andanisotropic analysis of teleseismic, regional and local earth-quake waveforms. This extensive database provides a back-drop against which magmatic processes inferred by geo-chemical data can be evaluated effectively.

3. METHODS

Multiple samples were collected from twenty-one localities of cindercones and flows from the Debre Zeyit and Butajira region (Fig. 1). Thealignment of the cinder cones, which is broadly the same as the localrift valley alignment (NE-SW), dictated the locations of sampling.Xenoliths and megacysts were identified in multiple samples from eachof nine distinct localities (Fig. 1).

Elemental analysis of minerals in mafic lavas and xenoliths from theButajira (BJ series) and Debre Zeyit (DZ series) volcanics were un-dertaken at The Pennsylvania State University using a Cameca SX-50electron microprobe equipped with 4 wavelength spectrometers andone energy dispersive spectrometer (Tables 1–4). The analyses used a12 nA sample current and a 15 kV accelerating voltage. The width ofthe electron beam was 20 �m for analysis of all minerals excludingspinels where a beam width of �1–3 �m was used. Counting timeswere 20 s per element peak and 10 s background for major elementswhile minor elements (0.1%–1%) had a counting time of 60 s per peakand 30 s background.

Twenty-one samples (11 from Debre Zeyit and 10 from Butajira)were chosen for whole rock major, minor and trace element analy-sis. The samples were cut into slabs, trimmed to remove surfacealteration and polished to remove saw marks. These slabs were thencrushed using a porcelain jaw-crusher and powdered with a tung-sten-carbide disk mill. These powders were prepared and dissolvedusing the methods outlined in Knight (2002). Samples were ana-lyzed at Duke University by G. Dwyer using a VG PlasmaQuad-3ICP-MS for trace elements (including REE) and an ARL-FisonsSpectraspan 7 DCP for major elements, Sr and Ba. Natural basaltrock standards were used. For trace element analysis, inter-runprecision based on duplicate analysis is generally better than 2.5%while intra-run precision based on duplicate analysis of standards isgenerally better than 4%. Data relevant to this publication are
presented in Table 6.
4. RESULTS

4.1. Petrography

4.1.1. Host Lavas

Samples from lava flows are broadly hypocrystalline andseriate to porphyritic textured with occasional glomeroporphy-ritic intergrowths of plagioclase and clinopyroxene. Scoriasamples contain phenocrysts up to 1 cm in length that arefrequently cracked. Phenocrysts in both scoria and lava flowsconsist of various combinations of olivine, pyroxene and felds-par. Olivine is consistently the first phenocryst phase to growand has abundant Cr-spinel inclusions. Clinopyroxene is thesecond phenocryst phase to nucleate and is often pervasivelyresorbed by plagioclase. An-rich feldspars are the last silicatephase to crystallize, with magnetite and sparse ilmenite crys-tallizing in a groundmass of olivine, clinopyroxene and plagio-clase.

Olivine crystals range from euhedral to anhedral, with skel-etal textures also common. Plagioclase overgrowths on olivineoccur in many samples but associations of olivine with clinopy-roxene are restricted to rare inclusions in large clinopyroxenephenocrysts. Olivine phenocrysts (0.5–10 mm) are observed inall samples and are generally normally-zoned with core com-positions ranging from Fo68–90 (Fig. 2). Matrix olivine (0.15–0.5 mm) is of composition Fo71–82 and Fo68–81 in Debre Zeyitand Butajira respectively. The cores of olivine phenocrysts insamples DZ-1006, DZ-1013 and DZ-1014 have Fo contentsthat are significantly lower than those measured in the matrix ofthe host lava (Table 1). Reverse zoning is present occasionally(e.g., DZ-1007, DZ-1008, BJ-1045). The presence of euhedralFo�85 olivine in most samples indicates a primitive parentalmafic liquid that has undergone minor olivine fractionation.

Clinopyroxene phenocrysts range in size from 0.5–15 mm,are generally brown to almost colorless under transmitted light,and display a limited compositional range that straddles thediopside-augite boundary (Fig. 2). A second population ofpyroxene that is green and pleiochroic under transmitted lightoccurs as a phenocryst in BJ-1051 while a similar green butnonpleiochroic variety is observed in BJ-1044. The composi-tional zoning observed optically is not manifest in significantchemical heterogeneity (Table 2). Pervasive intergrowths ofAn-rich plagioclase feldspar and clinopyroxene are common inmost samples. This association is observed both in the matrix(predominantly a 0.01 mm intergrowth of diopside-augite andlabradorite) and also as a resorption texture in larger pheno-crysts. Rarely, large clinopyroxene and plagioclase crystalsoccur together with ophitic texture (BJ-1044). Occasional cli-nopyroxenes (0.75–1 mm) that are not associated with plagio-clase feldspars appear stressed, exhibiting wavy extinction andmosaic texture (e.g., DZ-1009, BJ-1044).

Individual localities have substantially heterogeneous plagio-clase feldspar populations. Feldspar phenocrysts are generallysmall (0.25 mm), compositionally An60–70 (Fig. 3), and fresh withoccasional larger crystals up to 2 mm that exhibit minor zoning(Table 3). Pervasive intergrowths with and replacement of cli-nopyroxene occur in all samples except BJ-1044 as noted above.Feldspar phenocrysts are generally either inclusion-free or containsmall olivine crystals. However, an inclusion of apatite was ob-
served in a large phenocryst in sample DZ-1006.


Table 1. Representative microprobe analysis of olivine from Butajira and Debre Zeyit.a

Location Analysis MnO FeO NiO Cr2O3 CaO SiO2 MgO Total n Fo Fa Tpb Mg#c

Host lava1003 Euhedral center 0.13 9.59 0.26 0.11 0.28 42.70 47.82 100.89 89.77 10.10 0.14 89.88

0.03 0.25 0.03 0.01 0.02 0.39 0.29 0.76 121003 Euhedral edge 0.18 13.12 0.22 bdl 0.27 42.12 44.97 100.88 85.77 14.04 0.20 85.93

0.01 0.90 0.02 bdl 0.01 0.50 0.78 0.49 31006 Core 0.20 13.64 0.22 bdl 0.27 40.59 44.47 99.38 85.14 14.64 0.22 85.32

0.02 0.31 0.05 bdl 0.01 0.14 0.32 0.56 61006 Edge 0.25 15.65 0.21 bdl 0.28 40.21 42.89 99.50 82.78 16.94 0.28 83.00

0.02 1.20 0.01 bdl 0.01 0.27 1.05 0.46 61006 Core 0.28 17.15 0.12 bdl 0.21 40.12 41.95 99.82 81.10 18.59 0.31 81.34

0.03 1.12 0.01 bdl 0.01 0.41 0.96 0.25 61006 Edge 0.22 14.79 0.25 bdl 0.26 40.79 44.01 100.31 83.95 15.82 0.23 84.14

0.02 0.21 0.02 bdl 0.00 0.21 0.14 0.24 61007 Unzoned 0.24 15.73 0.17 bdl 0.23 40.46 43.01 99.84 82.77 16.97 0.26 82.98

0.03 0.97 0.06 bdl 0.07 0.35 0.94 0.43 121007 Core 0.20 13.05 0.21 bdl 0.28 41.04 45.15 99.93 85.86 13.92 0.22 86.04

0.03 0.12 0.03 bdl 0.00 0.07 0.31 0.40 31007 Edge 0.31 18.85 0.16 bdl 0.27 39.77 40.42 99.77 79.00 20.66 0.34 79.26

0.07 2.61 0.07 bdl 0.00 0.61 1.97 0.44 31007 Skeletal 0.18 11.16 0.24 bdl 0.31 41.26 46.38 99.52 87.94 11.87 0.19 88.10

0.01 0.29 0.01 bdl 0.01 0.17 0.30 0.18 31008 Euhedral core 0.18 11.95 0.28 bdl 0.26 40.14 45.40 98.21 86.96 12.84 0.19 87.13

0.05 0.78 0.05 bdl 0.03 0.53 0.55 0.62 121008 Euhedral edge 0.20 14.13 0.19 bdl 0.28 40.21 43.72 98.73 84.46 15.31 0.22 84.65

0.04 0.98 0.05 bdl 0.02 0.54 0.70 0.33 91009 Core 0.40 13.71 0.51 0.28 0.48 40.98 44.95 101.32 85.02 14.54 0.43 85.39

0.03 0.22 0.02 0.02 0.01 0.18 0.05 0.41 31009 Rim 0.41 21.44 0.11 bdl 0.35 39.06 37.27 98.64 75.25 24.28 0.47 75.60

0.05 1.52 0.01 bdl 0.06 0.54 2.33 1.31 31042 Core 0.33 22.60 bdl bdl 0.17 38.66 37.51 99.27 74.46 25.16 0.37 74.73

0.03 0.18 bdl bdl 0.01 0.26 0.45 0.88 31042 Rim 0.42 25.31 bdl bdl 0.20 38.00 35.13 99.07 70.88 28.64 0.49 71.21

0.04 0.88 bdl bdl 0.02 0.25 0.87 0.66 31043 Skeletal 0.25 17.66 0.15 bdl 0.22 40.13 41.68 100.09 80.57 19.15 0.27 80.79

0.02 0.64 0.03 bdl 0.02 0.17 0.41 0.52 91043 Core 0.17 11.52 0.21 bdl 0.27 41.32 46.28 99.77 87.58 12.23 0.19 87.74

0.02 0.18 0.04 bdl 0.01 0.07 0.25 0.49 31044 Core 0.29 19.20 0.13 bdl 0.25 39.68 39.34 98.89 78.25 21.43 0.32 78.50

0.01 0.67 0.02 bdl 0.02 0.07 0.35 0.40 31044 Edge 0.33 21.21 0.10 bdl 0.27 39.20 38.14 99.25 75.94 23.69 0.37 76.21

0.01 0.13 0.02 bdl 0.02 0.18 0.23 0.17 31044 Core 0.21 15.74 0.26 bdl 0.21 40.11 42.15 98.69 82.49 17.28 0.24 82.68

0.01 0.05 0.02 bdl 0.01 0.08 0.29 0.40 31044 Edge 0.28 19.87 0.11 bdl 0.26 39.40 38.83 98.75 77.45 22.23 0.32 77.69

0.02 0.23 0.01 bdl 0.01 0.23 0.57 0.57 31045 Core 0.19 13.47 0.28 0.15 0.26 40.29 45.37 100.00 85.55 14.24 0.20 85.72

0.01 0.29 0.07 0.13 0.02 0.31 0.33 0.28 31045 Edge 0.22 16.17 0.15 bdl 0.26 39.96 43.51 100.27 82.55 17.21 0.24 82.74

0.02 0.35 0.01 bdl 0.02 0.26 0.61 0.75 31047 Core 0.24 18.06 0.17 bdl 0.19 39.91 41.98 100.55 80.35 19.38 0.26 80.56

0.01 0.32 0.04 bdl 0.01 0.14 0.26 0.13 31047 Edge 0.41 25.50 0.09 bdl 0.27 38.44 35.80 100.51 71.11 28.42 0.46 71.44

0.02 1.25 0.01 bdl 0.05 0.47 1.01 0.40 31049 Core 0.25 18.66 0.08 bdl 0.17 39.92 41.49 100.56 79.65 20.09 0.27 79.85

0.02 0.29 0.01 bdl 0.01 0.21 0.39 0.85 31049 Edge 0.48 25.91 0.04 bdl 0.26 38.46 34.95 100.11 70.22 29.23 0.55 70.62

0.09 2.16 0.01 bdl 0.02 0.34 1.85 0.16 31053 Core 0.20 14.31 N/A N/A 0.21 40.93 44.19 99.84 84.44 15.34 0.22 84.63

0.02 0.33 N/A N/A 0.02 0.26 0.61 0.44 31053 Edge 0.32 21.66 N/A N/A 0.24 39.19 37.49 98.91 75.25 24.39 0.36 75.52

0.08 3.65 N/A N/A 0.04 0.72 3.22 0.17 31053 Skeletal 0.27 18.15 N/A N/A 0.23 40.38 41.53 100.55 80.08 19.63 0.29 80.31

0.08 1.51 N/A N/A 0.01 0.60 1.46 0.36 2Group A xenoliths and xenocrysts

1006 Core 0.97 33.80 bdl bdl 0.15 36.28 27.84 99.03 58.80 40.04 1.16 59.480.04 0.20 bdl bdl 0.00 0.14 0.19 0.45 4

1007 In xenolith 0.34 22.18 bdl bdl 0.16 38.03 37.83 98.55 74.97 24.65 0.38 75.240.04 0.44 bdl bdl 0.02 0.46 0.31 0.58 11

1008 Rounded olivine 0.49 30.34 bdl bdl 0.09 36.50 31.73 99.15 64.72 34.71 0.56 65.08


Small opaque oxides (� 0.01 mm) form inclusions withinolivine and comprise a significant portion of the matrix. Chrome-rich (� 38% Cr2O3) and alumina-rich (� 55% Al2O3) spinelsform inclusions in olivines in samples from both Debre Zeyit andButajira (Table 4). Sparse larger crystals (up to 1 mm) that fingerinto the matrix include magnetite in BJ-1049 and magnetite andrarer ilmenite in BJ-1042 and BJ-1045.

Table 1

Location Analysis MnO FeO NiO Cr2O3

0.03 0.14 bdl bdl1008 Megacryst Core 0.43 26.25 bdl bdl

0.04 0.22 bdl bdl1008 Megacryst Edge 0.34 21.94 bdl bdl

0.13 4.29 bdl bdl1009 In xenolith 0.21 14.17 0.19 bdl

0.03 0.38 0.03 bdl1013 In xenolith 0.66 30.65 N/A N/A

0.08 0.18 N/A N/A1044 In xenolith 0.28 19.68 0.12 bdl

0.03 0.58 0.03 bdl1047 Core 0.66 32.19 bdl bdl

0.01 0.32 bdl bdl1047 Edge 0.52 28.00 bdl bdl

0.05 1.18 bdl bdl1051 In xenolith 0.34 21.78 bdl bdl

0.02 0.28 bdl bdl1003 In xenolith 0.32 21.25 bdl bdl

0.02 1.67 N/A N/A1003 In xenolith 0.27 21.10 N/A N/A

0.03 0.91 N/A N/AGroup C xeno

1007 Unzoned 0.26 15.99 N/A bdl0.05 1.27 N/A bdl

1008 Unzoned 0.15 10.54 0.34 0.100.02 0.15 0.03 0.01

1043 Core 0.12 9.05 0.41 0.140.01 0.18 0.01 0.01

1043 Edge 0.19 13.99 0.26 bdl0.04 2.65 0.09 0.02

1044 Core 0.14 9.95 0.39 0.140.01 0.08 0.00 0.04

1044 Edge 0.28 20.30 0.07 bdl0.04 0.53 0.00 bdl

1045 Core 0.17 10.91 0.30 bdl0.03 0.22 0.03 bdl

1045 Edge 0.20 14.10 0.25 bdl0.04 3.22 0.05 bdl

a All results presented as wt%. N/A � not analyzed; bdl � below db Tp is the tephroite component (Mn/(Mn � Mg � Fe)) * 100.c Mg# � (Mg/(Mg � Fe)) * 100.

Fig. 2. Core compositions of pyroxene (squares) and olivine ((A) Phenocrysts. (B) Al-augite xenocrysts.

4.1.2. Group A: Al-Augite Xenoliths, Xenocrysts, andMegacrysts

Xenoliths (5–30 mm) and megacrysts (up to 30 mm) areobserved in many locations within the study area (DZ-1003,DZ-1005, DZ-1006, DZ-1007, DZ-1008, DZ-1009, DZ-1013,DZ-1014, BJ-1042, BJ-1043, BJ-1044, BJ-1045, BJ-1047, BJ-

inued)

SiO2 MgO Total n Fo Fa Tpb Mg#c

0.48 0.25 0.39 337.12 35.17 99.14 70.14 29.37 0.49 70.480.14 0.13 0.19 3

38.09 38.03 98.62 75.26 24.35 0.39 75.540.69 3.71 0.21 3

40.83 44.30 99.96 84.59 15.18 0.23 84.780.08 0.27 0.52 3

37.58 31.25 100.25 64.02 35.22 0.76 64.500.01 0.55 0.84 2

39.38 39.15 98.87 6 77.77 21.92 0.31 78.000.23 0.49 0.36

36.71 29.97 99.62 62.06 37.31 0.77 62.390.08 0.18 0.31 3

37.48 33.60 99.78 67.73 31.67 0.60 68.140.32 1.09 0.36 3

40.09 37.52 99.90 75.14 24.47 0.39 75.420.21 0.20 11

39.62 37.35 98.81 75.53 24.10 0.37 75.800.59 2.18 0.80 3

40.35 38.18 100.10 76.11 23.59 0.30 76.330.43 1.01 0.63 6

d xenocrysts40.18 40.96 97.63 81.80 17.91 0.29 82.030.44 1.03 0.38 18

41.39 46.28 99.07 88.52 11.31 0.17 77.310.17 0.34 0.54 6

41.83 48.21 100.04 90.37 9.51 0.12 90.470.17 0.11 0.32 3

40.80 44.12 99.59 84.73 15.07 0.21 84.900.32 2.25 0.06 3

41.38 46.88 99.16 89.23 10.62 0.15 89.360.21 0.28 0.45 3

39.60 38.92 99.43 77.12 22.56 0.32 77.360.27 0.36 0.58 3

41.28 47.61 100.55 88.46 11.37 0.18 88.610.22 0.19 0.64 3

40.45 43.52 99.05 84.25 15.52 0.22 84.610.45 4.83 1.67 3

n limit (0.1%). Italicized numbers are standard deviation.

. (Cont

CaO

0.000.160.010.220.080.250.000.120.010.260.010.100.010.180.030.160.020.260.070.190.02liths an

0.240.060.260.010.280.010.230.010.280.010.260.000.300.020.540.49

etectio

circles) crystals in mafic Butajira and Debre Zeyit lavas.


Table 2. Representative microprobe analysis of pyroxene from Butajira and Debre Zeyit.a

Location Analysis Cr2O3 MnO TiO2 K2O FeO1 CaO Na2O Al2O3 SiO2 MgO Total n Wo En Fs Ac

Host lava1003 Sieve 0.13 0.13 0.44 bdl 6.97 20.92 0.60 2.63 53.68 14.12 99.62 44.29 41.60 11.81 2.30

0.02 0.02 0.09 bdl 0.21 0.21 0.03 0.40 0.50 0.15 0.33 61003 Orthopyroxene bdl 0.34 0.45 bdl 16.81 1.55 0.07 3.27 53.27 23.28 99.03 3.25 68.11 28.36 0.27

bdl 0.05 0.08 bdl 0.24 0.04 0.01 0.10 0.44 0.17 0.42 61006 Rounded Cpx Core 0.15 0.23 1.18 bdl 8.05 20.86 0.38 3.46 50.88 14.29 99.47 43.56 41.52 13.49 1.43

0.15 0.07 0.14 bdl 1.25 0.47 0.09 0.56 0.80 0.81 0.47 61006 Rounded Cpx Edge 0.20 0.24 1.53 bdl 7.93 21.02 0.39 4.61 49.69 13.53 99.14 44.78 40.13 13.57 1.51

0.15 0.07 0.24 bdl 0.82 0.39 0.06 0.76 0.76 0.65 0.52 61007 Core resorbed 0.34 bdl 1.21 bdl 5.67 20.92 0.45 4.90 51.18 14.48 99.15 45.08 43.43 9.71 1.77

0.02 bdl 0.24 bdl 0.52 0.06 0.09 0.17 0.51 0.49 0.61 31007 Edge resorbed 0.47 0.15 1.60 bdl 6.73 21.18 0.35 4.31 51.10 14.11 100.00 45.22 41.91 11.52 1.35

0.14 0.03 0.12 bdl 0.56 0.39 0.04 0.75 0.67 0.27 0.10 31008 Core bdl 0.14 2.12 bdl 8.63 19.73 0.72 8.11 46.92 12.42 98.79 43.71 38.30 15.11 2.88

bdl 0.04 0.10 bdl 0.17 0.04 0.03 0.03 0.21 0.12 0.20 31008 Edge 0.57 0.20 1.32 bdl 6.77 21.15 0.38 4.56 49.62 14.11 98.68 45.09 41.86 11.59 1.47

0.16 0.06 0.04 bdl 0.41 0.71 0.09 1.04 0.89 0.07 0.21 31008 Medium sized 0.28 0.19 1.02 bdl 7.04 19.16 0.69 8.12 48.72 14.19 99.39 41.93 42.97 12.31 2.79

0.01 0.05 0.05 bdl 0.10 0.13 0.03 0.12 0.16 0.08 0.17 31009 Core 0.50 0.13 1.57 bdl 6.02 21.67 0.34 5.86 49.41 13.59 99.10 47.08 41.10 10.47 1.35

0.01 0.03 0.07 bdl 0.03 0.08 0.03 0.09 0.21 0.06 0.07 31009 Edge 0.58 bdl 1.36 bdl 5.64 21.55 0.32 5.05 50.43 14.11 99.05 46.60 42.47 9.66 1.27

0.17 bdl 0.22 bdl 0.21 0.03 0.03 0.64 0.59 0.32 0.25 31043 Sieve Core bdl 0.13 1.57 bdl 7.29 20.36 0.34 5.65 50.02 14.19 99.55 43.75 42.44 12.47 1.34

bdl 0.02 0.15 bdl 0.23 0.12 0.04 0.56 0.68 0.31 0.36 31043 Sieve Edge 0.42 0.16 1.35 bdl 6.86 20.32 0.35 4.30 50.95 14.61 99.31 43.43 43.45 11.76 1.36

0.04 0.05 0.15 bdl 0.33 0.23 0.02 0.20 0.30 0.22 0.34 31044 Core 0.51 0.13 2.41 bdl 7.70 20.10 0.46 6.72 47.39 12.85 98.27 44.73 39.80 13.61 1.86

0.03 0.05 0.12 bdl 0.16 0.17 0.05 0.46 0.43 0.17 0.06 31044 Edge 0.19 0.19 2.89 bdl 8.58 20.29 0.45 5.59 47.69 12.40 98.28 44.86 38.16 15.16 1.82

0.01 0.06 0.68 bdl 0.31 0.37 0.07 1.25 1.70 1.01 0.53 21045 Core 0.14 0.17 0.84 bdl 7.56 17.83 0.69 5.61 49.79 15.96 98.59 37.70 46.96 12.70 2.64

0.02 0.03 0.10 bdl 0.08 0.22 0.02 0.03 0.11 0.13 0.36 31045 Edge 0.31 0.20 1.54 bdl 7.70 19.43 0.58 6.71 48.71 14.20 99.38 41.86 42.59 13.28 2.26

0.12 0.07 0.17 bdl 0.45 1.08 0.18 2.13 0.97 0.50 0.51 31045 Core bdl 0.20 1.91 bdl 9.14 18.22 0.82 8.67 46.20 13.01 98.15 40.45 40.20 16.08 3.28

0.01 0.03 0.10 bdl 0.25 0.30 0.06 0.04 0.37 0.07 0.63 31045 Edge 0.57 0.15 1.99 bdl 6.89 21.35 0.39 5.64 47.20 13.88 98.06 45.59 41.21 11.67 1.52

0.17 0.08 0.45 bdl 0.37 0.36 0.06 1.42 1.47 0.61 0.17 31045 Core 0.57 0.16 0.51 bdl 5.02 17.29 0.46 4.96 52.52 17.96 99.45 36.64 52.99 8.61 1.75

0.05 0.05 0.02 bdl 0.12 0.17 0.02 0.13 0.28 0.46 0.16 31045 Edge 0.69 0.13 0.61 bdl 4.86 18.70 0.47 5.76 51.72 16.58 99.51 40.19 49.58 8.40 1.83

0.20 0.06 0.02 bdl 0.08 0.18 0.01 0.56 0.66 0.13 0.53 31049 Orthopyroxene bdl 0.52 0.47 bdl 18.94 1.54 0.07 2.23 53.95 23.59 101.30 3.09 66.02 30.64 0.25

bdl 0.09 0.14 bdl 0.51 0.02 0.01 0.24 0.24 0.28 0.40 31049 Orthopyroxene bdl 0.43 0.44 bdl 17.21 1.66 0.07 2.52 53.33 24.57 100.22 3.35 68.71 27.69 0.25

bdl 0.07 0.04 bdl 0.12 0.07 0.01 0.27 0.48 0.16 0.20 4Group A xenoliths and xenocrysts

1007 In xenolith 0.19 0.20 1.67 bdl 8.07 19.94 0.63 7.82 47.11 12.84 98.48 43.95 39.36 14.16 2.530.08 0.04 0.24 bdl 0.49 0.38 0.07 0.69 0.70 0.35 0.37 17

1008 Megacryst 0.64 0.18 0.95 bdl 5.94 19.81 0.59 7.83 49.38 13.84 99.17 44.08 42.87 10.66 2.390.13 0.05 0.10 bdl 0.32 0.30 0.04 0.30 0.20 0.14 0.20 4





1045 Megacryst bdl 0.37 1.31 bdl 10.78 17.64 0.82 5.45 49.61 12.88 98.85 38.56 39.16 19.03 3.24bdl 0.05 0.14 bdl 0.32 0.18 0.04 0.41 0.35 0.36 0.45 8

1003 Megacryst bdl 0.27 1.11 bdl 10.01 17.73 0.56 5.21 50.74 13.47 99.09 38.93 41.15 17.70 2.21bdl 0.04 0.08 bdl 0.80 0.34 0.25 0.69 0.53 0.21 0.44 7

1003 Orthopyroxene bdl 0.34 0.45 bdl 16.81 1.55 0.07 3.27 53.27 23.28 99.03 3.25 68.11 28.36 0.27bdl 0.05 0.08 bdl 0.24 0.04 0.01 0.10 0.44 0.17 0.42 6

1013 In xenolith N/A 0.23 1.54 bdl 8.94 19.67 0.68 6.10 49.54 12.77 99.46 42.91 38.76 15.64 2.69N/A 0.05 0.29 bdl 0.34 0.27 0.06 0.54 0.53 0.34 0.34 6

Group B xenoliths and xenocrysts1003 Orthopyroxene in

Xenolithbdl 0.43 0.07 bdl 25.84 0.25 0.03 2.12 52.26 17.81 98.82 0.54 54.10 45.24 0.13

bdl 0.06 0.05 bdl 0.45 0.07 0.02 0.10 0.69 0.34 0.97 10Group C xenoliths and xenocrysts

1007 Combined xenolithcpx

0.66 0.10 1.01 bdl 5.17 20.94 0.35 5.88 50.66 14.09 98.86 46.20 43.26 9.15 1.39

0.15 0.05 0.21 bdl 0.42 0.52 0.07 0.64 0.42 0.46 0.28 24

a 1
All results are presented as wt%. N/A � not analyzed, bdl � below detection limit (0.1%). Italicized numbers are standard deviation. Total Ironis presented as FeO.


1049, BJ-1051, BJ-1053). This material comprises three dis-tinct textural and chemical varieties. Group A, the most com-mon inclusions, are defined as Al-augite (Wilshire andShervais, 1975) or as Type II mantle xenoliths (Frey and Prinz,1978). Group B is the second most commonly observed xeno-lithic material and is composed of norite, almost always founddisaggregated into low An-feldspar megacrysts and rare or-thopyroxene. Finally, disaggregated components (olivine, cli-nopyroxene and spinel) of Cr-diopside lherzolite are observedin rare cases (Group C). Group A and B xenoliths occurtogether at four localities (DZ-1003, DZ-1013, BJ-1047 andBJ-1051).

Megacrysts of clinopyroxene (up to 20 mm), olivine (up to10 mm) and plagioclase feldspar (up to 30 mm) are observed inmany locations from both Debre Zeyit and Butajira. Suchmegacrysts have not been previously reported in the MER, butare commonly observed in alkaline basaltic rocks (e.g., Irvingand Frey, 1984; Schulze, 1987; Shaw and Eyzaguirre, 2000).Al-rich (5.1–7.8% Al2O3) megacrysts of clinopyroxene typifyGroup A. These xenoliths occur primarily as amalgamations oflarge mafic crystals (Table 5), occasionally embedded in amatrix of sutured plagioclase feldspar (BJ-1044, BJ-1051, DZ-1003, DZ-1005). The amalgamations frequently exhibit meta-morphic textures such as 120o intergranular contacts and irreg-ular boundaries with reaction rims. Rutilated clinopyroxene isobserved in two of xenoliths (DZ-1013, BJ-1044).

Clinopyroxene megacrysts (Wo42–47; Fig. 2, Table 2) fre-quently contain inclusions of plagioclase and olivine. Group Aclinopyroxene has been analyzed at nine localities amountingto twenty-two crystals (9 megacrysts or inclusions in othermegacrysts and 13 in xenoliths). Many megacrysts have sievetexture where plagioclase replaces clinopyroxene, and exhibitstrain textures such as recrystallization, wavy extinction ortwinning. In samples where clinopyroxene megacrysts are ob-served, smaller clinopyroxenes that exhibit similar twinning orwavy extinction features are occasionally observed in the hostlava (e.g., DZ-1014). Clinopyroxene megacrysts frequently arechemically indistinguishable from phenocrysts in the hostmagma (e.g., DZ-1007, DZ-1008; Table 2) and in such casesare identified by size, inclusion in a xenolith or association withother xenocrysts.

Olivine megacrysts and xenolith components exhibit a widerange of core compositions though intrasample variation is

Fig. 3. Phenocryst and xenocryst feldspar zoning and composition inthe Debre Zeyit and Butajira areas.

generally restricted to �10% of the Fo content (Table 1). The

cores of some large olivines are Fo-poor relative to smallolivines present in the matrix, requiring the larger crystals to beclassified as xenocrysts. Thirty Group A olivine crystals (15megacrysts or inclusions in other megacrysts and 15 in xeno-liths) have been analyzed at 10 localities. Many of these largeolivine crystals exhibit stress features such as kink-banding andwavy extinction (Fig. 4) while others have recrystallized toform composite crystals (DZ-1007).

Twenty-seven Group A feldspar crystals (14 megacrysts orinclusions in other megacrysts and 13 in xenoliths) from 11locations have been analyzed (Table 3). Feldspar megacrystsand xenolith components occupy a range of core compositionscentered on An55 (Fig. 3). These feldspars frequently displayreverse zoning from An55 to compositions similar to those ofmatrix plagioclase. Small apatite inclusions are observed inxenocryst plagioclase at DZ-1006.

Opaque minerals in these xenoliths are primarily titaniferousmagnetites that occur as inclusions in clinopyroxene and oli-vine (Table 4). They are poorer in Cr than the inclusions foundin olivines in the host lava (0.04%–12% vs. 2%–38% Cr2O3).

The compositional similarity of Al-augite xenoliths andmegacrysts (e.g., DZ-1013, DZ-1003) suggests that individualmegacrysts result from disaggregation of Al-augite xenoliths assuggested in other megacryst localities (e.g., Shaw and Eyza-guirre, 2000; Shaw, 2004). Disaggregation of the xenoliths islikely accompanied by the dissolution of orthopyroxene andpartial to complete diffusive reequilibration of clinopyroxeneand olivine (Shaw, 2004). While it is possible that some cli-nopyroxene and olivine megacrysts may be cognate with thehost magma (Irving and Frey, 1984), they may equally havebeen crystalized from liquids that are more differentiated (Irv-ing and Frey, 1984; Schulze, 1987). FeO-MgO exchange co-efficients (KD � (FeOcrystal* MgOhost)/(FeOhost* MgOcrystal))for megacrysts observed in the DZBJ lavas frequently falloutside the equilibrium range (0.22–0.40; Irving and Frey,1984) suggesting a noncognate origin. Moreover, the reversezoning of some feldspar and olivine xenocrysts and high mag-nesium number of the host lavas suggest the Al-augite materialis not in equilibrium with the host lava.

4.1.3. Group B: Norite Xenoliths/Megacrysts

Group B xenoliths comprise An-poor (An�30) feldspar andorthopyroxene megacrysts. A single norite xenolith (DZ-1003)indicates that An-poor feldspar and rare orthopyroxene xenoc-rysts and megacrysts are derived from disaggregation of noritexenoliths. Orthopyroxene xenocrysts are observed in xenoliths(En54) at DZ-1003 and as individual crystals (En68) at DZ-1003and BJ-1049 (Table 2). Orthopyroxenes in a sample fromlocality BJ-1049 are mantled with a rim of small olivine crys-tals (Fo70). Apatite inclusions are also observed in orthopyrox-ene at DZ-1003.

Plagioclase feldspar megacrysts of core composition An�30

found in six locations (DZ-1003, DZ-1013, BJ-1042, BJ-1047,BJ-1049 and BJ-1051) sometimes exhibit reverse zoning tomatrix feldspar compositions (typically An60–70). One xenolithoccurrence of this feldspar zoning pattern was observed atDZ-1003 (Table 5). Sieve texture is common, though not
ubiquitous.


Table 3. Representative feldspar microprobe analysis.a

Location Analysis TiO2 K2O FeO CaO Na2O Al2O3 SiO2 MgO Total n An Ab Or

Host Lava1006 Core 0.09 0.21 0.59 13.64 3.38 30.80 51.34 0.10 100.15 68.18 30.58 1.24

0.02 0.01 0.06 0.05 0.05 0.14 0.21 0.01 0.01 31006 Edge 0.10 0.30 0.66 13.20 3.65 30.52 51.94 0.10 100.46 65.46 32.78 1.76

0.09 0.15 0.11 0.46 0.08 0.45 0.40 0.02 0.33 31008 Matrix 0.12 0.34 0.70 12.80 3.53 30.33 51.80 0.13 99.75 65.54 32.60 2.06

0.04 0.11 0.08 0.84 0.42 0.84 1.05 0.04 0.45 71008 Core 0.06 0.30 0.45 12.54 3.86 30.70 52.88 0.06 100.85 63.07 35.12 1.81

0.01 0.02 0.02 0.08 0.02 0.07 0.14 0.01 0.19 31008 Edge 0.06 0.17 0.52 14.74 2.83 32.11 49.94 0.11 100.47 73.48 25.50 1.02

0.02 0.02 0.07 0.16 0.04 0.24 0.48 0.02 0.71 31013 Matrix 0.12 0.23 0.57 13.44 3.46 30.77 52.27 0.12 100.99 67.25 31.36 1.39

0.05 0.02 0.10 0.23 0.11 0.29 0.28 0.02 0.29 81014 Core 0.11 0.20 0.56 14.14 3.13 31.43 51.48 0.09 101.13 70.52 28.29 1.19

0.02 0.01 0.13 0.02 0.11 0.26 0.32 0.02 0.26 31014 Edge 0.15 0.22 0.71 13.63 3.39 31.04 52.19 0.07 101.40 68.05 30.62 1.33

0.05 0.07 0.10 0.76 0.37 0.88 0.78 0.02 0.49 31044 Core 0.09 0.19 0.63 12.83 3.50 30.18 51.99 0.12 99.54 66.16 32.68 1.16

0.01 0.02 0.05 0.02 0.03 0.23 0.36 0.03 0.41 31044 Edge 0.13 0.19 0.64 12.78 3.56 29.99 52.08 0.11 99.47 65.72 33.14 1.15

0.05 0.02 0.07 0.10 0.08 0.48 0.32 0.02 0.63 31044 Core 0.10 0.20 0.61 12.93 3.50 30.13 51.94 0.12 99.53 66.31 32.45 1.24

0.03 0.02 0.04 0.11 0.04 0.15 0.04 0.02 0.12 31045 Core 0.10 0.19 0.57 14.35 3.16 31.25 50.21 0.15 99.97 70.73 28.15 1.12

0.03 0.01 0.07 0.06 0.05 0.21 0.24 0.01 0.25 31045 Edge 0.13 0.18 0.59 14.07 3.34 30.75 49.54 0.14 98.75 69.18 29.75 1.06

0.06 0.02 0.02 0.21 0.04 0.22 0.10 0.01 0.37 31049 Core 0.10 0.30 0.37 12.17 4.37 30.00 54.36 0.07 101.74 59.58 38.67 1.75

0.09 0.02 0.02 0.23 0.12 0.23 0.46 0.01 0.24 31049 Edge 0.12 0.32 0.62 11.73 4.50 29.60 54.53 0.09 101.50 58.03 40.22 1.87

0.09 0.03 0.16 0.66 0.39 0.61 0.90 0.00 0.20 3Group A xenoliths and xenocrysts

1007 In xenolith 0.05 0.33 0.43 11.89 4.24 29.53 53.54 0.05 100.06 59.56 38.45 1.990.05 0.01 0.02 0.22 0.08 0.34 0.30 0.00 0.47 3

1007 In xenolith 0.05 0.23 0.40 13.76 3.33 31.04 50.24 0.06 99.12 68.60 30.02 1.380.02 0.02 0.05 0.36 0.17 0.17 1.25 0.01 1.25 9

1008 In xenolith 0.08 0.26 0.31 13.50 3.16 30.83 51.05 0.08 99.27 3 69.15 29.29 1.550.07 0.06 0.03 0.70 0.39 0.80 1.05 0.02 0.14

1008 In xenolith 0.11 0.25 0.50 12.95 3.65 30.67 53.09 0.10 101.31 64.85 33.63 1.520.05 0.06 0.14 0.77 0.48 0.39 1.10 0.06 0.47 6

1044 In xenolith 0.08 0.17 0.49 13.31 3.30 30.64 51.44 0.13 99.54 68.30 30.66 1.040.07 0.02 0.05 0.24 0.05 0.34 0.26 0.03 0.36 7

1044 In xenolith 0.08 0.31 0.26 11.46 4.40 29.36 53.34 0.03 99.22 57.92 40.22 1.860.04 0.04 0.02 0.51 0.22 0.39 1.21 0.01 0.97 3

1044 In xenolith 0.08 0.16 0.51 13.76 3.15 31.07 50.59 0.13 99.44 70.02 29.02 0.960.03 0.00 0.07 0.23 0.10 0.15 1.10 0.01 0.94 3

1045 Megacryst Core 0.08 0.45 0.39 9.98 5.12 28.35 56.07 0.05 100.49 50.48 46.81 2.710.01 0.03 0.02 0.35 0.21 0.46 0.56 0.00 0.55 3

1045 Megacryst Edge 0.09 0.17 0.56 13.86 3.13 31.45 51.05 0.14 100.44 70.26 28.74 1.000.02 0.03 0.08 0.34 0.20 0.31 0.54 0.02 0.18 3

1051 In xenolith 0.07 0.21 0.45 12.29 3.83 30.50 52.52 0.07 99.94 63.10 35.62 1.280.04 0.03 0.04 0.32 0.19 0.31 0.39 0.02 —b 12

1003 In xenolith 0.07 0.26 0.53 11.91 4.31 30.17 55.21 0.11 102.58 59.48 38.95 1.570.04 0.02 0.22 0.41 0.24 0.19 0.68 0.03 0.56 6

Group B xenoliths and xenocrysts1013 Core 0.06 1.18 0.21 5.74 7.25 24.42 62.54 0.00 101.41 28.33 64.74 6.93

0.05 0.19 0.03 0.37 0.05 0.26 0.44 0.01 0.23 31013 Edge 0.08 1.47 0.61 5.44 7.20 24.18 62.66 0.06 101.69 26.91 64.45 8.64

0.13 0.19 0.57 0.52 0.28 0.22 1.26 0.07 0.23 31042 Core 0.08 0.64 0.16 7.33 6.62 26.21 59.39 0.01 100.42 36.52 59.69 3.79

0.05 0.09 0.05 0.22 0.05 0.04 0.54 0.01 0.31 31042 edge 0.11 0.31 0.66 12.06 3.99 30.00 52.97 0.05 100.16 61.34 36.77 1.89

0.02 0.02 0.04 0.16 0.09 0.19 0.40 0.03 0.54 31049 Core 0.10 0.89 0.28 7.58 6.63 26.15 59.96 0.02 101.60 36.40 58.14 5.12

0.03 0.06 0.03 0.07 0.02 0.12 0.24 0.01 0.11 31049 Edge 0.12 0.26 0.58 12.65 4.02 30.42 53.24 0.09 101.39 62.54 35.93 1.53

0.07 0.06 0.12 0.27 0.07 0.32 0.12 0.03 0.48 31003 In xenolith 0.04 0.98 0.17 6.43 7.02 25.30 61.40 0.01 101.35 31.67 62.61 5.72

0.05 0.08 0.03 0.06 0.20 0.30 0.69 0.01 1.17 5

a All values presented as wt%. Italicized numbers are standard deviation.
b All values normalized to 100.


4.1.4. Group C: Lherzolite Xenoliths and Xenocrysts

Several xenoliths of clinopyroxene-olivine granular aggre-gates (1–2 cm) are observed in sample DZ-1007. These xeno-liths have interacted with the host lava resulting in clinopyrox-ene reaction textures and the loss of all orthopyroxene. Olivines(Fo82) in these xenoliths are small and unstrained but coexist-ing clinopyroxene (Wo46) exhibits kink-banding. These tex-tures suggest that some the primitive olivine and clinopyroxenexenocrysts observed in the DZBJ volcanics are derived fromdisaggregation of Cr-diopside lherzolite xenoliths. Primitiveolivine typically occurs as small euhedral crystals within the

Table 4. Representative micropro

Sample Analysis TiO2 Al2O3 FeO

HDZ-1007 Spinel in skeletal olivine 0.64 16.98 17.3

0.38 1.84 0.0DZ-1007 Spinel in skeletal olivine 0.62 15.29 16.5

1.13 2.80 0.0DZ-1007 Spinel in olivine 0.72 54.87 17.9



0.64 1.80 0.1BJ-1042 Large magnetite 6.00 2.79 80.8

6.19 1.68 4.4BJ-1044 Spinel in olivine 1.92 31.89 34.6

1.26 1.94 0.2BJ-1049 Large Fe-Ti oxide 18.04 5.40 67.7

0.27 0.22 0.2BJ-1049 Large Fe-Ti oxide 18.04 3.83 68.9

0.75 0.24 0.9BJ-1049 Illmenite with Fe exosolution 43.72 0.70 46.5

0.31 0.17 0.0BJ-1049 Fe rich exosolution 20.82 3.69 66.6



3.12 0.24 0.1Group A xeno

DZ-1003 Magnetite in megacryst 13.93 7.75 63.70.72 0.17 0.1

BJ-1044 Magnetite in xenolith 13.61 9.87 54.70.42 0.01 0.3

Group C xenoDZ-1008 Spinel in olivine 1.00 31.25 21.2






1.06 0.34 0.0

a All analyses are wt%. Italicized numbers are standard deviations.b Total Iron presented as FeO.c Low totals reflect Fe2O3.d Cr# � (Cr/(Cr � Al)) * 100.e Mg# � (Mg/(Fe � Mg)) * 100.

host lava. However, rare large, primitive and anhedral strained

olivine xenocrysts can occur in the same samples as theseeuhedral primitive olivine phenocrysts (Table 1). Primitiveolivine megacrysts (3.25 mm Fo88) exhibiting kink bands areobserved in BJ-1045 and similar primitive olivine is observedin DZ-1008 and DZ-1007. Large (2.5 mm) strained olivines(Fo86) and mosaic recrystallization textures are observed inBJ-1044. Spinel inclusions within these olivines are Cr2O3 rich(12%–37%) and have a range of Mg# from 42 to 69 (Table 4).

Two groups of clinopyroxenes have geochemical featurescharacteristic of Cr-diopsides (e.g., higher Cr2O3, MgO andlower TiO2, FeO than Al-augites; Irving, 1980; Griffin et al.,

sis of spinels and Fe-Ti oxides.a

nO MgO NiO Cr2O3 Totalc n Cr#d Mg#e

a.20 28.73 0.19 19.02 100.58 42.90 74.73.40 0.05 0.09 2.08 1.06 3.20 29.16 0.15 17.66 99.21 43.65 75.82.69 0.01 0.07 1.92 2.69 2.16 18.12 0.36 7.17 99.60 8.05 64.31.10 0.02 0.05 0.38 0.36 3.25 37.50 0.15 6.19 99.39 33.88 80.83.04 0.02 0.00 0.08 0.12 2.17 19.72 0.17 29.08 98.54 53.48 61.23.19 0.02 0.02 3.23 2.01 3.30 2.31 Bdl 0.34 92.79 7.64 4.84.04 0.11 Bdl 0.59 0.73 3.26 10.98 0.19 17.60 97.60 27.02 36.09.06 0.06 0.07 0.34 0.30 2.50 3.84 Bdl Bdl 95.72 0.92 9.19

Bdl 0.08 Bdl 0.14 0.32 3.51 3.81 Bdl Bdl 95.37 1.09 8.95

Bdl 0.11 Bdl 0.20 0.47 3.37 4.73 Bdl Bdl 96.20 6.34 15.33

Bdl 0.01 Bdl 0.12 0.29 3.45 3.85 Bdl 0.10 95.69 1.75 9.32.01 0.00 Bdl 0.11 0.46 2.18 14.55 0.21 34.93 99.38 44.80 56.66.09 0.01 0.02 0.06 0.02 2.18 14.84 0.26 34.25 98.22 43.63 59.26.51 0.03 0.02 0.96 0.83 3d xenocrysts.38 5.15 Bdl 0.50 91.72 4.18 12.58.00 0.05 Bdl 0.09 0.46 3.38 5.23 0.12 12.85 96.86 46.63 14.55.17 0.01 0.01 0.07 0.41 3d xenocrysts.19 14.32 0.17 28.86 97.67 38.26 54.53.04 0.03 0.08 0.85 0.42 2.30 12.07 0.17 36.61 96.97 53.32 45.82.91 0.01 0.04 0.08 0.49 2.16 21.81 0.20 15.71 99.73 22.85 69.18.02 0.02 0.06 3.12 0.82 2.39 14.46 0.17 35.62 98.06 49.59 54.54.52 0.08 0.02 0.14 0.91 2.18 23.12 0.18 12.37 99.25 26.22 62.72.68 0.02 0.02 3.73 0.31 3.26 10.33 0.13 37.31 97.02 53.16 41.51.38 0.04 0.03 0.13 0.45 3

below detection limits (0.1%).

be analy

b M

ost lav1 03 27 06 32 07 05 00 06 02 29 04 06 06 10 069 095 089 04 03 02 07 04 1liths an5 07 01 02 0liths an8 05 13 03 01 00 28 09 39 07 23 06 0

Bdl �

1984; Ghorbani and Middlemost, 2000): (a) clinopyroxene that


Table 5. Xenolith and host lava components with temperature and pressure estimates (in parentheses).a

SampleClinopyroxene composition

(pressure, kb)Olivine composition

(temperature, °C) Feldspar compositionb

Xenolith/MegacrystDZ-1013 Xenolith Wo42-En40 (4.4–5.1) Fo64 (1120) An49-Ab48

DZ-1007 Composite Wo45-En39 Fo79–81 (1288–1318) An60-Ab38

DZ-1008 Clinopyroxene megacryst Wo44-En42 Fo74 (1155) An65-Ab34 to An69-Ab29

RimDZ-1007 Xenolith Wo44-En39 (6) Fo75-Fo81 (1172–1261) An69-Ab30

DZ-1008 Olivine megacryst Wo42-En42 (0.2) Fo70C (1150)-Fo75R (1212) An54-Ab34

DZ-1008 Clinopyroxene megacryst Wo44-En43 (7.8) — —BJ-1045 Feldspar megacryst — Fo80 An51-Ab47C to An70-Ab29RBJ-1045 Clinopyroxene megacryst Wo40-En40 (8.8) Fo82 (1341) An53-Ab44

BJ-1045 Clinopyroxene megacryst Wo39-En39 (4.6) — —BJ-1044 Xenolith Wo43-En40 (2.7) Fo77 (1223) An68-Ab31 & An57-Ab40

BJ-1044 Xenolith Wo45-En43 (3.6) Fo78–81 (1182–1230) An70-Ab29

BJ-1044 Clinopyroxene megacrysts Wo44-En44 Fo78–83 (1140–1213) An70-Ab29

DZ-1003 Xenolith Wo39-En40 Fo76–78 (1283–1319) An59-Ab39

DZ-1003 Norite xenolith En54-Fs45 — An31-Ab53

DZ-1007 Lherzolite fragment Wo46-En43 Fo82 —Host Lava

DZ-1006 (0.90) —(1.24) —(0.36) —

DZ-1007 — (1126)DZ-1008 (4.25) —

(5.93) —(1.16) —(3.44) —(9.42) —

DZ-1009 — (1213)(2.70) —

DZ-1014 (3.89) —(2.96) —(5.54) —(1.47) —(3.82) —(3.55)(4.79) —(4.61) —

BJ-1044 (3.55) (1276)(2.76) —(3.40) —(4.09) —(3.49) —(2.04) (1267)

BJ-1045 (6.93) —(5.63) (1326)c

— (1303–1326)(8.17) —(0.80) —(9.99) —(4.78) —(0.32) —(4.05) —

— (1362)c

(8.08) —(6.76) —

a Pressure calculated using Nimis and Ulmer (1998b) with a standard error of 1.7 kb. Temperatures calculated using geothermometer of Loucks(1996). C � core; R � rim of crystal. Ranges of temperatures are given in the case of more than two crystals. The standard error of the geothermometeris � 6°C while analytical uncertainty is generally less than � 35°.

b Not used in PT calculations.
c From Lindsley (1983).

includ


coexists with primitive anhedral (or strained) olivine and (b)clinopyroxene in lherzolite xenoliths. These xenocrystic cli-nopyroxenes that exhibit Cr-diopside characteristics haveclearly been metasomatised but are difficult to distinguish fromother clinopyroxene phenocrysts based on optical characteris-tics. Such xenocrysts and are grouped with the host lava inTable 2.

4.2. Geochemical Characterization of Mafic Host Lavas

Mafic scoria from cinder cones and associated vesicularflows from this region are primarily basalts accompanied by afew trachy-basalts (Fig. 5). Their bulk compositions plot withinthe range of Quaternary Ethiopian rift lavas, and define a mildlyalkaline liquid line of descent. All samples are mildly neph-eline-normative (�7%) excluding five that are hypersthene-normative. The clear trend of decreasing CaO/Al O with de-

Table 6. Whole rock major, minor, and trace

Location SiO2 Al2O3 Fe2O3 MgO CaO Na2O K2O TiO2 MnO

DZ-1003 49.42 16.04 12.18 6.84 8.98 3.34 1.11 2.07 0.19DZ-1004 48.71 15.78 10.74 8.50 10.05 3.07 1.51 1.82 0.17DZ-1005 49.62 16.07 10.94 7.16 9.57 3.15 1.60 1.97 0.17DZ-1006 50.01 16.96 10.75 6.96 9.10 3.22 1.58 1.63 0.17DZ-1007 47.42 15.48 11.28 9.26 10.68 2.99 1.13 1.87 0.18DZ-1008 47.39 15.51 11.38 8.96 10.75 3.00 1.33 1.96 0.19DZ-1009 48.10 13.76 10.98 10.21 11.16 3.05 1.25 1.81 0.18DZ-1012 47.08 16.10 12.13 8.06 10.71 2.91 0.99 2.28 0.19DZ-1013 48.53 16.01 11.74 7.52 9.45 3.40 1.41 2.20 0.18DZ-1014 47.31 16.01 11.28 7.76 10.73 3.49 1.50 2.00 0.18DZ-1015 49.97 15.81 10.28 6.99 10.04 3.29 1.85 1.87 0.17BJ-1042 48.68 17.11 12.66 5.20 7.69 4.01 1.76 2.92 0.20BJ-1043 47.32 14.92 12.38 8.79 10.22 2.97 0.89 2.59 0.18BJ-1044 47.24 14.89 13.25 8.02 9.67 3.21 1.07 3.05 0.19BJ-1045 47.76 15.26 11.71 9.09 10.20 2.85 1.03 2.44 0.17BJ-1047 47.99 16.56 12.17 6.67 8.91 3.42 1.50 2.72 0.19BJ-1048 48.10 15.17 12.53 8.26 9.78 2.98 0.97 2.37 0.19BJ-1049 48.17 16.73 12.75 5.97 8.27 3.74 1.54 3.02 0.19BJ-1051 47.58 15.20 12.06 8.95 9.68 3.28 0.97 2.37 0.18BJ-1052 48.21 15.79 12.07 7.83 9.55 3.23 1.16 2.44 0.18BJ-1053 49.17 15.35 11.56 8.15 9.87 3.09 1.09 2.38 0.16

Location Ce Pb Nd Sm Zr Hf Eu Gd Tb

DZ-1003 62.1 3.49 33.1 6.75 175 3.83 2.06 5.91 0.99DZ-1004 66.1 3.08 29.9 5.81 186 4.21 1.77 5.48 0.82DZ-1005 71.4 3.31 32.8 6.30 218 4.91 1.97 6.09 0.93DZ-1006 68.3 2.87 30.9 5.87 192 4.35 1.78 5.89 0.89DZ-1007 62.1 2.48 31.0 5.93 175 3.81 1.83 5.25 0.89DZ-1008 59.9 2.19 28.8 5.71 165 3.99 1.88 5.74 0.86DZ-1009 51.2 1.85 25.4 5.10 134 3.15 1.79 5.01 0.73DZ-1012 34.9 1.52 17.3 3.62 97 2.19 1.18 3.49 0.54DZ-1013 69.5 3.64 33.3 6.47 207 4.73 2.13 6.27 0.94DZ-1014 65.1 2.54 31.1 6.03 170 4.10 2.11 5.89 0.86DZ-1015 81.9 4.50 36.9 6.91 226 5.34 2.03 6.63 1.01BJ-1042 78.3 2.56 39.4 7.93 207 5.01 2.60 7.37 1.09BJ-1043 46.7 1.57 27.1 5.61 130 2.98 1.88 5.00 0.83BJ-1044 55.8 1.98 31.8 6.68 148 3.49 2.58 6.57 0.96BJ-1045 47.7 1.82 25.4 5.43 131 3.19 1.89 5.08 0.79BJ-1047 67.1 2.18 34.6 6.95 166 3.84 2.40 6.76 0.95BJ-1048 50.6 2.20 28.1 5.93 149 3.55 2.11 5.97 0.91BJ-1049 66.0 2.79 35.2 7.17 178 4.06 2.41 6.80 1.01BJ-1051 54.3 1.81 30.2 6.11 151 3.37 2.02 5.31 0.88BJ-1052 49.7 2.01 26.5 5.57 133 3.45 1.92 5.19 0.78BJ-1053 43.5 1.47 22.9 4.62 118 2.91 1.63 4.54 0.68

a Major and minor elements (including Ba and Sr) were analyzed byPrecision based on replicate analyses and are generally better than �1%Sc, V, Co, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Hf, and Ta; �5% P2Otrace elements are given as ppm. Standards used were USGS natural rstandard 688; Lamont Doherty Earth Observatory standards K1919a a

b The sum is all major and minor oxides obtained by DCP and also

2 3

creasing MgO indicates strong compositional control of

clinopyroxene fractionation on melt composition. Negative cor-relations of Cr, Ni and Sc with MgO suggest that olivine andclinopyroxene fractionation controls the observed variations incompatible trace element concentrations. Ni contents of theDZBJ lavas are consistently �145 ppm, less than the valueexpected for primitive mantle-derived liquids, requiring someremoval of olivine during differentiation. However, the pres-ence of xenocryst clinopyroxene and olivine indicate that bothof these phases have also been mechanically added to the hostlava.

Other major and incompatible trace element variability as afunction of MgO indicate that the Debre Zeyit and Butajiraseries are not related strictly through fractional crystallizationof a common parental basalt (Fig. 6). The Debre Zeyit andButajira suites may be distinguished by the greater abundanceof the more incompatible trace elements (e.g., La, Zr, Rb and

nt analysis results for the DZBJ host lavas.a

SUMb Sr Ba Cs Rb Th U Nb Ta La

100.80 597 522 0.04 16.9 2.29 0.43 34.9 1.98 30.7100.81 618 449 0.34 31.6 3.74 0.96 45.6 2.76 34.4100.75 579 506 0.12 32.8 3.99 0.90 46.6 2.86 36.4100.82 500 380 0.10 31.1 4.06 0.69 42.2 2.53 34.9100.86 601 408 0.28 24.0 2.87 0.67 41.8 2.50 31.2100.94 571 412 0.39 28.6 2.91 0.78 42.9 2.58 30.0100.89 698 427 0.30 26.8 2.62 0.74 35.5 2.12 25.9100.76 334 241 0.09 10.8 1.69 0.42 23.5 1.38 17.5101.00 697 500 0.29 29.0 3.54 0.91 46.8 2.78 35.0100.82 723 449 0.26 30.4 3.42 0.89 48.1 2.82 33.3100.82 639 493 0.50 42.9 5.63 1.40 55.2 3.49 42.3100.84 817 581 0.29 29.5 3.96 1.01 56.2 3.61 38.1100.90 693 347 0.10 14.9 2.08 0.52 29.8 1.80 22.0101.30 728 424 0.16 17.3 2.36 0.55 36.9 2.34 26.0100.94 679 362 0.21 17.7 2.30 0.62 32.5 2.16 23.1100.75 782 515 0.20 22.9 3.21 0.60 47.3 2.88 32.6100.89 577 409 0.10 15.7 1.88 0.49 27.0 1.65 25.7101.04 781 492 0.24 23.6 2.93 0.66 44.4 2.71 31.3100.92 635 387 0.17 19.4 2.46 0.59 36.6 2.14 26.3100.91 656 365 0.14 19.0 2.42 0.63 33.7 2.21 23.7101.19 589 297 0.17 16.1 2.02 0.58 29.5 1.86 20.9

Y Ho Er Yb Lu Co Cr Ni V Sc

31.1 1.11 2.99 2.71 0.40 54.2 178 82 238 26.124.2 0.91 2.37 2.12 0.32 60.9 370 112 235 28.027.1 1.02 2.71 2.43 0.38 59.0 236 69 241 26.527.9 1.01 2.70 2.48 0.41 58.1 219 101 211 26.025.5 0.94 2.50 2.20 0.34 57.8 307 108 233 27.724.5 0.91 2.40 2.21 0.32 60.7 336 107 239 28.719.8 0.74 1.87 1.56 0.23 63.6 606 145 249 28.915.4 0.57 1.46 1.26 0.21 38.6 128 61 161 16.327.3 1.03 2.57 2.40 0.38 66.7 214 172 248 27.023.5 0.88 2.34 1.93 0.29 60.5 188 89 236 25.228.5 1.06 2.73 2.56 0.40 57.2 229 80 210 23.530.3 1.09 2.82 2.60 0.39 45.4 16 20 216 19.323.4 0.86 2.21 1.82 0.28 62.2 341 125 267 27.326.2 0.96 2.54 2.01 0.31 63.0 295 103 294 28.921.5 0.81 2.04 1.86 0.29 60.9 397 139 277 28.626.5 0.97 2.49 2.19 0.34 55.8 129 54 244 23.027.3 0.99 2.61 2.18 0.32 69.1 356 119 268 28.428.1 0.99 2.72 2.26 0.35 53.6 38 33 240 20.725.9 0.95 2.50 2.14 0.34 62.7 357 134 241 25.822.2 0.84 2.19 1.89 0.29 56.8 215 99 224 23.618.6 0.70 1.83 1.50 0.23 53.6 227 86 219 22.8

except P2O5 and all trace elements, which were analysed by ICP-MS.Sr, Y, Zr, Nb, La, Ce; �3% for other major elements, Ba, Sr, Rb, Cr,b, Lu, Pb, Th, and �8% Cs and U. All oxides are presented as wt %,dards W-2, DTS-1, DNC-1, AGV-1, G-2, SDC-1, BIR-1; NBS/NIST92.es P2O5, which was analyzed by ICP-MS.

eleme

P2O5

0.620.460.490.440.560.490.390.310.570.560.550.620.640.700.420.620.560.650.650.440.37

Dy

5.664.605.045.084.884.543.872.865.104.555.445.824.645.264.255.064.975.494.944.323.51

DCP,SiO2,

5, Ni, Yock stannd All-

Nb) in Debre Zeyit relative to Butajira lavas of comparable

5 show


MgO content. In contrast, TiO2 is more enriched in Butajirasamples (Fig. 6). The less incompatible trace elements (e.g.,Yb, Y) do not show such variability at common MgO content.Primitive-mantle normalized incompatible trace element pro-files (Fig. 7) of Debre Zeyit and Butajira samples are homoge-

Fig. 4. Photomicrograph of sample BJ-104

Fig. 5. IUGS classification (Le Bas et al., 1986) of the Debre Zeyitand Butajira series. Dashed line � MER Quaternary products; stippledfield � Oligocene lavas associated with the onset of plume volcanism
(Hart et al., 1989; Pik et al., 1999; Furman et al., (2005) and referencestherein).
neous overall. They are characterized by positive Ba anomaliesand low normalized abundances of Pb, Th and U. The observeddifferences between the Debre Zeyit and Butajira suites may beattributed to partial melting processes, which exert a strongcontrol on the compositions of the DZBJ mafic lavas. Thebehavior observed for Nb, La and Ba (Fig. 8), for example, canbe explained by a combination of partial melting and fractionalcrystallization (Hoernle and Schmincke, 1993; Furman, 1995).In contrast, the behavior of TiO2 (Fig. 8) is consistent with ahigh degree of compatibility in the residuum during melting(for example, buffered by an accessory phase) as describedbelow.

5. DISCUSSION

5.1. Conditions of Melting and Melt Segregation

The major element compositions of the DZBJ basalts, back-corrected to Mg# 72 by step-wise addition of equilibriumolivine, indicate that parental mafic liquids were generated atpressures between 15 and 25 kb (Fig. 9). They are amongst themost shallowly-derived products of the EARS, and plot in afield similar to Quaternary mafic lavas from Turkana, N. Kenyathat are interpreted as moderate (5%–7%) degree melts offertile spinel peridotite (Furman et al., 2004). The back-corre-lated FeO* and SiO2 contents of the DZBJ primary liquids plotalong a trend consistent with melting in the presence of residualclinopyroxene (Fig. 9), suggesting melt percentages below�7% (Kushiro, 1996; Wasylenki et al., 2003). Trace elementabundances are consistent with a depth of melt generationwithin the spinel lherzolite field. Given the moderate degree of

ing kink banding in an olivine crystal.

partial melt and pressure range of melt generation, a tempera-


Fig. 6. Selected major element oxide abundances (wt%) and incompatible trace element abundances (ppm) plotted againstMgO (wt%) for Debre Zeyit and Butajira mafic lavas. Samples from Debre Zeyit have consistently higher abundances of
LREE and LILE than lavas from Butajira at comparable MgO contents, although abundances of HFSE and HREE such asYb overlap in samples with �8 wt% MgO. See text for discussion.
Fig. 7. Primitive mantle normalized (Sun and McDonough, 1989) trace element profiles. Gray fields enclose the DebreZeyit and Butajira series, with the average pattern for each location also shown. See text for discussion.


ture range from �1270 to 1425°C (based on experimentalstudies; Kushiro, 1996) would be expected. This range is con-sistent with the maximum temperature estimate (1362°C; Table5) for the DZBJ lavas.

5.2. Genesis of Mafic Host Lavas

The bulk geochemistry of the DZBJ lavas does not support asignificant role for crustal contamination. Both Ce/Pb and K/Rbvalues fall within the range of mantle derived basalts (Fig.

Fig. 8. Abundances of elements expected to behave incompatiblyduring partial melting of anhydrous spinel-lherzolite are shown herenormalized to Al2O3 contents and plotted against Zr/Al2O3 followingthe method of Hoernle and Schmincke (1993). Positive linear correla-tions observed for Nb and La pass through origin and are interpreted asindicating entirely incompatible behavior during melting in DebreZeyit/Butajira source region. The plot for TiO2 however, has a shallownegative slope, indicating that abundances of this element are con-trolled by a phase present throughout the melting interval, interpretedhere to be titaniferous clinopyroxene.

10A,B). In samples that are distal from the Quaternary silicic

centres (DZ-1009; DZ-1012), the Ba, Th, U and Pb abundancesare identical to those observed in other DZBJ lavas, makingassimilation of silicic material an unlikely process by which theobserved incompatible element patterns were generated. Thehost lavas plot within the ocean island basalt (OIB) field of theRb/Cs–K/Rb variation diagram (Fig. 10B; Hart and Reid, 1991)which would argue against both lower-crustal granulite andupper-crustal rhyolite contamination. The source of the DZBJlavas is unclear without isotopic information; however, trace-element ratios (La/Nb-Ba/Nb) may be used to compare theselavas with other suites in the region (Fig. 10C). The DZBJ lavashave incompatible trace element abundances similar to those ofmodern plume-derived basalts from Djibouti and Erta’Ale.These geochemical signatures are clearly distinct from those ofthe HT Oligocene basalts (Fig. 10C), many of which likelyincorporate contributions from sublithospheric source regions(Pik et al., 1999).

The geochemical data constrains the possible mineralogy of

Fig. 9. FeO* vs. SiO2 for selected EARS mafic suites and experi-mental melts in equilibrium with mantle peridotite (Baker and Stolper,1994; Kushiro, 1996). Dotted fields encircle compositions of experi-mental melts coexisting with cpx � oliv � opx; dashed fields encloseexperimental melts coexisting with olivine � orthopyroxene afterhigher degrees of melt removal. Major element compositions wererenormalized with total iron as FeO, then corrected for olivine frac-tionation to correspond to experimentally derived melts of fertile pe-ridotite (Mg# �72); only samples with MgO �8 wt% were correctedusing this procedure. Experimental data are plotted as reported and areshown in the enclosed fields. Corrected compositions of the DZBJvolcanics indicate melting of cpx-bearing peridotite at pressures of15–25 kb. For comparison, Turkana basalts define two fields: (1) TPR(Turkana Pliocene-Recent) lavas indicate melting conditions similar tothose inferred for the DZBJ series, and extend to slightly lower pres-sures consistent with the unusually thin crustal section in this area(Furman et al., 2004); (2) TOM (Turkana Oligo-Miocene) picrites andbasalts record pressures of 25–30 kb, and also have much higher FeO*contents than the other sample groups, consistent with melting extentsgreat enough to deplete the residuum in clinopyroxene entirely. Pres-sures estimated on the basis of this diagram are approximate becauseisobaric experiments are not directly analogous to natural systemsgenerated by adiabatic decompression melting; however, the relativepressures inferred for EARS suites are useful for comparative purposes.

the mantle source region for the DZBJ basalts. At the inferred


Fig. 10. Whole-rock major and trace element ratios from the DZBJ volcanics compared with other regional mafic (�5.5%MgO) basalts: Oligocene high-Ti and low-Ti flood basalts (Pik et al., 1999; Kieffer et al., 2004); Quaternary plume-associatedmagmatism in Erta’Ale (Barrat et al., 1998) and Djibouti (Deniel et al., 1994); MER Quaternary Wonjii basalts from Gedemsa(Peccerillo et al., 2003); and Quaternary Turkana basalts (Furman et al., 2004). (A) Ce/Pb vs. SiO2 of DZBJ host lavas comparedwith other basalts in the region. DZBJ lavas fall broadly within the range expected for mantle-derived lavas (25 � 5; Hofmannet al., 1986). Other basalts in the region, such as the Oligocene flood basalts and Quaternary samples from Gedemsa, show signsof crustal or lithospheric contamination. (B) Typical variation in Rb, Cs, and K with the fields for OIB and MORB outlined (Hartand Reid, 1991). Clustering of the DZBJ data within the fields of OIB and sublithospherically derived Quaternary Turkana basaltssuggest minimal contribution from lithospheric sources for the DZBJ lavas. Note that the Oligocene flood basalts, in contrast,show substantial evidence for lithospheric or crustal involvement. Those Oligocene samples with the highest K/Rb ratio also havethe lowest Ce/Pb, suggesting again crustal or lithospheric influence. (C) Ba/Nb vs. La/Nb variation diagram indicating thedifferences between the DZBJ lavas and other regional basalt suites. Clustering of the DZBJ basalts in the Quaternary plume field(compared with Oligocene flood basalts) suggests a significant shared source component with the plume-derived basalts, without
substantial crustal contribution.


pressure of melt extraction (15–25 kb), clinopyroxene andspinel are the dominant phases that contribute to partial melts.The compatible behavior of Ti over the entire melting interval(Fig. 8) points to a phase that is not exhausted during the partialmelting process, which we interpret to be clinopyroxene. Cli-nopyroxene in the DZBJ suite has up to 3.7 wt% TiO2 whilespinel inclusions in lherzolitic olivine (Table 4) are character-ized by low TiO2 (� 1.1%) and high Cr2O3 (12%–37%). Thepresence of phases rich in TiO2 such as rutile or illmenite is notsupported by evidence indicating that Nb (Fig. 8), Y and Ta areincompatible throughout the melting interval. There is no evi-dence of hydrous metasomatism (amphibole and phlogopite) inthe host lavas or xenoliths. Their presence in the magmaticsource is almost wholly precluded by the host basalt Rb/Sr(0.02–0.07) and Ba/Rb (11.5–30.9) ratios, both of which fallgenerally close to values inferred for the primitive mantle (Sunand McDonough, 1989). Similarly, no evidence of carbonatitemetasomatism such as low Ti/Eu values or high Zr/Hf values(Dupuy et al., 1992; Rudnick et al., 1993; Yaxley et al., 1998;Gorring and Kay, 2000) is observed in the DZBJ basalts. Wetherefore propose a spinel lherzolite source mineralogy inwhich clinopyroxene is unusually titaniferous.

5.3. Genesis of Debre Zeyit-Butajira Xenoliths

The DZBJ lavas have entrained fragments of lithosphericmaterial over a range of depths. The xenoliths have a crustaland/or mantle provenance that varies from shallow cumulatesand moderate-pressure dyke-vein systems to lithospheric-man-tle peridotites. Each xenolith group is mineralogically andchemically distinct and therefore a multi-stage, polybaric, ori-gin is required to explain their diversity. As outlined below,lithospheric processes related to mechanical rifting and magmasupply generated the Al-augite and norite xenoliths whereas thelherzolites are fragments of the shallowest lithospheric mantle.

5.3.1. Xenolith Thermobarometry

The wide variety of crystalline phases in these samplesprovides an opportunity to develop a clear understanding of thesource regions for both mafic lavas and xenoliths. Magmatictemperatures were estimated from the compositions of coexist-ing olivine-clinopyroxene pairs (Loucks, 1996) and two-py-roxenes pairs (Lindsley, 1983) where appropriate (Table 5).Temperature estimates from xenoliths define a wide range ofnear-magmatic temperatures: 1120 to 1341°C. Clinopyroxenegeobarometry (Nimis and Ulmer, 1998a,b) typically indicatespressures of crystallization for samples from both Debre Zeyitand Butajira of less than 6 kb (Table 5), although two samples(DZ-1008 and BJ-1045) yield pressure estimates of 7.8 � 1.7and 8.8 � 1.7 kb respectively.

The variability in temperature at a given pressure makes itdifficult to construct a geothermal profile for the region (Fig.11). It is clear that the significant range in pressure estimates(even within a single locality) is inconsistent with derivation ofall xenoliths from a single midcrustal magma chamber. Indeedsuch variation implies greater complexity within the plumbingsystem, e.g., many small magma chambers or dykes/veins largeenough to accommodate fractionation.

Pressure-temperature studies of other xenolith suites in Ethi-

opia have recorded substantially lower temperatures at similarpressures (e.g., �1045°C, 11.3 kb; Conticelli et al., 1999) thanthat inferred from lavas and xenoliths in the Debre Zeyit andButajira regions (Fig. 11). The disparity between the hot Al-augite/host lava and the cooler wall rocks represented in otherEthiopian xenolith suites suggests that these mafic lavas may bean important heat transfer mechanism in the rift.

5.4.2. Origin of Group A: Al-Augite Xenoliths

Pressure estimates from the Al-augite xenoliths indicate thatthey are derived from depths ranging from �1–30 km (Table5). It is unlikely that xenoliths derived from even a largecumulate underplate would exhibit such a range of pressures.Similarly, the wide range in both temperatures and xenocrystcomposition argues against their derivation from a homogenousbody. We suggest the Group A xenoliths are derived from asystem of lithospheric dykes and veins. The magma that pro-duced the observed Al-augite material was too evolved to havebeen directly generated by partial melting of the lithosphericmantle as suggested at other Al-augite occurrences (e.g., Irving,1980). The range of olivine megacryst compositions (Fo64–80)and the presence of An50–70 feldspar indicate the influence ofa differentiation process. Zonation and crystal fractionationwithin Al-augite dykes has been observed in lherzolites of theBalmuccia peridotite massif (Mukasa and Shervais, 1999) sug-gesting a likely process by which the megacryst compositionsevolved (Irving, 1980; Mukasa and Shervais, 1999). Occasional“reaming” of the conduits by primitive magma would prefer-entially entrain the more evolved megacrysts and xenoliths(Irving, 1980). The lack of wall-rock attached to the Al-augitexenoliths is not unexpected: the pervasive disequilibrium tex-tures indicate the veins were not fully solidified at the time ofentrainment (see also Litasov et al., 2000; Luhr et al., 2001).Al-augite megacrysts and xenoliths have been observed in all

Fig. 11. Cartoon of pressure-temperature variations for DZBJ xeno-liths and host lavas plotted from Table 5. The estimated depth andtemperature of melt generation of the DZBJ host lavas is derived fromFigure 9. For comparison, xenoliths from the northern plateau andsouthern rift (Conticelli et al., 1999) are also plotted. Moho depthestimates are for the central MER from Dugda et al. (2005). A sche-matic of dyking/veining is shown with fractional crystallization ondyke/vein walls.

modern continental rift localities (Litasov et al., 2000; Luhr et


al., 2001; this study). Not surprisingly, veining of the lowerlithosphere has been linked to lithospheric extension or thin-ning (Foley, 1992; Bodinier et al., 2004). We suggest thisprocess of dyke/vein intrusion represents an essential phase inmagma assisted rifting (e.g., Buck, 2004)

5.3.3. Group B: Norite Xenoliths

Group B xenoliths are interpreted as samples of cumulatesfrom shallow magma chambers. Sieve textures in clinopyrox-ene and plagioclase both in the host lava and in Group Axenoliths, in addition to the reverse zoning of the andesine/oligoclase feldspar at some locations, suggests prolonged res-idence time within a magma chamber after the incorporation ofthe Al-augite material. Orthopyroxene is abundant both withinthe norite xenolith and in host lava samples from DZ-1003.Surprisingly, however, orthopyroxene is absent at all otherlocalities save BJ-1049, where a lack of reaction textures re-quires minimal interaction between the host lava and group Bxenolith. Thus we infer that residence time within these shallowcrustal chambers is typically sufficient to dissolve orthopyrox-ene (Shaw, 2004) and develop resorption textures.

The inferred shallow chamber environment is not sampled asoften as the regional dyke network that yields the Al-augitematerial. Indeed, the only norite xenolith locality (DZ-1003) isparticularly close to the silicic centre of Mt. Yerer and thebasaltic host lava may have sampled material associated withYerer. Shallow magma chambers have been suggested else-where along the Main Ethiopian Rift axis, where the process oflow-pressure fractionation (�5 kb) of primitive melts andassimilation of Pan-African crust of variable composition hasbeen invoked to explain Pb isotope compositions of differen-tiated Quaternary lavas (Hart et al., 1989; Trua et al., 1999).Though the precise mechanisms by which the evolved productsare generated remain controversial (Ayalew, 2000; Deniel etal., 2000), it is clear that shallow magma chambers exist in theregion, as indicated by the occurrence of calderas at mostwithin-rift volcanoes and by widespread ash and pumice de-posits.

5.3.4. Group C: Xenoliths and the Metasomatic Impact ofAl-Augite Dykes and Veins

In the Debre Zeyit and Butajira areas, silicate melt associatedwith Al-augite dykes and vein growth may have resulted inAl-Fe-Ti-alkali metasomatism of the wall rocks. This interac-tion has been observed at other localities (Irving, 1980; Griffinet al., 1984, 1988; O’Reilly and Griffin, 1988; Mukasa andShervais, 1999), although the products of the interaction are oflimited extent and difficult to distinguish from the crystalizedproducts of the intruding silicate dyke/vein (Irving, 1980).Clinopyroxenes in Group C xenolith DZ-1007 display clearlythe effects of silicate melt metasomatism: increasing TiO2 withdecreasing Mg# (Fig. 12; Irving, 1980; Griffin et al., 1984;Mukasa and Shervais, 1999). The Group C clinopyroxenexenocrysts have the highest Mg# and Cr2O3 but lie along atrend that includes both the bulk host lava and Al-augite.Similarly, olivine in the xenolith and xenocrysts also displayincreased Fe, indicated by their lower Fo numbers (Table 1).
Ti-enrichment of lherzolites by Al-augite veining has been
observed in lherzolitic spinel and clinopyroxene elsewhere(Woodland et al., 1996; Tartarotti et al., 2002); we suggest thatthis process buffers magmatic TiO2 concentrations during par-tial melting of Ti-rich clinopyroxene in the DZBJ source re-gion.

The Debre Zeyit and Butajira regions lie within the radius ofthe proposed flattened Afar plume head (Schilling et al., 1992),suggesting that plume-related magmatism controls xenolith andlava compositions. In the absence of isotopic data it is notpossible to determine the source of the silicate dykes and veins.However, pervasive metasomatism of peridotites by the perco-lation of plume-derived mafic silicate melts has been inferredfor the Mega/Sidamo area (Bedini et al., 1997; Lorand et al.,2003) and a less pervasive plume-derived metasomatism eventis suggested from the Lake Tana xenoliths (Roger et al., 1999).Mafic lavas and Al-augite veins at Debre Zeyit and Butajiracommonly record temperatures in excess of �1280°C, themaximum potential temperature for ordinary mantle melts(MacKenzie, 1988). These values are �200 to 300°C hotterthan those inferred from xenoliths sampled by plateau basalts(Fig. 12), but still within the range of potential temperatures formid–ocean ridge basalt (Herzberg and O’Hara, 2002). Thesethermal considerations are consistent with involvement theAfar plume, but cannot resolve this question entirely given thesmall thermal anomaly and material flux inferred for the Afarplume (Schilling, 1991; Sleep, 2000).

The timing of the Al-augite metasomatic event in the centralMER is unconstrained; however, the lack of equilibrium tex-tures within many Al-augite xenoliths suggests a dynamic,ongoing process. Xenoliths throughout the EARS record rela-tively recent and widespread silicate and carbonatite metaso-matic events. The DZBJ xenoliths indicate a silicate metaso-matism event in the central MER, similar to those inferred forother Ethiopian xenolith localities such as the Pleistocene Si-damo area (Morten et al., 1992; Bedini et al., 1997; Lorand etal., 2003) and the Oligocene to Quaternary Lake Tana region(Roger et al., 1997; Roger et al., 1999; Ayalew et al., 2005).Silicate metasomatism in the EARS is not restricted to Ethio-pia: in Yemen and Saudi Arabia, silicate melt metasomatismreflects mantle processes associated with the rifting of the RedSea (Henjes-Kunst et al., 1990). To the south, Tanzanian xeno-liths exhibit veining that is thought to be the source of K-Fe-Ti-OH-REE metasomatism in the SCLM (Dawson and Smith,1988; Dawson, 2002). In addition to these silicate events, someevidence of carbonatitic metasomatism is evident locally. Shal-low mantle xenoliths in Yemen record a carbonatitic metaso-matism event (Chazot et al., 1996; Baker et al., 1998) during orshortly after the Oligocene flood volcanism in Yemen. Thisevent is interpreted to be related to the initial impact of the Afarplume (Baker et al., 1998). In Tanzania, recent carbonatiticfluid derived from the athenosphere has also had a metasomaticimpact on the regional lithospheric mantle (Rudnick et al.,1993).

5.4. Crustal and Lithospheric Structure

Recent seismic tomography in the central MER (Keranen etal., 2004) has identified high velocity anomalies underlying theactive volcanic zones, including the Debre Zeyit region (Fig.
13). These anomalies are segmented in a fashion similar to the

-1007. ((DZ-1


surface volcanism and may represent the early stages of aseafloor-spreading axis. Geophysical interpretation suggeststhat these anomalies represent mafic intrusions at 10 km depthoverlain by dykes. The highest-velocity anomalies correspondto the magmatic segments proposed from structural analysis byEbinger and Casey (2001) except in the case of the Kokasegment, where the typical NE-SW orientation of the high-velocity anomalies is perturbed towards the Debre Zeyit region(Fig. 13). Additionally, three-dimensional modelling of gravitydata in the Ethiopian rift has indicated the presence of anom-alously hot material beneath both Debre Zeyit/Butajira and theWonjii Fault Belt (Mahatsente et al., 1999). Xenolith evidencepresented here suggests that the large geophysical anomaliesbeneath the currently active volcanic regions reflect pervasivedyking/veining (i.e., to a depth of at least 30 km) rather than asingle large intrusion. These dykes are also consistent with theregional SKS splitting anisotropy, which has been interpretedas resulting from the alignment of melt-filled cracks or dykes,preferentially along the rift margins (Kendall et al., 2005).

The geochemical, geophysical and morphologic similaritiesbetween Debre Zeyit/Butajira and main rift-axis volcanism arestriking; the Debre Zeyit region in particular shares common-

Fig. 12. Microprobe clinopyroxene and olivine chemicaand xenoliths throughout the Debre Zeyit and Butajira regpresence of Group C fragments. (A) Clinopyroxene vClinopyroxene phenocrysts and xenocrysts in sample DZmaterial. (D) Group C clinopyroxene in a single xenolith

alities in the mode of basaltic volcanism, which is identical to

the main-rift axis (cinder cones and minor flows between largesilicic centres). Such similarities in addition to the presence ofan upper crustal high velocity anomaly (Keranen et al., 2004),and a low velocity P-wave anomaly at 75 km interpreted to bea melting zone (Bastow et al., 2005), raise the possibility thatthe Debre Zeyit region is a zone of extension and focuseddecompressional melting.

Questions remain as to the timing and source of the dykesfrom which the Al-augite xenoliths are derived. Seismic evi-dence of underplating (�7.4 km/s layer), interpreted to beOligocene in age, is evident beneath the adjoining westernplateau (Mackenzie et al., in press). The formation of high-velocity lower crust (�7.4 km/s, e.g., cumulate underplate), isconsidered a characteristic of plume-crust interaction alongvolcanic rifted margins (Farnetani et al., 1996; Condie, 1999;Menzies et al., 2002). Recent geophysical investigation usingreceiver function analysis has determined the depth to theMoho in the region to be between 27 and 35 km (Dugda et al.,2005). However, the large-scale replacement of the felsic crustwith mafic material in the MER north of �8°N (Dugda et al.,2005) make the geochemical distinction between the lowercrust and underlying lithospheric mantle unclear at best. Pres-

sitions plotted as a function of Mg# and MgO in host lava-1007 was chosen for more detailed attention because ofin xenocrysts and phenocrysts in DZBJ basalts. (B)

C) Olivine chemical variation in host lava and Al-augite007).

l compoion. DZariation

sure estimates derived from the xenoliths in this study indicate

shownregion


depths of origin of 0.7–31 km i.e., from the base of the crust tonear the surface, whereas cumulate underplate is expected atthe base of the crust (Farnetani et al., 1996). The proposeddepth of melt generation of the host lavas (�53–88 km) is wellbelow that of the suggested cumulate underplate and corre-sponds to a low velocity P and S wave anomaly in the upper100 km of the mantle interpreted as partial melt (Bastow et al.,2005). The data presented here, however cannot preclude theDebre Zeyit/Butajira xenoliths from having been emplaced asdykes/veins related to an Oligocene plume-head underplate.

6. SUMMARY

Xenoliths and host lavas from the Debre Zeyit and Butajiraregions in the central Main Ethiopian Rift provide importantinsight into the study of continental rifting and the transition tosea-floor spreading. The primitive host lavas contain composi-tional information on the current source of rift magmatismwhile the xenoliths document magmatic processes taking placewithin the rift. Specifically this study has found that:

i). The Quaternary basaltic products in the Debre Zeyit andButajira regions of the central Main Ethiopian Rift containAl-augite (Group A), norite (Group B) and rare lherzolitemantle xenoliths (Group C).

ii). The host magmas of these xenoliths are derived throughmoderate-degree partial melting at 15–25 kb of a fertileperidotite source that has not experienced significant car-

Fig. 13. Debre Zeyit Quaternary basalt region andtomography (Kernan et al., 2004). Central MER magmaSeismic tomography (Kernan et al., 2004) at 10 km depthNazret and Sabure magmatic segments in addition to the

bonatitic or hydrous metasomatism.

iii). The entrained Al-augite xenoliths and megacrysts indicatethat a process of dyking/veining affects the crust andlithospheric mantle in the central MER to a depth of atleast 30 km. The wide range in pressure estimates of theAl-augite xenoliths precludes their derivation solely froma cumulate underplate. Crystal fractionation within theseveins results in phenocryst compositions that are moreevolved than those found in host lavas that entrain thexenoliths. The norite xenoliths are derived from shallowcumulates that were entrained shortly before eruption.

iv). Al-augite dyking and veining has metasomatised the litho-spheric mantle by enriching lherzolite wall rocks in Ti andFe (e.g., Irving, 1980; Menzies et al., 1987). The dykes/veins record temperatures significantly hotter (�200°C)than the wall-rock and may be an important heat-transfermechanism within the rift.

v). The spatial distribution of Al-augite xenoliths is consistentwith recent geophysical evidence and suggests that thisveining is regionally pervasive, occurring in associationwith the recent zones of focused strain (the Silti-DebreZeyit Fault Zone and the Wonjii Fault Belt) in the MER.Al-Augite xenoliths have previously been observed inmodern regions of continental rifting: Durango, Lake Bail-kal and now in the Main Ethiopian Rift. The occurrence ofsuch dykes/veins in rift settings suggests a shared processassociated with continental rifting and may be a key pro-cess in the transition from continental rifting to seafloor

locations (starred) in comparison to shallow seismicents (Ebinger and Casey, 2001) are indicated (dotted).in grayscale. High-velocity anomalies coincide with the

of Debre Zeyit Quaternary basaltic activity.

sampletic segm

spreading.


The process of Al-augite veining of the lithosphere clearlyplays an important role in the transition of magmatic andtectonic styles observed in the MER, and we expect this processalso occurs in other continental rifts. Further investigation ofthese xenoliths (e.g., trace element information) and host lavas(e.g., isotopic constraints such as Sr, Nd and Pb systematics)can provide more detail such as the source of these veins, theirevolution, and how they may have interacted with the SCLM inaddition to constraints as to the current source(s) of rift magam-stism. Combining these geochemical techniques with furthergeophysical investigations in the Debre Zeyit and Butajiraregions will also reveal more clearly the processes associatedwith the continental-oceanic transition.

Acknowledgments—This research was supported by National ScienceFoundation grant EAR 0207764 (T. Furman) as part of the cross-disciplinary Ethiopia Afar Geoscientific Lithosphere Experiment (EA-GLE). A George H. Deike Jr. grant to T. Furman provided additionalsupport for analytical work. TR thanks Kassahun Ejeta, guide anddriver, and Roeland Doust, field assistant during fieldwork in Ethiopia.Margaret Nitz and Leigh Patterson were helpful in sample preparation.Brian LeVay performed some last-minute petrologic descriptions. Weare grateful to Mark Angelone for help with the Electron Microprobe atPenn State, and Emily Klein, Gary Dwyer and Mark Rudnicki of DukeUniversity for performing DCP and ICP-MS analysis. Thanks also toPeter Maguire, Graeme MacKenzie, Katie Keranen, Barry Hanan,William Hart and Martin Menzies for thoughtful reviews that improvedthe manuscript.

Associate editor: M. Menzies

REFERENCES

Abebe T., Mazzarini F., Innocenti F., and Manetti P. (1998) TheYerer-Tullu Wellel volcanotectonic lineament: A transtensionalstructure in central Ethiopia and the associated magmatic activity.J. Afr. Earth Sci. 26, 135–150.

Ayalew D. (2000) Crustal control in the genesis of Plio-Quaternarybimodal magmatism of the Main Ethiopian Rift (MER): Geochemi-cal and isotopic (Sr, Nd, Pb) evidences—Discussion. Chem. Geol.168, 1–3.

Ayalew D., Yirgu G., Ketefo E., Barbey P., and Ludden J. (2003)Intrusive equivalents of flood volcanics: Evidence from petrologyof xenoliths in Quaternary Tana basanites. Ethiopian J. Sci. 26,93–102.

Baker M. B. and Stolper E. M. (1994) Determining the composition ofhigh-pressure mantle melts using diamond aggregates. Geochim.Cosmochim. Acta 58, 2811–2827.

Baker J., Chazot G., Menzies M., and Thirlwall M. (1998) Metasoma-tism of the shallow mantle beneath Yemen by the Afar plume—Implications for mantle plumes, flood volcanism and intraplatevolcanism. Geology 26, 431–434.

Barrat J. A., Fourcade S., Jahn B. M., Cheminee J. L., and CapdevilaR. (1998) Isotope (Sr, Nd, Pb, O) and trace-element geochemistryof volcanics from the Erta’Ale range (Ethiopia). J. Volcanol. Geo-therm. Res. 80, 85–100.

Bastow I. D., Stuart G. W., Kendall J.-M., and Ebinger C. J. (2005)Upper-mantle seismic structure in a region of incipient continentalbreakup: Northern Ethiopian rift. Geophysical J. Int., in press.

Bedini R. M., Bodinier J. L., Dautria J. M., and Morten L. (1997)Evolution of LILE-enriched small melt fractions in the lithosphericmantle: A case study from the East African Rift. Earth Planet. Sci.Lett. 153, 67–83.

Bedini R. M. and Bodinier J. L. (1999) Distribution of incompatibletrace elements between the constituents of spinel peridotite xeno-liths: ICP-MS data from the East African Rift. Geochim. Cosmo-chim. Acta 63, 3883–3900.

Benvenuti M., Carnicelli S., Belluomini G., Dainelli N., Di Grazia S.,Ferrari G. A., Iasio C., Sagri M., Ventra D., Atnafu B., and Kebede

S. (2002) The Ziway-Shala lake basin (Main Ethiopian Rift,Ethiopia): A revision of basin evolution with special reference tothe Late Quaternary. J. Afr. Earth Sci. 35, 247–269.

Bodinier J.-L., Menzies M. A., Shimizu N., Frey F. A., and McPhersonE. (2004) Silicate, hydrous and carbonate metasomatism at Lherz,France: Contemporaneous derivatives of silicate melt-harzburgitereaction. J. Petrol. 45, 229–320.

Buck W. R. (2004) Consequences of asthenospheric variability oncontinental rifting. In Rheology and Deformation of the Lithosphereat Continental Margins (eds. G. Karner, B. Taylor, N. W. Driscolland D. L. Kohlstedt), pp. 1–30. Columbia University Press.

Chazot G., Menzies M., and Harte B. (1996) Silicate glasses in spinellherzolites from Yemen: Origin and chemical composition. Chem.Geol. 134, 159–179.

Chernet T., Hart W. K., Aronson J. L., and Walter R. C. (1998) Newage constraints on the timing of volcanism and tectonism in thenorthern Main Ethiopian Rift–southern Afar transition zone (Ethi-opia). J. Volcanol. Geotherm. Res. 80 (3–4), 267–280.

Condie K. C. (1999) Mafic crustal xenoliths and the origin of the lowercontinental crust. Lithos 46, 95–101.

Conticelli S., Sintoni M. F., Abebe T., Mazzarini F., and Manetti P.(1999) Petrology and geochemistry of ultramafic xenoliths and hostlavas from the Ethiopian Volcanic Province: An insight into theupper mantle under eastern Africa. Acta Vulcanol. 11, 143–159.

Coulie E., Quidelleur X., Gillot P. Y., Courtillot V., Lefevre J. C., andChiesa S. (2003) Comparative K-Ar and Ar/Ar dating of Ethiopianand Yemenite Oligocene volcanism: Implications for timing andduration of the Ethiopian traps. Earth Planet. Sci. Lett. 206, 477–492.

Courtillot V., Davaille A., Besse J., and Stock J. (2003) Three distincttypes of hotspots in the Earth’s mantle. Earth Planet. Sci. Lett. 205,295–308.

Dawson J. B. (2002) Metasomatism and partial melting in upper-mantle peridotite xenoliths from the Lashaine volcano, northernTanzania. J. Petrol. 43, 1749–1777.

Dawson J. B. and Smith J. V. (1988) Metasomatized and veinedupper-mantle xenoliths from Pello Hill, Tanzania—Evidence foranomalously-light mantle beneath the Tanzanian sector of the East-African Rift-Valley. Contrib. Mineral. Petrol. 100, 510–527.

Deniel C., Vidal P., Coulon C., and Vellutini P. J. (1994) Temporalevolution of mantle sources during continental rifting—The volca-nism of Djibouti (Afar). J. Geophys. Res. Solid Earth 99(B2),2853–2869.

Deniel C., Trua T., and Mazzuoli R. (2000) Crustal control in thegenesis of Plio-Quaternary bimodal magmatism of the Main Ethi-opian Rift: Geochemical and isotopic (Sr, Nd, Pb) evidence—Reply. Chem. Geol. 168, 5–7.

Dugda M. T., Nyblade A. A., Julia J., Langston C. A., Ammon C. J.,and Simiyu S. (2005) Crustal structure in Ethiopia and Kenya fromreceiver function analysis. J. Geophys. Res. B Solid Earth Planets110 (B1),B01303.

Dupuy C., Liotard J. M., and Dostal J. (1992) Zr/Hf fractionation inintraplate basaltic rocks—Carbonate metasomatism in the mantlesource. Geochim. Cosmochim. Acta 56, 2417–2423.

Ebinger C. J. and Casey M. (2001) Continental breakup in magmaticprovinces: An Ethiopian example. Geology 29, 527–530.

Farnetani C. G., Richards M. A., and Ghiorso M. S. (1996) Petrologicalmodels of magma evolution and deep crustal structure beneathhotspots and flood basalt provinces. Earth Planet. Sci. Lett. 143,81–94.

Foley S. (1992) Vein-plus-wall-rock melting mechanisms in the litho-sphere and the origin of potassic alkaline magmas. Lithos 28,435–453.

Frey F. A. and Prinz M. (1978) Ultramafic inclusions from San Carlos,Arizona: Petrologic and geochemical data bearing on their petro-genesis. Earth Planet. Sci. Lett. 38, 129–176.

Furman T. (1995) Melting of metasomatized subcontinental litho-sphere—Undersaturated mafic lavas from Rungwe, Tanzania. Con-trib. Mineral. Petrol. 122, 97–115.

Furman T. and Graham D. (1999) Erosion of lithospheric mantlebeneath the East African Rift system: Geochemical evidence from
the Kivu volcanic province. Lithos 48, 237–262.


Furman T., Bryce J. G., Karson J., and Iotti A. (2004) East African Riftsystem (EARS) plume structure: Insights from Quaternary maficlavas of Turkana, Kenya. J. Petrol. 45, 1069–1088.

Furman T., Bryce J. G., Rooney T., Hanan B. B., Yirgu G., and AyalewD. (2005) Heads and tails: 30 million years of the Afar Plume. JGeol. Soc. London. Accepted.

Gasparon M., Innocenti F., Manetti P., Peccerillo A., and Tsegaye A.(1993) Genesis of the Pliocene to recent bimodal mafic-felsicvolcanism in the Debre Zeyt area, central Ethiopia: Volcanologicaland geochemical constraints. J. Afr. Earth Sci. 17, 145–165.

George R., Rogers N., and Kelley S. (1998) Earliest magmatism inEthiopia: Evidence for two mantle plumes in one flood basaltprovince. Geol. 26, 923–926.

Ghorbani M. R. and Middlemost E. A. K. (2000) Geochemistry ofpyroxene inclusions from the Warrumbungle Volcano, New SouthWales, Australia. Am. Mineral. 85, 1349–1367.

Gorring M. L. and Kay S. M. (2000) Carbonatite metasomatizedperidotite xenoliths from southern Patagonia: Implications for litho-spheric processes and Neogene plateau magmatism. Contrib. Min-eral. Petrol. 140, 55–72.

Griffin W. L., Wass S. Y., and Hollis J. D. (1984) Ultramafic xenolithsfrom Bullenmerri and Gnotuk Maars, Victoria, Australia: Petrologyof a sub-continental crust-mantle transition. J. Petrol. 25, 53–87.

Griffin W. L., O’Reilly S. Y., and Stabel A. (1988) Mantle metasoma-tism beneath Western Victoria, Australia. 2. Isotopic geochemistryof Cr-diopside iherzolites and Al-augite pyroxenites. Geochim.Cosmochim. Acta 52, 449–459.

Hart W. K., WoldeGabriel G., Walter R. C., and Mertzman S. A.(1989) Basaltic volcanism in Ethiopia: Constraints on continentalrifting and mantle interactions. J. Geophys. Res. B Solid EarthPlanets 94, 7731–7748.

Hart S. R. and Reid M. R. (1991) Rb/Cs fractionation—A link betweengranulite metamorphism and the S-process. Geochim. Cosmochim.Acta 55, 2379–2383.

Henjes-Kunst F., Altherr R., and Baumann A. (1990) Evolution andcomposition of the lithospheric mantle underneath the WesternArabian Peninsula—Constraints from Sr-Nd isotope systematics ofmantle xenoliths. Contrib. Mineral. Petrol. 105, 460–472.

Herzberg C. and O’Hara M. J. (2002) Plume-associated ultramaficmagmas of phanerozoic age. J. Petrol. 43, 1857–1883.

Hoernle K. and Schmincke H.-U. (1993) The petrology of the tholeiitesthrough melilite nephelinites on Gran Canaria, Canary Islands:Crystal fractionation, accumulation and depths of melting. J.Petrol. 34, 573–597.

Hofmann A. W., Jochum K. P., Seufert M., and White W. M. (1986)Nb and Pb in oceanic basalts—New constraints on mantle evolu-tion. Earth Planet. Sci. Lett. 79, 33–45.

Irving A. J. (1980) Petrology and geochemistry of composite ultramaficxenoliths in alkalic basalts and implications for magmatic processeswithin the mantle. Am. J. Sci. 280-A, 389–426.

Irving A. J. and Frey F. A. (1984) Trace element abundances inmegacrysts and their host basalts: Constraints on partition coeffi-cients and megacryst genesis. Geochim. Cosmochim. Acta 48,1201–1221.

Kendall J.-M., Stuart G. W., Ebinger C., Bastow I. D., and Keir D.(2005) Magma-assisted rifting in Ethiopia. Nature 433, 146–148.

Keranen K., Klemperer S. L., Gloaugen R., and the EAGLE workinggroup. (2004) Three-dimensional seismic imaging of a protoridgeaxis in the Main Ethiopian Rift. Geology 32, 949–952.

Kieffer B., Arndt N., Lapierre H., Bastien F., Bosch D., Pecher A.,Yirgu G., Ayalew D., Weis D., Jerram D. A., Keller F., andMeugniot C. (2004) Flood and shield basalts from Ethiopia: Mag-mas from the African superswell. J. Petrol. 45, 793–834.

Knight K. M. (2002) Tertiary chemical structure of the Afar mantleplume: Evidence from primitive mafic lavas from Turkana, N.Kenya. Master’s thesis. The Pennsylvania State University.

Korme T., Chorowicz J., Collet B., and Bonavia F. F. (1997) Volcanicvents rooted on extension fractures and their geodynamic implica-tions in the Ethiopian Rift. J. Volcanol. Geotherm. Res. 79, 205–222.

Kushiro I. (1996) Partial melting of a fertile mantle peridotite at high
pressures: An experimental study using aggregates of diamond. InEarth Processes: Reading the Isotopic Code (eds. A. Basu and S.
Hart), pp. 109–122. Geophysical Monograph 95. AmericanGeophysical Union.

Le Bas M. J., Le Maitre R. W., Streckeisen A., and Zanettin B. (1986)A chemical classification of volcanic rocks based on total Alkali-Silica diagram. J. Petrol. 27, 745–750.

Lindsley D. H. (1983) Pyroxene thermometry. Am. Mineral. 68, 477–493.

Litasov K. D., Foley S. F., and Litasov Y. D. (2000) Magmaticmodification and metasomatism of the subcontinental mantle be-neath the Vitim volcanic field (East Siberia): Evidence from traceelement data on pyroxenite and peridotite xenoliths from Miocenepicrobasalt. Lithos 54, 83–114.

Lorand J.-P., Reisberg L., Bedini R. M., Horan M. F., Brandon A. D.,and Neal C. R. (2003) Platinum-group elements and melt percola-tion processes in Sidamo spinel peridotite xenoliths, Ethiopia, EastAfrican Rift. Chem. Geol. 196, 57–75.

Loucks R. R. (1996) A precise olivine-augite Mg-Fe-exchange geo-thermometer. Contrib. Mineral. Petrol. 125, 140–150.

Luhr J. F., Henry C. D., Housh T. B., Aranda-Gomez J. J., andMcIntosh W. C. (2001) Early extension and associated mafic alka-lic volcanism from the southern Basin and Range Province: Geol-ogy and petrology of the Rodeo and Nazas volcanic fields, Du-rango, Mexico. Geol. Soc. Am. Bull. 113, 760–773.

Macdonald R. (1994) Petrological evidence regarding the evolution ofthe Kenya Rift valley. Tectonophysics 236, 373–390.

Mahatsente R., Jentzsch G., and Jahr T. (1999) Crustal structure of theMain Ethiopian Rift from gravity data: 3-dimensional modeling.Tectonophysics 313, 363–382.

Mazzarini F., Abebe T., Innocenti F., Manetti P., and Pareschi M. T.(1999) Geology of the Debre Zeyt area (Ethiopia) (with a geologicmap at scale 1:100,000). Acta Vulcanol. 11, 131–141.

McKenzie D. and Bickle M. J. (1988) The volume and composition ofmelt generated by extension of the lithosphere. J Petrol. 29, 625–679.

MacKenzie G., Thybo H., and Maquire P. (in press) Crustal velocitystructure across the Main Ethiopian Rift: Results from 2-dimen-sional wide-asle seismic modelling. Geophys. J. Intern.

Menzies M., Rogers N., Tindle A. and Hawkesworth C. J. (1987)Metasomatic enrichment processes in lithospheric peridotites, aneffect of athenosphere-lithosphere interaction. In Mantle Metaso-matism (eds. M. Menzies and C. J. Hawkesworth), pp. 313–361.Academic Press.

Menzies M. A., Klemperer S. L., Ebinger C. J., and Baker J. (2002)Characteristics of volcanic rifted margins. Special Paper 362, pp.1–14. Geological Society of America.

Mohr P. A. (1961) The geology, structure and origin of the Bishoftuexplosion craters. Bull. Geophys. Observ. Addis Ababa 2, 65–101.

Mohr P. A. (1967) Major volcano-tectonic lineament in the Ethiopianrift system. Nature 213, 664–665.

Morten L., Defrancesco A. M., Bonavia F. F., Haileselassie G., Bar-gossi G. M., and Bondi M. (1992) A new mantle xenolith localityfrom Southern Ethiopia. Min. Mag. 56, 422–425.

Morton W. H., Rex D. C., Mitchell J. G., and Mohr P. (1979) Riftwardyounging of volcanic units in the Addis Ababa region, EthiopianRift valley. Nature 280, 284–288.

Mukasa S. B. and Shervais J. W. (1999) Growth of subcontinentallithosphere: Evidence from repeated dike injections in the Balmuc-cia lherzolite massif, Italian Alps. Lithos 48, 287–316.

Nimis P. and Ulmer P. (1998a) Clinopyroxene geobarometry of mag-matic rocks, part 1: An expanded structural geobarometer for an-hydrous and hydrous, basic and ultrabasic systems. Contrib. Min-eral. Petrol. 133, 122–135.

Nimis P. and Ulmer P. (1998b) Clinopyroxene geobarometry of mag-matic rocks, part 1: An expanded structural geobarometer for an-hydrous and hydrous, basic and ultrabasic systems. Contrib. Min-eral. Petrol. 133, 314–314.

O’Reilly S. Y. and Griffin W. L. (1988) Mantle metasomatism beneathWestern Victoria, Australia. 1. Metasomatic processes in Cr-diop-side lherzolites. Geochim. Cosmochim. Acta 52, 433–447.

Peccerillo A., Barberio M. R., Yirgu G., Ayalew D., Barbieri M., andWu T. W. (2003) Relationships between mafic and peralkaline
silicic magmatism in continental rift settings: A petrological, geo-


chemical and isotopic study of the Gedemsa volcano, centralEthiopian rift. J. Petrol. 44, 2003–2032.

Pik R., Deniel C., Coulon C., Yirgu G., and Marty B. (1999) Isotopicand trace element signatures of Ethiopian flood basalts: Evidencefor plume-lithosphere interactions. Geochim. Cosmochim. Acta 63,2263–2279.

Roger S., Pik R., Dautria J. M., Coulon C., Yirgu G., Ayalew D., andLegros P. (1997) Rifting actif ou passif en Ethiopie? Elements dereponse apportes par l’etude des xenolites peridotitiques de laregion du lac Tana. Active or passive rifting in Ethiopia? Contri-bution of peridotitic xenoliths from the Lake Tana area. Comp.Rendu. Acad. Sci. Ser. II 324, 1009–1016.

Roger S., Dautria J. M., Coulon C., Pik R., Yirgu G., Michard A.,Legros P., and Ayalew D. (1999) An insight on the nature, com-position and evolution of the lithospheric mantle beneath the north-western Ethiopian Plateau: The ultrabasic xenoliths from the TanaLake Province. Acta Vulcanol. 11, 161–168.

Rogers N. W., James D., Kelley S. P., and de Mulder M. (1998) Thegeneration of potassic lavas from the eastern Virunga Province,Rwanda. J. Petrol. 39, 1223–1247.

Rudnick R. L., McDonough W. F., and Chappell B. W. (1993) Car-bonatite metasomatism in the Northern Tanzanian mantle—Petro-graphic and geochemical characteristics. Earth Planet. Sci. Lett.114, 463–475.

Schilling J.-G. (1991) Fluxes and excess temperatures of mantleplumes inferred from their interaction with migrating mid-oceanridges. Nature 352, 397–403.

Schilling J.-G., Kingsley R. H., Hanan B. B., and McCully B. L. (1992)Nd-Sr-Pb isotopic variations along the Gulf of Aden: Evidence forAfar mantle plume–continental lithosphere interaction. J. Geophys.Res. B Solid Earth Planets 97, 10927–10966.

Schulze D. J. (1987) Megacrysts from alkalic volcanic rocks. In MantleXenoliths (ed. P. H. Nixon), pp. 433–451. Wiley.

Shaw C. S. J. (2004) The temporal evolution of three magmatic systemsin the West Eifel volcanic field, Germany. J. Volcanol. Geotherm.Res. 131, 213–240.

Shaw C. S. J. and Eyzaguirre J. (2000) Origin of megacrysts in themafic alkaline lavas of the West Eifel volcanic field, Germany.
Lithos 50, 75–95.
Sleep N. H. (1990) Hotspots and mantle plumes—some phenomenol-ogy. J Geophys. Res. B Solid Earth Planets 95, 6715–6736.

Späth A., Le Roex A. P., and Opiyo-Akech N. (2001) Plume-lithos-phere interaction and the origin of continental rift-related alkalinevolcanism: The Chyulu Hills volcanic province, southern Kenya. J.Petrol. 42, 765–787.

Sun S.-S. and McDonough W. F. (1989) Chemical and isotopic sys-tematics of oceanic basalts: Implications for mantle compositionand processes. In Magmatism in the Ocean Basins (ed. A. D.Saunders), pp. 313–345. Special Publication 42. Geological Soci-ety.

Tartarotti P., Susini S., Nimis P., and Ottolini L. (2002) Melt migrationin the upper mantle along the Romanche Fracture Zone (equatorialAtlantic). Lithos 63, 125–149.

Trua T., Deniel C., and Mazzuoli R. (1999) Crustal control in thegenesis of Plio-Quaternary bimodal magmatism of the Main Ethi-opian Rift (MER): Geochemical and isotopic (Sr, Nd, Pb) evidence.Chem. Geol. 155, 201–231.

Wasylenki L. E., Baker M. B., Kent A. J. R., and Stolper E. M. (2003)Near-solidus melting of the shallow upper mantle: Partial meltingexperiments on depleted peridotite. J. Petrol. 44, 1163–1191.

Wilshire H. G. and Shervais J. W. (1975) Al-augite and Cr-diopsideultramafic xenoliths in basaltic rocks from western United States.Phys. Chem. Earth 9, 257–272.

WoldeGabriel G. (1988) Volcanotectonic history of the central sectorof the Main Ethiopian Rift: A geochronological, geochemical andpetrological approach. Ph.D. thesis. Case Western Reserve Univer-sity.

WoldeGabriel G., Aronson J. L., and Walter R. C. (1990) Geology,geochronology and rift basin development in the central sector ofthe Main Ethiopia Rift. Geol. Soc. Am. Bull. 102, 439–458.

Woodland A. B., Kornprobst J., McPherson E., Bodinier J. L., andMenzies M. A. (1996) Metasomatic interactions in the lithosphericmantle: Petrologic evidence from the Lherz Massif, FrenchPyrenees. Chem. Geol. 134, 83–112.

Yaxley G. M., Green D. H., and Kamenetsky V. (1998) Carbonatitemetasomatism in the southeastern Australian lithosphere. J. Petrol.
39, 1917–1930.

Structure of the Ethiopian lithosphere: Xenolith evidence in the Main Ethiopian Rift

Documents