Geochemical Journal Vol 40 pp 47 to 56 2006Copyright 2006 by The Geoc - PDF document

1 Geochemical Journal, Vol. 40, pp. 47 to
Geochemical Journal, Vol. 40, pp. 47 to 56, 2006Copyright © 2006 by The Geochemical Society of Japan. Sub-lithospheric source for Quaternary alkaline Tepi shield, southwest EthiopiaYALEWARTY P. B and EDepartment of Earth Sciences, Addis Ababa University, P.O. Box 729/1033, Addis Ababa, EthiopiaCRPG-CNRS, B.P. 20, F-54501 Vandoeuvre-les-Nancy Cedex, FranceReceived November 15, 2004; Accepted June 29, 2005Mineral chemistry, major and trace elements and Sr-Nd isotopes are presented for basalts and trachybasalts from theQuaternary alkaline Tepi shield from southwest Ethiopia. The lavas are variably porphyritic with phenocrysts of olivine), clinopyroxene (Wo) and Ti-magnetite set in a microcrystalline matrix.6.2 wt.%). The major element variations can be explained by olivine + clinopyroxene plagioclase fractionation, mostlikely consistent with low P fractionation. The geochemical variations observed in the Tepi lavas are interpreted in termsKeywords: alkaline lava, southwest Ethiopia, primitive mantle, Tepi shieldThe small number of samples makes it difficult to es-alkalic magmatism in Ethiopia and, perhaps, other por-tions of the east African rift system.The earliest Tertiary volcanism (basalt and associatedrhyolite) in east Africa is found in south and southwestmiddle Eocene ca. 49 Ma (Davidson and Rex, 1980) andhas continued episodically to recent times (Ayalew ., 1993; George WoldeGabriel Davidson and Rex (1980) broadly divided the TertiaryMa). The main sequence (bimodal basalt and rhyolite)which deposited on a peneplain during early Tertiary. Theneath them. The post-rift sequence (19 Ma to presen

2 t)., 1998; Ethiopia, Keiffer 2004). Thes
t)., 1998; Ethiopia, Keiffer 2004). These low volume alkaline rocks are either verypost-date the main tholeiitic eruptive phase. The originminor, still they can provide keys to the nature of themantle and important clues as to the EarthÕs works.why small degree, alkalic melts are exposed from Tanza-nia in the south to as far as Israel. The alkaline rocks ofeast Africa are important corollaries to the main phase ofTertiary volcanism. This paper presents geochemical datafor a suite of Quaternary alkaline lavas from southwestEthiopia (near Tepi, Fig. 1), a remote and poorly studiedregion of the Ethiopian flood basalt province. The sam-Geosciences Cooperation Project (Ethiopia 2000). These 48D. Ayalew Plio - Quaternary volcanicsMiocene basaltsOlig. basalt & rhyoliteBasement 38 km Gore Teferi Gulf ofAdenER Fig. 1b Throughout the Oligocene, a broad arch developedconcave to the northeast (Davidson and Rex, 1980). Thisuplift allowed erosion of the earlier volcanic cover, re-late Miocene ca. 13., 1993), al-though shallow proto-rifts may have existed somewhatQuaternary volcanism in the Tepi area is related tofaults, Davidson, 1983). The Quaternary volcanism cancones. There is no geochronological data available forEOCHEMISTRYPetrography and mineral chemistryThe Tepi lavas have aphyric to mildly porphyritic tex-(Wophase as the phenocryst assemblage. Phenocryst propor-tions range from 515 vol.% in the basalts to 10 vol.% inthe trachybasalts. Ti-magnetites occurring as smalleuhedral grains are an important phase of the matrix. Ta-clinopyroxene, plagioclase and Ti-magnetites embeddedFig. 1. (a) Location map of

3 the Afro-Arabian large igneous province
the Afro-Arabian large igneous province and the Afar triple rift junction. (b) Generalized geologimap of the southwest Ethiopia volcanic rocks and other lithologic units (modified after Merla et al., 1979). Small numbers (witthe prefix 98-) refer to sample numbers. Quaternary alkaline Tepi shield49Major and trace element data for the Tepi Quaternarybasalts and hawaiites are reported in Table 2. The lavasplot in the alkaline field on the TAS (total alkali-silica)diagram (Fig. 2), and they are nepheline normative. Thefrom 4.7 to 6.2 wt.% and Mg values (Mg# = Mg/(Mg +fractionation. Two alkali basalts (samples 98and Mg values (6465), and are inferred to have under-Bivariate major element plots (Fig. 3) show that thelavas display an evolutionary trend; (i) Al contents of�among samples with 4.8 wt.% MgO, then decreaseinfers deeper level processes, (iii) an increase in TiO in-dicating that Fe-Ti oxides were not involved in the behaves as an in-compatible element within the suite. The overall dataTrace elementsThe two high-MgO lavas have relatively elevated Nisamples have low contents of Ni of 2169 ppm and Cr ofstrong depletion in Ni (4.2 ppm), Cr (1.2 ppm) and Curole in the evolution of the rocks. However, the smallnumber of samples makes it difficult to be quantitativeabout fractionation processes. Variations of incompatibletrace element ratios (e.g., Rb/Sr-Ba/Rb, Fig. 4c) show that = 5.0 = 1.3Table 1. Compositions of representative phenocrysts from Tepi lavas PhaseOlivineClinopyroxeneTi-magnetitePlagioclase Sample98-1798-1998-1698-1798-1698-1698-1998-1698-1998-1798-16 39.9440.2336.27SiO49.144

4 7.0548.18SiO0.000.09SiO50.5650.4852.37 0
7.0548.18SiO0.000.09SiO50.5650.4852.37 0.000.000.12TiO1.571.672.46TiO20.1518.06Al31.1430.9528.90FeO12.7612.0329.22Al3.626.933.90Al3.033.41FeO0.500.540.78MnO0.360.220.75FeO7.728.839.46FeO62.6768.29CaO13.7914.2711.39MgO46.4147.1132.68MnO0.040.170.29MnO0.651.01NaO3.513.684.49CaO0.290.290.33MgO13.7812.8012.71MgO3.703.75KO0.200.140.55NiO0.220.160.02CaO22.2820.6621.43Cr7.531.15Total99.70100.0598.48Total99.98100.0499.39NaO0.410.950.60Total97.7395.76Total98.5699.0599.03%An67.767.756.5%Fo86.687.566.6%Ab31.131.540.3%Wo46.945.445.9%Or1.20.83.2%En40.339.137.8%Fs12.815.416.3 on microprobe at the University of PointcarŽ in Nancy, France. Typical working conditions were 15 kV accelerating potential, 10 nAcurrent and a beam diameter of 1 mm. Total for Ti-magnetite is low and that is because of Fe problem. 50D. Ayalew Table 2. Major (wt.%) and trace (ppm) element data for the Tepi lavasMajor elements were determined by ICP-AES and trace elements by ICP-MS at CRPG, Nancy, France. Total Fe as Fe Rock typeBasaltHawaiiteSample98-1998-1798-1198-1598-1298-1698-2098-18SiO46.746.747.847.346.048.947.448.7TiO1.871.902.403.553.612.082.212.3815.415.517.315.214.917.216.517.911.911.913.315.515.310.813.112.3MnO0.160.160.250.150.250.170.200.18MgO10.810.65.54.84.75.36.24.7CaO8.78.78.38.09.57.78.46.7O2.933.013.623.563.854.684.094.62O0.991.040.971.240.932.241.141.640.520.520.580.731.080.850.690.86Mg#64.564.045.138.137.849.648.843.4CIPW normPl49.54958.1751.849.646.351.8458.2Or5.796.035.677.335.513.16.689.57Ne0.390.940.0001.836.433.071.97Di11.111.37.7410.816.110.811.253.3Hy004.036.72000.000Ol2322.512.999.4412.112.115.8814.9Il3.513.57

5 4.526.726.863.894.164.48Mt5.65.515.525.5
4.526.726.863.894.164.48Mt5.65.515.525.555.555.55.525.61Ap1.181.181.341.692.481.951.601.97Ba213.7213.2355.5550.5485.8949.6265.4624.4Cr528.7528.226.319.61.273.1121.018.3Cs0.200.250.270.830.300.570.270.18Cu54.054.520.739.419.333.842.238.6Ga17.017.819.826.119.418.419.518.1Hf3.683.943.626.153.673.364.315.69Nb23.224.928.027.530.478.929.047.1Ni262.0270.428.045.04.238.169.321.3Pb2.202.663.424.763.884.683.093.61Rb18.219.617.722.017.861.423.224.6Sr451.9476.0575.7496.2701.21248.9597.0849.2Ta1.531.672.001.882.335.351.993.00Th1.882.202.532.992.837.412.623.61U0.480.550.390.690.661.260.700.97V173.0171.1255.2274.9221.6183.3182.3110.0Y23.425.227.938.830.026.928.833.1Zn87.687.9115.9172.4132.696.5106.389.5Zr182.8193.4149.8286.6150.5152.5215.8298.1La18.319.724.335.833.754.826.234.3Ce41.243.449.978.972.099.457.673.2Pr5.045.356.4610.069.1211.097.099.08Nd21.922.526.543.240.542.430.237.6Sm4.514.875.848.798.367.396.427.51Eu1.611.791.832.812.722.432.042.62Gd4.294.725.448.337.486.145.566.81Tb0.650.690.861.251.130.840.830.93Dy4.244.245.007.105.964.814.956.12Ho0.790.820.891.361.070.890.971.14Er1.922.062.513.432.762.492.312.88Tm0.320.320.390.500.380.300.360.48Yb1.832.142.543.012.202.262.212.72Lu0.300.330.400.510.410.360.350.50 Quaternary alkaline Tepi shield51 4045505560HawaiiteTephriteMugearite (wt.%)36912 0.30.40.50.6MgO (wt.%)CaO/Al TiO (wt.%) 0.40.60.81.01.236912 (wt.%) Fig. 2. Total alkalis v. silica (TAS, Le Bas et al., 1986) dia-gram for Tepi lavas, showing an alkaline composition. Alka- = 11.2), resulting in crossing REE pat-terns. This sample is also richer in incompatible elements(Ba, Nb, Rb, Sr, Ta, Th and U) tha

6 n any other lava.ible elements. The two
n any other lava.ible elements. The two primitive rocks do not have posi-tive Ba anomaly, but the more evolved samples showpositive Ba anomalies which call for the effects ofSr-Nd isotopesmeasured Sr and Nd isotopic ratios (Table 3) were notSm and theyconsidered as initial ratios. The Tepi lavas have ratios from 0.51280 to 0.51284. Although the Sr-Nd iso-Fig. 3. Bivariate major element plots for Tepi basalts (recalculated to 100%, volatile-free). Note Al show inflected trends, 52D. Ayalew As discussed in the previous section the major andtrace element variations observed in the Tepi lavas con-firm the importance of fractionation in the evolution ofthe rocks. This section focuses on the role of contamina-Lithospheric source region v. Crustal contaminationBefore discussing the source of the Tepi lavas, it iset al of Hofmann 5),lithospheric source material for mafic magmatism or al-uids. Then some strong case must be made for being ableto distinguish lithosphere from crust. The evidence be-The two most MgO-rich lavas plot at lower Ce/Pbdirectly from the lithosphere or sublithospheric sources.may be higher. In the plot of Sr against Ce/PbSr values have low Ce/Pb val-ues. This indicator points towards crustal involvement,erupted along the Main Ethiopian Rift). Variations of SrMantle source compositionoccurred to get to ~5 wt.% MgO, the two MgO-rich sam-a source composition. Although the two high-MgO sam-anomaly. The presence of positive Ba anomalies in theevolved lavas calls for the effects of fractionation. Theprimitive samples have elevated Ti/Eu values (6373carbonatite metasomatism. Melts derived

7 from carbonatite La (ppm) 36912MgO (wt.%
from carbonatite La (ppm) 36912MgO (wt.%)Ba (ppm) 0.020.030.040.050.065152535Ba/RbRb/Sr MgO-rich lavas Fig. 4. Bivariate trace element plots for Tepi lavas, showingdence for the state of the mantle source. Variations of Sr Quaternary alkaline Tepi shield53high Zr/Hf, Nb/La and La/Yb ratios and low Ti/Eu (e.g.,the Tepi data, compared with those of southern Ethiopia(George and Rogers, 1999) and Turkana (Furman 2004) shows that the Tepi samples are indeed close toREE profiles (e.g., La/Sm v. Sm/Yb) are particularlyperidotite (Lassiter and DePaolo, 1997). The two primi-ing curve (Fig. 8b). They have high La/Sm (4.03 Rocks/Chondrite La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu 98-17 98-19 98-16 98-11 98-18 98-12 98-15 98-20 Rocks/Primitive mantle 98-11 98-16 98-17 98-18 98-19 98-12 98-15 98-20 Cs Rb Ba Th U Nb Ta K La Ce Pb Sr P Nd Sm Zr Hf Eu Ti Tb Y Yb Lu Sample98-1598-1298-1698-18Sr/Sr0.703966 ± 770.703510 ± 120.703127 ± 480.703242 ± 48Nd/Nd0.512802 ± 280.512838 ± 13 Table 3. Sr and Nd isotopic compositions of Tepi lavasDetails of Sr-Nd analytical procedures have been described by Ayalew et al. (2002). Fig. 5. Multi-element plots for Tepi lavas: (a) rare earth elements normalized to chondritic meteorites (Nakamura, 1974). The 54D. Ayalew 0.70300.70350.70404.64.85.05.25.4MgO (wt.%)Sr/ 0.70300.70350.7040454647484950Sr/ 4681012crustal / lithospheric valuesmantle values (25 0.70280.70320.70360.70400.704416182022 0.00.51.01.52.0 Turkana Quaternary SE Ethiopia Tepi MgO-Rich La/NbBa/Nba 012345 Jebel Mara elting of garnetperidotite (low F)erid (low Sm/YbPM b Tosa Sucha Fig. 7. Plot of

8 Ce/Pb against MgO for the Tepi lavas, gi
Ce/Pb against MgO for the Tepi lavas, givinglithospheric/crustal values. Mantle compositional fields areUMMARYThe largest issue is whether the Tepi melts are repre-sentative of an enriched mantle source or not. With pos-sibly one or two exceptions, all the Tepi samples lie withinThe Tosa Sucha volcanics lie at the extreme periphery ofthey have La/Sm ratios of 9This is consistent with their derivation from a more traceelement-enriched source than primitive mantle. The factFig. 8. Variations of Ba/Nb v. La/Nb (a) and La/Sm v. Sm/Yb(b) for the Tepi samples, compared with those of the southernEthiopia (George and Rogers, 1999), Turkana (Furman et al.,2004), western Sudan (Davidson and Wilson, 1989) and west-ern Yemen (Baker et al., 1997) lavas. Also shown fractionalFig. 6. Variations of Sr isotopes against other chemical indi-) for the Tepi lavas. Note a broad negative correlation with MgO, respectively. Quaternary alkaline Tepi shield55Tepi rocks compared with the Tosa Sucha volcanics mayheterogeneous (i.e., the lithosphere is indeed differentbetween Tepi and Tosa Sucha), (ii) remobilisation of por-events which requires large-scale processes like those ofmetasomatism in east Africa may be restricted to the pe-may be lower.The geochemical variations observed within the Tepilavas are interpreted in terms of variable open system frac-tional crystallization of a magma sourced from primitiveH. Bureau is gratefully acknowledged foristry of Foreign Affairs and the Institut National des Sciencesby the Department of Earth Sciences of Addis Ababa Univer-sity and the French Embassy in Addis Ababa. The manuscr

9 iptbenefited from comments by R. George
iptbenefited from comments by R. George and A. le Roex. T.Ayalew, D., Yirgu, G. and Pik, R. (1999) Geochemical and iso-topic (Sr, Nd and Pb) characteristics of volcanic rocks fromJ. Afr. Earth SciAyalew, D., Barbey, P., Marty, B., Reisberg, L., Yirgu, G. andGeochim. Cosmochim. ActaBaker, J., Menzies, M., Thirlwall, M. and Macpherson, C.a, Yemen: implications for plume-lithosphere inter-action and polybaric melt hybridization. J. PetrolDavidson, A. (compiler) (1983) The Omo river project, recon-naissance geology and geochemistry of parts of Ilubabor,Kefa, Gemu Gofa and Sidamo, Ethiopia. Surv. BullDavidson, A. and Rex, D. C. (1980) Age of volcanism and riftingNatureDavidson, J. P. and Wilson, I. R. (1989) Evolution of an alkaliEbinger, C. J., Yemane, T., Woldegabriel, G., Aronson, J. L.and Walter, R. C. (1993) Late Eocene-Recent volcanism andEwart, A., Milner, S. C., Armstrong, R. A. and Ducan, A. R.geochemical evidence of early Cretaceous Tristan plumeJ. PetrolFurman, T. and Graham D. (1999) Erosion of lithospheric mantlebeneath the east African rift system; geochemical evidenceFurman, T., Bryce, J. G., Karson, J. and Iotti, A. (2004) EastQuaternary mafic lavas of Turkana, Kenya. J. PetrolGeorge, R. and Rogers, N. (1999) The petrogenesis of Plio-Pleistocene alkaline volcanic rocks from the Tosa Sucharegion, Arba Minch, southern main Ethiopian rift. VulcanolGeorge, R., Rogers, N. and Kelly, S. (1998) Earliest magmatismHofmann, A. W., Jochum, K. P., Seufert, M. and White, W. M.Irvine, T. N. and Baragar, W. R. A. (1971) A guide to the chemi-Kieffer, B., Arndt, N., Lapierre, H., Bastien, F., Bosch, D.,

10 Pecher, A., Yirgu, G., Ayalew, D., Weis,
Pecher, A., Yirgu, G., Ayalew, D., Weis, D., Jerram, D. A.,Keller, F. and Meugniot, C. (2004) The Transition from Pla-teau to Shield Volcanism in Ethiopia: a petrological andgeochemical study. J. PetrolLassiter, J. C. and DePaolo, D. J. (1997) Plume/lithosphere in-Large IgneousProvinces: Continental, Oceanic and Planetary FloodVolcanism (Mahoney, J. J. and Coffin, M. F., eds.), 355, Am. Geophys. Union, Washington,Latin, D., Norry, M. J. and Tarzey, R. J. E. (1993) Magmatismin the Gregory rift, east Africa: evidence for melt genera-J. Petrolle Roex, A. P. and Lanyon, R. (1998) Isotope and trace elementJ. Petrol, 11171146.LeBas, M. J., Le Maitre, R. W., Streckeisen, A. and Zanettin,B. (1986) A chemical classification of volcanic rocks basedJ. PetrolMcDonough, W. F. and Sun, S. S. (1995) The composition ofMerla, G., Abbate, E., Azzaroli, A., Bruni, P., Canuti, P.,Fazzuoli, M., Sagri, M. and Tacconi, P. (1979) A geologicalmap of major landforms. C.N.R., Firenze, Italy.Nelson, D. R., Chivas, A. R., Chappell, B. W. and McCulloch,M. T. (1988) Geochemical and isotopic systemedics inGeochim. Cosmochim. Acta 56D. Ayalew Rogers, N. W., De Mulder, M. and Hawkesworth, C. J. (1992)from Karisimbi volcano, Virugna volcanic province,Contrib. Mineral. Petrol111WoldeGabriel, G., Yemane, T., Suwa, G., White, T. and Asfaw,B. (1991) Age of volcanism and rifting in the Burji-Soyomaarea, Amaro horst, southern main Ethiopian rift: geo- andJ. Afr. Earth SciZanettin, B., Justin-Visentin, E., Nicoletti, M. and Petrucciani,C. (1978) The evolution of the Chencha escarpment andthe Ganjuli graben (lake Abaya) in southern Ethiopi