Abstract

New Pb, Sr, Nd, Hf, and He isotope data for Quaternary basalts, erupted from Debre Zeyit, Butajira, and the Wonji Fault Belt of the Main Ethiopian Rift, show systematic mixing relationships involving three distinct mantle sources. The Pb, Sr, Nd, and Hf isotopic arrays converge in a specific region of isotopic multi-space where they define the composition of the Afar mantle plume (centered about 206Pb/204Pb = 19·5, 87Sr/86Sr = 0·7035, εNd = +4·6, εHf =  +9·3, 3He/4He > 15 RA). This plume end-member has an identical composition to that observed previously in oceanic basalts. The distinct isotopic arrays for the various volcanic areas in the Main Ethiopian Rift vary spatially in a systematic manner, and may be viewed as pseudo-binary mixing arrays. This further suggests that the Afar mantle plume interacts with the local continental lithosphere and upper mantle asthenosphere (mid-ocean ridge basalt-like source) through an ordered sequence of mixing events. Simple mixing models require that the mass proportions of continental lithosphere and upper mantle involved in magma generation must be nearly constant within each volcanic area, but that the proportion of plume material decreases regularly with distance southwestward along the Main Ethiopian Rift, away from the central axis of the plume. This systematic behavior means that continental lithosphere can become detached and mixed into the shallow mantle prior to the flow of upwelling plume material beneath the developing rift system. Detachment and mixing into the asthenosphere during continental rift evolution is an important process for producing the range of ambient upper mantle compositions sampled by mid-ocean ridge volcanism away from island hotspots.

INTRODUCTION

Mantle plumes, anomalously hot and buoyant upwellings from the deep mantle, can strongly influence the chemical composition of basalts erupted at spreading ocean ridges (Morgan, 1972; Schilling, 1973a, 1985; Hanan et al., 1986; Detrick et al., 2002) as well as those in continental settings (Morgan, 1971; Burke & Wilson, 1972; Watts, 1976; Crough, 1978; Courtney & White, 1986; Campbell & Griffiths, 1990; McNutt & Judge, 1990; Griffiths & Campbell, 1991; Farnetani & Richards, 1994; Ingle et al., 2002; Ernst & Buchan, 2003; Jones & White, 2003; Camp & Ross, 2004; Pik et al., 2006; Saunders et al., 2007; Hanan et al., 2008; Camp & Hanan, 2008; Shervais & Hanan, 2008; Graham et al., 2009). Much of the spatial variation of Pb, Nd, Sr, and He isotopes in mid-ocean ridge basalts (MORB) from the South Atlantic appears to have been produced during broad radial dispersion of isotopically distinct plume material into the asthenosphere during rifting and formation of the ocean basin, prior to when the plumes were over-ridden (in the fixed hotspot reference frame; Duncan, 1981) by the migrating Mid-Atlantic Ridge (Hanan et al., 1986; Graham et al., 1992). Thus, one explanation for the observed range of upper mantle compositions is that it reflects isotopic heterogeneity introduced by hotspots derived from distinct source reservoirs in the deep mantle (e.g. Storey et al., 1989). However, geochemical studies of MORB and continental basalts also indicate that the upper mantle has been polluted by continental lithosphere and/or lower crust (Arndt & Goldstein, 1989; Kamenetsky et al., 2001; Hanan et al., 2004; Carlson et al., 2005; Escrig et al., 2005; Class & le Roex, 2006; Goldstein et al., 2008). The spectrum of isotopic signatures observed in ocean island (hotspot) basalts, ranging between EM I, EM II and HIMU (Zindler & Hart, 1986), is also observed in basalts and xenoliths that sample the continental lithosphere and lower continental crust (e.g. McKenzie & O'Nions, 1995). Therefore, a second explanation for the origin of isotopic heterogeneity in the upper mantle is that it reflects direct pollution and mixing with detached continental lithosphere. Here we address these two different possibilities (deep vs shallow sources of mantle isotopic heterogeneity) through isotopic characterization and modelling of the sources involved in plume–lithosphere interaction along the Main Ethiopian Rift. The Main Ethiopian Rift is the type example of a modern continental rift in a plume-influenced environment (Falcon et al., 1970; Girdler, 1970; Schilling, 1973b), and provides a natural laboratory for studying the interaction of a mantle plume with the continental lithosphere.

The Red Sea, Gulf of Aden, and Main Ethiopian Rift meet in the Afar triangle, where they delineate three arms of a rift system that began at ∼30 Ma over the upwelling Afar mantle plume (Fig. 1; Hofmann et al., 1997). The Main Ethiopian Rift is the least evolved of the three triple junction arms, is about 18 Myr younger than the Red Sea and Gulf of Aden rifts, and is thus not a primary feature of the Afar plume, but is instead related to plate kinematics (Wolfenden et al., 2004). The Main Ethiopian Rift began to propagate north from the Turkana region, a zone previously rifted in the Mesozoic and again in the Cenozoic, at ∼11 Ma, coincident in time with the propagation of the Kenya Rift (Ebinger et al., 2000).

Fig. 1.

(a) Map of the Main Ethiopian Rift study area with Quaternary sample locations shown along the Wonji Fault Belt, Debre Zeyit, and Butajira extensional zones. The Debre Zeyit, Butajira, and Wonji Fault Belt zones contain widely (20–40 km) spaced felsic volcanoes and calderas. Samples are from smaller basaltic cinder cones, associated flows, and occasional fissure eruptions that are interspersed between these volcanoes. The inset shows the study area location relative to the Afar triple junction. (b) Map of the Gulf of Aden showing the sample locations of Schilling et al. (1992). Backgrounds are reflectance images of East Africa and Yemen from the NASA/JPL Shuttle Radar Tomography Mission (USGS, 2004). The topography and bathymetry are derived from Geomapapp (http://www.geomapapp.org; Ryan et al., 2009).

Fig. 1.

(a) Map of the Main Ethiopian Rift study area with Quaternary sample locations shown along the Wonji Fault Belt, Debre Zeyit, and Butajira extensional zones. The Debre Zeyit, Butajira, and Wonji Fault Belt zones contain widely (20–40 km) spaced felsic volcanoes and calderas. Samples are from smaller basaltic cinder cones, associated flows, and occasional fissure eruptions that are interspersed between these volcanoes. The inset shows the study area location relative to the Afar triple junction. (b) Map of the Gulf of Aden showing the sample locations of Schilling et al. (1992). Backgrounds are reflectance images of East Africa and Yemen from the NASA/JPL Shuttle Radar Tomography Mission (USGS, 2004). The topography and bathymetry are derived from Geomapapp (http://www.geomapapp.org; Ryan et al., 2009).

Previous studies have described the isotopic diversity of the Main Ethiopian Rift magmas in terms of mixing between plume, depleted upper mantle, and continental lithosphere (Deniel et al., 1994; Barrat et al., 1998; Trua et al., 1999; Rogers et al., 2000; Furman, 2007), but only in very broad terms. Although the timing, extent and mechanisms of mixing between these mantle reservoirs are central to the development of an integrated geodynamic model of the rifting process, they have remained poorly understood. Schilling (1973b) and Schilling et al. (1992) reported rare earth element and Nd–Sr–Pb isotopic compositions of basalts dredged from along the Sheba Ridge in the Gulf of Aden and its extension into the Gulf of Tadjoura, and of subaerial basalts from the Asal Rift in eastern Afar (Fig. 1). This dataset provided an initial framework for interpreting melting processes associated with interaction of the Afar plume with the continental lithosphere and an emerging ocean basin. The mixing relationships (Schilling et al., 1992) define three mantle source end-members interpreted as Pan-African continental lithosphere, the Afar mantle plume, and the depleted upper mantle (MORB asthenosphere source).

In this study we refine and extend the original three-component mixing model of Schilling et al. (1992) to the Butajira, Debre Zeyit, and Wonji Fault Belt regions of the Main Ethiopian Rift, three extensional zones that together form a 250 km transect along the Main Ethiopian Rift (Fig. 1, Table 1). We focus on Quaternary basalts having ages <1 Ma, and present new Pb, Nd, Sr, Hf, and He isotope results, as well as new Hf isotope data for the same Gulf of Aden sample suite previously analyzed for Pb, Nd, and Sr isotopes by Schilling et al. (1992; Table 2). Collectively, we use these new data to quantitatively characterize the distribution of mantle source reservoirs along the Main Ethiopian Rift portion of the East African Rift, their relative contribution to volcanism, and their sequence of mixing.

Table 1:

Location and isotope data for basalts from the Main Ethiopian Rift

Location Lat. Long. Distance* 206Pb/204Pb 2σ√n 207Pb/204Pb 2σ√n 208Pb/204Pb 2σ√n 87Sr/86Sr 2σ√n 143Nd/144Nd 2σ√n εNd 176Hf/177Hf 2σ√n εHf 
 (°N) (°E) (km)               
Debre Zeyit                  
DZ-1003 8·84 39·00 377 18·1518 15·5978 38·3801 19 0·704539 2† 0·512631 −0·1 0·282839 2·4 
DZ-1004 8·80 39·00 380 18·5096 10 15·5891 10 38·6038 27 0·703833 14 0·512794 3·0    
DZ-1005 8·80 39·00 380 18·3151 10 15·5782 38·4726 19 0·703905 8† 0·512763 2·4    
DZ-1007 8·72 38·99 388 18·4088 15·5723 38·5149 13 0·703909 5† 0·512784 2·8 0·282997 8·0 
DZ-1008 8·72 38·98 387 18·4949 15·5716 38·5552 21 0·703810 2† 0·512808 3·3 0·283019 8·7 
DZ-1009 8·59 39·15 386 18·5007 15·5735 38·5330 17 0·704173 7† 0·512805 3·3 0·283039 9·4 
DZ-1013 8·62 38·92 401 18·3893 15·5742 38·4825 18 0·703978 7† 0·512775 2·7    
DZ-1014 8·60 38·91 402 18·5816 15·5714 38·6030 17 0·703825 1† 0·512850 4·1 0·283027 9·0 
Butajira                  
BJ-1053 8·19 38·45 470 18·5166 15·5594 38·4494 12 0·703652 17 0·512803 3·2    
BJ-1051 8·16 38·43 475 18·4916 15·5519 38·4200 19 0·703631 10 0·512839 6† 3·9 0·283007 8·3 
BJ-1043 8·14 38·44 476 18·4300 12 15·5519 11 38·3816 28 0·703719 0·512804 3·2 0·282993 7·8 
BJ-1049 8·14 38·39 480 18·2712 15·5433 38·1914 20 0·703582 0·512790 3·0    
BJ-1044 8·10 38·38 483 18·5186 15·5534 38·4340 13 0·703647 11 0·512837 3·9 0·282997 8·0 
BJ-1045 8·06 38·36 488 18·5223 15·5602 38·4697 16 0·703606 14 0·512820 7† 3·6 0·282973 7·1 
BJ-1048 8·00 38·33 495 18·1741 10 15·5389 13 38·1730 22 0·703786 15 0·512808 3·3    
Wonji Fault Belt‡ 
WFB-1026 9·48 40·20 238 18·9194 15·5803 38·8592 13 0·703681 9† 0·512896 5·0    
WFB-1023 9·14 39·95 284 18·7524 5 15·5837 4 38·7819 11 0·704102  0·512845 4·0 0·283049 9·8 
WFB-1027 9·48 40·20 238 18·8272 3 15·5807 4 38·8958 7 0·703722  0·512890 28 4·9 0·283075 10·7 
WFB-E99-2 8·91 39·82 313 18·6022 12 15·5854 13 38·7011 25 0·704117  0·512826  3·7    
WFB-N-18 8·91 39·83 313 18·7317 6 15·5969 6 38·8345 17 0·703961  0·512827  3·7 0·283019 8·7 
WFB-1037 8·86 39·75 323 18·5571 4 15·5840 4 38·6318 0·704251  0·512796 13 3·1 0·283009 8·4 
WFB-KO-12 8·82 39·68 330 18·6577 10 15·5934 8 38·8260 23 0·704303  0·512793  3·0    
WFB-KO-4B 8·82 39·68 330 18·8011 6 15·5960 5 38·8717 17 0·704015  0·512829  3·7    
WFB-N-20 8·83 39·69 329 18·8109 6 15·5923 5 38·8730 17 0·703996        
WFB-1035 8·80 39·68 332 18·6267 4 15·5847 4 38·7497 0·704198  0·512795 16 3·1 0·283007 8·3 
WFB-N-22 8·72 39·68 340 18·9449 14 15·5948 16 38·8959 56 0·703768  0·512866  4·4 0·283056 10·0 
WFB-N-15 8·78 39·63 337 18·6799 6 15·5883 5 38·8788 15 0·704203  0·512828  3·7    
WFB-1036 8·84 39·71 327 18·8083 4 15·5896 4 38·8616 10 0·703949  0·512827 3·7 0·283008 8·4 
WFB-N-21 8·78 39·68 334 18·6241 12 15·5856 12 38·7938 36 0·704228  0·512821  3·6 0·283009 8·4 
WFB-E99-20 8·38 39·33 393 18·6940 9 15·5908 7 38·7907 20 0·704117        
WFB-N-01 8·39 39·33 392 18·5185 4 15·5892 4 38·6860 15 0·704485  0·512758  2·3 0·282976 7·2 
Mojo                  
1010 8·56 39·23 384 17·8821 15·5875 38·2677 17 0·705289 5† 0·512551 −1·7    
Location Lat. Long. Distance* 206Pb/204Pb 2σ√n 207Pb/204Pb 2σ√n 208Pb/204Pb 2σ√n 87Sr/86Sr 2σ√n 143Nd/144Nd 2σ√n εNd 176Hf/177Hf 2σ√n εHf 
 (°N) (°E) (km)               
Debre Zeyit                  
DZ-1003 8·84 39·00 377 18·1518 15·5978 38·3801 19 0·704539 2† 0·512631 −0·1 0·282839 2·4 
DZ-1004 8·80 39·00 380 18·5096 10 15·5891 10 38·6038 27 0·703833 14 0·512794 3·0    
DZ-1005 8·80 39·00 380 18·3151 10 15·5782 38·4726 19 0·703905 8† 0·512763 2·4    
DZ-1007 8·72 38·99 388 18·4088 15·5723 38·5149 13 0·703909 5† 0·512784 2·8 0·282997 8·0 
DZ-1008 8·72 38·98 387 18·4949 15·5716 38·5552 21 0·703810 2† 0·512808 3·3 0·283019 8·7 
DZ-1009 8·59 39·15 386 18·5007 15·5735 38·5330 17 0·704173 7† 0·512805 3·3 0·283039 9·4 
DZ-1013 8·62 38·92 401 18·3893 15·5742 38·4825 18 0·703978 7† 0·512775 2·7    
DZ-1014 8·60 38·91 402 18·5816 15·5714 38·6030 17 0·703825 1† 0·512850 4·1 0·283027 9·0 
Butajira                  
BJ-1053 8·19 38·45 470 18·5166 15·5594 38·4494 12 0·703652 17 0·512803 3·2    
BJ-1051 8·16 38·43 475 18·4916 15·5519 38·4200 19 0·703631 10 0·512839 6† 3·9 0·283007 8·3 
BJ-1043 8·14 38·44 476 18·4300 12 15·5519 11 38·3816 28 0·703719 0·512804 3·2 0·282993 7·8 
BJ-1049 8·14 38·39 480 18·2712 15·5433 38·1914 20 0·703582 0·512790 3·0    
BJ-1044 8·10 38·38 483 18·5186 15·5534 38·4340 13 0·703647 11 0·512837 3·9 0·282997 8·0 
BJ-1045 8·06 38·36 488 18·5223 15·5602 38·4697 16 0·703606 14 0·512820 7† 3·6 0·282973 7·1 
BJ-1048 8·00 38·33 495 18·1741 10 15·5389 13 38·1730 22 0·703786 15 0·512808 3·3    
Wonji Fault Belt‡ 
WFB-1026 9·48 40·20 238 18·9194 15·5803 38·8592 13 0·703681 9† 0·512896 5·0    
WFB-1023 9·14 39·95 284 18·7524 5 15·5837 4 38·7819 11 0·704102  0·512845 4·0 0·283049 9·8 
WFB-1027 9·48 40·20 238 18·8272 3 15·5807 4 38·8958 7 0·703722  0·512890 28 4·9 0·283075 10·7 
WFB-E99-2 8·91 39·82 313 18·6022 12 15·5854 13 38·7011 25 0·704117  0·512826  3·7    
WFB-N-18 8·91 39·83 313 18·7317 6 15·5969 6 38·8345 17 0·703961  0·512827  3·7 0·283019 8·7 
WFB-1037 8·86 39·75 323 18·5571 4 15·5840 4 38·6318 0·704251  0·512796 13 3·1 0·283009 8·4 
WFB-KO-12 8·82 39·68 330 18·6577 10 15·5934 8 38·8260 23 0·704303  0·512793  3·0    
WFB-KO-4B 8·82 39·68 330 18·8011 6 15·5960 5 38·8717 17 0·704015  0·512829  3·7    
WFB-N-20 8·83 39·69 329 18·8109 6 15·5923 5 38·8730 17 0·703996        
WFB-1035 8·80 39·68 332 18·6267 4 15·5847 4 38·7497 0·704198  0·512795 16 3·1 0·283007 8·3 
WFB-N-22 8·72 39·68 340 18·9449 14 15·5948 16 38·8959 56 0·703768  0·512866  4·4 0·283056 10·0 
WFB-N-15 8·78 39·63 337 18·6799 6 15·5883 5 38·8788 15 0·704203  0·512828  3·7    
WFB-1036 8·84 39·71 327 18·8083 4 15·5896 4 38·8616 10 0·703949  0·512827 3·7 0·283008 8·4 
WFB-N-21 8·78 39·68 334 18·6241 12 15·5856 12 38·7938 36 0·704228  0·512821  3·6 0·283009 8·4 
WFB-E99-20 8·38 39·33 393 18·6940 9 15·5908 7 38·7907 20 0·704117        
WFB-N-01 8·39 39·33 392 18·5185 4 15·5892 4 38·6860 15 0·704485  0·512758  2·3 0·282976 7·2 
Mojo                  
1010 8·56 39·23 384 17·8821 15·5875 38·2677 17 0·705289 5† 0·512551 −1·7    

*Radial distance from Lake Abhe, Afar (11·25°N, 41·42°E).

Prior to dissolution rock chips were cleaned with distilled 2 N HCl, 1 N HBr, and ultrapure H2O. Sr, Nd, and Pb were separated for analysis following the procedures of Hanan & Schilling (1989) and Blichert-Toft et al. (1997) for Hf. Sr isotopes were measured on a VG Sector 54, seven-collector thermal ionization mass spectrometer in multi-dynamic mode. Hf, Nd, and Pb were analyzed on a Nu Plasma HR multi-collector inductively coupled plasma mass spectrometry (MC-ICP-MS) system using a Nu Instruments desolvating nebulizer equipped with a 100 µl min−1 nebulizer (DSN100). The 87Sr/86Sr isotope ratios were normalized to 88Sr/86Sr = 0·1194 to correct for mass fractionation, and the Sr isotopic data are reported relative to NIST SRM 987 = 0·710250. The Nd and Hf isotope ratios were corrected for instrumental mass fractionation and machine bias by applying a discrimination factor determined by bracketing sample runs with analyses of the SDSU AMES Nd standard using a value of 0·512130 and the JMC 475 standard (every two or three samples) with 176Hf/177Hf of 0·282160. Hf isotope sample data are normalized to 179Hf/177Hf = 0·7325 and reported relative to JMC 475 = 0·282160. The CHUR value for 176Hf/177Hf of 0·282772 of Blichert-Toft & Albarède (1997) was used to calculate the εHf values. Sample data were normalized to 146Nd/144Nd = 0·7219 and are reported relative to the laboratory value of 143Nd/144Nd = 0·512130 for the SDSU AMES Nd standard. The measured value of La Jolla Nd at SDSU = 0·511844. Pb was separated on two anion exchange columns, a 500 µl followed by a 200 µl column to maximize the Pb separation for optimum ICP-MS analyses. The lead yields were determined to be ∼100%, therefore the columns do not have a fractionation effect. The in-run mass discrimination for Pb was determined with NIST SRM 997 Tl with 205Tl/203Tl = 2·3889 and NIST SRM 981 (Todt, 1996) using the method of White et al. (2000). The Pb isotope ratios were corrected for machine bias and drift by bracketing the sample analyses with analyses of the NIST SRM 981 (every two samples) corrected for fractionation. Total procedural blanks were: <25 pg Hf, <90 pg Pb, <200 pg Nd, <250 pg Sr. No blank corrections were applied to the data because they are insignificant. The reported uncertainties are 2 standard errors of the mean (2σ√n) for the in-run precision, except for some Nd and Sr samples (†), where the uncertainties represent 2σ√n for 2–3 replicate analyses, and thus provide a measure of sample reproducibility. The NIST SRM 987 standard averaged 87Sr/86Sr = 0·710257 ± 8 2σ√n, n = 16, AMES Nd standard averaged 143Nd/144Nd = 0·512118 ± 2 2σ√n, n = 22, JMC475 standard averaged 0·282151 ± 0·000001 2σ√n, n = 21, and NIST SRM 981 averaged 206Pb/204Pb = 16·9361 ± 0·0004, 207Pb/204Pb = 15·4914 ± 0·0004, 208Pb/204Pb = 38·6981 ± 0·0012 2σ√n, n = 17.

‡The analyses in italics for the Wonji Fault Belt were published previously (Furman et al., 2006). The Pb isotopes for the Furman et al. (2006) Wonji Fault Belt data were analyzed at SDSU under the same conditions as reported here. We have added the fourth decimal place and errors (from the SDSU data archives) to be consistent with the new Main Ethiopian Rift data. The analytical methods and procedures for Nd and Sr have been reported by Furman et al. (2006). The Nd isotope ratios for these samples have been renormalized to the SDSU La Jolla = 0·511844 and the Sr isotope ratios are normalized to NIST SRM 987 = 0·710250.

Table 2:

Location and Hf isotope data for the Gulf of Aden basalts

Sample Latitude (°N) Longitude (°E) Distance* (km) 176Hf/ 177Hf 2σ√n εHf 
Asal Rift       
DJ-74 11·60 42·48 122 0·283122 12·4 
DJ-42 11·60 42·49 122 0·283161 13·8 
DJ-26 11·60 42·50 123 0·283130 12·7 
DJ-18 11·60 42·51 124 0·283115 12·1 
DJ-9 11·60 42·51 125 0·283141 13·0 
DJ-5 11·60 42·52 126 0·283144 13·2 
DJ-2 11·60 42·52 126 0·283141 13·0 
Gulf of Tadjoura       
VD-11 11·51 42·85 158 0·283070 10·5 
V-60 g 11·75 43·02 183 0·283077 10·8 
VD-39 11·59 43·10 187 0·283122 10 12·4 
V-55 11·87 43·27 213 0·283083 11·0 
V3307 41D-1 11·94 43·41 229 0·283098 11·5 
Tadjoura Trough       
V3307 42D-1 g 11·94 43·69 259 0·283134 12·8 
V3307 44D-2 g 12·08 43·93 289 0·283263 17·4 
V3307 46D-1 g 12·06 44·71 370 0·283230 16·2 
West Sheba Ridge       
V3307 47D-1 g 12·07 44·83 382 0·283186 14·6 
V3307 48D-2 g 12·04 45·21 422 0·283138 12·9 
V3307 49D-1 12·04 45·51 454 0·283008 8·3 
V3307 50D-1 g 12·10 45·85 492 0·283105 11·8 
V3307 51D-1 g 12·16 46·29 539 0·283126 12·5 
V3307 52D-1 g 12·27 46·71 587 0·283274 17·8 
V3307 53D-1 12·42 47·10 632 0·283304 18·8 
V3307 54D-1 g 12·42 47·25 648 0·283225 16·0 
V3307 55D-2 g 12·47 47·84 712 0·283270 17·6 
V3307 56D-1 g 12·65 48·27 762 0·283262 17·3 
V3306 39D-1 g 12·75 48·49 788 0·283246 16·8 
V3307 57D-1 g 12·90 48·99 844 0·283284 18·1 
V3307 58D-1 12·90 49·36 883 0·283309 12 19·0 
V3307 59D-1 g 13·04 49·39 890 0·283253 17·0 
V3307 60D-1 13·16 49·69 925 0·283302 18·7 
V3307 62D-1 g 13·36 49·81 942 0·283315 19·2 
V3307 61D-2 g 13·26 49·97 958 0·283277 17·9 
V3307 63D-3 g 13·29 50·32 995 0·283295 18·5 
V3307 64D-3 g 13·34 50·69 1036 0·283339 20·1 
V3307 65D-1 g 13·21 51·02 1068 0·283349 20·4 
East Sheba Ridge       
V3306 35D-1 14·65 52·41 1254 0·283293 18·4 
V3307 66D-1 g 14·39 52·68 1275 0·283163 13·8 
V3307 67D-1 g 14·37 53·58 1368 0·283257 17·2 
V3306 25D-1 14·48 53·91 1406 0·283134 12·8 
V3306 24D-1 g 14·75 54·80 1508 0·283187 14·7 
V3307 68D-1 g 14·74 55·71 1603 0·283189 14·7 
V3307 69D-1 g 14·47 56·42 1672 0·283201 15·2 
Sample Latitude (°N) Longitude (°E) Distance* (km) 176Hf/ 177Hf 2σ√n εHf 
Asal Rift       
DJ-74 11·60 42·48 122 0·283122 12·4 
DJ-42 11·60 42·49 122 0·283161 13·8 
DJ-26 11·60 42·50 123 0·283130 12·7 
DJ-18 11·60 42·51 124 0·283115 12·1 
DJ-9 11·60 42·51 125 0·283141 13·0 
DJ-5 11·60 42·52 126 0·283144 13·2 
DJ-2 11·60 42·52 126 0·283141 13·0 
Gulf of Tadjoura       
VD-11 11·51 42·85 158 0·283070 10·5 
V-60 g 11·75 43·02 183 0·283077 10·8 
VD-39 11·59 43·10 187 0·283122 10 12·4 
V-55 11·87 43·27 213 0·283083 11·0 
V3307 41D-1 11·94 43·41 229 0·283098 11·5 
Tadjoura Trough       
V3307 42D-1 g 11·94 43·69 259 0·283134 12·8 
V3307 44D-2 g 12·08 43·93 289 0·283263 17·4 
V3307 46D-1 g 12·06 44·71 370 0·283230 16·2 
West Sheba Ridge       
V3307 47D-1 g 12·07 44·83 382 0·283186 14·6 
V3307 48D-2 g 12·04 45·21 422 0·283138 12·9 
V3307 49D-1 12·04 45·51 454 0·283008 8·3 
V3307 50D-1 g 12·10 45·85 492 0·283105 11·8 
V3307 51D-1 g 12·16 46·29 539 0·283126 12·5 
V3307 52D-1 g 12·27 46·71 587 0·283274 17·8 
V3307 53D-1 12·42 47·10 632 0·283304 18·8 
V3307 54D-1 g 12·42 47·25 648 0·283225 16·0 
V3307 55D-2 g 12·47 47·84 712 0·283270 17·6 
V3307 56D-1 g 12·65 48·27 762 0·283262 17·3 
V3306 39D-1 g 12·75 48·49 788 0·283246 16·8 
V3307 57D-1 g 12·90 48·99 844 0·283284 18·1 
V3307 58D-1 12·90 49·36 883 0·283309 12 19·0 
V3307 59D-1 g 13·04 49·39 890 0·283253 17·0 
V3307 60D-1 13·16 49·69 925 0·283302 18·7 
V3307 62D-1 g 13·36 49·81 942 0·283315 19·2 
V3307 61D-2 g 13·26 49·97 958 0·283277 17·9 
V3307 63D-3 g 13·29 50·32 995 0·283295 18·5 
V3307 64D-3 g 13·34 50·69 1036 0·283339 20·1 
V3307 65D-1 g 13·21 51·02 1068 0·283349 20·4 
East Sheba Ridge       
V3306 35D-1 14·65 52·41 1254 0·283293 18·4 
V3307 66D-1 g 14·39 52·68 1275 0·283163 13·8 
V3307 67D-1 g 14·37 53·58 1368 0·283257 17·2 
V3306 25D-1 14·48 53·91 1406 0·283134 12·8 
V3306 24D-1 g 14·75 54·80 1508 0·283187 14·7 
V3307 68D-1 g 14·74 55·71 1603 0·283189 14·7 
V3307 69D-1 g 14·47 56·42 1672 0·283201 15·2 

*Radial distance from Lake Abhe, Afar.

Hf isotope data reported for the Gulf of Aden sample suite were analyzed by MC-ICP-MS (Plasma 54) at the Ecole Normale Supérieure in Lyon following the procedures of Blichert-Toft et al. (1997). The JMC-475 Hf standard, which was run every two samples, averaged 0·282160 ±  0·000010 during the analytical run sessions of the Gulf of Aden samples. The total procedural Hf blank was less than 20 pg. The CHUR value for 176Hf/177Hf of 0·282772 of Blichert-Toft and Albarède (1997) was used to calculate the εHf values. The Pb, Sr, and Nd isotope compositions have been reported by Schilling et al. (1992).

BACKGROUND

Sample selection

A subset of the most primitive Main Ethiopian Rift samples (6–10 wt % MgO) previously analyzed for major and trace elements (Rooney et al., 2005, 2007, 2011; Furman et al., 2006; Rooney, 2010) were selected for Hf, Sr, Nd, Pb (n = 15), and He (n = 12) isotopic analysis. The major element and trace element data for two samples, WFB-1026 and WFB-N-14, that were analyzed but not reported by Furman et al. (2006), are reported in Appendix Table A1. Hf isotope data are also reported for a suite of basalts dredged from the Sheba Ridge axis in the Gulf of Aden and its extension into the Gulf of Tadjoura, and for subaerial basalts from the Asal Rift of eastern Afar. The Pb, Sr, and Nd isotope data for these samples were reported by Schilling et al. (1992), and are the basis for the three-component mixing model proposed therein. The analytical details for the Main Ethiopian Rift and Gulf of Aden isotope data are given in the legends of Tables 1–3. The complete dataset is provided as Supplementary Data (available for downloading at http://www.petrology.oxfordjournals.org/).

Table 3:

Location and He isotope data for basalts from the Main Ethiopian Rift

Sample Lat. Long. Distance Phase Wt 3He/4He 2σ√n [He] 
 (°N) (°N) (km)  (mg) R/RA   
Debre Zeyit         
DZ-1014 8·60 38·91 402 cpx 598·1 5·61 0·34 1·66E – 09 
DZ-1009 8·59 39·15 386 olivine 442·1 8·30 0·40 2·59E – 09 
DZ-1007 8·72 38·99 388 olivine 744·6 8·66 0·14 1·12E – 08 
DZ-1008 8·72 38·98 387 olivine 581·5 8·80 0·16 1·06E – 08 
Butajira         
BJ-1045 8·06 38·36 488 olivine 422·4 7·52 0·30 4·22E – 09 
BJ-1045 8·06 38·36 488 cpx 420·0 7·53 0·16 1·14E – 08 
BJ-1053 8·19 38·45 470 cpx 541·4 7·70 0·24 4·91E – 09 
BJ-1043 8·14 38·44 476 olivine 430·4 8·11 0·36 2·55E – 09 
Wonji Fault Belt         
WFB-1027 9·48 40·20 238 olivine 428·6 12·66 0·98 7·49E – 10 
WFB-1036 8·84 39·71 327 olivine 631·4 12·16 0·24 8·02E – 09 
WFB-N-21 8·78 39·68 334 olivine 537·8 12·82 0·26 6·40E – 09 
WFB-N-14 8·76 39·61 340 olivine 871·4 15·12 0·30 3·13E – 09 
Sample Lat. Long. Distance Phase Wt 3He/4He 2σ√n [He] 
 (°N) (°N) (km)  (mg) R/RA   
Debre Zeyit         
DZ-1014 8·60 38·91 402 cpx 598·1 5·61 0·34 1·66E – 09 
DZ-1009 8·59 39·15 386 olivine 442·1 8·30 0·40 2·59E – 09 
DZ-1007 8·72 38·99 388 olivine 744·6 8·66 0·14 1·12E – 08 
DZ-1008 8·72 38·98 387 olivine 581·5 8·80 0·16 1·06E – 08 
Butajira         
BJ-1045 8·06 38·36 488 olivine 422·4 7·52 0·30 4·22E – 09 
BJ-1045 8·06 38·36 488 cpx 420·0 7·53 0·16 1·14E – 08 
BJ-1053 8·19 38·45 470 cpx 541·4 7·70 0·24 4·91E – 09 
BJ-1043 8·14 38·44 476 olivine 430·4 8·11 0·36 2·55E – 09 
Wonji Fault Belt         
WFB-1027 9·48 40·20 238 olivine 428·6 12·66 0·98 7·49E – 10 
WFB-1036 8·84 39·71 327 olivine 631·4 12·16 0·24 8·02E – 09 
WFB-N-21 8·78 39·68 334 olivine 537·8 12·82 0·26 6·40E – 09 
WFB-N-14 8·76 39·61 340 olivine 871·4 15·12 0·30 3·13E – 09 

Helium isotope analyses were carried out by in vacuo crushing of mineral separates from the Ethiopian basalts. Analytical procedures have been described elsewhere (Graham et al., 1998). The reported uncertainties are 2 standard errors, computed from the quadrature sum of uncertainties associated with the sample analysis, air standard analyses and blank determinations. Distance is measured radially in kilometers.

Crustal contamination indicators (e.g. La/Nb and Ce/Pb) for the samples selected for isotopic analysis and modeling show no evidence for significant crustal assimilation (Rooney et al., 2005, 2007) and the samples fall largely within the range observed globally for mantle-derived basalts (Hofmann et al., 1986, 2005). We also examined the relationship between K/P and Ti/Yb and the Nd and Sr isotope ratios for the Main Ethiopian Rift lavas. Concurrent increases in K/P with increasing 87Sr/86Sr (or decreasing 143Nd/144Nd or 176Hf/177Hf) are indicative of silicic crustal assimilation (Hart et al., 1989). We observed no consistent relationships between Sr, Nd, Hf, or Pb isotope ratios and K/P. The Main Ethiopian Rift lavas have Ti/Yb ratios of about 6000–10 000 (except for sample 1003, which has Ti/Yb = 4500 and contains mafic crustal xenoliths; Rooney et al., 2005). Low K/P <6·5 accompanied by high Ti/Yb (>6000) for the Main Ethiopian Rift lavas rule out significant crustal contamination (Leeman & Hawkesworth, 1986; Van Calsteren et al., 1986; Hart et al., 1989). Furthermore, the He isotope data presented below show that the signature of the Afar plume, as well as the upper mantle, dominates the He isotope variations in the Main Ethiopian Rift basalts selected for this study, which indicates that any potential crustal contamination is minor. The lack of crustal assimilation in these magma suites is consistent with focused recent volcanism in the Wonji Fault Belt and Silti-Debre Zeyit Fault Zone (e.g. WoldeGabriel et al., 1990; Ebinger & Casey, 2001; Casey et al., 2006). Magmas have re-utilized earlier conduits, incorporating xenocrysts and cumulate xenoliths from the walls, with apparently little interaction with the continental lithosphere on their way to the surface (Rooney et al., 2005).

Isotope signatures, mantle components, and mixing

The isotopes of Pb, Sr, Nd, Hf, and He are powerful tracers of mantle dynamics, because mantle-derived basalts fall into distinct subdivisions based on their radiogenic isotope compositions (White & Hofmann, 1982; Zindler & Hart, 1986; Hofmann, 1997). These subdivisions are related to distinct processes (pelagic vs terrigenous sediment subduction, hydrothermal alteration, etc.) as well as to distinct tectonic or petrogenetic environments (e.g. mid-ocean ridge, oceanic plateau, island arc, etc.). As one example, DUPAL signatures (Hart, 1984) may arise from the recycling of oceanic and/or continental lithosphere, ranging from ancient (EM I) to recent (EM II) in age, may be associated with, respectively, ancient and recent sediments, and may include metasomatized lithosphere (e.g. HIMU) (Zindler & Hart, 1986; Mahoney et al., 1989, 1992, 1996; McKenzie & O'Nions, 1995; Rehkämper & Hofmann, 1997; Lustrino, 2005; Willbold & Stracke, 2010). The source of MORB is best constrained in terms of its location in the mantle, and corresponds to the trace-element depleted upper mantle MORB source (White & Hofmann, 1982; Zindler & Hart, 1986). The common isotopic component C, defined by the convergence of trends in Pb isotope compositional space (at relatively radiogenic Pb; 206Pb/204Pb = 19·6 ± 0·4) for MORB from the Atlantic, Pacific and Indian Oceans, has been attributed to the presence of recycled oceanic lithosphere in the mantle source region of many MORB (Hanan & Graham, 1996). In contrast, the mantle component FOZO, defined by the convergence of global ocean island basalt isotopic trends, has been attributed to the presence of deep, relatively primitive mantle (Hart et al., 1992; Hauri et al., 1994), although this interpretation has since been modified (Stracke et al., 2005). Mixing models involving these mantle components and end-members, however, often cannot define a unique source for the basalts. In particular, for continental rift basalts it is difficult to distinguish a deep mantle component derived by melting recycled lithosphere from shallower melting involving old subcontinental lithosphere. Although this is a challenging problem, the systematics of mixing during petrogenesis can be deciphered using a multi-isotopic approach. For example, the order of mixing between three end-members may be revealed through the combined use of 208Pb–207Pb–206Pb isotopes (Hanan et al., 1986; Hanan & Schilling, 1997; Douglass & Schilling, 2000). Systematic spatial and temporal variations in this isotope space provide further constraints on the mixing sequence. Mixing models typically use exactly chosen values for end-member compositions to avoid the mathematical complexity that arises from the mixing of heterogeneous end-members. However, mantle sources are, by their very nature, heterogeneous to varying degrees, and, hence, simple mixing models assuming homogeneous end-members should be viewed only as approximations; they cannot be expected to account for all the natural variability observed in most sample suites. The approach that we adopt below also involves a multi-component mixing model with fixed end-member compositions. Nevertheless, we impose rigorous mathematical and statistical tests to evaluate the ability of the model to account for the observed geochemical variability in the Main Ethiopian Rift sample suite. The details and mathematical development of this multi-component mixing model are well documented (e.g. Hanan et al., 1986, 2000; Schilling et al., 1992; Hanan & Schilling, 1997; Douglass & Schilling, 2000), and its algorithms and rationale have been given by Schilling et al. (1992), Albarède (1995) and Douglass & Schilling (2000) and hence are not repeated here.

RESULTS

Pb–Sr–Nd–Hf–He spatial isotope variations

The Pb–Sr–Nd–Hf–He isotope variations along the axis of the Main Ethiopian Rift, relative to the presumed location of the center of the Afar plume beneath the Lake Abhe region, are shown in Fig. 2a and b. The location of the plume center near Lake Abhe is based on a synthesis of independent structural, tectonic, magnetic, and geochemical data that all focus toward this region (Schilling, 1973b). The new Pb, Sr, Nd, and Hf isotope ratios for the Main Ethiopian Rift lavas all become more radiogenic upon approaching the inferred plume center, except for 87Sr/86Sr and 207Pb/204Pb in two Wonji Fault Belt samples closest to the plume, which appear to be slightly low (Table 1; Fig. 2). In addition, 143Nd/144Nd and 176Hf/177Hf data for Butajira basalts are not quite as low as predicted by a monotonic linear trend (Fig. 2b). There is a general trend of increasing 3He/4He towards Lake Abhe (Table 3). However, there is also significant variation on a shorter length scale; for example, the highest 3He/4He in our sample suite (15·1 RA) occurs in a Wonji Fault Belt lava at ∼340 km distance (sample WFB N-14), similar to what is observed for 206Pb/204Pb (Fig. 2a). The bulk composition of sample N-14, which contains abundant plagioclase and olivine phenocrysts relative to the other samples, is unlikely to represent a magmatic liquid because it falls off the major element trends of the combined dataset (e.g. CaO/Al2O3 and MgO).

Fig. 2.

(a) Variation of 206Pb/204Pb (r2 = 0·48), 207Pb/204Pb (r2 = 0·60), 208Pb/204Pb (r2 = 0·83), and 3He/4He R/RA (r2 = 0·53), and (b) 87Sr/86Sr, εNd, and εHf in Quaternary lavas of the Main Ethiopian Rift (Wonji Fault Belt, Debre Zeyit, and Butajira) as a function of radial distance from the inferred present-day location of the Afar plume axis in the Lake Abhe region (Table 1). (c) Variation of 87Sr/86Sr (r2 = 0·18), εHf (r2 = 0·14), εNd (r2 = 0·10), and 206Pb/204Pb (r2 = 0·036) as a function of radial distance from the Afar plume center for the Asal Rift, Gulf of Tadjoura, Tadjoura Trough and the Sheba Ridge. The prominent shift in εHf across the Alula-Fartak fracture zone as the East Sheba Ridge steps northward towards Yemen should be noted. The trend lines [in (a) and (c); dashed black lines] represent the least-squares linear regression for the Main Ethiopian Rift [from (a) and (b)]. The ‘by eye’ trends for the Gulf of Aden data are shown by a thin dashed line in (c). The Gulf of Aden Hf isotope data are from Table 2 and the Pb, Sr, and Nd isotope data are from Schilling et al. (1992). Also shown are the Main Ethiopian Rift data from (a) and (b) and Table 1. Negative distances signify samples from along the Gulf of Aden, positive distances indicate samples along the East African Rift. E- and W-Torus refer to the position of the torus plume head beneath the Main Ethiopian Rift and Gulf of Aden as proposed by Schilling et al. (1992). Debre Zeyit sample DZ-1003 is labeled.

Fig. 2.

(a) Variation of 206Pb/204Pb (r2 = 0·48), 207Pb/204Pb (r2 = 0·60), 208Pb/204Pb (r2 = 0·83), and 3He/4He R/RA (r2 = 0·53), and (b) 87Sr/86Sr, εNd, and εHf in Quaternary lavas of the Main Ethiopian Rift (Wonji Fault Belt, Debre Zeyit, and Butajira) as a function of radial distance from the inferred present-day location of the Afar plume axis in the Lake Abhe region (Table 1). (c) Variation of 87Sr/86Sr (r2 = 0·18), εHf (r2 = 0·14), εNd (r2 = 0·10), and 206Pb/204Pb (r2 = 0·036) as a function of radial distance from the Afar plume center for the Asal Rift, Gulf of Tadjoura, Tadjoura Trough and the Sheba Ridge. The prominent shift in εHf across the Alula-Fartak fracture zone as the East Sheba Ridge steps northward towards Yemen should be noted. The trend lines [in (a) and (c); dashed black lines] represent the least-squares linear regression for the Main Ethiopian Rift [from (a) and (b)]. The ‘by eye’ trends for the Gulf of Aden data are shown by a thin dashed line in (c). The Gulf of Aden Hf isotope data are from Table 2 and the Pb, Sr, and Nd isotope data are from Schilling et al. (1992). Also shown are the Main Ethiopian Rift data from (a) and (b) and Table 1. Negative distances signify samples from along the Gulf of Aden, positive distances indicate samples along the East African Rift. E- and W-Torus refer to the position of the torus plume head beneath the Main Ethiopian Rift and Gulf of Aden as proposed by Schilling et al. (1992). Debre Zeyit sample DZ-1003 is labeled.

The similar profiles for Pb, Sr, Nd, Hf, and He isotopes are not expected, because mantle-derived lavas usually show broad anti-correlation between 143Nd/144Nd (or 176Hf/177Hf) and 87Sr/86Sr (or 206Pb/204Pb) reflecting long-term in situ radiogenic production. The signatures observed here, therefore, rather reflect recent mixing.

The variations in Nd–Sr–Pb–Hf isotope compositions and La/Sm ratios along the Gulf of Aden spreading center define an 800 km long gradient eastward from the Afar region, based on samples from the Asal Rift, Gulf of Tadjoura, Tadjoura Trough, and along the Sheba Ridge to about 50°E (Schilling et al., 1992; Table 2). The Asal Rift and Gulf of Tadjoura represent a nascent stage of rift evolution, characterized by diffuse extension. In contrast, the more developed Sheba Ridge is approaching steady-state seafloor spreading. The differences in the rift development stage are reflected in the Pb, Sr, Nd, and Hf isotope characteristics of the basalts. The West Sheba Ridge shows an enrichment peak for Pb, Sr, Nd, and Hf isotopes and large ion lithophile elements (LILE) centered around 46°E, about 500 km east of the presumed plume center (Table 2). East of 48°E, Pb, Sr, Nd, and Hf isotope compositions show much less variation. The 206Pb/204Pb, 87Sr/86Sr, εNd, and εHf display similar spatial variation (Fig. 2c) consisting of well-defined alternating peaks and troughs. Although the peaks and troughs appear to be slightly out of phase, there is no apparent fracture zone control on the Pb, Sr, and Nd isotope compositions (Fig. 2; Schilling et al., 1992). However, the new Hf isotope data show a step at the Alula-Fartak fracture zone toward less radiogenic, more continental-like signatures, as the ridge jumps northeastward toward Yemen. The Asal Rift and Gulf of Tadjoura show a much smaller enrichment spike in Pb, Nd, and Hf isotope ratios relative to the Sheba Ridge. However, 206Pb/204Pb is more radiogenic, εNd and εHf are less radiogenic than observed in the LILE-depleted basalts of the East Sheba Ridge and all overlap the range of the West Sheba Ridge samples. The range in 87Sr/86Sr for the Asal Rift and Gulf of Tadjoura data is distinctly higher than observed along the entire Sheba Ridge. The anomaly at 46°E was previously attributed to the interaction of the Sheba Ridge with a torus-like mantle plume (Schilling et al., 1992). In the torus plume model, the Gulf of Tadjoura and Asal Rift overlie a mantle region located to the inside of a plume-head torus, where the mantle source is proposed to be dominated by Pan-African continental lithosphere and the upper mantle asthenosphere (Fig. 2c; Schilling et al., 1992). The plume torus overlies the enrichment peaks in radiogenic Pb and Sr and corresponding troughs in εNd and εHf located on the Sheba Ridge at about 500 km from the plume center. In the Schilling et al. (1992) model, the opposite side of the torus would be located at a similar distance from the plume center, along the southwestern Main Ethiopian Rift.

Pb–Pb isotope correlations

The new data for the Main Ethiopian Rift were collected from three spatially resolvable sections of the Main Ethiopian Rift: Debre Zeyit, Butajira, and the Wonji Fault Belt. The Debre Zeyit and Butajira volcanic fields are two major zones of volcanism and tectonism separated by a 50 km gap that comprises the northernmost portion of the Silti-Debre Zeyit Fault Zone (WoldeGabriel et al., 1990; Rooney et al., 2011). The most distinctive Quaternary magmatic feature in the Ethiopian Rift is the Wonji Fault Belt, which extends from around Lake Abhe into southern Ethiopia (Mohr, 1967). The three sample populations (Debre Zeyit, Butajira, and Wonji Fault Belt) have different, but overlapping ranges in Pb isotope ratios, and appear to form distinct arrays in Pb–Sr–Nd–Hf multi-isotope space (Figs 3a, b and 4; Table 4). This raises the question of whether the three spatially distinct datasets can be treated as a single population, or whether they constitute three distinct sub-populations.

Fig. 3.

(a) Covariation of 208Pb/204Pb and 207Pb/204Pb with 206Pb/204Pb, and 207Pb/206Pb with 208Pb/206Pb in the Main Ethiopian Rift Quaternary and Gulf of Aden sample suites. The dashed lines are least-squares regression lines for the Gulf of Aden, Wonji Fault Belt, Debre Zeyit, and Butajira volcanic zones. (b) Covariation of 143Nd/144Nd and 87Sr/86Sr with 206Pb/204Pb, and 143Nd/144Nd with 87Sr/86Sr in the Wonji Fault Belt, Debre Zeyit, and Butajira volcanic zones and Gulf of Aden sample suites. The dashed lines represent the least-squares linear regression lines of the datasets for the Main Ethiopian Rift and Gulf of Aden data. Debre Zeyit samples Mojo 1010 and DZ-1003 are labeled.

Fig. 3.

(a) Covariation of 208Pb/204Pb and 207Pb/204Pb with 206Pb/204Pb, and 207Pb/206Pb with 208Pb/206Pb in the Main Ethiopian Rift Quaternary and Gulf of Aden sample suites. The dashed lines are least-squares regression lines for the Gulf of Aden, Wonji Fault Belt, Debre Zeyit, and Butajira volcanic zones. (b) Covariation of 143Nd/144Nd and 87Sr/86Sr with 206Pb/204Pb, and 143Nd/144Nd with 87Sr/86Sr in the Wonji Fault Belt, Debre Zeyit, and Butajira volcanic zones and Gulf of Aden sample suites. The dashed lines represent the least-squares linear regression lines of the datasets for the Main Ethiopian Rift and Gulf of Aden data. Debre Zeyit samples Mojo 1010 and DZ-1003 are labeled.

Fig. 4.

(a) Triangular Pb isotope plot (after Hanan et al., 1986) of the relative abundances of 206Pb, 207Pb, and 208Pb in Quaternary lavas from Afar and the northern Main Ethiopian Rift (Deniel et al., 1994). Only lavas having MgO > 7 wt % are included, to limit the potential variation related to shallow crustal processes. The least-squares trend for the Red Sea and Gulf of Aden are shown for comparison. Data for the Red Sea are from Altherr et al. (1988), Rogers (1993), and Volker et al. (1993, 1997), and for the Gulf of Aden from Schilling et al. (1992). (b) Triangular Pb isotope plot (after Hanan et al., 1986) of the relative abundances of 206Pb, 207Pb, and 208Pb in lavas from the Main Ethiopian Rift and the model end-members depleted mantle, Pan-African lithosphere, and the Afar plume. Basalts from Butajira, Debre Zeyit, and the Wonji Fault Belt define three distinct pseudo-binary mixing arrays that converge at the composition of the Afar plume, which lies within the range of compositions for the common component ‘C’ deduced from oceanic basalts (Hanan & Graham, 1996). The depleted ends of the mixing arrays are characterized by three ratios of depleted mantle to Pan-African lithosphere, each of which is constant over the range of mixing seen in the three arrays. The proportion of the Afar plume is largest for the Wonji Fault Belt basalts (∼240–440 km from the Afar plume center), smallest for the Butajira basalts (∼470–500 km distant), and intermediate for the Debre Zeyit basalts (∼370–400 km distant). Sample Mojo-1010 is labeled. The ternary grid is drawn at 0·25% intervals. The ternary grid interval is 0.25%.

Fig. 4.

(a) Triangular Pb isotope plot (after Hanan et al., 1986) of the relative abundances of 206Pb, 207Pb, and 208Pb in Quaternary lavas from Afar and the northern Main Ethiopian Rift (Deniel et al., 1994). Only lavas having MgO > 7 wt % are included, to limit the potential variation related to shallow crustal processes. The least-squares trend for the Red Sea and Gulf of Aden are shown for comparison. Data for the Red Sea are from Altherr et al. (1988), Rogers (1993), and Volker et al. (1993, 1997), and for the Gulf of Aden from Schilling et al. (1992). (b) Triangular Pb isotope plot (after Hanan et al., 1986) of the relative abundances of 206Pb, 207Pb, and 208Pb in lavas from the Main Ethiopian Rift and the model end-members depleted mantle, Pan-African lithosphere, and the Afar plume. Basalts from Butajira, Debre Zeyit, and the Wonji Fault Belt define three distinct pseudo-binary mixing arrays that converge at the composition of the Afar plume, which lies within the range of compositions for the common component ‘C’ deduced from oceanic basalts (Hanan & Graham, 1996). The depleted ends of the mixing arrays are characterized by three ratios of depleted mantle to Pan-African lithosphere, each of which is constant over the range of mixing seen in the three arrays. The proportion of the Afar plume is largest for the Wonji Fault Belt basalts (∼240–440 km from the Afar plume center), smallest for the Butajira basalts (∼470–500 km distant), and intermediate for the Debre Zeyit basalts (∼370–400 km distant). Sample Mojo-1010 is labeled. The ternary grid is drawn at 0·25% intervals. The ternary grid interval is 0.25%.

Table 4:

Characteristics of slopes fitted to the linear arrays in Pb isotope space for Main Ethiopian Rift magmas

Pb isotopes Region Test Value for Value for other 
   single belt two belts pooled 
206–208 WFB Slope 0·54 0·69 
  Intercept 28·75 25·67 
 DZ Slope 0·52 1·12 
  Intercept 28·87 17·90 
 BJ Slope 0·88 0·78 
  Intercept 22·23 24·25 
207–206 WFB Slope 0·01 0·00 
  Intercept 15·49 15·55 
 DZ Slope –0·05 0·08 
  Intercept 16·44 14·12 
 BJ Slope 0·05 0·01 
  Intercept 14·58 15·34 
Pb isotopes Region Test Value for Value for other 
   single belt two belts pooled 
206–208 WFB Slope 0·54 0·69 
  Intercept 28·75 25·67 
 DZ Slope 0·52 1·12 
  Intercept 28·87 17·90 
 BJ Slope 0·88 0·78 
  Intercept 22·23 24·25 
207–206 WFB Slope 0·01 0·00 
  Intercept 15·49 15·55 
 DZ Slope –0·05 0·08 
  Intercept 16·44 14·12 
 BJ Slope 0·05 0·01 
  Intercept 14·58 15·34 

Differences between the slopes and intercepts can be utilized to examine differences and similarities between the groups (BJ, Butajira; DZ, Debre Zeyit; WFB, Wonji Fault Belt) using statistical methods (F-test; see Table 5).

Least-squares linear regressions of the three datasets show that their trends are significantly different at the 95% confidence level for both the 206Pb/204Pb vs 208Pb/204Pb and 207Pb/204Pb binary projections (Table 4). The three sub-populations (Wonji Fault Belt, Debre Zeyit, and Butajira) define linear arrays in Pb isotopic space, with r2 values of 0·68, 0·93, and 0·97, respectively, in the 208Pb/204Pb vs 206Pb/204Pb diagram (Fig. 3a). In the 207Pb/204Pb vs 206Pb/204Pb diagram the trends for Debre Zeyit and Butajira have r2 values of 0·43 and 0·89, respectively. The Wonji Fault Belt data form a near horizontal array with r2 = 0·02. The r2 statistic measures how significantly the slope of a fitted line differs from zero, and is not necessarily a measure of good fit. To further test the validity of subdividing the data on the basis of spatial location, we performed analysis of the covariance of the Pb–Pb least-squares regression lines for each of the sub-populations against the combined datasets of the other two areas. The rationale for this comparison is that if all the data fall along the same regression, then removing a subset of the data should not affect the pooled regression. We tested the null hypothesis that the least-squares regression for a sub-population was the same as the regression for the other two groups combined. The F-test results for 208Pb/204Pb and 207Pb/204Pb vs 206Pb/204Pb, using the slopes and intercepts of the least-squares linear regressions for the three sub-populations (Wonji Fault Belt, Debre Zeyit, and Butajira) relative to the data subsets containing the other two groups, allow rejection of the null hypothesis at the 95% confidence level (Table 5). Therefore, these three sample suites, from spatially different volcanic belts, have distinct Pb isotope correlations, or trends, and define three statistically distinct sub-populations that make up the complete Main Ethiopian Rift sample suite of this study.

Table 5:

F-test analysis of the similarities of the WFB, Butajira and Debre Zeyit in Pb isotope space

Pb isotopes Region Test F-value Degrees of freedom F-crit. (95% confidence) Is the group different from pooled? 
206–208 WFB Slope 0·58 27 4·2 No 
  Elevation 19·33 28 4·2 Yes 
 DZ Slope 5·94 27 4·2 Yes 
  Elevation 2·43 28 4·2 No 
 BJ Slope 0·37 27 4·2 No 
  Elevation 60·2 28 4·2 Yes 
207–206 WFB Slope 0·01 27 4·2 No 
  Elevation 9·79 28 4·2 Yes 
 DZ Slope 15·68 27 4·2 Yes 
  Elevation 5·44 28 4·2 Yes 
 BJ Slope 2·9 27 4·2 No 
  Elevation 67·48 28 4·2 Yes 
Pb isotopes Region Test F-value Degrees of freedom F-crit. (95% confidence) Is the group different from pooled? 
206–208 WFB Slope 0·58 27 4·2 No 
  Elevation 19·33 28 4·2 Yes 
 DZ Slope 5·94 27 4·2 Yes 
  Elevation 2·43 28 4·2 No 
 BJ Slope 0·37 27 4·2 No 
  Elevation 60·2 28 4·2 Yes 
207–206 WFB Slope 0·01 27 4·2 No 
  Elevation 9·79 28 4·2 Yes 
 DZ Slope 15·68 27 4·2 Yes 
  Elevation 5·44 28 4·2 Yes 
 BJ Slope 2·9 27 4·2 No 
  Elevation 67·48 28 4·2 Yes 

The primary question posed was whether each of the three recognized geographical groups (BJ, Butajira; DZ, Debre Zeyit; WFB, Wonji Fault Belt) were distinctive on the basis of their isotopic results. As discussed in the text, Pb isotopes form distinctive linear arrays and hence can be used to examine differences and similarities between the groups. To establish whether each of the groups is distinctive from the dataset as a whole, we examined the slope and elevation of the arrays generated by 206Pb/204Pb–208Pb/204Pb and 206Pb/204Pb–207Pb/204Pb. The F-test specifically asked: Are the Wonji Fault Belt, Debre Zeyit and Butajira samples distinctive from the remaining dataset in terms of slope and elevation? F-values were calculated using established methods (Snedecor & Cochran, 1989). F-crit. is the critical F-value (Davis, 1973) based on the degrees of freedom of the dataset; where the F-value is above the F-crit. value, the null hypothesis (that the slope and/or intercept of the respective datasets are identical) can be rejected at the 95% confidence level.

Nd–Hf isotope correlations

In εNd vs εHf isotope space (Fig. 5a) the Main Ethiopian Rift data lie along and above the mantle array, trending between the Afar plume, as represented by the C field, and compositions typical of ancient lithosphere. In contrast, the new data from the Sheba Ridge axis in the Gulf of Aden and its extension into the Gulf of Tadjoura and the Asal Rift plot at more radiogenic values of εHf and εNd, similar to Indian Ocean MORB with predominantly positive values of ΔεHf, rather than negative ΔεHf values such as those of most Pacific MORB (Fig. 5; ΔεHf is the vertical displacement from the mantle array in εHf units for a data point in the εNd vs εHf diagram). Both the Main Ethiopian Rift and the Gulf of Tadjoura–Asal Rift data have distinct high Δ208Pb/204Pb (where Δ208Pb/204Pb is relative to the 208Pb/204Pb of the Sheba Ridge array at any given 206Pb/204Pb; Fig. 3a). These observations are consistent with the different tectonic settings. The Sheba Ridge data reflect an origin dominated by plume–oceanic lithosphere interaction (e.g. Schilling et al., 1992), whereas the Main Ethiopian Rift and the Gulf of Tadjoura–Asal Rift data indicate strong influence from a continental lithosphere component, in line with a continental rift setting.

Fig. 5.

(a) εHf vs εNd and (b) 206Pb/204Pb vs εHf for the Quaternary Main Ethiopian Rift sample suite and for the Gulf of Aden basalts [Nd and Pb isotopic data from Schilling et al. (1992)]. Model end-members (see Table 6) are shown as black open stars for the Afar plume and depleted mantle. Pan-African lithosphere plots off the diagram at this scale. The dashed-line field represents the range of values for C (Hanan & Graham, 1996; Hanan et al., 2000). The fields for HIMU, depleted Indian Ocean MORB from the Australian–Antarctic Discordance region of the Southeast Indian Ridge, and depleted Pacific MORB from the East Pacific Rise Garrett Fracture Zone are also shown for reference (data from Chaffey et al., 1989; Salters & White, 1998; Wendt et al., 1999; Hanan et al., 2004). The mantle array in (a) (dashed line) is represented by εHf = 1· 4εNd + 2·8 (J. Blichert-Toft, unpublished compilation of >3000 oceanic basalts). Symbols are the same as in previous plots. (c) Δ208Pb/204Pb vs ΔεHf for the most depleted MORB from the Pacific and Indian Ocean basins, represented by the Garrett Fracture Zone (Wendt et al., 1999) and the Australian–Antarctic Discordance on the Southeast Indian Ridge (after Hanan et al., 2004). Pacific MORB have Δ208Pb/204Pb < 20, whereas Indian Ocean MORB have Δ208Pb/204Pb > 20. In terms of ΔεHf, the Pacific data are confined to the SW quadrant of the diagram with ΔεHf < 1. Depleted Indian Ocean MORB have positive ΔεHf, and plot in the NE quadrant with ΔεHf > 1. The distinction between the Pacific and Indian Ocean MORB derives from pollution of the mantle source of the latter by continental material (Hanan et al., 2004). The data for the Main Ethiopian Rift and the Gulf of Aden show Hf and Pb isotope signatures similar to those of Indian Ocean MORB derived from mantle sources polluted with continental material. Debre Zeyit sample DZ-1003 is labeled.

Fig. 5.

(a) εHf vs εNd and (b) 206Pb/204Pb vs εHf for the Quaternary Main Ethiopian Rift sample suite and for the Gulf of Aden basalts [Nd and Pb isotopic data from Schilling et al. (1992)]. Model end-members (see Table 6) are shown as black open stars for the Afar plume and depleted mantle. Pan-African lithosphere plots off the diagram at this scale. The dashed-line field represents the range of values for C (Hanan & Graham, 1996; Hanan et al., 2000). The fields for HIMU, depleted Indian Ocean MORB from the Australian–Antarctic Discordance region of the Southeast Indian Ridge, and depleted Pacific MORB from the East Pacific Rise Garrett Fracture Zone are also shown for reference (data from Chaffey et al., 1989; Salters & White, 1998; Wendt et al., 1999; Hanan et al., 2004). The mantle array in (a) (dashed line) is represented by εHf = 1· 4εNd + 2·8 (J. Blichert-Toft, unpublished compilation of >3000 oceanic basalts). Symbols are the same as in previous plots. (c) Δ208Pb/204Pb vs ΔεHf for the most depleted MORB from the Pacific and Indian Ocean basins, represented by the Garrett Fracture Zone (Wendt et al., 1999) and the Australian–Antarctic Discordance on the Southeast Indian Ridge (after Hanan et al., 2004). Pacific MORB have Δ208Pb/204Pb < 20, whereas Indian Ocean MORB have Δ208Pb/204Pb > 20. In terms of ΔεHf, the Pacific data are confined to the SW quadrant of the diagram with ΔεHf < 1. Depleted Indian Ocean MORB have positive ΔεHf, and plot in the NE quadrant with ΔεHf > 1. The distinction between the Pacific and Indian Ocean MORB derives from pollution of the mantle source of the latter by continental material (Hanan et al., 2004). The data for the Main Ethiopian Rift and the Gulf of Aden show Hf and Pb isotope signatures similar to those of Indian Ocean MORB derived from mantle sources polluted with continental material. Debre Zeyit sample DZ-1003 is labeled.

Hf–Pb isotope correlations

In εHf vs 206Pb/204Pb space the West Sheba Ridge array extends between the fields for the most depleted MORB from the Southwest Indian Ridge and the C-like Afar plume. In contrast, the Main Ethiopian Rift data extend between the Pan-African lithosphere and the Afar plume end-member (Fig. 5b). The Asal Rift and Tadjoura data have affinity to, and overlap with, the East Sheba Ridge data in a region of the diagram between the Main Ethiopian Rift and the West Sheba Ridge arrays, displaced toward the Pan-African lithosphere component. Overall, the combined dataset, consisting of the Main Ethiopian Rift and the Gulf of Aden samples, has Indian Ocean-like Hf and Pb isotope signatures, distinct from the Pacific asthenosphere MORB source (Fig. 5c) and consistent with a continental-like mantle source component.

DISCUSSION

Afar plume–continental lithosphere–depleted mantle mixing model

The Arabian–Nubian shield represents Precambrian microplate accretion of inter-continental arcs, slivers of back-arc oceanic lithosphere, and older Precambrian terranes reactivated during Pan-African thermotectonic events (Schilling et al., 1992; Be'eri-Shlevin et al., 2010). The Pan-African sub-continental lithosphere is of variable mean age, origin, and isotopic composition; there are very few published xenolith isotope data from the Arabian–Nubian shield. Radiogenic isotope analyses of mantle xenoliths from beneath Arabia (Yemen and Kod Ali Island in the Red Sea; Baker et al., 2002) indicate that the lithosphere has been extensively modified by interaction with the plume and the asthenosphere, and in fact could be modelled by these same three end-members. The rejuvenated Pan-African lithosphere is not conducive to preserving ancient high 3He/4He. Lithosphere with high time-integrated 3He/(U + Th) that would allow preservation of ancient 3He enrichments cannot exist within or below the continents, and is unlikely to exist within the upper mantle, based on 3He/4He in MORB (Day et al., 2005). The source origin of the Main Ethiopian Rift basalts cannot be solely within the sub-continental lithosphere. In the 208Pb/204Pb and 207Pb/204Pb vs 206Pb/204Pb isotope diagrams the new combined data do not fall along a single linear correlation, ruling out a simple binary source mixing origin for the lavas. At least three end-member components are required to account for the entire dataset. The three-component mixing model invoked by Schilling et al. (1992) for the mantle source of the lavas from the Gulf of Aden, its extension into the Gulf of Tadjoura, and subaerial basalts from the Ardoukoba Rift in east Afar calls upon interaction between a mantle plume, EM-like Pan-African continental lithosphere and the MORB-source asthenosphere, to explain the Nd–Sr–Pb isotope compositions. In general, these three end-members enclose the entire range of the Nd–Sr–Pb isotope ratios of the new Main Ethiopian Rift data presented here. Schilling et al. (1992) associated the Afar plume with a HIMU radiogenic Pb isotope signature. However, the Butajira, Debre Zeyit, and Wonji Fault Belt trends converge at a lower 206Pb/204Pb (∼19·6) than the characteristic HIMU value (206Pb/204Pb >20; Zindler & Hart, 1986), consistent with an Afar plume composition similar to ‘C’ (Fig. 4), the ‘common’ isotopic composition observed in oceanic basalts (Hanan & Graham, 1996). This observation is further supported by the εNd vs εHf diagram, in which the Wonji Fault Belt, Debre Zeyit, and Butajira data trends all fall above and parallel to the mantle array, and do not show mixing relationships toward the HIMU composition, which plots below the mantle array (Fig. 5). In terms of three-component mixing, in addition to the relatively high 206Pb/204Pb plume end-member, the Main Ethiopian Rift data distribution requires the other two end-members to have lower 206Pb/204Pb, with one having relatively higher 207Pb/204Pb than the other. These constraints are consistent with the Pan-African continental lithosphere (high 207Pb/204Pb) and the asthenosphere (MORB-source mantle with low 207Pb/204Pb) end-members of the Schilling et al. (1992) model for the Gulf of Aden, Gulf of Tadjoura, and Afar basalts. Here we retain the isotopic composition of the Pan-African lithosphere and depleted mantle end-members proposed by Schilling et al. (1992). We further extend and modify the three-component mixing model between the Afar plume, depleted upper mantle, and Pan-African lithosphere proposed by Schilling et al. (1992) and apply the model to the new Main Ethiopian Rift data (Tables 1–6).

Table 6:

Isotope composition and elemental abundances for the model end-members

 Sr Nd Pb 87Sr/86Sr 143Nd/144Nd 206Pb/204Pb 207Pb/204Pb 208Pb/204Pb 
 (ppm) (ppm) (ppm)      
Afar plume 160 14 29 0·7035 0·512875 19·5 15·6 39·2 
DM 75 0·7022 0·51335 17·5 15·3 36·6 
Pan-African 225 22 39 0·7075 0·5121 17·85 15·75 39·75 
 Sr Nd Pb 87Sr/86Sr 143Nd/144Nd 206Pb/204Pb 207Pb/204Pb 208Pb/204Pb 
 (ppm) (ppm) (ppm)      
Afar plume 160 14 29 0·7035 0·512875 19·5 15·6 39·2 
DM 75 0·7022 0·51335 17·5 15·3 36·6 
Pan-African 225 22 39 0·7075 0·5121 17·85 15·75 39·75 

These values are identical to those reported by Schilling et al. (1992) with the exception of the Afar plume end-member. Based on the present study, the Afar plume is now considered to be of a C mantle composition (Hanan & Graham, 1996) and the average isotopic values of C are used here. DM, depleted mantle; Pan African, Pan-African lithosphere. [See the Appendix of this paper and Schilling et al. (1992, appendix) for more model details.]

In summary, the three end-member isotope compositions for the Main Ethiopian Rift ternary mixing model are dictated by the following constraints (see also Schilling et al., 1992 for additional detail): (1) they must span the entire Main Ethiopian Rift dataset in Pb–Nd–Hf–Sr multi-isotope space (i.e. define a triangular space that surrounds the Main Ethiopian Rift dataset); (2) the convergence of the three pseudo-binary mixing trends for the Debre Zeyit, Butajira, and Wonji Fault Belt sub-populations constrains the composition of the radiogenic Pb end-member (the Afar plume); (3) the Afar plume has the isotopic composition of C (Hanan & Graham, 1996; Hanan et al., 2000); (4) the least radiogenic of the two low 206Pb/204Pb end-members is assumed to represent the depleted asthenosphere sampled along normal mid-ocean ridge segments; (5) the low 206Pb/204Pb with a relatively high 207Pb/204Pb lithospheric-like end-member represents Pan-African lithosphere and is EM-like in its isotopic composition. We retain the Nd, Sr, and Pb isotopic compositions and elemental concentrations of the Pan-African lithosphere and depleted asthenosphere components used by Schilling et al. (1992). The Main Ethiopian Rift mixing model differs from the previous model in the choice of the Afar plume isotope composition, as outlined above. The Hf isotope data (Tables 1 and 2; Figs 4 and 5a) for the Gulf of Aden and the Main Ethiopian Rift strongly support the suggestion that the Afar plume composition is C-like rather than of HIMU composition.

The choice of end-members and the validity of the three-component mixing model deduced from the Pb isotope data were tested using principal component analysis (Fig. 6; for an explanation of principal component analysis applied to basalts see, for example, Schilling et al., 1992; Hanan & Schilling, 1997; Blichert-Toft et al., 2003, 2005; Agranier et al., 2005; Debaille et al., 2006). Considering only the Pb isotopes, 99·5% of the population variance is accounted for within the plane of the first two eigenvectors. Principal component analysis using 206Pb/204Pb–207Pb/204Pb–208Pb/204Pb–87Sr/86Sr–143Nd/144Nd yields an equivalent result, with the first two eigenvectors accounting for 98·1% of the variance. Hf isotopes were not included given the smaller dataset and the principal component analysis requirement that each sample has the same number of variables. Furthermore, we have not assigned Hf isotope values to the end-members because (1) there are few or no Hf isotope data available to characterize the continental lithosphere in this region, and (2) the sparse Hf isotope data for the Main Ethiopian Rift and East African Rift in general do not currently show enough spread to determine accurate mixing trajectories. The principal component analysis results suggest that hyperbolic mixing of the geochemical end-members is minimal; that is, the elemental ratios of Pb, Sr, and Nd, in the hybrid end-member (comprising asthenospheric upper mantle and a small proportion of continental lithosphere) are approximately equal to those in the Afar plume end-member, resulting in linear mixing trends in multi-isotope ratio plots (e.g. Hanan & Schilling, 1997; Douglass & Schilling, 2000). The linear arrays in multi-isotope space define converging pseudo-binary mixing trends (Figs 3, 4 and 6). The converging trends for the Wonji Fault Belt, Debre Zeyit, and Butajira lavas apparently result from a special case of ternary mixing (case 3 described by Douglass & Schilling, 2000) in which the mass proportions of two of the end-members (depleted upper mantle and continental lithosphere) co-vary systematically along the rift, maintaining an essentially fixed ratio within each extensional zone. This type of mixing (in which the ratio of asthenosphere to continental lithosphere mass fraction is nearly constant, whereas the ratio of plume to asthenosphere markedly varies) is difficult to attribute to a single-stage mixing event, because in that case all of the components should be equally well mixed (Douglass & Schilling, 2000). An ordered, two-stage mixing process is required, during which the asthenosphere and continental lithosphere interact to form a hybrid, before mixing with the Afar plume.

Fig. 6.

PC1–PC2 projection of a principal component analysis for 206Pb/204Pb, 207Pb/204Pb, 208Pb/204Pb, εNd, and εSr using the Main Ethiopian Rift data and the end-members defined in Table 6. The lines are least-squares regressions through the three sub-populations of data from Debre Zeyit, Butajira, and the Wonji Fault Belt. The labeled point (DZ-1003) was included in the principal component analysis, but omitted from the Debre Zeyit regression because it contains norite xenoliths and may have been affected by crustal assimilation associated with volcanism at a nearby silicic volcanic center (Rooney et al., 2005). The eigenvectors PC1 = 70·9% and PC2 = 27·2% account for >98·1% of the total variance. The three-component mixing model is justified because two principal components (and thus, three geochemical end-members) sufficiently account for the Sr–Nd–Pb isotopic variability in the Main Ethiopian Rift lavas, and the three end-members lie within a single plane of mixing. The approximate field of the C-like Afar plume is shown as a dashed-line circle.

Fig. 6.

PC1–PC2 projection of a principal component analysis for 206Pb/204Pb, 207Pb/204Pb, 208Pb/204Pb, εNd, and εSr using the Main Ethiopian Rift data and the end-members defined in Table 6. The lines are least-squares regressions through the three sub-populations of data from Debre Zeyit, Butajira, and the Wonji Fault Belt. The labeled point (DZ-1003) was included in the principal component analysis, but omitted from the Debre Zeyit regression because it contains norite xenoliths and may have been affected by crustal assimilation associated with volcanism at a nearby silicic volcanic center (Rooney et al., 2005). The eigenvectors PC1 = 70·9% and PC2 = 27·2% account for >98·1% of the total variance. The three-component mixing model is justified because two principal components (and thus, three geochemical end-members) sufficiently account for the Sr–Nd–Pb isotopic variability in the Main Ethiopian Rift lavas, and the three end-members lie within a single plane of mixing. The approximate field of the C-like Afar plume is shown as a dashed-line circle.

We carried out the three-component Pb, Sr, and Nd isotope mixing model using basic principles described previously (Schilling et al., 1992; Hanan & Schilling, 1997; Douglass & Schilling, 2000; Hanan et al., 2000) (Fig. 7). The modeling indicates that the mixtures are dominated by the asthenosphere (depleted mantle component) (mass fraction = 49–80%), and that the mass fraction of the Afar plume (C-like Afar plume component) (10–43%) is larger than that of the continental lithosphere (Pan-African lithosphere component) (6–17%). The mass fractions of the plume and depleted mantle components co-vary (r2 = 0·92), and the contribution from the Afar plume decreases systematically away from the hypothesized plume center beneath Lake Abhe (Schilling et al., 1992) (Fig. 7). The 3He/4He ratio (Table 3) broadly co-varies positively with the mass proportion of the Afar plume in the three-component mixture, providing further evidence that the converging pseudo-binary mixing arrays in isotope–isotope space can be used to constrain the Pb–Sr–Nd isotopic composition of the Afar plume component.

Fig. 7.

(a) Correlation diagrams for mixing model solutions. With this model, a complete solution providing the percentage of plume, depleted mantle, and continental lithosphere for a single sample can only be determined using two isotope ratios at a time. Thus each ‘model’ uses a pair of isotopes from each sample to yield percentages of each component (PA, Pan-African; AP, Afar plume; DM, depleted mantle). We have compared the results of using one pair of isotopes with an alternate pair to test the model's self-consistency. Model A = 207Pb/204Pb–206Pb/204Pb, Model B = 208Pb/204Pb–206Pb/204Pb, Model C = 143Nd/144Nd–206Pb/204Pb, Model D = 87Sr/86Sr–206Pb/204Pb. The model (A–D) end-member proportions for each Main Ethiopian Rift sample from Table 1 are presented in Appendix Table A2. The slopes vary from 0·87 to 1·07 and r2 is consistently 0·99, indicating that the mixing model is robust and adequately accounts for the isotope variability of the Main Ethiopian Rift samples (Wonji Fault Belt, Debre Zeyit, Butajira). Details of the three-component mixing model are provided in the Appendix. For any sample, the percentage values of the three end-members that are given in the main text are an average of the values derived from the four models (A–D) presented above. (b) Spatial variation of the mass proportions of Afar plume, depleted mantle and Pan-African lithospheric components deduced from our modeling along the East African Rift. The distance in kilometers (Table 1) represents the great circle radial distance from the plume center in the Lake Abhe region (Schilling et al., 1992). There is a systematic decrease in plume mass proportion and increase in depleted mantle proportion away from the plume center.

Fig. 7.

(a) Correlation diagrams for mixing model solutions. With this model, a complete solution providing the percentage of plume, depleted mantle, and continental lithosphere for a single sample can only be determined using two isotope ratios at a time. Thus each ‘model’ uses a pair of isotopes from each sample to yield percentages of each component (PA, Pan-African; AP, Afar plume; DM, depleted mantle). We have compared the results of using one pair of isotopes with an alternate pair to test the model's self-consistency. Model A = 207Pb/204Pb–206Pb/204Pb, Model B = 208Pb/204Pb–206Pb/204Pb, Model C = 143Nd/144Nd–206Pb/204Pb, Model D = 87Sr/86Sr–206Pb/204Pb. The model (A–D) end-member proportions for each Main Ethiopian Rift sample from Table 1 are presented in Appendix Table A2. The slopes vary from 0·87 to 1·07 and r2 is consistently 0·99, indicating that the mixing model is robust and adequately accounts for the isotope variability of the Main Ethiopian Rift samples (Wonji Fault Belt, Debre Zeyit, Butajira). Details of the three-component mixing model are provided in the Appendix. For any sample, the percentage values of the three end-members that are given in the main text are an average of the values derived from the four models (A–D) presented above. (b) Spatial variation of the mass proportions of Afar plume, depleted mantle and Pan-African lithospheric components deduced from our modeling along the East African Rift. The distance in kilometers (Table 1) represents the great circle radial distance from the plume center in the Lake Abhe region (Schilling et al., 1992). There is a systematic decrease in plume mass proportion and increase in depleted mantle proportion away from the plume center.

Quaternary basalts (MgO >7 wt %) from the Djibouti and Red Sea region [e.g. the Quaternary Moussa Ali volcano and the Maranda-Inkir volcanic centers located NE of the Afar plume center, and from the Sept Frères volcanic islands in the Red Sea (Deniel et al., 1994)] provide additional support for the model of radial dispersion from a plume centered beneath the Lake Abhe region (Fig. 4). The Djibouti and Red Sea data fall along extensions of the converging Pb isotope trends for the Main Ethiopian Rift lavas, within the region defined by the three-component mixing model end-members.

Plume–lithosphere interaction in the East African Rift system

Previous studies have shown that the extent of upper mantle pollution in the vicinity of an upwelling and dispersing plume decreases systematically with radial distance from the plume center (Schilling, 1985; Hanan et al., 1986; Hanan & Schilling, 1989). Afar plume influence has been observed to extend from Afar for ∼700 km along the Gulf of Aden, into the Red Sea, and along the Ethiopian sector of the East African Rift (Barrat et al., 1990; Schilling et al., 1992; Marty et al., 1993, 1996; Deniel et al., 1994; Scarsi & Craig, 1996; Moreira et al., 1999; Hopp et al., 2004). The Afar plume head is envisioned to have flattened beneath the lithosphere like a pancake, and now has a radial influence of the order of 700–1400 km (Schilling et al., 1992; Volker et al., 1997).

Based on the laboratory models of Griffiths & Campbell (1990), Schilling et al. (1992) proposed that the flattened Afar plume head has a torus-like structure, containing an outer ring of deeply derived material located about 450 km (± 150 km) from the center of the upwelling, and an inner region of entrained, depleted mantle. In this model the plume tail appears to be tilted toward Afar (Schilling et al., 1992). Along the Sheba Ridge an anomaly in radiogenic Pb, Sr, Nd, and Hf isotopes is located above the proposed torus ring. East of the anomaly (48°E) the MORB are LILE-depleted and show a slight increase in radiogenic Pb and Sr isotope compositions (Schilling et al., 1992). To the west of the anomaly, lavas from the Gulf of Tadjoura and Asal Rift are more similar to the Main Ethiopian Rift lavas compared with the West Sheba Ridge MORB, and have Pb, Sr, Nd, and Hf isotope compositions that indicate a greater involvement of continental lithosphere in their genesis. The western part of the proposed torus would underlie the Main Ethiopian Rift in southernmost Ethiopia, where the plume tail is supposedly tilted toward the west. Geophysical studies have not imaged a distinct tail structure, but instead show a broad low-velocity structure tilted westward (Benoit et al., 2006). According to the torus model, our new Main Ethiopian Rift sample locations would correspond to a traverse from inside the plume torus towards and across the edge of the proposed location of the western ring. The observed gradient of decreasing radiogenic Pb, Sr, and Nd away from the plume center and toward the proposed ring location is opposite to what would be predicted by the torus model. Furthermore, melting in the center of the torus plume was suggested to be dominated by lithospheric mantle and depleted asthenosphere entrained by the plume. However, based on the regular spatial variations in the Pb–Sr–Nd–Hf–He isotope ratios along the Main Ethiopian Rift, and the results of our mixing model, it is apparent that all three mantle source components, Pan-African lithosphere, depleted mantle, and the Afar plume, contribute to the petrogenesis of the Main Ethiopian Rift basalts located in the region of the proposed inner ring. The Asal rift and the Gulf of Tadjoura lavas have Pb isotope ratios consistent with radial dispersion of the plume and their distance of about 150–300 km from the plume center. We observe a gradient in plume influence (Figs 2 and 7) along the Main Ethiopian Rift that is expressed by a systematic co-variation between plume and depleted asthenosphere mass fractions (Fig. 7). The near-constant ratio of African continental lithosphere to depleted upper mantle for each of the extensional zones, revealed by the pseudo-binary arrays for the Wonji Fault Belt, Debre Zeyit, and Butajira subgroups, requires two stages of mixing (Douglass & Schilling, 2000). The first stage involves mixing of asthenospheric mantle and continental lithosphere, followed by a second stage during which this hybrid upper mantle mixes with the Afar plume. Detachment and foundering of lithosphere during rifting is one mechanism for contaminating the upper mantle with continental material (Arndt & Goldstein, 1989). This process may occur when the lithosphere becomes ductile, as a result of thermal or flexural weakening during extension (Burov et al., 2007). The hot Afar plume head beneath the rift also contributes to thermal weakening by conductive heating. As the lithosphere is heated and thinned during plume impingement and incipient rifting (e.g. Kent et al., 1992), portions of the uppermost mantle beneath the lithospheric mechanical boundary layer (McKenzie & Bickle, 1988) are admixed with lithospheric material. This upper mantle pollution with continental lithospheric material occurs prior to when the uppermost mantle becomes a sink for the buoyant and dispersing plume (Sleep, 2008). This process is influenced by the underside topography of the continental lithosphere, which may channel plume material into ‘thin spots’ without involving plume-driven thinning (Thompson & Gibson, 1991; Ebinger & Sleep, 1988; Sleep et al., 2002; Shervais & Hanan, 2008). For example, the isotopic anomaly on the West Sheba Ridge adjacent to the transition between continental and oceanic lithosphere may relate to a step decrease in lithospheric thickness that acts as a sink for plume material (Fig. 2c). The two-stage ordered mixing along the Main Ethiopian Rift suggests that the sub-continental lithosphere acted as a levee against plume flow along the rift until the tectonic forces and geometry of extension associated with the triple junction creation led to formation of the Main Ethiopian Rift arm.

Implications for the composition of the upper mantle

The sequence of mixing described above has implications for the origin of upper mantle heterogeneity, as well as for the composition and origin of mantle plumes. Asthenospheric mantle that was once associated with continental rifting may retain evidence of continental lithosphere contamination (McKenzie & O'Nions, 1983; Hawkesworth et al., 1986; Arndt & Goldstein, 1989; Hanan & Graham, 1996; Hanan et al., 2004). Pollution of the upper mantle by continental lithosphere (and/or lower crust) has been observed in the Tertiary basalts of Iceland (Hanan & Schilling, 1997; Hanan et al., 2000), along the Mid-Atlantic Ridge north of Iceland (Andres et al., 2004; Blichert-Toft et al., 2005) and along the Gakkel Ridge in the Arctic (Goldstein et al., 2008), in the South Atlantic (Hawkesworth et al., 1986; Carlson et al., 1996; Kamenetsky et al., 2001; Andres et al., 2002) and along the Southwest (Escrig et al., 2005; Janney et al., 2005; Meyzen et al., 2007) and Southeast Indian Ridges (Hanan et al., 2004; Graham et al., 2006). Our observations and model for the Ethiopian sector of the East African Rift indicate that mantle plume material interacts with continental lithosphere and upper mantle only after thermo-mechanical thinning of the lithosphere has progressed sufficiently to form a sink for the rising and dispersing plume material. Lithospheric detachment and mixing into the asthenosphere during continental rift evolution may, therefore, account for much of the range of ambient upper mantle compositions sampled by mid-ocean ridge volcanism away from island hotspots. A corollary is that the isotopic composition of plume-related lavas, such as continental flood basalts and some oceanic islands, may not always represent the isotopic composition of the deep mantle plume source alone, but instead carry an integrated record of plume interaction and mixing with upper mantle that has been variably polluted with continental material such as the Yellowstone Plume (Hanan et al., 2008) and eastern Atlantic cases (Geldmacher et al., 2011).

CONCLUSIONS

The isotopic variations of basalts erupted along the Main Ethiopian Rift at Debre Zeyit, Butajira, and the Wonji Fault Belt result from mixing between the Afar plume, the asthenospheric upper mantle and the African continental lithosphere. The isotopic composition of the Afar plume resembles the ‘C’ component on the basis of converging arrays in multi-isotope space. The influence of this plume signature systematically decreases along the Main Ethiopian Rift away from Afar into southern Ethiopia. Mixing between the plume, lithosphere, and ambient upper mantle follows an ordered sequence of events. Within each volcanic belt the Afar plume mixes with a hybrid reservoir that is composed of different proportions of lithosphere and depleted mantle. Within each volcanic belt, the mass proportions of continental lithosphere and upper mantle are relatively constant, requiring that the mixing event that generated these hybrid reservoirs occurred prior to their mixing with the Afar plume.

Detachment and dispersal of continental lithosphere into the upper mantle during plume–rift interaction follows an ordered sequence of mixing events. This mixing produces large-scale pollution of the upper mantle by continental lithosphere that is recorded in the ‘background’ geochemical signatures of MORB, even those erupting far away from mantle hotspots or plumes. Furthermore, the distinct isotopic compositions of plume-related lavas may not necessarily represent intrinsic mantle source signatures, but rather a complex history of mixing involving components derived from both the upper and lower mantle.

FUNDING

This work was supported by National Science Foundation grants to B. Hanan, D. Graham, and T. Furman, and a George H. Deike Jr grant to T. Furman. J. Blichert-Toft received financial support from the French Institut National des Sciences de l'Univers.

SUPPLEMENTARY DATA

Supplementary data for this paper are available at Journal of Petrology online.

ACKNOWLEDGEMENTS

We thank Kassahun Ejeta, Roeland Doust, Gezahegn Yirgu, and Dereje Ayalew for field assistance and discussion in Ethiopia, Margaret Nitz and Leigh Patterson for help with sample preparation, and Joan Willis and Jasper Konter for their assistance in the clean room. We thank John Lupton for access to the helium isotope laboratory in Newport, OR, which is supported by the NOAA Vents Program. Comments from Nick Rogers helped to clarify early versions of the paper. Helpful comments were provided by Andrew Kerr, two anonymous reviewers, and editor Dominique Weis. Finally, we would like to thank the EAGLE group and in particular Ian Bastow for stimulating discussions.

REFERENCES

Agranier
A
Blichert-Toft
J
Graham
D
Debaille
V
Schiano
P
Albarède
F
The spectra of isotopic heterogeneities along the mid-Atlantic Ridge
Earth and Planetary Science Letters
 , 
2005
, vol. 
238
 
1–2
(pg. 
96
-
109
)
Albarède
F
Introduction to Geochemical Modelling
 , 
1995
Cambridge
Cambridge University Press
pg. 
543
 
Altherr
R
Henjes-Kunst
F
Puchelt
H
Baumann
A
Volcanic activity in the Red Sea axial trough—evidence for a large mantle diapir
Tectonophysics
 , 
1988
, vol. 
150
 
1–2
(pg. 
121
-
133
)
Andres
M
Blichert-Toft
J
Schilling
J G
Hafnium isotopes in basalts from the southern Mid-Atlantic Ridge from 40°S to 55°S: Discovery and Shona plume–ridge interactions and the role of recycled sediments
Geochemistry, Geophysics, Geosystems
 , 
2002
, vol. 
3
 pg. 
8502
  
doi:10.1029/2002gc000324
Andres
M
Blichert-Toft
J
Schilling
J G
Nature of the depleted upper mantle beneath the Atlantic: evidence from Hf isotopes in normal mid-ocean ridge basalts from 79°N to 55°S
Earth and Planetary Science Letters
 , 
2004
, vol. 
225
 
1–2
(pg. 
89
-
103
)
Arndt
N T
Goldstein
S L
An open boundary between lower continental-crust and mantle—Its role in crust formation and crustal recycling
Tectonophysics
 , 
1989
, vol. 
161
 
3–4
(pg. 
201
-
212
)
Baker
J
Chazot
G
Menzies
M A
Thirlwall
M
Menzies
M A
Klemper
S L
Ebinger
C
Baker
J
Lithospheric mantle beneath Arabia: A Pan-African protolith modified by the Afar and older plumes, rather than a source for continental flood volcanism
Volcanic Rifted Margins
 , 
2002
, vol. 
362
 
Geological Society of America
Special Papers
(pg. 
65
-
80
)
Barrat
J A
Jahn
B M
Joron
J L
Auvray
B
Hamdi
H
Mantle heterogeneity in Northeastern Africa—Evidence from Nd isotopic compositions and hygromagmaphile element geochemistry of basaltic rocks from the Gulf of Tadjoura and Southern Red Sea Regions
Earth and Planetary Science Letters
 , 
1990
, vol. 
101
 
2–4
(pg. 
233
-
247
)
Barrat
J A
Fourcade
S
Jahn
B M
Cheminee
J L
Capdevila
R
Isotope (Sr, Nd, Pb, O) and trace-element geochemistry of volcanics from the Erta'Ale range (Ethiopia)
Journal of Volcanology and Geothermal Research
 , 
1998
, vol. 
80
 
1–2
(pg. 
85
-
100
)
Be'eri-Shlevin
Y
Katzir
Y
Blichert-Toft
J
Kleinhanns
I C
Whitehouse
M J
Nd–Sr–Hf–O isotope provinciality in the northernmost Arabian–Nubian Shield: implications for crustal evolution
Contributions to Mineralogy and Petrology
 , 
2010
, vol. 
160
 
2
(pg. 
181
-
201
)
Benoit
M H
Nyblade
A A
VanDecar
J C
Upper mantle P-wave speed variations beneath Ethiopia and the origin of the Afar hotspot
Geology
 , 
2006
, vol. 
34
 
5
(pg. 
329
-
332
)
Blichert-Toft
J
Chauvel
C
Albarède
F
Separation of Hf and Lu for high-precision isotope analysis of rock samples by magnetic sector multiple collector ICP-MS
Contributions to Mineralogy and Petrology
 , 
1997
, vol. 
127
 
3
(pg. 
248
-
260
)
Blichert-Toft
J
Weis
D
Maerschalk
C
Agranier
A
Albarède
F
Hawaiian hot spot dynamics as inferred from the Hf and Pb isotope evolution of Mauna Kea volcano
Geochemistry, Geophysics, Geosystems
 , 
2003
, vol. 
4
 pg. 
8704
  
doi:10.1029/2002GC000340
Blichert-Toft
J
Albarède
F
The Lu-Hf isotope geochemistry of chondrites and the evolution of the mantle-crust system
Earth and Planetary Science Letters
 , 
1997
, vol. 
148
 (pg. 
243
-
258
)
Blichert-Toft
J
Agranier
A
Andres
M
Kingsley
R
Schilling
J G
Albarède
F
Geochemical segmentation of the Mid-Atlantic Ridge north of Iceland and ridge–hot spot interaction in the North Atlantic
Geochemistry, Geophysics, Geosystems
 , 
2005
, vol. 
6
 pg. 
Q01E19
  
doi:10.1029/2004GC000788
Burke
K
Wilson
J T
Is the African Plate stationary?
Nature
 , 
1972
, vol. 
239
 (pg. 
387
-
390
)
Burov
E
Guillou-Frottier
L
d'Acremont
E
Le Pourhiet
L
Cloetingh
S
Plume head–lithosphere interactions near intra-continental plate boundaries
Tectonophysics
 , 
2007
, vol. 
434
 
1–4
(pg. 
15
-
38
)
Camp
V E
Hanan
B B
A plume-triggered delamination origin for the Columbia River Basalt Group
Geosphere
 , 
2008
, vol. 
4
 (pg. 
480
-
495
)
Camp
V E
Ross
M E
Mantle dynamics and genesis of mafic magmatism in the intermontane Pacific Northwest
Journal of Geophysical Research
 , 
2004
, vol. 
109
 
B08204
pg. 
14
  
doi:10.1029/2003JB002838
Campbell
I H
Griffiths
R W
Implication of mantle plume structure for the evolution of flood basalts
Earth and Planetary Science Letters
 , 
1990
, vol. 
99
 (pg. 
79
-
93
)
Carlson
R W
Esperanca
S
Svisero
D P
Chemical and Os isotopic study of Cretaceous potassic rocks from southern Brazil
Contributions to Mineralogy and Petrology
 , 
1996
, vol. 
125
 
4
(pg. 
393
-
405
)
Carlson
R W
Pearson
D G
James
D E
Physical, chemical, and chronological characteristics of continental mantle
Reviews of Geophysics
 , 
2005
, vol. 
43
 
1
pg. 
RG1001
  
doi:10.1029/2004RG000156
Casey
M
Ebinger
C
Keir
D
Gloaguen
R
Mohamed
F
Yirgu
G
Ebinger
C
Maguire
P
Strain accomadation in transitional rifts: Extension by magma intrusion and faulting in Ethiopian rift magmatic segments
The Afar Volcanic Province within the East African Rift System
 , 
2006
, vol. 
259
 
Geological Society, London
Special Publications
(pg. 
143
-
164
)
Chaffey
D J
Cliff
R A
Wilson
B M
Saunders
A D
Norry
M J
Characterization of the St Helena magma source
Magmatism in the Ocean Basins
 , 
1989
, vol. 
42
 
Geological Society, London
Special Publications
(pg. 
257
-
276
)
Class
C
le Roex
A P
Continental material in the shallow oceanic mantle—How does it get there?
Geology
 , 
2006
, vol. 
34
 
3
(pg. 
129
-
132
)
Courtney
R
White
R
Anomalous heat flow and geoid across the Cape Verde Rise: evidence for dynamic support from a thermal plume in the mantle
Geophysical Journal of the Royal Astronomical Society
 , 
1986
, vol. 
87
 (pg. 
815
-
867
)
Crough
S T
Thermal origin of mid-plate hotspot swells
Geophysical Journal of the Royal Astronomical Society
 , 
1978
, vol. 
55
 (pg. 
451
-
469
)
Davis
J C
Statistical and Data Analysis in Geology
 , 
1973
New York
John Wiley
pg. 
550
 
Day
J M D
Hilton
D R
Pearson
D G
Macpherson
C G
Kjarsgaard
B A
Janney
P E
Absence of a high time-integrated He-3/(U + Th) source in the mantle beneath continents
Geology
 , 
2005
, vol. 
33
 
9
(pg. 
733
-
736
)
Debaille
V
Blichert-Toft
J
Agranier
A
Doucelance
R
Schiano
P
Albarède
F
Geochemical component relationships in MORB from the Mid-Atlantic Ridge, 22–35°N
Earth and Planetary Science Letters
 , 
2006
, vol. 
241
 
3–4
(pg. 
844
-
862
)
Deniel
C
Vidal
P
Coulon
C
Vellutini
P J
Temporal evolution of mantle sources during continental rifting—the volcanism of Djibouti (Afar)
Journal of Geophysical Research: Solid Earth
 , 
1994
, vol. 
99
 
B2
(pg. 
2853
-
2869
)
Detrick
R S
Sinton
J M
Ito
G
Canales
J P
Behn
M
Blacic
T
Cushman
B
Dixon
J E
Graham
D W
Mahoney
J J
Correlated geophysical, geochemical, and volcanological manifestations of plume–ridge interaction along the Galápagos Spreading Center
Geochemistry, Geophysics, Geosystems
 , 
2002
, vol. 
3
 pg. 
8501
  
doi:10.1029/2002GC000350
Douglass
J
Schilling
J G
Systematics of three-component, pseudo-binary mixing lines in 2D isotope ratio space representations and implications for mantle plume–ridge interaction
Chemical Geology
 , 
2000
, vol. 
163
 
1–4
(pg. 
1
-
23
)
Duncan
R A
Hotspots in the Southern Oceans—an absolute frame of reference for motion of the Gondwana continents
Tectonophysics
 , 
1981
, vol. 
74
 (pg. 
29
-
42
)
Ebinger
C J
Casey
M
Continental breakup in magmatic provinces: An Ethiopian example
Geology
 , 
2001
, vol. 
29
 
6
(pg. 
527
-
530
)
Ebinger
C J
Sleep
N H
Cenozoic magmatism throughout east Africa resulting from impact of a single plume
Nature
 , 
1988
, vol. 
395
 (pg. 
788
-
791
)
Ebinger
C J
Yemane
T
Harding
D J
Tesfaye
S
Kelley
S
Rex
D C
Rift deflection, migration, and propagation: Linkage of the Ethiopian and Eastern rifts, Africa
Geological Society of America Bulletin
 , 
2000
, vol. 
112
 
2
(pg. 
163
-
176
)
Ernst
R E
Buchan
K L
Recognizing mantle plumes in the geologic record
Annual Review of Earth and Planetary Sciences
 , 
2003
, vol. 
31
 (pg. 
469
-
523
)
Escrig
S
Schiano
P
Schilling
J G
Allegre
C
Rhenium–osmium isotope systematics in MORB from the Southern Mid-Atlantic Ridge (40°–50°S)
Earth and Planetary Science Letters
 , 
2005
, vol. 
235
 
3–4
(pg. 
528
-
548
)
Falcon
N L
Gass
I G
Girdler
R W
Laughton
A S
A discussion on the structure and evolution of the Red Sea and the nature of the Red Sea, Gulf of Aden, and Ethiopia Rift Junction
Philosophical Transactions of the Royal Society of London, Series A
 , 
1970
, vol. 
267
 (pg. 
131
-
142
)
Farnetani
C G
Richards
M A
Numerical investigation of the mantle plume initiation model for flood basalt events
Journal of Geophysical Research
 , 
1994
, vol. 
99
 (pg. 
13813
-
13833
)
Furman
T
Geochemistry of East African Rift basalts: an overview
Journal of African Earth Sciences
 , 
2007
, vol. 
48
 
2–3
(pg. 
147
-
160
)
Furman
T
Bryce
J G
Rooney
T
Hanan
B B
Yirgu
G
Ayalew
D
Yirgu
G
Ebinger
C
Maguire
P
Heads and tails: 30 million years of the Afar plume
The Afar Volcanic Province within the East African Rift System
 , 
2006
, vol. 
259
 
Geological Society, London
Special Publications
(pg. 
95
-
120
)
Geldmacher
J
Hoernle
K
Hanan
B B
Blichert-Toft
J
Hauff
F
Gill
J B
Schmincke
H-U
Hafnium isotopic variations in East Atlantic intraplate volcanism
Contributions to Mineralogy and Petrology
 , 
2011
, vol. 
162
 (pg. 
21
-
36
)
Girdler
R W
A review of the Red Sea heat flow
Philosophical Transactions of the Royal Society of London, Series A
 , 
1970
, vol. 
267
 (pg. 
191
-
204
)
Goldstein
S L
Soffer
G
Langmuir
C H
Lehnert
K A
Graham
D W
Michael
P J
Origin of a ‘Southern Hemisphere’ geochemical signature in the Arctic upper mantle
Nature
 , 
2008
, vol. 
453
 
7191
(pg. 
89
-
93
)
Graham
D W
Jenkins
W J
Schilling
J G
Thompson
G
Kurz
M D
Humphris
S E
Helium isotope geochemistry of mid-ocean ridge basalts from the South Atlantic
Earth and Planetary Science Letters
 , 
1992
, vol. 
110
 
1–4
(pg. 
133
-
147
)
Graham
D W
Larsen
L M
Hanan
B B
Storey
M
Pedersen
A K
Lupton
J E
Helium isotope composition of the early Iceland mantle plume inferred from the Tertiary picrites of West Greenland
Earth and Planetary Science Letters
 , 
1998
, vol. 
160
 
3–4
(pg. 
241
-
255
)
Graham
D W
Blichert-Toft
J
Russo
C J
Rubin
K H
Albarède
F
Cryptic striations in the upper mantle revealed by hafnium isotopes in southeast Indian ridge basalts
Nature
 , 
2006
, vol. 
440
 
7081
(pg. 
199
-
202
)
Graham
D W
Reid
M R
Jordan
B T
Grunder
A L
Leeman
W P
Lupton
J E
Mantle source provinces beneath the Northwestern USA delimited by helium isotopes in young basalts
Journal of Volcanology and Geothermal Research
 , 
2009
, vol. 
188
 
1–3
(pg. 
128
-
140
)
Griffiths
R W
Campbell
I H
Stirring and structure in mantle starting plumes
Earth and Planetary Science Letters
 , 
1990
, vol. 
99
 
1–2
(pg. 
66
-
78
)
Griffiths
R W
Campbell
I H
Interaction of mantle plume heads with the Earth's surface and onset small-scale convection
Journal of Geophysical Research
 , 
1991
, vol. 
96
 (pg. 
18275
-
18310
)
Hanan
B B
Graham
D W
Lead and helium isotope evidence from oceanic basalts for a common deep source of mantle plumes
Science
 , 
1996
, vol. 
272
 
5264
(pg. 
991
-
995
)
Hanan
B B
Schilling
J G
Easter microplate evolution—Pb isotope evidence
Journal of Geophysical Research: Solid Earth and Planets
 , 
1989
, vol. 
94
 
B6
(pg. 
7432
-
7448
)
Hanan
B B
Schilling
J G
The dynamic evolution of the Iceland mantle plume: the lead isotope perspective
Earth and Planetary Science Letters
 , 
1997
, vol. 
151
 
1–2
(pg. 
43
-
60
)
Hanan
B B
Kingsley
R
Schilling
J G
Pb isotope evidence in the South Atlantic for migrating ridge–hotspot interactions
Nature
 , 
1986
, vol. 
322
 (pg. 
137
-
144
)
Hanan
B B
Shervais
J W
Vetter
S K
Yellowstone plume–continental lithosphere interaction beneath the Snake River Plain
Geology
 , 
2008
, vol. 
36
 
1
(pg. 
51
-
54
)
Hanan
B B
Blichert-Toft
J
Kingsley
R
Schilling
J G
Depleted Iceland mantle plume geochemical signature
Geochemistry, Geophysics, Geosystems
 , 
2000
, vol. 
1
 pg. 
1003
  
doi:10.1029/1999GC000009
Hanan
B B
Blichert-Toft
J
Pyle
D G
Christie
D M
Contrasting origins of the upper mantle revealed by hafnium and lead isotopes from the Southeast Indian Ridge
Nature
 , 
2004
, vol. 
432
 
7013
(pg. 
91
-
94
)
Hart
S R
A large-scale isotope anomaly in the southern-hemisphere mantle
Nature
 , 
1984
, vol. 
309
 
5971
(pg. 
753
-
757
)
Hart
S R
Hauri
E H
Oschmann
L A
Whitehead
J A
Mantle plumes and entrainment; isotopic evidence
Science
 , 
1992
, vol. 
256
 
5056
(pg. 
517
-
520
)
Hart
W K
Woldegabriel
G
Walter
R C
Mertzman
S A
Basaltic volcanism in Ethiopia—Constraints on continental rifting and mantle interactions
Journal of Geophysical Research: Solid Earth and Planets
 , 
1989
, vol. 
94
 
B6
(pg. 
7731
-
7748
)
Hauri
E H
Whitehead
J A
Hart
S R
Fluid dynamic and geochemical aspects of entrainment in mantle plumes
Journal of Geophysical Research: Solid Earth
 , 
1994
, vol. 
99
 
B12
(pg. 
24275
-
24300
)
Hawkesworth
C J
Mantovani
M S M
Taylor
P N
Palacz
Z
Evidence from the Parana of South Brazil for a continental contribution to Dupal basalts
Nature
 , 
1986
, vol. 
322
 
6077
(pg. 
356
-
359
)
Hofmann
A W
Mantle geochemistry: the message from oceanic volcanism
Nature
 , 
1997
, vol. 
385
 (pg. 
219
-
229
)
Hofmann
A W
Hemond
C
Sarbas
B
Jochum
B
Yes, there really is a lead paradox
EOS Transactions, American Geophysical Union
 , 
2005
, vol. 
86
 
52
 
(Fall Meeting Supplement), Abstract V23D-05
Hofmann
A W
Jochum
K P
Seufert
M
White
W M
Nb and Pb in oceanic basalts—New constraints on mantle evolution
Earth and Planetary Science Letters
 , 
1986
, vol. 
79
 
1–2
(pg. 
33
-
45
)
Hofmann
C
Courtillot
V
Feraud
G
Rouchett
P
Yirgu
G
Ketefo
E
Pik
R
Timing of the Ethiopian flood basalt event and implications for plume birth and global change
Nature
 , 
1997
, vol. 
389
 
6653
(pg. 
838
-
841
)
Hopp
J
Trieloff
M
Altherr
R
Neon isotopes in mantle rocks from the Red Sea region reveal large-scale plume–lithosphere interaction
Earth and Planetary Science Letters
 , 
2004
, vol. 
219
 
1–2
(pg. 
61
-
76
)
Ingle
S
Weis
D
Scoates
J S
Frey
F A
Relationship between the early Kerguelen plume and continental flood basalts of the paleo-Eastern Gondwanan margins
Earth and Planetary Science Letters
 , 
2002
, vol. 
197
 (pg. 
35
-
50
)
Janney
P E
Le Roex
A P
Carlson
R W
Hafnium isotope and trace element constraints on the nature of mantle heterogeneity beneath the central Southwest Indian Ridge (13°E to 47°E)
Journal of Petrology
 , 
2005
, vol. 
46
 
12
(pg. 
2427
-
2464
)
Jones
S M
White
N
Shape and size of the starting Iceland plume swell
Earth and Planetary Science Letters
 , 
2003
, vol. 
216
 (pg. 
271
-
282
)
Kamenetsky
V S
Maas
R
Sushchevskaya
N M
Norman
M D
Cartwright
I
Peyve
A A
Remnants of Gondwanan continental lithosphere in oceanic upper mantle: Evidence from the South Atlantic Ridge
Geology
 , 
2001
, vol. 
29
 
3
(pg. 
243
-
246
)
Kent
R W
Storey
M
Saunders
A D
Large igneous provinces—Sites of plume impact or plume incubation
Geology
 , 
1992
, vol. 
20
 
10
(pg. 
891
-
894
)
Leeman
W P
Hawkesworth
C J
Open magma systems—trace-element and isotopic constraints
Journal of Geophysical Research: Solid Earth and Planets
 , 
1986
, vol. 
91
 
B6
(pg. 
5901
-
5912
)
Lustrino
M
How the delamination and detachment of lower crust can influence basaltic magmatism
Earth-Science Reviews
 , 
2005
, vol. 
72
 (pg. 
21
-
38
)
Mahoney
J
Le Roex
A P
Peng
Z
Fisher
R L
Natland
J H
Southwestern limits of Indian Ocean ridge mantle and the origin of low 206Pb/204Pb mid-ocean ridge basalt: isotope systematics of the central Southwest Indian Ridge (17°–50°E)
Journal of Geophysical Research
 , 
1992
, vol. 
97
 (pg. 
19771
-
19790
)
Mahoney
J J
White
W M
Upton
B G J
Neal
C R
Scrutton
R A
Beyond EM-1: lavas from Afanasy-Nikitin Rise and the Crozet Archipelago, Indian Ocean
Geology
 , 
1996
, vol. 
24
 (pg. 
615
-
618
)
Mahoney
J J
Natland
J H
White
W M
Poreda
R
Bloomer
S H
Fisher
RL
Baxter
A N
Isotopic and geochemical provinces of the western Indian Ocean spreading centers
Journal of Geophysical Research
 , 
1989
, vol. 
94
 (pg. 
4033
-
4052
)
Marty
B
Appora
I
Barrat
J-A A
Deniel
C
Vellutini
P
Vidal
P
Kusakabe
M
Shinohara
H
He, Ar, Sr, Nd and Pb isotopes in volcanic rocks from Afar; evidence for a primitive mantle component and constraints on magmatic sources
Geochemical Journal
 , 
1993
, vol. 
27
 
4–5
(pg. 
219
-
228
)
Marty
B
Pik
R
Gezahegn
Y
Helium isotopic variations in Ethiopian plume lavas; nature of magmatic sources and limit on lower mantle contribution
Earth and Planetary Science Letters
 , 
1996
, vol. 
144
 
1–2
(pg. 
223
-
237
)
McKenzie
D
Bickle
M J
The volume and composition of melt generated by extension of the lithosphere
Journal of Petrology
 , 
1988
, vol. 
29
 
3
(pg. 
625
-
679
)
McKenzie
D
O'Nions
R K
Mantle reservoirs and ocean island basalts
Nature
 , 
1983
, vol. 
301
 
5897
(pg. 
229
-
231
)
McKenzie
D
O'Nions
R K
The source regions of ocean island basalts
Journal of Petrology
 , 
1995
, vol. 
36
 
1
(pg. 
133
-
159
)
McNutt
M K
Judge
A V
The superswell and mantle dynamics beneath the South Pacific
Science
 , 
1990
, vol. 
248
 (pg. 
969
-
975
)
Meyzen
C M
Blichert-Toft
J
Ludden
J N
Humler
E
Mevel
C
Albarède
F
Isotopic portrayal of the Earth's upper mantle flow field
Nature
 , 
2007
, vol. 
447
 
7148
(pg. 
1069
-
1074
)
Mohr
P A
Major volcano-tectonic lineament in the Ethiopian rift system
Nature
 , 
1967
, vol. 
213
 
5077
(pg. 
664
-
665
)
Moreira
M
Doucelance
R
Kurz
M D
Dupre
B
Allegre
C J
Helium and lead isotope geochemistry of the Azores Archipelago
Earth and Planetary Science Letters
 , 
1999
, vol. 
169
 
1–2
(pg. 
189
-
205
)
Morgan
W J
Convection plumes in the lower mantle
Nature
 , 
1971
, vol. 
230
 (pg. 
42
-
43
)
Morgan
W J
Reginald
S
Hargraves
R B
Morgan
W J
Van Houten
F B
Burk
C A
Holland
H D
Hollister
L C
Plate motions and deep mantle convection
Studies in earth and space sciences: A memoir in honor of Harry Hammond Hess. Geological Society of America, Memoirs
 , 
1972
, vol. 
132
 (pg. 
7
-
22
)
Pik
R
Marty
B
Hilton
D R
How many mantle plumes in Africa? The geochemical point of view
Chemical Geology
 , 
2006
, vol. 
226
 
3–4
(pg. 
100
-
114
)
Rehkämper
M
Hofmann
A W
Recycled ocean crust and sediment in Indian Ocean MORB
Earth and Planetary Science Letters
 , 
1997
, vol. 
147
 (pg. 
93
-
106
)
Rogers
N
Macdonald
R
Fitton
J G
George
R
Smith
M
Barreiro
B
Two mantle plumes beneath the East African Rift system; Sr, Nd and Pb isotope evidence from Kenya Rift basalts
Earth and Planetary Science Letters
 , 
2000
, vol. 
176
 
3–4
(pg. 
387
-
400
)
Rogers
N W
Prichard
H M
Alabaster
T
Harris
N B W
Neary
C R
The isotope and trace element geochemistry of basalts from the volcanic islands of the southern Red Sea
Magmatic Processes and Plate Tectonics
 , 
1993
, vol. 
76
 
Geological Society, London
Special Publications
(pg. 
455
-
467
)
Rooney
T
Furman
T
Yirgu
G
Ayalew
D
Structure of the Ethiopian lithosphere: Xenolith evidence in the Main Ethiopian Rift
Geochimica et Cosmochimica Acta
 , 
2005
, vol. 
69
 
15
(pg. 
3889
-
3910
)
Rooney
T
Furman
T
Bastow
I D
Ayalew
D
Gezahegn
Y
Lithospheric modification during crustal extension in the Main Ethiopian Rift
Journal of Geophysical Research, B, Solid Earth and Planets
 , 
2007
, vol. 
112
 pg. 
B10201
  
doi:10.1029/2006JB004916
Rooney
T O
Geochemical evidence of lithospheric thinning in the southern Main Ethiopian Rift
Lithos
 , 
2010
, vol. 
117
 
1–4
(pg. 
33
-
48
)
Rooney
T O
Bastow
I D
Keir
D
Insights into extensional processes during magma assisted rifting: evidence from aligned scoria cones and maars
Journal of Volcanology and Geothermal Research
 , 
2011
, vol. 
201
 
1–4
(pg. 
83
-
96
)
Ryan
W B F
Carbotte
S M
Coplan
J O
O'Hara
S
Melkonian
A
Arko
R
Weissel
R A
Ferrini
V
Goodwillie
A
Nitsche
F
Bonczkowski
J
Zemsky
R
Global Multi-Resolution Topography synthesis
Geochemistry, Geophysics, Geosystems
 , 
2009
, vol. 
10
 pg. 
Q03014
  
doi:10.1029/2008gc002332
Salters
V J M
White
W M
Hf isotope constraints on mantle evolution
Chemical Geology
 , 
1998
, vol. 
145
 (pg. 
447
-
460
)
Saunders
A D
Jones
S M
Morgan
L A
Pierce
K L
Widdowson
M
Xu
Y G
Regional uplift associated with continental large igneous provinces: The roles of mantle plumes and the lithosphere
Chemical Geology
 , 
2007
, vol. 
241
 (pg. 
282
-
318
)
Scarsi
P
Craig
H
Helium isotope ratios in Ethiopian Rift basalts
Earth and Planetary Science Letters
 , 
1996
, vol. 
144
 
3–4
(pg. 
505
-
516
)
Schilling
J-G
Iceland mantle plume: geochemical evidence along Reykjanes Ridge
Nature
 , 
1973
, vol. 
242
 (pg. 
565
-
571
)
Schilling
J G
Afar Mantle Plume—rare-earth evidence
Nature, Physical Science
 , 
1973
, vol. 
242
 
114
(pg. 
2
-
5
)
Schilling
J G
Upper mantle heterogeneities and dynamics
Nature
 , 
1985
, vol. 
314
 
6006
(pg. 
62
-
67
)
Schilling
J G
Kingsley
R H
Hanan
B B
Mccully
B L
Nd–Sr–Pb isotopic variations along the Gulf of Aden—Evidence for Afar mantle plume continental lithosphere interaction
Journal of Geophysical Research: Solid Earth
 , 
1992
, vol. 
97
 
B7
(pg. 
10927
-
10966
)
Shervais
J W
Hanan
B B
Lithospheric topography, tilted plumes, and the track of the Snake River–Yellowstone hot spot
Tectonics
 , 
2008
, vol. 
27
 
5
pg. 
TC5004
  
doi:10.1029/2007TC002181
Sleep
N H
Channeling at the base of the lithosphere during the lateral flow of plume material beneath flow line hot spots
Geochemistry, Geophysics, Geosystems
 , 
2008
, vol. 
9
 pg. 
Q08005
  
doi:10.1029/2008GC002090
Sleep
N H
Ebinger
C J
Kendall
J-M
Fowler
C M R
Ebinger
C J
Hawkesworth
C J
Deflection of mantle plume material by cratonic keels
The Early Earth: Physical, Chemical and Biological Development
 , 
2002
, vol. 
199
 
Geological Society, London
Special Publications
(pg. 
135
-
150
)
Snedecor
GW
Cochran
WG
Statistical Methods
 , 
1989
Ames
Iowa State University Press
pg. 
503
 
Storey
M
Saunders
A D
Tarney
J
Gibson
I L
Norry
M J
Thirlwall
M F
Leat
P
Thompson
R N
Menzies
M A
Contamination of Indian Ocean asthenosphere by the Kerguelen–Heard mantle plume
Nature
 , 
1989
, vol. 
338
 (pg. 
574
-
576
)
Stracke
A
Hofmann
A W
Hart
S R
FOZO, HIMU, and the rest of the mantle zoo
Geochemistry, Geophysics, Geosystems
 , 
2005
, vol. 
6
  
Q05007, doi:10.1029/2004GC000824
Thompson
R N
Gibson
S A
Subcontinental mantle plumes, hotspots and pre-existing thinspots
Journal of the Geological Society, London
 , 
1991
, vol. 
148
 (pg. 
973
-
977
)
Todt
R A
Basu
A
Hart
S
Evaluation of a 202Pb–205Pb double spike for high-precision lead isotope analysis
Earth Processes: Reading The Isotopic Code: Geophysical Monograph
 , 
1996
Washington DC
American Geophysical Union
(pg. 
429
-
437
)
Trua
T
Deniel
C
Mazzuoli
R
Crustal control in the genesis of Plio-Quaternary bimodal magmatism of the Main Ethiopian Rift (MER); geochemical and isotopic (Sr, Nd, Pb) evidence
Chemical Geology
 , 
1999
, vol. 
155
 
3–4
(pg. 
201
-
231
)
USGS
Shuttle Radar Topography Mission 1 arc second scene unfilled, unfinished 2.0
 , 
2004
College Park, MD
Global Land Cover Facility, University of Maryland
Van Calsteren
P W C
Harris
N B W
Hawkesworth
C J
Menzies
M A
Rogers
N W
Carswell
D A
Hall
J
Wedepohl
K H
Xenoliths from southern Africa: A perspective on the lower crust
The Nature of the Lower Continental Crust
 , 
1986
Oxford
Blackwell Scientific
(pg. 
351
-
362
)
Volker
F
McCulloch
MT
Altherr
R
Submarine basalts from the Red Sea—New Pb, Sr, and Nd isotopic data
Geophysical Research Letters
 , 
1993
, vol. 
20
 
10
(pg. 
927
-
930
)
Volker
F
Altherr
R
Jochum
K-P
McCulloch
M T
Quaternary volcanic activity of the southern Red Sea; new data and assessment of models on magma sources and Afar plume–lithosphere interaction
Tectonophysics
 , 
1997
, vol. 
278
 
1–4
(pg. 
15
-
29
)
Watts
A B
Gravity and bathymetry in the central Pacific Ocean
Journal of Geophysical Research
 , 
1976
, vol. 
81
 (pg. 
1533
-
1553
)
Wendt
J I
Regelous
M
Niu
Y L
Hekinian
R
Collerson
K D
Geochemistry of lavas from the Garrett Transform Fault: insights into mantle heterogeneity beneath the eastern Pacific
Earth and Planetary Science Letters
 , 
1999
, vol. 
173
 
3
(pg. 
271
-
284
)
White
W M
Hofmann
A W
Sr and Nd isotope geochemistry of oceanic basalts and mantle evolution
Nature
 , 
1982
, vol. 
296
 
5860
(pg. 
821
-
825
)
White
W M
Albarède
F
Télouk
P
High-precision analysis of Pb isotope ratios by multi-collector ICP-MS
Chemical Geology
 , 
2000
, vol. 
167
 (pg. 
257
-
270
)
Willbold
M
Stracke
A
Formation of enriched mantle components by recycling of upper and lower continental crust
Chemical Geology
 , 
2010
, vol. 
276
 (pg. 
188
-
197
)
WoldeGabriel
G
Aronson
J L
Walter
R C
Geology, geochronology, and rift basin development in the central sector of the Main Ethiopia Rift
Geological Society of America Bulletin
 , 
1990
, vol. 
102
 
4
(pg. 
439
-
458
)
Wolfenden
E
Ebinger
C
Yirgu
G
Deino
A
Ayalew
D
Evolution of the northern Main Ethiopian rift: birth of a triple junction
Earth and Planetary Science Letters
 , 
2004
, vol. 
224
 
1–2
(pg. 
213
-
228
)
Zindler
A
Hart
S
Chemical geodynamics
Annual Review of Earth and Planetary Sciences
 , 
1986
, vol. 
14
 (pg. 
493
-
571
)

APPENDIX

Ternary mixing model

The isotope composition of a multi-component mixture is a weighted average of the components contributing to the mixture:  

(1)
formula
where R is the isotope ratio (of Pb, Nd or Sr), subscript i refers to component a, b or c (here a indicates the Afar plume, b depleted mantle, c Pan-African lithosphere), m is the mixture (sample), ci is the Pb, Nd, or Sr concentration in component i, and wi is the mass fraction contributed by each component i to the mixture.

We also have the closure condition  

(2)
formula
Therefore, when all end-member Ri and ci values are known, the measurement of two isotope ratios (such as 206Pb/204Pb and 143Nd/144Nd) for any sample leads to three equations (one for each Rm, plus the closure equation) and three unknowns (wi values), and the system is fully solvable.

The chosen elemental concentrations and isotopic end-members are identical to those of Schilling et al. (1992), except the plume component, which was given a value corresponding to the C mantle reservoir (Hanan & Graham, 1996). A rationale for the end-member isotopic compositions has been given in the text and by Schilling et al. (1992). The derivation of the mixing model, rationale, and method for determining the end-member concentrations have been given in detail by Schilling et al. (1992, appendix).

Table A1:

Major and trace element data for samples from the Wonji Fault Belt following the methods outlined by Furman et al. (2006)

Location Lat. Long. Distance* SiO2 MgO CaO MnO TiO2 Fe2O3 Na2Al2O3 K2P2O5 Rb Sr Ba Zr        
 (°N) (°E) (km)                       
WFB-N-14 8·76 39·61 340 47·01 6·67 11·34 0·16 2·11 11·39 2·91 18·31 0·43 0·28 7·1 514 179 87 18·1        
WFB-1026 9·48 40·20 238 47·16 5·42 9·70 0·23 3·14 14·28 3·44 15·07 0·65 0·63 14·1 395 246 159 31·3        
Location Lat. Long. Distance* SiO2 MgO CaO MnO TiO2 Fe2O3 Na2Al2O3 K2P2O5 Rb Sr Ba Zr        
 (°N) (°E) (km)                       
WFB-N-14 8·76 39·61 340 47·01 6·67 11·34 0·16 2·11 11·39 2·91 18·31 0·43 0·28 7·1 514 179 87 18·1        
WFB-1026 9·48 40·20 238 47·16 5·42 9·70 0·23 3·14 14·28 3·44 15·07 0·65 0·63 14·1 395 246 159 31·3        
 Cs Hf Th Nb Ta Sc Co Ni Cr La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Yb Lu Pb 
WFB-N-14 0·05 0·26 0·97 15·5 0·94 28·9 46 64 90 306 11 25·2 3·5 15·3 3·59 1·42 4·01 0·66 3·32 0·62 1·69 1·42 0·21 1·1 
WFB-1026 0·15 3·8 0·47 1·67 28·8 1·84 29·7 42 34 59 305 18·9 44·3 6·1 26·8 6·4 2·32 6·87 1·06 6·17 1·2 3·2 2·75 0·41 1·5 
 Cs Hf Th Nb Ta Sc Co Ni Cr La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Yb Lu Pb 
WFB-N-14 0·05 0·26 0·97 15·5 0·94 28·9 46 64 90 306 11 25·2 3·5 15·3 3·59 1·42 4·01 0·66 3·32 0·62 1·69 1·42 0·21 1·1 
WFB-1026 0·15 3·8 0·47 1·67 28·8 1·84 29·7 42 34 59 305 18·9 44·3 6·1 26·8 6·4 2·32 6·87 1·06 6·17 1·2 3·2 2·75 0·41 1·5 

*Radial distance from Lake Abhe, Afar.

Table A2:

Model mass proportions of the end-members for the Main Ethiopian Rift data from Table 1

Sample % Afar plume using model:
 
% Depleted mantle using model:
 
% Pan-African lithosphere using model:
 
 Av. Av. Av. 
WFB-1023 33 31 31 37 33 57 61 61 49 57 10 14 10 
WFB-1027 36 36 33 36 35 55 54 61 55 57 
WFB-1026 41 39 37 41 40 52 54 57 51 54 
WFB-E99-2 26 24 25 28 26 62 67 66 57 63 12 10 14 11 
WFB-N-18 34 31 31 34 32 53 60 60 54 57 13 12 11 
WFB-1037 24 22 24 27 24 63 70 65 56 64 12 11 16 12 
WFB-KO-12 30 28 29 34 30 57 61 60 50 57 13 11 11 17 13 
WFB-KO-4B 38 35 34 39 36 51 56 57 49 53 12 13 11 
WFB-N-20 37 35 n.a. 39 37 52 56 n.a. 49 52 11 n.a. 12 11 
WFB-1035 27 26 27 31 28 61 65 62 54 61 12 11 15 12 
WFB-N-22 46 41 41 44 43 44 52 53 47 49 10 
WFB-N-15 30 30 28 34 31 58 59 63 51 58 12 11 15 12 
WFB-1036 37 35 35 38 36 53 57 56 50 54 11 12 10 
WFB-N-21 27 26 26 31 28 61 63 64 53 60 12 10 10 16 12 
WFB-E99-20 31 29 n.a. 33 31 57 62 n.a. 52 57 12 n.a. 14 12 
WFB-N-01 23 21 23 27 24 63 69 64 53 62 13 10 13 20 14 
DZ-1003 11 10 12 12 11 71 81 68 67 72 18 20 21 17 
DZ-1004 23 20 22 22 22 63 72 67 67 67 13 11 11 11 
DZ-1005 15 14 15 15 15 71 77 71 72 73 13 13 13 12 
DZ-1013 18 16 18 18 17 70 76 70 69 71 12 12 13 11 
DZ-1014 24 22 23 25 24 66 71 68 64 67 10 11 
DZ-1009 21 19 21 24 21 68 74 68 61 67 11 11 15 11 
DZ-1007 18 16 18 18 18 70 75 70 69 71 12 12 12 11 
DZ-1008 21 19 21 21 21 68 73 68 68 69 11 11 11 10 
BJ-1044 20 19 21 21 20 71 76 69 70 72 
BJ-1048 10 11 11 10 80 84 78 78 80 11 12 10 
BJ-1049 13 11 14 13 13 78 83 74 78 78 12 
BJ-1051 19 18 20 20 19 72 77 70 71 73 
BJ-1045 21 19 22 21 21 70 75 68 70 71 10 
BJ-1053 21 19 22 21 21 70 76 67 70 71 11 
BJ-1043 17 16 19 18 18 74 78 70 71 73 11 10 
Sample % Afar plume using model:
 
% Depleted mantle using model:
 
% Pan-African lithosphere using model:
 
 Av. Av. Av. 
WFB-1023 33 31 31 37 33 57 61 61 49 57 10 14 10 
WFB-1027 36 36 33 36 35 55 54 61 55 57 
WFB-1026 41 39 37 41 40 52 54 57 51 54 
WFB-E99-2 26 24 25 28 26 62 67 66 57 63 12 10 14 11 
WFB-N-18 34 31 31 34 32 53 60 60 54 57 13 12 11 
WFB-1037 24 22 24 27 24 63 70 65 56 64 12 11 16 12 
WFB-KO-12 30 28 29 34 30 57 61 60 50 57 13 11 11 17 13 
WFB-KO-4B 38 35 34 39 36 51 56 57 49 53 12 13 11 
WFB-N-20 37 35 n.a. 39 37 52 56 n.a. 49 52 11 n.a. 12 11 
WFB-1035 27 26 27 31 28 61 65 62 54 61 12 11 15 12 
WFB-N-22 46 41 41 44 43 44 52 53 47 49 10 
WFB-N-15 30 30 28 34 31 58 59 63 51 58 12 11 15 12 
WFB-1036 37 35 35 38 36 53 57 56 50 54 11 12 10 
WFB-N-21 27 26 26 31 28 61 63 64 53 60 12 10 10 16 12 
WFB-E99-20 31 29 n.a. 33 31 57 62 n.a. 52 57 12 n.a. 14 12 
WFB-N-01 23 21 23 27 24 63 69 64 53 62 13 10 13 20 14 
DZ-1003 11 10 12 12 11 71 81 68 67 72 18 20 21 17 
DZ-1004 23 20 22 22 22 63 72 67 67 67 13 11 11 11 
DZ-1005 15 14 15 15 15 71 77 71 72 73 13 13 13 12 
DZ-1013 18 16 18 18 17 70 76 70 69 71 12 12 13 11 
DZ-1014 24 22 23 25 24 66 71 68 64 67 10 11 
DZ-1009 21 19 21 24 21 68 74 68 61 67 11 11 15 11 
DZ-1007 18 16 18 18 18 70 75 70 69 71 12 12 12 11 
DZ-1008 21 19 21 21 21 68 73 68 68 69 11 11 11 10 
BJ-1044 20 19 21 21 20 71 76 69 70 72 
BJ-1048 10 11 11 10 80 84 78 78 80 11 12 10 
BJ-1049 13 11 14 13 13 78 83 74 78 78 12 
BJ-1051 19 18 20 20 19 72 77 70 71 73 
BJ-1045 21 19 22 21 21 70 75 68 70 71 10 
BJ-1053 21 19 22 21 21 70 76 67 70 71 11 
BJ-1043 17 16 19 18 18 74 78 70 71 73 11 10 

n.a., no data exist for one or more of the isotopes necessary to complete this system. Av. is the average of the four models and is used to plot Fig. 7. Model A = 207Pb/204Pb–206Pb/204Pb, Model B = 208Pb/204Pb–206Pb/204Pb, Model C = 143Nd/144Nd–206Pb/204Pb, Model D = 87Sr/86Sr–206Pb/204Pb.

Supplementary data