Mantle dynamics and petrogenesis of Gomer basalts in the Northwestern Ethiopia: A geochemical perspective

2.1 Location and accessibility

The study area, Gomer (Figure 1), is in the west Gojam Zone of Amhara National Regional State, about 565 km west of Addis Ababa, covering about 507.7 km². The study area can be accessed using an asphalted road, which runs from Addis Ababa through Dejen, Debre-Markos, to Bure, and from Bure to Gomer using a gravel road.

Figure 1

Location of the study areas: (a) geographical map of the northern eastern Africa and Arabia (NASA STRM30) showing the location of the three rifts (Main Ethiopian Rift [MER], Gulf of Aden, and Red Sea) together with the Ethiopian (NW plateau) and Somalian plateaus (SE plateau), surrounding the Afar depression. (b) Location and accessibility map of the study area with its drainage pattern.

2.2 Methodology for geological mapping, petrographic analysis, and geochemical assessment

2.2.1 Geological field mapping and sample collection

The detailed geological field mapping of the study area was conducted based on ground traverses, which selected areas where one can get good stratigraphic sections (such as river or stream sections and road sections). Fieldwork activities involve describing and recognizing the rock types from the representative outcrops, collecting a representative fresh rock sample from each rock type, labeling rock samples, and recording their geographical coordinates. It also involved identifying, describing, and mapping geological structures observed in rocks. Photographs are taken from representative outcrops, which show a peculiar feature. Finally, all data collected from the field were compiled in the notebook, and a detailed geological map of the area was produced (Figure 2).

Figure 2

Geological map of the study area with representative sample location.

2.3 Geological setting

The Ethiopian Cenozoic volcanisms, plateau uplift, and the development of the Main Ethiopian and Afar rift systems have been attributed to the upwelling of the Afar mantle plume, impinging at the base of the Ethiopian lithosphere [9,12,27–29] with a potential link to the African superplume beneath Southern Africa [30,31]. The volcanism covers an area of greater than 600,000 km² [2,32] and was generally dominated by up to 300,000 km³ of basaltic flood lavas with minor pyroclastic products and rhyolitic lava flows, mainly forming the Ethiopian plateau [3]. Most of these basalts and associated felsic rocks have erupted in a short time interval (<5 Ma), with the most significant eruption rates at about 31–29 Ma [5]. They are unconformably underlain by Mesozoic sedimentary sequences or Precambrian metamorphic rocks [33].

The overall volcanites of Ethiopia have been studied to be associated with the various stages of rift development of the East African Rift system [34] and subdivided into five significant provinces [35] based on lithological development, type of activity, age of effusion, and frequency of volcanic centers: (1) volcanites of the northern plateau; (2) volcanites of the southern plateau; (3) volcanites of the Somali plateau; (4) Afar volcanites; and (5) MER volcanites. The first three groups represent a significant part of the Ethiopian volcanites, collectively known as “Traps” [36]. Among these volcanoes, the fine-grained stratified flood Paleocene basalts represent the more significant portion. The Afar and MER volcanites are referred to as the Aden Series [36], related to Neocene continental fragmentation, and have a more limited extension.

The northwestern Ethiopian Plateau is extensively covered by Eocene-Miocene CFBs [37], which comprise the Ashenge and Aiba basaltic and Alaje rhyolitic formation with an estimated local thickness of 500–2,000 m [2]. The emplacement of these flood basalts and the associated ignimbrites were concomitant with the continental rifting of the Red Sea-Gulf of Aden (late Oligocene) but predates the development of the Main Ethiopian Rift (late Miocene) in association with the first appearance of the Afar plume head [5,9,38]. They erupt through fissures in the underlying metamorphic and sedimentary succession [2]. The plateau basalts are commonly tholeiitic to alkaline in composition [6,7,27], and they are spatially grouped rather than temporally based on their contrasting geochemical signatures into low-Ti (LT), high-Ti1 (HT1), and high-Ti2 (HT2) basalts [6,7,12]. The LT basalts are restricted to the northwestern part of the province, while the HT basalts are found in the eastern and southern parts of the northwestern plateau. However, the mineralogy and geochemistry of such plateau basalts are similar, and most rocks are described as aphyric to sparsely phyric and contain plagioclase and clinopyroxene phenocrysts with rare or without olivine [4].

The northwestern flood basalt eruption was followed by the development of several low-angle, less-voluminous, large shield volcanoes (e.g., Guna ∼10 Ma; Choke and Guguftu ∼23 Ma; [4] on the surface of the plateau from 30 Ma to ∼10 Ma, whose compositions and age match the underlying flood basalts with a thinner and less continuous lava flows than the underlying flood basalts [4]. They have a relief of about 1–2 km above the original plateau surface [4,39]. A systematic age variation is observed among shield volcanoes from the northwest plateau (Choke and Guguftu ∼22.4–23.3 Ma [4] and the south-east part of the northwest plateau (near Gebra Guracha) [39] to the southeastward direction in central Ethiopia (∼15–10 Ma) [40]. The shield volcanoes are bimodal in nature and contain sequences of alternating basalts, rhyolitic and trachytic lava flows, tuffs, and ignimbrites, particularly near their summits, dominated by transitional to Na-alkaline basalts [3]. During the Quaternary period, minor alkaline basaltic eruptions (with Scoria cones, tuff cones, tuff rings, and maars) have also occurred on the Ethiopian plateau, south and southwest of Tana Lake, which are considered to be related to the newly developed rifts and fractures (Tana rift) [41,42].

3.1 Geology of the study area

The study area is part of the northwestern Ethiopian plateau and is dominated by differently textured basalts such as vesicular basalt, amygdaloidal basalt, and massive basalt with minor thinly bedded scoriaceous basalt, agglomerate, and slightly metamorphosed granitic rocks.

3.1.1 Basaltic unit

Basaltic rocks are exposed throughout the study area in many places, such as at hilltops, hillsides, along roads, and river cuts, with irregular distributions in thickness (Figure 3c). This rock unit covers a higher proportion of up to 420 km² and comprises vesicular basalt, amygdaloidal basalt, scoriaceous basalt, and massive basalt [43]. The rock unit is characterized by dark to light gray fresh colors and pink to pinkish gray weathered surface. This unit consists of many horizontally stratified massive basaltic flows of 1–2 m thickness and experienced an exfoliation type of weathering. At the study area’s northeastern and southeastern parts, this unit contains vesicles of variable size (about ∼10 cm) and shapes ranging from spherical to elongated. The vesicles are usually elongated in the N–S direction [44]. The amygdaloidal basalt is irregularly distributed and spattered in many places with small (∼12%) aerial coverage. It is primarily exposed in the study area’s central, northern, and northeastern parts (Figure 3a) and forms flat topography [42]. Sometimes, vesicles of basalts from many outcrops of the study area are filled with whitish secondary minerals, such as quartz, zeolites, calcite, and dolomite (Figure 3b). It has dark to gray fresh colors and a pink to light gray weathered surface. These secondary minerals are usually elongated in the N–S direction with variable sizes of about <2 cm, similar to their vesicles. In the southern part of the study area, meta granites unconformably underlie this rock unit and are conformably overlain by a scoriaceous basaltic unit. The lower part of this unit is marked by laterally continuous reddish color and thin (about 1.5 m thick) paleosol layers [45].

Figure 3

Field photographs of amygdaloidal basalt from different representative outcrops. (a) The amygdaloidal basalt from the hilltop, with minerals of quartz, zeolites, calcite, and dolomite infills. (b) The amygdaloidal basaltic unit from the quarry site has quartz, zeolites, calcite, and dolomite minerals in the fills. (c) Vesicular basalt from the hill top.

3.1.2 Scoriaceous basalt

This unit is rare, intercalated with thin layers of not more than 2 m with basaltic flows of the study area. It is also sparsely spattered on the surface of basaltic lavas (Figure 4a). It is small to map but mainly exposed in the south and northwestern parts of the study area along hillsides and quarries. The first episode of volcanism of the scoriaceous basalt is underlain by basement rock (meta-granite) as its boundary is marked by a 1.5 m-thick paleosol derived from a granitic rock unit (Figure 4b). This unit has deep gray to brown, weathered, and reddish fresh colors. Sometimes, vesicles are filled with white secondary minerals such as quartz, calcite, dolomite, and zeolites.

Figure 4

Field photographs of the scoria unit from different representative outcrops: (a) the scoria unit is sandwiched between the two massive stratified basaltic units. (b) The first episodic volcanism of the scoria unit is underlain by reddish-colored, about 1.5 m-thick paleosoil, and overlain by a massive stratified basaltic unit.

3.1.3 Agglomerate

This rock unit is exposed in scattered (patchy) form at road and stream cuts and is found mainly in the southern parts of the area between the two basaltic units (Figure 5a). It is overlain by a 35 m-thick, columnar-jointed dark gray aphanitic basalt and shows a sharp contact with this basaltic unit (Figure 5b). It has an average thickness of about 10 m with N–S orientation. Moreover, this rock unit is characterized by pinkish, reddish, greenish gray, and yellowish weathered colors and by the presence of subangular to subrounded pebbles, blocks, and bombs of scoria and basalt rock fragments in a fine ash and lapilli matrix. The clasts are commonly silicified, epidotized, unconsolidated, and poorly sorted.

Figure 5

(a) Field photographs of agglomerate from different representative outcrops. (a) The agglomeratic unit from the hillside with fragments of volcanic rocks having different shapes ranging from sub-angular to sub-rounded pebbles, blocks, and bombs of scoria and basalt rock fragments within a fine ash and lapilli matrix; this is overlain by stratified basaltic lava. (b) The agglomeratic unit from a hilltop with different-sized and shaped clasts of volcanic rocks and basalts ranges from sub-angular to sub-rounded pebbles, blocks, and bombs of scoria and basalt rock fragments within a fine ash and lapilli matrix.

3.1.4 Meta-granites

Meta-granitic rocks are mainly exposed in the southern, southwestern, and southeastern parts (Figure 6a) of the study area, where intensive surface erosion was higher (stream cuts) and sometimes, at road cuts (Figure 6a and b). Outcrops of this rock unit are mostly found relatively at the lowest elevation, with a boulder-like appearance in association with non-metamorphosed, highly weathered (fragmented) volcanic rocks and soils of the area. The meta-granites are also overlain by scoriaceous basalt marked by a 1.5–4 m thick paleosol. The meta-granites are pinkish to gray, medium to extremely coarse-grained, and foliated rock. At places, the foliation of this unit is defined by a precise alignment of mafic (hornblende, biotite, and chlorite) and felsic (quartz and both feldspars) minerals in the N–S general direction. The association of foliation with this unit implies that this rock is part of the basement rock. These meta-granitic rocks are interpreted as part of the Precambrian basement exposed through erosional windows in the southern parts of the study area. Their spatial distribution at low elevations and the observation that they are unconformably overlain by scoriaceous basalts with paleosol development confirm that they represent the pre-volcanic basement complex, rather than intrusive bodies related to the Tertiary volcanism. The rock mainly comprises orthoclase feldspars, about ∼70%, with minor quartz, plagioclase feldspars, hornblende, muscovite, and biotite. Figure 7 shows the geological map of the study area.

Figure 6

Field photographs of meta-granites (a and b) from different representative outcrops, surrounded by transported and residual regolith.

Figure 7

Geological map of the study area with representative sample location.

3.2 Petrography

Petrographically, all rock samples exhibit porphyritic texture with a highly variable phenocryst assemblage of olivine, clinopyroxene, and plagioclase. Plagioclase and pyroxene form the main phenocryst phases, and small amounts of Olivine phenocrysts are found in these rocks. Accordingly, two petro-types are recognized: Plagioclase-Pyroxene phyric basalt and plagioclase phyric basalt. Samples, such as SH-1, SH-2, SH-5, SH-8, SH-7, GO-1, and GO-3, only contain plagioclase (<10%) phenocrysts, while those KO-8, WO-1, GO-33, KO-22, WO-2, and KO-1 have abundant pyroxene (5–20%) and plagioclase phenocrysts (10–20%) and minor olivine phenocrysts (<2%).

3.2.1 Plagioclase-pyroxene phyric basalt

The plagioclase-pyroxene phyric samples mainly comprise olivine, clinopyroxene, and plagioclase minerals [46]. Samples also show highly porphyritic to glomeroporphyritic textures (Figure 8a) with abundant, euhedral to subhedral phenocrysts of clinopyroxene (1–3 mm) and plagioclase (2–4 mm) and subordinate anhedral olivine (1–2 mm) set in intergranular textured groundmass of the same mineral assemblage as the phenocrysts, with some Fe–Ti oxide (opaques) and glasses (Figure 8b) (Roex et al., [47]). The modal composition of the samples is 66–72% microcrystalline groundmass and 24–32% phenocrysts. The proportion of the phenocrysts is 65–74% plagioclase, 15–25% clinopyroxene, 1–3% olivine, and <1% Fe–Ti oxides. The groundmass is dominated by microcrystalline plagioclase, clinopyroxene, olivine, and Fe–Ti oxides (Figure 8c) [48]. Therefore, in general, the modal composition of this rock unit is dominated by plagioclase, followed by clinopyroxene and olivine. Plagioclase shows polysynthetic twinning. Besides the porphyritic to glomeroporphyritic textures, the clinopyroxene crystals included the anhedral olivine crystals (poikilitic texture) [49] (Figure 8d).

Figure 8

Microphotograph of plagioclase-pyroxene phyric basalt under optical microscope, at 4× magnifications, in cross-polarized light (XPL) (a and c) and Plane-polarized light (PPL) (b and d) views. The labels of the identified minerals are as follows: Ol, olivine; Cpx, clinopyroxene; An, anorthite; and Opa, opaques. (a) The clusters of plagioclase and clinopyroxene phenocrysts are embedded in a glassy groundmass of the same mineral as phenocryst with some olivine and opaque (probably Fe–Ti oxides); anhedral olivine crystals are included in the subhedral clinopyroxene; the subhedral to euhedral Anorthite shows a polysynthetic twinning and zoning. (b) A groundmass of opaques embedded in the plagioclase and clinopyroxene phenocrysts. (c) Glomerophyric textured phenocrysts of plagioclase and clinopyroxene are in contact with one another and are embedded in a glassy groundmass of the same mineral as phenocryst with some olivine and opaque (probably Fe–Ti oxides). (d) Ground mass of opaque (probably Fe–Ti oxides) embedded the glomerophyric textured phenocrysts of plagioclase and clinopyroxene crystals.

3.2.2 Plagioclase phyric basalt

The plagioclase phyric samples comprise clinopyroxene and plagioclase with very minor olivine minerals [50]. They show a highly porphyritic to glomeroporphyritic, locally trachytic texture with phenocrysts of euhedral to subhedral plagioclase (1–3 mm) and very subordinate, euhedral to anhedral clinopyroxene (1–2 mm) set in intergranular and interseptal textured groundmass of plagioclase laths (<0.5 mm), granular clinopyroxene (<0.5 mm), euhedral to subhedral opaque (Fe–Ti oxides) (<0.5 mm), and/or glasses (Figure 9a). The modal composition of the samples is 75–85% microcrystalline groundmass and 15–25% phenocrysts. The phenocryst phases mainly comprise plagioclase (∼90%) and clinopyroxene (∼7%) with minor Fe–Ti oxide microphenocrysts (Figure 9b). The groundmass is dominated by microcrystalline plagioclase, clinopyroxene, olivine, and Fe–Ti oxides (Figure 9c). Most plagioclase phenocrysts are elongated and randomly oriented, often showing a glomerophyric texture with clusters of plagioclase crystals (Figure 9d). Furthermore, the plagioclase phenocrysts show compositional zoning and fine-grained opaque phases within their crystals, indicating a resorption texture [51].

Figure 9

Microphotograph of plagioclase phyric basalt under an optical microscope, at 4× magnifications, in XPL (a and c) and PPL (b and d) views. The labels of the identified minerals are as follows: Ab, albite; An, anorthite; Cpx, clinopyroxene; and Opa, opaques. (a) Phenocrysts of plagioclase crystals are embedded in a glassy groundmass of the same mineral as phenocryst with clinopyroxene and some opaque (probably Fe–Ti oxides); the subhedral to euhedral anorthite shows a polysynthetic twinning and zoning. (b) The inclusion of opaque (Fe–Ti oxides) in albite phenocrysts (resorption texture). (c) Glomerophyric textured phenocrysts of plagioclases are in contact with each other and are embedded in a glassy groundmass of the same mineral as phenocryst with some clinopyroxene and opaque (probably Fe–Ti oxides) groundmass. (d) Ground mass of opaque (probably Fe–Ti oxides) with clinopyroxene has embedded the glomerophyric textured phenocrysts of plagioclase crystals.

3.3 Whole-rock geochemistry

The results of whole-rock major and trace elements and CIPW norm analyses for seven samples from Gomer basalts are presented in Table 1, reflecting their nature. These analyses are very important for interpreting geochemical and petrologic data trends. They can also be used to classify and name volcanic rocks that may have few identifiable minerals [52]. The whole-rock trace element geochemical data for the Gomer volcanic rock samples exhibit notable variations across several trace elements and ratios. Scandium (Sc) and vanadium (V) contents are relatively consistent, with V peaking in WO-1 at 480 ppm, while chromium (Cr) and cobalt (Co) display significant fluctuations, particularly with WO-1 showing a very low Cr concentration of 10 ppm compared to other samples.

Key trace elements like rubidium (Rb), strontium (Sr), and yttrium (Y) show variability, with Rb highest in WO-1 and SH-8, and Sr highest in WO-1 at 356 ppm. Zirconium (Zr) ranges from 81 ppm in GO-1 to 219 ppm in WO-1, while niobium (Nb) content also varies, peaking in WO-1. Other elements, such as barium (Ba) and hafnium (Hf), show similar trends, with Ba reaching 315 ppm in SH-8. REEs like lanthanum (La) and cerium (Ce) are notably enriched in WO-1, showing higher concentrations in comparison to other samples. Europium (Eu) anomalies (Eu/Eu*) indicate slight to moderate depletion across samples, with ratios ranging from 1.56 to 2.07. Trace element ratios further reveal petrogenetic insights, with Nb/Y ratios ranging from 0.224 to 0.413, suggesting source characteristics and degree of partial melting. Nb/U and Ce/Pb ratios indicate variability in the source composition, with a high Ce/Pb in GO-1 pointing toward possible crustal influences. These variations collectively point to a complex petrogenetic history for the Gomer volcanic rocks, with notable fractional crystallization and mantle source characteristics.

3.3.1 Geochemical classification

In the total alkali (Na₂O + K₂O) vs silica (SiO₂) diagram (Figure 10a) [53], all samples fall within the field of basalt and show sub-alkaline composition. In the Zr/TiO₂ vs Nb/Y diagrams (Figure 10b) [54], samples are plotted in the subalkaline basalt field. The samples also fall in the field of tholeiitic series both in the FeO_t/MgO vs SiO₂ diagram (Figure 10c) [55], as well as in the (Na₂O + K₂O)–FeO_t–MgO (AFM) ternary diagram (Figure 10d) [56]. Samples are hypersthene and quartz normative (quartz-tholeiites), consistent with the above classifications. These basalts’ tholeiitic and hypersthene–quartz normative character indicates derivation from a Fe-rich, moderately depleted mantle source under relatively dry conditions. This supports partial melting within the spinel stability field, characteristic of enriched intraplate settings such as those associated with early-stage continental rifting.

Figure 10

(a) SiO₂ vs (Na₂O + K₂O) (TAS) diagram (after Cox et al. [53]), the Gomer rock samples plotted within the basalt field (b): Zr/TiO₂ vs Nb/Y diagram after Winchester and Floyd [54], the Gomer rock samples plotted within the subalkaline basalt field. (c) FeO_t/MgO vs SiO₂ variation diagram (after Miyashiro [55]), and (d) Na₂O + K₂O–FeO_t–MgO ternary diagram (after Irvine and Baragar [56]), the Gomer rock samples plotted within the tholeiitic series field.

The geochemical analysis of the Gomer volcanic rock samples shows a silica (SiO₂) content range of 47–52 wt%, with major oxides including Fe₂O₃, MgO, CaO, and Na₂O. Notably, TiO₂ varies between samples, peaking at 3.14 wt% in sample WO-1 and lowering to 1.56 wt% in sample GO-1. The CaO/Al₂O₃ ratio remains stable, ranging from 0.63 to 0.75. CIPW norm mineralogy highlights quartz, orthoclase, albite, and anorthite as dominant normative minerals, with higher quartz levels in SH-8 and increased anorthite in GO-1 and KO-1. Diopside and hypersthene are present, though olivine and wollastonite are absent in all samples. Minor components include consistent titanite and hematite across samples, while apatite is most concentrated in WO-1. These results point to a tholeiitic composition, with mineral variations suggesting fractional crystallization as a key factor in the petrogenesis of these rocks.

Pik et al. [6] classified the northwestern Ethiopian Plateau basalts into three distinct magma types based on their TiO₂ abundances: low-Ti (LT), high-Ti1 (HT1), and high-Ti₂ (HT2) basalts. The LT lavas display the lowest TiO₂ contents (1–2.6 wt%) commonly found in CFB provinces. On the other hand, the two other lava groups (HT1 and HT2) have much higher TiO₂ contents. The concentration of TiO₂ content in HT1 magmas varies from 2.5 to 4.2 wt%, while in HT2, it ranges from 2.6 to 5 wt%. Similarly, Pik et al. [6] also suggested that the LT basalts have relatively low Ti/Y (288–437) and Nb/Y (0.1–0.41) values, whereas in the HT1 basalt, Ti/Y value ranges from 352 to 814 and Nb/Y from 0.52 to 1.1. The HT2 basalts are characterized by high values of Ti/Y and Nb/Y that range from 670 to 885 and 0.9 to 1.44, respectively. In the plots of Ti/Y vs Nb/Y, tholeiitic samples from the Gomer area fall in the LT field.

3.3.2 Variation diagram

Bivariate major oxide plots against MgO as a differentiation index are illustrated in Figure 11 [57], which recommends the use of MgO as a differentiation index in the variation diagrams when the range of SiO₂ (43.1–51.7 wt%) contents between samples is small, and the rock series includes abundant mafic members. The basalt samples depict a continuous decrease in Na₂O, K₂O, and P₂O₅ contents with increasing MgO and an increase in TiO₂ and CaO contents, while Al₂O₃ and FeO_t show slope variations in their trends; from MgO values 4 to 5.5 wt%, Al₂O₃ shows positive correlation, while FeO_t shows negative correlation. From MgO values 5.5–6.15 wt%, both oxides show the reverse events. The co-variations of FeO_t and CaO with MgO in Gomer basalts confirm the suggestion of fractional crystallization of olivine, clinopyroxene ± spinel.

Figure 11

Plots of major oxides vs Mg₂O for Gomer basalt samples.

Samples also show a great variation in compatible and incompatible trace element concentrations. They have low Ni (26–72 ppm), Co (22–51 ppm), and Cr (10–150 ppm) contents. Some selected compatible and incompatible trace elements against MgO plots for Gomer basalts are shown in Figure 12. Ni, Cr, and Co exhibit a positive correlation with MgO. In contrast, compatible elements (e.g., V) and incompatible trace elements (Sr, La, Zr) show negative correlations with a well-defined evolution trend with MgO. Eu shows a positive correlation from MgO values 4–5.5 wt% and a negative correlation from MgO values 5.5–6.15 wt%.

Figure 12

Selected compatible and incompatible trace element abundances (ppm) vs MgO (wt%) for the Gomer basalts.

3.3.3 Spider diagram

The primitive mantle-normalized multi-element diagram of the representative samples is shown in Figure 13. The diagram for all samples shows a right declining trend with decreasing incompatibility of trace elements similar to E-MORB [58]. All lavas also show enrichment in Ba, U, K, Pb, Sr, Nd, and Ti and depletion in Rb, Th, Ta, Eu, and P, with most samples showing slightly negative Ce and Nb anomalies. The pronounced troughs at P, Nb, and Ta, and peaks at Pb, U, and Ba, observed in the multi-element spider diagram (Figure 13), along with elevated La/Ybₙ ratios ranging from 2.19 to 3.85, strongly suggest interaction with crustal components during magma evolution [53]. E-MORB reference values have been plotted alongside the Gomer basalt samples to strengthen the geochemical comparison with potential mantle sources. In addition, a supplementary diagram has been included with all samples normalized directly to E-MORB, providing a clearer basis for evaluating mantle source characteristics and the degree of enrichment.

Figure 13

(a) Primitive mantle-normalized multi-element spider diagrams for Gomer basalts and basalts from the Northwestern Ethiopian plateau and Ethiopian rifts, plotted for comparison; the REE element abundances of oceanic basalts (N-MORB, E-MORB, and OIB) from [58] are also plotted for references; (a) multi-element spider diagram for Gomer basalts; (b) multi-element spider diagram for Gomer basalts and oceanic basalts (N-MORB, E-MORB and OIB); (c) multi-element spider diagram for Gomer basalts and Oligocene flood basalts from [6,7] (d) multi-element spider diagram for Gomer basalts and Afar basalts from [59] (e) multi-element spider diagram for Gomer basalts and MER basalts from [31]. Normalization values are from [58].

3.3.4 REE diagram

The chondrite-normalized REE patterns of the representative samples are given in Figure 14. All samples are characterized by variably moderate enrichment in LREE (La/SmN = 1.19–1.71) and less fractionated sub-parallel REE (La/YbN = 2.19–3.85) and HREE patterns. The analyzed samples show pronounced small negative Eu anomalies (Eu/Eu* = 1.56–2.7) (Where the subscript N means chondrite-normalized [58]. All samples show similar trends, suggesting similar magma sources and REE fractionation degrees. They are well correlated with Enriched Mid Oceanic Ridge Basalt and are distinct from Ocean Island basalt (OIB). Furthermore, an E-MORB-normalized REE diagram is included in the supplementary figures to directly compare with this mantle end member, supporting the discussion of source enrichment and melting conditions.

Figure 14

Chondrite-normalized REE diagrams for Gomer basalts and basalts from Northwestern Ethiopian plateau and Ethiopian rifts, plotted for comparison; the REE element abundances of oceanic basalts (N-MORB, E-MORB, and OIB) from [58] are also plotted for references; (a) REE diagram for Gomer basalts; (b) REE diagram for Gomer basalts and oceanic basalts (N-MORB, E-MORB and OIB); (c) REE diagram for Gomer basalts and Oligocene flood basalts from [6,7]; (d) REE diagram for Gomer basalts and Afar basalts from [59]; (e) REE diagram for Gomer basalts and MER basalts from [31]. Normalization values are from [58].

4.1 Fractional crystallization

The analyzed basaltic samples from the Gomer area are not primary mantle melts known from their low MgO (400–500 ppm, Cr: >1,000 ppm, and MgO: 10–15 wt% [61]. These characteristics suggest that they have undergone olivine, clinopyroxene ± spinel fractionation from parental magmas either in magma chambers or en route to the Earth’s surface. The low to moderate concentrations of Ni (26–75 ppm) and Cr (10–150 ppm) and the co-variations of Ni, Cr, CaO, FeO_t, and Al₂O₃ with MgO in Gomer basalts confirm the suggestion of fractional crystallization of olivine, clinopyroxene ± spinel. The porphyritic to glomeroporphyritic textures with plagioclase, pyroxene, and Olivine phenocrysts in Gomer basalts imply that the magma was crystallizing these mineral phases before reaching the earth’s surface. Extremely coarse-grained phenocrysts of plagioclase, pyroxene, and Olivine surrounded by a fine-grained matrix of the same minerals indicate bimodal crystallization of Gomer basalts began from a deep-seated magma chamber with high-pressure and temperature conditions.

The mineral assemblages dominated by olivine, clinopyroxene, and plagioclase, and the absence of garnet-phase influence in REE patterns, suggest that the Gomer magmas fractionated within the spinel peridotite stability field, corresponding to depths shallower than ∼80 km. Mineral-melt equilibrium models (e.g., Putirka [62]) support crystallization pressures of ∼0.4–0.8 GPa, typical for lower to middle crustal magma chambers. The relatively high (Tb/Yb)ₙ values (1.29–1.70) are still below the garnet stability threshold (>1.8), reinforcing the view that no residual garnet was involved during fractional crystallization. These data collectively support a low-pressure, crustal magma storage regime, with differentiation likely occurring in shallow-seated reservoirs, followed by ascent and final crystallization in feeder conduits or subvolcanic bodies.

The presence of abundant clinopyroxene and plagioclase phenocrysts, particularly in the form of glomerocrysts and poikilitic textures, indicates a prolonged crystallization history at depth prior to eruption. Their modal proportions (clinopyroxene 5–20%, plagioclase 10–20%) and crystal size distribution are consistent with equilibrium crystallization at moderate to high pressures, likely within mid- to lower-crustal magma chambers. These features, along with the lack of pervasive zoning or disequilibrium textures, support a model of polybaric fractional crystallization, where minerals began crystallizing under deep-seated conditions and continued through magma ascent. Although detailed mineral chemistry was not conducted, the petrographic characteristics provide indirect evidence for high-pressure crystallization regimes. Furthermore, the positive correlation of CaO/Al₂O₃ with MgO contents in the lavas also indicates that clinopyroxene fractionation plays a significant role in the differentiation of Feldspars. Apatite and Ti–Fe oxides are not a significant fractionation mineral assemblage given the negative correlation between K₂O, Na₂O, P₂O_5, and TiO₂ with MgO and the positive anomalies in the trace element profiles.

Moreover, the positive Sr and Eu anomalies (Eu/Eu* = 1.56–2.07) on the normalized spider diagrams for these samples and the increase in Al₂O₃ and Sr concentrations with decreasing MgO also imply that plagioclase fractionation from the mafic lavas was not undergone in the early stages of differentiation. The positive Eu and Sr anomalies in all samples may also indicate the selective retention (accumulation) of plagioclase crystals in magmas during differentiation or contamination of mantle-derived melts by lower-crustal material [63]. As the major and trace element abundances change regularly and predictably with progressive differentiation [64,65], the overall major and trace element data trends for Gomer basalts generally show coherent trends; there are no considerable scatter plots. Thus, this suggests that significant and trace element variations within the suite are due to fractional crystallization processes [65].

4.2 Crustal contamination

Mantle-derived magmas in continental settings can be subjected to variable degrees of contamination by different crustal components during either their ascent through or stagnation in a thick and compositionally variable region, which leads to their compositional diversity [66,67]. The different mechanisms by which crustal contamination may arise include the bulk assimilation of crustal material, the assimilation of partial melt-derived crustal materials, and the selective exchange of specific elements aided by the transfer of fluids from crust to melt [57].

The pronounced troughs at P, Nb, and Ta and peaks at Pb, U, and Ba of the Gomer tholeiitic samples with a La/YbN (2.19–3.85) ratio of greater than one and with no Zr Hf anomaly suggest that their parent magmas were contaminated by crustal material. Moreover, the Ce/Pb and Nb/U ratios of ocean basalts (OIBs and MORBs) are reasonably well-constrained and are relatively uniform (Ce/Pb = 25 ± 5 and Nb/U = 47 ± 10) [5], reflecting the composition of the Earth’s mantle and therefore, are sensitive indicators of crustal contamination of mantle-derived mafic magmas. Basaltic lavas with lower Ce/Pb and Nb/U values reflect a crustal overprint to the primary mantle-derived signature [5,58]. The basalt samples from the Gomer area have lower Ce/Pb (2.9–9) and Nb/U (19.05–35.33) values, a low-to-the-range typical for oceanic basalts [5]. Yet, the values are much more similar to those of continental crust (CC) [68] than mantle composition, which supports the idea of crustal contamination in the analyzed tholeiitic samples. Similarly, in Nb/U vs SiO₂ plot of Gomer basalts showing a trend away from the oceanic MORB-OIB array toward lower Nb/U ratios reflecting the role of lithospheric mantle in the genesis of parent magma and magma-crust interactions.

In addition, La/Ta ratios for primary mantle-derived melts are well-defined (La/Ta < 22) [69] and are sensitive to contamination, making them valuable indicators for crustal assimilation of mafic rocks [31]. Thus, the crustal contamination effects of the basaltic lavas from the Gomer area were also confirmed by their higher La/Ta (17–40.3) ratios than the La/Ta ratio for the primary mantle-derived melts (La/Ta < 22) [69].

Figure 15a compares Nb/U ratios to SiO₂ contents across Gomer basalt samples. Most samples plot below the typical mantle Nb/U range (∼47 ± 10), particularly those with higher SiO₂ values. This trend is characteristic of crustal contamination, as the CC has lower Nb/U ratios than the mantle. The data suggest that while some samples (e.g., Wo-1, SH-7) retain near-mantle Nb/U values, others (e.g., Ko-1, SH-2) have likely undergone interaction with the crust, lowering their Nb/U ratios. This is consistent with assimilation-fractional crystallization processes, especially in samples showing evolved compositions.

Figure 15

(a) Variation of Nb/U vs Ce/Pb of Gomer basalts. Mantle compositional fields are after Hofmann et al. [5]. The average composition of the upper and LCC (av. UCC and av. LCC, respectively) is from Rudnick and Gao [68]. (b) Nb/U vs SiO₂ plot of Gomer basalts showing a trend away from the oceanic MORB-OIB array toward lower Nb/U ratios reflecting the role of lithospheric mantle in the genesis of parent magma and magma-crust interactions.

The plot (Figure 15b) uses Ce/Pb and Nb/U ratios to distinguish between mantle-derived and crustally modified basalts. The shaded fields represent average mantle values (Nb/U ≈ 47 ± 10, Ce/Pb ≈ 20 ± 5). The Gomer basalt samples plot significantly below both mantle fields, clustering near the average lower continental crust (LCC) and upper continental crust (UCC) values. This indicates that the magma experienced significant crustal contamination, likely during its ascent through thick lithosphere or stagnation in crustal magma chambers.

4.3 Degree of partial melting

The overall data trends, for example, patterns on normalized multi-element and REE diagrams for Gomer basalts, the non-scattered plots on the primary and trace element variation diagrams, and the similar incompatible trace elements ratios between samples indicate that the volcanic groups are identical in their source and their parental magmas were derived from the same mantle sources by the same degrees of partial melting. The hypersthene-quartz normative nature of the Gomer basaltic suite supported that their parental magma generation was attributed to the high degree of partial melting from shallow mantle sources [70]. Furthermore, similar sub-parallel patterns on normalized multi-element and REE diagrams for Gomer samples suggest the same mantle source (co-genetic) [71]. In addition, the rare-earth elements (REEs) are also the most vital informants to understand the mantle melting conditions, as their relative concentration in mantle-derived magmas reflects the degree of partial melting and the nature of the aluminous phase (spinel or garnet) in the mantle source [72]. The HREE (e.g., Yb) is compatible with garnet and has high garnet/melt partition coefficients, whereas MREE (e.g., Sm and Gd) and LREE (e.g., La) are incompatible and have low garnet/melt partition coefficients [57,72]. La/Yb and Sm/Yb are strongly fractionated when melting occurs in the garnet stability field, and on the other hand, in contrast to this, La/Yb is slightly fractionated, and Tb/Yb are nearly unfractionated during melting in the spinel peridotite domain. Low (La/Yb) N ratios (2.19–3.85) and relatively flat HREE patterns [(Tb/Yb) N = 1.29–1.7 and (Dy/Yb) N = 1.27–1.57)] for the Gomer basalts are consistent with a higher degree of partial melting of an E-MORB-type of the mantle in the spinel lherzolite stability field. Moreover, the ratios of Zr/Y and Nb/Y are less sensitive to crustal assimilation and are considered indicators of residual garnet in the mantle source from which basaltic melts are derived and, thus, a proxy for depth [73]. The high field strength element (HFSE) (Zr/Y ≤ 4.6 and Nb/Y ≤ 0.5) ratios of the Gomer basalts are relatively lower and consistent with a spinel peridotite source [73,74].

Figure 16a shows the discrimination between partial melting in the presence of garnet peridotite (deep mantle) vs spinel peridotite (shallower mantle). All the Gomer basalt samples plot below the garnet–spinel boundary, in the spinel peridotite field, with (Tb/Yb)ᴺ ratios ranging from approximately 1.2–1.6. The (La/Dy)ᴺ values vary from ∼1.5 to 3.0, suggesting an enriched mantle source but not one influenced by garnet residue. Figure 16b shows the separation of the melting signatures of spinel vs garnet peridotite. Again, all samples fall in the spinel peridotite field, confirming that the residual garnet phase was not involved during partial melting. The (La/Sm)ᴺ values reflect light REE enrichment, supporting derivation from an enriched spinel lherzolite source, consistent with an E-MORB-type mantle.

Figure 16

(a) (Tb/Dy)N vs (La/Dy)N and b) (Tb/Yb)N vs (La/Sm) N variation plots for Gomer basalts. Chondrite normalization values. The horizontal solid line in (a) and (b) separates the fields of garnet-bearing peridotite and of spinel-bearing peridotite (from Wang et al. [76]), respectively. In both diagrams, all samples from the Gomer area are plotted in spinel peridotite fields.

4.4 Mantle source characteristics

The patterns on primitive mantle-normalized multi-element and chondrite-normalized REE plots (for Gomer basalts are more likely consistent with a contribution from E-MORB [58]. All samples are characterized by their enrichments in highly and moderately compatible trace element abundances (e.g., LREE abundances), ranging from a few tens of the chondritic abundances to hundreds is challenging to reconcile strictly with a depleted mid-ocean ridge basalt (MORB) mantle source. These patterns are usually thought to be typical of magmas generated by the partial melting of the asthenospheric upper mantle [75].

The multi-element spider diagram of the Gomer basalts shows a typical E-MORB signature, with moderate LILE enrichments and distinct negative anomalies at Nb and Ta, alongside positive spikes at Pb and U. These patterns are characteristic of CFBs and are often attributed to either subduction-modified mantle sources or crustal contamination [77]. Comparisons with other global LIP systems reveal similar features; for example, Afar flood basalts display comparable Nb–Ta depletions and Pb enrichments associated with plume-lithosphere interaction [59], while Deccan Traps basalts show these signatures due to significant crustal input (Mahoney et al. 2000). In the Gomer basalts, the La/Ybₙ range of 2.19–3.85, with elevated Ba, U, and Pb and suppressed Nb and Ta, underscores the role of enriched mantle source melting coupled with minor crustal assimilation. These geochemical characteristics emphasize mantle heterogeneity beneath the northwestern Ethiopian plateau, indicating a composite source involving an E-MORB-like mantle influenced by lithospheric input and early rift-related magmatism. The enrichment characteristics in the incompatible trace elements observed in MORB (E-MORB) may originate in material entrained from the lower mantle [78], or as plumes or hotspots rising from below the asthenosphere [79], or from subducted materials distributed in a streaky fashion throughout the upper mantle [79].

Furthermore, the ratios of the highly incompatible trace element having similar incompatibility (e.g., Zr/Nb, Nb/Th, Ba/La, Ba/Nb, Ba/Th, Rb/Nb, Th/Yb, Zr/Y, and La/Nb) are often thought to be more directly representatives of source compositions and can indicate the mantle source heterogeneity as they are the least susceptible to fractionate during partial melting and not significantly changed by limited degrees of low-pressure crystallization of melts [71,80,81]. The Gomer basalts have incompatible trace element ratios (e.g., Zr/Nb = 12.4–19.8, La/Nb = 1.13–1.5) similar to E-MORB [58]. [82,83] have also emphasized the importance of the Zr/Nb ratio as an indicator of the “depleted” or “enriched” nature of the source region for basaltic lavas. The studied samples have lower Zr/Nb (4.9–19.8) ratios than N-MORB (Zr/Nb > 30), suggesting an enriched mantle source. However, plume-type basalts generally have lower Zr/Nb ratios than N-MORB (Zr/Nb > 30). Therefore, the Gomer basalts’ Zr/Nb values (4.9–19.8) exhibit a comparatively lower range, reflecting a plume signature of the magma that generated these lavas. In addition, samples also have low Y/Nb (0.62–4.46) and Ti/Zr (69.52–121.7) ratios and high Zr/Y (3.79–7.96) ratios, which confirm their derivation from enriched mantle regions [83]. In the Zr/Nb vs Nb/Th plot (Figure 17a), the studied samples have a strong tendency to fall between the values of PM and E-MORB mantle array, reflecting the role of PM and E-MORB type mantle in the genesis of parent magma with minor magma crust interactions. In addition, in the Ba/Th vs Rb/Nb and Ba/Nb vs Ba/La plots of Figure 17 (Weaver [84]), the Gomer basalts show high affinity to EM-I with minor involvement of EM-II and PM mantle end members.

Figure 17

(a) Zr/Nb vs Nb/Th, (b) Ba/La vs Ba/Nb, and (c) Ba/Th vs Rb/Nb plot of Gomer basalts relative to the mantle compositional components. The fields for PM, HIMU, EMI, and EMII are based on data from Weaver [84]. The data for PM, N-MORB, E-MORB, and OIB are from Sun and McDonough [58]. The data for the continental, lower, and UCC are from Rudnick and Gao [68]. PM, primitive mantle; HIMU, high (U/Pb) source; EM1 and EM2, enriched mantle sources; N-MORB, normal ocean ridge basalt; E-MORB, enriched ocean ridge basalt; OIB, oceanic island basalt; LCC, UCC, and CC.

The trace element ratios of the Gomer basalts provide significant insights into their mantle source characteristics. The Zr/Nb ratios range from 12.4 to 19.8, and La/Nb ratios from 1.13 to 1.5, consistent with an enriched mantle source, similar to E-MORB compositions. These ratios suggest a mantle source undergoing prior enrichment processes [85]. Additionally, the Ti/Y ratios, which vary between 450 and 700, further support the involvement of an enriched mantle component. We have utilized REE-based discrimination diagrams to further constrain the mantle source’s depth and nature. The (La/Yb)ₙ vs (Tb/Yb)ₙ plot is particularly informative, distinguishing between garnet and spinel peridotite sources. The (Tb/Yb)ₙ ratios in this study’s samples are consistently below 1.8, indicating melting within the spinel peridotite facies.

4.5 Tectonic settings

The HFSE abundances and ratios are considered suitable parameters to discriminate the tectonic environment for the eruption of basaltic magmas. All basalt samples are plotted in the Ti vs Zr diagram [81] in the field of the within-plate basalt (WPB). The plots of these samples also cluster WPBs on the Zr vs Zr/Y [83]. The Gomer basalts are geographically located in the northwestern Ethiopia plateau, have a stratified nature, and are classified as low Ti basalt of the northwestern Ethiopia plateau based on TiO₂ contents and incompatible trace element abundance and ratios, implying the Gomer basalts belong to pre-rift basalts of the Northwestern Ethiopia plateau. These geochemical characteristics – particularly the E-MORB-like trace element signatures and the low TiO₂ content – support derivation from an enriched mantle source associated with early-stage continental rifting in the western Ethiopian Plateau, contributing valuable insights into the geodynamic evolution of the Ethiopian Rift System (Figure 18).

Figure 18

(a) Ti–Zr discrimination diagram (Pearce [86]) shows the Gomer rock samples plotted within-plate lavas field. (b) Ti/Y–Zr discrimination diagram [82] also shows the Gomer rock samples plotted within-plate lavas field.

4.6 Comparison of Gomer basalts with other Ethiopian basaltic provinces: Northwestern Ethiopian plateau basalts, Afar basalts, and Main Ethiopian Rift basalts

While geochemical discrimination diagrams (e.g., Ti–Zr, Ti/Y-Zr) plot the Gomer basalts within the WPB field, this classification must be contextualized within the broader tectonic setting. The location of the Gomer region on the western margin of the Ethiopian plateau – along the evolving East African Rift System – suggests a tectonomagmatic environment shaped by early-stage continental rifting. The E-MORB-like trace element signatures, low (La/Yb)ᴺ ratios, and spinel-peridotite-derived melting patterns are consistent with partial melting due to asthenospheric upwelling and decompression in an extending lithosphere. These features point to a transitional setting influenced by plume–lithosphere interaction and rift initiation rather than mature rift magmatism.

Incompatible trace element data from previous works, such as for Northwestern Ethiopian plateau Oligocene basalts [6,7], MER basalts [31], and Afar rift basalts [59], have been plotted on spider and REEs diagrams together with the primary data of this study to compare and determine possible petrogenetic relationships among them. Incompatible trace element data with low LOI and higher MgO (>6 wt%) values were selected to plot.

The primitive mantle normalized multi-element patterns for Gomer basalts are similar to the LT NW Ethiopian plateau basalts, with a typical peak at Ba, K, Sr, Pb, and Ti and troughs at Th and Nb. The chondrite-normalized REE patterns are similar to the LT NW Ethiopian plateau basalts, except for the LREE depletion in the LT Oligocene plateau basalts. However, the Gomer basalts are more enriched in the most incompatible elements than the LT basalts. The TiO₂ contents and the Ti/Y and Nb/Y ratios for Gomer basalts are typical of the LT Oligocene plateau basalts [6]. Although the Gomer basalts have many similar incompatible element characteristics with the LT plateau basalts, some other consistent differences are also found between patterns of those lavas that require distinct histories of the petrogenetic process.

On the other hand, the Gomer basalts and MER basalts show similar patterns on REEs and spider diagrams with different enrichments in incompatible elements and have typical peaks at Ba, Pb, Sr, and Ti. However, both basalts show opposed anomalies at Nb and U. Nb shows negative anomalies, whereas U is positive in Gomer Basalts, as opposed to the MER basalts. From this, we can conclude that both magmas are from enriched mantles with different petrogenetic processes acting on them. The Gomer basalts also show similar enrichments in incompatible elements on REEs and spider diagrams with no more common peaks and troughs, except both show peaks at U. Although both magma suites have similar enrichments in incompatible elements, they show their peaks and troughs differently, thus implying that they are derived from enriched mantle sources, with no common history of petrogenesis.