Meso-Neoproterozoic Maﬁc Sills along the South-Eastern Margin of the Siberian Craton, SE Yakutia: Petrogenesis, Tectonic and Geochemical Features

: We report major and trace element concentrations, along with Nd isotope compositions, for Late Mesoproterozoic to Early Neoproterozoic dolerite sills from the Sette-Daban ridge (southern Verkhoyansk, south-east Siberia). Based on their major element composition, all rocks correspond to low-Ti ( < 3 wt% TiO 2 ) moderately alkaline basalts. The intrusions can be subdivided into two groups based on their trace element compositions. One group includes sills mainly distributed in the southern part of the study area (Yudoma group), with mid-ocean ridge basalt (MORB) trace element patterns enriched in aqueous ﬂuid mobile incompatible (FMI) elements (Sr, Pb, Ba, U). The second group includes sills mostly distributed in the northern part of the study area, enriched in immobile incompatible (II) elements (Th, Nb, light rare earth elements (LREE)) and to a lesser extent, in aqueous ﬂuid mobile elements. The Nd isotope signatures of the dolerites characterize a depleted mantle source, with a small enrichment from recycled continental crust. The geochemical characteristics of these igneous rocks are analogous to low-Ti basalts of large intraplate provinces (e.g., the Karoo and Siberian Traps). We propose that they formed by rifting-induced melting of the heterogeneous metasomatized shallow spinel-bearing mantle zone. We suggest that two di ﬀ erent melting sources were involved in the generation of the two geochemically distinct sill groups, including the addition of two di ﬀ erent subduction components. The southern sills were formed by melting of depleted lithospheric mantle enriched with FMI elements, corresponding to subduction-induced metasomatic alteration by ﬂuids at shallow depths. The northern dolerites were formed by melting of depleted lithospheric mantle enriched with II elements, associated with the melting of subducted sediments at deeper depths. Craton The latest Mesoproterozoic to early Neoproterozoic mafic magmatic event, commonly known as the Sette-Daban event [8], was dated by U-Pb baddeleyite and Sm-Nd isochron methods [6,10,11], but very limited information has been published on the geochemical and isotopic compositions of the associated igneous rocks [6,9]. main goal of study is to present new data on the geochemical and isotopic compositions of the latest Mesoproterozoic to early Neoproterozoic mafic intrusions and shed new light on their tectonic setting.


Introduction
The sedimentary succession deposited along the south-eastern margin of the Siberian Craton ( Figure 1) includes carbonate and clastic rocks ranging in age from Mesoproterozoic to Mesozoic, and records the long-term evolution from an intracratonic basin to a passive margin, and finally to a fold-and-thrust belt [1][2][3][4]. Several mafic magmatic events accompanied the pre-Mesozoic evolution Neoproterozoic, early Paleozoic, and Devonian events [5][6][7]. Paleozoic mafic magmatic events have been described and discussed in many papers, and are mainly reported from the eastern margin of the Siberian Craton [2,5,6,8,9]. The latest Mesoproterozoic to early Neoproterozoic mafic magmatic event, commonly known as the Sette-Daban event [8], was dated by U-Pb baddeleyite and Sm-Nd isochron methods [6,10,11], but very limited information has been published on the geochemical and isotopic compositions of the associated igneous rocks [6,9]. The main goal of this study is to present new data on the geochemical and isotopic compositions of the latest Mesoproterozoic to early Neoproterozoic mafic intrusions and shed new light on their tectonic setting.  [12], modified), and (e) composite stratigraphic section (after [12][13][14] and observations by authors) of the Meso-and Neoproterozoic succession in the Maya-Kyllakh zone. Sample locations, numbers, and specific colors used in the following diagrams are shown in Figure 1d. U-Pb baddeleyite and Sm-Nd isochron ages in Figure 1e are from [7,10,11].  [12], modified), and (e) composite stratigraphic section (after [12][13][14] and observations by authors) of the Meso-and Neoproterozoic succession in the Maya-Kyllakh zone. Sample locations, numbers, and specific colors used in the following diagrams are shown in Figure 1d. U-Pb baddeleyite and Sm-Nd isochron ages in Figure 1e are from [7,10,11].

Geological Setting and Petrography
The Verkhoyansk fold-and-thrust belt (Verkhoyansk FTB) is located along the eastern margin of the Siberian Platform, extending for 2500 km from the coast of the Laptev Sea in the north to the Sea of Okhotsk in the south. The outer zone of the Verkhoyansk FTB is divided (from north to south) into the Olenek, West Verkhoyansk, and South Verkhoyansk sectors [2]. The South Verkhoyansk sector The sills were intruded parallel to the bedding planes of host rocks and are characterized by chilled margins ranging in thickness from a few millimeters to 10 centimeters. The width of the baked margin zone is dependent on the thickness of the sill and typically ranges up to a few meters, but can also reach 15-20 m where dolerite sills are located in close proximity to each other. In the largest sills, the grain size increases from fine-grained to coarse-grained towards the upper contact of the intrusion [14].
A variety of textures are recognized in thin section (Figure 3), depending on intrusion size and position within the intrusion, including intergranular, poikilophitic, ophitic, doleritic, microdoleritic, and micrographic textures. The sills were intruded parallel to the bedding planes of host rocks and are characterized by chilled margins ranging in thickness from a few millimeters to 10 centimeters. The width of the baked margin zone is dependent on the thickness of the sill and typically ranges up to a few meters, but can also reach 15-20 m where dolerite sills are located in close proximity to each other. In the largest sills, the grain size increases from fine-grained to coarse-grained towards the upper contact of the intrusion [14].
A variety of textures are recognized in thin section (Figure 3), depending on intrusion size and position within the intrusion, including intergranular, poikilophitic, ophitic, doleritic, microdoleritic, and micrographic textures. Hydrothermal alteration can be recognized in most samples but the degree of alteration is variable. In approximately half of the samples, the alteration processes are not widely distributed and less than 10% of the rock-forming minerals are altered to secondary minerals. Plagioclase is locally replaced by sericite, albite, and epidote, whereas clinopyroxene is replaced by chlorite and amphibole (uralite). Plagioclase is typically more altered than clinopyroxene. Hydrothermal alteration can be recognized in most samples but the degree of alteration is variable. In approximately half of the samples, the alteration processes are not widely distributed and less than 10% of the rock-forming minerals are altered to secondary minerals. Plagioclase is locally replaced by sericite, albite, and epidote, whereas clinopyroxene is replaced by chlorite and amphibole (uralite). Plagioclase is typically more altered than clinopyroxene.

Whole-Rock Major and Trace Element Measurements
Whole-rock geochemical analyses were undertaken at the All Russian Geological Research Institute (VSEGEI) in St. Petersburg, Russia. Major element concentrations were determined by XRF using an ARL 9800 spectrometer, with trace and rare earth element (REE) compositions determined by inductively coupled plasma-mass spectrometry (ICP-MS) using an ELAN 6100 DRC ICP-MS. Measured concentrations were significantly above detection limits, with analytical uncertainties of <3% for major elements and 4-10% for trace elements and REE, except for Ni which had an uncertainty of about 15%. The international standards OU-6 and SBC-1 were used in this study. Concentrations of the major elements, trace elements, and REE, with detection limits for the studied samples, are given in Appendix A.

Whole-Rock Sm-Nd Isotope Measurements
The 147 Sm/ 144 Nd ratios and Nd isotope compositions were analyzed on a ThermoFisher Triton multi-collector TIMS at the Institute of Precambrian Geology and Geochronology (St. Petersburg), using the method described in [19]. The measured 143 Nd/ 144 Nd ratios were normalized to 146 Nd/ 144 Nd = 0.7219 and adjusted to the La Jolla Nd standard ( 143 Nd/ 144 Nd = 0.511860). Total blank measurements during the analyses were 0.03-0.2 ng for Sm and 0.1-0.5 ng for Nd. Analytical error (2σ) was ±0.5% for the Sm and Nd concentrations, ±0.5% for the 147 Sm/ 144 Nd ratios, and ±0.005% for the 143 Nd/ 144 Nd ratios. ε Nd (t) was calculated using the present-day ratios 143 Nd/ 144 Nd = 0.512638 and 147 Sm/ 144 Nd = 0.1967 in the chondritic uniform reservoir (CHUR) [20].
Sm-Nd isotope analyses for samples AS18-21 and AS18-22 were performed on a Finnigan MAT 262 mass spectrometer at the Geological Institute of the Kola Science Center RAS (Apatity), using the method described in [21]. The measured 143 Nd/ 144 Nd ratios were normalized to 146 Nd/ 144 Nd = 0.7219 and adjusted to the La Jolla Nd standard ( 143 Nd/ 144 Nd = 0.511860). Total blank measurements during the analyses were 0.06 ng for Sm and 0.3 ng for Nd. Analytical error (2σ) was ±0.5% for the Sm and Nd concentrations, ±0.5% for the 147 Sm/ 144 Nd ratios, and ±0.005% for the 143 Nd/ 144 Nd ratios. ε Nd (t) was calculated using the present-day ratios 143 Nd/ 144 Nd = 0.512638 and 147 Sm/ 144 Nd = 0.1967 in the CHUR [20]. The Sm-Nd isotope results are shown in Table 1.

Major Elements
The intrusive igneous rocks associated with the Sette-Daban magmatic event display small variations in SiO 2 (46.76-51.99 wt%) and K 2 O + Na 2 O (2.1-4.8 wt%), but larger variations in MgO (1.85-9.16 wt%), plotting in the moderately alkaline basalt field on the total alkali-silica (TAS) diagram ( Figure 4a). All dolerites plot in the tholeiitic basalt field on the alkali-FeO tot -MgO diagram (Figure 4b), lining up along the FeO-MgO axis. The Sakhara dolerites are characterized by the strongest Fe tholeiite signature, while the Yudoma dolerites are more calc-alkaline. The Mg# increases from the Sakhara (0.13-0.38) to the Yudoma (0.41-0.51) samples, indicating the latter to be closer to the composition of primitive mantle melts. Dolerites from both the western and eastern Allakh-Yun river do not plot as two distinct groups, with their compositions overlapping those of the Sakhara and Yudoma samples.

Major Elements
The intrusive igneous rocks associated with the Sette-Daban magmatic event display small variations in SiO2 (46.76-51.99 wt%) and K2O + Na2O (2.1-4.8 wt%), but larger variations in MgO (1.85-9.16 wt%), plotting in the moderately alkaline basalt field on the total alkali-silica (TAS) diagram ( Figure 4a). All dolerites plot in the tholeiitic basalt field on the alkali-FeOtot-MgO diagram (Figure 4b), lining up along the FeO-MgO axis. The Sakhara dolerites are characterized by the strongest Fe tholeiite signature, while the Yudoma dolerites are more calc-alkaline. The Mg# increases from the Sakhara (0.13-0.38) to the Yudoma (0.41-0.51) samples, indicating the latter to be closer to the composition of primitive mantle melts. Dolerites from both the western and eastern Allakh-Yun river do not plot as two distinct groups, with their compositions overlapping those of the Sakhara and Yudoma samples.

REE and Other Incompatible Elements
Major element compositions in a residual melt evolve as the crystallization of minerals such as plagioclase, olivine, or pyroxene takes place, hampering an interpretation of the original source of the melt based on major element compositions alone. However, the concentrations of highly incompatible large ion lithophile elements (LILE) and high field strength elements (HFSE), including rare earth elements (REE), are much less affected by fractional crystallization. Moreover, ratios of highly incompatible LILE are unaffected by fractional crystallization, with ratios primarily controlled by the degree of partial melting [24]. Therefore, in order to constrain the partial melting process of the mantle source, we focus on ratios of highly incompatible LILE in the discussion below. Samples

REE and Other Incompatible Elements
Major element compositions in a residual melt evolve as the crystallization of minerals such as plagioclase, olivine, or pyroxene takes place, hampering an interpretation of the original source of the melt based on major element compositions alone. However, the concentrations of highly incompatible large ion lithophile elements (LILE) and high field strength elements (HFSE), including rare earth elements (REE), are much less affected by fractional crystallization. Moreover, ratios of highly incompatible LILE are unaffected by fractional crystallization, with ratios primarily controlled by the degree of partial melting [24]. Therefore, in order to constrain the partial melting process of the mantle source, we focus on ratios of highly incompatible LILE in the discussion below.
Primitive mantle-(PM-) normalized trace element patterns (Figure 6b) of the Sakhara and eastern Allakh-Yun dolerites exhibit some variations in LILE and a generally pronounced enrichment in LREE, a wide range in Th and U, weak negative Ta-Nb anomaly, negative Sr, and positive Pb anomaly. The PM-normalized trace element patterns for the Sakhara and eastern Allakh-Yun dolerites demonstrate a closer similarity to island arc basalts (IAB) [24] than to ocean island basalts (OIB) or mid-ocean ridge basalts (MORB). The negative Sr and weak negative Eu anomalies in those dolerites can be explained by the buffer effect of crystallizing plagioclase. The positive Pb anomaly is related to recycled sediment in the melting mantle source [25].
The Yudoma and western Allakh-Yun dolerites display flat REE patterns and lack a negative Eu anomaly, similar to the composition of normal mid-ocean ridge basalt (N-MORB) (Figure 6a). The lower (La/Yb) N ratio (0.73-1.18) in the Yudoma and western Allakh-Yun dolerites, compared to the Sakhara and eastern Allakh-Yun dolerites, may be related to differences in the degree of partial melting of the mantle source. enrichment in LREE, a wide range in Th and U, weak negative Ta-Nb anomaly, negative Sr, and positive Pb anomaly. The PM-normalized trace element patterns for the Sakhara and eastern Allakh-Yun dolerites demonstrate a closer similarity to island arc basalts (IAB) [24] than to ocean island basalts (OIB) or mid-ocean ridge basalts (MORB). The negative Sr and weak negative Eu anomalies in those dolerites can be explained by the buffer effect of crystallizing plagioclase. The positive Pb anomaly is related to recycled sediment in the melting mantle source [25]. The Yudoma and western Allakh-Yun dolerites display flat REE patterns and lack a negative Eu anomaly, similar to the composition of normal mid-ocean ridge basalt (N-MORB) (Figure 6a). The lower (La/Yb)N ratio (0.73-1.18) in the Yudoma and western Allakh-Yun dolerites, compared to the Sakhara and eastern Allakh-Yun dolerites, may be related to differences in the degree of partial melting of the mantle source.
The dolerites from Yudoma and western Allakh-Yun are characterized by large ranges in concentrations of most incompatible elements. With the exception of two samples, all dolerites from these areas more closely resemble N-MORB in relation to immobile (IM) incompatible trace element concentrations (Th, Nb). However, at the same time, they are characterized by a very high enrichment in high-fluid-mobility (HFM) incompatible trace elements (Ba, U, Sr, Pb) [28,29], which could reflect a shallow subducted sediment component [24,28]. The dolerites from Yudoma and western Allakh-Yun are characterized by large ranges in concentrations of most incompatible elements. With the exception of two samples, all dolerites from these areas more closely resemble N-MORB in relation to immobile (IM) incompatible trace element concentrations (Th, Nb). However, at the same time, they are characterized by a very high enrichment in high-fluid-mobility (HFM) incompatible trace elements (Ba, U, Sr, Pb) [28,29], which could reflect a shallow subducted sediment component [24,28].

Nd Isotopes
All four dolerite groups do not show significant variations in Nd isotope compositions (Figure 7). ε Nd (t) values range from 3.6 to 7.5, indicating a depleted MORB source with a minor contribution from an enriched source. The Nd isotope compositions of most samples do not exhibit a clear correlation with SiO 2 contents (Figure 7b), implying that direct crustal contamination, as suggested by the range of Nd isotope values, is very unlikely. However, samples from the Yudoma river group demonstrate a weak negative correlation between ε Nd (t) and SiO 2 . Therefore, we cannot completely exclude the possibility of a minor amount of direct crustal contamination. Nevertheless, the small variation in ε Nd (t), as well as in the chemical composition of the dolerites, allow us to suggest that the effect of possible crustal contamination on the primitive geochemical pattern of the dolerites is likely to be insignificant. demonstrate a weak negative correlation between εNd(t) and SiO2. Therefore, we cannot completely exclude the possibility of a minor amount of direct crustal contamination. Nevertheless, the small variation in εNd(t), as well as in the chemical composition of the dolerites, allow us to suggest that the effect of possible crustal contamination on the primitive geochemical pattern of the dolerites is likely to be insignificant.

LILE Variations in Contributions from Aqueous Fluids
Alteration of the mafic rocks is minor, and when optically discernible, appears to mostly affect plagioclase, forming sericite, albite, and epidote. Such alteration is too small to affect the chemical composition of the rocks. Nevertheless, it could affect the primordial trace element abundance in the studied dolerites, especially fluid mobile trace elements such as Rb, Sr, Ba, and U. We used the loss on ignition (LOI) value to test the effect of post-crystallization alteration and weathering on the primordial content of highly mobile elements in the dolerites (Figure 8). The absence of a correlation between mobile elements and LOI indicates that no alteration occurred after crystallization, therefore the variation in Ba-Sr-U could be related to metasomatic alteration of the melting source. At the same time, several samples (e.g., ##AS18-21, AS18-22, АС17-16*) contain anomalously high concentrations of Sr and Ba, along with high LOI values. In addition, significant enrichments in Ba, Sr and U in those samples are not accompanied by enrichments in other highly incompatible elements (such as Rb, K, Th, U, La, Ce). Consequently, the geochemical peculiarities of these samples could have been generated by secondary alteration processes. Therefore, samples with an anomalously high enrichment in Ba, Sr, and U should be excluded from the discussion concerning HFM incompatible trace elements. The other samples demonstrate no obvious correlation of LILE concentration with LOI value and can be assumed to have undergone no post-crystallization alteration.
The enrichment in, and wide variation of, HFM elements in the observed dolerites could be explained as a result of the fertilization of lithospheric mantle by aqueous fluids at shallow depths prior to melting [24,30]. The ratios of mobile-to-immobile incompatible trace elements which have similar bulk partitioning coefficients (e.g., Ba/Th, Sr/Th) do not considerably fractionate during mantle melting and following fractional crystallization [31]. Therefore, any variation in their ratios can be interpreted to reflect geochemical heterogeneity of the mantle. On the other hand, LILE are

LILE Variations in Contributions from Aqueous Fluids
Alteration of the mafic rocks is minor, and when optically discernible, appears to mostly affect plagioclase, forming sericite, albite, and epidote. Such alteration is too small to affect the chemical composition of the rocks. Nevertheless, it could affect the primordial trace element abundance in the studied dolerites, especially fluid mobile trace elements such as Rb, Sr, Ba, and U. We used the loss on ignition (LOI) value to test the effect of post-crystallization alteration and weathering on the primordial content of highly mobile elements in the dolerites (Figure 8). The absence of a correlation between mobile elements and LOI indicates that no alteration occurred after crystallization, therefore the variation in Ba-Sr-U could be related to metasomatic alteration of the melting source. At the same time, several samples (e.g., ##AS18-21, AS18-22, AC17-16*) contain anomalously high concentrations of Sr and Ba, along with high LOI values. In addition, significant enrichments in Ba, Sr and U in those samples are not accompanied by enrichments in other highly incompatible elements (such as Rb, K, Th, U, La, Ce). Consequently, the geochemical peculiarities of these samples could have been generated by secondary alteration processes. Therefore, samples with an anomalously high enrichment in Ba, Sr, and U should be excluded from the discussion concerning HFM incompatible trace elements. The other samples demonstrate no obvious correlation of LILE concentration with LOI value and can be assumed to have undergone no post-crystallization alteration.
The enrichment in, and wide variation of, HFM elements in the observed dolerites could be explained as a result of the fertilization of lithospheric mantle by aqueous fluids at shallow depths prior to melting [24,30]. The ratios of mobile-to-immobile incompatible trace elements which have similar bulk partitioning coefficients (e.g., Ba/Th, Sr/Th) do not considerably fractionate during mantle melting and following fractional crystallization [31]. Therefore, any variation in their ratios can be interpreted to reflect geochemical heterogeneity of the mantle. On the other hand, LILE are soluble in aqueous fluids and any variations in Ba/Th and Sr/Th ratios are likely to reflect the secondary impact of aqueous fluids (metasomatism) on the composition of the mantle source [32,33]. On the plots of Ba/Th and Sr/Th versus Th/Nd (Figure 9), the analyzed dolerites form similar negatively correlated trends, indicating the importance of aqueous fluids for the generation of basaltic magma. The Yudoma and some of the western Allakh-Yun dolerites are characterized by higher Ba/Th and Sr/Th compared to those from the Sakhara and eastern Allakh-Yun, suggesting that they may be associated with larger contributions from aqueous fluids in the melting source. We suggest that the variations in contributions from aqueous fluids are associated with different depth levels of subduction-related alteration of the lithospheric mantle, including shallow depths for the source of the Yudoma dolerites, and deeper levels for the source of the Allakh-Yun and Sakhara dolerites [24]. basaltic magma. The Yudoma and some of the western Allakh-Yun dolerites are characterized by higher Ba/Th and Sr/Th compared to those from the Sakhara and eastern Allakh-Yun, suggesting that they may be associated with larger contributions from aqueous fluids in the melting source. We suggest that the variations in contributions from aqueous fluids are associated with different depth levels of subduction-related alteration of the lithospheric mantle, including shallow depths for the source of the Yudoma dolerites, and deeper levels for the source of the Allakh-Yun and Sakhara dolerites [24].

Immobile Incompatible Trace Element Variations and the Mantle Source
On the Th/Yb-Nb/Yb plot, the Sette-Daban sills can be subdivided into two groups based on distinct geochemical compositions (Figure 10a). The first group plots within the mantle MORB-OIB array, mostly comprising dolerites from the Yudoma and western Allakh-Yun rivers, but also a few samples from Sakhara and eastern Allakh-Yun. Variations within the mantle array on the Th/Yb-Nb/Yb plot reflect differences in the degree and depth of melting, whereby the Sakhara sample is the most differentiated and the Yudoma samples are the most primitive. The second group of samples mostly comprises dolerites from the Sakhara and eastern Allakh-Yun rivers, along with a few samples from Yudoma and western Allakh-Yun, and are shifted out of the mantle array towards the volcanic arc array. This contrasting geochemical composition suggests that the mantle source of the second dolerite group experienced crustal contamination or subduction-related modification [25]. The second group also plots within the volcanic arc field on the Nb/Th-Zr/Nb plot (Figure 10b). The western and eastern Allakh-Yun dolerites appear to encompass the range of compositions

Immobile Incompatible Trace Element Variations and the Mantle Source
On the Th/Yb-Nb/Yb plot, the Sette-Daban sills can be subdivided into two groups based on distinct geochemical compositions (Figure 10a). The first group plots within the mantle MORB-OIB array, mostly comprising dolerites from the Yudoma and western Allakh-Yun rivers, but also a few samples from Sakhara and eastern Allakh-Yun. Variations within the mantle array on the Th/Yb-Nb/Yb plot reflect differences in the degree and depth of melting, whereby the Sakhara sample is the most differentiated and the Yudoma samples are the most primitive. The second group of samples mostly comprises dolerites from the Sakhara and eastern Allakh-Yun rivers, along with a few samples from Yudoma and western Allakh-Yun, and are shifted out of the mantle array towards the volcanic arc array. This contrasting geochemical composition suggests that the mantle source of the second dolerite group experienced crustal contamination or subduction-related modification [25]. The second group also plots within the volcanic arc field on the Nb/Th-Zr/Nb plot (Figure 10b). The western and eastern Allakh-Yun dolerites appear to encompass the range of compositions defined by the Sakhara and Yudoma sills, but the eastern sills more closely resemble Sakhara signatures, whereas the western sills more closely resemble Yudoma signatures. It is difficult to explain the observed variations in ratios of immobile incompatible trace elements (especially Nb/Th) by partial melting, but these variations could be explained by the addition of incompatible trace elements to the lithospheric mantle at deep depths during subduction, analogous to the modern Mariana Trough [24]. Another key difference between the two groups of dolerites is their contrasting REE patterns ( Figure 6), best illustrated by their LREE/MREE (middle REE) fractionation systematics. Fractionation of LREE/MREE and MREE/HREE allows us to understand the melting regime in the mantle [37]. All of the dolerite samples demonstrate obvious fractionation between the LREE/MREE ratios, while fractionation between the MREE/HREE ratios is not observed (Figure 6a). The La/Sm and Sm/Yb ratios do not correlate with Mg-number and their variations cannot be explained by fractional crystallization or crustal contamination processes [37].
The diagram in Figure 11 displays model trends of peridotite melting corresponding to different physical parameters and representing compositions of aggregated melts. OIB compositions form at high pressures and high potential temperatures, characteristic of a mantle plume, while MORB compositions reflect melting of the mantle at relatively shallow depths and lower temperatures. The Sette-Daban dolerites are characterized by an almost flat Sm/Yb trend, corresponding to melting at relatively low pressures, with increasing temperatures from the eastern Allakh-Yun and Sakhara dolerites to the Yudoma and western Allakh-Yun dolerites. All samples plot well within the MORB field, which is consistent with their low alkali and low Ti compositions, which are noticeably distinct from OIB compositions on the Th/Yb-Nb/Yb plot (Figure 10a). The Sm/Yb and La/Sm ratios also exclude the presence of garnet in their melt source, suggesting melting of a shallower spinel-bearing source [37] with different degrees of melting or subduction-related modification, as inferred from the Th/Yb-Nb/Yb diagram (Figure 10a).  [34,35]. The composition of the Siberian Traps (dark gray field) is taken from [36], while the composition of the Karoo Province is taken from [37]. (b) Zr/Nb-Nb/Th plot [38]. Another key difference between the two groups of dolerites is their contrasting REE patterns (Figure 6), best illustrated by their LREE/MREE (middle REE) fractionation systematics. Fractionation of LREE/MREE and MREE/HREE allows us to understand the melting regime in the mantle [37]. All of the dolerite samples demonstrate obvious fractionation between the LREE/MREE ratios, while fractionation between the MREE/HREE ratios is not observed (Figure 6a). The La/Sm and Sm/Yb ratios do not correlate with Mg-number and their variations cannot be explained by fractional crystallization or crustal contamination processes [37].
The diagram in Figure 11 displays model trends of peridotite melting corresponding to different physical parameters and representing compositions of aggregated melts. OIB compositions form at high pressures and high potential temperatures, characteristic of a mantle plume, while MORB compositions reflect melting of the mantle at relatively shallow depths and lower temperatures. The Sette-Daban dolerites are characterized by an almost flat Sm/Yb trend, corresponding to melting at relatively low pressures, with increasing temperatures from the eastern Allakh-Yun and Sakhara dolerites to the Yudoma and western Allakh-Yun dolerites. All samples plot well within the MORB field, which is consistent with their low alkali and low Ti compositions, which are noticeably distinct from OIB compositions on the Th/Yb-Nb/Yb plot (Figure 10a). The Sm/Yb and La/Sm ratios also exclude the presence of garnet in their melt source, suggesting melting of a shallower spinel-bearing source [37] with different degrees of melting or subduction-related modification, as inferred from the Th/Yb-Nb/Yb diagram (Figure 10a).  [41], using mineral-melt distribution coefficients for rare elements from [42]. The evolutionary trends calculated for the melting of peridotite correspond to DM composition. The composition of depleted mantle (yellow star) is taken from [43]. The composition of primitive mantle (green star) is taken from [44]. BCC (violet star) represents the composition of bulk continental crust [45]. The dark gray field represents the compositional range of MORB basalts, while the pale grey field represents the compositional range of OIB basalts [46].
The studied samples can be geographically separated into the "northern" (Sakhara) and "southern" (Yudoma) groups based on their incompatible trace element ratios (Th/Yb, Nb/Yb, Nb/Th, La/Sm; Figures 10 and 11). A transitional zone between the two groups includes the sublatitudinal profile of the Allakh-Yun river, encompassing geochemical features of both the Sakhara and Yudoma areas. The geochemical differences between the "southern" and "northern" dolerite patterns are also apparent in the mobile LILE concentrations (Ba, Sr, U, Pb) (Figure 9), suggesting lateral heterogeneity of the lithospheric mantle/melting source. By integrating all of the geochemical characteristics, we suggest that the Yudoma (southern) mantle source composition was modified by metasomatic enrichment induced by hydrous fluid alteration and expressed in fluid mobile LILE, possibly reflecting ancient subduction at shallow depths [24]. On the other hand, the Sakhara (northern) mantle source is enriched in immobile incompatible elements (Th, Nb and LREE) induced by the melting of recycled sediments, possibly reflecting alteration by ancient subduction at deeper levels [24]. The variation in initial Nd isotope signatures of the studied dolerites can be explained by subduction-related alteration of their mantle source.

Tectonic Settings of Sette-Daban Magmatism
The dolerite sills of the Sette-Daban magmatic event are distributed along the south-eastern margin of the Siberian Craton and were intruded into Meso-Neoproterozoic sedimentary cover of the Siberian passive margin, based on ages of around 975-1000 Ma determined from U-Pb dating of baddeleyite [1-4,6,10]. During this time, Siberia formed part of a supercontinent, along with Laurentia and Baltica [47][48][49]. The Sette-Daban region was located far from any lithospheric plate Figure 11. La/Sm-Sm/Yb plot with P-T conditions of melting. Solid grey lines represent segregated melt compositions formed during the melting of mantle peridotite at various pressures (gray rhombs connected by black lines), and correspond to different potential temperatures. The model composition of segregated peridotite melts was calculated in the AlfaMelts program code [41], using mineral-melt distribution coefficients for rare elements from [42]. The evolutionary trends calculated for the melting of peridotite correspond to DM composition. The composition of depleted mantle (yellow star) is taken from [43]. The composition of primitive mantle (green star) is taken from [44]. BCC (violet star) represents the composition of bulk continental crust [45]. The dark gray field represents the compositional range of MORB basalts, while the pale grey field represents the compositional range of OIB basalts [46].
The studied samples can be geographically separated into the "northern" (Sakhara) and "southern" (Yudoma) groups based on their incompatible trace element ratios (Th/Yb, Nb/Yb, Nb/Th, La/Sm; Figures 10 and 11). A transitional zone between the two groups includes the sub-latitudinal profile of the Allakh-Yun river, encompassing geochemical features of both the Sakhara and Yudoma areas. The geochemical differences between the "southern" and "northern" dolerite patterns are also apparent in the mobile LILE concentrations (Ba, Sr, U, Pb) (Figure 9), suggesting lateral heterogeneity of the lithospheric mantle/melting source. By integrating all of the geochemical characteristics, we suggest that the Yudoma (southern) mantle source composition was modified by metasomatic enrichment induced by hydrous fluid alteration and expressed in fluid mobile LILE, possibly reflecting ancient subduction at shallow depths [24]. On the other hand, the Sakhara (northern) mantle source is enriched in immobile incompatible elements (Th, Nb and LREE) induced by the melting of recycled sediments, possibly reflecting alteration by ancient subduction at deeper levels [24]. The variation in initial Nd isotope signatures of the studied dolerites can be explained by subduction-related alteration of their mantle source.

Tectonic Settings of Sette-Daban Magmatism
The dolerite sills of the Sette-Daban magmatic event are distributed along the south-eastern margin of the Siberian Craton and were intruded into Meso-Neoproterozoic sedimentary cover of the Siberian passive margin, based on ages of around 975-1000 Ma determined from U-Pb dating of baddeleyite [1][2][3][4]6,10]. During this time, Siberia formed part of a supercontinent, along with Laurentia and Baltica [47][48][49]. The Sette-Daban region was located far from any lithospheric plate boundaries [3], and its magmatic activity can be considered as intraplate volcanism induced by continental rifting. Despite the paleo-tectonic position of the studied intrusions, trace element patterns of the dolerites have common features to those characteristic of island arc settings (IAB). IAB-like geochemical features could be generated by a post-orogenic mafic magmatic event or intraplate LIP. Post-orogenic rifting and delamination is often followed by the development of mafic magmatism bearing IAB-like geochemical features [50]. Based on numerous paleogeographic reconstructions, Siberia was connected to Laurentia at 1000 Ma when the large Grenvillian orogenic event occurred [10,11,47,49]. Nonetheless, post-orogenic mafic magmatic rocks are typically characterized by high alkalinity (presumably potassic), along with high Ti and LILE concentrations (e.g., [51]), far exceeding those of the Sette-Daban dolerites.
On the other hand, trace element patterns of the Sette-Daban dolerites resemble those of low-Ti basic lavas and dikes of intraplate large igneous provinces (LIP), such as the Early Jurassic Karoo Province and Permian-Triassic Siberian Traps flood basalts [36,37] (Figure 10a). A number of models have been proposed for the origins of LIP ( [52] and references therein), with at least two models proposed for the origins of the Karoo and Siberian provinces. The low-Ti basalts of the Karoo province are considered to have formed as a result of plume-lithosphere interaction. The low Ti content and IAB-like trace element pattern could be explained by the geochemical peculiarity of the sub-continental lithospheric mantle, acquired prior to the development of an intracontinental tectonic setting [37]. Thus, the metasomatic enrichment of sub-continental lithospheric mantle could be associated with ancient subduction. This enriched mantle was apparently preserved beneath the craton until continental rifting initiated melting during the Sette-Daban event.
A more recent model for the Siberian Traps flood basalt magmatism assumes long-term subduction beneath the lithospheric plate on which the intraplate magmatic event occurred. This model invokes mantle melting and subsequent intraplate magmatism, as a result of deep water and fluid recycling due to ongoing crustal subduction [36,52]. Geochemical features, such as negative Ta-Nb anomalies and contamination by incompatible elements (Th, Pb), could be explained by the recycling of terrigenous sediments aboard the subducting slab. Support for the involvement of a previous subduction event in the mantle source of the Sette-Daban magmatism can be acquired from the preceding Grenville active margin, with subsequent continental collision along convergent plate boundaries located across Laurentia, Baltica and Siberia from 1.4 to 1.0 Ga [10,47,53]. On the contrary, the geodynamic context for intraplate mafic magmatism can explain the shallow depth of melting revealed by the La/Sm and Sm/Yb ratios, as well as the flat REE patterns of the studied dolerites. In light of our new data, we suggest that the Sette-Daban intrusions were formed by melting of depleted MORB mantle (DMM), enriched by subduction-related components. To test the various hypotheses proposed above, further geochemical, Sr-Nd-Pb isotope, and geochronological data are urgently required.

Conclusions
Our new geochemical and Nd isotope data for dolerite sills from Sette-Daban (South Verkhoyansk region, eastern Siberia), provide new constraints on the mantle source which formed the latest Mesoproterozoic to early Neoproterozoic mafic intrusions of the Sette-Daban province.
The major and trace element data, along with initial Nd isotope compositions, obtained for the Sette-Daban dolerites allow us to propose the following conclusions: • The Sette-Daban dolerite sills resemble low-Ti lavas of intraplate flood basalt provinces (e.g., Karoo, Siberian Traps) and possess IAB-like trace element patterns.

•
The studied dolerites are the magmatic products of shallow-depth melting of depleted sub-continental lithospheric mantle, due to continental rifting.

•
The trace element compositions of intrusions across the study area reveal striking geochemical differences between northern and southern sills, reflecting lateral heterogeneity in the composition of their mantle source.

•
The source for both northern and southern sills was depleted lithospheric mantle, metasomatically enriched at different depths by ancient subduction.