2026 October LIP of the Month

Ancient sources and complex magma plumbing systems in the Deccan Volcanic Province

Josua J. Pakulla, Carsten Münker

Institute of Geology and Mineralogy, University of Cologne, Germany

Emails: j.j.pakulla@googlemail.com; c.muenker@uni-koeln.de

Extracted and modified after Pakulla et al., 2023 (Chemical Geology, Vol. 640) and Pakulla et al., 2025 (EPSL, Vol. 653). For a more detailed read, including the full discussion about different plumbing systems and mantle source endmembers, we refer to the respective publications.

Abstract

Identifying the sources and plumbing system of the ~66 Ma Deccan Volcanic Province (DVP) has long been debated. Especially, two questions have been widely discussed: 1. What are the relative contributions of the Réunion mantle plume and upper mantle melts?; 2. Does a DVP-wide plumbing system exist that is represented by the current chemostratigraphy? By using state-of-the-art high-precision analyses of ¹⁸²W/¹⁸⁴W and ¹⁴²Nd/¹⁴⁴Nd that record ancient Hadean differentiation events, coupled with high-precision isotope dilution concentration analyses of high-field strength elements (W, Th, U, Nb, Zr, Ta) as well as the long-lived isotope systems Rb-Sr, ¹⁴⁷Sm-¹⁴³Nd, U-Th-Pb, and Lu-Hf, we try to answer these questions.

The DVP chemostratigraphy and plumbing system

The widely accepted 12 chemostratigraphic formations of the DVP along the Western Ghats Escarpment (Fig. 1) were originally identified by variations in trace element and isotope systematics across the stratigraphic succession (Beane et al., 1986; Cox and Hawkesworth, 1985). The chemical variations have been explained by assimilation, fractional crystallisation, and periodic melt recharges (RAFC) within (potentially a single) large magma chamber (e.g., Cox and Hawkesworth, 1985). However, recent studies suggest more complex magma chamber plumbing systems without significant AFC processes, but with increased magma recharge events and mixing of melts from distinct sources in separated and short-lived transcrustal plumbing systems (Hoyer et al., 2023; Pakulla et al., 2023; Basu et al., 2020; Marzoli et al., 2026; Mittal and Richards, 2021).

However, if these plumbing systems are separated and relatively short-lived, why should lava flows of single chemostratigrahpic formations record the same sources and petrogenetic processes across the DVP? In fact, some previous studies pointed out distinct chemostratigraphic orders in the northern DVP (Mahoney et al., 2000; Peng et al., 1998), whereas others question the resolving power of the trace element ratios that the chemostratigraphy is built upon (Kale et al., 2019).

Reinvestigation of literature data, combined with our new data, supports that several elements are not useful for geochemical investigation (i.e., K₂O, Rb, Ba, Sr), likely due to weathering under tropical conditions and auto-hydrothermal processes (Pakulla et al. 2023). Thus, assuming the current difficulties of a simple chemostratigraphy, we reinvestigated the current data as a whole (not assuming different chemostratigraphic formations) using elements that are not significantly affected by fractional crystallisation, melting, or weathering (i.e., Nb, Th, W, Ta, Hf, Sm, Nd, Lu). We find distinct chemical groups for Nb/Th and Th/Ta (and less strongly Ce/Pb) versus MgO, indicating distinct sources that were not altered by AFC processes (Fig. 2). These groups can be independently distinguished by using a k-means clustering approach for MgO, Nb/Th, and Th/Ta (as well as including ⁸⁷Sr/⁸⁶Sr and ²⁰⁶Pb/²⁰⁴Pb in a second run). At least two DVP groups are represented in every chemostratigraphic formation, implying short-lived magmatic pulses and separated plumbing systems, advocating for a revision of the current chemostratigraphy. The different groups 1 to 3 are characterised by different trace element and isotope compositions and require distinct sources (Fig. 2).

DVP sources

Group 1 displays the lowest Nb/Th (Fig. 2) and the most radiogenic ⁸⁷Sr/⁸⁶Sr (Fig. 3) and is strongly associated with the Bushe Fm. These isotope and trace element systematics represent a dominant contribution of lithospheric mantle components and continental crust assimilation to a Réunion-type melt. Still, isotope and trace element mixing calculations indicate a larger lithospheric mantle contribution than assimilation of continental crust (e.g., Pakulla et al., 2023; Marzoli et al., 2026). Group 2’s Nb/Th ranges between Group 1 and Group 3 (Fig. 2). Long-lived isotope compositions of Group 2 appear to derive from a Réunion-type source and extend towards lower εNd and εHf, higher ⁸⁷Sr/⁸⁶Sr and variable ²⁰⁶Pb/²⁰⁴Pb. The ²⁰⁷Pb/²⁰⁴Pb compositions are usually more enriched, similar to Indian carbonatites and alkaline rocks associated with the DVP (Fig. 3). These systematics indicate derivation from a Réunion-type mantle source and addition of compositionally heterogeneous lithospheric mantle components for Group 2 melts. Group 3 samples represent melts least influenced by lithospheric mantle or continental crust and display the highest Nb/Th, εHf, and εNd, and lowest ⁸⁷Sr/⁸⁶Sr, overlapping with La Réunion melts. However, ²⁰⁶Pb/²⁰⁴Pb and ²⁰⁷Pb/²⁰⁴Pb are more unradiogenic, overlapping with Indian MORB, extending to even more unradiogenic compositions. Previously, this was suggested to imply a stronger contribution of an upper depleted mantle source and potentially assimilation by mafic lower continental crust (e.g., Pakulla et al., 2023; Hoyer et al., 2023; Peng et al., 1994; Lightfoot and Hawkesworth, 1988). However, µ¹⁸²W and µ¹⁴²Nd isotope data now draw a new picture.

DVP sources from µ¹⁸²W and µ¹⁴²Nd

Variations in µ¹⁸²W and µ¹⁴²Nd (ppm-level differences of ¹⁴²Nd/¹⁴⁴Nd and ¹⁸²W/¹⁸⁴W relative to a terrestrial standard solution) could only have been produced during the first 60 and 500 Myrs after solar system formation by fractionation of ¹⁸²Hf from ¹⁸²W and ¹⁴⁶Sm from ¹⁴²Nd, respectively. This is due to the relatively short half-lives of ¹⁸²Hf (~9 Myr) and ¹⁴⁶Sm (~95 Myr). Variations of ¹⁸²W and ¹⁴²Nd in Phanerozoic mantle-derived rocks have so far only been observed in ocean island basalts (OIBs) that are related to mantle plumes that were seismically imaged to reach the lowermost mantle and core-mantle boundary (e.g., Mundl et al. 2017; Mundl-Petermeier et al. 2020, Jackson et al. 2020, Peters et al. 2021). These variations were explained by core-mantle interaction (for ¹⁸²W, e.g., Mundl et al. 2017; Rizo et al. 2019) or Hadean silicate differentiation or melting events (Peters et al. 2021; Tusch et al. 2022). They are now also observed for lavas from the DVP and La Réunion (Jansen et al. 2022; Pakulla et al. 2025; Peters et al. 2021)

The OIBs from La Réunion display resolvable µ¹⁸²W and µ¹⁴²Nd variations, clearly showing the preservation of a Hadean reservoir in the deep mantle source of this plume (Peters et al., 2021; Rizo et al., 2019; Jansen et al., 2022). We can expand this finding to the DVP (Pakulla et al., 2025). Group 2 and Group 3 lavas display significant variations in µ¹⁸²W. However, we could not corroborate the µ¹⁴²Nd anomalies for the DVP or La Réunion (Peters et al., 2021). Foremost, this depicts a lower mantle plume origin for the DVP. Further statistical investigation shows that the µ¹⁸²W compositions of the apparently (in Pb isotopes) MORB-like Group 3 show larger µ¹⁸²W variations (-12 ± 2.3 95%CI) than Group 2 lavas (-4.2 ± 3.0 95%CI) that overlap with the La Réunion average (-4.9 ± 1.5 95%CI). Based on the µ¹⁸²W composition of upper mantle-derived lavas (~0) and the lack of µ¹⁸²W variations in basement rocks (Pakulla et al. 2025; Ravindran et al. 2026), admixture of MORB sources or crustal contamination can therefore be excluded and does not fit the µ¹⁸²W isotope composition of Group 3 melts. Consequently, the low ²⁰⁶Pb/²⁰⁴Pb and ²⁰⁷Pb/²⁰⁴Pb samples reflect a deep mantle component that differs from the modern La Réunion plume source. Our preferred model explains the difference in µ¹⁸²W and Pb isotope composition and thus the temporal evolution of the Deccan-La Réunion plume system, by the addition of a recycled Hadean mafic protocrustal restite to the La Réunion plume source. This appears to agree with the isotope and trace element systematics (W/Th, Nb/Th, εNd, εHf, ^206,207,208Pb; Pakulla et al. 2025; Fig. 5, models 2 and 3). Yet other scenarios are also able to explain the data (Fig 5, model 1).

Conclusion

Our recent studies (Pakulla et al. 2023; 2025) support a fast-paced magma plumbing system with limited crustal assimilation but more significant lithospheric mantle assimilation, in agreement with other studies (e.g., Marzoli et al., 2026). In fact, our data imply the simultaneous eruption of lavas from distinct asthenospheric and lithospheric mantle sources, thus challenging the current chemostratigraphy and supporting the results of previous studies (e.g., Kale et al., 2019; Peng et al. 1998; Mahoney et al., 2000). Further, the combination of traditional long-lived isotope data and short-lived µ¹⁸²W and µ¹⁴²Nd shows that the apparent “upper mantle component” (low ²⁰⁶Pb/²⁰⁴Pb and ²⁰⁷Pb/²⁰⁴Pb, high εNd and εHf) is actually a distinct mantle component originating from a deep plume-derived mantle source. This result further advocates for a temporal evolution of the lower mantle source of the Deccan-La Réunion mantle plume.

References