May 2026 LIP of the Month

Large volcanic eruptions are mostly sourced above mobile basal mantle structures

Annalise Cucchiaro¹, Nicolas Flament¹, Maëlis Arnould², Noel Cressie³

¹ Environmental Futures, School of Earth, Atmospheric and Life Sciences, University of Wollongong, Wollongong, NSW 2522, Australia’. Email: A. Cucchiaro: ac576@uowmail.edu.au

² Université Claude Bernard Lyon 1, École Normale Supérieure de Lyon, CNRS, Université Jean Monnet, LGL-TPE, UMR 5276, Villeurbanne, France

³ NIASRA, School of Mathematics and Physics, University of Wollongong, Wollongong, NSW 2522, Australia

Extracted and modified from the publication:

Basal mantle structures, mantle plumes and LIPs

Large Igneous Provinces (LIPs) are the surface expression of mantle plumes (Bryan & Ernst, 2008; Ernst, 2014), which transport hot mantle material from the deep Earth to the surface. Seismic imaging of the lower mantle reveals two large continent-sized regions characterised by low velocity (Garnero & McNamara, 2008) called Large Low Velocity Provinces (LLVPs). Mantle plumes are thought to originate from these structures, but it is debated whether mantle plumes are generated preferentially in a plume generation zone along the edges of these basal mantle structures (Doubrovine et al., 2016; Torsvik et al., 2010), or from their interiors (Austermann et al., 2014; Davies, 2015), or from both. Some studies also propose that basal mantle structures have been stationary for hundreds of millions of years (Torsvik et al., 2010). However, mantle flow models that are either fully-dynamic or forced by plate tectonic histories both suggest that basal mantle structures have been mobile through time, deformed by sinking oceanic lithosphere over tens of millions of years (McNamara & Zhong, 2005). Here, we refer to the modelled equivalent of LLVPs as Big LOwer Mantle Basal Structures, or BLOBS.

Associations in space and time have been proposed to exist between 1/ LIPs, kimberlites and fixed LLVP edges (Torsvik et al., 2010), and 2/ LIPs, kimberlites and mobile BLOBS (Flament et al., 2022); however, previous studies have not explicitly considered mantle plumes; they are only implied as the mechanism in which LIPs are emplaced.

Here, we measure the distance between LIPs and plume conduits predicted by paleogeographically constrained mantle flow models at 1,040 km depth (Figure 1). We also measure the distance between LIPs and a suite of seismically-imaged LLVP (tomographic model cases T1–T4, from previous work (Auer et al., 2014; French & Romanowicz, 2014; Ritsema et al., 2011; Simmons et al., 2010), and between LIPs and BLOBS edges predicted by mantle flow model cases C1–C6 with varying basal layer density (between 0–2% greater than ambient mantle). We consider three LIP databases developed to assess distinct aspects of the volcanic record: D16 (Doubrovine et al., 2016), which consists of 26 LIPs associated with plume ‘heads’ representative of an updated selection proposed to be of deep origin; EY17 (Ernst & Youbi, 2017), which consists of 74 mafic and silicic LIPs associated with plume ‘heads’; and J18 (Johansson et al., 2018), which includes 167 volcanic products associated with both plume ‘heads’ and ‘tails’(Coffin et al., 2006).

Success metrics for LLSVPs and predicted BLOBS and plume conduits

We summarise angular distances (θ (°)) between large volcanic eruptions from 300 Ma and mantle structures (LLSVPs, BLOBS, and plumes) (Fig.2a) as cumulative distribution functions which we refer to as ‘sample distributions’ (Fig.2b). We repeat this for 1,000 random sets of LIPs with the same temporal resolution as the sample LIP database (Fig.2b). We then compute the p-value, denoted P, for a Monte Carlo significance test (Hope, 1968) based on the Kolmogorov-Smirnov statistic (Massey, 1951) (Fig.2b). This test allows us to compare the sample distance distribution with the random distance distribution. The null hypothesis of this test is that the distances between LIPs and mantle structures are the result of a sample of ‘uniform random locations’ where P < 0.05 indicates a departure from the null hypothesis (Casella & Berger, 2002). We report the mean of all angular distances within the sample distribution (θ_s) and the grand mean of all angular distances obtained from the 1,000 samples of uniform random locations (θ_r).

Relationships between LIPs and model plumes

LIPs in the database J18 that contains both plume head and plume tail products are closer to plume conduits than random points for all mantle flow model cases (C1–C6; Fig. 3a,b). Through Monte Carlo significance testing we find a statistical-dependence relationship between modelled plume conduits and the J18 database for all model cases (Fig. 3c). We attribute this to the inclusion of plume ‘tail’ products in J18, which results in more numerous products of hotspot volcanism and clustering due to the motion of plates above a plume conduit. Coincidentally, plume tail products in database J18 are also more closely comparable to the plume conduits detected in our models. Conversely, databases EY17 and D16 containing silicic LIPs and selected plume ‘head’ products respectively, resulted in no statistical relationships for any model case (Fig.3c).

Relationships between LIPs and: LLVPs and BLOBS

To assess the relationship between LIPs and basal structure edges, we consider LIPs above the interiors and exteriors of LLVPs and of BLOBS separately. For LIPs above the exterior of the LLVPs, we find a statistically significant relationship between LIPs in one database and LLVP edges in only one tomographic model (T1). There is no statistically significant relationship for any other permutation of LIPs in any database (D16, EY17, J18), and LLVP edges in any other tomographic model (T1–T4), either above the interior or the exterior of LLVPs. We attribute the success of tomographic model T1 to the overall smaller area of LLVPs, and to the presence of numerous small-scale basal mantle structures that are absent in other tomography models. For moving BLOBS, there is no statistical relationship between LIPs in any database and BLOBS interiors. However, there is a statistically significant relationship between J18 LIPs and BLOBS exteriors when BLOBS are between 1% and 1.6% denser than ambient mantle (Fig. 4c). This result supports a view in which plumes are not limited to a plume generation zone at the edge of fixed LLVPs (Doubrovine et al., 2016; Torsvik et al., 2010) but are instead anywhere above the interior of mobile BLOBS, or within ~5º to the outside of their edges due to plume tilting.

References: