Australia: The Land Where Time Began

A biography of the Australian continent 

Earth’s longest continental Hotspot Track – Lithospheric Controls on Magma Composition

Hotspots are anomalous volcanic regions at the surface of the Earth that are not obviously associated with boundaries of tectonic plates. Included among the classic examples are the Hawaiian-Emperor chain and the Yellowstone-Snake River Plain province. It is believed the majority form as the tectonic plates of the Earth move over mantle plumes that are long-lived: buoyant upwellings that bring hot material from the deep mantle to the surface of the Earth (Morgan, 1971). That the rise height of plumes is limited by the thickness of the overlying lithosphere has been recognised for a long time (Davies, 1974; Farnetani & Richards, 1995; White & McKenzie, 1995) and, thereby, their minimum melting pressure. According to Davies et al. it should, therefore, have a controlling influence on the geochemistry of magmas that are plume-related, though there is so far a lack of evidence that is unambiguous for this. In this paper Davies et al. integrate observational constraints from surface geology, geochronology, reconstruction of plate motion, geochemistry and seismology to determine the depths at which the plume melts beneath the longest continental hotspot track, which is a 2,000 km track in eastern Australia displaying a record of volcanic activity between 33 Ma and 9 Ma (Cohen et al., 2008; Cohen at al., 2013), the Causgrove Track,. The analysis by Davies et al. highlights a strong correlation between lithospheric thickness and the composition of the magma along this track, with:

  1. Standard composition of basalt in regions of lithosphere where the thickness of the lithosphere is less than 110 km;

  2. Volcanic gaps where the thickness of the lithosphere exceeds 150 km; and

  3. Low volume volcanism that is leucitite-bearing in regions of intermediate lithospheric thickness.

Samples collected along this track had trace element concentrations that support the suggestion that these compositional variations, which resulted from different degrees of partial melting, are controlled by the thickness of the overlying lithosphere. The first observational constraints obtained from the results of this study on the sub-continental depth of melting of mantle plumes, and provide direct evidence that the thickness of the lithosphere has a dominant influence on the volume and chemical composition of magmas that are derived from plumes.

Plate theory has been successful in describing how the lithosphere, the rigid outermost shell of the Earth, consists of a mosaic of segments moving and interacting across the surface of the Earth. It also accounts for most of volcanism on Earth, which is concentrated at plate boundaries. There is, however, an important class of volcanism which occurs within plates or across the boundaries of plates, which often form linear volcanic chains that are older in the direction of the plate motion. Most of these so-called hotspots are believed to mark the surface expression of mantle plumes that are upwelling (Morgan, 1994; Duncan & Richards, 1991).

There are about 50 hotspots that have been identified at the surface of the Earth (Steinberger, 2000; Courtillot, Davaille, Besse & Stock, 2003). Only about 20 % of these occur on continents and, therefore, most of the knowledge of mantle plumes has come from the tracks of hotspots in oceanic settings. Oceanic lithosphere is, however, regularly recycled into the mantle through subduction, so in order to understand volcanism that is related to plumes before about 200 Ma, which constitutes most of the geological record of the Earth (Campbell & Griffiths, 1992), it is necessary to learn;

1)    How the plumes interact with the lithosphere of the continents; and

2)    How the chemical composition and volume erupted of lavas at the surface is affected by this interaction.

In this paper Davies et al. first combined observational constraints from surface geology, geochronology and the histories of the motion of the plates to identify the longest continental hotspot track on Earth which is in eastern Australia. They subsequently integrated constraints from seismology and geochemistry to determine how the variations in regional thickness of the lithosphere influence the volume and composition of magmas that are derived from plumes along this track.


 in Australia from the Cainozoic Era represents one of the most extensive intraplate volcanic regions of the Earth (Wellman & McDougall, 1974). There are 3 types of volcano that were identified in the classification of Wellman & McDougall (Wellman & McDougall, 1974), that is used widely, which was also adopted by Davies et al. in this paper:

1)    Central volcanoes, which predominantly have a basaltic composition though they have felsic lava flows or intrusions, with lavas typically being produced from central vents, which often build large volcanic complexes;

2)    Lava fields, which are basaltic, extensive and thin, and are often characterised by an abundance of small scoria, lava cones and maars;

3)    The leucitite suite, dominated by low volume, leucitite-bearing lavas which are rich in potassium.

These classes of volcanoes are principally distinguished by petrology, with central volcanoes being distinguished from lava-field volcanoes by the presence of felsic rock, and both distinguished from leucitite suite by the absence of leucitite (Johnson (ed), 1989). However, there are also considerable differences in age trends in these classes: 40Ar-39Ar and K-Ar geochronological studies demonstrated that the central volcanoes and the leucitite suite both defines tracks that are age-progressive, becoming younger in the south. The tracks that have been identified so far include:

1)    Comboyne – a track that is about 770 km long which extends from Fraser Island in Queensland to Comboyne in New South Wales, which displays a record of volcanic activity from about 32 Ma to 16 Ma;

2)    Canobolas – a track that is 760 km long, that extends from Bunya in Queensland to Canobolas in New South Wales, which records volcanism from about 24 Ma to 12 Ma;

3)    A track leading from Cape Hillsborough in Queensland to Buckland that records volcanism from about 34 Ma to 27 Ma; and

4)    A track bearing leucitite that is about 65 km long that extends from Bokhara River in New South Wales to Causgrove in Victoria that displays a record of volcanic activity from about 17 Ma to 9 Ma.

It is widely believed that these tracks mark the passage of mantle plumes beneath the Australian Plane that is migrating to the north (Cohen et al., 2008; Cohen et al., 2013; Knesel, Cohen, Vasconcelos & Thiede, 2008; Southerland et al., 2012). On the other hand, lava field volcanoes do not show such a progression and are believed to be generated by an alternative process, with the model that is driven by age being suggested for the formation of the Newer Volcanic Province (NVP) (Davies & Rawlinson, 2014; King & Anderson, 1998).

It has been considered that the central volcanoes of central Queensland are not related to the leucitite suite of New South Wales and Victoria (Cohen et al., 2008; Cohen et al., 2013; Wellman & McDougal, 1974; Southerland et al., 2012; Ewart, Campbell & Menzies, 1988), mainly because the volcanic province that was identified in each have dramatically different compositions and eruptive volume, and they are separated by a volcanic gap of more than 650 km. It is suggested by their relative locations and ages, however, that they may be the surface expression of the same mantle plume, therefore constituting a single hotspot track. This hypothesis has been tested by predicting volcanic locations along this track, by the reconstruction of absolute motion of the Australian Plate (Torsvik, 2010). Specifically, Davies et al. mapped locations at 15 volcanic centres that had been dated by 40Ar-39Ar techniques, and predict their location that at that time were associated with the next volcanic centre that had been dated. The estimates of the uncertainty in predicted locations, which arises through a combination of:

1)    Uncertainties in the diameter of the underlying mantle plume and the extent of the associated melt region (Farnetani & Richards, 1995; Leitch & Davies,2001);

2)    The potential for drift of the plume (Courtillot et al., 2003; Tarduno et al., 2003; Davies & Davies, 2009);

3)    The uncertainty that is introduced via preferential melt extraction pathways (Sleep, 1996); and

4)    Uncertainties in the 40Ar-39Ar ages of volcanic centres that have been dated.

Begargo Hill of the leucitite suite, is the only volcanic centre that is consistently located further south than is predicted by the reconstruction of Davies et al. They speculate that this indicates either a rapid phase of motion of the plume to the south that exceeds 4 cm/yr from about 17 Ma to 15 Ma, a change in the pathway of the extraction melt, or a combination of both. It was noted by Davies et al. in support of that:

1)    Variable migration of a plume have been observed elsewhere (Davies & Davies, 2009) and are also predicted in simulations of global mantle convection (Steinberger, 2008); and

2)    Begargo Hill is located to the south of a region where the lithospheric thickness if thickened, which may focus sub-lithospheric plume material, as well as any associated melt, southwards, as a result of the rapid motion of the Australian continent to the north.

It is most notable, however, that the northernmost leucitite-bearing volcano satisfies the location criterion, which confirms that the central volcanoes of central Queensland and the leucitite suite of News South Wales and Victoria are the surface expression of a single mantle plume. Combined, they constitute the longest hotspot track on Earth, which is known as the Cosgrove Track. This then leads to further questions. Specifically, these volcanic centres are the surface expression of the same mantle plume, so why does the Cosgrove Track have so many gaps? What is the driver of the considerable variations in volume and chemical composition of magmas that are derived from the plume between the central volcanoes and the leucitite Suit? Are these 2 characteristics related? Davies et al. combined the observational constraints from the seismology and geochemistry to answer these fundamental questions.

First, they generated map of lithospheric thickness, combining constraints from recent 3-D body-wave tomography results (Davies & Rawlinson, 2014; 23) with the regional Australian Seismological reference Model (AsSREM) (Southerland et al., 2012). The main features that are evident in Fig. 1b of the Methods and extended data include:

1)    The contrast between the lithosphere in the centre of Australia and the thinner lithosphere to the east, which is consistent with transition from Precambrian central Australia to Phanerozoic eastern Australia, and the oceanic lithosphere outboard of the margin of the continent;

2)    A zone of thin lithosphere, which is bound to the east by a zone of lithosphere that is of intermediate thickness that is of similar width extending southwestwards from about 30oS through central New South Wales into northern Victoria (Davies & Rawlinson, 2014); and

3)    Considerable changes in the thickness of the lithosphere over distances that are relatively short’

It is generally accepted that these lithospheric ‘steps’ will produce complex flows (Farrington, 2010) and, as previously noted, this resulted in a suggestion of a model that is edge-related for the formation of the NVP (Davies & Rawlinson, 2014; King & Anderson, 1998). Such edge-related mechanisms are probably also applicable to other areas of lava-field volcanism in the region, as it has been found that all lava-field volcanic provinces are adjacent to substantial steps in the thickness of the lithosphere, above lithosphere that is comparatively thin, and so providing a favourable setting (Davies & Rawlinson, 2014; King & Anderson, 1998).

The mechanism by which mantle plumes interact with these variations in the thickness of the lithosphere, specifically, how the variations in lithospheric thickness influence the volume and composition of magmas that are derived from plumes has remained poorly understood. Along the Cosgrove Track intriguing trends are evident:

1)    Volcanic gaps occur in regions where the lithospheric thickness exceeds about 150 km;

2)    The basaltic and felsic central volcanoes in central Queensland occur in regions of lithospheric thickness than about 110 km; and

3)    Volcanism of low volume, leucite-bearing  to the south in regions where the lithosphere is of intermediate thickness, with volcanic gaps within the leucitite suite, which also coincides with regions where the lithosphere is thicker.

It is suggested by these unambiguous trends that the thickness of overlying lithosphere dictates the volume and composition of magmas that are derived from plumes, by limiting the height the underlying plume rises to and, therefore, the degree of partial melting. Davies et al. infer that the underlying mantle plume:

1)    Cannot rise to depths that are shallow enough to induce melting by decompression in regions where the thickness of the lithosphere exceeds about 150 km, therefore providing an explanation for the volcanic gaps along the Cosgrove Track, and places the first observational constraint on the maximum depth of melting of mantle plumes beneath continents (excluding ultra-volatile melts that form kimberlites and carbonates);

2)    Undergoes a high degree of partial melting beneath lithosphere that is comparatively thin to produce basaltic and central volcanoes along the northern segment of the Cosgrove Track; and

3)    Undergoes very low-degree partial melting in regions where the thickness of the lithosphere is intermediate, and therefore facilitating the production of leucitite-bearing volcanoes of low volume towards the southern end of the Cosgrove Track.

In order to determine whether or not these inferences are compatible with geomechanical observations, Davies et al. collated trace element data that had been published previously from outcrops along the Cosgrave Track (Ewart, Chappelle & Menzies, 1988; Paul, Hergt & Woodhead, 2005). This data set provided a basis for testing the hypothesis put forward by Davies et al., though this data set is limited, with only 8 data points, 4 obtained from the central volcanoes of central Queensland and 4 from the leucitite suite. Incompatible trace elements such as barium (Ba) are transferred into the melt preferentially, as the mantle material is subjected to partial melting. The concentrations of the incompatible elements are subsequently diluted as the degree of partial melting increases (Hofmann, 2003). Accordingly if the available trace-element data supported the inferences that were presented earlier by Davies et al., higher concentrations of incompatible trace-elements, when compared to the samples from the basaltic central volcano, would be displayed by leucitite-bearing volcanics. According to Davies et al. it is apparent that such trends are present in the trace-element concentrations plotted in their Fig. 2: barium concentrations  are displayed by leucitite samples that are in excess of basaltic samples by a factor of about 3. The inferences by Davies et al. that the melt fraction along the Cosgrove Track is controlled by the thickness of the lithosphere, by which the rise height of plumes are limited, should also leave a signal that is discernible in the trace element concentrations illustrated in Fig. 2 of Davies et al. When melting occurs at greater depth, in the presence of garnet at higher concentrations (Ringwood, 1975), the heavy rare earth elements such as Lutetium (Lu), will be sequestered with respect of the middle rare earth elements such as gadolinium (Gd). Consequently, leucitite suite samples should display elevated Gd/Lu ratios of about 6, compared to about 2-3 for basalts of central Queensland. The hypothesis of Davies et al. therefore is supported by the trace element data that is available.

The question of why the basaltic and felsic volcanoes of central Queensland do not re-emerge to the south of Cosgrove, in a region where the lithosphere is comparatively thin, is an aspect of the Cosgrove Track that has not been addressed. Davies et al. suggest it is possible the underlying plume faded at about 8 Ma, as mantle plumes have finite lifetimes (Steinberger, 2000; Davies & Davies, 2009), though this would be an unlikely coincidence. Davies et al. speculate that an alternative mechanism is at play: it has been demonstrated previously (Davies & Rawlinson, 2014) that variations in the 3-D thickness of lithosphere, coupled with the rapid northwards drift to the north of the Australian plate, gives rise to a focused edge-driven convection cell to the west of the Cosgrove Track, near the NVP, which therefore provides a mechanism for the localisation of lava field volcanism to this region. An explanation for the onset of NVP volcanism, at about 5 Ma, has, however, remained elusive: The variations of the lithosphere thickness that drives the EDC in this region are probably long-lived (Rawlinson et al., 2014), which is not easy to reconcile with volcanism that is comparatively recent. The reconstructions by Davies et al., however, place the mantle plume that generated the Cosgrove hotspot track less than 50 km to the east of the NVP, from about 6.5 Ma to 5 Ma. It has been speculated by Davies et al. that the capture and entrainment of this plume, into an EDC cell that was existing previously, was the trigger for magmatism within the NVP and is an explanation for the lack of a hotspot track to the south of Cosgrove. It was noted by Davies et al. that in support of these ideas, that although EDC is expected to occur on all lithospheric steps, which includes those to the east of the predicted passage of the plume, in this region the dominant cell lies directly beneath the NVP (Davies & Rawlinson, 2014). According to Davies et al. it was to be expected that preferential westwards flow, and therefore entrainment of plume material, into the region of NVP, was evident in their previous model (Davies & Rawlinson, 2014). As far as Davies et al. knew the interaction between mantle plumes and EDC has been documented elsewhere. However, this process:

1)    has important implications for the surface expression of mantle plumes in the vicinity of step changes in the thickness of the lithosphere; and

2)    provides a solution to the global problem of why step changes in the thickness of the lithosphere, which occur along the edges of cratons as well as at passive margins, only produce volcanism at isolated locations, their study complements the previous study of Davies et al. (Davies & Rawlinson, 2014).

Davies et al. note, finally, that the present-day location of the mantle plume that generated the Cosgrove hotspot track lies to the northwest of Tasmania, coinciding with a region where there was recent seismicity and is at the western limit of the so called East Australian Plume System that had been previously imaged by the use to finite frequency tomography (Montelli, Nolet, Dahlen, 2006).


  1. Davies, D. R., et al. (2015). "Lithospheric controls on magma composition along Earth/'s longest continental hotspot track." Nature 525(7570): 511-514.





Author: M. H. Monroe
Last Updated 08/09/2017
Journey Back Through Time
Experience Australia
Aboriginal Australia
National Parks
Photo Galleries
Site Map
                                                                                           Author: M.H.Monroe  Email:     Sources & Further reading