Roadmap | The review process | Home
   ... more LIP discussion
.... more discussion on LIP classification
... this is a continuation of discussion of the LIP classification scheme proposed by Sheth (“Large Igneous Provinces (LIPs)”: Definition, recommended terminology, and a hierarchical classification) and the counter proposal by Bryan & Ernst (Proposed Revision to Large Igneous Province Classification).

To go to the beginning of the Discussion, click here.

4th April, 2006, Rajat Mazumder
I welcome Hetu´s effort to classify LIPs. His is a simple and easy-to-use system. Classification means grouping on some logical basis that should be simple so that one can follow and refer an item one wants to talk about and communicate easily within the scientific community. From that point of view it has enough logic to separate the LIPs into LVP and LPP. Some workers have criticised grouping LIPs into volcanic and plutonic categories. Then why should we classify igneous rocks themselves into volcanic and plutonic categories, and why not consider gabbros and basalts to be the same?

Regarding size, it is really problematic which dimension/area one should consider. Richard Ernst´s idea of 100,000 km2 is not a good one because in that case one should not consider the majority of the Proterozoic LIPs whose original dimensions are very difficult to calculate and their present distribution is much much smaller than their original one because of erosion!

Also I have a problem with the original definition of LIP. The definition included a phrase like "plate-tectonic processes unrelated….". Classification should not be interpretation-based because it can create confusion.

5th April, 2006, Sarajit Sensarma
This discussion is based on Hetu’s LIP scheme of classification, as well as the counter proposal suggested by Bryan & Ernst. I thank these colleagues for bringing out the confusion and inconsistency that has crept into the terminology and definitions of several aspects of LIPs.

Size: What about the size of ancient Precambrian LIPs? Since these terranes are variably deformed and subjected to prolonged erosion and alteration and resultant shrinkage, it may hardly be possible to estimate erupted volumes of melts and/or aerial extents of emplacement with reasonable certainty. Thus, Precambrian belts in many cases are unlikely to satisfy the minimum size limit suggested i.e. 50,000 km2 (Sheth) or 100,000 km2 (Bryan & Ernst), but still may otherwise maintain the character(s) of LIPs. Bryan et al. (2002) seem to have emphasized other criteria collectively more than size for an igneous province to be LIP. Ernst (10/03/06) also rightly opined “If events at 250 Ma and 120 Ma can be so uncertain in size, then the many Paleozoic and Proterozoic LIPs dominated by dykes and sills are even more uncertain in their extents”. In this respect, I find merit in Don’s and Kamal’s suggestions that in characterizing a LIP by size should be used loosely or not at all. Alternatively, the proposal of Bleeker & Ernst (2006) to use different size limits may be considered for now, but with caution pending further investigations and reports of other occurrences. Also, large volumes of intrusive (e.g., dykes etc.) and lavas associated in space may or may not be genetically/fundamentally related (e.g., Bondre et al., 2006). Thus, neither should be integrated immediately for size estimation in a given province, a point which Hetu has tried to address.

Bimodal LIPs: Bryan (2002) and Bryan & Ernst (2006) claim that no LIP having both substantial mafic and silicic components (>10% in both cases) is known. I would like to draw attention to the already-reported Dongargarh Group, a palaeoproterozoic LIP in central India, where about 8-10 km of preserved volcanic sequence comprises sub-equal proportions of both silicic (high temperature rhyolites and high Si rhyolites) and mafic volcanic rocks (high Mg basalts and tholeiites) (Sensarma, 2005; Sensarma et al., 2004). There is no reason why many such occurrences should not be discovered in future. Therefore, in my view, Hetu’s suggestion for a bimodal rhyolite-basalt province deserves consideration and retention, besides having separate slots for mafic LIPs and SLIPs. It is conceivable that silicic LIPs may make important contributions to understanding mafic LIPs (Bryan (12/03/06)) and vice-versa.


11th April, 2006, Vic Camp
I applaud Hetu Sheth's initiative to revise the LIP classification with more useable and more meaningful terms.  I do have a concern, however, in the proposal to lump continental flood basalts, oceanic basalt plateaus, and oceanic seamount chains into a single category (LBPs, for Large Basaltic Provinces).  As touched on by earlier contributors, the hallmark of CFBs is the eruption of extraordinary volumes of basalt over extraordinarily short periods of time, in contrast to oceanic seamount chains which have erupted smaller volumes over very long time intervals.  Independent of size or genesis, I believe that the unusually large magma-supply rates for CFBs, and some (but not all) oceanic plateaus, should take precedence in their classification.  I would cast my vote for a separate category for the CFBs. 

On the subject of LIP sizes, I would like to point out a minor mistake in Sheth's webpage concerning the size of the Columbia River Basalts.  The oft-quoted area of 164,000 km2 for the CRBs has been revised upward to ~215,000 km2, with a revised volume of ~234,000 km3, based on recent mapping and the incorporation of the Oregon Steens basalts into the CRB stratigraphy (Camp et al., 2003; Camp & Ross, 2004, Hooper et al., accepted).  The CRBs may still be the smallest of the continental flood basalts, but now they are just a bit closer in size to their larger cousins!



I would like to comment on the volume of the Siberian Traps. This matter has arisen from my current work on a P4 chapter and is also related to the following remark by Richard Ernst ‘we are really at early days in understanding the size of LIP – even the youngest ones: The Siberian traps nearly doubled in known size a few years ago, from about 2.6 million km3 to >3.9 million km3 with the dating of basalts under the West Siberian Basin (Reichow & Saunders and others …

First, note that km3 should be km2.

Second, it is not true that the size of the Siberian traps has been doubled by recent work. The volcanic rocks in the West Siberian Basin were recognized as a part of the Siberian traps long ago. For example, Fedorenko et al. (1996) write: ‘Milanovsky (1976) concluded that the original extent of the Siberian traps was ~4 x 106 km2 and that their volume exceeded 2 x 106 km3. We believe that even this volume may be underestimated …’. This is already the same as the ‘doubled’ values but described as underestimates!

Masaitis (1983) estimated the original extent of Siberian Traps to be about 7 x 106 km2 and the volume as much as 4 x 106 km3. Therefore, the ‘doubled ... a few years ago’ value is only a half that already suggested by Russian geologists in the 1980s. A new value currently circulating in Russian geological literature is 16 x 106 km3; 7 times the ‘doubled’ value (see ref. 22 to Dobretsov in Vasil’ev et al., 2000). This value includes traps from the Siberian platform, and buried traps from the West Siberian Basin, Kara and Barents undersea basins.

All such estimates, however, are suspicious, because they are made in an oversimplistic way; multiplying the size by an average thickness of volcanic rocks. Thickness varies in some areas from tens of meters for sills to kilometers for lava piles. An attempt at accurate volume estimation made done by Vasil’ev et al. (2000). These authors focused on the Siberian platform and calculated the preserved volume of lava, volcanoclastic and intrusive rocks using geological survey and drilling data. The size of the Siberian Traps erupted on the platform according to Vasil’ev et al. (2000) is 4.3 x 106 km2, which of the same order as the size assumed by Reichow et al. (2002) for the whole LIP. The volume estimated by Vasil’ev et al. (2000) is 1.752 x 106 km3.

One may pose the question, what is the real size of the Siberian traps? In my view, the volume of order of 4 x 106 km3 proposed by Masaitis (1983) looks close to the true value. It is a bit more than double the precise volume of present-day volcanic remnants on the Siberian platform. It includes various types of rocks from ultrabasic to acidic with basalts as the major rock type. Basalts are dominant on the Siberian platform, whereas rhyolites and dacites are abundant in the West Siberian Basin (Masaitis, 1983; Medvedev et al., 2003).

Reference for doubling the volume:

  • Reichow, M.K., Saunders, A.D., White, R.V., Pringle, M.S., Al'mukhamedov, A.I., Medvedev, A.I. and Kirda, N.P., 2002. 40Ar/39Ar dates from the West Siberian Basin: Siberian flood basalt province doubled. Science, 296, 1846-1849.

Other cited references:

  • Fedorenko, V.I., Lightfoot, P.C., Naldrett, A.J., Czamanske, G.K., Hawkesworth, C.J., Wooden, J.L. and Ebel, D.S., 1996. Petrogenesis of the flood-basalt sequence at Noril’sk, North Central Siberia. Int. Geol. Rev., 38, 99-135.
  • Masaitis, V.L., 1983. Permian and Triassic volcanism of Siberia: problems of dynamic reconstructions. Zapiski Vserossiiskogo Mineralogicheskogo Obshestva, 4, 412-425. (In Russian)
  • Medvedev A.Ya., Al'mukhamedov A.I. and Kirda N.P. (2003) Geochemistry of Permo-Triassic volcanic rocks of West Siberia. Geologiya i Geofizika 44, 86-100.
  • Vasil'ev, Yu.R., Zolotukhin, V.V., Feoktistov, G.D. and Prusskaya, S.N., 2000. Evaluation of the volumes and genesis of Permo-Triassic trap magmatism on the Siberian Platform, Geologiya i Geofizika, 41, 1696-1705. (In Russian)

13th April, 2006, Scott Bryan & Richard Ernst: Reply to Discussion & Comments by Rajat Mazumder & Sarajit Sensarma

It cannot be stressed enough that the term Large Igneous Province (LIP), should not be used to include every igneous terrane or province of local or even regional significance. Correct identification of LIP events is critical for identifying, among other aspects:

  1. major or catastrophic mantle events through Earth history (e.g., arrival of core-mantle boundary-derived plume, mantle overturn or delamination, mantle penetration and melting by boloidal impact, or edge convection driven by rapid continental rifting),
  2. major episodes of new crustal addition from the upper mantle,
  3. episodes of continental breakup and supercontinent cycles,
  4. those events that will have significantly impacted on the biosphere and atmosphere leading to climate shifts and mass extinctions, and
  5. the formation of major mineral provinces (e.g., Ni-Cu-PGE deposits for the mafic LIPs and epithermal Au-Ag bonanza mineralisation for silicic LIPs).

It is important to develop a definition and classification for LIPs that will direct us towards their origin and recognise those with regional to global effects. Several critical points are emphasized below.

1. Minimum event size should be at least 100,000 km2 or km3 (if not larger), not 50,000 km2 or km3
Despite the concerns of Mazumder & Sensarma that the original dimensions for the majority of Proterozoic LIPs have proved difficult to calculate because of losses due to erosion, as we discussed in our "Proposed Revision to Large Igneous Province Classification", and as shown by other workers (e.g., Ernst & Buchan, 2001a), the 100,000 km2 minimum extent is in fact met by many of the Proterozoic LIPs; this has been calculated largely on the areal extent of the intrusive components. Additionally, a separate grouping (‘waiting room’) may be required for those smaller-scale igneous provinces that could have been a LIP or part of a LIP, but currently do not meet the dimension criteria for LIP definition as a result of a lack of data or size limitations due to erosion or burial (cf. Bleeker & Ernst, 2006). This is because new LIPs are identified as additional age data are obtained allowing correlation of what were previously considered unrelated igneous events in different and widely separated tectonic terranes (e.g., the late Mesoproterozoic Warakurna LIP; Windgate et al., 2004).

2. LIP Events should be “brief”
An issue we see with the classification scheme of Sheth (2006) is that there is no time consideration or limitation, such that any mass of igneous rock of >50,000 km2 areal extent can be defined as a LIP. Given sufficient time and space, all plate boundary processes generating magma (ie. MORs, subduction zones, continental rifts) will produce igneous rock of LIP-scale dimensions. It is well-recognised for example, that major continental batholiths, which can meet the areal definition of LIPs, comprise intrusive suites that range considerably in age (up to 100's of Myr) and intruded under a range of different tectonic regimes - they are composite features.

The 2.5–2.2 Ga Dongargarh Group is a case in point. This Group represents a 300 Myr history of magmatism and sedimentation (see Table 1 of Sensarma et al., 2004). This long duration is diametrically opposed to the whole basis of LIPs that are recognised as geologically brief episodes of rapid magma eruption. At most, LIPs appear to have an overall age duration of up to 60 Myr (Ernst & Buchan, 2001b), but in these cases, most magma volume was likely emplaced in pulse(s) over much shorter periods (~< 10 Myr).

3) Compositional bimodality vs volumetric bimodality
Most, if not all LIPs emplaced into continental regions are compositionally bimodal, and it may be possible for a LIP to be volumetrically bimodal. However, the preserved record of most if not all LIPs is volumetrically, dominated by either mafic or silicic igneous rock, a fact which is constrained by their crustal setting, crustal source regions and nature of the large-scale magmatic processes in operation during LIP events (see Table 3 of Bryan et al., 2002).

Sensarma (and offered by Sheth as an example) states that the 2.5–2.2 Ga Dongargarh Group, a Palaeoproterozoic volcano-sedimentary province in central India, is an example of a bimodal LIP where equal proportions of mafic and silicic igneous rock were emplaced, and by implication, emplaced coevally (ie. interbedded). The Dongargarh Group comprises many different formations (volcanic and sedimentary; see Table 1 of Sensarma et al., 2004) and that although the stratigraphic Group can be described as bimodal, the bimodalism is stratigraphically (and temporally) separate. The progression from an early silicic volcanic phase followed by fundamentally mafic volcanism is a common compositional progression in continental rifting (either back-arc or intraplate), and even in rifted oceanic arcs (see Fackler Adams & Busby, 1998). The presence of andesite formations within the Group may offer support for a subduction-related setting. The different stratigraphic formations of the Dongargarh Group therefore likely represent entirely different events that may or may not include a LIP event(s) and may or may not be related (see point 4 below). The basal rhyolitic stratigraphic formation has a minimum volume of 8000 km3, thus requiring the extrusive volume of overlying and temporally related basaltic formation(s) to be >90,000 km3, for consideration as a LIP. In this case, the Dongargarh Group would not be volumetrically bimodal.

An important point is that the eruptive stratigraphies of LIPs are incompletely preserved yet many mafic LIPs show an increasing proportion of silicic volcanism up-section. Consequently, the proportion of silicic volcanic products may be underestimated (because of erosion), but also, their occurrence late in the evolution of a mafic LIP may be an artefact of preservation (Bryan et al., 2002). Constraining the true proportion of silicic to mafic igneous rock (and total volume) must also include the eroded portion and hidden intrusive component that in general, remain largely unknown.

4) Magmatic units grouped as a LIP must be presumed genetically related
Sensarma states: "large volumes of intrusive (e.g., dykes etc.) and lavas associated in space may or may not be genetically/fundamentally related, and .... neither should be integrated immediately for size estimation in a given province, a point which Hetu has tried to address."

We believe it is imperative to establish clear temporal and genetic associations for the igneous rocks upon which a LIP will be defined. Different genetic processes and source regions (e.g., mantle versus crust) are involved for the mafic and silicic magmas in LIPs. However, genetic links are being established for mafic and silicic magmatism in LIPs, such as the recognition of low and high-Ti-type basalts and rhyolites (e.g., Peate, 1997; Marsh et al., 2001). Where genetic relationships may be less apparent, the igneous rocks are still relatable by their stratigraphic association and ages. Establishing temporal relationships and the rapidity of emplacement of such huge volumes of magma have been the foundation stone to the whole concept of LIPs.


  • Bleeker W, Ernst R (2006) Short-lived mantle generated magmatic events and their dyke swarms: The key unlocking Earth's paleogeographic record back to 2.6 Ga. In: Hanski E, Mertanen S, Rämö T, Vuollo J (eds) Dyke Swarms - Time Markers of Crustal Evolution. A.A. Balkema Publishers, Rotterdam, 2006.
  • Bryan SE, Riley TR, Jerram DA, Leat PT, Stephens CJ (2002) Silicic volcanism: an under-valued component of large igneous provinces and volcanic rifted margins. In: Menzies MA, Klemperer SL, Ebinger CJ, Baker J (eds) Magmatic Rifted Margins. Geological Society of America Special Paper, 362: 99-120.
  • Ernst RE, Buchan KL (2001a) The use of mafic dike swarms in identifying and locating mantle plumes. In: Ernst, R.E., Buchan, K.L. (Eds.), Mantle Plumes: Their Identification Through Time, Special Paper, vol. 352. Geological Society of America, Boulder, CO, pp. 247– 265.
  • Ernst RE, Buchan KL (2001b) Large mafic magmatic events through time and links to mantle-plume heads. In: Ernst RE, Buchan KL (Eds) Mantle Plumes: Their Identification Through Time. Geological Society of America Special Paper 352, 483- 575.
  • Fackler-Adams B & Busby C (1998) Structural and stratigraphic evolution of extensional oceanic arcs. Geology, 26: 735-738.
  • Marsh JS, Ewart A, Milner SC, Duncan AR, Miller RMcG (2001) The Etendeka igneous province; magma types and their stratigraphic distribution with implications for the evolution of the Paraná-Etendeka flood basalt province. Bulletin of Volcanology, 62: 464 486.
  • Peate DW (1997) The Paraná-Etendeka Province. In: Mahoney JJ, Coffin MF (Eds), Large Igneous Provinces: Continental, Oceanic, and Planetary Flood Volcanism. American Geophysical Union, Geophysical Monograph 100, pp. 217-245.
  • Sheth H (2006) Large Igneous Provinces (LIPs): Definition, recommended terminology, and a hierarchical classification. Journal of Volcanology & Geothermal Research (in press).
  • Wingate, MTD, Pirajno F, Morris PA (2004) Warakurna large igneous province: a new Mesoproterozic large igneous province in west-central Australia. Geology 32, 105–108.

13th April, 2006, Hetu Sheth
With my proposed LIP terminology and classification, I seem to have opened a can of worms. But it's very well that we sort problems out once and for all.

Will everyone please decide first whether "large" (as in LIP) should be defined on the basis of size, or eruption/magma supply rates? To me, it's size. It's true that the LBPs have both large sizes and possibly high eruption/magma supply rates. But the basic criterion is size, not the latter. Therefore, the fact that some of the large batholiths may have formed over tens of hundreds of Myr does not render them unsuitable for the LIP category, which is what Bryan and Ernst imply.

If this is not acceptable, dispense with the term LIP, and coin new terms that address the high eruption/magma supply rates. In nature, there is no arbitrary cut-off. If a province formed over tens of Myr, while a CFB formed over 1 Myr or less, and both are the same size and large, both are LIPs to me. The time scales reflect mantle processes (melting) plus lithospheric processes, and all these are quite variable.

The statement by Bryan and Ernst that any "normal" process like seafloor spreading will produce an LIP over a long enough time is correct (and this doesn't matter), but I showed that SFS can produce LIP-sized areas of volumes over 5 Myr, which is in fact the duration of many CFBs. Where do we draw the line?

To me, the meaning of "large igneous province" is simple. I do not see that the confusion is cleared simply by using the terms "mafic LIP" and "silicic LIP". Even the word "silicic" should be replaced with "felsic", as, strictly, "silicic" does NOT mean silica-rich (which is what is implied) but simply silica-bearing. It must be admitted that the ocean floor is a LIP and LVP and LBP. To me, the exclusion of the ocean floor makes no sense as far as the dictionary meanings of "large", "igneous", and "province" go. If the expression LIP is to be retained, give a place to the ocean floor.

Thanks everyone for their input.

18th April, 2006, Scott Bryan
We feel that Sheth misrepresents the term "Silicic LIP", but in suggesting that "silicic" should be replaced by "felsic" he provides an opportunityfor us to clarify why we (Bryan et al., 2002) have used the term "silicic". Before suggesting replacing or abandoning terms, we need to start with the correct definitions (Bates & Jackson, 1987):

  • Silicic is to describe a silica-rich rock or magma, where silica constitutes at least 65% or 2/3 of the rock – granite and rhyolite are typical silicic rocks;
  • Felsic is derived from fel (feldspar & feldspathoid) and sic (silica), to describe igneous rocks having abundant light coloured minerals in its mode, and also to describe those minerals (quartz, feldspar, feldspathoids, muscovite) as a group.

From these definitions, Felsic is a term very much based on the mode and mineralogy; Silicic is based on the whole-rock (silica) composition. Felsic can be applied to highly differentiated alkaline rocks (trachytes, phonolites) that have abundant light-coloured minerals, but are not silica-rich (phonolites have SiO2 contents of ~60 wt%). Therefore, 'felsic' is a much broader term that encompasses rock compositions ranging from silica undersaturated and highly alkaline, to silica-oversaturated peraluminous to peralkaline rhyolites & high-silica rhyolites. Highly differentiated and alkaline felsic rocks such as phonolites & trachytes are not present in Silicic LIPs.

The ignimbrites and other silicic rocks that comprise Silicic LIPs display considerable variation in crystal content and phenocryst assemblage; some are aphyric. In general, the ignimbrites contain the phenocryst assemblage of plagioclase, quartz, and Fe-Ti oxides with alkali feldspar uncommon, and the ferromagnesian phases are dominated by pyroxene, biotite and/or hornblende (Cameron et al., 1980; Wark, 1991; Ewart et al., 1992; Riley et al., 2001). Ignimbrites from the Whitsunday and Sierra Madre Occidental Silicic LIPs are predominantly pyroxene rhyolites (Ewart et al., 1992; Cameron et al., 1980). Although plagioclase is abundant, for the Whitsunday Silicic LIP, plagioclase is modally dominant in all volcanic compositions (basalt to high-silica rhyolite; Bryan et al., 1997).

In general, interpreting whole-rock compositions from the mineralogy can be a highly misleading practice. Our purpose in using the word "Silicic" was, and is, simply to emphasise the volumetric dominance of igneous compositions with >65% wt SiO2 in the LIPs (ie dacite to rhyolite); the term does mean silica-rich. Silicic is equally applicable to describe similar rocks in the mafic LIPs (e.g., Marsh et al., 2001). Although some workers and other definitions consider felsic and silicic synonymous or interchangeable, felsic is inappropriate because:

  1. whole-rock compositions can be incorrectly interpreted from mineralogy and modes;
  2. derivation of the term includes the feldspathoid minerals, which are not present in any of the Silicic LIP rocks;
  3. of the light coloured minerals, only plagioclase is dominant – quartz and alkali feldspar are uncommon, and muscovite is extremely rare, yet plagioclase is also very abundant in the mafic rocks; and probably most importantly,
  4. the term is unnecessarily broad to describe the igneous rocks of Silicic LIPs.


19th April, 2006, Hetu Sheth
Thanks to Scott for giving us his view of the term "silicic". I do not see major problems with his definition. However, I should note that, conventionally, the opposite of "mafic" is "felsic". Thus, mafic rocks and felsic rocks are opposites. So maybe, as Scott says, "felsic" should be replaced by "silicic" in igneous petrology in general, not just the LIP business. I haven't known many people use "silicic" (and, I might add, I thoroughly abhor the word "acid" in common use for rocks like rhyolites and dacites). For minerals, recall, we have the terms "femic" and "salic", which correspond to "mafic" and "felsic" for the rocks. One more thing. Above, in the previous comment, I said that the terms LIP, LBP etc. should relate to size, not high eruption rate. If the aim was to represent volcanic provinces with high melt eruption rate, we should coin new terms that express these. We already have such terms, however – "flood basalt" is precisely that, so why not simply use that instead of the "mafic LIP" of Bryan and Ernst? And why not use "flood rhyolites" and "flood trachytes" instead of "silicic LIP"? The last two are very much in use in the literature too.

19th April, 2006, Scott Bryan
To answer Hetu Sheth's query about using "flood" to describe the sheet-like expanses of basalt lavas and rhyolites that form LIPs, it is important to understand the origin of the term "flood basalt", which is largely a model-driven description of the basalt lavas (see Shaw & Swanson, 1970). I direct those interested to read the excellent paper of Self et al. (1997) that summarises the basis for the early model of flood basalt emplacement, and documents very well, evidence for the now widely accepted "inflated pahoehoe" model. Some of the relevant sections of the Self et al. (997) paper are summarised below.

Shaw & Swanson (1970) interpreted the flood basalt lavas to have been emplaced rapidly (in days to weeks), by thick, extensive turbulent flows (noting that the deposit thickness approximated the flow thickness), with flow velocities of several km/hr. This was to account for the following features:

  1. that the flood basalt lavas do not show evidence for measurable crystallisation during transport, and were not significantly cooled over their run-out distances of up to 500 km;
  2. the glassy selvages, which were interpreted to indicate the lavas had erupted at temperatures well above their liquidus temperatures (this allowed cooling, but without crystallisation).

As consequence of the high temperatures, magma viscosities were thought to be very low, and permitted rapid flowage and emplacement. The general picture from these early studies then was of cataclysmic floods of lava charging across the landscape, hence the term "flood basalt".

Several studies have now shown this to have been highly unlikely, and flood basalt lavas are now thought to have been emplaced as inflated pahoehoe-like lavas that are relatively slow advancing (as low as 0.2-1.4 m/s), from low eruptive rate (e.g., ~4000 m3/s average total eruption rate for Roza Member) and long-lived eruptions (years to decades; see Self et al. (1997) & later publications).

The term "flood", although embedded in our everyday usage and description of basalt lavas in LIPs, has a model-driven beginning, and would be more appropriate for describing the voluminous rhyolites formed by ignimbrite-forming eruptions than the basalts. However, its use should not be encouraged simply to describe a very extensive, sheet-like geometry for the rhyolites or other eruptive units. "Flood basalt" is only applicable to the mafic eruptive units; as we have been discussing, there are also very voluminous intrusive units with sheet-like geometries and similar extents that occur in LIPs, and the term "flood basalt" (or rhyolite) was never meant to describe these.


  • Self S, Thordarson T, Keszthelyi L (1997) Emplacement of continental flood basalt lava flows. In: Mahoney, JJ, Coffin MF (Eds) Large Igneous Provinces: Continental, Oceanic, and Planetary Flood Volcanism. American Geophysical Union, Geophysical Monograph 100, Washington DC, pp 381-410.
  • Shaw HR, Swanson DA (1970) Eruption and flow rates of flood basalts. In: Gilmour EH, Stradling D (eds) Procedings, Second Columbia River Basalt Symposium. Eastern Washington State College Press, Cheney, pp 271-299.

26th April, 2006, Hetu Sheth
I have studied the Shaw and Swanson and Self et al. works. The definition of "flood basalt" is something I also explored in my LIP terminology and nomenclature paper, though I didn't include it in the webpage. Here is a
relevant excerpt from my paper:

In this section, I discuss the desirability of retaining the terms “flood basalt” and “continental flood basalt” (CFB). The first use of the term “flood basalt” is not clear. The New Penguin Dictionary of Geology (Kearey, 1996) defines flood basalt as “an extrusion of low viscosity basaltic magma of very large volume”. Sigurdsson (1999) defines flood basalts as “laterally extensive deposits of basaltic lava flows, resulting from outpouring of vast volumes of magmas during fissure eruptions.” Bardintzeff & McBirney (2001) define flood basalt as “a voluminous, laterally extensive lava flow, normally erupted from a fissure”.

Many individual lava flows in provinces like the Deccan and Columbia River are thick (>100 m) and laterally extensive (>100 km), with volumes exceeding 1,000 km3 (see Bondre et al., 2004 and references therein). The Columbia River basalt province was one of the first in the world to be studied in detail by modern methods. Shaw & Swanson (1970) presented a model of turbulent, rapid emplacement for these large lava flows. Based on features such as glassy selvages in the basalts at great distances (100 km) from the source vents, they correctly inferred insignificant cooling in these lavas and therefore proposed very rapid emplacement over days to weeks. As Self et al. (1997) pointed out, however, lack of heat loss in the Columbia River basalts need not mean rapid, turbulent emplacement. Based on observations of modern Hawaiian lavas and those of Iceland, Hon et al. (1994), Self et al. (1997) and Thordarson & Self (1998) argued that the large flood basalt flows would have formed over long time periods (months to years) by insulated lava transport and internal growth by inflation. With insulation by a frozen surface crust, lava can be transported several hundred kilometres with almost negligible heat loss. In the model of Self et al. (1997), individual lava flows in the flood basalt provinces, though orders of magnitude larger than Hawaiian flows, would also have formed through inflation, over months to years.

Early proponents of rapid emplacement of CFBs, such as Shaw & Swanson (1970), in all probability did NOT envisage these lavas flowing like torrents in a river experiencing flooding, and thereby covering 100’s of kilometres without significant cooling. The word FLOOD does not necessarily imply a powerful torrent, but generally inundation. Flood basalts are analogous to flood waters in the sense that they both fill and inundate low-lying topography. It is not the sheetlike form of lavas over flat surfaces, but the fact that they convert originally uneven topography into flat topography, that is well represented by the term flood basalt. The term therefore is a suitable and valuable scientific term for lava flows of large volume and high fluidity that produce essentially flat landscapes by inundating and filling pre-existing topography.

29th May, 2006, Bernard Bonin, Chairman of the IUGS Subcommission on Systematics of Igneous Rocks
Old terms, such as silicic, felsic, salic, etc. should be used correctly. Definitions on igneous rock features are recalled hereafter. They are taken from: Le Maitre (Editor), 2002. Igneous rocks. A classification and Glossary of Terms. 2nd Edition. Recommendations of the International Union of Geological Sciences Subcommission on Systematics of Igneous Rocks. Cambridge University Press, 236 pages.

Chemical definitions, updated from old definitions by Judd (1881), Abich (1841), Judd (1886), and Abich (1841), respectively:

  • Ultrabasic: rocks having less than 45 wt% SiO2 (Le Maitre, 2002, page 35)
  • Basic: rocks having from 45 to 52 wt% SiO2 (Le Maitre, 2002, page 35)
  • Intermediate: rocks having from 52 to 63 wt% SiO2 (Le Maitre, 2002, page 35)
  • Acid: rocks having more than 63 wt% SiO2 (Le Maitre, 2002, page 35)

Silicic and Siliceous are not defined in Le Maitre (2002). In the Bates and Jackson Glossary of Geology, silicic stands for rocks having more than either 65 wt% SiO2, or comprising two thirds of the rocks (i.e. 66.67 wt%). Thus, a SILICIC rock is always ACID, but the least acid rocks are NOT silicic. Siliceous stands for rocks having free silica in their mode. The definition is, therefore, modal, not chemical.

Modal definitions, taken from the seminal paper in which the terms were created (Cross, Iddings, Pearson and Washington, 1912. Modification of the quantitative system of classification of igneous rocks. Journal of Geology, Chicago, volume 20, page 561):

  • Felsic: a collective term for MODAL quartz, feldspars and feldspathoids, which was introduced to stop the normative term SALIC being used incorrectly for that purpose. The first definition by CIPW concerns clearly minerals, not rocks. Surprisingly, in the Bates and Jackson Glossary of Geology, the term 2 is applied, first, to a group of rocks having "abundant" (how much?) felsic minerals and, second, to the group of minerals. In that definition, a silicic rock should be felsic. But, as felsic minerals include feldspathoids, some felsic rocks can have ultrabasic compositions, e.g., urtite with more than 70 % nepheline and yielding about 43 wt% SiO2 (Sorensen, 1974. Chapter II.3. Nephelinites and ijolites. The Alkaline Rocks. John Wiley and Sons, London, New York, page 56).
  • Mafic: a collective term for MODAL ferromagnesian minerals, such as olivine, pyroxene, etc., which was introduced to stop the normative term FEMIC incorrectly being used for that purpose.

CIPW-normative original definitions (Cross, Iddings, Pearson and Washington, 1902. A quantitative chemico-mineralogical classification and nomenclature of igneous rocks. Journal of Geology, Chicago, volume 10, page 573):

  • Salic: a name used in the CIPW normative classification for one of the two major groups of normative minerals, which includes quartz, feldspars and feldspathoids, as well as zircon, corundum and the sodium salts. The Bates and Jackson Glossary of Geology applies the term salic first to a group of minerals, then to rocks having one or more of salic minerals as major components of the norm, e.g., a glassy rhyolitic obsidian is salic and also silicic. Glassy phonolite with more than 80 % CIPW-normative (alkali feldspars + nepheline) is salic, though yielding no more than about 57 wt% SiO2 (Sorensen, 1974. Chapter II.2. Alkali syenites, feldspathoidal syenite and related lavas. The Alkaline Rocks. John Wiley and Sons, London, New York, page 33).
  • Femic: a name used in the CIPW normative classification for one of the two major groups of normative minerals, which includes the Fe and Mg silicates, such as olivine and pyroxene, as well as the Fe and To oxides, apatite and fluorite.

Thus, it is incorrect to state that felsic stands for ROCKS and salic for MINERALS. Actually, both terms stand for MINERALS, felsic for minerals occurring in the mode of the rocks and salic for minerals that are calculated in the norm. These terms were used for ROCKS only after they were defined for minerals, and remain imprecise as no quantified amounts ("abundant", "major component", etc.) of felsic, mafic, salic, femic minerals are offered for felsic, mafic, salic, femic rocks.

As stressed by several contributors, the correct use of terms, following accepted definitions, is important for scientific communication thoughout the world. These notes, as formal as they could appear, are given only to help to understand better the terms that are currently circulating in the literature.

9th June, 2006, Hetu Sheth
Dear WM, I welcome Dr. Bonin’s valuable comment and thank him for pointing out the correct usage of some petrological terms. However, when he writes that “whether LIP terminology should be expanded, more precise and more detailed than the current definition of LIPs, or should retain its current somewhat loose definition, is a matter of personal philosophy”, I would point out that the distinction between granite and granodiorite may then also be considered a matter of personal philosophy [Ed: and the term "plume", also, perhaps; see Plume definitions page]. If loose terms are okay, then granite is a very good and perfectly satisfactory "loose" term for granite, granodiorite, trondhjemite, tonalite and charnockite etc.

Nature is a continuum; any classification is necessarily artificial, and yet classifications and correct terminology are essential for a uniform scientific language and understanding. Denying this means denying the great practical utility of classifications such as mine, the alternative one by Bryan and Ernst, and all the very helpful igneous rock classifications by the IUGS Subcommission on Systematics of Igneous Rocks.

5th September, 2006, Romain Meyer
I agree with the authors and discussion contributions that recent LIP research has illustrated that LIPs are more varied than the initial definitions embrace. However, in Bryan & Ernst's LIP 2 proposed classification the NAIP is a continental (YES!!) Mafic (but not unanimously!!) LIP. Such a classification in mafic/silicic LIP's will thus bring the scientific community against the problem that for some (older) Paleozoic, Proterozoic and Archean LIPs a correct classification is not possible. This is mainly due to fact that a huge percentage of igneous rocks has been eroded away. Imagine, for example, the NAIP after erosion of the flood basalts. The remaining magmas will all be sub-NAIP rocks (Scotland, lower series from SE Greenland and mid-Norwegian margin) and indicate a silicic LIP.

In my opinion it will be quite difficult to strictly separate mafic from silicic LIPs. Following the work of Bowen we all know that we will normally see both basaltic primary magmas and differentiated as well as contaminated/assimilated silicic magmas. Mantle/continetal crust interactions could also be a geochemical key in the understanding of mantle processes (temperature, source compositon etc.) due to anatectic and hybridization processes in the crust. So the silicic parts of LIPs will become of greater scientific interest. The actual mafic to silicic interpretation of LIPs includes also a bias due to:

  1. logistical and accessibility limitations in many LIPs, and
  2. over-sampling of the same outcrops/boreholes.

As a result I would prefer toclassify LIPs into continental LIPs and oceanic ones and transitional LIPs continental volcanic rifted margins on the basis of their plate tectonic context at the time of formation. If this is not known, they should be classified simply as unclassified LIPs. Thus:

  1. oceanic LIPs
  2. continental LIPs
  3. transitonal LIPs

11th September, 2006, Sami Mikhail
This discussion on LIP classification has demonstrated that there are many ways in which one can classify LIPs. There appeas to be more ways to classify LIPs than there are models to explain their formation. What I see as evident is that no two LIPs are the same (perfectly stated by Saunders 2005), we can say for sure that the Whitsunday volcanic province is dominated by silicic material whereas the Ontong Java plateau is dominated by mafic material. Also that the Whitsunday is associated with continental rifting classing it as a VRM and Ontong Java is an oceanic plateau not a ‘continental’ VRM but has MORs surrounding it. BUT, the NAIP is both; it is associated with continental rifting which has evolved into a MOR making part of it, technically, an oceanic plateau (the lavas on the oceanic basin surrounding Iceland and enveloped by the UK and Greenland).

As interesting as it is to classify and sub-divide LIPs, the real question still remains ‘how do we get such high magma fluxes in such short time scales?’

Stratifying LIPs through time is essential to observe patterns which may show us something about their origin. Yale & Carpenter (1998) showed that LIP clusters coincide with GDS and follow supercontinent assembly and suggest that incubation of the mantle beneath supercontinents can (possibly) initiate top-down plumes simply by trapping heat and even suggest that periods of relatively dispersed continents such as from 725 to 250 Ma may explain gaps in the LIP record. [Ed: See also Anderson, D.L., 1994, Superplumes or supercontinents: Geology, 22, p. 39-42.) As well as stratifying LIPs, classification does shine light on the formation of LIPs. Bryan & Ernst (2005) have devised a simple ‘family tree’ for LIPs (Fig 1 in their web page on this site) which is useful in showing that SLIPs can be grouped and may share a petrogenic relationship which is the opposite of mafic VRMs. The pre-SLIP crust may have just been fertile! Then the question still remains;

  • Did the fertile crust melt due to decompressional melting caused by exhumation which in turn is caused by rifting?
  • Or, did a plume head supply the heat to melt the fertile crust and initiate the rifting?

Aside from composition it would be even more interesting to make a classification scheme where there are three branches based on available evidence (both physical and chemical; where possible) for and against plumes:

  1. Evidence strongly suggests a plume,
  2. Evidence for a plume is weak and top-down models fit better, and
  3. Both plume and top-down models share the distribution of evidence

Finally, I feel that the best way to move forward is to deal with each LIP as it comes. We must as scientists classify them to an extent, but to truly classify them, their petrogenic history is essential and more important.

  • SAUNDERS, A. D. 2005. Large Igneous Provinces: Origin and Environmental consequence. Elements. 1.
  • LESLIE B. YALE., SCOTT J. CARPENTER. 1998. Large igneous provinces and giant dike swarms: proxies for supercontinent cyclicity and mantle convection. Earth and Planetary Science Letters 163, 109–122.
  • BRYAN, SE., ERNST, R. 2006. Proposed revision to Large Igneous Province Classification.

15th September, 2006, Romain Meyer
Sami Mikhail writes above "What I see as evident is that no two LIPs are the same”. This conclusion is identical with Jason Morgan’s comment at the Great Plume Debate in Ft. William: “All mantle plumes are different“ (Final scientific report). However with such conclusions the scientific community can never come to a situation (as desired by Sami Mikhail) where we will be able “to make a classification scheme where there are three branches based on available evidence (both physical and chemical, where possible) for and against plumes”. The major LIP problems should not only be likened to a hypothesis, that every LIP is totally different, but can fully reflect a) a lack of observations, b) unknown processes, c) statistical significance of the available datasets, and …

I cannot exclude that some LIPs are different. That’s possible and can never be excluded. A valuable classification should be free of any model interpretation (because else we are again at the border where a model becomes reality for a part of the community and students)! A model can only be a best fit, to explain some observations. But will and can never be the reality, as it is a human mind product. A classification is only useful if it is based on founded OBSERVATIONS.

I based my classification on the location of the LIPs relative to the tectonic plates. As a result transitional LIPs are in no case linked to mineralogy/geochemistry! Transitional LIPs could also be (maybe better) named MARGIN LIPs, due to the fact that this group includes the volcanic rifted margins (e.g. the NAIP). The transition from continental to oceanic setting as volcanic rifted margins start as being continental and end as being oceanic. By the way Foulger (2006) postulated that maybe an extension of the Jan Mayen microcontinent could contribute to the Iceland crust. This reflects again the need for more data, before e.g. Sami can interpret this part of the NAIP as being fully oceanic.

Composition based classifications are classifying by the DOMINANT composition but as I showed in my NAIP example such a classification is quite difficult. Initial LIP magmatism is often silicic in continental settings due to crustal mantle interactions etc.. Such LIPs will have after erosion, a higher silicic proportion (and in the worst case switch into a silicic class). This is mainly a problem with older provinces!

An another question is “Have silicic LIPs always been mainly silicic?”

A first indication to the answer of this question may be found on the webpage "Proposed revision to Large Igneous Province classification". “Silicic LIPs are expected to have ” … “more mafic igneous underplating at lower crustal depths”. This dense mafic material is not buoyant enough to reach the surface, and as a result started to crystallize and/or began partially to differentiate into silicic magmas being able to rise again into higher crustal levels (today visible in outcrops)! However major parts of silicic LIPs are clearly MAFIC underplated bodies, being still today unreachable for sampling.

And so I totally agree with Sami Mikhail: “We must as scientists classify them to an extent, but to truly classify them, their petrogenic history is essential and more important.”