Roadmap | The review process | Home

The Emperor and Hawaiian Volcanic Chains: How well do they fit the plume hypothesis?

Kilauea Lava, Hetu Sheth

G. R. Foulger1 & Don L. Anderson2

1Dept. Earth Sciences, University of Durham, Durham DH1 3LE, U.K.

2Seismological Laboratory, California Institute of Technology, Pasadena, CA 91125

Click here for News & Discussion of this page

A note on Pele

Pele was born of the female spirit Haumea, or Hina, who, like all other important Hawai`i gods and goddesses, descended from the supreme beings, Papa, or Earth Mother, and Wakea, Sky Father. Pele was among the first voyagers to sail to Hawai`i, pursued, legends say, by her angry older sister, Na-maka-o-kaha`i because Pele had seduced her husband. Pele landed first on Kauai`i, but every time she thrust her o`o (digging stick) into the earth to dig a pit for her home, Na-maka-o-kaha`i, goddess of water and the sea, would flood the pits. Pele moved down the chain of islands in order of their geological formation, eventually landing on the Big Island's Mauna Loa, which is considered the tallest mountain on Earth when measured from its base at the bottom of the ocean (Betty Fullard-Leo, 1999;

Historical background

The distinctive northwest-southeast alignment of the Hawaiian chain was known to the early Hawaiians. Their legends clearly reveal that they recognized that the islands are progressively younger from the northwest to the southeast.

The first geologic study of the Hawaiian Islands (1840-1841) was directed by James Dwight Dana who deduced that the islands young to the southeast from the differences in their degree of erosion. He also suggested that some other island chains in the Pacific showed a similar general decrease in age from northwest to southeast.

The Hawaiian chain apparently consists of two strands of volcanoes located along distinct but parallel curving pathways. Multiple volcanoes line up to form each strand. Dana coined the terms Loa and Kea series for two of the prominent trends. The Kea trend includes the volcanoes of Kilauea, Mauna Kea, Kohala, East Maui (Haleakala) and West Maui. The Loa trend includes Lo`ihi, Mauna Loa, Hualalai, Mahukona (a submerged volcano), Kaho`olawe, Lana`i, and West Moloka`i (Figure 1). A pair of volcano trends may exist all the way along the Hawaiian and Emperor chains, though this is less clear amongst the older islands and seamounts [Clague & Dalrymple, 1987].

Figure 1: The Hawaiian islands and principle volcanoes.

The alignment of the Hawaiian Islands, Dana proposed, reflected localized volcanic activity along segments of a major fissure zone on the ocean floor. Dana's “great fissure” origin for the islands served as a working hypothesis for subsequent studies until the mid-20th century.

Figure 2: The Hawaiian leeward islands

In 1963, Tuzo Wilson pointed out that the time-progressive volcanism along the Hawaiian chain could be explained by the lithosphere moving across a stationary hot spot in the mantle [Wilson, 1963]. This suggestion gave rise, a decade later, to the theory of deep mantle plumes when Morgan [1971; 1972] proposed that a) this hot spot was continually supplied by a plume from the deep mantle, and b) there are approximately 20 such plumes in the mantle. Fixity relative to one-another, time-progressive volcanic tracks, and a high rate of volcanism were considered to be the primary characteristics of volcanic regions fuelled by deep-mantle plumes.

The Hawaiian - Emperor system appears superficially to fit the fixed deep mantle plume hypothesis well, and indeed, the hypothesis was inspired by it. The most compelling and widely quoted evidence today is still geometric considerations such as fixity, perceived parallelism with other volcanic chains and the regular time progression of volcanism, along with the high melt productivity. Lnear time-progression of volcanism and high magmatic productivity can be explained by other mechanisms such as propagating cracks and high mantle fertility. Nevertheless, for about 20 years there has been no serious challenge to the fluid-dynamic, deep thermal mantle plume hypothesis for the origin of the Emperor and Hawaiian islands and seamounts.

There are, however, a substantial number of aspects of the chains that are not predicted by the plume hypothesis and fit it poorly. These must give clues to alternative genesis models that would be interesting to test, and might satisfy the observations better. This review summarizes those, and other, aspects of the Emperor and Hawaiian chains.

Aspects of the Emperor and Hawaiian chains unpredicted by the plume hypothesis

1. The “bend” did not result from a change in direction of Pacific plate motion: The Emperor and Hawaiian chains differ in trend by about 60°. The two trends intersect at about 30°N, near the Mendocino fracture zone (Figures 3 & 4). It is often assumed that this resulted from a change in direction of motion of the Pacific plate at ~ 50 Ma. However, such a change in plate direction did not occur, as is shown by magnetic lineations, fracture zone orientations and plate motion reconstructions [e.g., Raymond et al., 2000]. This “nonevent” [Norton, 1995] is often attributed to the collision of India with Asia or to changes in circumpacific geometry [Gordon et al., 1978]. In any case, reorganisation of global plate movements resulting from continental collision is not expected to be a geologically instantaneous change from one stable state to another as would be required if the bend is to be explained in this way [Duncan & Richards, 1991; Lithgow-Bertelloni & Richards, 1995; Richards & Lithgow-Bertelloni, 1996]. Rapid and large changes in plate motion are not expected, given current understanding of plate forces. Convective forces can change instantaneously, but plate driving forces cannot.

Figure 3: Map of the Pacific ocean showing sea floor age as determined from magnetic lineations. It can be seen clearly from the continuity of fracture zones that no change in the direction of motion of the Pacific plate occurred at the time of the bend at ~ 50 Ma.

Figure 4: Locations of seamounts and volcanoes of various ages on the Emperor and Hawaiian chains, along with their their theoretical positions, calculated from reconstructions of global plate motions, assuming a fixed locaation of the “hotspot” relative to a fixed Indo-Atlantic hotspot reference frame [from Raymond et al., 2000].

In addition to the change in trend, the locus of volcanism moved south by ~ 800 km relative to the geomagnetic and biofacies reference frames while the Emperor Seamount chain formed. This conclusion is supported by data from bioclastic sediment on Emperor seamounts, which corresponds to higher latitudes than present-day Hawaii [Butt, 1980; McKenzie et al., 1980]. Prior to the bend, the hotspot migrated at ~ 7 cm/year relative to the Pacific sea floor. The rate of propagation changed at the time of the bend to ~ 9 cm/year.

The Hawaiian and Emperor chains appear thus to be separate phenomena in some sense. Their orientations might be controlled by the direction of regional stress, the fabric of the seafloor, or stresses caused by previously erupted volcanoes [Hieronymus & Bercovici, 1999]. Fluctuations also occur in the directions and volumes of individual chains and volcanoes [Clague & Dalrymple, 1987].

2. The Emperor chain began near a ridge: 86Sr/87Sr is uniform along most of the volcanic chain, but decreases to MORB-like values on approaching the ~ 80 Ma Detroit seamount, the oldest (and largest) of the dated Emperor seamounts. Pacific plate palinspastic reconstructions [Engebretson et al., 1985; Smith, 2003] (see Pacific page), and the elastic thickness of the lithosphere beneath the northern Emperors [Watts, 1978], are consistent with the Detroit seamount having been erupted on or very close to a spreading ridge.

Figure 5: Pacific bathymetry clearly shows the Emperor and Hawaiian volcanic chains.

This is a coincidence in the plume hypothesis, and is not unique to Hawaii. Many other hotspots are on ridges or their tracks started on ridges. The “primary hotspots” of Courtillot et al. [2003] are mainly on ridges, plate boundaries, triple junctions, or fracture zones near plate boundaries.

3. There is no Emperor/Hawaiian “plume head”: Flood-basalt provinces and oceanic plateaus are commonly thought to represent the initial stages of plume volcanism [Campbell & Griffiths, 1990; Campbell & Griffiths, 1993]. If the Hawaiian-Emperor chain began with such extensive volcanism, the plateau so produced is no longer present on the Pacific plate. Oceanic plateaus are not thought to be subductable, and thus an oceanic “plume head” for the Emperor chain would presumably have been scraped off, accreted or obducted onto the Aleutian/Kurile/Kamchatka arc. There is no evidence for such material associated with the Emperor chain although other accreted plateaus, e.g. Wrangellia in the Pacific NE, have been identified. Some alternative hypotheses, e.g., propagating cracks, do not require a large igneous province at the beginning of the chain. Many other hotspot tracks also lack a “plume head” (e.g., the Easter hotspot), and many oceanic plateaus and large igneous provinces on land lack an obvious “tail”, or “hotspot track” (e.g., the Siberian Traps).

4. The magmatic rate is highly variable: The eruption rate along the Emperor Seamount chain averaged ~ 0.01 km3/yr [Bargar & Jackson, 1974] (Figure 6). The eruption rate was very low, almost zero, for the initial ~ 5 Myr of the Hawaiian chain, but the average along the Hawaiian chain has been 0.017 km3/yr. Over the last ~ 6 Myr it has been higher than ever before, or 0.095 km3/yr, but the average rate for the last 1 Myr is even higher, or 0.21 km3/yr [Robinson & Eakins, 2006]. These rates may be compared with the current Pu'u 'O'o eruption rates of 0.113 km3/yr and a value also attributed to the shield stage of Hawaiian volcanism. The average magmatic production rate at a mid-ocean ridge spreading segment is ~ 0.02 km3/yr.

The eruption rate may correlate with the propagation rate of the melt locus, as this has approximately doubled over the last 2 Myr [Shaw et al., 1980; Clague & Dalrymple, 1987]. The southeast end of the Hawaiian chain is surrounded by a broad area of seafloor containing smaller-volume contemporaneous volcanism. These large variations in magmatic rate illustrate that the current magmatic rate is highly anomalous and should not be considered to be "typical" hot spot or plume behaviour. They are not currently explained.

Figure 6: Calculated volumes and magma supply rates. (A) Histogram along the Emperor Seamounts and Hawaiian Ridge at 5-my intervals (from Bargar and Jackson, 1974). Dotted lines mark the long-term average magma supply rates. (B) Histogram of the Hawaiian Islands at 1-my intervals. Intervals roughly correspond to individual island complexes and volcanoes: (a) Hawaii Island, (b) Maui Nui, (c) Koolau, (d) Waianae, (e) Niihau and Kaula, (f) Kauai. Gray bars show volumes and magma supply rates for this study; diagonal lines show results from Bargar and Jackson (1974). Both studies show the same trend of increasing magma supply rates towards the present (from Robinson & Eakins, 2006).

No quantitative model has explained the flux rate and its variation at Hawaii and along the Hawaiian and Emperor chains. No thermal model has explained how high flux rates can occur beneath thick plates, where the top of the productive part of the melting column is missing. The plume-head/plume-tail hypothesis predicts a large initial rate that declines subsequently, the opposite of that observed along the Hawaiian chain.

5. There is no heatflow anomaly: Lithosphere underlain by a thermal plume is expected to be thinner and to have higher heat flow than the average for lithosphere of the same age elsewhere. Heatflow across the Hawaiian swell, however, shows no significant anomaly [von Herzen et al., 1989; Stein & Stein, 1992,1993; McNutt, 2000] (see Heatflow page). The bathymetric swell surrounding the southernmost part of the Hawaiian chain cannot therefore be explained as a thermal effect, and in the plume model must be attributed to dynamic or compositional effects with no discernible thermal effect shallower than the base of the ~ 80-km-thick lithosphere [Liu & Chase, 1989; Sleep, 1994] (Figure 7). Anomalous heat flow and thermal rejuvenation (i.e., thinning of the plate) are predicted by the plume hypothesis [Crough, 1983] but not by some alternative hypotheses e.g., propagating cracks.

Figure 7: Intermediate wavelength geoid anomalies near the Hawaiian chain.

6. Mantle temperature is elevated by up to ~ 200°C: Petrology can be used to infer the temperature where surface-erupted melts were last in equilibrium with the mantle source, and Tp, the mantle potential temperature. (see Temperature page) The results for Hawaii are variable, and whether or not a high temperature anomaly is inferred depends on what “average” mantle temperature is considered to be. “Average” mantle temperature is mostly studied at mid-ocean ridges (MORs), since this is where magma of mantle origin is mostly erupted. The temperatures required for plumes to rise are debated. Peridotic plumes from the core-mantle boundary would require temperature anomalies of the order of 600 K [Watson & McKenzie, 1991; Cordery et al., 1997]. Temperature anomalies of ~ 250 K have been suggested for crust-contaminated plumes rising from the transition zone [e.g., Courtney & White, 1986], but it is debatable whether such a phenomenon is physically possible as the base of the upper mantle is not thought to be a thermal boundary layer.

Combinations of geophysical and geochemical arguments (e.g., “garnet signature” in rare-Earth elements, olivine-liquid geothermometers have been used to infer temperatures beneath Hawaii of 150 - 200 K higher than MORs. Also, study of the MgO contents of picritic glass found on the Puna ridge (the submarine part of Kilauea's east rift zone) indicate that if they represent primary melts from a lherzolite source in the presence of H2O and CO2, their minimum potential temperature of generation was ~ 1420°C [Gudfinnsson & Presnall, 2002]. This suggests formation at a potential temperature 140-160°C hotter than at MORs if an average Tp of 1260-1280°C is assumed [e.g., McKenzie & Bickle, 1988; Presnall et al., 2002].

This conclusion is dependent on the “average” mantle Tp assumed. Study of basalt petrogenesis by high-pressure experimental methods, in which both the progressive melting behaviour of peridotite and (C-H-O) is investigated, has been applied to both Hawaii and MORs. The results, coupled with data from olivine phenocrysts in Hawaiian picrites, suggest temperatures of formation of ~ 1325°C for both Hawaiian and MOR magmas, and a temperature difference no more than ~ 20 K [Green et al., 2001] (see Mantle temperature page).

Hawaii is the only oceanic island where picrite glass has been found. It is not abundant, however, and is only known from a few grains of turbidite sand [Clague et al., 1991; Clague et al., 1995]. The presence of picrite glass suggests that Hawaii is hotter than localities that lack picrite glass. However, Hawaii is better sampled than almost any other oceanic hotspot on Earth, and given the extreme rarity (one location) of this lithology there, it is not safe to conclude from this that all other chains are cooler.

The bottom line today is that the majority of studies suggest that there is a temperature anomaly beneath Hawaii today, compared with MORs, but it is less than 200 K, which is somewhat less than that generally assumed to be required in the coolest of plumes. The temperature variations in the mantle expected from normal plate tectonic processes are ± 180ÚC [Kaula, 1983].

Lava at Kilauea. Photo by Hetu Sheth
7. The melt originates from the shallow asthenosphere: The petrology of Hawaiian lavas requires that much of the melt was last in equilibrium with its mantle host in the shallower part of the garnet stability field. It rules out an origin deep in the garnet stability field, meaning that the melt probably comes from ~ 80 – 120 km depth, or near the base of the lithosphere. (e.g., T. Sisson, unpublished results) The greater the depth of melting and the larger the melt drainage area to Hawaiian volcanoes, the larger the potential melt volume, for whatever mechanism is proposed.
An isolated one-dimensional volcanic feature such as Hawaii can drain a volume larger than a two-dimensional structure such as a ridge, which must share its magma sources with adjacent ridge segments.

8. OIB geochemistry is ambivalent regarding depth of origin: Ocean-island basalt (OIB) geochemistry by itself does not require a mantle plume, or a deep or hot source. The ~ 80 – 120 km depth of melting indicates that the immediate source of the “enriched” signature of OIB is the shallow mantle. What is disputed is the ultimate origin of material - whether it is the deepest mantle or the mantle wedge, the asthenosphere or a shallower layer which collects subduction zone products. Hawaiian lavas exhibit both enriched and depleted signatures showing chemical similarities to continental and marine sediments, ocean crust and ridge basalts. This collection of signatures defines OIB geochemistry, two of the main characteristics of which are a large statistical variance in isotopic ratios, and rapid lateral and temporal differences.

OIB geochemistry has been attributed to the incorporation of subducted oceanic crust or sediments into the mantle source [Hofmann & White, 1982]. In the deep-plume model, these materials are carried to the core-mantle boundary, through the homogeneous depleted upper mantle, and then they ascend all the way back up in the core of the plume. This mechanism is inconsistent with the observation that OIB is widespread throughout the Pacific, occurring on thousands of seamounts where plumes are neither expected nor reasonable to postulate.

The geochemistry of lavas from the Hawaiian and Emperor chains changes:

  • geographically, where the chain crosses fracture zones, e.g., the Mendocino, Murray and Molakai fracture zones [Basu & Faggart, 1996], mainly by increase in the variance,
  • temporally, e.g., Mukhopadhyay et al. [2003] report variations in 3He/4He of up to 8 Ra during a single century in Kauai volcano, that correlate with variations in radiogenic isotope ratios. This suggests rapid changes in composition or contamination of the melt supplying the magma reservoir, and
  • spatially, from volcano to volcano, such that different volcanoes do not appear to be fed by the same magma source.

“End-member” and principal-component interpretations require at least four different source components to explain the geochemistry of Hawaiian lavas. This result, along with the variability described above, suggests a spatially distributed, compositionally inhomogenous and temporally variable source that is also sensitive to shallow lithospheric features. A chemically heterogeneous shallow mantle, due to recycling [Meibom & Anderson, 2003] or a zoned plume stem have been proposed.

The maximum 3He/4He observed at Hawaii is 35 Ra in samples from Loihi [Graham, 2002]. Such high maximum ratios are often cited as unambiguous evidence for lower-mantle affinity and a deep-mantle plume. This model is disputed (see Helium Fundamentals page).

9. Seismology has not detected a plume: The most promising seismic methods for detecting plume-like structures are:

  • teleseismic tomography, which can image the three-dimensional wave-speed in the upper few hundred kilometers,
  • receiver functions (waves converted at upper mantle discontinuities), which can reveal topography on the 410- and 650-km discontinuities, and the thickness of the transition zone, and
  • the study of ScS multiples. (see Seismology & Transition zone pages)

A hot column of mantle should reveal low wave speeds wherever earthquake waves cross it, and should depress the 410 km discontinuity and elevate the 650 km discontinuity, because of the effect of temperature on phase boundaries. A large active upwelling would spread out below the lithosphere, and be dragged to the NW. Actual tomography shows extension of the low-velocity zone to the SE, toward the East Pacific Rise.

Early teleseismic tomography on the big island of Hawaii revealed little structure in the upper ~ 80 km, and poorly resolved lower velocities below this [Ellsworth, 1977]. An experiment using a network ~ 600 km long, involving digital seismic stations on several of the islands detected a low-wave-speed anomaly beneath the islands of Maui and Molokai, 250 km NW of the big island, but no low-wave-speed anomaly beneath the big island down to the maximum depth of good resolution there at ~ 150 km [Wolfe et al., 2002] (Figure 8).

Figure 8: Seismic anomalies beneath the Hawaiian island archipelego at 150 km depth [from Wolfe et al., 2002].

In order to probe the upper mantle in detail throughout much of its thickness, a ~ 600-km-wide network including many ocean-bottom seismometers is required. Such an experiment is currently underway (see

Tomography in the central Pacific can only provide relatively good resolution of upper-mantle wave speeds if surface waves and multiple ScS waves are used, and can resolve structures on a spatial scale of a few hundred kilometers. Cross sections through the best available model of this sort - S20RTS [Ritsema et al., 1999] are shown in Figure 9.

Figure 9: Six cross sections passing through the big island of Hawaii and the whole-mantle tomography model of Ritsema et al. [1999]. The big island is exactly in the centre of the cross sections.

A strong low-wave-speed anomaly lies to the NW of the big island of Hawaii, which might be associated with that detected by Wolfe et al. [2002]. There is apparent continuity of low-wave-speed material between the surface in the region of Hawaii and the core-mantle boundary beneath a large swath of the Earth ranging from the New Hebrides and Samoa, throughout the south Pacific and north along the East Pacific Rise, depending on the line of cross section selected. These connections involve various anomaly contortions and tilts. From the standpoint of the whole-mantle plume model, the problem lies in choosing which of many candidate anomalies to advocate, rather than explaining why one is not detected, as is the case for other hotspots such as Iceland and Tristan da Cunha.

A magnitude 6.2 earthquake occurred on the big island in 1973, and generated shear (S) waves that travelled directly downwards, reflected from the boundary of the core and, upon their return, registered on a seismometer on the big island (ScS waves). Six multiple echoes were well recorded. The times of these waves indicate that the average S-wave speed of the mantle beneath the Hawaii region [Katzman et al., 1988] is somewhat higher than the average beneath the southwestern Pacific Ocean [Best et al., 1975], and that the propagation efficiency is high, contrary to expectations for regions of high temperature or partial melting. (see Julian & Foulger, 2003).

Transition zone thickness beneath the Hawaii region is estimated to be 229 km [Gu & Dziewonski, 2001]. This is ~ 13 km thinner than the global average of 242 km, but not significantly thinner than the transition zone throughout much of the central Pacific and other oceans (Figure 10) (see also Transition zone page).

Figure 10: Thickness of the transition zone. Residuals are interpolated using a spherical harmonic expansion up to degree 12. The long-wavelength features are fairly consistent among these maps, which are dominated by low-degree harmonics [from Gu & Dziewonski, 2001].

In summary, the results of seismic studies of the mantle beneath the Hawaiian region to date suggest that:

  1. the Pacific as a whole, especially the south Pacific, is a region of anomalously low wave speeds on a global scale,
  2. low wave-speed anomalies at the surface in the Hawaii region are continuous down to the core-mantle boundary along numerous paths that embrace almost the entire south Pacific,
  3. ironically, the average wave speed immediately beneath the Hawaiian island region itself is not unusually low, and
  4. the transition zone is somewhat thinner than the global average but typical of the central Pacific as a whole.

Tomography & Geochemistry

Variations in the geochemistry of Hawaiian lavas have been interpreted in the context of a zoned plume model. This approach predicts a plume less than 50 kilometers in diameter [DePaolo et al., 2001] (Figure 11). Such a plume would be undetectable seismically, except in the upper 50-100 km, where a low-wave-speed anomaly is not detected in tomographic studies. If such a plume were downward-continuous with a similar diameter, it would be too small to be detected by teleseismic tomography using a broad ocean-bottom seismic network.


Figure 11: Geochemical "plume maps" (from DePaolo et al., 2001)

Other possible constraints or coincidences

  • The Hawaiian chain is bounded by a large gravitational moat and arch structure that is most strongly evident in bathymetry around its southeastern end (Figure 7).
  • Some Hawaiian basalts and xenoliths have extreme 3He/4He ratios compared with the MORB mean. In common with other “hotspots” these also exhibit high variance and have lower 3He abundances than MORB glasses.
  • The volcanoes of the Hawaiian chain, and possibly also the Emperors, feature long, nested, lateral rift zones which are rarely seen on other oceanic islands (see Samoa page).
  • Hawaiian volcanoes are distributed along the paired Loa and Kea trends that comprise arrays of sinusoidal volcano loci. These trends appear to be controlled by the loads of the volcanic edifices [Hieronymus & Bercovici, 1999].
  • The volcanic constructs (groups of volcanoes) along the chains are mostly oriented at a high angle to the overall trend of the volcanic chain. The orientations of the rift systems are mostly oblique to structure on the underlying seafloor such as ridge segments and transform faults.
  • Hawaii is about as far away from other active volcanic chains, hotspots, plate boundaries of all types and continents as it is possible to be, very close to the center of the colossal Pacific plate.
  • The oldest end of the Emperor chain is subducting almost exactly into the cusp of the Aleutian/Kurile trench – the “NW Pacific pocket”. The cause and effect of this are debated.
  • The great bend occurs close to where the chain crosses the Mendocino fracture zone.


What, then, is the evidence in support of a deep mantle plume? The Hawaiian hotspot has a very regular time-progressive track. Clearly, during the last ~ 50 Myr the melt anomaly has remained approximately stationary with respect to the North American plate, beneath which the Pacific plate is subducting at the Aleutian trench. (Since the Pacific plate is moving much faster than any other plate on Earth, the Hawaiian hotspot is essentially fixed with respect to every other plate also.) The magmatic rate is currently very high, and is characterised by OIB geochemistry. This includes high maximum 3He/4He isotope ratios, commonly considered to be a “smoking gun” for a plume from the lower mantle.

Alternative mechanisms have been proposed for all these observations (see also Cracks & Stress page). Other Pacific volcanic chains roughly parallel to the Hawaiian chain may (e.g., the eastern end of the Louisville chain) or may not (e.g., the Line Islands) be time-progressive. High rates of melt production have been explained by fertile mantle composition, perhaps resulting from remelting subducted oceanic crust, and this would also explain OIB geochemistry. Shallow models for high 3He/4He isotope ratios, e.g., release of old helium stored for 100s of Myr in subducted, lower-crustal, olivine-rich cumulates, have also been proposed and not challenged to date [e.g., Anderson, 1998a; Anderson, 1998b; Meibom et al., 2003] (see Helium Fundamentals page). However, a persuasive, alternative model for the Emperor and Hawaiian volcanic chains remains to be fully quantified.

Any satisfactory theory for Hawaiian volcanism must explain (or rationalize) the:

  • change in migration direction of the melting locus at the bend,
  • association of the great bend with the Mendocino fracture zone,
  • change in migration rate at the bend,
  • apparent commencement of the volcanic chain near a ridge,
  • absence of a “plume head”,
  • large variations in magmatic production, and a current magmatic rate about 3 times greater than the next most productive hotspots,
  • absence of a significant heat flow anomaly,
  • absence of lithospheric thinning,
  • absence of a strong high-temperature signal in the erupted basalts,
  • production of very large volumes of magma even though the depth to the top of the melting column is exceptionally large compared with MORs,
  • spatial and temporal variation in the composition of erupted lavas on a variety of scales,
  • remote location of Hawaii, near the center of a very large plate,
  • location of the oldest end of the chain with respect to the “Pacific pocket”,
  • unique rift zones,
  • paired Loa and Kea trends,
  • seismic whole-mantle mantle structure that is apparently normal compared with the Pacific ocean elsewhere, and
  • occurrence of a bathymetric swell (a moat and “arch”) along the eastern two-thirds of the Hawaiian chain and wrapping around its southeastern end, with alkalic basaltic volcanism occurring at some places along it.

In conclusion, Hawaii is not fully explained by any current hypothesis. It is impressive that a region of the Earth so extensively studied for so many years, by so many Earth scientists with so many techniques could remain so intransigent to full understanding. Many of the numerous features that are not yet fully understood, and the parameters of alternative hypotheses, are not currently being studied, but they offer exciting research opportunities.

Acknowledgments: We gratefully acknowledge the input and help of Jim Natland, Dean Presnall, Bruce Julian, Warren Hamilton and Seth Stein in drafting this web page. This acknowledgement does not necessarily imply agreement with its contents. We also thank Barry Eakins, Joanne Stock, Jim Natland and Jeroen Ritsema for supplying figures.

News & Discussion

"Fixed" Hawaiian hotspot not fixed


  • Bargar, K.E., and E.D. Jackson, Calculated volumes of individual shield volcanoes along the Hawaiian-Emperor chain, U.S. Geological Survey Journal of Research, 2, 545-550, 1974.
  • Basu, A.R., and Faggart, Temporal Variation in the Hawaiian mantle plume: The Lanai anomaly, the Molokai fracture zone and a seawater-altered lithospheric component, in Hawaiian volcanism, in Earth Processes: Reading the Isotopic Code, edited by A. Basu, and S. Hart, pp. 149-159, 1996.
  • Best, W.J., L.R. Johnson, and T.V. McEvilly, ScS and the mantle beneath Hawaii, EOS Trans. AGU, 1147, 1975.
  • Butt, A., Biostratigraphic and paleo-environment analyses of the sediments at the Emperor Seamounts, DSDP Leg 55, Northwestern Pacific: Cenozoic foraminifers, in Initial Reports, DSDP, edited by E.D. Jackson, I. Koizumi, and others, pp. 289-325, U.S. Government Printing Office, Washington, 1980.
  • Campbell, I.H., and R.W. Griffiths, Implications of mantle plume structure for the evolution of flood basalts, Earth planet. Sci. Lett., 99, 79-93, 1990.
  • Campbell, I.H., and R.W. Griffiths, The evolution of the mantle's chemical structure, Lithos, 30, 389-399, 1993.
  • Clague, D.A., and G.B. Dalrymple, The Hawaiian-Emperor volcanic chain, in Volcanism in Hawaii, edited by R.W. Decker, T.L. Wright, and P.H. Stauffer, pp. 5-54, U.S. Government Printing Office, 1987.
  • Clague, D.A., J.G. Moore, J.E. Dixon, and W.B. Friesen, Petrology of submarine lavas from Kilauea's Puna ridge, Hawaii, J. Pet., 36, 299-349, 1995.
  • Clague, D.A., W.S. Weber, and J.E. Dixon, Picritic glasses from Hawaii, Nature, 353, 553-556, 1991.
  • Cordery, M.J., G.F. Davies, and I.H. Campbell, Genesis of flood basalts from eclogite-bearing mantle plumes, J. geophys. Res., 102, 20,179-20,197, 1997.
  • Courtillot, V., A. Davaillie, J. Besse, and J. Stock, Three distinct types of hotspots in the Earth's mantle, Earth planet. Sci. Lett., 205, 295-308, 2003.
  • Courtney, R.C., and R.S. White, Anomalous heat flow and geoid across the Cape Verde Rise: evidence fro dynamic support from a thermal plume in the mantle, Geophys. J.R. Astron. Soc., 87, 815-867, and microfiche GJ 87/1, 1986.
  • Crough, S.T., Hotspot swells, Ann. Rev. Earth Planet. Sci., 11, 165-193, 1983.
  • DePaolo, D.J., J.G. Bruce, A. Dodson, D.L. Shuster, and B.M. Kennedy, Isotopic evolution of Mauna Loa and the chemical structure of the Hawaiian plume, Geochemistry, Geophysics, Geosystems, 2, 2000GC000139, 2001.
  • Duncan, R.A., and M.A. Richards, Hotspots, mantle plumes, flood basalts, and true polar wander, Rev. Geophys., 29, 31-50, 1991.
  • Ellsworth, W.L., Three-dimensional structure of the crust and mantle beneath the island of Hawaii, Ph.D. thesis, Massachusetts Institute of Technology, 1977.
  • Engebretson, D.C., A. Cox, and R.G. Gordon, Relative motions between oceanic and continental plates in the Pacific basin, in Geol. Soc. Am. Spec. Pap., pp. 1-59, 1985.
  • Gordon, R.G., A.A. Cox, and C.E. Harter, Absolute motion of an individual plate estimated from its ridge and trench boundaries, Nature, 274, 752-755, 1978.
  • Graham, D.W., Noble gas isotope geochemistry of mid-ocean ridge and ocean island basalts: Characterization of mantle source reservoirs, Noble gases in Geochem. Cosmochem., Rev. Min. Geochem., 47, 247-317, 2002.
  • Green, D.H., T.J. Falloon, S.M. Eggins, and G.M. Yaxley, Primary magmas and mantle temperatures, Eur. J. Min, 13, 437-451, 2001.
  • Hieronymus, C.F., and D. Bercovici, Discrete alternating hotspot islands formed by interaction of magma transport and lithospheric flexure, Nature, 397, 604-607, 1999.
  • Hofmann, A.W., and W.M. White, Mantle plumes from ancient oceanic crust, Earth planet. Sci. Lett., 57, 421-436, 1982.
  • Katzman, R., L. Zhao, and T.H. Jordan, High-resolution, two-dimensional vertical tomography of the central Pacific mantle using ScS reverberations and frequency-dependent travel times, J. geophys. Res., 103, 17,933-17,971, 1988.
  • Kaula, W.M., Minimal upper mantle temperature variations consistent with observed heat flow and plate velocities, J. geophys. Res., 88, 10,323-10,332, 1983.
  • Lithgow-Bertelloni, C., and M.A. Richards, Cenozoic plate driving forces, Geophys. Res. Lett., 22, 1317-1320, 1995.
  • Liu, M., and C.G. Chase, Evolution of midplate hotspot swells - numerical solutions, J. geophys. Res., 94, 5571-5584, 1989.
  • McKenzie, D., and J. Bickle, The volume and composition of melt generated by extension of the lithosphere, J. Pet., 29, 625-679, 1988.
  • McKenzie, J.A., D. Bernoulli, and S.O. Schlanger, Shallow-water carbonate sediments from the Emperor Seamounts: Their diagenesis and paleogragraphic significance, in Initial Reports, DSDP, edited by E.D. Jackson, I. Koizumi, and others, pp. 415-455, U.S. Government Printing Office, Washington, 1980.
  • Morgan, W.J., Convection plumes in the lower mantle, Nature, 230, 42-43, 1971.
  • Morgan, W.J., Deep mantle convection plumes and plate motions, Bull. Am. Assoc. Pet. Geol., 56, 203-213, 1972.
  • Mukhopadhyay, S., J.C. Lassiter, K.A. Farley, and S.W. Bogue, Geochemistry of Kauai shield-stage lavas: Implications for the chemical evolution of the Hawaiian plume, Geochemistry, Geophysics, Geosystems, 4, 10.1029/2002GC000342, 2003.
  • Norton, I.O., Plate motions in the North Pacific; the 43 Ma nonevent, Tectonics, 14, 1080-1094, 1995.
  • Presnall, D.C., G.H. Gudfinnsson, and M.J. Walter, Generation of mid-ocean ridge basalts at pressures from 1 to 7 GPa, Geochim. Cosmochim. Acta, 66, 2073-2090, 2002.
  • Raymond, C.A., J.M. Stock, and S.C. Cande, Fast Paleogene motion of the Pacific hotspots from revised global plate circuit constraints, in History and Dynamics of Plate Motions, edited by M.A. Richards, R.G. Gordon, and R.D. van der Hilst, pp. 359-375, 2000.
  • Richards, M.A., and C. Lithgow-Bertelloni, Plate motion changes, the Hawaiian-Emperor bend, and the apparent success of dynamical models, Earth planet. Sci. Lett., 137, 19-28, 1996.
  • Ritsema, J., H.J. van Heijst, and J.H. Woodhouse, Complex shear wave velocity structure imaged beneath Africa and Iceland, Science, 286, 1925-1928, 1999.
  • Shaw, H.R., E.D. Jackson, and K.E. Bargar, Volcanic periodicity along the Hawaiian-Emperor chain, Am. J. Sci., 280-A, 667-708, 1980.
  • Sleep, N.H., Lithospheric thinning by midplate mantle plumes and the thermal history of hot plume material ponded at sublithospheric depths, J. geophys. Res., 99, 9327-9343, 1994.
  • Smith, A.D., A re-appraisal of stress field and convective roll models for the origin and distribution of Cretaceous to Recent intraplate volcanism in the Pacific basin, Int. Geol. Rev., 45, in press, 2003.
  • Stein, C.A., and S. Stein, A model for the global variation in oceanic depth and heat-flow with lithospheric age, Nature, 359, 123-129, 1992.
  • Stein, C.A. and Stein, S., Constraints on Pacific midplate swells from global depth-age and heat flow-age models, in The Mesozoic Pacific: Geology, Tectonics, and Volcanism, Geophysical Monograph. American Geophysical Union, Washington, D.C., pp. 53-76, 1993.
  • von Herzen, R.P., M.J. Cordery, R.S. Detrick, and C. Fang, Heat-flow and the thermal origin of hot spot swells - the Hawaiian swell revisited, J. geophys. Res., 94, 13,783-13,799, 1989.
  • Watson, S., and D. McKenzie, Melt generation by plumes: a study of Hawaiian volcanism, J. Pet., 32, 501-537, 1991.
  • Watts, A.B., An analysis of isostasy in the world's oceans; 1, Hawaiian-Emperor seamount chain, J. geophys. Res., 83, 5989-6004, 1978.
  • Wilson, J.T., A possible origin of the Hawaiian Islands, Can. J. Phys., 41, 863-870, 1963.
  • Wolfe, C.J., S.C. Solomon, P.G. Silver, J.C. VanDecar, and R.M. Russo, Inversion of body-wave delay times for mantle structure beneath the Hawaiian islands: results from the PELENET experiment, Earth planet. Sci. Lett., 198, 129-145, 2002.
last updated 11th March, 2006