Pacific Hotspots Research Articles

Geoarchaeological context of Holocene subsidence at the Ferry Berth Lapita site, Mulifanua, Upolu, Samoa

The 1973 discovery of an underwater archaeological site during dredging for a ferry landing at Mulifanua on Upolu raised important unanswered questions about the prehistory of Samoa, particularly the evolution of Holocene shorelines and relative local sea levels. A cultural horizon yielding Lapita potsherds, the only decorated Lapita assemblage yet found in Samoa and dating to ca. 2.8 ka, lies at a depth of 2.25 m below modern sea level beneath a capping of cemented paleobeachrock. With fluctuating hydro-isostatic sea level taken into account, the sherd occurrence implies subsidence of a former coastline by ca. 4 m at a mean rate of 1.4 mm/yr. Shoreline features on both Upolu and nearby Savai'i are fully compatible with bulk Holocene subsidence. We attribute the observed subsidence to downflexure of the lithosphere from volcano loading centered on the Savai'i locus of historic volcanism, and conclude that any other Lapita sites that may exist in Samoa have subsided by a comparable amount. Although the Samoan linear volcanic chain resembles other Pacific hotspot tracks where active volcano loading is confined to their southeastern ends, the most voluminous Holocene eruptions in Samoa have occurred on Savai'i at the northwestern end of the exposed island chain. Samoan volcanism has evidently been influenced by lateral flexure of the Pacific plate as it moves past the northern extension of the Tonga subduction zone, and the active volcanism is apparently controlled by a longitudinal rift, which transects both Upolu and Savai'i and is superimposed upon older volcanic edifices that may record earlier hotspot volcanism. Early archaeological sites in American Samoa display a variable record of subsidence and possible uplift apparently related to volcano loading in the Manu'a Islands and possibly to the location of Tutuila in a position to be affected by uparching of lithosphere between downflexures beneath more active volcanic islands to the east and west. Bulk subsidence of Upolu and Savai'i may be the fundamental cause of damaging coastal erosion in modern Samoa. © 1998 John Wiley & Sons, Inc.

Paleomagnetic evidence for motion of the Hawaiian hotspot during formation of the Emperor seamounts

The bend in the Hawaiian-Emperor chain is the best example of a change in plate motion recorded in a fixed-hotspot frame of reference. Alternatively, the bend might record primarily differences in motion of the Hawaiian hotspot relative to the Pacific lithosphere. New paleomagnetic data from the Emperor chain support the latter view. Althouth the rate of motion is difficult to constrain because of uncertainties posed by true polar wander and limited sampling of the chain, the best available paleomagnetic data suggest Pacific hotspots may have moved at rates comparable to those of lithospheric plates (> 30 mm yr −1) in late Cretaceous to early Tertiary times (81-43 Ma). If correct, this requires a major change in how we view mantle dynamics and the history of plate motions. In the early to mid-Cretaceous (128-95 Ma), hotspots in the Atlantic moved at similar rates. These episodes during which groups of hotspots appear to move rapidly are separated by times of much slower motion, such as the past 5 m.y.

The Petrogenetic Evolution of Lavas from Easter Island and Neighbouring Seamounts, Near-ridge Hotspot Volcanoes in theSE Pacific

Major and trace element, mineralogical and petrographical data are presented for submarine and subaerial lavas from Easter Island and two neighbouring seamounts. The samples can be divided into three groups based on their major element composition and incompatible element enrichment, typified by their (La/Sm)N ratios. Tholeiitic samples with (La/Sm)N of ∼1.2 are comparable with lavas constituting three young volcanic fields closer to the spreading axis. The other two lava series from Easter Island and the two seamounts are transitional to slightly alkaline, one having an intermediate enrichment with (La/Sm)N of 1.5–2, the other being even more enriched [(La/Sm)N ∼2.3]. Large phenocrysts and xenocrysts of olivine and plagioclase in the submarine lavas indicate a crustal reservoir with crystallizing magma. Compared with the relatively mafic submarine samples, the more fractionated lavas on Easter Island suggest the presence of another, more shallow magma chamber. The extreme differentiation of rhyolites and trachytes probably occurred as a result of decreasing magma supply during waning phases of volcano activity. Each volcano of the Easter Hotspot probably evolved through at least two stages. A tholeiitic stage appears to form large volcanoes at depths <1000 m and is succeeded by a transitional group of lavas which can be observed on Easter Island. A third stage with the most enriched lavas may form small eruptive centres on the flanks of older volcanoes or as flank cones on the seafloor. The degree of partial melting decreases from the tholeiitic to the most enriched alkaline series. The occurrence of different stages of volcanism at the Easter Hotspot not only resembles the magmatism of other Pacific hotspots such as Hawaii but also the volcanism of non-plume near-ridge seamounts at the East Pacific Rise. We suggest that the development of each magmatic stage depends on the age of the overlying plate, which determines (1) the amount of depleted mid-ocean ridge material mixed into the magma source and (2) the degree of partial melting.

Pacific Plate motion and undulations in geoid and bathymetry

Previous studies have shown that the Pacific geoid and gravity fields exhibit lineated anomalies, trending approximately in the direction of absolute plate motion over the underlying mantle. Because the undulations obliquely cross fracture zones they have often been attributed a convective origin. Recently, lithospheric boudinage caused by diffuse extension has been proposed as a possible mechanism. We have examined the undulations in the free-air anomalies, geoid and bathymetry over a portion of the Pacific Plate to determine quantitatively how the undulations are related to plate motion. We compare the observed data to an axisymmetric, sinusoidal undulation defined in an arbitrary frame of reference; in particular, we seek the north pole of this reference frame that maximizes the correlation between data and model. Poles that are close to the Pacific hotspot pole represent copolar undulations possibly related to plate motion. The distance between the best-fitting poles and the hotspot pole is determined as a function of undulation wavelength and reveals several minima (with distance < 10°) for discrete geoid wavebands centered on wavelengths of 160 km, 225 km, 287 km, 400 km, 660 km, 850 km, 1000 km and 1400 km. Bathymetry data have copolar bathymetric expressions as well, giving an implied admittance of 2–3 m/km. The most co-polar geoid/bathymetry undulations (with poles within 2–3° of the average Pacific Euler pole) have wavelengths of ∼ 280 km and ∼ 1050 km, respectively. The latter could have a convective origin or be related to the spacing of hotspot swells. The former may reflect lithospheric boudinage formed in response to diffuse extension, but could also have a dynamic origin since flexural dampening may only have attenuated the bathymetric amplitude by 50% or less. Radiometric dating of volcanic ridges found in the troughs of prominent gravity lineations gives ages that correlate well with documented changes in Pacific and Indo/Australian Plate motion, suggesting the ridges formed in response to intermittent plate boundary stresses and not as a direct consequence of small-scale convection or diffuse extension.

Large-scale motion between Pacific and Atlantic hotspots

STUDIES of true polar wander (TPW), the rotation of the solid Earth with respect to the spin axis1, have suggested that there has been 10–15° of relative motion over the past 130 Myr (refs 2–4). In such studies, the orientation of the spin axis is recovered from continental palaeomagnetic poles (corrected for relative plate motions), and compared with a deep-mantle reference frame defined by hotspot locations. But deducing relative plate motions becomes increasingly difficult for older (Mesozoic) time periods, hindering tests of TPW on timescales comparable to those of large-scale mantle convection; moreover, the assumption of hotspot fixity is controversial5,6. We examine here a more direct approach7,8, using palaeolatitudes derived from Pacific guyots. Contrary to predictions from TPW models, these data suggest only minor latitudinal shifts of Pacific hotspots during the Cretaceous period. Instead of TPW, relative motion between the Atlantic and Pacific hotspot groups9 is required at a velocity of approximately 30 mm yr−1, more than 50% larger than previously proposed5.

The possible reflection of mantle discontinuities in Pacific geoid and bathymetry

Geoid anomalies over the Pacific plate show lineated undulations approximately oriented in the direction of absolute plate motion. Conventional spectral analyses have revealed a broad range of dominant wavelengths, but as filtering is direction‐dependent, results are difficult to interpret. Here, we present a new approach designed to quantify the correlation between the geoid undulations and Pacific plate motion. We calculate the spatial correlation between geoid lineations and arbitrarily oriented, axisymmetric sinusoidal undulations of given wavelength. The pole ( i.e., the orientation) of the sinusoid which maximizes the correlation is determined. The distance between this pole and the Pacific hotspot pole is calculated and represents a measure of how well a geoid lineation of a given wavelength aligns with Pacific plate motion. This distance varies with undulation wavelength and has discrete minima for several wavebands. These minima, moreover, occur at wavelengths whose values are remarkably close to the depths of seismically inferred mantle discontinuities. A similar analysis of available bathymetry corroborates these findings. If these correlations are significant, they suggest that geoid (and bathymetry) undulations are influenced by mantle structure and thus are likely to reflect mantle dynamics.

Manganese, methane, iron, zinc, and nickel anomalies in hydrothermal plumes from Teahitia and Macdonald volcanoes

Hydrothermal vents discharging turbid water have recently been discovered at Teahitia and Macdonald seamounts, which are situated at the southeast ends of two South Pacific hotspot systems. Mixtures of the plumes with ambient seawater were sampled during the 1989 CYAPOL cruise and found to contain significant amounts of CH 4, Mn, Fe, Zn, and Ni, which are comparable in concentration to those found over hydrothermal fields on the EPR and in megaplumes over the Juan de Fuca Ridge. Above Teahitia volcano, CH 4 shows maximum concentrations of 8 nmol/kg, Mn concentrations of 19 nmol/kg, and Fe concentrations of 125 nmol/kg. The concentrations were much higher during an earlier SO-47 cruise in 1987, reaching maximum values of 14 nmol CH 4/kg and 60 nmol Mn/kg. By contrast, the Macdonald seamount plume shows an opposite trend reaching very high concentrations of 350 nmol/kg CH 4, 250 nmol Mn/kg, and 3920 nmol Fe/kg during the CYAPOL cruise when the volcano was at an eruptive stage with high seismic activity. The Fe Mn ratio of the plume is variable between the two active volcanoes and at different hydrocast stations, reflecting the balance between input and removal of these two elements. Mapping of the element concentrations in the water column shows the plume ceiling, or anomaly maxima, to occur to several hundred metres above the seafloor and up to several kilometres (> 5 km) from the vents on Macdonald seamount. Multiple maxima in hydrocast profiles indicate the presence of several distinct hydrothermal vents at different water depths on Teahitia and on the western flank of Macdonald seamount. REE data from a single hydrocast over the active Macdonald summit indicate a mobility of the REE, which probably results from high-temperature basalt-seawater interaction.

Reference frames for plate tectonics and uncertainties

Paleomagnetic, hotspot, and mean-lithosphere frameworks are compared for the Cenozoic Era. Errors in the relative motions of plate pairs accumulate to give uncertainties in reference frames. Uncertainties in the poles of the mean-lithosphere framework are estimated to be less than 1° and for the hotspot framework 1.1°, assuming a 50 km error in location of individual hotspots. The A 95 's for paleomagnetic data range from 1° to 5°. With these uncertainties the differences between the paleomagnetic and mean-lithosphere coordinate systems are not found to be statistically significant, and the hotspot frame barely differs from the mean-lithosphere frame during the early Tertiary. On the other hand, the hotspot and paleomagnetic axes diverge in opposing directions in the mean-lithosphere framework, and can be distinguished. This difference can be attributed to the Pacific hotspots having more than 10° motion relative to the mean-lithosphere, whereas the African set of hotspots does not statistically differ from the mean-lithosphere frame. One possible explanation is a shift of the Pacific hotspots relative to those in the Atlantic.

Pacific plate reconstructions and uncertainties

The Pacific plate is positioned into the hotspot reference frame, first directly relative to the Pacific hotspots, and then indirectly relative to the African hotspots by using seafloor spreading data. The reconstructions are repeated for 10, 15, 37, 48, and 61 Ma. The errors associated with each link for both approaches are expressed by covariance matrices which can be combined to yield the overall error in the case of the relative plate motion circuit. Assuming the hotspots are fixed, and that the reconstructions record the true relative positions of the plates, the reconstructions should be equivalent within their uncertainties. In fact, however, there is agreement within the uncertainties only for 10 Ma, with the misfit increasing with age for the earlier reconstructions. These misfits can he accounted for by three alternative explanations: large hotspot location uncertainties, a shift in time of the Pacific group of hotspots relative to the African group, or an additional plate boundary in the relative plate motion circuit, which we assume to be in Antarctica. By examining each of these possibilities in turn, we conclude that the uncertainties in individual hotspot locations required to bring the reconstructions into agreement are unreasonably large, and that the data available at present do not allow us to distinguish which of the remaining two possibilities may be the primary source of reconstruction errors.

The paleomagnetism of eastern Nazca plate seamounts

Paleomagnetism of eastern Nazca plate seamounts defines Nazca and Farallon absolute plate motion during Cenozoic times. Magnetic and bathymetric surveys are presented for two eastern Nazca plate seamounts in the Chile Basin and they are used to calculate paleomagnetic poles with uniform and nonuniform magnetic modeling. The paleopole for Piquero-2 seamount is coincident with the earth's pole, suggesting a young seamount. The paleopole for Piquero-1 seamount indicates that the Nazca plate moved 23° northward during 0–50 ma. This is 13° more latitudinal motion than predicted by a Pacific hotspot reference frame and 20 ° more motion than predicted by DSDP sediment and basalt paleomagnetism.

The origin of the Marquesas fracture zone ridge and its implications for the nature of hot spots

The Marquesas Island chain departs by 20–30° from the trends of other late Tertiary hot-spot traces and appears more intermittent in activity than most others in having produced volcanos only in the period 1.4–6.4 Ma. In order to explain the Marquesas data, it has been proposed that hot spots can move relatively rapidly with respect to one another and that most plumes are episodic phenomena. We reevaluate these hypotheses based on geophysical data from the Marquesas fracture zone ridge collected during a recent oceanographic expedition to the Marquesas Islands. Our preferred explanation based on analysis of Seabeam bathymetric, seismic, gravity, and magnetic data across the ridge is that it is a young volcanic construct, 20 km wide and almost 2 km high, formed recently as the Marquesas hot spot encountered the Marquesas fracture zone. Examination of regional bathymetric maps reveals that topographically identical ridges also appear along the Galapagos fracture zone and the Nova-Canton Trough at positions corresponding to the predicted trace of the Marquesas plume based on the trends of the Hawaiian and Society island chains. Using these observations, we conclude that the Marquesas hot spot is not moving rapidly with respect to the other Pacific hot spots, nor is it intermittent in nature. We prefer a model in which the Marquesas plume is simply too weak to penetrate normal oceanic lithosphere unless given an easy conduit to the surface, such as a fracture zone. The surface expression of weak plumes such as the Marquesas is therefore strongly controlled by the thermal and mechanical state of the overlying oceanic lithosphere, an effect which must be taken into account before interpreting hot-spot traces in terms of intrinsic plume properties.

Latitudinal shift of Pacific hotspots during the late Cretaceous and early Tertiary

Two frames of reference, the geomagnetic field and hotspots, have been widely used to measure the past motions and interactions of lithospheric plates. Several studies have found that these reference frames have undergone as much as 19–22° of relative motion, often called true polar wander (TPW), in the past 200 Myr1–4. Moreover, it also appears that this offset did not accumulate at a uniform rate. Some evidence suggests that spurts of TPW can occur at rates as high as 1–2° per Myr3–5. Rates even an order of magnitude less could pose serious problems for tectonic studies based on the hotspot reference frame because velocities of this magnitude cannot be considered insignificant compared to those of the plates. Unfortunately, the imprecision of most palaeomagnetic data is such that large error bounds are common for many estimates of TPW rates. In this report, we present a record of TPW derived from Pacific palaeomagnetic data that are relatively well constrained in age and palaeolatitude. They show that offset between the two reference frames accumulated at ∼0.3° per Myr–1 from late Cretaceous to early Tertiary.

Models for true polar wander from palaeomagnetism, geoid and hotspot studies

Various analyses of palaeomagnetic data have indicated that the configuration of the time averaged geomagnetic field approximates an offset rather than a geocentric dipole. The displacement of the dipole from the geocentre' relative to present-day Eurasia is inferred to be directed, irrespective of age, toward the western equatorial Pacific—an area which is also characterized by the largest positive geoid height. It is argued that the close association between this geoid high and offset of the magnetic centre has a common cause, a long-standing quadrupolar heterogeneity which is connected ultimately to the Earth's core. With this assumption a conceptual model of true polar wander is outlined which predicts that the above referred mass anomaly was aligned with the Earth's rotation axis in the early Palaeozoic and was moved smoothly to the equator by the Goldreich-Toomre mechanism before the late Cretaceous. If seen from the mantle related hotspot frame, which has now been extrapolated back to include the Permian, a more complex model emerges. As a premise to the model, it is shown that the Pacific geoid high is composed of two positive height anomalies of which only the eastern one is correlatable with the Pacific hotspot province. Thus, only this and the Atlantic-African positive geoidal feature are hotspot related and the western Pacific positive residual is assumed, essentially as above, to be originating from the core. Given that the geoid anomalies in question are stable against time, the hotspot related model requires a 30° true polar wander with respect to the mantle as well as the core in roughly opposite directions from early Permian to late Cretaceous. During this time the mantle moved counterclockwise about an equatorial pole at approximate longitude 37° and the core clockwise about a nearby pole both with respect to the rotation axis. For times preceding the Permian the extrapolation can be done at least in principle on similar grounds as with the simpler model.

Late Cretaceous apparent polar wander of the Pacific Plate: Evidence for a rapid shift of the Pacific hotspots with respect to the spin axis

Paleomagnetic data from the Pacific plate show that the average azimuth of the Pacific plate decreased by ∼12° between 90 Ma and 81 Ma, whereas trends of Late Cretaceous seamount chains suggest that the Pacific plate moved northwestward relative to the hotspots with little change in azimuth during the same time interval. This difference is used to infer a 13°±4° shift of the spin axis relative to the hotspots in the Pacific Ocean basin. This shift occurred within 5 m.y. of the occurrence of anomalous behavior of the paleomagnetic field recorded in anomalies 33 and 34. This coincidence in timing suggests there may be a causal link between these three phenomena.

Geochronology and origin of the Pratt‐Welker Seamount Chain, Gulf of Alaska: A new pole of rotation for the Pacific Plate

40K‐40Ar and fission‐track dating of four seamounts near the southeast end of the Pratt‐Welker seamount chain in the Gulf of Alaska, in conjunction with previously published K‐Ar and fission‐track ages near the northwest end of the chain, documents the complex origin of this seamount chain. Transitional basalts from the adjacent guyots Hodgkins, Davidson, and Denson are dated as 14.3 to 18.2 m.y. These ages, only slightly younger than the ages of the underlying crust, indicate formation of these three seamounts at or very near a spreading center. In contrast, alkalic series lavas (alkali olivine basalts and trachytes) from Kodiak, Giacomini, Dickins, and Hodgkins fit a systematic linear age progression: 23.9±0.6 m.y., 20.9±0.4 m.y., 4.0±0.2 m.y., and 2.8±0.2 m.y., respectively. Hodgkins has apparently experienced two generically different episodes of volcanism, separated by about 12 m.y. The age progression among dated alkali basalts is consistent with the hot spot hypothesis and suggests that for the last 24 m.y. the Pacific plate has moved northwest at 4.4±0.4 cm/yr with respect to the Pratt‐Welker hot spot. This volcanic propagation rate, together with the rates from other parallel Neogene Pacific chains, allows an improved estimate of the pole and rate of rotation of the Pacific plate relative to hot spots: 70°N, 95°W, and 0.88°±0.10°/m.y. We conclude that no significant motion of the Pratt‐Welker hot spot with respect to other Pacific hot spots has yet been detected. However, the Pratt‐Welker age data may alternatively be explained by either the longitudinal roll or propagating crack hypothesis. New K‐Ar ages from Horton guyot, in the Cobb seamount chain, indicate alkalic volcanism 20.7± .0 m.y. ago, consistent with a predicted age of 20 m.y. based on the hot spot hypothesis. Guyot depths from Horton and the dated Pratt‐Welker seamounts are consistent with the K‐Ar ages and normal subsidence of oceanic crust.