The association of Hawaiian-Emperor volcanism with a large-scale central Pacific anisotropy anomaly at ~150 km depth can be explained by tapping of shallow melt sources in a perisphere/LLAMA (layer of lateral advection of mass and anisotropy) model. The origin of the anisotropy anomaly can be traced to the formation of a phlogopite-garnet-pyroxenite assemblage in the perisphere beneath an island arc on the Stikine terrane of the North American Cordillera in the Carboniferous. The pyroxenites were formed when subduction-related melts invaded the mantle wedge at ~150–200 km depth. The enriched region inherited the thermal profile of the mantle wedge, along with a solar-like noble gas isotopic composition from earlier fluxing of hydrothermal fluids between interplanetary dust particle–bearing deep-sea sediments and ultramafic layers of the oceanic crust prior to subduction. After termination of subduction, the enriched perisphere was displaced to the northeast beneath the Farallon plate, and then to the northwest beneath the Izanagi and Pacific plates, eventually becoming distorted into the shape of the present-day central Pacific anisotropy anomaly. During the thermal equilibration time, estimated at ~170 m.y., the phlogopite-garnet-pyroxenite assemblage followed a horizontal trajectory in pressure-temperature ( P-T ) space. As the P-T path crossed the solidi for volatile-bearing pyroxenite compositions, diabatic partial melting generated carbonatitic to alkaline melts which began to ascend and metasomatize shallower levels of the perisphere, carrying with them the geochemical signature of the original pyroxenites. The present central Pacific anisotropy anomaly is the current manifestation of the metasomatized domain. The latter was tapped from the Late Cretaceous to the present, by propagating fractures induced by large-scale plate reorganizations in the northwest of the Pacific Basin, to produce the Hawaiian-Emperor volcanic chain.

The Northwest Hawaiian Ridge is a classic example of a large igneous province. The morphology and geology of the ridge is poorly characterized, although it constitutes the longest segment (~47%) of the Hawaiian-Emperor Chain. Here we present a new bathymetric compilation, petrographic and X-ray fluorescence (XRF) data for lavas from 12 volcanoes along the Northwest Hawaiian Ridge, and review literature data for the age and isotopic variation of the ridge. The bathymetric compilation revealed that the Northwest Hawaiian Ridge consists of at least 51 volcanoes. The 45 new XRF analyses show that the Northwest Hawaiian Ridge contains tholeiitic and alkalic lavas with compositions typical of lavas from the Hawaiian Islands. The absolute ages and duration of volcanism of individual Northwest Hawaiian Ridge volcanoes are poorly known, with modern 40 Ar/ 39 Ar ages for only 10 volcanoes, mostly near the bend in the chain. We infer the initiation age of the Hawaiian-Emperor Bend to be ca. 49–48 Ma, younger than the age for the onset of island arc volcanism in the western Pacific (52–51 Ma). Thus, the kink in the Hawaiian-Emperor Chain and the onset of arc volcanism were not synchronous. Isotopic data are sparse for the Northwest Hawaiian Ridge, especially for Pb and Hf. Two transitional lavas from just south of the bend have Loa trend type Pb and Sr isotopic ratios. Otherwise, the available chemistry for Northwest Hawaiian Ridge lavas indicates Kea-trend source compositions. The dramatic increase in melt flux along the Hawaiian Ridge (~300%) may be related to changes in melting conditions, source fertility, or plate stresses.

In order to better define the late Eocene clinopyroxene-bearing (cpx) spherule layer and to determine how the ejecta vary with distance from the presumed source crater (Popigai), we searched for the layer at 23 additional sites. We identified the layer at six (maybe seven) of these sites: Deep Sea Drilling Project (DSDP) and Ocean Drilling Program (ODP) Holes 592, 699A, 703A, 709C, 786A, 1090B, and probably 738B. The cpx spherule layer occurs in magnetochron 16n.1n, which indicates an age of ca. 35.4 ± 0.1 Ma for the layer. We found the highest abundance of cpx spherules and associated microtektites in Hole 709C in the northwest Indian Ocean, and we found coesite and shocked quartz in the cpx spherule layer at this site. We also found coesite in the cpx spherule layer at Site 216 in the northeast Indian Ocean. This is the first time that coesite has been found in the cpx spherule layer, and it provides additional support for the impact origin of this layer. In addition, the discovery of coesite and shocked quartz grains (with planar deformation features [PDFs]) supports the conclusion that the pancake-shaped clay spherules associated with quartz grains exhibiting PDFs are diagenetically altered cpx spherules. An Ir anomaly was found associated with the cpx spherule layer at all four of the new sites (699A, 709C, 738B, 1090B) for which we obtained Ir data. The geometric mean of the Ir fluence for the 12 sites with Ir data is 5.7 ng/cm 2 , which is ~10% of the fluence estimated for the Cretaceous-Tertiary boundary. Based on the geographic distribution of the 23 sites now known to contain the cpx spherule layer, and 12 sites where we have good chronostratigraphy but the cpx spherule layer is apparently absent, we propose that the cpx spherule strewn field may have a ray-like distribution pattern. Within one of the rays, the abundance of spherules decreases and the percent microtektites increases with distance from Popigai. Shocked quartz and coesite have been found only in this ray at the two sites that are closest to Popigai. At several sites in the Southern Ocean, an increase in δ 18 O in the bulk carbonate occurs immediately above the cpx spherule layer. This increase may indicate a drop in temperature coincident with the impact that produced the cpx spherule layer.

Alteration of undisturbed igneous material used for argon dating work often results in inaccurate estimates of the crystallization age. A new quantitative technique to detect alteration has been developed (see Baksi, this volume , Chapter 15), utilizing the 36 Ar levels observed in rocks and minerals. The method is applied to data in the literature for rocks linked to hotspot activity. For subaerial rocks, argon dating results are critically examined for the Deccan Traps, India. The duration of volcanic activity and its coincidence in time with the K-T boundary are shown to be uncertain. The bulk of dated seafloor material (recovered from the Atlantic, Indian, and Pacific oceans) proves to be altered. Ages determined using large (hundreds of milligram) samples are generally unreliable, due to inclusion of altered phases. Such analyses include studies suggesting an age of ca. 43 Ma for the bend in the Hawaiian-Emperor chain. More recent attempts, using much smaller sub-samples (∼10 mg) that have been acid leached to remove alteration products, are generally of higher reliability. Plagioclase separates sometimes yield reliable results. However, many whole-rock basalts from the ocean floor yield ages that are, at best, minimum estimates of the time of crystallization. Most “rates of motions,” calculated from hotspot track ages, are shown to be invalid. Seafloor rocks are recovered at considerable expense but often are not suitable for dating by the argon methods. Most are severely altered by prolonged contact with sea-water. A method is recommended for testing silicate phases prior to attempts at argon dating. This involves a quantitative determination of the 36 Ar content of the material at hand; dating phases without pretreatment—leaching with HNO 3 for material containing ferromagnesian phases, and HF for feldspars—is strongly discouraged.

If mantle plumes (hotspots) are fixed in the mantle and the mantle reference frame does not move relative to the spin axis (i.e., true polar wander), a model of plate motion relative to the hotspots should predict the positions of past paleomagnetic poles. Discrepancies between modeled and observed poles thus may indicate problems with these assumptions, for example, that the hotspots or spin axis have shifted. In this study, I compare paleomagnetic and hotspot-model–predicted apparent polar wander paths (APWP) for the Pacific plate. Overall, the two types of APWP have similar shapes, indicating general agreement. Both suggest ∼40° total northward drift of the Pacific plate since ca. 123 Ma. Offset between paleomagnetic and hotspot-predicted poles is small for the past ca. 49 Ma, consistent with fixed hotspots during that time, but the offsets are large (6–15°) for earlier times. These differences appear significant for the Late Cretaceous and early Cenozoic. During the period 94–49 Ma, the hotspot model implies the paleomagnetic pole should have drifted ∼20° north without great changes in rate. Measured paleomagnetic poles, however, indicate rapid polar motion between 94 and 80 Ma and a stillstand from 80 to 49 Ma. Comparison with global synthetic APWP suggests that the 94- to 80-Ma polar motion may be related to true polar wander. The stillstand indicates negligible northward motion of the Pacific plate during the formation of the Emperor seamounts. This observation is drastically different from most accepted Pacific plate motion models and requires rethinking of western Pacific tectonics. If the Emperor seamounts show relative motion of the plate relative to the Hawaiian hotspot, the implied southward hotspot motion is ∼19°. Lack of a diagnostic coeval phase of polar wandering in global APWP and consideration of the significance of the Hawaiian-Emperor bend imply that true polar wander is probably not the cause. Likewise, mantle-flow models do not readily explain the large southward drift of the hotspot or its inferred large westward velocity component. Thus, current models for the formation of the Emperor seamounts appear inadequate, and new ideas and further study are needed. Comparison of the Pacific APWP with a global APWP, both rotated into the Antarctic reference frame, shows an offset of ∼10°, implying problems with plate circuits connecting Antarctica with surrounding plates. This result suggests that caution is required when predicting trends of hotspot seamount chains using plate circuits through Antarctica.

Current interpretations of Cretaceous tectonic evolution of the northwest Pacific trace interactions between the Pacific plate and three other plates, the Farallon, Izanagi, and Kula plates. The Farallon plate moved generally eastward relative to the Pacific plate. The Izanagi and Kula plates moved generally northward relative to the Pacific plate, with Izanagi the name given to the northward-moving plate prior to the Cretaceous normal polarity superchron and the name Kula applied to the postsuperchron plate. In this article I suggest that these names apply to the same plate and that there was only one plate moving northward throughout the Cretaceous. I suggest that the tectonic reorganization that has previously been interpreted as formation of a new plate, the Kula plate, at the end of the superchron was actually a plate boundary reorganization that involved a 2000 km jump of the Pacific–Farallon–Kula/Izanagi triple junction. Because this jump occurred during a time of no magnetic reversals, it is not possible to map or date it precisely, but evidence suggests mid-Cretaceous timing. The Emperor Trough formed as a transform fault linking the locations of the triple junction before and after the jump. The triple junction jump can be compared with an earlier jump of the triple junction of 800 km that has been accurately mapped because it occurred during the Late Jurassic formation of the Mesozoic-sequence magnetic lineations. The northwest Pacific also contains several volcanic features, such as Hawaii, that display every characteristic of a hotspot, although whether deep mantle plumes are a necessary component of hotspot volcanism is debatable. Hawaiian volcanism today is apparently independent of plate tectonics, i.e., Hawaii is a center of anomalous volcanism not tied to any plate boundary processes. The oldest seamounts preserved in the Hawaii-Emperor chain are located on Obruchev Rise at the north end of the Emperor chain, close to the junction of the Aleutian and Kamchatka trenches. These seamounts formed in the mid-Cretaceous close to the spreading ridge abandoned by the 2000 km triple junction jump. Assuming that Obruchev Rise is the oldest volcanic edifice of the Hawaiian hotspot and thus the site of its initiation, the spatial and temporal coincidence between these events suggests that the Hawaii hotspot initiated at the spreading ridge that was abandoned by the 2000 km jump of the triple junction. This implies a tectonic origin for the hotspot. Other volcanic features in the northwest Pacific also appear to have tectonic origins. Shatsky Rise is known to have formed on the migrating Pacific-Farallon-Izanagi triple junction during the Late Jurassic–Early Cretaceous, not necessarily involving a plume-derived hotspot. Models for the formation of Hess Rise have included hotspot track and anomalous spreading ridge volcanism. The latter model is favored in this article, with Hess Rise forming on a ridge axis possibly abandoned as a result of a ridge jump during the superchron. Thus, although a hotspot like Hawaii could be associated with a deep mantle plume today, it would appear that it and other northwest Pacific volcanic features originally formed as consequences of shallow plate tectonic processes.

Reconstruction of the tectonic evolution of the Pacific basin indicates a direct relationship between intraplate volcanism and plate reorganizations, which suggests that volcanism was controlled by fracturing and extension of the lithosphere. Middle Jurassic to Early Cretaceous intraplate volcanism included oceanic plateau formation at triple junctions (Shatsky Rise, the western Mid-Pacific Mountains) and a diffuse pattern of ocean island volcanism (Marcus Wake, Magellan seamounts) reflecting an absence of any well-defined stress field within the Pacific plate. The stress field changed in the Early Cretaceous when accretion of the Insular terrane to the North American Cordillera and the Median Tectonic arc to New Zealand stalled migration of the Pacific-Farallon and Pacific-Phoenix ocean ridges, leading to the generation of the Ontong Java, Manahiki, Hikurangi, and Hess Rise oceanic plateaus. Plate reorganizations in the Late Cretaceous resulted from the breakup of the Phoenix and Izanagi plates through collision of the Pacific-Phoenix ocean ridge with the southwest margin of the basin and development of island arc–marginal basin systems in the northwestern part of the basin. The Pacific plate nonetheless remained largely bounded by spreading centers, and intraplate volcanism followed preexisting lines of weakness in the plate fabric (Line Islands) or resulted from fractures generated by ocean ridge subduction beneath island arc systems (Emperor chain). The Pacific plate began to subduct under Asia in the Early Eocene as inferred from the record of accreted material along the Japanese margin. Further changes to the stress field at this time resulted from abandonment of the Kula-Pacific and the North New Guinea (Phoenix)–Pacific ridges and from development of the Kamchatkan and Izu-Bonin-Mariana arcs, leading to the generation of the Hawaiian chain as a propagating fracture. The final major change in the stress field occurred in the Late Oligocene as a result of breakup of the Farallon into the Cocos and Nazca plates, which caused a hiatus in Hawaiian volcanism; initiated the Sala y Gomez, Foundation, and Samoan chains; and terminated the Louisville chain. The correlations with tectonic events are compatible with shallow-source models for the origin of intraplate volcanism and suggest that the three principal categories of volcanism, intraplate, arc, and ocean ridge, all arise from plate tectonic processes, unlike in plume models, where intraplate volcanism is superimposed on plate tectonics.

The lithosphere crack model, the main alternative to the mantle plume model for age-progressive magma emplacement along the Hawaiian-Emperor volcano chain, requires the maximum horizontal tensile stress to be normal to the volcano chain. However, published stress fields calculated from Pacific lithosphere tractions and body forces (e.g., subduction pull, basal drag, lithosphere density) are not optimal for southeast propagation of a stress-free, vertical tensile crack coincident with the Hawaiian segment of the Hawaiian-Emperor chain. Here we calculate the thermoelastic stress rate for present-day cooling of the Pacific plate using a spherical shell finite element representation of the plate geometry. We use observed seafloor isochrons and a standard model for lithosphere cooling to specify the time dependence of vertical temperature profiles. The calculated stress rate multiplied by a time increment (e.g., 1 m.y.) then gives a thermoelastic stress increment for the evolving Pacific plate. Near the Hawaiian chain position, the calculated stress increment in the lower part of the shell is tensional, with maximum tension normal to the chain direction. Near the projection of the chain trend to the southeast beyond Hawaii, the stress increment is compressive. This incremental stress field has the form necessary to maintain and propagate a tensile crack or similar lithosphere flaw and is thus consistent with the crack model for the Hawaiian volcano chain.