Liang Liu

Oceanic trench bending over 1000 km occurs globally, with notable along-trench variations in mantle wedge thermochemistry. However, the mechanisms influencing both shallow trench geometry and deep mantle state remain unclear. Here we focus on the Mariana Trench as a representative case. Our 3D model successfully replicates its curved geometry. Modeling results further reveal that extensive intra-oceanic trench retreat and bending occur when the incoming plate features sharp density variations along the trench accompanied by a large buoyant segment (e.g., the Caroline Plate). In this scenario, significant dynamic pressure fluctuations develop in the mantle wedge, generating vigorous toroidal flow extending >1000 km from the buoyant segment, even under non-linear mantle rheology. The model-predicted toroidal flow transports warm mantle materials, reshaping local topography, rejuvenate the mantle wedge, and heating the slab surface to facilitate volatile release. These experiments highlight toroidal flow as a key regular of both large-scale trench bending and along-trench thermochemical variations.

The process‐based surface magmatic expression of mantle plumes is typically a large igneous province (LIP) induced by the mantle plume head, followed by subsequent age‐progressive volcanic tracks resulting from plate drifting above the plume tail. However, although plume‐like low velocity structures have long been observed below Hainan, the magmatic evidence for the mantle plume head or tail remains absent in the surrounding area, which left the existence of the Hainan mantle plume debated yet. Here, we identify volcanism derived from a hybrid source of Hainan plume and lithospheric base, 700 km away from the Hainan plume conduit, based on the submarine seamount basalt obtained by a submersible. This volcanism coincides with the most extensive eruption episode of the Hainan basalt and occurs beneath a thicker lithospheric lid than that of the Hainan basalt. Consistent with the seismic tomography of low velocity zone beneath the lithosphere, this seamount basalt serves as the key magmatic evidence for the interaction between the Hainan mantle plume head and the northern South China Sea (N‐SCS) continental lithosphere. We further demonstrate that the basalt that primarily fractionated within the lower crust was seismically detected as a high‐velocity body. Therefore, it is highly likely that a mantle plume head is responsible for the large volume of igneous rocks on the N‐SCS (at least 150,000 km ³ ), comparable with LIPs (>100,000 km ³ ). The N‐SCS region now may be under the mantle plume head stage and, together with the Hainan basalts, belong to an active LIP.

Hotspot swells have typically been associated with mantle plume upwelling, yet the origins of much broader but lower superswells remain unclear. This study first reveals that swell width increases while height decreases at younger seafloor ages; it is possible for a single hot spot swell size to approach that of a superswell. Compressible numerical models replicate the observed trend. Modeling results suggest that both hot spot swells and superswells can form from interactions between mantle plumes and oceanic plates. Beneath young seafloors, the lateral redistribution of restitic, hence buoyant and relatively low-viscosity, lower lithospheric mantle due to plume impingement explains the formation of superswell-like morphology. These models effectively demonstrate lithospheric lid effects on mantle melting and support a reassessment of conventional methods for estimating core heat flux.

The thermal mechanism that drives the prolonged volcanism on the Moon, especially after the major pulse of Imbrian-aged eruptions, remains unknown. Here, we present a petrological and geochemical study of two types of young farside mare basalts, the ~2.8–billion year (Ga) low-Ti and ~2.9-Ga very-low-Ti basalts, collected during the Chang’e-6 mission. The results of our study show that these basalts have pyroxenitic sources and originate from a depth of ~60 to 80 kilometers and ~120 kilometers, respectively. The depth of their source that became progressively shallower over time, combined with thermal modeling results, suggests that magmatic underplating beneath the ilmenite-bearing cumulate (IBC) that escaped the mantle overturn phase could be a thermal driver for the young lunar magmatism. Global remote-sensing investigations further reveal different TiO ₂ contents between young basalts from each side of the Moon, attributable to asymmetric composition and thickness of the IBC in the uppermost mantle.

Classical mantle convection models predict a broad surface uplift over a lower mantle upwelling. However, recent studies have identified anomalously localized surface subsidence above seismically imaged lower mantle upwellings, particularly in regions where upwellings are impeded by subducted/delaminated blocks not currently connected to a subducting/delaminating lithosphere ('remnant blocks' for simplicity), e.g., in the western USA and the South China Sea. Known geological processes cannot fully explain the observed localized subsidence, and its locality implies a strong association with the underneath upwelling. Here, we use numerical models to quantitatively explore the contribution of lower mantle upwelling to surface topography evolution. Our results demonstrate that the divergent mantle flow caused by lower-mantle upwelling can stretch the overlying lithosphere, inducing broad subsidence that can be reversed when the upwelling reaches the lithospheric bottom. Moreover, the interaction between the remnant block and upwelling can extend the duration of subsidence and focus the subsidence into a narrow region, which we call the lens effect of remnant blocks. Similar abnormal subsidence events in northeastern Asia further show the potential broad applicability of the proposed mechanism, although, in specific research areas, other shallow mechanisms could also have played important roles.

Deep carbon fundamentally modifies the physical properties of solids and melts, thereby affecting partial melting and compositional differentiation within the mantle. Experimental petrological studies suggest that the carbonated oceanic crust in the subducted slab may undergo carbon‐induced partial melting and decarbonation in the deep upper mantle. Although slab geotherms in cold subduction zones fall below the experimentally obtained solidus of the carbonated oceanic crust, as many subducted slabs can stagnate within the mantle transition zone (MTZ), they could be warmed up during the elongated residence time. However, quantitative geodynamic studies regarding the effect of such slab stagnation on slab decarbonation through partial melting are lacking. To fill this research gap, we employ 2D numerical modeling to investigate the potential decarbonation behaviors of subducted slabs due to carbon‐induced melting during their flattening within the MTZ. Our results demonstrate that the decarbonation rate of a stagnated slab that is hot and carbonate‐rich can surpass 75%, whereas for a cold and carbonate‐poor slab, for example, the present‐day Pacific plate, the decarbonation rate is limited to a few percent. Hence, we suggest that geochemical signatures of Cenozoic intraplate basalts in eastern China and seismological observations of the low‐velocity layer in the Northeast Asian upper mantle can be traced back to an early phase of subduction when the subducted slab was moderately hot and carbonate‐rich. Our geodynamic model provides a quantitative constraint on the deep carbon cycle associated with subduction, especially in the deep upper mantle.

In the evolutionary history of the Tethys tectonic realm, numerous continental fragments progressively split from the southern hemisphere's Gondwana continent and “unidirectionally” converged and assembled with the northern hemisphere's Eurasian continent, ultimately shifting the center of the Earth's continental masses from the southern hemisphere in the late Paleozoic to the present northern hemisphere. Previous studies have vividly summarized this seemingly unidirectional process of plate fragmentation and reassembly as the “Tethys one-way train.” As a “welcoming ceremony” for this train's arrival, the upper plates' lithosphere, such as in the Tibetan Plateau and Anatolia from different geological eras, records unique tectonic-magmatic responses. For example, tectonic-magmatic activity may first appear in the interior, thousands of kilometers away from the convergence boundary, then expand from the inside out. This can also develop into a “piston-like” cycle of transformations: crustal compression + magmatic quiescence → crustal extension + magmatic peak → crustal compression + magmatic quiescence → and so on. Addressing these typical geological phenomena of the Tethys tectonic realm and combining the tectonic background revealed by plate reconstruction with the contemporaneous multiple episodes of block assembly, we employ forward numerical simulation to interpret the deep driving processes and mechanisms behind these phenomena. By utilizing geological, geochemical, and geophysical observations to constrain model results, we propose that the abrupt changes in the lower-plate movement characteristics (such as subduction angle and rate) caused by the subduction of high-buoyancy blocks significantly control the rapid transition of tectonic-magmatic patterns in regions like Tibet and Anatolia. The multiple episodes of block assembly can explain the accordion-like tectonic-magmatic cycles of the active continental margins. Given that the high-buoyancy blocks require continuous northward driving forces during their journey from Gondwana's fragmentation to their convergence with the Eurasian continent, we further calculated the temperature distribution in today's upper mantle using previous global seismic wave attenuation models to establish a forward geodynamic model, exploring the deep driving mechanisms of the convergence process in the Tethys tectonic realm. The modeling results indicate that the current temperature structure of the upper mantle, with colder northern regions and warmer southern regions, can create sufficiently large lateral mantle density contrasts and trigger the initiation of oceanic plate subduction towards the low-temperature areas, essentially starting the engine of the express train. Subsequently, the demise of secondary ocean basins during convergence often accompanies the subduction and rebound of high-buoyancy blocks, which rapidly returning fragments strongly collide with the rear oceanic plates, triggering a new round of oceanic subduction and further cooling the northern hemisphere's upper mantle, thereby giving the convergence process a chain reaction characteristic. Therefore, although the continental blocks fragmented from Gondwana may be seen as “passengers” of the one-way train, they have played a significant role in both the welcoming ceremony and the sustainable operation of the train.

The type of lithosphere subducted between India and Tibet since the Paleocene remains controversial; it has been suggested to be either entirely continental, oceanic, or a mixture of the two. As the subduction history of this lost lithosphere strongly shaped Tibetan intraplate tectonism, we attempt to further constrain its nature and density structure with numerical models that aim to reproduce the observed history of magmatism and crustal thickening in addition to present-day plateau properties between 83˚E and 88˚E. By matching time-evolving geological patterns, here we show that Tibetan tectonism away from the Himalayan syntaxis is consistent with the initial indentation of a craton-like terrane at 55±5 Ma, followed by a buoyant tectonic plate with a thin crust, e.g., a broad continental margin (Himalandia). This new geodynamic scenario can explain the seemingly contradictory observations that had led to competing hypotheses like the subduction of Greater India versus largely oceanic subduction prior to Indian indentation.