The seismic behavior of subduction megathrusts varies spatially and is influenced by the properties of subducting plates, including their sedimentary cover. Characterizing these subduction inputs is essential for understanding the mechanisms behind fault slip variability. The Hikurangi subduction margin exhibits well-documented variations in fault slip ranging from large earthquakes to slow slip events and aseismic creeping. The variations coincide with along-margin changes in convergence rate, trench sediment thickness, and accretionary wedge development. Here, we integrate regional seismic reflection profiles with drill cores from the International Ocean Discovery Program Expeditions 372 and 375, and with other seafloor samples. We find that the Hikurangi Trough hosts some of Earth's thickest trench sediments, at up to 10 km, including pelagic sediments and a >6 km thick siliciclastic trench wedge that is younger than previously thought. Seismic imaging reveals that the shallow megathrust depth increases by 3.5-4.5 km southward along the margin in parallel with increasing trench sediment thickness despite a "step up" of the fault's stratigraphic position to a different assemblage of host rocks. In contrast to the irregular northern megathrust associated with subducting seamounts and heterogeneous volcaniclastic and pelagic host rocks, the southern megathrust is smooth, hosted in calcareous pelagic rocks, and straddles inherited polygonal faults. These along-strike variations in physical properties and megathrust geometry correspond to observed changes in fault slip behavior. Our findings provide new insights into the factors controlling fault slip variability, with implications for seismic hazard assessment at subduction zones worldwide and for understanding subduction margin evolution.

Understanding Earth's subsurface is critical for energy transition, natural hazard mitigation, and planetary science. Yet subsurface analysis remains fragmented, with separate models required for structural interpretation, stratigraphic analysis, geobody segmentation, and property modeling-each tightly coupled to specific data distributions and task formulations. We introduce the Geological Everything Model 3D (GEM), a unified generative architecture that reformulates all these tasks as prompt-conditioned inference along latent structural frameworks derived from subsurface imaging. This formulation moves beyond task-specific models by enabling a shared inference mechanism, where GEM propagates human-provided prompts-such as well logs, masks, or structural sketches-along inferred structural frameworks to produce geologically coherent outputs. Through this mechanism, GEM achieves zero-shot generalization across tasks with heterogeneous prompt types, without retraining for new tasks or data sources. This capability emerges from a two-stage training process that combines self-supervised representation learning on large-scale field seismic data with adversarial fine-tuning using mixed prompts and labels across diverse subsurface tasks. GEM demonstrates broad applicability across surveys and tasks, including Martian radar stratigraphy analysis, structural interpretation in subduction zones, full seismic stratigraphic interpretation, geobody segmentation, and property modeling. By bridging expert knowledge with generative reasoning in a structurally aware manner, GEM lays the foundation for scalable, human-in-the-loop geophysical AI-transitioning from fragmented pipelines to a vertically integrated, promptable reasoning system. Project page: https://douyimin.github.io/GEM

Along the Chilean continental margin, where the active Nazca plate subducts under the South American plate, large deposits of gas hydrates trapped in marine sediments can be observed in seismic profiles (from Valparaíso to Patagonia), which may be affected by the increase in deep-water temperature due to climate change. This could trigger a massive release of methane gas into the marine environment and the atmosphere with negative effects on the environment, mainly through dissociation of the hydrates and/or direct migration to the seafloor through faults.In this study, new multichannel seismic data located off Chiloé Island (∼42°S) are presented, and for the first time a BSR parallel to the continental margin (north-south) is analysed. Here, an efficient sediment subduction occurs where the young oceanic Nazca plate (5-25 Ma) subducts beneath the continental plate along Chiloé Island. In addition, an integrated analysis to better understand the hydrological and hydrothermal system in this area included previous geophysical and geological datasets, such as a) bathymetry, b) geothermal gradient calculated from gas hydrate distribution boundaries, and c) thermal conductivities. The results show a continuous and strong BSR along the entire seismic profile MGL1701-27 (approximately 27 km long), where large deformations, fractures and faults occur, favouring the flow of carbon-rich fluids from the depths, which subsequently form the gas hydrate layer. The large amount of gas hydrate present in the marine sediments off the coast of Chiloé Island is impressive. Velocity analysis of seismic profile MGL1701-27 indicates gas hydrate concentration values of up to 25% of the total rock volume, the highest estimates obtained at the Chilean margin. In addition, a peak in the geothermal gradient is related to deep faults that allow the passage of hot fluids from deeper strata to the seafloor. Finally, a large amount of free gas is estimated beneath the BSR, which enhances its intensity and continuity in the seismic profile.Our results provide valuable information for current and future studies related to climate change (methane storage), hydrothermal circulation, seismicity, gas hydrate stability and fluid venting in this natural laboratory, since Chiloé Island is located at the southern boundary of the rupture zone of the largest earthquake in recorded history (the Mw 9.5 Valdivia earthquake), the implications of which have yet to be fully elucidated.

Four decades of seismic reflection, onshore-offshore and ocean-bottom seismic data are integrated to constrain a high-resolution 3-D P-wave velocity model of the Hikurangi subduction zone. Our model shows wavespeeds in the offshore forearc to be 0.5-1 km/s higher in south Hikurangi than in the central and northern segments (VP <= 4.5 km/s). Correlation with onshore geology and seismic reflection data sets suggest wavespeed variability in the overthrusting plate reflects the spatial distribution of Late Jurassic basement terranes. The crustal backstop is 25-35 km from the deformation front in south Hikurangi, but this distance abruptly increases to similar to 105 km near Cape Turnagain. This change in backstop position coincides with the southern extent of shallow slow-slip, most of which occurs updip of the backstop along the central and northern margin. These relationships suggest the crustal backstop may impact the down-dip extent of shallow conditional stability on the megathrust and imply a high likelihood of near/trench-breaching rupture in south Hikurangi. North of Cape Turnagain, the more landward position of the backstop, in conjunction with a possible reduction in the depth of the brittle ductile transition, reduces the down-dip width of frictional locking between the southern (similar to 100 km) and central Hikurangi margin by up-to 50%. Abrupt transitions in overthrusting plate structure are resolved near Cook Strait, Gisborne and across the northern Raukumara Peninsula, and appear related to tectonic inheritance and the evolution of the Hikurangi margin. Extremely low forearc wavespeeds resolved north of Gisborne played a key role in producing long durations of long-period earthquake ground motions.

Chile is the world's leading producer of many terrestrial mineral resources; however, the potential of the country's marine mineral resources has been largely overlooked. Within its continental shelf (up to 200 nautical miles from the baselines from which the breadth of the territorial sea is measured), Chile has favorable geological characteristics for hosting and forming marine minerals and energy resources.During the last decades, several novel studies have demonstrated the potential of gas hydrate reservoirs in Chile (between 33° and 56°S) as an energy resource and source of greenhouse gases, which has attracted the attention of the Chilean scientific community. In addition, some studies have highlighted the potential value of marine minerals in the Chilean continental shelf, mainly due to the increasing demand for minerals for low-carbon energy production, such as cobalt-rich ferromanganese crusts, polymetallic nodules, and massive sulfides on the seafloor.The goal of this study is to review all information on the non-conventional energy and mineral resources of the Chilean continental shelf. Furthermore, we provide data (e.g., core samples, seismic profiles, or research from related papers) on marine deposits in the Chilean seabed. Here, we show unpublished seismic images of a previously unidentified massive hydrate deposit in the southernmost part of the 1960 Valdivia earthquake rupture zone, which was the largest earthquake recorded in history. We also present geological data that suggests the presence of nodules, sulfides, and crusts on the Chilean continental shelf. Collectively, these findings represent the most important but least explored resources for critical elements and base metals in the country.This study provides a primer for policymakers to apprise them of future research needed to develop potential mineral and energy resources within prospective deep-sea areas. It also includes advice on developing an environmental baseline for future environmental impact assessment. The new understanding of mineral and energy resources presented here greatly expands Chile's position beyond that as a source of terrestrial minerals to include potentially vast marine-based energy and metal resources, which are highly valued by industry. Based on these considerations, we encourage decision makers to promote and support studying marine deposits to further protect and evaluate future exploration.

Seamounts and basaltic basement can influence deformation and mass fluxes within subduction zones. We examined seamounts and volcanic units across the western Hikurangi Plateau, near the Hikurangi subduction margin, New Zealand, with seismic reflection images. Volcanism at the Hikurangi Plateau occurred in at least three phases that we attribute to (1) Early Cretaceous large igneous province formation, the top of which is marked by laterally continuous and dipping wedges of reflections that we interpret as lava flows; (2) Late Cretaceous seamounts and volcaniclastics that erupted onto the crust of the Hikurangi Plateau and make up the majority of seamount volume and basement relief; and (3) late-stage, Pliocene volcanics that erupted through and adjacent to Cretaceous seamounts and younger sediments of the north-central Hikurangi Plateau. The Pliocene volcanoes do not appear to be strongly welded to the plateau basement and may be petit spot volcanoes that are related to the displacement and accumulation of hydrous transition zone melts. Large seamounts and volcaniclastic units are evenly distributed across most of the Hikurangi Plateau near the Hikurangi margin but are absent from the Pegasus Basin. Although faults are imaged throughout the basement of the Pegasus Basin, contemporary normal faulting of the Hikurangi Plateau is uncommon, except for a zone of Quaternary normal faults near the Pliocene volcanics. These trends indicate that the Hikurangi megathrust may be more influenced by volcanic structures in the north and central Hikurangi margin, where plateau rifting and voluminous seamount eruptions have more substantially overprinted the original Early Cretaceous basement.

Abstract Fluid flow and pore‐pressure cycling are believed to control slow slip events (SSEs), such as those that frequently occur at the northern Hikurangi margin of New Zealand. To better understand fluid flow in the forearc system we examined the relationship between several physical properties of Cretaceous‐to‐Pliocene sedimentary rocks from the Raukumara peninsula. We found that the permeability of the deep wedge is too low to drain fluids, but fracturing increases permeability by orders of magnitude, making fracturing key for fluid flow. In weeks to months, plastic deformation, swelling, and possibly not‐yet‐identified mechanisms heal the fractures, restoring the initial permeability. We conclude that overpressures at the northern HM might partly dissipate during SSEs due to enhanced permeability near faults. However, in the months following an SSE, healing in the prism will lower permeability, forcing pore pressure to rise and a new SSE to occur.

The southern Hikurangi subduction zone exhibits significant along‐strike variation in convergence rate and obliquity, sediment thickness and, uniquely, the increasing proximity of southern Hikurangi to, and impingement on, the incoming continental Chatham Rise, an ancient Gondwana accretionary complex. There are corresponding changes in the morphology and structure of the Hikurangi accretionary prism. We combine widely spaced multichannel seismic reflection profiles with high resolution bathymetry and previous interpretations to characterize the structure and the history of the accretionary prism since 2 Ma. The southern Hikurangi margin can be divided into three segments. A northeastern segment (A) characterized by a moderately wide (∼70 km), low taper (∼5°) prism recording uninhibited outward growth in the last ∼1 Myr. Deformation resolvable in seismic reflection data accounts for ∼20 % of plate convergence, comparable with the central Hikurangi margin further North. A central segment (B) characterized by a narrow (∼30 km), moderate taper (∼8°) prism, with earlier (∼2‐∼1 Ma) shortening than segment A. Outward prism growth ceased coincidentally with development of major strike‐slip faults in the prism interior, reduced margin‐normal convergence rate, and the onset of impingement on the incoming Chatham Rise to the south. A southwestern segment (C) marks the approximate southern termination of subduction but widens to ∼50 km due to rapid outward migration of the deformation front via fault reactivation within the now‐underthrusting corner of the Chatham Rise. Segment C exhibits minimal shortening as margin‐normal subduction velocity decreases and plate motion is increasingly taken up by interior thrusts and strike‐slip faults.

Full-waveform inversion (FWI) utilizing long-offset ocean bottom node (OBN) data has become a key velocity model building tool for resolving complex velocities for deeper targets. Acquiring such OBN datasets with dense shot and receiver sampling remains costly. Sparse OBN data has been studied in recent years to reduce acquisition expense while maintaining desired benefits for FWI. Whereas sparse OBN data can provide velocities comparable to dense OBN data, it alone cannot provide seismic images and gathers fit for interpretation purposes. We demonstrate a cost-effective approach using joint sparse OBN and streamer FWI that effectively updates the deep velocity of a complex subduction zone in the Hikurangi margin near New Zealand's East Coast. The streamer data is then migrated to provide detailed and accurate images fit for interpretation.

Most conceptual models for how fluids and sediment influence slip behavior and uplift along subduction margins are poorly constrained by geophysical observations. Given the complexity of subduction systems, overcoming this gap in knowledge will require a systems-level approach which uses high quality geophysical constraints. We present wide-angle, onshore-offshore seismic data collected along the northern Hikurangi margin, New Zealand, from which P-wave velocities were calculated using active- and passive-sources. A gravity model and reflection profiles were also assembled to create a complete, ~400 km long transect which images the incoming plate, down going slab, overthrusting forearc, and backarc rift. Velocities and gravity modelling help to constrain the lithology of the forearc basement to ~20 km depth. Upper plate lower crustal velocities and reflectivity point to the presence of underplated sediments immediately above the lithospheric mantle nose, suggesting that underplated sediments are driving uplift of the forearc. Comparing these results to geophysical images from the southern Hikurangi margin, we suggest that the backarc rift influences along-strike changes in the compressional stresses experienced by the forearc, driving changes in bending stresses within the subducting slab.

In addition to the largest documented earthquake, the Mw 9.5 Valdivia earthquake of 1960, the Chilean subduction zone has experienced three great (>Mw 8) earthquakes since 2010 for which detailed models of deformation prior to, during and after the earthquake are available. The 1960 and 2010 Mw 8.8 Maule earthquake affected the south-central Chilean margin, where the trench is sediment filled, whereas the 2014 Mw 8.2 Iquique and 2015 Mw 8.3 Illapel earthquakes occurred farther north where there is little sediment in the trench. We summarize recent results from three marine seismological projects designed to elucidate the impact that geologic structure may have had on slip during these events. Because analysis of these datasets is ongoing, this report should be considered as an interim summary rather than a comprehensive synthesis. None-the-less, some patterns can be discerned. That we can resolve 5 m of margin uplift by differencing swath bathymetric data highlights the value of obtaining ground truth marine geophysical data from potential seismic gaps. Controlled source seismic imaging indicates that subduction of nearly all incoming sediment in the sediment-filled section of the trench is a widespread characteristic of the subduction zone segment that ruptured in the two largest events and is likely a generalizable characteristic that can be used to anticipate whether a subduction zone is likely to rupture in great tsunamigenic earthquakes. While this is not a new idea, our data provide additional insights into plate boundary evolution at depth in this scenario. We also find that subduction of anomalously rough basement topography similar to 2 million years ago may have led to indentation, large scale internal deformation, and erosion of the overlying wedge north of the 2010 rupture zone and that the process of wedge recovery may have resulted in an abrupt offset along strike of the plate boundary that served as a barrier to slip propagation in 2010. In general, structures generating wedge disruption may no longer be detectable following their subduction deep beneath the forearc wedge, but their passage can be inferred from residual structures mapped within the wedge, and their impact on slip during large earthquakes may be long lasting. We anticipate that further analysis of the data from these experiments, and from future experiments that build on these results, will further elucidate the relationship between geologic structure and slip such that images of subsurface structure become an additional tool in the toolbox of observations used to evaluate future hazard.

Recurring slow slip along near-trench megathrust faults occurs at many subduction zones, but for unknown reasons, this process is not universal. Fluid overpressures are implicated in encouraging slow slip; however, links between slow slip, fluid content, and hydrogeology remain poorly known in natural systems. Three-dimensional seismic imaging and ocean drilling at the Hikurangi margin reveal a widespread and previously unknown fluid reservoir within the extensively hydrated (up to 47 vol % H2O) volcanic upper crust of the subducting Hikurangi Plateau large igneous province. This ~1.5 km thick volcaniclastic upper crust readily dewaters with subduction but retains half of its fluid content upon reaching regions with well-characterized slow slip. We suggest that volcaniclastic-rich upper crust at volcanic plateaus and seamounts is a major source of water that contributes to the fluid budget in subduction zones and may drive fluid overpressures along the megathrust that give rise to frequent shallow slow slip.

Abstract Marine multichannel and wide‐angle seismic data constrain the distribution of seamounts, sediment cover sequence and crustal structure along a 460 km margin‐parallel transect of the Hikurangi Plateau. Seismic reflection data reveals five seamount up‐to 4.5 km high and 35–75 km wide, with heterogeneous internal velocity structure. Sediment cover decreases south‐to‐north from ∼4.5 km to ∼1–2 km. The Hikurangi Plateau crust (VP 5.5–7.5 km/s) is 11 ± 1 km thick in the south, but thins by 3–4 km further north (∼7–8 km). Gravity models constructed along two seismic lines show the reduction in crustal thickness persists further east, coinciding with a bathymetric scarp. Gravity data suggest the transition in crustal thickness may reflect spatial variability in deformation and lithospheric extension associated with plateau breakup. Variability in the thickness of subducting crust may contribute to differences in megathrust geometry, upper‐plate stress state and high‐rates of contraction and uplift along the southern Hikurangi margin.

Large seamounts and basement relief cause permanent deformation when they collide with the overriding plate at subduction zones. The resulting structural and compositional heterogeneities have been implicated as controlling factors in megathrust slip behaviour. Subducting seamounts may temporarily lock plates, favouring subsequent large earthquakes. Alternatively, seamounts may redistribute stress, reducing seismic slip. Here we present three-dimensional seismic data from the seamount-studded subducting Hikurangi Plateau along New Zealand’s North Island. We find that one well-imaged seamount, the Pāpaku Seamount, locally uplifts the overriding plate and leaves a tube-shaped lens of sediment trailing in its wake. Anomalously low seismic velocities within and below the Pāpaku lens and along the megathrust fault are consistent with the presence of unconsolidated, overpressured fluid-rich sediments. Similar observations from an older sediment lens, which corresponds to the location of a 2014 slow-slip rupture event, suggest that such overpressures can persist along the megathrust due to delayed drainage out of the subducting plate. The collocation of the 2014 slow-slip earthquake with this sediment lens suggests that these fluid-rich regions define zones that enable slow slip. We hypothesize that sediment lenses left behind by subducting seamounts can create low-effective-stress patches within transitionally stable marine sediment along the megathrust that are conducive to slow slip.

Marine multichannel and wide-angle seismic data constrain crustal structure along a 530 km margin-parallel transect of the Hikurangi subduction zone. The subducting Hikurangi Plateau crust (V-P 5.0-7.4 km/s) is similar to 1 km thicker (11 +/- 1 km) and mantle velocities are similar to 0.2 km/s higher (V-P 8.3-8.5 km/s) beneath south/central Hikurangi relative to north Hikurangi. In the overthrusting plate, an abrupt 0.5 km/s south-to-north reduction in forearc wavespeeds occurs in concert with a change in seismic reflection character. We analyze legacy seismic data to show that the forearc transition likely reflects lateral variability in the updip extent of the Torlesse Backstop. Furthermore, we map this unit along-strike and note a broad correlation between the backstop and down-dip extent of shallow slow-slip. We propose that the geological architecture of the overthrusting plate contributes to spatial variability in the location of shallow frictional transitions along the Hikurangi margin, impacting both seismic and tsunami hazard.

The NZ3D OBS experiment performed in 2017-2018 in the northern Hikurangi margin off the east coast of North Island, New Zealand, provided the highest-resolution seismic refraction/wide‐angle reflection data with multi-azimuth ray coverage in subduction zones to date (Arai et al., 2020). The study area extending 60 km in the trench-normal direction and 14 km in the trench-parallel direction covers source regions of a variety of slow earthquake phenomena, such as shallow slow slip events and tectonic tremor (e.g., Wallace, 2020), and thus offers an ideal location to link our understanding of structural and hydrogeologic properties at subduction faults to slip behavior. We applied an anisotropic traveltime tomography analysis to this active-source dataset from 97 ocean bottom seismographs deployed with an average spacing of 2 km on four parallel lines and dense air gun shooting with a 25 m interval, and succeeded in quantitatively constraining the P-wave velocities (Vp) of the upper plate forearc and the subducting slab and their azimuthal anisotropy in three dimensions. The velocity models revealed some locations with significant Vp azimuthal anisotropy over 5 % near the splay faults in the low-velocity accretionary wedge and the deformation front. This finding suggests that the anisotropy is not ubiquitous and homogeneous within the upper plate, but more localized in the vicinity of active thrust faults. While the fast axes of Vp are mostly oriented in the trench-normal direction in the accretionary wedge, which is interpreted as results of preferentially oriented cracks in a compressional stress regime associated with the plate convergence, they are rotated to the trench-parallel direction on the seaward side of the trench and in the landward backstop. This regional variation is consistent with the results of shear-wave splitting analysis (Zal et al., 2020) and the directions of maximum horizontal stress inferred from the borehole breakouts at two IODP drilling sites (Wallace et al., 2019). The significant magnitudes of anisotropy may indicate that in addition to the crack orientation, clay-rich sedimentary sequences that stack and form coherent strata along the accretionary wedge also contribute to seismic anisotropy in the subduction margin.

We present velocity images across a bottom-simulating reflector (BSR) recovered using 3D high-resolution full-waveform inversion (FWI) and discuss its use as a tool for understanding the nature of the BSR. FWI is a seismic imaging technique which generates highly resolved physical property models of the subsurface. FWI uses the full recorded waveform for inversion which leads to a superior resolution compared to other imaging methods, but also makes it computationally more expensive. Relative to 2D inversions, 3D FWI leads to image improvements due to an increase in azimuthal coverage and ability to map out-of-plane arrivals to the correct location, which is particularly important in a complex geological setting. Therefore, next to the advantage of a fully resolved 3D structure, the model will also be more accurate. Caused by gas hydrate in an upper layer and/or free gas in a lower layer, a BSR indicates the base of the gas hydrate stability zone. This significant change of the physical properties in the upper few hundred meters of the marine sediment produces a distinct reflection, i.e. the BSR, that can be seen in the seismic image. We are imaging and investigating a BSR at Puke Ridge, a thrust ridge on the accretionary wedge of the northern Hikurangi subduction margin, offshore the North Island of New Zealand. We are using seismic multichannel streamer data, belonging to the NZ3D dataset collected in 2018, to invert for the P-wave velocity. The resolved velocity model displays the geometry and the structure of a BSR characterised by a velocity increase followed by a sudden decrease and provides us with accurate velocities which we can use for rock physics modelling and interpretation.

Subduction megathrusts exhibit a range of slip behaviors spanning from large earthquakes to aseismic creep, yet what controls spatial variations in the dominant slip mechanism remains unresolved. We present multichannel seismic images that reveal a correlation between the lithologic homogeneity of the megathrust and its slip behavior at a subduction zone that is world renowned for its lateral slip behavior transition, the Hikurangi margin. Where the megathrust exhibits shallow slow-slip in the central Hikurangi margin, the protolith of the megathrust changes ~10 km downdip of the deformation front, transitioning from pelagic carbonates to compositionally heterogeneous volcaniclastics. At the locked southern Hikurangi segment, the megathrust forms consistently within pelagic carbonates above thickened nonvolcanic siliciclastic sediments (unit MES), which subduct beyond 75 km horizontally. The presence of the MES layer plays a key role in smoothing over rough volcanic topography and establishing a uniform spatial distribution of lithologies and frictional properties that may enable large earthquake ruptures.

Abstract High‐resolution bathymetry and three‐dimensional seismic data along the Cocos Ridge reveal a 245 km2 field of ∼1–4 km in diameter seafloor depressions. The seafloor depressions are part of a two‐tiered honeycomb pattern. The lower‐tier depressions have steep faults that truncate strata with chaotic internal reflections consistent with sediment collapse into the depression. These extend into a lens shaped interval just above igneous basement. Overlying these depressions is a second broader set with rough seafloor morphology with gently dipping boundaries defined by pinch‐out stratigraphic patterns. Drilling results indicate that the lens‐shaped zones that host the deeper depressions represent anomalous regions of high porosity, low velocity, and low density within calcareous rich sediment. Analysis of nannofossils from IODP Site U1414 suggests the collapse structures formed during the late Miocene, whereas the younger shallower depressions likely formed between the early Pliocene and the Pliocene‐Pleistocene boundary. Geochemical and petrological analysis at Site U1414 suggests that hydrothermal circulation during the late Miocene led to carbonate dissolution and collapse. Following collapse, focused fluid‐flow and bottom current scouring resulted in formation of the overlying set of depressions and a honeycomb seafloor morphology. Similar sets of depressions along the Carnegie Ridge to the south support the hypothesis that two‐tiered depressions formed in response to processes that occurred broadly across the Panama Basin between the late Miocene and the Pliocene‐Pleistocene transition. Geochemical results at Site U1414, combined with geophysical data, suggest this two‐tiered system of depressions currently guides ongoing fluid outflow.