Magnetic minerals serve a dual role in Earth sciences, acting as key tracers in the global iron cycle and fundamental recorders of sedimentary paleomagnetic signals. However, in marine settings, such signals are often compromised by diagenetic alteration. Hydrogenetic ferromanganese (Fe-Mn) crusts, which grow on sediment-free seamounts throughout the global ocean, provide a unique archive that avoids post-depositional overprinting. Here we present high-resolution rock magnetic data from a 4.9 Ma Fe-Mn crust recovered from a seamount in the South China Sea. Our results reveal three distinct phases of magnetic mineral assemblage evolution. From 4.9 to 3.6 Ma, biogenic magnetite dominated, corresponding to low magnetic susceptibility. Between 3.6 and 2.6 Ma, magnetic susceptibility and magnetic grain size increased substantially, reflecting a rising contribution of terrigenous detrital multidomain (MD) magnetite alongside the biogenic component. After 2.6 Ma, both susceptibility and grain size declined as MD magnetite content decreased. We attribute the enhanced MD magnetite flux during 3.6-2.6 Ma to intensified physical weathering of continental source rocks, triggered by strengthened East Asian Summer Monsoon precipitation following the onset of Northern Hemisphere glaciation (NHG). This monsoon-driven erosion likely increased fluvial supply of bioavailable Fe(II)-bearing magnetite to the ocean, potentially stimulating phytoplankton productivity and contributing to atmospheric CO2 drawdown and global cooling.

Reconstructing dissolved oxygen (DO) variability in the deep ocean is critical for understanding the interplay between oceanic carbon storage and changes in climate over glacial-interglacial cycles. However, quantitative proxies that reliably resolve past DO fluctuations remain limited. Here, we develop an empirical Gompertz model framework to establish a nonlinear relationship between hydrogenetic iron concentrations in ferromanganese (Fe-Mn) crusts and ambient seawater DO levels. The derived relationship is expressed as yFe201 wt% e eDO () . 1 207 0 013 with . . = & times; - & times;- & times;, R2 = 0.704. We demonstrate that Fe-oxyhydroxide precipitation in these crusts reflects a dynamic equilibrium governed by surface ocean iron fluxes and DO-dependent oxidation kinetics. Applying this proxy framework, we present the first quantitative reconstruction of deep-water DO variations in the western Pacific over the past 450 k.y. Our record reveals recurrent and pronounced DO depletions during glacial periods and terminations across the last three glacial cycles. Quantitative estimates further indicate substantially enhanced respired carbon sequestration in the abyssal western Pacific during the Last Glacial Maximum. These results constrain the spatiotemporal heterogeneity of DO variability and respiratory carbon distribution across Pacific water masses, thereby advancing our mechanistic understanding of oceanic respiratory CO2 partitioning during glacial cycles.

The formation and evolution of the oceanic lithosphere are controlled by complex magmatic processes that remain subjects of ongoing research. Studies on Plagioclase Ultraphyric Basalt (PUB) have greatly expanded our understanding of these processes over the past two decades. In this study, we report a novel type of PUB sampled from the ultraslow-spreading Southwest Indian Ridge (SWIR), revealing two distinct mineral assemblages that coexist within individual samples. Group 1 minerals, dominated by high-anorthite (An) plagioclase macrocrysts with core An values up to 86 and high-Mg olivine with core Fo (molar Mg/(Mg+Fe) x 100) values up to 88, are characterized by cotectic crystallization. They represent deep, anorthositic mush zones that have undergone multiple magma replenishments. In contrast, Group 2 minerals, consisting of clinopyroxene oikocrysts and plagioclase chadacrysts, display features of fractional crystallization from shallower gabbroic mush zones. The broad range of 87Sr/86Sr ratios in Group 1 plagioclase macrocrysts (0.70280-0.70375) and Group 2 plagioclase chadacrysts (0.70285-0.70388), coupled with a narrower isotopic range in matrix plagioclase laths (0.70308-0.70384), suggests that magma diversity was well preserved within these crystal mushes. Moreover, significant magma mixing from variable sources occurred at mid-crustal levels. These findings illustrate the presence of polybaric, heterogeneous mush zones, with slow crystallization occurring at various depths beneath the SWIR. This study highlights the intricate interplay among multiple magma sources, replenishment events, and crystallization environments, providing new insights into the formation and variability of the oceanic crust at ultraslow-spreading ridges.

At mid-ocean ridges, volcanic activity is predominantly marked by the voluminous effusion of tholeiitic basaltic lavas, with sporadic occurrences of mildly alkalic basalts. However, the genetic link between voluminous tholeiitic basalts and small-volume alkali basalts remains enigmatic. We report both alkaline and tholeiitic volcanism at the Marion Rise segment of the ultraslow spreading (14 mm/yr) Southwest Indian Ridge (SWIR). In contrast to the effusive tholeiitic (MORB) volcanism in Dr 27 along the Marion Rise, the Dr 30 samples consist of transitional to alkalic glass and alkaline scoria with E-MORB-like affinity (e.g., with high K/Ti, volatile contents (e.g., CO2 and H2O), La/Sm and radiogenic heavy isotopes). Critically, the petrological and geochemical evidence of melt inclusions and their host minerals suggests shallow mixing of a volatile-rich low-viscosity alkaline magma with entrained MORB crystal mush. The MORB mush represents a later volatile-poor viscous melt erupted effusively from beneath the axial valley onto the seafloor. This indicates bimodal magmatism beneath ocean ridges with the generation of early-formed alkaline melts that ascend independently of the far more voluminous tholeiites, which then interact and mix in the melt storage region in the ocean lithosphere. Mixing of small volume alkaline melts with more voluminous tholeiitic melts then explains the local major element uniformity of MORB and its isotopic and trace element diversity. We suggest this may apply globally as at magmatically more robust ridges the role of alkali basalt is likely masked by the far more voluminous MORB.

Information about the microbiota in deep-sea seamount sediments is important because the microbiota and their activities in sediments affect deep-sea ecosystems. To evaluate deep-sea seamount microbial diversity, we performed 16S rRNA gene amplicon sequencing on sediment samples from 10 stations in the South China Sea.

Tectonic extension at slow- to ultraslow-spreading mid-ocean ridges is predominantly accommodated by large- offset detachment faults. Nonetheless, the interplay between these detachment faults and melt supply remains an area of active research. In our study, we undertook a comprehensive analysis of a narrow shear zone and its wall rock, both retrieved from the International Ocean Discovery Program (IODP) Hole U1473A at the Atlantis Bank Oceanic Core Complex (OCC) on the Southwest Indian Ridge. Our methodology encompassed microstructural observations, electron probe analysis, and electron backscatter diffraction. Our findings indicate that an oxide- saturated melt permeated the semi-brittle lithosphere and the fault plane, acting as a lubricant and significantly diminishing the fault's frictional strength. Subsequent deformation-triggered melt segregation led to a 3-4-fold reduction in the viscosity of the footwall gabbros. This reduction likely induced a listric geometry, driven by the reorientation of local stress fields and unusually elevated melt pressure. In contrast to numerical model predictions, our results suggest that the development of oceanic detachment faults is chiefly controlled by synmagmatic strain localization, potentially linked to the rate of melt emplacement and the thickness of the axial lithosphere.

Seafloor hydrothermal venting may be an important source of marine Cu and affect the biogeochemical cycling of Cu in the oceans. The distribution of Cu and its isotope compositions (delta Cu-65) can provide insight into seafloor hydrothermal processes and their role in the mass balance of global Cu. To date, there are no published Cu isotope data for hydrothermal plumes and very few reports on Cu concentration distributions. This study presents both Cu concentrations and dissolved Cu isotope (delta Cu-65) in hydrothermal plumes from back-arc volcanoes in the Northeast Lau Basin. The dissolved Cu (dCu) concentrations range from 2.14 to 5.66 nM most of which are higher than the deep seawater concentrations. However, the concentrations for some plume samples were lower than the deep seawater dCu concentrations due to adsorption onto particles in the plumes. The delta Cu-65 of hydrothermal plumes vary from 0.25 to 1.05 parts per thousand. As plumes dispersed, the Cu isotopes from high-temperature Mata Fitu and Mata Ua vents shifted towards light values. In contrast, the delta Cu-65 in plumes from low-temperature East Mata and West Mata vents show increasingly higher values with plume dispersal. Rayleigh distillation models are built based on the adsorption of dCu onto Fe particles and complexation with organic ligands to describe the dCu isotope evolution in hydrothermal plumes. The results suggest that the adsorption and organic complexation may be the likely explanations for the observed dCu isotope compositions in plumes from high- and low-temperature venting, separately, besides the mixing with background seawater. Our measured delta Cu-65 in three fluid samples (0.08, 0.09 and 0.20 parts per thousand) are lower than that of the characterized sinks (similar to 0.3 parts per thousand) in the oceans, indicating that hydrothermal fluids might be a source of light Cu isotopes. However, the delta Cu-65 (0.53 parts per thousand on average) in plume samples are higher than 0.3 parts per thousand and these seafloor hydrothermal plumes do not seem to be a source of light Cu isotope of dCu, at least near the venting sites. Our study first reveals the Cu isotope compositions and evolution in hydrothermal plumes and provides new hints about the impact of seafloor hydrothermal venting on the modern oceanic Cu isotope budget.

新生代深海铁锰矿床的大规模成矿是地质历史上特有的现象,其形成的海底铁锰结核/结壳因富含巨量的有用金属而备受关注.水成型铁锰成矿的胶体成因模型自20世纪90年代中期提出以来已被广泛接受并采用.随着近20年来纳米地球科学的迅速发展,人们意识到纳米颗粒作为胶体的最小部分,能够以其独特的性质显著影响铁锰成矿过程.通过总结已有研究,发现铁氧化物与锰氧化物会以纳米颗粒的形式普遍共存于多种表生地质环境,还证实了水成型铁锰结核/结壳中的主要铁锰矿物(如水羟锰矿和水铁矿)都是纳米颗粒.铁氧化物纳米颗粒对二价锰[Mn(Ⅱ)]的表面催化氧化可能是水成型铁锰矿物通常在纳米尺度密切共生的原因.此外,在铁锰结壳中还观测到大量在以往研究中被普遍忽视的三价锰[Mn(Ⅲ)]矿物,其含量在结壳顶部最高,随深度增加逐渐下降,四价锰[Mn(Ⅳ)]矿物的含量则呈相反的变化趋势.不同价态锰氧化物纳米颗粒的表面能差异导致Mn(Ⅲ)矿物在Mn(Ⅱ)的氧化过程中最先沉淀,并可能在沉淀之后逐渐转化为Mn(Ⅳ)矿物.相信随着纳米地球科学与高精度原位实验技术的发展,必将不断深化对海水铁锰循环及海底铁锰成矿的认识.

Although there are approximately 25 million seamounts in the ocean, surprisingly little is known about seamount microbial ecology. We provide evidence that seamounts are island-like habitats harboring microbial communities distinct from those of nonseamount habitats, and they exhibit a distance-decay pattern.

铁锰结核中主微量元素的赋存形式研究对于理解铁锰结核的形成及其中元素的进入过程具有重要意义.选择采集自南海蛟龙海山水深约3 300m的3个铁锰结核,根据其内部剖面上原位X荧光光谱分析(XRF)的主量元素含量差异分层进行了分层取样.对于分层样品开展了淋滤实验,分别提取样品中的碳酸盐相、锰矿物相、铁矿物相和残渣相组分,并将该4种组分以及原样中主微量元素加以分析比较.结果表明,蛟龙海山结核多数主要元素赋存方式基本上类似于大洋水成型结核结壳,但相对于大洋水成型结核结壳,Al、K、Mg、Li、Ti在碎屑相中赋存比例更高,呈现出边缘海水成型结核元素赋存方式的特点.圈层之间部分元素赋存形式略有变化,其中Mg、Cu、Ni、Zn从核心到边缘在锰矿物相中的赋存比例随着Mn/Fe比值的升高而增加,体现了不同生长阶段结核元素和矿物组成的变化.即Mn/Fe比值越高,说明锰氧化物/铁氧化物比值越高,而Mg、Cu、Ni、Zn主要赋存于锰氧化物中,所以Mg、Cu、Ni、Zn赋存于锰矿物相/铁矿物相的比例增加.此外,通过与前人关于南海蛟龙海山结核淋滤结果的对比发现,实验试剂与反应时间对于结核淋滤结果有较大影响.

Deep-sea hydrothermal venting is an important source of dissolved iron (dFe) to the oceans. Fe isotopes can be used as a potential tool to trace the dispersal of hydrothermal plumes. However, Fe isotope fractionation and its relation with Fe speciation as hydrothermal plumes disperse is still poorly constrained. In this study, we determined the Fe speciation and total and dissolved Fe isotope composition (delta 56tFe, delta 56dFe) for several hydrothermal plumes from backarc volcanoes in the Northeast Lau Basin. This combined approach provides important insights into the evolution of Fe isotopes in hydrothermal plumes. The results suggest delta(56)tFe variation in plumes is related to the loss of particulate Fe-sulfides or Fe-oxyhydroxides (FeOOH), both of which are dependant on the H2S concentrations and Fe/H2S in the source hydrothermal fluids. delta(56)dFe compositions in the hydrothermal plumes increase during plume dispersal/dilution and can be as high as 0.85 parts per thousand, demonstrating that hydrothermal plumes can export dissolved Fe with a significantly heavier delta(56)dFe than hydrothermal fluids. The reasons may be ascribed to the organic Fe complexes (FeL) and colloidal FeOOH in the dissolved phase. Another interpretation might be associated with the low pH in volcanic arc hydrothermal systems rich in magmatic CO2 and SO2, which decreases the Fe(II) oxidation rate. Further, we demonstrate for the first time that the delta(56)dFe is positively correlated with the conditional stability constants of FeL (logK'(FeL)). A Rayleigh distillation model is presented based on the mass balance of the determined FeL, and colloidal FeOOH in hydrothermal plumes, which can explain the observed Fe isotope compositions in hydrothermal plumes. Our data show how Fe isotopes are transformed within a hydrothermal plume above arc volcanoes and how these may differ from that of the original vent fluids. It adds to our understanding of the processes that have an impact on the Fe speciation and isotope composition in deep-sea hydrothermal plumes.(c) 2022 Elsevier Ltd. All rights reserved.

Geochemical heterogeneities observed in the mantle are usually attributed to recycling of oceanic lithosphere through subduction. However, it remains hotly debated where recycled material stagnates, and how quickly it can be liberated back to surface. This knowledge gap hinders our understanding of mantle circulation and the chemical evolution of the Earth. Here we address these questions using a combination of geochronology and geochemistry from South China Sea (SCS) seamounts. The Shixingbei seamount lavas formed during active seafloor spreading at c. 19.1 Ma show limited geochemical variability, whereas the Zhenbei-Huangyan seamount chain formed during the post-spreading stage at c. 7.8 Ma and displays a wide range of compositions. However, melt inclusions in olivine and plagioclase from the Zhenbei-Huangyan basalts show considerably greater isotopic variability than seen in the whole rock compositions of both the SCS syn- and post-spreading lavas. A previously unidentified third mantle source component (FOZO) revealed by olivine-hosted melt inclusions along with both depleted (DMM) and enriched (EMII) mantle components is required in the source region to explain the observed isotopic and chemical variability. On the basis of our results, the age of the recycled ocean crust and sediments in this region are estimated to be c. 120 – 350 Ma. We infer that these enriched components in the SCS lavas come from the mantle transition zone. Variations in mantle source heterogeneity coupled with melting process control spatial–temporal (spreading vs. post-spreading stage) geochemical variations of lavas from the SCS and surrounding areas. Together with the results from published studies, we propose that marginal basins are one of the major locations on Earth where oceanic and/or continental lithosphere is transferred into the upper mantle and transition zone, representing an important source of upper mantle heterogeneity. We provide a simple conceptual model linking plate subduction and upper mantle heterogeneity and the volcanism in the SCS and surrounding areas.

深海铁锰结壳的定年对其记录的百万年尺度古海洋环境变化研究至为关键.综合运用10Be/9Be、Co经验公式、230Thex/232Th和磁性地层学,对采自加瓜海脊的铁锰结壳样品开展了系统的年代学对比研究.结果表明:相对于开阔大洋的铁锰结壳,较多的陆源物质输入造成了不同定年方法获得的年龄或生长速率的明显差异.其中,因为大量陆源物质携带的232Th以及对Co含量的稀释,铁锰结壳表层的230Thex/232Th初始通量以及样品部分层位的Co通量出现显著变化,230Thex/232Th定年方法与Co经验公式获得的结果受到碎屑物质的影响最为显著.尽管10Be/9Be初始通量也受到了陆源物质输入的影响,但是10Be/9Be初始通量变化很小,应该是本研究中最为可信的结果.而古地磁地层学定年法需要参考其他定年结果,最后也只能得到几个年龄控制点.最终得出加瓜海脊该铁锰结壳样品的年龄为7.09 Ma,而不同核素在铁锰结壳中的赋存状态应该是今后值得深入研究的一个重要方向.

海底热液成因含金属沉积物广泛分布于全球各大洋与弧后活动扩张中心、大洋玄武岩上覆沉积层的底部以及板内火山的顶部等区域.块状硫化物烟囱体经氧化蚀变发生再沉积作用、热液羽流的扩散和沉降作用或低温弥散流的直接沉淀均可形成含金属沉积物.尽管不同热液区的含金属沉积物在矿物和化学组成上具有一定的差异,但其相对正常远洋沉积物均表现为富含Fe、Mn并亏损Al和Ti等组分,其中的主要矿物通常为结晶程度较差的铁锰氧化物/氢氧化物和富铁蒙脱石(绿脱石).自二十世纪七十年代末发现现代海底热液活动以来,大量的研究不仅基本明确了含金属沉积物的形成机制,还在与之相关的微生物矿化作用和自生黏土矿物的成因研究等方面取得了重要的进展.对含金属沉积物开展综合研究,确定其鉴别分类标准,既能丰富人们对海底热液循环系统、热液活动对全球海洋热和化学通量的贡献以及对海底深部生物圈的认识,也可为寻找多金属硫化物矿床、揭示古板块中类似矿床的成因和分布规律以及探索古海洋环境演化等提供重要信息.

安山岩是俯冲带岩浆弧中重要的岩石类型,其成因至今仍是国际地质学界研究的热点之一.根据安山岩分布的地质背景,岩浆弧安山岩可以简单划分为陆弧安山岩和洋弧安山岩,二者在化学成分和空间分布等方面存在明显差异.自20世纪20年代末以来,大量的研究成果丰富了人们对岩浆弧安山岩成因的认识,逐渐将其归纳为玄武质岩浆输入和安山质岩浆输入2种模型.玄武质岩浆输入模型认为形成岩浆弧安山岩的初始岩浆成分为玄武质,强调分离结晶、同化混染和岩浆混合等壳内过程;而安山质岩浆输入模型认为地幔源区可以直接形成安山质熔体,强调俯冲板片来源的流体/熔体—地幔橄榄岩的交代反应和沉积物底辟等壳下过程.虽然岩浆弧安山岩的成因研究取得了一定的进展,但每一个模型都有亟待完善之处.地幔交代岩的实验岩石学研究、安山岩与大陆地壳形成与演化间的关系、理论计算和模拟的应用等都是未来需要研究的领域.

Basaltic lavas sample recycled crustal materials from their mantle source. Constraining the location and residence time of these recycled materials in the mantle is critical to understand global mantle dynamics. In this study, we present new whole‐rock major and trace element abundances, Sr‐Nd‐Mo‐Os isotopes, water contents and He isotopes of volcanic glasses, U‐Pb ages of zircons, and compositions of melt inclusions, spinels and olivines from the South China Sea (SCS) seamounts lavas. These new data are compared with literature data from intraplate volcanism of similar age from Southeast (SE) Asia. The isotope data of late Cenozoic lavas from the SCS seamounts and SE Asia can be explained by mixing between enriched mantle 2 (EM2) and depleted mid‐ocean ridge basalt mantle components. Our data are consistent with the EM2 signature of late Cenozoic lavas derived from recycled young oceanic crust and sediments. The compositions of olivine phenocrysts indicate an olivine‐dominated (peridotitic) mantle source. There is currently no evidence for a high‐ 3 He/ 4 He mantle plume component beneath the SCS. Our results combined with geophysical data and plate reconstructions suggest that the late Cenozoic magmatism is related to the upwelling of instabilities from the mantle transition zone (MTZ) triggered by a stagnant slab. The SCS seamount lavas sample an enriched MTZ containing young recycled materials, consistent with regional past subduction. Our study provides additional evidence that storage and recycling of crustal materials in or near the MTZ is an important mechanism to develop global mantle heterogeneities sampled by intraplate volcanoes.

Here we report the discovery of mosaic zircons in Hole U1473A on the Atlantis Bank in the Southwest Indian Ridge. Oxygen isotope, U-Pb dating, and geochemical analyses of the zircons were carried out directly in petrographic thin-sections. The mosaic zircons are from diorite and oxide gabbro at shallow depths, whereas the nonmosaic zircons occur in oxide gabbronorite and tonalite at greater depths. The majority of zircons have delta O-18 of 5.5 +/- 0.1 parts per thousand, which are mantle-like values, reflecting formation by simple fractional crystallization. No correlation exist between U-Pb ages, delta O-18 values, and the mosaic microstructures. The mosaic subdomains in a single zircon are well correlated with some trace elements and fractures. Therefore, the variable trace element contents were produced during crystallization, and associated with secondary brittle deformation due to movement on mid-ocean ridge detachment faults. The pervasive brittle deformation weakened the originally rigid zircon and triggered fracturing. The fractures became highly permeable pathways, allowing rapid grain-boundary diffusive loss of Pb, which resulted in the relatively young age of 11.42 +/- 0.41 Ma for the mosaic zircons as compared with 12.16 +/- 0.14 Ma for the non-mosaic zircons. The zircon trace element patterns are all indistinguishable from global oceanic zircons. However, the trace element abundances and ratios require a significantly depleted normal-type mid-ocean ridge basalt mantle source. (C) 2021 Elsevier B.V. All rights reserved.

The tectonic history of the Philippine Sea plate is an essential piece in understanding the tectonic evolution of Southeast Asia, but it is still unclear and controversial. We present the first geochemical data obtained from lavas from the Gagua Ridge (GR) within the Philippine Sea. The GR lavas exhibit geochemical signatures typical of subduction-related arc magmatism. Plagioclase Ar-Ar ages of ca. 124–123 Ma and subduction-related geochemical signatures support the formation of GR lavas in the vicinity of an arc during the Early Cretaceous induced by subduction of the oceanic plate along East Asia. The ages of trapped zircon xenocrysts within the GR lavas cluster at 250 Ma, 0.75 Ga, and 2.45 Ga and match well the ages of zircons recovered from the Cathaysian block, southern China. Our results imply that the GR basement is partially composed of continental material that rifted away from the Eurasian margin during opening and spreading of the Huatung Basin. The depleted mantle wedge-derived magmas evolved and picked up the continental zircons during ascent. The youngest zircon ages and the GR lava Ar-Ar ages (ca. 124–123 Ma) presented in this study newly constrain an Early Cretaceous age for the Huatung Basin. Our study provides further evidence that the Huatung Basin is a remnant of a Mesozoic-aged ocean basin that dispersed from southern China during the Cretaceous. Transport of continental slivers by growth and closure of marginal seas along the East Asia margin may have been more prevalent than previously recognized.

Abstract Drilling 809‐m Hole U1473A in the gabbro batholith at the Atlantis Bank Oceanic Core Complex (OCC) found two felsic vein generations: late magmatic fractionates, rich in deuteric water, hosted by oxide gabbros, and anatectic veins associated with dike intrusion and introduction of seawater‐derived volatiles. Microtextures show a change from compressional to tensional stress during vein formation. Temperatures and oxidation state were obtained from amphibole‐plagioclase and oxide pairs in the adjacent gabbros. Type I veins generally have reverse shear‐sense, with restricted ΔFMQ, high Mt/Ilm ratios, and low‐amphibole Cl/F indicating deuteric fluids. They formed during percolation and fractionation of Fe‐Ti‐rich melts into the primary olivine gabbro. Type II veins are usually hosted by olivine gabbro, occur at dike contacts and the margins of normal‐sense shear zones. They are undeformed or weakly deformed, with highly variable ΔFMQ, low Mt/Ilm ratios, and high‐amphibole Cl/F, indicating seawater‐derived fluids. The detachment fault on which the gabbro massif was emplaced rooted near the base of the dike‐gabbro transition beneath the rift valley. The ingress of seawater volatiles began at >800°C and penetrated at least ~590 m into the lower crust during extensional faulting in the rift valley and adjacent rift mountains. The sequence of the felsic vein formation likely reflects asymmetric diapiric flow, with a reversal of the stress regime, and a transition from juvenile to seawater‐derived volatiles. This, in turn, is consistent with fault capture leading to the large asymmetries in spreading rates during OCC formations and heat flow beneath the rift mountains.