The persistence of mountainous topography in tectonically quiescent settings remains one of the outstanding questions in geomorphology. Along the northern passive continental margin of the South China Sea (SCS), although topography is expected to be entirely eroded under a theoretical relief-decay threshold of ca 1 cm/kyr, mountainous terrain has persisted since the Oligocene without significant tectonic uplift. This long-term persistence of topography is attributed either to the preservation of relict features under subdued erosion rates or to a dynamic equilibrium between rock uplift and spatially variable erosion. Distinguishing between these two scenarios remains difficult due to a lack of multi-domains erosion rate constraints across the mountainous interior. In this study, we address this issue by utilizing four in-situ cosmogenic 10Be depth profiles to quantify erosion rates across representative geomorphic domains along the northern coast of the SCS, specifically flat-top surfaces, bare-rock and regolith-mantled hillslopes, and pediment platforms. We evaluate three numerical inversion models, including steady-state erosion, continuous exposure, and abrupt mass loss, using chisquare statistics, solution stability, and field observations. The optimal solutions show that erosion rates exceed the relief-decay threshold on regolith-mantled hillslopes (ca. 8.12 cm/kyr) and pediment platforms (ca. 3.38 cm/ kyr), match the threshold on flat-top surfaces (ca. 1.1 cm/kyr) and fall below it on bare-rock hillslopes (ca. 0.73 cm/kyr). These results demonstrate that erosion rates across the geomorphic domains exhibit pronounced spatial variability, with the majority of these geomorphic domains actively eroding at rates above the theoretical reliefdecay threshold. The above-threshold erosion, combined with the order-of-magnitude convergence among catchment-wide erosion rates, thermochronology-based erosion rates, and theoretical isostatic rock uplift rates, demonstrates that a dynamic equilibrium between rock uplift and spatially variable erosion is more likely to drive the long-term persistence of mountainous topography along the northern passive continental margin of the SCS.

Seismic electric precursors (SEPs) and their related earthquakes are two different geophysical phenomena observed on the ground surface. The temporal relationship between SEPs and earthquakes observed from several historical cases in China and Japan suggests that both were resulted from the same geological process, and as evidenced experimentally, in association with stress accumulation and release in a deeper crust. Here a model named the fault electric field (FEF) is proposed to account for this geological process and its relationship with SEPs and earthquakes. The FEF model indicates that slip along an undulated fault plane causes the asperities on both hanging wall and footwall of fault to close together, thus results in stress concentration around the asperities, forming a seismogenic region (SR); simultaneously, a pair of rock batteries develop on the either side of fault plane, producing a fault electric field (FEF) centered around the SR, which is related to the SEP observed on the ground. When the deflection angle of the normal stress axis exceeds a threshold, rapid slip of the fault with sudden release of elastic energy during earthquake happens. The FEF model effectively explains the geophysical observations on the SEP-earthquake relationship and the experimental investigations on the mechanic-electric effect of rocks. For testing the model, five FEF surveillance stations have been employed on different faults in the Pearl River Delta, South China, and the monitor results from those stations demonstrate the existence of the FEF in each surveyed fault. This research provides a feasible way to detect the process of stress concentration in the deeper segment of a seismogenic fault, which is significant for the earthquake prediction.

The formation of continental red beds is generally considered to be related to an arid climate. Heating experiments (performed by L.J. and G.C.) using dried black mud sediment also demonstrate that the reddening may be caused by the transformation of goethite to haematite that begins at approximately 150 degrees C under anhydrous conditions, and increasing the temperature to 450 degrees C is positively correlated with the red colour and peak value of haematite. If this process applies to continental red beds, it implies a thermal origin of red beds as a result of high diagenetic temperatures rather than as the cause of their deposition under an arid climate. Namely, subsiding red -bed basins are heated from below rather than warmed from above. Here, we further strengthen this idea by new evidence from borehole cores drilled from red beds in SE China, showing a clear geological section from the surface soil to red beds to bottom granite. The data reveal that the continental red beds formed at least at a temperature within 150 - 400 degrees C, and the underlying granite usually formed at temperatures greater than 600 degrees C. Our results imply a possible relationship between continental red bed events and Earth ' s thermal cycles. Copyright (c) 2024, Guangzhou Institute of Geochemistry. CAS. Published by Elsevier BV. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Tectonically induced liquefaction and the resulting soft-sediment deformation structure (SSDS) can provide useful information on paleo-earthquakes, which is vital for the assessment of geohazard susceptibility in tectonically active regions. In this study, we combined sedimentary and chronological methods to the reveal detailed characteristics of the mottled clay in the Huizhou Quaternary Basin. The dating results suggest that mottled clay usually developed during the late Pleistocene, overlying the fluvial deposit or embedding homogeneous aeolian yellow silt. Mottled clay has a typical bimodal frequency distribution with modal sizes at 5 and 80–90 μm, which are identical to those of the yellow silt and the underlying fluvial sand, respectively. Micro-X-ray fluorescence mapping revealed high concentrations of Fe and Si in the red and white fraction, respectively. In addition, the red fraction of mottled clay has a high hematite content, similar to loess-like yellow silt, whereas the white fraction and the underlying fluvial sediments are dominated by goethite. This sedimentary evidence together suggests that the mottled clay could be an admixture of aeolian yellow silt and the underlying fluvial sand. Furthermore, diverse deformed structures (e.g., fragmented structures, sand veins, sand dykes and flame structures) were observed in mottled clay. Therefore, we suggest that the mottled clay structure in the Huizhou Basin is a product of liquefaction-induced SSDS. Tectonic activity was considered to have triggered the liquefaction and SSDS, which is supported by the close spatial relationship between the mottled clay and regional faults. We propose that the SSDS of mottled clay could be a potential indicator of paleo-earthquakes in the coastal Quaternary basins of the northern South China Sea.

The Pearl River Delta (PRD, China) has undergone complex geological development within a multi-island faulted basin, shaped by the interplay of regional tectonic movements, Quaternary sea-level fluctuations, and fluvial-marine interactions. Despite a great number of studies on the Holocene sedimentary sequences and spatial differences of lithofacies and environments, scant attention has been paid to the overarching human influence on deltaic evolution and coastline modifications since the Neolithic epoch. To further elucidate the spatial variation in Holocene sedimentation and its underlying basement topography shaped during the Last Glacial Maximum (LGM), we compiled a comprehensive dataset incorporating borehole data from over 2800 cores (the maximum depth can reach 92.5 m) within the PRD. Subsequently, high-resolution isobath maps of Quaternary deltaic deposits were generated, offering unprecedented insights into sediment distribution. This dataset facilitated a nuanced reconstruction of pre-Holocene topography, revealing a zone characterized by elongated, deep-incised valleys governed by NW-SE fault orientations. Further, we delineated coastline shifts since the period of maximum Holocene transgression (~7000 years BP), contributing to an enhanced understanding of the formation and evolutionary patterns of the delta and river network oscillations. Our findings illuminate an increasing anthropogenic impact on the rate of fluvial sedimentation and land growth, particularly accentuated over the last two millennia, favoring deltaic accretion.

Crustal anatexis is an important process in the tectonic evolution of many orogenic systems. In the Wuyi-Yunkai orogen in the South China block, the duration of partial melting and its relationship with orogenesis are poorly constrained. In this study, we present a multifaceted approach to determine the timing of anatexis and unravel the petrogenesis of Fuhuling migmatites in the Yunkai region of the southwestern South China block. Geochemical analyses indicate that the migmatites have (meta-)sedimentary protoliths. The absence of anhydrous peritectic minerals but the presence of microstructural and outcrop-scale indicators of partial melting suggest that the Fuhuling migmatites experienced fluid-present melting. Complex zoning and variable trace element concentrations in newly formed zircons in migmatites reflect their evolutionary histories during partial melting. Anatectic melt was present at Fuhuling in the Yunkai region from ca. 449–427 Ma during early Paleozoic Wuyi-Yunkai orogenesis. The wide variety of morphologies observed in the Fuhuling migmatites implies that migmatites in the Yunkai region experienced incipient partial melting, melt segregation, and melt migration. Combining new and previous results, we argue that the Yunkai region experienced two stages of crustal anatexis during the early Paleozoic, which may have been triggered by crustal thickening followed by rapid exhumation and orogenic collapse during the intra-plate Wuyi-Yunkai orogeny in the South China block.

While understanding the long-term slip rate of active normal faults is essential for the comprehensive assessment of seismic activity, it is difficult due to the absence of age control in the erosional bedrock region. The preserved sequence of wave-cut platforms in granite allows exploration of the long-term slip rate in the footwall of some normal faults. We investigated wave-cut platforms in the southern Pearl River Delta (PRD), a coastal delta transected by the seismically active Littoral Fault Zone (LFZ) in the northern South China Sea, to derive slip rates and their impacts on the seismic hazard potential. We mapped a flight of four wave-cut platforms (T 1 –T 4 ), dated the T 2 and T 4 platforms by 10 Be cosmogenic nuclide dating, and used the absolute age to correlate the un-dated platform to global sea-level highstands. Our results allocate the ages of 128 ka, 197 ka, and 239 ka to the upper three wave-cut platforms and yield temporally various uplift rates ranging from 0.30 to 0.38 mm/a during 239–128 ka to 0.09 mm/a since 128 ka. A decrease in the uplift rate, which coincided with a decreased subsidence rate within the PRD in previous work, implied a weakened differential uplift onshore of the LFZ system. Our findings infer that the transgression event occurred as early as marine isotope stage (MIS) 7 in the PRD, consistent with the view that Pleistocene sedimentation began in MIS 5 or earlier in the PRD.

Hematite and goethite, the two most common iron oxides in nature, are widely distributed in sediments. Their relative content relationship can reflect the sedimentary environment and provide provides a basis for origin discrimination. Due to the complex operation and low efficiency of traditional methods, it is difficult to quickly and accurately determine iron the species and content of iron oxide within the sediments. Recently, diffuse reflectance spectroscopy (DRS) based on ultraviolet-visible-near infrared spectrophotometer has been widely used in sediments because of its simple operation, fast test and low detection limit. A set of Last Glacial yellow silt, sometimes mixed with red and gray and known as "mottled clay", is widely developed in the late Quaternary basins of Fujian and Guangdong Provinces in the coastal areas south China. This layer was often attributed to exposed weathering of the underwater sediments during the global low sea-level period. However, there is no transition between mottled clay and its underlying deposit, which is difficult to explain by weathering. Moreover, marine fossils rich in the underlying layer are not found in the mottled clay layer, indicating great differences in the sedimentary environment and provenance between these two layers. In order to further determine the sedimentary environment and origin of the mottled clay, four Quaternary drill cores in the Pearl River delta with the method of DRS are analyzed from the perspective of iron mineral characteristics in this study. The results show that the peak value of hematite within the mottled clay is higher than that of goethite, suggesting that the sample is rich in hematite and relatively low in goethite. This trend is opposite to that of the underlying sediments. Hematite is formed in a dry, warm and onshore oxidation environment, where as goethite is the product of long-term wet and underwater reduction conditions. Hence, the mottled clay had not undergone long-term hydration transformation and is therefore not formed by weathering of in-situ underwater deposition but constitutes a subaerial exotic dust accumulation. The small coefficient of variation of the two iron mineral peak values and the similar DRS first derivative curves from the top to bottom of the mottled clay layer in every drill coreindicate that the composition of the mottled clay in different depths is uniform, and the samples had suffered a sufficient mixing and sorting before accumulation. It gives new evidence for the determination of mottledaeolian clay. It can be seen that the DRS method provides not only technical support for iron oxide identification of sediment but also contributes new ideas for the determination of sedimentary environment and origin.

A late Quaternary red sandy sediment called the Old Red Sand is widely distributed in coastal South China. Most studies have considered it a single sand body composed of wind-transported beach sand. However, the Old Red Sand also contains silt and clay. To determine the implications of this fine fraction for the origin of the Old Red Sand, four sections were studied using various depositional analyses. Under a scanning electron microscope, quartz particles in the fine fraction are well rounded, with abundant aeolian marks on their surfaces. The grain size is homogeneous and comparable to that of typical loess. The diffuse reflectance spectroscopy results suggest a higher content of haematite than goethite within the fine fraction, indicating subaerial deposition without strong hydration. The geochemical composition of the fine fraction is close to that of the upper continental crust and comparable to that of typical aeolian deposits, indicating the extensiveness of material sources, with terrestrial dust being fully mixed by wind over a large area. The depositional characteristics, sedimentary environment and provenance of the fine fraction are markedly different from those of the coarse fraction, which is composed of near-source beach sand. Therefore, the Old Red Sand is not a deposit with a single source. Both near-source coarse beach sand and exotic fine dust contributed to the formation of the deposit in the late Pleistocene, especially the last glacial period. The fine fraction is a key factor contributing to the cementation and redness of the sand body.

The ancient fluvial terrace is one of the keys to understanding how a large river develops in long-term landform evolution. However, surface erosion usually dissects the once extensive terrace into discontinuous segments and makes it difficult to explore the evolution of a large river worldwide. Fortunately, an exception is found in the Jinshaan Gorge of the Yellow River (JGYR), which preserves many Late Miocene-Late Quaternary fluvial terraces and thus can help understand the origin and evolution of a large river. Detailed investigation of the high terraces along the JGYR indicates a broad valley that was later superposed by a V-shaped valley due to rapid incision. To constrain the timeframe in which the broad valley persisted, as well as the onset time of the V-shaped valley formation, we select ten typical fluvial terrace profiles, apply terrestrial cosmogenic nuclide (TCN) Al-26/Be-10 burial dating to yield the ages of terrace deposits, and correct the results by considering both inherited and postburial nuclide production during slow red clay and loess deposition. Integrated with previous studies of magnetostratigraphic ages, our chronological results show that the lateral erosion and valley widening of the infant JGYR formed a broad valley during ca. 8-3.7 Ma; subsequent downward incision began ca. 3.7 Ma, accelerated to form a V-shaped valley and incising meanders into the initial broad valley since ca. 2.5 Ma. Moreover, most terrace ages plot in either periodic or short-lived arid conditions during ca. 8-3.7 Ma, while a distinctive downward incision has occurred under strengthened oscillating arid conditions since ca. 2.5 Ma, suggesting that climate aridification favours terrace aggradation, and tectonic uplift mainly dominated incision, which was consistent with the onset of the extension of the Shanxi Graben System ca. 8 Ma and the enhanced uplift of the Ordos Block since ca. 3.7 Ma. (C) 2021 Elsevier B.V. All rights reserved.

The crystalline basement of Cathaysia Block is commonly believed to have formed before the Middle Neoproterozoic; however, the basement rock has seldom been reported in the eastern Cathaysia. Recently, numerous siliceous and femic rock blocks distributed in the diatexitic host granite were observed on the rocky beach of Fuhu Hill in the southwestern Guangdong, China. The results of geological and geochemical investigations of these rock blocks revealed that the protolith of siliceous rock blocks (SRBs) was marine silicalite deposited in a pelagic environment between the ridge‐proximal and continental margin, and the femic rock blocks (FRBs) geochemically corresponded to trachybasalt and monzogabbro, pertaining to OIB/E‐MORB types in the Ce/Y‐Zr/Nb and Zr/Y/Nb systems. 206Pb/238U ages of 995 and 1,002 Ma were obtained from zircon U–Pb dating of the FRBs. The host rocks of the SRBs and FRBs were mainly diatexitic granite that was intruded and sometimes hosted by massive granite. The zircon 206Pb/238U age gained from the diatexitic granite was 991 Ma whereas that from the massive granite was 239 Ma. It may be concluded that SRBs and FRBs were derived from a spilite‐keratophyre formation formed during 1,005–995 Ma and disintegrated around 991 Ma due to crustal anataxis. These SRBs, FRBs, and their diatexitic host granite blocks might have been brought up to the shallow crust by granitic magma in the Indosinian period or due to fault activity in the late stage. Moreover, the southwest part of the eastern Cathaysia possessed a similar basement as the western Cathaysia.

Laser particle size analysis integrated laser, photoelectric and computer technologies has recently become a mainstream method of grain size testing. As particles with different sizes produce scattered light at different angles to the incident laser, the particle size distribution of the sample can be calculated by measuring the intensity of scattered light at different angles. Because of simple operation, rapid test and high precision, it has an important application in sedimentology. A layer of "mottled clay" is widely developed between the Late Pleistocene and Holocene in Quaternary basins in the coastal areas of Guangdong and Fujian provinces. Current research attributes its origin to exposure weathering of Late Pleistocene marine/fluvial deposits during the last glacial maximum. However, our studies find that the mottled clay has no transition in color, structure and composition with its underlying layer and is therefore not formed from weathering. The mottled clay is silty, easily raised by the wind and similar to typical loess. In order to ascertain the character and origin of the mottled clay, three drill cores in the Pearl River delta area were chosen in this study with the method of laser grain size analysis. The results show that grain size composition is characterized by the modal grain size group of coarse silt(10 similar to 50 mu m), followed by the group of clay grain (<5 mu m), both of which are typical particle compositions of an aeolian deposit. All grain size parameters are in accordance with that of an aeolian deposit. Both particle size parameter scatter diagrams and index distribution range of the mottled clay are consistent with typical loess but different from the underlying deposits. The discriminant analysis exhibits an aeolian origin of the mottled clay. The phase analysis also shows that the mottled clay points coincide with the range of typical loess, but has no genetic correlation with its underlying deposits. It is concluded that the mottled clay is not a weathering product of its underlying sediments but an exotic aeolian deposit. This conclusion is of great scientific significance for reconstructing the paleoenvironment of the last glacial period in Fujian and Guangdong coastal areas in the future. The laser particle size method based on optical scattering provides effective scientific evidence for judging the sedimentary environment and origin of the sediments.

普遍认为珠江三角洲存在上、下两个海侵—海退旋回,对下旋回有30～50 ka(MIS3)和约120 ka(MIS5)之争.通过近1万个钻孔资料,提取了20余个典型钻孔岩心数据,运用生物地层学、年代地层学和层序地层学等方法,重新划分了层序及沉积相,并与深海氧同位素阶段进行对比,识别出SB3(基岩深度风化壳)、SB2(花斑黏土层2)和SB1(花斑黏土层1)等层序边界,分别对应于MIS6,4(或4～2),2等3次低海面期,并将三角洲沉积分出下、中、上3层.认为250 ka以来珠江三角洲最大平均下沉速率约为0.34 mm/a.下沉前,为星罗棋布的低矮山地+宽阔台地+"过路河"的亚热带准平原地貌景观,掀斜下沉后,被厚度不等的晚更新世以来松散沉积层覆盖.

How post-burial terrestrial in situ cosmogenic nuclides (TCN) production is build-up in slow burial condition remains uncertain. In this study, we dated high fluvial terrace sediments that overlain by loess with a slow deposition rate that is based on the assumption of steady accumulation rate via Al-26/Be-10 burial dating method. For sample in deep and rapid buried condition prior to slow overlying, post-burial production shows low component (4-6%) in measured concentration. However, post-burial production occupied 39-79% of measured Al-26 and Be-10 concentrations, which suggests approximately 0.2-0.3 Ma underestimation in simple burial dating mode if post-burial production is neglected. Corrected by post-burial production, we yielded burial ages of two top terraces (ca. 1.7-2.5 Ma and 1.5 Ma, respectively) below the Tangxian planation surface along the Jinshan Gorge of the Yellow River, implying that the Jinshan Gorge initially incised at ca. 1.7-2.5 Ma, corresponding well to regional tectonic activities, such as extension in the Fenwei Graben, accelerative uplift of the Tangxian planation surface, vigorous dissection of North China, and significant tectonic deformation in northeastern Tibetan Plateau. The slow burial mode is probably a potential method for plateau paleo-elevation estimation taking advantage of height sensitivity of TCN production rate if overlying loess magnetostratigraphic age is known.

Large evaporite provinces (LEPs) represent prodigious volumes of evaporites widely developed from the Sinian to Neogene. The reasons why they often quickly develop on a large scale with large areas and thicknesses remain enigmatic. Possible causes range from warming from above to heating from below. The fact that the salt deposits in most salt-bearing basins occur mainly in the Sinian-Cambrian, Permian-Triassic, Jurassic-Cretaceous, and Miocene intervals favours a dominantly tectonic origin rather than a solar driving mechanism. Here, we analysed the spatio-temporal distribution of evaporites based on 138 evaporitic basins and found that throughout the Phanerozoiceon, LEPs occurred across the Earth’s surface in most salt-bearing basins, especially in areas with an evolutionary history of strong tectonic activity. The masses of evaporites, rates of evaporite formation, tectonic movements, and large igneous provinces (LIPs) synergistically developed in the Sinian-Cambrian, Permian, Jurassic-Cretaceous, and Miocene intervals, which are considered to be four of the warmest times since the Sinian. We realize that salt accumulation can proceed without solar energy and can generally be linked to geothermal changes in tectonically active zones. When climatic factors are involved, they may be manifestations of the thermal influence of the crust on the surface.

Unobservability of the seismogenic process in a causative fault that makes earthquake (EQ) prediction difficult. Although the relationship between the preseismic electric anomaly (PSEA) and the mainshock indicates that both the PSEA and EQ may originate from same course proceeding in the seismogenic zone, as evidenced by experiments on stressed granite, geological interpretation of those observations, and experiments was limited by the traditional granite formation theory. Based on new information from studies of granite genesis and geotransects, we present a synthetic model, the fracture electric field (FEF), to elucidate the seismogenic process on a causative fault and its logical linkage with the PSEA and EQ. The model is constrained with various data from the 1975 Haicheng EQ and verified by survey data from an FEF monitor station constructed in 2012 in Guangzhou, China. The main conclusions are as follows:(1) An uneven or undulant fault-plane is the prerequisite for stress-accumulation in the locality of the plane to form a seismogenic zone;(2) The position of the continental seismic layer corresponds to that of the crustal granite layer, suggesting that the seismogenic process of any causative fault in continents may produce the PSEA;(3) A normal FEF exists in a causative fault and can be measured beyond the preseismic situation. Thus, it is possible to detect the seismogenic process of a fracture through monitoring the variation of its FEF when the fracture enters a preseismic situation.

Last Glacial Maximum (LGM) mottled clay occurs widely in Late Quaternary basins in south China coastal areas. Current research attributes its origin to exposure weathering of Late Pleistocene marine/fluvial deposits during the LGM. However, field data suggest that this is not the case as there is no gradual transition in lithology, grain size, structure and material composition among these layers. Instead, the mottled clay possesses sedimentary characteristics of exotic dust. In this study, three typical drill cores in the Pearl River Delta were studied using grain size analysis, diffuse reflection spectroscopy (DRS) and geochemical analysis to ascertain the clay's sedimentary characteristics and origin. Grain size distribution patterns and parameters of the mottled clay were similar to those of a typical loess, indicating aeolian origin. In DRS curves, the peak height of hematite > goethite, indicating that the mottled clay had not experienced strong hydration and constitutes a continental product. This conforms to a typical loess but differs from the underlying marine/fluvial deposits. The chemical composition of the mottled clay was homogeneous in the vertical and planar directions. Upper continental crust (UCC) normalized curves of major and trace elements of the mottled clay were close to the average UCC and were consistent with typical aeolian deposits. The spatial and temporal distribution characteristics and relationship with the underlying layer suggest that the mottled clay was a loess-like deposit during the LGM and its mottled structure originated from strong modification of oxidation during the postglacial period after homogeneous dust had accumulated.

A layer of yellow silt with thickness from tens of centimeters to a few meters is widely distributed on low hills and terraces in the coastal area of south China. The yellow silt has different sedimentological characteristics from other common types of Quaternary sediments and the bedrock weathering material in the area. Five representative sections were studied using scanning electron microscopy, grain size analysis, diffuse reflection spectroscopy, geochemical analysis and optically-stimulated luminescence (OSL) dating to reveal the sedimentary characteristics, deposition environment, origin and age of the yellow silt. Subrounded-subangular quartz grains showed typical aeolian impact pits, such as dish-like and crescent pits, and chemical precipitation and corrosion on grain surfaces. The grain size composition, grain size parameters and unimodal frequency distribution curves of the yellow silt were similar in all sections examined and consistent with those of north and central China loess. The iron mineral content of the yellow silt was dominated by hematite, indicating a similar sedimentary environment to that of typical Chinese loess. The major oxide composition of the yellow silt with respect to normalized upper continental crust (UCC) distribution patterns was close to that of the UCC and similar to north China loess and other aeolian deposits in China. However, the data suggested that the yellow silt in the coastal area of south China had suffered strong weathering after deposition. The OSL ages of the yellow silt indicated a deposition age of late Pleistocene. The results of the various analytical methods all showed an origin correlation and regular transition of loess deposits from northwest to southeast China. Hence, it was concluded that the yellow silt widely distributed in the coastal area of south China is a loess-like deposit formed during the late Pleistocene. It probably mostly originated from northwest inland China and constitutes the south extension of other loess in China.

Integration of the published experimental data of partially melted amphibolite, orthogneiss, granite and aplite reveals the existence of three rapid strength drops of the rock-assemblage at melt fraction around 7%, 21% and 41% respectively. The first and the last drops are equivalent to the well-known 'melt connectivity transition' (MCT) and 'solid to liquid transition' (SLT). The drop at melt fraction around 21% is newly identified and termed as the 'framework-melting transition' (FMT). The FMT is defined as the transition from eutectic quartz-feldspars melting to refractory biotite melting. The three rheological transitions are well constrained by the geological data from the Fuhu profile in Guangdong Province, SE China. This profile is a 15 m high cliff. Rocks exposed on the profile are zoned as mottled, stripped, narrow-banded and wide-banded migmatites and diatexitic granite from the top of the profile downwards. The leucosome proportions of the zones measured on the outcrops combined with the experimental data of the partially melted rocks have significant implications for understanding the role of rheological transitions during crustal melting.The MCT, which corresponds to the boundary between the mottled and the narrow-banded migmatites, suggests the segregation of melt within the partially melted system along with the compaction of matrix. Thus, the MCT is deemed to be the transition of the partially melted system from matrix-supported (melt stored in pore-space) to framework-supported (melt stored in foliation/bedding-space of protolith).The FMT corresponding to the boundary between the narrow- and the wide-banded migmatites denotes the inception of melting of the solid framework. Mechanically, the FMT can be regarded as the transition of the framework from compaction to fusion, and geochemically, the transition from eutectic melting of quartz-feldspars to that of refractory biotite. The increasing melt due to framework-melting after the FMT is basically limited within the solid framework until the melt fraction exceeding the SLT.The SLT is geologically constrained by the in-situ metatexite-diatexite transition that is also termed as Magma Interface (MI). The SLT (MI) represents the upper limit of a crustal magma layer. Because of the downward increasing temperature and melt fraction, there should be no consecutive layer of felsic rock existed in the space between the SLT (MI) and the mafic lower crust except the isolated rock-blocks within the magma layer. (C) 2017 Elsevier Ltd. All rights reserved.

Unobservability of the rock-forming process of granite has resulted in a long-term controversy of granite genesis. Numerical techniques combined with super-computer create the possibility of digital reconstruction of the thermodynamic process of granite formation. Here we first look back the advance in study of physical-chemical parameter of rock-melting, introduce the concept of crustal 'average strength' and re-determine the positions of the known rheological transitions, i. e. the MCT, FMT and SLT on the relationship of crustal average strength vs. melt fraction. Secondly, a review of advance in physical and numerical simulations of the various models of magma emplacement is given, and the un-universal characteristics of digital models constructed on different emplacement modes are discussed, which are mainly ascribed to the separation of source and room of magma in the intrusion model. Finally, we introduce the 2-D numerical simulation of large-scale crustal melting in the Tianhe-2 super-computer on the basis of in-situ melting model by Chen and Grapes (2007), preliminarily reconstructing the thermodynamic process of formation of granite and migmatite. The modeling result indicates that thermal convection within a crustal partial melting region is essential for formation of granite magma; Roof-stopping results in the upward motion of the MI (SLT) and thus thickening the convection magma layer; And prerequisite for development of a crustal magma layer is not a very high temperature, but a sustained energy input to maintain the convection state of the magma system.