Local Decoupling Research Articles

Neural-aware Decoupling Fusion based Personalized Federated Learning for Intelligent Sensing

Personalized federated learning (PFL) framework that targets individual models for optimization provides better privacy and flexibility for clients. However, in challenging intelligent sensing applications, the heterogeneous client’s data distributions make the aggregation of local models in the server unstable or even hard to converge. To deal with the performance degradation caused by the above problem, existing PFL methods focus more on how to fine-tune the global model but ignore the impact of the global model fusion algorithm on the results. In this paper, we propose a new explainable neural-aware decoupling fusion based PFL framework, p-FedADF , to address the above challenges. It contains two carefully designed modules. The local decoupling module deployed on the client, utilizes the architecture disentangle technique to decouple the feature extractors in the client’s local model into sub-network according to data categories. It obtains the inference process of feature extraction for different categories of data by training. The global aggregation module, deployed on the server, aligns the sub-networks positions for multiple clients and implements a fine-grained generic feature extractor aggregation. In addition, we provide a mask encoding scheme to reduce the communication overhead of transmitting the sub-network sets between the server and clients. Our p-FedADF obtains 1.6%, 0.2%, 2.3%, and 4.5% improvement on a real-world dataset and three benchmark datasets, compared to state-of-the-art (SOTA) methods.

Paradiplomacy and international conflict: disengagement from China by Swedish local governments

This article analyzes the development of paradiplomatic relations between subnational governments in China and Sweden. The empirical material comprises official documents, media reports, and interviews, and correspondence with local decision makers. Following steady growth in the 1990s and 2000s, the number of paradiplomatic agreements between the two countries declined dramatically at the end of the 2010s. To our knowledge, this extensive “local decoupling” is unprecedented, at least in the history of modern paradiplomacy between Europe and China. The decline can primarily be explained by three interrelated factors: worsening bilateral relations between Sweden and China, growing Swedish concern about the human rights situation in China, and the lack of activity in some of the cooperation projects. Rather than pursuing goals such as peacebuilding and economic benefit, generally described as the main drivers behind local governments’ international activities, several subnational governments in Sweden deemed it appropriate to terminate cooperation with their Chinese counterparts due to other concerns. This development demonstrates the usefulness of a more constructivist perspective of paradiplomacy.

Bright is the moon over my home village: Chinese local executives and environmental decoupling

This study investigates whether and how local executives' hometown identities affect enterprises' environmental decoupling. Environmental decoupling refers to the discrepancy between a company's environmental sustainability promises and its actual environmental sustainability performance. Based on the place attachment theory, we expect local executives to engage in fewer corporate environmental decoupling actions than non-local executives. Using the annual data for Chinese companies listed on the A-share market between 2009 and 2021, we tested this hypothesis. Our findings reveal that local executives' sense of identification and belonging to their hometowns lessens the likelihood of businesses engaging in environmental decoupling. This result is more pronounced for companies located in regions with low population mobility, a slow marketization process, strict environmental regulations, and companies with high financing constraints and low corporate governance levels. Furthermore, we find that local executives decrease underreporting of the environmental performance of firms, thereby reducing firms' environmental decoupling. Following several robustness tests, the primary conclusions remain valid. This study offers new evidence that executive characteristics affect company sustainability from an environmental performance decoupling standpoint. It shows that identifying the executives' hometowns motivates them to uphold a sincere commitment to corporate social responsibility.

Distinct brain network patterns in complete and incomplete spinal cord injury patients based on graph theory analysis.

To compare the changes in brain network topological properties and structure-function coupling in patients with complete spinal cord injury (CSCI) and incomplete spinal cord injury (ICSCI), to unveil the potential neurobiological mechanisms underlying the different effects of CSCI and ICSCI on brain networks and identify objective neurobiological markers to differentiate between CSCI and ICSCI patients. Thirty-five SCI patients (20 CSCI and 15 ICSCI) and 32 healthy controls (HCs) were included in the study. Here, networks were constructed using resting-state functional magnetic resonance imaging to analyze functional connectivity (FC) and diffusion tensor imaging for structural connectivity (SC). Then, graph theory analysis was used to examine SC and FC networks, as well as to estimate SC-FC coupling values. Compared with HCs, CSCI patients showed increased path length (Lp), decreased global efficiency (Eg), and local efficiency (Eloc) in SC. For FC, ICSCI patients exhibited increased small-worldness, clustering coefficient (Cp), normalized clustering coefficient, and Eloc. Also, ICSCI patients showed increased Cp and Eloc than CSCI patients. Additionally, ICSCI patients had reduced SC-FC coupling values compared to HCs. Moreover, in CSCI patients, the SC network's Lp and Eg values were significantly correlated with motor scores, while in ICSCI patients, the FC network's Cp, Eloc, and SC-FC coupling values were related to sensory/motor scores. These results suggest that CSCI patients are characterized by decreased efficiency in the SC network, while ICSCI patients are distinguished by increased local connections and SC-FC decoupling. Moreover, the differences in network metrics between CSCI and ICSCI patients could serve as objective biological markers, providing a basis for diagnosis and treatment strategies.

Hyperspectral image classification based on local feature decoupling and hybrid attention SpectralFormer network

ABSTRACT Classification of hyperspectral image (HSI) is an essential step in various remote sensing tasks. Currently, transformer-based models have made significant contributions to HSI classification due to their remarkable ability to model the relation among long-range sequence data. However, HSI data exhibit strong nonlinear coupling and strong correlation, leading to a decrease in the classification performance of transformer-based models. The SpectralFormer network has achieved excellent performance in HSI classification, but its cross-layer adaptive fusion mechanism cannot effectively highlight critical information, affecting the representation ability of spectral-spatial features. To address the above issues, this paper proposes an HSI classification method based on local feature decoupling and SpectralFormer with a hybrid attention module (HAMSF) network. Specifically, this paper first uses the histogram of oriented gradient algorithm to preprocess HSI coupled data and achieve preliminary nonlinear decoupling of HSI data. Secondly, the hybrid attention module is embedded in a cross-layer adaptive fusion mechanism, generating refined neighbouring information before each information fusion to improve the spectral-spatial feature representation ability. Finally, the processed data are input into the HAMSF network for the final decoupling and classification of HSI. The experiment is carried out on three benchmark hyperspectral datasets, and the experimental results confirm that the proposed method has better classification performance compared to several transformer-based and classical HSI classification methods.

Multi-ensemble metrology by programming local rotations with atom movements

Current optical atomic clocks do not utilize their resources optimally. In particular, an exponential gain in sensitivity could be achieved if multiple atomic ensembles were to be controlled or read out individually, even without entanglement. However, controlling optical transitions locally remains an outstanding challenge for neutral-atom-based clocks and quantum computing platforms. Here we show arbitrary, single-site addressing for an optical transition via sub-wavelength controlled moves of atoms trapped in tweezers. The scheme is highly robust as it relies only on the relative position changes of tweezers and requires no additional addressing beams. Using this technique, we implement single-shot, dual-quadrature readout of Ramsey interferometry using two atomic ensembles simultaneously, and show an enhancement of the usable interrogation time at a given phase-slip error probability. Finally, we program a sequence that performs local dynamical decoupling during Ramsey evolution to evolve three ensembles with variable phase sensitivities, a key ingredient of optimal clock interrogation. Our results demonstrate the potential of fully programmable quantum optical clocks even without entanglement and could be combined with metrologically useful entangled states in the future.

Characteristics of cellular structure of detonation waves propagating in annular channels

This study investigates the characteristics of stable and unstable cells and wavefronts of detonation waves propagating in annular channels with different inner radii and channel widths using two-dimensional Euler equations along with a two-step induction-exothermic reaction kinetics. The results reveal that the effect of annular channels on the detonation cell structure depends on both the inner radius and channel width. To quantify this effect, a parameter σ is introduced, representing the ratio of the inner and outer radii of the channel. We have discovered that for values of the parameter σ exceeding a critical value σs, the detonation wavefront demonstrates characteristics similar to those observed in a straight channel scenario. On the contrary, when σ is below σs, the wavefront becomes distorted, potentially leading to Mach reflection as σ decreases further to another critical value σm. Additionally, the interaction among expansion waves induced by the inner walls leads to an augmented induced length and the potential occurrence of localized decoupling of the detonation wave, particularly for unstable detonation waves. However, it is worth noting that the re-initiation of the detonation wave may be triggered by the formation of hotspots resulting from the interaction between transverse shock waves and the detonation wave. This study aims to characterize the propagation characteristics of detonation waves within annular channels, with the objective of providing valuable insights for the design and optimization of annular chamber configurations in systems involving detonation.

Local structural-functional connectivity decoupling of caudate nucleus in infantile esotropia.

Abnormal brain structural and functional properties were demonstrated in patients with infantile esotropia (IE). However, few studies have investigated the interaction between structural and functional connectivity (SC-FC) in patients with IE. Structural network was generated with diffusion tensor imaging and functional network was constructed with resting-state functional magnetic resonance imaging for 18 patients with IE as well as 20 age- and gender- matched healthy subjects. The SC-FC coupling for global connectome, short connectome and long connectome were examined in IE patients and compared with those of healthy subjects. A linear mixed effects model was employed to examine the group-age interaction in terms of the coupling metrics. The Pearson correlation between coupling measures and strabismus degree was evaluated in IE patients, on which the regulatory effect of age was also investigated through hierarchical regression analysis. Significantly decreased SC-FC coupling score for short connections was observed in left caudate nucleus (CAU) in IE patients, whereas no brain regions exhibited altered coupling metrics for global connections or long connections. The group-age interaction was also evident in local coupling metrics of left CAU. The age-related regulatory effect on coupling-degree association was distinguishing between brain regions implicated in visual processing and cognition-related brain areas in IE patients. Local SC-FC decoupling in CAU was evident in patients with IE and was initiated in their early postnatal period, possibly interfering the visual cortico-striatal loop and subcortical optokinetic pathway subserving visual processing and nasalward optokinesis during neurodevelopment, which provides new insight into underlying neuropathological mechanism of IE.

Eulerian–Lagrangian modeling of deflagration to detonation transition in n-decane/oxygen/nitrogen mixtures

In this work, to promote deflagration to detonation transition (DDT), a designed hot jet in a pre-detonator is produced to initiate detonations in the main detonation tube. We perform two-dimensional simulations of the DDT process for low-volatile fuel (n-decane) mixed with nitrogen and oxygen based on the Eulerian–Lagrangian approach. The effects of fuel atomization, vaporization, and shock focusing on the flame acceleration and DDT are discussed under different nitrogen dilution ratio and droplet size conditions. The results show that the flame acceleration process can be divided into slow and fast deflagration stages. Additionally, initiation times are mainly determined by the fuel atomization and evaporation in the slow deflagration stage, which dominates the entire DDT time. Furthermore, there are different intensities of hot jets rather than stable detonation waves formed at the pre-detonator exit. Moreover, local decoupling and re-initiation events are detected near the internal wall of the U-bend, inducing the overdriven detonation decaying into stable detonation waves in the smooth tube. The results also demonstrate that the detonation pressure and velocity decrease by 13.56% and 12.55%, respectively, as the nitrogen dilution ratio increases from 0.5 to 2. In particular, as the nitrogen dilution ratio continued to increase to 2.25, the development in DDT is similar, but the jet intensity is significantly weakened. While as the droplet size increases from 10 to 40 μm, the detonation pressure and velocity decrease only by 2.69% and 1.49%, respectively.

Evaluation of the local decoupling of livestock and cropland in the Huang-Huai-Hai region

Decoupling livestock and cropland production at regional scale have poor resource-use efficiency and detrimental effects on environment in China. It is therefore necessary to identify the decoupled livestock and cropland production system and make recommendations to recouple livestock and cropland. This study used the indexes of land carrying capacity (LCC), animal manure absorption capacity (AMAC), and risk warning value (R) to evaluate the coupling between cropland and livestock at the local scale in the Huang-Huai-Hai region. The decoupling of cropland and livestock in the case of Beijing (SY_BJ) was found assessed with lower theoretical value of LCC and higher theoretical value of AMAC compared with local actual situation, categorized as grade IV with a high R value (above 1). Contrary results were found that the livestock and cropland production systems were coupled at the local scale in the cases located in Hebei and Shandong Provinces, categorized as grade I or II. Two measures were used to optimize the decoupled case by adjusting the ratio of manure to fertilization or reducing breeding quantity. The decoupled case of SY_BJ could be optimized by adjusting the ratio of manure to fertilization (95.34% based on nitrogen and 81.97% based on phosphorus, respectively). The breeding quantity in this case should be reduced by at least 46% to recouple the livestock and cropland at the local level to manage nutrient surpluses from livestock and poultry breeding.

Non-GDANets: Sports small object detection of thermal images with Non-Glodal decoupled Attention.

Because thermal infrared sport targets have rich and complex semantic information, there is a high coupling between different types of features. In view of these limitations, we propose a Non-Glodal decoupled Attention, namely,local U-shaped attention decoupling network (LUANets), which aims to decompose the coupling relationship of different sport target features in thermal infrared images and establish effective spatial dependence between them. This method takes the captured multi-scale initial features according to different levels and inputs them into the local decoupling module with U-shaped attention structure to realize the decomposition of semantic details. At the same time, considering the correlation between different targets, in the process of feature decomposition, using prior knowledge as guiding information many times to establish effective spatial dependence. Secondly, we design a two-way cross-aggregation FPN module to cross-aggregate information flows in the front and back directions to achieve feature interaction while further reducing the coupling between different types of features. The evaluation results on data such as TIIs,SportFCs and FLIR show that the LUANets method we proposed has achieved the best detection performance, with mAP of 68.72%,59.51% and 65.29%, respectively.

Holocene Fires and Ecological Novelty in the High Colombian Cordillera Oriental

Rapid climate changes and the increasing presence of humans define the Holocene Epoch (11.6 calibrated kiloyears before present – hereafter kyr BP), when biological systems have faced the most recent and abrupt environmental changes. Understanding how biodiversity responds to extrinsic factors requires determining the effects of varying climatic conditions, changes in disturbance regimes, and increasing anthropogenic impacts. Despite being one center for biodiversity, the potential synergies of long-term anthropogenic and climate changes in shaping areas of high Andean biodiversity have yet to be explored fully. Here we present new pollen and charcoal records from the Pantano de Monquentiva (hereafter Monquentiva) on the highlands of the eastern flank of the Colombian Cordillera Oriental (CCO) to document relationships between climate, vegetation, and fire through the Holocene. We found compositional transitions at 8.7, 6.1, and 4.1 kyr BP at Monquentiva resulting from the interaction of climate, fire, and human occupation. Reduced moisture and temperature caused a compositional shift in Páramo vegetation from ca. 8.7 kyr BP. Fire activity was recorded throughout the Holocene and increased slightly during the Mid-Holocene when regional and local fire decoupling suggested human activities as the source of ignition. Mid-Holocene fires had a large effect on the vegetation composition at Monquentiva which recorded a rapid shift at ca. 6.8 kyr BP. Fire activity increased sharply from 4.1 kyr BP, promoting the reorganization of plant communities at 3.8 kyr BP. This shift in fire activity was likely related to more severe ENSO events and subsequently intensified by human activities after 3.8 kyr BP. Although high climatic sensitivity explains most Holocene vegetation changes in the eastern flank of the CCO, our study highlights the relevance of fire activity, uneven distribution of climatic variables, and human intervention to the composition of the vegetation we see today.

A Refinement of the Processes Controlling Dissolved Copper and Nickel Biogeochemistry: Insights From the Pan‐Arctic

AbstractRecent studies, including many from the GEOTRACES program, have expanded our knowledge of trace metals in the Arctic Ocean, an isolated ocean dominated by continental shelf and riverine inputs. Here, we report a unique, pan‐Arctic linear relationship between dissolved copper (Cu) and nickel (Ni) present north of 60°N that is absent in other oceans. The correlation is driven primarily by high Cu and Ni concentrations in the low salinity, river‐influenced surface Arctic and low, homogeneous concentrations in Arctic deep waters, opposing their typical global distributions. Rivers are a major source of both metals, which is most evident within the central Arctic's Transpolar Drift. Local decoupling of the linear Cu‐Ni relationship along the Chukchi Shelf and within the Canada Basin upper halocline reveals that Ni is additionally modified by biological cycling and shelf sediment processes, while Cu is mostly sourced from riverine inputs and influenced by mixing. This observation highlights differences in their chemistries: Cu is more prone to complexation with organic ligands, stabilizing its riverine source fluxes into the Arctic, while Ni is more labile and is dominated by biological processes. Within the Canadian Arctic Archipelago, an important source of Arctic water to the Atlantic Ocean, contributions of Cu and Ni from meteoric waters and the halocline are attenuated during transit to the Atlantic. Additionally, Cu and Ni in deep waters diminish with age due to isolation from surface sources, with higher concentrations in the younger Eastern Arctic basins and lower concentrations in the older Western Arctic basins.

Solvent Reduced Extrusion‐Based Anode Production Process Integrating Granulate Coating, Drying, and Calendering

A novel electrode production process, based on a minimal solvent content, is developed for highly viscous water‐based graphite anodes, using a solid mass content of 50% or more compared to conventionally processed anodes. The electrode paste is prepared by mixing the solids with solvent inside of a twin‐screw extruder. Afterward, the extrudate is cut with a strand pelletizer to granules. By inserting the granules into the gap of a two‐roller calender, the current collector foil is coated directly. It is shown that the highly viscous pastes are capable of extended storage stability over several weeks, whereby a temporal and local decoupling of paste preparation and electrode manufacturing can be realized. To enable this innovative electrode processing, a combined development incorporating a selection of new anode binder materials, machinery design, and process parameters has been essential due to strong material‐process interaction, as known in the field of battery electrode production. Because the binder is exposed to greater shear stresses and a short duration time during extrusion mixing, the selection of novel anode binder materials is particularly important. This then needs to be combined with machine design and process parameter optimization, in order to establish this innovative electrode processing method.

Visual experimental research on the propagation instabilities in a plane-radial rotating detonation engine

Using visualization method, assisted by synchronous pressure and ion measurements, the characteristics of four instabilities and operating ranges are investigated in a plane-radial rotating detonation engine. At fuel-lean limits, the rotating detonation wave (RDW) propagates in an incoherent mode. The propagation is unstable with obvious interruption characterized by local flameout, decoupling, re-initiation and propagation direction change. At low mass flow rates, the quasi detonation mode with fairly low pressure peak but high velocity is obtained, in which the outer circumferential flame and central flame occur alternately. The modes switching of RDW operates at critical operating conditions. Two modes are obtained in one test and sudden change in the wave-head number exists in the combustor. An asymmetric dual-wave mode is obtained at special conditions with low mass flow rates and fuel-rich conditions, and the intensities of two asymmetric RDWs alternately increase and decrease with time. The phenomena of decoupling and re-initiation occur periodically, and an amplitude-frequency low-frequency oscillation is observed in this instability. In these instabilities, the pre-combustion phenomenon ahead of detonation wave occurs on the deflagration surface, which promotes the development of RDW. Furthermore, the pressure rising rates is greatly affected by combustion modes and wave-head number. The pressure rising rates of quasi detonation mode is an order of magnitude lower than that of stable detonation mode.

A Compact Integrated High-Voltage Pulser Insensitive to Supply Transients for 3-D Miniature Ultrasound Probes

In this paper, a compact high-voltage (HV) transmit circuit for dense 2D transducer arrays used in 3D ultrasonic imaging systems is presented. Stringent area requirements are addressed by a unipolar pulser with embedded transmit/receive switch. Combined with a capacitive HV level shifter, it forms the ultrasonic HV transmit circuit with the lowest reported HV transistor count and area without any static power consumption. The balanced latched-based level shifter implementation makes the design insensitive to transients on the HV supply caused by pulsing, facilitating application in probes with limited local supply decoupling, such as imaging catheters. Favorable scaling through resource sharing benefits massively arrayed architectures while preserving full individual functionality. A prototype of 8 x 9 elements was fabricated in TSMC 0.18 μm HV BCD technology and a 160 μm x 160 μm PZT transducer matrix is manufactured on the chip. The system is designed to drive 65 V peak-to-peak pulses on 2 pF transducer capacitance and hardware sharing of 6 elements allows for an area of only 0.008 mm2 per element. Electrical characterization as well as acoustic results obtained with the 6 MHz central frequency transducer are demonstrated.

Secular crustal deformation characteristics prior to the 2011 Tohoku-Oki earthquake detected from GNSS array, 2003–2011

In order to reveal the secular deformation evolution and strain accumulation progress along subduction zone, we focus on high-precision Global Navigation Satellite System (GNSS) data processing, GNSS time series analysis and regional spatiotemporal filtering. The GNSS position time series is modeled as a combination of the long-term plate movements, short-term noise and the frequency-dependent variations and tectonic deformation information. Common Mode Errors (CME) is removed and the deformation signals is then extracted from the residual time series by Principal Component Analysis (PCA). Combing the filtered displacement array and the dynamic regional strain array derived from GNSS baseline, a spatiotemporal tectonic evolution map is exhibited obviously. We process the data recorded by GEONET over the 2003–2011 period in middle Japan island and reveal the dynamic evolution of crustal deformation along subduction zone. The results show that the regional reference frame transformation by constraining one site location and one stable GNSS baseline direction can further improve the Signal-to-Noise Ratio (SNR) of GNSS observations. A integrated analysis combining the baseline length with azimuth change can show the crustal motion feature more comprehensively. The observed behaviors agree well with the simulation experiment results of the rock rupture in the laboratory. We divide the pre-seismic deformation into four stages, the stable linear strain accumulation stage, the formation stage of the local locked area, local decoupling stage, and strain release stage. We also give an appropriate dynamical model for the interplate coupling to explain the observed deformation characteristics. Research result provides some of the observational new evidence for inter-seismic deformation anomaly detection and the medium- to longer-term seismic hazard assessments.

The Mixed-Electrode Concept for Understanding Growth and Aggregation Behavior of Metal Nanoparticles in Colloidal Solution

The growth and aggregation behavior of metal nanoparticles can be modulated by surfactants and different other additives. Here the concept of how open-circuit mixed electrodes helps to understand the electrical aspects of nanoparticle growth and the consequences for the particle geometries is discussed. A key issue is the self-polarization effect of non-spherical metal nanoparticles, which causes a local decoupling of anodic and partial processes and asymmetry in the local rates of metal deposition. These asymmetries can contribute to deciding to the growth of particles with high aspect ratios. The interpretation of electrochemical reasons for particle growth and behavior is supported by experimental results of nanoparticle syntheses supported by microfluidics which can supply high yields of non-spherical nanoparticles and colloidal product solutions of high homogeneity.

Marine Forearc Extension in the Hikurangi Margin: New Insights From High‐Resolution 3‐D Seismic Data

AbstractUpper‐plate normal faults are a widespread structural element in erosive plate margins. Increasing coverage of marine geophysical data has proven that similar features also exist in accretionary margins where horizontal compression usually results in folding and thrust faulting. There is a general lack of understanding of the role and importance of normal faulting for the structural and tectonic evolution of accretionary margins. Here we use high‐resolution 2‐D and 3‐D seismic reflection data and derived seismic attributes to map and analyze upper‐plate normal faulting in the marine forearc of the accretionary Hikurangi margin, New Zealand. We document extension of the marine forearc over a wide area along the upper continental slope. The seismically imaged normal faults show low vertical displacements, high dip angles, a preference for landward dip, and often en echelon patterns. We evaluate different processes, which may cause the observed extension, including (1) stress change during the earthquake cycle, (2) regional or local uplift and decoupling of shallow strata from compression at depth, and (3) rotation of crustal blocks and resulting differential stresses at the block boundaries. The results suggest that normal faults play an important role in the structural and tectonic evolution of accretionary margins, including the northern Hikurangi forearc.

Formation and evolution of multistage magmatic-hydrothermal fluids at the Yulong porphyry Cu-Mo deposit, eastern Tibet: Insights from LA-ICP-MS analysis of fluid inclusions

The giant Eocene Yulong porphyry Cu-Mo deposit in eastern Tibet, China was produced by multistage veining, Cu-Mo sulfide deposition, and hydrothermal alteration associated with pulsed magmatic intrusions (i.e., early, transitional, and late stages) lasting for a prolonged period of time. In this study, fluid inclusion microthermometry and laser ablation-ICP-MS microanalysis are combined with textural observations on magmatic apatite and various hydrothermal quartz vein assemblages to constrain the formation and evolution of the multistage magmatic-hydrothermal fluids producing the Yulong porphyry Cu-Mo deposit.Fluid inclusions hosted in quartz veins of the early and transitional stages have very similar compositions, indicating that the initial single-phase intermediate-density (ID) ore-forming fluids of these two mineralizing stages were exsolved from similarly evolved magma reservoirs in the underlying magma chamber. During their ascent, decompression and cooling, the single-phase ID input fluids (∼9 wt% NaClequiv. with ∼1000 ppm Cu and ∼20 ppm Mo) entered the two-phase field and condensed into a small amount of metal-rich brines (∼42 wt% NaClequiv. with ∼9300 ppm Cu and ∼330 ppm Mo) coexisting with a large amount of vapors (vapor/brine mass ratio of ∼4). The condensation of brines and their accumulation at shallow level could be an efficient mechanism to concentrate and then precipitate Cu ± Mo in a small rock volume as represented by the mineralized porphyries. The sequential deposition of Mo and Cu from the condensed brine phase of the early stage caused the local Cu-Mo decoupling in shallow parts of the deposit, whereas the deeply located Mo-rich mineralization of the transitional stage was probably caused by the early precipitation of molybdenite from the single-phase ID fluids before phase separation. It is concluded that the formation of a late Mo-rich mineralization (similar to the transitional-stage mineralization at Yulong) in many other porphyry Cu deposits is due to the subsolidus hydrothermal processes rather than the progressive increase of the Mo/Cu ratio in residual parental melts as magma crystallization proceeds.The high Cs + Rb ± B concentrations of melt and fluid inclusions in sulfur-rich magmatic apatite (0.3–1.4 wt% SO3) suggest a pegmatitic environment for the exsolution of the fluids. The similar element/K ratios of liquid-rich inclusions in the late-stage pyrite-quartz veins to the fluid inclusions in the sulfur-rich apatite suggest that the late-stage fluids were also likely derived from highly evolved melt fractions in felsic magmas during the solidification of the magma chamber.