Search

Abstract In this work, we establish a direct connection between generative diffusion models (DMs) and stochastic quantization (SQ). The DM is realized by approximating the reversal of a stochastic process dictated by the Langevin equation, generating samples from a prior distribution to effectively mimic the target distribution. Using numerical simulations, we demonstrate that the DM can serve as a global sampler for generating quantum lattice field configurations in two-dimensional ϕ 4 theory. We demonstrate that DMs can notably reduce autocorrelation times in the Markov chain, especially in the critical region where standard Markov Chain Monte-Carlo (MCMC) algorithms experience critical slowing down. The findings can potentially inspire further advancements in lattice field theory simulations, in particular in cases where it is expensive to generate large ensembles.

Dust is a major source of atmospheric aerosols. Its chemical composition is often assumed to be similar to the parent soil. However, this assumption has not been rigorously verified. Here, we generated dust aerosols from soils to determine if there is particle-size-dependent selectivity of heavy metals in the dust generation. Mn, Cd, Pb and other heavy metals were found to be highly enriched in fine-dust (PM2.5) aerosols, which can be up to ∼ 6.5-fold. To calculate the contributions of dust to atmospheric heavy metals, regional air quality models usually use the dust chemical profiles from the U.S. Environmental Protection Agency's (EPA) SPECIATE database, which does not capture the correct size-dependent selectivity of heavy metals in dust aerosols. Our air quality modeling for China demonstrates that the calculated contribution of fine-dust aerosols to atmospheric heavy metals, as well as their cancer risks, could have significant errors without using proper dust profiles.

Due to the rapid development of industrialization and a substantial economy, China has become one of the global hotspots of nitrogen (N) and sulfur (S) deposition following Europe and the USA. Here, we developed a dataset with full coverage of N and S deposition from 2005 to 2020, with multiple statistical models that combine ground-level observations, chemistry transport simulations, satellite-derived vertical columns, and meteorological and geographic variables. Based on the newly developed random forest method, the multi-year averages of dry deposition of oxidized nitrogen (OXN), reduced nitrogen (RDN), and S in China were estimated at 10.4, 14.4, and 16.7 kg N/S ha−1 yr−1, and the analogous numbers for total deposition were respectively 15.2, 20.2, and 25.9 kg N/S ha−1 yr−1 when wet deposition estimated previously with a generalized additive model (GAM) was included. The dry to wet deposition ratio (Rdry/wet) of N stabilized in earlier years and then gradually increased especially for RDN, while that of S declined for over 10 years and then slightly increased. The RDN to OXN deposition ratio (RRDN/OXN) was estimated to be larger than 1 for the whole research period and clearly larger than that of the USA and Europe, with a continuous decline from 2005 to 2011 and a more prominent rebound afterwards. Compared with the USA and Europe, a more prominent lagging response of OXN and S deposition to precursor emission abatement was found in China. The OXN dry deposition presented a descending gradient from east to west, while the S dry deposition a descending gradient from north to south. After 2012, the OXN and S deposition in eastern China declined faster than the west, attributable to stricter emission controls. Positive correlation was found between regional deposition and emissions, while smaller deposition to emission ratios (D/E) existed in developed eastern China, attributed to more intensive human activities and thereby anthropogenic emissions.

We conducted samplings of total and particulate 234Th, along with particulate organic carbon (POC), in the summer of 2017 to examine nutrient-dependent structures of export productivity within the euphotic zone (Ez) of the oligotrophic basin of the South China Sea (SCS). Nitrate concentrations throughout the study area were below detection limits in the nutrient-depleted layer (NDL) above the nutricline, while they sharply increased with depth in the nutrient-replete layer (NRL) across the nutricline until the base of the Ez. Based on our vertical profilings of 234Th–238U disequilibria, this study estimated for the first time POC export fluxes both out of the NDL and at the horizon of the Ez base. The total 234Th deficit relative to 238U was determined in the NDL at all study sites. By contrast, 234Th was mostly in equilibrium with 238U in the NRL, except at the northernmost station, SEATS (SouthEast Asian Time-series Study; 18∘ N, 116∘ E), where the 234Th deficit was also observed in the NRL. By combining 1D steady-state 234Th fluxes and POC/234Th ratios, we derived vertical patterns of POC export fluxes. The POC export fluxes at station SEATS were 1.6 ± 0.6 mmol C m−2 d−1 at the NDL base, accounting for approximately half of that at the base of the Ez. For the rest of the sampling sites, the POC export fluxes at the NDL base (averaged at 2.3 ± 1.1 mmol C m−2 d−1) were comparable with those at the base of the Ez (1.9 ± 0.5 mmol C m−2 d−1), suggesting rapid export of POC out of the NDL. This finding fundamentally changes our traditional view that the NDL, being depleted in nutrients, would not be a net exporter of POC. Furthermore, our results revealed a significant positive correlation between POC export fluxes at the NDL base and the potential of subsurface nutrient supplies, indicated by nutricline depth and nutrient concentrations obtained from both in situ measurements and numerical modeling. POC export fluxes (averaged at 3.4 ± 1.2 mmol C m−2 d−1) at the NDL base at stations with shallow nutriclines and high levels of subsurface nutrients approximately doubled those (averaged at 1.6 ± 0.5 mmol C m−2 d−1) at other stations. We subsequently used a two-endmember mixing model based on the mass and 15N isotopic balances to evaluate the relative contribution of different sources of new nitrogen that support the observed particle export at stations SEATS and SS1, located respectively in the northern and southern basin of the SCS with different hydrological features. We showed that more than 50 % of the particle flux out of the NDL was supported by nitrate sources likely supplied from depth and associated with episodic intrusions other than atmospheric deposition and nitrogen fixation. However, the exact mechanisms and pathways for subsurface nutrients to support the export production from the NDL merit additional careful and dedicated studies.

We designed a low-cost expandable current profiler (XCP) including software and hardware. An XCP is an observation instrument that rapidly measures currents based on the principle that currents cut the geomagnetic field to induce electric fields. The cost of an XCP must be reduced because it is a single-use device. The digitization of the previously developed XCP is carried out underwater, which requires the probe to contain not only analogue circuits for acquiring signals but also digital circuits and digital chips, which are relatively expensive. In this study, an XCP was developed that adopts signal modulation and demodulation to transmit analogue signals on an enamelled wire, and the signal digitization occurs above the surface of the water. The cost of the instrument was effectively reduced by half while maintaining its ability to measure parameters such as sea current and temperature in real time. After comparison with data processed from laboratory tests, the acquisition circuit showed an accuracy within 0.1 % and the XCP analogue circuit developed for the overall system was stable and reliable. The system exhibited an acquisition accuracy higher than 50 nV for 16 Hz, and the quality of the acquired signal satisfied the requirements for an XCP instrument.

Targeted nanoparticle delivery to sites of interest is important for targeted therapeutics. Here, the authors improve the targeting efficiency of antibodies on nanoparticles using a monobody adapter to correctly orientate the antibody to preserve targeting function and demonstrate application in targeting the vascular endothelium of human kidneys.

Achieving high thermoelectric performance in thin film heterostructures is essential for integrated and miniatured thermoelectric device applications. In this work, we demonstrate a mechanism and device performance of enhanced thermoelectric performance induced by interfacial effect in a transition metal dichalcogenides-SrTiO3 (STO) heterostructure. Owing to the formed conductive interface and elevated conductivity, the ZrTe2/STO heterostructure presents large thermoelectric power factor of 3.7 × 105 μWcm−1K−2 at 10 K. Formation of quasi-two-dimensional conductance at the interface is attributed for the large Seebeck coefficient and high electrical conductivity, leading to high thermoelectric performance which is demonstrated by a prototype device attaining 3 K cooling with 100 mA current input to this heterostructure. This superior thermoelectric property makes this heterostructure a promising candidate for future thermoelectric device.

The high-density electrical method is a primary method used in shallow geophysical prospecting. Due to the rapid industrial development that has taken place in recent years, the function and performance of high-density electrical instruments have been considerably improved in several aspects. However, most of the electrical instruments currently available on the market still exhibit some shortcomings, such as being bulky, heavy, limited in their data acquisition accuracy, and difficult to connect to the Internet for remote monitoring. To address these problems, this study developed a new multifunctional four-dimensional (4D) high-density electrical instrument based on remote wireless communication technology. The system is small and lightweight, includes an integrated transceiver, has high data acquisition accuracy, and is capable of remote wireless real-time control. In this study, the hardware circuit was designed. The Arm all-in-one (AIO) LJD-eWinV5-ST7 with a 154.4 cm × 87 cm, 800 × 480 high-brightness wide-temperature-range display is used as the host computer, which has the advantages of small size, low power consumption, and abundant hardware resources. Internet of things (IoT) technology is incorporated in the system, and a 4G module is employed to provide a real-time remote control and data acquisition monitoring system based on the cloud platform. Tests showed that this instrument is stable and convenient to use and can meet the requirements for use in field prospecting.

The desire to increase the energy density of stoichiometric layered LiTMO_{2} (TM = 3d transition metal) cathode materials has promoted investigation into their properties at high states of charge. Although there is increasing evidence for pronounced oxygen participation in the charge compensation mechanism, questions remain whether this is true O-redox, as observed in Li-excess cathodes. Through a high-resolution O K-edge resonant inelastic x-ray spectroscopy (RIXS) study of the Mn-free Ni-rich layered oxide LiNi_{0.98}W_{0.02}O_{2}, we demonstrate that the same oxidized oxygen environment exists in both Li-excess and non-Li-excess systems. The observation of identical RIXS loss features in both classes of compounds is remarkable given the differences in their crystallographic structure and delithiation pathways. This lack of a specific structural motif reveals the importance of electron correlation in the charge compensation mechanism for these systems and indicates how a better description of charge compensation in layered oxides is required to understand anionic redox for energy storage.

With the development of economy, land reclamation by dredger fill has become an effective measure to alleviate the shortage of land resources. However, the accompanying subsidence has always been a challenge to the safe use of soil in dredger fill area. In this study, Chongming East Shoal, China, where dredger filling activities are going on in recent years was selected as the study area. SBAS-InSAR was applied to monitor the variation of land subsidence and deformation in the recent two years. Furthermore, a total of 25 undisturbed soil samples including dredger fill and underlying soil were collected from 5 boreholes (maximum depth 55 m), and the land at each borehole had different a formation time. The physical properties and compressibility were tested by laboratory tests. Results show that for the current state, fast to slow subsidence velocity was observed in the reclamation area close to the coastline, which is controlled by building load and geological features of soil layers. The building load is the main factor affecting the land subsidence and special attention should be paid. It is the poor drainage condition of the soil layer in the offshore area resulting in slow subsidence. Consolidation degree and final settlement of soil can be obtained from monitoring data of land subsidence. Based on the settlement-time curve obtained by SBAS-InSAR, the estimated final settlement of typical settlement area is −27.03 to −38.96 mm, and the corresponding consolidation degree is 58.95 % on average. It still takes a long time to achieve stability. In conclusion, land subsidence is essentially the macro-accumulation of drainage consolidation of all the soil layers, so it is controlled by soil structure and engineering geological properties of both dredger fill and underlying soil layer. The research combined with field investigation, laboratory testing can provide a mechanism explanation for monitoring results. Future research will focus on longer monitoring time and a higher sampling frequency to enrich and improve the research.

The failure of adjustable speed drive (ASD), which is sensitive to transient low voltage disturbance such as voltage sag and short duration interruption, will bring considerable loss to industrial users. Evaluating the disturbance tolerance ability of adjustable speed drives is the key to improve its tolerance. The voltage tolerance curve (VTC) is a tool to quantify the tolerance of voltage sag sensitive equipment including adjustable speed drives. However, the traditional test strategy for obtaining this curve is time-consuming and may introduce errors caused by the performance deterioration of the tested equipment. This paper proposes an analytical model of adjustable speed drive based on its response mechanism analysis to voltage sags. The model constructs a group of practical formulas for calculating the knee point of the voltage tolerance curve. The knee point can be used to quickly estimate the contour of the curve. This model can help engineers to conduct a first-cut assessment of the sag tolerance for drives to select equipment or design planning. Experimental tests are carried out to verify the effectiveness of the proposed method. Moreover, an application of the results is presented to improve the traditional test strategy.

Up-to-date 3D building models are important for many applications. Airborne very high resolution (VHR) images often acquired annually give an opportunity to create an up-to-date 3D model. Building segmentation is often the first and utmost step. Convolutional neural networks (CNNs) draw lots of attention in interpreting VHR images as they can learn very effective features for very complex scenes. This paper employs Mask R-CNN to address two problems in building segmentation: detecting different scales of building and segmenting buildings to have accurately segmented edges. Mask R-CNN starts from feature pyramid network (FPN) to create different scales of semantically rich features. FPN is integrated with region proposal network (RPN) to generate objects with various scales with the corresponding optimal scale of features. The features with high and low levels of information are further used for better object classification of small objects and for mask prediction of edges. The method is tested on ISPRS benchmark dataset by comparing results with the fully convolutional networks (FCN), which merge high and low level features by a skip-layer to create a single feature for semantic segmentation. The results show that Mask R-CNN outperforms FCN with around 15% in detecting objects, especially in detecting small objects. Moreover, Mask R-CNN has much better results in edge region than FCN. The results also show that choosing the range of anchor scales in Mask R-CNN is a critical factor in segmenting different scale of objects. This paper provides an insight into how a good anchor scale for different dataset should be chosen.

The coming Big Data Era requires progress in storage and computing technologies. As an emerging memory technology, Resistive RAM (RRAM) has shown its potential in the next generation high-density storage and neuromorphic computing applications, which extremely demand low switching voltage and power consumption. In this work, a 10 nm-thick amorphous GeS2 thin film was utilized as the functional layer of RRAM in a combination with Ag and Pt electrodes. The structure and memory performance of the GeS2-based RRAM device was characterized — it presents high on/off ratio, fast switching time, ultralow switching voltage (0.15 V) and power consumption (1.0 pJ and 0.56 pJ for PROGRAM and ERASE operations, respectively). We attribute these competitive memory characteristics to Ag doping phenomena and subsequent formation of Ag nano-islands in the functional layer that occurs due to diffusion of Ag from electrode into the GeS2 thin film. These properties enable applications of GeS2 for low energy RRAM device.

Semantic segmentation, especially for buildings, from the very high resolution (VHR) airborne images is an important task in urban mapping applications. Nowadays, the deep learning has significantly improved and applied in computer vision applications. Fully Convolutional Networks (FCN) is one of the tops voted method due to their good performance and high computational efficiency. However, the state-of-art results of deep nets depend on the training on large-scale benchmark datasets. Unfortunately, the benchmarks of VHR images are limited and have less generalization capability to another area of interest. As existing high precision base maps are easily available and objects are not changed dramatically in an urban area, the map information can be used to label images for training samples. Apart from object changes between maps and images due to time differences, the maps often cannot perfectly match with images. In this study, the main mislabeling sources are considered and addressed by utilizing stereo images, such as relief displacement, different representation between the base map and the image, and occlusion areas in the image. These free training samples are then fed to a pre-trained FCN. To find the better result, we applied fine-tuning with different learning rates and freezing different layers. We further improved the results by introducing atrous convolution. By using free training samples, we achieve a promising building classification with 85.6 % overall accuracy and 83.77 % F1 score, while the result from ISPRS benchmark by using manual labels has 92.02 % overall accuracy and 84.06 % F1 score, due to the building complexities in our study area.

Representation of scenes on the Earth surface by using voxels is gaining attention because of its suitability for integrating heterogeneous data sources in simulations and quantitative models. Computation of shadows in such models is needed, for example, to obtain crop suitability of agricultural fields in the presence of trees and buildings, or to analyze urban heat island causes and effects. We present an efficient algorithm to compute which of the voxels in a dataset receive direct sunlight, given the solar azimuth and elevation angles. The algorithm can work with multiple (sparse and dense) voxel storage strategies.

Various kinds of urban applications require true orthophotos. True orthophoto generation requires a DSM (Digital Surface Model) to project the photo orthogonally and minimize geometric distortion due to topographic variance. DSMs are often generated from airborne laser scan data. In urban scenes, DSM data may fail to deliver sharp and straight building roof edges. This will affect the quality of the resulting orthophotos. Therefore, it is necessary to incorporate good quality building outlines as breaklines during DSM interpolation. This study proposes a data-driven approach to construct building roof outlines from LiDAR point clouds by a workflow consisting of the following steps: given roof segments, roof boundary points are extracted using a concave hull algorithm. Straight edges may be difficult to find in complex roof configurations. Therefore, two ingredients are combined. First, RanSAC corner point preselection, and second, DBSCAN-based clustering of edge points. The method is demonstrated on an area of ±1.2 km2 containing 42 buildings of different characteristics. A quality assessment shows that the proposed method is able to deliver 92 % of building lines with acceptable geometric accuracy in comparison to a building line in the base map.

Change detection is an essential step to locate the area where an old model should be updated. With high density and accuracy, LiDAR data is often used to create a 3D city model. However, updating LiDAR data at state or nation level often takes years. Very high resolution (VHR) images with high updating rate is therefore an option for change detection. This paper provides a novel and efficient approach to derive pixel-based building change detection between past LiDAR and new VHR images. The proposed approach aims notably at reducing false alarms of changes near edges. For this purpose, LiDAR data is used to supervise the process of finding stereo pairs and derive the changes directly. This paper proposes to derive three possible heights (so three DSMs) by exploiting planar segments from LiDAR data. Near edges, the up to three possible heights are transformed into discrete disparities. A optimal disparity is selected from a reasonable and computational efficient range centered on them. If the optimal disparity is selected, but still the stereo pair found is wrong, a change has been found. A Markov random field (MRF) with built-in edge awareness from images is designed to find optimal disparity. By segmenting the pixels into plane and edge segments, the global optimization problem is split into many local ones which makes the optimization very efficient. Using an optimization and a consecutive occlusion consistency check, the changes are derived from stereo pairs having high color difference. The algorithm is tested to find changes in an urban areas in the city of Amersfoort, the Netherlands. The two different test cases show that the algorithm is indeed efficient. The optimized disparity images have sharp edges along those of images and false alarms of changes near or on edges and occlusions are largely reduced.

Exploiting nonlinear impact force induced by a stopper has been widely proposed to broaden the bandwidth of energy harvesters from mechanical vibrations, but the previous studies just directly employed the stopper and did not consider the effect of inherent factors of the stopper like its type and material. This letter devotes to efficiently utilize the stopper to broaden the resonance region of the energy harvester mainly from perspective of stopper type and material. Four stopper types are taken into account including the stopper with rigid and soft materials and the spring stopper. Experimental results show that two sides followed stopper type can greatly increase the bandwidth which is much better than other considered types. Importantly, using the stopper with soft materials or the spring stopper is followed by a wider resonance region and a higher output voltage of the energy harvester compared to that with rigid materials. Moreover, agreements between theoretical predictions and experimental measurements indicate the feasibility of using the trilinear spring model to describe the nonlinear impact force which is determined by stopper type and material. The present study delivers to essentially improve the output performance of energy harvesters by selecting the efficient stopper.

(Semi)-automatic facade reconstruction from terrestrial LiDAR point clouds is often affected by both quality of point cloud itself and imperfectness of object recognition algorithms. In this paper, we employ regularities, which exist on façades, to mitigate these problems. For example, doors, windows and balconies often have orthogonal and parallel boundaries. Many windows are constructed with the same shape. They may be arranged at the same lines and distance intervals, so do different windows. By identifying regularities among objects with relatively poor quality, these can be applied to calibrate the objects and improve their quality. The paper focuses on the regularities among the windows, which is the majority of objects on the wall. Regularities are classified into three categories: within an individual window, among similar windows and among different windows. Nine cases are specified as a reference for exploration. A hierarchical clustering method is employed to identify and apply regularities in a feature space, where regularities can be identified from clusters. To find the corresponding features in the nine cases of regularities, two phases are distinguished for similar and different windows. In the first phase, ICP (iterative closest points) is used to identify groups of similar windows. The registered points and a number of transformation matrices are used to identify and apply regularities among similar windows. In the second phase, features are extracted from the boundaries of the different windows. When applying regularities by relocating windows, the connections, called chains, established among the similar windows in the first phase are preserved. To test the performance of the algorithms, two datasets from terrestrial LiDAR point clouds are used. Both show good effects on the reconstructed model, while still matching with original point cloud, preventing over or under-regularization.

According to the study by Mayer and Jakowski (2009), periods of E layer dominated ionosphere (ELDI) are defined as being characterized by vertical electron density profiles having a maximum density at E layer altitudes. In this paper, characteristics of ELDI intervals have been investigated, focusing on their temporal variations, using field-aligned measurements from the EISCAT and ESR radars during the interval 2009–2011. ELDI events were identified using simple but reasonable criteria, in which a minimal duration was required to exclude possible "fake" events induced by random errors in measurements. It was found that ELDIs were observed more often in winter and earlier spring than other seasons, especially in the auroral zone. The occurrence of ELDI intervals peaks around geomagnetic midnight at auroral latitudes, while it reaches a maximum around geomagnetic local noon at the latitude of the ESR. Our results imply that ELDI intervals appear to be a sporadic rather than a regular phenomenon, in disagreement with previous results inferred from radio occultation measurements. The discrepancy between the typical durations of ELDI events observed by the two radars is remarkable, being 30 min on average at Tromsø but about a half of this at Svalbard. During intervals of ELDI, the mean thicknesses of the E layer are quite close at the two sites, as are the values of HmE and the ratio of NmE / NmF2. Case studies confirm that either extra E layer ionization or F layer density depletion alone could lead to the presence of ELDIs. Based on a careful check on ELDI intervals of various types, however, we suggest that both of them play a critical role in ELDI formation.

The causes for a productive upwelling region to be a source of CO2 are usually referred to the excess CO2 supplied via upwelling of high dissolved inorganic carbon (DIC) from deep water. Furthermore, we hypothesize that microbial activity plays a significant role on top of that. To test this hypothesis, multiple biogeochemical parameters were investigated at two cyclonic-eddy-induced upwelling sites, CE1 and CE2, in the western South China Sea. The data showed that upwelling can exert significant influences on biological activities in the euphotic zone and can also impact on particulate organic carbon (POC) export flux depending on upwelling conditions, such as the magnitude, timing, and duration of nutrient input and consequent microbial activities. At CE2, the increase of phytoplankton biomass caused by the upwelled nutrients resulted in increase of POC export flux compared to non-eddy reference sites, while at CE1 the microbial respiration of organic carbon stimulated by the upwelled nutrients significantly contributed to the attenuation of POC export flux. These results suggest that on top of upwelled DIC, microbial activities stimulated by upwelled nutrients and labile organic carbon produced by phytoplankton can play a critical role for an upwelling area to be outgassing or uptaking CO2. We point out that even though an upwelling region is outgassing CO2, carbon sequestration still takes place through the POC-based biological pump as well as the refractory dissolved organic carbon (RDOC)-based microbial carbon pump.

We evaluated the export fluxes of Particulate Organic Carbon (POC) in the Subtropical Frontal zone (STF) of the SW Pacific sector of the Southern Ocean. The site is characterized by enhanced primary productivity, which has been suggested to be stimulated through so-called natural iron fertilization processes where iron-depleted subantarctic water (SAW) mixes with mesotrophic, iron-replete subtropical water (STW). We adopted the small-volume 234Th method to achieve the highest possible spatial sampling resolution in austral late autumn-early winter, May–June, 2008. Inventories of chlorophyll-a, particulate 234Th and POC observed in the upper 100 m were all elevated in the mid-salinity water type (34.5 < S < 34.8), compared with low salinity waters (S < 34.5) which were of SAW origin with high macronutrients and high (S > 34.8) salinity waters which were of STW origin with low macronutrients. However, Steady-State 234Th fluxes were similar across the salinity gradient being, 25 ± 0.78 ((1.5 ± 0.047) × 103) in the mid-salinity, and 29 ± 0.53 ((1.8 ± 0.032) × 103) and 22 ± 1.1 Bq m−2 d−1 ((1.3 ± 0.066) × 103 dpm m−2 d–1) in the high and low salinity waters respectively. Bottle POC/Th ratios at the depth of 100 m were used to convert 234Th fluxes into POC export fluxes. The derived POC flux did not appear to be enhanced in mid-salinity waters where the primary productivity was inferred to be the highest at the time of sampling, with a flux of 11 ± 0.45 mmol C m−2 d−1, compared to 14 ± 0.39 mmol C m−2 d−1 in high salinity waters and 8.5 ± 0.66 mmol C m−2 d−1 in low salinity waters. This study thus implied that natural iron fertilization does not necessarily lead to an enhancement of POC export in STF regions.