Non-negative Matrix Factorization Research Articles

Source-oriented stochastic health risk assessment of toxic metals in soil via a hybrid model and Monte Carlo simulation

Toxic metal contamination in soils poses significant hazards to the environment and human health; thus, quantitative assessment of the sources and risks of metal contaminants are urgently needed. A hybrid model that integrates the positive matrix factorization (PMF) and random forest (RF) methods was proposed to quantify the sources of toxic metals in soils by combining diverse environmental variables (source proxies) in this study. In addition, a health risk assessment and Monte Carlo simulations were integrated to estimate the source-oriented stochastic health risk. The results suggested that, except for Ni, which exhibited moderate contamination, other toxic metals (As, Cd, Cr, Hg and Pb) presented slight contamination. Four sources (agricultural activities loaded heavily by As, atmospheric deposition loaded heavily by Hg and Pb, natural sources and mining activities loaded heavily by Cr and Ni, and industrial activities loaded heavily by Cd) were defined and explained 23.44 %, 26.65 %, 30.13 % and 19.78 % of the total variance in toxic metals, respectively. The principal route of exposure (i.e., ingestion), the population at highest risk (i.e., children), and the most hazard-inducing metals (i.e., As and Cr) were determined. Agricultural activities and the combination of natural sources and mining activities demonstrated certain degrees of noncarcinogenic risk to children, with exceedance ratios of 2.20 % and 2.56 %, respectively. Additionally, the combination of natural sources and mining activities demonstrated probabilities for significant carcinogenic risk to adults and children of 0.59 % and 3.76 %, respectively. To reduce the health risks of toxic metals in soils and to protect food and ecological safety, strict regulations should be established to control the discharge of waste from mining and agricultural activities.

Long-range transport of air pollutants increases the concentration of hazardous components of PM2.5 in northern South America

Abstract. Long-range transport (LRT) of air pollutants from a range of sources can substantially enhance background pollution levels, especially in urbanized regions, which can exacerbate high-pollution episodes. In the Aburrá Valley (AV), Colombia, and other cities in northern South America, biomass burning (BB), dust, and volcanic degassing have been identified as sources of long-range aerosol transport. However, the impact of these sources on air quality and the characterization of these sources have yet to be thoroughly studied. This work investigates the influence of these sources on the chemical composition of PM2.5 during annual and intra-annual high-load aerosol events in the AV. We identified, tracked, and meteorologically characterized LRT events and evaluated their influence on PM2.5 concentration and chemical composition. We found that the LRT of aerosols from BB, dust, and volcanic degassing influenced approximately 13 %, 8 %, and 13 % of days in the year, respectively. We applied the positive-matrix-factorization (PMF) statistical model to quantify PM2.5 concentrations and chemical compositions for the different LRT event types (e.g., BB). For BB events, we identified large contributions from organic carbon (OC1 and OC2), F−, and secondary aerosol tracers (SO42- and NO3-). For dust LRT events, crustal mineral components, along with Ti and Ca, were the primary contributors to aerosol composition, while SO42-, Na, Al, and Ca were the primary contributors during volcanic events. The concentrations of some ions and toxic heavy metals (Cr, Mn, Cd, and Ni) were also elevated during BB and volcanic-degassing events. BB contributed the most to PM2.5 levels during the LRT events (∼11 µg m−3), while contributions from aerosols arising from dust and volcanic events were also substantial (<7 µg m−3). Our study identifies the Orinoco and the Middle Magdalena Valley as sizable sources of BB aerosols and Nevado del Ruiz as a source of volcanic aerosols. Additionally, we found that African dust reached the Andean region via the Caribbean route. As a result, we identify the need for future chemical-transport modeling studies in the region and new support strategies to manage internal and external pollution sources that degrade air quality in the AV and the surrounding region.

Adversarial nonnegative matrix factorization for temporal link prediction

Temporal link prediction has been extensively studied and widely applied in various applications, aiming to predict future network links based on the historical networks. However, most existing methods ignore the behavior of previous network updating information in temporal networks. To address these issues, we propose a novel link prediction model based on adversarial nonnegative matrix factorization, which fuses graph representation and adversarial learning to perform temporal link prediction. Specifically, we add a bounded adversary matrix to the input matrix to provide the robustness against real perturbations. Then, our model fully exploits the impact of snapshots by using communicability. Simultaneously, we utilize the cosine similarity to extract the node similarity and map it to low-dimensional latent representation to preserve the local structure. Additionally, we provide effective updating rules to learn the parameters of this model. Extensive experiments results on six real-world networks demonstrate that the proposed method outperforms several classical and the state-of-art matrix-based methods.

Textual analysis and credit scoring: a new matrix factorization approach

Credit scoring models are important for financial institutions’ credit decisions. This study examined how variables are extracted from loan statements and whether textual variables can improve the accuracy of the default model. We used a combination of forward selection and non-negative matrix factorization to extract variables from loan statements. We also built a credit scoring model using both loan statement and numerical data. The results show that in the comparative analysis, the credit scoring model built using the optimal cut-off logistic regression model and the two types of data had the highest accuracy. Moreover, compared with the credit scoring model constructed using the deep learning method based on word vectors, the credit scoring model in this study had better interpretation. The regression analysis revealed that the variables from the loan statement have a significant effect on the default status.

Source apportionment and ecological risk assessment of antibiotics in Dafeng River Basin using PMF and Monte-Carlo simulation.

Antibiotics, prevalent in aquatic ecosystems, pose a grave threat to human health and the ecological well-being. This paper performed a case study on Dafeng River Basin in southern China. Specifically, techniques including positive matrix factorization (PFM) and Monte-Carlo simulation were employed to comprehensively investigate the spatial variations, possible sources, and ecological risks of antibiotics in four groups: sulfonamides (SAs), macrolides (MLs), quinolones (QNs), and tetracyclines (TCs). The major findings were as follows: first, 43 and 39 antibiotics were detected in the surface water and sediments of the basin, respectively, where the respective total content were ND-490.08ng/L and ND-144.34μg/kg, and the QNs and TCs were the two dominating groups. Second, the highest antibiotic content in surface water (441.43ng/L) was observed in the midstream area, whereas the highest concentration in sediments (68.41μg/kg) was found in the upstream region. Third, the investigation identified five sources of antibiotics discharged to surface water: domestic sewage, agricultural drainage, livestock discharge, sewage treatment plants, and aquaculture; three sources were detected for antibiotics in sediments: aquaculture, sewage treatment plants, and livestock discharge. Fourth, QNs had a significantly higher ecological risk than the other three groups of antibiotics, and livestock discharge (31.4% contribution) and aquaculture (23.4% contribution) were the main sources of risks of antibiotic contamination in Dafeng River Basin. This study is expected to provide some reference for control and risk management of antibiotic pollution in Dafeng River Basin.

Unsupervised learning of spatially-resolved ARPES spectra for epitaxially grown graphene via non-negative matrix factorization

This study proposed an unsupervised machine-learning approach for analyzing spatially-resolved ARPES. A combination of non-negative matrix factorization (NMF) and k-means clustering was applied to spatially-resolved ARPES spectra of the graphene epitaxially grown on a SiC substrate. The Dirac cones of graphene were decomposed and reproduced fairly well using NMF. The base and activation matrices obtained from the NMF results reflected the detailed spectral features derived from the number of graphene layers and growth directions. The spatial distribution of graphene thickness on the substrate was clearly visualized by the clustering using the activation matrices acquired via NMF. Integration with k-means clustering enables clear visualization of spatial variations. Our method efficiently handles large datasets, extracting spectral features without manual inspection. It offers broad applicability beyond graphene studies to analyze ARPES spectra in various materials.

Composition and Biogeochemical Effects of Carbohydrates in Aerosols in Coastal Environment

We adopted a simple and rapid measurement method to analyze the concentrations of monosaccharides (MCHO) and polysaccharides (PCHO) in carbohydrates, a subset of organic carbon found in size-fractionated atmospheric particles. Seasonal and source-related factors influenced carbohydrate concentrations, with total water-soluble carbohydrates (TCHO) accounting for approximately 23% of the water-soluble organic carbon (WSOC) in spring when biological activity was high. We observed that the mode of aerosol transport significantly influenced the particle size distribution of carbohydrates, with MCHO exhibiting relatively high concentrations in fine particles (<1 μm) and PCHO showing higher concentrations in coarse particles (>1 μm). Moreover, our results revealed that MCHO and PCHO contributed 51% and 49%, respectively, to the TCHO concentration. This contribution varied by approximately ±19% depending on the season, suggesting the importance of both MCHO and PCHO. Additionally, through the combined use of principal component analysis (PCA) and positive matrix factorization (PMF), we determined that biomass burning accounts for 30% of the aerosol. Notably, biomass burning accounts for approximately 52% of the WSOC flux, with MCHO accounting for approximately 78% of the carbon from this source, indicating the substantial influence of biomass burning on aerosol composition. The average concentration of TCHO/WSOC in the atmosphere was approximately 18%, similar to the marine environment, reflecting the relationship between the biogeochemical cycles of the two environments. Finally, the fluxes of MCHO and PCHO were 1.10 and 5.28 mg C m−2 yr−1, respectively. We also found that the contribution of atmospheric deposition to marine primary productivity in winter was 15 times greater than that in summer, indicating that atmospheric deposition had a significant impact on marine ecosystems during nutrient-poor seasons. Additionally, we discovered that WSOC accounts for approximately 62% of the dissolved organic carbon (DOC) in the Min River, suggesting that atmospheric deposition could be a major source of organic carbon in the region.

Co-tracing of groundwater emerging and conventional contaminants in karst area of Southwest China: Distribution, source, and source-specific health risks

Groundwater pollution, especially the co-contamination of emerging contaminants (ECs) and conventional contaminants, is increasingly recognized as a critical global concern. Karst aquifers are found throughout the world, providing critical water sources and sustaining rivers and ecosystems, and are particularly vulnerable to pollution. In this study, the characteristics of the contamination, potential sources, and source-specific health risks related to ECs and trace elements (TEs) in groundwater were investigated in Guiyang (a typical karst city in southwest China). The distribution of TEs followed a descending pattern from urban to suburban to rural areas. Interestingly, the median concentrations (217 ng/L) of Caffeine (CAF) in rural areas surpassed those in urban areas (145 ng/L) and even exceeded these in densely populated areas. Additionally, ECs and TEs in groundwater are significantly influenced by population density and land use types within a 1000-meter buffer zone around the sampling sites. Hierarchical cluster analysis and positive matrix factorization suggested that five principal sources of contamination were identified: domestic wastewater (29.7 %), industrial production (25.2 %), agricultural activities (16.3 %), natural and industrial sources (14.9 %), and pharmaceuticals (13.9 %). Monte Carlo simulations revealed that the non-carcinogenic risks to children in urban areas are significant, with 30.0 % of these risks exceeding the safety threshold. Source-specific health risk assessments indicated that industrial production is the primary source of risk. Furthermore, this study introduces a co-tracing method to trace the sources of groundwater contamination in karst systems. The co-tracing method based on the ECs and TEs allows for a deeper understanding of contaminant transfer and its fate in karst groundwater.

Immunotherapy and pan-apoptotic characterization of the tumor microenvironment in gastric cancer (STAD): a single-cell multidimensional analysis

BackgroundThe aim of this study was to elucidate the critical role of autophagy-related gene aggregation in gastric cancer tumor microenvironment cells and to investigate their major roles in cellular functions. In particular, the expression of these genes in tumor-associated fibroblast subtypes was scrutinized in an attempt to explain their cell-subpopulation-specific roles in cell–cell communication and regulation of cellular functions.MethodsIn this study, single-cell RNA sequencing data were first analyzed in multiple steps, including data preprocessing, cell clustering, and cell classification. Cell subpopulations and gene expression patterns were identified and analyzed using unsupervised non-negative matrix factorization (NMF) techniques. The dynamic expression of autophagy-related gene aggregates in various cell types was deciphered by pseudotime trajectory analysis (PTA). Intercellular communication analysis was performed using the CellChat R software package, revealing the intricate communication patterns and exchange of key signaling molecules between cell subpopulations, and SCENIC analysis was used to identify gene regulatory networks and reveal the mechanisms behind cellular heterogeneity.ResultCell subpopulations associated with pan-apoptosis were identified by NMF decomposition and SCENIC analysis. Cell–cell communication analysis revealed intricate communication patterns and exchange of key signaling molecules between cell subpopulations. Dynamic expression of autophagy-related genes aggregated in the pseudotemporal trajectory of STAD was observed by PTA. In the fibroblast subtype, different ligand-receptor interactions and their key roles in immunomodulation were observed.ConclusionBy deeply analyzing and comparing gene expression patterns within cellular subpopulations and intercellular communication, this study provides new insights into the role of pan-apoptosis-related genes in regulating immune responses and cellular functions in gastric cancer. These findings pave the way for further exploration of the role of these genes in tumorigenesis and immune regulation, as well as laying the foundation for potential therapeutic strategies.

Evaluation Method of Teaching Effect of Environmental Art and Design in Universities Based on Multi-Source Data Fusion Technology

The evaluation of teaching reform satisfaction is crucial for improving teaching quality. In the evaluation of satisfaction with teaching reforms in environmental art and design, issues such as the low reliability of evaluation indicator data can lead to unsatisfactory results. To address this, this paper proposes a new method for evaluating satisfaction with teaching reforms in environmental art and design based on multi-source data integration. By establishing principles for constructing a multi-source indicator system for evaluating satisfaction with teaching reforms in environmental art and design, a four-level multi-source indicator system was determined. This system mainly includes one first-level indicator, four second-level satisfaction evaluation indicators, and several third- and fourth-level evaluation indicators. The construction of the multi-source indicator system for evaluating satisfaction with teaching reforms in environmental art and design was completed. Using a four-part graph model and entropy weight calculation, the importance of the evaluation indicators for satisfaction with teaching reforms in environmental art and design was classified. The expert scoring method was introduced to correct the multi-source indicator data for evaluating satisfaction with teaching reforms in environmental art and design. Based on standardized indicator data, a penalty rule was added to the loss function of the multiple linear regression model to remove data noise, achieving pre-processing of the multi-source indicators for evaluating satisfaction with teaching reforms in environmental art and design. After completing the above process, first collect multi-source data from questionnaire surveys, student feedback, etc., and preprocess and map it to a unified ontology class. Subsequently, data features are extracted through non-negative matrix factorization, and the similarity between object types is calculated. Finally, based on the importance, redundancy, and interaction of data sources, an integrated quality model is designed to effectively integrate multiple sources of data, construct a satisfaction evaluation model, and output the score of satisfaction evaluation. The experimental results show that the proposed method has the highest satisfaction index, reaching about 0.99, indicating a high degree of recognition for the reform; when the amount of integrated data changes, the proposed method maintains the best integratison reliability. Even with an increase in data volume, its reliability only slightly decreases and stabilizes at a high level; meanwhile, the average absolute error of the proposed method is the lowest, consistently below 2.5%, indicating a high level of stability and accuracy in its evaluation results. The above results confirm that the proposed method can effectively evaluate the satisfaction of environmental art design teaching reform and improve teaching quality.

Organophosphate esters (OPEs) in the air and dust of new vehicle cabins: Concentrations, sources and contributions

Vehicles cabin represents unique indoor environment characterized by its limited space and multiple interior materials. For the first time, OPEs levels in vehicles and OPEs content in cabin materials were measured, and the material contributions were also analyzed. Between May and July in 2023, air, dust and materials samples were collected from 32 new vehicles. The detection rates were as high as 71%–100 % in dust and 100 % in selected cabin materials. The median of ∑OPEs in air-phase (both gas and particles) was significantly higher at 589.26 ng/m3 compared to those in other vehicle investigations. Material content analysis revealed that the top three were cushion textile, mat and lubricant, with ∑OPEs ranging from 85.94 μg/g to 1406.04 μg/g, which were three orders of magnitude higher than those in toys. Chlorinated forms such as TCIPP and TDCIPP were predominant among all air, dust and material samples analyzed. Particle-phase OPEs provided statistic diversity on manufactures and seat materials. Positive Matrix Factorization (PMF) analysis found that the combination of cushion textile, seat leather, and carpet substrate textile contributed significantly, accounting for 68 % of cabin OPEs.

Unsupervised Learning for the Automatic Counting of Grains in Nanocrystals and Image Segmentation at the Atomic Resolution.

Identifying the grain distribution and grain boundaries of nanoparticles is important for predicting their properties. Experimental methods for identifying the crystallographic distribution, such as precession electron diffraction, are limited by their probe size. In this study, we developed an unsupervised learning method by applying a Gabor filter to HAADF-STEM images at the atomic level for image segmentation and automatic counting of grains in polycrystalline nanoparticles. The methodology comprises a Gabor filter for feature extraction, non-negative matrix factorization for dimension reduction, and K-means clustering. We set the threshold distance and angle between the clusters required for the number of clusters to converge so as to automatically determine the optimal number of grains. This approach can shed new light on the nature of polycrystalline nanoparticles and their structure-property relationships.

Temporal variation, sources, fluxes, and risk assessment of heavy metals and arsenic in rainwater from Zhanjiang Bay, Northern South China Sea: Impact of typhoons Lion and Kompasu

This study investigates the impact of typhoon-induced rainfall on coastal pollution dynamics in Zhanjiang Bay during the 2021 wet season, focusing on typhoons Lion and Kompasu. Typhoon-induced rainfall (217 mm) contributed 23 % of the total wet season precipitation. Concentrations of Cu, Zn, Cd, and As in rainwater during typhoons were significantly diluted, showing decreases of 48 %, 48 %, 54 %, and 42 %, respectively. In contrast, Pb concentrations remained consistent (29.5 vs. 29.3 μg L−1), indicating that increased local emissions offset the dilution effect. The deposition fluxes of these elements during typhoons accounted for 12–21 % of total wet season deposits. A positive matrix factorization model identified six primary sources, highlighting a rise in coal combustion contributions during typhoons. Overall, while typhoons reduced risks for Cu, Cd, and As, risks for Zn and Pb increased upon deposition on Zhanjiang Bay, illustrating the complex impact of typhoon-induced rainfall on coastal pollution dynamics.

Two different approaches for source apportionment of ambient black carbon in highly polluted environments

The aethalometer model (AM) is widely used for source apportionment (SA) of black carbon (BC) in regions with mixed BC sources despite being initially developed for a relatively simplistic and low-pollution environments. The present study interrogates the applicability of AM in highly polluted metropolitan environments by comparing its results with the more nuanced Positive Matrix Factorization (PMF) model. The measurements were conducted in Delhi during the winter and summer season. PMF apportions the BC into diverse sources by taking help from complementary trace elemental measurements, thereby acknowledging the complex pollution landscape of the Delhi region. The AM estimates BCbb (BC from biomass burning) and BCff (BC from fossil fuel combustion) contributions as 48.3% and 51.7% during the winter and 16.6% and 83.4% during the summer, respectively.In contrast, the PMF model-derived biomass burning factor is the dominant source of BC during both winter and summer seasons, contributing 53.9% and 44% of the total BC, respectively. The decrease in light absorption at UV wavelengths of biomass-burning aerosols owing to escalated ambient aging is posited to be the reason for BCbb underprediction by the AM model during summers. Furthermore, while the AM model identifies fossil fuel combustion as the only other BC source apart from biomass burning, the PMF model apportions BC to five additional sources during winter, including vehicle emissions (22.9%), Pb-rich factor (10%), power plant (5.7%), waste incineration (4%) and industrial emission (3.6%). The contribution of these BC sources during summer is vehicular emission (16.5%), power plant (14.5%), waste incineration (11.5%), Pb-rich factor (9.5%), and industrial emission (4%). Additionally, the spectral variation of the light absorption properties of black carbon (bBCabs) and brown carbon (bBrCabs), delta-C effect, and sensitivity of the AM are reported for the study period. The present study cautions that BC source apportionment can be complex in highly polluted metropolitan environments, and complementary tracer measurements are recommended for reliable results.

Analysis of muscle synergies and muscle network in sling exercise rehabilitation technique

The study assessed motor control strategies across the four sling exercises of supine sling exercise (SSE), prone sling exercise (PSE), left side-lying sling exercise (LLSE), and right side-lying sling exercise (RLSE) positions base on the muscle synergies and muscle network analyses. Muscle activities of bilateral transversus abdominis (TA), rectus abdominis, multifidus (MF), and erector spinae (ES) were captured via surface electromyography. Muscle synergies were extracted through principal components analysis (PCA) and non-negative matrix factorization (NNMF). Muscle synergies number, muscle synergies complexity, muscle synergies sparseness, muscle synergies clusters and muscle networks were calculated. PCA results indicated that SSE and PSE decomposed into 2.88 ± 0.20 and 2.82 ± 0.15 synergies respectively, while the LLSE and RLSE positions decomposed into 3.76 ± 0.14 and 3.71 ± 0.11 muscle synergies, respectively, which were more complex (P = 0.00) but less sparse (P = 0.01). Muscle synergies clusters analysis indicated common muscle synergies among different sling exercises. SSE position demonstrated specific muscle synergies with a strong contribution of the bilateral TA. LLSE-specific synergy has a strong contribution of the left erector spinae (ES). The RLSE-specific synergy has significant contributions from the right ES and multifidus. Muscle networks were functionally organized, with clustering coefficient (F(1.5, 24) = 6.041, P = 0.01) and global efficiency of the undirected network (F(1.5, 24) = 6.041, P = 0.01), and betweenness-centrality of the directed network (F(2.7, 44) = 6.453, P = 0.00). Our research highlights the importance of evaluating muscle synergies and network adaptation strategies in individuals with neuromuscular disorders and developing targeted therapeutic interventions accordingly.

Integrated analysis and experiments uncover the function of disulfidptosis in predicting immunotherapy effectiveness and delineating immune landscapes in uterine corpus endometrial carcinoma.

Recently, a novel type of metabolic-regulated cell demise titled disulfidptosis has been discovered. Studies have demonstrated its importance in immune responses against cancer and its impact on the proliferation of cancer cells. Nonetheless, the precise mechanism and roles of disulfidptosis are not fully understood, particularly regarding the prognosis for individuals with uterine corpus endometrial carcinoma (UCEC). In this research, a distinctive disulfidptosis pattern was developed in UCEC, and by utilizing Non-negative Matrix Factorization (NMF) on 23 disulfidptosis related genes within the TCGA database, 3 distinct subgroups were distinguished. To collect data, we acquired gene expression profiles, somatic mutation information, copy number variation data, and corresponding clinical data from the TCGA and GEO database, specifically from UCEC patients. Cell line experiments and immunohistochemical (IHC) staining were conducted to validate the role of the LRPPRC in proliferation, migration and invasion. The genetic features and immune microenvironment of these subgroups were examined. It is worth mentioning that these subgroups offer important insights into comprehending the tumor microenvironment (TME) and the response of patients to immunotherapy and chemotherapy. Moreover, a disulfidptosis model was developed and validated, demonstrating a high level of accuracy in predicting the prognosis and outcomes of immunotherapy in UCEC patients. Additionally, a novel biomarker, LRPPRC, was identified, which can server as a promising predictor for forecasting prognosis in UCEC patients, with validation through tissue microarray staining and cell line experiments. This study has designed a classification system and a disulfidptosis model for UCEC, in addition to identifying a new biomarker, LRPPRC, for UCEC. These advancements serve as reliable and positive indicators for predicting outcomes and the efficacy of immunotherapy for each UCEC patient.

Hierarchical Biclustering of Mouse Pancreas Mass Spectrometry Imaging Data Using Recursive Rank-2 Non-negative Matrix Factorization.

One of the main challenges in mass spectrometry imaging data analysis remains the analysis of m/z-spectra displaying a low signal-to-noise ratio caused by their low abundance, sample preparation, matrix effects, fragmentation, and other artifacts. Additionally, we observe that molecules with a high abundance suppress those with lower intensities and misdirect classical tools for MSI data analysis, such as principal component analysis. As a result, the observed significance of a molecule may not always be directly related to its abundance. In this work, we present a recursive rank-2 non-negative matrix factorization (rr2-NMF) algorithm that automatically returns spectral and spatial visualization of colocalized molecules, both highly and lowly abundant. Using this hierarchical decomposition, our method finds spatial and spectral correlations on different levels of abundances. The quality of the analysis is evaluated on MALDI-TOF data of healthy mouse pancreatic tissue for the annotation of molecules of interest in the lower abundances. The results show interesting findings regarding the functioning and colocalization of certain molecules.

Characteristics and Source Apportionment of Atmospheric Volatile Organic Compounds in Zhengzhou During O3 Campaign Period

An online gas chromatograph （GC5000） was used to monitor the volatile organic compounds （VOCs） in the atmospheric environment of Zhengzhou City during the ozone campaign period from May to September of 2022. The relationship between O3 and its precursors as well as meteorology was analyzed and the pollution characteristics of VOCs during the O3 exceeding and non-exceeding the standard days were compared and explored. Different VOC activity evaluation methods of OFP and L·OH were utilized to compare and analyze the key active components and species and the ratio analysis （RA） and positive matrix factorization （PMF） analysis models were used to study the apportionment contribution of VOCs. The results showed that the O3 pollution in June and September in Zhengzhou was mainly due to the adverse meteorological conditions of high temperature and low humidity, strong radiation, and low wind speed, superimposed by the prominent concentrations of local VOCs and NO2, resulting in frequently high and excessive O3 occurrences. The VOCs concentration in Zhengzhou during the campaign period was an average of （68.3 ± 18.4） μg·m-3, whereas it was 75.7 μg·m-3 during O3 exceeding standard days and 13.4 μg·m-3 during O3 non-exceeding days, respectively. Among the VOC species, the OVOCs was 31.6%, accounting for the highest mass fraction, followed by halogenated hydrocarbon, alkane, and aromatic hydrocarbon, and the major species were ethane, n-butane, dichloromethane, propane, isopentane, toluene, chloromethane, 1,2-dichloroethane, and acetylene. VOC diurnal variation indicated that the emission of VOC pollution sources in the morning, evening peak, and at night should be paid more attention. The contribution of VOCs to OFP during the campaign period was （130.5 ± 46.4） μg·m-3, and the L·OH was （6.5 ± 2.9） s-1, among which the top 15 species with high activity were primarily acetaldehyde, isoprene, ethylene, m/p-xylene, toluene, hexal, isopentane, propanal, propylene, trans-2-butene, etc. In particular, the contributions of acetaldehyde, isoprene, ethylene, and hexal species were prominent during the O3 exceeding days. Ratio analysis showed that the B/T ratio in Zhengzhou from May to September ranged from 0.05 to 5.3, with an average value of 1.1 ± 0.6, and the regional VOCs was mainly controlled by the aging air mass with possible long-distance transports. The analysis of the PMF model showed that the major pollution sources to VOC concentration in Zhengzhou were motor vehicle exhaust emission sources and industrial solvent and secondary conversion sources, contributing 25.6% and 25.8%, respectively. The contribution rates of solvent coating sources, oil and gas volatile sources, plant emission sources, industrial solvents, and secondary conversion sources during O3 exceeding days were 5.4%, 4.7%, 3.3%, and 0.7% higher than those during O3 non-exceeding days, respectively. The research showed that the management of VOCs and NOx pollution sources should be strengthened to reduce their contribution to the O3 generation when O3 exceeds the standard.

Pollution of Heavy Metals in Topsoil of Remote Mountainous Areas in Western Yunnan-Guizhou Plateau

Heavy metals from anthropogenic emissions have had a negative impact on the ecological environment in remote regions. A total of 69 topsoil samples were collected from 13 remote mountainous areas in the western Yunnan-Guizhou Plateau at altitudes of 2 563-4 037 m, and the concentrations of ten heavy metals in the samples were determined. Enrichment characteristics and pollution sources of heavy metals in topsoil were discussed by referencing the enrichment factor （EF）, positive matrix factorization （PMF）, and Pb isotopes. The results showed that the average concentrations of Al, Fe, Cu, V, and Zn in the topsoil were lower than the soil background values in Yunnan Province； the average concentrations of Ni and Pb were similar to the background values； and the average concentrations of Cd, Cr, and Hg were 1.8-3.6 times higher than the background values. The average EF values of Pb, Cr, and Ni were 3.8, 3.4, and 2.3, respectively, showing moderate enrichment according to the EF classification criteria； the average EF values of Cd and Hg were 15.2 and 10.0, reflecting significant enrichment； and the average EF values of the other metals ranged from 1.1 to 1.9, displaying none-weak enrichment. Combining the comparisons of heavy metal concentrations and ratios in topsoil and bedrock and the EF and PMF results, Al, Fe, V, Cr, Cu, Ni, and Zn in topsoil were considered primarily from detrital sources, and the spatial concentration variations of the metals should have been mainly regulated by the parent material. Cadmium, Hg, and Pb were obviously polluted by anthropogenic emissions, and the main sources were non-ferrous metal smelting and coal combustion. The areas with relatively high Cd, Hg, and Pb pollution were mainly distributed in the Jiaozi snow mountain, Bitahai watershed, Luoji Mountain, and Laojun Mountain areas. Anthropogenic emissions contributed 23.8% of the accumulation of heavy metals in the topsoil.

Pollution Characteristics and Source Apportionment of Volatile Organic Compounds in Typical Solvent-using Industrial Parks in Beijing

The BCT-7800A PLUS VOC online monitor system was employed to measure ambient volatile organic compounds （VOCs） in a typical solvent-using industrial park in Beijing. From January to June 2023, the pollution characteristics, source apportionment, and ozone formation potential（OFP）of VOCs were studied, and the results of a comparative analysis were also discussed between heating and non-heating periods. The results indicated that VOC concentrations from January to June 2023 were （104.21 ± 91.31） μg·m-3 on average. The concentrations of TVOCs under the influence of southerly and northerly winds were （214.18 ± 202.37） μg·m-3 and （197.56 ± 188.3） μg·m-3, respectively. Alkanes were the species with the highest average concentration and proportion, respectively （45.53 ± 41.43） μg·m-3. The VOC concentration during the heating period was higher than those during the non-heating period, with values of （111.57 ± 83.96） μg·m-3 and （87.92 ± 75.03） μg·m-3, respectively. Propane and ethane were the species with the highest average concentration during the heating period. Compared with those in the non-heating period, the average concentrations of three species （propane, ethane, and n-butane） in the top ten species increased during the heating period, with average concentrations increasing by 51.94%, 54.64%, and 26.32%, respectively. The source apportionment results based on the positive matrix factorization （PMF） model indicated that the major sources of VOCs in the park during the monitoring period were printing emission sources （4.95%）, oil and gas evaporation sources （9.52%）, fuel combustion sources （15.44%）, traffic emissions sources （18.97%）, electronic equipment manufacturing （24.59%）, and industrial painting sources （26.52%）. Therefore, industrial painting sources, electronic equipment manufacturing sources, and traffic emissions sources were the emission sources that the park should focus on controlling. Compared with those during non-heating periods； industrial painting, traffic emission, and fuel combustion sources contributed more during the heating period, with VOC concentrations increasing by 15.02%, 16.53%, and 24.98%, respectively. The average OFP of VOCs from May to June during the monitoring period was 198.51 μg·m-3 and OVOCs, olefins, and aromatic hydrocarbons contributed the most to OFP, which were 47.41%, 22.15%, and 18.41%, respectively. The electronic equipment manufacturing source was the largest contributor to the summer OFP of the park and its contribution rate was 30.11%, which should be strengthened in the future.