PARAFAC Model Research Articles

PARAFAC study of L-cys@CdTe QDs interaction to BSA, cytochrome c and trypsin: An approach through electrostatic and covalent bonds

Utilizing fluorescence spectroscopy, non-covalent and covalent interactions of L-cys@CdTe quantum dots to bovine serum albumin (BSA), cytochrome c and trypsin were investigated. L-cys@CdTe QDs with the emission maximum at 530nm and an average diameter of 2.6nm were synthesized in the aqueous medium. Formaldehyde, N-(3-dimethylaminopropyl)-N-ethylcarbodiimide hydrochloride (EDC) with N-hydroxysuccinimide (NHS), and glutaraldehyde was applied as cross-linkers. In the case of both electrostatic and covalent strategies PARAFAC, as a powerful multi-way chemometrics technique, was utilized to analyze fluorescence excitation-emission (EEM) spectra. For non-covalent and covalent bonding, two and three significant components composed the PARAFAC models. Resolved EEM shows that in the presence of formaldehyde, a new component with an emission peak similar to BSA was obtained. Using EDC-NHS cross-linker, the fluorescence peak of the newly formed component was in a distinct wavelength with similar emission intensity, compared to L-cys@CdTe QDs and BSA. Employing glutaraldehyde, a distinguished component was easily detected at emission wavelengths higher than that of L-cys@CdTe QDs and proteins. It was concluded that the choice of cross-linker is a critical step to create different emission spectra when dealing with nano-bio-conjugations. This study shows that glutaraldehyde cross-linker leads to increase sensitivity, selectivity, and accuracy of protein analysis.

Factor Uniqueness of the Structural Parafac Model

Factor analysis is a well-known method for describing the covariance structure among a set of manifest variables through a limited number of unobserved factors. When the observed variables are collected at various occasions on the same statistical units, the data have a three-way structure and standard factor analysis may fail. To overcome these limitations, three-way models, such as the Parafac model, can be adopted. It is often seen as an extension of principal component analysis able to discover unique latent components. The structural version, i.e., as a reparameterization of the covariance matrix, has been also formulated but rarely investigated. In this article, such a formulation is studied by discussing under what conditions factor uniqueness is preserved. It is shown that, under mild conditions, such a property holds even if the specific factors are assumed to be within-variable, or within-occasion, correlated and the model is modified to become scale invariant.

Three‐way clustering around latent variables approach with constraints on the configurations to facilitate interpretation

AbstractThe set‐up of comprehensive studies in life sciences involving a longitudinal dimension—as appears in time‐scale metabolomics—calls for the use of dimension reduction techniques for three‐way data structures (e.g., samples by variables by time points). For this purpose, a clustering around latent variables for three‐way data approach, CLV3W, has been proposed. CLV3W aims at both partitioning the variables into nonoverlapping clusters and estimating within each cluster a rank‐one Parafac model consisting of a latent component (resp. a weighting system) associated with the first mode (resp. third mode) and a vector of loadings reflecting the degree of closeness of each variable of the second mode to its cluster. In this paper, two constrained CLV3W models are discussed. First, a nonnegativity constraint is defined implying that clusters are composed of positively correlated variables. Second, it is proposed to constrain the weighting system to be the same for all clusters. These two constraints aim at providing more parsimonious models with configurations that are easier to interpret. The appropriateness of both constraints is evaluated in a simulation study and illustrated on two case studies pertaining to sensory evaluation and metabolomics data. Regarding the first case study, CLV3W yields the identification of two consumer segments together with one common emotional pleasantness dimension associated with coffee aromas. CLV3W analysis of human preterm breast milk metabolomics data provided three clusters of lipid species that are responsible for specific functions (i.e., milk fat globules membrane‐constituents, fatty acid oxidation‐products, lipid mediators as eicosanoids and endocannabinoids).

Simultaneous determination of benzenediol isomers in tap water by second-order calibration and voltabsorptometry

K4[Fe(CN)6] and benzenedial isomers: catechol, resorcinol and hydroquinone, were studied by voltoabsortometry, absorption spectroscopy versus potential sweep, to generate second order data sets. Linear sweep voltammetry was used for K4[Fe(CN)6] and differential pulse voltammetry for benzenediol isomers. Quantitative models were constructed using the PARAFAC algorithm in samples of tap water with the addition of catechol, resorcinol and hydroquinone and RMSEP of 0.06, 0.07 and 0.03 mmolL−1 were obtained, respectively. For the U-PLS/RBL algorithm, the RMSEP were 0.05, 0.10 and 0.03 mmolL−1, respectively for catechol, resorcinol and hydroquinone. When the data set is derived with respect to the potential, PARAFAC recovers a derived kinetic profile that fits the voltammogram in the model system, K4[Fe(CN)6], but presents displacement in the benzenediol system and a break in the trilinearity of resorcinol. The RMSEP obtained for the derived data for PARAFAC were 0.06 and 0.05 mmolL−1 for catechol and hydroquinone, and 0.07, 0.04 and 0.07 mmolL−1 for U-PLS/BRL, for catechol, resorcinol and hydroquinone, respectively. Degenerate solutions were observed for PARAFAC models and confirmed by MCR-ALS with soft modeling closure constraint. It is observed that degenerate solutions should be expected in higher-order quantitative applications involving spectroelectrochemistry data. Although degenerate solutions are not a pure component solution, they do not affect quantitative models.

Temporal and Spatial Distribution Characteristics and Difference Analysis of Chromophoric Dissolved Organic Matter in Sediment Interstitial Water from Gangnan Reservoir

The chromophoric dissolved organic matter (CDOM), the main component of dissolved organic matter, affects the morphological characteristics, migration, and conversion of pollutants in water. Based on UV-vis spectra and excitation emission matrix spectroscopy (EEMs) combined with the parallel factor analysis (PARAFAC), the spatial distribution and spectral characteristics were investigated and source analysis of CDOM was performed. Thus, the spatiotemporal differences in the CDOM in Gangnan Reservoir were analyzed. Results showed that a254, a260, a280, and a355 exhibited significant seasonal differences in Gangnan Reservoir, and the order of CDOM concentrations was summer > spring > autumn > winter. There are significant seasonal differences in the E2/E3, E3/E4, E4/E6, and SR of interstitial water CDOM. The concentrations of E2/E3, E3/E4, E4/E6, and SR were high in winter and low in summer. E2/E3 and E3/E4 in autumn and winter were significantly higher than those in spring and summer, and the E3/E4 in autumn and winter was greater than 3.5, which indicates that the CDOM of the autumn and winter sediments has a smaller molecular weight and a lower degree of humification. Protein-like substances (C1), short-wave fulvic acid (C2), and degraded humic substances (C3) were identified by the PARAFAC model, and there was a significant positive correlation among the three fluorescent components (P<0.001). The total fluorescence intensity of CDOM and the fluorescence intensity of each fluorescent component show significant seasonal differences. The total fluorescence intensity and the fluorescence intensity of each component show the highest levels in spring, followed by autumn and winter, and the lowest levels in summer. The proportion of each fluorescent component in autumn and winter and that of each fluorescent component in spring and summer showed no significant difference. There was a significant difference in the proportion of each fluorescent component between autumn/winter and spring/summer. The BIX and FI of CDOM for autumn and winter were higher than those for spring and summer, indicating that the autogenous source of CDOM in autumn and winter is stronger than that in spring and summer, which was consistent with the result of HIX. PCA and Adonis analysis showed that the spectral characteristics of CDOM exhibited obvious seasonal differences (P<0.001). Moreover, the C1, C2, and C3 and water quality parameters (NH4+, NO3-, NO2-, TDN, and TDP) exhibited significant correlation based on linear regression. The results could provide technical support for the control of organic carbon pollution sources and water quality management in Gangnan Reservoir.

A Comparison of Non-negative Tucker Decomposition and Parallel Factor Analysis for Identification and Measurement of Human EEG Rhythms

Abstract Analysis of changes in the brain neural electrical activity measured by the electroencephalogram (EEG) plays a crucial role in the area of brain disorder diagnostics. The elementary latent sources of the brain neural activity can be extracted by a tensor decomposition of continuously recorded multichannel EEG. Parallel factor analysis (PARAFAC) is a powerful approach for this purpose. However, the assumption of the same number of factors in each dimension of the PARAFAC model may be restrictive when applied to EEG data. In this article we discuss the potential benefits of an alternative tensor decomposition method – the Tucker model. We analyze situations, where in comparison to the PARAFAC solution, the Tucker model provides a more parsimonious representation of the EEG data decomposition. We show that this more parsimonious representation of EEG is achieved without reducing the ability to explain variance. We analyze EEG records of two patients after ischemic stroke and we focus on the extraction of specific sensorimotor oscillatory sources associated with motor imagery during neurorehabilitation training. Both models provided consistent results. The advantage of the Tucker model was a compact structure with only two spatial signatures reflecting the expected lateralized activation of the detected subject-specific sensorimotor rhythms.

Natural organic matter-enhanced transportation of iodine in groundwater in the Datong Basin: Impact of irrigation activities

Local residents in the Datong Basin of northern China are exposed to groundwater with elevated iodine concentrations. Natural organic matter (NOM) has been linked to the heterogeneous distribution of elevated iodine in groundwater used for irrigation purposes, but little is known about the effects of hydrologic fluctuations and NOM characteristics on the transport and enrichment of iodine in the groundwater. Cl/Br molar ratios in Datong Basin groundwater range widely from 133 to 2099. A rapid increase in Cl/Br molar ratio with increasing Cl content indicates hydrologic fluctuations from the upper groundwater to the deeper aquifer due to large-scale irrigation activities in the Basin. A two end-member model of groundwater δ2H and δ18O values suggests the contribution of upper water recharging groundwater ranges from 20.7 to 49.5%. This vertical recharge process predominantly controls iodine enrichment and distribution in the groundwater. Additionally, the correlation between DOC concentration and δ18O signatures indicates considerable fresh organic matter is imported into the aquifer during the vertical recharge process. Iodine mobilization is likely promoted by young carbon transported to the deeper aquifer in the organo‑iodine form. Excitation-emission matrix (EEM) results indicate humic-like substances dominate NOM in the groundwater. Evidence from a PARAFAC model suggests organic matter in groundwater samples is associated with microbially-mediated degradation processes in an anaerobic environment. The drawdown migration of organic matter from the upper soil/sediments or surface could provide an extra energy source that promotes microbial activity. Buried sedimentary iodine coupled with anaerobic microbial respiration of subsurface organic carbon within the aquifer could lead to the release of iodine into the groundwater. These findings pave the way for a more comprehensive assessment of the susceptibility of drinking water aquifers, thereby supporting the management of groundwater resources.

Characterization of the fate and changes of post-irradiance fluorescence signal of filtered anthropogenic effluent dissolved organic matter from wastewater treatment plant in the coastal zone of Gapeau river.

Anthropogenic effluent dissolved organic matter (DOM) plays an important role in coastal zone pollution. The objectives of the present study were to characterize the fluorescence signal of anthropogenic effluent DOM from wastewater treatment plant and to evaluate the effect of solar irradiation on the fluorescence signal in the coastal zone. Solar irradiation experiments were conducted to evaluate the effect photochemical degradation using excitation-emission matrix (EEM) method combined with parallel factor analysis (PARAFAC). Results showed high fluorescence of DOM before irradiation and the intensity tends to decrease after 4th and 15th day of irradiation. Rapid photochemical degradation of humic-like fluorophores and appearance of a post-irradiance dominant anthropogenic effluent DOM fluorophores were also observed after irradiation. Our experiments showed a sharp reduction in fluorescence intensity which occurred after 4th day of solar irradiation and the fluorescence signal did not disappeared after 15th day indicating the formation of a specific signal due to solar irradiation. PARAFAC model divided the bulk EEM spectra into three individual fluorescent components with C1 "terrestrial humic-like" and C2 "humic-like of longer wavelength" and C3 is a noisy component with two emission maxima. Multilinear regression of PARAFAC components contribution with mixing composition was most suitable according to the equation C*i = AWWi,0 + AWWi,1.fSW + AWWi,2.fRW, where C*i is the normalized contribution of PARAFAC component number i in a given irradiation day; AWWi,0, AWWi,1, AWWi,2 are the multilinear regression coefficients and contain implicitly the effect of fWW; and WW, SW, and RW are treated wastewater, sea water, and river water respectively. The values of AWWi,0, AWWi,1, and AWWi,2 fitted second-order kinetics with irradiation process with kinetic constant of 9.68, - 987.35, and - 977.67 respectively for C1 equation and the same trend for C2 and no values for C3 due to its noisy character indicating the rapid degradation with increase of fSW and fRW and the predominance of the residual fluorescence coming from fWW which is the content fraction of anthropogenic effluent DOM because AWWi,0 was 100 times less sensitive to photobleaching. A suitable model for predicting the fluorescence EEMs as a function of mixing composition was developed.

Isothermal Chemical Denaturation: Data Analysis, Error Detection, and Correction by PARAFAC2.

Characterization of a protein's conformational stability is a key step in the development of biotherapeutics, where protein unfolding leads to adverse properties, such as aggregation and loss of efficacy. Isothermal chemical denaturation (ICD) can be applied to determine chemical stability, aiming to identify the optimal solvent conditions, in terms of pH, salt concentration, and added excipients. For seven monoclonal antibodies, this study investigates the observed intrinsic protein fluorescence emission spectra as a function of denaturant concentration. Protein formulations are screened in two experimental series. We show how the peak shapes of folded and unfolded proteins are preserved under added salt (0-140 mM NaCl) and added excipients concentrations, as typically found in biotherapeutic formulations and that only minor effects in tryptophan fluorescence peak tailing are observed over a large pH range (5.5-9.0). The data of seven mAbs, where GuHCl was a suitable denaturant, are modeled using PARAFAC2. PARAFAC2, a linear decomposition method, is well suited for the data and yields robust, valid, and automated models that allow for the detection of erroneous measurements. Analysis of the errors show correlation with the well-based experimental setup, and differences in observed errors between the two experimental series. We additionally show a correction method for these outliers based on PARAFAC2 model scores, such that full transition curves can be retrieved, increasing the accuracy of any subsequent analysis.

The First Identification of the Uniqueness and Authentication of Maltese Extra Virgin Olive Oil Using 3D-Fluorescence Spectroscopy Coupled with Multi-Way Data Analysis.

The potential application of multivariate three-way data analysis techniques, namely parallel factor analysis (PARAFAC) and discriminant multi-way partial least squares regression (DN-PLSR), on three-dimensional excitation emission matrix (3D-EEM) fluorescent data were used to identify the uniqueness and authenticity of Maltese extra virgin olive oil (EVOO). A non-negativity constrained PARAFAC model revealed that a four-component model provided the most appropriate solution. Examination of the extracted components in mode 2 and 3 showed that these belonged to different fluorophores present in extra virgin olive oil. Application of linear discriminate analysis (LDA) and binary logistic regression analysis on the concentration of the four extracted fluorophores, showed that it is possible to discriminate Maltese EVOOs from non-Maltese EVOOs. The application of DN-PLSR provided superior means for discrimination of Maltese EVOOs. Further inspection of the extracted latent variables and their variable importance plots (VIPs) provided strong proof of the existence of four types of fluorophores present in EVOOs and their potential application for the discrimination of Maltese EVOOs.

TASTE: Temporal and Static Tensor Factorization for Phenotyping Electronic Health Records.

Phenotyping electronic health records (EHR) focuses on defining meaningful patient groups (e.g., heart failure group and diabetes group) and identifying the temporal evolution of patients in those groups. Tensor factorization has been an effective tool for phenotyping. Most of the existing works assume either a static patient representation with aggregate data or only model temporal data. However, real EHR data contain both temporal (e.g., longitudinal clinical visits) and static information (e.g., patient demographics), which are difficult to model simultaneously. In this paper, we propose Temporal And Static TEnsor factorization (TASTE) that jointly models both static and temporal information to extract phenotypes. TASTE combines the PARAFAC2 model with non-negative matrix factorization to model a temporal and a static tensor. To fit the proposed model, we transform the original problem into simpler ones which are optimally solved in an alternating fashion. For each of the sub-problems, our proposed mathematical re-formulations lead to efficient sub-problem solvers. Comprehensive experiments on large EHR data from a heart failure (HF) study confirmed that TASTE is up to 14× faster than several baselines and the resulting phenotypes were confirmed to be clinically meaningful by a cardiologist. Using 60 phenotypes extracted by TASTE, a simple logistic regression can achieve the same level of area under the curve (AUC) for HF prediction compared to a deep learning model using recurrent neural networks (RNN) with 345 features.

PARAFAC model as an innovative tool for monitoring natural organic matter removal in water treatment plants.

The increase of fluorescent natural organic matter (fNOM) fractions during drinking water treatment might lead to an increased coagulant dose and filter clogging, and can be a precursor for disinfection by-products. Consequently, efficient fNOM removal is essential, for which characterisation of fNOM fractions is crucial. This study aims to develop a robust monitoring tool for assessing fNOM fractions across water treatment processes. To achieve this, water samples were collected from six South African water treatment plants (WTPs) during winter and summer, and two plants in Belgium during spring. The removal of fNOM was monitored by assessing fluorescence excitation-emission matrices datasets using parallel factor analysis. The removal of fNOM during summer for South African WTPs was in the range 69-85%, and decreased to 42-64% in winter. In Belgian WTPs, fNOM removal was in the range 74-78%. Principal component analysis revealed a positive correlation between total fluorescence and total organic carbon (TOC). However, TOC had an insignificant contribution to the factors affecting fNOM removal. Overall, the study demonstrated the appearance of fNOM in the final chlorinated water, indicating that fNOM requires a customised monitoring technique.

Effects of digestate DOM on chemical behavior of soil heavy metals in an abandoned copper mining areas

The influence of digestate dissolved organic matter (DOM) on chemical behavior of soil heavy metals (HMs) in an abandoned copper mining areas was explored by fluorescence quenching titration and heavy metal extracting experiment. Five fluorescent components were obtained from digestate DOM by PARAFAC model combined with the EEM data. The stability constant (log KM) values were in the range of 4.95–5.53, 5.05–5.29, 5.21–6.00, and 4.12–4.75 for DOM-Cr(III), DOM-Cu(II), DOM-Fe(III) and DOM-Pb(II) complexes, respectively. Alcohols, ethers and esters in digestate DOM were preferentially combined with Fe(III), Cu(II) and Zn(II). However, phenolic hydroxyl groups were more likely to combine with Cr(III) and Pb(II). The speciation distribution of HMs indicated that mining resulted in a higher concentration of Cu(II) in the grassland soil (GS) than those in the agricultural soil (AS) and forest land soil (FS). Fe–Mn oxides and organic forms of Pb(II) increased dramatically due to mining. Digestate DOM extraction can increase the content of Cr(III), Fe(III) and Pb(II), and decrease the content of Cu(II) and Zn(II) in the AS, GS, and FS. However, the contents of HMs in the mining soil (MS) and slag soil (SS) decreased due to the application of digestate DOM, except for Cu(II) in the SS.

Missing data recovery combined with Parallel factor analysis model for eliminating Rayleigh scattering in the process of detecting pesticide mixture.

Excitation-Emission Matrix (EEM) fluorescence spectroscopy combined with Parallel factor analysis (PARAFAC) provides a widely used method to extract useful information containing unknown components. However, the inherent scattering especially Rayleigh scattering will influence the accuracy of PARAFAC so that appropriate procedure to the scattering becomes an essential problem when processing the EEM data. Many methods have been proposed to solve the problems about eliminating scattering. Missing data recovery combined with PARAFAC model has been discussed in this paper. For EEM data, this method extracted the signal values in the missing area which can effectively correct scattering region. It can eliminate Rayleigh scattering effectively by choosing Gaussian contour constraint. The results of the correlation coefficient (R), the root mean square error of prediction (RMSEP) and average recovery rate (AR) are better which can prove that the combined method is easier to implement and provide better concentration prediction results in detecting pesticide mixture.

Assessment of C-DBP and N-DBP formation potential and its reduction by MIEX® DOC and MIEX® GOLD resins using fluorescence spectroscopy and parallel factor analysis

This study investigated the applicability of parallel factor analysis (PARAFAC) of fluorescence excitation-emission matrices (EEM) spectra to assess the formation potentials (FP) of carbonaceous and nitrogenous disinfection byproducts (C-DBP and N-DBP) and the FP reduction by the magnetic ion exchange resins, MIEX® DOC and MIEX® GOLD. Two source waters of different nature — a surface water and a secondary treated wastewater effluent — were studied. The samples were analyzed for formation potentials of trihalomethanes (THM4), haloacetonitriles (HAN4), haloketones (HK2), and chloropicrin (CPN). A 4-component PARAFAC model was developed from 150 EEM samples generated from the raw source waters and their treatment with MIEX® resins. Components C1, C2, and C3 corresponded to humic-like dissolved organic matter (DOM) while C4 corresponded to protein-like DOM. Both MIEX® resins preferentially removed components C1, C2, and C3 over C4, indicating affinity with humic materials. MIEX® resins were shown to be more effective to treat surface water than secondary effluent, including effective removal of DBP precursors with extended bed volume treatment. Among all parameters investigated, THM4-FP strongly correlated with humic-like component C3, while HAN4-FP strongly correlated with protein-like component C4 (ρ > 0.89 and p < 0.01); CPN-FP and HK2-FP both correlated with anthropogenic DOM C2 (ρ > 0.89 and p < 0.01). Our results indicate that EEM–PARAFAC was valuable for assessing DBP formation potentials and removal of their precursors by MIEX® resins in different water sources.

Three-way analysis of pH-UV absorbance dataset for the determination of paracetamol and its pKa value in presence of excipients.

A three-way analysis method, parallel factor analysis (PARAFAC) model was applied to the pH-absorbance dataset for the simultaneous determination of paracetamol and its acid-base dissociation constant in presence of excipient interference in a syrup formulation without using chemical pretreatment or chromatographic separation step. The UV spectroscopic data matrices of calibration set, validation and unknown samples were obtained from the absorbance measurements at the five different pH media, considering conjugate acid/base properties of the related drug. Their pH-absorbance data matrices were arranged as a cubic data array (wavelength x sample x pH) (425x52x5). Three-way array of pH-absorbance dataset was decomposed into a trilinear set of spectral, pH and relative concentration profiles of paracetamol and excipients in the commercial syrup using PARAFAC model. In the PARAFAC implementation, paracetamol in the commercial syrup formulation and its pKa value were simultaneously predicted from the relative concentration and pH profiles, respectively. In the method validation step of this study, the performance of PARAFAC model was checked by analyzing the validation samples in terms of selectivity, sensitivity, accuracy and precision of the method. The determination results of paracetamol and its pKa value provided from PARAFAC application were compared to those obtained by a newly developed ultra-performance liquid chromatography (UPLC) method, in terms of simplicity, applicability, interpretability with low cost and short analysis time.

A novel PARAFAC model for continental hot springs reveals unique dissolved organic carbon compositions

Dissolved organic carbon in hot springs reflects a range of sources and biogeochemical processes. We evaluated ~200 continental hot spring samples, with a range in pH and temperature, collected from the Tengchong hydrothermal region, Yunnan Province, China and Yellowstone National Park, Wyoming, USA. Dissolved organic carbon concentrations ranged from 16.7 µM to 2.97 mM. Acidic springs displayed the highest values and widest range in carbon concentration. Alkaline springs had a narrower range and lower average concentrations. Carbon composition was evaluated using ultraviolet absorption and 3D-fluorescence spectroscopy. Total fluorescence was correlated (p < 0.05) with dissolved organic carbon (DOC) concentration. Fluorescence excitation-emission matrices were deconvolved using parallel factor analysis. We validated a five-component model that represented >97% of the total fluorescence. Our model includes three humic-like components, one protein-like component, and one novel component exclusively observed in highly acidic springs. The closest spectral match to the novel component is an acid-soluble lignin produced during high-temperature, acid digestion of wood pulp. Humic-like components were dominant in mid-pH springs (4 < pH < 7) indicating these springs had greater terrestrial carbon input. Acidic springs also exhibited evidence for terrestrial carbon input. Alkaline springs, in contrast, consistently had low dissolved organic carbon content and low fluorescence intensity suggesting that these springs had little terrestrial input. This absence of terrestrial carbon implies a predominantly hydrothermal fluid source. A comparison of the traditional fluorescence indices with our five model components suggest that these indices may have limited utility in continental hot springs with multiple organic matter sources and alteration processes.

Fluorescence components of natural dissolved organic matter (DOM) from aquatic systems of an Andean Patagonian catchment: Applying different data restriction criteria for PARAFAC modelling.

Fluorescence spectroscopy is a widely applied technique to characterize the composition of the fluorescent fraction of dissolved organic matter (DOM), allowing to infer sources and diagenetic state of soil, marine and freshwater DOM samples. The analysis of fluorescent DOM (FDOM) is often carried out by multi-way models such as parallel factor analysis (PARAFAC), which allows decomposing excitation-emission matrices (EEMs) obtained from DOM samples into their underlying chemical components. Some aspects of the performance/accuracy of the EEM-PARAFAC technique regarding the use of selective vs. non-selective EEM data are still in discussion. In this investigation, we evaluated the outcome of two different approaches (non-selective and selective) applied to study the composition of DOM from four headwater streams (Case 1) and from two neighboring shallow lakes (Case 2), all belonging to the same Andean watershed within Nahuel Huapi National Park in North Patagonia (Argentina). In both cases, the outcome of the PARAFAC performed to non-selective data (EEM datasets from all the streams and the two lakes) vs. selective datasets (EEMs from each stream and each lake treated separately) is compared on the basis of modelled fluorescent components. Regardless of the restriction criteria applied for the analyses, the results obtained indicated similar component loadings in the four streams and in the two lakes. The similarity of the outcomes likely relates to the low internal variation of the EEMs, since these are located in the same catchment, influenced by similar soils and vegetation which are the main DOM sources. Therefore, we conclude that the use of a small selective EEM dataset may not condition the validation of the FDOM components and their temporal dynamics.

Seasonal variation and ecological risk assessment of dissolved organic matter in a peri-urban critical zone observatory watershed

Peri-urban ecosystems are among the most intensive areas in terms of competition between different ecosystem components. Dissolved organic matter (DOM) plays a significant role in aquatic carbon cycling. The chemical composition of DOM and associated potential ecological risks in peri-urban aquatic ecosystems are poorly understood. Herein, we used fluorescence excitation-emission matrix and parallel factor analysis (EEM-PARAFAC) to characterize DOM in a peri-urban critical zone observatory watershed in Eastern China. According to the theory of natural disaster risk formation, we calculated the ecological risk of DOM in the peri-urban watershed. Seasonal variation in DOM concentrations was observed, whereas fluorescent DOM concentrations were site-specific across four sub-watersheds. The analysis of DOM absorption properties revealed the presence of DOM components with high aromatic content and large molecular weight in the watershed. Four fluorescent components (two humic-like and two protein-like substances) were identified using the PARAFAC model. Spatial distribution analysis showed that DOM quality was mainly influenced by human activities, and the proportion of protein-like substance (C3) was strongly correlated with anthropogenic parameters. The distribution of optical indices indicated diverse sources of DOM in the watershed. Ecological risk related to DOM was greater in the dry season than the wet season. There was a slight risk in most areas, with an extreme risk in areas experiencing the most intensive human disturbance, which were also extremely or heavily vulnerable. The results emphasize the strong influence of human disturbance on the ecological risk of DOM in peri-urban aquatic ecosystems. Our study provides useful information for ecological risk assessment of DOM that is difficult to obtain using traditional chemical analysis.

Application of parallel factor analysis model to decompose excitation-emission matrix fluorescence spectra for characterizing sources of water-soluble brown carbon in PM2.5

The applicability of parallel factor analysis (PARAFAC) model for identifying potential sources of water-soluble brown carbon (BrC) in fine particulate matter (PM2.5) using seasonal and annual excitation-emission matrix (EEM) fluorescence spectra data was investigated. The uncertainties related to the application of PARAFAC model to water-soluble BrC analysis were evaluated and the physicochemical meanings of PARAFAC-derived components were clearly interpreted. EEM spectra were obtained from water-soluble extractions of PM2.5 samples, which were collected at an urban and a suburban site in Chongqing, southwest of China during four one-month periods, each representing a different season in 2015. The measured EEM spectra were decomposed into three individual fluorescence components using PARAFAC algorithm, and the potential sources of BrC were identified based on the fingerprinting characteristics of PARAFAC-derived components. Each of the individual component exhibited similar spectral profiles in different seasons except in summer at the urban site; however, the relative intensities between the components varied with season, suggesting seasonal dependent source intensity of BrC. The relative contributions of the individual fluorescence components to the total fluorescence intensity varied largely from 0 to 89.2% at different excitation and emission wavelengths. Therefore, the relative abundance of each individual component based on the maximum fluorescence intensity (Fmax) should be used carefully for source apportionment analysis of BrC.