Priority screening on emerging contaminants in sediments of the Yangtze River, China

Abstract

Background

Screen the priority of emerging contaminants (ECs) from sediments is essential for risk assessment to aquatic environment and human health. Currently, priority approaches mainly focus on contaminant identification, exposure analysis, risk assessment, and hazard properties. However, there is still far from the reality due to, for instance, limitations on lack of occurrence data and uncertainty analysis. In this study, the multi-criteria screening method on the basis of hazard potential (HP) and exposure potential (EP) integrating with uncertainty analysis was developed for prioritization of 185 ECs, which have been reported to be widely found in the Yangtze River sediment. The HP based on the ecological risk and human health, and the EP according to the occurrence were both quantitatively analyzed. The priority index of these 185 chemicals was the product of the normalized HP and the normalized EP.

Results

According to the priority ranking scheme, 20 chemicals were identified as the top-priority, and 58 compounds as high-priority, respectively. After uncertainty scoring for each chemical based on data availability, there were 7 compounds (5 pesticides and 2 PFASs) recommended as the major priority ECs. In addition, the current study also emphasized that necessary for further studying some ECs, such as PFAS alternatives, as the data limitation may lead to reduce accurate prioritization.

Conclusions

Overall, this study provides an efficient approach for screening priority ECs, which is useful for river ecosystem health management.

Background

Emerging contaminants (ECs), such as plasticizers, antibiotics, per- and polyfluoroalkyl substances (PFASs), pesticides, and flame retardants, have caused growing concerns for both environments and human beings. The ECs usually consist all or part of the characteristics, which are great harm, hidden risk, environmental persistence, extensive sources, and complex management [1]. These chemicals are necessary and continually used in industry, agriculture and personal products. ECs access the sediment environment by diverse means, including agricultural production activities [2, 3], municipal waste [4], reclaimed wastewater irrigation [5, 6], and atmospheric deposition [7]. Due to continuous consumption, ECs have been frequently detected in various sediments at 10^–10 to 10^–6 g g⁻¹ [8,9,10,11,12]. Studies have reported that ECs can induce ecological risk and human health even at concentrations of 10^–9 g∙g⁻¹ [13,14,15]. Hence, ECs have received significant attention owing to their harmful effects on ecosystem and human health [16]. However, most ECs are not regulated in many countries around the world [17].

The Yangtze River is the third longest river in the world and its basin covers about one-fifth of China's total land area [18]. More than 40 percent of China’s population life in the watershed area [19]. In recent decades, a large number of treated sewage and industrial wastewater have been discharged into the Yangtze River (25 billion tons/year), accounting for 42 percent of the country's total sewage discharge [20, 21]. The surrounding soil and sediment of the river have been contaminated accordingly. Previous studies indicated that sediment may act as both sink and source of pollutants in aquatic systems [22]. Since water regime, for instance, the fluctuation of water level could cause pollutants such as ECs in sediment either return to the water body or suspend, both of which could bring potential risk to ecological environment and even human health. Therefore, understanding the potential risk of ECs in river ecosystem, in particular for sediments is essential for pollution control and ecosystem health maintain.

Based on the large amount of organic contaminants (about 10⁸ chemicals in 2014) in the environment [23], regulated and monitored chemicals, however, are only minor part of those massive amount of the chemicals present in environments [24]. Therefore, it is necessary to prioritize the compounds to ensure efforts on controlling and reducing potential threats. Currently, methods on identifying priority ECs from water bodies have been developed. For instance, the NORMAN network used a decision tree to divided ECs into 6 categories [25], such a decision tree-based approach is difficult to easy-to-implement. The Ministry of Ecology and Environment of China have published “list of key controlled emerging contaminants (2023 version)” [26] by combing semi-quantitative methods and specialist suggestions, leading to subjective influences on the final list [27]. Besides, methods relying on ecotoxicity or human health effects have been established to screen pollutants for priority controlling [28, 29]. At present, the prioritization systems typically include contaminant identification, exposure analysis, risk assessment, and hazard properties [30]. Whereas, these methods have not taken environmental occurrence into consideration [31], which may lead to loss of accuracy on prioritization. To improve prioritization strategies, ideas on ranking pollutants by multi-criteria analysis approach have been recommended [32]. For example, the EU Water Framework Directive (WFD) has developed a screening method relying on a weighted average of diverse effect scores of the exposure and effect index, such as hydrotoxicity, bioaccumulation and human health hazards [33]. By this, determination on the relative materiality of each criterion is possible, while the intrinsic connections among the criteria are ignored. The priority ranking of pollutants in China’s water bodies has been studied, in which estimating substance concentrations [34] by non-determinism analysis, and focusing on ecological risks [35] with narrowing scope of pollutant categories [36]. In fact, most of the existing priority screening of ECs are mainly aimed at water bodies, with rare studies focus on sediments and soils [37, 38].

Therefore, to address the limitations on ECs prioritization in sediments, this study developed the multi-criteria screening method that considered both exposure potential (EP) and hazard potential (HP) [30]. The evaluation parameters were persistence, bioaccumulation, ecotoxicity, human health effects, concentration and detection frequency of ECs. The relative materiality of each standard was judged by multiple linear regression model analysis, and the selected ECs were prioritized according to the results of priority analysis and uncertainty analysis. Finally, the ECs (plasticizers, PFASs, pesticides, flame retardants, antibiotics, PCBs) in the Yangtze River sediment were used for multi-class prioritization, which provides useful methodology for ECs management in large rivers worldwide.

Materials and methods

Data collection

Data were collected for six indicators of persistence, bioaccumulation, ecotoxicity, human health effects, concentration, and detection frequency. The detailed workflow for the ECs priority ranking is shown in Fig. 1.

Fig. 1

The flow of the multi-criteria approach for screening priority emerging contaminants (ECs). PC1 principal component 1, HP hazard potential, EP exposure potential, Groups I–IV top, high, moderate, and low priority group

Occurrence data

The occurrence data of ECs in sediments of the Yangtze River during 2013–2023 were obtained from government reports and/or publications by searching in Web of Science with the topic words “phthalates”, “Per- and polyfluoroalkyl substances”, “pesticides”, “flame retardants”, “antibiotics” or “PCBs” in combination with “Yangtze River”, “soil” or “sediment”. The topic words of the ECs belong to six categories, which have been widely detected in the Yangtze River basin, and their ecotoxicity and human health risk effects have been reported [39,40,41,42,43,44].

Data preprocessing efforts involved [30]: (1) the medians of all concentration and detection frequency, less affected by outliers, are used; (2) the means of concentration and detection frequency are used when median values were incalculable or not detected (ND); (3) employing half of the limit of detection (LOD) or half of the method detection limit (MDL) when ECs were not detected.

Persistence and bioaccumulation

The degradation half-life (DHL), which was determined by the BIOWIN v4.1 module [45], was used to indicate persistence. The sole criterion to judge the persistence time was the ultimate biodegradability, and the values for ultimate biodegradability of each chemical corresponded to the time as follows: five for hours; four for days; three for weeks; two for months; and one for longer period. The bioaccumulation was associated with octanol–water partition coefficient, using KOWWIN v4.1 [45] in the EPI component for calculation.

Ecotoxicity

Ecological environment effects were expressed using predicted no-effect concentrations of ecological (PNEC_eco), which was obtained by dividing the using the lowest lethal median concentration (LC₅₀) or half-effect concentration (EC₅₀) by the appropriate assessment factor (AF). The conversion of corresponding data to the PNEC value in the sediment was conducted using Eqs. (1) and (2):

$${\text{PNEC}}_{{{\text{sediment}}}} = f_{{{\text{oc}}}} \cdot k_{{{\text{oc}}}} \cdot {\text{PNEC}}_{{{\text{aqua}}}} ,$$

(1)

$${\text{PNEC}}_{{{\text{aqua}}}} = \frac{{{\text{EC}}50/{\text{LC}}50}}{{{\text{AF}}}}.$$

(2)

PNEC_aqua represents the PNEC value of water; f_oc is the weight fraction of organic carbon in sediment, f_oc = 0.1.

Depending on data availability, EC₅₀/LC₅₀ of the chemical was used in this study and AF was set to 1000. The EC₅₀/LC₅₀ values were gained from the US EPA ECOTOX knowledgebase [46], the US Department of Agriculture, Agriculture Research Service Pesticide Properties Database [47], published articles [11, 14, 48,49,50,51,52,53,54], and the US EPA Ecological Structure Activity Relationships (ECOSAR) model [45].

The RQ_eco value is calculated as following Eq. (3) [55]:

$${\text{RQ}}_{{{\text{eco}}}} = \frac{{{\text{MEC}}}}{{{\text{PNEC}}_{{{\text{eco}}}} }},$$

(3)

where RQ is the ecological risk value of EC compound; MEC is the actually measured EC concentration; and PNEC_eco is the predicted no-effect concentration.

Human health effects

The evaluation of human health effects considered adults exposure through rice and vegetables, utilizing PNEC_hum (ng g⁻¹). Based on Eq. (4), the PNEC_hum values were determined considering acceptable daily intake (ADI, mg·kg⁻¹·day⁻¹), minimal risk level (MRL) or reference dose (RfD):

$${\text{PNEC}}_{{{\text{hum}}}} = \frac{{1000 \times {\text{ADI}} \times {\text{BW}} \times {\text{AT}}}}{{\left( {{\text{IngRr}} + {\text{IngRv}}} \right) \times {\text{BCF}} \times {\text{EF}} \times {\text{ED}}}},$$

(4)

where 1000 was a conversion factor (ng·ug⁻¹); when ADI is unavailable, either RfD or MRL value would be utilized; the body weight (BW) of an average Chinese adult was set at 63 kg [56]. AT was the average exposure time for adults, setting at 10,500 d [14]. IngRr was the adult rice ingestion rate, setting at 279.2 g·d⁻¹, IngRv was the adult vegetable ingestion rate, setting at 92.3 g·d⁻¹ [14]. The bioconcentration factor for ECs in terrestrial organism was noted as BCF. In the current study, soil adsorption allocation coefficient (K_oc) was used to estimate BCF, deriving from Kenaga and Goring [57], ${\text{lgBCF}}=1.12{\text{lgKoc}}-1.58$, the Koc value is calculated using KOCWIN v4.1 [45] in the EPI component. EF denoted the exposure frequency, which was set at 350 d·a⁻¹, and ED was the exposure duration, setting at 30 years [30].

The ADI values of pesticides were prepared with reference to the national food safety standard (GB 2763–2021) [58], and that for antibiotics were prepared with reference to the national food safety standard (GB 31650–2019) [59]. The other ADI values of plasticizers and PCBs were derived from Zhong et al. [30] and Ossai and Sun et al. [60, 61], respectively. The MRLs of PFASs and flame retardants were acquired from the U.S. Department of Health and Human Services’ Agency for Toxic Substances and Disease Registry [62]. Some of the ADI/RfD/MRL values that were not available were replaced by the median values of the category chemicals.

The health risk assessment (RQ_hum) value of the ECs is calculated as following Eq. (5):

$${\text{RQ}}_{{{\text{hum}}}} = \frac{{{\text{MEC}}}}{{{\text{PNEC}}_{{{\text{hum}}}} }},$$

(5)

where RQ_hum is the health risk assessment value of each EC; MEC is the actually measured EC concentration; and PNEC_hum is the predicted no-effect concentration.

Prioritization

Normalization of criteria-specific data

Due to the fact that data sources were from literature, government report or publications, the min–max normalization method was approved to normalize data into dimensionless items in the range of 0–1 (Additional file 1: Table S1). The min–max normalization method, on one hand can maintain the distribution and relative size relationship of the original data, on the other hand, it has less impact on outliers and thus reduces its impact on the overall data. The method is derived from Kumar and Zhong [30, 32]. The respective orders of magnitude for environmental EC concentrations, PNEC_eco values, and PNEC_hum terms were 6, 8, and 12. To reduce the data discreteness and facilitate data calculation, the PNEC and concentration values were, respectively, converted to log_10- and log₂-due to the wide distributions. In order to supply a logical distribution of values for each particular criterion, the utility function carefully selected the highest and lowest values to ensure the dimensionless utility function terms would cover the range 0 to 1 across all ECs. These utility functions were utilized chemical scores.

Multivariate analysis

Principal component analysis (PCA) is used in this research, in which the largest variation is captured by the PC1, which can be applied as a new cumulative variable for ECs screening and sorting [63]. The degradability of volatile organic pollutants has been explained using the term PC1 in prior research [64], the persistence, bioaccumulation, and toxicity (PBT) characteristics of contaminants [65], and comprehensive aquatic toxicity at different nutrient levels [66, 67]. The current study used PCA to analyze the standardized data related to four hazard indicators: persistence, bioaccumulation, ecotoxicity, and human health effects. The PC1_hazard was determined as a HP value which the four hazard impacts were evaluated comprehensively. Likewise, standardized criteria-specific data for two exposure factors (pollutant concentration and detection frequency) were analyzed using PCA. The EP was illustrated by utilizing the PC1_exposure.

Scoring

The utility function was used to convert the EP and HP into dimensionless terms within the range of 0–1. By performing the multiplication of the normalized EP and normalized HP, the priority index for pollutant classification was determined. The dimensionless EP, HP value and priority index were calculated according to the following utility functions (6), (7) and (8):

$$U\left( {{\text{EP}}} \right) = \frac{{{\text{EP}} - {\text{EP}}_{{{\text{min}}}} }}{{{\text{EP}}_{{{\text{max}}}} - {\text{EP}}_{{{\text{min}}}} }},$$

(6)

$$U\left( {{\text{HP}}} \right) = \frac{{{\text{HP}} - {\text{HP}}_{{{\text{min}}}} }}{{{\text{HP}}_{{{\text{max}}}} - {\text{HP}}_{{{\text{min}}}} }},$$

(7)

$${\text{Priority}}\;{\text{index}} = U\left( {{\text{EP}}} \right) \times U\left( {{\text{HP}}} \right),$$

(8)

where EP_max is the maximum EP in the overall list of candidate ECs, EP_min is the lowest EP value in the overall list of candidate ECs; HP_max is the maximum HP in the overall list of candidate ECs, HP_min is the lowest HP value in the overall list of candidate ECs; the priority index determines the priority ranking of chemicals.

In this study, a multiple linear regression model was used to quantitatively represent the relationship between concentration, detection frequency and EP. Similarly, multiple linear regression model was used to quantify the relationship between persistence, bioaccumulation, ecotoxicity, human health effects and HP. Standardized values are used for the above parameters.

Uncertainty analysis

Regarding uncertainty analysis, the uncertainty score should be derived from the accessibility of monitoring data (Table 1). The occurrence data of this study were obtained from previous studies, in which when the data of occurrence were obtained from a minimum of 4 provinces and 50 sites, the uncertainty scores of pollutant concentrations and detection frequencies were 0. If data of occurrence were from fewer than 4 provinces or less than fifty sites, the uncertainty scores were set at 0.25. In case of occurrence data were not available, the uncertainty scores were established as 0.5. The lack of experimental data, the model-based toxicity, ADI values, and BCF values provided a significant level of uncertainty regarding ECs. The ecotoxicity and human health effects uncertainty scores were 0 when experimental data were used for PNEC calculations, yet they increased to 0.25 after incorporating model-based evaluations, and further rose to 0.5 without any experimental or evaluated data. As all the data of human health effects were from model calculation, all chemicals had an uncertainty score basis of 0.25 in the human health effect criteria. About the remaining criteria, chemical data availability and unavailable determined the assignment of uncertainty scores as either 0 or 0.5. Finally, the aggregate uncertainty scores were decided by using the arithmetic mean of the each uncertainty scores of the 6 criteria.

Table 1 Split-point values assigned to uncertainty categories I–IV for ECs in the Yangtze River sediment

Results

Concentration and detection frequency of ECs

Overall, a total of 185 ECs including priority controlled chemicals in China (Additional file 1: Table S2), were selected from 2399 sites in 10 provinces or municipalities (Additional file 1: Figure S1). The median concentration of these 185 compounds ranged from 5 × 10^–5 ng g⁻¹ to 835 ng g⁻¹, with 4 compounds (DBP, DEHP, TBOEP, 6:2 FTOH) exhibiting median concentration (> 100 ng g⁻¹). The plasticizers displayed a higher median concentration (1.8 ng g⁻¹), followed by pesticides (0.89 ng g⁻¹), and PCBs (0.71 ng g⁻¹). The flame retardants and PFASs exhibited low median concentration, with the mean values up to 0.49 and 0.25 ng g⁻¹, respectively. Antibiotics had the lowest median concentration which was below 0.1 ng g⁻¹. The average ECs detection frequency was greater than 35% for all categories except antibiotics (21.9%) (Fig. 2).

Fig. 2

The log₂concentration and detection frequency of the 185 compounds (the names of these chemicals are listed in Additional file 1: Table S2)

Persistence, bioaccumulation, ecotoxicity and human health

The 4 criteria values of 185 ECs can be seen in Additional file 1: Table S4. The DHL values representing persistence for the 185 selected compounds ranged from − 2.36 to 3.66, and their LogK_ow values ranged from − 3.21 to 12.11. The PNEC_eco value ranged from 0.01 to 1.08 × 10⁵ ng g⁻¹, while the PNEC_hum value were between 1.36 × 10^–7 and 3.29 × 10⁶ ng g⁻¹. All data are presented in Fig. 3. Overall, plasticizers revealed comparatively high DHL with a median value of 3.11, followed by antibiotics (1.99), flame retardants (1.83), pesticides (1.74), PCBs (1.44) and PFASs (0.74). PCBs displayed higher LogK_ow with a median value of 6.98, followed by flame retardants (5.88). The other categories observed a sequential drop in their median LogK_ow as follows: plasticizers (4.61), pesticides (4.56), PFASs (4.46) and antibiotics (0.43). PCBs and pesticides showed relatively low PNEC_eco with the median values of 60.18 ng g⁻¹ and 61.00 ng g⁻¹ followed by flame retardants (159.16 ng g⁻¹) and plasticizers (234.31 ng g⁻¹). The order of increased median PNEC_eco of other categories were as follows: antibiotics (670.35 ng g⁻¹), PFASs (1297.53 ng g⁻¹). The PCBs showed relatively low PNEC_hum with median values of 0.22 ng g⁻¹, followed by flame retardants (599.05 ng g⁻¹). The progression of increased median PNEC_eco for the remaining categories unfolded in the sequence: pesticides (2027.60 ng g⁻¹), PFASs (3614.23 ng g⁻¹), plasticizers (9.8 × 10⁵ ng g⁻¹) and antibiotics (1.27 × 10⁶ ng g⁻¹).

Fig. 3

The persistence, bioaccumulation, logPNEC_eco and logPNEC_hum of plasticizers, pesticides, PFASs, flame retardants, antibiotics and PCBs

Hazard and exposure assessment

In the current study, a PCA analysis was conducted on these 185 compounds. As shown in Fig. 4, PC1_hazard played a major part with an explanation rate of 54.9%, PC2_hazard constituted 25.0% and explaining the further hazard parameters. For instance, compounds with greater ecotoxicity and human health effects were located at the top right of the PCA score chart, yet ECs with greater persistence and bioaccumulation were situated in the bottom right of the PCA score chart (Fig. 4a). The 4 hazard criteria completely enhanced with PC1_hazard, indicating prospective tendencies for the HP parameter. Ranging from − 2.65 to 4.37, the HP values of the 185 chemicals are listed in Additional file 1: Table S5. On the whole, PCBs displayed higher median HP value (1.61), followed by flame retardants (0.58) and PFASs (0.33). The median HP value of pesticides (0.20) located in the middle. The plasticizers (− 1.12) and antibiotics (− 1.55) showed relatively low HP. The relative importance of the 4 hazard parameters to the HP was quantitatively analyzed using a multiple linear regression analysis, and the proportions of DHL, LogK_ow, PNEC_eco and PNEC_hum were 2.674, 3.086, 1.292 and 3.109, respectively.

Fig. 4

PCA for the four hazard parameters (a) (PC1_hazard denoted the integrated HP) and PCA of the 2 exposed effect parameters (b) (PC1_exposure denoted the integrated EP)

The PCA analysis results of the exposure parameters for the 185 ECs could be obtained from Fig. 4b and Additional file 1: Table S6. The explanation of the total variance was divided between PC1_exposure (72%) and PC2 _exposure (27%). The range of EP values of the 185 ECs varied between − 3.17 to 3.28. Actually, the EP of plasticizers stood at 1.33, marking the highest median value, followed by PFASs at 0.66. The medium median EP of the categories were pesticides (0.20) and flame retardants (0.20). The lower median EP of the categories were PCBs (− 0.03) and antibiotics (− 1.11). The relationship between EP and the 2 exposure parameters was investigated through multiple linear regression analysis, and the proportions of concentration and frequency were 4.856 and 1.96, respectively.

Priority index

The listing of the priority indices for the concerning ECs is displayed in Additional file 1: Table S7. The 185 concerning ECs were divided into 4 groups based on the priority index distribution (Fig. 5a): Group I (consisting of 20 chemicals in top priority), Group II (comprising 58 chemicals in high priority), Group III (including a total of 69 chemicals in moderate priority) and Group IV (containing 38 chemicals in low priority). Due to the variable threshold size, the distribution of priority index was done on a relative scale. Based on this, the prioritization of the future ecological environment monitoring and regulation should focus on the chemicals in Group I and Group II. Compounds in Group III should also be included when additional data on toxicity and risk assessment are sufficient. Each priority ranking group in Fig. 5b showed the numbers and category of compounds. PFASs, PCBs, flame retardants, pesticides, and plasticizers contributed for 50%, 25%, 15%, 5%, and 5%, respectively, among the compounds in Group I. With Group II, PCBs, pesticides, flame retardants, PFASs and plasticizers contributed for 37%, 26%, 19%, 16% and 2%, respectively. Therefore, PFASs, PCBs, flame retardants, and pesticides were classified as top-priority category, which were chosen for future priority environmental monitoring and pollutant treatment researches with their contamination in Group I and II representing serious human health risks and ecotoxicity.

Fig. 5

Priority index and rank of the concerning ECs (a) and the numbers of compounds of each category in each ranking group (b). The respective priority index split-points for Group I, II, and III were set at 0.35, 0.21, and 0.06

Flame retardants (i.e., BDE-209, TCPP, and BDE-154), PFASs (i.e., PFTeDA, PFTrDA, PFDoDA, 8:2Cl-PFAES, PFOA, 6:2FTOH, PFNA, PFUnDA, 6:2Cl-PFAES, PFHxDA), PCBs (i.e., PCB206, PCB153, PCB28, PCB52, PCB138), pesticides (i.e., p,p'-DDT), plasticizers (i.e., DNP) were the top-priority chemicals in Group I. Specific category ranking displayed of the 185 concerning ECs were manufactured by priority indices in Additional file 1: Table S8. The list of chemicals in each category belonging to top-priority Group was as follows: dinonyl phthalate (plasticizer), perfluorotetradecanoic acid (PFAS), dichlorodiphenyltrichloroethane (pesticide), decabromodiphenyl ether (flame retardant), styrene (antibiotic), nonachlorobiphenyl (PCB).

Comparison between the ranking approaches

The ranking of the top 78 ECs in the 5 different prioritization schemes (RQ_hum, RQ_eco, EP, HP, priority index) is shown in Table 2. All the ranking lists are accessible through Additional file 1: Table S9. The Pearson’s correlation (Fig. 6) varied between − 0.084 (RQ_eco and HP) to 0.778 (HP and priority index) for these 5 ranking approaches, and the priority index correlated greatly with HP and EP (0.511).

Table 2 Top 78 priority ECs according to the 5 disparate prioritization schemes, namely EP, HP, RQ_eco, RQ_hum and priority index

Fig. 6

Correlation among the five ranking methods (EP, HP, RQ_eco, RQ_hum, and Priority index)

Four prioritization strategies selected 6:2 FTOH, TCPP, PCB61, PFOA, p,p'-DDT, p,p'-DDE, PCB28, PFTrDA, PFTeDA, TEHP, PCB52, HCB, BDE-154, PCB153, BDE-99, PCB138, cypermethrin in the top 78 chemicals. Three prioritization schemes included DBP, TDCIPP, 8:2 FTOH, BDE-209, 6:2Cl-PFAES, DNP, chlorpyrifos, 6:2 FTS, PFNA, PFDoDA, 8:2Cl-PFAES, PFUnDA, 8:2 FTS, p,p'-DDD, PCB66, PCB18, disulfoton, parathion, thionazin, BDE-28, phorate, PCB180, PCB206, PCB126, PCB189, PCB157, PCB156, PCB167, PCB169, BDE-138, BDE-100, PCB118, PCB123, PCB114, PCB105, PCB101, PCB81, PCB77, BDE-85, cyhalothrin, o,p'-DDT, deltamethrin, PCB209, ROX, aldrin, BDE-71 in their ranking of the top 78 chemicals. According to the 5 different prioritization schemes, the ranking on the top of 78 ECs consist of 14 plasticizers, 28 PFASs, 30 pesticides, 24 flame retardants, 13 antibiotics and 27 PCBs, amounting from 29.55% (antibiotics) up to 100% (PCBs) of the selected chemicals in each types. Hence, different prioritization scheme can result in different ranking patterns of chemicals.

Uncertainty score

The whole uncertainty scores of concerning ECs are displayed in Additional file 1: Table S10. On the basis of the origin of uncertainty, compounds were divided into 4 categories: Category I (26 chemicals with plenty ecotoxicity and occurrence data); Category II (65 chemicals with incomplete toxicity data); Category III (36 chemicals with incomplete human health effect and occurrence data); Category IV (58 chemicals with insufficient data on both occurrence and toxicity).

Antibiotics were distributed across all the 4 uncertainty categories, accounting for 57.7%, 23%, 16.7%, and 13.8% of Category I to IV, respectively. Flame retardants are mainly distributed in categories II and III, accounting for 17% and 33.3% of Category II and III, respectively. PCBs were generally included in Category II, accounting for 26.2%. Most pesticides were identified into Category III, accounting for 50%. PFASs and plasticizers were representative greatly in Category IV, accounting for 39.7% and 13.8%, respectively. The PFASs in Category IV were mainly substitutes for PFASs. In general, 31.35% of the ECs were included into Category IV. (Fig. 7).

Fig. 7

The quantity (a) and percentum (b) of compounds through each EC category in the uncertainty category

Results for priority control

Combining the results of 4 priority groups and 4 uncertainty categories, the selected 185 ECs were separated to 16 subgroups for comprehensive ranking. These chemicals were ranked using a priority index in each uncertainty category, and the results are available from Additional file 1: Table S11. The uncertainty categories I, II, III, and IV exhibited respective amounts of 26, 66, 36, and 57 chemicals. Finally, 7 chemicals were selected as the final priority ECs, which are p,p'-DDT, PFOA, p,p'-DDE, p,p'-DDT, p,p'-DDD, PFOS, and α-HCH. The detailed uncertainty categories of chemicals showed in the priority groups can be found in Table 3.

Table 3 The detailed uncertainty categories of chemicals showed in priority groups

Discussion

Occurrence, fate and bioaccumulation

Plasticizers (PAEs) could be degraded by light [68], therefore, the highest median concentration might be related to the widespread detection campaigns in the Yangtze River Basin [44]. The high intensity of agricultural activities in the Yangtze River Basin resulting in pesticides detectable in high concentrations [69, 70]. Previous studies mainly focused on long-chain PFASs (PFOA and PFOS), while their alternatives with short-chain structures were rarely studied, hence the lowest median concentration of PFASs were found [42]. It is also possible that PFASs in surface water mainly come from Waste Water Treatment Plant [71], and the concentration of PFASs would be reduced after treatment. The ECs categories of the current study were widely distributed in environment [72], having high detection frequency. The median DHL value of plasticizers was 3.11, indicating that the duration of completely mineralization last for weeks. While the median DHL values of pesticides, flame retardants, PCBs and antibiotics were closed to 2, revealing the complete mineralization lasted for months, and that for PFASs (DHL = 0.74) were for many years. Overall, PFASs, have been categorized as persistent organic pollutants, showing the strongest environmental persistence [73]. Comparing the median LogK_ow values of these six categories, it was found that the value of PCBs (6.98) was the highest. Previously study had shown that the enrichment factor of PCBs was higher [74]. Due to the fact that chemical properties of each compounds varied significantly, the priority ranking based on multiple criteria is of importance. The PNEC_eco values and PNEC_hum values of PCBs were in the lowest level, indicating that the toxicity of PCBs were the strongest ones.

The potential effects on exposure and hazard

As could been seen from Fig. 4a, PCBs were generally situated at the higher PC1_hazard values, representing the greatest integrated hazard and hence should be considered to be the most concerning EC category. Pesticides and PFASs were situated at the mid-range PC1_hazard values. Nevertheless, pesticides were primarily composed of positive PC2_hazard values, in contrast to PFASs which had PC2_hazard values were primarily negative. These results demonstrated higher ecotoxicity and human health effect for pesticides, while higher persistence and bioaccumulation potential for PFASs, as compared to the other selected ECs. Antibiotics were situated at the lower PC1_hazard values, indicating the minimal hazard influences. However, antibiotics might still represent primary concerns. Flame retardants and plasticizers were widely scaled, nearly one half were in positive PC1_hazard which indicated human health effect and bioaccumulation potential. The study [30] also showed that PFAS and pesticide had the higher HP values, and plasticizer and antibiotic had the lower HP values; whereas, the higher HP value of flame retardants in the current study was in contrast, which might be due to the PNEC_hum derived from simulation and existed differences with experimental data.

As PC1_exposure raised, there was a corresponding rise in the cumulative exposure potential. ECs categories included a broad variety of characteristics, for instance, PFASs, pesticides, and flame retardants were widely dispersed across the PC1_exposure range, suggesting that their properties varied widely in the environment. Antibiotics and PCBs were principally situated at the lower PC1_exposure values, manifesting most of them were absence of detection in the actual monitoring or the existing methods were insufficient to detect. Plasticizers were mainly situated at higher PC1_exposure position which was related to higher concentration and detection. A study also showed that plasticizers and flame retardants had the higher EP values, and antibiotic had a lower EP value [75]. The EP data for PFASs and pesticides in this research were in contrast, which might be related to industry and agriculture distribution differed in the Yangtze River basin.

Priority index

Due to the fact that 50% compounds in Group I were PFASs, indicating that PFASs might be the most hazardous. The DNP, p,p'-DDT and PCB138 were also listed at the top level in a previously study [27]. Evidence has shown that the p,p′-DDT is harmful with health risk to human [76]. For other chemicals in Group I such as BDE-209, special focus should be given because its high concentration and detection frequency in environment, causing endocrine disorders, hepatotoxicity and cardiovascular toxicity [77, 78]. BDE-154 was detected in various environment and showed high bioaccumulation [79]. Nowadays, the global concern for PFASs has gained widely recognition [75]. Compared with the study [30], it is found that PFOA was also listed in the Group I. Simultaneously, PFOA is a group of chemicals that used for industrial production with strong persistence, bioaccumulation, and toxic effects. In 2013, the International Agency for Research on Cancer (IARC) included PCBs as Group 1 carcinogens for humans [80]. BDE-209, PCB206, PFOA, PCB153, PCB28, PCB52, and PCB138 are included into Stockholm Convention on Persistent Organic Pollutants [81] and China’s List of Key Emerging Contaminants under Control (2023) [82]. BDE-154 is contained in China’s List of Key Emerging Contaminants under Control (2023) and p,p'-DDT is involved into EU POPs Control List [83]. Insides, the EU is considering PFNA, PFUnDA, PFDoDA, and PFTrDA as potential candidates for the control of POPs under the Stockholm Convention [81]. Importantly, TCPP, PFTeDA, 8:2 Cl-PFAES, 6:2 FTOH, 6:2 Cl-PFAES, PFHxDA, DNP, which were not contained in any list so far, were identified as the top-priority ECs in our study. Dinonyl phthalate, perfluorotetradecanoic acid, dichlorodiphenyltrichloroethane, decabromodiphenyl ether, styrene, and nonachlorobiphenyl, which are the compounds ranked first in each category, may be classified as priority pollutants for the future environmental monitoring and estimate of sediment treatment process. Therefore, the current approach provide new insights to priority the ECs.

Methodology feasibility and uncertainty analysis

The HP and EP showed a strong correlation with the priority index (0.778 and 0.511), implying that the advantage of our approach considered effects of ECs to both human health and ecological risks. The identification of 78 ECs in Groups I or II based on priority index highlighted the availability for explaining both occurrence and toxicity factors.

The distribution of antibiotics in uncertainty categories showed that there was a lack of research on certain classes of antibiotics. The main distribution of flame retardants in categories II and III indicated a lack of simultaneous monitoring campaigns and toxicity evaluation of the same chemical. The general distribution of PCBs in Category II suggested data gaps related to ecotoxicity. Most pesticides were identified into Category III, indicating the necessary for strengthen monitoring programs in the Yangtze River sediment. PFASs and plasticizers were representative greatly in Category IV, suggesting the urgent need to strengthen monitoring campaigns and hazard assessments. The PFASs in Category IV were mainly substitutes for PFASs, indicating that there was a major need to strengthen research on alternatives to PFASs. The greater mobility and equivalent persistence of short-chain PFASs, as opposed to legacy PFASs, resulting in stronger long-range transport availability [84]. The plasticizers in Category IV were not included in the priority control contaminants [85], and the DNP was in the priority Group I, thus further studies is needed to narrow this gap. In general, there are many ECs were included into Category IV, which indicated principal study gaps for occurrence and toxicity of ECs in the Yangtze River sediment.

Recommendations for priority control

Rigorous hazard assessments of the 65 compounds in uncertainty Category II are recommended, while the 65 compounds in uncertainty Category III require an intensive monitoring. Simultaneously, additional monitoring activities and hazard assessments for the 58 compounds in uncertainty Category IV are recommended.

About the 7 priority ECs shown in priority group I/uncertainty Category I and priority group II/uncertainty Category I, the chemical number only accounting for 8.9% without uncertainty analysis, routine environmental monitoring, setting relevant emission standards, and establishment of control measures are suggested. Similarly, especial concern should also be given to the 19 ECs in the priority group III/uncertainty Category I and the priority group IV/uncertainty Category I. Chemicals of remaining uncertainty categories should be the candidates for subsequent environmental monitoring and toxicity tests. When new occurrence and toxicity data become available, these ECs should be included for re-evaluation of the priority control list.

Conclusion

In the current study, a multi-criteria analysis approach on the basis of HP and EP was developed to rank the priority of 185 selected ECs in the Yangtze River sediment, belonging to plasticizers, PFASs, pesticides, flame retardants, antibiotics and PCBs. Of which, an integrated priority index of concerning chemicals was computed by combining their hazard index and exposure index. The results showed that PCBs, flame retardants and PFASs exhibited higher HP values, however, plasticizers and antibiotics were with low HP values. Additionally, the plasticizers and PFASs showed relatively high EP values, while that for PCBs and antibiotics were low, as revealed by exposure analysis. The priority index listed 20 chemicals that were top-priority and 58 chemicals as high-priority, in which PFASs amounted the highest proportion at both top-priority and high-priority groups. The PCBs, pesticides, and flame retardants were within the high-priority groups. After uncertainty analysis and categorization, a total of 7 ECs were recognized as priority chemicals and suggested to be controlled as the first target. Hence, the study highlights the necessary to provide priority screening on emerging contaminants in different regions. Besides, further studies are needed for the alternatives of PFASs and plasticizers, which could overcome data limitation and thus optimize the approach.