Bayesian statistical modeling of disinfection byproduct (DBP) bromine incorporation in the ICR database

Using Compliance Data to Understand Uncertainty in Drinking Water Lead Levels in Southwestern Pennsylvania

The historical use of lead in potable water plumbing systems has caused significant public health challenges. The Lead and Copper Rule requires utilities to take action if the 90^th percentile lead concentration exceeds the action level (AL) of 15 ppb. Assessment of the AL is based on a sample of homes representing a relatively small fraction of connections. Due to the intentional nonrepresentative sampling approach, the full set of conditions influencing lead concentrations in a large distribution system may be poorly characterized. Further, there is uncertainty in assessing statistical parameters such as the 90^th percentile concentration. This work demonstrates methods to compute the uncertainty in the 90^th percentile statistic and assesses the associated effect on compliance outcomes. The method is demonstrated on four utilities in southwest Pennsylvania (referred to as A, B, C, and D). For Utility A, evaluation of the 90^th percentile showed an increase over time in observed and estimated values and the value's uncertainty. This type of change in the uncertainty might have served as an early warning of the exceedance that followed. This could have triggered more timely review of operational changes in order to avoid the effects of noncompliance on utility costs and consumer confidence.

Risk associated with increasing bromide in drinking water sources in Yancheng City, China

The bromide concentration in water source (WS) of Yancheng City in China increased unexpectedly due to industrial discharge and saltwater intrusion, which leads to the formation of trihalomethane (THMs) in finished water of water treatment plants (WTP), especially brominated THMs. In Yancheng City, drinking water is supplied by WTP1 and WTP2, primarily sourced by WS1 and WS2, respectively. In this paper, the seasonal variations of bromide in WS1 and WS2 and THMs species in WTP1 and WTP2 were analyzed and compared. The effects of bromide in WS on THMs formation in finished water of WTP in terms of bromine substitution factor (BSF) were simulated by statistical linear model. Although the THMs concentrations in WTP1 were approximate to that in WTP2, the brominated THMs concentrations in WTP1 were higher than that in WTP2 due to higher bromide concentration in WS1 than WS2. The cancer risk analysis indicated that THMs' species of DBCM is the dominant THMs for WTP1 as well as WTP2, which can provide more information for WTPs with higher bromide concentration in water source.

Disinfection byproduct regulatory compliance surrogates and bromide-associated risk

Natural and anthropogenic factors can alter bromide concentrations in drinking water sources. Increasing source water bromide concentrations increases the formation and alters the speciation of disinfection byproducts (DBPs) formed during drinking water treatment. Brominated DBPs are more toxic than their chlorinated analogs, and thus have a greater impact on human health. However, DBPs are regulated based on the mass sum of DBPs within a given class (e.g., trihalomethanes and haloacetic acids), not based on species-specific risk or extent of bromine incorporation. The regulated surrogate measures are intended to protect against not only the species they directly represent, but also against unregulated DBPs that are not routinely measured. Surrogates that do not incorporate effects of increasing bromide may not adequately capture human health risk associated with drinking water when source water bromide is elevated. The present study analyzes trihalomethanes (THMs), measured as TTHM, with varying source water bromide concentrations, and assesses its correlation with brominated THM, TTHM risk and species-specific THM concentrations and associated risk. Alternative potential surrogates are evaluated to assess their ability to capture THM risk under different source water bromide concentration conditions. The results of the present study indicate that TTHM does not adequately capture risk of the regulated species when source water bromide concentrations are elevated, and thus would also likely be an inadequate surrogate for many unregulated brominated species. Alternative surrogate measures, including THM₃ and the bromodichloromethane concentration, are more robust surrogates for species-specific THM risk at varying source water bromide concentrations.

Kinetic removal of haloacetonitrile precursors by photo-based advanced oxidation processes (UV/H2O2, UV/O3, and UV/H2O2/O3)

The objective of the study is to evaluate the performance of conventional treatment process (i.e., coagulation, flocculation, sedimentation and sand filtration) on the removals of haloacetonitrile (HAN) precursors. In addition, the removals of HAN precursors by photo-based advanced oxidation processes (Photo-AOPs) (i.e., UV/H₂O₂, UV/O₃, and UV/H₂O₂/O₃) are investigated. The conventional treatment process was ineffective to remove HAN precursors. Among Photo-AOPs, the UV/H₂O₂/O₃ was the most effective process for removing HAN precursors, followed by UV/H₂O₂, and UV/O₃, respectively. For 20min contact time, the UV/H₂O₂/O₃, UV/H₂O₂, and UV/O₃ suppressed the HAN formations by 54, 42, and 27% reduction. Increasing ozone doses from 1 to 5 mgL^-1 in UV/O₃ systems slightly improved the removals of HAN precursors. Changes in pH (6-8) were unaffected most of processes (i.e., UV, UV/H₂O₂, and UV/H₂O₂/O₃), except for the UV/O₃ system that its efficiency was low in the weak acid condition. The pseudo first-order kinetic constant for removals of dichloroacetonitrile precursors (k'_DCANFP) by the UV/H₂O₂/O₃, UV/H₂O₂ and standalone UV systems were 1.4-2.8 orders magnitude higher than the UV/O₃ process. The kinetic degradation of dissolved organic nitrogen (DON) tended to be higher than the k'_DCANFP value. This study firstly differentiates the kinetic degradation between DON and HAN precursors.

Current and Potential Future Bromide Loads from Coal-Fired Power Plants in the Allegheny River Basin and Their Effects on Downstream Concentrations

The presence of bromide in rivers does not affect ecosystems or present a human health risk; however, elevated concentrations of bromide in drinking water sources can lead to difficulty meeting drinking water disinfection byproduct (DBP) regulations. Recent attention has focused on oil and gas wastewater and coal-fired power plant wet flue gas desulfurization (FGD) wastewater bromide discharges. Bromide can be added to coal to enhance mercury removal, and increased use of bromide at some power plants is expected. Evaluation of potential increases in bromide concentrations from bromide addition for mercury control is lacking. The present work utilizes bromide monitoring data in the Allegheny River and a mass-balance approach to elucidate bromide contributions from anthropogenic and natural sources under current and future scenarios. For the Allegheny River, the current bromide is associated approximately 49% with oil- and gas-produced water discharges and 33% with coal-fired power plants operating wet FGD, with 18% derived from natural sources during mean flow conditions in August. Median wet FGD bromide loads could increase 3-fold from 610 to 1900 kg/day if all plants implement bromide addition for mercury control. Median bromide concentrations in the lower Allegheny River in August would rise to 410, 200, and 180 μg/L under low-, mean-, and high-flow conditions, respectively, for the bromide-addition scenario.

Development of quantitative structure activity relationship (QSAR) model for disinfection byproduct (DBP) research: A review of methods and resources

Quantitative structure-activity relationship (QSAR) models are tools for linking chemical activities with molecular structures and compositions. Due to the concern about the proliferating number of disinfection byproducts (DBPs) in water and the associated financial and technical burden, researchers have recently begun to develop QSAR models to investigate the toxicity, formation, property, and removal of DBPs. However, there are no standard procedures or best practices regarding how to develop QSAR models, which potentially limit their wide acceptance. In order to facilitate more frequent use of QSAR models in future DBP research, this article reviews the processes required for QSAR model development, summarizes recent trends in QSAR-DBP studies, and shares some important resources for QSAR development (e.g., free databases and QSAR programs). The paper follows the four steps of QSAR model development, i.e., data collection, descriptor filtration, algorithm selection, and model validation; and finishes by highlighting several research needs. Because QSAR models may have an important role in progressing our understanding of DBP issues, it is hoped that this paper will encourage their future use for this application.

Predictive models for water sources with high susceptibility for bromine-containing disinfection by-product formation: implications for water treatment

This study examines a matrix of synthetic water samples designed to include conditions that favour brominated disinfection by-product (Br-DBP) formation, in order to provide predictive models suitable for high Br-DBP forming waters such as salinity-impacted waters. Br-DBPs are known to be more toxic than their chlorinated analogues, in general, and their formation may be favoured by routine water treatment practices such as coagulation/flocculation under specific conditions; therefore, circumstances surrounding their formation must be understood. The chosen factors were bromide concentration, mineral alkalinity, bromide to dissolved organic carbon (Br/DOC) ratio and Suwannee River natural organic matter concentration. The relationships between these parameters and DBP formation were evaluated by response surface modelling of data generated using a face-centred central composite experimental design. Predictive models for ten brominated and/or chlorinated DBPs are presented, as well as models for total trihalomethanes (tTHMs) and total dihaloacetonitriles (tDHANs), and bromide substitution factors for the THMs and DHANs classes. The relationships described revealed that increasing alkalinity and increasing Br/DOC ratio were associated with increasing bromination of THMs and DHANs, suggesting that DOC lowering treatment methods that do not also remove bromide such as enhanced coagulation may create optimal conditions for Br-DBP formation in waters in which bromide is present.

Enhanced coagulation with powdered activated carbon or MIEX secondary treatment: a comparison of disinfection by-product formation and precursor removal

The removal of both organic and inorganic disinfection by-product (DBP) precursors prior to disinfection is important in mitigating DBP formation, with halide removal being particularly important in salinity-impacted water sources. A matrix of waters of variable alkalinity, halide concentration and dissolved organic carbon (DOC) concentration were treated with enhanced coagulation (EC) followed by anion exchange (MIEX resin) or powdered activated carbon (PAC) and the subsequent disinfection by-product formation potentials (DBP-FPs) assessed and compared to DBP-FPs for untreated samples. Halide and DOC removal were also monitored for both treatment processes. Bromide and iodide adsorption by MIEX treatment ranged from 0 to 53% and 4-78%, respectively. As expected, EC and PAC treatments did not remove halides. DOC removal by EC/PAC was 70 ± 10%, while EC/MIEX enabled a DOC removal of 66 ± 12%. Despite the halide removals achieved by MIEX, increases in brominated disinfection by-product (Br-DBP) formation were observed relative to untreated samples, when favourable Br:DOC ratios were created by the treatment. However, the increases in formation were less than what was observed for the EC/PAC treated waters, which caused large increases in Br-DBP formation when high Br-DBP-forming water quality conditions occurred. The formation potential of fully chlorinated DBPs decreased after treatment in all cases.

Effects of pH on the speciation coefficients in models of bromide influence on the formation of trihalomethanes and haloacetic acids

This study investigated effects of pH, bromide and natural organic matter (NOM) level on yields and speciation of trihalomethanes (THMs) and haloacetic acids (HAAs) in chlorinated water. Experimental data were obtained using two water sources, one with a medium (DOC = 1.4 mg/L and SUVA = 2.60 L mg(-1) m(-1)) and the other with higher (DOC = 7.7 mg/L and SUVA = 4.26 L mg(-1) m(-1)) organic carbon level. The experiments employed the simulated distribution system (SDS) procedure at varying bromide concentrations and pH values of 7.0, 8.5 and 10. The speciation of THMs and dihalogenated HAAs (DHAAs) was interpreted based on the modelling of mixed halogenation yields via dimensionless ratios of bromination/chlorination reaction rates at each halogen incorporation node. The approach allowed precise modelling of the speciation of THMs and DHAAs at all examined pHs. In the case of DHAA, the dimensionless ratios of the bromination/chlorination reaction rates were not consistently affected by pH variations. For THMs, increase of pH caused the values of the dimensionless bromination/chlorination reaction rates to decrease in the case of halogenation of the initial reaction sites indicating a decreasing preference toward bromination at this reaction node. A similar trend was observed for the reactivity of dichlorinated reaction intermediate denoted as SCl2 whose formation precedes the release of CHCl3 and CHBrCl2. A similar but less consistent trend was observed for intermediate SBrCl whose halogenation yields both CHBrCl2 and CHBr2Cl. An opposite trend of increasing preference towards bromination at higher pHs was observed monobrominated intermediate SBr and in some extent dibrominated intermediate SBr2. These results help develop detailed DBP speciation models which needed to better understand the generation and potential health effects of THMs and HAAs at varying operating conditions and ultimately to adopt measure to minimize their levels in drinking water systems.

Modeling bromide effects on yields and speciation of dihaloacetonitriles formed in chlorinated drinking water

This study examined effects of bromide on yields and speciation of dihaloacetonitrile (DHAN) species that included dichloro-, bromochloro- and dibromoacetonitriles generated in chlorinated water. Experimental data obtained using two water sources, varying concentrations and characters of Natural Organic Matter (NOM), bromide concentrations, reaction times, chlorine doses, temperatures and pHs were interpreted using a semi-phenomenological model that assumed the presence of three kinetically distinct sites in NOM (denoted as sites S1, S2 and S3) and the occurrence of sequential incorporation of bromine and chlorine into them. One site was found to react very fast with the chlorine and bromine but its contribution in the DHAN generation was very low. The site with the highest contribution to the yield of DHAN (>70%) has the lowest reaction rates. The model introduced dimensionless coefficients (denoted as φ1(DHAN), φ2(DHAN) and φ3(DHAN)) applicable to the initial DHAN generation sites and their monochlorinated and monobrominated products, respectively. These parameters were used to quantify the kinetic preference to bromine incorporation over that of chlorine. Values of these coefficients optimized for DHAN formation were indicative of the strongly preferential incorporation of bromine into the engaged NOM sites. The same set of φ(i)(DHAN) coefficients could be used to model the speciation of DHAN released from their kinetically different precursors. The dimensionless speciation coefficients φ(i)(DHAN) were determined to be site specific and dependent on the NOM content and character as well as pH. The presented model of DHAN formation and speciation can help quantify in more detail the generation of DHAN and provide more insight necessary for further assessment of their potential health effects.

A Bayesian Multiple Imputation Method for Handling Longitudinal Pesticide Data with Values below the Limit of Detection

Environmental and biomedical research often produces data below the limit of detection (LOD), or left-censored data. Imputing explicit values for values < LOD in a multivariate setting, such as with longitudinal data, is difficult using a likelihood-based approach. A Bayesian multiple imputation (MI) method is introduced to handle left-censored multivariate data. A Gibbs sampler, which uses an iterative process, is employed to simulate the target multivariate distribution within a Bayesian framework. Following convergence, multiple plausible data sets are generated for analysis by standard statistical methods outside of a Bayesian framework. With explicit imputed values available variables can be analyzed as outcomes or predictors. We illustrate a practical application using longitudinal data from the Community Participatory Approach to Measuring Farmworker Pesticide Exposure (PACE3) study to evaluate the association between urinary acephate concentrations (indicating pesticide exposure) and self-reported potential pesticide poisoning symptoms. Additionally, a simulation study is used to evaluate the sampling property of the estimators for distributional parameters as well as regression coefficients estimated with the generalized estimating equation (GEE) approach. Results demonstrated that the Bayesian MI estimates performed well in most settings, and we recommend the use of this valid and feasible approach to analyze multivariate data with values < LOD.

Multi-level modelling of chlorination by-product presence in drinking water distribution systems for human exposure assessment purposes

During drinking water treatment and distribution, chlorine reacts with organic matter occurring in water to form various chlorination by-products (CBPs) such as trihalomethanes (THMs) and haloacetic acids (HAAs). This paper presents the occurrence of THMs and HAAs in different water distribution systems (DS) of the same region and their modelling for exposure assessment purposes. This study was conducted in eight DS supplying chlorinated water to the population of Québec City, Canada. These systems differ in type of water source (i.e. surface, ground or mixed water), in treatment applied at the plant, and in size and structure of the DS. Two spatio-temporal databases for THMs and HAAs were implemented, one for model development and the other for model validation. The analysis of the data demonstrates significant seasonal and spatial variations of these compounds. A multi-level statistical modelling approach was applied to estimate the ranges for occurrence of THMs and HAAs in the eight DS (i.e. a single model for the study region for each CBP species). The modelling approach integrates available or easily measurable parameters. For both THMs and HAAs, a two-level model considering a sampling-site random effect was selected among various models initially developed. The model capacity for estimating the presence of THMs and HAAs in drinking water and its usefulness for exposure assessment purposes in the studied region was demonstrated.