Journal of Environmental Engineering

This study develops and tests a method for multiobjective optimization of water-quality monitoring networks in river-reservoir systems. The optimization method identifies optimal sampling locations to detect the sudden release of contaminants to a reservoir, and meets two objectives: (1) minimizing the prediction error of methyl tert-butyl ether (MTBE) at the reservoir’s outlet valve; and (2) minimizing the average time during which MTBE is detected at sampling locations. The optimization method considers 36 contaminant scenarios, corresponding to three volumes of contaminant release, three release locations, and four different seasonal release times. The MTBE pollutant chemograph was simulated at the outlet valve of the reservoir and at 16 possible sampling locations with the CE-QUAL-W2 model for each of the 36 scenarios of contaminant release. A support vector regression (SVR) tool is coupled to the nondominated sorting genetic algorithm II (NSGAII) to optimize the water-quality sampling locations. Implementation of the NSGAII-SVR method demonstrates its capacity to design water-quality monitoring networks that meet multiple objectives in a river-reservoir system.

Chlorine decay modeling in drinking water transport and distribution systems is a major tool for the management of disinfectant concentration. For such purposes, the models to be used in water quality simulators must describe chlorine decay in the systems as accurately as possible, requiring a minimum number of parameters and still be sufficiently robust to predict chlorine residuals under changing operational conditions. In this paper, two second-order decay models (single reactant and two reactants) were examined for their ability to predict chlorine residuals in surface raw and treated waters at a range of initial chlorine and natural organic matter concentrations. The two reactants model is necessary for chlorine residuals prediction in raw waters. However, for treated waters it only performed marginally better than the simpler single reactant model, with half the number of parameters that needed to be estimated. In addition, correlations were observed between both models parameters for reactive natural organic matter (NOM) concentration and measured fluorescence and ultraviolet (UV) spectroscopic data associated with humic-like compounds. Such findings are likely to open the door for a more exact determination of the models’ parameters and, thus, for improving the chlorine decay models’ adequacy and accuracy.

Investigations were conducted of oxygen transfer by multiple jets in tandem configuration discharging into water flowing in a rectangular channel based on the volumetric oxygen transfer coefficient, $K_L{a}_{20}$. Results from this study show that water depth in the receiving channel, number of jets, fall height, and jet velocity influence oxygen transfer. Consistent with previous research, an optimum depth was observed corresponding to maximum oxygen transfer, which deviated from estimations using the equation for hydraulic structures due to the low velocities of water in the channel. Since increasing the number of jets increases $K_L{a}_{20}$, but not essentially standard oxygen transfer efficiency (SOTE) for the same jet power per unit volume ($P/V$), the authors introduced the concept of an optimum $P/V$. This optimum was constant for any number of jets investigated, and at this value, both $K_L{a}_{20}$ and SOTE increase with increasing number of jets. Ultimately, a relationship is proposed for estimating $K_L{a}_{20}$ for multiple plunging jets in tandem with modifications for practical application to spatially varied flow in gutters.

Building on previous batch studies, this study further investigated the removal of dissolved organic $P$ (DOP) using small (15 cm) columns with water treatment residual (WTR) and alum as two promising amendments to bioretention soil media (BSM). BSM modified with alum and WTR exhibited higher sorption capacity. Compared with unmodified BSM, significant improvements were observed for adenosine monophosphate (AMP), myo-inositol hexakisphosphate (IHP), and inorganic $P$ (SRP) removal using 10% alum-modified WTR-amended BSM (WTR-BSM), by factors of 26, 7, and 150, respectively. The 10% alum-modified WTR-BSM reduced DOP $\mathrm{levels}<0.01\mathrm{mg}/\mathrm{L}$ within 1 min of contact time, demonstrating more-rapid DOP sorption processes compared with unamended media. The oxalate ratio (OR), which is based on oxalate-extractable Al, Fe, and P, was employed to provide a criterion for BSM performance with respect to $P$ sorption. Strong correlations ($R^2=0.92$, 0.96, and 0.65 for AMP, IHP, and SRP, respectively) between the media OR and $P$ sorption capacity of media was observed, which supports the use of the OR to predict $P$ sorption performance of BSM.

Particulate matter, nutrients, and gross solids from roads contribute nonpoint-source pollution to waterways. To inform road runoff management, a field monitoring study was undertaken at eight road sites across North Carolina. Particle size distributions (PSD) of edge-of-pavement runoff samples were analyzed to understand the granulometry of the particulate matter. Knowledge of PSDs is critical when attempting to understand sediment fate and transport through stormwater control measures (SCMs), especially when the primary removal mechanism is particle settling. For 43 road runoff events, median particle size varied from 31 to 144 μm. The median PSD from hot mix asphalt (HMA) was 2.6% clay, 44.8% silt, and 52.6% sand. PSD was not correlated to roadway classification or ecoregion; however, PSD was significantly correlated to the presence of a permeable friction course (PFC) overlay, which is a layer of porous asphalt placed over traditional HMA. The median $d_{90}$ for PFC (131 μm) was significantly smaller than for HMA (428 μm), with 11.5% less sand fraction present in the PFC runoff. Smaller particles emitted from PSD will diminish performance for downslope highway SCMs, including swales and filter strips. PFC also emitted lower median effluent total suspended solids (TSS) concentrations ($9\mathrm{mg}/\mathrm{L}$) than HMA ($23\mathrm{mg}/\mathrm{L}$). Gross solids may represent a substantial portion of the total nutrient load from roads, but they are often entirely overlooked. Dry mass of leaf litter from roads in North Carolina was correlated to rainfall depth and peak rainfall intensity. Total nitrogen (TN) loads from gross solids varied from 0.2 to $0.9\mathrm{kg}/\mathrm{ha}/\mathrm{year}$, while total phosphorus (TP) loads ranged between 0.03 and $0.12\mathrm{kg}/\mathrm{ha}/\mathrm{year}$, approximately one order of magnitude less than reported TN and TP nutrient loads in stormwater runoff. Given the typical fraction of TN and TP leaching into stormwater as leaves and organic matter break down, gross solids represented about 0.6% of TN and 3.6% of TP yearly nutrient loading for the evaluated highway sites, suggesting gross solids were a relatively inconsequential portion of the total nutrient load from roads.

Two of the most common and simple stormwater control measures are swales and filter strips. However, an overly simplistic one-size-fits-all design approach typically is used for these practices. To provide more flexibility in design, a coupled hydraulics and particle-settling model was created to predict swale and filter strip total suspended solids (TSS) reduction as a function of catchment area, longitudinal slope, side slope, cross section type (triangular swale, trapezoidal swale, or filter strip), and length. The hydraulics and hydrology models were based on Manning’s equation and the rational method, respectively, with the underlying requirement that the water quality design storm does not exceed the height of the grass. The particle-settling model was underpinned by the Aberdeen equation. The model predicts that triangular swales produce the least and filter strips the most TSS removal because of increased hydraulic retention time; trapezoidal swales had on average 10% greater TSS removal than triangular swales with similar design characteristics. Grass filter stormwater controls performed better with decreasing slope, increasing length, smaller catchment area, and shallower side slopes. Filter strips registered the majority of their TSS reduction within the first meter of their width and were insensitive to increases in longitudinal slope. This water quality design storm-abating modeling approach could be coupled with modeling for infrequent-return-interval storms to obtain dual benefits of sediment reduction and conveyance. This water quality model may provide regulatory agencies a tool to provide variable TSS credit to swales and filter strips as a function of design.

In this work, laboratory experiments and numerical simulations were used to investigate the transport of solute through stratified porous media. Experiments were performed using nonreactive solute transport through an experimental model of stratified porous media of different lengths. The aim of conducting the experiments was to evaluate the effect of the length of mobile–immobile domain on estimation of first-order mass transfer coefficient. The velocity was also changed for different lengths of stratification for understanding the behavior of mass transfer coefficient. The estimated values of mass transfer coefficient were analyzed for interdependency with length, velocity, and dispersion coefficient. It was found that mass transfer coefficient has an inverse relation with dispersion coefficient and also varies with the length of the mobile–immobile region and velocity. Finally, the variable behavior of mass transfer coefficient is obtained when the transport media are under the influence of physical nonequilibrium.

This paper presents the environmental and health threats caused by petroleum products as well as the results of remediation work on five air bases where biological cleaning methods developed by the authors were implemented. The aviation fuel leakages endangered the health of people and surface water streams at the sites as well as in distant locations. The aim of the research was to verify remediation techniques developed in laboratory tests. The initial concentration of total petroleum products in the soils reached $\mathrm{2,320}–\mathrm{8,770}\mathrm{mg}/\mathrm{kg}$ of dry matter, which presented an extremely high risk to the population and the environment. The autochthonous bacterial strains, which included Pseudomonas fluorescens, Acinetobacter calcoaceticus, Pseudomonas putida I, and Pseudomonas putida II, were proven to be very effective in degrading the xenobiotics. The applied microorganisms in conjunction with the authors’ remediation techniques produced very successful results. In the technical ex situ and in situ tests, the concentration of gasoline range organics C6–C12 at the considered bases was reduced to $14–58\mathrm{mg}/\mathrm{kg}$, allowing safe usage of the facilities and helping prevent propagation of the contaminants.