Journal of Environmental Engineering

Aquatic colloids are ubiquitous in nature and are composed of inorganic and organic material. During treatment of drinking water they foul granular filters, membranes, and granular activated carbon and are likely to hinder oxidation processes. The organic fraction of aquatic colloids was isolated from 14 samples (three wastewaters, five reservoirs, three rivers, two biological laboratory reactors, and one treated reservoir water). Transmission electron microscopy revealed nanoscale cellular debris and fibrous material. On the basis of advanced spectroscopic techniques (Fourier transform infrared spectroscopy, ${}^{13}\mathrm{C}$ nuclear magnetic resonance, and transmission electron microscopy) organic colloids were found to contain peptidoglycan, lipids, carbohydrates, and proteins. Saccharides and hydrolysable amino acids accounted for 44 to 81% of organic carbon. The reactivity of these colloids toward chlorine was also evaluated. Disinfection by-product yields were comparable to other organic matter fractions. During simulation of conventional water treatment, organic colloids were well-removed (42% or $0.41\mathrm{mg}\mathrm{\text{-}}\mathrm{C}/\mathrm{L}$), and the organic colloid disinfection by-product yield was reduced by 32% for trihalomethanes and 25% for haloacetic acids, illustrating that conventional treatment was highly effective at physical removal. The composition of organic colloids relates to aquatic biological activity and is likely to pose treatment challenges when biologically active waters (e.g., wastewater reuse, algae-impacted surface water) enter drinking water treatment plants.

Drinking water treatment research can be complicated by difficulties in shipping large water quantities and changing natural organic matter (NOM) characteristics over time. To overcome these issues, it is advantageous to have a reliable method for concentrating NOM with minimal loss and maximum shelf-life. NOM concentration and preservation by lyophilization (freeze-drying) has been practiced for many years; however, little information for lyophilizing and reconstituting NOM exists in the literature. The purpose of this research was to both evaluate and improve the lyophilization process of a concentrated NOM solution and to determine the appropriate conditions (pH, mixing time, and concentration factor) under which to reconstitute lyophilized NOM for maximum total organic carbon (TOC) and ultraviolet absorbance at 254-nm $(\mathrm{U}{\mathrm{V}}_{254})$ recovery. To achieve this, a new lyophilization method was developed, in which NOM concentrate is supercooled below its freezing point and heat transfer is carefully regulated to avoid ice melt-back. The lyophilized NOM was then reconstituted in 36 aliquots, at every unique combination of three pH values (6, 8, and 10), three mixing times (1, 4, and 24 h), and four concentration factors ($1\times$, $10\times$, $100\times$, $1000\times$). The overall mean TOC recovery was 101% ($\pm 3.3\%$). The improved method recovered 11% more TOC than a conventional procedure. No statistically significant ($p<0.05$) differences in TOC concentration were observed with respect to pH, mixing time, or concentration factor. The mean recovery of $\mathrm{U}{\mathrm{V}}_{254}$ was 91% with significantly better recovery at pH 10 than at pH 8 or 6. Mixing time did not significantly affect $\mathrm{U}{\mathrm{V}}_{254}$ recovery. No statistically significant differences in $\mathrm{U}{\mathrm{V}}_{254}$ were observed between reconstitutions at $1\times$, $10\times$, and $100\times$. A small but significant drop (4.8%, $p=0.029$) in $\mathrm{U}{\mathrm{V}}_{254}$ recovery was observed between $100\times$ and $1000\times$ samples. Finally, it was found that lyophilization and reconstitution reduces the dissolved silicon concentration in the final sample

The iron (II)-ethylenediaminetetraacetic acid (EDTA) and iron (III)-EDTA activation of persulfate was compared using reactant-specific probe compounds: the combined sulfate radical and hydroxyl radical probe anisole, the hydroxyl radical probe nitrobenzene, and the reductant/nucleophile probe hexachloroethane. Iron (II)-EDTA and iron (III)-EDTA were equally effective for activating persulfate decomposition and for generating reductants/nucleophiles at pH 5, while iron (III)-EDTA was a more effective activator than iron (II)-EDTA for generating oxidants. Use of the combined sulfate radical and hydroxyl radical scavenger isopropanol and the hydroxyl radical scavenger tert-butanol demonstrated that approximately 86 and 73% of the oxidation activity in iron (II)-EDTA and iron (III)-EDTA-activated persulfate systems, respectively, was attributable to hydroxyl radical activity. The generation of hydroxyl radical in iron-EDTA-activated persulfate systems at pH 5 was confirmed using electron spin resonance spectroscopy. The results of these pathway analyses, in conjunction with confirmation of the results using the common groundwater contaminant trichloroethylene (TCE), show that iron (II)-EDTA and iron (III)-EDTA-activated persulfate may be an effective system for the in situ remediation of contaminated groundwater.

Fluidized-bed pretreatment of a rapid sand filter with alum [$\mathrm{A}{\mathrm{l}}_2(\mathrm{S}{\mathrm{O}}_4{)}_3·14{\mathrm{H}}_2\mathrm{O}$] has been shown to significantly improve performance during the initial portion of filter operation. Two alternative mechanisms are hypothesized by which alum pretreatment acts to improve particle removal: (1) $\mathrm{Al}(\mathrm{OH}{)}_{3(\mathrm{am})}$ coats the sand filter medium, and alters its porosity. (2) Precipitated $\mathrm{Al}(\mathrm{OH}{)}_{3(\mathrm{am})}$ embedded within the media pores acts as an additional filter medium that increases the particle capture efficiency. Conceptual models for these two mechanisms are developed and compared with experimental data for particle removal and head loss. Model predictions indicate that particle removal by filtration through $\mathrm{Al}(\mathrm{OH}{)}_{3(\mathrm{am})}$ flocs is the mechanism that significantly enhances the filter performance.

Sorption of endosulfan and chlorpyrifos was studied through batch experiments with three agricultural soils (organic carbon 0.8, 1.37, and 2%) and different initial concentrations of pesticides below the limits of their solubility. The effect of the presence of one pesticide on sorption of the other was also examined. The sorption of both pesticides followed the Lagergren kinetic model, and more than 95% of sorption was completed within 4–6 hours for endosulfan and 6–12 hours for chlorpyrifos. The equilibrium sorption of both pesticides was primarily on the soil organic matter, but the isotherms were nonlinear, especially at low-solute concentrations. The nonlinearity was the result of the sorption onto the hard carbon portion of soil organic carbon, which was also indicated by the simulations with a dual reactive domain model (DRDM). The sorption of either pesticide was suppressed in the presence of the other, and the isotherm nonlinearity decreased with increasing concentration of the cosolute. The effect of a cosolute was more prominent for the sorption of endosulfan compared with that of chlorpyrifos. The simulations with DRDM indicated a competitive sorption in the portion of organic carbon showing site-limited sorption. No competition was observed in the portion showing linear partition behavior. Among the two isomers of endosulfan, the $\beta$ isomer was preferentially partitioned over the $\alpha$ isomer.

Improved solutions for the porosity and specific surface of uniform porous media with the attached film are provided for four regular packing arrangements. These solutions overcome the limitations of previous solutions, which did not cover the full range of film thicknesses to complete occlusion of the pore space in the porous media. The void spaces within the clogged porous media are analyzed geometrically. The exact solutions for the volume and surface area of overlapping shapes and isolated void spaces are formulated in terms of integrals. New polynomial solutions are proposed to replace the numerical integration in the complete solutions to allow easy implementation in numerical models. With the new solutions, both the porosity and specific surface gradually decrease from the initial values to zero when the film increases to complete pore occlusion. These solutions have application in numerical modeling of the growth of organic mass and deposition of inorganic mass in leachate collection systems and other similar systems in which there maybe growth of a semipermeable film on the surface of the particles in a relatively uniformly graded granular porous medium.

Accurate knowledge of the movement of contaminants in porous media is essential to track their trajectory and later extract them from the aquifer. A two-dimensional flow model is implemented and then applied on a linear contaminant transport model in the same porous medium. Because of different sources of uncertainties, this coupled model might not be able to accurately track the contaminant state. Incorporating observations through the process of data assimilation can guide the model toward the true trajectory of the system. The Kalman filter (KF), or its nonlinear invariants, can be used to tackle this problem. To overcome the prohibitive computational cost of the KF, the singular evolutive Kalman filter (SEKF) and the singular fixed Kalman filter (SFKF) are used, which are variants of the KF operating with low-rank covariance matrices. Experimental results suggest that under perfect and imperfect model setups, the low-rank filters can provide estimates as accurate as the full KF but at much lower computational effort. Low-rank filters are demonstrated to significantly reduce the computational effort of the KF to almost 3%.

Long-term groundwater monitoring data for 48 retail gasoline sites were analyzed to define the characteristics of affected groundwater plumes containing benzene, methyl tert-butyl ether (MTBE), and tert-butyl alcohol (TBA). Results of this analysis were used to determine the observed range and statistical distribution of current plume lengths, plume stability conditions, constituent concentration trends and attenuation rates, and the remediation timeframe for this population of sites. The goal of this evaluation was to characterize plume behavior as observed across a variety of hydrogeologic settings, on the basis of detailed groundwater monitoring records, rather than to define the site-specific factors controlling plume behavior. The results indicate that MTBE plumes in groundwater underlying a majority of these underground storage tank sites that were monitored for five years or longer (1) have significantly diminished in concentration over time, (2) are comparable in length to benzene plumes, (3) are, like benzene plumes, principally stable or shrinking in size and concentration, and (4) are on track to achieve remedial goals within a timeframe comparable to or faster than that of benzene plumes. At these same sites, TBA plumes were found to be comparable to benzene and MTBE plumes in terms of plume length. However, whereas most TBA plumes are also stable or shrinking, the percentage of TBA plumes that are currently stable or shrinking (68%) is less than that for benzene plumes (95%) or MTBE plumes (90%), likely reflecting the temporary build-up of TBA concentrations in groundwater attributable to methyl tert-butyl ether (MTBE) biodegradation. Nevertheless, overall trends for TBA concentrations in groundwater indicate that TBA is attenuating at rates comparable to benzene and MTBE and can be expected to meet applicable remediation goals in a similar timeframe as the other gasoline constituents.

This work was carried out to investigate the response of Escherichia coli, Enterococcus faecalis, and spores of Bacillus subtilis to ultraviolet (UV) irradiation after freezing. The effect of freezing temperature and freeze thaw cycles on UV inactivation efficiency was investigated using a collimated beam apparatus. Freezing affected the response of all test microbes; overall, a lower UV inactivation was observed in the freezing-treated bacteria. The greatest UV inactivation decrease, approximately 1.0 log unit, occurred in the freezing-treated E. coli. The average inactivation level of E. faecalis after freezing was 0.3–0.5 log unit lower than that of the control, and the difference reached approximately 0.1–0.2 log for B. subtilis spores. Under certain conditions, the significantly lower UV inactivation was observed in E. faecalis after surviving freezing, and the freezing-treated spores became more resistant to UV. Freezing temperature did not have a profound effect on the response of freezing-treated test organisms to UV. Freeze thaw cycles affected the response of E. coli and E. faecalis but not B. subtilis. The results suggested that UV inactivation could be less effective on waterborne microorganisms preexposed to freezing.

In recent decades, there has been a growing interest in using constructed wetlands (CWs) for the treatment of contaminated waters. There has been an increase in the use of wetland technology because of its low greenhouse effects, low maintenance and operational costs, and energy efficiency. However, there exists a level of uncertainty in regards to the performance of CWs and in some cases, CWs act as sources of contaminants. To address the issue, this paper presents the removal efficiencies for different heavy metal elements and identifies the factors affecting the removal processes. Using the San Joaquin Marsh constructed wetlands, removal efficiencies for four heavy metal elements—Cd, Cu, Pb and Zn—were evaluated. It was found that the effluent metal concentrations were not substantially lower than the influent. The removal efficiencies of 23.9%, 10.6%, and 17.6% were found for Cd, Cu, and Zn, respectively. No significant reduction was observed for concentrations of Pb. Different aspects of the removal mechanisms and possible factors controlling them were addressed and some recommendations were made to increase the removal efficiency of such systems. The results indicated that factors such as low metal loadings and algal growth in the marsh could be responsible for low removal rates of Pb.

Instrumental analyses were combined with the community bureau of reference (BCR) sequential extraction procedure (SEP) to study heavy metal pollution in sediments exposed to different pollutant compositions (primarily inorganic, primarily organic, and an upstream control sample). Heavy metals accumulated primarily in the nonresidual fractions of sediments as revealed by sequential extraction. The discharge of waste water with different pollutant compositions differentially affected the distribution of heavy metal species in extracts from sediments. Sediment polluted by organic waste water had a higher proportion of heavy metals in the oxidizable fraction. A higher residual fraction was detected in the sample that was polluted by high fluorine waste water. For all sediment samples, most Pb was found in the reducible fraction. The oxidizable fraction contained most of the chemically extractable Cu and Cr (31.6–81.5%); whereas the highest percentage of Zn was found in the residual fraction for all samples (46.9, 37.8, and 65.4%, respectively). The mobility order of the heavy metals studied in polluted sediments was $\mathrm{Cr}>\mathrm{Cu}>\mathrm{Pb}>\mathrm{Zn}$, much different from the order $\mathrm{Cu}>\mathrm{Pb}>\mathrm{Cr}>\mathrm{Zn}$ found in the control unpolluted sediment sample.

Solids are widely identified as a carrier of harmful pollutants in stormwater runoff exerting a significant risk to receiving waters. This paper outlines the findings of an in-depth investigation on heavy metal adsorption to solids surfaces. Pollutant build-up samples collected from 16 road sites in residential, industrial, and commercial land uses were separated into four particle size ranges and analyzed for a range of physicochemical parameters and nine heavy metals including Iron (Fe), Aluminum (Al), Lead (Pb), Zinc (Zn), Cadmium (Cd), Chromium (Cr), Manganese (Mn), Nickel (Ni), and Copper (Cu). High specific surface area (SSA) and total organic carbon (TOC) content in finer particle size ranges were noted, thus confirming strong correlations with heavy metals. On the basis of their physicochemical characteristics, two different types of solids originating from traffic and soil sources were identified. Solids generated by traffic were associated with high loads of heavy metals, such as Cd and Cr, with strong correlation with SSA. This suggested the existence of surface dependent bonds, such as cation exchange between heavy metals and solids. In contrast, Fe, Al, and Mn, which can be attributed to soil inputs, showed strong correlation with TOC suggesting strong bonds such as chemsorption. Zn was found to be primarily attached to solids by bonding with the oxides of Fe, Al, and Mn. The data analysis also confirmed the predominance of the finer fraction, with 70% of the solids being finer than 150 µm and containing 60% of the heavy metal pollutant load.

Artificial intelligence techniques, such as neural networks, are modeling tools that can be applied to analyze urban runoff quality issues. Artificial neural networks are frequently used to model various highly variable and nonlinear physical phenomena in the water and environmental engineering fields. The application of neural networks for analyzing the performance of combined permeable pavement and earth energy systems is timely and novel. This paper presents the application of back-propagation neural networks and the testing of the Levenberg-Marquardt, Quasi-Newton, and Bayesian Regularization algorithms. The neural networks were statistically assessed for their goodness of prediction with respect to the biochemical oxygen demand (BOD), ammonia-nitrogen, nitrate-nitrogen, and ortho-phosphate-phosphorus by numerical computation of the mean absolute error, root-mean-square error, mean absolute relative error, and the coefficient of correlation for the prediction compared with the corresponding measured data. The three neural network models were assessed for their efficiency in accurately simulating the effluent water quality parameters from various experimental pavement systems. The models predicted all key parameters with high correlation coefficients and low minimum statistical errors. The back-propagation and feed-forward neural network models performed optimally as pollutant removal predictors with regard to these two sustainable technologies.

A recent study showed that, in addition to precipitating sulfides, ${\mathrm{Fe}}^{3+}$ addition to sewers reduces the sulfate-reducing and methanogenic activities of anaerobic sewer biofilms. The dynamic response of these activities to the commencement and termination of ferric dosage is investigated in this study. Both the sulfate-reduction and methane-production rates of anaerobic sewer biofilms decreased upon the addition of ferric ions, with a maximum inhibition of approximately 50% achieved after 3 and 7 days, respectively. In comparison, the sulfate-reducing activity of suspended biomass in the ${\mathrm{Fe}}^{3+}$ receiving reactor was completely inhibited. The volatile solids-to-total solids ($\mathrm{VS}/\mathrm{TS}$) ratio of sewer biofilms decreased from $91\pm 3\%$ before ${\mathrm{Fe}}^{3+}$ addition to $54\pm 6\%$ after a full adaptation of the biofilms to ${\mathrm{Fe}}^{3+}$ addition. Energy-dispersive spectroscopy profiles showed higher iron and sulfur contents in biofilms adapted to ${\mathrm{Fe}}^{3+}$ addition, suggesting FeS deposition in biofilms. We hypothesize that FeS precipitation primarily occurred in the upper layer of sewer biofilms causing inhibition to sulfate-reducing bacteria (SRB) and methanogenic archaea (MA) in this layer (and in the suspended phase). However, SRB and MA were likely to be active in the deep layer. The sulfate-reducing activity recovered to approximately 90% that of the reference reactor in 5 weeks after ${\mathrm{Fe}}^{3+}$ injection was stopped, whereas there was no recovery of the methanogenic activity in the same period. The results suggest that a temporary suspension of ferric dosage would not lead to immediate recovery of the sulfate-reducing and methanogenic activities. The recovery of sulfate-reducing activity was found to be linearly correlated with the recovery of biofilm $\mathrm{VS}/\mathrm{TS}$ ratio, following the termination of ${\mathrm{Fe}}^{3+}$ addition.

Dissolved oxygen (DO) is essential to maintaining flora and fauna in aquatic ecosystems. The DO replenishment through the surface reaeration mechanism is commonly described using the reaeration coefficient ($K_2$). This paper presents a new approach to modeling instream DO and estimating $K_2$. The new approach includes an extension of the Variable Residence Time (VART) model and optimization algorithms for inverse modeling. The VART model is modified in this paper to incorporate the DO reaeration mechanism across the air-water interface, forming the VART-DO model. A major advantage of the VART-DO model is that it is capable of simulating DO exchange across the water-sediment interface through the hyporheic exchange mechanism in addition to the air-water exchange. A sensitivity analysis is conducted to investigate the relative importance of key model parameters to DO modeling. It is found that the $K_2$ value increases markedly with the dispersion coefficient. The simplex-simulated annealing and the genetic algorithm are used to optimize the estimation of the parameter $K_2$. A comparison of the optimized $K_2$ value from the VART-DO model with those from existing empirical equations reveals that $K_2$ value could increase 1.5 to 6 times when the effects of the dispersion and hyporheic exchange mechanisms on DO are taken into account. The results from the VART-DO model provide new insights into the DO variation in streams and rivers. The VART-DO model also provides a general and effective tool for total maximum daily load (TMDL) calculations of DO.