On-site filtration of large sample volumes improves the detection of opportunistic pathogens in drinking water distribution systems.

In this study, we compared conventional vacuum filtration of small volumes through disc membranes (effective sample volumes for potable water: 0.3-1.0 L) with filtration of high volumes using ultrafiltration (UF) modules (effective sample volumes for potable water: 10.6-84.5 L) for collecting bacterial biomass from raw, finished, and tap water at seven drinking water systems. Total bacteria, Legionella spp., Legionella pneumophila, Mycobacterium spp., and Mycobacterium avium complex in these samples were enumerated using both conventional quantitative PCR (qPCR) and viability qPCR (using propidium monoazide). In addition, PCR-amplified gene fragments were sequenced for microbial community analysis. The frequency of detection (FOD) of Legionella spp. in finished and tap water samples was much greater using UF modules (83% and 77%, respectively) than disc filters (24% and 33%, respectively). The FODs for Mycobacterium spp. in raw, finished, and tap water samples were also consistently greater using UF modules than disc filters. Furthermore, the number of observed operational taxonomic units and diversity index values for finished and tap water samples were often substantially greater when using UF modules as compared to disc filters. Conventional and viability qPCR yielded similar results, suggesting that membrane-compromised cells represented a minor fraction of total bacterial biomass. In conclusion, our research demonstrates that large-volume filtration using UF modules improved the detection of opportunistic pathogens at the low concentrations typically found in public drinking water systems and that the majority of bacteria in these systems appear to be viable in spite of disinfection with free chlorine and/or chloramine.IMPORTANCEOpportunistic pathogens, such as Legionella pneumophila, are a growing public health concern. In this study, we compared sample collection and enumeration methods on raw, finished, and tap water at seven water systems throughout the State of Minnesota, USA. The results showed that on-site filtration of large water volumes (i.e., 500-1,000 L) using ultrafiltration membrane modules improved the frequency of detection of relatively rare organisms, including opportunistic pathogens, compared to the common approach of filtering about 1 L using disc membranes. Furthermore, results from viability quantitative PCR (qPCR) with propidium monoazide were similar to conventional qPCR, suggesting that membrane-compromised cells represent an insignificant fraction of microorganisms. Results from these ultrafiltration membrane modules should lead to a better understanding of the microbial ecology of drinking water distribution systems and their potential to inoculate premise plumbing systems with opportunistic pathogens where conditions are more favorable for their growth.

Animal Feedlots and Domestic Wastewater Discharges are Likely Sources of N-Nitrosodimethylamine (NDMA) Precursors in Midwestern Watersheds.

N-nitrosodimethylamine (NDMA) precursor concentrations along four major rivers in Minnesota, USA were quantified and correlated with watershed land cover types, anthropogenic activity, and organic matter characteristics. River water samples (36 in total) were chloraminated under uniform formation conditions (UFC) before and after lime-softening treatment, and the resulting NDMA concentrations were quantified (NDMAUFC). Regarding land cover, NDMAUFC in raw river water exhibited weak positive correlations with urban land (ρ = 0.33, p = 0.05) and cropland coverage (ρ = 0.35, p = 0.04). For anthropogenic activity, NDMAUFC in raw river water positively correlated with the number of feedlots (ρ = 0.57), total weight of animals (ρ = 0.68), and total number of domestic wastewater treatment plants (WWTPs; ρ = 0.63) with p < 0.01. NDMAUFC positively correlated with region IV fluorescence intensity from fluorescence excitation-emission spectra (ρ = 0.70, p < 0.01). Lime softening of river water typically increased NDMAUFC and preferentially removed organic matter that fluoresces in region V, suggesting that the organic matter in this region decreases NDMAUFC by competing for available chloramines. Overall, animal feedlots, along with domestic WWTPs, are predominant sources of NDMA precursors in the studied watersheds, while croplands and urban runoff are of lesser importance.

Spatiotemporal Variability in N-Nitrosodimethylamine Precursor Levels in a Watershed Impacted by Agricultural Activities and Municipal Wastewater Discharges and Effects of Lime Softening.

The Crow River, a tributary of the Mississippi River in Minnesota, U.S.A., that is impacted by agricultural activities and municipal wastewater discharges, was sampled approximately monthly at 12 locations over 18 months to investigate temporal and spatial variations in N-nitrosodimethylamine (NDMA) precursor levels. NDMA precursors were quantified primarily by measuring NDMA formed under the low chloramine dose uniform formation conditions protocol (NDMAUFC) and occasionally using the high dose formation potential protocol (NDMAFP). Raw water NDMAUFC concentrations (2.2 to 128 ng/L) exhibited substantial temporal variation but relatively little spatial variation. An increase in NDMAUFC was observed for 126 of 169 water samples after lime-softening treatment. A kinetic model indicates that under chloramine-limited UFC test conditions, the increase in NDMAUFC can be attributed to a decrease in competition between precursors and natural organic matter (NOM) for chloramines and reduced interactions of precursors with NOM. NDMAUFC concentrations correlated positively with dissolved nitrogen concentration (ρ = 0.44, p < 0.01) when excluding the spring snowmelt period and negatively correlated with dissolved organic carbon concentration (ρ = -0.47, p < 0.01). Overall, NDMA precursor levels were highly dynamic and strongly affected by lime-softening treatment.

Multi-scale Investigation of Ammonia-Oxidizing Microorganisms in Biofilters Used for Drinking Water Treatment.

Ammonia-oxidizing microorganisms (AOMs) include ammonia-oxidizing bacteria (AOB), archaea (AOA), and Nitrospira spp. sublineage II capable of complete ammonia oxidation (comammox). These organisms can affect water quality not only by oxidizing ammonia to nitrite (or nitrate) but also by cometabolically degrading trace organic contaminants. In this study, the abundance and composition of AOM communities were investigated in full-scale biofilters at 14 facilities across North America and in pilot-scale biofilters operated for 18 months at a full-scale water treatment plant. In general, the relative abundance of AOM in most full-scale biofilters and in the pilot-scale biofilters was as follows: AOB > comammox Nitrospira > AOA. The abundance of AOB in the pilot-scale biofilters increased with increasing influent ammonia concentration and decreasing temperature, whereas AOA and comammox Nitrospira exhibited no correlations with these parameters. The biofilters affected AOM abundance in the water passing through the filters via collecting and shedding but exhibited a minor influence on the composition of AOB and Nitrospira sublineage II communities in the filtrate. Overall, this study highlights the relative importance of AOB and comammox Nitrospira compared to AOA in biofilters and the influence of filter influent water quality on AOM in biofilters and their release into the filtrate.

Flushing Temporarily Improves Microbiological Water Quality for Buildings Supplied with Chloraminated Surface Water but Has Little Effect for Groundwater Supplies.

Microbial communities in premise plumbing systems were investigated after more than 2 months of long-term stagnation, during a subsequent flushing event, and during post-flush stagnation. Water samples were collected from showers in buildings supplied with chlorinated groundwater, untreated groundwater, and chloraminated surface water. The building supplied with chlorinated groundwater generally had the lowest bacterial concentrations across all sites (ranging from below quantification limit to 5.2 log copies/L). For buildings supplied with untreated groundwater, bacterial concentrations (5.0 to 7.6 log copies/L) and microbial community diversity index (ACE) values were consistent throughout sampling. Nontuberculous mycobacteria (NTM) and Legionella pneumophila were not detected in any groundwater-supplied buildings. Total bacteria, Legionella spp., and NTM were abundant in the surface water-supplied buildings following long-term stagnation (up to 7.6, 6.2, and 7.6 log copies/L, respectively). Flushing decreased these concentrations by ∼1 to >4 log units and reduced microbial community diversity, but the communities largely recovered within a week of post-flush stagnation. The results suggest that buildings supplied with disinfected surface water are more likely than buildings supplied with treated or untreated groundwater to experience deleterious changes in microbiological water quality during stagnation and that the water quality improvements from flushing with chloraminated water, while substantial, are short-lived.

Microbiome of Drinking Water Biofilters is Influenced by Environmental Factors and Engineering Decisions but has Little Influence on the Microbiome of the Filtrate.

Bacterial communities in biofilters can improve drinking water quality through the biodegradation of dissolved contaminants but also pose potential risks by harboring and shedding microbes into the drinking water distribution system. In this study, pilot-scale granular activated carbon (GAC)-sand and anthracite-sand pilot-scale biofilters were investigated to determine the effects of filter design and operation on the microbiome of the filter media and its relationship to the microbiome in the filter effluent water. Bacterial abundance in the biofilters was relatively stable over time. Bacterial community composition exhibited spatial variation (i.e., with bed depth) and temporal variation linked to water quality changes. Bacterial community composition was significantly affected by the media type (GAC vs anthracite) and backwashing strategy (chloraminated water vs nonchloraminated water). The biofilters reduced bacterial abundance in the water (∼70%) but had only a minor effect on the bacterial community composition in the filtrate. Overall, our results suggest that the bacterial communities growing on biofilters affect filtered water quality primarily through the biotransformation of pollutants and nutrients rather than by altering the microbial community composition of the water as it passes through the filter.

Flushing of Stagnant Premise Water Systems after the COVID-19 Shutdown Can Reduce Infection Risk by Legionella and Mycobacterium spp.

There is concern about potential exposure to opportunistic pathogens when reopening buildings closed due to the COVID-19 pandemic. In this study, water samples were collected before, during, and after flushing showers in five unoccupied (i.e., for ∼2 months) university buildings with quantification of opportunists via a cultivation-based assay (Legionella pneumophila only) and quantitative PCR. L. pneumophila were not detected by either method; Legionella spp., nontuberculous mycobacteria (NTM), and Mycobacterium avium complex (MAC), however, were widespread. Using quantitative microbial risk assessment (QMRA), the estimated risks of illness from exposure to L. pneumophila and MAC via showering were generally low (i.e., less than a 10-7 daily risk threshold), with the exception of systemic infection risk from MAC exposure in some buildings. Flushing rapidly restored the total chlorine (as chloramine) residual and decreased bacterial gene targets to building inlet concentrations within 30 min. During the postflush stagnation period, the residual chlorine dissipated within a few days and bacteria rebounded, approaching preflush concentrations after 6-7 days. These results suggest that flushing can quickly improve water quality in unoccupied buildings, but the improvement may only last a few days.

Prediction of Photochemically Produced Reactive Intermediates in Surface Waters via Satellite Remote Sensing.

Absorption of solar radiation by colored dissolved organic matter (CDOM) in surface waters results in the formation of photochemically produced reactive intermediates (PPRIs) that react with pollutants in water. Knowing the steady-state concentrations of PPRIs ([PPRI]ss) is critical to predicting the persistence of pollutants in sunlit surface waters. CDOM levels (a440) can be measured remotely for lakes over large areas using satellite imagery. Laboratory measurements of [PPRI]ss and apparent quantum yields (Φ) of three PPRIs (3DOM*, 1O2, and •OH) were made for 24 lake samples under simulated sunlight. The total rate of light absorption by the water samples (Ra), the rates of formation (Rf), and [PPRI]ss of 3DOM* and 1O2 linearly increased with increasing a440. The production rate of •OH was linearly correlated with a440, but the steady-state concentration was best fit by a logarithmic function. The relationship between measured a440 and Landsat 8 reflectance was used to map a440 for more than 10 000 lakes across Minnesota. Relationships of a440 with Rf, [PPRIs]ss, and Ra were coupled with satellite-based a440 assessments to map reactive species production rates and concentrations as well as contaminant transformation rates. This study demonstrates the potential for using satellite imagery for estimating contaminant loss via indirect photolysis in lakes.

Nontuberculous Mycobacteria in Two Drinking Water Distribution Systems and the Role of Residual Disinfection.

Nontuberculous mycobacteria (NTM) are frequently found in chloraminated drinking water distribution systems (DWDSs) due to their chloramine tolerance. NTM were investigated in the water-main biofilms and drinking water of a chloraminated DWDS in the United States (initial chloramine residual = 3.8 ± 0.1 mg L-1) and a DWDS in Norway with minimal residual disinfectant (0.08 ± 0.01 mg L-1). Total mycobacteria and Mycobacterium avium complex (MAC) were quantified by qPCR targeting, respectively, atpE genes and the internal transcribed spacer region. Mycobacteria concentrations in drinking water did not differ between the two systems (P = 0.09; up to 6 × 104 copies L-1) but were higher in the biofilms from the chloraminated DWDS (P = 5 × 10-9; up to 5 × 106 copies cm-2). MAC were not detected in either system. Sequencing of mycobacterial hsp65 genes indicated that the chloraminated DWDS lacked diversity and consisted almost exclusively of M. gordonae. In contrast, there were various novel mycobacteria in the no-residual DWDS. Finally, Mycobacterium- and Methylobacterium-like 16S rRNA genes were often detected simultaneously, though without correlation as previously observed. We conclude that, though residual chloramine may increase mycobacterial biomass in a DWDS, it may also decrease mycobacterial diversity.

Occurrence of Legionella spp. in Water-Main Biofilms from Two Drinking Water Distribution Systems.

The maintenance of a chlorine or chloramine residual to suppress waterborne pathogens in drinking water distribution systems is common practice in the United States but less common in Europe. In this study, we investigated the occurrence of Bacteria and Legionella spp. in water-main biofilms and tap water from a chloraminated distribution system in the United States and a system in Norway with no residual using real-time quantitative polymerase chain reaction (qPCR). Despite generally higher temperatures and assimilable organic carbon levels in the chloraminated system, total Bacteria and Legionella spp. were significantly lower in water-main biofilms and tap water of that system ( p < 0.05). Legionella spp. were not detected in the biofilms of the chloraminated system (0 of 35 samples) but were frequently detected in biofilms from the no-residual system (10 of 23 samples; maximum concentration = 7.8 × 104 gene copies cm-2). This investigation suggests water-main biofilms may serve as a source of Legionella for tap water and premise plumbing systems, and residual chloramine may aid in reducing their abundance.

Effects of Chloramine and Coupon Material on Biofilm Abundance and Community Composition in Bench-Scale Simulated Water Distribution Systems and Comparison with Full-Scale Water Mains.

The vast majority of bacteria in drinking water distribution systems (DWDSs) reside in biofilms on the interior walls of water mains. Little is known about how water quality conditions affect water-main biofilms because of the inherent limitations in experimenting with drinking water supplies and accessing the water mains for sampling. Bench-scale reactors permit experimentation and ease of biofilm sampling, yet questions remain as to how well biofilms in laboratory reactors represent those on water mains. In this study, the effects of DWDS pipe materials and chloramine residual on biofilms were investigated by cultivating biofilms on cement, polyvinyl chloride, and high density polyethylene coupons in CDC reactors for up to 28 months in the presence of chloraminated or dechlorinated tap water. The bench-scale biofilm microbiomes were then compared with the microbiome on a water main from the full-scale system that supplied the water to the reactors. The presence of a chloramine residual (1.74 ± 0.21 mg/L) suppressed biofilm accumulation and selected for Mycobacterium-like and Sphingopyxis-like operational taxonomic units (OTUs) while the destruction of the chloramine residual resulted in a significant increase in biomass quantity and a shift toward a more diverse community dominated by Nitrospira-like OTUs, which, our results suggest, may be complete ammonia oxidizers (comammox). Coupon material, however, had a relatively minor effect on the abundance and community composition of the biofilm bacteria. Although biofilm communities from the chloraminated water reactor and the water mains shared some dominant populations (namely, Mycobacterium- and Nitrosomonas-like OTUs), the communities were significantly different. This manuscript provides novel insights into the effects of dechlorination and pipe material on biofilm community composition. Furthermore, to our knowledge, it is the first study to compare biofilm in a tap water-fed, bench-scale simulated distribution system to biofilm on water mains from the full-scale system supplying the tap water.

The Bacterial Communities of Full-Scale Biologically Active, Granular Activated Carbon Filters Are Stable and Diverse and Potentially Contain Novel Ammonia-Oxidizing Microorganisms.

The bacterial community composition of the full-scale biologically active, granular activated carbon (BAC) filters operated at the St. Paul Regional Water Services (SPRWS) was investigated using Illumina MiSeq analysis of PCR-amplified 16S rRNA gene fragments. These bacterial communities were consistently diverse (Shannon index, >4.4; richness estimates, >1,500 unique operational taxonomic units [OTUs]) throughout the duration of the 12-month study period. In addition, only modest shifts in the quantities of individual bacterial populations were observed; of the 15 most prominent OTUs, the most highly variable population (a Variovorax sp.) modulated less than 13-fold over time and less than 8-fold from filter to filter. The most prominent population in the profiles was a Nitrospira sp., representing 13 to 21% of the community. Interestingly, very few of the known ammonia-oxidizing bacteria (AOB; <0.07%) and no ammonia-oxidizing Archaea were detected in the profiles. Quantitative PCR of amoA genes, however, suggested that AOB were prominent in the bacterial communities (amoA/16S rRNA gene ratio, 1 to 10%). We conclude, therefore, that the BAC filters at the SPRWS potentially contained significant numbers of unidentified and novel ammonia-oxidizing microorganisms that possess amoA genes similar to those of previously described AOB.

Sulfate Reducing Bacteria and Mycobacteria Dominate the Biofilm Communities in a Chloraminated Drinking Water Distribution System.

The quantity and composition of bacterial biofilms growing on 10 water mains from a full-scale chloraminated water distribution system were analyzed using real-time PCR targeting the 16S rRNA gene and next-generation, high-throughput Illumina sequencing. Water mains with corrosion tubercles supported the greatest amount of bacterial biomass (n = 25; geometric mean = 2.5 × 10(7) copies cm(-2)), which was significantly higher (P = 0.04) than cement-lined cast-iron mains (n = 6; geometric mean = 2.0 × 10(6) copies cm(-2)). Despite spatial variation of community composition and bacterial abundance in water main biofilms, the communities on the interior main surfaces were surprisingly similar, containing a core group of operational taxonomic units (OTUs) assigned to only 17 different genera. Bacteria from the genus Mycobacterium dominated all communities at the main wall-bulk water interface (25-78% of the community), regardless of main age, estimated water age, main material, and the presence of corrosion products. Further sequencing of the mycobacterial heat shock protein gene (hsp65) provided species-level taxonomic resolution of mycobacteria. The two dominant Mycobacteria present, M. frederiksbergense (arithmetic mean = 85.7% of hsp65 sequences) and M. aurum (arithmetic mean = 6.5% of hsp65 sequences), are generally considered to be nonpathogenic. Two opportunistic pathogens, however, were detected at low numbers: M. hemophilum (arithmetic mean = 1.5% of hsp65 sequences) and M. abscessus (arithmetic mean = 0.006% of hsp65 sequences). Sulfate-reducing bacteria from the genus Desulfovibrio, which have been implicated in microbially influenced corrosion, dominated all communities located underneath corrosion tubercules (arithmetic mean = 67.5% of the community). This research provides novel insights into the quantity and composition of biofilms in full-scale drinking water distribution systems, which is critical for assessing the risks to public health and to the water supply infrastructure.

Feasibility of using a particle counter or flow-cytometer for bacterial enumeration in the assimilable organic carbon (AOC) analysis method.

Assimilable organic carbon (AOC) is one of the major determinants of microbial growth and stability in drinking water distribution systems. Nevertheless, AOC measurements are rarely conducted in practice owing, in part, to the tedious and time-consuming nature of the bioassay. Herein, we compared three alternative cell count approaches [flow cytometry with staining (FC-S), flow cytometry without staining (FC-NS), and particle counting (Coulter counter; CC)] for bacterial enumeration as a means to expedite the AOC bioassay. Our results suggest that of the three methods only FC-S provides a suitable alternative to plate counting for rapid and accurate enumeration of both P17 and NOX in the AOC bioassay. While the cell counts obtained by FC-NS were linearly correlated with those obtained using the traditional heterotrophic plate count (HPC) method (FC-NS: R(2) = 0.89-0.96), the AOC values obtained by FC-NS were overestimated by 18-57 %. The CC approach was unsuccessful in enumerating Spirillum strain NOX cells because of the relatively small size of that organism. The CC counts were linearly correlated with HPC for Pseudomonas fluorescens strain P-17 (P17) cells (R(2) = 0.83) but like FC-NS, the CC approach also overestimated the AOC values (for P-17). The advantage of the FC-S method over the other two is improved sensitivity and the ability to specifically enumerate whole cells (and likely viable) as opposed to non-viable cells, cell debris, and other contaminating particles introduced by the test water itself or sample handling.

Residual disinfectant effectively suppresses Legionella species in drinking water distribution systems supplied by surface water in Minnesota, USA.

Seven public water systems in Minnesota, USA were analyzed from one to five times over a two-year period to assess temporal changes in the concentrations of total bacteria, Legionella spp., and Legionella pneumophila from source (i.e., raw water) through the water treatment process to the end water user. Bacterial biomass was collected by filtering large volumes of raw water (12 to 425 L, median: 38 L) or finished and tap water (27 to 1205 L, median: 448 L) using ultrafiltration membrane modules. Quantitative PCR (qPCR) was then used to enumerate all bacteria (16S rRNA gene fragments), all Legionella spp. (ssrA), and Legionella pneumophila (mip). Total coliforms, Escherichia coli, and L. pneumophila also were quantified in the water samples via cultivation. Median concentrations of total bacteria and Legionella spp. (ssrA) in raw water (8.5 and 4.3 log copies/L, respectively) decreased by about 2 log units during water treatment. The concentration of Legionella spp. (ssrA) in water collected from distribution systems inversely correlated with the total chlorine concentration for chloraminated systems significantly (p = 0.03). Although only 8 samples were collected from drinking water distribution systems using free chlorine as a residual disinfectant, these samples had significantly lower concentrations of Legionella spp. (ssrA) than samples collected from the chloraminated systems (p = 5 × 10-4). There was considerable incongruity between the results obtained via cultivation-independent (qPCR) and cultivation-dependent assays. Numerous samples were positive for L. pneumophila via cultivation, none of which tested positive for L. pneumophilia (mip) via qPCR. Conversely, a single sample tested positive for L. pneumophilia (mip) via qPCR, but this sample tested negative for L. pneumophilia via cultivation. Overall, the results suggest that conventional treatment is effective at reducing, but not eliminating, Legionella spp. from surface water supplies and that residual disinfection is effective at suppressing these organisms within drinking water distribution systems.

Root exudate enhanced contaminant desorption: an abiotic contribution to the rhizosphere effect.

Despite reports in the literature of superior contaminant degradation in the root-zone of plants, this phenomenon, known as the rhizosphere effect, is poorly understood. We investigated whether root exudates could enhance desorption of residual pollutants, thus improving bioavailability and subsequent biodegradation potential. Root exudates were harvested from three species of hydroponically grown plants, and artificial root exudates (AREs) were created using a literature recipe. Aliquots of the exudates were metabolized by soil bacteria to investigate whether biotransformed exudates exhibited different chemical characteristics or had different effects on contaminant bioavailability than 'raw exudates.' Slurries of naphthalene-aged soil containing raw exudates had a significantly lower soil-water distribution coefficient (Kd) than slurries with metabolized exudates or no-exudate controls, exhibiting median reductions of 50% and 55%, respectively. Raw exudates had a significantly lower surface tension while not increasing overall solubility, indicating the presence of surface-active compounds below the critical micelle concentration; this is a newly observed mechanism of the rhizosphere effect. Exudate samples were characterized by specific UV absorbance, spectral slope, fluorescence index, and excitation-emission matrices. Substantial changes in organic carbon character pre- and postmetabolism, and between harvested exudates and AREs, suggest that AREs are not chemically representative of plant root exudates. Overall, we present evidence that enhanced contaminant desorption in the presence of exudates provides an abiotic contribution to the rhizosphere effect.

Effect of strain rate on the mechanical properties of Staphylococcus epidermidis biofilms.

The quantification of biofilm mechanical properties can serve as a basis for understanding biofilm resilience and for developing biofilm control strategies. One aspect of tensile testing that is likely to be important for a viscoelastic material such as bacterial biofilm, but unfortunately is often overlooked (i.e., not controlled or reported), is the strain rate used during testing. Thus, we performed tensile testing on intact S. epidermidis biofilms using the microcantilever method at 12 strain rate values ranging over approximately 3 orders of magnitude (0.013-9.07 s(-1)). Ultimate or cohesive strength, elastic modulus, and toughness increased with increasing strain rate and approached a plateau at approximately 1.3 s(-1). Failure strain, on the other hand, did not exhibit any trend with strain rate. Given that the mean values of some parameters increased by as much as 1 order of magnitude over the strain rate range used in this work, we suggest that the strain rate during tensile testing should be carefully controlled and reported to facilitate comparisons among different studies. Furthermore, the quantitative expressions developed in this work that relate mechanical property values with strain rate may be useful for modeling the deformation of bacterial biofilms under applied loads.

Fate of naphthalene in laboratory-scale bioretention cells: implications for sustainable stormwater management.

Bioretention cells are increasingly popular in low-impact development as a means to sustainably mitigate the environmental problems associated with stormwater runoff. Yet, much remains to be known regarding the removal and ultimate fate of pollutants such as petroleum hydrocarbons in bioretention cells. In this work, laboratory-scale bioretention cells were constructed inside sealed glass columns. The columns were periodically spiked with (14)C-naphthalene over a 5-month period and the fate of this representative hydrocarbon and the influence of vegetation on naphthalene fate was studied. Three column setups were used: one planted with a legume (Purple Prairie Clover, Dalea purpureum), one planted with grass (Blue-Joint Grass, Calamagrostis canadensis), and one unplanted (i.e., control). Overall naphthalene removal efficiency was 93% for the planted columns and 78% for the control column. Adsorption to soil was the dominant naphthalene removal mechanism (56-73% of added naphthalene), although mineralization (12-18%) and plant uptake (2-23%) were also important. Volatilization was negligible (<0.04%). Significant enrichment of naphthalene-degrading bacteria occurred due to contaminant exposure and plant growth as evidenced by increased biodegradation activity and increased naphthalene dioxygenase gene concentrations in the bioretention media. This research suggests that bioretention is a viable solution for sustainable petroleum hydrocarbon removal from stormwater, and that vegetation can enhance overall performance and stimulate biodegradation.

Development and testing of a novel microcantilever technique for measuring the cohesive strength of intact biofilms.

Cohesive strength is an important parameter for understanding and modeling the mechanics of biomass detachment from bacterial biofilms. It is challenging to measure the mechanical properties of biofilms, however, because biofilms may desiccate when removed from liquid medium and they are inherently fragile. Poppele and Hozalski (Poppele and Hozalski, 2003, J Microb Methods 55:607-615) presented a microcantilever method for measuring the tensile strength of detached biofilm fragments while submersed in liquid medium. Here we present a modification of the microcantilever method to quantify the strength of intact bacterial biofilms. Initial testing was performed on Pseudomonas aeruginosa biofilms and on Staphylococcus epidermidis biofilms grown in rotating disk reactors. The cohesive strength values were highly variable (i.e., coefficients of variation ranging from 71% to 143%) and ranged from 59 to 18,900 Pa for the P. aeruginosa biofilms and from 61 to 5,840 Pa for the S. epidermidis biofilms. The biofilms also appeared to be isotropic as strength did not vary with angle of testing relative to the direction of applied shear. Strength testing using both the intact and fragment methods was performed on five samples of P. aeruginosa biofilms, and the strength populations were not from the same distribution in three cases. Equivalent diameters for the fragments detached from biofilms during strength testing ranged from 5 to 500 microm, which is within the range of size of biofilm fragments observed in the effluents of lab-scale and full-scale bioreactors. The microcantilever is a simple yet powerful tool for measuring the cohesive strength of intact biofilms at a relevant scale.