Imidacloprid is one of the most frequently used neonicotinoid insecticides in crop fields, which is considered as a safer alternative to organochlorine, organophosphate, and carbamate pesticides. However, imidacloprid exhibits high toxicities towards insect pollinators (e.g., bees) and uncertain chronic effects on human health. The potentially high toxicity and frequent occurrence of imidacloprid in drinking water supplies justify its regular monitoring, particularly at the sites near agricultural fields. Unfortunately, the traditional methods for imidacloprid analysis are usually based on expensive and time-consuming solid-phase extraction and liquid chromatography–tandem mass spectrometry. Surface-enhanced Raman spectroscopy (SERS) is a promising alternative to the standard analytical methods because of its fast detection speed, high sensitivity, and potential for field deployment. To achieve rapid and inexpensive detection of imidacloprid, a “mixing-and-detecting” approach was employed using the benchmark citrate-coated gold nanoparticle (AuNP) colloid as the SERS substrate. We systematically investigated the reproducibility of the SERS spectra of imidacloprid as a function of its concentrations in water. We also examined the influence of drinking water matrices on the sensitivity and reproducibility of SERS for imidacloprid analysis. Our results demonstrated a 100% reliability of SERS for the detection of imidacloprid at 10 nM in Madison tap water. In addition, we successfully quantify imidacloprid using principal component analysis based on its concentration-dependent spectral patterns. This rapid and low-cost analytical method can be used for high-throughput prescreening of imidacloprid in drinking water supplies at an unprecedented spatiotemporal resolution.

The dissolved organic matter (DOM) in aquatic environments is influenced by natural and anthropogenic processes related to land use. We investigated how the fluorescent DOM (FDOM) composition varies in two stream networks that are impacted by compromised wastewater infrastructure and surrounded by different land uses. Monthly water samples were collected from 27 sites across rural-to-urban flow gradients in Baltimore, Maryland. The samples were analyzed for FDOM by excitation-emission matrix spectroscopy with parallel factor analysis. A four-component model was validated. Assessment of the FDOM components indicated that the relative presence of humic- and fulvic-acid like fluorescence components (C1, C2, C3) was fairly consistent across the sampling period. In contrast, the protein-like fluorescence component (C4), which we previously identified as a wastewater indicator, presented high variability in specific urban tributaries with a high proportion of impervious surfaces and sewer lines. Principal component analysis confirmed that sewer lines, impervious surfaces, and C4 clustered together at urban sites; moreover, Spearman correlations yielded significant positive relationships for C4 with impervious surfaces (ρ=0.602, p<0.001) and sewer lines (ρ=0.461, p<0.001). These findings were validated via co-detection of sucralose and caffeine, wastewater indicators, in water samples at concentrations as high as 1655 and 1735 ng L⁻¹, respectively. Taken together, these results not only reinforced the potential for C4 to serve as a wastewater indicator and proxy for compromised sewer infrastructure in urban streams, but also suggested that geospatial land use and sewer infrastructure data can be used to predict which tributaries are most vulnerable to sewer leaks.

The wide use of most organochlorine pesticides (OCPs) was banned in the 1970s in the North America. However, due to their persistence and bioaccumulation, some of them were still frequently detected in the sediment of the Great Lakes. This work aims at revealing the spatial distribution and the temporal trends of selected OCPs in the sediment of Lakes Michigan, Superior and Huron. A total of 28 cores (>500 core segments) and 112 Ponar surface grab sediment samples collected from 2010-2012 were analyzed for 15 OCPs using gas chromatography coupled with either electron impact ionization triple quadruple or electron capture negative ionization single quadruple mass spectrometry. All samples were also characterized for various physicochemical properties, and the cores were dated using multiple radionuclides. Based on the data from the surface grab samples, five compounds were detected with relatively high concentrations with medians of 0.043, 0.029, 0.045, 0.024, and 0.013 ng/g for chlorothalonil, trans-nonachlor, chlordanes, dieldrin, and dacthal, respectively. Highest concentrations of most compounds were found at sites near Sleeping Bear Dunes, Duluth, and Saginaw Bay in the three lakes, respectively. Core profiles showed that the net fluxes of hexachlorocyclohexanes (HCHs), hexachlorobenzene (HCB), chlordanes, trans-nonachlor, chlorothalonil, and dieldrin were peaking mostly in either the 1960s or 1980s and decreasing after then, whereas net fluxes of pentachloronitrobenzene (PCNB) and dacthal were increasing up until the 2010s due probably to their continuous use during sampling periods. Further research on these legacy pollutants is warranted.

Sanitary sewer lines are aging and failing in many urban areas in the United States, and these infrastructure failures have had measurable influences on surface water quality in urban streams. In San Diego, CA, USA, sanitary sewer overflows and interflow flushing soils around sanitary sewer lines have been shown to contribute to fecal contamination of waterways, especially during storm events. An isolated pool, which consistently appeared in the vicinity of a buried sanitary sewer line was investigated as a suspect site for sewer exfiltration. Compared to an adjacent (within 30 m) brackish stream, Alvarado Creek, pool samples consistently had significantly higher concentrations of total coliforms, Escherichia coli and enterococci and lower salinity. Results from 16S sequencing further showed that the relative abundances of microbial classes of the isolated pool samples were more similar to untreated wastewater than to other surface water samples. When the pool was no longer present during the dry season, moist soils collected from the site had significantly higher (p < 0.05) total coliform and enterococci concentrations than soils from a nearby, uncontaminated control site, reflecting the expected influence of sewage contamination. E. coli concentrations were, however, not measurable and HF183 concentrations were not significantly different than controls, which may be explained by the lower persistence of E. coli and HF183, compared to enterococci, in the natural environment. These results indicate that even in the absence of standing water, soil samples collected at sites of potential sewer exfiltration may serve to pinpoint sites of sanitary sewer failures.

Monitoring water quality at high frequency is challenging and costly, which motivates a search for strategies to reduce the required sampling frequency for effective load estimation and watershed characterization. Compressed sensing (CS) offers a potential approach to reconstruct high-frequency water quality data from limited measurements, given that water quality signals are commonly “sparse” in the Fourier frequency domain. In this study, we investigated the sparsity of stream flow and concentration time-series and tested reconstruction with CS. All stream signals were characterized as sparse using 15-minute discrete time-series transformed to the Fourier domain. Stream temperature, conductance, dissolved oxygen, and nitrate concentration were sparser than discharge, turbidity, and phosphorus. CS effectively reconstructed these signals with only 5-10% of measurements needed, meaning that 15-minute data were reconstructed with a 2.5-5-hour average sampling interval (and daily with less than weekly). CS may have broad applications for efficient sensing of intermittent environmental and geophysical signals.