Suspect-screening analysis of a coastal watershed before and after Hurricane Florence using high-resolution mass spectrometry

https://doi.org/10.1016/j.scitotenv.2021.146862Get rights and content

Highlights

  • Hurricane Florence delivered 686 mm rainfall to a forested water-reuse site in 2018.

  • Suspect screening was used to analyze chemicals in waste, ground and surface water.

  • On-site chemical reservoirs were diluted after the storm vs. upstream and off-site.

  • This site was functionally and hydrologically robust to extreme precipitation events.

  • Similar sites may likewise be robust to such events which are expected to increase.

Abstract

On September 14, 2018, Hurricane Florence delivered ~686 mm rainfall to a 106 km2 watershed in coastal North Carolina, USA. A forested land treatment site comprises one third of the watershed wherein municipal wastewater effluent is spray-irrigated onto 8.9 km2 of forest. This communication provides insight for land treatment function under excess water duress as well as changes in organic chemical composition in on- and off-site waters before (June 2018) and after (September & December 2018) Hurricane Florence's landfall. We compare the numbers and relative abundances of chemical features detected using suspect screening high resolution mass spectrometry in waste-, ground-, and surface water samples. Values for upstream and receiving waters in September were lower than for sampling events in June and December, indicating an expected dilution effect across the watershed. Chemical diversity was greatest for all surface water samples in December, but only upstream surface water showed a dramatic five-fold increase in relative chemical abundance. Chemical abundance in on-site water and downstream surface water was equal to or lower than the September storm dilution effect. These data suggest that the land treatment system is functionally and hydrologically robust to extreme storm events and contributed to dilution of upstream chemical reservoirs for downstream receiving waters for months after the storm. Similar systems may embody one water reuse strategy robust to the increasing occurrence of extreme precipitation events.

Introduction

Hurricanes (Kossin et al., 2017; Melillo and Richmond, 2014) and extreme rainfall (Knight and Davis, 2009) are projected to increase in intensity and frequency as evidenced by recent events in coastal U.S. communities from Texas to North Carolina (Moore et al., 2015; Paerl et al., 2019; Prat and Nelson, 2016). These extreme events can impact the transport and distribution of organic chemical contaminants via the deposition of aerosolized contaminants (Noyes et al., 2009), the disturbance and resuspension of sediments in water (Fisher et al., 2016a; Van Metre et al., 2006), surface run-off, infiltration of flood waters to groundwater (Bonneau et al., 2017; Jasechko et al., 2017; Jasechko and Taylor, 2015), and overflows from wastewater treatment plants, industrial and agricultural waste lagoons, and septic systems (Fisher et al., 2016a; Fisher et al., 2016b). The resultant disruption of chemical reservoirs therefore merits further understanding of both the extent of chemical redistribution and potential risk to surface and groundwater resources.

To date, studies of hurricanes have focused on chemical disturbance of nutrients (Fisher et al., 2016b; Bales et al., 2000; Presley et al., 2006), metals (Van Metre et al., 2006; Presley et al., 2006; Fox et al., 2009; Pardue et al., 2005), polycyclic aromatic hydrocarbons (Fisher et al., 2016a; Van Metre et al., 2006; Bales et al., 2000; Fox et al., 2009; Pardue et al., 2005), pesticides (Bales et al., 2000), pharmaceuticals and hormones (Fisher et al., 2016a; Fisher et al., 2016b), and other organic contaminants of concern (COCs) (Fisher et al., 2016a; Van Metre et al., 2006; Presley et al., 2006) using targeted analytical methods. We utilized suspect screening analysis via high resolution mass spectrometry (HRMS) versus targeted analysis to evaluate a broader array of regulated and non-regulated chemicals present in watershed sources before and after Hurricane Florence. This approach involves qualitative screening for a variety of chemical features (defined as detected m/z, retention time, and abundance), in which spectral output is compared with compound-specific parameters contained within curated chemical databases for tentative identification (Aceña et al., 2015; Gosetti et al., 2016). Our aim was to use HRMS to explore chemical features and tentatively identified chemicals prior to and after extreme rainfall and flooding from Hurricane Florence. Established sampling locations provided a unique opportunity to compare chemical profiles in ground- and surface waters temporally and spatially across a watershed with diverse reservoirs of chemicals including a municipal wastewater land treatment site.

On September 14, 2018, the eye of Hurricane Florence passed 32 km southwest of the study watershed near Jacksonville, North Carolina. The City of Jacksonville Land Treatment Site (LTS; Fig. A.1) comprises a large portion of the watershed; it irrigates secondary treated municipal wastewater across 890 ha of a 2193 ha mixed hardwood and loblolly pine forest (Birch et al., 2016; Hedgespeth et al., 2019; McEachran et al., 2016). Two weeks prior to Hurricane Florence's arrival, the LTS initiated emergency spraying to lower wastewater reservoir levels. The LTS did not irrigate on September 14, 2018, when Hurricane Florence delivered 686 mm of rainfall to the site and raised groundwater elevation to within 0.04–0.45 m below ground surface (Figs. A.2, A.3). On September 15, 2018, the facility resumed operation and emergency spraying continued until April 2019 (Jacksonville Land Treatment Site, personal communication). Emergency spraying involves more frequent rotation of irrigation zones, thus increasing hydraulic and chemical loading for spray fields. Permit requirements (e.g. avoiding pooling of irrigation water on soil surface) must be ensured, and daily surface water quality monitoring is required for microbiological, nutrient, and select macro-chemical parameters. Prior hydrological measurements and isotopic modeling have shown that irrigated LTS wastewater sustains surface water baseflow during inter-storm periods and constitutes a large fraction of streamflow during storm response peaks via groundwater discharge and lateral subsurface flow (Birch et al., 2016). Peak stream flow during a 2015 tropical storm resulted in peak export of select pharmaceuticals from the LTS (McEachran et al., 2017).

While lagoons did not overflow and accidental discharge of wastewater did not occur due to the storm at the LTS, we expected the extreme rainfall of Hurricane Florence to mobilize chemical reservoirs from LTS soils and groundwater to receiving surface water. However, we were cognizant of similar dynamics in upstream regions of the watershed where land use is dominated by residential development and septic systems, agricultural row crops, and a municipal airport (Birch et al., 2016). Because coastal forests provide important ecosystem services through stormwater mitigation and groundwater recharge (Callahan et al., 2017), suspect-screening evaluation of waters at the LTS and the larger watershed provides unique insight to pre- and post-hurricane changes in COC release for waste-, ground-, and surface waters for the facility and surrounding watershed.

Section snippets

Sampling

Grab samples were collected from LTS wastewater effluent (post-chlorination; i.e. irrigation water) and from two groundwater monitoring wells on the LTS site (LTS wells A and B) and LTS surface water. Reference watershed groundwater was collected to the immediate west and hydrologically-upgradient of the LTS (off site) as well as surface waters upstream and downstream of the LTS (Fig. A.1) (Hedgespeth et al., 2019). Samples were collected prior to Hurricane Florence on June 10–11, 2018, as part

Results and discussion

We detected ~4000 unique chemical features in LTS wastewater consistently across sampling events (Fig. 1a). Non-targeted analysis does not permit direct knowledge of compound concentration, but QC indicated relative stability of injections and matrix effects; we therefore make qualitative comparisons of relative chemical abundance, as a proxy for chemical concentrations, among individual chemicals and chemical use categories. Total relative abundance of all chemical features was nearly 30%

CRediT authorship contribution statement

Melanie L. Hedgespeth: Conceptualization, Investigation, Formal analysis, Visualization, Writing – original draft, Writing – review & editing. James P. McCord: Investigation, Resources, Writing – review & editing. Katherine A. Phillips: Resources, Writing – review & editing. Mark J. Strynar: Resources, Writing – review & editing. Damian Shea: Conceptualization, Writing – review & editing. Elizabeth Guthrie Nichols: Conceptualization, Investigation, Visualization, Resources, Writing – review &

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

The authors would like to thank the City of Jacksonville and the Jacksonville LTS for their assistance. Special thanks to Amy Palaschak, Martina Gonzalez Bertello, Nancy Gibson, and Joshua Pil. This study was funded by USDA NIFA Award # 2016-68007-25069.

This document has been reviewed by the U.S. Environmental Protection Agency, Office of Research and Development, and approved for publication. The views expressed in this article are those of the authors and do not necessarily represent the

References (36)

  • J. Aceña et al.

    Advances in liquid chromatography–high-resolution mass spectrometry for quantitative and qualitative environmental analysis

    Anal. Bioanal. Chem.

    (2015)
  • Bales, J. D.; Oblinger, C. J.; Sallenger, A. H. Two Months of Floodings in Eastern North Carolina, September–October...
  • A.L. Birch et al.

    Hydrologic impacts of municipal wastewater irrigation to a temperate forest watershed

    J. Environ. Qual.

    (2016)
  • T.J. Callahan et al.

    Coastal forests and groundwater: using case studies to understand the effects of drivers and stressors for resource management

    Sustainability

    (2017)
  • T.H. Epps et al.

    Characterization of storm flow dynamics of headwater streams in the South Carolina Lower Coastal Plain1

    JAWRA Journal of the American Water Resources Association

    (2013)
  • M. Fox et al.

    Potential for chemical mixture exposures and health risks in New Orleans Post-Hurricane Katrina

    Hum. Ecol. Risk. Assess.

    (2009)
  • P. Gago-Ferrero et al.

    Extended suspect and non-target strategies to characterize emerging polar organic contaminants in raw wastewater with LC-HRMS/MS

    Environmental Science & Technology

    (2015)
  • F. Gosetti et al.

    Contaminants in water: non-target UHPLC/MS analysis

    Environ. Chem. Lett.

    (2016)
  • Cited by (8)

    • Target and suspect analysis of liquid crystal monomers in soil from different urban functional zones

      2023, Science of the Total Environment
      Citation Excerpt :

      Therefore, it is necessary to determine whether novel LCMs are present in the environment. High-resolution mass spectrometry (HRMS) has been applied to explore unpredicted contaminants in various environments (Aalizadeh et al., 2021; Baesu et al., 2021; Caballero-Casero et al., 2021; Hedgespeth et al., 2021; Gong et al., 2022; Ren and Su, 2022; Ye and Su, 2022), due of its high mass-spectral resolving power, wide mass range, and high mass accuracy. Based on the above background, we combined the target and suspect analysis techniques, in allusion applied them to 39 target LCMs and 1173 suspect LCMs (Su et al., 2022), which were used for the synthetic estimation of the contamination status of LCMs in soil samples collected from five different functional zones.

    • Non-target and suspect-screening analyses of hydroponic soybeans and passive samplers exposed to different watershed irrigation sources

      2022, Science of the Total Environment
      Citation Excerpt :

      Because the land treatment system stores wastewater in large open reservoirs for 10–14 days and these reservoirs receive rainfall, consistent rainfall events can dilute the number of chemical features in wastewater to equivalent numbers detected in upstream surface waters (Hedgespeth et al., 2021). Aggregated ToxPI scores of TICs from multi-sampling campaigns were similar for chemicals of greatest abundance between upstream surface water samples and wastewater samples (Hedgespeth et al., 2021). After a catastrophic hurricane in 2018, the number of chemical features and total abundance in upstream surface water and off-site groundwater samples far exceeded those numbers determined for on-site and downstream surface waters and groundwater at two week and three months beyond the hurricane event (Hedgespeth et al., 2021).

    View all citing articles on Scopus
    View full text