Concurrent Assessment of Diffusive and Advective PAH Movement Strongly Affected by Temporal and Spatial Changes

Abstract

Chemical movement influences exposure, remediation and interventions. Understanding chemical movement in addition to chemical concentrations at contaminated sites is critical to informed decision making. Using seepage meters and passive sampling devices we assessed both diffusive and advective flux of bioavailable polycyclic aromatic hydrocarbons (PAHs) at three time points, across two seasons, at a former creosote site in St. Helens, Oregon, United States. To our knowledge, this is the first time both diffusive and advective flux have been measured simultaneously at a contaminated site. Concentrations of 39 parent PAHs were determined by gas chromatography triple quadrupole mass spectrometry. Across both seasons and all sites, diffusive flux of PAHs was up to three orders of magnitude larger than advective flux. Release of PAHs from sediments and water were identified, likely from legacy contamination, as well as deposition from the air into the site from contemporary and other sources. The majority of PAH movement was comprised of three and four ring PAHs. Chemical movement on the site was found to be spatially and temporally variable. Volatilization decreased and atmospheric deposition increased from summer to fall. At the locations with higher levels of contamination, sum PAH release from sediments decreased by more than two orders of magnitude from summer to late fall. These data reflect the spatial heterogeneity and temporal variability of this site and demonstrate the importance of seasonality in assessing chemical movement at contaminated sites. Results from this study can inform future legacy site assessments to optimize remediation strategies and assess remediation effectiveness.

Graphical Abstract

1. Introduction

Rapid industrialization and a lack of regulation on chemical discharge prior to 1972 has led to a large number of contaminated sites across the United States^1,2 including many abandoned without remediation¹. These legacy sites continue to be a secondary source of contamination to sediments, water, and the surrounding environment³. One chemical class found at legacy contaminated sites is polycyclic aromatic hydrocarbons (PAHs), a class of abundant, widespread organic contaminants arising from both natural and anthropogenic sources. PAHs are defined as having at least two fused aromatic rings and are mainly semi-volatile, with aqueous solubility and vapor pressure decreasing with increasing molecular weight⁴. They are slow to degrade and the wide range of PAH sources leads to their abundance and persistence in the environment. Previous studies have linked PAH exposure to carcinogenicity, developmental and neurological effects, respiratory problems and suppressed immune function^5–11. Assessment and effective remediation of legacy contaminated sites is important for reducing exposure to PAHs.

Bioavailable chemicals are in constant motion, based on their physicochemical properties and environmental conditions^12,13. The movement of a chemical may occur within a single environmental compartment, or across multiple interfaces as part of the larger biogeochemical cycle. Chemical movement can be function of a single type of movement (such as diffusive flux) or multiple types of movement simultaneously (such as diffusive and advective flux)¹³. Because chemical movement will influence exposure, remediation and interventions^14-16, it is important to study direction, magnitude and type of chemical movement beyond the simple chemical concentrations present at contaminated sites. Unlike fugacity, which only provides net direction of chemical movement, flux provides a magnitude and direction vector for a given compound. Measuring flux can therefore provide a quantitative measure of contaminant load from a source into a receiving system, for example sediment into water.

Two common types of chemical movement are diffusive and advective flux. Diffusive flux is a passive transport process resulting from the movement of chemicals from environmental compartments with high concentrations to those of low concentrations, attempting to reach a state of equilibrium¹⁷. This movement between environmental compartments occurs across a diffusion-controlled film, known as the boundary layer. Advective flux refers to the bulk movement of chemicals with flowing air or water. This measure is based on chemical concentration and flow velocity, with chemicals moving in the same direction as air or water flow¹⁸. While diffusive and advective flux both provide a magnitude and direction vector for a given compound, these two types of movement are unique, and are driven by different processes. Diffusive flux is directly proportional to concentration gradients^{17, 19, 20}. In contrast, advective flux is additionally driven by the movement of water and can be influenced by factors such as wave or tidal pumping, or groundwater flow ^{21, 22}. Therefore, the type of chemical movement occurring may influence which remediation approach is most effective at contaminated sites. For example, previous work has shown that PAHs may break through a passive remediation cap in a short period of time in the presence of hyporheic flows²². Despite their different transport processes and potential implications for designing remediation strategies, to our knowledge diffusive and advective flux have not yet been measured simultaneously at contaminated sites.

In any dynamic system, chemical movement may also change depending on season. Environmental conditions such as temperature, wind speed, precipitation, water table levels and changes in water flow can influence movement of persistent pollutants, affecting site assessment and remediation decisions^23-26. In this study, we assess temporal influence on diffusive and advective flux at a legacy creosote site. Measuring both diffusion and advection of chemical movement temporally provides a more complete understanding of chemical fate in the environment.

When investigating chemical exposure and efficient remediation strategies, it is important to consider the bioavailable fraction, which is a more biologically relevant measure than total concentrations. Traditional ”grab samples” require more intensive laboratory manipulation to estimate the bioavailable fraction that can lead to increased cost or loss of analytical sensitivity²⁷. Passive sampling devices (PSDs) are an advantageous alternative, because they non-selectively sequester organic compounds in a biomimetic manner²⁸, a diffusional process between the membrane and the environment. Passive samplers also represent a time-weighted average measure across a given deployment period, rather than a discrete measure. Research spanning multiple decades has demonstrated that a variety of polymers can be used as passive samplers in various applications across all environmental compartments deployed from hours to months at a time with little to no maintenance^{27, 29–31}. The ease of implementation also allows for co-deployment across multiple environmental compartments, such as air, waters and sediments, to calculate chemical movement in addition to concentration. Therefore, passive samplers offer a simple and direct approach to measure the bioavailable fraction.

This study assesses bioavailable parent PAH movement between air, water and sediments at a former creosote site. Spatial and temporal changes of chemical movement were determined, as well as the portion of chemical movement by diffusion or advection processes. The implications for policies such as remediation strategies are discussed. The objectives of this study were to; 1) assess bioavailable PAH movement across the air-water and water-sediment interfaces; 2) evaluate spatial and temporal changes in PAH movement and 3) assess contributions of PAH movement from either diffusive or advective processes.

2. Materials and Methods

2.1. Sampling location

Scappoose Bay is a tidal estuary located along the Multnomah Channel of the Willamette River near the confluence with the Columbia River near the town of St. Helens, Oregon, United States. Industrial use in this area dates back to the early 1900s, and primarily consisted of wood product manufacturing³². The 42-acre former Pope & Talbot Inc. wood treatment facility operated in this area from 1912 to 1960³³. At this facility, wood was preserved using coal tar creosote, which can contain as much as 85% alkylated and parent PAHs^34,35. The property was purchased by the Port of St Helens (now the Port of Columbia County) in 1963. Preliminary assessments in 2000 found the subsurface soil and sediment contaminated with PAHs as well as pentachlorophenol, arsenic, chromium and petroleum hydrocarbons³³. The Port removed the remaining structures and filled much of the property’s terrestrial land with dredge sands between 1963 and 1973. The remedial investigation at the site is complete, and the feasibility study was published in 2022³².

Oregon is located in the Pacific Northwest region of the United States, and is generally divided into six major agroclimatic areas ⁶⁰. St. Helens, OR is located in the western portion of the state in the Willamette Valley region, which is characterized as having cool, wet winters and warm, dry summers⁶¹. Air temperatures are generally mild due to proximity to the Pacific Ocean, with mean temperatures averaging 3.3° C in January and 19° C in July⁶⁰. Average annual precipitation ranges from 89 to 160 cm⁶².

In this study, we sampled during summer, early fall and late fall seasons. Summer sampling captured the warm, dry season, when there was no rain and the average temperature was 20° C (±0.1° C). Early fall captured the Willamette Valley’s transition from warm/dry to cool/wet. During early fall sampling there was 5 cm of precipitation, and the average temperature was 12°C (±0.1° C). Late fall captured Western Oregon’s cooler, rainy season, with an average temperature of 6.5° C (±0.1° C) and 14 cm of precipitation.

2.2. Field sampling methodology

Deployment of PSDs and seepage meter readings occurred August 23rd through September 23rd (summer) 2019, September 23rd through October 23rd (early fall) 2019, and November 9th through December 9th (late fall) of 2020. Paired air-water and water-sediment PSD deployments averaged 30 days. PSDs and seepage meters were co-deployed at 11 locations along the shoreline of the site (Figure 1).

Figure 1.

a) Samplers were deployed at eleven locations during each deployment. Samples were grouped by region of the site and averaged. Regions of the site include upstream, cove, peninsula and pilings. Yellow pins represent locations where water and sediment porewater passive samplers along with a seepage meter were deployed. Green pins represent the locations where air and water passive samplers were deployed. Blue line represents approximate site boundaries. b) a schematic of diffusive and advective sampling units co-deployed at each site. Components include; passive sampler deployment cages for air (1), water (2), sediment samplers (3) and Dock-Block^® floating platform (4). Seepage meter includes bottomless cylinder (5), tubing (6), water collection bag (7) and anchor weight (8). See SI Figure S1 for photographs of passive sampler deployment cages.

Samplers were deployed at 11 locations for each time period. The center peninsula site was set up as a triplicate set of samples for air and water during two of the three deployments. The 11 individual sites were further grouped by region and named as “upstream” (two sites), “cove” (three sites), “peninsula” (three sites and replicates, a total of five) and “pilings” (three sites). A total of 140 passive samples were deployed at the site. One porewater sample was lost in the late fall time period. The most upstream location is approximately 80 meters upriver of the wood treatment facility (Figure 1).

Advective flow between surface water and sediment porewater was measured using a seepage meter design based on previous studies^36–38. Briefly, the seepage meter consisted of an approximately 32 cm section cut from the top of a metal 55-gallon barrel, connected to a 2-meter-long piece of 2 cm (outer diameter) clear PVC tubing containing a gate valve leading to a seepage meter bag (Figure 1, Figure S1). The 2-meter tubing allows manipulation of the seepage meter bag without disturbing the barrel and surrounding sediment. Clear tubing allows for removal of any air bubbles that will affect seepage readings, and the gate valves allows the system to be closed when removing or connecting bags. The barrel is placed into the sediment at least 5 cm deep forming a “seal” with the sediment and left undisturbed for 24-hours to allow for equilibration. To begin sampling, a pre-weighed, sealed hydration bladder containing 100 to 300 mL of water is attached to the tubing. After 24 hours, the bag is removed and reweighed to measure water loss or gain. Compiled seepage measurements and seepage water flow calculations are provided in SI and Table S1. Environmental concentrations of bioavailable PAHs were measured using low-density polyethylene (LDPE) PSDs described below in Passive sampler preparation.

To measure water-air diffusive flux, air cages containing five LDPE strips were deployed from Dock-Blocks^® one meter above the surface of the water. Below each Dock-Block^®, a water cage containing 5 LDPE strips was deployed one meter below the surface of the water from a steel cable (Figure S1). Dock-Blocks^® were deployed using anchoring systems allowing for movement with tide and water flow, while not disturbing sediment porewater-water PSDs and seepage meters. Sediment porewater-water diffusive flux utilized water cages containing 5 LDPE strips deployed from a steel cable attached to a weight and buoy system (Figure S1). Water cages were deployed about 0.3 meters above the sediment to ensure stability in the water column and avoid contact with the sediment. Immediately next to the water cage, a sediment probe containing 1 strip of LDPE was submerged into the sediment using a slide hammer (Figure S1). The probe measures sediment porewater spanning approximately 1 to 25 cm below the sediment surface. For assessment of water advective flux, seepage meters were placed within 2 meters of the co-deployed water cage and sediment probe PSDs.

2.3. Passive sampler preparation

One-meter passive sampling strips were constructed from additive free LDPE (Brentwood Plastics Ltd) and cleaned with hexanes as previously described^{39, 40}. Prior to deployment, three performance reference compounds (PRCs) were added to the LDPE to allow for calculations of in-situ uptake rates (fluorene-_d10, pyrene-_d10 and benzo[b]fluoranthene-_d12). Analyte sampling rates were then derived from the loss of a PRC with a retained fraction between 10 and 90%. Initial performance reference compound concentrations were determined by averaging three quality control infused LDPE strips (Table S2).

Recovered LDPE samples were transported in sealed amber glass jars in coolers, and stored in the laboratory at −20°C. Air samples were cleaned using isopropanol to remove particulate matter and moisture⁴¹. Water and sediment porewater samples were cleaned using dilute hydrochloric acid (1N) and isopropanol to remove superficial fouling and water⁴². Samples were then extracted in n-hexane and quantitatively concentrated as previously described ⁴¹. All solvents were Optima grade (Fisher Scientific, Pittsburg, PA) or equivalent. Prior to extraction, known concentrations of seven surrogates (naphthalene-_d8, acenaphthylene-_d8, phenanthrene-d10, fluoranthene-_d10, chrysene-_d12, benzo[a]pyrene-_d12 and benzo[ghi]perylene-_d12) were spiked onto samples for calculation of extraction efficiency based on surrogate recovery. Deuterated surrogates were from CDN Isotopes (Pointe-Claire, Quebec Canada). Average extraction surrogate recovery was 87.0% (57.0–117%). Sample extracts were stored at −20°C.

2.4. QC Samples

A standard containing all target analytes was used as a continuing calibration verification (CCV) approximately every 10 samples as well as before and after each batch. The data quality objectives of ±30% was met for all CCVs, sample duplicates and sample overspikes. Quality control samples were collected from the field (trip and field blanks) and the laboratory (PSD construction, lab processing, post-deployment cleaning and reagent blanks) to account for potential background concentrations. QC blank samples were used for sample concentration correction via background subtraction. Concentrations of QC samples used for background subtraction are provided in Table S3. At one of the three peninsula locations, triplicate samples were deployed in air and water to assess field variability (Tables S4-5).

2.5. Sample Analysis

Samples were analyzed for 39 individual parent PAHs^{43, 44}. Analysis was performed with an Agilent 7890A gas chromatograph, Agilent 7000C triple quadrupole mass spectrometer (GC-MS/MS) and Agilent Select PAH column (30m x 250 μm x 0.15 μm). At least a 3-point (3-7) calibration was employed with correlations ≥0.99. Perylene-_d12 was used as the instrumental internal standard. PAH physicochemical properties and instrument parameters are detailed in Tables S6-7. GC-MS/MS data was analyzed using MassHunter Quantitative Analysis v.B.06.00 SP1 build 6.0.388.1 (Agilent Corp. Wilmington, DE) software. Instrument concentrations were surrogate corrected to account for any losses during sample processing in the laboratory.

2.6. Calculations

Time-weighted average concentrations of freely dissolved water, porewater and air vapor-phase PAHs were determined using an empirical uptake model and in-situ sampling rates derived from PRCs as described by Huckins et al.⁴⁵. Detailed equations are provided in Supplementary Information.

Calculations for diffusive flux (ng/m²/d) are expressed in units of mass/area/time as a function of the concentration gradient (dC) across the x-axis (dx) and the molecular diffusion coefficient (D):

$$F_x=\text{D}\left(\frac{dC}{dx}\right)$$ (1)

The molecular diffusion coefficient is determined by both the physicochemical properties of the target analyte and the environmental conditions at the time of sampling. This equation follows the form of Fick’s First Law, where a negative flux represents deposition and a positive flux represents release or volatilization.

Diffusive flux across the water-air interface (F_w-a) follows a Whitman 2-film model as described by Bamford et al. 1999⁴⁶:

$$F_{w-a}=k_{ol}\left(C_w-\frac{c_a}{H^T}\right)$$ (2)

where C_w (ng/m³) is the freely dissolved concentration in water, C_a (ng/m³) is the vapor phase concentration in air, H^T is the unitless Henry’s law constant, corrected for temperature using the van’t Hoff equation and k_ol is the mass-transfer coefficient between air and water. Compiled water-air diffusive flux measures are provided in Table S8.

Diffusive flux between sediment porewater and water (F_pw-w) is expressed through the following equation ^{20, 47}:

$$F_{pw-w}=-\frac{D_w}{{\delta }_L}\left(C_w-C_{pw}\right)$$ (3)

where C_w (ng/m³) is the freely dissolved concentration in water, C_pw (ng/m³) is the freely dissolved concentration in sediment porewater, D_w (m²/d) is the compound specific diffusion coefficient in water and ${\delta }_L$ (m) is the width of the boundary layer. Compiled porewater-water diffusive flux measures are provided in Table S9.

Diffusive flux was only calculated for compounds that were above detection limits in at least one environmental compartment. Analytes below method quantification limits were assigned the method detection limit. Detailed equations are provided in Supplementary Information.

Calculations for advective flux (ng/m²/d) are expressed in units of mass/area/time. The amount of contaminant being transported is a function of its concentration in sediment porewater and the quantity of groundwater flowing⁴⁸. Advective flux between sediment porewater and surface water is described by the following equation:

$$F_{advenction}=C_{pw}.v$$ (4)

where C_pw (ng/L) is the freely dissolved concentration in sediment porewater and v (L/m²/d) is the advective flow of water.

Passive samplers were deployed for 30 days to provide a time-weighted average concentration. Advective flow was measured using seepage meters, and provides a discrete measure across 24-hours. In this study, we collected 24-hour seepage meter readings at the beginning and end of each PSD sampling period, and averaged them together to provide an estimate of advective flow for each season. Compiled advective flux measurements are provided in Table S10.

2.7. Statistical analysis

Propagation of error across the water-air and sediment porewater-water interfaces followed Minick et al. and Liu et al.^{20, 49}, assuming uncertainties of 30% for mass transfer coefficients and 20% for unitless Henry’s law constant. Error in advective flux across the sediment porewater-water interface was calculated in a similar manner, using an uncertainty of 20% in the advective flow measured from the seepage meter based on previous studies^{38, 50, 51}. Detailed equations are provided in Supplementary Information. Propagation of error results for flux measurements are provided in Tables S8-10. Analysis was performed using Microsoft Excel, JMP Pro 15.0.1 and R statistical software.

3. Results and Discussion

3.1. Air-water diffusive flux

Across locations and times, the atmosphere was primarily a source of PAHs to surface water (median and range of sum PAH release: 80 (0.9 – 3400) ng/m2/day, median and range of sum PAH deposition: 5900 (370 – 24000) ng/m2/day). We observed concurrent deposition and volatilization of different PAHs across the study area and time course (Figure 1A). Certain compounds, like phenanthrene, were in deposition during all sampling events while the volatilization of many PAHs decreased over time. Sum volatilization of PAHs was highest in the summer and decreased by more than two orders of magnitude from summer to late fall at each of the locations. At the Upstream and Cove locations during the summer many two and three ring PAHs were volatilizing. However, in the early and late fall these compounds were in deposition throughout the site (Figure 2, SI Figure S6). Previous studies in similar temperate climates have shown that higher temperatures and lower precipitation during summer months can lead to more volatilization of PAHs^{20, 46, 63, 64}. While seasonal trends in air-water diffusive flux are a combination of many factors, temperature, has a large effect on the volatility of PAHs. As air temperatures decreased from summer to late fall the volatility decreased by a factor of 3-15 depending on the compound (Temperature-corrected unitless Henry’s Law constants, Figure S5). Consequently, in early fall the air represents an increasing proportion of the contaminant load at the site. In concordance with the results presented here, McDonough et al.⁶³ also observed the highest rates of phenanthrene deposition in the Lower Great Lakes during the summer, decreasing in the subsequent time period. In the present study, the PAHs with the largest magnitudes of diffusive release across the air-water interface were acenaphthene, naphthalene and pyrene while the compounds with the largest magnitude of deposition were naphthalene, phenanthrene and acenaphthene (Table S8). The majority of four ring PAHs were volatilizing across the study area and seasons, though at a lower magnitude than two and three ring compounds. PAHs with greater than four rings contributed less than 1% to overall chemical movement across the air-water interface. While individual PAHs displayed spatial variability in volatilization, the sum of PAH release was well conserved across the site, varying by less than an order of magnitude across locations within each timepoint. Magnitudes reported in this study are similar to those reported for Lake Superior in Ruge et al.⁵⁷ who reported flux values for phenanthrene from −14000 ng/m2/day to below detection limits and acenaphthene from −5700 ng/m2/day to below detection limits. In the present study, the fluxes of these compounds range from −5400 to −5 ng/m2/day and −1500 to 600 ng/m2/day for phenanthrene and acenaphthene, respectively. However, in the present study many three and four ring PAHs were found to be in release while Ruge et al. reported few PAHs in release and at lower magnitudes.

Figure 2.

Diffusive flux between air, water and sediment porewater across locations and season in ng/m²/d (n=33). Positive values represent release from sediments to water or water to air. Negative values represent deposition from water to sediments or air to water. A total of thirteen samples for each time point are shown, eleven individual sites were composited into the upstream, cove, peninsula and pilings regions for each time point. Error bars represent one standard deviation after propagation of error (SI, Table S8-S9). Panels A-E present a subset of analytes. See SI Figure S6 for diffusive fluxes of all individual PAHs.

3.2. Porewater-water diffusive flux

Sediments were primarily a source of PAHs to overlying water (sum PAH release 2400 to 1.4x10^7 ng/m2/day and deposition from 13 to 7300). With the exception of two samples (Upstream Summer, Peninsula Early Fall), the majority of PAHs were in release from sediments to surface water (Figure S6). We observed greater magnitudes of release at the Cove and Pilings sites (Figure 2). The high rates of release from sediments at these two locations correspond to known areas of contamination where structures were built and activities were taking place when the former creosote site was in operation³³. From summer to late fall the release of PAHs decreased by approximately two orders of magnitude at Cove and Pilings (Figure 2A). These temporal trends were driven by decreasing porewater concentrations of three and four ring PAHs which constitute the majority of diffusive flux between sediments and water (Figure S6). PAHs with greater than five rings had lower magnitudes of release (from <1 to 5500ng/m2/day) compared to three ring PAHs (from <1 to 6.6x10^6 ng/m2/day). The PAHs with the largest magnitude of release across the sediment water interface were three ring PAHs, acenaphthene, phenanthrene and fluorene. The PAHs with the greatest magnitude of deposition to sediments were pyrene, fluoranthene and naphthalene. The movement of three and four ring PAHs like acenaphthene, phenanthrene and fluorene are consistent with PAH profiles of creosote^{34, 56} and diffusive flux results from Minick et al.²⁰ in samples measured under the sediment cap at another legacy creosote facility, the McCormick and Baxter Superfund site on the Willamette River in Oregon.

3.3. Spatial and Temporal Patterns of Advective Flux

Transport of PAHs by bulk water movement across the sediment-water interface occurred in both directions (sum PAH flux from −460 to 2900 ng/m2/day). The advective transport of PAHs was dominated by 3 ring PAHs like acenaphthene, phenanthrene and fluorene (Figure S7). We observed higher magnitudes of advective flux at Cove and Pilings in the summer and early fall as a result of the higher porewater concentrations measured in those locations. The highest advective fluxes were observed during the summer as porewater concentrations reach their maximum (Table S10). Observed changes in magnitude of advective flux across time were driven by changing porewater concentrations rather than changing flow rates. For example, sum PAH porewater concentrations decreased by approximately one and two orders of magnitude from summer to late fall for Cove and Pilings, respectively, while flow rates changed by approximately a factor of two. These results highlight the importance of sediment porewater concentrations in driving chemical movement across the sediment-water interface, through both advective and diffusive processes, at contaminated sites.

While the magnitude of advective flux is primarily driven by porewater concentrations, the direction is determined by the flow of bulk water (Equation 4). At Cove and Peninsula seepage meter flows were higher and typically positive while flows at Upstream and Cove were typically negative (Figure 3D). Across time, flows at Peninsula and Upstream were stable. In contrast, at Cove and Peninsula, the magnitude of flows decreased consistently from Summer to Late Fall (Figure 3D). Seasonal patterns of advective flow may be driven by a variety of factors such as the effect of temperature, precipitation, and runoff on surface water and groundwater³⁷. Previous studies measuring seasonal variability of advective flow have identified many factors including changes in velocity due to temperature changes between groundwater and surface water^{65, 66}.

Figure 3.

A-C). Advective flux between sediment porewater and surface water in ng/m²/d on a base 10 logarithmic scale (n=33). Positive values represent release from sediments to surface water. Negative values represent deposition from water to sediments. Error bars represent one standard deviation after propagation of error (SI, Table S10). D). Flows between surface water and sediment porewater measured with seepage meter.

Advective flux is sensitive to sediment characteristics and topography ⁵², and previous studies have shown that seepage rates can vary by more than a factor of two when seepage meters are installed one meter apart ³⁷. In general, sediment tends to be more heterogeneous than air or water. Previous studies have noted that sediment composition can influence chemical movement, with sandy sediments facilitating increased advective flow in contrast to finer grained, silt and clay sediments where processes are typically dominated by diffusion⁵². While Scappoose Bay contains silt and fine sediments, the underlying bedrock is basalt. The varying depth and permeability of basaltic rock could contribute to the spatial variability we observed in porewater concentrations and flow. Additionally, each sampling location across the entire site was impacted to a different degree by the former creosote operations. In the pilings area of the site, wood and other industrial debris are buried throughout the sediment, which likely influence the flow of water between groundwater and Scappoose Bay. Therefore, the variability in porewater concentrations and flows measured in our study reflect the heterogeneity of the sediments and water dynamics across the site.

3.4. Synthesis of air-water, porewater-water and advective flux

These concurrent measures across the air-water and sediment-water interface demonstrate that Scappoose Bay receives inputs of two and three ring PAHs both from legacy contamination in the sediment porewater and local or regional PAH sources via atmospheric deposition. Minick et al.²⁰ observed similar trends for the Willamette River when measuring diffusive flux between air, water and sediment across the McCormick and Baxter Superfund Site. However, in Minick et al.²⁰, modern activity was found to be a larger source of two and three ring PAHs than legacy pollution, likely due to the closer proximity of the Portland Metro-region and the Portland Harbor Superfund and the Superfund Site’s previous remediation.

Our results demonstrate that both water-air diffusive flux and sediment porewater-water diffusive flux at this legacy contaminated site are strongly temporally influenced. Temporal dependence of chemical movement has important implications for many site assessment evaluations, including stakeholders monitoring the effectiveness of remediation strategies at contaminated sites.

While a small number of studies have utilized passive samplers to estimate diffusive fluxes of organic pollutants, to the authors’ knowledge this is the first study to evaluate the relative contribution of diffusive and advective flux using passive sampler derived fluxes. In this study, advective flux accounted for less than 1% of the parent PAH chemical movement regardless of site or season (Table 1). Sediment porewater-water diffusive flux of sum PAHs was 2 to 4 orders of magnitude larger than advective flux (Figure 2A,3A). Our results demonstrate that in the areas measured in the present study, advective flux is an extremely minor contributor to overall chemical movement across the sediment-water interface. However, bulk transport may play a greater role in other areas of the site along the shoreline where a subsurface creosote plume meets the water table. Understanding the relative contribution of mechanisms of chemical movement can have important implications for policy decisions and remediation. In cases where chemical movement between sediments and water occurs primarily via diffusive flux, passive capping of sites using neutral materials such as sand and crushed rock is usually selected to act as a physical barrier between two environmental compartments. However, when contaminant transfer is driven via advective flux, a passive cap will not contain bulk water^{21, 22}.Therefore, active capping using technology such as organoclay or activated carbon may need to be selected^{53, 54}, imposing additional costs and complexity^{53, 55}. By understanding the contribution of both diffusive and advective flux, responsible parties can optimize the remediation design to be cost-effective and meet clean up goals.

Table 1.

Diffusion as a percentage of total flux across sediment-water interface

	Summer	Early Fall	Late Fall
Upstream	100%	99.9%	100%
Cove	100%	100%	100%
Peninsula	99.9%	99.8%	99.7%
Pilings	100%	99.9%	100%

Highlights:

• PAH movement measured between air, water and sediments via diffusion and advection

• Combined seepage meters and passive sampling devices to measure advective flux

• Atmosphere and sediment found to be predominantly a source of PAHs to surface water

• Diffusive and advective PAH inputs from sediments decreased from summer to fall

• Diffusive fluxes more than two orders of magnitude larger than advective fluxes