3

Problem Formulation: Sources, Settings, and Ecological Receptors

This chapter begins with a brief overview of problem formulation and its use in ecological risk assessment. The chapter then identifies key information that initiates potential environmental exposures to UV (ultraviolet) filters, specifically discussing point and nonpoint sources of UV filters reaching aquatic ecosystems and, to a lesser extent, terrestrial systems. After describing sources of UV filters reaching ecosystems, this chapter outlines conceptual models for potential UV filter exposure in different environmental settings and how these are used to frame the state of knowledge for multiple aquatic environments (to be elaborated upon in subsequent chapters). Next, the chapter discusses exposures to UV filters in the context of ecological receptors (i.e., specific species) and ecosystem services. Gaps in available information are highlighted and summarized at the conclusion of the chapter.

PROBLEM FORMULATION

Problem formulation is an early and essential step in an ecological risk assessment (Figure 3.1). When developing the Framework for Ecological Risk Assessment (EPA, 1992) and subsequent guidance (EPA, 1998), the U.S. Environmental Protection Agency (EPA) placed emphasis on understanding the matter at hand and planning the assessment. EPA (1998) describes problem formulation as “a process for generating and evaluating preliminary hypotheses about why ecological effects have occurred, or may occur, from human activities.” During problem formulation

• the purpose for the assessment is articulated,
• the problem is defined, and
• a plan for analyzing and characterizing risk is determined.

Problem formulation may start with the “integration of available information on sources, stressors, effects, and ecosystem and receptor characteristics.”

The committee is charged with reviewing the fates and effects of UV filters in aquatic environments to provide information on the state of the science for future application in an ecological risk assessment. Problem formulation is the starting point for the committee to ensure that the appropriate information is brought together, reviewed, and organized in a manner that facilitates the eventual conduct of an ecological risk assessment. This chapter of the report helps frame the assessment process by characterizing sources of UV filters, environmental settings where UV filters may enter aquatic environments, and ecological receptors that may be exposed to the UV filters.

SOURCES AND INPUTS OF UV FILTERS INTO THE ENVIRONMENT

UV filter use in various consumer products and personal care products, as discussed in Chapter 2, results in a wide variety of potential routes of entry into the environment that include direct “point sources” (e.g., rinse-off from bathers in surface waters, municipal wastewater treatment effluent, industrial discharges) and more distributed “nonpoint sources” (e.g., stormwater in communities with combined sewer overflows [CSOs], runoff from beach areas, leaching into surface waters from communities with on-site wastewater systems). Different sources can result in continuous discharge or episodic “pulse” inputs, causing variable loading of UV filters into the environment (both aquatic and terrestrial). The following section summarizes the state of knowledge regarding key sources and inputs of UV filters into the environment. While it would likely be difficult to confirm that observed UV filters originated from sunscreens specifically, the exposure routes explored here are those most likely to be relevant to sunscreens.

Direct Release of UV Filters During Recreational and Other Direct Surface Water Contact Activities

Release of UV filters from bathers and other aquatic recreational activities (e.g., swimming, surfing, diving, boating) is explored here as a potentially important input to surface waters. UV filters may also be released into the environment during sunscreen application, particularly if aerosol/spray formulations are used. Application near water (i.e., on the beach) may result in aerosol dispersion with direct deposition onto water, deposition and binding to sand where it may run off into the water during flushing events (i.e., rainfall or tides). For example, one study reports 50 to 60 percent of the sunscreen used during application of spray sunscreen on skin may enter the environment (Broussard et al., 2020). Loading in the environment will be dependent on the amount and type of sunscreen applied, the duration that people spend in the water or rinsing off on the beach (if this water runs to the shore), and the rate at which the sunscreen is rinsed or sloughed off the skin. Collectively, this information can be used for risk modeling in combination with information about the number of bathers, the rates of sunscreen application per bather, and the location specific characteristics.

Sunscreen use, body coverage, and re-application rate (all of which sum to the amount of sunscreen used) varies based upon a broad range of human behavior factors (see Chapter 7). A maximal usage scenario can be described using the American Academy of Dermatology (AAD) recommendation, which recommends the use of 1 ounce (28 g) of sunscreen for the entire adult body and frequent reapplication (i.e., every 2 hours, or more frequently in circumstances of sweating or water exposure). The recommended application of sunscreen is 2 mg/cm² (Petersen and Wulf, 2014). Sunscreen application is recommended over the entire body area, which varies from 1 to 2 m² of skin area for children to adult men (e.g., Mosteller, 1987). Thus, a single application would use 20 to 40 g per person per day, and two applications would use 40 to 80 g per person per day. UV filter content (from Chapter 2) ranges from 3 to 25 weight percent (wt%) of most sunscreens. Taking the inorganic UV filters as an example, which have a maximum of 25 wt%, an upper bound (i.e., maximal use) of UV filter application may be 20 g/day/person. This is complicated by the many combinations of UV filters at various percentages that can compose a sunscreen formulation.

Accurate rates of rinse-off from human skin during aquatic recreational activities vary by the chemical properties of the UV filter, the formulation of the product (Saxe et al., 2021) such as its water-resistance (Stokes and Diffey, 1999), and the application rate. The rates of rinse-off may also depend on the environmental context (e.g., salinity, water temperature, presence of dissolved or particulate organic matter), although research examining these factors has not yet been published to the committee’s knowledge.

Some understanding of the dynamics of rinse-off are essential to accurately comprehend the degree to which direct inputs of UV filters are present in any given aquatic ecosystem. Saxe et al. (2021) proposed a laboratory assay approach for examining the extent of rinse-off for various compounds and formulations. This approach uses a standard skin analog developed in the laboratory for testing purposes and demonstrates that different chemicals rinse off with differing rates, and formulations influence the degree of rinse off. The rates of rinse-off in seawater in this study ranged from less than 2 percent for octisalate, homosalate, and octocrylene to as high as 24 percent for oxybenzone in a lotion. The authors highlighted analytical issues for some UV filters (e.g., a high loss of octocrylene to the glass of test containers; Saxe et al., 2021). If procedural and analytical losses are not corrected for, they may lead to under- or overestimations of these reported release rates. Moreover, as discussed in Chapter 4, the form of the UV filter at the point of rinse-off (i.e., as single discrete organic molecules or as macromolecular assemblages with other formulation components) may significantly impact the environmental fate processes, particularly if any macromolecular associations persist in aquatic environments. Prior to these recent developments of a standard approach for estimating rinse-off rates, and in the absence of detailed information for all of the formulations of currently marketed sunscreens, researchers have assumed that some fraction of applied sunscreen rinses off during aquatic recreational activities.

A number of studies have assumed a fraction, from 20 to 100 percent, of applied sunscreens rinse off directly into surface water (see Sharifan et al., 2016). A controlled laboratory study examining the extent of water resistance retention (WRR) by sunscreen found those claiming to be water-resistant generally have 50 to 80 percent WRR after four immersions in water, as compared to 10 to 30 percent WRR after one immersion by those that did not claim water resistance (Stokes and Diffey, 1999, as reported by Lambropoulou et al., 2002). Applied sunscreen was found to wash off as a whole formula rather than as individual ingredients. Limitations from these results come from uncertainties introduced in real-life situations and differences in water immersions of individual people during bathing activities compared to the laboratory setting (Lambropoulou et al., 2002; Wright et al., 2001).

Some studies have used estimates of rinse-off rates, in combination with the number of swimmers visiting and entering the water at various locations and sunscreen application and reapplication rates, to estimate the amounts of UV filters that enter an aquatic system (e.g., Sharifan et al., 2016). Labille et al. (2020b) estimated fluxes of UV filters to the water at a beach in France where 3,000 beachgoers were present. They estimate that there were 633 and 83 g/day for titanium dioxide (TiO₂) and zinc oxide (ZnO), respectively. They also estimate an additional 2,669, 1,899, 1,789, and 1,073 g/day for avobenzone, octocrylene, octisalate, and homosalate, respectively. The total mass of UV filters was estimated to range from 4,982 to 31,829 g/day. Overall, ranges of inputs for UV filters based on sunscreens applied to the skin could be modeled using a worst-case scenario (based on AAD application/reapplication rates) as an upper-bound to more data-informed releases based upon demography, behavior, water

chemistry, activity type/level, and sunscreen formulation. However, predictive measures of the latter releases do not appear to be well developed in the literature.

In addition to direct rinse-off of UV filters from human skin, UV filters absorbed through the skin and eliminated through excretory processes can be present in human urine (Bury et al., 2019; Hiller et al., 2019a,b; Huang et al., 2019; Janjua et al., 2004, 2008; Kang et al., 2019; Klotz et al., 2019; Li et al., 2019b). Thus, urination in water bodies may present an additional source to aquatic environments during recreation. Rates of urination while swimming have been well studied in swimming pools (Zheng et al., 2017). While some human urine biomarkers and pollutants readily degrade, other chemicals present in human urine (e.g., artificial sweeteners like sucralose) can be used to estimate the amount of urination occurring in a water body because they persist for long durations in water (Tovar-Sánchez et al., 2013). Upper limits for urinary concentrations of benzophenone (BP)-type UV filters are on the order of 25 μg/L (Adoamnei et al., 2018), and upper ranges of daily urination volumes are 2 L/day. Thus, daily excretion of BP-type UV filters may have an upper range of 50 μg/person/day. Thus, human urine inputs of UV filters appear to be orders of magnitude lower than direct rinse-off from human skin during bathing.

The cumulative input of sunscreens during aquatic recreation have been studied in a few surface waters, which have associated UV filter measurements with estimates of the numbers of recreational users in the area. Some studies have correlated UV filter presence to seasonal patterns of recreational activity (Bargar et al., 2015; Poiger et al., 2004; Tsui et al., 2017) and others have correlated it to distance from shore and recreational activity (Bachelot et al., 2012; Mitchelmore et al., 2019; Schaap and Slijkerman, 2018). For example, a study of two Swiss lakes found higher concentrations in lake water during summer months, when recreation and UV filter use was presumed to be highest. In addition, this study found that the pattern of the occurrence of UV filters was the opposite of contaminants known to be associated with sewage, making wastewater a less likely source of the UV filters (Poiger et al., 2004). In another study, the U.S. Geological Survey (USGS) sampled water concentrations of UV filters in the U.S. Virgin Islands to investigate the influence of direct inputs to surface waters via recreation (Bargar et al., 2015). In the bays sampled, UV filters were detected more commonly in June (when swimmers were present) than in April (when fewer swimmers were present).

In Lac Bay, Bonaire, the measured concentrations of UV filters declined with distance from recreational bathing zones (Schaap and Slijkerman, 2018). Similarly, in Trunk and Hawksnest Bays, U.S. Virgin Islands, concentrations of a number of UV filters sampled declined as distance from the shoreline increased. UV filters were also present along the coastal areas of the U.S. Virgin Islands, despite the dilution of the surrounding ocean waters (Bargar et. al., 2015). In addition, using mussels as sentinels for pollution patterns, researchers found a correlation between increasing concentrations of UV filters in mussel tissues and recreational activities (Bachelot et al., 2012). These results suggest that rinse-off directly from human use of sunscreens is a dominant source of UV filters to aquatic environments.

Other studies (Mitchelmore et al., 2019; Tsui et al., 2017) have also looked at relationships between UV filter occurrence in surface waters and the number of people actively recreating in the water. Diurnal patterns in UV filter concentrations in surface waters provide additional evidence of the direct inputs of sunscreen from recreational bathing (Tovar-Sánchez et al., 2013). The diurnal patterns in concentrations, which demonstrate increased concentrations during the day versus the nighttime hours, suggest direct rinse-off as a source of UV filters to surface waters and show that concentrations in a particular location are highly variable. Tsui et al. (2014b, 2017) also highlighted spatial differences within the water column as concentrations were observed to be 40 times lower in samples at depth than from the surface, although it should be noted these samples were collected at different times. Conflicting and little evidence is available on UV filter concentrations at the surface microlayer, which is important to investigate given that sensitive life stages of organisms may be present close to this layer (Mitchelmore et al., 2021).

Rinse-off of UV filters has been studied in contained water systems, namely swimming pools. Water from swimming pools is periodically replaced, with discharge of water to sewers, drainages, or other outdoor locations. They also can provide useful study systems to understand sunscreen wash-off from swimmers. A study of TiO₂ from sunscreens in swimming pools found total titanium concentrations [Ti] ranged between 21 μg/L and 60 μg/L (Holbrook et al., 2013). Concentrations increased throughout the sampling period though concentrations in tap water (water supply) remained constant. Most (82 to 98 percent) Ti was found in the dissolved phase (< 1 kDa),

and both total and microparticulate forms were higher following pool use. The authors found that experimental analysis and simple models suggest that sunscreen may be an important source of particulate and microparticulate Ti. The authors also noted that pool-system filtration systems may have removed some of the particulate TiO₂. Thus, unless such water treatment technology is specifically considered, this may influence the utility of swimming pool studies as a means to estimate UV filter input rates in outdoor aquatic environments.

Occurrence measurements of UV filters in water have been made using releases from bathers into water plus biogeochemical transformations and physical hydrologic mixing or dilution. Overall, the combined approaches of estimating inputs and validating concentrations in aquatic systems with well-defined hydrology and loading can provide case study locations to validate release rates as well as UV filter fate and transport mechanisms.

Direct Release of UV Filters from Stormwater and Wastewater

As described in Chapter 2, sunscreen is only one product type that contains UV filters and other product uses may also contribute to UV filters in stormwater and wastewater, which can be centralized or on site. Usage in other personal care products (shampoos, body washes, hair sprays, fragrances, toothpastes) will likely have the most similar fate to UV filters from sunscreens, though other product usages can also make their way into wastewater, such as UV filters used in synthetic resins, inks, paints, and coatings (Keller and Lazareva, 2014). Information to trace UV filters that enter stormwater and wastewater systems only from sunscreens and determine their percentage contribution to the total loading was not available. Instead, this section describes the transport and abatement of UV filters in stormwater and wastewater, with the presumption that sunscreens will be a contributor, given known sources and uses, but not the only contributor.

Centralized Stormwater and Wastewater Sewer Systems

There are three types of sewer systems: sanitary sewers, stormwater sewers, and combined sewers. Sanitary sewer systems collect wastewater from private residences, commercial buildings, industrial manufacturing facilities (with and without on-site treatment requirements prior to discharge), and other buildings and they convey wastewater by gravity and occasional pump stations to centralized wastewater treatment facilities. Municipal wastewater from homes, hotels, and other buildings may contain UV filters due to sunscreen rinse-off from shower/bathing (due to use of sunscreen for daily activities, outdoor recreation, or remaining on the skin after aquatic activities), rinse-off during shower/bathing of other personal care products containing the UV filters, and/or discharges from manufacturers/industry that incorporate UV filters into products. In most communities, it is typical to have 5 percent to 20 percent of the wastewater coming from industrial manufacturing, with the remaining percentage being equally distributed among private residences and other sources. Concentrations of UV filters in sanitary sewer water will be presented later in the chapter. Some amount of infiltration (i.e., rainwater, groundwater) and exfiltration (i.e., leakage from sewers into groundwater) can occur during conveyance (Nguyen et al., 2021), and may serve as a source of UV filters into sewers or source of release into the environment. Leakage and discharge/overflow from sewers can be quite important; one study in Pittsburgh, Pennsylvania, found that 70 percent of a stream length was located within a potential sewer leakage “hotspot,” and these “hotspots” correspond to older sewer pipe sections (Hopkins and Bain, 2018). During a single storm event on Oahu, Hawaii, heavy rains resulted in over 322,000 gallons of wastewater discharge from at least nine sewers (Hawaii Department of Health Clean Water Branch, 2021).

An additional potential source of UV filters to wastewater treatment facilities may be the manufacturing facilities that produce and use these compounds in product formulations. To our knowledge, this has not been assessed for UV filters, but in the case of other traditional contaminants and contaminants of emerging concern, manufacturing facilities can be a hotspot in terms of environmental concentrations (e.g., Anliker et al., 2020; Scott et al., 2018). More research to understand this potential source to the environment is needed to ascertain whether this is a significant source of UV filters to the environment.

Stormwater sewers can receive rain or snow runoff from streets, lawns, parking areas, and buildings. Whether UV filters might be present in stormwater from sunscreens specifically is completely unknown; theoretically, they could be present if sunscreens are included in discarded and mismanaged waste or in runoff in recreational areas

(e.g., beaches, athletic fields). A few studies have measured inorganic UV filters in stormwater (e.g., Baalousha et al., 2020), and no studies have systematically measured stormwater concentrations of organic UV filters. However, there are other sources of inorganic UV filters that contribute to stormwater loads and would complicate attributions to sunscreens, namely building facades, other surface coatings, and concrete (Al-Kattan et al., 2013; Boonen and Beeldens, 2014). Overland flow across grasses and soil, on-site catchment, and other required stormwater mitigation measures could trap UV filters before entering sewers or receiving waters (Baalousha et al., 2013), though hydrophilic (low log Kow) UV filters may not be as readily removed (Spahr et al., 2020). Stormwater sewers convey runoff to receiving waters (e.g., creeks, ponds, lakes, rivers, estuaries, ocean outfalls), most commonly without any treatment.

The final type of sewers are “combined sewers,” which occur in some communities but has not commonly been used in urban water infrastructure since passage of the Clean Water Act, collect runoff, domestic sewage, and industrial wastewater into one pipe. Under normal conditions, all of the wastewater is transported to a centralized sewage treatment plant. However, the volume of wastewater can sometimes exceed the capacity of the system or treatment plant (e.g., during heavy rainfall events or snowmelt) and thus untreated stormwater and untreated wastewater discharges directly to nearby streams, rivers, estuaries, and other water bodies as combined sewer overflows (CSOs). EPA estimates that this results in about 850 billion gallons of untreated wastewater being discharged into water bodies per year from CSOs (EPA, 2004b). The committee found no data on UV filter concentrations in CSO water, but as with on-site wastewater systems, it may be possible to make some assumptions based on pharmaceutical or personal care product (PPCP) data that act as surrogates for UV filters in wastewaters. For example, although CSO discharges represented only 10 percent of the total water discharged (CSO plus treated WWTP effluent discharges) over a 13-month period in Burlington, Vermont, CSOs contributed 40 to 90 percent of the annual load of wastewater-based micropollutants (i.e., including PPCPs) that can be well treated (i.e., > 90 percent removal within the WWTP) and 10 percent of the annual load for micropollutants with lower removal efficiencies. While sewage flows are relatively constant, stormwater flows are highly variable and micropollutant concentrations decrease with increasing flow because rainfall runoff dilutes the sewer flows. CSOs threaten inland and coastal waters from both pollutants contained within the CSOs and from hydraulic shock from the increased volume of water (Marsalek et al., 1998). Overall, the impact of CSOs on the environment will vary by location (Petrie, 2021).

Treatment and Discharge of Municipal Wastewater

Nearly 80 percent of the U.S. population is served by centralized municipal sewers, which collect water from residential, commercial, industrial, and other building types (EPA, 2004a). Collected sewage is treated on-site or at centralized wastewater treatment plants (WWTPs; Table 3.1), which use a range of lower- to higher-technology based processes to achieve variable levels of regulation-driven water quality standards based on the size, use, and classification of the receiving surface waters (lakes, rivers, ocean) that receive the treated wastewater. While most treated wastewater is discharged to surface waters, in some regional cases treated wastewater is infiltrated or injected into groundwater, used in industrial cooling facilities, evaporated from lined ponds, or applied to agricultural crops. In addition to the liquid effluent discharge, sewage solids (biosolids) are also produced, which may contain many of the influent chemicals or degradates of them. Across the United States, roughly 55 percent of sewage solids are land applied, 30 percent are disposed of to landfills, or incinerated (about 15 percent) (EPA, 2019a,b).

The sewage solids applied to agricultural crops or pelleted are used as fertilizer or soil amendment, and there is potential risk of chemicals present within sewage solids leaching into aquatic systems (e.g., Ramos et al., 2021). Currently there are no regulations for the levels of UV filters in WWTP effluent discharges or sewage solids.

Where not served by centralized or regulated WWTPs, the remaining portion of the U.S. population primarily relies upon on-site septic systems or cesspools plus leaching fields into shallow groundwater for disposal of sewage from homes, individual buildings, and (occasionally) groups of dwellings. The costs of sewer piping, pumping, and associated infrastructure is often a limiting factor for communities that have decentralized or on-site collection systems and treatment, with lower population density and more sparsely spatially distributed homes being more costly to install sewer systems to collect wastewater. Consequently, centralized wastewater treatment is more

TABLE 3.1 Common Wastewater Treatment Systems and Discharge Locations

^a Each system type will also produce bacterial sewage solids of some form; these are continuously produced and are disposed of daily, monthly, or annually depending upon size and type of facility. Common disposal options for sewage solids are landfills, land application, or incineration.

common in more densely populated communities. Regulatory oversight of on-site septic systems varies by state and jurisdiction, ranging from inspection during construction or sale of the land on which the system is located to periodic sampling of water quality or maintenance (usually less frequently than annually).

The following sections describe the potential removal of UV filters in effluent through the different types of wastewater treatment systems, starting with on-site treatment systems and then progressing to larger community-based treatment systems. In most pollutant exposure scenarios, wastewater discharges occur as point source releases of water into the environment. These discharges are continuous throughout the year, but exhibit diurnal and seasonal fluctuations based on water use patterns. In some scenarios (e.g., large numbers of septic systems located along a small creek) contributions of water and pollutants may be considered as a “nonpoint” or distributed sources of pollutants along the length of the river.

On-site Wastewater Treatment Systems

Dwellings with on-site wastewater treatment are regulated by local and state agencies, which often rely on building codes to design, construct, and install below-ground treatment systems and local ordinance to inspect, repair, and clean below-ground treatment systems upon selling and transfer of property rights. Figure 3.2 shows the distribution of frequency of use of on-site wastewater systems across the United States as of 1990, the last time the U.S. Census collected such information. The distribution and density of septic systems varies by state from 9.8 percent (California) to 55 percent (Vermont) (EPA, 2002; U.S. Census Bureau, 1990). A more recent survey by the U.S. Census Bureau found that about 19 percent of U.S. homes used septic systems in 2011 (U.S. Census Bureau, 2013).

Septic system failure rates can be frequent. A survey reported by the U.S. Census Bureau in 1997 estimated that 403,000 homes experienced a septic system breakdown in a three-month period, and an EPA review reported failure rates ranging from 10 to 20 percent, but generally information about septic failures is lacking (as reported in EPA, 2002). Failure is common in part due to aging or underperforming technology.

Common on-site wastewater treatment systems include septic systems or cesspools. Figure 3.3 shows the general configuration of a septic system that includes a fully contained tank where bacteria degrade pollutants, larger particles settle, and oil/grease rise to the surface. Leaching fields distribute flow over a large area, where water slowly percolates through the soil. Figure 3.4 contrasts septic system design to cesspools that usually have a perforated tank and infiltrate into the groundwater without a leaching field. Capacity and maintenance of both types of systems influence their performance to remove pollutants. For a properly sited and functioning on-site

septic system, UV filters would be generally well removed (likely through sorption and degradation). However, multiple systems are not properly sited or functioning, potentially resulting in additional sources of UV filters into groundwater. For example, in Hawaii, there are large numbers of smaller on-site cesspools that serve small numbers of people (Figure 3.5). Approximately 19 percent of homes have on-site wastewater treatment systems in Hawaii (U.S. Census Bureau, 1990). Here, cesspools have been identified as a risk of pollution to groundwater, surface water, and the ocean due to release of pathogens (e.g., bacteria, protozoa, and viruses that can cause gastroenteritis, hepatitis A, leptospirosis, salmonella, and cholera), an array of chemical contaminants, and nutrients that stimulate the growth of harmful algae (Hawaii Department of Health, 2022).

There are a limited number of published studies on the occurrence of UV filters in septic, cesspool, and groundwater systems. Table 3.2 provides information from one study on detected concentrations entering the onsite treatment systems, removal potential by different types of on-site treatment process, and possible mechanisms for their removal. Seven UV filter chemicals were detected at 32 on-site wastewater systems at concentrations of < 1 to 146 μg/L, and up to 4.9 μg/L of oxybenzone were observed in gray water effluents (Leal et al., 2010). The effluent from these washing activities can enter septic systems in on-site treatment systems. Separately, in a review considering many different personal care products from 20 distinct studies of conventional versus alternative onsite treatment systems (e.g., sand filters, aerobic treatment units, wetlands, peat-based systems, and biofilters), oxybenzone concentrations in the effluents were reported; median (maximum) values for septic systems (n = 8), leaching fields (n = 1), or alternative on-site systems (n = 12) were 8 (16) μg/L, 3 (6.3) μg/L, and 0.2 (1.2) μg/L (Schaider et al., 2017 with data from Matamoros et al., 2009). They reported 41 to 67 percent removal of oxybenzone in drainfield systems and higher removal (82 to 99 percent) in alternative treatment systems. The presumed

mechanisms for removal of the UV filters varied by chemical structure (Table 3.2). On-site wastewater treatment systems are often designed to have both oxygenated (aerobic) and very low oxygen containing (anaerobic) zones within their design to support growth of different types of bacteria that are capable of degrading different nutrient species, pathogens, or other chemicals. For the broad range of trace organic chemicals, removal efficiencies during on-site wastewater treatment observed no consistent relationship with values for hydrophilic chemicals (log Kow < 4), but higher removal efficiency (about 90 percent) for organic chemicals with values ≥ 3.5 (likely due to enhanced sorption and subsequent removal on solids). Comparatively, for organic chemicals with values < 3.5, the removal efficiency was lower (60 percent). Chemical properties (e.g., pKa, log Kow) and predictions for relative rates of biodegradability have been applied for removal of trace organic chemicals in on-site wastewater systems (Schaider et al., 2017) and may serve as a strategy to predict removal percentages of organic UV filters.

While few studies have measured occurrence, removal, and release of organic UV filters themselves through septic systems, there is compelling evidence that other PPCPs occur in septic systems at concentrations much above those at municipal wastewater treatment plants. In the absence of other information, PPCPs are used here as surrogates or “markers” for potential treatment and off-site migration of UV filter chemicals from on-site systems. This can lead to insights into potential impacts of on-site wastewater treatment systems on freshwater and marine environments. A review of studies of removal of organic contaminants from on-site systems found that removal varied widely and that septic systems are potentially of greatest concern in areas with dense residential development when septic systems are common, in sand and gravel aquifers with shallow water tables, and where nutrient

TABLE 3.2 Assessment of UV-Filter Removal from Grey Water

^a removal higher than 80% (+ +), between 50% and 80% (+), less than 50% (−), possible removal mechanisms.

DATA SOURCE: Leal et al., 2010.

contamination from septic systems is known to affect surface water quality (Schaider et al., 2017). For example, on a densely populated but seasonally occupied recreational island in New England the groundwater samples collected near an extended health care facility that relies upon a single large septic system and near many small septic systems found PPCP micropollutant concentrations totaling 1 to 20 μg/L, including in shoreline wells (Phillips et al., 2015). A separate study near headwater streams in the Colorado Rocky Mountain region found 88 percent of on-site wastewater treatment systems sampled contain one or more of 24 PPCPs, with nonresidential sources generally having higher concentrations than residential sources (Conn et al., 2006). Removal efficiencies varied from < 1 percent to > 99 percent based upon the PPCP and type of biological treatment. They estimated PPCP loadings on the order of 20 mg/m²/day into the ground. Other studies showed on-site wastewater treatment contributed to measurable concentrations of PPCPs in coastal groundwaters along the east coast of the United States (Swartz et al., 2006). These studies demonstrate linkages between use of personal care products in homes, which enter septic systems and then are detectable in groundwater, and their recharge into surface or estuary waters. It is reasonable to presume that similar pathways and sources of UV filters are occurring, although detailed studies are lacking.

Removal mechanisms differ for inorganic UV filters during biological treatment and passage through the subsurface. Aggregation and some internalization within bacterial biofilms of TiO₂ and ZnO will occur (Kiser et al., 2012). ZnO will also slowly dissolve and release zinc ions. However, under anaerobic conditions ZnO may also transform through reactions with iron (Gomez-Gonzalez et al., 2019) or form a sulfide shell-like layer on the ZnO to make it insoluble in water (Abbas et al., 2020; Lombi et al., 2012). The rate and extent of ZnO dissolution depends upon pH, oxygen, sulfur, and other conditions present in water (see Chapter 4). Removal of ZnO during wastewater treatment is less frequently reported than other nanoparticles (e.g., copper or silver based; Bolaños-Benítez et al., 2020; Liu et al., 2018; Taylor and Walker, 2016). Studies on ZnO have addressed their potential detrimental effects on normal functioning of biological wastewater systems (e.g., nitrogen removal) due to ZnO toxicity to certain classes of bacteria (Zheng et al., 2011). In other work, minimal effects of ZnO were reported, and laboratory studies to simulate wastewater treatment observed about 70 percent removal of ZnO during sedimentation and near complete removal from the effluent during simulated biological treatment (aeration + sedimentation cycles) (Hou et al., 2013). Consequently, there was little adverse impact on nitrogen removal in the wastewater system. TiO₂ in nanoparticle form also had minimal adverse effects on wastewater treatment efficiencies, and accumulated into settled biological solids (Qiu et al., 2016). After exposure in biological processes, on-site wastewater treatment usually involves infiltration through soil. TiO₂ particles have been experimentally shown to be retained in porous media, indicating the likelihood of significant retention of particles in soil before reaching groundwater (Waller

et al., 2018). However, they noted that surface functionalized TiO₂ may behave differently and may migrate into groundwater; the surface coatings help with emulsifying the product and preventing the skin-damaging, reactive oxygen species from being produced. TiO₂ is unlikely to dissolve in septic systems (or most other environmental water matrices) and requires intensive acid and heating to degrade into titanium ions during chemical analysis (Peters et al., 2018). This research suggests that inorganic UV filters are likely to be effectively retained within on-site biological treatment systems or removed from the effluent in subsequent soil infiltration or leaching fields and remain in the sewage solids. There is insufficient research available to make similar assessments regarding organic UV filters.

Overall, septic systems can be locally important sources of pollutants to adjacent aquifers and surface waters, indicating this may be the case for UV filters as well. However, there is a significant data gap in measurement of concentrations, removal efficiencies, and off-site groundwater migration assessments for the broad range of organic UV filter chemicals. Poorly maintained on-site treatment systems can increase the release of UV-filters into groundwater, and subsequently impact nearby surface waters.

Centralized Wastewater Treatment

Centralized wastewater treatment plants (WWTPs; or publicly owned treatment works) treat water from sanitary sewers and a portion of the flow in combined sewers. As suggested by Table 3.1, a wide range of different types of processes are employed at WWTPs, which vary based on the population served and local effluent discharge limits. For example, in Hawaii, there is a small number of centralized wastewater collection and treatment systems that serve large numbers of people. Of Hawaii’s 199 regulated wastewater treatment systems, roughly 25 percent are publicly operated¹ and approximately 80 percent of the population may be served by these regulated facilities (EPA, 2002). Treated wastewater from these facilities are discharged to, or enter, freshwater (surface and groundwater) as well as the ocean, sometimes including pipelines to discharge treated wastewater further away from the coast. While regulated for numerous biological and chemical pollutants, there is no regulation of UV filters in the treated wastewater.

Natural treatment processes (e.g., constructed wetlands) have low capital and operating costs. Generally, as the number of people served by a WWTP increases, more intensive and engineered treatment processes are employed. However, the common feature among nearly all WWTPs is reliance on biological treatment processes to degrade organic matter, which produces bacterial biomass that can adsorb nonbiodegradable materials. Biomass separated from water through sedimentation and/or filtration is called sewage solids. After biomass separation, water is disinfected prior to discharge. Micropollutants, including organic UV filters, can be volatilized, biotransformed, adsorbed to solids that settle out of the water and produce sewage solids, or oxidized during wastewater treatment. Figure 3.6 depicts the treatment process for removal of titanium from wastewater, though this process is applicable to other UV filters as well.

Estimates of micropollutant removal efficiency at WWTPs can be readily predicted by models that rely on physical and chemical properties (e.g., EPI Suite™ from EPA, 2022a; Tran et al., 2018). The extent of removal depends on the affinity of the chemical for solid phases (i.e., as approximated by log Kow values) as well as the biodegradability of the chemical and the type of biological and other treatment processes employed at the WWTP, and thus impacts the concentrations of UV filters in treated wastewater effluent discharged into the environment (e.g., lakes, rivers, estuaries, oceans, and occasionally groundwater or evaporation basins). Available estimates of removal from wastewater treatment for the UV filters approved for use in sunscreen in the United States are shown in Table 3.3.

An EPA standard method (OPPTS 835.1110) for estimating soluble pollutant removal during wastewater treatment is widely used to estimate adsorption of organic pollutants on sewage solids. This method uses freeze-dried instead of fresh, hydrated sewage solids. Because of the important role of lipid bilayers and other physical properties of sewage solids, EPA methods to predict removal of TiO₂ and ZnO nanoparticles were found to be

___________________

¹ See http://geoportal.hawaii.gov/datasets/wastewater-treatment-plants/explore.

TABLE 3.3 Predicted Removal of Nondeprotonated Forms of Organic UV Filter from Wastewater Treatment

NOTE: EPI Suite™ has not been updated or validated to predict the fate of particles in WWTPs for 2 organic UV filters (ecamsule and trolamine salicylate) and 2 inorganic UV filters (titanium dioxide and zinc oxide). SOURCE: EPI Suite™ from EPA, 2022a (data accessed November 2021).

more representative when fresh (not freeze-dried) sewage solids were used. Freeze-dried sewage solids underpredicted the potential removal of TiO₂ and other inorganic colloids/nanomaterials (Kiser et al., 2012; OECD, 2021).

UV filters that adsorb onto biomass (i.e., sewage solids) are removed from the water, and their fate is determined by the sewage solids disposal methods and use. Sewage solids from WWTPs are most commonly disposed of to landfills (30 percent), land applied (55 percent), or incinerated (about 15 percent) (EPA, 2019a,b). Organic chemical contaminants are often highly concentrated in these sewage solids. Land application, often on agricultural fields, may result in uptake by plants and animals, soil contamination, or percolation through soil contaminating groundwater and/or entering other aquatic systems through land runoff (e.g., wind and rain events). These solids will contain organic and inorganic UV filters and other pollutants (Heidler and Halden, 2008; Westerhoff et al., 2015). Figure 3.7 summarizes UV filter chemical occurrence data in WWTP sewage solids (note that it includes UV filters not approved for use in sunscreens in the United States) as well as influent and effluent. Leaching of

pollutants from sewage solids that are land applied has been well studied. Inorganic UV filters tend to accumulate in soils beneath where they are land applied (e.g., Labille et al., 2020a; Yang et al., 2014).

Even low concentrations of UV filters can result in significant daily or annual loads. Typical wastewater flow rates per person served in a community is on the order of 50 to 75 gallons/capita/day (190–280 L/capita/day). Loads can be calculated by multiplying UV filter mass per person by a per capita wastewater volume and number of people served by the centralized WWTP. For example, one review found that oxybenzone was detected in wastewater from six countries (WWTP influents: 7–3,975,000 ng/L; WWTP effluents: 1.1–2,196,000 ng/L) (Kasprzyk-Hordern et al., 2009; Magi et al., 2013; Trenholm et al., 2008 as reported by Montes-Grajales et al., 2017). In another study, oxybenzone was present in raw and treated wastewater in Switzerland at concentrations up to 7,800 ng/L and 700 ng/L, respectively (Fent et al., 2008). Oxybenzone can adsorb and accumulate in sewage solids (Gago-Ferrero et al., 2011b; Rodil et al., 2009). A recent study found a commercial sewage solid containing 3,404 ± 92 ng/g representing a sum of 6 UV filters, with octocrylene and drometrizole trisiloxane the major contributors at 1,519 ± 22 and 1,629 ± 54 ng/g dry weight, respectively (Ramos et al., 2021). Sewage solids (i.e., sludge) are mostly bacterial cells, onto which UV filters (parents or degradates), or other contaminants, plus associated mineral precipitates and other particulates are brought into WWTPs with the wastewater.

In another comprehensive study, concentrations of oxybenzone and its derivatives (including dioxybenzone) in influent and emissions (as sludge or effluent into a river) were investigated in two WWTPs in the Albany area of New York State (Wang and Kannan, 2017). The median concentrations of oxybenzone and sum of its four derivatives in influents were 0.035–0.049 and 0.124–0.145 μg/L, respectively. The authors found biodegradation to be the main pathway of removal for oxybenzone and its derivatives, with 75.2 to 82.5 percent of the chemicals transformed or lost during treatment and 13.2 to 15.7 percent sorbed to sludge, and it appeared a higher proportion of oxybenzone was transformed compared to the derivatives. The daily mass loadings (25.7–81.4 and 76.1–194 mg/d/1,000 people) and environmental emissions (10.5–17.5 and 44.5–76.2 mg/d/1,000 people) for oxybenzone and its derivatives in WWTPs were also estimated. The authors estimated that approximately 11 percent and 20 percent of the total production of oxybenzone and benzophenone-1 (a first-formed metabolite) in the United States reached WWTPs, while percent and 15 percent of the loaded amounts were emitted through WWTP discharges. These removal ranges are supported by other research reporting > 80 percent removal of oxybenzone when sufficient sludge retention time is incorporated into biological WWTP processes (Oppenheimer et al., 2007).

A national environmental risk assessment of oxybenzone in the United States focused on the role of WWTPs on releases of the UV filter into freshwater systems (Burns et al., 2021). Using an estimated 71 to 86 percent removal across WWTPs based on 10 studies, mostly due to biodegradation, the authors estimated a 90th percentile exposure concentration of 150 ng/L in streams receiving wastewater effluents. UV filter occurrence data in the United States are otherwise limited, compared with more widely reported data in other countries (Ramos et al., 2016). Liu et al. (2012) measured oxybenzone concentrations of 2,086 ± 1,027 ng/L in influent, and up to 153 ± 121 ng/L in effluent, respectively in WWTPs in South Australia. In the sewage solids, octocrylene occurred as one of the highest concentrations at 465 ± 65 ng/g. Sorption onto sewage solids played a dominant role in the removal for UV filters, especially for two non-U.S. UV filters and octocrylene, which accounted for 54 to 92 percent of influent loads. O’Malley et al. (2020) measured the annual release of select UV filters through effluent from wastewater treatment plants in Australia. The authors detected UV filters in 96 percent of effluent samples; in addition to non-U.S. UV filters (4-methylbenzylidene camphor and benzophenone-1), ensulizole, oxybenzone, and sulisobenzone were detected above the limit of quantification (LOQ) and octocrylene and dioxybenzone were below the LOQ.

In a study of five WWTPs in Hong Kong, avobenzone, benzophenone-1 (an expected degradate; see Chapters 4 and 5), oxybenzone, sulisobenzone, octinoxate (out of 12 UV filters included in the study) were detected in all influent samples from the four largest WWTPs (mean concentrations ranging from 23 to 1,290 ng/L) (Tsui et al., 2014a). They were frequently detected in effluent with mean concentrations from 18 to 1,018 ng/L. Higher concentrations in this, and other studies (Balmer et al., 2005; Cunha et al., 2015; Li et al., 2007; Magi et al., 2013; Pedrouzo et al., 2010 as reported by Ramos et al., 2016), were found during the wet (summer) season, except for sulisobenzone, which was the most abundant compound detected in Tsui et al. (2014a). Fewer than 20 percent of compounds experienced high removal (> 70 percent) when undergoing primary treatment and > 55 percent of compounds experienced high removal with secondary treatment; reverse osmosis resulted in the highest removal efficiencies (Tsui et al., 2014a).

Although not currently being widely considered nor regulated in the United States, other countries are starting to consider adding additional physico-chemical treatment processes at the end of existing biological processes used in WWTPs for the specific reason to remove pharmaceuticals, personal care products, or other trace organics (Völker et al., 2019). The use of ozone chemical oxidation processes or adsorption by powder activated carbon are now being integrated into final treatment processes at WWTPs to remove trace organics (Eggen et al., 2014). Whereas ozonation transformation by-products may pose ecological risk, overall trends appear to be reduced adverse effects upon ozonation of wastewaters (Völker et al., 2019). Ozone’s effectiveness has been studied for a limited number of UV filters (Hopkins et al., 2017; Margot et al., 2013). Powder or granular activated carbon can adsorb trace organics from wastewater (Bourgin et al., 2018) without forming any oxidation byproducts, but only have been demonstrated for a limited number of organic UV filters (Gutiérrez et al., 2021; Margot et al., 2013).

In one study of removal of titanium in WWTPs, concentrations in raw sewage and treated effluent were determined for 10 representative WWTPs within the United States. Raw sewage titanium concentrations ranged from 181 to 1,233 μg/L (median of 26 samples was 321 μg/L). Using common sequence of WWTP treatment processes (i.e., Figure 3.6), the WWTPs removed more than 96 percent of the influent titanium, and all WWTPs had effluent titanium concentrations of less than 25 μg/L (Westerhoff et al., 2011, 2013). Unlike TiO₂, which undergoes minimal dissolution, ZnO can dissolve during wastewater treatment but most of the zinc ions then adsorb or otherwise become incorporated into sewage solids (Ma et al., 2014).

Biodegradability of UV Filters in Wastewater Treatment Plants

In addition to removal onto solids, a key process affecting the ultimate release of organic chemicals from WWTPs to receiving waters is the potential for biodegradation within the WWTP. Thus, assessment of biodegradability—particularly under conditions representative of WWPs—is commonly employed for many organic chemicals.

Studying chemical biodegradation under conditions of wastewater treatment is important to regulatory definitions of environmental persistence. Standard laboratory studies using benchtop WWTP reactors have been used in chemicals legislation to address biodegradation potential of chemicals. A number of definitions, largely related to the extent of degradation and mineralization, are used in this context. OECD test guidelines distinguish six forms of biodegradation, as follows (OECD, 1981b, 1991, 1992a,b, 2001, 2002, 2004a, 2008) (from ECETOC, 2013, Technical Report 123):

1. Ultimate biodegradation (mineralization): The level of degradation achieved when the test compound is totally utilized by microorganisms resulting in the production of carbon dioxide, water, mineral salts, and new microbial cellular constituents (biomass).
2. Primary biodegradation (biotransformation): The alteration in the chemical structure of a substance, brought about by biological action, resulting in the loss of a specific property of that substance.
3. Ready biodegradable: An arbitrary classification of chemicals which have passed certain specified screening tests for ultimate biodegradability; these tests are so stringent that it is assumed that such compounds will rapidly and completely biodegrade in aquatic environments under aerobic conditions.
4. Inherent biodegradable: A classification of chemicals for which there is unequivocal evidence of biodegradation (primary or ultimate) in any test of biodegradability.
5. Half-life (t0.5): The time taken for 50 percent transformation of a test substance when the transformation can be described by first-order kinetics; it is independent of the initial concentration.
6. Disappearance time 50 (DT50): The time within which the initial concentration of the test substance is reduced by 50 percent.

The most commonly used test guideline is that addressing ready biodegradability in the series of design options under OECD TG 301 (OECD, 1992b). Under this guideline, a chemical is added to a system with non-adapted wastewater organisms at low wastewater solids concentration. This design forces microbes to rely on the test chemical as the sole source of carbon for energy and growth. A killed control provides background correction

and a positive control, using a well-known biodegradable compound, is used to demonstrate the adequacy of the inoculum. Under these constrained conditions, mineralization to carbon dioxide and water is measured and results are expressed as a percentage of theoretical CO₂ of the test compound. If 60 percent or greater of the test chemical is biodegraded in the course of the 28-d study it is considered “ready biodegradable” and under some regulatory interpretations, this must happen within a 10-d window beginning with the onset of degradation (typically inferred at 10 percent of theoretical CO₂ evolved). Ready biodegradable compounds are considered to have limited potential for environmental dispersion and are consequently considered to be “non-persistent.” This does not mean they would not be present in wastewater, as many chemicals enter waste streams at a more or less constant rate of input. Thus, it is sometimes also of interest to study the combined effect of degradation and sorption on wastewater treatment plant removal as well as degradation in other environmental systems such as surface water, sediment, soil, and marine waters. Importantly, however, biodegradability test data may lack realism compared to real-world situations due to extremely high dose rates used (i.e., orders of magnitude above limits of solubility).

Of the 15 organic UV filters approved for use in sunscreen in the United States, 12 have been assessed in at least one of the OECD test guideline studies (i.e., good laboratory practice [GLP] studies) conducted by various contract labs and designed to determine ready, ultimate, or inherent biodegradability (Table 3.4). Five of these (avobenzone, dioxybenzone, octocrylene, ensulizole, and ecamsule) have been determined to be nonbiodegradable with little to no evidence of biotransformation. Aminobenzoic acid, octisalate, octinoxate, and padimate O have clear evidence of ready biodegradability and thus, when discharged into WWTPs are highly likely to degrade. The conservative nature of the OECD 301 assay would suggest removal would be at a much higher rate given solids load at a higher rate in wastewater treatment. Interestingly, octisalate, octinoxate, and padimate O are also high log Kow (an indicator of the likelihood to adsorb to solids) compounds, meaning removal may also occur via wastewater solids given the sorptive nature of these compounds. Homosalate, oxybenzone, and sulisobenzone have inconsistent or inconclusive results regarding their ready biodegradability. Homosalate is inherently biodegradable with one of two available results suggesting ultimate biodegradability failing to meet the 10-d window requirement. The 10-days window is the prescribed duration within which the level of mineralization must reach 60 percent that follows a recorded level of 10 percent mineralization; if less than 60 percent mineralization is observed in the subsequent 10 days, the chemical is declared inherently degradable, but not ready biodegradable. However, this result was based on biological oxygen demand (BOD), a measure with relatively poor precision compared to other metrics for biodegradability. Oxybenzone’s lone result is based on BOD with a range of biodegradation cited (60 to 70 percent, at the margin of ready biodegradable). The available documentation of the study does not allow a deep enough review of the result. Sulisobenzone is amongst the most studied of the UV filters; however, several of the studies were performed in a non-optimal design (Zahn-Wallens test) where both sorption and biodegradation influence removal and DOC removal, a relatively low-resolution endpoint, was used. The results are thus inconsistent and clear conclusions are not drawn. Table 3.4 also provides a summary of the overall conclusions of these WWTP studies, integrating the studies on WWTP, surface water, and sediment for a reasoned conclusion of overall biodegradability when several studies are present.

A handful of studies are available in less common forms of the biodegradation assays (Table 3.4). Avobenzone, octinoxate, and ensulizole have been assessed for their anaerobic biodegradability in anaerobic sludge WWTP test systems. Octinoxate appears to be biotransformed and is concluded to be ultimately biodegradable under anaerobic conditions. Avobenzone has been the most extensively studied with additional data on biodegradation in surface freshwater and freshwater sediment environments. The formation of metabolites in these media suggest some level of biodegradability, which is contraindicated in WWTP systems. Octocrylene biodegradability in sediment has been studied with a conclusion of inherent biodegradability even though the chemical is clearly nonbiodegradable in wastewater. The UV filters have complex metabolic profiles and improved evaluations of biodegradability under more diverse environmental situations beyond wastewater treatment would be beneficial. Data are absent for all the UV filters for soil and marine environmental biodegradability and all information is under freshwater conditions.

Trolamine salicylate, cinoxate, and meradimate do not have formal ready biodegradability assessments (degradation in waste treatment) and TiO₂ and ZnO are inorganic and hence will not biodegrade. Thus, they are also unstudied.

TABLE 3.4 Formal Studies Regarding Ready Biodegradability Derived from (primarily) EU REACH Registrations of Organic UV Filters

NOTES: Green indicates a consistent conclusion of ready biodegradable, orange indicates inconsistent conclusions or inconclusive evidence of biodegradability, and red indicates nonbiodegradable. Results are then interpreted as robust conclusions regarding biodegradability under conditions of municipal wastewater treatment plant removal. R = ready; I = inherent; U = ultimate; N = nonbiodegradable. All data presented are under freshwater conditions.

There is no data available for three organic UV filters (i.e., trolamine salicylate, cinoxate, meradimate). This table is not applicable to the two inorganic UV filters considered in this report (i.e., titanium dioxide and zinc oxide). DATA SOURCE: https://echa.europa.eu.

Potential for Natural Sources of UV filters

Benzophenones, related to the benzophenone class UV filters, can occur naturally (Wu et al., 2014b). As a class of compounds, natural benzophenones contain more than 300, structurally diverse members with a range of bioactive properties. These compounds are largely plant-derived, with nearly 80 percent coming from the Clusiaceae family. Other, non-plant, sources have been isolated from fungus (Wu et al., 2014b). Despite the occurrence of benzophenone in nature (Wu et al., 2014b), there are no reports of natural sources of oxybenzone, dioxybenzone, or sulisobenzone entering the aquatic environment. However, this could be better informed by focused studies with analytical methods to identify natural sources.

The inorganic minerals used as UV filters do have potential natural sources. Titanium is the ninth most frequently occurring element in Earth’s crust, and various forms of titanium oxides are present as minerals in terrestrial systems (e.g., anatase, rutile, ilmenite) (Woodruff et al., 2017). These minerals are also mined directly, or produced from mined precursors and widely used as colorants in food, plastics, paints, and many other products (Keller and Lazareva, 2014; Weir et al., 2012). Likewise, ZnO can be naturally occurring as the mineral zincite. Magneli phase titanium oxide can occur naturally and has recently been viewed as an environmental pollutant originating from energy production during coal-combustion (Yang et al., 2017). While some research attempted to use trace elements within sunscreen-based TiO₂ to differentiate sunscreen from natural sources (Gondikas et al., 2014; Philippe et al., 2018), the approach has not been widely deployed and differentiating (i.e., allocating) contribution of inorganic sunscreens remains an analytical challenge and research gap.

ENVIRONMENTAL SETTINGS AND ROUTES OF UV FILTER EXPOSURE

Routes of UV Filter Exposure

Exposures of aquatic ecological receptors to UV filters are situational and depend on a combination of several factors. These factors pertain to the physico-chemical properties of the chemical and its interaction with, and ultimate fate in, environmental matrices as well as the ecological characteristics or the aquatic biota. First is the nature of the source, the way in which that source delivers the chemicals to receiving waters, and the concentration and frequency of input at which the chemical is present. As described previously, there is a variety of nonpoint and point sources of UV filters into the environment. Bathing beaches and water-based recreational areas with application and rinse-off of UV filters can result in passage of a fraction of skin-applied UV filters in sunscreens directly to receiving waters. WWTPs can capture or treat some compounds, lowering the levels in effluents relative to influents, and septic tanks and leaching fields can be a source to groundwater that may enter surface waters. Runoff from urban and agricultural areas may contribute, although there are limited data on these as a source. The strength of the source (i.e., the relative importance of the source) will depend on the mass of the UV filters being used and their propensity for either washing off skin at bathing beaches, passing through publicly owned treatment works, or reaching surface waters via leaching into soils (especially if sewage solids are applied) and subsequent groundwater transport or direct land runoff (after wind and storm events).

Second, the physico-chemical characteristics of the UV filters—primarily water solubility, degradation, and adsorption potential (Kow, Kd)—influence the fate of UV filters released into the environment and distribution among environmental compartments (i.e., air, water, sediment, and biota). The available, albeit limited, data on measurements of concentrations of UV filters in the environment point to combinations of physico-chemical fate characteristics and environmental settings that give rise to higher and lower exposure concentrations of the UV filters. While the physico-chemical fate characteristics are somewhat generally known or predictable (see Chapter 4), the interpretation of environmental exposure concentrations can be confounded by spatial and temporal variations in UV filter inputs (e.g., sunscreen use and contribution from other UV filter-containing product sources) and by data reliability constraints. These constraints arise from the common and novel challenges associated with sampling and analysis of environmental media. Analytical challenges are described in Chapter 4.

Figure 3.8 illustrates how factors related to environmental settings converge to influence exposure. A multitude of situations reflect combinations of sources, fate and transport dynamics, and receptors’ locations and ecologies.

At a conceptual level, the role of three predominant factors—source strength, proximity to source, and residence time/dispersion—can be envisioned as combining in various ways to affect exposure concentrations experienced by receptors.

Source strength.

Recreational use of sunscreens and sewage outfalls (including untreated stormwater) are likely among the most important direct inputs of UV filters from sunscreens into aquatic systems. This is especially the case when small water bodies with limited water exchange—ponds, lakes, and coves—are used for recreational swimming and there is a high density of people engaged. Other more diffuse sources include septic systems near water bodies, and nonpoint sources including urban and agricultural runoff. The magnitudes of sunscreen-associated UV filters entering the aquatic environment depend in large part on the number of people engaged in recreational activities or population sizes being serviced by WWTPs. There may also be seasonal patterns in certain areas based on climate; temperate-zone recreational areas can undergo peaks in the summer, while tropical recreational areas and wastewater outfalls may be impacted more continuously. Indeed, evidence for temporal variability (i.e., seasonal, diurnal) in UV filter concentrations has been shown in previous studies (described earlier in this chapter).

Proximity of ecological receptors to sources.

Proximity of a receptor to a source depends on UV filter source locations in relation to the locations and the ecology (e.g., habitat requirements, life history, and behavior) of the receptors. For sessile or highly localized organisms, the proximity to sources is largely a function of the distance (i.e. distance from source such as the beach), including the three dimensions of length, depth (e.g., surface, deeper waters), and time (e.g., tidal influence). Examples of sessile or highly localized organisms include reef communities (e.g., oyster and corals), attached or sessile animals (e.g., limpets, mussels, freshwater clams, other benthic invertebrates, crustacea, and other animals associated with specific habitats) and plants (e.g., seagrasses, seaweeds, freshwater macrophytes). In some cases, these sessile benthic organisms will be found only at the water depth or slightly above it. In other cases, animals such as corals and oysters can form reefs that can approach the water surface, with depths varying in tidal locations. Furthermore, many organisms (e.g., coral, fish) have early life stages (e.g., gametes, larvae) that reside in upper surface layers close to the microlayer. Other organisms are located in the pelagic (main water body) area but can transverse randomly or in diurnal cycles between surface and deeper

waters. Water depth also becomes a part of dispersion considerations because it allows for greater vertical mixing of inshore with offshore waters, resulting in dilution of concentrations at receptor exposure locations. The water depth at which receptors are located also influences the magnitudes of exposure to UV filters that are dissolved and/or associated with particles. Other organisms may not be sessile, but their reliance on localized aquatic areas for part of their life history, such as reproduction (e.g., amphibians), could cause them to have life stages in close proximity to a source.

Residence time.

Residence time of water is governed by dispersion (e.g., local mixing processes) and advection (e.g., current flows). The greater the residence time, the greater the likelihood for and magnitude of exposure depending on source input (i.e., continuous, pulsed, or single “spill” input) and total volume of the receiving body of water in comparison to the input volume. This applies to any aquatic habitat and can be measured, modeled, or otherwise estimated from physical principles. These principles are related to fluid mechanics and hold for all aquatic systems. Where freshwater and saltwater meet, though, the differences in salinity create limitations to potential dispersion and mixing of UV filters; this may result in other changes or degradation. These aquatic systems can be modeled to reflect the influence of the freshwater–saltwater mixing dynamics. Variations occur because of the relative magnitudes of processes such as wind-driven circulation, tides, seiches in large lakes, and inflows or circulation influences from other systems. This means that embayments or water bodies that are relatively quiescent, with shallow or low water volumes, or have longer residence times, are more prone to exposures of receptors than are aquatic environments characterized by higher energy physical processes (e.g., high flushing and/or advection rates).

Environmental Settings

As described in the preceding sections, UV filters from multiple point- and nonpoint sources can enter the environment, leading to variable amounts of dilution in the receiving waters and hence exposure potential to aquatic organisms. Different sources—such as those depicted in three representative scenarios shown in Figure 3.9—carry different “loads” of UV filters into an ecosystem, which can be the result of continuous discharge or episodic “pulses” of UV filters entering aquatic systems. The conceptual scenarios laid out in Figure 3.9 will be used in this and subsequent chapters as example frameworks for understanding how UV filters, including those from sunscreens, make their way to particular receptors and identifying data gaps that require addressing to perform ecological risk assessments. Developing distinct scenarios is useful because the particular characteristics of receptor environmental settings significantly influence the fate and transport processes UV filters are subject to once in aquatic environments, and ultimately their bioavailability and potential toxicity. Briefly, the three following scenarios are intended to contextualize the potential sources of UV filters.

1. Large water bodies with open water exchange (e.g., large lakes, estuaries, open-coast shorelines) with input from the land via rivers and nonpoint sources (e.g., urban or rural stormwater runoff, beach runoff, marinas) as well as point sources (e.g., industrial discharges, municipal wastewater discharges, combined sewer overflows, recreational activities). These water bodies have dilution capacity with water throughout the lake or with the open ocean but depend highly on direction and intensity of water movement and mixing, as well as total suspended sediments and organic matter.
2. Water bodies with minimal water exchange (e.g., small lakes or shallow, enclosed coastal bays) where discharges from large point sources are uncommon but impacts from nonpoint sources (e.g., septic system leaching, stormwater, land or beach runoff) and point source release during recreational activities (e.g., bathing, boating) is potentially more important than in Scenario I (depending on the extent of source input) because there is less dilution (i.e., water volume and movement).
3. Waterways with rapid water movement (e.g., creeks, streams, rivers) where spatial and temporal impacts of different nonpoint (e.g., stormwater, septic leaching) and point sources (e.g., municipal wastewater,

1. industrial discharge, combined sewer discharge, recreational activities) on aquatic exposures will vary with amount of dilution (i.e., streamflow)—which can vary seasonally, organic carbon transport, and population density present through the riverine environment.

The factors illustrated in Figure 3.8 (source magnitude, residence time/dispersion, and proximity/distance from source) along with the conceptual scenarios illustrated in Figure 3.9 provide a framework for considering how exposures might vary and how information on exposure from one type of location could be used to inform an assessment of exposures in another system. For example, a near-shore, semi-enclosed, shallow fringing reef system such as Hanauma Bay, Oahu, Hawaii, would be expected to have a larger source magnitude, experience longer residence times, and be nearer to the source than the offshore, open, deep barrier reef system bordering the Florida Keys, Florida (Figure 3.10). Similarly, near shore littoral areas of a large lake, like Lake George, New York, would be expected to be at the higher end of source magnitude, residence time, and proximity to source than the open, deeper waters and offshore benthic environment of the lake (Figure 3.11). The images in Figures 3.10 and 3.11 are real-world examples of the conceptual frameworks described in Figure 3.9.

ECOLOGICAL RECEPTORS AND ECOSYSTEM SERVICES

Ecological Receptors

Exposures are considered with reference to ecological receptors. Therefore, the selection of receptors for evaluating fate, exposure, effects, and risks is a key early step in ecological risk assessment. In accordance with EPA guidance, this is carried out as part of problem formulation and involves the development of assessment endpoints (EPA, 1998). Assessment endpoints are “explicit expressions of the actual environmental value that is to be protected, operationally defined by an ecological entity” (i.e., the receptor) and its attributes. The guidelines provide three selection criteria for identifying entities: ecological relevance, susceptibility (exposure plus sensitivity), and relevance to management goals. While assessment endpoints typically include receptors at various levels of biological organization ranging from organisms to ecosystems as well as critical habitat (EPA, 2003b), endpoints can also include ecosystem services (EPA, 2016) such as shoreline protection, food production, and carbon sequestration.

Many conceptual models can be constructed to convey exposure pathways for ecological receptors in the diverse aquatic and marine systems illustrated in Figure 3.9. One example is given for a coastal aquatic system with an upstream freshwater component (Figure 3.12). Exposure pathways for all aquatic systems include a combination of water, sediment, and food. The relative importance of these routes of exposure to receptors depends on the nature of the source, characteristics of the UV filters, and the ecology, life history and habitat of receptors.

Aquatic systems contain myriad microorganisms, plant and animal species, communities, and associated ecosystem services, processes, and functions. Some receptors are highly valued for the ecosystem services they provide

(e.g., coastal wetlands and oyster and coral reefs); some are considered more vulnerable to exposure because they are sessile (e.g., oysters, mussels, and corals); and some are keystone species (e.g., beavers, corals, sea grasses, mangroves). Threatened and endangered species are particularly important to identify to meet management needs under the Endangered Species Act. Likelihood of exposure to elevated levels of a contaminant (such as presence in recreational areas) and expected resiliency of populations to exposure and effects are also considerations. Resiliency of populations refers to a receptor’s capacity to withstand exposures and effects over the long term. Resiliency depends on factors such as biological turnover and recovery rates: the faster these rates, the more resilient the populations may be to short-term perturbations (Kindsvater et al., 2016), such as for species of plankton (Field et al., 1998). Turnover times and reproductive rates of populations are important considerations with respect to exposures and effects. Following from Kindsvater et al. (2016), animals that are longer-lived with longer turnover times are likely more susceptible to adverse population-level effects of exposures than are shorter-lived species with higher turnover rates although there are evolutionary advantages to longer lifespans. Phytoplankton are examples of organisms that undergo relatively rapid turnover and populations may recover rapidly following a perturbation to return in a following growing season. On the other hand, corals are complex holobionts (i.e., a sensitive relationship of animal, algal, and microbial components) with life history traits that may make them vulnerable, for a long time, to impacts from pollution or destructive events; this is because their ability to recover populations and reestablish may take decades or longer. There may also be less obvious potential exposure scenarios related to sunscreen and beach use (e.g., maternal transfer, inhalation, direct absorption across permeable membranes, ingestion). A number of species including various crustaceans and meiofauna utilize beach sands as habitat and marine turtles use beaches for laying their eggs which could be exposed depending on site-specific factors.

Figure 3.12 depicts organisms covering a range of sensitivities, from sensitive sessile species (e.g., endangered mussels) to less sensitive species with high turnover rates (e.g., plankton). Trophic transfer may occur from plankton to fish (and across fish species) to mammals and birds. Some organisms, like amphibians, have sensitive aquatic life stages (e.g., eggs and tadpoles).

Exposure of aquatic organisms to UV filters will partially depend on the fate characteristics of the UV filter and the ecology of the species. Upon entering aquatic environments, UV filters will partition into various environmental media including air, water, suspended solids, surface microlayers, and sediments. The dissolved phase of a UV filter in water is an important route for direct exposure to membranes such as those used for respiration and nutrition (e.g., gills and other surfaces that permit exchanges of dissolved gasses, nutrients, and other chemicals between the animal and the environment). For this route of exposure, which has received the greatest attention for toxicity testing as discussed in Chapter 6, aqueous solubility of a UV filter governs the amount of a UV filter that may be present in the water and available for exposure. Some of the more insoluble UV filters may adsorb onto organic matter such as suspended particulates, including phytoplankton, and other aquatic plants can also accumulate such chemicals into their tissues. Animals that feed on suspended particulate organic matter (i.e., filter feeders) could be exposed to UV filters in this manner. While there is little information on this exposure route for UV filters, it is a well-documented exposure pathway for other lipophilic organic compounds. Another exposure route is exemplified by corals, which, in addition to being filter feeders, utilize mucus to entrap particles (Brown and Bythell, 2005). This mucus could also potentially absorb UV filters.

Some animals feed directly on sediments and are broadly known as deposit feeders; these include sea cucumbers as well as a number of polychaete and crustacean species. To the extent that UV filters partition to the sediments, deposit feeding would be an exposure route for these species. For UV filters that adsorb to, or accumulate within, larger algae and vascular plants, there are myriad species that feed on these including invertebrates, turtles, aquatic birds, and some mammals. Little is known about this exposure pathway for UV filters. However, as discussed in Chapter 5, trophic transfer is an exposure route for animals that prey on animals that have accumulated chemicals into their tissues. Finally, some marine animals (e.g., dolphins and whales) breathe at the interface of air and water, and could inhale some water as part of this respiratory process. As described in Chapter 6, most effects data for UV filters are for water exposures. An ERA may need to utilize information on fate characteristics to evaluate exposures associated with the other routes of exposure.

Ecosystem Services

Human health and well-being are affected by actions taken to protect human health (in the case of this report, sunscreen use) as well as actions taken to protect ecological receptors and associated ecosystem services (MEA, 2005). Identification of ecosystems that provide valuable ecosystem services is one criterion for identifying receptors of concern. Coral and oyster reef ecosystems, seagrass beds, wetlands, and freshwater near-shore littoral areas provide a few examples.

Coral reefs are highly diverse marine ecosystems of substantial economic, ecological, and cultural value. A recent valuation study estimated the annual value of flood risk reduction alone provided by U.S. coral reefs is flood protection for more than 18,000 people and damage avoidance valued at $1.805 billion in 2010 U.S. dollars, which does not include fisheries and other ecological and cultural services (Storlazzi et al., 2019). Reef-based recreational fisheries are valued at about $100 million per year (Florida Keys National Marine Sanctuary, 2021). Coral reefs are also of substantial cultural value and importance, particularly in the Pacific Islands (e.g., Gregg et al., 2015). In Hawaiian culture, coral reefs are the site for a number of communal activities like fishing (e.g., McCoy et al., 2018), and ultimately fish sharing, which plays a key role in community cohesion (Glazier et al., 2013; Severance et al., 2013; Vaughan and Vitousek, 2013).

Oyster reefs are of great ecological, economic, cultural, and coastal resilience value. In addition to their importance in recreational and commercial fisheries, they provide habitat for numerous species, and are essential for a number of ecosystem services, including shoreline stabilization, water quality, and denitrification (Grabowski et al., 2012). Oyster reefs have an approximate value between $5,500 and $99,000 per hectare per year (Grabowski et al., 2012).

Seagrass beds, including in their role as habitat for other important species, provide ecosystem services including provision of food, coastal protection and erosion control, water purification, maintenance of fisheries, carbon sequestration, and tourism and recreation (Barbier et al., 2011; Nordlund et al., 2016).

Freshwater environments (e.g., lakes, rivers, and wetlands) play a key role in sustaining the health and welfare of people and the environment (UNEP, 2021). These ecosystems are valuable because they provide numerous ecosystem services (e.g., water supply for drinking water, irrigation, recreation, nutrient cycling, and waste assimilation; IPBES 2019). Services provided by lakes and rivers have been valued at $12,512/ha (Kubiszewski et al., 2017). The near-shore littoral areas of freshwaters (such as wetlands, which have been valued at $140,174/ha; Kubiszewski et al., 2017) also provide spawning and nursery grounds for many fish species that are used by myriad wildlife species (Peters and Lodge, 2009) and humans, too. For example, in 2016 in the United States, more than 35 million people went fishing (most anglers in the United States fish in freshwaters). Collectively, these individuals spent over $46 billion on recreational fishing (DOI, 2016).

FINDINGS AND KNOWLEDGE GAPS

There is a general weight of evidence to support the focus of direct recreational input and wastewater effluent as sources of UV filters into the aquatic environment, which is reinforced by some limited information about UV filters or similar contaminants. However, exposures of aquatic ecological receptors to UV filters depend on a combination of factors pertaining to the physico-chemical properties of the chemical and its interaction with, and ultimate fate in, environmental settings. UV filters vary in their aqueous solubility, lipophilicity, and degradation rates. Fate and transport also depend on initial mixing processes and the physical characteristics of the receiving waters. These will be discussed in greater length in Chapter 4. The knowledge gaps identified here are priorities based on the committee’s review of the information in this chapter.

Finding: Highly variable concentrations of some UV filters have been correlated with the time, location, and intensity of recreational activity, indicating sunscreen use as a source of UV filters to the aquatic environment.

Finding: In regard to decentralized wastewater treatment, inorganic UV filters are likely to be retained within on-site biological treatment systems (e.g., cesspools, septic systems) or removed from the effluent as the water

subsequently infiltrates leaching fields or permeates soil. There is insufficient research available to make similar assessments regarding organic UV filters.

Finding: Through engineered processes at most wastewater treatment facilities, homosalate, meradimate, octocrylene, octinoxate, octisalate, and padimate O, as well as the inorganic UV filters titanium dioxide and zinc oxide, are most likely to be highly removed from the effluent. Studies have shown their presence in sewage solids that are collected and disposed of off-site (landfills, land applied, incinerated).

Knowledge Gap: Predictive measures of data-informed releases based upon demography, behavior, water chemistry, activity type/level, and sunscreen formulation would enable quantification of rinse-off rates and exposure scenarios associated with recreational activities.

Knowledge Gap: Concentrations of UV filters in septic system discharge and off-site groundwater migration, in leakages from aging infrastructure, and in combined sewer overflows would improve understanding of inputs of UV filters into the environment from wastewater.

Knowledge Gap: Information on removal efficiencies for UV filters by specific wastewater treatment technologies could inform design of WWTPs that adequately remove UV filters from wastewater influent.

This page intentionally left blank.