Periphyton as an integrative indicator of pesticide bioaccumulation, nutrient pollution and fatty acid alterations in agricultural streams

TL;DR

Periphyton in agricultural streams can accumulate many pesticides and reflect nutrient pollution and fatty acid changes, making it a useful indicator of environmental health.

Contribution

This study demonstrates periphyton's role as an integrative bioindicator of pesticide exposure, nutrient pollution, and fatty acid alterations in agricultural streams.

Findings

Periphyton bioaccumulated up to 30 pesticides, with 18 exceeding thresholds for bioaccumulative substances.

Nutrient enrichment increased algal abundance and essential fatty acids like eicosapentaenoic and docosahexaenoic acid.

Diflufenican was identified as a high-risk pesticide due to its persistent, bioaccumulative, and toxic properties.

Abstract

Agricultural activities are a major source of pesticides and nutrients in freshwater ecosystems. However, little is known about the bioaccumulation and toxic impacts of pesticide mixtures on periphyton, an often-overlooked community of benthic microalgae. This study investigates pesticide bioaccumulation in periphyton and its links to nutrient enrichment and shifts in fatty acid profiles, highlighting its role as an integrative indicator in agricultural streams. Periphyton colonized on artificial substrates in three watercourses in southern Sweden was sampled and analyzed over a three-month summer period, representing sites with varying pesticide contamination. Periphyton bioaccumulated up to 30 pesticides, 17 of which were also present in surface water, with distinct profiles between matrices. Bioaccumulation was persistent over time and showed site-specific patterns. Eighteen pesticides exceeded the REACH threshold for very bioaccumulative substances (BCF > 5000). Diflufenican was identified as a high-risk compound, characterized by a high BCF and with persistent, bioaccumulative and toxic properties. Elevated nutrients coincided with greater algal abundance, notably diatoms, and higher levels of essential fatty acids, including eicosapentaenoic acid and docosahexaenoic acid, suggesting nutrient enrichment strongly shapes periphyton and may obscure subtle pesticide effects. Assessing pesticide accumulation in periphyton demonstrates its value as a passive sampler providing complementary insight into chemical exposure and ecological status beyond surface-water monitoring.

This study shows that periphyton in agricultural streams bioaccumulates a wide range of pesticides, while nutrient enrichment strongly shapes its community and fatty acid profiles, demonstrating periphyton’s value as a complementary bioindicator of chemical exposure and ecological status beyond surface-water monitoring.

Figures

(A) Pesticide concentrations (ng/g AFDM) -grouped into classes, fungicides, herbicides, and insecticides- in periphyton collected from REF stream, Streams 1 and 2 in July, August, and September 2022. (B) Pesticide concentrations (µg/l) in surface water from Streams 1 and 2 as mean values for July, August, and September 2022 from weekly time-integrated sampling. Water concentrations for the REF stream are omitted because they were collected using a different sampling method (grab sampling; see Figure S4 for details).

(A) Bioconcentration factors (BCF) for 30 pesticides, calculated as the ratio of concentrations in periphyton to surface water (y-axis is log transformed). Dashed lines indicate REACH thresholds: 5000 (very bioaccumulative, the upper dashed red line) and 2000 (bioaccumulative, the lower dashed orange line). (B) Linear regression of BCF (log transformed) vs. Log KOW. A significant correlation was found (P < 0.001) when excluding pyroxsulam as an outlier.

(A) Chlorophyll a concentration (mg/mg AFDM ± SD) in periphyton of each stream and sampling month. (B) Relative abundance of photosynthetic pigments in periphyton (normalized to chlorophyll a) of each stream and sampling month. Note that pigments associate to green algae are indicated in green, those associated to cyanobacteria in blue, and those associated to diatoms in brown.

(A) Sum concentrations of the essential omega-3 fatty acids EPA and DHA in periphyton (µg/g AFDM) from the REF stream, Streams 1, and 2. Bars represent mean ± SD; n = 3. (B) Relationship between fucoxanthin concentrations and the summed concentrations of EPA and DHA (µg/g AFDM) in periphyton across all sites. Symbols represent different sites: circles = Stream 1, rectangles = Stream 2, diamonds = REF stream. Linear regressions illustrate the relationship between total EPA+DHA concentrations and fucoxanthin across varying pesticide exposures.

Linear relationships of concentrations of fatty acids EPA+DHA (µg/g AFDM) and diatom maker fucoxanthin (µg/g AFDM) vs. total nitrogen (mg/l), total phosphorus (µg/l), and sum of total pesticides concentrations in periphyton (ng/g AFDM, dark green).

Tables

Table 1. Ranking of pesticides that had BCF > 5000 and >2000 in periphyton, their maximum and mean BCF values, number of samples for which BCF could be calculated (in brackets number of samples with BCF above the 5000/2000 threshold), their Log KOW, PNEC values for aquatic organisms (EFSA guidelines), water solubility, and Soil DT50 values (both retrieved from PPDB).

Ranking by BCF max	Substances	BCF max value	BCF mean value	No. of samples (no. > 5000)	Log Kow	PNEC (µg/l)	Water solubility (mg/l)	Soil DT50 (days)
BCF > 5000
1	Pyroxsulam	626 000	68 400	13 (12)	-1.01	0.26	3200	3.3
2	Clomazone	59 800	21 600	6 (3)	2.58	5.7	1212	22.6
3	Ethofumesate	50 800	12 600	12 (5)	2.70	16	50	21.6
4	Prosulfocarb	34 400	11 400	30 (16)	4.48	4.2	13	11.9
5	Diflufenican	32 100	26 400	13 (13)	4.20	0.025	0.05	94.5
6	Propamocarb	20 900	13 300	3 (2)	0.84	630	900 000	14.00
7	Tebuconazole	16 500	4920	12 (6)	3.70	1.0	36	63
8	Methabenzthiazuron	10 000	3400	9 (1)	2.64	NA	60	135
9	Bixafen	9600	3000	25 (9)	3.30	0.46	0.49	500
10	Benzoin diflupyr	9520	4440	21 (8)	4.30	0.035	0.98	908
11	Mandipropamid	8210	3120	14 (3)	3.20	7.6	4.2	49.1
12	Fluopyram	7960	2280	16 (1)	3.30	5.0	16	309
13	Isoproturon	7730	2160	30 (3)	2.50	1.3	70	12
14	Dimethomorph	7390	2450	29 (3)	2.68	5.6	29	72.7
15	Thiacloprid	7280	4190	16 (6)	1.26	0.077	184	0.88
16	Phenmedipham	6460	2160	14 (1)	2.70	0.5	1.8	12
17	Fluopicolide	5540	1790	12 (1)	2.90	2.9	2.8	271
18	Prochloraz	5100	2160	8 (1)	3.50	0.55	27	120
BCF > 2000
19	Pirimicarb	4990	1220	31 (6)	1.70	0.09	3100	73.6
20	Boscalid	4500	2880	6 (4)	2.96	13	4.6	484.4
21	Picoxystrobin	4280	3050	7 (7)	3.60	0.057	3.1	24.4
22	Fluxapyroxad	3910	731	25 (3)	3.13	2.9	3.44	183.0
23	Propyzamide	2600	1550	10 (3)	3.27	2.1	9.0	50.5
24	Fenpropimorph	2600	1870	4 (1)	4.50	0.2	4.32	35
25	Pyraclostrobin	2500	1260	30 (6)	3.99	0.042	1.9	41.9
26	Acetamiprid	2000	1560	5 (1)	0.80	0.024	2950	1.6

Equations

$$\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} \begin{eqnarray*} BCF\ = \frac{{\textit{Concetration}\ in\ \textit{periphyton}}}{{\textit{Concentration}\ in\ \textit{surface}\ \textit{water}}}. \end{eqnarray*}\end{document}$$

Full text

Introduction

Aquatic ecosystems face many human-induced stressors. Chemical pollution, especially from chemical pesticides, is a major driver of biodiversity loss and ecosystem degradation in small agricultural streams (Spycher et al. 2018, Sigmund et al. 2023). Pesticide residues primarily enter these streams through runoff, spray drift, and leaching, potentially impacting non-target aquatic organisms (European Environmental Agency 2023). Throughout the growing season, herbicides, fungicides, and insecticides are applied based on crop requirements, pest pressure, and prevailing weather conditions. This usage results in pesticide pollution in agricultural streams and often occurs as concentration peaks following application or rainfall events, with levels fluctuating rapidly over time and consisting of complex mixtures of various compounds (Nowell et al. 2018, Spycher et al. 2018, Halbach et al. 2021). Some of these pesticides may include older, now-banned substances with high persistence that have accumulated in the landscape and are slowly leaching from soil and groundwater (Rasmussen et al. 2015, Rheinheimer dos Santos et al. 2020). Others may be modern pesticides, including contaminants of emerging concern (CECs) such as organochlorines and neonicotinoids (Fadare et al. 2024). In parallel with pesticide inputs, agricultural runoff also contributes substantial nutrient loads, particularly nitrogen and phosphorus, which can stimulate algal growth and further alter stream biological structure and ecosystem functions (Ardón et al. 2021).

Periphyton is a benthic community of photosynthetic organisms including algae and cyanobacteria, that live together with heterotrophic bacteria, fungi, protozoa, and meiofauna. This close association creates a complementary relationship where phototrophs produce organic compounds that support heterotrophic metabolism, while heterotrophs release CO₂ through respiration, which phototrophs then use for photosynthesis (Battin et al. 2003). Periphyton refers to the photosynthetically active microbial community attached to submerged surfaces like cobbles and stones in aquatic ecosystems, often visible and complex. In contrast, biofilm is a broader term for any microbial community embedded in an extracellular polymeric substances (EPS) matrix on a surface, found in many environments beyond aquatic systems (Bonnineau et al. 2021). Periphyton also facilitates essential ecological functions such as primary production, respiration, organic matter decomposition, and contaminant turnover in aquatic environments (Guasch and Subater 1995, Battin et al. 2003, Romaní et al. 2008, Fechner et al. 2012, Battin et al. 2016, Wood et al. 2019). Periphyton responds to factors such as light availability, water discharge (i.e. flow conditions), and the trophic status of the environment (Tlili et al. 2011, Battin et al. 2016, Zubrod et al. 2019). The functions and community structure of periphyton can be changed and adversely affected by pesticide toxicity (Tlili et al. 2011, Gardeström et al. 2016, Zubrod et al. 2019, Nagai 2020, Gómez-Martínez et al. 2024). The algal component of periphyton is particularly sensitive to the toxic pressure of herbicides, for example isoproturon, diflufenican, or methazachlor have modes of action associated with photosynthesis inhibition or disruption of membrane lipids (Copin and Chèvre 2015, Gómez-Martínez et al. 2023). Pesticide impacts extend beyond periphyton itself, influencing the broader aquatic food web, as periphyton constitutes a critical basal food source for consumers at higher trophic levels (Bonnineau et al. 2021, Izma et al. 2024). In addition to altering periphyton structure and function, pesticides can accumulate within these communities (Bonnineau et al. 2021).

Although still limited, several studies have demonstrated that a wide range of pesticides can bioaccumulate in periphyton (Fernandes et al. 2020, Rheinheimer dos Santos et al. 2020, Rooney et al. 2020, Ijzerman et al. 2024). In this study, the term bioaccumulation refers to the concentration of pesticide residues measured in periphyton, i.e., the extent to which pesticides are retained within the periphyton community. In contrast, bioconcentration is described through calculated bioconcentration factors (BCFs), which provide a standardized measure of the partitioning of a compound between biological material and the surrounding water.

Bioconcentration of pesticides depends on fundamental physical and chemical properties of each pesticide such as water solubility, its affinity to organic matter, and its reactivity, as well as environmental conditions (Boxall 2012, Vonk and Kraak 2020, Bonnineau et al. 2021). Bioconcentration has been shown to correlate with the octanol-water partition coefficient Log K_OW_ (Ijzerman et al. 2024), whereas other studies have not observed such relationship (Mahler et al. 2020, Rooney et al. 2020). Also properties such as water solubility and soil degradation DT50 can be used to discuss occurrences of single compounds in periphyton (Fernandes et al. 2023). Understanding and predicting bioaccumulation is a key component of chemical risk assessment within the persistence, bioaccumulation, and toxicity (PBT) framework, established under EU REACH and adopted by the US EPA (EU REACH 2011, US EPA 2016). PBT substances are of particular concern because they persist in the environment, accumulate in biota, and can exert toxic effects even at low concentrations. Within this framework, the bioaccumulation (B) criterion is assessed using BCF, bioaccumulation (BAF), or biomagnification factors (BMF), with substances exhibiting a BCF above 2000 in aquatic organisms classified as bioaccumulative (EU REACH 2011). In agricultural streams, such substances can accumulate within algal and microbial components of periphyton and biomagnify through food webs, posing risks to periphyton communities and higher trophic levels. Incorporating bioaccumulation data from natural periphyton communities can improve predictions of trophic transfer (Rogers et al. 2016, Rheinheimer dos Santos et al. 2020) and chronic exposure, which are often underestimated by water- monitoring programs (Fernandes et al. 2023).

In addition to chemical stressors like pesticides, nutrient availability is a key driver shaping the structure and function of periphyton communities (Carpenter et al. 1998, Munn et al. 2010). While pesticide exposure can adversely affect periphyton composition and function, nutrient enrichment, particularly from agricultural runoff, can stimulate algal biomass and also alter community composition (Taulbee et al. 2005, Munn et al. 2010). Moderate-to-high nutrients availability promotes algal biomass and the proliferation of diatoms taxa and green-algae (Chase et al. 2017). The trophic status also affects cyanobacteria abundance, being especially sensitive to phosphorus levels (Sabater et al. 2016, Pacheco et al. 2022). One key function as part of primary production of periphyton is the production of essential fatty acids (FA) by its algal components, such as diatoms. Essential FA—i.e. omega-3 and omega-6 polyunsaturated FA (ω3 and ω6 PUFA) with ≥20 carbon, are needed for many physiological processes that affect growth, reproduction, immunity, and overall performance of aquatic animals; but animals generally lack the ability to synthesize these FA de novo and have to rely on dietary intake (Sargent et al. 1995, Brett and Müller-Navarra 1997, Sargent et al. 2003). Thus, FA have been widely used as biomarkers to assess food quality, consumer diets, and trophic-transfer pathways in food webs (Müller-Navarra et al. 2000, Brett et al. 2009, Lau et al. 2009, Brett et al. 2017). Algae such as diatoms that can produce large amounts of essential FA, such as eicosapentaenoic acid (EPA, 20:5ω3) and docosahexaenoic acid (DHA, 22:6ω3), are particularly high-quality basal resources for efficient trophic transfer and consumer production in aquatic ecosystems (Brett and Müller-Navarra 1997).

Periphyton serves as a bioindicator of water quality due to its rapid response to environmental and water quality changes such as nutrient pollution (Burns and Ryder 2001, Wood et al. 2019) However, its potential as a bioindicator of pesticide contamination and its integration into pesticide monitoring programmes remains less established, despite a few studies highlighting its importance (Fernandes et al. 2020, Mahler et al. 2020, Rheinheimer dos Santos et al. 2020, Rooney et al. 2020, Fernandes et al. 2023, Morin and Artigas 2023, Ijzerman et al. 2024). The ecological impacts of pesticide exposure on periphyton, particularly on its community structure and FA-based nutritional quality, under realistic field scenarios remain poorly understood. In agricultural pesticides applications, exposure scenarios shift continuously since pesticide applications vary across the crops growing season (Spycher et al. 2018). This in turn, will potentially influence periphyton’s pesticide bioaccumulation, community structure, algal biomass, and nutritional quality. Furthermore, in agricultural streams, high nutrient loads often co-occur with pesticide residues, potentially masking the specific effects of pesticides (Alexander et al. 2013, Rossi et al. 2018). Thus, the interplay between nutrient enrichment and pesticide exposure can significantly influence periphyton responses, with nutrient availability potentially exacerbating or mitigating the toxic effects of pesticides (Murdock et al. 2013, Rossi et al. 2018).

This study aims to assess the role of periphyton as an integrative indicator linking pesticide bioaccumulation, nutrient enrichment, and changes in fatty acid profiles in agricultural streams. Specifically, the objectives were (1) to evaluate the capacity of periphyton to bioaccumulate pesticides by comparing pesticide profiles across sites and sampling periods, estimating BCFs, and relating these patterns to compound hydrophobicity (Log K_OW_). And (2) to determine whether shifts in algal community composition and fatty acid profiles are associated with water quality and pesticide bioaccumulation.

We hypothesize that periphyton in agricultural streams bioaccumulates a diverse range of pesticides, particularly those with a high potential to partition into organic material (i.e. high Log K_OW_), whereas surface water is expected to contain more water-soluble pesticides. We expect the algal assemblages in stream periphyton receiving agricultural runoff to differ from those in the reference site, both in algal composition and biomass because of the different levels of pesticide and nutrients among sites. If this hypothesis is correct, we anticipate that the composition and concentration of essential fatty acids in periphyton will differ between agricultural streams and the reference site, potentially as a consequence of elevated ambient nutrient levels and the associated increase in periphyton biomass. By exploring these interconnected aspects, the study will provide a more comprehensive understanding of how pesticide exposure affects the functional integrity and ecological role of periphyton in aquatic ecosystem. This understanding is essential for monitoring the potential ecological risks posed by pesticide contamination in those systems.

Method

Study sites

The study was carried out in three streams in Sweden, experiencing different levels of pollution. Two of the streams are located in small catchments with intensive agricultural activities, while one is located in a nature reserve and serves as the reference stream for this study (Figure S1). The two impacted streams are part of Sweden’s national monitoring programs for pesticides and nutrient losses which facilitated their selection as case studies of pesticide-impacted agricultural streams (Kyllmar et al. 2014, Boye et al. 2019). Agricultural land use exceeds 89% in both catchments. One of these streams (Stream 1) is situated in Östergötland County, in southeastern Sweden, with a catchment area of 16.4 km². The other (Stream 2) is located in Skåne, the southernmost region of the country, with a catchment area of 8.2 km². In both catchments, the dominant crops are autumn- and spring-sown cereals, and more than 95% of the arable land is artificially drained (Kyllmar et al. 2014). The average pesticide use between 2002 and 2016 was estimated at 0.8 kg/ha for Stream 1 and 1.72 kg/ha for Stream 2 per year (Boye et al. 2019). Sampling occurred 1–2 km downstream of the underground drainage outlets of both streams. A third stream (REF stream), had no major pollution sources, located within the Häckeberga Nature Reserve in Skåne. This stream is located ∼23 km north of Stream 2. The REF stream is a small, groundwater-fed waterway that flows through a semi-forested, grassy landscape managed by cattle grazing as part of the reserve’s habitat management. The Häckeberga Nature Reserve covers 165 hectares and strictly prohibits the use of chemical pesticides. It features a diverse mosaic of habitats, including beech-dominated deciduous forests, mixed woodlands, pasture meadows, dry grasslands, and wetlands. This stream has served as reference stream for non-pesticide pollution in several studies (Wessberg 2015, Håkansson 2017). The width of all three streams was ∼3 m, with generally low water flow in summer. Flow was particularly low during the June sampling, with an average water depth of 0.15 m.

The bottom substrate of Stream 1 consisted of 75%–80% silt/clay (<0.063 mm), with gravel and sand each contributing 10%. A riparian buffer strip of at least 6 m in width surrounded Stream 1, comprising dense grass, semi-grass, and bush vegetation. Stream 2 had a more heterogeneous substrate, composed of 50% gravel (2–63 mm), 20% sand (0.0063–2 mm), 15% cobbles (63–200 mm), 10% boulders (200–630 mm), and 5% large boulders (630–2000 mm). The riparian buffer strip exceeded 10 m in width and was dominated by nettles and black alder trees. The REF stream had a substrate of 70% silt (0.002–0.063 mm), 20% sand, and 10% gravel. Its riparian zone (0–5 m) was dominated by ferns and black alder.

Sampling

Artificial substrates made of unglazed ceramic tiles (14.5 × 14.5 × 0.5 cm), were deployed in the benthos of each sampling site for periphyton colonization and growth for ∼3 months (from end of June to mid-October 2022). Ceramic tiles were chosen because the streams were shallow, and the tiles could be placed just a few centimetres above the bottom of the streams. Ceramic tiles were mounted on metal frames with 90° angles and fixed in place using small plastic clamps to allow periphyton colonization under natural flow conditions. To minimize grazing pressure, the tile edges were coated with a thin film of Vaseline to deter invertebrate access A total of 15 tiles were deployed in each stream over a 50 m reach, arranged in five groups of three tiles (Figure S1). The groups were spaced along the reach, with tiles within each group separated by 1–3 m. Approximately every 30 days over the three-month study period, one tile from each group was collected, resulting in five replicate samples per sampling occasion. Sampling occurred at three time points corresponding to July, August, and late September/mid-October. For simplicity, we refer to these as the July, August, and September collection periods, although the final sampling took place in mid-October. This 50 m reach size was designed to minimize within-stream variation in habitat conditions (e.g. flow, riparian shading, etc.) for biofilm growth and in pesticide effects due to potential longitudinal dilution of pesticide concentrations. We treat stream reach as the primary experimental unit. During the collection the tiles were gently rinsed in the stream water to remove invertebrates or debris from the surface. The collected tiles were shipped to the laboratory at SLU Uppsala within 24 h in dark, cooled boxes. Once in the laboratory, the periphyton was brushed off from the tiles using a toothbrush and rinsed with ultrapure water (Milli-Q water) into a tray. The periphyton slurry was then collected in 250 ml bottles and stored at −18°C. Samples were then freeze-dried for 7 days and homogenized using a mortar and pestle and kept at −18°C afterward. We note that differences in pesticide concentrations among sampling periods may reflect both the timing of pesticide applications and the developmental stage of the periphyton community.

Pesticide concentrations in surface water were measured using time-integrated weekly composite samples collected from Streams 1 and 2, under the Swedish monitoring following the methodology described by Boye et al. (2019). Briefly, samples were collected every 90 min using an automatic ISCO sampler (6712FR), with 20 ml of water taken per sampling event. Each sample was split evenly, with 10 ml stored in a glass bottle and 10 ml in a plastic bottle. The bottles were kept refrigerated and replaced weekly. The REF stream, however, is not part of the national monitoring program. Therefore, two surface water grab samples were taken, one at the time of tile deployment end of June and another at the end of the three-month sampling period mid-October in 2022. These samples were kept at 4°C during transport to the laboratory.

Flow-proportional samples for total nutrient and phosphorus analysis were collected at frequencies controlled by dataloggers installed at the water discharge stations under the national nutrient monitoring program (Kyllmar et al. 2014). Water was drawn through suction tubes positioned at defined depths within the measuring sections. The sub-samples were stored under refrigerated conditions in composite bottles. For the REF stream, only a single water sample for nutrient analysis was collected on the day the tiles were placed (28 June 2022).

Sample preparation and analysis

Pesticide analysis

A total of 109 pesticide compounds were analysed at the OMK laboratory at SLU. The list of compounds and some of their chemical characteristics can be found in Table S1 of the supplemental material. The term pesticides in this study referred to active substances of chemical plant protection products.

Periphyton

We used a validated biota analysis method developed in-house for pesticide residue determination. For each replicate, 50–100 mg dry weight (DW) of periphyton was transferred into 7 ml Precellys mixing tubes containing 2.8 mm ceramic beads. If individual replicates contained less than 50 mg DW, two replicates were combined to meet the required sample weight. An internal standard solution (IS, 50 µl of 50 ng/ml stable isotopically-labelled compounds in acetonitrile, ACN) was then added to the samples, which were left to dry for 20 min. Next, 2 ml of ACN was added to each tube, and the samples were homogenized and extracted using a Bertin Precellys® Evolution Touch instrument. The homogenization process involved two 30-s cycles at 6000 r/m, with a 30-s pause in between. The tubes were then centrifuged at 3000 × g for 2 min, and the resulting supernatant was transferred into 15 ml polypropylene tubes. This extraction process was repeated with an additional ultra-sonication step (using a Sonics Vibracell, 6 mm probe, 30% amplitude, for 30 s) before centrifugation. The Precellys tubes were refilled with 2 ml of ACN, vortexed, centrifuged, and the supernatant was once again transferred into the 15 ml polypropylene tubes. The combined extracts were evaporated to dryness in a 40°C water bath under a gentle stream of nitrogen gas. The dried extracts were then reconstituted in 100 µl of ACN. After mixing on a Heidolph mixer for 20 min, the samples were centrifuged at 3000 × g for 2 min, and the supernatant was transferred to 2 ml LC vials containing 250 µl glass inserts.

Surface water

Surface water samples, including grab samples from the REF stream, were processed and analyzed at the the accredited Organic Environmental Chemistry (OMK) laboratory at the Swedish University of Agricultural Sciences (SLU), following ISO/IEC 17025-accredited methods (ISO 2017). The analytical procedure is detailed elsewhere (Jansson and Kreuger 2010). Briefly, the pH of the samples was adjusted to pH 5 for ES(+) analysis and pH 3.5 for ES(–) analysis using diluted acetic acid. An aliquot of 100 ml internal standard solution (5 ng/ml methanol) was added to 5.0 ml of the pH-adjusted sample. After thorough mixing, the samples were filtered using RC syringe filters attached to a 5 ml polypropylene plastic single-use syringe. The filtered samples were then transferred to 6 ml clear screw-cap vials (Agilent Technologies) for subsequent instrumental analysis. For this study, median surface water concentrations were calculated for each month from July to October.

Instrumental analysis

Pesticide concentrations in periphyton and surface water samples were measured using high-performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS) with an on-line solid-phase extraction (SPE) system and an electrospray ionization interface as described elsewhere (Jansson and Kreuger 2019). In brief, samples were injected for positive (ESI+) and negative (ESI-) ionisation on an Agilent 6470 instrument (Agilent Technologies, USA). The on-line SPE system consisted of two connected columns: a Strata C18 (20 μm, 20 × 2 mm, Phenomenex, USA) and a Strata X (25 μm, 20 × 2 mm, Phenomenex, USA). In total 109 polar and semi-polar pesticides were quantified (Table S1). During analysis, the sample was loaded onto the SPE columns, after which the target compounds were back-flushed into the analytical column, an Eclipse Plus C18 (3.0 × 100 mm, 3.5 μm, Agilent Technologies, USA). Separation was achieved using a methanol gradient in a 10 mmol/l ammonium formate buffer (pH 4.2) containing 2% methanol and 6% isopropanol. For each pesticide, at least two MS/MS transitions were monitored, one transition was used for quantification, while the other served as a qualifier ion to confirm the compound’s identity.

Quality control for periphyton analysis

Each of the three analytical batches included a six-point calibration curve, a matrix QC sample (periphyton spiked with native analytes and IS), and blank samples (method blank and solvent blank with IS). For validation and final concentration determination, relative recoveries (i.e. the native signal divided by the IS signal in spiked sample matrix, in relation to the native and IS signals in calibration samples) were assessed to correct for extraction recovery and matrix effects in the LC-MS interface (ion source). This was determined through matrix spike tests, in which duplicates of periphyton samples were spiked with native compounds to a nominal concentration of 10 ng/g. The measured concentrations for these relative recovery samples were background subtracted (concentration in the non-spiked matrix) and divided by the nominal concentration, giving the relative recovery for each compound and matrix. Furthermore, absolute recoveries were assessed by spiking periphyton samples in duplicate with native compounds prior to extraction, and comparing them to a sample spiked after extraction, which served as a reference. The IS solution was in this experiment added to the final extracts. A non-spiked sample was extracted to determine any background signals that were subtracted from the absolute recovery samples and the reference sample. The average total recovery for periphyton samples was 65%. However, 17 compounds exhibited total recoveries below 20% (Table S2); these compounds could not be reliably quantified and were also not detected in any of the periphyton samples. Method limits of detection (LODs) were estimated from spiking experiments and method limits of quantification (LOQs) were then set to 3.33 times that concentration (Table S2). Final concentrations of all periphyton samples were then correct by their organic content (Figure S2). Information about the quality control of the surface water analysis can be found in Jansson and Kreuger (2010).

Pigment analysis

For pigment analysis 10–20 mg DW of each periphyton sample were weight into 15 ml polypropylene tubes. Pigments were extracted by resuspending the dried periphyton in 4 ml of solvent (80/20 acetone/methanol v/v) and incubating for 1 h at −20°C in dark conditions. Ultrasonication was then performed on the solution at 4°C for 3 min. The extracts were then filtered through 0.45 μm Target2™ Nylon Syringe 4 mm diameter Filters (Thermo Scientific™, cat n°F2504-1) and stored at −20°C in dark glass vials until analyzed by high-performance liquid chromatography (HPLC; Shimadze Prominence HPLC Systems) according to Corcoll et al. (2019). A total of eight photosynthetic pigments were identified, including chlorophyllides (chlorophyll a and chlorophyll b), and xanthophylls (fucoxanthin, violaxanthin, zeaxanthin, lutein, diadinoxanthin, and diatoxanthin). Pigments were identified using internal standards, and relative abundance was estimated according to Jeffrey and Wright (2006). The qualitive analysis included photosynthetic pigments in periphyton from green algae, cyanobacteria, and diatoms, collectively referred to here as the algal assemblage. Although cyanobacteria, also known as blue-green algae, are prokaryotes and not true algae in a taxonomic sense, they are included in the term “algal assemblage” due to their ecological and functional similarity to eukaryotic algae. Like other algal groups, cyanobacteria perform oxygenic photosynthesis and contribute significantly to primary production in aquatic ecosystems. Chlorophyll a and fucoxanthin could be determined additionally quantitively and used as markers for total algal biomass and diatom biomass, respectively.

Fatty acid analysis

We used a method modified from Grieve and Lau (2018) for FA analysis. FA from ca. 5 mg DW of each periphyton sample were extracted using 1000 µl of extraction buffer (2 : 1 v: v chloroform: methanol) containing the IS deuterium-labelled pentadecanoic-d_29_ acid (10 ng/µl; Sigma–Aldrich Sweden AB, Stockholm, Sweden). A tungsten bead was added and the solution was shaken at 30 Hz for 3 min in a mixer mill (Mixer Mill MM 400, Retsch GmbH, Haan, Germany). The bead was removed and 200 µl of 0.15 M NaCl was added. After 2 min vortexing, the samples were left at room temperature for 40 min, and then centrifuged at 18 845 × g for 10 min at 4°C. For each sample, 400 µl of the lower phase was transferred to two micro-vials (subsamples A and B), i.e. each with 200 µl for later methylation or transmethylation. All extracts were then dried using a nitrogen concentrator and stored at −80°C. Blank samples (without any periphyton) were also prepared in the same way as the periphyton samples. A control sample, made of pooled extracts from all periphyton samples was prepared and aliquoted to later run several times evenly distributed during the analysis for quality control.

Subsample A was methylated for analysis of free FA, and subsample B was transmethylated for analysis of FA bound in lipid compounds. For methylation of subsample A, 200 µl 1% trimethylsilyldiazomethane in isopropanol: dichlormethane (1 : 5 v: v) was added, then vortexed for 10 min on a multi-tube vortexer (VX-2500, VWR International AB, Kista, Sweden). The solution was evaporated in a fume hood for ∼16 h to complete dryness and afterward added with 60 µl deuterium-labelled methyl heptadecanoate-d_33_ (10 ng/µl in heptane) (Sigma–Aldrich Sweden AB). For transmethylation of subsample B, 100 µl deuterium-labelled methyl heptadecanoate-d_33_ (25 ng/µl in heptane) and 100 µl of sodium methylate: dimethyl carbonate (1 : 1 v: v) were added. The solution was vortexed for 10 min on a multi-tube vortexer and then left for phase separation at room temperature for 1 h. Afterward, 40 µl of the upper phase was transferred to a new vial for analysis.

Concentrations of FA methyl esters were analysed with a gas chromatography–mass spectrometry (7890A GC, Agilent Technologies Inc., California, USA; 7000C QQQ MS, Agilent Technologies Inc.) installed with a Zebron ZB-FAME column (length 20 m, internal diameter 0.18 mm, cyanopropyl phase thickness 0.15 µm; Phenomenex, California, USA). The Supelco 37 Component FAME Mix (Sigma–Aldrich Sweden AB) was used as the standard to identify individual FA. 1 µl of each sample was injected, using split mode (1 : 5) and splitless mode for methylated and transmethylated samples, respectively. Injector temperature was 250°C. Helium was used as the carrier with a constant flow rate of 1 ml/min. Column temperature was set at 80°C for 1.5 min, then increased by 40°C/min–160°C, 5°C/min–185°C, then 30°C/min–260°C which was finally maintained for 1 min. Column effluent was introduced into the electron impact ion source of the mass spectrometry. Transfer line and ion source temperatures were 250°C and 230°C, respectively. Quantification of sample FA was performed using specific ions with MassHunter™ Quantitative Analysis QQQ (Agilent Technologies Inc.). Data of individual FA from methylation and transmethylation were then summed for each sample.

Ash-free dry mass

To determine the ash-free dry mass (AFDM), all periphyton samples were first freeze-dried and weighed to measure its dry weight. Periphyton samples were then burned in a muffle oven at 505°C for 90 min, following a 15-min ramp-up period. After combustion, the remaining inorganic material was weighed to calculate the ash-free dry mass. The pesticide concentration in periphyton, pigment concentrations and fatty acid concentrations were all corrected by its organic fraction (Figure S2).

Nutrient analysis

Total nitrogen and phosphorus analyses of the water samples were conducted at the accredited Geochemical Laboratory at SLU. For total nitrogen analysis, samples were treated with HCl and subjected to catalytic oxidation to convert nitrogen compounds into nitrogen oxides, which were then measured via chemiluminescence using a Shimadzu TOC-VCPH instrument, following the standard SS-EN ISO 20236 : 2021 (ISO 2021). Total phosphorus analysis was performed on unfiltered samples according to SS-EN ISO 6878 : 2005 (ISO 2005). Phosphorus was oxidized with peroxodisulfate to form dissolved orthophosphate, which was subsequently quantified using a molybdate reaction followed by spectrophotometric determination.

Bioconcentration factor

Under the term bioaccumulation in this study we refer to the pesticide residue concentration measured in the periphyton; hence how much is retained in periphyton relative to input mass. Bioconcentration factors (BCFs) is a standardized measure of partitioning between biological material and water. They were calculated by dividing the pesticide concentration in periphyton by the corresponding concentration in ambient surface water:

\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} \begin{eqnarray*} BCF\ = \frac{{\textit{Concetration}\ in\ \textit{periphyton}}}{{\textit{Concentration}\ in\ \textit{surface}\ \textit{water}}}. \end{eqnarray*}\end{document}

The BCF was calculated for Streams 1 and 2 with surface water concentrations determined via time integrated sampling. For pesticides detected in periphyton but not detected in surface water, BCFs were estimated using half the limit of detection (LOD) for water concentrations.

Pesticide characteristics

Predicted No Effect Concentrations (PNECs) for pesticides on aquatic organisms were obtained from the European Food Safety Authority (EFSA) guidelines, specifically from the EFSA conclusions for the respective pesticides. Pesticide characteristics, values like Log K_OW_, water solubility, and soil degradation were retrieved from the Pesticide Properties DataBase (PPDB).

Statistical methods

Differences in periphyton pesticide concentrations, algal biomass proxies (chlorophyll a and fucoxanthin concentrations), and algal pigment signals (peak areas) among sites and months were evaluated using a two-way ANOVA, using site and month as fixed factors and site replicates as a random effect. We conducted post hoc pairwise comparisons using the Tukey’s honest significance test when ANOVA detected significant effects of site and/or month. Linear regression analyses were used to examine the relationship between i) log-transformed pesticide BCF and log K_OW_, and ii) between nutrient levels (total nitrogen, total phosphate), sum of pesticide concentration in periphyton and the combined EPA+DHA as well as fucoxanthin. Comparisons of nutrient concentrations between Streams 1 and 2 were performed using the non-parametric Mann–Whitney test, as the data were not normally distributed. Data from the REF stream were excluded from the analysis because only one observation was available. Non-metric multidimensional scaling (NMDS) was used to visualize relationships between pigments profile in algae in periphyton samples and between pesticide concentrations in surface water in Streams 1 and 2. NMDS was conducted using Vegan package in R (version 4.5.0 2025–04–11 ucrt), while other statistical analyses were conducted using JMP Pro 16.0.0. Statistical significance was set at 0.05.

Results

Pesticide residues in periphyton and surface water

A total of 30 pesticide residues were detected in periphyton (Fig. 1A and Table S3). In samples from Stream 1, 24 pesticides were identified, including 15 fungicides, 6 herbicides, and 3 insecticides reaching a maximal concentration of 436 ng/g AFDM. In Stream 2, 20 pesticides were detected, comprising 9 fungicides, 9 herbicides, and 2 insecticides, reaching a maximal concentration of 1231 ng/g AFDW. In the REF stream, 15 pesticides were found, consisting of 7 fungicides, 6 herbicides, and 2 insecticides, reaching a maximal concentration of 75 ng/g AFDW. Azoxystrobin, dimethomorph, pyraclostrobin, isoproturon, and prosulfocarb were detected in all periphyton samples. Among the pesticides with the highest concentrations were azoxystrobin, diflufenican, prosulfocarb, pyroxsulam, and tri-allate. Pesticide concentrations ranged from 0.06 to 1144 ng/g AFDM, with the highest concentration observed for pyroxsulam in periphyton collected in July from Stream 2 (Table S3). Pesticide concentrations in periphyton showed significant variation among sites and months (two-way ANOVA). For the majority of detected pesticides, concentrations in Stream 1 were significantly higher than those in Stream 2 and the REF stream (Tukey’s HSD test, P < 0.05). Pesticide concentrations in periphyton were not significantly different between Stream 2 and the REF stream overall, although that in July seemed much higher in Stream 2 than in the REF stream (Fig. 1A). Among the streams significantly higher concentrations were observed in July compared to September (P < 0.05), with August concentrations intermediate between the two months. Pesticide composition and concentration levels in pesticides accumulated in periphyton differed between streams (Fig. 1A and Figure S3). In Stream 1, fungicides was the predominant pesticide class, while in Stream 2, herbicides and fungicides were equally present with herbicides having overall higher concentrations in both streams. In the REF stream, the profile was similar to that of Stream 2, with herbicide most dominant. Additionally, differences between individual pesticide concentrations were higher in Stream 2 than in Stream 1 and REF stream. Between the different month, the chemical profile, occurrence, and concentrations of detected pesticides, are rather similar at each site (Fig. 1A).

(A) Pesticide concentrations (ng/g AFDM) -grouped into classes, fungicides, herbicides, and insecticides- in periphyton collected from REF stream, Streams 1 and 2 in July, August, and September 2022. (B) Pesticide concentrations (µg/l) in surface water from Streams 1 and 2 as mean values for July, August, and September 2022 from weekly time-integrated sampling. Water concentrations for the REF stream are omitted because they were collected using a different sampling method (grab sampling; see Figure S4 for details).

In surface water, 22, 29, and 6 pesticides were detected in Streams 1 and 2 and REF stream, respectively during July–September 2022 (Fig. 1B, Table S3 and Figure S4). Herbicide concentrations were generally more prominent than those of insecticides and fungicides. The chemical profiles exhibited temporal variability, with most detected pesticides characterized by high water solubility [e.g. metamitron, 2,6-dichlorobenzamide (BAM), and amidosulfuron, as listed in Table S1]. When comparing pesticides in periphyton and surface water, a greater number of transformation products were identified in surface water than in periphyton. Figure S5 shows the distribution of detected pesticides in water and periphyton, categorized into four Log K_OW_ ranges over the three-month period for Stream 1. In surface water, up to 50% (Stream 1) and 40% (Stream 2) of pesticides had Log K_OW_ < 2, while in periphyton, this proportion was ≤24%. In contrast, 48%–58% of pesticides in periphyton had Log K_OW_ > 3 values. To illustrate monthly fluctuations in the number of detected pesticides in surface water and periphyton, a Venn diagram was created (Figure S6). In Stream 1, more pesticides were consistently detected in periphyton than in water across all three months, with 6–9 pesticides shared between both compartments. In contrast, Stream 2 showed higher pesticide counts in water each month, with 4–7 shared detections. Noticeable was the difference at the REF stream with 15 detected pesticides in periphyton vs. 6 in surface water from grab samples.

When comparing the detected pesticides with their regulatory status in Sweden (as of 2025), we found that 38% of the compounds in water and 30% in periphyton were banned pesticides. In this context, pesticides approved only as biocides were also classified as banned, since this study focused exclusively on chemical pesticides and the sampling sites receive no wastewater effluent or other potential sources of biocides. Among herbicides, approximately half of the detected compounds were banned in both matrices (11 of 22 in water and 5 of 11 in periphyton). The majority of fungicides detected, 85% in water and 81% in periphyton, were still approved for use. Insecticides were less frequently detected: only two were found in water, of which one (imidacloprid) is banned as a pesticide but still approved as a biocide, and three were found in periphyton, with one (thiacloprid) classified as banned.

Bioconcentration factors (BCFs) were calculated for the 30 detected pesticides in periphyton (Fig. 2A). Most compounds showed bioconcentration potential. Eighteen pesticides had at least one BCF value exceeding 5000, indicating a high bioconcentration potential, while an additional eight pesticides had at least one BCF value above 2000, meeting the EU REACH criteria for bioaccumulation (EU REACH 2011; see Table 1). Importantly, this does not imply that all measurements for these substances exceeded the respective thresholds (Fig. 2A). Rather, individual samples for which BCFs could be calculated reached or surpassed these values. For example, diflufenican consistently showed BCF values above 5000, whereas prosulfocarb exceeded 5000 in 16 out of 30 detections (see Table 1). Significant linear correlation was observed between BCF and pesticide hydrophobicity (Log K_OW_) across all 30 pesticides (Figure 2B; P < 0.001) and for the 17 pesticides detected in both water and periphyton (P < 0.001; Figure S7). Pyroxsulam was excluded from both correlations as an outlier, following Ijzerman et al. (2024). Despite its low Log K_OW_ (–1.01), indicating hydrophilic properties, pyroxsulam exhibited unusually high concentrations in periphyton from Stream 2 in July, resulting in a high BCF.

(A) Bioconcentration factors (BCF) for 30 pesticides, calculated as the ratio of concentrations in periphyton to surface water (y-axis is log transformed). Dashed lines indicate REACH thresholds: 5000 (very bioaccumulative, the upper dashed red line) and 2000 (bioaccumulative, the lower dashed orange line). (B) Linear regression of BCF (log transformed) vs. Log KOW. A significant correlation was found (P < 0.001) when excluding pyroxsulam as an outlier.

In Table 1 are ranked the pesticides by maximum BCF and presented they key properties, including toxicity (predicted no-effect concentrations, PNEC), persistence (Log K_OW_ and soil DT₅₀), and water solubility. The top-ranking pesticides based on BCF were pyroxsulam, clomazone, ethofumesate, prosulfocarb, and diflufenican. Among them, diflufenican showed high BCF for all detected samples, toxicity (lowest PNEC), persistence, and low water solubility.

Algae assemblages

Mean chlorophyll a concentration ranged from 0.05 to 4.44 mg/mg AFDM in periphyton (Fig. 3A). The total amount of chlorophyll a, a proxy of algal biomass, was significantly lower at the REF stream compared to Streams 1 (Tukey’s test P < 0.05) and not significantly different to Stream 2 (Tukey’s test P = 0.064). No significant changes in chlorophyll a concentrations were observed among months at each site. Mean fucoxanthin concentrations (a biomass marker for diatoms) ranged from 0.02 to 1.48 µg/mg AFDM in periphyton (Figure S8), with significantly higher concentrations in Stream 1 compared to the REF stream (Tukey’s test, P < 0.05), and a difference that approached significance in Stream 2 (Tukey’s test, P = 0.051).

(A) Chlorophyll a concentration (mg/mg AFDM ± SD) in periphyton of each stream and sampling month. (B) Relative abundance of photosynthetic pigments in periphyton (normalized to chlorophyll a) of each stream and sampling month. Note that pigments associate to green algae are indicated in green, those associated to cyanobacteria in blue, and those associated to diatoms in brown.

The pigment profile shows that fucoxanthin and diadinoxanthin, markers for diatoms, were the most abundant pigments in the algae assemblages in Streams 1 and 2, whereas lutein and chlorophyll b, markers for green algae were more abundant in the REF stream (Fig. 3B and S8). For five out of seven pigments there are significant differences between Streams 1 and 2 to the REF stream (ANOVA P < 0.01, Tukey’s test Stream 1 and 2 ≠ REF stream P < 0.05), but not between month (ANOVA P > 0.05). NMDS analysis of pigment composition across sites (Streams 1 and 2) indicates greater similarity in pigment profiles among months and replicates within Stream 2, as reflected by the clustering of data points (Figure S9). In contrast, data points for Stream 1 are more dispersed, suggesting higher variability in pigment composition.

Nutrients and trophic status of the environment

Total nitrogen concentrations ranged from 0.512 to 15 mg/l in Stream 1 and from 6.69 to 18.4 mg/l in Stream 2 between June and October 2022 (Figure S10). The nitrogen concentrations are significant higher in Stream 2 (Mann–Witney test P < 0.005). Total phosphorus concentrations ranged from 24 to 192 µg/l in Stream 1 and from 241 to 2400 µg/l in Stream 2 over the same period. Both streams substantially exceed the good/moderate boundary concentrations for rivers in Europe, indicating that they do not meet the criteria for Good Ecological Status (GES) and are therefore classified as eutrophic (Nikolaidis et al. 2022). The REF stream sampled in June exhibited a total nitrogen concentration of 0.785 mg/l and a total phosphorus concentration of 23 µg/l, which fall within the acceptable nutrient target ranges.

Fatty acids in periphyton

Values of summed mean concentrations of all FA range between 529 and 1227 µg/g AFDM (Figure S11). Monounsaturated fatty acid and short chain saturated fatty acid had the highest concentrations. Omega-6 (ω6) and omega-3 (ω3) polyunsaturated FA were detected in all samples and in concentration up to 124 and 96 µg/g AFDM, respectively (Figure S10). Values of summed mean concentrations of two essential omega-3 FA, EPA and DHA ranged from 7 to 80 µg/g AFDM (Fig. 4A). Significant differences were observed among concentrations of summed EPA and DHA between the three sites (ANOVA, P < 0.01), but not between months. Concentrations of summed EPA and DHA in Stream 2 over the three month were higher than those in Stream 1 and the REF stream (Tukey’s test: Stream 1 > Stream 2 = REF stream). A significant linear relationship was observed between the concentration of EPA+DHA and the diatom pigment fucoxanthin (P < 0.001, Fig. 4B).

(A) Sum concentrations of the essential omega-3 fatty acids EPA and DHA in periphyton (µg/g AFDM) from the REF stream, Streams 1, and 2. Bars represent mean ± SD; n = 3. (B) Relationship between fucoxanthin concentrations and the summed concentrations of EPA and DHA (µg/g AFDM) in periphyton across all sites. Symbols represent different sites: circles = Stream 1, rectangles = Stream 2, diamonds = REF stream. Linear regressions illustrate the relationship between total EPA+DHA concentrations and fucoxanthin across varying pesticide exposures.

Positive but mostly non-significant linear relationships were observed between EPA + DHA fatty acids, the pigment marker for diatoms fucoxanthin, and total nitrogen, total phosphorus, and total pesticide concentrations in periphyton (Fig. 5). Only the relationship between fatty acids and the pesticide concentrations was statistically significant (Fig. 5; P < 0.05). The analysis is constrained by the limited dataset, which includes only data from three sampling months and three sites.

Linear relationships of concentrations of fatty acids EPA+DHA (µg/g AFDM) and diatom maker fucoxanthin (µg/g AFDM) vs. total nitrogen (mg/l), total phosphorus (µg/l), and sum of total pesticides concentrations in periphyton (ng/g AFDM, dark green).

Discussion

A variety of pesticides was detected in periphyton

Our results demonstrate that periphyton can accumulate a wide variety of pesticides, and that pesticides detected in surface water do not necessarily reflect those accumulated in periphyton. In total, up to 30 different compounds belonging to the classes of herbicides, fungicides, and insecticides, were detected in periphyton versus 37 detected in the surface water. Chemical profiles of periphyton and surface water varied among streams but remained relatively consistent across months within each stream. Notably, total pesticide concentrations were higher in periphyton in Stream 1, while surface water concentrations were higher in Stream 2. This may be due to the greater distance between periphyton sampling and the drainage from the agricultural fields in Stream 2, resulting in more diluted exposure or differnece in organic content of the periphyton. Pesticide concentrations in surface water from the REF stream were low as expected, with only six compounds detected. However, 15 pesticides were found in periphyton, with concentrations similar than those in Stream 2 for August and September (not July), which was located 23 km away. The overall insignificant differences in periphyton pesticide concentrations between Stream 2 and REF might partly be due to the lowered statistical power (for detecting differences) associated with the unequal numbers of replicates used for pesticide residue analysis in periphyton, which was a result from the pooling of samples to meet analytical mass requirements. Also, the REF stream might have been influenced by agricultural activities in the surrounding landscape of the nature reserve, via atmospheric deposition of pesticides and/or inputs from polluted groundwater, as the stream was predominantly groundwater-fed.

When examining the proportion of detected pesticides that are currently banned, 38% of those found in water and 30% of those found in periphyton are no longer approved for use in Sweden. This suggests persistence of these compounds in the environment or inputs from diffuse sources such as atmospheric deposition. The slightly higher proportion of banned pesticides in water compared to periphyton may indicate that water reflects more recent or transient inputs, or it could be related to differences in detection limits, where lower limits in water allowed a greater number of compounds to be detected.

Our findings support the hypothesis that more hydrophilic pesticides and transformation products were primarily detected in surface water, while hydrophobic pesticides were more predominant in periphyton. Recent studies have reported similar numbers of pesticides detected in periphyton, including 10–16 in samples from nine agricultural streams across southern Ontario, Canada (Ijzerman et al. 2024), 20 in samples from a costal marsh on Lake Erie, Canada downs stream from agricultural areas (Rooney et al. 2020), and up to 34 in samples from 46 streams with different land-use in California, USA (Mahler et al. 2020). Notably, the fungicide tebuconazole was the only pesticide analyzed and detected across all studies, including the present one, indicating broad use, environmental occurrence and persistence (Mahler et al. 2020, Rooney et al. 2020, Fernandes et al. 2023, Ijzerman et al. 2024). Tebuconazole is a fungicide and ecotoxicological effects on aquatic fungi have been detected at environmental concentrations elsewhere (Bertrans-Tubau et al. 2023, Gómez-Martínez et al. 2024).

The calculation of BCFs revealed that a substantial proportion of these compounds exhibit significant accumulation potential. Notably, 18 out of 30 of detected pesticides exceeded the high bioconcentration threshold of 5000, and additional eight surpassed the 2000-threshold defined under the EU REACH bioaccumulation criterion (EU REACH 2011). Under REACH, substances with a BCF above 2000 in aquatic species are considered bioaccumulative. Exceedance of these thresholds, observed for 87% of the 30 pesticides detected in periphyton, signals elevated bioaccumulation risk and highlights the importance of including periphyton in monitoring programs, as water-only measurements may underestimate exposure to hydrophobic or persistent substances. The ranking of pesticides by BCF values further emphasizes substances of concern, such as the herbicide diflufenican, which combines high bioaccumulation potential, high toxicity on algae (Gómez-Martínez et al. 2023) and overall low predicted no-effect concentrations (PNEC), strong persistence, and low water solubility characteristics that align with PBT criteria. This highlights its potential ecological risk, particularly to algal communities within the periphyton matrix. Diflufenican is widely used in agricultural systems and is frequently detected in Swedish environmental monitoring programs, underscoring its relevance for freshwater risk assessment and the need to track its fate and effects in periphyton and associated food webs.

The observed relationship between bioconcentration factors (BCFs) and pesticide hydrophobicity (Log K_OW_) supports the general expectation that physicochemical properties influence bioaccumulation. Previous studies have suggested that this relationship plays a central role in the bioaccumulation of pesticides (Boxall 2012, Lundqvist et al. 2012, Ijzerman et al. 2024). In our study, the overall correlation of BCF and Log K_OW_ across all 30 pesticides was significant, indicating that Log K_OW_ can be a useful, though not exclusive, predictor of accumulation in periphyton. Pyroxsulam represents a noteworthy exception, as despite its low Log K_OW_ and high water-solubility, it showed exceptionally high concentrations in periphyton, likely due to localized temporal high application patterns in the catchment of Stream 2. This anomaly illustrates that additional factors (e.g. amount of organic matter, species composition, pH etc.), beyond hydrophobicity, may govern pesticide accumulation in periphyton, warranting further investigation. While Ijzerman et al. (2024) reported a significant correlation between partitioning and BCFs for 29 compounds, other studies have questioned the strength or consistency of this relationship (Mahler et al. 2020, Rooney et al. 2020), highlighting the complexity of predicting bioaccumulation in natural systems.

Our findings confirm that periphyton can bioconcentrate a wide range of pesticides, supporting its role as a passive sampler in aquatic environments (Lundqvist et al. 2012, Mahler et al. 2020, Rooney et al. 2020, Ijzerman et al. 2024). This highlights the utility of periphyton for detecting bioaccumulative and PBT substances often missed by water-based monitoring alone.

Algal assemblages are impacted by nutrients

The results indicate that agricultural streams exhibit higher algal biomass compared to the REF stream, with diatoms being the dominant algal group in impacted streams. This pattern is likely driven by increased nutrient availability, particularly elevated nitrogen and phosphorus concentrations, which are known to promote algal growth in freshwater ecosystems (Dodds and Smith 2016). The pigment analysis further reinforces these trends, with fucoxanthin, diatoxanthin, and diadinoxanthin, pigments characteristic of diatoms, being dominant in agricultural streams, while lutein and chlorophyll b, markers for green algae, were more abundant at the REF stream. This shift in community composition aligns with previous studies showing that diatoms can thrive in nutrient-enriched environments, whereas green algae are more prevalent in oligotrophic conditions (Stevenson et al. 2010, Vadeboncoeur et al. 2021, Yuan et al. 2023). No significant variations in pigment composition over the three months study were detected within each stream, suggesting that in this environmental study the pigments profile mainly respond to governing and stable environmental conditions such as light availability, water flow and nutrient levels (Corcoll et al. 2012, Ponsatí et al. 2016). However, NMDS analysis revealed greater homogeneity in pigment composition in Stream 2, whereas Stream 1 exhibited more variability. This suggests that Stream 1 may experience more fluctuating environmental conditions, possibly linked to variations in nutrient input, e.g. total nitrogen concentrations varied more in Stream 1 between July and October compared to Stream 2 (Figure S8), or hydrological dynamics (Stevenson et al. 2006, Coffey et al. 2018). Nutrient concentrations in the agricultural streams were high, with total nitrogen in both Streams 1 and 2 surpassing the nutrient criteria for river water under the European Water Framework Directive (Poikane et al. 2019, Nikolaidis et al. 2022). Additionally, phosphorus levels in Stream 2 exceeded eutrophication thresholds, particular that may contribute to the observed high algal biomass and the dominance of diatoms (Nikolaidis et al. 2022).

Pesticide exposure, fatty acid quality, and algal biomass: ecological implications

Our results indicate that periphyton in agricultural streams contain higher concentrations of essential polyunsaturated FA (PUFAs), particularly EPA and DHA, compared to the REF stream. The strong positive correlation between EPA+DHA and fucoxanthin indicates that diatom biomass drives omega-3 fatty acid production, supporting fucoxanthin as an indicator of high-quality omega-3 availability in periphyton. It is well stablished that diatoms are the few life forms able to produce PUFA (Yi et al. 2017).

Correlations between EPA + DHA fatty acids and the diatom biomarker fucoxanthin with environmental variables, including total nitrogen, total phosphorus, and total pesticide concentrations, revealed generally positive but non-significant relationships. These trends suggest potential linkages between nutrient enrichment, primary producer quality, and chemical exposure in the studied streams. However, the lack of statistically significant correlations indicates that nutrient availability or pesticide exposure alone may not be the primary drivers of EPA + DHA or fucoxanthin variation under the current conditions. The limited temporal and spatial coverage (three sampling months and three sites) constrain the statistical power and generality of the analysis, reducing the ability to capture broader seasonal or site-specific dynamics. Our results aligns with other studies, supporting that environmental conditions, particularly nutrient availability and light, play a critical role in regulation the PUFA content in diatoms (Cointet et al. 2019). Contrary to our initial hypothesis, total pesticide concentrations in periphyton did not show negative impact on omega-3 fatty acids (EPA + DHA) or diatom biomass (fucoxanthin). Instead, nutrient availability seems to play a dominant role in controlling both periphyton biomass and fatty acid production, potentially masking subtle or compound-specific inhibitory effects of pesticides. Previous studies have reported that nutrients can alleviate the harmful effects of toxic chemical on periphyton (Aristi et al. 2016). Another limitation is that total pesticide concentrations are usually not a good predictor of toxicity, as it ignores differences in compound toxicity and unmeasured contaminants. At low concentrations, pesticides may not substantially impact periphyton biochemical processes, while high nutrient levels can stimulate diatom growth.

However, 67% of the pesticides detected in periphyton in this study exhibited bioconcentration potential. Toxicity risk varies among pesticides; for example, diflufenican has a low toxic threshold (PNEC = 0.025 µg/l), whereas pyroxulam was detected at higher concentrations but has a comparatively higher toxic threshold (PNEC = 0.260 µg/l). Consequently, exposure risk depends on the specific toxicity and concentration of each pesticide. Moreover, once pesticide-specific thresholds are exceeded, or through the combined presence of multiple pesticides detected in periphyton that may exert additive or synergistic effects, pesticide accumulation in periphyton can alter nutritional quality, community composition, and key functional processes, including fatty acid biosynthesis, photosynthesis, growth, and gene expression (Demailly et al. 2019). Experimental evidence supports such mechanisms. Malbezin et al. (2024) showed that 14-day exposure of periphyton to gradients of atrazine and S-metolachlor increased chlorophyll a fluorescence in both cyanobacteria and diatoms, suggesting shifts in community structure and nutritional quality, as cyanobacteria, unlike diatoms, do not produce essential fatty acids such as EPA and DHA (Napolitano 1999, Müller-Navarra et al. 2000). At higher concentrations, both herbicides also altered periphyton fatty acid profiles, reducing their nutritional value. Such effects were not observed in our study, as EPA + DHA and diatom biomass (fucoxanthin) were elevated in Streams 1 and 2 compared to the REF stream. Similarly, Chaumet et al. (2017) showed that exposure of periphyton to the herbicide diuron inhibited photosynthesis at 50 µg/l, demonstrating functional impairment of periphyton (Chaumet et al. 2017). In contrast, in our study, both EPA + DHA and fucoxanthin were elevated in the nutrient-rich streams compared to the reference site, suggesting that nutrient enrichment may have outweighed pesticide-related stress.

Furthermore, the high bioconcentration factors (BCF > 2000, REACH criteria) observed for most detected pesticides in this study indicate a potential for uptake and trophic transfer within aquatic food webs. Periphyton are a key primary food source in freshwater food webs, linking both autotrophic (green) and detrital (brown) energy pathways. Their nutritional quality and contaminant burden directly influence grazers such as insects, gastropods, and fish (Bonnineau et al. 2021). Persistent organic pollutants and metals can accumulate in periphyton and be transferred to higher trophic levels through dietary exposure (Bonnineau et al. 2021). While trophic transfer of metals is well documented, evidence for organic pollutants remains scarce. Reported cases include bioaccumulation and transfer of flame retardants (Ruhí et al. 2016) and PCBs (Walters et al. 2011). Few studies have addressed pesticide transfer specifically. In one example, stormwater-exposed periphyton fed to Neocloeon triangulifer and Planorbella pilsbryi reduced consumer survival and growth, with snail performance negatively correlated with periphyton pesticide burden (Izma et al. 2024). These findings suggest that periphyton-associated pesticides can contribute to dietary toxicity, although co-occurring contaminants may also play a role (Izma et al. 2024). Overall, periphyton represents an underrecognized pathway for contaminant transfer in aquatic food webs, and the trophic implications of pesticide accumulation in periphyton warrant further investigation.

In summary, this study demonstrates that while pesticides can accumulate in periphyton, their ecological effects are difficult to isolate from those of co-occurring stressors such as nutrient enrichment. The results indicate that nutrient availability may enhance periphyton biomass and fatty acid production, potentially masking subtle or compound-specific pesticide effects. Consequently, assessing pesticide impacts in natural systems requires experimental designs that disentangle nutrient-driven stimulation from contaminant-induced stress. Furthermore, as periphyton form a key link between primary production and higher trophic levels, their role as vectors for contaminant transfer represents an important but underexplored pathway in aquatic food webs. Future research should therefore combine field observations with controlled mesocosm studies to clarify how pesticide exposure influences both periphyton function and ecosystem-level processes.

Periphyton as passive sampler

Estimating ecological risk of pesticides in aquatic environments is challenging due to ecosystem complexity and the numerous factors that influence it (Rico et al. 2025). One way is to use indicators derived from surface water concentrations. For example, toxic units (TUs) that are based on surface water concentrations is one approach to ecological risk assessment (Nowell et al. 2014). However, the distinct differences in chemical profiles between surface water and periphyton in this study demonstrate that aquatic organisms experience different exposure conditions than those estimated solely from surface water concentrations. Periphyton revealed numerous bioconcentrated and banned pesticides, some of which were not detected in surface water, highlighting its value as a complementary indicator. These findings strongly suggest that current monitoring approaches should be improved to include biological recipients, such as periphyton, for better assessment of the ecological risks posed by pesticides.

Incorporating periphyton monitoring as a complement to water monitoring could provide a more accurate representation of long-term pesticide exposure, as periphyton not only accumulates hydrophobic contaminants but also plays a crucial role in aquatic ecosystems (e.g. as primary producers), as previous studies suggested (Sabater et al. 2007, Fernandes et al. 2023, Mahler et al. 2020, Fernandes et al. 2023, Morin and Artigas 2023). Ijzerman et al. (2024) demonstrated that periphyton analysis could distinguish the long-term pesticide usage history in catchments and detect pesticide influx pulses immediately after application. This level of detail was not achievable through active water sampling or suspended sediment monitoring. Further Mahler et al. (2020) argues that periphyton is a more effective matrix for assessing the effects of chemicals on aquatic invertebrate communities for two key reasons. First, invertebrates are directly exposed to contaminants through biofilm consumption (Izma et al. 2024). Second, periphyton accumulates a broader range of chemicals compared to sediment, providing a more comprehensive assessment of pollutant exposure. Using generalized additive models (GAMs), the study showed that periphyton was a more effective pesticide stressor than sediment for assessing chemical effects on aquatic invertebrate communities (Mahler et al. 2020). To better evaluate the ecological consequences of pesticide in aquatic environments, multiple diagnostic approaches could be employed to provide complementary lines of evidence under complex field conditions (Rico et al. 2025).

In short, our study showed that periphyton in agricultural areas bioconcentrates numerous pesticides, with a chemical profile distinct from that of water samples and contains high concentrations of important biochemical nutrients (EPA + DHA). Periphyton also plays key ecological roles, serving as a primary food source for various organisms and contributing to nutrient cycling. Assessing pesticide accumulation in periphyton, as passive samplers, can therefore enhance our understanding of ecological risks and strengthen environmental risk assessment strategies (Bonnineau et al. 2021), as it has been suggest for other contaminants such as PAHs and metals (Froehner et al. 2012, Leguay et al. 2016).

Conclusion

This study demonstrates that periphyton in agricultural streams effectively bioconcentrates a broad range of pesticides, confirming its value as a passive sampler and bioindicator of chemical exposure in aquatic ecosystems. The distinct pesticide profiles observed in periphyton compared to water and the high number of pesticides detected that showed high bioconcentration potential emphasize the need to include periphyton monitoring for detecting bioaccumulative and persistent compounds often overlooked by conventional water analyses.

Elevated nutrient concentrations in agricultural streams were associated with higher diatom biomass and increased levels of essential fatty acids (EPA + DHA), reflecting nutrient-driven enhancement of primary producer quality. While pesticides accumulated in periphyton, their ecological effects were difficult to separate from those of nutrient enrichment and potential other environmental factors (e.g. light irradiance, hydrology of the streams), suggesting that high nutrient availability may mask subtle or compound-specific toxic impacts. To better assess ecosystem responses, future studies should employ experimental approaches that disentangle nutrients and pesticide stressors.

Given periphyton’s key position at the base of aquatic food webs, its role in contaminant accumulation and trophic transfer represents an important but underexplored pathway of exposure. Future research should combine field monitoring with controlled mesocosm and trophic transfer experiments to clarify how pesticide accumulation affects periphyton function, biochemical composition, and food web dynamics. Additionally, further investigation is needed to quantify the temporal integration of contaminant uptake in biofilms under varying chemical and environmental conditions. Such advances will strengthen ecological risk assessment by capturing the dual role of periphyton as both a sensitive bioindicator and a vital trophic link in freshwater systems.

The Swedish pesticide monitoring program currently surveys surface waters and conducts annual assessments of community composition using diatom and aquatic invertebrate indices. Incorporating periphyton as an additional bioindicator within such programs could provide a more comprehensive assessment of ecological exposure, capturing both chemical accumulation and biological responses. Such advances will strengthen ecological risk assessment by reflecting the dual role of periphyton as both a sensitive integrator of environmental stressors and a vital trophic link in freshwater systems.

Supplementary Material

fiag016_Supplemental_File