Science of the Total Environment 838 (2022) 156012

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Pesticides in ambient air, influenced by surrounding land use and weather,

pose a potential threat to biodiversity and humans
⁎
Johann G. Zaller a, , Maren Kruse-Plaß b, Ulrich Schlechtriemen c, Edith Gruber a, Maria Peer a, Imran Nadeem d,
Herbert Formayer d, Hans-Peter Hutter e, Lukas Landler a
a
University of Natural Resources and Life Sciences Vienna (BOKU), Department of Integrative Biology and Biodiversity Research, Institute of Zoology, Gregor Mendel Straße 33, 1180 Vienna, Austria
b
TIEM Integrated Environmental Monitoring, 95615 Marktredwitz, Germany
c
TIEM Integrated Environmental Monitoring, Hohenzollernstr. 20, 44135 Dortmund, Germany
d
University of Natural Resources and Life Sciences Vienna (BOKU), Department of Water, Atmosphere and Environment, Institute of Meteorology and Climatology, Peter-Jordan Straße 82, 1180
Vienna, Austria
e
Department of Environmental Health, Center for Public Health, Medical University Vienna, Kinderspitalgasse 15, 1090 Vienna, Austria

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Passive air sampling detected 67 different

pesticides.
• These included 24 herbicides, 30 fungi-
cides, and 13 insecticides.
• Pesticides increased with increasing ara-
ble land in the surroundings.
• Pesticides also found in two National
Parks and a City Center.
• Potential toxicity for non-target organisms
and humans was considerable.

A R T I C L E I N F O A B S T R A C T

Editor: Daniel Wunderlin Little is known about (i) how numbers and concentrations of airborne pesticide residues are influenced by land use,
interactions with meteorological parameters, or by substance-specific chemo-physical properties, and (ii) what poten-
Keywords: tial toxicological hazards this could pose to non-target organisms including humans. We installed passive air samplers
Agrochemicals (polyurethane PUF and polyester PEF filter matrices) in 15 regions with different land uses in eastern Austria for up to
Off-target area
8 months. Samples were analyzed for 566 substances by gas-chromatography/mass-spectrometry. We analyzed rela-
Pesticide drift
Exposure
tionships between frequency and concentrations of pesticides, land use, meteorological parameters, substance proper-
Agriculture ties, and season. We found totally 67 pesticide active ingredients (24 herbicides, 30 fungicides, 13 insecticides) with
Human toxicology 10–53 pesticides per site. Herbicides metolachlor, pendimethalin, prosulfocarb, terbuthylazine, and the fungicide
Air pollution HCB were found in all PUF samplers, and glyphosate in all PEF samplers; chlorpyrifos-ethyl was the most abundant
Toxic loads insecticide found in 93% of the samplers. Highest concentrations showed the herbicide prosulfocarb (725 ±
Environmental risk assessment 1218 ng sample−1), the fungicide folpet (412 ± 465 ng sample−1), and the insecticide chlorpyrifos-ethyl (110 ±
98 ng sample−1). Pesticide numbers and concentrations increased with increasing proportions of arable land in the
surroundings. However, pesticides were also found in two National Parks (10 and 33 pesticides) or a city center (17
pesticides). Pesticide numbers and concentrations changed between seasons and correlated with land use, tempera-
ture, radiation, and wind, but were unaffected by substance volatility. Potential ecotoxicological exposure of mam-
mals, birds, earthworms, fish, and honeybees increased with increasing pesticide numbers and concentrations.
Human toxicity potential of detected pesticides was high, with averaged 54% being acutely toxic, 39% reproduction

⁎ Corresponding author.
E-mail address: johann.zaller@boku.ac.at (J.G. Zaller).

http://dx.doi.org/10.1016/j.scitotenv.2022.156012
Received 8 February 2022; Received in revised form 8 April 2022; Accepted 12 May 2022
Available online 18 May 2022
0048-9697/© 2022 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
J.G. Zaller et al. Science of the Total Environment 838 (2022) 156012

toxic, 24% cancerogenic, and 10% endocrine disrupting. This widespread pesticide air pollution indicates that current
environmental risk assessments, field application techniques, protective measures, and regulations are inadequate to
protect the environment and humans from potentially harmful exposure.

1. Introduction in eastern Austria, but also included sites in two national parks and a
city center. We distributed samplers at high densities to maximize spa-
Global pesticide use has increased dramatically in recent decades, tial resolution while studying a relatively small region. This strategy
resulting in increasing exposure of all organisms and the environment allowed us to analyze the detailed effects of landscape, weather, and
(Tang et al., 2021; Zaller, 2020). Areas near pesticide-intensive agriculture season on pesticide levels at minimal local differences in pesticide-
are particularly affected, as reported for the USA (Córdoba Gamboa et al., independent conditions.
2020; Fenske et al., 2002), China (Liu et al., 2014), Costa Rica (van We hypothesized that: (i) pesticide exposure correlates with spatial and
Wendel de Joode et al., 2012), Japan (Kawahara et al., 2005), South- temporal use in agricultural fields in the surroundings, (ii) high volatility of
Africa (Dalvie et al., 2014), northern Italy (Linhart et al., 2019; Linhart pesticides, higher temperatures, less rainfall, and low wind speeds promote
et al., 2021), Germany (Brühl et al., 2021a; Kruse-Plaß et al., 2021), and pesticide occurrence in air samples; and (iii) higher pesticide levels lead to
many other countries. Various media can be analyzed to monitor pesticide higher toxic loads of non-target organisms including humans. To the best of
contamination and it is clear that aerial dispersal is an important pathway our knowledge, the present study is novel because it interactively examines
for the spread of pesticides in the environment (Bidleman and Leone, multiple factors affecting pesticide exposure and its potential risk to nontar-
2004; Gibbs et al., 2017; Lu et al., 2000). However, little is known about get organisms, including humans.
how the number and concentration of pesticide residues in the air are af-
fected by land use, meteorological parameters, or compound-specific 2. Materials and methods
chemical-physical properties, and what toxicological hazards this might
pose to non-target organisms, including humans. 2.1. Sampling locations
In general, pesticides in air are the result of application-related drift
from (i) pesticide droplets that evaporate before they hit the target and We selected 15 sampling sites in a cultural landscape in eastern Austria
travel with fine spray particles, (ii) post application drift of vapors from to represent a land use gradient from 0% to 100% agricultural use within a
the target area, and (iii) drift from wind or wind-blown soil particles 1 km radius. Surrounding land use included arable crops, managed grass-
(Mueller, 2015). Off-target pesticide contamination in public places has lands, vineyards, or apple orchards. Additionally, one site was located in
been shown to correlate with distance from their application fields and a city center 100% surrounded by settlements, and two sites located within
the proportion of pesticide treated fields in the landscape (Linhart et al., two different National Parks of which one was surrounded by 100% forest
2019). In addition, inherent chemical properties may be important for and the other one by 73% grassland (Fig. 1). In order to protect privacy of
drift (Schampheleire et al., 2009). Some chemicals are inherently volatile the landowners, the exact location of the selected sites cannot be given
and can evaporate depending on the vapor pressure of the substance in here.
the atmosphere (the lower, the more volatile), temperature (the higher, Field mapping was carried out in July 2015. As a reference we used the
the more volatile) and relative humidity of the air (the higher humidity, official Austrian land utilization map from 2012 (Integriertes Verwaltungs-
the lower the evaporation) (Bish et al., 2021). Precipitation can reduce pes- und Kontrollsystem, INVEKOS).
ticide vaporization but increase pesticide leaching (Zaller et al., 2021).
Therefore, it can be hypothesized that lower ambient temperatures and a
low-volatility of a pesticide would reduce pesticide contamination in air
arable crop grassland settlement
samples. vineyard/apple forest water
Little research has been conducted on interactions between land use, 100
meteorology and substance properties and their toxicological consequences
of airborne pesticide exposure of wildlife or humans. Using pesticide sales
data, total potential toxic loads for honeybees (expressed as total LD50
Proportion of surrounding land use (%)

doses applied) increased even though pesticide use substantially decreased 80

over the last 25 years in the UK (Goulson et al., 2018). Potential toxic expo-
sure of aquatic insects and pollinators to insecticides has also increased over
the past 26 years in the USA as the classes of chemicals used have shifted,
60
even as the amounts applied have decreased (Schulz et al., 2021). Depend-
ing on the pesticides considered, a decrease in pesticide use may also lead to
lower toxic loads, as has been demonstrated for farmland bird species in UK
(Tassin de Montaigu and Goulson, 2020). Pesticide contamination is not 40
limited to agricultural land, but may also pose a toxicological threat to
crops cultivated near application sites, non-target plants, livestock, sensi-
tive nature conservation areas (Brühl et al., 2021a), fish and wildlife
(Tang et al., 2021). Pesticide drift can also expose farm workers, bystanders 20
and contaminate public places (Linhart et al., 2019; Linhart et al., 2021;
Paglia et al., 2021; Sapcanin et al., 2016).
We conducted this study to: 1) assess environmental exposure and sea-
sonal distribution patterns of pesticides via passive air sampling at 15 sites 0
that differed in their surrounding land use; 2) investigate the influence of 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
land use, meteorological parameters and specific chemical properties Location number
of pesticides on local contamination levels; and 3) evaluate potential
toxicological hazards of these exposures of various non-target organ- Fig. 1. Proportion of land use categories surrounding air sampler locations within a
isms and humans. The study was conducted in an agricultural landscape 1 km radius based on the Austrian land use map and orthophotos.

2
J.G. Zaller et al. Science of the Total Environment 838 (2022) 156012

Land use types were mapped within a 1 km radius of the sampler site synergists, auxiliary materials, and compounds unrelated to pesticides that
based on publicly available CORINE Land Cover database (UBA Wien and are known to exert adverse health effects and may be unintentionally pres-
EEA, 2018) and EUNIS Habitat Classification (Davies et al., 2004). The ent in agricultural products, such as polychlorinated biphenyls (PCB ́s),
most recent land cover database available was from 2018, while pesticide hexachlorobenzene (HCB), anthraquinone, dichlorobenzophenone (DCBP-
sampling was in 2020. In addition, freely available orthophotos from pp), and piperonyl butoxide (Kruse-Plaß et al., 2021).
2020 with an accuracy of 29 cm were used to adjust for potential land Calculations and statistical analyses for the current study only focused on
use changes (Geoland, 2020). Accordingly, the following land use types pesticide active ingredients and their metabolites. Of the 566 substances
were distinguished: arable crops, vineyards and apple orchards, grassland, analyzed, 192 were insecticides (34%), 141 herbicides (25%), and 116
forest, settlements, and water bodies. Mapping and analysis were per- fungicides (20%), 65 substances were metabolites (24%), and 52 other sub-
formed using ArcGis 10.2.195, QGis 2.8.1, FRAGSTATS 4.296 and stances (nematicides, rodenticides, acaricides, bactericides, molluscicides,
CHLOE201297. growth regulators, inhibitors, or safeners). This analysis tailored for
While we do not have data on the actual application of pesticides to ag- Germany included 43% of insecticides, 63% of herbicides, and 63% of fun-
ricultural fields during the study period, the land use categories indirectly gicides approved in Austria during the study year.
provide information on the seasonal patterns of pesticide application. As- Following the protocol DIN EN 15662 (July 2018) chemical substances
suming that fields were managed according to good agricultural practices, were analyzed by gas chromatography with mass spectrometry coupling
this included pesticide application, and we should be able to detect it in (GC–MS) and/or liquid chromatography-tandem mass spectrometry (LC-
the air. For example, arable crops consisting mainly of cereals are sprayed MS/MS) in PEF after acetonitrile extraction/partitioning and purification
until harvest in June/July, while vineyards and apple orchards are sprayed with dispersive SPE sample preparation (QuEChERS) (BVL, 2018). Extrac-
over a longer period from March/April until harvest in September/October. tion of PUF was performed with dichloromethane in a Soxhlet extractor
For grassland, forests, water bodies and settlements, we assume that pesti- (Estellano et al., 2015; Yusà et al., 2009) and analyzed accordingly. In a
cides are applied on spots only. separate analysis by LC-MS/MS after extraction with aqueous hydrochloric
acid and derivatization with fluorenylmethoxycarbonyl, the PEF matrix
2.2. Passive air sampling was analyzed for glyphosate and its primary degradation product
aminomethylphosphonic acid (AMPA).
Passive air sampling was conducted in accordance with protocols such The limit of quantification (LQ) was 10 ng sample−1 for most sub-
as the Global Atmospheric Passive Sampling (GAPS) network (Schuster stances. Exceptions were a LQ of 20 ng sample−1 for folpet, boscalid
et al., 2021). We used passive air samplers consisting of two matrices per (PEF), dimoxystrobin, dodine, fluopyram (PEF), fluxapyroxad (PEF),
location: a polyurethane foam (PUF) and a polyester foam (PEF) sampler. imidacloprid, prosulfocarb (PEF), tebuconazole (PEF) and thiacloprid. A
Both matrices were placed at about 1.8 m height to simulate a height LQ of 30 ng sample−1 was used for chlorothalonil, methylene cyclopropane
most related to the inhalation route in humans. The mean distance between acetic acid (MCPA), mecoprop-P, desmedipham, pendimethalin (PEF),
samplers was 99.3 ± 49.0 km (mean ± SD) with a minimum distance be- phenmedipham, propamocarb and prothioconazole-desthio (PEF). For
tween samplers of 5.0 km and a maximum distance of 196.8 km. The PUF chlorotoluron the LQ was 100 ng sample−1, and for glyphosate and
sampler consisted of a disk (diameter 14 cm, height 1 cm) under a metal AMPA 5 ng sample−1.
dome to avoid direct precipitation of dust and rain. The PEF sampler
consisted of four disks (each diameter 8 cm, height 2 cm) placed under 2.4. Meteorological data
the dome and exposed to the air, also collecting particles. PUF disks were
obtained from Tisch Environmental Inc. (Cleves, OH, USA), PEF disks The meteorological data are based on the INCA dataset (Haiden et al.,
from Freudenberg Filtration Technologies (Weinheim, Germany). The com- 2011) of the Austrian weather service ZAMG. INCA provides gridded mete-
bined sampler consisting of PUF and PEF disks was built by TIEM Integrated orological data with a temporal resolution of 1 h for the whole of Austria
Environmental Monitoring (Dortmund, Germany) and used to detect both and a spatial resolution of 1 × 1 km. For the 15 locations, the representa-
volatile and particle-bound substances such as glyphosate (Kruse-Plaß tive INCA grid was selected and the mean was calculated for the vegetation
et al., 2021). In contrast to the German study (Kruse-Plaß et al., 2021) period (10 March to 20 November) of 2020. For 6 locations, 5 additional
PEF samples were analyzed for the same spectrum of substances as PUF subperiods within the vegetation period were calculated. Period averages
samples. Passive air sampling with PUF matrices has been used in many were calculated for temperature, wind-speed, relative humidity, and radia-
studies (Gouin et al., 2008; Hayward et al., 2010; Herkert et al., 2018; tion, and totals were calculated for precipitation.
Koblizkova et al., 2012; Zhang et al., 2013).
Prior to sampling, PUF media were purified using acetone, petroleum 2.5. Ecotoxicological assessment
ether and methanol (Shoeib et al., 2008). Non-exposed samples were ana-
lyzed for both PUF and PEF matrices to account for possible contamination. All pesticides or their metabolites were characterized with respect to po-
Fifteen PUF and six PEF matrices were installed for up to eight months tential harmfulness to the environment and humans according to lists of the
(March to November 2020), six PUF matrices were replaced approximately Pesticides Properties Database (Lewis et al., 2016), the Pesticide Action
every two months during this period to account for seasonal variations in Network International (PAN, 2016), or the Austrian Federal Office for
pesticide contamination (Supplementary Table S1). Matrices were changed Food Safety (BAES, 2021) (Supplementary Table S2). Information on
by trained members of the research team using nitrile gloves and forceps ac- pesticide-specific human toxicity was assessed only for pesticides, as no suf-
cording to clear instructions, stored in a cooler, and then placed in a freezer ficient information was available for metabolites.
(−18 °C) until laboratory analysis. We converted the concentrations sample−1 into concentrations m−2 as-
suming that the area of PUF and PEF matrices exposed to air was equal to
2.3. Chemical analyses the land area where the pesticides would be deposited. A total of 311 cm2
of one PUF matrix and 205 cm2 of four combined PEF matrices were
Chemical analysis of the matrices was performed by the laboratory exposed to ambient air.
KWALIS (Fulda, Germany) registered with the German Accreditation Body In order to assess potential toxic loads of detected substances to organ-
(Deutsche Akkreditierungsstelle). Samples were analyzed for 566 chemical isms and ecosystems, we followed common approaches (DiBartolomeis
substances based on the active ingredients listed for plant-based foods in the et al., 2019; Goulson et al., 2018; Kudsk et al., 2018; Schulz et al., 2021;
official multi-method of the German Federal Office of Consumer Protection Tassin de Montaigu and Goulson, 2020) and divided the concentrations
and Food Safety (ASU L 00.00–115, 2018–10; BVL (2018). Chemical sub- m−2 by the reported toxicity/lethality effects for each class of organisms:
stances to be analyzed included pesticides, their metabolites, safeners, mammals (acute oral LD50 in mg kg−1 for rats), birds (acute LD50 mg

3
J.G. Zaller et al. Science of the Total Environment 838 (2022) 156012

kg−1 for different bird spp.), fish (acute 96 h LC50 mg l−1 different fish To test the effects of the meteorological and land-use variables on total
spp.), bees (contact acute LD50 μg bee−1 for Apis mellifera) and earthworms pesticide numbers and concentrations, we performed model selection pro-
(chronic no observable effect concentration, NOEC, on reproduction of cedures using either pesticide number per sample or total concentration
Eisenia fetida in mg kg−1). as dependent variables. The independent variables were the land-use and
Our toxic load calculations are based upon LD50, LC50 and NOEC data meteorological components. In this case, a Gaussian error distribution
for each substance from the Pesticide Properties Database accessed in with a logarithmic link (function glm) was used because an initial check
2021 (Lewis et al., 2016). Because the LD50 captures only mortality, our ap- showed a good fit. The dredge function (package MuMIn; Bartoń (2020))
proach does not evaluate sublethal effects or indirect effects, such as dietary was used to compare the AIC of all models, in addition we removed models
depletion. When the reported available LD50 values exceeded a certain with a maximum correlation coefficient between predictor variables of
value (e.g., >2000 mg kg−1 of body weight), we used the minimum value rmax > 0.5 in order to avoid issues regarding multi-collinearity (Jaffé
(here, 2000 mg kg−1). This may overestimate the toxicity of substances (2019) for R code). Model result tables were generated using the function
with very low toxicity, but we consider this a negligible effect on on the tab_model (package sjPlot (Lüdecke, 2020)). Again, model predictions
overall calculations. were estimated using ggpredict. Only data obtained from PUF samplers
The calculation of toxic loads is based on the number of LD50 doses were used in this analysis.
(in their respective units) of each substance in ambient air. This approach as- We tested the hypothesis that increasing number and/or concentra-
sumes that the substance was absorbed in its entirety by the organism. In re- tion of pesticides leads to increased toxicity by running linear models
ality, however, only a small percentage of the pesticides used will interact with total toxic loads (TTL) as the response and number of pesticides
with non-target species, perhaps none at all. We do not know the proportion or pesticide concentration in the PUF samples as the predictor variable
of each compound exposed to non-target organisms; therefore, it is impor- (function glm with a Gaussian error distribution, predictions as
tant to reiterate that this is not an attempt to estimate actual death of the above). This was done using data of calculated ecotoxicological param-
non-target organisms under consideration. We also note that the effects of eters for mammals (acute), birds (acute), fish (acute), bees (contact) and
pesticides on organisms can often be indirect (Brühl and Zaller, 2021), earthworms (chronic).
e.g., by depleting their food supply, and that sublethal effects can occur at To analyze seasonal effects, we used the temporally resolved PUF sam-
much lower doses than the LD50, which we also do not examine here. None- ples and total pesticide concentration or number as the dependent variable.
theless, we believe this is a useful approach to highlight pesticides that may Similar to the previous analyses, we performed model selection using the
be worth investigating further in terms of their potential effects. buildglmmTMB function (using a Gaussian error distribution with a loga-
Human toxicological hazards of detected pesticides was based on the in- rithmic link, after initial distribution comparisons). The full model included
terpretation given in the Pesticide Properties Database (Lewis et al., 2016) the predictor variables land use and meteorological components, and co-
and the EU pesticide database (EC, 2021a). Categories distinguished were sine and sine of time of the year in radians. Sampling location was included
a substances' cancerogeneity, reproduction toxicity, endocrine disruption in the model as random factor. Cosine and sine of time of the year provide a
(EDC), acute toxicity, specific target organ toxicity STOT RE/SE (repeated/ way to examine the inherently cyclical time variable in linear models
single exposure), skin irritation, skin sensitization, eye irritation. It is (Pewsey et al., 2013). To facilitate interpretation of the results, we present
important to note that the interpretations of these databases are a summary sine(time of the year) and cosine(time of the year) as Sept <> March and
of the main human health concerns across a number of issues. However, Dec <> June, respectively. This represents the minimum (left) and maxi-
both use a ‘weight-of-the-evidence’ approach that emphasizes caution mum (right) of the cyclic behavior of these variables. Thus, a positive effect
(Lewis et al., 2016). of “Sept <> March” would mean higher predicted values in March. Predic-
tions were calculated using ggpredict, and tables were generated using
2.6. Statistical analyses tab_model.
We also plotted the human toxicity of the detected pesticides using the
All statistical analysis were performed in R (R Core Team 2020). In a radarchart function in the package fmsb (Nakazawa, 2021).
first step, the meteorological and land-use parameters (separately) were re-
duced using principal component analysis (PCA, function in R: princomp) 3. Results
to avoid potential multicollinearity issues. The principal components
were then further used as predictor variables in the respective model selec- 3.1. Pesticides detected
tion approaches. This was done for each of the data-sets used, i.e. PUF, PEF
and time resolved PUF sampling. Only substances > LQ were included in In total, we found 67 pesticides and 4 metabolites in air sampler matri-
the statistical analysis. ces which is 27% of the 247 active ingredients approved in Austria in the
To test the effects of chemo-physical properties, meteorological, and
land-use effects on the concentration of individual pesticides, a model selec-
tion approach was performed with pesticide concentration (ng sample−1) Table 1
as dependent variable and half-life (DT50), bio-concentration factor Overview of number of pesticides detected with passive air sampler using
(l kg−1), volatility (mPa), land-use and meteorological components, and polyurethan foam (PUF) and polyethylene foam (PEF) matrices. Residues above
specific pesticide chemical (later referred to as “pesticide”), pesticide and below detection limit are listed.
class, and interaction between pesticide and pesticide class with all other Parameter/ Sampling method
parameters as predictor variables. Sampling location was included in the Number of… PUF PEF
model as a random factor. The function buildglmmTMB (buildmer package;
Samples 15 6
Voeten (2020)) was used for automatic AIC-based model selection. After Detected substances 56 38
comparing the performance of a number of different distributions Detected substances site−1 (median) 10–35 (23) 7–35 (16)
(Tweedie, Gaussian – with/without loglink - and negative binomial), the Substances detected only in this matrix 39 20
Metabolites 2 3
Tweedie distribution provided the best fit (lowest AIC). All model fits
Pesticides / metabolites detected in total 67 / 4
performed were visually checked using QQ plots generated with the Fungicides / herbicides / insecticides 30 / 24 / 13
DHARMa package (Hartig, 2019). Model predictions for the selected Volatility low (<5 mPa) / moderate / high (>10 mPa)a 58 / 1 / 5
models were performed using the ggpredict function in the ggeffects pack- Not approved substances in 2020 9
age (Lüdecke, 2018), and ANOVA tables were generated using Anova. Substances banned under the Stockholm conventionb 9

glmmTMB. This approach was performed separately for both the PUF and a
Without categorization: chlorotoluron, O,P′-DDT und P,P′-DDT.
b
PEF matrices. Stockholm Convention on persistent organic pollutants (UNEP, 2019).

4
J.G. Zaller et al. Science of the Total Environment 838 (2022) 156012

study year (Table 1). Of the pesticides detected, 45% were fungicides, 36% model selection. It can therefore be assumed that their influence, if any, is
herbicides, and 19% insecticides. The sampling locations varied consider- small (see Supplementary Table S4).
ably in total number of pesticides detected from 10 to 53. In a National
Park surrounded by 100% forest 10 pesticides were detected, in another
3.3. Relations with land use, meteorological parameters
National Park with mainly grassland in the surrounding 33 pesticides,
and in a city center surrounded by 100% settlements 17 pesticides were de-
After the model selection procedure, the final model for pesticide num-
tected (Supplementary Table S3). Nine substances (13% of 67 pesticides
bers included the land use axis: settlement vs. arable land; forest vs. arable
found) were not legally approved in Austria during the study period:
land; and an axis dominated by arable land and settlement (Fig. 2, Table 3).
carbendazim, chlorfenvinphos, cycloate, O,P'-DDT, P,P'-DDT, dichlorprop,
Thus, arable land indicates the use of pesticides. In addition, a meteorolog-
dimethenamid, HCB, gamma-HCH, metolachlor, permethrin, tri-allat
ical axis with precipitation and humidity vs. temperature and solar radia-
(Supplementary Table S4); nine substances were also banned under the
tion significantly influenced pesticide number. The final model for
Stockholm Convention on Persistent Organic Pollutants (HCB, gamma-
pesticide concentration included two land-use axes, with only the forest
HCH, 4 PCBs, 3 DDTs).
vs. arable land axes showing significance.
Most frequently found were five herbicides (metolachlor,
Total number of pesticides detected increased steeply with increasing
pendimethalin, prosulfocarb and terbuthylazine in all samples; and 2,4-D-
proportion of arable land in the surroundings and decreased with increas-
ethylhexyl in 93% of samples), two fungicides (HCB in all, and
ing forest area (Fig. 2A; Table 3). Pesticide numbers increased only slightly
chlorothalonil in 93% of samples), and one insecticide (chlorpyriphos-
(but significantly) with increasing arable land vs. settlement area (Fig. 2C,
ethyl – in 93% of samples; Table 1). The metabolite most frequently
Table 3). Regarding meteorological parameters, the number of pesticides
found in 93% of samples was prothiconazole-desthio, the metabolite of
increased with increasing air temperature and decreasing precipitation
the fungicide prothioconazole.
(Fig. 2B; Table 3).
Rankings based on mean concentrations differed from frequency
Total concentrations of pesticides increased steeply with the proportion
ranks: the highest mean concentrations were for the herbicide
of arable land and decreasing forest in the surroundings (Fig. 2D; Table 3).
prosulfocarb (725 ± 1218 ng sample−1), the fungicide folpet (412 ±
465 ng sample −1 ), the insecticide chlorpyrifos-ethyl (110 ± 98 ng
sample−1) and the metabolite prothioconazole-desthio (75 ± 125 ng 3.4. Seasonal fluctuations in pesticide contamination
sample−1) (Table 2).
Pesticide exposure showed significant seasonal cycles with the highest
3.2. Analysis of physico-chemical parameters pesticide numbers in the first half of the year (Fig. 3A) and highest pesticide
concentrations in spring (Fig. 3B).
Predicted concentrations derived from statistical models, holding all The best models for predicting seasonal variation included similar land
influencing factors (land use, meteorology) constant, showed (Supplemen- use axes as for the analysis over the entire year: proportion of settlement,
tary Fig. S1) a predominant influence of a few, highly concentrated forest, arable land and vineyards (Supplementary Table S5). As in the pre-
fungicides (mainly folpet and chlorothalonil in PUF) and herbicides vious, the proportion of arable land was found to be positively correlated
(prosulfocarb and pendimethalin in PUF; prosulfocarb and glyphosate in with pesticide concentration. For pesticide number, this was true only in
PEF). Insecticides were detected in consistently low concentrations in an interaction with time of the year (Dec <> June). Moreover, seasonal pat-
both air sampling matrices (Supplementary Fig. S1). Physicochemical prop- terns in pesticide numbers and concentrations were clearly time-dependent
erties did not appear to be related to either concentration or occurrence of and interacted with radiation, precipitation and wind speed (Fig. 3C-F). For
the measured substance, as such variables were excluded during automated both, pesticide concentration and number, lower wind combined with

Table 2
Top five most frequently found substances (only PUF matrix considered). Min values refer to lowest concentrations detected, when the substance was present. See Supple-
mentary Table S4 for all detected substances.
Substance Frequency found in Rank based on Concentration (ng sample−1) Rank based on

% samples frequency min max Mean ± SD mean conc.

Herbicides
Metolachlor 100 1 12.3 382.6 117.2 ± 98.1 3
Pendimethalin 100 1 44.9 3932.4 650.4 ± 1044.9 2
Prosulfocarb 100 1 13.7 4758.8 724.7 ± 1217.9 1
Terbuthylazine 100 1 15.5 583.6 95.3 ± 139.3 5
2,4-D-ethylhexyl 93.3 2 12.0 71.0 30.9 ± 23.5 9

Fungicides
HCB 100 1 14.0 43.4 28.1 ± 8.7 4
Chlorothalonile 93.3 2 30.6 554.4 193.6 ± 177.7 2
Folpet 86.7 3 35.5 1665.2 411.9 ± 465.3 1
Tebuconazole 80.0 4 10.4 67.7 23.6 ± 21.5 5
Tetraconazole 53.3 5 11.1 61.3 17.1 ± 21.0 7

Insecticides
Chlorpyrifos-ethyl 93.3 1 24.7 287.0 110.3 ± 98.3 1
Chlorpyrifos-methyl 53.3 2 15.5 126.5 23.8 ± 34.4 3
Tefluthrin 33.3 3 18.5 91.8 13.1 ± 25.0 4
Gamma-HCH 26.7 4 10.5 27.0 4.0 ± 7.8 5
P,P′-DDT 13.3 5 12.5 15.5 1.9 ± 5.0 7
Permethrin 13.3 5 17.3 37.5 3.7 ± 10.4 6

Metabolites
Prothioconazole-desthio 93.3 1 14.4 501.0 74.9 ± 124.7 1
P,P′-DDE 73.3 2 11.6 109.5 31.4 ± 34.6 2

5
J.G. Zaller et al. Science of the Total Environment 838 (2022) 156012

Fig. 2. Model predictions of total pesticide numbers (A,B,C) and total concentrations (D) detected with passive air samplers (PUF only) in relation to land use in the
surroundings (A,C,D) and meteorological parameters (B). Datapoints show real measurement values (above detection limits), the lines represent the predicted mean,
shaded areas 95%-confidence intervals.

higher solar radiation or lower precipitation resulted in the highest pre- time-resolved samples (four samples over the whole period) were used at
dicted values (Fig. 3C-F). fewer sites (six locations), so the statistical power to detect land use influ-
It is important to note that the primary focus of this specific analysis was ences was lower in comparison to the analysis described in the previous
to understand time-dependent and meteorological patterns. Therefore, subchapter.

Table 3 3.5. Ecotoxicology and human toxicity of detected substances

Statistical analysis of total numbers and concentrations of pesticides in passive air
samplers (PUF only) in response to land use in the surroundings and meteorological Ecotoxicological assessments based on LD50, LC50, or NOEC properties
parameters. Significant effects in bold. showed significantly increasing toxic loads with increasing pesticide
Predictors t p numbers (Supplementary Fig. S2A) and concentrations (Fig. S2A) for
most organisms. The exception was the total toxic load for bees, where no
Number of pesticides
Settlement <> arable land1 2.68 0.023 significant relationship with pesticide concentration was observed
Forest <> arable land 8.94 <0.001 (Fig. S2H).
Arable land & settlement −4.05 0.002 With the exception of bees, the slopes between toxic loads and concen-
Precipitation & humidity <> temperature & radiation 3.51 0.006 trations were steeper than those of pesticide numbers.
Concentrations of pesticides Of the 67 pesticides detected, 36 (53.7%) were classified as acutely
Settlement <> arable land 1.85 0.089 toxic to humans, 30 (44.8%) as skin irritants, 29 (43.3%) as eye irritants,
Forest <> arable land 2.91 0.013 26 (38.8%) as reproductive toxicants, 22 (32.8%) were classified as skin
<> denotes the axis range in the model, i.e. settlement arable land together with a pos- sensitizing, 16 (23.9%) as cancerogenic, 15 (22.4%) as specific target
itive t-value means that number of pesticides increases with proportion of arable land. organ toxic, and 7 (10.4%) as endocrine active (Supplementary Table S6

6
J.G. Zaller et al. Science of the Total Environment 838 (2022) 156012

Fig. 3. Pesticide numbers (A) and concentrations (B) detected with passive air samplers (PUF only) in different sampling periods. Model predictions for pesticide numbers and
concentrations including radiation and wind (C,D) or precipitation and wind (E,F) during the measurement periods. Shades in C\\F show 95%-confidence intervals.

and Table S7, Fig. 4). Four pesticides of the category STOT RE/SE cause re- Detected insecticides had the highest average human toxicity, with 92%
spiratory irritation (SE3- hazard code H335): Chlorothalonil, 2,4-D of the 13 active ingredients (a.i.) classified as acutely toxic, 54% as repro-
ethylhexyl, cypermethrin, and permethrin. duction toxicants and skin irritants, 46% as specific target organ toxicants

7
J.G. Zaller et al. Science of the Total Environment 838 (2022) 156012

Cancerogenic
approval expired), (ii) we found a strong positive correlation between agri-
100% cultural land use and pesticide levels and a negative correlation with forest
80%
area. Our analyses also showed that these patterns were driven by a few
Reprotox. Eye irrit.
highly concentrated herbicides and fungicides; insecticides were found in
60% much lower concentrations. Therefore, we assume that either drift during
application, or secondary drift after application via volatilization and dust
from soil were responsible for pesticide levels in air samples (Deziel et al.,
2015; Ward et al., 2006).
0% Surprisingly, we found 9 pesticides that were not approved in Austria
during the sampling period. Some of these pesticides have been banned
EDC Skin sens.
for decades (e.g., cycloate, DDT, HCB, permethrin) and their origin is there-
fore unclear. It could be that (i) applicators illegally imported pesticides or
depleted their stocks (Zaller, 2020), (ii) officially not approved pesticides
were used based on temporary emergency authorizations (mainly involving
neonicotinoids), (iii) samples include pesticides with very long half-lives
that could be detected decades after application, (iv) contamination from
stocks stored at farms (Veludo et al., 2022), and (v) ambient air was con-
Acute tox. Skin irrit.
taminated by non-agricultural sources (Linhart et al., 2021). Because only
Fungicides 19% of the detected pesticides were approved for non-professional users,
Herbicides
Insecticides STOT RE/SE
we concluded that the majority of residues in our samples came from agri-
cultural applications. None of the detected pesticides were approved in
organic farming. There is also the possibility that pesticides not approved
Fig. 4. Percentage contribution to human toxicological classifications of fungicides
(30 active ingredients), herbicides (24 a. i..), and insecticides (13 a.i.) detected with in Austria entered the study areas from neighboring countries, as the dis-
passive air samplers (PUF only). Only unambiguous classifications were considered. tance to the national border of these countries was only between 0.5 and
Interpretation of the human health hazards based on the Pesticide Properties 4 km, depending on the sample location.
Database (Lewis et al., 2016) and EU Pesticide Database (EC, 2021a). A relationship between pesticides in the air and their agricultural use in
the surrounding area has also been shown for peach in Israel (Zivan et al.,
2017), apple and vine orchards in Northern Italy (Linhart et al., 2019;
and eye irritants, 38% as cancerogenic, 31% as endocrine disruptors Linhart et al., 2021) or a mixed agricultural landscape in the Tuscany region
(EDCs), and 23% as skin sensitizers (Fig. 4). in Italy (Estellano et al., 2015). The frequency of pesticide classes found in
Detected herbicides ranked second in average human toxicity, with our air samples most likely depended on their use in surrounding agricul-
50% of the 24 a.i. classified as eye irritant, 46% as acute toxic and skin irri- ture. Fungicides and herbicides are the dominant pesticide classes in arable
tant, 42% as skin sensitizing, 33% as reproductive toxic, 21% as specific tar- farming in Austria (AGES, 2020) and these were also the most frequently
get toxic, 17% as cancerogenic, and 4% as EDCs. found substance classes in our air samples. In landscape more dominated
Detected fungicides were the least toxic to humans: 43% of the 30 a.i. by apple orchards and viticulture fungicides were more dominant in the
classified as acutely toxic, 40% as skin irritants, 37% as eye irritants and re- surroundings (Linhart et al., 2019; Linhart et al., 2021).
productive toxicants, 30% as skin sensitizers, 23% as cancerogenic, and 7%
as EDCs. 4.2. Properties of detected pesticides

4. Discussion Five herbicides were detected in all samples; one herbicide, two fungi-
cides, an insecticide, and the metabolite of a fungicide in 93% of samples.
We found that the numbers and concentrations of pesticides (and some This indicates that chemical (adjuvants), technical (noozles to regulate drop-
metabolites) in the ambient air of a cultural landscape in Austria depend let sizes), and regulatory (no-spray restrictions at high wind speeds) mea-
strongly on the surrounding agricultural use. Analysis of seasonal patterns sures to prevent pesticide drift (Katzman et al., 2021; Mesnage, 2021) were
showed a peak in pesticide numbers and concentrations in spring which apparently not sufficient to prevent air contamination in our study region.
reflected the main pesticide application activities in the surroundings We also found glyphosate, the most commonly used herbicide world-
agricultural fields. Pesticide numbers and concentrations also showed a wide, in all samples of the PEF matrix at mean concentrations of 116 ±
positive relationship with temperature and radiation, but a negative rela- 90 ng sample−1. This is comparable to a study using similar air samplers
tionship with precipitation and relative humidity. Contrary to our expecta- where glyphosate was the only among 500 analyzed substances that was
tions, the inherent chemical volatility of pesticides did not affect pesticide found in all locations across Germany (Kruse-Plaß et al., 2021). In general,
numbers or concentrations (i.e., was excluded during the model selection there were large differences in the pesticides detected in the current study
procedure). Ambient air contamination with multiple pesticides was not a compared to the German study (Kruse-Plaß et al., 2021), indicating sub-
problem limited to agricultural lands. We found 10 pesticides in samples stantial differences in air pollution from pesticides even in similarly struc-
from a National Park surrounded with 100% forest, 33 pesticides in another tured landscapes. Analyzing the PEF samples for the same range of
National Park with 73% grassland surrounding, and 17 pesticides in pesticides as the PUF samples increased the number of detected substances
samples of a city center surrounded by 100% settlements. per site. Thus, such a combined analysis allowed the detection of pesticide
occurrence comparable to active sampling data (Kruse-Plaß et al., 2021).
4.1. Factors affecting pesticides in ambient air The neonicotinoid insecticide in our sample, imidacloprid, had a grace
period until 2022, while other neonicotinoids have been banned for out-
Of the 67 pesticides in the air samples, 90% were approved for use in ar- door use in the EU since 2018 (EC, 2021b). Neonicotinoids have been
able farming, 82% in horticulture, 46% in fruit orchards, 45% in ornamen- linked to detrimentally affect insects (Simon-Delso et al., 2015; van der
tals, 37% in vineyards, 19% for private use in home and small gardens, 9% Sluijs, 2020), soil organisms (van Hoesel et al., 2017), and were found in or-
in forestry, and 6% in grassland in Austria (BAES, 2021). Hence, the main ganically farmed soils and crops (Humann-Guilleminot et al., 2019) and in
source of pesticides in air samples was most likely conventional agriculture. carnivorous and insectivorous birds (Humann-Guilleminot et al., 2021).
This can be concluded because (i) the majority of detected pesticides have Emergency authorizations in Austria (and other EU member states) still
been approved for use in conventional agriculture (but for some the allowed their use for short periods in justifiable cases.

8
J.G. Zaller et al. Science of the Total Environment 838 (2022) 156012

Overall, we were surprised to find so many pesticides in ambient air, as Potential impacts of these usually low contaminations are poorly under-
only 7% of the 67 pesticides were classified as highly volatile. In addition, stood, but it can be expected that organisms in these ecosystems might be
the most commonly found herbicides (metolachlor, pendimethalin, especially sensitive to additional exposures. Although, the amounts found
prosulfocarb, terbuthylazine and 2,4-D-ethylhexyl), fungicides (HCB, appear negligible, it is important to emphasize that environmental risk
chlorothalonile), and insecticides (chlorpyrifos-ethyl) are all classified as assessments are conducted primarily for single substances and that interac-
low volatile. This is noteworthy because the risk assessment of pesticides tions among multiple pesticides, such as those found in the current study,
in Europe is also partly based on the volatility of a substance (EFSA, remain unclear.
2014). Thus, our data show that these models seem to underestimate the
real situation for non-target organisms, users, bystanders, and neighbors 4.4. Potential human health impacts
(Clausing, 2020b).
Drift of the herbicides pendimethalin and prosulfocarb causes problems It was worrying to see that half of the detected pesticides (36 out of 67)
when they contaminate neighboring fields, especially when organic prod- were classified as acutely toxic, 39% with reproductive toxicity, 24% as
ucts contaminated with these pesticides can no longer be marketed cancerogenic, and 10% as endocrine active. The risks of pesticide exposure
(Hofmann and Schlechtriemen, 2017). Additionally, sensitive crops of humans especially in agricultural regions, have been documented from
(Centner, 2021a) or nature conservation (Brühl et al., 2021a) areas in the various regions including the USA (Lu et al., 2000), Costa Rica (van
surroundings can be negatively affected by pesticide drift. Although the Wendel de Joode et al., 2012), Italy (Linhart et al., 2021), India (Gupta,
current study was conducted in Austria, the phenomenon of pesticide pollu- 2004), and South Africa (Chetty-Mhlanga et al., 2021).
tion is a global one (Zaller, 2020) and it has been estimated that 64% of the Quantifying the hazard posed by airborne pesticides is difficult. Our ap-
world's agricultural land is at risk of pesticide contamination by more than proach was to evaluate the detected pesticides based on the hazard state-
one active ingredient (Tang et al., 2021). ments provided by the manufacturers but it is difficult to extrapolate
those data to realistic scenarios of pesticide exposure of humans living in
4.3. Toxic load for the environment the study region. However, the point we wanted to make is that humans
in the study regions were exposed to substances with different human
Pesticides generally have a strong impact on various non-target health risks. Such qualitative decisions without thresholds are also made
organisms and are discussed as drivers of biodiversity decline (Brühl and in the approval of active substances in Europe (Clausing, 2020a). As soon
Zaller, 2019; Habel et al., 2019a; Jactel et al., 2020; Poisson et al., 2021; as a substance is classified as cancerogenic, genotoxic or toxic to reproduc-
Wagner, 2020). Therefore, we expressed the toxic load of airborne pesti- tion no approval may be granted and a quantitative risk assessment is not
cides to mammals, birds, fish, bees and earthworms based on research by necessary (Clausing et al., 2018). Of particular concern for airborne con-
several authors (DiBartolomeis et al., 2019; Goulson et al., 2018; Kudsk tamination are pesticides, which are known to irritate the respiratory
et al., 2018; Schulz et al., 2021; Tassin de Montaigu and Goulson, 2020). tract (STOT RE/SE - hazard code H335). Of these, we found the fungicide
Overall, this is a coarse approach and provides an indication of the potential chlorothalonil and the herbicide 2,4-D ethylhexyl in almost all samples
toxic exposure, but does not imply that a specific number of animals will be (93%) and, less frequently, the insecticides permethrin (13%) and
killed. Clearly, field-accurate, spatially referenced use data paired with bio- cypermethrin (7%). Inhaled pesticides can act differently in the human
diversity monitoring would allow for more accurate risk assessments body when ingested via contaminated food (Clausing, 2020b). However,
(Mesnage et al., 2021). toxic assessment via ambient air is insufficiently considered during pesti-
Our models showed that the toxic load for all groups of organisms con- cide approval (Clausing, 2020a) and it is argued that the registration pro-
sidered (with the exception of pesticide concentration and bees) increased cess for pesticides is generally a rather outdated assessment system that
significantly with increasing number of pesticides and concentration. This undervalues health effects (Centner, 2021b).
is a non-trivial result, as pesticides vary greatly in toxicity and a few very We also found 9 pesticides listed in the Stockholm Convention of Persis-
toxic substances can be more toxic than a high number of harmless sub- tent Organic Pollutants (UNEP, 2019). Persistent organochlorinated pesti-
stances (Schulz et al., 2021; Cech et al., 2022). cides such as the fungicide HCB, which was found in 100% of our
Non-target effects of pesticides have been reported for groups of organ- samples, pose a public health problem due to their lipophilic properties,
isms as diverse as plants (Schmitz et al., 2013), soil biota (Gaupp- which cause bioaccumulation and biomagnification (Mrema et al., 2013).
Berghausen et al., 2015; Gunstone et al., 2021; van Hoesel et al., 2017; The pesticides with the highest average concentrations across sample
Zaller et al., 2021), bees (Sanchez-Bayo and Goka, 2014; Sánchez-Bayo sites were the herbicide prosulfocarb (725 ± 1218 ng sample−1), the
et al., 2016; Sánchez-Bayo and Tennekes, 2020), amphibians (Adams fungicide folpet (412 ± 465 ng sample−1) and the insecticide
et al., 2020; Baier et al., 2016; Brühl et al., 2013; Leeb et al., 2020a), bats chlorpyrifos-ethyl (110 ± 98 ng sample−1). Prosulfocarb is contained in
(Stahlschmidt and Brühl, 2012; Stahlschmidt et al., 2017), butterflies at 7 products approved in Austria and widely used in arable farming and hor-
all life stages (Krishnan et al., 2021), and other organisms important for ticulture; folpet is contained in 27 fungicides applied in vineyards, cereals
the functioning of agroecosystems (Zaller and Brühl, 2019). In addition to and horticulture. Chlorpyrifos-ethyl, was widely used in arable farming,
direct toxic effects, amphibians showed avoidance behavior to fungicides fruit and wine orchards and horticulture but has been banned in EU since
and herbicides (Leeb et al., 2020b). Spray drift of herbicides and other pes- 2020.
ticides had a significantly negative effect on wildflowers (Brühl and Zaller, Although the amounts found in ambient air were low, it is of concern
2021; Hahn et al., 2015; Strandberg et al., 2021) that affect floral resources that 7 (10%) of the pesticides (dimoxystrobin, HCB, bromoxynil,
for insect pollinators (Habel et al., 2019b; Kratschmer et al., 2021). chlorfenvinphos, permethrin, O,P'-DDT, thiacloprid) were endocrine active
We were surprised to find agricultural pesticides in two National Parks or endocrine disrupting chemicals (EDCs), and 25 pesticides are possibly
and a city center as well. This is concerning because National Parks are sup- EDCs. These substances affect hormone balance in fetuses, children and ad-
posed to protect biodiversity and urban biodiversity is already stressed by olescents (Kahn et al., 2020) and can be effective at very low doses
air pollution, noise and light. Pesticide contamination has also been found (Mesnage and Antoniou, 2021). Importantly, even very low levels of pesti-
in other pristine environments such ice cores of alpine glaciers in Europe cide exposure over years can affect children's development, health, and
(chlorpyrifos; Rizzi et al. (2019), in the Arctic (Balmer et al., 2019) and in behavior (Pascale and Laborde, 2020). Simultaneous exposure of a variety
insects collected from nature conservation areas in Germany (Brühl et al., of pesticides could also trigger synergistic effects (Laetz et al., 2009), which
2021a). Using passive air samplers similar to this study, pesticides (endo- are extremely difficult to study in risk assessments (Damalas and
sulfan, cypermethrin, chlorpyrifos) have even been detected in high alti- Eleftherohorinos, 2011). Because of the low-dose effect, the non-
tude national parks at >2200 m altitude in Brazil (Guida et al., 2018) or monotonic dose-response relationship and the interaction of EDCs with en-
in the Bolivian Andes at altitudes up to 5200 m (Estellano et al., 2008). dogenous hormones or other EDCs, it is questionable whether a low dose

9
J.G. Zaller et al. Science of the Total Environment 838 (2022) 156012

can ever be considered safe (Munn and Goumenou, 2013; Vandenberg Consent for publication
et al., 2012).
Among the pesticides detected, the insecticides chlorpyrifos-ethyl and All authors approved the final version of this manuscript for publication.
chlorpyrifos-methyl were among the most hazardous. Since 2020, these
two insecticides have been banned in the EU and airborne contamination Ethics approval and consent to participate
should be reduced in the future. Residues of these chemicals have also
been found on children's playgrounds near intensively managed apple Not applicable.
orchards (Linhart et al., 2019) and other public lands (Linhart et al.,
2021), and in urine samples of agricultural workers and residents living Funding
in agricultural areas (Paglia et al., 2021). Even very low doses of chlorpyr-
ifos have been shown to cause brain abnormalities in fetuses and children The association “Verein zur Förderung einer enkeltauglichen Umwelt in
(Rauh et al., 2012), and impair locomotor activity, behavior, and neuro- Österreich” funded the chemical analyses but had no role in study design,
transmitter systems in rats (Perez-Fernandez et al., 2020). data analysis, decision to publish, or preparation of the manuscript.
All the sites we studied had multiple contamination with pesticides, but
the effects of multiple exposures are rarely studied. Such studies would be Availability of data and materials
necessary since it is obvious that people living in such areas are simulta-
neously exposed by breathing the contaminated air. The datasets used and/or analyzed during the current study are avail-
able from the corresponding author on reasonable request.
5. Conclusions
References
We detected multiple pesticides in the air in an Austrian cultural land-
scape not only in agricultural areas, but also in two National Parks and Adams, E., Gerstle, V., Brühl, C.A., 2020. Dermal fungicide exposure at realistic field rates in-
duces lethal and sublethal effects on juvenile European common frogs (Rana temporaria).
even inmidst of a large city. Several of these pesticides potentially pose a Environ. Toxicol. Chem. 40 (5), 1289–1297 pp. 22pp.
toxicological risk to organisms, including humans, but the effects of such AGES, 2020. Pflanzenschutzmittel-Verwendungsstatistik. Austrian Agency for Health and
contamination are not well known. One aspect that is rarely addressed in Food Security accessed 04Oct2021 https://www.ages.at/themen/landwirtschaft/
pflanzenschutzmittel/forschung/pflanzenschutzmittel-verwendungsstatistik/.
scientific and public debate is that pesticides found in air samples may BAES, 2021. Pflanzenschutzmittel-Register - Verzeichnis der in Österreich zugelassenen/
also be deposited in private gardens, nature reserves, or organic farm fields. genehmigten Pflanzenschutzmittel. Bundesamt für Ernährungssicherheit. Fachbereich
Our findings also suggest that regulatory guidance on operator, worker, Pflanzenschutzmittel. https://psmregister.baes.gv.at/psmregister/faces/main?_afrL
oop=4739298204069462&_afrWindowMode=0&_adf.ctrl-state=begqx7oro_4.
resident, and bystander exposure (EFSA, 2014) is inadequate. Therefore,
Baier, F., Gruber, E., Hein, T., Bondar-Kunze, E., Ivanković, M., Mentler, A., et al., 2016. Non-
we recommend a general reduction of pesticide exposure as foreseen in target effects of a glyphosate-based herbicide on common toad larvae (Bufo bufo, Am-
the European form-to-fork strategy for sustainable agriculture (Brühl phibia) and associated algae are altered by temperature. PeerJ 4, e2641. https://doi.
et al., 2021b; EC, 2019). A prerequisite for further research would be sys- org/10.7717/peerj.2641.
Balmer, J.E., Morris, A.D., Hung, H., Jantunen, L., Vorkamp, K., Rigét, F., et al., 2019. Levels
tematic monitoring of pesticides in air and other media, as already exists and trends of current-use pesticides (CUPs) in the arctic: an updated review, 2010–2018.
for pesticide monitoring of food (EPRS, 2018). Overall, our results show Emerg. Contam. 5, 70–88.
that pesticide application practices should be improved as precautionary Bartoń, K., 2020. MuMIn: Multi-Model Inference. Version R package version 1.43.17. https://
CRAN.R-project.org/package=MuMIn.
measures to protect human and environmental health from uncontrolled Bidleman, T.F., Leone, A.D., 2004. Soil–air exchange of organochlorine pesticides in the
pesticide exposure. Southern United States. Environ. Pollut. 128, 49–57.
Supplementary data to this article can be found online at https://doi. Bish, M., Oseland, E., Bradley, K., 2021. Off-target pesticide movement: a review of our cur-
rent understanding of drift due to inversions and secondary movement. Weed Technol.
org/10.1016/j.scitotenv.2022.156012. 35, 345–356.
Brühl, C.A., Zaller, J.G., 2019. Biodiversity decline as a consequence of an inappropriate envi-
CRediT authorship contribution statement ronmental risk assessment of pesticides. Front. Environ. Sci. 7, 177. https://doi.org/10.
3389/fenvs.2019.00177.
Brühl, C.A., Zaller, J.G., 2021. Indirect herbicide effects on biodiversity, ecosystem functions,
MKP, US planned and organized sampling and analyses; MKP, EG, MP, and interactions with global changes. In: Mesnage, R., Zaller, J.G. (Eds.), Herbicides:
IN, HF, H-PH, JGZ performed landscape analyses, contributed meteorolog- Chemistry, Efficacy, Toxicology, and Environmental Impacts. Elsevier, Amsterdam, NL,
pp. 231–272.
ical data, compiled information on substance properties and assessed their
Brühl, C.A., Schmidt, T., Pieper, S., Alscher, A., 2013. Terrestrial pesticide exposure of am-
potential effects; LL performed statistical analyses and made the graphs. phibians: an underestimated cause of global decline? Sci. Rep. 3, 1135.
JGZ, and LL wrote the first manuscript draft; all authors edited the final Brühl, C.A., Bakanov, N., Köthe, S., Eichler, L., Sorg, M., Hörren, T., et al., 2021a. Direct pes-
manuscript. ticide exposure of insects in nature conservation areas in Germany. Sci. Rep. 11, 24144.
Brühl, C.A., Zaller, J.G., Liess, M., Wogram, J., 2021b. The rejection of synthetic pesticides in
organic farming has multiple benefits. Trends Ecol. Evol. https://doi.org/10.1016/j.tree.
Declaration of competing interest 2021.11.001.
BVL, 2018. Untersuchung von Lebensmitteln - Multiverfahren zur Bestimmung von
Pestizidrückständen mit GC und LC nach Acetonitril-Extraktion/Verteilung und
The authors declare that they have no known competing financial inter- Reinigung mit dispersiver SPE in pflanzlichen Lebensmitteln - Modulares QuEChERS-
ests or personal relationships that could have appeared to influence the Verfahren. Beuth Verlag, Berlin, Germany.
work reported in this paper. Cech, R.M., Jovanovic, S., Kegley, S., Hertoge, K., Leisch, F., Zaller, J.G., 2022. Reducing overall
herbicide use may reduce risks to humans but increase toxic loads to honeybees, earth-
worms and birds. Environ. Sci. Eur. 34, 44. https://doi.org/10.1186/s12302-022-00622-2.
Acknowledgements Centner, T.J., 2021a. Creating a compensation program for injuries from dicamba spray drift
and volatilization. Appl. Econ. Perspect. Policy https://doi.org/10.1002/aepp.13167.
Centner, T.J., 2021b. Pesticide usage is compromising people's health in the United States:
We are grateful to the association “Verein zur Förderung einer ideas for reducing damages. Agriculture 11, 486.
enkeltauglichen Umwelt in Österreich” (translated to grandchildren-fit en- Chetty-Mhlanga, S., Fuhrimann, S., Basera, W., Eeftens, M., Röösli, M., Dalvie, M.A., 2021. As-
vironment for Austria) for funding sampling, chemical analyses and data sociation of activities related to pesticide exposure on headache severity and
neurodevelopment of school-children in the rural agricultural farmlands of the Western
analyses. Thanks to Gabriele Moder for information on organic farming
Cape of South Africa. Environ. Int. 146, 106237.
products. LL was supported by the Austrian Science Fund FWF (P32586). Clausing, P., 2020a. Baumrinden-Monitoring der Pestizid-Belastung über die Luft: Eine
Open access funding was provided by University of Natural Resources toxikologische Bewertung. Bündnis für eine enkeltaugliche Landwirtschaft. https://
and Life Sciences Vienna (BOKU). www.enkeltauglich.bio/wp-content/uploads/2020/02/Tox-Bewertung-Peter-Clausing-
02.pdf pp. 22pp.
Thanks to Peter Clausing (PAN Germany), and Lara Petschik (University Clausing, P., 2020b. Wertlose Werte. In: Forster, M., Schümann, C. (Eds.), Das Gift und Wir.
of Landau, Germany) for discussing various aspects with us. Westend Verlag, Frankfurt/Main, Germany, pp. 98–106.

10
J.G. Zaller et al. Science of the Total Environment 838 (2022) 156012

Clausing, P., Robinson, C., Burtscher-Schaden, H., 2018. Pesticides and public health: an anal- earthworms, soil microorganisms, and litter decomposition. Front.Plant Sci. 8, 215.
ysis of the regulatory approach to assessing the carcinogenicity of glyphosate in the https://doi.org/10.3389/fpls.2017.00215.
European Union. J. Epidemiol. Community Health 72, 668–672. Hofmann, F., Schlechtriemen, U., 2017. Biomonitoring der Immissionsbelastung von
<collab>UBA Wien, E.E.A.collab, 2018. CORINE-Landbedeckung 2018, Erhebungszeitraum Glyphosat, Glufosinat und AMPA sowie weiteren PSM-Wirkstoffen mittels Luftgüte-
212-2018, Nomenklatur: level 3 (28/41 Klassen in Österreich). https://www.data.gv. Rindenmonitoring-Ergebnisse Screening 2017. https://doi.org/10.13140/RG.2.2.
at/katalog/dataset/76617316-b9e6-4bcd-ba09-e328b578fed2 pp. 22pp. 12733.15842.
Córdoba Gamboa, L., Solano Diaz, K., Ruepert, C., van Wendel de Joode, B., 2020. Passive Humann-Guilleminot, S., Binkowski, Ł.J., Jenni, L., Hilke, G., Glauser, G., Helfenstein, F.,
monitoring techniques to evaluate environmental pesticide exposure: results from the In- 2019. A nation-wide survey of neonicotinoid insecticides in agricultural land with impli-
fant's Environmental Health study (ISA). Environ. Res. 184, 109243. cations for agri-environment schemes. J. Appl. Ecol. 56, 1502–1514.
Dalvie, M.A., Sosan, M.B., Africa, A., Cairncross, E., London, L., 2014. Environmental monitor- Humann-Guilleminot, S., Laurent, S., Bize, P., Roulin, A., Glauser, G., Helfenstein, F., 2021.
ing of pesticide residues from farms at a neighbouring primary and pre-school in the Contamination by neonicotinoid insecticides in barn owls (Tyto alba) and Alpine swifts
Western Cape in South Africa. Sci. Total Environ. 466–467, 1078–1084. (Tachymarptis melba). Sci. Total Environ. 785, 147403.
Damalas, C.A., Eleftherohorinos, I.G., 2011. Pesticide exposure, safety issues, and risk assess- Jactel, H., Imler, J.L., Lambrechts, L., Failloux, A.B., Lebreton, J.D., Le Maho, Y., et al., 2020.
ment indicators. Int. J. Environ. Res. Public Health 8, 1402–1419. Insect decline: immediate action is needed. C. R. Biol. 343, 267–293.
Davies, C.E., Moss, D., Hill, M.O., 2004. EUNIS Habitat Classification Revised 2004. European Jaffé, R., 2019. dredge_mc: Assess multicollinearity between predictors when running the
Topic Centre on Nature Protection and Biodiversity, Paris, p. 310. dredge function (MuMIn - R). https://github.com/rojaff/dredge_mc.
Deziel, N.C., Friesen, M.C., Hoppin, J.A., Hines, C.J., Thomas, K., Freeman, L.E.B., 2015. A re- Kahn, L.G., Philippat, C., Nakayama, S.F., Slama, R., Trasande, L., 2020. Endocrine-disrupting
view of nonoccupational pathways for pesticide exposure in women living in agricultural chemicals: implications for human health. Lancet Diabetes Endocrinol. 8, 703–718.
areas. Environ. Health Perspect. 123, 515–524. Katzman, D., Bohbot-Raviv, Y., Dubowski, Y., 2021. Does polyacrylamide-based adjuvant ac-
DiBartolomeis, M., Kegley, S., Mineau, P., Radford, R., Klein, K., 2019. An assessment of acute tually reduce primary drift of airborne pesticides? Sci. Total Environ. 775, 145816.
insecticide toxicity loading (AITL) of chemical pesticides used on agricultural land in the Kawahara, J., Horikoshi, R., Yamaguchi, T., Kumagai, K., Yanagisawa, Y., 2005. Air pollution
United States. PLOS ONE 14, e0220029. and young children's inhalation exposure to organophosphorus pesticide in an agricul-
EC, 2021a. EU Pesticides Database. https://ec.europa.eu/food/plant/pesticides/eu-pesticides- tural community in Japan. Environ. Int. 31, 1123–1132.
database/public/?event=homepage&language=EN. (Accessed 21 May 2021). Koblizkova, M., Genualdi, S., Lee, S.C., Harner, T., 2012. Application of sorbent impregnated
EC, 2021b. Some facts about neonicotinoids. European Commission. https://ec.europa.eu/food/ polyurethane foam (SIP) disk passive air samplers for investigating organochlorine pesti-
plants/pesticides/approval-active-substances/renewal-approval/neonicotinoids_en#current- cides and polybrominated diphenyl ethers at the global scale. Environ. Sci. Technol. 46,
status-of-the-neonicotinoids-in-the-eu. 391–396.
EC, 2019. The European Green Deal: Communication from the Commission to the Kratschmer, S., Pachinger, B., Gaigher, R., Pryke, J.S., van Schalkwyk, J., Samways, M.J., et
European Parliament, The European Council, The Council, The European Economic al., 2021. Enhancing flowering plant functional richness improves wild bee diversity in
and Social Committee and the Committee of the Regions. COM(2019) 640 final vineyard inter-rows in different floral kingdoms. Ecol. Evol. 11, 7927–7945.
2019. https://eur-lex.europa.eu/resource.html?uri=cellar:b828d165-1c22-11ea- Krishnan, N., Zhang, Y., Aust, M.E., Hellmich, R.L., Coats, J.R., Bradbury, S.P., 2021. Monarch
8c1f-01aa75ed71a1.0002.02/DOC_1&format=PDF. butterfly (Danaus plexippus) life-stage risks from foliar and seed-treatment insecticides.
EFSA, 2014. Guidance on the assessment of exposure of operators, workers, residents and by- Environ. Toxicol. Chem. 17.
standers in risk assessment for plant protection products. EFSA J. 12 (10), 3874 55pp. Kruse-Plaß, M., Hofmann, F., Wosniok, W., Schlechtriemen, U., Kohlschütter, N., 2021. Pesti-
EPRS, 2018. Directive 2009/128/EC on the sustainable use of pesticides . http://www. cides and pesticide-related products in ambient air in Germany. Environ. Sci. Eur. 33, 114.
europarl.europa.eu/RegData/etudes/STUD/2018/627113/EPRS_STU(2018)627113_EN. Kudsk, P., Jørgensen, L.N., Ørum, J.E., 2018. Pesticide load—a new Danish pesticide risk in-
pdf. (Accessed 13 October 2019). dicator with multiple applications. Land Use Policy 70, 384–393.
Estellano, V.H., Pozo, K., Harner, T., Franken, M., Zaballa, M., 2008. Altitudinal and seasonal Laetz, C.A., Baldwin, D.H., Collier, T.K., Hebert, V., Stark, J.D., Scholz1, N.L., 2009. The syn-
variations of persistent organic pollutants in the Bolivian Andes mountains. Environ. Sci. ergistic toxicity of pesticide mixtures: implications for risk assessment and the conserva-
Technol. 42, 2528–2534. tion of endangered Pacific salmon. Environ. Health Perspect. 117, 348–353.
Estellano, V.H., Pozo, K., Efstathiou, C., Pozo, K., Corsolini, S., Focardi, S., 2015. Assessing Leeb, C., Brühl, C., Theissinger, K., 2020a. Potential pesticide exposure during the post-
levels and seasonal variations of current-use pesticides (CUPs) in the Tuscan atmosphere, breeding migration of the common toad (Bufo bufo) in a vineyard dominated landscape.
Italy, using polyurethane foam disks (PUF) passive air samplers. Environ. Pollut. 205, Sci. Total Environ. 706, 134430.
52–59. Leeb, C., Kolbenschlag, S., Laubscher, A., Adams, E., Brühl, C.A., Theissinger, K., 2020b.
Fenske, R.A., Lu, C., Barr, D., Needham, L., 2002. Children's exposure to chlorpyrifos and para- Avoidance behavior of juvenile common toads (Bufo bufo) in response to surface contam-
thion in an agricultural community in central Washington State. Environ. Health ination by different pesticides. PLOS ONE 15, e0242720.
Perspect. 110, 549–553. Lewis, K.A., Tzilivakis, J., Warner, D.J., Green, A., 2016. An international database for pesti-
Gaupp-Berghausen, M., Hofer, M., Rewald, B., Zaller, J.G., 2015. Glyphosate-based herbicides cide risk assessments and management. Hum. Ecol. Risk Assess. Int. J. 22, 1050–1064.
reduce the activity and reproduction of earthworms and lead to increased soil nutrient Linhart, C., Niedrist, G.H., Nagler, M., Nagrani, R., Temml, V., Bardelli, T., et al., 2019. Pesti-
concentrations. Sci. Rep. 5, 12886. cide contamination and associated risk factors at public playgrounds near intensively
Geoland, 2020. Orthofoto TileCache of Austria. published by geoland.at https://www.data. managed apple and wine orchards. Environ. Sci. Eur. 31, 28.
gv.at/katalog/dataset/254757be-69ef-4a6c-a4c1-1432815d7522. Linhart, C., Panzacchi, S., Belpoggi, F., Clausing, P., Zaller, J.G., Hertoge, K., 2021. Year-round
Gibbs, J.L., Yost, M.G., Negrete, M., Fenske, R.A., 2017. Passive sampling for indoor and out- pesticide contamination of public sites near intensively managed agricultural areas in
door exposures to chlorpyrifos, azinphos-methyl, and oxygen analogs in a rural agricul- South Tyrol. Environ. Sci. Eur. 33, 1.
tural community. Environ. Health Perspect. 125, 333–341. Liu, P., Wu, C.-h., Chang, X.-l., Qi, X.-j., Zheng, M.-l., Zhou, Z.-j., 2014. Assessment of chlorpyr-
Gouin, T., Wania, F., Ruepert, C., Castillo, L.E., 2008. Field testing passive air samplers for cur- ifos exposure and absorbed daily doses among infants living in an agricultural area of the
rent use pesticides in a tropical environment. Environ. Sci. Technol. 42, 6625–6630. Province of Jiangsu,China. Int. Arch. Occup. Environ. Health 87, 753–762.
Goulson, D., Thompson, J., Croombs, A., 2018. Rapid rise in toxic load for bees revealed by Lu, C., Fenske, R.A., Simcox, N.J., Kalman, D., 2000. Pesticide exposure of children in an ag-
analysis of pesticide use in Great Britain. PeerJ 6, e5255. ricultural community: evidence of household proximity to farmland and take home expo-
Guida, Yd.S., Meire, R.O., Torres, J.P.M., Malm, O., 2018. Air contamination by legacy and sure pathways. Environ. Res. 84, 290–302.
current-use pesticides in Brazilian mountains: an overview of national regulations by Lüdecke, D., 2018. Ggeffects: tidy data frames of marginal effects from regression models.
monitoring pollutant presence in pristine areas. Environ. Pollut. 242, 19–30. J. Open Source Softw. 3, 772. https://doi.org/10.21105/joss.00772.
Gunstone, T., Cornelisse, T., Klein, K., Dubey, A., Donley, N., 2021. Pesticides and soil inver- Lüdecke, D., 2020. sjPlot: Data Visualization for Statistics in Social Science Version R package
tebrates: a hazard assessment. Front. Environ. Sci. 9, 643847. https://doi.org/10.3389/ version 2.8.3.
fenvs.2021.643847. Mesnage, R., 2021. Coformulants in commercial herbicides. In: Mesnage, R., Zaller, J.G.
Gupta, P.K., 2004. Pesticide exposure - Indian scene. Toxicology 198, 83–90. (Eds.), Herbicides: Chemistry, Efficacy, Toxicology, and Environmental Impacts. Elsevier,
Habel, J.C., Samways, M.J., Schmitt, T., 2019a. Mitigating the precipitous decline of terres- Amsterdam, NL, pp. 87–112.
trial European insects: requirements for a new strategy. Biodivers. Conserv. 28, Mesnage, R., Antoniou, M., 2021. Mammalian toxicity of herbicides used in intensive GM
1343–1360. crop farming. In: Mesnage, R., Zaller, J.G. (Eds.), Herbicides: Chemistry, Efficacy, Toxicol-
Habel, J.C., Ulrich, W., Biburger, N., Seibold, S., Schmitt, T., 2019b. Agricultural intensifica- ogy, and Environmental Impacts. Elsevier, Amsterdam, NL, pp. 143–180.
tion drives butterfly decline. Insect Conserv. Divers. 12, 289–295. Mesnage, R., Straw, E.A., Antoniou, M.N., Benbrook, C., Brown, M.J.F., Chauzat, M.-P., et al.,
Hahn, M., Schotthöfer, A., Schmitz, J., Franke, L.A., Brühl, C.A., 2015. The effects of agro- 2021. Improving pesticide-use data for the EU. Nat. Ecol. Evol. 5, 1560. https://doi.org/
chemicals on Lepidoptera, with a focus on moths, and their pollination service in field 10.1038/s41559-021-01574-1.
margin habitats. Agric. Ecosyst. Environ. 207, 153–162. Mrema, E.J., Rubino, F.M., Brambilla, G., Moretto, A., Tsatsakis, A.M., Colosio, C., 2013. Per-
Haiden, T., Kann, A., Wittmann, C., Pistotnik, G., Bica, B., Gruber, C., 2011. The integrated sistent organochlorinated pesticides and mechanisms of their toxicity. Toxicology 307,
nowcasting through comprehensive analysis (INCA) system and its validation over the 74–88.
Eastern Alpine region. Weather Forecast. 26, 166–183. Mueller, T.C., 2015. Methods To Measure Herbicide Volatility. Weed Science 2015,
Hartig, F., 2019. DHARMa: Residual Diagnostics for Hierarchical (Multi-level/mixed) Regres- pp. 116–120 Special Issue.
sion Models R package version 02 4. Munn, S., Goumenou, M., 2013. Thresholds for Endocrine Disrupters and Related Uncer-
Hayward, S.J., Gouin, T., Wania, F., 2010. Comparison of four active and passive sampling tainties: Report of the Endocrine Disrupters Expert Advisory Group. Institute for Health
techniques for pesticides in air. Environ. Sci. Technol. 44, 3410–3416. and Consumer Protection, Publications Office.
Herkert, N.J., Spak Scott, N., Smith, A., Schuster, J.K., Harner, T., Martinez, A., et al., 2018. Nakazawa, M., 2021. fmsb: Functions for Medical Statistics Book with some Demographic
Calibration and evaluation of PUF-PAS sampling rates across the Global Atmospheric Pas- Data. Version R package version 0.7.1. https://CRAN.R-project.org/package=fmsb.
sive Sampling (GAPS) network. Environ. Sci. Processes Impacts 20, 210–219. Paglia, G., Del Greco, F.M., Carli, C., Sigurdsson, B.B., Smarason, S., Wegher, M., et al., 2021.
van Hoesel, W., Tiefenbacher, A., König, N., Dorn, V.M., Hagenguth, J.F., Ua, Prah, et al., Longitudinal assessment of chlorpyrifos exposure in farmers and residents of an Italian al-
2017. Single and combined effects of pesticide seed dressings and herbicides on pine region. ExpoHealth 13, 651–659. https://doi.org/10.1007/s12403-021-00409-5.

11
J.G. Zaller et al. Science of the Total Environment 838 (2022) 156012

PAN, 2016. PAN International List of Highly Hazardous Pesticides. Pesticide Action Network Stahlschmidt, P., Brühl, C.A., 2012. Bats at risk? Bat activity and insecticide residue analysis of
International. http://www.pan-germany.org/download/PAN_HHP_List_161212_F.pdf. food items in an apple orchard. Environ. Toxicol. Chem. 31, 1556–1563.
Pascale, A., Laborde, A., 2020. Impact of pesticide exposure in childhood. Rev. Environ. Stahlschmidt, P., Hahn, M., Brühl, C.A., 2017. Nocturnal risks-high bat activity in the agricul-
Health 35, 221–227. tural landscape indicates potential pesticide exposure. Front. Environ. Sci. 5, 62.
Perez-Fernandez, C., Morales-Navas, M., Guardia-Escote, L., Colomina, M.T., Giménez, E., Strandberg, B., Sørensen, P.B., Bruus, M., Bossi, R., Dupont, Y.L., Link, M., et al., 2021. Effects
Sánchez-Santed, F., 2020. Postnatal exposure to low doses of chlorpyrifos induces long- of glyphosate spray-drift on plant flowering. Environ. Pollut. 280, 116953.
term effects on 5C-SRTT learning and performance, cholinergic and GABAergic systems Tang, F.H.M., Lenzen, M., McBratney, A., Maggi, F., 2021. Risk of pesticide pollution at the
and BDNF expression. Exp. Neurol. 330, 113356. global scale. Nat. Geosci. 14, 206–210.
Pewsey, A., Neuhäuser, M., Ruxton, G., 2013. Circular Statistics in R. Oxford University Press, Tassin de Montaigu, C., Goulson, D., 2020. Identifying agricultural pesticides that may pose a
Oxford, UK. risk for birds. PeerJ 8, e9526.
Poisson, M.C., Garrett, D.R., Sigouin, A., Belisle, M., Garant, D., Haroune, L., et al., 2021. UNEP, 2019. In: UNEP (Ed.), Stockholm Convention on Persistent Organic Pollutants (POPs).
Assessing pesticides exposure effects on the reproductive performance of a declining ae- UNEP, Stockholm Convention pp. 79pp http://www.pops.int/TheConvention/O
rial insectivore. Ecol. Appl. 31, e02415. verview/TextoftheConvention/tabid/2232/Default.aspx.
Rauh, V.A., Perera, F.P., Horton, M.K., Whyatt, R.M., Bansal, R., Hao, X., et al., 2012. Brain Vandenberg, L.N., Colborn, T., Hayes, T.B., Heindel, J.J., Jacobs, D.R., Lee, D.H., 2012. Hor-
anomalies in children exposed prenatally to a common organophosphate pesticide. mones and endocrine-disrupting chemicals: low-dose effects and nonmonotonic dose re-
Proc. Natl. Acad. Sci. 109, 7871–7876. sponses. Endocr. Rev. 33, 378–455.
Rizzi, C., Finizio, A., Maggi, V., Villa, S., 2019. Spatial-temporal analysis and risk characteri- Veludo, A.F., Martins Figueiredo, D., Degrendele, C., Masinyana, L., Curchod, L., Kohoutek, J.,
sation of pesticides in Alpine glacial streams. Environ. Pollut. 248, 659–666. et al., 2022. Seasonal variations in air concentrations of 27 organochlorine pesticides
Sanchez-Bayo, F., Goka, K., 2014. Pesticide residues and bees – a risk assessment. PLOS ONE (OCPs) and 25 current-use pesticides (CUPs) across three agricultural areas of South
9, e94482. Africa. Chemosphere 289, 133162.
Sánchez-Bayo, F., Tennekes, H.A., 2020. Time-cumulative toxicity of neonicotinoids: experi- Voeten, C., 2020. buildmer: Stepwise Elimination and Term Reordering for Mixed-Effects
mental evidence and implications for environmental risk assessments. Int. J. Environ. Regression. Version R package version 1.5. URL https://CRAN.R-project.org/
Res. Public Health 17, 1629. https://doi.org/10.3390/ijerph17051629. package=buildmer.
Sánchez-Bayo, F., Goulson, D., Pennacchio, F., Nazzi, F., Goka, K., Desneux, N., 2016. Are bee Wagner, D.L., 2020. Insect declines in the Anthropocene. In: Douglas, A.E. (Ed.), Annual Re-
diseases linked to pesticides? — a brief review. Environ. Int. 89–90, 7–11. view of Entomology. vol. 65, pp. 457–480.
Sapcanin, A., Cakal, M., Imamovic, B., Salihovic, M., Pehlic, E., Jacimovic, Z., et al., 2016. Ward, M.H., Lubin, J., Giglierano, J., Colt, J.S., Wolter, C., Bekiroglu, N., et al., 2006. Proxim-
Herbicide and pesticide occurrence in the soils of children's playgrounds in Sarajevo, ity to crops and residential exposure to agricultural herbicides in Iowa. Environ. Health
Bosnia and Herzegovina. Environ. Monit. Assess. 188. Perspect. 114, 893–897.
Schampheleire, M., Nuyttens, D., Baetens, K., Cornelis, W., Gabriels, D., Spanoghe, P., 2009. van Wendel de Joode, B., Barraza, D., Ruepert, C., Mora, A.M., Córdoba, L., Öberg, M., et al.,
Effects on pesticide spray drift of the physicochemical properties of the spray liquid. Pre- 2012. Indigenous children living nearby plantations with chlorpyrifos-treated bags have
cision Agric. 10, 409–420. elevated 3,5,6-trichloro-2-pyridinol (TCPy) urinary concentrations. Environ. Res. 117,
Schmitz, J., Schäfer, K., Brühl, C.A., 2013. Agrochemicals in field margins - assessing the im- 17–26.
pacts of herbicides, insecticides, and fertilizer on the common buttercup (Ranunculus Yusà, V., Coscollà, C., Mellouki, W., Pastor, A., de la Guardia, M., 2009. Sampling and analysis
acris). Environ. Toxicol. Chem. 32, 1124–1131. of pesticides in ambient air. J. Chromatogr. A 1216, 2972–2983.
Schulz, R., Bub, S., Petschick, L.L., Stehle, S., Wolfram, J., 2021. Applied pesticide toxicity Zaller, J.G., 2020. Daily Poison. Pesticides - An Underestimated Danger. Springer Nature,
shifts toward plants and invertebrates, even in GM crops. Science 372, 81–84. Cham, Switzerland.
Schuster, J.K., Harner, T., Eng, A., Rauert, C., Su, K., Hornbuckle, K.C., et al., 2021. Tracking Zaller, J.G., Brühl, C.A., 2019. Non-target effects of pesticides on organisms inhabiting
POPs in global air from the first 10 years of the GAPS network (2005 to 2014). Environ. agroecosystems. Front. Environ. Sci. 7. https://doi.org/10.3389/fenvs.2019.00075.
Sci. Technol. 55, 9479–9488. Zaller, J.G., Weber, M., Maderthaner, M., Gruber, E., Takács, E., Mörtl, M., et al., 2021. Effects
Shoeib, M., Harner, T., Lee, S.C., Lane, D., Zhu, J., 2008. Sorbent-impregnated polyurethane of glyphosate-based herbicides and their active ingredients on earthworms, water
foam disk for passive air sampling of volatile fluorinated chemicals. Anal. Chem. 80, infiltration and glyphosate leaching are influenced by soil properties. Environ. Sci. Eur.
675–682. 33, 51.
Simon-Delso, N., Amaral-Rogers, V., Belzunces, L.P., Bonmatin, J.M., Chagnon, M., Downs, C., Zhang, X., Brown, T.N., Ansari, A., Yeun, B., Kitaoka, K., Kondo, A., et al., 2013. Effect of wind
et al., 2015. Systemic insecticides (neonicotinoids and fipronil): trends, uses, mode of ac- on the chemical uptake kinetics of a passive air sampler. Environ. Sci. Technol. 47,
tion and metabolites. Environ. Sci. Pollut. Res. 22, 5–34. 7868–7875.
van der Sluijs, J.P., 2020. Insect decline, an emerging global environmental risk. Curr. Opin. Zivan, O., Bohbot-Raviv, Y., Dubowski, Y., 2017. Primary and secondary pesticide drift pro-
Environ. Sustain. 46, 39–42. files from a peach orchard. Chemosphere 177, 303–310.

12