Atmospheric Transport of Adulticides Used to Control Mosquito Populations across an Urban Metropolitan Area

: Pesticides registered with the U.S. EPA for mosquito control are called adulticides and are released directly into the atmosphere as aerosols to target ﬂying mosquitos. This adulticide application approach is different from traditional (agricultural) pesticide applications, yet the fate and transport of adulticides in large metropolitan areas is largely unknown. The Houston Metropolitan Area encompasses eight counties, many of which require county-level mosquito control programs that utilize adulticides. Malathion and permethrin are the primary adulticides used by Harris County (HC) in Houston, TX, USA. Houston, like many other metropolitan areas, has an urban atmosphere supporting the oxidation of both gas and particle phase pollutants. During the summer mosquito season of 2016, we collected atmospheric total suspended particulate matter (PM) samples at Jones Forest (JF), located in Montgomery County (directly north of HC) to investigate the atmospheric transport and oxidation of adulticides in an urban atmosphere. Despite HC alternating the adulticide treatment schedule, we measured permethrin, malathion, and malaoxon (oxidation product of malathion), throughout the sampling campaign. These consistent measurements, in conjunction with 12 h backward trajectories, support the conclusion that JF is inﬂuenced by other county-level mosquito-control programs and agricultural pesticide use. This cross-county transport may impact adulticide effectiveness by supporting pesticide resistance in mosquito populations due to repeated exposures to pesticides. This study highlights the need for mosquito control collaborations between counties, especially in areas of urban expansion overlapping with agricultural activities.


Introduction
Mosquito-borne diseases (i.e., malaria, West Nile virus, dengue, Zika, yellow fever, and Chikungunya) [1] cause millions of deaths globally every year [2,3].For example, dengue is endemic in 100 countries, and 390 million annual dengue infections are caused by disease-carrying mosquitoes [4].Research also suggests that different aspects of climate change (e.g., rising global temperatures) will increase mosquito-borne disease transmission rates and extend mosquito population geographic ranges [5][6][7][8].Worldwide, the main strategy for vector control is through the use of pesticides [9], and many of the most populated metropolitan areas regularly utilize pesticides to prevent mosquito-borne disease outbreaks (e.g., Houston, TX, USA, Los Angeles, CA, USA, Chicago, IL, USA, and Miami, FL, USA) [9][10][11][12][13], as well as larger geographical areas throughout East Asia (e.g., Seoul, South Korea [14] and Shanghai, China) [15], South America (e.g., Lima, Peru) [16], and Africa (e.g., Lagos, Nigeria) [17].Pesticides registered by the U.S. EPA to control adult mosquito populations are called adulticides [18].Note: registering a pesticide with the U.S. EPA is a stringent and lengthy process, including scientific, legal, and administrative procedures, which makes it difficult for new chemicals to be registered.In Harris County, TX (home to the Houston metropolitan area and the third most populous county in the U.S. at ~4 million inhabitants), the Harris County Public Health Mosquito Control Division (PHMCD) utilizes truck-mounted, ultra-low volume (ULV) sprayers to release adulticides directly into the atmosphere as aerosols (5-30 µm aerodynamic diameter) during dusk and dawn to kill flying mosquitoes on contact.Adulticide applicators often utilize pesticide drift (i.e., atmospheric transport) to target mosquito habitat that is inaccessible, which is different from traditional application techniques that limit atmospheric drift (e.g., agricultural pesticide use).Large-scale adulticide applications that utilize aerosol atmospheric transport should result in different atmospheric concentrations and degradation pathways as compared with traditional pesticide applications.Due to urban expansion, the Houston metropolitan area is the fifth most populous metropolitan area in the United States (~6.7 million inhabitants in 2016), and includes Harris County, Montgomery County, Liberty County, Waller County, Chambers County, Fort Bend County, Galveston County, and Brazoria County (Figure 1, left) [19].These counties have similar mosquito seasons and chemical control strategies (i.e., the use of adulticides) to the Harris County PHMCD [10,[20][21][22][23][24].Additionally, due to urban sprawl, agricultural activities in many of these counties are now adjacent to areas of recent urban growth (Figure 1, right) [25].It is important to note that agricultural pest control efforts, which may utilize similar (or the same) classes of insecticides [26], would manage agricultural pests separately from public health mosquito control efforts [27].
Atmosphere 2023, 14, x FOR PEER REVIEW 2 of 15 Seoul, South Korea [14] and Shanghai, China) [15], South America (e.g., Lima, Peru) [16], and Africa (e.g., Lagos, Nigeria) [17].Pesticides registered by the U.S. EPA to control adult mosquito populations are called adulticides [18].Note: registering a pesticide with the U.S. EPA is a stringent and lengthy process, including scientific, legal, and administrative procedures, which makes it difficult for new chemicals to be registered.In Harris County, TX (home to the Houston metropolitan area and the third most populous county in the U.S. at ~4 million inhabitants), the Harris County Public Health Mosquito Control Division (PHMCD) utilizes truck-mounted, ultra-low volume (ULV) sprayers to release adulticides directly into the atmosphere as aerosols (5-30 µm aerodynamic diameter) during dusk and dawn to kill flying mosquitoes on contact.Adulticide applicators often utilize pesticide drift (i.e., atmospheric transport) to target mosquito habitat that is inaccessible, which is different from traditional application techniques that limit atmospheric drift (e.g., agricultural pesticide use  1, left) [19].These counties have similar mosquito seasons and chemical control strategies (i.e., the use of adulticides) to the Harris County PHMCD [10,20-24].Additionally, due to urban sprawl, agricultural activities in many of these counties are now adjacent to areas of recent urban growth (Figure 1, right) [25].It is important to note that agricultural pest control efforts, which may utilize similar (or the same) classes of insecticides [26], would manage agricultural pests separately from public health mosquito control efforts [27].Globally, however, the development of insecticide resistance (IR) is a growing threat to the efficacy of adulticides for mosquito control strategies [9,[28][29][30][31][32][33][34][35][36].As part of integrated pest management strategies to prevent the development of IR in mosquito populations, the Harris County PHMCD alternates weekly between using two of the most commonly used adulticides, malathion (an organophosphate) and permethrin (a pyrethroid), to control mosquito populations, with-it is important to note-periods of no treatment Globally, however, the development of insecticide resistance (IR) is a growing threat to the efficacy of adulticides for mosquito control strategies [9,[28][29][30][31][32][33][34][35][36].As part of integrated pest management strategies to prevent the development of IR in mosquito populations, the Harris County PHMCD alternates weekly between using two of the most commonly used adulticides, malathion (an organophosphate) and permethrin (a pyrethroid), to control mosquito populations, with-it is important to note-periods of no treatment between adulticide type [10, [37][38][39].However, as mentioned before, this alternating schedule is determined separately from adjacent agricultural pest treatment schedules, as well as from neighboring county-level mosquito control programs (within the Houston Metropolitan area), which presents opportunities for increased development of IR due to pesticide atmospheric transport.For example, Dunbar et al. (2018) identified increased IR in areas of overlap between agricultural and public health pesticide use (i.e., county-level use) [27].The Harris County PHMCD sprays malathion as a pure solution (i.e., 97% malathion active ingredient), while permethrin is sprayed as a mixture [10].The permethrin mixture includes mineral oil and a synergistic compound, piperonyl butoxide (PBO), which increases the toxicity of permethrin.
The environmental fate of these two adulticides in an urban atmosphere is a combination of their individual physical-chemical properties and their atmospheric lifetimes (due to being released as aerosols).Permethrin and malathion are classified as semi-volatile organic compounds (SVOCs) [40,41]; based on their physical-chemical properties, they can be present in both the gas and particulate phase in the atmosphere.The dispersion of adulticides in both the gas and particle phases is impacted by meteorological conditions (i.e., wind speed and direction, temperature, and relative humidity), while atmospheric transformation is driven by processes such as photolysis and oxidation.
To the best of the authors' knowledge, no risk assessment studies have been completed that specifically focus on inhalation exposure scenarios resulting from county-level adulticide applications, despite concerns of adverse public health effects associated with chronic exposures to these chemicals.The few existing studies focus on adverse health effects (i.e., toxicity tests) or risk assessments associated with human exposures to domestic-use pesticides [74] or insecticides used for ornamental plants, horticulture, and agricultural crops [75][76][77].This may be, in part, due to the lack of atmospheric measurements of county-level adulticide use in urban areas (i.e., pesticides released as aerosols directly into the atmosphere).Furthermore, the reference concentration (RfC) for inhalation exposure routes of neither malathion nor permethrin are available in the U.S. EPA Integrated Risk Information System database; thus, accurate risk calculations are difficult to estimate [78].As mentioned above, chamber-based studies have observed the oxidation of malathion and permethrin in an isolated setting; however, few real-world measurements of these adulticides and oxidation products exist.Given the increased predicted toxicities of the oxidation products, more real-world measurements of adulticides need to be completed to estimate the exposure levels and assess public health risks.
A previous study reported the spatial and temporal variability in atmospheric malathion, malaoxon, and permethrin concentrations in particulate matter (PM) throughout the Houston metropolitan area [79,80].Measured atmospheric adulticide concentrations were consistently higher in nighttime PM samples (i.e., ranging from <MDL to 41 ng m −3 ) and lower in daytime PM samples (i.e., ranging from <MDL to 14 ng m −3 ), which strongly reflects the nighttime applications of the adulticides [79].Using diurnal atmospheric concentrations of malathion and its oxidation product, malaoxon, the nighttime atmospheric half-life of malathion was estimated to be <24 h, which is 40-90% shorter than half-lives reported previously [55,79].The increased degradation rate raises concerns of using malathion in heavily NO x polluted atmospheres (such as in Houston) because malaoxon is more toxic to humans [81,82].Importantly, the underlying oxidation chemistry supported by NO 2 , O 3 , and fresh NO emissions is nuanced; the heterogeneity of this chemistry varies across the Houston Metropolitan area and mosquito season, and therefore will influence each application (and subsequent atmospheric processing) differently.
Ultimately, the environmental fate is important for (1) mosquito control within the targeted areas (i.e., adulticide effectiveness), (2) potential adverse human and environmental impacts, and (3) IR through chronic sublethal exposure in neighboring counties' mosquito populations (via cross-county transport).Currently, there are knowledge gaps in our understanding of cross-county adulticide concentrations in areas of recent urban expansion, which raises a question regarding an increase in IR in mosquito populations in the Houston Metropolitan area from overlapping pesticide use.Thus, accurate measurements of outdoor adulticide concentrations are necessary to understand opportunities to develop resistant mosquito population in urban areas, which ultimately drives adulticide effectiveness and subsequently public health.The goals of this study were to: (1) collect PM samples at a county neighboring Harris County, ( 2) analyze PM samples for atmospheric concentrations of malathion, malaoxon, and permethrin, and (3) assess opportunities for cross-county atmospheric transport of adulticides used for mosquito control.

Sampling
PM samples were collected using total suspended particulate (TSP; n = 16) samplers at a ground-based sampling site from 15 August to 23 August 2016.The Jones Forest (JF) sample site was located in a suburban area north of Harris County (i.e., Montgomery County; ~56 km N of downtown Houston at 30.2367, −95.4827; Figure 1) to represent a potential downwind (based on historical predominant wind direction) receptor from the application sites within the county.Samples were collected as "day" samples (6:00 a.m.-7:30 p.m. local Central Daylight Time) and "night" samples (8:00 p.m.-5:30 a.m.), which corresponds to adulticide application periods.It is important to note that Montgomery County does not use permethrin or malathion, but instead uses deltamethrin applied using a ULV truck-mounted sprayer (personal communications with Montgomery County).This allows JF to serve as a receptor site for other counties that utilize permethrin and malathion as adulticides.Sample media were quartz fiber filters (QFF), which were baked at 550 • C for 12 h in individual foil packets prior to sampling.QFF were stored in freezers (−10 • C) pre-and post-sampling.TSP samples were collected on 20 cm × 25 cm QFF (Pall Corporation, Port Washington, NY, USA).Sampler flow rates were calibrated prior to field deployment.TSP samples were collected using a high-volume sampler (Tisch Environmental, Cleves, OH, USA) with a flow rate of 940 L min −1 .Field blank samples (n = 4) were taken throughout the sampling campaign by placing an unsampled QFF in a filter holder and placing it in the sampler momentarily before removal.Field blanks were treated in a manner identical to those used for sampled filters.

NOAA HYSPLIT Trajectory Model
The NOAA Air Resources Laboratory HYSPLIT Trajectory Model was used to calculate backward trajectories from JF during the sampling campaign [83].These backward trajectories were used to confirm visually wind direction and adulticide source locations during sampling periods.It is important to note that the model was not used to validate atmospheric concentrations measured in this study.Table S1 and S2 in the Supplementary Materials provides model inputs for day and night samples.The backward trajectories had a total run time of either 12 or 24 h, with a new trajectory starting every hour of the sampling period (i.e., Day sample backward trajectories included 14 total trajectories due to the 6:00 a.m.-7:30 p.m. sampling period, and Night sample backward trajectories included 10 total trajectories due to the 8:00 p.m.-5:30 a.m.sampling period).
All backward trajectories terminated at the sampling site at a 10-m height.Day sample backward trajectories can be found in Supplementary Materials Figures S8-S10.

Chemicals and Consumables
Analytical standards, materials, and solvents were purchased from commercially available vendors: unlabeled permethrin (cis-and trans-) from ChemService (West Chester, PA, USA), unlabeled malathion and malaoxon from Accustandard (New Haven, CT, USA), and labeled d 10 -malathion and 13 C 6 -trans-permethrin from Cambridge Isotopes (Cambridge, MA, USA).Vendors of other materials and solvents have been previously described [79].

Chemical Analysis
Permethrin, malathion, and malaoxon were extracted from TSP atmospheric QFF samples using pressurized liquid extraction (2:1 v/v dichloromethane:acetone at 100 • C with three 5 min static cycles, 80% flush volume, and 100 s N 2 purge) [84].An aliquot of each sample was placed in an extraction cell and spiked with isotopically-labeled surrogate standards ( 13 C 6 -trans-permethrin and d 10 -malathion) prior to extraction to correct for dayto-day variability in sample preparation.Extracts were concentrated to ~125 µL under a gentle stream of N 2 at 35 • C using a Zymark TurboVap II (Biotage, Charlotte, NC, USA).
Extracts were spiked with isotopically-labeled internal standards (d 12 -Benzo (e) pyrene) prior to analysis via gas chromatography-mass spectrometry (GCMS) with electron ionization (EI) and selective ion monitoring.GCMS methods, and analyte identification and quantitation methods, have been previously described [79].A continuous calibration verification standard (near the midpoint of the calibration solutions) was analyzed before and after each GCMS sample batch.Solvent blanks were run before and after each check standard.Labeled d 10 -malathion was monitored via m/z 125, 93, and 183; unlabeled malathion was monitored via m/z 125, 173, and 93; and malaoxon was monitored via m/z 127, 99, 55, and 109.Ions (m/z) are presented as quantification ions first, followed by the qualification ion(s).Labeled 13 C 6 -trans-permethrin was monitored via m/z 189; unlabeled cis-permethrin was monitored via m/z 183, 163, and 165; and unlabeled trans-permethrin was monitored via m/z 183.Calibration solutions ranged from 800 ppb to 17,000 ppb for cis-permethrin and trans-permethrin, 2700 to 23,000 ppb for malathion, and 4000 to 34,000 ppb for malaoxon.All permethrin concentrations are reported as the addition of cisand trans-isomers.

Quality Assurance/Quality Control
A known amount of Standard Reference Material (SRM) 1649b (Urban Dust) was loaded onto a blank unsampled QFF, spiked with isotopically-labeled surrogate standards, and extracted with each batch of samples for quality assurance and quality control (QA/QC) and method validation.Blank unsampled QFF were analyzed to assess laboratory contamination/cross-contamination associated with sample preparation and were exposed to the entire analytical method.Method detection limits (MDLs) for cis-and trans-permethrin, malathion, and malaoxon were 14, 16, 19, and 4.6 ppb, respectively [79].Average sample surrogate recoveries of 13 C 6 -trans-permethrin and d 10 -malathion for this method were 92 ± 19% and 91 ± 11%, respectively (see Table S3 in Supplementary Materials).Average surrogate recoveries spiked in the SRM (at ppb concentrations) were 102 ± 19% (n = 5).Field blanks were extracted and analyzed under the same conditions as samples, with negligible concentrations; thus, no adjustments were made to measured concentrations.

Results and Discussion
To inform the public of the treatment, the Harris County PHMCD provides proposed treatment maps-an example of which is given in Figure 2-publicly, daily, online; these show the ZIP codes within the county that will be sprayed with adulticides on a specific day.It is important to note that the other counties surrounding Harris County do not provide/archive proposed treatment maps for their mosquito control activities; thus, those  3 for sampling date and adulticide application date correspondence.Based on ULV practices, it is assumed the adulticide concentrations will be highest at the point of release, but will decrease in concentration with time via dispersion, degradation (e.g., oxidation and photolysis), partitioning to the gas phase, and deposition [40].Malathion and permethrin were present in 94% and 100% of the TSP collected at the JF, respectively (Figure 3).Malaoxon, the oxidation product of malathion, was also measured in 19% of the TSP samples collected from JF (Figure 3A).In this study, the highest measured atmospheric concentration of malathion was 1. Malathion and permethrin were present in 94% and 100% of the TSP collected JF, respectively (Figure 3).Malaoxon, the oxidation product of malathion, was also ured in 19% of the TSP samples collected from JF (Figure 3A).In this study, the h measured atmospheric concentration of malathion was 1.8 ng m −3 (night sample, gust), of malaoxon was 2.2 ng m −3 (day sample, 23 August), and of permethrin was m −3 (day sample, 18 August).The peak measurement of malathion in the 22 Augus sample aligns with the Harris County PHMCD proposed adulticide treatment sc (i.e., malathion was applied 21-23 August), while the peak measurement of permet the 18 August day sample may suggest agricultural adulticide activities outside Harris County PHMCD (discussed further in Section 3.2) [10].
In comparison with the measurements from the 2013 study, the present stud cates lower and more consistent adulticide concentrations in both nighttime and d samples.This study provides additional nuance to the 2013 Houston adulticide which showed higher nighttime adulticide concentrations (ranging from below de limit to 41 ng m −3 ) and lower daytime concentrations (ranging from below detectio to 14 ng m −3 ) [79], by considering other adulticide activities across the Houston me itan area (e.g., neighboring county-based mosquito control efforts and agricultura cide applications) in addition to the assessment of adulticide-use specifically asso with the Harris County PHMCD.

Malathion and Malaoxon
During this study, TSP atmospheric concentrations of malathion and ma ranged from below detection limit to ~1.8 ng m −3 and below detection limit to ~2.2 (15 August night through 23 day), respectively (Figure 3A).The peak malaoxon pheric concentration occurred during a daytime sample (23 August), and was high the peak malathion atmospheric concentration (22 August).First, this indicates d oxidation of malathion to malaoxon, which has been reported previously [79].Sec higher concentrations of malaoxon compared to malathion highlights the importa measuring oxidation products of adulticides; public exposure levels of adulticides appear consistently low (and potentially acceptable) if oxidation products are not ured.Lastly, while the peak measurements of malathion and malaoxon occurred the Harris County malathion treatment period (21-23 August), notably, In comparison with the measurements from the 2013 study, the present study indicates lower and more consistent adulticide concentrations in both nighttime and daytime samples.This study provides additional nuance to the 2013 Houston adulticide study, which showed higher nighttime adulticide concentrations (ranging from below detection limit to 41 ng m −3 ) and lower daytime concentrations (ranging from below detection limit to 14 ng m −3 ) [79], by considering other adulticide activities across the Houston metropolitan area (e.g., neighboring county-based mosquito control efforts and agricultural pesticide applications) in addition to the assessment of adulticide-use specifically associated with the Harris County PHMCD.

Malathion and Malaoxon
During this study, TSP atmospheric concentrations of malathion and malaoxon ranged from below detection limit to ~1.8 ng m −3 and below detection limit to ~2.2 ng m −3 (15 August night through 23 day), respectively (Figure 3A).The peak malaoxon atmospheric concentration occurred during a daytime sample (23 August), and was higher than the peak malathion atmospheric concentration (22 August).First, this indicates daytime oxidation of malathion to malaoxon, which has been reported previously [79].Secondly, higher concentrations of malaoxon compared to malathion highlights the importance of measuring oxidation products of adulticides; public exposure levels of adulticides might appear consistently low (and potentially acceptable) if oxidation products are not measured.Lastly, while the peak measurements of malathion and malaoxon occurred during the Harris County malathion treatment period (21-23 August), notably, these measurements are 1-2 orders of magnitude lower than measurements associated with Harris County adulticide activities in the 2013 study [79].
Malathion and malaoxon were measured throughout the sampling campaign, despite the Harris County alternating treatment schedule (i.e., permethrin treatments on 16-18 August, no mosquito control activities during 19-20 August, and malathion treatments on 21-23 August, Figure 3A).As mentioned earlier, Montgomery County does not use permethrin or malathion as part of their county-level mosquito control efforts [21], and previous studies [40] show measurements of residentially used pesticides to be in the 10-100 s pg m −3 range, which is below MDL for this study.This suggests that malathion and malaoxon measured throughout the campaign may have been a result of cross-county transport of adulticide use from a neighboring county.For example, samples collected on 18 August measured malathion and malaoxon, even though this was a permethrin treatment date in Harris County.Interestingly, the daytime malaoxon concentration on the 18th was higher than both the nighttime and daytime concentrations of malathion.
NOAA HYSPLIT Trajectory Model backward trajectories from JF on 18 August show air masses arriving from the east which would pass over Liberty County (Figures 1 and  4C) and not Harris County.Liberty County has mosquito control operations in individual cities and agricultural activities within the county that likely contributed to the malathion measured in this sample [26,85].The daytime measurement of malaoxon could be a result of oxidation of daytime agricultural use of malathion in Liberty County, or transport (and oxidation) from adulticides used for mosquito control in Liberty County the night before (See Supplementary Materials Figure S8C for daytime backward trajectories).While it is difficult to identify the exact source of the malathion and malaoxon measured, it does suggest evidence of large-scale adulticide use (i.e., mosquito control or agricultural activities) influencing atmospheric pesticide concentrations in a neighboring county.
Malathion and malaoxon were measured throughout the sampling campaign, despite the Harris County alternating treatment schedule (i.e., permethrin treatments on August 16-18, no mosquito control activities during 19-20 August, and malathion treatments on 21-23 August, Figure 3A).As mentioned earlier, Montgomery County does not use permethrin or malathion as part of their county-level mosquito control efforts [21], and previous studies [40] show measurements of residentially used pesticides to be in the 10-100 s pg m −3 range, which is below MDL for this study.This suggests that malathion and malaoxon measured throughout the campaign may have been a result of cross-county transport of adulticide use from a neighboring county.For example, samples collected on 18 August measured malathion and malaoxon, even though this was a permethrin treatment date in Harris County.Interestingly, the daytime malaoxon concentration on the 18th was higher than both the nighttime and daytime concentrations of malathion.
NOAA HYSPLIT Trajectory Model backward trajectories from JF on 18 August show air masses arriving from the east which would pass over Liberty County (Figures 1 and  4C) and not Harris County.Liberty County has mosquito control operations in individual cities and agricultural activities within the county that likely contributed to the malathion measured in this sample [26,85].The daytime measurement of malaoxon could be a result of oxidation of daytime agricultural use of malathion in Liberty County, or transport (and oxidation) from adulticides used for mosquito control in Liberty County the night before (See Supplementary Materials Figure S10C for daytime backward trajectories).While it is difficult to identify the exact source of the malathion and malaoxon measured, it does suggest evidence of large-scale adulticide use (i.e., mosquito control or agricultural activities) influencing atmospheric pesticide concentrations in a neighboring county.Malathion was also measured in TSP samples on 19 and 20 August (both day and night samples), despite those dates being the "no-treatment" period in Harris County (Figure 3A).Therefore, the malathion measured in these samples is likely not from the Harris County PHMCD.Backward trajectories (12 h) from JF on 20 August suggest that air masses arrived from the south from Brazoria County, which uses malathion as part of their county-level mosquito control operations and is likely the source of malathion measured in this sample (Figure 5C) [23].Interestingly, however, malaoxon was not measured in these samples, which would be expected after traveling over Houston (has high oxidation capacity).
(16-18 August) [83] Malathion was also measured in TSP samples on 19 and 20 August (both day and night samples), despite those dates being the "no-treatment" period in Harris County (Figure 3A).Therefore, the malathion measured in these samples is likely not from the Harris County PHMCD.Backward trajectories (12 h) from JF on 20 August suggest that air masses arrived from the south from Brazoria County, which uses malathion as part of their county-level mosquito control operations and is likely the source of malathion measured in this sample (Figure 5C) [23].Interestingly, however, malaoxon was not measured in these samples, which would be expected after traveling over Houston (has high oxidation capacity).

Permethrin
During this study, TSP concentrations of permethrin ranged from below detection limit to ~7.6 ng m −3 (15 August night through 23 day) (Figure 3B).Similarly, to malathion, permethrin was measured throughout the sampling campaign at JF, despite the Harris County alternating treatment schedule (i.e., permethrin treatments on 16-18 August, no mosquito control activities during 19-20 August, and malathion treatments on 21-23 August, Figure 3B).Because Montgomery County does not use permethrin or malathion as part of their county-Permethrin was also measured during the Harris County malathion treatment period (21-23 August) (Figure 3B).For example, the night sample collected on 22 August represents the highest nighttime measurement of permethrin even though it is during a Harris County malathion treatment period (Figure 3B).Backward trajectories (12 h) from JF on 22 August show air masses arriving from Galveston County, in addition to Chambers County (Figures 1 and 6B).Harris County was utilizing malathion on 22 August, which suggests Galveston County and Chambers County as potential sources of permethrin in this sample.
m ) [79].Based on this, it is possible the permethrin did come from Harris County, even though the proposed treatment schedule suggests otherwise.Conversely, it is possible that the malathion and permethrin both came from Brazoria County but from separate activities (i.e., malathion from mosquito control and permethrin from agricultural pest control, or vice versa).Notably, however, atmospheric concentrations of pesticides following dispersion from agricultural sites have been reported in the 10-100 s pg m −3 range [56,87,88].
Permethrin was also measured during the Harris County malathion treatment period (21-23 August) (Figure 3B).For example, the night sample collected on 22 August represents the highest nighttime measurement of permethrin even though it is during a Harris County malathion treatment period (Figure 3B).Backward trajectories (12 h) from JF on 22 August show air masses arriving from Galveston County, in addition to Chambers County (Figures 1 and 6B).Harris County was utilizing malathion on August 22, which suggests Galveston County and Chambers County as potential sources of permethrin in this sample.

Conclusions
Globally, millions of deaths every year are attributed to mosquito-borne diseases, which continues to be exacerbated by climate change as disease transmission rates increase and mosquito population geographic ranges extend.Concerns are further heightened as the increasing development of IR threatens the efficacy of using pesticides for mosquito control.Despite this worldwide issue, this study highlights that multiple county level mosquito control groups are acting independently during the mosquito control season, rather than coordinating efforts.Houston is a large metropolitan area that is representative of other large urban areas dealing with issues pertaining to air quality and public health mosquito control (e.g., Los Angeles, CA, USA, Chicago, IL, USA, and Miami, FL, USA).During this relatively short campaign, Jones Forest, in Montgomery County, served as an adulticide receptor site for multiple counties across the Houston Metropolitan area.Downwind site measurements showed consistent 1-5 ng m −3 concentrations of adulticides, despite Harris County's "on/off" usage strategy.Twelve-hour backward trajectories support the conclusion that Jones Forest serves as a receptor site for adulticide use by

Conclusions
Globally, millions of deaths every year are attributed to mosquito-borne diseases, which continues to be exacerbated by climate change as disease transmission rates increase and mosquito population geographic ranges extend.Concerns are further heightened as the increasing development of IR threatens the efficacy of using pesticides for mosquito control.Despite this worldwide issue, this study highlights that multiple county level mosquito control groups are acting independently during the mosquito control season, rather than coordinating efforts.Houston is a large metropolitan area that is representative of other large urban areas dealing with issues pertaining to air quality and public health mosquito control (e.g., Los Angeles, CA, USA, Chicago, IL, USA, and Miami, FL, USA).During this relatively short campaign, Jones Forest, in Montgomery County, served as an adulticide receptor site for multiple counties across the Houston Metropolitan area.Downwind site measurements showed consistent 1-5 ng m −3 concentrations of adulticides, despite Harris County's "on/off" usage strategy.Twelve-hour backward trajectories support the conclusion that Jones Forest serves as a receptor site for adulticide use by multiple counties in the Houston Metropolitan Area, and that these counties' adulticide treatment schedules are determined on an individual county basis.As a result of such independent activities, the relatively consistent permethrin and malathion concentrations in Montgomery County may lead to an increase in the development of IR in mosquito populations due to the repeated exposure to adulticides.This is especially relevant for Harris County, which is geographically central in the Houston Metropolitan area, and is therefore downwind of multiple neighboring counties with similar mosquito control operations.We assume that this phenomenon is happening throughout the season (i.e., 5-8 months) and across the multiple counties that make up the Houston metropolitan area.This cross-county transport of adulticides also adds a level of complexity to risk assessments.The crosscounty transport creates a wide range of adulticides and oxidation product concentrations across the metropolitan area, which may result in differing effectiveness in mosquito control (i.e., support development of IR) as well as different public health exposure scenarios to consider within a larger population (i.e., ~6.7 million people).To prevent counties from working against each other unintentionally, mosquito control programs should operate on a larger scale than the county level.The findings of this study highlight the need for communication and collaboration between county-level mosquito control programs that utilize the same pesticides, especially in large, populated areas of urban expansion that overlap with agricultural activities (e.g., midwestern and southeastern states in the U.S.).

Figure 1 .
Figure 1.(left): Texas counties surrounding Harris County with county−level mosquito−control programs utilizing adulticides.The black star indicates the Houston city center.The red star indicates the 2016 sampling site of the present study.(right): Greater Houston percentage population change between 2010 and 2018.The black star indicates the Houston city center.Figure modified from original [25].Greater Houston absolute numerical population change can be found in supplementary information (SI) (Figure S1).

Figure 1 .
Figure 1.(left): Texas counties surrounding Harris County with county−level mosquito−control programs utilizing adulticides.The black star indicates the Houston city center.The red star indicates the 2016 sampling site of the present study.(right): Greater Houston percentage population change between 2010 and 2018.The black star indicates the Houston city center.Figure modified from original[25].Greater Houston absolute numerical population change can be found in Supplementary Information (SI) (FigureS1).
are not available for comparison.Harris County covers 1778 square miles and has 230 ZIP codes.Depending on mosquito population counts, the Harris County PHMCD sprays adulticides 5-8 months of the year [10].In August 2016, the adulticides were applied on 23 of the 31 days, and the number of ZIP codes treated per application day ranged from 0 to 25 [10].On application days, Harris County treatment occurs "through the night" beginning at 8 PM and ending at 4:30 AM (personal communications with the Harris County PHMCD), which is the common use period for adulticides due to increased mosquito activity at dusk and dawn, and aims to reduce public exposure to applications [10].Malathion is applied at an hourly rate of 6.53 kg h −1 , and permethrin is applied at an hourly rate of 0.96 kg h −1 (personal communications with the Harris County PHMCD) [10].During campaign dates, permethrin was applied 16-18 August, and malathion was applied 21-23 August (personal communications with the Harris County PHMCD) [10].See Figure

Atmosphere 2023 ,
14, x FOR PEER REVIEW 6 of 15show the ZIP codes within the county that will be sprayed with adulticides on a specific day.It is important to note that the other counties surrounding Harris County do not provide/archive proposed treatment maps for their mosquito control activities; thus, those are not available for comparison.Harris County covers 1778 square miles and has 230 ZIP codes.Depending on mosquito population counts, the Harris County PHMCD sprays adulticides 5-8 months of the year [10].In August 2016, the adulticides were applied on 23 of the 31 days, and the number of ZIP codes treated per application day ranged from 0 to25 [10].On application days, Harris County treatment occurs "through the night" beginning at 8 PM and ending at 4:30 AM (personal communications with the Harris County PHMCD), which is the common use period for adulticides due to increased mosquito activity at dusk and dawn, and aims to reduce public exposure to applications[10].Malathion is applied at an hourly rate of 6.53 kg h −1 , and permethrin is applied at an hourly rate of 0.96 kg h −1 (personal communications with the Harris County PHMCD) [10].During campaign dates, permethrin was applied 16-18 August, and malathion was applied 21-23 August (personal communications with the Harris County PHMCD)[10].See Figure3for sampling date and adulticide application date correspondence.Based on ULV practices, it is assumed the adulticide concentrations will be highest at the point of release, but will decrease in concentration with time via dispersion, degradation (e.g., oxidation and photolysis), partitioning to the gas phase, and deposition[40].

Figure 2 .
Figure 2. Example of Harris County PHMCD adulticide proposed treatment map on 16 August 2016 (http://www.hcphes.org/(accessed on 16 August 2016) [10].The yellow highlighted areas indicate the zip codes proposed to receive adulticide treatment on that day; this example has 25 proposed zip codes.The remaining proposed treatment maps throughout the sampling campaign are available in the Supplementary Materials (Figures S2-S8).

Figure 2 .
Figure 2. Example of Harris County PHMCD adulticide proposed treatment map on 16 August 2016 (http://www.hcphes.org/(accessed on 16 August 2016) [10].The yellow highlighted areas indicate the zip codes proposed to receive adulticide treatment on that day; this example has 25 proposed zip codes.The remaining proposed treatment maps throughout the sampling campaign are available in the Supplementary Materials (Figures S2-S6).

Figure 3 .
Figure 3. Atmospheric concentrations of (A) malathion and malaoxon, and (B) permethrin ured in TSP samples (n = 16) collected at JF. Daytime samples were collected from 6 p.m. to 7 Nighttime samples were collected from 8 p.m. to 5:30 a.m.Note: a green * (malathion) and o (permethrin) indicates a day sample was not collected on that date (i.e., 15 August), while (malathion) and yellow * (permethrin) indicates a night sample was not collected on that d 23 August).

Figure 3 .
Figure 3. Atmospheric concentrations of (A) malathion and malaoxon, and (B) permethrin, measured in TSP samples (n = 16) collected at JF. Daytime samples were collected from 6 p.m. to 7:30 p.m. Nighttime samples were collected from 8 p.m. to 5:30 a.m.Note: a green * (malathion) and orange * (permethrin) indicates a day sample was not collected on that date (i.e., 15 August), while a blue * (malathion) and yellow * (permethrin) indicates a night sample was not collected on that date (i.e., 23 August).

Figure 4 .
Figure 4.The NOAA HYSPLIT Trajectory model output used to visually confirm wind direction and approximate adulticide source locations during Harris County permethrin treatment period (16-18 August) [83].(A) 16 August 2016 (night sample period) backward trajectory (12 h) from Jones Forest showing air sources coming from Liberty County and Hardin County.(B) 17 August 2016 (night sample period) backward trajectory (12 h) from Jones Forest showing air sources coming from Harris County and Liberty County.(C) 18 August 2016 (night sample period), backward trajectory (12 h) from Jones Forest showing air sources coming from Liberty County.Corresponding day sample backward trajectories are in Supplementary Materials Figure S8.

Figure 5 .
Figure 5.The NOAA HYSPLIT Trajectory model output used to visually confirm wind direction and approximate adulticide source locations during Harris County no−treatment periods (15, 19 and 20 August) [83].(A) 15 August 2016 (night sample period) backward trajectory (12 h) from Jones Forest showing air sources coming from Liberty County, Hardin County, and San Jacinto County.(B) 19 August 2016 (night sample period) backward trajectory (12 h) from Jones Forest showing air sources coming from Harris County and Brazoria County.(C) 20 August 2016 (night sample period), backward trajectory (12 h) from Jones Forest showing air sources coming from Harris County and Brazoria County.Corresponding day sample backward trajectories are in Supplementary Materials Figure S11.

Figure 5 .
Figure 5.The NOAA HYSPLIT Trajectory model output used to visually confirm wind direction and approximate adulticide source locations during Harris County no−treatment periods (15, 19 and 20 August) [83].(A) 15 August 2016 (night sample period) backward trajectory (12 h) from Jones Forest showing air sources coming from Liberty County, Hardin County, and San Jacinto County.(B) 19 August 2016 (night sample period) backward trajectory (12 h) from Jones Forest showing air sources coming from Harris County and Brazoria County.(C) 20 August 2016 (night sample period), backward trajectory (12 h) from Jones Forest showing air sources coming from Harris County and Brazoria County.Corresponding day sample backward trajectories are in Supplementary Materials Figure S9.

Figure 6 .
Figure 6.The NOAA HYSPLIT Trajectory model output used to visually confirm wind direction and approximate adulticide source locations during Harris County malathion treatment period (21-23 August) [83].(A) 21 August 2016 (night sample period) backward trajectory (12 h) from Jones Forest showing air sources coming from Liberty County and Hardin County.(B) 22 August 2016 (night sample period) backward trajectory (12 h) from Jones Forest showing air sources coming from Galveston County and Chambers County.Corresponding day sample backward trajectories are in Supplementary Materials Figure S12.

Figure 6 .
Figure 6.The NOAA HYSPLIT Trajectory model output used to visually confirm wind direction and approximate adulticide source locations during Harris County malathion treatment period (21-23 August) [83].(A) 21 August 2016 (night sample period) backward trajectory (12 h) from Jones Forest showing air sources coming from Liberty County and Hardin County.(B) 22 August 2016 (night sample period) backward trajectory (12 h) from Jones Forest showing air sources coming from Galveston County and Chambers County.Corresponding day sample backward trajectories are in Supplementary Materials Figure S10.
. (A) 16 August 2016 (night sample period) backward trajectory (12 h) from Jones Forest showing air sources coming from Liberty County and Hardin County.(B) 17 August 2016 (night sample period) backward trajectory (12 h) from Jones Forest showing air sources coming from Harris County and Liberty County.(C) 18 August 2016 (night sample period), backward trajectory (12 h) from Jones Forest showing air sources coming from Liberty County.Corresponding day sample backward trajectories are in Supplementary Materials Figure S10.