Spatio-temporal impacts of aerial adulticide applications on populations of West Nile virus vector mosquitoes

Spatio-temporal impacts of aerial adulticide applications on populations of West Nile virus vector mosquitoes | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Spatio-temporal impacts of aerial adulticide applications on populations of West Nile virus vector mosquitoes Karen M. Holcomb, Robert C. Reiner, Christopher M Barker This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-92266/v2 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 24 Feb, 2021 Read the published version in Parasites & Vectors → Version 2 posted 9 You are reading this latest preprint version Show more versions Abstract Background: Aerial applications of insecticides that target adult mosquitoes are widely used to reduce transmission of West Nile virus to humans during periods of epidemic risk. However, estimates of the reduction in abundance following these treatments typically focus on single events, rely on pre-defined, untreated control sites, and can vary widely due to stochastic variation in population dynamics and trapping success unrelated to the treatment. Methods: To overcome these limitations, we developed generalized additive models fitted to mosquito surveillance data from CO 2 -baited traps in Sacramento and Yolo counties, California from 2006-2017. The models accounted for the expected spatial and temporal trends in the abundance of adult female Culex tarsalis and Culex pipiens in the absence of aerial spraying. Estimates for the magnitude of deviation from baseline abundance following aerial spray events were obtained from the models. Results: One-week post-treatment with full spatial coverage of the trapping area by pyrethroid or pyrethrin products, Cx. pipiens abundance was reduced by a mean of 52.4% (95% CI: -65.6, -36.5%) while the use of at least one organophosphate pesticide resulted in a 76.2% (95% CI: -82.8, -67.9%) reduction. For Cx. tarsalis one-week post-treatment with full coverage resulted in a 30.7% (95% CI: -54.5, 2.5%) reduction; pesticide class was not a significant factor contributing to reduction. In comparison, repetition of spraying over three to four consecutive weeks resulted in similar estimates for Cx. pipiens and a somewhat smaller magnitude for Cx. tarsalis. Conclusions: Aerial adulticides are effective for rapid short-term reduction of the abundance of the primary West Nile virus vectors, Cx. tarsalis and Cx. pipiens . A larger magnitude of reduction is estimated in Cx. pipiens , possibly due to the species’ focal distribution. Effects of aerial sprays on Cx. tarsalis populations are likely modulated by the species’ large dispersal ability, population sizes, and vast productive larval habitat present in the study area. Our modeling approach provides a new way to estimate effects of public-health pesticides on vector populations using routinely collected observational data and accounting for spatio-temporal trends and contextual factors like weather and habitat. It does not require pre-selected control sites and expands upon past studies that have focused on effects of individual aerial treatment events. Parasitology Adulticide Aerial spraying Culex tarsalis Culex pipiens GAM Generalized additive models Spatial-temporal model Mosquitoes Mosquito-borne disease West Nile virus Figures Figure 1 Figure 2 Figure 3 Figure 4 Background West Nile virus (WNV; genus Flavivirus , family Flaviviridae ) causes a potentially fatal, neuroinvasive mosquito-borne disease [1]. It is maintained in an enzootic cycle between birds and mosquitoes [2,3], predominantly in the genus Culex [4], and can spillover to infect horses and humans, both of which are dead-end hosts vulnerable to disease [5]. Culex (Cx.) tarsalis and Cx. pipiens complex mosquitoes are the primary enzootic and epizootic vectors in California [6,7]. While 80% of human infections are asymptomatic, clinical manifestations can include acute febrile illness, encephalitis, flaccid paralysis, and death [8]. Often the severe form results in long-term physical and mental disabilities [9]. WNV invaded California in 2003 and has become endemic [7]. An average of 238 neuroinvasive cases occur statewide annually with approximately one-third occurring in the Central Valley, where the landscape is dominated by large-scale agriculture punctuated by cities and small towns [10]. As no human vaccine exists, prevention of human diseases relies primarily on personal protective measures (i.e. wearing long sleeves, using insect repellent, and avoiding the dawn/dust periods when mosquitoes bite) and vector control by local vector control districts or health departments [11,12]. In periods of epidemic risk when large numbers of WNV-infected Culex mosquitoes are detected near human population centers, large-scale aerial applications of insecticides are utilize to rapidly reduce the abundance of adult mosquitoes and disrupt virus transmission cycles, thereby reducing zoonotic transmission risk [13]. Three main classes of pesticide products have been licensed for use in aerial spray applications in California; pyrethrins, pyrethroids, and organophosphates [12,14,15]. Pyrethrins are naturally derived insecticides from chrysanthemum flowers ( Chrysanthemum cineriaefolium ) that inactivate sodium channels in the insect nervous system, resulting in paralysis and death [15]. Pyrethroids are synthetically derived pyrethrins with a similar mode of action and longer half-life. Organophosphates inhibit acetylcholinesterase, affecting neurotransmission and causing uncontrolled nerve activation and death in insects [14]. A standard method for evaluating the efficacy of an aerial spray event compares pre- to post-treatment mosquito trap counts inside the treatment zone versus changes for the same period in an adjacent unsprayed control area [16]. This method, first proposed by Mulla et al. [17], has been adapted to and widely used in evaluating the efficacy of aerial spraying for reducing the abundance of female mosquitoes and has been extended to assess changes in other indicators of risk, namely infection prevalence in mosquitoes, human cases, and reported dead birds with WNV infection [16]. However, reported estimates vary widely, with some studies even indicating occasional increases in trap counts following spray events [18–21]. Despite its wide use, the assumptions behind Mulla’s formula are often violated, resulting in confounded estimates. First, treatment and control sites are often not independent due to the spatial connectivity of populations with mosquito dispersal and immigration [22]. With the connectivity and potential drift of pesticides via wind, there is the potential that insecticide sprays have wider population impacts than just the targeted spray zone [20]. Second, the difference in pre- to post- trap count ratios in and between areas are not solely due to control measures, but rather are impacted by weather, seasonality in mosquito populations, differential presence of larval breeding sources or simply stochastic variation in trapping success [18,20,21]. Overall, Mulla’s formula neglects the spatio-temporal structure of mosquito populations and external factors impacting the random volatility of mosquito trapping success. To overcome the limitations of assessing the efficacy of aerial sprays on the individual spray event basis, we paired long-term surveillance and vector control records (12 years) from Sacramento-Yolo Mosquito and Vector Control District (SYMVCD) in California to capture baseline spatio-temporal mosquito population dynamics and estimate the magnitude and duration of the impacts of aerial sprays on the abundance of Cx. pipiens and Cx. tarsalis , the predominant WNV vectors in California. We chose a generalized additive modeling (GAM) framework to capture the nonlinear population dynamics and associations inherent to mosquito collections. Methods Study Area The study area encompasses Sacramento and Yolo counties, California (Fig 1) which have a combined area of approximately 5,126 km 2 and a population of approximately 1.73 million people in 2016 [23]. Sacramento County is 34.12% urban and 65.88% rural with the majority of urban areas consisting of the concentrated Sacramento urban center and surrounding suburbs. In comparison, Yolo County is 4.61% urban and 95.39% rural with smaller, more dispersed urban areas [24]. These counties, located in the northern part of California’s Central Valley, are characterized by a Mediterranean climate with hot, dry summers (Jul mean temperature: 25.8 °C, May-Sep mean total rainfall: 3.18 cm) and mild, rainy winters (Jan mean temperature: 9.6 °C, Oct-Apr mean total rainfall: 47.30 cm) [25] and extensive irrigated agriculture, especially rice and row crops such as tomatoes. Sacramento-Yolo Mosquito and Vector Control District (SYMVCD), established in 1946 to protect the public from nuisance mosquito biting and mosquito-borne diseases, manages mosquito populations in Sacramento and Yolo counties [26]. Aerial Treatments and Mosquito Collections SYMVCD provided the spatial polygons (Fig 1b) and associated data detailing the date, area targeted for spraying, number of consecutive nights of spraying in the same location, and pesticide product used for all aerial sprays during the study period (1,021 unique nights of spraying during 930 spray events). Mosquito collection records for CDC CO 2 -baited EVS traps [27] from SYMVCD for the years 2006-2017 (Fig 1c) were obtained with permission through the CalSurv Gateway [28], an online database hosting data from California vector control agencies. Any records that indicated trap malfunctions or which unfeasibly ran longer than one night were excluded. Each record ( n = 24,344) contained latitude, longitude, date, number of traps employed, and total female Cx. tarsalis and Cx. pipiens captured. Distribution of traps and spray events by year (Additional file 1: Figure S1) and season (Additional file 2: Figure S2) are presented in the Supplementary Information. Any records that indicated trap malfunctions or which were operated for more than one night were excluded. For each species separately, we removed the collection reports corresponding to those greater than two standard deviations above the mean in each week to remove the influence of outliers (i.e., large singular deviations from broader abundance trends) on smooth functions subsequently estimated by the models. All geographic data were projected from geographic to planar coordinates (Albers conic equal-area, EPSG 3310, NAD83) using the rgdal package in R statistical software (version 3.3.2; [29,30]) for all data processing and analysis. Covariate Development To isolate the effects of aerial insecticide treatments within out final statistical model, we first developed a set of spatio-temporal and environmental covariates to explain the long- and short-term trends in Cx. tarsalis and Cx. pipiens abundance. Inclusion of these covariates established a counterfactual basis in the models for the expectation in abundance in the absence of control, leaving the additional terms characterizing aerial insecticide sprays to explain any deviations attributable to the treatments. For each remaining trap collection ( n = 23,707 for Cx. pipiens ; n = 23,678 for Cx. tarsalis ), we derived a set of temperature variables to capture the effect of weather on trap collections. The mean temperature during the host-seeking period (dusk to dawn) and 30-year monthly average temperature were determined for each collection using 4-km resolution data provided by the PRISM Climate Group [31]. We calculated the deviation in temperature from the monthly average during the host-seeking period to capture activity rates on the night of trapping (i.e., warmer/colder than ‘normal’ resulting in higher/lower mosquito activity and resulting trap counts). As mosquito developmental rates are highly impacted by temperature [32,33], we also calculated the average temperature during the two-week period immediately preceding the trap collection to capture short-term effects of weather on mosquito abundance. Rainfall was not considered because amounts were negligible in the study area during the season when aerial insecticide applications occurred. To characterize the larval habitat present around a trap location and to incorporate the sharp changes in land use across the study area, we used the 30x30m gridded land cover data from the 2011 National Land Cover database [34]. We used the classification of all pixels within a 5km radius area surrounding each trap to determine the proportion of three non-overlapping land use categories: urban, cultivated crops, and natural (Fig 1a). Land use categories were chosen to represent larval habitat and bionomics of the Cx. pipiens and Cx. tarsalis [35]. The radius was chosen based on known dispersal distances for the species [36,37]. The ‘urban’ category encompasses all levels of developed land (i.e. structures, roads, and constructed materials). The ‘crops’ category represents annual irrigated crops which are predominantly rice in the study area. The remaining classifications were combined to create the ‘natural’ category. We quantified the spatio-temporal intersections between spray zone polygons and trap locations to quantify the degree to which antecedent spray events impacted mosquito collections. Spatial coverage of each trap was quantified as the average proportion of the area from which a trap collects mosquitoes (‘collection area’) that was covered by aerial treatment zones during the four weeks preceding the collection. Using a conservative estimate on Culex flight distance [36–39] and to account for insecticide drift during application, we used a 5km radius collection area for both species. The temporal sequence of overlapping sprays during the four weeks preceding a collection (modeled as a factor for each unique sequence) was used to capture any lagged effects and impacts of repeated spray events. For each week preceding a collection, the total proportion of the collection area overlapping with the treatment zone was assessed. When multiple treatment zones overlapped with a single collection area in a week, we assumed an additive effect, summing the proportions of overlap from each unique spray up to a maximum of 1.0 that represented complete coverage. We used the average spatial coverage of targeted spray zones during all weeks with at least one overlapping spray event to quantify the spatial impact of sprays for each specified temporal sequence of spraying. Traps > 5km from all treatment zones had a spatial coverage of zero and corresponded to the reference level of the temporal sequence factor. These traps were included to capture baseline spatio-temporal mosquito dynamics in the absence of aerial treatments. Therefore, the impact of aerial spray events on collections was quantified with a two-fold approach, namely with the factor corresponding to the sequence of weeks when spraying overlapped during the preceding four weeks and the average proportion of the collection area that overlapped under that sequence. According to guidelines from the California Department of Public Health, periods of high risk for arbovirus transmission are characterized in part by abnormally high mosquito abundance [12]. In order to capture this dramatic deviation from ‘normal’ abundance that our smoothed modeling framework could not capture, but that precipitated aerial spray events, we also applied a prospective assessment to identify spray events closely following each collection. To account for the time required to respond to a high-risk period, we assessed the presence of overlapping treatment zones with a collection in the following four weeks on the weekly scale, similar to the above retrospective assessment of sprays. To capture potential differences between broad classes of pesticides used (organophosphate vs. pyrethrin and pyrethroids combined), we included a binary indicator variable for whether at least one spray event associated with a particular trap collection used an organophosphate. Sample sizes were too small to further investigate differences between pyrethrins and pyrethroids or between individual insecticide products. We considered time in a variety of ways to capture trends in mosquito abundance in two parts: typical annual seasonality and coarser spatial-temporal trend over the twelve-year study period. Using the trap collection dates, ‘week’ (number of weeks from the start of the study period; range 1-626) and ‘day’ (range 1-365) variables were created to capture a continuous yearly effect and seasonality, respectively. We interacted the ‘day’ variable with each category of land use (i.e. ‘urban’, ‘crops’, and ‘natural’) to capture the seasonal trend in these different habitats. Each individual seasonality curve was weighted by the proportion of land use in that category within 5 km of the trap collection to produce a single unified seasonal trend that reflected the specific habitat composition for that collection. Statistical Analysis We developed generalized additive models (GAMs) to relate nightly trap counts of female mosquitoes, either Cx. tarsalis or Cx. pipiens , to aerial adulticide applications, adjusted for variation in trap counts due to spatio-temporal mosquito dynamics. We chose GAMs because of the flexible parameterization of smooth functions of covariates to explain spatial and temporal trends [40,41]. Covariates considered to explain baseline mosquito dynamics were day or week of the year, year, location, land use, two-week average temperature, and nightly deviations from average temperature during trapping, the presence of a spray event in the following one to four weeks (high risk period), and pesticide class used in the aerial spray. Either a smooth function or a factor were used in fitting the covariates with the choice between these forms, along with spline and basis dimension if a smooth function was chosen, based on model fit and biological relevance. Cyclic cubic regression splines were used to prevent discontinuity between the ends of the smooth representing the seasonal patterns (aka between Dec 31 and Jan 1). Thin plate regression splines were chosen for most covariates because they are isotropic and have been shown to be the optimal smoother of any given basis dimension [42]. A cubic regression spline was used in the spatio-temporal surface due to its superior performance over thin plate regression splines for the large amount of observations [43]. We fit negative binomial GAMs using the gam function in R (version 3.3.2; package mgcv ) [29,44] with restricted estimation maximum likelihood (REML) as the smoothing parameter estimation method. We choses a negative binomial function to account for the over-dispersed nature of trap count data. We used backward selection guided by AIC [45] to reach our final model. In each model, we included an on offset term for the number of traps operated per trapping event. All covariates were included in the initial model and choices of interactions between covariates entered into the initial model were guided by biological relevance. We used concurvity, a measure of collinearity for smooth functions (range 0-1; 43), and visually examined deviance residuals for consistency in space and time to assess the final model fit. Using the other covariates to establish the expected abundance of each species in the absence of aerial spraying, we estimated the mean change in predicted abundance across the range of spray regimes observed in the data, using the Bayesian posterior covariance matrix for the parameters that accounted for smoothing parameter uncertainty [43]. We simulated 10,000 random draws from the posterior distribution of the fitted model, a multivariate normal distribution with mean equal to the estimated model coefficients and covariance matrix of the parameters, to predict the abundance of each species across the spatial and temporal sequences of sprays observed in the data. For each draw, we then calculated the mean change in abundance from the baseline no-spray scenario at each point on the spatio-temporal surface, along with the corresponding 95% confidence interval. Estimates of efficacy from the model were compared with those derived from Mulla’s formula [16,17]. An R script outlining the workflow of parameter development, model fitting, and estimating change in abundance across the spatio-temporal surface presented in Additional file 3: Text S1. Results Data overview and model selection The relative abundance (number of females per trap-night) of Cx. pipiens and Cx. tarsalis varied spatially during the peak WNV season from late Jun to early Oct when aerial sprays occurred (Fig 2). Typically, higher abundance of Cx. pipiens was observed in urban areas whereas higher abundance of Cx. tarsalis was typically in non-urbanized areas near irrigated agriculture. The final model for each species included an offset for the number of traps run per collection event; and smooth functions of space by time (on the weekly timescale across the twelve years), day of the year by each land use category (‘urban’, ‘crops’, and ‘natural’), two-week average temperature, nightly deviations in average temperature during trapping, and spatio-temporal impacts of aerial spraying. Our choice of cut-off for removing outliers during model fitting did not significantly change our results. Removing the top 0%, 4.5%, or 10% of data in each week for each species resulted in minor shifts in confidence interval widths and magnitude of some estimates, but no change to inference. The resulting smooth functions for each species are illustrated in Fig 3 (see Additional files 4-5: Figure S3-S4 for spatio-temporal surfaces for all years for each species). Construction of smooth functions used in the final models is outlined in Table S1 (Additional file 6). All smooth functions were highly significant ( P < 0.0001). A random intercept for site location was included to account for repeated collections at the same location, fitted using coefficients penalized by a ridge penalty [46]. We also retained, based on reductions in AIC, parameters for the presence of sprays in the 1 & 4 and 1, 2, & 3 weeks following the trap collection for Cx. tarsalis and Cx. pipiens , respectively (Additional file 7: Table S2). Based on reduction in AIC, only the model for Cx. pipiens retained the term indicating the presence of at least one spray event with an organophosphate pesticide during the previous four weeks, as compared to all sprays using combinations of pyrethrin or pyrethroid products (-48.9% change in abundance for ≥ 1 organophosphate, P < 0.001). The largest magnitude of variability in the baseline abundance for both species was due primarily to the seasonality covariates, followed by the spatio-temporal surface. The temperature covariates contributed the smallest magnitude to establishing abundance, but all smoothed functions were highly significant ( P < 0.0001). In all smooth functions, positive estimates correspond to increases in the population, negative estimates correspond to decreases in the population, and 0 indicates no modulation in abundance at that covariate value. The distribution of the final model residuals was right-skewed for both species indicating the model underestimated extreme trap counts. However, no spatial or temporal pattern remained in the deviance residuals (Additional files 8-9: Figures S5-S6). Both models had low estimated concurvity values [43] for the aerial spraying smooth function with the rest of the model parameters ( Cx. pipiens: 0.145; Cx. tarsalis: 0.148). This indicates that the smooth estimates for the impact of aerial spraying were not confounded by other parameters. In addition, the relatively high deviance explained value for both models ( Cx. pipiens : 44.0%; Cx. tarsalis : 62.3%) indicates good model fit despite for the complex dynamics inherent in mosquito populations. Effects of aerial insecticide treatments A smooth surface of the spatio-temporal impacts on Cx. pipiens abundance is presented for treatments with only pyrethrin or pyrethroid products (Fig 4a) or with at least one organophosphate product (Fig 4b). For Cx. tarsalis , the difference in impact by broad pesticide class was not retained in the final model, so a single smooth is presented (Fig 4c). Overall, the models estimated a lower magnitude of change in Cx. tarsalis abundance as compared to Cx. pipiens . For example, following aerial spraying with full spatial coverage (i.e. 100% coverage of the area within 5 km of the trap), we estimated a mean one-week Cx. pipiens abundance change of -52.4% (95% CI: -65.6, -36.5%) if all spray events had used pyrethroid or pyrethrin products. If at least one organophosphate product had been used, we estimated a -76.2% (95% CI: -82.8, -67.9%) mean change in Cx. pipiens abundance. In contrast, Cx. tarsalis populations with full spatial coverage by aerial sprays showed an estimated -30.7% (95% CI: -54.5, 2.5%) mean change one-week post-spraying regardless of the product class. For both species, larger reductions in abundances were estimated in areas with higher spatial coverage of aerial sprays (large proportion of spatial overlap) than those on the fringes (low proportion of spatial overlap). Sprays occurring closer in time to collections were generally estimated to result in larger reductions in abundance compared to those longer ago. At longer time lags (i.e., two to four weeks post-spraying), higher than expected abundance for both species was estimated, with the increase only occurring in Cx. pipiens populations following sprays with pyrethrins and pyrethroids. The majority of temporal spray sequences lacked data across the full range of spatial overlap (0-100%); we did not estimate the change in abundance for these areas. Data were sparser or lacking for higher spatial coverage during multiple weeks of spraying. For regions of the spatio-temporal surface with data support, the reduction one-week post-spraying with full spatial coverage was the largest reduction predicted for Cx. tarsalis. In contrast, Cx. pipiens populations on the fringes of spray events for the preceding four weeks results in a similar reduction as for populations with full spatial coverage by aerial sprays one-week ago (all pyrethrin/pyrethroids: -54.3% (95% CI: -81.0, -6.2%) change; at least one organophosphate: -77.2% (95% CI: -90.4, -53.9%) change). Comparison to conventional estimates As a comparison for our model results, we applied the conventional approach of Mulla’s formula [16,17] to the combined trapping and control records from 2006-2017 to estimate the efficacy of sprays. Considering trapping one-week before and after a spray event and using a 5km buffer around the targeted spray zone as the adjacent comparison area, we were able to calculate the effect for 36 spray events (3.87%) for Cx. pipiens and Cx. tarsalis ; the majority of spray events lacked traps in all of the required spatial and temporal locations for the calculation. Most of the estimates for the change in abundance indicated a reduction in trap counts, but estimates varied widely, ranging from complete population elimination (100% decrease) up to 11,000% increases following a spray event (Additional file 10:Figure S7). Most estimates for Cx. pipiens indicated varying degrees of reduction while those for Cx. tarsalis spanned reductions to increases. Discussion This study found that aerial insecticide treatments achieve strong short-term reductions in both Cx. tarsalis and Cx. pipiens populations. Previous studies have assessed the short-term impact of aerial spraying on mosquito abundance and highlighted the volatility of estimates across space and time. In order to overcome the limitations of using single events to estimate the efficacy of aerial spraying on reducing the abundance of WNV vector mosquitoes, we used a large dataset of surveillance and control records together with GAM models. This modeling framework allowed us to establish baseline mosquito adult abundance and identify deviations from expected nightly abundance attributed to aerial spraying (i.e. counterfactual basis) as well as the spatial and temporal impacts of aerial applications. Our results indicate that aerial sprays do achieve population reduction for both Cx. pipiens and Cx. tarsalis with heterogeneity in the magnitude and pattern of reduction between species and pesticide product. The differences in the magnitude and pattern of estimated response between species can be attributed partially to the different bionomics of the individual mosquito species. In the study area, Cx. pipiens are predominantly peridomestic with larval habitats limited primarily to backyard sources and stormwater systems in urbanized areas [35,47]. Thus, areas targeted by aerial insecticide treatments would span a large fraction of any particular population, leaving few adults to repopulate the treated area from proximal unsprayed locations. In contrast, Cx. tarsalis breed in agricultural areas and may disperse into surrounding agricultural and urban areas [35,48,49]. This species also has a larger typical dispersal distance and achieves higher population densities than Cx. pipiens [35–37]. Aerial insecticide treatments typically target only a small fraction of the total available habitat for Cx. tarsalis, often near urbanized areas, and any effect of aerial sprays could be moderated by immigration of adult Cx. tarsalis from surrounding unsprayed locations, potentially from distant locations [38,50]. Therefore, the best suppression for Cx. tarsalis populations would be achieved in isolated areas, as has been reported previously [22]. Additionally, repetition of sprays on the weekly scale is less effective at controlling Cx. tarsalis than Cx. pipiens because of the rapid immigration and emergence of new adults from large areas of productive larval habitat. While our estimated reduction in abundance of Culex mosquitoes show some similarity to previous published estimates of aerial spray events from Sacramento and Yolo counties (Table 1), our methodology also accounted for contextual factors, resulting in more robust estimates of the average effect of spraying. Previous estimates exhibited spatial heterogeneity [19]. Utilizing covariates to capture the spatial structure, temperature deviations, and varying distribution of larval habitats removed the confounding impact of these factors on our model results. Similarly, previous estimates have varied, in part because Mulla’s formula cannot fully capture spatio-temporal nuances of mosquito population dynamics. For example, Lothrop et al. [20] observed 73% increases in Cx. tarsalis abundances post-spraying despite observing mortality in caged sentinel mosquitoes and large reductions during previous spray events. The authors attributed the estimated increase to the dynamics of Cx. tarsalis populations at the time of the study, particularly the large emergence of Cx. tarsalis following the annual flooding of the nearby wetlands that was not captured by the Mulla’s formula framework and the fact that the sprays were not impacting mosquitoes in the productive larval habitats. Previous estimates also use differing time interval lengths to estimate mosquito abundance before and after spray events. The heterogeneity in these time intervals combined with the heterogeneity of resulting estimates (Table 1) highlights the need to use consistent time intervals to improve generalizability of estimates between studies. As standardization of the time interval largely depends on the operational capacity of vector control districts for trapping and responding to epidemic conditions, no single recommendation may be feasible across all studies. However, we recommend that mosquito control agencies should keep the timeframe consistent across their evaluations to increase comparability of intra-agency control efforts.A similar range of estimates for change in Culex abundance following adulticide treatments have been reported outside California. Most are in broad agreement with our findings, although none used Mulla’s formula. In Chicago, Illinois, a reduction of 54% in Cx. pipiens trap counts within the spray zone vs. the baseline pre-spray abundance was reported in contrast with a 153% increase outside the spray zone following two, single-night aerial spray events with a pyrethroid seven days apart [51]. An average 65.3% reduction in Cx. pipiens/restuans populations was observed within 24 hours of truck-mounted applications of a pyrethroid [52]. In contrast, no significant changes in Cx. pipiens abundance were observed following single-night truck-mounted applications of a pyrethroid in three communities near Boston, Massachusetts [53]. Up to 75% reduction in the two-day counts of female Cx. quinquefasciatus was reported during a month-long period with truck-mounted pyrethroid sprayed five days a week in Dubai, United Arab Emirates [54]. Without untreated comparison locations, it is hard to directly compare these results to those of our study. Our model structure most closely compares to the study design used by Elnaiem et al. [21] where a timescale of one week before and after a spray event with a pyrethroid pesticide was chosen when assessing mosquito abundance (Table 1). Our estimates for reduction for Cx. pipiens (-52.4%) and Cx. tarsalis (-30.7%) are lower than the observed reductions ( Cx. pipiens: - 75%; Cx. tarsalis : -48.7%). While qualitatively similar, the differences in magnitude may be due to differences in analytical methods, shifts from pyrethrin to pyrethroids over time, or the longer twelve-year time period of our study that could have yielded a more conservative estimate of average spray effects. Our approach to causal inference using observational data builds upon earlier contributions from the fields of environmental science and epidemiology. Mulla’s formula can be considered an extension of the Before/After and Control/Impact (BACI) analysis framework. Originally defined by Green [55] and extended and applied by others [56–59], BACI originated in environmental science literature to distinguish natural variability from the impact of an anthropogenic disturbance and has been applied to mosquito larvicide evaluations [60,61]. The BACI methodology compares an impact and at least one separate control location, sampled at various time points before and after the impact, to detect changes in the natural history of the environment due to the impacts [56–58]. An ANOVA test is used to detect a significant difference in the trajectories before vs. after the disturbance in the impact area as compared to the control area. Location and timing of sampling in each is chosen to ensure each are independent across space and time and increase the evidence that a detected change was attributable to the disturbance itself [56,59]. Our GAM framework extends BACI using a three-dimensional spatio-temporal function and other covariates to capture entire spatio-temporal context as a way to estimate the expected mosquito abundance in the absence of spraying. The functions also capture trends in the impact over spatial and temporal combinations, while the BACI framework may miss significant changes due to the sampling timeframe chosen [62]. Additionally, our methods do not depend on the ANOVA assumptions of independence and homoscedasticity of samples [63], as the spatial and temporal correlation and the non-normal distribution of trap collections are accounted for through the covariates in the negative binomial GAMs. Our statistical approach for estimating the effects of mosquito control on abundance relies on counterfactual theory that has been applied in the field of epidemiology as a conceptual basis for understanding measures of effect [64-66]. Counterfactual theory as a basis for causal inference is premised on the idea that for any unit being observed, there are multiple potential exposures but only one actually occurs, and outcomes under other alternative exposures exist only as potential outcomes that would have occurred if an alternative exposure had been applied. Because the alternative exposures did not occur, these are contrary to fact, or counterfactual. In this study, our units of study are trapping locations, and we are seeking to understand the effect of aerial spraying by statistically relating the observed mosquito abundance following spray events to the mosquito abundance in the same place and time that would have been observed in the absence of the spray. BACI and Mulla’s formula approaches utilize untreated control sites to establish expectations for the unsprayed condition. Our approach instead aims to estimate the counterfactual expectation for mosquito abundance directly at the same place and time using spatio-temporal trends and contextual variables (i.e., weather and land use). This approach offers two clear advantages for estimating the effects of public-health pesticide use: (1) it allows for use of rich observational data sets that exist already and capture pesticide usage in real operational contexts, as opposed to experimental settings that are often closer to ideal conditions and (2) it does not rely on pre-selected, untreated control sites, which is helpful because vector management programs are rarely willing to withhold treatments in experimental control sites if their public-health action thresholds are met. The smooth functions associated with land-use categories in the final GAMs accurately captured known seasonal and population dynamics. Cultivated crops were the primary source of Cx. tarsalis with smaller contributions from other non-urban land types during the warmest months of the year, accurately representing the presence of highly productive larval habitats in clean, recently created water sources characteristic of cultivated crops [35,48,67]. Urbanized areas did not produce large numbers of Cx. tarsalis during the WNV season as they contain few suitable larval habitats for this species. The estimated peak in abundance occurred in late July, but remained high through September, capturing the variation in timing of the peak across the years of the study. Populations of Cx. tarsalis in the Sacramento Valley are greatest from July-September [35,68,69]. The steep slope of the curve up to the peak mimicked the rapid increase in Cx. tarsalis observed at the start of the planting season [35,69]. For Cx. pipiens , urbanized areas largely contributed to abundance throughout the year with additional contributions in natural areas (aka non-cultivated croplands) later in the season, reflecting the presence of larval habitats in artificial structures like storm drains or dairy wastewater lagoons [70,71]. Crops generally had lower Cx. pipiens abundance across the season, reflecting the general lack of high quality suitable larval habits in these areas. Temperature anomalies during trapping and the two-week average antecedent temperature prior to trapping contributed to the overall abundance of both species, albeit relatively weakly in the presence of the other spatio-temporal terms. Their inclusion in the model was required to fully account for mosquito dynamics and night-to-night fluctuations in trapping success. Concordant with previous experiments, extremes in the average temperatures reduced abundance for both species illustrating the negative impacts on mosquito developmental rates and adult survivorship [32,33]. The estimated region of positive contribution to abundance for both species ( Cx. pipiens : 19.2-25.4°C ; Cx. tarsalis : 18.9-27.0°C) was narrower than the thermal tolerance of the species, but contains the observed regions of rapid developmental and high reproduction rates and the typical temperature ranges during the summer. As expected, small anomalies in average temperature on the night of trapping made relatively small contributions to change in abundance while extreme deviations result in much more marked change, highlighting the non-linear relationship underlying temperature and trap success. It is interesting to note the additional marked reduction in Cx. pipiens abundance when at least one organophosphate product was used, especially as compared to the lack of a similar difference in Cx. tarsalis populations. As mentioned above, the class of product used may be more important for Cx. pipiens due to their focal distributions and more limited dispersal [35,72]. As an aerial spray will likely impact a large proportion of the localize Cx. pipiens population at once, there will be limited immigration from unsprayed segments of the population in the nearby proximity. Therefore, the full effect of an aerial spray is discernable. In contrast, the dispersed nature of Cx. tarsalis populations facilitates rapid immigration from surrounding unsprayed locations [35,36,38], thus diluting any difference in effect between product classes; any difference is not discernable against the background population dynamics accounted for in our modeling framework. Another factor contributing to the difference by species could be insecticide resistance, as resistance to pyrethroids and organophosphates have been reported for both species in California [14,73]. If Cx. pipiens populations in the study area were more resistant to pyrethroids than Cx. tarsalis as has been previously reported in the Central Valley [74–76], this could explain the increased efficacy of organophosphates for Cx. pipiens . However, since we found a stronger effect of pyrethroids on Cx. pipiens as compared to Cx. tarsalis , resistance does not fully explain the observed difference. Alternatively, a single organophosphate spray may be insufficient to produce a marked difference in Cx. tarsalis populations; a repetition may be required. The underlying shape of the smooth function of spatio-temporal impacts of aerial spraying for either species likely differs between product classes and the specific timing and number of different products used, but sparse data prevented us from including an interaction to assess these dynamics. A potential population rebound effect occurred for both species at more distant time lags from spraying where abundance was estimated to be higher than expected two to four ( Cx. pipiens under pyrethrin and pyrethroid sprays) and three to four ( Cx. tarsalis ) weeks post-spray. Appropriately spaced treatments in time may be required to maintain a long-term reduction in population abundance. However, such a rebound does not negate the potential value of aerial treatments for achieving short-term reductions in the abundance of WNV-infected adult mosquitoes during periods of epidemic risk. The increase in abundance at low spatial coverages of sprays for both species could reflect excito-repellency of pesticides at the fringes of targeted areas. Excito-repellency, a form of behavioral avoidance, combines two forms of sub-lethal exposure that results in mosquito movement away from a chemical source; contact excitation (increased activity upon contact) and non-contact spatial repellency [77–79]. These non-toxic behavioral impacts of pesticides were first identified in Anopheles mosquitoes in response to DDT and later with insecticide-treated bed nets and indoor residual spraying [80–82]. Populations of Cx. quinquefasciatus, another species in the Cx. pipiens complex, exhibit strong contact excitation and poor spatial repellency to pyrethroid, organophosphate, and carbamate pesticides [83,84]. No study has investigated excito-repellency in Cx. tarsalis populations. The behavioral avoidance of Culex to pesticides could be pushing mosquitoes out of the spray zones, resulting in an increase in abundance around the fringes of sprays, consistent with our estimates for spatial coverages < 40% for Cx. pipiens following sprays with pyrethrin or pyrethroids only in the previous week. A limitation of choosing a GAM framework is that we were only able to capture the average effects of covariates on nightly mosquito trap counts and unable to fully account for large stochastic fluctuations inherent to mosquito populations. However, GAMs easily allowed us to incorporate nonlinear relationships between abundance and covariates without having to constrain relationships with a priori knowledge. In particular, we were able to capture the higher-order relationships and correlation across space and time with the three-dimensional spatio-temporal function. The form of the smooth relationships in the final model did appear to approximate what is observed in nature. Other strengths of our modeling approach are that it takes into account regional differences in mosquito populations, population dynamics and seasonality in different land use types, and the impacts of short-term (night) and longer-term (two week) weather, resulting in robust estimation of the baseline expected abundance in the absence of spray effects, allowing us to isolate the deviations in abundance due to aerial spraying. This counterfactual basis of the model enables estimation in the absence of an independent control. However, even with the large amount of data available, data were inadequate to estimate the impact of aerial spraying reliably for certain time lags or spatial coverages that were rare or absent in the data. This is primarily due to typical SYMVCD spraying and trapping practices due to logistical and financial constraints. SYMVCD concentrated their mosquito collections efforts near urban areas to maximize the sensitivity for assessing the risk in proximity to human populations while minimizing time and costs associated with large-scale mosquito surveillance. Additionally, in an effort to control mosquitoes in known problem areas (highly productive larval habitats in proximity to the margins of urban areas) and reduce aerial applications over urban areas, the majority of areas receiving repeated sprays across the WNV season are in more rural areas where the mosquito trapping is sparser. This limited the data and statistical power to quantify the full range of spatial overlap with aerial sprays. We were also unable to account explicitly for drift outside the target zones during sprays, as has been previously described [20]. Our use of a continuous variable to measure spray coverage within the 5km-radius collection areas surrounding each trap partially accounts for this effect. As such, however, we are unable to fully parse out the effect of aerial spraying on populations outside aerial spray zones and limited our analysis to only assess spatial coverage of traps within targeted spray zones. Additionally, we were unable to estimate the relative effects of different lengths of multi-night spray events (one vs two vs three consecutive nights) due to data limitations and our analytical choice to aggregate all sprays on the weekly scale to achieve the balance between robust estimates and operationally relevant information for vector control districts. These limitations of observational studies such as this one could be addressed in future experimental field trials. Conclusions Aerial adulticides were shown to achieve short-term reductions in the abundance of the primary West Nile virus vectors, Cx. tarsalis and Cx. pipiens . A greater reduction was estimated for Cx. pipiens , likely due to its focal distribution in urbanizes areas and limited dispersal. The use of organophosphate products versus a combination of pyrethrins and pyrethroids increased the magnitude of reduction estimated for Cx. pipiens while the difference by broad insecticide class was not significant for Cx. tarsalis . The effects of aerial sprays on Cx. tarsalis populations were likely moderated by the species broad dispersal ability, large population sizes, and vast expanses of productive larval habitat in the study area. Therefore, the best control of Cx. tarsalis would be achieved in areas with isolated or highly spatially segmented populations. For both species, aerial spraying reduced abundance at high spatial coverage while reductions were also estimated at lower spatial coverage, at albeit greatly reduced magnitudes, indicating that aerial sprays had some impacts beyond the target zone. There was also evidence for population rebounds at periods of two to four weeks post-spraying. Our modeling approach allowed us to utilize observational data to isolate aerial treatment effects while taking into account contextual factors like spatio-temporal relationships, weather, and habitat that contribute to stochastic variation in nightly trap counts. This is an important advance that complements experimental trials and expands upon conventional observational approaches that summarize population changes following aerial treatments at individual time points. Further work should expand upon these methods to estimate the change in WNV transmission potential and resulting human infections following aerial spray events. List Of Abbreviations GAM: generalized additive model: SYMVCD: Sacramento-Yolo Mosquito and Vector Control District: WNV: West Nile virus Declarations Acknowledgements We would like to thank Ruben Rosas, Marcia Reed, Sarah Wheeler, Samer Elkashef and Gary Goodman from Sacramento-Yolo Mosquito & Vector Control District for providing trapping and aerial treatment data used in this study and for their insight on mosquito control practices. We also thank the two anonymous reviewers for their helpful comments which improved the manuscript. Ethics approval and consent to participate Not applicable Consent for publication Not applicable Availability of data and materials The data that support the findings of this study were obtained from the Sacramento-Yolo Mosquito and Vector Control District. These data were used with permission for the current study and are not publicly available. Data are however available from the authors upon reasonable request and with permission from the Sacramento-Yolo Mosquito and Vector Control District. Competing interests The authors declare that they have no competing interests. Funding KH acknowledges funding support from the Floyd & Mary Schwall Fellowship in Medical Research at UC Davis, and KH and CB acknowledge support from the Pacific Southwest Center of Excellence in Vector-Borne Diseases funded by the U.S. Centers for Disease Control and Prevention (Cooperative Agreement 1U01CK000516). KH acknowledges supported by the National Center for Advancing Translational Sciences, National Institutes of Health, through grant number UL1 TR001860 and linked award TL1 TR001861. Authors’ contributions KH performed the analysis and wrote the initial draft of the manuscript. CB conceived the project idea, supervised the research, and contributed to revisions of the manuscript. RR provided statistical guidance. All authors read and approved the final manuscript. References Hayes EB, Sejvar JJ, Zaki SR, Lanciotti RS, Bode A V, Campbell GL. Virology, pathology, and clinical manifestations of West Nile virus disease. Emerg Infect Dis. 2005;11(8):1174–9. McLean RG, Ubico SR, Docherty DE, Hansen WR, Sileo L, McNamara TS. West Nile virus transmission and ecology in birds. Ann N Y Acad Sci. 2001;951:54–7. Kilpatrick AM, LaDeau SL, Marra PP. Ecology of West Nile virus transmission and its impact on birds in the western hemisphere. Auk. 2007;124(4):1121–36. Turell MJ, Dohm DJ, Sardelis MR, O Guinn ML, Andreadis TG, Blow JA. An update on the potential of North American mosquitoes (Diptera : Culicidae) to transmit West Nile virus. J Med Entomol. 2005;42(1):57–62. Kramer LD, Styer LM, Ebel GD. A global perspective on the epidemiology of West Nile virus. Annu Rev Entomol. 2007/07/25. 2008;53:61–81. Goddard LB, Roth AE, Reisen WK, Scott TW. Vector competence of California mosquitoes for West Nile virus. Emerg Infect Dis. 2002/12/25. 2002;8(12):1385–91. Reisen WK, Lothrop HD, Chiles RE, Madon MB, Cossen C, Woods L, et al. West Nile virus in California. Emerg Infect Dis. 2004;10:1369–78. Mostashari F, Bunning ML, Kitsutani PT, Singer DA, Nash D, Cooper MJ, et al. Epidemic West Nile encephalitis, New York, 1999: results of a household-based seroepidemiological survey. Lancet. 2001;358(9278):261–4. Hughes JM, Wilson ME, Sejvar JJ. The long-term outcomes of human West Nile virus infection. Clin Infect Dis. 2007;44(12):1617–24. California Department of Public Health. Human West Nile virus activity, California, 2003-2019. [Internet]. 2019. Available from: http://westnile.ca.gov/ Gubler DJ, Campbell GL, Nasci R, Komar N, Petersen L, Roehrig JT. West Nile virus in the United States: guidelines for detection, prevention, and control. Viral Immunol. 2001/02/24. 2000;13(4):469–75. California Department of Public Health, Mosquito and Vector Control Association of California, University of California. California mosquito-borne virus surveillance & response plan. 2019;(April). Available from: www.westnile.ca.gov Rose RI. Pesticides and public health: integrated methods of mosquito management. Emerg Infect Dis. 2001/03/27. 2001;7(1):17–23. California Department of Health Services Vector-Borne Disease Section. Overview of mosquito control practices in California. [Internet]. 2005. Available from: https://www.cdph.ca.gov/Programs/CID/DCDC/CDPH Document Library/OverviewofMosquitoControlinCA.pdf Schleier JJ, Peterson RKD. Pyrethrins and pyrethroid insecticides. In: Lopez O, Fernfmdez-Bolafios JG, editors. Green Trends in Insect Control. Royal Society of Chemisty; 2011. p. 94–131. Reisen WK. Using “Mulla’s Formula” to estimate percent control. In: Atkinson PW, editor. Vector Biology, Ecology, and Control. New York: Springer Science+Business Media B.V.; 2010. p. 127–38. Mulla MS, Norland RL, Fanara DM, Darwezeh HA, McKeen DW. Control of chironomid midges in recreational lakes. J Econ Entomol. 1971;64:300–7. Nielsen CF, Reisen WK, Armijos V, Wheeler S, Kelley K, Brown D. Impact of climate variation and adult mosquito control on the West Nile virus epidemic in Davis, California during 2006. Proc Pap Mosq Vector Control Assoc Calif. 2007;75:125–30. Macedo PA, Nielsen CF, Reed M, Kelley K, Reisen WK, Goodman GW, et al. An evaluation of the aerial spraying conducted in reponse to West Nile virus activity in Yolo county. Proc Pap Mosq Vector Control Assoc Calif. 2007;75:107–14. Lothrop H, Lothrop B, Palmer M, Wheeler S, Gutierrez A, Lothrop H, et al. Evaluation of pyrethrin aerial ultra-low volume applications for adult Culex tarsalis control in the desert environments of the Coachella Valley, Riverside county, California. J Am Mosq Control Assoc. 2007;23(4):405–19. Elnaiem DA, Kelley K, Wright S, Laffey R, Yoshimura G, Reed M, et al. Impact of aerial spraying of pyrethrin insecticide on Culex pipiens and Culex tarsalis (Diptera: Culicidae) abundance and West Nile virus infection rates in an urban/suburban area of Sacramento County, California. J Med Entomol. 2008;45(4):751–7. Reisen WK, Milby MM, Reeves WC, Eberle MW, Meyer RP, Schaefer CH, et al. Aerial adulticiding for the suppression of Culex tarsalis in Kern County, California, using low volume propoxur: 2. Impact on natural populations in foothill and valley habitats. J Am Mosq Control Assoc. 1985;1(2):154–63. U. S. Census Bureau. State and County QuickFacts [Internet]. 2019 [cited 2019 Oct 14]. Available from: https://www.census.gov/quickfacts United State Census Bureau. Percent urban and rural in 2010 by state and county. [Internet]. Available from: https://www.census.gov/geo/reference/ua/urban-rural-2010.html Cline G, Neigher A, Bellinder A. Climate of Sacramento, California. National Weather Service Office. Sacramento, California; 2010. Sacramento-Yolo Mosquito & Vector Control District. Mission & Vision [Internet]. Elk Grove, CA: Sacramento-Yolo Mosquito & Vector Control District; 2018 [cited 2019 Oct 5]. Available from: http://www.fightthebite.net/about/about-us/ Newhouse VF, Chamberlain RW, Johnson JG, Sudia WD. Use of dry ice to increase mosquito catches of the CDC miniature light trap. Mosq News. 1966;26(1):30–5. California Vectorborne Disease Surveillance System (CalSurv). [Internet]. 2018. Available from: https://vectorsurv.org/ R Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statisitcal Computing; 2020. Available from: https://www.r-project.org/ Bivand R, Keitt T, Rowlingson B. rgdal: bindings for the “geospatial” data abstraction library. 2019. Available from: https://cran.r-project.org/package=rgdal PRISM Climate Group, Oregon State University. Daily mean and monthly average temperature datasets. [Internet]. 2018. [Accessed 2018 May 12]. Available from: http://prism.oregonstate.edu Ciota AT, Matacchiero AC, Kilpatrick AM, Kramer LD. The effect of temperature on life history traits of Culex mosquitoes. J Med Ent. 2014;51(1):55–62. Reisen WK. Effect of temperature on Culex tarsalis (Diptera: Culicidae) from the Coachella and San Joaquin valleys of California. J Med Entomol. 1995;32(5):636–45. Multi-resolution land characteristics consortium. NLCD Land Cover (CONUS) Database [Internet]. 2011 [cited 2018 May 20]. Available from: https://www.mrlc.gov/data?f%5B0%5D=category%3Aland cover&f%5B1%5D=year%3A2011 Reisen WK, Reeves WC. Bionomics and ecology of Culex tarsalis and other potential mosquito vector species. In: Reeves WC, editor. Epidemiology and control of mosquito-borne arboviruses in California, 1943-1987. Sacramento, CA: California Mosquito and Vector Control Association; 1990. p. 254–329. Bailey SF, Eliason DA, Hoffmann BL. Flight and dispersal of mosquito Culex tarsalis Coquillett in Sacramento Valley of California. Hilgardia. 1965;37(3):73–113. Reisen WK, Milby MM, Meyer RP, Pfuntner AR, Spoehel J, Hazelrigg JE, et al. Mark-release-recapture studies with Culex mosquitoes (Diptera: Culicidae) in southern California. J Med Entomol. 1991;28(3):357–71. Dow RP, Reeves WC, Bellamy RE. Dispersal of female Culex tarsalis into a larvicided area. Am J Trop Med Hyg. 1965;14(4):656–70. Reisen WK, Milby MM, Meyer RP. Population dynamics of adult Culex mosquitoes (Diptera: Culicidae) along the Kern River, Kern county, California, in 1990. J Med Ent. 1992;29(3):531–43. Hastie T, Tibshirani R. Varying-coefficient models. J R Stat Soc Ser B-Methodological. 1993;55(4):757–96. Hastie T, Tibshirani RJ. Generalized additive models. Stat Sci. 1986;1(3):297–318. Wood SN. Thin plate regression splines. J R Stat Soc Ser B. 2003;65(1):95–114. Wood SN. Generalized additive models: an introduction with R. Boca Raton, FL: Chapman & Hall/CRC; 2006. Wood SN. Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. J R Stat Soc Ser B-Statistical Methodol. 2011;73:3–36. Aike HAI. A new look at the statistical model identification. IEEE Trans Automat Contr. 1974;19(6):716–23. Wood SN. Fast stable direct fitting and smoothness selection for generalized additive models. J R Stat Soc Ser B-Statistical Methodol. 2008;70:495–518. Hopkins CC, Hollinger FB, Johnson RF, Dewlett HJ, Newhouse VF, Chamberlain RW. The epidemiology of St. Louis encephalitis in Dallas, Texas, 1966. Am J Epidemiol. 1975;102(1):1–15. Wekesa JW, Yuval B, Washino RK. Spatial distribution of adult mosquitoes (Diptera:Culicidae) in habitats associated with the rice agroecosystem of northern California. J Med Entomol. 1996;33(3):344–50. Mitchell CJ, Kilpatrick JW, Hayes RO, Curry HW. Effects of ultra-low volume applications of malathion in Hale county, Texas II: mosquito populations in treated and untreated areas. J Med Ent. 1970;7(1):85–91. Reisen WK, Yoshimura G, Reeves WC, Milby MM, Meyer RP. The impact of aerial applications of ultra-low volume adulticides on Culex tarsalis populations (Diptera: Culicidae) in Kern County, California, USA, 1982. J Med Entomol. 1984;21(5):573–85. Mutebi JP, Delorey MJ, Jones RC, Plate DK, Gerber SI, Gibbs KP, et al. The impact of adulticide applications on mosquito density in Chicago, 2005. J Am Mosq Control Assoc. 2011;27(1):69–76. Clifton ME, Xamplas CP, Nasci RS, Harbison J. Gravid Culex pipiens exhibit a reduced susceptibility to ultra-low volume adult control treatments under field conditions. J Am Mosq Control Assoc. 2019;35(4):267–78. Reddy MR, Spielman A, Lepore TJ, Henley D, Kiszewski AE, Reiter P. Efficacy of resmethrin aerosols applied from the road for suppressing Culex vectors of West Nile virus. Vector-Borne Zoonotic Dis. 2006;6(2):117–27. Holmes PR. A study of population changes in adult Culex quinquefasciatus Say (Diptera: Culicidae) during a mosquito control programme in Dubai, United Arab Emirates. Ann Trop Med Parasitol. 1986;80(1):107–16. Green RH. Sampling design and statistical methods for environmental biologists. Chichester, UK: Wiley Interscience; 1979. Underwood AJ. Beyond BACI: the detection of environmental impacts on populations in the real, but variable, world. J Exp Mar Bio Ecol. 1992;161:145–78. Stewart-Oaten A, Murdoch WW, Parker KR. Environmental impact assessment: “pseudoreplication” in time? Ecology. 1986;67(4):929–40. Stewart-Oaten A, Bence JR. Temporal and spatial variation in environmental impact assessment. Ecol Monogr. 2001;71(2):305–39. Underwood AJ. On Beyond BACI: Sampling Designs that Might Reliably Detect Environmental Disturbances. Ecol Appl. 1994;4(1):3–15. Rochlin I, Iwanejko T, Dempsey ME, Ninivaggi D V. Geostatistical evaluation of integrated marsh management impact on mosquito vectors using before-after-control-impact (BACI) design. Int J Health Geogr. 2009;8(1):1–20. Wolfram G, Wenzl P, Jerrentrup H. A multi-year study following BACI design reveals no short-term impact of Bti on chironomids (Diptera) in a floodplain in Eastern Austria. Environ Monit Assess. 2018;190(12). Smokorowski KE, Randall RG. Cautions on using the Before-After-Control-Impact design in environmental effects monitoring programs. Facets. 2017;2(1):212–32. Stewart-Oaten A, Bence JR, Osenberg CW. Assessing effects of unreplicated perturbations: no simple solutions. Ecology. 1992;73(4):1396–404. Lewis D. Causation. J Philos. 1973;70:556–67. Rubin DB. Comment: Neyman (1923) and Causal Inference in Experiments and Observational Studies. Stat Sci. 1990;5. Greenland S, Rothman KJ, Lash TL. Measures of effect and measures of association. In: Rothman KJ, Greenland S, Lash TL, editors. Modern Epidemiology, 3 rd edition. Lippincott, Williams and Wilkins Philadelphia; 2008. p. 51–70. Reisen WK, Meyer RP, Shields J, Arbolante C. Population ecology of preimaginal Culex tarsalis (Diptera: Culicidae) in Kern County, California. J Med Entomol. 1989/01/01. 1989;26(1):10–22. Loomis EC, Meyers EG. California encephalitis surveillance program: mosquito population measurement and ecologic considerations. Am J Epidemiol. 1960;71(3):378–88. Barker CM, Eldridge BF, Reisen WK. Seasonal abundance of Culex tarsalis and Culex pipiens complex mosquitoes (Diptera: Culicidae) in California. J Med Ent. 2010;47(5):759–68. Loomis EC, Eide RN, Caton JR, Merritt DA. Physical and cultural control: Mosquito and fly problems in dairy waste-water systems. Calif Agr. 1980;34(3):37–8. Su TY, Webb JR, Meyer RR, Mulla MS. Spatial and temporal distribution of mosquitoes in underground storm drain systems in Orange County, California. J Vector Ecol. 2003;28(1):79–89. Reisen WK. The contrasting bionomics of Culex mosquitoes in Western North America. J Am Mosq Control Assoc. 2012;28(4):82–91. California Department of Public Health Vector-Borne Disease Section. California mosquito pesticide resistance summary. [Internet]. 2005. Available from: http://westnile.ca.gov/downloads.php?download_id=3137&filename=Pesticide_Resistance.pdf Reed M, Macedo PA, Brown D. Increased tolerance to permethrin in Culex pipiens complex population from Sacramento County, California. Proc Pap Mosq Vector Control Assoc Calif. 2012;80:56–8. Reed M, Macedo PA, Goodman G, Brown D. Tough mosquitoes - why they should be everyone’s problem. Proc Pap Mosq Vector Control Assoc Calif. 2013;81:75–9. Sorensen MA, Stevenson JA. An adulticide resistance “heat map” for Culex pipiens and Culex tarsalis (Diptera: Culididae) in Placer County. Proc Pap Mosq Vector Control Assoc Calif. 2015;83:8–11. Dethier VG, Browne LB, Smith CW. The designation of chemicals in terms of the responses they elicit from insects. J Econ Entomol. 1960;53(1):134–6. Roberts DR, Chareonviriyaphap T, Harlan HH, P H. Methods ot testing and analyzing excito-repellency responses of malaria vectors to insecticides. J Am Mosq Control Assoc. 1997;3(1):13–7. Miller JR, Siegert PY, Amimo FA, Walker ED. Designation of chemicals in terms of the locomotor responses they elicit from insects: an update of Dethier et al. (1960). J Econ Entomol. 2009;102(6):2056–60. Lindsay SW, Adiamah JH, Miller JE, Armstrong JRM. Pyrethroid-treated bednet effects on mosquitoes of the A nopheles gambiae complex in The Gambia. Med Vet Entomol. 1991;5:477–83. Kennedy JS. The excitant and repellent effects on mosquitos of sub-lethal contacts with DDT. Bull Entomol Res. 1945;37(4):593–607. Davidson G. Experiments on the effect of residual insecticides in houses against Anopheles gambiae and A. funestus . Bull Entomol Res. 1953;44:231–54. Boonyuan W, Bangs MJ, Grieco JP, Tiawsirisup S, Prabaripai A, Chareonviriyaphap T. Excito-repellent responses between Culex quinquefasciatus permethrin susceptible and resistant mosquitoes. J Insect Behav. 2016;29:415–31. Sathantriphop S, Ketavan C, Prabaripai A, Visetson S, Bangs MJ, Akratanakul P, et al. Susceptibility and avoidance behavior by Culex quinquefasciatus Say to three classes of residual insecticides. J Vector Ecol. 2006;31(2):266–74. Macedo PA, Schleier JJ, Reed M, Kelley K, Goodman GW, Brown DA, et al. Evaulation of efficacy and human health risk of aerial ultra-low volume applications of pyrethrins and piperonyl butoxide for adult mosquito management in response to West Nile virus activity in Sacramento county, California. J Am Mosq Control Assoc. 2010;26(1):57–66. Tables Table 1 Estimated change in Culex mosquito populations following aerial spray events in California using Mulla’s formula. Location City, county Year Nights Sprayed # Product Class Comparison Length † Species % Change ‡ Reference Davis, Yolo 2006 2 pyrethrin 2 days before, 2 days after Cx. pipiens - 58.0 [19] Cx. tarsalis - 25.6 Woodland, Yolo 2006 2 pyrethrin 2 days before, 2 days after Cx. pipiens - 77.7 Cx. tarsalis - 46.8 Sacramento, Sacramento 2005 3 pyrethrin 7 days before, 7 days after Cx. pipiens - 75.0 [21] Cx. tarsalis - 48.7 Sacramento, Sacramento 2006 3 pyrethrin 3 days before, 3 days after Cx. pipiens - 39.3 [82] Cx. tarsalis - 57.3 Coachella valley, Riverside 2005 (Mar) 3 alt^ pyrethroid 5 days before, 1 day after Cx. tarsalis - 93.0 [20] 2005 (Jun) 3 alt^ pyrethroid 5 days before, 1 day after Cx. tarsalis - 77.0 2005 (Sep) 3 alt^ pyrethroid 5 days before, 1 day after Cx. tarsalis 73.0 # Number of consecutive nights sprayed in spray event. † Length of time before and after an aerial spray event used when comparing of trap counts. ‡ Percent change in Culex abundance following an aerial spray event as calculated using Mulla’s formula [16,17] . ^ Aerial spraying occurred on 3 alternate nights. Supplementary Files Additionalfile1.tif Additional file 1 (.tif) Additional file 1: Figure S1. Location of CO2-baited mosquito trapping events and aerial spray events stratified by year (2006-2017). A random jitter of ≤ 1 km applied to trapping locations for visualization of repeated events at the same site. Each spray event polygon represents the area targeted during a single aerial spray application. Additionalfile2.tif Additional file 2 (.tif) Additional file 2: Figure S2. Location of CO2-baited mosquito trapping events and aerial spray events stratified by season. Season defined into three-month intervals. A random jitter of ≤ 1 km applied to trapping locations for visualization of repeated events at the same site. Each spray event polygon represents the area targeted during a single aerial spray application. Additionalfile3.docx Additional file 3 (.doc) Additional file 3: Text S1. R script outlining our workflow of covariate development, GAM model fitting, and estimated change in abundance. Additionalfile4.tif Additional file 4 (.tif) Additional file 4: Figure S3 Spatial surface for Cx. pipiens at midpoint of the typical WNV season (week of Aug 1). Surface presented for each year (2006-2017) reflects the relative abundance of the species and is a slice from the three-dimensional spatio-temporal smoothed function. Contours applied for visualization of estimates. Additionalfile5.tif Additional file 5 (.tif) Additional file 5: Figure S4. Spatial surface for Cx. tarsalis at midpoint of the typical WNV season (week of Aug 1). Surface presented for each year (2006-2017) reflects the relative abundance of the species and is a slice from the three-dimensional spatio-temporal smoothed function. Contours applied for visualization of estimates. Additionalfile6.docx Additional file 6 (.doc) Additional file 6: Table S1. Smooth functions included in the final GAMs for both Cx. tarsalis and Cx. pipiens. Table outlining the construction of smooth function used in the final models, including spline type and basis dimensions chosen. Additionalfile7.docx Additional file 7 (.doc) Additional file 7: Table S2. Change (%) in nightly abundance from expected for collections preceding aerial spraying. Table indicating the final model estimates for change in expected trap-counts for collections in the one to four weeks preceding an aerial spray event. Additionalfile8.tif Additional file 8 (.tif) Additional file 8: Figure S5. Spatial and temporal distribution of model deviance residuals for Cx. pipiens for 2006-2017. Residuals presented spatially at the associated trapping location (random jitter of ≤ 1 km applied for visualization of repeated events). Additionalfile9.tif Additional file 9 (.tif) Additional file 9:Figure S6. Spatial and temporal distribution of model deviance residuals for Cx. tarsalis for 2006-2017. Residuals presented spatially at the associated trapping location (random jitter of ≤ 1 km applied for visualization of repeated events). Additionalfile10.tif Additional file 10 (.tif) Additional file 10:Figure S7. Estimated percent change in Cx. pipiens and Cx. tarsalis populations with Mulla’s formula. Change estimated for the 36 aerial sprays in Sacramento and Yolo counties, CA (2006-2017) with associated trap collections within the targeted zone (treated) and an adjacent 5km buffer (control) within one-week before and one-week following spraying. Cite Share Download PDF Status: Published Journal Publication published 24 Feb, 2021 Read the published version in Parasites & Vectors → Version 2 posted Editorial decision: Accept 27 Jan, 2021 Review # 2 received at journal 23 Jan, 2021 Reviewer # 2 agreed at journal 19 Jan, 2021 Reviewer # 1 agreed at journal 15 Jan, 2021 Review # 1 received at journal 15 Jan, 2021 Editor assigned by journal 12 Jan, 2021 Reviewers invited by journal 12 Jan, 2021 Submission checks completed at journal 12 Jan, 2021 Editor invited by journal 12 Jan, 2021 You are reading this latest preprint version Show more versions Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-92266","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research","associatedPublications":[],"authors":[{"id":8876263,"identity":"b63786b1-b8b0-48f9-9b38-815b60d6ba44","order_by":0,"name":"Karen M. Holcomb","email":"","orcid":"","institution":"University of California Davis School of Veterinary Medicine","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Karen","middleName":"M.","lastName":"Holcomb","suffix":""},{"id":8876264,"identity":"c0b3d31f-be71-4953-affa-67131db86240","order_by":1,"name":"Robert C. Reiner","email":"","orcid":"","institution":"Institute for Health Metrics and Evaluation","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Robert","middleName":"C.","lastName":"Reiner","suffix":""},{"id":8876265,"identity":"e42a0cad-a175-4ce4-a04b-4f53f35148e6","order_by":2,"name":"Christopher M Barker","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA4klEQVRIiWNgGAWjYFACHiA+wMDAD6QkwIiBwYA4LZINJGsxOABVT1AL/7SzBz9XnLHJN76R/PDGzzYLewb25m0S+LRI3M5LljxzI81y2400Y8veNonEBp5jZXi1GEjnGEg2fDhsYHYjh02C54xEAoNEjhkhLcY/Gz78NzCekcMm+eeMhD2D/BuCWswkG24cMDCQyGGT5qmQYGyQ4MGvBeiXNMuGM8kGEmeeGVvLVEgktvGkFVvg08I/O/fwzYZjdgb87ckPb74xqLPnZz+88QY+LZiAjTTlo2AUjIJRMAqwAQAGr0S8vxMxpgAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0002-7941-346X","institution":"University of California Davis","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Christopher","middleName":"M","lastName":"Barker","suffix":""}],"badges":[],"createdAt":"2020-10-13 22:25:00","currentVersionCode":2,"declarations":"","doi":"10.21203/rs.3.rs-92266/v2","doiUrl":"https://doi.org/10.21203/rs.3.rs-92266/v2","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s13071-021-04616-6","type":"published","date":"2021-02-24T15:00:32+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":5335354,"identity":"92d174f8-9ebe-4466-8b53-f2bf605af706","added_by":"auto","created_at":"2021-01-28 12:36:48","extension":"tif","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":266057,"visible":true,"origin":"","legend":"Land cover, aerial sprays, and mosquito collections (2006-2017) in Sacramento \u0026 Yolo counties, California. (a) Distribution of cultivated crops (primarily rice), urban, and other natural land cover types across the study area. Land cover categories were derived from the 2011 National Land Cover database [34]. Inset highlights the location of these counties in the state of California. Location of (b) zones targeted for aerial treatment applications and (c) CO2-baited mosquito traps during 2006-2017 in Sacramento and Yolo counties. Each polygon (b) and point (c) represents a single spray or trapping event, respectively. A random shift of ≤ 1 km applied to trap locations for visualization of repeated trapping at the same location across time.","description":"","filename":"Fig1.tif","url":"https://assets-eu.researchsquare.com/files/rs-92266/v2/7c1c7fdc6d4f205cee95c8b0.tif"},{"id":5335185,"identity":"d37c244d-e2d4-461d-b78d-54767b372707","added_by":"auto","created_at":"2021-01-28 12:30:49","extension":"tif","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":605799,"visible":true,"origin":"","legend":"Collections of (a) Cx. pipiens and (b) Cx. tarsalis during peak WNV season. Plus signs (+) indicate the location and number of female mosquitoes per trap-night for each collection during the period when aerial spraying occurred (late-Jun to early-Oct). Colors represent abundance quintiles by species for non-zero collections. A random shift of ≤ 1 km was applied to trap locations to aid visualization of repeated collections at the same locations during the study timeframe.","description":"","filename":"Fig2.tif","url":"https://assets-eu.researchsquare.com/files/rs-92266/v2/cb59c0f06806b5d42b7a9af3.tif"},{"id":5335184,"identity":"979e4688-9a1a-49e3-aff8-21787a0ccabf","added_by":"auto","created_at":"2021-01-28 12:30:49","extension":"tif","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":189970,"visible":true,"origin":"","legend":"Smooth covariate functions explaining nightly abundance of (a) Cx. pipiens and (b) Cx. tarsalis. Smooth functions from final GAMs are shown for the spatio-temporal surface (top), seasonality in a fully urban area, seasonality in a fully crop area, seasonality in a fully natural area, deviation (°C) from the 30-year monthly average temperature on the night of trapping, and the average temperature (°C) during the two-weeks prior to trapping (bottom). The shaded region represents 95% confidence interval for one-dimensional functions. A representative slice of the three-dimensional spatio-temporal surface is presented for 2011 at the midpoint of the typical WNV season (week of Aug 1). Spatio-temporal surfaces are plotted on individual axes for each species to resolve the spatial scale.","description":"","filename":"Fig3.tif","url":"https://assets-eu.researchsquare.com/files/rs-92266/v2/f3d1d871c6b42a6ed4562b2f.tif"},{"id":5335224,"identity":"4aa0c417-df9c-4281-99f9-ecd34e5de2b4","added_by":"auto","created_at":"2021-01-28 12:33:48","extension":"tif","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":311844,"visible":true,"origin":"","legend":"Mean changes in abundance following aerial spraying, as compared to no-spray baseline. Estimates are shown for (a, b) Cx. pipiens and (c) Cx. tarsalis abundance changes with respect to antecedent sequence and average spatial coverage of aerial treatments. For Cx. pipiens, estimates are shown for (a) sprays that used only pyrethrin or pyrethroid products and for (b) sprays that utilized an organophosphate product at least once. For Cx. tarsalis, estimates are with any product class. Horizontal axes represent the average proportion of the 5km buffer surrounding a trap covered by a spray event and vertical axes represent the temporal sequence of aerial sprays during the four weeks preceding the trapping event. Presence (1) or absence (0) of sprays in the 1, 2, 3, and 4 weeks (R to L) prior to trap are indicated by the 4-digit sequence. The sequences of sprays are ordered from the fewest number and temporally most distant spray events (bottom) to the largest number and temporally closest spray events (top). Estimates are truncated to the range observed with the available data (grey squares indicate points present in data). Areas enclosed in a black border represent the portion of the spatio-temporal surface with significant estimates (P \u003c 0.05). ","description":"","filename":"Fig4.tif","url":"https://assets-eu.researchsquare.com/files/rs-92266/v2/3eff095d758cd1d36adec649.tif"},{"id":13652077,"identity":"dd767b5d-e506-44ff-baa2-cc097365edd0","added_by":"auto","created_at":"2021-09-17 09:48:11","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3117473,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-92266/v2/0a56977e-0be5-49b1-8c7a-8877a845d7f6.pdf"},{"id":5335355,"identity":"3b9aa6ca-d913-4cbe-a0c7-bf856b806d8a","added_by":"auto","created_at":"2021-01-28 12:36:48","extension":"tif","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1487520,"visible":true,"origin":"","legend":"Additional file 1 (.tif)\nAdditional file 1: Figure S1. Location of CO2-baited mosquito trapping events and aerial spray events stratified by year (2006-2017). A random jitter of ≤ 1 km applied to trapping locations for visualization of repeated events at the same site. Each spray event polygon represents the area targeted during a single aerial spray application.\n","description":"","filename":"Additionalfile1.tif","url":"https://assets-eu.researchsquare.com/files/rs-92266/v2/3b786288df092be7dce27a97.tif"},{"id":5335356,"identity":"1ae57335-aaec-44f8-ab25-3fd3a3faa7c2","added_by":"auto","created_at":"2021-01-28 12:36:49","extension":"tif","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":987882,"visible":true,"origin":"","legend":"Additional file 2 (.tif)\nAdditional file 2: Figure S2. Location of CO2-baited mosquito trapping events and aerial spray events stratified by season. Season defined into three-month intervals. A random jitter of ≤ 1 km applied to trapping locations for visualization of repeated events at the same site. Each spray event polygon represents the area targeted during a single aerial spray application.\n","description":"","filename":"Additionalfile2.tif","url":"https://assets-eu.researchsquare.com/files/rs-92266/v2/faffc27632326aeb9b36bfe2.tif"},{"id":5335220,"identity":"34d69501-cfe5-41ad-a8d4-8065d8b4eeac","added_by":"auto","created_at":"2021-01-28 12:33:47","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":25195,"visible":true,"origin":"","legend":"Additional file 3 (.doc)\nAdditional file 3: Text S1. R script outlining our workflow of covariate development, GAM model fitting, and estimated change in abundance.\n","description":"","filename":"Additionalfile3.docx","url":"https://assets-eu.researchsquare.com/files/rs-92266/v2/6c1d4f50916a96004fcc0d83.docx"},{"id":5335178,"identity":"2717c574-22b7-4cf0-b8d9-7fa6c895eb6b","added_by":"auto","created_at":"2021-01-28 12:30:48","extension":"tif","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":12839574,"visible":true,"origin":"","legend":"Additional file 4 (.tif)\nAdditional file 4: Figure S3 Spatial surface for Cx. pipiens at midpoint of the typical WNV season (week of Aug 1). Surface presented for each year (2006-2017) reflects the relative abundance of the species and is a slice from the three-dimensional spatio-temporal smoothed function. Contours applied for visualization of estimates.\n","description":"","filename":"Additionalfile4.tif","url":"https://assets-eu.researchsquare.com/files/rs-92266/v2/fd8df092f2e10897aa4071c5.tif"},{"id":5335189,"identity":"c07d2aa5-e840-4dea-9016-e22c88f15fb8","added_by":"auto","created_at":"2021-01-28 12:30:49","extension":"tif","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":14334180,"visible":true,"origin":"","legend":"Additional file 5 (.tif)\nAdditional file 5: Figure S4. Spatial surface for Cx. tarsalis at midpoint of the typical WNV season (week of Aug 1). Surface presented for each year (2006-2017) reflects the relative abundance of the species and is a slice from the three-dimensional spatio-temporal smoothed function. Contours applied for visualization of estimates.\n","description":"","filename":"Additionalfile5.tif","url":"https://assets-eu.researchsquare.com/files/rs-92266/v2/77559b35f0fe4178d1aed7d3.tif"},{"id":5335357,"identity":"d402cafe-9fff-4335-9318-91b1914248d8","added_by":"auto","created_at":"2021-01-28 12:36:49","extension":"docx","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":17710,"visible":true,"origin":"","legend":"Additional file 6 (.doc)\nAdditional file 6: Table S1. Smooth functions included in the final GAMs for both Cx. tarsalis and Cx. pipiens. Table outlining the construction of smooth function used in the final models, including spline type and basis dimensions chosen.\n","description":"","filename":"Additionalfile6.docx","url":"https://assets-eu.researchsquare.com/files/rs-92266/v2/9f1efde9b895898f5ac72800.docx"},{"id":5335222,"identity":"9a60b578-a85d-430a-b90a-5146a681c23c","added_by":"auto","created_at":"2021-01-28 12:33:48","extension":"docx","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":13589,"visible":true,"origin":"","legend":"Additional file 7 (.doc)\nAdditional file 7: Table S2. Change (%) in nightly abundance from expected for collections preceding aerial spraying. Table indicating the final model estimates for change in expected trap-counts for collections in the one to four weeks preceding an aerial spray event.\n","description":"","filename":"Additionalfile7.docx","url":"https://assets-eu.researchsquare.com/files/rs-92266/v2/d461a9db016eae1169e7dc49.docx"},{"id":5335223,"identity":"9e6270f6-a24b-498f-bacd-e3809c5b05a2","added_by":"auto","created_at":"2021-01-28 12:33:48","extension":"tif","order_by":8,"title":"","display":"","copyAsset":false,"role":"supplement","size":9222120,"visible":true,"origin":"","legend":"Additional file 8 (.tif)\nAdditional file 8: Figure S5. Spatial and temporal distribution of model deviance residuals for Cx. pipiens for 2006-2017. Residuals presented spatially at the associated trapping location (random jitter of ≤ 1 km applied for visualization of repeated events). \n","description":"","filename":"Additionalfile8.tif","url":"https://assets-eu.researchsquare.com/files/rs-92266/v2/985eabb52ae37632cd475c66.tif"},{"id":5335226,"identity":"6352ad49-59ab-43a9-8e20-d9a59aaa9072","added_by":"auto","created_at":"2021-01-28 12:33:49","extension":"tif","order_by":9,"title":"","display":"","copyAsset":false,"role":"supplement","size":9081012,"visible":true,"origin":"","legend":"Additional file 9 (.tif)\nAdditional file 9:Figure S6. Spatial and temporal distribution of model deviance residuals for Cx. tarsalis for 2006-2017. Residuals presented spatially at the associated trapping location (random jitter of ≤ 1 km applied for visualization of repeated events). \n","description":"","filename":"Additionalfile9.tif","url":"https://assets-eu.researchsquare.com/files/rs-92266/v2/e9f52bb221908f044d97841f.tif"},{"id":5335188,"identity":"a34ff44a-7dd5-4e87-b8dd-8dac64c9a471","added_by":"auto","created_at":"2021-01-28 12:30:49","extension":"tif","order_by":10,"title":"","display":"","copyAsset":false,"role":"supplement","size":451302,"visible":true,"origin":"","legend":"Additional file 10 (.tif)\nAdditional file 10:Figure S7. Estimated percent change in Cx. pipiens and Cx. tarsalis populations with Mulla’s formula. Change estimated for the 36 aerial sprays in Sacramento and Yolo counties, CA (2006-2017) with associated trap collections within the targeted zone (treated) and an adjacent 5km buffer (control) within one-week before and one-week following spraying.\n","description":"","filename":"Additionalfile10.tif","url":"https://assets-eu.researchsquare.com/files/rs-92266/v2/6542a207c05e0dc0d75ae9f6.tif"}],"financialInterests":"","formattedTitle":"Spatio-temporal impacts of aerial adulticide applications on populations of West Nile virus vector mosquitoes","fulltext":[{"header":"Background","content":"\u003cp\u003eWest Nile virus (WNV;\u0026nbsp;genus \u003cem\u003eFlavivirus\u003c/em\u003e, family \u003cem\u003eFlaviviridae\u003c/em\u003e) causes a potentially fatal, neuroinvasive mosquito-borne disease\u0026nbsp;[1]. It is maintained in an enzootic cycle between birds and mosquitoes\u0026nbsp;[2,3], predominantly in the genus \u003cem\u003eCulex\u003c/em\u003e [4], and can spillover to infect horses and humans, both of which are dead-end hosts vulnerable to disease\u0026nbsp;[5]. \u003cem\u003eCulex (Cx.) tarsalis\u0026nbsp;\u003c/em\u003eand \u003cem\u003eCx. pipiens\u003c/em\u003e complex mosquitoes are the primary enzootic and epizootic vectors in California\u0026nbsp;[6,7]. While\u0026nbsp;80% of human infections are asymptomatic, clinical manifestations can include acute febrile illness, encephalitis, flaccid paralysis, and death\u0026nbsp;[8]. Often the severe form results in long-term physical and mental disabilities\u0026nbsp;[9]. WNV invaded California in 2003 and has become endemic\u0026nbsp;[7]. An average of 238 neuroinvasive cases occur statewide annually with approximately one-third occurring in the Central Valley, where the landscape is dominated by large-scale agriculture punctuated by cities and small towns\u0026nbsp;[10].\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAs no human vaccine exists, prevention of human diseases relies primarily on personal protective measures (i.e. wearing long sleeves, using insect repellent, and avoiding the dawn/dust periods when mosquitoes bite) and vector control by local vector control districts or health departments\u0026nbsp;[11,12]. In periods of epidemic risk when large numbers of WNV-infected \u003cem\u003eCulex\u0026nbsp;\u003c/em\u003emosquitoes are detected near human population centers,\u0026nbsp;large-scale aerial applications of insecticides are utilize to rapidly reduce the abundance of adult mosquitoes and disrupt virus transmission cycles, thereby reducing zoonotic transmission risk\u0026nbsp;[13].\u003c/p\u003e\n\u003cp\u003eThree main classes of pesticide products have been licensed for use in aerial spray applications in California; pyrethrins, pyrethroids, and organophosphates\u0026nbsp;[12,14,15]. Pyrethrins are naturally derived insecticides from chrysanthemum flowers (\u003cem\u003eChrysanthemum cineriaefolium\u003c/em\u003e) that inactivate sodium channels in the insect nervous system, resulting in paralysis and death\u0026nbsp;[15]. Pyrethroids are synthetically derived pyrethrins with a similar mode of action and longer half-life. Organophosphates inhibit acetylcholinesterase, affecting neurotransmission and causing uncontrolled nerve activation and death in insects\u0026nbsp;[14].\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eA standard method for evaluating the efficacy of an aerial spray event\u0026nbsp;compares pre- to post-treatment mosquito trap counts inside the treatment zone versus changes for the same period in an adjacent unsprayed control area\u0026nbsp;[16]. This method, first proposed by Mulla et al.\u0026nbsp;[17], has been adapted to and widely used in evaluating the efficacy of aerial spraying for reducing the abundance of female mosquitoes and has been extended to assess changes in other indicators of risk, namely infection prevalence in mosquitoes, human cases, and reported dead birds with WNV infection\u0026nbsp;[16].\u0026nbsp;However, reported estimates vary widely, with some studies even indicating occasional increases in trap counts following spray events\u0026nbsp;[18\u0026ndash;21].\u003c/p\u003e\n\u003cp\u003eDespite its wide use, the assumptions behind Mulla\u0026rsquo;s formula are often violated, resulting in confounded estimates. First, treatment and control sites are often not independent due to the spatial connectivity of populations with mosquito dispersal and immigration\u0026nbsp;[22]. With the connectivity and potential drift of pesticides via wind, there is the potential that insecticide sprays have wider population impacts than just the targeted spray zone\u0026nbsp;[20]. Second, the difference in pre- to post- trap count ratios in and between areas are not solely due to control measures, but rather are impacted by weather, seasonality in mosquito populations, differential presence of larval breeding sources or simply stochastic variation in trapping success\u0026nbsp;[18,20,21]. Overall, Mulla\u0026rsquo;s formula neglects the spatio-temporal structure of mosquito populations and external factors impacting the random volatility of mosquito trapping success. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo overcome the limitations of assessing the efficacy of aerial sprays on the individual spray event basis, we paired long-term surveillance and vector control records (12 years) from Sacramento-Yolo Mosquito and Vector Control District (SYMVCD) in California to capture baseline spatio-temporal mosquito population dynamics and estimate the magnitude and duration of the impacts of aerial sprays on the abundance of \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003eand \u003cem\u003eCx. tarsalis\u003c/em\u003e, the predominant WNV vectors in California. We chose a generalized additive modeling (GAM) framework to capture the nonlinear population dynamics and associations inherent to mosquito collections.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eStudy Area\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe study area encompasses Sacramento and Yolo counties, California (Fig 1) which have a combined area of approximately 5,126 km\u003csup\u003e2\u003c/sup\u003e and a population of approximately 1.73 million people in 2016\u0026nbsp;[23]. Sacramento County is 34.12% urban and 65.88% rural with the majority of urban areas consisting of the concentrated Sacramento urban center and surrounding suburbs. In comparison, Yolo County is 4.61% urban and 95.39% rural with smaller, more dispersed urban areas\u0026nbsp;[24]. These counties, located in the northern part of California\u0026rsquo;s Central Valley, are characterized by a Mediterranean climate with hot, dry summers (Jul mean temperature: 25.8 \u0026deg;C, May-Sep mean total rainfall: 3.18 cm) and mild, rainy winters (Jan mean temperature: 9.6 \u0026deg;C, Oct-Apr mean total rainfall: 47.30 cm)\u0026nbsp;[25]\u0026nbsp;and extensive irrigated agriculture, especially rice and row crops such as tomatoes. Sacramento-Yolo Mosquito and Vector Control District (SYMVCD), established in 1946 to protect the public from nuisance mosquito biting and mosquito-borne diseases, manages mosquito populations in Sacramento and Yolo counties\u0026nbsp;[26].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAerial Treatments and Mosquito Collections\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSYMVCD provided the spatial polygons (Fig 1b) and associated data detailing the date, area targeted for spraying, number of consecutive nights of spraying in the same location, and pesticide product used for all aerial sprays during the study period (1,021 unique nights of spraying during 930 spray events).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMosquito collection records for CDC CO\u003csub\u003e2\u003c/sub\u003e-baited EVS traps\u0026nbsp;[27]\u0026nbsp;from SYMVCD for the years 2006-2017 (Fig 1c) were obtained with permission through the CalSurv Gateway\u0026nbsp;[28], an online database hosting data from California vector control agencies. Any records that indicated trap malfunctions or which unfeasibly ran longer than one night were excluded. Each record (\u003cem\u003en\u0026nbsp;\u003c/em\u003e= 24,344) contained latitude, longitude, date, number of traps employed, and total female \u003cem\u003eCx. tarsalis\u003c/em\u003e and \u003cem\u003eCx. pipiens\u003c/em\u003e captured. Distribution of traps and spray events by year (Additional file 1: Figure S1) and season (Additional file 2: Figure S2) are presented in the Supplementary Information.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAny records that indicated trap malfunctions or which were operated for more than one night were excluded. For each species separately, we removed the collection reports corresponding to those greater than two standard deviations above the mean in each week to remove the influence of outliers (i.e., large singular deviations from broader abundance trends) on smooth functions subsequently estimated by the models.\u003c/p\u003e\n\u003cp\u003eAll geographic data were projected from geographic to planar coordinates (Albers conic equal-area, EPSG 3310, NAD83) using the \u003cem\u003ergdal\u0026nbsp;\u003c/em\u003epackage in R statistical software (version 3.3.2;\u0026nbsp;[29,30]) for all data processing and analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCovariate Development\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo isolate the effects of aerial insecticide treatments within out final statistical model, we first developed a set of spatio-temporal and environmental covariates to explain the long- and short-term trends in \u003cem\u003eCx. tarsalis\u0026nbsp;\u003c/em\u003eand \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003eabundance. Inclusion of these covariates established a counterfactual basis in the models for the expectation in abundance in the absence of control, leaving the additional terms characterizing aerial insecticide sprays to explain any deviations attributable to the treatments.\u003c/p\u003e\n\u003cp\u003eFor each remaining trap collection (\u003cem\u003en\u0026nbsp;\u003c/em\u003e= 23,707 for \u003cem\u003eCx. pipiens\u003c/em\u003e; \u003cem\u003en\u0026nbsp;\u003c/em\u003e= 23,678 for \u003cem\u003eCx. tarsalis\u003c/em\u003e), we derived a set of temperature variables to capture the effect of weather on trap collections. The mean temperature during the host-seeking period (dusk to dawn) and 30-year monthly average temperature were determined for each collection using 4-km resolution data provided by the PRISM Climate Group\u0026nbsp;[31]. We calculated the deviation in temperature from the monthly average during the host-seeking period to capture activity rates on the night of trapping (i.e., warmer/colder than \u0026lsquo;normal\u0026rsquo; resulting in higher/lower mosquito activity and resulting trap counts). As mosquito developmental rates are highly impacted by temperature\u0026nbsp;[32,33], we also calculated the average temperature during the two-week period immediately preceding the trap collection to capture short-term effects of weather on mosquito abundance. Rainfall was not considered because amounts were negligible in the study area during the season when aerial insecticide applications occurred.\u003c/p\u003e\n\u003cp\u003eTo characterize the larval habitat present around a trap location and to incorporate the sharp changes in land use across the study area, we used the 30x30m gridded land cover data from the 2011 National Land Cover database\u0026nbsp;[34]. We used the classification of all pixels within a 5km radius area surrounding each trap to determine the proportion of three non-overlapping land use categories: urban, cultivated crops, and natural (Fig 1a). Land use categories were chosen to represent larval habitat and bionomics of the \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003eand \u003cem\u003eCx. tarsalis\u0026nbsp;\u003c/em\u003e[35]. The radius was chosen based on known dispersal distances for the species\u0026nbsp;[36,37]. The \u0026lsquo;urban\u0026rsquo; category encompasses all levels of developed land (i.e. structures, roads, and constructed materials). The \u0026lsquo;crops\u0026rsquo; category represents annual irrigated crops which are predominantly rice in the study area. The remaining classifications were combined to create the \u0026lsquo;natural\u0026rsquo; category.\u003c/p\u003e\n\u003cp\u003eWe quantified the spatio-temporal intersections between spray zone polygons and trap locations to quantify the degree to which antecedent spray events impacted mosquito collections.\u0026nbsp;Spatial coverage of each trap was quantified as the average proportion of the area from which a trap collects mosquitoes (\u0026lsquo;collection area\u0026rsquo;) that was covered by aerial treatment zones during the four weeks preceding the collection. Using a conservative estimate on \u003cem\u003eCulex\u0026nbsp;\u003c/em\u003eflight distance\u0026nbsp;[36\u0026ndash;39]\u0026nbsp;and to account for insecticide drift during application, we used a 5km radius collection area for both species. The temporal sequence of overlapping sprays during the four weeks preceding a collection (modeled as a factor for each unique sequence) was used to capture any lagged effects and impacts of repeated spray events. For each week preceding a collection, the total proportion of the collection area overlapping with the treatment zone was assessed. When multiple treatment zones overlapped with a single collection area in a week, we assumed an additive effect, summing the proportions of overlap from each unique spray up to a maximum of 1.0 that represented complete coverage. We used the average spatial coverage of targeted spray zones during all weeks with at least one overlapping spray event to quantify the spatial impact of sprays for each specified temporal sequence of spraying. Traps \u0026gt; 5km from all treatment zones had a spatial coverage of zero and corresponded to the reference level of the temporal sequence factor. These traps were included to capture baseline spatio-temporal mosquito dynamics in the absence of aerial treatments. Therefore, the impact of aerial spray events on collections was quantified with a two-fold approach, namely with the factor corresponding to the sequence of weeks when spraying overlapped during the preceding four weeks and the average proportion of the collection area that overlapped under that sequence.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAccording to guidelines from the California Department of Public Health, periods of high risk for arbovirus transmission are characterized in part by abnormally high mosquito abundance\u0026nbsp;[12]. In order to capture this dramatic deviation from \u0026lsquo;normal\u0026rsquo; abundance that our smoothed modeling framework could not capture, but that precipitated aerial spray events, we also applied a prospective assessment to identify spray events closely following each collection. To account for the time required to respond to a high-risk period, we assessed the presence of overlapping treatment zones with a collection in the following four weeks on the weekly scale, similar to the above retrospective assessment of sprays.\u003c/p\u003e\n\u003cp\u003eTo capture potential differences between broad classes of pesticides used (organophosphate vs. pyrethrin and pyrethroids combined), we included a binary indicator variable for whether at least one spray event associated with a particular trap collection used an organophosphate. Sample sizes were too small to further investigate differences between pyrethrins and pyrethroids or between individual insecticide products.\u003c/p\u003e\n\u003cp\u003eWe considered time in a variety of ways to capture trends in mosquito abundance in two parts: typical annual seasonality and coarser spatial-temporal trend over the twelve-year study period.\u0026nbsp;Using the trap collection dates, \u0026lsquo;week\u0026rsquo; (number of weeks from the start of the study period; range 1-626) and \u0026lsquo;day\u0026rsquo; (range 1-365) variables were created to capture a continuous yearly effect and seasonality, respectively. We interacted the \u0026lsquo;day\u0026rsquo; variable with each category of land use (i.e. \u0026lsquo;urban\u0026rsquo;, \u0026lsquo;crops\u0026rsquo;, and \u0026lsquo;natural\u0026rsquo;)\u0026nbsp;to capture the seasonal trend in these different habitats. Each individual seasonality curve was weighted by the proportion of land use in that category within 5 km of the trap collection to produce a single unified seasonal trend that reflected the specific habitat composition for that collection.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistical Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe developed generalized additive models (GAMs) to relate nightly trap counts of female mosquitoes, either \u003cem\u003eCx. tarsalis\u0026nbsp;\u003c/em\u003eor \u003cem\u003eCx. pipiens\u003c/em\u003e, to aerial adulticide applications, adjusted for variation in trap counts due to spatio-temporal mosquito dynamics. We chose GAMs because of the\u0026nbsp;flexible parameterization of smooth functions of covariates to explain spatial and temporal trends\u0026nbsp;[40,41]. Covariates considered to explain baseline mosquito dynamics were day or week of the year, year, location, land use, two-week average temperature, and nightly deviations from average temperature during trapping, the presence of a spray event in the following one to four weeks (high risk period), and pesticide class used in the aerial spray. Either a smooth function or a factor were used in fitting the covariates with the choice between these forms, along with spline and basis dimension if a smooth function was chosen, based on model fit and biological relevance. Cyclic cubic regression splines were used to prevent discontinuity between the ends of the smooth representing the seasonal patterns (aka between Dec 31 and Jan 1). Thin plate regression splines were chosen for most covariates because they are isotropic and have been shown to be the optimal smoother of any given basis dimension\u0026nbsp;[42]. A cubic regression spline was used in the spatio-temporal surface due to its superior performance over thin plate regression splines for the large amount of observations\u0026nbsp;[43].\u003c/p\u003e\n\u003cp\u003eWe fit negative binomial GAMs using the \u003cem\u003egam\u003c/em\u003e function in R (version 3.3.2; package \u003cem\u003emgcv\u003c/em\u003e)\u0026nbsp;[29,44]\u0026nbsp;with restricted estimation maximum likelihood (REML) as the smoothing parameter estimation method. We choses a negative binomial function to account for the over-dispersed nature of trap count data. We used backward selection guided by AIC\u0026nbsp;[45]\u0026nbsp;to reach our final model.\u0026nbsp;In each model, we included an on offset term for the number of traps operated per trapping event.\u0026nbsp;All covariates were included in the initial model and choices of interactions between covariates entered into the initial model were\u0026nbsp;guided by biological relevance.\u0026nbsp;We used concurvity, a measure of collinearity for smooth functions\u0026nbsp;(range 0-1; 43), and visually examined deviance residuals for consistency in space and time to assess the final model fit.\u003c/p\u003e\n\u003cp\u003eUsing the other covariates to establish the expected abundance of each species in the absence of aerial spraying, we estimated the mean change in predicted abundance across the range of spray regimes observed in the data, using the Bayesian posterior covariance matrix for the parameters that accounted for smoothing parameter uncertainty\u0026nbsp;[43]. We simulated 10,000 random draws from the posterior distribution of the fitted model, a multivariate normal distribution with mean equal to the estimated model coefficients and covariance matrix of the parameters, to predict the abundance of each species across the spatial and temporal sequences of sprays observed in the data. For each draw, we then calculated the mean change in abundance from the baseline no-spray scenario at each point on the spatio-temporal surface, along with the corresponding 95% confidence interval. Estimates of efficacy from the model were compared with those derived from Mulla\u0026rsquo;s formula\u0026nbsp;[16,17].\u003c/p\u003e\n\u003cp\u003eAn R script outlining the workflow of parameter development, model fitting, and estimating change in abundance across the spatio-temporal surface presented in Additional file 3: Text S1.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eData overview and model selection\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe relative abundance (number of females per trap-night) of \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003eand \u003cem\u003eCx. tarsalis\u0026nbsp;\u003c/em\u003evaried spatially during the peak WNV season from late Jun to early Oct when aerial sprays occurred (Fig 2). Typically, higher abundance of \u003cem\u003eCx. pipiens\u003c/em\u003e was observed in urban areas whereas higher abundance of \u003cem\u003eCx. tarsalis\u0026nbsp;\u003c/em\u003ewas typically in non-urbanized areas near irrigated agriculture.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe final model for each species included an offset for the number of traps run per collection event; and smooth functions of space by time (on the weekly timescale across the twelve years), day of the year by each land use category (\u0026lsquo;urban\u0026rsquo;, \u0026lsquo;crops\u0026rsquo;, and \u0026lsquo;natural\u0026rsquo;), two-week average temperature, nightly deviations in average temperature during trapping, and spatio-temporal impacts of aerial spraying. Our choice of cut-off for removing outliers during model fitting did not significantly change our results. Removing the top 0%, 4.5%, or 10% of data in each week for each species resulted in minor shifts in confidence interval widths and magnitude of some estimates, but no change to inference.\u0026nbsp;The resulting smooth functions for each species are illustrated in Fig 3 (see Additional files 4-5: Figure S3-S4 for spatio-temporal surfaces for all years for each species).\u0026nbsp;Construction of smooth functions used in the final models is outlined in Table S1 (Additional file 6).\u0026nbsp;All smooth functions were highly significant (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.0001).\u0026nbsp;A random intercept for site location was included to account for repeated collections at the same location, fitted using coefficients penalized by a ridge penalty\u0026nbsp;[46]. We also retained, based on reductions in AIC, parameters for the presence of sprays in the 1 \u0026amp; 4 and 1, 2, \u0026amp; 3 weeks following the trap collection for \u003cem\u003eCx. tarsalis\u0026nbsp;\u003c/em\u003eand \u003cem\u003eCx. pipiens\u003c/em\u003e, respectively (Additional file 7: Table S2). Based on reduction in AIC, only the model for \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003eretained the term indicating the presence of at least one spray event with an organophosphate pesticide during the previous four weeks, as compared to all sprays using combinations of pyrethrin or pyrethroid products (-48.9% change in abundance for \u0026ge; 1 organophosphate, \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe largest magnitude of variability in the baseline abundance for both species was due primarily to the seasonality covariates, followed by the spatio-temporal surface. The temperature covariates contributed the smallest magnitude to establishing abundance, but all smoothed functions were highly significant (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.0001). In all smooth functions, positive estimates correspond to increases in the population, negative estimates correspond to decreases in the population, and 0 indicates no modulation in abundance at that covariate value.\u003c/p\u003e\n\u003cp\u003eThe distribution of the final model residuals was right-skewed for both species indicating the model underestimated extreme trap counts. However, no spatial or temporal pattern remained in the deviance residuals (Additional files 8-9: Figures S5-S6). Both models had low estimated concurvity values\u0026nbsp;[43]\u0026nbsp;for the aerial spraying smooth function with the rest of the model parameters (\u003cem\u003eCx. pipiens:\u0026nbsp;\u003c/em\u003e0.145; \u003cem\u003eCx. tarsalis:\u0026nbsp;\u003c/em\u003e0.148). This indicates that the smooth estimates for the impact of aerial spraying were not confounded by other parameters. In addition, the relatively high deviance explained value for both models (\u003cem\u003eCx. pipiens\u003c/em\u003e: 44.0%; \u003cem\u003eCx. tarsalis\u003c/em\u003e: 62.3%) indicates good model fit despite for the complex dynamics inherent in mosquito populations.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEffects of aerial insecticide treatments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA smooth surface of the spatio-temporal impacts on \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003eabundance is presented for treatments with only pyrethrin or pyrethroid products (Fig 4a) or with at least one organophosphate product (Fig 4b). For \u003cem\u003eCx. tarsalis\u003c/em\u003e, the difference in impact by broad pesticide class was not retained in the final model, so a single smooth is presented (Fig 4c). Overall, the models estimated a lower magnitude of change in \u003cem\u003eCx. tarsalis\u0026nbsp;\u003c/em\u003eabundance as compared to \u003cem\u003eCx. pipiens\u003c/em\u003e. For example, following aerial spraying with full spatial coverage (i.e. 100% coverage of the area within 5 km of the trap), we estimated a mean one-week \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003eabundance change of -52.4% (95% CI: -65.6, -36.5%) if all spray events had used pyrethroid or pyrethrin products. If at least one organophosphate product had been used, we estimated a -76.2% (95% CI: -82.8, -67.9%) mean change in \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003eabundance. In contrast, \u003cem\u003eCx. tarsalis\u0026nbsp;\u003c/em\u003epopulations with full spatial coverage by aerial sprays showed an estimated -30.7% (95% CI: -54.5, 2.5%) mean change one-week post-spraying regardless of the product class.\u003c/p\u003e\n\u003cp\u003eFor both species, larger reductions in abundances were estimated in areas with higher spatial coverage of aerial sprays (large proportion of spatial overlap) than those on the fringes (low proportion of spatial overlap). Sprays occurring closer in time to collections were generally estimated to result in larger reductions in abundance compared to those longer ago. At longer time lags (i.e., two to four weeks post-spraying), higher than expected abundance for both species was estimated, with the increase only occurring in \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003epopulations following sprays with pyrethrins and pyrethroids.\u003c/p\u003e\n\u003cp\u003eThe majority of temporal spray sequences lacked data across the full range of spatial overlap (0-100%); we did not estimate the change in abundance for these areas. Data were sparser or lacking for higher spatial coverage during multiple weeks of spraying. For regions of the spatio-temporal surface with data support, the reduction one-week post-spraying with full spatial coverage was the largest reduction predicted for \u003cem\u003eCx. tarsalis.\u0026nbsp;\u003c/em\u003eIn contrast, \u003cem\u003eCx. pipiens\u003c/em\u003e populations on the fringes of spray events for the preceding four weeks results in a similar reduction as for populations with full spatial coverage by aerial sprays one-week ago (all pyrethrin/pyrethroids: -54.3% (95% CI: -81.0, -6.2%) change; at least one organophosphate: -77.2% (95% CI: -90.4, -53.9%) change).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eComparison to conventional estimates\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAs a comparison for our model results, we applied the conventional approach of Mulla\u0026rsquo;s formula\u0026nbsp;[16,17]\u0026nbsp;to the combined trapping and control records from 2006-2017 to estimate the efficacy of sprays. Considering trapping one-week before and after a spray event and using a 5km buffer around the targeted spray zone as the adjacent comparison area, we were able to calculate the effect for 36 spray events (3.87%) for \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003eand \u003cem\u003eCx. tarsalis\u003c/em\u003e; the majority of spray events lacked traps in all of the required spatial and temporal locations for the calculation. Most of the estimates for the change in abundance indicated a reduction in trap counts, but estimates varied widely, ranging from complete population elimination (100% decrease) up to 11,000% increases following a spray event (Additional file 10:Figure S7). Most estimates for \u003cem\u003eCx. pipiens\u003c/em\u003e indicated varying degrees of reduction while those for \u003cem\u003eCx. tarsalis\u003c/em\u003e spanned reductions to increases.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study found that aerial insecticide treatments achieve strong short-term reductions in both \u003cem\u003eCx. tarsalis\u0026nbsp;\u003c/em\u003eand \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003epopulations. Previous studies have assessed the short-term impact of aerial spraying on mosquito abundance and highlighted the volatility of estimates across space and time. In order to overcome the limitations of using single events to estimate the efficacy of aerial spraying on reducing the abundance of WNV vector mosquitoes, we used a large dataset of surveillance and control records together with GAM models. This modeling framework allowed us to establish baseline mosquito adult abundance and identify deviations from expected nightly abundance attributed to aerial spraying (i.e. counterfactual basis) as well as the spatial and temporal impacts of aerial applications. Our results indicate that aerial sprays do achieve population reduction for both \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003eand \u003cem\u003eCx. tarsalis\u003c/em\u003e with heterogeneity in the magnitude and pattern of reduction between species and pesticide product.\u003c/p\u003e\n\u003cp\u003eThe differences in the magnitude and pattern of estimated response between species can be attributed partially to the different bionomics of the individual mosquito species. In the study area, \u003cem\u003eCx. pipiens\u003c/em\u003e are predominantly peridomestic with larval habitats limited primarily to backyard sources and stormwater systems in urbanized areas\u0026nbsp;[35,47]. Thus, areas targeted by aerial insecticide treatments would span a large fraction of any particular population, leaving few adults to repopulate the treated area from proximal unsprayed locations. In contrast, \u003cem\u003eCx. tarsalis\u0026nbsp;\u003c/em\u003ebreed in agricultural areas and may disperse into surrounding agricultural and urban areas\u0026nbsp;[35,48,49]. This species also has a larger typical dispersal distance and achieves higher population densities than \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003e[35\u0026ndash;37]. Aerial insecticide treatments typically target only a small fraction of the total available habitat for \u003cem\u003eCx. tarsalis,\u0026nbsp;\u003c/em\u003eoften near urbanized areas, and any effect of aerial sprays could be moderated by immigration of adult \u003cem\u003eCx. tarsalis\u0026nbsp;\u003c/em\u003efrom surrounding unsprayed locations, potentially from distant locations\u0026nbsp;[38,50]. Therefore, the best suppression for \u003cem\u003eCx. tarsalis\u0026nbsp;\u003c/em\u003epopulations would be achieved in isolated areas, as has been reported previously\u0026nbsp;[22]. Additionally, repetition of sprays on the weekly scale is less effective at controlling \u003cem\u003eCx. tarsalis\u0026nbsp;\u003c/em\u003ethan \u003cem\u003eCx. pipiens\u003c/em\u003e because of the rapid immigration and emergence of new adults from large areas of productive larval habitat.\u003c/p\u003e\n\u003cp\u003eWhile our estimated reduction in abundance of \u003cem\u003eCulex\u0026nbsp;\u003c/em\u003emosquitoes show some similarity to previous published estimates of aerial spray events from Sacramento and Yolo counties (Table 1), our methodology also accounted for contextual factors, resulting in more robust estimates of the average effect of spraying. Previous estimates exhibited spatial heterogeneity\u0026nbsp;[19]. Utilizing covariates to capture the spatial structure, temperature deviations, and varying distribution of larval habitats removed the confounding impact of these factors on our model results. Similarly, previous estimates have varied, in part because Mulla\u0026rsquo;s formula cannot fully capture spatio-temporal nuances of mosquito population dynamics. For example, Lothrop et al.\u0026nbsp;[20]\u0026nbsp;observed 73% increases in \u003cem\u003eCx. tarsalis\u0026nbsp;\u003c/em\u003eabundances post-spraying despite observing mortality in caged sentinel mosquitoes and large reductions during previous spray events. The authors attributed the estimated increase to the dynamics of \u003cem\u003eCx. tarsalis\u0026nbsp;\u003c/em\u003epopulations at the time of the study, particularly the large emergence of \u003cem\u003eCx. tarsalis\u003c/em\u003e following the annual flooding of the nearby wetlands that was not captured by the Mulla\u0026rsquo;s formula framework and the fact that the sprays were not impacting mosquitoes in the productive larval habitats. Previous estimates also use differing time interval lengths to estimate mosquito abundance before and after spray events. The heterogeneity in these time intervals combined with the heterogeneity of resulting estimates (Table 1) highlights the need to use consistent time intervals to improve generalizability of estimates between studies. As\u0026nbsp;standardization of the time interval largely depends on the operational capacity of vector control districts for trapping and responding to epidemic conditions,\u0026nbsp;no single recommendation may be feasible across all studies. However, we recommend\u0026nbsp;that mosquito control agencies should keep the timeframe consistent across their evaluations to increase comparability of intra-agency control efforts.A similar range of estimates for change in \u003cem\u003eCulex\u0026nbsp;\u003c/em\u003eabundance following adulticide treatments have been reported outside California. Most are in broad agreement with our findings, although none used Mulla\u0026rsquo;s formula. In Chicago, Illinois, a reduction of 54% in \u003cem\u003eCx. pipiens\u003c/em\u003e trap counts within the spray zone vs. the baseline pre-spray abundance was reported in contrast with a 153% increase outside the spray zone following two, single-night aerial spray events with a pyrethroid seven days apart\u0026nbsp;[51]. An average 65.3% reduction in \u003cem\u003eCx. pipiens/restuans\u0026nbsp;\u003c/em\u003epopulations was observed within 24 hours of truck-mounted applications of a pyrethroid\u0026nbsp;[52]. In contrast, no significant changes in \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003eabundance were observed following single-night truck-mounted applications of a pyrethroid in three communities near Boston, Massachusetts\u0026nbsp;[53]. Up to 75% reduction in the two-day counts of female \u003cem\u003eCx. quinquefasciatus\u003c/em\u003e was reported during a month-long period with truck-mounted pyrethroid sprayed five days a week in Dubai, United Arab Emirates\u0026nbsp;[54]. Without untreated comparison locations, it is hard to directly compare these results to those of our study.\u003c/p\u003e\n\u003cp\u003eOur model structure most closely compares to the study design used by Elnaiem et al.\u0026nbsp;[21]\u0026nbsp;where a timescale of one week before and after a spray event with a pyrethroid pesticide was chosen when assessing mosquito abundance (Table 1). Our estimates for reduction for \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003e(-52.4%) and \u003cem\u003eCx. tarsalis\u003c/em\u003e (-30.7%) are lower than the observed reductions (\u003cem\u003eCx. pipiens: -\u003c/em\u003e75%; \u003cem\u003eCx. tarsalis\u003c/em\u003e: -48.7%). While qualitatively similar, the differences in magnitude may be due to differences in analytical methods, shifts from pyrethrin to pyrethroids over time, or the longer twelve-year time period of our study that could have yielded a more conservative estimate of average spray effects.\u003c/p\u003e\n\u003cp\u003eOur approach to causal inference using observational data builds upon earlier contributions from the fields of environmental science and epidemiology. Mulla\u0026rsquo;s formula can be considered an extension of the Before/After and Control/Impact (BACI) analysis framework. Originally defined by Green [55] and extended and applied by others [56\u0026ndash;59], BACI originated in environmental science literature to distinguish natural variability from the impact of an anthropogenic disturbance and has been applied to mosquito larvicide evaluations [60,61]. The BACI methodology compares an impact and at least one separate control location, sampled at various time points before and after the impact, to detect changes in the natural history of the environment due to the impacts [56\u0026ndash;58]. An ANOVA test is used to detect a significant difference in the trajectories before vs. after the disturbance in the impact area as compared to the control area. Location and timing of sampling in each is chosen to ensure each are independent across space and time and increase the evidence that a detected change was attributable to the disturbance itself [56,59]. Our GAM framework extends BACI using a three-dimensional spatio-temporal function and other covariates to capture entire spatio-temporal context as a way to estimate the expected mosquito abundance in the absence of spraying. The functions also capture trends in the impact over spatial and temporal combinations, while the BACI framework may miss significant changes due to the sampling timeframe chosen [62]. Additionally, our methods do not depend on the ANOVA assumptions of independence and homoscedasticity of samples [63], as the spatial and temporal correlation and the non-normal distribution of trap collections are accounted for through the covariates in the negative binomial GAMs.\u003c/p\u003e\n\u003cp\u003eOur statistical approach for estimating the effects of mosquito control on abundance relies on counterfactual theory that has been applied in the field of epidemiology as a conceptual basis for understanding measures of effect [64-66]. Counterfactual theory as a basis for causal inference is premised on the idea that for any unit being observed, there are multiple potential exposures but only one actually occurs, and outcomes under other alternative exposures exist only as potential outcomes that would have occurred if an alternative exposure had been applied. Because the alternative exposures did not occur, these are contrary to fact, or counterfactual. In this study, our units of study are trapping locations, and we are seeking to understand the effect of aerial spraying by statistically relating the observed mosquito abundance following spray events to the mosquito abundance in the same place and time that would have been observed in the absence of the spray. BACI and Mulla\u0026rsquo;s formula approaches utilize untreated control sites to establish expectations for the unsprayed condition. Our approach instead aims to estimate the counterfactual expectation for mosquito abundance directly at the same place and time using spatio-temporal trends and contextual variables (i.e., weather and land use). This approach offers two clear advantages for estimating the effects of public-health pesticide use: (1)\u0026nbsp;it allows for use of rich observational data sets that exist already and capture pesticide usage in real operational contexts, as opposed to experimental settings that are often closer to ideal conditions and (2) it does not rely on pre-selected, untreated control sites, which is helpful because vector management programs are rarely willing to withhold treatments in experimental control sites if their public-health action thresholds are met.\u003c/p\u003e\n\u003cp\u003eThe smooth functions associated with land-use categories in the final GAMs accurately captured known seasonal and population dynamics. Cultivated crops were the primary source of \u003cem\u003eCx. tarsalis\u003c/em\u003e with smaller contributions from other non-urban land types during the warmest months of the year, accurately representing the presence of highly productive larval habitats in clean, recently created water sources characteristic of cultivated crops\u0026nbsp;[35,48,67]. Urbanized areas did not produce large numbers of \u003cem\u003eCx. tarsalis\u003c/em\u003e during the WNV season as they contain few suitable larval habitats for this species. The estimated peak in abundance occurred in late July, but remained high through September, capturing the variation in timing of the peak across the years of the study. Populations of \u003cem\u003eCx. tarsalis\u0026nbsp;\u003c/em\u003ein the Sacramento Valley are greatest from July-September\u0026nbsp;[35,68,69]. The steep slope of the curve up to the peak mimicked the rapid increase in \u003cem\u003eCx. tarsalis\u0026nbsp;\u003c/em\u003eobserved at the start of the planting season\u0026nbsp;[35,69]. For \u003cem\u003eCx. pipiens\u003c/em\u003e, urbanized areas largely contributed to abundance throughout the year with additional contributions in natural areas (aka non-cultivated croplands) later in the season, reflecting the presence of larval habitats in artificial structures like storm drains or dairy wastewater lagoons\u0026nbsp;[70,71]. Crops generally had lower \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003eabundance across the season, reflecting the general lack of high quality suitable larval habits in these areas.\u003c/p\u003e\n\u003cp\u003eTemperature anomalies during trapping and the two-week average antecedent temperature prior to trapping contributed to the overall abundance of both species, albeit relatively weakly in the presence of the other spatio-temporal terms. Their inclusion in the model was required to fully account for mosquito dynamics and night-to-night fluctuations in trapping success. Concordant with previous experiments, extremes in the average temperatures reduced abundance for both species illustrating the negative impacts on mosquito developmental rates and adult survivorship\u0026nbsp;[32,33]. The estimated region of positive contribution to abundance for both species (\u003cem\u003eCx. pipiens\u003c/em\u003e: 19.2-25.4\u0026deg;C\u003cem\u003e; Cx. tarsalis\u003c/em\u003e: 18.9-27.0\u0026deg;C) was narrower than the thermal tolerance of the species, but contains the observed regions of rapid developmental and high reproduction rates and the typical temperature ranges during the summer. As expected, small anomalies in average temperature on the night of trapping made relatively small contributions to change in abundance while extreme deviations result in much more marked change, highlighting the non-linear relationship underlying temperature and trap success.\u003c/p\u003e\n\u003cp\u003eIt is interesting to note the additional marked reduction in \u003cem\u003eCx. pipiens\u003c/em\u003e abundance when at least one organophosphate product was used, especially as compared to the lack of a similar difference in \u003cem\u003eCx. tarsalis\u0026nbsp;\u003c/em\u003epopulations. As mentioned above, the class of product used may be more important for \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003edue to their focal distributions and more limited dispersal\u0026nbsp;[35,72]. As an aerial spray will likely impact a large proportion of the localize \u003cem\u003eCx. pipiens\u003c/em\u003e population at once, there will be limited immigration from unsprayed segments of the population in the nearby proximity. Therefore, the full effect of an aerial spray is discernable. In contrast, the dispersed nature of \u003cem\u003eCx. tarsalis\u003c/em\u003e populations facilitates rapid immigration from surrounding unsprayed locations\u0026nbsp;[35,36,38], thus diluting any difference in effect between product classes; any difference is not discernable against the background population dynamics accounted for in our modeling framework. Another factor contributing to the difference by species could be insecticide resistance, as resistance to pyrethroids and organophosphates have been reported for both species in California\u0026nbsp;[14,73]. If \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003epopulations in the study area were more resistant to pyrethroids than \u003cem\u003eCx. tarsalis\u003c/em\u003e as has been previously reported in the Central Valley\u0026nbsp;[74\u0026ndash;76], this could explain the increased efficacy of organophosphates for \u003cem\u003eCx. pipiens\u003c/em\u003e. However, since we found a stronger effect of pyrethroids on \u003cem\u003eCx. pipiens\u003c/em\u003e as compared to \u003cem\u003eCx. tarsalis\u003c/em\u003e, resistance does not fully explain the observed difference. Alternatively, a single organophosphate spray may be insufficient to produce a marked difference in \u003cem\u003eCx. tarsalis\u003c/em\u003e populations; a repetition may be required. The underlying shape of the smooth function of spatio-temporal impacts of aerial spraying for either species likely differs between product classes and the specific timing and number of different products used, but sparse data prevented us from including an interaction to assess these dynamics.\u003c/p\u003e\n\u003cp\u003eA potential population rebound effect occurred for both species at more distant time lags from spraying where abundance was estimated to be higher than expected two to four (\u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003eunder pyrethrin and pyrethroid sprays) and three to four (\u003cem\u003eCx. tarsalis\u003c/em\u003e) weeks post-spray. Appropriately spaced treatments in time may be required to maintain a long-term reduction in population abundance. However, such a rebound does not negate the potential value of aerial treatments for achieving short-term reductions in the abundance of WNV-infected adult mosquitoes during periods of epidemic risk.\u003c/p\u003e\n\u003cp\u003eThe increase in abundance at low spatial coverages of sprays for both species could reflect excito-repellency of pesticides at the fringes of targeted areas. Excito-repellency, a form of behavioral avoidance, combines two forms of sub-lethal exposure that results in mosquito movement away from a chemical source; contact excitation (increased activity upon contact) and non-contact spatial repellency\u0026nbsp;[77\u0026ndash;79]. These non-toxic behavioral impacts of pesticides were first identified in \u003cem\u003eAnopheles\u0026nbsp;\u003c/em\u003emosquitoes in response to DDT and later with insecticide-treated bed nets and indoor residual spraying\u0026nbsp;[80\u0026ndash;82]. Populations of \u003cem\u003eCx. quinquefasciatus,\u0026nbsp;\u003c/em\u003eanother species in the \u003cem\u003eCx. pipiens\u0026nbsp;\u003c/em\u003ecomplex, exhibit strong contact excitation and poor spatial repellency to pyrethroid, organophosphate, and carbamate pesticides\u0026nbsp;[83,84]. No study has investigated excito-repellency in \u003cem\u003eCx. tarsalis\u003c/em\u003e populations. The behavioral avoidance of \u003cem\u003eCulex\u003c/em\u003e to pesticides could be pushing mosquitoes out of the spray zones, resulting in an increase in abundance around the fringes of sprays, consistent with our estimates for spatial coverages \u0026lt; 40% for \u003cem\u003eCx. pipiens\u003c/em\u003e following sprays with pyrethrin or pyrethroids only in the previous week.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eA limitation of choosing a GAM framework is that we were only able to capture the average effects of covariates on nightly mosquito trap counts and unable to fully account for large stochastic fluctuations inherent to mosquito populations. However, GAMs easily allowed us to incorporate nonlinear relationships between abundance and covariates without having to constrain relationships with \u003cem\u003ea priori\u003c/em\u003e knowledge. In particular, we were able to capture the higher-order relationships and correlation across space and time with the three-dimensional spatio-temporal function. The form of the smooth relationships in the final model did appear to approximate what is observed in nature. Other strengths of our modeling approach are that it takes into account regional differences in mosquito populations, population dynamics and seasonality in different land use types, and the impacts of short-term (night) and longer-term (two week) weather, resulting in robust estimation of the baseline expected abundance in the absence of spray effects, allowing us to isolate the deviations in abundance due to aerial spraying. This counterfactual basis of the model enables estimation in the absence of an independent control. However, even with the large amount of data available, data were inadequate to estimate the impact of aerial spraying reliably for certain time lags or spatial coverages that were rare or absent in the data. This is primarily due to typical SYMVCD spraying and trapping practices due to logistical and financial constraints. SYMVCD concentrated their mosquito collections efforts near urban areas to maximize the sensitivity for assessing the risk in proximity to human populations while minimizing time and costs associated with large-scale mosquito surveillance. Additionally, in an effort to control mosquitoes in known problem areas (highly productive larval habitats in proximity to the margins of urban areas) and reduce aerial applications over urban areas, the majority of areas receiving repeated sprays across the WNV season are in more rural areas where the mosquito trapping is sparser. This limited the data and statistical power to quantify the full range of spatial overlap with aerial sprays.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe were also unable to account explicitly for drift outside the target zones during sprays, as has been previously described\u0026nbsp;[20]. Our use of a continuous variable to measure spray coverage within the 5km-radius collection areas surrounding each trap partially accounts for this effect. As such, however, we are unable to fully parse out the effect of aerial spraying on populations outside aerial spray zones and limited our analysis to only assess spatial coverage of traps within targeted spray zones.\u003c/p\u003e\n\u003cp\u003eAdditionally, we were unable to estimate the relative effects of different lengths of multi-night spray events (one vs two vs three consecutive nights) due to data limitations and our analytical choice to aggregate all sprays on the weekly scale to achieve the balance between robust estimates and operationally relevant information for vector control districts. These limitations of observational studies such as this one could be addressed in future experimental field trials.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eAerial adulticides were shown to achieve short-term reductions in the abundance of the primary West Nile virus vectors, \u003cem\u003eCx. tarsalis \u003c/em\u003eand \u003cem\u003eCx. pipiens\u003c/em\u003e. A greater reduction was estimated for \u003cem\u003eCx. pipiens\u003c/em\u003e, likely due to its focal distribution in urbanizes areas and limited dispersal. The use of organophosphate products versus a combination of pyrethrins and pyrethroids increased the magnitude of reduction estimated for \u003cem\u003eCx. pipiens\u003c/em\u003e while the difference by broad insecticide class was not significant for \u003cem\u003eCx. tarsalis\u003c/em\u003e. The effects of aerial sprays on \u003cem\u003eCx. tarsalis\u003c/em\u003e populations were likely moderated by the species broad dispersal ability, large population sizes, and vast expanses of productive larval habitat in the study area. Therefore, the best control of \u003cem\u003eCx. tarsalis \u003c/em\u003ewould be achieved in areas with isolated or highly spatially segmented populations. For both species, aerial spraying reduced abundance at high spatial coverage while reductions were also estimated at lower spatial coverage, at albeit greatly reduced magnitudes, indicating that aerial sprays had some impacts beyond the target zone. There was also evidence for population rebounds at periods of two to four weeks post-spraying. Our modeling approach allowed us to utilize observational data to isolate aerial treatment effects while taking into account contextual factors like spatio-temporal relationships, weather, and habitat that contribute to stochastic variation in nightly trap counts. This is an important advance that complements experimental trials and expands upon conventional observational approaches that summarize population changes following aerial treatments at individual time points. Further work should expand upon these methods to estimate the change in WNV transmission potential and resulting human infections following aerial spray events.\u003c/p\u003e"},{"header":"List Of Abbreviations","content":"\u003cp\u003eGAM: generalized additive model: SYMVCD: Sacramento-Yolo Mosquito and Vector Control District: WNV: West Nile virus\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe would like to thank Ruben Rosas, Marcia Reed, Sarah Wheeler, Samer Elkashef and Gary Goodman from Sacramento-Yolo Mosquito \u0026amp; Vector Control District for providing trapping and aerial treatment data used in this study and for their insight on mosquito control practices. We also thank the two anonymous reviewers for their helpful comments which improved the manuscript.\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data that support the findings of this study were obtained from the Sacramento-Yolo Mosquito and Vector Control District. These data were used with permission for the current study and are not publicly available. Data are however available from the authors upon reasonable request and with permission from the Sacramento-Yolo Mosquito and Vector Control District.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eKH acknowledges funding support from the Floyd \u0026amp; Mary Schwall Fellowship in Medical Research at UC Davis, and KH and CB acknowledge support from the Pacific Southwest Center of Excellence in Vector-Borne Diseases funded by the U.S. Centers for Disease Control and Prevention (Cooperative Agreement 1U01CK000516). KH acknowledges supported by the National Center for Advancing Translational Sciences, National Institutes of Health, through grant number UL1 TR001860 and linked award TL1 TR001861.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026rsquo; contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eKH performed the analysis and wrote the initial draft of the manuscript. CB conceived the project idea, supervised the research, and contributed to revisions of the manuscript. RR provided statistical guidance. All authors read and approved the final manuscript.\u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eHayes EB, Sejvar JJ, Zaki SR, Lanciotti RS, Bode A V, Campbell GL. Virology, pathology, and clinical manifestations of West Nile virus disease. Emerg Infect Dis. 2005;11(8):1174\u0026ndash;9.\u003c/li\u003e\n\u003cli\u003eMcLean RG, Ubico SR, Docherty DE, Hansen WR, Sileo L, McNamara TS. West Nile virus transmission and ecology in birds. Ann N Y Acad Sci. 2001;951:54\u0026ndash;7.\u003c/li\u003e\n\u003cli\u003eKilpatrick AM, LaDeau SL, Marra PP. Ecology of West Nile virus transmission and its impact on birds in the western hemisphere. Auk. 2007;124(4):1121\u0026ndash;36.\u003c/li\u003e\n\u003cli\u003eTurell MJ, Dohm DJ, Sardelis MR, O Guinn ML, Andreadis TG, Blow JA. An update on the potential of North American mosquitoes (Diptera : Culicidae) to transmit West Nile virus. J Med Entomol. 2005;42(1):57\u0026ndash;62.\u003c/li\u003e\n\u003cli\u003eKramer LD, Styer LM, Ebel GD. A global perspective on the epidemiology of West Nile virus. Annu Rev Entomol. 2007/07/25. 2008;53:61\u0026ndash;81.\u003c/li\u003e\n\u003cli\u003eGoddard LB, Roth AE, Reisen WK, Scott TW. Vector competence of California mosquitoes for West Nile virus. Emerg Infect Dis. 2002/12/25. 2002;8(12):1385\u0026ndash;91.\u003c/li\u003e\n\u003cli\u003eReisen WK, Lothrop HD, Chiles RE, Madon MB, Cossen C, Woods L, et al. West Nile virus in California. Emerg Infect Dis. 2004;10:1369\u0026ndash;78.\u003c/li\u003e\n\u003cli\u003eMostashari F, Bunning ML, Kitsutani PT, Singer DA, Nash D, Cooper MJ, et al. Epidemic West Nile encephalitis, New York, 1999: results of a household-based seroepidemiological survey. Lancet. 2001;358(9278):261\u0026ndash;4.\u003c/li\u003e\n\u003cli\u003eHughes JM, Wilson ME, Sejvar JJ. The long-term outcomes of human West Nile virus infection. Clin Infect Dis. 2007;44(12):1617\u0026ndash;24.\u003c/li\u003e\n\u003cli\u003eCalifornia Department of Public Health. Human West Nile virus activity, California, 2003-2019. [Internet]. 2019. Available from: http://westnile.ca.gov/\u003c/li\u003e\n\u003cli\u003eGubler DJ, Campbell GL, Nasci R, Komar N, Petersen L, Roehrig JT. West Nile virus in the United States: guidelines for detection, prevention, and control. Viral Immunol. 2001/02/24. 2000;13(4):469\u0026ndash;75.\u003c/li\u003e\n\u003cli\u003eCalifornia Department of Public Health, Mosquito and Vector Control Association of California, University of California. California mosquito-borne virus surveillance \u0026amp; response plan. 2019;(April). Available from: www.westnile.ca.gov\u003c/li\u003e\n\u003cli\u003eRose RI. Pesticides and public health: integrated methods of mosquito management. Emerg Infect Dis. 2001/03/27. 2001;7(1):17\u0026ndash;23.\u003c/li\u003e\n\u003cli\u003eCalifornia Department of Health Services Vector-Borne Disease Section. Overview of mosquito control practices in California. [Internet]. 2005. Available from: https://www.cdph.ca.gov/Programs/CID/DCDC/CDPH Document Library/OverviewofMosquitoControlinCA.pdf\u003c/li\u003e\n\u003cli\u003eSchleier JJ, Peterson RKD. Pyrethrins and pyrethroid insecticides. In: Lopez O, Fernfmdez-Bolafios JG, editors. Green Trends in Insect Control. Royal Society of Chemisty; 2011. p. 94\u0026ndash;131.\u003c/li\u003e\n\u003cli\u003eReisen WK. Using \u0026ldquo;Mulla\u0026rsquo;s Formula\u0026rdquo; to estimate percent control. In: Atkinson PW, editor. Vector Biology, Ecology, and Control. New York: Springer Science+Business Media B.V.; 2010. p. 127\u0026ndash;38.\u003c/li\u003e\n\u003cli\u003eMulla MS, Norland RL, Fanara DM, Darwezeh HA, McKeen DW. Control of chironomid midges in recreational lakes. J Econ Entomol. 1971;64:300\u0026ndash;7.\u003c/li\u003e\n\u003cli\u003eNielsen CF, Reisen WK, Armijos V, Wheeler S, Kelley K, Brown D. Impact of climate variation and adult mosquito control on the West Nile virus epidemic in Davis, California during 2006. Proc Pap Mosq Vector Control Assoc Calif. 2007;75:125\u0026ndash;30.\u003c/li\u003e\n\u003cli\u003eMacedo PA, Nielsen CF, Reed M, Kelley K, Reisen WK, Goodman GW, et al. An evaluation of the aerial spraying conducted in reponse to West Nile virus activity in Yolo county. Proc Pap Mosq Vector Control Assoc Calif. 2007;75:107\u0026ndash;14.\u003c/li\u003e\n\u003cli\u003eLothrop H, Lothrop B, Palmer M, Wheeler S, Gutierrez A, Lothrop H, et al. Evaluation of pyrethrin aerial ultra-low volume applications for adult \u003cem\u003eCulex tarsalis \u003c/em\u003econtrol in the desert environments of the Coachella Valley, Riverside county, California. J Am Mosq Control Assoc. 2007;23(4):405\u0026ndash;19.\u003c/li\u003e\n\u003cli\u003eElnaiem DA, Kelley K, Wright S, Laffey R, Yoshimura G, Reed M, et al. Impact of aerial spraying of pyrethrin insecticide on \u003cem\u003eCulex pipiens \u003c/em\u003eand\u003cem\u003e Culex tarsalis \u003c/em\u003e(Diptera: Culicidae) abundance and West Nile virus infection rates in an urban/suburban area of Sacramento County, California. J Med Entomol. 2008;45(4):751\u0026ndash;7.\u003c/li\u003e\n\u003cli\u003eReisen WK, Milby MM, Reeves WC, Eberle MW, Meyer RP, Schaefer CH, et al. Aerial adulticiding for the suppression of \u003cem\u003eCulex tarsalis\u003c/em\u003e in Kern County, California, using low volume propoxur: 2. Impact on natural populations in foothill and valley habitats. J Am Mosq Control Assoc. 1985;1(2):154\u0026ndash;63.\u003c/li\u003e\n\u003cli\u003eU. S. Census Bureau. State and County QuickFacts [Internet]. 2019 [cited 2019 Oct 14]. Available from: https://www.census.gov/quickfacts\u003c/li\u003e\n\u003cli\u003eUnited State Census Bureau. Percent urban and rural in 2010 by state and county. [Internet]. Available from: https://www.census.gov/geo/reference/ua/urban-rural-2010.html\u003c/li\u003e\n\u003cli\u003eCline G, Neigher A, Bellinder A. Climate of Sacramento, California. National Weather Service Office. Sacramento, California; 2010.\u003c/li\u003e\n\u003cli\u003eSacramento-Yolo Mosquito \u0026amp; Vector Control District. Mission \u0026amp; Vision [Internet]. Elk Grove, CA: Sacramento-Yolo Mosquito \u0026amp; Vector Control District; 2018 [cited 2019 Oct 5]. Available from: http://www.fightthebite.net/about/about-us/\u003c/li\u003e\n\u003cli\u003eNewhouse VF, Chamberlain RW, Johnson JG, Sudia WD. Use of dry ice to increase mosquito catches of the CDC miniature light trap. Mosq News. 1966;26(1):30\u0026ndash;5.\u003c/li\u003e\n\u003cli\u003eCalifornia Vectorborne Disease Surveillance System (CalSurv). [Internet]. 2018. Available from: https://vectorsurv.org/\u003c/li\u003e\n\u003cli\u003eR Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statisitcal Computing; 2020. Available from: https://www.r-project.org/\u003c/li\u003e\n\u003cli\u003eBivand R, Keitt T, Rowlingson B. rgdal: bindings for the \u0026ldquo;geospatial\u0026rdquo; data abstraction library. 2019. Available from: https://cran.r-project.org/package=rgdal\u003c/li\u003e\n\u003cli\u003ePRISM Climate Group, Oregon State University. Daily mean and monthly average temperature datasets. [Internet]. 2018. [Accessed 2018 May 12]. Available from: http://prism.oregonstate.edu\u003c/li\u003e\n\u003cli\u003eCiota AT, Matacchiero AC, Kilpatrick AM, Kramer LD. The effect of temperature on life history traits of \u003cem\u003eCulex\u003c/em\u003e mosquitoes. J Med Ent. 2014;51(1):55\u0026ndash;62.\u003c/li\u003e\n\u003cli\u003eReisen WK. Effect of temperature on \u003cem\u003eCulex tarsalis\u003c/em\u003e (Diptera: Culicidae) from the Coachella and San Joaquin valleys of California. J Med Entomol. 1995;32(5):636\u0026ndash;45.\u003c/li\u003e\n\u003cli\u003eMulti-resolution land characteristics consortium. NLCD Land Cover (CONUS) Database [Internet]. 2011 [cited 2018 May 20]. Available from: https://www.mrlc.gov/data?f%5B0%5D=category%3Aland cover\u0026amp;f%5B1%5D=year%3A2011\u003c/li\u003e\n\u003cli\u003eReisen WK, Reeves WC. Bionomics and ecology of \u003cem\u003eCulex tarsalis \u003c/em\u003eand other potential mosquito vector species. In: Reeves WC, editor. Epidemiology and control of mosquito-borne arboviruses in California, 1943-1987. Sacramento, CA: California Mosquito and Vector Control Association; 1990. p. 254\u0026ndash;329.\u003c/li\u003e\n\u003cli\u003eBailey SF, Eliason DA, Hoffmann BL. Flight and dispersal of mosquito \u003cem\u003eCulex tarsalis \u003c/em\u003eCoquillett in Sacramento Valley of California. Hilgardia. 1965;37(3):73\u0026ndash;113.\u003c/li\u003e\n\u003cli\u003eReisen WK, Milby MM, Meyer RP, Pfuntner AR, Spoehel J, Hazelrigg JE, et al. Mark-release-recapture studies with \u003cem\u003eCulex\u003c/em\u003e mosquitoes (Diptera: Culicidae) in southern California. J Med Entomol. 1991;28(3):357\u0026ndash;71.\u003c/li\u003e\n\u003cli\u003eDow RP, Reeves WC, Bellamy RE. Dispersal of female \u003cem\u003eCulex tarsalis \u003c/em\u003einto a larvicided area. Am J Trop Med Hyg. 1965;14(4):656\u0026ndash;70.\u003c/li\u003e\n\u003cli\u003eReisen WK, Milby MM, Meyer RP. Population dynamics of adult \u003cem\u003eCulex\u003c/em\u003e mosquitoes (Diptera: Culicidae) along the Kern River, Kern county, California, in 1990. J Med Ent. 1992;29(3):531\u0026ndash;43.\u003c/li\u003e\n\u003cli\u003eHastie T, Tibshirani R. Varying-coefficient models. J R Stat Soc Ser B-Methodological. 1993;55(4):757\u0026ndash;96.\u003c/li\u003e\n\u003cli\u003eHastie T, Tibshirani RJ. Generalized additive models. Stat Sci. 1986;1(3):297\u0026ndash;318.\u003c/li\u003e\n\u003cli\u003eWood SN. Thin plate regression splines. J R Stat Soc Ser B. 2003;65(1):95\u0026ndash;114.\u003c/li\u003e\n\u003cli\u003eWood SN. Generalized additive models: an introduction with R. Boca Raton, FL: Chapman \u0026amp; Hall/CRC; 2006.\u003c/li\u003e\n\u003cli\u003eWood SN. Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. J R Stat Soc Ser B-Statistical Methodol. 2011;73:3\u0026ndash;36.\u003c/li\u003e\n\u003cli\u003eAike HAI. A new look at the statistical model identification. IEEE Trans Automat Contr. 1974;19(6):716\u0026ndash;23.\u003c/li\u003e\n\u003cli\u003eWood SN. Fast stable direct fitting and smoothness selection for generalized additive models. J R Stat Soc Ser B-Statistical Methodol. 2008;70:495\u0026ndash;518.\u003c/li\u003e\n\u003cli\u003eHopkins CC, Hollinger FB, Johnson RF, Dewlett HJ, Newhouse VF, Chamberlain RW. The epidemiology of St. Louis encephalitis in Dallas, Texas, 1966. Am J Epidemiol. 1975;102(1):1\u0026ndash;15.\u003c/li\u003e\n\u003cli\u003eWekesa JW, Yuval B, Washino RK. Spatial distribution of adult mosquitoes (Diptera:Culicidae) in habitats associated with the rice agroecosystem of northern California. J Med Entomol. 1996;33(3):344\u0026ndash;50.\u003c/li\u003e\n\u003cli\u003eMitchell CJ, Kilpatrick JW, Hayes RO, Curry HW. Effects of ultra-low volume applications of malathion in Hale county, Texas II: mosquito populations in treated and untreated areas. J Med Ent. 1970;7(1):85\u0026ndash;91.\u003c/li\u003e\n\u003cli\u003eReisen WK, Yoshimura G, Reeves WC, Milby MM, Meyer RP. The impact of aerial applications of ultra-low volume adulticides on \u003cem\u003eCulex tarsalis\u003c/em\u003e populations (Diptera: Culicidae) in Kern County, California, USA, 1982. J Med Entomol. 1984;21(5):573\u0026ndash;85.\u003c/li\u003e\n\u003cli\u003eMutebi JP, Delorey MJ, Jones RC, Plate DK, Gerber SI, Gibbs KP, et al. The impact of adulticide applications on mosquito density in Chicago, 2005. J Am Mosq Control Assoc. 2011;27(1):69\u0026ndash;76.\u003c/li\u003e\n\u003cli\u003eClifton ME, Xamplas CP, Nasci RS, Harbison J. Gravid \u003cem\u003eCulex pipiens \u003c/em\u003eexhibit a reduced susceptibility to ultra-low volume adult control treatments under field conditions. J Am Mosq Control Assoc. 2019;35(4):267\u0026ndash;78.\u003c/li\u003e\n\u003cli\u003eReddy MR, Spielman A, Lepore TJ, Henley D, Kiszewski AE, Reiter P. Efficacy of resmethrin aerosols applied from the road for suppressing \u003cem\u003eCulex\u003c/em\u003e vectors of West Nile virus. Vector-Borne Zoonotic Dis. 2006;6(2):117\u0026ndash;27.\u003c/li\u003e\n\u003cli\u003eHolmes PR. A study of population changes in adult \u003cem\u003eCulex quinquefasciatus\u003c/em\u003e Say (Diptera: Culicidae) during a mosquito control programme in Dubai, United Arab Emirates. Ann Trop Med Parasitol. 1986;80(1):107\u0026ndash;16.\u003c/li\u003e\n\u003cli\u003eGreen RH. Sampling design and statistical methods for environmental biologists. Chichester, UK: Wiley Interscience; 1979.\u003c/li\u003e\n\u003cli\u003eUnderwood AJ. Beyond BACI: the detection of environmental impacts on populations in the real, but variable, world. J Exp Mar Bio Ecol. 1992;161:145\u0026ndash;78.\u003c/li\u003e\n\u003cli\u003eStewart-Oaten A, Murdoch WW, Parker KR. Environmental impact assessment: \u0026ldquo;pseudoreplication\u0026rdquo; in time? Ecology. 1986;67(4):929\u0026ndash;40.\u003c/li\u003e\n\u003cli\u003eStewart-Oaten A, Bence JR. Temporal and spatial variation in environmental impact assessment. Ecol Monogr. 2001;71(2):305\u0026ndash;39.\u003c/li\u003e\n\u003cli\u003eUnderwood AJ. On Beyond BACI: Sampling Designs that Might Reliably Detect Environmental Disturbances. Ecol Appl. 1994;4(1):3\u0026ndash;15.\u003c/li\u003e\n\u003cli\u003eRochlin I, Iwanejko T, Dempsey ME, Ninivaggi D V. Geostatistical evaluation of integrated marsh management impact on mosquito vectors using before-after-control-impact (BACI) design. Int J Health Geogr. 2009;8(1):1\u0026ndash;20.\u003c/li\u003e\n\u003cli\u003eWolfram G, Wenzl P, Jerrentrup H. A multi-year study following BACI design reveals no short-term impact of Bti on chironomids (Diptera) in a floodplain in Eastern Austria. Environ Monit Assess. 2018;190(12).\u003c/li\u003e\n\u003cli\u003eSmokorowski KE, Randall RG. Cautions on using the Before-After-Control-Impact design in environmental effects monitoring programs. Facets. 2017;2(1):212\u0026ndash;32.\u003c/li\u003e\n\u003cli\u003eStewart-Oaten A, Bence JR, Osenberg CW. Assessing effects of unreplicated perturbations: no simple solutions. Ecology. 1992;73(4):1396\u0026ndash;404.\u003c/li\u003e\n\u003cli\u003eLewis D. Causation. J Philos. 1973;70:556\u0026ndash;67.\u003c/li\u003e\n\u003cli\u003eRubin DB. Comment: Neyman (1923) and Causal Inference in Experiments and Observational Studies. Stat Sci. 1990;5.\u003c/li\u003e\n\u003cli\u003eGreenland S, Rothman KJ, Lash TL. Measures of effect and measures of association. In: Rothman KJ, Greenland S, Lash TL, editors. Modern Epidemiology, 3\u003csup\u003erd\u003c/sup\u003e edition. Lippincott, Williams and Wilkins Philadelphia; 2008. p. 51\u0026ndash;70.\u003c/li\u003e\n\u003cli\u003eReisen WK, Meyer RP, Shields J, Arbolante C. Population ecology of preimaginal \u003cem\u003eCulex tarsalis \u003c/em\u003e(Diptera: Culicidae) in Kern County, California. J Med Entomol. 1989/01/01. 1989;26(1):10\u0026ndash;22.\u003c/li\u003e\n\u003cli\u003eLoomis EC, Meyers EG. California encephalitis surveillance program: mosquito population measurement and ecologic considerations. Am J Epidemiol. 1960;71(3):378\u0026ndash;88.\u003c/li\u003e\n\u003cli\u003eBarker CM, Eldridge BF, Reisen WK. Seasonal abundance of \u003cem\u003eCulex tarsalis \u003c/em\u003eand\u003cem\u003e Culex pipiens \u003c/em\u003ecomplex mosquitoes (Diptera: Culicidae) in California. J Med Ent. 2010;47(5):759\u0026ndash;68.\u003c/li\u003e\n\u003cli\u003eLoomis EC, Eide RN, Caton JR, Merritt DA. Physical and cultural control: Mosquito and fly problems in dairy waste-water systems. Calif Agr. 1980;34(3):37\u0026ndash;8.\u003c/li\u003e\n\u003cli\u003eSu TY, Webb JR, Meyer RR, Mulla MS. Spatial and temporal distribution of mosquitoes in underground storm drain systems in Orange County, California. J Vector Ecol. 2003;28(1):79\u0026ndash;89.\u003c/li\u003e\n\u003cli\u003eReisen WK. The contrasting bionomics of \u003cem\u003eCulex\u003c/em\u003e mosquitoes in Western North America. J Am Mosq Control Assoc. 2012;28(4):82\u0026ndash;91.\u003c/li\u003e\n\u003cli\u003eCalifornia Department of Public Health Vector-Borne Disease Section. California mosquito pesticide resistance summary. [Internet]. 2005. Available from: http://westnile.ca.gov/downloads.php?download_id=3137\u0026amp;filename=Pesticide_Resistance.pdf\u003c/li\u003e\n\u003cli\u003eReed M, Macedo PA, Brown D. Increased tolerance to permethrin in\u003cem\u003e Culex pipiens\u003c/em\u003e complex population from Sacramento County, California. Proc Pap Mosq Vector Control Assoc Calif. 2012;80:56\u0026ndash;8.\u003c/li\u003e\n\u003cli\u003eReed M, Macedo PA, Goodman G, Brown D. Tough mosquitoes - why they should be everyone\u0026rsquo;s problem. Proc Pap Mosq Vector Control Assoc Calif. 2013;81:75\u0026ndash;9.\u003c/li\u003e\n\u003cli\u003eSorensen MA, Stevenson JA. An adulticide resistance \u0026ldquo;heat map\u0026rdquo; for \u003cem\u003eCulex pipiens \u003c/em\u003eand\u003cem\u003e Culex tarsalis\u003c/em\u003e (Diptera: Culididae) in Placer County. Proc Pap Mosq Vector Control Assoc Calif. 2015;83:8\u0026ndash;11.\u003c/li\u003e\n\u003cli\u003eDethier VG, Browne LB, Smith CW. The designation of chemicals in terms of the responses they elicit from insects. J Econ Entomol. 1960;53(1):134\u0026ndash;6.\u003c/li\u003e\n\u003cli\u003eRoberts DR, Chareonviriyaphap T, Harlan HH, P H. Methods ot testing and analyzing excito-repellency responses of malaria vectors to insecticides. J Am Mosq Control Assoc. 1997;3(1):13\u0026ndash;7.\u003c/li\u003e\n\u003cli\u003eMiller JR, Siegert PY, Amimo FA, Walker ED. Designation of chemicals in terms of the locomotor responses they elicit from insects: an update of Dethier et al. (1960). J Econ Entomol. 2009;102(6):2056\u0026ndash;60.\u003c/li\u003e\n\u003cli\u003eLindsay SW, Adiamah JH, Miller JE, Armstrong JRM. Pyrethroid-treated bednet effects on mosquitoes of the A\u003cem\u003enopheles gambiae \u003c/em\u003ecomplex in The Gambia. Med Vet Entomol. 1991;5:477\u0026ndash;83.\u003c/li\u003e\n\u003cli\u003eKennedy JS. The excitant and repellent effects on mosquitos of sub-lethal contacts with DDT. Bull Entomol Res. 1945;37(4):593\u0026ndash;607.\u003c/li\u003e\n\u003cli\u003eDavidson G. Experiments on the effect of residual insecticides in houses against \u003cem\u003eAnopheles gambiae \u003c/em\u003eand\u003cem\u003e A. funestus\u003c/em\u003e. Bull Entomol Res. 1953;44:231\u0026ndash;54.\u003c/li\u003e\n\u003cli\u003eBoonyuan W, Bangs MJ, Grieco JP, Tiawsirisup S, Prabaripai A, Chareonviriyaphap T. Excito-repellent responses between \u003cem\u003eCulex quinquefasciatus\u003c/em\u003e permethrin susceptible and resistant mosquitoes. J Insect Behav. 2016;29:415\u0026ndash;31.\u003c/li\u003e\n\u003cli\u003eSathantriphop S, Ketavan C, Prabaripai A, Visetson S, Bangs MJ, Akratanakul P, et al. Susceptibility and avoidance behavior by\u003cem\u003e Culex quinquefasciatus\u003c/em\u003e Say to three classes of residual insecticides. J Vector Ecol. 2006;31(2):266\u0026ndash;74.\u003c/li\u003e\n\u003cli\u003eMacedo PA, Schleier JJ, Reed M, Kelley K, Goodman GW, Brown DA, et al. Evaulation of efficacy and human health risk of aerial ultra-low volume applications of pyrethrins and piperonyl butoxide for adult mosquito management in response to West Nile virus activity in Sacramento county, California. J Am Mosq Control Assoc. 2010;26(1):57\u0026ndash;66.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cstrong\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003eTable 1\u003c/span\u003e\u003c/strong\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003e\u0026nbsp;Estimated change in \u003cem\u003eCulex\u0026nbsp;\u003c/em\u003emosquito populations following aerial spray events in California using Mulla\u0026rsquo;s formula.\u003c/span\u003e\u003c/p\u003e\n\u003ctable style=\"width:100.06%;margin-left:-.25pt;border-collapse:collapse;border:none;\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width:15.06%;border-top:solid windowtext 1.0pt;border-left:none;border-bottom:solid windowtext 1.0pt;border-right:none;padding:0in 5.4pt 0in 5.4pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003eLocation\u003c/span\u003e\u003c/p\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003eCity, county\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:8.3%;border-top:solid windowtext 1.0pt;border-left:none;border-bottom:solid windowtext 1.0pt;border-right:none;padding:0in 5.4pt 0in 5.4pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003eYear\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:11.58%;border-top:solid windowtext 1.0pt;border-left:none;border-bottom:solid windowtext 1.0pt;border-right:none;padding:0in 5.4pt 0in 5.4pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003eNights Sprayed\u003csup\u003e#\u003c/sup\u003e\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:13.0%;border-top:solid windowtext 1.0pt;border-left:none;border-bottom:solid windowtext 1.0pt;border-right:none;padding:0in 5.4pt 0in 5.4pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003eProduct Class\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:15.02%;border-top:solid windowtext 1.0pt;border-left:none;border-bottom:solid windowtext 1.0pt;border-right:none;padding:0in 5.4pt 0in 5.4pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003eComparison Length\u003csup\u003e\u0026dagger;\u003c/sup\u003e\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:11.94%;border-top:solid windowtext 1.0pt;border-left:none;border-bottom:solid windowtext 1.0pt;border-right:none;padding:0in 5.4pt 0in 5.4pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003eSpecies\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:12.4%;border-top:solid windowtext 1.0pt;border-left:none;border-bottom:solid windowtext 1.0pt;border-right:none;padding:0in 5.4pt 0in 5.4pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e% Change\u003csup\u003e\u0026Dagger;\u003c/sup\u003e\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:12.7%;border-top:solid windowtext 1.0pt;border-left:none;border-bottom:solid windowtext 1.0pt;border-right:none;padding:0in 5.4pt 0in 5.4pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003eReference\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" style=\"width:15.06%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003eDavis, Yolo\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width:8.3%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e2006\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width:11.58%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e2\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width:13.0%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003epyrethrin\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width:15.02%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e2 days before, 2 days after\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:11.94%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cem\u003e\u003cspan style='font-size:16px;line-height: 150%;font-family:\"Times New Roman\",serif;'\u003eCx. pipiens\u003c/span\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:12.4%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e- 58.0\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"4\" style=\"width:12.7%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003e[19]\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width:11.94%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cem\u003e\u003cspan style='font-size:16px;line-height: 150%;font-family:\"Times New Roman\",serif;'\u003eCx. tarsalis\u003c/span\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:12.4%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e- 25.6\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" style=\"width:15.06%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003eWoodland, Yolo\u0026nbsp;\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width:8.3%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e2006\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width:11.58%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e2\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width:13.0%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003epyrethrin\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width:15.02%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e2 days before, 2 days after\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:11.94%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cem\u003e\u003cspan style='font-size:16px;line-height: 150%;font-family:\"Times New Roman\",serif;'\u003eCx. pipiens\u003c/span\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:12.4%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e- 77.7\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width:11.94%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cem\u003e\u003cspan style='font-size:16px;line-height: 150%;font-family:\"Times New Roman\",serif;'\u003eCx. tarsalis\u003c/span\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:12.4%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.6pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e- 46.8\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" style=\"width:15.06%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003eSacramento, Sacramento\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width:8.3%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e2005\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width:11.58%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e3\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width:13.0%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003epyrethrin\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width:15.02%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e7 days before, 7 days after\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:11.94%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cem\u003e\u003cspan style='font-size:16px;line-height: 150%;font-family:\"Times New Roman\",serif;'\u003eCx. pipiens\u003c/span\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:12.4%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e- 75.0\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width:12.7%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003e[21]\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width:11.94%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cem\u003e\u003cspan style='font-size:16px;line-height: 150%;font-family:\"Times New Roman\",serif;'\u003eCx. tarsalis\u003c/span\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:12.4%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e- 48.7\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" style=\"width:15.06%;border:none;padding:0in 5.4pt 0in 5.4pt;height:13.45pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003eSacramento, Sacramento\u0026nbsp;\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width:8.3%;border:none;padding:0in 5.4pt 0in 5.4pt;height:13.45pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e2006\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width:11.58%;border:none;padding:0in 5.4pt 0in 5.4pt;height:13.45pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e3\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width:13.0%;border:none;padding:0in 5.4pt 0in 5.4pt;height:13.45pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003epyrethrin\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width:15.02%;border:none;padding:0in 5.4pt 0in 5.4pt;height:13.45pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e3 days before, 3 days after\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:11.94%;border:none;padding:0in 5.4pt 0in 5.4pt;height:13.45pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cem\u003e\u003cspan style='font-size:16px;line-height: 150%;font-family:\"Times New Roman\",serif;'\u003eCx. pipiens\u003c/span\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:12.4%;border:none;padding:0in 5.4pt 0in 5.4pt;height:13.45pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e- 39.3\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width:12.7%;border:none;padding:0in 5.4pt 0in 5.4pt;height:13.45pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003e[82]\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width:11.94%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cem\u003e\u003cspan style='font-size:16px;line-height: 150%;font-family:\"Times New Roman\",serif;'\u003eCx. tarsalis\u003c/span\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:12.4%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e- 57.3\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" style=\"width:15.06%;border:none;border-bottom:solid windowtext 1.0pt;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003eCoachella valley, Riverside\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:8.3%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e2005 (Mar)\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:11.58%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e3 alt^\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:13.0%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003epyrethroid\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:15.02%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e5 days before, 1 day after\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:11.94%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cem\u003e\u003cspan style='font-size:16px;line-height: 150%;font-family:\"Times New Roman\",serif;'\u003eCx. tarsalis\u003c/span\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:12.4%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e- 93.0\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"3\" style=\"width:12.7%;border:none;border-bottom:solid windowtext 1.0pt;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003e[20]\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width:8.3%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e2005 (Jun)\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:11.58%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e3 alt^\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:13.0%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003epyrethroid\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:15.02%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e5 days before, 1 day after\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:11.94%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cem\u003e\u003cspan style='font-size:16px;line-height: 150%;font-family:\"Times New Roman\",serif;'\u003eCx. tarsalis\u003c/span\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:12.4%;border:none;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e- 77.0\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width:8.3%;border:none;border-bottom:solid windowtext 1.0pt;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e2005 (Sep)\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:11.58%;border:none;border-bottom:solid windowtext 1.0pt;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e3 alt^\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:13.0%;border:none;border-bottom:solid windowtext 1.0pt;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003epyrethroid\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:15.02%;border:none;border-bottom:solid windowtext 1.0pt;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e5 days before, 1 day after\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:11.94%;border:none;border-bottom:solid windowtext 1.0pt;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cem\u003e\u003cspan style='font-size:16px;line-height: 150%;font-family:\"Times New Roman\",serif;'\u003eCx. tarsalis\u003c/span\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:12.4%;border:none;border-bottom:solid windowtext 1.0pt;padding:0in 5.4pt 0in 5.4pt;height:12.55pt;\"\u003e\n \u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;text-align:center;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family: \"Times New Roman\",serif;'\u003e73.0\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003csup\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003e#\u003c/span\u003e\u003c/sup\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003e\u0026nbsp;Number of consecutive nights sprayed in spray event.\u003c/span\u003e\u003c/p\u003e\n\u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003csup\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003e\u0026dagger;\u0026nbsp;\u003c/span\u003e\u003c/sup\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003eLength of time before and after an aerial spray event used when comparing of trap counts.\u003c/span\u003e\u003c/p\u003e\n\u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003csup\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003e\u0026Dagger;\u0026nbsp;\u003c/span\u003e\u003c/sup\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003ePercent change in Culex abundance following an aerial spray event as calculated using Mulla\u0026rsquo;s formula\u0026nbsp;\u003c/span\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003e[16,17]\u003c/span\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003e.\u003c/span\u003e\u003c/p\u003e\n\u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003e^ Aerial spraying occurred on 3 alternate nights.\u003c/span\u003e\u003c/p\u003e\n\u003cp style='margin-top:0in;margin-right:0in;margin-bottom:0in;margin-left:0in;line-height:150%;font-size:15px;font-family:\"Calibri\",sans-serif;'\u003e\u003cstrong\u003e\u003cspan style='font-size:16px;line-height:150%;font-family:\"Times New Roman\",serif;'\u003e\u0026nbsp;\u003c/span\u003e\u003c/strong\u003e\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"parasites-and-vectors","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"parv","sideBox":"Learn more about [Parasites \u0026 Vectors](http://parasitesandvectors.biomedcentral.com/)","snPcode":"13071","submissionUrl":"https://submission.nature.com/new-submission/13071/3","title":"Parasites \u0026 Vectors","twitterHandle":"@bugbittentweets","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Adulticide, Aerial spraying, Culex tarsalis, Culex pipiens, GAM, Generalized additive models, Spatial-temporal model, Mosquitoes, Mosquito-borne disease, West Nile virus","lastPublishedDoi":"10.21203/rs.3.rs-92266/v2","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-92266/v2","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground:\u003c/strong\u003e Aerial applications of insecticides that target adult mosquitoes are widely used to reduce transmission of West Nile virus to humans during periods of epidemic risk. However, estimates of the reduction in abundance following these treatments typically focus on single events, rely on pre-defined, untreated control sites, and can vary widely due to stochastic variation in population dynamics and trapping success unrelated to the treatment.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e To overcome these limitations, we developed generalized additive models fitted to mosquito surveillance data from CO\u003csub\u003e2\u003c/sub\u003e-baited traps in Sacramento and Yolo counties, California from 2006-2017. The models accounted for the expected spatial and temporal trends in the abundance of adult female \u003cem\u003eCulex tarsalis\u003c/em\u003e and \u003cem\u003eCulex pipiens \u003c/em\u003ein the absence of aerial spraying. Estimates for the magnitude of deviation from baseline abundance following aerial spray events were obtained from the models.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e One-week post-treatment with full spatial coverage of the trapping area by pyrethroid or pyrethrin products, \u003cem\u003eCx. pipiens\u003c/em\u003e abundance was reduced by a mean of 52.4% (95% CI: -65.6, -36.5%) while the use of at least one organophosphate pesticide resulted in a 76.2% (95% CI: -82.8, -67.9%) reduction. For \u003cem\u003eCx. tarsalis\u003c/em\u003e one-week post-treatment with full coverage resulted in a 30.7% (95% CI: -54.5, 2.5%) reduction; pesticide class was not a significant factor contributing to reduction. In comparison, repetition of spraying over three to four consecutive weeks resulted in similar estimates for \u003cem\u003eCx. pipiens\u003c/em\u003e and a somewhat smaller magnitude for \u003cem\u003eCx. tarsalis.\u003c/em\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eConclusions:\u003c/strong\u003e Aerial adulticides are effective for rapid short-term reduction of the abundance of the primary West Nile virus vectors, \u003cem\u003eCx. tarsalis \u003c/em\u003eand \u003cem\u003eCx. pipiens\u003c/em\u003e. A larger magnitude of reduction is estimated in \u003cem\u003eCx. pipiens\u003c/em\u003e, possibly due to the species’ focal distribution. Effects of aerial sprays on \u003cem\u003eCx. tarsalis\u003c/em\u003e populations are likely modulated by the species’ large dispersal ability, population sizes, and vast productive larval habitat present in the study area. Our modeling approach provides a new way to estimate effects of public-health pesticides on vector populations using routinely collected observational data and accounting for spatio-temporal trends and contextual factors like weather and habitat. It does not require pre-selected control sites and expands upon past studies that have focused on effects of individual aerial treatment events.\u003c/p\u003e","manuscriptTitle":"Spatio-temporal impacts of aerial adulticide applications on populations of West Nile virus vector mosquitoes","msid":"","msnumber":"","nonDraftVersions":[{"code":2,"date":"2021-01-28 12:30:45","doi":"10.21203/rs.3.rs-92266/v2","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Accept","date":"2021-01-28T00:00:00+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2021-01-24T00:00:00+00:00","index":2,"fulltext":"Recommendation: Reviewer's comments unavailable due to the journal's policy.\n"},{"type":"reviewerAgreed","content":"","date":"2021-01-20T00:00:00+00:00","index":2,"fulltext":""},{"type":"reviewerAgreed","content":"","date":"2021-01-16T00:00:00+00:00","index":1,"fulltext":""},{"type":"editorInvitedReview","content":"","date":"2021-01-16T00:00:00+00:00","index":1,"fulltext":"Recommendation: Reviewer's comments unavailable due to the journal's policy.\n"},{"type":"editorAssigned","content":"","date":"2021-01-13T00:00:00+00:00","index":"","fulltext":""},{"type":"reviewersInvited","content":"","date":"2021-01-13T00:00:00+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2021-01-12T23:00:00+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2021-01-12T23:00:00+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"parasites-and-vectors","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"parv","sideBox":"Learn more about [Parasites \u0026 Vectors](http://parasitesandvectors.biomedcentral.com/)","snPcode":"13071","submissionUrl":"https://submission.nature.com/new-submission/13071/3","title":"Parasites \u0026 Vectors","twitterHandle":"@bugbittentweets","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}},{"code":1,"date":"2020-10-19 14:31:10","doi":"10.21203/rs.3.rs-92266/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Major Revision","date":"2020-11-19T00:00:00+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2020-11-04T00:00:00+00:00","index":1,"fulltext":"Recommendation: Reviewer's comments unavailable due to the journal's policy.\n"},{"type":"editorInvitedReview","content":"","date":"2020-11-02T00:00:00+00:00","index":2,"fulltext":"Recommendation: Reviewer's comments unavailable due to the journal's policy.\n"},{"type":"reviewerAgreed","content":"","date":"2020-10-22T12:00:00+00:00","index":3,"fulltext":""},{"type":"reviewerAgreed","content":"","date":"2020-10-18T12:00:00+00:00","index":2,"fulltext":""},{"type":"reviewerAgreed","content":"","date":"2020-10-16T12:00:00+00:00","index":1,"fulltext":""},{"type":"reviewersInvited","content":"","date":"2020-10-15T12:00:00+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2020-10-13T12:00:00+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2020-10-13T12:00:00+00:00","index":"","fulltext":""},{"type":"submitted","content":"","date":"2020-10-12T12:00:00+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2020-10-12T12:00:00+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"parasites-and-vectors","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"parv","sideBox":"Learn more about [Parasites \u0026 Vectors](http://parasitesandvectors.biomedcentral.com/)","snPcode":"13071","submissionUrl":"https://submission.nature.com/new-submission/13071/3","title":"Parasites \u0026 Vectors","twitterHandle":"@bugbittentweets","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"311a09cd-4a84-42b5-ad44-4cbe7e75ada6","owner":[],"postedDate":"January 28th, 2021","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":820664,"name":"Parasitology"}],"tags":[],"updatedAt":"2021-02-28T15:00:51+00:00","versionOfRecord":{"articleIdentity":"rs-92266","link":"https://doi.org/10.1186/s13071-021-04616-6","journal":{"identity":"parasites-and-vectors","isVorOnly":false,"title":"Parasites \u0026 Vectors"},"publishedOn":"2021-02-24 15:00:32","publishedOnDateReadable":"February 24th, 2021"},"versionCreatedAt":"2021-01-28 12:30:45","video":"","vorDoi":"10.1186/s13071-021-04616-6","vorDoiUrl":"https://doi.org/10.1186/s13071-021-04616-6","workflowStages":[]},"version":"v2","identity":"rs-92266","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-92266","identity":"rs-92266","version":["v2"]},"buildId":"FbvkV6FR0MCFSLy54lSbu","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}