Spatiotemporal analysis of West Nile virus infection in the human population based on arboviral detection testing of blood donations

Spatiotemporal analysis of West Nile virus infection in the human population based on arboviral detection testing of blood donations | 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 Article Spatiotemporal analysis of West Nile virus infection in the human population based on arboviral detection testing of blood donations Benoit Talbot, Antoinette Ludwig, Sheila F. O’Brien, Steven J. Drews, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4714418/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 28 Dec, 2024 Read the published version in Scientific Reports → Version 1 posted 16 You are reading this latest preprint version Abstract West Nile virus (WNV) is a mosquito-borne zoonotic flavivirus which often causes asymptomatic infection in humans but may develop into a deadly neuroinvasive disease. In this study, we aimed to investigate variables potentially associated with human WNV infection using human and mosquito WNV surveillance and monitoring datasets, established over 20 years, from 2003 to 2022, across the province of Ontario, Canada. We combined climatic and geographic data, mosquito surveillance data (n=3,010 sites), blood donation arboviral detection testing data in the human population, and demographic and socio-economic data from Canadian population censuses. We hypothesized that spatio-temporal indices related to mosquito vector habitat and phenology, in addition to human demographic and socio-economic factors, were associated with WNV infection in the human population. Our results show that habitat suitability of the main WNV vector in this region, Cx. pipiens/restuans (IRR = 2.0), and variables related to lower income (IRR = 2.8), and shelter infrastructure spending (IRR = 0.7), were key risk factors associated with WNV infection among blood donors from 2003 to 2022 across Ontario (R 2 = 0.67). These results may inform points of entry for practical intervention aimed at reducing risk of mosquito-borne pathogens in Canada. Biological sciences/Ecology/Ecological epidemiology Health sciences/Diseases/Infectious diseases/Viral infection Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Human infections with West Nile virus (WNV; family Flaviviridae) are most frequently mild or asymptomatic. Symptomatic infections result in an illness referred to as ‘West Nile fever’, which, in rare cases, particularly in older age groups, may develop into severe and sometimes fatal neuroinvasive disease 1 . The virus, first introduced to North America in 1999, rapidly spread across the continent in the following years 2,3 , and is now considered an endemic mosquito-borne pathogen in much of North America 4,5 . In Canada, detection tests for WNV have been regularly conducted on most blood donations since 2003, to avoid transmission from donors to recipients 6 . While the blood donor population isn’t absolutely synonymous with the general population, data from WNV detection tests on blood donations offer a representative snapshot of spatio-temporal variations in WNV infection risk in the human population 7 . Human case reporting in surveillance is also a useful source of information on WNV risk in North America 8 , despite having some limitations, including underreporting due to asymptomatic infections, misdiagnoses, and uncertainty as to the precise size of the exposed population 9,10 . While such limitations may be avoided by using active pathogen surveillance data on in the human population 11–13 , data spanning a large area and timeframe are scarce. Mosquito species of the genus Culex , such as Culex pipiens and Culex restuans , are recognized as important vectors maintaining the enzootic cycle of WNV among avian hosts in northeastern North America. Culex pipiens and Cx. restuans also occasionally feed on mammals, bridging transmission between birds and humans 14–19 . Aedes vexans is very common in northeastern North America, is competent for the transmission of WNV and is highly opportunistic in its blood feeding preferences, meaning they can also effectively bridge transmission between birds and humans 20–23 . Environmental factors that influence mosquito development and virus transmission are important for understanding the drivers of human WNV infection risk. Larval development time for Cx. pipiens and Cx. restuans mosquitoes is estimated to be 19 and 14 days on average, respectively, while their adult lifespan is approximately 30 and 27 days, respectively 24 . Development time for Ae. vexans is 10 days on average 25 , and lifespan is 20 days on average 23 . WNV infection can be detected using nucleic acid tests to detect viral RNA by polymerase chain reaction (PCR) during viremia from around Day 1 to Day 13 after bite by an infected mosquito, and using serological tests to detect IgM or IgG by enzyme-linked immunosorbent assay (or ELISA) from around Day 5 to 9 (respectively for IgM and IgG) after bite by an infected mosquito, for approximately 5 to 7 months on average (respectively for IgM and IgG) 26 . Exposure of humans, including blood donors, to infected mosquitoes depends on their abundance and infection prevalence. Mosquito abundance varies seasonally and from year-to-year according to longer term impacts of weather and climate affecting survival over winter and shorter-term effects of weather on mosquito reproduction and activity, and WNV replication in infected mosquitoes 27–32 . Inter-annual variations in cycles of transmission amongst avian reservoirs, and mosquitoes changing their behaviour to include mammals as sources of blood meals also impact seasonal and inter-annual variations 33 . Together, this means that weather over a period of weeks to months, combined with local environmental impacts on mosquito reproduction and WNV transmission, determine the risk of human infections, and of WNV outbreaks, each year. Current effects of climate change and land use change in Canada and elsewhere may contribute to an increase in the risk of zoonotic diseases, including mosquito-borne pathogens, such as WNV 4,34–40 . Therefore, there is a pressing need for an enhanced understanding of the environmental hazard posed by WNV for public health. Comprehensive temporal and spatial databases of a large variety of climatic, geographic and ecological factors are needed to correctly characterize the distribution and dynamics of mosquito vector populations and mosquito-borne disease transmission. Additionally, certain demographic and socio-economic factors may contribute to both mosquito vector density and exposure to mosquito-borne disease risk 41–47 . Indeed, lower socio-economic status seem to be strongly associated with mosquito-borne disease risk, through variation in population density 48 , quality of household infrastructure 44 , and education, potentially associated with level of risk perception 46 . Occupation 49 , age and sex 50 may also influence exposure to mosquitoes. Therefore, regularly updated population censuses offer invaluable information to combine with other data sources to identify modifiable risk factors and better target interventions to reduce mosquito-borne disease risk. In this study, we aimed to investigate variables potentially associated with human WNV exposure and infection using a large data collection, established over 20 years, from 2003 to 2022, across most of the province of Ontario, Canada. We combined climatic data, geographic data, mosquito surveillance data, arbovirus testing data from human blood donations, and demographic and socio-economic data from Canadian population censuses. We hypothesized that a mixture of spatial and temporal variables, impacting mosquito habitat and phenology, and demographic and socio-economic variables, which are associated with increased mosquito habitat and also increased exposure to mosquitoes, are the main predictors of WNV infection in the human population. Results Study area and duration The study area in Ontario, Canada spanned more than 428,000 km 2 with a northern limit (latitude ~ 50–51° N) that matched the extent of most datasets used in this study (Fig. 1). Major urban population centers in our study area are situated in the southeastern section, where a warmer agricultural and residential landscape predominates, and which represents the most populated corridor in Canada from the cities of Windsor to Ottawa. The northern section is dominated by a mixture of forests, wild grasslands and wetlands, with colder temperatures, and ranging from dry in the west at the border with the province of Manitoba, to humid in the center and towards the border with the province of Québec in the east (Fig. 1). Mosquito collection data Based on the available mosquito surveillance data from all 34 Ontario public health units, sampling sites (n = 3,010) were visited at least once, and at most 876 times, from 2003 to 2022. On average, each site was visited 48 times across the study period (Table 1). Sampling sites were visited mostly from May to October when both Cx. pipiens/restuans and Ae. vexans were most likely to be observed in the study area. Out of a total of 145,102 sampling site visits, 99.9% were made from May to October. About twice as many Ae. vexans than Cx. pipiens/restuans individuals were collected, in total and on average per sampling site visit. Around 60% of sampling visits yielded each species on average (Table 1). Average abundance and occurrence per sampling site visit generally decreased across years for Ae. vexans , but not for Cx. pipiens/restuans (Fig. 2). Average abundance and occurrence per sampling site visit increased from May to August, and then decreased in the following months to October (Fig. 2). The number of mosquito pools tested for West Nile virus (WNV) was higher for Cx. pipiens/restuans due to higher priority of Cx. pipiens/restuans for WNV testing, and the total number of positive pools was also higher for Cx. pipiens/restuans (Table 1). However, the total number of Cx. pipiens/restuans individuals tested was lower, which is due to fewer collected Cx. pipiens/restuans individuals for that species on average compared to Ae. vexans , and minimum infection rate (MIR) was much higher for Cx. pipiens/restuans (Table 1). There was an overall higher MIR infection rate from 2013 to 2022 (1.9), compared to 2003 to 2012 (1.3), where 2012 was the year with the highest MIR (7.9; Fig. 2). The MIR was also much higher in August and September, compared to July and October (Fig. 2). No WNV-positive mosquito pools were observed from November to May of any year. Land cover and climatic data Land cover data collated from Agriculture Canada and United States Geological Survey consisted of a total of 10 classes, namely two classes of humid cover (open water and wetlands), three classes of open cover (natural and anthropogenic vegetated, and barren), one class of forested cover, and four classes of residential cover (containing varying degrees of vegetation cover), at a resolution of 120 meters. The predominance of each class varied considerably across the study area (Fig. 1). The four most common landscape classes containing a mosquito sampling site were the medium green residential, high green residential, agricultural cropland and forested classes, and the three least common were non-vegetated residential, exposed, and wetlands (Table S1 ; Fig. 1). The 20-year average total daily precipitation varied between 1.7 to 4.0 mm across the study area. The western part of the study area received less precipitation on average, while a few areas in the center and center-east received more precipitation on average (Fig. 1). The 20-year average mean daily temperature varied between 0.7 and 10.4°C, and the southern part of the study area was much warmer on average than the northern part, with the warmest area being the corridor from Windsor to Toronto in the extreme south near Lake Erie, and the coldest area being the center north and northeast of the study area (Fig. 1). Average total precipitation and mean temperature at cells containing a sampling site were 2.6 mm and 7.5°C, respectively (Table S1 ). Ecological niche modeling analysis The correlation coefficient of 20-year average total daily precipitation and 20-year average mean daily temperature was low ( r < 0.1). Overall, the receiver operating characteristic’s area under the curve (AUC) values were lower for Ae. vexans models compared to Cx. pipiens/restuans and West Nile virus models. Out of 100 models, the best performing model had an AUC = 0.82 for Cx. pipiens/restuans , 0.79 for West Nile virus, and 0.72 for Ae. vexans (Table 2). The ensemble model contained 100, 100 and 8 models, respectively for WNV, Cx. pipiens/restuans , and Ae. vexans , with a mean AUC higher than 0.9 for mean habitat suitability index (HSI) values and committee averaging for all species (Table 2). Explanatory variable importance (on a scale of 0 to 1, obtained from 100 permutations of the ensemble niche model) of temperature in the ensemble model was higher than 0.6 and higher than the importance of the other two explanatory variables for both mosquito species and WNV. The importance of precipitation was nearly the same as that of temperature for Ae. vexans , but was low, at less than 0.2, for Cx. pipiens/restuans and moderate, at less than 0.4, for West Nile virus. Comparatively, land cover was of moderate importance for all species, at less than 0.4 (Table 2). Projected HSI of the best performing model for all species led to an area of highest HSI value (100, 924 and 883, for WNV, Ae. vexans and Cx. pipiens/restuans , respectively, with highest possible maximum value of 1000) concentrated in the extreme southern part of the study area for all species. This was centered around the corridor from Windsor to Toronto near Lake Erie (Fig. 1) for both mosquito species, and was confined to major urban centers for WNV. Another area of high HSI was present for both mosquito species in the southeast around the cities of Kingston and Ottawa near Lake Ontario, where HSI was overall higher for Ae. vexans than for Cx. pipiens/restuans . Ae. vexans also displayed high HSI in areas where Cx. pipiens/restuans did not, particularly in the center and center-east of the study area north of Lake Huron, and in several areas in the extreme west of the study area west of Lake Superior (Fig. 1). Mean HSI and committee averaging showed high suitability of wetlands and two residential land cover classes, low green and medium green, for both mosquito species and WNV (Fig. 3). Wild grasslands had high suitability for the two mosquito species, but low for WNV. High green residential showed high suitability only for Ae. vexans . Open water, forested, and exposed land cover classes showed low suitability for WNV (Fig. 3). Other land cover classes either showed low confidence, as displayed by values near 0.5, or conflict between mean HSI and committee averaging values (Fig. 3). Mid-range values of total daily precipitation, i.e. between roughly 2.5 and 3.0 mm, were associated with the highest suitability values for both mosquito species, and lowest values, below 2.5 mm, were associated with the highest suitability values for WNV, according to both mean HSI and committee averaging (Fig. 3). Higher mean daily temperature values, i.e. above roughly 6°C for both mosquito species and 9°C for WNV, were associated with the highest suitability values, according to both mean HSI and committee averaging (Fig. 3). Socio-economic and demographic data Using Statistics Canada census datasets from 2016, we identified a total of 14 demographic and socio-economic variables of interest in our analyses, at the level of the census subdivision. One variable captured information on the sex ratio (percent male residents); three variables on age structure (percent residents younger than 15, percent residents older than 64, and percent residents older than 84); two variables on ethnicity (percent residents self-identified as immigrant, percent residents self-identified as Indigenous); one variable on population density (number of residents per km 2 ); two variables on income (mean income, and percent of population earning less than $ 20,000), two variables on shelter (percent shelters needing major repairs, percent residents spending 30% or more of their income on shelter); one variable on education (percent residents with no secondary education); and two variables on occupation (percent residents working in trades, and percent residents working with natural resources). Blood donation arboviral testing data A total of more than 6.5 million blood donations from more than 900,000 donors were tested by Canadian Blood Services across the province of Ontario, Canada, from 2003 to 2022 (Table 3). Many donations were coming from the same donor multiple times throughout the study period. We hereby refer use ‘donor’ to refer to each unique individual who donated blood either once or multiple times, and the term ‘donation’ to reflect each individual donation given. There were more donations from male than female donors, and more donations from donors aged 39 to 65 compared to younger and older (Table 3). However, there were slightly more individual female donors than males, and slightly more individual donors younger than 39 compared to 39 and older. Donors aged older than 65 represented the lowest number of individual donors, among all age groups (Table 3). There was a total of 102 donations with a positive test result for WNV infection based on nucleic acid testing. The bulk of WNV-positive blood donations were from male donors and from donors aged 39 to 65. The cumulative WNV infection rate was slightly higher for male than female donors, and much higher for donors older than 65 than for other age groups (Table 3). The WNV infection rate was also much higher in some census subdivisions and some years compared to others. There was an overall higher WNV infection rate in southern Ontario (Fig. 4). There was also an overall higher WNV infection rate from 2013 to 2022 (2.7 per 100,000), compared to 2003 to 2012 (0.8 per 100,000), where 2018 was the year with the highest WNV infection rate (10.6 per 100,000; Fig. 2). The incidence of WNV infection amongst blood donors was also much higher in August and September, compared to July and October (Fig. 2). No positive WNV blood sample was observed from November to June of any year. Spatiotemporal WNV infection analyses At the analysis level of the census subdivision, we considered a total of 30 variables in negative binomial regression models to identify variables associated with WNV incidence: 14 demographic and socio-economic variables, 10 land cover variables, two climatic variables, two mosquito vector habitat suitability index variables, one WNV habitat suitability index variable, and census subdivision area (Table S2 ). Two groups of variables had r > 0.7. Population density and all residential land cover classes were highly correlated, and therefore we chose to only keep population density, and dropped the four residential land cover variables. Mean daily temperature and habitat suitability index of both Ae. vexans and Cx. pipiens/restuans were highly correlated, so we chose to keep only habitat suitability index of Cx. pipiens/restuans , and dropped the temperature variable and habitat suitability index of Ae. vexans . Six variables were dropped due to absence of significant association with WNV infection in univariable models (Table S2 ). We ran a multivariable model with the remaining 18 variables. After model selection, the final model contained thirteen variables, three of which had an IRR that was significantly different from 1 (Table 4). The proportion of households earning less than $ 20,000 (low income) had an IRR around 2.8 (Table 4), suggesting a strong positive association of this variable with human WNV infection. The proportion of residents spending 30% or more of their income on their shelter (high spending on shelter) had an IRR around 0.7 (Table 4), suggesting a strong negative association of this variable with WNV infection. Habitat suitability index of Cx. pipiens/restuans had an IRR value around 2.0 (Table 4), suggesting a strong positive association of this variable with WNV infection. All other variables did not display a significant association with WNV infection. The R 2 value of the final model was around 0.67, which suggests strong statistical power in the final model. At the analysis level of the individual blood donation, we considered a total of 4 variables in logistic regression models to identify factors associated with WNV infection: donor age and sex, and year and month of detection test (Table S2 ). No pair of variables had r > 0.7. One variable was dropped due to absence of significant association with WNV infection in univariable models (Table S2 ). The final model contained three variables, all of which had a statistically significant IRR, with values very close to 1 (Table 4). Donor age, year of detection test and month of detection test had an IRR value between 1.002 and 1.003 (Table 4), suggesting a positive association of these variables with WNV infection. The R 2 value of the final model was lower than 0.001, which suggests poor statistical power in the final model, potentially stemming from extremely low variation in the outcome variable, i.e. small number of cases. Discussion Our study on spatiotemporal effects of climatic, geographic, ecological, demographic and socio-economic variables on West Nile virus (WNV) infection in the human blood donor population identified multiple modifiable and non-modifiable risk factors that may be practically useful to inform disease prevention and control efforts. First, we identified most regions of Southern Ontario along Saint Lawrence River, Lake Ontario, Lake Erie and south of Lake Huron, to be the main habitat for both Cx. pipiens/restuans and Ae. vexans , which is also where climate is warmest, somewhat more humid and mostly agricultural and urban. Habitat for WNV itself was much narrower, being confined to the Greater Toronto Area and the Windsor region. Some regions in Northern Ontario were also suitable for Ae. vexans . Positive WNV cases in the human blood donor population were mostly detected in Southern Ontario, with very few cases in Northern Ontario. Across years, the abundance and occurrence of Ae. vexans tended to decrease, but remained mostly unchanged for Cx. pipiens/restuans . We observed a peak during the month of August in abundance and occurrence of Cx. pipiens/restuans and Ae. vexans , the two main WNV vectors in our study area, and mosquito vector WNV infection prevalence. Positive mosquito vector pools and infection cases in the blood donor population mostly occurred in the second half of the study period, during the months of August and September. These results mostly support previous literature 40,51–57 . However, our study sheds light on the spatiotemporal interplay in abundance and/or occurrence between the two main WNV vectors in northeastern North America, and how this affects WNV infection in the vector populations and in the human population, using longitudinal surveillance data over long timeframe and over a large study area. In addition, our study leverages blood donor testing data to provide spatiotemporally widespread arboviral detection in the Ontario population, which to date has not been investigated to such an extent in Canada, as opposed to other countries such as the United States 7 . The importance of climatic and geographic variables in the ecological niche modeling analyses for WNV vector species showed similar results from a recent study in eastern Ontario, which performed similar analyses at a smaller but overlapping spatial and temporal scale using mosquito surveillance data from 2011 to 2020 56 . Whereas the previous study identified a mediocre to weak importance of all variables, our study herein suggests a strong effect of temperature for both mosquito vector species and for WNV, and a strong effect of precipitation for Ae. vexans . This is to be expected given the larger climatic variation across this study area compared to the previous study. In this study, the threshold of temperature on habitat suitability for WNV was higher compared to that of the two mosquito species. Optimal amounts of precipitation were also lower compared to those of the two mosquito species alone. Interestingly, land cover had a moderate relative importance for both species, despite land cover classes being all relatively well represented in terms of sampling sites in the present study. Associations between habitat suitability and specific land cover classes were mostly similar between Ae. vexans , Cx. pipiens/restuans and WNV, except for wild grasslands which were unsuitable for WNV, and high levels of vegetation in urban landscape are not suitable for Cx. pipiens/restuans and WNV. Finally, urban landscapes with low to moderate levels of vegetation and wetlands were suitable for both mosquito species and for WNV. These results are largely supported by previous literature, where high temperatures, and vegetated urban and wetland cover were predictors of both WNV vector habitat and WNV transmission 55,56,58–61 . Our statistical models identified habitat suitability index for Cx. pipiens/restuans as a strong predictor of WNV infection in the human blood donor population. Due to their strong correlation with habitat suitability index for Cx. pipiens/restuans , habitat suitability index for Ae. vexans and averaged mean daily temperature couldn’t be included in multivariable statistical models at the level of the census subdivision. However, they are similarly strong predictors of WNV infection. Canada is currently experiencing effects of climate change on vector-borne disease risk, including WNV 4,34–40 . Given the strong effect of temperature on mosquito WNV vector habitat, which in turn affects WNV infection in the human population, rigorous surveillance of southerly locations is needed to effectively predict large upticks in human WNV cases across Canada. Our statistical models identified habitat suitability index for Cx. pipiens/restuans as a strong predictor of WNV infection in the human blood donor population, which may also be expanded to two highly correlated variables: habitat suitability index for Ae. vexans and averaged mean daily temperature. Our statistical models also identified the proportion of low-income households and proportion of households where 30% or more of the residents’ income is spent on shelter being positively and negatively, respectively, associated with WNV infection in the human population. Household wealth is widely known to be associated with mosquito vector density, potentially through greater perception of risk and access to mosquito control methods 44,46,62,63 . In a previous analysis at a smaller but completely overlapping spatial and temporal scale in the city of Ottawa, Ontario, from 2007 to 2014, proportion of 60-years-old-and-older shelters was associated with higher WNV risk. A similar result was also observed in Chicago during a WNV outbreak in 2002 63 . This variable is possibly linked to availability of suitable habitats, i.e. breeding sites in suboptimal drainage systems 64 , used by WNV vector species. Population density and an urban environment were previously associated with higher WNV risk in Ottawa, Ontario, from 2007–2014 55 , and in Chicago and Detroit during a WNV outbreak in 2002 59 . However, in our study here there was no significant effect of either population density or residential land cover on WNV infection in the human population, despite using extensive and high-resolution datasets of both types of variables. Certain residential land cover classes do seem to affect habitat suitability of the two mosquito WNV vector species and WNV itself, albeit with an importance moderate or low compared to weather variables, but this effect does not seem to translate to significantly higher WNV infection in humans, as demonstrated by our multivariable statistical models. These results either suggest limited power to identify an effect of population density or residential land cover by our multivariable statistical models, or that spurious associations in previous studies would have been better explained by alternative unmeasured variables that have been included in this study herein. Our study has a few limitations. Nucleic acid tests we used have a possibility of cross-reacting with other members of the Japanese encephalitis serocomplex, such as Japanese encephalitis, Murray Valley encephalitis, Saint Louis encephalitis and Kunjin virus 65–68 (leading to a modification to the “National case definition: West Nile virus – Canada.ca” in 2024), which is relevant for donors with a certain travel history or due to vaccination with Japanese encephalitic virus vaccine. Also, most data points in the mosquito sampling data and the human donor data are highly aggregated around urban and metropolitan areas of the study area. However, outcomes of these limitations on the main results of our analyses are likely to be minimal. In conclusion, results from our study point to several modifiable risk factors that may be used as points of entry for practical intervention aimed at reducing risk of mosquito-borne pathogens in Canada, in a context of increasing mosquito-borne pathogen exposure and illness. For example, our study supports the need for government education campaigns and incentives facilitating renovations aimed at reducing mosquito habitat and/or exposure to mosquitoes, especially in areas with highly suitable habitat for Cx. pipiens/restuans in Southern Ontario, such as the “Prevention of West Nile virus – Canada.ca” program. Such programs and incentives would also be useful in practical intervention measures against Aedes albopictus , an invasive mosquito vector for several pathogens, including dengue virus, which is detected periodically in Southern Ontario. Our study is one of the few using arboviral detection tests over a large area and a large period in the human population to identify factors predictive of WNV infection. Such studies are not impeded by the same spatial, temporal and clinical (e.g., under reporting) biases as those using disease case reports, but are uncommon due to the sheer amount of work needed to produce representative databases. Materials and Methods Study area and duration The study area is situated across most of the province of Ontario, in Canada, at and below latitudes 50–51° N (Fig. 1). The study spans the years 2003 to 2022. Mosquito collection data Since 2002, Public Health Ontario (PHO) has collected data on mosquito surveillance activities conducted by public health units in Ontario, which comprise mosquito capture and identification of 22 species/species groups. Capture is performed using the protocol as described in Talbot et al. (2023) 56 . Briefly, Culex pipiens/restuans are prioritized over all other species, and Ae. vexans is also of high priority. RNA is extracted from each pool using RNeasy Mini Kit (Qiagen, Hilden, Germany). RNA extracts are tested by quantitative polymerase chain reaction (PCR) for arbovirus presence 69 . We used all mosquito capture data available across the entire province of Ontario, Canada, comprising data from all 34 public health units, for the years 2003 to 2022. For each mosquito pool, we calculated the WNV minimum infection rate (MIR), which is the test outcome (positive = 1; negative = 0), divided by the number of mosquitoes present in the pool, multiplied by 1000. Land cover and climatic data We followed the approach of Talbot et al. (2023) 56 to process land cover and climate data for subsequent ecological niche models. Land cover data for the year 2013 were obtained from Agriculture and Agri-Food Canada (AAFC) and the United States Geological Survey (USGS) with a resolution of 30 meters across our study area. Land cover data for the year 2013 were chosen because it is approximately the mid-point of our study period from 2003 to 2022. Annual crop inventory data from AAFC 70 comprise seven land cover classes: open water, wetlands, agricultural croplands, natural grasslands, forests, exposed surface and residential areas. Residential areas were subdivided into four categories, according to the normalized difference vegetation index (NDVI) from USGS, created using Landsat 8 data (collection 2, level 2, maximum 50% clouds) from May to October 2013, from USGS 71,72 . The goal of this procedure was to subdivide urban environments according to the presence of vegetation, which may affect habitat selection by the studied species. Residential areas with NDVI 0.15 and 0.30 and 0.60 as high green residential areas. Data from AAFC and USGS used the same resolution with matching cell frames, and therefore merging of the two datasets could be performed manually. Therefore, we considered temperature and precipitation data across the entire study duration. Temperature and precipitation data were obtained from the National Aeronautics and Space Administration (NASA). Given their ease of access, we extracted data on annual total precipitation, and annual maximum and minimum temperature from the annual surface weather and climatological summaries from NASA 73 with a resolution of 1,000 meters. These data were averaged across the 20 years of the study duration, and then divided by number of days in a year to obtain 20-year averaged mean daily temperature and 20-year averaged total daily precipitation for the total period across our study area. Ecological niche modeling analysis We followed the approach of Talbot et al. (2023) 56 . The two studied mosquito species were observed at least once in the vast majority of the 3,010 sampling locations over the study period (2,783 for Ae. vexans , or 92%, and 2,686 for Cx. pipiens/restuans , or 90%). At least one mosquito vector species was present in a total of 2,810 sites, for which a WNV detection test could be performed, and lead to at least one positive outcome in a total of 650 sites. Therefore, the number of visits in which each species was observed at least once at each site was calculated, and divided by the total number of visits at each site over the study period. The resulting value is the frequency of observed presence, and is a form of aggregated performance measure often used in species distribution models 74 . Sites were then dichotomized into a 0/1 distribution for each mosquito species and for WNV, which is a requirement of the approach.: Sites with 50% or more species occurrence for both mosquito species were classified as ‘presence, and sites with less than 50% as ‘absence. Sites with at least one positive test outcome for WNV were classified as ‘presence’, and others as ‘absence’. The random forest algorithm was considered suitable for our analyses because the presence and absence of a species in a given sampling site visit are likely to be affected by the same sampling bias 75,76 . This decision tree-based approach performs as well as the maximum entropy approach 77,78 , and better than traditional regression-based approaches when using large datasets sampled over a long duration and a large spatial scale 79 . We performed the analyses using the ‘biomod2’ package 80 in R 4.2.1 (R Development Core Team, Vienna, Austria). We projected all land cover and climatic datasets to Albers Conic Equal Area, which was the original projection of the land cover dataset from Agriculture Canada. All explanatory variables were resampled to a cell size of 120 x 120 meters and set to be at the same cell frame to reduce spatial bias caused by unequal resolution with the mosquito dataset 75 . We computed the Pearson’s correlation coefficient among climatic variables at sampling sites in R 4.2.1. Potential collinearity problems were considered if r > 0.7. We set the prevalence parameter to 0.5 was specified, meaning “presence” and “absence” distributions are considered in equal proportions in the analysis 81 . For each species, we trained 100 replicate models using 80% of data. To evaluate each model, we computed a receiver operating characteristic’s area under the curve (AUC) using the remaining 20% of data. Data was selected randomly in each model for training versus testing. We used the final model, trained by the 100 replicate models and using 100% of data, to generate a habitat suitability index (HSI) map in the study area. We kept all other parameters at default values. We used all models with AUC above 0.7 to generate an ensemble niche model 80 . We generated response plots of the mean HSI across models and committee averaging 82 , for each explanatory variable. We calculated variable importance for each explanatory variable, which varies from 0 to 1, using a procedure of 100 permutations from the ensemble niche model. Lastly, we created a projected HSI map from the model with the highest AUC for each mosquito species and for WNV. Socio-economic data Statistics Canada collects data on a large variety of socio-economic variables during the Census of Population every five years across Canada. Data for the years 2011 and 2016 were initially considered because they are near the mid-point of our study period from 2003 to 2022. Given their easier access and processing, we chose data for the year 2016, which we extracted using the Beyond 20/20 Professional Browser software. We selected all variables that relate to either sex, age, ethnicity, population density, income, shelter infrastructure, education, and occupation, to be included in analyses on the determinants of WNV infection in the human population. We uploaded data at the level of the census subdivision, to match blood donation arboviral testing data structure. Blood donation arboviral testing data Across the study area, since 2003, Canadian Blood Services (CBS) have tested blood donations across Ontario. These tests were performed on all blood donations from 2003 to 2015. As of December 2015, all donations from June to November were tested, but only donations from travelers to certain destinations were tested from December to May. These months usually see much less mosquito activity, and therefore likelihood of exposure to an infected mosquito is negligible. Groups of six donations were tested in 6-unit minipools. Positive minipools were then retested separately for each donation from the corresponding minipool, along with all donations from surrounding areas for the next two weeks. From June 2003, testing was performed using TaqScreen WNV test IUO (F. Hoffmann-La Roche AG, Basel, Switzerland). From June 2007, testing was performed using the IND cobas TaqScreen WNV test for use with the cobas s 201 system. From June 2008, testing was performed using the licensed cobas TaqScreen WNV test for use with the cobas s 201 system. From December 2017, testing was performed using the cobas® WNV – Nucleic acid test for use on the cobas® 6800/8800 Systems. At donor registration, the donor’s date of birth, sex and residential address are recorded. For confidentiality purposes, we used only the donor’s reported sex; year of birth, to obtain approximative age of the blood donor at the time of donation; census subdivision of residence from the Canadian Census of Population boundaries for the year 2016; and unique donor identifier. To avoid sampling biases leading to spurious associations in our analyses of determinants of WNV infection in the human population, we chose to exclude all data from census subdivisions where fewer than 1000 WNV tests on blood donations were available, which are mostly located in sparsely populated areas mostly in northeastern parts of the study area, and north of the northern limit of the study area. This exclusion step lead to the removal of 126 subdivisions out of the total of 417 (30%) where blood donor data is available. In the remaining 291 subdivisions, we chose to exclude 4 census subdivisions where total number of residents was lower than 500, because most socio-economic variables were missing for confidentiality reasons. The final dataset contained 287 census subdivisions. Two of these census subdivisions, named “Kenora, Unorganized” and “Thunder Bay, Unorganized” had boundaries above the northern limit of the study area. However, the large majority of blood donations in these subdivisions were conducted in the south, near the Kenora and Thunder Bay townships, respectively, which is also where the large majority of the residents of these subdivisions live. For these reasons, we considered only the portion in these two subdivisions that lie within our study area (Fig. 1, Fig. 4). Spatiotemporal WNV infection analyses To attain our main objective, we investigated the effect of a wide range of factors related to land cover, climate, mosquito habitat and socio-economic status of residents on WNV infection in the human population, using data from WNV tests on blood donations. These factors were carefully chosen a priori to include variables most likely affecting WNV risk in the human population, including aspects related to mosquito occurrence, mosquito activity, and characteristics of residents linked to higher exposure to mosquito bites. We conducted these analyses as two different levels: one at the census subdivision, and one at the individual blood donation. For the first level of analysis, we gathered all arboviral testing data from blood donations, for which residence information is available, from 2003 to 2022, and grouped them within the respective census subdivision of residence of the blood donor. We conducted a simple generalized linear regression, using the ‘lme4’ package 83 in R 4.2.1, for each variable separately, namely selected socio-economic variables, land cover variables (percent cover of each class in census subdivisions), climatic variables (20-year averaged daily total precipitations and 20-year averaged daily mean temperature, averaged across census subdivisions), habitat suitability index variables (for each mosquito vector species and for WNV, averaged across census subdivisions), and lastly, the census subdivision geographic area (in km 2 ), which can be stochastically associated with number of positive cases. Prior to these analyses, we subtracted the mean and divided by the standard deviation of all values for numerical variables, namely test outcome, donor age, month of test, and year of test. We used the zero-inflated negative binomial modeling family, using the ‘pscl’ package 84 in R 4.2.1, where we used the number of positive cases within census subdivisions as outcome variable, and number of donations tested within census subdivisions as offset variable. For the second level of analysis, we gathered all WNV testing data from blood donations, for which residence information is available, from 2003 to 2022. We conducted a simple mixed-effects generalized linear regression, using the ‘lme4’ package 83 in R 4.2.1, for each variable separately, namely donor sex, donor age, month of test, and year of test, against test outcome, where 1 is positive and 0 is negative. We used a hierarchical random-effects term, which is the unique donor number nested within the census subdivision of residence. Prior to these analyses, we subtracted the mean and divided by the standard deviation of all values for numerical variables, namely test outcome, donor age, month of test, and year of test. Given the data structure is binomial, the ‘binomial’ modeling family was most intuitive, but the extremely small number of positive cases compared to negative cases lead to numerous fitting errors. Therefore, we chose the more general ‘gaussian’ modeling family, for which no modelling problem occurred. For both analysis levels, we used Pearson’s correlation coefficient in R 4.2.1, to identify correlation among all selected variables, and dropped one variable from each pair of variables displaying r > 0.7. Next, we selected all variables displaying a significant univariable association ( P < 0.01), with WNV infection into a multivariable generalized linear regression analysis, using the same packages as previously in R 4.2.1. For the analysis at the level of the census subdivision, given the large number of variables included, we proceeded with a two-step model selection approach using the Bayesian Information Criterion (BIC), i.e. the Akaike Information Criterion using the logarithm of the number of observations as the k parameter. We computed a BIC value for the full model and for all combinations of the full model excluding one variable. A BIC value decrease of 2 or more is considered positive evidence for a variable displaying little effect on the response variable 85 . We ran a new multivariable regression model, but this time excluding all variables that, when dropped from the full model, caused a decrease of the BIC value of more than 2. We then reapplied the same model selection approach as a second step on this new model. Using this model selection approach, we ensured that only the most important variables were retained in the resulting final model. In both analysis levels, we ran the final model using the full dataset, and computed incidence rate ratio (IRR) values, i.e. the exponents of the slope coefficients, 95% confidence intervals for the IRR values, and P values for each variable. Finally, we computed the R 2 of the final model, using the ‘modelsummary’ package 86 in R 4.2.1. Declarations Acknowledgements We thank Canadian Blood Services (in particular Bill Ferguson), Public Health Ontario (in particular Curtis Russell), Statistics Canada, Agriculture and Agri-Food Canada, United States Geological Survey National Aeronautics and Space Administration, for access to blood donor, mosquito capture, socio-economic, geographic and climatic data used in this study. We also thank Maxime Rioux-Rousseau for their help collecting and processing geographic and climatic data sources. This project was supported by grants from the Canadian Institutes of Health Research and Natural Science and Engineering Research Council of Canada to M.A.K. Author contribution B.T. contributed to the conception of the study, performed analyses, interpreted results, and wrote the manuscript draft. M.A.K. and A.L. contributed in the conception of the study, interpreted results and edited the manuscript draft. N.H.O., S.F.O. and S.J.D. contributed in the conception of the study and edited the manuscript draft. Data availability statement Blood donor data from Canadian Blood Services is not available due to privacy reasons according to our data sharing agreement between University of Ottawa and Canadian Blood Services. All other data are freely available on the webpage of the relevant authority: Public Health Ontario (https://www.publichealthontario.ca/en/Data-and-Analysis/Using-Data/Data-Requests) for mosquito capture data, Statistics Canada (https://www150.statcan.gc.ca/n1/en/type/data) for socio-economic data, Agriculture and Agri-Food Canada (https://www.agr.gc.ca/atlas/apps/aef/main/index_en.html?AGRIAPP=23) and United States Geological Survey (https://earthexplorer.usgs.gov/) for geographic data, and National Aeronautics and Space Administration (https://www.earthdata.nasa.gov/) for climatic data. Additional Information We declare no competing interests. References Gubler, D. J. The continuing spread of West Nile Virus in the western hemisphere. Clinical Infectious Diseases 45 , 1039–1046 (2007). Reisen, W. K. Ecology of West Nile virus in North America. Viruses 5 , 2079–2105 (2013). Chancey, C., Grinev, A., Volkova, E. & Rios, M. The global ecology and epidemiology of West Nile Virus. BioMed Research International 2015 , 1–20 (2015). Chen, C. et al. Climate change and West Nile virus in a highly endemic region of North America. IJERPH 10 , 3052–3071 (2013). Shing, E. et al. The direct healthcare costs attributable to West Nile virus illness in Ontario, Canada: a population-based cohort study using laboratory and health administrative data. BMC Infect Dis 19 , 1059 (2019). Todoric, D. et al. An overview of the National West Nile Virus Surveillance System in Canada: A One Health approach. CCDR 48 , 181–187 (2022). Busch, M. P. et al. West Nile virus infections projected from blood donor screening data, United States, 2003. Emerg. Infect. Dis. 12 , 395–402 (2006). Murray, J. & Cohen, A. L. Infectious Disease Surveillance. in International Encyclopedia of Public Health 222–229 (Elsevier, 2017). doi:10.1016/B978-0-12-803678-5.00517-8. Doyle, T. J. Completeness of notifiable infectious disease reporting in the United States: an analytical literature review. American Journal of Epidemiology 155 , 866–874 (2002). Amodio, E., Zarcone, M., Casuccio, A. & Vitale, F. Trends in epidemiology: the role of denominator fluctuation in population based estimates. AIMS Public Health 8 , 500–506 (2021). Won, E. J. et al. Under-diagnosis of vector-borne diseases among individuals suspected of having Scrub Typhus in South Korea. PLoS ONE 18 , e0286631 (2023). Keita, A. K. et al. The detection of vector-borne-disease-related DNA in human stool paves the way to large epidemiological studies. Eur J Epidemiol 30 , 1021–1026 (2015). Koliopoulos, P. et al. Multiplex-RT-PCR-ELISA panel for detecting mosquito-borne pathogens: Plasmodium sp. preserved and eluted from dried blood spots on sample cards. Malar J 20 , 66 (2021). Patrican, L. A. et al. Host-feeding patterns of Culex mosquitoes in relation to trap habitat. Emerg. Infect. Dis. 13 , 1921–1923 (2007). Hamer, G. L. et al. Host selection by Culex pipiens mosquitoes and West Nile virus amplification. Am. J. Trop. Med. Hyg. 80 , 268–278 (2009). Hamer, G. L. et al. Culex pipiens (Diptera: Culicidae): a bridge vector of West Nile virus to humans. Journal of Medical Entomology 45 , 125–128 (2008). Ebel, G. D., Rochlin, I., Longacker, J. & Kramer, L. D. Culex restuans (Diptera: Culicidae) relative abundance and vector competence for West Nile virus. me 42 , 838–843 (2005). Farajollahi, A., Fonseca, D. M., Kramer, L. D. & Marm Kilpatrick, A. “Bird biting” mosquitoes and human disease: a review of the role of Culex pipiens complex mosquitoes in epidemiology. Infection, Genetics and Evolution 11 , 1577–1585 (2011). Andreadis, T. G. The contribution of Culex pipiens complex mosquitoes to transmission and persistence of West Nile virus in North America. Journal of the American Mosquito Control Association 28 , 137–151 (2012). Tiawsirisup, S. et al. Vector competence of Aedes vexans (Diptera: Culicidae) for West Nile virus and potential as an enzootic vector. Journal of Medical Entomology 45 , 452–457 (2008). Molaei, G. & Andreadis, T. G. Identification of avian- and mammalian-derived bloodmeals in Aedes vexans and Culiseta melanura (Diptera: Culicidae) and its implication for West Nile virus transmission in Connecticut, U.S.A. Journal of Medical Entomology 43 , 1088–1093 (2006). Greenberg, J. A., Lujan, D. A., DiMenna, M. A., Wearing, H. J. & Hofkin, B. V. Identification of blood meal sources in Aedes vexans and Culex quinquefasciatus in Bernalillo County, New Mexico. Journal of Insect Science 13 , 1–12 (2013). Anderson, J. F., Main, A. J. & Ferrandino, F. J. Horizontal and vertical transmission of West Nile virus by Aedes vexans (Diptera: Culicidae). Journal of Medical Entomology 57 , 1614–1618 (2020). Moser, S. K. et al. Scoping review of Culex mosquito life history trait heterogeneity in response to temperature. Parasites Vectors 16 , 200 (2023). Shaman, J. Using a dynamic hydrology model to predict mosquito abundances in flood and swamp water. Emerg. Infect. Dis. 8 , 8–13 (2002). Busch, M. P. et al. Virus and antibody dynamics in acute West Nile virus infection. J INFECT DIS 198 , 984–993 (2008). Dohm, D. J. & Turell, M. J. Effect of incubation at overwintering temperatures on the replication of West Nile virus in New York Culex pipiens (Diptera: Culicidae). me 38 , 462–464 (2001). Dohm, D. J., O’Guinn, M. L. & Turell, M. J. Effect of environmental temperature on the ability of Culex pipiens (Diptera: Culicidae) to transmit West Nile virus. J Med Entomol 39 , 221–225 (2002). Roiz, D., Ruiz, S., Soriguer, R. & Figuerola, J. Climatic effects on mosquito abundance in Mediterranean wetlands. Parasites Vectors 7 , 333 (2014). Giménez, J. O., Fischer, S., Zalazar, L. & Stein, M. Cold season mortality under natural conditions and subsequent hatching response of Aedes ( Stegomyia ) aegypti (Diptera: Culicidae) eggs in a subtropical city of Argentina. J. Med. Entomol. 52 , 879–885 (2015). Eisen, L. et al. The impact of temperature on the bionomics of Aedes ( Stegomyia ) aegypti , with special reference to the cool geographic range margins. J Med Entomol 51 , 496–516 (2014). Shocket, M. S. et al. Transmission of West Nile and five other temperate mosquito-borne viruses peaks at temperatures between 23°C and 26°C. eLife 9 , e58511 (2020). Ozdenerol, E., Taff, G. & Akkus, C. Exploring the spatio-temporal dynamics of reservoir hosts, vectors, and human hosts of West Nile Virus: a review of the recent literature. International Journal of Environmental Research and Public Health 10 , 5399–5432 (2013). Allan, B. F., Keesing, F. & Ostfeld, R. S. Effect of forest fragmentation on Lyme disease risk. Conservation Biology 17 , 267–272 (2003). Canadian Paediatric Society. West Nile virus in the context of climate change. Paediatrics & Child Health 13 , 399–400 (2008). Kim, B. I. et al. A conceptual model for the impact of climate change on fox rabies in Alaska, 1980-2010. Zoonoses Public Health 61 , 72–80 (2014). Gardner, A. M., Lampman, R. L. & Muturi, E. J. Land use patterns and the risk of West Nile virus transmission in Central Illinois. Vector-Borne and Zoonotic Diseases 14 , 338–345 (2014). Hoover, K. C. & Barker, C. M. West Nile virus, climate change, and circumpolar vulnerability. WIREs Clim Change 7 , 283–300 (2016). Ludwig, A. et al. Increased risk of endemic mosquito-borne diseases in Canada due to climate change. CCDR 45 , 91–97 (2019). Rakotoarinia, M. R. et al. Effects of land use and weather on the presence and abundance of mosquito-borne disease vectors in a urban and agricultural landscape in Eastern Ontario, Canada. PLoS ONE 17 , e0262376 (2022). Elmqvist, T. et al. Urbanization, Biodiversity and Ecosystem Services: Challenges and Opportunities . (Springer Netherlands, Dordrecht, 2013). doi:10.1007/978-94-007-7088-1. Li, Y. et al. Urbanization increases Aedes albopictus larval habitats and accelerates mosquito development and survivorship. PLoS Negl Trop Dis 8 , e3301 (2014). Kjellstrom, T. et al. Urban environmental health hazards and health equity. J Urban Health 84 , 86–97 (2007). Donnelly, M. A. P., Kluh, S., Snyder, R. E. & Barker, C. M. Quantifying sociodemographic heterogeneities in the distribution of Aedes aegypti among California households. PLoS Negl Trop Dis 14 , e0008408 (2020). Little, E. et al. Socio-ecological mechanisms supporting high densities of Aedes albopictus (Diptera: Culicidae) in Baltimore, MD. J. Med. Entomol. 54 , 1183–1192 (2017). Whiteman, A. et al. Aedes mosquito infestation in socioeconomically contrasting neighborhoods of Panama City. EcoHealth 16 , 210–221 (2019). Dowling, Z. et al. Linking mosquito infestation to resident socioeconomic status, knowledge, and source reduction practices in suburban Washington, DC. EcoHealth 10 , 36–47 (2013). Fuentes-Vallejo, M. et al. Territorial analysis of Aedes aegypti distribution in two Colombian cities: a chorematic and ecosystem approach. Cad Saude Publica 31 , 517–530 (2015). Vonesch, N. et al. Emerging zoonotic viral infections of occupational health importance. Pathogens and Disease 77 , ftz018 (2019). Ellwanger, J. H., Cardoso, J. D. C. & Chies, J. A. B. Variability in human attractiveness to mosquitoes. Current Research in Parasitology & Vector-Borne Diseases 1 , 100058 (2021). Ripoche, M., Campagna, C., Ludwig, A., Ogden, N. H. & Leighton, P. A. Short-term forecasting of daily abundance of West Nile virus vectors Culex pipiens-restuans (Diptera: Culicidae) and Aedes vexans based on weather conditions in Southern Québec (Canada). Journal of Medical Entomology 56 , 859–872 (2019). Giordano, B. V., Kaur, S. & Hunter, F. F. West Nile virus in Ontario, Canada: a twelve-year analysis of human case prevalence, mosquito surveillance, and climate data. PLOS ONE 12 , e0183568 (2017). Giordano, B., Turner, K. & Hunter, F. Geospatial analysis and seasonal distribution of West Nile virus vectors (Diptera: Culicidae) in Southern Ontario, Canada. International Journal of Environmental Research and Public Health 15 , 614 (2018). Mallya, S. et al. Factors associated with human West Nile virus infection in Ontario: a generalized linear mixed modelling approach. BMC Infectious Diseases 18 , 1–9 (2018). Talbot, B., Ardis, M. & Kulkarni, M. A. Influence of demography, land use, and urban form on West Nile virus risk and human West Nile virus incidence in Ottawa, Canada. Vector-Borne and Zoonotic Diseases 19 , 533–539 (2019). Talbot, B. et al. Ecological niche and positive clusters of two West Nile virus vectors in Ontario, Canada. EcoHealth 20 , 249–262 (2023). Talbot, B., Caron-Lévesque, M., Ardis, M., Kryuchkov, R. & Kulkarni, M. A. Linking bird and mosquito data to assess spatiotemporal West Nile virus risk in humans. EcoHealth 16 , 70–81 (2019). Ludwig, A., Rousseu, F., Kotchi, S. O., Allostry, J. & Fournier, R. A. Mapping the abundance of endemic mosquito-borne diseases vectors in southern Quebec. BMC Public Health 23 , 924 (2023). Ruiz, M. O., Walker, E. D., Foster, E. S., Haramis, L. D. & Kitron, U. D. Association of West Nile virus illness and urban landscapes in Chicago and Detroit. International Journal of Health Geographics 6 , 10 (2007). Johnson, B. J. et al. The roles of mosquito and bird communities on the prevalence of West Nile virus in urban wetland and residential habitats. Urban Ecosystems 15 , 513–531 (2012). Skaff, N. K. & Cheruvelil, K. S. Fine-scale wetland features mediate vector and climate-dependent macroscale patterns in human West Nile virus incidence. Landscape Ecology 31 , 1615–1628 (2016). Carrasquilla, M. C. et al. Entomological characterization of Aedes mosquitoes and arbovirus detection in Ibagué, a Colombian city with co-circulation of Zika, dengue and chikungunya viruses. Parasites Vectors 14 , 446 (2021). Ruiz, M. O., Tedesco, C., McTighe, T. J., Austin, C. & Kitron, U. Environmental and social determinants of human risk during a West Nile virus outbreak in the greater Chicago area, 2002. Int J Health Geogr 3 , 8 (2004). Ozdenerol, E., Bialkowska-Jelinska, E. & Taff, G. N. Locating suitable habitats for West Nile Virus-infected mosquitoes through association of environmental characteristics with infected mosquito locations: a case study in Shelby County, Tennessee. International Journal of Health Geographics 7 , 12 (2008). Drews, S. J. et al. A case series of inactivated Japanese encephalitis virus vaccination associated with positive West Nile virus blood donor screening nucleic acid tests. Transfusion 60 , 1097–1103 (2020). O’Brien, S. F. et al. West Nile virus in 2006 and 2007: the Canadian Blood Services’ experience. Transfusion 50 , 1118–1125 (2010). Vamvakas, E. C., Kleinman, S., Hume, H. & Sher, G. D. The development of West Nile virus safety policies by Canadian Blood Services: guiding principles and a comparison between Canada and the United States. Transfusion Medicine Reviews 20 , 97–109 (2006). Postler, T. S. et al. Renaming of the genus Flavivirus to Orthoflavivirus and extension of binomial species names within the family Flaviviridae. Arch Virol 168 , 224, s00705-023-05835–1 (2023). Lanciotti, R. S. et al. Rapid detection of West Nile virus from human clinical specimens, field-collected mosquitoes, and avian samples by a TaqMan reverse transcriptase-PCR assay. J. Clin. Microbiol. 38 , 4066–4071 (2000). Agriculture and Agri-Food Canada. Annual Crop Inventory. https://www.agr.gc.ca/atlas/apps/metrics/index-fr.html?appid=aci-iac&lang=en (2022). Masek, J. G. et al. A Landsat surface reflectance dataset for North America, 1990–2000. IEEE Geosci. Remote Sensing Lett. 3 , 68–72 (2006). Vermote, E., Justice, C., Claverie, M. & Franch, B. Preliminary analysis of the performance of the Landsat 8/OLI land surface reflectance product. Remote Sensing of Environment 185 , 46–56 (2016). Thornton, M.M. et al. Daymet: daily surface weather data on a 1-km grid for North America, Version 4. 0 MB (2020) doi:10.3334/ORNLDAAC/1840. Halvorsen, R. A gradient analytic perspective on distribution modelling. Sommerfeltia 35 , 1–165 (2012). Sillero, N. & Barbosa, A. M. Common mistakes in ecological niche models. International Journal of Geographical Information Science 35 , 213–226 (2021). Phillips, S. J. et al. Sample selection bias and presence-only distribution models: implications for background and pseudo-absence data. Ecological Applications 19 , 181–197 (2009). Zhao, Z., Xiao, N., Shen, M. & Li, J. Comparison between optimized MaxEnt and random forest modeling in predicting potential distribution: a case study with Quasipaa boulengeri in China. Science of The Total Environment 842 , 156867 (2022). Beeman, S. P., Morrison, A. M., Unnasch, T. R. & Unnasch, R. S. Ensemble ecological niche modeling of West Nile virus probability in Florida. PLoS ONE 16 , e0256868 (2021). Mondal, R. & Bhat, A. Comparison of regression-based and machine learning techniques to explain alpha diversity of fish communities in streams of central and eastern India. Ecological Indicators 129 , 107922 (2021). Thuiller, W., Lafourcade, B., Engler, R. & Araújo, M. B. BIOMOD - a platform for ensemble forecasting of species distributions. Ecography 32 , 369–373 (2009). Barbet-Massin, M., Jiguet, F., Albert, C. H. & Thuiller, W. Selecting pseudo-absences for species distribution models: how, where and how many? Methods in Ecology and Evolution 3 , 327–338 (2012). Araujo, M. & New, M. Ensemble forecasting of species distributions. Trends in Ecology & Evolution 22 , 42–47 (2007). Bates, D., Mächler, M., Bolker, B. & Walker, S. Fitting linear mixed-effects models Using lme4 . J. Stat. Soft. 67 , (2015). Zeileis, A., Kleiber, C. & Jackman, S. Regression Models for Count Data in R . J. Stat. Soft. 27 , (2008). Raftery, A. E. Bayesian model selection in social research. Sociological Methodology 25 , 111 (1995). Arel-Bundock, V. modelsummary : Data and Model Summaries in R . J. Stat. Soft. 103 , (2022). Tables Table 1. Descriptive statistics relating to the abundance (number of mosquitoes per sampling visit) and occurrence (proportion of presence across sampling visits) for adult mosquitoes from both studied species, Culex pipiens/restuans and Aedes vexans , collected across 3,010 sampling locations in Ontario, Canada, 2003-2022. Species Culex pipiens/restuans Aedes vexans Total sampled adult mosquitoes 934,951 1,778,876 Number of visits/site (mean ± standard deviation) 48 ± 80 Abundance/visit (mean ± standard deviation) 5 ± 8 12 ± 18 Occurrence/visit (mean ± standard deviation) 0.57 ± 0.32 0.65 ± 0.29 Total number of WNV detection tests 102,938 82,134 Number of positive pools [% of total for all species] 2,949 [88] 182 [5] Total number of adult mosquitoes tested for West Nile virus 916,673 1,147,874 Number of adult mosquitoes tested/test (mean ± standard deviation) 9 ± 12 14 ± 15 Total minimum infection rate (positive pools/number tested x1000) 3.22 0.16 Table 2. Parameters used to generate ensemble niche models for both studied species, Culex pipiens/restuans and Aedes vexans , collected across 3,010 sampling locations in Ontario, Canada, 2003-2022. Number of models included out of the 100 random forest models generated, response operating curves for the best-performing model, for the mean habitat suitability index (HSI) averaged across models, and committee averaging of HSI across models and variable importance calculated with a 100-permutation procedure for the three explanatory variables are shown. Species Aedes vexans Culex pipiens/restuans WNV Number of models 8 100 100 Area under the curve Best-performing model 0.72 0.82 0.79 Mean HSI values 0.93 0.91 0.93 Committee averaging 0.92 0.90 0.92 Variable importance Land cover 0.32 0.21 0.29 Precipitations 0.62 0.19 0.39 Temperature 0.64 0.78 0.62 Table 3. Descriptive statistics relating to the number of blood donors and donations to the Canadian Blood Services in Ontario, Canada, 2003-2022, tested for presence of West Nile virus (WNV). Age / Sex 11 to 38 39 to 65 66 to 93 Female Male Total Number of tests on blood donations 2,200,777 3,990,529 389,440 2,776,604 3,804,165 6,580,769 Number of unique blood donors 576,752 439,512 52,525 504,580 467,814 972,394 Mean number of donations per donor 3.8 9.1 2.5 5.5 8.1 6.8 Number of WNV-positive donations 16 77 9 38 64 102 WNV-positive rate (x100,000 donations) 0.7 1.9 2.3 1.4 1.7 1.6 Table 4. Regression incidence rate ratio (IRR) values, 95% confidence intervals [95% CI], the P value for each variable, and the R 2 for the model, for the multivariable modelling analyses on the WNV infection in the human donor population in Ontario from 2003 to 2022, at the analysis levels of the individual donor and the 2016 Census subdivision. HSI refers to habitat suitability index of the corresponding mosquito vector species. Analysis level Variable (units) IRR [95% CI] P R 2 Census subdivision Low income (%) 2.829 [2.061, 3.885] <0.001 0.677 High spending on shelter (%) 0.702 [0.526, 0.939] 0.017 Cx. pipiens/restuans (HSI) 1.950 [1.160, 3.280] 0.012 West Nile virus (HSI) 1.179 [0.845, 1.645] 0.333 Precipitations (mm) 0.872 [0.558, 1.363] 0.548 Population density (residents/km 2 ) 0.908 [0.720, 1.145] 0.414 Mean income ($) 1.316 [0.941, 1.839] 0.108 Residents older than 64 (%) 0.878 [0.584, 1.321] 0.533 Occupation in trades (%) 0.807 [0.479, 1.357] 0.418 Occupation in natural resources (%) 1.389 [0.927, 2.081] 0.111 Land cover of wetlands (%) 0.654 [0.347, 1.234] 0.190 Land cover of wild grasslands (%) 1.397 [0.909, 2.148] 0.127 Land cover of forests (%) 1.332 [0.680, 2.612] 0.403 Individual donation Donor age (years since birth) 1.002 [1.001, 1.003] <0.001 <0.001 Year of detection test (date) 1.003 [1.002, 1.003] <0.001 Month of detection test (date) 1.002 [1.001, 1.003] <0.001 Additional Declarations No competing interests reported. Supplementary Files TableS1.docx TableS2.docx Cite Share Download PDF Status: Published Journal Publication published 28 Dec, 2024 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Revision requested 26 Aug, 2024 Reviewers agreed at journal 11 Aug, 2024 Reviewers agreed at journal 08 Aug, 2024 Reviews received at journal 08 Aug, 2024 Reviews received at journal 04 Aug, 2024 Reviews received at journal 26 Jul, 2024 Reviews received at journal 26 Jul, 2024 Reviewers agreed at journal 18 Jul, 2024 Reviewers agreed at journal 17 Jul, 2024 Reviewers agreed at journal 17 Jul, 2024 Reviewers agreed at journal 17 Jul, 2024 Reviewers invited by journal 16 Jul, 2024 Editor assigned by journal 12 Jul, 2024 Editor invited by journal 12 Jul, 2024 Submission checks completed at journal 10 Jul, 2024 First submitted to journal 09 Jul, 2024 You are reading this latest preprint version 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-4714418","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":335438013,"identity":"6fe78061-2897-4813-98d0-d3675a1b0d2a","order_by":0,"name":"Benoit Talbot","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+ElEQVRIiWNgGAWjYBACAxCRACLYeYBEBdFaEhgkGJiBWg6cYWDgYSNGCwNMy8E2IrSYs/cefPDwB0MdfzPvwccf5x2Ws5dvfibxg8FOHpcWy55zyQYgh0kc5ks2OLjtsDEPG5uZZA9DsmEDLofdyDGTAPvlMI+ZBFBLYg8bg9kNHoYDjDi13H8D0SIP1jIHpIX9280/DAfscdvCA9FiANbSANLCY3YbaEsiLi2WPTnGBglpEpIbQX45cyzdmOdYTvlvGYPkZFxazNnPGD78YWPDL3ccGHQVNdZy7M3HNxu+qbCzxaUFCiQwHIxf/SgYBaNgFIwC/AAAHCVRpgLivpMAAAAASUVORK5CYII=","orcid":"","institution":"University of Ottawa","correspondingAuthor":true,"prefix":"","firstName":"Benoit","middleName":"","lastName":"Talbot","suffix":""},{"id":335438014,"identity":"94b44397-203d-4d8e-b682-6d048153626b","order_by":1,"name":"Antoinette Ludwig","email":"","orcid":"","institution":"Public Health Agency of Canada, St-Hyacinthe","correspondingAuthor":false,"prefix":"","firstName":"Antoinette","middleName":"","lastName":"Ludwig","suffix":""},{"id":335438015,"identity":"3accb68a-6356-4ab6-8fc1-54d757ac41a0","order_by":2,"name":"Sheila F. O’Brien","email":"","orcid":"","institution":"Canadian Blood Services","correspondingAuthor":false,"prefix":"","firstName":"Sheila","middleName":"F.","lastName":"O’Brien","suffix":""},{"id":335438016,"identity":"a8e82a77-f0da-492b-b269-7ca29ad3884a","order_by":3,"name":"Steven J. Drews","email":"","orcid":"","institution":"Canadian Blood Services","correspondingAuthor":false,"prefix":"","firstName":"Steven","middleName":"J.","lastName":"Drews","suffix":""},{"id":335438017,"identity":"987919aa-470b-4f05-84b6-f1f434e171e9","order_by":4,"name":"Nicholas H. Ogden","email":"","orcid":"","institution":"Public Health Agency of Canada, St-Hyacinthe","correspondingAuthor":false,"prefix":"","firstName":"Nicholas","middleName":"H.","lastName":"Ogden","suffix":""},{"id":335438018,"identity":"950a60ff-2fd1-4414-a097-f28555edd4c5","order_by":5,"name":"Manisha A. Kulkarni","email":"","orcid":"","institution":"University of Ottawa","correspondingAuthor":false,"prefix":"","firstName":"Manisha","middleName":"A.","lastName":"Kulkarni","suffix":""}],"badges":[],"createdAt":"2024-07-09 21:38:30","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4714418/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4714418/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-024-82739-3","type":"published","date":"2024-12-28T15:57:14+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":61763114,"identity":"1ba02050-57f9-4ada-9446-d92c24d7860d","added_by":"auto","created_at":"2024-08-05 09:51:04","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1933817,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial representations of land cover classes for the year 2013 according to Agriculture and Agri-Food Canada and United States Geological Survey (A); total daily precipitations averaged from 2003 to 2022 according to National Aerospace Space Agency (B); mean daily temperature averaged from 2003 to 2022 according to National Aerospace Space Agency (C); and projected habitat suitability index for\u003cem\u003e Aedes vexans\u003c/em\u003e (D), \u003cem\u003eCulex pipiens/restuans\u003c/em\u003e (E) and West Nile virus-infected mosquito vector pools (F). Habitat suitability indices were created using mosquito collection data from Public Health Ontario eaHat 3,010 sampling locations from 2003 to 2022 (D-E), and a reduced dataset of 2,810 sampling locations from 2003 to 2022 excluding locations where the two vectors were not observed at all (F).\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-4714418/v1/348bb7f335bd99e59600401f.png"},{"id":61763813,"identity":"84d9c6bd-6bd0-4b53-91ed-af97eb8355c6","added_by":"auto","created_at":"2024-08-05 09:59:04","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":291132,"visible":true,"origin":"","legend":"\u003cp\u003eHistograms showing average abundance across sampling visits (A-B), average occurrence (% divided by 10) across sampling visits (C-D) of \u003cem\u003eAedes vexans\u003c/em\u003e and \u003cem\u003eCulex pipiens/restuans\u003c/em\u003e adult mosquitoes, and West Nile virus (WNV) minimum infection rate (MIR), sampled and tested by Public Health Ontario at 3,010 sampling locations from 2003 to 2022, and number of blood donations to the Canadian Blood Services that tested positive for WNV, for each year (A, C), and for each month from June to October (B, D).\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-4714418/v1/508a709c41d061830c5cbd9a.png"},{"id":61763120,"identity":"92d1aa2c-5eab-4538-ab1c-1de96c7a9bb7","added_by":"auto","created_at":"2024-08-05 09:51:04","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":84226,"visible":true,"origin":"","legend":"\u003cp\u003eGraphs of habitat suitability index (black dots/lines) and committee averaging (red dots/lines) of the ensemble niche models for\u003cem\u003e Aedes vexans\u003c/em\u003e (A-C), \u003cem\u003eCulex pipiens/restuans \u003c/em\u003e(D-F) and West Nile virus (G-I), specifically for land cover classes (A, D, G), total daily precipitations averaged from 2003 to 2022 (B, E, H), and mean daily temperature averaged from 2003 to 2022 (C, F, I). Habitat suitability indices were created using mosquito collection data from Public Health Ontario eaHat 3,010 sampling locations from 2003 to 2022 (A-F), and a reduced dataset of 2,810 sampling locations from 2003 to 2022 excluding locations where the two vectors were not observed at all (G-I).\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-4714418/v1/67a1a22b8c9593ef592f72be.png"},{"id":61763117,"identity":"476512e9-83d1-4201-82ff-264a651b7f57","added_by":"auto","created_at":"2024-08-05 09:51:04","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":204800,"visible":true,"origin":"","legend":"\u003cp\u003eMap of rates of West Nile virus (WNV) infection per 100,000 tested blood donations given to the Canadian Blood Services across Ontario, Canada, from 2003 to 2022.\u003c/p\u003e","description":"","filename":"Figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-4714418/v1/437e2f2b68ca6743ff8bce16.png"},{"id":72640528,"identity":"91ded993-ccf9-4758-8315-e8afe0db5ecb","added_by":"auto","created_at":"2024-12-30 16:06:28","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3471353,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4714418/v1/a505bcd9-081e-4c7e-9761-f048d2b50eb8.pdf"},{"id":61763115,"identity":"c9dea3da-c8d6-4e30-90d6-d4a6fb2afbe4","added_by":"auto","created_at":"2024-08-05 09:51:04","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":13390,"visible":true,"origin":"","legend":"","description":"","filename":"TableS1.docx","url":"https://assets-eu.researchsquare.com/files/rs-4714418/v1/f4862a8ab1947b1e9e4f0237.docx"},{"id":61763118,"identity":"016b562a-49c5-48f3-8d74-089c2471ac10","added_by":"auto","created_at":"2024-08-05 09:51:04","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":16168,"visible":true,"origin":"","legend":"","description":"","filename":"TableS2.docx","url":"https://assets-eu.researchsquare.com/files/rs-4714418/v1/f5f2b4c0ce430c2cd13f71e4.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Spatiotemporal analysis of West Nile virus infection in the human population based on arboviral detection testing of blood donations","fulltext":[{"header":"Introduction","content":"\u003cp\u003eHuman infections with West Nile virus (WNV; family Flaviviridae) are most frequently mild or asymptomatic. Symptomatic infections result in an illness referred to as \u0026lsquo;West Nile fever\u0026rsquo;, which, in rare cases, particularly in older age groups, may develop into severe and sometimes fatal neuroinvasive disease \u003csup\u003e1\u003c/sup\u003e. The virus, first introduced to North America in 1999, rapidly spread across the continent in the following years \u003csup\u003e2,3\u003c/sup\u003e, and is now considered an endemic mosquito-borne pathogen in much of North America \u003csup\u003e4,5\u003c/sup\u003e. In Canada, detection tests for WNV have been regularly conducted on most blood donations since 2003, to avoid transmission from donors to recipients \u003csup\u003e6\u003c/sup\u003e. While the blood donor population isn\u0026rsquo;t absolutely synonymous with the general population, data from WNV detection tests on blood donations offer a representative snapshot of spatio-temporal variations in WNV infection risk in the human population\u003csup\u003e7\u003c/sup\u003e. Human case reporting in surveillance is also a useful source of information on WNV risk in North America \u003csup\u003e8\u003c/sup\u003e, despite having some limitations, including underreporting due to asymptomatic infections, misdiagnoses, and uncertainty as to the precise size of the exposed population \u003csup\u003e9,10\u003c/sup\u003e. While such limitations may be avoided by using active pathogen surveillance data on in the human population \u003csup\u003e11\u0026ndash;13\u003c/sup\u003e, data spanning a large area and timeframe are scarce.\u003c/p\u003e \u003cp\u003eMosquito species of the genus \u003cem\u003eCulex\u003c/em\u003e, such as \u003cem\u003eCulex pipiens\u003c/em\u003e and \u003cem\u003eCulex restuans\u003c/em\u003e, are recognized as important vectors maintaining the enzootic cycle of WNV among avian hosts in northeastern North America. \u003cem\u003eCulex pipiens\u003c/em\u003e and \u003cem\u003eCx. restuans\u003c/em\u003e also occasionally feed on mammals, bridging transmission between birds and humans \u003csup\u003e14\u0026ndash;19\u003c/sup\u003e. \u003cem\u003eAedes vexans\u003c/em\u003e is very common in northeastern North America, is competent for the transmission of WNV and is highly opportunistic in its blood feeding preferences, meaning they can also effectively bridge transmission between birds and humans \u003csup\u003e20\u0026ndash;23\u003c/sup\u003e. Environmental factors that influence mosquito development and virus transmission are important for understanding the drivers of human WNV infection risk. Larval development time for \u003cem\u003eCx. pipiens\u003c/em\u003e and \u003cem\u003eCx. restuans\u003c/em\u003e mosquitoes is estimated to be 19 and 14 days on average, respectively, while their adult lifespan is approximately 30 and 27 days, respectively \u003csup\u003e24\u003c/sup\u003e. Development time for \u003cem\u003eAe. vexans\u003c/em\u003e is 10 days on average \u003csup\u003e25\u003c/sup\u003e, and lifespan is 20 days on average \u003csup\u003e23\u003c/sup\u003e. WNV infection can be detected using nucleic acid tests to detect viral RNA by polymerase chain reaction (PCR) during viremia from around Day 1 to Day 13 after bite by an infected mosquito, and using serological tests to detect IgM or IgG by enzyme-linked immunosorbent assay (or ELISA) from around Day 5 to 9 (respectively for IgM and IgG) after bite by an infected mosquito, for approximately 5 to 7 months on average (respectively for IgM and IgG) \u003csup\u003e26\u003c/sup\u003e. Exposure of humans, including blood donors, to infected mosquitoes depends on their abundance and infection prevalence. Mosquito abundance varies seasonally and from year-to-year according to longer term impacts of weather and climate affecting survival over winter and shorter-term effects of weather on mosquito reproduction and activity, and WNV replication in infected mosquitoes \u003csup\u003e27\u0026ndash;32\u003c/sup\u003e. Inter-annual variations in cycles of transmission amongst avian reservoirs, and mosquitoes changing their behaviour to include mammals as sources of blood meals also impact seasonal and inter-annual variations \u003csup\u003e33\u003c/sup\u003e. Together, this means that weather over a period of weeks to months, combined with local environmental impacts on mosquito reproduction and WNV transmission, determine the risk of human infections, and of WNV outbreaks, each year.\u003c/p\u003e \u003cp\u003eCurrent effects of climate change and land use change in Canada and elsewhere may contribute to an increase in the risk of zoonotic diseases, including mosquito-borne pathogens, such as WNV \u003csup\u003e4,34\u0026ndash;40\u003c/sup\u003e. Therefore, there is a pressing need for an enhanced understanding of the environmental hazard posed by WNV for public health. Comprehensive temporal and spatial databases of a large variety of climatic, geographic and ecological factors are needed to correctly characterize the distribution and dynamics of mosquito vector populations and mosquito-borne disease transmission. Additionally, certain demographic and socio-economic factors may contribute to both mosquito vector density and exposure to mosquito-borne disease risk \u003csup\u003e41\u0026ndash;47\u003c/sup\u003e. Indeed, lower socio-economic status seem to be strongly associated with mosquito-borne disease risk, through variation in population density \u003csup\u003e48\u003c/sup\u003e, quality of household infrastructure \u003csup\u003e44\u003c/sup\u003e, and education, potentially associated with level of risk perception \u003csup\u003e46\u003c/sup\u003e. Occupation \u003csup\u003e49\u003c/sup\u003e, age and sex \u003csup\u003e50\u003c/sup\u003e may also influence exposure to mosquitoes. Therefore, regularly updated population censuses offer invaluable information to combine with other data sources to identify modifiable risk factors and better target interventions to reduce mosquito-borne disease risk.\u003c/p\u003e \u003cp\u003eIn this study, we aimed to investigate variables potentially associated with human WNV exposure and infection using a large data collection, established over 20 years, from 2003 to 2022, across most of the province of Ontario, Canada. We combined climatic data, geographic data, mosquito surveillance data, arbovirus testing data from human blood donations, and demographic and socio-economic data from Canadian population censuses. We hypothesized that a mixture of spatial and temporal variables, impacting mosquito habitat and phenology, and demographic and socio-economic variables, which are associated with increased mosquito habitat and also increased exposure to mosquitoes, are the main predictors of WNV infection in the human population.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy area and duration\u003c/h2\u003e \u003cp\u003eThe study area in Ontario, Canada spanned more than 428,000 km\u003csup\u003e2\u003c/sup\u003e with a northern limit (latitude\u0026thinsp;~\u0026thinsp;50\u0026ndash;51\u0026deg; N) that matched the extent of most datasets used in this study (Fig.\u0026nbsp;1). Major urban population centers in our study area are situated in the southeastern section, where a warmer agricultural and residential landscape predominates, and which represents the most populated corridor in Canada from the cities of Windsor to Ottawa. The northern section is dominated by a mixture of forests, wild grasslands and wetlands, with colder temperatures, and ranging from dry in the west at the border with the province of Manitoba, to humid in the center and towards the border with the province of Qu\u0026eacute;bec in the east (Fig.\u0026nbsp;1).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eMosquito collection data\u003c/h2\u003e \u003cp\u003eBased on the available mosquito surveillance data from all 34 Ontario public health units, sampling sites (n\u0026thinsp;=\u0026thinsp;3,010) were visited at least once, and at most 876 times, from 2003 to 2022. On average, each site was visited 48 times across the study period (Table\u0026nbsp;1). Sampling sites were visited mostly from May to October when both \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e and \u003cem\u003eAe. vexans\u003c/em\u003e were most likely to be observed in the study area. Out of a total of 145,102 sampling site visits, 99.9% were made from May to October. About twice as many \u003cem\u003eAe. vexans\u003c/em\u003e than \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e individuals were collected, in total and on average per sampling site visit. Around 60% of sampling visits yielded each species on average (Table\u0026nbsp;1). Average abundance and occurrence per sampling site visit generally decreased across years for \u003cem\u003eAe. vexans\u003c/em\u003e, but not for \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e (Fig.\u0026nbsp;2). Average abundance and occurrence per sampling site visit increased from May to August, and then decreased in the following months to October (Fig.\u0026nbsp;2). The number of mosquito pools tested for West Nile virus (WNV) was higher for \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e due to higher priority of \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e for WNV testing, and the total number of positive pools was also higher for \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e (Table\u0026nbsp;1). However, the total number of \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e individuals tested was lower, which is due to fewer collected \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e individuals for that species on average compared to \u003cem\u003eAe. vexans\u003c/em\u003e, and minimum infection rate (MIR) was much higher for \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e (Table\u0026nbsp;1). There was an overall higher MIR infection rate from 2013 to 2022 (1.9), compared to 2003 to 2012 (1.3), where 2012 was the year with the highest MIR (7.9; Fig.\u0026nbsp;2). The MIR was also much higher in August and September, compared to July and October (Fig.\u0026nbsp;2). No WNV-positive mosquito pools were observed from November to May of any year.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eLand cover and climatic data\u003c/h2\u003e \u003cp\u003eLand cover data collated from Agriculture Canada and United States Geological Survey consisted of a total of 10 classes, namely two classes of humid cover (open water and wetlands), three classes of open cover (natural and anthropogenic vegetated, and barren), one class of forested cover, and four classes of residential cover (containing varying degrees of vegetation cover), at a resolution of 120 meters. The predominance of each class varied considerably across the study area (Fig.\u0026nbsp;1). The four most common landscape classes containing a mosquito sampling site were the medium green residential, high green residential, agricultural cropland and forested classes, and the three least common were non-vegetated residential, exposed, and wetlands (Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e; Fig.\u0026nbsp;1). The 20-year average total daily precipitation varied between 1.7 to 4.0 mm across the study area. The western part of the study area received less precipitation on average, while a few areas in the center and center-east received more precipitation on average (Fig.\u0026nbsp;1). The 20-year average mean daily temperature varied between 0.7 and 10.4\u0026deg;C, and the southern part of the study area was much warmer on average than the northern part, with the warmest area being the corridor from Windsor to Toronto in the extreme south near Lake Erie, and the coldest area being the center north and northeast of the study area (Fig.\u0026nbsp;1). Average total precipitation and mean temperature at cells containing a sampling site were 2.6 mm and 7.5\u0026deg;C, respectively (Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eEcological niche modeling analysis\u003c/h2\u003e \u003cp\u003eThe correlation coefficient of 20-year average total daily precipitation and 20-year average mean daily temperature was low (\u003cem\u003er\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.1). Overall, the receiver operating characteristic\u0026rsquo;s area under the curve (AUC) values were lower for \u003cem\u003eAe. vexans\u003c/em\u003e models compared to \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e and West Nile virus models. Out of 100 models, the best performing model had an AUC\u0026thinsp;=\u0026thinsp;0.82 for Cx. \u003cem\u003epipiens/restuans\u003c/em\u003e, 0.79 for West Nile virus, and 0.72 for \u003cem\u003eAe. vexans\u003c/em\u003e (Table\u0026nbsp;2). The ensemble model contained 100, 100 and 8 models, respectively for WNV, \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e, and \u003cem\u003eAe. vexans\u003c/em\u003e, with a mean AUC higher than 0.9 for mean habitat suitability index (HSI) values and committee averaging for all species (Table\u0026nbsp;2). Explanatory variable importance (on a scale of 0 to 1, obtained from 100 permutations of the ensemble niche model) of temperature in the ensemble model was higher than 0.6 and higher than the importance of the other two explanatory variables for both mosquito species and WNV. The importance of precipitation was nearly the same as that of temperature for \u003cem\u003eAe. vexans\u003c/em\u003e, but was low, at less than 0.2, for \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e and moderate, at less than 0.4, for West Nile virus. Comparatively, land cover was of moderate importance for all species, at less than 0.4 (Table\u0026nbsp;2).\u003c/p\u003e \u003cp\u003eProjected HSI of the best performing model for all species led to an area of highest HSI value (100, 924 and 883, for WNV, \u003cem\u003eAe. vexans\u003c/em\u003e and \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e, respectively, with highest possible maximum value of 1000) concentrated in the extreme southern part of the study area for all species. This was centered around the corridor from Windsor to Toronto near Lake Erie (Fig.\u0026nbsp;1) for both mosquito species, and was confined to major urban centers for WNV. Another area of high HSI was present for both mosquito species in the southeast around the cities of Kingston and Ottawa near Lake Ontario, where HSI was overall higher for \u003cem\u003eAe. vexans\u003c/em\u003e than for \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e. \u003cem\u003eAe. vexans\u003c/em\u003e also displayed high HSI in areas where \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e did not, particularly in the center and center-east of the study area north of Lake Huron, and in several areas in the extreme west of the study area west of Lake Superior (Fig.\u0026nbsp;1).\u003c/p\u003e \u003cp\u003eMean HSI and committee averaging showed high suitability of wetlands and two residential land cover classes, low green and medium green, for both mosquito species and WNV (Fig.\u0026nbsp;3). Wild grasslands had high suitability for the two mosquito species, but low for WNV. High green residential showed high suitability only for \u003cem\u003eAe. vexans\u003c/em\u003e. Open water, forested, and exposed land cover classes showed low suitability for WNV (Fig.\u0026nbsp;3). Other land cover classes either showed low confidence, as displayed by values near 0.5, or conflict between mean HSI and committee averaging values (Fig.\u0026nbsp;3). Mid-range values of total daily precipitation, i.e. between roughly 2.5 and 3.0 mm, were associated with the highest suitability values for both mosquito species, and lowest values, below 2.5 mm, were associated with the highest suitability values for WNV, according to both mean HSI and committee averaging (Fig.\u0026nbsp;3). Higher mean daily temperature values, i.e. above roughly 6\u0026deg;C for both mosquito species and 9\u0026deg;C for WNV, were associated with the highest suitability values, according to both mean HSI and committee averaging (Fig.\u0026nbsp;3).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eSocio-economic and demographic data\u003c/h2\u003e \u003cp\u003eUsing Statistics Canada census datasets from 2016, we identified a total of 14 demographic and socio-economic variables of interest in our analyses, at the level of the census subdivision. One variable captured information on the sex ratio (percent male residents); three variables on age structure (percent residents younger than 15, percent residents older than 64, and percent residents older than 84); two variables on ethnicity (percent residents self-identified as immigrant, percent residents self-identified as Indigenous); one variable on population density (number of residents per km\u003csup\u003e2\u003c/sup\u003e); two variables on income (mean income, and percent of population earning less than \u003cspan\u003e$\u003c/span\u003e20,000), two variables on shelter (percent shelters needing major repairs, percent residents spending 30% or more of their income on shelter); one variable on education (percent residents with no secondary education); and two variables on occupation (percent residents working in trades, and percent residents working with natural resources).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eBlood donation arboviral testing data\u003c/h2\u003e \u003cp\u003eA total of more than 6.5\u0026nbsp;million blood donations from more than 900,000 donors were tested by Canadian Blood Services across the province of Ontario, Canada, from 2003 to 2022 (Table\u0026nbsp;3). Many donations were coming from the same donor multiple times throughout the study period. We hereby refer use \u0026lsquo;donor\u0026rsquo; to refer to each unique individual who donated blood either once or multiple times, and the term \u0026lsquo;donation\u0026rsquo; to reflect each individual donation given. There were more donations from male than female donors, and more donations from donors aged 39 to 65 compared to younger and older (Table\u0026nbsp;3). However, there were slightly more individual female donors than males, and slightly more individual donors younger than 39 compared to 39 and older. Donors aged older than 65 represented the lowest number of individual donors, among all age groups (Table\u0026nbsp;3).\u003c/p\u003e \u003cp\u003eThere was a total of 102 donations with a positive test result for WNV infection based on nucleic acid testing. The bulk of WNV-positive blood donations were from male donors and from donors aged 39 to 65. The cumulative WNV infection rate was slightly higher for male than female donors, and much higher for donors older than 65 than for other age groups (Table\u0026nbsp;3). The WNV infection rate was also much higher in some census subdivisions and some years compared to others. There was an overall higher WNV infection rate in southern Ontario (Fig.\u0026nbsp;4). There was also an overall higher WNV infection rate from 2013 to 2022 (2.7 per 100,000), compared to 2003 to 2012 (0.8 per 100,000), where 2018 was the year with the highest WNV infection rate (10.6 per 100,000; Fig.\u0026nbsp;2). The incidence of WNV infection amongst blood donors was also much higher in August and September, compared to July and October (Fig.\u0026nbsp;2). No positive WNV blood sample was observed from November to June of any year.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eSpatiotemporal WNV infection analyses\u003c/h2\u003e \u003cp\u003eAt the analysis level of the census subdivision, we considered a total of 30 variables in negative binomial regression models to identify variables associated with WNV incidence: 14 demographic and socio-economic variables, 10 land cover variables, two climatic variables, two mosquito vector habitat suitability index variables, one WNV habitat suitability index variable, and census subdivision area (Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). Two groups of variables had \u003cem\u003er\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.7. Population density and all residential land cover classes were highly correlated, and therefore we chose to only keep population density, and dropped the four residential land cover variables. Mean daily temperature and habitat suitability index of both \u003cem\u003eAe. vexans\u003c/em\u003e and \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e were highly correlated, so we chose to keep only habitat suitability index of \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e, and dropped the temperature variable and habitat suitability index of \u003cem\u003eAe. vexans\u003c/em\u003e. Six variables were dropped due to absence of significant association with WNV infection in univariable models (Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). We ran a multivariable model with the remaining 18 variables. After model selection, the final model contained thirteen variables, three of which had an IRR that was significantly different from 1 (Table\u0026nbsp;4). The proportion of households earning less than \u003cspan\u003e$\u003c/span\u003e20,000 (low income) had an IRR around 2.8 (Table\u0026nbsp;4), suggesting a strong positive association of this variable with human WNV infection. The proportion of residents spending 30% or more of their income on their shelter (high spending on shelter) had an IRR around 0.7 (Table\u0026nbsp;4), suggesting a strong negative association of this variable with WNV infection. Habitat suitability index of \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e had an IRR value around 2.0 (Table\u0026nbsp;4), suggesting a strong positive association of this variable with WNV infection. All other variables did not display a significant association with WNV infection. The R\u003csup\u003e2\u003c/sup\u003e value of the final model was around 0.67, which suggests strong statistical power in the final model.\u003c/p\u003e \u003cp\u003eAt the analysis level of the individual blood donation, we considered a total of 4 variables in logistic regression models to identify factors associated with WNV infection: donor age and sex, and year and month of detection test (Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). No pair of variables had \u003cem\u003er\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.7. One variable was dropped due to absence of significant association with WNV infection in univariable models (Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). The final model contained three variables, all of which had a statistically significant IRR, with values very close to 1 (Table\u0026nbsp;4). Donor age, year of detection test and month of detection test had an IRR value between 1.002 and 1.003 (Table\u0026nbsp;4), suggesting a positive association of these variables with WNV infection. The R\u003csup\u003e2\u003c/sup\u003e value of the final model was lower than 0.001, which suggests poor statistical power in the final model, potentially stemming from extremely low variation in the outcome variable, i.e. small number of cases.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eOur study on spatiotemporal effects of climatic, geographic, ecological, demographic and socio-economic variables on West Nile virus (WNV) infection in the human blood donor population identified multiple modifiable and non-modifiable risk factors that may be practically useful to inform disease prevention and control efforts. First, we identified most regions of Southern Ontario along Saint Lawrence River, Lake Ontario, Lake Erie and south of Lake Huron, to be the main habitat for both \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e and \u003cem\u003eAe. vexans\u003c/em\u003e, which is also where climate is warmest, somewhat more humid and mostly agricultural and urban. Habitat for WNV itself was much narrower, being confined to the Greater Toronto Area and the Windsor region. Some regions in Northern Ontario were also suitable for \u003cem\u003eAe. vexans\u003c/em\u003e. Positive WNV cases in the human blood donor population were mostly detected in Southern Ontario, with very few cases in Northern Ontario. Across years, the abundance and occurrence of \u003cem\u003eAe. vexans\u003c/em\u003e tended to decrease, but remained mostly unchanged for \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e. We observed a peak during the month of August in abundance and occurrence of \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e and \u003cem\u003eAe. vexans\u003c/em\u003e, the two main WNV vectors in our study area, and mosquito vector WNV infection prevalence. Positive mosquito vector pools and infection cases in the blood donor population mostly occurred in the second half of the study period, during the months of August and September. These results mostly support previous literature \u003csup\u003e40,51\u0026ndash;57\u003c/sup\u003e. However, our study sheds light on the spatiotemporal interplay in abundance and/or occurrence between the two main WNV vectors in northeastern North America, and how this affects WNV infection in the vector populations and in the human population, using longitudinal surveillance data over long timeframe and over a large study area. In addition, our study leverages blood donor testing data to provide spatiotemporally widespread arboviral detection in the Ontario population, which to date has not been investigated to such an extent in Canada, as opposed to other countries such as the United States \u003csup\u003e7\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eThe importance of climatic and geographic variables in the ecological niche modeling analyses for WNV vector species showed similar results from a recent study in eastern Ontario, which performed similar analyses at a smaller but overlapping spatial and temporal scale using mosquito surveillance data from 2011 to 2020 \u003csup\u003e56\u003c/sup\u003e. Whereas the previous study identified a mediocre to weak importance of all variables, our study herein suggests a strong effect of temperature for both mosquito vector species and for WNV, and a strong effect of precipitation for \u003cem\u003eAe. vexans\u003c/em\u003e. This is to be expected given the larger climatic variation across this study area compared to the previous study. In this study, the threshold of temperature on habitat suitability for WNV was higher compared to that of the two mosquito species. Optimal amounts of precipitation were also lower compared to those of the two mosquito species alone. Interestingly, land cover had a moderate relative importance for both species, despite land cover classes being all relatively well represented in terms of sampling sites in the present study. Associations between habitat suitability and specific land cover classes were mostly similar between \u003cem\u003eAe. vexans\u003c/em\u003e, \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e and WNV, except for wild grasslands which were unsuitable for WNV, and high levels of vegetation in urban landscape are not suitable for \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e and WNV. Finally, urban landscapes with low to moderate levels of vegetation and wetlands were suitable for both mosquito species and for WNV. These results are largely supported by previous literature, where high temperatures, and vegetated urban and wetland cover were predictors of both WNV vector habitat and WNV transmission \u003csup\u003e55,56,58\u0026ndash;61\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eOur statistical models identified habitat suitability index for \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e as a strong predictor of WNV infection in the human blood donor population. Due to their strong correlation with habitat suitability index for \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e, habitat suitability index for \u003cem\u003eAe. vexans\u003c/em\u003e and averaged mean daily temperature couldn\u0026rsquo;t be included in multivariable statistical models at the level of the census subdivision. However, they are similarly strong predictors of WNV infection. Canada is currently experiencing effects of climate change on vector-borne disease risk, including WNV \u003csup\u003e4,34\u0026ndash;40\u003c/sup\u003e. Given the strong effect of temperature on mosquito WNV vector habitat, which in turn affects WNV infection in the human population, rigorous surveillance of southerly locations is needed to effectively predict large upticks in human WNV cases across Canada.\u003c/p\u003e \u003cp\u003eOur statistical models identified habitat suitability index for \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e as a strong predictor of WNV infection in the human blood donor population, which may also be expanded to two highly correlated variables: habitat suitability index for \u003cem\u003eAe. vexans\u003c/em\u003e and averaged mean daily temperature. Our statistical models also identified the proportion of low-income households and proportion of households where 30% or more of the residents\u0026rsquo; income is spent on shelter being positively and negatively, respectively, associated with WNV infection in the human population. Household wealth is widely known to be associated with mosquito vector density, potentially through greater perception of risk and access to mosquito control methods \u003csup\u003e44,46,62,63\u003c/sup\u003e. In a previous analysis at a smaller but completely overlapping spatial and temporal scale in the city of Ottawa, Ontario, from 2007 to 2014, proportion of 60-years-old-and-older shelters was associated with higher WNV risk. A similar result was also observed in Chicago during a WNV outbreak in 2002 \u003csup\u003e63\u003c/sup\u003e. This variable is possibly linked to availability of suitable habitats, i.e. breeding sites in suboptimal drainage systems \u003csup\u003e64\u003c/sup\u003e, used by WNV vector species. Population density and an urban environment were previously associated with higher WNV risk in Ottawa, Ontario, from 2007\u0026ndash;2014 \u003csup\u003e55\u003c/sup\u003e, and in Chicago and Detroit during a WNV outbreak in 2002 \u003csup\u003e59\u003c/sup\u003e. However, in our study here there was no significant effect of either population density or residential land cover on WNV infection in the human population, despite using extensive and high-resolution datasets of both types of variables. Certain residential land cover classes do seem to affect habitat suitability of the two mosquito WNV vector species and WNV itself, albeit with an importance moderate or low compared to weather variables, but this effect does not seem to translate to significantly higher WNV infection in humans, as demonstrated by our multivariable statistical models. These results either suggest limited power to identify an effect of population density or residential land cover by our multivariable statistical models, or that spurious associations in previous studies would have been better explained by alternative unmeasured variables that have been included in this study herein.\u003c/p\u003e \u003cp\u003eOur study has a few limitations. Nucleic acid tests we used have a possibility of cross-reacting with other members of the Japanese encephalitis serocomplex, such as Japanese encephalitis, Murray Valley encephalitis, Saint Louis encephalitis and Kunjin virus \u003csup\u003e65\u0026ndash;68\u003c/sup\u003e (leading to a modification to the \u0026ldquo;National case definition: West Nile virus \u0026ndash; Canada.ca\u0026rdquo; in 2024), which is relevant for donors with a certain travel history or due to vaccination with Japanese encephalitic virus vaccine. Also, most data points in the mosquito sampling data and the human donor data are highly aggregated around urban and metropolitan areas of the study area. However, outcomes of these limitations on the main results of our analyses are likely to be minimal.\u003c/p\u003e \u003cp\u003eIn conclusion, results from our study point to several modifiable risk factors that may be used as points of entry for practical intervention aimed at reducing risk of mosquito-borne pathogens in Canada, in a context of increasing mosquito-borne pathogen exposure and illness. For example, our study supports the need for government education campaigns and incentives facilitating renovations aimed at reducing mosquito habitat and/or exposure to mosquitoes, especially in areas with highly suitable habitat for \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e in Southern Ontario, such as the \u0026ldquo;Prevention of West Nile virus \u0026ndash; Canada.ca\u0026rdquo; program. Such programs and incentives would also be useful in practical intervention measures against \u003cem\u003eAedes albopictus\u003c/em\u003e, an invasive mosquito vector for several pathogens, including dengue virus, which is detected periodically in Southern Ontario. Our study is one of the few using arboviral detection tests over a large area and a large period in the human population to identify factors predictive of WNV infection. Such studies are not impeded by the same spatial, temporal and clinical (e.g., under reporting) biases as those using disease case reports, but are uncommon due to the sheer amount of work needed to produce representative databases.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eStudy area and duration\u003c/h2\u003e \u003cp\u003eThe study area is situated across most of the province of Ontario, in Canada, at and below latitudes 50\u0026ndash;51\u0026deg; N (Fig.\u0026nbsp;1). The study spans the years 2003 to 2022.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eMosquito collection data\u003c/h2\u003e \u003cp\u003eSince 2002, Public Health Ontario (PHO) has collected data on mosquito surveillance activities conducted by public health units in Ontario, which comprise mosquito capture and identification of 22 species/species groups. Capture is performed using the protocol as described in Talbot et al. (2023) \u003csup\u003e56\u003c/sup\u003e. Briefly, \u003cem\u003eCulex pipiens/restuans\u003c/em\u003e are prioritized over all other species, and \u003cem\u003eAe. vexans\u003c/em\u003e is also of high priority. RNA is extracted from each pool using RNeasy Mini Kit (Qiagen, Hilden, Germany). RNA extracts are tested by quantitative polymerase chain reaction (PCR) for arbovirus presence \u003csup\u003e69\u003c/sup\u003e. We used all mosquito capture data available across the entire province of Ontario, Canada, comprising data from all 34 public health units, for the years 2003 to 2022. For each mosquito pool, we calculated the WNV minimum infection rate (MIR), which is the test outcome (positive\u0026thinsp;=\u0026thinsp;1; negative\u0026thinsp;=\u0026thinsp;0), divided by the number of mosquitoes present in the pool, multiplied by 1000.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eLand cover and climatic data\u003c/h2\u003e \u003cp\u003eWe followed the approach of Talbot et al. (2023) \u003csup\u003e56\u003c/sup\u003e to process land cover and climate data for subsequent ecological niche models. Land cover data for the year 2013 were obtained from Agriculture and Agri-Food Canada (AAFC) and the United States Geological Survey (USGS) with a resolution of 30 meters across our study area. Land cover data for the year 2013 were chosen because it is approximately the mid-point of our study period from 2003 to 2022. Annual crop inventory data from AAFC \u003csup\u003e70\u003c/sup\u003e comprise seven land cover classes: open water, wetlands, agricultural croplands, natural grasslands, forests, exposed surface and residential areas. Residential areas were subdivided into four categories, according to the normalized difference vegetation index (NDVI) from USGS, created using Landsat 8 data (collection 2, level 2, maximum 50% clouds) from May to October 2013, from USGS \u003csup\u003e71,72\u003c/sup\u003e. The goal of this procedure was to subdivide urban environments according to the presence of vegetation, which may affect habitat selection by the studied species. Residential areas with NDVI\u0026thinsp;\u0026lt;\u0026thinsp;=\u0026thinsp;0.15 were referred to as non-vegetated, with NDVI\u0026thinsp;\u0026gt;\u0026thinsp;0.15 and \u0026lt;\u0026thinsp;=\u0026thinsp;0.30 as low green, with NDVI\u0026thinsp;\u0026gt;\u0026thinsp;0.30 and \u0026lt;\u0026thinsp;=\u0026thinsp;0.60 as medium green and with NDVI\u0026thinsp;\u0026gt;\u0026thinsp;0.60 as high green residential areas. Data from AAFC and USGS used the same resolution with matching cell frames, and therefore merging of the two datasets could be performed manually.\u003c/p\u003e \u003cp\u003eTherefore, we considered temperature and precipitation data across the entire study duration. Temperature and precipitation data were obtained from the National Aeronautics and Space Administration (NASA). Given their ease of access, we extracted data on annual total precipitation, and annual maximum and minimum temperature from the annual surface weather and climatological summaries from NASA \u003csup\u003e73\u003c/sup\u003e with a resolution of 1,000 meters. These data were averaged across the 20 years of the study duration, and then divided by number of days in a year to obtain 20-year averaged mean daily temperature and 20-year averaged total daily precipitation for the total period across our study area.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eEcological niche modeling analysis\u003c/h2\u003e \u003cp\u003eWe followed the approach of Talbot et al. (2023)\u003csup\u003e56\u003c/sup\u003e. The two studied mosquito species were observed at least once in the vast majority of the 3,010 sampling locations over the study period (2,783 for \u003cem\u003eAe. vexans\u003c/em\u003e, or 92%, and 2,686 for \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e, or 90%). At least one mosquito vector species was present in a total of 2,810 sites, for which a WNV detection test could be performed, and lead to at least one positive outcome in a total of 650 sites. Therefore, the number of visits in which each species was observed at least once at each site was calculated, and divided by the total number of visits at each site over the study period. The resulting value is the frequency of observed presence, and is a form of aggregated performance measure often used in species distribution models \u003csup\u003e74\u003c/sup\u003e. Sites were then dichotomized into a 0/1 distribution for each mosquito species and for WNV, which is a requirement of the approach.: Sites with 50% or more species occurrence for both mosquito species were classified as \u0026lsquo;presence, and sites with less than 50% as \u0026lsquo;absence. Sites with at least one positive test outcome for WNV were classified as \u0026lsquo;presence\u0026rsquo;, and others as \u0026lsquo;absence\u0026rsquo;.\u003c/p\u003e \u003cp\u003eThe random forest algorithm was considered suitable for our analyses because the presence and absence of a species in a given sampling site visit are likely to be affected by the same sampling bias \u003csup\u003e75,76\u003c/sup\u003e. This decision tree-based approach performs as well as the maximum entropy approach \u003csup\u003e77,78\u003c/sup\u003e, and better than traditional regression-based approaches when using large datasets sampled over a long duration and a large spatial scale \u003csup\u003e79\u003c/sup\u003e. We performed the analyses using the \u0026lsquo;biomod2\u0026rsquo; package \u003csup\u003e80\u003c/sup\u003e in R 4.2.1 (R Development Core Team, Vienna, Austria). We projected all land cover and climatic datasets to Albers Conic Equal Area, which was the original projection of the land cover dataset from Agriculture Canada. All explanatory variables were resampled to a cell size of 120 x 120 meters and set to be at the same cell frame to reduce spatial bias caused by unequal resolution with the mosquito dataset \u003csup\u003e75\u003c/sup\u003e. We computed the Pearson\u0026rsquo;s correlation coefficient among climatic variables at sampling sites in R 4.2.1. Potential collinearity problems were considered if \u003cem\u003er\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.7. We set the prevalence parameter to 0.5 was specified, meaning \u0026ldquo;presence\u0026rdquo; and \u0026ldquo;absence\u0026rdquo; distributions are considered in equal proportions in the analysis \u003csup\u003e81\u003c/sup\u003e. For each species, we trained 100 replicate models using 80% of data. To evaluate each model, we computed a receiver operating characteristic\u0026rsquo;s area under the curve (AUC) using the remaining 20% of data. Data was selected randomly in each model for training versus testing. We used the final model, trained by the 100 replicate models and using 100% of data, to generate a habitat suitability index (HSI) map in the study area. We kept all other parameters at default values. We used all models with AUC above 0.7 to generate an ensemble niche model \u003csup\u003e80\u003c/sup\u003e. We generated response plots of the mean HSI across models and committee averaging \u003csup\u003e82\u003c/sup\u003e, for each explanatory variable. We calculated variable importance for each explanatory variable, which varies from 0 to 1, using a procedure of 100 permutations from the ensemble niche model. Lastly, we created a projected HSI map from the model with the highest AUC for each mosquito species and for WNV.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eSocio-economic data\u003c/h2\u003e \u003cp\u003eStatistics Canada collects data on a large variety of socio-economic variables during the Census of Population every five years across Canada. Data for the years 2011 and 2016 were initially considered because they are near the mid-point of our study period from 2003 to 2022. Given their easier access and processing, we chose data for the year 2016, which we extracted using the Beyond 20/20 Professional Browser software. We selected all variables that relate to either sex, age, ethnicity, population density, income, shelter infrastructure, education, and occupation, to be included in analyses on the determinants of WNV infection in the human population. We uploaded data at the level of the census subdivision, to match blood donation arboviral testing data structure.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eBlood donation arboviral testing data\u003c/h2\u003e \u003cp\u003eAcross the study area, since 2003, Canadian Blood Services (CBS) have tested blood donations across Ontario. These tests were performed on all blood donations from 2003 to 2015. As of December 2015, all donations from June to November were tested, but only donations from travelers to certain destinations were tested from December to May. These months usually see much less mosquito activity, and therefore likelihood of exposure to an infected mosquito is negligible. Groups of six donations were tested in 6-unit minipools. Positive minipools were then retested separately for each donation from the corresponding minipool, along with all donations from surrounding areas for the next two weeks. From June 2003, testing was performed using TaqScreen WNV test IUO (F. Hoffmann-La Roche AG, Basel, Switzerland). From June 2007, testing was performed using the IND cobas TaqScreen WNV test for use with the cobas s 201 system. From June 2008, testing was performed using the licensed cobas TaqScreen WNV test for use with the cobas s 201 system. From December 2017, testing was performed using the cobas\u0026reg; WNV \u0026ndash; Nucleic acid test for use on the cobas\u0026reg; 6800/8800 Systems. At donor registration, the donor\u0026rsquo;s date of birth, sex and residential address are recorded. For confidentiality purposes, we used only the donor\u0026rsquo;s reported sex; year of birth, to obtain approximative age of the blood donor at the time of donation; census subdivision of residence from the Canadian Census of Population boundaries for the year 2016; and unique donor identifier.\u003c/p\u003e \u003cp\u003eTo avoid sampling biases leading to spurious associations in our analyses of determinants of WNV infection in the human population, we chose to exclude all data from census subdivisions where fewer than 1000 WNV tests on blood donations were available, which are mostly located in sparsely populated areas mostly in northeastern parts of the study area, and north of the northern limit of the study area. This exclusion step lead to the removal of 126 subdivisions out of the total of 417 (30%) where blood donor data is available. In the remaining 291 subdivisions, we chose to exclude 4 census subdivisions where total number of residents was lower than 500, because most socio-economic variables were missing for confidentiality reasons. The final dataset contained 287 census subdivisions. Two of these census subdivisions, named \u0026ldquo;Kenora, Unorganized\u0026rdquo; and \u0026ldquo;Thunder Bay, Unorganized\u0026rdquo; had boundaries above the northern limit of the study area. However, the large majority of blood donations in these subdivisions were conducted in the south, near the Kenora and Thunder Bay townships, respectively, which is also where the large majority of the residents of these subdivisions live. For these reasons, we considered only the portion in these two subdivisions that lie within our study area (Fig.\u0026nbsp;1, Fig.\u0026nbsp;4).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e\u003cem\u003eSpatiotemporal WNV infection analyses\u003c/em\u003e\u003c/h2\u003e \u003cp\u003eTo attain our main objective, we investigated the effect of a wide range of factors related to land cover, climate, mosquito habitat and socio-economic status of residents on WNV infection in the human population, using data from WNV tests on blood donations. These factors were carefully chosen \u003cem\u003ea priori\u003c/em\u003e to include variables most likely affecting WNV risk in the human population, including aspects related to mosquito occurrence, mosquito activity, and characteristics of residents linked to higher exposure to mosquito bites. We conducted these analyses as two different levels: one at the census subdivision, and one at the individual blood donation.\u003c/p\u003e \u003cp\u003eFor the first level of analysis, we gathered all arboviral testing data from blood donations, for which residence information is available, from 2003 to 2022, and grouped them within the respective census subdivision of residence of the blood donor. We conducted a simple generalized linear regression, using the \u0026lsquo;lme4\u0026rsquo; package \u003csup\u003e83\u003c/sup\u003e in R 4.2.1, for each variable separately, namely selected socio-economic variables, land cover variables (percent cover of each class in census subdivisions), climatic variables (20-year averaged daily total precipitations and 20-year averaged daily mean temperature, averaged across census subdivisions), habitat suitability index variables (for each mosquito vector species and for WNV, averaged across census subdivisions), and lastly, the census subdivision geographic area (in km\u003csup\u003e2\u003c/sup\u003e), which can be stochastically associated with number of positive cases. Prior to these analyses, we subtracted the mean and divided by the standard deviation of all values for numerical variables, namely test outcome, donor age, month of test, and year of test. We used the zero-inflated negative binomial modeling family, using the \u0026lsquo;pscl\u0026rsquo; package \u003csup\u003e84\u003c/sup\u003e in R 4.2.1, where we used the number of positive cases within census subdivisions as outcome variable, and number of donations tested within census subdivisions as offset variable.\u003c/p\u003e \u003cp\u003eFor the second level of analysis, we gathered all WNV testing data from blood donations, for which residence information is available, from 2003 to 2022. We conducted a simple mixed-effects generalized linear regression, using the \u0026lsquo;lme4\u0026rsquo; package \u003csup\u003e83\u003c/sup\u003e in R 4.2.1, for each variable separately, namely donor sex, donor age, month of test, and year of test, against test outcome, where 1 is positive and 0 is negative. We used a hierarchical random-effects term, which is the unique donor number nested within the census subdivision of residence. Prior to these analyses, we subtracted the mean and divided by the standard deviation of all values for numerical variables, namely test outcome, donor age, month of test, and year of test. Given the data structure is binomial, the \u0026lsquo;binomial\u0026rsquo; modeling family was most intuitive, but the extremely small number of positive cases compared to negative cases lead to numerous fitting errors. Therefore, we chose the more general \u0026lsquo;gaussian\u0026rsquo; modeling family, for which no modelling problem occurred.\u003c/p\u003e \u003cp\u003eFor both analysis levels, we used Pearson\u0026rsquo;s correlation coefficient in R 4.2.1, to identify correlation among all selected variables, and dropped one variable from each pair of variables displaying \u003cem\u003er\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.7. Next, we selected all variables displaying a significant univariable association (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01), with WNV infection into a multivariable generalized linear regression analysis, using the same packages as previously in R 4.2.1. For the analysis at the level of the census subdivision, given the large number of variables included, we proceeded with a two-step model selection approach using the Bayesian Information Criterion (BIC), i.e. the Akaike Information Criterion using the logarithm of the number of observations as the k parameter. We computed a BIC value for the full model and for all combinations of the full model excluding one variable. A BIC value decrease of 2 or more is considered positive evidence for a variable displaying little effect on the response variable \u003csup\u003e85\u003c/sup\u003e. We ran a new multivariable regression model, but this time excluding all variables that, when dropped from the full model, caused a decrease of the BIC value of more than 2. We then reapplied the same model selection approach as a second step on this new model. Using this model selection approach, we ensured that only the most important variables were retained in the resulting final model. In both analysis levels, we ran the final model using the full dataset, and computed incidence rate ratio (IRR) values, i.e. the exponents of the slope coefficients, 95% confidence intervals for the IRR values, and \u003cem\u003eP\u003c/em\u003e values for each variable. Finally, we computed the R\u003csup\u003e2\u003c/sup\u003e of the final model, using the \u0026lsquo;modelsummary\u0026rsquo; package \u003csup\u003e86\u003c/sup\u003e in R 4.2.1.\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank Canadian Blood Services (in particular Bill Ferguson), Public Health Ontario (in particular Curtis Russell), Statistics Canada, Agriculture and Agri-Food Canada, United States Geological Survey National Aeronautics and Space Administration, for access to blood donor, mosquito capture, socio-economic, geographic and climatic data used in this study. We also thank Maxime Rioux-Rousseau for their help collecting and processing geographic and climatic data sources. This project was supported by grants from the Canadian Institutes of Health Research and Natural Science and Engineering Research Council of Canada to M.A.K.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor\u003c/strong\u003e \u003cstrong\u003econtribution\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eB.T. contributed to the conception of the study, performed analyses, interpreted results, and wrote the manuscript draft. M.A.K. and A.L. contributed in the conception of the study, interpreted results and edited the manuscript draft. N.H.O., S.F.O. and S.J.D. contributed in the conception of the study and edited the manuscript draft.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Blood donor data from Canadian Blood Services is not available due to privacy reasons according to our data sharing agreement between University of Ottawa and Canadian Blood Services. All other data are freely available on the webpage of the relevant authority: Public Health Ontario (https://www.publichealthontario.ca/en/Data-and-Analysis/Using-Data/Data-Requests) for mosquito capture data, Statistics Canada (https://www150.statcan.gc.ca/n1/en/type/data) for socio-economic data, Agriculture and Agri-Food Canada (https://www.agr.gc.ca/atlas/apps/aef/main/index_en.html?AGRIAPP=23) and United States Geological Survey (https://earthexplorer.usgs.gov/) for geographic data, and National Aeronautics and Space Administration (https://www.earthdata.nasa.gov/) for climatic data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAdditional Information\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe declare no competing interests.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eGubler, D. J. The continuing spread of West Nile Virus in the western hemisphere. \u003cem\u003eClinical Infectious Diseases\u003c/em\u003e \u003cstrong\u003e45\u003c/strong\u003e, 1039\u0026ndash;1046 (2007).\u003c/li\u003e\n\u003cli\u003eReisen, W. K. Ecology of West Nile virus in North America. \u003cem\u003eViruses\u003c/em\u003e \u003cstrong\u003e5\u003c/strong\u003e, 2079\u0026ndash;2105 (2013).\u003c/li\u003e\n\u003cli\u003eChancey, C., Grinev, A., Volkova, E. \u0026amp; Rios, M. The global ecology and epidemiology of West Nile Virus. \u003cem\u003eBioMed Research International\u003c/em\u003e \u003cstrong\u003e2015\u003c/strong\u003e, 1\u0026ndash;20 (2015).\u003c/li\u003e\n\u003cli\u003eChen, C. \u003cem\u003eet al.\u003c/em\u003e Climate change and West Nile virus in a highly endemic region of North America. \u003cem\u003eIJERPH\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 3052\u0026ndash;3071 (2013).\u003c/li\u003e\n\u003cli\u003eShing, E. \u003cem\u003eet al.\u003c/em\u003e The direct healthcare costs attributable to West Nile virus illness in Ontario, Canada: a population-based cohort study using laboratory and health administrative data. \u003cem\u003eBMC Infect Dis\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e, 1059 (2019).\u003c/li\u003e\n\u003cli\u003eTodoric, D. \u003cem\u003eet al.\u003c/em\u003e An overview of the National West Nile Virus Surveillance System in Canada: A One Health approach. \u003cem\u003eCCDR\u003c/em\u003e \u003cstrong\u003e48\u003c/strong\u003e, 181\u0026ndash;187 (2022).\u003c/li\u003e\n\u003cli\u003eBusch, M. P. \u003cem\u003eet al.\u003c/em\u003e West Nile virus infections projected from blood donor screening data, United States, 2003. \u003cem\u003eEmerg. Infect. Dis.\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 395\u0026ndash;402 (2006).\u003c/li\u003e\n\u003cli\u003eMurray, J. \u0026amp; Cohen, A. L. Infectious Disease Surveillance. in \u003cem\u003eInternational Encyclopedia of Public Health\u003c/em\u003e 222\u0026ndash;229 (Elsevier, 2017). doi:10.1016/B978-0-12-803678-5.00517-8.\u003c/li\u003e\n\u003cli\u003eDoyle, T. J. Completeness of notifiable infectious disease reporting in the United States: an analytical literature review. \u003cem\u003eAmerican Journal of Epidemiology\u003c/em\u003e \u003cstrong\u003e155\u003c/strong\u003e, 866\u0026ndash;874 (2002).\u003c/li\u003e\n\u003cli\u003eAmodio, E., Zarcone, M., Casuccio, A. \u0026amp; Vitale, F. Trends in epidemiology: the role of denominator fluctuation in population based estimates. \u003cem\u003eAIMS Public Health\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 500\u0026ndash;506 (2021).\u003c/li\u003e\n\u003cli\u003eWon, E. J. \u003cem\u003eet al.\u003c/em\u003e Under-diagnosis of vector-borne diseases among individuals suspected of having Scrub Typhus in South Korea. \u003cem\u003ePLoS ONE\u003c/em\u003e \u003cstrong\u003e18\u003c/strong\u003e, e0286631 (2023).\u003c/li\u003e\n\u003cli\u003eKeita, A. K. \u003cem\u003eet al.\u003c/em\u003e The detection of vector-borne-disease-related DNA in human stool paves the way to large epidemiological studies. \u003cem\u003eEur J Epidemiol\u003c/em\u003e \u003cstrong\u003e30\u003c/strong\u003e, 1021\u0026ndash;1026 (2015).\u003c/li\u003e\n\u003cli\u003eKoliopoulos, P. \u003cem\u003eet al.\u003c/em\u003e Multiplex-RT-PCR-ELISA panel for detecting mosquito-borne pathogens: \u003cem\u003ePlasmodium\u003c/em\u003e sp. preserved and eluted from dried blood spots on sample cards. \u003cem\u003eMalar J\u003c/em\u003e \u003cstrong\u003e20\u003c/strong\u003e, 66 (2021).\u003c/li\u003e\n\u003cli\u003ePatrican, L. A. \u003cem\u003eet al.\u003c/em\u003e Host-feeding patterns of \u003cem\u003eCulex\u003c/em\u003e mosquitoes in relation to trap habitat. \u003cem\u003eEmerg. Infect. Dis.\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 1921\u0026ndash;1923 (2007).\u003c/li\u003e\n\u003cli\u003eHamer, G. L. \u003cem\u003eet al.\u003c/em\u003e Host selection by \u003cem\u003eCulex pipiens\u003c/em\u003e mosquitoes and West Nile virus amplification. \u003cem\u003eAm. J. Trop. Med. Hyg.\u003c/em\u003e \u003cstrong\u003e80\u003c/strong\u003e, 268\u0026ndash;278 (2009).\u003c/li\u003e\n\u003cli\u003eHamer, G. L. \u003cem\u003eet al.\u003c/em\u003e \u003cem\u003eCulex pipiens\u003c/em\u003e (Diptera: Culicidae): a bridge vector of West Nile virus to humans. \u003cem\u003eJournal of Medical Entomology\u003c/em\u003e \u003cstrong\u003e45\u003c/strong\u003e, 125\u0026ndash;128 (2008).\u003c/li\u003e\n\u003cli\u003eEbel, G. D., Rochlin, I., Longacker, J. \u0026amp; Kramer, L. D. \u003cem\u003eCulex restuans\u003c/em\u003e (Diptera: Culicidae) relative abundance and vector competence for West Nile virus. \u003cem\u003eme\u003c/em\u003e \u003cstrong\u003e42\u003c/strong\u003e, 838\u0026ndash;843 (2005).\u003c/li\u003e\n\u003cli\u003eFarajollahi, A., Fonseca, D. M., Kramer, L. D. \u0026amp; Marm Kilpatrick, A. \u0026ldquo;Bird biting\u0026rdquo; mosquitoes and human disease: a review of the role of \u003cem\u003eCulex pipiens\u003c/em\u003e complex mosquitoes in epidemiology. \u003cem\u003eInfection, Genetics and Evolution\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 1577\u0026ndash;1585 (2011).\u003c/li\u003e\n\u003cli\u003eAndreadis, T. G. The contribution of \u003cem\u003eCulex pipiens\u003c/em\u003e complex mosquitoes to transmission and persistence of West Nile virus in North America. \u003cem\u003eJournal of the American Mosquito Control Association\u003c/em\u003e \u003cstrong\u003e28\u003c/strong\u003e, 137\u0026ndash;151 (2012).\u003c/li\u003e\n\u003cli\u003eTiawsirisup, S. \u003cem\u003eet al.\u003c/em\u003e Vector competence of \u0026lt;I\u0026gt;Aedes vexans\u0026lt;/I\u0026gt; (Diptera: Culicidae) for West Nile virus and potential as an enzootic vector. \u003cem\u003eJournal of Medical Entomology\u003c/em\u003e \u003cstrong\u003e45\u003c/strong\u003e, 452\u0026ndash;457 (2008).\u003c/li\u003e\n\u003cli\u003eMolaei, G. \u0026amp; Andreadis, T. G. Identification of avian- and mammalian-derived bloodmeals in \u003cem\u003eAedes vexans\u003c/em\u003e and \u003cem\u003eCuliseta melanura\u003c/em\u003e (Diptera: Culicidae) and its implication for West Nile virus transmission in Connecticut, U.S.A. \u003cem\u003eJournal of Medical Entomology\u003c/em\u003e \u003cstrong\u003e43\u003c/strong\u003e, 1088\u0026ndash;1093 (2006).\u003c/li\u003e\n\u003cli\u003eGreenberg, J. A., Lujan, D. A., DiMenna, M. A., Wearing, H. J. \u0026amp; Hofkin, B. V. Identification of blood meal sources in \u003cem\u003eAedes vexans\u003c/em\u003e and \u003cem\u003eCulex quinquefasciatus\u003c/em\u003e in Bernalillo County, New Mexico. \u003cem\u003eJournal of Insect Science\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 1\u0026ndash;12 (2013).\u003c/li\u003e\n\u003cli\u003eAnderson, J. F., Main, A. J. \u0026amp; Ferrandino, F. J. Horizontal and vertical transmission of West Nile virus by \u003cem\u003eAedes vexans\u003c/em\u003e (Diptera: Culicidae). \u003cem\u003eJournal of Medical Entomology\u003c/em\u003e \u003cstrong\u003e57\u003c/strong\u003e, 1614\u0026ndash;1618 (2020).\u003c/li\u003e\n\u003cli\u003eMoser, S. K. \u003cem\u003eet al.\u003c/em\u003e Scoping review of Culex mosquito life history trait heterogeneity in response to temperature. \u003cem\u003eParasites Vectors\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 200 (2023).\u003c/li\u003e\n\u003cli\u003eShaman, J. Using a dynamic hydrology model to predict mosquito abundances in flood and swamp water. \u003cem\u003eEmerg. Infect. Dis.\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 8\u0026ndash;13 (2002).\u003c/li\u003e\n\u003cli\u003eBusch, M. P. \u003cem\u003eet al.\u003c/em\u003e Virus and antibody dynamics in acute West Nile virus infection. \u003cem\u003eJ INFECT DIS\u003c/em\u003e \u003cstrong\u003e198\u003c/strong\u003e, 984\u0026ndash;993 (2008).\u003c/li\u003e\n\u003cli\u003eDohm, D. J. \u0026amp; Turell, M. J. Effect of incubation at overwintering temperatures on the replication of West Nile virus in New York \u003cem\u003eCulex pipiens\u003c/em\u003e (Diptera: Culicidae). \u003cem\u003eme\u003c/em\u003e \u003cstrong\u003e38\u003c/strong\u003e, 462\u0026ndash;464 (2001).\u003c/li\u003e\n\u003cli\u003eDohm, D. J., O\u0026rsquo;Guinn, M. L. \u0026amp; Turell, M. J. Effect of environmental temperature on the ability of \u003cem\u003eCulex pipiens\u003c/em\u003e (Diptera: Culicidae) to transmit West Nile virus. \u003cem\u003eJ Med Entomol\u003c/em\u003e \u003cstrong\u003e39\u003c/strong\u003e, 221\u0026ndash;225 (2002).\u003c/li\u003e\n\u003cli\u003eRoiz, D., Ruiz, S., Soriguer, R. \u0026amp; Figuerola, J. Climatic effects on mosquito abundance in Mediterranean wetlands. \u003cem\u003eParasites Vectors\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 333 (2014).\u003c/li\u003e\n\u003cli\u003eGim\u0026eacute;nez, J. O., Fischer, S., Zalazar, L. \u0026amp; Stein, M. Cold season mortality under natural conditions and subsequent hatching response of \u003cem\u003eAedes\u003c/em\u003e (\u003cem\u003eStegomyia\u003c/em\u003e) \u003cem\u003eaegypti\u003c/em\u003e (Diptera: Culicidae) eggs in a subtropical city of Argentina. \u003cem\u003eJ. Med. Entomol.\u003c/em\u003e \u003cstrong\u003e52\u003c/strong\u003e, 879\u0026ndash;885 (2015).\u003c/li\u003e\n\u003cli\u003eEisen, L. \u003cem\u003eet al.\u003c/em\u003e The impact of temperature on the bionomics of \u003cem\u003eAedes\u003c/em\u003e (\u003cem\u003eStegomyia\u003c/em\u003e) \u003cem\u003eaegypti\u003c/em\u003e, with special reference to the cool geographic range margins. \u003cem\u003eJ Med Entomol\u003c/em\u003e \u003cstrong\u003e51\u003c/strong\u003e, 496\u0026ndash;516 (2014).\u003c/li\u003e\n\u003cli\u003eShocket, M. S. \u003cem\u003eet al.\u003c/em\u003e Transmission of West Nile and five other temperate mosquito-borne viruses peaks at temperatures between 23\u0026deg;C and 26\u0026deg;C. \u003cem\u003eeLife\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, e58511 (2020).\u003c/li\u003e\n\u003cli\u003eOzdenerol, E., Taff, G. \u0026amp; Akkus, C. Exploring the spatio-temporal dynamics of reservoir hosts, vectors, and human hosts of West Nile Virus: a review of the recent literature. \u003cem\u003eInternational Journal of Environmental Research and Public Health\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 5399\u0026ndash;5432 (2013).\u003c/li\u003e\n\u003cli\u003eAllan, B. F., Keesing, F. \u0026amp; Ostfeld, R. S. Effect of forest fragmentation on Lyme disease risk. \u003cem\u003eConservation Biology\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 267\u0026ndash;272 (2003).\u003c/li\u003e\n\u003cli\u003eCanadian Paediatric Society. West Nile virus in the context of climate change. \u003cem\u003ePaediatrics \u0026amp; Child Health\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 399\u0026ndash;400 (2008).\u003c/li\u003e\n\u003cli\u003eKim, B. I. \u003cem\u003eet al.\u003c/em\u003e A conceptual model for the impact of climate change on fox rabies in Alaska, 1980-2010. \u003cem\u003eZoonoses Public Health\u003c/em\u003e \u003cstrong\u003e61\u003c/strong\u003e, 72\u0026ndash;80 (2014).\u003c/li\u003e\n\u003cli\u003eGardner, A. M., Lampman, R. L. \u0026amp; Muturi, E. J. Land use patterns and the risk of West Nile virus transmission in Central Illinois. \u003cem\u003eVector-Borne and Zoonotic Diseases\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 338\u0026ndash;345 (2014).\u003c/li\u003e\n\u003cli\u003eHoover, K. C. \u0026amp; Barker, C. M. West Nile virus, climate change, and circumpolar vulnerability. \u003cem\u003eWIREs Clim Change\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 283\u0026ndash;300 (2016).\u003c/li\u003e\n\u003cli\u003eLudwig, A. \u003cem\u003eet al.\u003c/em\u003e Increased risk of endemic mosquito-borne diseases in Canada due to climate change. \u003cem\u003eCCDR\u003c/em\u003e \u003cstrong\u003e45\u003c/strong\u003e, 91\u0026ndash;97 (2019).\u003c/li\u003e\n\u003cli\u003eRakotoarinia, M. R. \u003cem\u003eet al.\u003c/em\u003e Effects of land use and weather on the presence and abundance of mosquito-borne disease vectors in a urban and agricultural landscape in Eastern Ontario, Canada. \u003cem\u003ePLoS ONE\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, e0262376 (2022).\u003c/li\u003e\n\u003cli\u003eElmqvist, T. \u003cem\u003eet al.\u003c/em\u003e \u003cem\u003eUrbanization, Biodiversity and Ecosystem Services: Challenges and Opportunities\u003c/em\u003e. (Springer Netherlands, Dordrecht, 2013). doi:10.1007/978-94-007-7088-1.\u003c/li\u003e\n\u003cli\u003eLi, Y. \u003cem\u003eet al.\u003c/em\u003e Urbanization increases \u003cem\u003eAedes albopictus\u003c/em\u003e larval habitats and accelerates mosquito development and survivorship. \u003cem\u003ePLoS Negl Trop Dis\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, e3301 (2014).\u003c/li\u003e\n\u003cli\u003eKjellstrom, T. \u003cem\u003eet al.\u003c/em\u003e Urban environmental health hazards and health equity. \u003cem\u003eJ Urban Health\u003c/em\u003e \u003cstrong\u003e84\u003c/strong\u003e, 86\u0026ndash;97 (2007).\u003c/li\u003e\n\u003cli\u003eDonnelly, M. A. P., Kluh, S., Snyder, R. E. \u0026amp; Barker, C. M. Quantifying sociodemographic heterogeneities in the distribution of \u003cem\u003eAedes aegypti\u003c/em\u003e among California households. \u003cem\u003ePLoS Negl Trop Dis\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, e0008408 (2020).\u003c/li\u003e\n\u003cli\u003eLittle, E. \u003cem\u003eet al.\u003c/em\u003e Socio-ecological mechanisms supporting high densities of \u003cem\u003eAedes albopictus\u003c/em\u003e (Diptera: Culicidae) in Baltimore, MD. \u003cem\u003eJ. Med. Entomol.\u003c/em\u003e \u003cstrong\u003e54\u003c/strong\u003e, 1183\u0026ndash;1192 (2017).\u003c/li\u003e\n\u003cli\u003eWhiteman, A. \u003cem\u003eet al.\u003c/em\u003e \u003cem\u003eAedes\u003c/em\u003e mosquito infestation in socioeconomically contrasting neighborhoods of Panama City. \u003cem\u003eEcoHealth\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 210\u0026ndash;221 (2019).\u003c/li\u003e\n\u003cli\u003eDowling, Z. \u003cem\u003eet al.\u003c/em\u003e Linking mosquito infestation to resident socioeconomic status, knowledge, and source reduction practices in suburban Washington, DC. \u003cem\u003eEcoHealth\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 36\u0026ndash;47 (2013).\u003c/li\u003e\n\u003cli\u003eFuentes-Vallejo, M. \u003cem\u003eet al.\u003c/em\u003e Territorial analysis of \u003cem\u003eAedes aegypti\u003c/em\u003e distribution in two Colombian cities: a chorematic and ecosystem approach. \u003cem\u003eCad Saude Publica\u003c/em\u003e \u003cstrong\u003e31\u003c/strong\u003e, 517\u0026ndash;530 (2015).\u003c/li\u003e\n\u003cli\u003eVonesch, N. \u003cem\u003eet al.\u003c/em\u003e Emerging zoonotic viral infections of occupational health importance. \u003cem\u003ePathogens and Disease\u003c/em\u003e \u003cstrong\u003e77\u003c/strong\u003e, ftz018 (2019).\u003c/li\u003e\n\u003cli\u003eEllwanger, J. H., Cardoso, J. D. C. \u0026amp; Chies, J. A. B. Variability in human attractiveness to mosquitoes. \u003cem\u003eCurrent Research in Parasitology \u0026amp; Vector-Borne Diseases\u003c/em\u003e \u003cstrong\u003e1\u003c/strong\u003e, 100058 (2021).\u003c/li\u003e\n\u003cli\u003eRipoche, M., Campagna, C., Ludwig, A., Ogden, N. H. \u0026amp; Leighton, P. A. Short-term forecasting of daily abundance of West Nile virus vectors \u003cem\u003eCulex pipiens-restuans\u003c/em\u003e (Diptera: Culicidae) and \u003cem\u003eAedes vexans\u003c/em\u003e based on weather conditions in Southern Qu\u0026eacute;bec (Canada). \u003cem\u003eJournal of Medical Entomology\u003c/em\u003e \u003cstrong\u003e56\u003c/strong\u003e, 859\u0026ndash;872 (2019).\u003c/li\u003e\n\u003cli\u003eGiordano, B. V., Kaur, S. \u0026amp; Hunter, F. F. West Nile virus in Ontario, Canada: a twelve-year analysis of human case prevalence, mosquito surveillance, and climate data. \u003cem\u003ePLOS ONE\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, e0183568 (2017).\u003c/li\u003e\n\u003cli\u003eGiordano, B., Turner, K. \u0026amp; Hunter, F. Geospatial analysis and seasonal distribution of West Nile virus vectors (Diptera: Culicidae) in Southern Ontario, Canada. \u003cem\u003eInternational Journal of Environmental Research and Public Health\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 614 (2018).\u003c/li\u003e\n\u003cli\u003eMallya, S. \u003cem\u003eet al.\u003c/em\u003e Factors associated with human West Nile virus infection in Ontario: a generalized linear mixed modelling approach. \u003cem\u003eBMC Infectious Diseases\u003c/em\u003e \u003cstrong\u003e18\u003c/strong\u003e, 1\u0026ndash;9 (2018).\u003c/li\u003e\n\u003cli\u003eTalbot, B., Ardis, M. \u0026amp; Kulkarni, M. A. Influence of demography, land use, and urban form on West Nile virus risk and human West Nile virus incidence in Ottawa, Canada. \u003cem\u003eVector-Borne and Zoonotic Diseases\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e, 533\u0026ndash;539 (2019).\u003c/li\u003e\n\u003cli\u003eTalbot, B. \u003cem\u003eet al.\u003c/em\u003e Ecological niche and positive clusters of two West Nile virus vectors in Ontario, Canada. \u003cem\u003eEcoHealth\u003c/em\u003e \u003cstrong\u003e20\u003c/strong\u003e, 249\u0026ndash;262 (2023).\u003c/li\u003e\n\u003cli\u003eTalbot, B., Caron-L\u0026eacute;vesque, M., Ardis, M., Kryuchkov, R. \u0026amp; Kulkarni, M. A. Linking bird and mosquito data to assess spatiotemporal West Nile virus risk in humans. \u003cem\u003eEcoHealth\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 70\u0026ndash;81 (2019).\u003c/li\u003e\n\u003cli\u003eLudwig, A., Rousseu, F., Kotchi, S. O., Allostry, J. \u0026amp; Fournier, R. A. Mapping the abundance of endemic mosquito-borne diseases vectors in southern Quebec. \u003cem\u003eBMC Public Health\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 924 (2023).\u003c/li\u003e\n\u003cli\u003eRuiz, M. O., Walker, E. D., Foster, E. S., Haramis, L. D. \u0026amp; Kitron, U. D. Association of West Nile virus illness and urban landscapes in Chicago and Detroit. \u003cem\u003eInternational Journal of Health Geographics\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e, 10 (2007).\u003c/li\u003e\n\u003cli\u003eJohnson, B. J. \u003cem\u003eet al.\u003c/em\u003e The roles of mosquito and bird communities on the prevalence of West Nile virus in urban wetland and residential habitats. \u003cem\u003eUrban Ecosystems\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 513\u0026ndash;531 (2012).\u003c/li\u003e\n\u003cli\u003eSkaff, N. K. \u0026amp; Cheruvelil, K. S. Fine-scale wetland features mediate vector and climate-dependent macroscale patterns in human West Nile virus incidence. \u003cem\u003eLandscape Ecology\u003c/em\u003e \u003cstrong\u003e31\u003c/strong\u003e, 1615\u0026ndash;1628 (2016).\u003c/li\u003e\n\u003cli\u003eCarrasquilla, M. C. \u003cem\u003eet al.\u003c/em\u003e Entomological characterization of \u003cem\u003eAedes\u003c/em\u003e mosquitoes and arbovirus detection in Ibagu\u0026eacute;, a Colombian city with co-circulation of Zika, dengue and chikungunya viruses. \u003cem\u003eParasites Vectors\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 446 (2021).\u003c/li\u003e\n\u003cli\u003eRuiz, M. O., Tedesco, C., McTighe, T. J., Austin, C. \u0026amp; Kitron, U. Environmental and social determinants of human risk during a West Nile virus outbreak in the greater Chicago area, 2002. \u003cem\u003eInt J Health Geogr\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, 8 (2004).\u003c/li\u003e\n\u003cli\u003eOzdenerol, E., Bialkowska-Jelinska, E. \u0026amp; Taff, G. N. Locating suitable habitats for West Nile Virus-infected mosquitoes through association of environmental characteristics with infected mosquito locations: a case study in Shelby County, Tennessee. \u003cem\u003eInternational Journal of Health Geographics\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 12 (2008).\u003c/li\u003e\n\u003cli\u003eDrews, S. J. \u003cem\u003eet al.\u003c/em\u003e A case series of inactivated Japanese encephalitis virus vaccination associated with positive West Nile virus blood donor screening nucleic acid tests. \u003cem\u003eTransfusion\u003c/em\u003e \u003cstrong\u003e60\u003c/strong\u003e, 1097\u0026ndash;1103 (2020).\u003c/li\u003e\n\u003cli\u003eO\u0026rsquo;Brien, S. F. \u003cem\u003eet al.\u003c/em\u003e West Nile virus in 2006 and 2007: the Canadian Blood Services\u0026rsquo; experience. \u003cem\u003eTransfusion\u003c/em\u003e \u003cstrong\u003e50\u003c/strong\u003e, 1118\u0026ndash;1125 (2010).\u003c/li\u003e\n\u003cli\u003eVamvakas, E. C., Kleinman, S., Hume, H. \u0026amp; Sher, G. D. The development of West Nile virus safety policies by Canadian Blood Services: guiding principles and a comparison between Canada and the United States. \u003cem\u003eTransfusion Medicine Reviews\u003c/em\u003e \u003cstrong\u003e20\u003c/strong\u003e, 97\u0026ndash;109 (2006).\u003c/li\u003e\n\u003cli\u003ePostler, T. S. \u003cem\u003eet al.\u003c/em\u003e Renaming of the genus Flavivirus to Orthoflavivirus and extension of binomial species names within the family Flaviviridae. \u003cem\u003eArch Virol\u003c/em\u003e \u003cstrong\u003e168\u003c/strong\u003e, 224, s00705-023-05835\u0026ndash;1 (2023).\u003c/li\u003e\n\u003cli\u003eLanciotti, R. S. \u003cem\u003eet al.\u003c/em\u003e Rapid detection of West Nile virus from human clinical specimens, field-collected mosquitoes, and avian samples by a TaqMan reverse transcriptase-PCR assay. \u003cem\u003eJ. Clin. Microbiol.\u003c/em\u003e \u003cstrong\u003e38\u003c/strong\u003e, 4066\u0026ndash;4071 (2000).\u003c/li\u003e\n\u003cli\u003eAgriculture and Agri-Food Canada. Annual Crop Inventory. https://www.agr.gc.ca/atlas/apps/metrics/index-fr.html?appid=aci-iac\u0026amp;lang=en (2022).\u003c/li\u003e\n\u003cli\u003eMasek, J. G. \u003cem\u003eet al.\u003c/em\u003e A Landsat surface reflectance dataset for North America, 1990\u0026ndash;2000. \u003cem\u003eIEEE Geosci. Remote Sensing Lett.\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, 68\u0026ndash;72 (2006).\u003c/li\u003e\n\u003cli\u003eVermote, E., Justice, C., Claverie, M. \u0026amp; Franch, B. Preliminary analysis of the performance of the Landsat 8/OLI land surface reflectance product. \u003cem\u003eRemote Sensing of Environment\u003c/em\u003e \u003cstrong\u003e185\u003c/strong\u003e, 46\u0026ndash;56 (2016).\u003c/li\u003e\n\u003cli\u003eThornton, M.M. \u003cem\u003eet al.\u003c/em\u003e Daymet: daily surface weather data on a 1-km grid for North America, Version 4. 0 MB (2020) doi:10.3334/ORNLDAAC/1840.\u003c/li\u003e\n\u003cli\u003eHalvorsen, R. A gradient analytic perspective on distribution modelling. \u003cem\u003eSommerfeltia\u003c/em\u003e \u003cstrong\u003e35\u003c/strong\u003e, 1\u0026ndash;165 (2012).\u003c/li\u003e\n\u003cli\u003eSillero, N. \u0026amp; Barbosa, A. M. Common mistakes in ecological niche models. \u003cem\u003eInternational Journal of Geographical Information Science\u003c/em\u003e \u003cstrong\u003e35\u003c/strong\u003e, 213\u0026ndash;226 (2021).\u003c/li\u003e\n\u003cli\u003ePhillips, S. J. \u003cem\u003eet al.\u003c/em\u003e Sample selection bias and presence-only distribution models: implications for background and pseudo-absence data. \u003cem\u003eEcological Applications\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e, 181\u0026ndash;197 (2009).\u003c/li\u003e\n\u003cli\u003eZhao, Z., Xiao, N., Shen, M. \u0026amp; Li, J. Comparison between optimized MaxEnt and random forest modeling in predicting potential distribution: a case study with \u003cem\u003eQuasipaa boulengeri\u003c/em\u003e in China. \u003cem\u003eScience of The Total Environment\u003c/em\u003e \u003cstrong\u003e842\u003c/strong\u003e, 156867 (2022).\u003c/li\u003e\n\u003cli\u003eBeeman, S. P., Morrison, A. M., Unnasch, T. R. \u0026amp; Unnasch, R. S. Ensemble ecological niche modeling of West Nile virus probability in Florida. \u003cem\u003ePLoS ONE\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, e0256868 (2021).\u003c/li\u003e\n\u003cli\u003eMondal, R. \u0026amp; Bhat, A. Comparison of regression-based and machine learning techniques to explain alpha diversity of fish communities in streams of central and eastern India. \u003cem\u003eEcological Indicators\u003c/em\u003e \u003cstrong\u003e129\u003c/strong\u003e, 107922 (2021).\u003c/li\u003e\n\u003cli\u003eThuiller, W., Lafourcade, B., Engler, R. \u0026amp; Ara\u0026uacute;jo, M. B. BIOMOD - a platform for ensemble forecasting of species distributions. \u003cem\u003eEcography\u003c/em\u003e \u003cstrong\u003e32\u003c/strong\u003e, 369\u0026ndash;373 (2009).\u003c/li\u003e\n\u003cli\u003eBarbet-Massin, M., Jiguet, F., Albert, C. H. \u0026amp; Thuiller, W. Selecting pseudo-absences for species distribution models: how, where and how many? \u003cem\u003eMethods in Ecology and Evolution\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, 327\u0026ndash;338 (2012).\u003c/li\u003e\n\u003cli\u003eAraujo, M. \u0026amp; New, M. Ensemble forecasting of species distributions. \u003cem\u003eTrends in Ecology \u0026amp; Evolution\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, 42\u0026ndash;47 (2007).\u003c/li\u003e\n\u003cli\u003eBates, D., M\u0026auml;chler, M., Bolker, B. \u0026amp; Walker, S. Fitting linear mixed-effects models Using \u003cem\u003elme4\u003c/em\u003e. \u003cem\u003eJ. Stat. Soft.\u003c/em\u003e \u003cstrong\u003e67\u003c/strong\u003e, (2015).\u003c/li\u003e\n\u003cli\u003eZeileis, A., Kleiber, C. \u0026amp; Jackman, S. Regression Models for Count Data in \u003cem\u003eR\u003c/em\u003e. \u003cem\u003eJ. Stat. Soft.\u003c/em\u003e \u003cstrong\u003e27\u003c/strong\u003e, (2008).\u003c/li\u003e\n\u003cli\u003eRaftery, A. E. Bayesian model selection in social research. \u003cem\u003eSociological Methodology\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 111 (1995).\u003c/li\u003e\n\u003cli\u003eArel-Bundock, V. \u003cstrong\u003emodelsummary\u003c/strong\u003e : Data and Model Summaries in \u003cem\u003eR\u003c/em\u003e. \u003cem\u003eJ. Stat. Soft.\u003c/em\u003e \u003cstrong\u003e103\u003c/strong\u003e, (2022).\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cstrong\u003eTable 1.\u0026nbsp;\u003c/strong\u003eDescriptive statistics relating to the abundance (number of mosquitoes per sampling visit) and occurrence (proportion of presence across sampling visits) for adult mosquitoes from both studied species, \u003cem\u003eCulex pipiens/restuans\u003c/em\u003e and \u003cem\u003eAedes vexans\u003c/em\u003e, collected across 3,010 sampling locations in Ontario, Canada, 2003-2022.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"595\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"49.16107382550336%\"\u003e\n \u003cp\u003eSpecies\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.201342281879196%\"\u003e\n \u003cp\u003e\u003cem\u003eCulex pipiens/restuans\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.63758389261745%\"\u003e\n \u003cp\u003e\u003cem\u003eAedes vexans\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"49.16107382550336%\"\u003e\n \u003cp\u003eTotal sampled\u0026nbsp;\u003c/p\u003e\n \u003cp\u003eadult mosquitoes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.201342281879196%\"\u003e\n \u003cp\u003e934,951\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.63758389261745%\"\u003e\n \u003cp\u003e1,778,876\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"49.2436974789916%\"\u003e\n \u003cp\u003eNumber of visits/site\u003c/p\u003e\n \u003cp\u003e(mean\u0026nbsp;\u0026plusmn;\u0026nbsp;standard deviation)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"50.7563025210084%\" colspan=\"2\"\u003e\n \u003cp\u003e48\u0026nbsp;\u0026plusmn;\u0026nbsp;80\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"49.16107382550336%\"\u003e\n \u003cp\u003eAbundance/visit\u003c/p\u003e\n \u003cp\u003e(mean\u0026nbsp;\u0026plusmn;\u0026nbsp;standard deviation)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.201342281879196%\"\u003e\n \u003cp\u003e5\u0026nbsp;\u0026plusmn;\u0026nbsp;8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.63758389261745%\"\u003e\n \u003cp\u003e12\u0026nbsp;\u0026plusmn;\u0026nbsp;18\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"49.16107382550336%\"\u003e\n \u003cp\u003eOccurrence/visit\u003c/p\u003e\n \u003cp\u003e(mean\u0026nbsp;\u0026plusmn;\u0026nbsp;standard deviation)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.201342281879196%\"\u003e\n \u003cp\u003e0.57\u0026nbsp;\u0026plusmn;\u0026nbsp;0.32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.63758389261745%\"\u003e\n \u003cp\u003e0.65\u0026nbsp;\u0026plusmn;\u0026nbsp;0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"49.16107382550336%\"\u003e\n \u003cp\u003eTotal number of WNV\u0026nbsp;\u003c/p\u003e\n \u003cp\u003edetection tests\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.201342281879196%\"\u003e\n \u003cp\u003e102,938\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.63758389261745%\"\u003e\n \u003cp\u003e82,134\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"49.16107382550336%\"\u003e\n \u003cp\u003eNumber of positive pools\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e[% of total for all species]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.201342281879196%\"\u003e\n \u003cp\u003e2,949 [88]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.63758389261745%\"\u003e\n \u003cp\u003e182 [5]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"49.16107382550336%\"\u003e\n \u003cp\u003eTotal number of adult mosquitoes tested for West Nile virus\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.201342281879196%\"\u003e\n \u003cp\u003e916,673\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.63758389261745%\"\u003e\n \u003cp\u003e1,147,874\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"49.16107382550336%\"\u003e\n \u003cp\u003eNumber of adult mosquitoes tested/test\u003c/p\u003e\n \u003cp\u003e(mean\u0026nbsp;\u0026plusmn;\u0026nbsp;standard deviation)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.201342281879196%\"\u003e\n \u003cp\u003e9\u0026nbsp;\u0026plusmn;\u0026nbsp;12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.63758389261745%\"\u003e\n \u003cp\u003e14\u0026nbsp;\u0026plusmn;\u0026nbsp;15\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"49.16107382550336%\"\u003e\n \u003cp\u003eTotal minimum infection rate\u003c/p\u003e\n \u003cp\u003e(positive pools/number tested x1000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.201342281879196%\"\u003e\n \u003cp\u003e3.22\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.63758389261745%\"\u003e\n \u003cp\u003e0.16\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2.\u003c/strong\u003e Parameters used to generate ensemble niche models for both studied species, \u003cem\u003eCulex pipiens/restuans and Aedes vexans\u003c/em\u003e, collected across 3,010 sampling locations in Ontario, Canada, 2003-2022. Number of models included out of the 100 random forest models generated, response operating curves for the best-performing model, for the mean habitat suitability index (HSI) averaged across models, and committee averaging of HSI across models and variable importance calculated with a 100-permutation procedure for the three explanatory variables are shown.\u003c/p\u003e\n\u003cdiv align=\"center\"\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"614\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"36.97068403908795%\" colspan=\"2\"\u003e\n \u003cp\u003eSpecies\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.03257328990228%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eAedes vexans\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"29.153094462540718%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eCulex pipiens/restuans\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.843648208469055%\" valign=\"top\"\u003e\n \u003cp\u003eWNV\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"36.97068403908795%\" colspan=\"2\"\u003e\n \u003cp\u003eNumber of models\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.03257328990228%\" valign=\"top\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"29.153094462540718%\" valign=\"top\"\u003e\n \u003cp\u003e100\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.843648208469055%\" valign=\"top\"\u003e\n \u003cp\u003e100\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"15.309446254071661%\" rowspan=\"3\"\u003e\n \u003cp\u003eArea under the curve\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"21.661237785016286%\" valign=\"top\"\u003e\n \u003cp\u003eBest-performing model\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.03257328990228%\" valign=\"top\"\u003e\n \u003cp\u003e0.72\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"29.153094462540718%\" valign=\"top\"\u003e\n \u003cp\u003e0.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.843648208469055%\" valign=\"top\"\u003e\n \u003cp\u003e0.79\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.576923076923077%\" valign=\"top\"\u003e\n \u003cp\u003eMean HSI values\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.653846153846153%\" valign=\"top\"\u003e\n \u003cp\u003e0.93\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"34.42307692307692%\" valign=\"top\"\u003e\n \u003cp\u003e0.91\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.346153846153847%\" valign=\"top\"\u003e\n \u003cp\u003e0.93\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.576923076923077%\" valign=\"top\"\u003e\n \u003cp\u003eCommittee averaging\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.653846153846153%\" valign=\"top\"\u003e\n \u003cp\u003e0.92\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"34.42307692307692%\" valign=\"top\"\u003e\n \u003cp\u003e0.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.346153846153847%\" valign=\"top\"\u003e\n \u003cp\u003e0.92\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"15.309446254071661%\" rowspan=\"3\"\u003e\n \u003cp\u003eVariable importance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"21.661237785016286%\" valign=\"top\"\u003e\n \u003cp\u003eLand cover\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"20.03257328990228%\" valign=\"top\"\u003e\n \u003cp\u003e0.32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"29.153094462540718%\" valign=\"top\"\u003e\n \u003cp\u003e0.21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.843648208469055%\" valign=\"top\"\u003e\n \u003cp\u003e0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.576923076923077%\" valign=\"top\"\u003e\n \u003cp\u003ePrecipitations\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.653846153846153%\" valign=\"top\"\u003e\n \u003cp\u003e0.62\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"34.42307692307692%\" valign=\"top\"\u003e\n \u003cp\u003e0.19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.346153846153847%\" valign=\"top\"\u003e\n \u003cp\u003e0.39\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.576923076923077%\" valign=\"top\"\u003e\n \u003cp\u003eTemperature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"23.653846153846153%\" valign=\"top\"\u003e\n \u003cp\u003e0.64\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"34.42307692307692%\" valign=\"top\"\u003e\n \u003cp\u003e0.78\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.346153846153847%\" valign=\"top\"\u003e\n \u003cp\u003e0.62\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eTable 3.\u003c/strong\u003e Descriptive statistics relating to the number of blood donors and donations to the Canadian Blood Services in Ontario, Canada, 2003-2022, tested for presence of West Nile virus (WNV).\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"624\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"21.12%\"\u003e\n \u003cp\u003eAge / Sex\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.16%\"\u003e\n \u003cp\u003e11 to 38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.8%\"\u003e\n \u003cp\u003e39 to 65\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.64%\"\u003e\n \u003cp\u003e66 to 93\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.76%\"\u003e\n \u003cp\u003eFemale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.76%\"\u003e\n \u003cp\u003eMale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.76%\"\u003e\n \u003cp\u003eTotal\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"21.12%\"\u003e\n \u003cp\u003eNumber of tests on blood donations\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.16%\"\u003e\n \u003cp\u003e2,200,777\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.8%\"\u003e\n \u003cp\u003e3,990,529\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.64%\"\u003e\n \u003cp\u003e389,440\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.76%\"\u003e\n \u003cp\u003e2,776,604\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.76%\"\u003e\n \u003cp\u003e3,804,165\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.76%\"\u003e\n \u003cp\u003e6,580,769\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"21.12%\"\u003e\n \u003cp\u003eNumber of unique blood donors\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.16%\"\u003e\n \u003cp\u003e576,752\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.8%\"\u003e\n \u003cp\u003e439,512\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.64%\"\u003e\n \u003cp\u003e52,525\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.76%\"\u003e\n \u003cp\u003e504,580\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.76%\"\u003e\n \u003cp\u003e467,814\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.76%\"\u003e\n \u003cp\u003e972,394\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"21.12%\"\u003e\n \u003cp\u003eMean number of donations per donor\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.16%\"\u003e\n \u003cp\u003e3.8 \u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.8%\"\u003e\n \u003cp\u003e9.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.64%\"\u003e\n \u003cp\u003e2.5 \u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.76%\"\u003e\n \u003cp\u003e5.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.76%\"\u003e\n \u003cp\u003e8.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.76%\"\u003e\n \u003cp\u003e6.8 \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"21.12%\"\u003e\n \u003cp\u003eNumber of WNV-positive donations\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.16%\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.8%\"\u003e\n \u003cp\u003e77\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.64%\"\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.76%\"\u003e\n \u003cp\u003e38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.76%\"\u003e\n \u003cp\u003e64\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.76%\"\u003e\n \u003cp\u003e102\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"21.12%\"\u003e\n \u003cp\u003eWNV-positive rate (x100,000 donations)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.16%\"\u003e\n \u003cp\u003e0.7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.8%\"\u003e\n \u003cp\u003e1.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.64%\"\u003e\n \u003cp\u003e2.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.76%\"\u003e\n \u003cp\u003e1.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.76%\"\u003e\n \u003cp\u003e1.7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.76%\"\u003e\n \u003cp\u003e1.6\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eTable 4.\u003c/strong\u003e Regression incidence rate ratio (IRR) values, 95% confidence intervals [95% CI], the \u003cem\u003eP\u003c/em\u003e value for each variable, and the R\u003csup\u003e2\u003c/sup\u003e for the model, for the multivariable modelling analyses on the WNV infection in the human donor population in Ontario from 2003 to 2022, at the analysis levels of the individual donor and the 2016 Census subdivision. HSI refers to habitat suitability index of the corresponding mosquito vector species.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"15.088282504012842%\" valign=\"top\"\u003e\n \u003cp\u003eAnalysis level\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"43.98073836276083%\" valign=\"top\"\u003e\n \u003cp\u003eVariable (units) \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.74317817014446%\" valign=\"top\"\u003e\n \u003cp\u003eIRR \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;[95% CI]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.593900481540931%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eP\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.593900481540931%\" valign=\"top\"\u003e\n \u003cp\u003eR\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"15.088282504012842%\" rowspan=\"13\"\u003e\n \u003cp\u003eCensus subdivision\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"43.98073836276083%\"\u003e\n \u003cp\u003eLow income (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.74317817014446%\"\u003e\n \u003cp\u003e2.829\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e[2.061, 3.885]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.593900481540931%\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.593900481540931%\" rowspan=\"13\"\u003e\n \u003cp\u003e0.677\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"59.1792656587473%\"\u003e\n \u003cp\u003eHigh spending on shelter (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"26.565874730021598%\"\u003e\n \u003cp\u003e0.702\u003c/p\u003e\n \u003cp\u003e[0.526, 0.939]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.254859611231101%\"\u003e\n \u003cp\u003e0.017\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"59.1792656587473%\"\u003e\n \u003cp\u003e\u003cem\u003eCx. pipiens/restuans\u003c/em\u003e (HSI)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"26.565874730021598%\"\u003e\n \u003cp\u003e1.950\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;[1.160, 3.280]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.254859611231101%\"\u003e\n \u003cp\u003e0.012\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"59.1792656587473%\"\u003e\n \u003cp\u003eWest Nile virus (HSI)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"26.565874730021598%\"\u003e\n \u003cp\u003e1.179\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;[0.845, 1.645]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.254859611231101%\"\u003e\n \u003cp\u003e0.333\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"59.1792656587473%\"\u003e\n \u003cp\u003ePrecipitations (mm)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"26.565874730021598%\"\u003e\n \u003cp\u003e0.872\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;[0.558, 1.363]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.254859611231101%\"\u003e\n \u003cp\u003e0.548\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"59.1792656587473%\"\u003e\n \u003cp\u003ePopulation density (residents/km\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"26.565874730021598%\"\u003e\n \u003cp\u003e0.908\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;[0.720, 1.145]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.254859611231101%\"\u003e\n \u003cp\u003e0.414\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"59.1792656587473%\"\u003e\n \u003cp\u003eMean income ($)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"26.565874730021598%\"\u003e\n \u003cp\u003e1.316\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;[0.941, 1.839]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.254859611231101%\"\u003e\n \u003cp\u003e0.108\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"59.1792656587473%\"\u003e\n \u003cp\u003eResidents older than 64 (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"26.565874730021598%\"\u003e\n \u003cp\u003e0.878\u003c/p\u003e\n \u003cp\u003e[0.584, 1.321]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.254859611231101%\"\u003e\n \u003cp\u003e0.533\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"59.1792656587473%\"\u003e\n \u003cp\u003eOccupation in trades (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"26.565874730021598%\"\u003e\n \u003cp\u003e0.807\u003c/p\u003e\n \u003cp\u003e[0.479, 1.357]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.254859611231101%\"\u003e\n \u003cp\u003e0.418\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"59.1792656587473%\"\u003e\n \u003cp\u003eOccupation in natural resources (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"26.565874730021598%\"\u003e\n \u003cp\u003e1.389\u003c/p\u003e\n \u003cp\u003e[0.927, 2.081]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.254859611231101%\"\u003e\n \u003cp\u003e0.111\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"59.1792656587473%\"\u003e\n \u003cp\u003eLand cover of wetlands (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"26.565874730021598%\"\u003e\n \u003cp\u003e0.654\u003c/p\u003e\n \u003cp\u003e[0.347, 1.234]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.254859611231101%\"\u003e\n \u003cp\u003e0.190\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"59.1792656587473%\"\u003e\n \u003cp\u003eLand cover of wild grasslands (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"26.565874730021598%\"\u003e\n \u003cp\u003e1.397\u003c/p\u003e\n \u003cp\u003e[0.909, 2.148]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.254859611231101%\"\u003e\n \u003cp\u003e0.127\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"59.1792656587473%\"\u003e\n \u003cp\u003eLand cover of forests (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"26.565874730021598%\"\u003e\n \u003cp\u003e1.332\u003c/p\u003e\n \u003cp\u003e[0.680, 2.612]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.254859611231101%\"\u003e\n \u003cp\u003e0.403\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"15.088282504012842%\" rowspan=\"3\"\u003e\n \u003cp\u003eIndividual donation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"43.98073836276083%\"\u003e\n \u003cp\u003eDonor age (years since birth)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.74317817014446%\"\u003e\n \u003cp\u003e1.002\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e[1.001, 1.003]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.593900481540931%\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"10.593900481540931%\" rowspan=\"3\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"59.1792656587473%\"\u003e\n \u003cp\u003eYear of detection test (date)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"26.565874730021598%\"\u003e\n \u003cp\u003e1.003\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e[1.002, 1.003]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.254859611231101%\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"59.1792656587473%\"\u003e\n \u003cp\u003eMonth of detection test (date)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"26.565874730021598%\"\u003e\n \u003cp\u003e1.002\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e[1.001, 1.003]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.254859611231101%\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\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":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-4714418/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4714418/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eWest Nile virus (WNV) is a mosquito-borne zoonotic flavivirus which often causes asymptomatic infection in humans but may develop into a deadly neuroinvasive disease. In this study, we aimed to investigate variables potentially associated with human WNV infection using human and mosquito WNV surveillance and monitoring datasets, established over 20 years, from 2003 to 2022, across the province of Ontario, Canada. We combined climatic and geographic data, mosquito surveillance data (n=3,010 sites), blood donation arboviral detection testing data in the human population, and demographic and socio-economic data from Canadian population censuses. We hypothesized that spatio-temporal indices related to mosquito vector habitat and phenology, in addition to human demographic and socio-economic factors, were associated with WNV infection in the human population. Our results show that habitat suitability of the main WNV vector in this region, \u003cem\u003eCx. pipiens/restuans\u003c/em\u003e (IRR = 2.0), and variables related to lower income (IRR = 2.8), and shelter infrastructure spending (IRR = 0.7), were key risk factors associated with WNV infection among blood donors from 2003 to 2022 across Ontario (R\u003csup\u003e2\u003c/sup\u003e = 0.67). These results may inform points of entry for practical intervention aimed at reducing risk of mosquito-borne pathogens in Canada.\u003c/p\u003e","manuscriptTitle":"Spatiotemporal analysis of West Nile virus infection in the human population based on arboviral detection testing of blood donations","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-08-05 09:50:59","doi":"10.21203/rs.3.rs-4714418/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-08-26T08:27:27+00:00","index":"","fulltext":""},{"type":"reviewerAgreed","content":"200406578311476136972993862554110797322","date":"2024-08-12T02:39:42+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"85691482719087149994866804027443447018","date":"2024-08-09T02:46:22+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-08-08T11:34:46+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-08-04T21:58:24+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-07-26T21:19:32+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-07-26T09:19:01+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"220128365451079083719155090922571333519","date":"2024-07-19T03:37:19+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"67485614494537711051561410035674391603","date":"2024-07-17T16:10:29+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"289314746785124171689070383855889893434","date":"2024-07-17T12:56:55+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"231578997053303480295705369968114314320","date":"2024-07-17T10:12:01+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-07-17T01:16:36+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-07-12T05:17:51+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2024-07-12T04:56:37+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-07-11T03:58:42+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2024-07-09T21:29:20+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"1f9b08bb-5b21-4b90-8c29-3ac41dbe5f12","owner":[],"postedDate":"August 5th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":35528780,"name":"Biological sciences/Ecology/Ecological epidemiology"},{"id":35528781,"name":"Health sciences/Diseases/Infectious diseases/Viral infection"}],"tags":[],"updatedAt":"2024-12-30T15:59:55+00:00","versionOfRecord":{"articleIdentity":"rs-4714418","link":"https://doi.org/10.1038/s41598-024-82739-3","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2024-12-28 15:57:14","publishedOnDateReadable":"December 28th, 2024"},"versionCreatedAt":"2024-08-05 09:50:59","video":"","vorDoi":"10.1038/s41598-024-82739-3","vorDoiUrl":"https://doi.org/10.1038/s41598-024-82739-3","workflowStages":[]},"version":"v1","identity":"rs-4714418","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4714418","identity":"rs-4714418","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}