West Nile virus (family Flaviviridae, genus Flavivirus) is a multi-host and multi-vector zoonotic arbovirus system maintained between mosquito vectors (primarily in the genus Culex) and avian hosts (mainly in the Passeriformes and Columbiformes orders) (Colpitts et al., 2012; Reisen, 2013). The virus has global relevance with the broadest geographic distribution of any arbovirus and it continues to expand in geographic range (Ciota, 2017). WNV is the leading cause of mosquito-borne human arbovirus disease in the U.S. and has a 10% mortality rate among infected people who develop neuroinvasive disease (Ciota, 2017). After its arrival to North America in 1999, substantial resources have been invested to understand the ecology of the WNV system in the U.S., including multiple studies examining landscape correlations to identify where transmission hazard may be greatest (DeGroote and Sugumaran, 2012; Reisen, 2013).
Landscapes provide important habitat for mosquito vectors and avian hosts. The composition and configuration of the landscape can affect the presence and opportunities for contact between vectors and hosts required for virus transmission (Pavlovsky 1966; Lambin et al., 2010; Kilpatrick, 2011; Reisen, 2013). WNV mosquito vectors and avian hosts inhabit a broad range of landscapes, from urban to rural and agricultural settings (Reisen, 2013). This factor, along with regional variations in the distribution of mosquito vectors and amplifying hosts, have presented challenges to identifying consistent and generalizable landscape predictors of WNV transmission hazard (Colpitts et al., 2012; Rochlin et al., 2019). The result is that landscape studies of WNV often lead to inconsistent or contradictory outcomes (Bowden et al., 2011; DeGroote and Sugumaran, 2012).
For example, increased urbanization in Dallas County, TX, where Culex quinquefasciatus Say (Diptera: Culicidae) mosquitoes are abundant, contributed to increased risk for WNV-infected mosquito pools during the 2012 human epidemic (Poh et al., 2020), and in Chicago, IL, and Detroit, MI, high incidence of WNV in humans occurred in suburban neighborhoods, areas with moderate vegetation, and moderate population density, where Culex pipiens Linnaeus (Diptera: Culicidae) is often abundant (Ruiz et al., 2007). However, in California, transmission patterns were linked with fast-growing rice fields or more generic cropland preferred by Culex tarsalis Coquillett (Diptera: Culicidae) in the Central Valley and with developed landscapes and neglected swimming pools preferred by Cx. quinquefasciatus mosquitoes in Southern California (Scott et al., 2001; Goddard et al., 2002; Reisen, Barker, et al., 2008; Reisen, Takahashi, et al., 2008; Kwan et al., 2010; Kovach and Kilpatrick, 2018; Snyder et al., 2020).
In addition, irrigated agricultural landscapes and rural landscapes, which are the preferred habitat of Cx. tarsalis, were correlated with WNV transmission in the Pacific Northwest, Iowa, and Colorado (DeGroote et al., 2008; Eisen et al., 2010; Crowder et al., 2013; Dunphy et al., 2019), while in South Dakota, irrigated agriculture was not correlated with human incidence (Chuang et al., 2012). In the southeastern U.S., urban and semi-urban landscapes, where Cx. quinquefasciatus are abundant, have been linked to elevated WNV seroprevalence in birds in Georgia with a strong association with human housing density (Gibbs et al., 2006; DeGroote and Sugumaran, 2012; Levine et al., 2013), while population density had a negative correlation with WN incidence in Louisiana where Culex nigripalpus Theobald (Diptera: Culicidae) may play a role in transmission (Ezenwa et al., 2007). Variation in these results suggest that landscape studies at regional to local scales may be needed to reveal operationally useful information to mosquito control and public health agencies (Bowden et al., 2011; DeGroote and Sugumaran, 2012).
In Florida, USA, little is known about landscape correlations and WNV. Florida encompasses < 1% of the WNV reported cases in North America since it was first detected in the state 2001 (CDC, 2023). However, the virus is endemic, transmission occurs statewide, and constitutes a continual health risk to humans and equines (FDOH, 2023). Culex nigripalpus and Cx. quinquefasciatus mosquitoes are considered the main mosquito vectors in Florida (Rochlin et al. 2019, Vitek et al. 2014). However, additional species including Culex coronator (Dyar & Knab) (Diptera: Culicidae) and Culex erraticus (Dyar & Knab) (Diptera: Culicidae) may also contribute to maintenance and/or spillover of the virus in the natural environment (Burkett-Cadena et al., 2012; Alto et al., 2014). These species occupy a range of urban and rural habitats, using artificial containers, roadside ditches, and agricultural furrows, among others as breeding sites (Day et al., 2015; Wilkerson et al., 2021). Culex nigripalpus and Cx. quinquefasciatus feed upon a wide variety of vertebrate hosts in Florida, including birds, mammals, and reptiles (Edman, 1974; Hancock and Camp, 2022; Reeves and Burkett-Cadena, 2022). In addition, multiple Passeriformes birds that can serve as WNV hosts are found across a broad range of landscapes in Florida for all or a portion of the year, including American Robin (Turdus migratorius) and Common Grackle (Quiscalus quiscula) (Hostetler and Archer, 2017), and several songbirds previously implicated as WNV hosts, including Red-Winged Blackbird (Agelaius phoeniceus) and Cedar Waxwing (Bombycilla cedrorum), are widely found in agricultural landscapes (Audubon, 2023a; Audubon, 2023b), which may be important to WNV transmission in Florida.
The Florida Department of Health (FDOH) in conjunction with mosquito control programs in the state, has conducted long-term monitoring for WNV through an extensive sentinel chicken (Gallus gallus domesticus Linnaeus) (Galliformes: Phasianidae) program since its first detection in 2001 in Jefferson County (Lloyd et al., 2018). This surveillance program provides a valuable opportunity to investigate landscape correlations with WNV transmission, but few studies have leveraged this information (Lloyd, 2018). Previous landscape studies focused primarily on land cover composition to predict the potential distribution of different components of the WNV system using ecological niche models (Sallam et al., 2016; Beeman et al., 2021). However, land cover configuration, including the amount of edge habitat in an area, may be an important predictor of WNV transmission hazard in the state. Edge habitats can alter opportunities for host/vector interactions (Rizzoli et al., 2019) and affect vector borne disease system dynamics in an area (Gottwalt, 2013; Reiskind et al., 2017; Ribeiro Prist et al., 2022).
Here, we leverage 2018 Florida Sentinel Chicken Program (FSCP) data to investigate landscape correlations with WNV seroconversion rates at individual chicken coops across the state and at coops within five bioclimatic regions within Florida. In 2018 the FSCP recorded 814 total sentinel chicken WNV seroconversions, the greatest number in 10 years, and several human and equine WNV cases were reported in multiple Florida counties and regions (Morrison et al., 2018). Although little is known about specific landscapes associated with WNV transmission in Florida, we expect that not only the composition but the configuration of landscapes surrounding sentinel chicken coops will be important determinants of the proportions of positive WNV seroconversion. In addition, we expect that greater percentages of more fragmented suburban or semi-rural landscapes that support both Cx. quinquefasciatus and Cx. nigripalpus mosquito vectors and multiple WNV avian hosts will be correlated with higher seroconversion in sentinel chickens. We also expect to find regional differences in the importance of specific land cover classes to WNV seroconversion and that greater percentages of agricultural composition in central Florida will have a positive correlation with WNV seroconversion. Because the scale at which landscapes affect WNV seroconversion is unknown, we examine landscape composition and configuration summarized across multiple distances from sentinel chicken coops. Understanding these linkages has the potential to reveal new information about WNV transmission hazard in this region, while contributing to a broader understanding of multivector/multi-host vector-borne disease distributions across heterogeneous landscapes.