Developing operational algorithms using linear and non-linear squares estimation in Python® for the identification of Culex pipiens and Culex restuans in a mosquito abatement district (Cook County, Illinois, USA)

In this research, community level spatial models were developed for determining mosquito abundance and environmental factors that could aid in the risk prediction of West Nile virus (WNv) outbreaks. Adult <em>Culex pipiens</em> and <em>Culex restuan</em> mosquitoes and multiple habitat covariates were collected from nine sites within Cook County, Illinois, USA, to provide spatio-temporal information on the abundance of WNv vectors from 2002 to 2005. Regression analyses of the sampled covariates revealed that the adult <em>Culex</em> population was positively associated with temperature throughout the sampling frame. The model output also indicated that precipitation was negatively associated to mosquito abundance in 2002, 2003 and 2005 (P &lt;0.05), but positively associated in 2004 (P &lt;0.05). A land use land cover classification, based on QuickBird visible and near infra-red data, acquired at 0.61 m resolution, was used to investigate possible associations between geographical features and the abundance of sampled <em>Culex</em> oviposition surveillance sites. A maximum likelihood unsupervised classification in ArcInfo 9.2® revealed that the highest overall mosquito abundance was found in sites having a low-to-moderate range of built environment (40%) and high forest composition. A set of propagation equations were then designed to model the calibration uncertainties, which revealed that normalized difference vegetation index (NDVI), and two NDVI variants, were informative markers for the sampled mosquito data. Spatial dependence of the covariates of <em>Cx. restuans</em> and <em>Cx. pipien</em>s oviposition sites were indexed using semivariograms, which suggested that all main effects of the explanatory variables were statistically significant in the model. Additionally, a multispectral classification and digital elevation model-based geographical information system method were able to evaluate stream flow direction and accumulation for identification of terrain covariates associated with the sampled habitat data. These results demonstrate that remotely sensed operational indices can be used to identify parameters associated with field-sampled <em>Cx. pipiens</em> and <em>Cx. restuans</em> aquatic habitats