Culicoides Midge Abundance across Years: Modeling Inter-Annual Variation for an Avian Feeder and a Candidate Vector of Hemorrhagic Diseases in Farmed Wildlife.

ABSTRACT

(1) Background: Epizootic hemorrhagic disease virus (EHDV) and bluetongue virus (BTV) are orbiviruses that cause hemorrhagic disease (HD) with significant economic and population health impacts on domestic livestock and wildlife. In the United States, white-tailed deer (Odocoileus virginianus) are particularly susceptible to these viruses and are a frequent blood meal host for various species of Culicoides biting midges (Diptera Ceratopogonidae) that transmit orbiviruses. The species of Culicoides that transmit EHDV and BTV vary between regions, and larval habitats can differ widely between vector species. Understanding how midges are distributed across landscapes can inform HD virus transmission risk on a local scale, allowing for improved animal management plans to avoid suspected high-risk areas or target these areas for insecticide control. (2) Methods: We used occupancy modeling to estimate the abundance of gravid (egg-laden) and parous (most likely to transmit the virus) females of two putative vector species, C. stellifer and C. venustus, and one species, C. haematopotus, that was not considered a putative vector. We developed a universal model to determine habitat preferences, then mapped a predicted weekly midge abundance during the HD transmission seasons in 2015 (July-October) and 2016 (May-October) in Florida. (3) Results: We found differences in habitat preferences and spatial distribution between the parous and gravid states for C. haematopotus and C. stellifer. Gravid midges preferred areas close to water on the border of well and poorly drained soil. They also preferred mixed bottomland hardwood habitats, whereas parous midges appeared less selective of habitat. (4) Conclusions: If C. stellifer is confirmed as an EHDV vector in this region, the distinct spatial and abundance patterns between species and physiological states suggest that the HD risk is non-random across the study area.