Comparative Analysis of Mosquito Trap Counts In the Peruvian Amazon: Effect of Trap Type and Other Covariates On Counts and Diversity.

Efficient detection of multiple species of adult mosquitoes in various habitats using effective traps is a crucial 1st step in any disease prevention program. Novel trap types that target tropical vectors of human diseases require field testing in the habitat of the vector-disease system in question. This paper analyzes a series of mosquito trapping studies conducted at Mapacocha, San Juan Bautista District, Loreto, Peru, during August-September 2013 and April-May 2014. Six trap configurations were evaluated in forest and rural locations. Adult mosquito counts were analyzed using full Bayesian inference of multilevel generalized linear models and posterior probability point estimates of the difference of means of the combined trap catch by trap type comparisons of all species. Light traps (Centers for Disease Control and Prevention [CDC] incandescent, white light-emitting diode [LED], and ultraviolet LED) caught greater numbers of mosquitoes compared with traps baited with yeast-generated CO and Biogents Sentinel traps (battery powered traps without light and passive box traps). However, diversity measures (species richness, evenness, and similarity) were consistently nearly equal among trap types. Arbovirus vectors were more common in forest locations, while malaria vectors were more common near human habitations. Location had a significant effect on trap effectiveness and mosquito diversity, with traps from forest locations having greater numbers and greater species richness, compared with traps set near human habitations. The results of this study will inform mosquito surveillance trap choices in remote regions of central South America, including regions with emerging tropical diseases, such and dengue and Zika virus.