CO-OCCURRENCE OF MOSQUITO COMMUNITIES IN DISTURBED ENVIRONMENTS USING MARKOV RANDOM FIELDS (MRFs) IN ST. JOHNS COUNTY, FLORIDA

Abstract

The distribution of mosquito communities is predicted by complex micro- and macrohabitat systems. While macrohabitat variables are significant in modeling the distribution of individual mosquito species, the distribution of mosquito communities in disturbed urban and semi-urban environmental gradients was overlooked in most of the previous models. In our study, we used conditional Markov Random Fields (CRF) to evaluate spatial co-occurrence patterns between mosquito vectors of eastern equine encephalitis (EEEV) and west Nile virus (WNV) in a disturbed urban environment in Saint John’s County, Florida. We aimed to 1) quantify the strength and direction of spatial unconditional and conditional correlations between mosquito assemblages in disturbed environments, and 2) evaluate whether the strength of correlations between mosquito assemblages is conditional on landscape or climate variables. We leveraged the longitudinal surveillance effort using Biogents sentinel traps (BGS) conducted by Anastasia Mosquito Control Districts in disturbed urban environments during 2017-2020. The distribution of high mosquito abundance, especially Aedes albopictus, Ae. aegypti, Ae. vexans, Ae. taeniorhynchus, Culex nigripalpus, Cx. salinarius, and Cx. quinquefasciatus, were conditionally correlated with other EEEV and WNV vector species in reduced woody and herbaceous wetlands and evergreen forests (-54.44%), and in urban developed landscapes (3.44%) during 2019 and 2020. Moreover, conditional correlations between mosquito species pairs were positively associated with increased total precipitation and in areas with high average minimum and maximum temperatures. Our results show that the micro- and macrohabitat characteristics demonstrated spatial effects on distribution and correlations between species pairs of EEEV and WNV mosquito vectors across disturbed environments. Our findings could be used to better understand the joint effects of drivers on mosquito diversity at a specific locality, interspecific interactions among mosquito assemblages, and how this diversity changes across environmental gradients.