Fig. 3

Effect sizes of selected abiotic environmental covariates used to describe the spatial distribution of the four mosquito species estimated using a full model. Points give the maximum a posteriori estimates of effect sizes and grey lines give the associated 95 % credible intervals. Credible intervals which do not cross zero signify that the posterior probability that the effect size is 0 or of the opposite sign is less than 0.05, analagous to statistical significance at the 5 % level in a frequentist statistical test. Effect sizes are divided into those for continuous variables (depth, temperature, oxidation-reduction potential (ORP) and salinity) and for discrete variables (plant type), since effect sizes cannot be directly compared for the two different types of data