Many mosquito control agencies use carbon dioxide-baited traps as surveillance tools for adult vector populations. However, decisions regarding the number and location of trap sites and the frequency of collections are often based on logistical issues, and not on the bionomics or spatial distribution of the target species. Therefore, with the aim of providing practical information for adult mosquito surveillance programs, we used an array of 81 carbon dioxide- and octenol-baited lights traps to obtain weekly samples of adult mosquitoes in Redland Shire in southeastern Queensland, Australia. The spatial patterns of four different mosquito species were examined, and positive spatial autocorrelation in trap counts was evident for Ochlerotatus vigilax (Skuse), Coquillettidia linealis (Skuse), and Culex annulirostris Skuse, but not for the container species Ochlerotatus notoscriptus (Skuse). Of the three species that exhibited spatially correlated trap counts, the autocorrelation was greatest in Oc. vigilax at a lag distance of 0–1.5 km, with Moran’s I values of 0.30–0.64. Moran’s I indices were also positive and statistically significant (P < 0.05) at lag distances of 1.5–3.0 and 3.0–4.5 km on each of the 15 sampling occasions. However, at 3.0–4.5 km the Moran’s I values were low, which indicated only weak spatial autocorrelation in trap counts. Universal kriging was used to estimate the numbers of each species at unsampled locations throughout the study area, and leave-one-out cross validation analyses indicated that this was a robust method for Cq. linealis and Oc. vigilax. In contrast, trap counts for the container-breeding species Oc. notoscriptus were randomly distributed and the interpolated counts were not reliable. Comparisons of weekly contour maps of adult mosquito counts indicated a consistent spatial pattern for Oc. vigilax and Cq. linealis. Particular geographic areas had consistently high or low numbers of vectors, and these patterns were stable from year to year. Definition of geographic areas with consistently high or low numbers of vectors may allow control activities to be focused in areas with the greatest risk of arbovirus transmission.