Little is known regarding the comparative source–sink relationships between primary mosquito breeding sites (source) and neighboring (sink) environments in heterogeneous landscapes. An exploration of those relationships may provide unique insights into the utility of open-space buffer zone mitigation strategies currently being considered by urban planners to reduce contact between mosquitoes and humans. We investigated the source–sink relationships between a highly productive mosquito habitat and adjacent residential (developed) and rural (undeveloped) coastal environments. Our results suggest that source–sink relationships are unaffected by environment. This conclusion is supported by the high level of synchronicity in daily saltmarsh mosquito abundance observed among all surveyed environments (β = 0.67–0.79, P < 0.001). This synchronicity occurred despite the uniqueness of each surveyed environment and the considerable distances of open water and land (2.2–2.6 km) between them. Trap catches, which we interpret as expected mosquito biting nuisance, were high in both residential and rural coastal landscapes (309.4 ± 52.84 and 405.3 ± 62.41 mosquitoes/day, respectively). These observations suggest that existing and planned open-space buffer zones will do little to reduce the biting burden caused by highly vagile saltmarsh mosquitoes. This strengthens the need for empirically informed planning guidelines that alert urban planners to the real risks of human residential encroachment on land that is close to highly vagile mosquito habitat.
Mosquito nuisance and the threat of disease is often given little consideration by urban planners, despite considerable research on the topic and the substantial costs to individuals and government (Dale 2010, Medlock and Vaux 2015, Dwyer et al. 2016). In the Australian southeast, a major source of mosquito nuisance in coastal developments is uninhabited islands dominated by a mixture of mangrove and saltmarsh habitat favorable to highly vagile Aedes vigilax (Skuse) (Knight et al. 2012). This and other saltmarsh-associated species represent a significant public health threat, since they are important vectors of the zoonotic arboviruses Ross River virus (RRV) and Barmah Forest virus (BFV) (Jacups et al. 2008, Naish et al. 2011). Owing to the combined health and nuisance risks that these mosquitoes present, considerable efforts are made to control their numbers through area-wide larviciding campaigns. However, inundating tides, dense mangrove canopy, and the occurrence of staggered brood emergence makes complete control extremely challenging (Johnson et al. 2020a). Urban planners are therefore considering alternative mitigation strategies such as open-space buffer zones (Webb and Russell 2019) and vegetation barriers that interrupt or reduce adult mosquito movement between larval habitats and adjacent residential areas.
The use of buffer zones assumes that the presence of refugia-free land around residential developments will discourage dispersal, yet despite their mention in regional mosquito management guidelines (Scott 2002), no firm guidelines on distance, area, positioning, or habitat type are provided, and the practice has limited empirical support (Webb and Russell 2019). Further, little is currently known regarding the comparative source–sink relationships between primary mosquito breeding habitat (source) and neighboring (sink) environments in heterogeneous urban and rural landscapes. This is important, since environments that contain a greater variety of refugia and harborage vegetation may serve as greater population sinks relative to more developed environments (Abella-Medrano et al. 2015, Roiz et al. 2015) by affecting the magnitude and frequency of mosquito dispersal events. Investigating such relationships will fill valuable knowledge gaps regarding the utility of buffer zones against highly vagile saltmarsh mosquitoes in coastal landscapes.
Here, we investigate source–sink relationships between a productive offshore (island) breeding site and adjacent residential and rural sites on the mainland. All study sites were located in the area of Redland City, Queensland, Australia, and separated by 2.2–2.6 km of land or open water (Fig. 1A, 1B). Patterns of daily mosquito abundance and host-seeking activity were analyzed in each study site using emerging smart trap technology (BG-Counter®; Biogents AG, Regensburg, Germany). Trap deployment, operation, and statistical analyses followed the methods outlined in Johnson et al. (2020b). Since the BG-Counter cannot differentiate mosquito species, estimated abundances of saltmarsh and freshwater species were determined by multiplying daily collection totals by the community dominance (%) of each species observed at each surveillance interval (2–10 days).
In total, data were collected for 66 complete days from February to May 2020. Repeated measures analysis of variance (one-way ANOVA) of daily collections revealed no significant (F64,128= 2.7, P = 0.08) difference in estimated saltmarsh mosquito abundance (Ae. vigilax, Culex sitiens Weid., and Ae. alternans (Westwood)) across all sites, but a stronger influence of freshwater species (predominantly Cx. quinquefasciatus Say and Cx. annulirostris Skuse) in the undeveloped site (23.72 ± 3.01% vs. 49.55 ± 9.65%; Fig. 2A, 2C). This is not surprising, since the area received a substantial amount of rainfall (626.2 mm) during the study period, which coincided with spring tide events that reduced saltmarsh emergence by heavy inundation and flooding. No notable collections of freshwater mosquitoes occurred within the island site. Estimated daily saltmarsh abundance was 324.2 ± 64.16, 309.4 ± 52.84, and 405.3 ± 62.41 for the island, residential, and undeveloped sites, respectively. Host-seeking activity (attraction to CO2 baited traps) was also similar among all sites (Fig. 2B), although elevated trap catches were observed during the period from 6:00 p.m. to 12:00 p.m. in the undeveloped site. This observation may be attributable to the higher percentage of freshwater species at this site, since freshwater Culex spp. have been observed to exhibit longer periods of host-seeking activity relative to saltmarsh Aedes spp. (Nasci and Edman 1981, Reisen et al. 1997). Of greatest note was the high degree of synchronicity in estimated daily saltmarsh mosquito abundance observed among all surveyed environments (β = 0.67–0.79, P < 0.001; Fig. 2D–F). The degree of population synchronicity is all the more impressive considering the uniqueness of each surveyed environment and the considerable distances between them.
The results from this study indicate that coastal land use (urban or rural) has little impact on the dispersal of saltmarsh mosquitoes from adjacent offshore mosquito habitat. This suggests that green space does not serve as a significant population sink for dispersing saltmarsh mosquitoes and that the incorporation of open-space buffer zones will do little to reduce nuisance from such species. However, further study is needed, since buffer zones may have some utility against other potential RRV vectors (Jansen et al. 2019) associated with brackish/freshwater swamps such as Verrallina funerea (Theobald) that disperse relatively short distances (<1 km; Dwyer et al. 2016). It is therefore imperative that the mitigation strategies proposed by urban planners are empirically informed and guided by local and regional mosquito ecologists and control professionals. This will require more effective communication between regional planning bodies, housing developers, and mosquito control groups. Such consultation would help revise planning directives that currently do little to consider the impacts of urban encroachment on or near productive mosquito habitat.
We thank the staff of the Redland City Council Pest Management team for their assistance with trapping mosquitoes and for providing transport to field sites. The Mosquito and Arbovirus Research Committee (Inc.), an independent Australian body involved in mosquito research made up of representatives from local governments, state government agencies, industry and scientific organizations, provided the funding for the study.
Mosquito Control Laboratory, QIMR Berghofer Medical Research Institute, 300 Herston Road, Brisbane, Queensland, 4006, Australia.
Pest Management, Redland City Council, Redland City Council, Cleveland, Queensland, 4163, Australia.