Vol. 34, no. 2 Journal of Vector Ecology 317 A geospatial evaluation of Aedes vigilax larval control efforts across a coastal wetland, Northern Territory, Australia N. Kurucz1, P. I. Whelan1, J.M. Carter1, and S.P. Jacups2 1Medical Entomology, Centre for Disease Control, Department of Health and Families, Darwin, Northern Territory, Australia 2School for Environmental Research, Institute of Advanced Studies, Charles Darwin University, Darwin, Northern Territory, Australia Received 19 May 2009; Accepted 18 September 2009 ABSTRACT: Adjacent to the northern suburbs of Darwin is a coastal wetland that contains important larval habitats for Aedes vigilax (Skuse), the northern salt marsh mosquito. This species is a vector for Ross River virus and Barmah Forest virus, as well as an appreciable human pest. In order to improve aerial larval control efforts, we sought to identify the most important vegetation categories and climatic/seasonal aspects associated with control operations in these wetlands. By using a generalized linear model to compare aerial control for each vegetation category, we found that Schoenoplectus/ mangrove areas require the greatest amount of control for tide-only events (30.1%), and also extensive control for tide and rain events coinciding (18.2%). Our results further indicate that tide-affected reticulate vegetation indicated by the marsh grasses Sporobolus virginicus and Xerochloa imberbis require extensive control for Ae. vigilax larvae after rain-only events (44.7%), and tide and rain events coinciding (38.0%). The analyses of vector control efforts by month indicated that September to January, with a peak in November and December, required the most control. A companion paper identifies the vegetation categories most associated with Aedes vigilax larvae population densities in the coastal wetland. To maximize the efficiency of aerial salt marsh mosquito control operations in northern Australia, aerial control efforts should concentrate on the vegetation categories with high larval densities between September and January. Journal of Vector Ecology 34 (2): 317-323. 2009. Keyword Index: Aedes vigilax, aerial mosquito control, landscape ecology. INTRODUCTION mosquito larval control programs were initiated in 1986 to reduce the risk of arbovirus transmission and the number of Darwin is the largest city in the Northern Territory pest complaints in the adjacent residential areas (Brogan et (NT) of Australia, located on the north coast (Figure 1), al. 2002, Whelan 2007). This paper outlines the vegetation with a population of approximately 105,991 (Australian categories and seasonal conditions most associated with Bureau of Statistics 2006). The Darwin area experiences a aerial control efforts for Ae. vigilax larvae in this wetland. monsoonal climate, with a high average annual rainfall of A companion paper addresses vegetation category 1,708 mm (Australian Bureau of Meteorology 2008), and associations with Ae. vigilax larvae densities in the wetland most rainfall occurs during the wet season from November (Jacups et al. 2009). By identifying the most significant to April. Some of Darwin’s northern suburbs are adjacent vegetation categories and seasonal conditions associated to an extensive coastal marsh and upper mangrove wetland with Ae. vigilax breeding, survey time can be reduced and that experiences seasonal tidal inundation providing the efficiency of aerial salt marsh mosquito larval control suitable breeding conditions for the northern salt marsh can be maximized. mosquito Aedes vigilax (Skuse) (Whelan 1989) (Figure 1). Aedes vigilax is recognized as a major human pest species MATERIALS AND METHODS (Mackerras 1926, Marks 1967, Gislason and Russell 1997, Webb and Russell 1999) and is also a major vector for GIS analysis Ross River virus and Barmah Forest virus in the Northern Aerial larval mosquito survey and control operations Territory (Merianos et al. 1992, Tai et al. 1993, Whelan et al. have been conducted in the coastal wetlands within a 5 1997, Russell 1998, Jacups et al. 2008). These two arboviruses km radius of Darwin’s northern suburbs margin; namely constitute the majority of arbovirus infections in Australia Leanyer, Holmes Jungle, Micket Creek, and Shoal Bay (Russell 1994, 1995). Thus, control of Ae. vigilax is of great swamps (Figures 1 and 2). We determined the vegetation importance for disease prevention and nuisance reduction categories most associated with Ae. vigilax aerial control in the NT and other areas in Australia. from July, 2000 to June, 2007. Aerial control primarily Medical Entomology in Darwin conducts integrated used large volume water-based Bacillus thuringiensis var. mosquito control measures for Ae. vigilax breeding in the israelensis (Bti) mist applications applied by helicopter. wetland. Extensive engineering works completed in 1985 Control operations for other mosquito species were have led to improved drainage; both aerial and ground excluded. 318 Journal of Vector Ecology December 2009 Figure 1. Study area proximity to Darwin urban residential area within Australia. The wetland was divided into discernible vegetation category represented. Thus, a generalized linear model using categories based on previous vegetation surveys and the use a binominal distribution with logit link was applied to the of recent aerial photographs with ground truthing (Figure aerial control data. Three models were created, one for each 2) (Whelan 1989, Hayes 2000). Vegetation samples of each of the control datasets; tide-only events (Model 1), rain- category were formally identified by the NT Herbarium only events (Model 2), and rain and tide events coinciding (Table 1) (Northern Territory Government 2008). (Model 3). The vegetation category least associated withAe. Helicopter surveys were conducted after monthly high tides vigilax larval abundance “lower mangroves”, was chosen or rain events (>10 mm) for retained water and the presence as the reference category (Jacups et al. 2009). Similarly, of mosquito larvae. Details of aerial control were recorded March had the lowest aerial control activity recorded, for each vegetation category and subsequently mapped and and thus became the reference month for aerial control recorded as feature themes using ArcGIS (ArcGIS 3.2). The measures for vegetation categories by month. For ease of area of each vegetation category controlled for mosquito comparison between vegetation categories, we report odds breeding was calculated by intersecting each vegetation ratios (OR) rather than coefficients. All statistical analysis category with the control feature. Control areas were were performed using Intercooled Stata 10.0 (Stata Corp., separated into tide-only, rain-only, and tide and rain events College Station, TX). coinciding, with the latter defined as tide events plus >10 mm rain occurring within four days of the control event RESULTS (Table 1). All salt marsh mosquito larval control operations from July, 2000 to June, 2007 were included. Cumulated total areas controlled for Ae. vigilax larvae using aerial spraying are presented (Table 1). Schoenoplectus/ Environmental data mangroves (i) received the highest amount of control for Sea level data were provided by the Bureau of tide-only control efforts by total area (765.18 ha, 30.1%), Meteorology (BOM), with tidal maximums observed followed by the tide-affected reticulate (b) category during aerial survey operations and topographical data (indicated by the marsh grasses Sporobolus virginicus and used to indicate tide inundation levels. High tides variably Xerochloa imberbis) (679.7 ha, 26.7%) (Table 1, Figure 2). inundated the various vegetation categories in the wetland; Tide-affected reticulate vegetation (b) received the highest with a low (> 7.4 m), medium (> 7.6 m) and high (> 7.8 m) level of control (908.87 ha, 44.7%) for rain-only and for monthly tide. In addition, highest average tides >7.5 m were rain and tides coinciding control efforts (919.63 ha, 38.0%) calculated for each month using BOM tide data over the (Table 1, Figure 2), indicating that tide-affected reticulate study period (Australian Bureau of Meteorology 2006). vegetation is the most important area for salt-marsh mosquito control with respect to total area sprayed. Data analyses- vector control Results of the binomial models adjusted for the total Total areas controlled by aerial larvicide applications area for each vegetation category are presented in Table were cumulatively collated over the study period (Table 2. Control measures were most strongly associated with 1). Control efforts for each vegetation category were then Schoenoplectus/mangrove (i) vegetation (OR = 46.8) for statistically compared as a proportion of the total area for tide-only events (Model 1, Table 2). All vegetation categories that vegetation category, independent of the total area each received significantly more control efforts (P<0.01) than the Vol. 34, no. 2 Journal of Vector Ecology 319 Table 1. Vegetation categories and areas controlled in the northern Darwin coastal wetlands, July, 2000 to June, 2007. Total area Vegetation Dominant plant species Tide* (ha) Total area (ha) Rain** (ha) T&R*** (ha) controlled categories present (% total tide) (ha) Model 1 2 3 Lower Rhizophora stylosa a 908.35(28.9%) 173.66(6.8%) 45.87(2.3%) 62.44(2.6%) 281.97 mangroves Avicennia marina Reticulate tide- Xerochloa imberbis b 190.7(6.1%) 679.7(26.7%) 908.87(44.7%) 919.63(38.0%) 2508.2 affected Sporobolus virginicus Sporobolus virginicus Reticulate Xerochloa imberbis c 64.03(2.0%)