|Home | About | Journals | Submit | Contact Us | Français|
A mark-recapture project examined dispersal and flight distances of Culex mosquitoes from a wastewater treatment plant in Albany, NY, during 2007 and 2008. A self-marking device was constructed to mark egressing mosquitoes with fluorescent marking powder. Mosquitoes were recaptured using 30 CDC miniature light traps located within a 2.0 km radius of the marking site. A total of 13 and 10 marked Culex mosquitoes were recaptured in 2007 and 2008, respectively. Culex mosquitoes traveled a minimum of 0.16 km, a maximum of 1.98 km and, following correction for decreasing trap density with distance, had a mean distance traveled of 1.33 km. Characterizing the dispersal patterns of these mosquitoes is important for understanding the distribution of West Nile virus and other pathogens.
Determining the flight dispersal of vector mosquitoes is required to understand the epidemiological patterns of mosquito-borne pathogens and to inform control measures. In the northeastern United States, Culex pipiens L. and Cx. restuans Theobald are the main enzootic vectors of West Nile virus (WNV) (Andreadis et al. 2001, 2004; Bernard et al. 2001; Kilpatrick et al. 2005). Mosquito surveillance for WNV and control therefore focuses on these two species. Despite this, very little is known about the flight dispersal of Cx. pipiens and Cx. restuans in this region. A recent mark-release-recapture study with the Cx. pipiens sibling species Cx. pipiens pallens Coquillet in an urban area in Japan demonstrated a mean distance traveled (MDT) of 0.42 km and a maximum of 1.2 km (Tsuda et al. 2008), yet the population and habitat differences between this study and the Northeastern United States are likely to significantly alter dispersal patterns. A previous study of Cx. nigripalpus in central Florida suggests this species generally travels <0.4 km (Nayar et al. 1980). The majority of Culex dispersal studies have examined flight patterns of Cx. quinquefasciatus Say or Cx. tarsalis Coquillet, which represent the primary vectors of WNV in Southern and Western United States. Studies with both species in Kern County, CA, demonstrated that marked mosquitoes most frequently traveled 1.0 km or less, yet flights as long as 12.6 km were documented, with the distance traveled per day estimated at 0.6–1.0 km (Reisen et al. 1991, 1992). A follow-up long-term study in Coachella Valley, CA, estimated daily travel at just 0.2 km/d and also demonstrated seasonal variance in mean distance traveled (Reisen and Lothrop 1995). Additional studies with Cx. quinquefasciatus, which is itself a member of the Cx. pipiens species complex, have reported maximum flight distances ranging from <1.0 to 2.1 km (Fussell 1964, Lindquist et al. 1967) and MDTs of 1.27 km (Schreiber et al. 1988) and 1.64 km (LaPointe 2008). Taken together, these studies demonstrate that dispersal patterns of different species and populations of similar species are variable, and suggest that spatial and temporal variations in both habitat and environmental conditions may shape such patterns. We assessed dispersal patterns and flight distances of Culex mosquitoes from a wastewater treatment facility in Albany County, NY. We used a self-marking device that, despite limiting our ability to track exact timing and number of marked mosquitoes, removed many of the biases common to mark-release-recapture studies such as ingrained flight patterns and the effect of time of day on dispersal (Reisen et al. 2003). Wastewater treatment facilities have previously been indicated as sources of mosquitoes in residential areas (Morris et al. 1991). The wastewater treatment site used in this study has historically produced a relatively large number of Culex mosquitoes per trap night (L. D. Kramer, unpublished data). Results provide insight into the patterns of WNV vector dispersal from source populations and the potential impact of control strategies on Culex populations in this area.
Mosquitoes were marked at a wastewater treatment plant using a modification of a self-marking device described by Niebylski and Meek (1989) (Fig. 1). Briefly, the base of the device was constructed of 2.54 cm PVC piping and a piece of 1.27 cm conduit was placed in PVC cross-pieces to provide support. The upper frame was made of 1.27 cm PVC and this frame supported the exit grid. Fiberglass screening was placed over the entire device with a hole to allow mosquitoes to leave through the exit grid. The exit grid was made of 5.08 × 5.08 cm pieces of wood with holes drilled to support 46.4 cm steel rods. The wood and rods were held together by wing nuts and locking bolts placed on opposite ends of two 1.27 cm threaded rods. Cheesecloth impregnated with bioluminescent powder (Bioquip, Rancho Dominguez, CA) was used to mark mosquitoes. Cloth was folded in half and a 2.54 cm wide piece of fabric was sewn on to create a pocket, which allowed the cloth to be easily slid onto the rods. Mosquitoes leaving through the exit grid were marked with fluorescent powder as they landed on or brushed into the cheesecloth. A 1.83 × 0.61 m tarpaulin was suspended over the marking device to prevent powder from getting wet.
Two such devices were constructed and placed over a 1.83 × 21.34 m overflow tank at a wastewater treatment facility, which was previously observed to be a Culex oviposition site (Fig. 1). This site is located in the Sewer District in the Northeast corner of the city of Albany, NY. This is an urban location that is surrounded by rural areas to the north and east and has historically produced large numbers of Culex per trap night during peak activity (up to 24.5 Culex per trap night in ground level and 66.0 in elevated Centers for Disease Control [CDC] light traps baited with dry ice; L. D. Kramer, unpublished data). Any open space between the devices was covered using fiberglass screen and tied down with rope and tent stakes. Bricks were also placed on the screen to help secure and prevent mosquitoes from escaping. In addition, a wind direction and velocity meter (Onset, Bourne, MA), which measured at 15 min intervals, was placed at the water treatment facility for the entirety of the study.
The marking efficiency and adult mortality caused by the device were pretested in a laboratory setting. A 2 × 1 foot collection container was secured to the self marking device to capture mosquitoes as they exited. One hundred Cx. pipiens larvae were placed underneath the marking device where they were allowed to emerge and exit the marking device into the collection container. Mosquitoes were removed from the collection container daily and examined for fluorescent powder until all adults had emerged.
This project was carried out over two time periods in 2007 and one time period in 2008. In 2007, early season (05 June through 21 June) and peak season (24 July through 22 August) mark-recapture trials were completed, and in 2008, a single trial was completed during peak activity (22 July through 16 August). A different colored powder was used on the marking device each week (yellow, red, and blue in weeks 1–3, respectively) to separate weekly cohorts to estimate survival and distance traveled since marking. One variation to this weekly change was made when the yellow marking was extended through 4 August in 2007 (rather than 31 July) because of the high number of marked mosquitoes resting at the site. This allowed time for mosquitoes to disperse before initiating the next color scheme. During each time period the devices were set up on Mondays and removed on Friday mornings to allow additional mosquitoes to lay eggs in the tank. Thirty CDC light traps baited with dry ice were placed at 1–2 m heights within a 2.0 km radius of the marking location and were operated on three consecutive nights (Tuesday through Thursday) each week, with the exception of the last week of the 2007 study when traps were only run for two nights (20–21 August 2007). Traps were dispersed throughout the study area and the locations were fixed within and among years. From nearest to most distant, a total of 5, 6, 8, and 11 traps, respectively, were set within each 0.5 km annuli. Collections were transported to the Wadsworth Center Arbovirus laboratory where mosquitoes were immobilized with dry ice and transferred to a chill table. All mosquitoes were examined for the presence of fluorescent powder using a UV flashlight (Bioquip) and a microscope in a dark room. Female mosquitoes were identified morphologically using the identification keys of Stojanovich (1961) and Means (1979, 1987). In 2007, marked Culex mosquitoes were placed in individual tubes filled with 100% EtOH, whereas all other mosquitoes were sorted by gender and species, pooled in groups of 50 or fewer, placed in labeled microcentrifuge tubes, and stored at −70°C. Cx. pipiens, Cx. restuans, and Cx. salinarius are difficult to distinguish as adults and were therefore combined (Culex spp.). In 2007, molecular identifications were performed using polymerase chain reaction with species-specific primer sets to determine the relative abundance of each species of Culex (Crabtree et al. 1995, Drummond et al. 2006).
The trapping area surrounding the release point was divided into four annuli, which were each separated by 0.5 km (Figs. 2). The MDT was calculated using a correction factor (Lillie et al. 1981, 1985; White and Morris 1985). Specifically, MDT is equivalent to the summation (for each annulus) of the estimated recaptures * the median distance of the annulus/the total number of estimated recaptures, where the estimated recaptures = the number of observed recaptures/the number of traps * the correction factor, and the correction factor = the area of the annulus * the number of traps in the area/the total trapping area. This method allows the determination of the MDT without knowledge of the number of mosquitoes marked and a correction for densities of traps in each annulus (Morris et al. 1991). In addition, Culex captured per trap-night was calculated.
In the first period of 2007 (05–21 June), 19,809 mosquitoes were captured and sorted by species. Among these were 524 Culex spp., representing 2.6% of the total catch. The most prevalent species captured during this trial was Aedes sticticus Meigen, representing 53% of the catch. Other species trapped included Ae. canandensis Theobald, Ae. cinereus Meigen, Ae. japonicus Theobald, Coquillettidia perturbans Walker, Psorophora ferox Humbodt, Ae. provocans Walker, An. punctipennis Say, Anopheles quadrimaculatus Say, Ae. stimulans Walker, Cx. territans Walker, Ae. triseriatus Say, Ae. trivittatus Coquillet, Uranotaenia sapphirina Sacken, and Ae. vexans Meigen.
During the second period (24 July through 22 August) a total of 21,166 mosquitoes were captured including 3,326 Culex spp. This represented 15.7% of the total catch, a substantial increase from the June trial, both in absolute number and Culex captured per trap-night (Fig. 3). In total, 84% of the Culex spp. were successfully speciated using molecular techniques, indicating 934 (34%) were Cx. pipiens, 513 (18%) were Cx. restuans, and 1,337 (48%) were Cx. salinarius Coquillet. The proportion of each species was variable among different trapping locations (Fig. 4). In particular, trapping sites along the river tended to be predominantly Cx. salinarius, whereas urban sites to the south were dominated by Cx. pipiens and more rural sites to the north and west tended to yield more Cx. restuans. All species trapped in June also were trapped during July and August, with the addition of Ae. atropalpus Coquillett and Culiseta minnesotae Barr.
Laboratory testing of the marking device demonstrated that marking efficiency was high (>95%) for exiting mosquitoes. In addition, no immediate mortality or differences in longevity relative to unmarked mosquitoes were apparent. In the first period of 2007, 12 mosquitoes marked with fluorescent powder were recaptured, including 6 Cq. perturbans, 3 Ur. sapphirina, 2 Ae. sticticus, 1 Ae. canadensis, and a single Culex spp. trapped at 0.32 km from the marking site. Many non-Culex species that became marked were unlikely to have emerged from holding tanks as it is not typical habitat for oviposition or hatching of eggs from, for example, floodwater species. These species may have acquired fluorescent powder when resting on the marking device after emergence from other sites. For this reason, MDT calculations were only reported for Culex spp.
Twenty-eight marked mosquitoes were captured during the second trial of 2007. These included 13 Culex spp. mosquitoes (10 Cx. pipiens and 3 Cx. salinarius; Table 1). Cx. pipiens were recaptured at 10 different sites, with a minimum distance traveled of 0.16 km and a maximum of 1.98 km. The MDT for Cx. pipiens in 2007 was 1.37 km. Cx. salinarius were captured at three different sites with a minimum flight distance of 0.32 km and a maximum of 1.34 km. The MDT for Cx. salinarius was 1.10 km (Table 1). Eleven of the 13 marked Culex mosquitoes trapped during the second trial of 2007 were marked with yellow powder, indicating that they were marked between 24 July and 4 August. These mosquitoes were trapped 8 August through 14 August. Two Culex mosquitoes marked with red powder were trapped in 2007, indicating they were marked 7 August through 11 August. These mosquitoes were captured on 13 August and 14 August, at 1.34 and 0.32 km, respectively. As demonstrated with these recaptures, there was no general trend between date recaptured and distance from the marking site. Although blue marking powder was used 14 August through 18 August in 2007 and trapping continued through 22 August, there were no captures of Culex or mosquitoes marked with blue fluorescence.
During peak Culex mosquito activity in 2007 (trial 2), the mean wind direction was 32° from the NW and the mean velocity was 3.4 km/h. On 12 of the 13 d in which marked mosquitoes were trapped in the second period of 2007, the prevailing winds were northerly. Despite this, just 4 of the 12 marked mosquitoes were trapped south of the marking location. Temperatures during this period were close to historical averages (mean of 22.3°C) with significant precipitation (>0.02”) occurring on just four of the 30 d of trapping and a total of 3.24” of rainfall.
In an effort to increase the success of tracking Culex dispersal, a single mark-recapture trial was completed in 2008 that corresponded to the period of peak activity in 2007 (22 July through 16 August). During this period 31,361 mosquitoes were trapped, including 4,164 Culex spp. Although the absolute number of mosquitoes was substantially higher in 2008 (10,195 more mosquitoes trapped in 2008), the percent of Culex spp. in the catch (13.3%) was similar to the same period in 2007. The total number of Culex caught was comparable to 2007, yet because trapping was carried out over 3 wk (2008) rather than four (2007), there was a substantial increase in Culex captured per trap-night in the former study (Fig. 3). The overall increase in absolute number of trapped mosquitoes in 2008 was attributed primarily to the abundance of Ae. trivittatus, which accounted for 53% of the catch and increased by 9,683 relative to 2007. All species trapped in 2007 were again trapped in 2008, with the exception of Cx. territans and Ae. provocans.
In total, 13 marked mosquitoes were captured in 2008 (Table 1). Among these were 10 Culex spp., 1 Ae. stimulans, 1 Ae trivittatus, and 1 Ae. punctipennis. In 2008, the minimum distance traveled for a Culex mosquito was 0.2 km, the maximum distance was 1.59 km, and the MDT was 1.35 km. When 2007 and 2008 data are combined, the MDT based on the 23 marked Culex spp. captured was 1.33 km (Table 1).
Unlike 2007, only two Culex spp. mosquitoes marked with yellow powder were trapped in 2008, yet the yellow marking period lasted just one week in 2007 rather than two in 2008. These two mosquitoes were trapped on 24 July at 0.63 and 0.20 km from the marking sight. Because yellow marking occurred between 22 July and 26 July, these mosquitoes were marked at most 2 d before trapping. Red powder was used for marking from 29 July to 2 August, and between 30 July and 12 August, five Culex spp. mosquitoes with red fluorescence were trapped. Similar to 2007, date trapped did not necessarily correspond to days since marking. On 30 July 2008, a Culex mosquito marked with red powder was captured 1.34 km from the marking site. Because red marking commenced on 29 July, this mosquito was marked at most 1 d before capture. Two Culex spp. mosquitoes marked with blue fluorescence were trapped in 2008. Marking with blue powder occurred 5 August through 9 August and captures occurred on 14 August and 15 August at 1.59 and 0.71 km, respectively. Similar to 2007, distances traveled during different marking periods were not indicative of a temporal change in MDTs.
During the 2008 trapping period the average wind direction was 33° NW and the mean velocity was 4.9 km/h, which was very similar to the same period in 2007. As with 2007, wind directions on days of or preceding trapping of marked mosquitoes did not necessarily correspond to the location of the trap relative to the marking location. For example, the marked mosquito trapped on 12 August was caught 1.27 km southeast of the marking sight, yet the prevailing winds on 8 August through 14 August were northerly. Temperatures during the 2008 trapping period were similar to 2007 and historic averages (mean temperature of 21.8°C), yet precipitation was substantially greater, with eight total days of significant rainfall and a total of 8.17”.
The mosquito marking device we created (Fig. 1; modification of Niebylski and Meek, 1989) successfully marked Cx. pipiens at a water filter plant in upstate NY to allow a determination of flight distance. However, the efficiency of recapture using this method in this area was only effective during peak Culex activity. Only one marked Culex mosquito was captured during the first trapping period of 2007; this is likely because of the relatively low population of Culex present, the existence of multiple sources of Culex throughout the trapping area, and the relatively low trapping efficiency resulting from both density of traps and frequency of trapping. The weekly average of Culex spp. captured per trap-night (30 CDC light traps) was just 2.1 during the first trapping period of 2007, but increased to 8.2 in the second (July and August; Fig. 3). In conjunction with this rise in overall trapping, an increase from 1 to 13 marked Culex spp. recaptures was obtained. This demonstrated the need to focus on peak activity to track dispersal in 2008. During the 2008 trapping period, Culex per trap-night increased to 14.1 (Fig. 3). This overall increase in Culex spp. in 2008 can likely be attributed to more favorable environmental conditions for these mosquitoes, possibly as a result of an increase in heavy precipitation followed by dry periods. Despite this general population rise, only 10 marked Culex were captured in 2008. Given the population increase in 2008, this resulted in a twofold decrease in the number of marked Culex trapped per total Culex trapped. Although the total number of marked mosquitoes could not be quantified using these methods, this decrease is likely because of an increased number of source populations resulting from the abundance of standing water in 2008.
The recapture of 23 Culex spp. mosquitoes during peak activity in 2007 and 2008 allowed for the calculation of MDT, which has not been reported previously in the United States for these species. Because calculated MDTs for both years were comparable, data were combined to increase the sample size and overall accuracy of the estimation. The MDT for all Culex mosquitoes captured was 1.33 km (Table 1), but captures occurred at distances as close as 0.16 km and as far away as 1.98 km (Fig. 2). Both the MDT and maximum flight distance are modestly higher than those reported for Cx. quinquefasciatus by Schreiber et al. (1988) and Fussel (1964) and substantially higher than the MDT of 0.42 km reported for Cx. pipiens pallens in urban Japan. Milby et al. (1991) reported capturing Cx. quinquefasciatus at distances as great as 10.9 km in Kern County, CA, although the landscape ecology would have been very different from upstate NY. This information along with the fact that we captured marked Culex spp. at the outer limits of our trapping area suggests that the maximum distance traveled by marked females was likely greater than we observed.
In 2007, speciation of all Culex spp. trapped was done to compare both trap specific proportions and MDTs among species. The fact that 10 of 13 marked mosquitoes captured in 2007 were Cx. pipiens was not surprising given that the marking sight was relatively urban and such sites are generally dominated by Cx. pipiens (Diuk-Wassner et al. 2006). This was further supported by our speciation efforts of all trapped Culex spp., which demonstrated traps in urban areas were dominated by Cx. pipiens and those in more rural areas by Cx. restuans (Fig. 4). Cx. salinarius, however, tended to be more prevalent in close proximity to the Hudson River and surrounding areas of exposed surface water, which is consistent with previous reports of the preferred habitat of this species (Slaff and Crans 1982, Diuk-Wassner et al. 2006). The recapture of three marked Cx. salinarius allowed us to compare MDTs among two Culex species in 2007. These data suggested that Cx. pipiens may on average travel greater distances than Cx. salinarius (1.37 vs. 1.10 km in 2007), yet more data will be required to confirm this result and increase its precision.
The self-marking device that we used in this study allowed for natural dispersal without the biases which often arise from controlled releases. Although these methods did not allow for the exact timing or quantification of the number of marked mosquitoes, the use of different marking colors did provide some insight into the rates of dispersal. In 2007, yellow marked mosquitoes trapped on 14 August had been marked 10 or more days prior. This does not necessarily indicate 10 or more days of dispersal, as mosquitoes were often observed resting on or near the device following marking. In addition to this, no blue marked mosquitoes were trapped in 2007, despite trapping efforts continuing 8 d beyond the commencement of blue marking. However, this was not always the case, as a blue marked mosquito was trapped in 2008 just 1 d after the commencement of marking 1.34 km from the marking site. This indicated that Culex mosquitoes were capable of traveling at least this distance (and likely further) in a single day. However, the fact that marked Culex mosquitoes were trapped as close as 0.16 km multiple days after marking clearly demonstrated the variability in daily distances traveled. Although other studies have reported daily travel of Culex mosquitoes ranging from 0.2 to 1.0 km (Reisen et al. 1990, Reisen and Lothrop 1995), the variation in results here, even with the limited number of recaptures, cautions against attempting to estimate such values.
Culex mosquitoes generally traveled away from their marking location along the Hudson River (Fig. 2). This result is similar to that reported by Milby et al. (1991) in which Cx. tarsalis generally dispersed along the river channel. Additional studies with Cx. tarsalis suggest that dispersal patterns are primarily correlated with wind (Bailey et al. 1965, Reisen and Lothrop 1995), while others demonstrate that dispersal may be retarded by elevated winds (Reisen et al. 2003). Results here demonstrate dispersal patterns which are not consistently correlated with wind patterns, arguing against wind-assisted travel as the primary mechanism of dispersal. Others have reported that in the laboratory, Cx. pipiens exhibited a directional flight preference for the northeast regardless of the light regime, perhaps as a result of magnetic field navigation (Tanabe 1972). In our study, Culex mosquitoes most often dispersed either to the northeast or to the southwest, suggesting that additional ecological and meteorological influences, likely related to cues from the surrounding river environment, drive dispersal in this area.
Our study suggests that wastewater treatment facilities can act as a source of Culex mosquitoes for surrounding areas as far as 1.98 km away and likely further. The 10 marked Cx. pipiens recaptured in 2007 accounted for 0.89% of the overall catch of Cx. pipiens. Although this represents less than one percent of the catch, the tank we used was just one of many likely breeding locations within the facility. In addition, the decrease in trapping of marked mosquitoes relative to total population from 2007 to 2008 demonstrates, not surprisingly, that control efforts at such facilities would be more effective in years when natural oviposition sites are more limited. Therefore, depending upon both the seasonal conditions and the proximity of water treatment facilities to populated areas, control at such facilities may be warranted, especially in areas where pathogens such as WNV are endemic.
The authors thank Caroline Lego, Pamela Chin, and Jennifer Dawson for technical assistance. This work was supported by National Institutes of Health Contract #N01 (AI)-25490.