1 Journal of Medical Entomology Saul Lozano-Fuentes
2 SAMPLING, DISTRIBUTION, Department of Microbiology, Immunology
3 DISPERSAL and Pathology,
4 Colorado State University,
5 LOZANO-FUENTES ET AL.: 1690 Campus Delivery
6 TEMPORAL CHANGES IN Fort Collins, CO 80523
7 Aedes aegypti ABUNDANCE Phone: (970) 491 8745
8 E-mail: [email protected]
9
10 Intra-Annual Changes in Abundance of Aedes (Stegomyia) aegypti and Aedes (Ochlerotatus)
11 epactius (Diptera: Culicidae) in High-Elevation Communities in México
12
13 SAUL LOZANO-FUENTES1,4, CARLOS WELSH-RODRIGUEZ2, ANDREW J.
14 MONAGHAN3, DANIEL F. STEINHOFF3, CAROLINA OCHOA-MARTINEZ2, BERENICE
2 3 1 15 TAPIA-SANTOS , MARY H. HAYDEN , AND LARS EISEN
16 17 1 Department of Microbiology, Immunology and Pathology, Colorado State University, 3185
18 Rampart Road, Fort Collins, CO 80523.
19 2 Centro de Ciencias de la Tierra, Universidad Veracruzana, Calle Francisco J. Moreno 207,
20 Colonia Emiliano Zapata, Xalapa, Veracruz, Mexico C.P. 91090.
21 3 National Center for Atmospheric Research, P.O. Box 3000, Boulder, CO 80307.
22 4 Corresponding author, e-mail: [email protected]
23
1
24 Abstract
25 We examined temporal changes in the abundance of the mosquitoes Aedes (Stegomyia) aegypti
26 (L.) and Aedes (Ochlerotatus) epactius Dyar & Knab from June – October 2012 in one reference
27 community at lower elevation (Rio Blanco; ~1,270 m) and three high-elevation communities
28 (Acultzingo, Maltrata, and Puebla City; 1,670-2,150 m) in Veracruz and Puebla States, México.
29 The combination of surveys for pupae in water-filled containers and trapping of adults, using
30 BG-Sentinel traps baited with the BG-Lure, corroborated previous data from 2011 showing that
31 Ae. aegypti is present at low abundance up to 2,150 m in this part of México. Data for Ae.
32 aegypti adults captured through repeated trapping in fixed sites in Acultzingo – the highest
33 elevation community (~1,670 m) from which the temporal intra-annual abundance pattern for Ae.
34 aegypti has been described – showed a gradual increase from low numbers in June to a peak
35 occurring in late August and thereafter declining numbers in September. Aedes epactius adults
36 were collected repeatedly in BG-Sentinel traps in all four study communities; this is the first
37 recorded collection of this species with a trap aiming specifically to collect human-biting
38 mosquitoes. We also present the first description of the temporal abundance pattern for Ae.
39 epactius across an elevation gradient: peak abundance was reached in mid-July in the lowest
40 elevation community (Rio Blanco) but not until mid-September in the highest elevation one
41 (Puebla City). Finally, we present data for meteorological conditions (mean temperature and
42 rainfall) in the examined communities during the study period, and for a cumulative measure of
43 the abundance of adults over the full sampling period.
44 Keywords: Aedes aegypti, Aedes epactius, abundance, temporal changes, high elevation, México
2
45 Introduction
46 The mosquito Aedes (Stegomyia) aegypti (L.) is a primary vector of dengue, yellow fever and
47 chikungunya viruses (Gratz 1999, Gubler 2002, Weaver and Reisen 2010). We reported
48 previously on the collection of Ae. aegypti at high elevation (1,600-2,100 m above sea level) in
49 Veracruz State and Puebla State, México, based on recovery of immatures from water-holding
50 containers on residential premises and species identification conducted after rearing to the adult
51 stage (Lozano-Fuentes et al. 2012a). Those collection records extended the known elevation
52 range for Ae. aegypti in México by >300 m, as previously published studies had not reported this
53 mosquito species, or local dengue virus transmission, above 1,750 m (Ibanez-Bernal 1987,
54 Herrera-Basto et al. 1992). Our collections from high-elevation communities also commonly
55 included another mosquito species: Aedes (Ochlerotatus) epactius Dyar & Knab (Lozano-
56 Fuentes et al. 2012b). The females of Ae. epactius reportedly are aggressive blood feeders
57 (O’Meara and Craig 1970, Farajollahi and Price 2013), and a laboratory strain of this mosquito
58 was shown to be capable of transmitting Jamestown Canyon virus (Heard et al. 1991).
59 Because our previous study, with fieldwork conducted from July – August 2011, included
60 as many as 12 communities along an elevation/climate gradient from Veracruz City (sea level) to
61 Puebla City (above 2,100 m), it was restricted to a single sampling occasion per community.
62 Herein, we report on a follow-up study with repeated sampling, conducted from June – October
63 2012, that focused on a subset of four communities, of the 12 examined in 2011, located at
64 elevations ranging from 1,200 – 2,100 m. These communities represented areas where the
65 abundance of Ae. aegypti in the previous year ranged from moderate (Rio Blanco; ~1,270 m;
66 estimated proportion of premises with Ae. aegypti present in 2011 of 0.62) to low (Acultzingo;
3
67 ~1,670 m; 0.26) and very low (Maltrata; ~1,720 m; 0.06; and Puebla City; ~2,150 m; 0.05)
68 (Lozano-Fuentes et al. 2012a).
69 The primary aim of the present study was to determine temporal changes in the
70 abundance of Ae. aegypti adults in high-elevation communities, through bi-weekly trapping of
71 mosquitoes using the BG-Sentinel trap, during the wet and warm part of the year (June to late
72 September / early October). Secondary aims included: (1) corroborate, through a combination of
73 trapping of adults and surveys for pupae, the presence of Ae. aegypti in two high-elevation
74 communities (Maltrata and Puebla City) where it was collected and identified conclusively to
75 species in very low numbers (7 and 3 specimens, respectively) in the previous year (Lozano-
76 Fuentes et al. 2012a); and (2) determine if Ae. epactius adults can be collected with the BG-
77 Sentinel trap and, if so, examine temporal changes in the abundance of this species along the
78 elevation gradient from Rio Blanco to Puebla City.
79
80 Materials and Methods
81 Study environment. Studies were conducted in four communities in Veracruz and
82 Puebla States (Figure 1). Population size, elevation, and basic meteorological characteristics of
83 the study communities are given in Table 1 and Figure 2. The community of Rio Blanco, at
84 ~1,270 m, was included as a lower elevation reference for comparison with the high elevation
85 communities located near the elevational range margin for Ae. aegypti (Acultzingo, Maltrata, and
86 Puebla City; 1,670-2,150 m).
87 To facilitate comparison among communities, the study focused on neighborhoods
88 dominated by middle-income homes with small to medium-sized yards. Neighborhoods
89 dominated by the following premises types were excluded from the study: high income premises 4
90 and low income “fraccionamiento” style premises, which typically are small homes clustered
91 closely together and with very small yards. Based on a survey of the characteristics of the
92 premises that were included in the study (data not shown), the typical home was a one-story
93 house constructed from concrete, brick, or cinder blocks, and with a roof made of concrete. The
94 vast majority of the study homes (> 95%) had piped water (albeit with varying water outage
95 schedules) and regular trash removal services but lacked air conditioning. The average number
96 of rooms per house was 4.8 and the average lot size was ~400 m2. Shrubs and trees were
97 common in the yards, and potential water-holding containers were frequently observed outdoors
98 on the premises (averages of 56 potentially water-holding containers and 8.2 actual water-
99 holding containers per premises, with no distinction made between containers filled by rain
100 versus human action was made per premises).
101 Imagery available through Google Earth (Google, Mountain View, CA), typically <3 yr
102 old, was used to select four clusters within each of the four communities. A cluster was defined
103 as an area of ~1 km2 including blocks (groups of houses surrounded by streets or roads)
104 considered suitable for inclusion in the study. When possible, with the exception of the small
105 communities of Acultzingo and Maltrata, clusters were separated by a distance of at least 1 km,
106 which exceeds the typical flight range (<100 m) of Ae. aegypti (Harrington et al. 2005). Adult
107 trapping and pupal surveys were both performed within the selected clusters, but pupal surveys
108 were not done on the specific premises where adults were trapped.
109 Trapping of adults. Trapping of adult mosquitoes was done with battery-operated BG-
110 Sentinel (BGS) traps equipped with the BG-Lure (a combination of lactic acid, ammonia, and
111 fatty acids, especially caproic acid; Biogents AG, Regensburg, Germany). The battery runs a fan
112 located inside the trap that circulates air between the trap and the environment. The air coming 5
113 out of the trap disperses the lure chemicals while the air coming in forces attracted insects into a
114 collection net. The rationale for using this particular trap is that it has emerged as a standard for
115 collection of adult Ae. aegypti, our target species (Krockel et al. 2006, Maciel-de-Freitas et al.
116 2006, Williams et al. 2006, Barrera et al. 2011, Johnson et al. 2012, Salazar et al. 2012). Each of
117 the four examined communities held 10 fixed trap locations on 10 different residential premises
118 (1 trap per premises). No single cluster within a community had more than four fixed trap
119 locations. Specific trap locations were under roof cover to prevent rain from damaging the traps
120 or the mosquito catches, and in the backyard to minimize the risk of traps or batteries being
121 stolen or vandalized. To maximize catches of Ae. aegypti adults as they enter or exit homes,
122 traps were placed in proximity to windows or doors.
123 Trapping was conducted every two weeks from early June to late September / early
124 October 2012, for a total of nine sampling occasions for each of the 40 fixed trap locations
125 (Table 1). This time period typically includes the warmest and wettest months in the targeted
126 communities (e.g., Fernandez-Eguiarte et al. 2014). For each sampling occasion, the traps were
127 operated over a 48-hr period, with the battery and mosquito catch bag replaced after the first 24
128 hr. The total number of trap-days during the study was 720 (9 sampling occasions x 40 trap
129 locations x 2 days of trapping per sampling occasion).
130 Mosquitoes recovered from the traps were stored dry in tubes together with a desiccant
131 (t.h.e. Desiccant 100% Indicating; EMD Chemicals, Waltham, MA) prior to identification. The
132 adults were identified, using the key of Darsie and Ward (2005), to the following taxonomic
133 entities: (1) Ae. aegypti, (2) Ae. epactius, or (3) a grouping consisting of any other mosquito
134 species (hereinafter referred to as other mosquito species).
6
135 Pupal surveys. We focused solely on pupae in the surveys because the pupal stage has
136 lower mortality than the larval stage and pupal abundance therefore is a better proxy for the
137 abundance of emerging adults compared to abundance of both immature stages combined (Tun-
138 Lin et al. 1996, Focks and Chadee 1997, Knox et al. 2010). The minimum target number of
139 premises to examine per community and sampling occasion was 50; these fell within 3 – 4
140 different clusters per community. A larger number of premises, >100, were examined per
141 sampling occasion in Puebla City because of the very low abundance of Ae. aegypti in that
142 community in our previous survey for immatures (Lozano-Fuentes et al. 2012a). No more than 5
143 premises were examined within a single block. Survey teams started at the northeastern corner
144 of a block and then proceeded in a clockwise direction, sampling every household for which
145 someone was present to permit entry. Only homes within a targeted block that presented obvious
146 safety concerns for the survey teams were excluded.
147 Pupal surveys were conducted on two occasions in each community, between early July
148 and mid-August (Table 1). As shown in Table 1, pupal surveys proceeded from the lowest
149 elevation community (Rio Blanco) to the highest one (Puebla City) within each of the two
150 sampling rounds; this was designed to minimize the potential confounding effect of increasing
151 mosquito numbers over time on the comparison of presence and abundance of Ae. aegypti among
152 communities. Out of the total of 550 premises visits to conduct pupal surveys, only 10
153 individual premises were surveyed during both sampling occasions.
154 Water-holding containers located outdoors on the study premises were examined for
155 presence of mosquito pupae. Container types were classified following the scheme described
156 previously by Lozano-Fuentes et al. (2012a, b). The following container types were excluded
157 from the examination based on safety concerns or difficulty of access: plastic roof water tanks, 7
158 rain gutters, and septic tanks. All mosquito pupae were collected from small (≤5 liter) to
159 medium-sized (5-200 liters) containers, whereas we followed the methodology described by
160 Romero-Vivas et al. (2007) for sampling of larger ones (≥ 200 liters), such as barrels/drums or
161 cement tanks, with a sweep net mounted on a pole. This methodology also estimates the total
162 number of pupae in a large container based on those collected with a single sweep of the net and
163 a multiplication factor determined by the container water capacity (less or more than 1,000 liters)
164 and the water fill level (1/3 full, 2/3 full or full) (Romero-Vivas et al. 2007).
165 Collected pupae, separated by premises and container type, were transferred to plastic
166 bags with water and placed inside coolers for transport to the laboratory. The pupae were then
167 placed in emergence chambers (Mini Mosquito Breeder; Bioquip, Rancho Dominguez, CA) and
168 allowed to emerge to adults. Adults were stored and identified as described previously for those
169 recovered from BGS traps.
170 Estimation of mosquito abundance based on trapping of adults. We calculated the
171 mean daily abundance of Ae. aegypti and Ae. epactius adults per community based on the BGS
172 trap catches. The mean daily abundance of a given species for a community and sampling
173 occasion is calculated as the total number of adults collected in the 10 BGS traps during the 48 hr
174 collecting period divided by the total number of trap-days (n=20). We also calculated the
175 geometric mean daily abundance, but those data are not shown herein because they resulted in
176 similar bi-weekly abundance patterns as for the arithmetic mean.
177 In addition, we calculated the area under the curve (AUC) for the mean daily abundance
178 plots of Ae. aegypti and Ae. epactius adults over the full sampling period (i.e., from the first to
179 the last sampling date) in a given community. The overall AUC was calculated by first
180 determining and then summing AUCs between subsequent trapping occasions (i.e., from 5 to 19 8
181 June, from 19 June to 3 July, etc.). As an example, Figure 3 shows the AUCs for the 14 August
182 to 11 September period for Ae. aegypti in Acultzingo. Estimates for the days within the
183 sampling periods when we did not conduct mosquito trapping were based on a linear
184 relationship. The AUC combines mosquito abundance and time into an expression that describes
185 the intensity and duration of a mosquito infestation, similar to the concept of insect-days used in
186 economic entomology (Ruppel 1983, Dittrich et al. 1985, Archer and Bynum 1992, Beckendorf
187 et al. 2008, Ohnesorg et al. 2009).
188 The AUC between sampling occasions is calculated from the definitive integration of a
189 linear function (f(x) = mx + b) that describes the relationship between collection days. Using the
190 fundamental theorem of calculus we can say that:
191 And the indefinite integration of , where m and b are constants, is:
192 Thus:
193 where ti is the day-of-year for which the mosquito density estimation was performed.
194 Summation of AUCs is calculated as:
195 where: 9
196
197 Estimation of mosquito abundance based on pupal surveys. Of the 2,500 pupae
198 recovered in the surveys, 1,110 (44%) were successfully reared to the adult stage and identified.
199 To account for the remaining non-identified pupae we proportionally allocated them to Ae.
200 aegypti versus Ae. epactius or other mosquito species based on the identified adults from the
201 same container type and premises. Twenty five premises (with a total of 262 non-identifiable
202 pupae) out of the 550 premises surveyed were excluded from further analyses because
203 collections from these premises failed to produce any identifiable adults.
204 To account for complete sampling of small and medium-sized containers versus partial
205 sampling of very large containers (barrels/drums, water tanks and water cisterns) a multiplication
206 factor was applied. Container types with complete sampling were uniformly assigned a neutral
207 multiplication factor of 1, whereas multiplication factors ranging from 1.9 – 3.5 were used for
208 very large container types with partial sampling based on their water volume and water fill level
209 (Romero-Vivas et al. 2007). The final step in the estimation of abundance of pupae for a given
210 premises was to sum the estimated number of pupae across the encountered container types.
211 Meteorological data. Temperature and relative humidity (RH) data for the study period
212 (May – October 2012) were obtained from HOBO® (Onset Computer Corporation, Bourne,
213 MA) data loggers set up in each examined community. Rainfall data for the locations where the
214 HOBO loggers were placed were obtained from the 0.07° gridded Climate Prediction Center
215 Morphing technique (CMORPH) version 1.0 dataset (Joyce et al. 2004), which uses precipitation
216 estimates derived exclusively from low orbiter satellite microwave observations and features
10
217 transported via spatial propagation information obtained from geostationary satellite IR data.
218 CMORPH provides some of the most reliable estimates for tropical summer rainfall compared to
219 other satellite- and model-based rainfall products (Ebert et al. 2007). Weekly averages for mean
220 temperature and rainfall are shown in Figure 2 for the period from 1 May to 31 October 2012 in
221 the four study communities.
222
223 Results
224 Meteorological data. As expected, the mean temperature for the study communities was
225 negatively associated with their elevation. From 1 May to 31 October 2012, the seasonal mean
226 temperature was 20.3 °C for the lowest elevation community of Rio Blanco (just below 1,300
227 m), 18.6 and 19.1 °C, respectively, for the mid-range elevation communities of Acultzingo and
228 Maltrata (~1,700 m), and 17.8 °C for the highest elevation community of Puebla City (above
229 2,100 m). The 1 May – 31 October 2012 seasonal mean temperature departures from the 1951 –
230 2010 averages (Table 2) were -0.7 °C, -0.3 °C, +1.6 °C, and -0.8 °C for Rio Blanco, Acultzingo,
231 Maltrata, and Puebla City, respectively, indicating slightly cooler than normal conditions in all
232 communities except for Maltrata. The highest mean temperatures in the study communities were
233 recorded during early May and early June (Figure 2). Mean temperatures in a given community
234 then were relatively stable from late June to early September, albeit varying by 0.9–2.4 °C
235 among weeks in a given community, before decreasing in late September and October (Figure 2).
236 Rainfall patterns were variable among the study communities (Figure 2). The total
237 rainfall from 1 May to 31 October 2012 was higher in Rio Blanco and Puebla City (784 mm in
238 both communities) compared to Acultzingo and Maltrata (463 and 501 mm, respectively). The 1
239 May – 31 October 2012 total rainfall departures from the 1951 – 2010 averages (Table 2) were - 11
240 883 mm, -44 mm, -161 mm, and -97 mm for Rio Blanco, Acultzingo, Maltrata, and Puebla City,
241 respectively, indicating drier-than-normal conditions, especially in Rio Blanco. There were
242 distinct peaks in weekly rainfall during July for Puebla City and from late July to early
243 September for Rio Blanco, whereas weekly rainfall from late June to early September was less
244 variable in Acultzingo and Maltrata (Figure 2).
245 Summary of mosquito collections. The overall numbers of Ae. aegypti and Ae. epactius
246 collected and conclusively identified to species during the study are summarized in Table 2.
247 Trapping of adults – based on 180 trap-days per community – produced 245 Ae. aegypti in the
248 lowest elevation community of Rio Blanco, and 28 and 1, respectively, for the mid-range
249 elevation communities of Acultzingo and Maltrata (Table 2). No Ae. aegypti adults were
250 collected by trapping in the highest elevation community of Puebla City. Pupal surveys
251 produced specimens later identified as Ae. aegypti in the adult stage from three of the four study
252 communities, with the greatest numbers from Rio Blanco (n=133), followed by Acultzingo
253 (n=21) and Puebla City (n=1) (Table 2). None of the conclusively identified mosquitoes from
254 Maltrata that originated from pupal surveys were Ae. aegypti. Our estimates for the mean
255 abundance of Ae. aegypti pupae per examined premises – including allocation of non-identified
256 pupae as described in the Materials and Methods – in the study communities (Table 2) further
257 underscore the trend of this species occurring most commonly in the lowest elevation community
258 of Rio Blanco, in lower numbers in the mid-range elevation community of Acultzingo, and in
259 very low numbers in the mid-range elevation community of Maltrata and the highest elevation
260 community of Puebla City. Aedes epactius was collected through both trapping of adults and
261 pupal surveys in all four communities (Table 2).
12
262 Temporal patterns of abundance, and cumulative abundance over the study period,
263 of Ae. aegypti and Ae. epactius based on repeated trapping of adults. The bi-weekly patterns
264 of abundance for adults over the study period in 2012 could be examined only for Rio Blanco
265 and Acultzingo for Ae. aegypti (due to collection of only zero or a single specimen in Maltrata
266 and Puebla City) but for all four communities for Ae. epactius. Aedes aegypti was collected on
267 all (Rio Blanco) or nearly all (Acultzingo) sampling occasions from early June to late September,
268 with the peak recorded abundance occurring in late August in both communities (Table 3, Figure
269 3). However, the bi-weekly abundance pattern differed between the two communities in that
270 abundances near the peak recorded value occurred already in early June for the lower elevation
271 community of Rio Blanco, whereas the abundance increased gradually from June to late August
272 in Acultzingo (Figure 3). In both communities, the abundance of Ae. aegypti declined over the
273 month of September (Figure 3). The bi-weekly patterns of abundance for Ae. aegypti adults, in
274 relation to weekly averages for mean temperatures and rainfall, are shown for Rio Blanco and
275 Acultzingo in Figure 4. Notably, an unusual rainfall event in early May, combined with warm
276 weather in the latter part of that month, was associated with a substantial but short-lived spike in
277 abundance of Ae. aegypti adults in early June in Rio Blanco and a minor spike in abundance in
278 Acultzingo. The abundance then declined in Rio Blanco and did not peak until late August,
279 following substantial rainfall in late July and early August. Rainfall occurring later, in early
280 September, had no similar association with increased abundance of Ae. aegypti adults as their
281 numbers declined and remained very low until late September in both Rio Blanco and
282 Acultzingo.
283 Aedes epactius displayed variable patterns of bi-weekly abundance across the study
284 communities (Table 3, Figure 3). The greatest contrast is seen when comparing the lowest and 13
285 highest elevation communities: Rio Blanco and Puebla City. For example, peak abundance
286 occurred in mid-July in Rio Blanco but not until mid-September in Puebla City (Figure 3). One
287 of the mid-range elevation communities – Acultzingo – displayed an intermediate pattern with
288 peak abundance recorded in early August, whereas the temporal abundance pattern in the other
289 mid-range elevation community – Maltrata – was more similar to that for Puebla City (Figure 3).
290 By the last sampling occasion in late September / early October, the abundance of Ae. epactius
291 had decreased to low levels in all four study communities (Figure 3).
292 To provide a quantitative estimate for total adults over the full study period, which is
293 more informative for human risk of encountering adult mosquitoes compared to the peak value
294 recorded during the study period, we calculated the area under the curve (AUC) for Ae. aegypti
295 and Ae. epactius adults in the study communities when this was possible for a given mosquito
296 species (Table 3). Calculations based on mean daily abundances per trap-day, calculated from
297 20 trap-days per sampling occasion, resulted in an AUC estimate of 75.8 for Ae. aegypti for the
298 full sampling period in Rio Blanco (20 x 75.8 = 1,516 total estimated adults over the full
299 sampling period based on our level of trapping effort) compared to only 8.0 for Acultzingo (160
300 total estimated adults) (Table 3). This represented a 9.4-fold difference between the estimates
301 for the two communities (75.8/8.0) and differentials between the two communities of 67.8 for the
302 AUC estimate (75.8-8.0) and 1,356 for total estimated adults (1,516-160). For Ae. epactius, the
303 corresponding AUC estimates for the full sampling period were more similar across the four
304 communities, ranging from 6.8 (136 total estimated adults) in Puebla City to 14.2 (284 total
305 estimated adults) in Rio Blanco (Table 3). Figure 3 appears to show a similar pattern, where the
306 highest abundance of Ae. epactius adults can be seen to shift among communities for different
14
307 sampling occasions rather than consistently being higher in one community (as seen in Figure 3
308 for Ae. aegypti adults in Rio Blanco versus Acultzingo).
309
310 Discussion
311 Our most notable findings are: (1) the description of the temporal abundance pattern for
312 Ae. aegypti adults at high elevation (~1,670 m); (2) the corroboration of data from Lozano-
313 Fuentes et al. (2012a) showing that Ae. aegypti occurs in low numbers at elevations up to 2,150
314 m in México; (3) the first recorded collection of Ae. epactius with a trap designed to collect
315 human-biting mosquitoes; and (4) the first description of the temporal abundance pattern for Ae.
316 epactius across an elevation gradient. The main weaknesses of the study were that we were able
317 to examine the temporal abundance patterns only for a single year, and that the sampling did not
318 adequately capture the start of the active season in the lower elevation reference community of
319 Rio Blanco. Use of a greater number of traps in each study community also would be beneficial
320 in future studies targeting locations near the cool range margin where Ae. aegypti is present but
321 occurs only in low numbers.
322 In a previous study (Lozano-Fuentes et al. 2012a), we reported the collection of Ae.
323 aegypti immatures in 2011 in high-elevation communities (1,670–2,150 m) in México based on
324 surveys for larvae and pupae on residential premises. However, many of the specimens were
325 collected in the larval stage and then raised to adults for identification in the laboratory. This
326 transfer of larvae to potentially more favorable temperature conditions for continued
327 development raised the question of whether they could have progressed to the pupal and adult
328 stages in the examined high-elevation communities. The present study combined trapping of Ae.
329 aegypti adults with surveys of pupae, which are far more likely to emerge as adults compared to 15
330 larvae (Knox et al. 2010) and can complete the emergence to adults in water temperatures as low
331 as 13–16 °C (Bar-Zeev 1958). These collection efforts confirmed the presence of Ae. aegypti for
332 a second consecutive year in the high-elevation communities of Acultzingo (~1,670 m;
333 collection of 28 Ae. aegypti adults and 21 pupae in 2012), Maltrata (~1,720 m; collection of a
334 single adult), and Puebla City (~2,150 m; collection of a single pupa). The observed abundance
335 patterns in 2012 among these high-elevation communities and the lower elevation, warmer
336 reference community of Rio Blanco (Tables 1–2; Figure 2) was similar to that seen in 2011
337 (Lozano-Fuentes et al. 2012a), with large numbers of Ae. aegypti collected in Rio Blanco,
338 moderate numbers in Acultzingo, and very low numbers in Maltrata and Puebla City.
339 Perhaps the most intriguing finding is the discrepancy between Acultzingo and Maltrata,
340 which are located within 11 km of each other, have similar elevations (~1,670 and ~1,720 m),
341 and experienced similar meteorological conditions with regards to mean temperature and rainfall
342 during the study period in 2012 (Figure 2). Pursuing the question of why Acultzingo appears to
343 consistently support a larger Ae. aegypti population compared to nearby Maltata could be
344 fruitful. Both communities have an abundance of water-holding containers (data not shown) but
345 there could be differences in the likelihood of annual introductions of Ae. aegypti immatures as
346 Acultzingo is an often used rest stop just below a mountain pass used by commercial traffic to
347 and from Puebla City, whereas Maltrata is bypassed by the major road leading from Veracruz to
348 Puebla City. Moreover, there may be some subtle climatic differences of which we are still
349 unaware that negatively impact the bionomics of Ae. aegypti in Maltrata; for example,
350 temperatures during part of the year falling just above a critical biological threshold in
351 Acultzingo and just below the threshold in Maltrata. This line of thinking is supported by the
352 fact that BGS traps located in the lower altitude (1,623–1,630 m) portion of Acultzingo yielded 16
353 more Ae. aegypti adults compared with traps in the higher altitude (1,678–1,700 m) portion of
354 the community (17 total adults collected over 54 trap-days for lower altitude traps versus 11 total
355 adults over 126 trap-days for higher altitude traps. Wilcoxon rank sum test W = 3891; df = 54,
356 126; P = 0.002). The small community of Acultzingo has homogenous housing but is located
357 on a slope and encompasses an elevation range of nearly 100 m. We speculate that this
358 community includes microclimates with temperatures near important biological development or
359 survival thresholds for Ae. aegypti, and that the observed decrease in abundance of Ae. aegypti
360 adults in the higher altitude portion of the community result from the negative effects of
361 meteorological factors.
362 The observed temporal pattern for abundance of Ae. aegypti adults in the high-elevation
363 community of Acultzingo – unimodal with a gradual increase from late spring to a summer peak
364 followed by a decline in late summer and early fall – is commonly seen for mosquitoes near the
365 cool margin of their ranges, and a similar temporal abundance pattern was reported previously
366 for Ae. aegypti from Buenos Aires City in Argentina (Vezzani et al. 2004, de Majo et al. 2013).
367 To provide quantitative estimates for total adults over the full study period, we calculated
368 cumulative mean abundances of Ae. aegypti in the lower elevation reference community of Rio
369 Blanco and the high-elevation community of Acultzingo. We estimate that the cumulative
370 abundance of adults over the ~17 wk study period was 9.4-fold greater in Rio Blanco versus
371 Acultzingo. This type of estimate – similar to the concept of insect-days used in agricultural
372 entomology (Ruppel 1983, Dittrich et al. 1985, Archer and Bynum 1992, Beckendorf et al. 2008,
373 Ohnesorg et al. 2009) – can be useful when comparing different geographical locations over the
374 same time period or the same location in different years, or when exploring potential climate-
17
375 driven change in mosquito populations in a given location, which could impact both the number
376 of days within the year during which adults can be active and the population size.
377 In our previous paper on Ae. epactius (Lozano-Fuentes et al. 2012b), we presented data
378 for immatures collected from water-holding containers and reviewed what little is known about
379 the biology of this mosquito. The present study adds new knowledge about the seasonality of the
380 adult stage. The females of Ae. epactius reportedly are aggressive blood feeders (O’Meara and
381 Craig 1970, Farajollahi and Price 2013), and here we present the first evidence that they can be
382 collected from residential premises with the BG-Sentinel equipped with the BG-Lure, designed
383 specifically to collect human-biting mosquitoes. It therefore would be interesting to determine in
384 future studies how commonly Ae. epactius females feed on humans versus other vertebrate
385 animals in the study area.
386 To the best of our knowledge, we present the first description of the temporal abundance
387 pattern for Ae. epactius across an elevation/climatic gradient. Our previous study (Lozano-
388 Fuentes et al. 2012b) showed that Ae. epactius is encountered at elevations ranging from near sea
389 level in Veracruz City on the Gulf of México to above 2,100 m in Puebla City, and that the
390 mosquito is most abundant at elevations from 1,250–1,750 m and then decreases in abundance
391 above 1,800 m. This geographical abundance pattern was related to meteorological conditions,
392 as we found statistically significant parabolic relationships between the percentage of premises in
393 a community with Ae. epactius pupae present (peaking at mid-range elevations) and temperature-
394 related factors (Lozano-Fuentes et al. 2012b). Here we present additional information showing a
395 linkage between the timing of peak recorded abundance of adults and elevation-dependent
396 meteorological conditions: the peak occurred by mid-July in the warmest study community (Rio
397 Blanco) but not until mid-September in the coolest one (Puebla City) (Figures 2–3). Similar to 18
398 the curious geographical pattern for abundance of Ae. aegypti, for which Maltrata is more similar
399 to the distant Puebla City than the proximate Acultzingo, the peak recorded abundance of Ae.
400 epactius adults occurred later in Maltrata and Puebla City (mid-September) than in Acultzingo
401 (early August). This finding strengthens our speculation that some yet-to-be-determined aspects
402 of the local climatic conditions in Acultzingo versus Maltrata may have important impacts on
403 mosquito biology.
404
405 Acknowledgments
406 We thank Eric Hubron, Elena Rustrian, Selene Tejeda, Marco Aurelio Morales, Selene
407 Janitzio Perez and Jesus Yair Zamora of Universidad Veracruzana for field and laboratory
408 assistance. We also thank the Ministries of Public Health, Education and Civil Protection of
409 Veracruz Government for their support. Finally, we are grateful to the involved home owners
410 for granting us access to collect mosquitoes from their properties. This study was funded by a
411 grant from the National Science Foundation to the University Corporation for Atmospheric
412 Research (GEO-1010204).
413
414 References Cited
415 Archer, T. L., and E. O. Bynum, Jr. 1992. Economic injury level for the Russian wheat aphid
416 (Homoptera: Aphididae) on dryland winter wheat. J. Econ. Entomol. 85: 987–992.
417 Barrera, R., M. Amador, and A. J. MacKay. 2011. Population dynamics of Aedes aegypti and
418 dengue as influenced by weather and human behavior in San Juan, Puerto Rico. PLoS Negl.
419 Trop. Dis. 5: e1378.
420 Bar-Zeev, M. 1958. The effect of temperature on the growth rate and survival of the immature 19
421 stages of Aëdes aegypti (L.). Bull. Entomol. Res. 49: 157-163.
422 Beckendorf, E. A., M. A. Catangui, and W. E. Riedell. 2008. Soybean aphid feeding injury
423 and soybean yield, yield components, and seed composition. Agron. J. 100: 237–246.
424 Darsie, Jr., R. F., and R. A. Ward. 2005. Identification and Geographical Distribution of the
425 Mosquitoes of North America, north of Mexico. University Press of Florida, Gainesville, FL.
426 De Majo, M. S., S. Fischer, M. Otero, and N. Schweigmann. 2013. Effects of thermal
427 heterogeneity and egg mortality on differences in the population dynamics of Aedes aegypti
428 (Diptera: Culicidae) over short distances in temperate Argentina. J. Med. Entomol. 50: 543-
429 551.
430 Dittrich, V., S. O. Hassan, and G. H. Ernst. 1985. Sudanese cotton and the whitefly: a
431 casestudy of the emergence of a new primary pest. Crop Protection 4: 161–176.
432 Ebert, E. E., J. E. Janowiak, and C. Kidd. 2007. Comparison of near-real-time precipitation
433 estimates from satellite observations and numerical models. Bull. Amer. Meteor. Soc. 88: 47-
434 64.
435 Farajollahi, A., and D. C. Price. 2013. A rapid identification guide for larvae of the most
436 common North American container-inhabiting Aedes species of medical importance. J. Am.
437 Mosq. Contr. Assoc. 29: 203-221.
438 Fernandez-Eguiarte, A., J. Zavala-Hidalgo, and R. Romero-Centeno. 2014. Atlas Climático
439 Digital de México. Centro de Ciencias de la Atmósfera. Universidad Nacional Autónoma de
440 México. http://uniatmos.atmosfera.unam.mx/.
441 Focks, D. A., and D. D. Chadee. 1997. Pupal survey: an epidemiologically significant
442 surveillance method for Aedes aegypti: an example using data from Trinidad. Am. J. Trop.
443 Med. Hyg. 56: 159-167. 20
444 Gratz, N. G. 1999. Emerging and resurging vector-borne diseases. Annu. Rev. Entomol. 44: 51-
445 75.
446 Gubler, D. J. 2002. The global emergence/resurgence of arboviral diseases as public health
447 problems. Arch. Med. Res. 33: 330-342.
448 Harrington, L. C., T. W. Scott, K. Lerdthusnee, R. C. Coleman, A. Costero, G. G. Clark, J.
449 J. Jones, S. Kitthawee, P. Kittayapong, R. Sithiprasasna, and J. D. Edman. 2005.
450 Dispersal of the dengue vector Aedes aegypti within and between rural communities. Am. J.
451 Trop. Med. Hyg. 72: 209-220.
452 Heard, P. B., M. Zhang, and P. R. Grimstad. 1991. Laboratory transmission of Jamestown
453 Canyon virus and snowshoe hare virus (Bunyaviridae: California serogroup) by several
454 species of mosquitoes. J. Am. Mosq. Contr. Assoc. 7: 94-102.
455 Herrera-Basto, E., D. R. Prevots, M. L. Zarate, J. L. Silva, and J. Sepulveda-Amor. 1992.
456 First reported outbreak of classical dengue fever at 1,700 meters above sea level in Guerrero
457 State, Mexico, June 1988. Am. J. Trop. Med. Hyg. 46: 649-653.
458 Ibanez-Bernal, S. 1987. Nuevo registro altitudinal de Aedes (Stegomyia) aegypti (Linnaeus,
459 1762) (Diptera: Culicidae) en Mexico. Folia Entomol. Mex. 72: 163-164.
460 Johnson, P. H., V. Spitzauer, and S. A. Ritchie. 2012. Field sampling rate of BG-Sentinel traps
461 for Aedes aegypti (Diptera: Culicidae) in suburban Cairns, Australia. J. Med. Entomol. 49:
462 29-34.
463 Joyce, R. J., J. E. Janowiak, P. A. Arkin, and P. Xie. 2004. CMORPH: A method that
464 produces global precipitation estimates from passive microwave and infrared data at high
465 spatial and temporal resolution. J. Hydromet. 5: 487-503.
21
466 Knox, T. B., Y. T. Nguyen, N. S. Vu, B. H. Kay, and P. A. Ryan. 2010. Quantitative
467 relationships between immature and emergent adult Aedes aegypti (Diptera: Culicidae)
468 populations in water storage container habitats. J. Med. Entomol. 47: 748-758.
469 Krockel, U., A. Rose, A. E. Eiras, and M. Geier. 2006. New tools for surveillance of adult
470 yellow fever mosquitoes: comparison of trap catches with human landing rates in urban
471 environment. J. Am. Mosq. Contr. Assoc. 22: 229-238.
472 Lozano-Fuentes, S., M. H. Hayden, C. Welsh-Rodriguez, C. Ochoa-Martinez, B. Tapia-
473 Santos, K. C. Kobylinski, C. K. Uejio, E. Zielinski-Gutierrez, L. Delle Monache, A. J.
474 Monaghan, D. F. Steinhoff, and L. Eisen. 2012a. The dengue virus mosquito vector Aedes
475 aegypti at high elevation in México. Am. J. Trop. Med. Hyg. 87: 902-909.
476 Lozano-Fuentes, S., C. Welsh-Rodriguez, M. H. Hayden, B. Tapia-Santos, C. Ochoa-
477 Martinez, K. C. Kobylinski, C. K. Uejio, E. Zielinski-Gutierrez, L. Delle Monache, A. J.
478 Monaghan, D. F. Steinhoff, and L. Eisen. 2012b. Aedes (Ochlerotatus) epactius along an
479 elevation and climate gradient in Veracruz and Puebla States, México. J. Med. Entomol. 49:
480 1244-1253.
481 Maciel-de-Freitas, R., E. A. Eiras, and R. Lourenço-de-Oliveira. 2006. Field evaluation of
482 effectiveness of the BG-Sentinel, a new trap for capturing adult Aedes aegypti (Diptera:
483 Culicidae). Mem. Inst. Oswaldo Cruz 101: 321-325.
484 O'Meara, G. F., and G. B. Craig, Jr. 1970. Geographical variation in Aedes atropalpus
485 (Diptera: Culicidae). Ann. Entomol. Soc. Am. 63: 1392-1400.
486 Ohnesorg, W. J., K. D. Johnson, and M. E. O’Neal. 2009. Impact of reduced-risk insecticides
487 on soybean aphid and associated natural enemies. J. Econ. Entomol. 102: 1816–1826.
488 Romero-Vivas, C. M. E., Llinas H., and A. K. I. Falconar. 2007. Three calibration factors, 22
489 applied to a rapid sweeping method, can accurately estimate Aedes aegypti (Diptera:
490 Culicidae) pupal numbers in large water-storage containers at all temperatures at which
491 dengue virus transmission occurs. J. Med. Entomol. 44: 930-937.
492 Ruppel, R. 1983. Cumulative insect-days as an index of crop protection. J. Econ. Entomol. 76:
493 375–377.
494 Salazar, F. V., N. L. Achee, J. P. Grieco, A. Prabaripai, L. Eisen, W. Suwonkerd, and T.
495 Chareonviriyaphap. 2012. Evaluation of a peridomestic mosquito trap for integration into
496 an Aedes aegypti (Diptera: Culicidae) push-pull control strategy. J. Vector Ecol. 37: 8-19.
497 Tun-Lin, W., B. H. Kay, A. Barnes, and S. Forsyth. 1996. Critical examination of Aedes
498 aegypti indices: correlations with abundance. Am. J. Trop. Med. Hyg. 54: 543-547.
499 Vezzani, D., S. M. Velázquez, and N. Schweigmann. 2004. Seasonal pattern of abundance of
500 Aedes aegypti (Diptera: Culicidae) in Buenos Aires City, Argentina. Mem. Inst. Oswaldo
501 Cruz 99: 351-356.
502 Weaver, S. C., and W. K. Reisen. 2010. Present and future arboviral threats. Antiviral Res. 85:
503 328-345.
504 Williams, C. R., S. A. Long, R. C. Russell, and S. A. Ritchie. 2006. Field efficacy of the BG-
505 Sentinel compared with CDC backpack aspirators and CO2 baited EVS traps for collection of
506 adult Aedes aegypti in Cairns, Queensland, Australia. J. Am. Mosq. Contr. Assoc. 22: 296-
507 300.
23
508 Table 1. Characteristics of study communities in Veracruz and Puebla States, México, and
509 overview of mosquito sampling efforts from June–October 2012.
Community Rio Blanco Acultzingo Maltrata Puebla City Characteristics of study communities Mean elevation (m)a 1,275 1,677 1,719 2,150 Mean annual temperature (C)b 19.7 17.6 16.7 17.2 Mean max. / min. July temp. (C)b 25.6 / 15.6 24.8 / 12.4 22.5 / 11.1 25.3 / 11.6 Mean max. / min. Jan. temp. (C)b 23.1 / 10.2 20.5 / 6.9 21.5 / 6.9 23.0 / 4.9 Mean annual rainfall (mm)b 1,903 579 783 961 Mean May–Oct. temp. (C) / 21.0 / 1,667 25.3 / 507 17.5 / 662 18.6 / 881 rainfall (mm)b Population estimatec 40,000 7,040 11,840 1,434,000 Trapping of adults First / last trapping date 5 June / 26 Sept. 5 June / 26 Sept. 12 June / 3 Oct. 12 June / 3 Oct. No. fixed trap locations 10 10 10 10 No. sampling occasions 9 9 9 9 Total no. trap-days 180 180 180 180 Pupal surveys – 1st occasion Survey dates 9-10 July 16-17 July 10-11 July 23-24 July No. examined premises 56 48 50 129 Pupal surveys – 2nd occasion Survey dates 30 July 31 July - 1 Aug. 30-31 July 13-15 Aug. No. examined premises 52 50 54 111 510 aFor the premises included in pupal surveys or trapping of adults; bBased on 1951-2010 climate 511 normals obtained from Mexico’s Servicio Meteorológico Nacional; cBased on data for 2010 512 obtained from Mexico’s Instituto Nacional de Estadística y Geografía;
24
513 Table 2. Summary of collections of Ae. aegypti and Ae. epactius from June–October 2012. Total no. mosquitoes collected and Estimated mean abundance of pupae per examined conclusively identified to species premises (standard error of the mean) Community Ae. aegypti Ae. epactius Ae. aegypti Ae. epactius (elevation in Adult Pupal Adult Pupal 1st surveyd 2nd surveyd 1st surveyd 2nd surveyd meters) trapsa surveysb,c trapsa surveysb,c Rio Blanco (1,275) 245 133 44 115 2.2 (0.9) 2.7 (1.6) 3.1 (1.5) 1.5 (0.7) Acultzingo (1,677) 28 21 25 260 0.7 (0.5) 0.0 10.7 (3.4) 1.8 (0.9) Maltrata (1,719) 1 0 40 64 0.0 0.0 0.5 (0.2) 4.1 (3.0) Puebla City (2,150) 0 1 9 58 0.008 (0.008) 0.0 0.9 (0.3) 0.4 (0.2) 514 aBased on 180 trap-days per community; bOnly 44% of collected pupae emerged as adults and could be conclusively identified to taxa; 515 cThe total number of premises examined ranged from 98 in Acultzingo to 104 in Maltrata, 108 in Rio Blanco, and 240 in Puebla City; 516 dCommunity-specific survey dates are given in Table 1.
25
517 Table 3. Aspects of the temporal occurrence of Ae. aegypti and Ae. epactius adults in the study communities based on trapping of 518 adults from early June to late September / early October 2012. Full extent Sampling occasion with b AUC (Total estimated Community of 2012 Sampling occasion with peak recorded Sampling occasion with c adults) (elevation in sampling first collection abundance last collection meters) period Ae. aegypti Ae. epactius Ae. aegypti Ae. epactius Ae. aegypti Ae. epactius Ae. aegypti Ae. epactius Rio Blanco 5 June to 26 75.8 14.2 5-6 June 5-6 June 28-29 Aug. 17-18 July 25-26 Sept. 25-26 Sept. (1,275) Sept. (1,516) (284) Acultzingo 5 June to 26 31 July-1 8.0 8.0 5-6 June 5-6 June 28-29 Aug. 25-26 Sept. 25-26 Sept. (1,677) Sept. Aug. (160) (160) Maltrata 12 June to 3 12.8 NDa 12-13 June NDa 18-19 Sept. NDa 2-3 Oct. NDa (1,719) Oct. (256) Puebla City 12 June to 3 6.8 NDa 26-27 June NDa 18-19 Sept. NDa 2-3 Oct. NDa (2,150) Oct. (136) 519 aNot determined, due to no or a single specimen trapped over the full sampling period; bArea under the curve for the full sampling 520 period based on mean daily abundances per trap. See Materials and Methods for a description of how this was calculated. cBased on 521 overall effort with 20 trap-days per sampling occasion.
26
522 Figure legends
523
524 Figure 1. Locations of study communities in Veracruz State and Puebla State, México.
525
526 Figure 2. Weekly averages for selected meteorological variables – mean temperature (oC; left)
527 and mean rainfall (mm/day; right) – for May – October 2012 in the four study communities.
528
529 Figure 3. Temporal abundance pattern for Ae. aegypti (left) and Ae. epactius (right) adults from
530 early June to late September / early October 2012 in the study communities. The shading
531 illustrates the area under the curve (AUC) for the mean daily abundance for Ae. aegypti in
532 Acultzingo for the 14 August to 11 September period.
533
534 Figure 4. Temporal abundance pattern for Ae. aegypti adults from early June to late
535 September 2012 in Rio Blanco (left) and Acultzingo (right), in relation to weekly averages of
536 mean temperature (oC) and mean rainfall (mm/day). The break interval for the secondary Y-
537 axis is set from 12 to 16
538
27
Figure 1.
28
Figure 2.
29
Figure 3.
30
Figure 4.
31