Oviposition ecology and species composition of Aedes spp. and Aedes aegypti dynamics in variously urbanized settings in arbovirus foci in southeastern Côte d’Ivoire

Background Aedes mosquito-transmitted outbreaks of dengue and yellow fever have been reported from rural and urban parts of Côte d’Ivoire. The present study aimed at assessing Aedes spp. oviposition ecology in variously urbanized settings within arbovirus foci in southeastern Côte d’Ivoire. Methods Aedes spp. eggs were sampled using a standard ovitrap method from January 2013 to April 2014 in different ecosystems of rural, suburban and urban areas. Emerged larvae were reared until the adult stage for species identification. Results Aedes spp. oviposition ecology significantly varied from rural-to-urban areas and according to the ecozones and the seasons. Species richness of Aedes spp. gradually decreased from rural (eight species) to suburban (three species) and urban (one species) areas. Conversely, emerged adult Aedes spp. mean numbers were higher in the urban (1.97 Aedes/ovitrap/week), followed by the suburban (1.44 Aedes/ovitrap/week) and rural (0.89 Aedes/ovitrap/week) areas. Aedes aegypti was the only species in the urban setting (100 %), and was also the predominant species in suburban (85.5 %) and rural (63.3 %) areas. The highest Ae. aegypti mean number was observed in the urban (1.97 Ae. aegypti/ovitrap/week), followed by the suburban (1.20 Ae. aegypti/ovitrap/week) and rural (0.57 Ae. aegypti/ovitrap/week) areas. Aedes africanus (9.4 %), Ae. dendrophilus (8.0 %), Ae. metallicus (1.3 %) in the rural, and Ae. vittatus (6.5 %) and Ae. metallicus (1.2 %) in the suburban areas each represented more than 1 % of the total Aedes fauna. In all areas, Aedes species richness and abundance were higher in the peridomestic zones and during the rainy season, with stronger variations in species richness in the rural and in abundance in the urban areas. Besides, the highest Culex quinquefasciatus abundance was found in the urban areas, while Eretmapodites chrysogaster was restricted to the rural areas. Conclusions Urbanization correlates with a substantially higher abundance in Aedes mosquitoes and a regression of the Aedes wild species towards a unique presence of Ae. aegypti in urban areas. Aedes wild species serve as bridge vectors of arboviruses in rural areas, while Ae. aegypti amplifies arbovirus transmission in urban areas. Our results have important ramifications for dengue and yellow fever vector control and surveillance strategies in arbovirus foci in southeastern Côte d’Ivoire. Electronic supplementary material The online version of this article (doi:10.1186/s13071-016-1778-9) contains supplementary material, which is available to authorized users.


Background
Several Aedes mosquito species are involved in the transmission of arboviral diseases, including dengue and yellow fever, responsible for major health burdens worldwide [1][2][3]. In the mid-1990s, yellow fever was controlled in Francophone Africa by vaccination while both yellow fever and dengue were eliminated in the Americas by effective control of Aedes aegypti [4]. However, in recent years, there has been a dramatic resurgence of dengue fever worldwide [5][6][7][8] and re-emergence of yellow fever in some parts of Africa [9]. In addition, other arboviruses vectored by Aedes mosquitoes, such as chikungunya [10], Rift valley fever [11] and Zika virus [12] are emerging or re-emerging in Africa, particularly in West Africa [13][14][15]. The patterns of arboviral disease transmission and its geographic expansion are likely a result of intensive urbanization [1,2,6,16]. However, dengue and yellow fever originated in enzootic (sylvatic) cycles associated with wild Aedes vectors in rural areas. Enzootic cycles are linked to urban transmission cycles by a major domestic vector, Ae. aegypti [17].
In Côte d'Ivoire, single and dual epidemics of dengue and yellow fever involving several wild Aedes species and the major urban vector, Ae. aegypti, have been reported in both rural and urban areas [18]. Sylvatic dengue virus circulation, without human infections, was documented by isolation of DENV-2 serotype from wild Aedes vectors, including Aedes africanus, Aedes furcifer, Aedes luteocephalus, Aedes opok and Aedes vittatus in rural areas in the 1980s [19]. Since then, several sporadic outbreaks of dengue DENV-1, DENV-2 and DENV-3 serotypes and yellow fever sometimes resulting in fatal outcomes have occurred [20][21][22]. The outbreaks were mostly concentrated in surrounding villages and periurban areas of Abidjan, the economic capital and the most densely populated city of Côte d'Ivoire [21,22].
The comprehension of the shaping patterns of immature Aedes mosquito ecology along the urbanization gradient is of paramount importance in determining their role in maintenance of epidemic arboviral diseases transmission [1,2]. Knowledge of such patterns may therefore enable a more effective deployment of vector control measures for the benefit of public health. Aedes mosquitoes are readily adapted to a broad range of ecological settings and are expected to vary according to natural and urbanized environment [2,9]. Certain Aedes mosquito species are confined and limited to sylvatic, rural or urban areas, whereas others have a large distribution and colonize almost every environment, such as the wild, rural and urbanized settings, the domestic and peridomestic premises, the types of landscapes and the microhabitats [2,9]. Those species of Aedes occurring in transitional environments may serve as bridge vectors between enzootic diseases and humans in rural areas.
Moreover, Aedes mosquitoes are the main reservoirs of arboviruses and the longest link of the transmission chain since they host the viruses during longer duration compared to humans and wild animals [23]. These Aedes vector species show both oral and transovarial infection [23,24]. The extent to which eggs are resistant against desiccation varies between species and strains, and depends on climatic conditions [25,26]. Otherwise, Aedes mosquito species can be associated with other mosquito species for different interaction purposes such as predation, competition and symbiosis [9]. Eretmapodites chrysogaster is a predaceous mosquito and lays its eggs in Aedes species breeding sites [27]. Aedes and Culex species, mainly Culex quinquefasciatus and Ae. aegypti, are sympatric and co-occur in the same containers [28].
Aedes aegypti is an urban species and a major vector of dengue and yellow fever by amplifying epidemics among the urban populations [9,17]. This species consists of two subspecies, Ae. aegypti aegypti and Ae. aegypti formosus that are morphologically [29], behaviourally and genetically distinct [30][31][32]. However, there are ambiguities resulting in confusion over morphological distinction between the two subspecies of Ae. aegypti in West Africa [13,33].
Urbanization could potentially modify Aedes mosquito ecology by changing the composition and dynamics of species, and increasing the abundance of their breeding sites due to environmental changes, and thus contribute to arbovirus outbreaks [2]. However, Aedes mosquito egg laying ecology is unknown in arbovirus foci located in variously urbanized settings of southeastern Côte d'Ivoire. To fill this gap, our study explored Aedes mosquito egg laying patterns, species composition and dynamics in Ehania-V1, Blockhauss and Treichville representing rural, suburban and urban settings of southeastern Côte d'Ivoire, respectively. Because immature mosquitoes are sensitive to environmental changes [2,25,26], we hypothesized that Aedes mosquito oviposition ecology and species composition, and the dynamics of Ae. aegypti change from rural to suburban and urban settings. Field surveys of Aedes mosquito egg were performed using a highly sensitive sampling method, namely the standardized World Health Organization (WHO) ovitraps [23,24], larval rearing in the laboratory and adult stage identification were conducted to test our hypothesis. The findings provide valuable information on Aedes mosquito egg laying patterns, species composition and Ae. aegypti dynamics in different urbanized ecosystems.
The key results open new perspectives for improving current vector control and surveillance strategies for dengue and yellow fever that are tailored for specific settings of southeastern Côte d'Ivoire.

Study area
The study was conducted in three settings in southeastern Côte d'Ivoire: Ehania-V1, Blockhauss and Treichville, representing rural, suburban and urban areas, respectively (Fig. 1). The village of Ehania-V1 (5°18′N, 3°04′W) belongs to the district of Aboisso some 140 km east of Abidjan. Ehania-V1 is a rural area with a population density of approximately 48 people/km 2 and unpaved roads. The residencies are composed of traditional and ordinarily modern houses. This area is surrounded by industrial oil palm plantation (Elaesis guineensis) of 11,444 ha and 100 ha of preserved primary rainforest. The rainforest provides strong vegetation with dense canopy cover, trees with holes and bamboos and hosts non-human primates and birds.
Blockhauss (5°19′N, 4°00′W) is located within Abidjan bordered in its northern part by Banco National Park with over 3750 ha of rainforest. This setting is a suburban area with approximately 750 people/km 2 and paved roads. The land use comprises a mixture of residential buildings, hospitals and schools. The residencies are ordinarily modern houses and some blocks with flats. Urbanization is underway in untapped spaces.
Treichville (5°18′N, 4°00′W) is situated in central Abidjan and separated from Blockhauss by the Ebrié Lagoon that has a width of approximately 4 km. This setting is an urban area with more than 1800 people/km 2 and paved roads. The density of the population greatly increases during the daytime due to the convergence of people from other municipalities of Abidjan for trading, businesses and sports. The land use is essentially residential, commercial, cultural and sportive buildings, seaport, and public services such as schools and hospitals, filled with green spaces set apart. The residencies are mostly composed of blocks of flats and some ordinarily modern houses. Urbanization is completed due to the lack of availability of additional space for the construction of new houses.
In southeastern Côte d'Ivoire, the climate is warm and humid and overall, transitional from equatorial to tropical with four seasons. The seasons are more clearly distinguished by rainfall than temperature. The two rainy seasons are separated by a dry season. The main rainy season extends from May to July, while a short rainy season occurs in October and November. The main dry season extends from December to April. This classic sequence of seasonality has been disrupted due to strong climate variability [34]. The average annual precipitation ranges from 1200 to 2400 mm. The annual temperature is around 26.5°C and the annual relative humidity (RH) ranges between 78 and 90 %.

Design of sample collection
Aedes spp. eggs were sampled using the standard WHO ovitrap method [23,24]. Ovitraps were metallic boxes cut with 400 cm 3 volume and covered with black paint to attract gravid female Aedes mosquitoes in search of egg laying grounds. They were filled (75 % full) with distilled water mixed with rainwater and 10 % infusions of herbs (Panicum maximum) to increase the attractiveness of the ovitraps [35]. A 5 × 7 × 0.3 cm paddle made of hardboard, rough on one side and serving as oviposition substrate, was plunged in each box and left for a one-week period during each of six surveys. The rural area was divided into three ecozones: domestic, peridomestic and sylvatic. The suburban and urban areas were divided into domestic and peridomestic zones because there were no sylvatic zones. According to Cordellier et al. [23], domestic zone refers to the human-inhabited space, the peridomestic zone covers the vegetated environment surrounding the domestic zone in which humans usually discard artificial items that serve as breeding sites for Aedes mosquitoes, the sylvatic zone is primarily an undisturbed environment free of discarded containers that host forests with natural containers (e.g. tree holes) and wild animals. In our study, the peridomestic zone extended from the edge of the domestic zone to 300 m, while the sylvatic zone was located from 300 to 800 m around the village. During each of the six surveys, 50 ovitraps were repeatedly placed in the same location in each defined ecozone. In the human-inhabited zone, the ovitraps were suspended at 1.5 m above the ground to secure and protect them. In total, 900, 600 and 600 ovitraps were deployed in the rural, suburban and urban areas, respectively, from January 2013 to April 2014. In addition, from April to July 2013, we conducted socioecological investigations in 50 households representative of each study area in which ovitraps were placed to identify their sociogeographic status.

Key socio-geographic characteristics
The socio-ecological investigations showed that the surveyed households varied highly along the increasing urbanization gradient. The mean number (mean ± standard error) of people per household was 5.9 ± 2.8 in the rural, 8.6 ± 2.1 in the suburban and 11.9 ± 3.2 in the urban areas. The potential resident containers were mainly discarded items (cans, pots, barrels, tyres) (n = 50; 58.0 %) and natural containers (coconut, tree hole, bamboo, snail shell) (46. Fig. 1 Map of the areas located in southeastern Côte d'Ivoire where the ecological study of Aedes mosquitoes was carried out: a Ehania-V1 (covers the villages of Ehania-V1 and Akakro and represents the rural area without major and secondary paved roads. The site is close to the primary rainforest reserve). b Blockhauss (comprises the villages of Blockhauss and Petit-Cocody and represents the suburban area with only secondary paved roads. It is about 5 km away from the rainforest of Banco National Park). c Treichville (includes the sections of Jacques-Aka and Biafra and is the urban area with numerous major and secondary paved roads. It is located in the centre of Abidjan and is separated from Blockhauss by Ebrié lagoon) 50) in the rural, 98.0 % in the suburban and 100 % in the urban areas.

Sample laboratory treatment
In the field, the paddles, Culex egg rafts and remaining water from the ovitraps were separately stored in plastic boxes and then transported in a cold box to the laboratory. The paddles were dried during a period of 5 days at a temperature of 25 ± 1°C, RH of 80-90 % and a 12:12 h light:dark photoperiod. They were screened with white, insect-proof, nylon netting to prevent eventual egg laying from other mosquitoes and potential predators. The paddles were then separately immersed in plastic cups (6 × 9 × 15 cm) 75 % filled with distilled water for attached egg hatching. The process was repeated three times after flooding of 5 days to maximize egg hatching. Emerging larvae were counted and recorded.
As there was no reliable larval identification key, the larvae were reared until adult stage under identical conditions as paddle drying. To avoid overcrowding and limit mortality, at most 20 emerging larvae were placed per 200 ml plastic cup filled to 75 % with distilled water. Each plastic cup was labeled with a unique number of the sample, the study area and the date of collection. Larvae were fed each morning (07:00-08:00 h) with Tetra-Min Baby Fish Food®. Emerging pupae were transferred to new plastic cups using plastic pipettes for adult emergence. The cups containing the pupae were netted to avoid draining the newly hatched adults. Culex egg rafts were not dried but were stored at 15 ± 1°C to avoid desiccation [36]. In addition, the larvae hatched from Culex egg rafts and the larvae of Aedes, Culex and Eretmapodites found in the collected water from the ovitraps were also separately reared until adult stage, under the same conditions as described above. During rearing, emerging adult Aedes, Culex and Eretmapodites specimens were identified to the species level using morphological criteria [23,24,37]. The mosquito individuals were stored at subgenus, species and sex levels and data were recorded in an entomology collection database.

Statistical analysis
The measures of Aedes species proportions were expressed as the percentage of specimens among Aedes fauna for each study area and analysed using Fisher's exact test to look at the relationship between the species composition and the study area and ecozone, and followed by the Proportion-test. Fisher's exact test was used because expected numbers of specimens were equal or less than five.
Aedes species richness was assessed as the number of collected species in each study area and compared using a one-way analysis of variance (ANOVA), followed by Bonferroni's correction. The species diversity and dominance of Aedes spp. were estimated by Shannon-Weaver's index [38] and Simpson's index [39] and analysed by Kruskal-Wallis test because the log-transformed data exhibited significant deviations from normality. The abundance of Aedes spp. and Ae. aegypti was expressed as the mean number of specimens per ovitrap and analysed using repeated measures approaches in a generalized linear mixed model (GLMM) framework in order to take into account the possible interactions between the variables "month", "study site" and "ecozone" [35]. To account for overdispersion due to excessive numbers of zeroes, the data were logtransformed [log (number of specimens + 1)] [36]. The logtransformed data were subjected to GLMM procedures and analysed as follows [35]. We compared the mean numbers of Aedes mosquito specimens per ovitrap between the study areas, the ecozones and the months using mixedeffects regression (xtmixed command), performed the joint tests of the interactions and the main effects of the study sites, the ecozones and months (contrast command) to understand the significant interactions, followed up the simple effects of each study area and ecozone over the months by pairwise comparisons (margins and pwcompare commands) and the post-hoc test of the trends (contrast p. operator) and the post-hoc test of the partial interaction (contrast a. operator). The mortality of the larvae during rearing was compared using negative binomial error. The extra sub-site, sylvatic zone, was excluded from the analysis when performing the comparisons between the study areas, and only included when the comparisons were conducted among the ecozones in the rural area. A significance level of 5 % was set for statistical testing. All data were analysed using Stata version 14.0 (Stata Corporation; College Station, TX, USA).

Species composition of emerged adult mosquitoes
The mortality of the larvae hatched from Aedes spp. eggs during the rearing to adult stage was not statistically significant (all P > 0.05) thus making the comparison of emerged adults possible. Table 1 shows the species composition of adult Aedes spp. emerged from eggs collected from the different study areas. Totals of 2441, 2440 and 3098 adult Aedes spp. emerged from the eggs collected in the rural, suburban and urban areas, respectively. Aedes species belonged to three subgenera (Stegomyia, Aedimorphus and Diceromyia) in the rural areas, two subgenera (Stegomiya and Aedimorphus) in the suburban areas and a single subgenus (Stegomyia) in the urban areas. The species richness of Aedes spp. gradually decreased from the rural (eight species) to the suburban (three species) and urban (one species) areas. Fisher's exact test indicated that Aedes species richness significantly varied from one study area to another (all P < 0.001). Proportion-testing indicated that there was a significant difference in Aedes species proportions in the rural (χ 2 = 9411.15, df = 7, P < 0.0001) and the suburban (χ 2 = 5052.86, df = 2, P < 0.0001) areas.
Aedes aegypti was the predominant species with significantly higher proportions among Aedes fauna collected in the rural (Z = 18.91, P < 0.001) and suburban areas (Z = 7.83, P < 0.001), and the sole Aedes species in the urban areas. Aedes africanus and Ae. dendrophilus in the rural areas and Ae. vittatus in the suburban areas were found in significantly higher proportions. Aedes metallicus represented more than 1 % of the total Aedes fauna in the rural and the suburban areas whereas Ae. furcifer, Ae. fraseri and Ae. luteocpehalus were collected in lower proportions in the rural areas. Non-Aedes mosquito species were also sampled in all study areas. Totals of 277, 108 and 67 specimens of Culex spp. were sampled from the rural, suburban and urban areas, respectively. In the rural area, Culex spp. was composed of three species, Cx. nebulosus (n = 277; 49.4 %), followed by Cx. quinquefasciatus (28.2 %) and Cx. poicilipes (22.4 %). The diversity of Culex spp. then decreased to a single species, Cx. quinquefasciatus, in the suburban (n = 108) and urban (n = 133) areas. Eretmapodites spp. was restricted to the rural area and composed of only one species, Er. chrysogaster, with 274 specimens.
Richness, diversity and dominance of Aedes spp. Table 2 presents the species richness, diversity and dominance of Aedes spp. in all of the study areas and different ecozones. Aedes spp. species richness was significantly different among the study areas (F = 18.60, df = 2, P = 0.0001) and ecozones (F = 9.24, df = 6, P < 0.0001), with higher numbers of species in the rural area and the peridomestic zone of the same area. The species diversity of Aedes spp. was statistically different among the study areas (χ 2 = 14.00, df = 2, P = 0.0009) and ecozones (χ 2 = 27.65, df = 6, P = 0.0001), with higher values for both diversity indices in the rural area and the sylvatic zone of the rural area. Moreover, Aedes spp. species dominance was significantly different among the study areas (χ 2 = 13.86, df = 2, P = 0.0011) and ecozones (χ 2 = 28.00, df = 6, P = 0.0001), with higher Simpson's index values in the urban area and both peridomestic and domestic zones of the urban area.

Dynamics of Aedes spp. numbers
The highest mean numbers of emerged adult Aedes spp. were found in the urban setting (1.97 ± 0.10 Aedes/ovitrap/week), followed by the suburban (1.44 ± 0.09 Aedes/ ovitrap/week) and rural (0.89 ± 0.06 Aedes/ovitrap/week) areas. The mean numbers of emerged adult Aedes spp. were significantly different between the rural and urban areas (Z = 5.01, P < 0.001). The effects and the interactions among the study areas, the ecozones and months, and the trends of Aedes spp. numbers over the months were statistically significant (Table 3). Table 4 summarises the geographical variation of adult Aedes species collected in each of the three study areas. In the rural areas, specimens of Ae. africanus, Aedes dendrophilus, Aedes metallicus and Aedes fraseri were collected in the domestic zone, while significant numbers of Ae. aegypti were sampled in the sylvatic zone. Emerged adult Aedes spp. mean numbers were significantly higher in the peridomestic zone with 1.36 ± 0.14 Aedes/ovitrap/week in the rural (Contrast = 0.50, Z = 5.16, P < 0.001), suburban (2.10 ± 0.15 Aedes/ovitrap/week; Contrast = -4.89, Z = -4.81, P < 0.001) and urban (2.80 ± 0.21 Aedes/ovitrap/week; Contrast = -0.49; Z = -4.85, P < 0.001) areas.
Additional file 1: Table S1 indicates the seasonal variation of emerged adult Aedes spp. in all of the different study areas. Aedes metallicus, Aedes usambara, Ae. fraseri, Ae. luteocephalus and Ae. furcifer were not collected in January 2014 and April 2014. However, Ae. aegypti was sampled in all surveys in each study area. In all of the study areas, higher numbers of emerged adult Aedes spp. were found in July 2013 with 1.47 ± 0.18 Aedes/ovitrap/week in the rural, 2.31 ± 0.29 Aedes/ovitrap/week in the suburban and 4.06 ± 0.28 in the urban areas (Fig. 2). Conversely, the significantly respective lowest numbers of Aedes spp. were recorded in January 2014 with 0.47 ± 0.13 (all P < 0.05), 0.43 ± 0.17 (all P < 0.05) and 0.47 ± 0.11 (all P < 0.001) Aedes/ovitrap/week.

Dynamics of Ae. aegypti
A total of 1754 (n = 2421; 63.3 ± 1.2 %) adult Ae. aegypti emerged from the eggs collected from the rural areas, 2159 (n = 2440; 85.5 ± 0.8 %) from the suburban and 3098 (n = 3098; 100 %) from the urban areas (Table 1). The highest mean numbers of Ae. aegypti were found in the urban setting, with 1.97 ± 0.10 Ae. aegypti/ovitrap/ week). Considerably lower mean numbers were recorded in the rural and suburban settings, with 0.57 ± 0.05 and 1.20 ± 0.09 Ae. aegypti/ovitrap/week, respectively. Aedes aegypti mean numbers were significantly different between the urban and rural (Z = 6.23, P < 0.001), and the suburban and rural (Z = 2.15, P < 0.05) areas. The effects of, and interactions among the study areas, ecozones and months, and the trends of Ae. aegypti numbers over the months were statistically significant (Table 3). Figure 3 shows the geographical variations of adult Ae. aegypti mean numbers and frequencies. Significantly higher mean numbers per ovitrap of Ae. aegypti were found in the peridomestic zones with 0.85 ± 0.12 Ae. aegypti/ovitrap/week in the rural (Contrast = 0.48, Z = 5.68, P < 0.001); 1.64 ± 0.14 Ae. aegypti/ovitrap/ week in the suburban (Contrast = -0.36, Z = -3.65, P < 0.001); and 2.80 ± 0.21 Ae. aegypti/ovitrap/week in the urban (Contrast = -0.49, Z = -5.04, P < 0.001) settings.   Aedes aegypti was collected in all of the ecozones of each study area. Its frequencies gradually increased from the sylvatic zone of the rural area (n = 395; 33.4 %) to the domestic zone of the urban area (n = 1160; 100 %) (Z = 31.43, P < 0.001). Figure 4 shows the monthly variations of emerged adult Ae. aegypti mean numbers in relation to the rainfall. Emerged adult Ae. aegypti mean numbers significantly varied as a function of rainfall fluctuation in all study areas. The highest mean numbers were found during the rainy season in July 2013 with 0.96 ± 0.14 Ae. aegypti/ovitrap/week in the rural and 4.06 ± 0.28 Ae. aegypti/ovitrap/week in the urban areas, and in October 2013 with 1.65 ± 0.25 Ae. aegypti/ovitrap/week in the suburban areas. In urban areas, Ae. aegypti mean numbers dramatically declined in January 2014 (0.47 ± 0.11 Ae. aegypti/ovitrap/week) compared to July 2013 (Contrast = -1.25, Z = -7.88, P < 0.001). In the same study area, Ae. aegypti numbers were significantly higher in January 2013 (3.08 ± 0.11 Ae. aegypti/ovitrap/week) compared to January 2014 (Contrast = -1.02, Z = -6.57, P < 0.001).

Discussion
To our knowledge, this is the first study exploring oviposition ecology of Aedes mosquitoes in variously urbanized settings of Côte d'Ivoire. Importantly, several species of Aedes were involved in previous dengue and yellow fever outbreaks in Côte d'Ivoire [18][19][20][21][22]. The outbreaks might be explained by the paucity of vector control strategies. A deeper understanding of the modifications induced by urbanization in the ecology of Aedes mosquitoes is crucial. Our data suggest that Aedes oviposition ecology and species composition, and Ae. aegypti dynamics differ from rural to suburban and urban areas in southeastern Côte d'Ivoire.
Our data highlighted that the mean numbers of emerged adult Aedes spp. increase from rural to urban areas. An increase in Aedes species prevalence and abundance by urbanization was indicated by Li et al. [2]. This phenomenon is probably due to elevated numbers of Aedes breeding sites such as tyres, discarded cans or water storage containers, provided by urbanizing environment [2]. In addition, an urbanized environment accelerates Aedes mosquito development and survivorship [2]. By increasing Aedes spp. abundance, urbanization could potentially aggravate epidemic risk factors for arbovirus.
Our results showed that urbanization alters Aedes mosquito species composition towards dominance of Ae. aegypti in the urban areas, while rural and suburban areas favour other wild Aedes species, including Ae. vittatus, Ae. dendrophilus, Ae. africanus, Ae. luteocephalus, Ae. furcifer, Ae. metallicus, Ae. usambara and Ae. fraseri. Aedes aegypti eggs are expected to be more desiccation-resistant [23,24]; this might raise their ability to survive in a deforested environment such as the urban areas exposed to direct sunlight and thus increase the species geographical invasion. Conversely, the wild Aedes species collected only in the rural and suburban settings probably originated from a natural environment such as the preserved rainforest and the Banco National Park forest, respectively. The disappearance of wild Aedes species in the urban settings might be explained by the destruction of the natural environment for building houses and other infrastructure. The removal of vegetation due to house constructions and other infrastructure developments results in direct exposure of Aedes spp. breeding sites to solar radiation. The wild Aedes species eggs from rural settings could be protected against solar radiation by rainforest canopy [40] since they are laid in tree holes [41] and bamboo internodes [42] filled by rainwater and maintained under low temperature. It is conceivable that wild Aedes species that lay more fragile and desiccation-sensitive eggs remain confined to the rural areas, mainly in the rainy forest [23,24]. Additional field manipulations and experiments under controlled laboratory conditions testing the different Aedes species egg desiccation-resistance levels may be useful to better understand the segregation among the species and the population growth rates. Indeed, the forest-dwelling Aedes species that are still present in the rural areas may play a key role as bridge vectors between the sylvatic cycles of dengue, yellow fever and other viruses among non-human primates  Letters indicate the results of one-way ANOVA test followed by Bonferroni correction (richness) and Kruskal-Wallis test (Shannon diversity index, Simpson dominance index). Groups that do not share the same letter are significantly different (P < 0.05) 1 Ecozone in the rural area 2 Ecozone in the suburban area 3 Ecozone in the urban area and humans [17]. The vector role of these Aedes species is subtle and difficult to trace, and often remains undetected because there are no traditional epidemiological risk indicators such as the house index, container index or Breteau index [43]. However, the exclusive existence of predators such as Er. chrysogaster in the rural areas might influence the abundance of Aedes species [27,42]. Er. chrysogaster is also suspected to transmit arboviruses in tropical Africa [23]. In summary, the segregation induced by urbanization in Aedes species diversity is consistent with the known arbovirus transmission cycles in tropical Africa [17] and merits further consideration for dengue and yellow fever surveillance.
Our results suggested that the geographical and seasonal variations of Aedes spp. are associated with urbanized settings. The preference of Aedes spp. to lay eggs in the peridomestic vicinity confirms previous findings from urban areas in Brazil [44] and Vietnam [45]. Peridomestic premises are in close proximity to human residencies, and hence the principal blood-meal sources of adult Aedes mosquitoes. Furthermore, they also provide ideal ecosystems such as dense vegetation favourable for Aedes spp. refugia [46] and natural breeding sites such as tree holes [39,42,47] and artificial containers as discarded cans and old vehicle tyres [48,49]. Regarding the seasonal variation, Aedes spp. mean numbers were strongly associated with rainfall patterns, history, variability and intensity. The fluctuations in Aedes spp. counts could be influenced by seasonal flooding-drying cycles as reported in Côte d'Ivoire [18] and Brazil [44]. Aedes spp. eggs probably enter into a dormant  stage to withstand desiccation periods during the dry season, while precipitations might flood the breeding sites and increase the abundance of Aedes spp. [50]. However, the sudden decline of Aedes spp. numbers in October 2013 in the urban setting might be due to heavy precipitations and exacerbated flushing of their eggs because of the lack of protective vegetation in the built-up environment [51].
Finally, our findings revealed that Ae. aegypti is the most common species along the increasing urbanization gradient and the unique Aedes species in urban settings, thus suggesting particular attention on its egg laying patterns and population dynamics. Aedes aegypti is an urban species that preferentially feeds on humans [52] and is well adapted to live in close proximity to human habitats [53]. Such highly anthropophilic behaviour may enhance human-to-human transmission of arboviruses and trigger dengue and yellow fever outbreaks. The dominance of Ae. aegypti in still urbanizing and already urbanized areas of Africa is well documented [8,9,18] and is possibly due to its plastic oviposition behaviour allowing the colonization of natural and artificial environments [9,42]. Otherwise, the rising occurrence of Ae. aegypti was also coupled with the increasing presence of another urban, anthropophilic and sympatric species, Cx. quinquefasciatus [54][55][56] and the lack of predators such as Er. chrysogaster [27] in the urban area. In contrast, the specimens of Ae. aegypti unexpectedly collected in the sylvatic zone are, perhaps, members of the Aedes aegypti formosus, the ancestral progenitor of Aedes aegypti aegypti and the only sylvan form known in West Africa [30,31]. Aedes aegypti formosus has no white scales on the first abdominal tergite and a dark or black cuticle. This subspecies is exophilic, preferentially feeds on wild animals and breeds in natural containers such as tree holes [32,57,58], whereas, Ae. aegypti aegypti has scales on the first abdominal tergite and a lightly tanned cuticle and tends to be endo-and anthropophilic and breed in man-made containers [31]. However, contrary to East Africa [29,30,57], the scaling and behavioural patterns do not match with the discrete genetic differences in allozymes and microsatellites for Ae. aegypti collected in West Africa [13,31,33,40,59]. This results in confusion over morphological distinction between the two forms [13]. Due to these ambiguities, we were not able to confirm which Ae. aegypti form was represented among the sylvan specimens collected in the rural area. Above all, the urban and sylvan forms of Ae. aegypti are both competent arbovirus vectors in West Africa [13]. Urbanization continues at a rapid pace in Côte d'Ivoire, particularly in the southeastern part resulting in drastic segregation among Aedes species by favouring Ae. aegypti and restricting wild Aedes species to rural areas. These trends were paralleled by recurrent resurgences of yellow fever and dengue in recent years. However, yellow fever is historically well known as a key factor having forced the transfer of the colonial capital of Côte d'Ivoire from Grand-Bassam to Abidjan in 1899 [60]. Despite this historical and present background, the resurgence of yellow fever and dengue outbreaks is not resolved and their sporadic occurrence creates major public health concerns [60]. Between 2001 and 2007, 1468 suspected, 41 confirmed and 26 fatal cases of yellow fever were reported. During the period of 2007-2001, 111 suspected with 31 confirmed and 43 deadly cases of yellow fever were notified. The incidence of yellow fever gradually increased and peaked in 2011 with 79 cases and 35 deaths. In 2008, nine cases of yellow fever and two cases of dengue DENV-3 were recorded. In 2010, 13 confirmed and two fatal cases of yellow fever, and one deadly case of DENV-1 were reported. The strengthened warning systems and the operated vector control are usually performed in urban areas, mainly in Abidjan. Our study suggests that while vector control should focus on urban areas, rural areas are important as they may serve as transition zones for (re-)introduction of arboviral diseases through sylvatic bridge vectors. Because rural areas host various wild vectors, they act as a potential reservoir and originator of arboviruses from which urban areas are (re-)infected. Therefore, rural areas also need to be considered when elaborating and applying arbovirus vector surveillance and control strategies. Aedes species control strategies could apply the lethal ovitrap [61] and autocidal [62] gravid ovitrap-based on mass trapping method.

Conclusions
In arbovirus foci of the southeastern Côte d'Ivoire, urbanized environment correlates with a substantially higher abundance in Aedes species and a regression of the Aedes wild species towards a unique presence of Ae. aegypti. Aedes aegypti is expected to drive arbovirus transmission in the urban areas, while other species probably serve as potential bridge vectors between sylvatic and urban cycles of human arboviral infections in the rural areas. Our findings provide valuable information on Aedes spp. ecology patterns in variously urbanized settings and therefore suggest that the rural areas also need to be considered when implementing arbovirus vector surveillance and control strategies.

Additional file
Additional file 1: Table S1. Ethics approval and consent to participate The study protocol received approval from the local health and administrative authorities. In addition, all entomological surveys and sample collections carried out on private lands or private residential areas were done with the permission and written informed consent of the owners or residents. These studies did not involve endangered or protected species.
Author details