Winter Refuge for Aedes aegypti and Ae. albopictus Mosquitoes in Hanoi during Winter

Dengue occurs throughout the year in Hanoi, Vietnam, despite winter low temperatures <10°C. During July 2010 to March 2012, we surveyed monthly for Aedes larvae and pupae in 120 houses in 8 Hanoi districts. Aedes albopictus preferred discarded containers in summer and pupal density drastically decreased in winter. Aedes aegypti preferred concrete tanks and this preference increased in winter. Even in winter, the lowest water temperature found in concrete tanks was >14°C, exceeding the developmental zero point of Ae. aegypti. Although jars, drums and concrete tanks were the dominant containers previously (1994–97) in Hanoi, currently the percentage of residences with concrete tanks was still high while jars and drums were quite low. Our study showed that concrete tanks with broken lids allowing mosquitoes access were important winter refuge for Ae. aegypti. We also indicate a concern about concrete tanks serving as foci for Ae. aegypti to expand their distribution in cooler regions.


Introduction
Dengue fever and dengue haemorrhagic fever are very important viral diseases, with an estimated 50 million infections every year and around 3.6 billion people living in areas at risk [1]. Dengue virus is transmitted by Aedes aegypti (L.) and in Southeast Asia to a lesser extent also by Aedes albopictus (Skuse) (Diptera: Culicidae) [2]. In Hanoi northern Vietnam, there were two major outbreaks in 1998 and 2009 and dengue incidence has been increasing yearly since 1999 [3]. Transmission peaks in late summer; however, many dengue cases occur in Hanoi every winter [4,5].
Although Hanoi has a sub-tropical climate, winters are cool, with minimum temperatures around 10uC [6]. Both Ae. aegypti and Ae. albopictus are found in Hanoi [3]. In Hanoi, these species may experience seasonal constraints in activity due to cold temperature, as the developmental zero point of Ae. aegypti is 13.3uC [7] and that of Ae. albopictus is 12.0uC [8].
Mosquitoes in temperate zones have efficient overwintering mechanisms and hibernation in the egg stage is practiced by most Aedes species [9]. In short day-length conditions, Ae. albopictus collected from East Asia (Shanghai, Beijing, Japan and Korea) enters into diapause [10]. Though Ae. albopictus from Southeast Asia (Taipei, Hong Kong, Thailand, and Malaysia) does not enter into diapause in short day conditions [10], little is known about overwintering of Ae. albopictus in Vietnam.
Within the urban environment in many developing countries, densely packed housing and inadequate drinking-water supplies, poor garbage collection services, and surface-water drainage systems combine to create favorable habitats for the proliferation of insect vectors and reservoirs of communicable diseases [11]. In Hanoi, urbanization has been proceeding rapidly and infrastructure has been developing. During a 1994-1997 survey in Hanoi, the major habitat for Aedes immatures were ceramic jars, drums, and concrete tanks [12]. However, in the 15 years since this survey, the situation may have changed, and urbanization may change Aedes habitat and behavior.
Overall in southeast China, the winter minimum and maximum temperatures have been increasing as has diurnal temperature range. Where rapid urbanization has occurred, land-use changes from urbanization have created urban heat islands [13]. Vietnam has also experienced dramatic economic growth since the Doi Moi Policy started in 1986, also contributing to rapid urbanization in Hanoi. In Hanoi, warming via urbanization may result in greater mosquito survival during winter. However, little is known about winter activity of dengue vectors, as there is little transmission then. Our purpose was to investigate the seasonal activity of Ae. aegypti and Ae.albopictus in Hanoi, especially in winter.

Study Area
The study area included eight districts (Hoan Kiem, Dong Da, Hai Ba Trung, Hoang Mai, Thanh Xuan, Thanh Tri, Tu Liem, and Ha Dong) in Hanoi City, Vietnam. These districts had many dengue cases in 2008 and 2009 [4,5]. We selected 15 houses randomly in each district, a total of 120 houses monthly.

Entomological Survey
All indoor and outdoor containers were surveyed to confirm the presence of Aedes immatures. Pupae were brought to the laboratory and reared to adults for species identification. Buckets, drums, jars, concrete tanks, toilet concrete tanks, and wells were sampled by sweeping them five times with a small net. Bonsai, trees grown in containers, plant saucers, and aquarium without fish were included in others. The number of pupae was estimated from mean interval percentage recovery [14], and small containers were emptied to collect pupae. All water-holding containers, both indoors and outdoors, were inspected only after obtaining the necessary permission from the occupants.

Temperature Measurement
Our preliminary survey showed that residents seldom use ceramic jars and drums for storing water and concrete tanks are preferred. January is the coldest month in Hanoi [6]. We measured water temperatures in seven concrete tanks and one metal tank by data loggers ( We also obtained the outside air temperature data measured in Dong Da district from January 2010 to October 2012 [15].

Statistical Analysis
The lowest temperature and January mean temperature data in 2011 and 2012 were subjected to analysis of variance (ANOVA) to compare the values of different medium and location. When the Ftest was significant (p,0.05), the treatment means were compared using Tukey's significantly different test. To examine which type of container was used by mosquitoes in summer and winter, habitat preference was analyzed by G-test. Similar types of containers were grouped if expected frequencies were ,3. Multiple correspondence analysis was conducted to analyze the relationship between species (Ae. aegypti and Ae. albopictus), season (summer and winter) and container type. Mutiple correspondence analysis is a graphic technique that shows each category as a point in a type of scatter plot. The positions of the category-points on the graph mean similarity or association between categories. All data were analyzed using R (version 2.6.2) software.

Results
The lowest Hanoi outside air temperature in January was 9.0uC in 2011 and 8.0uC in 2012 [15]. The mean outside air temperatures were lower than room temperatures, metal tank, and the concrete tank temperatures ( Table 1). The mean concrete tank water temperature was higher than other temperatures in 2011 and 2012. The difference was not significant between the mean of outside air temperature and metal tank water temperature. However, outside air temperature and room temperature were significantly different. The mean outside air temperature was lowest in January 2011 and 2012, though it was not significantly different from means of the room temperature and metal tank water temperature (Table 2). Concrete tanks were warmer in January 2011 and 2012 compared to outside and room temperatures.
Peaks of Ae. aegypti pupation were seen in October and December, 2010, and in May and October, 2011 (Fig. 2). Pupation of Ae. albopictus peaked only twice October 2010 and May 2011. Compared to Ae. albopictus, the density of Ae. aegypti was higher during winter.
Concrete tanks, flower vases, and discarded containers were abundant both in summer and winter (Table 3). Discarded Although most Ae. albopictus were collected from discarded containers, the frequency of preferred containers in summer was not concentrated on a particular container, as was the case for concrete tanks in Ae. aegypti. However, in the winter, most Ae. albopictus were collected from concrete tanks (41%) with 38% in other containers. There was a significant difference in habitat preference between summer and winter for Ae. albopictus (G = 124.37, d.f. = 5, P,0.001).
Multiple correspondence analysis enabled 68.99% of total inertia to be explained by two dimensions, with the first accounting for 38.09% and the second 30.90% (Fig. 3). The analysis showed that Ae. albopictus was closely related to discarded containers. Meanwhile, concrete tanks were related to both Ae. aegypti and winter.

Discussion
Cold temperatures may act as a strong ecological constraint over possible range expansions of Ae. aegypti and Ae. albopictus [16]. Certain populations of Ae. albopictus may extend their summer range considerably north of the 10uC cold month isotherms, but they are not able to survive during cold winter months, which prevents the establishment of permanent populations [9]. Climatic variability and extreme temperatures will have more important effects on reproductive success and population dynamics than winter warming [17].
It may be presumed that Ae. aegypti and Ae. albopictus use different strategies to survive in winter in Hanoi. Aedes aegypti does not enter into diapause [18]. Our results show that Ae. aegypti continued to develop throughout the winter in containers that kept water warm, as the development zero point of Ae. aegypti is 12-13uC [7,19,20].    However, it is hard for Ae. aegypti adults to fly and search for a food in winter, because they cannot fly below 17uC [18]. Winter can impose a severe physiological stress on insects, and overwintering sites that are above the ground are generally colder and experience more variable temperature regimes than those underground [21]. Winter severity varies both temporally and spatially and many insects are adapted to take advantage of the milder conditions that can occur [21]. In our study, the average temperature of a room in January was below 17uC in Hanoi, with outside temperatures much lower. As a consequence, the time and space for Ae. aegypti adults to fly and feed might be limited in winter. Aedes albopictus originating from Taipei, Hong Kong, and Thailand do not enter into diapause [10], which implies that Ae. albopictus from Hanoi may also be a non-diapause race. However, there were fewer Ae. albopictus in winter in Hanoi, compared with Ae. aegypti. Eggs of Ae. albopictus collected from Hanoi delayed hatching under short day length conditions [Tsunoda et al., unpublished data]. These results suggest that most Ae. albopictus enter into diapause and others continue to grow, an apparent bedhedging strategy [22].
Migration may be an alternative strategy to escape from severe winters. In the Shanghai area of China, diapausing adults of Culex tritaeniorhynchus are blown southward for at least 35 km, possibly as far as 200 km, in autumn [23]. The maximum flight distance of Ae. aegypti is usually 100-200 m [24][25][26][27], although, in some particular situations, this species can fly distances .400-600 m [28,29]. The maximum flight distance of Ae. albopictus is longer than that of Ae. aegypti, but at most 100-1,000 m [30,31]. Therefore, both Ae. aegypti and Ae. albopictus might stay near their natal habitat in winter.
Our results show that Ae. albopictus preferred discarded containers mostly in summer but not in winter. It is likely that water temperature in a bottle and other discarded outside containers fluctuates with outside air temperature. It seems to be difficult for Ae. albopictus to grow in discarded containers in winter, because the average outside temperature in January is close to its developmental zero point, 11-12uC [8,19].
Although Ae. aegypti is highly endophilic, Ae. albopictus is exophilic in Thailand [32], which suggests for habitat segregation between the species. However, as room air was warmer than outside air in winter, it will be advantageous for Ae. albopictus adults to fly into and take a blood meal in a house. We should point out that the house-entering behavior of Ae. albopictus may be influenced by the presence of Ae. aegypti.
Over the last 15 years, the distribution of containers for Aedes mosquitoes has changed in Hanoi. The entomological survey conducted in 1994-1997 showed that concrete tanks (38.9% in total containers), jars (30.2%), and drums (26.0%) were abundant [12]. However, in our entomological survey, the relative percentages of jars (3.3-3.6%) and drums (0.6-1.0%) were low, though concrete tanks were still high (20.8-24.4%). The reason why residents do not currently use jars and drums is attributed to the increasing popularity of tap water over the last 15 years. Residents usually keep tap water in an underground or above-ground concrete tank. The water is automatically pumped up to a rooftop tank, since tap water pressure is very weak. As a result, people retain existing tanks when they remodel a home or build a new one if a new home is constructed. Since some concrete tanks may be sealed with putty or covered by plastic sheeting and are difficult to detect, our estimate of the number of tanks may be low.
Our study showed that concrete tanks are important habitat for overwinter-survival of Ae. aegypti in Hanoi. Here, as well as potentially in other countries, if such domestic water systems are encouraged, these concrete tanks could serve as foci for range expansion of Ae. aegypti in cooler regions.