Vertical stratification of adult mosquitoes (Diptera: Culicidae) within a tropical rainforest in Sabah, Malaysia

Studying the anthropophily of simian malaria vectors in the canopy of tropical forests is essential because the hosts are primarily arboreal. In order to develop, sustain or adapt a good control programme, it is important to monitor mosquito populations as well as their hosts and host-seeking preference, distribution and behaviour. Although previous studies in Southeast Asia have used monkey-baited traps at different canopy heights [12, 40, 41], this is the first study to attempt human landing catches, using this method, in the canopy. This study found that there was a higher abundance and human landing rate of mosquitoes at ground level, where people tend to be, than in the canopy where the simian hosts reside. This trend was driven by An. balabacensis, a key malaria vector in Sabah, and highlights the potential importance of this species in transmitting Plasmodium species from simian to human hosts.

Anopheles balabacensis was the most abundant mosquito in all sampled areas, accounting for 70 % of all collected species. Anopheles balabacensis is considered the most important vector of human malaria parasites on Banggi Island and mainland Sabah, Malaysia [18, 19, 30, 42]. In Sabah, An. balabacensis was found to be mainly exophagic, but could also be endophagic and exophilic [30, 43]. These behaviours cause An. balabacensis to be an effective vector of P. knowlesi from human to primate hosts. There were also two distinct subpopulations, one more zoophilic and one more anthropophilic [44–46]. Anopheles balabacensis occurs in forested areas, and readily bites human and monkey hosts, making it an ideal vector of simian malaria [20, 47].

Currently insecticide-treated bed nets and indoor residual spraying are the two main control methods in Malaysia [48]. This study showed An. balabacensis bites as early as 18:00 h in the Tawau Division. Other studies have shown the species bites as early as 18:00 h in recent years in comparison to late night biters in previous decades [44, 49–51]. Given that An. balabacensis is early evening biting, highly anthropophilic, exophagic and exophilic, current control methods are not sufficient to break the transmission cycle of P. knowlesi [52]. In Vietnam and Cambodia, long-lasting insecticidal hammocks (LLIH) were shown to reduce malaria incidence and prevalence in forested areas, and may prove to be an additional effective tool in reduction of malaria in Malaysia [52–54]. The use of repellents have been used for malaria control, but need to be tested in forest and plantations areas.

This study found a different community composition of mosquitoes in the canopy to that at ground level. Different mosquito species have particular flight distributions, with certain species flying and feeding close to ground, some species showing a preference for higher canopy heights, while others show a random distribution [55, 56]. The percentage biting at different canopy heights can be affected by microclimate conditions, such as relative humidity, temperature, wind speed and rainfall [57, 58], but may also change according to time of day [59].

Moonlight appeared to have a significant impact on mosquito activity, with human landing rates increasing on bright nights. Although some studies have shown moonlight increases relative abundance of biting vectors [60–64], others have shown a decrease [65–68] or no effect at all [69]. Collection bias was reduced in this study by collecting in each area under different phases of the moon.

This study also showed how forest disturbance affected mosquito abundance, species richness and human landing rates. Vector abundance was greater in the lightly modified virgin jungle reserve and heavily modified logged forest than in the unmodified primary forest. These results may be explained by the availability of larval breeding sites. Wheel tracks in logged areas due to logging activities can provide breeding sites for a range of mosquito species, whereas wheel tracks are not present within primary forests or virgin jungle reserves [31]. Species richness, estimated by the Chao1 index and ACE, differed across forest disturbance and height, with logged forest and ground level having a higher species richness than primary forest and canopy.