Aquatic habitats of the malaria vector Anopheles funestus in rural south-eastern Tanzania

Anopheles funestus has been a major malaria vector in many east and southern African countries for several years [1‐4]. In south-eastern Tanzania, they have been implicated in more than 85% of malaria transmission events across several villages [4‐6]. Its dominance in pathogen transmission [4, 7] is attributable to factors such as: (a) being predominantly anthropophilic (i.e. strong preference for blood from humans over other vertebrates) and endophilic (i.e. strong preference for biting and resting indoors than outdoors) [8, 9], (b) their resistance to some of the commonly-used pyrethroid insecticides in locations such as south-eastern Tanzania [10‐14], and (c) their superior daily survival probabilities as reflected in the higher parity rates compared to other Anopheles species [4, 5, 7].

The supremacy of An. funestus in malaria transmission has been observed even in areas where they occur at far lower densities compared to other malaria vectors, such as Anopheles arabiensis [4, 5, 15]. In such settings, the infrequent occurrence partly explains why their behaviours are relatively understudied in the field. More generally, An. funestus is also far easier to find as adults than as larvae. As a result, this species rarely features in larval surveys of Anopheles species. Researchers, therefore, sometimes rely on adult collections rather than larval collections to obtain enough samples for insecticide resistance testing [4], which according to the World Health Organization (WHO) protocols require F1 offspring with synchronized age groups [16].

It has previously been suggested that an in-depth ecological understanding, followed by improved targeting of An. funestus could potentially improve their control, and significantly reduce malaria transmission in areas where the vector dominates [4]. Given the strong resistance of some An. funestus populations to insecticides commonly applied on insecticide-treated nets (ITNs) and/or indoor residual spraying (IRS) [10‐14], supplementary measures targeting the aquatic stages of the mosquitoes are critical for more effective control of An. funestus. This requires rigorous surveys to identify and characterize preferred larval habitats for An. funestus [17]. Strategies such as targeted larviciding—a component of larval source management could indeed significantly improve control efforts and accelerate progress towards malaria elimination, especially in communities where the aquatic habitats are fixed, few and findable [18, 19].

A previous study in western Kenya reported that An. funestus prefers to oviposit in large semi-permanent water bodies containing aquatic vegetation and algae [20]. A separate study in coastal Kenya observed these species breeding in vegetated aquatic habitats that are stable and permanent, and were along river streams [21]. In Cameroon, it was demonstrated that An. funestus habitats were often found in open savannas instead of deep or degraded forests [22, 23]. These habitats had greater exposure to sunlight and high temperatures, and remained productive for longer, often with peaks after the start of the dry season. Unfortunately, in south-eastern Tanzania where the species now dominates transmission, there have not been detailed studies of its natural aquatic habitats and responses to interventions. This situation is complicated by difficulties in colonizing the species inside laboratories, which would enable such studies.

Although An. funestus are among the most important vectors of malaria in Africa, little is known regarding their larval ecology and development. This crucial gap needs an urgent solution, but is perpetuated by the inability of most mosquito biologists to create laboratory colonies of this vector species. Understanding the basic environmental parameters that influence mosquitoes breeding and oviposition can improve the planning, development and deployment of new interventions to control malaria transmission [27]. This study identified and characterized larval habitats of An. funestus in south-eastern Tanzanian villages of Ulanga and Kilombero districts, where this mosquito species has been implicated in most malaria-infective bites [4, 6].

The study examined more than 100 potential habitats across six villages and identified three main habitat types. First were small water wells with well-defined edges and were spring-fed, some of which were also used by locals as domestic water sources (Fig. 4b). These habitats were often occupied by multiple species of the An. funestus group, and in some cases, they were shaded by large trees. The second type of habitat was medium-sized ponds, for which the central part retained water for all or most of the year. These habitats often had surface vegetation (Fig. 4a) and were occupied by multiple other Anopheles species, such as An. arabiensis. Third was the riverside habitats consisting of the slow-moving waters on the rivers or river tributaries, also with vegetation (Fig. 4c). These habitats were mostly found at altitudes above 400 m above sea level, unlike the other two habitats which were more common at lower altitudes below 300 m (Fig. 5). In summary, An. funestus in this area appears to prefer permanent and semi-permanent aquatic habitats with stagnant or slow-moving waters, emergent vegetation e.g. algae on swamp surfaces, clear waters at depths exceeding 50 cm and nearness to human dwellings.

This study provides a basis for designing future surveys and control operations targeting malaria, especially in places such as south-eastern Tanzania where An. funestus and An. arabiensis play a major role in malaria transmission [5, 6, 28‐30]. This study has suggested that permanent or semi-permanent habitats characterized by emergent vegetation play a major role in the ecology of An. funestus. The findings are concurrent with past evidence from earlier investigations in Kenya [20, 21, 31]. Although this current study did not assess the seasonality of An. funestus larvae densities in the different habitats, the observed preference of permanent and semi-permanent water bodies explains the known seasonality of its adult densities in the same study villages as observed in recent entomological surveys [4, 30]. The adult densities of An. funestus tend to peak after the rains just before the dry seasons begin, and are sustained by the large permanent water bodies [20]. Although no detailed studies have been done in this area targeting An. funestus aquatic habitats, early accounts by Gillies and DeMeillon [9], as well as limited surveys done nearly 50 years ago in the Ifakara area (which neighbours the current study site) already suggested an association between the late peaks in An. funestus densities and the large perennial habitats [32].

Although there was no clear statistical association, the An. funestus habitats had depths greater than 50 cm and were located within 100 m from the human dwellings. This is likely due to the anthropophagic nature of these mosquitoes [33], and further explains the importance of this species in malaria transmission in these areas. Other Anopheles species, such as An. gambiae, which breed in open sunlit stagnant water pools [9, 20, 34] are also highly anthropophagic and generally occur near human habitations [35]. The ability of An. funestus to breed in the river waters is not unique to Tanzania, but has also been demonstrated in other places such as coastal Kenya [9, 21], and may be due to the higher levels of aeration and dissolved oxygen in such waters. Additional investigations are therefore required to further examine these details. A similar ecological niche has been described for Anopheles pseudopunctipennis in South America, which was successfully controlled by clearing the river waters of the algal blooms [36]. While it is unclear whether clearing the identified habitats of emergent vegetation would be suitable for control of An. funestus in Tanzania, it will be important to investigate it as a potential environmentally-friendly approach, in which community members could be engaged to achieve effective disease prevention. Besides, it will be important to ascertain the importance of these habitat types across multiple sites and settings. For instance, in one area in the north of Tanzania, Dida et al. [37] found no mosquito larvae near the main rivers, suggesting the dominant malaria vectors may be breeding elsewhere in such settings.

Understanding the physicochemical characteristics of mosquito larval habitats is also important in understanding their overall ecological needs, and assessing options for manipulation. It is probably that the physicochemical parameter levels observed in this study might have been influenced by agricultural practices and pesticide use, which is widespread in the valley [38]. Emerging adult mosquitoes from these habitats might become more resistant towards the insecticides having the same chemical formula used in mosquito vector control [35, 39, 40].

Habitats most productive of An. funestus were those in higher altitude villages, which were probably less affected by agricultural insecticidal deposits [41, 42], than habitats at the floor of the valley. Nonetheless, the mosquito species from the same study villages are known to be already strongly resistant to insecticides used for public health, including pyrethroids and carbamates [43], a situation potentially related to the widespread use of pesticides in both agriculture and public health. Similar to other studies on Anopheles mosquitoes, the An. funestus habitats in this study area had weak acidity pH [40, 44]. The main habitats had pH ranging from 6.5 to 6.7, turbidity from 26.6 to 54.8 NTU and total dissolved solids from 60.5 to 80.3 mg/L, all of which are similar to most observations of habitats of Anopheles mosquitoes in previous studies [45, 46]. Anopheles funestus mosquitoes were collected from the habitats with different concentrations of nitrate, but it remains unclear whether this might influence larval development as earlier described [40].

One limitation of this study was that some characteristics such as water temperature, though included in this analysis are subject to change during the day. Future studies should consider laboratory investigations and also the use of field data collected multiple times a day to determine the suitable temperature ranges and other physicochemical characteristics for optimal survival of this mosquito species.

Overall, this study has provided a basic description of An. funestus habitats in rural south-eastern Tanzanian districts of Ulanga and Kilombero. There were three main habitat types occupied by An. funestus, namely: (a) small spring-fed pools with well-defined perimeters, (b) medium-sized natural ponds retaining water most of the year, and (c) slow-moving waters along river tributaries particularly important at higher altitudes at the edge of the valley. The habitats generally had clear waters with emergent surface vegetation, depths greater than 0.5 m and distances less than 100 m from human dwellings. They were permanent or semi-permanent, retaining water most of the year. Effective control measures for this species should consider understanding their behaviour and ecology including characteristics of their aquatic habitats so that they can be targeted during their immature stages. Given the rarity of the An. funestus habitats and the observed characteristics, these habitats fit the description of being fixed, few and findable. Future studies should, therefore, investigate the potential of using larviciding or larval source management to improve malaria control in settings where An. funestus dominate.