Insect migration is the mass movement of insects, usually in search of more appropriate conditions. The malaria mosquito, genus Anopheles, is not thought to be a migratory insect: generally, it doesn’t travel (‘disperse’) more than about five km, and long-range movements that have been recorded are thought to be ‘accidental’ and un-important in terms of the movement of disease. Anopheles are weak fliers, and any long-distance dispersal would need to be assisted – i.e., wind-aided dispersion. However, in other weak-flying insects, it has been suggested that wind-aided dispersion is governed by changes in the insect’s pattern of activity, perhaps due to physiological mechanisms set in motion in response to particular environmental cues. Therefore, to an extent at least, under the insect’s control.

Mosquitoes require an aquatic habitat for larval development. It is surprising, therefore, that in areas where surface water can be absent for many months of the year, mosquito populations, and the associated disease, surge back when the rainy season starts.

The Sahel is a semi-arid region stretching east to west across the south-central regions of North Africa. Malaria transmission in the Sahel deceases markedly during the (3 – 8 -month) dry season, in which the mosquito vectors are not present. The reappearance of mosquitoes alongside the rains is currently attributed to mosquito aestivation (particularly An. coluzzii), or long-distance migration from regions with year-round larval habitats (An. gambiae sensu strictu and An. arabiensis). Although the dispersal of mosquitoes (here referred to as ‘migration’) has been documented, a recent study by Huestis and colleagues indicates that migration of malaria mosquitoes may be of greater significance than previously appreciated.

To sample the migrating mosquitoes, four villages in the Sahel of Mali were chosen, from which sticky insect mesh traps hanging from large, helium-filled balloons were launched into the air (Figure 1). Secured to the ground by a giant block of concrete, with the cord attached to a re-purposed garden hose reel for bringing it up and down, insect traps were then attached to the cord of the balloon to sample at three heights – initially 40, 120 and 160; then 90, 120 and 190 m above ground. Two additional stations at 240 and 290 m were added later. The traps were large rectangles (3×1 m) of mesh, painted with insect glue, and the balloons were sent skywards for about 10 consecutive nights per month between 2013 and 2015 (total 617 nights).

Figure 1. Large rectangular sticky insect traps were suspended, at different heights, from helium-filled balloons in four villages in the Sahel of Mali. After each trap night, trap catch was identified morphologically, and by molecular methods (An. gambiae s. l. only).

In total, ten species of Anopheles – malaria mosquitoes – were caught. These included the primary malaria vectors An. coluzzii and An. gambiae s. s., as well as vectors of secondary importance (An. pharoensis, An. coustani, An. squamous and An. rufipes) and three undetermined An. species. A total of 235 anopheline mosquitoes was caught. Although this total is relatively low (total number of the five most-captured anophelines: An. squamosus, 100; An. pharoensis, 40; An. coustani, 30; An. rufipes, 24; An. coluzzii, 23), when modelling the number of mosquitoes predicted to cross a 100-km line linking the sampling sites (perpendicular to the prevailing wind direction), the annual migration figures are impressive: more than 80,000 An. gambiae, 6 million An. coluzzii and 44 million An. squamosus mosquitoes at that transect alone.

Among captured anophelines, the majority were female (4:1), and more than 90 % of these had previously taken a bloodmeal (i.e. were blood-fed, or the bloodmeal-dependent development of eggs could be seen [gravid/semi-gravid]). These findings highlight the significance of the captured mosquitoes in terms of disease transmission – although none were found to harbour the malaria parasite (all anophelines were tested), it is probably a sample size issue. That they were blood-fed indicates they could have been exposed to infection. Assuming a Plasmodium (Pl.) infection rate of 1 % across all anophelines, there was a possibility of seeing only 2.35 infected mosquitoes. Using a Pl. infection rate of 1 % for An. coluzzii, An. gambiae, An. coustani and An. pharoensis, and 0.1 % for the other (known) Anopheles species, the authors calculated that more than 286,000 infected migrant mosquitoes would cross their 100-km line, at altitude, per year.

Figure 2. (A) The relationship of height of trap (y axis) with mosquito catch for the five most-common anopheline species. Mosquito density given in panel density (panel = sticky trap, usually 3 per balloon) and aerial density (panel density divided by the volume of air sampled), with bubble size proportional to mosquito density. (B) Monthly panel density for the five most-common anopheline species (An. squamosus divided by three to preserve the scale), overlaid by the length of migration period (dashed lines). The sampling month for species that were collected only once or twice is shown by letters: c1, An. namibensis; g, An. gambiae; m1, Anopheles Mali species 1; m2, Anopheles Mali species 2; nC, An. sp. nr concolor.

Using the most accurate meteorological data available, the authors then modelled 2-hr and 9-hr flight trajectories of all caught anophelines, revealing the possible origin of the mosquitoes as well as the estimated nightly distance travelled (Figure 3). For 9-hr nightly flights, the maximum distance (‘displacement’) was 257 – 295 km for all anopheline species with sample size of more than 20 insects. Further, the mean distance for the 2/9-hr flights was 30 and 120 km respectively (maxima of 70 and 295 km). Most predictions gave a southwestern origin for the mosquitoes, which relates to the prevailing wind direction at peak migration times (Figure 2B).

Figure 3. For all anophelines caught, backward flight trajectories were modelled using meteorological data. Here 9-hour backwards flight trajectories are shown. Each line represents one of four simulated trajectories of one (or more) mosquitoes intercepted at that location and on that night. The area encompassed by the four trajectories is shown by shadow, and migration season by line colour. The Anopheles species is indicated above each panel.

Given the large predicted total annual migrations of Anopheles, and therefore movement of infected vectors, these intriguing findings bear considerable significance for disease control. As we push for malaria elimination in some regions, possible re-invasion by Pl.-infected mosquito vectors may present a bigger threat than previously understood. It may be that figuring out – and then controlling – possible sources of immigrating mosquitoes, by modelling meteorological data and not-too-distant vector populations, could considerably aid in the drive to elimination.