Dispersion pattern of Mansonia in the surroundings of the Amazon Jirau Hydroelectric Power Plant

Mansonia spp. are voracious hematophagous mosquitoes whose mature stages usually breed in freshwater bodies containing aquatic vegetation. The reduction in water flow leads to a proliferation in aquatic plants, increasing their populations. Besides, some species are potential vectors of pathogens such as arboviruses and microfilariae. We evaluated the degree of active dispersion of females of Mansonia spp. in the surrounding area of the Jirau hydroelectric power plant in the Amazon, Rondônia, Brazil, using mark-release-recapture techniques. The flight behavior of the recaptured specimens was summarized with a set of average and maximum distances traveled. We show that the dispersal movement of Mansonia spp. is predominantly performed by random, low, and short flights, with a tendency to remain near the breeding sites in certain vegetation fragments. However, the maximum distances traveled were 2000 m from the release point for Mansonia amazonensis during 2018 and Mansonia humeralis during 2019.

Study area. The sampling planning considered exploring the site and its surroundings, choosing the most representative capture points and the most efficient equipment, as well as selecting the methods of sample preservation.
The study was carried out in the surrounding area of the Jirau HPP, located in the Amazon Forest biome. The Jirau HPP (9° 16′ 16.8″ S; 64° 38′ 25.9″ W) is located near the city of Porto Velho. The sampling area near the release and recapture points had larval habitat with a high concentration of aquatic vegetation of the following plant species: Eichornia crassipes (Mart.) Solms-Laubach, Eichornia azurea (Sw.) Kunth, Ceratopteris pteridoides (Hook.) Hieron, Pistia sp., Salvinia sp., Paspalum sp. Hymenachne sp., Oxicarium cubensis, Ludwigia helminthorriza, Phyllantus fluitans, Eleocharis sp., and grasses. The respective larval habitats were characterized by a team specializing in macrophytes and another in entomology. Subsequently, they were confirmed by in-situ observation according to the descriptions of Sioli 13 . The region's Köppen climate classification is tropical (type Aw), with a distinguishable dry season in winter from May to October (July is the driest month) and a rainy season in the summer from November to April. The average annual temperature is 25.6 °C, with monthly averages as low as 16 °C in the coldest month and over 34 °C in the warmest ones. Precipitation is over 2000 mm per year 14 .
The capture, release, and recapture points were located in Jaci-Paraná, a district of Porto Velho, the capital of Rondônia. The specimens were captured in fragments of the closed forest (9° 13′ 02.1″ S 64° 30′ 06.6″ W) and released and recaptured in an open field (9° 14′ 50.3″ S 64° 28′ 06.5″ W). An aerial image was used to delineate the mark-release-recapture (MRR) experiment area. This image was georeferenced using a Garmin® GPS unit (Fig. 1).
Specimen capture method. Sampling was carried out in May, July, October, and December 2018, and in April, July, September, and November 2019. The captures were conducted between 6:00 p.m. and 8:00 p.m. on three alternate days using Shannon light traps and electric aspirators used similarly to a manual suction tube to optimize the sampling effort. A 12-V rechargeable battery operated the electric aspirator. The simultaneous use www.nature.com/scientificreports/ of two devices makes them more efficient at capturing mosquitoes in areas with a high density of adults, aspirating inactive specimens in the Shannon trap and active ones in flight. Additionally, we used the Castro catcher to capture samples, which was used for five minutes at intervals of 30 min, totaling 20 min. The purpose was to relate the species captured in the electric aspirator occurring during the sampling since this collector ensures greater preservation of morphological and anatomical characteristics of the specimens, leading consequently to greater reliability in identification. These specimens were confined in cages and not tagged or released and therefore sacrificed and placed in conical tubes until identification. Meanwhile, the specimens captured with the electric aspirator were kept in a polyvinyl chloride tube measuring 100 mm in diameter and 30 cm in length, which was changed every 5 min to keep the specimens alive until they were released. These tubes were closed with a plastic cap at one end and kept in a moist chamber until their release. Finally, all specimens were counted. The sampling effort with Castro catcher and electric aspirator were as follows: Sampling 1-(5.6%; 94.4%), Sampling 2-(7.2%; 92.8%), Sampling 3-(8.4%; 91.6%), Sampling 4-(17.5%; 82.5%), Sampling 5-(3.4%; 96.6%), Sampling 6-(5.6%; 92.8%), Sampling 7-(5.6%; 93.6%), Sampling 8-(5.6%; 94.4%), respectively. Specimen mark and release method. Specimens were marked with fluorescent powder (BioQuip®) about an hour before the time of release and then released approximately two hours after capture on the same night in the study area. The fluorescent powder was sprayed onto the specimens using a small hand pump to mark them.
Specimen recapture method. Recaptures were carried out with CDC and MF 60® light traps. Initially, wooden stakes of ≈ 1 m were driven into the ground so as to suspend the light traps. Then, 28 CDC-type traps and 28 MF60 traps were installed at distances between 30 m and 2000 m to the north, south, east, or west of the release site. The capture and recapture points remained in the same place throughout the sampling season; traps were monitored and changed every 24 h (Fig. 2). The recaptures in the sample area extended over six subsequent days.

Identification. Species of Mansonia were identified by direct observation of morphological characters
under a stereomicroscope (Zeiss®) and an optical microscope with ultraviolet light (Nikon®). We used dichotomous keys 15 , to consult the species descriptions and associated biological characteristics and behaviors. Aiming to ensure the specific identification of Mansonia females, the arrangement pattern of the spines on the last abdominal tergite was compared. All specimens from each sample period had their eighth abdominal segment sectioned. The structures were washed in distilled water and neutral detergent to loosen all scales from the structure. Then, the structures were dehydrated with 70% to 95% alcoholic series in absolute alcohol. The eighth abdominal segment was mounted dorsally with Canada balsam, between slide and coverslip. The specimens were subsequently deposited in the Entomological Collection of the Instituto Oswaldo Cruz, under the title "Amazon Collection, UHE-JIRAU".
The abbreviations of genera and subgenera followed the norms suggested by Reinert 16 , modified according to the proposed nomenclature for mosquitoes.
A set of dispersion measures composed of the mean (MDT) and maximum (MAX) distance traveled were used to summarize the flight behavior of the recaptured culicids 17,18 .

Results
A total of 38,113 females of Mansonia spp. were captured, of which 26,677 (70%) were marked and released, and 11,436 (30%) were preserved for species identification. This allowed for an approximate representation of the species captured, marked, and released in each experiment.
The highest number of recaptures of Mansonia spp. occurred at a distance of 300 m (N = 29) from the release site. The maximum dispersal distance of 2000 m was observed during sampling periods 4, 7, and 8 for all taxa. Sampling period 8, performed in the rainy season, had the highest number of recaptured specimens (N = 57) ( Table 1).
Throughout the study period, specimens tagged with fluorescent powder were found in traps 25 times. Sampling periods 3 and 7 had traps with tagged mosquitoes at five different dispersion distances (Table 1). Tagged mosquitoes were found in traps at distances of 300, 500, 1000, and 2000 m in three sampling periods (

Discussion
The movement of Mansonia species observed in the surrounding area of the Jirau HPP was consistent with dispersion patterns observed by a previous study 19 . The authors analyzed the dispersal behavior of three mosquito species in two habitats. They found that 70% Ivanova-Kazas 20 reported that mosquitoes move from their breeding sites when searching for a blood-feeding source and then return to their breeding sites for oviposition. Displacement movements from breeding sites to human dwellings were also observed for Anopheles maculipennis Meigen, 1818. These movements seem to be influenced by the topography and prevailing winds that carry attractive odors. Mansonia perturbans (Walker) also shows upward and downward displacements from the forest canopy at dusk and dawn 21 .
The vegetation type that makes up the background landscape of the study area, with the formation of forest mosaics in the Amazon, favors the outward movement of adults. The occurrence of forests in the vicinity of the experimental area inhibits the movement of specimens in this direction. In addition, the high availability of food sources beyond the boundary of the experimental area changes the return rates of individuals. Thus, the specimens of Mansonia spp. may disperse more than 2 km from the adult's point of emergence, with a slight tendency to remain in the vegetation patch under stable conditions, i.e., in preserved environments with nearby food sources.
Gorayeb and Ribeiro 22 studied the Tabanidae fauna from the eastern Amazon to define its species displacement autonomy. They found that the disparity in the number of species captured was influenced by seasonality. In addition, they point to intrinsic mechanisms of tabanid species in terms of finding and attacking hosts, which www.nature.com/scientificreports/ are clearly different over small and large distances. They also emphasize that wind is unlikely to play a role in local tabanid displacement behavior in searching for horses as blood-feeding sources.
Few studies have explored the flight intervals in Mansonia species. Wharton 23 reported that Ma. bonneae Edwards, 1930 and Ma. dives (Schiner, 1868) can disperse the considerable distances of 1.6 to 3.2 km in the forest environment when searching for hosts as blood-feeding sources. In this sense, the depletion of food sources can change the population dynamics of Mansonia spp. by widening the dispersion range of the taxa through the search for food in other ecological niches.
The number of recaptured specimens that we recorded is consistent with other similar studies on mosquitoes [24][25][26][27] . The results from our eight recapture reference samples therefore within the expected ranges for our experimental design.   19 . This was also confirmed by Bailey and Gould 28 using CDC light traps, demonstrating that the marked individuals did not fly more than 375 m from the release point.
The dispersal behavior of immature and adult specimens is an intrinsic characteristic of all species. Taxa disperse spontaneously, stimulated by several factors, and the movements occur due to the natural ability of subsistence related to the physiological needs of each species 29 .
We conclude that the dispersal movement of Mansonia spp. is predominantly performed by random, low (≅ 1 m high), and short flights maintaining a home range within a radius of approximately 30 to 100 m from the adults' point of emergence, showing a tendency to remain near the breeding sites in certain fragments of vegetation. Even so, our observations confirm that specimens of Mansonia spp. are capable of covering distances greater than 2 km from the adults' emergence point, despite their general tendency to remain in the same vegetation patch under stable conditions of food sources and environmental preservation.