Sand fly (Diptera, Psychodidae, Phlebotominae) abundance and diversity in areas affected by the São Francisco River transposition project in Ceará State, Brazil

Entomological surveillance of sand fly vectors was carried out to support leishmaniasis prevention and control measures in areas affected by the São Francisco River Transposition Project. Sand flies were collected monthly between May 2011 and December 2014 in seven municipalities: Missão Velha, Brejo Santo, Lavras da Mangabeira, Iguatu, Mauriti, Jaguaribe and Jaguaretama, in dwellings, peridomicile and forest areas for three consecutive days, for a period of 12 h each day (18:00 to 06:00 h). Differences in species composition between sites were tested with Shannon’s diversity index, the similarity between habitats was estimated by the Sørensen’s qualitative similarity index and, for the most abundant species in each municipality, a standardized index of species abundance was applied. The influence of climatic factors on sand fly population densities was analyzed using Spearman’s correlation coefficients. A total of 214,213 sand fly specimens belonging to 18 species were captured. The most abundant species in all municipalities was Lutzomyia longipalpis (Lutz & Neiva, 1912). The municipalities of Mauriti and Missão Velha stand out in terms of high species richness, with the latter exhibiting the greatest diversity. The number of sand flies in the Iguatu, Jaguaribe and Jaguaretama municipality was higher during the rainy months, whereas the populations declined in the drier months; the sand fly population density in other municipalities (Missão Velha, Brejo Santo, Lavras de Mangabeira and Mauriti) showed negative correlation with rainfall. This study confirms the presence of several Leishmania spp. vectors in the seven municipalities affected by the São Francisco River Transposition Project, with Lu. longipalpis being the most abundant species at all study sites. Vector populations in these municipalities should be monitored, ultimately to assess the associations between environmental changes and sand fly population dynamics and leishmaniasis transmission risk.


Background
Leishmaniasis poses significant public health problems in Brazil. Both American cutaneous leishmaniasis (ACL) and American visceral leishmaniasis (AVL) are expanding and new epidemiological scenarios are emerging. AVL, in particular, has become urbanized, becoming endemic in medium and large cities in different regions of the country [1,2]. Approximately 90% of human AVL cases registered in South America occur in Brazil [2,3]. ACL has been recorded in every state in Brazil, and is currently in a territorial expansion phase and changing its epidemiological profile [1]. Studies suggest that the geographic expansion of leishmaniasis stems from anthropic activities, by altering the environmental conditions that favor sand fly populations and increase their contact with humans [3,4].
In northeastern Brazil, AVL is endemic to the State of Ceará (CE), with outbreaks recorded in several municipalities. Fortaleza, Sobral, Juazeiro, Barbalha and Caucaia are priority municipalities for control activities; in fact, according to the classification by the Ministry of Health, Fortaleza is an intense transmission city [5,6]. ACL, caused by Leishmania (Viannia) braziliensis (Vianna, 1911), is unquestionably one of the major public health concerns in Ceará. It occurs in several areas, especially in mountainous regions and those adjacent to the coast, with municipalities such as Uruburetama, Itapajé, Barbalha and Missão Velha being the areas at greatest risk [7].
The São Francisco River Transposition Project aims at transforming the socio-economic reality in northeast Brazil by ensuring water supply to several states, including Ceará, where the transposition began in 2008. However, some negative impacts have been ascribed to this project, such as modifications of ecosystems that favor the expansion of vector populations that consequently lead to increase in vector-borne diseases such as leishmaniasis.
Both AVL and ACL are endemic to the State of Ceará; thus, the Ministry of Health recommended an evaluation of the populations of sand fly vectors, and monitoring the emergence of new likely sites of infection in municipalities under directly influence in the area affected by this (still ongoing) work. Entomological surveillance aids in increasing the knowledge of the sand fly fauna in general, and indicates the occurrence of potential leishmaniasis vectors.
The present study focused on entomological surveillance of sand flies aiming to support leishmaniasis prevention and control measures in the areas affected by the project. This enabled a global analysis of sand fly communities in dwellings, peridomicile and forest areas, and allowed characterization of the fauna in seven municipalities of Ceará under direct influence of the São Francisco River Transposition Project, in terms of species richness, diversity, dominance, abundance and similarity.

Methods
Monitoring of sand flies was performed in the areas of influence of Section I, in the North axis of the São Francisco River Transposition Project (Fig. 1).
The current physical execution rate of the São Francisco River Transposition Project is 89.9% (90.7 and 88.7% in the North and East axes, respectively), and three of the project's pumping stations have engines in operation. The impact area of the present study is included in Key Goal 3 N (81 km) of the North axis, extending from Boi II reservoir in the municipality of Brejo Santo, Ceará State to the Engenheiros Ávidos reservoir in the municipality of Cajazeiras, Paraíba State, with 94% of physical implementation [8].  [9] were employed in each MS, at a height of approximately 1 m from the ground, for three consecutive days, for a period of 12 h every day (18:00 to 06:00 h). The taxonomic diagnosis of sand flies followed Galati's proposal [10,11]. Abbreviations of generic and subgeneric names follow those proposed by Marcondes [12].
The Shannon's diversity index (H′) was used in the evaluation and the comparison of differences in composition within the sand fly community in each area of the study [13]. To assess the presence of significant differences between the diversity indices, Student's t-test was applied, with significance level of 5%, using the software DivEs® [14]. The diversity index (H′) was chosen for its suitability in studies comprising random samples of species within a community or subcommunity. Species dominance (D%) was defined by: D% = (i/t) × 100, where i is the total of individuals of a particular species and t is the total number of specimens collected. To evaluate the most abundant species in each MS, we applied the index of species abundance (ISA), which was normalized to values between zero and one by the standardized index of species abundance (SISA), according to the definitions proposed by Roberts & Hsi [15]. SISA values close to 1 correspond to the most abundant species. These indices were calculated using the formulae: where a is the number of samples in which the species was absent multiplied by c; c is the highest value of n obtained, considering all samples, plus 1; for each sample, species should be ranked from 1 to n (1 assigned to the most frequent species); Rj is the sum of the positions of each species; and K is the number of samples.
The similarity between habitats with respect to the number of species was estimated by the Sorensen-Dice qualitative similarity index (SI), based on the presence or absence of the species [16]. Kruskal-Wallis analysis, with significance level of 5%, was used to evaluate the occurrence of significant difference between sand fly populations in dwellings, peridomicile and forest areas, for each municipality, using the software IBM® SPSS® Statistics Version 23.
The monthly frequency was represented as total number of specimens captured per municipality. These data were obtained from the most abundant species, as well as the main vectors of leishmaniasis, considering the total numbers of specimens in dwellings, peridomicile and forest areas of each municipality, per month during the 44 months of captures.
The influence of rainfall on sand fly population densities was analyzed by Spearman's correlation coefficient with 95% and 99% confidence level, using the software
Ten species and 34,429 specimens were collected in the municipality of Jaguaribe, of which 99.4% were Lu. longipalpis, which was also the most abundant species (SISA = 0.977). Thus, all other species collected in this municipality had percentage values lower than 1%. A similar occurrence was observed in the municipality of Jaguaretama, where 14,008 specimens belonging to eight species were captured, of which 99.6% were Lu. longipalpis, which was the most abundant species (SISA = 1.000) (Tables 1 and 2).
The diversity indices showed that the Missão Velha municipality had the greatest diversity (H′ = 0.535). On the other hand, Jaguaribe and Jaguaretama showed lower Shannon's diversity index values (Table 1). These index values not only are influenced by the lower richness than that observed in other municipalities, but also reflect the strong dominance of Lu. longipalpis, which represented almost 100% of all specimens. However, those values did not differ significantly between MSs (P > 0.05) ( Table 3).
The similarity index indicates how similar two areas are in terms of species richness. Since areas with values greater than 0.5 are considered similar, all areas in the present study were also considered, with Missão Velha and Mauriti (SI = 0.89) and Jaguaribe and Jaguaretama (SI = 0.89) the most similar to one another, whereas Brejo Santo and Jaguaribe were the least similar (SI = 0.50). The Kruskal-Wallis analysis did not indicate significant differences in the number of sand flies captured in the dwelling, peridomicile and forest from all MSs (Missão Velha: P = 0.100; Brejo Santo: P = 0.506; Lavras de Mangabeira: P = 0.581; Iguatu: P = 0.654; Mauriti: P = 0.451; Jaguaribe: P = 0.771; and Jaguaretama: P = 0.412).
The sand fly density was correlated with rainfall, when compared with the month of capture and with previous month, and very little in the previous two months (Table  3). Sand fly fauna in the Iguatu municipality, including Lu. longipalpis, was influenced by regional climatic conditions: the highest population density occurred during the rainy months, with population peaks between January and June, whereas in the dry months (July to October), the population decreased. In Missão Velha, the sand fly fauna, and specifically the species Ev. evandroi, Ev. lenti, Lu. longipalpis, Mg. migonei and Ny. intermedia, showed negative correlation with rainfall. For these species, the negative correlation occurred with rainfall in the month of capture (Month 0), as well as in the previous month (Month 1), and in two previous months (Month 2). This negative correlation with rainfall also occurred with the total number of specimens (Month 1) and with Ev. evandroi (Month 1), Ev. lenti Mauriti. In Jaguaribe and Jaguaretama municipalities, the sand fly density, and specifically Lu. longipalpis, presented positive correlation with rainfall of the previous two months (Month 2) ( Table 4, Fig. 4).

Discussion
The environmental impacts of anthropic activities such as continuous deforestation, construction of hydroelectric plants, works and projects of different nature, migration, implementation of major agricultural projects, and military activities can provide new ACL epidemiological profiles in several American countries [1][2][3]. In this context, they can also contribute to the urbanization of diseases, since the transmission cycles tend to occur in dwellings and peridomicile [3].
The hypothesis that alterations caused by any environment-modifying activity might affect the population dynamics of leishmaniasis sand fly vectors should be considered. Especially, the alterations associated with deforestation and considerable environmental changes, thinking the competence of the vector in adapting to the urban environment. In this last aspect, it is known that sand flies are opportunistic and have eclectic feeding habits. In fact, all environmental change processes, in some way, are determinant in enabling a close contact between humans and leishmaniasis vectors.
The Brazil's San Francisco River project is already a reality in Paraíba and Pernambuco, Brazilian states that are very close to Ceará. The Brazilian Government estimates that the project will be completed in Ceará in the next months. These municipalities will continue to be monitored, by the State Department of Health and by our team, to assess the effects of environmental impacts not only on Lu. longipalpis populations, but also on the incidence of visceral leishmaniasis in dogs and of AVL and ACL in humans.
The differences in population curves, for each municipality, could be related to some characteristics. Physiogeographic aspects and soil type can influence sand fly populations, by affecting breeding sites of immature forms. The presence of Lu. longipalpis was observed in every month throughout the year, with the most visible population peaks occurring in the months following the rainy season; an exception was the Iguatu municipality, wherein the population peaks occurred during the rainy period. In this particular case, some other environmental factors may have favored this different behavior during the captures. For this species, we found correlation with rainfall in the months of capture and up to two months before. According to Deane [17], the tendency of Lu. longipalpis predominance in the rainy season has already been observed in the semi-arid areas of north-eastern Brazil, where populations of L. longipalpis increased during rainy seasons. However, in other biogeographic areas, this pattern may change, for example, in eastern Costa Rica, an endemic area of AVL, where the climate is hot and rainfall low [18], similar to the north-eastern Brazil, where greater abundance of Lu. longipalpis occurs in the dry season. It should also be considered that the establishment of this vector in peridomicile areas depends on some factors such as abundant organic matter for the development of larvae and availability of food sources for adults.
In fact, it is still questionable whether the density of sand fly vector species has an impact on the occurrence of human cases. Keeping in mind that high density of the vector can increase the number of specimens infected with Leishmania spp. and thereby increase the risk of transmission, this is a hypothesis that should not be disregarded. The records and population densities of Lu. longipalpis, the main AVL vector in Brazil, stand out for their relevance in all municipalities investigated. The strong dominance of this sand fly as compared to other species in all seven municipalities, mainly in dwellings and peridomicile, is notorious. Although highly anthropized, the seven municipalities exhibited average specific  Two-tailed test, with a 99% and 95% confidence limit *P < 0.05, **P < 0.01 richness but low diversity of sand flies. The ability of Lu. longipalpis to feed often on domestic and synanthropic animals and its remarkable anthropophilia, favor its adaptation to changing environments, enabling the maintenance of the visceral leishmaniasis transmission cycle in rural areas and its spread to urban areas [19].
Early studies on AVL in Ceará were performed in the city of Sobral, where 100,000 inhabitants died during an outbreak in 1953; researchers were able to identify the vector, Lu. longipalpis, as well as domestic and wild reservoirs of Leishmania (Leishmania) infantum [17,20]. Since then, AVL has been reported as endemic to the state. In the period corresponding to the present study (between 2011 and 2014) the highest incidence of AVL occurred in 2011, with 611 new cases. Among the investigated municipalities, those considered of intense transmission (Missão Velha, Brejo Santo and Mauriti) recorded 30, 26 and 39 cases of AVL, respectively.
Evidence from studies conducted in Serra de Baturité, Ceará, suggested Ny. whitmani as an ACL vector due to the finding of natural infection with L. (V.) braziliensis, the high density of the species in transmission sites, and its high degree of anthropophilia [21][22][23]. dwellings, naturally infected with L. (V.) braziliensis [21][22][23] and anthropophilic behavior, was also suggested as an ACL vector. Both species may possibly share the transmission of the parasite. Psychodopygus wellcomei Fraiha, Shaw & Lainson, 1971 is also present in forested areas in this municipality, but so far, no evidence of its participation in the transmission of ACL has been found [24].
Among the municipalities where ACL cases were recorded over the study period, Missão Velha showed the highest number of cases (n = 23), 12 of which in 2014. Albeit in a smaller number, this locality recorded the occurrence of Ny. intermedia (n = 305), Mg. migonei (n = 53), and Ny. whitmani (n = 77).
Previous studies have reported a coincidental distribution of the disease and of Ny. whitmani and Mg. migonei in Ceará, suggesting these sand flies were acting as ACL vectors within the state [22,25,26]. Considering the other municipalities with records of ACL transmission, Iguatu did not record the presence of any of these vectors; thus, it would be interesting to investigate whether registered cases are actually coming from this municipality, and if so, to assess whether any of the species found in the municipality took part in the transmission. While Ny. intermedia and Mg. migonei occurred in Brejo Santo, Lavras da Mangabeira and Mauriti, Ny. whitmani was found solely in Mauriti.
In fact, environmental changes can alter the epidemiology of leishmaniasis, affecting mostly vectors and reservoirs, leading to a considerable increase in the number of people exposed to the risk of infection [1][2][3].
The São Francisco River Transposition Project, regarding the environmental impact and particularly the record of both AVL and ACL in directly affected municipalities, should monitor the vector population curve, and the possibility that the occupation of new habitats might increase the transmission risk area. Above all, health managers should evaluate the need for surveillance and prevention measures against these diseases.
According to the policies of the Brazilian Ministry of Health for surveillance and control of leishmaniasis [1,2] and following the guidance from the World Health Organization [3], considering the diversity of sand fly vectors of ACL and AVL, an entomological surveillance measures have been recommended. It is important monitor environmental changes that may affect the behavior of sand fly vectors, especially near and within transmission risk areas.

Conclusions
This study confirms the presence of several Leishmania spp. vectors in the seven municipalities affected by the São Francisco River Transposition Project, with Lu. longipalpis being the most abundant species all study sites. Vector populations in these municipalities should be monitored, ultimately to assess the associations between environmental changes and sand fly population dynamics and leishmaniasis transmission risk.