Malaria Transmission around the Memve’ele Hydroelectric Dam in South Cameroon: A Combined Retrospective and Prospective Study, 2000–2016

Dam constructions are considered a great concern for public health. The current study aimed to investigate malaria transmission in the Nyabessan village around the Memve’ele dam in South Cameroon. Adult mosquitoes were captured by human landing catches in Nyabessan before and during dam construction in 2000–2006 and 2014–2016 respectively, as well as in the Olama village, which was selected as a control. Malaria vectors were morphologically identified and analyzed for Plasmodium falciparum circumsporozoite protein detection and molecular identification of Anopheles (A.) gambiae species. Overall, ten malaria vector species were identified among 12,189 Anopheles specimens from Nyabessan (N = 6127) and Olama (N = 6062), including A. gambiae Giles (1902), A. coluzzii Coetzee (2013), A. moucheti Evans (1925), A. ovengensis Awono (2004), A. nili Theobald (1903), A. paludis Theobald (1900), A. zieanni, A. marshallii Theobald (1903), A. coustani Laveran (1900), and A. obscurus Grünberg (1905). In Nyabessan, A. moucheti and A. ovengensis were the main vector species before dam construction (16–50 bites/person/night-b/p/n, 0.26–0.71 infective bites/person/night-ib/p/n) that experienced a reduction of their role in disease transmission in 2016 (3–35 b/p/n, 0–0.5 ib/p/n) (p < 0.005). By contrast, the role of A. gambiae s.l. and A. paludis increased (11–38 b/p/n, 0.75–1.2 ib/p/n) (p < 0.01). In Olama, A. moucheti remained the main malaria vector species throughout the study period (p = 0.5). These findings highlight the need for a strong vector-borne disease surveillance and control system around the Memve’ele dam.


Study Design
The study combined retrospective and prospective malaria entomological surveys. Retrospective data were collected from the results of entomological surveys conducted from 2000 to 2006 and stored in a mosquito database at the Malaria Research Laboratory of Organisation de Coordination pour la lute contre les Endémies en Afrique Centrale (OCEAC). Prospective data were collected through cross-sectional surveys conducted in Nyabessan and Olama villages from 2014 to 2016 (July-December), two to four years after the launch of the Memve'ele dam construction in 2012.

Study Sites
The study was carried out in Nyabessan (2 •  2016 (July-December), two to four years after the launch of the Memve'ele dam construction in 2012.
Nyabessan is located on the edge of the Memve'ele dam across the Ntem River, at 220 km from Yaoundé, the capital city of Cameroon. The local human population has increased from ≈200 inhabitants in 2012 to ≈600 inhabitants in 2016, following the influx of workers involved in the dam construction. Most of the dwellings are made of temporary materials, but some houses are built with cement material. The area is characterized by the equatorial climate, with two rainy seasons from March-June and from September-November, and two dry seasons from December-February and from July-August. The average annual rainfall is 1500 mm and the average annual temperature is ≈25 °C. In 2000 and 2001, mosquito samples were collected in Oveng (2°10′ N, 10°30′ E), a village situated in the Nyabessan neighborhood, 5 km upstream along the Ntem river, and sheltering less than 100 inhabitants.
Olama, located on the edge of the Nyong River, at 150 km from the Memve'ele dam, was selected as the control village with no nearby hydraulic infrastructures. This village shelters ≈300 inhabitants and the houses are mostly built with clay. Similar to Nyabessan, this area is characterized by the equatorial climate.  Nyabessan is located on the edge of the Memve'ele dam across the Ntem River, at 220 km from Yaoundé, the capital city of Cameroon. The local human population has increased from ≈200 inhabitants in 2012 to ≈600 inhabitants in 2016, following the influx of workers involved in the dam construction. Most of the dwellings are made of temporary materials, but some houses are built with cement material. The area is characterized by the equatorial climate, with two rainy seasons from March-June and from September-November, and two dry seasons from December-February and from July-August. The average annual rainfall is 1500 mm and the average annual temperature is ≈25 • C. In 2000 and 2001, mosquito samples were collected in Oveng (2 • 10 N, 10 • 30 E), a village situated in the Nyabessan neighborhood, 5 km upstream along the Ntem river, and sheltering less than 100 inhabitants.
Olama, located on the edge of the Nyong River, at 150 km from the Memve'ele dam, was selected as the control village with no nearby hydraulic infrastructures. This village shelters ≈300 inhabitants and the houses are mostly built with clay. Similar to Nyabessan, this area is characterized by the equatorial climate.

Mosquito Collection and Morphological Identification
Adult mosquitoes were collected by human landing catches (HLC) performed by volunteers living in the prospected study sites. We used a stratified random sampling approach to select the houses and volunteers for mosquito collection. Each village was divided into three to five collection units, and one house was randomly selected in each unit for mosquito collection. At each selected house, one volunteer collected mosquitoes inside the house whereas another one collected mosquitoes from outside the house. In each village, mosquito collection was performed per night by 12-20 volunteers working in two teams of 6-10 collectors. The first team collected from 19:00 to 01:00 and the second team collected from 01:00 to 05:00. To avoid bias due to the attractiveness and skills of the mosquito collectors, the volunteers rotated between collecting indoors and outdoors every two hours. Mosquitoes were collected through cross-sectional entomological surveys conducted during small rainy seasons (Mars, April, May), small dry seasons (June, July), big rainy seasons (August, September), and big dry seasons (December). The total number of person-nights for mosquito collection in each study sites was 109 in Nyabessan and 106 in Olama (Table 1). Collected mosquitoes were morphologically identified using the standard identification keys [32,33], and individually stored in labeled 1.5 mL Eppendorf tubes containing silica gel desiccant for further analysis.

Molecular Identification
All female mosquitoes morphologically identified as belonging to the A. gambiae complex were subjected to molecular identification. DNA was extracted from the legs and wings using CTAB (Cetyl trimethyl ammonium bromide) 2% as described by Collins et al. [34] and used for identification of sibling species by means of the SINE200 (short interspersed elements) polymerase chain reaction (PCR) protocol as described by Santolamazza et al. [35]. SINEs are homoplasy-free and co-dominant genetic markers for differentiation of A. gambiae and A. coluzzii.

Determination of Plasmodium Falciparum Infectivity
For each female anopheline mosquito, the head and thorax were tested for the presence of circumsporozoite protein (CSP) of Plasmodium falciparum using enzyme-linked immunosorbent assay (ELISA) as described by Burkot et al. [36] and later modified by Wirtz et al. [37].

Statistical Analysis
For both retrospective and prospective data, the following entomological indicators were calculated: • The Human Biting Rate (HBR) which represents the average number of bites received per person per night; • The Infection Rate (IR) that measured the proportion of mosquitoes positive for P. falciparum Circumsporozoite antigene; • The Entomological Inoculation Rate (EIR), which is the number of infective bites received per person per night; EIR is calculated as the product of the HBR and IR.
The 2-sample test for equality of proportions was used to compare the proportions of each species before and after the launch of the dam construction and the ANOVA to compare the average HBR, IR, and EIR at 95% confidence interval. Data were analyzed using the R 3.5.0 software (R Development Core Team, Vienna, Austria, 2018).

Ethics Approval and Consent to Participate
The study was conducted under the ethical clearance no. 2016/01/685/CE/CNERSH/SP delivered by the Cameroon National Ethics (CNE) Committee for Research on Human Health. All volunteers participating in human landing catches signed a written informed consent form indicating their willingness to take part in the study. They also received free malaria prophylaxis.

Composition and Abundance of Anopheline Fauna in Nyabessan and Olama
A total of 12,189 female anophelines were collected in both study sites from 2000-2016. The distribution of species in the two study sites is shown in Table 1.
vector species/groups of species found in the two study sites, including A. nili (Table 1). However, A. moucheti remained the most abundant malaria vector species (69%-99%) regardless of the collection period, although its proportion significantly decreased in 2014 (69%) (p < 0.0001) and was balanced with A. ziemanni (29%) (Figure 2). The proportion of each other species including A. gambiae s.l. remained less than 3% during the seven entomological surveys conducted from 2000-2016 (p > 0.05).

Night Biting Rates Indoors and Outdoors
For each of the six major vector species (A. gambiae s.l., A. marshallii, A. moucheti, A. ovengensis, A. paludis, and A. ziemanni) that were identified, the average in/outdoor HBR before and during MDM in Nyabessan versus Olama are presented in Figure 3 and the effect plot in Figure  4. The ANOVA table for the linear model of HBR (Table 2) revealed that the variables period, species, interaction between period and species, hours, interactions between period and position (in/out), species and position, location and hour significantly explained variations in HBR. Q-Q plot of standardized residuals versus theoretical quantiles for this model revealed a slight deviation of the residuals from normal, with quite heavy tails. Over time, the proportions of A. moucheti, A. ovengensis, and A. marshallii significantly decreased (p < 0.0001), while the proportions of A. gambiae s.l. increased from 5% before MDC to 15% after MDC (p < 0.0001). PCR analysis of 164 specimens of A. gambiae complex collected during three consecutive years (2014-2016) revealed changes in the proportions of two sibling species, from 93% A. gambiae s.s. and 7% A. coluzzii in 2014 to 61% A. gambiae s.s. and 39% A. coluzzii in 2016. Similar to A. coluzzii, which was nearly absent before 2006 (all the 37 specimens tested as A. gambiae), A. ziemanni and A. coustani occurred during the implementation of activities associated with dam construction, although at low frequencies (<3%).
In Olama, 6062 anophelines were caught by 106 person-nights (PSN) (70 PSN before MDC, 36 PSN after MDC). The composition of anopheline fauna in Olama was similar to Nyabessan, with 8 vector species/groups of species found in the two study sites, including A. nili (Table 1). However, A. moucheti remained the most abundant malaria vector species (69%-99%) regardless of the collection period, although its proportion significantly decreased in 2014 (69%) (p < 0.0001) and was balanced with A. ziemanni (29%) (Figure 2). The proportion of each other species including A. gambiae s.l. remained less than 3% during the seven entomological surveys conducted from 2000-2016 (p > 0.05).

Night Biting Rates Indoors and Outdoors
For each of the six major vector species (A. gambiae s.l., A. marshallii, A. moucheti, A. ovengensis, A. paludis, and A. ziemanni) that were identified, the average in/outdoor HBR before and during MDM in Nyabessan versus Olama are presented in Figure 3 and the effect plot in Figure 4. The ANOVA table for the linear model of HBR (Table 2) revealed that the variables period, species, interaction between period and species, hours, interactions between period and position (in/out), species and position, location and hour significantly explained variations in HBR. Q-Q plot of standardized residuals versus theoretical quantiles for this model revealed a slight deviation of the residuals from normal, with quite heavy tails. The ANOVA model of human biting rates as a function of different time slots and species is provided in Table 3.
In Nyabessan, the mean HBRs of A. gambiae were 3-6 bites/person/night before and during MDC ( Figure 3). For A. moucheti and A. ovengensis, the HBRs decreased from 25 bites/person/hour and 40 bites/person/hour before MDC to 18 bites/person/hour and 15 bites/person/hour after MDC respectively (p < 0.001). The species A. paludis, which was scarce before MDC and after MDC until 2015, became one of the major species in 2016 with up to 35 bites/person/night (p = 0.003). Conversely, A. marshallii was found before MDM (0.3-4.4 bites/person/night), but became scarce later on after MDC (only 0.4 bite/person/night in 2014), although the difference was not significant (p = 0.72). The species A. ziemanni was only found after MDC, although at a very low density, less than 1 bite/person/night.
In Olama, the mean HBR of A. moucheti was more than three times higher before MDC (84 bites/person/night respectively) than later on during MDC (23 bites/man/night from) ( Figure 3). Nevertheless, the HBR of this species was still high compared with that of the five remaining vectors. The average HBRs of the remaining vectors (e.g., A. paludis and A. marshallii) were frequently less than 1 bite/person/night (Figure 3), except A. ziemanni, which reached 6 bites/person/night in 2014.

Indoor and Outdoor Biting Cycles
The effect plot showing the mean HBRs outdoors/indoors in Olama and Nyabessan is given in Figure 5. The average indoor/outdoor HBR of the six major vectors was not significantly different before MDC (60 bites/person/night) versus during MDC (52 bites/person/night) (p = 0.48) ( Table 2). However, there was a significant difference of HBR between outdoor and indoor collections at species level (p = 0.04), especially in A. moucheti, A. ovengensis, and A. paludis as compared with A. gambiae (p < 0.004) ( Table 3). The HBR was not significantly different between Nyabessan and Olama (p = 0.1) ( Table 2).
Throughout the nights from 19:00 to 05:00, variable trends of biting activity were noted either in

Indoor and Outdoor Biting Cycles
The effect plot showing the mean HBRs outdoors/indoors in Olama and Nyabessan is given in Figure 5. The average indoor/outdoor HBR of the six major vectors was not significantly different before MDC (60 bites/person/night) versus during MDC (52 bites/person/night) (p = 0.48) ( Table 2). However, there was a significant difference of HBR between outdoor and indoor collections at species level (p = 0.04), especially in A. moucheti, A. ovengensis, and A. paludis as compared with A. gambiae (p < 0.004) ( Table 3). The HBR was not significantly different between Nyabessan and Olama (p = 0.1) ( In Olama, A. moucheti displayed high HBRs indoors before MDC (up to 25 bites/person/night) versus outdoors (up to 15 bites/person/night), with a peak of activity between 23:00 and 01:00 as in Nyabessan. The peak biting activity outdoors at this period was recorded between 01:00 and 03:00. During MDC, the HBRs was less than 7 bites/person/night throughout the nights, either indoors or outdoors. The biting cycles of the remaining anopheline species could not be analyzed due to their absence or very low densities.   Figure 7. The biting cycles varied according to the survey period (before/after MDC), the biting position (indoor/outdoor), and the anopheline species. In general, the biting rates were significantly higher between 23:00 and 01:00 or between 01:00 and 03:00 (p < 0.02) as compared with 19:00 and 21:00 (Table 3). Four major vector species (A. gambiae, A. moucheti, A. ovengensis, and A. paludis) showed an ability to indifferently bite indoors and outdoors throughout the study period (p = 0.48) ( Table 2).

Infection Rates of Malaria Vectors in Nyabessan and Olama
Data on P. falciparum infection rates (IR) by vector species in Nyabessan and Olama are presented in Figure 8 and the ANOVA model in Table 4.
In Nyabessan, four anopheline species, namely A. gambiae s.l., A. marshallii, A. moucheti, and A. ovengensis were mainly responsible for Plasmodium transmission irrespective of the survey year, followed by A. paludis, which arose as the local malaria vector species in 2016. No significant difference of IR was observed between the period before (1.1%) and during dam construction (1.0%) (p = 0.16). In 2015, the IR for all five species was 0.0%. Some variations of IR were observed among the vector species, although theses variations were not significant (p = 0.07) (Figure 8). None of the variables period, species, location, and their interaction significantly explained variations in IR (Table In Nyabessan, A. gambiae hourly HBRs before MDC were flat either outdoors or indoors. However, during MDC, the peaks of biting activity occurred between 01:00 and 03:00 indoors, or after 05:00 outdoors (4 bites/person/hour). With A. moucheti, the biting cycle indoors or outdoors displayed similar tendencies before and during MDC, with peaks of hourly HBRs between 23:00 and 01:00 indoors or earlier (21:00 and 23:00) outdoors. Furthermore, a significant decrease of HBRs was clearly observed from 5-15 bites/person/hour in/outdoors before MDC to less than 5 bites/person/hour during MDC, no matter the position of mosquito collection (Figures 5 and 7). With A. ovengensis, variable shapes of biting cycles were observed. Before MDC, the HBRs gradually decreased from 10 to 15 bites/person/hour (indoors and outdoors) during the first half of the night to less than 5 bites/person/hour (indoors and outdoors) during the second half of the night (Figures 5 and 7). During MDC, the HBRs remained less than 5 bites/person/hour during the whole night. More importantly, we have found A. paludis with a high endophagic tendency, and a peak of HBR between 03:00 and 05:00 (up to 10 bites/person/night). The biting cycle of this species outdoors was flat.
In Olama, A. moucheti displayed high HBRs indoors before MDC (up to 25 bites/person/night) versus outdoors (up to 15 bites/person/night), with a peak of activity between 23:00 and 01:00 as in Nyabessan. The peak biting activity outdoors at this period was recorded between 01:00 and 03:00. During MDC, the HBRs was less than 7 bites/person/night throughout the nights, either indoors or outdoors. The biting cycles of the remaining anopheline species could not be analyzed due to their absence or very low densities.

Infection Rates of Malaria Vectors in Nyabessan and Olama
Data on P. falciparum infection rates (IR) by vector species in Nyabessan and Olama are presented in Figure 8 and the ANOVA model in Table 4.

Entomological Inoculation Rates of Malaria Vectors in Nyabessan and Olama
Data on entomological inoculation rates (EIR) by vector species in Nyabessan and Olama are presented in Figure 9 respectively and the ANOVA model in Table 5.
In Nyabessan, the entomological inoculation rate for each of the five main vector species was mostly less than 1 infective bite/person/night (ib/p/n), with no significant difference between the two study periods (p = 0.06). None of the variables period, species, location, and their interaction significantly explained variations in EIR (Table 5). Q-Q plot of standardized residuals versus theoretical quantiles for this model revealed very little deviation of the residuals from normal. However, the EIR of A. paludis significantly increased from 0.0 ib/p/n before MDC to 1.2 ib/p/n during MDC (Figure 9). The EIRs of each of the remaining species were less than 1 ib/p/n throughout the two study periods.
Similar to the IR, the overall EIR for all five species was not significantly different between Olama and Nyabessan (p = 0.06). The highest EIRs in Olama were recorded in A. moucheti before  In Nyabessan, four anopheline species, namely A. gambiae s.l., A. marshallii, A. moucheti, and A. ovengensis were mainly responsible for Plasmodium transmission irrespective of the survey year, followed by A. paludis, which arose as the local malaria vector species in 2016. No significant difference of IR was observed between the period before (1.1%) and during dam construction (1.0%) (p = 0.16). In 2015, the IR for all five species was 0.0%. Some variations of IR were observed among the vector species, although theses variations were not significant (p = 0.07) (Figure 8). None of the variables period, species, location, and their interaction significantly explained variations in IR (Table 4, Figure 8). Q-Q plot of standardized residuals versus theoretical quantiles for this model revealed slight deviation of the residuals from normal, with quite heavy tails.
No significant difference of IR was observed between Olama and Nyabessan (p = 0.68) (Table 4). However, in Olama, malaria was mainly transmitted by A. moucheti (0.7% to 2.5% IR) throughout the study period, followed by A. gambiae s.l. The IR of A. gambiae s.l. decreased from 6% before MDC to 0.0% during MDC, but this decrease was not significant as shown on the effect plot (Figure 8). For the remaining vector species, the IR was almost less than 1.5%.

Entomological Inoculation Rates of Malaria Vectors in Nyabessan and Olama
Data on entomological inoculation rates (EIR) by vector species in Nyabessan and Olama are presented in Figure 9 respectively and the ANOVA model in Table 5.

Discussion
The data presented in this study were collected during seven years since year 2000, 12 years before the launch of the Memve'ele dam construction in 2012, then for 3 years during the construction until the impoundment of the dam in 2016. Although entomological surveys were not conducted for consecutive years or during the entirety of the dry and rainy seasons, the current study brings up-todate information on the variable roles of An. moucheti together with other anopheline vector species in the evolution of malaria transmission, in a dual context of intensive environmental modifications and Long Lasting Insecticidal Net (LLIN) interventions in Nyabessan versus Olama, and similar localities in the equatorial regions of Cameroon.
Of the forty malaria vector species previously identified in Cameroon [38,39], ten species were recorded, among which A. gambiae, A. coluzzii, A. moucheti, and A. nili known as the major malaria vectors in the equatorial forest region, alongside A. funestus, which was not recorded either in Nyabessan or in Olama. Each Anopheles species was characterized by typical habitat preferences,  In Nyabessan, the entomological inoculation rate for each of the five main vector species was mostly less than 1 infective bite/person/night (ib/p/n), with no significant difference between the two study periods (p = 0.06). None of the variables period, species, location, and their interaction significantly explained variations in EIR (Table 5). Q-Q plot of standardized residuals versus theoretical quantiles for this model revealed very little deviation of the residuals from normal. However, the EIR of A. paludis significantly increased from 0.0 ib/p/n before MDC to 1.2 ib/p/n during MDC (Figure 9). The EIRs of each of the remaining species were less than 1 ib/p/n throughout the two study periods.
Similar to the IR, the overall EIR for all five species was not significantly different between Olama and Nyabessan (p = 0.06). The highest EIRs in Olama were recorded in A. moucheti before MDC (≈1.5 ib/p/n), afterward it decreased to less than 1.0 ib/p/n and remained at that level until 2016. For the remaining vector species, the EIR was almost 0.0 ib/p/n.

Discussion
The data presented in this study were collected during seven years since year 2000, 12 years before the launch of the Memve'ele dam construction in 2012, then for 3 years during the construction until the impoundment of the dam in 2016. Although entomological surveys were not conducted for consecutive years or during the entirety of the dry and rainy seasons, the current study brings up-to-date information on the variable roles of An. moucheti together with other anopheline vector species in the evolution of malaria transmission, in a dual context of intensive environmental modifications and Long Lasting Insecticidal Net (LLIN) interventions in Nyabessan versus Olama, and similar localities in the equatorial regions of Cameroon.
Of the forty malaria vector species previously identified in Cameroon [38,39], ten species were recorded, among which A. gambiae, A. coluzzii, A. moucheti, and A. nili known as the major malaria vectors in the equatorial forest region, alongside A. funestus, which was not recorded either in Nyabessan or in Olama. Each Anopheles species was characterized by typical habitat preferences, including exposure to sunlight, turbidity of the water, and presence of vegetation, pH or nitrate and phosphate concentrations of the water [40]. These environmental factors are specific for each dam and its shoreline. Furthermore, with the presence of a dam, flooding is common upstream from the dam. During flooding, water covers a vast area of land and when the water recedes, little ponds are left within the flooded area which are suitable breeding sites for mosquitoes. While fluctuations were observed in the distribution of four major vector species in Nyabessan before and during dam construction across the Ntem River, A. moucheti remained the major malaria vector species in Olama during the whole study period, followed by A. gambiae. Either in Nyabessan or in Olama, malaria vector biting rates were highly variable between indoors and outdoors depending on the period and species. In Nyabessan, the most interesting trend was that of A. paludis, which in 2016 when its HBR indoors was higher when compared to other malaria vector species as well as with its HBR outdoors. By contrast, A. ovengensis showed an exophilic tendency. These findings emphasize the variability of malaria vectors' biting behavior depending on the environmental factors and the high risk of Plasmodium transmission indoors and outdoors as well, as previously stressed [41,42]. Improved socio-economic variables, effective vector control programs or changes in health-seeking behavior are also pivotal in determining whether a water project triggers an increase in malaria transmission [43]. From the time when the first nationwide free distribution of LLINs was conducted in 2011-2012 and the second one in 2015, the general use of LLINs for malaria prevention might lead to temporary changes in vector biting behavior and malaria transmission system. The reduction of man-vector contact and increase of outdoor biting rates underline the need for outdoor vector control measures to complement LLNs interventions in these areas.
The increase of malaria transmission intensity in 2016 in Nyabessan was led by A. gambiae and A. paludis in the progressive replacement of A. moucheti. This switch of vector species distribution is consistent with previous studies conducted in some areas of the equatorial region of Cameroon that encountered considerable environmental modifications [17,[44][45][46]. In contrast to Olama where no significant ecological change occurred during the last decade, malaria was permanently transmitted by A. moucheti, as previously reported by Antonio-Nkondjio et al. [42]. A. moucheti is widely distributed across West and Central Africa, playing a major role in malaria transmission in villages located close to rivers or slow-moving streams in the equatorial region [17,[45][46][47]. The larvae of this species are frequently associated with lentic rivers, low temperatures, and the abundance of aquatic vegetation at the edge of the river [48]. Increased urbanization and deforestation as well as lower-scale landscape modification such as river banks cleaning for gardening and/or recreational purposes were shown to be highly detrimental to this species, fostering changes in the malaria vector system composition with a higher prevalence of An. gambiae, taking the lead over An. moucheti [48].
Anopheles paludis is known as a secondary malaria vector, essentially found in the equatorial forest and savanna from West Africa to East Africa [49]. In Cameroon, this anopheline species has been recorded in 21 villages [50], its role in malaria transmission in this country has been confirmed since 1964 [51]. However, very little was known about the increase of its frequency and its important role in malaria transmission in Nyabessan. The common breeding sites of this mosquito species are expanses of clear water with aquatic or semi-aquatic vegetation, most often sheltered from the forest canopy. Larvae are found in pools, ponds, backwaters, upright vegetation, and even along grassy banks of rivers. Previous studies revealed that A. paludis is very anthropophilic, exophilic, and exophagous [52]. Its biting cycle showed either an activity distributed over the night, or a peak of aggressiveness at dusk. This Anopheles species was reported to bite in the undergrowth during the day in Cameroon and the Central African Republic [53]. The current study provided up-to-date information on the important and increasing role of A. paludis in malaria transmission in Nyabessan among the vectors that were previously predominant in the study area such as A. moucheti and A. ovengensis as well as those that were scarce, such as A. gambiae s.l. The high frequency among the collected anopheline samples, the infection rate, biting cycles outdoors and indoors as well as the EIR reported here reveal that A. pludis may be responsible for high malaria risk in Nyabessan during the coming years. These data are comparable with those reported by Kharch et al. [52] in the Bandundu region of the Democratic Republic of Congo during the dry season. The peak biting activity of A. paludis in Nyabessan occurred between 03:00 and 05:00, suggesting significant daytime biting rates. Although such biting behavior is consistent with previous reports [53], the daytime biting behavior may also be related to high coverage of LLINs in households. The risk of malaria transmission might therefore become very high when people are awake and outside LLINs.
Furthermore, the current study revealed a peak of A. gambiae s.l. biting activity between 03:00 and 05:00. Together with A. paludis, A. gambiae was found to increasingly occur in Nyabessan with densities rising over time, with the highest value obtained in 2016. Conversely, the densities of A. gambiae in Olama remained low and constant throughout the collection period. Environmental modifications for habitat construction led by the massive arrival of dam workers in Nyabessan might have created suitable breeding sites for A. gambiae. This species is known to colonize aquatic habitats resulting from housing and agricultural activities in peri-urban areas in sub-Saharan Africa [42,[54][55][56][57]. Molecular identification of specimens belonging to the A. gambiae complex revealed that samples collected in Nyabessan were composed of two species, namely A. gambiae s.s. (76.3%) and A. colluzzii (1.43%), while only A. gambiae s.s. was previously found in this locality [58]. Interestingly, the few specimens collected in Olama from 2014-2016 (N = 7) were identified as A. gambiae s.s. The apparently new occurrence of A. colluzzi in Nyabessan may be associated with anthropization of the environment, opportunistic behavior of this species, as well as haphazard and uncontrolled human activities [59,60]. Furthermore, A. paludis, which was scarce before dam construction in Nyambessan, became wide spread in 2016, while in Olama, its proportion during the same period remained very low (0.43%). However, seasonal variations observed in the density of this species may also be related to rainfall or stream flow [38].
Lastly, the intermittent presence of A. ziemanni in at both study sites at low densities and the Plasmodium infection rates in Olama highlights its secondary role in local malaria transmission. Previous studies revealed A. ziemanni as a major malaria vector in Northern Cameroon [61]. This species typically breeds in swampy areas and large water bodies, such as those resulting from environmental modifications around the Ntem River for the Memve'ele dam construction.

Conclusions
Although the composition of anopheline fauna was similar in Nyabessan and Olama, variations observed in malaria transmission patterns (changes in vector species distribution, increased A. paludis abundance and entomological inoculation rates) following large dam construction in Nyabessan may lead to an increase in malaria cases. The data reported here are of a great importance for the implementation of additional measures to sustain malaria prevention in Nyabessan. The peak of biting activity between 03:00 and 05:00 and outdoor biting rates underlines the need for complementary vector control measures, especially those targeting outdoor malaria transmission. Further entomological surveys are needed to monitor the malaria entomological profile following the impoundment of the Memve'ele dam. Furthermore, future water resource developments should include a comprehensive assessment of potential health impacts and a multifaceted action plan for health improvement.

Conflicts of Interest:
The authors declare that there is no conflict of interest.