Hence,plant benches the impact of water development projects on malaria transmission is variable and likely depends on the ecology of local mosquito vectors, underlying ecological factors, epidemiologic setting, socioeconomic conditions, and existing malaria control measures . Thus, its complexity can only be understood through site-specific evaluation of these parameters. Insecticide-based vector control interventions mainly long-lasting insecticidal nets and indoor residual spraying have been implemented to reduce malaria transmission with significant impacts. These tools have resulted in dramatic reduction in the proportion of endophagic and anthropophilic malaria vector species such as Anopheles gambiae, An. coluzzii, and An. funestus and a proportionate increase in An. arabiensis, which tend to be exophagic and less anthropophilic. However, previous studies indicate that vectors can develop resistance to insecticides or adapt to the presence of insecticides by becoming partially zoophilic and exophilic. Hence with the scale up of LLINs and widespread use of IRS, there is likely to be a shift in vector dominance from the highly endophilic An. gambiae/An. coluzzi and An. funestus to the more zoophilic and exophilic An. arabiensis . There is a pressing need to enhance our understanding on the effect of irrigation in a site where there is malaria vector control. The study aims to assess the effect of a recently established irrigation scheme in Homa Bay, Kenya, on malaria vector bionomics and transmission. Vector control intervention using LLINs and IRS with organophosphate, pirimiphos-methyl was being undertaken during the study period. Long-term success of the current malaria control efforts, ITNs and IRS, is dependent on continuous operational surveillance of the mosquito vectors, thus an effective mosquito sampling tool is required. Hence, the secondary goal was to compare the trap effectiveness of Centers for Disease Control and Prevention light traps against the gold standard, human landing catches . Results of this study will serve as the baseline vector bionomics and malaria transmission pattern for the evaluation of the success of core vector interventions and inform policymakers in planning and guiding future interventions especially in irrigated areas where there is scale up of LLINs distribution and application of IRS.Seasonal surveys were conducted in the dry and wet seasons in 2019 using five different trapping methods .

Indoor and outdoor host-seeking vector collections using CDC light traps and HLCs were undertaken in two randomly selected clusters in each zone. There were 160 trap-nights for each trap. Indoor and outdoor resting vector collections using pyrethrum spray catches , clay pots , and pit shelters were undertaken in four randomly selected clusters in each zone. There was a total of 320 trap-nights for each trap and 144 for pit shelters. Longitudinal adult vector surveillance was conducted using PSCs in four clusters in each zone for malaria vectors population dynamic research between 2018 and 2019 .Human landing catches were conducted both indoor and outdoor to assess vector host-seeking behavior . In each compound, vectors were collected indoors and outdoors 5 m away from the sentinel indoor collection house . Collection was undertaken by four volunteers, two in each of the indoor and outdoor stations, who alternated after 6 h. Hourly collections were done from 6 p.m. to 6 a.m. each night, with 45 min of collection and a 15 min break per hour. Each hourly collection was placed in individually labeled paper cups and maintained with a 10% sugar solution pad and then placed in a cool box. All collections were supervised by a team leader. Vector collections were undertaken in fve randomly selected houses in each cluster for four consecutive nights once per season. All collectors were provided with anti-malarial chemo-prophylaxis during the study period.Data were entered in Microsoft Excel 2010 data sheets and analyses were done using R statistical software . Mean density and proportions were calculated for vector populations. The density of adult anopheline mosquitoes was calculated as the average number of female mosquitoes per house per night . Several models were evaluated for the analysis of vector density, and the model with the lowest akaike information criterion and variables of interest was selected as the best model . In the analysis of seasonal data, a negative binomial mixed model was fitted to analyze Anopheles densities outdoor, indoor, and trapping method . Zone and trapping methods were fitted as the fixed variables in the outdoor and indoor models while zone, trapping methods, and location were considered as fixed variables in the trapping method model. In the indoor and trapping method models,gardening rack house number and date were used as covariates whereas house number and cluster were used as covariates in the outdoor model. In the analysis of longitudinal data, a NBMM with repeated measures were fitted to compare Anopheles densities and seasonality in the two zones by adjusting for months . Zone and season were fitted as the fixed variables and year: , date: cluster, cluster considered as covariates. The chi-square test was used to compare differences in vector species gonotrophic stage proportions between indoor and outdoor collections and also the zones. The human blood index for each mosquito species was calculated as the proportion of mosquito samples that had fed on humans out of the total number tested .

Sporozoite rates were calculated as the proportion of Anopheles mosquito samples positive for Plasmodium spp. out of the total number tested. The human biting rate was calculated as the product of blood-fed females per person per night and the human blood index. Annual entomological inoculation rates were calculated as the product of the sporozoite rate and the human biting rates .This study investigated the effect of concrete canal and food irrigation systems on species composition, malaria vector abundance and their seasonality, vector behavior, and malaria transmission in irrigated and non-irrigated sites where vector control using LLINs and IRS was being undertaken during the study period. Findings of the study demonstrated that An. arabiensis was the dominant anopheline species and was more abundant in irrigated zone compared to the non-irrigated zone. The high vector density of An. arabiensis together with their potential to transmit P. falciparum confirms the significant risk of malaria transmission in populations living within the irrigation scheme. The secondary aim of the study was to compare the trap effectiveness of CDC light traps against the gold standard, HLC. The results indicated that CDC light traps were equally effective in sampling malaria vectors as the HLC. This indicates the usefulness of this tool for continuous operational surveillance of the mosquito vectors within the study site. The distribution of An. arabiensis is generally concentrated in the drier savannah environments where rainfall is<1000 mm . Anopheles funestus was scarcely collected in the study site. Other anopheline species that were reported, and occurred only in the irrigated zone, were An. coustani and An. pharoensis. These two mosquito species have previously been considered as secondary vectors of malaria in Africa but recent studies have shown that they could play an important role in malaria transmission . Thus, it is prudent to integrate them in malaria vector surveillance and control strategies particularly where they are abundant. A significant variation in vector density was observed in the irrigated and non-irrigated zones which is consistent with previous studies that the introduction of irrigation schemes leads to an increase in vector density and abundance . In the irrigated zone, the irrigated canals, seepage areas, and fooded irrigated fields serve as the main larval habitats and provide stable mosquito breeding habitats during the dry season when other larval habitats dry up. In contrast, the low An. arabiensis density in the non-irrigated zone may be due to the temporary andparched nature of aquatic habitats during the dry season. These observations have also been reported in similar studies in the Mwea Irrigation Scheme and the neighboring non-irrigated agroecosystems .

The indoor resting density of An. funestus was generally low; however, their indoor abundance was relatively high during the dry season of 2018 in the irrigated zone prior to the application of the IRS in the study site. Thereafter, An. funestus was rarely collected from our study. This can be attributed to the application of Actellic® 300CS IRS which has been shown to significantly reduce the indoor resting density of An. funestus in the same area . Indoor residual spraying is known to be highly effective on endophilic and anthropophilic mosquito species such as An. funestus due to high exposure to the wall sprayed insecticides. The indoor and outdoor host-seeking density of An. arabiensis varied significantly with the highest biting densities collected indoors. Studies conducted over three decades ago by Githeko et al. showed that An. arabiensis was more likely to bite indoor than outdoor before the scale up of vector control intervention in western Kenya . In the present study, the endophagic tendency of An. arabiensis was still observed despite the high LLINs coverage and application of IRS in the study sites. This could be attributed to behavior of this species whereby it may enter a house protected with malaria vector control interventions in search of unprotected host, but exit without fatal exposure to insecticide-treated surfaces . Nonetheless, outdoor biting behavior was observed for this mosquito species which is consistent with other studies within its distribution range . It is worth mentioning that when comparing the traps deployed in the study, the mean vector density varied significantly between traps. The outdoor density of An. arabiensis was significantly higher in pit shelters and clay pots than for CDC light traps; in contrast, the indoor density of An. arabiensis was not significantly different between traps. Such variations were likely driven by differences in vector behavior, vector species composition, and history of malaria interventions rather than differences in the efficiency between the traps. Wide variation in the vector density of each trapping method has also been observed by Degefa et al. in a study conducted in western and attributed this variations to vector behavior . Human landing catches have been considered the gold standard method for estimating mosquito-human contact. However, it is a labor-intensive procedure requiring highly trained collectors, extensive supervision, variation in the skill of the collectors or their individual attractiveness to mosquitoes, and ethical concerns associated with potential exposure to infectious mosquito bites . Our results indicate that CDC light trap is an effective trapping alternative to HLC for continuous operational surveillance of mosquito vectors within the study sites. In a recent study conducted in western Kenya and southwestern Ethiopia, human-odor-baited CDC light traps collected twice the number of outdoor host-seeking An. arabiensis and An. funestus compared to non-baited CDC light traps . Thus, it will be important to evaluate the effectiveness of this tool in the study sites that could be a better outdoor surveillance tool than the non-baited CDC light trap.