Environmental Determinants of Malaria Transmission in Agricultural Communities Around Large Dams in Ethiopia

Abstract

Dams are key to ensuring food security and promoting economic growth in sub-Saharan Africa. However, the potential adverse public health impacts of dams, such as malaria, could undermine their intended benefits. Understanding the influence of dams on the distribution of malaria transmission in different ecological settings is thus crucial to devise tailor-made malaria control tools. This study assessed the impact of dams on malaria transmission at different eco-epidemiological settings, and evaluates the potential of optimized dam management for malaria control.<br/> To determine levels of malaria transmission around dams at different ecological settings, entomological and epidemiological surveys were conducted at three Ethiopian dams located in lowland, midland and highland areas. Larval and adult anopheline mosquitoes were collected from dam and non-dam (control) villages between October 2013 and July 2014. Female anophelines were tested for malaria sporozoite infection and blood meal sources. Five years of monthly malaria case data (2010-2014) were also analyzed. Mean monthly malaria incidence was two- and ten-fold higher at the lowland dam than at the midland and highland dams, respectively. Shoreline puddles and irrigation canals were consistently the major mosquito breeding habitats. Densities of larval and adult anophelines were also highest at the lowland dam village, followed by the midland and highland dam villages. Anopheles arabiensis was the predominant malaria vector species, followed by An. pharoensis and An. funestus sensu lato (s.l.) which were largely collected from lowland and midland dam villages. The annual Entomological Inoculation Rate (EIR) of An. arabiensis, An. funestus s.l. and An. pharoensis at the lowland dam village was 157.7, 54.6 and 48.6 infective bites per person per year, respectively. The annual EIR of An. arabiensis and An. pharoensis was 5.4 and 3.1 times higher at the lowland dam village than at the midland dam village. These data indicate that increased malaria associated with dams is higher in the lowland than midland and highland ecological settings.<br/> Factors linked to malaria transmission around the study dams were examined using environmental (elevation, distance from the reservoir shoreline, Normalized Difference Vegetation Index (NDVI), monthly average reservoir water level and monthly changes in water level) and meteorological (precipitation, and minimum and maximum temperature) data. Multiple regression analysis demonstrated that village distance from reservoir shoreline (lagged by 1 month) was negatively associated with malaria incidence around all three dams, while average monthly reservoir water level (lagged by 2 months) and monthly precipitation (lagged by 1 or 2 months) were positively associated with malaria incidence only at the lowland and midland dams. Similarly, minimum air temperature and monthly change in reservoir water level when lagged by 2 months were positively associated with malaria incidence at the highland dam. Maximum temperature did not show any correlation with malaria incidence at any of the study dams. These results suggest that reservoir factors (monthly average reservoir water level and reservoir water level change) were important predictors of malaria incidence.<br/> Different water level drawdown rates were tested in an experimental field setting to evaluate the potential of using reservoir water level management for larval mosquito control. Twelve experimental dams were constructed on the foreshore of the midland Koka Dam, and grouped into one of four daily water drawdown treatments: 0 (control), 10, 15 and 20 mm.day-1. Larval sampling was conducted weekly during the main transmission season (October to November 2013) and subsequent dry season (February to March 2014). Mean weekly larval density was highest in the control experimental dams throughout the study, and decreased significantly with increasing water drawdown rates in both seasons. The results indicate that faster water level drawdown rates help reduce larval vector abundance around dams.<br/> The results of the experimental work were then used to evaluate the potential of water level management for malaria control at the reservoir-scale. Digital elevation models were constructed for the three study dams to estimate reservoir parameters (surface area and perimeter of wetted shoreline) at different reservoir capacities (70, 75, 80, 85, 90, 95 and 100% full capacity). Water level drawdown rates of 10, 15 and 20 mm.day-1 were applied and larval abundance, entomological inoculation rate (EIR) and malaria prevalence were estimated for each reservoir capacity scenario. At the lowland dam, larval abundance increased with increasing reservoir volume and wetted shoreline area, although the opposite pattern was observed at the midland and highland dams due to differences in reservoir topography. Estimated EIR, malaria prevalence, malaria treatment and economic costs generally decreased when water level drawdown rate increased from 10 to 15 and 20 mm.day-1. The results indicate that increasing water level drawdown rate will reduce malaria transmission and the associated economic impacts around dams during the main transmission season. Findings of the present study highlight that by regulating the persistence of shallow shoreline breeding habitats, reservoir water level management could serve as a potential malaria vector control tool around African dams.<br/> This study underscores the benefits of optimized dam management by incorporating malaria vector control into reservoir management practices. Mosquito larval control using reservoir water level manipulation could therefore supplement the existing vector control measures to dramatically reduce malaria around dams. Future research should assess the practicability of dam management for malaria control in diverse African settings, including the social and economic costs of optimized dam operations on hydropower generation and downstream agriculture.