Impact of irrigation expansion on the inter-epidemic and between-season transmission of Rift Valley fever in Bura Sub-County, Tana River County, Kenya

Abstract

The government of Kenya has prioritized food production through revitalized and sustained agricultural expansion as a means of achieving food security for the rapidly growing population, uplifting the living standards of rural communities by providing them with livelihood opportunities as well as jumpstarting overall economic growth for employment and wealth creation through the export of excess produce and by products. Key among several policy directives issued in this regard was the harnessing of the country’s vast irrigation potential, through the expansion of existing irrigation schemes as well as establishment of new ones along the Tana and Athi river basins as well as along the shoreline of Lake Victoria. The development of these schemes however has the potential to alter local environmental and ecological conditions that may influence the risk of RVF disease transmission. Given that a thorough understanding of the risk factors precipitating the occurrence of any infectious disease is vital for its effective control, this study aimed to investigate whether these land-use changes associated with the development of irrigation schemes had any impact on the transmission dynamics of RVF virus. This study involved the spatio-temporal evaluation and comparison of the abundance, distribution and species diversity of potential vectors of RVF virus across three villages with differing ecological habitats in Bura, Tana River County, Kenya. These included the irrigated ecosystem represented by Bura irrigation scheme located near Bura township, the riverine ecosystem represented by Husingo village that is adjacent to Tana River and the dry, pastoral ecosystem represented by Chifiri village which is located further inland, away from both the irrigated and riverine ecosystems. A concurrent longitudinal study was also undertaken to measure and compare the risk of RVF on the local population of sheep and goats kept by households, and between the three ecosystems. Representative samples of mosquitoes were collected from all ecosystems and morphologically identified using taxonomic keys. The sources of blood meals were also examined in order to identify the host spectrum of engorged females while sampling of sentinel animals for RVF antibodies was undertaken as a direct measure of risk. Comparative and multivariable analysis between these ecosystems revealed that the irrigated and riverine ecosystems were similar in terms of mosquito abundance and seasonality, despite one being naturally occurring while the other being man-made. Further, the irrigated ecosystem maintained a constant and minimum presence of mosquitoes throughout all seasons, while the riverine ecosystem exhibited much more sensitivity to rainfall, with mosquito abundance significantly increasing during the wet season. The host spectrum of blood fed females revealed that most blood meals came from easily available and accessible hosts such as goats, sheep and humans. Screening of sheep and goats for RVFV antibodies detected several seroconversions in the riverine and irrigated ecosystem, with those within the riverine ecosystem all occurring in one month approximately eight weeks after the start of the rainy season while those within the irrigated ecosystem were spread out over several months thereafter. The seroconversion study, which was a direct measure of RVF disease risk, indicated that most seroconversions occurred during the rainy season within the riverine ecosystem, while those within the irrigated ecosystem were spread out over several months. No seroconversions were detected in the much drier pastoralist ecosystem. The findings of this study suggest that further expansion of the irrigation scheme or establishment of new ones in this region will lead to a gross increase in the abundance and diversity of total mosquitoes, as well as potential vectors of RVF virus. The results further imply that with increased numbers of vectors, in the presence of low numbers of animal hosts may pose an increased risk of spillover infection to humans as opportunistic hosts during large epidemics. In addition to irrigation expansion, rainfall and flooding still remains a significant risk factor for the transmission of RVF virus in this county, especially in the non-irrigated riverine and pastoral ecosystems, and particularly in the presence of large numbers of naive animal hosts. Practical implications of these findings include targeted vector surveillance especially of known vectors of RVF virus as well as the formulation and implementation of integrated vector and environmental control programs. Vector competence studies of other mosquito species identified that might be potential vectors are recommended in future in order to improve on current outbreak prediction models as well as evaluate the success of potential surveillance and control options.