Abstract:
The SolarMal project on Rusinga Island, Western Kenya, aims to eliminate malaria by reducing the mosquito population, using a new developed mosquito trap which will be installed at all households on the island. The effect of the installed traps is monitored by sampling mosquitoes from randomly selected selected households on the island during the whole duration of the project. This thesis project aims to support the SolarMal project by performing spatial analysis to the distribution of mosquitoes in relation to the environment on Rusinga Island. Environmental variables were searched that are, according to literature, potential determinants of mosquito presence. Second, potential determinants were validated against the mosquito distribution on Rusinga Island. Third, the fitness of the available spatial data, both environmental and mosquito catches, to relate with each other, was examined. In order to study the environment of the island, an elevation map (30 m. ASTER) and a satellite image (2,4 m QuickBird) were available for the creation (in ArcGIS 10.2) of 14 environmental variables, for example the slope and topographical wetness index of the area. These environmental variables were related (within the R environment) to a spatial dataset consisting one year of mosquito catches. This study shows that there are only weak correlations (R2 < 0,11) between the studied environmental variables and the mosquito catches. The spatial distribution of malaria vector mosquitoes is highly varying over time and shows small preferences for specific areas on the island. It is concluded that the available spatial data was not suitable for explaining the spatial distribution of adult mosquitoes on Rusinga Island. An explanation for this is found in the small scale breeding sites that are commonly found on Rusinga Island like tire tracks, footprints of cattle in drenched grass, or dumped waste in bushes, which were not detectable. The spatial data for the creation of environmental variables are limited in the spatial resolution to detect the small scale breeding sites (< 0,5 m) and are especially in the temporal resolution too limited for detection of temporal breeding sites. Continuations of this study are highly recommended to increase the spatial resolution of the data in combination with the inclusion of other environmental variables that are indicators for the chance on the mentioned small scale breeding sites. In order to increase the temporal resolution, inclusion of precipitation data is recommended