Spatio-temporal Dynamics of Wetlands and Malaria in the Ethiopian Highlands Using Multi-sensor Satellite Observations

Tropical wetlands provide various beneficial ecosystem services; however, they can also facilitate the transmission of vector-borne diseases. Because wetlands serve as breeding habitats for Anopheles mosquitoes, particularly during the dry season, they are critical eco-hydrologic elements for malaria transmission. The overarching hypothesis of this study is that landscape and regional patterns of wetlands are associated with malaria risk in the Amhara region of Ethiopia. To test this hypothesis, we developed a random forest decision tree model to map seasonal and permanent wetlands in the Amhara region. Wetland training and validation data were acquired from high-resolution imagery in Google Earth and ground surveys. We evaluated the effectiveness of three random forest models using the following sets of predictor variables: (1) topographical indices from 30 m SRTM data, (2) individual reflectance bands and multispectral wetness indices from Landsat TM/ETM+ imagery, and (3) combined spectral and topographic data. The combined model produced the most accurate wetland maps, and we used it to map wetlands across the study area for 2000, 2005, and 2010. We found spatial associations between indicators of malaria risk from historical surveillance data and metrics of wetland cover at a district level. We also quantified seasonal moisture variability among three different land use land cover types (permanent wetland, seasonal wetland, and cropland) using Actual Evapotranspiration (ETa) over a ten year period (2001-2010) derived from MODIS imagery. We found that permanent and seasonal wetlands have peak moisture during the major malaria transmission season (September-November), whereas the permanent wetlands retain moisture and potentially sustain mosquito populations during the low transmission season (December-March). These findings about the spatial and temporal associations of malaria risk and wetlands can help to highlight areas that likely sustain transmission during the dry season and assist public health workers in planning and implementing malaria prevention, control, and elimination strategies for the region.