Capabilities of multispectral thermal data for identification of Anopheles gambiae mosquito larval habitats in African urban environments

The satellite-based Multi-spectral Thermal Imager (MTI) employs a 15-waveband sensor that measures radiation in the visible and infrared spectrums. The MTI exhibits 5 in visible to 20 m thermal resolutions. This research examines whether available MTI image data provide details signalling mosquito larval habitats in urban locales of Malindi and Kisumu, Kenya. MTI data encompassing visible wavebands 2, 3 and 4 were selected to synthesize images containing positive Anopheles gambiae larval habitats that ranged from swamps to tree holes. Nine observers with varying degrees of experience (0-6 years) in interpreting remotely sensed data were asked to visually inspect 100 constructed land cover images and discern aquatic larval habitats. Twenty images of negative control habitats were included. Specific habitat status was contemporaneously verified by field tests. As a group, the observers correctly identified 54 of 900 (6%) observations of the positive larval habitats and 126 of 180 (70%) observations of the negative control habitats. Habitats most correctly identified were puddles, although aquatic habitat size did not influence the approach to identification. These results suggest that MTI images assembled from optical wavebands 2, 3 and 4 do not offer adequate spatial information for remote visual detection of Anopheles gambiae larval habitats due to inadequate display of surface features. © 2005 Taylor & Francis Ltd.