Satellites, space, time and the African trypanosomiases

Abstract

The human and animal trypanosomiases of Africa provide unique challenges to epidemiologists because of the spatial and temporal scales over which variation in transmission takes place. This chapter describes how our descriptions of the different components of transmission, from the parasites to the affected hosts, eventually developed to include geographical dimensions. It then briefly mentions two key analytical techniques used in the application of multi-temporal remotely sensed imagery to the interpretation of field data; temporal Fourier analysis for data reduction, and a variety of discriminant analytical techniques to describe the distribution and abundance of vectors and diseases. Satellite data may be used both for biological, process-based models and for statistical descriptions of vector populations and disease transmission. Examples are given of models for the tsetse Glossina morsitans in the Yankari Game Reserve, Nigeria, and in The Gambia. In both sites the satellite derived index of Land Surface Temperature (LST) is the best correlate of monthly mortality rates and is used to drive tsetse population models. The Gambia model is then supplemented with a disease transmission component; the mean infection rates of the vectors and of local cattle are satisfactorily described by the model, as are the seasonal variations of infection in the cattle. High and low spatial resolution satellite data have been used in a number of statistical studies of land cover types and tsetse habitats. In addition multi-temporal data may be related to both the incidence and prevalence of trypanosomiasis. Analysis of past and recent animal and human trypanosomiasis data from south-east Uganda supports the suggestion of the importance of cattle as a reservoir of the human disease in this area; mean infection prevalences in both human and animal hosts rise and fall in a similar fashion over the same range of increasing vegetation index values. Monthly sleeping sickness case data from the districts and counties of south-east Uganda are analysed and often show significant correlations with local LST. Case numbers increase with LST in areas that are relatively cooler than average for this part of Uganda, but decrease with LST in areas that are on average warmer. This indicates different seasonal cycles of risk across the region, and may be related to the differing vectorial roles of the two local tsetse, G. fuscipes and G. pallidipes. Finally, the increasing pace of change, and the likelihood of new or reemerging vector-borne diseases, highlight the need for accurate and timely information on habitat changes and the impacts these will have on disease transmission. The next generation of satellites will have significantly more spectral and spatial resolution than the current satellites, and will enable us to refine both statistical and biological predictions of trypanosomiasis and other vector-borne diseases within disease early warning systems.