Remotely sensed data from satellite images were used to study environmental changes associated with changing risk of cutaneous leishmaniasis (CL) and malaria in southeastern Turkey. A large irrigation project involving the Province of Sanliurfa has dramatically altered land use, cropping patterns, migration, and population density. Incidence of both CL and malaria appear to be increasing. Disease-specific application of remotely sensed data of different resolutions were employed to relate spatial variation and temporal changes in the environment to transmission incidence.
Cutaneous Leishmaniasis The environmental determinants of anthroponotic CL due to Leishmania tropica, unlike those for zoonotic CL, are not well understood. Anthroponotic CL transmission appears to be primarily urban, but many determinants of risk, including breeding sites for Phlebotomus sandfly breeding vectors are unknown. Epidemic CL has recently reemerged in the city of Sanliurfa (population 400,000), where disease incidence has increased ten-fold from 1991 (411) to 1994 (3968 cases). The spatial distribution of 1994 incidence was calculated by neighborhood (n=52) to range from 0 to 121 per 1000 residents. Using a Landsat MSS image (4 August 1992), an unsupervised classification of the city and surrounding areas identified four land cover types: urban, vegetation, fields, and bare soil/rock. Enlarged RGB color prints of the unclassified sub-scene, were used for local ground-truthing; GPS-derived locations were obtained and used to verify cover classes and determine neighborhood borders. The resulting vector overlay of borders then was applied to the image to calculate neighborhood-specific land cover statistics and relate these to CL incidence. Population (1990 census) and housing quality classification (city planning office) were included in the analysis. A positive correlation was found between the percentage of urban land cover and log10 CL classes (r=0.43; p=0.002). Conversely, vegetation was negatively associated with CL incidence (r=-0.51; p<0.001). Quality of housing also was correlated negatively with CL incidence and positively with population density (all p's <0.001), suggesting that greater density does not increase CL if housing and sanitation are adequate. Thus, disease incidence was greatest in neighborhoods that were classified as urban with least vegetation, regardless of population density.
Malaria Sanliurfa province also has experienced increased malaria by roughly an order of magnitude from 1990 (786 cases) to 1995 (6,377). Virtually all cases, however, come from cities and villages outside the Provincial capital. Our recent studies (unpublished ) have shown that Anopheles species are very rare in or near Sanliurfa city, while being patchily abundant elsewhere in the region. We attempted to define the spatial variation of malaria, and determine whether changing environmental characteristics were related to changing disease incidence. Increased irrigation for agriculture during the summer dry season was hypothesized to augment Anopheline mosquito breeding habitat that might be related to malaria risk. We retrospectively evaluated the association between extent of summer vegetation (NDVI from satellite image data) and temporal or spatial variation in malaria incidence. All malaria cases diagnosed and treated in Sanliurfa province from 1980-1995 (n about 30,000) were obtained from provincial Malaria Control Center records. Cases were grouped by village of residence and the 10 district centers. For this macrogeographic comparison, Pathfinder AVHRR data (1km and 8km composites) were analyzed. We used NDVI and a channels 4, 2, and 1 unsupervised land cover classification to identify vegetation and estimate annual percentage change by district. First, three test regions with historically different summer cultivation patterns were compared using MSS images to confirm that summer vegetation distribution and change was proportional to water use. Vegetation estimates also were found to correspond with those from the AVHRR images. Then, AVHRR data were used to compare changes in summer vegetation during 1985-1995 in the 10 districts. A positive association between irrigation and malaria incidence was suggested in areas where irrigation and cultivation have recently begun.
Back to
Leishmaniasis in Turkey Project Introduction