1. A geographic information and remote sensing based model for prediction of Oncomelania hupensis habitats in the Poyang Lake area, China
- Author
-
Dong Li, Ming-Gang Chen, Vounatsou Penelope, Hong-Qing Zhu, Anderegg Daniel, Fei Hu, Jia-Gang Guo, Utzinger Jürg, Zhan-ying He, Tanner Marcel, Rong Zhu, and Chun-Li Cao
- Subjects
Oncomelania ,Geographic information system ,Range (biology) ,Veterinary (miscellaneous) ,Snails ,Population ,Disease Vectors ,Environment ,Models, Biological ,Normalized Difference Vegetation Index ,Dry season ,Animals ,Humans ,education ,Demography ,Remote sensing ,education.field_of_study ,biology ,business.industry ,Satellite Communications ,biology.organism_classification ,Infectious Diseases ,Habitat ,Schistosomiasis japonica ,Insect Science ,Oncomelania hupensis ,Geographic Information Systems ,Environmental science ,Parasitology ,business - Abstract
A model was developed using remote sensing and geographic information system technologies for habitat identification of Oncomelania hupensis, the intermediate host snail of Schistosoma japonicum, in the Poyang Lake area, China. In a first step, two multi-temporal Landsat TM 5 satellite images, one from the wet and the second from the dry season, were visually classified into different land-use types. Next, the normalized difference vegetation index was extracted from the images and the tasseled-cap transformation was employed to derive the wetness feature. Our model predicted an estimated 709 km2 of the marshlands in Poyang Lake as potential habitats for O. hupensis. Near-ground temperature measurements in April and August yielded a range of 22.8-24.2 degrees C, and pH values of 6.0-8.5 were derived from existing records. Both climatic features represent suitable breeding conditions for the snails. Preliminary validation of the model at 10 sites around Poyang Lake revealed an excellent accuracy for predicting the presence of O. hupensis. We used the predicted snail habitats as centroids and established buffer zones around them. Villages with an overall prevalence of S. japonicum below 3% were located more than 1200m away from the centroids. Furthermore, a gradient of high-to-low prevalence was observed with increasing distance from the centroids. In conclusion, the model holds promise for identifying high risk areas of schistosomiasis japonica and may become an important tool for the ongoing national schistosomiasis control programme. The model is of particular relevance for schistosome-affected regions that lack accurate surveillance capabilities and are large enough to be detected at most commercially available remote sensing scales.
- Published
- 2005