Application Of High Resolution Remote Sensing Images To Identify Snail Habitats In Lake And Marshland Regions

Schistosomiasis due to Schistosoma japonicum is one of the most serious parasitic causing serious damage to people’s health.In China, it remains endemic in lake and marshland regions in Anhui, Hubei, Hunan, Jiangxi, Jiangsu Provinces and Mountains areas in Sichuan and Yunan Provinces. In spite of the remarkable achievements in the national schistosomiasis control during the past six decades,the new century will indeed be confronted by the great challenges. In this study, TM images and high-resolution CBERS images were complemented to detect potential snail habitats of in the lake and marshland regions based on geographic information system (GIS) and remote sensing (RS) technology, in order to meet the needs of the work under the new situation of schistosomiasis control.In our study, the data of disease, meteorological, hydrological and geographical information from42sample villages across Xingzi County were collected to establish geographic information system database of spatial distribution of schistosomiasis and snails. Then spatial pattern of schistosomiasis risks were maped and factors associated with geographical variation in infection patterns were identified,which will help local decision-makers to develop a more sustainable control strategy.Part I Identification of the snail habitats in lake and marshland regions based on the joint application of normalized difference vegetation and water indicesObjective With the Lansant TM remote sensing images, the snail habitats in lake and marshland regions were identified by using two indices of modified normailzed difference water index(MNDWI) and normailzed difference vegetation index (NDVI). Methods Two TM remote sensing images images of wet season and drought season in Xingzi County region were obtained. MNDWI and NDVI were extracted from the images, respectively and the regions of "water in summer and land in winter" and vegetation coverage were got. By overlaying those two regions, we obtained the areas of potential snail habitats. Finally, the sensitivity and specificity of the recommended methods were assessed, which were also compared to that of the traditional methods to extract the areas of snail habitats. Results The threshold values of MNDWI for the wet and drought seasons was0.34and0.58, respectively and the threshold value for the NDVI of the drought season was0.02. Our method’s sensitivity and specificity was95%and100%, respectively, which are higher than the conventional approach (se=75%,sp=100%).Conclusion The joint application of MNDWI and NDVI is a better method to detect the snail habitats and can be used for the quantitative and automatic snail surveillance.Part Ⅱ Object-oriented snail habitats extraction from high-resolution remote sensing imageObjective To explore the application of remote sensing image classification method based on object-oriented in detecting the snail habitats in lake and marshland regions. Methods Images with high resolutions are first fused from CBERS-02B HR and CCD images of wet season and drought season covered Xingzi County in2008, and then eCognition were used to high-resolution remote sensing image segmentation and classification to extracted potential snail habitats. Results The area of snail habitats of Xingzi County is98.12Km2,which densely located between Beng Lake and Poyang Lake.ConcIusion The snail habitats extracted from high-resolution remote sensing image based on object-oriented classification method with a higher classification precision,which applicable to schistosomiasis prevention and control research studies in complex environment or smaller area.Part III Recognition of high risk snail habitatsObjective Study high risk snail habitats of Xingzi County region to help local decision-makers formulate a more sustainable control strategy. Methods Parasitological data from(?)tandardized surveys were available for36,208locals (aged between6-65years) from42sample villages across the county and used in combination with environmental, meteorological, hydrological and geographical information were collected to establish geographic information system database. Model-based geostatistical model was used to map high risk regions of schistosomiasis and potential snail habitats were overlaied to identify high risk snail habitats. Results Model coefficients shows that distance to snail habitats and wetland, maximum and minimum rainfall, and mean hours of sunshine were significantly negatively associated with schistosomiasis risk while the maximum LST at daytime, the maximum and minimum LST at night, and mean NDVI were significantly positively associated with schistosomiasis.Conclusion According to application of geostatistics to predict the spatial patterns of S. japonicum infection in Xingzi County, which help understand its epidemiological characteristics. More importantly, we identified locations of high risk snail habitats where control efforts should be taken as a strategic priority.