Obtaining elevation of Oncomelania Hupensis habitat based on Google Earth and it’s accuracy evaluation: an example from the Poyang lake region, China

Abstract

AbstractSchistosomiasis japonicum is a major zoonosis that seriously harms human health and affects social and economic development in China. The control of Oncomelania Hupensis, the only intermediate host of schistosome japonicum, is one of the integrated measures for schistosomiasis control in China. Acquiring updated elevation data of snail habitat environment, as well as it’s spatial analysis, play an important role for the risk evaluation and precise control of schistosomiasis transmission and prevalence. Currently, the elevation database of snail habitat environment in schistosomiasis epidemic areas has not been available in the world, which affects the development of research and application work regarding to snail control. Google Earth(GE) can provide massive information related to topography, geomorphology and ground objects of a region due to its indisputable advantages such as wide use, free charge and rapidly updating. In this paper, taking the Poyang lake region as a example, we extracted elevation data of snail-inhabited environment of the lake from GE and established a elevation correction regression model(CRM) for acquiring accurate geospatial elevations, so as to provide a decision-making reference for snail control and risk evaluation of schistosomiasis in China. We developed a GE Application Programming Interface(API) program to extract elevation data from GE, which was compared with the actual elevation data obtained from topographic map of the Poyang Lake bottom. Then, a correction regression model was established and evaluated by 3 index, Mean Absolute Error(MAE), Root Mean Squared Error(RMSE) and Index of Agreement(IOA) for the accuracy of the model. The elevation values extracted from GE in 15086 sample grid points of the lake ranged from 8.5 m to 24.8 m. After the sample points were divided randomly to three groups, the mean elevations of three groups were 13.49 m, 13.52 m and 13.65 m, respectively, with standard deviation ranged from 2.04–2.06. The mean elevation among three groups has no statistic difference (F = 1.536, P = 0.215). A elevation correction regression model was established as y = 6.228 + 0.485×. the evaluation results for the accuracy of the model showed that the MAE and RMSE before correction was 1.28 m and 3.95 m respectively, higher than that after correction, which were 0.74 and 1.30 m correspondingly. The IOA before correction (−0.40)was lower than that after correction(0.34). Google Earth can directly or indirectly get access to massive information related to topography, geomorphology and ground objects due to its indisputable advantages. However, it still needs to be converted into more reliable and accurate data by combining with pre-processing tools. This study used self-developed API program to extract elevation data from GE through precisely locating and improved the accuracy of elevation by using a correction regression model, which can provide reliable data sources for all kinds of spatial data researches and applications.