协同克立格和标化植被指数在广东省登革热分布特征研究中的应用

目的　探索用标化植被指数 (NDVI)预测登革热 (denguefever,DF)流行和媒介种群的空间分布的可行性。方法　收集广东省 1995年各市县DF发病资料、同期的伊蚊媒介监测资料及广东省县界数字化地图。用ERDAS8.5软件从卫星图像中提取各监测点NDVI。用ArcGIS8.1空间分析软件进行协同克立格 (Co Kriging)分析。 结果　DF流行、种群媒介及NDVI三者的空间分布呈现地域一致性 ;从Co Kriging可见 ,用NDVI对DF发病和BI进行局部估计作图时 ,也呈现了一致性。用NDVI对DF发病、BI同时替代进行三者协同时也取得了一致性的效果。分布图的交叉核验时 ,DF发病、媒介和NDVI的平均预测误差 (MPE)接近于 0 ,估计的方差 (RMSE)与平均标准误 (ASE)都较小 ,且极为接近 ,预测误差的变异程度 (RMSSE)接近于 1。结论　Co Kriging方法是描述DF空间分布特征的较好方法 ,NDVI可以用来作为DF发病和媒介的预测替代

Combined application of Co-Kriging and NDVI for studying the distribution of dengue fever in Guangdong Province

Objective To explore the feasibility of normalized difference vegetation index (NDVI) to forecast spatial distribution of dengue fever prevalence and vector population. Methods Data of dengue fever from all the cities and counties were collected in 1995, aedes vector data and the digitalized map of Guangdong Province divided by county boundaryin the same period of time were collected. NDVI was distilled from those images of remote sensing secondary planet by the ERDAS8.5 software. Co-Kriging analysis was supported by the spatial analysis software of ArcGIS8.1. Results The areas presented consistency, the areas including the transmission intensity of dengue disease, vector population and NDVI spatial distribution map in Guangdong Province in 1995. When the distribution maps were made respectively by NDVI instead of dengue fever, BI, the areas presented consistency, too which was seen form Co-Kriging. The same result was obtained by NDVI instead of dengue fever and BI. When test of cross-validation was used, mean prediction error (MPE) of dengue, its vector and NDVI was close to zero, and root-mean-square error (RMSE) and average standard error (ASE) were all lesser and almost close, and root-mean-square standardized error (RMSSE) was close to 1. Conclusion Co-Kriging is a better method in producing the spatial distribution map of dengue fever and vector, and NDVI may predict dengue fever and its vector.