|
(65-05) Spatial extrapolation of local measured PAH immission by use of neural network. Bartel, Martina1, Klein, Roland*,1, 1 University of Trier, Trier, Germany ABSTRACT- The wish and need of having exhaustive data available for purpose of environmental assessment is opponents to the local character of measuring methods of the most environmental monitoring programs. Therefore, spatial extrapolation of data is needed but still a problem. The aim of the study is to present a new approach of spatial extrapolation by the use of neural network. Starting point is the data assessment of the german environmental specimen banking program (ESB) in so-called representative sampling areas belonging to the main ecosystem types of Germany. By the assumption that there exist causal and functional relations between eco-structural components and the chemical "burdening" of ecosystems an spatial extrapolation within the same ecosystem types should be possible. For this purpose, the PAH immission data measured in one-year-old-spruce shoots from sampling areas of the german ESB over more than 15 years, data of Corine Land Cover, climatic, and topographic parameters were analysed. A functional relation between ecosystem type and PAH "burdening" was found. The extrapolation should be done by using neural network to relate PAH data and eco-structural components in the next two months. Key words: spatial extrapolation, PAH immission, German Environmental Specimen Banking Program, Neural Network |
