Fuzzy model of soil polygons for managing the imprecision

Abstract

At 16:00 UTC on October 23, 1994, 340 kg of perfluoromethylcyclohexane (PMCH) were released into the air from Monterfil in Brittany, France. Air samples were collected at 168 stations in 17 European countries for a period of 90 hours from the start of the release. The European Tracer Experiment (ETEX) was initiated with the aim of collecting data for validating long range transport and dispersion models used for emergency response applications [4, 13]. Another release was made a month later under different meteorological conditions. Although the data have been used in numerous mechanistic atmospheric dispersion studies, only previously was a geostatistical analysis performed in order to provide some basis for a spatial interpolation [5]. Such an analysis applied fractional Brownian motion models in order to summarise the spatial correlation structure in terms of the power exponent of the variogram, which is directly related to the fractal dimension. Because of the distressing nature of the data set, which is highly skewed and required a logarithmic transformation in the Dubois et al study, this paper attempts to apply more robust variography on the raw data in order to extract some order out of the chaos. We use the term "robust" in this paper specifically to refer to the stability of variogram in the presence of a strong direct proportional effect, which characterises the ETEX-1 dataset. © 2009 Springer Berlin Heidelberg.