INTRODUCTION TO THE SPATIAL INTERPOLATION COMPARISON (SIC) 2004 EXERCISE AND PRESENTATION OF THE DATASETS

The

The Spatial Interpolation Comparison (SIC) 2004 exercise was organised during the summer 2004 to assess the current know-how in the field of "automatic mapping".The underlying idea was to explore the way algorithms designed for spatial interpolation can automatically generate maps on the basis of information collected regularly by monitoring networks.Participants to this exercise were invited to use some prior information to design their algorithms and to test them by applying the software code to two given datasets.Estimation errors were used to assess the relative performances of the algorithms proposed.Participants were not only invited to minimize estimation errors but also to design the algorithms so as to render them suitable for decision-support systems used in emergency situations.The data used in this exercise were daily mean values of gamma dose rates measured in Germany.This paper presents the exercise and the data used more in detail.

2. 3
Figure 3Frequency histogram of the first data used for estimations and theoretical normal distribution (line).

Figure 4
Figure 4Frequency histogram of the subset used as the "joker" data set used for estimations and theoretical normal distribution (line).
Figure 5 3D model of the dispersion process displayed on the top of the background radiation map (vertical scale in nSv/h).

Table 1 .
The smooth curve depicts the normal distribution.Statistics for the 10 datasets used to train the algorithm used in SIC2004.Measurement units are nSv/h.
Summary statistics of the first dataset used for the SIC2004 exercise.