Analysis and quantification of monitoring networks structure (non-homogeneity or clustering) is an essential task for the characterization of networks' ability to detect phenomena under study. Clustering of monitoring networks has a considerable influence on the quality of univariate data analysis (distributions modelling) and on the consequent spatial predictions and simulations. In the present paper fractal measures/indexes are used in order to quantify monitoring networks. Comparison of the real Swiss Indoor Radon Monitoring Network (SIRMN) with simulated random networks on different validity domains shows the high degree of clustering of the network and the relevancy of the use of validity domains to predict the phenomenon. Multifractal analysis of real monitoring network has been used to reveal the distribution of spatial patterns of radon elevated concentrations.