Spatiotemporal analysis of catalogues declustered by different methods in Northeastern Italy region

The objective of this paper is to analyse the spatiotemporal patterns of the background seismicity identified by Nearest Neighbour (NN) and Stochastic Declustering (SD) methods, in Northeastern Italy. This analysis uses as statistical tools the Allan Factor (AF), Markov Modulated Poisson Process (MMPP), and Morisita Index (MI), to assess the performance of the aforementioned declustering methods in removing the spatiotemporal clustering patterns from the full catalogue. Our results show that the time correlation and the space clustering are reduced but not totally eliminated in the resulting declustered catalogues. Namely, the time clustering structure is significantly reduced at the short timescale, displaying a nearly Poissonian behaviour, but it is not totally eliminated at the intermediate long timescale for NN method, and long timescale for SD method. The MMPP is then confirmed to be a suitable model for describing the background seismicity. The space clustering structure is significantly reduced, though with slightly different proportions, by both the declustering methods; however it is not totally eliminated from the catalogue, due to the natural clustering of seismicity along existing fault systems.