Probabilistic Completeness Assessment of the Past 30 Years of Seismic Monitoring in northeastern Italy

International audience; We investigate detection probabilities and recording completeness of the seismic network in northeastern Italy, operated by the OGS (Istituto Nazionale di Oceanografia e Geofisica Sperimentale) during the years 1977-2007, using the Probability-based Magnitude of Completeness (PMC) method by Schorlemmer & Woessner (2008). Completeness of the dataset is varying in space and time due to the evolution of network geometry, instrumental characteristics, and monitoring and processing strategies over time: this is a common feature for all the regional and national instrumental catalogues that should be properly accounted for in seismicity rates evaluations. For the first time, we quantify with the new PMC method detection probabilities of a regional network operating since the late 70's, also including paper records. PMC is based on empirical data and requires the earthquake catalog containing phase picks, the station locations and on-/off-times, and the attenuation relation used to compute local magnitudes. In the 30-year period, we identified four time windows, roughly corresponding to the main changes in the acquisition system. We reconstructed on-/off-times by merging the available information on instruments with the analysis of inter-pick times at each station. We revised the relationship between duration and local magnitudes as the calibration of duration magnitude demonstrated to be not homogeneous among the different acquisition systems. Moreover, we introduced some constraints on the application of the PMC method to improve its performance for networks with sparse data and show the role of missing events and of temporary networks on regional completeness maps. The results demonstrate that the OGS network is detecting earthquakes completely down to magnitude 1.5 for a large part of the Friuli-Venezia-Giulia region since the earliest stages of its functionality; the OGS instrumental catalogue is therefore the most precise and complete dataset available for this area. Our analysis shows that the change from analog to digital systems does not necessarily correspond to an improvement in detection capabilities. The computed time series of completeness maps are available and should be considered for any seismicity study involving data from the OGS network.