Enhanced, High-Quality, Semi-Automated Processing for the Engineering Strong-Motion Database (ESM) and Associated Data Products.

ORFEUS (Observatories and Research Facilities for European Seismology; https://www.orfeus-eu.org/) promotes the coordinated development of waveform seismology data services in the Euro-Mediterranean region, including the rapid raw strong‐motion (RRSM; https://orfeus-eu.org/rrsm/) and the engineering strong‐motion (ESM; https://esm-db.eu/) databases and associated web interfaces and webservices. The RRSM uses only on-scale, automatically-processed recordings available from the European Integrated Waveform Data Archive (EIDA; https://www.orfeus-eu.org/data/eida/), whereas the ESM data collects both EIDA and off-line data and results are presently published online after expert revision and approval. 

As seismic stations density continues to increase and the scope of seismic hazard & risk applications evolves to include magnitude/distance/amplitude/instrumentation ranges beyond traditional ‘strong-motion’ seismology, manually reviewed waveform processing in the ESM needs to be strongly supported and eventually replaced by high-quality automated processing. To this end, we introduced ESMpro, a modular Python-based software designed for the updated processing framework of ESM. The software is currently available in a stand-alone Beta version on GitLab (https://shake.mi.ingv.it/esmpro/) and it is in development stage on the ESM database. An enhanced  graphic interface is also in beta testing.

ESMpro automates waveform trimming and filtering, identifies poor-quality data and multiple events, and classifies records into quality categories based on detected features and warnings. This reduces the number of records that require manual revision to a significantly smaller subset. Further, the reproducibility of the dataset means the ESM archive will be ML-ready, allowing the development of emerging strategies for data processing and waveform feature detection based on ML techniques. Additionally, ESMpro has a modular and flexible design that allows integration of different processing schemes. Currently, besides the standard processing adopted in ESM, it also includes the eBasco scheme, specifically tailored to process near-source records featuring fling-step.

Testing conducted on manually processed ESM waveforms demonstrates a strong alignment between automatic and manual data processing, strongly supporting the shift toward fully automated procedures for large-scale data processing that will consolidate RRSM and ESM databases into a single infrastructure in the future.