Dataset for "Classification of Seismic Events in Mainland China Based on Spectrograms and Model Interpretability"

Dataset DescriptionThis dataset contains seismic waveform data collected from China Digital Seismograph Network for earthquake event classification and machine learning research. The dataset supports the ResWaveQuake model described in the associated publication on time-frequency based earthquake event classification in mainland China.Data OverviewThe dataset spans from January 2013 to May 2024, covering mainland China and adjacent regions. It includes 4,566 seismic events with 132,142 three-component waveform traces, categorized into natural earthquakes (1,158 events), explosions (972 events), collapses/subsidences (740 events), and noise (1,696 events).Data Characteristics and ProcessingAll waveform data are sampled at 100 Hz with three components (BHZ, BHN, BHE) representing vertical, north-south, and east-west directions. Each record contains 200 seconds of data (50 seconds before P-wave arrival plus 150 seconds after) in MiniSEED format, band-pass filtered to 0.1-25 Hz with signal-to-noise ratio ≥ 2. All waveform data underwent standardized preprocessing including instrument response removal, detrending, band-pass filtering, taper window application, and maximum amplitude normalization. Quality assurance measures include retaining only events with ≥3 valid waveform records, systematic quality control using STA/LTA detection for noise samples, consistent preprocessing standards across all regional data, and manual expert annotations for all P-wave arrivals. The dataset is organized using event-level stratified sampling with 80% allocated for training and 20% for testing, while maintaining complete event structure. Data were collected from 650 seismic stations across mainland China, covering epicentral distances from 0-500 km and magnitudes from 0-5.Research ApplicationsThis dataset is designed for earthquake event classification and discrimination, deep learning model development for seismic signal processing, time-frequency analysis of seismic waveforms, automated seismic monitoring system development, and model interpretability studies in seismology.Data Source and AcknowledgmentThe seismic waveform data were obtained from China Digital Seismograph Network. When using this dataset, please acknowledge the data source and cite the associated research paper if applicable.