Supplementary material to "Subduction Parameters Controlling the Occurrence of Shallow and Deep Slow-Slip Events (SSEs)"

This dataset contains the code, processed data, and notebooks used to reproduce the results of the study “Subduction Parameters Controlling the Occurrence of Shallow and Deep Slow-Slip Events (SSEs)”, which investigates the subduction zone parameters driving the occurrence of SSEs, and the potential for SSEs in global subduction margins using geospatially derived features and machine learning models. The workflow combines GIS-based geospatial preprocessing with machine learning and probabilistic inference. Geospatial data processing was performed using GIS software (e.g., ArcGIS Pro and/or QGIS) and includes the construction of transect-based and slab-derived features (shallow and deep configurations). The resulting processed datasets are provided to ensure reproducibility of the modeling framework. The repository includes Jupyter notebooks covering:1) preprocessing of ML-ready datasets from GIS-derived inputs.2) Jupyter notebooks of the machine learning models3) Jupyter notebooks of the for the inference maps based on the combination of probabilities for SSE from the different approaches. All processed datasets required to reproduce the main results are included. Raw geospatial datasets are not redistributed due to size and/or licensing constraints, but data sources and processing steps are fully documented. All processed datasets required to reproduce the main results are included. Raw geospatial datasets are not redistributed due to size and/or licensing constraints, but data sources and processing steps are fully documented.This dataset enables full reproduction of the machine learning pipeline and inference results presented in the associated publication, without requiring repetition of the GIS preprocessing steps.