FORECASTING EARTHQUAKES WITH THE SAGE MODEL

Every year, millions of seismic tremors shake the earth, but only a few can be anticipated with precision—highlighting the urgent global need for advanced earthquake forecasting tools. This study focused on evaluating the SAGE Model (System dynamics, Bayesian inference, and Geoelectric analysis) as an integrated forecasting framework in the earthquake-prone state of Tennessee, USA. The justification for this research lies in the increasing unpredictability and economic devastation caused by earthquakes, and the limitations of conventional predictive models that fail to capture nonlinear tectonic behavior or incorporate real-time geophysical anomalies. The main objective was to assess whether the SAGE Model could improve forecasting accuracy and early warning potential using secondary data alone. A quantitative methodology was employed, drawing on 210 seismic events from USGS, NOAA, and NASA databases between 2000 and 2024. The system dynamics simulation achieved an 88% match between predicted and historical magnitudes. Bayesian inference models using ETAS principles yielded a 23% improvement in probabilistic accuracy (p < 0.05), and geoelectric anomaly analysis captured signal deviations up to 72 hours before quakes. Overall, the SAGE Model recorded a forecasting correlation coefficient of 0.79, indicating a strong predictive relationship. These results imply that integrating deterministic, probabilistic, and geophysical data enhances earthquake risk models' validity and timeliness. Based on these findings, the study recommends wider institutional adoption of hybrid models for public safety, policy reform to support advanced seismic monitoring, and further machine-learning enhancements for geoelectric filtering.