Scientific Scenarios of Climate Change for Decadal Forecasts of Earth’s Surface Movements in Germany

Abstract:

Climate change has been proven to exacerbate the ongoing deformations of the Earth's surface in Germany. Also, human activities such as mining, fluid extraction, and reservoir-induced seismicity cause local surface deformations. Therefore, long-term forecasts of Earth's surface movements are needed for infrastructure planning, hazard mitigation, and the sustainable management of natural resources in Germany. By applying Machine Learning (ML) and statistical analyses, we develop scientific scenarios of climate change to forecast surface movements in Germany over the next two decades. Together with Global Navigation Satellite Systems (GNSS), data from five interdisciplinary fields, including the Sun and Moon ephemerides, polar motions, surface loadings, gravity variations, and meteorology, are utilized as features for training ML-based forecast models. Our results indicate that the accuracy of regression ML models reaches millimeter levels, and the decadal forecast models produce fewer than 2% extreme values in the total predictions per year. Based on climate change scenarios, the findings reveal that the average intra-plate motions in Germany will accelerate from ~1.2 mm/yr to ~1.5mm/yr over the next two decades. The annual variations across the 346 GNSS monitoring stations are predicted to increase from 4.7mm to 5.1mm. Surface deformations will be more severe in the southeastern regions and river basins such as the Elbe, Weser, Ems, and Rhine. Significant extensions are expected in the Eifel volcanic region, while notable compressions may occur along the Upper Rhine Graben and the Saxony region in the next twenty years. Additionally, experimental functions showing the statistical distribution of Earth's surface deformation trends in Germany over the next two decades have been proposed. Potentially, the methodology in this study can also be adapted to forecast surface movements related to climate change in polar regions.

Keywords:

Climate change, Surface deformation, Movement forecast, Machine learning, GNSS