Akademik Veri Yönetim Sistemi
EGU General Assembly 2025, Vienna, Avusturya, 27 Nisan - 02 Mayıs 2025, ss.1, (Özet Bildiri)
This research uses hybrid machine learning technique to show the regional trends of sea level change along the Australian coastline. The study uses a combination of machine learning algorithms to capture regional heterogeneity in sea level changes by combining data from 43 tidal gauge stations and grid satellite altimetry solutions covering the years 1993–2023. A robust regional evaluation is provided by the hybrid modeling framework, which combines spatial interpolation approaches with algorithms such as Random Forest (RF), Decision Tree (DT), Support Vector Machines (SVM), and Gaussian Process Regression (GPR).
The results show that the hybrid approach is able to effectively capture both temporal trends and spatial patterns in sea level variations, especially when DT is combined with spatial analysis. Regional variations in sea level trends were identified, and the northern regions exhibit slightly different patterns compared to the southern coastal areas. The model explained up to 76% of the variance in the tide gauge data while giving very accurate predictions of regional trends with average rates of 3.55-4.06 mm/year for tide gauge data and 3.13-3.99 mm/year for satellite altimetry data.
A new regional classification is proposed, which is based on the patterns of sea level behavior and delineates well-defined coastal zones characterized by similar features of the trend. This is a very useful regional categorization for local coastal management strategies and also pinpoints areas that need special attention in climate adaptation planning. These results clearly show the importance of regional variation in sea level trends and hybrid machine learning methods for efficient monitoring of coastal environments.