Analyzing Future Spatiotemporal Changes of Precipitation in Yodo River Basin Using Clustering-Based Assessment

Key Points

• The aim is to analyze how precipitation patterns will change over time in the Yodo River Basin using machine learning methods.
• Analyzed current and future climate data from d4PDF5km over 720 years.
• Identified and extracted the top 10 rainfall events annually based on intensity.
• Utilized k-means, self-organizing maps (SOM), and autoencoder clustering methods for analysis.
• Applied statistical tests (chi-squared and proportion tests) for verification.
• Increased concentration of rainfall events in the northern and eastern regions of the basin was noted.
• Shorter duration rainfall events have become more prevalent over time.