[U-02] 人新世・第四紀の気候および水循環

[U02-08] Five-Century Monthly Streamflow Reconstruction of the Upper Brahmaputra River Reveals Unprecedented Recent Drought

*Wenzheng Nie^1,2、Dunxian She²、Jun Xia²、Kanon Kino¹ (1.University of Tokyo、2.Wuhan University)

キーワード：Streamflow , Reconstruction , Brahmaputra River, Drought-Flood Abrupt Alternation, Monthly

Long-term streamflow records are essential for understanding changes in extreme hydrological events, especially drought and flood. Most existing long-term streamflow data are available at annual resolution, with monthly records being scarce, which limits the ability to reveal the seasonal variability of extreme hydrological events at monthly time scale. In this study, we developed a hierarchical weighting strategy (HWS) based on Mass Balance Regression (MBR) framework to reconstruct monthly streamflow data using tree-ring data. The upper Brahmaputra River (UBR) is chosen as a case study area, and the monthly streamflow data for the last five centuries (1500–2010) is reconstructed using a multi-species tree-ring network. The results showed that HWS significantly enhances the model skill of monthly streamflow reconstruction, with R² improving by 37.3% compared to MBR. Further investigation using the reconstructed monthly streamflow data demonstrates that recent drought and pluvial events are highly unusual within the context of the last five centuries. Our study offers critical context and fundamental data for hydrological risk analysis and water resources management in a perspective of centuries of historical time scale.