Decomposition-based teleconnection between monthly streamflow and global climatic oscillation

Abstract

Hydrometeorological teleconnections play a vital role in hydrological processes, as they can reflect the influence of large-scale atmospheric circulations. Understanding and accessing such teleconnections can help reveal the hydrometeorological phenomenon and its associated mechanisms. Therefore, a decomposition-based teleconnection framework (DBTF) is developed to identify the hydrological processes related to the teleconnections under climate change. It consists of three modules, namely, hydrological series decomposition, teleconnection patterns decomposition, and attribution analysis. Specifically, the streamflows are separated into the stationary and nonstationary components using the SSA decomposition technique. Then, the teleconnection patterns (ENSO, AO, PDO, and Sunspots) are decomposed into multi-layer signals by the EEMD decomposition technique. The multiple complex correspondences of different components from hydrological and climate variables are studied by factorial analysis, wavelet coherence transformation, maximum information coefficient, and the partial autocorrelation coefficients. Basins from the humid, semi-humid and semiarid, and arid climate zones are estimated by the framework. The results show that: (1) high-frequency of teleconnection patterns has the highest correlation with the stationary components of streamflow, while low-frequency and trend have the most considerable association with the nonstationary components; (2) the stationary components of streamflow show more pronounced correlations with the teleconnection patterns than the nonstationary components; (3) teleconnection patterns demonstrate more substantial on the decomposed components than the raw streamflow. The contribution is to provide new insight into mechanisms in the effects of hydrometeorological teleconnections on hydrological variations.