Abstract
Understanding the teleconnections between hydro-meteorological data and the El Niño–Southern Oscillation cycle (ENSO) is an important step towards developing flood early warning systems. In this study, the concept of mutual information (MI) was applied using marginal and joint information entropy to quantify the linear and non-linear relationship between annual streamflow, extreme precipitation indices over Mekong river basin, and ENSO. We primarily used Pearson correlation as a linear association metric for comparison with mutual information. The analysis was performed at four hydro-meteorological stations located on the mainstream Mekong river basin. It was observed that the nonlinear correlation information is comparatively higher between the large-scale climate index and local hydro-meteorology data in comparison to the traditional linear correlation information. The spatial analysis was carried out using all the grid points in the river basin, which suggests a spatial dependence structure between precipitation extremes and ENSO. Overall, this study suggests that mutual information approach can further detect more meaningful connections between large-scale climate indices and hydro-meteorological variables at different spatio-temporal scales. Application of nonlinear mutual information metric can be an efficient tool to better understand hydro-climatic variables dynamics resulting in improved climate-informed adaptation strategies.
Highlights
For many water resources planning and management studies, reliable preliminary estimates of dependence between two hydroclimatic variables are extremely important
El Niño–Southern Oscillation (ENSO) has a direct influence on rainfall anomalies over the tropical and subtropical regions, only a portion of the variation in the annual flow of rivers located in these regions is associated with ENSO events [31]
This study discusses in detail the possible dependences between different quarters of ENSO indices and hydro-meteorological dataset over Mekong river basin
Summary
For many water resources planning and management studies, reliable preliminary estimates of dependence between two hydroclimatic variables are extremely important. Knowledge of dependence between large-scale climate patterns such as El Niño–Southern Oscillation (ENSO) [1], the Pacific Decadal Oscillation (PDO) [2], and the Atlantic Multi-decadal Oscillation (AMO) [3] with local precipitation, temperature, or streamflow has resulted in improved longer lead-time forecasting models [4,5,6]. Several studies indicated the presence of a significant relationship between large-scale climate phenomena and hydrologic extremes, such as extreme precipitation events [8,9,10], droughts [11,12,13], and floods [14,15]. The predictability of these large-scale climate patterns much in advance is extremely important to improve the design of early warning systems of extreme events [19]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
