Unravelling Individual and Joint Effects of Large-Scale Climate Modes and Surface Weather Features on Streamflow in the Murray River, Australia

Abstract

Considerable effort has been expended on finding linkages between regional hydroclimate and large-scale climate variability modes on interannual to decadal time scales. Most of these studies have investigated the influence of modes as a set of independent individuals rather than as a system of possibly interacting variables. Moreover, the impacts of interactions between modes and local-scale weather features are rarely explored or placed in a modelling framework capable of unravelling multivariate complexities in the hydroclimatic system. This study examines the influence of climate modes and surface weather features on monthly streamflow in the Murray River Basin, Australia, over a 124-year period (July 1895–December 2019). A Bayesian network analysis is used to extract the key modes and surface weather features and quantify the strengths and directions of cross-variable relationships. Expanding window and block bootstrap methods are used to ascertain the sensitivity of the model structure and parameter estimates to trends and background hydroclimatic variability over the full study period and a shorter sample (March 1896–February 2004), respectively. It is found that antecedent flow conditions, subtropical ridge intensity and average zonal sea level pressure (SLP) gradient have a direct and robust effect on Murray River flow. The influences exerted by the Indian Ocean Dipole and Southern Annular Mode were sensitive to the period selected for analysis, as was the set of variables that defined the initial state of the hydroclimatic system as characterised by the selected Bayesian networks. These results indicate potential long-term changes in the influence of climate drivers on the Murray River flow. The selected network for the full study period explained some 78% of the total variance in the streamflow series. This result indicates that the network has high explanatory power. These findings could be useful for future applications such as guidance on long-term climate outlooks for the achievement of desired social, environmental and cultural benefits in the Murray River Basin.