Oxygen depletion in coastal and estuarine waters has been increasing rapidly around the globe over the past several decades, leading to decline in water quality and ecological health. In this study we apply a numerical model to understand how salt wedge dynamics, changes in river flow and temperature together control oxygen depletion in a micro-tidal riverine estuary, the Yarra River estuary, Australia. Coupled physical–biogeochemical models have been previously applied to study how hydrodynamics impact upon seasonal hypoxia; however, their application to relatively shallow, narrow riverine estuaries with highly transient patterns of river inputs and sporadic periods of oxygen depletion has remained challenging, largely due to difficulty in accurately simulating salt wedge dynamics in morphologically complex areas. In this study we overcome this issue through application of a flexible mesh 3-D hydrodynamic–biogeochemical model in order to predict the extent of salt wedge intrusion and consequent patterns of oxygen depletion. The extent of the salt wedge responded quickly to the sporadic riverine flows, with the strength of stratification and vertical density gradients heavily influenced by morphological features corresponding to shallow points in regions of tight curvature ("horseshoe" bends). The spatiotemporal patterns of stratification led to the emergence of two "hot spots" of anoxia, the first downstream of a shallow region of tight curvature and the second downstream of a sill. Whilst these areas corresponded to regions of intense stratification, it was found that antecedent conditions related to the placement of the salt wedge played a major role in the recovery of anoxic regions following episodic high flow events. Furthermore, whilst a threshold salt wedge intrusion was a requirement for oxygen depletion, analysis of the results allowed us to quantify the effect of temperature in determining the overall severity and extent of hypoxia and anoxia. Climate warming scenarios highlighted that oxygen depletion is likely to be exacerbated through changes in flow regimes and warming temperatures; however, the increasing risk of hypoxia and anoxia can be mitigated through management of minimum flow allocations and targeted reductions in organic matter loading. A simple statistical model (<i>R</i><sup>2</sup> > 0.65) is suggested to relate riverine flow and temperature to the extent of estuary-wide anoxia.