River deltas commonly have a heterogeneous substratum of alternating peat, clay and sand deposits. This has important consequences for the river bed development and in particular for scour hole formation. When the substratum consists of an erosion resistant top layer, erosion is retarded. Upon breaking through a resistant top layer and reaching an underlying layer with higher erodibilty, deep scour holes may form within a short amount of time. The unpredictability and fast development of these scour holes makes them difficult to manage, particularly where the stability of dikes and infrastructure is at stake. In this paper we determine how subsurface lithology controls the bed elevation in net incising river branches, particularly focusing on scour hole initiation, growth rate, and direction. For this, the Rhine-Meuse Estuary forms an ideal study site, as over 100 scour holes have been identified in this area, and over 40 years of bed level data and thousands of core descriptions are available. It is shown that the subsurface lithology plays a crucial role in the emergence, shape, and evolution of scour holes. Although most scour holes follow the characteristic exponential development of fast initial growth and slower final growth, strong temporal variations are observed, with sudden growth rates of several meters per year in depth and tens of meters in extent. In addition, we relate the characteristic build-up of the subsurface lithology to specific geometric characteristics of scour holes, like large elongated expanding scour holes or confined scour holes with steep slopes. As river deltas commonly have a heterogeneous substratum and often face channel bed erosion, the observations likely apply to many delta rivers. These findings call for thorough knowledge of the subsurface lithology, as without it, scour hole development is hard to predict and can lead to sudden failures of nearby infrastructure and flood defence works.

Variable flows and fast morphological changes characterize the river system of Bangladesh, which includes the downstream reaches and delta of the Ganges and Brahmaputra rivers, two of the largest rivers in the world. In contrast, fresh water supply around the country largely depends on much smaller distributaries that take off from those large rivers. With the arrival of the dry season and the drop of water levels in the rivers, some of the distributaries become disconnected during several months from their parent rivers because of aggradation at the offtake during the monsoon season. Analysing the evolution of such offtakes from a morphodynamic perspective is fundamental for the definition of effective measures to prevent their closure. However, bed elevation data required to perform such analyses are rarely available, and bathymetric surveys of large rivers are costly and quickly outdated by fast morphological changes. Physics-based numerical models provide a way to fill the gap of unavailable data, while also allowing to explore river morphodynamics beyond the setting of existing rivers.