Harnessing the power of graphs to model subglacial sediment transport networks and basal sediment evolution

A better knowledge of sediment transport is needed to understand the distribution of sediments beneath ice, and the signals of cryosphere change that may be detected in glacial sediments deposited offshore. We present here an updated graph-analysis approach to enable further exploration of the sedimentary consequences of hydrological change and allows for a quantitative estimate of water and sediment fluxes in the subglacial environment, and associated basal till evolution and properties, impacting on glacier sliding and hydrogeology processes. The analysis is based on the outputs of physical models, including an ice sheet model output and a subglacial hydrology model output. The approach defines catchment-scale graphs as ‘communities’ of the subglacial hydrology network, from which nested subgraphs are defined optimised to the problem at hand: Such representations greatly reduce the model size and operating in parallel allow efficient development of an ensemble result. The subgraphs may be defined from prior information, such as known catchments, or ad-hoc definition during run-time based on stochastic, probabilistic, or adaptive algorithms. For the subglacial environment the models resolve where sediment is unlikely to be preserved, contrasting with areas of high survivability and deposition. Key properties are defined including thickness of the sediment layer, the grain size of sediment and sediment transport history. These properties may be used to further understand the mechanism of basal sliding and the opportunity for hydrogeological processes. We demonstrate the approach for synthetic examples and catchment-scale studies of real systems.