Identifying variabilities in grain-scale sediment transport thresholds and implications for river stability and form

Understanding the processes which define the geometry of gravel bedded river channels is important in predicting their future geomorphological adjustment and addressing contemporary river management challenges in response to climate change (e.g. flood risk). The geometry of stable river channels and the thresholds of channel adjustment can be predicted using an understanding of the mobility of bed material, which is typically quantified using the dimensionless critical shear stress parameter, '*c. In most applications, '*c is assumed spatially and temporally constant for a channel reach, which fails to consider the influence of bed material composition and arrangement which are known to vary within a channel and through time. This thesis quantifies the uncertainty of '*c assumptions in an applied river management context, explores novel approaches towards improving estimates of '*c at a grain-scale using force-balance approaches, and investigates the spatial and temporal drivers of grain arrangement and mobility.Fundamental hydraulic geometry theory was used to evaluate the uncertainties associated with basic '*c assumptions in river management contexts. Predictions of channel (dis)equilibrium were obtained by quantifying deviations from expected equilibrium geometry for 45 upland UK channel reaches, and were largely consistent with observed evidence of recent channel adjustment and could therefore give predictive insight into hotspots of geomorphic risk. However, predictions of stable channel geometry were sensitive to small variations in '*c. To incorporate grain-scale mobility controls such as grain protrusion into estimates of '*c, force-balance models were tested; these combined grain microtopography data quantified using Structure-from-Motion topography with measurements of particle resistance force (FR/FW) taken in the field. Median estimates of '*c generated from the force-balance approaches (0.05 - 0.15) were consistent with typical values of '*c (0.02 - 0.1), but were advantageous because they evaluated a distribution of likely '*c values for localised patches, rather than a single value for an entire reach. The force-balance methodology could also be automated at no real detriment to the '*c estimates, albeit further validation is required. To evaluate the spatial controls on bed mobility across different rivers, distributions of FR/FW were compared against metrics of bed texture, including imbrication, sorting and surface roughness. It was found that all sampled beds could be categorised into one of four broad bed types, based on combinations of surface texture and grain shape. Each bed type yielded different FR/FW measurements. For example, well sorted, imbricated surfaces had high FR/FW and were therefore least mobile, whereas poorly sorted, non-imbricated surfaces had low FR/FW and were therefore most mobile. Given that bed texture was an important control on grain mobility, the influence of flow history on the temporal variability of grain arrangement properties was further investigated by repeat sampling an exposed channel bed (River Alport, Peak District) over a nine-month monitoring period. Observations revealed evidence for the self-reorganisation of bed material following large destructive events, following periods of inundation. This gave insight into the timescales of bed recovery, for example, highlighting the role of flood sequencing on the cumulative destruction or restabilisation of river gravels, depending on the timing of flows of varying magnitudes.These findings are important in river management applications, because they can help to develop estimates of '*c which can better reflect the variability of natural river beds. Further applications of the datasets and approaches outlined in this thesis should be used to develop guidelines which can be used by river managers to adjust '*c based on easy-to collect qualitative channel bed observations and/or flow history records. This would help to improve evaluations of river channel stability and form over large spatial scales.