Development of a mobile sensor for robust assessment of river bed grain forces

The forces experienced by sediment grains at entrainment and during transport, and those exerted on river beds, are significant for the development of river systems and landscape evolution. The assessment of local grain forces has been approached using two different methodologies. The first approach uses static impact sensors at points or cross-sections to measure velocity and/or acceleration. A second approach uses mobile natural or artificial 'smart' pebbles instrumented with inertia micro-sensors for directly measuring the local forces experienced by individual grains. The two approaches have yielded significantly different magnitudes of impact forces. Static sensors (piezoelectric plates connected to accelerometers) temporally smooth the impacts from several grains and infrequently detect the higher forces (up to ×100g) generated by direct single-grain impacts. The second method is currently unable to record the full range of impacts in real rivers due to the low measurement range of the deployed inertia sensors (×3g). Laboratory applications have required only low-range accelerometers, so excluding the magnitude of natural impacts from the design criteria. Here we present the first results from the development of a mobile sensor, designed for the purpose of measuring local grain-forces in a natural riverbed. We present two sets of measurements. The first group presents the calibration of a wide range micro-accelerometer from a set of vertical drop experiments (gravitational acceleration) and further experiments on a shaking table moving with pre-defined acceleration. The second group of measurements are from incipient motion experiments performed in a 9m x0.9m flume (slope 0.001 to 0.018) under steadily increasing discharge. Initially the spherical sensor grain was placed on an artificial surface of hemispheres of identical diameter to the sensor (111mm). Incipient motion was assessed under both whole and half-diameter exposure for each slope. Subsequently, the sensor was placed on a bed of natural gravel of equivalent mean diameter under low slope conditions (0.001). Incipient motion was monitored over a fully covered stable bed and over a partially covered bed developed over an artificial surface constructed to simulate a natural bedrock surface. Statistical analysis of the results describes the relationship between flow conditions and pre-entrainment grain vibration and the acceleration threshold for incipient motion. Finally we perform a preliminary analysis to assess the degree of dependency of the same threshold on the different degrees of alluvial coverage of a river bed and so illustrate the potential to evaluate existing models describing grain entrainment and transport.