5 Sediment Delivery to Streams: Mechanisms, Volumes, and Timing

Friday, 29 April 2005
8:00 AM - 12:00 PM

LITHOLOGIC INFLUENCE AND EXPERIMENTAL VARIABILITY IN GRAVEL ABRASION: IMPLICATIONS FOR PREDICTING RATES OF DOWNSTREAM FINING OF RIVER BED SEDIMENTS

Paper number: 5-9
Program: THEME SESSIONS
Time: 10:40 AM – 11:00 AM

Authors

Abstract

Recent work on gravel abrasion and differential transport in rivers has neglected rock strength/lithology in parameterizing fining models. Here we report results of tumbling experiments focusing on the effects of general lithologic composition on gravel durability. We consider three separate questions: 1) can rock tensile strength be used to predict differences in bulk fining rates across a wide spectrum of rock types; 2) does variability in rock durability among gravel clasts of the same lithologic composition and initial grain size lead to an evolution of grain size distribution that differs significantly from the predictions of simple fining models; and 3) what is the uncertainty in abrasion coefficients determined by replicate experiments? We used a steel barrel tumbler with excellent control of rotational velocity. Tested rocks were from the Franciscan Formation, Redwood Creek Watershed, California; and from sedimentary and plutonic rocks of the Henry Mountains, Utah. We used the ‘Brazilian' tensile splitting test to measure core strengths of the type and locality tumbled. We determined abrasion coefficients by weighing individual clasts after at least five runs, with typical ‘travel' distances of 4-6 km.

We find a systematic variation in abrasion rate with rock tensile strength and evolution of grain size sensitive to lithology. Abrasion of sedimentary rocks produced a large proportion of sand and silt, presumably due to failure between sediment grains. In contrast, microcrystalline rock, including chert and serpentinite, produced more fine gravel by clast splitting and trended toward a bimodal grain size distribution. Finally, we quantified significant uncertainty in experimental abrasion coefficients for replicate runs with the same initial conditions. Overall, the data are well fit with an exponential law relationship. We conclude that predictive models of downstream fining by abrasion must incorporate lithologic sources of both systematic and stochastic variability in the rate of bulk mass loss and the resulting evolution of grain size distributions.