Investigating effects of initial grain size distribution on rock wear rate and shape change due to abrasion

Abstract

This study investigates how varying initial grain size distributions (GSD) of sediment transported along a river change in shape from angular to round and undergo a reduction in size as mass is lost. Two sets of experiments were done were in order to examine how rocks change in shape, through the use of a laboratory rock tumbling mill which simulates rock abrasion and attrition in natural river environments, following methods based on previous related studies (Arabina & Sklar, 2016;Sklar et al. 2017). The first, consisted of an unsteady state rock tumbling experiment utilizing limestone rock selected according to 4 different initial grain size distributions and tumbled to a target mass loss of 10 percent. The second, involved conducting a longer unsteady state rock tumbling experiment where a single narrow grain size distribution was selected, with a target mass loss of 6 percent during each tumbling run. In the steady state experiment, the mass that was lost from the initial rocks during each tumbling run was subsequently replaced with the addition of fresh angular rocks before reintroduction to the rock tumbler. This approach was taken in order to see how a rock population in a river is affected by the addition of angular rock grains into a population that is progressively rounding. Methods used during both experimental approaches consisted of a combination of manual physical measurements in conjunction with digital measurements utilizing photo image analysis. For our first approach, results showed that initial grain size distribution has a small influence on the efficiency in which rocks erode, with some GSDs such as the narrow GSD traveling shorter distances to reach 10% mass loss compared to the very narrow GSD. This was also reflected in the variation in wear rate/alpha value range associated with each GSD, further supporting that initial GSD has a small influence in the percent of mass lost per km (1/km). Despite variation in wear rate, all GSDs experienced comparable levels of rounding and final mass allocation over the course of the experiment. Most of the initial mass was retained in the original rocks (over 1 g), followed by fine sediment which reflected most of the mass lost, and small fraction of the mass allocated in fragments (<1g). Fragments produced between each GSD varied indicating that initial grain size distribution may affect how rocks break down. For our steady state experiment, we found that wear rates were highest in the first 5 km of distance traveled, followed by a notable reduction in wear rate as travel distance increased. The highest wear rates were observed during the first few km of distance traveled, followed by a stabilized wear rate at further distance traveled such that the subsequent addition of new grains neither increased or decreased the wear rate, leaving it within the same order of magnitude. An increase in fragment production and circularity was observed for the duration of the experiment. This indicates that, despite the addition of new angular rock grains between each tumbling run, the rounding of grains persisted. Results also showed a steady exponential decline in average diameter as travel distance increased. Mass allocation was similar to the unsteady state experiment in which most of the mass distribution remained in original rock particles greater than 1g, followed by fine sediment, and the smallest fraction to fragments. Our results from our steady state experiment align with previous work showing two phase abrasion and the exponential fining of rock diameter with increased distance traveled. The third component of this study aimed to quantify how variations in photo image processing affect shape parameter, in particular circularity, by using photo analysis methods in ImageJ and Photoshop that built off of previous studies (Miller et al., 2014). This involved taking high resolution images of our rock grains with the Lidar Camera of the iPhone 12 Pro and transposing them into binary images to extract rock shape parameters using Dr. Brays MATLAB application CobbleApp. Analysis of photo imaging methods shows that the resolution of rock photos will influence the outline of the rock perimeter and the resulting estimates of circularity.