LONG-TAILED SEDIMENT STORAGE TIME DISTRIBUTION FROM THE MEANDERING POWDER RIVER, MONTANA

As sediment is transported through river corridors, it can experience periods of immobility in alluvial storage reservoirs. The time spent within these storage reservoirs can greatly exceed that of the time spent in transport. Thus, the timescales of sediment transport from source to sink are often dominated by storage timescales, which has implications towards the fate and transport of contaminants, the development of effective watershed restoration plans, and the delivery of sediment signals. Our understanding of the probability distribution of storage times is largely limited to short timescales (on the order of a few hundred years) or to results from physical and numerical models, with a notable lack of field data over geologic timescales. In this study, we quantify the storage time distribution for Powder River from 1998-2013 over a 30-km reach by determining the age distribution of eroded sediment. Our approach integrates surveyed cross-sections, analysis of historical aerial imagery, aerial LiDAR obtained in 2016, geomorphic mapping of lateral-accretion elements, and age control provided by optical stimulated luminescence (OSL) and dendrochronology. Laterally eroded sediment dated using dendrochronology provides the first 150 years of the distribution. The remainder of the distribution is defined from OSL samples from banks with sediment ages ranging from 0 to nearly 6000 years, and erosion volumes determined at cross-sections that have been surveyed repeatedly from 1975 through 2019. The complete storage time distribution is heavy-tailed with mean and median storage times of 2100 and 700 years, respectively. An exponential function does not accurately capture the distribution of storage times, indicating that Powder River’s floodplain does not act as a well-mixed reservoir. Instead, Power law, Pareto, and Weibull distribution functions appear to better capture the nature of the storage times for these alluvial storage reservoirs. All of these heavy-tailed distributions fit our observations reasonably well, suggesting that the distribution of alluvial storage times may be adequately represented by several different mathematical functions.