Correction of deposit ages for inherited ages of charcoal: implications for sediment dynamics inferred from random sampling of deposits on headwater valley floors

Abstract

Inherited age is defined herein as the difference between times of carbon fixation in a material and deposition of that material within sediments from which it is eventually sampled in order to estimate deposit age via radiocarbon dating. Inheritance generally leads to over-estimation of the age by an unknown amount and therefore represents unquantified bias and uncertainty that could potentially lead to erroneous inferences. Inherited ages in charcoal are likely to be larger, and therefore detectable relative to analytic error, where forests are dominated by longer-lived trees, material is stored for longer periods upslope, and downstream post-fire delivery of that material is dominated by mass movements, such as in the near-coastal mountains of northwestern North America. Inherited age distribution functions were estimated from radiocarbon dating of 126 charcoal pieces from 14 stream-bank exposures of debris-flow deposits, fluvial fines, and fluvial gravels along a headwater stream in the southern Oregon Coast Range, USA. In the region, these 3 facies are representative of the nearly continuous coalescing fan-fill complexes blanketing valley floors of headwater streams where the dominant transport mechanism shifts from debris-flow to fluvial. Within each depositional unit, and for each charcoal piece within that unit, convolution of the calibrated age distribution with that of the youngest piece yielded an inherited age distribution for the unit. Fits to the normalized sums of inherited age distributions for units of like facies provided estimates of facies-specific inherited age distribution functions. Finally, convolution of these distribution functions with calibrated deposit age distributions yielded corrections to published valley-floor deposit ages and residence time distributions from nearby similar sites. Residence time distributions were inferred from the normalized sums of distributions of ∼30 deposit ages at each of 4 sites: 2 adjacent valley reaches ∼103 m long and within ∼102 m of 2 tributary confluences. Mean inherited ages from the observed distributions are 666, 688, and 1506 yr for debris-flow deposits, fluvial fines, and fluvial gravels, respectively. On average, correction reduced estimates of individual deposit age means by a factor of 0.71 (0.56–0.94) and increased standard deviations by a factor of 6.1 (0.97–43). Across sites, mean residence times decreased by 24.0% and standard deviations by 12.5% on average. Corrected residence time distributions have thicker tails, as indicated by gamma-distribution fits with smaller shape factors, and these changes are significant relative to the bootstrapped 95% confidence limits representing potential error in the sampling for inherited ages. The ratio of the means of sediment age and residence time ranged from 1.03 to 1.80 across sites before correction and 1.21 to 2.18 after correction, where a value of one implies that probability of evacuation from the “reservoir” comprising valley-floor deposits is independent of time since deposition. Corrected values of this ratio therefore indicate that evacuation favors younger deposits at all sites, whereas uncorrected results implied age-independent evacuation from the more downstream valley reach.