Lost and found: Maximizing the information from a series of bedload tracer surveys

Abstract

Bedload particle tracking is a technique used to better understand sediment dynamics in rivers. Despite technical advances, tracers may be missed in field surveys. The missed tracers may bias the study results even where recovery rates are high, for example if they are preferentially buried close to the seeding site or transported downstream of the surveyed reach. The goal of the current study is to demonstrate that more information can be extracted from a series of bedload tracer surveys by carefully considering the fate of missing and found tracers and implementing a set of strategies to include the (incomplete) information on sediment displacement metrics. A set of open-source Matlab algorithms collectively called PITtrack are described that perform the calculations. Results from two tracer datasets show that commonly used sediment displacement metrics are sensitive to the inclusion of the missing tracers, even for cases with high recovery rates. Metrics that describe the variance and skewness of the tracers as they disperse are particularly sensitive. The recommended strategy is to include (a) inferred positions of tracers that are missing but unmoved, (b) likely positions of tracers that are missing, moved, and movement can be attributed to a survey period within the uncertain period that meets a dominant flood criterion, and (c) last known positions of tracers considered lost because they go missing and are never found again. Overall the results offer a method to include all available information on missing tracers to better understand sediment dispersion. Future work should be done to assess the classification system for a wider range of field sites and further refine classification based on spatial or other information.

Flow Strength and Bedload Sediment Travel Distance in Gravel Bed Rivers

“…This software ensured that the analysis was completed in a consistent manner for all sites using similar assumptions with regard to the handling of missing tracers. Following the recommendations of MacVicar and Papangelakis (2022), the event‐based metrics ($${p}_{i}$$ and $$\overline{{L}_{{\imath}}}$$) included three classes of tracers: intersect tracers are those found in surveys both before and after the flood, inferred tracers are those missing in the survey after the flood but later re‐found close to their position before the flood and so were immobile during the flood period, and likely tracers are those missing and later re‐found in a new position, but a dominant flood occurred during the uncertain period to which the tracer movement could be ascribed. For this study, a dominant flood was defined as a survey period during which the $$\overline{{L}_{{\imath}}}$$ of the intersect tracers was at least twice that of other survey periods in the uncertain period. …”

“…From these results, stochastic models have been used to show how the combination of rapid transport over the surface of the bed combined with subsequent burial will lead to tracer slowdown as the tracers are mixed with the channel bed (Pelosi et al., 2016). Tracer dispersion models built from field data sets, however, have not consistently accounted for tracer slowdown and are further complicated by the loss of tracers and diminishing recovery rates as the study progresses (MacVicar & Papangelakis, 2022).…”

Comparing Field, Probabilistic, and 2D Numerical Approaches to Assess Gravel Mobility in a Gravel‐Bed River

“…Following MacVicar and Papangelakis (2022), we added inferred tracers to the tracer population in S i , that is, tracers missing in S i that were found in both S i –1 (or S i –2 ) and S i +1 and remained immobile throughout. This substantially increased the number of tracers in the 2014 partial survey (+92 inferred tracers) and 2015 survey (+18 inferred tracers).…”

Modelling Coarse-Sediment Propagation Following Gravel Augmentation: The Case of the Rhône River at Péage-De-Roussillon (France)

“…Table 1 summarizes the recovery statistics by size fraction of the different surveys. To maximize information on lost tracers, we employed the PITtrack2 Matlab code developed by MacVicar and Papangelakis (2022). This tool maximizes the information that can be obtained from a series of particle tracking surveys when the tagged stones are intermittently lost and re-found.…”