The variability of grain size metrics in gravel-bed rivers
[EN] Grain size in gravel-bed rivers can be highly variable at the reach scale. Quantitative characterisation of the different sources of this spatial variability has many implications. For example, for how grain size sampling is programmed in gravel-bed rivers, or how this variability is considered...
| Autores: | , |
|---|---|
| Formato: | artículo |
| Estado: | Versión publicada |
| Fecha de publicación: | 2025 |
| País: | España |
| Recursos: | Consejo Superior de Investigaciones Científicas (CSIC) |
| Repositorio: | DIGITAL.CSIC. Repositorio Institucional del CSIC |
| OAI Identifier: | oai:digital.csic.es:10261/399494 |
| Acesso em linha: | http://hdl.handle.net/10261/399494 https://api.elsevier.com/content/abstract/scopus_id/105002259658 |
| Access Level: | acceso abierto |
| Palavra-chave: | Gravel-bed rivers Bedload Fluvial geomorphology Grain-size distribution Grain-size variability |
| Resumo: | [EN] Grain size in gravel-bed rivers can be highly variable at the reach scale. Quantitative characterisation of the different sources of this spatial variability has many implications. For example, for how grain size sampling is programmed in gravel-bed rivers, or how this variability is considered in the propagation of uncertainties associated with the use of models and equations that employ grain size metrics as input parameters (e.g. bedload models, flow friction equations). In this paper we have undertaken a re-analysis of a large database of 462 grain size distributions (GSDs) of gravel-bed rivers compiled from previous studies. In a first step, we explore this database to identify which distribution model best describes the GSD of the collected data. Five different probability distribution models (PDFs) (lognormal, log-logistic, Weibull, Pareto, and log-raised cosine) and one empirical GSD model (Recking's similarity model) are tested. Weibull and Recking's models are the ones showing a best fit to the general shape of the GSD. This first analysis allows us to identify the minimum parameters we need to focus on to adequately characterise grain-size variability, namely the central tendency (e.g., D<inf>50</inf>), the variability around this central tendency, a parameter characterising the tail towards the fine terms (e.g. % of sediment < 2 mm) and another for the coarse tail (e.g. D<inf>84</inf>). In a second step, we do a statistic analysis on those datasets where more than one GSD sample was collected across the same river reach, with the aim of identifying trends in reach-scale variability in these parameters. For each shape-defining and scaling parameters of the different distribution models we report a good fit to a normal or log-normal distribution at the reach-scale. Finally, we show how the information derived from our analysis can be combined with statistical resampling methods for maximising the information that can be obtained from most conventional grain size characterisations in gravel-bed rivers. |
|---|