New Mexico Geological Society Annual Spring Meeting — Abstracts

Bedload transport is notoriously difficult to monitor. Devices used to capture bedload can alter hydraulics or may quickly fill to capacity, while the high intensity flows that most effectively transport bedload can be dangerous and rare. Equations that rely on measurements of excess shear to predict bedload, such as the Meyer-Peter and Müller equation, typically give reasonable results, but may be orders of magnitude off from simultaneous direct bedload flux measurements. Broad grain size distributions, turbulent variability, local supply limitation, and rarified particle transport all contribute to the challenges.

Fluvial seismology provides a potential solution, by monitoring ground vibrations caused by bedload impacts. The direct relation between the process and the physical measurement gives a new, independent metric to include in analyses. We present here a methodology that leverages a moderate number of direct bedload flux measurements to train an empirical equation that predicts bedload flux using seismic power spectral density (PSD) and bed shear estimated from flow depth. Using this method, long-term bedload monitoring can be conducted with simply a bankside seismometer and a water level pressure transducer.

We provide a proof-of-concept example from the Arroyo de los Pinos, an ephemeral tributary of the Rio Grande in New Mexico. At our site, bedload is monitored with automatic Reid-type slot samplers, stage is measured with vented pressure transducers, and seismic data are collected with inexpensive stand-alone seismic nodes. The poorly-sorted and unarmored bed of this ephemeral channel produces very high bedload flux rates during flow events, with measured fluxes up to 16 kg m^-1 s^-1. The conditions of this test site make some aspects of the development of a bedload equation more challenging (e.g., wide range of interacting particle sizes, rapid changes in water stage), while other aspects become trivial (e.g., no armor, no supply limitation). At our site, the trained hydraulic-seismic equation performs better than classic hydraulic equations alone. Yet method testing in contrasting fluvial environments is necessary to determine if other factors that contribute to the failure of hydraulic bedload equations (e.g., armoring and grain packing) can also be overcome.