Susan L. Bilek

Bedload transport dynamics during rapid increases of stage remain poorly constrained, particularly in ephemeral streams where such conditions are common. We combined two cutting-edge monitoring techniques – smartrocks and seismic measurements – to investigate bedload transport patterns during rapid increase of stage in two ephemeral channels with different morphologies. The latter technique was used to characterize bedload activity through the Power Spectral Density (PSD) of recorded seismic signals. Our observations reveal three distinct stages of bedload response: (1) At shallow relative depth (h/d84≤0.9), rapid increase of stage enhanced bed material activity compared to steady flow, with PSD ratios (PSD _{rapid stage rise} / PSD _{steady flow} ) above unity and a higher prevalence of vibrational movement under rapid stage rise conditions relative to steady flow; (2) At intermediate relative depths (0.9≤h/d84≤2.5), the effect of rapid increase of stage on bedload activity diminished; (3) At greater relative depths (h/d84≥2.5), bedload activity is once again enhanced during rapid increase of rise, with both seismic energy and particle motion exceeding values observed under steady flow conditions. The transitions between these stages occurred at similar relative depths in both channels despite their different morphologies, suggesting that channel roughness strongly influences how rapid stage rises affect bedload transport.

In the southwestern United States, flash flooding frequently occurs when localized thunderstorms generate rainfall with intensities that exceed the infiltration capacity of dry soils. During these events, overland flow occurs, coalescing in rills and eventually larger channels where ephemeral flows can create downstream hazards and drive significant geomorphic change. Here, we explore the opportunistic use of distributed acoustic sensing (DAS) to monitor high-frequency strain oscillations along the length of a kilometer-long fiber-optic cable for the purpose of identifying flow in small rills with widths less than 1 m immediately following rainfall events. We demonstrate that for sufficiently high rainfall intensities, rill-flow in hectare-scale basins can be identified at multiple locations simultaneously with DAS. Our observations suggest that DAS can be used, alongside rainfall measurements, to determine the rainfall intensity threshold at which runoff and the resulting rill-flow begins for a particular site. By analyzing a season of DAS data alongside rainfall measurements, we determined a preliminary rainfall intensity threshold of 0.5 mm/min with 5-min duration to generate runoff at our study site. Although we recorded only four runoff events throughout the season, this method highlights the potential future value of DAS for determining hillslope hydrologic properties such as the threshold of runoff generation at relevant spatial scales for modeling and management. We also explore spectral properties of rill-generated DAS signals, finding that rill-flow produces primarily low frequencies (<50 Hz) with some signals observed as high as 200 Hz. This suggests that the rill-driven signals we observed are primarily a result of turbulence, with secondary influence from small (<2 mm) bedload grains impacting the fiber-optic cable and adjacent bed. These findings demonstrate that DAS can be a useful tool for determining high-resolution hydrologic and geomorphic properties across multiple hillslopes with a single instrument.

Quantifying the ocean soundscape is crucial for ocean-based seismoacoustic monitoring; it sets a baseline for the kinds and sizes of signals that can be detected above the background noise. As sea ice recedes, human activity in and around the Arctic Ocean is increasing, elevating sound levels and heightening the urgency of monitoring. Here, seven years of passive acoustic recordings from a National Oceanographic and Atmospheric Administration hydrophone in the Beaufort Sea are analyzed, focusing on frequencies between 1 and 125 Hz, a band of interest for detecting regional earthquakes and similar events. Sound is related to geophysical and anthropogenic sources, and seasonal and intraseasonal variations in the soundscape are examined. Sea ice emerges as a keystone feature of the Arctic Ocean acoustic environment, controlling or contributing to ambient sound levels at all frequencies studied. During low-ice months, sound levels are dominated by wind and sea surface waves, and in ice-covered months, wind-driven ice noise dominates. Seismic air gun surveys are prominent during low-ice periods, with sound levels decreasing with increasing distance and bathymetric complexity along the propagation path. The implications of this baseline soundscape for event detection in the Alaskan Arctic are discussed.

Bedload flux estimation from reach‐averaged hydraulic conditions, while tractable and generally reliable over long integration periods, struggles to capture the intrinsic variability of transport in turbulent flow and the strong influence of local sediment size and morphological heterogeneity. Here we suggest an empirical equation to simplify the relationship between seismic power spectral density (PSD) and bedload flux relative to a full physics‐based seismic model. We posit that adding seismic PSD as a predictor to hydraulics‐based bedload equations improves bedload flux predictions by accounting for short‐ and medium‐term flux variations independent of flow conditions. We present a new calibrated empirical equation combining seismic PSD and excess shear stress to predict bedload flux at high temporal resolution (minute‐scale). The calibrated coefficients for the shear contribution are consistent with existing hydraulics‐based equations (e.g., Meyer‐Peter and Müller) that have been calibrated across a broad range of channels, suggesting that the values for the seismic parameter might also be broadly applicable across channels. In comparison to field data from a sandy‐gravel‐bed alluvial river in New Mexico, USA, the locally‐trained equation reduces scatter in bedload flux predictions relative to methods solely using either seismic PSD or shear stress. We further validate the equation with independent flow events from the same river and from a separate gravel bed channel, the Nahal Eshtemoa in Israel. Notably, the seismic‐hydraulic equation was able to be calibrated on low‐transport data and reliably predict high‐transport data. Likewise, the seismic‐hydraulic equation reduced apparent overestimations of bedload flux by the hydraulics‐based equation during high‐shear conditions.

Although the importance of building meaningful connections between earthquake science research and local communities is widely recognized, the pathways to achieving such collaboration are not always clear, actionable, or well-supported in current dominant scientific culture. We offer strategies and tools to help bridge this gap in earthquake science, drawing on the Collective Impact Model, defined by principles of developing a common agenda, mutually reinforcing activities, centralized support, shared measurement, and continuous communication. We identify lessons learned and offer tangible tools based on the coauthors’ collective experience through the primarily U.S.-based Center for Collective Impact in Earthquake Science. We also discuss the unique role of Minority-Serving Institutions (MSIs) in bridging this gap. The five lessons learned include: 1) allocate time; 2) take stock and focus on small, tangible steps; 3) embrace an interdisciplinary and multidisciplinary team and approach; 4) invest in centralized support; 5) be flexible to ensure that community needs guide research directions. The three tangible tools and practices include: 1) an initial project evaluation tool to develop community-engaged earthquake science research projects, 2) a formative collective impact evaluation rubric tailored to natural hazards and earthquake science for evaluating in-progress projects, and 3) practical recommendations for communication practices. Through collective impact, the earthquake science research community can engage in more inclusive research inquiry that is centered on real-world impact for populations at risk from earthquake hazards.

Bedload transport dynamics during rapid increases of stage remain poorly constrained, particularly in ephemeral streams where such conditions are common. We combined two cutting-edge monitoring techniques – smartrocks and seismic measurements – to investigate bedload transport patterns during rapid increase of stage in two ephemeral channels with different morphologies. The later technique was used to characterize bedload activity through the Power Spectral Density (PSD) of recorded seismic signals. Our observations reveal three distinct stages of bedload response: (1) At shallow relative depth (h /d₈₄ ≤ 0.9), rapid increase of stage enhanced bed material activity compared to steady flow, with PSD ratios (PSD_{rapid stage rise} /PSD_{steady flow}) above unity and a higher prevalence of vibrational movement under rapid stage rise conditions relative to steady flow; (2) At intermediate relative depths (0.9 ≤ h /d₈₄ ≤ 2.5), the rapid increase of stage effect on bedload activity diminished; (3) At greater relative depths (h /d₈₄ ≥ 2.5), bedload activity is once again enhanced during rapid increase of rise, with both seismic energy and particle motion exceeding values observed under steady flow conditions. The transitions between these stages occurred at similar relative depths in both channels despite their different morphologies, suggesting that channel roughness strongly influences how rapid stage rises affect bedload transport.

Bedload transport dynamics during rapid increases of stage remain poorly constrained, particularly in ephemeral streams where such conditions are common. We combined two cutting-edge monitoring techniques – smartrocks and seismic measurements – to investigate bedload transport patterns during rapid increase of stage in two ephemeral channels with different morphologies. The later technique was used to characterize bedload activity through the Power Spectral Density (PSD) of recorded seismic signals. Our observations reveal three distinct stages of bedload response: (1) At shallow relative depth (h /d₈₄ ≤ 0.9), rapid increase of stage enhanced bed material activity compared to steady flow, with PSD ratios (PSD_{rapid stage rise} /PSD_{steady flow}) above unity and a higher prevalence of vibrational movement under rapid stage rise conditions relative to steady flow; (2) At intermediate relative depths (0.9 ≤ h /d₈₄ ≤ 2.5), the rapid increase of stage effect on bedload activity diminished; (3) At greater relative depths (h /d₈₄ ≥ 2.5), bedload activity is once again enhanced during rapid increase of rise, with both seismic energy and particle motion exceeding values observed under steady flow conditions. The transitions between these stages occurred at similar relative depths in both channels despite their different morphologies, suggesting that channel roughness strongly influences how rapid stage rises affect bedload transport.

Bedload transport is a key fluvial process. Yet, it is challenging to measure in natural streams, and continuous time series are rare. Seismic methods show promise for continuous monitoring of bedload transport. However, comparisons with direct bedload measurements are required for calibration and validation, and few such data sets exist. We collected high temporal resolution seismic and bedload flux data from 19 flash flood events that occurred during 5 years at the Arroyo de los Pinos experimental watershed in central New Mexico. We use this data set to investigate the variation of seismic signals with bedload flux across events. To compare seismic and bedload data, we first isolated seismic power‐spectral density (PSD) in the 30–80 Hz frequency range, largely to exclude local noise sources. Next, we computed the linear regression of recorded PSD with bedload flux measurements to determine the robustness of correlation. We found a good correlation between median PSD in the 30–80 Hz range and bedload flux (Spearman's ρ = 0.77). To further investigate variation in the bedload‐PSD relationship, we sieved sediment captured in bedload samplers and determined variations in PSD according to grain size. We found a significant positive correlation between PSD and the median grain size of bedload in motion, but the peak frequencies did not correlate with grain size.

After deploying a seismic network above a karst aquifer that feeds Bear Spring in Minnesota, USA, three injection experiments were conducted, which involved pouring ~10,000 L of water just above an overflow spring to induce and characterize ambient seismic noise from subsurface water flow. A rainstorm event followed the second experiment and caused the overflow spring to start flowing with meters of water level rise, increasing the total spring discharge (perennial and overflow emanations) from ~100 to 300 L/s. Water level, electrical conductivity, and temperature data were used to estimate the conduit hydraulic diameter of ~0.06 m from the second injection experiment and ~0.5 m from the recharge event, which suggested both open channel and full pipe conduit flow. The rainstorm led to the largest seismic signals over a 5 s period while water levels rose, producing ground motion up to ~0.5 mm/s, possible oscillatory behavior, and 1–150 J of radiated seismic energy at frequencies of ~30–50 and 70–90 Hz (similar to signal frequencies when the large water slug from the second injection experiment reached a sump or water-filled passage). The seismic sources originated from multiple directions. The calculations suggest these signals came from conduits with air pockets that were compressed from rising water levels and later released via ventilation pathways. Seismic records of this process enable identification of the change in mode of water level propagation and potential delineation of unknown subsurface pathways, facilitating karst aquifer characterization beyond what is possible with traditional monitoring strategies.