A Seismic Approach to Flood Detection and Characterization in Upland Catchments

Abstract

River floods are a major hazard, especially in narrow upland valleys that cover up to 10% of Europe (cf. Supporting Information S1). While Alpine communities are commonly tuned to this hazard, awareness is lower in other upland regions (Schneiderbauer et al., 2021). There, floods are typically thought to be associated with larger rivers (fluvial floods), with day long warning times and inundation damage by high water levels. Under a warming climate with higher atmospheric moisture concentrations and more frequent stationary pressure fields (IPCC, 2018;Nikogosian et al., 2021), European upland regions have already been but will become increasingly exposed to persistent strong precipitation that may cause hazardous floods in smaller catchments (pluvial floods). Communities in such catchments require efficient and reliable flood warning.Existing hydrometric station networks, even dense ones in economically strong countries, are not designed to adequately anticipate, track and characterize rapidly evolving floods in catchments of 10 1 -10 2 km 2 . The majority of stream gauges measure only one metric: water level, and do so once in 15 min. With tens of km between stations on a stream, estimates of flood propagation speed have a time delay of an hour at best and are not available for the valley segment upstream of the first station. Moreover, water levels may well rise by up to a meter between stage measurements, and in-stream gauging stations are inherently prone to intra-flood erosion and deposition effects (Cook et al., 2018) as well as destruction by floodwater or the coarse debris it transports. Often this debris and the accompanying sediment load add substantially to the damage caused by flood water alone.The destructive power of rapidly evolving floods was evident on 14-15 July 2021, after a stationary atmospheric low had discharged up to 250 mm of rain within three days in the Eifel upland region of Germany (Junghänel et al., 2021). The uplifted Eifel plateau is dissected by steep and narrow valleys, feeding the headwaters of several larger rivers. The 900 km 2 catchment of the Ahr River, a tributary of the Rhine, experienced a fast rising flood with peak discharge estimates at 1,200 m 3 /s (Roggenkamp & Herget, 2022) inundating parts of 15 towns,

Anatomy of an Alpine Bedload Transport Event: A Watershed‐Scale Seismic‐Network Perspective

“…This result is positive and gives some confidence in the performance of the FMI since all required parameters were determine beforehand in the field, the inversion was constrained without using measured data, and the seismic‐inverted time‐series was not adjusted by any empirical factor to better fit the SPG data. Given the low cost, ease‐of‐deployment and ease‐of‐maintenance of seismometers (Bakker et al., 2020; K. L. Cook & Dietze, 2022; K. L. Cook et al., 2018; Dietze et al., 2022a, 2022b), seismic data collected at sensor S1 results in a relatively reliable time‐varying estimate of bedload transport, at a factor 100 cheaper than the SPG monitoring for the VdN case. The performance of the seismic inversion approach at S1 is also substantially better than outcomes from capacity‐based bedload transport equations in this environment, with at least an order of magnitude of error typically reported (Ancey, 2020a, 2020b; Recking, 2013; Schneider et al., 2015; Yager et al., 2015), and roughly two orders of magnitude for daily bedload transport predictions in the specific VdN case (Antoniazza et al., 2022).…”

“…Third, field estimates of the nine parameters needed to constrain FMI are also subject to uncertainty. A Monte Carlo simulation was run (Text S7 in Supporting Information S1) on the FMI model parameters (Dietze et al., 2022b) for seismic sensor S1 located nearby the SPG monitoring station, to assess possible changes in bedload transport estimates with changing model parameters. The first three runs (500 random combinations each) investigated the effect of channel width $$W$$, of the GSD parameters ($${D}_{50}$$ and $${\sigma }_{g}$$), and of the ground seismic property parameters ($${v}_{\mathrm{p}0}$$, $$\xi $$, $${K}_{0}$$, and $$\eta $$), on the FMI model outputs.…”

“…To illustrate it, the mean empirical spectrum recorded by each of the 24 seismometers is provided in Text S4 in Supporting Information S1, following Dietze et al. (2022b). It indeed shows that seismic energy in the frequency band specific of bedload transport (25–60 Hz) outweighs the seismic power in the frequency band specific of turbulence (<20 Hz) at each of the 24 seismic sensor locations.…”

“…L. Cook & Dietze, 2022;K. L. Cook et al, 2018;Dietze et al, 2022aDietze et al, , 2022b, seismic data collected at sensor S1 results in a relatively reliable time-varying estimate of bedload transport, at a factor 100 cheaper than the SPG monitoring for the VdN case. The performance of the seismic inversion approach at S1 is also substantially better than outcomes from capacity-based bedload transport equations in this environment, with at least an order of magnitude of error typically reported (Ancey, 2020a(Ancey, , 2020bRecking, 2013;Schneider et al, 2015;Yager et al, 2015), and roughly two orders of magnitude for daily bedload transport predictions in the specific VdN case (Antoniazza et al, 2022).…”

The risk may not be limited to the flooding – Polluted high flood sediments pose a health threat to the unaware public

“…While research [88, [94][95][96] and reports [97] focus on the shocking amounts of debris transported by the rivers, ultimately exacerbating the event by clogging bridges that failed later on, little attention has been paid to the ne fraction of the ood sediments. However, as the transport capacity for ne materials is particularly high, they are spread over larger areas.…”