EXPLORE THE ARCHIVE

David Fee

Publications and source records attributed to David Fee.

2 recordsLinked to original sources

USGS2026

Rapid seismic and infrasound assessment of large landslides: A case study from Denali National Park and Preserve (Alaska)

Large, rapid landslides are a global hazard that can occur in remote, mountainous areas. Eyewitness reports of landslides and satellite imagery can often be limited or delayed, particularly during inclement weather. However, landslide-generated seismic and infrasound (low-frequency atmospheric sound) waves can be remotely detected in near real-time. This information can significantly expedite characterization and possible landslide response activities. Here, we highlight these capabilities using a > 4 million m³ ice–rock avalanche in Denali National Park and Preserve (Alaska). This event was detected via a landslide-specific seismic location and volume estimation algorithm deployed in Alaska, and — notably — by standard earthquake monitoring systems. Following rapid detection of this event, we combined its seismic and infrasound dataset with optical, synthetic aperture radar, and oblique aerial imagery, multitemporal digital elevation models, and a numerical flow model to reconstruct its failure timeline and dynamics. We apply array processing to infrasound signals traveling > 250 km and find that two precursory events occurred minutes prior to the main failure. We use long-period seismic signals to infer the force exerted by the landslide on the Earth and constrain the rheological parameters of our numerical flow simulation with this result and deposit morphology. The main failure produced a steeply-dipping impulsive initial downward force and reached speeds exceeding 60 m/s. This impulsive force generated relatively strong seismic body waves, which contributed to the earthquake system detection. This large, remote Alaska landslide underscores the key value of seismic and infrasound analysis for rapid landslide assessment and motivates efforts to further operationalize these approaches.

Alaska↗

USGS2025

A generalized deep learning model to detect and classify volcano seismicity

Volcano seismicity is often detected and classified based on its spectral properties. However, the wide variety of volcano seismic signals and increasing amounts of data make accurate, consistent, and efficient detection and classification challenging. Machine learning (ML) has proven very effective at detecting and classifying tectonic seismicity, particularly using Convolutional Neural Networks (CNNs) and leveraging labeled datasets from regional seismic networks. Progress has been made applying ML to volcano seismicity, but efforts have typically been focused on a single volcano and are often hampered by the limited availability of training data. We build on the method of Tan et al. [2024] ( 10.1029/2024JB029194 ) to generalize a spectrogram-based CNN termed the VOlcano Infrasound and Seismic Spectrogram Neural Network ( VOISS-Net ) to detect and classify volcano seismicity at any volcano. We use a diverse training dataset of over 270,000 spectrograms from multiple volcanoes: Pavlof, Semisopochnoi, Tanaga, Takawangha, and Redoubt volcanoes\replaced (Alaska, USA); Mt. Etna (Italy); and Kīlauea, Hawai`i (USA). These volcanoes present a wide range of volcano seismic signals, source-receiver distances, and eruption styles. Our generalized VOISS-Net model achieves an accuracy of 87 % on the test set. We apply this model to continuous data from several volcanoes and eruptions included within and outside our training set, and find that multiple types of tremor, explosions, earthquakes, long-period events, and noise are successfully detected and classified. The model occasionally confuses transient signals such as earthquakes and explosions and misclassifies seismicity not included in the training dataset (e.g. teleseismic earthquakes). We envision the generalized VOISS-Net model to be applicable in both research and operational volcano monitoring settings.

Volcanica↗