We introduce an algorithm for declustering earthquake catalogs based on the nearest‐neighbor analysis of seismicity. The algorithm discriminates between background and clustered events by random thinning that removes events according to a space‐varying threshold. The threshold is estimated using randomized‐reshuffled catalogs that are stationary, have independent space and time components, and preserve the space distribution of the original catalog. Analysis of catalog produced by the Epidemic Type Aftershock Sequence model demonstrates that the algorithm correctly classifies over 80% of background and clustered events, correctly reconstructs the stationary and space‐dependent background intensity, and shows high stability with respect to random realizations (over 75% of events have the same estimated type in over 90% of random realizations). The declustering algorithm is applied to the global Northern California Earthquake Data Center catalog with magnitudes
Seismology is witnessing explosive growth in the diversity and scale of earthquake catalogs. A key motivation for this community effort is that more data should translate into better earthquake forecasts. Such improvements are yet to be seen. Here, we introduce the Recurrent Earthquake foreCAST (RECAST), a deep‐learning model based on recent developments in neural temporal point processes. The model enables access to a greater volume and diversity of earthquake observations, overcoming the theoretical and computational limitations of traditional approaches. We benchmark against a temporal Epidemic Type Aftershock Sequence model. Tests on synthetic data suggest that with a modest‐sized data set, RECAST accurately models earthquake‐like point processes directly from cataloged data. Tests on earthquake catalogs in Southern California indicate improved fit and forecast accuracy compared to our benchmark when the training set is sufficiently long (>104events). The basic components in RECAST add flexibility and scalability for earthquake forecasting without sacrificing performance.
more » « less- Award ID(s):
- 1761987
- NSF-PAR ID:
- 10462172
- Publisher / Repository:
- DOI PREFIX: 10.1029
- Date Published:
- Journal Name:
- Geophysical Research Letters
- Volume:
- 50
- Issue:
- 17
- ISSN:
- 0094-8276
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
more » « less
Abstract m ≥ 4 during 2000–2015; a Southern California catalog withm ≥ 2.5, 3.5 during 1981–2017; an area around the 1992 Landers rupture zone withm ≥ 0.0 during 1981–2015; and the Parkfield segment of San Andreas fault withm ≥ 1.0 during 1984–2014. The null hypotheses of stationarity and space‐time independence are not rejected by several tests applied to the estimated background events of the global and Southern California catalogs with magnitude ranges Δm < 4. However, both hypotheses are rejected for catalogs with larger range of magnitudes Δm > 4. The deviations from the nulls are mainly due to local temporal fluctuations of seismicity and activity switching among subregions; they can be traced back to the original catalogs and represent genuine features of background seismicity. -
more » « less
Abstract Seismic phase association is a fundamental task in seismology that pertains to linking together phase detections on different sensors that originate from a common earthquake. It is widely employed to detect earthquakes on permanent and temporary seismic networks and underlies most seismicity catalogs produced around the world. This task can be challenging because the number of sources is unknown, events frequently overlap in time, or can occur simultaneously in different parts of a network. We present PhaseLink, a framework based on recent advances in deep learning for grid‐free earthquake phase association. Our approach learns to link phases together that share a common origin and is trained entirely on millions of synthetic sequences of
P andS wave arrival times generated using a 1‐D velocity model. Our approach is simple to implement for any tectonic regime, suitable for real‐time processing, and can naturally incorporate errors in arrival time picks. Rather than tuning a set of ad hoc hyperparameters to improve performance, PhaseLink can be improved by simply adding examples of problematic cases to the training data set. We demonstrate the state‐of‐the‐art performance of PhaseLink on a challenging sequence from southern California and synthesized sequences from Japan designed to test the point at which the method fails. For the examined data sets, PhaseLink can precisely associate phases to events that occur only ∼12 s apart in origin time. This approach is expected to improve the resolution of seismicity catalogs, add stability to real‐time seismic monitoring, and streamline automated processing of large seismic data sets. -
more » « less
Abstract Unrecognized deterioration of COVID-19 patients can lead to high morbidity and mortality. Most existing deterioration prediction models require a large number of clinical information, typically collected in hospital settings, such as medical images or comprehensive laboratory tests. This is infeasible for telehealth solutions and highlights a gap in deterioration prediction models based on minimal data, which can be recorded at a large scale in any clinic, nursing home, or even at the patient’s home. In this study, we develop and compare two prognostic models that predict if a patient will experience deterioration in the forthcoming 3 to 24 h. The models sequentially process routine triadic vital signs: (a) oxygen saturation, (b) heart rate, and (c) temperature. These models are also provided with basic patient information, including sex, age, vaccination status, vaccination date, and status of obesity, hypertension, or diabetes. The difference between the two models is the way that the temporal dynamics of the vital signs are processed. Model #1 utilizes a temporally-dilated version of the Long-Short Term Memory model (LSTM) for temporal processes, and Model #2 utilizes a residual temporal convolutional network (TCN) for this purpose. We train and evaluate the models using data collected from 37,006 COVID-19 patients at NYU Langone Health in New York, USA. The convolution-based model outperforms the LSTM based model, achieving a high AUROC of 0.8844–0.9336 for 3 to 24 h deterioration prediction on a held-out test set. We also conduct occlusion experiments to evaluate the importance of each input feature, which reveals the significance of continuously monitoring the variation of the vital signs. Our results show the prospect for accurate deterioration forecast using a minimum feature set that can be relatively easily obtained using wearable devices and self-reported patient information.
-
more » « less
Volcanic earthquake catalogs are an essential data product used to interpret subsurface volcanic activity and forecast eruptions. Advances in detection techniques (e.g., matched-filtering, machine learning) and relative relocation tools have improved catalog completeness and refined event locations. However, most volcano observatories have yet to incorporate these techniques into their catalog-building workflows. This is due in part to complexities in operationalizing, automating, and calibrating these techniques in a satisfactory way for disparate volcano networks and their varied seismicity. In an effort to streamline the integration of catalog-enhancing tools at the Alaska Volcano Observatory (AVO), we have integrated four popular open-source tools: REDPy, EQcorrscan, HypoDD, and GrowClust. The combination of these tools offers the capability of adding seismic event detections and relocating events in a single workflow. The workflow relies on a combination of standard triggering and cross-correlation clustering (REDPy) to consolidate representative templates used in matched-filtering (EQcorrscan). The templates and their detections are then relocated using the differential time methods provided by HypoDD and/or GrowClust. Our workflow also provides codes to incorporate campaign data at appropriate junctures, and calculate magnitude and frequency index for valid events. We apply this workflow to three datasets: the 2012–2013 seismic swarm sequence at Mammoth Mountain (California), the 2009 eruption of Redoubt Volcano (Alaska), and the 2006 eruption of Augustine Volcano (Alaska); and compare our results with previous studies at each volcano. In general, our workflow provides a significant increase in the number of events and improved locations, and we relate the event clusters and temporal progressions to relevant volcanic activity. We also discuss workflow implementation best practices, particularly in applying these tools to sparse volcano seismic networks. We envision that our workflow and the datasets presented here will be useful for detailed volcano analyses in monitoring and research efforts.
-
more » « less
Abstract The effects of strong ground shaking on hillslope stability can persist for many years after a large earthquake, leading to an increase in the rates of post earthquake land sliding. The factors that control the rate of post‐earthquake land sliding are poorly constrained, hindering our ability to reliably forecast how landscapes and landslide hazards and risk evolve. To address this, we use a unique data set comprising high‐resolution terrestrial laser scans and airborne lidar captured during and after the 2010–2011 Canterbury Earthquake Sequence, New Zealand. This earthquake sequence triggered thousands of rock falls, and rock and debris avalanches (collectively referred to as “rockfall”), resulting in loss‐of‐life and damage to residential dwellings, commercial buildings and other infrastructure in the Port Hills of Christchurch, New Zealand. This unique data set spans 5 years and includes five significant earthquakes. We used these data to (a) quantify the regional‐scale “rockfall” rates in response to these earthquakes and the postearthquake decay in rockfall rates with time; and (b) investigate the site‐specific factors controlling the location of seismic and nonseismic rockfalls using frequency ratios and logistic regression techniques. We found that rockfall rates increased significantly in response to the initial earthquake that generated the strongest shaking in the sequence—The MW6.2 22 February 2011 event—Compared to the long‐term background rates derived from the dating of pre‐2010 talus piles at the toe of the slopes. Non seismic rockfall rates also increased immediately after the 22 February 2011 earthquake and decayed with time following a power‐law trend. About 50% of the decay back to the pre‐earthquake rockfall rates occurred within 1–5 years after the 22 February 2011 earthquake. Our results show that the short‐term decay in rockfall rates over time, after the initial earthquake, was attributed to the subsequent erosion of seismically damaged rock mass materials caused by environmental processes such as rain. For earthquake‐induced rockfall at the regional‐scale, the peak ground accelerations is the most significant variable in forecasting rockfall volume, followed by the relative height above the base of the slope. For both earthquake and non‐seismic conditions at the site‐specific scale, the probability of rockfall increases when the adjacent areas have failed previously, indicating that accrued damage preconditions localized areas of the slope for subsequent failure. Such preconditioning is a crucial factor driving subsequent rockfalls; that is, future rockfalls are likely to cluster near areas that failed in the past.
