skip to main content


Title: Detecting Slow Slip Events From Seafloor Pressure Data Using Machine Learning
Abstract

Detecting slow slip events (SSEs) at offshore subduction zones is important to understand the slip behavior on offshore subduction megathrusts, where tsunamis can be generated. The most widely used method to detect SSEs is to measure the vertical seafloor deformation caused by SSEs using seafloor pressure data. However, due to the small signal‐to‐noise ratio and instrumental drift, such detection is very difficult. In this study, we trained a machine learning model using synthetic data to detect SSEs and applied it to real pressure data in New Zealand between 2014 and 2015. Our method detected five events, two of which are confirmed by the onshore GPS records. Besides, our model performs better than the traditional matched filter method. We conclude that machine learning could be used to detect SSEs in real seafloor pressure data. The method can be applied to other regions, especially where near trench GPS is not available.

 
more » « less
NSF-PAR ID:
10455775
Author(s) / Creator(s):
 ;  ;  ;  
Publisher / Repository:
DOI PREFIX: 10.1029
Date Published:
Journal Name:
Geophysical Research Letters
Volume:
47
Issue:
11
ISSN:
0094-8276
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Abstract

    Seafloor pressure sensor data is emerging as a promising approach to resolve vertical displacement of the seafloor in the offshore reaches of subduction zones, particularly in response to slow slip events (SSEs), although such signals are challenging to resolve due to sensor drift and oceanographic signals. Constraining offshore SSE slip distribution is of key importance to understanding earthquake and tsunami hazards posed by subduction zones. We processed seafloor pressure data from January to October 2019 acquired at the Hikurangi subduction zone, offshore New Zealand, to estimate vertical displacement associated with a large SSE that occurred beneath the seafloor array. The experiment included three self‐calibrating sensors designed to remove sensor drift, which, together with ocean general circulation models, were essential to the identification and correction of long‐period ocean variability remaining in the data after applying traditional processing techniques. We estimate that long‐period oceanographic signals that were not synchronous between pressure sensors and reference sites influenced our inferred displacements by 0.3–2.6 cm, suggesting that regionally deployed reference sites alone may not provide sufficient ocean noise correction. After incorporating long‐period ocean variability corrections into the processing, we calculate 1.0–3.3 cm of uplift during the SSE offshore Gisborne at northern Hikurangi, and 1.1–2.7 cm of uplift offshore the Hawke's Bay area at central Hikurangi. Some Hawke Bay displacements detected by pressure sensors near the trench were delayed by 6 weeks compared to the timing of slip onset detected by onshore Global Navigation Satellite System sites, suggesting updip migration of the SSE.

     
    more » « less
  2. null (Ed.)
    International Ocean Discovery Program (IODP) Expedition 372 combines two research topics, slow slip events (SSEs) on subduction faults (IODP Proposal 781A-Full) and actively deforming gas hydrate–bearing landslides (Proposal 841-APL). Our study area on the Hikurangi margin east of New Zealand provides unique locations for addressing both research topics. Gas hydrates have long been suspected of being involved in seafloor failure; not much evidence, however, has been found to date for gas hydrate–related submarine landslides. Solid, icelike gas hydrate in sediment pores is generally thought to increase seafloor strength, as confirmed by a number of laboratory measurements. Dissociation of gas hydrate to water and overpressured gas, on the other hand, may destabilize the seafloor, potentially causing submarine landslides. The Tuaheni Landslide Complex on the Hikurangi margin shows evidence for active, creeping deformation. Intriguingly, the landward edge of creeping coincides with the pinchout of the base of gas hydrate stability (BGHS) on the seafloor. We therefore hypothesize that gas hydrate may be linked to creeping by (1) repeated small-scale sliding at the BGHS, in a variation of the conventional model linking gas hydrates and seafloor failure; (2) overpressure at the BGHS due to a permeability reduction linked to gas hydrates, which may lead to hydrofracturing, weakening the seafloor and allowing transmission of pressure into the gas hydrate stability zone; or (3) icelike viscous deformation of gas hydrates in sediment pores, similar to onshore rock glaciers. The latter two processes imply that gas hydrate itself is involved in creeping, constituting a paradigm shift in relating gas hydrates to submarine slope failure. Alternatively, creeping may not be related to gas hydrates but instead be caused by repeated pressure pulses or linked to earthquake-related liquefaction. We have devised a coring and logging program to test our hypotheses. SSEs at subduction zones are an enigmatic form of creeping fault behavior. At the northern Hikurangi subduction margin (HSM), they are among the best-documented and shallowest on Earth. They recur about every 2 y and may extend close to the trench, where clastic and pelagic sediments about 1.0–1.5 km thick overlie the subducting, seamount-studded Hikurangi Plateau. The northern HSM thus provides an excellent setting to use IODP capabilities to discern the mechanisms behind slow slip fault behavior, as proposed in IODP Proposal 781A-Full. The objectives of Proposal 781A-Full will be implemented across two related IODP expeditions, 372 and 375. Expedition 372 will undertake logging while drilling (LWD) at three sites targeting the upper plate (midslope basin, proposed Site HSM-01A), the frontal thrust (proposed Site HSM-18A), and the subducting section in the trench (proposed Site HSM-05A). Expedition 375 will undertake coring at the same sites, as well as an additional seamount site on the subducting plate, and implement the borehole observatory objectives. The data from each expedition will be shared between both scientific parties. Collectively, the LWD and coring data will be used to (1) characterize the compositional, structural, thermal, and diagenetic state of the incoming plate and the shallow plate boundary fault near the trench, which comprise the protolith and initial conditions for fault zone rock associated with SSEs at greater depth, and (2) characterize the material properties, thermal regime, and stress conditions in the upper plate above the SSE source region. These data will be used during Expedition 375 to guide the installation of CORK observatories at the frontal thrust and in the upper plate above the SSE source to monitor temporal variations in deformation, fluid flow, seismicity, and physical and chemical properties throughout the SSE cycle (Saffer et al., 2017). Together, these data will test a suite of hypotheses about the fundamental mechanics and behavior of SSEs and their relationship to great earthquakes along the subduction interface. 
    more » « less
  3. Abstract

    In subduction zones worldwide, seafloor pressure data are used to observe tectonic deformation, particularly from megathrust earthquakes and slow slip events (SSEs). However, such measurements are also sensitive to oceanographic circulation‐generated pressures over a range of frequencies that conflate with tectonic signals of interest. Using seafloor pressure and temperature data from the Alaska Amphibious Community Seismic Experiment, and sea surface height data from satellite altimetry, we evaluate the efficacy of various seasonal and oceanographic pressure signal proxy corrections and conduct synthetic tests to determine their impact on the timing and amplitude prediction of ramp‐like signals typical of SSEs. We find that subtracting out the first mode of the complex empirical orthogonal functions of the pressure records on either the shelf or slope yields signal root‐mean‐square error (RMS) reductions up to 73% or 80%, respectively. Additional correction with proxies that exploit the depth‐dependent spatial coherence of pressure records provides cumulative variance reductions up to 83% and 93%, respectively. Our detectability tests show that the timing and amplitude of synthetic SSE‐like ramps can be well constrained for ramp amplitudes ≥4 cm on the shelf and ≥2 cm on the slope, using a fully automated detector. The principal limits on detectability are residual abrupt changes in pressure that occur as part of the transition to and from summer to winter conditions but are not adequately characterized by our seasonal corrections, as well as the inability to properly account for instrumental drift, which is not readily separated from the seasonal signal.

     
    more » « less
  4. null (Ed.)
    International Ocean Discovery Program (IODP) Expedition 372 combined two research topics, slow slip events (SSEs) on subduction faults (IODP Proposal 781A-Full) and actively deforming gas hydrate–bearing landslides (IODP Proposal 841-APL). Our study area on the Hikurangi margin, east of the coast of New Zealand, provided unique locations for addressing both research topics. SSEs at subduction zones are an enigmatic form of creeping fault behavior. They typically occur on subduction zones at depths beyond the capabilities of ocean floor drilling. However, at the northern Hikurangi subduction margin they are among the best-documented and shallowest on Earth. Here, SSEs may extend close to the trench, where clastic and pelagic sediments about 1.0–1.5 km thick overlie the subducting, seamount-studded Hikurangi Plateau. Geodetic data show that these SSEs recur about every 2 years and are associated with measurable seafloor displacement. The northern Hikurangi subduction margin thus provides an excellent setting to use IODP capabilities to discern the mechanisms behind slow slip fault behavior. Expedition 372 acquired logging-while-drilling (LWD) data at three subduction-focused sites to depths of 600, 650, and 750 meters below seafloor (mbsf), respectively. These include two sites (U1518 and U1519) above the plate interface fault that experiences SSEs and one site (U1520) in the subducting “inputs” sequence in the Hikurangi Trough, 15 km east of the plate boundary. Overall, we acquired excellent logging data and reached our target depths at two of these sites. Drilling and logging at Site U1520 did not reach the planned depth due to operational time constraints. These logging data will be augmented by coring and borehole observatories planned for IODP Expedition 375. Gas hydrates have long been suspected of being involved in seafloor failure; not much evidence, however, has been found to date for gas hydrate–related submarine landslides. Solid, ice-like gas hydrate in sediment pores is generally thought to increase seafloor strength, as confirmed by a number of laboratory measurements. Dissociation of gas hydrate to water and overpressured gas, on the other hand, may weaken and destabilize sediments, potentially causing submarine landslides. The Tuaheni Landslide Complex (TLC) on the Hikurangi margin shows evidence for active, creeping deformation. Intriguingly, the landward edge of creeping coincides with the pinch-out of the base of gas hydrate stability on the seafloor. We therefore hypothesized that gas hydrate may be linked to creep-like deformation and presented several hypotheses that may link gas hydrates to slow deformation. Alternatively, creeping may not be related to gas hydrates but instead be caused by repeated pressure pulses or linked to earthquake-related liquefaction. Expedition 372 comprised a coring and LWD program to test our landslide hypotheses. Due to weather-related downtime, the gas hydrate-related program was reduced, and we focused on a set of experiments at Site U1517 in the creeping part of the TLC. We conducted a successful LWD and coring program to 205 mbsf, the latter with almost complete recovery, through the TLC and gas hydrate stability zone, followed by temperature and pressure tool deployments. 
    more » « less
  5. null (Ed.)
    Finite-fault models for the 2010 M w 8.8 Maule, Chile earthquake indicate bilateral rupture with large-slip patches located north and south of the epicenter. Previous studies also show that this event features significant slip in the shallow part of the megathrust, which is revealed through correction of the forward tsunami modeling scheme used in tsunami inversions. The presence of shallow slip is consistent with the coseismic seafloor deformation measured off the Maule region adjacent to the trench and confirms that tsunami observations are particularly important for constraining far-offshore slip. Here, we benchmark the method of Optimal Time Alignment (OTA) of the tsunami waveforms in the joint inversion of tsunami (DART and tide-gauges) and geodetic (GPS, InSAR, land-leveling) observations for this event. We test the application of OTA to the tsunami Green’s functions used in a previous inversion. Through a suite of synthetic tests we show that if the bias in the forward model is comprised only of delays in the tsunami signals, the OTA can correct them precisely, independently of the sensors (DART or coastal tide-gauges) and, to the first-order, of the bathymetric model used. The same suite of experiments is repeated for the real case of the 2010 Maule earthquake where, despite the results of the synthetic tests, DARTs are shown to outperform tide-gauges. This gives an indication of the relative weights to be assigned when jointly inverting the two types of data. Moreover, we show that using OTA is preferable to subjectively correcting possible time mismatch of the tsunami waveforms. The results for the source model of the Maule earthquake show that using just the first-order modeling correction introduced by OTA confirms the bilateral rupture pattern around the epicenter, and, most importantly, shifts the inferred northern patch of slip to a shallower position consistent with the slip models obtained by applying more complex physics-based corrections to the tsunami waveforms. This is confirmed by a slip model refined by inverting geodetic and tsunami data complemented with a denser distribution of GPS data nearby the source area. The models obtained with the OTA method are finally benchmarked against the observed seafloor deformation off the Maule region. We find that all of the models using the OTA well predict this offshore coseismic deformation, thus overall, this benchmarking of the OTA method can be considered successful. 
    more » « less