Seismic moment tensor results in support of "Observational evidence of a very-low-frequency earthquake (Mw 3.8) leading to an earthquake (Mw 4.2): Minto Flats strike-slip fault zone, central Alaska"

Seismic moment tensor results in support of “Observational evidence of a verylow-frequency earthquake (Mw3.8) leading to an earthquake (Mw4.2): Minto Flats strike-slip fault zone, central Alaska” Amanda McPherson December 25, 2025 Attribution: This collection is prepared for a manuscript in review (McPherson et al., 2026). If you use these results, please cite the published paper, in addition to this Zenodo collection. Overview This collection includes figures and results from a catalog of 5 estimated moment tensors for events in Minto Flats fault zone. The results were obtained using MTUQ (Thurin et al., 2025), in addition to FK (Zhu and Rivera, 2002) and Obspy (Krischer et al., 2015). All source mechanisms were assumed to be double couples. For the two 2025 events (earthquake and VLFE), we estimate moment tensors for a set of 55 different bandpasses, and the full results are provided here. Description of files 1. McPherson2026_main.pdf (this file) 2. McPherson2026_mech.txt Text file summarizing the 5 moment tensors. Details can be found within the header lines, which also refer to Kanamori (1977); Silver and Jordan (1982); Tape and Tape (2012). 3. Waveform_fits.zip A zipped set of waveform fits figures for the 5 moment tensors. A subset of waveform fits for one moment tensor is shown in Figure 1; the features of the plot are described in detail below. 4. MTUQ_weights.zip 3 MTUQ weight files for P waves, Rayleigh waves, and Love waves. Both events in 2025 use the same weights file. 5. McPherson2026bp_2025eq_mech.txt and McPherson2026bp_2025vlfe_mech.txt Two text files, each containing 55 moment tensors. One set is for the 2025 earthquake, the other is for the 2025 VLFE. The 55 solutions are for the bandpasses depicted in Figure 4. 6. Bandpass_iterations Zipped file of all output figures for the bandpass run of 2 ×55 moment tensors. 1 Details for figures of waveform fits An example waveform fit figure is shown in Figure 1. Black are observed waveforms; red are synthetic waveforms. The waveforms are fit separately within five time windows: two for P waves and three for surface waves. The P-wave time windows are expressed in the vertical (Z) and radial (R) components. The surface wave time windows are represented in the Rayleigh wave (Z and R) and Love wave (transverse, T) windows. The beachball in the top left represents the best solution, i.e., the global minimum of the misfit function. The beachball is plotted as a lower-hemisphere projection (standard seismological convention) of the moment tensor. Here is a header for the example event in Figure 1: 2025-08-20T20:65:27 64.74°N 149.08°W Mw 4.15 Depth 18.0 km model: tactmod solver: FK misfit (L2): 7.544e-01 body waves: 1.5 - 4.0 s (15.0 s), Rayleigh: 16.0 - 30.0 s (120.0 s), Love: 16.0 - 30.0 s (120.0 s) strike dip slip: 210 52 -3, lune coords γ δ : 0 0, N-Np-Ns: 43-41-42 The four header lines are as follows: 1. 2025-08-20T20:65:27 64.74°N 149.08°W Mw 4.15 Depth 18.0 km Date and origin time 2022-01-15T04:14:45, longitude and latitude of the event (−149.08,64.74), best-fitting Mw, and depth. In this case, the depth was obtained from a grid search (Figure 3b). 2. model: tactmod solver: FK misfit (L2): 7.544e-01 Reference model in which the Green’s functions are computed. Solver denotes the numerical solver used to compute the Green’s functions (could be Specfem, FK, etc.) or indicates if the Green’s functions were downloaded from the Syngine data service (Krischer et al., 2017), which build upon Instaseis (van Driel et al., 2015) and the spectral-element solver AxiSEM (Nissen-Meyer et al., 2014). L2 denotes the misfit measures (L1, L2 and hybrid norms are implemented) and the associated misfit value. A new feature since the publication of (Thurin et al., 2025) is that the misfit shown here is normalized by the data norm, making this a unitless value. In this case, the Earth model is the 1D model tactmod (Beaudoin et al., 1992; Ratchkovski and Hansen, 2002), the software FK was used to compute the Green’s functions (Zhu and Rivera, 2002), and the L2 waveform misfit is 0.7544. 3. body waves: 1.5 - 4.0 s (15.0 s), Rayleigh: 16.0 - 30.0 s (120.0 s), Love: 16.0 - 30.0 s (120.0 s) The body waves, if used in the inversions, were filtered 1.5–4.0 s and have a 15 s time window, the Rayleigh waves were filtered 16–30 s, and the Love waves were filtered 16–30 s s. Both Rayleigh and Love wave data time windows are 120 s long. 4. strike dip slip: 210 52 -3, lune coords γ δ : 0 0, N-Np-Ns: 43-41-42 The parameters of the best-fitting moment tensor. The first three values are strike, dip, and rake (e.g., slip) angles, which define the orientation of the moment tensor. The next two value are lune longitude γand lune latitude δ, which define the moment tensor source 2 type (Tape and Tape, 2012). For our events, we constrain the solution to be double-couple (γ= 0◦, δ = 0◦). The station rows are sorted by epicentral distance, with AK.NEA2 being the closest station (17 km). The numbers below each station label are: 1. source–station epicentral distance, km 2. station azimuth, in degrees The three numbers below each pair of waveforms are 1. the cross-correlation time shift ∆T=Tobs −Tsyn′required for matching the static-shifted synthetics s′(t) with the data u(t). A positive time-shift means that the static-shiftedsynthetics arrive earlier than the data and that the assumed velocity model is faster than the actual earth structure. The synthetic seismograms shown in the example waveform plot (Figure 1) have an allowable time shift of ±3 s for the body waves, and ±15 s for the surfaces waves. Most of the synthetic surface waves only need ±3 s of shift. 2. the maximum cross-correlation percentage between u(t) and s′(t−∆T) 3. the percentage of the total misfit this window contributes Figures of big beachballs Four beachball plots are shown in Figure 2. (The plot for the 2025 earthquake for the bandpass 12.5–27.5 s surface-wave-only inversion is not shown.) Figures of grid searches over depth Four depth grid searches are shown in Figure 3. For the 2025 VLFE, the best-fitting depth is 16 km, but we opted to use the earthquake depth of 18 km. No depth searches were performed for the fixed-bandpass (12.5–27.5 s) runs of the 2025 VLFE and earthquake (fixed depth 18 km). The depth increment is 1 km, and the white arrow marks the depth obtained from the moment tensor inversion. One moment tensor is missing in each plot at 11 km depth. The 1D model used to make the inversions (tactmod) has a layer boundary at 11 km, and thus no inversion can be done at that depth. The event magnitude is free to change at each depth, and generally increases with depth for the best-fitting solution, as we might expect. 3 Acknowledgments This project was supported by the National Science Foundation grant EAR 2342129 and the TREMOR project funded by Air Force Research Laboratory contract FA9453-24-9-0001. References Beaudoin, B. C., G. S. Fuis, W. D. Mooney, W. J. Nokleberg, and N. I. Christensen, Thin, low-velocity crust beneath the southern Yukon-Tanana terrane, east central Alaska: Results from Trans-Alaska Crustal Transect refraction/wide-angle reflection data, J. Geophys. Res., 97(B2), 1921–1942, 1992. Kanamori, H., The energy release in great earthquakes, J. Geophys. Res.,82, 2981–2987, 1977. Krischer, L., T. Mengies, R. Barsch, M. Beyreuther, T. Lecocq, C. Caudron, and J. Wassermann, ObsPy: a bridge for seismology into the scientific Python ecosystem, Computational Science & Discovery,8(1), 014003, doi:10.1088/1749-4699/8/1/014003, 2015. Krischer, L., A. R. Hutko, M. van Driel, S. St¨ahler, M. Bahavar, C. Trabant, and T. Nissen-Meyer, On-demand custom broadband synthetic seismograms, Seismol. Res. Lett.,88 (4), 1127–1140, doi:10.1785/0220160210, 2017. McPherson, A. M., C. Tape, and Y. Kaneko, Observational evidence of a very-low-frequency earthquake (Mw3.8) leading to an earthquake (Mw4.2): Minto Flats strike-slip fault zone, central Alaska, Geophys. Res. Lett. (in prep.), 2026. Nissen-Meyer, T., M. van Driel, S. C. St¨ahler, K. Hosseini, S. Hempel, L. Auer, A. Colombi, and A. Fournier, AxiSEM: broadband 3-D seismic wavefields in axisymmetric media, Solid Earth, 5, 425–445, doi:10.5194/se-5-425-2014, 2014. Ratchkovski, N. A., and R. A. Hansen, New constraints on tectonics of interior Alaska: Earthquake locations, source mechanisms, and stress regime, Bull. Seismol. Soc. Am.,92(3), 998– 1014, doi:10.1785/0120010182, 2002. Silver, P. G., and T. H. Jordan, Optimal estimation of scalar seismic moment, Geophys. J. R. Astron. Soc.,70, 755–787, 1982. Tape, C., et al., Earthquake nucleation and fault slip complexity in the lower crust of central Alaska, Nature Geoscience,11, 536–541, doi:10.1038/s41561-018-0144-2, 2018. Tape, W., and C. Tape, A geometric setting for moment tensors, Geophys. J. Int.,190, 476–498, doi:10.1111/j.1365-246X.2012.05491.x, 2012. Thurin, J., R. Modrak, C. Tape, A. McPherson, F. R. Rodr´ıguez-Cardozo, J. Kintner, L. Ding, Q. Liu, and J. Braunmiller, MTUQ: a framework for estimating moment tensors, point forces, and their uncertainties, Geophys. J. Int.,241, 1373–1390, doi:10.1093/gji/ggaf080, 2025. van Driel, M., L. Krischer, S. C. St¨ahler, K. Hosseini, and T. Nissen-Meyer, Instaseis: instant global seismograms based on a broadband waveform database, Solid Earth,6, 701–717, doi: 10.5194/se-6-701-2015, 2015. Zhu, L., and L. A. Rivera, A note on the dynamic and static displacements from a point source in multilayered media, Geophys. J. Int.,148, 619–627, doi:10.1046/j.1365-246X.2002.01610.x, 2002. 4 Figure 1: Example plot of waveform fits. 5 (a) 2025 VLFE (b) 2025 earthquake (c) 2019 earthquake (d) 2022 earthquake Figure 2: Station coverage for the four events examined in this study. The stations are plotted at their ray-path piercing point on the lower-hemisphere of the source mechanism. Upward ray paths are depicted by the X symbols and plotted at their antipode. Downward ray paths are depicted by the circle symbols. (The radial coordinate of each symbol depends on the choice of 1D Earth model.) All stations are within 300 km of each epicenter. The 2019 event occurred during the FLATS deployment (Tape et al., 2018), whose stations were directly above the earthquake, as indicated by the X symbols near the center of the beachball in (c). 6 (a) 2025 VLFE (b) 2025 earthquake (c) 2019 earthquake (d) 2022 earthquake Figure 3: Depth grid searches for the four events in our study. The depth search increment is 1 km, the magnitude increment is ∆Mw= 0.05. The white triangle marks the best-fitting depth, and the uncertainty is depicted by the horizontal line, which marks a misfit value 5% higher than the global minimum. Note that in (a), the horizontal line lies outside the limits of the plot. 7 (a) 2025 VLFE (b) 2025 earthquake Figure 4: Variations in best-fitting source mechanisms as a function of bandpass for the 2025 VLFE (a) and the 2025 earthquake (b). The x-axis is the minimum period of the bandpass, and the y-axis is the maximum period. In total, there are 55 different bandpasses. Each source mechanism beachball is colored by its misfit. The allowable time shifts are ±3 s. The beachball inside the black square is the overall lowest misfit solution among all the bandpasses. Our choice of 12.5–27.5 s is based on the VLFE results in (a). 8