SCEC Award Number 15157 View PDF
Proposal Category Collaborative Proposal (Integration and Theory)
Proposal Title Alternative Rupture Basis for UCERF4
Investigator(s)
Name Organization
Glenn Biasi University of Nevada, Reno Timothy Dawson California Geological Survey
Other Participants
SCEC Priorities 4e, 2e, 6b SCEC Groups WGCEP, CS
Report Due Date 03/15/2016 Date Report Submitted 03/14/2016
Project Abstract
 We present an alternate, smaller rupture basis for use in future Grand Inversion (GI) based rupture rate forecasts on California faults. The smaller rupture set is a subset of the full UCERF3 set selected to model all possible ground motions, rather than all possible ruptures. Ruptures span all possible ground motions by assigning lengths set by small magnitude increments dM instead of small length increments dL as in UCERF3. Magnitude scales as exp(L), so fewer than 30 ruptures from a subsection represent ground motion for longest faults, compared to 130 or more for UCERF3. By construction the method preserves all shorter ruptures, so hazard resolution is maintained throughout. We show that rupture counts on faults with continuity to the San Andreas and San Jacinto shrink by 92 to 98%. Point checks confirm that every subsection still has a wide and redundant set of ruptures to represent hazard. Com-pared to the FM3.1 full set of 253,706 ruptures, the largest filtered rupture set with 32,932 ruptures is only 13%, but still has a factor of ~3 more ruptures than are used by the GI for individual runs. GI inversions begin final convergence an order of magnitude sooner, or ~100 s compared to 1800 s, and fit the data slightly better on average. The smaller rupture set seems to better focus the inversion. Ground-motion filtered rupture sets increase in size slowly with finer fault discretization, and thus could be used to improve hazard resolution from moderate earthquakes.
Intellectual Merit This research shows how order of magnitude improvements may be possible in the convergence time of future versions of the Uniform California Earthquake Rupture Forecast. Savings are realized by changing the focus of the Grand Inversion from rates of all possible ruptures to the rates of all relevant magnitudes of earthquakes. The improved convergence may be taken as computational savings or improved fault system detail in future models.
Broader Impacts Improvements to UCERF developed in this work will allow more detailed fault models to be considered in the future, including more uniform coverage of moderate earthquake magnitudes where seismic hazard is concentrated. Better earthquake rupture forecasts improve public safety and preparation for seismic hazards.
Exemplary Figure Figure 3. (Upper) Example convergence rates for the full UCERF3 rupture set and the “Both” subset. Energy is a measure of total data misfit. The filtered rupture subset starts the final convergence phase in an order of magnitude sooner than the full UCERF3 set and converges to a similar final fit. (Lower) The three filtered subsets converge at similar rates, indicating a general stability in the filtering process.