Transcript Magnitude Limits: Implications of the 2011 Tohoku Earthquake

Comments on physical
simulator models
David D. Jackson, UCLA
Forecast capabilities
• Statistical distributions of earthquakes
on prespecified faults
– Magnitude frequency
– Temporal (recurrence) for initiation,
participation, and combinations
– Spatial
• Paleoseismic rate, sigma, slip rate,
displacement distribution
Off-fault earthquakes
• Not defined in phys sim models
• Contribute to stress, and presumably to
earthquake probability, on faults
• Are counted in actual earthquake
statistics (magnitude, temporal, spatial)
but not in phys sim models
• Participate in clustering models as
sources and receivers, but not in phys
sim models
UCERF2 Assignment of historic earthquakes to faults
UCERF2 Assignment of instrumental earthquakes to faults
Summary, UCERF2 Earthquakes
on faults
total
M>7
After 1931
Yes
15
7
11
no
8
2
1
maybe
23
6
7
Total
46
15
19
Limits to forecasting power
• Phys sim models are explicitly time-dependent
(conditional on history), but history can’t be input or
matched in simulations
• Can’t be tested because they could forecast only onfault quakes, which are not defined.
• Paleoseis record is presumably recording on-fault
quakes, but is not complete.
• ?Clustering studies show that small earthquakes are
important in triggering, but phys sim models can
include them only with massive calculations?.
• Models depend on Coulomb stress, which hasn’t yet
demonstrated clear forecasting ability. But
simulators could help make the case by setting up
initial conditions.
Conditional Stopping Probability in 10km segment
boundary according to WGCEP2008
Elsinore Fault
1
Conditional Probability
WG07 Fault Rupture Model
0.8
0.6
Null hypothesis
0.4
0.2
0
0
50
100
150
Distance from North End, km
200
250
Smoothed seismicity and faults
Stochastic earthquake simulation,
100 m=6.5+ events