CISM RSQSim Parameters

From SCECpedia
Jump to navigationJump to search

RSQSim Parameter Sweep Notes

➢ List of main RSQSim parameters for use in deciding on the set of simulations for the parameter sweep. Some parameters work independently, but for several it is the combination that controls important aspects of the simulated catalogs. This list of parameters isn’t meant to be a direct list of the parameters to vary, but a list of the main parameters (and the general range of values for RSQSim calculations) and how they affect untuned RSQSim catalogs.

Physical Parameters:
1. Rate constitutive parameter (A_1)1
    a. Range of values is ~0.001 – 0.01
    b. Strong control on aftershock productivity
2. State constitutive parameter (B_1)1
    a. Range of values is ~0.004 – 0.0015
    b. (b-a) > 0 results in slip weakening behavior, and (b-a) < 0 results in steady creep
3. Initial normal stresses (sigma0_1)2
    a. Range of values is ~100MPa – 200MPa
    b. Scales stress drop, which scales the recurrence times of large events
4. Initial shear stress (tau0_1)2
    a. Range of values is ~55MPa – 120MPa
    b. Shear stress evolves quickly during simulations
5. Coefficient of friction (mu0_1)
    a. Range of values is ~0.5 – 0.7
    b. Simulations are not very sensitive to changes in the coefficient of friction
6. Critical slip distance (Dc_1)
    a. Range of values is ~1e-4 – 1e-5
    b. Increasing Dc by a factor of 10 slightly increases the b-value

Simulation Parameters:
1. Rupture slip speed (ddotEQ)
    a. Range of values is ~0.5 m/s – 3 m/s
    b. Increasing ddotEQ resulted in higher stress drops for single fault simulations
2. Stress Overshoot Factor
    a. Range of values is ~0.01 – 0.3
    b. Small affect on b-value, but increasing it increases the number large events, particularly M6.3-6.5
3. A-reduction Factor (fA)
    a. Range of values is ~0.001 – 0.3
    b. Increasing fA, increases the rate of small events (and thus b-value)


1 Stress drop scales as (b-a)* σ, so perhaps only changing a or b is best.
2 These values are altered during the tuning process, but shouldn’t necessarily be tested independently. During tuning, the normal stress is changed to adjust the stress drop and therefore control the recurrence times of supraseismogenic events. When changing the normal stress however, the shear stress is changed by the same factor to avoid instantaneous nucleation of entire fault segments.


Default Parameters:
• A_1 = 0.001
• fA = 0.1
• B_1 = 0.008
• Dc_1 = 1.0e-05
• mu0_1 = 0.6
• tau0_1 = 55
• sigma0_1 = 100
• ddotEQ_1 = 1
• stressOvershootFactor = 0.1

Optional Slip Weakening Parameters*:
• muAmp = 0.2
• muInvDist = 1.0
cohesion = 6

  • When using the additional slip weakening friction (i.e. setting them to nonzero values), smaller values of a and b are preferred (a = 0.003-0.006, b=0.004-0.006); additionally the cohesion parameter is still optional with the additional weakening parameters.

Parameter Sweep Plan:
• Tuned stresses can be input as initial stresses for new parameter sweep models
    - Perturb list of parameters
    - Perturb combinations of parameters to keep other values constant (specifically stress drop)
    - Compare with original, tuned, million-year catalog
    - Possibly retune the catalogs


Main Parameters for Sweep: (Each Million-year catalog requires ~10k SU’s on Blue Waters)
1. A_1 (0.003-0.01)
2. A_1 and B_1 simultaneously (to keep stress drop constant; (b-a)*sigma)
3. Mu0_1 (0.5-0.7)
4. Dc_1 (1e-4 – 1e-5)
5. ddotEQ_1 (0.5 – 3)
6. stressOvershootFactor (0.01 – 0.3)
7. fA (0.01 – 0.3)

Other potential model variables:
• Parameter heterogeneity
   - Varying parameters stochastically
   - Varying parameters at depth
• Structural heterogeneity
   - Rough faults
   - Fractally segmented faults
• Different fault models
• Deep creep