Earthquake Impact Simulator

Simulation view

Scenario: Cascadia Subduction Zone

Estimated consequence window—Adjust the scenario, then run the model.

Severe shakingStrong shakingFelt areaAftershocksTsunami/offshore

Wave propagation timeline

P-wave · S-wave · surface-wave / lifeline delay

T+0 secRupture nucleation

Fault/location presets

Fast scenario cards in the style of an impact map console

Science, formulas, and calculation trace

All estimates are visible: magnitude energy, rupture scaling, intensity decay, casualties, damage, aftershocks, tsunami/volcano/outage gates.

Mathematical model

Seismic moment and energy

M₀ = 10^(1.5Mw + 9.1) N·m
log₁₀E(J) = 1.5Mw + 4.8

Moment expresses source size; radiated energy gives intuitive blast-scale comparison.

Rupture scaling

log₁₀L = a + bMw
strike: a=-3.55,b=.74 · reverse/subduction: -2.86,.63 · normal: -2.57,.62 · subduction: -2.86,.63

Wells-Coppersmith-style regression provides the physical rupture length. The user scale slider is visual-only and cannot override magnitude-derived source physics.

Intensity field

PGA = GMPE(Mw, R, Vs30)
MMI = Worden (2012) GMICE(PGA)

Shaking is a distance-dependent field, not a single peak value: a BSSA14-style GMPE gives PGA at each distance, converted to MMI by the Worden et al. (2012) ground-motion-to-intensity equation. Site class (Vs30, set by exposure density) controls soil amplification.

Loss model (PAGER-style)

fatalities = Σⱼ ν(MMIⱼ)·Pop(MMIⱼ)
ν(S) = Φ[ (1/β)·ln(S/θ) ]

Jaiswal & Wald (2010) lognormal fatality rate applied to population distributed across intensity bins by isoseismal area — not peak intensity applied to everyone. θ, β set by vulnerability class (modern / mixed / fragile). Calibrated against Northridge, Loma Prieta, Kobe, Tōhoku, Türkiye and Wenchuan to within ~2–3×; Haiti-class URM collapse is under-predicted, as in global models.

Aftershock productivity

N(M≥m) ≈ 10^(a − b m), b≈1
Mmax_after ≈ Mw − 1.1 + productivity×0.5

Aftershock output is scenario-level, based on Gutenberg-Richter behavior and productivity slider.

Rupture kinematics (3D view)

D̄ = M₀ / (μ·L·W)
W: log₁₀W = −1.01+.32Mw (crust) · −0.86+.35Mw (interface)
Vᵣ ≈ 0.8Vₛ · Tᵣ = 2.03×10⁻⁹·M₀^⅓ (dyn·cm)

The 3D fault cinema solves mean slip from moment balance (μ = 33–40 GPa), propagates a circular rupture front from the hypocenter at 0.8Vₛ, applies Somerville-style rise time, and separates hanging wall / footwall along the style-dependent dip and rake. Block offsets are exaggerated for visibility and labeled as such.

Tsunami / volcano / lifelines

Tsunami score = offshore × thrust/subduction × Mw × shallow depth
Volcano score = volcanic zone × Mw × depth gate
Outage = population × fragility × MMI × duration

These are gate scores, not warnings. Tsunami warnings require official tsunami-warning centers.

Run or generate a report to print the calculation trace.

Scientific limitations

LIMITS: This simulator uses a simplified (educational) GMPE, GMICE and PAGER-style loss model. It does not use published regulatory GMPE coefficient tables, real soil/Vs30 grids, building inventory, bridge/network topology, hospital capacity, live occupancy, landslide susceptibility rasters, official tsunami propagation, or volcanic observatory data. It is designed to be honest, transparent, and useful as a public-facing scenario visualizer.

References used for model structure

USGS real-time feeds

Live event button uses USGS GeoJSON feeds. See: earthquake.usgs.gov/earthquakes/feed/v1.0/geojson.php

PAGER / ShakeMap logic

Impact estimates follow the broad PAGER principle: population exposed to shaking intensity is compared against fatality and economic-loss models.

Wells & Coppersmith

Rupture length/area relationships are represented by simplified log-linear regressions for educational estimates.

Gutenberg-Richter

Aftershock scenario follows frequency-magnitude logic with b≈1 and a productivity control.

Seismic Hazard Analysis · DSHA + PSHA

GMPE-based deterministic scenario and full Cornell–McGuire probabilistic hazard: hazard curve, return-period ground motion, deaggregation, uniform hazard spectrum, and an epistemic logic tree.

HAZARD ENGINE

Site class Vs30 (m/s) Closest distance R (km) Fault length L (km) Max magnitude Mmax G–R b-value Activity rate ν (M≥5 /yr) Hazard-curve ordinate Design return period

Run the hazard analysis to compute DSHA and PSHA outputs.

DSHA controlling eq—

DSHA median—

DSHA 84th %ile—

PSHA design—

Return period—

Annual rate λ—

Deagg mean M—

Deagg mean R—

Hazard curve — annual rate of exceedance (mean + logic-tree envelope)

Uniform Hazard Spectrum vs DSHA

Deaggregation by magnitude

Calculation trace

Run the hazard analysis to print the trace.

Ground-motion model (GMPE)

ln Y = e1 + F_M(M) + F_D(R,M) + F_S(Vs30)
F_D=(c1+c2(M−4.5))·ln√(R²+h²)+c3(R−1)
σ_lnY ≈ 0.60

Active-crustal Boore–Atkinson / BSSA14-style functional form: magnitude saturation about a hinge at M 6.2, geometric spreading, anelastic attenuation, and a Vs30 site term. The coefficient set is a compact educational parameterization tuned to reproduce NGA-West2-style median scaling and dispersion — not a verbatim regulatory table.

DSHA

Y_med = GMPE(Mmax, R_min, Vs30)
Y_84 = Y_med · exp(σ_lnY)

Controlling earthquake = Mmax at the closest distance. Reports the median and the 84th-percentile (median + 1σ) response spectrum — the standard deterministic deliverable for critical facilities.

PSHA — Cornell (1968) / McGuire

λ(Y>y)=Σ ν·∫∫ P(Y>y|m,r)·f_M(m)·f_R(r) dm dr
f_M: truncated exponential (bounded G–R)
P(Y>y|m,r)=1−Φ((ln y−μ)/σ), |ε|≤3

Total-probability integral over a bounded Gutenberg–Richter magnitude PDF and a planar-fault distance distribution, with lognormal aleatory variability truncated at 3σ. Evaluated numerically over discretised (m,r) bins.

Return period & Poisson

P(Y>y in t)=1−e^(−λt)
10%/50 yr → λ=2.11e-3 (475 yr)
2%/50 yr → λ=4.04e-4 (2475 yr)

The mean hazard curve is inverted at the target annual rate to give the design ground motion. Deaggregation reports the mean/modal (M,R) controlling that level; the UHS is the equal-exceedance ground motion across spectral periods.

Epistemic logic tree

b ∈ {b−.1, b, b+.1} (0.3/0.4/0.3)
Mmax ∈ {Mmax, Mmax−.3} (0.6/0.4)
mean λ = Σ w·λ_branch

Six weighted branches separate epistemic (knowledge) uncertainty from the aleatory σ carried inside the GMPE. The shaded band on the hazard curve is the branch envelope.

References

Cornell (BSSA, 1968); McGuire (2004, EERI monograph); Boore & Atkinson (2008, Earthquake Spectra); Boore, Stewart, Seyhan & Atkinson (2014); Wells & Coppersmith (1994); Gutenberg & Richter (1944); Baker, Introduction to PSHA (2013). Coefficients are simplified for transparency — use OpenQuake (GEM) or the USGS NSHM for engineering-grade work.

Earthquake Impact Simulator PRO