SEISMIC · FORECAST

We Don't Predict Earthquakes.
We Calculate Seismic Risk with Millimetric Precision.

Free weekly seismic risk forecasts for every active fault zone. Need the full picture for your exact address? Get a professional 5-page PDF report — fault analysis, 5-year probabilities, and AI risk scores across all magnitude bands — delivered to your inbox for €5.

0.619–0.994 M4-5 AUC Range
8 Regions Tested
127 Backtest Years
0.099 Area Skill Score*
USGS Earthquake Catalog
ISC Bulletin
NGL GNSS Network
GEM Active Faults
AFAD Catalog
Ambraseys Catalog
01Use Cases
Eren Bostan·[TAL-118]Map the people each forecast serves

Who Needs Seismic Intelligence?

🏠
Homeowners & Renters
Know your property's true seismic exposure before buying, building, or renewing your insurance. Location-specific risk — not just a regional average.
🏗️
Real Estate & Investors
Quantify location-specific seismic risk for site selection, due diligence reports, and portfolio-level exposure management.
📊
Insurers & Risk Analysts
Multi-band probability data for premium calculation, reserve modeling, and regulatory stress testing.
02Free Weekly Forecasts
Talivio Engine·[TAL-204]Stream 30-day regional forecasts

Latest Regional Forecasts

See All Forecasts →
M6-7 · 30-day window
30%
Low Risk
Explore Free → →
M6-7 · 30-day window
30%
Low Risk
Explore Free → →
M5-6 · 30-day window
45%
Moderate Risk
Explore Free → →
M5-6 · 30-day window
45%
Moderate Risk
Explore Free → →

Precision @ 0.7

Near-Zero False Positive Rate

Classic AI models cry wolf at every minor tremor, inflating their accuracy numbers. Talivio works differently. Our system is deliberately under-confident — it never triggers an alert when the evidence is uncertain.

When our system issues a "Risk > 0.7" warning, seismic activity in the target region is confirmed 91% of the time across 15 walk-forward folds.

Say goodbye to false alarms. Know only the moments that truly require action.

03Benchmark
Eren Bostan·[TAL-152]Compare against published forecasting systems

Beyond Classical ML and State-of-the-Art

Feature / Metric Traditional Institutions (USGS) Classical Academic AI Talivio (Regional ML)
Core Logic Pure Statistics (only knows aftershocks) Black-box Classification (memorises) Physics-Informed AI (Spatial/Temporal)
Spatial Resolution 10–50 km (very broad) Province or region level H3 Res 8/9 (0.7 km² pinpoint)
Time Horizon Usually 24 hours Single & uncertain 7, 30 & 90-day dynamic probabilities
False Positive Medium Very high (up to 80% false alarms) Near-zero ([email protected] = 91%)
Probability Calibration (ECE) None Usually not measured 0.032 (excellent calibration)
Brier Skill Score (BSS) 1.5–3% (reference) Negative (worse than random) 23.9% (far above industry standard)
04Global Comparison
Eren Bostan·[TAL-160]Rank Area Skill Score among global systems

World Earthquake Prediction Benchmark Ranking

Talivio's performance compared to published earthquake forecasting systems worldwide. Our AUC values use hard same-region negatives — the most rigorous evaluation protocol.

SystemMetricValue ASS† WindowNegativesValidation
ETAS Italy (OEF-Italy) Area Skill Score 0.7 0.700 1-day same_region Prospective
RELM California (CSEP) Probability Gain vs Random 10 5-year same_region Prospective
CSEP California (2011–2020) IGPE vs Benchmark 0 5-year same_region Prospective
FCN Deep Learning (California) Area Skill Score 0.882 0.882 15–90 day same_region Retrospective
DeVries 2018 (Google Brain) AUC 0.849 static global_mixed Retrospective
ETAS Japan (CSEP daily) CSEP Pass Rate 0.9 1-day same_region Prospective
EEPAS New Zealand IGPE 0.64 3-month same_region Prospective
Talivio v2 AUC + ASS (Molchan) 0.619–0.994 0.099 30-day Hard same-region Retrospective + Prospective

† ASS = Area Skill Score (Molchan, [0,1]). FCN=0.882, ETAS Italy=0.70 (published).
8 regions, 127 backtest years. Full methodology →

05Architecture
Talivio Engine·[TAL-031]Per-band algorithm competition pipeline

How It Works

Mechanical & Celestial Stress
Fault geometry (strike-slip/normal) and Coulomb stress transfers are mapped to compute physical fracture points.
Epidemic Clustering (ETAS)
Hundreds of thousands of past earthquakes are analysed via Morlet Wavelet transforms and temporal decay to separate background noise from genuine hazard signals.
Spatial Network Awareness
Using Uber's H3 hexagonal architecture, seismic stress transfer between neighbourhoods is analysed with graph-based logic.
06Walk-Forward Validation
Eren Bostan·[TAL-145]Expanding-window backtest engine

25 Years of Uninterrupted Validation

Our system has been tested with Walk-forward Validation and a 30-day purge gap protocol to prevent data memorisation.

0.809
Median M4-5 AUC
Across 8 regions, hard same-region negatives
0.994
Best Regional AUC
Highest M4-5 AUC across all regions
127
Backtest Years
Total years of expanding-window validation
0.099
Area Skill Score
Molchan ASS (prospective)
07Location-Specific Models
Talivio Engine·[TAL-189]Train location-specific regional models

Location-Specific ML Models

Each region trains its own ML model. Negative samples are drawn from the same tectonic zone — producing honest AUC values (0.62–0.99 range) with no geographic bias.

Istanbul & Marmara
0.619
AUC · M4-5
M3-4 M4-5 M5-6 M6-7 M7+
Kahramanmaras & EAF
0.796
AUC · M4-5
M3-4 M4-5 M5-6 M6-7 M7+
Yedisu Seismic Gap
0.738
AUC · M5-6
M3-4 M4-5 M5-6 M6-7 M7+
Izmir & Aegean
0.702
AUC · M4-5
M3-4 M4-5 M5-6 M6-7 M7+
Cascadia Subduction
0.940
AUC · M3-4
M3-4 M4-5 M5-6 M6-7 M7+
Tokyo Metropolitan
0.809
AUC · M4-5
M3-4 M4-5 M5-6 M6-7 M7+
Nepal-Tibet Himalayan
0.983
AUC · M4-5
M3-4 M4-5 M5-6 M6-7 M7+
San Andreas South
0.854
AUC · M4-5
M3-4 M4-5 M5-6 M6-7 M7+
Zagros Thrust Belt
0.894
AUC · M5-6
M3-4 M4-5 M5-6 M6-7 M7+
NZ Alpine Fault
0.994
AUC · M4-5
M3-4 M4-5 M5-6 M6-7 M7+
Mindanao & Philippine Trench
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Tohoku & Japan Trench
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Northern California — SF Bay & Mendocino
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Caribbean — Venezuela (El Pilar)
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Sulawesi — Palu-Koro
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Kamchatka–Kuril Subduction
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
PNG New Britain–Bismarck
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Tonga–Fiji
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Vanuatu
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Kermadec
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Kuril Islands
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Loyalty–S.Vanuatu
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Molucca–Halmahera
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Central Chile
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Banda Sea
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Tonga North
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Santa Cruz Islands
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Chiapas–Guatemala
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
PNG Huon
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Taiwan
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Western Aleutians
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
South Sandwich
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Fiji deep
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Costa Rica–Panama
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Bonin–Izu
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Bougainville
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Atacama
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
North Sumatra
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Flores Sea
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Mentawai
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Papua Birds Head
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Haida Gwaii
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Hindu Kush
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Guerrero–Oaxaca
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Mariana
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
South Peru
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Pacific-Antarctic Ridge
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Ryukyu
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
North Kuril
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Papua–Yapen
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Komandorski
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Sunda–Sumbawa
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Colombia
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Luzon
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Eastern Aleutians
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
South Chile
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Andaman–Nicobar
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Hellenic Arc–Crete
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Ecuador
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Jalisco–Colima
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Guam
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
NZ North–Raoul
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Balleny
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Java
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Atlantic Equatorial
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Argentina deep
Not trained
M3-4 M4-5 M5-6 M6-7 M7+
Alaska Peninsula
Not trained
M3-4 M4-5 M5-6 M6-7 M7+

Amber band = trained model exists · AUC = expanding-window backtest result (hard same-region negatives)

Seismic Intelligence

Latest from the Blog

Deep dives into earthquake science, AI forecasting models, and seismic risk analysis.

Visit Blog
Global M5+ Seismic Activity: Talivio's Weekly Analytical Recap (June 15-21)
Data Sources

Global M5+ Seismic Activity: Talivio's Weekly Analytical Recap (June 15-21)

This comprehensive recap from Talivio analyzes global M5+ seismic activity between June 15 and 21, highlighting significant events and their tectonic contexts. Our AI-driven platform continuously monitors Earth's movements, providing data-backed insights into recent seismicity patterns and contributing to a deeper understanding of our dynamic planet.

Jun 27, 2026 5 min read
Unveiling Earth's Tensions: How b-value Anomalies Drive Talivio's Earthquake Forecasts
Seismic Science

Unveiling Earth's Tensions: How b-value Anomalies Drive Talivio's Earthquake Forecasts

Earthquakes remain one of nature's most unpredictable phenomena, yet advancements in AI are bringing us closer to understanding their precursors. This post delves into how Talivio leverages b-value anomalies, a critical seismic indicator, to enhance its AI-powered earthquake forecasting models. We explore the science behind b-value changes and its integration into our sophisticated machine learning framework.

Jun 26, 2026 6 min read
Cascadia's Silent Giant: Unveiling Megathrust Risk with AI-Powered Insights
Risk Analysis

Cascadia's Silent Giant: Unveiling Megathrust Risk with AI-Powered Insights

The Cascadia Subduction Zone poses one of the most significant seismic hazards globally, capable of generating devastating megathrust earthquakes. Talivio leverages advanced AI and machine learning to continuously monitor this critical region, analyzing complex geophysical data to enhance our understanding of its unique tectonic characteristics and potential for future seismic events.

Jun 25, 2026 7 min read
08What's Inside
Eren Bostan·[TAL-099]Spec the risk report deliverable

Your Report: 5-Page Professional PDF

AI risk scores across all 5 magnitude bands (M3–4 to M7+)
Multi-horizon probability table: 7-day / 30-day / 90-day / 1-year / 5-year
Active fault segment analysis with strike, dip, and slip rates
Historical earthquake catalog (USGS + ISC + Ambraseys)
GNSS strain rate & seismic cycle phase analysis
Celestial tidal stress windows
Physical impact estimates & structural damage scenarios
AI model performance metrics (ROC-AUC, Brier score, calibration)
Delivered to your inbox within 1–5 minutes. View Free Sample Report →
09FAQ
Eren Bostan·[TAL-220]Answer the questions people actually ask

Frequently asked questions

Is this a guaranteed earthquake prediction?
No. Talivio produces a 30-day probabilistic risk score for a given region and magnitude band — not a guaranteed prediction. Scores are for scientific purposes and decision support.
Where does the data come from?
Public scientific sources including the USGS and ISC earthquake catalogs, AFAD, the NGL GNSS geodesy network, and the GEM active-fault database. All sources are listed on the methodology page.
How is accuracy measured?
With walk-forward, expanding-window backtests using AUC, Brier score, and the Molchan Area Skill Score — each region evaluated against hard same-region negatives.
What does a risk report include?
Regional risk distribution, magnitude-band probabilities, fault-segment analysis, the historical seismic cycle, and preparedness guidance. You can preview a sample report.
Which regions are covered?
Trained models exist for 15 regions and the list keeps growing. Location-specific models are trained on each region's own seismic history.

Location Reports

Street-Level Seismic Intelligence

€5
one-time · no subscription

One complete seismic risk analysis for your exact location — 5 magnitude bands, 5 time horizons, fault map, and historical catalog.

Get Your Report — €5

View Free Sample Report → →

This system provides statistical probability forecasts. Earthquake time, location and magnitude cannot be precisely predicted scientifically. Probabilities are based on historical patterns and physical models; they do not imply certainty.