Seismic load analysis of the North Anatolian Fault Marmara segment and probabilistic forecasts for 2026–2030 based on the Gutenberg-Richter law.
HIGH RISK ZONE — NAF Marmara Segment
M5-6
cumulative through 2030
100%
M5–6 magnitude probability
LGBMClassifier
M6-7
cumulative through 2030
89.5%
M6–7 magnitude probability
ExtraTreesClassifier
M7+
cumulative through 2030
52.8%
M7+ destructive probability
RandomForestClassifier
Cumulative Probability Timeline
2026–2030 Annual Risk Evolution
2026
M5-6
93.9%
M6-7
36.2%
M7+
13.9%
2027
M5-6
99.6%
M6-7
59.3%
M7+
25.9%
2028
M5-6
100%
M6-7
74.1%
M7+
36.2%
2029
M5-6
100%
M6-7
83.5%
M7+
45.1%
2030
M5-6
100%
M6-7
89.5%
M7+
52.8%
Data source:
GR-Poisson Model (1900–2024) —
Note: No live API forecasts found — values calculated using the Gutenberg-Richter Poisson model (λ: M5+ 2.8/yr, M6+ 0.45/yr, M7+ 0.15/yr).
Target Region
Istanbul & Marmara — Fault Map
Why So High?
Key Risk Factors
🔴
NAF Marmara Seismic Gap
The Marmara segment of the North Anatolian Fault has been accumulating strain for ~260 years since its last major rupture (1766). Estimated accumulated slip: 3–5 m.
📡
GNSS Strain Rate
GNSS measurements in the Marmara region show 20–26 mm/yr right-lateral slip rate. This is one of Turkey's highest locked fault segments.
🌊
Submarine Fault Geometry
The ~170 km fault segment beneath the Sea of Marmara extends directly under Istanbul. A potential M7.2–7.4 scenario also carries tsunami risk.
📊
Gutenberg-Richter Analysis
1900–2024 catalog data: M5+ annual frequency λ=2.8, M6+ λ=0.45, M7+ λ=0.15. Trend factor 2.0× for the last 10 years (increased activity).
Quarterly Forecasts
2026–2030 Model Outputs
Period
Band
Risk Score
Window
Method
Loading forecasts…
Corporate Risk Report
Detailed seismic risk analysis for Istanbul, fault segment rupture scenarios, structural damage estimates, and a complete forecast package through 2030.
These forecasts are statistical probability models. Exact earthquake timing and magnitude cannot be scientifically predicted. Data is for risk awareness and research purposes only. For information on GR+Trend+ML methodology, see the methodology page.