Earthquake · next 30 days
Where a large earthquake is more likely in the next 30 days
Earthquakes cannot be predicted. This forecast gives the chance of at least one magnitude 6+ earthquake in each 2° grid cell of the world over the next 30 days, so you can see where the odds are higher than usual.
Map
Quality checks: Published with warnings. The forecast is published; a non-blocking check flagged the following. Full check record
- Calibration on live forecasts not yet measured (too few closed forecasts with events)
For earthquake information, alerts and safety guidance, follow the USGS Earthquake Hazards Program and your national geological survey. HazardPulse publishes research forecasts, not warnings.
- 0.1–0.3%
- 0.3–1%
- 1–3%
- 3–10%
- 10% or more
Unshaded cells have a chance below 0.1%. Numbered pins match the list below.
After a large earthquake
A large earthquake is often followed by more large earthquakes nearby. Here is how often that has happened.
M7.0, 18 km deep, Xizang — (29 days before this forecast).
In the world’s record since 1990, after shallow earthquakes of M7.0–7.4, another M6+ struck the same 2° cell within the next 30 days in 15% of cases, counting from a week after it (95% interval 10%–19%; 245 earthquakes). This forecast gives that cell 2.6% for the next 30 days. Our aftershock model in test gives it 3.4% (not the published number).
This is the historical record, not a forecast for this earthquake. Our published model is known to give too low a chance in the first days after a large earthquake (earthquake program section 13). The aftershock model in test was checked on 88 large earthquakes it was not fitted on: a day after each, it gave the epicentre’s cell 25% where 25% had another M6+ (the published model: 13%). It is not yet the published number because it has not passed our check that it leaves ordinary forecasts no worse (earthquake program section 15). Large aftershocks can be as damaging as the first earthquake: follow your local authorities.
Cells with the highest chance
The 20 cells with the highest 30-day chance. Open a row for the cell, its recent earthquakes and one input of the forecast.
1Vanuatu Islands15°S 167°E · 11 M2.5+ earthquakes in the last 30 days, largest M6.212.5%
Rate-model input: In the clustering-rate model the forecast reads as one input, 77% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 12.5%
- Cell
- 2° × 2° centred on 15°S 167°E, 810 km N of Nouméa, New Caledonia
- Recent earthquakes in the cell
- 11 M2.5+ earthquakes in the last 30 days, largest M6.2
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 6.76×
- Gutenberg–Richter b-value
- 1.04
- Correlation length
- 25.1 km
- Spatial spread of recent events
- 70.4 km
- Depth trend
- −61.8 km
2Mindanao, Philippines5°N 125°E · 9 M2.5+ earthquakes in the last 30 days, largest M5.110.7%
Rate-model input: In the clustering-rate model the forecast reads as one input, 89% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 10.7%
- Cell
- 2° × 2° centred on 5°N 125°E, 130 km S of General Santos, Philippines
- Recent earthquakes in the cell
- 9 M2.5+ earthquakes in the last 30 days, largest M5.1
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 0.31×
- Gutenberg–Richter b-value
- 1.14
- Correlation length
- 41.4 km
- Spatial spread of recent events
- 84.9 km
- Depth trend
- −2.2 km
3New Britain Region, P.N.G.5°S 151°E · 4 M2.5+ earthquakes in the last 30 days, largest M5.08.8%
Rate-model input: In the clustering-rate model the forecast reads as one input, 69% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 8.8%
- Cell
- 2° × 2° centred on 5°S 151°E, 480 km ENE of Lae, Papua New Guinea
- Recent earthquakes in the cell
- 4 M2.5+ earthquakes in the last 30 days, largest M5.0
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 1.08×
- Gutenberg–Richter b-value
- 1.32
- Correlation length
- 89.2 km
- Spatial spread of recent events
- 90.5 km
- Depth trend
- −26.7 km
4New Ireland Region, P.N.G.5°S 153°E · 5 M2.5+ earthquakes in the last 30 days, largest M5.08.8%
Rate-model input: In the clustering-rate model the forecast reads as one input, 69% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 8.8%
- Cell
- 2° × 2° centred on 5°S 153°E, 690 km ENE of Lae, Papua New Guinea
- Recent earthquakes in the cell
- 5 M2.5+ earthquakes in the last 30 days, largest M5.0
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 1.22×
- Gutenberg–Richter b-value
- 1.27
- Correlation length
- 100.4 km
- Spatial spread of recent events
- 62.8 km
- Depth trend
- −10.4 km
5Vanuatu Islands13°S 167°E · 2 M2.5+ earthquakes in the last 30 days, largest M4.78.3%
Rate-model input: In the clustering-rate model the forecast reads as one input, 69% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 8.3%
- Cell
- 2° × 2° centred on 13°S 167°E, 860 km ESE of Honiara, Solomon Islands
- Recent earthquakes in the cell
- 2 M2.5+ earthquakes in the last 30 days, largest M4.7
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 6.45×
- Gutenberg–Richter b-value
- 1.02
- Correlation length
- 25.0 km
- Spatial spread of recent events
- 81.8 km
- Depth trend
- −87.8 km
6Near Coast of Chiapas, Mexico15°N 93°W · 11 M2.5+ earthquakes in the last 30 days, largest M4.58.1%
Rate-model input: In the clustering-rate model the forecast reads as one input, 85% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 8.1%
- Cell
- 2° × 2° centred on 15°N 93°W, 54 km SW of Escuintla, Mexico
- Recent earthquakes in the cell
- 11 M2.5+ earthquakes in the last 30 days, largest M4.5
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 0.74×
- Gutenberg–Richter b-value
- 0.90
- Correlation length
- 42.0 km
- Spatial spread of recent events
- 80.1 km
- Depth trend
- −10.3 km
7Talaud Islands, Indonesia3°N 127°E · 2 M2.5+ earthquakes in the last 30 days, largest M5.37.6%
Rate-model input: In the clustering-rate model the forecast reads as one input, 64% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 7.6%
- Cell
- 2° × 2° centred on 3°N 127°E, 250 km N of Ternate, Indonesia
- Recent earthquakes in the cell
- 2 M2.5+ earthquakes in the last 30 days, largest M5.3
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 0.31×
- Gutenberg–Richter b-value
- 1.31
- Correlation length
- 33.0 km
- Spatial spread of recent events
- 67.2 km
- Depth trend
- 12.7 km
8Philippine Islands Region5°N 127°E · 6 M2.5+ earthquakes in the last 30 days, largest M5.07.6%
Rate-model input: In the clustering-rate model the forecast reads as one input, 68% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 7.6%
- Cell
- 2° × 2° centred on 5°N 127°E, 240 km ESE of General Santos, Philippines
- Recent earthquakes in the cell
- 6 M2.5+ earthquakes in the last 30 days, largest M5.0
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 0.35×
- Gutenberg–Richter b-value
- 1.18
- Correlation length
- 48.8 km
- Spatial spread of recent events
- 78.8 km
- Depth trend
- −8.6 km
9Santa Cruz Islands11°S 165°E · 6 M2.5+ earthquakes in the last 30 days, largest M6.47.3%
Rate-model input: In the clustering-rate model the forecast reads as one input, 72% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 7.3%
- Cell
- 2° × 2° centred on 11°S 165°E, 580 km ESE of Honiara, Solomon Islands
- Recent earthquakes in the cell
- 6 M2.5+ earthquakes in the last 30 days, largest M6.4
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 1.34×
- Gutenberg–Richter b-value
- 1.17
- Correlation length
- 70.3 km
- Spatial spread of recent events
- 60.7 km
- Depth trend
- −1.5 km
10Philippine Islands Region7°N 127°E · 7 M2.5+ earthquakes in the last 30 days, largest M4.67.3%
Rate-model input: In the clustering-rate model the forecast reads as one input, 75% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 7.3%
- Cell
- 2° × 2° centred on 7°N 127°E, 150 km E of Davao, Philippines
- Recent earthquakes in the cell
- 7 M2.5+ earthquakes in the last 30 days, largest M4.6
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 0.44×
- Gutenberg–Richter b-value
- 1.14
- Correlation length
- 41.7 km
- Spatial spread of recent events
- 77.3 km
- Depth trend
- −8.2 km
11Vanuatu Islands19°S 169°E · 3 M2.5+ earthquakes in the last 30 days, largest M4.67.2%
Rate-model input: In the clustering-rate model the forecast reads as one input, 67% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 7.2%
- Cell
- 2° × 2° centred on 19°S 169°E, 450 km NE of Nouméa, New Caledonia
- Recent earthquakes in the cell
- 3 M2.5+ earthquakes in the last 30 days, largest M4.6
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 3.94×
- Gutenberg–Richter b-value
- 1.32
- Correlation length
- 20.5 km
- Spatial spread of recent events
- 69.8 km
- Depth trend
- −42.3 km
12Kermadec Islands, New Zealand29°S 177°W · 0 M2.5+ earthquakes in the last 30 days7.0%
Rate-model input: In the clustering-rate model the forecast reads as one input, 73% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 7.0%
- Cell
- 2° × 2° centred on 29°S 177°W, 1,110 km NE of Whangarei, New Zealand
- Recent earthquakes in the cell
- 0 M2.5+ earthquakes in the last 30 days
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 0.61×
- Gutenberg–Richter b-value
- 1.12
- Correlation length
- 64.5 km
- Spatial spread of recent events
- 75.0 km
- Depth trend
- 4.8 km
13Hokkaido, Japan Region41°N 143°E · 6 M2.5+ earthquakes in the last 30 days, largest M5.16.9%
Rate-model input: In the clustering-rate model the forecast reads as one input, 79% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 6.9%
- Cell
- 2° × 2° centred on 41°N 143°E, 130 km ENE of Hachinohe, Japan
- Recent earthquakes in the cell
- 6 M2.5+ earthquakes in the last 30 days, largest M5.1
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 0.40×
- Gutenberg–Richter b-value
- 0.99
- Correlation length
- 51.5 km
- Spatial spread of recent events
- 68.7 km
- Depth trend
- 4.7 km
14Off East Coast of Honshu, Japan39°N 143°E · 3 M2.5+ earthquakes in the last 30 days, largest M4.76.8%
Rate-model input: In the clustering-rate model the forecast reads as one input, 77% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 6.8%
- Cell
- 2° × 2° centred on 39°N 143°E, 180 km ESE of Morioka, Japan
- Recent earthquakes in the cell
- 3 M2.5+ earthquakes in the last 30 days, largest M4.7
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 0.41×
- Gutenberg–Richter b-value
- 1.06
- Correlation length
- 50.2 km
- Spatial spread of recent events
- 77.0 km
- Depth trend
- 1.1 km
15Philippine Islands Region9°N 127°E · 13 M2.5+ earthquakes in the last 30 days, largest M5.86.6%
Rate-model input: In the clustering-rate model the forecast reads as one input, 76% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 6.6%
- Cell
- 2° × 2° centred on 9°N 127°E, 160 km E of Butuan, Philippines
- Recent earthquakes in the cell
- 13 M2.5+ earthquakes in the last 30 days, largest M5.8
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 0.94×
- Gutenberg–Richter b-value
- 1.35
- Correlation length
- 92.5 km
- Spatial spread of recent events
- 69.0 km
- Depth trend
- −3.2 km
16Banda Sea7°S 129°E · 8 M2.5+ earthquakes in the last 30 days, largest M5.16.6%
Rate-model input: In the clustering-rate model the forecast reads as one input, 65% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 6.6%
- Cell
- 2° × 2° centred on 7°S 129°E, 380 km SSE of Ambon, Indonesia
- Recent earthquakes in the cell
- 8 M2.5+ earthquakes in the last 30 days, largest M5.1
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 0.99×
- Gutenberg–Richter b-value
- 1.31
- Correlation length
- 92.5 km
- Spatial spread of recent events
- 69.0 km
- Depth trend
- −12.6 km
17Tanimbar Islands Reg., Indonesia7°S 131°E · 10 M2.5+ earthquakes in the last 30 days, largest M6.66.6%
Rate-model input: In the clustering-rate model the forecast reads as one input, 77% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 6.6%
- Cell
- 2° × 2° centred on 7°S 131°E, 480 km SE of Ambon, Indonesia
- Recent earthquakes in the cell
- 10 M2.5+ earthquakes in the last 30 days, largest M6.6
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 0.91×
- Gutenberg–Richter b-value
- 1.30
- Correlation length
- 84.2 km
- Spatial spread of recent events
- 63.5 km
- Depth trend
- 3.9 km
18Near East Coast of Honshu, Japan37°N 141°E · 10 M2.5+ earthquakes in the last 30 days, largest M5.76.5%
Rate-model input: In the clustering-rate model the forecast reads as one input, 69% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 6.5%
- Cell
- 2° × 2° centred on 37°N 141°E, near Iwaki, Japan
- Recent earthquakes in the cell
- 10 M2.5+ earthquakes in the last 30 days, largest M5.7
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 1.25×
- Correlation length
- 83.4 km
- Spatial spread of recent events
- 69.8 km
- Depth trend
- −2.6 km
19Kermadec Islands Region31°S 177°W · 4 M2.5+ earthquakes in the last 30 days, largest M5.66.4%
Rate-model input: In the clustering-rate model the forecast reads as one input, 68% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 6.4%
- Cell
- 2° × 2° centred on 31°S 177°W, 960 km ENE of Whangarei, New Zealand
- Recent earthquakes in the cell
- 4 M2.5+ earthquakes in the last 30 days, largest M5.6
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 0.52×
- Gutenberg–Richter b-value
- 1.32
- Correlation length
- 51.5 km
- Spatial spread of recent events
- 55.8 km
- Depth trend
- 21.9 km
20Vanuatu Islands13°S 165°E · 150 M2.5+ earthquakes in the last 30 days, largest M6.86.4%
Rate-model input: In the clustering-rate model the forecast reads as one input, 95% of this cell’s 30-day rate comes from its short-term term — every earlier M5+ earthquake nearby since 1973, weighted by how long ago it happened, so large old earthquakes count as well as recent ones — and the rest from its long-term rate. This describes an input, not a breakdown of the chance above.
- Chance of an M6+ earthquake in this cell, next 30 days
- 6.4%
- Cell
- 2° × 2° centred on 13°S 165°E, 680 km SE of Honiara, Solomon Islands
- Recent earthquakes in the cell
- 150 M2.5+ earthquakes in the last 30 days, largest M6.8
Research statistics (not used by the forecast)
Standard seismicity statistics of the recent catalog in this cell, shown for research context only. They do not set the probability, and no test of them as forecasts is published on this site.
- Recent rate vs. this cell’s background
- 7.82×
- Gutenberg–Richter b-value
- 1.05
- Correlation length
- 24.4 km
- Spatial spread of recent events
- 90.4 km
- Depth trend
- −44.3 km
Every cell’s value is in the forecast record: eq_fcst_20261121_0000.json (the probability_grid field, row by row from the south-west corner).
About this forecast
A chance is not an alarm: a 20% chance still means that, of five such 30-day periods, four would pass without an M6+ earthquake in that cell. Compare a cell with others.
- What the model is
- Boosted trees over each cell’s long-term M5+ rate and its aftershock-style clustering after recent earthquakes nearby, combined with GEAR1, a published global model of long-term earthquake rates from crustal strain and past seismicity (Bird et al. 2015)
- How it tested, 2023–2025
- Across every cell, a cell that went on to have an M6+ earthquake was ranked above one that did not 96.9% (95% interval 96.4%–97.3%) of the time (a second look at those years).
- The harder question
- Among recently active cells only, where telling cells apart is hardest: 76.9% (95% interval 73.5%–79.9%).
- Calibration in the test
- It forecast 13% more cells with an M6+ earthquake than there were (ratio 1.13, 95% interval 1.03–1.25).
- Model version
eq_operational_S2_gear1_v2-ecb87f020f99- Full method
- Methods: earthquake
Check this forecast
See what produced it
The forecast’s record: its data sources, model version, and hashes of its inputs and output.
Open the evidenceSee the track record
How this model scored on past years, and how its live forecasts score as their 30-day windows close.
Open the track recordRe-run it yourself
Download this forecast’s frozen record: the published numbers, the model version, and the inputs it used (for earthquakes, the catalog window and its hash).
Download eq_fcst_20261121_0000.json