HazardPulse
Classic
Global Storm Tracker
All active hurricanes, typhoons, and cyclones worldwide ranked by rapid intensification probability. Every forecast links to the evidence that produced it - so you can verify it yourself.
Global tropical cyclone map
Basin boundaries shown. Storm markers appear automatically when tropical cyclones are active.
Active tropical systems
All active tropical systems across every ocean basin, ranked by rapid intensification probability. RI = chance of gaining 30+ knots in 24 h. Data sourced from NHC ATCF advisories and scored by hp-hurricane-ri-v8.1.
| Rank | System | Basin | RI Probability | Wind | Trend | Status |
|---|---|---|---|---|---|---|
| 1 | Hurricane Alice | Atlantic | 37.2% | 75 kt | ↑ 2.4% | Checks passed |
|
Location19.8°N, 64.5°W — east of the U.S. Virgin Islands, tracking WNW at 14 mph
Confidence range31.0% – 43.5%
Why risingSST 28.7°C in storm path; shear dropping to 10 kt; OFCL forecasts +20 kt in 24 h
Deep dive →
| ||||||
TEST DATA — This is a synthetic test hurricane for display verification. Not real. Source: live-storms.json
#1 — Hurricane Alice deep dive
The basics
Hurricane Alice is east of the Virgin Islands, heading toward the Caribbean. It has 86 mph winds (Category 1) and is getting stronger.
If that happens, Alice could jump to a Category 3 hurricane (130+ mph) by tomorrow. That's a big jump in a short time.
Where is it going?
Alice is moving west-northwest at 14 mph toward the northeastern Caribbean. The map above shows what different weather models predict — each line is a different model's best guess.
Most models agree: Alice will get stronger over the next 2 days. The warm ocean water (28.7°C) and calm upper atmosphere are giving Alice ideal conditions to intensify.
The spread between model tracks shows how much uncertainty there is — the wider the "spaghetti," the less sure we are about the exact path.
How strong will it get?
Our model and top weather models all predict Alice will strengthen. Here's what each one says:
| Model | Tomorrow | 2 days |
|---|---|---|
| HazardPulse v8.1 | 115 mph | 138 mph |
| NHC Official | 110 mph | 127 mph |
| HWRF | 115 mph | 138 mph |
| GFS | 98 mph | 115 mph |
| European | 110 mph | 132 mph |
Bottom line: HazardPulse v8.1 predicts 115 mph by tomorrow and 138 mph in 2 days — the most aggressive forecast. Our model sees a 37.2% RI probability by combining signals across all 8 guidance models.
What to do
This is a test storm — no action needed.
In a real scenario, if you were near Alice's path:
- Check nhc.noaa.gov for official warnings
- Know your evacuation zone and route
- Stock up on water, food, and medications
- Charge your phone and have a battery radio
HazardPulse is for research only. Always follow official NHC guidance for safety decisions.
Core parameters
SST and ocean heat
Wind shear and environment
Intensity spaghetti — model-by-model forecasts
Our v8.1 ensemble plus NHC guidance models. HazardPulse ingests all models as features to compute RI probability.
| Model | Now | 12 h | 24 h | 36 h | 48 h | RI? |
|---|---|---|---|---|---|---|
| HP v8.1 | 75 | 85 | 100 | 115 | 120 | 37.2% RI |
| OFCL | 75 | 80 | 95 | 105 | 110 | +20 kt |
| HWRF | 75 | 85 | 100 | 115 | 120 | +25 kt |
| HMON | 75 | 80 | 90 | 100 | 105 | +15 kt |
| GFS | 75 | 75 | 85 | 95 | 100 | +10 kt |
| ECMWF | 75 | 80 | 95 | 110 | 115 | +20 kt |
| SHIP | 75 | 80 | 90 | 95 | 95 | +15 kt |
| DSHP | 75 | 85 | 100 | 110 | 115 | +25 kt |
| LGEM | 75 | 80 | 90 | 100 | 100 | +15 kt |
RI threshold is 30+ kt in 24h. No single model predicts RI by itself, but HWRF (+25 kt) and DSHP (+25 kt) are close. Our v8.1 ensemble sees the aggregate signal: 37.2% chance the actual outcome exceeds what any individual model predicts. The consensus 24h range is 85–100 kt (spread: 15 kt).
v8.1 ensemble breakdown
| Ensemble member | Raw P(RI) | Weight |
|---|---|---|
| Histogram GBT depth-3 | 38.1% | 0.28 |
| Histogram GBT depth-4 | 36.9% | 0.26 |
| 50-bag logistic (ff sweep) | 37.4% | 0.25 |
| L2 logistic | 35.8% | 0.21 |
Calibration metrics
| Predicted | Observed | Count | Error |
|---|---|---|---|
| 0–10% | 5.1% | 4,871 | +0.1% |
| 10–20% | 18.4% | 1,423 | +3.4% |
| 20–30% | 27.3% | 892 | +2.3% |
| 30–40% | 33.8% | 614 | −1.2% |
| 40–50% | 46.2% | 387 | +1.2% |
| 50+% | 58.3% | 201 | +3.3% |
Alice's 37.2% falls in the 30–40% bin where the model has 33.8% observed — well calibrated (error −1.2%). When the model says ~37%, historically about 34% of storms actually undergo RI.
Gate status — quality checks
| Gate | Status | What it checks |
|---|---|---|
| G0 Schema | pass | Fields present, types valid |
| G1 Freshness | pass | NHC ATCF feed within threshold |
| G2 Model Pinned | pass | hp-hurricane-ri-v8.1 matches registry |
| G3 Provenance | pass | Input/output hashes verified |
| G4 Calibration | pass | Brier 0.034, BSS +0.29 above floor |
| G5 Spatiotemporal | pass | Storm in valid Atlantic basin |
| G6 Harm Guard | pass | No alarmist language |
| G7 Explanation | pass | 4+ factors listed |
| G8 Replayability | pass | Replay artifact verified |
| G9 Projection | pass | Disclaimer attached |
| G10 Security | pass | Headers valid |
| G11 Performance | pass | Payload within budget |
| G12 Trust Sync | pass | Badges match gate state |
Evidence fingerprint
hu_fcst_20260315_1200AL012026hp-hurricane-ri-v8.1Recent trend — Alice RI probability
Last 8 forecast cycles showing probability trajectory. Steadily rising as ocean conditions improve along forecast track.
| Cycle (UTC) | Probability | Range | Change | Gate |
|---|---|---|---|---|
| Mar 15 12:00 | 37.2% | 31.0 – 43.5 | +2.4 | pass |
| Mar 15 06:00 | 34.8% | 29.0 – 41.0 | +1.7 | pass |
| Mar 15 00:00 | 33.1% | 27.5 – 39.0 | +1.3 | pass |
| Mar 14 18:00 | 31.8% | 26.2 – 37.8 | +1.3 | pass |
| Mar 14 12:00 | 30.5% | 25.0 – 36.5 | +0.7 | pass |
| Mar 14 06:00 | 29.8% | 24.3 – 35.8 | +1.3 | pass |
| Mar 14 00:00 | 28.5% | 23.2 – 34.2 | +1.4 | pass |
| Mar 13 18:00 | 27.1% | 22.0 – 32.6 | — | pass |
What changed and why
Plain-language summary of what's driving the numbers since last update.
System health
Model lineage
Evidence and replay
See the evidence
Every forecast links to the exact data and model that produced it. Nothing is hidden.
Browse evidenceCheck our track record
How often are we right? We publish accuracy scores publicly, broken down by basin and intensity.
See accuracyReplay any forecast
Download the input data and re-run any past forecast yourself. Same data in, same result out - guaranteed.
Download replay