Research
Do proposed precursor signals show up before hazards?
Published studies have proposed that space weather, geomagnetic activity or lightning change before earthquakes, rapid hurricane intensification or tornadoes. These tests check those claims against the event record. They are research, separate from the published forecasts, and the results are published whatever they find.
Results
| Question | Sample | Difference before events | Result |
|---|---|---|---|
| Is geomagnetic activity (the Kp index, highest in the 72 hours before) higher before M6+ earthquakes? (Sobolev; Hayakawa) | 200 events, 193 matched times | 0.03 (95% interval −0.25 to 0.31) | No difference found |
| Are there more M- and X-class solar flares in the 72 hours before M6+ earthquakes? (Freund; Pulinets) | — | — | Not run: no data for these events in this window |
| Is the interplanetary magnetic field more strongly southward (most negative Bz, 72 hours) before M6+ earthquakes? | 200 events, 193 matched times | 0.15 (95% interval −0.52 to 0.82) | No difference found |
| Is there more lightning in a hurricane’s core before it intensifies rapidly? (Fierro et al. 2014) | — | — | Not run yet: needs a curated list of events |
| Does a storm’s lightning rate jump in the 15 minutes before a tornado? (Steiger et al. 2007; Schultz et al. 2011) | — | — | Not run yet: needs a curated list of events |
| Do solar storms (coronal mass ejections) arrive more often in the 72 hours before rapid intensification? (Thakur et al.) | — | — | Not run yet: needs a curated list of events |
How to read the results
Each test compares a measurement in the hours before events with the same measurement at matched times without an event (the same calendar dates a year earlier and later). An interval that excludes zero means a difference was found in this sample — it does not show cause, and with several tests run together one may cross that line by chance. An interval that contains zero means no difference was found at this sample size.
None of these signals is used by the published earthquake or hurricane forecasts. NOAA’s storm lightning measurements are among the inputs of the tornado model, which uses them as NOAA reports them.