All articles

Where EarthPulse data comes from

4 min read · updated 13 July 2026

EarthPulse doesn’t generate any observations of its own — the craft is in fusing the world’s best open feeds into one living picture, honestly. This page is the full inventory: what each source provides, its cadence, and what we do to it on the way to your screen.

NASA EONET — the event backbone

The Earth Observatory Natural Event Tracker curates natural events — wildfires, severe storms, volcanic activity, sea and lake ice — from NASA and partner agencies, with position tracks over an event’s lifetime. We poll it every two minutes and normalise every event into a common model: position, track history, magnitude where available, category, sources. Storm tracks you see arcing across the globe come from EONET’s multi-day position histories.

USGS — earthquakes within minutes

The United States Geological Survey publishes global seismic events in near real time; a significant quake typically appears within minutes of the shaking. We carry every event of magnitude 4.5 and above from the last 30 days — the threshold at which detection is globally reliable — including hypocentre depth and tsunami advisories where issued.

NASA FIRMS — fires seen from orbit

The Fire Information for Resource Management System distributes raw satellite fire detections from the MODIS and VIIRS instruments — tens of thousands of burning pixels a day. We cluster them on a ~55 km grid, require multiple confident detections per cluster (which filters agricultural burns and industrial flares), weight each cluster’s position by fire intensity, and report combined fire radiative power in megawatts. Selection is per-region, so a ferocious African burning season can’t crowd a Spanish wildfire off the map. Where a FIRMS cluster overlaps an officially tracked EONET fire, the named event wins.

Open-Meteo — weather, climate, air, places

Open-Meteo aggregates national weather models into a single open API. From it we draw current conditions and forecasts (refreshed every ten minutes), air quality (European AQI with component pollutants), worldwide place search, and — the quiet star — a ten-year historical archive from which we compute the seasonal baseline for any coordinate, powering every “above the seasonal norm” statement and the Heat Watch leaderboard. Reverse geocoding for dropped pins comes from BigDataCloud.

What we promise about the pipeline

Sources are fault-isolated: any one feed going dark never blanks the others. Everything is cached at honest lifetimes — minutes for live events, a full day for climate normals — so we’re gentle on the public services we depend on. And nothing is embellished: if a number is on the globe, it traces back to one of the agencies above. When data is unavailable, we say so rather than interpolate.

Keep reading

Watch it live on the globe →