River Discharge & Flood Forecast Scraper

Get daily river discharge forecasts for any river coordinate on Earth — in m³/s, with the full ensemble spread (mean, median, min, max, p25, p75) so you can see how confident the model is. Up to 210 forecast days ahead and 60 days of history, with no API key and no account anywhere.

The data comes from the Open-Meteo Flood API, which serves the GloFAS (Global Flood Awareness System) hydrological model — the same modelling family that European flood agencies use for early warning.

---

⚠️ Put your coordinates ON a river

This is the single most important thing to know: GloFAS is a river-routing model on a grid of river cells. Coordinates in the middle of a field, a city block far from water, or out at sea will either return tiny/near-zero discharge or no useful values at all.

Pick coordinates that sit on or very near the river channel you care about — a bridge, a gauging station, a river bend. Zoom into a map, drop the pin on the blue line, copy the lat/lon. Values then land in the physically sensible range (a small creek is < 5 m³/s, the Danube at Budapest is ~1,400 m³/s, the Amazon is in the tens of thousands).

---

What you get

One row per location, per calendar day:

Field	Description
location_name	Your label, or "lat,lon" if you did not give one
latitude, longitude	Model grid-cell coordinates actually used
date	ISO date (YYYY-MM-DD)
river_discharge_m3s	Deterministic river discharge forecast, m³/s
discharge_mean	Ensemble mean
discharge_median	Ensemble median
discharge_max	Ensemble maximum — the flood-risk upper bound
discharge_min	Ensemble minimum — the drought-risk lower bound
discharge_p25, discharge_p75	25th / 75th ensemble percentiles
uncertainty_range	p75 − p25 — one-number forecast spread
unit	Discharge unit reported by the API (m³/s)
source	Always open-meteo-flood
scraped_at	UTC timestamp of the run

Rows where every discharge value is null are dropped automatically, so you never pay for empty records.

---

Input

{
  "mode": "forecast",
  "locations": [
    "50.94,6.96,Cologne Rhine",
    "47.50,19.04,Budapest Danube",
    "-3.10,-60.02,Manaus Amazon"
  ],
  "forecastDays": 30,
  "pastDays": 7,
  "maxItems": 1000
}

Input	Type	Default	Notes
mode	select	forecast	Daily river discharge per location
locations	array	2 prefilled rivers	"lat,lon" or "lat,lon,Label" — on/near a river
forecastDays	integer	30	1–210 days ahead
pastDays	integer	0	0–60 days of history prepended
maxItems	integer	1000	Hard cap on pushed rows (1–10,000)

---

Example output

{
  "location_name": "Budapest Danube",
  "latitude": 47.525,
  "longitude": 19.02501,
  "date": "2026-07-28",
  "river_discharge_m3s": 1434.49,
  "discharge_mean": 1441.02,
  "discharge_median": 1434.49,
  "discharge_max": 1602.31,
  "discharge_min": 1380.11,
  "discharge_p25": 1410.55,
  "discharge_p75": 1478.20,
  "uncertainty_range": 67.65,
  "unit": "m³/s",
  "source": "open-meteo-flood",
  "scraped_at": "2026-07-28T13:45:19Z"
}

---

Use cases

• Flood risk monitoring — watch discharge_max against a known bankfull threshold for each site and alert days before water arrives.
• Insurance & reinsurance — screen portfolios of insured locations against rising discharge forecasts; quantify exposure with uncertainty_range.
• Hydrology & research — pull consistent multi-river time series for model comparison, low-flow/drought studies, or reservoir inflow work.
• Logistics & inland shipping — plan barge loading on the Rhine, Danube or Mississippi, where low water means reduced draft and high water means closed locks.
• Adjacent: hydropower scheduling, construction site planning, agriculture and irrigation, water utilities.

---

Pricing

Pay-per-event style economics: roughly $0.50 per 1,000 rows. A 30-day forecast for 10 rivers is 300 rows — about $0.15 per refresh. Schedule it daily and a full river-monitoring feed costs a few dollars a month.

---

Tips

• Run with pastDays: 7 to get context: is the river already rising before the forecast starts?
• A large uncertainty_range relative to river_discharge_m3s means the ensemble disagrees — treat that forecast as low confidence.
• forecastDays above ~30 switches into seasonal-ensemble territory: useful for trend, not for exact timing.
• Batch many rivers in one run — the actor fetches up to 50 locations concurrently.

---

Related actors

• USGS Water Data Scraper — observed US streamflow and gauge heights
• NOAA Weather Alerts Scraper — live US flood and severe-weather warnings
• MET Norway Weather Scraper — global precipitation forecasts (the driver of discharge)
• Earthquake Scraper — global seismic events

---

Notes & limits

• Data source: Open-Meteo Flood API (GloFAS reforecast + forecast). Free for non-commercial and low-volume use; check Open-Meteo's terms for heavy commercial usage.
• Discharge is modelled, not measured. For observed gauge readings use the USGS actor above (US only).
• No login, no API key, no proxy required.