Early warning & Preparedness
On-the-ground impact from online media, alongside your forecasts and warnings. See whether a flood is happening, where, and how it affects people, in real time and as a record afterwards.
On-the-ground impact, as it happens and as a record
Meteorological and hydrological services need live situational awareness as an event unfolds, and a way to verify whether their warnings matched what happened on the ground. That impact signal sits in social and news media, but checking it by hand is laborious and incomplete, so important reports are missed.
FloodTags collects, classifies and locates that media automatically, turning it into a fast, reliable and operational source of impact information. Independent, real-time validation of impact forecasts and warnings is a recognised gap, and the same data builds the ground-truth record the sector needs for its move to impact-based forecasting. Coverage is global, though the volume and type of media vary by country, so the service is set to your region, hazards and languages at the start.
How it's used
Situational awareness
Watch the live picture build across your region and see on-the-ground impact, whether or not the event was forecast. Forecasters browse the live map and the underlying reports for immediate awareness, rather than waiting for impact to be confirmed through official channels.
Real-time forecast validation
Confirm forecast accuracy as it happens, catching both false positives, where impact was forecast but is not occurring, and false negatives, where impact is occurring but was not forecast, while the event is still unfolding rather than only in hindsight afterwards.
Post-event verification
After an event, produce verification statistics and a post-event report that show how warnings performed against what actually happened, comparing what was forecast against what was observed. These feed back into forecasting-model review, so each event informs and improves the next.
Impact datasets and event databases
Build multi-year impact records from the same data, capturing where and how past events affected people and infrastructure. These datasets support the validation and recalibration of specialised hazard-impact models, and are available as a one-off tailored analysis on request.
One subscription, used live or as reports
For early warning and preparedness, clients most commonly use Operational Mode: the full live environment, with an interactive map, the underlying posts and tags, queries and reports on demand. The same subscription can also run in Reports Mode, which delivers the event reports and summaries on the rhythm you choose, from a quick summary to a detailed report with map, timeline, statistics and photos; the upgrade to live is always available. Tailored analyses, such as a multi-year Event Time-series, are a separate track on request.
This is our early warning and preparedness configuration, a ready starting point you can adjust to your preference.
- Real-time tag & event monitoring
- Event detection
- Reports, from summaries to detailed
- Confidence score
- Response-time guarantee
- Data export / feed
- API access
- Data retention
What it tells you: impacted locations and event times, event classification, failure of protection measures, impact on people and households, disruptions to logistics and infrastructure, and impact on critical infrastructure.
Validating warnings at the Flood Forecasting Centre
At the Met Office, FloodTags runs in the Services Operation Centre in Exeter, supplying filtered, geolocated impact information in near-real time during extreme weather. Hydrometeorologists at the Flood Forecasting Centre and Civil Contingency Advisors monitor live, gaining awareness whether or not a flood was forecast. When a warning is out, they verify the forecast in real time, checking for both false positives, where impact was forecast but did not occur, and false negatives, where impact occurred but was not forecast. Urgent observations are shared to an internal Teams channel reaching around 650 staff. After an event, the team produces verification statistics and post-event reports that feed back into forecasting-model review, complementing official forecasts and warnings.