When a flood warning goes out, behind that message lies a complex chain of sensors, mathematical models, protocols and human decisions that was set in motion hours or even days earlier. Understanding how that chain works is essential to appreciating what early warning systems do — and to knowing how to react when a warning reaches you.

This article explains the full workings of a flood warning system, from the collection of meteorological and hydrological data to the notification that reaches the public, taking in warning levels and the particular features of the Spanish system.

The warning chain: from data to action

A flood early warning system works as a chain with well-defined links. If any one of them fails, the whole chain loses its effectiveness. The essential components are:

  • Monitoring and data collection, which is observing the phenomenon.
  • Analysis and prediction, through modelling and forecasting.
  • Risk assessment and decision-making, using thresholds and protocols.
  • Dissemination of the warning to the public and the authorities.
  • Response, by the population and the emergency services.

Let us look at each link in turn.

Link 1: monitoring and data collection

Information is the foundation of any warning system. Forecasting floods requires two fundamental types of data.

Meteorological data

Rainfall is the primary trigger of most floods. Weather data comes from several sources:

  • Automatic rain gauges. The SAIH and AEMET networks provide point rainfall measurements in real time, at intervals of 5 to 15 minutes.
  • Weather radar. AEMET’s radar network — 18 operational radars across Spain — estimates rainfall over large areas at kilometre resolution every 10 minutes. Radar does not measure rain directly: it measures the reflectivity of water droplets in the atmosphere, which calibrated algorithms convert into a rainfall rate.
  • Weather satellites. Geostationary Meteosat satellites and polar-orbiting platforms supply cloud imagery, rainfall estimates and atmospheric humidity profiles.
  • Numerical weather models. The ECMWF, GFS and AEMET’s HARMONIE-AROME produce rainfall forecasts days ahead, making it possible to anticipate a risk episode.

Hydrological data

Rain alone does not determine whether there will be a flood. The state of the hydrological system matters just as much:

  • River levels. SAIH sensors measuring water depth in the main channels.
  • Discharge. The volume of water passing a point on the river per unit of time, in cubic metres per second.
  • Reservoir levels. The stored volume and the remaining freeboard determine whether a reservoir can absorb a flood wave or will have to release water.
  • Soil moisture. Soil already saturated by previous rain cannot absorb more, which increases surface runoff and flood risk. This is estimated with models or satellites such as SMOS and Sentinel-1.
Why antecedent conditions matter: Fifty millimetres of rain may cause no problem at all if the soil is dry and reservoirs have room to regulate. The same rain on saturated ground with full reservoirs can produce a flood. That is why warning systems look not only at forecast rainfall but at the prior state of the catchment.

Link 2: analysis and prediction

Raw data has to be turned into useful information through models and analysis. This is where hydrological modelling comes in.

Rainfall-runoff models

These mathematical models simulate how rain falling on a catchment becomes discharge in the river. They account for topography, soil type, vegetation cover, urban sealing and antecedent moisture. Models widely used in Spain include HEC-HMS, developed by the US Army Corps of Engineers, the TETIS model developed at the Polytechnic University of Valencia, and various proprietary models of the river basin authorities.

Flood routing models

Once the generated discharge has been estimated, routing models calculate how the flood wave travels downstream, when it will reach each point and what level it will reach. They matter especially for giving downstream communities time to react.

The forecasting window

A critical concept is lead time: how far in advance a reliable warning can be issued. It depends on the type of catchment:

  • Large basins such as the Ebro, Guadalquivir or Duero. The flood wave takes days to travel, allowing lead times of 24 to 72 hours or more.
  • Medium basins such as the Júcar or the middle Segura, with times of concentration of 6 to 24 hours.
  • Small basins and Mediterranean ravines, with times of concentration of one to three hours, sometimes less. Here the lead time is minimal, and the only way to gain any is to forecast the rain before it falls.
Flash floods: In small, steep catchments — like many ramblas of south-eastern Spain — the flood can arrive within an hour of the rain starting. These flash floods are the hardest to predict and the most dangerous, because they leave so little time to evacuate.

Link 3: warning levels and thresholds

Both AEMET, for meteorological warnings, and the river basin authorities, for hydrological ones, use colour-coded warning levels aligned with the European Meteoalarm system.

Green: no significant risk

No dangerous weather or hydrological phenomena are expected. River levels are within their usual range and no particular action is required.

Yellow: potential risk

Phenomena are expected that are unusual but do not pose a serious danger to the general population: significant rainfall totals, rising rivers still within their channels. The advice is to stay informed and follow developments. For AEMET this typically means 40 to 80 mm in 12 hours, depending on the area.

Orange: significant risk

Dangerous weather or hydrological phenomena are expected that may affect people and property: intense rain with a risk of minor channels overflowing, and localised flooding. Emergency plans are activated. In the Mediterranean region, AEMET thresholds for this level are usually 80 to 150 mm in 12 hours.

Red: extreme risk

Phenomena of exceptional intensity are expected, posing a serious and immediate danger to the population: extraordinary floods, widespread overflowing and the possibility of fatalities. The highest levels of emergency plans are activated, and confinement or evacuation orders may be issued. In the Mediterranean region this means rainfall above 150 to 180 mm in 12 hours.

Thresholds are not uniform: The thresholds for each warning level differ across Spain, adapted to the climate and vulnerability of each area. Sixty millimetres in 12 hours may justify an orange warning in Castile and León but only a yellow one in the Valencia region, where such totals are more routine.

Meteorological versus hydrological warnings

It is worth distinguishing between two related but different kinds of warning:

  • Meteorological warning (AEMET). It warns about forecast rainfall. It is issued before the rain falls, based on prediction models, so it is strongly predictive, and it covers whole provinces or wide areas.
  • Hydrological warning (river basin authorities and SAIH). It warns about the behaviour of rivers and reservoirs. It may be issued in advance, based on hydrological models, or during the flood, based on observed data. It is more localised and specific: it says which reach of which river will be affected.

The two are complementary. The meteorological warning gives a general early heads-up; the hydrological one turns the threat into specifics about which rivers will rise, by how much and when.

Link 4: getting the warning out

A warning that does not reach the affected population is useless. The dissemination mechanisms in Spain include the following.

Traditional channels

  • AEMET warnings, published on the AEMET website and distributed to the media, Civil Protection and social networks.
  • The 112 emergency service, which coordinates warnings at regional level and can issue notices through the media.
  • Sirens and public address systems. Some municipalities have audible warning systems, especially downstream of large dams.

The ES-Alert system

Since June 2023 Spain has had ES-Alert, based on Cell Broadcast technology. It sends mass alerts directly to every mobile phone within a defined geographical area, with no app installed and no registration required. The message appears as an emergency notification with a distinctive tone, even if the phone is on silent.

ES-Alert is run by Civil Protection and is activated when a situation poses an imminent danger to life. Its implementation was one of the requirements of the European Electronic Communications Code.

ES-Alert coverage: The system works with the three main Spanish mobile operators and reaches any compatible handset, which means the great majority of current smartphones. It uses no mobile data and needs no internet connection: the alert goes out through every mast in the affected area.

How WhatAWeather builds its warning view

WhatAWeather combines several data sources to give a complete picture of flood risk.

Integrated data sources

  • Official warnings issued by national authorities, including AEMET for Spain, shown directly on the map.
  • Open-Meteo forecasts for rainfall, temperature, wind and other parameters 48 to 168 hours ahead, refreshed every few hours.
  • Weather radar showing current rainfall and how it has moved over recent hours.
  • River flood forecasts from Open-Meteo’s flood models, covering discharge and overflow risk on the main rivers.
  • Reservoir monitoring, with current levels, trends, percentage full and available regulation capacity.

Threshold analysis

Besides displaying official warnings, WhatAWeather runs its own threshold analysis, checking whether forecast or current conditions exceed reference values. That covers rainfall intensity, wind speed, expected accumulation over the coming hours and river levels.

Link 5: the response

The last link in the chain — and arguably the most important — is how the public and the authorities respond. A technically perfect system is worth little if people do not know what to do.

Civil Protection and emergency plans

In Spain, the response to floods is organised through the Special Civil Protection Plans for Flood Risk, which exist at national and regional level. These plans define activation levels, the responsibilities of each body, the risk areas and the actions to be taken in each scenario.

What the public should do

The basic recommendations when a flood warning is issued are:

  • Stay informed through official sources: AEMET, 112 and Civil Protection.
  • Never try to cross a flooded area, on foot or in a vehicle.
  • Keep away from river channels, ravines and low-lying ground.
  • If you are in a flood-prone area, move to upper floors; never go down to basements or garages.
  • Do not drive on flooded roads: 30 cm of moving water can sweep a car away.
  • Follow the instructions of the authorities and evacuate if told to do so.
A sobering statistic: According to Civil Protection data, most flood fatalities in Spain occur inside vehicles that tried to cross flooded channels or roads. Never underestimate the force of moving water.

Challenges and the future of warning systems

Early warning systems have improved enormously in recent decades, but significant challenges remain:

  • Forecasting convective rainfall. The small-scale intense storms typical of the Mediterranean remain very hard to pin down in space and time.
  • Flash floods. In small catchments the lead time is minimal, which makes research into nowcasting — very short-range forecasting, zero to six hours — a priority.
  • Effective communication. The challenge is not only issuing the warning but making sure people understand it and act. Warning fatigue, caused by repeated alerts that come to nothing, and a weak culture of self-protection are real problems.
  • Machine learning. Learning algorithms offer real potential to improve flood forecasting, particularly in combining multiple data sources and detecting complex patterns.

Conclusion

A flood warning system is far more than a sensor and a siren. It is an integrated chain of technology, science, organisation and communication that requires investment, maintenance and, above all, an informed population ready to act. Tools such as WhatAWeather help democratise access to hydrological information, putting within everyone’s reach data that used to be available only to specialists.

The key is anticipation: every extra minute of warning can save lives. Maintaining and improving early warning systems is therefore not an expense but an investment in safety and resilience against a risk that climate change will only intensify.