Drought, low river levels and water scarcity may dominate the public debate during dry periods, but they should not push flood preparedness off the agenda. Germany’s experience shows why the two extremes increasingly have to be managed together. A new study by the Federal Institute for Research on Building, Urban Affairs and Spatial Development examines how sensors, digital twins, data platforms and AI can strengthen municipal flood preparedness. For an international audience, the German case is particularly instructive because flood risk management is distributed across federal, state and local authorities – and because many of the challenges it faces are becoming European ones.

When major rivers carry unusually little water, preparing for floods may sound counterintuitive. Yet climate resilience increasingly requires precisely this kind of thinking. A prolonged dry spell does not protect a city from an intense convective storm. Within hours – sometimes minutes – large volumes of rain can hit sealed urban surfaces, dry soils and drainage systems that were never designed for such short-lived extremes.

This is one of the central messages of a 2026 study, Digital Solutions for Flood Preparedness in Heavy Rainfall, published by Germany’s Federal Institute for Research on Building, Urban Affairs and Spatial Development, or BBSR. The study emerged from accompanying research for Germany’s “Model Projects Smart Cities” programme and examines how municipalities can use digital technologies across the full disaster-risk-management cycle.

Its particular focus is important: not only the Rhine, Elbe, Danube or other major rivers, but smaller watercourses and surface-water flooding triggered directly by intense rainfall.

The distinction matters. In Germany, as elsewhere in Europe, flood protection has traditionally been strongly associated with rivers overtopping their banks. Heavy rainfall creates an additional hazard. Water may accumulate or flow across the surface before ever reaching a river. Small streams can rise rapidly, drainage systems can be overwhelmed, and pluvial – rainfall-induced – flooding can occur at the same time as conventional fluvial flooding.

That changes both the geography and the speed of flood risk.

Why German flood governance needs some explanation

For readers outside Germany, the institutional structure is essential to understanding the digital projects now being developed.

Germany is a federal state consisting of 16 Länder. There is no single national authority responsible for every aspect of flood protection and disaster response. Instead, responsibilities are distributed among the federal government, the states, districts, independent cities, municipalities, water authorities, utilities, specialised water associations and, ultimately, property owners.

At federal level, the principal water-law framework is the Federal Water Resources Act – Wasserhaushaltsgesetz, or WHG. Germany used the Act to transpose the EU Floods Directive into national law, including requirements for systematic risk assessment, hazard and risk mapping and flood-risk-management planning. Federal spatial-planning rules also establish principles such as protecting retention areas, considering climate impacts and taking critical infrastructure into account.

Operational disaster protection, however, is constitutionally primarily a responsibility of the Länder. Each state has its own disaster-protection legislation. Districts and independent cities act as important lower disaster-protection authorities, while state flood centres monitor conditions and issue flood-related warnings.

For crises that cross state boundaries, the federal level can support coordination through the Joint Information and Situation Centre of the Federal Government and the Länder, the GMLZ, operated by the Federal Office of Civil Protection and Disaster Assistance.

Municipalities are crucial because German local government enjoys considerable planning autonomy. Cities and municipalities influence land-use planning, development plans, public infrastructure and local climate-adaptation measures. They can therefore integrate heavy-rainfall risks into urban development, commission their own detailed surface-water hazard maps and incorporate the results into emergency planning and public information.

The picture becomes still more complex around smaller rivers. Public-law water and soil associations can be responsible for maintaining and developing watercourses. Municipal utilities and other operators manage sewers, pumping stations, retention basins and wastewater-treatment plants. German law also places a degree of responsibility on property owners: people potentially affected by flooding are expected, within reasonable limits, to take their own preventive measures.

This fragmentation can be a strength because decisions can be tailored closely to local conditions. During rapidly developing heavy-rain events, however, it creates an obvious challenge: water, data and hazards move across administrative boundaries far more easily than responsibilities do.

This is one reason digital integration has become so important.

From flood defence to disaster-risk management

The BBSR study does not treat digitalisation primarily as a collection of smart-city gadgets. It places technology within four phases of disaster-risk management: prevention, preparedness, response and recovery.

Prevention is concerned with identifying hazards, reducing vulnerability and adapting settlements and infrastructure before a disaster occurs. Preparedness covers monitoring, early warning, emergency plans, training and operational readiness. Response is the management of the acute event, when protecting life and limiting damage take priority. Recovery includes reconstruction, evaluation and institutional learning.

Most importantly, this is a cycle rather than a sequence with a clear end. Information gathered during a flood should improve future planning; weaknesses discovered in emergency response should lead to revised preparedness; and recovery should reduce vulnerability rather than simply restore the pre-disaster condition.

Digital technologies can contribute at every stage.

GIS-based analysis can identify exposed areas. Simulations can show where water is likely to flow. Sensors can monitor rainfall and water levels. Dashboards can create operational situation pictures. Digital twins can visualise alternative scenarios. Apps can communicate warnings. Long-term data sets can eventually support machine-learning models and more sophisticated forecasting.

But the study adds an important qualification: whether digital technology actually increases resilience depends on the organisational and institutional structures surrounding it.

A sensor has little value if nobody is responsible for its data. A forecast does not help if it is disconnected from emergency procedures. And a highly detailed simulation changes little if its findings are ignored in land-use planning.

Berlin: using data before the water arrives

Berlin illustrates how digital flood resilience is moving upstream into urban planning.

The city’s SmartWater activities include a web-based planning tool for blue-green infrastructure – measures such as green spaces, infiltration areas, green roofs and other systems designed to retain, absorb or manage rainwater.

Instead of starting with a predetermined engineering measure, planners can use spatial data to identify vulnerabilities linked to heavy rainfall, heat, sealing of surfaces and water protection. They can then examine the feasibility of interventions and model their expected effects.

This is significant because the same urban landscape increasingly has to deal with opposite climatic pressures.

A conventional drainage philosophy aims to remove water as rapidly as possible. A climate-resilient city increasingly asks whether water can instead be absorbed, stored, retained or allowed to infiltrate. The same green or permeable space that reduces surface runoff during a cloudburst can help retain moisture during a dry period.

Germany commonly discusses this approach under the concept of the “sponge city” – Schwammstadt. The BBSR study describes water-sensitive urban development as an attempt to create a more natural urban water cycle and to retain, store, evaporate or infiltrate rainfall rather than simply channel it away.

Berlin also intends to make risk information more accessible to citizens through map-based tools that can indicate local exposure and provide advice on personal preparedness. Behind the visible applications lies an equally important component: a common data architecture using open interfaces and standards to connect different information sources.

The lesson is straightforward. Digital flood protection does not begin with an emergency alert. It can begin years earlier, when a planning authority decides what should – or should not – be built on a particular site.

Dresden: turning a city into a flood scenario

Dresden takes the idea further through an urban digital twin.

The eastern German city faces both rainfall-induced flooding and conventional river flooding from the Elbe. Its approach combines an existing three-dimensional city model with additional information on buildings, roads, drainage infrastructure, waterways, green areas and other hydrologically relevant features.

Rainfall, sensor, radar and modelling data can be integrated to simulate flood scenarios.

For an external observer, the most striking feature is the spatial resolution. A 3D web application is designed to visualise potential inundation depths across the city and down to individual buildings. For the administration, the system can support risk analysis and water-sensitive urban development; for the public, complex flood scenarios become far more understandable than they would be in a specialist hydrological data set.

This illustrates the real potential of a digital twin in civil protection.

A flood gauge provides a number.

An operational decision requires context.

Which road will become impassable? Which neighbourhood is located in a flow path? Which electricity installation, hospital, nursing home or public building might be affected? Which route remains available for emergency services?

Bringing different data layers together can turn a water-level measurement into decision-relevant information.

Not every digital system should meet the same standard

This is also where the security question begins.

A public visualisation platform designed to raise awareness does not need to meet exactly the same requirements as a system used by a fire service or emergency-management centre to make time-critical decisions.

The German study therefore argues that digital solutions should be developed from the specific use case, rather than starting with the technology itself.

A planning tool, citizen-information service and operational warning platform have different requirements for accuracy, latency, validation, availability and cybersecurity. When digital data begins to influence emergency deployments or critical-infrastructure decisions, resilience requirements rise substantially.

That distinction is particularly relevant as municipalities experiment with inexpensive IoT technologies. Efficient communications technologies such as LoRaWAN or NB-IoT can make dense sensor networks economically feasible, but their use does not automatically create a system suitable for every critical function. The BBSR study therefore highlights data quality, availability and IT security as important issues when municipal information systems become operationally significant.

In other words: a successful smart-city prototype is not automatically a crisis-proof system.

Bitburg-Prüm: putting small streams on the digital map

The Eifel district of Bitburg-Prüm demonstrates why Germany is investing in more localised monitoring.

The region was among those heavily affected by the catastrophic flooding of 2021. One of the lessons was that smaller second- and third-order watercourses had not always been monitored systematically even though they can respond extremely rapidly to intense rainfall.

The district has therefore installed 32 autonomous, battery-powered radar water-level sensors on selected smaller watercourses. Measurements are transmitted via LTE-M to an open-source data platform.

An internal dashboard provides authorised users from disaster protection, fire services and specialist departments with additional operational information, while a public dashboard makes current levels available to residents. Data are also supplied to the state of Rhineland-Palatinate and integrated into wider flood-information structures.

The project is now moving from observation towards prediction.

Water-level data are being combined with information including precipitation and soil moisture, while an AI-supported model is being developed to generate flood information several hours in advance.

This is a more credible role for artificial intelligence in flood protection than the notion of an autonomous algorithm somehow “predicting the weather”. AI can help detect relationships within growing quantities of local data and potentially produce earlier indications of dangerous developments – provided the underlying data are reliable and the model is integrated with hydrological expertise.

The more interesting aspect may ultimately be institutional rather than technological.

Bitburg-Prüm is part of a wider cross-municipal cooperation involving authorities in Rhineland-Palatinate and Saarland. A shared modular platform is intended to allow real-time comparison of water levels across administrative borders.

That matters for downstream protection. Rising levels upstream can become visible before the same water reaches the next municipality.

Here the central problem of flood governance becomes obvious again: river basins are hydrological systems, not administrative territories.

Solingen: making warnings more local

The city of Solingen in western Germany highlights another weakness of conventional warning structures: spatial precision.

A regional alert can be completely correct and still be too broad to tell residents in one valley or neighbourhood what is actually happening around them.

Solingen’s topography makes this particularly relevant. Smaller settlements and buildings in valleys can be exposed to rapidly rising local streams even when conditions elsewhere in the city are very different.

The city therefore uses sensors along streams and the River Wupper to observe conditions in very short intervals. Data transmission uses several routes, including fibre, mobile communications and LoRaWAN – an architecture designed to increase resilience if one communication channel fails.

Emergency services receive current local data while citizens can be informed through the city’s “Mensch, Solingen!” app and digital information displays. At the time examined by the study, the app had around 22,000 active users in a city of approximately 160,000 people.

The model is not intended to replace national or state warning channels. It adds local detail.

This distinction is essential for future warning architectures: national systems provide reach; local systems can provide relevance.

When the digital infrastructure itself floods

Digitalisation also introduces a less comfortable question: what happens if the technology supporting crisis management is itself affected by the disaster?

Solingen experienced a warning during the 2021 floods. Water threatened municipal server rooms, creating the possibility that IT systems could fail. Such an outage might have affected administrative services far beyond the immediate flood response, including the payment of social benefits.

That is a classic cascading-risk scenario.

Flooding damages IT infrastructure. The IT outage affects administration. Administrative disruption affects essential services and vulnerable groups.

As cities become more dependent on digital twins, cloud platforms, connected sensors and mobile applications, flood resilience therefore has to include digital resilience.

Systems intended for crisis management require resilient power supplies, communications redundancy, secure access, dependable data and tested fallback arrangements. Cybersecurity and physical climate protection increasingly converge.

A flood-warning platform cannot be considered resilient if its communications fail as soon as the emergency begins.

Data governance becomes part of civil protection

Technical interoperability presents another challenge.

German flood management may involve municipal environmental departments, planners, fire brigades, disaster-protection authorities, water associations, utilities and state agencies. If each actor operates an isolated data environment, digitalisation merely reproduces institutional silos electronically.

Open standards and interoperable interfaces can enable sensor data, geospatial information and official data to circulate between systems.

The BBSR examples show that local measurements can be integrated into state-level flood portals, although the difference between officially validated information and supplementary municipal data must remain transparent.

The technical issue quickly becomes a governance issue.

Who owns the data? Who verifies them? Who may access detailed information about critical infrastructure? Which information should be public? Who operates a platform once temporary project funding expires? Which organisation is responsible when a sensor provides an implausible reading in the middle of an emergency?

A municipal data strategy is therefore not an administrative luxury. As systems move closer to operational civil protection, it becomes part of the safety architecture.

From dashboards to political decisions

Perhaps the most important finding extends beyond emergency management.

Digital flood tools should not only help decision-makers during a crisis. They should influence decisions that determine whether the crisis becomes severe in the first place.

A simulation can show what additional development might do to surface runoff. A digital twin can identify an area where water retention would provide particular benefits. A risk map can reveal that a new critical facility is being considered in an exposed location.

Technology does not make the final decision. Planning remains a political and administrative process involving competing interests.

But digital tools can make the consequences of those choices far more explicit.

The BBSR study therefore recommends using digital information in political and planning processes, particularly where complex risks need to be translated into understandable options.

This may ultimately be more important than another warning app.

The most effective flood emergency is the one whose consequences have already been reduced through better land use, retention, drainage design and resilient infrastructure.

Learning after the water recedes

The district of Hof addresses another part of the cycle: learning.

Its 27 municipalities differ in size, resources, geography and risk exposure. A “Water Transfer Workshop” is being established to analyse local conditions, collect experience and share knowledge across the district. Digital tools and the existing hoferLand.digital twin provide a common platform through which water-related information and adaptation measures can be evaluated.

The example matters because recovery remains one of the less developed elements of digital disaster-risk management.

After a major flood, authorities need more than a physical reconstruction programme. They need to know which assumptions proved wrong, where warning chains failed, which infrastructure was vulnerable and which interventions worked.

Digital damage registers, monitoring data and operational records can create an institutional memory instead of allowing lessons to disappear when a project ends or staff move on.

The BBSR study notes that this feedback stage remains comparatively underrepresented in many smart-city projects.

Future resilience will depend on correcting that imbalance.

The technology is not the flood protection

The German projects ultimately lead to a simple conclusion.

A sensor does not retain water.

A digital twin does not unseal a paved surface.

An app does not stop a flash flood.

An AI model does not enlarge a retention basin.

Digital systems create value when they improve decisions that produce physical, organisational or behavioural change.

Monitoring can identify inadequate drainage. Modelling can support the development of blue-green infrastructure. Sensors can extend warning times. Dashboards can help emergency services deploy resources. Data from an event can improve future planning.

But the actual reduction of vulnerability comes from acting on that information: retaining more water in the landscape, protecting critical installations, adapting land-use planning, improving drainage, exercising emergency procedures and strengthening public preparedness.

The German experience therefore points towards a broader principle for digital civil protection:

technology should follow the risk, not the other way around.

The European perspective: from flood protection to water resilience

Germany is by no means alone in facing this transition.

Its flood-risk-management system already sits within a European legal architecture. The EU Floods Directive requires Member States to identify flood-prone areas, map risks and establish flood-risk-management plans. Germany incorporated those requirements into its Federal Water Resources Act.

The European debate has since moved further.

The European Environment Agency’s first European Climate Risk Assessment identified 36 major climate risks affecting sectors ranging from infrastructure and water resources to health and financial stability. The EEA warns that several risks have already reached critical levels and that adaptation is not keeping pace with the speed at which climate threats are increasing.

This is pushing Europe towards a more integrated understanding of water.

The European Water Resilience Strategy, presented by the European Commission in June 2025, explicitly treats floods, droughts, water scarcity and poor water management as questions not only of environmental policy but also of security and crisis preparedness. Its objectives include restoring and protecting the water cycle, improving water management and building a more resilient and water-smart European economy.

That wider perspective closely mirrors what German cities are beginning to do at local level.

Flood protection can no longer be reduced to higher dikes or larger drainage pipes. Nor can drought policy be considered separately from heavy-rain management. The strategic question is increasingly how landscapes and cities can retain water when it is available without becoming overwhelmed when too much arrives at once.

This strengthens the case for nature-based and blue-green infrastructure, water-sensitive urban planning, restoration of natural retention areas and more intelligent management of drainage and water systems.

But Europe adds another dimension that is particularly important for rivers such as the Rhine, Danube, Elbe, Oder and Meuse: transboundary cooperation.

Just as a flood does not stop at a German district boundary, it does not stop at a national border. Upstream sensor information, rainfall forecasts, reservoir operations, river-basin modelling and emergency warnings can affect communities hundreds of kilometres downstream in another Member State.

The European direction of travel should therefore be towards systems that are interoperable not only between municipal departments but across regions and countries.

That includes common data standards, compatible warning architectures, shared river-basin information and greater integration of climate adaptation, civil protection, water management and critical-infrastructure resilience.

The Commission’s implementation of its Water Resilience Strategy now includes structured dialogues with Member States aimed at accelerating implementation of EU water law and improving flood preparedness.

The German Smart City projects offer useful building blocks for this broader European challenge. Dense local sensor networks can close gaps that national monitoring systems cannot cover. Digital twins can translate abstract climate projections into specific local consequences. Open platforms can link previously isolated datasets. AI can assist forecasting. Apps can sharpen local warnings.

Yet the most transferable lesson is institutional rather than technological.

Europe will not become flood-resilient simply by installing more sensors.

It will become more resilient when information generated in one place can lead to coordinated action in another; when water management follows river basins rather than administrative borders; when urban development takes future flood paths seriously; and when drought and flooding are recognised as connected expressions of a changing water cycle.

That is why low-water periods should not be treated as a pause in flood protection.

They are preparation time.

The best moment to build the next flood-warning network, reassess retention space, protect digital infrastructure or redesign a vulnerable neighbourhood is not when the sirens are already sounding. It is when the river appears harmless – and there is still time to act. [ML]