Combining GIS, hydraulic modelling and public safety: an early warning system for flood risk management in Zemgale Region (Latvia)

The Zemgale region is a lowland area, highly exposed to flooding due to its flat terrain, dense river network and drainage inefficiencies The Lielupe River, one of the main rivers in the area, has a slow course due to the low gradient of its riverbed. The height of its floodplain does not exceed one metre above water level. Downstream of Jelgava, the river drops only 5–10 cm per kilometre, while the riverbed is much lower than the average level of the Baltic Sea. These conditions make the area naturally prone to water accumulation, high groundwater levels and drainage difficulties, especially in urban areas such as Jelgava and in flat rural landscapes.

Climate change is expected to further exacerbate these existing flood-related vulnerabilities. According to the Latvian Environment, Geology and Meteorology Centre (Ministry of Environmental Protection and Regional Development, 2019), by the end of the 21st century (2071-2100), annual precipitation in Latvia is projected to increase by about 13% under RCP4.5 (intermediate greenhouse gas emission scenario) and 16% under RCP8.5 (the worst emission scenario) compared to the 1961–1990 reference period. The strongest seasonal increase is expected in winter, with precipitation projected to rise by 24–37% under RCP4.5 and by 35–51% under RCP8.5 scenario (LVĢMC, 2017). Copernicus Climate Change Service (C3S projections, implemented by ECMWF, and accessible on Climate-ADAPT) also indicate increasing extreme precipitation under both RCP4.5 and RCP8.5 (see for example the maximum precipitation accumulated over five consecutive days and the extreme annual total precipitation) . Recent research on Zemgale further identified heavy rainfall events exceeding 30 mm in 24 hours as an important flood-related hazard and highlighted that changing snow and ice-melt patterns, together with more intense rainfall, may increase both riverine and pluvial flood risk in the region (Veidemane et al., 2026).

July 2024 was the third wettest July in Latvia’s observational record, with 132.1 mm of precipitation, 75% above the monthly norm (LVĢMC – July 2024 Latvian river regime overview). In Jelgava, prolonged rainfall can cause flooding and infrastructure damage, while low topography and high water tables make drainage and stormwater systems difficult to operate. In rural areas, these events can affect settlements, villages and fertile agricultural land, which plays an important role in the regional economy.

The challenge is not only to reduce flood exposure, but also to increase preparedness, forecasting accuracy and institutional response capacity. Extreme rainfall events are often short-term and spatially localised, which makes forecasting difficult and reduces warning times. At the same time, modelling floods in a flat basin with a complex hydrographic network is technically demanding. Forecast accuracy depends on the availability and quality of meteorological and hydrological data, terrain and riverbed modelling, and the capacity to process information rapidly.

A further challenge is the need to move from a local and partly manual early warning approach to a regional, automated and coordinated system. The Jelgava city experience provides a practical starting point. However, regional upscaling requires better data integration, cooperation between municipalities and national monitoring bodies, technical expertise and public trust. False alarms or missed events could undermine citizens’ confidence in the system. For this reason, the adaptation measure focuses on improving flood prediction, automating data flows and providing timely, property-level information to both public authorities and residents.

The case is framed within the European and Latvian policy framework on flood risk management, climate change adaptation and civil protection. At EU level, the Floods Directive establishes a common framework for assessing and managing flood risks, requiring Member States to prepare flood hazard maps, flood risk maps and flood risk management plans. These plans must address prevention, protection and preparedness, including flood forecasting and early warning systems.

In Latvia, flood risk management is implemented through river basin management and flood risk management plans. For the 2022–2027 planning period, Latvia has prepared integrated river basin management and flood risk management plans for the Daugava, Gauja, Lielupe and Venta river basin districts. These plans identify flood risk areas, potential flood damage and losses, and include programmes of measures for flood prevention and flood risk reduction. This is directly relevant for Zemgale, which is crossed by the Lielupe and Daugava river systems.

The Latvian National Plan for Adaptation to Climate Change until 2030 provides the national climate adaptation framework. The plan identifies climate-related risks for civil protection and disaster management, including extreme precipitation, changes in river run-off and flood-related risks. It also includes specific indications on early warning systems. It states that early warning should be responsive, regularly updated and able to ensure high-resolution modelling for forecasting the impacts of extreme weather conditions.

The measure is also consistent with Latvia’s civil protection framework. Under the Civil Protection and Disaster Management Law, disaster risk assessment includes risk identification, analysis and evaluation. The law recognises the right of people to receive early warnings and recommendations on how to act in case of disaster or threat. In this context, a flood early warning system supports both institutional preparedness and public risk communication.

At regional level, the Zemgale Planning Region Development Programme 2021–2027 and the Zemgale Planning Region Climate Change Adaptation Plan until 2030 provide the planning context for regional climate change adaptation.

The main objective of the measure is to enhance preparedness for flood risk management in the Zemgale region by upscaling the local flood early warning system in Jelgava City to the regional level and by introducing predictive flood modelling. The measure supports early flood warnings, decision making and civil protection through a regional system for real-time information exchange and analysis.

A specific objective is to enable the evaluation of flood risks before an event occurs and to provide public institutions and citizens with the information needed to select appropriate measures to protect people, health, the environment and economic activities effectively.

The system also improves cooperation between municipalities and national monitoring bodies, extends the flood risk early warning system to the entire region, and improves communication to decision makers and citizens.

The adaptation solution consists of an Early Warning System for Flood Risk Management. It combines flood modelling (HEC-RAS hydraulic model) with Geographic Information System (GIS)-based information, including terrain and river geometry data, hydrological forecasts, cadastral and address data, and flood maps, together with decision-support and warning functionalities (IMPETUS, Deliverable D4.3 – Enabling rapid and far-reaching change by Innovative Technologies.

The system has been designed to address the challenges related to flood risk prediction and preparedness in a region characterised by flat terrain and complex hydrological dynamics. The system uses the HEC-RAS 2D hydraulic model to simulate water flow in two dimensions. This model is used for flood modelling and forecasting and can be integrated with GIS, supporting data processing and visualisation. The system automatically reads hydrological forecast data from the Latvian Environment, Geology and Meteorology Centre’s forecast system.

The modelling results are used to identify potential flood areas. The system not only supports institutional decision-making but also directly informs the population, contributing to improved preparedness and response capacity. It currently covers approximately 93,000 households (around 225,000 residents) in the Zemgale region.

The web-based application has high social relevance since it allows residents to register and select up to three locations, such as a home, workplace or children’s school, in order to receive targeted alerts about potential flooding.

The system can currently predict floods one to two days before they occur. When a risk is identified, trained experts are alerted. They monitor the situation, using drones when needed to capture real-time images of threatened areas. These images are shared with the Civil Protection Commission of the affected cooperation territory which is also responsible for preparing local civil protection plans this commission assesses the risk and decides whether to prepare evacuation points and procedures or continue monitoring if the threat is limited.

Besides addressing emergencies, the system stores flood data that are relevant for long-term analysis. In the near future, this will allow authorities to assess how often floods occur, which areas are most affected and how risks are changing over time. These insights feed into Jelgava’s Climate Change Adaptation Plan that was developed in parallel and in coordination with the early warning system. The plan is expected to orient future choices for flood management. Such options include nature-based solutions such as wetland restoration.

A future expansion of the early warning system is foreseen. This can be done through AI-driven algorithms connected with the national monitoring system and other databases. AI can integrate multiple data sources and modelling processes into a coherent system capable of simulating flood events over extended periods and at high spatial resolution.

At the same time, the system is flexible. Based on the previous experience in Jelgava, it is designed to be further improved and potentially up-scaled. Potential future developments include extending the system to the national level and enhancing modelling capabilities to address rainfall-induced flooding beyond riverine contexts. Another possible future development might be making upgraded datasets accessible as open data to experts and society.

However, the robustness of the system is currently influenced by limitations in forecast accuracy. Those depend on the quality of meteorological data, terrain representation, and hydrological assumptions.

Since the system became operational in 2025, monitoring data on its actual use from residents and institutions alongside its performance is not available yet. However, they will become available at a later stage, once the system has been tested over a sufficient period of operation. To date, the system was validated through simulations, assessing its functionality and performance under modelled flood scenarios.

The early warning system requires continuous refinement and maintenance, particularly in relation to:

• improving the quality and density of input data (e.g. riverbed surveys, sensors),
• calibration of hydrological and hydraulic models,
• Optimisation of numerical settings to balance accuracy and computational cost.

These aspects indicate that the quality of the solution is not static. It requires ongoing technical updates to maintain and improve performance over time.

The early warning system is an outcome of collaboration between several actors. The Jelgava Digital Centre is a municipal digital institution that coordinates civil protection processes and developed the web application that acts as early warning system for the public. The Zemgale Planning Region is an institution which embraces six municipalities and coordinates the regional development and climate change adaptation planning under the supervision of the Ministry of Environmental Protection and Regional Development. It was responsible for development of HEC-RAS 2D hydraulic model. The Latvian Environment, Geology and Meteorology Centre collects, stores and provides environmental information, including the meteorological and hydrological data used by the model to predict flood areas. The Baltic Environmental Forum works on the interface between science, civil society, policy and administration. It coordinated stakeholder engagement for the development of the early warning system and the adaptation plan.

Stakeholder participation has been central to the development of the flood early warning system. The stakeholder engagement strategy involved groups representing ‘quadruple helix' categories (academia, government, industry and community) and included representatives from local, regional and national institutions and organisations. The process included informing, collaborating, exchanging information, consulting and co-creating of the solution. Experts from various fields were also consulted and involved in information exchange on flood risk prediction.

Success factors

Several factors supported the development and implementation of the early warning system. These include the strong interest of local and regional stakeholders in improving flood risk management, the growing awareness of climate-related flood risks, and the availability of good practice examples, such as digital twins.

A key enabling factor is the cooperation between the Jelgava Digital Centre, the Zemgale Planning Region, the Latvian Environment, Geology and Meteorology Centre, and the Baltic Environmental Forum. The system also builds on the sensor-based application already under operation in Jelgava City. This initial warning system relied primarily on river sensors to measure water levels and a simple notification mechanism to alert residents. While this approach proved valuable at the local level, its functionality was limited in scope and lacked predictive capacity. This previous experience provided a consolidated technical and knowledge base for extending the system to the wider Zemgale region.

The solution has clear upscaling potential. It is designed to be extended from Jelgava City to the Zemgale regional level and to be connected with the national monitoring network. The experience gained through the project, including handbook, can also support other municipalities in setting up and operating flood early warning systems.

Limiting factors

At the same time, some limiting factors remain. They mainly concern data availability, cooperation with municipalities, technical capacity and public trust. Potential barriers include incomplete datasets and difficulties in reaching agreement with local municipalities on data storage and related services.

Limited availability of highly skilled experts to operate HEC-RAS-based prediction systems at municipal level may also condition its operation, alongside its dependence on external GIS-based data sources. Mistrust among local residents towards warnings received through the early warning system is also possible, especially if false alarms are generated.

The action has received funding from the European Union’s Horizon 2020 research and innovation programme (IMPETUS project). No information on the specific costs of the measure is available

The main benefit is the extension of early warning information to a large number of residents and households. The system covers approximately 93,000 households, corresponding to around 225,000 residents in the Zemgale region. Thus, the measure is not limited to institutional users, but supports public preparedness directly by helping residents and authorities receive targeted information before flood events. Allowing citizens to select specific locations for targeted alerts, such as schools, homes, or workplaces, provides clear social value by directly addressing community needs.

From an economic perspective, the relevance of the measure is linked to the role of Zemgale in the Latvian economy and to the exposure of both urban and rural areas to flooding. In 2023, Zemgale generated EUR 3.1 billion of GDP, corresponding to 7.8% of Latvia’s GDP, with GDP per capita of EUR 13,626. Based on high soil fertility, the region has a strong agricultural and bioeconomy profile. Improved flood warning can therefore not only support the protection of households, infrastructure, and public services. It also benefits economic activities, including agriculture and food-related value chains.

The measure provides benefits for public planning and investment decisions. By improving the availability of flood risk information, it can help municipalities and regional authorities prioritise preparedness actions, emergency response and future adaptation measures. Over time, the collection and use of flood-related data can support a better understanding of flood frequency, exposed areas and changing risk patterns. This would, help to guide climate adaptation planning and the selection of suitable structural or nature-based measures.

The early warning system was developed over the four years of the IMPETUS project, funded under EU Horizon 2020 (2022-2025) and ended accordingly with the scheduled project time.

As the measure is a digital and operational service, its lifetime should be understood as continuous rather than limited to a predefined period.

The system is expected to remain operational as long as it is maintained, updated and used by the responsible authorities. Its continuity depends on regular technical maintenance, updating of input data and models, and coordination among the institutions involved. Over time, the system may be further improved to increase forecast accuracy or to adapt its functionalities to changing needs.

Marija Stocka
Manager at Project Management Division
Information Technology Development Department, Jelgava City Municipality Institution “Jelgava Digital Centre”

Mārīte Ieviņa
Head of Project Management Division
Information Technology Development Department, Jelgava City Municipality Institution “Jelgava Digital Centre”

jdc@jelgava.lv

Handbook for know-how to stakeholders on Early warning in case of floods

Adaptation pathways to address climate change-induced flood risk in the Zemgale region, Latvia

Climate Change adaptation plan in Zemgale region