The Keureuto Dam is located in Blang Pante Village (population 2.145 people and 530 households), Paya Bakong District (population 3.210 people and 570 households), North Aceh Regency, Indonesia (Badan Pusat Statistik Kabupaten Aceh Utara n.d.) (Figure 1). The reservoir extends across several nearby villages, including Plu Pakam (population 1.820 people and 450 households) and Makarti (population 2.055 people and 510 households) in Tanah Luas District, and Rusip (population 257 people; 73 households) and Tembolon (population 234 people; 69 households) in Bandar District, Bener Meriah Regency (Badan Pusat Statistik Kabupaten Bener Meriah n.d.). Geographically, the area lies between 05°01’57.58” N – 04°49’52.97” N and 97°16’19.37” E – 96°58’24.32” E (Variadi et al. 2024). The watershed covers 69.18 km² with a river length of 31.19 km, an inundation area of 896.39 ha, an average annual rainfall of 1619.58 mm and an average discharge of 13.53 m³/s. The situation map of the Keureuto Dam is shown in Figure 1.

The Keureuto Dam is an embankment-type structure with a height of 74 m and a crest length of 386 m. The total storage capacity is 215.94 million m³, divided into effective (167.22 million m³) and dead storage (33.95 million m³). The dam crest elevation is +107.51 m, the normal water level (NWL) +101.51 m and the flood water level +103.86 m. Additional specifications include a 358.8 m evacuation tunnel and a 372 m waterway tunnel.

The inflow discharges corresponding to return periods of 25 years, 1000 years and the probable maximum flood (PMF) are 707.78 m³/s, 1039.77 m³/s and 2680.88 m³/s, respectively. The sediment storage level is at +69.18 m, the low water level at +73.43 m, the normal water level at +101.51 m, the flood water level at +103.86 m and the flood control storage level at +98.01 m. The inundation area is 896.39 ha, with a total storage capacity of 215 million m³ (Balai Wilayah Sungai Sumatera 2022).

The technical specifications of the dam are as follows: It is a zonal-type embankment dam with a main dam height of 74 m and a total crest length of 386 m. The total storage capacity is 215.94 million m³, comprising an effective storage capacity of 167.22 million m³ and a dead storage capacity of 33.95 million m³. The reservoir covers an area of 896.39 ha at the normal water level (NWL). The dam crest elevation is +107.51 m, the intake elevation is +69.10 m, the spillway elevation is +101.511 m, the evacuation tunnel has a length of 358.80 m, and the waterway tunnel has a length of 372.00 m (Balai Wilayah Sungai Sumatera I 2024). The complete technical specifications of the Keureuto Dam are presented in Figure 2.

The basic concept of disaster risk reduction (DRR) encompasses four main components: Risk identification and assessment, which involves analysing hazards, vulnerabilities, capacities and community exposure to disaster threats; mitigation and prevention, which focus on reducing the potential impacts of disasters through spatial planning, the development of disaster-resilient infrastructure and environmental conservation; preparedness and early warning, which include information systems, community training and technology-based early warning mechanisms; and recovery and rehabilitation, which emphasise the principle of Build Back Better to ensure communities become more resilient after disasters (Seddiky, Giggins & Gajendran 2020).

Acceptable risk is the level of disaster risk considered tolerable by society, balancing safety and practicality. It varies according to social, economic, political, cultural, technical and environmental conditions, and guides engineering design, policies and disaster risk reduction measures to keep potential impacts within manageable limits (UN Office for Disaster Risk Reduction 2024).

The United Nations Office for Disaster Risk Reduction (UNDRR) defines DRR as a systematic effort to identify, assess and reduce disaster risks. This approach focuses not only on emergency response, but also on prevention and mitigation efforts before a disaster occurs (ed. Brears 2022). Disaster risk reduction is a systematic approach to identifying, assessing and reducing disaster risks through mitigation, preparedness and community resilience strategies.

The rainfall data used in this study were obtained from the Meteorology, Climatology and Geophysics Agency (BMKG) North Aceh meteorological observation station located in and around the river basin (Table 1). The annual maximum daily rainfall data over a 20-year period were used for hydrological frequency analysis and PMF estimation.

The statistical properties of the rainfall series summarised the number of data as 20 years, mean rainfall 176.1 mm, maximum rainfall 245 mm, minimum rainfall 112 mm, standard deviation 38.5 mm and coefficient of variation 0.22. The rainfall data were subsequently analysed using the Log Pearson Type III distribution to estimate design rainfall for various return periods.

In general, hydrological analysis is a fundamental component of the hydraulic structure design process. Subsequent analyses and design decisions depend heavily on the data and information derived from the hydrological study. The Nakayasu synthetic unit hydrograph method was applied to derive the probable maximum precipitation (PMP) and probable maximum flood (PMF) using the formula (Hershfield 1961) (Equation 1):

where Xm is the maximum rainfall value or PMP; x¯ P is the average of the n-year maximum rainfall data series; Km is the function value of the rainfall duration and the average annual maximum daily rainfall; and Sp is the standard deviation of the n-year maximum daily rainfall data.

Based on the probabilistic method, frequency analysis is conducted to determine the magnitude of the design rainfall. A distribution equality test was conducted to determine whether the results of the frequency analysis were acceptable. The distribution equality test was conducted using probability values of 1% and 5%. In this study, both distributions were used to measure the uniformity of the Chi-square and Smirnov–Kolmogorov distributions. The Hearfield method was used to calculate the maximum rainfall value (PMP) and obtain the design flood discharge (PMF) (Back & Bonfante 2021). Table 2 shows the results of the planned flood discharge calculations.

Manning’s roughness coefficients were assigned based on land-use classifications and established hydraulic references, including Ven Te Chow (1988) and the United States Army Corps of Engineers guidelines. Spatial variation in roughness values was incorporated to represent differences in flow resistance associated with each land-cover type (Table 3).

Higher Manning’s coefficients were assigned to forested and residential areas to account for increased hydraulic resistance caused by dense vegetation, buildings and infrastructure. In contrast, lower coefficients were applied to river channels and open areas characterised by relatively smooth surfaces.

A sensitivity analysis was conducted to evaluate the influence of roughness uncertainty on model outputs. The results indicate that a ± 10% variation in Manning’s coefficient produced changes of approximately 3% – 8% in the maximum inundation depth, with the most pronounced effects occurring in low-gradient floodplain areas.

In sediment failure analysis, the most uncertain parameters include fracture length, size, shape, position and formation time. The empirical equation developed by Froehlich, as documented in the HEC-RAS User’s Manual, has been widely applied in various dam safety assessments and breach analyses (Froehlich 1995).

In 2008, Froehlich conducted a study on 74 dam breach cases involving various types of embankments, including those with impervious (clay) cores and rockfill cores. The objective of the study was to develop empirical equations (Equation 2 and Equation 3) for estimating the average breach width, side slopes and failure time (Froehlich 2008).

where B is the average fracture width (m); t_f is the fracture formation time (second); Vw is the volume at failure (m³); hb is the final fracture height (m); and Ko is a constant (1.3 for overtopping and 1.0 for piping).

Most dam failures result from a combination of factors, including overtopping, piping and foundation failure, each of which contributes to structural damage of the dam (Chen et al. 2023; Zhou et al. 2022). However, in the context of this study, particular attention was given to overflow as the primary cause of dam failure. This focus was chosen because HEC-RAS version 5.0.7, the analytical software employed in this study, is capable of simulating dam failure mechanisms related only to piping and overflow, but not other types of failure. Dam failures frequently occur due to overflow, which happens when water overtops the dam crest as the upstream inflow exceeds the dam’s capacity to manage extreme flood discharges. Moreover, each sediment transport function is highly dependent on site-specific geometry, hydraulic conditions and sediment characteristics, making it unsuitable for generalisation to other contexts (Basri et al. 2020). Furthermore, piping the flow of water through gaps or voids in the dam structure or foundation is a significant cause of dam failure, particularly when sedimentation compromises structural stability.

Model consistency verification was conducted by comparing the hydraulic behaviour of the simulation results with previously published dam-break studies. The inundation distribution pattern, which follows river corridors and low-topographic areas, also demonstrates agreement with similar studies using the 2D HEC-RAS model. This consistency in hydraulic behaviour indicates that the model configuration is capable of realistically representing flood-wave propagation and can therefore be used for dam-break flood hazard analysis. The results of the consistency verification are presented in Table 4.

Moreover, the characteristics of the downstream flood hydrograph can be influenced by various sediment-related failure parameters. The dam breach was modelled in HEC-RAS 5.0.7 using the 2D unsteady flow module. Inputs included dam geometry, reservoir elevation, breach side slopes and Manning’s n roughness (Dam Failure Parameters, Water Overflow Scenario as shown in Table 5).

To provide a visual illustration of this phenomenon, the accompanying Figure 3 illustrates a dam failure caused by water overtopping.

The HEC-RAS 2D model simulated water propagation through the Keureuto River valley under a total breach scenario. The computational mesh resolution was 10 m × 10 m based on Digital Elevation Model Nasional Indonesia (DEMNAS) Geospatial data. Boundary conditions incorporated inflow hydrographs, outflow boundaries and roughness coefficients for land use types. The simulation generated time-series outputs for flood depth, velocity and extent, which were used to map hazard zones for risk mitigation.

A mesh resolution of 10 m × 10 m was selected to achieve an optimal balance between topographic representation and computational efficiency in the dam-break simulations. The topographic dataset was derived from DEMNAS, which has a spatial resolution of 10 m; therefore, the computational mesh was aligned with the native digital elevation models (DEM) resolution to minimise excessive interpolation and potential elevation distortion. This mesh resolution was considered sufficient to represent the morphology of the main river channels and floodplains, capture topographic variations within residential and agricultural areas and simulate stable flood-wave propagation.

Coarser meshes (> 20 m) tend to smooth critical topographic features, thereby reducing the accuracy of inundation depth and flood extent estimates. In contrast, finer meshes (< 5 m) substantially increase computational time and memory demand without providing a commensurate improvement in simulation accuracy relative to the quality of the available DEM data.

A sensitivity analysis was performed using mesh resolutions of 5 m, 10 m and 20 m. The results indicated that the difference in predicted inundation area between the 10 m and 5 m meshes was less than 5%, whereas the computational time for the 5 m mesh increased by more than twofold. Accordingly, the 10 m mesh resolution was adopted as the optimal configuration for the two-dimensional hydraulic simulations.

This article followed all ethical standards for research without direct contact with human or animal subjects.

The sediment breach parameters in HEC-RAS include breach geometry, failure timing and elevation, reservoir water level, and side-slope gradient. The dam material coefficient typically ranges from 0.5 to 0.6, whereas the overflow coefficient ranges from 2.6 to 3.0 (Li et al., 2021; Al-Iraqi, 2025).

The HEC-RAS simulation revealed extensive inundation downstream of the Keureuto Dam (Figure 4). Floodwaters spread rapidly through low-lying villages in Matang Kuli, Paya Bakong, Pirak Timu and Tanah Luas sub-districts. The simulation results illustrating the flood distribution caused by the Keureuto Dam collapse under overtopping conditions are presented in Figure 4.

The maximum flood depth reached 19.577 m at 01:30, followed by 18.807 m at 01:35 and 18.032 m at 01:40. The inundation area was approximately 49.91 km², covering 84 villages. Figure 5 presents the temporal variation in water depth.

Initially, flood depth rose sharply, indicating rapid dam breach erosion, then declined as discharge stabilised. Within about 3 h, depths reduced to below 2 m, and by noon, the floodwater receded to less than 1 m. This pattern is crucial for disaster preparedness, as it provides the time window available for evacuation and emergency actions in downstream communities (Mehta, Weeks & Tyquin 2020).

The dam-break simulation shows high flow velocities near the breach and along the main river channel, causing rapid flood-wave propagation and increasing the risk of severe erosion and damage to bridges, roads and residential buildings.

High flow velocities and deep inundation greatly increase hydrodynamic forces on structures and reduce evacuation time. Areas near the river corridor experience the highest depths and velocities, indicating very high hazard intensity.

The velocity distribution is influenced by topography and land-cover roughness. Flow slows in vegetated and agricultural areas due to higher Manning’s coefficients, while the main river channel maintains higher velocities because of lower hydraulic resistance.

From a disaster management perspective, flow velocity strongly affects human safety, vehicle stability and structural vulnerability. High velocities combined with deep inundation pose major risks to downstream settlements and infrastructure, highlighting the need for early warning systems, evacuation planning and flood-resilient infrastructure.

Figure 6 shows the flow velocity evolution after the breach. Peak velocity reached 1.389 m/s at 00:50, gradually decreasing below 1 m/s after 1:10.

These results indicate that the most dangerous period occurs within the first 90 min of dam failure. Such quantitative insights enable authorities to define critical response times and prioritise early-warning dissemination within that window. Velocity data also assist in assessing structural vulnerability, since buildings exposed to flows above 1 m/s face a high risk of damage or collapse (Elsanadedy et al. 2023; Song et al. 2024).

Hydrological modelling is a representation of hydrological cycle processes used to simulate the movement of water within a watershed. This model includes processes such as precipitation, infiltration, evapotranspiration, surface runoff and river flow. In the context of risk mitigation, hydrological modelling functions to predict extreme events such as floods and droughts, identify disaster-prone areas and support water infrastructure planning and spatial planning.

According to Ven Te Chow et al. (1988), hydrological models are essential tools for understanding hydrological responses to changes in environmental conditions and land use. Hydrological modelling plays a crucial role in disaster risk mitigation, particularly in flood mitigation, where models are used to simulate peak discharge, flood forecasting and the planning of dams and drainage systems. In drought mitigation, hydrological modelling is applied for water balance analysis, prediction of water deficits and irrigation planning. For watershed management, it is used to evaluate land use changes and analyse the impact of urbanisation on runoff.

Overall, hydrological modelling is a highly important tool in disaster risk mitigation, especially for floods and droughts. With its ability to simulate various scenarios, it supports better decision-making in infrastructure planning, water resources management and disaster risk reduction.

Although direct field validation is limited due to the dam’s recent construction, simulated peak discharges and floodplain extents were compared with historical floods in the region and found to be consistent with topographic conditions. Nevertheless, model accuracy may be affected by DEM resolution, simplified breach geometry and absence of sediment transport coupling (Harada et al. 2019; Tazaki, Harada & Gotoh 2022).

Despite these constraints, the integrated hydrological–hydraulic modelling approach provides a reliable framework for risk assessment and scenario planning. Future work should include multi-scenario simulations (e.g. piping, seismic failure) and integration with socioeconomic exposure data to support holistic risk reduction strategies.

To improve the comprehensiveness of the dam-break assessment, a piping failure scenario should be included alongside overtopping, since internal erosion is a common cause of embankment dam failure. Incorporating this scenario would allow comparison of flood characteristics under different failure mechanisms and strengthen hazard assessment and emergency planning.

Alternatively, if the piping scenario is excluded, a stronger technical justification should be explicitly stated. The exclusion can be justified by the structural and geotechnical characteristics of the Keureuto Dam, which indicate a relatively low susceptibility to internal erosion. The dam incorporates an impervious core, graded filter layers and internal drainage systems designed according to modern embankment dam safety standards to prevent seepage-induced erosion. In addition, geotechnical monitoring records and seepage evaluations conducted during the operational period have not indicated abnormal pore-water pressure, excessive seepage discharge or signs of internal material transport that could trigger piping initiation. Therefore, the present study prioritises the overtopping failure mechanism because extreme hydrological loading associated with the PMF represents the most critical and credible worst-case scenario for the Keureuto Dam. Nevertheless, future studies are recommended to incorporate multi-scenario failure analyses, including piping and combined failure mechanisms, to reduce uncertainty and improve dam safety risk evaluation.

The following recommendations are proposed based on the findings of this study:

By implementing these recommendations, government agencies such as the River Basin Agency of Sumatra I and the Aceh Regional Disaster Management Agency (BPBD) can enhance flood preparedness and reduce potential loss of life and property in future dam-related disasters.