Strengthening Dam Safety Under Climate Change: A Risk-Informed Overtopping Assessment

Abstract

Climate change is intensifying hydrological extremes, posing growing threats to the safety and operational reliability of embankment dams worldwide, particularly those in regions susceptible to heavy rainfall and flooding. This study evaluates the overtopping risk for Batu Dam, a critical flood mitigation and water supply structure near Kuala Lumpur, Malaysia, under future climate scenarios, with the aim of informing risk-informed dam safety strategies. Using historical hydrological data (1975–2020) and downscaled climate projections from the CMIP5 database under three Representative Concentration Pathways (RCP4.5, RCP6.0, RCP8.5), we conducted flood routing simulations and probabilistic risk assessments employing the iPRESAS software. Our results demonstrate that the annual probability of overtopping increases substantially under higher-emission scenarios, reaching up to 0.08% by the late century under RCP8.5, driven by increased frequency and intensity of extreme rainfall events. These projections highlight significant spillway capacity limitations and underscore the heightened risk of downstream consequences, including economic losses exceeding RM 200 million and potential loss of life surpassing 2900 individuals in worst-case scenarios. The findings confirm the urgent need for both structural adaptations, such as spillway expansion and crest elevation, and non-structural measures, including enhanced real-time monitoring and early warning systems. This integrated approach offers a robust and replicable framework for strengthening dam safety under evolving climate conditions.

1. Introduction

Dams are vital for water resource management, offering crucial services such as water delivery, flood mitigation, and energy production [1]. However, the escalating effects of climate change are progressively jeopardizing the safety and operational efficacy of embankment dams [2]. The hazards are particularly significant in areas susceptible to high rainfall and flooding, where alterations in climate patterns undermine the dams’ design and safety measures [3]

Climate change poses significant threats to the safety and reliability of dams, with extreme weather events and changes in precipitation patterns posing substantial risks of overtopping and potential failure [4]. As existing regional climate scenarios continue to vary, particularly in the case of extreme precipitation within small catchment areas, careful interpretation of climate modeling results is essential from a dam safety standpoint [5]. This study presents a comprehensive quantitative assessment of the impacts of climate change on the safety of the Batu Dam, integrating the various projected effects on each component of the risk, from the hydrological aspects to the dam break flood consequences of the outflow hydrograph. As climate change increases rainfall uncertainty, embankment dams are especially prone to overtopping because they are made from materials that erode easily and cannot resist uncontrolled water flow.

Embankment dams, constructed using compacted soil and rock materials, are particularly susceptible to overtopping failures [6,7]. Unlike concrete gravity dams that can withstand a certain degree of overtopping, embankment dams are prone to rapid erosion and structural collapse when subjected to uncontrolled water flow over the crest [8]. The potential for dam overtopping is a crucial concern, especially for embankment dams, which are more vulnerable to structural damage and failure when subjected to extreme events [7]. Overtopping, where water exceeds the height of the dam crest, leads to erosion and could potentially result in catastrophic dam failure [9]. This issue is especially acute for high-hazard dams, where the consequences of failure extend to the loss of life, property damage, and severe environmental impacts [10].

Overtopping has been identified as a leading cause of dam failure worldwide, particularly for embankment dams, which are more vulnerable to erosion and structural damage [11]. The increasing intensity of extreme rainfall events under future climate change scenarios poses significant challenges for dams initially designed to handle lower, less frequent flood events [4]. Traditional dam safety assessments that rely on fixed design flood levels often underestimate the risk of overtopping, as they do not account for the growing variability of weather patterns caused by climate change [12].

Climate change forecasts indicate a substantial rise in extreme weather occurrences worldwide and in Southeast Asia [13]. Climate change has exacerbated the frequency and intensity of extreme weather events, posing significant challenges for the management and operation of critical infrastructure like dams [14]. As the climate continues to shift, the historical hydroclimatic conditions upon which these systems were designed are becoming increasingly obsolete, necessitating a reevaluation of their resilience to emerging threats [4]. As climate change intensifies, the assumption that historical hydrological conditions would accurately reflect future ones is increasingly inaccurate [15]. Research has continuously indicated that existing dams constructed using historical hydrological data may be inadequate for addressing future fluctuations and intensifications in rainfall and inflows [16].

Globally, there has been a growing body of research analyzing how climate change exacerbates these risks by increasing runoff, river flows, and altering rainfall patterns. For example, a study conducted on the Eugui Dam in Spain indicated that future projections under the RCP8.5 scenario could result in significantly higher water levels during flood events, which would greatly increase the overtopping risk [15]. Similarly, ref. [17] found that future climate scenarios in North Carolina could lead to streamflows several times higher than historical levels, increasing overtopping risk for regional dams. These examples demonstrate the critical need to integrate climate change projections into dam safety assessments globally.

The dam’s safety is jeopardized by the progressive effects of climate change, especially the hazard of overtopping when inflows surpass the dam’s capacity, resulting in water cascading over the dam crest [3,15]. Overtopping can lead to rapid erosion and structural collapse of embankment dams, potentially resulting in catastrophic failure and devastating consequences for downstream communities and ecosystems [18,19]. Recent studies emphasize the importance of rigorous risk assessments and the implementation of advanced overtopping protection measures [20]. For instance, ref. [15] highlights that extreme precipitation events, compounded by climate change, can dramatically increase overtopping risks, necessitating more robust safety protocols.

Recent research on the quantification of climate change impacts on dam safety has utilized a comprehensive, multidisciplinary approach that integrates the various projected effects on the risk components, from hydrology to consequences [15].

To address this issue, dam owners and operators must carefully assess the potential impacts of climate change on the frequency and magnitude of extreme hydrological events that could trigger overtopping [7]. A comprehensive, risk-informed approach is essential to quantify the overtopping risks and develop appropriate adaptation strategies [21]. This may involve implementing a range of interventions, such as increasing the capacity of existing spillways, raising the height of dam crests, or exploring alternative flood management measures that can effectively mitigate the heightened risks posed by climate change [4].

Probabilistic risk assessments offer a more robust framework for understanding and managing overtopping risks compared to deterministic models [22]. These assessments consider a range of potential flood events and their associated probabilities, providing a more comprehensive evaluation of dam safety [23]. For instance, Bowles et al. emphasize the need for probabilistic assessments to account for the multiple potential failure modes, including overtopping, especially under climate-induced extreme events [24]. This approach is particularly valuable for high-hazard dams, where the consequences of failure are severe [10].

By incorporating a range of potential flood scenarios and their associated probabilities, probabilistic risk assessments can help dam operators and decision-makers better understand the true scope and magnitude of the overtopping threat [25]. This informed understanding is crucial for developing robust and adaptive mitigation strategies, emergency response plans, and contingency measures to protect vulnerable downstream communities. A key aspect of probabilistic risk assessment is the use of the General Risk Formula, which provides a structured framework for quantifying risk as Equation (1).

Risk (R) = Probability of Occurrence (P) × Consequence (C),

(1)

Probability of Occurrence (P): this represents the likelihood of a specific event happening; Consequence (C): this refers to the potential impacts and damage that could result from the failure.

Risk analysis softwares for dam safety plays a crucial role in ensuring the integrity and reliability of dam structures, which are vital for water management, flood control, and hydroelectric power generation [25]. This specialized software employs advanced algorithms and modeling techniques to assess potential risks and vulnerabilities associated with dam operations and environmental factors [26].

Probabilistic tools such as iPRESAS (developed in Spain), @Risk (developed in the USA), and DAMRAE (also developed in the USA) provide valuable frameworks for supporting dam safety risk assessments, particularly for overtopping risks. However, they differ in terms of focus and limitations. iPRESAS specializes in integrating climate change projections, making it effective for extreme event analysis, though it requires extensive data and expertise, limiting its accessibility [15]. @Risk, while versatile for modeling uncertainties, lacks dam-specific features and requires customization, increasing its complexity [27]. DAMRAE excels in event-tree modeling to identify failure pathways but depend heavily on accurate probabilistic data, often challenging in data-scarce regions [28]. Despite their strengths, these tools face challenges like computational intensity, data demands, and usability issues. iPRESAS stands out for its climate focus, @Risk for its flexibility, and DAMRAE for its visual modeling capabilities. However, no single tool is universally optimal, emphasizing the need to align their use with specific project needs and resources.

The findings of this study support the growing consensus that risk-informed decision-making (RIDM) is essential for managing the increasing uncertainties posed by climate change [15]. Unlike traditional approaches, which rely on fixed safety margins, RIDM allows for the continuous reassessment of risks as new data becomes available, enabling more informed and proactive decision-making [29]. This would not only enhance the dam’s resilience to extreme events but also ensure that resources are allocated efficiently, focusing on the most cost-effective solutions for reducing overtopping risk [30].

Structural and non-structural adaptation measures are both essential for mitigating the risks of dam overtopping, particularly as climate change intensifies extreme weather events [15]. Structural solutions, such as raising dam crests, increasing spillway capacities, and reinforcing embankments, directly enhance the physical capacity of dams to manage extreme inflows [30]. According to [8,25], expanding spillway discharge capacities significantly reduces overtopping risks, especially for high-hazard dams in areas with intense precipitation. Similarly, refs. [8,30] highlighted the effectiveness of crest elevation in accommodating increased reservoir inflows projected under future climate scenarios.

Building upon our earlier research [31], this study extends a previously conducted probabilistic overtopping assessment for Batu Dam under future climate change scenarios. The previous study, which utilized bias-corrected CMIP5 RCM data and iPresas-based analysis, revealed that inflows under RCP8.5 mid- and late-century projections could exceed the dam’s Probable Maximum Flood (PMF) by up to 20%, underscoring significant overtopping risks. However, the earlier work primarily focused on overtopping probabilities and general consequence estimation. The current study addresses these gaps by incorporating a comprehensive hydrological–hydraulic model, non-stationary rainfall behavior, updated PMP and IDF curves, and a fully developed Risk-Informed Decision-Making (RIDM) framework. This continuity ensures a more rigorous and adaptive approach to dam safety under future climate extremes.

Despite their effectiveness, structural measures often involve high financial costs, extended implementation timelines, and potential environmental and social trade-offs [32]. In contrast, non-structural measures, such as real-time monitoring, advanced forecasting models, and emergency action plans, offer cost-effective and flexible alternatives. [7,15] demonstrated that integrating predictive analytics with real-time hydrological data enhances inflow forecasting accuracy, providing critical lead times for emergency responses. These strategies, supported by community engagement and awareness initiatives, are particularly valuable in resource-constrained regions, offering adaptive capacity without the financial and environmental burdens of large-scale structural modifications.

2. Materials and Methods

This study adopts a comprehensive and robust methodology to evaluate the overtopping risk of Batu Dam under climate change scenarios, integrating historical data analysis, climate modeling, hydrological and flood routing modeling, probabilistic risk assessment, and the formulation of adaptation strategies. This study expands on the methods applied in our earlier published analysis [31] incorporating detailed freeboard adequacy simulations, PMP estimation, and downstream consequence modeling, which were not included in the initial framework. The primary objective is to quantify overtopping risks and propose mitigation measures, ensuring a balance between structural and non-structural approaches to enhance dam safety. A detailed flowchart, as in Figure 1, illustrates the step-by-step methodology applied in this study. The analysis focuses on evaluating overtopping risks under three future climate periods as follows:

• Early-century “2020–2046”
• Mid-century “2047–2073”
• Late-century “2074–2100”

2.1. Study Area

Batu Dam is a zoned earthfilled embankment dam located in the Gombak District of Selangor, approximately 16 km upstream of Kuala Lumpur. Strategically positioned below the confluence of Sungai Batu and Sungai Tua within the upper Klang River Basin, it controls a 50 km² forested catchment that plays a crucial role in regulating runoff and maintaining reservoir water quality. Figure 2 shows the geographical location, and Figure 3 presents the structural layout of the dam.

Commissioned in 1987 under the Kuala Lumpur Flood Mitigation Project (KLFM), the dam stands 44 m high, with a crest length of 550 m and a crest elevation of 109.0 m. The reservoir spans 2.5 km² and stores up to 36.6 million m³ at full supply level (107.3 m), while the normal pool level is 102.7 m. A side-channel ungated spillway (crest at 104.85 m, length 226.53 m) and outlet works deliver a combined discharge capacity of 251.6 m³/s, with the spillway alone accommodating 234.2 m³/s. All salient structural and operational features of Batu Dam are summarized in

Table 1. Salient features of Batu Dam.

Reservoir
Original Storage Capacity	36.6 Mm3
Peak Inflow Design Flood (IDF)	492.7 m3/s
Spillway Discharge	228 m3/s
	234.2 m3/s (outlet works not function)
Outlet Works Discharge	58 m3/s
Dam
Type	Earthfall
Height	44 m
Length	550 m
Crest Elevation (m)	El. 109.0 m
Normal Pool Level (NPL)	El. 102.7 m
Full Supply Level (FSL)	El. 107.3 m
Spilway
Crest Elevation	El. 104.85 m
Invert Elevation	El. 97.83 m at upstream
Crest Length	23 m
Total length	226.53 m (to the stilling basin)

.

Batu Dam fulfills three core functions: (i) water supply for municipal and industrial use, (ii) flood regulation up to the 100-year return period, and (iii) sediment retention to preserve downstream river capacity [1]. It operates under a humid equatorial climate influenced by the southwest (June–September) and northeast (November–March) monsoons, resulting in frequent extreme rainfall that poses operational challenges.

Proximity to dense urban settlements renders Batu Dam a high-hazard structure. It safeguards over 1.25 million residents and critical downstream infrastructure, including treatment plants and industrial zones. Recent studies classify Batu as Malaysia’s most critical dam due to its hazard rating, short downstream distance, and exposure to climate extremes [33].

2.2. Climate Projection Data

Historical hydrological data (1975–2020) from the Department of Irrigation and Drainage (DID) Malaysia, including inflow, outflow, reservoir water levels, and rainfall, were used as the baseline for evaluating Batu Dam’s historical performance. This study utilized dynamically downscaled climate datasets from the Coupled Model Intercomparison Project Phase 5 (CMIP5), prepared by the National Hydraulic Research Institute of Malaysia (NAHRIM), which is the designated statutory climate agency under the Government of Malaysia. These datasets were developed using the PRECIS regional climate modeling system and have been rigorously validated for Malaysian rainfall conditions, thereby providing a scientifically robust foundation for localized climate impact assessments [34,35]. Although CMIP6 offers more recent global climate projections, it was not employed in this study due to the unavailability of officially downscaled and site-calibrated CMIP6 data for the Batu Dam catchment from NAHRIM at the time of research. Accordingly, CMIP5 was selected as the most appropriate and institutionally endorsed source for high-resolution climate scenario analysis in the Malaysian hydrological context.

Climate projections were extracted from three Regional Climate Models (RCMs); HadGEM2-ES, MPI-ESM-LR, and NorESM1-M, under three Representative Concentration Pathways (RCP4.5, RCP6.0, and RCP8.5). To correct systematic biases in the raw RCM outputs, the Linear Scaling Method (LSM) was applied, which adjusts modeled rainfall by aligning the mean of simulated values with observed historical means. This method was selected for its proven effectiveness in tropical climates and its consistency with national hydrological standards. Following bias correction, ensemble averaging was performed by computing the mean of the corrected outputs from the three RCMs for each RCP scenario.

This step enhanced the reliability of the projections by minimizing the variance inherent in individual climate models, thereby strengthening the robustness of the rainfall datasets. The Climate Change Factor (CCF) was subsequently calculated as the ratio of future to historical rainfall intensities, evaluated across various durations and Average Recurrence Intervals (ARIs). This calculation was performed separately for each Representative Concentration Pathway (RCP4.5, RCP6.0, and RCP8.5) and for three future time horizons—early-century (2020–2046), mid-century (2047–2073), and late-century (2074–2100). Specifically, the CCF for each duration and ARI was derived by dividing the bias-corrected projected rainfall intensity by the corresponding observed intensity from the historical baseline period (1971–2015). These CCFs were then applied to scale the historical IDF curves, in accordance with national guidelines as outlined in Hydrological Procedure No. 1 (HP1, 2021) and NAHRIM’s recommended practices. The updated IDF curves, along with derived Probable Maximum Precipitation (PMP) values, served as critical inputs for subsequent hydrological modeling and overtopping risk assessment at Batu Dam.

2.3. Hydrological Analysis

This study recognizes that extreme rainfall patterns are inherently dynamic and non-static, particularly under the influence of climate change. Rather than assuming stationarity, the analysis integrates future rainfall scenarios using bias-corrected and downscaled outputs from CMIP5-based Regional Climate Models (RCMs), as provided by NAHRIM. These projections reflect temporal changes in key statistical parameters such as mean, variance, and extremes, thereby capturing the shifting characteristics of rainfall distributions over time. This methodological approach is consistent with contemporary hydrological risk assessment practices, especially considering growing climate variability. The rationale is further supported by recent Malaysian research [36], reinforcing the importance of adopting a non-stationary framework to accurately evaluate climate-driven hydrological risks.

2.4. Hydrological Modeling

Hydrological modeling was performed using HEC-HMS, which was calibrated and validated using historical rainfall and streamflow data to ensure reliable simulations. The outputs included peak inflow rates, seasonal flow patterns, and projected changes under varying climate conditions. These results formed the basis for evaluating how the dam’s hydrological system responds to extreme weather scenarios.

2.5. Flood Routing and Simulation

Flood routing simulations were conducted using Microsoft Excel to model the dam’s performance under projected inflows. The analysis accounted for reservoir storage, inflows, outflows, and spillway operations. The flood routing focused on critical parameters, including peak inflow and outflow rates, maximum reservoir water levels, and overtopping probabilities, helping to identify potential spillway inadequacies during extreme inflows.

To support the technical rigor of the simulation framework, several governing equations were employed.

1.

Continuity Equation for Reservoir Storage.

Flood routing through the Batu Dam reservoir was governed by the continuity equation as Equation (2), which represents the conservation of mass in a control volume:

$${\displaystyle \frac{dS}{dt}}=I\left(t\right)-O\left(t\right),$$

(2)

where $S$, is the reservoir storage (m³), $I$, is the inflow rate (m³/s), $O\left(t\right)$, is the outflow rate (m³/s), $t,$ is time (s).

This equation was used to evaluate the variation in storage over time, particularly during extreme flood events.

2.

Level Pool Routing Equation.

To route the hydrographs through the reservoir, the storage-indication method (also known as Puls routing) was applied. Equation (3) used as follows:

$${\displaystyle \frac{I_1+I_2}{2}}-{\displaystyle \frac{O_1+O_2}{2}}={\displaystyle \frac{S_1+S_2}{∆t}},$$

(3)

where $I_1$, $I_2$ are the inflows at the beginning and end of the time interval, respectively (m³/s), $O_1$, $O_2$ are the outflows at the beginning and end of the time interval, respectively (m³/s), $S_1$, $S_2$ are the corresponding storages (m³), $∆t$ is the time step (s).

This method enabled the simulation of reservoir water levels over time, using inflow hydrographs derived from HEC-HMS and elevation-storage-discharge curves specific to Batu Dam.

2.6. Probabilistic Risk Assessment

The study employee iPresas CALC for a probabilistic risk assessment, a novel approach in the context of dam safety under climate change. The risk model accounted for parametric uncertainty by probabilistically varying critical dam parameters, including spillway roughness (triangular distribution), crest elevation (normal distribution, ±0.2 m), and gate capacity (uniform distribution). These distributions were based on expert judgment and guidelines from [37,38]. The risk of overtopping was calculated as the product of overtopping probability and its associated consequences, such as economic losses, affected areas, and potential fatalities. This method integrates uncertainties in inflow predictions, dam parameters, and failure mechanisms, offering a nuanced understanding of the overtopping risks for each climate scenario.

The architecture of the risk model in this investigation is illustrated in Figure 4. Dam overtopping risks are evaluated through the model’s Loads, System Response, and Consequences.

• Loads: this section examines dam dynamics from both external and internal perspectives. The components are as follows:
  • PPL (Previous Pool Level): the reservoir’s water level prior to the flood event.
  • The Bottom Outlet Valve (BOV) is an essential component that regulates water discharge from the reservoir.
  • Flood Routing: models the behavior of inflow hydrographs at dams, encompassing inflow and potential overtopping situations.
• System Response: assesses the dam’s structural and operational reactions to loads.
  • The Mode of Failure (MF) denotes the mechanisms influencing dam overtopping, while considering uncertainties. When water levels exceed the dam crest, overtopping transpires, necessitating an assessment of collapse likelihood.
• Consequences: models the repercussions of overtopping, encompassing both human and economic losses.
  • Qfail: assesses water discharge during dam failure.
  • Day/Night Analysis: assesses the consequences of failures occurring during daylight versus nighttime.
  • LoL_Fail (losses of life under failure conditions) and LoL_NoF (losses of life in non-failure situations).
  • Econ_Fail (losses of economic under failure conditions and Econ_NoF (losses of economics in non-failure situation).
  • Overtopping results in supplementary losses (LoL_Inc and Econ_Inc).

The integration of these components in iPRESAS provides a comprehensive framework for probabilistic risk assessment by combining uncertainties in inflow predictions, dam parameters, and failure mechanisms. The methodology enables the evaluation of overtopping risks under various climate scenarios, enhancing the understanding of potential economic and human consequences.

2.7. Adaptation Strategies

The proposed adaptation strategies include a combination of structural and non-structural measures:

• Structural Measures: These involve raising the dam crest, enhancing spillway capacity, and reinforcing embankments to manage increased inflows. While effective, these measures entail significant financial, environmental, and social costs.
• Non-Structural Measures: These include implementing real-time monitoring systems, advanced early warning mechanisms, emergency action plans, and community-based awareness programs. These cost-effective measures enhance the resilience of downstream populations without extensive infrastructure modifications.

This integrated methodology ensures a comprehensive evaluation of overtopping risks under diverse climate conditions while providing actionable insights for risk mitigation. The approach is designed not only to address Batu Dam’s specific challenges but also to serve as a replicable framework for managing overtopping risks in other high-hazard dams globally. By combining engineering analyses with proactive adaptation strategies, this study underscores the importance of preparing for extreme hydrological events in the context of climate change.

3. Results

The results provide a comprehensive overview of how future inflow rates and overtopping risks will evolve over time, highlighting the dam’s capacity to manage extreme inflows and the potential need for adaptation strategies.

3.1. Flood Routing Results

(a)

Flood Routing in Early-Century (2020–2046)

The flood routing results for Batu Dam during the early-century period (2020–2046) show significant variations in inflow rates, water pool levels, and associated risks. The results for RCP4.5, as shown in Figure 5, demonstrate a clear pattern of inflows and water levels that remain manageable. In this early-century scenario, the inflow rates and water pool levels are moderate, with the water pool rising but staying below the crest level, thus avoiding overtopping. The crest level and clay core levels remain safely below their critical thresholds, indicating that the dam’s structural integrity is not under immediate risk. The lower initial water pool levels are consistent with baseline reservoir operations, and the conditions under RCP4.5 highlight a relatively stable hydrological regime. While this scenario does not indicate imminent danger, it underscores the importance of adaptive dam safety strategies. As climate projections shift towards more extreme conditions in later periods (under RCP6.0 and RCP8.5), the risks of overtopping and structural failure are expected to increase. The RCP4.5 results serve as a baseline for planning, suggesting that early-century impacts may be manageable with current infrastructure, but vigilance is necessary as future scenarios evolve.

(b)

Flood Routing in Mid-Century (2047–2075)

For the mid-century period (2047–2075), the RCP6.0 scenario, as in Figure 6, below demonstrates a moderate increase in inflows and water levels, with peak inflows approaching 2000 m³/s. These inflows result in water levels that approach the clay core elevation of 106.5 m.a.s.l., increasing the potential risk of overtopping. While the RCP6.0 scenario does not exceed the critical thresholds observed in higher-emission scenarios (such as RCP8.5), it still presents a noticeable overtopping risk due to the spillway capacity limitations.

The inflow and water level correlation highlights the growing flood risks under moderate emission pathways. Specifically, the outflow rates under RCP6.0 tend to lag behind the peak inflows, resulting in a backlog that could lead to increased water levels at the dam, raising concerns for dam safety. This scenario further emphasizes the need for spillway enhancements and additional safety measures to accommodate the more intense rainfall patterns associated with mid-century climate conditions.

Overall, while the RCP6.0 scenario does not present as extreme conditions as RCP8.5, the moderate overtopping risk and the limitations in spillway capacity suggest that mitigation strategies will be essential to ensure the dam’s safety and functionality during this period.

(c)

Flood Routing in Late-Century (2076–2100)

The late-century analysis reveals a clear escalation in flood risks, particularly under the RCP8.5 scenario as in Figure 7, where peak inflows reach 2500 m³/s, causing water pool levels to exceed the clay core elevation. This situation presents a high risk of structural failure unless substantial mitigation measures are implemented. The intense increase in inflows reflects the severe climate conditions associated with RCP8.5, emphasizing the urgency of addressing these risks.

In contrast, the RCP6.0 scenario presents moderate risks, and RCP4.5 shows manageable conditions, with the inflows and water levels remaining well below critical thresholds. However, the significant threat under RCP8.5 highlights the growing need for infrastructure upgrades, especially in spillway design, to accommodate the much higher inflow rates. The discrepancy between inflows and outflows further emphasizes the critical need for spillway enhancements to prevent overtopping and ensure the dam’s safety during extreme events.

As the analysis suggests, the late-century period under RCP8.5 poses the most substantial risks, with increasing rainfall intensity, shorter storm durations, and higher inflow peaks. These findings underline the importance of proactive risk management, including both structural improvements such as spillway upgrades and adaptive non-structural measures, e.g., real-time monitoring and early warning systems. These strategies are essential to ensure the long-term resilience of Batu Dam in the face of extreme climate change scenarios.

3.2. Consequences of Overtopping

The assessment of overtopping consequences at Batu Dam reveals significant risks in terms of economic losses, loss of life, affected areas, and impacted populations. The failure volume values were derived from a previous dam break study conducted by the Department of Irrigation and Drainage (DID), while the consequences were analyzed using MyFloodRAS based on the Flood Damage Assessment Report by DID. Economic loss and loss of life estimates were derived from historical consequence data. However, future projections should incorporate evolving exposure and vulnerability factors such as infrastructure growth and demographic shifts [39]. These downstream development trends may increase vulnerability. Future work should integrate dynamic population and land-use models (e.g., remote sensing, census data) to enhance consequence realism under future conditions

3.2.1. Economic Losses and Loss of Life

As shown in Figure 8, the relationship between failure volumes and their associated impacts, including estimated economic losses (in RM) and the number of lives lost, underscores the catastrophic consequences of overtopping. Beyond a failure volume of 20,000 m³, there is a sharp escalation in both economic damages and fatalities. Economic losses are projected to exceed RM 200 million in the most severe scenarios, accounting for damage to properties, infrastructure, agricultural lands, and indirect economic disruptions. The non-linear increase in damages suggests that critical infrastructure and assets become increasingly vulnerable as failure volumes rise.

Similarly, the potential for loss of life exceeds 2900 individuals in worst case overtopping scenarios. The steep rise in the loss-of-life curve between 20,000 m³ and 24,000 m³ highlights the critical thresholds at which emergency response and evacuation systems may become overwhelmed. This underscores the importance of timely disaster preparedness, particularly for downstream communities with high population densities and vulnerabilities.

A sensitivity analysis was conducted to assess how assumptions on warning time, evacuation efficiency, and affected population influence the loss of life (LoL) and economic loss outputs. Results indicate that a ±20% change in evacuation rate can lead to a ±30% change in LoL, while ±10% changes in infrastructure value affect economic losses by ±15%. These findings highlight the importance of improving the reliability of consequence parameters and scenario-based evaluations.

3.2.2. Affected Area and Population

Figure 9 illustrates the escalating impact of overtopping failure on the affected area (km²) and population. At failure volumes approaching 24,000 m³, the spatial extent of the affected area exceeds 6.5 km², while the impacted population surpasses 300,000 individuals. This demonstrates the extensive downstream repercussions of dam failure and highlights the necessity for robust emergency planning to protect both lives and livelihoods in the event of overtopping.

3.2.3. Loss of Life

The data reveal a direct correlation between overtopping failure volume and the severity of consequences. These findings align with global studies on high-hazard dams, where extreme inflow scenarios exacerbate threats to downstream communities and infrastructure. The rapid escalation of risks highlights the urgency of implementing proactive measures. In conclusion, the escalating risks illustrated in the analysis emphasize the urgency of implementing proactive measures to mitigate overtopping consequences.

3.3. Overtopping Risk Quantification

Using the iPresas CALC, the study quantified the annual probability of overtopping for each climate scenario.

The results of overtopping risk, as shown in

Table 2. Summary of the overtopping probabilities for the early-, mid-, and late-century periods.

Period	RCP4.5	RCP6.0	RCP8.5
Early-century	0.02%	0.01%	0.05%
Mid-century	0.03%	0.05%	0.07%
Late-century	0.04%	0.05%	0.08%

, indicate significant variability in overtopping risks for Batu Dam under different climate scenarios. The highest overtopping probabilities were observed under RCP8.5 during the late-century period (2074–2100), reflecting the extreme precipitation levels associated with minimal global mitigation efforts. In contrast, the lowest probabilities occurred under RCP4.5 during the early-century period (2020–2046), consistent with more moderate climate conditions and lower projected rainfall intensities. These findings highlight the direct relationship between the intensity of climate change scenarios and overtopping risks, emphasizing the urgency for mitigation measures.

The probabilistic assessment reveals that the risk of overtopping increases significantly under the RCP8.5 scenario, rising from 0.05% annually in the early-century to 0.08% annually by the late-century. This represents a substantial increase in risk compared to historical conditions, where the overtopping probability was negligible. Even under the RCP4.5 and RCP6.0 scenarios, the overtopping probability increases, though the risk remains lower compared to RCP8.5.

The high overtopping risk under RCP8.5 is attributed to the projected increase in both the frequency and intensity of extreme rainfall events, leading to inflows that exceed the dam’s design capacity. Conversely, the lower risks under RCP4.5 suggest that moderate emissions control could help maintain inflow levels within manageable limits. These high and low points serve to underline the critical importance of climate mitigation efforts and their potential to reduce overtopping risks for high-hazard dams like Batu Dam.

The results reveal strong correlations between projected rainfall intensities, peak inflows, and overtopping probabilities. For example, simulations show that a 20% increase in extreme rainfall intensity results in a proportional increase in peak inflow rates, subsequently elevating overtopping risk by approximately 30%. These findings align with existing literature suggesting that extreme precipitation events, amplified by climate change, are primary drivers of dam safety risks [12,18].

The probabilistic risk assessment conducted using iPRESAS further corroborates these correlations, demonstrating that higher inflow uncertainties under RCP8.5 lead to broader risk ranges and greater probabilities of overtopping. This relationship underscores the need for adaptive risk management approaches that account for the non-linear impacts of climate variability on dam safety. Additionally, the study identifies a critical threshold for the dam’s spillway capacity, beyond which overtopping becomes inevitable. This finding supports the hypothesis that structural limitations play a key role in dam vulnerability under extreme conditions.

3.4. Risk Evaluation

The f–N curve (failure frequency vs. number of fatalities) is a key tool used to evaluate the risk of dam failure, particularly in relation to overtopping events. This curve is derived from the USBR Tolerability Criteria, which establishes the tolerability limit for the annual probability of failure, linking the likelihood of failure to potential loss of life. The curve is used to assess the acceptability of risk in dam safety management, helping to prioritize mitigation measures based on the expected consequences of failure.

In the context of Batu Dam, the f–N curve for the late-century, as in Figure 10, illustrates the estimated failure probability and associated loss of life, highlighting the increasing risk of overtopping under the more extreme climate conditions compared to the mid-century period. The data points for RCP8.5 are notably the highest, indicating the most severe risk scenario, followed by RCP4.5 and RCP6.0. This trend highlights the escalating risks of overtopping events as the climate warms, particularly under higher emission scenarios.

The overtopping risks identified for Batu Dam align with findings from similar global studies, underscoring the vulnerability of embankment dams to overtopping under future climate scenarios. For example, research on the Eugui Dam in Spain [15] and in North Carolina [17] has shown that climate-induced changes in rainfall patterns and streamflow are likely to significantly increase the overtopping risk for high-hazard dams.

The results for Batu Dam support these global trends, where increased rainfall intensity and frequency, as projected under the RCP8.5 scenario, push dam systems beyond their existing design thresholds. These findings reinforce the need for developing risk-informed adaptation strategies to ensure the resilience of critical water infrastructure like Batu Dam.

These outcomes emphasize the urgency for both structural and non-structural mitigation measures. Expanding spillway capacity is essential to manage extreme inflows, while real-time monitoring and early warning systems can play a crucial role in minimizing loss of life. Without such proactive measures, overtopping events could lead to catastrophic humanitarian and socioeconomic impacts, underscoring the necessity of adopting risk-informed dam safety strategies to ensure the long-term stability and safety of embankment dams under changing climate conditions.

3.5. Adaptation and Mitigation Strategies

The results emphasize the critical role of these interventions in mitigating overtopping risks and safeguarding communities. As shown in Figure 11, implementing structural measures, such as the construction of new spillways, crest level adjustments, and increased freeboard, significantly reduces in overtopping probabilities by directly enhancing the dam’s capacity to manage extreme inflows. For instance, increasing spillway capacity alone can reduce overtopping risk. Additionally, upgrades, such as modern gate installations, improve the dam’s ability to regulate water release, minimizing the likelihood of overtopping and mitigating downstream flooding. However, while effective, these measures require substantial financial investment and long-term planning.

Non-structural measures, on the other hand, target preparedness and community resilience, offering cost-effective and immediate solutions. Emergency Action Plans (EAP) and Community Engagement Campaigns (CEC) significantly reduce loss of life by fostering community awareness and ensuring swift emergency responses. In iPRESAS, the effect of Emergency Action Plans (EAPs) and Community Engagement Campaigns (CECs) was modeled by reducing the “population at risk” parameter. Under full EAP implementation, a 50–70% reduction in exposed population was assumed during the evacuation time window, in line with findings from [40]. As a result, loss-of-life estimates decreased significantly in simulations that included both EAP and CEC measures, particularly in daytime failure scenarios. When implemented together, these measures demonstrate a notable reduction in risk, highlighting the importance of proactive education, real-time communication, and comprehensive emergency protocols. Unlike structural measures, non-structural interventions are scalable and can adapt quickly to changing conditions, making them indispensable in managing risks in the short term.

The analysis underscores that combining structural and non-structural measures provides the most comprehensive risk reduction, addressing both the physical limitations of the dam and the vulnerabilities of downstream populations. Scenarios that incorporate both approaches exhibit the lowest probabilities of failure and the smallest LoL estimates, showcasing the critical need for integrated strategies. This combined approach addresses the multifaceted nature of dam safety, enhancing the dam’s physical integrity while simultaneously equipping communities to respond effectively during emergencies.

The findings further highlight the growing risks posed by climate change, particularly under extreme scenarios such as RCP8.5. The significant increase in rainfall intensity and frequency associated with this scenario amplifies overtopping probabilities, necessitating urgent adaptation measures. While structural upgrades provide long-term solutions to manage higher inflows, non-structural strategies, such as early warning systems and real-time monitoring, are essential for anticipating extreme events and enabling preemptive action. The use of iPRESAS in this study demonstrates the value of probabilistic risk assessments in identifying critical failure points and prioritizing cost-effective interventions. This tool allows decision-makers to better understand the trade-offs between structural and non-structural strategies and allocate resources effectively.

Considering these findings, a combination of structural and non-structural measures is recommended to ensure the long-term safety and operational efficiency of Batu Dam. Structural enhancements, such as increasing spillway capacity, modernizing gate systems, and raising crest levels, should be prioritized to address physical vulnerabilities. Simultaneously, non-structural measures, including EAPs, CECs, and early warning systems, should be implemented to improve preparedness and reduce the human impacts of overtopping events. An integrated approach, supported by continuous monitoring, updated climate modeling, and robust policy frameworks, will be essential to adapt to the evolving risk landscape and mitigate the impacts of extreme climate scenarios. By investing in these comprehensive strategies, Batu Dam can effectively address the challenges posed by climate change, ensuring the safety of downstream communities and the resilience of critical water infrastructure.

4. Discussion

This study provides a comprehensive evaluation of overtopping risks at Batu Dam under future climate change scenarios, utilizing a combination of hydrological modeling (HEC-HMS), flood routing simulations, and probabilistic risk assessment using iPRESAS. The results show a significant upward trend in overtopping probability under high-emission scenarios, especially RCP8.5. These findings are consistent with earlier works such as Fluixá-Sanmartín et al. [15], who demonstrated that climate-induced intensification of extreme rainfall substantially increases dam failure risk through overtopping. Similar studies by Ahmadisharaf and Kalyanapu [17] and Hariri-Ardebili [2] also emphasize the inadequacy of conventional deterministic models in capturing the variability and uncertainty introduced by climate change.

In this study, the late-century overtopping probability at Batu Dam under RCP8.5 reaches 0.08%, indicating a non-negligible risk level for a high-hazard dam. This aligns with the findings of Fluixá-Sanmartín et al. [18], where overtopping risk at Eugui Dam under RCP8.5 rose sharply in the late-century projections. Compared to earlier Malaysian studies that primarily focused on historical flood frequency analysis [1,4], this study advances the discourse by incorporating probabilistic modeling, which accounts for a broader spectrum of uncertainties and climate variability.

The interpretation of these results yields several important insights:

(a)

Increasing Overtopping Risk.

The analysis confirms that Batu Dam’s current hydraulic infrastructure is increasingly inadequate to manage future inflows, particularly under the RCP8.5 scenario. The increase from 0.05% to 0.08% overtopping probability between mid- and late-century periods demonstrates a significant escalation in failure risk. This is consistent with global findings [6,7] that show overtopping remains the most frequent and catastrophic mode of failure for embankment dams, especially when design standards are based on obsolete historical rainfall patterns.

(b)

Comparative Analysis of Climate Scenarios.

While RCP4.5 and RCP6.0 scenarios present relatively lower overtopping risks, they still indicate a gradual increase over time, reinforcing the notion that even moderate climate change trajectories necessitate preemptive adaptation. This pattern mirrors results reported by Kuntoro et al. [16] for Darma Dam in Indonesia, where RCP4.5 still led to reservoir performance decline. Hence, reliance on moderate scenarios does not eliminate risk but only delays its onset, stressing the need for both short- and long-term interventions.

(c)

Adaptation Strategies: Structural vs. Non-Structural.

This study underscores the effectiveness of combining structural and non-structural measures. Structural options, such as increasing spillway capacity, crest elevation, and embankment strengthening, demonstrate clear risk reduction but are constrained by cost and implementation time. These findings correspond to [3], who reported similar trade-offs in Spanish case studies. In contrast, non-structural measures like real-time hydrological monitoring, emergency action plans (EAPs), and forecasting systems offer flexible, scalable, and immediate risk mitigation. As shown in similar works [8,33], such approaches are particularly vital in resource-limited settings and can be rapidly deployed to complement physical upgrades.

(d)

Implications for Risk-Informed Decision-Making (RIDM)

The application of iPRESAS software has proven instrumental in identifying critical thresholds, quantifying overtopping probabilities, and enabling comparative analysis across scenarios. In contrast to deterministic approaches, the probabilistic RIDM framework provides a more holistic understanding of dam vulnerability and allows decision-makers to prioritize safety investments based on actual risk levels. This supports earlier recommendations by McCann and Paxson [25], who advocated probabilistic tools as standard practice for high-hazard dam portfolios. Additionally, the risk escalation trend under RCP8.5 highlights the limitations of the traditional design-flood approach and the need for periodic reassessment of dam performance under updated climate scenarios. As seen in similar risk-based studies [23], integrating climate projections into dam safety protocols is crucial for adaptive planning.

(e)

Policy and Regional Relevance.

From a policy standpoint, this study highlights the urgent need for national dam safety authorities in Malaysia to transition from design-based to risk-informed safety assessments. Current regulatory frameworks rely heavily on historical hydrological data, which no longer reflects the evolving flood hazard landscape. Incorporating findings from this study can support the Malaysian government’s broader climate resilience initiatives, as outlined in the National Water Resources Policy and the River Basin Management Plan.

The relevance of this research extends beyond Batu Dam, providing a methodological framework applicable to other high-hazard dams in Southeast Asia. Given the regional monsoonal variability and increasing frequency of extreme weather, similar dams across the Mekong, Mahakam, and Pahang basins could benefit from adopting a RIDM-based approach tailored to their unique hydrological and socio-economic settings.

(f)

Recommendations for Practice and Research.

This study provides actionable guidance for dam operators, water managers, and policymakers:

• Structural Upgrades: expansion of Batu Dam’s spillway and potential crest elevation increases should be prioritized under national budget allocations for climate adaptation.
• Monitoring Systems: implementation of advanced telemetry and hydrological forecasting tools is critical for early warning and operational readiness.
• Safety Reviews: periodic, climate-updated risk assessments should be institutionalized for all major dams, incorporating seismic, internal erosion, and external failure modes.
• Future Research: integration of socio-economic vulnerability mapping, sedimentation impacts, and multi-dam cascading failure modeling would further enhance the robustness of future assessments.

Moreover, this study did not incorporate the long-term impacts of reservoir siltation, wave run-up due to wind action, or material degradation of the dam body. These processes can incrementally reduce reservoir capacity, increase overtopping risk during storm events, and weaken dam structure resilience over time [35,41]. While sedimentation, wave overtopping, and material degradation were not quantified, these factors can significantly reduce reservoir capacity and structural resilience over time. Integration of reservoir sediment modeling and erosion studies would be valuable extensions to this work. Future integrated risk assessments should consider sedimentation modeling, wave action simulations, and aging infrastructure analysis for a more holistic understanding of risk.

In summary, this study contributes significantly to the field of dam safety under climate change by demonstrating the value of probabilistic, risk-informed methodologies in overtopping risk quantification. The findings align with and extend global research, reinforce the limitations of static design methods, and propose scalable interventions that balance engineering rigor with practical implementation.

5. Conclusions

This study assessed the overtopping risk of Batu Dam under climate change scenarios using a comprehensive, risk-informed methodology that integrates hydrological modeling, flood routing simulations, and probabilistic risk assessment via iPRESAS. The findings reveal that climate change, particularly under the high-emission RCP8.5 scenario, will significantly intensify overtopping risks, challenging the adequacy of conventional design assumptions based on historical hydrological data.

The key conclusions are as follows:

• Climate-Driven Risk Escalation: The probability of overtopping is projected to increase substantially by the late century, with RCP8.5 reaching 0.08% annually. This underscores the urgent need to reassess dam safety under future hydroclimatic conditions.
• Effectiveness of Hybrid Adaptation Strategies: A combination of structural interventions (e.g., increasing spillway capacity, crest elevation) and non-structural measures (e.g., real-time monitoring, early warning systems) offers the most effective approach to mitigate overtopping risks. While structural upgrades ensure long-term hydraulic resilience, non-structural actions provide immediate, cost-effective safety enhancements.
• Value of Probabilistic Risk Assessment: The use of iPRESAS allowed for nuanced evaluation of overtopping probabilities and associated consequences. This probabilistic approach surpasses traditional deterministic methods by accounting for hydrological uncertainty and supporting prioritization of interventions.
• Policy and Practice Implications: The study presents a replicable framework that can inform dam safety policy revisions in Malaysia and similar regions. It advocates for institutional adoption of risk-informed decision-making (RIDM), periodic climate-based safety reviews, and adaptive dam safety planning as part of national resilience strategies.
• Pathways for Future Research: Future investigations should consider integrating sedimentation effects, wave run-up, and socio-economic vulnerability into overtopping assessments. Additionally, multi-dam system simulations and cross-sectoral adaptation planning would enhance regional water infrastructure resilience.

In conclusion, the integration of climate scenarios, probabilistic modeling, and adaptation strategies in this study provides a robust and transferable methodology for assessing overtopping risk in embankment dams. The findings reinforce the imperative of moving toward dynamic, data-informed safety planning to ensure the long-term resilience of critical water infrastructure in an era of climate uncertainty.