Abstract

This study analyzed the flood risk associated with potential dam break events at the Pidekso dam in Wonogiri Regency, Central Java Province, Indonesia. Embankment dams, such as the Pidekso dam, are susceptible to piping and overtopping, which can result in dam failure and severe natural disasters, causing loss of life and infrastructure damage. The study utilized HEC-HMS and HEC-RAS software to simulate dam break scenarios, generating maps of dam break flood discharge, flood inundation, and flood arrival time. The analysis revealed that overtopping scenarios resulted in a higher outflow discharge compared to piping scenarios, with a peak discharge of 14,821 m³/s. Flood inundation and arrival time maps were used to assess the risks to nearby villages. Moreover, based on the risk index calculation using the formula provided by the National Disaster Management Agency, the studied villages were classified into distinct risk levels. Specifically, one very low-risk, four low-risk, six medium-risk, and seven high-risk villages, with none classified as very high-risk. This study also recommends a combination of structural and non-structural solutions to mitigate the risks of a dam break. By implementing structural mitigation measures such as an emergency spillway and compound channel along the downstream river, the study achieved an 8.4% reduction in flood extent. While most villages showed no significant changes in their risk indices, Sinorboyo village, which was previously susceptible to flooding, benefited from enhanced protection measures.

1 Introduction

Indonesia possesses immense water resources potential, with a total annual water potential of 3.9 trillion cubic meters, of which 3.2 trillion cubic meters remain underutilized (Tamim et al. 2023). Currently, 79.6% of the utilized water is allocated for agricultural irrigation, while the remaining 20.4% is used for domestic, industrial, and raw water needs (Lukas 2023). To address water crisis and food security concerns, the Indonesian government, through the Ministry of Public Works and Housing, is committed to completing the construction of 61 dams between 2014 and 2024 (Rosytha and Suryana 2023). Among these projects is the Pidekso Dam in Wonogiri Regency, Central Java, which was completed in 2021. Dams like Pidekso play a pivotal role in fulfilling various needs, including irrigation and electricity generation (Altinbi̊lek 2002; Hadjerioua et al. 2015). However, the importance of dams is accompanied by potential threats if they are not operated and managed with meticulous care (Tran et al. 2020). The significance of dam safety in Indonesia was underscored by the failures of the Situ Gintung dam in 2009 (Nabilah et al. 2020; Pratiwi et al. 2020), which prompted the government to enhance disaster preparedness and planning to mitigate the impacts of dam failures (Pratiwi et al. 2020).

Dam failures present substantial risks, often arising from issues such as piping and overtopping, especially when construction, operation, or management practices deviate from safety standards (Froehlich 2008; Evans et al. 2000). Overtopping is a condition in which the water volume in the dam exceeds its capacity, causing water to overflow along the sides of the dam and then flow downstream (Singh 1996; Ralston 1987). Meanwhile, piping occurs when water flows through gaps in the soil or material around the dam foundation, causing erosion that can lead to structural failure (Chen et al. 2019). Both of these are dangerous conditions and can cause serious damage, including total dam failure and natural disasters (Singh 1996).

The Pidekso dam, as an embankment dam, has a greater risk of experiencing piping and overtopping issues (Laksono et al. 2020; Froehlich 2008). The history of dam collapse disasters shows that most dam failures involve earth and rockfill dam types (embankment dams) (Froehlich 2008). In addition, the number of casualties in dam failure incidents is directly related to the warning time available to evacuate people downstream of the dam (Sharma and Kumar 2013).

Therefore, risk analysis needs to be carried out to identify potential hazards and the likelihood of collapse incidents at Pidekso dam. Risk analysis is conducted to identify hazard vulnerabilities and capacities in a given area. To identify potential risk and the likelihood of disaster events at Pidekso dam, accurate and complete data on hazard vulnerabilities and capacities in the surrounding area is required (National Disaster Management Agency 2012).

Several studies have been conducted in Indonesia regarding risk analysis and dam break modeling. For instance, Adityawan et al. (2023) presented a two-dimensional dam break numerical modeling through an urban area. Mellivera et al. (2020) conducted numerical simulations using the Saint-Venant 1D-2D equations with a finite difference scheme for Fordward Time Center Space. Yakti et al. (2018) utilized a 2D model with HEC-RAS, while Magdalena et al. (2020) studied the dam break phenomenon over a movable bed using a mathematical model based on the finite volume method. Additionally, Pratiwi et al. (2020) performed experimental work on dam break flow through a single oblique obstruction in a straight rectangular channel. These studies serve as valuable references for developing risk analysis models for dam failure.

This study utilizes HEC-HMS for hydrological modeling related to dam failure and the 2D HEC-RAS numerical model to simulate floods resulting from dam failure. In the risk analysis, this study introduced modifications to certain parameters, with a focus on flood depth as a key factor for hazard assessment. Additionally, factors such as flood arrival time and the extent of the flooded area were considered to determine capacities. A modified methodology was employed, incorporating data from the Central Statistics Agency, to assess vulnerabilities, resulting in a more precise understanding of the risks associated with potential dam break events.

The aim of this study is to develop a risk map as a first step in efforts to mitigate risks at the dam. The novelty of this research lies in its comprehensive approach to risk analysis, incorporating modified parameters and utilizing data from diverse sources. Additionally, the study will model structural mitigation plans such as emergency spillways and compound channels along the river and assess the differences in risk analysis for the affected areas. With the existence of a risk map, it is expected to assist relevant parties in making appropriate decisions in efforts to manage and maintain the dam, as well as to increase public awareness of the dangers and risks associated with embankment dams like Pidekso dam.

2 Study location

The Pidekso dam is located administratively in Giriwoyo District, Wonogiri Regency, Central Java Province, Indonesia. Specifically, it is situated in the Pidekso and Tukulrejo Villages within Giriwoyo District. The dam's normal inundation area covers approximately 290 hectares, encompassing three villages: Pidekso, Tukulrejo, and Sendangsari. The data related to the Pidekso dam can be found in Table 1, and Figure 1 provides a map illustrating the dam's location. The dam lies about 25 km east of the Wonogiri Regency, and is bounded to the north and west by Tukulrejo Village, to the south by Pidekso Desa Village, and to the east by Sendangsari Village.

Downstream of the Pidekso dam lies the Gajah Mungkur Reservoir, which receives direct flow from the Pidekso dam. In the event of a dam failure, the water from the Pidekso dam would flow directly towards the Gajah Mungkur Reservoir, passing through the villages along the downstream area. This highlights the critical importance of ensuring the structural stability of the Pidekso dam, as any failure could have significant consequences for the communities residing in the vicinity of the downstream area.

Table 1 Technical data of Pidekso dam.

1	Watershed Area	57.15 km2
2	Normal Water Level (NWL)	+ 185 m
3	Area at (NWL)	209 Ha
4	Flood Water Level (FWL)	+ 186.6 m
5	Area at (FWL)	232 Ha
6	Low Water Level (LWL)	+ 174.5 m
7	Total Storage Volume	25 million m3
8	Effective Storage Volume	17 million m3
9	Dead Storage Volume	8 million m3
10	Dam Type	Embankment Dam
11	Dam Crest Length	387 m
12	Dam Crest Width	10 m
13	Dam Crest Elevation	+ 189 m
14	Upstream Slope	1 V : 3 H
15	Downstream Slope	1 V : 3 H
16	Spillway Type	Ogee Spillway
17	Design Flood PMF	604 m3/s
18	Spillway Elevation	+ 185 m
20	Spillway Width	56 m

Figure 1 Pidekso dam, Gajah Mungkur reservoir, and the villages in the downstream area.

3 Data and methods

3.1 Data collection

The data used in this paper were obtained from several relevant agencies. The precipitation data were obtained from the Ngancar station and the GPM IMERG Final Precipitation satellite product downloaded from the website https://disc.gsfc.nasa.gov. This satellite product has a spatial resolution of 0.1° x 0.1° and a temporal resolution of 1 day, spanning from 2000 to 2021, for a total of 22 years. Topographic data were obtained from DEMNAS (DEM National) with a resolution of 8.25 x 8.25 metres. Land use information was obtained from the Ministry of Environment and Forestry of the Republic of Indonesia for the year 2019. Technical specifications of dams were provided by the River Basin Management Agency of Bengawan Solo. In addition, socio-economic data was obtained from the Central Statistics Agency (Central Statistics Agency of Wonogiri Regency 2022), with the most recent data available from 2022. These datasets from different agencies serve as valuable sources of information for the analysis conducted in this paper.

3.2 Methodology

The stages of the work can be explained by the writing methodology, see Figure 2, below.

Figure 2 Methodology.

The workflow in this study includes several hydrological analyses. First, rainfall data from the GPM satellite were corrected against ground station data using correction factors based on probability curves. This was followed by a Thiessen polygon analysis of regional rainfall and a Probable Maximum Precipitation (PMP) analysis using the Hersfield method (Hershfield 1965). Effective rainfall was then determined using the area reduction factor and the SCS curve number for infiltration. Design runoff hydrographs were then developed using various methods including the Gama method (Harto 1993), the Nakayasu method (Natakusumah 2011) and the Soil Conservation Service  Unit Hydrograph method (United States Soil Conservation Service 1972). All hydrological analyses were performed in accordance with the guidelines of the Dam Management Agency (2017) and the National Standardization Agency (2016), especially the “Procedure for Calculating Planned Flood Discharge”.

Dam failure simulations were conducted using HEC-HMS to determine the scenario with the highest outflow discharge for the most severe risk analysis. The simulations considered overtopping and piping conditions. Subsequently, HEC-RAS was used to model a 2D area downstream of the dam post-failure. The flood mapping from HEC-RAS provides outputs such as flood area, depth, and arrival time. In this context, flood depth is used as the hazard parameter in risk analysis, while flood arrival time and area indicate capacity. Capacity is determined based on flood area indicators in each village, as well as the capacity for evacuation based on the arrival time of the flood. Vulnerability aspects are based on socio-economic analysis. Risk analysis is then conducted, considering hazard, capacity, and vulnerability parameters. Mitigation measures, including both structural and non-structural measures, are subsequently developed. Structural mitigation involves modeling several structures, such as emergency spillways and compound channels along the river. The effectiveness of the mitigation measures is evaluated by assessing the changes in flood extent.

3.3 Dam breach parameters

Dam breach parameters refer to the various factors that influence the behavior of water flow when a dam fails or collapses (Froehlich 2008).The physical description of the dam failure will consist of the height of the breach (H_b), starting water-surface elevation or critical overtopping depth at which breach formation begins (H_w), average width (B_ave), and side slopes in H:V (Froehlich 2008). These values represent the maximum breach size. A diagram describing the breach is shown in Figure 3 below. Additionally, the time taken for the complete development of the breach after the initiation phase, known as the breach formation time (t_f), needs to be determined.

Figure 3 Dam breach dimension.

Regression equations for the dimensions of the breach have been developed by several researchers. In this study, the regression equation proposed by Froehlich (2008) was utilized to estimate the breach dimensions. The data used by Froehlich for his regression analysis covered the following ranges:

1. Height of the dams:                               3.05 – 92.96 m
2. Volume of water at breach time:         0.0139 – 660.0 m³ x 10⁶

Froehlich's regression equations for average breach width and failure time are:

(1)

(2)

Where:

Bave	=	Average breach width (m)
K0	=	Constant (1.3 for overtopping failures, 1.0 for piping),
Vw	=	Reservoir volume at time of failure (m3),
hb	=	Height of the final breach (m),
g	=	Gravitational acceleration (m/s2), and
tf	=	Breach formation time (s).

3.4 Risk parameters

The concept of disaster risk can be defined as the combination of hazard, vulnerability, and capacity (Keaokiriya et al. 2022). In simple terms, risk can be formulated as the potential for adverse effects or losses resulting from the interaction of hazards and vulnerabilities, taking into account the capacity to cope with or manage the risk (National Disaster Management Agency 2012). The formulation of risk can be expressed as follows:

(3)

The flood depth hazard level is classified based on United Nations Platform for Space based Information for Disaster Management and Emergency Response (2016). Table 2 shows the hazard parameters index in this paper.

Table 2 Hazard parameters.

Component	Indicator	Score					Unit
		Very Low (0.1)	Low (0.3)	Medium (0.5)	High (0.7)	Very High (0.9)
Hazard	Flood Depth	<1	1-2	2-3	3-4	>4	m

Capacity is determined based on flood area indicators in each village, as well as the capacity for evacuation based on the arrival time of the flood. The arrival time of the flood is calculated from the beginning of the dam break. Table 3 shows the capacity parameters index in this paper.

Table 3 Capacity parameters.

Component	Indicator	Weight Factor	Score					Unit
			Very Low (0.1)	Low (0.3)	Medium (0.5)	High (0.7)	Very High (0.9)
Capacity	Flooded Area	0.2	80-100	60-80	40-60	20-40	<20	%
	Flood Time Arrivals	0.8	<30	30-60	60-90	90-120	>120	Minutes

Vulnerability is a condition of a community or society that leads to or causes an inability to deal with the threat of disaster. Table 4, below, presents the vulnerability parameter indices used in this study, including demographic, social, health, economic, infrastructure, and environmental vulnerabilities (Disaster Management and Climate Change Agency of Nahdlatul Ulama 2017).

Table 4 Vulnerability parameters.

Component	Indicator	Weight Factor	Score					Unit
			Very Low (0.1)	Low (0.3)	Medium (0.5)	High (0.7)	Very High (0.9)
Demographic vulnerability	Total population	5	<500	500-1000	1000-2000	2000-3000	>3000	People
	Percentage of female presence	6	<10	10-25	25 - 50	50-75	>75	%
	Percentage of non-PLN households	6	<10	10-25	25 - 50	50-75	>75	%
Health vulnerability	Number of cases of malnutrition	8	0	1	2	2-4	>4	Case
	Access to health services	8		Easy	Quite Hard	Very Hard		Type
	Drugstore access	7		Easy	Quite Hard	Very Hard		Type
Social vulnerability	Type of river use	9		Source of drinking water	Source of drinking, washing water	Source of drinking, bathing, washing water		Type
	Physical and mental diffabilities	9	0	0-5	5-20	20-40	>40	People
Infrastructure vulnerability	Number of base transceivers station	5	>3	3	2	1	0	Unit
	Early warning system	10	>3	3	2	1	0	Unit
Economic vulnerability	Number of market	10	>3	3	2	1	0	Type
Environmental vulnerability	Number of natural threats	7	0	1	2	2-4	>4	Case
	Location	10		Top Elevation	Hillside	Valley / Plains		Type

4 Results and discussion

4.1 Hydrological analysis

The Pidekso dam was originally designed based on the discharge estimation of the Probable Maximum Flood (PMF) using data from the Ngancar rainfall station, which is located outside the watershed. To account for the accurate representation of the watershed's rainfall, this study incorporated precipitation data from the GPM (Global Precipitation Measurement) Satellite rainfall station within the watershed area. Incorporating rainfall data exceeding the initial PMF design was essential for an effective modeling of the dam's overtopping. Therefore, the calculations in this study utilized precipitation data obtained from the GPM product, specifically relying on the IMERG Final Precipitation satellite data. Within the vicinity of the Pidekso dam, there exist four GPM satellite grid data points that intersect with both the watershed and the Ngancar Rainfall Station. These specific GPM stations, visually depicted in Figure 4, provide valuable insights into the spatial distribution of rainfall across the study area.

Figure 4 Rainfall station and GPM grid.

GPM 1, which is closest to the Ngancar station, will be used for the data correction process along with the Ngancar station. The first step in this correction process is to compare the two datasets using the correlation coefficient. According to the Dam Management Agency (2017) guidelines, satellite data can be corrected if the correlation coefficient between the GPM rainfall station and the Ngancar rainfall station is 0.6 or greater. The monthly rainfall data correction at the Ngancar station using GPM 1 yielded a correlation coefficient of 0.8, indicating that rainfall correction with a correction factor can be applied. The correction involves the probability curve of monthly maximum rainfall events within specific intervals. Calculations were based on monthly maximum rainfall data from 2010 to 2018. The equations and correction factors for each interval are shown in Table 5, along with the probability curve graphs before and after the correction process, as shown in Figure 5.

Table 5 Rainfall correction.

Rainfall Interval (mm)	Correction Equation	Correction Factor
GPM > 140	y = 0.0028 e0.0414x	1.374
98 < GPM < 140	y = 0.002 x + 0.7457	0.975
78 < GPM < 98	y = -0.085 ln x + 1.2952	0.920
66 < GPM < 78	y = -0.326 ln x + 2.3174	0.858
47 < GPM < 66	y = 0.446 ln x - 0.9389	0.583
31 < GPM < 47	y = 0.0439 x - 1.3284	0.263
GPM < 31	y = 0	0

Figure 5 Probability curve before correction and after correction.

Furthermore, the same correction procedure was applied to GPM 2, GPM 3, and GPM 4 stations using the equations derived from GPM 1. The rainfall data from GPM 2, GPM 3, and GPM 4 stations will be used for regional rainfall calculations using the Thiessen polygons. The results of the regional rainfall are presented in Table 6 below.

Table 6 Annual maximum regional rainfall.

Year		Annual maximum GPM regional rainfall (mm)
2000		81
2001		103
2002		97
2003		70
2004		132
2005		228
2006		119
2007		240
2008		128
2009		98
2010		116
2011		85
2012		100
2013		174
2014		72
2015		56
2016		105
2017		106
2018		53
2019		73
2020		63
2021		52

Next, the design rainfall Probable Maximum Precipitation (PMP) was calculated using the Hershfield method (Hershfield 1965) based on the regional maximum rainfall. The calculation results indicate a PMP rainfall of 651.7 mm. Subsequently, the water loss is determined using the ARF (Area Reduction Factor) and SCS CN (Curve Number) infiltration method (Dam Management Agency 2017). Afterwards, a flood discharge analysis is conducted for a 6-hour rainfall duration using the PSA 007 rainfall distribution (Dam Management Agency 2017). In the flood discharge analysis, the Gama (Harto 1993), Nakayasu (Natakusumah et al. 2011), and Soil Conservation Service Unit Hydrograph methods (United States Soil Conservation Service 1972) will be used to compute the Probable Maximum Flood (PMF) discharge. The results of this hydrograph analysis are shown in Figure 6 below.

Figure 6 PMP Effective Rainfall and PMF hydrograph.

However, based on the comparative test using the Creager curve (Creager et al. 1945) presented in Figure 7, the SCS method is selected as the most appropriate and closest to the Creager curve, thus used to estimate the flood discharge. Furthermore, the SCS Probable Maximum Flood (PMF) hydrograph will be utilized as input for dam-break modeling using HEC-HMS and HEC-RAS software.

Figure 7 Creager curve.

4.2 Dam break modeling

The analysis of the failure of Pidekso dam was conducted for 2 scenarios, namely the piping condition at normal water level elevation, and the overtopping condition during the PMF event.

The parameters for dam failure were calculated using the Froehlich equation (Froehlich 2008), and the results were obtained for piping condition. The average crack width (B_ave) was 80 m, and the failure time (T_f) was 0.87 hours. Meanwhile, in the case of overtopping, the average crack width (B_ave) was 104 m, and the failure time (T_f) was 1 hour. Froehlich’s paper further indicates that the recommended average side slopes are as follows: 1H:1V for cases of overtopping failures, and 0.7H: 1V for other situations such as piping or seepage.

Based on the modeling results depicted in Figure 8 and Figure 9, the peak discharge during piping is determined to be 8,314 m³/s, whereas during overtopping it reaches 14,821 m³/s. Prioritizing safety considerations, the larger discharge from the overtopping scenario is selected as the parameter for the next phase of modeling.

Figure 8 Piping outflow discharge.

Figure 9 Overtopping outflow discharge.

In the next stage, a 2D-area modeling was carried out using HEC-RAS to map the downstream areas that are potentially affected by the dam failure. The results of the modeling provides information on flood depths and flood arrival times in each affected area, which can be used to develop accurate disaster mitigation plans. The topographic data used in the dam failure modeling comes from the DEMNAS Indonesia data with a cell size of 8.25 x 8.25 m.

The simulation scheme entails defining the upstream boundary condition as the outflow discharge from the overtopping scenario flood, while the downstream boundary condition is set as the normal depth. To ensure the accuracy of calculations, a 4 x 4 m mesh was employed for the river and its embankments, while a 40 x 40 m mesh was utilized in peripheral areas. This mesh configuration optimizes accuracy around the riverbanks while maintaining computational efficiency. In this modeling, the default Manning value of 0.06 was used.

The computational process involves a time step of 0.5 seconds over a period of 23 hours following the dam failure. The computational framework is governed by the Shallow Water Equation (SWE) of Saint Venant, which is a fundamental model in hydraulic engineering for simulating unsteady flow in open channels (Chow 1959; Henderson 1966). The SWE describes the conservation of mass and momentum in a fluid flow, making it a cornerstone in simulating the complex dynamics of floods and their impact on the surrounding environment (Moussa and Bocquillon 2000).

4.3 2D-Area modeling simulation results

The HEC-RAS modeling results have produced flood depth information downstream of the dam. This flood depth data was classified based on the United Nations Platform for Space-based Information for Disaster Management and Emergency Response (2016) standards. The modeling output was then translated into visualizations, as shown in Figures 10, 11, and 12, below. In addition, the HEC-RAS modeling results provided data on the flood arrival time since the dam started to collapse. This data was utilized to generate a flood arrival map, which is an important parameter for creating a capacity map. Figure 12 displays the results of the flood arrival map.

Figure 10 Flood inundation map.

Figure 11 Flood depth map.

Figure 12 Flood arrival time map.

4.4 Hazard, vulnerability, capacity, and risk analysis

Assessing the potential risk of a dam break flood disaster necessitates the consideration of three primary factors: hazard, vulnerability, and capacity. The hazard index is calculated based on the flood depths obtained from the previous modeling, and the resulting hazard index for each village is compiled in Table 7. The vulnerability index is derived through a socio-economic analysis encompassing demographic data, social conditions, health indicators, economic factors, infrastructure conditions, and environmental aspects. The results of the vulnerability index for each village are summarized in Table 7. Furthermore, the capacity index is determined by analyzing flood inundation areas and the corresponding flood arrival time for each village, utilizing the modeling outcomes from the preceding sub-chapter. The summarized capacity index results for each village are presented in Table 7.

Table 7 Risk index calculation.

Village	Hazard	Vulnerability	Capacity	Risk	Risk Index
Balepanjang	0.90	0.57	0.68	0.54	Medium
Bulurejo	0.90	0.52	0.22	0.72	High
Bumiharjo	0.90	0.57	0.41	0.67	High
Gambiranom	0.50	0.58	0.92	0.29	Low
Gedongrejo	0.90	0.54	0.43	0.65	High
Giriwoyo	0.90	0.54	0.55	0.60	High
Glesungrejo	0.90	0.53	0.87	0.39	Low
Minggarharjo	0.70	0.52	0.89	0.34	Low
Ngancar	0.90	0.55	0.40	0.67	High
Pidekso	0.90	0.58	0.24	0.74	High
Sejati	0.90	0.56	0.59	0.59	Medium
Selomarto	0.90	0.57	0.59	0.60	Medium
Sendangagung	0.90	0.57	0.67	0.55	Medium
Sirnoboyo	0.10	0.52	0.92	0.16	Very Low
Tawangharjo	0.90	0.62	0.71	0.55	Medium
Tegalharjo	0.50	0.52	0.91	0.29	Low
Tukulrejo	0.90	0.62	0.25	0.75	High
Watuagung	0.50	0.50	0.60	0.46	Medium

These three factors were utilized to calculate the overall risk using the formula provided by the National Disaster Management Agency, as discussed in the previous sub-chapter. Risk classification is based on the risk score, with scores ranging from 0 to 0.2 categorized as very low risk, scores from 0.2 to 0.4 as low risk, scores from 0.4 to 0.6 as medium risk, scores from 0.6 to 0.8 as high risk, and scores from 0.8 to 1 as very high risk. The risk index calculation results for each village are visualized in Figure 13, and a summary of the calculations can be found in Table 7. The risk calculation results reveal the following distribution of risk indices among the studied villages: specifically, one village was classified as very low-risk, four villages were classified as low-risk, six villages were classified as medium-risk, seven villages were classified as high-risk, and none of the villages were found to have a very high-risk.

Figure 13 Risk index map.

This comprehensive assessment provides valuable insights for decision-makers and stakeholders in formulating appropriate mitigation strategies, emergency preparedness plans, and resource allocation to minimize the potential impact of a dam break flood disaster. By understanding the specific risk levels of each village, effective measures can be implemented to enhance the resilience and safety of the communities residing in the downstream areas of the dam.

4.5 Mitigation

Structural Mitigation

Several structural mitigation efforts that can be implemented to prevent dam breaks at the Pidekso dam include injection system (grouting), pump system, dam maintenance, and construction of an emergency spillway. One of the crucial factors in ensuring the safety of a dam is the presence of an emergency spillway. As stated by Risdiyanto and Prawito (2022), construction of an emergency spillway is a necessity to release excessive floodwaters safely and avoid catastrophic failure of the dam. It is crucial to design and construct an efficient and functional emergency spillway to ensure safety during flood events. Evans et al. (2000) also stated that failure of the dam can be attributed to a large hydrologic event and the combination of several factors: 1) inadequate spillway design, 2) lack of an emergency spillway, 3) loss of permanent pool capacity due to sedimentation, and 4) poor dam maintenance resulting in seepage piping failure.

Furthermore, additional structural mitigation efforts can be directed towards the downstream section of the dam, specifically through the construction of embankments along the river. By implementing dual embankments, a compound channel configuration is formed. Compound channels consist of multiple geometric shapes, with water predominantly flowing within the main channel during low and normal flows. However, during high flows, the water level rises and overflows into the floodplain or overbank areas.

This paper presents the modeling of two structural mitigation solutions for the dam break scenario: an emergency spillway and downstream river embankments, forming a compound channel. The emergency spillway is designed for a 100-year return period and consists of a radial gate with a width of 36 m. The impact of the emergency spillway on the outflow discharge can be visualized in Figure 14. As for the compound channel design, the main channel is designed for a 2-year return period during low and normal flows, with a width of 30 m and an embankment height of 2 m. During high flows, the flood plain channel is designed for a 100-year return period, with an embankment height of 6 m and a width of 100 m. An example of the compound channel cross-section used in the modeling can be seen in Figure 15.

Figure 14 Impact of emergency spillway.

Figure 15 Cross-section of compound channel design.

From Figure 14, it can be observed that the emergency spillway has a limited impact on the peak discharge, reducing it by 3.4% from 14,821 m³/s to 14,310 m³/s. Additionally, there is a slight delay in the timing of the peak discharge, shifting from 6:50 to 7:10. These findings indicate that the influence of the emergency spillway is not particularly significant. This can be attributed to the large inflow discharge entering the dam, resulting in overtopping despite the presence of the emergency spillway. However, the emergency spillway does provide additional spillway capacity, reducing the likelihood of overtopping.

Subsequently, flood modeling was conducted with the implementation of the emergency spillway and compound channel. The results demonstrate a reduction in the affected flood area resulting from dam failure. Initially, the impacted area was 1,598 hectares, but with the presence of the emergency spillway and compound channel, the flooded area has been reduced to 1,464 ha, representing an 8.4% reduction. Figures 16 and17 below, illustrate the difference in inundation areas following the implementation of structural mitigations such as the emergency spillway and compound channel.

Figure 16 Flood area before mitigations.

Figure 17 Flood area after mitigations.

The structural mitigations shown in Figures 16 and 17 had minimal impact on reducing the flood area. Subsequent risk assessments were carried out for each village after implementing these measures. In general, the risk index remained unchanged for most villages, except for Sinorboyo village, which was previously exposed to flooding but is now protected. The updated risk index for each village is visualized in Figure 18. This highlights the importance of prioritizing preventive actions to enhance dam safety. Regular monitoring and diligent maintenance of dams and spillways are vital in ensuring their proper functioning, effectiveness, and minimizing the risk of dam failures during flood events. By implementing these measures, the risk of catastrophic dam failure can be reduced, providing reassurance to the communities residing downstream of the dam.

Figure 18 Risk index after mitigations.

Non-structural mitigation

To prepare for potential dam failure at the Pidekso dam, non-structural efforts can be made through the preparation of an emergency action plan. This plan should involve the local government and related agencies in disaster management, with a focus on two groups of actions: dam protection and saving the community and environment (Saleh and Yusmanizar 2019). Additionally, efforts to prevent dam breaks should include the establishment of laws and regulations specifically aimed at reducing disaster risk, as well as socialization and education, and training of dam owners, managers, and the community (National Disaster Management Agency 2012). Mitigation efforts such as an early warning system, evacuation routes, and shelter should also be implemented (Sianturi et al. 2021; Yunita and Puspitosari 2014).

The early warning system for the possibility of dam break in the Pidekso dam incorporates various components. First, rainfall and water level monitoring are essential for continuous surveillance of water levels and rainfall in the vicinity of the Pidekso dam. This is achieved through the installation of gauges that provide real-time data on these parameters. Second, a deformation measurement is conducted to assess land deformation and dam stability. Shear pegs are installed at multiple locations, allowing for the monitoring of any potential shifts or changes. Additionally, V-notch instruments are used to monitor water seepage, ensuring early detection of any abnormal seepage patterns. Another crucial component is the wire sensor, which is positioned downstream of the dam to detect debris flow. When debris flows occur, it impacts the wire sensor, triggering an alarm system that activates sirens and alerts people to start evacuating. Finally, the local government has designated evacuation routes and installed information boards in the village, providing clear guidance for residents on the shortest and safest paths to follow during an emergency evacuation. These measures collectively enhance the risk management capabilities and preparedness in the event of a dam break disaster.

In the event of a dam break, a recovery plan should be in place, including a command post and management sector, food and non-food sector (logistics), and a victim search and evacuation sector (National Disaster Management Agency 2012). It is important to carefully carry out socialization efforts to avoid public unrest and continuously evaluate and update the emergency action plan to ensure its effectiveness.

5 Consclusion

Based on the results of the calculation and analysis of the Pidekso dam break, the following conclusions can be drawn:

Based on the calculation of the outflow discharge due to overtopping and piping, it is known that the outflow discharge due to overtopping is higher than due to piping. The scenario that is used as a reference is the scenario of a dam collapse due to overtopping where the peak discharge has the largest number, which is 14,821 m³/s.

The identification of flood hazard areas can be accomplished through the utilization of HEC-RAS, a flood modeling approach that employs Digital Elevation Models (DEM) data. This analysis yields valuable outputs in the form of flood inundation maps and a flood arrival time map, as shown in Figures 10, 11, and 12.

Based on the risk index calculation using the formula provided by the National Disaster Management Agency, the results indicate that one village exhibits a very low-risk index, four villages have a low-risk index, six villages are classified as medium-risk, and seven villages are categorized as high-risk. Notably, no villages are identified as having a very high-risk index. The risk index values for each village can be visualized in Figure 13, while Table 7 summarizes the calculation results.

To mitigate the potential consequences of the Pidekso dam failure, both structural and non-structural solutions are advised. Structural measures include implementing an injection system, constructing an emergency spillway, installing a pump system, ensuring regular dam maintenance, and constructing embankments along the river (compound channel). Non-structural measures encompass establishing an early warning system, developing regulations, conducting socialization and education campaigns, implementing training programs, adopting mitigation strategies, establishing evacuation procedures, and implementing recovery efforts.

The modeled structural solutions in this study focused on the implementation of an emergency spillway and compound channel along the downstream river, resulting in an 8.4% reduction in flood extent. Subsequent risk analysis indicated that most villages did not experience changes in their risk indices, except for Sinorboyo village, which was previously susceptible to flooding but is now protected.

References

1. Adityawan, M.B., C. Sandi, D. Harlan, M. Farid, A.A. Kuntoro, Widyaningtias, A.N. Chaidar, and J. Nugroho. 2023. “Numerical Modeling of Dam Break Induced Flow through Multiple Buildings in an Idealized City.” Results in Engineering 18 (June): 101060. https://doi.org/10.1016/j.rineng.2023.101060
2. Altinbi̊lek, D. 2002. “The Role of Dams in Development.” International Journal of Water Resources Development 18 (1): 9–24. https://doi.org/10.1080/07900620220121620
3. Central Statistics Agency of Wonogiri Regency. 2022. Kabupaten Wonogiri Dalam Angka 2022.
4. Chen, S-S., Q-M. Zhong, and G-Z. Shen. 2019. “Numerical Modeling of Earthen Dam Breach Due to Piping Failure.” Water Science and Engineering 12 (3): 169–78. https://doi.org/10.1016/j.wse.2019.08.001
5. Chow, V.T. 1959. Open-Channel Hydraulics. McGraw-Hill, New York, USA.
6. Creager, W.P., J. Hinds, and J. DeWitt Justin. 1945. Engineering for Dams. John Wiley & Sons, New York, USA.
7. Dam Management Agency, Ministry of Public Works and Public Housing, Directorate General of Water Resources. 2017. “Technical Guidelines for Calculating Flood Discharge at Dams.” Jakarta: Ministry of Public Works and Public Housing.
8. Disaster Management and Climate Change Agency of Nahdlatul Ulama. 2017. “Development of Vulnerability Mapping to Disasters.” Disaster Management and Climate Change Agency of Nahdlatul Ula-ma.
9. Evans, J.E., S.D. Mackey, J.F. Gottgens, and W.M. Gill. 2000. “Lessons from a Dam Failure.” Ohio Journal of Science 100 (5): 121–31.
10. Froehlich, D.C. 2008. “Embankment Dam Breach Parameters and their Uncertainties.” Journal of Hydraulic Engineering 134 (12): 1708–21.
11. Hadjerioua, B., A.M. Witt, K.M. Stewart, M.B. Acosta, and M. Mobley. 2015. “The Economic Benefits Of Multipurpose Reservoirs In The United States – Federal Hydropower Fleet.” ORNL/TM-2015/550. Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States). https://doi.org/10.2172/1237622
12. Harto, S. 1993. Analisis Hidrologi. Gramedia Pustaka Utama. https://lib.ui.ac.id
13. Henderson, F.M. 1966. Open Channel Flow. Macmillan Publishing Co., Inc., New York, USA.
14. Hershfield, D.M. 1965. “Method for Estimating Probable Maximum Rainfall.” Journal AWWA 57 (8): 965–72. https://doi.org/10.1002/j.1551-8833.1965.tb01486.x
15. Keaokiriya, P., S.V.K. Prabhakar, B.R. Shivakoti, T. Goto, H. Sakai, A.Dewi, S. Jayasinghe, et al. 2022. “Disaster Risk Reduction in the ASEAN Region: Understanding and Assessing Systematic Risks of Floods and Landslides in a River Basin Context.” In The UN Global Assessment Report on Disaster Risk Reduction 2022 (GAR 2022) Contributing Paper. United Nations Office for Disaster Risk Reduction (UNISDR).
16. Laksono, R.S., Y.S. Sundari, and Y. Subhy. 2020. “Analisa Stabilitas Lereng Dan Rembesan Bendungan Tipe Urugan Pada Bendungan Teritip Balikpapan.” Kurva S 11 (1): 245–60.
17. Lukas, M. 2023. “Tantangan Common Dilema Dalam Kelembagaan Pengelolaan Daerah Irigasi Bena.” Jurnal Administrasi Dan Demokrasi (Administration and Democracy Journal) 2 (01): 60–68. https://ejurnal.undana.ac.id/index.php/JAD/article/view/11883
18. Magdalena, I., M.B. Adityawan, and C. Jonathan. 2020. “Numerical Model for Dam Break over a Movable Bed Using Finite Volume Method.” International Journal of GEOMATE 19 (71): 98–105. https://doi.org/10.21660/2020.71.27074
19. Mellivera, A., K. Zain, M.B. Adityawan, D. Harlan, M. Farid, and B.P. Yakti. 2020. “Development of FTCS Artificial Dissipation for Dam Break 2D Modelling.” Jurnal Teknik Sipil 27 (1). https://doi.org/10.5614/jts.2020.27.1.1
20. Moussa, R., and C. Bocquillon. 2000. “Approximation Zones of the Saint-Venant Equations f Flood Routing with Overbank Flow.” Hydrology and Earth System Sciences 4 (2): 251–60. https://doi.org/10.5194/hess-4-251-2000
21. Nabilah, R.A., D. Sutjiningsih, E. Anggraheni, and S. Murniningsih. 2020. “Dam Break Analysis of Situ Gintung Dam Collapse Reconstruction.” IOP Conference Series: Earth and Environmental Science 599 (1).
22. Natakusumah, D.K., W. Hatmoko, and D. Harlan. 2011. “Prosedur Umum Perhitungan Hidrograf Satuan Sintetis Dengan Cara ITB Dan Beberapa Contoh Penerapannya.” Jurnal Teknik Sipil 18 (3). https://doi.org/10.5614/jts.2011.18.3.6
23. National Disaster Management Agency. 2012. “Head of National Disaster Management Agency Regulation Number 02 of 2012: General Guidelines for Disaster Risk Assessment.” Jakarta: National Disaster Management Agency.
24. National Standardization Agency. 2016. “SNI 2415-2016: Procedure for Calculating Planned Flood Discharge.” Jakarta: National Standardization Agency.
25. Pratiwi, V., M.S.B. Kusuma, H. Kardhana, and M. Farid. 2020. “Experimental Study of Dam Break Flow Generated by the Flap Gate in Horizontal Channel.” International Journal of GEOMATE 19 (75): 19–26. https://doi.org/10.21660/2020.75.16098
26. Ralston, D.C. 1987. “Mechanics of Embankment Erosion During Overflow.” Hydraulic Engineering, Proceedings of the 1987 ASCE National Conference on Hydraulic Engineering, Williamsburg, Virginia, p. 733–38.
27. Risdiyanto, A., and A. Prawito. 2022. “Dimension Analysis Of The Emergency Spillway of Tirawan Dam With The Application of the System Dynamic Model.” UKaRsT 6 (1). https://doi.org/10.30737/ukarst.v6i1.2477
28. Rosytha, A., and W.M. Suryana. 2023. “Peran Unit Pengelola Bendungan Dalam Pengelolaan Bendungan Berkelanjutan Di Satker OP BBWS Brantas (Studi Kasus UPB Bendungan Babjulmati Dan Bendungan Nipah).” Publikasi Riset Orientasi Teknik Sipil (Proteksi) 5 (1): 44–50. https://doi.org/10.26740/proteksi.v5n1.p44-50
29. Saleh, R., and Y. Yusmanizar. 2019. “Emergency Preparedness and Response Communication in Dealing with the Effects of Hazards during the Increase of the Water Level of the Bili-Bili Dam.” Proceeding of International Conference on Communication, Culture and Media Studies (CCCMS), May. https://journal.uii.ac.id/CCCMS/article/view/13528
30. Sharma, R.P., and A. Kumar. 2013. “Case Histories of Earthen Dam Failures.” Seventh International Conference on Case Histories in Geotechnical Engineering 8.
31. Sianturi, R.S., A.Y. Koswara, and I. Elysiyah. 2021. “Recommending Assembly Points, Evacuation Routes, and Standard Operating Procedures for Potential Flooding Due to Reservoir Dam Failures: A Case Study of Gondang Reservoir, Lamongan, East Java.” IOP Conference Series: Earth and Environmental Science 778, 012004. https://doi.org/10.1088/1755-1315/778/1/012004
32. Singh, V.P. 1996. Dam Breach Modeling Technology. Springer Science+Business Media, Dordrecht. https://doi.org/10.1007/978-94-015-8747-1
33. Tamim, T., M. Tumpu, P. Indrayani, M.S. Yusman, I. Djamaluddin, and E.R. Bungin. 2023. Pengembangan Sumber Daya Air Berbasis Lingkungan. Tohar Media.
34. Tran, D.Q., S. Nishimura, M. Senge, and T. Nishiyama. 2020. “Risk of Embankment Dam Failure from Viewpoint of Hydraulic Fracturing: Statistics, Mechanism, and Measures.” Reviews in Agricultural Science 8: 216–29. https://doi.org/10.7831/ras.8.0_216
35. United Nations Platform for Space-based Information for Disaster Management and Emergency Response. 2016. “Flood Hazard Assessment.” 2016. https://www.un-spider.org/advisory-support/recommended-practices/recommended-practice-flood-hazard-assessment/step-by-step
36. United States Soil Conservation Service. 1972. SCS National Engineering Handbook, Section 4: Hydrology. The Service, 1972 – Hydrology.
37. Yakti, B.P., M.B. Adityawan, M. Farid, Y. Suryadi, J. Nugroho, and I.K. Hadihardaja. 2018. “2D Modeling of Flood Propagation Due to the Failure of Way Ela Natural Dam.” MATEC Web of Conferences 147, 03009. https://doi.org/10.1051/matecconf/201814703009
38. Yunita, F.T., and D.A. Puspitosari. 2014. “Risk Management in Dam Break Disaster: Lessons Learned from Way Ela Natural Dam Break.” Jurnal Sosek Pekerjaan Umum 6 (3): 140–221.