Effects of hydrodynamic actions on the stability of slopes constructed with tropical soils: experimental analysis in a reduced physical model

Efeitos de ações hidrodinâmicas na estabilidade de taludes construídos com solos tropicais: análise experimental em modelo físico reduzido

ABSTRACT

This study investigated the stability of dam slopes built with tropical soils, focusing on the effects of wave action and rapid drawdown with the aim of testing the physical scale model. A 1:30 scale physical model was developed using a lateritic clayey tropical soil from Ouro Preto, MG, classified as MH (SUCS) and LG' (MCT). The models were subjected to controlled wave generation (varying height and period) and rapid water level drawdown. Monitoring included moisture sensors, 3D laser scanning for erosion assessment, and water turbidity measurements. Results revealed that progressive surface erosion, saturation-induced instability, and localized failure from rapid drawdown are the predominant failure mechanisms. A direct relationship was found between wave height and erosion rate, with critical wave heights above 1.8 m (real scale) inducing accelerated erosion. The study provides technical data for designing safer hydraulic structures with tropical soils against extreme hydrodynamic events.

Keywords:
Slope stability; Tropical soils; Wave action; Physical modeling; Erosion; Dams

RESUMO

Este estudo investigou a estabilidade de taludes construídos em modelos físicos reduzidos que representavam taludes de barragens, focado nos efeitos da ação de ondas e do rebaixamento rápido com objetivo de testar o modelo físico reduzido. Um modelo físico reduzido em escala 1:30 foi desenvolvido com solo tropical laterítico argiloso de Ouro Preto, MG, classificado como MH (SUCS) e LG' (MCT). Os modelos foram submetidos a condições controladas de geração de ondas (variando altura e período) e rebaixamento rápido. O monitoramento incluiu sensores de umidade, escaneamento a laser 3D para avaliação da erosão e medição da turbidez. Os resultados revelaram que a erosão superficial progressiva, a instabilidade por saturação e a ruptura localizada por rebaixamento rápido são os mecanismos de falha predominantes. Observou-se relação direta entre a altura da onda e a taxa de erosão, com alturas de onda críticas acima de 1,8 m (escala real) induzindo erosão acelerada. O estudo fornece subsídios técnicos para o dimensionamento de estruturas hidráulicas mais seguras com solos tropicais frente a eventos hidrodinâmicos extremos.

Palavras-chave:
Estabilidade de taludes; Solos tropicais; Ação de ondas; Modelagem física; Erosão; Barragens

INTRODUCTION

The stability of water dam slopes is a fundamental element in geotechnical and hydraulic engineering, crucial for the safety and longevity of these structures. The erosive action of waves is a primary agent of degradation, potentially compromising the structural integrity of slopes and leading to large-scale failures. This soil-water interaction is a dynamic phenomenon involving saturation, progressive erosion, and gradual changes in the soil's mechanical properties (Fiori, 2015). Slope stability is influenced by geotechnical factors and hydrodynamic conditions, including soil characteristics, water level fluctuations, compaction, and the cyclic action of waves, which can alter internal stresses and reduce shear strength (Gerscovich, 2016).

Earth dams are widely used in Brazil, accounting for approximately 82% of the country's dams, due to favorable topography and the availability of earthy materials (Espósito et al., 2010; Mariano & Silva, 2022). These structures often utilize tropical soils, which result from intense chemical weathering and possess unique characteristics, such as good mechanical strength and low water absorption. The MCT (Miniature, Compacted, Tropical) methodology has improved the geotechnical understanding of these materials (Marinho et al., 2020; Departamento Nacional de Infraestrutura de Transportes, 2023).

Available research focuses on estimating soil loss using hydrosedimentological modeling and sedimentation data (Rabelo et al., 2025; Khodja et al., 2025; Luz et al., 2025). Physical modeling at reduced scales has been successfully employed in geotechnical studies to investigate complex soil-water interactions under controlled conditions. Studies by Zhu et al. (2023) and Hou et al. (2021) demonstrated the effectiveness of scaled physical models for analyzing slope stability under hydraulic loading, while Jones & Anastasopoulos (2021) validated the use of reduced-scale wave generation systems for studying tsunami-infrastructure interaction. However, these works have focused primarily on temperate or sandy soils, and the application of reduced-scale modeling to tropical lateritic soils remains limited. This gap highlights the need for experimental research to simulate field conditions and understand the governing mechanisms. This research addresses this gap by hypothesizing that a reduced physical model can represent the erosive behavior of tropical soil slopes under wave action.

Furthermore, it is necessary to develop solutions in harmony with the United Nations (2015) Sustainable Development Goals (SDGs), as advocated in the 2030 Agenda. The SDGs aim to improve people's quality of life, protect the planet, increase the resilience of cities and communities, and ensure global prosperity by 2030. Although the SDGs are indivisible, the theme addressed in this research is directly aligned with SDG 9 (Industry, Innovation and Infrastructure), SDG 11 (Sustainable Cities and Communities), SDG 12 (Responsible Consumption and Production), SDG 14 (Life Below Water), and SDG 15 (Life on Land).

The primary objective of this research is to develop and test a reduced physical model to analyze the stability of water dam slopes constructed with tropical soils subjected to wave action. Specific objectives include: (1) analyzing the effects of wave action experimentally under controlled conditions, (2) evaluating the effects of rapid water level drawdown, and (3) identifying erosion and instability mechanisms through multi-parameter monitoring.

Erosion and slope instability processes

The erosion of the margins of water bodies is a natural process of landscape modeling as shown in Figure 1 and 2, driven by fluvial and wave-induced forces (Empresa Brasileira de Pesquisa Agropecuária, 2020; Ribeiro, 2021; Andrade et al., 2020). This process has been intensified by anthropogenic activities and extreme climatic events, posing significant risks to hydraulic infrastructure. The stability of a slope depends on the equilibrium between gravitational forces and the resisting forces of the soil and rock materials (Santos, 2019). According to Bertoni & Lombardi Neto (2010) instability occurs when the driving forces exceed the material's resistance, a process influenced by factors such as soil type (erodibility), rainfall intensity (erosivity), and topography.

Figure 1
Activities flowchart.

Figure 2
A) Schematic of the tank and slope dimensions. B) Wave tank, wave generator, and the constructed soil slope within the tank.

For dam slopes, instability can manifest through various mechanisms. A critical process is undercutting, where wave action scours material at the toe of the slope, potentially compromising the entire structure (Nguyen et al., 2020). The upstream slope is particularly vulnerable as it is in direct contact with the reservoir water and subject to wave action and fluctuations in the water level. The downstream slope, while typically dry, must also be designed for stability under various conditions. Design guidelines provide recommendations for slope inclinations based on soil type and dam height to ensure stability (Agência Nacional de Águas e Saneamento Básico, 2016a, 2016b; Eletrobras, 2003).

Reservoir volume loss due to sediment deposition is another problem associated with the erosion of slopes surrounding dam reservoirs. Several studies employing different analysis proposals estimate the average annual loss of reservoir volume due to sediment deposition, with values of approximately 1% per year, depending on the specific conditions of the study site (Wisser et al., 2013). Studies conducted by Eletrobrás/IPH in Brazil show that the annual loss of reservoir storage capacity is approximately 0.5% (Carvalho, 1994; Carvalho et al., 2000). The work of Araújo (2003), focusing on the semiarid region of Brazil, mentions that siltation reduces reservoir storage capacity by an average of 0.2% per year. Specific studies related to sedimentological monitoring in these semiarid regions of Brazil were developed by Bento et al. (2019), Gil et al. (2020), Barra et al. (2022), and Rabelo et al. (2023). Araújo et al. (2006); Rabelo et al. (2025) highlighted that siltation reduces water availability in reservoirs through two main mechanisms: an increase in the exposed surface area, leading to evaporation, and a reduction in the usable storage volume, thus intensifying “spillway losses”.

Influence of wave action on slope stability

The cyclic action of waves is a primary agent of degradation for dam slopes. Waves generate dynamic pressures on the slope surface, causing fluctuations in internal pore-water pressure that can reduce the soil's effective stress and, consequently, its shear strength (Hou et al., 2021). This phenomenon affects different soils in distinct ways; in sandy soils, the repetitive loading can induce liquefaction, while in cohesive clayey soils, it typically results in progressive surface erosion (Zhu et al., 2023). The magnitude and frequency of the waves are fundamental parameters that dictate the rate of structural degradation over time. The stability of the slope is therefore directly compromised not only by the physical removal of particles but also by the internal weakening of the soil mass (Yang et al., 2023).

Geotechnical properties of tropical soils

The study focuses on tropical soils, as highlighted by Gomes et al. (2022), these soils are abundant in Brazil and are the product of intense chemical weathering in hot and humid climates. This weathering process leads to a concentration of residual minerals, especially iron and aluminum oxides, which impart a characteristic reddish color and unique geotechnical properties (Almeida & Lima, 2000). Tropical soils are often found in an unsaturated condition with a high void ratio, yet can exhibit good mechanical strength and low water absorption.

Due to their distinct formation and structure, conventional soil classification systems developed for temperate regions are often inadequate. This led to the development of the MCT (Miniature, Compacted, Tropical) methodology, which is specifically designed to characterize the behavior of tropical soils for engineering purposes. The MCT classification categorizes soils into groups such as LG' (lateritic clayey soils) and NA' (non-lateritic sandy soils), providing a more reliable prediction of their performance. A crucial feature of unsaturated tropical soils is the presence of negative pore-water pressure, or matric suction, which provides an additional component of shear strength. This suction-induced strength is highly dependent on moisture content and is lost as the soil becomes saturated, a critical consideration for the stability of slopes exposed to wave action and varying water levels.

Physical modeling and similitude criteria

To investigate complex geotechnical phenomena under controlled conditions, reduced physical models serve as an essential experimental tool. The validity of a physical model relies on the theory of similarity, which requires that the model and the full-scale prototype maintain proportional relationships in terms of geometry, kinematics (motion), and dynamics (forces). For hydraulic systems where gravity is the dominant force, such as wave propagation, dynamic similarity is achieved by maintaining a constant Froude number (Fr) (Chanson, 2004). The Froude number represents the ratio of inertial forces to gravitational forces and is defined as:

$Fr=\frac{v}{\sqrt{gL}}$ (1)

Where: v is the flow velocity, g is the acceleration of gravity, and L is a characteristic length. By establishing a geometric scale factor (λL​) and ensuring the Froude number is consistent between the model and prototype, the scaling laws for other key parameters, such as time (λt​=λL^1/2​) and velocity (λv​=λL^1/2​), can be derived. This allows the experimental results from the model to be reliably extrapolated to predict the behavior of the real-world structure.

While the Froude similarity criterion is well-suited for modeling gravity-driven processes such as wave propagation, surface runoff, and free-surface flows, it presents certain inherent limitations in the context of geotechnical modeling. The primary advantages of Froude scaling include its simplicity in application, suitability for phenomena dominated by gravitational forces, and straightforward derivation of temporal and velocity scaling relationships. These characteristics make it the most appropriate criterion for the predominantly hydraulic phenomena investigated in this study.

However, several limitations must be acknowledged. The Froude criterion does not fully capture matric suction effects, which are scale-dependent and crucial for unsaturated tropical soil behavior. Matric suction contributes significantly to the apparent cohesion and shear strength of unsaturated soils, but its magnitude and distribution cannot be accurately scaled using geometric relationships alone. Additionally, permeability scaling introduces distortions, as hydraulic conductivity does not scale linearly with geometric dimensions; the actual scaling relationship for permeability follows λk = λL under idealized conditions, which may not hold for heterogeneous tropical soils with complex pore structures. Furthermore, viscous forces and capillary effects, which become relatively more important at smaller scales, may be overrepresented in the model compared to the prototype. Despite these limitations, Froude similarity remains the most appropriate and widely accepted criterion for the gravity-controlled processes of wave action and slope erosion that constitute the primary focus of this research.

MATERIAL AND METHODS

The study utilized a reduced physical model to simulate the interaction between waves and slopes made of compacted tropical soil. The experimental program depicted in Figure 1 was developed using a 1:30 scale physical model to investigate slope stability under controlled wave action and rapid drawdown conditions. The choice of scale was based on similarity considerations and laboratory constraints, ensuring adequate representation of the physical phenomena while maintaining practical feasibility for detailed monitoring and measurement.

Soil characterization

The material used was a lateritic clayey tropical soil from Ouro Preto, Brazil, collected from a natural slope at a depth of 0.5 to 1.5 meters. It was classified using both the Unified Soil Classification System (SUCS) and the MCT (Miniature, Compacted, Tropical) methodology. The MCT classification system, specifically developed for tropical soils, was applied to complement the traditional SUCS classification. This system considers the peculiarities of tropical soils, including their lateritic characteristics and specific behavior patterns not adequately captured by conventional classification systems. Using MCT Methodology, it is possible to evaluate the soil erodibility potential. In accordance with Nogami & Villibor (1979).

Most of the tests in the experimental program were carried out in the laboratories of the Technology Center for Geotechnics Applied (CTGA) of the Federal University of Ouro Preto (UFOP). The following tests were conducted according to ABNT (Brazilian Association of Technical Standards) and ASTM standards: Particle size analysis (Associação Brasileira de Normas Técnicas, 2018); Liquid limit (w_l) (Associação Brasileira de Normas Técnicas, 2016a) and Plastic limit (w_p) (Associação Brasileira de Normas Técnicas, 2017); Proctor compaction test (Associação Brasileira de Normas Técnicas, 2016b); MCT (Miniature, Compacted, Tropical) classification, including Mini-MCV and loss on immersion tests (Departamento Nacional de Estradas de Rodagem, 2023; Departamento Nacional de Estradas de Rodagem, 1994); Mineralogical analysis by X-ray Diffraction (XRD) and chemical analysis by X-ray Fluorescence (XRF); Hydraulic conductivity (Associação Brasileira de Normas Técnicas, 2021).

The minerals in the samples were identified using X-ray diffraction (XRD) tests. The equipment used was a Bruker D2 Phaser (2nd Generation) benchtop diffractometer equipped with a copper tube (CuKα), from the NanoLab/REDEMAT/UFOP.

Reduced physical model

The experimental setup was housed in a rectangular glass tank measuring 3.0 m x 0.8 m x 0.8 m following the scheme of Figure 2, and the final model dimensions were a 0.70 m base, 0.70 m height, and 0.80 m width. A geometric scale of 1:30 was adopted, allowing a prototype with a base and height of approximately 21 m to be represented. The Froude similarity criterion was used as it governs phenomena influenced by gravity, such as waves and open channel flow (Chanson, 2004).

Two identical slope models were constructed. The soil was prepared to its optimum moisture content, determined by the Proctor test, and left to rest for 12 hours for homogenization. The slope was built up in seven 10 cm layers inside the tank, with each layer compacted to achieve a density between 90% and 100% of the maximum Proctor density. The final block of soil was then carefully carved to the desired slope geometry of 45° (1V:1H).

Instrumentation and testing procedures

The models were instrumented with four moisture sensors (TDR type) installed at 15 cm spacing to monitor the advance of the saturation front (Figure 3). The first model (Slope 1) was used for exploratory tests, varying wave periods (0.4 s to 1.2 s) and exposure times (30 s to 1 hour) to define critical parameters. The second model (Slope 2) was tested under standardized conditions (4 Hz frequency, 10-minute cycles) for comparative analysis.

Figure 3
Placement of sensors.

The moisture sensors used in the experiments were previously calibrated using controlled soil volumes and moisture levels. Calibration involved comparing sensor readings with reference measurements (analytical balance and oven drying). The resulting calibration curve showed a high linear correlation (R² > 0.99), ensuring reliable and accurate data during testing.

Water level indicators were installed in Slope 2 at three different points (one at 10 and two at 20 cm from the slope face) to monitor water level variations during wave action and drawdown. The water level indicators were installed through small-diameter access holes (20 mm) drilled horizontally into the compacted slope. While this installation was necessary for monitoring internal hydraulic conditions, it inadvertently created discontinuities in the soil structure, contributing to the increased erodibility observed in Slope 2 compared to the undisturbed Slope 1.

A comprehensive testing program was developed for both slopes to systematically investigate slope response under various hydraulic loading conditions.

Wave generation was achieved using a programmable paddle-type wave generator with displacement amplitude control. For each test, waves were allowed to stabilize before measurements began.

Rapid drawdown tests were conducted using three valve configurations (small, large, and combined) to assess hydraulic behavior and induced deformations.

1. Small valve configuration: Single 12.5 mm diameter valve, producing gradual drawdown (0.69 cm/min)

2. Large valve configuration: Single 18 mm diameter valve, producing rapid drawdown (1.89 cm/min)

3. Combined configuration: Simultaneous opening of both valves, producing very rapid drawdown (2.40 cm/min)

Water turbidity was measured using a HACH TL2360 nephelometric turbidimeter (NTU) to indirectly quantify sediment transport. Erosion was also quantified by 3D laser scanning of the slope surface and by weighing the eroded sediment collected after the tests. Visual analysis of erosion and deformation was performed using high-resolution video recordings.

RESULTS AND DISCUSSION

Soil properties

The particle size distribution (Physical indices, Atterberg limits and geotechnical classification are summarized in Table 1. The particle size distribution was determined following standard procedures, with tests performed both in natural conditions (without deflocculant – W/D) and with sodium hexametaphosphate as deflocculant (D) to assess the degree of aggregation typical of tropical soils. According to the Unified Soil Classification System (USCS), the soil was classified as MH (silt of high plasticity). However, the MCT classification is LG' (lateritic clayey soil), which. According to the MCT Classification, resistant soils exhibit lateritic characteristics (LG', LA', and LA). On the other hand, erodible soils exhibit non-lateritic behavior (NS', NA', and NA), resulting in higher percentages of silt and fine sand, considered low-cohesion materials. greater resistance to erosion is directly related to higher clay content in the soil. Fine soils tend to be less erodible due to greater cementation between particles, which hinders the development of erosion (Cassol et al., 2018; Couto & Gomes, 2020). Coarse soils, on the other hand, do not have cohesive forces capable of preventing detachment, but may present apparent cohesion.

Table 1
Particle size distribution, physical indices, Atterberg limits and geotechnical classification.

Soil characteristics	Values		Units
Liquid limit (wL)	56.0		(%)
Plastic limit (wp)	46.0		(%)
Plastic index (PI)	10		(%)
Gs	2.69
kh	3.42 10-7		m/s
γd​máx	13.75		kN/m3
wot	32.5		%
Particle size distribution
	W/D	D
Clay (C)	38	10	(%)
Silt (M)	35	45	(%)
Sand (S)	24	42	(%)
Gravel (G)	3	3
Geotechnical classification
USCS	MH
MCT	LG’

The results in Table 1 show that the presence of deflocculant caused a significant change in the values found. This difference can be attributed to the environmental conditions that gave rise to the soil, giving tropical soils greater cementation and particle aggregation capacity, as demonstrated by comparing the tests without deflocculant (W/D) and with deflocculant (D). Similar results are presented in other studies of tropical soils (Baudson et al., 2024; Couto, 2020).

Mineralogical analysis (XRD) identified kaolinite, illite, quartz, and hematite as the main constituents, which explains its reddish color (Munsell Color Company, 1975), low plasticity, and non-expansive behavior.

The results of the chemical analysis (Table 2) by X-Ray Fluorescence (XRF) – quantification of major oxides (%) and Loss on Ignition (LOI) – indicate that the sample is predominantly composed of silica (SiO₂), alumina (Al₂O₃), and iron oxides (Fe₂O₃), which are typical characteristics of highly weathered tropical soils, with possible presence of minerals such as quartz, illite, kaolinite, muscovite, and potassium feldspars.

Table 2
Mineralogical Composition of the Soil by XRF.

Óxide	Content (%)
SiO2	49.00
Al2O3	22.10
Fe2O3	14.80
CaO	< 0.01
MgO	0.17
TiO2	1.14
P2O5	0.11
Na2O	< 0.10
K2O	1.91
MnO	0.03
Total Oxides Sum (with LOI)	99.85
Loss on Ignition (LOI)	10.48

The high LOI content (10.48%) suggests a significant presence of minerals with structural water (clay minerals and hydrated iron oxides), supporting the presence of hydrated mineral phases previously identified. The low concentrations of CaO and Na₂O further indicate the absence of carbonate minerals or sodium feldspars.

Behavior of the physical model under wave action

The tests carried out on the first slope were structured in sequential stages, with systematic variation in wave frequency (3.6 to 4.4 Hz), water level (50.5 to 56 cm), application times (30 seconds to 1 hour), and rest intervals (3 minutes to 6 hours) between cycles, with the aim of identifying the critical parameters for fine particle transport and erosion in compacted tropical soil.

In the initial stages, only superficial transport of fine particles was observed, without significant changes in the slope structure. From the intermediate stages onwards, the increase in frequency and application time of the waves resulted in intensified erosion and particle displacement, making the erosive process more noticeable. In the final stages, with longer exposure time and continuous application of the waves, a significant increase in erosion was recorded, configuring conditions of progressive instability in the slope. These tests allowed the definition of threshold values ​​for wave height and period, as well as exposure time, which served as a reference for adjusting the parameters to be used in subsequent tests.

Based on data from the first slope, tests on the second slope were standardized to ensure controlled evaluation conditions. A single frequency of 4 Hz, a constant water height, and standardized times of 10 minutes per cycle were used. This standardization allowed isolating the effects of waves on the compacted slope, keeping the other parameters constant, thus enabling a detailed analysis of the influence of agitation on erosion characteristics and fine particle displacement in the reduced physical model. This made it possible to compare the erosive behaviors between the two slopes (Table 3), highlighting differences in erosion resistance, hydraulic behavior, and stability under simulated agitation conditions. This contributed to defining safe operating parameters for reservoirs and to understanding the performance of slopes constructed with compacted tropical soils.

Table 3
Comparison between Slope 1 and Slope 2.

Parameter	Slope 1 (Exploratory)	Slope 2 (Standardized)
Objective	Define critical patterns of wave action, water level, and exposure times for subsequent standardization of tests.	Evaluate erosive behavior under controlled conditions, using the parameters defined in Slope 1.
Wave frequency	Variable (3.6 to 4.4 Hz).	Single frequency (4 Hz).
Water depth	Variable (50.5 to 56 cm).	Constant (54.5 cm).
Application time	Variable (30 s to 1 h).	10 minutes per cycle.
Number of cycles	Variable (1 to 10 cycles per stage).	10 cycles.
Rest interval	Variable (3 min to 6 h).	20 minutes.
Main observations	Identification of thresholds for fine particle transport and onset of progressive erosion with increased frequency and exposure time.	Isolation of wave effects under constant conditions, allowing clear evaluation of erosion and fine particle displacement.
Purpose	Define critical agitation and drawdown parameters for simulation in a physical model.	Analyze erosive behavior under controlled and replicable conditions.

The two slope models exhibited distinct responses to wave action. Slope 1, built with fresh material and without disturbance, showed considerable resistance to erosion, with initial rates of 0.12 g/min stabilizing around 0.05 g/min after 60 minutes. This resistance is attributed to the micro-aggregated structure provided by iron and aluminum oxides. In contrast, Slope 2, which was disturbed during water level indicator installation and constructed with reused soil, was significantly more susceptible to erosion. Its erosion rates started at 0.48 g/min and remained high (0.22 g/min) after 60 minutes, demonstrating that soil disturbance drastically increases erodibility. Figure 4 shows the visual state of the slopes before and after testing.

Figure 4
Comparison of erosion between Slope 1 (a, b) and Slope 2 (c, d) before and after testing.

Table 4 presents the average values of total eroded volume after the end of testing for each slope, considering the different wave configurations tested. The results show that, regardless of the wave characteristics, the erosion resistance ranking remained constant, although the absolute magnitude of erosion varied significantly as a function of the wave parameters.

Table 4
Average total eroded volume after the end of testing for each slope.

Parameter	Slope 1	Slope 2
Initial volume of trapezoidal prism (m3)	0.224	0.392
Prism mass (kg)	308	539
Eroded mass (kg)	3.43	36.6
Eroded percentage (%)	1.114	6.793
Extrapolation at 1:30 scale (ton)	92.67	988.09
Test duration in model (min)	60	150
Real equivalent time (hh:mm)	05:29	13:41
Erosion rate in model per minute (g)	0.06	0.24
Erosion rate in real scale per minute (ton)	1.54	6.58

Based on the established scaling relationships the test results were extrapolated to real-world conditions, allowing for the estimation of erosion rates and potential slope deformations at full scale. For the adopted scale (1:30), this implies that:

The image analysis technique (Figure 5) allowed for the identification and quantification (Table 5) of lateral displacement of the slope during the wave tests.

Figure 5
(a) Initial profile “yellow line” of the slope with 45° inclination, (b) Final profile “white line,” and (c) Final eroded area of slope 2.

Table 5
Final result from image analysis – Real Scale (1:30).

Parameter	Value	Real Value (Prototype)
Eroded volume	0.01771 m3	478.19 m3
Dry mass	21.92 kg	592.97 tons
Wet mass (32.5%)	29.06 kg	784.71 tons

Figure 6 shows the 3D laser scanning depicting the erosion patterns, revealing the progressive formation of a natural slope (accretion) at the base of the slope from the deposition of eroded material. A linear relationship was observed between wave height and erosion rate, with a critical threshold for wave heights above 0.06 m (model scale). Waves with longer periods (0.8 s and 1.2 s) caused more widely distributed erosion along the slope face, while short-period waves (0.4 s) concentrated the erosion near the waterline.

Figure 6
3D digital model of the slope 2 after erosion.

Moisture sensors tracked the advance of a saturation front from the slope face inwards, leading to a significant reduction in matric suction and, consequently, a loss of soil strength. Cyclic fluctuations in moisture were observed, corresponding to the run-up and run-down of the waves (see Figure 7 and 8).

Figure 7
Moisture fluctuations – Slope 1.

Figure 8
Moisture fluctuations – Slope 2.

Effects of rapid drawdown and combined actions

The rapid drawdown tests showed that the combined valve configuration produced the highest hydraulic gradients, leading to the fastest water level reduction. This condition induced predominantly horizontal deformations on the slope face (Figure 9), consistent with classic drawdown-induced instability. The magnitude of deformation was directly proportional to the drawdown speed and inversely proportional to the soil's permeability. A synergistic effect was observed when slopes were subjected to wave action before drawdown; the pre-existing erosion and increased moisture content made the slopes more susceptible to failure during the subsequent drawdown.

Figure 9
cracks and fissures that appeared on the slope surface following the rapid drawdown test.

Turbidity analysis

Turbidity measurements (Table 6) confirmed the erosion observations. Slope 1 had a low average turbidity of 22.86 NTU. In contrast, the disturbed Slope 2 produced a much higher average turbidity of 177.7 NTU, an increase of over 675%, reflecting its structural fragility and high susceptibility to erosion.

Table 6
Turbidity of Slope 1 (Record 1 and 2) and Slope 2 (Record 3).

Sample	1	2	3	4	5	6	7	8	9	10	Média
Record 1(NTU)	18	17.7	17.3	28.2	27.3	24.2	22.1	26	24.4	23.4	22.86
Record 2(NTU)	39.5	29.2	27.6	25.8	40.6	30.8	30.3	28.7	27.9	36.4	31.68
Record 3(NTU)	137	142	141	207	209	204	148	145	149	295	177.7

For Slope 1, two independent measurement series were performed at different stages of testing to assess temporal evolution. Record 1, taken after initial wave tests (total exposure time: 30 minutes), yielded an average turbidity of 22.86 NTU with relatively low variability (standard deviation: 4.00 NTU). Record 2, taken after extended testing (total exposure time: 60 minutes), showed a slight increase to 31.68 NTU (standard deviation: 5.23 NTU). This 38% increase between records reflects cumulative erosion effects and progressive surface degradation, but the absolute values remain low, indicating good erosion resistance of the undisturbed lateritic soil structure.

For Slope 2 (Record 3), turbidity measurements were taken during the cycles of wave action, producing a dramatically higher average value of 177.7 NTU with high variability (standard deviation: 50.93 NTU). This represents a 675% increase compared to Slope 1's initial readings (Record 1) and a 461% increase compared to Slope 1's extended testing values (Record 2). This order-of-magnitude difference is directly attributable to Slope 2's compromised structural integrity resulting from piezometer installation and soil reuse.

The large difference in turbidity values between slopes quantitatively confirms the visual erosion observations and direct mass loss measurements. High turbidity indicates sustained fine particle detachment and transport, which not only causes immediate erosion but also suggests progressive weakening of the remaining soil structure. Individual water samples from Slope 2 testing showed turbidity values ranging from 137 to 295 NTU.

This distinct response highlights the critical role of initial slope stability in containing erosive processes, even at reduced scales. While the first slope resisted hydrodynamic forces more effectively, the second, previously weakened, exhibited intensified erosion and significant fine particle detachment into the water body.

These results emphasize the importance of ensuring homogeneous and controlled construction conditions in physical slope models, especially when aiming to comparatively analyze hydro-sedimentological phenomena. Furthermore, the impact of structural integrity on turbidity generation underlines the need for continuous monitoring and preventive measures, particularly in real-world geotechnical and hydraulic engineering projects.

Turbidity was statistically analyzed using the Statistical Process Control (SPC) method, as described by Montgomery & Runger (2012), Table 7 presents the mean values $(\overline{x})$, sample standard deviations (s), and the corresponding upper (UCL) and lower (LCL) control limits for each measurement.

Table 7
Process Control Statistics for Slopes 1 and 2.

	Slope 1 – Reading 1	Slope 1 - Reading 2	Slope 2 - Reading 3
𝑥̄	22.86	31.68	177.70
s	4.00	5.23	50.93
UCL	34.87	47.37	330.48
LCL	10.85	15.99	24.92

Slope 1 showed moderate and stable turbidity levels, while Slope 2 had much higher and more variable turbidity. One sample from Slope 2 (Sample 10) approached the upper control limit but remained within acceptable limits.

CONCLUSION

This research successfully developed and validated a 1:30 scale reduced physical model for analyzing the stability of dam slopes constructed with a lateritic tropical soil under hydrodynamic loading. The experimental approach provided valuable insights into the complex failure mechanisms.

The predominant failure mechanisms identified were progressive surface erosion, instability due to saturation and loss of matric suction, and localized rupture from rapid drawdown. These mechanisms were determined through convergent evidence from multiple complementary data sources:

Progressive surface erosion is evidenced by continuous mass loss, increased water turbidity, and gradual retreat of the slope face, with distributed erosion patterns confirmed by 3D laser scanning. Saturation-induced instability is supported by moisture sensor data showing the advance of the saturation front, with accelerated erosion rates when moisture content exceeds 80%. Finally, failure induced by rapid drawdown is documented by the formation of horizontal tension cracks and lateral deformations, the magnitude of which is inversely proportional to soil permeability, confirming classical drawdown theory.

The study confirmed that soil disturbance and material reuse dramatically increase susceptibility to erosion.

A direct relationship between wave height and erosion rate was established, with a critical wave height of 1.8 m (real scale) identified for the tested conditions. Longer period waves resulted in more distributed erosion, while shorter period waves concentrated damage near the waterline. The combination of wave action followed by rapid drawdown proved to be a critical loading scenario, highlighting a synergistic effect that significantly increases instability risk.

The Froude similarity criterion was validated as appropriate for this type of modeling, allowing for the reliable extrapolation of results to real-scale structures. The findings provide essential experimental data for calibrating numerical models and refining design criteria for dams built with tropical soils, contributing to the development of safer and more sustainable hydraulic infrastructure.

The experimental findings directly support the achievement of UN Sustainable Development Goals, particularly SDG 9 (Industry, Innovation and Infrastructure) by providing technical data for safer hydraulic structure design, and SDG 11 (Sustainable Cities and Communities) through improved understanding of infrastructure resilience against extreme hydrodynamic events. The research also contributes to SDG 15 (Life on Land) by addressing soil erosion processes and their mitigation.

DATA AVAILABILITY STATEMENT

Research data is available in the body of the article.

REFERENCES

• Agência Nacional de Águas e Saneamento Básico – ANA. (2016a). Guidelines for the development of dam projects Retrieved in 2025, May 19, from https://www.snisb.gov.br/portal-snisb/documentos-e-capacitacoes/relatorios
  » https://www.snisb.gov.br/portal-snisb/documentos-e-capacitacoes/relatorios
• Agência Nacional de Águas e Saneamento Básico – ANA. (2016b). Practical guide for small dams Retrieved in 2025, May 19, from https://www.snisb.gov.br/portal-snisb/documentos-e-capacitacoes/relatorios
  » https://www.snisb.gov.br/portal-snisb/documentos-e-capacitacoes/relatorios
• Almeida, M. R., & Lima, J. C. (2000). Comportamento geotécnico de solos tropicais sujeitos à erosão e ações ambientais. Revista Brasileira de Engenharia Agrícola e Ambiental, 4(2), 123-130.
• Andrade, L. N. P. S., Santino, M. B. C., Souza, C. A., Silva, F. L., Sousa, J. B., & Miranda, C. C. (2020). Marginal erosion: Geomorphological changes in the Teles Pires River (MT) resulting from the construction of a dam. Caminhos de Geografia, 21(75), 154-171.
• Araújo, J. C. (2003). Assoreamento em reservatórios do semiárido: modelagem e validação. RBRH, 8(2), 39-52. https://doi.org/10.21168/rbrh.v8n2.p39-56
  » https://doi.org/10.21168/rbrh.v8n2.p39-56
• Araújo, J. C., Fernandes, L., Machado Junior, J. C., Oliveira, M. R. L., & Sousa, T. C. (2006). Sedimentation of reservoirs in semiarid Brazil. In T. Gaiser, M. Krol, H. Frischkorn, & J. C. Araújo (Eds.), Global change and regional impacts (pp. 205-216). Berlin: Springer Verlag.
• Associação Brasileira de Normas Técnicas – ABNT. (2016a). NBR 6459. Solo — Determinação do limite de liquidez Rio de Janeiro: ABNT.
• Associação Brasileira de Normas Técnicas – ABNT. (2016b). NBR 7182. Solo — Ensaio de compactação Rio de Janeiro: ABNT.
• Associação Brasileira de Normas Técnicas – ABNT. (2017). NBR 7180. Solo — Determinação do limite de plasticidade Rio de Janeiro: ABNT.
• Associação Brasileira de Normas Técnicas – ABNT. (2018). NBR 7181. Solo — Análise granulométrica Rio de Janeiro: ABNT.
• Associação Brasileira de Normas Técnicas – ABNT. (2021). NBR 14545. Solo — Determinação do coeficiente de permeabilidade de solos argilosos à carga variável Rio de Janeiro: ABNT.
• Barra, O. A. L., Rabelo, D. R., Vasconcelos, F. P., & Casemiro, M. B. (2022). Evolução morfológica de um sistema praial induzido: Praia Mansa, enseada do Mucuripe, Fortaleza/CE. Boletín Geográfico (Neuquén), 40, 84-101. https://doi.org/10.4025/bolgeogr.v40.a2022.e60897
  » https://doi.org/10.4025/bolgeogr.v40.a2022.e60897
• Baudson, D. R. S., Ferreira, L. D., & Bacellar, L. A. P. (2024). Geotechnical analysis on the erodibility of tailings from the Fundão dam collapse. Geotechnical and Geological Engineering, 42(6), 4453-4475. https://doi.org/10.1007/s10706-024-02791-8
  » https://doi.org/10.1007/s10706-024-02791-8
• Bento, N. L., Amorim, J. S., Barros, F. M., & Silva, D. P. (2019). Descarga sólida em suspensão em uma bacia hidrográfica com apredominância de pastagem. Gaia Scientia, 13(3), 57-73. https://doi.org/10.22478/ufpb.1981-1268.2019v13n3.45307
  » https://doi.org/10.22478/ufpb.1981-1268.2019v13n3.45307
• Bertoni, J., & Lombardi Neto, F. L. (2010). Conservação do solo (7. ed.). São Paulo: Ícone.
• Carvalho, N. O. (1994). Erosão crescente na bacia do rio Doradas (Estado de Tachira, Venezuela) Rio de Janeiro: Furnas, Eletrobrás, Cadafe.
• Carvalho, N. O., Filizola Junior, S., Lima, P. M. C., & Lima, J. E. F. W. (2000). Guia de avaliação de assoreamento de reservatórios Brasília: ANEEL.
• Cassol, E. A., Silva, T. S., Eltz, F. L. F., & Levien, R. (2018). Soil erodibility under natural rainfall conditions as the K factor of the universal soil loss equation and application of the nomograph for a subtropical ultisol. Revista Brasileira de Ciência do Solo, 42(0), https://doi.org/10.1590/18069657rbcs20170262
  » https://doi.org/10.1590/18069657rbcs20170262
• Chanson, H. (2004). The hydraulics of open channel flow: an introduction (2nd ed.). Oxford: Elsevier Butterworth-Heinemann.
• Couto, B. O. C. (2020) Estudo da erodibilidade em horizontes de taludes de corte rodoviário por meio de procedimentos de laboratório e campo (Doctoral dissertation). Federal University of Ouro Preto, Ouro Preto (in Portuguese). Retrieved in 2025, August 20, from http://www.repositorio.ufop.br/handle/123456789/12359
  » http://www.repositorio.ufop.br/handle/123456789/12359
• Couto, B. O. C., & Gomes, R. C. (2020). Application of MCT methodology to determine soil erodibility in road cutting slopes in Quadrilátero Ferrífero region. Anuário do Instituto de Geociências, 43(1), 191-198. https://doi.org/10.11137/2020_1_191_198
  » https://doi.org/10.11137/2020_1_191_198
• Departamento Nacional de Estradas de Rodagem – DNER. (1994). DNER-ME 256/1994. Solo - Determinação do índice de expansão Rio de Janeiro: DNER.
• Departamento Nacional de Estradas de Rodagem – DNER. (2023). DNER-ME 228/2023. Solo – Determinação da massa específica seca máxima e do teor de umidade ótimo de solos utilizando o equipamento de compactação miniatura Rio de Janeiro: DNER.
• Departamento Nacional de Infraestrutura de Transportes – DNIT. (2023). DNIT 259/2023 – CLA: Solos – Classificação de solos finos tropicais para finalidades rodoviárias utilizando corpos de prova compactados em equipamento miniatura – Classificação Brasília, DF: DNIT.
• Eletrobras. (2003). Civil design criteria for hydroelectric power plants Rio de Janeiro: Eletrobrás.
• Empresa Brasileira de Pesquisa Agropecuária – EMBRAPA. (2020). Compressed files – Map of soil vulnerability to water erosion in Brazil Retrieved in 2025, August 20, from https://geoinfo.dados.embrapa.br/documents/7067/metadata_detail
  » https://geoinfo.dados.embrapa.br/documents/7067/metadata_detail
• Espósito, T., Naghettini, M., Ladeira, J., & Caldeira, L. (2010). Análise por árvore de eventos: ferramenta para gestão de risco em barragens de terra. Geotecnia, (120), 3-24. https://doi.org/10.14195/2184-8394_120_1
  » https://doi.org/10.14195/2184-8394_120_1
• Fiori, A. P. (2015). Fundamentos de mecânica dos solos e das rochas: aplicações na estabilidade de taludes São Paulo: Oficina de Textos.
• Gerscovich, D. M. S. (2016). Estabilidade de taludes (2. ed.). São Paulo: Oficina de Textos.
• Gil, M. M. L., Araújo, J. C., Montenegro, S. M. G. L., & Valença, J. M. M. S. (2020). Increase in water-scarcity risk in a Brazilian dry-region reservoir. Revista Caatinga, 33(4), 1025-1036. https://doi.org/10.1590/1983-21252020v33n418rc
  » https://doi.org/10.1590/1983-21252020v33n418rc
• Gomes, E. J. S., Lima, I. S., Castanheira Neto, P. P., Silva, F., & Pires, R. C. S. (2022). A utilização de solos tropicais em obras na engenharia. Epitaya E-books, 1(15), 66-79. https://doi.org/10.47879/ed.ep.2022540p66
  » https://doi.org/10.47879/ed.ep.2022540p66
• Hou, H., Wang, C., Chu, J., & Liu, H. (2021). Experimental study on the erosion resistance of sandy slope treated by microbially induced carbonate precipitation under wave actions. Marine Georesources and Geotechnology, 39(6), 692-701. https://doi.org/10.1080/1064119X.2020.1787961.
• Jones, L., & Anastasopoulos, I. (2021). Miniaturised tsunami generator to model interaction of tsunami with coastal infrastructure. International Journal of Physical Modelling in Geotechnics, 21(3), 135-149. https://doi.org/10.1680/jphmg.19.00021
  » https://doi.org/10.1680/jphmg.19.00021
• Khodja, Z. Y., Khanchoul, K., Altschul, R., Belkendil, A., & Djamai, Z. (2025). Assessing soil erosion and sediment yield in the Bou Rouina watershed (Northeast of Algeria): a RUSLE-SDR modeling approach. RBRH, 30, e19. https://doi.org/10.1590/2318-0331.302520240030
  » https://doi.org/10.1590/2318-0331.302520240030
• Luz, A. G. D., Polli, B. A., Bleninger, T. B., Lipski, B., Peixoto, E. B. D. A., & Warcheski, A. L. (2025). Comparison of methods for estimating reservoir sedimentation. RBRH, 30, e33. https://doi.org/10.1590/2318-0331.302520240143
  » https://doi.org/10.1590/2318-0331.302520240143
• Mariano, D. C. L., & Silva, J. B. (2022). Barragens de terra: características de seus alteamentos. Research. Social Development, 11(11), e277111133469. https://doi.org/10.33448/rsd-v11i11.33469
  » https://doi.org/10.33448/rsd-v11i11.33469
• Marinho, M. S., Costa, S. C. F. E., & Almeida, H. C. (2020). Development of web application for geotechnical classification of tropical soils based on MCT methodology. Revista de Engenharia Civil IMED, 7(1), 71-87. https://doi.org/10.18256/2358-6508.2020.v7i1.3373
  » https://doi.org/10.18256/2358-6508.2020.v7i1.3373
• Montgomery, D. C., & Runger, G. C. 2012. Estatística aplicada e probabilidade para engenheiros (5. ed.). Rio de Janeiro: LTC.
• Munsell Color Company. (1975). Munsell soil color charts Baltimore, MD: Munsell Color Company.
• Nguyen, T. H. T., Park, S. W., & Ahn, J. (2020). Numerical method to determine upstream scour slope in relation to turbulence and particle movement. Journal of Coastal Research, 36(1), 189-195. https://doi.org/10.2112/JCOASTRES-D-18-00130.1
  » https://doi.org/10.2112/JCOASTRES-D-18-00130.1
• Nogami, J. S., & Villibor, D. F. (1979). Soil characterization of mapping units for highway purposes in a tropical area. Bulletin of Engineering Geology and the Environment, 19(1), 196-199. https://doi.org/10.1007/BF02600475
  » https://doi.org/10.1007/BF02600475
• Rabelo, D. R., Araújo, J. C. D., & Cavalcante, A. A. (2025). Impacts of erosion and sedimentation on reservoirs in the Seridó river basin: a hydrosedimentological assessment in the brazilian semiarid region. RBRH, 30, e32. https://doi.org/10.1590/2318-0331.302520250001
  » https://doi.org/10.1590/2318-0331.302520250001
• Rabelo, D. R., Cavalcante, A. A., & Araújo, J. C. (2023). Sediment yield in a basin in Brazilian Semiarid Regions: a discussion on positive allometry. Catena, 221, 106749. https://doi.org/10.1016/j.catena.2022.106749
  » https://doi.org/10.1016/j.catena.2022.106749
• Ribeiro, R. S. (2021). Resultados Preliminares do Atlas de Riscos Geológico e Hidrológico do Estado do Espírito Santo (Vol. 5, Erosão Continental, Erosão de Margem Fluvial e Erosão). Brasília: CPRM.
• Santos, M. S. (2019). Análise da suscetibilidade e vulnerabilidade aos processos erosivos da bacia hidrográfica do rio São Francisco no estado de Sergipe (Monograph). Retrieved in 2025, August 20, from http://ri.ufs.br/jspui/handle/riufs/12124
  » http://ri.ufs.br/jspui/handle/riufs/12124
• United Nations. (2015). Sustainable development goals Retrieved in 2025, August 20, from https://unfoundation.org/what-we-do/issues/sustainable-development-goals/
  » https://unfoundation.org/what-we-do/issues/sustainable-development-goals/
• Wisser, D., Frolking, S., Hagen, S., & Bierkens, M. F. (2013). Beyond peak reservoir storage? A global estimate of declining water storage capacity in large reservoirs. Water Resources Research, 49(9), 5732-5739. https://doi.org/10.1002/wrcr.20452
  » https://doi.org/10.1002/wrcr.20452
• Yang, W., Zhang, G., Yang, H., Lin, D., & Shi, P. (2023). Review and prospect of soil compound erosion. Journal of Arid Land, 15(9), 1007-1022. https://doi.org/10.1007/s40333-023-0107-3
  » https://doi.org/10.1007/s40333-023-0107-3
• Zhu, L., Cui, S., Pi, X., Luo, L., Cheng, J., Wang, H., Du, R., & Yang, Q. (2023). Quantitative investigation on localized deformation process of rocks by uniaxial test and digital image correlation. Environmental Earth Sciences, 82, 267.

Edited by

• Editor-in-Chief:
  Adilson Pinheiro

• Associated Editor:
  Iran Eduardo Lima Neto

Publication Dates

• Publication in this collection
  16 Mar 2026
• Date of issue
  2026

History

• Received
  30 Aug 2025
• Reviewed
  05 Jan 2026
• Accepted
  13 Jan 2026