Health Assessment of Zoned Earth Dams by Multi-Epoch In Situ Investigations and Laboratory Tests

Abstract

The long-term safety and operational reliability of zoned earth dams depend on the structural integrity of their internal components, including core, filters, and shell zones. This is particularly relevant for old dams which have been operational for a long period of time. Such existing infrastructure systems are exposed to various loading types over time, including environmental, seepage-related, extreme event, and climate change effects. As a result, even when they look intact externally, changes might affect their internal structure, composition, and possibly functionality. Thus, it is important to delineate a comprehensive and cost-effective strategy to identify potential issues and derive the health status of existing earth dams. This paper outlines a systematic approach for conducting a comprehensive health check of these structures through the implementation of a multi-epoch geotechnical approach based on a variety of standard measured and monitored quantities. The goal is to compare current properties with baseline data obtained during pre-, during-, and post-construction site investigation and laboratory tests. Guidance is provided on how to judge such multi-epoch comparisons, identifying potential outcomes and scenarios. The proposed approach is tested on a well-documented case study in Southern Italy, an area prone to climate change and subjected to very high seismic hazard. The case study demonstrates how the integration of historical and contemporary geotechnical data allows for the identification of critical zones requiring attention, the validation of numerical models, and the proactive formulation of targeted maintenance and rehabilitation strategies. This comprehensive, multi-epoch-based approach provides a robust and reliable assessment of dams’ health, enabling better-informed decision-making workflows and processes for asset management and risk mitigation strategies.

1. Introduction

Earth dams are critical infrastructure systems playing an important role for the global economy. Most zoned earth dams, in Western countries, including Italy, were built during the second half of the last century [1] and can be classified as “old structures” (they are now in their “old age stage”), as they have been in operation for decades. Foster et al. [2] show that under static conditions, damage to earth dams occurs immediately after the end of construction, or several years later. Furthermore, most earth dams were designed before the introduction of modern seismic codes/guidelines and performance-based design approaches [1]. As a result, they were designed according to older codes and often without considering seismic actions. This aspect, combined with the need for managing institutions to ensure dam functionality, emphasizes the importance of performing seismic re-evaluations using numerical models which need advanced constitutive models of the dams’ materials [3]. Therefore, for many dams that are in these conditions it is necessary to carry out new and updated field and laboratory tests. Zoned earth dams are complex structures with different materials (core, filters, and shells) designed for specific functions. Their behavior can change over time due to various stressing factors which can be identified as loading conditions and time-dependent factors, such as water level fluctuations, seismic events, aging, and environmental conditions. The evaluation of the behavior of earth dams usually relies upon real-time monitoring of measurable quantities (e.g., pore pressures and deformations). A truly comprehensive health assessment of zoned earth dams should integrate such monitoring programs with periodic re-evaluations of fundamental soil mechanics parameters which are influenced by the above-mentioned “forcing factors”. Such a re-assessment can differentiate expected responses from anomalous behaviors which can profoundly affect the stability and long-term integrity of the dam. When performing these analyses, it is not just about how a parameter changes over time at one point. The most important factor is related to changes across the various zones of the dam and how these changes evolve spatially over time [4,5,6]. This helps pinpoint localized issues, such as developing seepage paths or zones of unexpected and/or accelerated settlements [5]. In this re-assessment approach a fundamental key-point is represented by baseline data: large datasets of soil parameters obtained during initial pre-construction site investigation and as part of during-construction investigation programs. This data ensemble establishes the above-mentioned “as-built” properties. Such baseline properties will then be used to check whether values gathered afterwards show any significant changes that could impact dam safety. From this comparison it is necessary to establish whether the change in the soil parameter values of a specific dam zone is significant and if it is capable of influencing the overall behavior of the dam [4,5]. This re-assessment approach of soil mechanics parameters can be combined with monitoring data to check the current status of the dam and identify anomalies. Such an approach can lead to diagnosing triggering mechanisms of these changes and potentially design timely mitigation strategies.

As mentioned above, many existing zoned earth dams were built during the second half of the last century. As a result, for each of them, the quantity of available documentation is extremely large. However, because such structures were designed and constructed decades ago, and the people involved in such processes are now older or have passed away, the direct knowledge about the structure is typically limited to a reduced number of people [7]. According to Jappelli [8], when the age of these structures becomes significant, the archive information becomes rare and confused and the results of monitoring systems less accurate and incomplete. The Farneto del Principe dam, in Southern Italy, exemplifies these challenges. It was designed in the early 1960s, constructed between the late 1970s and early 1980s, and has been in operation since 1989. As expected, available data for this dam are restricted not only in terms of quantity but also in terms of quality. To perform a seismic re-evaluation, required by the new Italian regulations, between 2015 and 2017 a new substantial geotechnical investigation program was conducted. The data gathered during this post-construction geotechnical program, along with legacy data from pre- and during-construction investigations, constitute a large dataset of geotechnical characterization information. Currently, there is no guidance on how to perform such a re-assessment quantitatively and holistically. In this paper, a step-by-step approach, based on a well-documented case study, is presented and discussed. This approach is presented in the subsequent section. It is then applied to an existing infrastructure: the Farneto del Principe zoned earth dam. The goal is to compare the geotechnical parameters of the 2015–2017 investigations with the baseline data and identify any significant changes that could impact dam safety. The term “significant” in this context assumes a rather important role. Such significance, in past practices, was usually assessed holistically and qualitatively. In this study, a specific quantitative definition is presented and used. To support this definition, specific references from the proposed case study and the open literature are proposed and analyzed. From the approach it is possible to capture both the temporal evolution of parameters at specific locations and the spatial variation across different zones of the dam. This approach can be used to monitor the current “health status” of a dam and make informed decisions regarding maintenance, rehabilitation, and/or operational adjustments.

2. Re-Assessment Framework

Zoned earth dams are complex structures which combine different soil types in a specific arrangement to effectively retain water, control seepage, and maintain structural stability, making them a highly efficient and widely used type of dam. They are constantly subjected to various “forcing factors” or external loads and environmental conditions, including climate change, that can challenge their integrity. These factors can be categorized as (1) hydraulic factors including water level fluctuations, rapid drawdown, seepage forces, and overtopping (uncontrolled flow of water over the dam crest); (2) ground shaking during seismic events generating excess pore water pressure, potentially leading to reduced shear strength causing liquefaction in saturated granular soils, liquefaction-induced lateral spreading, seismic compression, cyclic softening in saturated clays, and landslide instability phenomena; and (3) environmental factors, such as rainfall, wind, weathering, aging, and animal activity.

These forcing factors can influence geotechnical parameters of soil materials. By evaluating the change in these parameters it is possible to check the dam’s health. Based on a vast experience gained as part of seismic re-evaluations performed on various earth dams in Southern Italy, it was possible to define a novel protocol to perform existing earth dams’ health assessments. To make it efficient and usable in future studies and applications, this protocol is presented as a step-by-step approach and summarized in the flow chart shown in Figure 1:

(1)

Collection and organization of available geotechnical data from pre- and during-construction investigation programs.

During the design and construction phases, extensive geotechnical investigations are conducted to determine geotechnical parameters of the various soils used to construct the dam and in the foundations. Such data forms the “as-built” baseline properties for each dam’s zone and for the foundation materials. As highlighted by various authors [7,8], such data could be inherently incomplete for “old dams”. These data are usually not available in a digital form, making their accessibility limited and their collection time consuming.

(2)

Collection and organization of data from periodic geotechnical investigations.

During the operational life of the dam, geotechnical investigations are usually performed for different reasons: routine maintenance, instrumentation installation, in response to observed anomalies, and updated technical standards mandating such investigations. Earth dams are usually long structures with extended longitudinal dimensions. As a result, new data from in situ investigations and laboratory tests (on samples gathered during the investigation program) should be collected following three criteria: (1) filling potential gaps at locations where such data are not available from step 1, (2) performing the same investigations at the same locations where data are available from step 1 to assess any multi-epoch changes, and (3) when performing new investigations in response to detected anomalies, such data should be targeting locations potentially related to the observed/perceived anomalies. When anomalies are detected, new investigations should be carefully designed to capture invisible criticalities, such as underseepage and permeability nonhomogeneity issues (e.g., by collecting samples at a depth potentially related to permeability anomalies and/or nonhomogeneities) [5].

(3)

Comparative multi-epoch analysis.

This step is needed to identify significant changes and variations in trends in time and space. The values of the soil parameters obtained during step 2 (investigations performed during the operational life of the dam) are compared with the baseline data from pre- and during-construction investigations. It is important to identify significant changes in the parameter values. The term “significant” in this context is related to the potential effects such variations can cause to the behavior and performance of the dam. As a result, this assessment phase is necessarily based on holistic engineering judgement and/or experience with similar dams. It is also important to compare the parameters across different zones at different epochs to identify areas of unusual behavior. Such an assessment helps with identifying temporal and spatial variations. Guidance on a quantitative assessment which can help with defining “significant” changes in measured parameters can be provided by standard deviation values on measured geotechnical quantities from the literature. In this approach, standard deviations provided by Jones et al. [9] were used to help develop a holistic data-informed assessment. Their uncertainties are specifically derived for performance-based design approaches. As such, they are well-suited to be used in this context. As more dam-specific data become available in the open literature, such uncertainty levels can be re-assessed and/or customized for these infrastructure systems, making the proposed framework even more efficient.

(4)

Identification of unexpected behaviors influenced by anomalous parameter variations.

In case of significant multi-epoch and/or spatial changes in one or more parameters, it is important to identify the potential phenomena that can be triggered by such changes. This assessment is based on the analysis of baseline vs. new data and that of potential spatial changes performed in step 3. This interpretation should be focused on determining the causes triggering parameter variations and evaluating the potential impact of the observed parameter changes on the dam’s behavior and performance. For instance, a localized increase in the water content or changes in grain size distribution in a particular area may indicate the development of seepage paths or material degradation. Likewise, progressive changes in Atterberg limits and shear strength can be a signal of mechanical weakening due to internal erosion, aging of materials, or shifts in the groundwater regime.

(5)

Correlation with monitoring quantities.

When a behavioral change is identified in step 4, parameter changes and monitored quantities should be analyzed together to find whether they are indicative of the same phenomenon. This analysis is integral to the comprehensive health check of the dam and should lead to diagnosing the behavior change cause.

(6)

Ad hoc integration of the monitoring system and issue-specific investigations.

To track the evolution of issues identified in step 5, it may be useful implementing new monitoring systems and/or new investigations. Such systems and investigations are rapidly evolving with the integration of cutting-edge technologies that improve accuracy, efficiency, and safety. Useful monitoring technologies which were recently used in the study of dams are those reliant upon radar-based satellite technologies, such as the analysis of multi-epoch synthetic aperture radar (SAR) data [10,11,12], robotic total stations and GNSS (Global Navigation Satellite System) [13], terrestrial laser scanning [14,15], Light Detection and Ranging (LiDAR) data [15], and combined-technique approaches from satellite and terrestrial data [16,17]. The combination of existing and new data (the latter installed specifically to track identified ongoing issues) provides a more holistic and robust picture of dam’s health and improves accuracy, reliability, and spatial coverage.

(7)

Health assessment and risk evaluation: recommendations and remedial actions.

The synthesis of all steps leads to drawing conclusions about the current dam’s conditions. In case of proven ongoing issues, appropriate actions should be proposed and priorities in potential interventions delineated. Decision makers would then take actions based on such data-informed recommendations.

The proposed multi-step re-assessment framework relies upon multi-epoch investigations of zoned earth dams. It was derived using data, experience, and knowledge acquired during seismic re-assessments of earth dams located in a high seismic hazard area (Southern Italy). However, it is intended as a general framework applicable to similar contexts (highly tectonized areas prone to seismic hazard) as well as to different geological and/or tectonic settings (e.g., moderate seismicity areas and stable continental zones). This generalizability is ensured by the fact that all major potential data-gathering techniques, investigation types, monitoring systems, nonhomogeneities, and instability mechanisms are taken into account and discussed for each step of the proposed framework. Furthermore, in highly anthropogenic geotechnical systems such as zoned earth dams, the original geological setting is not as important as for natural systems (such as natural slopes). The steps illustrated in this section are vital for a comprehensive health assessment. Following this approach would lead to a proactive identification of potential issues leading to failure mechanisms. Furthermore, it is possible to quantitatively assess risks and ensure the long-term safety and functionality of these critical infrastructure assets. Finally, using the data collected and organized following the proposed framework, it is possible to create digital twins. These digital models can then be updated and used as benchmarks when additional data become available over time. Digital twins can also be useful to create complex two- and three-dimensional numerical models. To illustrate how the proposed approach should be implemented, an existing dam is analyzed. The following sections provide details on the main characteristics of the case study and on the available data from investigations performed at various points in time and from the dam’s monitoring system.

3. The Farneto del Principe Dam

3.1. Dam Description

The Farneto del Principe dam is located in the Calabria region (Southern Italy), in the town of Roggiano Gravina (Latitude: 39.6515° N–Longitude: 16.1627° E). The dam is located in a seismically active region influenced by both shallow crustal and deep seismic activity, the latter associated with the subduction zone of the Calabrian arc [18]. It is a zoned dam, used for irrigation purposes and flow balancing, characterized by a central impervious core comprising compacted clay and silt and by two shells composed of compacted sand and gravel. Protective filters, each approximately one meter thick, are positioned on both sides of the core—sand on the core side and a combination of sand and gravel on the shell sides. The maximum normal operating water surface level is 136.30 m above sea level (a.s.l.), the crest elevation is 144.40 m a.s.l., and the maximum allowable level of the reservoir is 141.70 m a.s.l. The dam height is about 30 m, while its length is more than 1200 m.

The dam is founded on alluvial materials (high permeability sands and gravels) which in turn overlay a deep stiff clay bed. The thickness of the foundation materials is not constant but varies lengthwise, with a maximum thickness of around 43 m in the central area of the dam. A cut-off wall is present throughout the length of the dam to prevent underseepage. The cut-off wall was built using two construction strategies: (1) two slurry walls formed by panels excavated in the presence of the bentonite mud, and (2) a double line of close piles with a half-meter diameter, without injection of waterproof material. The cut-off wall is embedded at least 3 m into the clay bed for the whole longitudinal extension of the dam. Downstream of the core an inspection tunnel is located that is also used as a conduit for the collection of the drained water coming from the dam. Figure 2 shows a representative cross section while the main geometrical characteristics of the dam are reported in

Table 1. Main geometrical characteristics of Farneto del Principe dam.

Geometrical Property	Value
Water storage volume	46 Mm3
Average height	27.7 m
Crest length	1240 m
Crest width	7 m
Upstream face slopes	1:2.5; 1:3; 1:3.5
Downstream face slopes	1:1.185; 1:2.25
Crest elevation	144.20 m a.s.l.
Maximum allowable water level	141.60 m a.s.l.
Maximum authorized water level	136.30 m a.s.l.

.

3.2. Geotechnical Characterization During the Design and Construction Phases

During the design phase of the dam, several laboratory tests, such as grain size distribution, direct shear, and triaxial and oedometric tests, were performed on undisturbed and/or reconstituted samples of the foundation materials (both clay bed and alluvium) and on materials which were candidates for the construction of the dam body. During the construction phases, additional laboratory tests were performed on the actual materials used to build the dam body. The core material, after a period of exposure to natural weathering, was compacted in 30 cm thick layers by means of six runs of a 60-ton dump truck and additional runs of a 35-ton smooth roller. The continuity of the layers was ensured by a scarification of the current compacted surface prior to the placement of a new layer. Water contents were accepted in a relatively narrow range (−3% to +2% around optimum). Construction logs show values (obtained by means of oedometric tests) of the hydraulic conductivity ranging from 1 × 10⁻⁷ to 4 × 10⁻⁹ cm/s. The shell material was compacted in 80–100 cm thick layers by means of a vibrating roller of 8 ton with a frequency of 1500 rpm (revolutions per minute). Additional details on pre- and during-construction field investigations and laboratory tests are provided in [19,20].

3.3. The 2015–2017 Geotechnical Investigation Program

In the period 2015–2017, as part of the seismic re-evaluation of the Farneto del Principe dam, a new geotechnical field investigation and laboratory test program was conducted. This new investigation program had the objective of identifying any ongoing issues and/or anomalies. The step-by-step framework presented in the previous section and discussed in this paper originated from the experience of the Authors with this and similar activities. The 2015–2017 geotechnical investigation program was specifically designed to fill existing gaps in terms of the quantities and quality of the available as-built data. Furthermore, a portion of this new program was designed to mimic the tests performed during the pre- and during-construction geotechnical program. These data would then be used to make comparisons between the as-built baseline and the new data affected by aging and other forcing loadings that the dam experienced since its construction. Finally, new data were obtained from modern geophysical investigation types. Such data cannot be directly compared with their as-built counterparts as these technologies were not available back then. However, they contribute to expand and deepen the level of knowledge about the dam’s health status. This new investigation program was planned according to the three criteria reported in step 2 of the proposed framework. The 2015–2017 geotechnical investigation program comprises the following test types:

(a)

Boreholes with continuous sampling and standard penetration tests (SPTs);

(b)

Static laboratory tests on undisturbed clay samples;

(c)

Piezocone and seismic cone penetration tests (CPTu and SCPTu, respectively);

(d)

Seismic tomography tests;

(e)

Down-hole and cross-hole tests;

(f)

Multichannel analysis of surface waves (MASWs);

(g)

Dynamic laboratory tests on undisturbed specimens (resonant column and cyclic torsional shear tests);

(h)

Microtremor horizontal-to-vertical spectral ratio (HVSR) analysis.

The locations of the aforementioned tests are reported in Figure 3. Fifteen boreholes were drilled during the field investigation, eight on the dam crest, three on the landside shell, one near the guardhouse, one by the lateral overflow spillway, and two in the landside, in a free-field condition. For ten of the fifteen boreholes (S1, S2, S3, S4, S5, S6, S7, S8, S12, and S13) SPTs were performed and five boreholes were equipped with Casagrande piezometers (Figure 3).

In addition to boreholes, SPT, CPTu, and SCPT, eight seismic tomography tests were performed. Five of them capture the two-dimensional distribution of the shear wave velocity along longitudinal profiles: one is located along the crest, one near the guardhouse, two along the inspection gallery, and one in a free-field area on the landside of the dam. The remaining three tomographies are meant to characterize typical cross sections. One is located on the right bank (108 m long), one on the centerline (102 m long), and the last one on the left bank (96 m long). Additional geophysical tests (down-hole, cross-hole, and MASWs) were performed to define the shear wave velocity profiles of various zones of the dam. Five down-hole tests were performed: in boreholes S3 (landside shell), S4 (crest), S13 (guardhouse), and S1 and S8 (the last two in a free-field area). The S4 down-hole reached a depth of 26 m, investigating the clay core up until the cut-off wall. The S3 down-hole extended for 40 m, crossing both the dam shell and the alluvial foundations, while S1, S8, and S13 reached the clay bed limit. Four cross-hole tests were performed, using three boreholes (S4, S5, and S6—these boreholes have an inter-boring spacing of four meters). Two tests were performed, utilizing two different source–receiver permutations (i.e., one time the source was in S4 and the receivers in S5 and S6 and then the source in S5 and the receivers in S4 and S6). An inclinometer probe was used to ensure the verticality of the boreholes. The depth reached was 26 m, which is roughly the location of the concrete cut-off wall. The tests were performed following the ASTM Standard procedures (D 4428).

Several dynamic laboratory tests on undisturbed specimens were also conducted on the dam crest materials (the specimens were collected in the following boreholes, S4, S5, S6, S11, and PZ3, following the ASTM Standard procedures, D 4015). Resonant column (RC) and cyclic torsional shear (CTS) tests were used to estimate the shear modulus of the clay core, its variation with depth, and its reduction with increasing shear strain. The CTS tests were performed using a frequency of 0.1 Hz and various numbers of cycles (ranging from 5 to 20). A summary of all the investigations and tests performed is reported in

Table 2. Summary of the tests performed at the Farneto del Principe dam site performed during the 2015–2017 investigation program (adapted from [6]).

Test Type	Number	Location
Borehole	15	Core, shells, foundations
SPT	10	Core, shells, foundations
CPTu	4	Core
SCPTu	4	Core
Seismic tomography	8	Longitudinal and transversal profiles
Cross-hole	4	Core
Down-hole	5	Core, shells, foundations
MASW	1	Core
Resonant column	14	Core
Cyclic torsional shear	8	Core
Microtremor HVSR	5	Core, inspection gallery, foundation

(adapted from [6]).

4. Application of the Proposed Re-Assessment Framework to the Case Study

4.1. Comparison Between the Grain Size Distribution Curves

A zoned earth dam relies upon different materials with specific grain size distributions to perform their intended functions. Generally, fining or coarsening can occur. A shift towards finer particles (fining) could indicate a loss of coarser particles due to internal erosion or suffusion, while a shift towards coarser particles (coarsening) could indicate migration of fines. It is well-known that changing the grain size in a dam core can significantly impact its hydraulic performance and can have significant consequences on the stability of the dam (e.g., [4]).

Figure 4a shows the comparison between the grain size distribution curves for the dam core from the 2015–2017 investigations and those obtained during the construction of the dam (indicated as “as built”). The “as-built” values were obtained by tests performed during the construction of the dam, every 5000 m³ of installed material. The baseline as-built and the more recent curves (retrieved from tests in the period 2015–2017) are practically identical. Thus, the groundwater flow within the dam body did not alter the core particle size distribution. This result suggests that neither internal erosion nor the migration of soil particles took place [20]. The same comparison was performed for the shells of the dam body (Figure 4b) in which the values “as built” were obtained examining every 15,000 m³ of installed material. Comparisons were also carried out on the foundation materials (alluvium, Figure 4c; clay bed, Figure 4d). All the pre-/during-construction and 2015–2017 curves are essentially identical for all the dam’s zones. This comparison highlights that no significant soil migration/dragging phenomena occurred during the operational life of the dam since 1989 (year of its initial impounding).

4.2. Comparison Between Water Content, Plastic Limit, Liquid Limit, and Plasticity Index in the Dam Core

To obtain more information on the health status of the dam, comparisons were performed between the parameters that carry a strong influence on the dam’s behavior during construction (compaction) and in the subsequent operational phase. These parameters are water content, plasticity limit, liquid limit, and plasticity index.

4.2.1. Water Content

An increase in the water content in the impervious core over time, particularly if coupled with a rising phreatic line or increased seepage, could indicate increased saturation, reduced density (due to internal erosion or degradation), and reduced shear strength. Conversely, a significant decrease in the water content in saturated soil could suggest the development of desiccation cracking. Small fluctuations tied to reservoir levels are normal. Figure 5a shows the trend of the baseline as-built water content with the depth in the dam’s core along with that measured during the 2015–2017 geotechnical investigations. The “as-built” values were obtained as a synthesis of measurements performed every 500 m³ of installed material during the construction. The mean water content from the measurements performed in 2015–2017 is 20.5%, which is slightly higher than the mean water content measured during the construction phase (17.88%). During the construction phase, the water contents were accepted in a relatively narrow range (−3% to +2% around optimum), with the optimum ranging between 16 and 20%. The difference between the baseline data and those measured in 2015–2017 is narrow and comprises the optimum acceptance limit prescribed during construction. Furthermore, such a discrepancy is within a standard deviation of the inherent soil variability and two standard deviations of the measurement error variability according to Jones et al. [9] for this parameter.

4.2.2. Plastic Limit

Significant changes in the plastic limit (PL) over time can indicate significant alterations in the clay core material, which in turn affect its performance and stability. However, its value is related to the plasticity index. Therefore, its change must be seen in relation to other parameters. An increase in the plastic limit can be due to drying, piping, or internal erosion, while a decrease in this parameter can be due to external water coming into the system and/or chemical alterations. Figure 5b shows a comparison of the measured PL with depth. The mean value of the 2015–2017 PL is equal to 20.8%, similar to the value measured during the construction phase, which is equal to 19.18%. This slight discrepancy is well within a standard deviation of the inherent soil variability and of the measurement error variability according to Jones et al. [9] for this parameter. Based on these values, the water content and PL for this dam’s core assumes a behavior between the semi-solid and plastic states [20]. Furthermore, based on this multi-consideration assessment, no significant changes occurred to these parameters during the dam’s operational life.

4.2.3. Liquid Limit

A decrease in the liquid limit (LL) may be very concerning. A reduction in the LL (and, consequently, of the plasticity index, PI) indicates that the clay has become less plastic, more brittle, and potentially more prone to cracking. This can be caused by leaching or chemical alteration of clay minerals, and it can lead to increased susceptibility to cracking, to increased permeability, and to potentially altered shear strengths. An increase in the LL, which is a less common occurrence, could lead to increased compressibility and settlement and swelling potential. Figure 5c shows the LL with depth in the dam’s core. The mean value of the LL measured in 2015–2017 is 49.36%. The mean baseline as-built value is 45.4%. These values indicate high swelling [21]. Similar to the PL, this slight discrepancy is well within a standard deviation of the inherent soil variability and roughly within a standard deviation of the measurement error variability according to Jones et al. [9] for this parameter. As a result, no significant change has occurred to the LL. Therefore, none of the above-mentioned concerning phenomena are likely to be observed.

4.2.4. Plasticity Index

As mentioned earlier in this section, monitoring the plasticity index (PI) of different materials within a zoned earth dam through multi-epoch investigations is a cornerstone of an effective dam’s health assessment. As a result, this is an important step in the proposed framework. This parameter is related to the LL and PL but much more important than them taken individually because it is a more comprehensive indicator of the plasticity of a clay. It relates directly to several critical soil properties affecting the dam’s stability and performance, including shear strength, permeability, compressibility and swelling/shrinkage, susceptibility to internal erosion, and piping. A significant change in the PI over time can signal mineralogical alteration, chemical changes, and weathering or leaching. In case a significant change in the PI is observed over time, the following related parameters should be further investigated and analyzed to identify the source of the change being observed and prevent undesired developments: water content, piezometric data, seepage measurements, and chemical analysis of pore/seepage water.

Figure 5d shows the PI with depth. The 2015–2017 measurements produce values ranging between 24.7% and 32%, with a mean value equal to 28.56%. The baseline as-built mean PI value is 26.16%. This difference is well within a standard deviation of the inherent soil variability and of the measurement error variability according to Jones et al. [9]. Furthermore, these values fall within the range that characterizes the plasticity index (PI = 15–30%) of the core materials of five Italian earth dams examined by Lanzo et al. [22]. These values indicate a robust protection against cracking [23] and high swelling potential [24].

4.3. Casagrande Plasticity Chart and Activity Index

The analysis of the Casagrande plasticity chart is another useful step to understand the plasticity of soil and its potential for swelling and shrinkage, and its permeability, which are crucial for the dam’s stability and performance. The plasticity chart (Figure 6b) of the Farneto del Principe dam core materials indicates medium plasticity. No differences between the as-built data points and those from the 2015–2017 geotechnical tests are visible. Furthermore, they are consistent with those of the other five Italian dams [22].

The activity index of clay (often simply called “activity,” denoted as “A”) is a very important parameter in geotechnical engineering, particularly for assessing the behavior of fine soils like those used in a dam core and for identifying the swelling or shrinking potential. The activity index of clay in a dam core is generally considered a relatively stable intrinsic property determined during the design and construction phases and is not variable over time. The colloidal properties depend on both the amount of clay and mineralogy and therefore are related to parameter A. Figure 6a shows the activity chart of the clay comprising the Farneto del Principe dam’s core. The activity index values measured during the 2015–2017 geotechnical investigations present an upper bound of 0.96 and a lower bound of 0.61, with a mean of 0.79 and a standard deviation of 0.11. Based on the values of A, the core clay can be classified as inactive–normal [25]. The values of parameter A did not vary during the dam’s operational life (Figure 6a).

4.4. Void Ratio in the Dam Core

Analyzing the void ratio (e) is extremely useful in assessing the performance of an earth dam during the construction phases and during its operational life. It depends on the compaction degree and influences soil conductivity, compressibility, and shear strength. Initial compaction during construction aims at minimizing the void ratio, ensuring a dense and impermeable core. However, over time, seepage forces can lead to internal erosion, washing away fine particles and increasing the void ratio, potentially leading to instability and piping particularly in the lower part of the core. Freeze–thaw cycles can also alter the void ratio; in fact, compacted material may exhibit an increase in the void ratio.

Figure 7a shows the void ratio values with depth obtained during the 2015–2017 geotechnical investigation program versus the bounds of the same parameter from pre- and during-construction tests. A slight increase in the void ratio measured in the dam’s core is observed. However, by itself, this is not a critical indicator of potential dam instability. Since a change is observed in this parameter, further geotechnical investigation programs will need to target this quantity to check its multi-epoch stability in the future.

4.5. Degree of Saturation in the Dam Core

The degree of saturation (S) in a clay dam core generally undergoes a complex trend over time. This trend can be influenced by various factors. During construction, the phenomena of compaction, first impoundment, and consolidation influence S. Its long-term behavior is influenced by reservoir fluctuations due to seasonal changes, operational demands, or drought/flood cycles. A rising reservoir level generally increases S and pore water pressures within the core with delay, depending on the hydraulic conductivity of the core. Similarly, a rapid drawdown influences S because the phreatic line within the core may lag behind the external water level. Therefore, the actual trend of S over time is dynamic. Internal processes such as creep, deformation, aging, degradation, desiccation, and differential settlements can influence the pore structure, consequently potentially altering its distribution in space. Monitoring instruments such as piezometers are crucial for tracking pore water pressures and inferring saturation levels within the core to ensure the dam’s safety.

Figure 7b shows the trend of S with depth of the core obtained during the 2015–2017 geotechnical investigations program. Figure 7b shows a trend compatible with the filling of the reservoir that can be obtained only with an appropriate numerical groundwater flow analysis. The absence of anomalous values of S represents an indication of the good health of the dam.

4.6. Hydraulic Conductivity of the Dam’s Clay Core

Calibrating the hydraulic conductivity of the clay core is very important when designing a zoned earth dam. It ensures water tightness and guides the static performance of such massive infrastructure systems. While a decrease in hydraulic conductivity is beneficial, several processes can lead to an undesirable increase, such as desiccation cracking, differential settlement, hydraulic fracturing, internal erosion, freeze–thaw cycles (localized increase near crest), root intrusion, and animal burrowing.

Figure 8 shows the values of the vertical coefficient of hydraulic conductivity (kv) in the baseline as-built conditions and those measured during the 2015–2017 program. In this case, the “as built” range of the hydraulic conductivity should be considered less reliable because it was derived from the work management documents and not from the test results. This is the case where the archive information is unclear as predicted for the “old dam” by Jappelli [8]. In Figure 8, kv values are plotted versus the depth for five vertical effective stress levels used in the oedometer tests, ranging between 196 kPa and 3138 kPa. Figure 9a shows kv versus void ratio at different depth levels. Similar trends are shown in Figure 9b, where kv is plotted as a function of the vertical effective stress. As expected, at every depth, kv increases as the void ratio increases and the vertical effective stress decreases. In addition to measurements performed in the oedometer, other measurements of kv were performed during the 2015–2017 geotechnical program by means of falling head tests carried out in the oedometer cell for two selected depths, 4.0 m and 10 m, and are indicated in Figure 8 with blue squares. The measurements using this test type are consistent with those obtained by using the oedometer test at the same depths. Within the core of the Farneto del Principe dam, the hydraulic conductivity did not increase over time. It actually decreased slightly. This multi-epoch discrepancy lies within the bounds proposed by Jones et al. [9] for inherent soil variability. As such, even though this would be a beneficial modification, this is not a significant difference. Multi-epoch differences in this quantity are expected to have an opposite trend as, after compaction, kv is expected to increase over time. This effect can be explained by the fact that it takes several months for water, after the construction, to flow within the core. As a result, during this period, thixotropic effects can take place [20]. These results together are representative of a good overall health status of the dam’s core. It is well-known that higher LL and PI in the clay material used for a dam core generally contribute to achieving lower hydraulic conductivity. Benson and Trast [26] studying the influence of several factors (initial saturation, compactive effort, PI, LL, and clay content) on hydraulic conductivity of compacted clays showed trends between hydraulic conductivity and PI and LL (Figure 10a,b). Figure 10a shows the values of the hydraulic conductivity for different depth levels and vertical effective stresses as LL varies. Figure 10b, similarly, shows the values of the hydraulic conductivity for PI. The values of the hydraulic conductivity of the dam’s core obtained during the 2015–2017 investigations follow the trend proposed by Benson and Trast [26]. Figure 10 also shows that for some of the low values of the vertical effective stress, the hydraulic conductivity assumes higher values than those proposed by Benson and Trast [26]. In the other cases, the hydraulic conductivity assumes much lower values than their predicted counterparts.

Based on a machine learning approach, Tan et al. [27] performed a feature importance analysis showing that the void ratio after compaction, fines content, specific gravity, S after compaction, and PI of soils are the top-five factors (in descending order) that influence the hydraulic conductivity of compacted soil. This ranking should be considered when analyzing multi-epoch and multi-zone results for clay cores of existing earth dams.

In addition to analyzing the vertical coefficient of permeability (kv), we also analyzed the influence of permeability anisotropy within the core. Such an analysis relied upon measurements of the coefficient of the horizontal hydraulic conductivity (kh), normal to the direction of compaction forces. These measurements were performed during the 2015–2017 geotechnical investigation program by falling head tests in the oedometer cell, for various stress levels, for a sample taken at a depth of 22 m, and the following values were derived: kh = 4.6 × 10⁻⁹ to 9.7 × 10⁻⁹ cm/s. For this dam’s core at this depth, the kh is slightly greater than the kv. This outcome confirms that the core material, compacted at the optimum water content, shows a partially oriented structure, influenced by the water flow direction [20].

4.7. Compressibility of the Dam’s Clay Core

The compressibility of a dam’s clay core is very important when assessing its long-term performance. It directly influences how much the dam will settle over time, how stresses are distributed within the structure, and, ultimately, its safety and stability. The compressibility of clays in a dam’s core is a complex property influenced by several factors, such as clay mineralogy, plasticity (LL and PI), grain size distribution and fines content, and compaction parameters. The primary laboratory test for measuring the compressibility of clay is the one-dimensional consolidation test (oedometer test). Figure 11a shows the results of the compression tests performed during the 2015–2017 investigations on undisturbed samples from different depths in the dam core. Figure 11b summarizes the trend with depth of the Cc (compression index) and Cr (recompression index). It can be noted that Cc values fall within the range of 0.1 to 0.3 typical of low-plasticity clays (CL) and that Cr values are practically constant with depth. Figure 11c shows the overconsolidation ratio (OCR) trend with depth primarily due to the combined effects of compaction-induced stresses and increasing overburden pressure. In the shallowest parts (upper layers) of the core, the OCR is typically very high (high stress from compaction and low stress from overburden); in deeper portions of the dam core (greater depth), increasing current overburden produces a decreasing OCR. These trends are well-known in the literature since they are obtained from in situ tests, such as the flat dilatometer test (DMT) [28].

4.8. Effective Cohesion (c′) and Friction Angle ϕ’

Strength parameters (cohesion, c’, and friction angle, ϕ’) of compacted clays are crucial for dams’ long-term stability. In general, c’ tends to decrease significantly over time, primarily due to saturation and potentially due to wetting/drying cycles causing cracking and structural degradation. On the other hand, ϕ’ is generally more stable. Figure 12a,b show the trend of strength parameters within the Farneto del Principe dam’s core obtained by direct shear and triaxial CIU tests during the 2015–2017 investigations. A mean value of 22° was estimated for ϕ’. This value is larger than the value of 18° obtained and used for the stability analysis during the design phase of the dam. Similar considerations can be made for the cohesion. Analyzing these two parameters, no issues are evident in the dam’s long-time performance.

4.9. Groundwater Flow Analysis

Significant changes in the above soil parameters may result in increased seepage. As a result, direct measurements and insights into seepage quantities are extremely important. To ensure that the original water tightness of a zoned dam is still robust, measurements of the seepage water collected through the dam’s drainage system are necessary. In the Farneto del Principe dam, this quantity can be measured using the channeling and collection system located in the inspection tunnel, along the longitudinal axis of the dam. When the quantity of the discharged water increases over time, the dam’s water tightness may be reduced [4].

Figure 13 shows the values of the discharge seepage water (i.e., drained water collected in the inspection tunnel) in the period January 2013–December 2015. Ausilio et al. [19] showed that in a previous period of observation (e.g., between 1991 and 2012), the maximum measured discharged water was 0.53 l/s. Comparing these data points, it is possible to conclude that the measured quantities of the discharged water are consistent over time and that such a quantity is hardly influenced by the water level in the reservoir. The relatively low values and the constant trend over time of the discharged water is an indication that water tightness in this dam is still ensured. When this quantity evolves over time, there may be a red flag, requiring advanced analyses to identify whether an increased amount of discharged water is a localized occurrence or a pervasive feature. For this reason, real-time monitoring of this quantity is suggested and encouraged.

4.10. Chemical Analysis of Discharge Water

During their operational life, dams may undergo significant changes. Several anomalous variations in soil parameters can be caused by chemical alteration. This sub-section analyzes chemo-related changes, which are usually neglected in traditional geotechnical analyses. Fell et al. [29] and Tabatabaei [30] show that chemical reactions (e.g., sulphide oxidation) may play a detrimental role in the dam’s performances and stability. This effect is due to potential reductions in the dam’s shear strength parameters. To check whether such changes are an actual occurrence, two water samples were taken in 2015 from the reservoir and in the drainage tunnel, respectively, of the Farneto del Principe dam. These water samples were then analyzed using chemical–physical tests (

Table 3. Chemical–physical test.

Parameters	Water (Reservoir)	Water (Tunnel)
PH	8.1	8.09
Conductivity (mS/cm)	437	534
Total Hardness (°f)	21.2	23.8
Calcium Hardness (°f)	12.3	9.8
Magnesium Hardness (°f)	8.9	14
Ca2+ (mg/L)	49.3	39.28
Mg2+ (mg/L)	21.63	34.03
Alkalinity (mg/L CaCO3)	190	225

). The water in the drainage tunnel (i.e., discharged water after seepage through the dam’s shell and core is completed) presents higher values of conductivity, alkalinity, and magnesium ions. On the other hand, the water taken in the reservoir (i.e., water which did not interact with the dam’s material) presents a larger quantity of calcium ions. This result was not expected. However, this mechanism can be explained by the fact that calcium ions precipitate, forming calcium formations (stalactites) that are visible along the entire inspection tunnel. As a result, when analyzing chemical interactions between water and soil within a zoned earth dam, such occurrences (i.e., calcium formations) should be noted when present as they can be related to potentially counterintuitive outcomes of the chemical analyses.

4.11. Shear Wave Velocity Measurements

An important step in the proposed re-assessment framework is represented by the analysis of the results obtained by modern tests. Such tests (and related measured/inferred parameters and data) were usually not performed during the pre- and during-construction investigations. As such, for these quantities, which usually include shear wave velocity profiles of the dams’ materials, no multi-epoch comparisons are possible yet. However, such data can become a baseline that can be used in subsequent future analyses. This is convenient, especially because such quantities can be obtained using non-invasive techniques. To evaluate the performance of the shells and foundation materials, during the 2015–2017 geotechnical investigation program, down-hole tests were performed. Figure 14a,b show the shear wave velocity profiles (S1 and S8 locations, both in a free-field area). These profiles can be considered representative of the foundation alluvial material. The scatter in the data is related to the natural variability of this material and its intrinsic heterogeneity. It is important to notice that, along the longitudinal direction of the dam, the thickness of this material varies. As a result, these two profiles cannot be directly compared.

Figure 15 shows the shear wave velocity profiles in the dam shell and in the alluvial foundation (locations S3 and S1, respectively). As expected, in both cases the shear wave velocity increases with depth. The maximum measured value was 700 m/s in the alluvial material. For down-hole S3 there is an abrupt discontinuity in the profile at the boundary between the shell material and the alluvial foundation material. This discontinuity is not as abrupt in the data from down-hole S1. The absolute value of the shear wave velocity of the alluvial foundation material measured beneath the shells is not the same as that measured in the free-field areas. Figure 15 shows such a direct comparison for down-holes S1 and S3. The shear wave velocity of the alluvial material beneath the dam is substantially higher than that in the free-field area at the same elevation. It takes about 20 m for the shear wave velocity in the alluvial material in free-field conditions to reach the value measured in the same unit immediately beneath the dam. This comparison highlights the expected effect of overburden pressure on shear wave velocity profiles [6].

5. Conclusions

The analysis of multi-epoch geotechnical investigation data offers a great opportunity to assess the health and long-term performance of zoned earth dams. By comparing data collected over different time periods and in different zones of a dam, it is possible to draw conclusions regarding material integrity and to provide a deeper understanding of the dams’ condition, performance, and stability. In this paper a systematic step-by-step framework is proposed and outlined. This comprehensive approach captures both the temporal evolution of parameters at specific locations and the spatial variation across different zones of the analyzed dam. The proposed approach builds upon the extensive experience of the Authors with several zoned earth dams in Italy and constitutes a holistic framework that combines qualitative, judgement-based, data-informed, and data-based quantitative considerations. In summary, the systematic use of geotechnical parameters derived from multi-epoch field and laboratory investigations offers a powerful diagnostic tool for assessing the health status of existing earth dams. It also provides a robust technical foundation for timely operational decisions, for planning structural rehabilitation or safety upgrades, and for informing and interacting with non-technical audiences, such as decision makers and stakeholders. The integrated and augmented multi-epoch geotechnical knowledge gained following the proposed framework directly feeds into the dam’s risk assessments. By continuously monitoring and updating the dam’s “health status,” dam owners and engineers can make informed decisions regarding maintenance, rehabilitation, or operational adjustments. This can significantly reduce project risk exposure, future consequences, and post-disaster residual risk assessments.

The proposed approach is applied to a well-documented case study, the Farneto del Principe dam in Southern Italy. The soil mechanic properties which characterize the behavior of the zoned earth dam obtained during recent geotechnical investigations are compared with the correspondent properties obtained during the construction phase. The latter constitute the baseline as-built threshold. The comparison between the baseline data and the new investigations shows no “significant” changes in the values of most of the parameters for the analyzed dam. However, through the lens of this case study, this paper provides insights on what to look for when analyzing such multi-epoch data and illustrates potential damage mechanisms related to each potential significant variation.

In addition to the assessment of the current performance and health status of a zoned earth dam, the re-evaluation approach proposed in this paper allows one to update, calibrate, and validate numerical models and digital twins of such critical infrastructure systems. Resulting calibrated models can then predict dams’ behaviors under various conditions (e.g., full reservoir and seismic events). Multi-epoch data are also useful for continuous validation of these predictions against actual observed behaviors, allowing for refinement of the model and identification of discrepancies. Geotechnical models (e.g., finite element analysis, finite difference analysis, and limit equilibrium methods) are used to simulate dam behavior. Multi-epoch data are crucial for calibrating and validating these models. As new data become available, the models are updated and refined, leading to more accurate predictions of future performance and stability. This iterative process helps understanding the complex interactions between various factors affecting dam’s stability, such as seepage, stress distribution, and deformation.