Earth fissures are a significant threat to various types of infrastructure in the Southwestern United States and elsewhere in the world. They are caused by differing subsidence rates and magnitudes associated with groundwater pumping in alluvial basins. In this article, Ravi Murthy and his co-author Jon Benoist, both with the Arizona Department of Water Resources, examine the impact that earth fissuring has on embankment dams used for flood control. One of their main dilemmas as regulators is: "Is it possible to safely rehabilitate dams in fissure risk zones to continue to provide economical flood protection to the public, or does the risk of dam failure due to an undetected fissure present too great a threat to the public?"

EMBANKMENT DAMS IN EARTH FISSURE RISK ZONES - A REGULATOR'S DILEMMA

Jon M. Benoist, P.E., Arizona Department of Water Resources, Phoenix, AZ
Ravi Murthy, P.E., Arizona Department of Water Resources, Phoenix, AZ

Abstract

Excessive groundwater withdrawal from alluvium-filled basins in the southwest has triggered large-scale ground subsidence in the Southwestern United States. Differential compaction of the alluvium has induced tensile stresses in the ground. Earth cracks or fissures have developed at locations where the induced tensile stresses exceed the tensile strength of the soil. Earth fissures associated with groundwater withdrawal and ground subsidence have been identified in Arizona, California, and Nevada. Historically, agencies such as the U.S. Army Corps of Engineers (USCOE) and the Soil Conservation Service (SCS) have avoided constructing, or have suggested abandoning embankment fill dams across known fissures and in areas of known high potential fissure risk. Presumably, this approach was taken to avoid potential catastrophic dam failure from erosion along a fissure in the dam foundation, or an associated embankment crack. Despite this cautious and conservative stance, single purpose flood control dams remain operational in developing fissure risk zones in Arizona.

In 2005, the Arizona State dam safety regulator is faced with thousands of people living in housing located immediately downstream of existing flood control embankment dams located in recently developed or developing fissure risk zones. The primary dilemma facing the state dam safety regulator: Is it possible to safely rehabilitate dams in fissure risk zones to continue to provide economical flood protection to the public, or does the risk of dam failure due to an undetected fissure present too great a threat to the public?

Typically, failure of a dam in an urban environment presents a significant risk of losing a large number of lives and extensive property damage. Clearly, a very low level of risk for a dam failure must be the primary objective for the state dam safety regulator, given his responsibility to protect the lives and property of the public. However, if dam rehabilitation is deemed to be unsafe due to fissure risks, the dam safety regulator must require the community to remove the existing dam. The only remaining alternative for the community is to construct significantly more expensive flood control alternatives, such as large floodways or flood basins. The problem is this path also results in risks to the public as it is likely that it will be years before funding is available to provide safe flood protection. Given the increased costs and immediate loss of flood protection, the removal of an existing dam will put the dam safety regulator at odds with the public as well as other government entities sensitive to the immediate needs of the public for adequate and economical flood protection. The authors believe that today's dam safety regulator has a dilemma in that they cannot simply categorize the presence or potential presence of a fissure risk zone as a fatal flaw for rehabilitation of an existing flood control dam.

The variables that the dam safety regulator must consider with regard to fissure risk zones are complex and include such issues as understanding fissure development, erodibility of the foundation soils, adequacy of state-of-the-art computer models and laboratory testing to assess magnitudes of erosion, adequacy of state-of-the-art monitoring to detect existing and developing fissures, and the ability to implement safe repairs to the foundation fissures after they are detected. Where all the factors indicate a safe dam rehabilitation, the dam safety regulator must still appreciate the major unknowns in the present state of the art for fissure evaluation. Advances will occur in the science and engineering related to building and maintaining safe dams in fissure zones. Thus, an essential part of the regulatory approval of a dam rehabilitation in a known or potential fissure risk zone is a periodic full reassessment of all technical factors to confirm the safety of the dam.

This paper discusses these issues from a regulator's point of view in light of two proposed embankment rehabilitation projects at single-purpose flood control dams within fissure risk zones in Maricopa County, Arizona.

Introduction

Since the 1940s, federal agencies such as the Soil Conservation Service ([SCS], now Natural Resources Conservation Service [NRCS]) and the United States Army Corps of Engineers (COE) have constructed flood control dams in the Southwestern United States. For the most part, these dams were designed and built to protect downstream agricultural lands. Some of the more recently-constructed dams however, were multipurpose, serving as flood protection structures for agricultural lands as well as urban centers and military installations.

Over the past two to three decades, and since the construction of these flood control dams, urban centers downstream from these dams have grown significantly. This urban growth has significantly increased the downstream consequences related to potential failures of these flood control dams. The increased risk for major potential loss of life associated with potential dam failure scenarios has forced owners, operators, and regulatory agencies to pay closer attention to the safety of flood control dams in Arizona and other areas of the Southwest. The risk is further compounded by the deterioration of the dams themselves. Deterioration of the dams has been caused by:

• Transverse cracking due to desiccation and foundation settlement
• Loss of freeboard due to ground subsidence related to excessive groundwater extraction
• Earth fissures associated with differential ground subsidence

This paper focuses on the potential adverse impacts of earth fissures on the safety and long-term performance of single-purpose flood control embankment dams and the creative approaches taken by the dam owner, owner's engineer and the dam safety regulator to achieve safe dam rehabilitation alternatives.

Subsidence & Earth Fissuring

Mechanics of Subsidence

Ground subsidence is known to occur in the alluvium-filled valleys in Arizona. This land subsidence is a gradual settling of the earth's surface caused by subsurface movement of earth materials. The United States Geological Survey (USGS, 1999) estimates that over 80 percent of the subsidence reported in the United States is caused by over-drafting of groundwater resources. Studies indicate that subsidence in Arizona is also largely attributed to excessive groundwater withdrawal.

Historically, water needed for agricultural activities in the semi-arid Southwest has been obtained by pumping of groundwater wells installed within the very deep, alluvium filled valley aquifers. Groundwater withdrawal at rates in excess of the natural replenishment of these aquifers leads to the lowering of the groundwater table. At some locations in Arizona, groundwater levels have been lowered by over 300 feet. Dewatering of the alluvium results in an increase in inter-granular (effective) stress on lower layers, causing the alluvium to consolidate. This consolidation manifests itself as a subsidence bowl at the ground surface. The magnitude of subsidence is directly related to the subsurface geology, the thickness and compressibility of the alluvium, and the magnitude of groundwater declines. Ground subsidence due to groundwater overdraft is essentially irreversible; however, the rate of subsidence can be reduced or arrested by reducing or halting declines in groundwater levels (Bouwer, 1977).

Information pertaining to the major subsidence bowls in Arizona is summarized below:

Subsidence in the area of a flood control dam typically lowers the dam crest, decreasing the storage capacity of the reservoir and the available embankment freeboard for the design flood, producing an unsafe overtopping potential. The loss of freeboard develops over a period of years (if not decades) and rarely leads to an overtopping dam failure, providing the owner has an active crest monitoring and repair program. A more insidious impact of subsidence on embankment dams, however, is the formation of foundation earth fissures. Even with specific monitoring activities, earth fissures are more difficult to detect than subsidence itself. A rapid reservoir filling during a flood event combined with the presence of earth fissure in the foundation and potential associated cracking in the earth dam can trigger catastrophic erosion of the fissure and failure of the dam.

Mechanics of Earth Fissure Development

Earth fissures develop near the margins of subsidence bowls and/or in the vicinity of a buried rock pediment edge, where differential subsidence induces tensile stresses in the ground that exceed the tensile strength of the unconsolidated alluvium. While fissures are generally associated with ground water declines in excess of 300 feet, earth fissures have also been identified in areas of lesser groundwater declines. The location of earth fissures is primarily controlled by the configuration of the buried bedrock surface, variation in the basin fill stratigraphy, and the location and characteristics of the subsidence area.

Figure 1 depicts a generalized sequence of earth fissure development. Generally, it is believed that the fissure is initiated as a relatively narrow crack that is formed when the induced tensile stresses exceed the tensile strength of the soil. Published information indicates that uneroded fissures range from "hairline" to 2 inches in width, and are typically less than 1-inch wide. Figure 2 shows a typical earth fissure in the early stage of development, where the fissure is still hidden below the surface. Subsequently, seepage from surface infiltration erodes the sides and top of the fissure until it reaches the surface". Figure 3 shows the initial surface manifestation of an earth fissure which typically includes small depressions, potholes, and surficial cracks oriented along a generally liner alignment. Runoff into these surface features causes significant and rapid enlargement of the fissure through a combination of erosion and slumping of the fissure walls. Finer grained soils are washed vertically into the fissure that theoretically extends to the groundwater table. Generally, the depth of erosion at a fissure has been limited to several tens of feet, which may be due to larger gravels and cobbles plugging the fissure or due to an increase in the resistance of the soils to erosion with depth. Figure 4 depicts a mature fissure.

Figure 1: Stages in Fissure Development (from Larson and Pewe, 1986)

Figure 2: Uneroded Fissure

Figure 3: Surface Potholes and Cracks along a Fissure

Figure 4: Mature Earth Fissure

The surface erosion features associated with mature fissures are known to exceed 20 feet in depth and 30 feet in width. The exposed lengths of fissures at the ground surface are typically less than 1 mile, but one fissure in the Picacho Basin in South Central Arizona is more than 9 miles in length. Generally, the great length of the mature eroded fissure results from areal runoff into the fissure. However, at least one case of extensive lateral erosion of a fissure has been observed (Weeks, 2005) for hundreds of feet in each direction from a relatively small point source of water discharging into the fissure.

Earth Fissures & Potential Dam Failure Modes

The primary failure modes in embankment dams related to earth fissures results from seepage erosion along the fissure in the dam foundation and uncontrolled release of the reservoir. A second failure mode is the uncontrolled release of the reservoir due to erosion of the embankment where the expression of the fissure results in a corresponding crack in the dam. A third failure mode is overtopping of a section of dam due to slumping of the dam into the eroding foundation.

Observations at fissure sites in Arizona have shown that an open fissure can "accept" hundreds of cubic yards of soil generated by erosion of the sidewalls, prior to plugging of the fissure. Consider a fissure through the foundation soils at an embankment dam where the fissure extends into the dam reservoir, providing an access point for the impounded water to intersect the fissure. Water flowing along the fissure under hydraulic reservoir head most certainly results in significantly greater erosion than observed where surficial runoff is the primary mechanism of erosion.

For the condition where the dam embankment retains its integrity and is able to span the eroded fissure, the erosion forms a "tunnel" under the dam. Depending on the geotechnical characteristics of the embankment soils, a portion of the dam may also erode. If the embankment dam is unable to span the eroded foundation fissure, significant cracking and slumping of the upstream face and crest of the dam occurs. Reservoir water flowing along these cracks may lead to piping and breach of the dam in addition to the erosion breach of the dam foundation. Furthermore, depending on the freeboard available and the depth of impoundment, the crest of the dam may slump, resulting in overtopping and subsequent breaching of the dam.

An extensive review of the literature has not identified any conclusive case histories related to impacts of earth fissures on the safety and long-term performance of embankment dams. However, the failure of Picacho Dam in Arizona may have been triggered by an earth fissure. Figure 5 shows the failure tunnel eroded in Picacho Dam (ADWR; file photo).

Figure 5: Picacho Dam Breach

The Regulator's Dilemma

Massive urbanization is presently occurring in the Southwest and is consuming agricultural lands and low density rural lands previously protected by flood control dams constructed under programs administered by the SCS and USCOE. Most of these dams extend for miles, crossing numerous, typically dry, alluvial fan drainages. Historically, these desert agricultural lands have been irrigated by over-drafting groundwater. Present groundwater levels have declined hundreds of feet resulting in surface subsidence and associated ground fissures in close proximity to the existing dams. These older dams have become at-risk for failure due to recently developed or developing fissure risk zones. These dams are even more problematic to the state regulator because they are now being reclassified from low or significant hazard, when protecting agricultural lands, to high hazard due to the downstream urbanization.

The Arizona State dam safety regulator is faced with thousands of people living in or moving into new housing developments located immediately downstream of existing, singlepurpose, flood control embankment dams located in known and developing fissure risk zones. These homeowners are often unaware of the need for flood protection or even the existence of the upstream flood control dam. Although owners typically conduct public outreach programs in advance of major dam and flood control projects, it is likely that many individuals living downstream are not aware of the threat to their lives and property from a potential dam failure where these older dams are located in existing or potential fissure risk zones. These homeowners are also likely to be extremely unhappy and vocal if the dam safety regulator requires the removal of an unsafe dam and their homes are suddenly declared to be without flood protection.

The regulator's dilemma is complex and difficult to quantify: Can flood control dams in fissure risk zones be safely rehabilitated to the same level of risk associated with embankment dams located outside of fissure risk zones? If this is not practical, is a higher level of risk still acceptable? Should this risk be commensurate with the risk a community faces when it is without flood protection for a period of years, until funding is available for other more expensive types of flood control protection? Is it safe to rehabilitate these dams or must they be breached? The dam safety regulator obviously understands the need for community flood protection and will be at odds with the immediate community needs if it is determined that an existing flood control dam cannot be safely rehabilitated, especially if the community does not have readily available funding for other flood protection alternatives.

Historically, the construction of a new dam or the rehabilitation of an existing dam has not been allowed by regulating agencies in existing or developing fissure risk zones. Fissure occurrence in the foundation of an earth dam and or the potential extension of the fissure feature into the embankment has been considered a fatal flaw, since the dam will likely fail catastrophically due to erosion upon filling. Historically, there has not been a practical method of detecting fissures developing in the foundation of a dam and thus, there has been no practical means of monitoring the safety of a dam as impacted by the occurrence of a fissure. Since an unexposed fissure could not be detected, it was also not possible to implement the rehabilitation of a fissure in the foundation of a dam. The discovery of a fissure has typically been the sudden emergence of a large, linear surface depression associated with the collapse and erosion of the surface soils into the fissure from surface runoff. The dam safety regulator cannot accept a dam failure related to the sudden collapse and the erosion of an undetected fissure beneath a dam when a major flood suddenly fills the reservoir.

If dam rehabilitation cannot be safely accomplished, the dam safety regulator is forcing the community to pursue significantly more expensive alternatives such as a large floodway or flood basin project. Limited flood control budgets would stall development of control structures and any development would have to proceed such that the public would be in defined flood plains and incur millions of dollars in flood insurance costs. Given limited available funding and multiple projects requiring rehabilitation - is the greater public safety served by rehabilitating a dam to a failure risk that still may be of significance but is much less than that posed by the existing dam, or by waiting until public funding is available to construct a clearly low risk flood protection alternative, leaving the public exposed to the risk of flooding for a number of years before funding becomes available? Clearly, the dam safety regulator's primary responsibility and objective is to be confident that the risk of a dam failure is very low and there is not an unacceptable risk to loss of life and property. If the dam safety regulator perceives that the risk is unacceptable and the dam must be breached, it is highly likely that this decision will be at odds with the lay public and potentially other government entities that are sensitive to the dissatisfaction of the public.

Recognizing the immediate need that the ongoing massive urbanization process has for adequate and economical flood control protection, the dam safety regulator must search for new and creative means to provide safe dams for flood protection. To provide economical levels of adequate flood protection the regulator is forced to seriously consider the construction or rehabilitation of dams in fissure zones. It is no longer possible to simply consider the occurrence of a fissure as a fatal flaw. The dam safety regulator in Arizona has had to ask the question with respect to rehabilitation of dams in fissure risk zones: Are elements present today within the context of investigations, structural designs, monitoring, rehabilitation, and future developments in all technical areas related to fissures that, when incorporated together, comprise a safe flood control dam?

The variables that the regulator must consider today in locating a dam in a fissure risk zone include many related elements. These elements are delineated below and are investigated and evaluated in close coordination with the owner and his design engineer.

1. Are fissure development and occurrence understood? Is there a slow development or a sudden development of a fissure of a given width?
2. Can potential fissure risk zones be identified based on surface settlement and subsurface geology?
3. Is it is possible to remotely detect fissure development below the ground surface?
4. What width of fissure can be remotely detected below the ground surface?
5. At what depth below the ground surface can fissures be remotely detected?
6. Which types of earth materials will express a fissure as a result of differential settlement?
7. Which types and sizes of the components of earth materials are susceptible to large levels of erosion?
8. Are the mechanics of fissure erosion understood? Do the same mechanics apply for a small amount of runoff versus the volume of water contained in a reservoir?
9. Is modeling computer code developed to simulate erosion on a fissure? Are laboratory tests and field tests of soil erosion developed where the erosion parameters are accurate enough for input to the model computer code? Is the seepage flow gradient occurring initially and during erosion of a fissure understood and incorporated into the model computer code? Is cementation and micro-structure occurring in soils incorporated into the erosion parameters input to the model computer code?
10. Can structural elements such as hardened dam sections or cut off walls and upstream aprons be adequately modeled and evaluated for increasing the safety of the dam against erosion failure on a fissure?
11. Can state-of-the-art monitoring, such as GPS surveys, low sun-angle photography, time-domain reflectometry cables, etc. detect existing and developing fissures?
12. Is it practical to safely remediate the adverse impacts of a fissure after it is detected in the foundation or reservoir of a dam?

Investigative & Analytical Tools

Where known earth fissures, or the significant risk of an earth fissure zone are identified at the site of an existing dam, and the continued use of the dam is the alternative of choice, the regulator's first goal must be to have the owner attempt to detect the fissure and implement remedial measures before the fissure erodes due to exposure to surface runoff or hydraulic head. In addition, the owner must also demonstrate that in the event that the fissure is not detected, the geotechnical conditions and embankment characteristics at the site are such that erosion of the fissure will not result in a catastrophic release of the reservoir. With a strong responsibility to safety of the public, the regulator must be convinced, typically through the efforts of the owner and the owner's design engineer, that the investigative and analytical tools used to develop and evaluate the proposed designs meet both current state-of-the-practice and state-of-the-art at the time of design and construction.

As the entity responsible for review and approval of the proposed designs, the dam safety regulator must be aware of the current state-of-the-practice and technical advances in the fields of fissure detection, geotechnical site characterization, and erosion modeling. This section of the paper discusses some investigative and analytical tools that have been used at dam sites. The existence and application of some of these tools have been known to designers for some time now but may have a modified application in the field of fissure studies. This section also presents some new tools that have been developed exclusively for the study of earth fissures.

Surface Seismic Refraction Surveys

Rucker (2000) has developed a method to detect the presence of a fissure in the subsurface soils using seismic refraction techniques. The presence of an earth fissure or other discontinuity in the soil manifests itself as an anomaly in the time record for the geophones. The anomaly may be a delay in the arrival time of the compression wave, and/or attenuation in signal strength as the energy traverses the subsurface discontinuity. This technique has recently been used very effectively by Rucker to investigate cracks at two FCD dams in Arizona. This technique holds significant promise for fissure detection in that relatively large areas or alignments can be surveyed rapidly. The results of the surface seismic refraction survey is utilized to focus intrusive subsurface investigative techniques such as trenching to visually confirm and describe an existing fissure.

Geotechnical Investigations

Thorough and detailed geological and geotechnical characterization of the site is considered to be a critical element of evaluating the feasibility of embankment dam rehabilitation in a fissure risk zone. Site investigations play a crucial role in identifying earth fissures (if present), and also provide relatively undisturbed samples for evaluating the erosion and other geotechnical characteristics of the foundation soils. Accurate location and proper backfilling of test trenches and boring must be included within the investigation. This section presents a brief discussion on some of the investigative tools used in fissure risk zones.

Test Trenching

Test trenching provides an opportunity for direct visual observation of the subsurface soils and represents the best available technique to study earth fissures, especially those fissures that have no, or limited, surface expressions. In areas where the risk associated with earth fissures is high, continuous trenching within the fissure risk zone may be warranted. At other locations, the scope of trenching may be reduced through use of seismic refraction techniques discussed above. It is imperative that test trenches be carefully logged and documented by an engineering geologist with experience and expertise in evaluating ground subsidence and earth fissures.

Hollow Stem Auger Borings

Hollow stem auger borings can be used to explore and collect samples for general geotechnical characterization of the subsurface, but have very limited value in actual fissure detection. Furthermore, depending on the site-specific subsurface characteristics, it is generally difficult to obtain undisturbed soil samples for laboratory testing.

Soil Coring

Soil coring using a track-mounted drill rig equipped with a HQ triple tube coring system has been used successfully to core unconsolidated alluvial soil deposits. The HQ system obtains a 2.4" core sample and drills a 4" hole. This method was successfully used to obtain relatively undisturbed samples at four embankment dam sites in Arizona. Core recovery was generally 100 percent, except in zones of relatively clean, uncemented sands, where the soils appear to have been washed away by the drilling fluid.

Modified Pitcher Sampling

A modified Pitcher sampling technique has been used to collect relatively undisturbed samples of alluvial soils. The sampling system consists of an outer barrel fitted with a carbide cutting bit. A spring-mounted Shelby tube is housed inside the barrel. The cutting bit trims the soil and the soil sample is slowly pushed into the Shelby tube. This drilling technique works well when the soils are fine-grained, but recovery is hampered when the subsurface soil contains medium to coarse gravels.

Soil Erodibility

All potential failure modes related to an earth fissure underlying an embankment dam are associated with seepage along the discontinuity, and erosion of the walls and base of the earth fissure. Therefore, evaluating the erodibility of the foundation soils, both in terms of threshold of erosion and rate of erosion are critical when evaluating the safety and integrity of embankment dams in earth fissure risk zones. This section provides a brief discussion on some tests used to evaluate the erodibility of soils.

Seismic Refraction Studies

EHI (2005) attempted to use excavatability information from typical surface seismic refraction surveys to evaluate the erodibility of alluvial soils. The method combines Kirsten's rippability index approach (1982) with site specific seismic refraction data to evaluate the energy needed to initiate erosion in a soil. The method assumes that in general, a greater resistance to excavation and ripping as estimated from Kirsten's (1982) approach, the greater the resistance of the soil to erosion. The authors believe that the EHI approach has two drawbacks. Firstly, the method fails to account for the fact that many alluvial soils in the Southwest slake and soften when exposed to water. Secondly, as with most empirical methods, Kirsten (1982) developed his excavation relationship based on a conservative interpretation of his data set. This approach is certainly valid when evaluating excavation conditions; however, when using seismic velocity data to assess erodibility, the very assumptions that make Kirsten's approach conservative for excavation and rippability, result in unconservative interpretations of erodibility. Seismic velocities provide some general qualitative information on erodibility of alluvial soils, but additional confirmation and perhaps modification of the EHI relationship is required to correlate seismic velocities to quantitative thresholds and rates of erosion.

Vertical Jet Index Tests

The test procedure consists of a water jet impinging on the in-situ soil to be tested. Periodically, a pointer and gage mechanism mounted within the jet tube is used to measure the depth of erosion. The data collected in the field is reduced to estimate the threshold of erosion and the rate of erosion. The principal advantage of this test is that the test can be performed in the field on the soil in in-situ conditions. However, the test requires direct access for men and equipment to the soil strata to be tested and is usually limited to test pits.

Erosion Function Apparatus Test

The Erosion Function Apparatus (EFA) test was developed at Texas A&M University, College Station, Texas. Water is driven through a rectangular pipe by a pump. An undisturbed soil sample is gradually introduced into the flowing water until the soil protrudes 1 millimeter (mm) into the rectangular pipe. The time needed to erode the 1 mm of soil by the flowing water is recorded (Briaud et al. 2001). The process is repeated for increasing flow velocities. Soil disturbance, macro structure and scale effects can significantly affect the results and their interpretation using this test procedure. However, the authors believe this test procedure offers the most promise in quantifying soil erosion on a fissure.

Hole Erosion Test

The hole erosion test (HET) was developed at the University of New South Wales, Australia, and is documented in Wan and Fell (2004). A 6-millimeter hole is drilled along the longitudinal axis of the undisturbed sample to simulate a concentrated leak. The upstream hydraulic head is varied, and the flow rate is used as an indirect measurement of the diameter of the eroded hole, and the rate of erosion. The test needs to be conducted for a sufficiently long time in order to account for the effects of slaking (if any). The authors were also concerned that the shape of the test hole (circular) tends to mitigate the adverse impact of any soil macro-structure or scale effects that may result in higher erosion rates on a fissure.

Fissure Erosion Modeling

EHI (2003) described a MathCAD model developed to evaluate erosion along an earth fissure. The maximum extent of the breach and the time to breach completion are dependent on the magnitude and spatial distribution of the erosive power of water, and on the erosion characteristics of the soil. The model calculates variation in the erosive power of the water in a fissure as a function of time due to changes in water surface elevation within the reservoir (potential energy of the water), and as a function of space due to changes in the dimensions of fissure in general, and the width-to-depth ratio of the fissure in particular.

To the best of the authors' knowledge, this is the only model available to assess erosion along an earth fissure. The model can be used to evaluate the existing dam configuration, as well potential mitigation measures such as cutoff walls and aprons. This model is a useful tool to study the potential eroded width of an earth fissure at a specific dam site. One drawback of the model is that it incorporates a constant gradient such as used in porous media flow. The model also simplifies the erosion process along the fissure using a velocity-erosion relationship and does not include any erosion related to soil slumping from mechanical failures.

The authors envision that during the initial stages of seepage along an earth fissure, the flow of water is predominantly downward at a relatively high gradient. The soils would be exposed to relatively high erosive forces under this vertical flow. At some point, based on the fissure characteristics and the soil grain-size distribution, the fissure will plug, diverting the high gradient flow laterally along the fissure. This erosion mechanism could lead to rapid lateral migration of a void along a fissure extending below a dam.

Fissure Risk Identification

FCD and their consultant AMEC Earth & Environmental Inc. developed a protocol to qualitatively evaluate fissure risk at a given location. The process includes collecting and reviewing interferograms, low sun angle aerial photographs, selective ground truthing, and seismic refraction surveys to identify zones of low, moderate and high fissure risk. To the best of the authors' knowledge, this is one of the first attempts to systematically evaluate sitespecific conditions, and assign formal, qualitative fissure risk zonation along an embankment dam. This is a very useful tool in rehabilitating dams in areas potentially prone to earth fissure development in that the embankment can be "segmented" and prioritized for rehabilitation. Depending on stakeholders' risk aversion and funding capabilities, zones of high, or moderate and high risk may be immediately rehabilitated, while the remainder of the embankment can be monitored and rehabilitated as required at a future date.

Case Histories

The Flood Control District of Maricopa County (FCD) operates and maintains 22 singlepurpose flood control dams in central Arizona. Portions of two of these dams, White Tanks Flood Retarding Structure (FRS) No. 3 and McMicken Dam, are located in earth fissure risk zones. FCD is currently designing and/or constructing modifications to rehabilitate both dams to mitigate the risk posed by potential earth fissures at the two dam sites. The two dams are under the jurisdiction of the Arizona Department of Water Resources (Department), and both authors were deeply involved in the application review process for both dams. These two case histories illustrate the regulator's dilemma related to embankment dams in earth fissure risk zones and identify the variables that the regulator, owner and design engineer had to consider in developing a safe rehabilitation of two dams located in fissure risk zones.

White Tanks FRS No. 3

Description

White Tanks FRS No. 3 is a single-purpose flood control dam located in western Maricopa County. The dam is a homogeneous embankment designed and built by the Soil Conservation Service and has a length of approximately 7700 feet. The dam is located on the lower reaches of an alluvial fan on the eastern flanks of the White Tank Mountains. The foundation profile consists of deep unconsolidated alluvial deposits. Unconsolidated alluvium, primarily consisting of fine sandy silt, silty sand, with lesser amounts of clayey sands, sandy clays of low plasticity, and relatively clean sands extend to depths from approximately 100 feet on the southwest end of the dam to 500 feet on the northeast end of the dam (ADWR, 1998). Highly erodible Holocene soils extend to a typical depth of 10 feet. The underlying Pliestocene soils are generally lightly cemented with smaller zones of low to high cementation.

Subsidence & Earth Fissure Risk

Since construction of the dam in the mid 1950s, nearly 4 feet of differential subsidence has occurred over the length of the embankment, with nearly 4.5 feet of total subsidence at the north end of the dam, and less than a foot of subsidence at the south end. The subsidence was caused by drop in groundwater levels. The differential settlement of the north end of the dam relative to the south end has been attributed to changes in lithology along the length of the dam. AMEC Earth & Environmental Inc. (AMEC, 2003) performed studies to assess fissure risk as this site. The AMEC (2003) study identified a moderate to high risk of earth fissure development along a 2500-foot section of the embankment. During a risk assessment workshop, experts on ground subsidence and earth fissures opined that there was an 80 percent likelihood that earth fissures would develop at the site during the 100-year life of the project. It is important to note however, that no earth fissures have been identified at the site to date.

Proposed Remedial Design

FCD has chosen to remediate the section of embankment within the moderate to high fissure risk zone, where to date extensive investigations have not identified the presence of any fissures. A typical cross section of the proposed remedial design is shown in Figure 6. The proposed design (URS, 2005) includes excavating the Holocene surface soils to a depth of 8 to 10 feet to expose Pleistocene-age soils, constructing a new soil-cement embankment directly upstream of the existing earthfill embankment, and protecting the new soil-cement embankment with two 30-foot deep cutoff walls, one each at the upstream and downstream toes of the new soil cement embankment. The soil-cement embankment is to be buried under earthfill that will be landscaped to improve the aesthetic quality of the project.

Figure 6: White Tanks FRS No. 3 - Typical Cross Section (from URS, 2005)

The design philosophy is that removal of the Holocene soils removes the upper very erodible and potentially collapsible soil section. Early modeling indicated an upstream blanket of soil cement was less effective than the cutoff walls in reducing the potential erosion along a fissure in the upper Pliestocene soils. The fine-grained Pliestocene soils appear to be somewhat less erodible below a depth of about 30 feet. The two cutoff walls will divert seepage along a fissure below the base of the cutoff walls into the more erosion resistant strata. The walls also increase the flowpath length, reducing the initial flow gradient and associated erosive forces. The combination of increased soil erosion resistance and reduced erosive force is expected to limit the eroded width of the fissure below the cutoff walls, and limit the rate of uncontrolled discharge from the reservoir in the event a fissure forms at the dam site. The soil-cement embankment is a key structural element of the project that is provided to bridge across an eroded fissure and to prevent erosion along any cracking that may occur in the dam in association with a foundation fissure.

The Regulator's Dilemma(s)

Well into the regulatory review and approval process, the Department staff maintained concerns about both the adequacy of the subsurface profile and the characterization of the erodibility of the foundation soils being input to the fissure erosion model. We believed that some of the more erodible layers within the profile were not reflected in the small set of test data. Also we believed additional borings were required to increase the confidence in the subsurface profile being input to the erosion model. The designer agreed that erodible layers existed in the subsurface profile, but indicated they believed these layers were interbedded with areally-extensive erosion-resistant layers that would "throttle" the seepage and limit erosion along the fissure. The Department remained unconvinced, and requested the owner perform additional field investigation and laboratory testing to confirm the subsurface profile and erodibility data input to the model. In-situ field investigative techniques such as developing a borehole erosion test method and drilling large-diameter shafts to allow access for direct visual inspection of the subsurface soils were explored. Ultimately, the owner performed a supplementary field investigation using triple tube coring, and downhole geophysical logging of an agreed upon borehole program.

The Department's first dilemma was that the existing dam is located in a predicted high risk fissure zone. However, an extensive site investigation did not discover any indications of past or present fissures. Given this, could the Department withhold a permit to rehabilitate and significantly increase the safety of the existing dam?

The Department's second dilemma was we were not convinced that the present state of technology for the model or soil characterization truly simulates the actual erosion along a fissure. However, we believe that the proposed design protects against these unknowns to a large degree. We do believe the model allows an assessment of the positive impact of structural elements as well as the negative impact of increasing width of a potential fissure. The modeling performed by the owner's design engineer convinced the Department that the dam design must include a core of soil-cement and that the foundation requires cutoff walls. The modeling also convinced the Department that any fissure must be detected and remediated at very small crack widths (1/4 to 1/2 inches).

Since the model results indicated it was essential to detect very small fissure widths, the third dilemma for the Department related to whether monitoring of the dam for fissure development was feasible. The Department had to be convinced that the state of technology had advanced to the point where very accurate ground strains could be measured as a precursor to fissure development and that small-width fissures could be detected when there were no surface expressions. If this was not possible, we believed it would not be safe to rehabilitate this dam in a fissure risk zone. Through evidence obtained during several site investigations at FCD dams in Arizona and the experience of one of the owners' design engineers, the Department was convinced that there was a high degree of confidence that a well planned monitoring program could detect both the ground strain occurring as a precursor to fissure development and also any small-width fissures which may be developing below the ground surface.

Resolution

The Department believes the proposed dam design and the conditions of approval listed below will result in a safe dam that provides the greater public the required long-term flood protection.

1. At the time of foundation preparation, carefully examine and map the foundation, and perform a seismic refraction survey to confirm the absence of earth fissures at the dam site today.
2. Design and implement a detailed instrumentation and monitoring plan to detect the onset of tensile strains that might serve as a precursor to fissure formation. The monitoring program for ground strains will include high-precision GPS monitoring, installation of a TDR cable, periodic review of interferograms, and low-sun-angle aerial photography. Actual fissure cracks will be searched for using the surface geophysical techniques developed by Rucker (2000), discussed earlier in this paper. Ground truth will include test trenching and geological logging at selected limits of ground strain and/or when fissure signatures are obtained using the geophysical survey lines.
3. As part of the monitoring plan, design and implement a detailed site investigation plan to detect fissures when the ground strain limits are reached. The plan will include surface geophysical techniques developed by Rucker (2000) and test trenches.
4. Should a fissure be detected at the site, design and implement remedial measures that would isolate the fissure from water in the impoundment.
5. Recognizing that the state-of-the-art and -practice in terms of fissure detection, remediation, and general dam engineering will advance with time, review the design every ten years to confirm that the remedial measures implemented today continue to meet current dam safety practices and standards.

McMicken Dam

Description

McMicken Dam is located in Western Maricopa County in Central Arizona. The dam was designed and constructed by the U.S. Army Corps of Engineers in the mid 1950s. The dam is a homogenous earth embankment, approximately 9 miles in length, and a maximum height of approximately 27 feet. The dam is located on the lower reaches of an alluvial fan. The foundation profile consists of deep unconsolidated alluvial deposits consisting primarily of cemented sandy gravels to depths of over 100 feet in the project area on the southwest end of the dam. The alluvial deposits range in total depth from approximately 400 on the southwest end of the dam to over 1600 feet on the northeast end of the dam (Oppenheimer, 1980).

Subsidence & Earth Fissure Risk

Declines in groundwater levels have caused as much as 4 feet of subsidence at McMicken Dam. The maximum subsidence occurs on the northeast end of the dam over the deeper alluvial deposits. However, studies in the early 1980s (AMEC, 2003a) identified earth fissures in the downstream vicinity of the southwest end of the dam. In 1982, the nearest fissure to the dam was approximately 650 feet from the embankment (AMEC, 2003a). Recent studies identified surficial features such as potholes and small depressions within 125 feet of the embankment. Subsurface investigations identified fissure traces both upstream and downstream of the embankment. Results of subsidence modeling (AMEC, 2003b) indicate if groundwater levels remain constant, horizontal strains may increase only slightly with time. If groundwater levels were to drop significantly (on the order of 100 feet), the risk associated with either widening of existing fissures or development of new fissures would increase (AMEC, 2003b).

Proposed Design

FCD has chosen a design that includes elimination of the southern 7800 feet of dam currently located within a high fissure risk zone. The site investigation identified open fissures below the dam and hairline trace fissures extending under the dam into the reservoir area. This section of the dam will be replaced by constructing a new soil-cement embankment extension, oriented roughly perpendicular to the existing dam centerline to retain the flood pool. The new embankment will be outside the high fissure risk zone and will be founded on coarse-grained and/or cemented Pleistocene age soils. The shallow Holocene soils will be removed to eliminate any potential for collapsible soils in the foundation and to remove highly erodible soils. The design includes a 40-foot wide horizontal, upstream soil-cement blanket. The design also includes upstream channels to divert flows that reported to the south end of the old dam, and construction of a retention basin to provide the required flood protection. Figure 7 shows a cross section of the new soil-cement embankment.

Figure 7: McMicken Dam - Typical Cross Section (from AMEC, 2004)

The design philosophy is that the Pliestocene foundation soils at McMicken Dam have moderate to high erosion resistance, and except for very small zones the erosion resistance exceeds the erosive forces. Thus, very little erosion is predicted if seepage were to occur along a potential earth fissure. These soils are generally believed to be coarse grained and will also tend to plug a small fissure. Early modeling indicated an upstream soil cement blanket was more cost effective than cutoff walls to reduce the potential erosion along a foundation fissure. The 40-foot soil cement blanket upstream of the dam increases the flowpath length, reducing the initially available erosive forces. This reduced erosive force is expected to limit the eroded width of the fissure, and thus, limit the rate of uncontrolled discharge from the reservoir. The soil-cement embankment is a key structural element of the project that is provided to structurally bridge an eroded fissure and to prevent erosion along any cracking that may occur in the dam in association with a foundation fissure.

The Regulator's Dilemma(s)

Our concerns at McMicken Dam lay primarily with the characterization of the subsurface soils at the site. For the fissure erosion modeling, the designer assumed that the soils at the site were moderately to strongly cemented, with minimal susceptibility to slaking and weakening when exposed to water, and with moderate to high thresholds of erosion. The initial geotechnical investigations at the site appeared to support this assumption. Late during the design process however, the designer cored the subsurface soils using the triple tube technique discussed earlier in this paper. Core recovery was excellent, and zones of weaker cementation and slaking susceptibility were identified in the core. Furthermore, the Department was also concerned that the designer had performed no erosion testing at this site, and instead relied entirely on geophysical and excavation data to assign numerical values for erosion resistance. To address the more erodible soils within the profile, the Department requested the owner and designer to perform additional modeling using lower thresholds or erosion for the weaker soil strata. Even with these lower thresholds, the modeling demonstrated that seepage erosion along the fissure would be small.

The Department's first dilemma was that the existing dam is located in a known fissure zone. However, the rehabilitated portion of the dam was being relocated to a predicted moderate fissure risk zone, in which the site investigation identified several hairline cracks in the foundation. Given this, could the Department withhold a permit to rehabilitate and significantly increase the safety of the existing dam?

The Department's second dilemma was we were not convinced that the present state of technology for the model or soil characterization truly simulates the actual erosion along a fissure. However, we believe that the proposed design protects against these unknowns to a large degree. We do believe the model allows an assessment of the positive impact of structural elements as well as the negative impact of increasing width of a potential fissure. The modeling performed by the owner's design engineer convinced the Department that the dam design must include a core of soil-cement and that the foundation requires an upstream soil-cement blanket. The modeling also convinced the Department that any fissure must be detected and remediated at very small crack widths (1/4 to 1/2 inches).

Since the model results indicated it was essential to detect very small fissure widths, the third dilemma for the Department related to whether monitoring of the dam for fissure development was feasible. The Department had to be convinced that the state of technology had advanced to the point where very accurate ground strains could be measured as a precursor to fissure development and that small-width fissures could be detected when there were no surface expressions. If this was not possible, we believed it would not be safe to rehabilitate this dam in a fissure risk zone. Through evidence obtained during several site investigations at FCD dams in Arizona and the experience of one of the owners design engineers, the Department was convinced that there was a very high degree of confidence that a well planned monitoring program could detect both the ground strain occurring as a precursor to fissure development and also any small-width fissures which may be developing below the ground surface.

Resolution

The Department believes the foundation soils at McMicken Dam are erosion resistant because of the degree of cementation, and/or the primary presence of large-size particles within the subsurface. The Department believes proposed dam design in conjunction with the proposed monitoring program will result in a safe dam that provides the greater public the required long-term flood protection. As with White Tanks FRS No.3, the owner is required to review safety of the dam on a 10-year cycle to confirm that the remedial measures implemented today continue to meet current dam safety practices and standards.

Discussions & Closure

Earth fissures pose a significant safety risk to several existing flood control embankment dams in the Southwest. As population increases in the Southwest there is a commensurate demand for groundwater. Thus, subsidence and related earth fissures will continue to adversely impact dams and other infrastructure in the future, to an even greater extent than today. Thus, understanding the science of earth fissures will continue to gain greater importance with time.

The dilemma facing the state dam safety regulator today and what complicated the approval of the rehabilitation of the White Tanks FRS No. 3 Dam and McMicken Dam is the paucity of both information and analytical methods related to fissure erosion and the performance of embankment dams on earth fissure risk zones. The current state-of-thepractice related to dams in fissure risk zones is at an infant stage, forcing the dam owner, design engineer and dam safety regulator to exercise, as judged today, conservative approaches in the rehabilitation of a dam in a fissure risk zone. There is so little science and performance data that there is no assurance the rehabilitation approach taken today will prove conservative in the light of advances made in fissure science in the future.

As the regulatory agency charged with dam safety oversight responsibilities in Arizona, the Arizona Department of Water Resources is fully committed to the safe design, construction, and operation of dams in the State. With increasing urban growth in the state, the Department is acutely aware of the need for both safe and cost effective flood protection such as afforded by White Tanks FRS No. 3 Dam and McMicken Dam. While breaching or removing these two flood control dams because they are located in fissure risk zones would eliminate the risk from a dam safety point of view, the public will then be exposed to the risk of flooding. The Department believed the greater public need would be better served by assessing all aspects of the state of the practice with respect to the safety of embankment dams in fissure risk zones. This included working with the owner and the respective design engineering firms to evaluate every available consideration to see if there was a combination of safety measures which appeared to result in a safe rehabilitation of each dam.

The owner and the respective design engineering firms for these two dams achieved a safe rehabilitation design against the threat posed by the fissure risk zones by committing significant monetary and engineering resources to evaluate every aspect of each project. In order to develop each safe rehabilitation design, it was necessary to

1. Perform comprehensive field investigations using state of the practice techniques
2. Perform laboratory erosion testing using state of the practice equipment
3. Develop a new computer model to simulate fissure erosion
4. Incorporate several erosion resistant design elements to increase the safety of the rehabilitation against uncontrolled erosion along any undetected fissure
5. Incorporate state of the practice monitoring techniques to detect the development of a fissure, and
6. Commit to immediate remediation of any fissure that may be detected in the future.

A critical factor that the Department considered in the decision to permit the proposed rehabilitation projects for White Tanks FRS No. 3 and McMicken Dam was that both these dams are single-purpose flood control structures and are generally dry. Thus, conducting long-term monitoring for detection of a fissure prior to the occurrence of a major flood pool is feasible and can be included as a factor that will enhance the safety of the dam against the risk of failure from an undetected fissure. Furthermore, these types of dams also allow rehabilitation to be performed, with little risk of the reservoir filling while the dam is threatened by the presence of a fissure and this factor can also be included as a safety enhancement.

Finally, while all the factors indicate a safe dam rehabilitation, the dam safety regulator must still appreciate the many unknowns and assumptions being made in the present state of the art and knowledge related to the mechanics of fissure development and erosion, especially under the hydraulic head imposed by a full reservoir. Advances will occur in the science and engineering related to building and maintaining safe dams in fissure zones. Thus, an essential part of the present regulatory approval of the McMicken Dam and White Tanks FRS No. 3 Dam rehabilitations was a stipulation for a full reassessment of all technical factors, each 10 years, to confirm the safety of each dam in the fissure risk zones.

The science and state of the practice for an embankment dam in an earth fissure risk zone needs to be advanced by harnessing the available intellectual capital being developed around the world. Together these experts have the capability of advancing fissure science in areas related to the mechanisms of fissure development, understanding of potential failure modes related to embankment dams, requirements for comprehensive field investigations, and comprehensive laboratory and field testing to develop computer model erosion parameters, development of robust computer models to simulate fissure erosion, and the development of engineering designs that result in safe dams.

Acknowledgements

The Flood Control District of Maricopa County operates and maintains White Tanks FRS No. 3 and McMicken Dam. URS Corporation is the designer for White Tanks FRS No. 3 rehabilitation project, and AMEC Earth & Environmental Inc. is the designer for the McMicken Dam Fissure Risk Zone Remediation project. Photographs of fissures were provided by Ralph Weeks, P.G., Ray Harris, P.G. and Ken Euge, P.G. Their contributions are acknowledged and appreciated.

References

1. AMEC, 2003a: Final Zoning of Earth Fissure Risk & Determination of Parts of the Dam that Require Dam Safety Modifications, McMicken Dam Fissure Risk Zone Remediation Project, prepared by AMEC Earth & Environmental Inc. for the Flood Control District of Maricopa County, November 5, 2003.
2. AMEC, 2003b: Earth Fissure Investigation Report, McMicken Dam, prepared by AMEC Earth & Environmental Inc. for the Flood Control District of Maricopa County, April 11, 2003.
3. AMEC, 2004: Design Report, McMicken Dam Fissure Risk Zone Remediation Project, prepared by AMEC Earth & Environmental Inc. for the Flood Control District of Maricopa County, August 24, 2004.
4. URS, 2005: Design Report - White Tanks FRS No. 3 Remediation Project, Phase I; prepared by URS Corporation for the Flood Control District of Maricopa County, March 23, 2005.
5. USGS, 1999: Land Subsidence in the United States, Circular 1182, U.S. Department of the Interior.
6. Bouwer, H, 1977: Land Subsidence and Cracking Due to Ground-Water Depletion; Groundwater, Vol 15, No. 5. September - October, 1977.
7. EHI, 2005: McMicken Dam Multi-Layer Foundation Fissure Modeling, prepared by Engineering & Hydrosystems Inc. for AMEC Earth & Environmental Inc. and Flood Control District of Maricopa County, February 17, 2005.
8. Kirsten, H.A.D., 1982: A Classification for Excavation in Natural Materials; The Civil Engineer in South Africa, pp 292-308, (discussion in Vol 25, No. 5, May 1983).
9. Briaud, J.L, F.C.K. Ting, H.C. Chen, Y. Cao, S.W. Han and K.W. Kwak, 2001: Erosion Function Apparatus for Scour Rate Predictions, Journal of Geotechnical and Geoenvironmental Engineering, ASCE, February 2001.
10. Wan, C.F. & Fell, R., 2004: Investigation of Rate of Erosion of Soils in Embankment Dams, Journal of Geotechnical and Geoenvironmental Engineering, ASCE, April 2004.
11. Larson, M.K. and Pewe, T.L., 1986: Origin of Land Subsidence and Earth Fissuring, Northeast Phoenix, Arizona; Bulletin of the Association of Engineering Geologist, Vol XXIII, No. 2, pp 139 - 165.
12. Oppenheimer, J.M., Sumner, J.S., 1980: Depth-to-Bedrock Map, Basin and Range Province, Arizona; University of Arizona, Department of Geosciences, Laboratory of Geophysics, Scale 1:1,000,000.
13. Arizona Department of Water Resources, (ADWR) 1998: Depth to Water and Well Inventory Data in files and Groundwater Site Inventory Site search, http://www.adwr.state.az.us/adwr.
14. Weeks, R.E. 2005: Personal Communication.

TrackBack URI for this entry

Comments (0)

Subscribe to this comment's feed

Show/Hide comment form