December 13, 1999
W.O. D56865
Mr. Kevin L. Stock via Facsimile (206) 340-0902
Helsell Fetterman
1500 Puget Sound Plaza
1325 Fourth Avenue
Seattle, Washington 98111-3846
Subject: Comments on Mechanically Stabilized Earth (MSE) Wall
Third Runway Seattle-Tacoma International Airport
Dear Mr. Stock:
Per our agreement I have reviewed the documents your staff sent to me as well as other publications related to the general design and performance of mechanically stabilized earth (MSE) walls (list attached). My opinions regarding geotechnical issues that would affect the design and construction of the MSE wall proposed to support the planned third runway at Seattle-Tacoma International Airport are presented below. I also have included comments regarding materials not contained in the documents that I believe are necessary for one to make a rational decision regarding the construction of the proposed third runway.
It is my understanding that the Port Authority is considering constructing a third runway at Sea-Tac International Airport to the west of the existing runways. The new runway will require constructing an embankment 8500 feet long and 150 high along some portions of its alignment. Total quantities of imported fill will be on the order of 20,000,000 cubic yards. A portion of the embankment would encroach on wetlands areas and a stream (Miller Creek), if it were constructed with nominal 2:1 side slopes. In order to minimize the encroachment the Port Authority plans to utilize a four-tiered mechanically stabilized earth wall along a 1450-foot portion of the embankment. The wall would be about 100 feet from Miller Creek at its nearest point. The tallest section of the retaining wall will be about 150 feet high and 500 feet long. A wall of this height and length is a massive undertaking, and in my view will require bold engineering and construction techniques for it to be successfully constructed. It likely will be the highest, or one of the highest, walls of this type constructed in a seismically active area.
In general a mechanically stabilized earth (MSE) wall of some form might be appropriate for this project. However, the specific geotechnical constraints of the site must be known before any earth retaining system can be counted on to assure success. The following are some brief comments regarding design and construction of the proposed wall. My comments include both the positive aspects of the wall system and those aspects of the system, and embankments in general, that should be considered before moving forward with the project.
MSE wall systems have been used throughout the world for about 30 years. It is a simple and innovative concept developed in the early 1970s to retain earth embankments by the use of reinforcing ties and relatively thin facing panels of concrete or steel. In general the system relies on increasing the strength and stability of earth embankments by placing reinforcing straps, mesh, or fabric within the embankment as it is constructed. The reinforcing elements are being attached facing panels where embankments must stand vertically.
Slope Stability
There are some definite limits and constraints to an MSE wall or any other earth retaining system being considered for this site. The reports reviewed discussed the soil conditions at the site in detail and provided engineering parameters needed for analysis of the proposed embankment and the MSE wall. However, the soil properties presented were for the most part assumed strength parameters based on general soil index testing rather than on laboratory tests of the specific soils at the site. Index tests are a way of inferring the behavior of soils from tests that do not actually measure strength characteristics specifically. For example, the number of blows required to drive a standard soil sampler into the ground with a standard weight hammer (Standard Penetration Test SPT) often is used to estimate the density and the strength of soil. In many situations empirical relationships between the "index" test and true soil properties is adequate for design. However, in my judgement the structures needed for this project (embankments and walls) warrant specific and thorough laboratory testing of the soils at the site to determine their true engineering characteristics.
The 1999 Hart Crowser reports (HC) also discussed the methodology they planned to use for the design including their approach to sensitivity analyses. Their analytical approach is appropriate for MSE walls and sloped embankments of moderate height, and their proposed approach to a sensitivity analysis is appropriate for embankments or MSE walls of any height. However, in my opinion the analysis of a wall of the height proposed should include a more rigorous analysis than that proposed (e.g., finite element analysis and full dynamic analysis), and, as stated above, it should be based on measured soil properties, not assumed properties. Furthermore, the summary of their analyses presented as an appendix to the Revise Draft Wetland Functional Assessment and Impact Analysis by PARAMETRIX, INC. indicated some configurations of the embankment and/or assumed soil strengths result in factors of safety (F.S. = 1.2 to 2.0 static, and F.S. = 0.8 to 1.2 seismic) below normally acceptable levels. A factor of safety of 1.5 is typically sought for earth embankments under static conditions, and a factor of safety of 1.1 is usually considered minimal under seismic loading conditions. A factor of safety less than 1.0 indicates instability and failure. No mention was made to the reasons for the low factors of safety presented in the report, nor was there mention of any way to improve the factors of safety of the planned embankments. The analytical details of the slope stability or wall stability analyses were not presented in the reports reviewed, so no real assessment of the stability of the planned development could be made.
Foundation Soils
The proposed wall is acknowledged to be a very tall structure even by MES wall standards. Although it is likely the tie/reinforcing system and facing panels can be designed and constructed to perform adequately, it was not clear in the reports reviewed whether or not the foundation soils on which the wall (and the sloped, non-retained portions of the embankment) would be constructed can support the imposed loads adequately. The vertical pressure below the high sections of the wall/embankment are estimated to be on the order of 18,000 to 20,000 pounds per square foot a significant bearing requirement. The geotechnical reports reviewed indicate the existing near surface soils in the area are composed of fills, bog deposits (peat) and recessional deposits of loose to dense silty sands from the Vashon glacial episode. Ground water is present at shallow depths (5 to 10 feet) where the toe of both the embankment and the wall will be placed. Dense glacial tills underlie those soils at varying depths throughout the site ranging from 15 feet to as much as 30 feet. The dense tills should provide adequate bearing for the proposed embankment or MSE wall; however, the peats would have to be removed below the new runway embankment. Peat is extremely soft and compressible, and very weak. If the peat deposits are not removed, settlement on the order of several feet could take place. However, it is more likely the embankment and wall would fail due to gross bearing capacity failure, horizontal sliding failure or general slope stability failure before it was constructed to a significant height. For a wall and embankment as tall as anticipated to succeed the foundation soils must be competent. Peats are not. Therefore, they will have to be excavated and replaced with competent fill material.
The logs of the soil borings presented in the geotechnical investigations for the project indicate the recessional deposits (sand and silt) likely would not support the wall/embankment in their natural state. In my opinion they would have to be removed also. The PARAMETRIX report also acknowledges this possibility, and indicates some form of in situ (in-place) soil strengthening would be used to improve the bearing and sliding resistance of those soils. "Stone columns" were mentioned as a possible method. The stone column method is an approach to in situ soil improvement whereby "columns" of gravel are installed in the ground in a tight grid pattern throughout an area where the soils have poor engineering characteristics. Stone columns are constructed by jetting and vibrating gravel into and through weak soil deposits. The columns then act as piles to transfer vertical loads to stronger strata below. This method can be an appropriate way in which to improve the strength, settlement, and drainage characteristics of soil deposits in some circumstances. However, since no stability analyses were presented in the documents reviewed, it is not clear whether or not that approach to in situ ground improvement would improve the foundation soils enough at this site to support the very high foundation pressures and to provide adequate sliding resistance to maintain horizontal stability of the planned embankment.
Embankment Fills
In my judgement the HC reports addressed the material requirements for the embankment fill satisfactorily for a feasibility report. The authors addressed the need for select fill materials and adequate compaction in the reinforced zone immediately behind the wall panels and in the upper traffic support zone immediately below the new runway. Those are the critical areas of the proposed embankment. They also addressed the need for select drainage material within the embankment to control ground water seepage and the potential for the build-up of excess pore water pressures.
If some of the peat deposits below the proposed footprint of the embankment are in fact up to 30 feet deep, the edges of the peat excavation would either have to be sloped for stability during excavation and placement of the backfill or they would have to be retained by some temporary structure during that period. The HC reports address both approaches. The foundation soils and some of the proposed embankment fill soils are sensitive to water, especially when disturbed by construction activity. Therefore, control of ground water seepage and surface water runoff into the excavation and into the fill sections would be critical during construction of the embankment. The reports discusses the need for a permanent subsurface drainage system to be incorporated into the sloped and the MSE wall embankments. It is likely some of these drainage components also can be utilized to help control water intrusion during construction. Drainage is an issue that should be thoroughly addressed if the design moves forward.
Borrow Sources
The project is estimated to require approximately 20 million cubic yards of fill. That is an unusually large amount of material to be mined and transported through an urban setting. It certainly is feasible to accomplish such a task; however, if the haul route(s) utilize local streets, their useful life can be expected to decrease significantly. That is, the streets use as the haul routes likely would have to be resurfaced after the haul is completed, or their subgrade may even have to be reconstructed. Repairs will depend on their intended capacity and their condition before the haul begins.
An assessment of Borrow Areas 1, 3, and 4 was reviewed. Although the investigation appears thorough regarding site geology and ground water conditions, no analysis was presented regarding which soil units encountered could be used for the various zones of fill planned for the embankment. Furthermore, no evaluation of the quantities of the specific classes of materials required for the planned embankments would be present in the borrow sources. In my view each of those items should be addressed and verified. The need for additional or even completely different borrow sources could severely impact the construction of the project.
The reports spoke briefly about the seismic environment of the Sea-Tac area. Since I had limited knowledge regarding the seismic environment of the Puget Sound area before starting my review of this project, I researched USGS and other web sites to confirm and broaden my understanding of the seismicity of the area.
Earthquakes primarily are generated along the margins of the dozen or so tectonic plates that form the crust of the earth. These plates move slowly (on the order of inches or fractions of inches per year) relative to one another. This motion is commonly referred to as continental drift. The relative motion along plate boundaries generally is expressed as lateral sliding between the two plates, pulling apart or spreading, and direct collision between two plates resulting in one plate sliding below the other. The area where one of the earths crustal plates dips below another is called a subduction zone. The seismicity of the Pacific northwest region of the continental United States is dominated by the Cascadia subduction zone between the Juan de Fuca tectonic plate and the North American plate. Large earthquakes can be generated along plate boundaries when the strain that builds up between the plates for centuries is released suddenly within a period of only several seconds. It is my understanding recent paleoseismicity studies indicate the potential for great earthquakes (M 8+) along the subduction zone between the offshore Juan de Fuca tectonic plate and the North American plate is significant. As I recall, the average return period for a great event cited in the literature is on the order of 300 to 500 years. To put this in perspective an earthquake with return period of 500 years has about a 10% probability of occurring at least once during a 50-year period.
In addition to these deep focus interplate earthquakes, lesser intraplate crustal faults can generate earthquakes of moderate magnitude. Several shallow crustal faults have been identified in the state of Washington. The fault nearest the airport site (13 miles north at its closest approach) is the Seattle Fault. The USGS has identified this fault as being capable of producing a magnitude 7.1 earthquake every 5000 years on average. It is one of the seismic sources use by the USGS to produce their current series of earthquake hazard map for the continental United States.
The level of ground shaking depicted on the USGS earthquake hazard maps depict various amplitudes of horizontal ground shaking at a site underlain by rock or very dense/hard soil associated with three different exposure levels. The horizontal ground shaking shown on these maps is expressed as a percentage of the acceleration of gravity (e.g., 20%g, 30%g, 40%g, etc). Three separate maps representing different seismic exposures, or levels of risk, have been published. They present amplitudes of peak horizontal ground acceleration that have the probability of being equaled or exceeded (probability of exceedence) of 10% in 50 years (average return period of approximately 500 years), 5% in 50 years (average return period of approximately 1000 years) and 2% in 50 years (average return period of approximately 2500 years). Normally, projects of this nature are designed for the 500-year return period ground motions. This level of exposure forms the basis for seismic design contained in the building code (Uniform Building Code UBC) used in the State of Washington and the other western states. In my opinion that level of risk is appropriate for this project.
The USGS hazard map indicates the peak horizontal ground motion at the airport site rages from 31%g to 33%g. The geotechnical report by CIVILTECH recommends using 27%g for design of the MSE walls and the sloped embankments. Although the difference between these two seismic design parameters (27%g and 33%g) may not appear large, it could seriously affect the design of the planed embankments and walls especially their global stability.
As stated above, I know of no MSE walls with the height and lateral extent under consideration that has experienced earthquake generated strong ground shaking. I reviewed recent literature regarding the performance of MES walls during earthquakes during this assessment. The literature has several case histories of MES walls that experienced strong ground shaking in several parts of the world California (Loma Prieta, Northridge), Japan (Kobe), Taiwan and Europe. All the case histories presented indicate MSE walls survived the earthquake shaking with only minor damage, if any, even though some of them experienced peak ground shaking (up to 80% g) more severe than that proposed for the airport site. Most of the reported damage was limited to minor cracking and spalling of concrete facing panels and/or minor lateral displacements of the walls - on the order of a few inches or less. However, none of the walls discussed were very tall relative to the proposed Sea-Tac wall. Most of the walls that experienced strong ground shaking were 30 feet or less in height. None of the walls were over 60 feet tall.
The authors of some of the case histories indicated little or no amplification of the ground motions appeared to have occurred within the wall systems because of their short natural periods of vibration (less than (0.5 seconds). This assertion seems reasonable considering the heights of the walls examined. However, a rough estimate of the natural period of vibration of the proposed 150-foot high embankment indicates it period would be on the order of 1.0 to 1.5 seconds. Therefore, amplification of the proposed MSE wall/embankment could be a significant factor in its seismic stability. In my judgement a wall as high and as long as the one planned should have more than a simple pseudo-static analysis performed.
Failure of the wall system and retained embankment under static conditions or due to earthquake shaking could occur from any number of critical elements of the total system, including but not limited to:
The technical issues raised above may have straightforward solutions; however, in my view a realistic assessment cannot be made without more information regarding the subsurface conditions at the site. In my view stability issues with the embankments have not been clearly addressed or justified as yet.
You asked me to discuss the consequences of failure of the embankment and wall during construction or after construction. Clearly, the consequences of failure likely would be worse after the project is complete rather than during construction. In a very general sense one can assume a slide mass resulting from failure of the embankment would extent a distance in front of the wall approximately equal to the height of the wall (in this case about 200 feet in front of the wall). Under some conditions, say an extreme seismic event, the slide mass could project a distance twice the height of the wall or more depending on the mode of failure. Also, one can assume the face of a slide scarp at top of the embankment) could be located at a distance behind the original location of the face of the wall ranging from half the height to the total height of the wall. The impacts of the slide debris in front of the wall and the slide scarp behind the wall would depend on what existed at those locations prior to the slide. Certainly if structures existed in runout zone in front of the wall they would be destroyed. If creeks or ponds lie in the runout zone, they would be filled with the slide debris. If the slide mass at the head of the slide-intersected roads or the new runway, they would be destroyed within that zone or severely damaged. They would be out of service until the embankment was reconstructed.
I hope my comments have answered your questions, and given you some insight and appreciation for the technical aspects of the planned embankment. If you have any further questions regarding the contents of this letter or this matter, please call.
Sincerely,
DOWL Engineers
David A Cole, P.E.
Project Manager
List of Publications Reviewed