94 ■ Transportation Research Record 1736

Paper No. 00-1315

Stabilization of Embankment
Slope with Geofoam
Walter S. Jutkofsky, J. Teh Sung, and Dawit Negussey

A case history is presented describing the use of expanded polystyrene MATERIAL PROPERTIES
(EPS) geofoam blocks to treat an unstable roadway embankment slope
involving clayey soils. The selection of the geofoam treatment was based The geofoam specified for the Route 23A project in New York State
upon its ability to be constructed and have the least impact on both the was of 20-kg/m3 (1.25-pcf) density. Individual blocks were specified
environment and adjacent homeowners. The site subsurface conditions, to have dimensions of 0.6 by 1.2 by 2.4 m (2 by 4 by 8 ft), with each
engineering properties of EPS, design analysis, and construction phases block weighing approximately 36 kg (80 lb). This density is between
are reviewed. Potential traffic safety problems associated with differen- nominal values of 18 kg/m3 (1.15 pcf ) for Type VIII and 22 kg/m3
tial icing of roadways caused by the presence of geofoam blocks beneath (1.35 pcf) for Type II EPS materials given in ASTM-C-578. On the
the pavements were minimized by using a thicker subbase layer in the basis of laboratory unconfined compression test results (8), EPS geo-
geofoam-treated area. Data from an instrumentation program consist- foam samples generally behave like an elastoplastic strain-hardening
ing of an inclinometer, extensometers, and thermistors are presented. material. In general, compressive resistance at a given shear rate can
Pavement temperature readings collected from areas with and without be related to density by either of the following comparable equations:
geofoam treatment are compared to investigate potential differential
icing on the roadway. σ d = 7.3 ⴱ D − 47 (1)

σ d = 9.4 ⴱ D − 76 (2)
“Geofoam,” a generic term for rigid cellular polystyrene used in
geotechnical applications, has provided solutions worldwide to where ␴d is compressive resistance in kilopascals and D is density
many difficult subsoil conditions since it was introduced as a light- in kilograms per cubic meter.
weight highway construction material in Norway in 1972. Most The first equation relates compressive resistance at 5 percent strain
applications of geofoam in highway construction are for with density and is based on test data from the Geofoam Research
Center at Syracuse University. The second equation is from the Euro-
1. Embankment stability and settlement control (1–3), pean standard for geofoam (9) and relates density to compressive
2. Retaining wall backfill to reduce lateral earth pressure and resistance at 10 percent strain. Both expressions are based on tests
settlements (4 ), performed at a strain rate of 10 percent per minute. The 5 percent
3. Compressible inclusion for stress reduction on rigid culverts criterion generally results in a compressive resistance that is about
(5), and 10 percent lower than that for the 10 percent strain level. These rela-
4. Subgrade insulation against frost action (6 ). tionships are useful in that compressive resistance for other densities
can be estimated. For the Route 23A project, geofoam blocks sup-
Expanded polystyrene (EPS) and extruded polystyrene (XPS) are plied to the job site were randomly selected to check for density
the most common types of geofoam. EPS, the kind used in the fol- conformance. Compliance with the desired range of compression
lowing case history, is formed into low-density cellular plastic solids resistance of 90 kPa (13 psi) to 124 kPa (18 psi) was confirmed
that have been expanded as lightweight, chemically stable, environ- indirectly using density as an index. To limit long-term creep defor-
mentally safe blocks. The EPS manufacturing process begins by mation of the geofoam blocks, working stress levels due to dead
exposing polystyrene resin beads containing a hydrocarbon blowing load were limited to less than 30 kPa with an additional load of up
agent to steam and pressure. The polymer cell walls soften and the to 10 percent allowed for live load due to traffic (9).
blowing agent expands the beads to form “prepuff.” These pre- In addition to compression resistance, the interface strengths
expanded beads are later poured into large rectangular block molds. between geofoam and geofoam, between geofoam and bedding
Steam is injected into the molds, in which, under heat and pressure, sand, as well as between geofoam and concrete were of interest.
the beads expand further and fuse to form the final product. Density Figure 1 shows typical results for geofoam-to-geofoam interface
of EPS commonly ranges from 11 to 29 kg/m3 (0.7 to 1.8 pcf). Typi- shear strength tests at two normal loads. At the geofoam-to-sand
cally, a block measures 0.6 by 1.2 by 4.8 m (2 by 4 by 16 ft), with interface, the interface shear coefficient develops to be no less than
weight averaging about 72.6 kg (160 lb). AFM Corporation (7) and the tangent of the internal friction angle for the bedding material
Negussey (8) provide further details. because sand grains become embedded within the foam and force
the shearing to occur in the bedding material. The load distribution
concrete slab over the geofoam fill was cast in place. A relatively
W. S. Jutkofsky and J. T. Sung, New York State Department of Transportation,
Bldg. 7, State Campus, 1220 Washington Ave., Albany, NY 12232-0002.
strong adhesion bond and a rough texture develops between poured-
D. Negussey, Department of Civil Engineering, Syracuse University, 220 Hinds in-place concrete and geofoam surfaces, resulting in a much higher
Hall, Syracuse, NY 13244. interface strength than that between geofoam and geofoam.
Jutkofsky et al. Paper No. 00-1315 95

FIGURE 1 Interface shear resistance for geofoam-to-geofoam surfaces at two

stress levels.

CASE HISTORY roadway into the uphill shoulder, indicating that the movement
was large and the slip plane possibly deep seated. Twenty-two hori-
Background zontal drains installed in an attempt to arrest movement by lower-
ing the groundwater table were not successful. An inclinometer
In 1966, New York State reconstructed Route 23A to meet current installed in the middle of the movement area indicated that lateral
safety requirements and future traffic needs. The roadway embank- movement was occurring from 11 to 12.2 m (36 to 40 ft) beneath
ment crossing a small gully was shifted about 3 m (10 ft) toward the roadway surface at an average rate of 25 mm (1 in.) per year.
the Schoharie Creek and the grade raised nearly 1.5 m (5 ft), as This movement zone was also about 3 m (10 ft) below the nearby
shown in Figure 2. Shortly after reconstruction a 91.5-m (300-ft) creek-bottom elevation. The inferred slip plane (Figure 2) was
long section of the roadway embankment began moving toward the based on the surface scar and inclinometer data. Through the ensu-
creek. The main scarp, as shown in Figure 3, extended across the ing years the slow subsidence remained a continuing maintenance

FIGURE 2 Failure profile and inclinometers.

96 Paper No. 00-1315 Transportation Research Record 1736

FIGURE 3 Plan view of limits of movement and treatment.

problem and traffic hazard. Finally in 1994, a permanent solution eral earth pressure. From Norwegian experience (1) of using the fric-
was pursued (10). tion coefficient of 0.7 for interface sliding between geofoam layers,
The general subsurface profile beneath the roadway consisted of the available resistance to sliding between blocks was adequate.
compact gravelly silt, clayey (recent fill material), and overlayered Laboratory one-dimensional creep tests were performed on geo-
clayey silt, silty clay, underlain by clayey silt, gravelly. Soil mois- foam samples of 20 kg/m3 (1.25 pcf) density. Without considering
ture content ranged from 19 to 33 percent. Groundwater was found problems associated with scaling laboratory data to the prototype
from 1.5 to 4.6 m (5 to 15 ft) below the surface. and extrapolation of maximum test duration of 1,000 min to the
project design life of 20 years, the long-term creep deformation
under a constant stress of 30 kPa for the maximum geofoam thick-
Design Analysis ness of 3.3 m (11 ft) was estimated to be 40 mm (1.5 in.).
Slope stability analysis indicated that replacing 2.8 m (9 ft) of soil
Various stabilizing treatments, including a berm, lowering the grade, with geofoam would raise the safety factor to 1.25. An additional
realignment away from the failure area, lightweight aggregate, and 1.8 m (6 ft) of soil would have to be removed to accommodate a
stone columns, were considered but deemed impractical, environ- 0.6-m (2-ft) drainage blanket below the geofoam mass, and a 1.2-m
mentally sensitive, cost-prohibitive, or all three. The weight-reduction (4-ft) thick subbase layer was placed above the geofoam mass. The
and drainage option using geofoam was chosen to stabilize the area. subbase provides heat storage to counteract insulating effects of
On the basis of field observations, laboratory data, and incli- the geofoam on the subbase and pavement section and to minimize
nometer records, stability of the existing embankment slope was differential icing as described below.
analyzed using the simplified Bishop method to back-figure in situ
soil strengths for a safety factor of 1.0, which would represent mar-
ginal stability of the existing embankment. The back-figured soil- Effects of Geofoam Fill on
strength parameters were angle of internal friction (␾) = 20 degrees Pavement Temperatures
and cohesion = 0. A unit weight of 1.53 kg/m3 (120 pcf) was used
for both saturated and unsaturated soils. Geofoam was modeled as Geofoam is an excellent insulator, restricting heat transfer between
a surcharge with a moisture-absorbed unit weight of 0.08 kg/m3 the subbase-and-pavement zone and the subsoil zone. During the
(6 pcf ) and without contribution of shear strength to overall slope warm season, the surface temperature of the pavement structure in
stability. From the laboratory direct-shear tests of geofoam over Q2 the geofoam-treated area tends to be higher than the pavement struc-
sand, which has similar grain shape and gradation as those of Ottawa ture outside the non-geofoam-treated area. This temperature differ-
sand, a friction coefficient of 0.5 was used to obtain a factor of safety ence may lead to early deterioration of the pavement surface layer.
of 1.25 against sliding between the geofoam blocks and the crushed- During the cold season, the roadway surface placed over the geo-
stone drainage layer and within the crushed-stone layer itself under lat- foam area tends to cool and freeze before the roadway surface over
Jutkofsky et al. Paper No. 00-1315 97

TABLE 1 Actual Gradation of Subbase Material work of perforated drain pipe placed just below the geofoam and
(Item 304.53, Type 2) on Project exposed on the embankment-slope face. This stone layer also func-
tioned as a clean, stabilizing working platform and a level surface
for proper placement of the geofoam blocks.
The vertical portion of the stone underdrain filter material was
extended up behind the back side and both ends of the geofoam mass
to the top of the geofoam (Figure 4). The stone filter material pro-
tected the geofoam from damage and filled the void left behind as
the temporary sheet pile wall was extracted, which minimized any
postconstruction settlement along the extracted sheet pile line.

Instrumentation

The instrumentation program consists of piezometers, extensometers,

an inclinometer, and thermistors to monitor the following elements:
the adjoining untreated area. This phenomenon is called differential
icing, which is similar to the well-recognized “bridge deck” condi- 1. Seepage pressure in the proposed drainage layer,
tion. On the basis of experience in Norway (2), a 0.6-m (2-ft) mini- 2. Lateral displacements of the entire geofoam block mass caused
mum thickness of subbase and pavement material was placed over by lateral earth pressures along with possible internal separation of
the geofoam to minimize this phenomenon, but because of the the geofoam blocks from slope movements,
road’s banked and inclined geometry, up to 1.2 m (4 ft) of subbase 3. Subsurface slope movements, and
material had to be placed. 4. Temperature difference in pavement structures between the
To minimize differential icing of roadways, Gustafson (11) rec- geofoam-treated and untreated areas.
ommended using fine-grained gravel material with some fines as the
roadway subbase above the geofoam. Such material has better ther-
mal conductivity and could provide high heat capacity by absorbing Piezometers
and maintaining moisture in the layer. In 1989, when the Colorado
Department of Transportation (3) used geofoam as a lightweight fill Three vibrating-wire piezometers were installed at the bottom of the
to stabilize a slope, a 0.8-m (2.5-ft.) thick layer of subbase was used vertical drainage gallery, located behind the geofoam mass, to detect
to minimize potential differential icing. For the Route 23A project, possible seepage pressure buildup in the drainage layer if, over time,
a subbase material with a high content of material passing the the outlet area became plugged with vegetative growth.
6-mm (0.25-in.) sieve (25 to 60 percent) was chosen to provide a
high heat-sink capacity and minimize potential problems of differ-
ential icing. For the actual gradation of the subbase material used, Extensometers
see Table 1.
Four extensometers were installed between geofoam layers (Figure 4)
to detect and measure possible sliding between blocks or sliding at
Drainage the interface between the geofoam mass and the drainage blanket.
An extensometer is basically a platform-settlement gauge laid on its
An L-shaped subsurface drainage system, as shown in Figure 4, was side, consisting of a 6-mm (0.25-in.) inside diameter (ID) iron pipe
designed to lower the groundwater table and provide an adequate inserted through a 25-mm (1-in.) ID polyvinyl chloride (PVC) pipe
factor of safety against sliding. The horizontal portion consists of a and bolted to the center of a 19-mm (0.75-in.) by 0.6-m2 (2-ft2) piece
0.6-m (2-ft) thick layer of graded crushed stone with an internal net- of pressure-treated plywood (10). Line-of-sight readings were taken

FIGURE 4 Typical section of treatment.

98 Paper No. 00-1315 Transportation Research Record 1736

on a mark cut on the protruding iron pipe to measure the lateral 0.3-m (6-in.) perforated pipe, was placed and the stone roughly
movement. graded with a bulldozer. Several workers using a construction level,
stringline, and hand rakes then fine-graded the stone to within the
required tolerance of 5 mm/3 m (0.5 in./10 ft).
Inclinometer

Since the old inclinometer had been sheared off, a new inclinometer Phase 4: Geofoam Quality Control Check
was installed in the center of the failure area before commencement
of any construction activities to monitor any slope movement that To verify that delivered blocks conformed to density requirements
might occur before, during, or after construction. stated in the specifications (10), quality was checked. One randomly
chosen block per trailer load or one block per 76.5 m3 (100 yd3) was
weighed by suspending it by straps from an electronic digital scale
Thermistors hanging from a front-end loader. Density varied within the allowable
limits of 1.8± kg (4± lb).
As shown in Figures 3 and 4, 15 temperature-sensing devices were
installed at various depths and locations in the subbase material,
with 12 inside and 3 outside the geofoam area, to monitor and com- Phase 5: Placing Geofoam Blocks
pare temperature variation of the subbase and pavement over time.
In 1998, four additional thermistors were installed 25 mm (1 in.) Geofoam blocks, delivered on flatbed trailers, were easily unloaded
below the pavement surface (one at each group location) to monitor by two laborers and carried and placed by four other labors. On aver-
pavement surface temperature. age, it took 1 h to unload and place a trailer load of 40 blocks, each
measuring 0.6 by 1.2 by 2.4 m (2 by 4 by 8 ft). The blocks were strong
enough for workers to walk and work directly upon each unprotected
Construction Sequence layer surface, with no damage but minor scuffing. To promote inter-
lock within each layer and avoid continuous vertical joints, blocks
Construction followed a sequential order of operations, discussed were placed in a staggered pattern (Figure 4), with each layer oriented
here in terms of successive phases. perpendicular (90 degrees) to those in the underlying layer. The
blocks were held in position during placement by two metal gripper
plates placed on each block, one in the center and one near an edge,
Phase 1: Installing Temporary Sheet Pile Wall approximating a 1.2-m (4-ft) grid pattern. These galvanized steel
plates are 3-mm (0.13-in.) thick and 102-mm (4-in.) square with
Once the inclinometer was installed in the middle of the failure area 13-mm (0.5-in.) high barbs protruding from both sides (9). Also, as
(Figure 2), installation of the sheeting began. Shortly thereafter, the blocks were laid, the four extensometers were installed between
inclinometer readings identified a zone of movement between 6.1 the geofoam layers (Figure 4).
and 9.2 m (20 and 30 ft) below the surface as shown in Figure 2. To accommodate a curve in the segmented-arc design, some
Some movement had been anticipated during vibratory driving. At blocks were field-cut to various wedge shapes with a hot-wire appa-
the end of driving, about 25 mm (1 in.) of lateral shift had developed ratus consisting of a thin piano wire (the middle-C wire worked best)
within the identified zone. Simultaneously, vibrations from driving stretched between two insulated-handle grips connected to a 110-v
were monitored at the nearby house to ensure that ground velocities AC power supply (small portable generator). Current flow was reg-
were within the tolerable threshold of 50 mm/s (2 in./s), to prevent ulated by a hand-held switch and rheostat. With one worker at each
damage. end, the taut hot wire was pressed into the block. As the wire melted
its way along a premarked line, it left a smooth, straight-cut surface.
Hot-wire cutting leaves no dust or debris such as that which occurs
Phase 2: Excavation when hand or power saws are used.

As soil was removed from in front of the sheeting, the sheeting line
was observed to bow into the excavation up to 0.6 m (2 ft). In voids Phase 6: Backfilling
that developed behind the sheeting, standing water was visible.
Numerous 50-mm (2-in.) holes were then cut in the sheeting just After three geofoam courses were placed, backfilling began. While
above the bottom of the excavation to relieve the water pressure crushed stone was placed between the sheeting and geofoam blocks
against the sheeting. Water flowed full volume through several holes for the vertical section of the L-shaped drainage system, earth fill
during construction, and a steady flow came from the drainage blan- was placed over the stepped face of the block mass (Figure 4).
ket on the slope surface. Following drawdown of water and placing
of the 0.6-m (2-ft) thick drainage blanket, all sheeting movement
stopped. Phase 7: Concrete Capping

For load distribution and protection against petroleum spills, a 100-mm

Phase 3: Installing Drainage System (4-in.) reinforced-concrete slab (Figure 4) was placed over the geo-
foam blocks. Forming for the pour was achieved by setting geofoam
The first half of the drainage system, which consisted of a 1.2-m blocks around the area to be poured, each forming block being held in
(2-ft.) thick layer of fine crushed stone with an internal network of place by three or four gripper plates. Wire mesh was set and concrete
Jutkofsky et al. Paper No. 00-1315 99

then pumped over it and rough-screeded as a final finish. Undamaged of the slip plane was below the nearby creek-bottom elevation. As
forming blocks were reused in the geofoam mass, as designed. shown in Figure 5, approximately 110 mm (4.3 in.) of the slope move-
ment was recorded by the inclinometer during construction, mostly
occurring during vibratory installation and extraction of the sheeting.
Phase 8: Subbase Placement No postconstruction movement has occurred since completion of the
project in 1996.
After concrete curing and backfilling up to the top of the concrete
cap were completed, the graded crushed-stone subbase was placed
and the thermistors were installed (Figure 4). Following subbase Extensometers
placement, the sheeting was extracted with a vibratory hammer.
During this time period, approximately 50 mm (2 in.) of additional Of the four extensometers installed in the geofoam blocks, the one
subsurface movement was recorded by the slope inclinometer and placed on top of the first geofoam course, with the plywood of the
extensometers. extensometer directly contacting the vertical drainage layer, regis-
tered the maximum lateral displacement of 61 mm (2.4 in.), as shown
in Figure 6, the majority of which occurred during extraction of the
Phase 9: Paving sheeting. The extensometer placed at 3 m (10 ft) inward from the outer
(south) edge of the lowest geofoam course registered about 30 mm
Placing the standard 230-mm (9-in.) thick asphalt pavement (Figure 4) (1.2 in.) of lateral displacement. Readings from the other two exten-
in April 1996 completed this final phase. someters placed at a higher elevation are not included because of data
In all, 2819 m3 (3,685 yd3) of soil was replaced with geofoam, scattering.
resulting in net reduction of driving weight by about 5351 Mg It is difficult to quantitatively attribute the extensometer dis-
(5,900 tons). placements to different mechanisms that had taken place during
construction. However, it can be assumed that the displacement of
an extensometer is due to a combination of the following: (a) rel-
Field Observations ative movement of the geofoam blocks, placed downslope from
the plywood of the extensometer, resulting from closing up of the
The following field observations are based on results of instrumen-
gap between geofoam blocks during sheeting extraction; (b) lat-
tation monitoring for purposes previously described.
eral compression of the geofoam blocks under the lateral earth
pressure; (c) lateral sliding of the geofoam blocks along the hori-
zontal drainage layer under the lateral earth pressure; and (d) block
Piezometers
rotation triggered by the overall slope movement during sheeting
To date, the three piezometers have indicated no water pressure in extraction. It is conceivable that the geofoam blocks located closer
the drainage material. to the vertical drainage layer would undergo a greater amount of
lateral compression since they would be subjected to greater lat-
eral earth pressure than blocks located at the same elevation but
Inclinometers outward (south) toward the slope face. Extensometer movement
contributed by Item d may be roughly estimated on the basis of the
Readings from the newly installed inclinometer show that the slip amount of slope movement from the inclinometer during the sheet
plane at the inclinometer location was located at a depth of 9.5 m extraction and the location of the slip plane. However, because of
(30 ft) below the ground surface, confirming that the lower portion length limits, this analysis is not presented.

FIGURE 5 Slope movement versus time.

100 Paper No. 00-1315 Transportation Research Record 1736

FIGURE 6 Lateral displacements of geofoam blocks.

Thermistor Readings and Differential Icing est cooling, whereas pavement surface temperatures in the rest of
the geofoam-treated area were less than 1°C higher or lower than
Based on available thermistor readings, the maximum temperature those outside the geofoam area. This results in a maximum time lag
differences between the geofoam-treated and untreated areas are of 0.4 to 0.9 h in differential freezing as shown in Figures 8 and 9.
presented in Table 2. They indicate that the presence of geofoam The area with the fastest cooling described above has an average
affects heat gain in the pavement structure more than heat loss subbase thickness of 0.75 m (2.5 ft), providing less heat capacity
(i.e., the maximum temperature difference in the pavement surface than other geofoam-treated areas that have a subbase layer ranging
was greater in the warmer season than in the cold season) and that from 0.82 to 1.2 m (2.7 to 4 ft) in thickness. The roadway surface
the maximum temperature difference in the subbase layer was has periodically been inspected for possible differential icing by
greater than that in the pavement layer. Inside the geofoam-treated New York State Department of Transportation regional resident
area, ground temperatures near the bottom of the subbase layer engineers during the fall and winter seasons. No differential icing
remain nearly constant with time, not responding much to change has been found nor have complaints associated with this phenome-
in air temperature. However, these temperatures from thermistors non been filed by any local residents. Additional temperature data
placed at different locations along the roadway varied up to 2.1°C are being collected to investigate the effect of the subbase layer
and 4.5°C in winter and summer, respectively. This difference thickness on both the temperature profile in the pavement structure
mainly depends upon the thickness of the subbase layer and the dis- and the long-term pavement performance.
tance from the pavement surface to where the thermistor is placed
in the subbase layer.
A comparison of typical pavement surface temperatures near CONCLUSIONS
freezing along the roadway is shown in Figure 7. Surface pavement
temperatures at 7:00 and 6:00 p.m. on March 10 and 11, 1999, Use of geofoam to reduce the driving force of a slope proved effec-
respectively, were 1.5°C to 3°C above the freezing point and subse- tive in stabilizing the slope at the New York State Route 23A site.
quently cooled down at about the same rate to below 0°C in 2 h. Fig- No slope movement has occurred since the treatment was com-
ure 7 also illustrates that the thermistor placed at a distance of 58 m pleted in 1996. The issue of differential icing was addressed, and
(190 ft) from the west end of the geofoam-treated area had the great- no such a phenomenon has occurred on the roadway. The treated
embankment area continues to be monitored for slope movement
and temperature variations of the pavement structure at various
locations.
TABLE 2 Maximum Temperature Differences Between Geofoam-
Treated Area and Untreated Area

ACKNOWLEDGMENTS

John Iori and Richard Tedisco of the Geotechnical Engineering

Bureau (GEB) of the New York State Department of Transporta-
tion (NYSDOT) are gratefully acknowledged for their assistance in
FIGURE 7 Pavement surface temperatures before and after freezing in
geofoam-treated and untreated areas.

FIGURE 8 Variation of pavement surface temperature just above and below

freezing point on March 11, 1999.

FIGURE 9 Variation of pavement surface temperature just above and below

freezing point on March 10, 1999.
102 Paper No. 00-1315 Transportation Research Record 1736

developing and monitoring the instrumentation program. Assis- 6. Pavement Design for Seasonal Frost Conditions. Army TM 5-818-2,
tance in graph preparation by Gary Robinson of GEB is also greatly Air Force AFM 88-6. U.S. Departments of the Army and the Air Force,
1985.
appreciated. This work was supported by the NYSDOT.
7. Perform Guard EPS Geofoam: Lightweight Fill and Ground Stabiliza-
tion. Publication GEO 0496. AFM Corporation, Excelsior, Minn., 1996.
8. Negussey, D. Properties and Applications of Geofoam: Geofoam Light-
REFERENCES Weight Fill Application in Syracuse, NY. Society of the Plastics Industry,
Washington, D.C., 1997.
1. Aaboe, R. 13 Years of Experience with Expanded Polystyrene as a 9. Draft European Standard. European Committee for Standardization,
Lightweight Fill Material in Road Embankments. Publication No. 61. Brussels, 1998.
10. Jutkofsky, W. S. Geofoam Stabilization of an Embankment Slope, A
Norwegian Road Research Laboratory, Oslo, 1987, pp. 21–27.
Case Study of Route 23A in the Town of Jewett, Greene County. New
2. Expanded Polystyrene Used in Road Embankments: Design, Construc-
York State Department of Transportation, 1998.
tion and Quality Assurance. Form 482E. Norwegian Road Research
11. Gustafson, K. Investigations at Test Field Linkoping, Over the Period
Laboratory, Oslo, 1992.
1976–1980 (L. J. Grubber, translator). V.I. Report NR 216A. National
3. Yeh, S. T., and J. B. Gilmore. Application of EPS for Slide Correction, Road and Traffic Research Institute, Linkoping, Sweden, 1981.
Stability and Performance of Slopes and Embankments. Geotechnical
Special Publication No. 31. ASCE, New York, 1992, pp. 1444 –1456.
4. Negussey, D., and M. C. Sun. Reducing Lateral Pressure by Geofoam The opinions and findings expressed in this paper are those of the authors and
(EPS) Substitution. In Proc., International Symposium on EPS Con- not necessarily those of the New York State Department of Transportation and
struction Method, Tokyo, Japan, 1996. Syracuse University.
5. Load Reduction on Rigid Culverts Beneath High Fills: Long Term
Behavior. Norwegian Road Research Laboratory, Oslo, 1993. Publication of this paper sponsored by Committee on Soils and Rock Instrumentation.