Geotextiles andGeomernbranes 15 (1997) 235 253

© 1998 Elsevier Science Limited

Printed in Great Britain. All rights reserved
0266-1144/97 $17.00
ELSEVIER PII: S0266-1144(97)10006-1

A Performance Summary of Reinforced Soil Structures in the

Greater Los Angeles Area after the Northridge Earthquake

Dean Sandri
TC Mirafi, Lake Forest, CA 92630, USA.

ABSTRACT

This paper contains a summary of observations on segmental retaining walls

and geogrid reinforced soil slopes which were subjected to the shaking of the
Northridge earthquake. Eleven structures were visually examined for evidence
of distress resulting from the earthquake. The structures highlighted provide a
wide range of slope and wall systems using commercially available and widely
marketed products. Specific details of two structures located within 35 km (22
miles) of the epicenter and subjected to strong ground motions are discussed.
General comments on their construction and performance are provided. It is
concluded from the survey that reinforced soil structures perform excellently
even when subjected to seismic loads far in excess qf those for which the), were
designed. © 1998 Elsevier Science Ltd.

THEEARTHQUAKE

At 4:30 am on January 17, 1994, the Greater Los Angeles area was shaken
by a 6.7 Magnitude earthquake. The shaking was centered near Northridge,
California; a Los Angeles suburb, but was felt strongly in four counties
surrounding Los Angeles.
Reports on the Northridge event indicate that the duration of strong
shaking lasted only 10-15 seconds (Shakal et al., 1994) but, was more violent
and destructive than other tremblers lasting much longer and having greater
magnitudes. One of the reasons for the violent shaking and unusually high
destruction was the pattern of motion itself. Accelerations experienced
during the January 17, 1994 earthquake were often as strong vertically as

235
236 D. Sandri

~ ~ & -.-

0
~ ~ ~ 2
d

0 E

-H 44 44 44

.~

~.~

~o ,~
 Reinforced soil structures 237

0 0 0

e~

0
~=~

c~"~

"~.= ~ =
~ ~'- ~

-'~ ~ ~ ~ ~ ~2~N

e-,

o ,~

~.~ o
238 D. Sandri

they were horizontally (Shakal et al., 1994; Peterson et al., 1996). Most
structure analyses and codes dictating structure designs had simply not
anticipated the type of vertical shaking like that which occurred on January
17, 1994.

THE VIBRATIONS

In most earthquakes, measurements of shaking have recorded horizontal

ground accelerations of about 1.5 + times that of the vertical ground accel-
eration. However, the Northridge earthquake produced just the opposite.
The California Division of Mines and Geology indicated that recorded
vertical ground accelerations exceeded 1.5 + times those of the horizontal at
several locations (Shakal et al., 1994). The strength and uniqueness of this
earthquake provided an excellent opportunity to assess the performance of a
number of geosynthetically reinforced soil structures.
Table 1 provides a general information summary for each structure along
with the approximate acceleration experienced by each. Figure 1 provides a
map of the greater Los Angeles Area with approximate location of the
structures surveyed. Figure 2 provides a seismic acceleration contour map of
the greater LA area along with the location of two surveyed structures which
were closest to the epicenter.

THE STRUCTURES

The structures surveyed include all those known to exist in Los Angeles and
Orange Counties which exceed 4.6m (15ft.) in height at the time of the
earthquake. Only geogrid reinforced, segmental block walls and geogrid
reinforced soil slopes with geogrid reinforced, segmental block walls set into
them were surveyed and reported herein.

THE SURVEY

The survey generally consisted of examining the walls/slopes for evidence

of movement. Examination included visual inspection of the structure for
alignment, relative facia movements, facia bulging, facia cracking,
geogrid slippage (at face), soil sloughing, tension cracking of soil/pave-
ment near the slope/wall crest, or other signs which might suggest seis-
mic related movements. Several of the structures inspected are illustrated
in Figs 4-8..
 Reinforced soil structures 239

Fig. 1. Location of reinforced soil structures surveyed.

THE DESIGNS

The design methods utilized in the analysis of segmental walls and slopes
highlighted herein varied considerably. However, all of them were designed
and analyzed using procedures consistent with available technology at the
time of their development.
240 D. Sandri

.../"®_.

,,,,..,,'"

"~'~-- SURFACEPROJECTIOHOF

!
APPROXli,IATEf.AULT RUPI1JRFPLkHE.

- • $ ¢ ~ e • - , " . . " - - . - .
• " - 2" 0 2 ¢" 6 8~ - -"-. ~ ~ : [] ~-0.?-"~
• .J . . . . . " , • . "-.. oi / ~ "--.~_~ax !;~N

I : - : : : -i .......
• " . " . • . " " ¢' " " - • . ~.k ',, ~ : Q -
I
I " • - • . . . . . . . . . ~.~ , a.~o.,.__ \ I

. - " - . . . . " " - " .. .J . . . . . . . . . . . . . . . ~- ........... ( i . . . . - . g - . . "

i ..-... - ' , , • " • ~,-.,~ ! . "'--..:..... ::

. " ". " . "- 0 Soastation . " " " " " " # o " " . " " ' " "

|. Notl; ~mc~u/~ca~onwsspmvidedbylheovm~'~geno/exoe~tc~U~steUonsas~$~m~3.3.1. . - • ]
I SlblS w~e t y p i C a d l y (:Muill~l by sur~:~l g a m l a g y : h o w e v e r , ~ stations were classified b a s e d o n Sxd or. aL ". ~ I
I" " (19~) in which ~(:ial . . s ~ . V. > . mls -- ~m~gmilld as ".k" ~ . . .- ".'..J

Fig. 2. L o c a t i o n o f Valencia and G o u l d tank wall on c o n t o u r m ap o f m a x i m u m horizontal

accelerations (after Shakal et al., 1 9 9 4 ) ,

In general, the walls were analyzed using tied back wedge analysis meth-
ods with Rankine assumptions. Slopes were analyzed using Modified Bishop
Method of Slices circular failure analyses methods described in ceference
documents frequently utilized for design at the time the projects were devel-
oped (Christopher e t a l . , 1989; Simac, 1990). Where employed, seismic
analyses were conducted using pseudo-static procedures with horizontal
accelerations to model seismic loads. Specific details of individual analyses
 Rein]orced soil structures 241

were unique to each project and were somewhat specification and/or regula-
tion controlled.
Long term design strengths (working strengths) of the geosynthetic rein-
forcements were determined using GRI (Geosynthetics Research Institute,
1991), Task Force 27 (AASHTO-AGC-ARTBA, 1990) or other similar
standard of practice. No differentiation was made in 'long term design
strengths' for static vs. seismic analyses. Factors-of-safety applied to long
term design strengths (LTDS), global stability, and other aspects of the
structure analysis also varied with project and were highly affected by speci-
fication/regulation requirements.
Specific details for the Valencia water treatment plant and Gould walls are
discussed below. The two structures discussed herein were chosen based on
their relative close proximity to the epicenter (22 and 35 km {14 & 22 miles}),
exposure to strong shaking, availability of geotechnical, design, and
construction information, past experience of structural compromising activ-
ity, and function as public works projects.

VALENCIA WATER TREATMENT PLANT WALL

The Valencia wall consist of an approximate 8 km long wall constructed less

than 1 year prior to the Northridge event. The wall is located approximately
18 km from the epicenter and only 14 km northwest of the highly publicized
failed overcrossing at the I-5 and Antelope Valley Freeway (SR14) (White
and Holtz, 1994). Shaking magnitude at the Valencia site was estimated by
the USGS to be at an approximated Modified Mercalli Intensity (MMI) of
VII (White and Holtz, 1994). Joyner-Boore attenuation relationships suggest
peak horizontal accelerations of about 0.6 g with actual measured horizontal
accelerations of about 0.5 g as rigorously determined by Bathurst and Cai
(1995) and 0.35 as estimated from Fig. 2 (Stewart et al., 1994).
Figure 9 shows the approximate location of the walls relative to a highway
bridge which was closed as a result of damage incurred during the earth-
quake. A typical cross section of the Valencia water treatment plant wall is
shown in Fig. 3.
Calculation of reinforcement long-term design strength (LTDS) was
conducted in general accordance with GRI GG-4 (Geosynthetics Research
Institute, 1991) with the exception that an additional factor-of-safety for
'uncertainties' was required by the owner. Design procedures employed were
in general accordance with the tied back wedge analysis method.
The structure was constructed in the flood plain of a river channel and is
supported by native silty sand and clay deposits extending far beneath the
wall base (Harding Lawson Associates, 1990). Fill used in the reinforced soil
242 D. Sandri

z~

".
.< ,.4
 Reinforced soil structures 243

Fig. 4. Keystone faced SRW situated midway up 24.4 m (80 ft.) high reinforced soil slope at
Congregate Care Facility in Dana Point.

zone consist of native silty sand deposits. Since the wall would be inundated
during flood events, a relatively comprehensive geotextile wrapped drainage
blanket was incorporated beneath and behind the reinforced soil mass.
Provision for utility installation located behind the wall face and within
the upper 2 m (6 ft.) of structure was accomplished by limiting reinforcement
length in the upper 2 m (6 ft.) of the wall to 3 m (10 ft.) in length. In order to
accommodate this, the analysis was conducted to examine the effects of both
an independent 2 m (6 ft.) upper wall section analyzed as a surcharge and a
full height structure (lower primary wall plus 2 m (6 ft.) upper wall).
Design horizontal accelerations were estimated by the geotechnical engi-
neer to be 0.3 g (Harding Lawson Associates, 1990). The magnitude of lateral
pressures resulting from seismic loading was dictated by the geotechnical
244 D. Sandri

Fig. 5. Loffel faced, 11.6 m (38 ft)maximum height SRWs constructed at crest of unreinforced
slope at the Pelican Hills Housing site near Irvine, CA.

engineer to be 16 times the total wall height and applied as a uniformly

distributed horizontal pressure over the full height of the wall (Harding
Lawson Associates, 1990).
The design of the wall also included analysis for seasonal flooding to
within 0.6 m (2 ft.) of the walls surface with subsequent rapid draw-down.
Wall embedment was dictated by the owner and geotechnical engineer and
was specified as a means of scour protection. As the wall was constructed, fill
was placed on both sides of the wall face until the scour/embedment criteria
were satisfied. Soil placed in the river flood plain was graded at a gentle slope
away from the wall face with saturation anticipated during periods of the
year. Design flood stage flow velocities at the face of the wall were 4.6 m/s
(15 ft/s).
Quality control during construction was provided as required by the
 Reinforced soil structures 245

Fig. 6. 6.1m (20ft.)maximum height, Diamond faced SRW supporting tennis courts at a
residence in Diamond Bar, CA.

contract documents and generally included field density testing of compacted

soils and visual confirmation for proper placement of segmental retaining
wall (SRW) facia units as well as for reinforcement length, strength and
placement elevation. However, after the walls were completed, excavation
for power pole and utility line installation near the southern end of the walls
violated the upper 2.5m (8± ft.) of reinforcement at a distance of about
1.2 m (4 ft.) from the facing units. The area was backfilled after the utilities
were installed without any reinforcement repair being conducted prior to the
earthquake. As a result, the southern end of the wall was in a compromised
structural state at the time the earthquake occurred.

GOULD TANK

The segmental retaining walls at the Gould Tank site were constructed for
the purpose of supporting an access road adjacent to above-grade water
storage tanks. The site is located along the southern margin of the San
Gabriel Mountains near Pasadena, California. The Gould Tank site is situ-
ated about 35km (22 miles) from the epicenter and along the trace of a
potentially active splay of the Sierra Madre Fault.
Joyner-Boore attenuation relationships for the Gould Tank site suggest
246 D. Sandri

peak horizontal accelerations of about 0.3 g with contoured maximum hori-

zontal accelerations of slightly greater being determined by actual measure-
ments (Stewart et al., 1994).
Maximum wall heights at the Gould location were about 5.8 m (19 ft.) and
were designed using tied back wedge analysis methods. LTDS's in the wall design
were calculated in accordance with GRI GG-4. Fill used in wall construction
consist of native soils excavated from the site. The soils generally consist of
clayey silt and silty sand alluvial deposits with foundation materials consisting of
native crystalline rock. Surcharge pressures of 1.2 kN/m 2 (250 lb/ft 2) were incor-
porated into the design to model anticipated traffic loads which were to be
maintained a short distance from the wall crest. A typical detail of the Gould
Tank wall is provided in Fig. 3.
Seismic analyses were not required for the wall portion of the design.
However, the stability of slopes into which the walls were founded were
analyzed by pseudo-static methods using horizontal accelerations of 0.15 g.
Factors-of-safety for the wall portion of the design exceeded 1.5 for external
sliding and internal tensile overstress and pullout. Factors-of-safety for
external stability overturning and bearing capacity exceeded 2.0. Facia
connection was not expressly considered in the analysis. The target factors-
of-safety noted above are in general accordance with those promoted by the
manufacturers of geogrids at the time the designs were prepared. Factors-of-
safety for global stability of the slopes into which the walls were constructed
were maintained at 1.1 (pseudo static) and 1.5 (static) or greater.
Construction of the walls included monitoring by the owners representa-
tives. With the exception of some alignment problems, wall construction
proceeded without significant incident.
Subsequent to completion of the walls, but prior to completion of the
adjacent water storage tanks, construction traffic over-surcharged the walls
causing the upper portions to move outward. The over-surcharging created
an apparent 'reverse batter' in the wall. Reinforcement damage (if any)
caused by the surcharge loading may exist but was not able to be verified
during a follow-up remedial investigation.

THE POST E A R T H Q U A K E INSPECTION RESULTS

None of the structures surveyed were seriously affected by the earthquake

(Sandri, 1994). In fact, none of the geosynthetically reinforced soil structures
in Orange County (farthest from the epicenter) showed any signs of move-
ment/distress. The performance of all walls is considered to be excellent
especially considering the moderately high horizontal and vertical accelera-
tions to which they were exposed. Even the wall areas compromised by
 Reinforced soil structures 247

excavating through the reinforced soil sections (Valencia wall) and over-
surcharged sections (Gould tank wall) performed admirably during the
record shaking. Comments on both follow.

Valencia

Intermittent minor surface tension cracking ( < 6 m m , (1/4in.) in width) was

noted in the surface soils near the back of the reinforced soil mass. In the area
where the geogrid had been excavated, the surface tension cracking was again
noted to exist near the back of the shortened 'reinforced soil mass' (Fig. 11), but
was substantially greater (about 50 mm {2 in.} in width). The gently sloping soil
placed in front the wall for scour protection (river side) showed significant signs
of movement including consolidation, and surface fissures. It is likely that these
soils were saturated to shallow depth at the time of the earthquake. Figure 10
shows some of the settlement and tension cracking in the fill at the wall face.
Inspection of the block and block/geogrid interface indicated a total
absence of relative movement. Close visual inspection of dried mud exposed
on top of and between facia blocks indicated the mud to be in tact and
undisturbed. Similarly, geogrid exposed at the wall face (sandwiched between
block courses) and 'set' in dried mud was absent of any indications of
disturbance or relative movement.

Fig. 7. Keystone laced, triple tiered walls set into 1.5:1 unreinforced slopes in custom home
development in Walnut, CA.
248 D. Sandri

In light of the minimal damage experienced by the SRW, it is interesting to

note that on-site cantilever walls of two clarifier tanks located within 100 m
(330 ft.) of the segmental wall experienced significant cracking causing them
to be taken out of service for repairs.

Gould tank

The Gould Tank site was subjected to predicted horizontal acceleration of

about 2 times those for which it was analyzed.
Surface tension cracking was again noted near the back of the reinforced
soil mass. The surface tension cracking was apparent as an approximate
6ram (1/4in.) wide fissure intermittently located along the wall alignment
(Fig. 11). Inspection of the wall did not reveal any signs of relative move-
ments between facia blocks or geogrid/facia blocks interface. Increase in
the previously existing reverse batter due to earthquake activity was not
visually apparent.

SUMMARY

The reinforced soil structures surveyed provide a variety of critical and non-
critical structures which have been exposed to significant ground shaking of a

Fig. 8. Keystone faced, 6.5 m (21.3 ft)max, height SRW at Valencia Water Treatment Plant in
Valencia, CA. Walls are located in flood plain of Santa Clara River.
 ReinJbrced soil structures 249

Fig. 9. Location of Valencia Water Treatment Plant relative to nearby bridge. Note lateral
displacement along the center line of the highway bridge.

6.7 magnitude seismic event. Analysis methods used in design of each structure
varied, but was conducted in accordance with accepted procedures and/or
specification requirements available at the time the designs were prepared.
The structures located farthest from the epicenter lacked visual evidence of
damage while the two structures closest to the center of shaking exhibited
relatively minor signs of movement. The structures located nearest the
epicenter experienced measured horizontal ground accelerations of up to 2
times those for which they were designed.
Portions of structures which had previously been compromised by unau-
thorized excavation experienced greater movements while areas which had
been previously been over-surcharged performed equally as well as un-
damaged sections.
250 D. Sandri

Fig. 10. Surface soil settlement and tension cracks on low side (in Santa Clara River flood
plain) of Valencia Water Treatment Plant walls. Note black notebook set into fissure adjacent
to wall in foreground.

CONCLUSIONS

The behavior of the walls surveyed above appears to corroborate other

similar post earthquake surveys (Peterson et al., 1996; Tatsuoka et al., 1995a;
Tatsuoka et al., 1995b; White and Holtz, 1994) which conclude that geosyn-
thetically reinforced soil structures will perform well in seismic situations.
The concept of reinforcing a soil mass with extensible reinforcement to make
it act as a coherent unit during a seismic event appears to have been vali-
dated.
The observations presented above and reinforced by the generally excel-
lent behavior of reinforced soil structures reported on after recent seismic
 Reinforced soil structures 251

Fig. 11. Typical surface tension cracking noted near back of reinforced soil mass at Valencia
and Gould sites.

events in the Western US and Japan suggest that geosynthetically reinforced

soil structures provide an excellent choice for grade transitions. Concern
about facia blocks jumping off walls, facia blocks self destructing as they
bang into each other during shaking or reinforcement pulling away from
facia blocks during a seismic event having significant vertical and horizontal
accelerations was not observed or substantiated by observations reported in
the literature.
While the performance of the structures reviewed is considered to be
excellent, improvements in design methods which employ Coulomb
assumptions have been suggested (Bathurst and Cai, 1995; Bathurst and
Simac, 1994). Back-calculation using Coulomb assumptions has shown an
excellent correlation to the observed distress/acceleration scenarios presented
252 D. Sandri

by the Northridge Earthquake thus suggesting that it may prove more

appropriate for analyses of near vertically faced reinforced soil structures.
It is the author's opinion that significant conservatism is built into design
procedures for reinforced soil structures. The combination of design conser-
vatism and the inherent ability of reinforced soil structures to resist seismic
stresses has limited the financial and social disaster resulting from the most
recent California and Japan earthquakes. (Petersen et al., 1996).

REFERENCES

AASHTO-AGC-ARTBA (1990). Design Guidelines for Use of Extensible Reinfor-

cements (Geosynthetic) for Mechanically Stabilized Earth Walls in Permanent
Applications, In-situ Soil Improvement Techniques, Task Force 27 Report,
American Association of State and Highway Transportation Officials,
Washington, DC, USA, August 1990, 38.
Bathurst, R. J. & Cai, Z. (1995). Pseudo-Static Seismic Analysis of Geosynthetic-
Reinforced Segmental Retaining Walls. Geosynthetics International, 2(5), 787-830.
Bathurst, R. J. & Simac, M. R. (1994). Geosynthetic Reinforced Segmental Retain-
ing Wall Structures in North America, Proceedings of the Fifth International
Conference on Geotextiles, Geomembranes and Related Products, Singapore,
September 1994, 24.
Christopher, B. R., Gill, S. A., Giroud, J. P., Juran, I., Schlosser, F., Mirchell, J. K.,
& Dunnicliff, J. (1989). Reinforced Soil Structures: Volume L Design and
Construction Guidelines, Report No. FHWA-RD-89-043, Washington, DC,
USA, November 1989, 287.
Geosynthetics Research Institute (1991). GRI GG 4(b) Standard of Practice for
Determination of the Long-Term Design Strength of Flexible Geogrids, Drexel
University, Philladelphia, PA, January 30, 1991, 16.
Harding Lawson Associates (1990) Geotechnical Engineering Investigation, Valen-
cia Wastewater Reclamation Plant Expansion, Valencia, California, HLA Job
No. 9372025.16, February 1, 1990, 18.
Petersen, M. D., Cramer, C. H., Bryant, W. A., Reichle, M. S., & Toppozada, T. R.
(1996). Preliminary Seismic Hazard Assessment for Los Angeles, Ventura, and
Orange Counties, California, Affected by the 17 January 1994 Northridge
Earthquake, Bulletin of the Seismological Society of America, 86(1B), $247-
$261, February 1996.
Sandri, D. (1994). Retaining Walls Stand Up to the Northridge Earthquake,
Geotechnical Fabrics Report, June/July 1994, 3~31.
Simac, M. R. (1990). Design Methodology For Miragrid Reinforced Soil Retaining
Walls, Mirafi, Inc., Charlotte, NC, USA, 1990, 110.
Shakal, A., Huang, M., Darragh, R., Cao, T., Sherburne, R., Malhotra, P., Cramer,
C., Sydnor, R., Graizer, V., Maldonado, G., Petersen, C. & Wampole, J. (1994).
CSMIP Strong-Motion Records from the Northridge, California Earthquake of
January 17, 1994, California Department of Conservation, Division of Mines
and Geology, Office of Strong Motion Studies, Report No. OSMS 94-07,
February 18, 1994, 307.
 Reinforced soil structures 253

Stewart, J. P., Bray, J. D., Seed, R. B. & Sitar, N. (1994). Preliminary Report on the
Principal Geotechnical Aspects of the January 17, 1994 Northridge Earthquake,
Report No. UCB/EERC-04/08, June 1994.
Tatsuoka, F., Tateyama, M. & Koseki, J. (1995a). Performance of Geogrid-Rein-
forced Soil Retaining Walls During the Great Hanshin-Awaji Earthquake,
January 17, 1995, Proceedings of the First International Conference on Earth-
quake Geotechnical Engineering, IS-Tokyo '95, Tokyo, Japan, November 1995,
55-62.
Tatsuoka, F., Tateyama, M. & Koseki, J. (1995b). Pseudo-Static Seismic Analysis of
Geosynthetic-Reinforced Segmental Retaining Walls. IGS News, 11(2), 12-I 3.
White, D. M. & Holtz, R. D. (1994). Performance of Geosynthetic-Reinforced
Slopes and Walls During the Northridge, California Earthquake of January 17,
1994, Department of Civil Engineering, Univ. Of Washington, Seattle, WA,
1994.