FEDERAL EMERGENCY MANAGEMENT AGENCY 
FEMA 226 1 February 1992 
Collocation Impacts on the 
Vulnerability of Lifelines During
Earthquakes with Applications to
the Cajon Pass, California 
Issued in Furtherance of the Decade 
for Natural Disaster Reduction 
Earthquake Hazard Reduction Series 61 
Collocation Impacts on the 
Vulnerability of Lifelines During 
Earthquakes with Applications to the 
Cajon Pass, California 
Submitted to the Federal Emergency Management Agency 
Washington, D.C. 
Submitted by: 
INTECH Inc. 
11316 Rouen Dr. 
Potomac, MD 20854-3126 
Co-Principal Investigators: 
Dr. Phillip A. Lowe, INTEC, Inc. 
Dr. Charles F. Scheffey, INTEC, Inc. 
Mr. Po Lam, Earth Mechanics, Inc. 
COLLOCATION IMPACTS ON THE VULNERABILITY OF LIFELINES DURING
EARTHQUAKES WITH APPLICATION TO THE CAJON PASS, CALIFORNIA
TABLE OF CONTENTS 
Volume 1
ITEM PAGE 
Table of Contents.i
List of Figure. ii
List of Tables. v
1.0 
CONCLUSIONS 1
2.0 
INTRODUCTION 4
2.1 
Background 4
2.2 
Study Approach .
5
2.3 
Chapter 2.0 Bibliography 6
3.0 
SUMMARY 6
4.0 
ANALYSIS METHOD 8
4.1 
Data Acquisition 11
Lifeline and Geologic Information 11
Ground Shaking Intensity 13
Selection of the Earthquake Event 16
4.2 
Calculation of Lifeline vulnerability 16
Damage Assessment 16
Shaking Damage 18
Fault Displacement. 23
Soil Movement 24
Landslide 24
30
Liquefaction or Lateral Spread 
Highway and Railway Bridges 31
Times, to Restore the Lifeline to its Needed Service 37
Repair Time 37
Access Time 38
Equipment and Material Time 39
4.3 
Collocation Analysis 39
Impacts on Damage State. 42
Impacts on Probability of Damage 42
Impacts on Time to Restore Lifeline Service .43
4.4 
Interpretation of the Results. 44
4.5 
Chapter 4.0 Bibliography 45
5.0 
APPLICATION OF THE METHOD TO CAJON PASS 47
5.1 
Data Acquisition 47
5.2 
Lifeline Collocation Vulnerability Analysis Results 55
Communication Lifelines 61
Electric Power Lifelines 6.8
Fuel Pipeline Lifelines 75
Transportation Lifelines 81
5.3 
Chapter 5.0 Bibliography. 91
6.0 
FUTURE STUDY NEEDS 91
Appendix A Modified Mercalli Intensity (MMI) Index Scale
i 
LIST OF FIGURES
FIGURE PAGE
Figure 1 Flow Chart of the Analysis Method 9
Figure 2 The "Attenuation Parameter k" for Use in Calculating
Earthquake Shaking Intensity 15
Figure 3 Map of the U.S. Seismic Hazard Regions 22
Figure 4 Liquefaction Induced Damage 32
Figure 5 Map Showing the General Location of the Cajon
Pass Study Area 48
Figure 6 A Composite of the Lifeline Routes at Cajon Pass .50
Figure 7 Identification of Collocations At Cajon Pass 52
Figure 8 Lifeline Routes With Shaking Intensity and
Potential Landslide and Liquefaction Areas 54
Figure 9 Communication Lifeline Routes. 62
Figure 10 Fiber Optic Conduits On A Bridge On The Abandoned
Portion of Cajon Blvd. Extension 63
Figure 11 Details of the Fiber Optic Conduits of Figure 10 .63
Figure 12 Fiber Optic Conduits On A Concrete Culvert
That Passes Under I-15 64
Figure 13 Wall Support Details for the Fiber Optic Conduits
Shown in Figure 12 64
Figure 14 Surface Water Conditions Near The Toe of The I-15S
Retaining Crib Wall 65
Figure 15 Fiber Optic Conduits Under a Heavy Water Pipe, Both
on a Highway Bridge 65
Figure 16 Electric Power Lifeline Routes 69
Figure 17 A Landslide Scar With Power Towers in the Slide Area 
70 
71
Figure 18 Power Lines, Natural Gas and Petroleum Products Pipelines
Intersecting Over the San Andreas Fault Rift Zone 
76
Figure 20 Fuel Pipeline Lifeline Routes 
Figure 21 Typical Long Open Span on the 36-Inch Natural
Gas Pipeline 75
Figure 22 Two Natural Gas Pipelines Located Next to Highway 138 77
Figure 23 Natural Gas Pipeline Crossing Under Railroad Beds 78
Figure 24 Transportation Lifeline Routes 83
Figure 25 I-15 Bridge Over Railroads in Cajon Wash 84
Figure 26 I-15 Bridge Over Cajon Wash 84
Figure 27 Highway 138 Bridge Over I-15 85
Figure 29 Railroad Bridge Over Highway 138 (Collapse Expected) 86
Figure 30 Santa Fe Railroad Bubble Masonry Pier Bridge .8. 87
Figure 31 Union Pacific Railroad Bridge With Power Lines
Overhead.8. 7
Figure 32 Typical I-15 Box Bridge Over the Railroads .90Figure 33 I-15 and Access Road Bridges Over the Southern 
Pacific Railroad 90.90
Figure 19 Power Tower at a Ravine Edge in the San Andreas Fault 
72
Rift Zone 
Figure 28 Railroad Bridges in Cajon Wash (I-15 Bridge in
Background) 886. 
TECHNICAL ADVISORY GROUP 6
Table 2. ATC-13 DEFINITION OF LIFELINE DAMAGE STATE 17
Table 3 
SHAKING DAMAGE PROBABILITY MATRICES,
ATC-13 Tables 7.10 19
Table 4 
MMI ADJUSTMENTS FOR SHAKING DAMAGE EVALUATIONS TO ACCOUNT
FOR LOCAL CONDITIONS 23
Table 5 
LIFELINE DAMAGE STATE FOR FAULT SURFACE DISPLACEMENTS,
ATC-13 Table 8.9 24
Table 6 
SOIL PARAMETERS FOR CALCULATING LANDSLIDE POTENTIAL 26
Table 7 
LANDSLIDE SLOPE FAILURE PROBABILITY MATRICES,
ATC-13 Table 8.7 27
Table 8 
CONVERSION OF LANDSLIDE SLOPE FAILURE STATE TO DAMAGE
STATE, ATC-13 Table 8.8 29
Table 9 
RELATIONSHIP BETWEEN LIQUEFACTION SEVERITY INDEX (LSI)
AND DAMAGE STATE 30
Table 10 PROBABILITY OF LIQUEFACTION GROUND FAILURE, PERCENT 31
Table 11 RELATIONSHIP OF BRIDGE VULNERABILITY INDEX TO
BRIDGE DAMAGE STATE 33
Table 12 BRIDGE VULNERABILITY INDEX FOR EARTHQUAKE SHAKING
DAMAGE 35
Table 13 ESTIMATED LIFELINE REPAIR TIMES TO 100% OPERATING
CAPACITY, ATC-13 Table 9.11 38
Table 14 LIFELINE ZONES OF PHYSICAL INFLUENCE 41
Table 15 MATRIX OF LIFELINE COLLOCATION AND INTERACTIONS .56
LIST OF TABLES
TABLE PAGE
Table 1. CAJON PASS IMPACTS OF LIFELINE PROXIMITY: EXPERT
iv 

COLLOCATION IMPACTS ON THE VULNERABILITY OF LIFELINES DURING EARTHQUAKES WITH APPLICATION TO THE CAJON PASS, CALIFORNIA 
1.0 CONCLUSIONS 
The purpose of this study was to: develop a management screening tool that can be used by lifeline owners, designers and providers, operators, users, and regulators to sort through numerous collocation conditions to identify the critical locations and to provide an estimate of the increased risk that results when such collocated facilities are subjected to an earthquake event; and to analyze the Cajon Pass, California, situation to demonstrate how the screening tool can be used and to examine specific conditions at the Pass.
The resulting screening tool is an important development for several reasons: 1) it is the first documented method for examining multiple collocation conditions and it is applicable to all lifeline facilities. As improvements are made in the fundamental analysis methods for individual lifelines or earthquake conditions, they readily can be introduced into the screening tool to improve its predictive ability; 2) its use can identify the most critical collocation conditions at a specific study area, thereby allowing limited resources to be focused on the most important conditions for improving the overall ability of the lifelines to survive an earthquake event; 3) its use can identify technical areas of uncertainty and/or poor siting practices. This can identify the need for and lead to further research and studies to reduce the identified technical uncertainty or it can identify ways to mitigate siting practices that are more vulnerable to inducing collocation failure conditions; and 4) by being documented and made widely available by the Federal Emergency Management Agency, it is anticipated that it will stimulate the earthquake and lifeline communities to developed improvements in the analysis method or even to develop new, improved screening methods.
The development of the analysis methodology as well as, its test application to the Cajon Pass has highlighted several important conclusions.
* Lifeline collocation can produce both benefits and increased risk of failure during earthquake events. A benefit of closely located transportation lifelines is that the second lifeline can provide the detour or access route to the damaged sections of the first lifeline. However, intersecting lifelines generally result in the failure of one lifeline increasing the risk of failure of the lifeline(s) it crosses.
* It is understandable that topographic conditions have led to the routing of lifeline systems in corridors. However, manmade considerations that force the lifeline owners to use the same rights-of-way for widely different lifelines (for example, locating petroleum fuel pipeline and communication conduits next to each other, routing natural gas pipelines back and forth under a railroad bed, and having a mix of lifelines cross the earthquake fault zone at the same location) greatly increase the risk of failure for the individual lifelines and the complications that will be encountered during site restoration after an earthquake.
* As compared to buildings, ground movement is more important that ground shaking for lifeline components, especially buried lifelines and electrical transmission towers. This means that much of the technical data base on earthquake shaking intensity is not critical for lifeline analysis, whereas important ground movement data and analyses are not as well developed as the shaking intensity data. This suggests that future studies need to emphasize obtaining ground movement information.
* A very useful screening tool has been developed during this study. The tool can be used to identify the critical lifeline collocation locations and the conditions that make them critical. It can identify areas of technical uncertainty and poor siting practices, and its use can identify important research and development activities that can lead to lowered risk of collocation-induced lifeline failures. It will be of value to lifeline owners, designers and providers, operators, users, and regulators.
* The analysis tool has been successfully applied to the Cajon Pass, California. It has identified that for this semi-desert region that: The Cajon Junction, Lone Pine Canyon (which contains the San Andreas fault zone), Blue Cut, and the area just south of the interchange between I-15 and I-215 are the critical locations in terms of collocation impacts at the Cajon Pass.
Fuel pipeline failures have the greatest impact on the other lifelines during the immediate recovery period 2 after an earthquake.
Current siting practices for fiber optic cables indicates that more severe telephone communication failures than have been experienced in past earthquakes can be anticipated in future earthquakes when fiber optic systems have become more dominant in providing the basic telephone service.
Lifeline siting practices have not fully considered the impacts that a new lifeline will have on existing lifelines and, conversely, the impacts that the existing lifelines will have on the new lifeline.
Transportation lifeline restoration of service is highly dependent on sequentially repairing the lifeline damage as the lifeline itself is needed to provide access to the next damage location. Parallel repair operations are more probable for the other lifeline systems and fuel pipeline Communication, electric power, lifelines can generally be analyzed as a set of discrete collocation points. The restoration of service at any one point is not a strong function of the restoration work that is needed at other collocation points. Thus, if there is a restoration problem that will take a long time compared to the other locations, it becomes the "critical path" that sets the time period for the restoration of the entire lifeline system.
When multiple lifelines of the same class are collocated (such as installing all fiber optic cables or all fuel pipelines in the same or parallel trenches) or when multiple different lifelines intersect at a common point, the reliability of each individual lifeline decreases to the value of the "weakest link" of the combined lifeline systems. In addition, repair times increase because of local congestion and the concern that work on one lifeline component could lead to damage of the other different lifeline components.
There is a need for further collocation lifeline studies: to o apply the newly developed screening tool to other locations to assure that the methods can be transferred to other U.S.
locations and to analyze different lifelines, geographic, and earthquake conditions; and to develop data and approaches that can be used to further improve the predictive capabilities of the screening tool.

2.0 INTRODUCTION 
2.1 Background 
Lifelines (e.g., systems and facilities that deliver energy and fuel and systems and facilities that provide key services such as water and sewage, transportation, and communications are defined as lifelines) are presently being sited in "utility or transportation corridors" to reduce their right-of-way environmental, aesthetic, and cost impacts on the communities that rely upon them. The individual lifelines are usually designed, constructed, and modified throughout their service life. This results in different standards and siting criteria being applied to segments of the same lifeline, and also to different standards or siting criteria being applied to the separate lifelines systems within a single corridor.
Presently, the siting review usually does not consider the impact of proximity or collocation of the lifelines on their individual risk or vulnerability to natural or manmade hazards or disasters.
This is either because the other lifelines have not yet been installed or because such a consideration has not been identified as being an important factor for such an evaluation.
There have been cases when some lifeline collocations have increased the levels of damage experienced during an accident or an earthquake. For example, water line ruptures during earthquakes have led to washouts which have caused foundation damage to nearby facilities. In southern California a railroad accident (transportation lifeline) led to the subsequent failure of collocated fuel pipeline, and the resulting fire caused considerable property damage and loss of life. Loss of electric power has restricted, and sometimes failed, the ability to provide water and sewer services or emergency fire fighting capabilities.
In response to these types of situations, the Federal Emergency Management Agency (FEMA) is examining the use of such corridors, and FEMA initiated this study to examine the impact of siting multiple lifeline systems in confined and at-risk areas.
The overall FEMA project goals are to develop managerial tools that can be used to increase the understanding of the lifeline systems' vulnerabilities and to help identify potential mitigation approaches that could be used to reduce those vulnerabilities.
Another program goal is to identify methods to enhance the transfer of the resulting information to lifeline system providers, designers, builders, managers, operators, users, and regulators.
This report is the second of a series of three reports. The first report(presented an inventory of the major lifeline systems located at Cajon Pass, California, and it summarized the earthquake and geologic analysis tools available to identify and define the * The numbers in superscript are references found at the end of each chapter. 4 level of seismic risk to those lifelines. This report presents the analytic methods developed to define the collocation impacts and the resulting analyses of the seismic and geologic environmental loads on the collocated lifelines in the Cajon Pass. The assumed earthquake event is similar to the 8.3 magnitude, San Andreas fault, Ft. Tejon earthquake of 1857. In this, report a new analysis method is developed and applied to identify the increase in the vulnerability of the individual lifeline systems due to their proximity to other lifelines in the Cajon Pass. A third reports presents an executive summary of the study. The Cajon Pass Lifeline Inventory report and this present report taken together provide a specific example of how the new analysis method can be applied to a real lifeline corridor situation.
2.2 Study Approach 
The approach used to develop the information for this report was as follows. The Cajon Lifeline Inventory report('), additional information provided during direct meetings with the lifeline owners, site reconnaissance surveys to validate the information and to examine specific site conditions of interest to the study, and existing literature that describes lessons learned from actual earthquake events were compiled and thoroughly studied. The principal investigators then hypothesized an analysis method that could be applied to the Cajon Pass lifelines to estimate the impacts of proximity on their earthquake-induced performance and repairs.
This analysis method emphasizes building upon existing data bases and analytic methods. In applications, it is recommended that the analyses, studies, and information available from the lifeline owners be used whenever possible. In the event that sufficient data on the lifeline response to earthquakes and the expected time to restore the lifeline back to its, required service level are not available from the lifeline owners, the analytic methods, with some important modifications, of "Earthquake Damage Evaluation Data for California, ATC-13(3),are recommended as an appropriate alternative analysis method. In this project the "most probable restoration time" was defined as the analysis parameter that best could be used to define the impact of lifeline proximity on the individual lifeline's earthquake vulnerability.
The resulting method was then applied to the Cajon Pass lifelines.
The U.S. Geologic Survey's digitized topographic map of the Cajon Pass and the contiguous quadrangles were utilized. The commercial, computer aided, design program AutoCAD was used as it is readily available to the public, thus the methodology is not limited to being dependent upon a specialized or proprietary computer program.
With this tool, overlays of the lifeline routes with seismic and geologic information presented in the inventory reports were used to identify the conditions and locations where the individual lifelines were most vulnerable to the hypothesized earthquake. The 5 


analysis methods described in Section 4.0 of this report were then
applied to the lifelines and the results are presented in Section
5.0. Section 6.0 identifies future studies that could be
undertaken to further qualify the analysis methods and to improve
the details of the specific analysis activities. Section 3.0
provides a summary of the study.
As part of the study validation process, the draft results of the
study were submitted to the project advisors, see Table 1, for
their independent professional evaluation and to the lifeline
owners and regulators who provided information for the preparation
of the report or the Cajon Pass Inventory report. FEMA also sent
draft report copies to a select list of independent reviewers.
Each comment received was addressed, and this final report then was
prepared and submitted to FEMA.
Table 1
CAJON PASS IMPACTS OF LIFELINE PROXIMITY:
EXPERT TECHNICAL ADVISORY GROUP
William S. Bivins T.D. O'Rourke
James H. Gates Dennis K. Ostrom
Le Val Lund Kenneth F. Sullivan
John D. (Jack) McNorgan
2.3 Chapter 2.0 Bibliography
1. P. Lowe, C. Scheffey, and P. Lam, "Inventory of Lifelines in
the Cajon Pass, California", ITI FEMA CP 120190, August 1991.
2. P. Lowe, C. Scheffey, and P. Lam, "Collocation Impacts on
Lifeline Earthquake Vulnerability at the Cajon Pass, California, 
Executive Summary", ITI FEMA CP 050191-ES, August
1991. 
3. C. Rojahn and R. Sharpe, "Earthquake Damage Evaluation Data
for California", ATC-13, 1985.
3.0 SUMMARY
This report presents a systematic approach to calculate the impacts
due to the collocation or close proximity of one lifeline to
another during earthquake conditions. Specifically, the
collocation vulnerability impact is defined as the increase in the
most probable time to restore the lifeline to its intended level of
service. The analysis methods proposed are intended to be used in
screening analyses that determine which lifelines or lifeline
segments are most impacted by the collocation or close proximity of
other lifelines. Once the critical locations or conditions are
known, it may be equally important to reanalyze them using more
detailed analyses to further define the collocation impacts.
The methods proposed are to use the best available information to
determine the lifeline damage state, the probability that the
damage state or greater will occur, and the time to restore the 
lifeline to its intended service. Normally, such information is
obtained from the lifeline owner/operator. However, a alternative
method is proposed when such information is not available from that
source.
The alternative method is based on building upon existing
earthquake damage information and analysis methods which have been
compiled by the Applied Technology Council (ATC). In that manner,
the analysis results can be compared with earlier or future studies
that use the data base without the need to compare or justify the
data base. However, important improvements to the existing ATC
data base also are presented.
Collocation impacts, can be described in one of two broad terms: 1)
the resource impacts (i.e., the increase in personnel, equipment, 
and material resources) that are required to return the total
lifeline system to its needed operating capacity. This is
performed in the present method by summing the impacts at each
component along the entire lifeline route. 2) the resource impacts
at a specific location where multiple lifeline components are
located. In both cases, the present method uses the most probable
time to restore the lifeline component or system to its needed
operating capacity as the appropriate measure of the resource 
impacts.
The analysis method has been applied to the lifeline systems in the
Cajon Pass, California, as a test case. It is clear that the
communication, electric power, and fuel transmissions lifeline
systems that have the potential for collocation impacts are, in
general, not very sensitive to earthquake ground shaking for
shaking intensities represented by Modified Mercalli Intensity
indices of VIII or less (these are the values found at Cajon Pass
for the assumed earthquake event). They are, however, very 
sensitive to ground movement expressed as fault displacement, 
landslides, or lateral spreads. Bridges are sensitive to both
ground shaking and ground conditions (displacement, landslide,
lateral spread, and local liquefactions at their foundation
locations).
It is understandable that topographic conditions have led to the 
routing of lifeline systems into corridors. However, manmade 
considerations that force the lifeline owners to use the exact same
rights-of-way for widely different needs (for example, locating
petroleum fuel pipeline and communication conduits next to each
other, routing natural gas pipelines back and forth under a 
railroad bed, and having a mix of lifelines cross the earthquake
fault zone at the same location) greatly increases the individual
lifeline risks and the complications that will be encountered
during site restoration after an earthquake. 

The Cajon Pass example has identified that the communication,
electric power transmission, and fuel pipeline lifelines generally
can be analyzed as a set of discrete collocation points. The
restoration of service at any one point is not a strong function of
the restoration work that is needed at other collocation points.
Thus, if there is a restoration problem that will take a long time
compared to the other locations, it becomes the "critical path"
that sets the time period for the restoration of the entire
lifeline system. Transportation lifeline collocation points,
however, are sensitive to the damage that has occurred along the
route of the transportation system. That is, often it is necessary
for the heavy equipment and material needed to have access to the
damage location by traveling along the highway or railroad itself.
Thus, before access to a particular bridge can be made, it may be
necessary to first repair all the damage sites on the route prior
to that location.
4.0 ANALYSIS METHOD
In performing an analysis of the impacts of collocation or close
proximity on lifeline systems and components for earthquake or
other at-risk conditions, it is important that the most accurate
data and analyses be used to characterize the response of the
individual lifelines to the loads applied. Whatever method is
applied must be applicable to all the components within the
lifeline system, because the evaluation of the collocation impacts
requires comparing the calculated time to restore the lifeline to
its intended service for both the collocation and an assumed non-
collocation condition. The general methods for performing such an
analysis are shown in the flow chart of Figure 1. If owner-
supplied or site specific analysis methods are not available for
use in the detailed calculations, the following material (Sections
4.1, 4.2, 4.3, and 4.4) can be used as the alternative analysis
method. This is discussed more fully in the following material.
Figure 1 shows a four step approach that can be used to analyze any
lifeline under at-risk conditions (e.g., an natural or manmade
disaster condition). However, the present study only develops the
detailed information needed to analyze earthquake conditions. The
steps are:
1) Data Acquisition;
2) Calculation of Lifeline Vulnerability;
3) Collocation Analysis; and
4) Interpretation of the Collocation Impacts.
Briefly, these activities include:
Data Acquisition
This task is to assemble all of the information that defines the
lifelines and their routes as well as the geologic and seismic
Figure 1, FLOW CHART OF THE ANALYSIS METHOD 
FLOW CHART OF ACTIVITIES FOR 
CALCULATING COLOCATION-INDUCED LIFELINE 
VULNERABILITIES DURING EARTHQUAKES 
STEP 1 
DATA ACQUISITION 
a Lifelines and routes 
* Geologic and seismic conditions 
* Colocation points 
* Lifeline analysis segments 
STEP 2 
CALCULATION OF LIFELINE VULNERABILITIES 
Assuming no colocation 
& Damage state 
4 Probability of damage 
a Restoration time 
STEP 3 
COLOCATION ANALYSIS 
* Lifeline zones of Influence 
* Damage scenario 
* Recalculate new 
* Damage state 
* Probability of damage 
* Restoration time 
STEP 4 
impact OF COLOCATION 
Incremental Change In restoration time 
I Colocation damage probability 
RESULT 
Most probable Incremental 
change in restoration time 
conditions that will place loads on the lifelines. Some analysis
and organization of the resulting information is included in this
step to facilitate the application of the analysis method to the
specific conditions of interest. Such analyses include identifying
the collocation sites as well as dividing the lifelines into
consistent sections for subsequent analysis.
Calculation of Lifeline Vulnerability
The geologic conditions identified during the data acquisition are
used as input to a seismic analysis. Such data include the
topology of the area being studied, a description of the sediment
and rock structures, locations of water, and identification of
surface ground slopes. Seismic conditions include identifying the
location and type of the anticipated earthquake. These are used to
estimate the earthquake shaking intensities (it is recommended that
Modified Mercalli Intensity (MMI) indices be used to characterize
the shaking intensity) and earthquake-induced landslides and soil
liquefaction locations.
During this analysis step, the earthquake intensities and ground
movements are use to determine the vulnerability of each lifeline
at each collocation site as if it were the only lifeline at that
site (e.g., as if there were no collocation there). Based on the
design and placement of the lifeline component or segment and the
seismic loads placed on it, the resulting damage state, probability
that the damage state will occur, and the time required to restore
the lifeline to its intended service can be calculated. The
restoration time is the sum of the time to repair the lifeline
assuming all the equipment, material, and repair personnel are
available at the damage location, plus the access time required to
transport them to the damage location, plus the time required to
have them available to transport to the site.
If owner-supplied damage information is not available, it is
recommended that the analysis methods, as modified in this report,
of "Earthquake Damage Evaluation Data for California", ATC-13,
1985, (prepared by the Applied Technology Council of Redwood City,
California) be used. When a study is to be performed for locations
outside of California, professional judgment must be applied to
determine how to adjust, if at all, the data base of ATC-13. The
methods of "Seismic Vulnerability of Lifelines in the Conterminous
United State, ATC-25, (presently in print at the Applied Technology
Council, and identified as reference 20 in this report section) can
be considered for use. However, it is noted that the consistency
and validity of the ATC-25 approach has not been examined during
the present study, and thus the methods of that study can not be
recommended by the Principal Investigators of the present study.
It is identified here for information only.
Collocation Analysis
This analysis step builds upon the results obtained from the
previous two analysis steps. Based on the actual anticipated
damage states for each lifeline at the collocation site as
determined in the previous analysis steps a collocation interaction
scenario is postulated. The scenario can change either the damage
state, the probability that the damage will occur, the restoration
time (typically only the access time would be changed and the
repair time then would be a new calculation), or any combination of
those items. After the individual items are specified, the
remaining items (i.e., the non specified damage state, probability,
or repair time) are determined using the calculation method applied
in the previous analysis step.
Interpretation of the Collocation Impact
This analysis step uses the calculated information of the two
previous steps to characterize the impact of lifeline collocation.
The most realistic measure of the impact is the "most probable
incremental change in the restoration of service time"'. This is
defined as the product of the probability of collocation damage
occurring times the incremental increase in restoration of service
time (the incremental change in the time to restore service is the 
restoration time for collocation minus the restoration time with no
collocation considered).
Additional details on the recommended analysis, approach are
provided in Sections 4.1, 4.2, 4.3, and 4.4 below.
4.1 Data Acquisition
Lifeline and Geologic Information
Data acquisition is the first step of any lifeline vulnerability
analysis. Information is needed to define the lifelines and their
routes as well as to define the geologic and seismic conditions
that apply to the lifelines of interest.
Information on the lifelines can be obtained from a number of
sources. It is recommended that a site reconnaissance visit be 
conducted first to help the researchers understand the physical
conditions and to preliminarily define the lifelines of interest.
In addition, maps from the U.S. Geologic Survey (such as
topographic maps, published at the quadrangle scale of 1:24,000),
state departments of natural resources or mines and geology, the
U.S. Forest Service, and highway maps are excellent sources of
data. They often indicate lifeline components and routes as well
as identify geographic features. The U.S. and state geologic
surveys (or departments of mines and geologies etc.) will also
have maps and studies that characterize the earthquake faults,
ground units (e.g., the types of sediments and rock formations in
the areas of interest), landslide locations, water table data,
etc. State offices of emergency response (such as offices of
emergency preparedness or seismic safety offices), fire marshal
offices, state public utility commissions, water boards and
commissions, and the general professional literature on earthquakes
are other important sources of information on lifelines and the
potential geologic/seismic conditions of interest.
The single most important source for lifeline information is the
owner/operators. They will each have detailed route maps and
details on their design, construction, and installation. However,
as built drawings and construction information are frequently
different than the "design" information. Thus, it is important to
discuss the information received with the suppliers, and to
validate the understanding received with data from other sources
and site reconnaissance visits.
Once the applicable lifeline data has been assembled, the lifeline
collocation or close proximity locations in the study region should
be identified and given a reference number. Also, each lifeline
should be divided into convenient segments that are reasonably
uniform in their characteristics. These activities are done to aid
in the subsequent analysis steps. The application of the analysis
algorithms (to be described below) can be separately applied to
each lifeline collocation location, using the list of collocation
location points as a check that all the needed locations were
considered, and using the lifeline segments to identify the
physical conditions at the collocation point being analyzed.
The lifeline segments or divisions selected for analysis should be
reasonably "uniform" in that the lifeline components should be
similar within the segment, the shaking intensity (as measured by
the Modified Mercalli Intensity (MMI)) index should be similar, the
ground conditions should be similar (that is, areas of ground
movement should be analyzed separately from areas of stable
ground), and access for repair crews, equipment, and material to
the lifeline proximity points along the segment should be
reasonably the same. With this approach, lifelines, such as buried
pipelines or electrical transmission lines, can be divided into
long segments. Their division is primarily set by the ground
conditions and the MMI values. Other lifeline systems that have
frequent component changes in them, such as transportation systems
that include bridges separated by roadbeds, need to be separated by
component and access route, and sometimes the roadbed must be
further divided to account for ground condition or MMI changes.
Whenever possible, standard measures of earthquake events should be
used to characterize the seismic conditions in the study area. In
this way the results of the study more readily can be compared with
other published data, which allows the conclusions to be validated
by such other available information. Thus, earthquake magnitude or
the earthquake "size" can be represented by the Richter scale.
Ground Shaking Intensity
Several methods to characterize the intensity of the shaking of an
earthquake were considered. Items considered included the
magnitude and extent of the shaking. Although ground acceleration,
velocity, and displacement are more appropriate for evaluating
specific lifeline designs, the use of intensity scales are more
dominant in the literature. Rossi-Forell (RF) and Modified
Mercalli Intensity (MMI) scales are commonly used as a measure of 
intensity. MMI is recommended for use since it is more widely used
in the earthquake literature, although it is a subjective scale
that is dependent on individual interpretation of its meaning. 
Appendix A presents the detailed definitions of MMI.
The MMI scale includes 12 categories of ground motion intensity
from level I (not felt) to level XII (total damage). The use of
Roman numerals was done to discourage analysts from trying to
consider half scale values. This further implies that the MMI is a
broad measure of the shaking intensity. The individual MMI scales
are almost exclusively characterized in terms of building damage,
so their usefulness for modern lifeline structures and components
is somewhat restricted. ATC-l3 (2, provides a detailed estimate of 
lifeline damage probability as a function of the MMI scale. As an
example of potential interpretation problems, the MMI scale IX
identifies that "underground pipes are sometimes broken" while ATC13 
for MMI = IX estimates in California that pipe breaks will occur 
with a total probability of 91.3%. This illustrates the subjective
nature of the MMI scale. Nevertheless, it is commonly used to
characterize earthquake intensity, *and for consistency it
recommended as the proper characterization parameter for examining
the collocation impacts on lifeline vulnerability to earthquakes.
Although there are two computer models, (3,4 ,5) that calculate
earthquake intensity and that are applicable to the conterminous
U.S., the Evernden ) model is recommended because it has been 
verified by comparison with historical earthquakes, it incorporates
the local sediment conditions and such sediment conditions are
generally available in the national U.S. Geological Survey geologic
data base and in the data bases of the various state offices of
mines and geology or natural resources, it is easy to use, it is
readily available to researchers, lifeline owners, and to others
who may need to apply the methods of this study to other regions in
the U.S., and it facilitates comparisons of this research with that
of others7) who have used the Evernden Model. The Advisors to this
Project were concerned that the Evernden model may not be as
accurate near the earthquake fault location (it appears to
underestimate the EMI values there) as it is in predicting the far
field effects. Discussions with the staff of the California
Division of Mines and Geology confirmed that they had similar
concerns. The recommended solution is to increase the calculated
MMI value by one scale level at locations near the earthquake fault
zone. For most lifeline components this is expected to have a 
small impact, because the fault displacement effects there are
expected to overshadow the shaking effects represented by the
increased MMI value.
The Evernden model has been coded in a computer program, QUAK2NW3.
Appropriate input data files are available with the model, and they
were verified for use in the present study. They include:
(1) a fault data file that identifies the location of the 
geologic fault by a series of uniform point sources. 
They can be spaced as closely as desirable. 
(2) a ground condition file that identifies the soil 
conditions (soil and ground geologic units or 
descriptions). The spacing of these ground units 
provides the calculation grid for the program. Evernden 
typically organizes the ground condition into 0.5 minute 
latitude by 0.5 minute longitude grids, and they were 
used for the present study. 
(3) a pseudodepth term "C" which is chosen to give the proper 
near-field die-off of the shaking intensities. Evernden 
has previously analyzed earthquakes along the San Andreas 
fault, and his value of C (10 kilometers) was used in 
this study. Values for other faults can be selected with 
consultation with Dr. Evernden or by professional 
judgment. 
(4) an attenuation parameter "k" which controls the rate of 
die-off of peak acceleration as a function of distance 
from the fault being analyzed. Evernden has identified a 
value for coastal California, eastern California and the 
Mountain States, the Gulf and Atlantic Coastal plains, 
and the rest of the eastern U.S., and these values are 
shown in Figure 2. The coastal California value was used 
in this study, k = 1.75. 
With the above input, computes the acceleration associated
with the energy release along each position of the fault.
Earthquake intensities are calculated in terms of the Rossi Forell
scale, and then the MMI value is computed from a correlation that
Evernden developed for that purpose. The intensity is first
computed for a reference ground unit condition (e.g., saturated
alluvium), and then the intensity value at each grid point is
adjusted for the actual ground condition specified in the ground
condition file. The output of QUAK2NW3 can be used as input into a
digital plotting program, so that the regions of uniform MMI index
can be automatically plotted over the routes of each lifeline
system studied. In the present study the commercially available
program "AutoCAD" was used, although other similar programs would
be just as appropriate. An important criteria for the selection of
the plotting program is that it should be able to read the
Figure 2, THE "ATTENUATION PARAMETER k" FOR USE IN CALCULATING 
EARTHQUAKE SHAKING INTENSITY 
APPROXIMATE PATTERN OF ATTENUATION CHARACTERISTICS (k-VALUE DISTRIBUTION)
THROUGHOUT THE CONTERMINOUS UNITED STATES 
ditigitized files of the U.S. Geological Survey topographic maps.
Those maps include the routes of many of the lifelines and the
geographic elevation contours. Later when ground slopes are needed
to calculate landslide and liquefaction potential, the computer
program can be used to automatically perform the calculations.
Thus, a single program can conveniently incorporate and graphically
present all the key data: lifeline location, fault traces, MMI
values, and ground slopes.
Selection of the Earthquake Event
The next step is to identify the earthquake event for the analysis.
Based on the faults in or near the study region, the QUAK2NW3
program can be used to perform a sensitivity evaluation to identify
the appropriate earthquake event. All that is required is to input
various earthquake events (length and location of the fault
movement, the ground conditions, the depth of the earthquake, and
the attenuation parameter). The results of several analyses can
then be compared to identify the most realistic event for the
analysis. Key additional data that should be considered is the
prediction of the magnitude and the probability that an earthquake
will occur near or in the study region. Such predictions are
available from Federal and state seismologic offices.
4.2 Calculation of Lifeline Vulnerability-
Again, it is recommended that the lifeline owners/operators be
consulted to determine if they already have detailed calculations
on their lifeline's vulnerability to earthquake events. If so,
that approach may be the most detailed available. As an
alternative, the following sections identify how the ATC-13
information, with important modifications, should specifically be
used if such owner/operator information is not available.
Damage Assessment
To determine the potential damage state that occurs, the impacts of
shaking, fault displacement, and soil movement due to either
landslide or liquefaction conditions have to be considered. The
total damage state is the sum of these individual components;
however, if one of these components dominates the others it can be
used without adding the other damage states (this is often the
actual situation). However, when that is done a similar approach
must be used for both the analysis performed while assuming no
collocation impacts and for the analysis performed while assuming
collocation impacts. Also, adding the separate damage states may
over estimate the total damage state. Knowledge of the physical
situation and professional judgement must be applied to determine
the realistic total damage state.
There are seven categories of damage state defined in ATC-13. They
are shown in Table 2.
Table 2
ATC-13 DEFINITION OF LIFELINE DAMAGE STATE
Lifeline For Non Pipeline For Pipeline
Damage State Lifelines Lifelines
No. Description % Damage Breaks/kilometer % Damage
1 None 0 0 
2 Slight 0.5 0.25 0.6 
3 Light 5 0.75 2 
4 Moderate 20 5.5 14 
5 Heavy 45 15 3.8 
6 Major 80 30 75 
7 Destroyed 100 40 100 
In the present method, the important parameter is the
identification of the Damage State Number, a number from 1 to 7. 
Thus, percent damage or breaks per kilometer are not the needed
variable. The experts that developed ATC-13 used the following
definitions for damage state: percent damage meant the estimate of
the dollar value of the earthquake damage divided by the dollar
cost to replace the entire lifeline. However, for pipelines they
were asked to think in terms of breaks in a pipeline per kilometer 
of pipeline length. Within a kilometer segment, 15 breaks may 
actually cost the same as 40, since the expected procedure would be
to simply replace the entire kilometer length rather than to make
such a large number of individual repairs and still be concerned 
that an additional partial break was undiscovered and thus remained
unrepaired. Similarly, an electrical transmission tower with 45%
physical damage would probably be replaced entirely, as it would
not be worth the risk to the owner to make such extensive repairs
when a new tower may be less expensive to install and certainly
would be more reliable in the future. Thus, when the ATC-13
definition is applied to a large number of similar lifeline
components, then, on the average, the damage state may properly
predict the condition of the sum of the individual repair costs
divided by the total replacement costs for all the components.
However, in the present analysis method, the ATC-13 data will be
applied to individual lifeline components. It is acceptable to use
the data in this manner as it provides an expert knowledge base for
estimating the damage state, and the final result of interest in 
the present analysis method is not the damage state but a time to 
restore lifeline service. Its use for single lifeline components
would be less accurate if the desired result were the percent
damage to be used to calculate a cost of repair (that is, ATC-13 is 
more accurate for costs averaged over a large number of cases than
it would be for a single case). The proposed analysis method
could, however, be improved if anew expert opinion study of the
damage state and probability for that damage state for single
lifeline components were to become available.
The following material indicates how the data of ATC-13 are
proposed for use in evaluating the collocation impacts of lifelines
during earthquake events.
Shaking Damage
The shaking impact of the earthquake event can be estimated by
using Table 7.10 (pages 198-217) of ATC-13. For convenience, the
more frequently needed tables for lifeline analysis are reproduced
in this report as Table 3.
These tables present the collective judgement of the probability
that a class of lifeline components will incur a given damage state
level, as a function of the Modified Mercalli Intensity (MMI)
index. They were developed by using a modified Delphi method that
employed a large number of experts who provided their opinion as to
what was the probability that a damage level would be experienced
for a given imposed value of shaking intensity, MMI.
The trend in the probability data would normally be expected to
show that, as the MMI increases, more of the lifeline components
would be expected to experience higher damage states. Thus, for
increasing values of MMI, the shape of the probability curve should
be expected to have its peak value move towards higher damage
states and the magnitude of the peak value decrease as the width of
the probability curve increases. However, at MMI = XII the 
probability curve should again focus over the narrow band of damage
states 6 and 7. The information for bridges, highways, and buried
pipelines and conduits follow this pattern. It is less evident for
electrical transmission towers and railroads. The methodology for
calculating shaking damage collocation impacts, because it is based
on the ATC-13 data, will be less accurate for electrical
transmission towers and railroads, compared to buried pipeline and
conduits, highways, and bridges. Still, the Principal
Investigators and Advisors for this project judged that the data
was adequate for the analysis purposes proposed in this report.
In Table 3 the lifeline items are: Facility Class 24-multiple
single span bridges; Facility Class 25-continuous/monolithic
bridges; Facility Class 31-underground pipelines; Facility Class
47-railroads; Facility Class 48-highways; Facility Class 55electrical 
towers less than 100 feet high; and Facility Class 56electrical 
towers more than 100 feet high.
In this report the ATC-13 shaking damage data is used in the
following manner. For the lifeline component or segment being
considered, the appropriate table is entered using the MMI value at
the collocation being analyzed. The table is entered to identify
the greatest probability value in the column under the MMI listing.
In the sample below enter the table (on page 21) for MMI = VIII. 
Reading to the left of that maximum probability, the most probable
damage state is then read from the left most column.
Table 3 
SHAKING DAMAGE PROBABILITY MATRICES, ATC-13 Tables 7.10 
Damage Probability Matrices Based on Expert Opinion for 
Earthquake Engineering Facility Classes 
Damage 
State 
Modified Mercalli Intensity 
Multiple Single Span Bridges 
[S
vYI VIi I III II XIII 
1 3.0 iii fit Itt iii nj it 
2 57.0 12.3 in i} in IF tiF 
3 i1i 85.7 70., fi t ri ii iti 
4 fi i 29.1 71.1 iffi iM 
5 it Mii 28.9 82.4 *tit ii 
6 iiitt1i i-n 1H.9 100.0 int 
7 i i ii off ii t 100.0 
Continuous/Monolithic Bridges 
v] V II VIII II I II III 
1 93.6 8. 1 0.9 i ii itn ii 
2 6.4 77.8 17.6 itt it ii tt 
3 it 14.1 78.6 56.5 it i ifn 
4 tit tt 2.9 43.5 1.9 1.2 0.7 
5 i.E.! ii iii it 98.2 36.8 5.7 
6 fiti i tit fit in 61.9 39.1 
7 ti iii iti tit tin 0.1 54.5 
L Underground Pipelines and Conduits 
VI VII VIII II I II XII 
1 100.0 9.a, 20.9 8.7 tlIt ti it: 
2 it 0.2 54.1 34.2 1.3 tit tlit 
3 itn a:: 17.2 36.1 7.9 os5 i1t 
4 it: t11 7.8 21.9 ;89.5 64.5 4.5 
5 tit tit t1t 2t1 1.1 29.6 56.4 
6 Itt tit t1t it: 0.2 3.3 37.9 
7 ilI sit iI in Iti 0.1 1.2 
***Very small probability 
19 
Table 3 (Continued)
SHAKING DAMAGE PROBABILITY MATRICES, ATC-13 Tables 7.10
Damage Modified Mercalli Intensity 
State 
Railroads VI 
ViI Vill II I I1 III
1 94.1 9.8 0.1 M ',0 III 
2 5.9 55.4 12.3 0.3 fit 
tI 
3 III 34.9 87.0 73.9 35.5 10.2 0.4 
4 H, 0.6 25.8 64.1 80.8 25.5 
5 ttt 
oft, 
off off 
0.4 9.0 67.9 
6 HI off "I off Hit 6.2 
7 Ht 
it# it MI I. tt 
Highways
VI Yll Vill ll I II XII 
1 
3.3 18.9 2.9 1.0 it ilI itI
2
6.7 61.5 27.0 13.8 1.3 0.1
3 
4 ftt 19.7 68.8 75.4 59.0 20.5 4.6 
5 tt 1.4 9.8 39.1 65.2 50.2 
It itt 0.6 14.2 43.4
6 ft 
III off Ht H 1.8
HIII III lt 
Electrical Towers Less Than 100 Feet High 
VI VIl Vlll II I II Ill 
1 94.1 6.9 1.0 #iI III 
2 5.9 78.8 51.0 2.9 itI 
3 fit 14.3 48.0 96.3 63.7 10.6 0.5 
4 tt Itt itI 0.8 36.3 82.7 39.0 
5M ott tt If MI 6.7 59.2 
6 Ht HtI *I fiI itI 1.3 
MI #it MI 
Electrical Towers More Than 100 Feet High 
VI VII Vll 1X I 11 
Ill 
1 93.6 7.3 1.8 III oft Ftt III 
2 6.4 72.1 50.9 7.5 0.3
3 off 20.6 47.3 92.2 72.5 16.6 0.8 
4 off ttt off 0.3 27.2 79.4 38.2 
5 fit off Hf III iII 4.0 5B.8 
6 it tII iII It* 
H 2.2 
7 itI itt Mt it 
***Very small probability 
Sample ATC-13 Shaking Damage Matrix
Damage State Modified Mercalli Intensity Index
VI VII VIII IX X XI XII
100 99.8 20.9 8.7 1 
-.2 54.1 34.7 1.3 2 
3 17.2 36.1 7.9 .5 4 
-7.8 21.9 89.5 66.5 4.5 
5 1.1 29.6 56.4 
6 .2 3.3 37.9 
7 .1 1.2 
For a EMI = VIII, the largest probability is 54.1 (identified in 
bold); therefore the assumed damage state is damage state 2 (also 
in bold). The probability that the damage state or greater will
occur is the sum of its probability and all the probabilities for
larger damage at the MMI value of interest: (54.1 + 17.2 + 7.8) = 
79.1%, or 79% for use in the subsequent analyses.
The data represented in Table 3 was developed based on assuming the
facility construction methods were in California. Since California
has incorporated seismic design criteria in some of their codes and
standards, it raises a question as to how the data should be
applied to other U.S. regions. The most direct approach would be 
to consider the design and construction practices at the study area
in question, and to adjust the damage state predicted by Table 3 to
account for differences with respect to California.
Rojahn 2 03 has developed a different approach. He suggests that the
EMI value can be adjusted to account for the different design and
construction practices. Increasing the MMI value would imply that
the local practices are less conservative for earthquake
considerations than those used in California. Decreasing the XMI
value would imply the opposite. Figure 3 shows the U.S. divided
into seismic hazard regions. The Rojahn adjustments for are
presented in Table 4. He has used Figure 3 to divide the U.S. into
five broad regions: California region 7; Other U.S. areas,of region
7; California regions 3 to 6; Puget Sound region 5; and all other
U.S. regions.
Table 4 is provided for information purposes. Data for additional
lifeline components are provided in reference 20. Rojahn did not
justify the selection of the Table 4 values or explain why
adjustments are needed for California (recall that ATC-13 was based
on assuming that it applied to California). One of the important
recommended follow-on studies to the present work is to apply the
present screening tool to another U.S. location. One purpose of
such a study would be to examine the validity of the adjustments to
MMI recommended by Rojahn.
Figure 3, MAP OF U.S. SEISMIC HAZARD REGIONS
NEHRP Seismic Map Areas OTC, 1978; BSSC, 1988). 
Seismic Risk
Regions
Table 4
EMI ADJUSTMENT FOR SHAKING DAMAGE EVALUATION
TO ACCOUNT FOR LOCAL CONDITIONS
(the region numbers correspond to the numbers of Figure 3) 
Multiple Span Continuous Rail beds & 
Region Bridges Highways
California, #7 0 0 0
Other area, #7 I 1 0 
California, #3-6 1 
Puget Sound, #5 0 1 0 
Other U.S. regions 3 2 or 3 0
Railroad Water Trunk Water Pipe
Region and Number Bridges Lines Distribution
California, #7 1 O 1 
Other area, #7 0 0 1 
California, #3-6 1 0 1 
Puget Sound, #5 0 0 1 
Other U.S. regions 1 1 2 
Electrical 
Towers Over Towers Less than
Region and Number 100 ft. high 100 ft. high
California, #7 0 
Other area, #7 0 
California, #3-6 o 0 
Puget Sound, #5 O f 0 
Other U.S. regions '0 1 
Natural Gas Natural Gas Oil
Region and Number Transmission Distribution Pipelines
California, #7 1 '0 1 
Other area, #7 l 0 1 
California, #3-6 l 0 1 
Puget Sound, #5 1 1 1 
Other U.S. regions 1 1 
Fault Displacement
In ATC-13, the maximum fault surface displacement, D, in meters is
calculated from the equation:
Log D = 4.865 + 0.1719 x M; where M is the earthquake 
magnitude
ATC-13 identifies that the fault average displacement is typically
77% of the maximum, and that 30% of the maximum displacement on the
main fault is characteristic of the displacement on subsidiary
faults.
The damage states for the estimated displacement are obtained from
ATC-13 Table 8.9 and are presented in Table 5.
Table 5
LIFELINE DAMAGE STATE FOR FAULT SURFACE DISPLACEMENTS,
ATC-13 Table 8.9
Facility Type Damage State (% damage is given in the parentheses) 
and Location For Various Values of Displacement in meters
Displacement = 0.2 m 0.6 m 1 m 3.5 m 10 m 
Subsurface Structure 
In Fault Zone 5(50) 6(80) 7(100) 7(100) 7(100) 
In Drag Zone 4(20) 5(40) 5(60) 6(80) 7(100) 
Surface Structures 
In Fault Zone 3(10) 4(30) 6(70) 7(100) 7(100) 
In Drag Zone 0(0) 0(°) 3(2) 3(10) 4(20) 
The "Fault Zone" is defined as being within 100 meters of the fault
trace, the "Drag Zone" is defined as being within 100 to 200 meters
of the fault trace. If lifeline components are judged to have
failed because of fault displacement, then the collocation impact
would be only an increase in the time to restore the lifeline to
its needed level of operation (e.g., damage greater than
catastrophic is not meaningful). Such time increases would be
attributed to the construction activity and the need to assure that
construction on one lifeline does not lead to damage on
reconstructed other lifelines.
Soil Movement
Many texts separately define the impacts due to landslides and
lateral spread (or liquefaction). However, they may be thought of
as being part of a continuum of soil movement with the slope of the
topography being a parameter that identifies whether the movement
should be calculated as a landslide or a lateral spread (or
liquefaction). That is the approach proposed in the present
analysis method.
Landslide (landslides occur on slopes greater
than 50) 
It is proposed that the historical landslides in the study area be
identified and considered as potential landslide regions when the
collocation evaluation is made. Keefer and Wilsonsi0) and Sadler
and Morton"1')have identified that landslides are associated with
many historical earthquakes and that shaking is one of the main
triggering agents for landslides. Actual site reconnaissance
visits are recommended as a means to verify the location of
historical landslides for any area being studied. In the present
study, a comparison of the known slides with the geologic unit map
identified that many of the landslides were associated with areas
where Pelona Schist is the bedrock unit. Other researchers are
advised to examine the geologic sediments and rocks in the areas
where they intend to evaluate collocation and to be sensitive to
the location of Pelona Schist.
2) be used to
It is proposed that the method of Legg et. al. CY
identify additional areas where landslides may occur. It is based
on the sliding block model proposed by Newmark1 3); Wilson & 
Keeferc'4) have proposed a similar model. However, the Wilson and
Reefer model requires using recorded accelerograms or predictions
of ground acceleration while the Legg method is related to using
MMI. The Legg model is the method used in ATC-13 to define the
damage state and probability of damage for landslides. Also, it
will be easier to apply the Legg model to other regions in the
U.S. Because of these items, the Legg method was adopted for
predicting additional landslide areas.
The Legg method consists of the following basic steps:
Step I Solve for the "critical acceleration" of the slope
for a given combination of slope angle and soil
properties. A formula derived from the stability
solution of an infinite slope was used by Legg and
also by Wilson and Reefer, and it is provided below.
Step 2 Use the critical acceleration to enter a table of 
"slope failure state"' versus DMI value. The table
values identify the potential for the slope to move
as a landslide. The tables are provided as Table
8.7 of ATC-13 and are reproduced below as Table 7.
Step 3 The slope state is related to damage state in Table
8.8 of ATC-13, which is presented below as Table 8.
However, the ATC-13 Table 8.8 has been extended to
more accurately account for buried lifelines, based
upon expert opinion obtained during the present
study. 
The formula for the critical acceleration is given by:
ar/g = c/(yh) + cos e tan ˘ sin 0 ; where 
a= the critical acceleration, ft/sec2
g = the gravitational constant, 32.2 ft/sec2 
c = the effective soil cohesion factor, lb/ft2 
= the soil density, typically 100 lb/ft3 
h = the thickness of the soil block, typically 10 ft 
l = the slope angle, degrees 
= the angle of friction of the slope material, 
degrees
Note, this equation applies to dry slopes.
25
The soil parameters recommended for use are given in Table 6.
Table 6
SOIL PARAMETERS FOR CALCULATING LANDSLIDE POTENTIAL
Shear Strength Parameters
Geologic Unit Cohesion. c (Vfs) Friction Angle, (degree)
Paleozoic Rocks 300 35
Older Cenozoic Rocks 0 35
Older Alluvium 0 30
Young Alluvium at Shallow
Ground Water & Pelona Schist 0 20
Wilson and Keefer(14 have also developed an analysis for saturated
and dry slopes. They use a 35 degree friction angle for sands,
sandstones, and crystalline rocks, and 20 degrees for clayey soils
and shales. They present a graph of the critical acceleration as:
0.5 
Z Plots of critical acceleration (Ad) versus slope steepness
for three sets of lithologies: group A, strongly cemented rocks 
G% crystalline rock and well-cemented sandstone); group B, weakly 
cemented
rocks (sandy soil and poorly cemented sandstone); group
argillaceous rocks (clayey soil and shale). The cohesion factor, 
for group A assumes values of c'-=300 psf and 
God
The angle of internal friction (e )(peak strength, undrained 
conditions) is 35°for sands, sandstone, and crystalline rocks and 
200 for clayey soils and shales. The solid lines depict dry slope 
materials, and the dashed lines depict saturation from the slide 
plane to the surface. 
10˘ 20 300 i 400 500 
5% 15% 30% 50% 70% 100% 
SLOPE 
Either the Legg formula or the Wilson graph is acceptable for
determining the critical acceleration, it's numeric value will be
used in the Legg tables discussed below.
The appropriate soil parameters from Table 6 or other references
should be identified. The formula or graph is then used to
determine the value of the critical acceleration, which in turn
determines the slope stability (unstable, low, moderate, high,
stable, very stable) so that the ATC-13 Table 8.7 (Table 7 given
below) can be used to define the state of slope failure (Table 7
uses the Legg definitions for terms of slope failure state and
slope stability scale, and those definitions also are provided with
the table).
Table 7 
LANDSLIDE SLOPE FAILURE PROBABILITY MATRICES, ATC-13 Table 8.7 
Scope Fit flue Probability 
(Summer Conditions) 
SLOPE: UNSTABLE, .4 c .01 r SLOPE MIR, 0.3 p < a, c 0.5 r 
SLOPE YMI SLOPE 
FAILURE 
STATE VI VD vm TX x Xl 
FAILUR E 
STATE Vi VE vm 
0 0 0 a 0 0 0 LIGHT 100 100 100 95 &5 0 s0 
MODERATE a 0 a a a 0 0 MODERATE 0 0D 0 5 II 15 20 
HEAVYf00 50 40 30 20 5 a HEAVY 0 0 0 0 5 5 15 
SEVERI S0 40 45 50 55 60 S5 SEVERE C, o 0 0 
CATASTROPHIC l 1a i5 20 i5 15 50 CATASTROPHE 0o0 a 0 0 0 0 
˝ 10I% 100% 100% 30% 100% 100% 2100% Zp 100% 300%. 10a 100% 100% 1005S 10% 
LOW, .01 a, c 0 SLOPE SABUI STAEL., 0.5 c %( 0.7 
SLOZ MM
SLOPE MMI RFAGER
FAILURE
YADLUR! 
VI Y3 ix X XI XE STATE VI vu 1x x X1 XII
STATE 
40 25 15 10 5 0 0 LIGHT 10O 100 100 100 00 Is 75 
LIGHT 
O MODERATE 0 0 0 0 10 l0 15
moderate 50 30 35 l0 20 10 0 
HEAVY 25 35 40 40 35 35 30 HEAVY 0 0 0 0 C 5 l a0 
5 10 10 15 20 35 40 SEVERE 0 0 0 0 0 0 C
SEVERC 
CATASTROPHIC 0 0 a 5 10 20 30 CATASTROPHIC C 0 0 0 0 0 0 
10096 100% 100% 100% 100%
1 100% 300% 100% 100% 100% 100%: 100%.
SLOPE Stability XODIXATE, 0.1 a, 3E SIAPE Y=V 1TA l 0.1 E < 4e 
SLOPE MMI SLOPE MMI 
Failure 
STATE VI VM TX x xi STATE V a TX X n XE 
LIGHT 100 100 15 1 55 20 0 LIGHT 100 100 100 100 100 t SO 
MODERATE O a 10 20 25 Jo 10 MODERATE 0 0 0 0 0 1 15 
HEAVY 0 0 5 la 15 25 40 HEAVY I 0 0 0 0 0 5 
SEVERE 0 0 0 0 5 i5 20 SEVERE 0 0 0 0 0 0 1 
CATASTROPHIC a a 0 0 0 10 10 CATASTROPHIC 0 0 0 0 0 a 0 
Zp 100% l10 100% 100% 100% 300% l0 , 100 100 10100 %00% 100% 100% 100% 
Table 7 (Continued)
Definitions
SLOPE FAILURE STATE SCALE RELATIVE SEISMIC SLOPE
STABILITY SCALE
LIGHT-Insignificant ground
movement, no apparent V Very stable, not likely to 
potential for move under severe shaking,
landslide failure, ac 2 0.7g.
ground shaking effect
only. Predicted S Stable, may undergo slight
displacement less
movement under severe
than 0.5 cm. shaking, 0.5g < a < 0.7g.
MODERATE-Moderate ground
H High, may undergo moderate 
failure, small cracks movement under severe
likely to form, shaking, some landslides
cracks similar to related to steep slopes,
having a lurch saturated conditions, and
phenomena. Predicted adverse dips, 0.3g < ac <
displacement 0.5 to 0.5g.
5.0 cm.
M Moderate, may undergo
HEAVY-Major ground failure, major movement under
moderate cracks and severe shaking or moderate
landslide movement under moderate
displacements with shaking, numerous
effects similar to landslides, rock falls
liquefaction or abundant, unconsolidated
lateral spread.
material deforms and
Predicted fails, 0.lg < ac < 0.3g.
displacement 5.0 to
50 cm. L Low, may undergo major 
movement under moderate
SEVERE-Extreme ground shaking, abundant
failure, large cracks landslides of all types,
and landslide 0.0lg 5 ac < 0.1g.
displacements with
effects similar to U Unstable, may undergo
large-scale fault major movement under
displacement. slight shaking, most of
Predicted the area and/or material
displacement 50 to falls, ac < 0.01g.
500 cm.
cm = centimeter 
CATASTROPHIC-Total ground 
g = gravitational constant 
failure, with
predicted
displacement greater 
than 500 cm.
To use Table 7, it is necessary to enter it with the critical
acceleration, ac, and the MMI value. The critical acceleration
value determines which sub-table is used. Within that sub-table,
in the MMI column, identify the location with the peak probability.
The slope failure state is read from the left-most column at the
row that contains the peak probability value. The probability that
the condition or worse will exist is the sum of the individual
probabilities for that slope state and all worse slope state
conditions. This is similar to how the shaking damage state and
its probability were calculated.
Next, the slope failure status (light, moderate, heavy, severe,
catastrophic) is converted to a damage state (and also a percent 
damage) by using ATC-13 Table 8.8 (Table 8 below). ATC-13 provides
a single conversion value for all lifelines. This has been
expanded in Table 8 to account for key buried lifelines. The new
values were based on expert opinion obtained during the present 
study. 
Table 8
CONVERSION OF LANDSLIDE SLOPE FAILURE STATE TO DAMAGE STATE
Damage State and (% Damage) 
ATC-13 Values New Values Determine During This Study
Slope Failure for all High Strength Low Strength
State Lifelines 
Light 0-3 (0%) 0-2 (0%) 0-3 (0%) 
Moderate 4 (15%) 3 (0%) 4 (30%) 
Heavy 5 (50%) 4 (15%) 5 (60%) 
Severe 
Catastrophic 
(80%) 
(100%) 
(50%) 
(100%) 
(90S%) 
(100%) 
The definition of high strength buried lifelines used to determine
the damage state is: continuous steel pipelines constructed
according to modern quality control standards with full penetration
girth welds; welds and inspection performed according to API 1104
or equivalent.
The definition of the buried lifelines which should be represented
by the original ATC-13 definitions is: pipelines and conduits
constructed according to modern standards with average to good
workmanship, other than the high strength lifelines defined above.
Lifelines in this category are expected to include electric cables,
steel pipelines with welded slip joints, ductile iron pipelines,
telecommunication conduits, reinforced concrete pipe including
concrete steel cylinder pipe, and plastic pipelines and conduits.
Also, if the high strength lifelines are oriented so that the
landslide motion is expected to place them into compression, they
should be analyzed in this category. Other lifelines not included
in the High Strength or Low Strength definitions should be
evaluated using the ATC-13 column.
The definition of low strength buried lifelines is: pipelines and
conduits sensitive to ground deformation because of age, brittle
materials, corrosion, and potentially weak and defective welds.
Lifelines in this category include cast iron, riveted steel,
asbestos cement, and unreinforced concrete pipelines; pipelines
with oxyacetylene welds; and pipelines and conduits with corrosion
problems. If other non high strength buried lifelines are oriented
so that they are perpendicular to the expected landslide motion
(e.g., their orientation is such that they will be put into
compression by the landslide), then they should be analyzed as a
low strength lifeline rather than with the ATC-13 column.
Liquefaction or Lateral Spread (lateral spread
occurs on slopes of 1-5°)
It is proposed that the Liquefaction Severity Index (LSI) be used
to correlate the liquefaction or lateral spread damage and the
probability of damage. The LSI is defined in the work of Youd and
Perkins'15). The following material was developed from expert
consultive support provided during this study by Dr. T.D. O'Rourke
of Cornell University 
In a manner similar to the critical acceleration defined for 
landslides, a critical LSI is defined in Table 9 below. The basis 
for its use and the LSI damage probabilities of Table 10 is the
work of Harding 6) which has shown that substantial lateral
spreading can be triggered at a critical acceleration, ac, of 0.05
to 0.15 g.
Table 9
RELATIONSHIP BETWEEN LIQUEFACTION SEVERITY INDEX (LSI)
AND DAMAGE STATE 
Physical Lateral Equivalent 
Ground Movement LSI Damage State Condition 
< 0.5 inch <1 3 light 
0.5 to 5.0 inches 1-5 4 moderate 
5 to 30 inches 
30 to 90 inches 
5-30 
30-90 
heavy severe 
> 90 inches > 90 7 catastrophic 
O'Rourke has prepared a regression analysis of the observed
relationship between the MMI index and the LSI index for four
earthquakes; the 1906 San Francisco, the 1964 Alaska, the 1971 San
Fernando, and the 1979 Imperial Valley earthquakes. The
observations identified LSI values of 5 to 100 for MMI values of V
to X. The resulting regression curve (with an r2 = 0.68) is: 
LSI=O . 226x1 0 255 
The equation can be used to calculate the LSI number, and then
Table 9 can be used to define the damage state. Graphically, the 
relationship between MMI and damage state is presented in Figure 4
below.
The probability that the liquefaction damage state will occur is
given in Table 10. Table 10 (which replaces ATC-13 Table 8.4)
applies to soil environments in which liquefaction is likely to
occur under strong earthquake shaking. These environments include:
active flood plains, deltas, other areas of gently sloping late
Holocene fluvial deposits, and loose sandy fill below the water
table (which are generally placed by end dumping or hydraulic fill
methods). The table does not apply to late Pleistocene Alluvium,
for which the probabilities of liquefaction are negligible for
intensities equal to or less than MMI of X. Thus, the combination
of the LSI equation and Table 9 (or the use of Figure 4) with Table 
10 is analogous to landslide calculations for low stability material.
Table 10
PROBABILITY OF LIQUEFACTION GROUND FAILURE, PERCENT
Liquefaction Damage MMI Value
State VI VII VIII IX X XI XII
3 Light 75 50 20 10 0 0 0
4 Moderate 20 30 40 25 15 10 0 
5 Heavy 5 20 30 40, 25 25 20 
6 Severe 0 0 10 20 35 40 30 
7 Catastrophic 0 0 0 5 15 25 50 
The new method developed during this study adds details to the
level of analysis available from ATC-13. It identifies a range of
damage from light to catastrophic (compared to the assumed
catastrophic levels of ATC-13) and a full range of probabilities
that the damage state will occur. Since it is based on observed
liquefaction damage from California earthquakes, additional
evaluation of the recommended approach at other U.S. locations is
warranted.
Highway and Railroad Bridges
The ATC-13 shaking intensity matrices (Table 3) identify three
broad classes for bridges: multiple simple span bridges, continuous
and multiple span bridges, and long span or major bridges. It is
difficult to fit every railroad and highway bridge into one of
these broad classifications. One example of how owner-supplied
information can be used to improve upon the direct use of the ATC13 
guidance is found in the methods of the California Department of
Transportation( 118 ,1
7 , 9 )'(CALTRANS). CALTRANS has a method to identify
the priority for performing retrofits, to their bridges to reduce
their vulnerability to earthquakes. This improved data was
LIQUEFACTION INDUCED DAMAGE 
Range Of Field Observation Data _ 
I ii III IV V VI VIl IX: X Xi 
Modifed index 
integrated with the ATC-13 data to provide more discrimination
capabilities for evaluating railroad and highway bridges. The
resulting procedures (described below) are fully applicable to
locations outside of California if the needed data on the 
individual bridges are known. The CALTRANS method includes
factors, such as traffic loading and detour routes1 that are
important for making decisions about whether to spend money to
retrofit a bridge, but they are not important for determining the 
damage state of the bridge. However, other factors, such as the
bridge sub and superstructure, the design codes used, and the
bridge geometry can be related directly to the ability of the
bridge to resist earthquake damage.
The method being proposed in this report calculates a parameter
that can be used to adjust the damage state value for shaking as
determined by the ATC-13 matrices (Table 3 of this report). The 
evaluation is based on starting with the ATC-13 shaking probability
matrix for Continuous and Multiple Span Bridges. The procedures
discussed above on how to use Table 3 to define the damage state
and the probability that the damage state or greater will occur are
used to calculate a tentative damage state. A bridge vulnerability
index then is calculated and used to determine if the tentative
damage state should be changed (the probability is not changed).
The decision to adjust the Table 3 tentative damage state value is
based on the numeric values identified below in Table 11 (high
values of the Bridge Vulnerability Index mean that the damage will
be more severe than that predicted by Table 3).
Table 11
RELATIONSHIP OF BRIDGE VULNERABILITY INDEX TO
BRIDGE DAMAGE STATE
Change to Table 3 Continuous
Bridge Vulnerability & Multiple Span Bridge 
Index Value Damage State Value
0.0 0.2 Lower the Damage State by 
two increments
0.2 0.4 Lower the Damage State by 
one increment
0.4 0.6 No Change 
0.6 0.8 Increase the Damage State by 
one increment 
0.8 1.0 Increase the Damage State by 
two increments 
The numeric value of the Bridge Vulnerability Index is calculated
by multiplying a Raw Score by a Multiplying Factor (in the
CALTRANS' method, the terms were weighting factor and pre-weighting
factor, respectfully). The Raw Score is assigned by the importance
of the bridge factor being evaluated, the Multiplying Factor is a
weighting scale that determines how earthquake resistant the Raw
Score items is. Table 12 presents the numeric values of the Raw
Score and the Multiplying Factors.
There are seven categories that are analyzed: 1) abutments; 2)
piers; 3) soil type; 4) superstructure type; 5) design code or
specification used; 6) bridge height; and 7) bridge skew and
curvature. A separate number (the raw score times the multiplying
factor) is calculated for each of the seven categories and then the
individual numbers are summed. The sum is divided by 100 to give
the total Bridge Vulnerability Index value.
In applying the incremental change to a tentative damage state from
Table 3, if this results in a damage state less than 1 or greater
than 7 use those limit values. Damage states for long span (length
greater than 400 feet) and major bridges may be estimated using
this procedure, but it is recommended that such structures be
subjected to special studies whenever possible. It is emphasized
that the above Bridge Vulnerability Index is for shaking damage.
Special conditions, such as liquefaction, require additional
analysis. 
The analysis factors required to enter Table 12 can be obtained
from the general design drawings of the bridge or by field
reconnaissance. Some assumptions may have to be made with respect
to foundation design in the latter case.
Railway bridges have proved to be somewhat more resistant to ground
shaking than highway bridges, in spite of the fact that the
American Railway Engineering Association (AREA) specifications make
no specific recommendations with regard to earthquake forces. This
is probably due to the fact that railroad brides have an allowance
for lateral loads (originally, the allowance was to account for the
loads produced by steam locomotives). Prior to 1935, this
allowance was 5% of the live load (typically based on a Cooper E60
engine, or about 852,000 lbs. on a 109 ft. span), but not more than
400 lbs. per foot of track. In 1935, this was changed to provide
for a lateral load of 20,000 lbs. applied at the top of the rail at
any point in the span. In 1950, AREA provided for higher allowable
stresses, so that the allowance became somewhat less conservative.
The multiplying factors of Table 12 for railroad bridges reflects
these facts.
Table 12 
BRIDGE VULNERABILITY INDEX FOR EARTHQUAKE SHAKING DAMAGE 
Bridge Raw Multiplying 
Element Score Multiplying Factor Criteria Factor Value
SUBSTRUCTURE
Abutments 10 Integral with pile foundation 0.0 
Integral with spread footing 0.5 
Hinge seat with restraints 0.6 
Hinge seat, all other types 1. 0 
Piers 15 wall 0.2
multiple column bent 0.5 
single column bent
Note, if a spread footing foundation is used, add 0.2 to the
pier multiplying factor, if the columns have been reinforced 
to recent seismic codes, subtract 0.3 from the pier
multiplying factor.
Soil Type 15 Rock or soil with bearing of
more than 4 tons/ft2 0.0
Soil with bearing of 2 4 tons/f t 0.1 
Soil with bearing of less than 2 tons/ft2 0.5
SUPERSTRUCTURE
Type 20 Highway Bridges
Simple span, box or slab 
Single span, arches, reinforced concrete
or well constructed masonry O0 . 1 
Simple span, steel or concrete beams 0.5 
Simple span, steel truss 0.5 
Multiple spans, continuous with no hinge 0. 0 
Multiple spans, continuous with 1 hinge 0.5 
Multiple spans, simple beams 1. 0 
Multiple spans, continuous with 2 or
or more hinges 1 . 0 
Railroad Bridges
Simple spans, steel with full truss 0.3 
Simple spans, deck or half truss 0.4 
Simple spans, steel or concrete ballasted 0.5 
Simple spans, steel or concrete beams 1.0 
Multiple spans, fully continuous 0.0 
Multiple spans, simple beams. 1.0 
Multiple spans, continuous with hinges 1.0 
Note, for both highway and railroad bridges with hinges, 
subtract 0.4 from the multiplying factor if restrainers have
been added. Subtract an additional 0.3 if the columns have
been reinforced to resist earthquake forces.
Table 12 (Continued)
BRIDGE VULNERABILITY INDEX FOR EARTHQUAKE SHAKING DAMAGE
Bridge Raw Multiplying
Element Score Multiplying Factor Criteria Factor Value
DESIGN CODE OR SPECIFICATION
Code used 20 Highways
CALTRANS * after 1978 or AASHTO* after 1987 0.0
CALTRANS between 1972 and 1978 0.2 
CALTRANS prior to 1972 and 
AASHTO prior to 1950 0.5 
AASHTO from 1950 to 1987 1.0 
Note, AASHTO, from 1950 to 1987, leaves the earthquake
considerations to the States. If it is known that the State has
no such consideration, use 2.0 as the Multiplying Factor value.
Railroads
AREA* from 1935 to 1950 0.5
AREA from 1950 to present 0.7
AREA prior to 1935 0.8
Note, for the condition of bridge, modify the design code or
specification Multiplying Factor by adding the following to the
factor:
Good or fair condition 0.0
Poor condition 0.2
GEOMETRY
Height 10 Less than 5 feet 0.2
5 to 15 feet 0.7
15 to 25 feet 0.9
25 feet and greater 1.0
Skew* and 10 Skew less than 200 and
curvature radius greater than 1000 ft. 0.0
Skew 200-400 and/or
radius greater than 500 ft. 0.1
Skew greater than 400 and/or
radius less than 500 feet 0.4
Key *
AASHTO, American Association of State Highway & Transportation
Officials
AREA, American Railroad Engineering Association
CALTRANS, California Department of Transportation
Skew is defined as the angle that abutments and piers make with
respect to the normal to the highway (or railway) alignment.
That is, when the plane of the abutment or pier is aligned
parallel to the normal to the road (or rail bed) alignment, the
skew is 0.
Times to Restore the Lifeline to its Needed Service
Once the lifeline components of interest have been identified and
the damage state and probability that the damage condition or worse
will occur have been calculated from the above tables and formulas,
the time to restore the lifeline component or segment from the
total calculated damage state to the operating level needed has to 
be determined.
The restoration time is a combination of the time to repair the
lifeline segment or component assuming all the equipment, material,
and personnel are available at the damage site, plus the access
time to get the equipment and material to the damage site, plus the
delay time needed to obtain the equipment and material required for
making the repair. The way to calculate these items is given next. 
Repair Time to restore the damaged lifeline to service
With the damage state known, the time to rep-air the lifeline
component or segment (assuming the equipment and material are at
the damage location) can be calculated from Table 9.11 of ATC-13.
The key information of Table 9.11 is provided below as Table 13. 
If intermediate operating conditions (e.g., repair to less than
100% capacity) are acceptable, the intermediate repair times of the
ATC-13 tables can be used or the plots of those tables provided by
Rojahn(20) can be used to estimate such intermediate condition
repair times. The newer curves by Rojahn are curve fits of the
data of ATC-13l23, thus they are not exact replications of the data.
But they may be more convenient to use since they relate the repair
time to MMI instead of to the damage state as is done in ATC-13.
Also, if there is concern about the magnitude of the repair time
estimated, Table I.1 of Appendix I of ATC-13 can be used to
determine the range of repair times identified by the experts that
prepared Table 9.11. It is important to recognize that the actual
repair time is not used directly to estimate the impact of
collocation on the vulnerability of lifelines to earthquakes (as
will be shown below).
Eleven of the more important repair tables are presented in Table
13 (some of the tables were adjusted from the ATC-13 values to
account for expert opinion obtained during the present study). To
make a specific estimate of lifeline repair time, enter the proper
lifeline table at the row that identifies the damage state and move
to the right until the correct lifeline column is encountered.
Then read the time, in days, required to restore the lifeline to
full capacity from that damage state. The ATC-13 lifelines of
interest are: lac-petroleum transmission pipelines, 25a-highway
major bridges, 25c-highway conventional bridges, 25d-freeways and
highways, 26a-railroad bridges, 26c-railroad roadbeds, 29belectrical 
transmission towers, 30f-water trunk lines, 31a-sewer
lines, 32a-natural gas transmission lines, and 32d-natural gas
distribution lines. It should be recalled, however, that better
estimates of repair time are probably available from the individual
lifeline owners, as they may have site specific conditions included
in their estimates.
.Table 13
ESTIMATED LIFELINE REPAIR TIMES TO 100% OPERATING CAPACITY 
ATC-13 Table 9.11
(Times in Days)
Damage 
State 
1* 
Highway 
Bed 
1 
Railroad 
Bed** 
1 
Highway** 
Conventional 
Bridge 
1 
Highway 
Major 
Bridge 
1 
Railroad 
Bridge 
1 
Water 
Trunk 
Line 
1 
2 1 1 1 2 1 2 
3 7 2 8 7 8 3 
4 41 11 84 141 58 10 
5 147 41 303 392 213 25 
6 292 82 686 845 468 74 
7 437 120 752 947 606 156 
Natural Gas** Natural Gas Fiber Electrical
Damage Distribution & Transmission Optic** Transmission Sewer
State Petroleum Lines Pipelines Conduits Towers Lines
1*1 1 11 1 
21 11 1 3 
33 3 12 5
46 11 3 17 18 
5 19 25 10 49 63 
6 44 44
24 82 102
7 55 75 30 127 141
* Damage State 1 has a 1 day allowance to allow for 
inspection to determine the actual damage state that
exists at the lifeline
** These values were determined by expert opinion during 
this study
Access Time to get the equipment and material to the
damage site
Next it is necessary to estimate the time to get the equipment and
repair material to the site. This time is the time to get
construction equipment and material to the damage site, and it
should not be confused with the time it would take to get general
population traffic to the site or with the time it would take for
repair crews to get to the damage site. In many situations, and
especially for lifelines such as pipelines, fiber optics, and
electrical transmission towers, most of the necessary equipment and
material can be driven to the damage location either along the
highways, unpaved access roads, or cross country if the land is dry
and accessible. In some of the more rugged regions they can be
helicoptered to the site. An exception would be if wet ground or
large water bodies must be negotiated. Thus, in general, for those
lifelines the access time is one or two days, depending upon the
location of the segment or component of the lifeline system being
examined. If access along the highway is required it should be
calculated as described below for the railroads and the highways.
For many of the railroad or highway components and segments, the
access will have to be along the railroad or highway itself because
of the size and weight of the material and equipment that is
required. In such cases, it will be necessary to estimate the
repair times for damage along the route prior to the location being
studied. The individual repair times must then be added for each
disruption that occurs before the location being studied to obtain
a total estimated access time. Alternatively, detours can be used
to calculate a "by pass" time estimate.
Equipment and Material Time to have those items available
For many of the lifelines, the owners have their own operating
equipment and have prepositioned repair material along their
lifeline routes. When they don't have suitable repair equipment in 
their operating stock, they may have existing agreements with other 
firms, to provide such equipment during emergencies. Frequently,
utility lifeline owners have reciprocal agreements with other 
utilities to provide personnel and equipment during emergency
periods. This preplanning can decrease the time it takes to have
equipment and repair material available to transport to the damage
location.
The problem of material availability can be pronounced for railway
and highway bridge repairs. In those cases, the time required to
fabricate off site the needed components must be accounted for in
the estimation of the delays in having equipment and material
available.
In almost all cases, it can be assumed that the equipment will not 
be available during the emergency phase of the earthquake, since it
will be diverted to life-saving duty at that time. However, prior
earthquake response experience indicates that most equipment and
needed material will be made available within one or two days.
4.3 Collocation Analysis
Section 4.2 presented a number of analysis methods that can be used
to determine the damage state, the probability that the damage
state or worse will occur, and the estimated restoration time to
return each lifeline component or segment to its needed service
level.
In the collocation analysis activities, a collocation damage
scenario is developed and the unknown conditions (either damage
state, probability of damage, restoration time, or any combination
of those items) are recalculated for the assumed collocation damage
scenario using the methods of Section 4.2. The collocation damage
scenario should be based on the knowledge of how the individual
lifelines would have responded if they had be the only lifeline at
the collocation point, the estimate of the types of impacts that
one lifeline failure could impose on another nearby lifeline, and
the zone of influence that one lifeline has.
This process requires that technical judgements be applied, based
on knowing the expected damage states of the collocated lifelines,
the seismic and geologic conditions, information about the
lifelines themselves (such information as the design conditions,
construction history, repair and maintenance history, and other
pertinent facts), and other lessons learned from prior earthquakes.
It will be important to obtain as much information from the
lifeline owners as possible to help guide the collocation damage
scenario analyses.
It is also important to recognize that there is a zone of
influence, beyond which the impact of one lifeline on another would
be negligible. During this study, expert opinion was used to
estimate the appropriate radii of influence zones for the lifelines
found in the Cajon Pass. The results are given in Table 14.
Care must be taken to differentiate between the zone of influence
and the actual influence or damage caused. For example, the zone
of influence of a failed dam is based on the path of the water that
spills past the dam. It includes the actual pathway and the area
that the water would inundate. The actual impact of the failed dam
could be erosion of foundations of other lifelines (thereby causing
them to collapse) or the flooding of them (perhaps restricting
their ability to function). There may be no influence on one
lifeline, while the impact on another could be pronounced. Some of
the impacts may be subtle. For example, a failed communication
lifeline may have no immediate impact on the physical state or
condition of other nearby lifelines. Its impact, however, could be
tied to increasing the restoration time of nearby lifelines due to
the difficulty of maintaining communications with the repair
personnel. In the present context of lifeline vulnerability, the
impact of one lifeline on a collocated or nearby lifeline can be
the damage state, the probability of damage, or the restoration of
service time. Other impacts, although real, have no way to be
accounted for in the analysis method.
Although the values in Table 14 are considered appropriate for the
semi-desert region of the Cajon Pass, California, for which they
were prepared, it will be important to validate these values when
the lifeline zones of influence are evaluated for other at-risk or
collocation conditions.
Table 14
LIFELINE ZONES OF PHYSICAL INFLUENCE
Liquid Fuel Pipeline The drainage path and catchment area 
for any liquids spilled; two times
the pipe burial depth for any soil
cratering impacts due to pipeline
ruptures; 100 feet if explosion
impacts are estimated; ground erosion
paths for liquids spilled; and the
burn path if fires are estimated.
Natural Gas Pipeline Two times the burial depth for any 
soil cratering impacts due to
pipeline ruptures.; 100 feet if
explosion impacts are estimated; and
the burn path if fires are estimated.
Fiber Optic Cables Zero feet (e.g., no physical impact 
on other lifelines).
Roadways 40 feet from the road edge; a 
possible ignition source for fuel
lifelines.
Railroads 40 feet from the track edge; a 
possible ignition source for fuel
lifelines.
Overhead Electrical A radius equal to the height of the 
Transmission Towers & tower for physical contact; 
Power Lines a possible source of ignition for
fuel lifelines.
Bridges For an area centered on the bridge, 
twice the length of the bridge and 40
feet-on either side of the bridge.
Dams., Reservoirs & The drainage path and inundation 
Canals, areas for the spilled water.
Water & Sewer Lines The erosion area downstream of the 
break (sewers only if they are
pressurized); the catchment area for
the spilled fluids.
It is anticipated/ but not required, that collocation impact
scenarios will follow the following general guidance.
Impacts on Damage State
One of the easier direct impacts to hypothesize will be that the
collocation conditions will lead to an increase in the damage state
of one or both of the collocated lifelines (if there are more than
two collocated lifelines this applies to all of them). It is easy
to understand the damage state, as it relates to a physical
condition. Because the individual lifeline damage states assuming
no collocation are known, those values can be used to help
understand how the lifeline could impact another nearby lifeline.
If, for example, light damage of a pipeline had been calculated, it
would be expected to cause no direct change in the damage state of
a nearby bridge. However, if the bridge had been estimated to
collapse, it would be reasonable to estimate that within the
bridge's zone of influence it would lead to failure of the pipeline
(this example also illustrates that the impacts are not necessarily
reciprocal).
As another example of how collocation impacts on damage state can
be estimated, consider the condition of a pipeline and a fiber
optic conduit hung from a bridge. The earthquake vibration may not
be enough to cause serious damage to the bridge or to the pipeline
or conduit if they were not collocated with each other. However,
the vibrations may cause the anchors holding the heavy pipeline to
the bridge to fail. As the pipeline sags (but does not fail) it
could fall onto the lower conduit, causing it to fail. The
collocation damage state hypothesis would then be: no impact on the
bridge; a small increase in damage state of the pipeline to account
for the work required to rehang the pipeline; and catastrophic
failure of the fiber optic conduit.
Special attention should be given to the collocation of fuel
carrying lifelines with other lifelines that have the ability to
provide an ignition source. The resulting fire and/or explosion
could lead to significant collocation damage. Similarly, broken
pipelines which eject fluids could lead to foundation erosion
problems that would result in increased damage to nearby lifelines.
Impacts on Probability of Damage
The probability of damage does not directly enter into the
calculation of the damage state level or the time to repair the
damage. It is, as will be discussed below, a very important item
for determining the key result of the collocation analysis, the
probable incremental change in restoration of service time.
There are several ways to estimate the change in the probability
that damage will occur, none are exact and there are no statistics
available from the literature on earthquakes. However, there are
some insights available to guide the analysts.
If the probabilities for two lifelines, assuming no collocation
conditions, are P1 and P2, they represent an upper bound on the
For example, if probability that a collocation damage would occur. 
the probability that lifeline 1 would fail is PI, and it is, known
with a100% probability,
that if lifeline I fails it will cause, 
damage to lifeline 2, then the probability that lifeline 2 receives
(e.g., P1 x 100%). Similarly, the
collocation damage is also P1 
upper bound on the probability that lifeline 2 has damaged lifeline
1 is P2. 
As a practical matter, the collocation damage likely will be less,
since there is seldom a 100% chance that the collocation damage
scenario will occur. A useful measure of the probability that the
collocation event has occurred is the product of the two
probabilities that the single independent events that were used to
develop the collocation scenario have occurred (the independent
events are the estimate of the damage state of each lifeline
assuming there was no collocation). In the present case, that is
found by multiplying PI x P2. The product can be interpreted as
follows. It represents the increase in probability that the two
independent lifeline damages will occur during the same initiating
event. If both events must occur before the collocation damage
scenario can take place, then it is a measure of the probability of
the collocation damage scenario.
The actual probability that the collocation event will occur should
be a number between the numerical limits of P1 and (PI x P2) for
having lifeline 1 cause additional damage to lifeline 2, and P2 and 
(P2 x Pi) for having lifeline 2 cause additional damage to lifeline 
1. It is recommended that for calculational purposes, the product
Pi x P2 be used to characterize the hypothesized collocation damage
scenario.
Impacts on Time to Restore Lifeline Service
As discussed above, the time to restore lifeline service is
composed of the sum of the time to repair the lifeline damage, the
time to access the damage site with equipment and material, and the
time to obtain the equipment and material.
The hypothesized collocation damage scenario does not have to
assume a repair time. Once the collocation damage state is known,
the repair time can be obtained from Table 13.
However, it is reasonable to include in the collocation damage
scenario impacts on accessibility to the damage site, which has the 
impact of increasing the overall restoration of service time
estimate. In fact, this is probably one of the more significant
aspects of the collocation damage scenario, e.g., the estimation of
the additional direct delays that will be incurred because of the
collocation of the lifelines. The greater the level of damage
estimated for each of the separate lifelines, assuming that there
is no collocation, the greater the anticipated delays that will
result from their actually being collocated.
The following are offered as possible examples of how collocation
could create access delays that would increase the time to restore
the lifeline to service. General congestion at the collocation
location because there are multiple lifelines could delay the start
of repair work on a lifeline. Concern over the possibility of
leaking fuel may cause all work to be delayed until it can be
confirmed that it is safe to have workers in the area. Spilled
liquid fuels may have to be treated and/or removed before
construction vehicles and welding (which could provide an ignition
source for fuel vapors) would be allowed.
Work on a pipeline buried next to a railroad may be delayed while
debris about and on the railroad is removed by heavy equipment.
Then, because of the weight of the debris and/or the heavy
equipment, the entire pipeline may have to be exposed and inspected
before it is allowed to return to service. Often, power
transmission towers are replaced with temporary towers while repair
work on the damaged tower is performed. However, the use of a
temporary tower may limit the access of pipeline and transportation
lifeline repair crews because of the increased potential for
electrocution if heavy equipment is operated near the temporary
tower. Fires at collocation locations can 
increase the time
required to inspect the nearby lifelines to determine the extent,
if any, of damage caused by the fire. 
Water inundation can cause
delays until the water is drained and the surrounding ground dries
to a condition that allows the repair equipment and material to be
delivered to the damage site. Major damage to a lifeline may
result in a regulatory review about the suitability of rebuilding
(or repairing) the lifeline. While the regulatory review is
underway the repair on the lifeline may be delayed.
In summary, a collocation damage scenario must be developed, 
based
on the knowledge of the lifelines and their anticipated damage
state if they had been isolated or non-collocated. This will
result in the estimation of a new damage state, 
new access times,
or combinations of those items. With the damage state known, a new
repair time is calculated, and the repair time and access time are
used to determine the new time to restore service.
4.4 Interpretation of the Results
This is the activity that brings together all of the previous
analyses.
The most appropriate measure of the impact of lifeline collocation
because of an earthquake was judged to be the most probable
incremental increase in the time to restore the lifeline to its
needed service level. The restoration of service time is 
a broad
measure of the impact of lifeline damage on personnel, equipment,
and material resources, it does not measure the impact that the
44
loss of the lifeline has on the community that was relying upon it.
The difference between the restoration of service time assuming
collocation impacts and the shorter restoration time found by
assuming no collocation impacts gives the incremental impact that
collocation has caused to service restoration. The incremental
time impact is a better measure of collocation impacts as compared
to the estimated total time to restore service, because any biases
in the estimation procedures tend to be canceled by the subtraction
process.
It is important to multiply the incremental change in restoration
time by the probability that the collocation damage has occurred.
This recognizes, the uncertainties in the data base and analysis
methods provided in Section 4.2, and it also recognizes that in
actual earthquakes there is a real probability that a given level
of damage will occur, or conversely, will not occur. The product,
incremental change in restoration time multiplied by probability,
identifies the most probable incremental change in restoration
time.
There are two ways to use the final measure:
1) the most probable incremental change in restoration time
can be considered at a specific collocation site to evaluate 
the impacts at that site. This will provide an insight on the
vulnerabilities that occur when specific types of lifelines
are collocated at at-risk locations. That is, this type of
information will help identify which lifeline types or which
lifeline design or construction practices, when collocated
with other lifelines, lead to the greatest increases to the
other lifelines' level of vulnerability.
2) the most probable incremental change in restoration time 
can be summed along the route of a given lifeline to provide
an insight on the impacts that the specific lifeline route has
had on the vulnerability of the lifeline. This type of
information can be used to help identify undesirable routing
decisions. 
4.5 Chapter 4.0 Bibliography 
1. P. Lowe, C. Scheffey, and P. Lam, Inventory of Lifelines in
the Cajon Pass, California", ITI FEMA CP 120190, August 1991.
2. C. Rojahn and R. Sharpe, "Earthquake Damage Evaluation Data
for California l, ATC-13, 1985.
3. J. Evernden, et. al., "Interpretation of Seismic Intensity
Data", Bulletin of the Seismological Society of America, V 63, 
1973.
45
4. J. Evernden, et. al., "Seismic Intensities of Earthquakes of
Conterminous United States Their Predictions and 
Interpretations", U. S. Geological Survey Professional Paper
1223, 1981.
5. S. Algermissen et. al., "Development of a Technique for the
Rapid Estimation of Earthquake Losses", U.S. Geological Survey
Open File Report 78-441, 1978.
6. Harding Lawson Associates, et. al., "Marina District &
Sullivan Marsh Area Liquefaction Study, San Francisco,
California", Draft report prepared for the City & County of
San Francisco, June 1991.
7. J. Davis, et. al., "Earthquake Planning Scenario for a
Magnitude 8.3 Earthquake on the San Andreas Fault in Southern
California", California Division of Mines and Geology, Special
Publication 60, 1982.
8. S. Barlett and T. Youd, "Case Histories of Lateral Spreads
Caused by the 1964 Alaska Earthquake", US/Japan Case Histories
of Liquefaction. Large Ground Deformation and Effects on
Lifeline Facilities, NCEER, in press.
9. T.D O'Rourke, et. al., "Large Ground Deformation and Their
Effects on Lifeline Facilities: 1906 San Francisco
Earthquake", US/Japan Case Histories of Liquefaction, Large
Ground Deformation and Effects on Lifeline Facilities, NCEER,
in press.
10. D. Keefer and R. Wilson, "Predicting Earthquake-Induced
Landslides, with Emphasis on Arid and Semi-Arid Environments",
Landslides in a Semi-Arid Environment with Emphasis on the
Inland Valleys of Southern California, Edited by P. Sadler and
D. Morton, 1989.
11. P. Sadler and D. Morton, Editors, "Landslides in a Semi-Arid
Environment with Emphasis on the Inland Valleys of Southern
California", Publications of the Inland Geological Society,
Volume 2, 1989.
12. M. Legg, M., J. Slosson, and R. Eguchi, "Seismic Hazard for
Lifeline Vulnerability Analyses", Proceedings of the Third
International Conference on Microzonation, Seattle,
Washington, 1982.
13. 
N. Newmark, "Effects of Earthquakes on Dams and Embankments",
Geotechnique, v.15, No. 2, p. 139-160, 1965.
R. Wilson and D. Keefer, "Predicting Aerial Limits of
Earthquake Induced Landsliding", Evaluating Earthquake Hazards
in the Los Angeles Region, U.S. Geological Survey Professional

Paper 1360, 1985.
15. L. Youd and R. Perkins, "Mapping of Liquefaction Severity
Index", Journal of Geotechnical Engineering, Vol. 113, No. 11,
pp 1374-1392, 1987, and "Mapping Liquefaction Induced Ground 
Failure Potential", Journal of the Geotechnical Engineering
Division, ASCE, 104, GT4, pp 433-446, 1978. 
16. T.D O'Rourke, et. al., "Large Ground Deformation and Their
Effects on Lifeline Facilities: 1971 San Fernando Earthquake",
US/Japan Case Histories of Liguefaction, Larae Ground
Deformation and Effects on Lifeline Facilities, NCEER, in
press.
17. "Seismic Design Procedures and Specifications 1940 to 1968",
CALTRANS Division of Structures, material provided for the
Cajon Pass Lifeline Study, September 1990.
18. B. Maroney and J. Gates, "Seismic Risk Identification & 
Prioritization in the CALTRANS Seismic Retrofit Program",
material provided for the Cajon Pass Lifeline Study, September
1990. 
19. "CALTRANS Seismic Risk Algorithm For Bridge Structures", SASA
Division of Structures, June 30, 1990, material provided for
the Cajon Pass Lifeline Study, September 1990.
20. C. Rojahn, C. Scawthorn, and M. Khater, "Seismic Vulnerability
of Lifelines in the Conterminous United State", ATC-25, in
press.
21. T. Airman, et. al., "Pilot Study on Seismic Vulnerability of
Crude Oil Transmission Systems", NCEER-90-0008, May 1990. 
5.0 
APPLICATION THE METHOD TO CAJON PASS
5.1 
Data Acquisition
The following material demonstrates the application of the analysis 
method described in Section 4.0. The first step of the process is,
to assemble the data base that describes the lifelines and their
routes in the study area as well as the geologic and seismic
situation. The earlier Cajon Pass study) provides most of the
needed information. It should be consulted for specifics about
each lifeline and when it was installed.
Figure 5 shows the Cajon Pass study area and its relationship to
other cities in California. It is used with the permission of the
Automobile Club of Southern California (it is copied from the San
Bernardino County and Las Vegas Area map). It shows that the Cajon
Pass canyon (which is about 10 miles northeast of San Bernardino)
is a natural access route between the San Gabriel Mountains to the
Figure 5, MAP OF THE GENERAL LOCATION.OF THE CAJON PASS STUDY AREA 
west and the San Bernardino Mountains to the east. The Pass
connects the Los Angles Basin in the south to the high desert
regions to the north. The City of San Bernardino is about 10 miles
southeast of the mouth of the Pass.
U.S. Geological Survey quadrangle maps (7.5 minute series
topographic maps published in 1988) were used to obtain more detail
and to develop a plan for a site survey. The site survey was then
conducted. It identified additional lifelines that were not
identified on the 7.5 minute quadrangle maps, which emphasizes the
heed to conduct actual site surveys to validate the published
information on lifeline systems. With the map and site visit 
information as a background, the individual lifeline owners were
contacted and meetings were held with their staff to obtain more
details on the location, capacity, design basis, operating and
maintenance history, and emergency response systems in place for
each lifeline. The Cajon Pass site was revisited to validate our
understanding of the actual siting conditions, and in some cases 
this led to additional visits and discussions with the lifeline
owners to resolve questions. This emphasis on the lifeline data
acquisition and validation is very important, as there are over 100
discrete locations (which include over 250 separate combinations of
collocated lifeline components) in the Cajon Pass study area where
different lifeline components are in close enough proximity that it
was necessary to evaluate their potential for collocation impacts.
Figure 6 is a plot of the communication, electrical power 
transmission, natural gas pipelines, petroleum products pipelines,
railroad, and highway lifelines overlaid upon the U.S. Geological
Survey's quadrangle map of the study area. Figure 6 shows several 
important items. First, the Pass is crowded with the lifelines
traveling in a general north-south orientation through the middle 
of the study area. Second, the lifelines are clearly routed in a 
utility corridor. Since the bed of the Pass varies, from about 0.5
miles near Blue Cut (which is located in about the center of the
figure) to over several miles wide at most other regions, topology
requirements alone would not require the observed congestion. The
conclusion reached was that routing criteria such as aesthetic,
cost, land use, and environmental considerations have had the
controlling impact on the lifeline routing decisions.
There are especially congested areas near the intersection of
Highways I-15 and I-215 in the southeast corner of the study area,
near Blue Cut in the center portion of the study areas, and south
and separately north of the intersection of Highway I-15 and State
Highway 138. In addition, there are crowded areas for several of
the lifeline systems, for example, near the railroad summit of
Cajon Pass, where natural gas pipelines, fiber optic lines, and the
railroads are closely located. Also in the northern portion of the
study area it is crowded where the two petroleum product pipelines
and two fiber optic conduits parallel one set of high voltage power
lines and also along Baldy Mesa Road where the two petroleum
Figure 6, A COMPOSITE OF THE LIFELINE ROUTES AT CAJON PASS SCALE 
Larger Scale Figure
Located at 
EXPLANATION 
End of Document 
FLUME-PENSTOCK 50 
product pipelines, the fiber optic conduits, and a natural gas
pipeline all are routed alongside of the road bed. The two
petroleum product pipelines, a natural gas pipeline, and two high
voltage power lines cross the San Andreas fault in Lone Pine Canyon
at approximately the same region. The unfortunate routing for
several miles of the petroleum products pipelines along the San
Andreas fault's rift zone does not enter into the current study
since there are no collocated lifelines of interest along that
route. Finally, there are collocated railroad lines, power lines,
a natural gas pipeline, the petroleum products pipelines, and the
fiber optic conduits parallel to I-15 between the I-15/I-215
interchange and Blue Cut.
Figure 7 is another composite map of Cajon Pass. Each of the 101 
collocations that were analyzed during this study are shown on this 
figure. Within those 101 locations, over 250 individual
collocations occurred. This emphasizes how siting decisions have
resulted in crowded collocation conditions, even though there is
sufficient space to avoid most of them. Although there are several
broad grouping of lifeline intersections, it is clear that they
occur throughout the entire length of the study area.
The seismic and geologic information was also obtained during the
data acquisition phase of the study. A sensitivity evaluation of
six postulated earthquake events was performed to guide the
selection of the event for use in the study. Other 2 ,3) studies were
consulted to help select the earthquake events. The six events
were:
1) The 1857 Ft. Tejon earthquake on the San Andreas fault.
This was 300 km long fault with a magnitude 8.3 earthquake,
and with the southern edge of the surface displacement located
just north and west of Blue Cut.
2) An earthquake on the southern segment of the San Andreas
fault. This was a 200 km long fault of 7.8 magnitude. The
northern edge of the surface displacement was placed just
north and west of Blue Cut.
3) An earthquake similar to event 1, except that the southern
extreme of the surface displacement was moved about five mile
further east into the study region.
4) An earthquake similar to event 1, except that the length of
the fault was reduced to 105 km. This resulted in a 7.7
magnitude earthquake.
5) An earthquake similar to event 1, except that it was
centered about the Cajon Pass. This resulted in a 8.3
magnitude earthquake.
6) A earthquake of 94 km length, but placed on the San Jacinto 
fault. This resulted in a 7.5 magnitude earthquake.
Figure 7, IDENTIFICATION OF LIFELINE COLLOCATION AT CAJON PASS
SCALE
0 1 2 MIZS 
2 K2I WMKDITFUS j LargerScale Figure 
Located at 
EXPLANATION
End of Document 
1-15-INTERSTATE. 
2 PAVED HIGHWAY 
RAII.ROAD 
0 -P POWER UINS 
PETROLEUM PRODUCT PIPEUINES
I -NATURAL 
GAS PIPEUNES 
FIBER OPTIC CABLE 
If -v 
FLUME- 52
The sensitivity study was performed with the QUAK2NW3 computer
code 4,5) developed by the U.S. Geological Survey. Based on the 
study, the ground shaking intensities were relatively insensitive
to the changes in fault rupture length The conclusions reached was
that the 1857 Ft. Tejon earthquake was a reasonable choice for the
study. It did not produce the most intense shaking, nor was it the
least intensive. However, by using it for the present study, it
will be possible to compare our solutions for earthquake intensity
6K. That comparison showed
with those of previous researchers 5,
general agreement except at the fault rift zone. There the
QUAK2NW3 program predicted lower shaking intensities than those
reported by DlavisL6). After discussions with Davis, it was decided
to increase the predicted MMI shaking intensity along the San
Andreas fault zone by one level from VIII to IX. This accounts for
the greater impacts that are expected to be associated with the
fault displacement and is consistent with the work of Davis.
The areas of potential liquefaction were determined by examining
the water well data for the Cajon Pass, and supplementing it with
other regions high water table as determined by the site
reconnaissance visits. Regions of high water table were correlated
to alluvial deposits to identify the liquefaction susceptible
7)and the
regions. The historical landslides were identified 6.
method of Legg (see Section 4.2, Table 7) was applied. A computer-
based check of the soil conditions at the Cajon Pass was used to
assure that the Legg method was applied at each slope of interest.
The landslide predictions based on the Legg model agreed quite well
with the record of historical landslides (that is, the Legg model
prediction included the historical landslides, but it also
identified many more potential areas of landslide).
Figure 8 presents the summary of the calculated seismic and 
geologic conditions overlaid upon the lifeline routes. Although
the figure is complex and filled with data, it does highlight some
important information. In the figure the shaking intensities are
shown with various levels of shading. The highest intensities, MSI = IX, are along the San Andreas fault rift zone. On the map they 
are shown as solid lines where the fault is well located, dashed
where its location is estimated, and circled when it is hidden by
younger rocks. The potential landslides are predominantly south of
the San Andreas fault and lie in a southeast trend. There are four 
important regions of potential liquefaction: just south of Cajon 
Junction, at Blue Cut where they coincide with potential landslide
regions, southeast of Blue Cut about two miles northeast of the I-
15/I-215 intersection, and just south of the 1-15/I-215
intersection.
Figure 8 shows that many of the conditions of high M4I value,
landslide, and liquefaction overlap. This is important to note
because the lifeline components in the study area >(with the
exception of some bridges) are not very sensitive to shaking
damage. MMI values of VIII generally would only cause damage state
2 or 3 to occur (less than 5% damage expected). At those
conditions, collocation induced additional damage would be slight
or none at all. However, the facilities are very sensitive to
ground movement and there are a number of potential landslide or
lateral spread (or liquefaction) locations along the lifeline
routes that coincide with MMI = VIII. North of the San Andreas 
fault the earthquake risk comes predominantly from shaking, but
most of the peak shaking intensities are VII with a few isolated
locations of VIII.
Another important activity of the data acquisition phase of the
study is to divide each lifeline into segments for subsequent
analysis. Figures 6 and 7 were used to guide the segmentation 
process. The communication lifelines were divided into 12 to 14
segments, depending on the route of the individual fiber optic
conduits; the electrical power transmission lifelines were divided
into four segments (the Arrowhead Calelectric-Shannin line at the
southeast corner of the study area was analyzed as a single
segment); the natural gas pipeline lifelines were divided into five
or eight segments; the petroleum products pipeline lifelines were
divided into 10 segments; the railroad lifelines required 30
segments; and the interstate highway required 36 segments. The 
subsequent vulnerability calculations for the communication,
electrical, and fuel pipeline lifelines were performed by hand;
those for the transportation lifelines were performed with a
standard computer spreadsheet. These approaches recognized the
number of calculations that would be required.
5.2 Lifeline Collocation Vulnerability Analysis Results
Figure 7 identifies the locations of the 101 collocations that were 
subsequently analyzed in this study. These collocations involved
over 250 separate potential lifeline interactions. Table 15 was
prepared to identify the collocation and the lifelines that were
involved at that location. As the collocation evaluation was
prepared, the results were tabulated against the index, thus
assuring that all potential interactions were located and
evaluated.
Table 15 and Figure 7 identify several critical clusters of
interactions. With this in mind, a collocation damage scenario was
developed for each critical cluster of interaction. Using a 
standard collocation damage scenario at the critical clusters
helped assure the overall consistency of the interaction analyses4
The clusters where the standard collocation damage scenario was
used were:
1. The liquefaction zone south of the interchange of I-15 and
I-215. There were 10 separate potential interactions
involving I-15, railroad bridges, several fiber optic lines,
the 16-inch natural gas pipeline, and the 8-inch and 14-inch
55
INTERSECTION 
NO. 
1 
2 
3 
4 
5 
71 
10 
23 
11 
12
10X 
13 
14 
16 
17 
is 
19 
24 
25 
26 
2 7 
28 
29 
Table 15 
MATRIX OF LIFELINE COLLOCATIONS AND INTERACTIONS 
HIGH-RAIL-POWER FIBER NATURAL PETRO 
WAYS ROADS LINES OPTIC GAS LINES 
A B C D E A B C A B C D E F A B C D A B C A B 
X x x 
X X 
X X X 
X X X 
X X XX X 
X X XX
X X X 
X
x X 
X 

X 
X X 
x 
X X X X X X X 
X X 
X X X 
X 
x xX 
=X-X l X = XX 
X X 
= : x I XXXXx 
X X XX XXX 
X X X X 
I XI X X X 
Table 15 (Continued) 
MATRIX OF LIFELINE COLLOCATIONS AND INTERACTIONS 
FIBER NATURAL PETRO
INTER-HIGH-RAIL-POWER 
OPTIC GAS LINES
SECTION WAYS ROADS LINES 
A B
NO. A B CD E A B C A B C D E F A B C D A B C 
30 X X
31 X X 
32 
xx xx x
33 X X 
34 X ,X 
35 
x x H 
XX*36 XX XX x 
37 X X XX 
38 .XX xx 
3S XXXX XX 
40 X XX X XXXX 
41 X xX XX X X X X 
42 XX x
43 X XX 
44 X X XX i XX XXXX 
X45 X X 
46 X 
XXXX X
47 
I X X
48 X X X X
49i 
X X X .s5 
51 xx 
52 . X X XX 
53 X XX 
54 X X X
55 X X XXXX 
56 x xxxx xx 
57 X X X
58 X X XXXX XX 
Table 15 (Continued) 
MATRIX OF LIFELINE COLLOCATIONS AND INTERACTIONS 
INTER-HIGH-RAIL-. POWER FIBER NATURAL PETRO 
SECTION WAYS ROADS LINES OPTIC GAS LINES 
NO. A B C D E A B C A B C D E F A B C D A B C A 3 
59 X X X X X 
601 X X xXX X X 
62 x x x xxx X 
63 X
64 X X 
65 X X 
66 X XX 
67 XXX XX 
68 X X 
69 X X 
70 x x 
71 = X 
72s x x 
73 xxx x x 
74 x x x xx xx 
75 1 I 
76 x x x x 
77 81_X' X X X 
7 8 I 
E l2 x x x xx 
84 .x x x x 
86 X XXX 
87 X X X X 
58
Table 15 (Continued)
MATRIX OF LIFELINE COLLOCATIONS AND INTERACTIONS
INTER-HIGH-RAIL-POWER FIBER NATURAL PETRO 
SECTICON WAYS ROADS LINES OPTIC GAS LINES 
NO. A S C D E A B C A 8 C D E F A 3 C D A BC A B 
8X X X XX 
89 HX XX X XX 
90 X X X 
NOT USEDX 
93 H 
NOT USED = X XXX 
96 
9397 _X 
X 
HX I 
XX 
X X XX 
98 X 
99 X 
100 X X 
101 XX 
X X 
103 X X: 
petroleum products pipelines. All of the buried lifelines
were found to have incurred damage state 7 (catastrophic) with
The assumed
probabilities that the damage occurred of 40%. 
collocation interaction was that the petroleum pipelines could
drain 1-2 miles of pipe but that no secondary fires or
explosions would occur. This contributed to an additional 30
day delay to the site before repair could commence, due to the
requirements to assure that fire conditions and environmental
concerns could be alleviated before work could start on the
individual lifelines. An additional 10 days of delay were
hypothesized due to the need to coordinate the work on so many
individual lifelines.
2. A second cluster exists along the Cajon Blvd. extension
into the Cajon Pass from north of the I-151I-215 interchange
to Blue Cut. There are two separate locations where
landslides (with a probability of occurrence of 45%) 
and two liquefaction areas (probability of occurrence of 20% and 40%)
are possible, including eight separate collocation impact
areas where there are collocated power lines, railroads, and a
natural gas pipeline. At the two landslide areas along the
Cajon Blvd. extension (which was the prior highway 66 before
I-15 was built), a natural gas pipeline and the railroads are
located at the toe of the slide area, and landslide debris is
expected there. Debris removal for clearing the railroad
would be required before work on the pipeline (which is
located in the railroad right-of-way and sometimes under the
rail bed) could begin. The debris removal was assumed to
cause a 30 day delay before work on repair of the pipeline
could occur.
3. At Blue Cut itself there is another landslide and
liquefaction zone. Due to its proximity to the San Andreas
fault rift zone, a 70% probability of occurrence was
estimated. At that location, a power line, a natural gas
pipeline, and a railroad are in the flow area of the slide,
which could cause the pipeline to surface and rupture as well
as to be covered with debris. The collocation damage scenario
assumed that an explosion and fire could result, increasing
the damage to the power line and its repair time by an
additional 20 days, compared to the delays described for the
other slide area, to repair the more extensive damage the fire
caused. The potential for landslides blocking the Cajon Creek
with a slide dam, and the subsequent impact on downstream
lifelines if the dam should catastrophically fail were
considered to be outside the scope of this study. 
At Blue Cut itself, the liquefaction zone has a 50% probability of
occurrence.
4. In the San Andreas fault rift zone there are six
collocation points that involve the power lines, 
the railroads, the fiber optics, a natural gas pipeline, the two
petroleum products pipelines, and the Cajon Blvd. extension.
Fuel spills are assumed to require a 30 day delay for
alleviating environmental and fire concerns. An additional 60
day delay was assumed for the petroleum products pipelines
because the estimated extensive damage to their right-of-way
along the fault trace will require regulatory review and
acceptance before the pipeline can be worked on. 
An additional 30 day delay for the other lifelines was 
Assumed because of the general congestion in the area. Since the
fault displacement causes catastrophic failure of the
lifelines, the collocation damage scenario does not assume any
further damage.
With these scenarios in mind, the collocation vulnerability
analysis was performed for each separate lifeline system.
Communication Lifelines
Figure 9 shows the communication lifeline routes in the study area. 
The location of the photographs presented in this section of the
report are also shown on the figure. The microwave, radio, and
cellular phone communication towers are sited such that they are
not collocated with other lifelines. Thus, they do not enter into
the analysis of the impact of collocation. The impact of degraded
communications (if these towers should fail) on the ability to 
restore the other lifeline systems to acceptable delivery
conditions is beyond the scope of the present study..
There are five fiber optic systems located in the study area. They
include American Telephone and Telegraph (AT&T), Continental
Telephone (Contel), MCI Communications (MCI), WTG West (WTG,
formerly WilTel), and US Sprint.
The individual fiber optic cables are multi-layered with an inner
structure that allows the cable to be pulled and maintained in a
state of tension without placing tension on the individual glass
fibers. This assembly is then wrapped with various insulating
materials, including a metal sheath. In the fall of 1986, the U.S. 
Forest Service provided MCI and WTG right-of-ways on the basis that
they would each provide two conduits and that each conduit would be
four inches in diameter so that cables from two different systems
could be placed in each individual conduit (thus, provisions were
made to lay eight cable systems along the two routes through the
Forest Service land). Furthermore the Forest Service required that
the routes be combined as quickly as, practical. Thus, the MCI and
Contel systems enter the study area from the north along Baldy Mesa
Road, while the AT&T, WTG, and US Sprint systems enter from the
north along the access road to I-15. The routes join together just
south of the separation of I-15N and 1-15S (about 1.8 miles north
of the Cajon Junction of I-15 and Highway 138). From there they
travel together as a bundle of four conduits. Much of their route
is along the Cajon Blvd. extension where they are laid in the
median strip. Also routed along much of the median strip are the 
two petroleum products pipelines (the Cajon Blvd. extension was the
former divided highway 66, but only the western two lanes are still
maintained for traffic). For the purposes of this study, the fiber
optic cables are analyzed as buried conduits. Because of their
collocation, if one conduit fails, all fail.
When the conduits, (which are normally buried) are routed to a
bridge location, they are generally brought to the surface and hung
with light anchors from the bridge. Figure 10 shows them on a 
typical bridge crossing on the abandoned portion of Highway 66;
Figure 11 shows some of the details of the bridge hangers and the
conditions of the conduits; Figure 12 shows them hung from a
highway culvert wall just south of Cajon Junction; Figure 13 shows
some of the details of the wall anchors, near the culvert location
of Figure 12. Just south of the culvert shown in Figure 12 the
II COMMUNICATIONLIFELINEROUTS.
Figure 9 COMMUNICATION LIFELINE ROUTES 
SCALE 
Larger Scale Figure
Located at 
EXPLANATION 
End of Document 
Figure 10 Fiber Optic Conduits On A Bridge On The 
Abandoned Portion of Cajon Blvd. Extension 
Figure 11 Details of the Fiber Optic Conduits of Figure 10
Figure 12 Fiber Optic Conduits On A Concrete Culvert 
That Passes Under 1-15 
Figure 13 Wall Support Details For The Fiber Optic 
Conduits Of Figure 12 
Figure 14 Surface Water Conditions Near The Toe Of The 
1-15s Retaining Crib Wall 
Figure 15 Fiber Optic Conduits Under a Heavy Water 
Pipe, Both On A Highway Bridge 
fiber optic cables are buried at the toe of a crib wall used to
support the southbound alignment of 1-15. Figure 14 shows the
surface water in this region, which was identified as a potential
liquefaction region. Figure 15 shows the conduits hung under a
water distribution system pipe, both of which are supported from a
road bridge over a railroad. Other important collocations include
where the conduits are buried near buried fuel pipelines, a number
of which are in liquefaction or landslide prone regions.
In all of these cases, the fiber optic cables are at potential risk
because the light anchors used could be expected to fail due to
shaking. The heavy water pipe just above the fiber optic cable
conduit can be expected to fall on and fail those cables.
The analysis of the fiber optic cable systems indicates that they
are not at significant risk until the MMI values equal VIII or
more. In the Cajon Pass that occurs at liquefaction areas or at
the San Andreas fault where surface displacements of up to 12
meters are anticipated. Thus, shaking damage is not expected to be
significant compared to displacement-related damage. However, the
screening method data base does not directly account for the light
anchors and the conditions of the conduits as shown in the figures.
It is possible that they are more sensitive to shaking damage than
the screening method predicts.
At the San Andreas fault location (just north of and close to Blue
Cut) the fiber optic conduits are not collocated with other
lifelines within the zone of influence of the lifeline. Thus, for
purposes of this study there is no collocation impact there.
However, the large number of near by lifelines does suggest that
general construction congestion could delay the permanent repair of
the communication lifelines in this region. Also, it probably
would be possible to temporarily lay the fiber conduits on the
ground surface, restoring service on a temporary basis. These
types of changes to the assumed restoration of service time are
noted, but they were not used in the present study to estimate the
service restoration time.
Four liquefaction regions have the potential to impact the fiber
optic systems. At the liquefaction region just southeast of the I-
15/I-215 interchange, the fiber optic cables are in separate and
dispersed conduits. However, two cable conduits cross this zone
near the railroad beds. One cable conduit crosses the 16-inch
natural gas pipeline and the 8-inch petroleum products pipeline,
and then it runs parallel to the 8-inch line. Another cable conduit
crosses the liquefaction zone and is perpendicular to the 16-inch
natural gas pipeline, which is also in the liquefaction zone. The
liquefaction impact is calculated as a damage state 7
(catastrophic) but it only has a 40% probability of occurring. The
conduit repair time in this region is hypothesized to triple, based
on the delays required to repair the natural gas and the petroleum
products pipelines and the delays associated with the repair of the
nearby damaged I-15 and railroad bridges. This causes the fiber
optic conduits' most probable repair time to increase by about ten
days.
In the second liquefaction region the four cable conduits are
collocated and they are parallel and next to the 8-inch and 14-inch
petroleum products pipelines. In this region they cross over the
36-inch natural gas pipeline. The liquefaction impact is a damage
state 7 with a 20% probability of occurrence. The collocation 
damage scenario assumed that the petroleum product pipelines would
have to be repaired before the fiber optic cables could be replaced
within their conduits along their old route, causing a 55 day delay
in being able to access the five fiber optic cables. Another 10
days delay for equipment availability is expected. However, the
low probability of liquefaction resulted in the most probable
restoration time increasing by only about four days.
The third liquefaction zone is just south of Blue Cut. The
collocated lifelines include the fiber optic conduits, fuel
pipelines and a high voltage power line also crosses over the
region. The liquefaction probability at this locations is
estimated at 50%. Most of the repair activities for the power line
will not impact the fiber optic cable conduit repair. The only
impact could come when the power lines directly over the conduits
are being worked on. The most probable restoration time increased
by only about 12 days.
The fourth liquefaction zone is where the fiber optic conduits
cross the toe of an I-15 retaining wall crib and then are connected
along the concrete culvert under I-15 (see Figures 13 and 14). The
concrete culvert is a massive structure, and its damage is expected
to be small. Analysis of the crib wall found that its movement
would not substantially impact the highway, however, just partial
movement of the crib wall could severely damage the fiber optic
cables. In such a case they could not be replaced permanently
until the wall had been stabilized. Fortunately, the probability
of liquefaction is only 40% in this region, resulting in the 140
day crib wall-induced delay becoming a 22 day most probable
restoration time increase for the fiber optic conduits.
Where the AT&T and WTG cables are hung from the bridge over the
Southern Pacific railroad is the final collocation damage location
where the collocation impacts were found to be serious (see Figure
15). There, because the shaking intensity is only MMI=VII, not
much damage would normally be expected to the cable conduit itself.
However, the wall brace and anchor supports for the water pipe and
the fiber optic conduit are small and have not been sized for
earthquake conditions. Thus, it was assumed that they fail
allowing the fiber optic conduit to sag. The heavier water pipe
was assumed to fall on top of the fiber optic conduit which is
located directly under it, causing the fiber optic conduit to
rupture and fail. The probability of this, failure scenario is 
estimated to be 80%, causing the most probable restoration time to
increase by about 4 days.
Thus, of the 55 collocations analyzed for the fiber optic systems,
nine were estimated to lead to increased probable times to restore
service. Most of these conditions resulted from ground motion-
induced failures, and the impact on the fiber optic systems was
that the failures of the other collocated lifelines lead to
increases in the delay time before the fiber optic systems could be
repaired. The practice of collocating all of the fiber optic
conduits together, along with the practice of hanging them from
bridges and culverts with very light anchor bolts, suggests that in
future earthquake situations the loss of telephone communications
will be more severe than have been experienced in the past. That
is because the loss of a few hard-wired telephone lines in past
earthquakes has not been significant in terms of the ability of the
systems to handle call traffic. Fiber optic cables, however,
handle many more calls per line than does a hard-wired system, and
if one cable is lost then probably all of the collocated cables
will be lost in the same event.
The overall estimate of the impact of collocation on the
communication lifelines was:
Increase in Probable Increase in Probable 
Time to Restore Time to Restore 
Lifeline Service, days Service, % 
Fiber Optic 61 86 
Cables 
Electric Power Lifelines
Figure 16 shows the electric power transmission lifeline routes in
the study area. The location of the photographs presented in this
section of the report are also shown on the figure. Experience has
shown that power transmission towers are quite resistant to
earthquake shaking, principally because of the conservative wind
loading criteria used in their design. Thus, only fault
displacement, landslide, or liquefaction are expected to caused
significant levels of damage to the towers.
The electric power lifeline systems include a major transmission
system substation at Lugo in the northeast corner of the study
area, a hydroelectric generation station in Lytle Creek Canyon, and
four major high voltage transmission systems. The hydroelectric
station is not collocated with any of the lifeline components of
interest to this study. Although the substation is collocated with
two of the high voltage transmission systems, component failures in
the substation are not expected to lead to transmission line
failures, and visa versa. The transmission lines are not expected
to have any failures at the shaking intensity expected at Lugo
Figure 16, Electric POWER LIFELINE ROUTES 
SCALE Larger Scale Figure
Located at
End of Document 
EXPLANATION 
 (where MMI = VII). Thus, although the substation is a potential 
weak link in the overall electric power transmission system, the 
lack of collocation impacts means it was not be examined further in 
the present study. 
In general, power transmission lines are not impacted directly by 
the other lifelines, as they are either above or otherwise outside 
of the zone of influence of the other lifelines. However, they 
can be impacted by construction delays or fires induced by the 
failure of other lifelines. The general tower design and footings 
used are quite rugged and earthquake resistant (resistant to 
shaking damage). There have been some cases when shaking has 
caused the 
lines 
themselves to 
gallop or 
vibrate,
resulting in 
their coming 
close or even 
touching
other lines 
routed on the 
same tower. 
The resulting 
arc and 
electrical 
short can 
cause fires 
and/or drop 
both lines 
from service. 
However, this 
failure mode 
is not 
addressed in 
the available Figure 17 A Landslide Scar With Power Towers In The 
data base and Slide Area 
thus was not 
considered in 
the present study. 
Transmission lines often traverse more rugged areas, and as such 
they are susceptible to landslide damage. Figure 17 shows two 
transmission towers located in an old landslide scar (the original 
towers were damaged in the landslide). This location is an 
important collocation site, as a buried natural gas transmission 
pipeline and the railroad tracts are located just below the slide 
area. Figure 18 shows the location of two high voltage power 
transmission tower systems, a buried natural gas transmission 
pipeline (shown by a surface marker), and the 8-and 14-inch buried 
petroleum products pipelines (also shown by a surface marker). 
This location 
is in the 
fault rift 
zone of the 
San Andreas 
fault ( in 
Lone Pine 
Canyon).
Figure 19 is 
a photograph 
from the same 
location 
looking in 
the opposite 
direction. 
It shows a 
transmission 
tower located 
at the edge 
of a steep 
ravine where 
it is subject 
to possible Figure 18 Power Lines, Natural Gas & Petroleum
landslide Products Pipelines Intersection Over the San Andreas 
failure. 
This general Fault Rift Zone 
location (in 
the San Andreas fault and rift zone) is the most significant 
collocation condition for the electric and the fuel lifelines. 
The Southern California Edison (SCE) 115 kV Arrowhead: Calelectric 
Shannin transmission line (located in the southeast corner of the 
study area) is routed through some local landslide and liquefaction 
zones, however, it is not collocated at those regions. Its other 
collocation impacts were negligible. In the northern section of 
the study area the SCE Lugo-Vincent two-line, 500 kV transmission 
lines traverse from Lug0 station west and then northwest. Because 
of the low shaking intensity (low for power lines), they only 
experience damage state 1 or 2. Thus, these two SCE transmission 
systems are not of interest for collocation impacts in the present 
study. There are, however, a Los Angeles Department of Water and 
Power (LADWP) two-line transmission system and a three-line SCE 
transmission line system that resulted in collocation impacts. 
The LADWP Victorvil1e:Century 287.5 kV transmission line system (it 
has two full circuits) extends from the north (located at about the 
center of the northern boundary of the study area) to the south of 
the study region. It was constructed in 1936 to transmit power 
from the Hoover Dam in Nevada to the Los Angles Basin. It has been 
upgraded in 1970, 1974 and 1980 to allow switching between the 
287.5 and other 500 kV lines as well as to add new controls. Parts 
of the line have previously experience problems of interest to this 
study. At a section between
Highway 138 and the Lone
Pine Canyon, the lines
experienced slow foundation
movement or shifting at a '7 k 
region were a local bowl or I I' 
depression exists. The V; , 
cause was determined to be a
high water table fed by
surface waters collected in
I s 
the bowl that allowed the
tower foundations to slowly
respond to the tension
imposed by the lines
themselves. The solution
was found to be to cover the
ground with a concrete
mixture so that surface
waters would could not seep
into the water table at that
location. The towers now
appear to be stable. On
several other occasions,
local brush/forest fires
have heated the copper
conductors to the extent
that they partially annealed
and sagged. This problem
was resolved by retensioning
the line in those regions.
It is of importance because
it indicates the problems Figure 19 Power Tower At A Ravine Edge
that earthquake-induced
fires could have on this In The San Andreas Fault Rift Zone
lifeline system.
The LADWP power lines cross the San Andreas fault zone in Lone Pine
Canyon very close to where the petroleum products and a 36-inch
natural gas transmission line cross the fault trace (see Figures 18
and 19). The expected 12 meter fault displacement causes a damage
state 7 to all the lifeline components in this region, with a
probability of 100%. The collocation scenario assumes that the
resulting petroleum products spill (several miles of those
pipelines could drain from the rupture, depending on how many other
ruptures are assumed since the petroleum products pipelines run
parallel to the fault trace for several miles) causes a 30-day
delay in the repair activities while the resulting environmental
and fire hazards are evaluated and mitigated. The general
congestion in the area and the need to coordinate the use of heavy
equipment so that its use will not adversely impact the other
lifelines is assumed to add another seven days to the power line
repair times. Because of the high probability of damage at this
location, the probable restoration time increases by 37 days. This
accounts for about 75% of the collocation-induced delays in the
restoration of service to the LADWP power lines.
The other significant collocation region for the LADWP power lines 
is also at Blue Cut. The power lines cross the eastern edge of a
landslide zone, and at the toe of the zone at a lower elevation a 
36-inch natural gas pipeline is sited next to and in the right-of-
way of a rail line. The probability of a slide in this region is
70%, it produces a power line damage state of 5 (heavy damage).
The collocation scenario includes a seven day period while the gas 
line is prepared and tested for leaks, and a increase in damage
state to level 6 because of potential natural gas pipeline failure 
impacts. That scenario increased the repair time from 49 to 66 
days and the net impact on the change in restoration time was about
12 days.
There are three 500 kV circuits on SCE's Lugo -Mira Loma system. 
They were installed in the early 1960s for about 300 kV service and
Line 3 was added in
upgraded to 500 kV service in the early 1970s. 
1983.
Line 1 is to the west, line 2 is in the center, and line 3 is the
eastern line. Line 1 is routed south by southwest from the Lugo
substation. Lines 2 & 3 leave Lugo station on a single tower 
system for about 1.5 miles, then they divide into two separate
tower systems. The power lines cross the railroad lines a short
distance before they cross Highway 138. From there they head
generally south until they cross I-15. In this high desert region
the only earthquake load comes from shaking characterized by MMI = 
VII. At that intensity there is no appreciable damage to the
towers or the lines, and no collocation impacts were hypothesized.
At the I-15 crossing, lines 1 & 2 cross at the northern boundary of 
a local liquefaction zone. Because their power towers are located 
on local hill tops in this region, it is assumed that there are no
impacts due to the liquefaction or due to collocation. However,
they cross near a concrete culvert that crosses under the highway.
At the culvert (see Figure 12) location there is also a metal crib
wall (see Figure 13) that provides support to the road bed, the 
fiber optic cables cross at the same location, there are two
railroad bridges in the downstream path from the run off that
passes through the culvert, and there is a 36-inch natural gas line
which crosses I-15 in the same area. In addition, the crossing is
in an area of high water table and of surface water, indicating a 
potential liquefaction zone. Although the power towers and lines
are not expected to experience damage at this location, they will
cause some delays in responding to damage on the other lifelines
and other equipment
because of the need to work with large cranes 
that could get close enough to the power lines to cause the need
for caution to avoid potentials for electrocutions, etc... Line 3
crosses I-15 further south and is not impacted by the crowded
conditions described above.
After crossing I-15, lines 1 & 2 are routed as parallel lines.
Where they cross the San Andreas fault they will experience damage
state 7 due to the 12 meter displacement expected at the fault
trace. The two lines pose a collocation potential that is assumed
to increase their repair time by 15 days. This increases their
probable restoration time by 15 days because of the high
probability that the large fault displacement will occur. After
this, the lines again separate, with line 1 heading south, and line
2 heading southeast until it joins with line 3 at Blue Cut.
As line 1 heads south it enters a liquefaction zone that is north
and abuts against a landslide zone. Within the liquefaction zone
it crosses over the two petroleum product pipelines, the fiber
optic conduits, and Cajon Blvd. extension. Near the boundary
between the liquefaction and landslide zones it crosses over a 36inch 
natural gas pipeline that is itself next to and in the right-
of-way of a railroad line. The liquefaction zone results in a 50%
probability of damage state 7 occurring to the power towers. The
collocation scenario is a 20 day delay due to the general
congestion in the region, which results in a probable delay in
restoration of 4 days. There are no other significant collocation
regions for line 1 further south along its route.
After they join together, lines 2 & 3 head in a south by southeast-
direction. They cross the Cajon Blvd. extension at the northern
boundary of a local liquefaction zone and head up the steep slopes
to the higher elevations of the San Gabriel Mountains. Just above
the Cajon Pass floor as they rise into the mountains they enter a
landslide zone. Figure 17 shows that in the past they have
experienced slides that have required extensive repair. At the toe
of the slide a 36-inch natural gas pipeline is located next to and
in the right-of-way of a railroad bed. The landslide causes a
damage state 5 (with a probability of 45%). It is assumed that the
congestion and the need to shut down the power line when the gas
pipeline is to be tested for leaks will add 40 days to the repair
of the power towers. This makes their probable restoration time
increase by nine days at this location.
The typical collocation damage scenario is that other lifelines
have a minor physical impact on power lines because the power lines
are above or removed from the zone of influence of the other
lifelines. However, when fuel-based lifelines are involved, they
can cause important delays in the power line restoration. This is
to assure that the power lines do not become a source of ignition
for the fuel. Also, when the other lifelines are directly under
the power lines, the expected use of temporary support towers may
not be acceptable because of the increased risk of electrocution
when other large repair equipment is operated near a temporary
tower. As was the case for the communication lifelines, ground
movement was the principal cause of electrical transmission
lifeline damage. 
One hundred and four collocations involving the electric power 
lifelines were analyzed. The overall estimate of the impact of 
collocation on the electric power lifelines in the Cajon Pass was: 
Increase in Probable Increase in Probable 
Time to Restore Time to Restore 
Lifeline Service. days Service. % 
Los Angeles Dept. 
of Water & Power 
49 28 
Southern California 19 10 
Edison Co. Line #1 
Southern California 28 13 
Edison Co. Line #2&3 
Fuel Pipeline Lifelines 
Figure 20 shows the fuel pipeline lifeline routes in the study 
area. The location of the photographs presented in this section of 
the report are also shown on the figure. The lifelines include one 
8-and one 16-inch petroleum products pipelines, three 36-and one 
16-inch natural gas pipelines, and the associated valves for each 
line. Modern 
buried 
pipelines of 
the type 
installed at 
Cajon Pass 
are very 
resistant to 
shaking
damage.
Thus, the 
earthquake
conditions of 
most interest 
for the 
pipelines are 
the 
conditions 
where ground 
movement is 
expected.
However, when 
they are 
buried next 
to or under Figure 21 Typical Long Open Span On The 36-Inch 
another Natural Gas Pipeline 
Figure 20, FUEL PIPELINE LIFELINE ROUTES 
SCALE 
LargerScale 
figure
EXPLANATION Located at 
End of Document
1-15 INTIESAh 
a PAID SUBWAY 
-…-0-
0 -PZTUOLIM PRODUCT 76 
-0-NATURAL
DU PhP 
0 0 0ALM 
1 4 -F i 1 i ran f = a 
they are with
the
railroads),
there is an
additional
concern that
heavy
equipment
used to
remove debris
or otherwise
restore the
railroad
lifeline
could lead to
pipe wall
damage of the
nearby buried
fuel
pipeline.
Because they Figure 22 Two Natural Gas Pipelines Near Highway 138,
are buried, One Buried, One Exposed
there are no
convenient
features that can be shown with photographs. However, there are 18
locations along the route of the western most 36-inch natural gas
pipeline that are exposed spans which range from 10 to 138 feet in
length. Figure 21 shows one of the longer spans across a dry creek
bed north of Highway 138 and east of I-15. Figure 22 shows a
location about 40 feet from the recently realigned Highway 138. At
that location two 36-inch natural gas pipelines are routed parallel
to each other. The line on the left is exposed, the one on the
right is buried and marked with the surface sign. In general, open
spans are estimated to increase the potential shaking damage state
by up to 1 level, compared to a buried pipeline damage state
condition.
When the pipelines cross under railroads, roadbeds, and sometimes
power line rights-of-way, many of the right-of-way owners required
the pipeline to be cased inside a larger pipe in the belief that
this adds safety and/or it reduces the lifeline interactions if
damage on one lifeline should occur (there is a current technical
question whether the casing increases or reduces the overall safety
of the crossing). Figure 23 shows a location where a 36-inch
natural gas transmission line crosses under two different rail
beds. One railroad requires the use of the extra casing, the other
does not. If the use of a casing increases the safety of the
crossing, the close proximity of the railroads means that the extra
benefit is lost. If the use of a casing decreases the safety of
the crossing, then the railroad expecting extra safety has not
77
achieved it because of the close proximity in this situation. 
Two of the most critical collocations involving fuel pipelines were 
shown in Figures 17 and 18. In Figure 17 the pipeline and the 
railroads are located directly below a landslide prone area that 
includes electrical transmission towers. Figure 18 shows that both 
types of fuel pipelines and high voltage electrical transmission 
systems all cross the San Andreas fault rift zone at approximately 
the same location. The results of these conditions are discussed 
below. 
The western most 36-inch natural gas pipeline from valve station 14 
(near the Cajon railroad summit east of the I-lS/Highway 138 
interchange) south to the end of the study area was installed in 
1960. The eastern most 36-inch natural gas pipeline was installed 
in 1966. They operate at a mean operating pressure of 845 psig. 
The 16-inch distribution supply line branches off of those lines at 
Valve 15 (south and west of 
the 1-15/1-215 interchange) 
and provides service to San 
Bernardino. It operates at 
a mean operating pressure of 
350 psig. The line from the 
north of the study area 
(along Baldy Mesa Road) to 
valve station 14 was 
installed in 1976. It 
operates at a mean operating 
pressure of 936 psig. All 
of the lines were arc welded 
and constructed from high 
grade steel pipe. They
deliver 0.6-1.0 billion 
cubic feet of natural gas 
per day. 
The 8-inch petroleum 
products pipeline was 
installed in 1960 and the 
14-inch line was installed 
in 1969. In 1980, several 
miles of the 8-inch line 
were abandoned and a new 
line was routed from the 
west side of Cajon Creek to 
join the 14-inch line on the 
east side. On the east side 
of Cajon Creek they are 
routed along the Cajon Blvd. 
extension. The lines 
transport a number of ?igure23 Natural Gas Pipeline 
refined products, the Crossing Under Railroad Beds 
transport a number of refined products, the principal ones being
gasoline and jet fuel. They are made of high grade carbon steel.
They operate at 1060-1690 psig. The lines pump about 30,000
barrels of fuel per day.
Most of the pipeline highway or railroad bed crossings are cased
crossings. However, none of the Southern Pacific crossings are
cased, and a few of the I-15 and Highway 138 crossings are uncased.
That occurred because the Southern Pacific did not require casing
when they authorized the right-of-way crossing, and when I-15
replaced Highway 66 and when Highway 138 was recently realigned (in
1990-1991) they did not follow all of the old routes. At new
crossings the pipeline was already installed, and new casings were
not required to be retrofit.
In the regions of MMI = VII or less, the shaking damage state for a 
buried gas or petroleum products transmission line is 1, and no
collocation damage is estimated except for unusual conditions to be
described below.
Following the pipelines from the center of the northern part of the
study area and heading south, the first location for significant
damage occurs at the crossing of the San Andreas fault. Here the
12 meter fault displacement is assumed to rupture the lines. The
location includes the 36-inch natural gas pipeline and the two
petroleum products pipelines (all of which rupture) and the
Victorville-Century 287.5 kV lines 1 and 2. The damage scenario
includes assuming that an explosion/fire occurs in addition to the 
discharge of petroleum products. The need to mitigate the impact
of the fire and adverse environmental conditions and the delays 
caused by the general congestion in the area results in an assumed 
37 day delay before restoration work can begin. Because of the
high probability (100%) of the damage state, the incremental
increase in the lifeline restoration time is also 37 days.
At Blue Cut the natural gas pipeline is next to the railroad bed
and at the toe of a landslide area. High voltage power lines are
located above the gas line in the slide area. A damage state 4 is
predicted for the pipeline. The heavy equipment used to clear the
railroad bed is assumed to increase the damage state to level 5.
This increases the pipeline repair time by 18 days, and the debris
clean up adds another 20 day delay. The general congestion adds a 
7 day delay. This results in a change in the most probable 
restoration time of 12 days. Further south the pipeline follows
the railroad bed along the west edge of Cajon Creek. About 4 miles
further south, the pipeline and railroad are again at the toe of a 
landslide region. In this case only the 18 day increase in repair
time and the 20 day debris removal delay are assumed. This results
in an 8 day increase in the probable restoration time for that 
pipeline, for a total increase in time of 20 days.
Moving from the north to the south, the eastern most natural gas 
pipeline first experiences a potential collocation damage condition
where it crosses I-15 in the region where the fiber optic conduits
cross under I-15 in a concrete culvert and the electric power lines
cross over the highway. This is a region of potential
liquefaction. The damage state is 7 but the probability of damage
is only 20% for the pipeline (but 40% for the lifelines that are
right in the local creek bed). The pipeline is somewhat removed
from the other lifelines, but there are railroad bridges also in
the general area. It is assumed that the other lifeline repairs
will be based on the requirement that long lengths of the natural
gas pipeline will need to be exposed and examined to assure that a
leak/explosion potential does not exist while they work on the
other lifelines. This leads to a 20 day delay in starting the
pipeline repair, resulting in a probable restoration delay of 8
days. About 3.2 miles further south the pipeline passes through
another landslide zone near where it crosses I-15, leading to a
probable increase in the restoration time of 16 days.
Across Cajon Creek from where the western pipeline was exposed to a
potential landslide, the eastern natural gas pipeline crosses
perpendicular to the two petroleum product pipelines and the fiber
optic cables. This is a potential liquefaction zone that results
in a pipeline damage state of 7 with a probability of 20%. There
is a 30 day delay while the petroleum product spill is cleaned up
and another 7 day delay for general congestion resolution
questions. This results in an incremental change in the probable
restoration time of 2 day.
Just south of the I-15/I-215 interchange the 16-inch natural gas
pipeline crosses a local liquefaction zone. It is collocated with
the two petroleum products pipelines, the railroads, and two of the
fiber optic cables. The liquefaction results in a damage state 7
with a probability of 40%. The collocation damage scenario assumes
a 30 day delay to clean up the petroleum product spills, and
another 20 days to account for the congestion and delays
experienced because of the repair to the nearby bridges. The net
increase in restoration time is 8 days.
The two petroleum products pipelines enter the study area from the
south in the Cajon Creek Wash. The 14-inch lines is collocated
along the railroad right-of-way, the 8-inch is on the western side
of the wash. Just south of the I-15/I-215 interchange, the 14-inch
pipeline enters a local liquefaction zone, along with the 16-inch
natural gas pipeline, the railroads, and two fiber optic cables.
Damage state 7 (40% probability) occurs in this region. There is a
20 day delay due to the general level of congestion, and because
the 14-inch pipeline is collocated with the railroad, there will be
a requirement to expose and inspect the pipeline before it can be
put back into service. This will add another 40 days of delay.
This results in a 10 day increase in the probable time to restore
service. About 6 miles further north, both pipelines enter another
liquefaction region, along with a 36-inch natural gas pipeline and
the collocated four fiber optic conduits. The low probability of
liquefaction (20%) and the delay due to congestion (7 days) results
in a 1 day increase in the probable restoration time. The 
pipelines enter a third liquefaction zone near Blue Cut. There the 
petroleum products pipelines cause delays in the repair of the
fiber optic cables. The fiber optic cables do not impact the
pipelines, and there is no collocation impact on the pipelines.
At the San Andreas Fault the pipelines are collocated with high
voltage power lines and a 36-inch natural gas pipeline. Because
the petroleum pipelines are located for several miles along the
fault rift zone, there will be a lengthy delay of several months
while the suitability of allowing them to relay the pipeline along
that route is resolved with the regulatory authorities. However,
that is not a collocation issue unless the pipelines are to be
rerouted near the existing natural gas pipeline. However the
damage scenario includes an explosion/fire, which will increase the
amount of pipeline that must be exposed and inspected. In
addition, the general congestion and environmental mitigation
activities will cause a 30 day delay, for a probable increase in 
the pipeline restoration time due to collocation of 30 days. Thus,
the total increase in the probable time to restore service for
these pipelines is about 41 days, most of that impact is due to
conditions at the fault rift zone.
Ninety-three collocations involving fuel pipelines were analyzed
during this study. A summary of the collocation impacts is:
Increase in Probable Increase in Probable 
Time to Restore Time to Restore 
Lifeline Service, days Service., 
Western Natural Gas 57 86 
36-inch Pipeline 
Eastern Natural Gas 25 83 
36-inch Pipeline 
Natural Gas 8 80 
16-inch Pipeline 
Petroleum Products 41 63 
Pipelines 
Transportation Lifelines 
The Cajon Pass has been used for critical transportation routes
since early times. At present it is used by the Southern Pacific,
Santa Fe, and Union Pacific railroads, and Interstate Highway I-15.
In addition, there are connecting highways, including the I-215
spur into San Bernardino, State Route 138 coming from the west into
the Pass from Palmdale and continuing to the lake district in the
east, and a partially abandoned section of old Federal highway U.S.
66, now called the Cajon Blvd. extension. The routes of these
lifelines in Cajon Pass is shown in Figure 24. The location of the
photographs presented in this section of the report are also shown
on the figure.
Because these routes are discontinuous in the sense that each
bridge, each change in MMI intensity, and each local liquefaction
or landslide area must be separately checked. The railroads had to
be segmented into 30 separate analysis sections and the I-15
highway into 36 separate analysis sections. During the
vulnerability analysis of these facilities, it was necessary to
consider some extension outside the study area in order to provide
realistic estimates of time to restore service. In the case of the
Southern Pacific Railroad, for example, the route considered was
extended to the Highland Boulevard over-crossing in San Bernardino.
The assumptions with regard to equipment available to make repairs
differs from that used for the pipeline and communication lifelines
in that it is assumed that the railroads and highways both have
local active maintenance yards with their own heavy equipment for
construction activities. While some of this may be pressed into 
service for life saving activities in the early emergency phase, it
is not likely that this will prevent immediate inspection and
reconnaissance. Therefore, no delay time waiting for equipment
availability was assumed.
Moving this equipment to the most critical sites along the
transportation lifeline, on the other hand, may present a
significant problem of access because the equipment must move along
the lifeline facility itself. In each analysis, it is first
assumed that the equipment must work from one end or the other of
the Cajon Pass, repairing each section as it goes before it can
reach the next section. The probable access time to a given
section is the sum of the products of times to repair all sections
up to that point multiplied by their respective probability of
damage. For the conditions existing in the Cajon Pass, this leads
to very long access times for those sections remote from the Pass
entrances. A second analysis was therefore made in which the
possibility of construction of temporary by-passes around damaged
sections to permit the access of construction equipment to more
critical sites was considered (this, of course, only applies to the
highway portions of the transportation lifelines). The access time
to a given site in that situation became the sum of the products of
the by-pass times and the probability that each of them is required
because of the damage-in the section being analyzed. Some of the
highway bridges on I-15 have built in by-pass capability, since
they are part of "diamond" interchanges in which the ramps may
serve this purpose. In the generally dry conditions of Cajon Pass,
it is possible in many cases to simply drive across country in
tracked vehicles and lightly loaded four wheel drive trucks. Some
road bed material would need to be placed to support heavy highway
construction equipment.
Figure 24, TRANSPORTATION LIFELINE ROUTES 
SCALE 
4 Larger Scale Figure 
Located at
EXPLANATION 
End of Document 
2 ramrod 
The purpose 
of these as 
access time 
analyses was 
to find the 
critical 
total time 
for the 
lifeline 
section being 
evaluated, 
that is, the 
time to gain 
access to the 
site with 
repair 
equipment 
plus the time 
to carry out 
the needed 
repairs. As 
in the case Figure 25 I-15 Bridge Over The Railroads In Cajon Wash 
and power 
lines, each 
lifeline was first divided into sections, such that the conditions
within each
section were
homogeneous.
Because of
the presence
of many
bridges on
the highways
and
railroads,
this leads to
more sections
for the
roughly 25
miles of
length of
each separate
transportation
system.
Prior to the
detailed
analysis 
the lifeline I
section being Figure 26 I-15 Bridge Over Cajon Wash
considered, the Bridge Vulnerability Index of each of these 
structures had to be determined, and the probable extent and type 
of damage to each postulated. 
The highway bridges along Interstate 1-15 were, for the most part, 
built in the late 1960fs, except for the section over Lytle Creek 
Wash. Although they are in an area generally considered to be 
If California region 7", the status of retrofit is somewhat 
irregular. In this analysis, they are considered to be "California 
region 3-6" until proved to qualify for the higher degree of 
safety. Nevertheless, none of these bridges are expected to 
completely collapse, although partial collapse of several is 
possible. The 1-15 bridge over the railway lines (Figure 25) and 
the high level 1-15 bridge over Cajon Wash (Figure 26) are 
vulnerable, and there is some possibility of the partial collapse 
of the steel girders over 1-15 at its junction with highway 138 
(Figure 27). 
Many of the railroad bridges in the Cajon Pass are over 50 years
old, but most are in relatively good condition. As noted in the 
discussion of the development of the Bridge Vulnerability Index, 
many of these bridges have inherent resistance to lateral loads. 
There are, however, several multiple simple span bridges over poor 
soil conditions (including possible liquefaction), such as both the 
Southern Pacific and Santa Fe bridges over the lower end of Cajon
Wash (Figure 28). There are several more such crossings over the 
Cajon Creek and its branches. There is also a large two span, 
through
plate, girder 
bridge on the 
Southern 
Pacific 
railroad over 
Highway 138 
which is 
sharply
skewed, and 
which has 
bearings
which are 
vulnerable to 
loss (Figure 
29). It is 
expected that 
the multiple 
span 
structures 
will have one 
or more spans 
dislodged
where they 
are subjected Figure 27 Highway 138 Bridge Over 1-15 
Figure 28 Railroad Bridges In Cajon Wash (-15 Bridge
In The Background)
Figure 29 Railroad Bridge Over Highway 138
(Collapse Expected)
86
to MMI= VIII 
shaking
intensity 
zones or are 
on 
potentially
liquefiable
soils. The 
failure of 
the Bridge 
shown in 
Figure 29 was 
responsible
for much of 
the access 
time estimate 
for repair of 
1-15, since 
it blocked 
Highway 138 
from being an 
immediate 
detour route 
for equipment Figure 30 Santa Fe Railroad Rubble Masonry Pier Bridge 
needed for 
the 1-15 repairs. 
There is one 
bridge on the 
Santa Fe 
railroad just 
south of the 
1-15 truck 
Weighing station which 
is founded on 
rubble 
masonry piers 
on sandy soil 
with a high 
water table 
 (Figure 30).
Loss of one 
or more spans 
is 
anticipated.
That is 
contrasted to 
the Union I 
Pacific 
railroad 
bridge at the 
Figure 31 Union Pacific Railroad Bridge With Power 
Lines Overhead 
same location (Figure 31) which has steel column, pile piers.
Figure 14 shows surface conditions near these two bridges.
The calculations for the vulnerability analysis and time to restore
service involve only simple arithmetic, but they are quite time
consuming if carried out manually. For this report, the analyses
were done on a computer spreadsheet, which greatly aided an orderly
and efficient approach.
The primary objective of the study was to determine how the times
to restore full service would be affected by the collocation of
several types of lifelines in the same congested corridor. The
interaction scenarios have been discussed in this report in earlier
sections. There are, however additional problems because of the
highway-railroad interactions. For the Cajon Pass application,
these interactions occur at the same general areas as the
previously identified critical clusters:
(1) The area near the liquefiable zone just south of the
highway I-15 crossing over the railroads and the Cajon Wash,
near the junction with I-215. This includes intersection
points 8, 9, 10, 11, 13, 14, 15, 16, 18, and 19. A partial
collapse of highway bridge No. 54-0818 over the three rail
lines adds to the problems at this location. A delay time of
30 days due to this bridge problem (Figure 25) was added to
the previously noted 30 day delay due to pipeline damage and
hazards.
(2) The section of the steep slide prone slopes along the west
side of Cajon Canyon (see Figure 17) and the liquefiable zone
on the east side about one mile north of the junction of I-15
an I-215. This includes intersections 22, 25, 26, 27, 28, 29,
30 on the east side and 31, 32, 33, 34, and 35 on the west 
side of the canyon. The 30 day delay previously established 
appears adequate.
(3) The conditions in Blue Cut are so congested, combined with 
the expected explosion and/or fire, that an increase in the
expected damage states for the railroads by one level is
justified. Intersections 38, 39, 93, and 99 are involved. A 
60 day delay in access is also assumed. 
(4) At the San Andreas fault zone, intersections 37, 40, 41, 
and 91 are involved. Problems with fuel pipelines and power
lines have already been noted. The 30 day delay in initiating
repairs to other lifelines was applied to the railroads.
(5) The area just north of the section of Highway 138 and west
of Highway I-15. Problems with pipelines and power lines have
already been noted at intersection points 47, 48, 49, 51, 52,
and 53. There is also a possibility of partial collapse of
the Southern Pacific railroad bridge over Highway 138 (Figure 88
29). An additional delay time of 30 days was added for access
involving this bridge, over and above the 30 days to clear
fire hazards related to pipeline damage.
(6) There are several other minor critical areas:
(a) The area just west and south of the I-15 truck
weighing station. The crib retaining wall could slump 
(see Figure 14). The principle effect is on Southern 
Pacific railroad sections 15 (westbound) and 24
(eastbound). A 30 day delay was assumed.
(b) Highway Structures 0796, 0797 and 0827 which carry
1-15 over the rail lines at Alray and Gish at
intersections 55 and 57 (Figure 32). These structures 
may only be lightly damaged, but a 10 day delay in
railroad access is assumed to permit time for inspection
and temporary shoring if required.
(c) The I-15 bridge 0664 (Figure 33) over the Southern
Pacific tracks north of the pass at intersection 83 is
expected to be lightly damaged, but 10 days delay is 
allowed for inspection (also see Figure 15 which shows
details of this bridge).
The effect of these collocation delays on the restoration of the
transportation lifelines was evaluated by making a analysis with
collocation assumed. This is a "second pass" analysis with the
spreadsheet. A special problem developed in reassessment of the 
alternate route by way of highway 138, in that access to the 
connection point on the Santa Fe was blocked by the expected
partial bridge collapse on the Southern Pacific. For this reason
the delay time associated with this problem was added to the
previous estimated time to reach the connection point, that is 10
days plus 60 days delay time. This gives 70 days. The cumulative
access times for this route were then computed as before, working
both ways from this point.
A summary of the results of the study are presented below.
Increase in Probable Increase in Probable
Time to Restore Time to Restore
Lifeline Service, days Service, %
Highway I-15 35 22
Southern Pacific 17 8
Railroad
Atcheson Topeka & 85 33 
Santa Fe Railroad
The smaller
percent
in increase 
time to l 
restore l 
the Southern
Pacific
railroad
compared to
the other
transportation
lifelines.
is due in
part to its
more
favorable
location with
respect to
other
lifelines;
but it should
also be noted Figure 32 Typical I-15 Box Bridge Over the Railroads
probable
partial collapse of one the bridges on this line contributes to
large access
for the
others.
Figure 33 I-15 & Access Road Bridges Over the Southern
Pacific Railroad
90


5.3 Chapter 5.0 Bibliography
1. P. Lowe, C. Scheffey, and P. Lam, "Inventory of Lifelines in
the Cajon Pass, California", ITI FEMA CP 120190, August 1991.
2. R. Greensfelder, "Maximum Credible Rock Acceleration From
Earthquakes in California", California Division of Mines and
Geology Map Sheet 23, 1974. 
3. T. Hanks, "t The National Earthquake Hazards Reduction Program Scientific 
Status", U.S. Geological Survey Bulletin 1652, 
1985. 
4. J. Evernden, et. al., "Interpretation of Seismic Intensity
Data", Bulletin of the Seismological Society of America, V.
63, 1973.
5. J. Evernden, et. al.., "Seismic Intensities of Earthquakes of
Conterminous United States -Their Predications and 
Interpretations", U.S. Geological Survey Professional Paper 
1223, 1981.
6. J. Davis, et. al., "Earthquake Planning Scenario for a
Magnitude 8.3 Earthquake on the San Andreas Fault in Southern
California", California Division of Mines and Geology, Special
Publication 60, 1982.
7. S. Algermissen, et. al., "Development of a Technique for the
Rapid Estimation of Earthquake Loses", U.S. Geological Survey
Open File Report 78-441, 1978.
8. E. Bertugno and T. Spittler, "Geologic Map of the San
Bernardino Quadrangle", California Division of Mines and
Geology Geologic Map Series, Map No. 3A (Geology), 1986.
6.0 FUTURE STUDY NEEDS 
Recognizing that this is the first comprehensive analysis of the
impact of lifeline collocation on the individual lifeline's
vulnerability, it is recommended that the follow-on studies, be
performed.
1. The collocation analysis should be repeated at another
location outside of California. It will provide information
on the following items:
Is there enough data available to conduct the analysis,
or was the data base available in California unique?
Can the methods suggested by Rojahn to adjust the
California data to other regions be applied to develop
reasonable results?
The site can include water and sewer systems or
reservoirs to assure that the collocation analysis method
can properly treat the impacts of these lifelines, which
were not available at the Cajon Pass.
The study can check the suitability of the LSI-MMI
relationship developed for analyzing liquefaction-induced
damage, the Bridge Vulnerability Index method, and the
lifeline zones of influence, all of which were developed
with the Cajon Pass situation in mind.
If possible, the study site should include lifeline
passage over a large water body, or at least over wet
ground. This will help clarify the impacts of equipment
and material access time compared to lifeline repair
time, as the dry ground of the Cajon Pass did not impose
very restrictive "detour" conditions.
2. In parallel with the above study to further refine the
collocation analysis method, a second study is warranted. 
It should focus on presenting the material to a broad audience.
Special emphasis should be given to contacting lifeline owners
and operators to discuss the study and the results obtained.
Their perspective and response should provide valuable
information on where improvements in the analysis method would
clarify important issues that relate to the siting of
lifelines in "lifeline corridors". It should also help
identify mitigation approaches that reflect the operational
and economic needs of the lifeline providers.
3. A longer term study is needed to provide more detailed data
and expert opinion for lifelines. Most of the current data
emphasizes earthquake impacts on buildings and secondly on
bridges. Most of the present data (including most of the
lifeline data) in the data bases were obtained from the
building and bridge technical sectors. An new study to
examine the present data base presented in ATC-13, but with
full emphasis on lifelines, should be undertaken to allow the
lifeline portions of earthquake analysis to have the same
level of technical input that buildings and structures
presently have.

Appendix A Modified Mercalli Intensity (MMI) Index Scale 93 Modified Mercalli (MM) Intensity Scale I. Not felt-or, except rarely under especially favorable Circumstance e. Under certain conditions, at and outside the boundary of the area In which a great shock is felt: sometimes birds, animals, reported uneasy, disturbed; sometimes dizziness or nausea experienced; sometimes trees1 structures, liquids, bodies of water, may sway--doors mayo swing, very slowly.
II. Felt indoor by few, especially on upper floors, or by sensitive or nervous persons.
Also, s in grade I, but often more noticeably: sometimes hanging objects may swing,' especially been delicately suspended; sometimes trees, structures, liquids, bodies of water, may sway, 'doors may swing, very slowly; sometimes birds, animals, reported uneasy or disturbed; sometimes dizziness or nausea experienced.
III. Felt indoors by several, motion usually rapid vibration.
Sometimes not recognized to be in earthquake at first.
Duration estimated in some cases.
Vibration like that due to passing of light, or lightly loaded trucks, or heavy trucks some distance away.
Hanging objects may swing slightly.
Movements may be appreciable on upper levels of tall structures.
Rocked standing motor car slightly.
Felt indoors by many, outdoors by few.
Awakened few, especially light sleepers.
Frightened no one, unless apprehensive from previous experience.
Vibration like that due to passing of heavy, or heavily loaded trucks.
Sensation like heavy body striking building, or falling of heavy objects inside.
Rattling of dishes) windows, doors; glassware and crockery clink and clash.
Creaking of walls frame, especially in the upper range of this grade.
Hanging objects swung, in numerous instances.
Disturbed liquids in open vessels slightly.
Rocked standing motor cars noticeably.
V. Felt indoors by practically al, outdoors by many or most: outdoors direction estimated.
Awakened many, or most.
Frightened few-slight excitement, a few ran outdoors.
Buildings trembled throughout.
Broke dishes, glassware, o some extent.
Cracked windows-in some cases, but not Generally.
Overturned vases, small or unstable objects, in many Instances, with occasional fall.
Hanging objects, doors, swing generally or considerably.
Knocked pictures against walls, or swung them out of place.
Opened, or closed, doors, shutters, abruptly.
Pendulum clocks stopped,. started, or ran fast, or slow.
Moved small objects, furnishings, the latter to slight extent Spilled liquids in small amount a from well-filled open containers.
Trees, bushes, shaken slightly.
"Adapted from Sieberg's (1923) Mercalli-Cancani scale, modified and condensed. Quoted from Wood and Neumann (1931).
Vl. Felt by all, indoors and outdoors.' Frightened many, excitement general, some alarm, many ran outdoors.
Awakened &l.
Persona made to move unsteadily.
Trees, bushes, shaken slightly to moderately.
Liquid set In strong motion.
Small bells rang-church, chapel, school, etc.
Damage slight in poorly built buildings.
Fall of planter in small amount.
Cracked plaster somewhat, especially fine cracks chimneys In some instances.
Broke dishes, glassware, In considerable quantity, also some windows.
Fail of knick-knacks, books, pictures.
Overturned furniture in many instances.
Moved furnishings of moderately heavy kind.
. Frightened all-general alarm, all ran outdoors.
Some, or many, found it difficult to stand.
Noticed by persons driving motor cars.
Trees and bushes shaken moderately to strongly.
Waves on ponds, lakes, and running water.
Water turbid from mud stirred up.
In caving to some extent of sand or gravel stream banks.
Rang large church bells, etc.
Suspended objects made to quiver.
Damage negligible in buildings of good design and construction, slight to moderate in well-built ordinary buildings, considerable in poorly built or badly designed buildings, adobe houses, old walls (especially where laid up without mortar), spires, etc.
Cracked chimneys to considerable extent, walls to some extent.
Fall of plaster in considerable to large amount, also some stucco.
Broke numerous 'windows, furniture to some extent.
Shook down loosened brickwork and tiles.
Broke weak chimneys at the roofline (sometimes damaging roofs).
Fall of cornices from tower and high buildings.
Dislodged bricks and atones.
Overturned heavy furniture, with damage from breaking.
Damage considerable to concrete irrigation ditches.
Fright general-alarm approaches panic.
Disturbed persons driving motor ears.
Trees shaken strongly-branches, trunks, broken off, especially palm trees.
Ejected sand and mud In small amounts.
Changes: temporary, permanent; in flow of springs and wells; dry wells renewed flow; in temperature of spring and well waters.
Damage alight in structures (brick) built especially to withstand earthquakes.
Considerable In ordinary substantial buildings, partial collapse: racked, tumbled down, wooden houses in some cases; threw out panel walls in frame structures, broke off decayed piling.
Fall of walls.
Cracked, broke, solid stone walls seriously.
Wet ground to some extent, also ground on steep slopes.
Twisting, fall, of chimneys, columns, monuments, also factory stacks, towers.
Moved conspicuously, overturned, very heavy furniture.
IX. Panic general.
Cracked ground conspicuously.
Damage considerable in Masonry) structures built specially to withstand earthquakes: threw out of plumb some wood-frame houses built especially to withstand earthquakes; peat in substantial masonry) buildings, some collapse in large part; or wholly shifted frame buildings off foundations, racked frames; serious to reservoirs; underground pipes sometimes broken.
X. Cracked ground, especially when loose and wet, up to widths of several inches; fissures up to a yard in width ran parallel to canal and stream banks.
Landslides considerable from river banks and steep coasts.
Shifted sand and mud horizontally on beaches and flat land.
Changed level of water in wells.
Threw water on banks of canals, lakes rivers, etc.
Damage serious to dams, dikes, embankment Damage severe to well-built wooden strictures and bridges, some destroyed.
Developed dangerous cracks in excellent brick wails.
Destroyed most masonry and frame structures, also their foundations.
Bent railroad rails slightly.
Tore apart, or crushed endwise, pipe lines buried in earth.
Open cracks and broad wavy folds in cement pavements and asphalt road surfaces.
XI. Disturbances In ground many and widespread, varying with ground material Broad fissures, earth slumps, and land slips in soft, wet ground.
Ejected water in large amount charged with sand and mud.
Caused sea-waves (tidal waves) of significant magnitude.
Damage severe to wood-frame structures, especially near shock centers.
Great to dams, dikes, embankments, often for long distances.
Few, if any masonry), structures remained standing.
Destroyed large well-built bridges by the wrecking or supporting piers, or pillars.
Affected yielding wooden bridges less.
Bent railroad rails greatly, and thrust them endwise.
Put pipe lines burled in earth completely out of service.
XII. Damage total-practically all works of construction damaged greatly or destroyed.
Disturbances in ground great and varied, numerous shearing cracks.
Landslides, falls of rock of significant character, slumping of river banks, etc., numerous and extensive.
Wrenched loose, tore off, large rock masses.
Fault slips in firm rock, with notable horizontal and vertical offset displacements.
Water channels, surface and underground, disturbed and modified greatly.
Dammed lakes, produced waterfalls, deflected rivers, etc.
Waves seen on ground surfaces (actually seen, probably, in some cases).
Distorted lines of sight and level.
Threw objects upward Into the air.
1. Not felt except by a very few under especially favorable circumstances.
I. Felt only by a few persons at rest, especially on upper floors of buildings.
Delicately suspended objects may swing.
II. Felt quite noticeably indoors, especially on upper Boors of buildings, but many people do not recognize it an earthquake. Standing motor can may rock slightly. Vibration like passing of truck. Duration estimated.
TV. During the day felt Indoors by many, outdoors by few. At night some awakened. Dishes, windows, doors disturbed; walls made cracking sound. Sensation like heavy truck striking building. Standing motor cars rocked noticeably.
V. Felt by nearly everyone; many awakened. Some dishes, windows, etc., broken; a few instances of cracked plaster; unstable objects overturned.
Disturbance of trees, poles and other tall objects sometimes noticed.
Pendulum clocks may stop.
VI. Felt by all; many frightened and run outdoors 8ome heavy furniture moved; a few instances of fallen plaster or damaged chimneys. Damage slight.
VII. Everybody runs outdoors. Damage negligible in buildings of good design and construction; slight to moderate In well-built ordinary structures; considerable In poorly built or badly designed structures; some chimneys broken. Noticed by persons driving motor cars.
VIII. Damage slight in specially designed structures; considerable in ordinary substantial buildings with partial collapse; great In poorly built structures.
Panel walls thrown out of frame structures. Fall of chimneys, factory stacks, columns, monuments, walls. Heavy furniture overturned.
Sand and mud elected in small amounts. Changes In well water. Disturbed persons driving motor cars.
IX. Damage considerable in specially designed structures; well designed frame structures thrown out of plumb; great In substantial buildings, with partial collapse. Buildings shifted off foundations. Ground cracked conspicuously. Underground pipes broken.
X. Some well-built wooden structures destroyed; most masonry and frame strictures destroyed with foundations; ground badly cracked. Rails bent.
Landslides considerable from river banks and steep slopes. Shifted sand and mud. Water splashed (slopped) over banks.
XI. Few, if any (masonry), structures remain standing. Bridges destroyed.
Broad fissures in ground. Underground pipe lines completely out of service. Earth slumps and land slips in soft ground. Rails bent greatly.
XII. Damage total. Waves seen on ground surfaces. Lines of sight and level distorted. Objects thrown upward into the air.
INTERSTATE PAVED HIGHWAY COLLOCATION LOCATION RAILROAD POWER LINES PETROLEUM PRODUCT PIPELINES NATURAL GAS PIPELINES FIBER OPTIC CABLE FLUME-PENSTOCK 117 31 30 25 117 22 FIGURE 9, COMMUNICATION LIFELINE ROUTES SCALE MILES 0 12 KILOMETERS EXPLANATION INTERSTATE PAVED Highway FIGURE 16, ELECTRIC POWER LIFELINE ROUTES SCALE I 2 MILES EXPLANATION 15 Interstate 12 PAVED HIGHWAY POWER LINES BURIED AQUEDUCT 117 317 117 31 117 22 FIGURE 20, FUEL PIPELINE LIFELINE ROUTES SCALE 0 I 2 MILES 0 1 2 KLOMETMES EXPLANATION 1-15 INTERSTATE 2 PAVED EIGHWAY PETROLEUM PRODUCT PIPEANES NATURAL GAS PIPEIUNES 0 VALVES 117 31 117 31 30 117 22 FIGURE 24. TRANSPORTATION LIFELINE ROUTES SCALE EXPLANATION 1-15 INTERSTATE 2 PAVED HIGHWAY RAILROAD