November 2007

Revised October 2008

Technical Memorandum: UCPRC-TM-2007-08

Alligator Cracking Performance and

Life-Cycle Cost Analysis of Pavement
Preservation Treatments

Authors:
Charles Lee, William A. Nokes, and John T. Harvey

This work was completed as part of Partnered Pavement Research Program Strategic Plan Item 3.2.5:
Documentation of Pavement Performance Data for Pavement Preservation Strategies and Evaluation of Cost-
Effectiveness of Such Strategies

PREPARED FOR: PREPARED BY:

California Department of Transportation University of California

Division of Research and Innovation Pavement Research Center
Office of Roadway Research Davis and Berkeley
DOCUMENT RETRIEVAL PAGE Technical Memorandum No.:
UCPRC-TM-2007-08
Title: Alligator Cracking Performance and Life-Cycle Cost Analysis of Pavement Preservation Treatments

Author: C. Lee, W. A. Nokes, and J. T. Harvey

Prepared for: FHWA No. Work Submitted Date:

Caltrans CA091561B April 23, 2008 November 2007
Revised October 2008
Strategic Plan No: Status: Version No:
3.2.5 Stage 6, final 2
Abstract:
This memo describes work done to (1) develop performance estimates for pavement preservation treatments, and (2) estimate the
cost-effectiveness in pavement preservation implementation. Construction project histories were collected from Caltrans, and their
performance histories in terms of alligator cracking were extracted from the Pavement Condition Survey (PCS) database. The
UCPRC-developed algorithm resolved the dynamic segmentation issue in the PCS data. Methodology on how to select adequate
project data to use is documented. Median pavement lives to 10 percent and 25 percent Alligator B cracking for various flexible
pavement treatments are estimated by means of a Kaplan-Meier estimator.

Life-cycle cost analysis (LCCA) was performed with 20-year and 35-year analysis periods to try to determine (1) whether
pavement preservation should be applied, (2) when pavement preservation should be applied. LCCA could only be completed for
dense-graded asphalt overlays and chip seals with conventional binders due to limitations in complete datasets that could be
gathered from existing Caltrans data sources. Cost differences were compared between continuous application of pavement
preservation treatments or limiting these treatments to not more than twice in between rehabilitations. Findings from the LCCA
show that, in general, (1) it is more cost-effective to apply pavement preservation treatment than to only rely on rehabilitation and
(2) it is more cost-effective to apply pavement preservation at earlier stages of cracking rather than later.

However, due to major limitations on the data that could be gathered for the analysis and despite the major effort by the UCPRC
and by Caltrans HQ and Division of Maintenance staff, the results presented should only be considered as indications of
pavement performance and life-cycle cost
Keywords: life-cycle cost analysis, pavement preservation, pavement performance, PMS Database
Proposals for implementation: 1. Use pavement preservation treatments, as opposed to using rehabilitation alone. Place
pavement preservation treatments at lower levels of cracking than was the practice from 1988 to 2003, with the comparison in this
study being made between treatment at less than or equal to 10 percent Alligator B cracking versus more than 25 percent Alligator
B cracking. 2. Implement changes in PMS: implement system for collecting as-built information for rehabilitation and
maintenance work; use fixed segmentation of network as opposed to changing segmentation from year-to-year; make
recommended changes in PCS. 3. Establish pavement preservation test sections following guidelines in Pavement Preservation
Strategies Technical Advisory Guide (PPSTAG). 4. Use the framework presented in this report for future LCCA evaluations of
the cost-effectiveness of pavement treatments.
Related documents:
“Pavement Performance Data Extraction from Caltrans’ PMS” (UCPRC-TM-2007-06)

Work plan: “Performance Review of Pavement Preservation Treatments” (October 9, 2006). Prepared by the University of
California Pavement Research Center, Berkeley and Davis, for the California Department of Transportation Division of Research
and Innovation Office of Roadway Research and Division of Maintenance Office of Pavement Preservation
Signatures:

C. Lee W. Nokes, D. Spinner J. Harvey T. J. Holland

1st Author J. Harvey Editor Principal Caltrans Contract
Technical Review Investigator Manager

ii UCPRC-TM-2007-08
 DISCLAIMER
The contents of this report reflect the views of the authors, who are responsible for the facts and accuracy
of the data presented herein. The contents do not necessarily reflect the official views or policies of the
State of California or the Federal Highway Administration. This report does not constitute a standard,
specification, or regulation.
ACKNOWLEDGMENTS
The Caltrans Division of Research and Innovation (DRI) Contract Manager was Michael Samadian, under
the direction of Nick Burmas. The Pavement Standards Team Technical Lead was Shakir Shatnawi.

The authors would also like to acknowledge the help of the Caltrans Division of Maintenance Office of
Roadway Rehabilitation, under the direction of Susan Massey, which has provided continued guidance
and assistance on this project, including provision of historic construction files provided by Leo
Mahserelli, Rob Marsh, Brian Webber, and Ron Jones.

PROJECT OBJECTIVES
The purpose of this project was to develop performance estimates for pavement preservation treatments
following rehabilitation and to compare them with performance estimates for sections which did not
receive pavement preservation treatments after rehabilitation. The performance estimates were made
using condition survey data from the Caltrans Pavement Condition Survey (PCS) database and
maintenance and rehabilitation construction history data obtained from Caltrans. The focus of this research
was on flexible and composite (asphalt on concrete) pavements. The objectives of this work were as
follows:
• Identify pavement preservation treatments in the PCS data in the Caltrans Pavement Management
System (PMS) databases;
• Retrieve and review PCS data and other data from available sources in order to develop
performance estimates;
• Develop condition survey performance histories;
• Perform statistical analyses to identify performance probabilities;
• Analyze timing of pavement preservation treatments in terms of extent of cracking and their
effects on relative performance;
• Calculate life-cycle costs; and
• Report the results of the investigation.

UCPRC-TM-2007-08 iii
 CONVERSION FACTORS
SI* (MODERN METRIC) CONVERSION FACTORS
APPROXIMATE CONVERSIONS TO SI UNITS
Symbol Convert From Multiply By Convert To Symbol
LENGTH
in inches 25.4 Millimeters mm
ft feet 0.305 Meters m
AREA
2
in square inches 645.2 square millimeters mm2
ft2 square feet 0.093 square meters m2
VOLUME
3
ft cubic feet 0.028 cubic meters m3
MASS
lb pounds 0.454 Kilograms kg
TEMPERATURE (exact degrees)
°F Fahrenheit 5 (F-32)/9 Celsius C
or (F-32)/1.8
FORCE and PRESSURE or STRESS
lbf poundforce 4.45 Newtons N
2
lbf/in poundforce/square inch 6.89 Kilopascals kPa
APPROXIMATE CONVERSIONS FROM SI UNITS
Symbol Convert From Multiply By Convert To Symbol
LENGTH
mm millimeters 0.039 Inches in
m meters 3.28 Feet ft
AREA
2
mm square millimeters 0.0016 square inches in2
m2 square meters 10.764 square feet ft2
VOLUME
3
m cubic meters 35.314 cubic feet ft3
MASS
kg kilograms 2.202 Pounds lb
TEMPERATURE (exact degrees)
C Celsius 1.8C+32 Fahrenheit F
FORCE and PRESSURE or STRESS
N newtons 0.225 Poundforce lbf
kPa kilopascals 0.145 poundforce/square inch lbf/in2
*SI is the symbol for the International System of Units. Appropriate rounding should be made to comply with Section 4 of ASTM E380.
(Revised March 2003)

iv UCPRC-TM-2007-08
 TABLE OF CONTENTS
List of Tables............................................................................................................................................. vii
List of Figures .............................................................................................................................................ix
List of Abbreviations Used in the Text ......................................................................................................x
1 Project Overview...................................................................................................................................1
1.1 Background and Objectives......................................................................................................1
1.2 “Data-Driven” Approach Used for the Study ...........................................................................5
1.3 Chronology of Activity.............................................................................................................5
1.4 Order of Work Summary ..........................................................................................................6
2 Process for Selecting Usable Data........................................................................................................9
2.1 Dynamic Segmentation...........................................................................................................10
2.2 Project Selection .....................................................................................................................14
3 Summary of Database.........................................................................................................................17
3.1 Initial Database of Projects.....................................................................................................17
3.2 Database of Observations .......................................................................................................18
3.2.1 Adjusting for Traffic ....................................................................................................18
3.2.2 Adjusting for Missing Lane Prior Condition Data .......................................................19
3.2.3 Distribution of Projects and Observations by District..................................................22
4 Summary of Practice 1988 to 2003 ....................................................................................................25
5 Factors influencing Performance ......................................................................................................29
5.1 Check Correlation of Factors..................................................................................................29
5.2 Analysis of Factors .................................................................................................................30
6 Pavement Life Results ........................................................................................................................33
6.1 Procedures Used to Develop Probability of Failure ...............................................................33
6.2 Summary by Strategy and Program........................................................................................36
6.3 Summary of Factors Affecting Performance ..........................................................................38
6.4 Annualized Pavement Treatment Costs..................................................................................41
7 Life-Cycle Cost Analysis.....................................................................................................................47
7.1 Assumptions for LCCA ..........................................................................................................49
7.2 Factors Sensitivity Analysis ...................................................................................................49
7.3 Attempt to Predict Pavement Performance by Linear Extrapolation......................................55
7.4 Service Life Used for LCCA ..................................................................................................56
7.5 LCCA Scenarios .....................................................................................................................58
7.6 LCCA Results and Findings ...................................................................................................63

UCPRC-TM-2007-08 v
 7.6.1 LCCA Objective 1: Is It More Beneficial to Apply Pavement Preservation (HM-1)
Versus No Maintenance?...................................................................................................63
7.6.2 LCCA Objective 2: Should Pavement Preservation Be Applied at an Earlier or a Later
Stage of Cracking? ............................................................................................................65
7.6.3 LCCA Objective 3: Comparison of LCC of Different Combinations of Pavement
Treatments .........................................................................................................................67
8 Summary and Preliminary Conclusions ...........................................................................................71
8.1 Tracking Pavement Preservation Projects ..............................................................................71
8.2 Estimated Pavement Performance ..........................................................................................72
8.3 Life-Cycle Cost Analysis........................................................................................................73
8.4 Recommendations .................................................................................................................75
8.3.1 Pavement Preservation Practice ...................................................................................75
8.3.2 Caltrans PMS Data.......................................................................................................75
8.3.3 Pavement Preservation Test Sections...........................................................................76
8.3.4 Framework for Future LCCA Evaluations of Cost-Effectiveness of Pavement
Treatments .........................................................................................................................76
9. References............................................................................................................................................77
Appendix A: Glossary ...............................................................................................................................79
Appendix B: Definition of Data Censoring .............................................................................................81
Appendix C: Example Calculation of Probability of Failure Distribution for Figure 8. ....................83

vi UCPRC-TM-2007-08
 LIST OF TABLES
Table 1: Project Objectives, Deliverables, and Notes ...................................................................................2
Table 2: List of Deleted Project Histories ...................................................................................................17
Table 3: Project Distribution—One Observation per Selected Project .......................................................18
Table 4: Average Lane Distribution Factor of AADTT for Highway with Different Number of
Lanes in One Direction (4) ..................................................................................................................19
Table 5: Distribution of Observations (Considering Multiple Lanes and Both Directions)........................19
Table 6: Number of Observations with Missing PCS Data.........................................................................19
Table 7: Distribution of Observations After Adjusting for Missing Prior Condition Data .........................20
Table 8: Distribution of Observations for Final Set of Data Used for.........................................................21
Table 9: Distribution of Observations for Composite Sections...................................................................21
Table 10: Distribution of Observations by Program Type and District.......................................................22
Table 11: Distribution of Observations by Strategy and District ................................................................22
Table 12: Distribution of Projects and Observations by Districts ...............................................................23
Table 13: Mean Existing Cracking Levels When Each PP Strategy Performed .........................................25
Table 14: Existing Cracking Data for Each PP Strategy and Program........................................................27
Table 15: Correlation Table—Factors Affecting Pavement Performance...................................................30
Table 16: Descriptive Statistics of Factors Affecting Cracking Performance in Data ................................31
Table 17: Statistics of Factors Affecting Pavement Performance in Data ..................................................31
Table 18: Statistics of Factors Affecting Pavement Performance in Data without MinTemp ....................32
Table 19: Significance Level of Temperature Factor in Pavement Performance........................................32
Table 20: Expected Life for Each PP Strategy and Program Showing Years to Failure Levels,
not Pre-Existing Cracking....................................................................................................................37
Table 21: ACOL-DG HM-1 ........................................................................................................................38
Table 22: ACOL-DG CapM........................................................................................................................39
Table 23: ACOL-DG Rehab........................................................................................................................39
Table 24: ACOL-OG HM-1 ........................................................................................................................39
Table 25: ACOL-RAC HM-1......................................................................................................................40
Table 26: ACOL-RACO HM-1...................................................................................................................40
Table 27: Chip Seal .....................................................................................................................................40
Table 28: Digout..........................................................................................................................................41
Table 29: Slurry...........................................................................................................................................41
Table 30: ACOL-DG HM-1 Annualized Cost per Lane-Mile (x $1,000) ...................................................42
Table 31: ACOL-DG CapM Annualized Cost per Lane-Mile (x $1,000)...................................................43

UCPRC-TM-2007-08 vii
Table 32: ACOL-DG Rehabilitation Annualized Cost per Lane-Mile (x $1,000) ......................................43
Table 33: ACOL-OG HM-1 Annualized Cost per Lane-Mile (x $1,000) ...................................................43
Table 34: ACOL-RAC HM-1 Annualized Cost per Lane-Mile (x $1,000).................................................44
Table 35: Chip Seal Annualized Cost per Lane-Mile (x $1,000) ................................................................44
Table 36: Digout Annualized Cost per Lane-Mile (x $1,000).....................................................................44
Table 37: Slurry Annualized Cost per Lane-Mile (x $1,000)......................................................................45
Table 38: AADTT Sensitivity Test Table ...................................................................................................50
Table 39: Average Annual Precipitation Sensitivity Test Table .................................................................50
Table 40: Average Annual Minimum Temperature Sensitivity Test Table ................................................51
Table 41: Average Annual Minimum Temperature in January Sensitivity Test Table...............................52
Table 42: Average Annual Freeze-thaw Cycles Sensitivity Test Table ......................................................53
Table 43: ACOL-DG HM-1—Rural vs. Urban District Sensitivity Test Table..........................................54
Table 44: ACOL-DG Rehabilitation—Rural vs. Urban District Sensitivity Test Table .............................54
Table 45: Median vs. Estimated ACOL-DG Rehabilitation Lives..............................................................56
Table 46: Dense-Graded Asphalt Concrete Overlay (Rehabilitation) Performance...................................56
Table 47: Dense-Graded Asphalt Concrete Overlay (CapM) Performance ................................................56
Table 48: Rubber Asphalt Concrete Overlay (CapM) Performance ...........................................................57
Table 49: Dense-Graded Asphalt Concrete Overlay (HM-1) Performance ................................................57
Table 50: Rubber Asphalt Concrete Overlay (HM-1) Performance............................................................57
Table 51: Open-Graded Asphalt Concrete Overlay (HM-1) Performance..................................................58
Table 52: Chip Seal (HM-1) Performance ..................................................................................................58
Table 53: LCCA Scenario List for Major Scenario Set No. 1.....................................................................60
Table 54: LCCA Scenario List for Major Scenario Set No. 2.....................................................................60
Table 55: LCCA Scenario List for Major Scenario Set No. 3.....................................................................61
Table 56: LCCA Scenario List for Major Scenario Set No. 4.....................................................................62
Table 57: LCCA Scenario List for Major Scenario Set No. 5.....................................................................62
Table 58: LCCA Scenario List for Major Scenario Set No. 6.....................................................................62
Table 59: LCCA Scenario List for Major Scenario Set No. 7.....................................................................63
Table 60: LCC of Preventive vs. No Preventive Actions (Rehabilitation at >25% Alligator B
Cracking)—Dense-Graded Overlay ....................................................................................................64
Table 61: LCC of Preventive vs. No Preventive Actions (Rehabilitation at >25% Alligator B
Cracking)—Chip Seal..........................................................................................................................64
Table 62: LCC of Preventive Treatment at Early vs. Later Stage of Cracking (Rehabilitation at
>25% Alligator B Cracking)—Dense-Graded Overlay.......................................................................66

viii UCPRC-TM-2007-08
Table 63: LCC of Preventive Treatment at Early vs. Later Stage of Cracking (Rehabilitation at
>25% Alligator B Cracking)—Chip Seal-AC .....................................................................................66
Table 64: LCC of Two HM-1 vs. Unlimited HM-1 Preventive Actions (Rehabilitation at
>25% Alligator B Cracking)—Dense-Graded Overlay.......................................................................68
Table 65: LCC of Two HM-1 vs. Unlimited HM-1 Preventive Actions (Rehabilitation at
>25% Alligator B Cracking)—Chip Seal ............................................................................................68

LIST OF FIGURES
Figure 1: Flowchart of project activity. .........................................................................................................8
Figure 2: Example of dynamic segmentation versus project limits.............................................................11
Figure 3: Scenario 1.....................................................................................................................................12
Figure 4: Scenario 2.....................................................................................................................................12
Figure 5: Scenario 3.....................................................................................................................................13
Figure 6: Scenario 4.....................................................................................................................................13
Figure 7: Example of project selection process (4-SM-84-PM20-PM21.5)................................................16
Figure 8: Example of probability distribution for HM-1 ACOL-DG for 10 percent
Alligator A cracking. ...........................................................................................................................35
Figure 9: Example of probability distribution for HM-1 ACOL-DG..........................................................36
Figure 10: Process flowchart for LCCA......................................................................................................48
Figure 11: Example of pavement life prediction using linear regression. ...................................................55

UCPRC-TM-2007-08 ix
 LIST OF ABBREVIATIONS USED IN THE TEXT
AAC cracks Alligator A cracks
ABC cracks Alligator B cracks
ACOL Asphalt Concrete Overlay
AADTT Average Annual Daily Truck Traffic
ACOL-DG Dense-graded asphalt concrete overlay
ACOL-OG Open-graded asphalt concrete overlay
ACOL-RAC Rubberized asphalt concrete
ACOL-RACO Rubberized asphalt concrete open-graded
CapM Treatments performed under Caltrans’ Capital Preventive Maintenance Program
ChipSeal-AC Chip seal with conventional asphalt or emulsion
ChipSeal-AR Rubberized asphalt concrete chip seal
ChipSeal-PMA Polymer-modified asphalt concrete chip seal
ChipSeal-PME Polymer-modified emulsion asphalt concrete chip seal
EUAC Equivalent Uniform Annual Cost
EA Expenditure Authorization (Caltrans)
HM-1 Treatments performed under Caltrans’ Contract Highway Maintenance Program
LCCA Life-cycle Cost Analysis
PP Pavement Preservation

x UCPRC-TM-2007-08
1 PROJECT OVERVIEW

1.1 Background and Objectives

The Division of Maintenance of the California Department of Transportation (Caltrans) has identified a
need to compare the estimated performance of asphalt pavements that have undergone preservation
treatments with those that haven’t. This is a first step necessary to compare the life-cycle costs of various
pavement preservation treatments with each other, and with “control” cases meaning pavements where no
preservation treatment has been applied.

The University of California Pavement Research Center (UCPRC) has been working on Partnered
Pavement Research Center Strategic Plan Element 3.2.5 (PPRC SPE 3.2.5), titled “Documentation of
Pavement Performance Data for Pavement Preservation Strategies and Evaluation of Cost-Effectiveness of
Such Strategies,” since October 2005. The goal of PPRC SPE 3.2.5 is to use performance prediction to
recommend best pavement preservation practices, and the original objectives of PPRC SPE 3.2.5 are:

a. To track pavement preservation projects through performance databases;

b. To estimate pavement performance for various strategies and application times relative to
condition of the existing pavement; and

c. To analyze life-cycle costs and recommend optimum timing and strategy selection.

In July 2006 the Division of Maintenance provided an additional $50,000 to the Division of Research and
Innovation for the UCPRC to increase the scope of this research and to overcome problems in completing
the original objectives. The revised project objectives are as follows:
a. To identify pavement preservation treatments in the Caltrans Pavement Condition Survey (PCS)
data in the Caltrans Pavement Management System (PMS) databases;
b. To Retrieve and review PCS data and other data from available sources in order to develop
performance estimates;
c. To develop condition survey performance histories;
d. To perform statistical analyses to identify performance probabilities;
e. To analyze timing of pavement preservation treatments and their effects on relative
performance;
f. To calculate life cycle costs; and
g. To report the results of the investigation.

UCPRC-TM-2007-08 1
The revised objectives are stated in the “Work Plan for Performance Review of Pavement Preservation
Treatments” (1). Each objective, its associated deliverable, and notes regarding their completion are
presented in Table 1.

By extracting and analyzing the best possible information already in the Caltrans databases, it is expected
that present pavement preservation practices can be improved. This included documenting past practice for
pavement maintenance, and comparison with a “pavement preservation” approach to pavement
maintenance. Pavement preservation seeks the best use of maintenance resources through optimization of
the selection and timing of application of maintenance treatments by use of performance estimation and
life-cycle cost analysis.

Table 1: Project Objectives, Deliverables, and Notes

Objective Deliverables Report Notes
Section

Identify pavement List of pavement preservation (PP) 2.1, 2.2, 3.1 11 PP strategies, mostly
preservation treatments. treatment sections constructed 1993 to ACOL-DG.
2002; summary of gaps in the available
data.

Retrieve and review PCS Database of pavement performance data 3.1, 3.2, Database created
data and other data. (condition survey) extracted from database and containing data for
Pavement Condition Survey and UCPRC-TM- performance, climate, and
containing traffic and climate data. 2007-06 truck traffic.
Cost data collected based
on past Caltrans M&R
projects.
Develop condition survey Summary memorandum of pavement 2.2 and Plotted 2,500 project
performance histories. performance histories for sections in UCPRC-TM- sections (5,000 plots).
database. 2007-06

Perform statistical analysis Summary memorandum of statistical 4, 5, 6.1–6.3 718 project sections were
to identify performance analysis of pavement performance for adequate after visual
probabilities. sections in database. inspection.
Visual inspection showed
performance history did
not meet criteria on 1,800
sections.
Probability of failure for
the pooled dataset
complete for ACOL-DG
based on Alligator A+B
cracking.
Developing refined
probabilities based on

2 UCPRC-TM-2007-08
 Objective Deliverables Report Notes
Section
traffic/climate factors.

Calculate life-cycle costs. Summary memorandum of calculated 6.4 Complete.

life-cycle costs for pavement
preservation treatments and control
sections included in statistical analysis of
performance.

Analyze timing and Summary memorandum of historical 7 Complete.

relative performance. maintenance patterns and estimated
optimum timing of pavement
preservation treatments based on life-
cycle cost analysis.

Report. Report summarizing all the work This report Complete

completed.

With the additional funding, data was collected for a larger number of pavement preservation treatments in
the PCS database as well. The objectives were completed by the UCPRC working with the Division of
Maintenance Office of Pavement Preservation (OPP), with assistance from the Office of Roadway
Rehabilitation and the Office of Roadway Maintenance.

As of October 2006 it had been determined that the primary difficulties with completing the project were
those of extracting performance data from the Caltrans PCS database due to dynamic segmentation,
finding information regarding underlying pavement structures, and finding histories of rehabilitation and
maintenance necessary to explain the condition survey data.

By October 2006 the problem of dynamic segmentation had largely been solved by developing an
algorithm that appropriately weights condition survey data across the various annual network
segmentations included in the PCS database. The solution to extracting cracking data collected under
dynamic segmentation from the PCS database is documented in the technical memorandum titled
“Pavement Performance Data Extraction from Caltrans’ PMS” (UCPRC-TM-2007-06).

To solve the problem of lack of underlying structure information, a set of more than 300 sections was
created that consists of pavements that have been cored and/or evaluated with ground-penetrating radar
(GPR) to determine pavement layers and thicknesses. This set includes flexible, rigid, and composite

UCPRC-TM-2007-08 3
(concrete overlaid with asphalt) pavements. The results of the GPR study are presented in the report titled
“Pilot Project for Fixed Segmentation of the Pavement Network” (UCPRC–RR-2005-11). The condition
survey data for these sections have been extracted from the PCS database for the period of 1978 to 2004,
and performance curves for important surface distresses have been developed.

The problem of finding maintenance and rehabilitation histories to identify pavement preservation
treatments applied or not applied was solved by working with regional maintenance coordinators from the
Caltrans Office of Roadway Rehabilitation to match project construction histories in the databases they
had personally developed with the condition survey data organized by the UCPRC. A set of specific
questions regarding maintenance and rehabilitation histories was created by the UCPRC after review of
the distress performance curves for each section in the set. The list of questions was submitted to the
Division of Maintenance Office of Roadway Rehabilitation in August 2006. The regional Maintenance
advisors answered as many of these questions as they could with their existing databases in September
2006 and in follow-up meetings.

Performance estimates were then developed, using the data for flexible and composite pavement,
rehabilitation, and pavement preservation treatments for the sections for which construction histories can
be found. These performance estimates were to be used to compare life-cycle costs of rehabilitation
projects with and without pavement preservation treatments following the procedure recently developed
by the UCPRC and the Division of Design, based on the Federal Highway Administration software
RealCost. Only those segments with good cracking condition survey trends, and for which the condition
survey trends were reasonable with respect to maintenance and rehabilitation construction records, were
used for the final performance estimates. The details of the data extraction and the numbers of sections
with data used for these analyses are documented in UCPRC-TM-2007-06.

This technical memorandum is the final report completing the revised objectives and it is one part of the
documentation to be delivered under Strategic Plan Element 3.2.5 of the PPRC, which includes
“documentation of pavement performance data for pavement preservation strategies and evaluation of
cost-effectiveness of such strategies.”

4 UCPRC-TM-2007-08
 1.2 “Data-Driven” Approach Used for the Study

A key point embedded in the objectives for this project is that the performance estimates and resultant life-
cycle cost estimates and recommendations be “data-driven,” using pavement condition and construction
information in order to develop empirical probabilities of pavement performance and to analyze life-cycle
costs. No attempt was made to apply “expert judgment” where data was not available or not consistent
with expectations. This approach was applied for two reasons. First, no judgment was applied to the data
so as not to bias the life-cycle cost analysis. Second, it was desired that “expected results” based on
common wisdom and judgment be tested with actual performance data, both to review whether our
expectations were verified by the data and to identify problems with using Caltrans’ current databases to
provide good performance data useable for life-cycle cost analysis.

Adequate resources were applied to this process to attempt to complete the objectives. The results should
be considered to be close to the best information that can be extracted from the current PCS database.

1.3 Chronology of Activity

A timeline showing research activities and deliverables is shown below.

September to November 2006

• Requested pavement preservation (PP) project information from Caltrans HQ and District
offices.
• Collected information for approximately 2,500 PP projects (project dates from 1992 to 2003).
Many projects were missing information.
• Continued making requests for project information from HQ and Districts continued into early
June 2007. Help from S. Massey and District reviewers at Caltrans HQ Maintenance was
essential. They spent many hours meeting with UCPRC, answering questions, checking their
databases, and giving recommendations.

November to December 2006

• Extracted Pavement Condition Survey (PCS) data.
• Applied UCPRC algorithm to correct for dynamic segmentation.
• Identified 2,400 projects from 1988 to 2003 with sufficient information for data extraction.
• Chose the type of PP strategy for the list of projects.

UCPRC-TM-2007-08 5
February 2007
• Created PCS history plots for the 2,400 projects.
• Produced plots showing time versus Alligator A, Alligator B, Alligator A+B, and IRI
(International Roughness Index) for each project section. IRI was not used due to problems
with the data.

May 2007
• Performed labor-intensive, visual inspection of the PCS plot vs. project history for every
project (all 2,400) to confirm application of PP strategy.
• Selected usable project data for developing performance probability.
• Identified reason that section data was not usable (history mismatch, duplicate section,
insufficient history).
• Applied lane distribution factors; 718 projects became 1,421 observations.

June to July 2007

• Developed cumulative probability distributions.
• Delivered draft summary memo.

August to November 2007

• Performed life-cycle cost analysis.
• Delivered technical memo.
• Delivered database used in project to Caltrans Division of Maintenance.

1.4 Order of Work Summary

Work performed in this study consisted of three main parts: data processing, statistical analysis, and life-
cycle cost analysis (LCCA).

Data processing: This involved using data from the Caltrans Pavement Condition Survey database to
produce historical information about pavement performance. Section 2 through Section 5 describe in detail
the work performed to compile the database of useful historical pavement information.

Statistical analysis: Based on data collected from the preceding process, the Kaplan-Meier estimator from
a statistical survival model is used to generate the pavement’s probability of failure over time. The life of

6 UCPRC-TM-2007-08
the pavement is then defined as the year when over half the sample population fails. Section 6 presents the
expected pavement life determined for different pavement treatments.

Life-cycle cost analysis: With available information on pavement life, LCCA is performed for different
scenarios of pavement maintenance sequences to determine life-cycle costs of pavement preservation. A
detailed description of steps for the LCCA and an interpretation of results are presented in Section 7.

Figure 1 presents a flowchart of all processes done to complete this study. The flowchart includes a more
comprehensive presentation of the specific tasks performed along each of the three main process
categories.

UCPRC-TM-2007-08 7
 Figure 1: Flowchart of project activity.

8 UCPRC-TM-2007-08
2 PROCESS FOR SELECTING USABLE DATA

This section describes how it was determined which specific sections of roadway performance information
to retrieve. The algorithm and criteria created by UCPRC to overcome difficulties caused by the data
structure of the current PCS database are presented to show how a particular section’s performance is
determined to be useful or not for the purposes of this study.

The Caltrans HQ and District staff members who provided project construction histories for the research
are listed below. This project could not have been completed without their input.

District Name HQ Name

2 Lance Brown HQ Maint Brian Toepfer
4 Robert Carmago HQ Maint Nerissa Chin
7 Gotson Okereke HQ Maint Susan Massey
8 Basem Muallem HQ Maint Rob Marsh
8 John Cai HQ Maint Leo Mahserelli
9 John Fox HQ Maint Brian Weber
10 Alvin Mangandin HQ Maint Ron Jones
10 Long Huynh HQ Constr Jim Cotey
11 Al Herrera HQ DRI Joe Holland
11 Dave Pound HQ DRI Alfredo Rodriguez

HQ Maintenance recommended those District offices where historical project information might be
accessible. Ms. Kelly McClain in District 5 provided additional project histories after analysis of the
database, so those projects are not included in these results but are available for future study.

The project flowchart (Figure 1) shows that key steps were to extract PCS data, apply the dynamic
segmentation-weighting algorithm, plot PCS data for fixed segments of each project, and then visually
inspect segments first to determine whether PCS agrees with project construction history and then to
determine an analysis period based on reasonableness of PCS trends.

UCPRC-TM-2007-08 9
This study focused on cracking only. International Roughness Index (IRI) data were not used because
previous studies by UCPRC found several problems with the database. (2, 3)
• There was little or no IRI data between 1992 and 1997.
• There were equipment calibration problems around 2000.
• Most of the sites had a sudden drop in IRI in 2003.
• IRI shows a decreasing trend in many cases where no record of pavement maintenance could
be found.
• IRI in certain cases is very low (<60) even where there was no evidence of treatment.
Rutting was not included because it has been surveyed as a binary variable (0 if not present, 1 if present)
and did not show consistency from year to year. Similarly, transverse cracking collection practice changed
during the survey and was not consistent when data from successive years was plotted as a time history.

2.1 Dynamic Segmentation

During extraction of the PCS data, UCPRC applied an algorithm it had developed to adjust for the
dynamic segmentation used in the PCS (3).

Figure 2 shows an example of differences between project limits and PCS dynamic segments. The data are
from the truck lane on I-505, southbound, PM 1.45 to 0.00, at the Yolo–Solano county line. The thick line
shown without symbols indicates limits of the construction project being tracked through time. Triangles
indicate the limits of the segments in the PCS database in each year, using the dynamic segmentation
approach used by the PMS in the years included in the study. A + symbol indicates that the PCS rater
observed a different pavement surface in 1996 than in 1998 and afterward. Reasons for gaps in between
the rated segments are unknown.

10 UCPRC-TM-2007-08
 Project Section

PCS record of
surface Type I.

PCS record of
surface Type II.

1 2 3 4 5 6 7

Figure 2: Example of dynamic segmentation versus project limits.

Caltrans PCS pavement segments in a given year usually do not have exactly the same beginning and
ending post miles as the UCPRC-selected segments that are held consistent through the full time period.
During PCS data extraction, Caltrans PCS data segments were included in the calculation of the
percentage of the wheelpath that cracked if they fell within the four scenarios described below. The terms
used in these scenarios are:

PMSEND = End Post Mile of Caltrans PCS

segment
PMSBEG = Start Post Mile of Caltrans PCS
segment
CLBEG = Start Post Mile of construction
project segment (called UCPRC
segment)
CLEND = End Post Mile of construction
project segment (called UCPRC
segment)

UCPRC-TM-2007-08 11
The four scenarios are defined below and presented graphically in Figure 3 through Figure 6.

Scenario 1: The Caltrans PCS segment starts outside the UCPRC segment but ends within the UCPRC
segment, with the condition that the overlapping length is greater than 10 percent of the UCPRC segment,
i.e., (PMSEND-CLBEG) >= 0.1*(CLEND-CLBEG). The overlap condition was applied based on
engineering judgment, and is based on the fact that Caltrans PCS procedures call for walking evaluation of
the first 100 ft of each Caltrans PCS segment with the assumption that the rest of the Caltrans PCS
segment is similar to the first 100 ft. Greater than 10 percent overlap was desired as a minimum so that the
condition of the first 100 ft of the Caltrans PCS segment is similar to the UCPRC segment condition.

UCPRC Project Section

CLBEG CLEND

PMSBEG PMSEND
Caltrans PCS Section

Figure 3: Scenario 1.
Scenario 2: The Caltrans segment starts within the UCPRC segment, i.e., PMSBEG ≥ CLBEG. The 10
percent overlap requirement was not applied to this scenario because it was assumed that since the 100 ft
of the Caltrans PCS segment actually surveyed by walking is within the UCPRC segment, they should be
similar.

UCPRC Project Section

CLBEG CLEND

PMSBEG PMSEND
Caltrans PCS Section

Figure 4: Scenario 2.
Scenario 3: The Caltrans segment covers the entire length of the UCPRC project segment,
i.e., (PMSBEG≤CLBEG) AND (PMSEND≥CLEND).

12 UCPRC-TM-2007-08
 UCPRC Project Section
CLBEG CLEND

PMSBEG PMSEND
Caltrans PCS Section

Figure 5: Scenario 3.

Scenario 4: The Caltrans segment lies within the boundaries of the UCPRC project segment,
i.e., (PMSBEG≥CLBEG) AND (PMSEND≤CLEND).

UCPRC Project Section

CLBEG CLEND

PMSBEG PMSEND
Caltrans PCS Section

Figure 6: Scenario 4.

To associate the extracted PCS data with the construction project limits for the maintenance treatment,
data extraction and weighting was performed as described below.

1. Data Extraction Algorithm: Data was extracted from three Caltrans PCS databases (1978 to 1992,
1994 to 1997, and 1998 to 2004). For a given project, a query was set to include all PCS segments
in the database of segments to be used for further analysis that match any of the four cases
described above.

2. Weighting Algorithm: An algorithm for calculating a weighted average pavement distress extent
for a UCPRC project segment using the condition survey data for the Caltrans PCS segments in
each year was applied to the following distresses: Alligator A (AAC) cracks (percent of
wheelpath), Alligator B (ABC) cracks (percent of wheelpath) for segments selected for further

UCPRC-TM-2007-08 13
 analysis using the Data Extraction Algorithm in Step 1. The formula used to calculate the
weighted averages was:
n n
Weighted Average % wheelpath cracked = (∑ (distressi * CTsegmi _ length) /(∑ CTsegmi _ length) (1)
i i

where:
distressi = measured value of AAC or ABC for a Caltrans PCS segment that falls into one of the
four dynamic segmentation scenarios above;
CTsegmi_length = the length of the Caltrans PCS segment whose distress was reported as part of
the UCPRC segment and whose measured extent of cracking is included in calculation of UCPRC
segment cracking extent;
n = sequence number of the Caltrans segments crossing UCPRC project segment boundaries.

An example of the weighted average percent wheelpath cracked calculation is shown below, where three
Caltrans PCS segments are identified as being applicable to the UCPRC segment (based on Scenarios 1,
4, and 2, respectively from left to right).

UCPRC Project Section

Length 4 @ 40% crack

Length 8 @ 20% crack
Length 5 @ 30% crack

The weighted average for the example shown above is:

(30 × 5) + ( 20 × 8) + ( 40 × 4)
Weighted Average % wheelpath cracked = = 27.65
5+8+ 4

2.2 Project Selection

After preparing weighted distress extent for cracking for each segment in each year of the PCS database,
the data were reviewed, compared with available project construction histories, and selected for further
analysis if the construction history matched the cracking time history. Sections were selected for further
analysis using the criteria shown below.
• Use the segment if cracking drops close to zero within a year before or two years after the
identified construction date. The purpose of this grace period is that we have only the
approximate date of the start of the construction contract, not the date the construction was

14 UCPRC-TM-2007-08
 actually performed. This criterion is expected to result in conservative estimates of number of
years to predefined cracking levels.
• Check whether duplicate condition survey records are found for the segment, and eliminate
one segment record if duplicated.
• Do not use segment if significant improvement is noted in pavement condition but no records
of construction.
• Do not use segment if two years of condition survey data or less are available after a
construction project date.

An example of how these criteria were applied to project information to determine whether they should be
further analyzed is provided below. The example draws on information from a rehabilitation project
involving placement of a dense-graded asphalt concrete overlay (ACOL-DG) completed in 1996 on
4-SM-84-PM20-21.5 in Lane 1. Figure 7 contains PCS data from 1978 through 2004. The condition and
project history are characterized as follows for this study.
• Construction records show completion in 1996. Records do not indicate whether this is the
date of closeout of the contract or closing of the Caltrans Expenditure Authorization (EA).
The date is assumed not to be the construction completion date.
• The estimated rehabilitation construction date of between 1996 and 1997 correlates with
Alligator B cracking time history from the condition survey data, which shows that 15 percent
of the wheelpath cracked in 1996 and zero cracking occurred in the wheelpath in 1997. For
some projects, the beginning year was assumed even if cracking occurred after the assumed
construction year of up to 5 percent of the wheelpath, to account for sources of error, e.g.,
errors made during condition rating or rounding errors made when using the data extraction
algorithm and the condition survey-averaging algorithm.
• The portion of the total PCS history used for this project is from 1997 to 2004, i.e., the
estimated date of construction of the overlay to the last year available in the PCS database.

UCPRC-TM-2007-08 15
 Project History: ACOL-DG Rehab completed in 1996
(Lane 1 PCS)

60 2004 - End Year of

AAC
PCS data analysis
1997 - Begin Year
of PCS data
analysis

1996 - Existing
40
% Cracking

Allig. B Crack
ABC

20

AAC+A
BC

0
8

4
7

0
19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

20

20

20

20

20
Year

Figure 7: Example of project selection process (4-SM-84-PM20-PM21.5).

(Note: AAC is Alligator A cracking as percentage of the wheelpaths cracked; ABC is Alligator B
cracking; AAC+ABC is the combined cracking.)

16 UCPRC-TM-2007-08
3 SUMMARY OF DATABASE

This chapter describes how the size of the project database evolved from the study’s start to the smaller
sample database used for the performance analyses as the selection criteria described in Section 2 were
applied. This chapter describes further adjustments to the database and summarizes the projects in the final
database.

3.1 Initial Database of Projects

The UCPRC collected construction histories for 2,564 projects. A total of 1,846 projects were eliminated
for the reasons shown in Table 2. These reductions resulted in a database of 718 pavement preservation
(PP) projects identified as having adequate PCS histories for performance evaluation.

Table 2: List of Deleted Project Histories

Number of Projects Reason for Deletion
Project started in 2003 or later, so insufficient performance
941 history data.
567 Performance history does not match construction history.
178 Duplicate construction project entries.
160 No project completion date.

An additional 25 rehabilitation (control) projects and rehabilitation with pavement preservation treatment
projects from District 9 were investigated beyond the 718 projects. Unfortunately, most of these sections
either did not show distress at the time of final data collection or they had insufficient information
necessary to draw conclusions. They were not considered further in this investigation.

The remaining 718 projects were sorted by strategy (ACOL, chip seal, slurry, etc.) and program category
(CapM, HM-1, Rehab). Table 3 shows the distribution of the selected projects based on their strategy and
funding program. The funding program was needed to approximately identify asphalt overlay thicknesses,
since no other thickness information was available. The funding program was also needed to compare the
practice for placing the overlays, in terms of extent of distress, with the intention of the funding program,
either maintenance or rehabilitation. The project information did not consistently specify direction and
lanes treated. No layer thickness data for ACOL were available.

UCPRC-TM-2007-08 17
 Table 3: Project Distribution—One Observation per Selected Project
PP Strategy CapM HM-1 Rehab (unknown) Total
ACOL-DG 31 275 36 79 421
ACOL-OG – 51 – 18 69
ACOL-RAC – 8 – 9 17
ACOL-RACO – 2 – 2 4
ChipSeal-AC – 79 – – 79
ChipSeal-AR – 6 – 5 11
ChipSeal-PMA – 1 – – 1
ChipSeal-PME – 7 – 17 24
CrackSeal – 2 – 1 3
DigOut – 10 – – 10
SlurrySeal – 6 – 2 8
(blank) 3 14 – 54 71
Total 34 461 36 187 718
Note: 127 of the 718 data points are composite sections (approximately 18 percent).
Projects were identified as composite if any PCC distress was noted in PMS (from 1978).

It should be noted that a large majority of the asphalt overlays come from the HM-1 and CAPM programs,
which have strict limits on the thickness of overlay that can be applied. For this reason, it was decided that
the lack of overlay thickness data in the Caltrans records did not prevent preliminary estimates of
performance from being made. It should also be noted that “digout,” shown in Table 3 and elsewhere in
this technical memorandum, is not a pavement preservation treatment, but instead a repair performed on a
distressed pavement. It consists of milling out a portion of the thickness of cracked asphalt concrete in the
wheelpaths and inlaying it with asphalt concrete. Sometimes digouts are performed prior to placement of
an overlay, and sometimes they are done without accompanying work. The construction project data
available in the PMS and in personal databases kept by District and HQ staff were often not clear as to
what was done with the digouts. Identification of the work performed is not standardized or based on
codes, and is often included only as a comment in the “Notes” part of the PMS database.

3.2 Database of Observations

3.2.1 Adjusting for Traffic

The 718 projects include multiple lanes and two directions. By applying lane distribution factors to each
lane of each project, more observations (each lane having different performance linked to different traffic)
were made available for developing failure probabilities. Lane distribution information for Annual
Average Daily Truck Traffic (AADTT) was applied from Lu et al. (4). Lane distribution factors are shown
in Table 4, and it was assumed that these factors applied to all projects.

18 UCPRC-TM-2007-08
 Table 4: Average Lane Distribution Factor of AADTT for Highway with
Different Number of Lanes in One Direction (4)
Number of Lanes
Lane 1 Lane 2 Lane 3 Lane 4 Lane 5 Lane 6
in One Direction
1 1.00 0.00 0.00 0.00 0.00 0.00
2 0.09 0.91 0.00 0.00 0.00 0.00
3 0.06 0.33 0.67 0.00 0.00 0.00
4 0.05 0.09 0.44 0.49 0.00 0.00
5 0.00 0.05 0.27 0.58 0.25 0.00
6 0.00 0.00 0.00 0.16 0.50 0.34

Applying the lane distribution factors to those projects with multiple lanes produced a total of
1,508 observations from the 718 projects, as shown in Table 5.

Table 5: Distribution of Observations (Considering Multiple Lanes and Both Directions)

PP Strategy CapM HM-1 Rehab Total
ACOL-DG 136 600 250 986
ACOL-OG 5 131 1 137
ACOL-RAC 39 29 10 78
ACOL-RACO 3 3 8 14
ChipSeal-AC – 181 – 181
ChipSeal-AR – 21 – 21
ChipSeal-PMA – 2 – 2
ChipSeal-PME – 46 – 46
CrackSeal – 6 – 6
DigOut – 17 – 17
SlurrySeal – 20 – 20
Total 183 1,056 269 1,508

3.2.2 Adjusting for Missing Lane Prior Condition Data

Inner lanes are not surveyed every year, and projects lacking PCS data were deleted from the database.
PCS data prior to treatment were required to evaluate the effect of existing cracking on the performance of
treatments. Due to the fact that PCS data were not recorded every year, PCS data was sought that had
existing cracking within a few years before construction. By evaluating distress data three to five years
before the treatment (with at least one PCS data point in that period), it was possible to obtain the number
of observations shown in Table 6.

Table 6: Number of Observations with Missing PCS Data

Tracked Number Number of Observations Number of Observations
of Years Deleted Remaining
3 years prior 90 (6.0 percent) 1,418
4 years prior 42 (2.8 percent) 1,466
5 years prior 33 (2.2 percent) 1,475

UCPRC-TM-2007-08 19
It was decided to use four years as a limit after consideration of three, four, and five years. If three years
were used, the number of observations would have been approximately halved (reduced from 6% to 2.8%)
and a five-year limit would only have reduced the deleted observations by another 0.6%. This produced a
revised distribution as shown in Table 7.

Table 7: Distribution of Observations After Adjusting for Missing Prior Condition Data
PP Strategy CapM HM-1 Rehab Total
ACOL-DG 136 583 235 954
ACOL-OG 5 127 1 133
ACOL-RAC 39 29 10 78
ACOL-RACO 3 3 8 14
ChipSeal-AC – 175 – 175
ChipSeal-AR – 21 – 21
ChipSeal-PMA – 2 – 2
ChipSeal-PME – 46 – 46
CrackSeal – 6 – 6
DigOut – 17 – 17
SlurrySeal – 20 – 20
Total 183 1,029 254 1,466

Further adjustment was needed for observations (lanes within projects) that have no traffic data (AADTT)
available. This resulted in a database containing 1,421 unique segments consisting of a single lane with a
construction history, a condition survey history both before and after construction, and truck traffic data
that was to be used for developing failure probabilities. Projects included in the database and used for
further analysis span an actual period from 1988 to 2003, so they reflect the practices, policies, and
procedures (within HQ and the Districts) during that time. The final database is summarized in Table 8.

20 UCPRC-TM-2007-08
 Table 8: Distribution of Observations for Final Set of Data Used for
Failure Probability Analysis
PP Strategy CapM HM-1 Rehab Total
ACOL-DG 132 567 222 921
ACOL-OG 5 127 1 133
ACOL-RAC 39 29 10 78
ACOL-RACO 3 3 8 14
ChipSeal-AC 169 169
ChipSeal-AR 19 19
ChipSeal-PMA 2 2
ChipSeal-PME 44 44
CrackSeal 6 6
DigOut 17 17
SlurrySeal 18 18
Total 179 1,001 241 1,421

A higher proportion of the composite sections (asphalt overlays on PCC pavement) were rehabilitation
sections as opposed to maintenance sections than is seen in the full data set, as shown in Table 9.

Table 9: Distribution of Observations for Composite Sections

Action Category CAPM HM-1 REHAB Total
ACOL-DG 26 14 53 93
ACOL-OG 0* 9 0 9
ACOL-RAC 3 4 0 7
ACOL-RACO 0 2 8 10
ChipSeal-AC – 4 – 4
ChipSeal-AR – 0 – 0
ChipSeal-PMA – 0 – 0
ChipSeal-PME – 8 – 8
CrackSeal – 2 – 2
DigOut – 0 – 0
SlurrySeal – 8 – 8
Total 29 51 61 141
* Zero values indicate that there were observations for this program/strategy but none was composite.

It will be observed that many of the strategies have a small number of observations on which to base
conclusions. Note also that while composite pavements comprised 127 of the 718 projects in the database,
only 141 (10 percent) of the 1,421 observations were composite sections.

UCPRC-TM-2007-08 21
3.2.3 Distribution of Projects and Observations by District
Table 10 shows the distribution of observations (unique lanes with complete data) by program type and
District in the dataset. It shows that:
• Over half of the HM-1 projects in the database are in District 2, and nearly a quarter are in
District 8.
• CapM projects are distributed evenly between Districts 2 and 4.
• The majority of rehabilitation projects are in urban districts, including Districts 4, 7, and 8.

Table 10: Distribution of Observations by Program Type and District

Dis trict
P rogra m Type 1 2 3 4 5 6 7 8 9 10 11 12 Grand Total
CAP M 72 71 1 16 19 179
HM1 26 554 2 74 19 46 11 227 11 18 10 3 1001
REHAB 67 1 29 101 39 4 241
Grand Total 26 693 3 174 20 46 128 285 15 18 10 3 1421

The distribution of PP strategies that could be obtained by district is presented below. Table 11 shows that:
• For conventional AC overlays (ACOL-DG), over half are in District 2 and nearly half are in
Districts 4, 7, and 8.
• For rubberized overlays (ACOL-RAC), nearly all are in Districts 4, 7, and 8.
• For conventional chip seals, nearly three-quarters are in District 2.
• Other strategies have small sample sizes.

Table 11: Distribution of Observations by Strategy and District

Dis trict

P P S tra te gy 1 2 3 4 5 6 7 8 9 10 11 12 Total by
Strategy
ACOL-DG 2 513 94 7 22 100 167 6 7 3 921
ACOL-OG 20 9 3 33 6 6 2 49 3 2 133
ACOL-RAC 4 32 26 16 78
ACOL-RACO 9 5 14
ChipS e a l-AC 129 2 38 169
ChipS e a l-AR 2 5 10 2 19
ChipS e a l-P MA 2 2
ChipS e a l-P ME 4 7 2 18 4 1 8 44
Cra ckS e a l 2 4 6
DigOut 17 17
S lurryS e a l 14 4 18
Total by District 26 693 3 174 20 46 128 285 15 18 10 3 1421
Percent by District 2% 49% 0% 12% 1% 3% 9% 20% 1% 1% 1% 0% 100%

22 UCPRC-TM-2007-08
The database is a cross section of conditions statewide. For example, District 2 represents northern
California, with four climate regions and a combination of low-volume routes and high-volume Interstate-
5. In contrast, District 8 provides data from a dry/desert climate in southern California with both urban and
rural routes. District 4 provides data for urban and rural routes in a coastal environment.

To check the consistency in the composition of the original set of projects with the final database across
districts, the original database of 2,564 projects is compared to the selected 718 projects and to the final
data set of the 1,421 observations as shown in Table 12, which shows that:
• Other than District 2, the proportion of projects is fairly consistent for each of the data sets.
• Projects in District 2 became a higher percentage of observations during the screening process
because of the good agreement between project histories and PCS data.

Table 12: Distribution of Projects and Observations by Districts

Dis trict
Total
1 2 3 4 5 6 7 8 9 10 11 12
Origina l 2564 P roje cts 3% 24% 3% 18% 5% 8% 7% 19% 4% 5% 2% 2% 100%
718 P roje cts 2% 44% 0% 15% 1% 3% 7% 23% 2% 1% 1% 1% 100%
1421 Obs e rva tions 2% 49% 0% 12% 1% 3% 9% 20% 1% 1% 1% 0% 100%

UCPRC-TM-2007-08 23
24 UCPRC-TM-2007-08
4 SUMMARY OF PRACTICE 1988 TO 2003

This section summarizes the pavement preservation treatments used by Caltrans in the period of 1988 to
2003 and the timing of their application with respect to extent of cracking. The summary is based on data
collected and selected by the process described in the preceding sections.

From the adjusted database of observations, average (mean) existing distress levels before treatment were
determined for each pavement preservation (PP) strategy. The results are presented in Table 13.

Table 13: Mean Existing Cracking Levels When Each PP Strategy Performed
A+B Crack (%) A Crack (%) B Crack (%)
PP Strategy Mean Std. Dev. Mean Mean
ACOL-DG 26 24 10 16
ACOL-OG 25 20 10 15
ACOL-RAC 32 30 12 20
ACOL-RACO 3 5 1 1
ChipSeal-AC 17 18 8 10
ChipSeal-AR 18 15 5 13
ChipSeal-PMA 21 3 6 15
ChipSeal-PME 11 11 5 6
CrackSeal 28 22 9 19
DigOut 48 28 13 36
SlurrySeal 2 3 2 1

Existing cracking levels sorted by PP strategy as well as funding program category (CapM, HM-1, and
Rehabilitation) are presented in Table 14. The information shown in Table 13 and Table 14 indicates the
following:
• For rehabilitation using ACOL-DG, the database of observations shows a mean value of
existing Alligator B cracking (16 percent) that is below the PMS prioritization table trigger
value of 25 percent. Actual overlay thickness is not known, and it is unknown what other
activities were done on these projects in conjunction with the overlays that might affect
performance.
• In general, HM-1 program activities appear likely to be performed at higher existing cracking
than these strategies are intended, and at cracking levels that are similar to those of
rehabilitation projects. Drawing further inferences from this observation should be done
cautiously because of limitations and uncertainties in the database, such as the use of PCS data
acquired as long as three years before the placement of treatments.

UCPRC-TM-2007-08 25
 • Despite differences in the sizes of samples for various strategies and programs, treatment
strategies show relatively consistent levels of existing cracking (A, B, and A+B) when the
strategy was applied. Older Caltrans priority assignments for rehabilitation or maintenance
treatment (from the period applicable to this memorandum) indicate various identifiers for
“unacceptable” conditions with respect to cracking of 30 percent Alligator B cracking, or 10
percent Alligator B cracking combined with more than 10 percent patching, or no Alligator B
cracking combined with more than 15 percent patching. Patching is presumed to indicate
repaired alligator cracking. One outlier in this pattern is existing cracking ACOL-RAC
rehabilitation, which appears to be applied at higher levels of cracking than ACOL-DG.
Again, caution should be noted with regard to this inference because of the small sample of
projects (10).
• For digouts and crack sealing, existing cracking levels are near or well above the 25 percent
Alligator B cracking trigger used for rehabilitation.

These results, in particular the thin overlays and chip seals placed through the HM-1 program, indicate
that Caltrans maintenance practice during the period of 1988 to 2003 is not what is considered pavement
preservation because of the high levels of cracking present at the time of treatment. The results also
suggest that many overlays will fail due to reflection cracking, since cracks have already appeared at the
pavement surface at the time of overlay. Although well-calibrated mechanistic models of reflection
cracking are only now being developed, most studies in the literature indicate that overlays placed over
pavement with surface cracks will have shorter lives than pavements without surface cracking, because
cracks result in larger strains at the bottom of the overlay due to strain concentration.

26 UCPRC-TM-2007-08
 Table 14: Existing Cracking Data for Each PP Strategy and Program
Program Type
Existing Cracking  POOLED DATA
PP Strategy CAPM HM‐1 REHAB
Type
Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev.
Alligator A 9 13 10 11 9 11 10 11
ACOL‐DG Alligator B 12 14 17 18 16 20 16 18
Alligator A+B 21 22 27 24 25 26 26 24
Alligator A 12 7 10 11 8 10 11
ACOL‐OG Alligator B 16 8 14 16 19 15 15
Alligator A+B 28 15 25 21 27 25 20
Alligator A 10 12 9 11 26 12 12 13
ACOL‐RAC Alligator B 13 16 21 19 45 29 20 21
Alligator A+B 23 25 30 24 71 34 32 30
Alligator A 3 2 4 6 0 1 3
ACOL‐RACO Alligator B 4 6 2 1 0 1 3
Alligator A+B 7 7 6 7 0 3 5
Alligator A 8 10 8 10
ChipSeal‐AC Alligator B 10 12 10 12
Alligator A+B 17 18 17 18
Alligator A 5 4 5 4
ChipSeal‐AR Alligator B 13 13 13 13
Alligator A+B 18 15 18 15
Alligator A 6 1 6 1
ChipSeal‐PMA Alligator B 15 5 15 5
Alligator A+B 21 3 21 3
Alligator A 5 5 5 5
ChipSeal‐PME Alligator B 6 8 6 8
Alligator A+B 11 11 11 11
Alligator A 9 7 9 7
CrackSeal Alligator B 19 15 19 15
Alligator A+B 28 22 28 22
Alligator A 13 15 13 15
DigOut Alligator B 36 22 36 22
Alligator A+B 48 28 48 28
Alligator A 2 3 2 3
SlurrySeal Alligator B 1 1 1 1
Alligator A+B 2 3 2 3

UCPRC-TM-2007-08 27
28 UCPRC-TM-2007-08
5 FACTORS INFLUENCING PERFORMANCE

This chapter describes statistical analysis that helped determine the factors (both traffic and
environmental) that correlate with the cracking performance of the pavement preservation strategies.
Based on a UCPRC study of cracking initiation and progression models using the Washington State
Department of Transportation (WSDOT) database [6], temperature, precipitation, and AADTT/lane are
factors found to be important in future pavement performance. Existing Alligator B cracking (ExABC) is
also used as one of the main factors affecting pavement performance, for two reasons. First, existing
cracking condition was found to be one of the significant factors in the WSDOT study. Second, Caltrans
uses Alligator B cracking as a trigger for maintenance and rehabilitation.

5.1 Check Correlation of Factors

If two explanatory factors are highly correlated with each other (with correlation value close to 1 or -1),
that means the effect of one factor can probably be demonstrated by the presence of another factor, which
reduces the significance of the original factor itself. In this event, a transformation process to combine the
two factors into a new third factor is usually performed to try to capture the effect of the original two
factors by just this newly created one. Table 15 presents a correlation table for five factors (developed
using an Microsoft ExcelTM function) affecting pavement performance and their correlation value against
all other factors. It can be seen in the table that these factors are not highly correlated with each other.
Therefore, a transformation to combine any two or more factors is not necessary. The exceptions are the
annual number of freeze-thaw cycles and the minimum temperature, which are highly correlated with the
number of freeze-thaw cycles expected to increase in colder regions. Therefore only one of these two
variables can be used or they must be combined. Based on this comparison and on previous findings from
UCPRC work (5), minimum annual temperature was selected for use because it was previously found to
be a more significant factor in predicting cracking initiation.

Min Temp: Average minimum annual temperature in degrees Celsius (ºC)

Precip: Average annual precipitation (mm)
Ftavg: Average annual number of freeze-thaw cycles
exABC: Existing Alligator B cracking prior to treatment
AADTT: Average annual daily truck traffic

UCPRC-TM-2007-08 29
 Table 15: Correlation Table—Factors Affecting Pavement Performance

Min Temp Precip Ftavg exABC AADTT

Min Temp 1
Precip –0.05936 1
Ftavg –0.96333 0.146822 1
exABC 0.021578 –0.11972 –0.0122 1
AADTT 0.224235 –0.16123 0.25772 0.10635 1

5.2 Analysis of Factors

Table 16 presents the range, median, and mean values of the minimum annual temperature (MinTemp),
average annual precipitation (Precip), and AADTT/lane (AADTT). The table also shows the distribution
of observations within each range of the existing cracking (exABC).

Shown below are “factors” affecting performance and descriptive statistics for project locations. Factors
in the analysis are defined from the variables described in Section 5.1 as follows, based on their median
values:

MinTemp was divided into two categories:

cold = minimum annual air temperature less than or equal to -7.6°C, and
hot = minimum annual air temperature greater than -7.6°C.

Precip was divided into two categories:

dry = annual precipitation less than or equal to 464 mm/year, and
wet = annual precipitation greater than 464 mm/year.

AADTT was divided into three categories:

low traffic = less than or equal to 125 AADTT/lane, and
high traffic = greater than 125 AADTT/lane

exABC was divided into three categories

existing Alligator B cracking = less than or equal to 10%; more than 10% and less than or equal to
25%; more than 25%
The sample size for existing cracking in the three categories is shown in Table 16.

30 UCPRC-TM-2007-08
 Table 16: Descriptive Statistics of Factors Affecting Cracking Performance in Data
Min. Annual
Descriptive Statistic for Temperature Precipitation
Climate and Traffic Factors (ºC/ ºF) (mm / inches) AADTT/Lane
Min -26.3 / -15.3 0/0 0
Max 4.6 / 40.3 1,977 / 77.8 6,567
Median -7.6 / 18.3 464 / 18.3 125
Average -9.4 / 15.1 594 / 23.4 355
Distribution of Alligator B Cracking Extents
Existing Alligator B Crack ≤ 10% 10% < x ≤ 25% > 25%
Sample Size 743 361 317

Analysis to determine which factors are significant was based on data for the ACOL-DG strategy in the
HM-1 program category because this data set had the highest number of observations, and overlay
thicknesses are nearly always less than 60 mm (0.2 ft), thus ensuring that there was not a wide range of
overlay thicknesses in the data set. The Cox Survival model (a regression method that does not try to fit a
fixed mathematical form) was used to obtain the Z-statistic and p-value for determining the significance of
each factor independently for: minimum annual temperature (MinTemp), annual precipitation (Precip),
existing Alligator B cracking (exABC), and traffic (AADTT). A p-value of 0.10 or less is typically
regarded as indicating that an explanatory variable is significant, while a p-value of 0.05 indicates that an
explanatory variable is highly significant.

Table 17: Statistics of Factors Affecting Pavement Performance in Data

p-value
Variable Std. Error Z P>|z| 95% Conf. Interval
MinTemp 0.1576932 0.38 0.703 0.7904394 1.417407
Precip 0.2038066 2.59 0.01 1.092777 1.901946
exABC 0.1056617 2.45 0.014 1.043213 1.459346
AADTT 0.1639774 1.32 0.186 0.9163076 1.566823

Table 17 shows that when the four factors are grouped to check for statistical significance, only “Precip”
and “exABC” are significant at a 95 percent confidence level (based on P>|z| values), meaning that the p-
value is less than 0.05. The least significant variable, “MinTemp,” was removed and the remaining three
factors were tested again to see if their significance increased.

UCPRC-TM-2007-08 31
 Table 18: Statistics of Factors Affecting Pavement Performance in Data without MinTemp
p-value
Variable Std. Error Z P>|z| 95% Conf. Interval
Precip 0.2016425 2.56 0.01 1.088609 1.889077
exABC 0.1051345 2.43 0.015 1.041001 1.455048
AADTT 0.1602786 1.47 0.14 0.9379399 1.573247

Based on the P>|z| values in Table 18, “AADTT” continued to not be statistically significant even after
removing “MinTemp.” This is believed to be due to an insufficient dataset that is unable to capture the
significance of the effect of traffic load to pavement cracking. It may also be necessary to split AADTT
using some value other than the median to increase its significance. Based on widely accepted previous
findings and engineering judgment on the significance of traffic loading to pavement performance, it is
important to keep traffic as an explanatory variable despite the low p-value.

Results from this analysis can be compared with previous studies regarding the effects of temperature. The
life of thin dense-graded overlays with respect to fatigue cracking increases as the minimum annual
temperature increases, according to the cracking initiation model that UCPRC developed using WSDOT
data.

Statistical tests (t-test) also indicated that low temperatures have a greater impact on the initiation of
cracking than do maximum temperatures, as shown in Table 19. (5)

Table 19: Significance Level of Temperature Factor in Pavement Performance

Variable Coefficient t-statistic
Tmax* -5.63E-02 -1.21E+01
Tmin** -1.88E-01 -2.43E+01
* Tmax: Average monthly maximum temperature of the hottest month (July, in oC).
** Tmin: Average monthly minimum temperature of the coldest month (December, in oC).

With regard to crack propagation, a UCPRC study based on WSDOT data also found that climate
variables, particularly yearly precipitation and minimum temperature, play a significant role: the higher
the yearly precipitation, the higher the rate of crack progression. Higher minimum temperatures reduce the
rate of crack progression (6). The reason for the apparent lack of correlation between temperature and
cracking performance in the database assembled for this study is not known.

Based on this analysis, the factors used for classifying the Alligator cracking performance of pavement
preservation strategies in the current study are: existing Alligator B cracking, annual precipitation
(dry/wet), and traffic (high/low).

32 UCPRC-TM-2007-08
6 PAVEMENT LIFE RESULTS

This chapter first describes the process followed and the assumptions made to derive the life of pavement
treatments. It then presents the projected pavement life for the studied sections based on the statistical
analysis and general observations. At the end of the section, the cost of each type of pavement treatment is
presented to demonstrate how Caltrans can select the most cost-effective choice based on life-cycle cost
rather than only on expected pavement service life.

6.1 Procedures Used to Develop Probability of Failure

Knowing the level of cracking in each of the years within the analysis period is required, to determine the
likelihood of failure at specific points in time after construction. The following conditions were applied to
the PCS condition survey data of each of the selected projects.
• Once the beginning and ending years are defined for the project life, it is expected that the
cracking (sum of percentage of wheelpath with Alligator A and B cracking) will be non-
decreasing during the analysis period.
• If cracking data are missing, an interpolated value is calculated between the year before and
the year after the missing year.
• If the level of cracking drops in a year within the analysis period, then the cracking level from
the previous year is used.

The probability density function (PDF) under the context of this study is the point estimate of the
probability of reaching the threshold of pavement failure, e.g., 10 percent A+B Alligator cracking
(combined percentage of wheelpath with Alligator A and B cracking), in a given year, which is designated
as Pr(x=T).

x = time of interest (the particular year after treatment)

T = Time in year

The cumulative density function (CDF) is the probability of pavement failure at the predefined threshold
since the completion of treatment, Pr(x<T).

UCPRC-TM-2007-08 33
The Kaplan-Meier estimator is used to estimate the failure probability of the pavement. It is the closest to
a pure empirical estimator that can take “censored” data into account. An example of censored data in this
study includes sections that lasted longer than the time period available in the PCS database. (A more
detailed description of censored data is included in Appendix B: Definition of Data Censoring.) The
Kaplan-Meier estimator is often used to measure the proportion of patients’ surviving for a certain amount
of time after an operation, when a number of the patients have not yet died at the time of the study. That
scenario is very similar to this study, which attempts to measure the life of a pavement after it has been
treated.

The formula for computing the Kaplan-Meier estimate of the survivor function is:
k n −dj
Sˆ (t ) = ∏ ( j ) For tj-1 ≤ t <tj (2)
j =1 nj
where, nj = number of pavements not failed and that are still being observed (termed “at risk”)
before tj and
dj = number of pavements failed in the time interval tj-1≤ t < tj.

The failure probability of an item before t is then Pr(x<T) = 1-Ŝ(t).

An example is shown in Figure 8 for HM-1 ACOL-DG projects. (For details, see Appendix C: Example
Calculation of Probability of Failure Distribution for Figure 8.) The Cumulative Distribution Function
(CDF) for Alligator A cracking is indicated by the solid line. Each bar indicates the probability of failure
for each year. In this example, all of the sections (n) in Year 9 failed (d), so the CDF reaches a value of 1.0
in that year. In this example, using the 50th percentile CDF as the expected pavement life means that half
the projects will last seven years or less after construction, and the other half will last longer than seven
years.

34 UCPRC-TM-2007-08
 exABC<10%, Wet, Low AADTT - Fail @ 10%

0.8
Probability of Failure

0.6

0.4

0.2

0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Year after M&R
AAC PDF AAC CDF

Figure 8: Example of probability distribution for HM-1 ACOL-DG for 10 percent Alligator A
cracking.

A second example is presented to demonstrate the situation where the 50th percentile CDF is never
reached in the scenario, shown in Figure 9. This is a scenario with ACOL-DG HM-1 treatment where the
failure is defined as 25 percent Alligator A cracking. The highest CDF reached 45 percent by Year 12 but
it never reached the median failure requirement of 50 percent, hence there was no median pavement life
calculated for this particular treatment under this given set of conditions. Because there were sections that
had not failed (n) in the years after Year 12, but there were no more observations in later years, the CDF
stops in Year 12. The last CDF value for cases such as this represents the probability of failure in that year
and beyond, for which there were no observations. For this example, the high value shown in Year 13
jumps to more than 60 percent because it includes the CDF of all years beyond Year 12. This indicates
that more than 60 percent of the sections in the data set did not reach 25 percent Alligator A cracking by
the end of the twelfth year.

UCPRC-TM-2007-08 35
 exABC<10%, Dry, Low AADTT - Fail @ 25%

0.8
Probability of Failure

0.6

0.4

0.2

0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
AAC PDF AAC CDF Year after M&R

Figure 9: Example of probability distribution for HM-1 ACOL-DG.

6.2 Summary by Strategy and Program

Table 20 shows the pooled expected life (from the cumulative probability) for each type of PP strategy.
Presented in the table is the expected life (number of years) to last until 10 percent and 25 percent cracking
(B and A+B, where A+B is the percentage of Alligator A plus the percentage of Alligator B) after
applying each strategy. Values shown are the 50th percentiles from the cumulative probability
distributions. N/A indicates that the number of failures (reaching 10 percent or 25 percent cracking) did
not reach 50 percent of the number of observations in the dataset for a given PP strategy. The following
observations can be made from Table 20.

36 UCPRC-TM-2007-08
 • Trends are reasonable and consistent. Expected life to 25 percent cracking is always longer
than to 10 percent, and A+B always has a shorter life than B alone.
• For ACOL-DG, Rehabilitation projects last much longer than HM-1 projects.
• Chip seals provide four to six years of service before reaching 10 percent Alligator B.
• Slurry seals provide four years of service before 10 percent Alligator B.
• The time to increase from 10 percent to 25 percent Alligator B cracking is one to three years.
This may have implications for programming pavement rehabilitation, where delays occur
between the condition survey, programming, and actual construction.

Table 20: Expected Life for Each PP Strategy and Program Showing Years
to Failure Levels, not Pre-Existing Cracking
Alligator B Cracking A+B Cracking
Sample
PP Strategy Years to Years to Years Years
Size
10% 25% to 10% to 25%
ACOL-DG ALL 921 7 10 5 7

ACOL-DG HM-1 567 5 8 4 6

ACOL-DG REH 222 10 12 9 11

ACOL-DG CAPM 132 *N/A N/A N/A N/A

ACOL-OG HM-1 127 6 N/A 6 6

ACOL-RAC HM-1 29 10 N/A 8 N/A

ACOL-RACO HM-1 3 N/A N/A 4 N/A

ChipSeal-AC HM-1 169 6 N/A 3 8

ChipSeal-AR HM-1 19 4 5 3 4

ChipSeal-PMA HM-1 2 N/A N/A 1 N/A

ChipSeal-PME HM-1 44 N/A N/A 4 N/A

Crack Seal HM-1 6 N/A N/A N/A N/A

DigOut HM-1 17 2 N/A 1 N/A

Slurry Seal HM-1 18 4 6 3 4

* N/A indicates that the number of failures did not reach 50 percent of the number of observations
within each strategy.

UCPRC-TM-2007-08 37
 6.3 Summary of Factors Affecting Performance

Table 21 through Table 29 present the average (50th percentile) life expectancy for each type of pavement
preservation strategy and program at the failure trigger levels. The tables show life expectancy for various
levels of existing cracking, for wet or dry conditions, and for high or low traffic levels.

For ACOL-DG HM-1, the number of years to 10 percent Alligator B cracking is close to five years for all
traffic and precipitation levels where existing Alligator B cracking is less than or equal to 25 percent.

The effect of traffic level on the number of years to 10 percent Alligator B cracking is reasonable, with
high traffic resulting in one to two years of less service where existing Alligator B cracking is less than or
equal to 25 percent.

Results for precipitation contradict what is expected: Wetter conditions are anticipated to result in shorter
years of service life. This may be an artifact of using the median level of rainfall to produce comparably
sized datasets for analysis. For existing cracking, the trend is reasonable: Increasing levels of cracking
usually result in fewer years to reach the 10 percent and 25 percent cracking levels. These observations
apply to Alligator A, B, and A+B cracking.

For chip seals (AC), where there is low existing Alligator B cracking (below 10 percent), chip seals last at
least five years (until 10 percent Alligator B occurs). Increasing traffic can significantly reduce service
life. There is no clear effect of precipitation in the presence of higher cracking (10 to 25 percent).

Table 21: ACOL-DG HM-1

Alligator A+B Crack Alligator A Crack Alligator B Crack
Year to Year to Year to Year to Year to Year to
Existing B Crack
PP Strategy Dry/Wet Traffic 10% 25% 10% 25% 10% 25%
Pre-Construction
crack crack crack crack crack crack
Low 3 5 10 5 6
Dry
High 3 5 4
x ≤10%
Low 6 7 7 9 7 11
Wet
High 6 7 8 11 6 12
Low 3 5 12 15 5 6
Dry
ACOL-DG- High 3 5 7 4 5
10%<x≤25%
HM1 Low 4 7 7 7 7
Wet
High 4 5 6 5 7
Low 3 6 6 4 7
Dry
High 2 3 6 9 3 4
x> 25%
Low 3 4 4 9 8
Wet
High 3 6 6 4

38 UCPRC-TM-2007-08
 Table 22: ACOL-DG CapM
Alligator A+B Crack Alligator A Crack Alligator B Crack
Existing B
Year to Year to Year to Year to Year to Year to
Action Crack Pre-
Dry/Wet Traffic 10% 25% 10% 25% 10% 25%
Category Constructi
crack crack crack crack crack crack
on
Low 8 8 4 8
Dry
High 6 7 6
x ≤10%
Low 5 5
Wet
High
Low 6 7 5
Dry
ACOL-DG- 10%<x≤2 High
CAPM 5% Low 4 7 7 5
Wet
High
Low
Dry
High
x> 25%
Low
Wet
High

Table 23: ACOL-DG Rehab

Alligator A+B Crack Alligator A Crack Alligator B Crack
Year to Year to Year to Year to Year to Year to
Existing B Crack
PP Strategy Dry/Wet Traffic 10% 25% 10% 25% 10% 25%
Pre-Construction
crack crack crack crack crack crack
Low 11 11
Dry
High 7
x ≤10%
Low 4 6 5 6 7
Wet
High 6
Low 8 11
Dry
ACOL-DG- High 10 11 11 12
10%<x≤25%
REH Low
Wet
High 6 9 9
Low 9 11 11 10 12
Dry
High 8 12 12 12 8 13
x> 25%
Low
Wet
High 6 6

Table 24: ACOL-OG HM-1

Alligator A+B Crack Alligator A Crack Alligator B Crack
Year to Year to Year to Year to Year to Year to
Existing B Crack
PP Strategy Dry/Wet Traffic 10% 25% 10% 25% 10% 25%
Pre-Construction
crack crack crack crack crack crack
Low 3
Dry
High 5 5
x ≤10%
Low
Wet
High
Low 2 3 3 3
Dry
High 2 4 4 5 4
ACOLOG 10%<x≤25%
Low
Wet
High 6 7
Low 2 3 5 2 3
Dry
High 3 5 4 5
x> 25%
Low
Wet
High 2 3 2 9

UCPRC-TM-2007-08 39
 Table 25: ACOL-RAC HM-1
Alligator A+B Crack Alligator A Crack Alligator B Crack
Existing B Crack Year to Year to Year to Year to Year to Year to
PP Strategy Dry/Wet Traffic
Pre-Construction 10% crack 25% crack 10% crack 25% crack 10% crack 25% crack
Low
Dry
High
x ≤10%
Low 6 6
Wet
High 2 2 4
Low
Dry
High
ACOLRAC 10%<x≤25%
Low 8 8 10
Wet
High 1 2
Low
Dry
High 5
x> 25%
Low
Wet
High

Table 26: ACOL-RACO HM-1

Alligator A+B Crack Alligator A Crack Alligator B Crack
Existing B Crack Year to Year to Year to Year to Year to Year to
PP Strategy Dry/Wet Traffic
Pre-Construction 10% crack 25% crack 10% crack 25% crack 10% crack 25% crack
Low
Dry
High 5 4 4 4
x ≤10%
Low
Wet
High
Low
Dry
High
ACOLRACO 10%<x≤25%
Low
Wet
High
Low
Dry
High
x> 25%
Low
Wet
High

Table 27: Chip Seal

Alligator A+B Crack Alligator A Crack Alligator B Crack
Existing B Crack Year to Year to Year to Year to Year to Year to
PP Strategy Dry/Wet Traffic
Pre-Construction 10% crack 25% crack 10% crack 25% crack 10% crack 25% crack
Low 4 9
Dry
High 6 7 6
x ≤10%
Low 5 13 13 9
Wet
High 3 5 5 5 6
Low 1 3 1 3
Dry
High 1 2 2 5
CS-AC 10%<x≤25%
Low 2 6 6 3 7
Wet
High 2 3 5 2 4
Low 7 12 12 12 14 14
Dry
High 1 2 2 2 3
x> 25%
Low 1 2 2 1 7
Wet
High 1 3 7 1 3

40 UCPRC-TM-2007-08
 Table 28: Digout
Alligator A+B Crack Alligator A Crack Alligator B Crack
Existing B Crack Year to Year to Year to Year to Year to Year to
PP Strategy Dry/Wet Traffic
Pre-Construction 10% crack 25% crack 10% crack 25% crack 10% crack 25% crack
Low
Dry
High
x ≤10%
Low 1 1
Wet
High
Low 2 2
Dry
High 2 2
Digout 10%<x≤25%
Low 1 2 1 2
Wet
High 1 1 1 1 1 1
Low 1 1
Dry
High
x> 25%
Low 1 2 2
Wet
High 1 1 2 2 1 1

Table 29: Slurry

Alligator A+B Crack Alligator A Crack Alligator B Crack
Existing B Crack Year to Year to Year to Year to Year to Year to
PP Strategy Dry/Wet Traffic
Pre-Construction 10% crack 25% crack 10% crack 25% crack 10% crack 25% crack
Low 2 2
Dry
High
x ≤10%
Low 2 4 6
Wet
High 3 3 4 3 4
Low
Dry
High
Slurry 10%<x≤25%
Low
Wet
High
Low
Dry
High
x> 25%
Low
Wet
High

6.4 Annualized Pavement Treatment Costs

The material provided in this section is not part of the deliverable of this project. UCPRC believes that the
life-cycle cost analysis will provide definitive knowledge on whether, and at what stage of cracking,
pavement preservation treatments should be applied on a policy level. Given that, pavement service lives
and Equivalent Uniform Annual Cost (EUAC) figures can be used to select the most appropriate
treatment, depending on the program and budget planning constraints. The EUAC calculated here is the
initial construction cost discounted over the 50th percentile life calculated in the previous section.

Looking at the cost of various treatments for pavements in different conditions in terms of EUAC over
their service life period enables the most cost-effective treatment to be selected. If the expected life values
were known more completely with better data (in Table 20 through Table 29), EUAC costs (discounted

UCPRC-TM-2007-08 41
appropriately) could be used to plan annual budgets based on historical performance of preservation
treatments on the highway network. Table 30 through Table 37 present the agency’s construction cost
(pavement costs only) in terms of EUAC for the various pavement treatments (thousands of dollars per
lane-mile per year).

The EAUC is reasonable only if the expected life (in Table 20 to Table 29) is reasonable. Only agency
construction costs are used in the EUAC calculation, while the data (pavement only) are from Caltrans’
RealCost User Manual (7). Cost data in the manual are derived primarily from calculations performed for
the “State of the Pavement Report” by the Division of Maintenance (2006 version), which carry with them
the assumptions used in those calculations and which support cost multipliers from other sources. A more
detailed description of the cost data is included in the manual. Agency costs alone were considered for this
study because hourly traffic flows and construction schedules needed for calculating user costs were not
available for the projects included in the study, and also because the time necessary to attempt to estimate
those data for “typical” projects was outside the scope of this study.

Table 30: ACOL-DG HM-1 Annualized Cost per Lane-Mile (x $1,000)

Alligator A+B Crack Alligator A Crack Alligator B Crack
Year to Year to Year to Year to Year to Year to
Action Existing B Crack
Dry/Wet Traffic 10% 25% 10% 25% 10% 25%
Category Pre-Construction
crack crack crack crack crack crack
Low ($12.18) ($7.59) ($4.17) ($7.59) ($6.45)
Dry
High ($12.18) ($7.59) ($9.31)
x ≤10%
Low ($6.45) ($5.63) ($5.63) ($4.55) ($5.63) ($3.86)
Wet
High ($6.45) ($5.63) ($5.02) ($3.86) ($6.45) ($3.60)
Low ($12.18) ($7.59) ($3.60) ($3.04) ($7.59) ($6.45)
Dry
ACOL-DG- High ($12.18) ($7.59) ($5.63) ($9.31) ($7.59)
10%<x≤25%
HM1 Low ($9.31) ($5.63) ($5.63) ($5.63) ($5.63)
Wet
High ($9.31) ($7.59) ($6.45) ($7.59) ($5.63)
Low ($12.18) ($6.45) ($6.45) ($9.31) ($5.63)
Dry
High ($17.92) ($12.18) ($6.45) ($4.55) ($12.18) ($9.31)
x> 25%
Low ($12.18) ($9.31) ($9.31) ($4.55) ($5.02)
Wet
High ($12.18) ($6.45) ($6.45) ($9.31)
Construction cost is $33,800 per lane mile, according to the draft RealCost User Manual.

42 UCPRC-TM-2007-08
 Table 31: ACOL-DG CapM Annualized Cost per Lane-Mile (x $1,000)
Alligator A+B Crack Alligator A Crack Alligator B Crack
Year to Year to Year to Year to Year to Year to
Action Existing B Crack
Dry/Wet Traffic 10% 25% 10% 25% 10% 25%
Category Pre-Construction
crack crack crack crack crack crack
Low ($17.53) ($17.53) ($32.51) ($17.53)
Dry
High ($22.51) ($19.66) ($22.51)
x ≤10%
Low ($26.51) ($26.51)
Wet
High
Low ($22.51) ($19.66) ($26.51)
Dry
ACOL-DG- High
10%<x≤25%
CAPM Low ($32.51) ($19.66) ($19.66) ($26.51)
Wet
High
Low
Dry
High
x> 25%
Low
Wet
High

Construction cost is $118,000 per lane mile, according to the draft RealCost User Manual.
Table 32: ACOL-DG Rehabilitation Annualized Cost per Lane-Mile (x $1,000)
Alligator A+B Crack Alligator A Crack Alligator B Crack
Year to Year to Year to Year to Year to Year to
Action Existing B Crack
Dry/Wet Traffic 10% 25% 10% 25% 10% 25%
Category Pre-Construction
crack crack crack crack crack crack
Low ($34.13) ($34.13)
Dry
High ($49.82)
x ≤10%
Low ($82.37) ($57.04) ($67.16) ($57.04) ($49.82)
Wet
High ($57.04)
Low ($44.41) ($34.13)
Dry
ACOL-DG- High ($36.86) ($34.13) ($34.13) ($31.86)
10%<x≤25%
REH Low
Wet
High ($57.04) ($40.21) ($40.21)
Low ($40.21) ($34.13) ($34.13) ($36.86) ($31.86)
Dry
High ($44.41) ($31.86) ($31.86) ($31.86) ($44.41) ($29.94)
x> 25%
Low
Wet
High ($57.04) ($57.04)

Construction cost is $299,000 per lane mile, according to the draft RealCost User Manual.
Table 33: ACOL-OG HM-1 Annualized Cost per Lane-Mile (x $1,000)
Alligator A+B Crack Alligator A Crack Alligator B Crack
Year to Year to Year to Year to Year to Year to
Action Existing B Crack
Dry/Wet Traffic 10% 25% 10% 25% 10% 25%
Category Pre-Construction
crack crack crack crack crack crack
Low ($12.32)
Dry
High ($7.68) ($7.68)
x ≤10%
Low
Wet
High
Low ($18.13) ($12.32) ($12.32) ($12.32)
Dry
ACOLOG- High ($18.13) ($9.42) ($9.42) ($7.68) ($9.42)
10%<x≤25%
HM1 Low
Wet
High ($6.52) ($5.70)
Low ($18.13) ($12.32) ($7.68) ($18.13) ($12.32)
Dry
High ($12.32) ($7.68) ($9.42) ($7.68)
x> 25%
Low
Wet
High ($18.13) ($12.32) ($18.13) ($4.60)

Construction cost is $34,200 per lane mile, according to the draft RealCost User Manual.

UCPRC-TM-2007-08 43
 Table 34: ACOL-RAC HM-1 Annualized Cost per Lane-Mile (x $1,000)
Alligator A+B Crack Alligator A Crack Alligator B Crack
Action Existing B Crack Year to Year to Year to Year to Year to Year to
Dry/Wet Traffic
Category Pre-Construction 10% crack 25% crack 10% crack 25% crack 10% crack 25% crack
Low
Dry
High
x ≤10%
Low ($7.90) ($7.90)
Wet
High ($21.95) ($21.95) ($11.41)
Low
Dry
ACOLRAC- High
10%<x≤25%
HM1 Low ($6.15) ($6.15) ($5.10)
Wet
High ($43.06) ($21.95)
Low
Dry
High ($9.30)
x> 25%
Low
Wet
High

Construction cost is $41,400 per lane mile, according to the draft RealCost User Manual.
Table 35: Chip Seal Annualized Cost per Lane-Mile (x $1,000)
Alligator A+B Crack Alligator A Crack Alligator B Crack
Action Existing B Crack Year to Year to Year to Year to Year to Year to
Dry/Wet Traffic
Category Pre-Construction 10% crack 25% crack 10% crack 25% crack 10% crack 25% crack
Low ($4.88) ($2.38)
Dry
High ($3.38) ($2.95) ($3.38)
x ≤10%
Low ($3.98) ($1.77) ($1.77) ($2.38)
Wet
High ($6.38) ($3.98) ($3.98) ($3.98) ($3.38)
Low ($18.41) ($6.38) ($18.41) ($6.38)
Dry
High ($18.41) ($9.38) ($9.38) ($3.98)
ChipSeal-AC 10%<x≤25%
Low ($9.38) ($3.38) ($3.38) ($6.38) ($2.95)
Wet
High ($9.38) ($6.38) ($3.98) ($9.38) ($4.88)
Low ($2.95) ($1.89) ($1.89) ($1.89) ($1.68) ($1.68)
Dry
High ($18.41) ($9.38) ($9.38) ($9.38) ($6.38)
x> 25%
Low ($18.41) ($9.38) ($9.38) ($18.41) ($2.95)
Wet
High ($18.41) ($6.38) ($2.95) ($18.41) ($6.38)

Construction cost is $17,700 per lane mile, according to the draft RealCost User Manual.
Table 36: Digout Annualized Cost per Lane-Mile (x $1,000)
Alligator A+B Crack Alligator A Crack Alligator B Crack
Action Existing B Crack Year to Year to Year to Year to Year to Year to
Dry/Wet Traffic
Category Pre-Construction 10% crack 25% crack 10% crack 25% crack 10% crack 25% crack
Low
Dry
High
x ≤10%
Low ($35.88) ($35.88)
Wet
High
Low ($18.29) ($18.29)
Dry
High ($18.29) ($18.29)
Digout 10%<x≤25%
Low ($35.88) ($18.29) ($35.88) ($18.29)
Wet
High ($35.88) ($35.88) ($35.88) ($35.88) ($35.88) ($35.88)
Low ($35.88) ($35.88)
Dry
High
x> 25%
Low ($35.88) ($18.29) ($18.29)
Wet
High ($35.88) ($35.88) ($18.29) ($18.29) ($35.88) ($35.88)

Construction cost is $34,500 per lane mile, according to the draft RealCost User Manual.

44 UCPRC-TM-2007-08
 Table 37: Slurry Annualized Cost per Lane-Mile (x $1,000)
Alligator A+B Crack Alligator A Crack Alligator B Crack
Action Existing B Crack Year to Year to Year to Year to Year to Year to
Dry/Wet Traffic
Category Pre-Construction 10% crack 25% crack 10% crack 25% crack 10% crack 25% crack
Low ($8.27) ($8.27)
Dry
High
x ≤10%
Low ($8.27) ($4.30) ($2.98)
Wet
High ($5.62) ($5.62) ($4.30) ($5.62) ($4.30)
Low
Dry
High
Slurry 10%<x≤25%
Low
Wet
High
Low
Dry
High
x> 25%
Low
Wet
High

Construction cost is $15,600 per lane mile, according to the draft RealCost User Manual.

UCPRC-TM-2007-08 45
46 UCPRC-TM-2007-08
7 LIFE-CYCLE COST ANALYSIS

In this chapter, life-cycle cost analysis (LCCA) work is performed in the context of delivering the
following two objectives, as stated in the project work plan:

1. Calculate life-cycle costs (LCC):

Construction cost information will be collected from the State of the Pavement Report (SOPR)
and included in the draft life-cycle cost analysis manual developed by the UCPRC and the
Caltrans Division of Design. Information for certain pavement preservation treatments is not
included in the SOPR and will need to be provided by the Caltrans Division of Maintenance.

The typical life cycles and performance probabilities from this study will be used with the cost
information to calculate net present value (NPV) for each scenario with and without applying
pavement preservation treatments following the procedures and using information in the Caltrans
draft LCCA manual. The analysis will be performed considering agency cost only, unless
construction duration and scheduling information can be found for pavement preservation
treatments.

2. Analyze timing and relative performance:

If a sufficient range of pavement condition data prior to placement of the pavement preservation
treatment is found in the data available, an analysis will be made regarding the optimum timing of
treatments with respect to pavement condition in terms of life-cycle cost.

UCPRC-TM-2007-08 47
Figure 10 presents the order of work conducted in this section to accomplish the LCCA objectives.

Figure 10: Process flowchart for LCCA.

48 UCPRC-TM-2007-08
7.1 Assumptions for LCCA

The cost of various pavement treatment sequences over a predefined analysis period needs to be calculated
to achieve the above objectives. To limit the number of different treatment combinations, the following
assumptions are made for all treatment sequences:
1. All existing cracking before current treatment means the amount of Alligator B cracking.
2. LCCA scenarios are evaluated for 20- and 35-year analysis periods.
3. Discount rate is set at a fixed 4 percent.
4. Rehabilitation treatment consists of conventional dense-graded overlay only.
5. HM-1 treatments are performed with either all ACOL (DG, OG, RAC) or all chip seal.
6. All HM-1 treatments are performed at less than 10 percent Alligator B cracking.
7. Pavement lives are based on findings derived from Pavement Condition Survey data.
8. LCCA will be performed only for treatments for which the UCPRC is able to derive an expected
pavement life.

7.2 Factors Sensitivity Analysis

A new sensitivity analysis of factors affecting pavement life, here defined as the 50th percentile life to the
given extent of cracking, was performed to check for reasonableness and to determine if certain factors
could be consolidated with others or should be excluded. The factors analyzed consist of one load factor,
AADTT (Average Annual Daily Truck Traffic), and four climatic factors: average annual precipitation
(mm/yr); average annual minimum temperature (Cº), which is the minimum temperature on the coldest
day of the year average across many years; average annual minimum temperature (Cº) in January; and
average annual freeze-thaw cycles.

Sensitivity analyses are applied to HM-1 dense-graded overlay treatments, as this style of treatment has
the greatest available sample data. Each factor is divided into four groups according to each of the
quartiles from the collected sample database. The results show that none of these factors alone has a
dominant effect on the performance of the pavement, hence all climatic factors are dropped from the
LCCA work and only the load factor, AADTT, remains. Table 38 to Table 42 provide reasons why these
factors are not considered in developing the pavement performance estimate used for LCCA. The red-
bracketed values in each of the tables indicate examples of the conclusion drawn from each type of
sensitivity study summarized in the text below each table for Table 38 to Table 42.

UCPRC-TM-2007-08 49
 Table 38: AADTT Sensitivity Test Table

In Table 38, the total cracking (Alligator A+B cracking) does not show a consistently decreasing
pavement life as the truck traffic increases. For example, see values beside the brackets in Table 38. Some
life values increase with increased traffic, and some decrease. The limited and unbalanced database of
treatments and performance histories contributes to these trends.

Table 39: Average Annual Precipitation Sensitivity Test Table

50 UCPRC-TM-2007-08
In Table 39, cracking does not show a consistently decreasing trend in pavement life as the amount of
precipitation increases, which is inconsistent with the expectation that precipitation should have an
adverse effect on pavement life. As noted in Section 5 of this technical memo, precipitation is a critical
factor affecting pavement life based on observations from the WSDOT database, and pavement life should
decrease as precipitation levels increase. The WSDOT database has many times more observations, and
only one treatment: the equivalent of ACOL-DG.

Table 40: Average Annual Minimum Temperature Sensitivity Test Table

Again, as seen in Table 40, cracking does not show a consistently increasing trend in pavement life as the
average annual minimum temperature increases. It is expected that pavement life in terms of cracking
distress should increase as the minimum temperature rises. Again, based on the WSDOT experience,
higher minimum temperatures should reduce the rate of cracking progression (6).

UCPRC-TM-2007-08 51
 Table 41: Average Annual Minimum Temperature in January Sensitivity Test Table
Alligator A+B Crack Alligator A Crack Alligator B Crack
Avg.
Year to Year to Year to Year to Year to Year to
PP Existing B Crack Pre- Annual
10% 25% 10% 25% 10% 25%
Strategy Construction MinTemp
crack crack crack crack crack crack
in Jan.
≤-5.9 4 6 10 5
≤-0.1 5 6 7 6
A+B<10% x ≤10%
≤4 7 9 10 9 12
>4
≤-5.9 3 6 7 5 7
≤-0.1 6 8 8 9 8 10
A+B≥10% x ≤10%
≤4 4 4 6 9 5 9
ACOL-DG- >4 5 5 8 6 11
HM1 ≤-5.9 5 6 8 15 6 7
≤-0.1 6 6 7 6 7
10%<x≤25%
≤4 4 5 6 5
>4 6 5 6
≤-5.9 4 6 6 5 7
≤-0.1 5 6 6 9
x> 25%
≤4 3 3 4 3 4
>4 3 4 7 4 4

As with the average minimum annual temperature, the trend of pavement life seen in Table 41 for cracking
does not consistently increase as average minimum January temperature rises. This observation also
contradicts what was learned from the study on the WSDOT database, where higher minimum temperature
reduced the rate of crack progression (6).

52 UCPRC-TM-2007-08
 Table 42: Average Annual Freeze-thaw Cycles Sensitivity Test Table
Alligator A+B Crack Alligator A Crack Alligator B Crack
Avg
Annual Year to Year to Year to Year to Year to Year to
PP Existing B Crack Pre-
Free- 10% 25% 10% 25% 10% 25%
Strategy Construction
Thaw crack crack crack crack crack crack
cycle
≤6
≤45 2 4 5 7 7
A+B<10% x ≤10%
≤137 6 6 9 6 12
>137 5 9 10 9 12
≤6 3 4 4 4
≤45 5 7 6 7 8
A+B≥10% x ≤10%
≤137 4 6 7 6 7
ACOL-DG- >137 6 8 8 9 7
HM1 ≤6 2 5 4 3 6
≤45 4 5 8 4 7
10%<x≤25%
≤137 4 7 9 15 5 7
>137 5 6 7 6 7
≤6 2 4 4 2
≤45 4 9 5
x> 25%
≤137 4 6 6 5
>137 4 6 6 5 7

Pavement life should increase with fewer freeze-thaw cycles. However, analysis does not show a
consistent increase in pavement life versus decrease in number of freeze-thaw cycles, as seen in Table 42.
Also, based on the cracking progression model study using WSDOT data, an increase in the number of
freeze-thaw cycles increases the rate of cracking progression (6).

Before proceeding with the life-cycle cost analysis, a final evaluation was performed to check for bias
with regard to urban versus rural sources of project data. Recall from Table 10 that over half of the HM-1
projects were located in Districts 2 and 8; nearly three fourths of the Rehabilitation projects were located
in Districts 2 and 7. In this evaluation, Districts 7 and 8 represented urban projects, and District 2 projects
were considered to be rural projects.

Table 43 presents expected pavement life for HM-1 projects using the ACOL-DG treatment based on the
entire dataset (under the “Alligator B Crack” column heading) and based on projects from Districts 2 and
8 (noted in column headings). Differences are apparent but do not clearly show bias. Table 44 presents
expected pavement life in a similar format that compares results both for the entire dataset and for projects
in Districts 2 and 7. Bias is not indicated.

UCPRC-TM-2007-08 53
 Table 43: ACOL-DG HM-1—Rural vs. Urban District Sensitivity Test Table
Alligator B Alligator B
Alligator B
Crack Rural Crack Urban
Crack
(District 2 Only) (District 8 Only)
Existing B Year to Year to Year to Year to Year to Year to
PP Strategy Crack Pre- Dry/Wet Traffic 10% 25% 10% 25% 10% 25%
Construction Crack Crack Crack Crack Crack Crack
Low 5 6 3 6
Dry
High 4 2 6
x ≤10%
Low 7 11 7
Wet
High 6 12 6 10 4
Low 5 6 4 8 2 3
Dry
ACOL-DG- High 4 5 4 6 5 5
10%<x≤25%
HM1 Low 7 7 7 7 4 7
Wet
High 5 7 6 7 4 4
Low 4 7 3 7
Dry
High 3 4 5 7 3 3
x> 25%
Low 8 8 4 3
Wet
High 4 5 4 3

Table 44: ACOL-DG Rehabilitation—Rural vs. Urban District Sensitivity Test Table

Alligator B Alligator B
Alligator B
Crack Rural Crack Urban
Crack
(District 2 Only) (District 7 Only)
Existing B Year to Year to Year to Year to Year to Year to
PP Strategy Crack Pre- Dry/Wet Traffic 10% 25% 10% 25% 10% 25%
Construction Crack Crack Crack Crack Crack Crack
Low 11 11
Dry
High 5 5 5 5
x ≤10%
Low 6 7 5 7
Wet
High 5
Low 11 11
Dry
ACOL-DG- High 12
10%<x≤25%
REH Low
Wet
High 9
Low 10 12 10 12
Dry
High 8 13 8 8
x> 25%
Low
Wet
High 6 2 6

54 UCPRC-TM-2007-08
 7.3 Attempt to Predict Pavement Performance by Linear Extrapolation

This section describes an effort to use linear regression to estimate pavement service life for datasets
where the median did not reach cracking trigger levels of 10 percent and 25 percent. Figure 11 shows an
example with linear regression applied to dense-graded asphalt concrete overlay rehabilitation. In this
example, the linear regression is forced to intercept the origin, since a failure is not expected to happen
immediately after the completion of any treatment work. The actual cumulative distribution function
(CDF) from the data is plotted up to 11 years, ending at just below 50 percent, indicating that less than
half the dataset reached the trigger of 10 percent cracking. The 50th percentile of failure can be estimated
(approximately 12 years) from the trend-line.

ACOL - Rehabilitation, Time to 10% Alligator B Crack

1
Cumulative Probability of Failure at

y = 0.0405x
0.8
10% Crack

0.6

0.4

Alligator B Crack CDF

0.2 Linear (Alligator B Crack CDF)

0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20

Year after treatment

Figure 11: Example of pavement life prediction using linear regression.

Using regression analysis, median values were first estimated, then compared to actual data. Table 45
shows a comparison of actual pavement services lives, where available, to the service lives that were
estimated with linear regression and the absolute value of the percentage differences.

UCPRC-TM-2007-08 55
 Table 45: Median vs. Estimated ACOL-DG Rehabilitation Lives
Alligator B Crack
Year to Year to Year to Year to 25%
PP Existing B Crack Pre- % %
10% 10% crack 25% crack
Strategy Construction Difference Difference
crack (Trendline) crack (Trendline)
x≤10% N/A 13 N/A 11 16 45%
ACOL-DG-
10%<x≤25% 12 14 17% N/A 25 N/A
REH
x>25% 10 9 10% 12 13 8%

As shown in Table 45, the difference between the observed 50th percentile versus the linearly estimated
50th percentile value range from 8 percent to 45 percent. Given this wide range of error, pavement service
life estimated from linear regression was not used for LCCA.

7.4 Service Life Used for LCCA

All the climatic factors are eliminated in deriving pavement service lives for LCCA. Such lives are shown
in Table 46 through Table 52. Traffic factors are not included in Table 46 through Table 48 because of
insufficient sample sizes for each traffic category. Where climate and traffic factors are not shown in the
table, all categories of traffic and climate have been pooled.

Table 46: Dense-Graded Asphalt Concrete Overlay

(Rehabilitation) Performance

Table 47: Dense-Graded Asphalt Concrete Overlay (CapM) Performance

56 UCPRC-TM-2007-08
 Table 48: Rubber Asphalt Concrete Overlay (CapM) Performance

Table 49: Dense-Graded Asphalt Concrete Overlay (HM-1) Performance

Table 50: Rubber Asphalt Concrete Overlay (HM-1) Performance

UCPRC-TM-2007-08 57
 Table 51: Open-Graded Asphalt Concrete Overlay (HM-1) Performance

Table 52: Chip Seal (HM-1) Performance

7.5 LCCA Scenarios

Seven major sets of scenarios were created for comparison of life-cycle cost. Most of the major scenario
sets consist of several combinations of scenarios as listed in Table 53 to Table 59. These subscenarios are
listed according to the major scenario set to which they belong, as numbered in the first column of these
tables. The numbers in the first two columns of the tables simply show the number of subscenarios
included in each major scenario; for example, in Table 53 subscenarios 1.1 to 1.8 represent one subset.
These LCCA scenarios are established with the goals of discovering the following:
1. Is it more beneficial to apply pavement preservation (HM-1) than to perform no maintenance at
all?
2. Should pavement preservation be applied at an earlier or a later stage of cracking?
3. How do the life-cycle costs of various pavement treatment combinations compare?

All seven sets of scenarios are presented here for completeness, because they were considered for analysis.
The life of each treatment in the LCCA scenarios is based on those indicated in Section 7.4 of this memo

58 UCPRC-TM-2007-08
in Table 46 to Table 52. LCCA for all scenarios were run based on the lives shown in Section 7.4.
However, only Scenario 2 could produce outcomes that provided meaningful interpretation because all
other scenarios’ results were unusable either because of a lack of data (e.g., CapM) or because they were
not viable for practical application (e.g., performing rehabilitation at 10 percent Alligator B cracking).

An example of how a LCCA scenario is sequenced is demonstrated here, using a dense-grade AC overlay
case in major scenario Set No. 2:
1. Rehabilitation is performed at >25% Alligator B cracking.
2. Each stage of rehabilitation work is followed by two HM-1 dense-graded overlay treatments.
3. Each HM-1 dense-graded overlay treatment is performed at (≤10% Alligator B) and (≤10%
alligator A+B) cracking.
4. Each HM-1 action is performed under both a high and low AADTT load.
Under such a scenario, lives of the Rehabilitation and HM-1 treatments are collected from Table 46 and
Table 49. The life of Rehabilitation action is 10 years when it is performed at >25% existing Alligator B
cracking. The life of the dense-grade overlay HM-1 action is 6 years when it is performed at (≤10%
Alligator B) and (≤10% alligator A+B) cracking. Based on a 35-year analysis period, the sequence of this
particular scenario is then:

Rehab@Year0 Æ DGAC HM-1@Year10Æ DGAC HM-1@Year16 Æ Rehab@Year22 Æ DGAC HM-

1@Year32 Æ End@Year35 (with 3 years of life left in the last HM-1 action as salvage value).

UCPRC-TM-2007-08 59
 Table 53: LCCA Scenario List for Major Scenario Set No. 1
Scenario Rehab performed at <10% Alligator B cracking
1 Rehab@10% -- Rehab@10% -- Rehab@10% --
1st Set Each rehab followed by two HM-1 treatments, HM-1 performed at (≤ 10%B) & (A+B ≤ 10%)
Rehab@10% -- HM1@10% -- HM1@10% -- Rehab@10% -- HM1@10% -- HM1@10% --
1.1 to 1.8 x4 for four kinds of HM1 treatment (ACOL-DG, ACOL-RAC, ACOL-OG, AC Chipseal)
x2 for two levels of traffic level (High, Low)
2nd Set Each rehab followed by two HM-1 treatments, HM-1 performed at (≤ 10%B) & (A+B > 10%)
Rehab@10% -- HM1@10% -- HM1@10% -- Rehab@10% -- HM1@10% -- HM1@10% --
1.9 to 1.16 x4 for four kinds of HM1 treatment (ACOL-DG, ACOL-RAC, ACOL-OG, AC Chipseal)
x2 for two levels of traffic level (High, Low)
3rd Set Rehab followed by only HM-1 treatments, HM-1 performed at (≤ 10%B) & (A+B ≤ 10%)
Rehab@10% -- HM1@10% -- HM1@10% -- HM1@10% -- HM1@10% --
1.17 to 1.24 x4 for four kinds of HM1 treatment (ACOL-DG, ACOL-RAC, ACOL-OG, AC Chipseal)
x2 for two levels of traffic level (High, Low)
4th Set Rehab followed by only HM-1 treatments, HM-1 performed at (≤ 10%B) & (A+B > 10%)
Rehab@10% -- HM1@10% -- HM1@10% -- HM1@10% -- HM1@10% --
1.25 to 1.32 x4 for four kinds of HM1 treatment (ACOL-DG, ACOL-RAC, ACOL-OG, AC Chipseal)
x2 for two levels of traffic level (High, Low)

Set No. 1 scenarios focused on life-cycle cost (LCC) if rehabilitation were performed at earlier stages of
Alligator B cracking (less than 10 percent), which is much lower than the current Caltrans trigger of
greater than 25 percent Alligator B cracking. This set of scenarios was not studied due to lack of data and
lack of viability for practical application.

Table 54: LCCA Scenario List for Major Scenario Set No. 2
Scenario Rehab performed at >25% Alligator B cracking
2 Rehab@25% -- Rehab@25% -- Rehab@25% --
1st Set Each rehab followed by two HM-1 treatments, HM-1 performed at (≤ 10%B) & (A+B ≤ 10%)
Rehab@25% -- HM1@10% -- HM1@10% -- Rehab@25% -- HM1@10% -- HM1@10% --
2.1 to 2.8 x4 for four kinds of HM1 treatment (ACOL-DG, ACOL-RAC, ACOL-OG, AC Chipseal)
x2 for two levels of traffic level (High, Low)
2nd Set Each rehab followed by two HM-1 treatments, HM-1 performed at (≤ 10%B) & (A+B > 10%)
Rehab@25% -- HM1@10% -- HM1@10% -- Rehab@25% -- HM1@10% -- HM1@10% --
2.9 to 2.16 x4 for four kinds of HM1 treatment (ACOL-DG, ACOL-RAC, ACOL-OG, AC Chipseal)
x2 for two levels of traffic level (High, Low)
3rd Set Rehab followed by only HM-1 treatments, HM-1 performed at (≤ 10%B) & (A+B ≤ 10%)
Rehab@25% -- HM1@10% -- HM1@10% -- HM1@10% -- HM1@10% --
2.17 to 2.24 x4 for four kinds of HM1 treatment (ACOL-DG, ACOL-RAC, ACOL-OG, AC Chipseal)
x2 for two levels of traffic level (High, Low)
4th Set Rehab followed by only HM-1 treatments, HM-1 performed at (≤ 10%B) & (A+B > 10%)
Rehab@25% -- HM1@10% -- HM1@10% -- HM1@10% -- HM1@10% --
2.25 to 2.32 x4 for four kinds of HM1 treatment (ACOL-DG, ACOL-RAC, ACOL-OG, AC Chipseal)
x2 for two levels of traffic level (High, Low)
5th Set Each rehab followed by two HM-1 treatments, HM-1 performed at (> 25%B)
Rehab@25% -- HM1@25% -- HM1@25% -- Rehab@25% -- HM1@25% -- HM1@25% --
2.33 to 2.40 x4 for four kinds of HM1 treatment (ACOL-DG, ACOL-RAC, ACOL-OG, AC Chipseal)
x2 for two levels of traffic level (High, Low)
6th Set Rehab followed by only HM-1 treatments, HM-1 performed at (> 25%B)
Rehab@25% -- HM1@25% -- HM1@25% -- HM1@25% -- HM1@25% --
2.41 to 2.48 x4 for four kinds of HM1 treatment (ACOL-DG, ACOL-RAC, ACOL-OG, AC Chipseal)
x2 for two levels of traffic level (High, Low)

Scenarios in set No. 2 focus on the LCC of pavements when rehabilitation activities are performed at
greater than 25 percent Alligator B cracking. For example, subscenarios 2.1 to 2.8 would be:

60 UCPRC-TM-2007-08
 2.1 Under low traffic, each rehabilitation is performed at 25% cracking and isfollowed by
two ACOL-DG actions at 10% cracking at HM-1 level.
2.2 Under low traffic, each rehabilitation is performed at 25% cracking and is followed by
two ACOL-RAC action at 10% cracking at HM-1 level.
2.3 Under low traffic, each rehabilitation is performed at 25% cracking and is followed by
two ACOL-OG actions at 10% cracking at HM-1 level.
2.4 Under low traffic, each rehabilitation is performed at 25% cracking and is followed by
two chip seal actions at 10% cracking.
2.5 Under high traffic, each rehabilitation is performed at 25% cracking and is followed
by two ACOL-DG actions at 10% cracking at HM-1 level.
2.6 Under high traffic, each rehabilitation is performed at 25% cracking and is followed
by two ACOL-DRAC actions at 10% cracking at HM-1 level.
2.7 Under high traffic, each rehabilitation is performed at 25% cracking and is followed
by two ACOL-OG actions at 10% cracking at HM-1 level.
2.8 Under high traffic, each rehabilitation is performed at 25% cracking and is followed
by two chip seal actions at 10% cracking.

This set of scenarios also reflects the current Caltrans practice by triggering a rehabilitation action only
when Alligator B cracking exceeds 25 percent. Subscenarios (5th and 6th sets) included under major
scenario No. 2 demonstrate the LCC of pavement under the current Caltrans practice of performing most
of the HM-1 treatments when Alligator B cracking has exceeded 25 percent versus the expected LCC if
HM-1 treatments are performed at lower levels of cracking.

Table 55: LCCA Scenario List for Major Scenario Set No. 3
Scenario CapM performed at <10% Alligator B cracking
3 CapM@10% -- CapM@10% -- CapM@10% -- CapM@10% --
1st Set Each CapM followed by two HM-1 treatments, HM-1 performed at (≤ 10%B) & (A+B ≤ 10%)
CapM@10% -- HM1@10% -- HM1@10% -- CapM@10% -- HM1@10% -- HM1@10% --
3.1 to 3.8 x4 for four kinds of HM1 treatment (ACOL-DG, ACOL-RAC, ACOL-OG, AC Chipseal)
x2 for two levels of traffic level (High, Low)
2nd Set Each CapM followed by two HM-1 treatments, HM-1 performed at (≤ 10%B) & (A+B > 10%)
CapM@10% -- HM1@10% -- HM1@10% -- CapM@10% -- HM1@10% -- HM1@10% --
3.9 to 3.16 x4 for four kinds of HM1 treatment (ACOL-DG, ACOL-RAC, ACOL-OG, AC Chipseal)
x2 for two levels of traffic level (High, Low)
3rd Set CapM followed by only HM-1 treatments, HM-1 performed at (≤ 10%B) & (A+B ≤ 10%)
CapM@10% -- HM1@10% -- HM1@10% -- HM1@10% -- HM1@10% --
3.17 to 3.24 x4 for four kinds of HM1 treatment (ACOL-DG, ACOL-RAC, ACOL-OG, AC Chipseal)
x2 for two levels of traffic level (High, Low)
4th Set CapM followed by only HM-1 treatments, HM-1 performed at (≤ 10%B) & (A+B > 10%)
CapM@10% -- HM1@10% -- HM1@10% -- HM1@10% -- HM1@10% --
3.25 to 3.32 x4 for four kinds of HM1 treatment (ACOL-DG, ACOL-RAC, ACOL-OG, AC Chipseal)
x2 for two levels of traffic level (High, Low)

UCPRC-TM-2007-08 61
Scenario set No. 3 presents the expected LCC of pavement maintenance sequences when activities consist
of only CapM treatments and their respective HM-1 treatments’ effects. Inadequate data for CapM
projects precluded LCC evaluation.

Table 56: LCCA Scenario List for Major Scenario Set No. 4
Scenario All CapM performed at >25% Alligator B cracking
4 CapM@25% -- CapM@25% -- CapM@25% -- CapM@25% --

Scenario set No. 4 reflects the expected LCC of pavement when only CapM maintenance treatments are
performed and they are applied when Alligator B cracking exceeds 25 percent. This scenario would
demonstrate the LCC difference when CapM treatments are performed in the later stages of cracking
versus scenario set No. 3, when pavements are treated at an earlier stage of cracking. Lack of CapM data
prevented analysis.

Table 57: LCCA Scenario List for Major Scenario Set No. 5
Scenario Rehab and CapM alternating at <10% Alligator B cracking
5 Rehab@10% -- CapM@10% -- Rehap@10% -- CapM@10% -- Rehab@10% --
1st Set Each rehab/CapM followed by two HM-1 treatments, HM-1 performed at (≤ 10%B) & (A+B ≤ 10%)
Rehab@10% -- HM1@10% -- HM1@10% -- CapM@10% -- HM1@10% -- HM1@10% -- Rehab@10% --
5.1 to 5.8 x4 for four kinds of HM1 treatment (ACOL-DG, ACOL-RAC, ACOL-OG, AC Chipseal)
x2 for two levels of traffic level (High, Low)
2nd Set Each rehab/CapM followed by two HM-1 treatments, HM-1 performed at (≤ 10%B) & (A+B > 10%)
Rehab@10% -- HM1@10% -- HM1@10% -- CapM@10% -- HM1@10% -- HM1@10% -- Rehab@10% --
5.9 to 5.16 x4 for four kinds of HM1 treatment (ACOL-DG, ACOL-RAC, ACOL-OG, AC Chipseal)
x2 for two levels of traffic level (High, Low)

In scenario set No. 5, instead of having the same major maintenance treatments throughout the analysis
period, rehabilitation and CapM actions are assumed to alternate within each analysis sequence. The two
sets of subscenarios (1st Set vs. 2nd Set) demonstrate the cost-effectiveness of performing HM-1
treatments earlier versus later in the cracking stage when major maintenance treatments are not unique.
This set was not included due to lack of CapM data and the impracticality of rehabilitation at only 10
percent cracking.

Table 58: LCCA Scenario List for Major Scenario Set No. 6
Scenario Rehab followed by one CapM, alternating at >25% Alligator B cracking
6 Rehab@25% -- CapM@25% -- Rehap@25% -- CapM@25% -- Rehab@25% --

Scenario set No. 6 focuses on alternating use of major maintenance activities in the later stages of cracking
(greater than 25 percent Alligator B cracking) versus scenarios in set No. 5, in which the pavement is
treated at less than 10 percent of Alligator B cracking. Lack of CapM data prevented analysis.

62 UCPRC-TM-2007-08
 Table 59: LCCA Scenario List for Major Scenario Set No. 7
Scenario Rehab and CapM alternating at >25% Alligator B cracking
7 Rehab@25% -- CapM@25% -- CapM@25% -- Rehap@25% -- CapM@25% -- CapM@25% -- Rehab@25% --
1st Set Each rehab/CapM followed by two HM-1 treatments, HM-1 performed at (≤ 10%B) & (A+B ≤ 10%)
Rehab@25% -- HM1@10% -- HM1@10% -- CapM@25% -- HM1@10% -- HM1@10% -- CapM@25% --
7.1 to 7.8 x4 for four kinds of HM1 treatment (ACOL-DG, ACOL-RAC, ACOL-OG, AC Chipseal)
x2 for two levels of traffic level (High, Low)
2nd Set Each rehab/CapM followed by two HM-1 treatments, HM-1 performed at (≤ 10%B) & (A+B > 10%)
Rehab@25% -- HM1@10% -- HM1@10% -- CapM@25% -- HM1@10% -- HM1@10% -- CapM@25% --
7.9 to 7.16 x4 for four kinds of HM1 treatment (ACOL-DG, ACOL-RAC, ACOL-OG, AC Chipseal)
x2 for two levels of traffic level (High, Low)

The scenarios in set No. 7 further expand the various pavement treatment combinations. Each
rehabilitation is followed by two CapM treatments. The two subsets (1st Set vs. 2nd Set) are used to
identify the cost-effectiveness of performing HM-1 treatments to pavements earlier in the cracking stages.
Lack of CapM data prevented analysis.

7.6 LCCA Results and Findings

Results from a life-cycle cost analysis of those scenarios evaluated show the following:
1. It is more cost-effective to apply preservation treatments than to perform no maintenance between
rehabilitations.
2. It is more cost-effective to apply preservation treatments at earlier stages of cracking.
3. Comparing different combinations of treatments was severely hampered by limited data.

Details of the LCCA results are discussed in the following three sections.

7.6.1 LCCA Objective 1: Is It More Beneficial to Apply Pavement Preservation (HM-1) Versus
No Maintenance?
Based on Table 60, for rehabilitation done at greater than 25 percent Alligator B cracking, the following
was observed when compared with the subscenario of only rehabilitation within no maintenance
treatments (first row in the table):
• When preventive actions are performed at less than 10 percent Alligator B cracking, a
minimum saving of 27 percent is achieved over a 20-year analysis period.
• When preventive actions are performed at less than 10 percent Alligator B cracking, a
minimum saving of 26 percent is achieved over a 35-year analysis period.
• Scenarios with preventive actions done at the very late stages of cracking (greater than
25 percent Alligator B cracking) reflect Caltrans practice during the period studied. All
scenarios show that at least a 13 percent saving in LCC is achieved by applying preventive
maintenance (dense-graded asphalt overlays) even when cracking is very high (greater than 25
percent).

UCPRC-TM-2007-08 63
 Table 60: LCC of Preventive vs. No Preventive Actions (Rehabilitation at >25% Alligator B Cracking)—Dense-Graded Overlay
HM-1 Treatment = DGAC overlay
Scenario Set No.2 - Rehabilitation Treated at Existing B Crack Pre- LCC over 20 Years in $1000/mi/ln LCC over 35 Years in $1000/mi/ln
>25% Alligator B Crack Construciotn (Cost reduction vs. no HM-1's) (Cost reduction vs. no HM-1's)
Low AADTT High AADTT Low AADTT High AADTT
Rehab performed at >25% Alligator B cracking No HM-1's 456.6 456.6
Each rehab followed by two HM-1 treatments, HM-1
333.3 (27%) 426.1 (29%)
performed at (≤ 10%B) & (A+B ≤ 10%)
A+B<10% x≤10%
Rehab followed by only HM-1 treatments, HM-1
333.3 (27%) 369.6 (39%)
performed at (≤ 10%B) & (A+B ≤ 10%)
Each rehab followed by two HM-1 treatments, HM-1
332 (27%) 333.8 (27%) 425.4 (29%) 447.7 (26%)
performed at (≤ 10%B) & (A+B > 10%)
A+B≥10% x≤10%
Rehab followed by only HM-1 treatments, HM-1
332 (27%) 333.8 (27%) 360.2 (40%) 369.6 (39%)
performed at (≤ 10%B) & (A+B > 10%)
Each rehab followed by two HM-1 treatments, HM-1
396.5 (13%) 484.3 (20%)
performed at (> 25%B)
Rehab followed by only HM-1 treatments, HM-1 X>25%
performed at (> 25%B) 343.7 (25%) 387.7 (36%)

Table 61: LCC of Preventive vs. No Preventive Actions (Rehabilitation at >25% Alligator B Cracking)—Chip Seal
HM-1 Treatment = AC Chipseal
Scenario Set No.2 - Rehabilitation Treated at Existing B Crack Pre- LCC over 20 Years in $1000/mi/ln LCC over 35 Years in $1000/mi/ln
>25% Alligator B Crack Construciotn (Cost reduction vs. no HM-1's) (Cost reduction vs. no HM-1's)
Low AADTT High AADTT Low AADTT High AADTT
Rehab performed at >25% Alligator B cracking No HM-1's 456.6 456.6
Each rehab followed by two HM-1 treatments, HM-1
309.9 (32%)
performed at (≤ 10%B) & (A+B ≤ 10%)
A+B<10% x≤10%
Rehab followed by only HM-1 treatments, HM-1
309.9 (32%) 319.2 (30%) 318.2 (47%) 336 (44%)
performed at (≤ 10%B) & (A+B ≤ 10%)
Each rehab followed by two HM-1 treatments, HM-1
performed at (≤ 10%B) & (A+B > 10%)
A+B≥10% x≤10%
Rehab followed by only HM-1 treatments, HM-1
337 (26%) 337 (26%) 368.1 (39%) 368.1 (39%)
performed at (≤ 10%B) & (A+B > 10%)
Each rehab followed by two HM-1 treatments, HM-1
307.8 (33%) 398.4 (13%) 315.4 (31%) 477.8 (21%)
performed at (> 25%B)
Rehab followed by only HM-1 treatments, HM-1 X>25%
performed at (> 25%B) 307.8 (33%) 328.7 (28%) 315.4 (31%) 359.8 (40%)

64 UCPRC-TM-2007-08
Results from LCC calculations shown in Table 61 for chip seals with conventional binders (Chip Seal-AC)
led to the following observations when compared with the subscenario of only rehabilitation with no
maintenance treatments (first row in the table):
• When preventive actions are performed at less than 10 percent Alligator B cracking, a
minimum saving of 26 percent is achieved over a 20-year analysis period.
• When preventive actions are performed at less than 10 percent Alligator B cracking, a
minimum saving of 21 percent is achieved over a 35-year analysis period.
• Asphalt chip seals applied even at >25% Alligator B cracking result in at least a 13 percent
saving in LCC versus rehabilitation without subsequent preservation treatment.
In general, regardless of when the preventive maintenance action is performed, it is more cost-
effective to apply preventive maintenance actions to pavements than not to apply treatments. This is true
for both dense-graded asphalt overlays and conventional chip seals. The cost savings from the 35-year
analysis are higher than those from the 20-year analysis. The range of values and the trends in these results
indicate the order of magnitude and the trend of likely life-cycle cost savings. This appears reasonable,
based on the limited data available.

7.6.2 LCCA Objective 2: Should Pavement Preservation Be Applied at an Earlier or a Later

Stage of Cracking?

Costs shown in Table 62, where a preservation treatment (HM-1 dense-graded asphalt concrete overlay) is
applied at either A+B and B cracking less than or equal to 10 percent, A+B greater than or equal to 10
percent and B cracking less than or equal to 10 percent, and B cracking greater than 25 percent, after
rehabilitation was done at Alligator B cracking greater than 25 percent, resulted in the following
observations:
• Trends from the limited data available suggest that waiting until a later stage of cracking
results in life-cycle costs of up to 16 percent higher (for 20-year analysis) than if treatments
are placed at an earlier stage of cracking; some cases did not show any higher cost.
• In the 35-year analysis, delaying treatment until later-stage cracking increases life-cycle costs
up to 12 percent.

UCPRC-TM-2007-08 65
 Table 62: LCC of Preventive Treatment at Early vs. Later Stage of Cracking
(Rehabilitation at >25% Alligator B Cracking)—Dense-Graded Overlay
HM-1 Treatment = DGAC overlay
LCC over 20 Years in $1000/mi/ln LCC over 35 Years in $1000/mi/ln
Scenario Set No.2 - Rehabilitation Treated at Existing B Crack Pre-
(Cost increase from main't at very low (Cost increase from main't at very low
>25% Alligator B Crack Construciotn
cracking) cracking)
Low AADTT High AADTT Low AADTT High AADTT
Each rehab followed by two HM-1 treatments, HM-1
333.3 426.1
performed at (≤ 10%B) & (A+B ≤ 10%)
A+B<10% x≤10%
Rehab followed by only HM-1 treatments, HM-1
333.3 369.6
performed at (≤ 10%B) & (A+B ≤ 10%)
Each rehab followed by two HM-1 treatments, HM-1
332 333.8 (0%) 425.4 447.7 (5%)
performed at (≤ 10%B) & (A+B > 10%)
A+B≥10% x≤10%
Rehab followed by only HM-1 treatments, HM-1
332 333.8 (0%) 360.2 369.6 (0%)
performed at (≤ 10%B) & (A+B > 10%)
Each rehab followed by two HM-1 treatments, HM-1
396.5 (16%) 484.3 (12%)
performed at (> 25%B)
Rehab followed by only HM-1 treatments, HM-1 X>25%
performed at (> 25%B) 343.7 (3%) 387.7 (5%)

Table 63: LCC of Preventive Treatment at Early vs. Later Stage of Cracking
(Rehabilitation at >25% Alligator B Cracking)—Chip Seal-AC
HM-1 Treatment = AC Chipseal
Scenario Set No.2 - Rehabilitation Treated at Existing B Crack Pre- LCC over 20 Years in $1000/mi/ln LCC over 35 Years in $1000/mi/ln
>25% Alligator B Crack Construciotn (Cost increase from main't at very low (Cost increase from main't at very low
cracking) cracking)
Low AADTT High AADTT Low AADTT High AADTT
Each rehab followed by two HM-1 treatments, HM-1
309.9
performed at (≤ 10%B) & (A+B ≤ 10%)
A+B<10% x≤10%
Rehab followed by only HM-1 treatments, HM-1
309.9 319.2 318.2 336
performed at (≤ 10%B) & (A+B ≤ 10%)
Each rehab followed by two HM-1 treatments, HM-1
performed at (≤ 10%B) & (A+B > 10%)
A+B≥10% x≤10%
Rehab followed by only HM-1 treatments, HM-1
337 (8%) 337 (5%) 368.1 (14%) 368.1 (9%)
performed at (≤ 10%B) & (A+B > 10%)
Each rehab followed by two HM-1 treatments, HM-1
307.8 (-1%) 398.4 315.4 477.8
performed at (> 25%B)
Rehab followed by only HM-1 treatments, HM-1 X>25%
performed at (> 25%B) 307.8 (-1%) 328.7 (3%) 315.4 (-1%) 359.8 (7%)

66 UCPRC-TM-2007-08
Based on calculated costs in Table 63 for chip seals with conventional binders, the following was
observed:
• Life-cycle costs can increase up to 14 percent by waiting to apply chip seals beyond 10
percent of the wheelpath with A + B cracking.
• Unexpected results, showing nominally unchanged life-cycle costs when chip seals are
applied when Alligator B cracking exceeds 25 percent (versus applying the seals at a lower
stage of cracking), are attributed to the limited data available for analysis.

Values and trends in the LCCA for dense-graded AC are expected to better represent the effects of
applying treatments earlier, when cracking levels are lower. Results from the LCCA for chip seals indicate
the limitations of the data available.

7.6.3 LCCA Objective 3: Comparison of LCC of Different Combinations of Pavement Treatments

The costs in Table 64 enable comparison of two alternative application sequences of preservation
treatments (dense-graded asphalt concrete overlay), where (1) rehabilitation is followed by two HM-1
treatments and then another rehabilitation or (2) rehabilitation is followed successively only by HM-1
treatments. Results led to the following observations:
• In the 20-year analysis, life-cycle costs are the same, except for pavements on which the HM-
1 overlays are applied at greater than 25 percent Alligator B cracking, where successive
application of HM-1 results in a 13 percent reduction in LCC.
• In the 35-year analysis, LCC savings range from 13 to 20 percent when HM-1 treatments are
applied successively.

UCPRC-TM-2007-08 67
 Table 64: LCC of Two HM-1 vs. Unlimited HM-1 Preventive Actions
(Rehabilitation at >25% Alligator B Cracking)—Dense-Graded Overlay
HM-1 Treatment = DGAC overlay
LCC over 20 Years in $1000/mi/ln LCC over 35 Years in $1000/mi/ln
Scenario Set No.2 - Rehabilitation Treated at Existing B Crack Pre-
(Cost decrease from two HM-1 to (Cost decrease from two HM-1 to
>25% Alligator B Crack Construciotn
unlimited HM-1 after Rehab) unlimited HM-1 after Rehab)
Low AADTT High AADTT Low AADTT High AADTT
Each rehab followed by two HM-1 treatments, HM-1
333.3 426.1
performed at (≤ 10%B) & (A+B ≤ 10%)
A+B<10% x≤10%
Rehab followed by only HM-1 treatments, HM-1
333.3 (0%) 369.6 (13%)
performed at (≤ 10%B) & (A+B ≤ 10%)
Each rehab followed by two HM-1 treatments, HM-1
332 333.8 425.4 447.7
performed at (≤ 10%B) & (A+B > 10%)
A+B≥10% x≤10%
Rehab followed by only HM-1 treatments, HM-1
332 (0%) 333.8 (0%) 360.2 (15%) 369.6 (17%)
performed at (≤ 10%B) & (A+B > 10%)
Each rehab followed by two HM-1 treatments, HM-1
396.5 484.3
performed at (> 25%B)
Rehab followed by only HM-1 treatments, HM-1 X>25%
performed at (> 25%B) 343.7 (13%) 387.7 (20%)

Table 65: LCC of Two HM-1 vs. Unlimited HM-1 Preventive Actions
(Rehabilitation at >25% Alligator B Cracking)—Chip Seal
HM-1 Treatment = AC Chipseal
LCC over 20 Years in $1000/mi/ln LCC over 35 Years in $1000/mi/ln
Scenario Set No.2 - Rehabilitation Treated at Existing B Crack Pre-
(Cost decrease from two HM-1 to (Cost decrease from two HM-1 to
>25% Alligator B Crack Construciotn
unlimited HM-1 after Rehab) unlimited HM-1 after Rehab)
Low AADTT High AADTT Low AADTT High AADTT
Each rehab followed by two HM-1 treatments, HM-1
309.9
performed at (≤ 10%B) & (A+B ≤ 10%)
A+B<10% x≤10%
Rehab followed by only HM-1 treatments, HM-1
309.9 (0%) 319.2 318.2 336
performed at (≤ 10%B) & (A+B ≤ 10%)
Each rehab followed by two HM-1 treatments, HM-1
performed at (≤ 10%B) & (A+B > 10%)
A+B≥10% x≤10%
Rehab followed by only HM-1 treatments, HM-1
337 337 368.1 368.1
performed at (≤ 10%B) & (A+B > 10%)
Each rehab followed by two HM-1 treatments, HM-1
307.8 398.4 315.4 477.8
performed at (> 25%B)
Rehab followed by only HM-1 treatments, HM-1 X>25%
performed at (> 25%B) 307.8 (0%) 328.7 (17%) 315.4 (0%) 359.8 (25%)

68 UCPRC-TM-2007-08
Based on costs shown in Table 65, in the case of an chip seals with conventional binders, the following
was observed:
• Very little data is available to identify trends.
• Life-cycle costs may be lower (possibly as much as 25 percent) if chip seals are applied
successively as opposed to placing a rehabilitation treatment after two preservation treatments.

Some of the differences in results observed from the 20- and 35-year analysis periods are due to the fact
that the 20-year analysis duration does not reflect the cost of the second rehabilitation (based on the
expected life of each treatment used in this analysis, shown in Table 46 to Table 52). Comparing different
combinations of preservation treatments was severely hampered by limitations in available data.
Substantially more projects with maintenance and rehabilitation histories and reasonably matched
pavement performance data are needed to draw more meaningful conclusions.
70 UCPRC-TM-2007-08
8 SUMMARY AND PRELIMINARY CONCLUSIONS

The purpose of this project is to develop performance estimates for pavement preservation treatments on
flexible pavements. This project is part of a larger UCPRC project (SPE 3.2.5) that has three main
objectives, with deliverables as follows:
• Track pavement preservation projects through performance databases.
o List of pavement preservation treatments
o Database of projects, pavement performance data, traffic, climate, and cost
• Estimate pavement performance for various strategies and application times.
o Pavement performance histories for each section
o Statistical analysis to identify performance probabilities
• Analyze life-cycle costs; recommend optimum timing and strategy selection (for those treatments
for which sufficient performance data is available).
o Calculated life-cycle costs
o Analysis of timing and relative performance
The final deliverable is documentation that shows results of the research. This technical memorandum
summarizes the results of the project, which are most likely the best ones that can be obtained from the
current Caltrans PMS database. Several gaps that exist in the database, which are summarized below, have
made all the conclusions presented in this report “preliminary” in nature. Changes are currently underway
to address these gaps in an updated PMS.

8.1 Tracking Pavement Preservation Projects

The UCPRC worked with HQ Division of Maintenance and contacted all Districts to identify projects and
retrieve project histories for treatments within the HM-1, CapM, and Rehabilitation programs. The
problem of finding maintenance and rehabilitation histories (i.e., what was done and when it was done) to
identify pavement preservation treatments applied or not applied still has not been solved despite
significant staff effort by Caltrans Maintenance to help find this information. Starting with 2,564 projects,
over two-thirds had to be omitted. Over 900 were not old enough to have sufficient performance history
(treatment performance in 2003 or later). A weighting algorithm developed by UCPRC to overcome or
mitigate the dynamic segmentation of the network in the PMS database was used to extract performance
data to try to link it with construction histories. Over 500 projects (beyond the 900 mentioned above) were
omitted because their performance history did not match construction information. Over 300 projects were

UCPRC-TM-2007-08 71
omitted due to duplicate construction entries or no completion data. As-built information was not available,
including the thickness of overlays, which must be a significant variable. Project funding programs (HM-1,
CapM, and Rehabilitation) were used as surrogates for overlay thickness.

The final database includes 718 projects. Most (82 percent) of the projects are located in three Caltrans
districts (2, 4, and 8). Nearly two-thirds of the treatments in the database are conventional dense-graded
asphalt overlays. The lack of construction histories for other treatments limits what can be concluded
about their performance and cost-effectiveness. The resulting database is the most complete available but
remains a limited and unbalanced representation of pavement treatments and performance histories.

It can be concluded from the exercise of putting together this database that significant changes are needed
in both pavement condition data collection and as-built data collection, as well as in the management of
Caltrans pavement-related databases, if Caltrans is going to be able to do a better job of determining
pavement performance and life-cycle costs than was possible in this study. Recommendations for changes
in practice are included at the end of this chapter.

8.2 Estimated Pavement Performance

Performance estimates were developed for sections where construction histories could be found, and they
reasonably matched expected pavement performance trends. For the sections included in the analysis,
pavement preservation treatments were applied at varying levels of cracking. The average (mean) value of
existing Alligator B cracking is 17 percent when placing HM-1 asphalt overlays, 16 percent for
rehabilitation overlays, and 12 percent for CapM. This suggests that pavement preservation using thin
overlays was being applied at high levels of cracking where CapM or rehabilitation may be more suitable.
This also indicates that the application of HM-1 asphalt overlays during the period 1988 to 2003 would not
generally be considered pavement preservation, because a relatively high level of cracking was present
when many overlays were placed.

The average value of existing Alligator B cracking was 10 percent when placing chip seals with
conventional asphalt, 13 percent for asphalt rubber chip seals, and 15 percent for polymer-modified chip
seals. This suggests that rubber and polymer-modified chip seals were placed on pavements with more
cracking. Comparatively few sections in the database deal with chip seals that use rubber and polymer-
modified binders. More sections are needed for inferences that have greater reliability.

72 UCPRC-TM-2007-08
After statistical evaluation of factors expected to affect pavement performance, it was decided to stratify
expected pavement life (number of years until 10 percent cracking and number of years until 25 percent
cracking), by existing cracking at the time of treatment application (less than 10 percent cracking, 10
percent to 25 percent cracking, and more than 25 percent cracking), rainfall (wet/dry), and traffic
(high/low). A statistical method used for estimating failure probability distributions was used to calculate
a median value of expected life for those preservation treatments with sufficient valid sections. Most
treatments have inadequate performance histories to produce a substantial number of median values. For
those with values, insufficient data prevented clear, consistent, and reasonable trends for some cases.

The median number of years to 10 percent Alligator B cracking for HM-1 overlays ranged from three to
eight years, with four and five years occurring most frequently (for all traffic and rainfall conditions)
whether existing Alligator B cracking at the time of treatment is less than 10 percent or more than
25 percent. Slight but reasonable trends were found, such as an association between high traffic and
shorter pavement life. However, unexpected trends also appear, such as wet conditions being associated
with longer pavement life. These are likely due to imbalances in the database.

Chip seals (based on a pooled dataset for conventional and modified binders), when applied where
Alligator B cracking is low, last at least five years (median) until 10 percent Alligator B cracking occurs.
As expected, high traffic substantially reduces service life. The effect of rainfall on chip seal performance
is not discernible.

8.3 Life-Cycle Cost Analysis

Analysis of life-cycle costs, based on the limited and biased database, enables order of magnitude
estimates and suggests trends. A more complete database of project maintenance and rehabilitation
histories is needed to obtain more precise estimates.

The LCCA focused on answering the following three questions:

• Is it more beneficial to apply pavement preservation than to perform no maintenance?
• Should pavement preservation be applied at an earlier or a later stage of cracking?
• How do life-cycle costs compare for different combinations of preservation treatment?

UCPRC-TM-2007-08 73
Answers to these questions were drawn only from data for dense-grade asphalt overlays and chip seals
with conventional binders because of lack of data for other treatments. The LCCA was performed based
on analysis periods of 20 and 35 years and produced similar trends for both periods.

Expected performance life (drawn from statistical analysis of observed performance histories) and inputs
from the Caltrans RealCost manual were used to calculate net present value. Scenarios for LCCA
compared projected performance of rehabilitation projects followed by (1) no maintenance until future
rehabilitation, (2) only two HM-1 treatments until future rehabilitation, and (3) only HM-1 treatments
without any future rehabilitation. These scenarios were evaluated for varying existing alligator cracking
and low vs. high traffic where sufficient data were available.

Results show savings in life-cycle costs from applying pavement preservation. The trends and magnitude
of savings are similar for both asphalt overlays and chip seals. Application of pavement preservation
between rehabilitations shows life-cycle costs 13 percent to 47 percent lower than rehabilitation without
preservation. This range is seen as indicating the order of magnitude and trend of life-cycle cost savings.
This level of detail is the best that appears reasonable, based on the limited data available.

Where preservation treatments are applied, LCCA was used to estimate cost savings from applying
treatments (thin overlay or chip seal) at an earlier or a later stage of cracking. For cases where sufficient
data are available, waiting until later stages of cracking results in life-cycle costs up to 14 percent higher
than if treatments are placed at an earlier stage of cracking. Some cases did not show any higher cost. As
with other conclusions, this may be due to the limited database.

Comparing various combinations of preservation treatments was severely hampered by the data limitations
mentioned above. Life-cycle costs were estimated for two cases. In the first case, rehabilitation is followed
by two treatments and another rehabilitation. In the second case, rehabilitation is followed only by
successive preservation treatments throughout the analysis period. For preservation with only thin overlays,
the second case (all preservation, no future rehabilitation) showed costs savings ranging from zero up to
20 percent. Where only chip seals were used, the second case cost savings ranged from zero up to percent
to 25 percent.

Substantially more projects with maintenance and construction histories and reasonably matched
pavement performance trends are needed to make more-meaningful and higher-confidence conclusions
about combinations of preservation treatments, timing of applying treatments, and quantitative benefits of
pavement preservation.

74 UCPRC-TM-2007-08
8.4 Recommendations

8.3.1 Pavement Preservation Practice

The estimates of pavement performance and life-cycle cost analysis based on the very limited data that
could be gathered and analyzed in this study provide some indication that life-cycle cost savings can be
obtained by use of the following pavement preservation practice:
• Use pavement preservation treatments, as opposed to using rehabilitation alone.
• Place pavement preservation treatments at lower levels of cracking than was the practice
from 1988 to 2003, with the comparison in this study being made between treatment at
less than or equal to 10 percent Alligator B cracking versus more than 25 percent
Alligator B cracking.

8.3.2 Caltrans PMS Data

A number of lessons were learned, or reemphasized from previous studies, throughout the pavement
preservation performance and life-cycle cost analysis. Subsequently, the following recommendations are
made to improve future studies of this type:
• Since one of the purposes of a PMS is to develop performance models relating maintenance and
rehabilitation to changes in the condition of the pavement, it is essential that additional data be
included in the PMS database that is not currently included. Such data consists of underlying
structure of the pavement, quality control records, and as-built information, including (at a
minimum) material type, thickness, and actual date of placement on the road. Project construction
histories were not available for maintenance performed directly by Caltrans forces, which means
that pavement distresses appear to be reduced in the condition survey data without any explainable
treatment in the construction information available. Inclusion of maintenance histories, including
treatment type, location, and date, will greatly increase the ability to connect pavement condition
to the work performed. It is also important that traffic data be tied to PMS data through relational
database rules. Linking of traffic data to other data must currently be done by hand via
spreadsheets, and traffic data must typically be downloaded by hand from PDF files.
• Improve quality of the pavement condition data by
o Conducting pavement survey on a fixed segmentation rather than current dynamic
segments, which will help in creating time histories necessary for developing pavement
performance models, and in identifying missing data from year to year and discrepancies
in data from year to year (such as pavement that improves in condition without any
treatment), and

UCPRC-TM-2007-08 75
 o Wherever possible, performing the pavement condition survey using measurements of
distress condition as opposed to identifying it as a binary variables (0 if present, 1 if not),
which do not have clear criteria and are evaluated visually.
o Wherever possible, take a more systematic approach, such as using an automated device
rather than the currently used “eyeball” visual inspection. Improved accuracy on
pavement condition data will help in extracting more-realistic performance estimates.

8.3.3 Pavement Preservation Test Sections

The Pavement Preservation Studies Technical Advisory Guide (PPSTAG) (8) guidelines were developed
for Caltrans by the UCPRC for use on controlled experiments that compared pavement preservation
treatments both to each other and to control sections (with no pavement preservation). Use of these
guidelines is recommended to help each Caltrans District in future efforts to set up field evaluation of
control and test sections. These experiments can provide more-detailed information than PMS data alone,
and can be used in conjunction with improvements in the PMS to perform more-accurate and more-
comprehensive analysis of pavement preservation life-cycle cost analysis than were possible with the data
available for the study presented in this technical memorandum.

8.3.4 Framework for Future LCCA Evaluations of Cost-Effectiveness of Pavement Treatments

It is recommended that the framework for performing life-cycle cost analysis evaluations of cost-
effectiveness of pavement treatments presented in this memorandum be used in the future by Caltrans.
Once the major gaps in the data identified in this project are addressed through improvements to the
Caltrans PMS—primarily consistent condition survey segmentation and better tracking of construction
histories and overlay thicknesses—this project should be repeated to draw more definitive conclusions.

76 UCPRC-TM-2007-08
 9. REFERENCES

1. University of California Pavement Research Center. Work Plan for “Performance Review of
Pavement Preservation Treatments.” September 1, 2006.
2. Lea, J., and Harvey, J. T. August 2002 (Revision December 2004). “Data Mining of the Caltrans
Pavement Management System (PMS) Database.” Draft report prepared for the California
Department of Transportation. Pavement Research Center, Institute of Transportation Studies,
University of California, Berkeley. UCPRC-RR-2002-04.
3. Kannekanti, V., Lee, C., Popescu, L., Harvey, J., and Nokes, W. November 2007. “Pavement
Performance Data Extraction from Caltrans’ PMS.” UCPRC-TM-2007-06.
4. Lu, Q., Harvey, J., Lea, J., Quinley, R., Redo, D., and Avis, J. June 2002. “Truck Traffic
Analysis Using Weigh-in-Motion (WIM) Data in California.” University of California
Pavement Research Center, Institute of Transportation Studies, University of California,
Berkeley. UCPRC-RR-2002-01.
5. Madanat, S., Nakat, Z., and Sathaye, N. October 2005. “Development of Empirical-Mechanistic
Pavement Performance Models Using Data from the Washington State PMS Database.” Davis
and Berkeley, California. UC Pavement Research Center. UCPRC-RR-2005-05.
6. Madanat, S., and Nakat, Z. “Development of an Empirical-Mechanistic Model of Overlay Crack
Progression Using Data from the Washington State PMS Database,” page 49. Davis and
Berkeley, California. UC Pavement Research Center. (In progress.)
7. See Interim RealCost User Manual updated April 2007 from Caltrans web
site: www.dot.ca.gov/hq/esc/Translab/OPD/LCCA_Manual_MASTER_4_1_07.pdf. Accessed
August 28, 2007.
8. Jones, D January 2007. “Pavement Preservation Studies Technical Advisory Guide.” Guideline
prepared for the California Department of Transportation, Division of Research and Innovation,
Office of Roadway Research, by the University of California Pavement Research Center, UC
Davis and Berkeley. UCPRC-GL-2005-01.

UCPRC-TM-2007-08 77
78 UCPRC-TM-2007-08
APPENDIX A: GLOSSARY

AADTT: Annual Average Daily Truck Traffic.

ACOL: Asphalt Concrete Overlay—placing layers of asphalt and inner membranes over an existing
roadway.

Alligator (fatigue) cracking: Cracks in asphalt that are caused by repeated traffic loadings. The cracks
indicate fatigue failure of the asphalt layer. When cracking is characterized by interconnected cracks, the
cracking pattern resembles that of an alligator’s skin.

Alligator A—A single crack or two parallel longitudinal cracks in the wheelpath; cracks are not
spalled or sealed; rutting or pumping is not evident.

Alligator B—An area of interconnected cracks in the wheelpath forming a complete pattern;
cracks may be slightly spalled; cracks may be sealed; rutting or pumping may exist.

CapM: Capital Preventive Maintenance: Use of heavy maintenance treatments such as intermediate
thickness asphalt blankets to provide five to seven years of additional pavement life. CapM projects are
thinner than Rehabilitation projects and add structural strength to the pavement.

HM-1: The highway program that funds routine and major maintenance on the state highway network.
The HM-1 program is preventive pavement repair work intended to preserve the system, retard future
deterioration, and prolong the service life of the pavement.

IRI—International Roughness Index: A standardized method of measuring the roughness of the pavement
surface, expressed in inches per mile or centimeters per kilometer, developed by the World Bank.

Type of pavement treatments:

• ACOL-DG: Dense-graded Asphalt Concrete Overlay
• ACOL-OG: Open-graded Asphalt Concrete Overlay
• ACOL-RAC: Rubberized Asphalt Concrete Overlay—a mixture of asphalt concrete containing rubber
“crumbs” and synthetic binders
• ACOL-RACO: Rubberized Open-graded Asphalt Concrete Overlay

UCPRC-TM-2007-08 79
• ChipSeal-AC: Asphalt Concrete Chip Seal—a surface treatment in which the pavement is sprayed
with asphalt and then immediately covered with aggregate and rolled with a pneumatic tire roller
• ChipSeal-AR: Rubberized Asphalt Concrete Chip Seal
• ChipSeal-PMA: Polymer Modified Asphalt Concrete Chip Seal
• ChipSeal-PME: Polymer Modified Emulsion Asphalt Concrete Chip Seal
• CrackSeal: Sealant applied to the crack lines on a pavement surface
• DigOut: Localized portion of the distressed pavement is removed, then patched with asphalt concrete
• SlurrySeal: A petroleum-based emulsion sealing coat (with embedded fine aggregates) applied to the
pavement surface

80 UCPRC-TM-2007-08
APPENDIX B: DEFINITION OF DATA CENSORING

Summarized and paraphrased from Leemis, L. M. Reliability: Probabilistic Models and Statistical
Methods. Prentice-Hall, 1995.

A data set is called complete when all failure times are known and censored if there are one or more
censored observations. A data set is called censored if there are one or more censored observations. There
are categories of censored data: right censored, left censored, and interval censored. Right censoring
occurs in long-lived pavements when one or more pavement sections do not have a final failure
observation. This typically occurs when data is being analyzed before all of the sections have failed (this is
a major limitation of the FHWA LTPP set of pavements, for example).

Left censoring occurs when data collection began after the pavements were constructed and failure
occurred prior to the first observation, when the time of construction is unknown, or when condition
survey procedures were changed and new variables were collected only for a certain, usually later, period
of time.

UCPRC-TM-2007-08 81
Interval censoring occurs when the pavement is checked periodically for failure, with gaps in the normal
data collection schedule, for example if an annual condition survey is not performed for several years. The
information available for the pavement section indicates that it failed sometime during the interval prior to
when failure was detected.

82 UCPRC-TM-2007-08
APPENDIX C: EXAMPLE CALCULATION OF PROBABILITY OF
FAILURE DISTRIBUTION FOR FIGURE 8.

Example: for HM-1 ACOL-DG performance under the following conditions:

(exABC<10%, Wet, LowAADTT)
Failure definition is set at 10% Alligator A crack.

Year (j) nj dj Kaplan-Meier Failure Rate Failure Rate

Estimate (CDF) (PDF)
S(t) F(t) = 1-S(t) f(t) = F(t)-F(t-
1)
0 0 0 1.000 0.000
1 77 2 0.974 0.026 0.026
2 75 5 0.909 0.091 0.065
3 62 3 0.865 0.135 0.044
4 45 5 0.769 0.231 0.096
5 29 3 0.689 0.311 0.080
6 23 4 0.570 0.430 0.120
7 10 1 0.513 0.487 0.057
8 3 1 0.342 0.658 0.171
9 2 2 0.000 1.000 0.342
10 2 0
11 2 0
12 2 0
13 2 0
14 2 0
15 2 0
16 2 0
17 2 0
18 2 0
19 2 0
20 2 0

Where n and d are defined by Equation 2 in Section 6.1. Ŝ(t) is the Kaplan-Meier survival rate estimator.

k
n −dj
Sˆ (t ) = ∏ ( j ) For tj-1 ≤ t <tj
j =1 nj
where, nj = number of pavements not failed and that are still being observed (termed “at risk”)
before tj and
dj = number of pavements failed in the time interval tj-1≤ t < tj.
The failure probability of an item before t is then Pr(x<T) = 1-Ŝ(t)

UCPRC-TM-2007-08 83