December 2005
Rev. February 2006
Research Report: UCPRC– RR- 2005- 11
Piilloott Prroojjeecctt ffoorr Fiixxeedd SSeeggmeennttaattiioonn ooff
tthhee Paavveemeenntt Neettwoorrkk
Authors:
E. Kohler, N. Santero, and
J. Harvey
Work Conducted as part of Partnered Pavement Research Center
Strategic Plan Element No. 3.2.4: Development of Integrated
Databases to Make Pavement Preservation Decisions
PREPARED FOR:
California Department of Transportation
Division of Research and Innovation
Office of Materials and Infrastructure
PREPARED BY:
University of California
Pavement Research Center
UC Davis, UC Berkeley
Stage 5 Distribution
UCPRC- RR- 2005- 11
ii
DOCUMENT RETRIEVAL PAGE Research Report: UCPRC– RR- 2005- 11
Title: Pilot Project for Fixed Segmentation of the Pavement Network
Author: Erwin Kohler, Nick Santero, and John Harvey
Prepared for:
Caltrans
FHWA No.:
CA081072B
Date:
December 2005
Revised February 2006
Strategic Plan No:
3.2.4
Status:
Draft
Version No:
Stage 5
Abstract:
The goal of this pilot project was to study a small sample of the Caltrans network to determine the feasibility of
expanding the pilot approach to the entire network. The project’s work included evaluating the effectiveness of
ground penetrating radar ( GPR) and limited coring for measuring pavement layer thicknesses and types, application
of an algorithm for determining “ fixed” segmentation of the pilot network, population of a database for the pilot
network, then assessing costs of these activities. Fixed segmentation for use in the Pavement Management System
( PMS) is required to develop the capability to do pavement performance modeling. Background information
summarizing the experiences of several other states in using GPR is included. The pilot network consisted of a total
of eight roadways: three interstate highways, four state routes, and one U. S. highway. GPR data was collected on
681 lane- miles of the pilot network and analyzed for 305 lane- miles. The research team collected coring data for
some of the locations on the pilot network and collected available as- built information and maintenance records, and
Pavement Condition Survey ( PCS) data. It then used the data collected to develop fixed segmentation for the pilot
network, resulting in 236 segments for the 305 lane- miles analyzed, with an average segment length of 1.27 miles.
Comparison of cores with the layer types and thicknesses identified by the GPR showed that the GPR data was
reliable, especially for the top two layers of the pavement. Extrapolation of the costs on the pilot network results in
an estimate of approximately $ 7.0 million of contracted field work consisting of GPR use and coring ( including
collection and analysis), plus 12.3 person- years of additional analysis work to complete the segmentation for the
entire Caltrans 49,000 lane- mile network.
Keywords:
Pavement segmentation, Ground Penetrating Radar, pavement structural inventory, road network evaluation,
Proposals for implementation:
- Conduct pavement structure inventory of full Caltrans network using GPR technology coupled with limited
traditional coring, to determine materials and thickness to populate a new Pavement Management System
- Implement a new type of pavement condition survey for PMS purposes. Document as- built information in the
structural database in order to keep the database accurate, as future maintenance, rehabilitation, and
reconstruction work occurs.
Related documents: Madanat, S., Nakat, Z., Sathaye, N. October 2005. Development of Empirical- Mechanistic
Pavement Performance Models using Data from the Washington State PMS Database. University of California
Pavement Research Center, Davis and Berkeley. UCPRC- RR- 2005- 05.
Signatures:
E. Kohler
1st Author
J. Harvey
Technical Review
D. Spinner
Editor
J. Harvey
Principal
Investigator
M. Samadian
Caltrans Contract
Manager
Stage 5 Distribution
UCPRC- RR- 2005- 11
iii
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.
PROJECT OBJECTIVES
The purpose of this research is to help Caltrans in developing a Pavement Management System ( PMS)
capable of performance prediction. The elements of the work are:
1. Identify existing and needed elements of databases;
2. Recommend database structures and data collection methodologies;
3. Populate databases for portion of the network to develop estimated costs, procedures, and
benefits for full network; and
4. Recommend final data collection and database operations.
Stage 5 Distribution
UCPRC- RR- 2005- 11
iv
TERMS AND ABBREVIATIONS USED IN THIS REPORT
Asphalt Concrete ( AC)
Asphalt- treated permeable base ( ATPB)
Average annual daily traffic ( AADT)
Bituminous base ( BB)
California Department of Transportation ( Caltrans)
Central Coast ( CC) climate region
Distance- measuring instrument ( DMI)
Dynamic Cone Penetrometer ( DCP)
Global Positioning System ( GPS)
Ground Penetrating Radar ( GPR)
Hot- mix asphalt ( HMA)
International Roughness Index ( IRI)
Inland Valley ( IV) climate region
Jointed plain concrete pavement ( JPCP)
Long- term Pavement Performance ( LTPP)
Maintenance and Rehabilitation ( M& R)
Open- graded asphalt concrete ( OGAC)
Partnered Pavement Research Center ( PPRC)
Pavement Condition Survey ( PCS)
Pavement Management System ( PMS)
Portland Cement Concrete ( PCC)
Strategic Highway Research Program ( SHRP),
Texas Transportation Institute ( TTI)
University of California Pavement Research Center ( UCPRC)
Washington State Department of Transportation ( WSDOT)
Stage 5 Distribution
UCPRC- RR- 2005- 11
v
EXECUTIVE SUMMARY
This report presents the results of the study, “ Pilot Project for Fixed Segmentation of the Pavement
Network.”
The goal of this pilot project was to study a small sample of the California Department of
Transportation ( Caltrans) network to determine the feasibility of expanding the pilot approach to the
entire pavement network. The project’s work included evaluating the effectiveness of ground penetrating
radar ( GPR) and limited coring for measuring pavement layer thicknesses and types, application of an
algorithm for determining “ fixed” segmentation of the pilot network, population of a database for the pilot
network, then assessing costs of these activities.
Fixed segmentation for use in the Pavement Management System ( PMS) is required to develop
the capability to do pavement performance modeling, which is essential for the following pavement
management tasks:
• Predicting future performance of segments of the network, and
• Identifying the most cost- effective maintenance and rehabilitation ( M& R) strategies based on
life- cycle costs.
Pavement layer- type and thickness data are also needed to develop effective pavement
performance models and to conduct effective condition surveys of composite pavements ( asphalt overlays
of PCC pavement). The data are also useful for project- level engineering.
Background information summarizing the experiences of several other states in using GPR for
pavement work is also presented.
The pilot network consisted of a total of eight roadways: three interstate highways ( I- 5, I- 505, and
I- 80), four state routes ( SR- 16, SR- 45, SR- 99, and SR- 113), and one U. S. highway ( US- 50). The
roadways chosen are mostly in District 3, except for the I- 80 section and part of the I- 505 section, which
are both in District 4. GPR data was collected on 681 lane- miles of the pilot network and analyzed for 305
lane- miles. Traffic data was obtained from Caltrans. Climate regions were determined from a recent map
developed by Caltrans and the University of California Pavement Research Center ( UCPRC).
The UCPRC collected coring data for some of the locations on the pilot network. Some of the
cores were provided to Infrasense, Inc., for GPR calibration and some were retained by the UCPRC for
checking the accuracy of the layer thicknesses and types that Infrasense determined from the GPR data.
Stage 5 Distribution
UCPRC- RR- 2005- 11
vi
The UCPRC also collected available as- built information and maintenance records, and the most recent
Pavement Condition Survey ( PCS) data from Caltrans.
The UCPRC then used the data collected to develop fixed segmentation for the pilot network,
resulting in 236 segments for the 305 lane- miles analyzed, with an average segment length of 1.27 miles.
Comparison of the cores retained by the UCPRC with the layer types and thicknesses identified
by the GPR showed that the GPR data was reliable, especially for the top two layers of the pavement.
Extrapolation of the costs on the pilot network for data collection and analysis and segmentation
results in an estimate of approximately $ 7.0 million of contracted field work consisting of GPR use and
coring ( including collection and analysis), plus 12.3 person- years of additional analysis work to complete
the segmentation for the entire Caltrans 49,000 lane- mile network.
If Caltrans moves ahead with collection of pavement structure data and fixed segmentation, it will
be important to document as- built information in the structural database as future maintenance,
rehabilitation, and reconstruction work occurs, in order to keep the database accurate.
Work beyond this pilot study is underway to determine:
• Whether PMS performance data can be used with the fixed segments to develop reasonable
performance histories for the segment, and
• Whether the performance models developed by the UCPRC from Washington State
Department of Transportation ( WSDOT) PMS data can be verified with Caltrans PMS
performance histories using the fixed network segments and other necessary data developed
in this pilot project.
Stage 5 Distribution
UCPRC- RR- 2005- 11
vii
TABLE OF CONTENTS
Terms and Abbreviations Used in this Report ............................................................................................. iv
Executive Summary ............................................................................................................................... ...... v
List of Figures ............................................................................................................................... .............. ix
List of Tables ............................................................................................................................... ................ x
1.0 Introduction................................................................................................................... .................. 1
1.1 Purpose of Work.......................................................................................................................... 2
1.2 Pilot Project ............................................................................................................................... . 2
1.3 Scope, Schedule, and Status of Project Tasks ............................................................................. 3
2.0 Background..................................................................................................................... ................ 7
2.1 Network Segmentation for Pavement Management .................................................................... 7
2.1.1 Performance Modeling............................................................................................................ 8
2.1.2 Challenges to Pavement Modeling Using the Current PMS Studied in This Project.............. 8
2.2 GPR Technology ....................................................................................................................... 10
2.2.1 Brief Description of the Technology..................................................................................... 10
2.2.2 Recent Experience with GPR by Caltrans and Other State DOTs ........................................ 11
3.0 Draft Segmentation ........................................................................................................................ 14
3.1 Administrative Boundaries ........................................................................................................ 17
3.2 Traffic........................................................................................................................ ............... 17
3.3 Pavement Structure.................................................................................................................... 17
3.4 Climate Region......................................................................................................................... 19
3.5 Condition Survey....................................................................................................................... 19
4.0 Data Requirements and Resources Involved ................................................................................. 21
4.1 Pilot Study ............................................................................................................................... . 21
4.1.1 Traffic Database .................................................................................................................... 21
4.1.2 GPR Data .............................................................................................................................. 21
4.1.3 Coring......................................................................................................................... .......... 24
4.1.4 As- Builts ............................................................................................................................... 24
4.1.5 Climate ............................................................................................................................... .. 25
4.1.6 Condition Survey and IRI ..................................................................................................... 25
4.1.7 Database ............................................................................................................................... 25
4.1.8 Summed Effort ...................................................................................................................... 26
4.2 Extrapolated Cost and Effort to Whole Network ...................................................................... 26
Stage 5 Distribution
UCPRC- RR- 2005- 11
viii
5.0 Discussion of Results from the Pilot Project and Remaining Work .............................................. 28
5.1 Utilization of Coring Data ......................................................................................................... 28
5.1.1 Core Sites .............................................................................................................................. 28
5.1.2 Determining Exact Core Locations ....................................................................................... 29
5.1.3 Core Results .......................................................................................................................... 30
5.2 Comparison of Current Caltrans Maintenance Dynamic Segmentation Versus Fixed Length
Segmentation ............................................................................................................................... .. 33
5.3 Collection of Pavement Condition Indicators for Performance Modeling in PMS................... 34
6.0 Conclusions, Future Work, and Recommendations....................................................................... 36
6.1 Conclusions ............................................................................................................................... 36
6.2 Recommendations ..................................................................................................................... 36
7.0 References..................................................................................................................... ................ 38
APPENDICES ............................................................................................................................... ............ 39
Appendix A. Minutes - 8/ 30/ 04 Meeting On Developing Objectives for theHighway Network
Segmentation & Data Collection In D- 3 Using GPS...................................................................... 40
Appendix B. GPR Survey Summary ............................................................................................. 45
Appendix C. Charts with GPR Structure Results and Data from the 2003 Pavement Condition
Survey ............................................................................................................................... .. 46
Appendix D. Segmentation Results ............................................................................................... 59
Appendix E. GPR Data and UCPRC Core Comparison: Plots ..................................................... 66
Appendix F. GPR Data and UCPRC Core Comparison: Tables................................................... 71
Appendix G. Recommendations for Changes to Caltrans Pavement Condition Survey ............... 76
Stage 5 Distribution
UCPRC- RR- 2005- 11
ix
LIST OF FIGURES
Figure 1. Location of roadways in relation to the entire network............................................................... 15
Figure 2. Location of roadways in North Central, California. ................................................................... 16
Figure 3. Example of GPR data taken on a GPR section of I- 5 in Sacramento County. ........................... 19
Figure 4. GPR equipment used on this project. ......................................................................................... 22
Figure 5. GPR versus core ( UCPRC) thicknesses. ..................................................................................... 32
Figure 6. Histograms of sections versus length with Caltrans segmentation............................................. 33
Figure 7. Histogram of sections versus length with fixed- length segmentation. ....................................... 34
Figure 8. GPR cross section with selected PCS data – Sacramento SR- 99 SB ( outside lane), CAL009.... 47
Figure 9. GPR cross section with selected PCS data – Sacramento I- 5 SB ( outside lane), CAL011. ........ 48
Figure 10. GPR cross section with selected PCS data – Sacramento and Sutter SR- 99 NB ( outside lane),
CAL013......................................................................................................................... .................... 49
Figure 11. GPR cross section with selected PCS data – Yolo SR- 113 NB ( outside lane), CAL015.......... 50
Figure 12. GPR cross section with selected PCS data – Sacramento I- 5 SB ( inside lane), CAL017. ........ 51
Figure 13. GPR cross section with selected PCS data – Solano I- 80 WB ( outside lane), CAL031. .......... 52
Figure 14. GPR cross section with selected PCS data – Sacramento and Yolo I- 15 NB
( outside lane), CAL033. ..................................................................................................................... 53
Figure 15. GPR cross section with selected PCS data – Colusa and Yolo SR- 16 WB
( outside lane), CAL035. ..................................................................................................................... 54
Figure 16. GPR cross section with selected PCS data – Solano I- 80 WB ( inside lane), CAL041. ............ 55
Figure 17. GPR cross section with selected PCS data – Sacramento US- 50 EB ( ouside lane), CAL047. . 56
Figure 18. GPR cross section with selected PCS data – Yolo SR- 45 SB ( outside lane), CAL049a........... 57
Figure 19. GPR cross section with selected PCS data – Yolo and Solano I- 505 SB
( ouside lane), CAL050. ...................................................................................................................... 58
Figure 20. GPR/ core thickness - Solano 505 SB, CAL050- 1a. .................................................................. 66
Figure 21. GPR/ core thickness - Yolo 113 NB, CAL015- 5. ...................................................................... 66
Figure 22. GPR/ core thickness - Yolo 113 NB, CAL015- 5a...................................................................... 67
Figure 23. GPR/ core thickness - Sacramento 50 EB, CAL047- 9............................................................... 67
Figure 24. GPR/ core thickness – Sacramento 50 EB, CAL047- 10. ........................................................... 68
Figure 25. GPR/ core thickness - Yolo 45 SB, CAL049- 11. ....................................................................... 68
Figure 26. GPR/ core thickness - Yolo 45 SB, CAL049- 12. ....................................................................... 69
Figure 27. GPR/ core thickness - Yolo 45 SB, CAL049- 12a. ..................................................................... 69
Figure 28. GPR/ core thickness - Sacramento 99 NB, CAL009- 14,15........................................................ 70
Figure 29. GPR/ core thickness - Colusa 16 WB, CAL035- 16,17,18. ........................................................ 70
Stage 5 Distribution
UCPRC- RR- 2005- 11
x
LIST OF TABLES
Table 1. Initial Sections for Segmentation Pilot Project ..................................................................... 14
Table 2. Personnel Needed and Direct Cost for Segmentation of Pilot Study and Estimated for Entire
Caltrans Network........................................................................................................................ ....... 27
Table 3. Final List of GPR Coring Locations ..................................................................................... 28
Table 4. Selected Coring Locations and Physical Reference Points ................................................... 30
Table 5. Average Thickness Differences ( Absolute) and Standard Deviations .................................. 32
.
Stage 5 Distribution
UCPRC- RR- 2005- 11
1
1.0 INTRODUCTION
This report presents the findings of the study, “ Pilot Project for Fixed Segmentation of Pavement
Network.”
The goal of this pilot project was to study a small sample of the California Department of
Transportation ( Caltrans) network to determine the feasibility of expanding the pilot approach to the
entire pavement network. The project’s work included evaluating the effectiveness of ground penetrating
radar ( GPR) and defining “ fixed” pavement segments, then assessing costs of these activities.
The work was conducted as part of the Partnered Pavement Research Center ( PPRC) Strategic
Plan Element 3.2.4 (“ Development of Integrated Databases to Make Pavement Preservation Decisions”)
for the following objectives.
• Identify Caltrans pavement data business practices and the elements of the pavement
databases that already exist, and work with Caltrans pavement organizations to perform a
needs analysis for pavement data.
• Develop recommended changes to pavement data business practices. Develop recommended
database tables and dictionaries for these databases, determine which variables are missing
and those that are currently being collected unnecessarily, and identify key issues ( such as the
linear reference system and Caltrans information technology requirements) that must be
resolved before the databases can be integrated.
• Based on the objectives listed above, prepare a report summarizing the findings and making
recommendations for changes.
• Populate databases with existing data and perform preliminary analyses.
• Develop recommendations for ongoing collection and database management procedures to be
implemented and operated by Caltrans functional units.
Work underway on analysis of the data collected is part of, or coordinated with, activities in
PPRC Strategic Plan Elements 3.2.5 (“ Documentation of Pavement Performance Data for Pavement
Preservation Strategies and Evaluation of Cost- Effectiveness of Such Strategies”) and 4.5 (“ Calibration of
Mechanistic- Empirical Design Models”).
Stage 5 Distribution
UCPRC- RR- 2005- 11
2
1.1 Purpose of Work
The general purpose of the work presented in this report is to support Caltrans Maintenance in its
development of an improved Pavement Management System ( PMS). Specific objectives focus on helping
Caltrans Maintenance develop the capability to do pavement performance modeling, which is essential for
the following pavement management tasks:
• Predicting future performance of segments of the network.
• Identifying the most cost- effective maintenance and rehabilitation ( M& R) strategies based on
life- cycle costs.
More specific purposes of the work address three key challenges to performance modeling using
the current Caltrans PMS.
1. The use of “ dynamic segmentation,” which has logistical benefits but masks the true
performance of fixed segments and confounds performance modeling. The current system
uses a “ dynamic segmentation” procedure in which the pavement is not evaluated over fixed
lengths but is divided into segments that have similar distress at the time of each assessment.
Consequently, both the segment’s length and its starting and ending points change from year
to year, and a given pavement section is identified as appearing in a different segment from
one year to the next.
2. The PMS database lacks subsurface pavement structure data, which is a key variable in
explaining pavement performance. Pavement structure cross- section data is not available in
any central or district Caltrans database and it is not routinely updated when rehabilitation
and maintenance activities are performed.
The nonexistence of quantification ( severity and/ or extent) for some pavement distresses, which
means that these distresses are observed and identified in the Pavement Condition Survey ( PCS), but they
are not measured.
1.2 Pilot Project
To meet the stated objectives, a pilot project was developed in which a small representative sample of the
Caltrans network in Districts 3 and 4 was selected for field testing and other data collection and analyses.
These efforts aimed at evaluating:
Stage 5 Distribution
UCPRC- RR- 2005- 11
3
• The effectiveness of using ground penetrating radar ( GPR) data, limited coring, and available
collected office data to provide an uninterrupted measurement of pavement thickness and
layer type on a variety of pavement types.
• The effectiveness of establishing static, well- defined ( fixed) network segments using the GPR
and other data collected on the pavement structures — combined with traffic, climate, and
condition survey, and roughness data.
• The costs of collecting the data and performing the segmentation and extrapolation of those
costs to the entire pavement network.
This report presents the data collected and the results of the analyses performed to complete these
three evaluations. Work will continue outside this pilot project, with additional analyses to be performed
to definitively conclude:
• Whether PMS performance data can be used with the fixed segments to develop reasonable
performance histories for the segment, and
• Whether the performance models developed by the UCPRC from Washington State
Department of Transportation ( WSDOT) PMS data can be verified with Caltrans PMS
performance histories using the fixed network segments and other necessary data developed
in this pilot project.
1.3 Scope, Schedule, and Status of Project Tasks
Specific tasks to be completed for this pilot project were identified in Meeting Minutes from August
30, 2004, “ On Developing Objectives for the Highway Network Segmentation & Data Collection in
District 3 Using GPR” ( Appendix A). The initial project scope was shown as follows:
Stage 5 Distribution
UCPRC- RR- 2005- 11
4
1. Collect GPR data on identified sections in Districts 3 and 4.
• Collect approximately 1,000 lane- miles of data, analyze approximately 300 lane- miles, and
retain the remaining raw data for potential analysis later.
• Include ( a) low- volume and high- volume traffic segments, and ( b) rigid, flexible, and
composite pavement structures.
􀃆 Tasks completed and items are presented in this report. Routes identified to include 1,000 lane- miles
actually consisted of about 681 lane- miles when measured ( see Appendix B).
2. Collect other data, including:
• The Caltrans Office of Pavement Rehabilitation’s studies of deflections,
• Project as- builts ( headquarters [ HQ] data, retrieval [ intranet] of District data),
• Data from moisture sensitivity studies, and
• Data from the Pavement Performance Evaluation Phase I ( Stantec Project) 1
􀃆 Tasks completed and utilized in this report.
• Coring Data
• Some samples are to be provided to the GPR contractor to calibrate the GPR data, and
others are to be held by the PPRC to verify GPR measurements
• Perform coring at only few locations, and only in sections where the GPR data has been
analyzed.
􀃆 Tasks completed for selected sites in Districts 3 and 4.
3. Perform analyses
• Analyze GPR data for thickness and layer type,
• Map the structures,
1 Stantec, Inc. was awarded a research project by the Caltrans Division of Research and Innovation to evaluate the performance
of in- service pavements in California and hence, the success of Caltrans' pavement design and rehabilitation procedures. As part
of this project, a large number of sections distributed throughout the state of California covering different districts and
environmental zones are being tested and many pavement related data attributes are being collected. The test sections include
rigid pavements, composite pavements and new and rehabilitated flexible pavements. Phase II of this project is currently
underway and is expected to be completed in the summer of 2006.
Stage 5 Distribution
UCPRC- RR- 2005- 11
5
• Revise GPR structures results based on coring data in areas where GPR identification is
questionable,
• Compare verification data with analyzed GPR data, and
• Analyze the costs.
􀃆 Task completed and results are included in this report.
Tasks A to C were scheduled to be completed in June 2005 and to be followed by:
4. Segment the 300 analyzed lane- miles by following a procedure ( described in the minutes)
that accounts for administrative units, pavement structure, climate region, traffic loading, and
condition survey, and ride quality data. Complete this item in August 2005.
􀃆 Task completed. PPRC performed a preliminary segmentation of the network based on traffic, climate,
pavement structure ( based on GPR data and verified by selected as- builts and GPR core data), condition
survey, and International Roughness Index ( IRI) data. The results are included in this report.
Additional scope added to the project later by Caltrans Maintenance includes the following tasks.
5. Extract historical condition survey and IRI data from the Caltrans PMS database for the 300
analyzed lane- miles.
􀃆 This task has been completed using the last available Caltrans Pavement Condition Survey ( 2003–
2004) based on the fixed segmentation completed as part of Task D and included in this report.
Additional condition survey and IRI data are being extracted as part of the PPRC Strategic Plan Item
3.2.5 from previous years and maintenance and rehabilitation histories. A separate report will be
delivered.
6. Check the accuracy of performance prediction models being developed as part of Item 4.5 of
the Strategic Plan for asphalt overlays on asphalt pavement, and IRI of flexible and rigid
pavement against extracted condition survey and IRI performance histories.
Stage 5 Distribution
UCPRC- RR- 2005- 11
6
7. Completion of this task is not guaranteed because of the dynamic segmentation present in the
California PMS condition survey data. The data collected under Task E as part of the PPRC
Strategic Plan item 3.2.5 will be used in the attempt to complete this task, which is scheduled
to be completed in March 2006.
Stage 5 Distribution
UCPRC- RR- 2005- 11
7
2.0 BACKGROUND
Adequate segmentation of a highway network is fundamental for the successful utilization of a Pavement
Management System ( PMS), in particular for the use of pavement deterioration models. The
homogeneous segments resulting from the segmentation process need to have a consistent traffic level
and a comparable pavement structure, and need to correspond to a single climate region. ( Section 2.1
presents a detailed discussion of pavement segmentation.)
A key part of the segmentation process is the pavement structure, in terms of materials and layer
thicknesses. Since Caltrans does not presently have adequate inventory information about the pavement
structure throughout the network, the feasibility of using ground- penetrating radar ( GPR) for this purpose
is being evaluated. A brief literature review on GPR is presented in Section 2.2.
2.1 Network Segmentation for Pavement Management
Long pavement segments in a PMS will generally be less uniform in composition ( i. e., there will be more
variation in pavement structure, condition, and other attributes within a segment) than short segments.
However, short segments require more data storage space because of the increased number of segments.
The final decision on size and method of segmenting should be based on selecting pavement segments
that Caltrans will consider as single entities when planning maintenance and rehabilitation. The smallest
number of segments that can adequately define the road network will be the most economical and easiest
to maintain.
As outlined in a previous report ( 1), Caltrans first implemented a PMS in 1977, when the concept
of pavement management was relatively new and computers were not as powerful as they are today. Over
the subsequent twenty- five years, advances in computer technology and significant changes in the theory
and practice of pavement management have changed the way pavements are maintained by Caltrans.
These changes have led to the slow evolution of the Caltrans PMS database and its use within the agency.
In today’s PMS literature, the Caltrans system would be referred to as a maintenance management system
because it is geared toward providing information for short- term maintenance activities rather than long-term
pavement performance assessment and modeling as well as optimization of expenditures for the
pavement network.
Stage 5 Distribution
UCPRC- RR- 2005- 11
8
2.1.1 Performance Modeling
Performance modeling using PMS field data is essential for continuous improvement of two key Caltrans
pavement management tasks at the network level:
• Predicting future performance of segments of the network. “ Performance” refers to pavement
surface distress in the annual condition survey and ride quality ( IRI). Future performance is
predicted using models of distress and ride quality as functions of existing condition,
structure, traffic, and climate, and maintenance and rehabilitation strategy selection.
• Identifying the most cost- effective maintenance and rehabilitation strategies based on life-cycle
costs. Life- cycle costs can be calculated for different conditions across the state
network, but the calculation requires the models described above to predict performance at
the network level plus cost data for each strategy.
At the network level, performance models derived from observations are “ empirical.” A
pavement performance model becomes “ empirical- mechanistic” when the explanatory variables are
selected based on the mechanics of pavement damage. To make these models useful for Caltrans
management, they must be calibrated using PMS field data. Compared to project- level design, inputs for
network performance modeling ( structure, traffic, and climate) need a lower level of detail. Collecting
data across the network with project- level detail would be cost- prohibitive.
Project- level PMS data for specific segments of the network is needed for calibrating
“ mechanistic- empirical” design procedures, which rely more heavily on pavement damage mechanics
theory. Detailed data for pavement structure, traffic, climate, materials, and construction quality must be
collected from the segments in order to predict their performance. Those models must then be calibrated
using historical PMS condition survey and ride quality data.
2.1.2 Challenges to Pavement Modeling Using the Current PMS Studied in This Project
As mentioned at the beginning of this report, three crucial aspects of the current PMS are addressed in
this project to enable performance modeling ( at the network level) and to calibrate design procedures ( at
the project level).
1. The use of “ dynamic segmentation,” which constantly shifts frame of reference;
2. Lack of inputs needed for modeling, because the PMS does not contain data about subsurface
pavement structure; and
Stage 5 Distribution
UCPRC- RR- 2005- 11
9
3. Inadequate quantification of pavement distresses, as some parameters in the Pavement
Condition Survey ( PCS) do not relate to pavement distress mechanisms or, if observed, are
only identified as present but are not measured.
The current Caltrans PMS staff inherited a “ dynamic segmentation” procedure established in
1977 in which pavement is not evaluated over fixed lengths. Instead, the pavement is divided into
segments that have similar distress at the time of each assessment. Consequently, both the segment’s
length and its starting point and its ending point change from year to year. As a result, a given pavement
section is often identified as appearing in a different segment from one year to the next. Often,
segmentation from year to year changes based solely on the effects of short- lived maintenance treatments
that do not change the pavement cross section. Therefore measured distresses and ride quality in the PMS
database can vary as a function of segmentation, depending on which sections of pavement are grouped
together within the segment. Although this is a good approach for scheduling maintenance, it does not
lend itself to statistical sampling of observed performance data or to predicting performance over time. It
may also result in inefficiencies for scheduling the rehabilitation of parts of a section as they fail over
several years; in reality, the entire section of which they are a part might be failing. Effective performance
modeling requires a network of “ fixed segments,” reasonably consistent pavement variables ( e. g.,
structure, traffic, climate), and similar maintenance and rehabilitation history.
The second challenge arises from the biggest problem with extracting pavement performance
information from the database: the database contains little information regarding pavement structure. In
some cases it contains data specifying whether the pavement surface is flexible ( asphalt concrete) or rigid
( portland cement concrete). In other cases, the database contains a generic description of apparent mix
type, such as open- graded or dense- graded asphalt. Missing are data about the true materials and layer
thicknesses beneath the surface, which are among the most important variables that explain pavement
performance. Without these, pavement performance models often give useless results or incorrect results.
The third challenge arising from the current PCS that Caltrans uses comes from its inclusion of
several variables whose presence or absence is noted but not measured, and from others that have no
meaning in terms of pavement distress mechanisms. This challenge can be met by making some relatively
minor changes in the PCS.
Stage 5 Distribution
UCPRC- RR- 2005- 11
10
2.2 GPR Technology
2.2.1 Brief Description of the Technology
Ground- penetrating radar ( GPR) pavement- related technology, which was developed during the Strategic
Highway Research Program ( SHRP), operates by transmitting short pulses of electromagnetic energy into
the pavement. These pulses are reflected back to the radar antenna with an amplitude and arrival time that
is related to the thickness and material properties ( dielectric constant) of the pavement layers ( 2).
GPR technology has the potential of being extremely useful for pavement management, allowing
highway agencies to quickly collect inventory data on all pavements under their jurisdiction ( 3, 4).
Because GPR data collection is nondestructive, it substantially reduces the need for frequent full- depth
pavement coring. Thickness determination of existing pavement layers employing GPR is standardized in
ASTM D4748. ( 5)
GPR is a high- resolution geophysical technique that utilizes electromagnetic radar waves to scan
shallow subsurfaces, to provide information on pavement layer thickness or to locate targets. The
frequency of the GPR antenna affects the depth of penetration into the pavement. Lower- frequency
antennas penetrate further than higher frequency ones do, but the latter type yield higher resolution. To
successfully provide pavement thickness information or to scan an interface, the following conditions
have to be present ( 6): ( 1) The physical properties of the pavement layers must allow for penetration of
the radar wave, ( 2) the interface between pavement layers must reflect the radar wave with sufficient
energy for it to be recorded, and ( 3) there must be a significant difference in the physical properties of the
layers separated by interfaces.
In NCHRP Synthesis 255 ( 7) the capabilities of GPR systems are listed as2:
• Asphalt layer thickness determination: GPR results are used to estimate thickness to within
10 percent; GPR accurately measures thicknesses of up to 0.5 m.
• Base thickness determination: Thicknesses are estimated, provided that a dielectric contrast
between the base and subgrade exists. ( The best results occur when the subgrade is made up
of clay soils, which are highly conductive compared to sands or gravels.)
• Concrete thickness determination: Depth constraints and accuracy are not yet well defined.
This is because portland cement concrete attenuates GPR signals more than asphalt does;
PCC conductivity changes as the cement hydrates; reinforcing steel contained in slabs makes
2 NCHRP Synthesis 225 was published in 1998 and it therefore reflects the state of the GPR technology as it was
more than eight years ago.
Stage 5 Distribution
UCPRC- RR- 2005- 11
11
interpretation difficult; and the dielectric contrast between PCC and the base may not be
adequate for reflection detection.
• Void detection: Although GPR has detected air- filled voids as thin as 6 mm, the detection of
water- filled voids is more problematic.
2.2.2 Recent Experience with GPR by Caltrans and Other State DOTs
As nondestructive testing has become an integral part of pavement evaluation and rehabilitation strategies
in recent years, Caltrans and other state highway agencies have looked into GPR technology for network
inventory and at the project level.
2.2.2.1 Caltrans
An evaluation of GPR and other non- destructive techniques for pavement thickness evaluation was
carried out for Caltrans by Infrasense, Inc ( 8). The work focused on determining quality control accuracy
in newly constructed asphalt and concrete pavements. The work involved theoretical analysis, laboratory
testing on small slabs and simulated pavement materials, testing at full- scale testing facilities, and actual
testing on recently constructed pavement sections in California. The actual testing was carried out on
eleven selected pavement sections, six of asphalt and five of concrete. Test sections were 305 meters
( 1,000 feet) long. The asphalt sites were selected to represent three main conditions: ( a) thick and thin
asphalt on aggregate base; ( b) asphalt on concrete; and ( c) thick and thin asphalt overlays. The concrete
sites were selected to represent variations in concrete thickness and age. Age was selected as a variable
because of its influence on GPR penetration and on the mechanical wave velocity. The asphalt sites were
tested with the horn antenna ( typical GPR test) method and the common midpoint, ( or CMP, a semi- static
GPR) method. The concrete pavements were evaluated with two different impact- echo devices, along
with the CMP method. After this evaluation, cores were taken for comparison with the test data. Twenty
cores were taken at each asphalt site and ten at each concrete site. The thickness values determined from
the various test methods were compared to the core values. The comparison showed generally good
correlation, but at each site a calibration was also needed. One core location per site was selected for
calibration.
For asphalt pavement, the GPR was found capable of measuring the average section thickness to
within 2.5 mm ( 0.1 inches) of the average core value. This level of accuracy was not achieved on concrete
pavements.
Stage 5 Distribution
UCPRC- RR- 2005- 11
12
2.2.2.2 Indiana
In 2001 the Indiana DOT ( 6) conducted experimental evaluation of the GPR for network inventory by
taking measurements at sections in five interstate highways ( I- 64, I- 65, I- 69, I- 70, and I- 74), five U. S.
highways, and nine state routes. GPR was used to test the truck lane for both directions of traffic ( east-west
or north- south) of each selected roadway at highway speed. Although GPR can display pavement
layer thickness continuously, it was decided to collect thickness data at only five incremental locations
( every 1,000 ft, or 300 m) of each mile. As part of the study, the researchers also obtained an estimate of
the total pavement thickness using FWD testing, which complemented data from the GPR tests regarding
the thickness of the top surface portion of the combined surface layers. Top surface portion thickness
information is very important for situations in which mill- and- fill operations are needed. The GPR
estimates of concrete pavement thickness, of hot mix asphalt ( HMA) thickness of flexible pavements, and
HMA thickness of composite pavements matched almost perfectly. GPR thickness estimate of pavement
layers underneath these layers was not as accurate and needs adjustment through very limited coring.
GPR did not provide any estimate of unbound pavement layers or of total pavement thickness.
The relevant conclusions of the study are the following:
• Network- level testing employing the FWD and GPR is a worthwhile, technically sound
program that provides a baseline of the structural capacities of in- service pavements.
• GPR is not expected to completely eliminate the need for coring, although GPR can be used
to establish the coring requirements, fill the gaps in thickness estimation, and verify thickness
results.
2.2.2.3 Virginia
Al- Qadi et al. ( 9) report that GPR was used to evaluate the layer thicknesses of seventeen pavement sites
of different types ( flexible, continuously reinforced, and jointed plain concrete) and different pavement
ages ( up to five years old, between ten and fifteen years old, older than twenty years with a surface less
than ten years old; and older than twenty years with a surface older than ten years). The sites were located
in different parts of Virginia on major interstates and high traffic- volume roads.
Analysis of the GPR data collected from all sites showed that for flexible pavements, the GPR
thickness error increased with pavement age ( 4.4 percent error for pavements up to five years old to 5.8
percent error for pavements older than twenty years with surfaces older than ten years). Comparison of
sites of the same age but with different pavement types showed that flexible pavements had a relatively
high thickness error, while the jointed plain concrete pavement ( JPCP) had the lowest thickness error.
Stage 5 Distribution
UCPRC- RR- 2005- 11
13
This could be mainly due to the presence of thin HMA layers in flexible pavements ( these layers are
significantly smaller than the GPR signal’s wavelength) as well HMA layers of different ages. GPR
considers layers with the same dielectric constant as one homogeneous layer, thus sometimes introducing
an error in the thickness computation.
The study concluded that the error produced in predicting the thickness of HMA and concrete is
very reasonable, and that GPR accuracy in predicting pavement layer thicknesses surpasses other
available techniques — with the exception of coring, which is time- consuming, has a very low coverage
area, and is considered a destructive technique that requires traffic closures.
2.2.2.4 North Carolina
In North Carolina ( 13), thirteen LTPP ( Long Term Pavement Performance) sites were tested one or more
times with GPR to obtain layer thickness variability over 152.4- m ( 500- ft) test sites. Duplicate runs were
made on the same day on one of the sites, and these paired tests were compared after the GPR data were
processed. Five of the sites showed good agreement with a Student’s t- test. Asphalt layers for the sites
varied in average thickness between 89 mm and 292 mm ( 3.5 and 11.5 in.). Thinner asphalt layers tended
to have lower coefficient of variation when the asphalt thickness was less than 152 mm ( 6 in.). The
standard deviation was generally less than 25 mm ( 1 in.).
2.2.2.5 Other DOT agencies
Other DOT agencies recently involved in verification of GPR technology are New Jersey ( 10), Missouri
( 11), and Kentucky ( 12). They report good results for thickness determination. The Florida and Texas
Departments of Transportation both own GPR equipment. The Florida DOT uses GPR primarily to
establish pavement thicknesses. In Texas, the Materials Division of the Texas Transportation Institute
( TTI) has developed performance specifications and test procedures for GPR systems. TTI has also
developed a GPR training program that has been used to train Texas DOT personnel in the two state
districts that own and use GPR.
Stage 5 Distribution
UCPRC- RR- 2005- 11
14
3.0 DRAFT SEGMENTATION
The initial step in the segmentation pilot project was to identify highways and routes to be analyzed in the
study. A total of eight roadways were selected: three interstate highways ( I- 5, I- 505, and I- 80), four state
routes ( SR- 16, SR- 45, SR- 99, and SR- 113), and one U. S. highway ( US- 50). The roadways chosen are
mostly in District 3, except for the the I- 80 section and the southern part of the I- 505 section, which are in
District 4. The research team selected these roads, which span five counties, for the pilot program because
they believed that the extent and diversity of their pavement sections fairly represented the entire state
network. Only one lane per selected route was chosen. For GPR purposes, the eight routes were converted
into the twelve sections — totaling 305 lane- miles — listed in Table 1. Figure 1 shows the pilot sections
with respect to the whole state network; Figure 2 shows the exact testing locations on a partial map of the
state overlaid with a GPS generated map ( using GPS coordinates obtained during the GPR testing).
The segmentation process consisted of dividing the pilot network into homogeneous segments
based on administrative boundaries, traffic load, climate, pavement structure, and pavement condition.
Table 1. Initial Sections for Segmentation Pilot Project
ID Route County Description Dir. Lane Length
( mi)
CAL009 SR- 99 Sacramento US- 50 to San
Joaquin Co line SB Out 25
CAL011 I- 5 Sacramento US- 50 to San
Joaquin Co line SB Out 24
CAL013 SR- 99 Sacramento/ Sutter From I- 80 to SR-
70 split NB Out 16
CAL015 SR- 113 Yolo From Davis to
Woodland NB Out 11
CAL017 I- 5 Sacramento Yolo/ Colusa line
SR- 113 SB In 21
CAL031 I- 80 Solano Solano County WB Out 45
CAL033 I- 5 Sacramento/ Yolo SR- 113 to SR- 99
split NB Out 13
CAL035 SR- 16 Colusa/ Yolo Woodland to SR-
20 WB Out 48
CAL041 I- 80 Solano Solano County WB In 45
CAL047 US- 50 Sacramento Sunrise Blvd. to
El Dor. Co line EB Out 11
CAL049 SR- 45 Yolo Yolo County SB Out 13
CAL050 I- 505 Solano/ Yolo I- 5 to I- 80 SB Out 33
Stage 5 Distribution
UCPRC- RR- 2005- 11
15
Figure 1. Location of roadways in relation to the entire network.
Stage 5 Distribution
UCPRC- RR- 2005- 11
16
Figure 2. Location of roadways in North Central, California.
The segmentation procedure consisted of five consecutive classification passes using five
different criteria to progressively break down the entire length of the roadways into segments that share
common attributes. Section 4 presents the details of the data utilized in the segmentation of the pilot
project and the effort involved completing the tasks. ( Note: The segmentation process was modified with
respect to the minutes of the meeting on August 30, 2004 [ see Appendix A]). Review of the data collected
resulted in a change in the order of the segmentation passes and in the decision not to use condition
survey data in the process. Condition survey data was not used because, in its current form, it was found
to be inapplicable. Once better pavement condition data is available ( i. e., data that can be tracked for
cracking over time) it should be used as part of the segmentation.
Stage 5 Distribution
UCPRC- RR- 2005- 11
17
3.1 Administrative Boundaries
The first segmentation pass consists of dividing the roadways into units based on district and county
boundaries. This step is based on past Caltrans practice of programming rehabilitation at the district and
county levels, which has resulted in different structures on each side of boundaries.
In the pilot project, this pass meant dividing I- 505 at the line between District 4 and District 3,
and dividing SR- 99 and I- 5 at the Sacramento/ Sutter and Sacramento/ Yolo county lines, respectively.
This step increased the number of pavement segments to fifteen from twelve.
3.2 Traffic
The researchers divided segments within counties if there was a significant change in traffic loading
between them, hence major intersections served as natural boundaries between sections. Intersections are
permanent physical reference points that also help in locating the sections in the field and can be used for
assigning names to the sections they separate. Traffic data is also required for assignment of priorities
during the selection of rehabilitation projects. The current Caltrans highway traffic database was used.
Dividing the network according to changes in traffic increased the number of sections from 15 to
173. The process included intersections that do not currently affect traffic in the route but that could
eventually grow and become significant. The length of the new segments ranged from 0.10 miles to 7.12
miles, with approximately 50 percent of them being less than 1.25 mile.
3.3 Pavement Structure
The next step was to divide sections with similar traffic into units that had comparable pavement
structure. This includes the surface types and the number and thickness of the layers that constituted the
pavement. Ideally the construction history would have been used to identify the materials and the age of
the pavement sections, but this information was not available for most sections. Sources checked included
as- built records, deflection study reports, and major maintenance archive files.
The thicknesses obtained through the GPR testing on all the roadways, combined with some as-built
drawings and existing knowledge of the pavements, permitted the research team to differentiate the
sections at the points of change in their structure. The method used at this stage for identification of the
point of change was visual and without a statistical analysis because in most cases the GPR data showed a
clear distinction between sections that needed to be separated. Statistical algorithms for automatic
Stage 5 Distribution
UCPRC- RR- 2005- 11
18
detection of changes based on GPR pavement structure data will be tested later when checking
performance of segments. 3
Figure 3 shows an example of GPR thickness and material for the section on I- 5 SB ( southbound)
in Sacramento County. At postmile 21.80 there was a change in AC thickness ( thickness on one side is
approximately 5.7 inches and on the other it is 8.4 inches).
The figure also shows the IRI and cracking data from the 2003– 2004 PCS. The figure shows
alligator cracking data from the PMS database, which illustrates a problem for the Pavement Condition
Survey caused by the lack of structural data in the PMS. Alligator cracking was surveyed because the
surface of the pavement is asphalt; however, alligator cracking can never occur in this pavement because
it is a composite pavement consisting of an asphalt overlay of PCC. The pavement condition surveyor has
no way of knowing that this is a composite pavement from the information available in the PMS. There is
no option in the PCS for evaluating a pavement as a Composite pavement, only Rigid or Flexible.
Composite pavements make up a significant portion of the Caltrans network, and made up 20 percent of
the lane- miles in this pilot project. Reflection cracking, the most common distress occurring on composite
pavements, is not included in the PCS.
After the segmentation by pavement structure was done, the total number of segments increased
from 173 to 236. The length of the new segments remained between 0.10 and 7.12 miles, but the average
length decreased from 1.68 miles to 1.27 miles. Approximately 50 percent of the resulting segments at
this point were shorter than 1.05 miles.
Segmentation based on pavement structure is difficult on pavement sections where layer
thicknesses vary wildly over short distances. Figure 8 and Figure 15 in Appendix C are good examples of
this. This lack of a uniform structure is usually found ( although not necessarily) on older AC pavements.
If no structural pattern is apparent ( such as in Figure 15), then segmenting based on pavement structure is
not recommended and segments should be based entirely on other factors, such as traffic breaks and
climate region. However, some pavements ( such as that in Figure 8) include short structural sections that
may span as little as a half mile. In this case, the pavement section should be broken down into as many
structural sections as the decision- maker can discern from the data and plots. In the case of Figure 8, as
many as ten structural sections can be identified and used as structural breaks in the segmentation process.
The exact number and location of the breaks is left to the discretion of the decision- maker.
3 Such algorithms will be investigated as part of project 3.3 on pavement management systems.
Stage 5 Distribution
UCPRC- RR- 2005- 11
19
0
5
10
15
20
25
30
35
40
28
27
26
25
24
21
20
19
18
17
16
15
14
13
12
10
Post Mile
Depth ( in)
0
50
100
150
200
250
300
350
400
IRI ( 0- 400 in/ mi) and Cracking ( 0- 100%)
Base
PCC
AC
IRI
AC - Alligator B
Cracking
Figure 3. Example of GPR data taken on a GPR section of I- 5 in Sacramento County.
3.4 Climate Region
Differing climate regions ( per the Caltrans Climate Region Map, June 2005) were used as a segmentation
pass. Most of the sections were contained within the Inland Valley ( IV) climate region. The exception
was the westernmost ten miles of I- 80 in Solano County, which is in the Central Coast ( CC) climate
region. This pass resulted in one additional segment, increasing the total to 237.
3.5 Condition Survey
The last pass of the segmentation process was to divide the section into homogeneous units from the
standpoint of pavement condition. However, analysis of the condition survey data indicated that to
rationally partition segments, consistent condition survey data over several years would be necessary, and
rehabilitation and maintenance records were needed to explain changes in observed distresses. These
could not be obtained within the schedule for this project. In the end, distress and IRI data may not be
needed to further divide the segments if the performance and histories show that the segmentation- based
Stage 5 Distribution
UCPRC- RR- 2005- 11
20
administrative boundaries, traffic, pavement cross section, and climate region result in reasonably
homogeneous sections with relatively uniform performance within them. Charts with GPR structure
results and data from the 2003 Pavement Condition Survey are presented in Appendix C.
The decision not to segment based on condition survey data is further supported by the
temporariness of certain maintenance procedures that may conceal existing damage. For example, a slurry
seal on a portion of a segment that has uniform structure, traffic, and climate region, and has alligator
cracking across its entire length, may show zero alligator cracking in the PCS data for a portion of it
because of the seal. However, the cracking remains and will come through the slurry seal after several
years. Segmentation based on PCS data such as this may add inaccuracies to the process as time wears
through various temporary maintenance solutions.
The segmentation is presented in Appendix D in the form of a table containing the postmiles and
the physical references for the resulting segments.
Stage 5 Distribution
UCPRC- RR- 2005- 11
21
4.0 DATA REQUIREMENTS AND RESOURCES INVOLVED
A variety of data were collected and analyzed for the segmentation process. Sources included private
contractors, Caltrans documents, field work data, and project records.
4.1 Pilot Study
Resources employed in the segmentation of the pilot network are as follows.
4.1.1 Traffic Database
A record of the most recent traffic counts can be found on the Caltrans website. 4 Included in the database
is the average annual daily traffic ( AADT) at certain intersections, political boundaries, and other unique
landmarks, along with the corresponding postmiles. The points defined in the traffic log created definitive
segments: in the urban areas, these segments tended to be between 0.1 mile and 4.0 miles long; in rural
areas, the segments could extend over 30 miles. For the GPR sections covered by the PPRC in this study,
the traffic sections typically remained small and only a few extended beyond 5.0 miles long. None of
them was over 10 miles long.
This data was used as the second pass for the segmentation, as explained in Section 3.0. A
Microsoft Excel version of the database is available on the web so no conversions are needed in order to
manipulate the data for this project. Once downloaded, locating the desired sections is straightforward and
takes very little time. For the twelve GPR sites considered in this pilot study, the process took about three
person- hours.
4.1.2 GPR Data
4.1.2.1 GPR Data Collection and Equipment
GPR data was collected at a density of one scan per linear foot of travel. Although this may seem
excessive for network- level work, this data rate is desirable for two reasons: ( 1) according to the
contractor, pavement type ( JPC, CRCP, AC/ PCC, etc.) is more easily identified with denser data; and ( 2)
4http:// www. dot. ca. gov/ hq/ traffops/ saferesr/ tradata
Stage 5 Distribution
UCPRC- RR- 2005- 11
22
the data will be available for future project work where the denser scan spacing might be more desirable.
The data from this project have already been used for project- level analysis of SR20, providing
thicknesses for backcalculation of foamed asphalt stiffnesses.
The GPS system operated concurrently with the GPR data collection. GPS coordinates were
recorded once per second with the current GPR scan number in a separate position log file.
Data was collected at speeds of up to 60 mph. Two- hundred- and- fifty- six samples were taken
during 20 nanosecond scans using 16- bit data resolution. The 20- nanosecond range provided the potential
for layer- depth information capability down to 36 inches. This depth generally exceeds the penetration
capability of the GPR equipment.
The GPR equipment used on this project included a GSSI SIR- 20 radar control and data
acquisition unit, a Model 4108 1- GHz horn antenna, mounting equipment, and an electronic distance-measuring
instrument ( DMI) attached to the vehicle wheel ( see Figure 4). The DMI had a resolution of
500 pulses per foot. The GPR equipment was approved and licensed by the FCC. Also included was
Trimble Model AG114 GPS, or an equivalent system, for recording GPS coordinates. This GPS system
provided submeter accuracy when used in a differential mode in conjunction with the Omnistar service.
According to the manufacturer’s specifications, the GPS data obtained with this service is in NAD83-
compatible format.
Figure 4. GPR equipment used on this project.
Stage 5 Distribution
UCPRC- RR- 2005- 11
23
GPR data was analyzed by the contractor at 0.1- mile intervals — based on the vehicle DMI —
beginning at the county line or other marked reference point in each test section. GPS coordinates were
reported for each GPR data point analyzed. When the 0.1- mile interval point fell on a bridge deck ( this
was easily identified in the GPR data), a neighboring location on either side of the deck was selected.
The results of the GPR analysis were provided in Microsoft Excel data files, one for each section.
The data reported at each location represents 200 feet of pavement, ± l00 feet on either side of the reported
location. Exceptions to the 200- foot length occurred when there was a bridge deck or other anomaly in the
pavement structure within the ± l00- foot interval. Where this occurred, the interval was shortened to
include only the pavement representative of the local area.
Within each file, there are five columns for each analyzed layer, and up to four layers analyzed.
The five columns for each layer are described as follows:
• Layer type ( e. g., AC, PCC, base),
• Layer thickness ( average of 200 feet, in inches),
• Layer dielectric constant ( average of 200 feet, no units),
• Layer thickness standard deviation over 200- foot length ( inches), and
• Layer confidence.
The contractor assigned a number from one to four to each analyzed data point to reflect his
degree of confidence. An explanation of the numbering code follows:
• Layer boundary and type is clear.
• Layer boundary is unclear – calculated thickness may be affected.
• Layer type is unclear – best assessment, but it is possible that identified type is incorrect. For
example, assigning a “ 3” to layer 2 when it is suspected to be AC but might be base.
• A combination of 2 and 3.
4.1.2.2 GPR Cost
Infransense, Inc., the contractor providing the GPR information, charged $ 30,923 for 305 lane- miles of
data, which included planning, mobilization, and the collection and analysis of the raw data. The per- mile
cost of data collection was $ 16.48, while the per- mile cost of analysis was $ 51.15. The cost of planning
and setup was $ 1,720; the cost of mobilization and demobilization was $ 8,575.
Stage 5 Distribution
UCPRC- RR- 2005- 11
24
4.1.2.3 Plotting Results for Segmentation
The GPR data received from the GPR contractor were easily plotted using Microsoft Excel. Creation of
charts showing cross sections of the twelve sections took eight person- hours.
4.1.2.4 Identification and Segmentation of Structure Changes
Depth trends in the plotted GPR data are visually evident in most cases, making identification of major
structure changes possible by inspection. Material- type recognition by dielectric constant is not an error-proof
process and therefore uncalibrated GPR results do not always properly identify material type.
Structure changes based only on material type are difficult to distinguish.
Once structure changes were identified on the charts, the exact corresponding postmile was
located in the GPR data and recorded in the segmentation database. Visually identifying the structure
changes, locating the precise point of the structure change in the GPR database, and segmenting based on
the structure changes took approximately 30 person- hours.
4.1.3 Coring
Coring was completed for thirteen sites: Twelve in District 3 and one in District 4. The sites were cored
on nine days between July 7 and September 16, 2005. Closures were performed by Caltrans district
Maintenance personnel. The internal cost of these closures to Caltrans is not known. From UCPRC
experience, a private contractor would charge approximately $ 2,000–$ 3,000 per day for the closures.
A crew of six people was necessary for this work. The crew was responsible for running the
coring machine, using the Dynamic Cone Penetrometer ( DCP), recording data, and backfilling the core-holes.
Including travel, setup, and breakdown, this took approximately 60 person- hours. The DCP
provided data regarding layer thicknesses below the depth of cores. Details on the coring are presented in
Section 5.1.
4.1.4 As- Builts
An attempt was made to collect as- built information for the GPR sections. Caltrans provided UCPRC
with a limited number of as- builts, and District offices were visited to find additional ones. However,
records for many of the as- builts were unavailable because documentation was lost or because records
were not kept when work was performed. Depending on the organization of records and the availability
of the necessary documents, collecting this information could take up to 16 person- hours. An attempt was
Stage 5 Distribution
UCPRC- RR- 2005- 11
25
made at collection of as- built information for the GPR sections. Caltrans provided UCPRC with a limited
number of as- builts and District offices were visited to find additional ones. However, many segments did
not have as- built records because of lost documents and work that has been performed but not recorded.
Depending on the organization of records and the availability of the necessary documents, this task could
take up to 16 person- hours.
4.1.5 Climate
Most GPR segments for this project were located in the Inland Valley climate region, with two sections
split between the Inland Valley and Central Coast regions . Segmentation based on climate boundaries is
simplified by the Caltrans Climate Map, making time- demand for this step negligible ( zero person hours).
Caltrans Maintenance has developed a map that defines the exact postmiles that define boundaries
between climatic regions on each route for nearly the entire state.
4.1.6 Condition Survey and IRI
Condition surveys, which include the IRI, are available from the Caltrans Pavement Management System
( PMS) database. Though the GPR sections have not been segmented based on the condition survey data,
the pavement condition has been entered into the GPR database and it has been used to generate charts for
comparison showing pavement condition alongside the GPR results. The plotted data includes the IRI,
alligator B cracking ( AC), and third stage cracking ( PCC).
The raw data from the PMS database needed to be converted into a manageable format, which
took about 10 person- hours. This task was completed for the whole state highway network. Loading the
PMS data into the GPR database and outputting the resulting plots took another 20 person- hours. In sum,
the condition survey data took 30 person- hours.
4.1.7 Database
Development and population of the database for the pilot segmentation project took place at the same
time that the data was being retrieved from all the sources. A nominal one person- hour is being accounted
for database handling.
Information collected for segmentation is currently stored in Excel with location identifiers tied to
the distance measured from nearest physical reference, such as structures or paddles, for which the exact
GPS coordinates have been obtained. Soon, the data will be loaded into a relational database ( Access) and
delivered to Caltrans.
Stage 5 Distribution
UCPRC- RR- 2005- 11
26
4.1.8 Summed Effort
The estimated time spent on the segmentation process for the twelve GPR sites sums to 148 person- hours.
Other costs include the contract costs for the GPR ($ 30,923 for 305 lane- miles), lane closures ( estimated
to be between $ 14,000 and $ 21,000 if done by a private contractor), materials for coring ( bits, backfill,
etc.) and various travel costs.
4.2 Extrapolated Cost and Effort to Whole Network
The Caltrans 2003 State of the Pavement Report states that there are over 49,000 lane- miles of pavement
in the California highway network. If segmentation of 305 lane- miles requires 148 person- hours, then the
whole network would take nearly 24,000 person- hours to complete. This amount is approximately 12.30
PY ( assuming 1,940 hours per year). At an estimated rate of $ 94.43 per mile, the GPR data collection,
analysis, and calibration, the cost for contracting the GPR testing over the entire network would be
approximately $ 4.63 million, not including mobilization. If mobilization is assumed to be 12 percent of
the cost of testing, then the total estimated cost of GPR testing and analysis can be considered $ 5.2
million.
The cost of lane closures needs to be added to that amount. At an assumed rate of $ 3,000 per day,
and considering about 600 days of closures to complete all the required coring, the cost would be $ 1.8
million. This brings the total direct cost to an estimated $ 7.0 million. The direct cost and personnel
needed are shown in Table 2.
Stage 5 Distribution
UCPRC- RR- 2005- 11
27
Table 2. Personnel Needed and Direct Cost for Segmentation of
Pilot Study and Estimated for Entire Caltrans Network
Item
Pilot Study
305 Lane- Miles
( actual)
Caltrans Network
49,000 Lane- Miles
( extrapolated)
Personnel 0.076 PY 12.30 PY
Direct cost $ 48,500 $ 7,000,000
The actual direct and personnel cost, both for field ( coring and GPR) and office work, will likely
be less than the figures stated above. Time spent retrieving data and segmenting based on that data will
drop significantly as personnel become increasingly proficient at the process. Also, the cost per lane- mile
of GPR measurement and analysis will decrease if a bid system to determine the lowest price can be
implemented.
It must be noted that the pavement structure database for the PMS that could be created by a GPR
project would lose its value over time unless it is routinely updated with accurate information regarding
the changes to pavement structures caused by future rehabilitation, maintenance, and reconstruction.
Stage 5 Distribution
UCPRC- RR- 2005- 11
28
5.0 DISCUSSION OF RESULTS FROM THE PILOT PROJECT AND REMAINING
WORK
5.1 Utilization of Coring Data
5.1.1 Core Sites
The coring for the GPR was completed on September 16, 2005. A total of 43 cores were extracted from
13 sites in Districts 3 and 4. The original plan called for 16 sites with a total of 65 cores. The difference
between these numbers is due to scheduling problems for the Caltrans Maintenance force and time
constraints that arose in the field. A summary of the coring locations is shown in Table 3.
Table 3. Final List of GPR Coring Locations
Closure
No.
Section
ID County Route Direction Start End
Coring
Date
No. of
Cores
Data Given to
Infrasense?
1a CAL050 Solano 505 SB 8.10 8.40 9/ 16/ 2005 4
X
1b CAL050 Solano 505 SB 5.00 5.40 Cancelled – time constraints in the field
2 CAL013 Sacramento 5 NB 27.70 28.22 Cancelled – could not get closure
3 CAL013 Sutter 99 NB 5.68 6.18 Cancelled – could not get closure
5 CAL015 Yolo 113 NB 2.89 3.20 8/ 22/ 2005 6 X
5a CAL015 Yolo 113 NB 8.40 8.70 8/ 25/ 2005 3
9 CAL047 Sacramento 50 EB 17.20 17.50 7/ 7/ 2005 4 X
10 CAL047 Sacramento 50 EB 20.01 20.31 7/ 11/ 2005 4
11 CAL049 Yolo 45 SB 10.80 11.10 8/ 24/ 2005 4 X
12 CAL049 Yolo 45 SB 7.82 8.12 8/ 25/ 2005 4
12a CAL049 Yolo 45 SB 9.00 9.30 8/ 24/ 2005 4
14 CAL009 Sacramento 99 SB 8.86 8.96 7/ 13/ 2005 2 X
15 CAL009 Sacramento 99 SB 6.26 6.36 7/ 21/ 2005 2
16 CAL035 Colusa 16 WB 3.04 3.14 7/ 25/ 2005 2 X
17 CAL035 Colusa 16 WB 1.84 1.94 7/ 25/ 2005 2 X
18 CAL035 Colusa 16 WB 0.64 0.74 7/ 25/ 2005 2
Most sites were chosen by Infrasense, Inc., and confirmed by the UCPRC. Sites were chosen
based on abrupt changes in the apparent pavement structure and uncertainties in the GPR data. Two sites
( Closures 5a and 12a) were chosen strictly by the UCPRC for control purposes.
Stage 5 Distribution
UCPRC- RR- 2005- 11
29
5.1.2 Determining Exact Core Locations
Determining exact core locations was critical to the success of the project. Cores taken in the field needed
to be matched up with the GPR results for exactly the same location so that an accurate comparison of the
two could be made.
Infrasense provided location data relative to a local physical reference at each site. This data
included a unique local reference point, distances from the reference point, and GPS coordinates. Physical
references were Caltrans postmile paddles, bridge decks, and obvious changes in surface material ( i. e.,
from PCC to AC overlay). Examples of this data appear in Table 4.
In the field, core locations were marked using a digital survey wheel taken from a given local
reference point. GPS measurements were taken at each core and used as a distance check once the data
was entered into the database. The database shows discrepancies between Infrasense’s GPS coordinates
and the UCPRC’s field GPS coordinates ranging from 4.7 feet to 158.6 feet. These values can be found in
Table 9 of Appendix F. Possible reasons for the discrepancies include:
• Inherent inaccuracies in the UCPRC GPS receiver, which did not have differential capability
( this device provided a typical error of ± 3m, with extreme error up to ± 30m at some
locations.);
• Inexact physical reference locations measured by Infrasense;
• On- the- fly measurements may be prone to inaccuracies;
• Mileposts were not necessarily in the same location in each direction;
• Equipment malfunction;
• Survey wheel was decommissioned by the UCPRC after CAL015 sites because of
malfunctions that might have affected previous sites; or
• Other human errors.
These discrepancies stem from the fact that the GPR and coring were done at different times and therefore
required extensive use of multiple measurement metrics ( GPS coordinates, postmiles, and distances from
physical references) in order to pinpoint the exact location of the GPR measurement. Alternatively, if the
coring crew was present during the GPR process, the GPR crew could guide the coring crew to the exact
location to ensure that the core would align with the GPR data. This alleviated the need to describe the
locations using a combination of measurement metrics.
Stage 5 Distribution
UCPRC- RR- 2005- 11
30
Table 4. Selected Coring Locations and Physical Reference Points
The “ Approximate Postmiles” were used as a check to ensure coring was done in the right
vicinity. They were also used to coordinate the closures with Caltrans maintenance yards. Infrasense
calculated the postmiles as a distance from certain physical features, such as paddles or county lines.
Because of the complexities in the Caltrans postmile system ( equations, inaccuracies, etc.), coring
locations were recorded independently of the approximate postmiles.
5.1.3 Core Results
Core layer thicknesses, layer material types, and DCP results were disclosed to Infrasense for the seven
sites noted in Table 3. This data was used by Infrasense to verify and calibrate both their thickness and
material- type results. After a review of the core data disclosed to them, Infrasense determined that a
Closure
No.
Section
ID County Route Dir.
Approx
PM
Local
Physical
Reference
Dist
from
Ref.
( ft)
Latitude
( deg)
Longitude
( deg)
9 CAL047 SAC 50 EB 17.20 E Joint
Bridge
Deck
377.00 38° 38.402 121° 11.694
9 CAL047 SAC 50 EB 17.30 E Joint
Bridge
Deck
905.00 38° 38.416 121° 11.584
9 CAL047 SAC 50 EB 17.40 E Joint
Bridge
Deck
1433.00 38° 38.429 121° 11.475
9 CAL047 SAC 50 EB 17.50 E Joint
Bridge
Deck
1961.00 38° 38.443 121° 11.366
10 CAL047 SAC 50 EB 20.01 SAC RP
20
57.00 38° 38.514 121° 08.507
10 CAL047 SAC 50 EB 20.11 SAC RP
20
585.00 38° 38.519 121° 08.397
10 CAL047 SAC 50 EB 20.21 SAC RP
20
1113.00 38° 38.524 121° 08.287
10 CAL047 SAC 50 EB 20.31 SAC RP
20
1641.00 38° 38.529 121° 08.177
Stage 5 Distribution
UCPRC- RR- 2005- 11
31
systematic calibration was not necessary. However, the following changes were made to CAL15- 5 by
Infrasense:
• Layer 2 thickness ( PCC): Reduced by 12 percent;
• Revised GPR data locations to match UCPRC core locations — shifted ~ 0.1 miles to account
for GPS discrepancies in a few cases.
Minor changes were recommended by Infrasense for two sites that had not been disclosed to
them:
• CAL15- 5a: Reduce thickness of PCC layer by 12 percent
• CAL035- 18: Change layer 3 material to “ base”
After the changes were made, the GPR results were compared to the remaining cores. DCP
results were used to estimate underlying layer materials and very approximate thicknesses. The
comparisons can be found in Appendices E ( plots) and F ( tables).
Review of the results shows that the GPR technology is effective for determining layer
thicknesses for all layers. The accuracy level decreases with depth, with layers 1 and 2 being more
accurate than layers 3 and 4. The average thickness difference ( absolute percentage of total layer) and
accompanying standard deviations are presented in Table 5. A comparison of the GPR versus core
( UCPRC) thicknesses is plotted in Figure 5.
Some of the more extreme values in Figure 5 may be a result of the discrepancy between the GPR
readings location and core locations discussed in Section 5.1.2. At some sites, layer thicknesses are highly
variable over small areas, so even a small difference in between the GPR reading and the core location
can result in a large difference in layer thicknesses. These extreme values affect the averages and standard
deviations expressed in Table 5.
Layer types as indicated by the GPR reading matched up well with the UCPRC cores. Deeper AC
was sometimes recorded as “ Base” or “ BB” ( bituminous base). These layers sometimes exhibited aging
effects ( such as the breakdown of materials) that may have caused the misnaming. The GPR was unable
to differentiate between base types, including cemented bases ( LCB or CTB) or asphalt- treated permeable
bases ( ATPB). Open- graded AC ( OGAC) layers were not distinguished from other AC types and were
grouped together with the underlying AC layers. For example, if a layer consisted of 25 mm OGAC and
100 mm DGAC, the GPR output would be 125mm AC.
Stage 5 Distribution
UCPRC- RR- 2005- 11
32
Table 5. Average Thickness Differences ( Absolute) and Standard Deviations
Layer 1 Layer 2 Layer 3 Layer 4
No. of Cores 31 22 15 3
Average Difference 12.62% 10.17% 27.88% 20.89%
Standard Deviation 11.2% 15.0% 23.4% 11.3%
0
5
10
15
20
0 5 10 15 20
GPR Thickness ( in)
Core Thickness ( in)
Layer 1
1: 1 Line
0
5
10
15
20
0 5 10 15 20
GPR Thickness ( in)
Core Thickness ( in)
Layer 2
1: 1 Line
0
5
10
15
20
0 5 10 15 20
GPR Thickness ( in)
Core Thickness ( in)
Layer 3
1: 1 Line
0
5
10
15
20
0 5 10 15 20
GPR Thickness ( in)
Core Thickness ( in)
Layer 4
1: 1 Line
Figure 5. GPR versus core ( UCPRC) thicknesses.
Stage 5 Distribution
UCPRC- RR- 2005- 11
33
5.2 Comparison of Current Caltrans Maintenance Dynamic Segmentation Versus Fixed
Length Segmentation
The length of segments that Caltrans uses for evaluation of pavement condition was obtained from the
program, “ Pavement Condition Reporting System.” 5 Statistics were obtained for data in the years 2000
and 2004 for about 305 miles of roadway for the pilot segmentation study. Histograms with the number of
segments in 0.1- mile intervals are presented in Figure 6 for each of these two years. The charts show that
the number of segments identified by the Caltrans dynamic segmentation for the pilot network increased
from 225 to 431 between 2000 and 2004, and that average length dropped from 1.17 to 0.66 miles. One-mile
segments seem to be the most common survey unit.
The same chart was prepared for the fixed segmentation performed as part of this study and is
presented in Figure 7. It can be noted that the fixed- length segmentation produced segments whose
lengths are spread over a wider range. Figure 7 shows segments only up to 5 miles long, but there were
four additional sections between 5.0 and 7.3 miles long.
0
10
20
30
40
50
60
70
80
90
0.1 0.4 0.7 1.0 1.3 1.6 1.9 2.2 2.5
Section length ( miles)
Number of sections
Year 2004
Nr. of sections: 431
Average length: 0.66 mi
0
10
20
30
40
50
60
70
80
90
0.1 0.4 0.7 1.0 1.3 1.6 1.9 2.2 2.5
Section length ( miles)
Number of sections
Year 2000
Nr. of sections: 225
Average length: 1.17 mi
Figure 6. Histograms of sections versus length with Caltrans segmentation.
5 Version 3.0.0 March 17, 2005.
Stage 5 Distribution
UCPRC- RR- 2005- 11
34
0
10
20
30
40
50
60
70
80
90
0.1 0.4 0.7 1.0 1.3 1.6 1.9 2.2 2.5 2.8 3.1 3.4 3.7 4.0 4.3 4.6 4.9
Section length ( miles)
Number of sections
PPRC Segmentation
Nr. of sections: 236
Average length: 1.27 mi
Figure 7. Histogram of sections versus length with fixed- length segmentation.
A comparison between the 2004 Caltrans segmentation versus the fixed- length segmentation
indicates that there would be roughly 200 fewer segments to survey ( 237 instead of 431) in the pilot
network, and that the average segment length would be 1.27 miles rather than 0.66 miles. The validity of
these conclusions is limited until segmentation by surface condition is performed; however this indicates
that fixed segments could result in fewer segments to survey, reducing the cost of the Caltrans Pavement
Condition Survey.
5.3 Collection of Pavement Condition Indicators for Performance Modeling in PMS
As mentioned in Section 2.1, collection of pavement condition data depends on whether the information
is going to be used for PMS or for project- level maintenance. In order to collect the necessary data to
develop or calibrate empirical models for pavement management and to calibrate mechanistic- empirical
pavement design models, some minor changes to the PCS procedure have to be implemented. It seems
that there are two possible approaches.
1. The first option is to continue with the current scheme of condition surveys, but to use the
fixed- length segments as breakpoints ( PCS hits the same ends as the PMS segments) so the
results can be tracked year after year. Since more than one PCS segment is likely to be found
within a PMS segment, a weighted average of the condition in the segments, based on length,
can be obtained to represent the condition of the entire PMS segment.
2. The second alternative is to conduct condition surveys for PMS purposes, independent of the
Pavement Condition Survey for maintenance. Since the level of detail in a PMS condition
survey is lower, it is a common practice to report smaller segments with an equivalent
Stage 5 Distribution
UCPRC- RR- 2005- 11
35
condition simply as “ same as previous” because there is no need for extensive examination
and to reduce the cost of the field work.
This report contains a list of recommendations regarding the type of information necessary for
PMS purposes. The recommended items are shown in the table in Appendix G, in which data included in
Caltrans current Pavement Condition Survey Method is compared with information required for PMS
purposes. In that table, recommended items to be changed are shaded. There are several types of
distresses whose collection is not recommended for PMS because:
• They represent features not directly on the roadway, such as shoulder cracking,
• They are very difficult to observe consistently ( e. g., pumping or segregation), or
• They represent the final state of another distress ( e. g., potholes caused by cracking). Because
of the lengthy time needed to quantify these distresses, the analysis would no longer be useful
for programming any work.
Stage 5 Distribution
UCPRC- RR- 2005- 11
36
6.0 CONCLUSIONS, FUTURE WORK, AND RECOMMENDATIONS
6.1 Conclusions
The conclusions that can be drawn from the pilot segmentation study and the GPR testing are as follows:
1. Fixed- length segmentation is a process that involves analysis of roadway information from
various sources. Once the segmentation process is completed, the resulting segments will
provide a theoretically sound frame for future pavement condition data collection that would
allow for the development of performance models.
2. Segmentation of the pilot network showed that the best approach to break down segments of
roadway is by means of the following steps: ( a) administrative boundaries, ( b) traffic load, ( c)
pavement structure, ( d) climate region, and ( e) pavement condition ( if needed).
3. The direct cost to implement PMS segmentation and to collect GPR data for inventory of
pavement structure for the entire network is estimated — based on extrapolation from this
pilot project — at approximately $ 7 million of contracted field work, while the approximate
need for personnel for segmentation analysis is an additional 12.3 PY.
4. Ground- penetrating radar ( GPR) pavement- related technology has been evaluated by Caltrans
and by other DOTs, and it has been found to be reliable, both for project- level assessment,
and for network- level inventory.
5. GPR testing supplemented with limited coring and DCP data to populate the inventory
database with pavement structures throughout the California highway network appears
feasible. The information provided by the GPR contractor was easy to use and reliable, based
on the coring by the UCPRC. The available data indicates that GPR provides reasonable
cross- sections.
6.2 Recommendations
The following are the recommendations based on this project.
1. A condition survey for PMS purposes needs to be implemented. It will consist of either minor
changes to the current Pavement Condition Surveys or the implementation of a parallel data
collection unit, focused only on the variables needed for adequate performance modeling.
2. If funding the segmentation of the entire network is an issue, the process can be staged,
adding more roads each year to spread the costs over several years.
Stage 5 Distribution
UCPRC- RR- 2005- 11
37
3. After inventory information is generated for parts of the network ( using GPR), it is important
to document as- built information in the structural database as future maintenance,
rehabilitation, and reconstruction work occurs, in order to keep the database accurate.
4. The data collection practice and segmentation processes should be followed and applied to
the entire Caltrans highway network, implementing a relational database according to
Caltrans IT ( i. e., data dictionaries, data collection, populating and integrating databases).
Stage 5 Distribution
UCPRC- RR- 2005- 11
38
7.0 REFERENCES
1. 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, University of California, Davis and
Berkeley.
2. Maser, K. R. 2002. “ Use of Ground- Penetrating Radar Data for Rehabilitation of Composite
Pavements on High- Volume Roads.” Transportation Research Board 1808: 122– 126.
3. Mishalani, R. and Koutsopoulos, H. 1995. “ Uniform Infrastructure Fields: Definition and
Identification.” Journal of Infrastructure Systems 1, no. 1.
4. Thomas, F. 2003. “ Statistical Approach to Road Segmentation.” Journal of Transportation
Engineering 129, no. 3.
5. ASTM International. 2005. “ ASTM D4748- 98: Standard Test Method for Determining the Thickness
of Bound Pavement Layers Using Short- Pulse Radar.”
6. Noureldin, A. S., Zhu, K., Li, S, and Harris, D. 2003. “ Network Pavement Evaluation With Falling-
Weight Deflectometer And Ground- Penetrating Radar,” Transportation Research Record 1860:
90- 99.
7. NCHRP. 1998. “ Synthesis 225: Ground Penetrating Radar for Evaluating Subsurface Conditions for
Transportation Facilities.”
8. Infrasense, Inc. 2003. “ Non- Destructive Measurement of Pavement Layer Thickness.” Caltrans
Report No. 65A0074.
9. Al- Qadi, I., Lahouar, S., Jiang, K., MeGhee, K., and Mokarem, D. January 2005. “ Validation of
Ground Penetration Radar Accuracy for Estimating Pavement Layer Thicknesses.” Presented at the
84th annual meeting of the Transportation Research Board, Washington D. C.
10. Gucunski, N. ( 2004) “ Demonstration of Ground Penet4413rating Radar ( GPR) ( NJDOT Statewide
GPR Pilot Project) Rutgers University. Research Report No: GPR- RU4474
11. Cardimona, S., Brent Willeford, Doyle Webb, Shane Hickman, John Wenzlick, Neil Anderson ( 2003)
“ Automated Pavement Analysis in Missouri Using Ground Penetrating Radar” University of
Missouri – Rolla, Department of Geology and Geophysics
12. Willet, D ( 2002). “ Ground Penetrating Radar Pavement Layer Thickness Evaluation” Kentucky
Transportation Center, University of Kentucky. Research Report KTC- 02- 29 / FR101- 00- 1F.
13. Corley- Lay, J and Morrison, C. S. 2001. “ Layer Thickness Variability for Flexible Pavements in
North Carolina.” Transportation Research Board 1778: 107– 112.
Stage 5 Distribution
UCPRC- RR- 2005- 11
39
APPENDICES
Stage 5 Distribution
UCPRC- RR- 2005- 11
40
§ 3.2.41
Appendix A. Minutes - 8/ 30/ 04 Meeting On Developing Objectives for theHighway Network
Segmentation & Data Collection In D- 3 Using GPS2
Attendees:
Design: bill_ farnbach@ dot. ca. gov,
Construction: chuck_ suszko@ dot. ca. gov,
Geometronics: adrian_ davis@ dot. ca. gov, Jim Brainard/ D03/ Caltrans/ CAGov
Maintenance: carole_ harris@ dot. ca. gov, pattie_ pool@ dot. ca. gov, susan_ massey@ dot. ca. gov,
Research: alfredo_ b_ rodriguez@ dot. ca. gov, james_ n_ lee@ dot. ca. gov, michael_ m_ samadian@ dot. ca. gov,
t_ joe_ holland@ dot. ca. gov, Michael Essex
PPRC/ Dynatest: jtharvey@ ucdavis. edu, Nick Coetzee
Introduction
This meeting dealt with the development of objectives for the “ Highway Network Segmentation & Data
Collection In D- 3 Using GPS” or what is being called “ The Segmentation Pilot Project.” The objectives
developed during the meeting were broken down into five key areas: A) which highways/ routes and which
lanes to collect data from, B) the types of data to collect, C) the kinds of analysis to be performed, D) the
phases of segmentation and whether all five phases can be achieved, E) deliverables, and F) lane miles
versus cost option selection.
A key issue to be resolved by this project is whether ground- penetrating radar ( GPR) for the continuous
measurement of pavement thickness can be used effectively ( i. e., is the technology sufficiently developed
such that the use of GPR hardware and software generates measurements that are reproducible and
repeatable).
Background
In the last meeting John Harvey presented a plan and costs for testing pre- selected parts of the highway
network in District 3 ( Sacramento & Yolo counties). Since that meeting it was decided to revisit the
rationale behind what and how much of the network would be sufficiently representative to meet the main
objective of understanding how segmentation, data collection, population of data bases, and the subsequent
analyses can be done in future by Caltrans resources and whether addition resources will be required. These
1 “ Development of Integrated Databases to Make Pavement Preservation Decisions” – PPRC Strategic Plan 03/ 04.
2 The segmentation of highway networks and related data collection was not originally envisioned as part of the PPRC 03/ 04
Strategic Plan Section 3.2.4. It evolved as a logical next step that will precede the development and population of the integrated
databases originally intended.
Stage 5 Distribution
UCPRC- RR- 2005- 11
41
issues are addressed Sections A, B, C, & D with a final determination of the optimum amount of lane miles
versus cost is made in Section F.
A. Data Collection Locations
The parts of the D- 3 network ( highways and route) listed below were previously identified as being good
candidates that should include a sufficiently diverse set of roadway structural sections to be representative
of the overall highway systems within California.
Highways / Route
• I- 80 ( Solano Co./ Yolo Co./ Sacramento Co – 35 miles)
• I- 5/ US 99 ( Sacramento Co. -? X miles)
• SR 20 ( Lake Co. to Grass Valley -? X miles)
Lanes, Lane Directions, Measurement
• 1 to 2 lanes per direction
• 4- lane facility ( outside – 1 direction, inside – other direction
• 6- lane facility ( outermost 2 lanes in each direction)
• 2- lane facility ( 1 direction)
• Type of initial measurement using GPR:
• General thickness ( homogeneous sections)
• Changes in structural cross section ( need horizontal sub- meter precision – get information from Surveys)
B. Other Data Collection Needs
Office Data
• Office of Pavement rehabilitation deflection studies
• As- Builts ( HQ data, retrieval [ intranet] of District data)
• Data from Moisture Sensitivity studies
• Data from the Pavement Performance Evaluation Phase I ( Stantec project)
Coring Data
• Use for verification of GPR measurements
• Take in questionable areas ( visually distinct fro the surrounding pavement)
• Use to calibrate GPR units used in the pilot
• Criteria for sampling
• A few random sites
• Areas designated for analysis only
Stage 5 Distribution
UCPRC- RR- 2005- 11
42
Criteria to Define Changes In Pavement Structure
• Where the average thickness changes greater than 50 mm ( between 0.1 mile sections)
• Where the order of layer type changes
• Where independent METS GPR data shows significant changes
C. Analysis
Pick 1000 lane miles from 2850 lane miles ( narrow sections – contact Pat Kelley @ D- 3
Design)
Take office data and map out structures
Collect existing coring data
Analyze GPR data
Identify questionable areas and do coring
Compare verification data with GPR information from analysis
Do the economic analysis
D. Segmentation3
A successful segmentation plan will consist of five passes through the network, each one resulting in a
further segmentation:
1. In the first pass, administrative considerations will prevail. This will lead to dividing the highway
network according to districts and routes. For example, I- 5 would first be segmented according to
the Caltrans districts that it lies along.
2. In the second pass, segments within an administrative unit ( route and district) are further segmented
according to pavement structure ( AC on granular, PCC, AC on PCC, AC on LCB, AC on CTB,
etc), subgrade type, with each segment having a “ uniform” pavement structure with regard to type
and general thicknesses, and underlying subgrade type.
3. In the third pass, uniform pavement structure segments are broken if they cross a climate region
boundary.
4. In the fourth pass, segments are broken if there is a significant change in traffic loading ( which
means that major intersections are natural boundaries between sections).
5. In the fifth pass, segmentation is based on surface measurements. At this level, the objective is to
divide the highway into sections that are homogeneous in their current condition ( general state of
surface distresses and IRI).
For this pilot process the first three passes will be done and the fourth and fifth passes will be conducted
depending on time, budget, and availability of traffic data.
3 Segmentation process details are from “ A Plan for Segmentation of Highway Pavements for Use in Caltrans’ Pavement
Management System,” Samer Madanat, April 29, 2004.
Stage 5 Distribution
UCPRC- RR- 2005- 11
43
Deliverables ( due dates)
Pilot Project Technical Deliverables
1. ( C1) – Develop a list of the 1000/ 300 lane miles ( GPR measurements/ coring & other data
collection) [ 9/ 04]
2. ( C3) – Develop information on preliminary structures/ sections ( include available information from
databases and maps) [ 12/ 04]
3. ( C5) – Final structures/ sections information ( database information & maps) [ 3/ 04]
4. ( C6) – Write Tech Memo ( technical feasibility of segmenting highway network) [ 6/ 05]
5. ( C7) - Write Tech Memo ( Economic Analysis) [ 6/ 05]
6. ( D1) - Write Tech Memo ( Segmentation Pilot Project) [ 8/ 05]
Other Deliverables
1 Marketing plan for upper management ( with technical backup)
Lane Miles Involved Vs. Costs4
Plan
# Lane miles to be
measured with GPR
# Lane miles to collect
additional data on for analysis
Estimated cost
A 2,850 300 $ 76,000
B 1,000 1,000 $ 76,000
C 2,000 2,000 $ 147,000
D 300 300 $ 36,000
E 500 500 $ 50,000
F 1,001 300 $ 40,000
Recommendation: Go with Plan F
Post Meeting Information/ Discussion
The purpose of the Segmentation Pilot is to demonstrate the feasibility of segmenting the highway network
into homogeneous sections that will allow for accurate prediction of pavement performance and the
optimization of the Maintenance budget process. However it is not clear how actual future segmentation
activities will be performed or who will perform them. What is anticipated is the development of a data
warehouse that will incorporate a tremendous amount of data from a wide variety of sources. This will
include the following:
• Research databases including the HVS field and laboratory databases and several others.
• Databases from the Pavement Performance Evaluation project, Phases I & II – Phase II to be started in
late 2004.
• The Pavement Management System ( PMS) including the existing database and the new one to be
developed starting in 2005.
• METS database( s).
4 Cost estimates are based on a consultant’s estimate to do data collection and analysis for varying lengths of roadway
( Infrasense Inc. PPRC Pilot GPR Project Ground Penetrating Radar Survey in Sacramento and Yolo Counties, August 25, 2004.
Stage 5 Distribution
UCPRC- RR- 2005- 11
44
Decision/ Action Needed
The above projects need to be coordinated closely to assure that data collected is compatible in terms of
populating what could become the PMS data warehouse. This raises a number of issues that will need to be
addressed:
1. Who will be the lead to verify that the right kinds of data are being collected ( essential and helpful
variables)?
2. How will the meta data be developed and by whom?
3. How will data quality be assured?
4. Do we need a Department- wide data collection, preservation, and availability policy, i. e., should the
Districts and Headquarters be required by a Directive to actively participate in an enterprise
pavement system in which design, construction, maintenance, research, and traffic data is available
to all potential users of pavement data? This was answered previously ( more- or- less) in the
affirmative but no strategy was developed to address this issue.
Suggestion:
It has been suggested that either Research or the Pavement Standards PMS Team write a white paper for
review by the Acting Director and the Acting Deputy Director for Maintenance and Research.
Stage 5 Distribution
UCPRC- RR- 2005- 11
45
Appendix B. GPR Survey Summary
GPR
File #
Date
Collected Route
Map
Direction
Target
Lane
Type
( speed) Layout Description
Approx.
Survey
Length
( mi) GPR Data Characteristics
Recommended
for Analysis
CAL009 Mar. 7, 2005 SR- 99 SB low US- 50 to San J. Co line 25 Mostly AC/ PCC, some full depth AC, somewhat variable x
CAL010 Mar. 7, 2005 SR- 99 NB faster US- 50 to San J. Co line 25 Mostly AC/ PCC; fairly homogeneous
CAL011 Mar. 7, 2005 I- 5 SB low US- 50 to San J. Co line 24 PCC, very homogeneous, some radio interference at S end. x
CAL012 Mar. 7, 2005 I- 5 NB faster US- 50 to San J. Co line 22 PCC, very homogeneous, some radio interference at S end.
CAL013 Mar. 7, 2005 SR- 99 NB low From I- 80 to SR- 70 split 16 Mostly full depth AC, fairly homogenous x
CAL014 Mar. 7, 2005 SR- 99 SB faster From I- 80 to SR- 70 split 16 Mostly full depth AC, fairly homogenous
CAL015 Mar. 7, 2005 SR- 113 NB low From Davis to Woodland 11
Very homogeneous PCC; North section appears to have Bituminous
Base. Is this possible? x
CAL016 Mar. 7, 2005 I- 5 NB low Yolo/ Colusa line SR- 113 21
Very homogeneous AC/ PCC, with several local full depth AC patches,
especially near YOL RP 11
CAL017 Mar. 7, 2005 I- 5 SB faster Yolo/ Colusa line SR- 113 21 Very homogeneous AC/ PCC x
CAL030 Mar. 8, 2005 SR- 113 SB faster From Davis to Woodland 11
Very homogeneous PCC; North section appears to have Bituminous
Base. Is this possible?
CAL031 Mar. 8, 2005 I- 80 WB low Solano County 45
Long homogeneous sections of full depth AC and PCC, some AC/ PCC,
layer type interpretation clear except in some sections near western end.
CAL032 Mar. 8, 2005 I- 80 EB low Solano County 45
Long homogeneous sections of full depth AC and PCC, some AC/ PCC,
layer type interpretation clear except in some sections near western end.
CAL033 Mar. 8, 2005 I- 5 NB low SR- 113 to SR- 99 split 13 homogeneous, looks like AC/ PCC/ Base. Not sure about the PCC x
CAL034 Mar. 8, 2005 I- 5 SB faster SR- 113 to SR- 99 split 13 homogeneous, looks like AC/ PCC/ Base. Not sure about the PCC
CAL035 Mar. 8, 2005 SR- 16 WB low Woodland to SR- 20 48
Mostly full depth AC, extremely variable, numerous pavement layers, may
be difficult distinguishing bound from unbound layers x
CAL036 Mar. 8, 2005 SR- 20 EB low Lake Co. line Sutter RP9 47 Mostly full depth AC, somewhat variable x
CAL037 Mar. 8, 2005 SR- 20 EB low Sutter RP9 to Grass Valley 41
Mostly full depht AC, some homogeneous sections, other areas hightly
variable, may be difficult to distinguish bound from unbound layers in
some areas
CAL041 Mar. 9, 2005 I- 80 WB faster Solano County 45 same as low speed, includes CRCP section in Fairfield x
CAL042 Mar. 9, 2005 I- 80 EB faster Solano County 45 same as low speed
CAL043 Mar. 9, 2005 SR- 160 SB low From I- 80 to Rio Vista 46
Mix of full depth AC and AC/ PCC. Some long homog. sections, some
areas with high variability;
CAL047 Mar. 10, 2005US- 50 EB low Sunrise Blvd. to El Dor. Co. line 11
2 Miles of homogenous PCC; the rest full depth AC, with lots of layers,
and variable.. Maybe difficult to distinguish bound from unbound layers x
CAL048 Mar. 10, 2005US- 50 WB faster Sunrise Blvd. to El Dor. Co. line 11
2 Miles of homogenous PCC; the rest full depth AC, also mostly
homogeneous
CAL049 Mar. 10, 2005SR- 45 SB low Yolo County Sect. 13
Full depth AC. Homogeneous in the north end; extreme changes in
pavement structure in the south end. x
CAL050 Mar. 10, 2005I- 505 SB low I- 5 to I- 80 33 Mostly PCC, some AC/ PCC, very homogeneous x
CAL051 Mar. 10, 2005I- 505 NB faster I- 5 to I- 80 33 Mostly PCC, some AC/ PCC, very homogeneous
Total Lane Miles Data Collection = 681 Analysis Total = 307
Section
Analyzed
Stage 5 Distribution
UCPRC- RR- 2005- 11
46
Appendix C. Charts with GPR Structure Results and Data from the 2003 Pavement
Condition Survey
Stage 5 Distribution
UCPRC- RR- 2005- 11
47
0
5
10
15
20
25
30
35
40
24
21
20
19
18.5
16
15
13
12
11
9
8
7
5
4
3
2
1
0
Post Mile
Depth ( in)
0
50
100
150
200
250
300
350
400
IRI ( 0- 400 in/ mi) and Cracking ( 0- 100%)
Base, PCC
Base, PCC,
AC
PCC, AC,
Base
AC
IRI ( test)
AC - Alligator
B Cracking
PCC - 1st
Stage
Cracking
PCC - 3rd
Stage
Cracking
Figure 8. GPR cross section with selected PCS data – Sacramento SR- 99 SB ( outside lane), CAL009.
Stage 5 Distribution
UCPRC- RR- 2005- 11
48
0
5
10
15
20
25
30
35
40
22
21
20
19
18
16
15
14
13
11
10
9
8
7
6
5
4
3
2
1
Post Mile Depth ( in)
0
50
100
150
200
250
300
350
400
IRI ( 0- 350 in/ mi) and 3rd Stage Cracking ( 0- 100%)
Subbase
Base
PCC
AC
IRI
PCC - 1st
Stage
Cracking
PCC - 3rd
Stage
Cracking
Figure 9. GPR cross section with selected PCS data – Sacramento I- 5 SB ( outside lane), CAL011.
Stage 5 Distribution
UCPRC- RR- 2005- 11
49
0
5
10
15
20
25
30
35
40
45
27.5
28
29
33
34
35
36
36.86 / 0
2
4
5
6
7
Post Mile Depth ( in)
0
50
100
150
200
250
300
350
400
450
IRI ( 0- 450 in/ mi) and Cracking ( 0- 100%)
Subbase
Base, BB
PCC, AC,
Base, BB
AC
IRI
AC - Alligator B
Cracking
I- 5 Postmiles
Figure 10. GPR cross section with selected PCS data – Sacramento and Sutter SR- 99 NB ( outside lane), CAL013.
Stage 5 Distribution
UCPRC- RR- 2005- 11
50
0
5
10
15
20
25
30
35
40
1
2
3
5
6
7
8
9
10
Post Mile
Depth ( in)
0
50
100
150
200
250
300
350
400
IRI ( 0- 400 in/ mi) and Cracking ( 0- 100%)
Subbase
Base
PCC
IRI
PCC - 1st Stage
Cracking
PCC - 3rd Stage
Cracking
Figure 11. GPR cross section with selected PCS data – Yolo SR- 113 NB ( outside lane), CAL015.
Stage 5 Distribution
UCPRC- RR- 2005- 11
51
0
5
10
15
20
25
30
35
40
28
27
26
25
24
21
20
19
18
17
16
15
14
13
12
10
Post Mile
Depth ( in)
0
50
100
150
200
250
300
350
400
IRI ( 0- 400 in/ mi) and Cracking ( 0- 100%)
Base
PCC
AC
IRI
AC - Alligator B
Cracking
Figure 12. GPR cross section with selected PCS data – Sacramento I- 5 SB ( inside lane), CAL017.
Stage 5 Distribution
UCPRC- RR- 2005- 11
52
0
5
10
15
20
25
30
35
40
44
43
42
41
39
38
37
36
34
33
32
31
29
28
27
26
24
23
21
18
11
10
9
8
7
6
5
4
Post Miles Depth ( in)
0
50
100
150
200
250
300
350
400
IRI ( 0- 400 in/ mi) and Cracking ( 0 - 100%)
Base
Base, Subbase,
PCC, AC
AC, PCC,
Base, BB
AC, PCC
IRI
AC - Alligator B
Cracking
PCC - 1st
Stage Cracking
PCC - 3rd
Stage Cracking
Figure 13. GPR cross section with selected PCS data – Solano I- 80 WB ( outside lane), CAL031.
Stage 5 Distribution
UCPRC- RR- 2005- 11
53
0
5
10
15
20
25
30
35
31
32
34
34.66 / 0
1
2
3
4
5
6
7
Post Mile Depth ( in)
0
50
100
150
200
250
300
350
IRI ( 0- 350 in/ mi) and Cracking ( 0- 100%)
Base ( 2)
Base ( 1)
PCC
AC
IRI
AC - Alligator B
Cracking
PCC - 1st Stage
Cracking
PCC - 3rd Stage
Cracking
Figure 14. GPR cross section with selected PCS data – Sacramento and Yolo I- 15 NB ( outside lane), CAL033.
Stage 5 Distribution
UCPRC- RR- 2005- 11
54
0
5
10
15
20
25
30
35
40
40
38
36
35
34
33
31
30
29
27
26
25
24
23
22
20
19
18
16
15
13
12
11
10
9
7
6
5
4.02
2.98
2.02
0.98
0 / 7
6
4.02
3
2
1
0
Post Mile
Depth ( in)
0
50
100
150
200
250
300
350
400
IRI ( 0- 400 in/ mi) and Cracking ( 0- 100%)
Base
Base, AC
Base, BB,
AC
AC
IRI
AC -
Alligator B
Cracking
Figure 15. GPR cross section with selected PCS data – Colusa and Yolo SR- 16 WB ( outside lane), CAL035.
Stage 5 Distribution
UCPRC- RR- 2005- 11
55
0
5
10
15
20
25
30
35
40
44
43
42
41
39
38
37
36
34
33
32
31
29
28
27
26
24
23
21
18
11
10
9
8
7
6
5
4
Post Mile
Depth ( in)
0
50
100
150
200
250
300
350
400
IRI ( 0- 400 in/ mi) and Cracking ( 0- 100%)
Base ( 2)
Base ( 1)
AC ( 2)
AC ( 1)
IRI
AC - Alligator B
Cracking
1st Stage
Cracking
PCC - 3rd Stage
Cracking
Figure 16. GPR cross section with selected PCS data – Solano I- 80 WB ( inside lane), CAL041.
Stage 5 Distribution
UCPRC- RR- 2005- 11
56
0
5
10
15
20
25
30
35
40
18
19
20
21
Post Mile
Depth ( in)
0
50
100
150
200
250
300
350
400
IRI ( 0- 400 in/ mi) and Cracking ( 0- 100%)
Base, BB
Base, BB
AC
PCC, AC
IRI
AC - Alligator B
Cracking
PCC - 1st Stage
Cracking
PCC - 3rd Stage
Cracking
Figure 17. GPR cross section with selected PCS data – Sacramento US- 50 EB ( ouside lane), CAL047.
Stage 5 Distribution
UCPRC- RR- 2005- 11
57
0
5
10
15
20
25
30
35
40
12.9
11
10
9
8
7
6
5
4
3
2
1
Post Mile
Depth ( in)
0
50
100
150
200
250
300
350
400
IRI ( 0- 400 in/ mi) and Cracking ( 0- 100%)
Base
AC, Base, BB
AC
IRI
AC - Alligator B
Cracking
Figure 18. GPR cross section with selected PCS data – Yolo SR- 45 SB ( outside lane), CAL049a.
Stage 5 Distribution
UCPRC- RR- 2005- 11
58
0
5
10
15
20
25
30
35
40
45
50
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
5
4
3
2
1
0 / 10.7
6
5
4
3
2
Post Mile
Depth ( in)
0
50
100
150
200
250
300
350
400
450
500
IRI ( 0- 500 in/ mi) and Cracking ( 0- 100%)
Base, Subbase
Base, Subbase,
PCC
PCC, AC
AC
IRI
AC - Alligator B
Cracking
PCC - 1st Stage
Cracking
PCC - 3rd Stage
Cracking
Figure 19. GPR cross section with selected PCS data – Yolo and Solano I- 505 SB ( outside lane), CAL050.
Stage 5 Distribution
UCPRC- RR- 2005- 11
59
Appendix D. Segmentation Results
Stage 5 Distribution
UCPRC- RR- 2005- 11
60
PM from Traffic
Data Physical Reference
Section ID Route County Direction
Climate
Region Start End Start End
CAL009 99 Sacramento SB IV 24.35 23.13 SACRAMENTO, JCT. RTE. 51, NORTH JCT.
RTE. 50; END FREEWAY
SACRAMENTO, 12TH AVENUE
IV 23.13 21.94 SACRAMENTO, 12TH AVENUE SACRAMENTO, FRUITRIDGE ROAD
IV 21.94 21.57 SACRAMENTO, FRUITRIDGE ROAD MARTIN LUTHER KING JR. BOULEVARD
IV 21.57 21.46 MARTIN LUTHER KING JR. BOULEVARD observed structure change
IV 21.46 20.86 observed structure change 47TH AVENUE
IV 20.86 19.61 47TH AVENUE FLORIN ROAD
IV 19.61 17.66 FLORIN ROAD SACRAMENTO, MACK ROAD
IV 17.66 17.46 SACRAMENTO, MACK ROAD observed structure change
IV 17.46 17.24 observed structure change SACRAMENTO, STOCKTON BOULEVARD
IV 17.24 15.90 SACRAMENTO, STOCKTON BOULEVARD COSUMNES RIVER BOULEVARD/ CALVINE
ROAD
IV 15.90 15.66 COSUMNES RIVER BOULEVARD/ CALVINE
ROAD
observed structure change
IV 15.66 15.16 observed structure change observed structure change
IV 15.16 14.87 observed structure change SHELDON ROAD
IV 14.87 13.84 SHELDON ROAD LAGUNA BOULEVARD/ BOND ROAD
IV 13.84 12.76 LAGUNA BOULEVARD/ BOND ROAD ELK GROVE BOULEVARD
IV 12.76 11.26 ELK GROVE BOULEVARD observed structure change
IV 11.26 10.07 observed structure change GRANT LINE ROAD
IV 10.07 9.26 GRANT LINE ROAD observed structure change
IV 9.26 8.96 observed structure change ESCHINGER ROAD
IV 8.96 8.46 ESCHINGER ROAD observed structure change
IV 8.46 7.96 observed structure change observed structure change
IV 7.96 7.36 observed structure change DILLARD ROAD
IV 7.36 6.01 DILLARD ROAD ARNO ROAD
IV 6.01 4.39 ARNO ROAD MINGO ROAD
IV 4.39 3.56 MINGO ROAD observed structure change
IV 3.56 3.53 observed structure change TWIN CITIES, JCT. RTE. 104 EAST
IV 3.53 3.26 TWIN CITIES, JCT. RTE. 104 EAST observed structure change
IV 3.26 3.16 observed structure change observed structure change
IV 3.16 2.70 observed structure change WALNUT STREET
IV 2.70 2.26 WALNUT STREET observed structure change
IV 2.26 2.16 observed structure change observed structure change
IV 2.16 1.88 observed structure change GALT, PRINGLE AVENUE
IV 1.88 1.57 GALT, PRINGLE AVENUE GALT, SIMMERHORN ROAD
IV 1.57 1.26 GALT, SIMMERHORN ROAD observed structure change
IV 1.26 0.79 observed structure change GALT, C STREET
IV 0.79 0.33 GALT, C STREET GALT, FRONTAGE ROAD
IV 0.33 0.12 GALT, FRONTAGE ROAD SAN JOAQUIN- SACRAMENTO COUNTY LINE
CAL011 5 Sacramento SB IV 22.57 20.53 SACRAMENTO, JCT. RTE. 50 SACRAMENTO, SUTTERVILLE ROAD
IV 20.53 19.30 SACRAMENTO, SUTTERVILLE ROAD SACRAMENTO, SEAMAS AVENUE ( FRUITRIDGE)
IV 19.30 18.65 SACRAMENTO, SEAMAS AVENUE SACRAMENTO, 43RD AVENUE
Stage 5 Distribution
UCPRC- RR- 2005- 11
61
PM from Traffic
Data Physical Reference
Section ID Route County Direction
Climate
Region Start End Start End
( FRUITRIDGE)
IV 18.65 17.19 SACRAMENTO, 43RD AVENUE SACRAMENTO, FLORIN ROAD
IV 17.19 16.15 SACRAMENTO, FLORIN ROAD SACRAMENTO, POCKET/ MEADOWVIEW
ROADS
IV 16.15 15.65 SACRAMENTO, POCKET/ MEADOWVIEW
ROADS
observed structure change
IV 15.65 13.05 observed structure change observed structure change
IV 13.05 12.04 observed structure change LAGUNA BOULEVARD
IV 12.04 10.83 LAGUNA BOULEVARD ELK GROVE BOULEVARD
IV 10.83 8.49 ELK GROVE BOULEVARD HOOD- FRANKLIN ROAD
IV 8.49 4.65 HOOD- FRANKLIN ROAD Lambert Road
IV 4.65 2.13 Lambert Road TWIN CITIES ROAD
IV 2.13 0.02 TWIN CITIES ROAD SAN JOAQUIN- SACRAMENTO COUNTY LINE
CAL013 99 Sacramento/
Sutter
NB IV 26.72 26.76 SACRAMENTO, JCT. RTE. 80 ( I- 5 Postmile) observed structure change
IV 26.76 26.96 observed structure change observed structure change
IV 26.96 29.02 observed structure change SACRAMENTO, DEL PASO ROAD ( I- 5 Postmile)
IV 29.02 29.91 SACRAMENTO, DEL PASO ROAD ( I- 5 Postmile) SACRAMENTO, JCT. RTE. 99 NORTH ( I- 5 Postmile)
- Start SR 99 Postmiles
IV 32.12 32.67 SACRAMENTO, JCT. RTE. 99 NORTH ( I- 5
Postmile) - Start SR 99 Postmiles
observed structure change
IV 32.67 33.36 observed structure change ELKHORN BOULEVARD
IV 33.36 35.37 ELKHORN BOULEVARD ELVERTA ROAD
IV 35.37 36.86 ELVERTA ROAD Sacramento- Sutter County Line
IV 0.00 0.61 Sacramento- Sutter County Line observed structure change
IV 0.61 0.95 observed structure change RIEGO ROAD
IV 0.95 4.21 RIEGO ROAD observed structure change
IV 4.21 5.91 observed structure change observed structure change
IV 5.91 6.11 observed structure change observed structure change
IV 6.11 6.83 observed structure change observed structure change
IV 6.83 8.07 observed structure change JCT. RTE. 70 NORTH
CAL015 113 Yolo NB IV 0.42 1.08 HUTCHINSON DRIVE DAVIS, RUSSELL BOULEVARD
IV 1.08 2.08 DAVIS, RUSSELL BOULEVARD COUNTY ROAD 31
IV 2.08 4.11 COUNTY ROAD 31 COUNTY ROAD 29
IV 4.11 5.80 COUNTY ROAD 29 observed structure change
IV 5.80 6.11 observed structure change COUNTY ROAD 27
IV 6.11 7.66 COUNTY ROAD 27 COUNTY ROAD 25
IV 7.66 9.23 COUNTY ROAD 25 WOODLAND, GIBSON ROAD
IV 9.23 10.15 WOODLAND, GIBSON ROAD WOODLAND, EAST MAIN STREET
IV 10.15 10.72 WOODLAND, EAST MAIN STREET WOODLAND, JCT. RTE. 5
CAL017 5 Yolo SB IV 28.92 25.57 YOLO COUNTY- COLUSA COUNTY ( COUNTY COUNTY ROAD 6
IV 25.57 23.79 COUNTY ROAD 6 COUNTY ROAD 8
IV 23.79 22.61 COUNTY ROAD 8 JCT. RTE. 505 SOUTH
Stage 5 Distribution
UCPRC- RR- 2005- 11
62
PM from Traffic
Data Physical Reference
Section ID Route County Direction
Climate
Region Start End Start End
IV 22.61 21.80 JCT. RTE. 505 SOUTH observed structure change
IV 21.80 17.62 observed structure change ZAMORA INTERCHANGE, COUNTY ROAD 13
IV 17.62 12.34 ZAMORA INTERCHANGE, COUNTY ROAD 13 YOLO INTERCHANGE, COUNTY ROAD 17
IV 12.34 10.81 YOLO INTERCHANGE, COUNTY ROAD 17 JCT. RTE. 16, COUNTY ROAD 18
IV 10.81 9.41 JCT. RTE. 16, COUNTY ROAD 18 COUNTY ROAD 99/ WEST STREET
IV 9.41 8.26 COUNTY ROAD 99/ WEST STREET WOODLAND, JCT. RTE. 113 NORTH
CAL031 80 Solano WB IV 44.72 42.67 SOLANO- YOLO COUNTY LINE JCT. RTE. 113 NORTH
IV 42.67 41.90 JCT. RTE. 113 NORTH observed structure change
IV 41.90 40.30 observed structure change observed structure change
IV 40.30 39.74 observed structure change Pedrick
IV 39.74 38.60 Pedrick observed structure change
IV 38.60 38.21 observed structure change JCT. RTE. 113 SOUTH
IV 38.21 36.90 JCT. RTE. 113 SOUTH Pitt School Road
IV 36.90 35.55 Pitt School Road DIXON AVENUE/ GRANT ROAD
IV 35.55 32.62 DIXON AVENUE/ GRANT ROAD Midway
IV 32.62 31.36 Midway Meridian
IV 31.36 29.86 Meridian Leisure Town
IV 29.86 28.36 Leisure Town VACAVILLE, JCT. RTE. 505 NORTH
IV 28.36 27.24 VACAVILLE, JCT. RTE. 505 NORTH VACAVILLE, MONTE VISTA AVENUE
IV 27.24 26.46 VACAVILLE, MONTE VISTA AVENUE Mason/ Elmira
IV 26.46 26.01 Mason/ Elmira VACAVILLE, DAVIS STREET
IV 26.01 25.31 VACAVILLE, DAVIS STREET VACAVILLE, ALAMO DRIVE
IV 25.31 23.96 VACAVILLE, ALAMO DRIVE PLEASANT VALLEY/ Pena Adobe Road
IV 23.96 20.80 PLEASANT VALLEY/ Pena Adobe Road FAIRFIELD, NORTH TEXAS STREET
IV 20.80 19.18 FAIRFIELD, NORTH TEXAS STREET FAIRFIELD, AIRBASE PARKWAY
IV 19.18 17.92 FAIRFIELD, AIRBASE PARKWAY FAIRFIELD, TRAVIS BOULEVARD
IV 17.92 17.20 FAIRFIELD, TRAVIS BOULEVARD FAIRFIELD, WEST TEXAS STREET
IV 17.20 15.82 FAIRFIELD, WEST TEXAS STREET FAIRFIELD, EAST JCT. RTE. 12
IV 15.82 15.20 FAIRFIELD, EAST JCT. RTE. 12 observed structure change
IV 15.20 13.49 observed structure change FAIRFIELD, SUISUN VALLEY ROAD
IV 13.49 12.84 FAIRFIELD, SUISUN VALLEY ROAD FAIRFIELD, JCT. RTE. 680 SOUTH
IV 12.84 12.70 FAIRFIELD, JCT. RTE. 680 SOUTH observed structure change
IV 12.70 11.98 observed structure change FAIRFIELD, JCT. RTE. 12 WEST
IV 12.22 11.98 MILEPOST EQUATION = 12.20 FAIRFIELD, JCT. RTE. 12 WEST
IV 11.98 11.39 FAIRFIELD, JCT. RTE. 12 WEST FAIRFIELD, RED TOP ROAD
IV 11.39 9.65 FAIRFIELD, RED TOP ROAD observed structure change/ climate region change
CC 9.65 ~ 8.2 observed structure change observed structure change
CC ~ 8.2 8.10 observed structure change AMERICAN CANYON ROAD
CC 8.10 8.00 AMERICAN CANYON ROAD NAPA- SOLANO COUNTY LINE
CC 8.00 6.81 NAPA- SOLANO COUNTY LINE SOLANO- NAPA COUNTY LINE
CC 6.81 5.63 SOLANO- NAPA COUNTY LINE VALLEJO, JCT. RTE. 37 WEST
CC 5.63 ~ 5.2 VALLEJO, JCT. RTE. 37 WEST observed structure change
CC ~ 5.2 4.43 observed structure change VALLEJO, REDWOOD STREET
Stage 5 Distribution
UCPRC- RR- 2005- 11
63
PM from Traffic
Data Physical Reference
Section ID Route County Direction
Climate
Region Start End Start End
CC 4.43 3.49 VALLEJO, REDWOOD STREET VALLEJO, TENNESSEE STREET
CC 3.49 3.23 VALLEJO, TENNESSEE STREET VALLEJO, SPRINGS ROAD
CC 3.23 2.88 VALLEJO, SPRINGS ROAD VALLEJO, GEORGIA STREET
CC 2.88 2.22 VALLEJO, GEORGIA STREET VALLEJO, JCT. RTE. 780 SOUTHEAST
CC 2.22 1.78 VALLEJO, JCT. RTE. 780 SOUTHEAST VALLEJO, MAGAZINE STREET
CC 1.78 1.14 VALLEJO, MAGAZINE STREET VALLEJO, JCT RTE 29 NORTHWEST
CC 1.14 0.00 VALLEJO, JCT RTE 29 NORTHWEST SOLANO COUNTY ( CARQUINEZ BRIDGE)
CAL033 5 Sacramento/
Yolo
NB IV 29.91 32.73 SACRAMENTO, JCT. RTE. 99 NORTH AIRPORT BOULEVARD
IV 32.73 34.35 AIRPORT BOULEVARD observed structure change
IV 34.35 34.65 observed structure change Sacramento- Yolo County Line
IV 0.00 0.50 Sacramento- Yolo County Line observed structure change
IV 0.50 0.80 observed structure change observed structure change
IV 0.80 2.60 observed structure change observed structure change
IV 2.60 5.53 observed structure change COUNTY ROAD 102
IV 5.53 6.51 COUNTY ROAD 102 WOODLAND, EAST MAIN STREET
IV 6.51 7.09 WOODLAND, EAST MAIN STREET WOODLAND, JCT. RTE. 113 SOUTH
IV 7.09 8.26 WOODLAND, JCT. RTE. 113 SOUTH WOODLAND, JCT. RTE. 113 NORTH
CAL035 16 Colusa/ Yolo WB IV 40.57 39.56 WEST MAIN STREET/ COUNTY ROAD 98 COUNTY ROAD 97
IV 39.56 36.71 COUNTY ROAD 97 COUNTY ROAD 94B
IV 36.71 35.44 COUNTY ROAD 94B observed structure change
IV 35.44 32.34 observed structure change observed structure change
IV 32.34 31.87 observed structure change JCT. RTE. 505; MADISON, EAST
IV 31.87 31.03 JCT. RTE. 505; MADISON, EAST MADISON, COUNTY ROAD 89
IV 31.03 28.27 MADISON, COUNTY ROAD 89 COUNTY ROAD 21A
IV 28.27 27.96 COUNTY ROAD 21A GRAFTON STREET
IV 27.96 27.55 GRAFTON STREET ESPARTO, ORLEANS STREET
IV 27.55 26.37 ESPARTO, ORLEANS STREET COUNTY ROAD 85B
IV 26.37 25.15 COUNTY ROAD 85B CAPAY, CAPAY CANAL BRIDGE
IV 25.15 20.17 CAPAY, CAPAY CANAL BRIDGE COUNTY ROAD 78A
IV 20.17 19.43 COUNTY ROAD 78A INDIAN BINGO ROAD
IV 19.43 19.20 INDIAN BINGO ROAD WINNERS WAY
IV 19.20 18.78 WINNERS WAY COUNTY ROAD 78
IV 18.78 18.13 COUNTY ROAD 78 MOSSY CREEK BRIDGE
IV 18.13 ~ 14.4 MOSSY CREEK BRIDGE observed structure change
IV ~ 14.4 12.21 observed structure change GUINDA, COUNTY ROAD 57
IV 12.21 10.80 GUINDA, COUNTY ROAD 57 COUNTY ROAD 45
IV 10.80 7.15 COUNTY ROAD 45 RUMSEY, MANZANITA AVENUE ( TO
ARBUCKLE)
IV 7.15 0.00 RUMSEY, MANZANITA AVENUE ( TO
ARBUCKLE)
Yolo- Colusa County Line
IV 7.26 0.00 Yolo- Colusa County Line BEAR CREEK, JCT. RTE. 20
CAL041 80 Solano WB IV 44.72 42.67 SOLANO- YOLO COUNTY LINE JCT. RTE. 113 N