See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/245560975

Assessment of Overlay Roughness in Long-Term Pavement Performance Test

Sites: Canadian Case Study

Article  in  Transportation Research Record Journal of the Transportation Research Board · January 2004
DOI: 10.3141/1869-15

CITATIONS READS
13 188

2 authors, including:

Susan Tighe
University of Waterloo
371 PUBLICATIONS   2,551 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Long Term Pavement Performance Program View project

Precast Concrete Inlay Panel Rehabilitation Strategy View project

All content following this page was uploaded by Susan Tighe on 29 October 2015.

The user has requested enhancement of the downloaded file.

Assessment of Overlay Roughness in
Long-Term Pavement Performance
Test Sites
Canadian Case Study
James T. Smith and Susan L. Tighe
A study was conducted on asphalt pavement overlay performance in the
Canadian environment. It investigated the impact of asphalt overlay
thickness, climatic zone, and subgrade type on the progression of rough-
ness as described by the international roughness index (IRI). Data from
the Canadian Long-Term Pavement Performance (LTPP) test sites
were analyzed. As a result of the investigation, pavement factors that
significantly impact overlay performance in the Canadian environment
were identified. Data collected over the first 13 years of study were used
to show national and provincial roughness trends from 53 test sites. The
IRI data were statistically summarized (mean, standard deviation) for
each category by the age of the overlay section. With the summarized
data, regression analysis was used to determine an equation that best
describes the progression of roughness. Two-factor analysis of variance
was used to determine if there were any significant differences within
specific categories. The results of the regression analysis were com-
pared with the Canadian Strategic Highway Research Project LTPP to
confirm the validity of the roughness progression equations. Results
show that overlay thickness and climatic zones significantly impact the
roughness, while subgrade type has little influence on the IRI values.
The roughness progression equations achieved squared correlation
coefficients (R2) between 0.93 and 0.39, demonstrating the accuracy of
the model equations.
In 1987, as part of a comprehensive 20-year study of in-service
pavements, the Long-Term Pavement Performance (LTPP) program
was initiated. The purpose of this program was to develop an under-
standing of why some pavements perform better than others and
how to maintain a cost-effective highway system (1). The LTPP
program monitors 107 test sections within Canada. At each site, data
on distress, roughness, structural capacity, traffic, and other pave-
ment performance measures were collected. The individual test sites
are identified by a climatic zone, overlay thickness, and subgrade.
Figure 1 shows the distribution, identification numbers, and charac-
teristics of the 40 test sites in the LTPP study that have overlays and
are included in the data analysis for this paper.
The international roughness index (IRI) is a measurement scale
to evaluate pavement roughness. The index is based on a response-
Department of Civil Engineering, University of Waterloo, Waterloo, Ontario,
Canada, N2L 3G1.
Transportation Research Record: Journal of the Transportation Research Board,
No. 1869, TRB, National Research Council, Washington, D.C., 2004, pp. 126–135.
126
type road roughness measuring system (RTRRMS) in relationship
to the longitudinal profile of the road surface. IRI is calculated by
dividing the accumulated output of the RTRRMS with the profile
length (2). The IRI index is measured in units of meters per kilo-
meter. An absolutely perfect pavement profile, one with no vertical
displacements, has an IRI value equal to 0 m/km. As the index value
increases, the smoothness of the road decreases. There is no maxi-
mum limit to the scale; however, IRI values greater than 8 m/km are
classified as damaged pavements or rough unpaved roads.
SCOPE AND OBJECTIVES
This paper analyzes the relationship between pavement performance
measured by IRI and the age of the asphalt pavement overlay. All the
data used in the analysis were extracted from the LTPP National
Information Management System, DataPave Online Release 15.
Based on this extraction, the following analysis was carried out.
• National and provincial roughness trends were summarized.
• The effect of climatic zones [wet–freeze (WF), wet–no freeze
(WNF), and dry–freeze (DF)] on roughness progression was analyzed.
• The effect of overlay thicknesses (30 to 60 mm, 60 to 100 mm,
and 100 to 185 mm) on roughness progression was analyzed.
• The effect of subgrade types (coarse and fine) on roughness
progression was analyzed.
• The LTPP results were compared with the Canadian Strategic
Highway Research Project (C-SHRP) LTPP study.
Overall, the paper is directed at determining the pavement factors
that have the most significant impact on overlay performance in the
Canadian environment.
IMPORTANCE OF IRI AS PAVEMENT INDICATOR
Pavement roughness is the primary measure that is used by most
transportation agencies to establish the need for rehabilitation. Pave-
ment roughness affects driving comfort, vehicle operating costs, and
safety (3). The ASTM defines pavement roughness as “the deviation
of the surface from a true planar surface with characteristic dimen-
sions that affect vehicle dynamic, ride quality, dynamic loads, and
 0 sites
in NF
3 sites 10 sites 4 sites 11 sites
in BC in AB in SK in MB
6 sites 4 sites 0 sites
in ON in QC in PE

0 sites
2 sites in NS
in NB

(a)

Test Site Years of Overlay Thickness (mm) Climatic Region Subgrade Type
ID Data 0-60 60-100 100-185 WF WNF DF CG FG
82-1005 3 ✓ ✓ ✓
BC

82-6006 11 ✓ ✓ ✓
82-6007 9 ✓ ✓ ✓ ✓
81-502 12 ✓ ✓ ✓
81-503 12 ✓ ✓ ✓
81-504 12 ✓ ✓ ✓
81-505 12 ✓ ✓ ✓
81-506 12 ✓ ✓ ✓
AB

81-507 12 ✓ ✓ ✓
81-508 12 ✓ ✓ ✓
81-509 12 ✓ ✓ ✓
81-1804 6 ✓ ✓ ✓
81-1805 3 ✓ ✓ ✓
90-6410 4 ✓ ✓ ✓
90-6412 4 ✓ ✓ ✓
SK

90-A310 6 ✓ ✓ ✓
90-B310 8 ✓ ✓ ✓
83-502 12 ✓ ✓ ✓
83-503 12 ✓ ✓ ✓
83-504 12 ✓ ✓ ✓
83-505 12 ✓ ✓ ✓
83-507 11 ✓ ✓ ✓
MB

83-508 12 ✓ ✓ ✓
83-509 13 ✓ ✓ ✓
83-6450 13 ✓ ✓ ✓
83-6451 13 ✓ ✓ ✓
83-6452 13 ✓ ✓ ✓
83-A310 12 ✓ ✓ ✓
87-1620 8 ✓ ✓ ✓
87-1622 3 ✓ ✓ ✓
87-1680 5 ✓ ✓ ✓
ON

87-1806 5 ✓ ✓ ✓
87-A310 8 ✓ ✓ ✓
87-B310 8 ✓ ✓ ✓
89-1021 3 ✓ ✓ ✓
89-1125 3 ✓ ✓ ✓
QC

89-1127 5 ✓ ✓ ✓
89-A310 2 ✓ ✓ ✓
84-1684 4 ✓ ✓ ✓
NB

84-6804 7 ✓ ✓ ✓
(b)

FIGURE 1 Characteristics of LTPP test sites and identification numbers (BC  British
Columbia, AB  Alberta, SK  Saskatchewan, MB  Manitoba, ON  Ontario, QC  Quebec,
NB  New Brunswick, NS  Nova Scotia, PE  Prince Edward Island, NF  Newfoundland,
CG  course grained, FG  fine grained).
128 Transportation Research Record 1869

drainage.” The longevity of a pavement and required maintenance
is affected by the initial roughness. An asphalt overlay in Canada
has a typical expected service life of 8 to 15 years depending on
thickness, local conditions, traffic levels, and climate (3). For the
purpose of this paper, the trigger level or action point for restora-
tion of an overlay pavement should occur at an IRI value of 2.66
m/km. This value illustrated in Figure 2 corresponds to a pavement
condition index (PCI) value of 50 and present serviceability index
(PSI) value of 2.5 (4, 5).
PSI = PSR ± e = 5 × e( −0.26  IRI )
where PSR is the present serviceability rating.
METHODOLOGY FOR ANALYSIS
The null hypothesis (H0) states that the population means are equal.
The alternative hypothesis (Ha) is true if there is any difference
between the population means. If the variation between groups is
large compared to the variation within the groups, there is evidence
against the null hypothesis (6 ).
Analysis was conducted using a four-step process.
1. Sample means and standard deviations were calculated for all
of the groups.
2. Plots were made to provide an overview of the data. These
(1) plots were observed for any extreme deviations from normal.
3. Null and alternative hypotheses were formulated for each
option.
4. ANOVA analysis was conducted and H0 was accepted if FCalc
≤ FCrit or rejected if FCalc > FCrit.

Two-way analysis of variance (ANOVA) was used to compare pop-
ulation means based on a simple random sample. Each population
is assumed to be normal, possibly with different means and the same
standard deviation. ANOVA separates the total variation of the data
into variation between group means and variation within groups.
NATIONAL ROUGHNESS TREND
Figure 3 is a boxplot that shows the national roughness trend for all
the Canadian test site overlays, from 1989 to 2002, in the LTPP pro-
gram. For each year, six measures are used to describe roughness.

IRI PCI PSI

0 100 5 Perfect
Excellent

0.40

0.86 80 4
Satisfactory Condition and
Vehicle Operating Costs
Good

1.37

1.96 60 3
Fai r

2.66 Action Point for Restoration

3.52 40 2
User Costs Sharply Rising and
Complaints Starting About Conditions
Poor

4.63

User Costs Excessive and Frequent Complaints

6.19 20 1

Extreme Discomfort, Traffic Slowdowns

Very Poor

and Extremenly High User Costs

8.85

34.63 0 0 Impassable

FIGURE 2 IRI trigger level and pavement condition.

Smith and Tighe 129

3.0
y = 1.029e0.0314x
2.5
R2 = 0.9467
2.0

IRI (m/km) 1.5

1.0

0.5

0.0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Overlay Age (Years)

Quartile 1 (25%) Minimum Median Mean

Maximum Quartile 3 (75%) Expon. ( Mean)

FIGURE 3 National IRI trends (Expon.  exponential).

Minimum and maximum IRI values are used to illustrate the best
and worst performing test sections. The mean and median are also
given to describe the IRI distribution. The first and third quartiles
are provided to show the divergence of IRI values from the median
value and capture 50% of the test population.
Figure 3 shows an increase in the overlay roughness during
the study period. Over the study period, the average IRI measured
increased from 1.031 m/km to 1.630 m/km, representing a pavement
that is still smooth and functioning properly. The difference between
the first and third quartiles has remained fairly constant, approxi-
mately 0.4 m/km. National roughness progression is best explained
using an exponential regression equation as shown by Equation 2 and
accurately predicts the LTPP IRI data as evidence of the squared cor-
relation coefficient (R2) equal to 0.9467. The exponential regression
equation was chosen over a linear equation and a second-degree poly-
nomial because it provides the highest R2 value. Using an IRI trigger
level of 2.66 m/km for maintenance, rehabilitation, and recon-
struction, the estimated average life span for overlays in Canada is
30 years. Note that all factors have been aggregated in this analysis
and there is considerable variation (0.92 m/km) in the initial IRI
value following the asphalt overlay. However, the results show that
overlays can provide good performance.
IRI = 1.029e 0.0314 ( Age )
PROVINCIAL ROUGHNESS TREND
Provincial roughness trends were developed for the seven provinces
that contain LTPP monitored asphalt overlay test sections. Table 1
presents the developed provincial roughness progression equations
and an estimated life expectancy of the overlay using a trigger for
rehabilitation set at an IRI of 2.66 m/km. There was a high correla-
tion for the provinces of Alberta, Manitoba, and Ontario as demon-
strated by the high R2 values. The lowest correlation occurred in the
province of Quebec where the yearly IRI data were inconsistent
from year to year.
Figure 4 shows the boxplot roughness trends for the province of
Ontario. There are six sites, 10 test sections, and 41 observations
included in this analysis.
The provincial roughness trend shows a constant increase in the
average IRI value similar to the national roughness trend. The aver-
age IRI changed from 1.076 m/km to 1.852 m/km in 8 years. The
difference between the first and third quartiles did not remain con-
stant over the study period and varied between 0.055 m/km and
0.676. Ontario roughness progression is best explained using an
exponential regression equation as shown by Equation 3. The high
R2 value of 0.9165 shows that Equation 3 accurately predicts the
LTPP IRI data. Similar to the national trends, an exponential regres-
(2) sion equation was selected because it provided the highest R2 value.

TABLE 1 Provincial Regression Analysis

Province Regression Equation R2 NOBS. Life (Years)

0.0523x
British Columbia (BC) y = 0.8465e 0.544 23 22
0.0203x
Alberta (AB) y = 1.1176e 0.9068 104 42
2
Saskatchewan (SK) y = 0.0199x - 0.0699x + 1.1612 0.6213 22 11
2
Manitoba (MB) y = 0.0022x + 0.0187x + 0.9934 0.9087 148 24
0.0671x
Ontario (ON) y = 1.0288e 0.9165 41 14
Quebec (PQ) y = 0.0088x + 1.2467 0.0053 14 -
0.0516x
New Brunswick (NB) y = 0.7651e 0.6723 12 24
NOTE: NOBS. = number of observations.
130 Transportation Research Record 1869

3.0

2.5
y = 1.0288e0.0671x
R2 = 0.9165
2.0

IRI (m/km)
1.5

1.0

0.5

0.0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Overlay Age (Years)

Quartile 1 (25%) Minimum Median Mean

Maximum Quartile 3 (75%) Expon. ( Mean)

FIGURE 4 Ontario IRI trends (Expon.  exponential).

An overlay in the province of Ontario has an estimated useful life of
approximately 14 years. This significantly lower service life, com-
pared with the national value, could be attributed to Ontario having
very high traffic loads and a climate that includes large extremes in
temperature. Both of these factors could result in decreased service
life as a result of traffic and environmental related distresses.
IRI = 1.0288e 0.0671( Age )
REGRESSION ANALYSIS
Regression analysis was performed on pavement classes to fa-
cilitate a comparison between the LTPP and the C-SHRP LTPP
studies. The C-SHRP LTPP study found that overlay thickness
and climatic region have a significant impact on the progression
of roughness.
Table 2 presents the results of the regression analysis and estimated
life span of the overlay for all of the pavement categories. The equa-
tions developed through regression analysis provide a good estimate
of roughness progression based on the calculated R2 values. R2 values
ranged between 0.9862 for WF, coarse-grained medium overlay pave-
(3) ment and 0.5458 for WNF, coarse-grained medium overlay pave-
ment. By using the developed regression equations it can be shown
that the pavement overlays have an estimated useful life between
8 and 35 years. Because the regression equations are based on up to
13 years of data, special care must be taken in forecasting future per-
formance. For example, the DF, coarse-grained thick overlay pave-
ment was estimated to have an unrealistic life span of 44 years
because of the small number of observations that all occurred before
the overlay was 5 years old and started to experience degradation.

TABLE 2 Pavement Class Regression Analysis

2
Pavement Class Regression Equation R NOBS. Life (Years)
0.0671x
Wet–freeze/Fine grained/ Thin y = 1.0288e 0.9165 41 14
2
Wet–freeze/Fine grained/ Medium y = 0.0054x + 0.0047x + 1.0266 0.9719 25 17
0.0429x
Wet–freeze/Fine grained/ Thick y = 0.9115e 0.7949 37 25
2
Dry–freeze/Fine grained/ Thin y = 0.0282x - 0.0623x + 1.3508 0.8113 14 8
2
Dry–freeze/Fine grained/ Medium y = 0.012x - 0.0357x + 0.7539 0.9865 6 15
Dry–freeze/Fine grained/ Thick - - -- -
0.0307x
Wet–freeze/Coarse grained/ Thin y = 0.9736e 0.7692 72 33
2
Wet–freeze/Coarse grained/ Medium y = 0.005x - 0.0155x + 1.2152 0.9862 42 17
0.0286x
Wet–freeze/Coarse grained/ Thick y = 0.9923e 0.6202 83 35
Wet–no freeze/Coarse grained/ Thin y = 0.0562x + 0.807 0.8218 17 33
0.023x
Wet–no freeze/Coarse grained/ Medium y = 1.7035e 0.5458 3 20
Wet–no freeze/Coarse grained/ Thick - - -- -
2
Dry–freeze/Coarse grained/ Thin y = 0.005x - 0.0056x + 0.9368 0.9475 23 19
Dry–freeze/Coarse grained/ Medium - - - -
0.022x
Dry–freeze/Coarse grained/ Thick y = 1.0096e 0.8039 4 44
NOTE: NOBS. = number of observations.
Smith and Tighe 131

2.00

1.75

1.50

1.25
IRI (m/km)
1.00

0.75

0.50

0.25

0.00
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Overlay Age (Years)

> 60 mm 60 < 100 mm 100 < 180 mm

FIGURE 5 Effect of overlay thickness on roughness progression in wet–freeze climatic zones

with fine-grained subgrades.

OVERLAY THICKNESS EFFECTS CLIMATIC ZONE EFFECTS

Overlay thickness is a primary consideration for pavement designers.
Figure 5 illustrates the effects of overlay thickness on the progression
of roughness. The overlays were divided into three categories: thin
(30 to 60 mm), medium (60 to 100 mm), and thick (100 to 185 mm).
Overall, this resulted in six thin overlays, two medium overlays, and
three thick overlays. Wet–freeze, fine-grained pavements were used
for comparison because data were available for all three overlay pave-
ment categories and the results of the regression analysis yielded good
correlation coefficients allowing for a meaningful comparison.
Initially, over the first 8 years, the thin overlay had the greatest
increase in roughness, while the changes for moderate and thick
overlays remained almost identical. After the 8th year, the deteriora-
tion rate for the moderate overlay thickness was accelerated, whereas
the thick overlay increased gradually.
As the thickness of the overlay increases, the structural capacity of
the pavement increases. This allows the pavement to resist deterio-
ration and produce lower roughness values. The results must be
interpreted with caution, since thickness cannot increase indefinitely.
If the overlay thickness becomes excessive, correct compaction can-
not be achieved, which results in accelerated roughness progression
due to early failure of the pavement structure.
A two-factor ANOVA analysis conducted at an α level of 0.05
or 95% showed a significant change in the IRI values over time for
all overlay thicknesses. The results from the ANOVA analysis are
presented in Table 3.
The three climatic zones that are presented on the LTPP sites are
WF, WNF, and DF. This analysis is intended to isolate the impact
of climatic zone on performance. Overall, this resulted in 10 WF test
sections, two WNF test sections, and two DF test sections. Coarse-
grained thin overlay pavements were used for comparison because
data were available for all three climatic zone categories and the
results of the regression analysis yielded good correlation coeffi-
cients allowing for a meaningful comparison. Figure 6 presents the
relationship between climatic zone and roughness progression.
The roughness values varied widely from one year to another.
This is possibly explained by considering the time of year when the
roughness data were collected. However, regardless of the time of
year, it appears that DF zones exhibited the poorest performance in
this study.
Roughness progression is accelerated by freeze–thaw effects and
trapped water. This effect is shown by the WNF being the best per-
forming zone. The presence of water did not affect roughness pro-
gression because it does not freeze, causing additional stress and
deterioration to the pavement structure.
A two-factor ANOVA was performed to determine if the differ-
ences between the climatic zones for thin overlays on coarse-grained
subgrade were statistically significant. Two-factor ANOVA analysis
conducted at an α level of 0.05 showed a significant change in the IRI
values over time between the WF and WNF, and the DF and WNF,
climatic zones. The difference between the WF and DF climatic zones

TABLE 3 Overlay Thickness ANOVA Analysis

Pavement Classes Fcalc. Fcrit. df Sig.
Wet–freeze/Fine grained/Thin and Wet–freeze/Fine grained/Medium 7.738 5.987 13 Yes
Wet–freeze/Fine grained/Thin and Wet–freeze/Fine grained/Thick 13.655 6.608 13 Yes
Wet–freeze/Fine grained/Medium and Wet–freeze/Fine grained/Thick 12.615 4.747 25 Yes
NOTE: df = degree of freedom, Sig. = significant.
132 Transportation Research Record 1869

2.00

1.75

1.50

1.25
IRI (m/km)
1.00

0.75

0.50

0.25

0.00
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Overlay Age (Years)

Wet–Freeze (WF) Wet–No Freeze (WNF) Dry–Freeze (DF)

FIGURE 6 Effect of climatic zone on roughness progression in thin overlays with coarse-grained
subgrades.

is not statistically significant. The results from the ANOVA analysis
are presented in Table 4.
SUBGRADE EFFECTS
The next analysis in this research focused on examining subgrade
type on pavement performance. Two categories of subgrade were
used: coarse and fine. Coarse-grained subgrades are composed of
sands and gravels, whereas fine-grained subgrades are composed
of silts and clays (7 ). Overall, this resulted in two fine-grained
subgrades and four coarse-grained subgrades. WF medium overlay
pavements were used for comparison because data were available
for the two subgrade types and the results of the regression analysis
yielded good correlation coefficients, allowing for a meaningful
comparison. Figure 7 compares the effect that the roadway subgrade
has on the progression of roughness.
The progression of roughness for both coarse and fine subgrades
is similar. During the first 8 years the overlay with fine-grained sub-
grades preformed better than coarse-grained subgrades. This trend
was reversed during the second half of the life cycle.
Although the effect of subgrade type produced similar results for
pavements in a WF climatic zone and medium overlay thickness,
special attention must be given to matching the subgrade to the envi-
ronmental conditions. Pavements in areas that are frost susceptible
should avoid fine-grained subgrades because there are problems
associated with continuous freeze–thaw effects.
Two-factor ANOVA analysis conducted at an α level of 0.05
showed no significant change in the IRI values over time when
coarse- and fine-grained subgrades are compared. The results from
the ANOVA analysis are presented in Table 5.
ROUGHNESS AND C-SHRP
The C-SHRP LTPP program began in 1989, 2 years after the start
of the LTPP. The goal of the C-SHRP LTPP experiment was to
build upon the LTPP program focusing on in-service pavement per-
formance of rehabilitated pavements over 15 years at the national
and provincial levels (8).
To validate the roughness deterioration equations from the LTPP
study, they were compared to the roughness data taken from the

TABLE 4 Climate Zone ANOVA Analysis

Pavement Classes Fcalc. Fcrit. df Sig.
Wet–freeze/Coarse grained/Thin and Wet–no freeze/Coarse grained/Thin 8.125 6.608 11 Yes
Wet–freeze/Coarse grained/Thin and Dry–freeze/Coarse grained/Thin 7.188 6.608 11 Yes
Wet–no freeze/Coarse grained/Thin and Dry–freeze/Coarse grained/Thin 0.806 5.117 19 No
NOTE: df = degree of freedom, Sig. = significant.

TABLE 5 Subgrade ANOVA Analysis

Pavement Classes Fcalc. Fcrit. df Sig.
Wet–freeze/Fine grained/Medium and Wet–freeze/Coarse grained/Medium 0.870 4.747 25 No
NOTE: df = degree of freedom, Sig. = significant.
Smith and Tighe 133

2.00

1.75

1.50

1.25

IRI (m/km)
1.00

0.75

0.50

0.25

0.00
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Overlay Age (Years)

Fine Grained (FG) Coarse Grained (CG)

FIGURE 7 Effect of subgrade type on roughness progression in medium overlays with

wet–freeze climatic zones.

2
1.75
1.5
IRI (m/km)

1.25
1
0.75
0.5
0.25
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Overlay Age (Years)
LTPP C-SHRP LTPP-Regression Eqn

FIGURE 8 Roughness progression of thin overlays in wet–freeze climatic zones with fine-
grained subgrades (Eqn  equation).

2
1.75
1.5
IRI (m/km)

1.25
1
0.75
0.5
0.25
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Overlay Age (Years)
LTPP C-SHRP LTPP-Regression Eqn

FIGURE 9 Roughness progression of medium overlays in wet–freeze climatic zones with

fine-grained subgrades (Eqn  equation).
134 Transportation Research Record 1869

C-SHRP LTPP study and the R2 was calculated (9). Figure 8 presents
the comparison between the two LTPP studies for thin overlays in a
WF fine-grained subgrade. The second-degree polynomial regression
equation accurately predicts the C-SHRP roughness deterioration
with a R2 of 0.9347.
IRI = 0.0059( Age)2 + 0.0449( Age) + 1.0536
Figure 9 illustrates the comparison between the two LTPP studies
for moderate overlays in a WF fine-grained subgrade. The second-
degree polynomial regression equation accurately predicts the
C-SHRP roughness deterioration with a R2 of 0.7497.
– As the thickness of the overlay increased, the pavement per-
formance increased in areas characterized by WF climates and
fine-grained subgrade.
– WNF climatic regions had the best pavement performance,
while DF regions performed the worst for test sections with a thin
overlay and coarse-grained subgrade.
• Subgrade type has little influence on the IRI values for asphalt
( 4) overlays.
• Asphalt overlays in Canada should have a life span between 8
and 35 years.
• LTPP regression equations adequately explain roughness pro-
gression for the C-SHRP sites.

IRI = 0.0054( Age)2 + 0.0047( Age) + 1.0266
Figure 10 shows the comparison between the two LTPP studies
for thick overlays in a WF fine-grained subgrade. The exponential
regression equation moderately predicts the C-SHRP roughness
deterioration with a R2 of 0.3878.
IRI = 0.9115e 0.0429( Age )
CONCLUSIONS
The major conclusions and findings from the first 13 years of the
LTPP experiment can be summarized as follows:
• The progression of overlay roughness on a national level
increases steadily over time and is appropriately explained using an
exponential regression equation.
• Provincial trends follow the same trends as the national average.
• Overlay thickness and climatic zones significantly affect
roughness.
RECOMMENDATIONS
(5)
This study gives an initial look at the performance of overlay rough-
ness for Canadian LTPP test sites. Further study would be benefi-
cial in the following areas:
1. Continued study for the remaining study period;
2. Creation of smaller subcategories to better describe the test
section properties;
(6) 3. Investigation of the potential for other performance factors
(overlay type, initial overlay IRI, traffic level, etc.) to influence
roughness; and
4. Development of a single equation to explain roughness
progression that accounts for all the performance factors.
REFERENCES
1. FHWA. DataPave Online, U.S. Department of Transportation. www.
datapave.com. December 20, 2002.
2. Sayers, M. W., T. D. Gillespie, and W. D. Patterson. Guidelines for the
Conduct and Calibration of Road Roughness Measurements. World Bank
Technical Paper 46, Washington, D.C., 1986.
3. Transportation Association of Canada. Pavement Design and Management
Guide. Ottawa, Ontario, 1997.

2
1.75
1.5
1.25
IRI (m/km)

1
0.75
0.5
0.25
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Overlay Age (Years)
LTPP C-SHRP LTPP-Regression Eqn

FIGURE 10 Roughness progression of thick overlays in wet–freeze climatic zones with fine-
grained subgrades (Eqn  equation).
 Smith and Tighe
4. Haas, R., W. R. Hudson, and J. Zaniewski. Modern Pavement Manage-
ment. Krieger Publishing Company, Malabar, Fla., 1994.
5. Hudson, W. R., R. Haas, and W. Uddin. Infrastructure Management.
McGraw-Hill, New York, 1997.
6. Moore, D., and G. McCabe. Introduction to the Practice of Statistics,
2nd ed. W. H. Freeman, New York, 1996.
7. Craig, R. F. Soil Mechanics, 6th ed. E & FN Spon, New York, 1997.
8. Tighe, S., R. Haas, and N. Li. Overlay Performance in Canadian Strate-
gic Highway Research Program’s Long-Term Pavement Performance
View publication stats
135
Study. In Transportation Research Record: Journal of the Transporta-
tion Research Board, No. 1778, TRB, National Research Council,
Washington, D.C., 2001, pp. 191–200.
9. Haas, R., N. Li, and S. Tighe. Roughness Trends at C-SHRP LTPP Sites.
Canadian Strategic Highway Research Program (C-SHRP), March 1999.
Publication of this paper sponsored by Surface Properties–Vehicle Interaction
Committee.