ArticlePDF Available

Increased costs to US pavement infrastructure from future temperature rise

Article

Increased costs to US pavement infrastructure from future temperature rise

Abstract and Figures

Roadway design aims to maximize functionality, safety, and longevity. The materials used for construction, however, are often selected on the assumption of a stationary climate. Anthropogenic climate change may therefore result in rapid infrastructure failure and, consequently, increased maintenance costs, particularly for paved roads where temperature is a key determinant for material selection. Here, we examine the economic costs of projected temperature changes on asphalt roads across the contiguous United States using an ensemble of 19 global climate models forced with RCP 4.5 and 8.5 scenarios. Over the past 20 years, stationary assumptions have resulted in incorrect material selection for 35% of 799 observed locations. With warming temperatures, maintaining the standard practice for material selection is estimated to add approximately US$13.6, US$19.0 and US$21.8 billion to pavement costs by 2010, 2040 and 2070 under RCP4.5, respectively, increasing to US$14.5, US$26.3 and US$35.8 for RCP8.5. These costs will disproportionately affect local municipalities that have fewer resources to mitigate impacts. Failing to update engineering standards of practice in light of climate change therefore significantly threatens pavement infrastructure in the United States. © 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Content may be subject to copyright.
LETTERS
PUBLISHED ONLINE: 18 SEPTEMBER 2017 | DOI: 10.1038/NCLIMATE3390
Increased costs to US pavement infrastructure
from future temperature rise
B. Shane Underwood1*, Zack Guido2, Padmini Gudipudi1and Yarden Feinberg1
Roadway design aims to maximize functionality, safety, and
longevity1,2. The materials used for construction, however, are
often selected on the assumption of a stationary climate1,3.
Anthropogenic climate change may therefore result in rapid
infrastructure failure and, consequently, increased mainte-
nance costs, particularly for paved roads where temperature is
a key determinant for material selection. Here, we examine the
economic costs of projected temperature changes on asphalt
roads across the contiguous United States using an ensemble
of 19 global climate models forced with RCP 4.5 and 8.5
scenarios. Over the past 20 years, stationary assumptions
have resulted in incorrect material selection for 35% of
799 observed locations. With warming temperatures, main-
taining the standard practice for material selection is estimated
to add approximately US$13.6, US$19.0 and US$21.8 billion
to pavement costs by 2010, 2040 and 2070 under RCP4.5,
respectively, increasing to US$14.5, US$26.3 and US$35.8
for RCP8.5. These costs will disproportionately aect local
municipalities that have fewer resources to mitigate impacts.
Failing to update engineering standards of practice in light
of climate change therefore significantly threatens pavement
infrastructure in the United States.
Climate change may have widespread impacts on road, water,
rail, and air system4,5. These impacts will result from intense
precipitation, heat/cold stress, and other non-physical challenges
that degrade infrastructure quality and longevity1,2,4,6–9. Because
these transportation systems constitute large civil investments
(US$7.7 trillion in assets and US$45 billion annual expenditures10)
and underpin an economic vibrancy (five trillion kilometres of pub-
lic travel per year11 and private citizen expenditures equal to 8.9% of
GDP10), the impacts may be substantial. Transportation infrastruc-
ture is built to last decades, but engineering protocols in the United
States assume climate stationarity, which may result in accelerated
degradation and, consequently, increased costs. Additional costs are
a concern since the American Society of Civil Engineers estimates
that infrastructure needs US$3.6 trillion in the next decade, with
a large fraction of that currently unfunded12. At present, engineers
assume a stationary climate when selecting pavement materials,
meaning that they may be embedding an inherent negative per-
formance bias in pavements for decades to come. With warming
trends observed and accelerating across the United States13 , and with
scientific consensus for future warming, continued use of such data
is likely to expose some areas to more rapid degradation14–17.
The current standard of roadway design guides engineers to
use climate data from 1964–1995 to select materials. We quan-
tify the economic effects of the continued use of this climate
record by examining the impacts of non-stationarity on the asphalt
grade in asphalt pavements (constitute 90% of paved surfaces
in the United States and 99% worldwide18 ). This analysis makes
important advancements from previous studies7,19–28. First, we carry
out the economic analysis nationally, regionally, and on a state-
by-state basis after incorporating local engineering practices in
material selection. We base this analysis on an ensemble of global
climate models (GCMs) to probabilistically examine the costs
from greater maintenance and reduced pavement life, while also
accounting for the distinct types of roadways. We show results
regionally and by state and over several decades to highlight their
geospatial–temporal nature and to provide insights on the infras-
tructure impacts to decision makers. More details from the literature
and the specific advancements made by this study are given in
Supplementary Section 1.
In the United States, asphalts are used per the Superpave Perfor-
mance Grading (PG) system, which assigns a temperature-related
grade based on the maximum and minimum temperatures between
which that asphalt should exhibit adequate performance3. A typical
grade might be PG64-22, which means that the asphalt is Perfor-
mance Graded for temperatures between 64C and 22 C. Since
the asphalt grade is based on pavement temperature and linked
to pavement performance, it serves as a direct indication of how
climate impacts pavement performance. Determining the required
low and high temperatures for any location involves calculating the
average and standard deviation of the minimum pavement temper-
ature and the maximum seven-consecutive-day pavement temper-
ature over a multi-decade period. In practice, the climate record
used for this purpose is 1966–1995 and the averages are statistically
adjusted to account for extremely cold or hot years and rounded
to standard 6 C grade increments. In pavement engineering, other
climate records may be used for structural design, but even in these
cases the 1966–1995 climate record is the one used to select the
materials for the design analysis. Thus, adherence to this record does
have a substantial impact on the design and long-term behaviour of
pavements even though other records may be used for part of the
design process.
By using the stationary climate record, we find that asphalt
grades are already being improperly determined in many parts of
the United States Figure 1 shows locations from the United States
Historical Climate Network where the required asphalt grade based
on temperature data from 1966–1995 differs from the one based on
data from 1985–2014. In total, 35% of stations have a different high-
or low-temperature grade (6% high-temperature only, 26% low-
temperature only, and 3% both high- and low-temperature). High-
temperature grade changes are the primary performance concern
since these sites will experience faster degradation, require greater
maintenance, and possibly lead to earlier reconstruction. Underes-
timates of the low temperature value suggest that the location has
additional protection against low-temperature cracking, but implies
that agencies are paying higher costs for materials that withstand
lower temperatures than currently exist.
1Arizona State University, School of Sustainable Engineering and the Built Environment, PO Box 875306, Tempe, Arizona 85287-5306, USA. 2Institute of
the Environment, University of Arizona, Tucson, Arizona 85721, USA. *e-mail: shane.underwood@ncsu.edu,Shane.Underwood@asu.edu
NATURE CLIMATE CHANGE | ADVANCE ONLINE PUBLICATION | www.nature.com/natureclimatechange 1
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
LETTERS NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE3390
No change temperature grade
Change in low temperature grade only
Change in high temperature grade only
Change in both high and low temperature grade
Figure 1 | Weather stations evaluated to compare 1966–1995 climate
database and 1985–2014 climate databases. Locations where the asphalt
grade diers between the two data sets are coloured in red, blue, and black.
The red symbols show locations where the 1966–1995 data underestimate
the high-temperature grade required, whereas the blue symbols show those
locations where the low temperatures were underestimated (estimated to
be too low). The black symbols show both an overestimate and
underestimate of the high- and low-temperature grades respectively.
Locations with no change are left unfilled.
We examine future impacts by comparing the asphalt grades
based on the 1966–1995 database to those determined by
temperatures under future climate scenarios. The future scenarios
are evaluated with a multi-model ensemble of one model run
for each of 19 GCMs (see Supplementary Table 1) and for
Representative Concentration Pathways (RCP) 8.5 and 4.5. Figure 2
maps the geospatial changes in the future pavement temperature
for different periods of time based on the median of the model
ensemble. We present results for RCP 8.5 here and RCP 4.5 in
Supplementary Fig. 1. In both scenarios, there is an increasing trend
in the pavement temperature across the United States. We show
the effect of these temperatures on the required standard asphalt
grade in Supplementary Fig. 2. The locations with grade changes
generally reflect what is shown in Fig. 2. Locations where greater
temperature changes are projected will need two standard asphalt
temperature grade increments and those with less temperature
increase will need one standard grade increment.
There are three factors that produce the unique high-temperature
change in Fig. 2 and Supplementary Fig. 1. First, is the use of
the hottest seven-consecutive day temperature (instead of mean
air temperature or highest temperature). Like what is reported
elsewhere for heat waves in the United States29 , the model ensemble
predicts this increase will be the greatest in the upper Midwest,
extending through the Ohio Valley. Second, the variation in
maximum seven-consecutive day temperature from one year to the
next is expected to increase. This variability is already greater in the
upper Midwest and Ohio Valley regions, and the model ensemble
predicts the variation to continue increasing. Third, the relationship
between air temperature and pavement temperature is affected by
sun position, with southern latitudes receiving more direct sunlight.
Under future scenarios the Ohio Valley and Southeast regions
of the country are expected to experience the greatest temperature
increases. The Southwest and Pacific coastal regions show relatively
little change in the high-temperature ensemble median. However,
the low-temperature median is projected to increase, particularly
along the Rocky Mountains. The model-to-model variation is
high, but the trends are consistent across the model ensemble
(Supplementary Section 3).
Changing pavement engineering policy and practice is slow
and, in the absence of adaptations, more frequent rehabilitation
and maintenance will be required. We estimate that pavements
originally intended to last 20 years will require rehabilitation after
only 16–17 years when the pavement grade is wrong by one 6 C
increment and 14–16 years when the pavement grade is wrong by
two 6 C increments (see Methods and Supplementary Section 4).
These performance losses will bear a cost, which we estimate using
life cycle cost analysis (LCCA).
Table 1 summarizes the net present value of all activities con-
sidered in the life cycle of the pavements when the correct asphalt
grade is used (baseline case) and when the asphalt grade is wrong by
one or two standard grades. Greater costs occur on the higher-traffic
roadways, but a greater impact as a percentage of the cost exists with
the lower-traffic roadways because construction represents a smaller
proportion of the overall life cycle impact for lower-traffic roadways.
If the initial construction costs are not included in the cost estimate,
also shown in Table 1, then the percentage differences in costs are
greater for the higher-traffic roadways. Since lower-traffic roadways
are generally the responsibility of municipal agencies, it is likely that
city and county road agencies will see a disproportionate economic
impact to their road network. These agencies often work with more
constrained budgets and fewer options to raise revenues for repairs
and reconstruction, which further exacerbates the impacts.
When applied to the entire pavement network in the United
States we find that the implications of these costs are large. Pro-
jections are made for each of the models and from the median
of the model ensemble for the sequential 30-year windows (2010,
2040 and 2070) in the data set. The differences between these
cases and the baseline scenario that assumes a stationarity climate
represents our estimated impact from failing to adapt engineering
practices to climate changes. The cost impacts are shown in Fig. 3.
The estimated costs across the United States based on RCP 4.5
are US$13.6, US$19.0 and US$21.8 billion in 2010, 2040 and 2070,
respectively. Cost estimates for the same periods based on RCP 8.5
are US$14.5, US$26.3 and US$35.8 billion, respectively. The varia-
tion in these costs are also large (as low as US$8.8 billion in 2010
to as high as US$45.5 billion in 2070) owing to the variability in
model outcomes. To place the calculated impacts into perspective,
the cumulative baseline costs for the United States are approximately
US$419 billion. Thus, the impacts from temperature increases add
approximately 3–9% to the cost to build and maintain the infras-
tructure over each 30-year period. A more comprehensive discus-
sion of these variations is given in Supplementary Section 5 and a
comparison to other similar estimates in the literature is given in
Supplementary Section 6.
Since maintenance and rehabilitation is the responsibility of each
individual state, we also cumulate the data by each state individually
and by region. Regionally, the Ohio Valley and Southeast are
projected to be the most affected. Differences also exist with respect
to the range of costs projected within a given scenario as well as
the sensitivity to the future scenario. The Southeast, for example,
is projected to have nearly identical results for both RCP 4.5 and
RCP 8.5 scenarios, while the Upper Midwest, South, and Northeast
show larger differences with respect to the future scenario (see
Supplementary Section 5). Another example, Delaware, has a large
range in costs with the RCP 4.5 ensemble, while the projected
impacts from the RCP 8.5 ensemble are almost identical for all
models. At the state level, Texas, California, Illinois and Florida will
experience the highest impacts (Supplementary Fig. 5). These states
maintain and manage large transportation networks so the results
are not entirely surprising. Small states generally have less total cost
impact than larger states, but on a per-kilometre basis smaller states
such as Maryland and Delaware show effects equal to or greater than
larger states (Supplementary Fig. 6).
Our LCCA accounts for only a fraction of the total number
of paved miles of the residential road network because details for
these networks are not readily accessible in national data sources.
However, the City of Phoenix, Arizona, is one network for which
details are known. The city’s network has more than 5,600 km
of residential streets. By comparison, the cumulative interstate,
2
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE CLIMATE CHANGE | ADVANCE ONLINE PUBLICATION | www.nature.com/natureclimatechange
NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE3390 LETTERS
Change in high temperature (°C) Change in low temperature (°C)
0−1 2−3 4−5 6−7 8−9 0−2 3−4 5−6 7−8 9−10 11−12 13−14 15−16
ad
be
cf
Figure 2 | Expected median increases in pavement temperature based on the RCP 8.5 ensemble. af, Average 7-day maximum temperature (ac) and
average minimum temperature changes (df) for 2010–2039, 2040–2069 and 2070–2099 respectively relative to the 1966–1995 climatology. Darker
coloured contours indicate locations where the pavement temperatures will exceed current stationarity implied pavement temperatures by a greater level
and thus experience greater performance impacts.
Table 1 | Impacts by pavement type when using the correct and incorrect asphalt grade (percentage dierence from using the correct
asphalt grade shown in brackets).
Roadway type Correct asphalt grade Asphalt grade incorrect by one
increment
Asphalt grade incorrect by two
increments
Net present cost (US$ km1)
Interstate 1,183,702 1,270,095 (6.80%) 1,312,235 (9.80%)
National route 723,106 775,997 (6.80%) 807,514 (10.5%)
State route 403,589 444,591 (9.20%) 472,737 (14.6%)
Local road 231,742 257,804 (10.1%) 280,576 (17.4%)
Net present cost (not including the initial construction cost) (US$ km1)
Interstate 199,240 285,632 (43.4%) 327,773 (64.5%)
National route 132,429 185,319 (39.9%) 216,837 (63.7%)
State route 108,251 149,252 (37.9%) 177,398 (63.9%)
Local road 84,072 110,135 (31.0%) 132,906 (58.1%)
Bottom four rows of table show the cost impacts if the initial construction costs arenot included in the NPV (note actual impacts carry slightly higher precision than shown in the table so round-o
dierence may be present).
national route, and state route system in the entire state of Arizona is
9,900 km. These 9,900 km of roads are maintained and rehabilitated
with an annual budget of approximately US$1.1 billion, whereas
the City of Phoenix maintains its network on an annual budget of
approximately US$57million. Even when accounting for the greater
number of lanes in the state system, the State has about 8.5 times
more financing per lane km than Phoenix. At the same time, the
projected impact from climate change in Phoenix is substantial.
The median RCP 8.5 scenario suggests a total cost impact for the
period 2070–2100 of approximately US$0.15 billion in Phoenix
alone. During this same period, the projected impact for the entire
state of Arizona is US$0.53 billion.
We have detailed information on the length of residential streets
in Phoenix, but do not have data on residential streets nationwide.
To identify states that may experience greater effects from the
residential networks we use the relative proportion of local miles
to other roadway types as a surrogate measure of the relative
extent of the residential street network. Using this measure (see
Supplementary Fig. 7) suggests that some states (California in
particular) may be facing a substantially higher potential impact
from what is estimated here (see Supplementary Section 5).
In terms of both performance and cost, a failure to update
engineering standards of practice and adapt to climate change
may leave the pavement infrastructure in the United States at risk.
Based on the analysis here, we expect that the impacts will be
greatest in geographically larger states, central and southeastern
regions and local municipalities. The results of this analysis and
comparisons between it and projections of mean air temperature
rise across the United States29,30 show that the impacts of climate
change induced temperature rise cannot be uniquely related to the
NATURE CLIMATE CHANGE | ADVANCE ONLINE PUBLICATION | www.nature.com/natureclimatechange
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
3
LETTERS NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE3390
10
20
30
40
50
2010 2040 2070
Year
2010 2040 2070
Cost impact (US$, billions)
RCP 4.5 RCP 8.5
Figure 3 | National cost impact from failing to adapt asphalt grade. Range
of costs vary by year and RCP scenario considered. The projected costs are
similar by RCP for the 2010–2040 period, but increases substantially by
2070–2100 period. The boxed areas enclose the 75th and 25th percentile
range from the model ensemble. The horizontal line in these boxes is the
median and the error bars show the maximum and minimum costs from the
models in the ensemble.
absolute value of air temperature or the change of this temperature
in the future. The key contributors to this risk are: the increase
local in air temperature and year-to-year variation; the geospatial
location (notably the latitude), the current engineering practices
of the pavements inclusive of the current reliability of the asphalt
grade; and the density of the road network across roadway types.
This study highlights that in given the temporal scale with which
roadways are engineered to perform, in the future, it may be
important that engineering practice incorporate up-to-date climate
records and/or incorporate future climate projections to mitigate
economic impact.
Methods
Methods, including statements of data availability and any
associated accession codes and references, are available in the
online version of this paper.
Received 8 April 2017; accepted 17 August 2017;
published online 18 September 2017
References
1. Huang, Y. H. Pavement Analysis and Design (Prentice Hall, 1993).
2. Yoder, E. J. & Witczak, M. W. Principles of Pavement Design (John Wiley, 1975).
3. Asphalt Institute Superpave Performance Graded Asphalt Binder Specifications
and Testing Superpave Series No. 1 (SP-1) (Asphalt Institute, 2003).
4. National Research Council Committee on Climate Change (NRC) Potential
Impacts of Climate Change on US Transportation (National Research
Council, 2008).
5. Meyer, M. et al.Strategic Issues Facing Transportation, Volume 2: Climate
Change, Extreme Weather Events, and the Highway System: Practioner’s Guide
and Research Report NCHRP Report 750 (National Cooperative Highway
Research Program, 2014).
6. Anderson, T., Beck, C., Gade, K. & Olmsted, S. Extreme Weather Vulnerability
Assessment FHWA No. 0704-0188.FHWA (Arizona Department of
Transportation, 2015).
7. Cambridge Systematics Central Texas Extreme Weather and Climate Change
Vulnerability Assessment of Regional Transportation Infrastructure
FHWA No. 0704-0188 (Capital Area Metropolitan Planning
Organization, 2015).
8. Koetse, M. J. & Rietveld, P. The impact of climate change and weather on
transport: an overview of empirical findings. Transp. Res. 14, 205–221 (2009).
9. IPCC Climate Change and Cities: First Assessment Report of the Urban Climate
Change Research Network (eds Rosenzweig, C., Solecki, W. D., Hammer, S. A.
& Mehrotra, S.) (Cambridge Univ. Press, 2011).
10. United States Department of Transportation (USDOT) Transportation in the
United States (Bureau of Transportation Statistics, 2015).
11. United States Department of Transportation (USDOT) US Beyond Traffic 2045:
Trends and Choices (US Department of Transportation, 2015).
12. American Society of Civil Engineers (ASCE) 2013 Report Card for Americas
Infrastructure (2013); http://www.infrastructurereportcard.org
13. IPCC Climate Change Impacts in the United States: The Third National Climate
Assessment (eds Melillo, J. M., Terese, R. & Yohe, G. W.) 841 (US Global
Change Research Program, 2014).
14. Knutti, R. & Sedlacek, J. Robustness and uncertainties in the new CMIP5
climate model projections. Nat. Clim. Change 3, 369–373 (2013).
15. Mearns, L. et al. The North American regional climate change assessment
program: overview of phase I results. Bull. Am. Meteorol. Soc. 93,
1337–1362 (2012).
16. Woldemeskel, F. M., Sharma, A., Sivakumar, B. & Mehrotra, R. Quantification
of precipitation and temperature uncertainties simulated by CMIP3 and
CMIP5 models. J. Geophys. Res. 121, 3–17 (2016).
17. Wuebbles, D. et al. CMIP5 climate model analysis: climate extremes in the
United States. Bull. Am. Meteorol. Soc. 95, 571–583 (2014).
18. United States Department of Transportation (USDOT) National Transportation
Statistics (Bureau of Transportation Statistics, accessed March 2016);
https://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/publications/
national_transportation_statistics/index.html
19. Chinowsky, P. & Arndt, C. Climate change and roads: a dynamic
stressor-response model. Rev. Dev. Econ. 16, 448–462 (2012).
20. Chinowsky, P. S., Price, J. C. & Neumann, J. E. Assessment of climate change
adaptation costs for the US road network. Glob. Environ. Change 23,
764–773 (2013).
21. Daniel, J. S., Jacobs, J. M., Douglas, E., Mallick, R. B. & Hayhoe, K. Impact of
climate change on pavement performance: preliminary lessons learned
through the infrastructure and climate network (ICNet). Int. Symp. Climatic
Effects Pavement Geotechnical Infrastructure (Fairbanks AK, 2013).
22. Fletcher, C. G., Matthews, L., Andrey, J. & Saunders, A. Projected changes in
mid-twenty-first-century extreme maximum pavement temperature in
Canada. J. Appl. Meteorol. Climatol. 55, 961–974 (2016).
23. Harvey, M. et al.Impact of Climate Change on Road Infrastructure
Report No. AP-R243/04 (Austroads and the Bureau of Transport and Regional
Economics, 2004).
24. Meagher, W., Daniel, J. S., Jacobs, J. & Linder, E. Method for evaluating
implications of climate change for design and performance of flexible
pavements. Trans. Res. Record J. Trans. Board 2305, 111–120 (2012).
25. Mills, B. N., Tighe, S. L., Andrey, J., Smith, J. T. & Huen, K. Climate change
implications for flexible pavement design and performance in Southern
Canada. J. Trans. Eng. 135, 773–782 (2009).
26. Mndawe, M. B., Ndambuki, J. & Kupolati, W. K. Revision of the macro
climatic regions of Southern Africa. OIDA Int. J. Sustainable Dev. 6,
37–44 (2013).
27. Schweikert, A., Chinowsky, P., Espinet, S. & Tarbert, M. Climate change and
infrastructure impacts: comparing the impact on roads in ten countries
through 2100. Procedia Eng. 78, 306–316 (2014).
28. Viola, F. & Celauro, C. Effect of climate change on asphalt binder selection for
road construction in Italy. Transp. Res. 37, 40–47 (2015).
29. Kunkel, K. E., Liang, X. Z. & Zhu, J. Regional climate model projections and
uncertainties of US summer heat waves. J. Clim. 23, 4447–4458 (2010).
30. Maurer, E. P., Brekke, L., Pruitt, T. & Duffy, P. B. Fine-resolution climate
projections enhance regional climate change impact studies. Eos Trans. Am.
Geophys. Union 88, 504 (2007).
Acknowledgements
We acknowledge the World Climate Research Program’s Working Group on Coupled
Modeling, which is responsible for CMIP, and we thank the climate modelling groups
(listed in Supplementary Table 1) for producing and making available their model output.
For CMIP the US Department of Energy’s Program for Climate Model Diagnosis and
Intercomparison provides coordinating support and led development of software
infrastructure in partnership with the Global Organization for Earth System Science
Portals. We would also like to acknowledge the Climate Assessment for the Southwest
(CLIMAS) at the University of Arizona for providing support to Z. Guido.
Author contributions
B.S.U. designed the study, identified the data sources, created the scripts to analyse the
climate data, and developed the structure of the paper in collaboration with Z.G.
and P.G.; Z.G. provided inputs on climate modelling and ensemble interpretation and
review of the manuscript; P.G. reviewed the manuscript and discussed interpretation of
the data at length; Y.F. assisted in downloading, cataloguing, and running the climate
scripts. All authors contributed equally to developing the ideas in this paper.
Additional information
Supplementary information is available in the online version of the paper. Reprints and
permissions information is available online at www.nature.com/reprints. Publisher’s note:
Springer Nature remains neutral with regard to jurisdictional claims in published maps
and institutional affiliations. Correspondence and requests for materials should be
addressed to B.S.U.
Competing financial interests
The authors declare no competing financial interests.
4
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE CLIMATE CHANGE | ADVANCE ONLINE PUBLICATION | www.nature.com/natureclimatechange
NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE3390 LETTERS
Methods
Climate data. Two climate databases were used in this study: the United States
Historical Climatology Network (USHCN); and a global climate model (GCM)
ensemble of 19 climate models each under Representative Concentration Pathway
(RCP) 8.5 and 4.5. In both cases we use the 1966–1995 climate record as the
comparative reference because this is the current basis of binder selection in the
United States. The objective in this analysis is to quantify the impacts from
continuing to adhere to a static database, which means comparing future year
effects to the current state of the practice and the condition that will exist if
engineers continue to adhere to this practice.
USHCN Data Processing We accessed the USHCN database through the US
Department of Energy portal and downloaded the daily maximum and minimum
temperatures (http://cdiac.ornl.gov/ftp/ushcn_daily). We choose this database to
determine the impact of present day temperatures on PG because, although there
are fewer weather stations in the database than others, they are quality controlled
so that each station has minimal missing data and data records are available
covering the period time of interest. Only those weather stations with complete
daily temperature data from January 1, 1966 through December 31, 2014 are
considered. In total, 799 weather stations met this criterion, and their locations are
shown in Fig. 1 of the manuscript. For each station, the data for the years
1966–1995 as well as the most recent 30-year period available (1985–2014) are
extracted from the files using custom MATLAB scripts.
GCM data processing. We selected GCMs for the ensemble from those models
that participated in the Coupled Model Intercomparison Project 5 (CMIP5), had
daily maximum and minimum temperature data for RCP 8.5 and RCP 4.5
scenarios, and were available in 1/8resolution29,31,32. The data were downloaded
from the archives of the Climate Analytics Group
(ftp://gdo-dcp.ucllnl.org/pub/dcp/archive/cmip5/bcca). For analysis, projections
are grouped by 30-year periods. These periods begin in the first year of a decade
(2010, 2040 and 2070) and are staged in 30-year increments. Results presented as
‘2010’ are based on the temperature projections for the period from January 1, 2010
to December 31, 2039; data given as ‘2040’ are based on data for 2040–2069, and so
forth. For statistical analysis purposes, and to most easily compare current and
future scenarios, the downscaled data is geospatially interpolated to the coordinates
of the Superpave weather stations33. This extraction and interpolation is performed
using custom MATLAB scripts.
Superpave method of asphalt cement specification. To determine the asphalt
grade, we follow the standard performance grade (PG) method. Equations (1)
and (2) are then used to estimate the minimum pavement and
seven-consecutive-day average maximum pavement temperatures, respectively33 .
These temperatures are used in the Superpave method because they relate to either
thermal cracking (low temperature) or rutting (high temperature).
Tpav,low =7.191 +0.72(Tair,low z×σair,low)0.004 Lat2
z(4.4 +0.52air,low)2)0.5 (1)
Tpav,high =((Tair,high +z×σair,high)0.00618Lat2+0.2289 Lat
+42.2)(0.9545)17.78 (2)
where Tair,low =minimum average air temperature (C), Lat =latitude (decimal
degrees), z=standard normal deviate (50% reliability z=0, 98% reliability
z=2.055), σair,low =standard deviation of minimum air temperature (C), σair,high =
standard deviation of seven-consecutive day high temperature (C), Tpav,low =
pavement low temperature (C), and Tair,high =seven-consecutive-day high
temperature (C). Latitude and longitude for GCM ensemble are the same as the
Superpave weather stations and for the USHCN database they are extracted
directly from the datafile. The daily maximum and minimum air temperatures at
each location are extracted from the downloaded databases, arranged by year, and
processed to determine the minimum air temperature and the highest
seven-consecutive-day average maximum air temperature for each year in the
record. Then, the average and standard deviation for these annual values
are calculated.
The process accounts for exceptionally hot summers and cold winters by
embedding statistical uncertainty into equations (1) and (2). This is conventionally
termed reliability, and is mathematically defined as the probability (expressed as a
percentage) that the temperatures will not be exceeded in any given year. When the
average of the annual air temperatures is used in these calculations, there is a 50%
probability that a given year will exceed the average, and thus grades that are based
on the averages are referred to as the asphalt grade at 50% reliability. Generally,
engineers consider it to be too risky to use this grade, and by convention choose a
temperature that yields a 98% reliability.
The final step adjusts the calculated 50% or 98% reliability pavement
temperatures to standard, six-degree temperature increments. For the
high-temperature grade, these are: 82, 76, 70, 64, 58, 52 and 46 C and for the low
temperature they are 46, 40, 34, 28, 22, 16 and 10 C. This rounding
process increases the true reliability of the given asphalt grade, but by convention it
is still referred to as either the 50% or 98% reliability grade. Equations (1) and (2)
can be rearranged to determine the true low temperature (RLT) and high
temperature (RHT) reliability of a selected grade, equations (3) and (4).
RLT =0.5"1+erf 7.191 LT 0.004Lat2+0.72Tair,low
(0.72σair +(4.4 +0.52air)2)0.5 )2!#×100 (3)
RHT =0.5 "1+erf HT+17.78
0.9545 +0.00618Lat20.2289Lat 42.2 Tair,high
σair2!#
×100 (4)
where LT and HT =low- and high-temperature grades against which the reliability
is evaluated.
For the USHCN database we use equations (1) and (2) to calculate the standard
grade from both the 1966–1995 and 1985–2014 temperature record. We compare
these results to identify the stations where the two databases yield different grades.
For each model in the GCM ensemble and for each time window in the study
(2010, 2040 and 2070), we apply the method above to calculate the projected 98%
pavement temperature for each time period and for each location. We then
calculate the difference between these temperatures when using the 1966–1995
climate record and from the median of the GCM ensemble. These temperature
differences are graphically depicted for the RCP 8.5 scenario in the manuscript
(Fig. 2) and for RCP 4.5 scenario in Supplementary Fig. 1. The outcomes from
equations (1) and (2) are also used to compute the standard grade (that is, the grade
in the standard temperature increments). This grade is compared with the
currently specified grade and the results are shown in Supplementary Fig. 2.
Finally, we substitute the averages and standard deviations for the high and low
temperatures into equations (3) and (4) along with the current asphalt grade to
estimate the future true reliability. We present statistical analysis of this true
reliability in Supplementary Fig. 4.
Uncertainty analysis of projected impacts. We evaluate the impacts by state,
region (defined using the National Climatic Data Center regions, Supplementary
Fig. 3), and nation. The variability of the climate models is examined across regions
by using the beta function, Equation (5). Characterization of this function reveals
that the ensembles result in a skewed distribution of impacts, which we use to
justify selecting the median as the central tendency measure of the ensemble. The
median value is estimated by finding the parameters of the beta function (α,β,a,
and c) using the Pearson method34. Detailed descriptions of the findings by region
are presented in the manuscript and in Supplementary Section 3.
P=0(α +β )
0(α)0 (β) xa
caα11xa
caβ1
×100 (5)
where P=cumulative probability of given true reliability, x,αand β=beta
distribution parameters, aand c=the maximum and minimum values of the
distribution function, and 0() is the gamma function.
Impact assessment. To estimate the cost impacts, we perform four steps: we use
the nationally calibrated Pavement Design ME model to calculate performance
reductions resulting from the temperature-induced shortfalls in grades; we
estimate changes to the construction, maintenance, and rehabilitation activities
brought on by reduced performance and the life cycle costs of these scenarios; we
coordinate roadway lengths to the nearest corresponding weather station; finally,
we calculate the increased costs using life cycle cost analysis (LCCA).
Pavement ME design model. The American Association of State Highway and
Transportation Officials (AASHTO) Pavement Design ME simulation tool is used
to estimate the performance impacts from using an incorrect asphalt grade. We
elected to use Pavement Design ME because it is the only pavement analysis and
design tool that has undergone extensive national calibration and one of the only
that can directly consider asphalt grade in the performance prediction process. This
tool explicitly considers the individual and interactive effects of local temperature,
traffic, material (including the PG grade used), soil conditions, and the pavement
structural configuration of the roadway types. To integrate these factors, Pavement
Design ME uses a complex assemblage of models and routines that link the
response of a pavement under trafficking to cracking, rutting, and ride quality
changes over the lifetime of the pavement structure (approximately 20 years).
NATURE CLIMATE CHANGE |www.nature.com/natureclimatechange
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
LETTERS NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE3390
We simulate the performance of four roadway types (interstate,
national route, state route, and local roads) in cold, moderate, and warm climate
zones (Minneapolis, Minnesota; Raleigh, North Carolina; and Miami, Florida). The
pavements are simulated so that the correct asphalt has a service life of 18–22 years.
The relevant inputs for each of the simulations are given in Supplementary Table 3.
For variables not explicitly given in this table (asphalt content, air void content,
thermodynamic properties, and so on) the default parameters of the simulation tool
are used. The climate records for each of these cities are available within the support
files for the simulation tool and the soil properties are found using the soil survey
tool available at http://nchrp923b.lab.asu.edu (ref. 35). For climate and roadway
type, the simulations are first performed by inputting the correct asphalt grade
for the current climate. Then subsequent simulations are carried out with a grade
that is either one or two standard increments below the current grade. We elected
to follow this approach and use the predicted performance under current climate
and current asphalt grade as the reference condition for performance estimations
because of inconsistencies in using the GCM predicted climate input in the
Pavement Design ME model. The pavement model uses hourly temperature values
whereas the GCMs provide only daily maximum and minimum temperatures.
Projecting the future hourly temperature variations and analysing pavement
performance under future climate grades would introduce additional uncertainty.
The basic prediction process followed in Pavement Design ME to predict
fatigue cracking and rutting are given in Supplementary Section 4, and even more
detailed descriptions of the models can be found elsewhere36–42. For our analysis,
we focused on the cracking and rutting distresses because the predicted values have
greater certainty than others and because we were most interested in the structural
performance of the pavements. Thermal cracking is also predicted from the
Pavement Design ME model and is a distress that is directly related to climate.
However, we do not consider thermal cracking effects in our analysis because, as
the data in Fig. 2 in the manuscript and Supplementary Fig. 1 demonstrate, future
climate change suggests a warming of the yearly cold temperature, which would
result in less future thermal cracking. Other relevant performance measures do
exist (ravelling and pothole formation for example), and we recognize that future
climate changes will affect the mechanisms that cause these distresses. For example,
higher temperatures will result in faster oxidation of the asphalt, which can
embrittle the material and make it more likely to ravel. However, the science
describing the mechanics of these distresses has not produced comprehensive
mechanistically based models capable of reliably predicting the initiation and
growth of these distresses. Empirical models do exist, but these may combine
multiple confounding factors into single variables (for example, a temperature
variable in the empirical model may implicitly assume an asphalt type that is
associated with that temperature), which makes it difficult to consider the effects of
asphalt changes on the resulting performance.
For the structural inputs to the model we recognize that pavement design
methodologies can vary substantially between and within states. Even more,
agencies do not always keep accurate records of the in-place designs or the
standards that they follow. Although the specific designs for every roadway are
difficult to identify, most pavement designs in the United States use a common
paradigm: asphalt concrete is placed on a supporting layer of unbound and
compacted granular base, which then rests on compacted native soil. The thickness
of the pavement layers is a function of loading severity (both in terms of actual load
levels and the number of repetitions). Loading severity strongly correlates with the
roadway types: interstate (most severe loading), national route, state route, or local
road (least severe loading). In this analysis, we consider this effect by creating four
different representative pavement structures for the simulations. The thickness of
these representative pavement types varies as shown in Supplementary Table 3,
with the thinner asphalt concrete pavements used for the roads with fewer trucks.
We assume that structural failure of a pavement occurs when either the fatigue
cracking reaches 20% of the total lane area36 or total pavement rutting is equal to
12.5 mm (ref. 43). We estimate the loss of performance from an incorrect asphalt by
comparing the time to failure (to the half year) when the correct asphalt grade is
used to the time to failure when an incorrect asphalt grade is used, equation (6).
PL =tfailurebaseline tfailure1grade
tfailurebaseline ×100 (6)
Where, PL =performance loss, tfailurebaseline =years to failure when the correct
asphalt grade is used, tfailure1grade =years to failure when the asphalt grade used is
wrong by either one or two increments.
Maintenance and rehabilitation planning. To conduct the LCCA we develop a
30-year life cycle plan involving the timing of maintenance and rehabilitation
activities. We start by developing a schedule for the case where the correct asphalt is
used (the reference schedule). The basis of this schedule is our own engineering
experience, the guidelines for Indiana and New York44,45, and national guidance46,47 .
We then develop schedules for the case where the asphalt is wrong by one or two
grades by modifying the timing of the individual activities in the reference schedule
by the same amount as the performance loss. For example, we estimated that the PL
of an interstate pavement with an incorrect asphalt grade was 10%. In the reference
schedule, a major rehabilitation will occur at year 16, but based on the 10% loss in
performance this activity is now expected to occur in year 14.4 (16 ×0.9 =14.4),
which we round in our final adjusted schedule so that it takes place in year 14.
Explicit justification for using this linear scaling method is given in Supplementary
Section 4. PLs change by roadway type and grade deficiency and so the timing
varies accordingly. The performance losses are given in Supplementary Section 4.
Supplementary Table 2 lists the activities for each of the scenarios sequentially
where the numbers given in the table for each activity correspond to the year in
which the activities occur. The Correct Grade Schedule is the expected strategy
when the correct asphalt grade is used, whereas 1 and 2 Standard Grade
Schedules are when the asphalt grades are wrong one and two grades, respectively.
Coordinating roadways to weather stations. We identify the roadways associated
with each weather station by using the built-in functionality of ArcMap
(Version 10.3) to draw Thiessen polygons around the 5,417 weather stations in the
database. These polygons define the nearest geospatial areas to each weather
station. We then extract roadway segments from the National Highway
Performance Network (NHPN) database, which contains details on all interstates,
national routes, state routes, and paved local roads in the United States48. Using
Geographical Information Software, we then calculate the total length of each type
of roadway contained within each weather station polygon.
Life cycle cost analysis. We conduct LCCA analysis based on the maintenance and
rehabilitation schedules in Supplementary Table 2. The unit of the analysis is a
one-kilometre segment of the roadway type in question. The number of lanes
assumed for each roadway type is based on national averages: interstate and
national routes are four lanes wide, state routes are three lanes wide, and local
routes have two lanes17. Quantities of materials are estimated assuming that each
lane is 3.7 m wide. The costs associated with each activity are based on values used
by the North Carolina and Arizona Departments of Transportation. Both states
have extensive transportation networks with multiple suppliers so they provide an
overall representative indication of national costs. All costs are returned to the base
year and summed according to equation (7).
NPV =IC +
N
X
j=1
Rj1
(1+i0)njSalvage 1
(1+i0)N(7)
where NPV =net present value, IC =initial cost, Rj=rehabilitation expenditure
(single cost expenditure), Salvage =the salvage value at the end of the analysis
period, i0=the discount rate, assumed 4% (ref. 47), and nj=year of expenditure.
The salvage value is calculated by multiplying the construction cost by the
proportion of remaining life.
We estimate the total NPV associated with each weather station by first
determining which maintenance and rehabilitation schedule to follow. We then
multiply the number of lane kilometres for each type of roadway by the NPV for
the appropriate maintenance and rehabilitation schedule (see Table 1). So, for
example, a weather station with 10 kilometres of interstate would have a total
estimated NPV of US$11,837,020 when the correct asphalt was used, while the
same station would have a NPV of US$12,700,950 when the asphalt was wrong by
one grade. The impact is quantified by the difference between the future scenario
costs and the costs when all roadways have the correct asphalt grade. We cumulate
the results for all models and weather stations, and also disaggregate them by state
and region (see Fig. 3 in the manuscript and Supplementary Fig. 5). We estimate
the regional and state costs on a per lane kilometre cost by dividing the costs by the
lane kilometres (Supplementary Fig. 6). For the ensemble of models, we calculate
the median, maximum, minimum, and 75th (NPV75th percentile) and 25th
(NPV25th percentile) percentiles. For analysis by state and region, we identify outlying
model predictions using the interquartile range as shown in equation (8) (ref. 49).
Outlier =
NPV >NPV75th percentile +1.5 NPV75th percentile NPV25th percentile
OR
NPV <NPV25th percentile 1.5 NPV75th percentile NPV25th percentile
(8)
Data availability. The authors declare that the data sets used for this study are
available from http://cdiac.ornl.gov/ftp/ushcn_daily and from
ftp://gdo-dcp.ucllnl.org/pub/dcp/archive/cmip5/bcca. Also, data analysis files, and
Matlab and Labview scripts supporting the asphalt binder grade determination and
economic analysis are available from the corresponding author [B.S.U.]
upon request.
References
31. Climate Analytics Group Downscaled CMIP3 and CMIP5 Climate and
Hydrology Projections (US Department of the Interior, Bureau of Reclamation,
Technical Service Center, accessed September, 2015);
ftp://gdo-dcp.ucllnl.org/pub/dcp/archive/cmip5/bcca
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE CLIMATE CHANGE |www.nature.com/natureclimatechange
NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE3390 LETTERS
32. Reclamation Downscaled CMIP3 and CMIP5 Climate Projections: Release of
Downscaled CMIP5 Climate Projections, Comparison with Preceding
Information, and Summary of User Needs 116 (US Department of the
Interior, Bureau of Reclamation, Technical Service Center, 2013);
http://gdo-dcp.ucllnl.org/downscaled_cmip_projections/techmemo/
downscaled_climate.pdf
33. Huber, G. A. Weather Database for the Superpave Mix Design System
SHRP-A-648A (Strategic Highway Research Program, 1994).
34. Johnson, N. L. & Kotz, S. Continuous Univariate Distributions: Chapter 24 –
Beta Distributions (Houghton Mifflin Company, 1970).
35. Zapata, C. E. & Cary, C. E. Integrating the National Database of Subgrade
Soil-Water Characteristic Curves and Soil Index Properties With the MEPDG
Final Report NCHRP 9-23b (National Cooperative Highway Research
Program, 2012).
36. Applied Research Associates Guide for Mechanistic-Empirical Design of New
and Rehabilitated Pavement Structures Final Report NCHRP 1-37A
(National Cooperative Highway Research Program, National Research
Council, 2004).
37. Darter, M. I., Glover, L. T., Von Quintus, H., Bhattacharya, B. B. & Mallela, J.
Calibration and Implementation of the AASHTO Mechanistic-Empirical
Pavement Design Guide in Arizona FHWA-AZ-14-606 FHWA (Arizona
Department of Transportation, 2014).
38. Heitzman, M., Timm, D., Tackle, E. S., Herzmann, D. E. & Traux, D. D.
Developing MEPDG Climate Data Input Files for Mississippi
FHWA/MS-DOT-RD-11-232. FHWA (Mississippi Department of
Transportation, 2011).
39. Pierce, L. M. & McGovern, G. Implementation of the AASHTO
Mechanistic-Empirical Pavement Design Guide and Software
Project 20-05, Topic 44-06 (National Cooperative Highway Research
Program, 2014).
40. Smith, B. & Nair, H. Development of Local Calibration Factors and Design
Criteria Values for Mechanistic-Empirical Pavement Design FHWA/VCTIR
16-R1.FHWA (Virginia Center of Transportation, 2015).
41. VonQuintus, H. L. & Moulthrop, J. S. Mechanistic-Empirical Pavement Design
Guide Flexible Pavement Performance Prediction Models for Montana, Volume III:
Field Guide Calibration and Users Guide for the Mechanistic-Empirical Pavement
Design Guide FHWA/MT-07-008/8158-3.FHWA (Montana Department of
Transportation, 2007).
42. Witczak, M. W., Zapata, C. E. & Houston, W. N. Models Incorporated into the
Current Enhanced Integrated Climatic Model, Final Report (National
Cooperative Highway Research Program, 2006).
43. Fwa, T. F., Pasindu, H. R. & Ong, G. P. Critical rut depth for pavement
maintenance based on vehicle skidding and hydroplaning consideration.
J. Trans. Eng. 138, 423–429 (2011).
44. Lamptey, G., Ahmad, M. Z., Labi, S. & Kumares, C. S. Life Cycle Cost Analysis
for INDOT Pavement Design Procedures FHWA/IN/JTRP-2004/28 (Indiana
Department of Transportation, 2005).
45. New York Department of Transportation (NYDOT) Comprehensive Pavement
Design Manual: Chapter 5 Rehabilitation (NYDOT Design Division and
Technical Service Division, 2002).
46. Santero, N. Life Cycle Assessment of Pavements: A Critical Review of Existing
Literature and Research (Lawrence Berkeley National Laboratory, 2010);
https://www.osti.gov/scitech/servlets/purl/985846
47. Walls, J. & Smith, M. R. Life Cycle Cost Analysis in Pavement Design - Interim
Technical Bulletin FHWA-SA98-079 (Federal Highway Administration, 1998).
48. National Highway Planning Network (NHPN) National Highway Planning
Network - Tools - Processes - Planning (accessed March 2015);
http://www.fhwa.dot.gov/planning/processes/tools/nhpn
49. NIST/SEMATECH e-Handbook of Statistical Methods (National Institute of
Standards and Technology, 2012); http://www.itl.nist.gov/div898/handbook
NATURE CLIMATE CHANGE |www.nature.com/natureclimatechange
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
... Asphalt pavement is one of the most important transportation infrastructures [1][2][3][4][5]. Rutting is one of the most common distresses of asphalt pavement, seriously affecting driving comfort and safety [6][7][8][9][10]. As indicated from considerable studies, the temperature will continue to increase in the future [11][12][13][14][15][16][17], thus resulting in the aggravation of rutting [18][19][20]. ...
... The absolute ratio of d to the current maximal e 0 p is defined as R d with the expression in Eq. (6). If R d is lower than 1%, the tertiary point is calculated; otherwise, move to Step 3. ...
Article
Full-text available
The three-stage behavior is the most significant feature of permanent deformation of asphalt mixture under repeated loading, and the transition points, which refer to the dividing points of different stages, are commonly used as vital parameters to evaluate the rutting resistance. This study aims to propose a novel calculation method for transition points, which are with clear physical significance and independent of human factors. First, the effects of human factors on transition points of existing methods are assessed. As revealed from the research, the thresholds significantly impact the three-stage model method, while the termination condition and the iteration cycle significantly impact the two-step secant method and secant method with other steps, respectively. Subsequently, a three-step algorithm termed as a double-tangent method is proposed, and the physical significances of the transition points are clarified. The transition points obtained by the proposed method are proved to be of clear physical significance and independent of human factors. The slopes of the tangent line at the transition points equal to the strain rate value when the average strain rate equals to the instantaneous strain rate. Lastly, the differences of transition points and slopes between the proposed double-tangent method and existing methods are studied.
... Deterioration of the pavement structure of roadway networks is an expected phenomenon that roadway engineers plan and design for. While an asphalt pavement is typically designed to last for several decades, pavement conditions deteriorate much earlier than the design service life due to traffic, environment, material properties, and operational considerations (Underwood et al. 2017). Therefore, substantial investments are necessary to preserve, maintain, and extend the life of pavement networks (Pais et al. 2013;Gonzalo et al. 2018). ...
... Therefore, substantial investments are necessary to preserve, maintain, and extend the life of pavement networks (Pais et al. 2013;Gonzalo et al. 2018). However, due to the limited resource availability, it is critical to allocate budgets more effectively to satisfy pavement maintenance needs (TRIP 2016;ASCE 2017;Underwood et al. 2017). Consequently, pavement management systems (PMS) have received increasing attention as tools for supporting road agencies in making investment decisions and devising optimum maintenance strategies (AASHTO 2012;Pérez-Acebo et al. 2018). ...
Article
Predicting pavement rutting is associated with significant uncertainties that often lead to inefficient maintenance planning. The predictive performance of rutting models is exacerbated in local road agencies and developing countries that rely on generic and knowledge-based models which are typically unreliable if used without adaptation, validation, or calibration. This study aims at developing a probabilistic framework that employs Ensemble Kalman Filter (EnKF) techniques to update the parameters associated with generic rutting predictive models while accounting for the prevailing uncertainties. When coupled with a continuous influx of measured data, the EnKF framework sequentially updates the generic models and minimizes prediction errors in real-time. The robustness of the presented scheme is demonstrated through a numerical example, and its sensitivity to the use of different generic curves as starting points is examined. The results indicate that the EnKF framework improves the accuracy of rutting predictions by up to 60% and that accuracy remains within tolerable limits whilst varying the range of the uncertainty in the measurements or the initial states. The paper concludes with a discussion of how practitioners can integrate the outcomes of the presented framework to enact maintenance policies that minimize the financial cost at the project and network levels.
... In this research, the author noticed that there is an increase of 2-9% for fatigue cracking and 9-40% for AC rutting by the end of 20 years of service life [2]. Using the same models, Underwood et al. estimated the increased costs of pavement infrastructure from future temperature increase in the United States, which showed up to approximately US$19.0, and US$35.8 billion to pavement costs by 2040 for RCP 4.5 and 8.5 respectively [26]. To select the appropriate GCM for the pavement performance evaluation, Underwood et al. verified the detailed pavement performance analyses with various models in four different states. ...
Article
Over the past 50 years, climate change has threatened both existing and imminent pavement infrastructure. It is also forecasted to become more aggressive in the future. Therefore, pavement design procedure needs to be adapted accordingly, accounting for the climate change factors. The current study explores the projected climate change in Newfoundland using various climate change models. In addition, this study aims to quantify the influence of climate change on pavement design in Newfoundland and Labrador Province, Canada. To achieve that, a robust mechanistic-empirical pavement design tool (AASHTOWare ME Pavement Design) was used. A statistically downscaled climate change models were extracted from the coordinated regional downscaling (CORDEX) Experiments to predict the future climate. Then the minimum and maximum daily temperatures for five different locations in Newfoundland were extracted using a python code. However, there is a challenge to convert these daily temperatures into hourly temperatures. Therefore, two relatively new hourly data estimation procedures, such as the modified imposed morphing method (M-IOMM) and Sine (14R-1) were used. These hourly estimated data were utilized to further predict the pavement performance for all the five locations in Newfoundland, Canada. This study concludes that climate change accelerates pavement deterioration compared to baseline climate.
... Doing so will facilitate a true circular economy with respect to the transportation sector [8][9][10][11][12][13][14][15][16] . ...
Article
In the early 1960s, increasing traffic demanded that transportation agencies use progressively harder asphalt binders to prevent premature wheel path rutting. Using stiffer binders increased stress from cold temperature shrinkage, typically manifesting, for the first time, as rather regularly spaced transverse cracks. Trials to investigate this phenomenon confirmed the importance of binder consistency, temperature sensitivity, phase homogeneity and durability. Studies in the 1990s focused on evaluating the new Superpave™ specification, developed in the United States, to address the increasing use of modified binders in preceding decades. Air-blown materials were found to perform poorly and polymer modifiers only provided benefits in softer grades. Current trials, commissioned by the Ontario government in the 2000s, were designed to refine binder specification tests, by accounting for well-documented phase separation and transformation, associated thermoreversible changes in rheology (gel formation), and to investigate high polymer and fiber modification. Benefit/cost ratios vary by 2-3 fold for test sections of equal Superpave design but different rheological type (sol, sol/gel or gel). Long-life pavements can be designed through a judicious and balanced choice of binder properties. Albertan binders of high phase angle (sol-type), which are low in wax and modified with small amounts of polymer or fiber, outperform less stable materials, such as those with too much polymer, air-blown residue, or wax (gel-type). Improved binder specification will provide substantial savings and prepare the world’s 25 million kilometers of flexible pavements for climate change.
Conference Paper
Full-text available
Climate change is expected to affect the performance of highway pavements. However, the literature review to date does not consider how the pavement performance will be affected by extreme weather (i.e., extreme temperature and extreme rainfall). The quantification of pavement deterioration due to extreme weather can be beneficial to evaluate the effect of climate change on pavement performance. This study aims to develop a pavement deterioration model for jointed plain concrete pavement (JPCP) using the artificial neural network (ANN) modeling techniques. The models will be developed using data collected from the Long Term Pavement Performance (LTPP) database for wet, freeze climatic region. The input variables are initial pavement condition (i.e., initial IRI), pavement structural and mechanical properties (i.e., age, concrete pavement thickness, base/subbase thickness, average contraction spacing, and base/subbase materials type), traffic [Cumulative ESAL (CESAL)], and climate attributes (i.e., average annual air temperature, total annual precipitation, annual freezing index, and annual freeze-thaw), and the output is pavement deterioration indicator [international roughness index (IRI)]. The effect of different climate scenarios can be examined by conducting several models with inputs that reflect several climate scenarios. The best model will be used to investigate how much more pavement will deteriorate if climate attributes change to extreme conditions. This research will address a few gaps in the literature, including the scarcity of studies on long-term IRI prediction using LTPP data and studies on the effect of climate attributes in pavement deterioration. This analysis can be beneficial to the long-term policymaking in road infrastructure.
Article
There is strong evidence for climate change leading to a rise in temperatures and a change in precipitation trends. These environmental changes pose a threat to pavement infrastructure worldwide. Therefore, it is necessary to modify pavement design procedures to consider climate change. In addition, it is necessary to consider suitable pavement materials for future climate. The objective of this study is to develop a framework for selecting an appropriate adaptation strategy to mitigate climate change impact. To fulfill this, the influence of climate change on long-term pavement performance in Canada has been quantified over sixteen Canadian pavement sections located over various provinces in Canada. In addition, the fundamental causes of pavement deterioration due to climate change were determined using ten different climate change models. Various adaptation strategies such as upgrading asphalt binder grade, increasing the thickness of asphaltic concrete layer, increasing the base layer thickness, and using stabilized base layers were analyzed to reduce pavement deterioration and to extend the service life of the pavement. Unlike the other studies, pavement temperatures were determined using Enhanced Integrated Climate Model (EICM) to determine the change in binder grade for the future climate in the adaptation process. The study found that by 2070 all examined locations will require an upgrade in binder grade and the majority will require an upgrade in mixture gradation. Furthermore, the east and west coast will be more vulnerable to climate change and require additional measures in comparison to Central Canada. Newfoundland and Prince Edward Island are the only locations that will need to consider a change in asphalt thickness while British Columbia is the only location eventually requiring a stabilized base. This study emphasizes the necessity of climate change adaptation strategies for Canadian asphaltic concrete pavements.
Article
Full-text available
Anthropogenic forcing is driving energy accumulation in the Earth system, including increases in the sensible heat content of the atmosphere, as measured by dry‐bulb temperature—the metric that is almost universally used for communications about climate change. The atmosphere is also moistening, though, representing an accumulation of latent heat, which is partly concealed by dry‐bulb temperature trends. We highlight that, consistent with basic theory, latent heat gains are outpacing sensible heat gains over about half of the Earth's surface. The difference is largest in the tropics, where global “hotspots” of total heat accumulation are located, and where regional disparities in heating rates are very poorly represented by dry‐bulb temperatures. Including latent heat in climate‐change metrics captures this heat accumulation and therefore improves adaptation‐relevant understanding of the extreme humid heat and precipitation hazards that threaten these latitudes so acutely. For example, irrigation can lower peak dry‐bulb temperatures, but amplify latent heat content by a larger margin, intensifying dangerous heat stress. Based on a review of the research literature, our Perspective therefore calls for routine use of equivalent temperature, a measure that expresses the combined sensible and latent heat content of the atmosphere in the familiar units of °C or K. We recognize that dry‐bulb air temperature must remain a key indicator of the atmospheric state, not least for the many sectors that are sensitive to sensible heat transfer. However, we assert here that more widespread use of equivalent temperature could improve process understanding, public messaging, and adaptation to climate change. This article is categorized under: Assessing Impacts of Climate Change > Observed Impacts of Climate Change Paleoclimates and Current Trends > Earth System Behavior
Article
Full-text available
Future climate warming is virtually certain to bring about an increase in the frequency of heat extremes. Highway design and pavement selection are based on a temperature regime that reflects the local climate zone. Increasing heat extremes could, therefore, shift some areas into a different performance grade (PG) for pavement, and more-heat-resistant materials are associated with increased infrastructure costs. This study combines observations, output from global climate models, and a statistical model to investigate changes in 20-yr return values of extreme maximum pavement temperature TPmax. From a multimodel range of simulated TPmax, future changes in PG are computed for 17 major Canadian cities. Relative to a 1981–2000 baseline, summertime Canada-wide warming of 1°–3°C is projected for 2041–70. As a result, climate change is likely to bring about profound changes to the spatial distribution of PG, with the severity of the changes directly linked to the severity of the projected warming. Even under weak simulated warming, an increase in PG is projected for greater Toronto, which is Canada’s largest urban area; under moderate (strong) warming 7 of 17 (9 of 17) major cities exhibit an increase. The influence of model spatial resolution is evaluated by comparing the results from global climate models with output from a set of regional climate models focused on North America. With the exception of mountainous terrain in western Canada, spatial resolution is not a major determining factor for projections of future PG changes. Keywords: North America; Extreme events; Climate change; Climate models; Regional models; Transportation meteorology
Article
Full-text available
The database developed under NCHRP Project 9-23A, Development of a National Catalog of Subgrade Soil Water Characteristic Curves (SWCC) Default Inputs to Use in the MEPDG, included not only measured soil index properties needed in all hierarchical levels of the enhanced integrated climatic model but also SWCC parameters, which are key in the implementation of Level 1 analyses. A set of maps in portable document format displaying the location of every soil unit identified within the continental United States, Hawaii, Alaska, and Puerto Rico and a simple interface in MS Excel to aid in querying of data were also developed for the project. Under NCHRP Project 9-23B, Integrating the National Database of Subgrade Soil Water Characteristic Curves and Soil Index Properties with the MEPDG, a second research effort was directed to integrate an enhanced version of the geographic information system enabled database with the Mechanistic Empirical Pavement Design Guide. Specifically, NCHRP Project 9-23B aimed at the implementation of an interactive tool that allows the design guide users (currently Dar Win-ME users) to retrieve both appropriate soil unit maps and soil properties relevant to a particular user-specified location by entering either state milepost information or geographical coordinates. The final product was integrated into a public website that can be accessed by design guide users, agencies, industry, and academicians through a simple link. Details of the development of this useful search tool and its main features are presented.
Article
Magnetite occurs widely in natural environments in both inorganic and biogenic forms. Discrimination of the origin of magnetite has important implications, from searching for past microbial activity to interpreting paleomagnetic and environmental magnetic records in a wide range of settings. In this study, we present rock magnetic and electron microscopic analyses of marine sediments from the continental margin of Oman. Low-temperature magnetic data reveal two distinct Verwey transition (Tv) temperatures that are associated with the presence of biogenic and inorganic magnetite. This interpretation is consistent with room temperature magnetic properties and is confirmed by electron microscopic analyses. Our study justifies the use of two distinct Tv temperatures as a diagnostic signature for discriminating inorganic and biogenic magnetite. Simple low-temperature magnetic measurements, therefore, provide a tool to recognize rapidly the origin of magnetite within natural samples. In addition, our analyses reveal progressive down-core dissolution of detrital and biogenic magnetite, but with preservation of significant amounts of fine-grained magnetite within sediments that have been subjected to severe diagenetic alteration. We demonstrate that preservation of magnetite in such environments is due to protection of fine-grained magnetite inclusions within silicate hosts. Our results, therefore, also provide new insights into diagenetic processes in marine sediments.
Article
Assessment of climate change impacts on water resources is extremely challenging, due to the inherent uncertainties in climate projections using global climate models (GCMs). Three main sources of uncertainties can be identified in GCMs, i.e., model structure, emission scenario, and natural variability. The recently released fifth phase of the Coupled Model Intercomparison Project (CMIP5) includes a number of advances relative to its predecessor (CMIP3), in terms of the spatial resolution of models, list of variables, and concept of specifying future radiative forcing, among others. The question, however, is do these modifications indeed reduce the uncertainty in the projected climate at global and/or regional scales? We address this question by quantifying and comparing uncertainty in precipitation and temperature from 6 CMIP3 and 13 CMIP5 models. Uncertainty is quantified using the square root of error variance, which specifies uncertainty as a function of time and space, and decomposes the total uncertainty into its three constituents. The results indicate a visible reduction in the uncertainty of CMIP5 precipitation relative to CMIP3 but no significant change for temperature. For precipitation, the GCM uncertainty is found to be larger in regions of the world that receive heavy rainfall, as well as mountainous and coastal areas. For temperature, however, uncertainty is larger in extratropical cold regions and lower elevation areas.
Article
This work explores the influence of climate changes on the proper selection of asphalt binder for pavement construction purposes, according to the Performance Grade (PG) defined in the SUPERPAVE specifications. Based on temperature data at national level, it is possible to obtain thematic maps for the whole Italian territory, which is extremely useful for technicians and pavement engineers for selection of asphalt binder for road construction purposes. Furthermore, the statistical significant temperature trends’ knowledge enables deriving thematic maps which allows to include the effects of climate change in the asphalt binder design. It is argued that, due to climate change, the binders to be selected may be different from those commonly selected at the design stage of the infrastructure, since likely higher temperature determine more demanding constraints. The comparison among the PG grades necessary for covering the needs for construction in the different regions of Italy also call for highly performing binders, such as those obtained via specific modification with polymers. This will also imply the need for even more performing materials, in terms of mechanical properties and durability, to be used for modification of neat asphalt as well as the need for defining new specification and testing methods, specifically valid for these modified materials.
Article
Incluye bibliografía e índice