Online Resource 3: Additional Methodological Details forthe Estimation of Impacts and Adaptation Costs for Roads
Climatic Change article:
Climate Change Risks to US Infrastructure: Impacts on roads, bridges, coastal development, and urban drainage
James E. Neumann, Jason Price, Paul Chinowsky, Leonard Wright, Lindsay Ludwig, Richard Streeter, Russell Jones, Joel B. Smith, William Perkins, Lesley Jantarasami, Jeremy Martinich
Corresponding author:
James E. Neumann
Industrial Economics, Inc.
Additional Methodological Details for the Estimation of Impacts and Adaptation Costs for Roads
The methodology for assessing adaptation costs for roads, as described in Chinowsky et al. (2013), accounts for four specific effects: (1) rutting of paved roads from precipitation, (2) rutting of paved roads caused by freeze-thaw cycles, (3) cracking of paved roads during periods of high temperatures, and (4) erosion of unpaved roads from precipitation. Adaptation costs related to precipitation and freeze-thaw were modeled based on evidence suggesting that these stressors affect the frequency of periodic maintenance activities, which in turn affects annual maintenance costs. In contrast, the approach for heat focuses on changes in pavement design likely to result from climate change and the corresponding changes in re-paving costs. Depending on the nature of the changes in climate projected for a given area, the analysis of these effects may suggest a net cost or a net cost savings.
This analysis assumes that society will implement adaptation measures to avoid the adverse impacts of climate change. Economically, the optimal adaptation strategy would involve implementation of only those adaptation measures that yield a net benefit to society. That is, adaptation options that require resources exceeding the value of the benefits gained (i.e, the value of the climate change impacts avoided) would not be implemented. Thus, under the most efficient adaptation strategy, the total costs borne by society would include the costs of adaptation plus the value of residual damages remaining after the implementation of those adaptation measures that result in positive net benefits.[1] Due to the limited information available on the value society places on avoiding road degradation, we do not explicitly model this optimization process. Instead, we assume that adaptation measures will be implemented to maintain the current level of service such that residual impacts are zero, and report adaptation costs only.
To model adaptation costs for precipitation and freeze-thaw, we assessed the extent to which these stressors affect the frequency of periodic maintenance activities for paved and unpaved roads. Empirical evidence suggests that precipitation and freeze-thaw accelerate the rutting of paved roads (N.D. Lea International, 1995 and U.S. DOT, 2006), while precipitation leads to erosion of unpaved roads (Dubé et al., 2004). Changes in rutting for paved roads may affect how frequently these roads are re-sealed[2], while changes in erosion may affect the frequency with which unpaved roads are reshaped. As the frequency of these maintenance practices increases (decreases), annual maintenance costs will also increase (decrease). For example, if paved roads are re-sealed every five years under current climate but are re-sealed every four years due to climate change, this would imply that the proportion of roads re-sealed on an annual basis increased from 20 percent under current climate to 25 percent with climate change. The cost of adapting to climate change in this scenario is the cost of re-sealing an additional 5 percent of the paved road network each year. Abstracting from this example, we estimate adaptation costs related to precipitation and freeze thaw for each individual grid cell as follows:
(1)
WhereCAM = adaptation costs associated with changes in maintenance frequency (i.e., change in annual maintenance costs);
FRC = frequency of resealing paved roads (or reshaping unpaved roads) under the climate change scenario (i.e., every FRC years);
FRB = frequency of resealing (or reshaping) under current baseline climate (once every 10 years);
MTU = total miles of paved (unpaved) road; and
RU = re-sealing (or reshaping) unit cost.
Under this approach, the expression represents the change in the miles
of road re-sealed (or reshaped) on an annual basis. We obtained MTU from the DOT/Tele Atlas inventory for each grid cell and derived separate estimates of the change in the frequency of re-sealing for paved and unpaved roads.
As indicated by Equation 1 above, the change in the frequency of re-sealing paved roads is estimated based on the difference between the frequency of re-sealing under current climate and the re-sealing frequency with climate change. The frequency of re-sealing under current climate (FRB) varies depending on the type of seal coat employed, road use, and other local conditions. Based on literature review ranges and data from state departments of transportation, we assume that paved roads are, on average, re-sealed every seven years under current climate.
We estimate the frequency of re-sealing with climate change (FRC) based on the impact of precipitation and freeze-thaw on the condition of paved roads, as measured by the pavement condition index (PCI).[3] A road’s PCI rating is typically at or near 100 at the beginning of its lifecycle but degrades over time prior to rehabilitation.[4] The precise level of PCI degradation occurring over a road’s lifecycle varies, but for the purposes of this analysis, we assume that roads are rehabilitated once their PCI rating declines to 55, which represents the threshold between good and fair pavement conditions. Applying this assumption, a road’s PCI rating declines by approximately 45 points during its lifecycle prior to rehabilitation.
To assess the frequency of re-sealing with climate change, we assume that a maximum level of PCI degradation is acceptable between re-sealings. The ratio of this maximum level of degradation to the annual rate of PCI degradation with climate change represents the frequency of re-sealing with climate change, as outlined in Equation 2. As indicated in the numerator of the right-hand side of Equation 2, the maximum value of PCI degradation between re-sealings may be derived from the annual rate of PCI degradation under current climate (RB) and the frequency of re-sealing (FRB). In addition, as indicated in the denominator, the annual amount of PCI degradation with climate change is estimated as the sum of PCI degradation under current climate (RB) and the incremental change in PCI degradation associated with climate change (RC).
(2)
WhereFRC = frequency of re-sealing with climate change (i.e., every FRC years);
FRB = frequency of re-sealing under current baseline climate (once every 7 years);
RB = annual rate of PCI degradation under current climate; and
RC= incremental PCI degradation per year due to climate change.
The annual rate of PCI degradation with current climate (RB) is likely to vary, but we approximate this value based on the typical PCI degradation occurring during a road’s lifecycle (45 points, as indicated above) and the typical duration of a road’s lifecycle. Based upon prior analyses by the U.S. Department of Transportation (2006) and N.D. Lea International (1995) in which a road’s lifecycle was modeled over a period of 20 to 30 years, we assume a road lifespan of 20 years for roads located in wet climates (average monthly precipitation exceeding 5 cm) or areas that experience at least 50 freeze days per year.[5],[6] A lifecycle of 30 years is applied to roads located in all other areas where precipitation and freeze-thaw stress are not as significant.
Averaging the 45-point PCI degradation occurring during a road’s lifecycle over the 20 to 30-year lifespan range, we estimate that annual PCI degradation under current climate (RB) is 1.5 points in wet areas and moderate or high-freeze areas[7] and 2.25 points in other areas. Given this range of values for RB and a re-sealing frequency (FRB) of seven years under current climate (see above), we estimate that the acceptable level of PCI degradation between re-sealings (the numerator in Equation 2) ranges from 10.5 to 15.75 PCI points.
Calculating the frequency of re-sealing with climate change (FRC) based on Equation 2 requires estimation of the extent to which precipitation and freeze-thaw affect annual PCI degradation (RC). To estimate RC, we draw from prior studies examining the rutting associated with precipitation and freeze-thaw, and subsequently assess the changes in PCI associated with these rutting impacts. N.D. Lea International (1995) indicates that rut depth over a road’s lifecycle increases by approximately 3 millimeters with every 10 centimeter increase in mean monthly rainfall. In addition, U.S. DOT (2006) shows that rut depths in moderate freeze areas (50 to 400 freeze days per year) and high-freeze areas (more than 400 freeze days per year) are approximately 3.25 and 2.75 millimeters higher, respectively, than in no-freeze zones. From these figures, we also infer that lifetime rutting is approximately 0.5 millimeters higher in moderate freeze areas than in high freeze areas. These values and those for precipitation represent the incremental rutting occurring over a road’s lifetime. To convert to annual values, these estimates are divided by a road’s lifespan (i.e., 20 or 30 years).
These rutting effects are translated into PCI changes based on the level of PCI degradation per millimeter of rut depth. To derive this relationship, we use data for the total PCI degradation and rutting occurring over a road’s lifecycle. As indicated above, we assume PCI degradation of 45 points during a road’s lifespan. Measurements of the degree of rutting occurring over a road’s lifecycle vary across the identified studies. N.D. Lea International (1995) estimates 8 millimeters of rutting over a road’s lifecycle, while U.S. DOT (2006) estimates 5.75 millimeters of rutting. Based on these values, PCI degradation of 5.625 points per millimeter of rutting is estimated related to precipitation and 7.83 PCI points per millimeter of rutting related to freeze-thaw.[8]
Based on the estimated change in re-sealing frequency for paved roads, we estimate the corresponding change in re-sealing costs assuming a unit re-sealing cost of $13,500 per lane mile, as derived from Yamada and Dimas (1999).
Changes in precipitation patterns associated with climate change may affect the erosion of unpaved roads and the frequency with which transportation agencies re-shape these roads. To estimate this change in frequency, we follow an approach similar to that outlined above for paved roads. More specifically, we estimate, by grid cell, the cumulative level of unpaved road erosion that occurs during the typical reshaping cycle under current climate. Treating this level of erosion as the maximum amount of erosion acceptable between reshapings, we then calculate how quickly this cumulative amount of erosion occurs under each climate change scenario—this represents our estimate of the frequency of reshaping with climate change.
To apply this method, the annual rate of erosion is first estimated under current climate by grid cell. These values are then multiplied by the frequency of reshaping under current climate to derive estimates of the cumulative level of erosion that occurs between each reshaping. These cumulative values are represented as the numerator in Equation 3. In calculating the numerator value, we assume that the average frequency of reshaping under current climate (FRB) is once every 10 years.[9] Dividing the level of erosion that occurs between reshapings (FRB × EAB) by the annual erosion that occurs with climate change (EAC) yields the frequency of re-shaping under each climate change scenario (FRC).
(3)
Where FRC = frequency of reshaping under climate change scenario;
FRB = frequency of reshaping under current climate (i.e., every FRB years);
EAB = annual erosion with current baseline climate (tons/acre/yr) , and
EAC = annual erosion (tons/acre/yr) with climate change
Calculation of the EAB and EAC terms in Equation 3 requires estimation of annual unpaved road erosion. Erosion is estimated as a function of precipitation, based on the findings of Dubé et al. (2004) and Sheridan and Noske (2005):[10]
(4)
Where EA = annual erosion (tons/acre/yr), and
PA = annual precipitation (inches).
After estimating the change in reshaping frequency, the annual change in re-shaping costs is estimated based on a unit re-shaping cost of $24,200 per lane mile derived from R.S. Means (2008).[11]
To assess the cost of adapting roadways to changes in temperature, we examine the implications of climate change for the design specifications of asphalt pavements. In areas where maximum temperatures increase due to climate change, asphalt pavements will become susceptible to increased cracking. This impact may be avoided by using a different binder in the surface asphalt mix.[12] To model adaptation through the use of different pavement binders, guidelines for Superpave are utilized, which specify binder performance grades based on the maximum 7-day pavement temperature during the course of the year.[13] The thresholds between Superpave performance grades occur in 6-degree increments between pavement temperatures of 46 and 82 degrees Centigrade, as shown in Table A.
Table 1. Superpave binder performance gradesPerformance Grade / 7-day Maximum Pavement Temperature (˚C)
PG-46 / 46
PG-52 / 52
PG-58 / 58
PG-64 / 64
PG-70 / 70
PG-76 / 76
PG-82 / 82
Source: Washington State DOT (undated)
The cost (savings) of adapting paved roads to higher (lower) temperatures is estimated as the incremental cost of re-paving with a higher (lower) grade binder. As an initial step in this process, the daily pavement temperature is estimated for each 0.5 degree by 0.5 degree grid cell under current climate and under each climate change scenario using the following equation from Lavin (2003):
(5)
Where TP = Pavement temperature (˚C)
TA = Air temperature (˚C)
L = Latitude
Using daily pavement temperatures derived from Equation 5, the binder performance grade appropriate to each grid cell is identified under current climate and with climate change. As indicated above, the Superpave guidelines specify performance grades based on the maximum 7-day pavement temperature during the course of the year. The maximum daily temperature is used as an approximation of this value.
After specifying the binder performance grade appropriate to each grid cell under current climate and with climate change, we identified grid cells where the performance grade is likely to change due to climate change. For each of these areas, the incremental annual cost of re-paving is estimated as a function of the total lane miles scheduled to be repaved and the incremental unit cost of re-paving, as represented below:
(6)
WhereCR = Incremental annual cost of re-paving;
MT= Total miles of road;
FP = Frequency of re-paving (i.e., every FP years);
BC = Asphalt unit cost with binder required with climate change; and
BB = Asphalt unit cost with binder required under baseline current climate.
The term MT(1/FP) in Equation 6 represents the lane miles of road scheduled to be repaved each year, while the term BC - BB represents the change in the per unit cost of asphalt when switching from one performance grade to another.
In applying Equation 6, the frequency of re-paving (FP) is set equal to once every 10 years, consistent with observed practice.[14],[15]
We estimate the cost of asphalts with varying binder performance grades (e.g., BC and BB) based on construction project data from the Colorado Department of Transportation (2010), as summarized in Table 1. While the costs of asphalt vary regionally, the Colorado data provided a complete set of asphalt costs, including cost variance by performance grade. This level of detail was necessary to implement the approach summarized in Equation 6. Based on asphalt pricing data from Argus (2011), asphalt costs in Colorado are 10 to 15 percent lower than in many other parts of the U.S. Our adaptation cost estimates associated with switching to higher grade binders are therefore likely to be conservative.
Table 1.average costs of asphalt by performance grade
Performance Grade / Cost (year 2010$ per lane mile)
PG-46 / $197,000
PG-52 / $210,000
PG-58 / $225,000
PG-64 / $241,000
PG-70 / $258,000
PG-76 / $276,000
PG-82 / $295,000
Source: Derived from Colorado DOT (2010)
Figure 1: Distribution of annual adaptation costs for roads by cost type for the climate projections spanning the range of precipitation futures in the US (billions of 2005$).
References
American Society of Civil Engineers. (2011). Failure to Act: The Economic Impact of Current Investment Trends in Surface Transportation Infrastructure.
Argus (2011). Asphalt Report. November 11, 2011.
Chinowsky, P., J. Price, and J. Neumann (2013) Assessment of climate change adaptation costs for the U.S. road network. Global Environment Change. 23: 764–773.
Colorado Department of Transportation (2010). “2010 Cost Data Construction,” July 27, 2010.
Dubé, Kathy, Walt Megahan, and Marc McCalmon (2004). Washington Road Surface Erosion Model, prepared for State of Washington Department of Natural Resources. February 20, 2004.
N.D. Lea International, Limited (1995). Modelling Road Deterioration and Maintenance Effects in HDM-4, prepared for the Asian Development Bank. October 1995.
R.S. Means (2008). Heavy Construction Cost Data 2008 Book. Reed Construction Data. 2008.
Sacramento Area Council of Governments (2006). Metropolitan Transportation Plan 2035, October 2006.
U.S. Department of Transportation, Federal Highway Administration (2006). Long-Term Pavement Performance (LTPP) Data Analysis Support: National Pooled Fund Study TPF-5(013) - Effects of Multiple Freeze Cycles and Deep Frost Penetration on Pavement Performance and Cost, Publication No. FHWA-HRT-06-121. November 2006.
Wyoming Technology Transfer Center (2010). Gravel Roads Management Volume 1, report number FHWA-WY-10/03F, prepared for U.S. DOT Federal Highway Administration and State of Wyoming Department of Transportation. October 2010.
Yamada, Alan and Sam Dimas (1999). “Asphalt Seal Coat Treatments,” prepared for U.S. Forest Service. accessed January 19, 2011.