Semisequicentennial Transportation Conference Proceedings
May 1996, Iowa State University, Ames, Iowa

A Methodology for Determining Road Damage Due to Railroad Branchline Abandonment

Eugene R. Russell, Sr., Michael W. Babcock, and Curtis Mauler

E.R. Russell, Sr.,
Center for Transportation Research and Training,
Kansas State University,
Seaton Hall,
Manhattan, Kansas 66506-2905.

M.W. Babcock,
Economics Department,
Kansas State University,
Waters Hall,
Manhattan, Kansas 66506.

C. Mauler,
Wilson & Company,
3059 W. 13th Street, P.O. Box 850,
Wichita, Kansas 67203.

As of late, there has been a trend towards low-density rail abandonment. This trend is cause for great concern for the policy makers and planners of Kansas and other midwestern states. Abandonment of light density rail lines constitutes a major change in the method of transportation by which grain is transported in rural areas, not only at the local elevator level but also at the production level. Farmers, who generally sell their produce to elevators offering the highest bid price, would probably be inclined to ship their grain to elevators served directly by rail because of the tendency of these elevators to provide higher bid prices due to the cheaper shipping costs. In addition, with discontinued rail service, elevators are forced to truck their grain to other elevators with rail service or to terminal elevators. The impacts of these changes cause increased truck mileage which means additional use and damage to public roads and streets. Since highway pavements are structures with finite lives, they are designed to withstand a specific number of equivalent, 18,000 pound single axle loads (ESALs). The truck traffic consumption of ESAL design life, and increased highway infrastructure costs associated with it, can increase rapidly where significant volumes of rail traffic diversions to trucks are involved. This phenomenon not only occurs on those highway segments that were designed for high levels of truck traffic, but also occurs with greater consequence to rural highways which are often not designed to handle large truck volumes. In this study, a transportation estimating model developed by M. H. Chow was used to generate wheat flow data relevant to determining the impacts of railroad abandonment. Grain flow data in bushels of wheat was developed by Chow's program, upon completing above explained data input procedures, for two scenarios: 1) pre-abandonment scenario and 2) post-abandonment scenario. These data were then transformed into truck loads originating at the different farms and elevators. The proposed abandonment of ATSF branchlines in south central Kansas were found to produce an estimated additional eight percent of wheat that will be transported by truck out of local elevators to terminal elevator transit points. Road damage costs were calculated for both the before and after abandonment scenarios. This analysis provided that the abandonment of the considered ATSF rail lines would result in an estimated additional annual damage to Kansas highways of $1,004,590. The additional damage constituted a 48 percent increase in incremental truck damage over the before abandonment scenario caused by the rail abandonments. Key words: rail abandonment effects, road damage, hauling wheat by truck, Chow network model.

The current trend towards low-density rail abandonment is cause for great concern for the policy makers and planners of Kansas and other states. U.S. rail mileage decreased by 44 percent to 140,000 miles and the Kansas rail mileage was reduced by 24 percent to 7,086 miles in 1988 alone (1). Farmers, who generally sell their produce to elevators offering the highest bid price, would probably be inclined to ship their grain to elevators served directly by rail because of the tendency of these elevators to provide higher bid prices due to cheaper shipping costs. In addition, with discontinued rail service, elevators are forced to truck their grain to other elevators with rail service or to terminal elevators. The impacts of these changes cause increased truck mileage which means additional use and damage to the state's roads. This paper will cover the findings of a study to document the impact of rail line abandonment on road maintenance costs.

Highway pavements are structures with finite lives. They are designed to withstand a specific number of equivalent, 8165 kg (18,000 lb) single axle loads (ESALs). One rail car of grain or dry fertilizer is roughly equivalent to 3.7 semi-trucks. Consequently, the truck traffic consumption of ESAL design life, and increased highway infrastructure costs associated with it, can increase rapidly where significant volumes of rail traffic diversions to trucks are involved (2). If a road section was not designed for heavy axle loads, as many rural roads are not, it could be rendered inadequate in a matter of months or even weeks.


In November 1992 the Santa Fe railroad announced the sale of 800 miles of rural Kansas branchline to Broe Corporation which will operate this system as the Central Kansas Railway. Even if short lines are a financial success, the Kansas rural rail system is likely to continue to shrink in the 1990s (3). Some short lines may not be successful.

The Kansas Department of Transportation (KDOT) needed to know the impacts of possible abandonment of these lines, and contracted with Kansas State University (KSU) to conduct a study. This study focused on 1) the potentially increased truck transportation and 2) road maintenance costs of abandonment. This paper primarily covers the latter.


If rail service is withdrawn, additional trucking of wheat will occur. Some farmers may continue to deliver wheat to elevators on these abandoned rail lines. After abandonment, these elevators would be completely reliant on trucks for shipment of grain to markets. In other cases, farmers will transport their wheat over greater distance to an elevator that offers a higher price due to the existence of rail service at that location. It is impossible to determine before the fact how much additional trucking will occur as a result of abandonment. In this study, it is assumed that elevators and farmers use the transportation service that minimizes wheat transportation and handling costs. This will maximize farm price received and country elevator profit margins.

Given the above assumption, a model is required that describes the movement of wheat from study area farms to final markets at the least transportation and handling costs. To do this, it was decided to use a wheat logistics network model developed at Kansas State University (4,5). The model, commonly referred to as the "Chow network model," was employed to measure truck and rail shipment of wheat assuming no rail abandonment in the study area. Then the model was used assuming rail abandonment to determine the additional trucking of wheat that would occur if three former Santa Fe branchlines in the study area were abandoned. The incremental trucking caused by abandonment is the difference in truck shipments measured by the two simulations.

To use the Chow network model, 400 5 km x 5 km simulated farms were identified visually on county maps as origins of production. The amount of wheat produced from the various farms was determined by proportioning county wheat production data. Ninety percent of this wheat produced from the simulated farms was then assumed to be shipped by truck to one or more of several country elevators that were determined to be feasible as intermediate destinations. It was assumed that the grain could travel by ither truck or rail from the country elevators to the inland or export terminals.

For each simulation, two types of truck movements were identified. The first is farm to local elevator movements by single-unit, two axle, farm trucks (SU-2AX) over a combination of county, municipal, and state roads. The second set involves commercial truck, five axle (CO-5AX) movements from country elevators to terminal elevators over various Kansas roads. In each case, bushels transported were converted to truck trips by road segment. This was done by dividing the wheat volume moved by truck on each road segment by truck capacity. Payload capacities were assumed to be 810 bushels for commercial trucks and 256 bushels for farm trucks.

Trucks were routed via a combination of county, city, and state roads from origin to final destination. For each movement along the various road systems, the different highway sections utilized by the wheat trucks were visually identified from county maps and combined to obtain the shortest highway route. Deficient bridges were taken into consideration by not routing trucks over any bridges rated less than eight tons. After the road segments were identified, the number of annual truck trips needed to move the wheat volume determined by the Chow network model over each segment was determined and then converted into ESALs which could be used in conjunction with pavement damage costs to estimate an amount of damage caused to the different road segments involved. The difference in truck-accountable road damage cost between before and after abandonment scenarios is equal to the additional road maintenance cost that abandonment of the considered lines would incur.


Neglecting the effects of natural aging, the fundamental relationship between pavement life (PL) and its design considerations can be depicted as follows (6):

PL = f(N,C,SSN,STR) [1]


N = Cumulative passes of a given axle type and load

C = Climatic zone or regional factor

SSN = Soil support number or index

STR = Strength of the highway section (some function of D or SN, T1, and/or T2)


D = Slab thickness (for rigid pavements)

SN = Structural number—an abstract measure of pavement strength

T1 = Thickness of asphalt concrete layers

T2 = Thickness of the aggregate base

For a mixed traffic stream and defined values of C and SSN, the effects of different axle passes can be translated into equivalent, 8165 kg (18,000 lb) ESALs, used as a common unit of measurement. Therefore, the pavement life becomes a function of ESALs and equation [1] becomes (6 ):

PL = f(ESALs) [2]

Road damage techniques developed by Tolliver were followed (5,6). The technique is basically based on modified American

Association of State Highway and Transportation Officials (AASHTO) pavement damage equivalency equations and Highway Performance Monitoring System (HPMS) pavement functions. The pavement damage equivalency equations determine the

degree of road damage for various axle loads in terms of standard ESALs; the HPMS pavement functions relate road life to ESALs (7 ). The modified AASHTO traffic equivalence formulas are shown in Figure 1.

Flexible pavement:

LOG100(NR/NX) = 4.79 * LOG10(LX + 1) - 4.79

* LOG10(LR + 1) + G/bR - G/bX [1]

Rigid pavement:

LOG10(NR/NX) = 4.62 * LOG10(LX + 1) - 4.62

* LOG10(LR + 1) + G/bR - G/bX [2]

The AASHTO axle equivalence formulas for tandem axles are as follows:

Flexible pavement:

LOG10(NR/NX) = 4.79 * LOG10(LX + 2) - 4.79

* LOG10(LR + 1) - 4.33 * LOG10(2)

+ G/bR - G/bX [3]

Rigid pavement:

LOG10(NR/NX) = 4.62 * LOG10(LX + 2) - 4.62

* LOG10(LR + 1) - 3.28 * LOG10(2)

+ G/bR - G/bX [4]


LOG10(NR/NX) = Log of the traffic equivalency formula

LR = Reference axle weight (18 kips)

LX = Axle weight (kips)

PSR = Pavement serviceability rating

G = LOG10[(5 - PSR)/3.5]

b = A damage function coefficient expressed below for the two pavement types as:

Flexible pavement:

b = .40 + [.081 * (L1 + L2)^3.23]/[(SN + (6/SN)^.5)^5.19 * L2^3.23] [5]

Rigid pavement:

b = 1 + [3.63 * (L1 + L2)^5.20]/[D + 1)^8.46 * L2^3.62] [6]


L1 = Axle load (kips)

L2 = Axle type (where "1" = the single axle and :2" = the tandem axle)

D = Depth of pavement (inches)

NOTE:The damage function coefficient (b) is computed with respect to the reference axle (bR) and axle group (bX), i.e., single or tandem axle.

Pavements have a limited useful life in terms of the passage of a finite number of ESALs, i.e., each passage uses up a portion of the pavement life. The life of a typical highway section that is maintained to acceptable standards is comprised of a series of cycles. Pavements are rehabilitated or reconstructed when the pavement becomes "unacceptable" for normal traffic use in terms of ride comfort (Pavement Serviceability Rating or PSR) and is usually

improved prior to the full expiration of structural pavement life, i.e., a specified minimum PSR. (Note that it is more common for pavement engineers to use Present Serviceability Index (PSI), the analytical equivalent of PSR, as a measure of pavement condition; however, PSR is used in the method developed by Tolliver (6).)

The consumption of pavement life constitutes an economic cost which occurs whenever a portion of the remaining useful life of a pavement is consumed. Two types of economic costs are associated with pavement consumption (6): 1) marginal cost and 2) incremental cost.

Within the context of highway impact analysis, short-run marginal cost (SRMC) reflects the additional consumption of highway rideability (PSR) resulting from each additional ESAL applied to a highway section in its given current condition. The long-run marginal cost (LRMC) has nothing to do with the current condition of a highway section and instead is the cost of an increase in pavement strength necessitated by the summation of ESALs over the life of the pavement (6). The LRMC of an ESAL would be the additional layer of thickness required to maintain the service life of a highway as before the one ESAL addition.

Incremental pavement costs are the costs that are most relevant in this study. The concept of short run marginal costs was used to obtain these values because there is a key linkage between marginal and incremental cost. The cost of an increment of traffic is roughly the sum of the marginal costs incurred by the individual vehicles.

The marginal cost of an axle pass depend on two factors (6): 1) age and serviceability of the highway section, and 2) vehicle axle loads and configurations.

The general steps are:

  1. Measure pavement life of each road segment in ESALs using HPMS pavement functions.
  2. Measure road damage in ESALs for each truck using the modified AASHTO traffic equivalency functions.
  3. Estimate total pavement damage costs for a given number of trucks.

The decline in pavement serviceability rating (PSR) is a nonlinear function of traffic over time. Logically, then, the short-run marginal cost of an axle pass will vary with time, increasing with age and decreasing serviceability of the highway section. For the reference axle, 8165 kg (18,000 lb), the marginal cost at any point on the PSR decay curve, modeled by the HPMS pavement functions, is given by the derivative of pavement serviceability with respect to cumulative axle passes.

In the method used in this study, an effective ESAL life of each segment was estimated using the HPMS pavement functions. The following equations were used (1):

Flexible Pavement:

LOG10(ESAL) = 9.36 * LOG10[SN + [3]

(6/SN)^.5] - .2 + G/B

G = LOG10 [(5 - PSRt)/3.5] [4]

B = .40 + 1094/[SN + (6/SN)^.5]^5.19 [5]

Rigid Pavement:

LOG10(ESAL) = 7.35 * LOG10(D + 1) [6]

- .06 + G/B

G = LOG10 [(5 - PSRt)/3.5] [7]

B = 1 + 16,240,000/(D + 1)^8.46 [8]


SN = Pavement structural number

D = Pavement thickness

PSRt = Terminal pavement serviceability rating

G = Damage index which is a function of PSRt

B = Damage function coefficient as estimated by a regression analysis from road test data

In this study, pavements with pavement structural numbers (SNs) greater than 5.5 are classified as rigid pavements while those equal to or less than 5.5 are flexible pavements. PSRt of 2.5 was used for state highways and PSRt of 2.0 was used for local roads.

The HPMS damage function basically uses the AASHTO pavement damage functions with the following modifications (6):

  1. HPMS uses the pavement serviceability rating (PSR) rather than the present serviceability index (PSI). PSR involves a subjective rating scheme whereas the PSI involves a mathematical relationship.
  2. In HPMS, the design serviceability rating is set at its theoretical maximum (PSR=5.0) rather than at 4.2 as used in AASHTO calculations.
  3. The term "(SN+1)" in the AASHTO function is replaced by the term (SN+(6/SN)^.5) in the HPMS function.

Structural numbers (SN) and PSRs were obtained from KDOT's Bureau of Transportation Planning for state roads, and samples from the various county roads were acquired from the same source for the local road system. These data were first input into a spreadsheet which in turn used the above equations to calculate the effective ESAL life of each highway section on the truck routes.

The AASHTO traffic equivalency formulas, equations 1 through 6 given in Figure 1, were used to accomplish the task of converting truck axle load passes to ESALs. The average empty and loaded axle loads for SU-2AX and CO-5A trucks were obtained from available, KDOT state-wide truck weight data, then converted into marginal costs in ESALs for each axle weight, given the strength and condition rating of each highway section. Then an average marginal cost in ESALs was determined for a single pass of each truck type by summing the average loaded and empty ESAL values for each axle of the given truck type.

To summarize, a Lotus spreadsheet was used to perform the calculations of road damage which included (1):

1. The total number of annual, loaded truck trips over visually determined truck routes was calculated by dividing the

total volume of grain moved between corresponding points (determined by the chow network model) by the truck capacity for each type of truck.

2. The present, effective ESAL life of each section of each route was calculated using the highway's attributes and the HPMS damage function equations.

3. The marginal cost of an axle pass in ESALs was determined for the two different vehicles involved (CO-2AX and CO-5AX) and for each road section involved using AASHTO traffic equivalency formulas (as in No. 4).

4. The ESALs for empty and loaded trucks for each of the two vehicles' axle groups were determined by the AASHTO traffic equivalency formulas and summed for each vehicle to obtain the degree of road damage per round-trip VMT for each vehicle, i.e., it was assumed that there would be an empty truck trip (return) for each loaded truck trip.

5. The number of annual truck trips by each truck type was multiplied by the road damage in ESALs per round trip VMT to obtain the total annual damage for each road segment.

6. Pavement rehabilitation costs per mile obtained from KDOT were divided by each segment's present, effective ESAL life to obtain a pavement rehabilitation cost per ESAL mile.

7. The cost attributable to trucks for pavement damage was calculated for each segment by multiplying the road damage in annual ESALs by segment miles and the pavement rehabilitation cost per ESAL mile for that segment.

8. The truck-accountable costs for each segment were summed to obtain the cost of damage caused by trucks on any given route, road type or total system.

These calculations were performed for two scenarios: (1) with no rail line abandonment and (2) assuming rail line abandonment. The difference in the two scenarios constitute the impact that rail line abandonment would have on rural highways in south central Kansas.


It was assumed that SU-2AX trucks were used for truck movements from simulated farm to country elevator while CO-5AX trucks were employed for country elevator to final destination transits. The per mile costs of surfacing and shoulders for rural roads were (1): Interstate $568,000, Principal Arterial $424,000, Minor-Arterial $248,000, Collector $161,000, and local $58,000. For urban roads they were, respectively, $1,217,000, $963,000, $462,000 and $115,000.

For farm-to-country elevator wheat movements with no abandonment, total road damage costs would be $638,613. Assuming abandonment, these costs would increase to $911,972, a 43 percent increase. Thus, the total road damage cost that would be due to a Santa Fe abandonment is $273,359. Of this amount, $261,699 (96 percent of the total) in road damage cost would occur on state funded arterial and collector roads, and $11,468 would occur on county funded roads. These increases in road damage costs after abandonment is due to farmers trucking their grain over longer distances to elevators with rail service.

There also are road damage costs of truck wheat movements from study area country elevators to terminal elevators (i.e., intercity movements) (2). The roads used by trucks in the intercity wheat movements are generally of higher quality than those used in the farm-to-country elevator movements. However, the larger trucks moving over greater distances more than offset higher road quality and inflict much more damage. With no abandonment, total road damage cost attributable to trucks would be $1,451,494. Assuming abandonment these costs would increase to $2,182,725, a 50 percent increase. The truck attributable road damage cost that would result from abandonment, is the difference between the above two figures: $731,231. All of this cost would occur on state funded arterial and collector roads.

Total truck attributable road damage cost that would be due to abandonment is $1,004,590 ($273,359 + $731,231), a 48 percent increase from the no abandonment case (2). Of the total damage cost, 27 percent is due to farm-to-country elevator movements and 73 percent to country elevator to terminal market movements. The one million dollar cost is probably conservative since the network model is unable to incorporate rail movements of wheat to local area flour mills. After an abandonment, some of this wheat would also be diverted to commercial trucks.

  1. Association of American Railroads. Railroad Facts. 1993 edition, Washington, D.C., pp. 44–45.
  2. C.E. Mauler. The Impacts of Selected Atchison, Topeka and Santa Fe Railroad Abandonments on the Highway Infrastructure of South Central Kansas. M.S. thesis, Kansas State University, Department of Civil Engineering, Manhattan, Kansas, 1990.
  3. M.W. Babcock, E.R. Russell, and R.E. Burns. Economic Development and Transportation Impacts of Railroad Branchline Abandonment in South Central Kansas. Kansas Department of Transportation, Topeka, Kansas, 1992.
  4. M.H. Chow. Economic Impacts of Structural Changes in the Wheat Logistics System: Exporting Winter Wheat from Twelve Counties in Northwest Kansas. Ph.D. dissertation, Department of Economics, Kansas State University, 1984.
  5. M.H. Chow. A Capacitated Network Model Software Using TURBOPASCAL for Personal Computers. Kansas State University, Department of Economics, Manhattan, Kansas, 1985.
  6. D.D. Tolliver. The Impacts of Grain Subterminals on Rural Highways. Ph.D. dissertation, Department of Environmental Design and Planning, Virginia Polytechnic Institute and State University, Blacksburg, Virginia, 1989.

The study on which this paper is based was funded by the Kansas Department of Transportation under the K-TRAN research program. The views, opinions, and conclusions expressed are those of the authors only.

CTRE is an Iowa State University center, administered by the Institute for Transportation.

Address: 2711 S. Loop Drive, Suite 4700, Ames, IA 50010-8664

Phone: 515-294-8103
FAX: 515-294-0467


Iowa State University--Becoming the Best