Michael H. Schrader, P.E.

Wilbur Smith Associates, Inc.

Dallas, Texas



John R. Hoffpauer, AICP


Little Rock, Arkansas




In order to better allocate scarce regional transportation resources when considering the construction of grade-separated highway-railway crossings in Central Arkansas, Metroplan, the Metropolitan Planning Organization, created a methodology to provide a consistent quantitative evaluation of potential locations.  This methodology considered use of seven quantitative factors: noise, community cohesion, delay, accessibility, connectivity, geographic distribution, and safety.  Each factor yielded a value between 0 and 1, which allows an equally weighted comparison between the different factors, as well as a net sum of all factors.  The net result, then, is an aggregate value that can aid in the evaluation and prioritization of potential highway-railway grade separation locations.






Michael H. Schrader, P.E.

Wilbur Smith Associates, Inc.

4925 Greenville Avenue

Suite 915

Dallas, Texas 75206-4085


Phone:  (214) 890-4460

Fax: (214) 890-7521

e-mail:  mschrader@wilbursmith.com



John R. Hoffpauer, AICP


501 West Markham Street

Suite B

Little Rock, Arkansas  72201


Phone: (501) 372-3300

Fax: (501) 372-8060

e-mail:  jhoffpauer@metroplan.org




Paper No. 01-3051



Michael H. Schrader, P.E.

Wilbur Smith Associates, Inc.

Dallas, Texas


John R. Hoffpauer, AICP


Little Rock, Arkansas




One roadway element that has always been of concern to traffic practitioners since the mass popularization of the automobile as the preferred mode of travel is the at-grade railroad crossing.  Railroad crossings are perceived by many to be among the most dangerous, if not the most dangerous, of all the roadway elements.  Nowhere else on our roadway network is there a direct conflict between vehicles of such different size and operating capability than at a railroad crossing -- when a collision does occur, it is usually severe.  In addition to the safety concerns of at-grade crossings are the economic and environmental concerns.  Delays at at-grade crossings contribute to the loss of economic productivity, excessive wear-and-tear on vehicles, and the inefficient use of non-renewable natural resources (e.g. fuel).  In addition, inadequate maintenance of at-grade crossings increases operational costs for both motorists and railroads.  In short, at-grade crossings, and especially poorly maintained crossings, can be detrimental to the economic vitality of a community.

Because of these economic and safety issues, it is often beneficial to replace an at-grade crossing with a grade-separated one.  The Interstate system provided the upgrading of at-grade crossings along the nation’s trunk highways.  While the impact of the Interstate system has been substantial in terms of mobility and economic vitality, the Interstate system represents just a small fraction of the total road and highway mileage, and correspondingly, the total number of at-grade railroad crossings, across the nation.

Unfortunately, grade-separated rail crossings are relatively expensive, and thus the prevalent railroad crossing is at-grade.  For many jurisdictions, the cost of just one grade separation would consume several years’ worth of operating budgets.  The majority of the member communities of Metroplan, the Metropolitan Planning Organization (MPO) for Central Arkansas, as well as the MPO itself, are these types of jurisdictions.  Because of these fiscal constraints, it was necessary to create a methodology for evaluating and prioritizing proposed grade-separated rail crossings so that the limited financial resources available would be spent where they would have the greatest impact.



When deciding what kind of methodology to use in evaluating potential highway-railway grade separation locations, it was recognized that while benefit/cost ratios are important, they should not be the only determinant in selecting where rail grade separations should be located.  It was decided to attempt to develop and apply a range of evaluation factors in an effort to prioritize highway-railway grade separation needs in the Central Arkansas metropolitan area.

Seven factors, both quantitative and qualitative, were identified as preliminary evaluators, and quantitative equivalents were established for each.  These seven factors are:  noise, community cohesion, safety, delay, accessibility, geographic distribution, and connectivity.  Each factor is equated to yield comparable values between 0 and 1.


Factor 1:  Noise

Train horn noise is a significant quality-of-life issue in many communities across the country.  The Central Arkansas communities shared this concern.  The municipal buildings of several of the cities of Central Arkansas are so close to railroad tracks that phone calls and meetings are interrupted when a train passes because a human voice can not be comprehended over the noise of the train’s horn.

While the communities did not want to impose Draconian measures such as the banning of the blowing of a train’s horn within their corporate boundaries (due to the obvious safety concerns of such an action), they did want to try to address the noise issue somehow.  It was perceived by these communities that if at-grade crossings were replaced by grade-separated ones, the trains would no longer have to blow their horns, and thus the noise problem would be abated.

Because of these concerns about noise, it was imperative to take into account noise in any evaluation of potential highway-railway grade separation locations.  The Noise Factor, NF, is a quantification of the impact of train horn noise on communities.  The intent of this factor is to compare at-grade rail crossings on the basis of the extent to which surrounding communities are impacted by noise from the blowing of train horns.

The Noise Factor, NF, is defined by the following equation:

NF            =    (P0.5  x ADTT)  /  250                              (Eq. 1)


P0.5   = estimated population within 0.83 km (0.5 miles) of the railroad along the length that the whistle is blown and 0.42 km (0.25 miles) on either side of the railroad crossing (in thousands)

ADTT  =    average daily train traffic

   250 = constant to generate a number between 0 and 1[1]

As population density in the vicinity of the crossing is a component of this factor, NF for urban areas will be higher than NF for rural areas, as more people are impacted by the noise.


Factor 2:  Community Cohesion

Community cohesion is the sense of “oneness” of a community—the greater the cohesion, the greater this “oneness”.  There are three basic levels of community cohesiveness:  fully cohesive, non-cohesive, and semi-cohesive.  A fully cohesive community is one in which all sections of a community are interdependent, analogous to the interdependence of a married couple.  In Central Arkansas, the city of Conway is the best example of a cohesive community, as each half of the city on either side of the Union Pacific tracks needs the other for survival.  At the opposite end of the cohesiveness spectrum is the non-cohesive community in which each section of the community is independent of the others save for a common identity, analogous to adult siblings.  Cabot, Arkansas, is the best example of this type of community in Central Arkansas, as the east and west halves are polar opposites with respect to demographics, geography, and infrastructure.  A semi-cohesive community is one in which one section is dependent upon another like a child is dependent upon a parent.  Little Rock is a semi-cohesive community, as the southwest section is economically dependent upon the rest of the city. 

In the communities of Central Arkansas, railroad tracks are perceived to be reason for the lack of community cohesiveness, and highway-railway grade separations are perceived to be the solution to the cohesiveness problem.  Therefore, it was necessary to include community cohesiveness in any evaluation of potential grade separation locations.  The Community Cohesion Factor, CCF, is one way to quantify cohesiveness and assign priority to those communities with the least amount of cohesiveness.

The Community Cohesion Factor, CCF, is defined by the following equation:

CCF    =    1  -  (DA-B   /  DB-A)                                     (Eq. 2)


DA-B   = Desire to travel from A to B

DB-A   = Desire to travel from B to A

For a fully-cohesive community, CCF will equal 0, as DA-B will be equal to DB-A .  For a non-cohesive community, CCF will equal 1, and for a semi-cohesive community, CCF will fall somewhere in between.


Factor 3:  Delay

That motorists generally incur travel-time delay costs when waiting for trains to clear at-grade crossings is universally recognized.  However, the cost of time is not the only cost that may result from delaying motorists at rail crossings.  In communities without a highway-railway grade separation, the time it takes for a train to pass could be life threatening to a trauma victim on the way to a hospital emergency room.  Therefore, it is important to determine which rail crossings have higher vehicular delay when evaluating potential grade separation locations, due to the full range of social costs associated with delay at grade crossings.

Previous research by others has been incorporated into the Delay Factor (DF), which is the final derivative of a series of equations provided in NCHRP Report 288[2] based on the average train length, the average train speed in the crossing, the railroad crossing activation time, and the average annual train and vehicular traffic. (During development of the delay factor, problems were encountered in using some of the formulas reported in NCHRP Report 288.  After communicating these problems, changes were made to several of the report’s equations by one of its authors.[3] ) 

The DF is given by the following equation:

DF            =    TD / 480                                                    (Eq. 3.1)


TD = total vehicular delay (in hours)

TD    =    (V  x D) /   60                                            (Eq. 3.2)


V = number of vehicles delayed at crossing on an average day

V   =    P  x  AADT                                                    (Eq. 3.3)


AADT  =    annual average daily traffic

P = probability of vehicular delay during an average day

P    =    M   / 1440                                                        (Eq. 3.4)


M = number of minutes crossing is blocked on an average day

M   =    [ (L / S) (60) + (AT + 0.1667) ] x ADTT        (Eq. 3.5)


L = average train length (in km)

S = average train speed through crossing (in km/h)

60 = constant (minutes in an hour)

AT = typical RR crossing signal activation time (in minutes)

ADTT = average daily train traffic

From Eq. 3.2,

D = average duration of delay per vehicle delayed (in minutes)

D   =    M / ADTT / 2                                            (Eq. 3.6)


M = number of minutes crossing is blocked on an average day (from Eq. 3.5)

ADTT = average daily train traffic


Factor 4:         Accessibility

In urban areas, it is relatively simple to bypass an at-grade rail crossing that is occupied by a train.  At many urban locations, a grade-separated crossing is just a few blocks away.  In those locations, is it generally quicker to bypass to the grade separation than to wait for the train.  In these locations, then, the need for a grade separation, with respect to a motorist’s ability to get from one side of the tracks to the other when a train is occupying the crossing, is, for all intents and purposes, moot.

This is not the case in suburban and rural areas.  Outside of the urban cores, highway-railway grade separations are located miles apart, if they exist at all.  Thus, in these areas a motorist may have to drive many miles in order to bypass an at-grade rail crossing occupied by a train.  In these locations, it is not very practical to bypass to the nearest grade-separated crossing, as the time required to bypass would be much greater than the time required to wait for the train.  It is in these locations that accessibility becomes a significant issue.

The Accessibility Factor (AF) measures the difference in distance between a route using an at-grade crossing and a route using the nearest grade-separated crossing.  Note that the distance for both the at-grade crossing and the grade-separated crossing begins where the alternate route leaves and end where the alternate route rejoins.

AF       =    ( D1  -   D2  )   /      24                                (Eq. 4)


D1  =    detour distance -- distance using nearest grade-separated crossing

D2 =    through distance -- distance using the at-grade crossing

24 =     constant (maximum ratio difference)


Factor 5:         Connectivity

Because of the relatively high cost of grade-separated rail crossings versus at-grade crossings, it is important to locate these structures at the locations will they will receive the highest usage.  Intuitively, these locations will be on facilities that connect major trip generators, be they states, cities, or major employment centers within a community.  These facilities tend to be those of higher functional classification, such as arterials, expressways, and freeways.  As the functional classification of a facility is easy to determine, the Connectivity Factor (CF) uses functional classification as a measure of connectivity.  Specifically, CF is


CF       =    AADT / FC / 20,000                                 (Eq. 5)


FC       =    Functional Classification Value, where

FC = 1 for a freeway;

FC = 2 for an expressway;

FC = 3 for a principal arterial;

FC = 4 for a minor arterial;

FC = 5 for a collector;

FC = 6 for a local street

AADT = annual average daily traffic

20,000 = constant to generate value between 0 and 1


Factor 6:  Geographic Distribution

The jurisdiction of Metroplan consists of four counties in Central Arkansas: Faulkner, Lonoke, Pulaski, and Saline.  These four counties straddle the two primary transportation corridors in Arkansas: an east-west corridor along the Arkansas River, and a northeast-to-southwest corridor paralleling the geologic boundary between the mountain ranges of northwest Arkansas and the flat plains of southeast Arkansas. 

The east-west corridor connects Memphis and points east with Oklahoma City and points west.  Goods in this corridor are transported by barge via the Arkansas River; train, via a Union Pacific Railroad trunk line to the west that connects to the railroads of the eastern United States; automobile and truck via Interstate 40.  The northeast-to-southwest corridor connects Chicago and Saint Louis with Dallas and Houston and points south and west.  Transport through this corridor occurs via rail (a Union Pacific trunk line) and automobile and truck (via US 67 and Interstate 30).

There was a concern expressed by some in the Central Arkansas region of the possibility that all of the regionally funded highway-railway grade separations could be located in the same part of the region, or worse, along the same rail line.  In order to address that possibility, a Geographic Distribution Factor (GDF) was created as one of the evaluators of potential grade-separation locations.  The GDF is based on the ratio of the number of grade-separated crossings to the total number of crossings in a particular county along a particular rail line.  It should be emphasized that the factor not only considers each county independently, but also each sub-corridor within each county independently as well.  In Pulaski County, the two major corridors intersect, resulting in four sub-corridors—northeast, southwest, east, and west.  While there are an abundance of grade-separated crossing locations along the southwest corridor, there is a dearth of grade-separations along the northeast corridor.  Because of such differences within one county, it was decided that to truly reflect the geographic disparity of grade-separated crossing locations, each sub-corridor within a county would be evaluated independently of all others within that same county.

The GDF is defined as follows:

GDF   =    1  -  (  GS  /  TC  )                                     (Eq. 6)


GS = number of highway-railway grade separations per km (mile) along the railroad subdivision in the subject county

TC = total number of at-grade and grade-separated crossings per km (mile) along the railroad subdivision in the subject county

   1 =  constant to prioritize low number of grade separations


Factor 7:         Safety

None of the other factors can evoke the emotion, publicity, and urgency as the issue of safety.  Accidents, and especially fatalities, tend to be given extensive media coverage, thus evoking outrage in the community to fix the problem to prevent further loss of life and property.  Logically, then, safety has to be considered when evaluating possible highway-railway grade separation locations. 

The Safety Factor, SF, is given by Equation 7.

SF =    A / XD                                                             (Eq. 7)


XD = type of warning device at railroad crossing, where:

XD = 1 for Gates

XD = 2 for Flashing Lights

XD = 3 for Passive devices

A = predicted accidents per year (from FRA’s Railroad Crossing Accident Prediction Formulas[4] or comparable formulas). Data needed for the FRA formulas include:

1.      Type of warning device (passive, flashing lights, or gates)

2.      AADT (annual average daily traffic)

3.      ADTT (average daily train traffic)

4.      Peak trains per day (number of trains per day during daylight hours)

5.      Maximum timetable speed (in mph)

6.      No. of main tracks

7.      Highway paved? Yes or No

8.      Number of highway lanes

9.      Accident history (number of accidents in T years).



Although the factors were intended to generate numbers of approximately equal magnitude within the range of 0 to 1, neither outcome occurred; several values were slightly greater than one, and the magnitude of the values varied widely from factor to factor.  Thus, the inherent weighting of the various factors was uneven.  The values were normalized so that the value ranges for each factor were approximately equal, and all values were  within the intended range.

It should be noted, however, that the values for CCF were not normalized and that the CCF was removed from further analysis.  When the CCF values were computed, all values were equal to unity.  This was the result of the use of trips, rather than attractions, as a surrogate for “desire” in the equation.  Due to time constraints, this error could not be rectified.

The normalization of the raw values was a three-step process.  First, for each factor, the maximum value was set equal to one, and all other values for that factor were adjusted proportionally.  Second, the average of the values for each factor was set equal to the highest average of values for any factor, and all values were adjusted accordingly.  The third and final adjustment set the highest maximum value for any factor equal to one, and for all factors all other values were adjusted proportionally.  The resulting values ranged from 0 to 1 and  yielded factors of approximately equal weight.





This methodology produced results that were deemed to be equitable and useful to the many jurisdictions and other stakeholders that participate in the regional planning process in Central Arkansas.  By quantifying a range of variables, the methodology provided a valuable tool in prioritizing and planning for highway-railway grade separations and addressing issues not normally considered in a typical benefit/cost analysis.

Because the normalized values produce factors of equal weight, it is simple to apply multipliers to individual factors to produce customized weighting scenarios based on the priorities of the communities involved.  In the case of Central Arkansas, such a customized weighting scenario was ultimately applied, with SF weighted 50 percent, DF weighted 30 percent, the combination of AF, CF, and GDF weighted 20 percent, and NF weighted 0 percent. 

While the normalized values of the factors are useful in providing a snapshot comparison of different crossing locations, the raw values are useful in detecting trends at a particular location or for a particular factor.  This type of trend information is particularly useful when implementing long range comprehensive planning strategies.

Finally, although some of the constants used in the equations are based on data specific to Central Arkansas, this methodology is still applicable elsewhere.  The three-step adjustment process adopted after the raw numbers were computed ensures that each factor carries approximately the same weight as every other factor and that all values fall within the range of 0 to 1; thus, this methodology is applicable anywhere.




[1] Constants for NF, AF, and CF based on data specific to Central Arkansas metropolitan area.  Factors may yield results outside the intended range for other areas.

2 NCHRP Report 288, Evaluating Grade-Separated Rail and Highway Crossing Alternatives, Transportation Research Board, NRC, Washington, D.C., June 1987, “Appendix A, User’s Guide To The Decision-Making Framework”, pp. 19-38.

[3] In March 1996, Alan Brick-Turin of Frederic R. Harris, Inc. (Fairfax, Virginia), provided new equations for the calculation of crossing delay to replace those reported in NCHRP Report 288 on pages 34 and 35.  The new equation for T (amount of time the crossing is closed due to a single train movement) is T = [50  +  1.364 L (Sx + Se)] / 60 and the new equation for D (total delay experienced by vehicles delayed) is D = [(T / 2 + 10 ) N + (N / n)2] / 60.

[4] DOT/FRA/OS-87/10, Rail-Highway Crossing Resource Allocation Procedure –User’s Guide, Third Edition, U.S. Dept. of Transportation, Federal Railroad Administration, Washington, D.C., August 1987.