EVALUATION OF AASHTO RULES FOR IMPLEMENTATION OF CLIMBING LANES ON TWO-LANE HIGHWAYS

................................................................................................................. ii ACKNOWLEDGMENTS .......................................................................................... iii TABLE OF CONTENTS ............................................................................................ iv LIST OF FIGURES ................................................................................................... vii LIST OF TABLES .................................................................................................... viii CHAPTER 1 .................................................................................................................. 1 1.1 STATEMENT OF THE PROBLEM .............................................................. 1 1.2 JUSTIFICATION OF THE STUDY ............................................................... 2 1.3 ORGANIZATION OF THE RESEARCH ...................................................... 3 CHAPTER 2 .................................................................................................................. 4 2.1 CHARACTERISTICS OF TWO-LANE HIGHWAYS .................................. 4 2.2 TWO-LANE HIGHWAY PERFORMANCE MEASURES........................... 7 2.3 AASHTO GUIDELINES FOR CLIMBING LANE IMPLEMENTATION .. 9 2.4 STATE DEPARTMENT OF TRANSPORTATIONS’ GUIDELINES FOR CLIMBING LANE IMPLEMENTATION .............................................................. 11 2.5 CLIMBING LANE DESIGN CONSIDERATIONS .................................... 13 2.6 PERFORMANCE ASSESSMENT OF TWO-LANE HIGHWAYS WITH CLIMBING LANES ................................................................................................ 20 CHAPTER 3 ................................................................................................................ 21 3.1 DATA GENERATION ................................................................................. 21

This thesis will seek to evaluate parts of the guidelines for the implementation of climbing lanes on two-lane highways as promulgated by AASHTO, 2011.
Evaluation will probe the necessity for, and the sufficiency of, climbing lanes when developed per the guidelines' recommendations. The thesis focuses only on Class I rural two-lane highways, with segment lengths that result in a 30-mph speed decrease on upgrades ranging from 6% to 10%, which connect to flat terrain on the grade's approach and departure. The study assumes for the most part uniformly distributed variables.

JUSTIFICATION OF THE STUDY
Climbing and passing lanes are introduced on two-lane highways in the aim of increasing their performance levels, breaking down platoons, enhancing travel speed and safety (Roess et al., 2011). In practice, most state Departments of Transportation (DOTs) base their decisions about implementing climbing lanes on old guidelines by the American Association of State Highway and Transportation Officials (AASHTO).
These guidelines are a set of general conditions in which to consider a climbing lane.
They have existed for over twenty years now and remained unchanged over at least four different editions of "A Policy on Geometric Design of Highways and Streets" (AASHTO, 1994(AASHTO, , 2001(AASHTO, , 2004(AASHTO, , 2011. While these guidelines are useful, they do not account for the combination of variables known by the state-of-the-practice to determine specifically two-lane highway level of service (LOS).
This research shall determine if the AASHTO guidelines still apply in the face of enhanced methodologies to determine performance on two-lane highways or they need much improvement. If the guidelines prove questionable or inadequate, it is hoped that the research results may spur future research toward the development of enhanced guidelines for the implementation of climbing lanes on two-lane highways.
Thus, resulting enhanced access to, and mobility of, goods and persons may promote enhanced economic, defense and transportation systems' performances. In addition, the resulting enhanced maneuverability may promote enhanced transportation system safety. If the guidelines prove to be adequate, then surely transportation operators, Departments of Transportation (DOTs) in particular, would benefit from enhanced confidence in their use.

ORGANIZATION OF THE RESEARCH
This thesis comprises five chapters. The first chapter gives a brief overview of the purpose and the goals of the study. The second is a brief review of the literature. It To ensure safety, every point on a highway must provide the safe stopping sight distance to drivers at the selected design speed. Passing sight distance constitutes the minimum sight distance required to perform safely a passing maneuver. Not every point on the highway needs provide the safe passing sight distance contrarily to stopping sight distance. Where passing sight distance is insufficient to allow for passing safely, passing should simply be disallowed. Especially two-lane highways rely on this safety measure, as passing maneuvers occur in the opposing traffic lane.
Hence there must be "no-passing" zones or highway markings that prohibit passing where unsafe (Roess et al., 2011).
Effective methods to break down platoons, formed through lack of passing opportunities, or prevent their formations include turnouts, passing lanes at given intervals in each direction and climbing lanes on upgrades. To provide an exhaustive overview, the study revisits the fundamentals of the three methods, even though climbing lanes are the sole focus of this research.
Turnouts provide sufficient room for slower moving vehicles to pull out of the through traffic, and stop if necessary, to permit following vehicles to pass or provide room for emergency stops. The location of turnouts depends on the type of facility.
They are mostly provided where passing opportunities are limited, a high frequency of slow-moving vehicles exists and cost for a full auxiliary lane would be inappropriate for the effect it causes (Washington State DOT, 2017). Turnouts are widened shoulder areas on two-lane highways, that are rather short, generally less than 625 ft (Transportation Research Board, 2010).
In general, climbing lanes allow for passing slow trucks through the provision of a short added lane to the right side of upgrades for the exclusive use of trucks.
Climbing lanes are necessary, where slower moving vehicles like trucks or heavy, agricultural vehicles impede traffic flow as following drivers of faster vehicles have limited to no opportunities to overtake. In these situations, faster vehicles must follow closely behind slower ones until they are able to pass, resulting in platoon formations.
In consequence, performance and safety may deteriorate (AASHTO, 2011;Polus and Reshetnik, 1987). where passing opportunities are unavailable or very limited over a long stretch of highway" (Arizona DOT, 2015). Furthermore, a climbing lane is defined as an "additional lane on steep upgrades to facilitate the passing of trucks and slow moving vehicles whose speed drops because of the sustained grade rather than a lack of passing opportunity over a long stretch of highway" (Arizona DOT, 2015).

TWO-LANE HIGHWAY PERFORMANCE MEASURES
To convey the quality of traffic flow on two-lane highways, the Highway Capacity Manual's (HCM) Level of service (LOS) analysis incorporates three measures of effectiveness; ATS (the average travel speed), PTSF (the percent-time-spentfollowing) and PFFS (the percent free flow speed) (Transportation Research Board, 2010). In effect, the HCM conveys that travel speed and overall maneuverability, the both, can be indicators of highway performance or quality of service.
The HCM distinguishes between two classes of rural two-lane highways.
Class I highways are relatively high-speed roads, arterials, primary highways that afford mobility. ATS and PTSF define the LOS of these highways. On Class II highways, drivers do not expect to travel at high speeds. These roads are access routes to Class I facilities or serve short trips. Only PTSF determines their LOS. The Florida Department of Transportation (FDOT) defines an additional class, Class III highways.  DOT, 2015). The graphs classify into LOS "C" or better, LOS "D" and LOS "E" or worse for undeveloped and developed segments, respectively. Per the definitions provided for undeveloped and developed segments, it is safe to conclude that they represent Class I and II HCS two-lane highways, respectively.
"The undeveloped case applies to rural highways where relatively high speeds of travel are expected and that are major intercity routes, primary highways connecting major traffic generators, daily commuter routes, primary links in state/national highway networks or any facility that serves long-distance trips" (New York State DOT, 2015).
"The developed case applies to all highways with an urban Functional Class. It also applies to some highways with a Functional Class of rural, but that serve developed areas (for example, a village), or a scenic or recreational area where speeds are expected to be lower" (New York State DOT, 2015).
NYSDOT does not consider the potentially negative impacts of unusually intense truck traffic on LOS; input variables to the graphs only include the AADT, expressed in vehicle per hour rather than passenger cars per hour, and the prevailing 85 th percentile flow speed.

AASHTO GUIDELINES FOR CLIMBING LANE IMPLEMENTATION
Per the guidelines, a consideration of climbing lanes on two-lane highways is justified by the below-stated conditions.  DOT, 2017). Hence, WSDOT replaces "and" by "or" in the consideration of performance warrants, which become alternates. In addition, congestion (with platoon creation as possible manifestation) that may ensue from exceptional conditions, outside of warrants, may also serve as warrants.
Unfortunately, as with NYSDOT, WSDOT does not specify these exceptional conditions. The startup location of the climbing lane depends on the trucks' speeds when approaching the grade and on the sight distance limitations at the approach. When conditions permit, a climbing lane can be introduced beyond the beginning of an upgrade if the speed of trucks will not immediately reduce to an intolerable speed for following drivers up the grade. Still, no sight distance restriction or other limitations to the speed must be at play (AASHTO, 2011).
The critical length is defined as "the maximum length of a specific upgrade on which a loaded truck can operate without an unreasonable reduction in speed" (Illinois DOT, 2016). If the length of the upgrade is longer than the critical length, trucks are at risk to attain speed reductions and operations that are unacceptable (AASHTO, 2011).
Typically, drivers do not tolerate the speed reduction of trucks on an upgrade, when the critical length of grade is reached, i.e. truck operating speed is reduced about 10 mph (New York State DOT, 2017b).
SDDOT states in general that an upgrade should not exceed 2,000 ft to guarantee acceptable operations. SDDOT construes the critical length of grade to equal 2,000 ft, regardless of grade percentage. On grades longer than the critical length, consideration of an additional lane should be made (South Dakota DOT, n.d.).
" Fig. 2 shows the critical lengths of grades associated with varied upgrade slopes and for varied acceptable speed reductions given a representative truck of 200 lb/hp and a grade entry speed of 70 mph. The 10 mph speed-reduction graph needs to be paid special attention as this represents the general guideline for estimating critical length. In the past a speed reduction of 15 mph was used to determine this length of grade, however, the crash involvement of trucks experiences a significant increase when truck speed reduction is higher than 10 mph. This led to the recommendation to determine the critical length of grade by using the 10 mph curve" Nevertheless, if the upgrade in question immediately follows a previous upgrade, the truck speed may already be lower than the design speed. In this case, the critical length of grade will be smaller. This consideration applies equally to an upgrade with an immediately approaching downgrade, where it is known that truck drivers will accelerate to get a "running start" at it. The critical length of grade will be longer than with a level entrance in that case (Harwood et al., 2003). This research will only focus on level terrain approaches and departures to upgrades. The critical length of grade is derived as the length of a tangent grade. An approximate equivalent length of tangent grade must be used where a vertical curve is part of a critical length of grade. When there are vertical curve tangents with only positive or only negative grades and the algebraic difference in grades is not too great, the measurement of critical length of grade is derived between the vertical points of intersection (VPI). Where vertical curves with positive and negative tangents are involved, particularly where the algebraic difference in grades is appreciable, about one-quarter of the vertical curve length may be considered as part of the grade under consideration (AASHTO, 2011). As this research will only consider grade connections to flat terrain, where the algebraic difference in grades will have an average to small value, the former measurements of critical length applies.
AASHTO recommends that the climbing lane should be developed through a tapered entry section with a ratio of 25:1 and a length of at least 300 ft (AASHTO, 2011 (Maryland DOT, n.d;Missouri DOT, n.d;Colorado DOT, 2005;New Jersey DOT, 2015;New York State DOT, 2017b;Texas DOT, 2018). NJDOT, NDOT and SDDOT require 300 ft, whereas Michigan DOT even requires 500 ft long entry tapers (South Dakota DOT, n.d; Michigan DOT, 2011;Nebraska DOT, 2011;New Jersey DOT, 2015).
To be effective, the full-width climbing lane itself, excluding the tapers, should be at least 1,000 ft long per IDOT (IDOT, 2016). SDDOT states that the length of the full-width climbing lane should at least be 0.5 mi, 2,640 ft, (South Dakota DOT, n.d.) and NDOT demands 1,200 ft (Nebraska DOT, 2011). A study from ADOT states that the majority of passing and climbing lanes on two-lane highways in the state of Arizona have a length of 0.5 mi to more than 1.0 mi (Arizona DOT, 2015).
AASHTO recommends that lane and shoulder widths be maintained for roadway segments with climbing lanes (AASHTO, 2011). Still, "Whenever possible, maintain a shoulder width equal to that of the adjacent roadway segments (preserve shoulder width continuity). On two-way two-lane highways, the shoulder may be reduced to 4 ft. If the shoulder width is reduced to 4 ft document the reasoning for the decision in the design parameter sheets. If the shoulder width is reduced to less than 4 ft, a design analysis is required." (Washington State DOT, 2017) WSDOT further limits entrance speed for trucks to 60 mph, regardless of the posted speed limit, for assessing whether the AASHTO speed performance warrant is met (Washington State DOT, 2017). Trucks' approach speed at the grade should be estimated at 55 mph (Illinois DOT, 2016). WSDOT estimates approach speeds of 60 mph (Washington State DOT, 2017). AASHTO hints to the varied entry speeds as indicative of the variation in design speed for varied states.
Further, AASHTO legitimizes the practice of assuming varied entry speed to its critical length graphs, Fig. 2. Although Fig. 2 assumes a grade entry speed of 70 mph, it can be applied to any design speed (Montana DOT, 2007). The graphs can be viewed as entry speed insensitive and indicative only of speed reduction. Per AASHTO, if there is a difference in initial and minimum tolerable speeds because of a lower design speed, the critical length will still be the same for the 10 mph speed reduction with a design speed of 70 mph (AASHTO, 2011). Accordingly, a truck will travel the same critical length to experience a speed reduction of 10 mph, whether starting at 70 mph and ending at 60 mph or traveling at 60 mph and ending at 50 mph.
The climbing lane shall be extended beyond the crest, at which point trucks could gain a speed of at least 40 mph and that only varies up to 10 mph to the speed of the vehicles in the normal lane. As trucks need a long distance to accelerate to the  (Maryland DOT, n.d;South Dakota DOT, n.d;Colorado DOT, 2005;Nebraska DOT, 2011). MDT states that the exit taper rate should vary, depending on the design speed. Hence, a 50 mph design speed would require a ratio of 50:1, a design speed of 60 mph would require a 60:1 ratio and a 70 mph design speed would require a 70:1 ratio for the exit taper (Montana DOT, 2007).
WSDOT recommends to "begin climbing lanes at the point where the speed reduction warrant is met, and end them where the warrant ends, for multilane highways, and 300 ft beyond for two-lane highways. Consider extending the auxiliary lane over the crest to improve vehicle acceleration and sight distance" (WSDOT, 2018).

CLIMBING LANES
To assess the impact of climbing lanes on the operation of two-lane highways, the HCS analysis proceeds as with a passing lane in level or rolling terrain. Data analysis will determine whether the guidelines adequately predict the necessity and sufficiency of climbing lanes for the cases simulated given the HCS 7 predictions.
To classify or interpret the performance enhancements reached due to the implementation of climbing lanes for simulated cases, the study gauges an acceptable LOS on two-lane highway, LOS "C". Furthermore, it assesses the efficacy of the guidelines by interpreting their classification rates of sampled scenarios for the necessity and sufficiency of climbing lanes in view of the HCS 7 performance gains they achieve.

Highway Capacity Software (HCS 7)
For the generated scenarios, HCS 7 establishes the performance of hypothetical  percentage values in the range between 3.0% and 9.9%, the highest percentage used in study equals 9.9%, even though displayed at times in the following tables at its rounded value of 10%. = 300 + + 250

Input Variables
Main variables of interest to the software include slope, directional split, total flow, percentage of trucks and no passing zones.  Flow on two-lane highways has a high variance, as high and low flows are combined into one distribution (Gerlough and Huber, 1975 The proportional factor between the opposed flows (D-Factor) sways usually the interval of 0.65 to 0.85 for two-lane highways (Roess et al., 2011). The truck percentage varies between 2%, the default value of the HCS, and 20% considered a high percentage by NYSDOT, 2017a, and representing the upper limit of the Highway Capacity Manual (Transportation Research Board, 2010).

Random Generation
A random number generator derives performance-impacting input variables for varied scenarios on two-lane highways. These probabilities combine with the known distributions of the input variables to identify them uniquely. Variables are generated singly assuming their independence and an input entry covariance matrix with zero (0) entries. Random number generation is the generation of a sequence of numbers that is uniformly distributed and cannot be predicted reasonably better than by a random chance. A simple example is simulating the toss of a die (Hoover and Perry, 1989).
Microsoft Excel has two functions to generate random numbers, the RAND( ) and RANDBETWEEN() function. These functions are uniform. The study deploys both the functions to generate input values for the experiments, per their ranges stated in Table 3. The RAND( ) function generates a random decimal number between 0 and 1 and is used for the values of the D-Factor, no-passing zone and slope percentage. To It may well be that the study artificially restrained the D-Factor to sway a limited range. However, the relevant literature duly references the range utilized, although the same documents other ranges that would not necessarily limit opposing flow to its upper bound.
Further, to bypass flow conversion, the analysis directly inputs generated flows in pcph, kept within their directional capacity limits, as HCS 7 flow entries in vph.
Overshooting the required range of data input ensured its adequate coverage of the coveted range in pcph. Where generated flows within HCS 7, in pcph, exceed directional capacity within a scenario, the analysis eliminates this scenario and its results from further considerations. Table 8 in Appendix D documents every scenario's ATS and PTSF directional flows. Scenarios with flows exceeding the limits of 1,700 pcph in one direction or 1,500 pcph in the other (assuming higher opposite directional flows), or the total capacity of 3,200 pcph are marked with a "0" in the last column of Table 8, Capacity Compliance. Scenarios with flows within the capacity boundaries are marked with a "1".
Switching the direction of generated flow, upgrade versus downgrade, ensures that high flows are not limited to the analysis direction. Half of the software runs generate flows in the analysis direction and the other half in the opposite direction, leading to two different data streams. The analysis fused both the streams using a random scenario/record sorting method in Excel.

DESCRIPTIVE STATISTICS FOR GENERATED FLOW DATA
Flow input values to HCS 7 as randomly generated for the ATS analysis directional flows sway the range of 12 pcph to 1,697 pcph, with an average flow of 775 pcph.    Prior to the analysis, certain entry values need set in the Input Data Panel to their scenario values for both, the analysis and opposing directions of travel. This, in addition to using the HCS 7 default values cited in Section 3.1.2 above. The terrain must be set to "Specific Grade", which will enable the functional buttons for "grade percentage" and "length of slope". PHF, truck percentage, no-passing zone percentage and Segment Length need to be input. The study assumes that the climbing lane length is equal to the segment length, which is the length of the section analyzed.
The segment lengths obtain from Table 2. Flows obtained for the analysis and opposing directions need also be set accordingly in the Input Data Panel. A listing of these values is enclosed in Table 7 in Appendix B.   Further execution results can be obtained from the Free-Flow Speed Panel, as displayed in Fig. 9. The average travel speed, ATSd, helps assess/confirm whether AASHTO's Guideline 3-a holds true. HCS 7 reports on the input data and results of software runs. Appendix C 1-5 show example reports for scenarios 68, 197, 307, 347 and 501.

DATA PROCESSING AND DATA ANALYSIS PROCEDURES
A preliminary review of the AASHTO guidelines on the implementation of climbing lanes (AASHTO, 2010) helps determine the logical clauses to test toward assessing their efficacy. If efficient, a climbing lane should be both, necessary and sufficient. Assuming necessity, for climbing lane implementation or nonimplementation, then for a prevalence of cases the below logical clause should hold true.
1. If the guidelines are satisfied and no climbing lane exists, then unacceptable performance must prevail (proper classification).
2. If the guidelines are not satisfied and a climbing lane exists, then unacceptable performance must prevail (proper classification).
Further, sufficiency of the guidelines implies that the below logical clauses should hold true as well.
3. If the guidelines are satisfied and a climbing lane exists, then acceptable performance must prevail (proper classification).
4. If the guidelines are not satisfied and a climbing lane does not exist, then acceptable performance must prevail (proper classification).
Logical Clause 4 is not as restrictive as earlier clauses for conditions of low flows or given the violation of flow guidelines, as AASHTO considers the limited economic impact of engendered delays by such flows to advice against climbing lane implementation.
The analysis views the AASHTO guidelines as a scenario classifier for the efficacy of the implementation of a climbing lane. It tests logical clauses determining the necessity and the sufficiency of the AASHTO guidelines using the HCS 7 performance predictions for the quasi-representative sample scenarios earlier generated. If the clauses are untrue, the guidelines misguide and thus misclassify on the efficacy, necessity or sufficiency, of climbing lanes. If true, the guidelines classify correctly on the same. The analysis further determines a running percentage of the AASHTO guideline classification rates, proper classification and types I or II classification error rates, with large enough a sample size to stabilize the rates achieved. The extent of these rates bears witness to the efficacy, necessity and sufficiency, of the AASHTO guidelines taken verbatim as warrants, a rather typical DOT practice.
An acceptable level of service on two-lane highways must be gauged to classify or interpret the performance enhancements reached due to the implementation of climbing lanes for simulated cases. In this study, LOS "C" constitutes the minimum acceptable level for the research question.  for each scenario the satisfaction of the Logical Clause(s) 1, 2, 3 and 4 earlier specified, Section 3.3, which establish the necessity and the sufficiency of a climbing lane. As with Guidelines 1, 2 and 3, entries of "0" and "1" convey the dissatisfaction and the satisfaction of a logical clause, respectively. Satisfaction of Logical Clause 1 entails 1) the satisfaction of overall AASHTO guidelines ("1" entry, Column 24) AND 2) an unacceptable performance without a climbing lane ("0" entry, Column 25).
Scenarios that satisfy Logical Clause 1 generate an entry of "1" in Column 27.
Otherwise "0" entry registers in this column. Further, the satisfaction of Logical Clause 2 entails 3) the lack of satisfaction of overall AASHTO guidelines ("0" entry, Column 24) AND 4) an unacceptable performance with a climbing lane ("0" entry, Column 26).
Scenarios that satisfy Logical Clause 2 generate an entry of "1" in Column 28.
Otherwise a "0" entry registers in this column. Further, the satisfaction of Logical Clause 3 entails 1) the satisfaction of overall AASHTO guidelines ("1" entry, Column 24) AND 2) an acceptable performance with a climbing lane ("1" entry, Column 26).
Scenarios that satisfy Logical Clause 3 generate an entry of "1" in Column 29.
Otherwise a "0" entry registers in this column. Finally, satisfaction of Logical Clause 4 entails 1) the lack of satisfaction of the overall AASHTO guidelines ("0" entry, Column 24) AND 2) an acceptable performance without a climbing lane ("1" entry, Column 25) Scenarios that satisfy Logical Clause 4 generate an entry of "1" in Column 30.
Otherwise a "0" entry registers in this column.
Lastly Proper scenario classification ensues from recommendations for climbing lane implementation that lead to beneficial impacts for upgrade segment performance.
The proper classification rate thus equals the ratio of the number of scenarios that satisfy logical clause 1 and 3, when the guidelines are satisfied, plus the number that satisfy logical clauses 2 and 4, when the guidelines are not satisfied to the total number of scenarios analyzed.

FINDINGS / RESULTS
To assess the efficacy of the AASHTO guidelines, the HCS 7 was run 601 times with the randomly generated scenarios of Table 7 in Appendix B. In 197 scenarios, the flows exceed two-lane highway capacity, which leaves 404 scenarios to analyze. The wording "total number of studied scenarios" refers in the following to these 404 scenarios.

CLASSIFICATION RATES AND GUIDELINES' EFFICACY
The classification errors hinted above in Section 4.1 and 4.2 do not encompass those resulting from the interplay of necessity and sufficiency. Although AASHTO guidelines predict adequately necessity and sufficiency, it is not evident that they can predict their interplay successfully. To do so, the analysis derives proper and erroneous classification rates.   Sixty-six (66) Figure 12 displays the stabilization course of the type I error rate.

Figure 12: Type I Error Rate
One hundred ninety (190) Figure 13 displays the stabilization course of the type II error rate.    1. Note that the adjustment factor for level terrain is 1.00, as level terrain is one of the base conditions. For the purpose of grade adjustment, specific downgrade segments are treated as level terrain.  1. Note that the adjustment factor for level terrain is 1.00, as level terrain is one of the base conditions. For the purpose of grade adjustment, specific downgrade segments are treated as level terrain. 1. Note that the adjustment factor for level terrain is 1.00, as level terrain is one of the base conditions. For the purpose of grade adjustment, specific downgrade segments are treated as level terrain.  Bicycle LOS Score, BLOS 6.69 Bicycle LOS F Notes: 1. Note that the adjustment factor for level terrain is 1.00, as level terrain is one of the base conditions. For the purpose of grade adjustment, specific downgrade segments are treated as level terrain.