Skip to contentUnited States Department of Transportation - Federal Highway Administration FHWA Home
Research Home
REPORT
This report is an archived publication and may contain dated technical, contact, and link information
Publication Number: FHWA-HRT-11-067
Date: June 2012

 

Field Evaluation of A Restricted Crossing U-Turn Intersection

CHAPTER 3. FINDINGS

OPERATIONAL OBSERVATIONS

Traffic counts for the peak morning hour at the RCUT are shown in figure 7.

Figure 7. Illustration. Traffic count for peak hour at RCUT intersection. This illustration shows the intersection traffic counts of the peak hour that was observed at the restricted crossing U-turn (RCUT) intersection. From the southbound approach on U.S. 15, 34 vehicles turned right onto the minor road, 613 vehicles continued south, and 24 vehicles turned left onto the minor road. From the northbound approach on U.S. 15, 120 vehicles turned left, 472 vehicles proceeded north, and 11 vehicles turned right. From the minor road approach on the west side of the highway, 242 vehicles turned right, 10 vehicles crossed the highway (via U-turn) and continued traveling east, and 3 vehicles turned left (via the U-turn). No observations were recorded from the minor road approach on the east side of the highway.

Figure 7. Illustration. Traffic count for peak hour at RCUT intersection.

Right Turn at the RCUT

All right turns from U.S. 15 Business onto U.S. 15 were catalogued between 8 a.m. and 9:20 a.m. on September 24, 2009, except for a 20-min gap beginning at 8:12 a.m. due to an equipment failure. Complete records were obtained on 254 vehicles that turned right. One additional vehicle turned right but broke down before merging with through traffic. That vehicle stopped on the right shoulder beyond the end of the right-turn acceleration lane and was not included in the analyses.

Of the 254 right-turning vehicles, 248 traveled freely through the right-turn slip lane, where freely means that no vehicle ahead of them greatly impeded speed through the right turn movement. In six cases, merging vehicles were impeded by vehicles ahead that either stopped or slowed in the slip lane.

Most of the right-turning vehicles were destined to continue south on U.S. 15. Vehicles destined for the east were completing an indirect through movement, and vehicles destined for the north completed the indirect left-turn movement.

Conflicts

During the primary observation period, only two conflicts were observed involving vehicles turning right from the minor road. Both conflicts were judged to be of low severity. One conflict appeared to be the result of a closely spaced platoon of vehicles merging together from the rightturn acceleration lane. The other conflict appeared to be the result of speeding on the main line, which may have caused the merging driver to misjudge the time available for the merge.

In one of these conflicts, a vehicle approached a platoon of merging vehicles and braked lightly for 2.3 s. The vehicle that braked changed to the left lane and never closed to less than 2 s timeheadway. After changing lanes, the braking vehicle overtook the last vehicle in the platoon, but that vehicle remained in the merge lane as it was overtaken.

In the second conflict, a southbound through vehicle approached the intersection at approximately 70 mi/h. It braked for a merging vehicle that moved into the right through lane behind a truck. The merging vehicle left about 1 s time-headway between itself and the truck. The braking vehicle did not change from the right lane even though there were no vehicles in the left lane. The braking event had a duration of about 1.4 s, but the brake lights were masked by obstructions at the end of the braking event, so the braking event could have been as long as 2.2 s. The rate of deceleration was judged to be low, as there was no discernable body sway associated with the braking.

Acceleration Lane Use

The majority of vehicles bound for destinations south of the RCUT intersection merged into the right through lane. The majority of vehicles destined to continue on minor road or to use the U-turn to complete a left-turn movement merged into the left lane in one continuous movement. Table 2 shows that 83 percent of vehicles destined for the U-turn merged directly into the left lane, whereas 7 percent of vehicles that continued south merged directly into the through lanes.

Table 2. Location where vehicles completed right-turn merge as a function of lane merged into and destination.

Merge Location U-Turn U-Turn South South
Left Entry
Lane
Right Entry Lane Left Entry Lane Right Entry Lane
Cross gore 2 2 0 9
End of gore line 4 0 14 104
Midway 4 0 1 71
End of merge area 0 0 3 40

Table 2 also shows that most of the drivers making right turns utilized at least a portion of the acceleration lane. That is, most drivers travelled at least the first 140 ft in the acceleration lane. Only about 5 percent of drivers merged into the through lanes before that point, and neither of the conflicts discussed above was associated with an early merge. Two vehicles that crossed the edge line that designated the end of the acceleration lane are included in the end of merge area category in table 2.

Lags

A summary of means, minimums, and standard deviations in lag between the entry of rightturning vehicles into the through lanes and the next vehicle that was in the right lane at the time the vehicle entered is shown in table 3. Only lags less than 11 s were included in the summary. The short minimums shown in the table occurred when the through vehicle changed lanes to accommodate the merging vehicle. These data suggest that most drivers use the acceleration lane to achieve an acceptable lag.

Table 3. Lag between right-turning vehicles and arrival of next through vehicle (seconds).

Merge Location Mean Count Minimum Standard Deviation
Cross gore 4.4 7 2.3 2.0
End of gore line 5.7 58 0.9 2.7
Midway 5.5 27 1.1 2.1
End of merge area 5.3 23 0.0 3.5

Weaving

In 39 cases, a vehicle in the right through lane shifted to the left lane in apparent response to the presence of a vehicle in process of making a right turn. Thus, induced lane changes were associated with 15 percent of the merges.

Travel Time

Because of the low volume of through and left-turn movements and because reliable travel time estimates require more observations than were obtained during the peak travel period, travel time measurement was extended into the afternoon. Table 4 shows a travel time summary for through and left-turn movements made between 8 a.m. and 1:43 p.m.

Table 4. Travel time for left-turn and through movements (seconds).

Movement N Mean Standard Deviation Minimum Maximum
Through 28 83 14 68 131
Left 29 80 10 64 112

RCUT U-Turn

Between 6:57 a.m. and 12:21 p.m., 42 vehicles were observed making a U-turn at the southern end of the RCUT.

Conflicts

Only one of the 42 U-turns resulted in a conflict. In that case, a full-sized car crossed directly into the right lane and did not use any part of the acceleration lane. A northbound vehicle in the left lane braked in response to this vehicle turning in front of it. However, the northbound vehicle’s brake lights did not illuminate before the turning vehicle was clear of the left lane.

The lag between the first incursion of the turning vehicle into left lane and the arrival of the northbound vehicle at the point of incursion was 5.9 s. The brake lamp illumination lasted 0.4 s. The conflict was judged to be of low severity.

Acceleration Lane Use

It has been asserted that drivers do not use the acceleration lane at RCUT U-turn openings. [3] This was true of 30 of the 42 U-turn movements that were observed. However, in 18 of the 30 cases where the acceleration lane was not used, there was no northbound vehicle in the left lane that was within 11 s, and of those 18 merges where there was no approaching vehicle in the left lane, there were only 5 cases where there was a vehicle in the right lane that was within 11 s. In two cases in which the acceleration lane was not used, vehicles in the through lanes changed lanes to accommodate the merging vehicle. Thus, in the majority of the cases where the acceleration lane was not used, there was no compelling reason to use it. Furthermore, 2 of the 18 vehicles that did not use the acceleration lane were too large to stay within the turning radius of the U-turn. These were a school bus and a large van. In 10 of the 12 cases in which the acceleration lane was used, there was a compelling reason to do so. In 10 cases, the drivers allowed vehicles in the left and right through lanes to pass as they accelerated to highway speeds.

Lags

Short lags between merging and through vehicles were not observed. Lags between merging and through vehicles in the left lane are summarized in table 5. Lags between merging vehicles and through vehicles in the right lane are summarized in table 6. The shortest lags were observed between through vehicles and merging vehicles that accelerated to highway speed before merging into the through lanes.

Table 5. Lag between entry of U-turn vehicle into left through lane and arrival of next vehicle in left lane.

Merge Location Mean N Minimum Standard Deviation
Cross gore 9.7 17 5.3 3.8
End of gore line 0
Halfway 4.5 1 4.5

— Indicates insufficient cases to compute statistic.

Table 6. Lag between entry of U-turn vehicle into left through lane and arrival of next vehicle in right lane.

Merge Location Mean N Minimum Standard Deviation
Cross gore9.520 2.9 3.6
End of gore line7.23 3.7 4.0
Halfway5.05 1.3 2.4

Weaving

In two cases in which the acceleration lane was not used, vehicles in the through lanes changed lanes to accommodate the merging vehicle.

Conventional Intersection

Traffic counts for a peak hour at the conventional intersection are shown in figure 8.

Figure 8. Illustration. Peak hour traffic count for conventional intersection. This illustration shows the intersection traffic counts for the peak hour that was observed at the conventional intersection. From the southbound approach on U.S. 15, 47 vehicles turned right, 840 vehicles continued southbound, and 1 vehicle turned left. From the northbound approach on U.S. 15, 
4 vehicles turned left, 587 vehicles continued northbound, and 6 vehicles turned right. From the eastbound approach on the minor road, seven vehicles turned right, four vehicles continued across the highway, and two vehicles turned left. From the westbound approach on the minor road, 21 vehicles turned right, 3 vehicles continued across the highway, and 21 vehicles turned right.

Figure 8. Illustration. Peak hour traffic count for conventional intersection.

Because of the low peak hour volume of vehicles emerging from North Franklin, several hours of operation were analyzed. Data were reduced from video recorded on Thursday, October 8, 2009, between 6:53 a.m. and 1:43 p.m. Data from all 115 movements of vehicles originating from North Franklin were tracked and characterized. These movements should be comparable to movements originating from the RCUT intersection.

Table 7. Movement of vehicle originating from North Franklin Road.

Destination Movement Count
South Right 79
East Through 22
North Left 14

At a conventional intersection with a divided highway, drivers may make through or left-turn movements in one or two stages. In a two-stage movement, drivers first proceed across one direction of traffic to the median opening and then wait there for the opportunity to cross or merge with traffic in the other direction. Alternatively, drivers may look for simultaneous acceptable gaps in both directions and make a single-stage movement. As shown in table 8, the majority of through and left movements were made in a single stage.

Table 8. Single-stage versus two-stage crossings at conventional intersection.

Destination Crossing Type Count
East (through) Single stage 15
East (through) Two stage 7
North (left) Single stage 11
North (left) Two stage 3

Conflicts

Six conflicts involving vehicles merging from North Franklin were identified as follows:

Lags

Table 9 summarizes lags observed for vehicles originating from the west on the minor road. The first column of the table lists the lane entered, and the second column lists the turning movement. In this table, lag is between the entry of a vehicle into a lane and the arrival of another vehicle that was in the given lane at the beginning of the movement. For vehicles turning right into the right lane, it does not consider vehicles in the left lane. For vehicles turning left into the left northbound lane, it does not consider lags to the arrival of vehicles in the right northbound lane. No vehicles turned right into the left southbound lane when other vehicles were approaching in the left lane, and no vehicles turned left into the right northbound lane when vehicles were approaching in that lane.

Table 9. Lags for vehicles originating from conventional intersection by lane and destination.

Lane Movement Mean Median N Minimum Maximum
SB right Right turn 8.3 8.3 34 4.5 10.9
SB left Right turn 0
SB right Through 7.6 7.5 11 5.6 10.2
SB left Through 9.0 9.0 2 7.8 10.1
NB left Through 6.6 6.4 7 4.5 9.8
NB right Through 7.3 7.4 16 3.7 10.5
SB right Left 6.9 7.1 8 3.6 9.8
SB left Left 7.0 6.6 8 5.1 10.5
NB left Left 7.2 7.2 3 5.4 8.9
NB right Left 0

SB = Southbound; NB = Northbound.
– Indicates insufficient observations to compute statistics.

Table 10 shows lags for the next vehicle in the adjacent lane: the left lane for vehicles making the right turn into the right lane and the right lane for vehicles making the left turn into the left lane. The lags for left-turning vehicles may be of concern, given the difference in speed between vehicles in the through lanes, where the speed limit is 55 mi/h and the difference in speed between left-turning vehicles that have no acceleration lane and a tight turning radius.

Table 10. Lags to next vehicle in adjacent lane.

Movement N Mean Median Minimum Maximum
Right turn 8 7.5 6.8 4.6 11.0
Left turn 5 2.9 3.9 1.0 4.1

Weaving

No lane changes were observed in response to drivers making a through movement from the minor road. Drivers in the through lanes were observed to change lanes in response to 10 of 79 right turns from the minor in road. Drivers in the northbound through lanes were observed to change lanes in response to 2 of 14 left turns.

Travel Time

Travel times for minor road left-turn and through movements are summarized in table 11.

Table 11. Travel times for through and left-turn movements at conventional intersection (seconds).

Movement N Mean Standard Deviation Minimum Maximum
Through 22 19 9 9 36
Left 14 28 11 14 51

CRASH ANALYSES

All the crash analyses are based on nine RCUT intersections that were deployed in Maryland between 1988 and 2003. Six of these intersections were deployed on U.S. 15, and three were deployed on U.S. 301. The relevant portion of U.S. 15 is a rural four-lane divided highway that runs from just north of Frederick, MD, in the south to the Pennsylvania State line in the north. In 2009, average annual daily traffic (AADT) along this stretch of U.S. 15 ranged from about 45,000 vehicles per day near Hayward Road down to about 20,000 vehicles per day near U.S. 15 Business. The relevant portion of U.S. 301 is a rural four-lane divided highway that runs from Queenstown, MD, in the south to east of Massey, MD, in the north. In 2009, AADT along this stretch of U.S. 301 was about 26,000 vehicles per day for the two southern intersections and about 10,000 vehicles per day for the northernmost intersection. Table 1 lists the nine intersections.

The RCUT treatment was not the same at all locations. As previously shown in table 1, directional U-turns were installed both south and north of the main intersection at four locations. At the remaining locations, adjacent intersections or a two-way U-turn opening accommodated left-turn and through movements from the RCUTs’ main intersections.

The crash analyses were performed from crash data provided by the Maryland State Highway Administration. For U.S. 15, crash data were provided from January 1, 1980, through December 31, 1999, and covered periods that extend more than 5 years before and after the first and last RCUT treatments were applied. For U.S. 301, the crash data covered a period from January 1, 1996, through December 31, 2008. The crash data included the following attributes used in this analysis:

Simple Before-After Crash Analysis

The simple before-after analyses excluded crashes that occurred within 60 days before or after the RCUT deployment date. Crashes 60 days before the deployment date were excluded to avoid crashes that might have occurred during construction. Crashes 60 days after the deployment date were excluded to avoid the period during which drivers were becoming accustomed to the RCUT.

The following before-after periods were defined:

For U.S. 15, the longer periods included the data from 1980 through 1999. For U.S. 301, the longer periods included data from 1996 through 2008.

A short period may be more prone to regression to the mean than a longer analysis period. A longer period may be more prone to interpretation problems associated with historical changes such as increasing traffic volumes. If analyses for shorter and longer periods lead to the same conclusion, then greater confidence in the findings may be justified.

U.S. 15 at Hayward Road

Table 12 lists the mean annual crash counts for U.S. 15 at Hayward Road for the 3-year beforeafter analysis period. Table 13 lists the mean annual crash counts at that intersection for the longer before-after period. In these tables—and those for the intersections that follow—the first row under "Location" shows data for the crossover to the south of the subject main intersection. The second row shows data between the crossover to the south and the main intersection. The third row shows data at the main intersection. The fourth row shows data for the segment from the main intersection to the crossover to the north of the main intersection. The fifth row shows data for the crossover north of the main intersection. The sixth row (labeled "RCUT total") shows the total annualized count by collapsing over rows 2 through 4 (i.e., the RCUT intersection and the two adjacent segments). Because for some RCUTS the first and fifth rows may be represented in another RCUT, these rows were excluded from the totals so that comparability is maintained. The mean annual crash counts are provided separately for crashes classified in the crash reports as intersection related (Int) and those classified as non-intersection related (Non-Int).

Table 12. Short-period before-after mean annual crash counts for U.S. 15 at Hayward Road.

Location Before
1/18/85 to 1/17/88
After
1/14/88 to 1/14/91
Percent Change
Int Non-Int Int Non-Int Int Non-Int Total
Directional crossover 0.00 0.00 0.00 0.00
Segment 0.67 2.33 0.00 1.00 -100 -57 -67
Main intersection 4.00 0.33 2.33 1.00 -42 203 -23
Segment 0.00 1.67 0.00 1.00 -40 -40
RCUT crossover 0.00 0.00 0.00 0.00
RCUT total 4.67 4.33 2.33 3.00 -50 -31 -41

— Indicates undefined (division by zero).

Table 13. Long-period before-after mean annual crash counts for U.S. 15 at Hayward Road.

Location Before
1/1/80 to 1/17/88
After
1/14/88 to 12/31/99
Percent Change
Int Non-Int Int Non-Int Int Non-Int Total
Directional crossover 0.00 0.00 0.00 0.00
Segment 0.35 2.46 0.09 1.71 -74 -30 -36
Main intersection 4.09 0.23 4.13 0.45 1 96 6
Segment 0.00 1.64 0.09 2.07 26 32
RCUT crossover 0.00 0.00 0.18 0.00
RCUT total 4.44 4.33 4.31 4.23 -3 -2 -3

— Indicates undefined (division by zero).

Average annual crash counts at the main intersection decreased by 23 percent for the 3-year period but increased by 6 percent for the longer period. For the entire area of the RCUT, which includes the crossover locations and segments, there was a 41 percent decrease in average annual crashes over the short period and 3 percent decrease over the longer period.

U.S. 15 at Willow Road

Table 14 lists the mean annual 3-year before-after crash counts for U.S. 15 at Willow Road. Note that the northern crossover for Willow Road is the Biggs Ford main intersection. Table 15 lists the mean annual crash counts for the long before-after periods at the same intersection.

Table 14. Short-period before-after mean annual crash counts for U.S. 15 at Willow Road.

Location Before
1/17/89 to 9/16/92
After
1/14/93 to 1/13/96
Percent Change
Int Non-Int Int Non-Int Int Non-Int Total
Intersection 0.00 0.00 0.00 0.00
Segment 0.00 1.33 0.33 4.00 201 226
Main intersection 1.67 0.00 0.33 0.00 -80 -80
Segment 0.00 1.67 0.00 3.00 80 80
RCUT crossover 4.33 0.00 1.33 0.00 -69 -69
RCUT total 1.67 3.00 0.66 7.00 -60 133 64

— Indicates undefined (division by zero).

Table 15. Long-period before-after mean annual crash counts for U.S. 15 at Willow Road.

Location Before
1/1/80 to 9/16/92
After
1/14/93 to 12/31/99
Percent Change
Int Non-Int Int Non-Int Int Non-Int Total
Intersection 0.00 0.00 0.29 0.00
Segment 0.00 1.10 0.14 3.73 239 252
Main intersection 1.26 0.00 0.57 0.00 -55 -55
Segment 0.08 1.89 0.14 2.87 75 52 53
RCUT crossover 3.85 0.00 1.29 0.00 -66 -66
RCUT total 1.34 2.99 0.85 6.60 -37 121 72

— Indicates undefined (division by zero).

Crashes at the main intersection dropped by 80 percent for the short period and by 54 percent for the longer period. Summing the main intersection and segments (but not the RCUT crossovers counted in adjacent tables), there was a 64 percent increase and a 72 percent increase in average annual crashes for the short and long periods, respectively.

U.S. 15 at Biggs Ford Road

Table 16 lists the 3-year before-after crash statistics for U.S. 15 at Biggs Ford Road. Table 17 lists the U.S. 15 at Biggs Ford Road before-after crash statistics for the long before-after period.

Table 16. Short-period before-after mean annual crash counts for U.S. 15 at Biggs Ford Road.

Location Before
1/17/89 to 9/16/92
After
1/14/93 to 1/13/96
Percent Change
Int Non-Int Int Non-Int Int Non-Int Total
RCUT crossover 1.67 0.00 0.33 0.00 -80 -80
Segment 0.00 1.67 0.00 3.00 80 80
Main intersection 4.33 0.00 1.33 0.00 -69 -69
Segment 0.00 1.00 0.33 1.67 67 100
RCUT crossover 0.33 0.00 1.33 0.00 303 303
RCUT total 4.33 2.67 1.66 4.67 -62 75 -10

— Indicates undefined (division by zero).

Table 17. Long-period before-after mean annual crash counts for U.S. 15 at Biggs Ford Road.

Location Before
1/1/80 to 9/16/92
After
1/14/93 to 12/31/99
Percent Change
Int Non-Int Int Non-Int Int Non-Int Total
RCUT crossover 1.26 0.00 0.57 0.00 -55 -55
Segment 0.08 1.89 0.14 2.87 75 52 53
Main intersection 3.85 0.00 1.29 0.00 -66 -66
Segment 0.00 1.02 0.14 2.01 97 111
RCUT crossover 0.24 0.00 1.00 0.00 317 317
RCUT total 3.93 2.91 1.57 4.88 -60 68 -6

— Indicates undefined (division by zero).

There were decreases of 69 and 66 percent in average annual crashes at the main intersection for the short and long periods, respectively. With nearby segments included, the corresponding reductions were 10 and 6 percent.

U.S. 15 at Sundays Lane

Table 18 lists the mean annual short-period 3-year before-after crash counts for U.S. 15 at Sundays Lane. Table 19 lists the mean annual crash counts for long-period before-after mean at that intersection.

Table 18. Short-period before-after mean annual crash counts for U.S. 15 at Sundays Lane.

Location Before
1/17/89 to 9/16/92
After
1/14/93 to 1/13/96
Percent Change
Int Non-Int Int Non-Int Int Non-Int Total
RCUT crossover 4.33 0.00 1.33 0.00 -69 -69
Segment 0.00 1.00 0.33 1.67 67 100
RCUT 0.33 0.00 1.33 0.00 303 303
Segment 0.00 2.00 0.33 1.33 -34 -17
Intersection 0.33 0.00 1.00 0.00 203 203
RCUT total 4.99 3.00 4.32 3.00 -13 0 50

— Indicates undefined (division by zero).

Table 19. Long-period before-after mean annual crash counts for U.S. 15 at Sundays Lane.

Location Before
1/1/80 to 9/16/92
After
1/14/93 to 12/31/99
Percent Change
Int Non-Int Int Non-Int Int Non-Int Total
RCUT crossover 3.85 0.00 1.29 0.00 -66 -66
Segment 0.00 1.02 0.14 2.01 97 111
RCUT 0.24 0.00 1.00 0.00 317 317
Segment 0.00 1.10 0.14 1.29 17 30
Intersection 0.31 0.00 0.43 0.00 39 39
RCUT total 4.40 2.12 3.00 3.30 -32 56 94

— Indicates undefined (division by zero).

At Sundays Lane, there was an increase in the mean annual crash count at the main intersection for both the short and long periods. With the segments included, the increase was 50 and 94 percent for the short and long periods, respectively.

U.S. 15 at College Lane

Table 20 lists the 3-year mean annual crash counts for the 3-year before-after periods at the U.S. 15 and College Lane intersection. Table 21 lists the mean annual crash counts for the long periods.

Table 20. Short-period before-after mean annual crash counts for U.S. 15 at College Lane.

Location Before
1/17/91 to 1/16/94
After
01/14/94 to 1/13/97
Percent Change
Int Non-Int Int Non-Int Int Non-Int Total
Directional crossover 0.00 0.00 0.00 0.00
Segment 0.00 0.33 0.00 0.33 0 0
RCUT 3.67 0.00 0.33 0.00 -91 -91
Segment 0.33 0.67 0.33 0.33 0 -51 -34
Directional crossover 0.00 0.00 0.00 0.00
RCUT total 4.00 1.00 0.66 0.66 -84 -34 -74

— Indicates undefined (division by zero).

Table 21. Long-period before-after mean annual crash counts for U.S. 15 at College Lane.

Location Before
1/1/80 to 1/16/94
After
01/14/94 to 12/31/99
Percent Change
Int Non-Int Int Non-Int Int Non-Int Total
Directional crossover 0.00 0.00 0.00 0.00
Segment 0.00 0.76 0.00 0.19 -75 -75
RCUT 2.07 0.00 0.57 0.00 -72 -72
Segment 0.07 0.48 0.19 1.15 171 140 144
Directional crossover 0.00 0.00 0.00 0.00
RCUT total 2.14 1.24 0.76 1.34 -64 8 -38

— Indicates undefined (division by zero).

At the main intersection, there was a 91 percent decrease in crashes for the shorter period and a 72 percent decrease for the longer period. With crashes on the adjacent segments included, the decreases were 73 and 38 percent for the short and long periods, respectively.

U.S. 15 at U.S. 15 Business/Old Frederick Road

Table 22 lists the mean annual 3-year short-period before-after crash counts for U.S. 15 at U.S. 15 Business. Table 23 lists the mean annual crash counts for the long period.

Table 22. Short-period before-after mean annual crash counts for U.S. 15 at Old Frederick Road/U.S. 15 Business.

Location Before
7/18/85 to 7/17/88
After
11/14/88 to 11/13/91
Percent Change
Int Non-Int Int Non-Int Int Non-Int Total
Directional crossover 0.00 0.00 0.00 0.00
Segment 0.00 0.33 0.00 0.00 -100 -100
RCUT 3.33 0.33 0.67 1.00 -80 203 -54
Segment 0.00 0.33 0.00 0.67 103 103
Directional crossover 0.00 0.00 0.00 0.00
RCUT total 3.33 0.99 0.67 1.67 -80 69 -46

— Indicates undefined (division by zero).

Table 23. Long-period before-after mean annual crash counts for U.S. 15 at Old Frederick Road/U.S. 15 Business.

Location Before
1/1/80 to 7/17/88
After
11/14/88 to 12/31/99
Percent Change
Int Non-Int Int Non-Int Int Non-Int Total
Directional crossover 0.00 0.00 0.00 0.00
Segment 0.00 0.35 0.09 0.00 -100 -74
RCUT 4.21 0.12 1.62 0.36 -62 200 -54
Segment 0.00 0.82 0.00 1.53 87 87
Directional crossover 0.00 0.00 0.00 0.00
RCUT total 4.21 1.29 1.71 1.89 -59 47 -35

— Indicates undefined (division by zero).

At the U.S. 15 and U.S. 15 Business main intersection, mean annual crashes decreased by 54 percent in both the short and long periods. When crashes on the adjacent segments were included, the decreases were 46 and 35 percent, respectively.

U.S. 301 at Main Street

Table 24 lists the mean 3-year short-period before-after crash annual crash counts for the U.S. 301 at Main Street intersection. Table 25 lists the mean annual crash counts for the long period.

Table 24. Short-period before-after mean annual crash counts for U.S. 301 at Main Street.

Location Before
11/17/99 to 1/16/02
After
3/16/03 to 3/15/06
Percent Change
Int Non-Int Int Non-Int Int Non-Int Total
Segment 0.33 3.33 0.00 3.67 -100 10 0
Main intersection 2.67 0.67 0.67 0.67 -75 0 -60
Segment 0.00 1.33 0.33 1.67 26 50
RCUT crossover 5.33 1.67 0.33 0.67 -94 -60 -86
RCUT total 3.00 5.33 1.00 6.01 -67 13 -16

— Indicates undefined (division by zero).
Note: This intersection had no U-turn crossover at its southern terminous.

Table 25. Long-period before-after mean annual crash counts for U.S. 301 at Main Street.

Location Before
11/1/96 to 1/16/02
After
3/16/03 to 12/31/08
Percent Change
Int Non-Int Int Non-Int Int Non-Int Total
Segment 0.29 3.20 0.17 4.83 -41 51 43
Main intersection 2.33 0.58 0.52 0.52 -78 -10 -64
Segment 0.15 1.45 0.17 1.21 13 -17 -14
RCUT crossover 5.52 1.45 0.34 0.52 -94 -64 -88
RCUT total 2.77 5.23 0.86 6.56 -69 25 -7

This intersection had no U-turn crossover at its southern terminous. At the main intersection, mean annual crashes decreased by 60 and 64 percent in the short and long periods, respectively. With the adjacent segments included, the decreases were 16 and 7 percent, respectively.

U.S. 301 at Del Rhodes Avenue

Table 26 lists the 3-year short-period before-after mean annual crash counts for U.S. 301 at Del Rhodes Avenue. Table 27 lists the mean annual crash counts for the long periods.

Table 26. Short-period before-after mean annual crash counts for U.S. 301 at Del Rhodes Avenue.

Location Before
11/17/99 to 1/16/02
After
3/16/03 to 3/15/06
Percent Change
Int Non-Int Int Non-Int Int Non-Int Total
Intersection 2.67 0.67 0.67 0.67 -75 0 -60
Segment 0.00 1.33 0.33 1.67 26 50
Main intersection 5.33 1.67 0.33 0.67 -94 -60 -86
Segment 0.00 0.33 0.00 2.33 606 606
Directional crossover 0.00 0.00 0.00 0.00
RCUT total 5.33 3.33 0.66 4.67 -88 40 -38

— Indicates undefined (division by zero).

Table 27. Long-period before-after mean annual crash counts for U.S. 301 at Del Rhodes Avenue.

Location Before
1/1/96 to 1/16/02
After
3/16/03 to 1/1/09
Percent Change
Int Non-Int Int Non-Int Int Non-Int Total
Intersection 2.33 0.58 0.52 0.52 -78 -10 -64
Segment 0.15 1.45 0.17 1.21 13 -17 -14
Main intersection 5.52 1.45 0.34 0.52 -94 -64 -88
Segment 0.00 0.15 0.00 1.90 1,167 1,167
Directional crossover 0.00 0.00 0.00 0.00
RCUT total 5.67 3.05 0.51 3.63 -91 19 -53

— Indicates undefined (division by zero).

At the main intersection, the mean annual crash counts decreased by 86 and 88 percent for the short and longperiods, respectively. With adjacent segments included, the decreases were 38 and 53 percent, respectively.

U.S. 301 at Galena Road

Table 28 lists the 3-year short-period before-after crash mean annual crash counts for U.S. 301 at Galena Road. Table 29 lists the mean annual crash counts for the long period.

Table 28. Short-period before-after mean annual crash counts for U.S. 301 at Galena Road.

Location Before
1/17/98 to 1/16/01
After
3/16/02 to 3/15/05
Percent Change
Int Non-Int Int Non-Int Int Non-Int Total
Directional crossover 0.00 0.00 0.00 0.00
Segment 0.00 0.33 0.00 0.67 103 103
Main intersection 5.00 0.00 0.67 0.00 -87 -87
Segment 0.00 3.00 0.00 0.33 -89 -89
Directional crossover 0.00 0.00 0.00 0.00
RCUT total 5.00 3.33 0.67 1.00 -87 -70 -80

— Indicates undefined (division by zero).

Table 29. Long-period before-after mean annual crashes for U.S. 301 at Galena Road.

Location Before
1/17/98 to 1/16/01
After
3/16/02 to 3/15/05
Percent Change
Int Non-Int Int Non-Int Int Non-Int Total
Directional crossover 0.00 0.00 0.00 0.00
Segment 0.00 0.51 0.00 0.73 43 43
Main intersection 5.27 0.00 0.29 0.00 -94 -94
Segment 0.00 2.72 0.15 0.59 -78 -73
Directional crossover 0.00 0.00 0.00 0.00
RCUT total 5.27 3.23 0.44 1.32 -92 -59 -79

— Indicates undefined (division by zero).

At the main intersection, there was about an 87 percent decrease in average annual intersection crashes for the short period and a 94 percent decrease over the long period. When the segments are included, the reduction was 80 percent over the short period and 79 percent over the long period.

Before-After Summary

Table 30 summarizes the results listed for the individual RCUT intersections for the 3-year short period, and table 31 summarizes the results for the long period.

Table 30. Before-after average annual crash summary for RCUT intersections in the short period.

Location At Intersection Intersection and Adjacent Segments
Before After Decrease
(percent)
Before After Decrease
(percent)
U.S. 15 at Hayward Road 4.33 3.33 23 9.00 5.33 41
U.S. 15 at Willow Road 1.67 0.33 80 4.67 7.67 -64
U.S. 15 at Biggs Ford Road 4.33 1.33 69 7.00 6.33 10
U.S. 15 at Sundays Lane 0.33 1.33 -300 3.33 5.00 -50
U.S. 15 at College Lane 3.67 0.33 91 5.00 1.33 73
U.S. 15 at U.S. 15 Business 3.67 1.67 55 4.33 2.33 46
U.S. 301 at Main Street 3.33 1.33 60 8.00 7.00 13
U.S. 301 at Del Rhodes Avenue 7.00 1.00 86 7.67 3.33 57
U.S. 301 at Galena Road 5.00 0.67 87 8.33 1.67 80
Total 33.33 11.33 66 57.33 40.00 30

Table 31. Before-after average annual crash summary for RCUT intersections in the long period.

Location At Intersection Intersection and Adjacent Segments
Before After Decrease
(percent)
Before After Decrease (percent)
U.S. 15 at Hayward Road 4.33 4.58 -6 8.77 8.53 3
U.S. 15 at Willow Road 1.26 0.57 54 4.32 7.46 -73
U.S. 15 at Biggs Ford Road 3.85 1.29 66 6.84 6.46 6
U.S. 15 at Sundays Lane 0.24 1.00 -326 2.36 4.59 -95
U.S. 15 at College Lane 2.07 0.57 72 3.39 2.11 38
U.S. 15 at Old Frederick Road 4.33 1.98 54 5.50 3.59 35
U.S. 301 at Main Street 2.91 1.03 64 7.70 7.41 4
U.S. 301 at Del Rhodes Avenue 6.98 0.86 88 7.27 2.76 62
U.S. 301 at Galena Road 5.27 0.29 94 8.50 1.76 79
Total 31.23 12.19 61 54.66 44.68 18

The simple before-after analysis suggests that the RCUT treatment dramatically decreased crashes at the main intersection but increased crashes on the adjacent sections of road. These findings are consistent with a decrease in crossing-path crashes at the main intersection and an increase in merging and weaving crashes on the segments between the main intersection and the turnarounds.

Before-After Analysis with Controls

In this analysis, the controls were conventional intersections on the U.S. 15 and U.S. 301 corridors that were not converted to RCUTs. To maintain comparability between the converted intersections and the controls, the data from the same date ranges were used for all the intersections on the same corridor. The first RCUT treatment on U.S. 15 was in 1988, and the last was in 1994. Therefore, the before period selected on U.S. 15 was 1985 through 1987 when all of the intersections were conventional. The after period selected for U.S. 15 was 1995 through 1997 when all of the RCUT intersection conversions on that corridor had been completed. On U.S. 301, the RCUT conversions were performed in 2002 and 2003. Therefore, the before period selected for that corridor was 1999 through 2001, and the after period selected was 2004 through 2006. As a result of selecting common before-after periods for all intersections on a corridor, the analysis periods for the analysis with controls were different from those used in the simple before-after analysis reported above, and thus, the RCUT crash reduction results can be different from those reported above.

Table 32 lists the number of observed crashes in these before and after periods. Note that unlike the annualized numbers reported in the simple before-after analysis, the numbers reported here are not annualized; they are totals over the 3-year periods.

Table 32. Crashes at the RCUT intersections.

Location Intersection Crashes Intersection and Adjacent Segments
Before After Decrease (percent) Before After Decrease (percent)
U.S. 15 at Hayward Road 16.0 10.0 38 31.0 26.0 16
U.S. 15 at Willow Road 3.0 2.0 33 7.1 15.2 -114
U.S. 15 at Biggs Ford Road 8.0 6.0 25 14.5 15.4 -6
U.S. 15 at Sundays Lane 1.0 5.0 -400 5.4 9.4 -73
U.S. 15 at College Lane 2.0 1.0 50 8.0 4.0 50
U.S. 15 at U.S. 15 Business 9.0 7.0 22 12.0 12.0 0
U.S. 301 at Main Street 8.0 4.0 50 19.0 20.7 -9
U.S. 301 at Del Rhodes Avenue 20.0 4.0 80 25.0 15.3 39
U.S. 301 at Galena Road 12.0 1.0 92 20.0 7.0 65
Total 79.0 40.0 49 142.0 125.0 12

In table 32, the columns labeled "Intersection Crashes" refer to crashes that occurred at the main intersection or crashes that were classified as intersection-related and occurred within 0.03 mi of the main intersection. The columns labeled "Intersection Plus Adjacent" refer to all crashes that occurred at the intersection or in the adjacent sections of road and at the next upstream and downstream location where a U-turn was possible. In the case where a section of road connected two adjacent RCUTS, the crashes for that section were prorated across the two treatment locations. In the case where the upstream or downstream U-turn location was a treatment site, the crashes at those locations were not included in the intersection plus adjacent columns (because they were already tallied in the corresponding RCUT treatment location row). These steps prevented double counting of crashes.

There was a drop in intersection crashes at every treatment location except U.S. 15 at Sundays Lane. Overall, for the main RCUT intersections, the total number of crashes after the treatment was reduced 49 percent from the before period. There was an increase in the total number of crashes on the adjacent road segments and U-turn locations. However, the total effect of the RCUT conversions was a 13 percent decrease between the before and after periods.

For comparison, 10 intersections on U.S. 15 and U.S. 301 were selected where no RCUT treatment was applied. The same approach was used to identify intersection and intersection plus adjacent crashes, and the same 3-year time periods were used. Table 33 lists the number of observed crashes for the control intersections.

Table 33. Observed crashes at control intersections during two 3-year periods.

Location Intersection Crashes Intersection and Adjacent Segments
Before After Decrease (percent) Before After Decrease (percent)
U.S. 15 at Devilbiss Bridge Road 4.0 5.0 -25 8.0 16.0 -100
U.S. 15 at Angleberger Road 3.0 4.0 -33 16.5 20.5 -25
U.S. 15 at Auburn Road 8.0 10.0 -25 22.5 27.5 -22
U.S. 15 at Roddy Creek Road 4.0 6.0 -50 11.0 18.0 -64
U.S. 15 at Motters Station Road 3.0 3.0 0 11.0 9.0 18
U.S. 15 at Welty Road 9.0 11.0 -22 18.0 22.0 -22
U.S. 301 at Greenspring Road 4.0 9.0 -125 17.0 19.0 -12
U.S. 301 at Rolling Bridge Road 8.0 8.0 0 19.0 16.7 12
U.S. 301 at Ruthsburg Road 25.0 21.0 16 34.1 30.3 11
U.S. 301 at Barclay Road 1.0 9.0 -800 11.0 24.0 -118
Total 69.0 86.0 -25 168.1 203.0 -21

Because there were no changes at the control intersections that would affect the upstream and downstream intersections, the intersection plus adjacent columns only include crashes at the intersection and on adjacent road segments but not upstream and downstream U-turn locations or intersections. Overall, there was a 21 percent increase in crashes on the control sections.

If it is assumed that the percentage increase in the expected number of crashes in the RCUT sections would have been the same as at the control locations, then the effective decrease in crashes on the RCUT sections was 28 percent.

This assumption seems reasonable because the control intersections were on the same corridors and interspersed between RCUT treatment intersections. Thus, any changes in traffic volume that may have influenced crash rates would have been about the same for both the treatment and control intersections. However, the comparison of the control and RCUT intersections is not without limitations. In general, the intersections that were converted to RCUTs had higher crossstreet volumes than the controls. Also, because the control intersections did not require U-turn crossovers, the adjacent segments were somewhat longer than those of the RCUT intersections. Table 34 provides the adjacent segment lengths for all intersections used in this analysis. Although the control intersections are not exactly comparable to the intersections that were converted to RCUTs, the overall trends suggest that without the conversion, the RCUT intersections would have experienced considerably more crashes.

Table 34. Segment lengths of intersections included in the before-after with controls analysis (in miles).

Corridor Cross Street Southern Segment Northern Segment
U.S. 15 Hayward Road 0.28 0.27
U.S. 15 Willow Road 0.47 0.89
U.S. 15 Biggs Ford Road 0.89 0.25
U.S. 15 Sundays Lane 0.25 0.48
U.S. 15 College Lane 0.33 0.35
U.S. 15 Old Frederick Road 0.31 0.4
U.S. 301 Main Street   0.42
U.S. 301 Del Rhodes Avenue 0.42 0.19
U.S. 301 Galena Road 0.25 0.17
Mean RCUT Segment Length 0.4 0.38
U.S. 15 Devilbiss Bridge Road 0.24 0.71
U.S. 15 Angleberger Road 0.69 1.81
U.S. 15 Auburn Road 1.81 0.68
U.S. 15 Roddy Creek Road 0.04 0.2
U.S. 15 Motters Station Road 0.86 0.23
U.S. 15 Welty Road 0.71 0.41
U.S. 301 Greenspring Road 0.31 0.75
U.S. 301 Rolling Bridge Road 1.17 1.28
U.S. 301 Ruthsburg Road 1.28 0.35
U.S. 301 Barclay Road 1.48 1.94
Mean RCUT Segment Length 0.93 0.85

Note: The blank cell indicates that there was no U-turn crossover south of the Main Street intersection.

It should be noted that the intersections where the field observations were made were included in this analysis. The U.S. 15 at U.S. 15 Business intersection experienced no change in the number of crashes between before and after periods, whereas the U.S. 15 at Roddy Creek Road intersection, which was not treated, experienced a 64 percent increase in the number of crashes.

Before-After EB

EB analysis requires an SPF for estimating the expected number of crashes at the locations of interest and an estimate for the overdispersion parameter associated with this SPF. For this analysis, SPFs from the recently published Highway Safety Manual were used, as shown in figure 9. [5]

Figure 9. Equation. SPF used in EB analysis. Lambda subscript int equals C subscript local times CMF subscript skew times CMS subscript left times CMF subscript right times CMS subscript lighting times open parenthesis a times AADT subscript major superscript b times AADT subscript major superscript c close parenthesis.

Figure 9. Equation. SPF used in EB analysis.

The parameters in this equation are listed in table 35.

Table 35. SPF parameters for intersections.

Parameter Description Four-Way Intersection Three-Way Intersection
Clocal Calibration factor adjusting the SPF for local conditions 1.23 1.23
CMFSkew Crash modification factor for intersection angle
1 + 0.053 × Skew
_______________
1.43 + 0.53 × Skew
1 + 0.016 × Skew
_______________
0.98 + 0.16 × Skew
CMFLeft Crash modification factor for left-turn lane on major road 0.72 (one approach)
0.52 (two approaches)
0.56
CMFRight Crash modification factor for right-turn lane on major road 0.86 (one approach)
0.74 (two approaches)
0.86
CMFLighting Crash modification factor for lighting 0.896 0.895
a SPF model parameter -10.008 -12.526
b SPF model parameter 0.848 1.204
c SPF model parameter 0.448 0.236
k Overdispersion parameter 0.494 0.460

The AADTMajor and AADTMinor parameters are the annual average daily traffic on the major and minor legs of the intersection, respectively. The Skew parameter is the difference between the actual angle (in degrees) at which the minor street meets the major street and 90 degrees. Overdispersion parameter k indicates how much variability there is in the expected number of crashes. A large value for k indicates that the expected number of crashes at each intersection is very close to the model estimate. For this analysis, the skew was estimated from aerial photographs, as was the presence of right- and left-turn lanes. The crash modification factor (CMF) for lighting term was omitted because information was not available on the date lighting was applied to the sites.

For the segments adjacent to the locations at which the RCUT treatment was applied, the SPF shown in figure 10 was used.

Figure 10. Equation. SPF for segments adjacent to locations with RCUT treatment. Lambda subscript seg equals C subscript local times CMF subscript lane times CMF subscript shoulder times CMF subscript median times CMS subscript lighting times CMF subscript enforcement times open parenthesis e superscript a times AADT superscript b times L close parenthesis.

Figure 10. Equation. SPF for segments adjacent to locations with RCUT treatment.

The parameters for the RCUT adjacent sections equation are listed in table 36.

Table 36. SPF parameters for highway segments.

Parameter Description Value
Clocal Calibration factor adjusting the SPF for local conditions 1.23
CMFLane Crash modification factor for lane width
CMFShoulder Crash modification factor for shoulder width
CMFMedian Crash modification factor for median width
CMFLighting Crash modification factor for the presence of street lights
CMFEnforcement Crash modification factor for automated speed enforcement
L Length of highway segment (in miles)
a SPF model parameter -9.025
b SPF model parameter 1.049
c SPF model parameter 1.549

— Indicates that the actual segment length for the particular highway segment was used, that is, L was a variable rather than a constant.
Note: See the Highway Safety Manual for details.[5]

For the road segments in this analysis, State data indicated that lane, shoulder, and median widths were large enough that these CMFs for these factors were 1. Information on lighting and automated speed enforcement was not available.

The parameter L is the length of the highway segment. Overdispersion parameter k indicates how much variability there is in the expected number of crashes.

To use these SPFs, estimates were needed of the AADT for U.S. 15, U.S. 301, and the crossing streets where the RCUT treatments were applied, as well as at intersections and highway segments that were used to estimate the adjustment factor for local conditions. This was achieved by using a simple two-step rate equation shown in figure 11.

Figure 11. Equation. AADT estimation. If annual traffic volume is unchanged or increasing, then AADT(Y) equals AADT subscript 0 times open parenthesis 1 plus R subscript 0 close parenthesis superscript Y minus Y subscript 0. If annual traffic volume is decreasing, then AADT(Y) equals AADT subscript 0 times open parenthesis 1 plus R subscript 0 close parenthesis superscript Y subscript 1 minus Y subscript zero times open parenthesis 1 plus R subscript 1 close parenthesis superscript Y minus Y subscript 1.

Figure 11. Equation. AADT estimation.

In this equation, AADT0 is the estimate for the AADT in year Y0, and R0 is the annual percentage growth rate. Year Y1 is the year at which the growth rate changes, and R1 is the growth rate that applies after the year Y1. A two-step equation was used because traffic volumes on both U.S. 15 and U.S. 301 were increasing historically but started decreasing, depending on the location, between 2006 and 2008. This two-step equation represents an increasing period followed by a decreasing period, as required by the data.

In general, a separate regression analysis was performed for each location to calibrate the equation. For example, annual AADT values were downloaded from the Maryland Department of Transportation (MDOT) Web site for 16 locations on U.S. 301 and 23 locations on U.S. 15 from 1985 through 2009. To estimate the AADT at the locations needed for this analysis, interpolation of the nearest downloaded values was used to generate estimates for the AADT for a number of years. A regression was then performed to calibrate the two-step rate equation to the AADT data. The same approach was used for the cross-street traffic when it was available from the MDOT Web site.

For most cross streets of intersections at which the RCUT treatment was applied, annual AADT data were not available. However, detailed traffic count data were available for at least one day at five of the locations at which the RCUT treatment was applied. Therefore, the EB analysis was restricted to these five intersections. At these intersections, traffic count data was used to estimate the AADT for the year in which the traffic counts were taken and a typical growth rate was applied to extrapolate that value to other years. The typical growth rate was obtained by analyzing cross street AADT values at other intersections in the study area where such data were available.

Local calibration factors were estimated by comparing model forecasts and crash observations at four intersections and seven road segments that were not impacted by the RCUT treatments. This provided all of the data needed to apply a locally calibrated version of the Highway Safety Manual SPF to the five RCUTs for which minor approach traffic volumes were available. [5] The results of this analysis are shown in table 36. All crash count numbers and estimates are for 3-year periods and are not annualized.

Table 37. EB estimation of the expected number of crashes before RCUT treatment.

Location Years Crashes Model Over-Dispersion Weight EB Crashes Before
U.S. 15: 15.829 to 16.17 8 24 6.8 0.62 0.19 20.7
Hayward Road 8 37 55.8 0.49 0.04 37.7
U.S. 15: 16.18 to 16.51 8 13 6.7 0.64 0.19 11.8
U.S. 15: 17.07 to 18.02 12 23 28.9 0.22 0.13 23.8
Biggs Ford Rd 12 44 26.8 0.49 0.07 42.8
U.S. 15: 18.02 to 18.33 12 12 9.1 0.69 0.14 11.6
Sundays Lane 12 2 17.8 0.49 0.10 3.6
U.S. 15: 18.33 to 18.87 12 13 15.4 0.39 0.14 13.3
U.S. 15: 34.619 to 34.99 8 3 3.5 0.57 0.33 3.2
Old Frederick Road 8 34 12.1 0.49 0.14 30.9
U.S. 15: 35.02 to 35.477 8 7 3.6 0.46 0.37 5.7
U.S. 301: 43.36 to 43.67 6 3 2.2 0.69 0.40 2.7
Galena Road 6 31 10.9 0.49 0.16 27.9
U.S. 301: 43.67 to 43.905 6 16 1.6 0.90 0.40 10.2

Note: Intersections are in italics, and adjacent highway segments are shown in normal text.

In table 37 and table 38, the intersections are in italics, and adjacent highway segments are shown in normal text. The "Years" column indicates the number of years of before data that was available for calibrating the model. The crashes column is the number of observed crashes during that period, and the model column is the number that would have been estimated using just the model.

The EB estimates for the effectiveness of the RCUT treatment are shown in table 38.

Table 38. EB estimation of the effectiveness of the RCUT treatment for intersections for which cross-street traffic counts were available.

Location EB Crashes Before Model After EB Crashes After Effectiveness
U.S. 15: 15.829 to 16.17 20.7 13.5 40.8 0.51
Hayward Road 37.7 120.6 81.4 0.39
U.S. 15: 16.18 to 16.51 11.8 13.1 23.1 -0.04
U.S. 15: 17.07 to 18.02 23.8 24.5 20.2 -0.04
Biggs Ford Road 42.8 27.6 44.1 0.80
U.S. 15: 18.02 to 18.33 11.6 7.8 9.9 -0.52
Sundays Lane 3.6 18.5 3.7 -0.87
U.S. 15: 18.33 to 18.87 13.3 13.2 11.4 0.13
U.S. 15: 34.619 to 34.99 3.2 6.8 6.1 0.84
Old Frederick Road 30.9 27.4 70.1 0.71
U.S. 15: 35.02 to 35.477 5.7 7.6 12.2 -0.39
U.S. 301: 43.36 to 43.67 2.7 2.4 2.9 -0.71
Galena Road 27.9 12.0 30.5 0.93
U.S. 301: 43.67 to 43.905 10.2 1.8 11.2 0.55

The EB analysis indicates that the number of crashes at these intersections dropped by about 62 percent after applying the RCUT treatment, while the number of crashes on the adjacent highway segments dropped by about 14 percent, for a cumulative decrease of about 44 percent. Crash Severity

Crash Severity

One of the presumed benefits of the RCUT design is to reduce the number of crossing-path crashes. However, by increasing the number of merge and weave movements, the RCUT design has the potential to increase crashes on segments between the main intersection and the U-turn locations. The sideswipe and rear-end crashes that occur in merging and weaving sections are expected to be less severe than right-angle crossing-path crashes. Thus, an overall reduction in the severity of crashes would be expected in the RCUT influence area. To assess this, the number of crashes for the RCUT intersections and for the adjacent sections of road were grouped into three bins and tallied, using the same before-after periods as the previous analysis with controls (1985 to 1987 for the before period and 1995 to 1997 for the after period). The three bins were PDO crashes, crashes involving a fatality, and crashes involving an injury but no fatality. These tallies are shown in table 39. Data on the severity of injury crashes were not available, so it was not possible to estimate the reduction in the severity of injuries. However, the reduction in fatal crashes suggests that injury severity was also likely to have decreased.

Table 39. Observed crashes by severity before and after the RCUT treatment.

Location Before Period After Period
PDO Fatal Injury PDO Fatal Injury
U.S. 15 at Hayward Road 32 1 41 36 0 59
U.S. 15 at Willow Road 29 1 22 27 0 22
U.S. 15 at Biggs Ford Road 38 1 46 21 1 10
U.S. 15 at Sundays Lane 13 0 12 17 0 9
U.S. 15 at College Lane 21 0 28 6 0 5
U.S. 15 at Old Frederick Road 23 1 21 23 1 16
U.S. 301 at Main Street 26 2 24 29 0 14
U.S. 301 at Del Rhodes Avenue 20 1 28 7 0 7
U.S. 301 at Galena Road 16 3 30 7 1 3
Total 218 10 252 173 3 145

In total, 55 percent of all crashes at these intersections involved an injury or fatality before the RCUT treatment was applied. After the RCUT treatment, this percentage dropped to 46 percent of all crashes. There was a 70 percent reduction in fatal crashes and a 42 percent reduction in injury crashes between the 3-year periods.

 

ResearchFHWA
FHWA
United States Department of Transportation - Federal Highway Administration