Guide for Identifying, Classifying, Evaluating, and Mitigating Truck Freight Bottlenecks (2017)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

87 C h a p t e r 8 This chapter describes how to incorporate truck freight bottleneck analysis into typical plan- ning study documents. This is followed by several examples of truck freight bottleneck analysis and how they were—or could be—incorporated into planning studies. 8.1 Incorporation into Study Documents There are several types of planning studies that can benefit from incorporation of a truck bottleneck analysis. These studies include: • Statewide, MPO, and local freight plans; • Statewide, MPO, and local general long-range transportation plans; • Freight-intensive corridor studies; • Local and regional truck route designation studies; • Modal diversion studies for freight; • Statewide, regional, or corridor-specific safety studies; • Multimodal bottleneck analyses; • Emissions estimation studies requiring detailed speed inputs; and • Economic development studies focused on infrastructure improvement. The tasks used to implement planning studies tend to fall into a set of activities that can be used as a pivot point with which to understand the relevance of truck bottleneck analysis. Table 8-1 shows how truck bottleneck analysis can be incorporated into tasks that are typically associated with planning studies. 8.2 I-95 Truck Bottleneck Analysis in North Carolina For the North Carolina DOT I-95 Economic Impact Study, a truck bottleneck analysis was con- ducted to identify bottleneck locations along the corridor. As a first step, a truck GPS dataset con- taining spot speeds for activity during June 2012 was produced, and all data points that fell along the I-95 corridor in North Carolina were compiled. The roadway was segmented bi-directionally at each mile of the 182 centerline miles to produce a shapefile with 364 bi-directional segments. The compiled data points were then matched to the 364 1-mile road segments. Within each of the 364 data bins, the data were separated further by day of week (Mo–Sun) and hour of day to produce 61,320 data bins. An average speed was produced for each bin and the results were scanned for congestion. The scan focused on data bins where average speeds within a segment fell below 85 percent of the free-flow speed at some point during a week. For this analysis, free-flow speed was considered to Incorporating Truck Bottleneck Analysis into the Planning Process

88 Guide for Identifying, Classifying, evaluating, and Mitigating truck Freight Bottlenecks Task in Planning Study Incorporation of Bottleneck Analysis Existing Conditions Essential data collected for bottleneck analysis (speed and volume) can be used as part of the description of existing conditions. See Chapters 3 and 4 of this document. Desktop analysis to identify and quantify bottlenecks (Chapter 5 of this document) can be used to describe existing conditions for trucks on the road network. Future Baseline Conditions Travel demand models can be augmented by using bottleneck analysis as the source of delay estimates in the base year, then increasing delay proportional to increases in V/C ratios provided by the travel demand model. Identification of Needs The causal analysis described in Chapter 6 can be used to identify needs in the system. For example, if a large percentage of truck bottlenecks are caused by crashes, then this indicates the need for safety improvements. Identification of Solutions to Consider Mitigation options described in Chapter 8 can be used as a source of solutions to consider for the planning study. Field analysis described in Chapter 7 can also be used to identify solutions. Analysis of Solutions and Development of Recommendations The ranking of causes of bottlenecks (see Chapters 6 and 7) can be used to prioritize solutions that are recommended. For example, if the majority of truck bottlenecks at a particular location are based on weather, then solutions that are targeted towards improving the road’s ability to handle inclement weather may be given a 30 percent increase across a scoring method for solutions. Outreach Draft results of bottleneck analyses should be presented to public- sector and private-sector stakeholders to validate locations of bottlenecks, severity of bottlenecks, potential causes of bottlenecks, and mitigation options to consider for addressing bottlenecks. Table 8-1. Incorporation of bottleneck analysis into planning studies using generic tasks. be the maximum average speed across all one hour time bins. Bins that fell below the 85 percent criteria were flagged for further congestion analysis, which included a calculation of average minutes of delay per week. Of the 61,320 bins, 1,491 showed this level of congestion. A total of 15 of 364 segments experienced delays greater than 5 minutes per week using the methodology described above. The locations where the delays occurred are shown in Figures 8-1 and 8-2. Figure 8-1 is a map that illustrates noticeable, measurable delay found during this scan.

Incorporating truck Bottleneck analysis into the planning process 89 Source: NCDOT I-95 Economic Impact Analysis. Figure 8-1. I-95 Truck bottleneck locations in North Carolina based on truck GPS data.

90 Guide for Identifying, Classifying, evaluating, and Mitigating truck Freight Bottlenecks Figure 8-2. I-95 truck bottlenecks in North Carolina by mile segment. Source: NCDOT I-95 Economic Impact Analysis. Figure 8-2 offers a more detailed look of where and why weekly minutes of delay occurred. Based on the data displayed in Figure 8-2, the areas where the greatest minutes of delay occurred on the corridor were two weigh stations. Based on overlapping crash data with the truck bottle- neck analysis periods, it was determined that delay also occurred due to an accident in Robeson County, as well as due to light congestion in Johnston County. On June 12 and 14, 2012, the corridor had the highest number of congested mile-hours (82) while June 24 had the least number of congested mile-hours (6). For context, there were a total of 8,736 mile-hours on the corridor in June 2012, meaning that on the most congested day, roughly one percent of mile-hours was congested. For the month, there were a total of 1,268 congested mile-hours out of 262,080 total mile-hours of travel. Most of the noticeable areas of congestion on the corridor are directly related to weigh stations and likely do not impact passenger vehicles. Four of the top five mile segments that have congestion are adjacent to a weigh station facility a few miles north of Lumberton. The lower speeds appear in the database at those locations as trucks slow down to exit or increase speed leaving the weigh station. It is possible that queues extending onto the highway at these weigh stations contribute to the lower speeds. These four weigh station segments taken together account for 396 congested mile-hours, which represents 30.9 percent of the total monthly congested mile-hours for the corridor. Mile segment 152 also contains a weigh station and contributed 154 congested mile-hours (12.1 percent of monthly total). Table 8-2 lists the 20 mile segments with the highest conges- tion levels. Regarding time of day, the highest levels of congestion occur between 10 a.m. and 3 p.m. As noted in the preliminary congestion scan, much of that congestion is related to weigh stations. Thus, given that weigh station activity is generally heaviest during the midday hours, this analysis further validates the findings of the preliminary congestion scan. A day-of-week analysis reveals that Tuesday has the highest number of congested mile-hours (273) and Sunday has the lowest number (39). Tables 8-3 and 8-4 describe these results further.

Incorporating truck Bottleneck analysis into the planning process 91 Top 20 Locations Mile Segment Number of Days with Some Congestion (0 to 30) Number of Hours with Some Congestion (0 to 720) 1 24_N 21 124 2 25_S 22 120 3 152_N 23 90 4 24_S 23 75 5 25_N 23 73 6 152_S 21 64 7 95_S 5 20 8 181_N 13 19 9 97_S 11 19 10 181_S 13 18 11 93_N 6 17 12 71_N 13 15 13 48_S 11 15 14 97_N 11 15 15 94_S 4 14 16 96_S 4 14 17 92_N 3 13 18 106_S 9 12 19 71_S 10 10 20 91_N 3 10 Source: NCDOT I-95 Economic Impact Analysis, 2013. Table 8-2. Top 20 congested locations on I-95 in North Carolina based on truck GPS data.

92 Guide for Identifying, Classifying, evaluating, and Mitigating truck Freight Bottlenecks Hour Number of Mile-Days of Congestion Begin Hour End Hour 0 1 3 1 2 6 2 3 7 3 4 11 4 5 9 5 6 28 6 7 49 7 8 61 8 9 84 9 10 98 10 11 121 11 12 112 12 13 124 13 14 120 14 15 111 15 16 90 16 17 99 17 18 51 18 19 35 19 20 22 20 21 12 21 22 10 22 23 2 23 24 3 Table 8-3. Congestion by hour of day on I-95 in North Carolina based on truck GPS data.

Incorporating truck Bottleneck analysis into the planning process 93 Day of Week Number of Mile-Hours of Congestion Monday 160 Tuesday 273 Wednesday 239 Thursday 258 Friday 210 Saturday 89 Sunday 39 Source: NCDOT I-95 Economic Impact Analysis, 2013. Table 8-4. Congestion by day of week on I-95 in North Carolina based on truck GPS data. 8.3 Mapping of Truck Speeds in Indianapolis Figure 8-3 shows truck speed data in a subarea of Indianapolis at the intersection on I-70 and I-465. This Figure shows the spot speed of thousands of truck speeds using truck GPS data pro- vided by the ATRI. This data has been mapped to aerial information, which allows for overlapping of land use data, truck count data, and other vehicle activity data. These data also show thousands of red dots in the subarea, which highlight truck parking loca- tions. This is a strong indication of the locations where internal-external and external-internal truck trips are being generated in the subarea. The facilities nearby these dots are the specific loca- tions that are most heavily impacted by the truck congestion that has been identified. 8.4 Truck Bottleneck Analysis in Downtown Valdosta, Georgia Truck speeds and delay in downtown Valdosta were measured using FHWA NPMRDS as part of a study of downtown truck traffic. The NPMRDS provided average truck and total vehicle speeds on NHS routes in the U.S. Both U.S. 84 and U.S. 41 are part of the NHS network. Truck congestion in the downtown area was analyzed using truck speed data during the after- noon peak period of 5:00 p.m. to 6:00 p.m. Figure 8-4 shows the average weekday truck speeds in April of 2015 during the afternoon peak period. Average truck speeds along U.S. 84 range from about 15 to 35 miles per hour. In down- town Valdosta, average speeds are generally under 25 miles per hour. Similarly, truck speeds along U.S. 41 Business from SR 31/Madison Highway south of downtown to SR 125/Bemiss Road north of downtown average between 15 and 35 mph. Average speeds along U.S. 41 Business south of Madison Highway are significantly higher as it is further removed from the core of the city. This compares to a range of posted speed limits on U.S. 41 that drops down to 25 mph within down- town and rises to 45 mph outside of downtown. Truck delay was then estimated by combining truck count data with truck speed data. Truck delay is measured as the difference between actual travel time and free-flow travel time multiplied by the hourly truck volume. The formula for calculating delay is as follows:  [ ]( ) ( )= −Truck Delay Distance Actual Truck Speed Distance Free-Flow Truck Speed Hourly Truck Volume

94 Guide for Identifying, Classifying, evaluating, and Mitigating truck Freight Bottlenecks Source: ATRI Truck GPS Data, 2015. Figure 8-3. Bottleneck map in Indianapolis subarea with roadway GIS attribute data, I-70 and I-465.

Incorporating truck Bottleneck analysis into the planning process 95 Source: Downtown Truck Traffic Mitigation Study, Valdosta-Lowndes County MPO, 2015. Figure 8-4. April 2015 average weekday truck speeds, 5–6 p.m. Truck delay through downtown Valdosta was found to be relatively low. As depicted in Fig- ure 8-5, for the month of April of 2015, delay along U.S. 84 is much higher outside of the core downtown area. The most significant delay on U.S. 84 occurs in two locations: (1) between SR 133/Street Augustine Road and I-75 and (2) between Clay Road and U.S. 41/Inner Perimeter Road as shown. 8.5 Truck Bottleneck Analysis in Idaho Statewide Freight Plan A truck GPS analysis was conducted in Idaho to identify truck bottlenecks. It was also over- lapped with truck volume data provided by the Idaho Transportation Department to identify the most critical truck bottleneck locations in the state. Based on the analysis, the stretch of I-84 from Caldwell through Boise is the only stretch of highway in Idaho that experiences congestion on a recurring basis. The section of I-84 east of this segment to the interchange with I-84 is also heavily utilized, with average daily truck traffic between 5,000 and 6,000 vehicles. These portions of the Interstate system serve the largest urban area in the state and link it to the Salt Lake City market, including intermodal facilities, and destinations further east. The remaining Interstate segments in Idaho all carry between 1,000 and 5,000 trucks per day. In addition to the stretch of

96 Guide for Identifying, Classifying, evaluating, and Mitigating truck Freight Bottlenecks Source: Downtown Truck Traffic Mitigation Study, Valdosta-Lowndes County MPO, 2015. Figure 8-5. Truck minutes of delay per day in evening peak, April 2015, 5:00–6:00 p.m. I-84 identified above, truck speeds are also low on I-90 at both the Oregon and Montana borders, as well as a short stretch of I-84 near Burley. The low truck volumes on I-90 for this time period indicate that the slow speeds are likely due to other vehicle congestion caused by rush hour (near Coeur d’Alene) and terrain (near Montana). Burley and Oakley to the south are important industrial and agricultural areas with numerous freight-reliant industries and truck AADT above 5,000, indicating that at least some of the slow speeds are likely due to freight-related congestion as trucks enter and leave the Interstate. The highest non-Interstate truck volume is found on U.S. 20 north of Idaho Falls. Segments of this route carry an average of 3,400 trucks per day. Areas with high truck AADT and low speeds include the Twin Falls area, eastern Boise, U.S. 20 north of Idaho Falls, and the Coeur d’Alene area. Figure 8-6 shows truck volume and locations where speeds were below 35 mph on the Interstate system, or below 25 mph on the state highway system, between 5 p.m. and 6 p.m. on weekdays in April 2015. The locations of truck bottlenecks in Idaho were later confirmed through a series of outreach efforts, including a combination of group stakeholder meetings and one-on-one interviews. Specifically, these maps were presented in both environments for comment by private-sector freight community members and public-sector transportation agency staff.

Incorporating truck Bottleneck analysis into the planning process 97 8.6 Truck Bottleneck Analysis in Arkansas Truck GPS data from March of 2015 were used to determine truck speed performance through- out Arkansas. The GPS data were used to determine average truck speeds during different periods throughout the day on both Interstates and non-Interstates. They also were used to determine difference between truck speeds during congested periods and free-flow periods. Figure 8-7 shows truck speeds on Interstates in Arkansas during the 5:00 p.m. to 6:00 p.m. afternoon peak period in March of 2015. Much of the truck congestion on the state’s highway system was found to be centered on the Little Rock metropolitan area. Average truck speeds in Little Rock generally are between 25 and 35 mph in the peak directions of travel. There also are truck mobility challenges in other population centers, notably northwest Arkansas (in the cities of Fayetteville and Bentonville) and the Jonesboro metropolitan area. In particular, I-49 in north- west Arkansas between Fayetteville and Bentonville has relatively significant truck congestion during the peak periods. Much of I-49 between these two cities consists of only two lanes in each direction, which is likely a contributing factor. I-555 in Jonesboro also shows some truck con- gestion during the peak periods, though not to the same extent as the Little Rock and northwest Arkansas regions. Much of I-555 also consists of two lanes in each direction. Truck GPS was also used to estimate truck speed reliability of the Arkansas Interstate system. Reliability is a measure of the variation of truck speeds over a long time period. Truck speed reli- ability is a critical operational issue for shippers and truck fleet operators. It causes trucks to build in a significant buffer time in to their delivery windows to ensure that they meet the desired level of on-time performance for their shipments. Source: NPMRDS, ITD. Figure 8-6. Truck speeds below 35 mph (Interstates) and 25 mph (non-Interstates) in p.m. peak and truck AADT.

98 Guide for Identifying, Classifying, evaluating, and Mitigating truck Freight Bottlenecks Reliability was calculated by using the average truck speed and standard deviation of truck speeds for each highway link. From these values, the percent deviation is calculated by dividing the standard deviation of truck speeds by the average truck speed and multiplying the resulting value by 100 percent using the following formula: ( )( )=Truck speed reliability Truck Speed Standard Deviation Average Truck Speed 100% Truck speed reliability values that are close to 0 percent indicate that truck speeds do not vary greatly during the observation period. Values that are close to 100 percent indicate that truck speeds vary significantly. Figure 8-8 shows the truck speed reliability on the Arkansas Interstate system during the p.m. peak hour. Locations of low truck reliability are similar to locations that exhibit high levels of congestion. In the Little Rock region, truck speeds on portions of the Interstate highway system are estimated to vary by as much as 75 percent to 100 percent during the p.m. peak period. Truck speeds show significant variation in some other parts of the state as well, particularly northwest Arkansas, Jonesboro, and the west Memphis regions. In northwest Arkansas, portions of I-49 near Bentonville show relatively high variations in truck speed, though not to the extent observed in Little Rock. Truck speeds along sections of I-40 near West Memphis and I-555 near Jonesboro exhibit truck speeds that vary by as much as 50 percent to 75 percent during peak periods. The I-40 corridor between Little Rock and Memphis is notable in that it exhibits much higher levels of unreliable truck speed locations relative to truck congestion locations. This indicates that, while congestion on the corridor is not a daily occurrence, the variation in truck speeds is causing significant impedance to truck activity on I-40. Source: ATRI truck GPS data, consultant analysis. Figure 8-7. Average p.m. peak-hour truck speeds on Interstates.

Incorporating truck Bottleneck analysis into the planning process 99 Figure 8-8. PM peak-hour truck speed reliability on Arkansas Interstates. Source: ATRI truck GPS data, consultant analysis. 8.7 Truck Parking Analysis in California Truck GPS was also used to conduct a truck parking analysis in California. Lack of truck park- ing is a potential cause of process-based truck delay through the additional time and distance that may be driven as truck drivers search for parking. In this example, every square mile in California was scanned to identify regions with high levels of truck parking. Medium- and heavy-duty trucks were scanned separately. As shown in Figure 8-9, medium-duty truck stops tend to be more dif- fuse and in urban areas. Separate analysis showed that heavy-duty truck parking tends to be along major highways. A “zoomed in” view was used to examine the heavy-duty parking locations. Figure 8-10 shows heavy-duty truck parking concentrations in northern Alameda County. The analysis was used to identify both expected and unexpected locations of industrial activity. Stakeholders at these locations are most heavily impacted by truck bottlenecks in the region. These stakeholders can be included in outreach activities that are used to determine the causes of bottlenecks and propose potential mitigation actions for these bottlenecks. 8.8 Truck Bottleneck Analysis in Georgia A truck bottleneck analysis was conducted as part of the Georgia Statewide Freight & Logistics Plan. Truck GPS data was used to estimate truck travel speeds throughout the state during four time periods. This example demonstrates the breadth of analyses that can be done using truck GPS data. Figure 8-11 shows the truck speeds mapped for the morning peak period for the state. It shows that the primary bottleneck locations are in the Atlanta metropolitan region, which were then featured in a series of maps such as the morning peak truck travel speed map shown

100 Guide for Identifying, Classifying, evaluating, and Mitigating truck Freight Bottlenecks Figure 8-9. Heat map of truck stops for medium-duty trucks in California. Source: Scanning California for Truck Stops. StreetLight Data. http://blog.streetlightdata.com/truck-stop-index. Accessed December 9th 2016. Figure 8-10. Heavy-duty truck stops in northern Alameda County. Source: Scanning California for Truck Stops. StreetLight Data. http://blog.streetlightdata.com/truck-stop-index. Accessed December 9th 2016.

Incorporating truck Bottleneck analysis into the planning process 101 in Figure 8-12. The truck speed analysis was also used to identify the key congested corridor segments in the Atlanta metropolitan region. These are shown in Figure 8-13. For each of the congested corridor segments, a series of maps was developed and analysis was conducted to provide detailed information on the nature of the bottlenecks. Table 8-5 shows the average truck speeds on each segment during each period of the day. Figure 8-14 shows the aver- age speed by time period by milepoint, truck travel speed reliability by milepoint, and reliability by time of day for the entire congested corridor in the westbound and eastbound direction for a 6-mile segment on I-20 in Atlanta. Figure 8-11. Average truck speeds as a percent of speed limit (morning peak period).

102 Guide for Identifying, Classifying, evaluating, and Mitigating truck Freight Bottlenecks Source: GDOT Freight & Logistics Plan, 2012. Figure 8-12. Average truck speeds as a percent of speed limit in Atlanta Metropolitan region using truck GPS data (morning peak).

Incorporating truck Bottleneck analysis into the planning process 103 Figure 8-13. Map of congested corridors in Atlanta metropolitan region based on truck GPS data.

104 Guide for Identifying, Classifying, evaluating, and Mitigating truck Freight Bottlenecks Table 8-5. Summary truck speed statistics on congested corridors in Atlanta metropolitan region based on truck GPS data. Corridor Direction A.M. Peak Average Speed Midday Average Speed P.M. Peak Average Speed Off-Peak Average Speed I-20 Miles 47-52 EB 38.2 52.6 54.6 58.7 I-20 Miles 47-52 WB 56.8 56.7 51.0 56.8 I-20 Miles 66-72 EB 59.5 58.2 39.9 56.9 I-20 Miles 66-72 WB 47.0 55.5 54.0 57.0 I-75 Miles 217-231 NB 55.9 59.5 55.0 61.7 I-75 Miles 217-231 SB 62.9 60.4 47.1 62.2 I-75 Miles 243-251 NB 40.1 52.5 39.7 55.7 I-75 Miles 243-251 SB 51.9 51.5 38.0 56.2 I-75 Miles 257-275 NB 61.7 60.2 39.3 60.1 I-75 Miles 257-275 SB 45.7 58.6 58.8 62.0 I-85 Miles 95-110 NB 60.6 59.9 48.3 60.4 I-85 Miles 95-110 SB 43.5 57.7 57.0 61.8 I-285 Miles 8-15 Inner Loop 54.5 58.9 55.7 59.5 I-285 Miles 8-15 Outer Loop 58.6 56.5 42.8 58.3 I-285 Miles 21-35 Inner Loop 50.9 56.6 37.0 57.5 I-285 Miles 21-35 Outer Loop 50.9 56.1 40.0 58.1 I-285 Miles 46-50 Inner Loop 60.5 60.5 58.0 61.6 I-285 Miles 46-50 Outer Loop 54.2 57.7 46.3 58.1 GA 400 Miles 7-20 NB 58.3 59.8 52.7 60.0 GA 400 Miles 7-20 SB 40.1 57.7 50.4 60.4

Incorporating truck Bottleneck analysis into the planning process 105 Figure 8-14. I-20 average speed, segment and time-of-day reliability based on truck GPS data.