Field Test Results of the Multimodal Level of Service Analysis for Urban Streets (2010)

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.

Page 9 3. Results Based upon the field tests and the feedback obtained from the various agencies participating in the field tests, the research team recommends the refinements to the MMLOS method and User’s Guide described in the sections below. Auto LOS Model An extensive field evaluation of the NCHRP 3-70 auto model and three alternative models was conducted (see attached evaluation white paper by Dr. Nagui Rouphail). Methodology The NCHRP 3-70 model and the current HCM auto LOS model produced LOS grades equal to or within 1 letter grade of service for 26 out of the 35 streets field tested. As shown in the Final Report (NCHRP 616) (reproduced in Exhibit 4 below) the NCHRP 3-70 auto model fits the laboratory results much better than the current HCM auto LOS model. The 3- 70 model matched the laboratory results for 69% of the clips, while the HCM 2000 model matched only 26% of the clip results. Consequently no changes are recommended for the NCHRP 3-70 auto LOS model. User Guide The research team spotted a few typos in Exhibit 9 of the User’s Guide. The parameters for the Adverse Signal Progression and No Signal Coordination rows have been switched. These parameters are used to predict stops per mile if the analyst cannot field measure stops or obtain a satisfactory analytical tool for predicting stops. The corrected table is provided below. Exhibit 3: Parameters for Auto Stops per Mile Equation Signal Progression Arrival Type A1 A2 A3 Adverse Signal Progression 1,2 0.636 5.133 0.051 No Signal Coordination 3 0.478 6.650 0.028 Good Signal Progression 4,5,6 0.327 9.572 0.013

Page 10 Exhibit 4: Evaluation of Proposed Auto LOS Models Art Spd Lim Actual Stops Left Ln Med Video HCM Model Model Clip # Street Class (mph) (mph) (stps/mi) (%) (1,2,3) LOS LOS #1 LOS #2 LOS 61 Rt 50 1 50 28 1.4 100% 0.00 A C B C 56 Sunset Hills Rd 2 40 23 2.0 100% 3.00 A C B A 2 Gallows Road 3 35 35 0.0 100% 3.00 B A B A 65 Lee Hwy 2 40 36 0.0 100% 2.00 B A B A 63 Rt 50 1 50 42 0.0 100% 3.00 B A B A 5 Wilson Blvd 3 35 30 0.0 100% 3.00 B B B A 62 Rt 50 1 50 37 0.0 100% 0.00 B B B A 13 Washington Blvd 3 35 25 0.0 0% 0.00 B B B A 7 Wilson Blvd 3 35 20 0.0 100% 1.00 B C B B 54 Lee Hwy 2 40 25 3.3 100% 2.00 B C B A 53 Prosperity 2 40 19 1.7 100% 3.00 B D B B 6 Clarendon 3 35 18 2.3 100% 1.00 B D B B 10 Washington Blvd 3 35 17 3.8 0% 0.00 B D C C 20 Rt 50 1 50 16 1.8 100% 3.00 B E B C 64 Rt 50 1 50 20 2.0 100% 3.00 B E B B 58 Sunrise Valley Rd 2 40 11 1.7 100% 3.00 B F B C 1 Rt 234 1 50 15 2.0 100% 3.00 B F B C 29 Rt 234 2 40 23 2.0 100% 3.00 C C B A 19 23rd St 4 30 16 5.8 0% 0.00 C C C C 12 Wilson Blvd 3 35 14 4.3 0% 0.00 C D C D 60 Lee Hwy 2 40 15 2.0 100% 2.00 C E B C 21 Rt 50 1 50 20 4.0 100% 3.00 C E C B 8 Wilson Blvd 3 35 14 4.1 100% 1.00 C E C C 52 M St 4 30 8 7.3 0% 0.00 C E D E 55 Braddock Rd 2 40 13 2.2 100% 3.00 C F B C 59 Sunset Hills Rd 2 40 12 4.9 0% 0.00 C F C E 15 Glebe Road 2 40 8 6.0 100% 3.00 C F C D 14 Glebe Road 2 40 11 6.0 100% 3.00 C F C C 57 Sunset Hills Rd 2 40 17 3.3 0% 0.00 D D C D 16 Fairfax Drive 3 35 12 7.3 100% 3.00 D F C C 51 M St 4 30 7 9.1 0% 0.00 D F D E 25 M St 4 30 11 3.7 0% 0.00 E D C D 23 M St 4 30 8 5.6 0% 0.00 E E C E 30 M St 4 30 7 14.5 0% 0.00 F F F E 31 M St 4 30 4 18.0 0% 0.00 F F F F % Exact Match To Video 100% 26% 69% 37% % Within 1 LOS of Video 100% 46% 94% 89% Source: NCHRP Report 616 Note that several different sections or time periods of the same street were used for many of the clips.

Page 11 Transit LOS Model The field tests indicated no issues with the output of the transit LOS model. The difficulties were generally on the data collection side. Engineers and planners unfamiliar with working with local transit agency personnel generally expressed the most concerns about gathering the transit service data. Methodology No changes are recommended for the transit LOS model. User Guide Obtaining field data on transit service characteristics for the specific section of the routes serving the analysis street section was a concern to many potential users of the transit LOS method. Over the course of several workshops various methods were developed for approximating field measurements through the use of data already being regularly collected by transit agencies for their own management needs. Many transit agencies regularly collect data on the peak passenger load points and on-time performance for each of the routes they operate. This data is not usually available by specific street segments. However, if one considers that the transit riders on any given street probably experienced the peak loading conditions and reliability of the route somewhere during their trip; it can be a reasonable approximation to apply this route data to the street being evaluated for transit LOS. This approach appeared to provide a cost-effective and reasonable substitute for measuring reliability and passenger load factors in the field. Similarly, rather than going to the expense of measuring bus speeds in the field an analyst can consult the published bus route schedules to obtain an average point to point speed for the portion of the bus route within which the analysis street section is located. The schedule speed will not be identical to the actual street section speed for the bus, but unless conditions on the analysis street section are very different from the rest of the bus route, it should be close enough to assess the transit level of service for the street. In the field tests, these approximations to field data collection appeared to be sufficiently accurate for planning purposes. The remaining transit data on bus stop amenities is relatively easy to gather in the field. Bicycle LOS Model Field application of the bike LOS model ran into some street measurement issues. Most of these measurement issues had long since been solved by Sprinkle Consulting. Thus the guidance for measurement of street widths for bike LOS is explained in a bit more detail below. Methodology The field tests did not indicate that changes were required for the bike LOS methodology, thus no changes are recommended.

Page 12 User Guide In San Diego the research team confronted problem of assessing bicycle LOS on a street where buses frequently stop in the only available travel lane for both bicycle and bus. The current method was not developed or calibrated for such a situation, so it is recommended that the MMLOS method not be applied in such situations to estimate bicycle LOS. The analyst might query bicycle riders on the bus street to obtain an assessment of bicycle LOS for those specific conditions. In an assessment of bicycle LOS for a residential street in Oakland, California, the research team noted that frequent single family driveways on the street caused the bicycle LOS to come out at a much poorer level than expected. The frequency of residential driveways caused the poor bicycle LOS to be unmitigatable short of closing the driveways. Consequently, it is recommended that users of the MMLOS method discount low volume single family driveways when computing bicycle LOS using the MMLOS method. The percentage discount would be left to the discretion of the analyst. Pedestrian LOS Model There are a couple of typos in the description of the Pedestrian LOS model in NCHRP 616 and web document 128. Both of these documents show the following equation for pedestrian intersection level of service: Pedestrian LOS for Signalized Intersections = 0.00569(RTOR+PermLefts) + 0.00013(PerpTrafVol*PerpTrafSpeed) + 0.0681(LanesCrossed0.514) + 0.0401ln(PedDelay) –RTCI(0.0027PerpTrafVol – 0.1946) + 1.7806 The highlighted parameters are incorrect. They should be 0.681 (instead of 0.0681) and 0.5997 (instead of 1.7806). This is equation 37, page 88 of NCHRP 616, and equation 22, Page 19 of the User’s Guide (Web Document 128). Sprinkle Consulting has also recommended the refinements described below for the NCHRP 3- 70 MMLOS pedestrian model. The original segment level Pedestrian Level of Service model was developed for FDOT in 2000 and presented at the Transportation Research Board’s Annual Meeting in 2001. That original model was developed based upon data obtained during an in-field Walk-for-Science event. As with other models used to evaluate transportation facilities, as the Pedestrian LOS model for segments was implemented by transportation practitioners, it was applied in roadway environments not captured during the original data collection event. This practical application of models often leads to refinements based upon insights obtained during application. For example, operational use of the Pedestrian LOS model for segments has led to one such refinement of the original model. The NCHRP 3-70 model evaluation process has included highly focused sensitivity analyses of the model in additional settings. Based upon the results of these analyses and other prior applications across the U.S., we suggest a couple of further refinements. Four minor refinements are discussed in this section. 1. A modification of the on-street parking effect coefficient,

Page 13 2. The inclusion of the impacts of shoulder striping on the lateral separation to motor vehicle traffic, 3. A maximum placed on the effect of additional sidewalk width, and 4. A low-volume roadway adjustment for streets without sidewalks. This paper first discusses the initial three potential revisions and the limited impact they would have on the form of the model. The final revision is discussed separately as it is typically associated with rural roadways or residential streets. Methodology On-Street Parking Coefficient The first recommended refinement to the model is to increase the on-street parking coefficient (fp) from 0.2 to 0.5. The original Walk-for-Science route used for data collection during the original Pedestrian LOS model (for a variety of course continuity/logistical reasons) did not have a wide range of traffic volumes along segments with on-street parking. The value of the on- street parking coefficient (fp = 0.2) was based upon those data points. Our application of the Pedestrian LOS model for segments across the country has led us to think that the influence of on-street parking on pedestrians’ perceptions of safety and comfort might be greater than is represented by the on-street parking adjustment currently in the model. During the NCHRP 3-70 Phase III analyses, the evaluating agencies tested additional locations along their roadways for the impacts of on-street parking in conditions with an increased upper range of adjacent traffic volumes. This yielded additional “data points” for the refinement of the segment level model. These “data points” confirmed our observations that a higher value for fp might be appropriate. Consequently, Sprinkle staff has now tested various values for fp and now recommend fp = 0.5 as a value for this coefficient to represent a greater range of adjacent traffic volumes. Impacts of Shoulder Striping The second proposed refinement is to represent expected improvements to pedestrian LOS from the inclusion of a bike lane or other paved space to the right of the travel lane. The original Pedestrian LOS data analysis and modeling did not reveal a significant correlation between the presence of a striped shoulder or a bike lane or a parking lane without cars and the perceptions of pedestrians. This is not to say that such a correlation did not exist, just that in the presence of the additional space it was not found to be statistically significant. Consequently, the overall pavement width from the edge of pavement to the left side of the rightmost lane, Wt, is used to represent the portion of the lateral separation term represented by pavement width. While providing “acceptable” and accurate results, this term does not capture subtle improvement in pedestrian LOS that may be expected by analysts or agencies contemplating including bike lanes (or possibly striping low use on-street parking). To illustrate this idea, imagine a total lane width, Wt, of 17 feet. If motorists drive in the center of the lane then cars would be centered 8.5 feet from the edge of pavement. If this outside lane width is striped as a 12-foot lane with a 5-foot bike lane, the motorists would drive centered in the 12-foot lane, centered 11 feet from the edge of the pavement. Additionally, on a roadway

Page 14 with a 20-foot outside lane with non-striped but allowed on-street parking, but no cars actually parking on it, motorists would likely track centered 10 feet from the curb. However if there are some parked cars, we’ve assumed 25% or greater, motorists would likely shift left about 10 feet and drive centered 15 feet from the edge of the roadway. As can be seen in the above examples, the potential additional separation is one-half any space provided to the right of the travel lane. We thus recommend replacing Wol, width of the outside lane, with Wt, total width of outside lane (and shoulder) pavement. This would also make the variables consistent with the Bicycle LOS segment model. To accommodate the additional separation between motor vehicles and pedestrians resulting from the striping of a shoulder, we also recommend adding 0.5Wl into the lateral separation term. Further, we recommend setting Wl =10 if un-striped on-street parking occupancy is 25% or greater. Maximum Effect of Sidewalk Width Several users of the segment model have noted that the sidewalk presence coefficient, fsw, reaches a maximum effect at 10 feet; application for wider sections reduces the sidewalk presence coefficient. The original (current) model development addressed this by setting fsw =3 for the infrequent cases when sidewalk widths (exclusive of buffers) exceed 10 feet. We recommend introducing this control condition into the sidewalk presence coefficient definition. Comparison of Pedestrian LOS for Segments Model with Proposed Refinement The current Pedestrian Level of Service Model for segments is as follows: PLOS = -I.2276 ln (Wol + Wl + fp x %OSP + fb x Wb + fsw x Ws) + 0.0091 (Vol15/L) + 0.0004 SPD2 + 6.0468 Where ln = Natural log Wol = Width of outside lane Wl = Width of shoulder or bicycle lane fp = On-street parking effect coefficient (=0.20) %OSP = Percent of segment with on-street parking fb = Buffer area barrier coefficient (=5.37 for trees spaced 20 feet on center) Wb = Buffer width (distance between edge of pavement and sidewalk, feet) fsw = Sidewalk presence coefficient (= 6 – 0.3Ws) Ws = Width of sidewalk Vol15 = Volume of motorized vehicles in the peak 15 minute period L = Total number of directional through lanes SPD = Average running speed of motorized vehicles traffic (mi/hr) The proposed refinement to the Pedestrian Level of Service Model for segments is provided below: PLOS = -I.2276 ln (Wt + 0.5Wl + fp x %OSP + fb x Wb + fsw x Ws) + 0.0091 (Vol15/L) + 0.0004 SPD2 + 6.0468 Where ln = Natural log

Page 15 Wt = total width of outside lane (and shoulder) pavement Wl = Width of shoulder, bicycle lane, and striped parking; or If there is un-striped parking and %OSP≥25 then Wl=10’ to account for lateral displacement of traffic fp = On-street parking effect coefficient (=0.50) %OSP = Percent of segment with on-street parking fb = Buffer area barrier coefficient (=5.37 for trees spaced 20 feet on center) Wb = Buffer width (distance between edge of pavement and sidewalk, feet) fsw = Sidewalk presence coefficient (fsw = 6 – 0.3Ws if Ws≤10, otherwise fsw = 3) Ws = Width of sidewalk Vol15 = Volume of motorized vehicles in the peak 15 minute period L = Total number of directional through lanes SPD = Average running speed of motorized vehicles traffic (mi/hr) Low Volume Roadways without Sidewalks Sprinkle Consulting has applied the Pedestrian LOS model for sidewalks on a wide variety of roadways across the United States including low volume rural roadways and residential streets without sidewalks. On these streets without sidewalks, we have observed that the fixed geometric definition of Wt and Wl seems to overestimate the impact of motorists on pedestrians when the volume of motor vehicles is relatively low; on very low volume streets or roads, the effective width approaches two times the geometric width as the traffic volume approaches zero. While this might be a relatively uncommon occurrence on urban arterial roadways, we feel it is worth consideration when the Pedestrian LOS model for segments is applied on across a rural network or neighborhood streets. Thus, just as is the case with the Bicycle LOS for segments model, when we apply the Pedestrian LOS model for segments to a network that includes low volume streets, we incorporate a low volume adjustment into our Pedestrian LOS calculations. The adjustment is the same as the adjustment factor for the Bicycle LOS for segments: Where the AADT ≤ 4000 vpd, Wv is substituted for Wt, and Wv = Wt*(2-0.00025*AADT) To accommodate the full range of potential roadway volumes, under conditions where no sidewalk exists, we feel this volume adjustment should be included in the Pedestrian LOS for segments model. User Guide In the field tests that occurred in central business districts, the users of the MMLOS method had frequent questions about the treatment of street furniture and planter boxes (as opposed to planter strips) in the pedestrian LOS model. The guidance given was for the user to assess the extent to which street furniture provided the same perceived degree of separation between pedestrian and traffic as a tree and to use an approximate equivalent value in the MMLOS method. If planter boxes were spaced so that they acted as the equivalent of a continuous planter strip (i.e. the pedestrians effectively use only the strip of the sidewalk further removed from the street), then the distance between the street curb and the planter boxes should be treated as the equivalent of a planter strip of that same width.