Gattis, Dammalapati, Cotton, Cotton 3rd Urban Street Symposium 1 June 24-27, 2007 Seattle, Washington CROSS SECTION WIDTH FOR PARALLEL PARKING by J. L. Gattis, Ph.D., P.E. Srinivas Dammalapati Joshua Cotton Joseph Cotton Corresponding Author: J. L. Gattis, Ph.D., P.E. Mack-Blackwell Transportation Center 4190 Bell Engineering Center Fayetteville, AR 72701 voice (501)575-3617 fax (501)575-7168 firstname.lastname@example.org ABSTRACT The growth of alternative roadway design approaches, such as Traditional Neighborhood Design and Context Sensitive Design, has focused new attention on urban street design standards. Some have called for urban street cross sections that are narrower than those recommended in established design references. One cross section design element for which there is some variation among different published guidelines and practices is the width provided for on-street parallel parking. This paper reports the findings from field measurements made to determine how much width was actually occupied by parallel-parked passenger cars, both in commercial and in residential areas. Gattis, Dammalapati, Cotton, Cotton 2 CROSS SECTION WIDTH FOR PARALLEL PARKING by J. L. Gattis, Ph.D., P.E., Srinivas Dammalapati, Joshua Cotton, Joseph Cotton INTRODUCTION The growth of alternative roadway design approaches, such as Traditional Neighborhood Design and Context Sensitive Design, has focused new attention on urban street design standards. Some adherents of these newer, alternative design perspectives have proposed and adopted urban street cross sections that are not as wide as some of the cross sections based on recommendations found in established references. One of the cross sectional elements about which the various references disagree, and for which one can find a range of recommended minimum acceptable values, is the width allowed for and allocated to on-street parallel parking. The lack of agreement about the width needed to accommodate on-street parallel parking prompted a study to determine how much width was actually being occupied by parallel-parked vehicles, both in commercial and in residential areas. Those involved in the study made field measurements of parallel-parked vehicles. Except where otherwise stated, the scope of the literature review and the field studies was confined to the passenger car or P design vehicle, which reflects the limiting attributes of most standard passenger cars (PC) such as sedans, coupes, station wagons, as well as minivans, sportutility vehicles (SUV), and pickup trucks (PU). It represents most of the non-commercial vehicle fleet that the general public owns and drives. LITERTURE REVIEW A review of available literature helps illustrate divergent points of view related to cross sectional width needed for on-street parallel parking. Both traditional and alternative sources were reviewed. Traditional Design Criteria The chief sources of “traditional” street cross section design criteria are the policy publications by AASHTO (American Association of State Highway and Transportation Officials), and the Manual on Uniform Traffic Control Devices (MUTCD) by the Federal Highway Administration. A Policy on Highway Types (Geometric), published by American Association of State Highway Officials (AASHO) in 1940 (1) stated the following about parking. “The width of parking lane depends upon the proposed method of parking and character of traffic. For parallel parking of passenger vehicles a width of 7 feet has been used extensively. It may be wide enough for side streets where few trucks use the streets and travel is at low speed. On through streets, lanes for parallel parking of passenger vehicles should be at least 8 feet wide if the adjacent traffic lanes are to be utilized without encroachment onto other traffic lanes.” The focus of the 1957 AASHO Red Book (2) was urban freeways and arterials. Unlike the later AASHTO Green Books, it did not contain chapters on local or collector street design. The lowest category listed in the table of contents was “major streets”. The discussion of onstreet parking for passenger cars on major streets included the following. “Passenger vehicles now in use are 5 ft.-8 in. to 6 ft.-10 in. in width. A vehicle parked alongside a curb likely will not be stopped against the curb but with the right edge 4 to 8 inches from it. Or it may be at a slight angle so that one Gattis, Dammalapati, Cotton, Cotton 3 extremity is 6 to 12 inches from the curb. Thus, the actual street space occupied 6.5 to 7.5 feet from the curb.” The discussion proceeded to suggest an additional clearance of 3 to 5 ft (0.9 to 1.5 m), for a total parking lane width of 10 to 12 feet (3.0 to 3.6 m). The 1973 Red Book (3), while briefly mentioning some aspects of collector and local streets, also did not contain chapters addressing design of these lesser functional classes. In the “Cross Section” chapter, the 1990 (4) and the 1994 (5) Green Books suggested an 8 ft (2.4 m) “desirable minimum width of a parking lane”. This dimension was recommended for both arterial and collector streets. The discussion noted that a width of 10 to 12 ft (3.0 to 3.6 m) was desirable to provide greater clearance, and this wider dimension should be used when bike routes are adjacent to the parked vehicles, to permit bicyclists to pass by open car doors. In the “Local Roads and Streets” chapter, AASHTO (5) mentioned a cross section width to accommodate a 12 ft (3.6 m) wide center travel lane with 7.2 ft (2.2 m) wide parking lane on either side. Lanes on the local street were unmarked. The “Collector Roads and Streets” chapter listed “a parallel parking lane from 7 to 10 ft (2.1 to 3.0 m) in width” in residential areas, and 8 to 10 ft (2.4 to 3.0 m) in commercial areas. The 2001 or fourth edition (6) of the Green Book stated that for urban arterial roadways, a “parking lane width as narrow as 2.4 m (8 ft) may be acceptable” and recommended a width of 3.0 to 3.6 m (10 to 12 ft). On urban collector streets in residential areas, a width of 2.1 to 2.4 m (7 to 8 ft) was suggested to accommodate on-street parking. The chapter on local streets gave a width for residential area parking of 2.2 m (7 ft) in one place and a minimum of 2.1 m (7 ft) in another. An 8 ft (2.4 m) width was the suggested minimum width in commercial areas. The 2004 Green Book “Cross Section Elements” chapter and other passages (7) contain similar statements. The discussion noted that a width of 10 to 12 ft (3.0 to 3.6 m) was desirable to provide greater clearance, and this wider dimension should be used when bike routes are adjacent to the parked vehicles, to permit bicyclists to pass by open car doors. A 7 ft (2.1 m) parking width is mentioned in the 1948 Manual on Uniform Traffic Control Devices (MUTCD) (8). The 1961 (Figure 2-10) (9), 1971 (Figure 3-16) (10), 1978 (Figure 3-16) (11), 1988 (Figure 3-16) (12), 2001 (Figure 3B-17) (13) , and 2003 (Figure 3B-18) (14) editions of the MUTCD showed an 8 ft (2.4 m) wide marked parallel parking space. Other Design Guides By their very nature, AASHTO and FHWA publications are positioned to have a wider distribution and acceptance than many other design guides are. Some alternative design guide sources were identified and reviewed. Residential Streets, jointly published in 1990 (15) by American Society of Civil Engineers (ASCE), National Association of Home Builders (NAHB), and the Urban Land Institute (ULI), showed a parking lane width of 6 to 7 ft (1.8 to 2.1 m) on an “access” street, and an 8 ft (2.4 m) width for subcollector and collector streets. The document states that for residential streets, “Designers should select the minimum width that will realistically satisfy all realistic needs.” A moving lane of 10 ft (3.0 m) width was shown for access, subcollector, and collector streets, for a total street width of 22 to 24 ft (6.6 to 7.2 m) for access streets, 26 ft (7.8 m) for subcollector streets, and 36 ft (10.8 m) for collector streets. The third edition of Residential Streets, published in 2001 (16), showed a parking lane width dimension of 6 to 7 ft (1.8 to 2.1 m) for the local streets and 8 ft (2.4 m) for residential collector streets. A moving lane of 11 to 12 ft (3.3 to 3.6 m) width was shown for local streets Gattis, Dammalapati, Cotton, Cotton 4 with parking not expected or restricted to one side and a 10 to 14 ft (3.0 to 4.2 m) wide moving lane for local streets with parking on both sides. The widths recommended in the third edition were somewhat greater than those in the second edition. A guide prepared for the state of Florida (17) promoted a number of relatively narrow street width design standards. It listed “7- or 8-foot parking lanes” for collectors. Design drawings prepared for the Portland metropolitan region (18) showed a width of 7 ft (2.1 m) allocated for parking in a number of streetscapes, including commercial areas. The design guidelines state “The preferred on-street parking lane width for parallel parking is 7 feet.” The Main Street handbook developed by the Oregon Department of Transportation (19) for Oregon communities mentioned a parallel parking width of 7 ft (2.1 m) on main streets. It also reported a scenario in which to increase the sidewalk width the parking width was reduced from 8 ft (2.4 m) to 7 ft (2.1 m). One of a series of articles discussing various aspects of traffic calming in an ITE Journal listed a minimum 7.2 ft (2.2 m) parking lane width for local streets, and an 8 ft (2.4 m) minimum for collectors (20). In another article (21), authors reported that a street standard for their city had been modified to reduce the parking lane width from 2.4 m (8 ft) to 2.1 m (7 ft). Given the context of the article, it can be assumed the reduced width applied to a residential street cross section. RESEARCH OBJECTIVES AND METHODS The objectives of the study were simple and straightforward. 1. measure the actual cross section width occupied by parked vehicles, in both commercial and residential environments 2. compare the measured dimensions with those proposed in various sources The data were collected in two cities, Mobile, Alabama and Fayetteville, Arkansas. Mobile is an older, established city with a population of about 200,000, in a growing metropolitan area of about 500,000 people. Fayetteville is a city of over 60,000 near the south end of a metropolitan area of over 200,000. In the 1990s, the area was one of the fastest growing metropolitan areas in the country. Street Environment and Data Pool The data were collected in downtown or older near-downtown commercial areas, in older residential neighborhoods, and near a university campus. These areas were targeted because onstreet parking regularly occurs on the streets in these areas. Many of these streets also possessed another attribute desired for the study, being rather narrow. It was hypothesized on a narrow street, any given driver would be less likely to position their parked vehicle randomly, and more likely to attempt to park close to the curb. The number of parked vehicles in the sample simply reflects the number of vehicles that were parallel-parked along the streets at the time the measurements were made. As data were collected, the vehicles were classified as either passenger car (PC), sport utility vehicle or van (SUV/Van), or pickup (PU). Table 1 lists the streets on which parked vehicle offsets from the curb were measured, and what types of vehicles were encountered. In the commercial areas, the individual parking spaces were delineated or marked. The streets also had marked center and/or lane lines. In the residential areas, the on-street parking spaces were not marked, and most of the studied residential area streets had no center or lane line Gattis, Dammalapati, Cotton, Cotton 5 markings. As the table shows, some parking spaces in the campus area were marked but most were not; center and lane lines were not present. TABLE 1 Description of Streets and Vehicle Types Street Name Nominal Width (ft) Parking Lanes No.of Marked Moving Lanes Vehicle types PC SUV/ PU Van Commercial streets - Mobile E. Daulphin St.(north) 16.3 E. Daulphin St.(south) 20.7 N. Jackson St. 21.5 Royal St. (east) 26.7 Royal St. (west) 26.7 State St. 22.9 Commercial streets - Fayetteville N. Block 22.0 E. Center 19.7 W. Center 20.0 N. East (south) 25.7 N. East (north) 16.3 E. Mountain 23.0 W. Mountain 22.3 Dickson Street 24.0 Residential streets - One-way Meadow 25.7 Watson 22.3 Residential streets - Two-way Boles 19.7 Lafayette 40.7 Sutton 23.7 Washington 30.7 Walnut 19.7 Willow (north) 27.0 Willow (south) 20.7 Campus – marked spaces Ark. Avenue (east) Ark. Avenue (west) Campus - unmarked spaces Douglas Lafayette Lindell Oakland Storer 1 1 1 1 1 1 2 2 2 2 2 2 6 32 2 1 4 6 2 16 1 1 3 1 0 6 1 1 2 3 2 2 2 2 2 2 2 2 2 2 2 2 1 2 2 1 8 23 7 5 20 20 10 27 6 6 0 2 5 4 2 12 5 8 1 1 5 6 1 6 1 1 none none 11 5 2 1 3 0 1 2 1 2 1 1 1 none 2 none none none none none 9 11 7 9 8 6 6 1 2 1 4 3 1 3 2 4 3 2 2 0 1 26.0 26.0 1 1 1 1 19 23 11 10 2 6 24.0 28.3 27.8 27.8 20.1 1 1 2 2 1 none 2 none none none 14 9 22 26 6 8 4 7 9 4 8 3 10 11 3 NOTE: “Nominal width” where parking lanes were marked was measured between or inside the parking lanes; where there were no parking space markings, width was measured between curb faces Gattis, Dammalapati, Cotton, Cotton 6 The streets on which data were collected in Fayetteville are in level to slightly rolling terrain. Curb faces are vertical, and curb heights typically do not exceed 6 in (150 mm). Extreme vertical differences between elevations of the street surface and the gutter surface, which could cause drivers to shy away from the curb, were typically not observed. Speeds on these streets are typically in the 20 to 35 mph (30 to 55 km/h) range, and traffic volumes were well below those that would cause congestion. Data Collection Two-person teams measured the offset distance from the curb face to the bottom of the side of the tire toward or facing the moving traffic. Measurements were made to the tire in order to avoid coming in contact with any part of the vehicle other than the bottom of the tire. A pole was employed to hold the tape end next to the tire. This allowed the team-member working on the moving-traffic side of the parked vehicle to stand up while making the measurement, and avoid bending over. Measurements were made at both front and rear tires, with the greater of the two offset distances used in subsequent data analysis. All measurements were made during the daytime in clear, dry weather. On one-way streets, dimensions for vehicles parked on both sides of the street were taken. Measuring to the tire does not reflect the entire width occupied by the parked vehicle, since the edge of the vehicle body may extend past the tire. To estimate a total offset (i.e., the entire width occupied by a parked vehicle), a four-person crew separately measured 50 vehicles parked on level terrain. A tape was stretched from the outer edge of the left front tire to the outer edge of the left rear tire. Then, a surveying pole with a bulls-eye level attached was placed in a vertical position extending from the pavement surface up to the vehicle body, and the offset distance from the pole to the stretched tape was measured. Distances from the tire edge to the both the edge of the vehicle body and to the outer edge of the left-side mirror were read. Measurements were made to the nearest 0.5 inch (13 mm). These dimensions were evaluated and standard values for the additional width out to the edge of the body and out to the edge of the mirror were determined. After other analyses had been performed, these standard values were added to measurements made from the curb to tire. Data Analysis Offset distances from the curb face to the outer edges of the left tires were measured for 88 vehicles parked in marked spaces in a commercial area of Mobile, Alabama. In Fayetteville, Arkansas, offsets of 202 vehicles in marked spaces in two commercial areas, 107 vehicles in unmarked spaces in residential areas, 71vehicles in marked spaces in a campus area, and 144 vehicles in unmarked spaces in a campus area were measured. The three handicapped spaces and the nine vehicles which were parked on a one-way street with no marked stalls on the other side of the street were excluded from the commercial-area data set of Fayetteville, Arkansas, which left 190 vehicles in that commercial area. Statistical tests were performed to determine if certain differences were statistically significant, with ∀ = 0.05. The following t-test equation comparing means from two samples of unequal size, where N is sample size and s is sample standard deviation, was used (22). offset 1 − offset 2 s1 s 2 = t × + N1 N 2 2 2 Gattis, Dammalapati, Cotton, Cotton 7 Table 2 describes the tested pairs and the outcome of the test for each pair. Based on these outcomes, the data from the campus area were kept separate from the residential area, while the data from the two Fayetteville commercial areas were combined. TABLE 2 Tests for Significant Differences Between Area Types Comparing the two means of Mean 1 Mean 2 Was the difference significant? Mobile commercial vs. Fayetteville commercial downtown Fayetteville commercial: downtown vs. near-downtown 6.81 6.93 no 6.93 6.84 no Mobile commercial vs Fayetteville commercial near-downtown Campus area unmarked vs. Residential areas unmarked 6.81 6.84 no 6.07 6.21 yes Table 3 shows the distribution of the three vehicle groups. Note the higher percentages of SUV/van in the campus-marked area, and the higher percentage of pickup trucks in the campus-unmarked area. As a check, data were examined to ascertain if any of the three vehicle classes were overor underrepresented on any of the street widths. Commercial and residential streets were analyzed separately. The Chi-square test of independence was performed on a contingency table. The numbers of vehicles within each class were aggregated by 1.0 ft (0.3 m) width increments. The outcomes of both the commercial and the residential group Chi-square tests found independence. That is, the numbers of each of the three vehicle classes were sufficiently proportional in each of the width increments. That fraction of the total street cross section intended for parking may be marked, or it may be unmarked and as such be co-mingled with overall street or outer lane width. In order to examine parking offset as a function of traveled-way width, some simplifying assumptions about streets with no marked center or lane lines were made. Based on knowledge of traffic on each street, the number of lanes that can and do normally move simultaneously on those streets lacking center or lane line markings was identified. A width of 7.5 ft (2.3 m) for each lane of onstreet parking that did occur was subtracted from the total street width. The remaining width was divided by number of lanes that can be observed moving when parking is present to arrive at the width of a moving lane. The assumed 7.5 ft (2.3 m) width was chosen as a “mid-way” value between the traditional 8.0 ft (2.4 m) width and the commonly-proposed alternative 7.0 ft (2.1 m) parking lane width. Gattis, Dammalapati, Cotton, Cotton 8 TABLE 3 Descriptors by Area and Vehicle Group Area type Vehicle group N % Mean offset (ft) 6.79 6.97 7.06 6.88 6.15 6.25 6.43 6.21 6.43 6.94 6.80 6.62 5.90 6.15 6.37 6.07 Standard. deviation (ft) 0.60 0.65 0.50 0.61 0.55 0.30 0.56 0.53 0.24 0.32 0.39 0.37 0.43 0.42 0.33 0.45 Commercial marked Passenger Car SUV/Van Pickup (PU) All Combined Passenger Car SUV/Van Pickup (PU) All Combined Passenger Car SUV/Van Pickup (PU) All Combined Passenger Car SUV/Van Pickup (PU) All Combined 171 61 46 278 72 18 17 107 42 21 8 71 77 32 35 144 62% 22% 17% 100% 67% 17% 16% 100% 59% 30% 11% 100% 53% 22% 24% 100% Residential unmarked Campus marked Campus unmarked Adjustment Factor As previously mentioned, four people measured 50 vehicles to obtain distances from the tire edge to the edge of the main vehicle body and to the edge of the left-side mirror. Since the tires of two vehicles actually protruded past the edge of the vehicle body, the measurements from the tire edge to the main body edge ranged from -0.125 to 0.25 ft (-0.038 to 0.076 m). For passenger cars only, most measured no more than 0.2 ft (0.06 m) from the tire to the body edge. For all vehicles, dimensions from the tire to the edge of the side mirror ranged from 0.25 ft (0.076 m) to 1.0 ft (0.305 m). Few passenger car side mirrors extended more than 0.65 ft (0.2 m) past the tire edge. The values in the 80th to the 90th percentile range (which did include passenger cars) were slightly above 0.65 ft (0.2 m). Vehicles seldom park perfectly parallel to the curb edge alignment, so it is common to observe a slight skew in a parked vehicle’s position. Therefore, the protrusion of a side mirror past the tire edge may effectively be less than indicated by the above measurements. Unless the vehicle is parked at an extreme skew to the curb edge, this difference will be slight. To reflect the additional width (i.e., past the tire edge) occupied by the vehicle body and the mirror, adjustment factors of 0.13 ft (0.04 m) and 0.52 ft (0.16 m) were added respectively to the measured offset width. These adjustments were included in cumulative plots of the data. RESULTS Table 4 presents, for each of the four area types, the number of vehicles measured and the minimum, average, and maximum widths observed. Gattis, Dammalapati, Cotton, Cotton 9 TABLE 4 Descriptive Statistics of Parked Vehicles Area Type Count Mean offset (ft) 6.88 6.81 6.92 6.21 6.62 6.07 Standard deviation (ft) 0.61 0.57 0.64 0.53 0.37 0.45 Minimum Maximum offset offset (ft) (ft) 5.3 5.3 5.6 5.3 5.8 5.0 9.7 9.2 9.7 8.5 7.5 7.0 Commercial marked Mobile Fayetteville Residential unmarked Campus marked Campus unmarked 278 88 190 107 71 144 Offset and Street or Traveled Lane Width The data were examined to determine if either narrower streets or narrower traveled lanes were associated with vehicles being parked closer to the curb. Among the area types in the study, the unmarked residential streets exhibited the greatest range of street widths. Examining the data in Table 5, there is no apparent trend between the measured offset values and width. Vehicles parked on the more narrow residential streets had measured offsets similar to those on wider streets. Likewise, no great differences were observed between offsets on a range of the wider and a range of the narrower traveled-lane widths. In fact, vehicles on both the wider streets and along the wider traveled lanes were positioned slightly closer to the curb. This could be a response by those parking their vehicles on the wider streets to the more typical volumes and speeds on the wider streets that were in the study (the narrow streets in the study have very low volumes and speeds). A linear regression analysis on the unmarked residential parking offset widths was performed, with the moving lane width as the independent variable and the parking offset width as the dependent variable. For residential area parking on streets with moving lane widths that ranged from 8.8 to 15.8 ft (2.68 to 4.82 m), the computed R2 value was 0.01. TABLE 5 Residential Parking Offset-to-Tire and Street Width Widths (ft) Low High Entire Street Width 30.1 to 40.0 21.9 to 26.5 19.1 to 20.3 19.1 to 40.0 (all) Unmarked Residential area offset to tire N avg. Min to Max 85th% (ft) (ft) (ft) 32 40 35 107 6.1 6.2 6.3 6.2 5.3 5.3 5.3 5.3 to to to to 7.1 8.0 8.0 6.6 6.5 6.6 7.0 6.6 Moving Lane Width (comparing “wide” with “narrow”) Wide – 12.6 to 15.8 27 6.3 5.4 to 8.0 6.6 Narrow – 8.8 to 9.5 23 6.4 5.6 to 8.5 6.8 Gattis, Dammalapati, Cotton, Cotton 10 Examination of the data revealed that pickup trucks were overrepresented in the upper range of offset values. This was observed in both the commercial and in the residential data. Cumulative Distribution Plots A review of a cumulative distribution graph can help identify patterns in and the outliers in a set of data. In the built world in which we live, it is not uncommon to design for all but the outliers. Figure 1 is a cumulative plot of the commercial area parking offsets, and Figure 2 shows the distribution of the residential area parking offsets. An inspection of the plot from the marked, commercial parking area shows the shape of the plotted line for the vehicle body does not break until around the 97th percentile value, about 8.0 ft (2.43 m). With additional width for the side mirror, this increases to about 8.4 ft (2.56 m). The break for the vehicle body in an unmarked, residential setting is not as well defined, and occurs somewhere at or above the 90th percentile value, perhaps between or 6.8 to 7.3 ft (2.07 to 2.23 m). Considering the side mirror, this value is about 7.7 ft (2.35 m). The plots of the campus-marked and campus-unmarked data did not show clear breaks. The 90th percentile values were about 7.3 ft (2.23 m) for marked spaces and 6.7 (2.04 m) for unmarked parking. With the additional width for the mirror, these values increase to 7.7 (2.35 m) and 7.1 ft (2.16 m). DISCUSSION The offsets from the curb face to the side of the tire facing moving traffic were measured for vehicles parked in commercial areas with marked spaces, residential areas with unmarked spaces, and a campus area with both marked and unmarked spaces. To arrive at a total parkedvehicle offset from the curb, and additional width of vehicle that extends past the tire was added to the dimension measured. The following points are noted. Χ In both the commercial and the residential areas, pickup trucks extended over 0.25 ft (0.081 m) farther out into the street than did passenger cars. A disproportionate number of pickups were observed in the upper range of the parking offset values. Χ Over the ranges of street widths and traveled lanes of the residential streets in this study, neither street width nor traveled lane width seemed to affect the parking offset width. Χ For all vehicles, the average offset of those parked in the marked commercial areas, 6.9 ft (2.1 m), extended further into the street than did the average offset of those parked in the unmarked residential areas, 6.2 ft (1.9 m). Χ The cumulative plots of parking offset widths suggest that in order to accommodate approximately 90% of the parked vehicles in this study, the marked commercial spaces would need to be about 8.0 ft (2.4 m) wide, and both the residential and the campus unmarked spaces about 7.2 ft (2.2 m) wide. Note that these dimensions do not provide any width to accommodate opening vehicle doors on the side of moving traffic. There are a number of possible explanations of why vehicles parked in the marked commercial spaces occupied a greater width than did those parked in the unmarked residential areas. Two of the following possible explanations were offered by persons who are not authors. 1. The difference was due to random chance. 2. The difference was due to unidentified factors peculiar to the area. Gattis, Dammalapati, Cotton, Cotton 11 CUMULATIVE PLOT - COMMERCIAL Car Body 100% 80% Rank (%) 60% 40% 20% 0% 5.0 6.0 7.0 8.0 9.0 10.0 11.0 Max. Distance Curb Face to Car Body (ft) Car Mirror FIGURE 1 Commercial parking offset width CUMULATIVE PLOT - RESIDENTIAL Car Body Car Mirror 100% 80% Rank (%) 60% 40% 20% 0% 5.0 6.0 7.0 8.0 9.0 10.0 11.0 Max. Distance Curb Face to Car Body (ft) FIGURE 2 Residential parking offset width Gattis, Dammalapati, Cotton, Cotton 12 FIGURE 3 Campus marked offset width FIGURE 4 Campus unmarked offset width Gattis, Dammalapati, Cotton, Cotton 13 3. 4. 5. 6. 7. The difference was due to the presence of parking meters and other objects close to the commercial streets, which were not present along the residential streets. The difference was due to a higher turnover rate in the commercial areas: drivers parking for a shorter duration may expend less effort aligning their vehicles. The difference was due to greater through lane vehicle volumes and/or speeds in the commercial area. Once they initially enter a parking space, drivers parking their vehicles in the commercial area were less prone to make repeated maneuvers to better position their vehicles, lest they strike passing through vehicles. But traffic volumes and speeds in the study environments may have been low enough so that this was not a factor. The difference reflects a “use all that is available” mentality. Since a marked 8 ft (2.4 m) width was often present, motorists making the parking maneuvers tended to use all of the space provided. If the marked spaces had been narrower, perhaps they would have positioned themselves closer to the curb. The difference reflects the difficulty in making a parallel parking maneuver as opposed to pulling over to the curb. In the areas studied, only a small percentage of the residential area curbsides are occupied by parked vehicles, so parking maneuvers usually consist of simply aiming the vehicle toward the curb, and then aligning the vehicle parallel to the curb while coming to a stop. In contrast, almost all of the commercial area parking spaces are typically occupied during the day, so parking in the marked spaces in the commercial area is much more likely to involve parallel parking maneuvers. Given the greater difficulty of the parallel parking maneuver, it is not surprising that vehicles parked in the marked spaces in the commercial area were positioned farther from the curb than those parked in the residential area. To test this possibility, parking offsets would have to be measured in an unmarked area where curb space is occupied to the extent that drivers usually have to make a parallel parking maneuver. CONCLUSION At the beginning of this study, the possibility that the influence of narrow streets would cause drivers to park their vehicles closer to the curb, and therefore take less width than called for in traditional parking width guidelines, was recognized. However, the measured data from the streets in this study do not support this hypothesis. An urban street cross section that provided 7.0 ft (2.1 m) or less of parking width would not accommodate 20% or more of the vehicles parallel parked along the curb. The traditional 8.0 ft (2.4 m) width did not appear to be excessive for the commercial area parking observed in this study. A minimum width of at least 7.2 ft (2.2 m) would accommodate most of those parked in unmarked residential areas. These dimensions do not include any width to accommodate opening vehicle doors on the side of moving traffic; to accommodate this need, it could be argued that additional width is needed along streets with higher volumes or speeds. ACKNOWLEDGMENT A special thank you goes to Joe Ruffer and James Foster of Mobile County Alabama, for providing parking width data. The support of the Mack-Blackwell National Rural Transportation Study Center at the University of Arkansas (through a grant from the U. S. Department of Transportation University Centers Program) made this research possible. The views expressed herein are those of the authors alone. Gattis, Dammalapati, Cotton, Cotton 14 REFERENCES 1. “A Policy on Highway Types (Geometric).” American Association of State Highway Officials, Washington, DC, 1940. 2. A Policy on Arterial Highways in Urban Areas. American Association of State Highway Officials, Washington, DC, 1957. 3. A Policy on Design of Urban Highways and Arterial Streets. American Association of State Highway and Transportation Officials, Washington, DC, 1973. 4. A Policy on Geometric Design of Highways and Streets. American Association of State Highway and Transportation Officials, Washington, DC, 1990. 5. A Policy on Geometric Design of Highways and Streets. American Association of State Highway and Transportation Officials, Washington, DC, 1994. 6. A Policy on Geometric Design of Highways and Streets. American Association of State Highway and Transportation Officials, Washington, DC, 2001, pp. 378, 379, 396, 397, 438, 482. 7. A Policy on Geometric Design of Highways and Streets. American Association of State Highway and Transportation Officials, Washington, DC, 2004, pp. 374, 375, 392, 393. 8. Manual on Uniform Traffic Control Devices. Joint Committee on Uniform Traffic Control Devices, Washington, DC, 1948. 9. Manual on Uniform Traffic Control Devices. National Joint Committee on Uniform Traffic Control Devices, Washington, DC, 1961. 10. Manual on Uniform Traffic Control Devices. National Joint Committee on Uniform Traffic Control Devices, Washington, DC, 1971. 11. Manual on Uniform Traffic Control Devices. Federal Highway Administration, Washington, DC, 1978. 12. Manual on Uniform Traffic Control Devices. Federal Highway Administration, Washington, DC, 1988. 13. Manual on Uniform Traffic Control Devices. Federal Highway Administration, Washington, DC, 2000. 14. Manual on Uniform Traffic Control Devices. Federal Highway Administration, Washington, DC, 2003. 15. Residential Streets (2nd Ed.). American Society of Civil Engineers, National Association of Home Builders, Urban Land Institute, Washington DC, 1990. 16. Kulash, W.M., ed. Residential Streets, 3rd Ed, American Society of Civil Engineers, Institute of Transportation Engineers National Association of Home Builders, Urban Land Institute, Washington, D.C., 2001, p.24. 17. Ewing, R., C. Heflin, M. DeAnna, and D. Porter. Best Development Practices. Florida Dept. of Community Affairs, Tallahassee, FL, May 1995. 18. Creating Livable Streets. Metro Regional Services, Portland, OR, November 1997. 19. Oregon Department of Transportation. Main Street…when a highway runs through it: A Handbook for Oregon Communities, 1999, p. 44, 90. 20. Leonard, J. D., and W. J. Davis. Urban Traffic Calming Treatments, ITE Journal, Vol. 67, No, 8, August 1997, p. 36. 21. West, J., and A. Lowe. Integration of Transportation and Land Use Planning through Residential Street Design, ITE Journal, Vol. 67, No. 8, Aug 1997, pp.48-51. 22. Ott, L. An Introduction to Statistical Methods and Data Analysis, 2nd ed., PWS Publishers, Boston, MA, 1984.