Journal of Transport Geography 19 (2011) 51–58
Contents lists available at ScienceDirect
Journal of Transport Geography journal homepage: www.elsevier.com/locate/jtrangeo
Time use, travel behavior, and the rural–urban continuum: Results from the Halifax STAR project Hugh Millward a,*, Jamie Spinney b a b
Department of Geography, Saint Mary’s University, Halifax, NS, Canada B3H 3C3 School of Geography and Earth Sciences, McMaster University, Hamilton, ON, Canada L8S 4K1
a r t i c l e Keywords: Travel behavior Time-use Rural Urban Commuting Travel mode
i n f o
a b s t r a c t This paper considers variations in time-related aspects of travel behavior along the urban–rural continuum, using the four categories of inner city, suburbs, inner commuter belt (ICB), and outer commuter belt (OCB). It employs geo-coded and GPS-validated data from the STAR survey conducted in the county-sized regional municipality of Halifax, Nova Scotia. Many significant inter-zonal differences are identified, and most travel variables are characterized by progressive urban-to-rural gradients, with large differences between inner-city and outer-commuter values. A clear break between city and country is seldom evident, however. Inner-city residents make most trips, but have trips of shortest duration, and spend least time in travel. Residents of the commuter belts spend most time in travel, and have trips of longest duration. While long trips and much driving were expected in commuter zones, there are significantly fewer trips in the OCB, which we attribute to lack of need, lack of opportunities, and adjustments in discretionary behavior. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction: the rural–urban continuum How and where people spend their time are vital indicators of local culture, lifestyles, and quality of life. Out-of-home time use also has important policy implications, and in particular the purpose, timing, location, and mode of travel are crucial to transportation and land-use planners (Shaw, 2006). Key concepts and empirical findings on time–space activity produced over the last 40 years are overwhelmingly oriented to urban milieux, and there has been insufficient attention to rural areas, or to rural–urban contrasts (Timmermans et al., 2002). In this paper we employ innovative data from the Halifax STAR (Space–Time Activity Research) project to investigate travel behavior along the entire urban-torural spectrum, which allows a more realistic and theoretically satisfactory analysis than a simple urban/rural dichotomy. Time-use data provide important insights into patterns of behavior, but the STAR data are particularly rich, since the project couples time diaries with GPS tracking of travel, and thus provides a complete picture of space–time activity. Such activity can be examined at various levels of aggregation, and related to various types of controls/influences. There are obvious demographic controls, such as age, sex, and income, but our focus in this paper is on how space–time activity varies spatially around a metropolitan * Corresponding author. Tel.: +1 902 420 5739; fax: +1 902 496 8213. E-mail addresses:
[email protected] (H. Millward),
[email protected] (J. Spinney). 0966-6923/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jtrangeo.2009.12.005
centre, along the rural–urban continuum. There are conceptual and theoretical reasons to expect considerable differences in time use and daily travel as we move outwards from the congested urban core to the rural periphery, but a rigorous empirical treatment has not been possible previously owing to lack of data. There are also practical reasons to relate time use and travel behavior to the rural–urban continuum. Further understanding of relationships between residential settings, transportation choices, and travel behavior should help improve the quality and accuracy of transportation models. It will also help inform debate on public policy relating to quality of life, health, education, rural development, and equitable access to public services. In particular, it may prove highly pertinent to the provision of schools and public transportation in rural areas (Furuseth, 1998), and to the issue of ‘walkability’/active transportation (Saelens and Handy, 2008) in urban areas. We begin with a brief discussion on notions of rurality, and how position on the rural–urban continuum might affect time-use aspects of travel behavior. The context, purpose, and methods of the STAR project are then introduced, and its zonal geography is discussed. Time-use data from this study are aggregated and analyzed, with emphasis on trip counts and travel duration for major activity categories, and on modal choice. Important differences between geographic zones are identified, and their implications are discussed. Traditionally, rural and urban ways of life were quite distinct, with country folk engaged in resource-based primary production,
52
H. Millward, J. Spinney / Journal of Transport Geography 19 (2011) 51–58
and town dwellers employed in the manufacturing or service sectors. Both groups lived close to their workplaces. Widespread use of automobiles, however (say, after 1950 in Canada), led to ‘time–space convergence’ (Janelle, 1969) which extended urban commuting fields (a.k.a. ‘daily urban systems’ or labor market areas) well beyond the built-up area, and greatly altered socio-economic characteristics within this ‘urban field’ (Berry and Neils, 1969; Friedmann and Miller, 1965; Plane, 1981; Russwurm, 1976; Smailes, 1947; Stabler and Olfert, 1996). The limit of this commuter zone is typically suggested as around 1 h’s drive from major urban employment nodes, which underlines the importance of time use in the structure of modern rural areas. Large-lot housing development can significantly alter the land-use structure and social character of the more intensively exurbanized portions of the commuter belt (Clark et al., 2009; Dahms, 1998; Lamb, 1983; Millward, 2000, 2002; Sharp and Clark, 2008). Pryor (1968), Bourne and Simmons (1982), Robinson (1990, particularly ch. 2), and Bryant et al. (2000) all provide useful discussions of the urban impact on the countryside and on rural ways of life. They agree with Pahl (1966) that there exists a ‘rural–urban continuum’, such that a simple urban/rural dichotomy is seldom useful or appropriate. Unlike Hoggart (1990), they see utility in retaining the term rural, but defining differing degrees of rurality based on social, economic, demographic, and land use criteria (Cloke, 1977; Harrington and Donoghue, 1998). Most national census bureaux also see utility in delineating urban and rural districts, based on both the urbanized (built-up) area and the commuting (labor market) area (e.g., Bourne and Simmons, 1982; Slifkin et al., 2004). Urbanized area is usually defined as a contiguous area exceeding a population density threshold (400/km2 in Canada), while the commuter-shed is a contiguous area within which a large proportion of the employed labor force commute to the central urbanized core (over 50% in Canada). The range of activity settings within easy travel range is often very limited in rural areas (Borgen, 1998; Furuseth, 1998; Pucher and Renne, 2005). Residents in remote or lightly-settled areas can be expected to avoid travel whenever possible, but to undertake lengthy trips when local alternatives are unavailable. Where resource-based employment in farming and fishing is still important, many workers may have little or no commuting distance, since much of this activity takes place at or near home. The average person’s journey-to-work may thus take no more time in remote areas than in the city, though the average participant’s may be somewhat longer. For journeys to shop and socialize, the dispersed nature of rural opportunities should be reflected in much longer journey distances than in the city, but journey times may be only moderately greater, since travel speeds are higher. We might also expect the frequency of discretionary trips to be less in the country, so that total distances traveled are similar to urban levels. The rural journey-to-school, however, is non-discretionary, and should definitely be longer in distance and duration than the urban one. No travel or time-use studies to date have satisfactorily examined differences in travel behavior along the entire rural–urban continuum. Several useful studies have examined urban-suburban differences, however: Cervero and Gorham (1995) contrasted trip generation in transit and automobile neighbourhoods of the Los Angeles and San Francisco Bay areas, Goudie (2002) compared trip-distances and modal choice (though not travel times) for inner, middle (suburban), and outer (fringe) zones in two small Australian cities, while Chen and McKnight (2007) contrasted travel times, purposes, and modes for homemakers in very dense, dense, and suburban zones around New York City. Variations in rural/urban travel behavior have been investigated by Pucher and Renne (2005), using the United States national household travel survey, and employing a dichotomous urban/rural split based on population density (500/sq. mile). They
report on trip lengths, modal choice, and trip purpose, but do not test for significant differences. Using data from the Canadian 1986 national time-use survey, Harvey (1994) examined time-use variations in travel for three geographic categories (metropolitan, large town, and rural/small-town), but failed to test for significant differences. Millward and Spinney (2009) used the published binary rural/urban coding for the 2005 Canadian time-use survey, which groups metropolitan (CMA) and census agglomeration (CA) respondents as ‘urban’, and rural/small town as ‘rural’. They report significant rural/urban differences in nine out of 10 major time-use categories, and also in key travel variables. They find that rural residents participate less in travel, and on average therefore spend less time in travel. Both in distance and time, rural journeys to work show more variation than do urban ones, though their average time duration is similar, for participants. The average rural respondent spends less time in journeys to shop, owing to lower participation. Though they shed much light on urban/rural differences in Canadian time use and travel, the studies by Harvey (1994) and Millward and Spinney (2009) crudely categorize urban labor markets (commuter zones) as urban. Yet these supposedly urban areas contain far more farmland and forest than urban development, and much of it would appear rural to the lay person. Certainly, too, most residents of exurban commuter belts would consider themselves rural, even when they work in town (Walker, 2003; Paquette and Domon, 2003).
2. Survey and methods This study employs data from the Halifax STAR project, which pertains to the county-sized regional municipality of Halifax, Nova Scotia, to divide the rural–urban continuum into five geographical zones, from inner city to remote rural. STAR (Space– Time Activity Research) is an innovative survey of both time use and travel activity, which employs precise GPS tracking to geo-reference respondent locations throughout a 48-h period (TURP, 2007). ‘‘The Halifax STAR project enjoys the designation of the largest GPS sample within the context of a household travel survey project undertaken to date” (Bricka, 2008, p. 3). Not only does this provide ground-breaking locational information, but the GPS tracking also greatly improves the quality of the timeuse data, as evidenced by the average number of daily travel trips reported (Spinney, 2008). The Halifax region was also the venue for an earlier geo-coded time survey directed by Andrew Harvey (Elliot et al., 1976), whose space–time data have been thoroughly analyzed (Goodchild and Janelle, 1984; Janelle and Goodchild, 1983). That study, however, focused only on the inner city and directly-adjacent suburbs. Conceptually, the zoning scheme was based on the rural–urban fringe concept (Bryant et al., 2000; Coppack et al., 1988; Furuseth and Lapping, 1999; Millward, in press, Fig. 1; Pryor, 1968; Wehrwein, 1942), with zones delimited operationally on the basis of both urban form (i.e. residential density and percentage of area developed) and commuting linkages to the urbanized area. The five zones are defined as follows: Inner city: the older (pre-1960) developed areas of Halifax and Dartmouth, within walking range (c. 5 km) of the downtowns. Suburbs: other contiguous built-up (‘urbanized’) areas within the urban service boundary (the area serviced by central water and sewerage systems). Inner commuter belt (ICB): all other areas within 25 km road distance of downtown Halifax. This transitional zone is highly impacted by large-lot subdivision development (exurbs), and within it most employees commute to the urbanized area.
H. Millward, J. Spinney / Journal of Transport Geography 19 (2011) 51–58
53
Fig. 1. Rural–urban settlement zones in Halifax regional municipality. (EPS file is provided separately: image shown here is for information only.)
Outer commuter belt (OCB): areas between 25 and 50 km road distance from downtown Halifax or another large town (over 10,000 population). These areas are less fully impacted by commuting, and remain somewhat reliant on local resource industries (particularly fishing). Remote rural: areas beyond 50 km by road from Halifax or another large town. Over half of HRM lies in this zone, but there are fewer than 5000 people. Most employment is in local resource industries (fishing, farming, and forestry). Fig. 1 maps the five zones for Halifax Regional Municipality (HRM). Note that the suburbs typically extend to a radius of 10 km from downtown, but there are extensions to Lower Sackville and Lakeside. The ICB and OCB are defined by 1.0 km buffers from the road network, so that certain unroaded/uninhabited areas close to the city are placed in the default category of remote rural. It is noteworthy that low density exurban sprawl is particularly evident in the Halifax region, owing to its glacially-scoured hardrock geology, which precludes farming and lowers land price, yet endows the area with attractive woodlands and lakes. Until very recently, there has been very little regulation to limit the extent or pattern of exurbanization (Millward, 2002).
3. Time use by rural–urban zones The STAR project sampled 1971 households, or about one household in 78 within HRM, during the period April, 2007–May,
2008. Primary respondents over the age of 15 completed 48-h time diaries and detailed questionnaire surveys. The sample design stratified for season, day of week, age, sex, and geographic zones, but it was not possible to obtain representative samples for all groups – younger adults in particular were under-sampled, and the ‘remote rural’ sample was too small for useful analysis. As Table 1 shows, the inner-city sample of diary-days represented an older population, with a large proportion of retirees. This zone’s sample was more highly educated than average, but also had fairly low incomes. Demographically the suburbs were most typical, while the commuter belts had the youngest and wealthiest samples. The ICB had the largest proportion of workers, and the smallest proportion of retirees. We found significant differences in mean daily time allocation between the other four zones, for all respondents and for participants (‘doers’) only. Differences in scores for adjacent zones were tested for statistical significance using the Mann–Whitney test. This non-parametric difference-of-ranks test is preferable to the t-test (ANOVA), since most variables are highly positively skewed.
3.1. Work activities Average time in Paid Work is very low in the inner city and suburbs, and very high in the inner commuter belt, where the participation rate is highest. For participants, suburban and OCB times considerably exceed inner city and ICB times, but for adjacent zones only the ICB/suburban difference is significant. Domestic
54
H. Millward, J. Spinney / Journal of Transport Geography 19 (2011) 51–58
Table 1 Percent of respondent-days in each settlement zone by demographic groups. All respondents (n)
Overall 100.0% (3919)
Inner city 20.2% (790)
Suburbs 53.8% (2109)
Inner commuter belt 17.5% (685)
Outer commuter belt 8.5% (335)
Age group 15–34 35–54 55+
8.9 49.7 41.4
6.6 42.1 51.3
9.7 45.6 44.7
9.1 62.8 28
8.4 66.6 25.1
Sex Female
54
57.3
51.7
53.4
62.1
Main activity Working Going to school Housework/childcare Retired
62 3 7.9 27.1
56 3.3 6.8 33.9
59.1 3.3 8.3 29.3
74.4 2.2 7.9 15.5
69.9 1.3 7.4 21.4
Primary driver Yes
78.9
69
78.9
87.2
86.3
Education University degree
40.2
54.3
36.7
37.5
34.2
Household income Under $40,000 $40,000–$59,999 $60,000–$79,999 $80,000 or more
14.9 17.5 18.6 49
21 16.1 17.2 45.8
15.3 20 17.7 47
10.1 13.7 20.4 55.8
8.9 13.1 23.4 54.6
Work occupies similar amounts of time, for all and participants, in the three inner zones. Time in household Care-giving (mostly in child care) increases outwards from the city, reflecting the ‘family status’ component identified in factorial ecology studies of urban social patterns (Davies and Murdie, 1993).
3.2. Other non-leisure activities Time spent in Shopping is very similar in all zones, both for all and participants. Inner-city residents spend significantly more time than residents of other zones in Personal Care (mostly for sleep); they are less encumbered by child-rearing responsibilities and also have low workforce participation. Participation rates for Education are low in all zones, since respondents were required to be 15 or older. Times for participants do not vary significantly, but the lower inner-city value may reflect a greater propensity for part-time attendance, owing to a lower travel-time penalty (see Table 3). Organizational activities refer to attendance and unpaid work in voluntary organizations such as churches, service clubs, and volunteer fire departments. Participation rates decline somewhat along the urban-to-rural gradient, presumably reflecting fewer local opportunities. Time per participant, however, is highest in the OCB: the church and fire-hall are important social centres in this zone, particularly for those who live nearby.
3.3. Leisure activities Discretionary leisure activities comprise Sports and Hobbies, Entertainment Events (out-of-home), and Media/Communications (TV, reading, computer games, etc.). Time spent on media is easily the largest segment, and this activity declines significantly from suburbs to commuter belt, for all and participants. Participation in entertainment events and sports/hobbies also declines outwards from the centre, but intriguingly average time at events for participants increases, and is very high in the outer commuter zone. This probably reflects high levels of child chauffering and attendance at little-league sports events (Tillberg Mattsson, 2002).
4. Travel by time-use activity 4.1. Summary data Globally, there were 6.5 activity-related trips (one-way O-D) by all modes, for a total of 84.5 min per day (averaged over the 7-day week). However, travel was also undertaken for non-transportation purposes: that is, driving, walking, or jogging purely for pleasure. The average respondent undertook 0.3 such trips, totaling 12.1 min per day. Across all geographic zones, therefore, the average respondent had 6.8 trips, totaling 96.6 min (Table 2). It should be noted that 6.8 trips is considerably higher than the number
Table 2 Summary travel data, by settlement zones, for all respondents and participants. Travel variable (per person/day)
Settlement zones (n = No. of diary days) All zones
Inner city n = 790
Suburbs n = 2109
Inner commuter belt n = 685
Outer commuter belt n = 335
All
Participants
All
Participants
All
Participants
All
Participants
All
Participation rate (%) Number of trips (O-D)
94.2 6.8
100 7.2
92.4 7.2
100 7.7
Total duration (mins)
96.6
102.5
92
Average trip duration (mins/person-day)
14.7
15.6
13.3
93.5
100 7.1 99
90.2 6.8
100 7.3
97
6.8 94
107
14
14.4
15.2
16.3
Underlined figures: ranks are signif. different from those of next zone inwards at p = 0.05 (Mann–Whitney, 2-tailed).
Participants
88.4
100
115
6.1 104
6.7 114
17.5
17.1
18.8
55
H. Millward, J. Spinney / Journal of Transport Geography 19 (2011) 51–58
reported in most traffic/transport surveys, even those employing GPS (Murakami and Wagner, 1999; Stopher et al., 2007). We attribute the higher value to GPS-assisted prompted recall interviewing technique, which is an innovative feature of the STAR survey. As expected, time expenditure on travel tends to increase as one progresses outwards from the inner city, both for all respondents and for participants only. Participation in travel, however, and number of trips, both decline, presumably as a response to fewer opportunities and increased distances (participation rates can be calculated as the mean for all divided by the mean for participants, multiplied by 100). Respondents partly compensate for high average trip durations in the commuter zones by reducing their number of trips: trip numbers decline significantly outwards, both for all respondents and participants only. Interestingly, trip totals in the ICB are remarkably similar to those in the suburbs, and the trip count falls significantly only in the OCB, where activity locations become highly scattered. Average trip durations vary significantly between all adjacent zones, increasing progressively outwards. Because of reduced trip numbers, however, there is less variation in total duration: there is a significant increase only between the suburbs and the inner commuter belt, whereas the ICB and OCB have very similar values. Travel participants in these commuter belts spend almost 2 h per day in travel. 4.2. Trip counts by activity From Table 3, we see that the largest group of trips is for shopping, when averaged for all respondents. In all zones, the average is
2.0–2.5 such trips per day. There is a high participation rate for shopping trips in all zones, which declines from 69% in the inner city to 59% in the OCB. Largely because of this decline, average trip-count per participant is lower in the commuter zones. Trips to/from paid work are frequent for participants, averaging between 3.0 and 3.2 per day (many people go home for lunch, and some have multiple jobs). Participation rates rise from 36% in the inner city to 45% in the ICB, and then fall back to 36% in the OCB. These rates partly reflect the youthful demography of the ICB, and its lower proportions of elderly and unemployed, but it may also indicate lack of accessible employment opportunities in the OCB. Also related to demography, trips for household-member care (mostly child care) have lowest participation in the inner city (19%), and highest in the ICB (27%). In all zones there is a high trip-count for participants, ranging from 2.7 in the OCB to 3.3 in the inner city. Participation in education is low in all zones, falling progressively from 3.2% in the inner city to 2.4% in the OCB. Trip counts for participants do not vary significantly between zones, presumably because of the small sample sizes. Trip counts for organizational, entertainment, and sports activities do not exhibit strong urban-to-rural gradients. Participation rates for all three categories are lower in the commuter zones, however, as one would expect for discretionary activities where opportunities are fewer and further apart. Inner-city respondents participate most in trips to entertainment events and sports, in part reflecting easier access to many downtown attractions. Participation in pleasure trips is high in the inner city and the OCB (27% and 25%, respectively), and lower in the suburbs
Table 3 Trip counts by activity group, by settlement zones, for all respondents and participants (means/person-day). Trip purpose
Settlement zones (n = No. of diary days) Inner city n = 790
During paid work To-from paid work Domestic Child–adult care Goods or services (shopping) Personal care Education Organization, volunteer, religious Entertainment events Sports, hobbies Media or communication Pleasure (driving, walking, hiking)
Suburbs n = 2109
Inner commuter belt n = 685
Outer commuter belt n = 335
All
Participants
All
Participants
All
Participants
All
Participants
0.5 1.1 0.1 0.6 2.5 0.1 0.1 0.6 0.7 0.4 0.0 0.4
4.1 3.2 1.8 3.3 3.6 2.2 3.1 2.7 2.5 2.3 1.6 1.6
0.2 1.2 0.1 0.7 2.5 0.0 0.1 0.6 0.6 0.4 0.0 0.3
3.1 3.2 1.7 3.0 3.7 1.8 3.2 2.6 2.5 2.3 1.6 1.4
0.5 1.3 0.1 0.9 2.2 0.0 0.1 0.5 0.5 0.3 0.0 0.3
4.2 3.0 1.8 3.2 3.5 1.7 3.0 2.6 2.3 2.5 1.7 1.4
0.4 1.2 0.1 0.6 2.0 0.0 0.1 0.5 0.6 0.3 0.1 0.3
3.3 3.2 1.7 2.7 3.3 1.4 3.6 2.8 2.5 2.0 1.9 1.3
Underlined figures: ranks are signif. different from those of next zone inwards at p = 0.05 (Mann–Whitney, 2-tailed). Table 4 Travel durations by activity group, by settlement zones, for all respondents and participants (mean minutes/person-day). Trip purpose
During paid work To-from paid work Domestic Child–adult care Goods or services (shopping) Personal care Education Organization, volunteer, religious Entertainment events Sports, hobbies Media or communication Pleasure (driving, walking, hiking)
Settlement zones (n = No. of diary days) Inner city n = 790
Suburbs n = 2109
Inner commuter belt n = 685
Outer commuter belt n = 335
All
Participants
All
Participants
All
Participants
All
Participants
9.1 14.4 1.6 6.2 22.6 2.3 1.2 6.0 8.2 5.4 0.3 14.5
75.0 40.6 23.8 32.2 32.8 57.4 36.7 27.4 30.9 30.7 11.0 54.9
5.1 19.2 1.7 7.8 23.9 1.2 1.5 7.6 8.2 5.2 0.3 12.0
70.0 50.3 21.6 35.3 36.3 46.9 47.9 32.4 34.7 32.7 10.9 53.5
11.4 24.4 1.6 11.8 24.6 0.8 1.2 7.1 9.1 4.8 0.3 10.0
90.5 54.3 26.7 43.5 39.2 27.4 48.2 38.3 40.0 43.0 10.4 52.1
7.0 25.3 2.1 8.6 24.8 0.2 2.1 7.0 10.1 5.1 0.4 11.1
60.4 70.0 28.0 37.8 42.4 15.0 86.5 38.2 44.6 38.6 14.5 45.3
Underlined figures: ranks are signif. different from those of next zone inwards at p = 0.05 (Mann–Whitney, 2-tailed).
56
H. Millward, J. Spinney / Journal of Transport Geography 19 (2011) 51–58
(22%) and ICB (19%). We speculate that these variations partly reflect the greater ‘time-pressure’ of households in the latter two zones, and partly reflects more walking opportunities in the inner city. 4.3. Travel durations by activity Table 4 shows average travel durations by activity group. Mean durations for travel during paid work are very long for participants in all zones, but particularly in the inner commuter belt, where results are highly positively skewed by long-distance air travel (many airline crew and frequent business flyers live in the Fall River area). For participants, travel to/from paid work has the next longest duration. Means increase significantly outwards, from 41 min in the inner city to 70 min in the OCB. Average one-way (O-D) trip durations (total duration divided by trip count, from Tables 2 and 3) increase smoothly through the four zones, from 12.7 min (inner city) to 15.7 (suburbs), 18.1 (ICB), and 21.9 (OCB). This gradient, while unsurprising, nevertheless confirms the importance of the rural–urban continuum to travel behavior. In a similar manner, time spent in travel to education also rises from inner to outer zones, for participants. The zonal differences are even more pronounced than for journey-to-work, but are not statistically significant owing to small sample sizes. Average time in shopping trips, for all respondents, is very similar for all zones, at around 24 min (Table 4). Participation, however, declines with distance from the centre, while means for participants rise progressively, from 33 min (inner city) to 42 min (OCB). Average one-way (O-D) trip durations (duration/trip count) also increase smoothly through the four zones, from 9.1 min (inner city) to 12.8 (OCB). Thus, residents of the commuter belts make fewer shopping trips, but those trips are considerably longer. The inner commuter belt shows the highest travel duration for childcare, presumably reflecting both its youthful demography (high participation rate) and the low density of childcare opportunities. Conversely, childcare travel durations are lowest in the inner city, owing to both lower participation and more conveniently located opportunities. Travel for organization- and leisure-related activities is typically of longer duration in the outer zones, whereas travel for pleasure is significantly longer in the inner city. 5. Travel modes 5.1. Trip counts by mode Exurban commuter zones are viewed as being entirely enabled by the automobile, in the popular mind and in the academic literature. The suburbs, too, are often seen as highly automobiledependent. The STAR data not only support these characterizations, but also underline the importance of walking as a travel mode, particularly in the inner zones. Many travel surveys fail to
record travel by walking or bicycling, or do so only through unreliable recall surveys (e.g. Transportation Tomorrow surveys of 2001 and 2006—see Data Management Group, 2003). Participation rates for travel by car as driver were lowest in the inner city (67.5%) and peaked in the ICB (82.9%), whereas participation in bus travel fell consistently outwards, from 8.1% in the inner city to 3.0% in the OCB. Low transit use reflects low response rates from those age and income cohorts most reliant on buses. Significant differences in mean trip counts for all respondents (Table 5) largely relate to participation differences, but we should note that there are also some significant differences for participants. Note in particular that car-drivers in the OCB take fewer trips than those in the ICB, presumably minimizing their discretionary trips. Also, the passenger trip count rises significantly outwards, peaking in the OCB. This is understandable, given the lack of walking or publictransit options in outer zones. Walking participation was very high in the inner city (59.9%), reflecting the high density of opportunities, but 36–42% elsewhere. Bicycling participation was 3.7% in the inner city, and below 1.3% elsewhere. Significant outward declines in trip counts for these ‘active transportation’ (AT) modes partly reflect lower participation, but note that even for participants the declines are significant. 5.2. Travel duration by mode Table 6 shows mean time spent in the main travel modes, in minutes per 24 h. As expected, urbanites (both inner-city and suburban) spend more time walking, bicycling, and using transit, and a smaller proportion of travel time in cars. Averaged over the 7-day week, inner-city respondents spend only 56 min per day in a car (45 as driver, 11 as passenger), whereas suburbanites spend 72 min, and inner commuters 91 min. This urban-to-rural gradient reverses slightly to 85 min in the OCB, however. Participants, of course, spend more time either at the wheel or in the passenger seat: drivers resident in the ICB average 93 min per day. Averaged for all respondents, transit (bus) use is low in all areas. Times for bus participants tended to rise outwards, though the OCB’s very high value is an artifact of data outliers. In other zones, times for bus riders were lower than those for car drivers, which is unexpected. Bicycle participation rates are low in the inner city and miniscule elsewhere, so that overall respondents average only 1.1 min/day. Durations for participants, however, are quite lengthy: they peak in the suburbs, but decline somewhat in the commuter zones. Average walking time in the inner city is 25.6 min, owing to high participation, but walking durations are much lower elsewhere. Even for participants, time spent walking declines progressively and significantly outwards. 6. Summary and conclusions Our premise in this study was that position along the rural–urban continuum (degree of urban-ness versus degree of rurality) re-
Table 5 Trip counts by travel mode, by settlement zones, for all respondents and participants (mean trips/person-day). Travel mode
Car as driver Car as passenger Bus Bicycle Walking
Settlement zones (n = No. of diary days) Inner city n = 790
Suburbs n = 2109
Inner commuter belt n = 685
All
Participants
All
Participants
All
Participants
All
Participants
4.0 0.8 0.2 0.2 2.0
5.9 3.0 1.9 4.1 3.4
4.6 0.9 0.1 0.0 1.2
5.8 3.4 2.1 2.1 2.8
5.1 0.8 0.1 0.0 0.8
6.2 3.5 1.8 1.6 2.2
4.2 0.8 0.1 0.0 0.9
5.4 3.9 2.9 1.0 2.1
Underlined figures: ranks are signif. different from those of next zone inwards at p = 0.05 (Mann–Whitney, 2-tailed).
Outer commuter belt n = 335
57
H. Millward, J. Spinney / Journal of Transport Geography 19 (2011) 51–58 Table 6 Duration in travel modes, by settlement zones, for all respondents and participants (mean minutes/person-day). Travel mode (mean mins/24 h)
Car as driver Car as passenger Bus Bicycle Walking
Settlement zones (n = No. of diary days) Inner city n = 790
Suburbs n = 2109
Inner commuter belt n = 685
Outer commuter belt n = 335
All
Participants
All
Participants
All
Participants
All
Participants
45.3 11.1 3.0 2.2 25.6
67.1 42.7 37.1 60.9 42.7
58.9 12.9 3.6 1.0 15.5
75.3 51.2 62.5 82.2 37.1
77.1 13.8 1.7 0.6 13.1
93.0 64.2 50.6 58.7 36.2
70.3 14.9 3.5 0.6 12.6
90.9 69.4 116.4 67.3 30.0
Underlined figures: ranks are signif. different from those of next zone inwards at p = 0.05 (Mann–Whitney, 2-tailed).
Table 7 Major findings, averaged for all respondent days. Major travel variables (per person/day)
Settlement zones
(H = high, M = medium, L = low)
Inner city
Suburbs
Inner commuter belt
Outer commuter belt
Total number of trips Total trip duration Average trip duration Number trips to/from work Number trips to/from shopping Mean duration, trips to/from work Mean duration, trips to/from shopping Number trips in car as driver Mean time in car as driver
H L L L H L L L L
M L M M H M M M M
M H H H M H H H H
L H H M L H H L H
mains important in how people order and shape their lives, and in particular that it affects travel behavior. No time-use or travel surveys to date have satisfactorily addressed this issue, but GPS-validated data from the Halifax STAR survey are ideally suited for the purpose. Four settlement zones were defined and delimited, based on density and accessibility criteria, and labeled as inner city, suburbs, inner commuter belt (ICB), and outer commuter belt (OCB). Inter-zonal variations were tested for significance using the Mann–Whitney difference-of-ranks test. Table 7 summarizes some of the key findings, for all respondent-days (a table for participants would look very similar). Each zone has been categorized as scoring low, medium, or high on each variable, relative to the overall survey value, and significant interzonal differences are indicated. Such differences are most evident between the suburbs and ICB, but this is largely attributable to their higher sample sizes. Even where Mann–Whitney tests do not show significance at a given inter-zonal boundary, note that most variables are characterized by progressive urban-to-rural gradients, with large differences between inner-city and outercommuter values. A clear break between city and country is seldom evident. The results demonstrate that a simple rural/urban dichotomy is insufficient to fully appreciate geographic variation in time use and travel. A more graduated range of settlement categories, based on accessibility and density criteria, provides more nuanced results, and greater insight. Our study is pertinent to several public policy issues of concern throughout North America. It relates particularly to policy debate on the relationship between neighbourhood density, transport mode choice, and healthy lifestyles: ‘active transportation’ (walking and bicycling) is shown to be largely confined to the inner city, with bus transport relatively most important in the suburbs, and the inner commuter belt most reliant on the automobile. Inner-city residents make most trips, but have trips of shortest duration, and spend least time in travel, while residents of the commuter belts spend most time in travel, and have trips of longest duration. Though not surprising, these findings confirm the lack of transportation choice in the commuter belt, and also demonstrate that res-
idents of the inner and outer belts operate under different constraints, and cope differently with lack of choice. Those in the ICB participate highly in the workforce and in commuting to the city, and thus drive longest and most frequently. Those in the OCB drive somewhat less, for reasons related to greater isolation and a more genuinely rural economy: they have lower workforce participation and a lower incidence of commuting to the city, and they tend to minimize non-essential (non journey-to-work) trips. Acknowledgement This research was supported through the Atlantic Innovation Fund from the Atlantic Canada Opportunities Agency, Project No. 181930, awarded to the Time Use Research Program of Saint Mary’s University (Principal Investigator: Andrew Harvey; CoInvestigator: Hugh Millward). References Berry, B., Neils, E., 1969. Location, size and shape of cities as influenced by environmental factors: the urban environment writ large. In: Perloff, H. (Ed.), The Quality of Urban Environment. Johns Hopkins University Press, Baltimore, pp. 257–302. Borgen, S., 1998. You can get there from here. Community Transportation 16, 10–33. Bourne, L., Simmons, W., 1982. Defining the area of interest: definitions of the city, metropolitan areas, and extended urban regions. In: Bourne, L. (Ed.), Internal Structure of the City, second ed. Oxford University Press, New York, NY, pp. 57–72. Bricka, S., 2008. Non-Response Challenges in GPS-based Surveys, Resource Paper prepared for the May 2008 International Steering Committee on Travel Survey Conferences Workshop on Non-response Challenges in GPS-based Surveys.
. Bryant, C., Coppack, P., Mitchell, C., 2000. The city’s countryside. In: Bunting, T., Filion, P. (Eds.), Canadian Cities in Transition, second ed. Oxford University Press, Don Mills, Ont, pp. 333–354. Cervero, R., Gorham, R., 1995. Commuting in transit versus automobile neighborhoods. Journal of the American Planning Association 61, 210–225. Chen, C., McKnight, C., 2007. Does the built environment make a difference? Additional evidence from the daily activity and travel behavior of homemakers living in New York City and suburbs. Journal of Transport Geography 15, 380– 395.
58
H. Millward, J. Spinney / Journal of Transport Geography 19 (2011) 51–58
Clark, J., McChesney, R., Munroe, D., Irwin, E., 2009. Spatial characteristics of exurban settlement pattern in the United States. Landscape and Urban Planning 90, 178–188. Cloke, P., 1977. An index of rurality for England and Wales. Regional Studies 11, 31– 46. Coppack, P., Russwurm, L., Bryant, C. (Eds.), 1988. Essays in Canadian Urban Process and Form 3: The Urban Field. University of Waterloo, Department of Geography, Publication Series No. 30, Waterloo. Dahms, F., 1998. Settlement evolution in the arena society in the urban field. Journal of Rural Studies 14, 299–320. Data Management Group, University of Toronto Joint Program in Transportation, 2003. Transportation tomorrow survey 2001: design and conduct of the survey (accessed 17.07.09). Davies, W., Murdie, R., 1993. Measuring the social ecology of cities. In: Bourne, L., Ley, D. (Eds.), The Changing Social Geography of Canadian Cities. McGillQueen’s University Press, Montreal, pp. 52–75. Elliot, D., Harvey, A., Procos, D., 1976. An overview of the Halifax time budget study. Society and Leisure 3, 145–159. Friedmann, J., Miller, J., 1965. The urban field. Journal of the American Institute of Planners 31, 312–319. Furuseth, O., 1998. Service provision and social deprivation. In: Ilbery, B. (Ed.), The Geography of Rural Change. Longman, Harlow, pp. 233–256. Furuseth, O., Lapping, M., 1999. Contested Countryside: The Rural–Urban Fringe in North America. Ashgate, Aldershot. Goodchild, M., Janelle, D., 1984. The city around the clock: space–time patterns of urban ecological structure. Environment and Planning A 10, 1273–1285. Goudie, D., 2002. Zonal method for urban travel surveys: sustainability and sample distance from the CBD. Journal of Transport Geography 10, 287–301. Harrington, V., Donoghue, D., 1998. Rurality in England and Wales 1991: a replication and extension of the 1981 rurality index. Sociologia Ruralis 38, 178–203. Harvey, A., 1994. Changing temporal perspectives and the Canadian metropolis. In: Frisken, F. (Ed.), The Changing Canadian Metropolis: A Public Policy Perspective, vol. 1. Institute of Governmental Studies Press, Berkeley, Cal, pp. 151–199. Hoggart, K., 1990. Let’s do away with rural. Journal of Rural Studies 6, 245–257. Janelle, D., 1969. Spatial reorganization: a model and concept. Annals of the Association of the American Geographers 59, 348–364. Janelle, D., Goodchild, M., 1983. Diurnal patterns of social group distributions in a Canadian city. Economic Geography 59, 403–425. Lamb, R., 1983. The extent and form of exurban sprawl. Growth & Change 14, 40– 48. Millward, H., 2000. The spread of commuter development in the Eastern Shore zone of Halifax, Nova Scotia, 1920–1988. Urban History Review 29, 21–32. Millward, H., 2002. Peri-urban residential development in the Halifax region 1960– 2000: magnets, constraints, and planning policies. The Canadian Geographer 46, 33–47. Millward, H., in press. ‘Exurban’ housing development in the Halifax commuter belt: processes, patterns, and policies. In: Beesley, K. (Ed.), The Rural–Urban
Fringe in Canada: Conflict and Controversy. Brandon University, Rural Development Institute, Brandon. Millward, H., Spinney, J., 2009. Time use and rurality: Canada 2005. Electronic International Journal of Time Use Research 6, 109–129. Murakami, E., Wagner, D., 1999. Can using global positioning system GPS improve trip reporting? Transportation Research Part C: Emerging Technologies 7 (2–3), 149–165. Pahl, R., 1966. The rural–urban continuum. Sociologia Ruralis 6, 299–327. Paquette, S., Domon, G., 2003. Changing ruralities, changing landscapes: exploring social recomposition using a multi-scale approach. Journal of Rural Studies 19, 425–444. Plane, D., 1981. The geography of urban commuting fields. Professional Geographer 33, 182–188. Pryor, R., 1968. Defining the rural–urban fringe. Social Forces 47, 202–215. Pucher, J., Renne, J., 2005. Rural mobility and mode choice: evidence from the 2001 National Household Travel Survey. Transportation 32, 165–186. Robinson, G., 1990. Conflict and Change in the Countryside. Belhaven, London. Russwurm, L., 1976. Country residential development and the regional city form in Canada. Ontario Geography 10, 79–96. Saelens, B., Handy, S., 2008. Built environment correlates of walking: a review. Medicine and Science in Sports and Exercise 40, S550–S566. Sharp, J., Clark, J., 2008. Between the country and the concrete: rediscovering the rural–urban fringe. City Community 7, 61–79. Shaw, S.-L., 2006. What about ‘‘time” in transportation geography? Journal of Transport Geography 14, 237–240. Slifkin, R., Randolph, R., Ricketts, T., 2004. The changing metropolitan designation process and rural America. Journal of Rural Health 20, 1–6. Smailes, A., 1947. The analysis and delimitation of urban fields. Geography 32, 151– 161. Spinney, J., 2008. Improving the number, timing, and location of trips: a GPSassisted prompted recall approach. Paper presented at the Canadian Association of Geographers, 57th Annual Meeting, Quebec City, Quebec, May 20–24. Stabler, J., Olfert, M., 1996. Spatial labor markets and the rural labor force. Growth and Change 27, 206–230. Stopher, P., Xu, M., FitzGerald, C., 2007. Assessing the accuracy of the Sydney Household Travel Survey with GPS. Transportation 34, 723–741. Tillberg Mattsson, K., 2002. Children’s (in)dependent mobility and parents’ chauffeuring in the town and the countryside. Tijdschrift voor Economische en Sociale Geografie 93, 443–453. Timmermans, H., Arentze, T., Joh, C.-H., 2002. Analysing space-time behaviour: new approaches to old problems. Progress in Human Geography 26, 175–190. TURP, 2007. Halifax Regional Space Time Activity Research (STAR) Survey: Project Description. Saint Mary’s University, Time Use Research Program, Halifax. Walker, G., 2003. Stabilizing the new residential frontier in the countryside. In: Beesley, K. et al. (Eds.), The New Countryside: Geographic Perspectives on Rural Change. Brandon University, Rural Development Institute, Brandon, pp. 414– 426. Wehrwein, G., 1942. The rural–urban fringe. Economic Geography 18, 217–228.