Impact of the residential location on the relation between travel time and activities duration
Résumé
Many studies have analysed travel times and activity times. For example it seems natural to observe a positive correlation between travel time and the time of activity at a given destination. The search for simplified representations of travel times began with the stability hypothesis in the 1970's, known as the “Travel-Time Budgets paradigm” (Goodwin, 1981), the “Brever law” (Hupkes, 1982), or the “Zahavi's hypothesis” (Zahavi, 1979). Less restrictive hypotheses have been proposed on travel times. One of them assumes that a travel time intensity can characterise activity times, therefore making travel times fully determined by activity times. It has been examined at the daily level on time expenditures. Kitamura & al. (1992) test the “principle of proportional assignment” (Beckmann & Golob, 1972). Under this principle, each activity time represents a fixed proportion of the total available time. “The ratio of the amounts of time assigned to two activities is invariant regardless of the total amount of time available” (Kitamura & al., 1992, p.135). Their results support that total daily travel time is proportional to the amount of daily available time.
More generally, concerning specifically activity, relationships between travel time and activity duration have been analysed and estimated in different estimation frameworks, such as linear model, structural equations model, duration model, etc (Van Wissen & Golob, 1991; Hamed & Mannering, 1993; Golob & McNally, 1997; Goulias & al. 1998; Kitamura, Chen & Narayanan 1998, Ma & Goulias 1998; Levinson, 1999; Lu & Pas, 1999; Joly, 2006). A concept equivalent to its proportionality has been addressed by some regarding travel time intensity. It hypothesises that travel time assigned to activity participation is related to activity duration, with the proportional form representing a particular case of the possible relationships.
In this context, Dijst & Vidakovic (2000) and Schwanen & Dijst (2002) propose and analyse travel-time ratio for work activities. Estimating a structural equations model for the activities duration Van Wissen & Golob (1991), Golob & McNally (1997), Lu & Pas (1999) analysed relationships between travel time and activity duration. Some of the estimators can be interpreted in terms of travel time intensities for different activity types. For example, Golob & McNally (1997, p.185) obtained a travel time intensity for maintenance that indicates that one hour of out-of-home maintenance activity requires on the average 7.8 minutes of travel time.
Studies on the urban mobility have shown that spatial characteristics have a significant effect on mobility and travel time (Timmermans et a. 2002),. We then propose to explore time use deduced from numerous households mobility surveys, in search for distinct relationships between activities duration.
This paper presents results of a comparative study of Swiss and French daily travel times based on mobility survey data of seven cities (Bern, Geneva, Grenoble, Lyon, Rennes, Strasbourg and Zurich) observed at two different dates. The cities chosen were guided by the disposability of the data and the following criterions:
∂ The form and spread of the peri-urbanisation process
∂ The level of equipment in heavy network of public transport
∂ The transport and urban policy of accessibility management
Swiss cities are clearly more transit and pedestrian oriented in terms of infrastructure networks and transport policy. In French cities, transport policies, such as restriction of automobile access to the centre and limitation of parking facilities, appeared relatively recently in comparison to Swiss.
Individuals and household socio-demographic characteristics and mobility on a given weekday are collected. Using, starting and stopping times and the types of activities at the origin and destination, the one-day out-of-home activities diary can be deduced from the first to the last trip of the day.
We focus on the relationships between travel and activity times (work, shopping, leisure). Following Kitamura & al. (1992), the proportional assignment paradigm, where daily available time is proportionally allocated to different activities with fixed proportions (which may vary across individuals), is tested. It is then adapted to test the proportionality of daily travel times for each activity type with the different activities daily times. It clearly indicates that travel times are influenced by duration of activities.
Furthermore, we pay special attention to the impact of the residential location, captured by a segmentation of the urban area in three zones: centre, suburban and periurban.
The final section presents estimations realised in the duration model framework, for the cities of Lyon and Grenoble, based on their more recent surveys. Models for the travel time budgets are built and lead to non-linear estimates of the travel time intensities. Influence of the residential location is captured dummies for the urban zone types and for the proximity with a highway access.
More generally, concerning specifically activity, relationships between travel time and activity duration have been analysed and estimated in different estimation frameworks, such as linear model, structural equations model, duration model, etc (Van Wissen & Golob, 1991; Hamed & Mannering, 1993; Golob & McNally, 1997; Goulias & al. 1998; Kitamura, Chen & Narayanan 1998, Ma & Goulias 1998; Levinson, 1999; Lu & Pas, 1999; Joly, 2006). A concept equivalent to its proportionality has been addressed by some regarding travel time intensity. It hypothesises that travel time assigned to activity participation is related to activity duration, with the proportional form representing a particular case of the possible relationships.
In this context, Dijst & Vidakovic (2000) and Schwanen & Dijst (2002) propose and analyse travel-time ratio for work activities. Estimating a structural equations model for the activities duration Van Wissen & Golob (1991), Golob & McNally (1997), Lu & Pas (1999) analysed relationships between travel time and activity duration. Some of the estimators can be interpreted in terms of travel time intensities for different activity types. For example, Golob & McNally (1997, p.185) obtained a travel time intensity for maintenance that indicates that one hour of out-of-home maintenance activity requires on the average 7.8 minutes of travel time.
Studies on the urban mobility have shown that spatial characteristics have a significant effect on mobility and travel time (Timmermans et a. 2002),. We then propose to explore time use deduced from numerous households mobility surveys, in search for distinct relationships between activities duration.
This paper presents results of a comparative study of Swiss and French daily travel times based on mobility survey data of seven cities (Bern, Geneva, Grenoble, Lyon, Rennes, Strasbourg and Zurich) observed at two different dates. The cities chosen were guided by the disposability of the data and the following criterions:
∂ The form and spread of the peri-urbanisation process
∂ The level of equipment in heavy network of public transport
∂ The transport and urban policy of accessibility management
Swiss cities are clearly more transit and pedestrian oriented in terms of infrastructure networks and transport policy. In French cities, transport policies, such as restriction of automobile access to the centre and limitation of parking facilities, appeared relatively recently in comparison to Swiss.
Individuals and household socio-demographic characteristics and mobility on a given weekday are collected. Using, starting and stopping times and the types of activities at the origin and destination, the one-day out-of-home activities diary can be deduced from the first to the last trip of the day.
We focus on the relationships between travel and activity times (work, shopping, leisure). Following Kitamura & al. (1992), the proportional assignment paradigm, where daily available time is proportionally allocated to different activities with fixed proportions (which may vary across individuals), is tested. It is then adapted to test the proportionality of daily travel times for each activity type with the different activities daily times. It clearly indicates that travel times are influenced by duration of activities.
Furthermore, we pay special attention to the impact of the residential location, captured by a segmentation of the urban area in three zones: centre, suburban and periurban.
The final section presents estimations realised in the duration model framework, for the cities of Lyon and Grenoble, based on their more recent surveys. Models for the travel time budgets are built and lead to non-linear estimates of the travel time intensities. Influence of the residential location is captured dummies for the urban zone types and for the proximity with a highway access.