Travel behaviour and Edinburgh’s Low Emission Zone: a cross-sectional survey

Patient and Public Involvement

A stakeholder group was established early in the TRAVEL study, which included representatives from City of Edinburgh Council, Public Health Scotland, Sustrans, NHS Lothian, Royal Town Planning Institute, Transport Scotland, and a local community member. This group provided input into the survey content and language, recruitment of potential survey respondents via social media advertising, and interpretation of results. This group will also assist with the dissemination of study findings.

Setting. Edinburgh has a population of 500,000 (CEC, 2023a) with an annual mean temperature of 13 °C and 128 days of rainfall on average (Met Office, 2024). Edinburgh has an extensive bus service with 110 million journeys recorded in 2023 (Lothian Buses, 2024). Bus travel is free for residents of Scotland under the age of 22 years, 60 years or older, or with a disability. Edinburgh also has a tramline and train station in the city centre.

LEZ intervention. The LEZ was officially implemented on 31^st May 2022 with a two-year grace period: enforcement of non-compliance fees (starting at £60 and doubling thereafter) commenced on 1^st June 2024 (CEC, 2024b; CEC, 2024c). The LEZ allows entry of petrol cars/vans registered from 2006 with Euro 4 emission standards (Euro 6 for diesel cars/vans registered from 2015) (CEC, 2023b). Although 95% of petrol cars are estimated to be compliant, only 50% of diesel cars meet the required standard (CEC, 2023c). Funds were made available to support low-income households with the disposal of vehicles that were not LEZ-compliant, as well as vouchers for the purchase of bikes or public transport tickets (CEC, 2024a). The LEZ is to ensure NO₂ concentrations are below regulatory limits, to encourage more sustainable and active journeys, and to contribute to achieving net zero emissions by 2030. The LEZ covers <3% of the city with a total area of 3.1 km² (CEC, n.d.), mainly including the city centre (Figure 1).

Figure 1.

A map presenting the boundary of the Edinburgh Low Emission Zone (LEZ). Basemap from ©OpenStreetMap contributors (www.openstreetmap.org), available under the Open Database License.

Survey design. To answer the two main research questions, the survey collected data on demographics, travel habits, and reasons for travel choices. To explore modal change, questions addressed travel behaviour prior to the LEZ enforcement (pre-June 2024), as well as at the time of survey submission (October–November 2024). We included, as a measure of self-reported health, personal wellbeing questions (ONS, 2016). The main focus of the survey was the commute to work or study, with questions included for non-commute (i.e., other) travel. In summary, the questionnaire included the following content:

We developed the questionnaire with a professional survey company (Civica). The survey was hosted on Civica’s external survey platform, Alchemer, for which Civica have a license. The survey was piloted for clarity and ease of completion with colleagues not on the project team and was designed to be completed within 15–20 minutes. The introductory webpage that potential respondents were directed to was used as the participant information sheet. Participants provided explicit written agreement to complete the survey before progressing to the questions. The study protocol was published on Research Registry on 30^th July 2024 (UIN: researchregistry10542), and the Reading Independent Ethics Committee approved the study on 23^rd August 2024 (reference: 230824).

Survey recruitment. The target audience was adults (≥18 years old) working or studying in Edinburgh. We aimed to obtain responses from 350 individuals, using quotas to ensure representativeness in age and sex, as well as to limit respondents who did not commute (e.g., remote work/study only), but travelled to the LEZ for other reasons. The sample size was estimated using figures from the 2011 Scotland census as the most recent data available: a population size of 300,000 for the number of workers in Edinburgh, proportion of commute mode ranging from 5% (cycle) to 33% (car), margin of error of 5%, and a 95% confidence interval.

Individuals were identified, screened, and recruited through an existing panel of potential survey respondents (managed by Dynata) living in the Edinburgh area. Individuals accessed the survey through a weblink, which opened to a webpage describing the study’s purpose, content, and use of data. The survey was also made available on the IOM website and was advertised by organisations represented on the stakeholder group; thus, there were two versions of the survey for ‘panel’ and ‘community’ respondents (full survey version available at https://osf.io/at8sv/). No additional payments or incentives were provided for completion of the survey; however, panel respondents do receive point incentives for each survey completed.

Statistical analysis. We were interested in comparing travel behaviour of those commuting inside the LEZ to that of those whose commute did not traverse the LEZ. For the purposes of the survey, inside the LEZ was defined as a journey that passed through or included any part of the LEZ, whereas journeys outside the LEZ did not enter the LEZ at any point. The main analysis involved summary statistics of main survey questions and comparing travel behaviour responses by inside/outside LEZ commuters. Statistical analysis was undertaken to compare differences in commuting behaviour and perspectives before/after LEZ enforcement and whether travel involved the LEZ. We used several different methods depending on whether the data was categorical (chi-squared), Normal (t-test)/non-Normal (Kruskal-Wallis), paired (Wilcoxon signed rank test), or proportions (equality of proportions). The modes of transport were grouped into public transport (train, tram, and bus), active travel (bike, scooter, walk, and run), and vehicle (all other modes). Area-level deprivation (i.e., by datazone) was assigned using deciles of the Scottish Index of Multiple Deprivation (SIMD) matched to postcode. Respondents had the opportunity to include as free text any further comments on the LEZ. These remarks were broadly classified into positive (i.e., in support of the LEZ), negative (i.e., opposed to the LEZ), or neutral (i.e., neither in support nor opposed).

Analysis was stratified for the panel and community respondents, given the different recruitment methods. Only fully completed surveys were included in the data analysis, which was performed using R (R Core Team, Vienna, Austria).

Carbon emissions. We analysed the total commuting carbon emissions before and after enforcement for single modes of those commuting inside the LEZ. Emission estimates were based on reported mode, duration, and weekly frequency. As we did not have data on distance travelled, we estimated speed by mode via published sources, including the Physical Activity compendium (Herrmann et al., 2024).

The following speeds were used for the single modes reported in the panel survey results:

The following list presents carbon footprints per km for each transport mode. These estimates are based on Our World in Data (2023), unless otherwise stated.

Traffic displacement. The minimum distance from the full home postcode to the nearest segment of the LEZ boundary was calculated for each respondent using the open-source software, QGIS (v3.30.0) (https://qgis.org). To examine any potential traffic displacement, taking account of the potential discrepancy between reporting of postcodes and living inside/outside the LEZ boundary, we compared two sets of respondents who reported “more traffic” or “less traffic” as a reason for any changes to frequency, duration, or mode in commute or other travel. The data for inside and outside commuters, as well as people who travel to the LEZ for other reasons, was used for this analysis.

One set of respondents included those who reported living inside or outside the LEZ, and the other set used postcodes to define living inside, ≤1 km outside the boundary, or >1 km outside the boundary. The latter set excluded any respondents who reported living inside the LEZ, but did not have an outcode within the LEZ (i.e., EH1, EH2, EH3, or EH8).

Travel behaviour inside and outside the Low Emission Zone (LEZ)

More respondents commuted inside the LEZ in the panel (n=178, 51%) and community (n=174, 67%) surveys than outside the LEZ boundary (panel n=89, 25%; community n=52, 20%). Table 1 presents current commuting frequency, mode, and duration separately by travel inside and outside the LEZ for both the panel and community surveys. Respondents who reported travelling to the LEZ for other reasons (panel n=83, 24%; community n=34, 13%) are presented separately in Table 2. The frequency of travel was different inside and outside the LEZ (panel only), and using a single transport mode was more common inside the boundary. Commute durations were higher inside the LEZ than outside, though with high variability (Table 1).

The most common five modes of transport reported in the panel survey were bus, car (alone), walk, train, and tram. The corresponding list in the community survey was slightly different, including walk, bus, car (alone), bike, and electric bike (see Figure 5 for the full list of selected transport modes).

Figure 5.

The transport modes reported in the a) panel (n=267) and b) community (n=226) surveys for commutes either inside or outside the LEZ.

For the panel survey, there was a clear differentiation (p<0.001) with more people using the car alone for their commute outside the LEZ. More people tended to use the bus inside the LEZ (p=0.010). There were borderline significant differences in the proportions inside and outside the LEZ using bus & walk (p=0.110) and walking (p=0.054), but these comparisons were based on few respondents. There were similar findings in the community survey to that of the panel relating to use of the car, bus, and walking, with no difference in the proportion of people who cycle inside and outside the LEZ (p=0.439) (see Figure 6).

Figure 6.

The most common reported modes of transport inside and outside the LEZ in the a) panel (inside LEZ, n=178; outside LEZ, n=89) and b) community surveys (inside LEZ, n=174; outside LEZ, n=52).

The reasons for using a given transport mode are depicted in Figure 7. For brevity, and since the three most common overall modes in the panel and community surveys were the same (Figure 5), we present the individual reasons for commuting by car (alone), bus, and walking. As suggested by Figure 7, there are generally more differences between modes than either commuting inside or outside the LEZ or between the panel and community surveys. While convenience is clearly an important consideration for a given mode, ‘better for mental health’ and ‘healthier’ were much more common reasons for walking than car or bus.

Figure 7.

The percentage of i) panel and ii) community survey respondents reporting reasons for using a) car, b) bus, and c) walking by those commuting inside and outside the LEZ.

Travel behaviour before and after LEZ enforcement

Most respondents did not change whether they commuted inside or outside the LEZ following enforcement in June 2024, though more travel changes were observed for those commuting inside the LEZ (see Table 3). Of the 159 inside LEZ commuters in the panel survey, 70 (44%) reported a change in the frequency, duration, or mode of transport versus 8 out of 77 (10%) for the commuters outside of the LEZ. A similar trend was noted in the community survey, though to a lesser extent: 41 out of 171 (24%) of inside LEZ commuters reported such changes compared to 4 out of 47 (9%) of the outside commuters. The reasons provided for any changes are presented in Figure 8.

Figure 8.

The reasons given for commute changes for a) inside or b) outside the LEZ for the i) panel and ii) community surveys (panel, inside n = 9; panel, outside n = 13; community, inside n = 5; community, outside n = 2).

For the frequency of commuting, there was no general trend following LEZ enforcement in either the panel or community surveys (Figure 9). However, the three most common reasons given in the panel survey (i.e., ‘quicker’, ‘more convenient’, ‘cheaper’) implied increasing frequency, whereas those in the community survey (i.e., ‘longer’, ‘less convenient’, ‘expensive’) suggested less frequent commuting (Figure 10). A similar trend of disparate results between the panel and community surveys was also evident for changes in commute duration. The most reported reasons in the panel survey were ‘less traffic’, ‘public transport more reliable’, and ‘cycle routes improved’, indicating faster commute durations, in contrast to ‘more traffic’, ‘public transport less reliable’, and ‘different route’ (omitting ‘other’), as noted in the community survey (Figure 10).

Figure 9.

The changes to the commuting frequency in commuters travelling inside the LEZ before and after enforcement in the a) panel (n=59) and b) community (n=19) surveys.

Figure 10.

Reported commute modes for inside LEZ commuters before and after enforcement who changed mode of transport in the a) panel and b) community surveys (panel n = 45; community n = 10).

The shift in commute mode showed a general decline in the reporting of private vehicle use following the LEZ enforcement (Figure 10). When categorising reported changes in commute modes for those commuting inside the LEZ before and after enforcement based on any active travel (i.e., including multiple modes), there was an increase in prevalence following enforcement in the panel (n=53/178, 30% vs n=62/178, 35%) (Wilcoxon signed rank test p=0.014) and to a lesser extent in the community survey (n=103/174, 59% vs n=108/174, 62%) (Figure 11).

Figure 11.

The number of participants who reported any changes to travel mode on the commute for those travelling inside the LEZ before and following enforcement in the a) panel (n=45) and b) community (n=10) surveys.

The survey question regarding multiple commute modes contained up to 10 different reported modes of transport, suggesting respondents may have entered all possible modes of transport rather than only those used for a typical commute. Therefore, we limited the analysis of active travel minutes to respondents reporting a single travel mode (i.e., participants reporting multiple travel modes were excluded) (Table 4). For insider commuters reporting a single mode, the median commute time was 22 minutes for both the panel and community survey (panel before [n=6] and after [n=5], community before [n=34] and after [n=34] LEZ enforcement). The weekly carbon emissions combined for respondents using a single mode prior to LEZ enforcement was 405 kg, which is only slightly higher than the combined emissions following enforcement (404 kg) (p=0.798) (see Figure 12).

Figure 12.

The carbon emissions estimated for those commuting inside the LEZ before and after enforcement for those using a single mode of transport (n=58) in the panel survey.

Impacts of the LEZ

The survey asked participants explicitly about any changes prompted by their knowledge of the LEZ. Although the most frequent response was ‘none of the above’ for inside and outside LEZ commuters, more than half (n=107, 60%) and over one third (n=63, 36%) of inside commuters changed one or more travel behaviours in the panel and community surveys, respectively. Such behaviours included using public transport more and walking/cycling more. About 40% of outside LEZ commuters in each survey reported such changes (see Figure 13). For shorter journeys (defined as up to 5 miles), a greater proportion of those commuting inside the LEZ made changes to these journey types compared to those commuting outside the LEZ in the panel (n=98/178, 55% vs n=19/89, 21%), but not community (n=36/174, 21% vs n=11/52, 21%) survey (see Figure 14).

Figure 13.

Reported actions taken as a result of knowledge of the LEZ in a) inside and b) outside commuters in the i) panel and ii) community surveys (inside: panel n=178, community n=174; outside: panel n=89, community n=52).

Figure 14.

Reported actions for shorter trips (up to 5 miles) taken as a result of knowledge of the LEZ in inside and outside commuters in the a) panel (inside n=178, outside n=89) and b) community (inside n=174, outside n=52) surveys.

The survey included questions on agreement related to seven specific impacts of the LEZ (e.g., reducing air and noise pollution). Responses to each of the individual impacts were broadly similar in the panel survey, with less than 25% disagreement and at least a 2:1 ratio of agreement to disagreement. The community survey suggested lower agreement overall and more variation across the different potential impacts; most agreement was indicated for ‘reduce air pollution’ and the most disagreement was for ‘encourage active travel’ (Figure 15).

Figure 15.

Agreement related to impacts of the LEZ for a) panel (n=351) and b) community (n=260) surveys.

Respondents who did not commute for work or study reported other travel to the LEZ area (panel n=83, 24%; community n=34, 13%). The purpose of these journeys was predominantly for leisure or appointments, and the frequency was less often than that reported for commuters. Less than a third of these respondents (panel n=20, 24%; community n=10, 29%) reported changes (i.e., frequency, duration, mode) to their travel since enforcement. As observed in the results of those commuting into the LEZ, there was a general shift away from private vehicle use (though based on a smaller sample) (other travel results are presented in Table 2 and Figure 16).

Figure 16.

Reported commute modes for other travel inside the LEZ before and after enforcement in the a) panel (n = 16) and b) community (n = 5) surveys.

There were 85 of the 351 (24%) respondents that reported living within the LEZ boundary. The main outcodes (i.e., first 3–4 digits of full postcode) within the LEZ are EH1, EH2, EH3 and EH8; 30 of the 85 reported living in one of these outcodes. Potential reasons for this discrepancy could be due to confusion about the specific location of the LEZ, having multiple home addresses, or reporting postcodes inaccurately. Overall, 55 people entered “More traffic” or “Less traffic” as a reason for a change to their commute/travel. Of these 55, 20 were excluded in the postcode analysis, as noted above. There did not appear to be evidence of more traffic outside the LEZ in either the postcode or self-reported data analysis (see Figure 17).

Figure 17.

The traffic displacement analysis based on selecting “more traffic” or “less traffic” comparing a) reported postcodes (n=35) and b) respondents who self-reported living within/outside the LEZ boundary (n=55). Respondents who commuted or travelled for other reasons were included in the analysis (panel survey only).

Free text comments were more prevalent in the community survey (n=147/260, 57%) than submitted in the panel version (n=128/351, 36%). The panel survey comments reflected more positive LEZ sentiments than those of the community respondents (n=56/128, 44% vs n=35/147, 24%).

Strengths and limitations

Our study has several strengths. The survey represents short-term results of the travel behaviour impacts and perspectives from commuters on the enforcement of the Edinburgh LEZ. We collected detailed information involving travel inside and outside the LEZ, and before and following enforcement using two recruitment methods. To help with understanding the impacts, we included questions on specific actions attributed to knowledge/awareness of the LEZ. Several limitations should be noted with the interpretation of our results. Our survey used a cross-sectional approach and relied on the recall of individuals of travel behaviour prior to LEZ enforcement. While self-reported travel behaviour has been shown to be reliable at the group level (Kelly et al., 2014), recall of pre-enforcement travel may have been less accurate, particularly if the start of LEZ enforcement was not a particularly salient event (Müggenburg, 2021). The use of objective data to assess traffic volumes or public transport use, for example, should be undertaken to compare with analyses based on subjective inputs (Tassinari, 2024). Impacts may have been underestimated, given the LEZ was implemented two years earlier (without issuing non-compliance fees); individuals may have undertaken specific changes prior to enforcement, which would not have been captured through the analysis of travel before and after enforcement. Nevertheless, we did ask about the replacement of non-compliant vehicles, which applied to a minority of respondents: four of eight individuals commuting inside the LEZ reported upgrading their non-compliant vehicle. Survey questions related to behaviour changes due to knowledge of the LEZ and agreement on potential LEZ impacts did not specify ‘enforcement’, so it was not possible to distinguish effects from implementations versus enforcement in those responses.

There were two notable issues with the collection of multi-modal data. First, it was apparent that the reporting of multiple modes may have represented separate commute journeys, which, if combined, would have overestimated commute duration. Second, the survey did not distinguish mode-specific duration of pre-enforcement travel. Thus, it is important for travel surveys to use explicit language for multi-modal travel to be able to answer the relevant research questions. There were inconsistencies between the locations of postcodes and whether a respondent reported living inside the LEZ. This weakened our ability to examine potential traffic displacement experienced by those living in close proximity to the LEZ boundary. Surveys were completed 4–5 months following the start of enforcement, so our results represent short-term findings, which should be compared with results over a longer follow-up period. The focus of our survey was on LEZ enforcement, though other concurrent policy/infrastructure changes may have influenced travel behaviour. For example, several respondents selected ‘cycle routes improved’ as the reason for changes in duration. Subsequent research should examine LEZ enforcement impacts in the broader context of other mobility-related policies in Edinburgh and other urban settings.