Modelling the energy savings of a 'Get Britain Cycling' scenario of modal shift

This is a presentation I gave on 29th April 2014 at the E-Futures Doctoral Training Centre in Sheffield

Transcript

1. Modelling the energy savings of a 'best case' shift to cycling Investigating the consequences of the 'Get Britain Cycling' scenario Robin Lovelace, University of Leeds Presented 29th April, 2014 For E-Futures, Sheffield Slides Available: robinlovelace.net
2. What I'm going to talk about 1) The Get Britain Cycling vision 2) My research context + approach 3) Data 4) Method + next steps 5) Discussion + Conclusions
3. 1) Introduction – the vision The most important impacts of cycling may only emerge in the future. Yet most cycling policy evaluation, and cycle promotion focusses on benefits in the here and now. Considering energy impacts encourage long-term thinking. Credit: Oil Drum Article (2009) http://campfire.theoildrum.com/node/5976
4. What is Get Britain Cycling? • GBC = optimistic vision of a cycling nation • Scenario with quantitative targets • Overview of policies needed to get there • Produced for MPs by the APPCG, April 2013 See more at: http://allpartycycling.org/
5. • “the long-term ambition should be to increase cycle use to 10% of all journeys in 2025, and 25% by 2050” • Case studies of growth (e.g. Devon) • Funding: needs > £10 person/yr The GBC report “For technical reasons, computer modelling and forecasting has played little role in assessing the future potential of the volume of cycling” (Goodwin 2013)
6. Limitations of GBC • Lacks timeline beyond targets for 2025 and 2050 • No geographical disaggregation of cycling uptake • Nothing on who would be cycling where, replacing which modes and for what trip types • Basically, great overview, scant on detail • So the first stage was to create scenarios
7. 2) Research context + approach
8. Dynamic modelling of populations From Lovelace et al. (2011) • Published in Energy Policy • Borrowed equation from population ecology • 3 scenarios of growth in Sheffield • Wide boundary energy savings explored P: population r: max. Growth rate K: carrying capacity
9. Limitations of Energy Policy paper • No geographic disaggregation • Little evidence on carrying capacity • Little evidence on growth rates • Scarce case studies in UK • Useful concepts • Now: better data
10. The importance of (Robin's Really Relevant) Replacement Ratio • Critical to environmental impact of a bike trip is how many car trips it replaces • Varies over space, time and for different trips • Difficult to estimate
11. The PhD: Adding more data Lovelace (2014). Thesis finally available online. http://etheses.whiter ose.ac.uk/5027/
12. The Geo-dimension Flows into Sheffield's 4 largest Employing Output Area zones
13. Local scenarios
14. The individual level
15. The scenarios in my PhD • “Go Dutch” (Bicycles) or “Go Finnish” (telecommuting – working from home)
16. Waiting for the 2011 census... Distance released March 26th 2014 but no cross tabulation with mode. Doh! Can be estimated (e.g. IPF)
17. 3) Input data
18. The latest time series data • National Travel Survey ('02-'08) • Rises since 2005 • But 0.024% pa... • How long would that take?
19. Linear extrapolation • Rate needed: 0.67% points pa for 35 yrs • ~30 times current rate • Realistic?
20. Things can be linear for a while... • Doubling in 8 yr • “Levelling off” • Proxy for number of cycle trips in London • From “sixty automatic cycle counters” • Adjusted to account for season
21. But we do not live in a linear world! From Lovelace and Philips (2014)
22. The “doubling in 10 yr” model • Realistic? • Almost there for 2050 target • Way off 2025 target
23. DfT's model • Department for Transport's National Transport Model (NTM) is used for long-term planning • Projects declines in proportion of trips by bike!
24. The DfT's position • Met the DfT (via CTC) to lobby for better cycling in their NTM • They will work on cycling – need more! • Very old, complex model based on GDP • No demographics • See CTC articles on the topic or my briefing paper http://www.ctc.org.uk/news/government-planning-to-fail-on-cycling http://robinlovelace.net/talks/2014/01/21/modelling-energy-cycling.html
25. The “Official” model in context • Clear need for better models • Even less realistic than our 'back of envelope' models! • Cyclist Touring Club in talks with DfT to rectify this
26. The logistic growth model • Adaptable • Rate limited and capped -> realism • Includes 'positive feedback' and fundamental limits to growth Reproducible R code: k = 27 B = 1.7 r = 0.15 time = 0:35 lgrowth <- (B * k * exp(r * time)) / (k + B * (exp( r * time) - 1) )
27. Increasing distance of bicycle trips • Increased potential to replace car trips? • Or just “Wiggo effect”? – Leisure/utility trips? • Clashes with DfT model Source: https://www.gov.uk/government/publications/national-travel-survey-2012
28. • Undercounting of cycle trips (according to “cycle census”) • 25% rush hour trips in central London are bicycles (TfL 2013) • Lack of 'dose-response' studies (we're working on it!) – Regression analysis of impact of new cycle lanes – Evaluations of organisation-level interventions Data issues
29. 4) Method
30. 4) Method • Let's just say we get to “25% of trips” - a 'snapshot' model • Bikes can only replace trips of a certain distance • Or, more specifically, p(T -> B) ~ f(d T ) – Probability of trip T switching is related to distance • Assume distribution of distances is similar •
31. Distance-decay varies with mode and reason for trip... Source: My Thesis, based on Iacono et al. (2011)
32. Distance decay overall
33. Distance of trip with Gaussian kernel density estimate
34. Distance decay by mode
35. Distance decay functions – look at the overlaps
36. Method: next steps • Decide best functional form for distance decay • Adjust distance decay functions by reason for trip (e.g. Shopping, school, work) • Disaggregate the results geographically – Regional data in NTS – Smaller scale estimates from commute patterns? • Calculate CO2 emissions based on energy saving estimates • Implement dynamic model spatially -> associated energy savings in time + space
37. What I expect the results to look like
38. Discussion • Need to factor-in possibility for behaviour change (e.g. People cycling further, telecommuting) • The “Amazon effect” - people shopping less outside • Electric bicycles: elephant in the room • Impact of demographic shift -> spatial microsimulation model
39. 5) Conclusions • It is hard to imagine, let alone model, a future of very high cycling uptake • Logistic growth seems most plausible • The National Transport Model should be reconciled with trends and the GBC scenario • The energy savings of cycling depends primarily on what modes cycling replaces • Modelling process has 2 major benefits: – Making visions of the future more tangible – Identification of long-term needs
40. References • Goodwin, P. (2013). Get Britain cycling: report from the inquiry. London. Retrieved from http://allpartycycling.files.wordpress.com/2013/04/get-britain-cycling_goodwi n-report.pdf • Lovelace, R., Beck, S. B. M. B. M., Watson, M., & Wild, A. (2011). Assessing the energy implications of replacing car trips with bicycle trips in Sheffield, UK. Energy Policy, 39(4), 2075–2087. doi:10.1016/j.enpol.2011.01.051 • Lovelace, R., & Philips, I. (2014). The “oil vulnerability” of commuter patterns: A case study from Yorkshire and the Humber, UK. Geoforum, 51(0), 169–182. doi:http://dx.doi.org/10.1016/j.geoforum.2013.11.005 • Lovelace, R. (2014). The energy costs of commuting: a spatial microsimulation approach. University of Sheffield. Retrieved from http://etheses.whiterose.ac.uk/5027/