Transcript 2 A Martinox

Quantitative modelling of the STOA
scenarios with ASTRA: potentials
and limits
Angelo Martino
TRT Trasporti e Territorio
The ASTRA model
 Tool for strategic assessment of transport policy and trends
 Integration of transport – environment – technology –
economy
 Long record of experiences in European and national projects
 A new version of the model is currently under development
on behalf of DG MOVE Unit A3 (as part of the ASSIST project)
 Capability to simulate several different policy measures, e.g.:





Technological improvements
Pricing and charging
Economic incentives (e.g. subsidies, feebate)
Targets (e.g. emission standards)
Regulation (e.g. speed limits)
STOA workshop – Brussels, 7th May 2013
2
The ASTRA model methodology
 Based on System Dynamics (implemented with Vensim®)
 Through computer simulation, SD allows to understand how
systems change over time
 The behaviour of systems is primary determined by its
feedback mechanisms
 System is continuously evolving: unlike CGE models or
transport assignment alghoritms equilibrium is never reached
 Time horizon: 1995 – 2050
 Spatial coverage: EU29 (EU27 + Norway and Switzerland)
 Geographical detail: mostly country level
 www.astra-model.eu
STOA workshop – Brussels, 7th May 2013
3
ASTRA model structure
STOA workshop – Brussels, 7th May 2013
4
Modelling technology improvements
 The implementation of technology measures affects:



Fuel/energy consumption of vehicles
Cost of vehicles (investment and operating costs)
Investments in R&D
 Implementing technology measures in ASTRA means:




Changing fuel/energy consumption factors
Changing CO2 emission factors (linked to fuel consumption)
Changing cost of vehicle for each fuel category (and size)
Adding investments to develop innovation
STOA workshop – Brussels, 7th May 2013
5
Modelling technology improvements
Investments
GDP
Unitary energy
consumption
Vehicle cost
Transport
demand
Energy
consumption
CO2 emissions
Operating cost
Government
revenues
Fleet size and
composition
Learning curves
Fuel duties
STOA workshop – Brussels, 7th May 2013
6
Modelling policies
 Changes generated by policy measures are reflected in three
main areas:
 User cost



Additional cost due to charges or taxes (road pricing or fuel taxes)
Variable operating costs (reduced fuel consumption/eco-driving)
Vehicle purchase cost (feebate)
 User time (speed limits, reserved lanes, etc.)


New regulation (speed limits)
New facilities (reserved lanes)
 Authority expenditure
 Subsidies (to reduce public transport fares)
 Increased public investment (new transport facilities)
 Increased expenditure for public activities (managing cordon pricing)
STOA workshop – Brussels, 7th May 2013
7
Modelling policies
Infrastructure
Public
expenditure
Aggregated
demand
User travel time
Transport
demand
GDP
User
generalised cost
by mode
Transport
demand
by mode
Energy
consumption
User travel cost
Cordon pricing
STOA workshop – Brussels, 7th May 2013
CO2 emissions
Public
revenues
8
GHG-TransPoRD reference scenario
 The GHG-TransPoRD Reference Scenario was based on two
main sources:


“EU energy trends to 2030 — update 2009” (PRIMES model, used for
assessment of the EC White Paper)
ADAM project reference scenario (after 2030)
 Key assumptions:





Moderate economic growth
Transport demand growing more or less like economy
Some improvement of transport energy efficiency
Transport CO2 emissions increasing as efficiency gains are more than
offset by demand growth
No penetration of innovative cars in the EU fleet
 The project was concluded in early 2012:
http://www.ghg-transpord.eu
STOA workshop – Brussels, 7th May 2013
9
ASTRA results explained - 1
 Scenario 1 - Cleaner modes: fuel cell cars climb to 61% of the
fleet




Vehicle fleet composition in ASTRA develops as new cars enter and old
cars are scrapped
New cars market shares (by technology) are modelled by a choice
algorithm based on (purchase and operating) cost, filling station
network availability and consumer preferences
Parameters are calibrated on observed fleet development in the past
Observed data on new technologies does not exist: validity of
parameters is a professional judgment
STOA workshop – Brussels, 7th May 2013
10
ASTRA results explained - 2
 Scenario 2 – Changing the modal split: car share falls only
from 64% to 58% and truck share from 57% to 50%






In ASTRA the mode split is computed by a logit algorithm applied
separately to several demand segments (distance band, purpose,
freight type…)
The algorithm is based on cost, time and other parameters reflecting
“alternative specific constants”
The algorithm is calibrated to replicate observed mode split
In last 15 years mode split has not changed dramatically
“Alternative specific constants” parameters reflect this situation and
do not allow a major mode shift
Also capacity constraints prevent large shift from car to other modes
STOA workshop – Brussels, 7th May 2013
11
ASTRA results explained - 3
 Scenario 3 - Reducing growth rates: despite halved transport
demand, the gain in CO2 emission reduction is much lower




Technological improvements become less effective because are
applied to a lower amount of traffic: the reduction of energy
consumption per km is applied to less km and then the total cut is
lower
Neither mode split nor fleet composition is drastically changed with
respect to AFS
Innovative vehicles penetration rate is higher because of extremely
high oil price which pushes the convenience of alternative fuels
Reduced freight volumes are modelled by reducing intra-EU trade and
this lead to negative economic output
STOA workshop – Brussels, 7th May 2013
12
Thanks for your attention
www.astra-model.eu
[email protected]
STOA workshop – Brussels, 7th May 2013
13