Assumptions | Energy

Scenario assumptions for the TIMER energy model

The assumptions for the different SRES energy scenarios in the TIMER model are based on the corresponding marker scenarios as published by IPCC (2000) and their story lines. The table summarizes the various assumptions.

A general explanation and specific assumptions are listed for:

Demand side
Supply side
Electricity generation
Trade.

Assumptions for the demand side

Energy tax levels

In the model, for all sectors, regions and fuels time-dependent energy taxes are added to the end-use prices. The taxes are added in terms of US$/GJ and are, in general, higher for electricity and transport fuels.

A1B, A1F, A1T Energy end-use taxes for all sectors, regions and fuels converge towards a 2100 level equal to 1995 USA levels (e.g., 4-6 US$/GJ in transport)

A2 In 2100, all end-use taxes reach a level in between the final A1B level and the current regional level

B1 Energy end-use taxes for all sectors, regions and fuels converge towards a 2100 level equal to 1995 Western European levels (e.g., 14-16 US$/GJ in transport)

B2 In 2100, all end-use taxes reach a level in between the final B1 level and the current regional level

Structural change parameter

Energy demand is calculated by first estimating the demand for energy services based on monetary activity indicators per sector. The relation between the demand for energy services and the monetary activity indicator is assumed to change of timer as result of structural change, e.g. changes in modal split in transport or shifts from heavy to light industry in the sector industry. By default, structural change in the model is described by an 'inverted U-shape curve' as a function of 'modernization', indicating that the demand for energy services at the first stages of 'modernization' tends to increase faster than activity, and at latter stages tends to increase slower than activities (measured in monetary terms). For 'modernization' per capita activity level is used as a proxy. The default formulation assumes saturation at high levels of modernization at a constant per capita activity. A scenario and time-dependent structural change parameter is used to multiply this default setting to model the impacts of more or less intensive material life-styles.
For the sectors services and residential separate structural change curves are used for heat and electricity. For all other sectors, first total energy demand is determined- which is in a next step subdivided into heat and electricity by means of a separate scenario file.

A1B, A1F Reaches a value 55-85% above default value to reflect material intensive life-styles

A1T Reaches a value 55-85% above default value to reflect material intensive life-styles

A2 Reaches a value 45-75% above default value to reflect material intensive life-styles

B1 Reaches a value 10-20% below default value to reflect transition towards less material-intensive lifestyles

B2 Reaches a value 25-35% above default value

Pay-back times

Also price-induced energy-efficiency improvement can reduce the amount of final energy needed for the same amount of energy services. The rate of investments in energy-efficiency in response to (increasing) energy prices is among others related to the pack-back times considered to be acceptable by the relevant investors. For instance, a pack-back time of 1 year indicates that investors are only interested to make investments in those options that have such a high rate of return that the original investment is paid back within 1 year. Generally, pay-back times tend to be lower in regions with relatively high uncertainties for investors and for sectors that are dominated by small-scale actors (transport, residential sector).

A1B, A1F, A1T, A2 Reaches for all regions the current high-income region values (1.2-3.3 years, depending on sector)

B1 Increases for all regions to levels twice the current high-income region values (2.8-6.2 years, depending on sector)

B2 Increased for all regions to a level half-way current values and B1 values

End-use conversion efficiency secondary fuels

In the conversion between the demand for energy services and final energy use end-use conversion efficiencies play an important role. In TIMER, these fuel, sector and time-dependent efficiencies are included as exogenous scenario files.

A1B, A1F, A1T, A2, B1, B2 Increases in all scenarios to a level of 80% for coal (85% for industry), 90% for oil, 94% for natural gas and 35% for traditional fuels (45% in industry). For electricity the conversion efficiency is equal to 1.0

Autonomous efficiency improvement (AEEI)

Autonomous efficiency improvement captures the increasing efficiency between the demand for energy services and the demand for final energy caused by price-independent technology development. Estimates of the historic importance of this parameter vis-à-vis price-induced energy efficiency improvement or structural change differ strongly - medium estimates are around 0.0-0.5 % per year. In TIMER it is assumed that AEEI shows a similar pattern as other forms of technology development, i.e. is relatively fast in the early stages of development and slowly reduces its speed along with cumulative output. The rate of AEEI in scenarios can be varied per region and time-period from the default settings by moving faster along the 'development' axis.

A1B, A1F Between 2000-2040 low income regions strongly catch up, in relation to high economic growth rates

A1T Between 2000-2040 low income regions very strongly catch up, reflecting technology optimism

A2 Between 2000-2040 low income regions slowly catch up

B1 Between 2000-2040 low income regions strongly catch up as result technology transfer and economic growth

B2 Between 2000-2040 low income regions moderately catch up

Non energy use

Fuels can also be used for uses other than for energy, in particular for feedstocks. In TIMER feedstocks are modelled as function of industrial activities, characterized by a certain intensity of use and an annual efficiency improvement.

A1B, A1F, A1T Default assumptions on intensity and efficiency, finally resulting in a global consumption of 52 EJ in 2100 for A1B and A1T and 58 EJ for A1F

A2 Default assumptions on intensity and efficiency, finally resulting in a global consumption of 15 EJ in 2100

B1 Default assumptions on intensity and efficiency, finally resulting in a global consumption of 21 EJ in 2100

B2 Default assumptions on intensity and efficiency, finally resulting in a global consumption of 27 EJ in 2100

Share electricity in total consumption

For the sectors industry, transport and other an autonomous scenario file per sector and region determines the share of electricity in total consumption.

A1B, A1F, A1T, A2, B1, B2 Electricity shares increase in all sectors. While there are regional differences - based on the current situation - on average the share of electricity reaches the following values: industry 50%, transport 10-20%, residential 30-70%, services 60-90%, other 20%. There are small differences between the scenarios, with B1 using slightly lower and A2 slightly higher values within the ranges mentioned above.

Premium factors final energy carriers

The end-use fuels 'solid fuels', 'liquid fuels' and 'gaseous fuels' compete for market share on the basis of their relative prices and an additional premium value (in some cases, mostly historically, sectors are shielded for penetration of certain fuels). These premium values capture other factors that determine market shares of fuels, in particular consumer preferences and environmental policies. If no preference or aversion for a certain fuel type exists, the premium value is set at 1. In most TIMER scenarios in particular premium values for coal in the transport, residential and services sectors are raised to reflect environmental and convenience considerations.

A1B Strong aversion from use of coal for health, convenience and environmental reasons

A1F Modest aversion from use of coal; problems related to coal use are solved differently

A1T Strong aversion from use of coal for health, convenience and environmental reasons

A2 Strong aversion from use of coal, but regional differences based on current values

B1 Very strong aversion from use of coal for environmental reasons

B2 Very strong aversion from use of coal for environmental reasons

Technology learning rate for price-induced energy efficiency (PIEEI).

Energy efficiency investments tend to become most cost-effective in time due to technological development. As for other technologies, in TIMER technology development for energy efficiency is modelled by a 'learning curve'. The actual rate of learning is a function of initial settings (assumptions about cumulative output before starting year of the simulation) and the progress ratio (1- reduction of costs for a doubling of cumulative output). In literature often progress ratios between 0.8 and 0.95 are reported. In TIMER most curves have been calibrated in such way that 0.9 is considered to be reasonable default value. For PIEEI the historic progress ratio is set at 0.85. Progress ratios can vary in time.

A1B Modest - normal learning rate (0.92 in 2100)

A1F Modest - normal learning rate (0.92 in 2100)

A1T High learning rate (0.88 in 2100)

A2 Modest - normal learning rate (0.93 in 2100)

B1 Normal learning rate (0.90 in 2100)

B2 Modest - normal learning rate (0.92 in 2100)

Assumptions for the supply side

Resources for fossil fuels

In TIMER depletion of fossil fuels is given in terms of supply-costs curves (assuming constant technology). For all fuels, these curves have been derived from a publication of Rogner (1997). The curves do not only include currently known reserves and conventional occurrences but also estimates for non-discovered and unconventional occurrences (e.g. methane hydrates and tar sands).

A1B, A1F, A1T, A2, B1, B2 Default values for fossil fuel resources and supply-cost curves have been used. These values and curves have been derived from Rogner (1997)

Technology learning rate for coal (surface coal only)

As for other technologies, in TIMER technology development in terms of the production costs for surface coal is modelled by a 'learning curve'. The actual rate of learning is a function of initial settings (assumptions about cumulative output before starting year of the simulation) and the progress ratio (1- reduction of costs for a doubling of cumulative output). In literature often progress ratios between 0.80 and 0.95 are reported. In TIMER most curves have been calibrated in such way that 0.90 is considered to be reasonable default value. For surface coal the historic progress ratio is set at 0.95. Progress ratios can vary in time. For underground coal, technology development is assumed to be unimportant for production costs.

A1B Slow (0.95)

A1F Normal (0.91)

A1T Very slow (0.97)

A2 Slow (0.95)

B1 Slow (0.95)

B2 Slow (0.96)

Technology learning rate for oil

As for other technologies, in TIMER technology development in terms of the production costs for oil is modelled by a 'learning curve'. The actual rate of learning is a function of initial settings (assumptions about cumulative output before starting year of the simulation) and the progress ratio (1- reduction of costs for a doubling of cumulative output). In literature often progress ratios between 0.8 and 0.95 are reported. In TIMER most curves have been calibrated in such way that 0.9 is considered to be reasonable default value. For oil the historic progress ratio is set at different values for different regions - but tend to converge around 0.9. Progress ratios can vary in time.

A1B Default (0.90)

A1F Fast (0.87)

A1T, A2, B1 Default (0.90)

B2 Slower (0.92)

Technology learning rate for natural gas

As for other technologies, in TIMER technology development in terms of the production costs for natural gas is modelled by a 'learning curve'. The actual rate of learning is a function of initial settings (assumptions about cumulative output before starting year of the simulation) and the progress ratio (1- reduction of costs for a doubling of cumulative output). In literature often progress ratios between 0.8 and 0.95 are reported. In TIMER most curves have been calibrated in such way that 0.9 is considered to be reasonable default value. For natural gas the historic progress ratio is set at different values for different regions - but tend to converge around 0.9. Progress ratios can vary in time.

A1B Default (0.90)

A1F Fast (0.86)

A1T, A2, B1 Default (0.90)

B2 Default (0.90)

Technology learning rate for biofuels

As for other technologies, in TIMER technology development in terms of the production costs for biofuels is modelled by a 'learning curve'. The actual rate of learning is a function of initial settings (assumptions about cumulative output before starting year of the simulation) and the progress ratio (1- reduction of costs for a doubling of cumulative output). In literature often progress ratios between 0.8 and 0.95 are reported. In TIMER most curves have been calibrated in such way that 0.9 is considered to be reasonable default value. For biofuels the historic progress ratio is around 0.88. Progress ratios can vary in time.

A1B Strong till 2040 (0.87), default from 2040 onwards (0.90)

A1F Slow to very slow (0.92-0.96)

A1T Strong till 2040 (0.86), default from 2040 onwards (0.90)

A2 Strong till 2020 (0.89), slow from 2020 to 2100 (0.96)

B1 Strong till 2040 (0.87), default from 2040 onwards (0.90)

B2 Strong till 2040 (0.87), default from 2040-2060, 2060-2100 slower (0.92)

Technology transfer parameter

Obviously, technology development is not independent in different regions around the world. In TIMER, this can be simulated by means of the 'technology transfer parameter' that result in transferring the technology progress ratio of the leading region to other regions (catching up). Historically, this parameter is switched off (set at zero).

A1B Modest

A1F Modest

A1T Modest

A2 Zero

B1 Modest

B2 Very low

Assumptions for the electricity supply

Technology learning rate for solar/wind

As for other technologies, in TIMER technology development in terms of the production costs for solar/wind is modelled by a 'learning curve'. The actual rate of learning is a function of initial settings (assumptions about cumulative output before starting year of the simulation) and the progress ratio (1- reduction of costs for a doubling of cumulative output). In literature often progress ratios between 0.8 and 0.95 are reported. In TIMER most curves have been calibrated in such way that 0.9 is considered to be reasonable default value. For solar/wind the historic progress ratio is set at 0.85. Progress ratios can vary in time.

A1B Very strong till 2050 (0.85); default between 2050-2100 (0.90)

A1F Slow to very slow (0.95 from 2050 onwards)

A1T Very strong till 2050 (0.82); very strong between 2050-2100 (0.85)

A2 Slow (0.93 from 2050 onwards)

B1 Very strong till 2050 (0.85); default between 2050-2100 (0.90)

B2 Very strong till 2050 (0.85) although low income regions have some delay; default between 2050-2100 (0.91)

Technology learning rate for nuclear

As for other technologies, in TIMER technology development in terms of the production costs for nuclear power is modelled by a 'learning curve'. The actual rate of learning is a function of initial settings (assumptions about cumulative output before starting year of the simulation) and the progress ratio (1- reduction of costs for a doubling of cumulative output). In literature often progress ratios between 0.8 and 0.95 are reported. In TIMER most curves have been calibrated in such way that 0.9 is considered to be reasonable default value. For nuclear power the historic progress ratio is set at average at 1.0 (as additional environmental and risk related conditions for nuclear power have offset costs reductions due to improved technology). Progress ratios can vary in time.

A1B Moderate (0.93)

A1F Slow (0.97)

A1T Moderate (0.93)

A2 Slow (0.96 - 0.97)

B1 Slow (0.95)

B2 Slow (0.96)

Use of hydropower

The use hydropower in TIMER is determined by two exogenous scenario files: 1) the share of the technical potential per region actually used for hydropower capacity and 2) the load factor of this capacity.

A1B Relatively high (40% of global potential)

A1F Default (35% of global potential)

A1T Relatively high (40% of global potential)

A2 Default (35% of global potential)

B1 Default (35% of global potential)

B2 Default (35% of global potential)

Premium values within thermal fuel production

The use of the different fuel types for electricity generation in thermal power plants is determined by their relative prices, the efficiency with which these fuels can be used and an additional premium value. These premium values capture other factors that determine market shares of fuels, in particular environmental policies and strategic considerations. If no preference or aversion for a certain fuel type exists, the premium value is set at 1. In most TIMER scenarios in particular premium values for are raised in certain scenarios to reflect environmental considerations.

A1B, A1F Very small aversion from coal, simulating small cost increase due to add-on technology

A1T Modest aversion from coal, simulating small cost increase due to add-on technology

A2 Regionally determined preference or aversion from coal, oil and natural gas based on 1995 situation and slow convergence to unity

B1 Very strong aversion from use of coal for environmental reasons (1.7)

B2 Regionally determined aversion from coal, based on 1995 situation, ranging from 1.3 to 8.0

Premium values for different electricity production forms

The use of the different forms electricity generation (thermal power plants, nuclear, renewables) is determined their relative prices and an additional premium value. These premium values capture other factors that determine market shares of fuels, in particular environmental policies and strategic considerations. If no preference or aversion for a certain fuel type exists, the premium value is set at 1. In most TIMER scenarios in particular premium values for are raised in certain scenarios to reflect environmental considerations.

A1B In 2100 almost indifferent

A1F In 2100 indifferent

A1T, A2 In 2100 almost indifferent

B1, B2 In 2100 preference for renewable electricity production; modest aversion towards nuclear; strong aversion towards fossil

Thermal efficiency electricity generation

An important parameter in thermal electricity generation is the efficiency by which primary fuels are converted into electricity. These efficiencies are a function of time, region and fuel type. The efficiencies can change over time, among others as result of technology development, additional environmental technologies etc.

A1B Increases to 0.47-0.49 for coal, 0.51-0.54 for oil and 0.56-0.58 for natural gas

A1F, A1T Increases to 0.48-0.50 for coal, 0.52-0.55 for oil and 0.57-0.59 for natural gas

A2 Increases to 0.43-0.47 for coal, 047-0.52 for oil and 0.51-0.56 for natural gas

B1, B2 Increases to 0.44-0.48 for coal, 0.49-0.53 for oil and 0.53-0.57 for natural gas

Assumptions for the energy trade

Trade constraints

All regions in TIMER can decide to import fossil fuels and biofuels from other region, in first instance based on differences in production and transportation costs. In addition, various forms of trade constraints exist. First of all, import and export constraints per region can simply prevent trade, even if costs consideration would indicate trade to be an attractive option. Secondly, region can favour domestic production by means of economic subsidies. Finally, a matrix between the different regions exist that can make trade between two specific regions less attractive. This matrix can first of all incorporate certain geographical difficulties for trade (e.g. the mountain ranges), but can also be used to introduce regional blocks.

A1B, A1F, A1T All import and exports constraints are removed

A2 Regional blocks introduced by assuming trade barriers

B1 All import and exports constraints are removed

B2 Regional blocks introduced by assuming trade barriers

Transport costs

The trade between regions is among other a function of transport costs. Transport costs on their turn are function of a regional distance matrix and transport costs per kilometre. The latter are an exogenously set model input.

A1B, A1F, A1T Decrease for natural gas; constant for other fuels

A2 Constant for all fuels

B1 Decrease for natural gas; constant for other fuels

B2 Constant for all fuels

Assumptions for the demand side
Energy tax levels
	In the model, for all sectors, regions and fuels time-dependent energy taxes are added to the end-use prices. The taxes are added in terms of US$/GJ and are, in general, higher for electricity and transport fuels.
		A1B, A1F, A1T	Energy end-use taxes for all sectors, regions and fuels converge towards a 2100 level equal to 1995 USA levels (e.g., 4-6 US$/GJ in transport)
		A2	In 2100, all end-use taxes reach a level in between the final A1B level and the current regional level
		B1	Energy end-use taxes for all sectors, regions and fuels converge towards a 2100 level equal to 1995 Western European levels (e.g., 14-16 US$/GJ in transport)
		B2	In 2100, all end-use taxes reach a level in between the final B1 level and the current regional level
Structural change parameter
	Energy demand is calculated by first estimating the demand for energy services based on monetary activity indicators per sector. The relation between the demand for energy services and the monetary activity indicator is assumed to change of timer as result of structural change, e.g. changes in modal split in transport or shifts from heavy to light industry in the sector industry. By default, structural change in the model is described by an 'inverted U-shape curve' as a function of 'modernization', indicating that the demand for energy services at the first stages of 'modernization' tends to increase faster than activity, and at latter stages tends to increase slower than activities (measured in monetary terms). For 'modernization' per capita activity level is used as a proxy. The default formulation assumes saturation at high levels of modernization at a constant per capita activity. A scenario and time-dependent structural change parameter is used to multiply this default setting to model the impacts of more or less intensive material life-styles. For the sectors services and residential separate structural change curves are used for heat and electricity. For all other sectors, first total energy demand is determined- which is in a next step subdivided into heat and electricity by means of a separate scenario file.
		A1B, A1F	Reaches a value 55-85% above default value to reflect material intensive life-styles
		A1T	Reaches a value 55-85% above default value to reflect material intensive life-styles
		A2	Reaches a value 45-75% above default value to reflect material intensive life-styles
		B1	Reaches a value 10-20% below default value to reflect transition towards less material-intensive lifestyles
		B2	Reaches a value 25-35% above default value
Pay-back times
	Also price-induced energy-efficiency improvement can reduce the amount of final energy needed for the same amount of energy services. The rate of investments in energy-efficiency in response to (increasing) energy prices is among others related to the pack-back times considered to be acceptable by the relevant investors. For instance, a pack-back time of 1 year indicates that investors are only interested to make investments in those options that have such a high rate of return that the original investment is paid back within 1 year. Generally, pay-back times tend to be lower in regions with relatively high uncertainties for investors and for sectors that are dominated by small-scale actors (transport, residential sector).
		A1B, A1F, A1T, A2	Reaches for all regions the current high-income region values (1.2-3.3 years, depending on sector)
		B1	Increases for all regions to levels twice the current high-income region values (2.8-6.2 years, depending on sector)
		B2	Increased for all regions to a level half-way current values and B1 values
End-use conversion efficiency secondary fuels
	In the conversion between the demand for energy services and final energy use end-use conversion efficiencies play an important role. In TIMER, these fuel, sector and time-dependent efficiencies are included as exogenous scenario files.
		A1B, A1F, A1T, A2, B1, B2	Increases in all scenarios to a level of 80% for coal (85% for industry), 90% for oil, 94% for natural gas and 35% for traditional fuels (45% in industry). For electricity the conversion efficiency is equal to 1.0
Autonomous efficiency improvement (AEEI)
	Autonomous efficiency improvement captures the increasing efficiency between the demand for energy services and the demand for final energy caused by price-independent technology development. Estimates of the historic importance of this parameter vis-à-vis price-induced energy efficiency improvement or structural change differ strongly - medium estimates are around 0.0-0.5 % per year. In TIMER it is assumed that AEEI shows a similar pattern as other forms of technology development, i.e. is relatively fast in the early stages of development and slowly reduces its speed along with cumulative output. The rate of AEEI in scenarios can be varied per region and time-period from the default settings by moving faster along the 'development' axis.
		A1B, A1F	Between 2000-2040 low income regions strongly catch up, in relation to high economic growth rates
		A1T	Between 2000-2040 low income regions very strongly catch up, reflecting technology optimism
		A2	Between 2000-2040 low income regions slowly catch up
		B1	Between 2000-2040 low income regions strongly catch up as result technology transfer and economic growth
		B2	Between 2000-2040 low income regions moderately catch up
Non energy use
	Fuels can also be used for uses other than for energy, in particular for feedstocks. In TIMER feedstocks are modelled as function of industrial activities, characterized by a certain intensity of use and an annual efficiency improvement.
		A1B, A1F, A1T	Default assumptions on intensity and efficiency, finally resulting in a global consumption of 52 EJ in 2100 for A1B and A1T and 58 EJ for A1F
		A2	Default assumptions on intensity and efficiency, finally resulting in a global consumption of 15 EJ in 2100
		B1	Default assumptions on intensity and efficiency, finally resulting in a global consumption of 21 EJ in 2100
		B2	Default assumptions on intensity and efficiency, finally resulting in a global consumption of 27 EJ in 2100
Share electricity in total consumption
	For the sectors industry, transport and other an autonomous scenario file per sector and region determines the share of electricity in total consumption.
		A1B, A1F, A1T, A2, B1, B2	Electricity shares increase in all sectors. While there are regional differences - based on the current situation - on average the share of electricity reaches the following values: industry 50%, transport 10-20%, residential 30-70%, services 60-90%, other 20%. There are small differences between the scenarios, with B1 using slightly lower and A2 slightly higher values within the ranges mentioned above.
Premium factors final energy carriers
	The end-use fuels 'solid fuels', 'liquid fuels' and 'gaseous fuels' compete for market share on the basis of their relative prices and an additional premium value (in some cases, mostly historically, sectors are shielded for penetration of certain fuels). These premium values capture other factors that determine market shares of fuels, in particular consumer preferences and environmental policies. If no preference or aversion for a certain fuel type exists, the premium value is set at 1. In most TIMER scenarios in particular premium values for coal in the transport, residential and services sectors are raised to reflect environmental and convenience considerations.
		A1B	Strong aversion from use of coal for health, convenience and environmental reasons
		A1F	Modest aversion from use of coal; problems related to coal use are solved differently
		A1T	Strong aversion from use of coal for health, convenience and environmental reasons
		A2	Strong aversion from use of coal, but regional differences based on current values
		B1	Very strong aversion from use of coal for environmental reasons
		B2	Very strong aversion from use of coal for environmental reasons
Technology learning rate for price-induced energy efficiency (PIEEI).
	Energy efficiency investments tend to become most cost-effective in time due to technological development. As for other technologies, in TIMER technology development for energy efficiency is modelled by a 'learning curve'. The actual rate of learning is a function of initial settings (assumptions about cumulative output before starting year of the simulation) and the progress ratio (1- reduction of costs for a doubling of cumulative output). In literature often progress ratios between 0.8 and 0.95 are reported. In TIMER most curves have been calibrated in such way that 0.9 is considered to be reasonable default value. For PIEEI the historic progress ratio is set at 0.85. Progress ratios can vary in time.
		A1B	Modest - normal learning rate (0.92 in 2100)
		A1F	Modest - normal learning rate (0.92 in 2100)
		A1T	High learning rate (0.88 in 2100)
		A2	Modest - normal learning rate (0.93 in 2100)
		B1	Normal learning rate (0.90 in 2100)
		B2	Modest - normal learning rate (0.92 in 2100)

Assumptions for the supply side
Resources for fossil fuels
	In TIMER depletion of fossil fuels is given in terms of supply-costs curves (assuming constant technology). For all fuels, these curves have been derived from a publication of Rogner (1997). The curves do not only include currently known reserves and conventional occurrences but also estimates for non-discovered and unconventional occurrences (e.g. methane hydrates and tar sands).
		A1B, A1F, A1T, A2, B1, B2	Default values for fossil fuel resources and supply-cost curves have been used. These values and curves have been derived from Rogner (1997)
Technology learning rate for coal (surface coal only)
	As for other technologies, in TIMER technology development in terms of the production costs for surface coal is modelled by a 'learning curve'. The actual rate of learning is a function of initial settings (assumptions about cumulative output before starting year of the simulation) and the progress ratio (1- reduction of costs for a doubling of cumulative output). In literature often progress ratios between 0.80 and 0.95 are reported. In TIMER most curves have been calibrated in such way that 0.90 is considered to be reasonable default value. For surface coal the historic progress ratio is set at 0.95. Progress ratios can vary in time. For underground coal, technology development is assumed to be unimportant for production costs.
		A1B	Slow (0.95)
		A1F	Normal (0.91)
		A1T	Very slow (0.97)
		A2	Slow (0.95)
		B1	Slow (0.95)
		B2	Slow (0.96)
Technology learning rate for oil
	As for other technologies, in TIMER technology development in terms of the production costs for oil is modelled by a 'learning curve'. The actual rate of learning is a function of initial settings (assumptions about cumulative output before starting year of the simulation) and the progress ratio (1- reduction of costs for a doubling of cumulative output). In literature often progress ratios between 0.8 and 0.95 are reported. In TIMER most curves have been calibrated in such way that 0.9 is considered to be reasonable default value. For oil the historic progress ratio is set at different values for different regions - but tend to converge around 0.9. Progress ratios can vary in time.
		A1B	Default (0.90)
		A1F	Fast (0.87)
		A1T, A2, B1	Default (0.90)
		B2	Slower (0.92)
Technology learning rate for natural gas
	As for other technologies, in TIMER technology development in terms of the production costs for natural gas is modelled by a 'learning curve'. The actual rate of learning is a function of initial settings (assumptions about cumulative output before starting year of the simulation) and the progress ratio (1- reduction of costs for a doubling of cumulative output). In literature often progress ratios between 0.8 and 0.95 are reported. In TIMER most curves have been calibrated in such way that 0.9 is considered to be reasonable default value. For natural gas the historic progress ratio is set at different values for different regions - but tend to converge around 0.9. Progress ratios can vary in time.
		A1B	Default (0.90)
		A1F	Fast (0.86)
		A1T, A2, B1	Default (0.90)
		B2	Default (0.90)
Technology learning rate for biofuels
	As for other technologies, in TIMER technology development in terms of the production costs for biofuels is modelled by a 'learning curve'. The actual rate of learning is a function of initial settings (assumptions about cumulative output before starting year of the simulation) and the progress ratio (1- reduction of costs for a doubling of cumulative output). In literature often progress ratios between 0.8 and 0.95 are reported. In TIMER most curves have been calibrated in such way that 0.9 is considered to be reasonable default value. For biofuels the historic progress ratio is around 0.88. Progress ratios can vary in time.
		A1B	Strong till 2040 (0.87), default from 2040 onwards (0.90)
		A1F	Slow to very slow (0.92-0.96)
		A1T	Strong till 2040 (0.86), default from 2040 onwards (0.90)
		A2	Strong till 2020 (0.89), slow from 2020 to 2100 (0.96)
		B1	Strong till 2040 (0.87), default from 2040 onwards (0.90)
		B2	Strong till 2040 (0.87), default from 2040-2060, 2060-2100 slower (0.92)
Technology transfer parameter
	Obviously, technology development is not independent in different regions around the world. In TIMER, this can be simulated by means of the 'technology transfer parameter' that result in transferring the technology progress ratio of the leading region to other regions (catching up). Historically, this parameter is switched off (set at zero).
		A1B	Modest
		A1F	Modest
		A1T	Modest
		A2	Zero
		B1	Modest
		B2	Very low

Assumptions for the electricity supply
Technology learning rate for solar/wind
	As for other technologies, in TIMER technology development in terms of the production costs for solar/wind is modelled by a 'learning curve'. The actual rate of learning is a function of initial settings (assumptions about cumulative output before starting year of the simulation) and the progress ratio (1- reduction of costs for a doubling of cumulative output). In literature often progress ratios between 0.8 and 0.95 are reported. In TIMER most curves have been calibrated in such way that 0.9 is considered to be reasonable default value. For solar/wind the historic progress ratio is set at 0.85. Progress ratios can vary in time.
		A1B	Very strong till 2050 (0.85); default between 2050-2100 (0.90)
		A1F	Slow to very slow (0.95 from 2050 onwards)
		A1T	Very strong till 2050 (0.82); very strong between 2050-2100 (0.85)
		A2	Slow (0.93 from 2050 onwards)
		B1	Very strong till 2050 (0.85); default between 2050-2100 (0.90)
		B2	Very strong till 2050 (0.85) although low income regions have some delay; default between 2050-2100 (0.91)
Technology learning rate for nuclear
	As for other technologies, in TIMER technology development in terms of the production costs for nuclear power is modelled by a 'learning curve'. The actual rate of learning is a function of initial settings (assumptions about cumulative output before starting year of the simulation) and the progress ratio (1- reduction of costs for a doubling of cumulative output). In literature often progress ratios between 0.8 and 0.95 are reported. In TIMER most curves have been calibrated in such way that 0.9 is considered to be reasonable default value. For nuclear power the historic progress ratio is set at average at 1.0 (as additional environmental and risk related conditions for nuclear power have offset costs reductions due to improved technology). Progress ratios can vary in time.
		A1B	Moderate (0.93)
		A1F	Slow (0.97)
		A1T	Moderate (0.93)
		A2	Slow (0.96 - 0.97)
		B1	Slow (0.95)
		B2	Slow (0.96)
Use of hydropower
	The use hydropower in TIMER is determined by two exogenous scenario files: 1) the share of the technical potential per region actually used for hydropower capacity and 2) the load factor of this capacity.
		A1B	Relatively high (40% of global potential)
		A1F	Default (35% of global potential)
		A1T	Relatively high (40% of global potential)
		A2	Default (35% of global potential)
		B1	Default (35% of global potential)
		B2	Default (35% of global potential)
Premium values within thermal fuel production
	The use of the different fuel types for electricity generation in thermal power plants is determined by their relative prices, the efficiency with which these fuels can be used and an additional premium value. These premium values capture other factors that determine market shares of fuels, in particular environmental policies and strategic considerations. If no preference or aversion for a certain fuel type exists, the premium value is set at 1. In most TIMER scenarios in particular premium values for are raised in certain scenarios to reflect environmental considerations.
		A1B, A1F	Very small aversion from coal, simulating small cost increase due to add-on technology
		A1T	Modest aversion from coal, simulating small cost increase due to add-on technology
		A2	Regionally determined preference or aversion from coal, oil and natural gas based on 1995 situation and slow convergence to unity
		B1	Very strong aversion from use of coal for environmental reasons (1.7)
		B2	Regionally determined aversion from coal, based on 1995 situation, ranging from 1.3 to 8.0
Premium values for different electricity production forms
	The use of the different forms electricity generation (thermal power plants, nuclear, renewables) is determined their relative prices and an additional premium value. These premium values capture other factors that determine market shares of fuels, in particular environmental policies and strategic considerations. If no preference or aversion for a certain fuel type exists, the premium value is set at 1. In most TIMER scenarios in particular premium values for are raised in certain scenarios to reflect environmental considerations.
		A1B	In 2100 almost indifferent
		A1F	In 2100 indifferent
		A1T, A2	In 2100 almost indifferent
		B1, B2	In 2100 preference for renewable electricity production; modest aversion towards nuclear; strong aversion towards fossil
Thermal efficiency electricity generation
	An important parameter in thermal electricity generation is the efficiency by which primary fuels are converted into electricity. These efficiencies are a function of time, region and fuel type. The efficiencies can change over time, among others as result of technology development, additional environmental technologies etc.
		A1B	Increases to 0.47-0.49 for coal, 0.51-0.54 for oil and 0.56-0.58 for natural gas
		A1F, A1T	Increases to 0.48-0.50 for coal, 0.52-0.55 for oil and 0.57-0.59 for natural gas
		A2	Increases to 0.43-0.47 for coal, 047-0.52 for oil and 0.51-0.56 for natural gas
		B1, B2	Increases to 0.44-0.48 for coal, 0.49-0.53 for oil and 0.53-0.57 for natural gas

Assumptions for the energy trade
Trade constraints
	All regions in TIMER can decide to import fossil fuels and biofuels from other region, in first instance based on differences in production and transportation costs. In addition, various forms of trade constraints exist. First of all, import and export constraints per region can simply prevent trade, even if costs consideration would indicate trade to be an attractive option. Secondly, region can favour domestic production by means of economic subsidies. Finally, a matrix between the different regions exist that can make trade between two specific regions less attractive. This matrix can first of all incorporate certain geographical difficulties for trade (e.g. the mountain ranges), but can also be used to introduce regional blocks.
		A1B, A1F, A1T	All import and exports constraints are removed
		A2	Regional blocks introduced by assuming trade barriers
		B1	All import and exports constraints are removed
		B2	Regional blocks introduced by assuming trade barriers
Transport costs
	The trade between regions is among other a function of transport costs. Transport costs on their turn are function of a regional distance matrix and transport costs per kilometre. The latter are an exogenously set model input.
		A1B, A1F, A1T	Decrease for natural gas; constant for other fuels
		A2	Constant for all fuels
		B1	Decrease for natural gas; constant for other fuels
		B2	Constant for all fuels