Add Useful Energy variables to the template

[jkikstra]

As discussed in energy breakout of ScenarioMIP today.
Ping amongst others @gunnar-pik @vruijven

Requirements:

• Complete set
• Key sectors for assessing efficiency
• Clear guidance on how to report / what unit.

Should include at least:

• Useful Energy|Transportation
• Useful Energy|Residential and Commercial
• Useful Energy|Industry

Probably better full list: directly the same as the Final Energy tree (which is not having the full 'components' list currently): https://github.com/IAMconsortium/common-definitions/blob/main/definitions/variable/energy/final-energy.yaml

[christophbertram]

What about Non-Energy Use and Bunkers? Probably we can leave them out. For the 3 sectors you mention above, the most important question would be whether or not to report in fuel-equivalent or not...? And I guess we do not need the subcomponents, or do we?

[jkikstra]

What about Non-Energy Use and Bunkers? Probably we can leave them out. For the 3 sectors you mention above, the most important question would be whether or not to report in fuel-equivalent or not...? And I guess we do not need the subcomponents, or do we?

Hmm yes that's the question, we need to have a comprehensive set of reporting. E.g., Agriculture, if reported separately under Final Energy, should it come under "Residential and Commercial" or "Industry"?

On 'whether or not to report in fuel-equivalent'; probably yes, but not sure, need to continue discussion with MESSAGE colleagues

[jkikstra]

So OK, my proposal for now is this.

Maybe @OFR-IIASA and/or @khaeru weigh in if you would think that there is a better way?

We follow:
(a) the sectoral structure of Final Energy (only top level), and adopt its definitions
(b) report in fuel-equivalent (which we add to the variable description, or the note)

variable	description	unit	tier
Useful Energy	Useful energy consumption by all end-use sectors and all fuels, including non-energy use, excluding transmission/distribution losses	EJ/yr	1
Useful Energy (w/o bunkers)	Useful energy consumption by all end-use sectors and all fuels including non-energy use, excluding international aviation and shipping (see 'Useful Energy|Bunkers|*') and excluding transmission/distribution losses	EJ/yr	2
Useful Energy|Agriculture	Useful energy consumption by the agriculture sector including fishing	EJ/yr	1
Useful Energy|Bunkers	Useful energy consumption by the international aviation and shipping	EJ/yr	1
Useful Energy|Carbon Management	Total energy use for carbon management, i.e., capture and/or removal of CO2	EJ/yr	2
Useful Energy|Commercial	Useful energy consumption by the commercial sector	EJ/yr	1
Useful Energy|Industry	Useful energy consumption by the industrial sector excluding non-energy use (e.g.feedstocks)	EJ/yr	1
Useful Energy|Non-Energy Use	Useful energy consumption in non-combustion processes	EJ/yr	2
Useful Energy|Other Sector	Useful energy consumption by the other sectors	EJ/yr	2
Useful Energy|Residential	Useful energy consumption by the residential sector	EJ/yr	1
Useful Energy|Residential and Commercial	Useful energy consumption by the residential and commercial sector	EJ/yr	1
Useful Energy|Transportation	Useful energy consumption by the transportation sector excluding international aviation and shipping (see 'Bunkers')	EJ/yr	1
Useful Energy|Transportation (w/ bunkers)	Useful energy consumption by the transportation sector including international aviation and shipping	EJ/yr	1

[khaeru]

As far as I can see you have at least the following:

#	MEASURE	SECTOR	SUBSECTOR	VARIABLE
1	Useful Energy	_T	_T	"Useful Energy"
2	Useful Energy	_T	_T1	"Useful Energy (w/o bunkers)"
3	Useful Energy	_T	_T2	"Useful Energy|Bunkers"
4	Useful Energy	TRAN	_T1	"Useful Energy|Transportation"
5	Useful Energy	TRAN	_T	"Useful Energy|Transportation (w/ bunkers)"

…wherein that the code list for SECTOR has the items, inter alia:

• TRAN (or whatever) "Transportation"
• _T "Total" = sum over all other labels

…and the code list for SUBSECTOR (or whatever ID) has, inter alia:

• "International Aviation"
• "Shipping"
• _T2 "Bunkers" = "International Aviation" + "Shipping".
• _T "Total" = sum of all other items, inclusive of both transport sub-sector and other sub-sectors.
• _T1 "Total ex bunkers" = _T - _T2.

In terms of coherence, some notes:

• The relationship of (1) and (2) is opposite the relationship of (4) and (5): in the first case the "plain"/shorter variable name is associated with SUBSECTOR=_T, whereas SUBSECTOR=_T1 is indicated by a parenthetical; in the second case it is the opposite.
• (2) is indicated by a parenthetical and (3) by a pipe character, even though the only distinction is in the same SUBSECTOR dimension.

I usually find writing out the keys in their entire dimensionality and then constructing the variable names from those leads to less confusion. However, if these are shadowing other variable names that also have the same issues, then I guess there is a choice between coherence and consistency.

Either way it's always good to record the full key as an annotation to the variable name.

[jkikstra]

As an additional comment, I just noticed that under "sdg.yaml" we also already have a few per-capita variables:

• Useful Energy|Transportation|Passenger [per capita]:
  description: Useful energy per capita for passenger transport
  unit: GJ/cap/yr
  sdg: 7
  weight: Population
• Useful Energy|Industry [per capita]:
  description: Useful energy per capita for industry
  unit: GJ/cap/yr
  sdg: 7
  weight: Population
• Useful Energy|Residential and Commercial [per capita]:
  description: Useful energy per capita for buildings
  unit: GJ/cap/yr
  sdg: 7
  weight: Population
• Useful Energy|Residential [per capita]:
  description: Useful energy per capita for buildings
  unit: GJ/cap/yr
  sdg: 7
  weight: Population
• Useful Energy|Commercial [per capita]:
  description: Useful energy per capita for buildings
  unit: GJ/cap/yr
  sdg: 7
  weight: Population

[jkikstra]

Thanks for the comment with your considerations, @khaeru

Indeed, I agree that the Final Energy tree is problematic; it should can be more clear in the variable naming (as you noted regarding bunkers) and should still specify components.

I had no interest to open that now, hence my suggestion for consistency (over coherence).

However, happy to hear other opinions still, if someone:

• wants to overhaul the final energy tree
• thinks it's better to discard the consistency with the final energy variables, and create a different, fully coherent, useful energy variable set

Note that, if we want reporting from IAM teams by March 1 (not sure that this is critical, but would be nice), we should make the decision and implement by Monday.
Otherwise, there is a bit more time, but teams may not be super happy about changes to the variable template in march/april

[vruijven]

Thanks for getting this started, @jkikstra. Few thoughts from my side

• Useful energy is an artificial construct that is not consistently defined, it is native to some models and alien to others. Hence, I would advice not to go overboard on detailed sector definitions and only assign Tier 1 to the small set of variables that you really need (Useful Energy|Industry , Useful Energy|Transportation and Useful Energy|Residential and Commercial)
• For several variables proposed here it is useful to define them now to have a placeholder once they should be collected in the future or for niche applications, but if they are not used in main calculations, they can be Tier 3 for now (Useful Energy|Agriculture, Useful Energy|Bunkers, Useful Energy|Other Sector)
• For some proposed variables I fail to understand their meaning or use case. What is the use case of Useful Energy|Carbon Management on top of its equivalent final energy variable and the amount of carbon captured? Or how is Useful Energy|Non-Energy Use different from Final Energy|Non-Energy Use?
• When thinking from the perspective of use cases, the more granular and directly related to service indicators the subsectors become. the more you can work directly with final energy intensity of services as expression for efficiency improvement (e.g. GJ/m2 or GJ/pkm) without the need for a useful energy variable.
• Fuel-equivalent should be defined such that the useful energy reflects model-endogenous efficiency improvements over time, which limits the options for calculating this ex-post (for example, in the table below, final to useful conversion efficiency increases from 20% to 25% between 2020 and 2050, but useful energy should be calculated such that it remains at 100 EJ/yr and does not follow the trend of reduction in final energy.

<style> </style>

	Useful Energy	Efficiency	Final Energy
2020	100 EJ/yr	20%	500 EJ/yr
2050	100 EJ/yr	25%	400 EJ/yr

[jkikstra]

Quick one on this:

For some proposed variables I fail to understand their meaning or use case. What is the use case of Useful Energy|Carbon Management on top of its equivalent final energy variable and the amount of carbon captured? Or how is Useful Energy|Non-Energy Use different from Final Energy|Non-Energy Use?

Only there for if they are needed for completeness/comprehensiveness - to not "miss" some of total UE if it is covered in total FE.
I think this depends on what the final energy tree looks like. Which variables need to be reported to be complete?

[vruijven]

Guess that is a matter of taste or preference, to either strive for a comprehensive set or to avoid meaningless variables. I think striving for comprehensiveness in Useful Energy doesn't work, because it is a too artificial construct.

On my final point above of how to calculate Fuel-equivalent; this is actually more diverse than I thought, because different teams and people define useful energy differently. If you define useful energy in transportation as the energy delivered to provide the actual service (moving people and/or goods from A to B), the changes in conversion efficiency should not only include improvements of the technology itself (e.g. the 20 to 25% in my example) but also in delivering the actual service (e.g. from one to four people in a car reflects an improvement of factor four in final to useful energy conversion).

(diving deeper, I find that: the conversion of final energy in end-use devices such as appliances, machines, and vehicles into useful energy such as the energy forms of kinetic energy or heat. Useful energy is measured at the crankshaft of an automobile engine, by the mechanical energy delivered by an industrial electric motor, by the heat of a household radiator or an industrial boiler, or by the luminosity of a light bulb. The application of useful energy provides energy services such as a moving vehicle (mobility), a warm room (thermal comfort), process heat (for materials manufacturing), or light (illumination) Hence, this does not include service provision efficiency.
https://previous.iiasa.ac.at/web/home/research/Flagship-Projects/Global-Energy-Assessment/GEA_Chapter1_primer_hires.pdf

Either way, collecting these variables is an important start, but probably also the beginning of a painful and confusing trajectory of harmonizing the way to calculate the values.