ΔE/GDP ofSegments

ExtractionΔE/GDP

PipelinesΔE/GDP

RefineriesΔE/GDP

-0.2-0.10.00.10.20.30.40.50.6

TotalIndustry

CompositionΔE/GDP ofSegments

ΔE/GDP

(MJ/

2002

$)

0.00.10.20.30.40.50.6

ExtractionΔE/GDP ΔEo/GDP ΔE/Eo

ΔE/GDP

(MJ/

2002

$)

-0.2-0.10.00.10.2

PipelinesΔE/GDP ΔEo/GDP ΔE/Eo

RefineriesΔE/GDP ΔEo/GDP ΔE/Eo

3B. Oil & gas showed an increase in E/GDP due to higherenergy costs for extraction. Pipeline and refinerysegments showed efficiency improvements:

Pipelines RefineriesExtraction

-0.6

3 -0.4

1

0.40

-0.2

5

-0.2

3

-0.1

3

-0.0

1

-0.2

0

0.04

-0.1

7

-0.8

-0.4

0.0

0.4

Total C & I Oil andGas

EnergyIntensiveIndustry

OtherIndustry

ElectricPower

FreightTransport

Resid-ential

PersonalTransport

Impa

ct o

n to

talΔ

E/GD

P(M

J/20

02$)

3. A net decrease in energy intensity (increase in energyefficiency) of Canada’s productive and household economies.

Note that large differences exist among sectors in thecontribution to net energy intensity…

…and within each sector, large differences exist in thefactors contributing to energy intensity change, as shownbelow.

- METHODS -

- ACKNOWLEDGEMENTS

- -

- RESULTS - - DISCUSSION / CONCLUSION-- INTRODUCTION -

-2.63

-0.63

-0.17

-0.62

-1.22

-3

-2

-1

0Total

Δ SectorEnergy

Intensities

Δ Per CapitaPersonalEnergy

ΔProductivity(GDP/capita)

Δ GDPComposition

by Sector

Chan

ge in

tota

l E/G

DP (M

J/20

02$)

0

2

4

6

8

10

12

14

0.6

1.0

1.4

1.8

1983 1993 2003 2013

Axis

Titl

e

Rela

tive

Valu

e(1

983

= 1.

0)

8

10

12

14

0.0

1.0

2.0

3.0

1983 1993 2003 2013

E/GD

P (M

J/20

02$)

Rela

tive

Valu

e(1

983

= 1.

0)

1995

-2.63 MJ/2002$(~23% reduction)

Why did Canada’sGDP and energyuse (E) diverge

between1995 & 2010?

2. Loss in manufacturing more than compensated for by anincrease in the Commercial & Institutional (C&I) sectors.

GDP/Capita

Population

GHG/E

E/GDP

GDP

E

E/GDP

-0.4

-0.2

0.0

0.2

0.4

0.6

O&

G ex

trac

tion

prod

uct m

ixFr

eigh

t tra

nspo

rt m

ode

shar

ePe

rson

al tr

ansp

ort P

KT/c

apita

O&

G ex

trac

tion

E/Eo

Min

ing

E/GD

PRe

siden

tial f

loor

are

a pe

r dw

ellin

gPe

rson

al tr

ansp

ort m

ode

shar

eRe

siden

tial h

ouse

hold

size

(peo

ple/

HH)

O&

G pi

pelin

e Eo

/GDP

O&

G in

dust

ry c

ompo

sitio

nO

&G

extr

actio

n Eo

/GDP

C&I b

uild

ing

type

shar

eRe

siden

tial d

wel

ling

type

shar

ePo

wer

gen

erat

ion

E/Eo

by

type

O&

G re

finer

y Eo

/GDP

O&

G pi

pelin

e pr

oduc

t mix

Ener

gy in

tens

ive

indu

stry

com

posit

ion

Chem

ical

s E/G

DPCe

men

t E/G

DPPo

wer

gen

erat

ion

mix

O&

G re

finer

y E/

EoFr

eigh

t tra

nspo

rt T

KT/G

DPPo

wer

gen

erat

ion

Eo/G

DPSm

eltin

g E/

GDP

Pulp

& p

aper

E/G

DPFr

eigh

t tra

nspo

rt E

/TKT

by

mod

eO

&G

pipe

line

E/Eo

by

prod

uct

C&I E

/flo

or a

rea

by b

uild

ing

type

Pers

onal

tran

spor

t E/P

KT b

y m

ode

Oth

er in

dust

ry E

/GDP

C&I f

loor

are

a/GD

PRe

siden

tial E

/flo

or a

rea

Impa

ct o

n to

talΔ

E/GD

P (M

J/20

02$)

0200400600800

1,000

GDP

(200

2$)

0

50

100

Ener

gyIn

tens

ity(M

J/20

02$)

Oil &Gas

PowerGen’n

EnergyIntensiveIndustry

FreightTransport C & I

OtherIndustry

A

B

Why did Canada’s GDP and Energy Use Diverge after 1995?Ralph Torrie1,2, Chris Stone1, David B. Layzell1,3

1 Canadian Energy Systems Analysis Research (CESAR) Init., U Calgary, AB; 2Torrie Smith Assoc. Coburg, ON; 3 Correspondance: dlayzell@ucalgary.ca

-0.5-0.4-0.3-0.2-0.10.0

TotalΔFloor

area/GDP ΔE/floor area

ΔE/GDP

(MJ/

2002

$)

3A. The C&I sector required less floor area per $GDP andhad lower energy use/m2, due to more efficient buildings.

ΔE/floor areaΔBuilding

mix

ΔE/floorarea by

building type

-0.21-0.13 -0.10

-0.04 -0.04

0.09

ΔE/GDPof

SegmentsPulp &Paper

Smelting&

Refining Cement Chemicals Mining

-0.3-0.2-0.10.00.1

TotalIndustry

compositionΔE/GDP ofSegments

ΔE/GDP

(MJ/

2002

$)

3C. The energy intensive industries all decreased intheir energy use/GDP except for mining:

3G. Personal transport increases in person-km travelled/yr(i.e. higher energy intensity) were largely offset by vehicleefficiency improvements.

-0.2-0.10.00.10.2

Total ΔPKT/capita ΔE/PKT

ΔE/GDP

(MJ/

2002

$)

ΔE/PKT ΔMode share ΔE/PKT by mode

3F. Residence sector had lower energy intensities due to lowerenergy use/m2, despite larger dwelling sizes:

-0.4-0.3-0.2-0.10.00.1

TotalΔPeople perhousehold

ΔE perhousehold

ΔE/GDP

(MJ/

2002

$)

ΔE perhousehold ΔHousing Mix ΔDwelling size ΔE/floor area

3E. Freight transport reductions in energy intensity(tonne(t)-km/GDP) were offset by more energy use per tshipped due to mode share change (more truck, less traintransport):

-0.2-0.10.00.10.20.3

Total ΔTKT/GDP ΔE/TKT

ΔE/G

DP(M

J/20

02$)

ΔE/TKT ΔMode share ΔE/TKT by mode

3D. Power generation showed a decrease in energy intensitydue to improved efficiency & a shift in the generation mix.

-0.2-0.10.00.1

Total ΔEo/GDP ΔE/Eo

ΔE/G

DP(M

J/20

02$)

ΔE/EoChange in

generation mixNet

efficiency

Growth in the national economies (measured as Gross DomesticProduct, GDP) are often correlated with increases in both fuel andelectrical energy use (E) and greenhouse gas (GHG) emissions. Thiswas the case in Canada between 1983-1995 (Fig. 1).

However, uncoupling GDP from E and GHG emissions can make animportant contribution to climate change mitigation.

Since 1995, government data on Canada’s energy systems hasshown a divergence in GDP and E (Fig 1B). We want to know Why.

Fig. 1. Changes in key parameters of Canada’s Energy Systems

- REFERENCES -

C&I3A

Oil &Gas3B Energy

IntensiveIndustries

3C

PowerGen3D

FreightTransp.

3E Resid-ential3F

Pers.Transp.

3G

[1] CanESS v6 from whatIf? Technologies Inc, Ottawa, ON[2] B.W. Ang, 2005 “The LMDI approach to decomposition analysis: a practicalguide”, Energy Policy 33 (2005) 867-871

We thank whatIf? Technologies Inc (Ottawa, ON) for allowing us to accessand use the extensive database resources that they have incorporated intotheir Canadian Energy Systems Simulator (CanESS) model.

Fig. 3. Changes (1995-2010) in energy intensity & GDP of Sectors in Canadian Economy

1. Richer Canadiansspend less of new$ on energyservices.

Fig. 5. Changes in ΔE/GDP of the C&I sectors between 1995 & 2010

Fig. 6. Changes in ΔE/GDP of the oil & gas sector between 1995 & 2010

Fig. 7. Changes in ΔE/GDP of the energy intensive sectors between 1995 & 2010 Fig. 11. Changes in ΔE/GDP of Personal transport between 1995 & 2010

Fig. 10. Changes in ΔE/GDP of Residential buildings between 1995 & 2010

Fig. 9. Changes in ΔE/GDP of the Freight Transport sector between 1995 & 2010

Fig. 8. Changes in ΔE/GDP of the Power Gen sector between 1995 & 2010

Fig. 2. Factors contributing to ΔE/GDP

Fig. 12. Contribution to the energy intensity (ΔE/GDP) of the Cdn economy (1995-2010from 32 underlying trends affecting the productive & household sectors.

Structural Changes in the Canadian economy (i.e. an increase role forCommercial & Institutional sectors that use little energy per GDP$)accounted for 46% of the economy’s observed decrease in the energyintensity (MJ//2002$) between 1995 and 2010.

An addition 24% of the observed decrease in energy intensity wasattributed to an increase in the wealth of Canadians who don’t spendtheir additional resources on energy services, at least in Canada.

The remaining 30% of the energy intensity change was attributed to anet improvement in energy efficiency in Canada’s productive andhousehold economies.We deconstructed the components of energy use and economicchange between 1995 & 2010 and found that some trends increasedand some decreased ΔE/GDP as shown in Fig. 12.

Most data for these analyses were obtained from the Canadian EnergySystems Simulator (CanESS) model [1] which integrates large amounts ofgovernment database resources. Energy use was separated into thesectors (i) associated with the productive economy (EP) & the householdeconomy (EH):

where Si=GDPi/GDP and is the structural factor (the share of GDPgenerated by sector i) and Ii=Ei/GDPi and is the intensity factor (the energyuse per dollar of value added in sector i).Decompositional analysis was conducted using the Logarithmic MeanDivisia Index (LMDI-I) method [2] which uses the above equations toproduce the following four factors that sum to the total change in energyintensity (∆E/GDP) over the study period (T1=1995 and T2=2010):Economic Structure Factor - changes in E/GDP due to changes inGDPi/GDP:

Productive Economy Intensity Factor - changes in E/GDP due tochanges in Ei/GDPi:

Personal Productivity Impact - changes in E/GDP due to changes inGDP/capita:

Household Per Capita Energy Intensity - changes in E/GDP due tochanges in per capita residential and personal transportation energy use:

Where

These equations were applied to annual data for the study period, usingannual chained analysis. A similar approach was used for subsequentdecomposition analyses within the various sectors.

1995 2010

SectoralEnergy

intensity

Per capitapersonal

energy use

SectoralComposition

of GDP

OtherInd.

Fig. 4. Breakdown of impacts from sectoral changes on total ΔE/GDP

Total

Trends increasingE/GDP

Trends decreasingE/GDP

This work shows that the observed decoupling of E and GDPbetween 1995 & 2010 was an emergent property of a large numberof changes in the Canadian economy and energy use.The study also identifies key trends in Canada’s energy systems thatcontribute to increases or decreases in ΔE/GDP. Some of thesecould provide policy levers to take action on climate change.

E

Productivity(GDP/capita)