Energy Studies Review

| Issue 2Volume 8 Article 2

10-31-1997

GHG Abatement in Central Canada withInterprovincial CooperationRichard Loulou

Amit Kanudia

Denis Lavigne

This Article is brought to you by DigitalCommons@McMaster. It has been accepted for inclusion in Energy Studies Review by an authorizedadministrator of DigitalCommons@McMaster. For more information, please contact esr@mcmaster.ca.

Recommended CitationLoulou, Richard; Kanudia, Amit; and Lavigne, Denis (1996) "GHG Abatement in Central Canada with Interprovincial Cooperation,"Energy Studies Review: Vol. 8: Iss. 2, Article 2.Available at: http://digitalcommons.mcmaster.ca/esr/vol8/iss2/2

GHG Abatement in Central Canada with Interprovincial Cooperation

AbstractThis paper reports on the use of an advanced multi-reg/ on, bottom-up model (Extended MARKAL) for anin-depth investigation of the responses by the Quebec-Ontario energy system to a series of increasingly severegreenhouse gas (GHG) emissions reduction targets. For each target, the responses are analyzed under fourpolicies resulting from the adoption (or not) of a joint emissions target and of electricity exchanges. Resultsindicate significant cost ,savings and a reduction in the need for nuclear energy in Ontario, which suggeststhat cooperative responses to GHG emissions caps should be seriously considered by the two provinces.

This article is available in Energy Studies Review: http://digitalcommons.mcmaster.ca/esr/vol8/iss2/2

This paper reports on the use of an advanced multi-re­g/on, bottom-up model (Extended MARKAL) for an in­depth investigation of the responses by the Quebec-On­tario energy system to a series of increasingly severegreenhouse gas (GHG) emissions reduction targets. Foreach target, the responses are analyzed under four poli­cies resulting from the adoption (or not) of a joint emis­sions target and ofelectricity exchanges. Results indicatesignificant cost ,savings and a reduction in the need fornuclear energy in Ontario, which suggests that coopera­tive responses to GHG emissions caps should be seriouslyconsidered by the two provinces.

Cet article utilise Ie modele multi-regional ExtendedMARKAL pour explorer en detail la reponse de La regionQuebec-Ontario a l'imposition de contraintes plus oumoins severes sur l'emission de gaz aelfet de serre (GES)par cette region. Pour ehaque eibIe de reduction, Lareponse est analysee sous quatre politiques eontrastees,chacune incluant ou excluant l'echange d'eleetrieite etl'echange de permis d'emission entre les deux provinces.Les resultats indiquent des economies substantielles surIe coiit total des reductions de GES, lorsque Ies deuxprovinces cooperent. De plus, la cooperation diminuefortement Ie besoin pour l'Ontario de developper sa ca­pacite nucleaire, lorsqu'une cible de reduction est im­posee. Ces resultats soulignent l'importanee de Ia co­operation interprovinciale future si des quotas d'emissionde GES etaient imposes.

This research was supported by EnvironmentCanada, NSERC (Canada), and FCAR (Quebec).

120

GHG Abatement inCentral Canada withInterprovincialCooperationRICHARD LOULOU,AMIT KANUDIA andDENIS LA VIGNE

1. Introduction

Canada is a signatory to the UNFCCC (UnitedNations Framework Convention on ClimateChange 1992), and has been an active player inthe subsequent meetings of the Conference ofParties (COP), a political-level group of na­tional delegates from all major United Nationscountries charged with devising global policieson the climate change issue. As an activemember of the COP, Canada has put green­house gas (GHG) emissions targets on its ownpolitical agenda, consisting of a return to the1990 level of emissions by year 2000, and asubsequent 20% reduction of Canadian GHGemissions by 2010. Whereas it appears that the2000 Canadian emissions stabilization will notmaterialize, other studies like those by the In­tergovernmental Panel on Climate Change(IPCC 1995 a, b, c) are indicating a high degreeof uncertainty in the setting of globaIly desir­able (optimal) reduction levels by OECD coun­tries. The range of possible targets is quitewide, leaving each country in a quandary as tothe setting of national policies. Faced with un­certainty, Canada would benefit from a thor­ough investigation of contingent plans to meeta wide range of targets.

Energy StudieS Review Vol. 8, No.2, 1996 Printed in Canada

Loulou et al.: GHG Abatement in Central Canada with Interprovincial Cooperation

Produced by The Berkeley Electronic Press, 1996

In this study, we focus on the analysis of acoordinated response to alternative GHG re­duction targets by Central Canada (i.e., the twopopulous Canadian provinces, Quebec andOntario). Thus defined, Central Canada com­prises slightly more than 60% of the Canadianpopulation and GDP. Although this study doesnot claim to be directly applicable to Canada asa whole, it provides a useful analysis of asignificant fraction of the country, andfurthermore, serves as an illustration of the po­tential benefits of interprovincial cooperationin dealing with GHG abatement. Since Quebecenjoys a substantial hydroelectric potential, itis of interest to examine how this could beused efficiently to reduce GHG emissions inCentral Canada as a whole.

In order to reflect the high degree of uncer­tainty regarding the amount and the timing offuture GHG abatement, we chose to examinefive alternative targets, viz. 0%,10%,20%,30%,and 40% cumulative emissions reductionscompared to the 1990 level. In addition, a base­case scenario with no emissions cap is also in­cluded in the analysis. Each reduction scenariois run four times, with each run assuming acombination of the following choices: with/without electricity exchanges between the twoprovinces, and with/without emissions trad­ing between the two provinces. In this way, wecan identify the potential benefits from coop­eration along the two dimensions of electricitytrading and of emissions trading.

In section 2, the methodology is outlined,and in section 3, some key results and analysesare developed. In section 4, we conclude thispaper and indicate further avenues for re­search and development.

2. Methodology

2.1 The Model

Over the years, we have developed twoMARKAL activity analysis models of the en­ergy/ industrial systems of the Quebec andOntario provinces. The most recent versions ofthese models include descriptions of very di­versified energy sources (extraction, imports),of energy transformation and distribution, and

of end-use processes and devices in all eco­nomic sectors, including a set of technologicaland energy conservation options in eachprovince. These elements of the energy systemare referred to as technologies in the MARKALjargon. Each provincial model has in excess of500 technologies (Loulou and Waaub 1992).MARKAL (Fishbone et al. 1981, 1983) is a de­mand-driven model based on the minimizationof the long-term discounted cost of a completeenergy system, including the production,conversion, distribution, and final use ofenergy forms. In MARKAL, each element ofthe energy system (such as a technology, afuel, or a conservation measure) is expliCitlyrepresented by a set of model variables, in­dexed by a time period. The model covers nineperiods of five years each - a 45-year horizon.The nine periods are centered on years 1995 to2035, so that the actual horizon covered is from1993 to 2037, inclUSively.

The model's engine is linear programming,which, by minimizing total discounted systemcost, in effect computes a partial equilibriumsolution for energy markets. The base de­mands for a large number of sectors and sub­sectors are specified exogenously for each sce­nario. MARKAL determines the values of allfuture investments and operations levels of thetechnologies at each time period, While re­specting emissions caps and a very detailed setof technical and logical constraints. In additionto technological and energy substitutions,MARKAL may also choose to adjust base de­mand levels endogenously, thanks to a set ofown-price elasticities. For example, whenemissions caps are imposed, the implicit pricesof some energy forms, and ultimately those ofsome economic goods and services, increase:MARKAL then has the option of reducingsome demand levels according to the specifieddemand curves. While the main outputs fromMARKAL are the values of the investmentsand capacities of the various technologies at alltime periods, additional output consists of theset of implicit prices of each energy form andof each demand category.

Each model's database has undergone in­depth revisions during the last two years. Inthe process, the set of GHGs modelled was in-

121

Energy Studies Review, Vol. 8 [1996], Iss. 2, Art. 2

http://digitalcommons.mcmaster.ca/esr/vol8/iss2/2

MARKAL

Table 1: Energy Carriers, Technologies, andDemand Segments in MARKAL Models

yearly GDP real growth of 2.1 % per year inQuebec and 2.4% in Ontario, until 2020. Thesegrowth rates are reduced by 0.4% per year(again in real terms) after 2020. The prices ofimported oil and natural gas grow by an aver­age of 1% (in real terms) per year until 2011,and then stagnate at their 2011 levels. Gasprices converge to oil prices by year 2005, andthe two remain equal thereafter. BecauseMARKAL is an integrated energy model thatmodels the energy supply, the prices of do­mestic energy forms such as electricity, bio­mass, and refined petroleum products are de­termined endogenously and are equal to theshadow prices of these energy forms in themodel. Finally, the real discount rate used is5% per year. All cost figures discussed in thisarticle are expressed in constant 1990 Canadiandollars.

As mentioned earlier, we adopted six alter­native emissions caps: the base case and fivelevels of cumulative GHG reductions equal to0%, 10%, 20%, 30%, and 40% of 1990 GHGemissions. For each reduction level, four runs

Quebec Ontario

131112

76

49

173623

89

1046768

1806

590

189

131412264

69

16362362

1067769

1244

517

184Energy Carriers

TechnologiesSupply Sources of Energy

Power GenerationOil RefiningOther EnergyProcessing

End-use ResidentialCommercialTransportationIndustryNon-energy Uses

End-Use Demand Segments

ResidentialCommercialTransportationIndustryNon-energy Uses

creased to include carbon dioxide (C02)'methane (CH4), and polychlorofluorides (CxF).In addition, emissions of acid gases (S02 andNOx) are also modelled. The technology setsare carefully designed to include all major ex­isting and new technologies, with special em­phasis on options with low or null GHG emis­sions, such as renewables. Following the cur­rent thinking on nuclear power, it has been as­sumed that there will be a de facto moratoriumon any decision to invest in new nuclear powerplants for another decade. Further, nuclearplants take about 10 years to build. Thus, theearliest that the models can set up new nuclearcapacity is in 2018.

As an indication of the level of detail in thetwo models, Table 1 provides a count of thedifferent energy carriers in the models, of thetechnologies present in each of the main sec­tors of the energy system, as well as the num­ber of separate demand segments in each end­use sector. Thus, MARKAL-Quebec has 184energy carriers, including imported energy, lo­cally produced primary energy forms, and allsecondary energy forms. The second part ofTable 1 indicates the richness of technologicaldetail in each MARKAL model. As an exam­ple, the transportation sector has a total of 69technologies (i.e., types of vehicles). Since themodel is U drawn" by demands for economicgoods and services, the third part of Table 1indicates the degree of disaggregation of eachbroad demand sector. For instance, there are 13demand segments in the residential sector,each capturing one homogeneous demand foran energy service (e.g., space heating for pre­1991 single-family dwellings). A detailed list­ing of technology and energy-carrier nameswould be too space consuming for inclusionhere; however, a copy of the database is avail­able upon request.

A Single economic scenario is used throughoutthis study, which corresponds to moderateeconomic growth until year 2020, slowingdown thereafter. The economic growth as­sumptions are close to those assumed in Natu­ral Resources Canada (1994), with an average

2.2 The Scenarios

122

Loulou et al.: GHG Abatement in Central Canada with Interprovincial Cooperation

Produced by The Berkeley Electronic Press, 1996

were performed, corresponding to the follow­ing cooperation policies:• NC-No cooperation (i.e., no electricity trad­

ing, no GHG permit trading);• EE-Electricity exchanges are allowed, but

GHG permit trading are not;• JE-Joint emissions target (i.e., GHG permit

trading is allowed, but electricity exchangesare not); and

• JEEE-Joint emissions target with electricityexchanges (i.e., electricity trading and emis­sion permit trading are both allowed).

To model Central Canada as one unit, weused the recent multi-regional feature of theCanadian MARKAL model (Kanudia andLoulou, 1997), where any number of MARKALmodels can be merged into a single larger-sizemodel. The number of models is limited onlyby the capacity of the numerical solver. Specialvariables are defined to represent the amountsof investment in electricity transport lines, andthe amounts of electricity traded at each periodbetween each pair of models. The size of themerged model reaches about 23,000 rows and40,000 columns, but the model remains quitetractable computationally as long as thenumber of regions remains reasonably small(two in our case).

The Central Canada model may now beused in various ways to simulate the four typesof runs listed above: in the NC run, eachprovince has its own GHG emissions con­straint, and all electricity exchange variablesare set to zero. In EE or JEEE runs, the electric­ity trading variables are left free for the modelto determine. In JE or JEEE runs, there is a sin­gle GHG emissions constraint, which themodel is free to allocate optimally to eachprovince, thus in effect simulating a permit ex­change system (the model also produces themarginal cost of the last tonne of GHG abate­ment, which is also the economic value of theone-tonne permit).

Since sales of electricity from Quebec orOntario to other regions (e.g., New York state)are set exogenously, these remain unaffectedwhenever the model endogenously determineselectricity exchanges between these twoprovinces. Therefore, there is no hidden cost oflost sales attached to increased interprovincial

exchanges.

3. The Impact of Cooperation onGHG Abatement

We shall examine in turn the cost aspects (andin particular the savings induced by tradingelectricity and/or permits), and the impacts onenergy supply and demand in each province.

3.1 The Dividends of Interprovincial Cooperation

What are the advantages ofa joint emissions targetand eIectricih) trading on the cost of abatement inQUlibec and Ontario (as compared to autarchy)?

The model described above was used to per­form four sets (one for each cooperation policyoutlined above) of six runs each (i.e., one un­constrained-emissions run, and five with cu­mulative GHG emissions reductions of 0%,10%, 20%, 30%, and 40%, respectively, whenmeasured with respect to the 1990 level ofemissions). These results were used to con­struct the tradeoff curves shown in Figure 1,where, for each policy, the system's discountedcost is plotted against cumulative GHGreductions. In Figure 1, the vertical axisrepresents the added cost of each scenario,over and above that of the base case (the NCscenario without any emissions control andwithout cooperation). The first observation isthat, under no obligation to reduce emissions,the four policies have almost identical systemcosts, as witnessed by the fact that each of thefour tradeoff curves starts with a zero ordinate.However, the EE and JEEE policies benefitfrom slight emissions reductions. When reduc­tions are imposed, system cost increases at anincreasing rate, as shown by the convexity ofeach tradeoff curve. For all reductions targets,there is evidently significant cost savings whenelectricity exchanges are allowed. However,joint emissions targets have only a marginalimpact on abatement costs. Electricity tradingis shown to result in total discounted savingsranging from $7 billion in the constant-emis­sions case to $10 billion in the 40% reductioncase, whereas in the unconstrained-emissionsscenario, the savings amount to only $0.48 bil-

123

Energy Studies Review, Vol. 8 [1996], Iss. 2, Art. 2

http://digitalcommons.mcmaster.ca/esr/vol8/iss2/2

-X-QCNC

-<>--ONNC

----tr--JEEE ~x

v X

~20 ~xu x,,10~c x"E' 0 41----t----t----t----;1""0 10 20 30 40

Cumulative C02 Emissions Reduction (%)

~

~ SOE'"~1: 40<I:

830u

70'2o~ 60~

140~ --a-NC-c 1200

--O-EE

e 100

'" ---tr-- JE'"'" 80'" "L -x-JEEE

(I":= 60-'"0~u 40

E:/""'"- 20

'" """(J)

o ,,,,""'+--+----+--t--!--f------i01234567

Cumulative Emissions Reduction (billion tonnes)

Figure 1: Cost/Emissions Reduction Tradeoff Curves

Figure 2: Marginal Cost of C02 Abatement

Table 2: Emissions Trading from Ontario to Quebec(million tonnes, cumulative)

sales of GHG permits by Ontario and Quebec'selectricity sales peak for the 20% reductiontarget. The reason is as follows: for moderatereductions, Quebec can afford to let its con­sumption sectors use more natural gas, thusfreeing some of its hydroelectriCity, which ismost useful in Ontario (where the main GHG­free alternative to Quebec's hydro is nuclearpower, an expensive energy source with longlead times). The penetration of natural gas inQuebec's residential sector explains the higherGHG emissions in that province, and hence thepurchase of permits from Ontario. However,when the reduction target is more stringent(30% or 40%), Quebec and Ontario both needGHG-free electricitlj, and it becomes moreadvantageous to use that resource close to itsproduction site so as to avoid transmissionlosses and investments in transmission lines.

Turning now to the JE case, Ontario buysemissions permits from Quebec so that it canshift some of its (expenSive) nuclear powergeneration to gas-based plants. Quebec im-

1/ Constant emissions

0%4.0

-277.1

10

121.5-138.5

20

170.0-98.3

Reduction

30

82.9-107.7

4032.8

-68.8JEEEjE

lion.A more subtle analysis of emISSIons and

electricity trading is made possible by examin­ing the behaviour of electricity exchanges onone hand, and of emissions trading on theother, across all reduction scenarios. Figure 2shows the marginal cost of CO2 abatement(which is the dual value of the cumulativeemissions constraint) under different coopera­tion scenarios, and Table 2 shows the cumula­tive amounts (in million tonnes) of GHG per­mits transferred from Quebec to Ontario, overthe model's 45-year horizon, in scenarios JEand JEEE. From Table 2, one observes that On­tario sells permits to Quebec in the JEEE sce­nario, whereas it buys permits from Quebecunder JE (in other words, Ontario either buyspermits or electricity from Quebec). Note thateven though there is a significant difference inthe marginal abatement costs in the twoprovinces (Figure 2), equilibrium is attainedwith a relatively small volume of emissionstrading (Table 2). Economic theory impliesthat, under an efficient permit exchange sys­tem, the price of a one-tonne GHG permitshould be equal to the marginal cost of abate­ment shown in Figure 2 (note that this article isnot concerned with the question of who paysfor the exchange of permits, since this is highlydependent upon the initial endowments inpollution rights of the two provinces - anethical/political issue not modelled here).

We first analyze the JEEE case, where both

124

Loulou et al.: GHG Abatement in Central Canada with Interprovincial Cooperation

Produced by The Berkeley Electronic Press, 1996

60

50

-+-EE

__ JE

-X-JEEE

........-<:>-- No Nuc NC

-X-JEEE

-----0--- No Nuc JEfE

10

5

25

0; 100aU

E2 sow~

'" o xx

40

E'"~ 30'" -----<>---- No Nuc JEEE~

G20 --D--JEEE

10 -b--NC

§ 250

E-;; 20000

'"Eu '50c

300

o1995 2000 2005 2010 2015 2020 2025 2030 2035

30

35

o+----+--+--t-----4---+-----l---f------1

1995 2000 2005 2010 2015 2020 2025 2030 2035

Year

Figure 5: Hydro Capacity in Quebec with 40%Emissions Reduction

~ 20'"~~ 1SG

01234567

Cumulative Emissions Reduction (billion tonnes)

Figure 4: Cost/Emissions Reduction TradeoffCurves (Nuclear-free Cases)

YearFigure 3: Nuclear Power Generation Capacity inOntario with 40% Emissions Reduction

Ontario's nuclear power capacity increasescontinuously as the emissions reduction growsmore and more stringent. In later periods, theinstalled capacity varies from 3.5 gigawatt(GW) in the unconstrained-emissions scenario,to more than 30 GW in the 40% reduction sce­nario. Figure 3 shows nuclear power capacityunder the 40% reduction targets with the fourdifferent cooperation scenarios. It is evidentthat interprovincial cooperation, besides re­ducing the joint cost of meeting reduction tar­gets, has the supplementary advantage of re­ducing the need for additional nuclear powerin Ontario. Given the immediacy of the debateon this issue, we conducted 10 additional runsto answer the following questions: haw muchmore does it cost to implement emissions reductionswithout investing in nuclear pawer?; and what isthe role ofelectricitlj trading in such a scenario?

The results are plotted as cost!emissionsreduction tradeoff curves in Figure 4. Eachcurve has six points (the base case plus the fivereduction targets). The new results concerntwo new policies: a 'No New Investment inNuclear' constraint under no cooperation (NoNuc NC); and a 'No New Nuclear' under jointemissions and electricity exchanges (No NucJEEE). Results for the NC and JEEE scenarioshave been included for purposes of compari­son.

The main observation is that a nuclearfreeze under NC more than doubles the cost ofmeeting the 40% reduction target (a $170 bil­lion increase over NC, in net discounted costterms). Whereas, under JEEE, the freeze costsjust about $20 billion more. Even for the moremoderate reduction targets, the advantage ofcooperation is very large, as shown by the dis­tance between the No Nuc NC and the No NucJEEE curves in Figure 4. In these cases, thebulk of the electricity used in the twoprovinces is generated by hydro plants inQuebec, as indicated by Figure 5. This is a

plements higher GHG emissions reductions bysubstituting alcohol for oil in the transportsector. This explains the negative signs on thesecond row of Table 2.

3.2 The Impact of Cooperation on Nuclear Capacity

125

Energy Studies Review, Vol. 8 [1996], Iss. 2, Art. 2

http://digitalcommons.mcmaster.ca/esr/vol8/iss2/2

major finding, since it provides an attractivealternative to nuclear in the event of the adop­tion of a GHG emissions target.

3.3 Other Energy/I'echnologJJ Implications

Let us now focus on the energy substitutionsresponsible for all the results documented inthe previous, subsection, Broadly speaking, al­cohol replaces oil in the transport sector in allreduction scenarios, Natural gas emerges as avery competitive option under mild reductionscenarios in both provinces. Under severe re­ductions, there is a heavy penetration of elec­tricity in the residential and commercial heat­ing demand segments,

Both emissions and electricity trading tendto reduce the need for expensive nuclearpower in Ontario, as pointed out earlier.Emissions trading (JE) is used to increase emis­sions in Ontario so that some of the nuclearpower can be replaced by gas-based poweLElectricity exchanges (EE) directly displacegeneration from nuclear plants. In the JEEEscenario, on top of the additional generation inQuebec, more electricity is released for exportfrom Quebec's residential and commercial sec­tors through gas substitution, Higher emis­sions reductions are implemented in Ontarioby delaying gas-based electricity generationand by the increased use of renewable energy,This is why electricity and emissions exportsfrom Quebec to Ontario peak in the same sce­nario (i.e., the one with a 20% emissions reduc­tion),

We now turn to a more detailed analysis ofthe NC scenario for the two provinces, andthen highlight the important impacts of thethree exchange policy scenarios, namely jointemissions target (JE), electricity exchanges(EE), and joint emissions target with electricityexchanges (JEEE) ,

THE NO COOPERATION SCENARIO (NC)

Under NC, the two provinces implement theirreduction targets in very different ways, Que­bec mainly uses substitutions in demand sec­tors, whereas in Ontario the supply sector alsoundergoes significant changes,

126

Alcohol replaces oil in the Quebec trans­port sector and comprises over one-half of thesectoral energy consumption in later periods,under 40% emissions reduction. There is a 10­15% increase in aggregate electricity consump­tion on account of residential and commercialend-use demands, In commercial heating, oil isdisplaced by gas in the mild mitigation scenar­ios, and by electricity in the severe ones, Natu­ral gas.is replaced by electricity for residentialheating in all reduction scenarios.

In Ontario, aggregate electricity consump­tion remains constant, except in the last peri­ods under the two most severe reduction lev­els, However, the electricity generation sectorundergoes profound changes, In the uncon­strained-emissions case, electricity generationis dominated by coal-based power and there isno fresh investment in nuclear power capacity,As GHG reductions are imposed, coal-basedplants disappear immediately and are replacedby natural gas plants and nuclear plants in themild reduction scenarios, The gas-based capac­ity peaks at 12 GW in the constant emissionscase. Nuclear capacity reaches over 30 GW inthe 40% reduction scenario, On the demandside, the transport sector shows the same alco­hol-oil substitution as Quebec, but here alcoholsupplies less than one-half the total energydemand for that sectoL As additional nuclearcapacity is not available in the first half of theplanning horizon, there is a substitution ofnatural gas for electricity in residential andcommercial heating demands, As a result, ag­gregate electricity consumption drops by 10­15%, There is also a significant penetration ofwood for residential heating in the later peri­ods,

EMISSIONS TRADING GE}

In the emissions trading (without electricityexchanges) scenario, there is a marked reduc­tion in nuclear power generation in Ontario, Toachieve this, the emissions restriction inOntario is relaxed by implementing a higherreduction in Quebec In the constant-emissionsscenario, emissions trade amounts to 9.4% ofQuebec's cumulative emissions over the mod­el's planning horizon of 45 years,

Loulou et al.: GHG Abatement in Central Canada with Interprovincial Cooperation

Produced by The Berkeley Electronic Press, 1996

Higher reductions are achieved in Quebecthrough substitution of alcohol for oil in thetransport sector. Ontario uses the relaxedemissions target to replace nuclear power bygas-based power. In the constant-emissionsscenario, generation from nuclear plants al­most assumes the unconstrained-emissionstrajectory (i.e., no fresh investment in nuclearcapacity). Aggregate electricity consumptionremains unaffected in both Ontario andQuebec.

ELECTRICITY EXCHANGES (EE)

When we allow electricity exchanges withoutemissions trading, Quebec sets up additionalcapacity and reduces consumption to exportelectricity to Ontario. Ontario, in turn, reducesits nuclear power generation and increases itsown electricity consumption.

Electricity production in Quebec increasesby up to 18% in the 20% emissions reductionscenario. Further, natural gas displaces elec­tricity for residential and commercial heatingto release 15-20% of aggregate electricity pro­duction for export. Electricity exports to On­tario reach a peak of about 300 petajoules inthe later periods under 20% emissions reduc­tion, which comprises over 40% of aggregateelectricity consumption in Ontario. The substi­tution away from electricity in Quebec resultsin higher emissions, which are compensated byadditional substitution of alcohol for oil in thetransport sector. The most important change inOntario is the reduction in nuclear powerrequirements. Even under 20% emissionsreduction, no fresh investments in nuclearcapacity are required. Electricity consumptionrises in the residential and industrial sectors,increasing overall consumption by up to 10%.This naturally gives an emissions advantage,which is offset by substituting oil for somerenewables in the transport sector.

EMISSIONS TRADING WITH ELECTRICITYEXCHANGES OEEE)

In this scenario, the effects of the EE scenarioand the reverse of the JE scenario are superim­posed. Quebec maintains its electricity produc-

tion at the higher levels reached in the EE sce­nario, and reduces its consumption even fur­ther. But instead of reducing emissions else­where, it transfers the burden to Ontario.

There is a further substitution of naturalgas for electricity in Quebec's residential andcommercial sectors, thus reducing aggregateelectricity consumption by another 4%. Elec­tricity exports peak in the 20% emissions re­duction scenario. The new feature over the EEscenario is that, instead of adjusting emissionsthrough the transport sector, these are in­creased further because of the replacement ofsome alcohol by oil; the increased emissionsburden is then transferred to Ontario. In thiscase, Ontario has a clearly defined job to do: ithas to bring about higher emissions reductionsthan in the EE scenario, and has some addi­tional electricity with which to do this. This isachieved by substituting renewable energyand, of course, electricity for natural gas. Pene­tration of gas-based power plants is delayed,alcohol is substituted for oil in the transportsector, and wood is substituted for natural gasin residential heating.

The discussion of this section is summa­rized in Table 3 below. In the NC column, wecharacterize the trends of the main energyforms as a function of the severity of the re­duction target. In the EE and JE columns, wecompare the energy trends to those of the NCpolicy. Finally, in the JEEE column, the trendsare compared to those of the JE policy.

4. Conclusion

In this paper, the advanced multi-region, bot­tom-up model, Extended MARKAL, was usedfor an in-depth investigation of the responsesby the Quebec-Ontario energy system to a se­ries of increasingly severe GHG reduction tar­gets, ranging from unconstrained emissions toa cumulative 40% reduction over the next 45years. For each target, the responses were ana­lyzed under four cooperation policies resultingfrom the adoption or not of a joint emissionstarget and of electricity exchanges. In the fullcooperation policy, the joint MARKAL modelendogenizes the trading of GHG emissionspermits, as well as the electricity exchanges

127

Energy Studies Review, Vol. 8 [1996], Iss. 2, Art. 2

http://digitalcommons.mcmaster.ca/esr/vol8/iss2/2

Table 3: Key Energy Substitutions under Different Exchange Policies

Policy

No Cooperation JOint Electricity JOint(NC) Emissions Exchanges Emissions &

Target GE) (EE) ElectricityExchanges

GEEE)behaviour when

compared to compared to compared toreduction targetbecomes more severe NC NC JE

Quebec

Aggregate Electricity Production increases no change increases no change

Petroleum Consumption in decreases decreases decreases increasesTransport

Alcohol Consumption in Transport increases increases increases decreases

Electricity Consumption in increases under no change decreases decreasesResidential and Commercial severe reduction

Natural Gas Consumption in increases under no change increases increasesResidential and Commercial mild reduction

Aggregate Electricity Consumption increases no change decreases decreases

Ontario

Generation from Nuclear Plants increases decreases decreases decreases

Generation from Gas-based Plants increases under increases decreases decreasesmild reduction

Petroleum Consumption in decreases no change increases decreasesTransport

Alcohol Consumption in Transport increases no change decreases increases

Electricity Consumption in decreases in mild; no change increases increasesResidential and Commercial increases in severe

Natural Gas Consumption in increases in mild; no change decreases decreasesResidential and Commercial decreases in severe

Wood Consumption in Residential increases no change decreases increasesHeating

Aggregate Electricity Consumption increases only under no change increases increasessevere targets

within the two provinces.The most dramatic effects of cooperation

were found to be: a) a marked reduction of thecost incurred to meet the desired GHG reduc­tion; and b) a much reduced need for nuclearenergy in Ontario, when a cooperative policy isadopted. Both findings suggest that coopera­tive responses to GHG emissions caps shouldbe seriously considered by the two provinces.Another interesting finding is that electricityexchanges playa more important role than dopermit exchanges in achieving large cost sav-

128

ings.Although the paper did not discuss the

precise sharing of the cooperation dividendsbetween the two provinces, it did establish thesize of the dividends to be shared. Under fullcooperation, the theory of cooperative gamesproposes several alternative schemes for a real­istic or an equitable sharing formula. One ofthe Simplest and most appealing sharingscheme is the Shapley value (Shapley, 1953),by which, in the case of two players, the bene­fits of cooperation should be shared equally.

Loulou et al.: GHG Abatement in Central Canada with Interprovincial Cooperation

Produced by The Berkeley Electronic Press, 1996

The Shapley value approach was used in simi­lar contexts in Berger et al. (1990a, 1990b). Oncea sharing scheme is agreed upon by the twopartners, it serves as a basis for the pricing ofelectricity sales and of permits.

Many additional energy and technologysubstitutions are used by the combined Que­bec-Ontario system in order to achieve a least­cost solution to the imposition of GHG reduc­tion targets. In this paper, the complex sys­temic responses have been explained, thanks tothe detailed nature of the models used. Suchsystem-wide effects would be impossible tocapture fully via simplified aggregated mod­elling. In addition, the ability always to exhibita technological rationale for the solutions ar­rived at by the model, constitutes a powerfuladditional validation of the bottom-up philos­ophy.

It would be quite possible to extend thepresent analysis to at least four Canadianprovinces for which separate MARKAL mod­els do exist (i.e., Alberta and Saskatchewan, inaddition to Quebec and Ontario). Of course,extensions to several countries is also possible,and is indeed being undertaken by the ETSAPgroup of MARKAL modellers, in particular byBahn et al. (1994).

References

Bahn, 0., A Haurie, S. Kypreos, I.-P. Vial(1994) A Decomposition Approach to Multire­gional Environmental Planning: A NumericalStudy, Technical Report No. 1994.4, Univer­sity of Geneva, Switzerland.

Berger, c., R. Dubois, R. Loulou, A Haurie, E.Lessard (1990a) 'Assessing the Dividends ofPower Exchange between Quebec and NewYork State,' International Journal of EnergtJResearch 14:3:253-73.

Berger, c., R. Dubois, A Haurie, R. Loulou(1990b) 'Modelling Electricity Trading in theNortheast,' pp. 304-15 in Energy Supply/Demand Balances: Option and Costs, Proceed­ings of the Twelfth Annual North AmericanConference, International Association forEnergy Economics, October 1-3, Ottawa.

Fishbone, L.G. and H. Abilock (1981)'MARKAL, A Linear Programming Model

for Energy Systems Analysis: Technical De­scription of the BNL Version,' InternationalJournal of EnergtJ Research 5:4:353-75.

Fishbone L.G., G. Giesen, H.A Hymmen, M.Stocks, H. Vos, D. Wilde, R. Zoelcher, C.Balzer, and H. Abilock (1983) 'Users Guidefor MARKAL: A Multi-period, Linear Pro­gramming Model for Energy Systems Anal­ysis,' BNL 51701 (Upton, NY: BNL; Julich,Germany: KFA).

IPCC (1995a) Climate Change 1995 - The Scienceof Climate Change, Contribution of WorkingGroup I to the Second Assessment Report ofthe Intergovernmental Panel on ClimateChange (Cambridge, UK: Cambridge Uni­versity Press).

- (1995b) Climate Change 1995-Impacts, Adapta­tions and Mitigation of Climate Change: Scien­tific-Technical Analyses, Contribution ofWorking Group II to the Second Assess­ment Report of the Intergovernmental Panelon Climate Change (Cambridge, UK: Cam­bridge University Press).

-(1995c) Climate Change 1995-Economic andSocial Dimensions of Climate Change, Contri­bution of Working Group III to the SecondAssessment Report of the Intergovernmen­tal Panel on Climate Change (Cambridge,UK: Cambridge University Press).

Kanudia, A. and R. Loulou (1997) 'ExtendedMARKAL: A Brief User Manual for theStochastic Programming and Multi-RegionFeatures,' Cahier du GERAD G-97-11,GERAD (Montreal: Ecole des Hautes EtudesCommerciales).

Loulou, R. and J.-P. Waaub (1992) 'C02 Con­trol with Cooperation in Quebec andOntario: A MARKAL Perspective,' EnergtJStudies Review 4:2:278-96.

Natural Resources Canada, (1994) Perspectivesenergiliques du Canada 1992-2020 (Ottawa:Supply and Services Canada).

Shapley, L.S. (1953) 'A Value for n-PersonGarnes,' pp. 307-17 in H.W. Kuhn and AW.Tucker (eds.) Annals of Mathematics Studies,8/2, Contributions to TheonJ of Games, Vol. II(Princeton, NJ: Princeton University Press).

United Nations Framework Convention onClimate Change (UNFCq (1992) (Geneva:IDCC/UNEP).

129

Energy Studies Review, Vol. 8 [1996], Iss. 2, Art. 2

http://digitalcommons.mcmaster.ca/esr/vol8/iss2/2