a sustainable energy future: construction of demand and renewable energy supply scenarios

INTERNATIONAL JOURNAL OF ENERGY RESEARCHInt. J. Energy Res. 2008; 32:436–470Published online 12 November 2007 in Wiley InterScience(www.interscience.wiley.com) DOI: 10.1002/er.1375

A sustainable energy future: Construction of demand and renewableenergy supply scenarios

Bent S�rensen*,y

Department of Environmental, Social and Spatial Change, Energy, Environment and Climate Research Group, Roskilde University,Universitetsvej 1, Bld. 11.1, P.O. Box 260, DK-4000 Roskilde, Denmark

SUMMARY

The creation of energy scenarios, usually describing future situations of interest, involves three steps: (1) Determiningthe activities in the target society that involves energy of one or another form. Examples of carrying out such an analysisare presented, with end-use demands distributed on energy forms (qualities) as the deliverable outcome. (2)Determining the available energy resources in the society in question. This is done for renewable energy resources andpresented as potential energy supply, with a discussion of the aggressivity of exploiting such sources. Finally (3)Matching demand and supply under consideration of the energy forms needed, with the use of intermediateconversions, storage and transmission and signalling unused surpluses that may be exported from the society inconsideration, or deficits that have to be imported. An example of such a matching is presented in an accompanyingarticle. Copyright # 2007 John Wiley & Sons, Ltd.

KEY WORDS: scenario technique; energy modelling; demand analysis; supply assessment; resource appraisal

1. ENERGY SYSTEMMODELLING}GENERAL

CONSIDERATIONS

The aim of the present work is to model a range ofoptions for the future energy system of a givensociety, say a country, with consideration of thesurrounding energy systems (such as those ofneighbouring countries) that may come into play

by exchange of energy including purchase of fuelsor other energy services. Owing to the considerableinertia in the system, caused by existing equipmentand infrastructure, the time horizon is chosen asaround 50 years, in order to capture the possibilityof a complete change in the mix of energy sourcesand the ways of converting, transmitting and usingenergy in society. It is thus the target of the studyto provide material for decision makers that may

*Correspondence to: Bent S�rensen, Department of Environmental, Social and Spatial Change, Energy, Environment and ClimateResearch Group, Roskilde University, Universitetsvej 1, Bld. 11.1, P.O. Box 260, DK-4000 Roskilde, Denmark.yE-mail: [email protected]

Contract/grant sponsor: Energy Research Programme of the Danish Energy Agency; contract/grant number: EFP-033001/33031-002

Received 10 January 2007Accepted 21 March 2007Copyright # 2007 John Wiley & Sons, Ltd.

help them in selecting an optimal energy solutionwith high economic benefits for the society inquestion. On the other hand, going thus far intothe future induces uncertainties and necessitatesthe formulation of assumptions that may turn outto be incorrect. For this reason, the methodologyselected is that of scenario analysis, renouncing onfinding a single optimal solution and insteadanalysing a number of alternative scenarios fortheir advantages and disadvantages. This willenable decision makers to apply their preferredweights describing the importance of variousfactors such as direct economy, environmentalimpacts including climate change, supply securityand robustness against (at least some) errors inassumptions. The scenarios contemplate a numberof solutions based on equipment or strategy notfully developed or tested. Although surprises interms of new solutions appearing within theplanning period are possible, the method em-ployed makes them less probable, because the50-year horizon is short enough to make it highlylikely that all technologies that can be made readywithin that period are already known at present, insome (possibly early) stage of development andreadiness.

2. ENERGY SYSTEMMODELLING}METHODOLOGY

2.1. Demand options

The definition of energy demand used here is trueend-use energy, i.e. the energy derived after thefinal conversion taking place at the end-user forsupplying some demanded energy service. Thisenergy can be defined rigorously once the demandof energy services is determined, on the basis of avision of the activities of the future society [1,Chapter 6].

This definition is independent of the efficiency ofthe possibly several energy conversion steps takingplace between primary energy and end-use energyservice. These important aspects will be discussedin Section 2.2, and together with the end-useenergy they determine the entailed requirement forprimary energy supplies.

In order to model energy trade between thecountry focussed upon and its neighbours, thepatterns of surplus and deficit must be determined(as function of time) for each region. This impliesthat demand scenarios and conversion efficiencyassumptions have to be made not only for sayDenmark, but also for the energy exchangepartners such as the Nordic countries and Ger-many, which are already connected to Denmarkby electric power grids. If the primary interest is ona single country, the demand models for theneighbouring countries do not have to be asdetailed as for the primary country, but in manycases, simultaneous studies of groups of countrieswith energy connections are the objectives.

The following is a list of relevant precursor end-use demand scenarios, using Denmark and itsneighbouring countries as an example but stillretaining a level of generality. By ‘precursorscenarios’ a set of preliminary scenarios is meant,out of which the final scenarios may be selected forcloser investigation. The bases for the precursorscenarios are the assumptions regarding desirableactivity levels and energy intensities of theactivities, as they have appeared in the energydebates in recent years, in Denmark and in Europebroadly. They claim no completeness, but try todisplay enough diversity to serve as a useful spanof the challenges facing the energy planner. Out ofthese, a more limited number of scenarios willtypically subsequently be selected or reformulatedfor closer discussion and concrete simulationefforts.

2.1.1. Run-away precursor scenario. In the run-away scenario, the energy demand grows at leastas quickly as the overall economic activity(measured e.g. by the gross national product).This has historically been the case during periodsof exceptionally low energy prices, notably in theyears around 1960. Conditions for this scenario, inaddition to low energy prices, would includemeasures such as encouraging transportation work(many passenger-kilometres facilitated by moreroads, cheap air connections and decentraliza-tion of the locations of homes, work places andleisure facilities, many ton-kilometres of freighthaul facilitated by decentralization of component

A SUSTAINABLE ENERGY FUTURE 437

Copyright # 2007 John Wiley & Sons, Ltd. Int. J. Energy Res. 2008; 32:436–470

DOI: 10.1002/er

production and shipment of small-size cargos). Inthe building sector, more square metres of livingspace and more square metres per unit ofeconomic activity, and in the electricity sector,more appliances and other equipment. Buildingstyle developments could create a perceived needfor air conditioning and space cooling. Forindustry, there could be increased emphasis onenergy-intensive production, although this ishardly relevant for countries such as Denmark.However, service sector activities and their energyuse could increase substantially, with retail shop-ping areas greatly increased and use of businesspromotion by light and other energy-demandingdisplays. For leisure activities, traditional naturewalks or swimming could be replaced by moto-cross, speedboat use and other energy-demandingactivities, similar to the habits already seen toexpand in North America.

2.1.2. High-energy-growth precursor scenario. Thehigh-energy-growth scenario is similar to the run-away scenario, but with a slower increase in energydemand. In the transportation sector this could bedue to a certain saturation tendency in transportactivities, due to higher value placed on the timelost in travelling on more congested roads and inmore congested air space. For industry, continueddecrease in energy-intensive production may leadto a demand growing less than the economicactivity. In buildings, heat use may increase lessthan floor area, due to zoning practices, etc.Generally, activity level and energy demand maysee a certain amount of decoupling, reflecting thefact that the primary demands of a society aregoods and services, and that these can be providedin different ways with different energy implica-tions. A certain effect of this type damps theenergy demand in the high-energy-growth scenarioas compared with the run-away scenario, but duemore to technological advances and alteredDanish industry mix than to a dedicated policyaimed at reducing energy demand.

2.1.3. Stability precursor scenario. The stabilityscenario assumes that the end-use energy demandstays constant, despite rearrangements in specificareas. Specifically, the energy demand in the

building sector is assumed to saturate (consideringthat the number of square metres per personoccupying the building, whether for work orliving, will not continue to increase, but reach anatural limit with enough space for the activitiestaking place, but not excessive areas to clean andotherwise maintain). In the industry sector, anincreasingly knowledge-based activity will reducethe need for energy-intensive equipment, replacingit primarily by microprocessor-based equipmentsuited for light and flexible production. Industrialenergy use will decline, although industries like theservice and private sectors will continue to addnew electronic equipment and computers. In othersectors, dedicated electricity demand will increasesubstantially, but in absolute terms more or lesscompensated for by the reductions in the industrysector. For transportation, saturation is assumedboth in number of vehicles and number ofpassenger- or ton-kilometres demanded for thereasons outlined above (in the section on the high-energy-growth scenario). Reasons for consideringthis possible could include the replacement ofconference and other business travel by videoconferencing, so that an increase in leisure tripsmay still be included. Presumably, there has to beplanned action for this to be realistic, includingabandoning tax rebates for commercially usedvehicles and for business travel, and possibly alsoefforts in city planning to avoid the current trendof increasing travel distances for everyday shop-ping and service delivery. The stability scenariowas used as the only energy-demand scenario in anearlier study on the possibilities for hydrogen inthe Danish energy system (see [2–4]).

2.1.4. Low-energy-demand precursor scenario. Inthe low-energy-demand scenario, full considera-tion is paid to the restructuring of industryassumed for countries such as Denmark, fromgoods orientation to service provision. Alreadytoday, many Danish enterprises only develop newtechnology and sometimes test it on a limitedDanish market: once the technology is ready forextended markets, the production is transferred toother companies, usually outside Denmark. Thischange in profit-earning activities has implicationsfor the working conditions of employees. Much

B. SØRENSEN438


DOI: 10.1002/er

work can be performed from home offices, usingcomputer equipment and electronic communica-tions technology and thereby greatly reducing thedemand for physical transportation. Also in theretail food and goods sector, most transactionsbetween commerce and customer will be madeelectronically, as it is already the case in a numberof sub-sectors today. An essential addition to thistype of trade is the market for everyday products,where the customers until now have made limiteduse of electronic media to purchase grocery andfood products, probably by reasons of a perceivedneed to, e.g. handle the fruit to see that it is ripebefore buying it. Clearly, better electronic tradearrangements with video inspection of actualproducts could change the reservations of currentcustomers. If everyday goods are traded electro-nically, the distribution of such goods will also bechanged to an optimal dispatch requiring con-siderably less transport energy than today’spersonal shopping. All in all, a quite substantialreduction in energy demand will emerge as a resultof these changes, should they come true. Theeconomic development is further de-coupled fromenergy use and may continue to exhibit substantialgrowth.

2.1.5. Catastrophe precursor scenario. In the cat-astrophe scenario, a reduced energy demand is dueto the failure to achieve a desirable economicgrowth. In the case of Denmark, reasons could bethe current declining interest in education, parti-cularly in those areas most relevant to a futureknowledge society. In this scenario there would bea lowering of those enabling skills necessary forparticipating in the international industry andservice developments, and the alternative ofimporting these intellectual skills is seen as havingbeen missed by an immigration policy unfavour-able to precisely the regions of the world produ-cing a surplus of people with technical and relatedcreative high-level education. Although there arepessimists who see this scenario as the default, i.e.the situation Denmark is moving towards unlessstrong policy measures are taken in the nearfuture, the stance here is that the traditionalopenness of the Danish society will also this timework to overcome the influence of certain negative

elements. A key reason that this may be likely isthe smallness of the Danish economy in the globalpicture, implying that even if Denmark shouldchoose to concentrate on less education-demand-ing areas such as coordination and planning jobsin the international arena (requiring primarilylanguage and overview skills), these could easilyprovide enough wealth to a nation of open-mindedindividuals willing to serve as small wheels inlarger international projects. This would make theeconomic decline a passing crisis to be followed bythe establishing of a small niche existence forDenmark, which in energy terms would implyreturning to one of the central scenarios describedabove. Only in case Denmark becomes interna-tionally isolated, will this option fade away and thecatastrophe scenario become reality. For othercountries, some aspects of these threats are alsoevident, but the discussion will have to be repeatedfor each country on its specific premises.

2.2. Energy conversion system

Countries such as Denmark have a long traditionfor placing emphasis on efficient conversion ofenergy. Following the 1973/1974 energy crises,particularly detached homes (where the occupantsare also the owners making decision on invest-ments) were retrofitted to such an extent that theoverall low-temperature heat use in Denmarkdropped by 30% over a decade. CO2 and pollutiontaxes on electricity have probably been a signifi-cant cause of the appliance-purchasing pattern,where the lowest-energy-consuming equipment hastaken a dominant part of the market. The sametrend is at least partially seen in automobilepurchasing, where a non-linear energy-efficiencydependent annual registration tax has made thesale of the most energy-efficient vehicles muchhigher in Denmark than in other Europeancountries with similar fuel costs. This trend is onlypartial, because there is still a substantial sale ofluxury cars and four-wheel-drive special utilityvehicles not serving any apparent purpose in acountry with hardly any non-paved roads. If theinitial registration tax on automobiles were simi-larly made energy efficiency dependent, the effectwould be much greater. The Danish utilities are



DOI: 10.1002/er

known for constructing some of the highest-efficiency conventional power and heat plants inthe world (using coal, natural gas and wood scrapor other biomass-based fuels), and Danish windturbines are also known for high efficiency. Thetransmission losses are fairly low, but Denmarkcurrently has a smaller coverage with undergroundcoaxial cables than many other European coun-tries, resulting in continued vulnerability duringstorms, for the remaining overhead lines. This,however, is finally in the process of beingremedied.

With this background, it is expected that theconversion efficiency will continue to improve inall areas, from primary over intermediate to end-use conversion. However, it might still be a goodidea to consider more than one scenario, forDenmark and particularly for other countries thatare currently not very far along the route to energyefficiency. The following are three precursorscenario suggestions.

2.2.1. Laisser-faire precursor scenario. In the lais-ser-faire scenario, conversion efficiencies are left tothe component and system manufacturers, whichwould typically be international enterprises (suchas vehicle and appliance manufacturers, powerstation and transmission contractors and thebuilding industry). The implication is that effi-ciency trends follow an international commondenominator, which at least in the past has meantlower average efficiency than suggested by actualtechnical advancement and sometimes even lowerthan the economic optimum at prevailing energyprices. Still, the efficiency does increase with time,although often for reasons not related to energy(for example, computer energy use has beenlowered dramatically in recent years, due to theneed to avoid component damage by the excessheat impact from high-performance processors).The gross inadequacies of the current system,deriving from the tax exemption of internationaltravel and shipping by sea or air (as imposed bythe WTO) and its impact on choice of transporta-tion technology, are assumed to prevail.

2.2.2. Rational investment precursor scenario. Inthe rational investment scenario, the selection of

how many known efficiency measures, which willactually be implemented through the technologieschosen at each stage in the time development ofthe energy system, will be based on a lifetimeeconomic assessment. This means that the effi-ciency level is not chosen according to a balancingof the cost of improving efficiency with the currentcost of energy used by the equipment, but with thepresent value of all energy costs incurred duringthe lifetime of the equipment. This assessmentrequires an assumption of future average energycosts and it is possible by choice of the cost profileto build in a certain level of insurance againstsurprises from higher-than-expected energy prices.The important feature of the rational investmentscenario is that it forces society to adopt a policyof economic optimization in the choice of energy-consuming equipment and processes. This ispartially implemented at present, e.g. through theenergy provisions in building codes, throughappliance labelling and through vehicle taxation.In the latter case, there is a distinction between theefficiency optimization for a vehicle with given sizeand performance, considered here, and the ques-tion of proper vehicle size and performancecharacteristics dealt with in the previous sectionon end-use energy. The current energy taxation forpassenger cars does not make such a distinction,and the tax reduction for commercial vehiclesactually counters rational economic considerationsand should rather be characterized as taxpayerssubsidising industry and commerce.

2.2.3. Maximum efficiency precursor scenario. Themaximum efficiency scenario could be based on theidea that every introduction of new energy-consuming equipment should be based on the bestcurrent technical efficiency. This would implyselecting the highest-efficiency solution availableat the marketplace, or even technology ready forbut not yet introduced into the commercialmarket. Higher-efficiency equipment under devel-opment and not fully proven would, however, notbe implemented, except for as parts of demonstra-tion programmes. This ‘best current technology’approach was used in several previous scenariostudies [1–6] with the reasoning that the currentlybest-efficiency technology would be a good proxy

B. SØRENSEN440


DOI: 10.1002/er

for the average-efficiency technology some 50years into the future. Depending on the assump-tions regarding future energy prices, the rationalinvestment scenario could be less efficient, ofsimilar efficiency or of higher efficiency than whatis offered by the ‘best current equipment’approach.

A maximum efficiency scenario better reflectingits name would in this light be defined by makingprojections regarding the typical average efficiencyimprovements over the planning period, and thenby insisting that the best technology at eachinstant in time be used for all new equipmentintroduced at that moment in time. While projec-tions on future efficiency of individual pieces ofequipment may be uncertain and sometimeswrong, it appears reasonable to assume thataverage efficiencies over groups of related equip-ment can be extrapolated more reliably.

2.3. Supply options

Basic energy supply is important in all scenarios,because the technologies for further conversiondepend on the type of primary energy used(through factors such as physical form (gas, liquidor solid), energy quality, temporal and geographicpatterns of provision). Further important consid-erations relate to reliability of supply, with issuessuch as resource depletion and stability of tradingpartners being crucial.

The present study is based on the premise thatfossil fuels are a temporary solution due to bothdepletion and their emission of pollutants andgreenhouse gases. In the case of oil, the concernover resource depletion is large, with North Searesources expected to fade out over the nextdecades, and substantial amounts being availableonly in the politically unstable Middle Easternregion. Even these substantial amounts may lastonly about half a century, especially if the newdemands in rapidly developing countries such asChina continue to rise [4, Chapters 5 and 7]. Alsonatural gas resources in Europe are expected todecline soon. There is substantial discussion of thereliability of global resource estimates, with apossible hope of discovering new gas fields inunexpected locations, but on the other hand, the

reserves in Russia may suffer from politicalinstabilities as much as the resources in the MiddleEast. Only for coal is the resource base substantialand the possibility of supply during more than 100years a realistic proposition. However, coal emitsmore greenhouse gases per unit of energy than theother fossil fuels, and a growing consensus isemerging, according to which continued use ofcoal (with smaller contributions from other fossilfuels) is acceptable only if the CO2 emissions canbe sequestered or avoided, e.g. by transferring theenergy content in coal to hydrogen before use. Allthis points to fossil fuels being of interest for futureenergy supply only if combined with a successfuldevelopment of hydrogen technologies [4, Chapter5]. Environmental and political impacts are furtherdiscussed below in Sections 2.6 and 2.7.

The nuclear fuels currently used in some parts ofthe world do not emit greenhouse gases but have anumber of other serious problems, related toinfrequent but large accidents (causing globalradioactive fallout such as in the Chernobylaccident), to radioactive waste accumulation(waste that has to be kept separate from thebiosphere for periods much longer than theaverage life time of countries or even civilizations)and finally to divergence of nuclear materials tobelligerent nations or terrorist organizations cap-able of manufacturing nuclear bombs. Efforts tomodify the nuclear technologies to avoid or reducethese problems have been ongoing for some time,but with slow progress. One fundamental problemis that known nuclear fuel reserves are no largerthan those of oil, if they are to be used in once-through nuclear cycles, and a contribution fromnuclear energy to the post-fossil era thus dependson successful development of breeder reactorswithout any of the mentioned problems. It is verydoubtful if there is any chance of meeting theserequirements, but if the R&D is undertaken and issuccessful, it likely again involves hydrogen as anintermediate energy carrier [4, Chapter 5].

The most promising successors to currentdepletable and environmentally dubious fossiland nuclear fuels are the renewable energytechnologies based on wind, direct solar conver-sion and biofuels, supplemented by existing hydro-power. Currently, wind power and some biofuel



DOI: 10.1002/er

technologies have direct costs similar to those offossil fuels, but at least in the case of wind withsubstantial indirect economy benefits from re-duced pollution as well as absence of greenhousegas emissions. Furthermore, the prospect of risingenergy prices due to global competition for thefinite oil resources will make more renewableenergy technologies economically attractive overthe coming years. With respect to resource size,coastal countries like Denmark have wind powerresources capable of supplying all the electricityneeds of all the demand scenarios considered. Athigh latitudes, the solar resources have timedistributions poorly matching the variations indemand and are therefore not expected to gain adominant role, although production of a storableintermediate fuel (such as hydrogen) duringperiods with high solar radiation is a possibility.At lower latitudes, solar power and heat couldbecome a major energy supplier, if the costsdevelop favourably.

Biomass is a major energy source in countriessuch as Denmark due to the substantial agricul-tural production, furnishing lots of residue-usageoptions. This can be supplemented by sustainableforestry residues. Current Danish food productiongreatly exceeds the Danish demand and is a keyexport article. Therefore, any new, dedicatedenergy crops would have to compete with foodproduction, where the average price of finalproducts is 5–10 times higher than current pricesfor the inherent energy content. This points tobiomass residues as the basis for energy uses,rather than the primary harvested crops (e.g.sugar). If the biomass production resembles thecurrent mix, then there will be about 10 times asmuch residues as food products (measured interms of energy content), and it is thus possible toderive about equal economic benefits from foodproducts and residues converted into energyproducts, at current prices. Presently, this isinsufficient to cover the additional cost of conver-sion to energy products, but again, the expectedenergy price increases can make it attractive in therelatively near future [1, 4]. Current use of biomassresidues for direct combustion is unlikely to bean acceptable solution in a future unhappy withthe pollution from fossil fuels. Particularly, the

dispersed use of biomass in home furnacescurrently gives rise to a much larger fraction ofundesired emissions than the fraction of energyproduced. A possible addition to the land-basedbiomass resources would be extensive aquaculture,e.g. on those off-shore areas set aside for windpower installations.

2.4. Role of energy storage

Mismatch between energy supply and demandmay be handled by a number of measures, ofwhich energy storage is an obvious one in cases,where the primary energy production cannot becontrolled, whether it is due to the fluctuations insolar or wind energy, or to built-in constraints inthe system such as fixed platform gas productionor bound heat–electricity ratios in combinedpower plants. A second possibility is demandmanagement, where tasks that are not time-wiseurgent can be postponed until it is favourable forthe energy system to satisfy them. There wouldnormally be limits to the length of displacement,and the final user should be able to see aneconomic advantage in subscribing to such ascheme. Some tasks have to be made on demand,and the demand management is therefore only apartial solution to variability of renewable sources,which includes periods of no supply at all, at leastfrom local resources.

The amount of wind variability with distance ofwind turbine dispersal has been studied, andalthough the power duration curve is flattened,the period of below-average supply remains nearlythe same as for a single, well-placed turbine [1]. Inorder to take advantage of variable wind regimes,wind turbine output should be combined (traded)over distances large enough to ensure passage ofseparate weather systems, which means 500 km ormore. For solar panels, the day-to-night variationcan only be smoothed by connecting the outputfrom panels placed at different longitudes aroundthe world, and the seasonal variation only byconnecting across hemispheres (or preferably placesolar panels near the Equator).

Similar considerations apply to many forms ofdemand management. For example, social habitsmake it difficult to disperse the period of hot meal

B. SØRENSEN442


DOI: 10.1002/er

cooking by more than 1–3 h, but demand manage-ment across different longitude regions could domuch better. This has actually been done for anumber of decades in Russia, where the nationalelectricity grid covers many longitudinal zones.Much more limited possibilities are offered by thechiefly North–South exchange of power in Wes-tern Europe. Also industry and commercial energyuse has limited flexibility, due to conventionalworking hours and the need to adapt to the timewhere customers like or are able to shop. Lightingand electronic equipments such as computers oraudio-visual devices show similar inflexibility(unless they are equipped with rechargeablebattery modules) because their use is determinedby the schedules of people. Only a few items suchas dish or clothes washing machines and driersallow the desired flexibility of 3–30 h of delay,being desirable as an alternative to short-termenergy storage.

Active energy storage can be based on alarge number of devices, which may broadly bedivided according to capacities measured interms of typical storage times: seconds to minutesfor reasons of system stability (flywheels, capaci-tors), hours to days for optimal use of fuel-basedpower systems (pumped hydro, compressed air,batteries), weeks to months for weather-dependentenergy production such as solar or windpower (seasonal water reservoirs, hydrogen,reversible phase change or chemical reactions)[1]. Seasonal water reservoirs are often the cheap-est among these, but suitable reservoirs exist onlyin special regions (e.g. in Norway but not inDenmark), and cheapness assumes that the envir-onmental costs of using such reservoirs are lowand that the other country does not overchargefor this service. In scenarios for future energysystems based on renewable energy supply, thelong-term storage options are particularly inter-esting, and one purpose of the present study is toexplore the possible roles that hydrogen can fill inthis connection. However, ‘long-term’ is typicallyonly a few weeks for wind power systems (see[2, 3]), whereas solar heat or power has both adiurnal and a seasonal (6 months) componentunder high-latitude conditions. This is one of thereasons that wind is seen as more appropriate than

solar energy for the North European energysystem.

2.5. Role of trade (agreements or pool auction)

To the extent that transmission capacity is alreadyin place or can be established at acceptable cost,exchange of power between different regions orcountries would appear an ideal way of handlingsurpluses and deficits in a given system. However,import requires that there is surplus productioncapacity in the neighbouring regions/countries,and export that the adjacent production can beadjusted downwards. Historically, a certain sur-plus capacity was normally present in any electricutility system, but with privatization of theproduction industry, a tendency to maintain thesmallest possible limit of extra capacity hasemerged, based on economic arguments incorpor-ating the smallness of eventual penalties for beingunable to satisfy demand during a (statistically)few hours a year. This policy has, e.g. in theUnited States of America, been blamed for majorblackouts occurring in the Eastern states.

Currently, Denmark can avoid active energystorage and deal with fluctuating production byrenewable energy sources (where Denmark hassome 20–30% of electricity generated by wind)through trade with the neighbouring countries. Ifall the countries concerned have large shares offluctuating sources in their power systems, theunconditional need for import and export adjust-ment would apply to all nearby countries, and to alarge extent surpluses would coincide in time anddeficits develop simultaneously (due to the char-acteristics of, e.g. wind systems discussed in theprevious section). In such a system, only theremaining fuel-based units and the hydropowerunits could serve to adjust the production up ordown, and ultimately, as renewable sources takeover, the fuel-based back-up would have to bebased on biofuels. Still, for a while one couldmaintain cheap fossil generators (gas turbines) foruse in these situations, as long as the periods ofusing this option are short and the annual fuel usecan be kept at insignificant levels.

In Scandinavia, the situation is better thanin most other regions, regarding trade as an



DOI: 10.1002/er

alternative to energy storage. The reason for this isthe very large component of hydropower based onseasonal water stores. Particularly in Norway,water reservoirs may contain enough water toserve the average supply for nearly a year (exceptin years of exceptionally low precipitation), andthe generating capacity is quite generous com-pared with the average production. The implica-tion is that adjustments of say Danish wind powersurpluses and deficits even in case of 100%coverage of Danish electricity supply with windwould produce minute variations in hydro reser-voir filling relative to the case of no exchange [7–9].However, being aware of its very special endow-ment, Norway will likely try to maximize theeconomic revenues from delivering such services toneighbouring countries.

Continuing this line of reasoning, Norway hashistorically maintained electricity prices far belowEuropean averages, and the current liberalizationof the power market has shown an expectedtendency to approach European price levels. Thismeans that the use of electric power by theNorwegian consumer is unlikely to remain at thecurrent level high above the consumption inneighbouring countries. It increasingly pays, asseen by the Norwegian customer, to invest in moreefficient energy use, and the result is that thealready installed Norwegian hydro energy capacity(which is unlikely to increase for environmentalreasons) will in the future likely be considerablyabove the indigenous Norwegian consumption,leaving room for substantial exports to theEuropean continent, provided that transmissionlines have sufficient capacity.

This very special situation in the North ofEurope will be an important consideration indiscussing the region’s options for dealing withfluctuating energy sources. As hinted at, theNorwegian effort to reach market prices mayalso imply that the use of power trade as a methodto handle the intermittency problem will reach aprice (unrelated to cost) which may be consider-ably higher than today. Assessing the likely futurelevel of such prices is an important part of thebalancing between the options of trade/exchangeand active energy storage, where estimates offuture price structures will be much more essential

that just considering the current costs (whether forpower exchange contracts or for Nordic powerpool trading on various conditions related towarning times). Still, it is important to investigatethe times of the day where wind surplusesand deficits are likely to occur, and compare themwith diurnal variations in expected future powerpool prices.

Other considerations are the seasonal tradingprice variations and at the next time scale, theeffects of Norwegian dry or wet years as basis forthe availability of hydropower. Also wind genera-tion varies from year to year, but due to the timebracket of at most a few weeks for ‘repaying’ thepower loaned, the hydro system would not beparticularly strained by such power exchange evenduring dry years, although the general price levelmay be higher. In this situation, a net export ofpower from Denmark to Norway (or to a lesserextent to Sweden) may by an economicallyinteresting option, requiring a certain overcapacityof wind turbines to be installed.

2.6. Environmental aspects

One basic reason for moving away from fossil andnuclear fuel-based energy systems is to avoid theenvironmental problems associated with them, dueto emissions, wastes, accidents, proliferation andin the case of carbon-containing fuels also climateimpacts. While technological improvements canreduce the impacts associated with pollution fromemissions, greenhouse gas emissions are betterhandled not by cleaning the flue gases, but by abasic transformation of the fuel before combustionor conversion, e.g. moving the energy from carbonto hydrogen. Large-scale application of suchmeasures will in any case create large amounts ofcarbon-containing waste to be disposed of. Work-able solutions to these problems are under study,but it is too early to decide which (if any) of thesuggested technologies may become environmen-tally acceptable. Removing CO2 from the fluegases after conversion (such as electricity produc-tion) is considered less attractive due to high costand a quite low efficiency, both in terms of energybalance and of the fraction of CO2 actuallycaptured [6, 10].

B. SØRENSEN444


DOI: 10.1002/er

Among the renewable energy sources, directsolar and wind systems have very small impacts,mainly associated with the materials used inmanufacturing the equipment}noise and visualimpacts being manageable and temporary. Hydro-power has severe environmental impacts if itinvolves the creation of reservoirs, altering thelandscape and biosphere over areas that are quitelarge, e.g. in the Norwegian case. For this reason,global large-scale hydro expansion has nearlyhalted, and only small run-of-the-river hydroplants are constructed, with careful integrationinto the natural setting [1].

Most environmental concerns over renewableenergy sources are associated with biomass use.Although biomass sources are usually taken asCO2 neutral due to the balance between previousCO2 intake during plant life and emission whenused for energy, many biomass uses have severeenvironmental impacts during their utilization forenergy purposes. This is particularly true forcombustion in small-scale furnaces, but also inlarge boilers where pre-treatment has often madethe biomass fuel more uniform, there are emis-sions, particularly during start-up. Whether indeveloping and amateur or in industrial settings,the burning of biomass cannot be consideredenvironmentally sustainable, and other ways ofusing the energy in biomass should be considered.

The fermentation routes offer ways of dealingwith biomass residues and waste from householdsand have gained popularity as a viable method ofwaste treatment, but often without exports ofsurplus energy out of the plant. The net energyproduction by fermentation depends on the energycost of collecting and transporting biomass re-sidues to the biogas reactor site, and there is a clearcompromise to be made between transportationcost and economy of scale for the fermentationplant [10].

Another route to gaseous biofuels is to gasifybiomass residues (both lignin-containing woodscrap and also agricultural waste of moderatewater content). The producer gas can be useddirectly in industrial furnaces, but distribution bypipeline to former natural gas customers, or use inthe transportation sector, requires purificationand/or reforming, if specific fuels such as hydrogen

are desired. Pollution from the involved industrialprocesses is believed to be containable, and thetransformation routes are thus preferable to directcombustion in boilers or furnaces.

A set of technologies estimated to gain anincreasing and perhaps dominating role in biomassconversion is the production of liquid biofuels.These include ethanol, methanol and biodiesel,already produced in some countries and mixedinto vehicle fuels. Most vehicle engines allow acertain amount (on the 10% level) of alcohols tobe mixed into the gasoline fuel without requiringengine adjustments. Also, many diesel engines canoperate on pure biodiesel fuel without modifica-tion. The current production of ethanol is basedon sugar, while methanol is based on wood scrap.More expensive catalytic production methods arebeing developed, which can use virtually anybiomass residue for the production of these fuels.The same is true for biodiesel, which today ismostly produced on the basis of grains and seeds.All these biofuels have associated air pollutionwhen used in, e.g. vehicles. This could be avoidedby further reforming to hydrogen, but therebyloosing the advantage of a high energy-densityliquid fuel similar to the ones used presently andthus with small infrastructure change requirementsentailed [4].

2.7. Non-economic factors

By non-economic factors are meant all effectsinvolving an impact on an economy that is hard toquantify, at least in monetary terms. An importantexample is supply security. This highlights con-siderations such as the variety of suppliersor supply options, and the risk of disruption ofsupply, for instance due to natural disastersor political instability, conflicts or warfare.

The situation here is that for petroleum pro-ducts, several of the North Sea deposits currentlyused by, e.g. Denmark are expected to decline overa period less than the 50-year planning horizon ofthe present study. High-latitude finds in Norwe-gian or Greenland seas may extend that period.Most of the world’s remaining oil deposits are inthe Middle East, in countries of highly unstablesituations, with dictatorial regimes challenged by



DOI: 10.1002/er

religious fundamentalists and with high risk ofinternal conflicts or civil wars. At the same time,the strongly growing oil demand from countriessuch as China and India and the increasingimports by the U.S.A. due to decline in its ownoil resources make it impossible to satisfy globaldemand without production being continuouslystepped up and more than doubling over theplanning period considered. These facts wouldseem nearly impossible to reconcile, and steeplyincreasing prices are likely to be the first sign of therising problem. The very nervous nature of the oilmarket will likely overlay the cost increase withfluctuations both up and down. For Europe, theinterest in oil substitution should thus be a toppriority. The U.S.A. could (in principle) stretch itsown oil resources by increasing its energy useefficiency to at least the present European level, apossibility that would require a reversal of bothpolitical and consumer attitudes.

For natural gas, the European situation issimilar to that of petroleum, with possible addi-tional resources primarily in the North, but also inRussia (including the non-European part). Recentevents have shown some signs of instability ofRussian policy towards gas exports (notably toEast European neighbours), but generally, Russiahas been a stable supplier, at least in the past.Substitution options for natural gas uses are betterthan for oil in the transportation sector, but willnot become available without effort (such asrapidly establishing a coal gasification pro-gramme). Owing to the pipeline type of supply,the final users are in a relatively inflexiblesituation, presently taken care of by havingestablished large underground gas stores. Danishgas storage facilities can furnish at least twomonths of supply, deemed necessary in case ofmajor ruptures in the undersea pipeline to theDanish North Sea platforms. Seen in the light of amajor substitution, this period is still too small tosupply proper supply security.

Finally, for coal, there is a geographically moreattractive distribution of resources, but still doubtsregarding the willingness to expand production arepresent, as would be needed if coal should takeover some of the oil and gas markets (after suitabletransformation). Because coal has a higher

emission of greenhouse gases per unit of energyproduced, the expanded use of coal is connectedwith aims to transform coal into hydrogen fuelbefore usage. This makes the future of coal highlyconnected to the successful development of hydro-gen technologies, including the fuel cell technolo-gies to be used in the transportation sector andpossibly in a much wider range of stationaryapplications.

Most uranium resources are in Niger and in theunstable region of Islamic former Soviet republics,although resources relatively evenly distributedover the rest of the world can sustain a modest useof nuclear energy.

Renewable energy sources (except hydro andhigh-temperature geothermal) are much moreevenly distributed over the regions of the world,with direct solar use having better prospects nearthe Equator (due to seasonal invariance), whilewind energy is most abundant under the mid-latitude jet streams. Biomass production is rela-tively similar all the way from the Equator tolatitudes of about 608; because the different solarradiation levels are compensated by oppositelyvarying levels of soil moisture and stability ofnutrient supply. As a result, a suitably selected mixof renewable energy resources can supply allenergy needs in nearly all parts of the world,provided that rural-to-city transmission and a levelof international trade are maintained at about thesame level as conducted today, but avoiding tradewith the politically more unstable parts of theworld [1, 5, 11]. Generally speaking, renewableenergy sources have the least indirect impacts orimpacts with uncertain or hidden economic costs.

3. SCENARIO CONSTRUCTION

In Section 2 above, some precursor scenario ideasfor the end-use demand and for the efficiencyefforts in conversions were discussed. A full energysystem scenario needs to add scenarios for choiceof energy sources and for the assignments ofenergy carriers to different energy-dependent tasksperformed in society. Previous studies for Den-mark [2, 12] and for the World [1, 4, 5] assumedmoderate activity increase combined with high

B. SØRENSEN446


DOI: 10.1002/er

energy efficiency to prove that any region in theworld could satisfy demands with 100% renewableenergy, with a different mix in different regionsand with continued energy exchange and trade(Denmark being an export country). Below I givean account of the update in demand and supplyscenarios used in the current investigation [13].

3.1. Demand scenario construction

The demand scenarios are constructed using thebottom-up approach [1, 2, 6]. Actual human de-mands are for services and products, all of whichmay (or may not) involve conversion of energy.The ways to deliver any needed energy arediscussed below in Section 3.2. Here all thedemands believed to have a relationship to energyuse are listed, with a core that is indisputable(‘basic demands’) and continuing to increasinglynegotiable wishes and desires (‘secondarydemands’). Because there is no point in guessingin great detail what secondary demands mayprevail in future societies, the discussion isgenerally held at an aggregate level. This isconsidered a better way to characterize futurechoices because it is a way that does not depend onidentifying the precise nature of new technologiesand new consumer products.

3.1.1. Biologically acceptable surroundings. Thisbasic demand requires access to indoor space witha temperature of about 208C in home and forindoor working. As in the previous work [1, 2, 6],we use an assumed floor area of housing/workspace to define this demand. In all scenarios, thevalue assumed is 60 m2 per person, such asfurnished by 40 m2 cap�1 in the home and 20 m2

cap�1 in the work area. Ceiling height is taken at2.3–2.6m. The implied space is not the minimumfor biological purposes, but includes a generoussecondary demand for a large indoor living space(160 m2 for a family of four) offering space forrelaxing and home activities. The work area mayalso seem generous compared with the currentaverage of closer to 10 m2 cap�1 in most officebuildings, but it includes common space in hall-ways, canteens, etc. The energy implications ofproviding this space with suitable temperature and

air exchange are detailed in Section 3.2. We do notconsider it necessary to provide ranges around the60 m2 cap�1; because this is already an average,and because contemporary family structures makelarger areas difficult to manage without help, andsmaller areas increasingly incompatible with typi-cal activity patterns.

3.1.2. Food, health and security. The energy infood (some 120W cap�1 average) has been in-cluded in previous renewable energy scenarios,because of the dual role of biomass in providingboth food and biofuels or simple heat by combus-tion. The general principle of providing food firstand using residues for energy is well established.Only marginal land unsuited for food productionis used for energy crops. This decision is alsoapplied to countries such as Denmark with surplusagricultural land (relative to its own food needs),considering that export of food is necessary inorder to satisfy demands in highly populatedcountries with insufficient agricultural land, andthat Denmark has a long tradition for supportingsuch export industries. The present study thus doesnot deviate from the previous ones, S�rensen et al.[2] for Denmark and S�rensen et al. [6] for theneighbouring countries, with respect to foodproduction and therefore just take over thenumbers derived in these studies. For biofuels,new technology has been forthcoming, whichsuggests a different mix of biofuels to be producedin the present scenarios, as compared with theprevious ones. The associated demands, particu-larly in the transportation sector, will be detailedbelow.

Food storage (refrigeration and cooling) andcooking are taken to imply an end-use energydemand of about 18W cap�1; as in the previousstudies [1]. Heat losses in freezers and refrigeratorsare added in Section 3.2.

Energy use for health includes hot water forpersonal hygiene and household uses, taken as 50l day�1 cap�1 heated to an average level 408Cabove the water supply temperature (an average88C in Denmark), yielding an average energydemand of 97W cap�1: Clothes washing anddrying plus dish washing are estimated at anaverage energy demand of 45W cap�1 [6]. These



DOI: 10.1002/er

numbers include energy spending requirements atthe waterworks for purification and pumping.Health institutions (hospitals, etc.) have heatrequirements assumed to be included in theestimates given above, and their electricity use isassumed to be included in the activity energy usegiven below.

Security needs (police, military and relatedinstitutions) are assumed to be covered throughspace conditioning and activity energy require-ments generally.

3.1.3. Human relations. In this category, electri-city use for lighting and audio/visual/computingequipment in the home or in other socialsurroundings is included, but at different levelsfor the different demand scenarios. The range of85–140W cap�1 was estimated in S�rensen et al.[6], assuming a saturation in energy-demandingleisure activities, using the argument that theseencourage individual relaxation (playing computergames, etc.) rather than a more desirable socialinteraction. This is clearly only one of severalpossible developments. In order to restrict thenumber of end-use demand scenarios from the fiveproposed in Section 2.1, the proposal is here to usethree variants, with energy demands for humanrelations gives as 100, 200 or 300W cap�1: Thefirst one, corresponding to the one used in theprevious study, reflects an energy-conscious im-plementation of desirable social interactions, thesecond being characterized by a larger expansionof activities in the area (making trips to energeticshows and spectacles, etc.), with the third onebeing an implementation of a little controlledscenario with human relations imbedded into veryindividualistic desires to show-off and compete(racing, offering communal car and boat trips,sporting events and competition involving energy-intensive equipment, etc.). For comparison, thecurrent level of end-use electric energy in thissector is below 100W cap�1 in Denmark [6].

There are transportation energy demands asso-ciated with recreational travel and social visits.Typical values for Denmark are of the order of10 000 km y�1 cap�1 [6], composed of many shortvisits (some 100 km roundtrip travel distance) anda few medium and long distance trips, such as in

the latter case taking a vacation in anothercontinent by means of air travel. The spread indemand for non-work-related travel is large,ranging from a few thousand kilometres to severaltens of thousands. The population-averaged de-mand will for the three end-use demand scenariosbe taken as 8000, 16 000 and 24 000 km y�1 cap�1:To this comes work-related travel and commuting,which will be dealt with below.

3.1.4. Human activities including derived ones. Hu-man activities include the acquiring of knowledgeas well as the social activities described above.However, other activities must be added to these,because an effort is required in order to supplybasic needs for food and shelter, as well as for theeducational and social endeavours. This meansestablishing production industries for agriculturalcultivation and processing as well as for theconstruction of equipment, buildings, roads andother infrastructure, which allows the necessaryproduction and distribution of goods and services,which again imply needs for supplying means oftransportation and equipment used by serviceproviders. Furthermore, the chain of derivedrequirements continues with a demand for materi-als and energy, a demand that in the past has beencatered to by mining and processing industries, butrecently has been supplemented by industries forrecycling and providing energy from renewablesources in addition to the traditional woodburning and hydro power plants.

The current net end-use energy used for theseactivities is somewhat less than 100W cap�1 [6].The future energy requirement of the agriculturaland construction sectors is unlikely to changedramatically, while that of manufacturing industryand services will depend on the direction of socialorganization and preferences. Current trends inDenmark have been towards less heavy industry,but this is not necessarily the case for all thecountries in the region. The current mineral-basedresource industry has a high level of energyconsumption, but also the recycling and renewableenergy equipment industries are energy intensive.The international goods manufacturing industryproduces a number of products from furniture tocomputers, of which some are or can be made

B. SØRENSEN448


DOI: 10.1002/er

considerably less energy intensive than the tech-nology being replaced. However, the quantity ofproducts increase, and new products appear on themarket every day. These opposing trends may leadto an end-use energy demand lower than today, ora good deal higher. The scenario end-use energyfor production of food and goods with theassociated materials and equipment industry maythen be taken to require 60, 120 or 180W cap�1:

To this comes transportation of materialsand products. The current exemption of interna-tional transport from taxation has made itextremely cheap to use materials shipped fromfar away, and to build equipment from partstravelling around the globe, sometimes severaltimes (e.g. parts being shipped from Europe to theFar East to have some other parts added, orwherever the cost of labour is lowest, and backagain). Also the transportation distances of thefinal products to the customer have increased, asthe point of production has become less relevantdue to the low transportation energy costs. With-out inexpensive fossil fuels, this pattern ofproduction will become less desirable, and oldvirtues of nearness to the customer may againcome into vogue. The uncertainty of how quicklythis attitude will penetrate the market behaviour ina world with rapidly increasing demand forproducts, e.g. in the large Asian economic growthcountries, makes it necessary to work withscenarios spanning a fairly large range of trans-portation energy demands. The scenarios are3000t km y�1 cap�1 (close to the present level;cf. [2]), 4000 and 5000 t km y�1 cap�1:

Also the amount of business travel (for sales,industry management and knowledge transfer) hasbeen greatly increasing during recent decades,despite the fact that technical options such asvideo conferencing has made it possible toessentially replace all business travel by near-zeroenergy alternatives. To this should be addedcommuting transport of employees, which hasalso been increasing due to abandoning thepreference for settling in homes close to the placeof work. A possible reduction in commuting needscould again technically be accomplished by mak-ing use of new communication technologies, whichallow many types of work to be carried out from a

home office. However, this also implies a change inattitude, after several generations of people havebeen brought up to value a strict division betweenwork and leisure. The scenarios thus work withnon-leisure transportation (of people) demandsincluding commuting based on annual travellingdistances of 10 000 km y�1 cap�1 (current levelequal to current transportation demand for socialrelations; cf. [2]), 40 000 and 70 000 km y�1 cap�1:This should include transportation work forshopping and for services, which in the low-energy-demand scenario would take advantage ofbicycle use (assuming nearness of shopping facil-ities), but in the high-energy-demand scenarioswould assume a centralization of shopping andservice facilities making the use of cars necessary.

One could consider differentiating end-use en-ergy needs between the countries and regionsincluded in the study, such as applying largertransportation needs in sparsely populated areas.However, the principle of not going into too highlevels of details when predicting possible usagepatterns 50 years into the future has spurred thedecision not to work with any regional differentia-tion of energy demands.

3.2. Scenarios for energy delivery to end-users

In Section 2.2, some scenario thoughts onthe energy conversion chains leading from primaryenergy sources to the end-user were presented.The detailed description of conversion lossesdepends on the way the entire energy system isput together and thus must be discussed foreach combined scenario. Of particular importanceis the fraction of energy having to go throughstorage cycles with the associated losses. Thefuture scenarios based on renewable energy willhave different types of losses as comparedwith those of the current system. Losses connectedwith transformation of fuels will occur for biofuelsand for hydrogen, whereas the losses in convertingwind energy into electric energy by wind turbinesis not usually included in the modelling, althoughit does of course influence the economic viabilityof wind turbines. Instead, one usually in this typeof system’s assessment considers wind electricity asdelivered at a specific production cost.



DOI: 10.1002/er

The following subsections estimate the energythat has to be supplied to the end-user in order toprovide the demanded energy service. Otherenergy losses between primary energy and deliveryto end-user will be discussed for each overallscenario. This is facilitated by assuming that thethree scenarios for attitudes towards conversionefficiency discussed in Section 2.2 will simplyfollow the end-use scenarios (now also reducedto three), so that the highest efficiency goestogether with the lowest end-use demand, themiddle efficiency together with the middle end-usedemand, and the laisser-faire conversion efficiencytogether with the high demand growth/run-awaydemand scenario. These three demand-conversionscenarios will be described below under the names‘highest efficiency scenario’, ‘improved efficiencyscenario’ (because the middle scenario assumes acontinued efficiency improvement trend withmodest legislative intervention, as it has materi-alized in the past) and finally the ‘unregulated-efficiency scenario’, named as such because itassumed not only the absence of new legislationaimed at improving efficiency, but also a measureof deregulating areas presently covered by legisla-tion (such as progressive automobile taxationlinked to efficiency).

3.2.1. Highest-efficiency scenario. Biologically ac-ceptable surroundings. In our previous work[1, 2, 6], a fixed relationship between indoortemperature T and space heating or coolingrequirement P (for power) was assumed:

P ¼ c� DT with cbased on assumption A

¼ 36W cap�1 8C�1

The ‘assumption A’ used was that the best currenttechnology would be the average future one, with‘future’ being year 2050, and the actual year wherethis value corresponded to the best technology wasaround 1980, based on detached houses with 25–30 cm mineral wool insulation, double-layeredglazing and high tightness requiring forced (andthus controlled) ventilation. The current besttechnology buildings (in Denmark) have at leastas good insulation, but with better control of coolbridges, and better energy glazing (low-conduc-

tance gas between panes), estimated to lead to avalue (‘assumption B’) of

cbased on assumption B ¼ 24W cap�1 8C

The methodology suggested for the highest-effi-ciency scenario (Section 2.2) was to assume afurther technology improvement to take place and,in the highest-efficiency scenario, become imple-mented by 2060. For this scenario, we shalltherefore assume a further improvement(‘assumption C’) to

cbased on assumption C ¼ 18W cap�1 8C

The use of this coefficient is to require a delivery ofspace heating and cooling at the rates of

Pheating ¼ 18� d � ð168C� TÞ

Pcooling ¼ 18� d � ðT � 248CÞ

as it is assumed that the dependence of the comforttemperature zone on outdoor temperature exhibitsa �48C interval due to the flexible influence ofindoor activities, as influenced by body heatand clothing. The factor d is the populationdensity, people per unit area ðcap m�2Þ and P thusin W m�2.

Figures 1 and 2 show seasonal variations intemperature and calculated space heating and cool-ing energy requirement for the geographical regionstudied. Temperatures used here are monthlyaverages from satellite observations [6, 14]. The moredetailed Danish study [2] uses hourly temperatures toobtain a more accurate value for the space heatingrequirements. Based on the average data, one findsthat except for Southern Germany, there is only aninsignificant requirement for space cooling in thisgeographical region, which does not show in thesetypes of figures, because periods of cooling need areusually much shorter than the month selected foraveraging. However, space heating is important andincreasingly so when moving towards the North(latitude effect) and the East (continental effect). Thepopulation in 2060 is nearly the same as in 2000, forthe countries considered. S�rensen et al. [2] givea detailed model for Denmark of populationdevelopment till 2050. The total population isconstant or slightly diminishing, if strict immigration

B. SØRENSEN450


DOI: 10.1002/er

policies prevail, and there is a modest movementaway from city centres, contrary to the situation indeveloping countries. The actual populations

assumed by 2060 are as in that study, and outsideDenmark as in S�rensen et al. [6]. The populationdensities d are depicted in Figure 3.

Figure 1. Average temperatures in January, April, July and October (based on data from Leemans and Cramer [14]).



DOI: 10.1002/er

Figure 2. Average end-use space heating demand in January, April (top: left, right), July and October (bottom: left,right), for highest-efficiency scenario (left-hand scale for all four maps) or the unregulated-efficiency scenario (right-

hand scale for all four maps).

B. SØRENSEN452


DOI: 10.1002/er

Figure 3. Above (a): Danish population density assumed in 2060 (number of people per 500 m� 500 m unit cell;S�rensen et al. [2]). Below (b): North European population density assumed in 2060 (number of people per km2;

S�rensen et al. [6]).



DOI: 10.1002/er

Food, health and security: Food energy, energyfor food storage and hot water for personalhygiene and for indoor cleaning are basicallythe same as at the end-use levels, with differencesshowing only in the conversion steps leading tothe end-user. However, there may be slightvariations between the end-use scenarios due toemerging technology and changes in habits. Thelatter would be difficult to predict, having to dowith, e.g. gastronomic preferences being linked tofood preparation times and temperatures as wellas to food choices implying different cookingrequirements. For instance, it takes lower energyat the end-user to sustain a diet on sushi thanon artichokes, and it takes more energy to sustaina diet on stews than on rare steaks. Also forhygiene, there are significant energy implicationsof shifting from bathtub use to shower use,and even more by using the low water-usageshowering cabins recently appearing on themarket. For cleaning, water usage is also a keyenergy-determining factor, as is the type ofchemical agents employed. The same is true forclothes washing, where simply stated one has achoice of using low temperatures and strongerdetergents or higher temperatures and simplersoaps. In this case, one may see a balancingissue involving environmental impacts from deter-gents ending up in the sewer systems versus energyuse for water heating. The tendency over the lastcentury has been towards using lower tempera-tures and more chemical aids for washing, butin recent decades, alternative chemical agentshave been introduced with less environmentallyadverse effects and still allowing lower washingtemperatures to be used.

We shall leave the energy use in this sectoridentical for all scenarios, except for a differentia-tion in hot water usage, where the sum of hotwater for hygiene and clothes washing was 142W cap�1 in the old study [2]. Here we shall use anaverage of 100W cap�1 for the high-efficiencyscenario, 142W cap�1 for the intermediate scenar-io and 200W cap�1 for the unregulated-efficiencyscenario, taking the new technology on-board inthe first case, and in the second case increasing hotwater use by, for example, more use of heatedswimming pools.

Human relations and activities: Following Sec-tion 3.2, relations and leisure activities at the end-user translate to 100W�1 cap�1 electricity use and8000 km y�1 cap�1 travel (and some heat energyfor heating venues of leisure activities, assumedincorporated into the building heat use), while theindirect activities for food and equipment produc-tion, distribution and sale, as well as for buildingand infrastructure construction and consequentlymaterials provision, were estimated for the mostefficient scenario to comprise some 60W cap�1 ofmainly high-quality energy (electric or mechanicalenergy), plus 3000 t km y�1 cap�1 of goods trans-portation plus 10 000 km y�1 cap�1 of work-re-lated passenger transportation. The transportationfigures translate into 400W cap�1 average poweruse for person transport and 300W cap�1 forfreight transportation.

Passenger transportation is based on the currentstate-of-the art vehicle (3 l of diesel fuel per100 km) with an average occupancy of 1.5 personson both leisure and business trips. Today, airtransport is about twice as energy intensive perpassenger as transport by road vehicles, but sincethe estimate pertains to a future situation, the highefficiency can technically be reached. In any case,the end-use demand scenario does not includeconversion losses and only uses the vehicleexample above as a proxy for the technicalminimum energy use for the given task, which isused as a proxy for end-use demand in cases whereno clear physical efficiency limit exists. Similarly,the proxy used for end-use requirements for freighttransport is 9 l of diesel fuel per 100 t km. The end-use demands include international transport ofpassengers and freight, which particularly in thelatter case is a major contributor.

3.2.2. Scenario with some efficiency emphasis.Biologically acceptable surroundings. The demandis as for the maximum-efficiency scenario de-scribed above in Section 3.2.1, except that theaverage 2060 building standard only correspondsto the best in 2005:

cbased on assumption B ¼ 24W cap�1 8C�1

This scenario reflects a possible outcome of apolicy like the current one of improving building

B. SØRENSEN454


DOI: 10.1002/er

energy codes with respect to energy at regularintervals, but with fairly modest steps taken.Graphically, the demand is like the one shown inFigures 1–4, because only the overall magnitude ischanged, not the geographical distribution.

Food, health and security: Following the discus-sion in Sections 3.2 and 3.2.1, we use average end-user energy requirements of 120W cap�1 for food,18W cap�1 for food storage and 142W cap�1 forhygiene-related tasks, again assuming small con-tributions for water supply, institutions andsecurity (police, military) as included.

Human relations and activities: Again followingthe assumptions made in Sections 3.2 and3.2.1, the end-use demands for human relations,leisure and all upstream construction, manu-facture and materials supply are 200 ðelectricityÞþ120 (agriculture and industry) þ1244 ðpassenger

transportÞ þ 400 ðfreight transportÞ ¼ 1964W cap�1:The high transportation demand involves globali-zation of the business activities as well as anextrapolation of numbers of vacation trips to othercontinents.

3.2.3. Unregulated-efficiency scenario. Biologicallyacceptable surroundings. The demand is as for themaximum-efficiency scenario described above inSection 3.2.1, except that the average 2060building standard only corresponds to the best in1980:

cbased on assumption A ¼ 36W cap�1 8C�1

Owing to the long life of many buildings, thisscenario is still quite likely in case of noprogressive improvement policy for buildingenergy standards. The best 1980 standard is still

Figure 4. End-use electricity demand in 2060 for the highest-efficiency scenario. Unsubstitutional electricity andelectricity used for convenience (e.g. for dishwashers, industrial furnaces) are included.



DOI: 10.1002/er

three times higher than the average 2000 buildingstandard [2].

Food, health and security: Again following thediscussion in Sections 3.2 and 3.2.1, we use averageend-user energy requirements of 120W cap�1 forfood, 18W cap�1 for food storage and 200W cap�1 for hygiene-related tasks, again assumingsmall contributions for water supply, institutionsand security (police, military) as included.

Human relations and activities: According toassumptions, the end-use demands in the unregu-lated scenario for human relations, leisure and allupstream construction, manufacture and materialssupply are 300 (electricity) þ180 (agricultureand industry) þ2089 (passenger transport) þ 500ðfreight transportÞ ¼ 3069W cap�1: Needless tosay, a significantly increased fraction of people’stime will in this scenario be used for transporta-tion, particular work related. It is one criticismagainst such a scenario that the business economymay decline if more paid time is spent unproduc-tively on travel, but at least when public transpor-tation is used, some of the travel time may be usedproductively. For freight, the figure reflects a life-cycle of products with components travelling backand forth between continents to become processedat the lowest possible labour cost, and to becomeultimately disposed of in the region where this canbe done cheapest. Clearly, this scenario only worksin case the cost of energy remains low comparedwith that of labour.

3.3. Choice of primary energy sources

The general situation for fossil, nuclear andrenewable supply options was briefly discussed inSection 2.3. The present scenario work is exploringthe options for 100% or near-100% coverage byrenewable energy sources, as a continuation of theefforts required by and performed as a result of theofficial Danish energy planning over the last 25years. The maximum energy yield from renewablesources such as wind power, biofuels and solarenergy has been investigated and is estimated inseveral previous studies [1, 2, 5, 6]. The numericalvalues will be discussed here only in connectionwith the individual scenarios in which they areused. Scenarios employing other types of primary

sources than renewable ones have been discussedin S�rensen [4].

3.4. Selecting intermediary energy carriers

Owing to the intermittency of renewable energysources such as wind and solar energy, the systemmust contain compensating carriers of energy,ready to step in when winds are low or when it isdark. The most obvious such energy carrier isbiofuel, because it is already a part of the proposedenergy system. The question is, to which extentthere is enough of it to fill the gaps betweendemand and production at the relevant times, andif the conversions to and from the desired energyforms can be performed without excessive losses.Investigations of the viability of biofuels to servethis purpose will be addressed in connection withone of the scenarios (see Section 3.5.2). Hydrogenis another suggested energy carrier, but one thatrequires substantial infrastructure modifications,depending on its penetration into the differentenergy sectors. The two scenarios described inSections 3.4.1 and 3.4.2 correspond to expecta-tions of either successful development of hydrogentechnologies in all sector applications, or a morelimited success in the introduction of hydrogen fora limited number of specialized applications,notably associated with the energy storage cyclesthat may be employed to deal with the variabilityof some of the renewable energy sources.

The choice of intermediate energy carriers andsystem layout (e.g. more centralized or moredecentralized) has implications for the efficiencyoptions regarding the conversion chain betweenprimary energy supply and end-user. These issueswill be dealt with as part of the construction ofeach scenario on the system level.

3.4.1. Scenario with successful development ofhydrogen technologies. In this scenario, it isassumed that fuel cells, the key technologies forthe employment of hydrogen for both transporta-tion and stationary applications, are successfullydeveloped to both economic and technical viabi-lity. Technical viability comprises performance(where the goal is 50–65% conversion efficiencyfrom hydrogen to electric power and near 100%

B. SØRENSEN456


DOI: 10.1002/er

for the reverse reaction), long lifetime (durabilityof fuel cell stacks rising from the present goal of 5years to around 20 years, like that of theequipment into which fuel cells are to be inte-grated) and low environmental impacts duringmanufacture and use. The economic goals includea competitive price relative to present or futurealternatives (such as diesel engines, gas turbines,etc.) capable of delivering energy in the desiredform and mode, albeit not necessarily usinghydrogen as their fuel. Current fuel cell pricesare an order of magnitude too high, and thelifetime an order of magnitude to short. If onedoes not achieve a 20-year fuel cell durability butonly 5 years (similar to that of technology-similaradvanced batteries), the break-even fuel cell cost isdiminished by a factor of 4. The viability issue maybe eased by insisting on including all life-cyclecosts, which usually is to the favour of hydrogen(as of renewable sources) and to the disfavour offossil fuels and to some extent biofuels.

In this successful hydrogen technology scenario,surpluses of variable renewable energy electricityproduction that cannot be used more profitably inother ways are converted into hydrogen, centrallyor decentrally. The decentralized version mostlikely requires successful commercial developmentof small-scale reversible fuel cells for buildingintegration. Some of this hydrogen would have tobe re-converted into electricity in order to coverperiods with deficit in renewable power produc-tion. However, when feasible, conversion lossesare minimized if the hydrogen can be used as such.This could be in vehicles using fuels cells toproduce power for electric motors, or in directhydrogen use for industrial process energy.

The centralized version of this scenario haspipeline or other distribution of centrally producedhydrogen to the sites where hydrogen is distrib-uted. For example, surplus wind power may beconverted to hydrogen near the wind turbine sites,or at central collection points, which couldconveniently be located where the hydrogenstorage facilities are present (salt intrusion oraquifer storage types as those used today fornatural gas). In the latter case, hydrogen transmis-sion pipelines would transport the hydrogen toautomobile and other vehicle filling stations, as

well as to industrial users. Where existing naturalgas lines can be upgraded to hydrogen quality,supply to individual building owners of fuel cellunits could also take place. Alternatively, if thebuilding-integrated fuel cells are reversible, theinput could be confined to electric power linetransfer of electricity, both for direct use and forhydrogen production and filling into vehiclesparked at the building. One possibility is to totallydispose of central hydrogen storage, if building-integrated storage types would become technicallyand economically feasible [2]. The advantage ofthe decentralized fuel cell placement is that thewaste heat can be used to cover the building’s heatrequirements. This is possible despite the highpower efficiency (and hence lower waste heatgeneration) of fuel cells, provided that the buildingheating needs are reduced by improved insulationstandards. Hot water needs would have to becovered in any case, which is totally within thecapability of fuel cells rated to cover the electricityneeds of the building. If there is a deficit of wasteheat for space heating during winter, this may becovered by electric heating, preferably throughheat pumps, as in the 2001 scenario. Precisematching between hydrogen stored in buildingsand the electric power demands during low-windconditions may not be feasible for each building,but this does not matter due to the electric gridconnection between buildings, as demonstrated bytime simulations in S�rensen et al. [2], however,only for the middle demand scenario.

In the present work, hydrogen storage isweighed against international power exchange.This implies that the market price of power soldin a given hour will determine if export iseconomically advantageous to producing hydro-gen for storage. Similarly, the import power pricewill determine whether to draw from the hydrogenstore (i.e. to re-generate electricity with theentailed additional energy loss) or to cover thedemand by imported electricity. In more sophisti-cated versions of this balancing, the degree offilling in the hydrogen stores could be consideredand held up against the expected cost developmenton the import–export market over a period oftime, in order to determine, if it might pay to useenergy from the stores or to hold it back for more



DOI: 10.1002/er

profitable usage later. Clearly, this type ofcalculation is time consuming as it requiresforward calculation at each time step, and it willonly be done for selected periods in order toillustrate the nature of this particular dispatchproblem.

If the re-generation of electric power duringperiods of insufficient renewable energy supply isperformed centrally, the fuel cell technologyutilized is likely that of high-temperature fuelcells, due to their superior conversion efficiency,while low-temperature fuel cells are the obviouschoice for building and vehicle applications, whereestablishment of operating temperatures of theorder of 8008C is considered inconvenient by mostindependent hydrogen scientists.

3.4.2. Scenario assuming failure in the developmentof fuel cell technologies. The no-fuel cell scenarioemerges if the technical or cost break-even targetsof fuel cells (low or high temperature) cannot bemet. The need to make up for fluctuating renew-able energy sources still exists, but must now besolved either by biofuels or by hydrogen in a pureenergy storage function.

The biofuel option involves converting biofuelto electricity, whenever fluctuating renewablesources are insufficient and import options areunavailable or more expensive than biofuel con-version. There will be a conflict between these usesof biofuels and the use as a vehicle fuel, which isdifficult to substitute. An assessment will be madeof whether the biomass resources are sufficient tosustain both functions. Surplus electricity wouldstill have to be exported, as conversion intobiofuels is not an option.

Such problems may be dealt with by conversioninto hydrogen for storage. Without new viable fuelcell technology the conversion of electricity intohydrogen for storage would have to be byconventional (alkaline) electrolysis. The regenera-tion of electric power in periods of insufficientdirect supply will in this case have to be made inconventional units such as gas turbines (but nowfuelled by hydrogen). The efficiency will be lowerthan if viable fuel cell technology were available,but not unacceptably so. In this scenario, biofuelsand hydrogen would both be available for uses in

the transportation sector, but hydrogen wouldserve the important task of filling supply–demandgaps in power production.

3.5. Constructing combined energy system scenarios

In this section, we combine the demand, storage,conversion and supply scenarios into completesystem scenarios. These are labelled according tothe technology characteristics of the alternativesdiscussed in Section 3.4, i.e. successful develop-ment of fuel cell technologies, hydrogen used onlyfor storage, and dealing with fluctuating produc-tion mainly through international power ex-change.

3.5.1. Complete renewable energy-hydrogen sce-nario. A completely renewable energy scenariowith the use of hydrogen as a major energy carrierand storable energy form will be constructedfor the following combinations of subsystemscenarios:

(I) Highest-efficiency demand and conversionscenarios}successful fuel cell scenario.

(II) Improved efficiency demand and conversionscenarios}successful fuel cell scenario.

(III) Unregulated demand and conversion effici-ency scenarios}successful fuel cell scenario.

From the earlier work, we know that scenario Iand probably II can be realized. For scenario IIIwith much higher demands on renewable resourcesthe aim is to investigate if such a scenario is at allpossible.

A discussion will be made of the possibilities fordecentralization based upon reversible fuel celltechnology. The preliminary investigation of thisoption in S�rensen et al. [2] can be furtherdiscussed at present due to the advancing com-mercial efforts to develop precisely the solutionsenvisaged in the 2001 scenario.

3.5.2. Renewable energy plus limited hydrogenenergy for storage scenario. The restriction ofhydrogen to storage uses with either re-generationof electricity and possibly minor industrial directuses of hydrogen makes the central question forthese scenarios one of whether international power

B. SØRENSEN458


DOI: 10.1002/er

exchange plus the use of biofuels in the transpor-tation sector can sustain the demand scenariosconsidered. Again, this will be investigated forthree scenarios corresponding to the ones listed inSection 3.5.1, but without fuel cells:

(IV) Highest-efficiency demand and conversionscenarios}no fuel cells; hydrogen storage.

(V) Improved efficiency demand and conversionscenarios}no fuel cells; hydrogen storage.

(VI) Unregulated demand and conversion effi-ciency scenarios}no fuel cells; hydrogenstorage.

As in Section 3.5.1, the viability of scenarios isquestionable, and probably even more so in thescenarios considered in this section, not havingavailable the high-efficiency fuel cells or theconvenient two-way conversion options offeredby the reversible fuel cells.

3.5.3. Renewable energy scenario with trade repla-cing storage. In these scenarios, internationalpower exchange is used whenever economical toavoid using the (lossy) storage cycle. The neigh-bouring countries, with which power is exchanged,are assumed to have gone through a transition torenewable energy sources like Denmark, althoughnot necessarily with the same mix of renewableenergy resources (in Norway, more hydro energy,in Sweden and Finland, more forestry-basedbiomass and in Germany, both wind and photo-voltaic energy).

These types of scenarios might be interestingboth in case of successful fuel cell development andin case of failure. It does not mean that hydrogenstorage should be completely avoided, but it mightbe expected that the storage demand will be lessthan in the Section 3.5.2 scenarios. However, this isby no means certain, as the maximum storagedemand could well turn out to be unaltered,because the time of such storage demand maycoincide with a moment where trade rules areunfavourable or where the neighbouring countriesdo not have any surplus electricity to sell. This isparticular likely if similar energy sources areemployed in the collaborating countries, because,e.g. wind deficits over demand would likely happensimultaneously throughout the region.

In practice, we will run all scenarios includingan option for power exchange, and thus the sixscenarios described in Sections 3.5.1 and 3.5.2 willbe performed for a geographical area extendedfrom Denmark to neighbouring countries. With-out the import/export possibility, the six scenarioscould have been simulated for Denmark inisolation. However, there is a question regardingthe rules of trade between the countries. Currentpower exchange consists of a part governed byfixed contracts, and another part traded on what iscalled the Nordic Power Pool, a bidding andexchange system allowing competitive trade in anumber of categories, ranging from forwardoptions over 24-h market bidding to a short-termspot market operated by a power balancing agentcharged with maintaining the stability of thesystem under conditions of technical problems,failure of parties to fulfil contractual agreementsor unexpected demand variations. The tools totake care of these tasks include a list of additionalbids from power producers willing to supplyadditional quantities of power to the spot market.Furthermore, a renewable energy-based systemwould like to add fluctuations in power productionto the list of problems to be attended to by thebalancing agent, and one would likely want tomodify the auction rules, e.g. by altering theperiod of time for which bidding is required andhence for which reliable production forecasts haveto be available.

In an earlier project, the frequency of incorrectforecasts in a 100% wind-based Danish systemwas estimated [8, 9]. The wind power averageproduction forecasts were found highly reliable forperiods of a few hours, they still have reasonablyhigh accuracy for 24-h forecasts (based on eithermeteorology or non-linear trend forecasting), butthen deteriorates quickly as the period of weatherfront passage times (typically some 3–7 days) isapproached. For the present trade rules requiring36-h forecast, we found that about 25% of the bidswere in error, but only 5% to such an extent that itcaused significant economic penalties with therules of the current pool system. The present studywill explore the effect of introducing trading rulesmore favourable to renewable energy systems,such as those that are already implemented in



DOI: 10.1002/er

some parts of the world (e.g. allowing bids anddecisions on bids to be made continuously). Thus,while the simulations outlined in Sections 3.5.1and 3.5.2 will be made with current pool rulesdeciding whether power exchange of hydrogenstorage is to be invoked, we will make anadditional set of simulations with new rulesdesigned to benefit renewable energy systems (inall the trading countries) rather than conventionalfuel-based ones:

(VII) Highest-efficiency demand and conversionscenarios}fuel cells; new pool trading rules.

(VIII) Unregulated demand and conversion effi-ciency scenarios}fuel cells; new pool trad-ing rules.

(IX) Highest-efficiency demand and conversionscenarios}no fuel cells; new pool tradingrules.

(X) Unregulated demand and conversion effi-ciency scenarios}no fuel cells; new pooltrading rules.

An adjacent article is carrying through thesimulation of a middle variant of the spread ofscenarios described above [15].

4. RENEWABLE RESOURCE MAPPING

An appraisal of renewable energy resources maybe found in S�rensen [1]. For off-shore wind, anew method of assessing the potential has beendeveloped. It will be described below. Solar, hydroand biomass assessments are essentially un-changed, except that off-shore biomass productionis an additional potential source considered here.It will also be described below.

4.1. Wind energy

Wind energy is the dominating renewable energyform in the future Danish energy system and isalso expected to play a significant role in othercountries with a high wind ocean coastline. InS�rensen et al. [2], the data used for simulation ofthe future Danish wind system were scaled upfrom current production, for which hourly datawere available. The scaling was done separately for

on-shore and off-shore wind, based on two year-2000 time series, each of which merged from theseparate time series published by the East- andWest-Danish power utilities. Because the windconditions are better and also more stable overwater, the two time series had somewhat differentcharacteristics. Added to this, there is the differ-ence originating from the fact that most land-based turbines are constructed to give maximumenergy over the year, whereas the current Danishoff-shore turbines use blade profiles yielding lessthan maximum energy over the year, but incompensation have considerably more productivehours over the year. This is already the case due tothe more persistent winds over sea, but wasemphasized by the manufacturers having em-ployed power curves peaking at lower wind speedsthan would have been dictated by a maximumenergy production [1]. Apart from this, the off-shore wind data were very suited for extrapolationto a large penetration of off-shore wind, becausethey were already based on turbines with hubheights of 50–70m. Because the data are fromactual production figures over a year with somenew wind turbine construction being commis-sioned and also decommissioning of old machines,on-shore, there is some question regarding theprecise validity of using the data as a proxy forfuture demand over a year.

For the on-shore data, one could further objectthat because they are a mixture of output fromolder and newer turbines, the data would containthe effects of some very small wind turbines builtover the past 30 years, turbines that wouldexperience winds in much lower heights thancontemporary and future, larger turbines. Theassumption in the scenario was that the totalnumber of land-based turbines would remain as itis today, but that older machines would graduallybe replaced by new ones of 2–4MW unit size. Still,the extrapolation of current production data islikely to be fairly reliable, because the oldermachines, despite large numbers, contribute fairlymodestly to the total production, which is thendominated by modern turbines much more similarto the ones envisaged for the future situation. Inany case, any increase in the overall Danishnumber of turbines was assumed to involve

B. SØRENSEN460


DOI: 10.1002/er

off-shore wind parks; hence, the off-shorecontribution to the hourly time series would inthe scenario future be much more important thanthe on-shore extrapolations.

The scenarios developed for the present projectassume that neighbouring countries, with whichDenmark exchanges power, expand their renew-able resource utilization including wind energy,and data for future wind production in all thesecountries are therefore required. One of the keyquestions asked is, whether the lulls and peaks inwind power production occur during the sameperiod in all countries, in which case wind powerexchange would not be able to contribute tosmoothing the effects of the variability in eachregion. It is well known that the passage ofweather front systems, and thereby also windpower production, is similar for regions of lineardimensions of the order of 500 km [1]. Because thedistance between, e.g. Denmark and the Northernparts of Scandinavia is more like 1500 km, onewould expect that some levelling could be accom-plished by export and import between theseregions. Furthermore, the neighbouring countrieswould not necessarily have wind as a dominatingrenewable energy source (as Denmark is supposedto have in the scenario future), but would be ableto offer power smoothing based on hydro andperhaps other renewable sources such as biomass.These possibilities will be investigated, and there-fore a certain amount of wind power will beassumed for each of the neighbouring countries,necessitating the construction of time series ofwind power production throughout the region ofScandinavia and Germany. However, the methodspresented below are applicable for any part of theWorld.

For another previous study (preliminary reportin [16]), wind data for the Nordic countries werecollected by Holttinen [17]. For Denmark, she alsoused total production data from the electricutilities (however, it seems, omitting the availableoff-shore data), while for the other Nordiccountries, time series from a very limited numberof operating turbines were used. In Norway andFinland, these were all located at coastal sites,while for Sweden, two inland sites at the far Northwere also included. For two additional Finnish

locations, meteorological data were used toestimate potential wind turbine output. The datafor these few sites with existing wind turbines werethen extrapolated to much higher levels of windenergy use in the countries involved. One mayquestion if this is a fair representation of possiblefuture wind energy use. In Denmark, coastal sitesare specifically excluded from wind turbine erec-tion due to environmental legislation. Similarlegislation may not be in place in the other Nordiccountries, but if expansion of wind power shouldmaterialize, it is likely that similar restrictions onthese recreationally important sites would beimposed. The inland sites in Norway, Swedenand Finland are all characterized by fairly lowwinds, due to shadowing from the Norwegianmountain ranges and due to the high roughnesscreated by the forest-covered areas in Sweden andFinland [1]. It would therefore seem more realisticto assume that any substantial use of wind energyin these countries will become based upon off-shore locations, where the wind conditions aremuch more favourable.

These considerations have spurred the presentproject to find alternative sources of wind energydata particularly suited for estimating off-shorepotentials. Two potential methods appear to beavailable. One is the wind atlas method, wherelimited data on geostrophic winds at a fewlatitude–longitude points are combined with sur-face roughness data to predict the production ofwind turbines of a given hub height [18]. Thismethod is mainly aimed at estimating monthly orannual production from a wind turbine erected ata proposed location, which is not sufficient for thecurrent project. The other potential method is todepart from the reanalysis of global measuredwind data, using circulation models to improveconsistency in areas of few measuring stations [19]and the model calculations to assess heightvariations up to the top of the atmosphere. Inorder to explore short-term fluctuations, addi-tional satellite measurements from suitable instru-ments can be used. The reanalysis data have untilrecently only been available at a very coursespatial resolution (some 250 km) and time resolu-tion (one month). Such data were previously usedto represent gross geographical variations in



DOI: 10.1002/er

potential wind production, using an interpolationbetween the two lowest atmospheric pressurelayers to represent data at a height of roughly70m [1]. Recent efforts have lowered the resolutionand have added new high-resolution satellitescatterometer-data over sea areas and blendedthe two kinds of data using novel reanalysismethods to obtain a time resolution of 6 h and aspatial resolution of 0:58 (56 km latitudinal widthmultiplied by the cosine to the latitude angle forlongitudinal width). The aim of these efforts hasbeen to study ocean–atmosphere interactions [20],and it is by no means obvious that they could beused to estimate wind power production. Becausethe satellite passes over a particular latitude–longitude location only a few times a day, theidea is to overlay the previous gross reanalysis datawith the high-resolution satellite data and extra-polate these to areas between the trails of satellitepassage, so that the final mixed data set containsthe high-frequency behaviour everywhere (at leastover water), but with correct time sequences onlyfor the (moving) locations of the satellite over itstrajectory. These new types of data would seemparticular interesting for off-shore wind estima-tion, as the ship and buoy data available are veryscarce.

It may seem a daring project to attempt to usethese data for short-term wind energy calculationsaimed at studying energy storage requirements,and the approach was adopted only after con-ducting a pilot study for on- and off-shorelocations in Denmark and the Netherlands duringa year with known wind power production at anumber of sites, and comparing these actual datawith the satellite data analysis, ideally for the sameperiod and the same sites [21]. Figures 5 and 6show the results of such a comparison, for alocation at 11:58E longitude, 55:08N latitude andcompared with measurements at Vindeby off-shorewind farm [1]. It was not possible to obtain datafor the same years as those covered by the off-shore scatterometer and on-shore reanalysis data;hence, only the overall impression of temporalvariations could be derived. A more preciseevaluation was possible for sites in the Nether-lands, where data for the exact same period in year2000 could be compared. These are presented in

Figures 7–9. It is seen, that the frequency ofvariations in power produced is very well repro-duced, although minor deviations between calcu-lated and measured data exist during the periodswhere the satellite did not pass over the Nether-lands area.

For one off-shore site, there is measured data ata height appropriate for modern wind turbines.This allows a comparison between model and datanot requiring the use of wind profile scaling for themeasured data but only for the blended scatte-rometer-reanalysis derived wind production. Theresult of this comparison is shown in Figure 7. Theimplication is that the sea surface scatterometerdata correspond to an effective height of about 3mover the average water surface.

Figure 8 shows a similar comparison forIjmuiden on the Western shore of Holland. TheDutch data are here measured at low height(18.5m), and the question of extrapolation towind turbine hub height (assumed to be 60–70m)has to be studied. A scaling factor is determined bystandard methods, assuming a profile correspond-ing to the roughness of the local surface (which ismeasured at all the Dutch locations). This scalingfactor was found to lie in a narrow region of 1.3–1.4 for the near-shore and off-shore sites studied.Theoretically, the scaling factor depends on bothstability of the air and on surface roughness(see Figure 10) [1], and the scaling factors usedcorrespond to typical values for a neutral atmo-sphere and a mesoscale (rather than strictly local)roughness length.

Figure 9 shows the comparison for an interiorlocation, on the border of Holland and Germany.Here the blended data are derived entirely fromreanalysis, and since no normalization has beenperformed, the effective height may not be thesame as for the scatterometer data over water.Because the reanalysis is supposed to reproduce10m measured data at selected stations, one wouldexpect the standard model of scaling to departfrom this height. However, this is not true.Because the Dutch data include measurements ofroughness, the scaling of measured data can beextended from off-shore to on-shore sites using thesame neutral scaling law. Because of the higherroughness over land, the scaling factors get larger,

B. SØRENSEN462


DOI: 10.1002/er

in addition to varying from place to place. Tryingto fit monthly wind production means derivedfrom the reanalysis to the Dutch data, it becomes

evident that the best scaling factor for thereanalysis data is about unity. This is surprising,but in accord with the questions regarding the

Figure 5. Measured time series of wind power output at the off-shore Danish location Vindeby during January 1995(top) and corresponding power calculated from scatterometer blended data for January 2000 (bottom). The time

resolution is 6 h [21].



DOI: 10.1002/er

Figure 6. Measured time series of wind power output at the off-shore Danish location Vindeby during July1995 (top) and corresponding power calculated from scatterometer blended data for July 2000 (bottom).

The time resolution is 6 h [21].

B. SØRENSEN464


DOI: 10.1002/er

interpretation of the reanalysis data asked, e.g. byMilliff et al. [23] and by Chelton and Freilich [24].The point is that even the new circulation modelcalculation with a mesh of some 50 km does not

have a spatial accuracy of more that 4–6 times thisdimension, and one should therefore not besurprised that the effects of local roughness arelost. On the other hand, the behaviour at larger

Figure 7. Measured time series of wind power output at the off-shore Dutch location Station XX during January 2000(top; [22]) and corresponding power calculated from scatterometer blended data for January 2000 (bottom; the

calculated average power is 346 Wm�2, very close to the measured one). The time resolution is 6 h [21].



DOI: 10.1002/er

heights (the circulation models uses some 100levels vertically through the atmosphere) is muchmore likely to be realistic, and the resolution of the

problem may simply be to regard the lowestlevel results as more representative for altitudesof some 40–80m above ground, i.e. exactly where

Figure 8. Measured time series of wind power output at the Dutch location Ijmuiden during January 2000 (top; [22])and corresponding power calculated from scatterometer blended data for January 2000 (bottom, scaling 1.0). The time

resolution is 6 h [21].

B. SØRENSEN466


DOI: 10.1002/er

the wind turbine hubs would typically be placed(cf. [21]).

Figure 11 shows the wind turbine potentialproduction map for Northern Europe obtainedfor 1.3-scaling. As mentioned, this scaling is

giving the best agreement with measurementsfor off-shore locations, but on land the properscaling varies on scales smaller than that ofthe map, and actual values could deviate by some30% (or of course more in case of particular

Figure 9. Measured time series of wind power output at the Dutch location Twente during January 2000(top; [22]) and corresponding power calculated from scatterometer blended data for January 2000

(bottom). The time resolution is 6 h [21].



DOI: 10.1002/er

obstacles to wind flow, such as cities or otherstructures).

4.2. Biofuels

The biofuel assessment is based on a biomass netproduction model [25] adapted to assess energyvalues [1]. Figure 12 shows the net biomassproduction for the North European region. Theagricultural residues are used to produce ethanolor bio-diesels at an assumed conversion efficiencyof 45%, while the forestry residues are used formethanol production at an assumed efficiency of50%. The consideration of aquaculture in thepresent study is new. Figure 13 shows the fractionof the grid cells used in the geographical coveragethat contain water surfaces. These comprise water-ways and lakes inland, as well as off-shore waters,to a distance of about 20 km from the shore.Because the Nordic countries have long coastlines,there is considerable potential for off-shore aqua-culture. In contrast to inland waterway aquacul-ture, this might be dedicated energy productionareas, assumed to be environmentally protected

Figure 10. Wind speed profiles for three types of atmo-spheric stability. The parameter L used to describe thenon-neutral curves is called the Monin–Obukhov length

(from S�rensen [1]).

Figure 11. Map of wind resources in Northern Europe, based on blended data model with a scaling factorof 1.3 and a power conversion curve typical of current wind turbines [21]. The unit is annual average

Wm�2 of swept turbine area.

B. SØRENSEN468


DOI: 10.1002/er

Figure 12. Potential net biomass production in Northern Europe, based on the model described inS�rensen [1]. The unit is Wm�2 of land.

Figure 13. The water fraction of each cell in the model geographical grid. Inland values contain lakes,rivers and other streams, while off-shore values basically indicate the areas of up to 20 km from the

coastline, which potentially could be utilized for aquaculture.



DOI: 10.1002/er

from interfering with the biology of open oceanwaters. Identification of the most suited plants oralgae for ocean farming aimed at fuel productionhas not been done (despite some work onhydrogen production from algae, cf. [4]); hence,this is an option lying some years (decades) ahead.

5. APPLICATION OF THE SCENARIO DATA

An example of using the data described in thepreceding section for the simulation of an entirefuture energy system is given in an accompanyingarticle [15].

ACKNOWLEDGEMENTS

Support from the Energy Research Programme of theDanish Energy Agency (EFP-033001/33031-0021) isacknowledged.

REFERENCES

1. S�rensen B. Renewable Energy (3rd edn). Elsevier Aca-demic Press: Burlington, 2004; 946pp.

2. S�rensen B, Petersen A, Juhl C, Ravn H, S�ndergren C,Simonsen P, J�rgensen K, Nielsen L, Larsen H,Morthorst P,Schleisner L, S�rensen F, Petersen T. Project report toDanish Energy Agency (in Danish): Scenarier for samletudnyttelse af brint som energibærer i Danmarks fremtidigeenergi-system, IMFUFA Texts No. 390, 2001; 226, RoskildeUniversity; report download site: http://energy.ruc.dk.

3. S�rensen B, Petersen A, Juhl C, Ravn H, S�ndergren C,Simonsen P, J�rgensen K, Nielsen L, Larsen H, MorthorstP, Schleisner L, S�rensen F, Petersen T. Hydrogen as anenergy carrier: scenarios for future use of hydrogen in theDanish energy system. International Journal of HydrogenEnergy 2004; 29:23–32.

4. S�rensen B. Hydrogen and Fuel Cells. Elsevier AcademicPress: Burlington, 2005; 464pp.

5. S�rensen B, Meibom P. A global renewable energyscenario. International Journal of Global Energy Issues2000; 13(1–3):196–276, based on Ref. 11.

6. S�rensen B, Kuemmel B, Meibom P. Long-term scenariosfor global energy demand and supply. IMFUFA Texts No.359, Roskilde University, 1999; 217pp.

7. S�rensen B. A combined wind and hydro power system.Energy Policy 1981; 51–55.

8. Meibom P, Svendsen T, S�rensen B. Import/eksport-politiksom redskab til optimeret udnyttelse af el produceret paVE-anlæg. IMFUFA Texts No. 343, 1997; 84pp.

9. Meibom P, Svendsen T, S�rensen B. Trading wind in ahydro-dominated power pool system. International Journalof Sustainable Development 1999; 2:458–483.

10. Kuemmel B, Nielsen S, S�rensen B. Life-Cycle Analysisof Energy Systems. Roskilde University Press: Roskilde,1997.

11. S�rensen B, Meibom P. A global renewable energyscenario. IMFUFA Texts No. 354, Roskilde University,1998; 112pp.

12. S�rensen B. Energy and resources. Science 1975; 189:255–260.

13. S�rensen B, Meibom P, Karlsson K, Nielsen L, Petersen A,Lindboe H. Comparative assessment of hydrogen storageand international electricity trade for a Danish energysystem with wind power and hydrogen/fuel cell technolo-gies. Project of the Danish Energy Research Programme,2007, in progress.

14. Leemans R, Cramer W. IMAGE2 climate database.Global Change Scenarios of the 21st Century, Alcamo J,Leemans R, Kreileman E (eds). Pergamon: Oxford, 1998.(Available from http://www.pik-potsdam.de/public/climate)(last accessed 1999).

15. S�rensen B. A renewable energy scenario for NorthernEurope. International Journal of Energy Research 2007;DOI: 10.1002/er.1376.

16. Meibom P, Morthorst P, Nielsen L, Weber C, Sander K,Swider D, Ravn H. Power system models. Deliverable3.2 for the EC Wilmar project. Ris� Report R-1441 (EN),2003.

17. Holttinen H. Hourly wind power variations and theirimpact on the Nordic power system operation. Thesis,Helsinki University of Technology, 2004.

18. Mortensen N, Landsberg I, Troen I. Wind Atlas Analysisand Application Program (WasP), Risoe National Labora-tory, Roskilde, Denmark, 1993.

19. Kalnay E et al. The NCEP/NCAR 40-year reanalysisproject. Bulletin of the American Meteorological Society1996; March issue.

20. Chelton D, Schlax M, Freilich M, Milliff R. Satellitemeasurements reveal persistent small-scale features inocean winds. Science 2004; 303:978–983.

21. S�rensen B. Wind energy futures for Europe. Inter-Academy Council International Workshop on Energy, Paris,2006, submitted.

22. KNMI. HYDRA Project, Royal Dutch MeteorologicalInstitute, Data available from website: http://www.knmi.nl/samenw/hydra, 2006.

23. Milliff R, Large W, Morzel J, Danabasoglu G, Chin T.Ocean general circulation model sensitivity to forcing fromscatterometer winds. Journal of Geophysical Research 1999;104:11337–11358.

24. Chelton D, Freilich M. Monthly Weather Review, Februaryissue, 409-429. Data available at the QSCAT/NCEP CoRAblended wind data website: http://dss.ucar.edu/datasets/ds744.4/data, 2004.

25. Melillo J, McGuire A, Kicklighter D, MooreIII B,Vorosmarty C, Schloss A. Global climate change andterrestrial net primary production. Nature 1993; 363:234–240.

B. SØRENSEN470


DOI: 10.1002/er

a sustainable energy future: construction of demand and renewable energy supply scenarios

Documents