Technology and behavior options screened

There is substantial modelling available on how a low emission future might be accomplished. In the IPCC scenario systematics these are the Relevant Concentration Pathways for a 2-degrees world: the RCP2.6 scenarios, see the extensive analysis and survey of scenario studies by (van Vuuren, Stehfest et al. 2011). These scenarios are peak-and-decline scenarios, first over-emitting relative to the 2-degrees climate goal till around 2050, requiring negative emissions thereafter.

The models used are global integrated assessment models with feedback between economy and environment and with regional disaggregation, and they include non-CO2 emissions and effects of land use change (LULUCF: Land Use, Land Use Change and Forestry). For the instrument discussion on Europe with focus on non-LULUCF CO2 emissions, three types of simpler models have been developed in CECILIA2050, giving more insight in mechanisms involved. They are an energy optimization model (Solano and Drummond 2014), the ETM-UCL Model; an instrument rich macro-economic model (Meyer, Meyer et al. 2014) the GINFORS Model; and a technology specific input-output scenario for 2050 (Koning, Huppes et al. 2014, De Koning, Huppes et al.

2015), the EXIOBASE Input-Output Model. The two economic modelling approches have a predictive element, assuming (abstract) technologies with certain characteristics and costs. With adequate input assumptionis the ETM-UCL can produce the desired emission reductions by 2050, with the Ginfors model coming only close, and the EXIOBASE Input-Output model not even coming close. The input-output scenario is free of dynamic mechanisms but looks at consistent future in the year 2050. The latter assumes high increases in efficiency and adds substantial

emission reducing technologies. But deeper changes are required. The outcomes of these three complementary approaches are analyzed in (Drummond 2014).

Emission reductions will have to come mainly from weaning off of fossil fuels. The other option is CCS, which is difficult to apply in mobile and small scale sources but may be applied to larger scale fixed installations. There it may play a role but seems not capable even at the micro level to reach the 95% emission reduction in electricity production as investigated here, see Table 1, indicating reduction options there in the order of 65% (in terms of greenhouse gases but referring mainly to CO2). The IPCC numbers in that table include upstream emissions.

Table 1 Global warming potential (GWP) per MWh of primary electricity sources

Technology Min Median Max

Currently commercially available technologies

Coal – PC 740 820 910

Gas – combined cycle 410 490 650

Solar PV – utility scale 18 48 180

Solar PV – rooftop 26 41 60

Concentrated solar power 8.8 27 63

Geothermal 6.0 38 79

Hydropower 1.0 24 2200

Wind offshore 8.0 12 35

Wind onshore 7.0 11 56

Nuclear 3.7 12 110

Pre‐commercial technologies

CCS – Coal – PC 190 220 250

CCS – Coal – IGCC 170 200 230

CCS – Gas- combined cycle 94 170 340

CCS – Coal-oxyfuel 100 160 200

Ocean (tidal and wave)) 5.6 17 28

Source (IPCC-WG3 (2014) Annex 3, p10)

These upstream emissions will become lower by 2050 as then also many of these emissions will have been reduced, see on this point the conceptually innovative paper by (Hertwich, Gibon et al. 2015). However, upstream infrastructure involves steel and cement, where emission reductions will be limited. So the role of CCS may be supplementary, for difficult cases where weaning off is not an acceptable option, like with steel. Also, the public perception of CCS is mixed, see the survey of the extensive literature by (L׳Orange Seigo, Dohle et al. 2014) and the underlying problem of where long term storage may be proven safe.

So, apart from the quite uncertain nuclear expansion, see the balanced view by (Lévêque 2014)it is only the renewables that can transform the energy domain towards deep emission reductions:

solar, wind, tidal, geothermal and hydro, and maybe a few more. Hydro-energy can be expanded

only to a limited extent in Europe, tidal has limited and unproven capacity, and geothermal may be restricted to hot beds close to the surface, so the resilience bet is on wind and solar, which have proven to be expandable at industrial scale. Rolling them out with adequate speed is one of the great challenges. Quantifications based on optimization models, see for a detailed example (Solano and Drummond 2014), tend to be bounded, with all technologies each playing a moderate role. Such outcomes are roughly indicative of options, as they assume cost prices which long term are not well known and market mechanisms which are highly conditional on policy instrumentation.

The IEA gives prospective costs of electricity supply in terms of levelized cost of electricity (LCOE), for 2020. The levelized cost concept has been developed for (US-type) highly regulated fossils based monopolistic electricity sector and has limited meaning in an open market situation with low short term variable costs of most renewables. In that old situation electricity prices are an input rather than an output of investors’ profitability calculations. To assess whether the cash flow of a new project is sufficient to reimburse the investment and capital costs the net present value (NPV) or internal rate of return is used, based on expected exogenous electricity prices, see (IEA 2015) Chapter 11 for a discussion. More relevant measures include system costs, always to be borne somehow, and take into account expected price developments, their variability, and their uncertainty over time, including policy induced effects though all sorts of market mechanisms. Predicting technologies long term therefore seems not well feasible, unless induced by assumed effective technology specific policies.

Reduction of energy use through energy efficiency increases plays a limited role in deep emission reduction in the long term, as opposed to the short and medium term, for structural reasons and due to policy limitations. By then all energy flows will mainly be emission free. Then using 40%

less solar PV hardly reduces emissions. With 5% of emissions left in the electricity system by 2050, an extreme energy reduction of 40% will help reduce emissions from 5% to 3%. That is a hardly relevant reduction but with a severe burden on society: energy is an essential ingredient in virtually all economic processes. There is a Herculean task ahead in climate policy. Confounding this task with an as yet not well-defined task in energy policy seems overcharging the political and economic system. Some have argued that energy efficiency improvement is a main technology entry for emission reduction. There is the ‘huge potential for improving energy efficiency’ (Grubb, Hourcade et al. 2014) p160 and similar (Hood 2011) Ch2 on ‘unused energy efficiency’. However, there is also is a huge potential to increase the energy intensity of production and consumption, partly fed by the rebound mechanisms of lowered costs of final energy use due efficiency increases. Faster transport alone can accommodate near endless increases in energy consumption, from fast trains and long haul normal aviation to supersonic flight and space tourism. That is a matter of energy prices and energy policy guiding technology, while the focus here is on having an economy with substantially reduced CO2 emissions. If supersonic flight is possible with near zero emissions, that reduction then is enough for climate policy. If climate policy kills off that option, tant pis for supersonic flight. But from a climate point of view there is nothing against supersonic flight as such.

Of course in the short and medium term, the overlap between energy and climate policy is environmentally and politically highly relevant, see (Hood 2011, Bausch, Roberts et al. 2014).

There are conflicting elements however. For supply security reasons diversification of fossils supply may create a barrier for effective climate policy. Also, adding energy efficiency

requirements as through standards may substantially increase the cost of emission reduction as compared to generic emission reduction policy, see the survey paper by (Parry, Evans et al. 2014) on the transport domain, and may help create lock-ins on long term detrimental technologies, like efficient combustion cars. If rising energy prices due to climate policy lead to a reduction of energy use, inducing somewhat lower emissions due to the volume effect, such an effect is of course fine from a climate policy point of view. But reducing specific fossil energy technologies will also lead to lower fossils prices, see (Meyer, Meyer et al. 2014), stimulating fossil energy use diffusely in all applications. Energy efficiency increases are always part of technology development but tend to be slow. They have proven to be difficult to speed up with policy measures, see (Tietenberg 2009), with emission pricing emerging as a major policy instrument.

Whatever the precise role of energy efficiency may be, also in a medium term instrument mix for climate policy, it is not part of the long term strategy delivering 90% emission reduction.

So for climate reasons, some mix of renewables, unknown as yet as to composition and volume, will have to take over, together with an also unknown volume of nuclear energy (with limited dispatchability) and some biobased energy as from agriculture. Biofuels are widely disputed as to their net effects on climate emissions, a discussion we will not go into, also related to land-use change. Their overall role will be limited, due to competition with food and materials for an increasingly wealthy global population. Solar and wind seem to have the option for the extreme capacity increase required, solar especially if printable-like mass production becomes available, integrated in infrastructure as in wall cladding and roof constructions and other surfaces that are to be built and cladded anyway. Somehow their intermittency is to be resolved, not only per day and week but also over the seasons, if heating and cooling would be substantially electricity based. Another option might be solar hydrogen or other solar fuel. We cannot know that in advance but policy instrumentation is to reckon with such options: by actively creating one option, or several, or by incentivizing any of them, including the provision of the required infrastructure.

The bottleneck for increasing renewables production hardly seems technical. A back-of-the-envelope quantification may indicate the size of the task. EU final energy use currently is around 8 petawatt-hour, of which now 3PWh is electric. Efficiency of electricity use as for heating can be a factor 3 higher than for natural gas, and similar efficiency gains are possible in person transport when shifting from fossil to electric drives. Primary electricity production in 2050 may be estimated at around 3TWh. At 300kWh per m² per year an area of solar cells then is required of 10 000km², with required land a factor of two higher at 20 000km2, around 0.5% of the EU28 land surface, with North Africa waiting. So the land requirement for shifting even to full only-PV production hardly seems a problem. The area of solar cells to be produced per year, assuming 5%

of the installed capacity in 2050, is 500km² per year³. Assuming some sort of printable or coating process to emerge, that surface can be compared to car coating. This PV surface is less than the surface for car coating. At 12.5 million cars per year with a multi-layered coated surface of on average 80m2 each (EGTEI 2005), the total coating surface is around 1000km² per year⁴. Again, the amount of solar cells to be produced would hardly pose a problem, certainly not if also

3 At 300kWh/m²/year a PV surface of 10 000km² delivers 3PWh/year. This requires a solar cell production rising to 500km²/year, assuming a life time of 20 years.

4 Current sales of 12.5 million cars at 80m² per car is 1000 million square meters, which is 1000km².

reckoning with wind, hydro, nuclear power, etc. Even this extreme PV-only option seems feasible.

Also the intermittency problem does not seem basic, as it might be solved using the capacity of the transport fleet alone, see BOX 1.

Whatever the exact role of PV and other sources may be, a major part of long term new electricity production will be highly intermittent, not linked to intermittency of demand. There are many ways in which supply and demand can be matched, real time clearance being a strict requirement in the electricity market. Shifting demand, storage and secondary production are the solutions, with many options available. Current swing producers, who now match supply to varying demand, will have left the market well before 2050, so the full problem is at the table. Again, however, even current technologies can already take care of market clearance over days at least: battery storage, hydro storage, secondary production as with fuel cells together can do so, and certainly in combination with smart grid developments to shave demand peaks, including with heat pumps with heat storage creating substantial flexibility over days, weeks and seasons. The IEA assumes in a highly assumption based study that a number of such technologies together may reduce peak demand by in the order of 30%, see (Heinen, Elzinga et al. 2011). At the peak supply side, just the electric drive passenger car park of 2050, with battery storage or with hydrogen storage with fuel cells, is more than enough to produce for any total daily peak demand, see BOX 1. This of course is not a prediction but goes to show that several technical solutions for load balancing are available in required quantities in principle. Climate policy instrumentation will determine which ones may come up, with innovation playing a key role, intermingled.

BOX 1 Integrating passenger vehicles in the electricity system

Beware: This is not a prediction but an investigation of one corner of feasible technology options relevant for climate policy instrumentation.

Peak load supply by passenger car fuel cells or car batteries

With primary electricity coming substantially from wind, solar PV and other intermittent sources there will be a regular non-match with peak demand for electricity, and a severe one at exceptional times, as when the sun does not shine (at night), solar PV is covered with snow, and when wind power is at a fraction of its installed capacity (stable high pressure over Europe) while electricity use is high due to extreme cold or extreme heat with high air conditioner use. Meeting peak demand is a key issue, to be always resolved.

There are several ways to resolve this problem as by centralized peak shaving, also as by using stronger East-West connections, and by decentralized peak shaving using the options of the internet of things as one line in smart grid development. The highly decentralized electricity system which may develop under the umbrella of an open real time electricity market can go many directions, not yet to be predicted. So there are solutions possible in many directions. But can they be sufficient? A simple technology scenario can help grasp the magnitude of feasible change, by picking out just one arbitrary but feasible option:

hydrogen fuel cells of cars connected to the grid when standing idle. Transport vehicles tend to be used more hours per day, but may have a similar role when idle.

Current peak demand in the EU is in the order of 800GW (Brauner, D’Haeseleer et al. 2013). With shifts towards electricity use in heating and transport this may rise to in the order of 3000GW, a very high estimate according to (Brauner, D’Haeseleer et al. 2013). Now imagine, exemplary, that all person cars and similar will be hydrogen fuel cell driven⁵. This is hardly probable but some form of non-fossil onboard

5 Hydrogen refilling stations are being rolled out by Shell in Germany now, 400 planned, with other countries following, see http://www.ft.com/intl/fastft/406871/shell-aims-boost-hydrogen-vehicles-germany. Toyota is

stored energy will be present in transport to drive the electromotors for the wheels. An average car will have a fuel cell power of at least 50kW, with some additional battery power for peak acceleration. (Tesla Model S has 310kW peak power now.) Assume there is an incentive to connect the fuel cell car to the grid, when the price is high enough to make this attractive for the individual car owner, or for the private fleet owner. This price must cover the cost of the hydrogen to be acquired, the additional cost of running the fuel cell, the cost of connecting to the grid, and the cost of the nuisance to go to this effort. Assume that half of all cars are connected at peak times. And assume there are only 100 million fuel cell vehicles, a very low estimate for the total number of passenger cars in the EU. Their combined power then is: 0.5 x 100.10⁶ x 50kW = 2500GW, roughly the total peak power demand for 2050, the high estimate. Overall, the efficiency of this hydrogen route would be relatively low, assuming the hydrogen is produced from renewable electricity. With a to-and-fro efficiency rising to 80% by then this would lead to an overall efficiency of 60%, to be paid for by the temporary high electricity price. However, there may be more attractive options for hydrogen production by then, like directly from solar (research for different technologies ongoing). Full electric cars would have a higher overall efficiency in delivering to the grid. But they can deliver a smaller amount if the gap between primary production and demand lasts. Many more options for secondary peak delivery exist, like pumped hydro, diverse static battery systems, and more esoteric systems like fly wheels, chemical storage, pressurized gases, etc. And of course there will not just be primary production with non-dispatchable renewables.

With relevant markets installed, covering the daily peak power problem of intermittent sources might be resolved with the car hydrogen fuel cell technology route alone, or by the battery car alone. Of course the future will be different, including as yet unknown options. The longer term mismatch between supply and demand - over weeks and seasons and in exceptional circumstances - may also be covered by the fuel cell car with hydrogen storage; not by the battery electric vehicle.

With the energy supply system mainly electrified, the next technical problem is how the electrification of the energy system is to work downstream, in applications formerly served by fossil electricity or fossil heat as with district heat from fossil fired power stations or dedicates fossil heating systems. For heating purposes, including hot water use, electricity based heat pumps with heat storage seem a most relevant option now available already. There is substantial experience with such systems. However, especially the seasonal energy storage may not have the right total capacity yet. Underground heat and cold storage in aquifers, for example, uses a larger underground area than the building area served above ground. Also ownership of the underground capacity is not well established yet, with first come first get a permit currently being the situation in the Netherlands. If such systems don’t develop, heat pumps may lack seasonal storage capacity and then may revert to air based heat pumps, as is the most used heating system now in Japan already. Substantial R&D and Demonstration seems due.

In industry, there is much diversity in energy use. With low temperature heat use the solutions are similar or even the same as for space heating and hot tap water. For higher temperature furnaces deeper adaptations may be required, requiring redesign and new investment in capital goods. The instrument mix will have to give the right impulses for such changes.

currently marketing its Mirai fuel cell car. How this will work out long term is unknown but learning curves will certainly be created.