I MPACT OF ENERGY SERVICES DEMAND PROJECTIONS ON FINAL ENERGY

- For water heating, and since both the ownership of equipment and the range of temperature variations is not wide, the highest uncertainty is on the assumptions on social structure (i.e. family size) of a household and on the consumer behavior (water use and number of days of consumption), with an impact that could vary from -25% to 70% in the energy service demand of 2050, when compared to REF.

- For the remaining end-uses, there is a pattern, where in general high variations in the specific energy needs related to changes in consumer behavior have a stronger impact in the ESD than the penetration rate of equipment. This holds true especially for the end-uses where the ownership of equipment is already high. Rosa-Flores and Gálvez (2010) also indicates technology penetration and utilization of equipment as very important variables affecting home energy use.

uses, the role of technology may significantly overweight behavioral practices and socio-economic changes.

The decreasing trend results for heating energy consumption until 2050 are consistent with those presented in other studies (e.g. Aguiar et al. (2002), Olonscheck et al. (2011) and Rosenberg et al.

(2012)). Straightforward comparisons are not possible since timeframes and baseline assumptions are not the same as in Aguiar et al. (2002), and diverse geographical and climate conditions are different as in Olonscheck et al. (2011) and Rosenberg et al. (2012). However, it should be considered that the Portuguese households have historically different characteristics regarding insulation measures that affect the thermal comfort levels. As demonstrated by the current energy per capita consumption, comparing to both EU₂₇ average and with countries with similar climate conditions, Portugal still has lower levels of consumption as well as thermal comfort levels inside households. This supports our results that despite future increasing demand for space heating energy services, this might outcome in a small decrease of final energy consumption in 2050 compared to 2005.

We find that energy consumption for cooling will keep increasing despite technological energy efficiency improvements. Young and Steemers (2011) also stated that the domestic cooling likely increase may undermine energy efficiency strategies related to the improvement of building design and fabric, and environmental systems for lighting, heating and cooling. Our findings for air conditioning evolution for Portugal (i.e. Mediterranean climate) are similar with the ones presented by others authors for similar climate conditions: Xu et al. (2012) for California identified that cooling energy consumption increase and heating energy consumption decrease over the next 100 years; Pilli- Sihvola et al. (2010) presented results for Southern Europe, where increases in cooling outweighed decrease in heating.

Concerning final energy demand for DHW, it is expected an increase since the energy services will be guaranteed with the use of less efficient renewable energy sources like boilers running on biomass pellets and solar collectors. Final energy growth rates for Other Electric equipment as refrigeration, kitchen appliances and small electric equipment suffer a consumption reduction due to efficiency improvements (e.g. refrigerators Class A or higher). Large appliances targeted by the EU Directives on labeling and the EU mandatory efficiency standards account for a decreasing share of the total consumption. This goes in line with Ruijven et al. (2010) findings, where technology improvement correlates negatively with future energy use. Smaller electric equipment (e.g. computers) raises their final energy consumption as a result of increasing equipment ownership, offsetting expected energy efficiency improvements. For these latest end-uses, and besides the likely continued improvements on energy efficiency, the policy focus and monitoring should be also devoted for consumers’ behavior.

Brounen et al. (2012) in their results also showed that the variation in residential energy consumption is a function of both technical characteristics of the dwelling and the composition and background of a household.

Figure 4.4 – Comparison between the evolution of the demand for energy services and final energy between 2005 and 2050 for REF

4.3.2.2 Sensitivity analysis

Twenty ESD scenarios corresponding to the highest and lowest impact from each parameter for each end-use were considered as input for TIMES_PT model to assess its impact on final energy consumption. The results for each end-use show that the uncertainty on the assumptions behind the parameters is mostly tilted for higher levels of consumption, as illustrated by Figure 4.5. Heating and other electric equipment are the end-uses showing stronger impacts on final energy consumption, due to their high share in the total Portuguese residential consumption, and the high uncertainty range of the parameters (i.e. thermal comfort levels and OE growth rates). For DHW the impact is more relevant on lower levels of consumptions, while for lighting is the opposite. In cooking the uncertainty is equally distributed to higher and lower consumption. Finally, despite a strong increase in the final energy consumption for cooling observed for REF, the impact of uncertainty range of parameters is not wide.

Figure 4.5 – Final energy consumption range between the REF and the Highest and Lowest variation scenario of each end-use in 2050

Sensitivity analysis results also indicate that extreme variations over energy services parameters from the REF could have a strong impact on both energy production and demand supply systems, namely by overestimating the production which leads to unnecessary economic costs or underestimated demand that induce problems regarding security of supply.

For illustration purposes, the energy consumption for heating is expected to decrease -0.2% per annum in the REF, while, for the High space heating demand scenario (T5) (+84%) the growth rate is around 1.5%/year. This trend has an impact on fuel consumption in 2050 compared to the REF:

electricity (≈+10%; 8 PJ); natural gas (≈+45%; 20 PJ) and oil products (≈+100%; 4 PJ) explained by increased adoption of heat pumps, electric heaters and natural gas and oil boilers. In the other hand, a Low space heating scenario (T1) (-47%), fallout in a consumption reduction of -2.3% per annum, with decrease use of electricity (-7%; 72 to 67 PJ), natural gas (-23%; 44 to 34 PJ) and oil products (-42%) in 2050 compared to the REF. No disruptive technological and fuel changes are expected to accomplish the ESD uncertainty range.

For lighting, results estimate for a REF scenario a reduction of -0.51% per annum of the final energy consumption. In a Low lighting demand scenario (-40%), the reduction is expected to be sharper (- 1.63%/year) with a 3% reduction of electricity consumption in 2050 compared to REF. In a High lighting scenario (+158%), the electricity consumption is expected to increase 10% in 2050 comparatively to REF, with an annual growth rate of 1.6% per annum for the period under analysis.

Fouquet and Pearson (2011) also considers that long run trends and the relative stability of the price elasticity estimates for lighting suggest that increased efficiency is likely to be eroded by rising income levels. This idea falls better in our High scenario assumptions where the number of light bulbs per household and the hours of use are higher.

Our results for both energy services and final energy consumption (REF and sensitivity analysis) support our assumption that improved knowledge on ESD drivers help to identify where policies should tackle to foster effective energy reduction - both energy efficiency increase as well as the drivers behind ESD, especially the ones more related to consumer behavior. As stated by Haas et al.

(2008) it is necessary a shift from historical trend where efficiency improvements are lower than the increase in service demand.