*4.1. Survey Data Standardisation*

As can be observed, Figure 2 shows the ratio between valid and non-valid data, as well as the absence of any answer for the energy consumption, GHG emission and RES production data given by the cities, which was standardised as previously explained in Section 3.1. Regarding energy consumption, the highest uncertainty corresponds to the case of liquid fuels, where only 37% of the cities provided reliable data. In the case of coal, all of the received answers were considered valid, since the consumption of this fuel is very low and most of the cities reported no consumption. Apart from coal, natural gas consumption data present the highest reliability, with 63% of the answers considered valid.

Regarding GHG emissions, 59% of the cities provided reliable data. As for RES installation and production, we observe a better knowledge in the case of electric RES (48% of answers valid) as compared to thermal, where only 41% of the cities provided an answer, with a validity rate of 33%.

With respect to the total energy consumption per capita, Figure 3 shows the data that we were able to collect. Note that, in order to calculate this total, only the final consumptions of electricity, natural gas, liquid fuels and coal were included in the survey. This, in some cases, resulted in a gap between the total real energy consumption of a city and the total energy consumption calculated in this assessment. As a further check of the validity of the results, the direct energy consumption data obtained from the surveys is compared with the national total primary energy supply (TPES) per capita of the corresponding countries in Figure 2. The ratios between these two range from 26% to 95%, meaning that this standardisation process has allowed us to map a 49% average TPES per inhabitant of a country. The gaps, on the one hand, may be due to the fact that a specific city is being compared to its national average reality. On the other hand, the energy losses in transformation (in order to produce

electricity and heat from coal, gas or biofuels and waste), the fuels employed for non-energy uses (crude oil for asphalt and oil products used in agriculture and chemical industry) and other fuels such as biofuels (biomass, biogas, bioethanol), butane gas or waste were not considered in the survey, which certainly led to differences.

**Figure 2.** Data standardisation.

**Figure 3.** The total energy consumption by inhabitant computed for the selected cities (by mPOWER project) and the national reality reflected by the International Energy Agency averages. The goal in energy reduction of 17.5 MWh·cap−1·yr−<sup>1</sup> has been indicated [30].

Another difficulty in the calculation of the total energy consumption was the case of those cities with significant production by Combined Heat and Power (CHP) plants, commonly fed by natural gas. In CHPs, energy consumption is measured, such as the consumed amount of natural gas, but similarly the electricity produced by the CHPs is also taken into account when energy consumed at homes is measured. Thus, this could generate small amount of double accounting that this project has not been

able to correct. For future research, a specific question could be included to understand the energy production from CHPs, and thus avoid this double accounting.

From this standardisation we obtained the data to be used in the evaluation of the results (Section 3.2), which is listed in Table A2 from the Appendix B.

#### *4.2. Survey Data Assessment*

Figure 4 shows some of the indicators related to energy democracy: the staff working on energy and transition, the municipal energy or transition campaigns, and the budget dedicated to energy transition (shown in Table A3 from Appendix B). These indicators are a measure of the resources dedicated to energy and transition by enrolled authorities at both the technical and the social level. With some exceptions, such as Barcelona, Frankfurt, Horst aan de Maas, Vienna and Zenica, most of the municipalities have from 0 to 20 employees working on energy and energy transition issues. Similarly, excluding Nis, Pamplona and Rijeka, most of the cities have from 5 to 15 annual campaigns. In the annual budget for the energy transition, the values differ much more. Whereas most of the cities dedicate several thousands of euros (within a broad range from 20,000 € to 792,000 €), some of the cities dedicate millions of euros to the energy transition. Such is the case of Mizil (1 M€) or Frankfurt (1.8 M€). The highest budgets correspond to Vila Nova de Gaia (5 M€), Manchester (7 M€) and Amsterdam (87.5 M€), which were left out of the Figure in order to make the rest of the cities visible. It is important to note that those budgets often depend on external projects (European or national, for instance), that make it difficult to define a fixed and constant annual budget. Some cities also pointed out the difficulty to define a budget solely related to energy transition, since this is normally spread out over the overall budget of the city.

**Figure 4.** Annual budget, number of campaigns led by the city councils and number of people working on energy transition.

Regarding the total energy consumption per capita, and going back to Figure 3, we can observe how energy consumption differs among the cities analysed. Thus, Zenica (Bosnia and Herzegovina) is the city with the lowest consumption per capita (6 MWh·cap−1·yr−1) while in Horst aan de Maas (Netherlands) the consumption is as high as 48 MWh·cap−1·yr<sup>−</sup>1, meaning that Zenica consumes 87.5% less energy than Horst aan de Maas. Comparing it to the previously cited target of 17.5 Mwh per person and year [30], 17 out of 27 cities do reach the target. From the ones that do not reach it, the total energy consumption of Frankfurt (Germany), Vienna (Austria), Aradippou (Cyprus), Nottingham (United Kingdom) and Metz (France) is especially high. Alternatively, when national TPES data are taken into account, most of the values are higher than the total energy consumption mapped by mPOWER. For instance, the values change to 22 MWh·cap−1·yr−<sup>1</sup> in the case of Zenica and 51 MWh·cap−1·yr−<sup>1</sup> in

the case of Horst aan de Maas, leading to a 57% smaller energy consumption in the former. In this case, all the consumptions are above the target level, meaning that none of the municipalities is able to reach it.

Figure 5 depicts the distribution of the municipal energy consumption by type of fuel, taking into account the national electric mix as that of the municipality. That assumption was made due to a lack of data on the RES production of some of the municipalities, based on the fact that the electricity consumed in each city is supplied by the national grid. This data changes considerably from city to city. It can be observed that in northern countries like Tampere (Finland) and Vaxjo (Sweden), around a 50% of the whole consumed energy is in the form of electricity. Other countries, such as the Netherlands (Amsterdam, Horst aan de Maas) and Austria (Vienna) present a stronger dependence on natural gas. It needs to be clarified that in Horst aan de Maas the high consumption of natural gas is due to the massive use of heated greenhouses for intensive vegetable production [47]. Finally, southern and eastern cities, such as Pamplona and Cadiz (Spain), Zenica or Krizevci (Croatia) make, in proportion, a higher use of liquid fuels, but this could be due to the generalised use of other fuels, such as biofuels or butane gas (instead of natural gas) that was not taken into account in the calculations. In Figure 5, the electric consumption has been disaggregated by source, according to the national electricity mix, in order to obtain the RES percentage of the total municipal energy consumption. It can be seen that only Vaxjo is above the target of 27% of renewable energy from the total energy consumption, with a 27.1% share. In the rest of the municipalities, renewable energies cover a maximum of 25.8% (Tampere) and a minimum of 1.5% (Horst aan de Maas) of the total energy consumed. Northern countries (Finland, Denmark and Austria) and Croatia show particularly high RES percentages. However, we have observed that a high renewable electricity mix (in green in Figure 5) does not assure low fossil fuel consumption. On the other hand, a lower energy consumption does not assure a high renewable share, i.e., none of the cities that are able to reduce their fossil fuel consumption to below 10 MWh·cap−1·yr−<sup>1</sup> (Aradippou -Cyprus-, Barcelona, Krizevci, Mizil -Romania-, Nis -Serbia-, Vila Nova de Gaia -Portugaland Zenica) have an integration of renewable energy above 17.3%.

**Figure 5.** Energy consumption by fuel within the direct and indirect (in electric vector) energy consumptions by municipality (note that the total consumption of cities obtained from mPOWER survey differs, as shown in Figure 3, from the national average data provided by International Energy Agency (IEA)).

In order to complement the information given in Figure 5 (taking into account the differences shown in Figure 3), Figure 6 has been created, where sectorial national direct energy consumption averages have been included from IEA balances [24], and national hidden energy flows (HEF) have also been added from previous analyses carried out by the authors [11]. In order to take into account the displacements that the impacts related to the consumption of the citizens generate elsewhere, energy embodied in imported products and services in each country have been included using the latest data from the year 2014 to obtain the difference between total primary energy footprint (TPEF) and TPES per capita at a national level, adding it to each city (HEF = TPEF/TPES). The calculations have been developed using global multi-regional input-output (GMRIO) methodology, and data are available in Appendix B, Table A4. The accuracy of these calculations could be improved with municipal hidden energy flows data instead of national average ones, but this would require an input-output analysis at a local level. Although a methodology for local input-output analysis is currently being developed by Cazcarro et al. [48] as well as by our team [49], it is currently beyond the scope of this paper.

This last figure shows how only a small percentage of the energy consumption, between 9 and 24%, is consumed by private households in terms of electricity and heat (green numbers in Figure 5), whereas from 76% to 91% is not consumed in the residential sector (imported and national products and services, transportation needs for humans and trade, and transformation and distribution losses of energy).

**Figure 6.** National energy consumptions by sector.

Figure 7 shows the relation between energy consumption (in red bars) and GHG emissions (in blue and green). The red bars correspond to the total per capita energy consumption calculated by mPOWER. The blue line corresponds to the per capita GHG emission data obtained from the survey and standardised using the IEA national values (as explained in Section 3.1). Finally, the green line corresponds to the GHG emissions calculated from the total energy consumption mapped from the surveys, taking into account the GHG emission intensity of each fuel given by the IPCC (tonnes of CO2eq per kWh), as well as the emission intensity of the national electricity mix. This way, when the IEA standardised values and the IPCC estimated values are on the same order of magnitude, it can be regarded as a further check of the validity of the results. We consider that both GHG emission values are on the same order of magnitude when the IEA standardised values are within the IPCC estimated error bars, which cover a range from 50 to 150% of the IPCC estimated value. It can be observed in general trends that energy consumption and the corresponding CO2eq emissions are related.

Even if the general trend is that the higher energy consumption, the higher the emissions, there are some exceptions, such as Litomerice, Manchester, Nottingham and Pamplona, which have a low IEA standardised GHG emission despite their high energy consumption. This can be explained by the use of different energy sources (a high natural gas and RES rate in the case of Manchester, Nottingham and Pamplona) or the uncertainty found in the data (low GHG emission data in the case of Litomerice). In some other cities, such as Frankfurt, Nis and Tampere, the opposite relationship is observed: the IEA standardised GHG emission data is high compared to the total energy consumption. This could be due to an overestimation of the GHG emissions or due to an underestimation of the energy consumption. In this last case, as well as in the case of Litomerice, we know the inconsistencies between IEA standardised GHG emissions and energy consumption are due to uncertainties in the data (and not due to the use of different energy sources), because the GHG emissions estimated from the IPCC intensities and those standardised with the IEA values are not on the same order of magnitude.

**Figure 7.** Energy consumption per capita by municipality, and the respective greenhouse gas (GHG) emissions obtained from the survey (IEA standardised) and calculated from the energy consumption and the IPCC emission intensities.

Finally, in order to understand the energy consumption differences in all the 27 cities, the total energy consumption detected by mPOWER has been related to the achieved national HDI [50] and national GDP [51], and also to physical conditions like the climate or the size of each city. All data were fitted to the equation "y = A ln(x) + B" following the methodology presented by Steinberger et al. [46], Arto et al. [10] and Akizu et al. [11]. Note that Horst aan de Maas was left out of the analysis because a large part of its energy consumption is used for industrial agriculture, making it difficult to correlate with the rest of the cities. This phenomenon could also occur in cities with a high presence of industrial production, but our results for rest of the analysed cities have not shown alterations as significant as those detected in Horst aan de Mass, so it has not been taken into account for the rest of the cities.

Figure 8 gives us a comparison of the consumed energy and the benefits obtained from it, using HDI and GDP indicators as they are the most commonly used in this respect [10,11]. The former, more related with human behaviours, allows us to understand how energy can affect education, life expectancy and economy, and the latter only focuses on national economy. The general trends among analysed cities show that life quality standards, measured in national HDI (Figure 8a) and GDP (Figure 8b), are directly related to consumed energy. Nevertheless, it can be observed that some cities can achieve high standards of living with markedly low energy (such as Vaxjo, Frankfurt and Frederikshavn). It must be noted that since HDI and GDP data are national averages, they are not fully sensitive to the realities of the cities, and it would be helpful to include city data for both indicators in the future.

**Figure 8.** Energy consumption of each city compared to the corresponding national Human Development Index (HDI) (**a**) and the national Gross Domestic Product (GDP) (**b**).

In Figure 9a, the climate of each city was taken into account by using the heating degree day (HDD) and cooling degree day (CDD) factors [52]. In the analysed cities, according to the obtained fitting and R2, there is a low correlation between the energy consumption and the HDD plus CDD, as shown in Figure 9a. Cities such as Zenica or Krizevci, with a high heating need, have a very low energy consumption per capita, whereas cities like Amsterdam and Pamplona have a higher energy consumption with a lower heating and cooling requirement. This could also be related to the difference in GDP. Figure 8b compares the size of each city (measured in inhabitants) and the energy consumption per capita, with the previous hypothesis that bigger cities might be more efficient than the smaller ones. However, Figure 9b shows that this assumption does not correspond to reality. Small cities like Aradippou show low energy consumption, whereas big cities such as Amsterdam are not especially efficient because of their large number of inhabitants.

**Figure 9.** Energy consumption of each city, and the corresponding heating and cooling need according to the climate of each city (**a**), and the corresponding inhabitants (**b**).



**Table**

**1.**

#### *Energies* **2020** , *13*, 1315

#### **5. Conclusions**

In relation to the difficulties in collecting municipal energy data, we have several considerations that may be relevant for the specific goals of the project and to be taken into account when boosting the general energy transition in Europe. The low quality of the data gathered by the municipalities (sometimes literally non-existent) is a clear sign of public disempowerment in energy issues. Despite the high effort of each municipality, there are not enough public up-to-date municipal data at the European level and the energy sector is mostly owned by private companies that manage the information according to their own interests [53]. Therefore, this project reflects how significant it is to have real energy consumption and production data in order to lead an energy transition at city level.

Nowadays, participation in different energy transition initiatives, such as the Covenant of Mayors, and thus, the sharing of energy consumption and GHG emission data with citizens is a voluntary act. However, in some regions, such as the Basque Country [54], the publication of data related to energy consumption is starting to become compulsory for their cities and villages. We claim this kind of law to provide citizens with information could boost the incoming energy transition. This knowledge could facilitate evaluating energy policies, analysing the real needs of each city, or creating roadmaps and energy plans for incoming sustainable energy transitions. We expect that this project will help in revealing these kinds of issues, enabling initiatives such as the creation of public databases, and empowering local public institutions in the management of low-carbon energy transition. Similarly, it is important to spread the know-how of the current energy reality among citizens and other agents in cities in order to boost citizen-led initiatives. As an example, and as an alternative to private energy management, renewable cooperatives could be an opportunity to start a transition towards a democratic and sustainable energy system [55–57].

Going back to the data collection, it has been observed that the mPOWER baseline survey has allowed us to map an average of 49% of the energy consumption per capita (in comparison with national average total primary energy supply values offered by the IEA) consumed by citizens. The remaining 52% is mainly due to transformation losses, and also, to a lesser extent, due to the lack of integration of non-common energy vectors such as biomass (like firewood or biofuels), butane gas, waste use for energy purposes and fuel oil boilers. In some cases, the use of CHPs could also generate alterations in results because of the double accountability they tend to cause when taking into account the gas they consume, but the electricity produced in homes is also taken into account.

The obtained results reveal that European cities still present a strong dependence on fossil fuels, ranging from 72% of fossil fuels in the total energy mix mapped in this paper to 98.4%. The low percentage of electricity in energy supply is also noteworthy, where renewable generation is still generally minimal. In addition, because of the small percentage of the total (renewable and non-renewable) energy that is consumed by private homes in the form of electricity and heat, it must be underlined that energy transition not only needs to be focused on the energy consumption of private dwellings, but it especially needs to challenge the current model of products and goods consumption. In this sense, this work provides a view of how consumers have the potential to improve the national energy system, partaking in shared responsibility with governments [58].

The comparison among the 27 cities analysed clearly relates the achieved human development index and gross domestic product in a city to the consumed energy, showing a dependence on energy consumption to maintain the current living standards, and improve them. However, some cities already show that they can achieve high GDP and HDI values with relatively low energy consumption (as Vaxjo and Plymouth); hence these cities should be taken as a reference. On the other hand, we can see that the climate and size of a city do not positively or negatively affect the energy efficiency, and thus, this gives various types of cities the opportunity to think about different strategies to improve their own energy systems.

Finally, in relation to the mPOWER project, this baseline has been shown as an effective tool to obtain an initial picture of the energy situation of different European cities. The results of the baseline and of this paper will help to improve the development of the project and can encourage participant cities to identify the above-mentioned key obstacles and information gaps in the transition to a participative low-carbon energy system.

**Author Contributions:** Conceptualization E.V., O.A.-G. and L.U.; Data curation E.V., O.A.-G. and O.A.; Formal analysis E.V. and O.A.-G.; Investigation E.V., O.A.-G., O.A., L.U., A.C.-C., I.B. and I.B.H.; Methodology E.V., O.A.-G. and O.A.; Writing—original draft E.V. and O.A.-G.; Writing—review & editing O.A., L.U., A.C.-C., I.B. and I.B.H. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the European Union's H2020 Research and Innovation programme under grant agreement 785171—mPOWER project. The research has been supported by 'Ekopol: Iraunkortasunerako Bideak' research group, recognised by the Basque Government (IT-1365-19) and the University of the Basque Country (GIC-18/22).

**Acknowledgments:** The authors would like to recognise the effort made by the technicians of the 27 analysed cities, during the data gathering process and also all the mPOWER research team from the University of Glasgow, Platform-London, Stitching Transnational Institute, The Society for the Reduction of Carbon Limited, Institute for Political Ecology and Energy Cities (Energy-Cities Association).

**Conflicts of Interest:** The authors declare no conflict of interest.
