**1. Introduction**

An increasing number of cities and local governments adhere to transnational initiatives that are active on climate change mitigation. Cities that adhere to transnational networks on climate change by making emission inventories and climate action plans publicly available, although in the absence of obligation, render themselves accountable both globally as well as locally [1]. Their performance and identity are increasingly scrutinized in terms of global impact and exploited in the scientific literature [2]. This includes factors influencing the cities' participation in the networks and multilevel governance models that include observed [3–12] drivers influencing the emissions and target setting [13–16], tools and strategies for the redaction of the climate action plans [17–22] and benchmarking methods [23–26]. The factors further involve the assessments of the global contribution of local climate mitigation actions [27–34].

In the European Union (EU), local authorities (LA) and cities have a crucial role in building public support for the EU's energy and climate goals, while deploying more decentralized and integrated energy systems. The Covenant of Mayors (CoM) is an EU based initiative, which started in 2008. The CoM has been a disrupting phenomenon in the arena of transnational initiatives, which have expanded tremendously over the past 10 years [35], covering more than 7,850 local authorities and 252 million inhabitants as of December 2018. The CoM at the time of writing the paper is part of the

Global Covenant of Mayors initiative, while this paper specifically addresses the experience gathered in the CoM in Europe between 2008 and 2018.

When LAs join the initiative, the CoM signatories (all of whom participate voluntarily) commit to reduce the levels of carbon dioxide (CO2) emissions in their territories by at least 20% by 2020 or at least 40% by 2030 through the implementation of a Sustainable Energy Action Plan (SEAP). Recently, actions on adaptation (climate risk assessment) have also been included in addition to those on mitigation. The combined plans are called Sustainable Energy and Climate Action Plans (SECAPs). In this paper, the focus is on climate mitigation action plans with commitment targets for 2020, i.e., the SEAPs. The CoM is a unique feature of multilevel polycentric governance that goes far beyond transnational city networking [11]. The initiative is supported by the European Commission and managed jointly by the Covenant of Mayors' Office (CoMO), a consortium of cities' networks and the Commission's Joint Research Centre (JRC), which provides scientific and methodological support. The latest overall assessment of the initiative by JRC [36] shows that the signatories' overall commitment to reducing greenhouse gas (GHG) emissions is 27% by 2020, i.e., 7 percentage points above the minimum requested target of 20%. Based on data from 315 implementation reports accompanied by a monitoring emissions inventory (MEI) (covering 25.5 million inhabitants and mainly for the period 2012–2014), a 23% overall reduction in emissions is observed to be already achieved.

The CoM reporting framework requires a three-step approach to the signatories: i) submission of emission inventories according to their standards, ii) setting a mitigation target as well as drawing a climate action plan, and lastly, iii) monitoring the progress towards the targets. The minimum requirement includes all the direct emissions that are produced within geographical boundaries (buildings and urban transport sectors), as well as indirect emissions associated with the final consumption of grid electricity and of heating/cooling networks. The local generation of energy and associated direct emissions are not part of the activity sectors that are included in the emission inventory but are considered in the calculation of the local emission factors to be applied to the consumption of grid supplied energy [37].

While accounting of direct emissions sources generally follows a coherent approach to those of the Intergovernmental Panel on Climate Change (IPCC), indirect emissions accounting is more complex and challenging as it requires methods on assessing the average emission factor of grid supplied energy. Grid supplied energy can take the form of electricity and district heating/cooling carriers. The identification of the emission sources of heat/cold production could be straightforward due to the limited number of generation units fueling the networks and certainly because heat can be transported effectively only for short distances from generation units to the users, compared to electricity. Therefore, in defining the criteria for the classification of local energy generation units, the focus herein is only on electricity, and the CoM methodology regarding accounting for the indirect emissions associated with grid supplied energy consumptions in cities is introduced.

Further within the method, a broader policy framework in which local action is taken is discussed based on the EU energy and climate policies. Over 80 exemplary good practices are overviewed across the technology areas of local energy generation ranging from photovoltaics, solar thermal, wind energy, hydroelectric power, bioenergy, geothermal energy, combined heat and power (CHP), district heating and/or cooling (DH/C) and smart grids, as well as energy generation from waste and wastewater based on the CoM Signatories' good practices. These good practices are associated with the urban climate governance options that have been put into action by the CoM signatories. Overall, the paper addresses multiple gaps in the literature by updating the approach for indirect emissions accounting and the linkage of good practices in the CoM Signatories' good practices database to the four modes of urban climate governance.

#### **2. Materials and Methods**

The methods are presented in the subsequent sections based on the definition of local energy generation, followed by the accounting of the indirect emissions in the CoM framework. The technical metrics that are important for an accurate and realistic accounting of indirect GHG emissions are further supported with the identification of good practices that are necessary to increase local energy generation. In this context, the method of the research work is summarized in Figure 1 with the aim of allowing authorities at the local level the opportunity to update and strengthen climate action. The broader policy framework in which local action is taken is also discussed based on EU energy and climate policies, the effective interaction of which is necessary for policy alignment.

**Figure 1.** Overview of the accounting context and the policy context for local energy generation. Source: own elaboration.

#### *2.1. Covenant of Mayors Definition of Local Power Generation*

The grid-supplied electricity is produced by a plurality of generation sources, ranging from distributed to centralized power plants, which can be located inside or outside the geographical boundaries of the cities' emission inventory.

Distributed generation is generally defined in the literature as electric power generation connected to the electrical distribution networks or on the customer side of the network, while usually, the connection of generation units to the transmission network is typical of centralised generation. The central idea of distributed generation, however, is to locate generation close to the load, hence on the distribution network or on the customer side of the meter. The distributed power facilities may differ according to "the purpose, the power scale, the power delivery, the technology, the environmental impact, the mode of operation, and the penetration of distributed generation" [38–41].

Based on the above definitions, a number of questions arise: what is local generation and how can it be differentiated from the national ones? Can it be assimilated into the definition of distributed generation? Generally, the term "distributed generation" refers to renewable energy technologies and it is often interchangeable with the term "local generation".

In relation to GHG emissions accounting, the present paper deals with the concept of local energy generation, which differs mainly from distributed generation. Within the CoM initiative that deals mainly with climate change actions at the local level, the environmental impact of power generation is of high importance, as well as the capability of the local governments to properly address it in their climate action plans. The regional context also affects the criteria, taking into account the national framework of energy and climate governance, characterized by different jurisdictions between state and non-state actors. Hence, there is a need to clearly define the criteria on which a power generation facility can be classified as local generation, and therefore to be accounted in the cities' emission inventories within the CoM framework.

The following main aspects were identified to be discussed in defining local production of electricity (LPE) within the scope of the CoM framework, and more precisely the location, the ownership, the source, the type of the technology and the capacity or power scale of the local generation facilities.

	- Local electricity production from renewable sources and combustible renewables are classified regardless of the technology and capacity, with the exclusion of the electricity sold to third parties that are located outside the local administrative boundaries and are identified through disclosed attributes. The rationale behind this is similar to the concept of the "residual mix" used by member states (MS) in the EU for assessing the grid emission factor. When determining the residual mix, MS often exclude the cancellation of electricity attributes (purchased via a Guarantee of Origin (GO) certificate in Europe) from the grid emission average [42]. The GOs are tracking instruments, introduced in 2009 by the Renewable Energy Directive (RED 2009/28/EC), that provide a means of demonstrating the origin of renewable electricity to consumers. The GOs system is a virtual one where the renewable attribute of energy trades separately from the physical energy. The usage is limited within 12 months of production of the corresponding energy unit, and is cancelled once it has been used.
	- Local electricity production from non-renewable sources, classified by types and capacity:
		- All combined heat and power plants, without capacity limit: the CHP system can be defined as local generation, as the second product (thermal energy) is consumed locally. Combined cycle gas turbines, internal combustion engines, combustion turbines, biomass gasification, geothermal, and Stirling engines, as well as fuel cells, are suitable for CHP processes. The heat demand usually drives the operation process, unless a back-up system for the heat production is in place.
		- Electricity-only with a capacity limit of 20 MW of thermal input: According to the principles that are laid out in the CoM, the inventory is not meant to be an exhaustive inventory of all emission sources in the territory but focuses on the energy consumption side and on the sectors and activities (buildings and transport) upon which the local authority has a potential influence. Large industrial power plants, covered by cap and trade schemes, such as the European Union Emission Trading Scheme (EU ETS), are not under LAs competence, but regulated by the ETS directive (2003/87/EC).

The amount of electricity to be reported in emission inventories as local electricity production will have a direct influence on the value of the local emission factor for electricity, and consequently, on the emissions that are associated with the consumption of electricity.

#### *2.2. Covenant of Mayors Methodology on Accounting the Indirect Emissions*

Two methods exist in the literature to allocate the emissions generated via electricity production to the final energy consumer of a given grid: location-based and market-based [43]. The location-based method is grounded on average generation emission factors for defined territories, considering local, subnational, or national boundaries, therefore reflecting the average emission intensity of the grid. In the marked-based method, the emission factors are derived from electricity purchases bundled with contractual instruments. The market-based method presents a higher degree of accuracy in the assessment of the emission factor. However, data collection in the context of market-based methods remains the main challenge when compared to location-based approaches, where local governments need to focus on assessing the indirect emissions of a multitude of consumers in their territories.

In order to calculate the indirect CO2 emissions that are to be attributed to the local consumption of electricity, JRC developed a specific methodology as described in this paper by estimating the local emission factor for electricity (EFE), taking into account both location- and market-based methods, but also an efficiency method for emission allocations in the case of CHPs.

#### 2.2.1. The Location-Based Method

The location-based method is used for assessing the amount of electricity and the associated GHG emissions excluding the electricity attributes (i.e., purchased via GOs certificates in Europe). This amount of electricity is associated with the physical electricity that is consumed in the local territory coming from the national energy mix and the local generation units and associated emission factors as below:


### 2.2.2. The Efficiency Method in Case of CHPs

Within the CoM reporting framework, the "efficiency" method for allocating CO2 emissions is the recommended method for LAs to be used in case of CHP power plants located within the geographical boundaries of the territory. The rationale behind this choice is to have a consistent method for emissions accounting with the method recommended in the European Energy Efficiency Directive (EED) (2012/27/EU) for determining the efficiency of the cogeneration process and primary energy savings. This method, called the "efficiency method", uses a reference system to allocate the output.

In a first step, the ratio of primary energy savings (PES) in comparison to a reference system is calculated. The methodology used to calculate the PES corresponds to the method defined in Annex II of the EED directive. According to Annex II of the EED directive, PES is defined as follows (Equation (1)):

$$PES = \left(1 - \frac{1}{\frac{\frac{\eta\_{CHP,heat}}{\eta\_{REF,heat}}}{\eta\_{REF,heat}} + \frac{\frac{\eta\_{CHP,cl}}{\eta\_{REF,cl}}}{\eta\_{REF,cl}}}\right) \tag{1}$$

where:


In a second step, the share of primary energy attributed to each of the two outputs electricity and heat can be calculated as follows. Regarding the share of allocation to heat, this amounts to a re-arrangement of Equation (1) based on the term (1 − PES) multiplied by the ratio of primary energy used for heat production in a reference scenario ( *PCHP*,*heat <sup>η</sup>REF*,*heat* ) on the primary energy used in the cogeneration scenario to produce the same amount of heat (PCHP,TOT), as ( *PCHP*,*heat <sup>η</sup>REF*,*heat* <sup>×</sup> *PCHP*,*TOT* ).

Therefore, the ratio of primary energy allocated to heat is obtained using Equation (2):

$$\text{Pratio allocated to heat} = (1 - PES) \times \left(\frac{P\_{CHP,heat}}{\eta\_{REF,heat} \times P\_{CHP,TOT}}\right) = \frac{\frac{P\_{CHP,heat}}{\eta\_{REF,heat}}}{\left(\frac{P\_{CHP,heat}}{\eta\_{REF,heat}} + \frac{P\_{CHP,cl}}{\eta\_{REF,cl}}\right)}\tag{2}$$

This method allocates the emissions based on the energy inputs required to produce separately (not in cogeneration) the same amount of outputs of heat and electricity (as in the CHP power plant output) as follows:

$$\text{CO2 emission allowed to heat} = \text{CO2}\_{\text{CHP/heat}} = \frac{\frac{\frac{P\_{\text{CHP/heat}}}{\eta\_{\text{REF/heat}}}}{\frac{P\_{\text{CHP/heat}}}{\eta\_{\text{REF/heat}}} + \frac{P\_{\text{CHP,cl}}}{\eta\_{\text{REF,el}}}} \times \text{CO2}\_{\text{CHP,TOT}} \tag{3}$$

$$\text{CO}\_2 \text{ emissions allowed to electricity} = \text{CO}\_{2\text{CHP};l} = \text{CO}\_{2\text{CHP}, \text{TOT}} - \text{CO}\_{2\text{CHP}; \text{heat}} \tag{4}$$

where:

