Contributing to carbon neutrality within water distribution services

IN

DEGREE PROJECT ENVIRONMENTAL ENGINEERING, SECOND CYCLE, 30 CREDITS

STOCKHOLM SWEDEN 2021 ,

Contributing to carbon neutrality within water distribution services

DAGHER NASSIM

KTH ROYAL INSTITUTE OF TECHNOLOGY

SCHOOL OF ARCHITECTURE AND THE BUILT ENVIRONMENT

Contributing to carbon neutrality within water

distribution services

Bidrag till klimatneutralitet för vattendistributionstjänster

DAGHER Nassim

Supervisor

FINNVEDEN Göran Examiner

BJÖRKLUND Anna

Supervisor at VEOLIA ROCHE Alain

Degree Project in Strategies for Sustainable Development KTH Royal Institute of Technology

School of Architecture and Built Environment

Department of Sustainable Development, Environmental Science and Engineering SE-100 44 Stockholm, Sweden

External

Logo

(if any)

ii

TRITA-ABE-MBT 2119

iii

Sammanfattning (summary in Swedish)

Vattenproduktion och distributionstjänster, har likt andra mänskliga aktiviteter, en påverkan på miljön. En indikator på denna påverkan är koldioxidavtrycket, vilket motsvarar den mängd växthusgas som släpps ut till följd av aktiviteten. För att motverka den globala uppvärmningen, måste växthusgasutsläppen minskas från dessa aktiviteter.

Vattentjänster avger växthusgaser i många steg, direkt eller indirekt: Detta innefattar bland annat produktionen av el som används för pumpning, produktionen och transport av kemiska reagens som används vid vattenbehandling, underhållsarbeten på nätverk och anläggningar, teknikernas dagliga pendling och resor för att driva tjänsten.

Denna mångfald av källor kräver en ännu större mångfald av möjliga åtgärder för att minska koldioxidavtrycket. Maximeringen av miljöfördelarna av sådana åtgärder där samtidigt genomförbarheten kan säkerställas och att samtliga typer av kostnader (ekonomiska, sociala, organisatoriska) begränsas kommer sannolikt att uppnås när en handlingsplan utformas.

Denna masteruppsats föreslår ett ramverk för en bedömning av möjliga åtgärder genom olika

kriterier och dess uppbyggnad för att skapa en effektiv handlingsplan. Bedömningskriterierna

eftersträvar en bättre förståelse av varje åtgärd och dess konsekvenser, så att åtgärden kan uppnå ett

effektivare resultat. Dessutom kan sammanhanget ha en stor inverkan på effektiviteten av ett antal

av dessa åtgärder, vilket understryker behovet av ett metodiskt ramverk som kan bidra till att dessa

aspekter beaktas.

iv

v

Abstract

Water production and distribution services, just like all human activity, have an impact on the environment. One indicator of that impact is its carbon footprint, which corresponds to the amount of greenhouse gas (GHG) emitted due to that activity. In order to counter global warming, mankind must reduce the GHG emissions of its activities.

Water services emit GHG at many steps, directly or indirectly: the production of the electricity used for pumping, the production and transport of chemical reagents used in water treatment, the maintenance works on the network and facilities, the daily commuting and travels of the technicians to operate the service, and more.

This diversity of sources calls for an even greater diversity of possible actions to reduce the carbon footprint. Maximizing the environmental benefits of such measures while ensuring their feasibility and limiting the costs of all sorts (financial, social, organizational) is most likely to be achieved when an action plan is designed.

This Master’s Thesis suggests a framework for the assessment of the possible actions through various criteria, and their arrangement into an effective action plan. The assessment criteria aim at a better understanding of each action and their consequences, so that the action is more likely to be effective.

Moreover, some elements of context can have a huge influence on the effectiveness of some actions, which highlights the need for a methodological framework that will help take these elements into account.

Keywords

water distribution service, carbon footprint, GHG emissions, action plan, framework

vi

vii

Acknowledgements

I would like to thank Alain ROCHE, my supervisor at VEOLIA, for the quality of his supervision, for the way he helped me organize my time and my interviews, and for his kindness and his care.

I also thank my supervisor at KTH, Göran FINNVEDEN, who made himself available for many meetings, and always gave me precious feedback and advice.

I would like to thank VEOLIA for entrusting me with this work, by hiring me for a six-months internship.

And last but not least, a huge thanks to everyone I worked with, or just spent time with, at the

company. Even though I cannot write your names in here, please know that discussing and working

with you was a pleasure. I very much enjoyed learning from all of you.

viii

ix

Table of contents

Sammanfattning (summary in Swedish) ... iii

Abstract ... v

Acknowledgements ... vii

Table of contents ... ix

1. Introduction ... 1

2. Aim & Scope ... 1

2.1. Aim ... 1

2.2. Scope ... 2

3. Background ... 2

3.1. Water production and distribution services ... 2

3.1.1. Water sources ... 2

3.1.2. The small water cycle ... 3

3.2. Carbon neutrality ... 5

3.3. Environmental impacts of water distribution services ... 6

3.3.1. Energy consumption ... 7

3.3.2. Construction and maintenance works ... 9

3.3.3. Mobility ... Erreur ! Signet non défini. 3.3.4. Chemical reagents and consumables ...10

3.3.5. Subcontractors and suppliers ... 11

3.3.5. Downstream usage ... Erreur ! Signet non défini. 4. Methodology ... 13

5. Results ... 16

5.1. The framework ... 17

5.1.1. System boundaries ... Erreur ! Signet non défini. 5.1.2. Data collection... 17

5.1.3. Indicator of global performance & goal setting ... 19

5.1.4. Assessment of solutions ... 20

5.1.5. Action plan development ... 23

5.1.6. Sensitivity analysis ... Erreur ! Signet non défini. 5.2. Applying the framework ... 25

5.2.1. Assessment of actions ... 26

5.2.2. Action plan ... 31

6. Limits & discussion ... 34

7. Conclusion ... 35

References... 37

x

1. Introduction

Global warming might be one of (if not the) most urgent and challenging crises of our millennium. In order to limit its extent and its impacts on our society, we must reduce the greenhouse gas emissions associated with human activity. Extracting primary resources, producing electricity, transforming raw material into goods, transporting goods and people from one place to another… everything we do has an impact on the environment. In order to reduce that impact, the first step is to account for its extent and to understand what role each and every component of our activity plays in this impact. Then solutions can be found, and their effects first estimated then measured and monitored.

Water is one of the resources that are most needed, for everything that ranges from our simple survival to our most advanced and luxurious comfort. From the glass on our table to the cooling pipes of a nuclear reactor, water is needed everywhere. Making it reach such a variety of places in a quality that is adequate to such a diversity of usages is no simple task, which requires both material infrastructure and energy. And of course, these come at a cost, which is both financial and environmental: water networks need investments to build and maintain, and come with an operating cost; and the system is responsible for GHG emissions, at different stages, from its construction to its operation and maintenance as well.

Methodologies and accounting frameworks already exist for establishing the list of carbon dioxide emissions (and other GHG) linked to the activity of a specific system: GHG protocol (The Greenhouse Gas Protocol, 2004), ISO standard 14064 and 14069 (ISO website, 2013 and 2018), Bilan Carbone (Association Bilan Carbone website, 2021)... However, the papers that suggest, apply, test and/or criticize a specific methodology for water services are few. In this context, elaborating a framework specifically for the water sector seemed necessary.

This thesis work was conducted alongside a six-months internship at Veolia, a French company operating in the fields of energy, waste management and water. The internship’s purpose was to identify ways of reducing CO

emissions for a water production and distribution service, and hierarchize them according to criteria that had yet to be determined.

2. Aim & Scope

2.1. Aim

The aim of this thesis is to define elements of framework to assess different actions for the reduction of GHG emissions, and arrange them into an action plan.

In order to do so, the framework takes as a starting point that the company already has an accounting framework for its GHG emissions, which implies that the hotspots (the main causes of GHG emissions) are already known. It also assumes that the possible actions that need to be assessed have already been identified.

The framework aims at granting its user a sound understanding of the available GHG reduction actions, of their advantages and drawbacks, as well as their context, what might alter their effectiveness, and what might be the obstacles to their implementation.

From this overall aim, four main research questions have been derived:

• What are the factors influencing the major hotspots in the system?

• How to assess and hierarchize the possible actions for the action plan?

• What aspects need to be considered when designing an action plan?

• What are the most effective actions for each hotspot?

2.2. Scope

This work covers GHG emissions for water production and distribution.

Even though the impacts of human activity on the environment can be of many types, this work will only cover GHG emissions. Godskesen et al. (2011) show that in the case of Copenhagen, greenhouse effect was the strongest environmental impact amongst all the indicators of the LCA methodology, after normalization. Their study covered water production for different treatment processes, along with transporting the water from the source to the distribution network.

This work takes the perspective of a private company willing to reduce its impact on global warming, for both environmental and commercial reasons.

This work does not deal with risk assessment nor risk management. The aspects of water service that involve notions of risk will be set aside, or addressed very briefly at most.

For confidentiality reasons, please note that I cannot publish the data that I have worked with. This is why this work will mainly be qualitative analysis. When an example is needed to illustrate a point of methodology, the data will be changed so that it does not represent any real-world situation. The location of the water service that I have worked on cannot be published either.

3. Background

In this section the most important elements of context needed to fully grasp the stakes of the framework that is the object of this thesis will be presented. The section is divided in four parts. The first one details the system of water production and distribution services. The second one introduces the concept of carbon neutrality, along with some GHG accounting principles and terminology. The third part focuses on the elements of the water production and distribution system that have already been identified as hotspots for GHG emissions in literature. Finally, the fourth part will briefly cover some of the existing framework for GHG accounting, assessment of actions for the reduction of GHG emissions, and development of action plans.

3.1. Water production and distribution services

The purpose of a water production and distribution service is to extract water from the environment, and deliver it to the consumers, while making sure that its quality meets the hygiene standards required. It comprises mainly a system for extracting water from the source, a treatment plant, pumping station, reservoirs, pipes, valves, along with a few more elements. Hazeltine (2003) provides in his guide a comprehensive overview of water production and distribution systems.

3.1.1. Water sources

Water can be extracted from various places, according to the specificities of the location.

Surface water is the water taken directly from water streams. As it is available on ground level and in

open air, no wells and only a little pumping are required. Glacier runoffs are also considered as surface

water. In mountain areas, surface water can be extracted upstream, at a higher altitude than the areas

that need to be supplied. In this situation, the distribution network is called a gravity-fed network, and the water is delivered to the consumers using only the force of gravity, with no additional pumping. On the other hand, surface water usually has a poor quality in relation to the hygiene standards, because it is in direct contact with its environment. This makes it necessary to treat the water thoroughly, using a bigger quantity of chemicals (Duguet and Gripois, 2007). Moreover, it can easily be contaminated by upstream industries or agricultural lands, which makes it dependent on upstream usages (which are often subject to another governance), and requires a good risk management plan as well as oversized treatment plants and/or backup procurement solutions (internal Veolia reports). In France, surface water represents 32% of tap water supply (Centre d’Information sur l’Eau website, n.d.)

Groundwater is the water located in porous underground geological layers, called aquifers.

Groundwater depth can range between a few meters down to a few hundred. Depending on that depth, groundwater extraction will require more or less deep wells and more or less powerful pumps.

Groundwater is relatively well isolated from potential causes of pollution, so it is globally of better quality than surface water, requiring less extensive treatment and chemicals (Duguet and Gripois, 2007). However, contamination is always possible, mostly from bad landfill waste management, leakage from pipes, or saline water infiltration. When an aquifer is polluted, it is more difficult to clean up, due to both the difficulty of access and the fact that it is more diffuse. In France, groundwater represents 68% of tap water supply (Centre d’Information sur l’Eau website, n.d.)

Finally, saline water can also be delivered to consumers after undergoing a desalination process, such as reverse osmosis or distillation. Such processes are highly energy-intensive, and hence both expensive and harmful for the environment (depending on the energy source). They are useful in coastal and arid areas, such as California, Florida, Mediterranean islands, and the Middle East (Horvath and Stokes, 2011). These processes are also useful in islands, which may not contain any other usable sources, and are difficult or impossible to connect to an existing inland network.

Public authorities tend to avoid relying on only one source, in order to secure the water supply, in terms of both quality and quantity (Grand Lyon Métropole, 2019). If something were to happen to one of the sources (failure in a water treatment plant, contamination of the source, unusual hydric stress…), the others could be operated more intensively to ensure the supply, while the issue is being addressed.

3.1.2. The small water cycle

The small water cycle (or domestic water cycle) is the name given to the full circuit travelled by water

so it can be used by human consumers. It comprises water production, water distribution, and

wastewater treatment. Quite ironically, it is not a true cycle per se, as the wastewater is never released

in the same water body that served as the source (it might be released in the same river, but always

further downstream). The term “small water cycle” is opposed to the “large water cycle”, which

corresponds to the “natural” water cycle (rivers, lakes, groundwater recharge, sea, evaporation,

rain…).

Figure 1. The small water cycle: production, distribution, treatment, and release (adapted from Syndicat des Eaux du Soiron website, n.d.)

Fig. 1 depicts the different steps that constitute the small water cycle:

1. water catchment: raw water (i.e. untreated water) is extracted from the source (a:

groundwater; b: surface water)

2. treatment plant: raw water undergoes one or several treatment processes so that it can meet health and hygiene standards (see section 3.3.4.)

3. pumping station: drinking-quality water is pumped in the main distribution pipe against a head that is appropriate to supply the corresponding area

4. storage (water tank): water is stored at an altitude higher than the area it needs to be distributed to. This allows the pumps to have a working schedule that is independent from the variation of demand, as long as the reservoirs are sufficiently filled

5. distribution (urban areas: residential, industrial) 6. distribution (rural areas)

7. wastewater directed for treatment

8. greywater reuse for municipal watering of green areas (after treatment)

9. wastewater treatment plant: wastewater is treated so it can safely be reinjected in a natural water body

10. treated water released in stream (downstream) 11. using sewer sludge as fertilizer

12. aquifer recharge (through wells or infiltration basins): this can only be done after a very thorough treatment process

This thesis focuses on elements 1 to 6 (Fig. 1, within the dotted polygon).

3.2. Carbon neutrality

In order to limit the magnitude and the consequences of climate change, ambitious targets must be met in terms of GHG emissions reduction. According to the IPCC’s 1,5 °C Global Warming Special Report (2018), emissions need to drop by 45% from 2010 levels by 2030 in order to stay below 1,5°C global warming, reaching net zero around 2050. The +2°C scenarios project a decline by about 25%

by 2030, reaching net neutrality around 2070. In France, a first Stratégie Nationale Bas-Carbone (National Low-Carbon Strategy in English) was adopted in 2015, which aimed at a 75% reduction of GHG emissions between 1990 and 2050. Revised in 2018-2019 and then 2020, the objective of the new NLCS is net carbon neutrality by 2050, with an intermediate objective of 40% reduction between 1990 and 2030 (in coherence with the Paris Agreements of 2015). These goals are also in line with the European 2030 Climate Target Plan, that calls for a cut of GHG emissions by at least 40% by 2030 (compared to 1990 levels), while a more ambitious target of 55% has been proposed by the EU Commission in September 2020. With this plan, the EU is aiming to reach “climate neutrality” by 2050.

Net carbon neutrality is a scenario in which total anthropogenic GHG emissions do not exceed the total capacity of CO2 sinks. As explained by Carbone 4 in their Net Zero Initiative framework (Carbone 4, 2020), such a scenario can only be reached collectively, i.e. at the scale of the whole planet. This is because the location of the CO2 fluxes is irrelevant: GHG emissions from all around the world contribute to global warming all the same, and the same reasoning goes for carbon sinks.

As depicted (Fig. 2), the process for achieving global carbon neutrality relies on three pillars that can

be worked on at an organizational scale: reducing own GHG emissions, reducing emissions outside

own perimeter, and developing carbon sinks.

Figure 2. The three pillars for carbon neutrality (adapted from Carbone 4, 2020)

On the vertical axis, in the bottom part of Fig. 2, three perimeters of action are shown. The figure depicts how the three pillars are applied to these three perimeters. There is a very important hierarchy in these perimeters, as companies and institutions need to start working on the environmental impacts of their own value chain, within their operational perimeter first and then upstream and downstream, before financing projects outside their value chain (Carbone 4, 2020). This order of priority will be discussed further in section 6.

Fig. 2 also mentions the three scopes of carbon accounting. As explained in the GHG Protocol Corporate Standard (The Greenhouse Gas Protocol, 2004), Scope 1 refers to direct GHG emissions that take place within the operational/organizational perimeter. Scope 2 refers to the GHG emitted for the production of the electricity and heat used for the operations. Finally, Scope 3 refers to all the indirect GHG emissions that are a consequence of the company’s activity, but occur outside the operational/organizational perimeter (e.g. emissions involved in the production of goods or consumables used in water production plants…).

3.3. Environmental impacts of water distribution services

A GHG hotspot is a portion of the system that is being studied (i.e. water production and distribution)

responsible for a significant share of the global GHG emissions of the system. Based on previous

scientific literature, methodologic guides, and internal Veolia reports, the usual hotspots for water production and distribution have been identified (Fig. 3), along with the major relationships to internal parameters and external factors.

Figure 3. Major hotspots in water production and distribution services and causal relationships to internal parameters and external factors. An arrow from A to B indicates that A has an influence on B. (own figure).

Fig. 3 covers the most important parts of water production and distribution services when it comes to GHG emissions. The different hotspots are detailed in the following subsections (3.3.1. to 3.3.5) .

3.3.1. Energy use

In literature, energy use is one of the most impactful hotspots in water production and distribution.

Some scientific papers even focus on energy use only (Boulos and Bros, 2010, Shrestha et al., 2011, Griffiths-Sattenspiel and Wilson, 2009), as being the main hotspot in the water production and distribution system. In the case study of Trondheim by Slagstad and Brattebø (Slagstad and Brattebø, 2013), energy represents 37% of GHG emissions of the water and wastewater system.

Electricity is the main energy vector used in water production and distribution. The production of this

electricity is one of the biggest contributors to the GHG emissions of the water production and

distribution sector. Hence, the importance of this hotspot depends heavily on the energy mix of the

country (see section 5.2.1.)

Electricity is used by the pumps needed to extract the water from underground sources, to elevate the water (so that it can reach areas located higher than the water source), and to guarantee a minimal standard pressure at the ground floor level of each building.

Electricity is also needed in the premises for heating, cooling, lighting, and ventilation, as well as for the computers and other appliances. This category is called auxiliary electricity consumption.

The biggest energy consumption usually takes place in the production phase. It is divided in two:

- the energy required to extract the water from the source (Fig. 3: pumping)

- the energy required to bring the water to a quality that meets the standards for drinking water (Fig. 3: water treatment)

The energy required to extract the water depends on:

- the depth of the source - the quantity of water needed - the efficiency of the pumps

The depth of the source directly determines the head requirements for the pump. For a specific pump, the deeper the water has to be extracted from, the more it will require energy. Similarly, the topography of the area that needs to be supplied also influences the amount of pumping that is required: pumping water up to the top of a hill requires more electricity than letting it flow downhill (gravity-fed networks).

Pumping more water will require either higher-capacity pumps, or more pumps. In both cases, more energy will be required. Whereas two pumps will require two times more energy as compared to one pump of the same model, going for a bigger pump allows for a better return to scale: it usually requires less energy to pump two times the volume with a bigger pump than with two smaller pumps.

Finally, the relationship between head and efficiency is not linear, and depends on the pump model and manufacturer. Every model has a theoretical optimal working spot on the efficiency curve provided by the manufacturer, which means it will provide a better energetical performance for that specific head, even though it can also pump against a lower or higher head. However, a different optimum is often observed empirically, this is why it is important to monitor the pumps in order to adapt their operating point.

These two points (scale efficiency, empirical optimal performance) highlight the importance of a good design. Huge amounts of energy can be saved by choosing the machines that are the most adapted to meet the purpose.

As for water treatment, different technologies exist. The appropriate technology and process are chosen depending on the origin of the water, which determines its hardness, as well as the presence of particles and/or pollutants. The various processes for drinking water treatment require either chemicals or additional pumping (to rush the water through filtration membranes), which will require more or less energy.

Additional pumping can also be required to recharge the aquifer artificially. This can be done either directly through a well connected to the aquifer, or through infiltration basins, in which water percolates through the bottom layers of sand to reach the aquifer. The process can be quite energy- intensive but is sometimes necessary to guarantee the aquifer does not reach a critical minimum level, and to ensure the quality of its water.

For the distribution of a given water volume, the required pumping depends solely on the morphology

of the area, and the relative position of the consumers to the water source. In order to distribute water

to an area that is located lower than the source, no pumping is required (gravity-fed network), whereas this is not the case for supplying water to the consumers living on the top of a hill (Fig. 4).

Figure 4. Water distribution levels (own figure).

3.3.2. Construction and maintenance works

Construction and maintenance works are also a hotspot for GHG emissions (internal Veolia reports).

These works include:

- replacements of water pipes - repair of leakages

- replacements of valves

- installations of new connections (connecting new buildings to the network) Sources of GHG emissions in works are:

- excavators and trucks exhaust emissions - extraction and transport of fill materials

- production and transport of the pipes/valves/connections that are being installed

GHG emissions of work are hence determined by the way the works are conducted (in Fig. 3:

operation). However, the amount of works also plays a role in the final amount of GHG emissions linked to works (in Fig. 3: planning).

The swiftness of the response to a detected leakage has a direct influence on the network’s global

performance, as repairing a leakage quickly limits the amount of water lost.

3.3.3. Chemical reagents and consumables

Water needs to be treated before being distributed, so that it meets the hygiene and quality standards.

Both the type of treatment and its intensity are determined according to the initial water quality, which directly depends on the characteristics of the water source (see section 3.1.1.). Sometimes, adding a bit of chlorine is enough to guarantee bacteria will not develop within the pipes, but water of poorer quality requires a heavier treatment, such as lime or soda decarbonation.

Duguet and Gripois (2007) have shown in their study that the weight of GHG emissions related to chemicals is largely dependent on the type of water source, as different types of sources provide water with different quality, which determines the types and amounts of chemicals to be used for treatment.

In their case study of Eau de Paris, the public water authority for the city of Paris, they found that raw materials inputs accounted for 46% of total GHG emissions (in front of energy with 41%). Back in 2007, water supply in Paris came from both groundwater and surface water, at roughly a fifty-fifty ratio. Duguet and Gripois analyzed the differences in GHG emissions between groundwater and surface water: groundwater required 2 to 3 times less energy for production and distribution, and 2 to 4 times fewer chemical reagents, compared to surface water. This difference shows the importance of context, which can make the hotspots differ completely from one system to the other, with chemical reagents sometimes being the most significant hotspot for GHG emissions, and sometimes represent barely a few percent (internal Veolia reports).

Figure 5. Different steps of water production for surface water (Centre d’Information sur l’Eau website, n.d.)

The chemical reagents are used in some of the steps in the treatment of raw water, which are (Fig. 5):

1. catchment

2. screening: bigger debris get removed by a grid

3. sifting: tighter grids for smaller debris

4. flocculation-coagulation (or decantation): using a coagulant, suspended particles (such as sand, silt, plankton, fine clay, bacteria, salt…) precipitate and sink at the bottom of the basin 5. sand filtration: water percolates at slow speed through a layer of sand, removing the last particles that can be seen by naked eye. At this stage, viruses and bacteria can still be present in the water

6. ozonation: ozone is used eliminate organic and non-organic matter, micropollutants (such as pesticides), and gets rid of tastes and smells

7. activated carbon filtration: activated carbon is the most adsorbent compound known to man, and is used to remove the remaining micropollutants, organic matter, and ammoniac, which gives the water better smell, taste, and color

8. chlorination: chlorine is used in small doses to prevent the development of bacteria within the water pipes during distribution

9. quality and sanitary control: ensuring the water meets quality and hygiene standards

3.3.4. Downstream usage

CDP’s Global 500 Climate Change Report (2013) shows that there is a mismatch between the scope 3 categories reported by the companies and their relative weight compared to their total GHG emissions. Downstream usage of the products sold is one of the aspects that is the most rarely considered in GHG accounting, despite being the most important source of GHG emissions by far, when other aspects, like business travels, employee commuting, or waste management, are far less impactful but far more documented. There are mostly three main reasons behind the fact that companies tend to disregard the emissions linked to downstream usages:

- it is often difficult to draw the line where the usage stops and what it implies, what should be of concern for the company and what should not

- collecting relevant data is often very difficult, as it implies studies on stakeholders that are usually outside the company’s perimeter. These studies consist most of the time in polls, which may yield results not robust enough. They can also take the form of experiments or models, which can sometimes be criticized on their external validity

- considering downstream usage, which is often a huge hotspot for GHG emissions, makes it more difficult for the company to appear environmentally friendly. From a marketing perspective, it becomes much more difficult to assume, and from a financial perspective, it increases the investments needed for carbon compensation required to claim carbon neutrality.

As for drinking water production and distribution services, GHG emissions related to downstream usages comprise the production and usage of energy for domestic hot water, the use of chemicals for cleaning (both domestic and public/industrial uses), and the emissions linked to the life cycles of all domestic appliances that work on tap water.

Energy consumption for heating domestic water has been estimated by an internal report by Veolia

to be up to four times higher than the whole energy needed in water production and distribution and

wastewater treatment, which makes even a few percent of energy savings in water heating

environmentally valuable.

3.3.5. Other sources of GHG emissions

In addition to the previously mentioned hotspot, a few elements are responsible for the rest of the GHG emissions. Their share is less significant than the one of each hotspot.

The operation and maintenance of a water network relies on many teams of technicians, that need to travel along the network (to check for leakages, operate the pumps, monitor the network, take care of customer service…). These technicians, as well as all of the company’s employees, also need to commute from home to work (and vice versa). All of this is mainly done by car, as it offers better flexibility than public transport, as well as more comfort. And of course, cars emit a fair amount of GHG. However, internal Veolia reports show that car travels are responsible for less GHG emissions than the previously mentioned hotspots (energy, works, chemicals, and downstream usage).The construction of the company’s premises, as well as the pumping stations and the treatment plants, is also a source of GHG emissions. It is usually accounted for as amortized over a long period of time (around 50 years for example).

Finally, many sources for GHG emissions are located outside of the perimeter of the water company, as the water services rely on many subcontractors and suppliers. Here is a list of usual subcontractors and suppliers:

- energy suppliers

- pipes and valves suppliers

- consulting companies (audits, management, strategy…) - chemicals and reagents suppliers

- suppliers for pumps, compressors, and other machines - water meters manufacturers

- cloud storage services Energy supply

As energy is one of the major hotspots in water production and distribution, the way it is produced weighs heavily on the final carbon footprint. Table 1 shows the emission factors (EF) for different means of production in France.

Table 1. Emission factors for different means of electricity production in France (ADEME website, n.d. a). These EF include construction (upstream) and operation, but not dismantling and recycling (downstream).

Means of production EF (g CO

/kWh)

Nuclear power plant 6

Gas power plant 418

Coal power plant 1058

Oil-fired steam power plant 730

Wind power 14 (inland), 15 (offshore)

Photovoltaic 55

Other suppliers of goods

It is no easy task to trace the carbon footprint of all the suppliers of a water company, as some of them may not have studied their carbon footprint in the first place, or if so, may communicate it in a non- transparent way. It takes time and expertise to assess these studies, as the methodology used by the suppliers is rarely known. For instance, they might exclude from their system boundaries one or more hotspots, whether it is on purpose or not.

The main factors influencing their carbon footprint are the type and origin of the raw materials, and the country in which the product is fabricated or assembled (national electricity mix, transport distances). Even without having a comparative LCA of two pipe manufacturers, it can be assumed that pipes made in eastern Europe are more likely to have a worse carbon footprint than pipes made in France (for a usage in France of course). Another example with activated carbon: it emits up to 4,4 times less GHG when made from coconut fibers, instead of coal (ADEME, 2018), showing how the choice of raw materials can make a difference in the final carbon footprint.

Cloud storage services

The amount of data stocked (telemetry, customer feedback, network warnings…) is not significant enough for the digital aspects to be a major hotspot. In fact, the electricity required to power up some internal servers at Veolia gives an order of magnitude that can be extrapolated for a whole water service, which is extremely small compared to the other energy usages (internal Veolia report).

3.4. Previous research

Some methodological frameworks already exist for the assessment of GHG reduction measures. The French Association Bilan Carbone has published a guide, “Guide méthodologique – Objectifs et principes de comptabilisation” (methodological guide – accounting objectives and principles) (Association Bilan Carbone, 2017), which defines a framework for GHG accounting, based on the physical fluxes going through the system. It defines key principles that are needed as a basis for the assessment of reduction actions.

The ADEME (the French agency for the environment and energy) published a guide in 2014 for creating, implementing and monitoring GHG reduction action plans (ADEME, 2014). It aims at granting a better understanding of the stakes of the action plan, and the suggested framework takes the form of many sets of questions (applying the what?-why?-how? tryptic over different aspects of the action plan), presented as tables. This document is general, and not sector-specific.

The ADEME also published a group of GHG emissions sectoral guides, along with actors from each sector. In particular, they published a guide for the sector of water production and wastewater treatment (ADEME, 2018). This guide lists all the key aspects to be considered when assessing the GHG emissions of a water service, and proposes a list of reduction actions, along with a short analysis and some examples.

4. Methodology

As previously explained in section 2.1., the aim of this thesis work is to design a framework for the

assessment of GHG reduction actions and the design of an action plan. The methodology relies on

two parts: literature study and interviews.

4.1. Articulation of the methodology

This methodology draws information from two sources: literature and interviews conducted during the internship (Fig. 6). Literature studies served as a source for general knowledge about GHG accounting and LCA, and about water production and distribution services. This provided the basis for a sound understanding of the background. This knowledge allowed for moving from general to specific, by combining these two aspects to lay the foundations of this thesis: GHG accounting and life cycle assessment (LCA) of water production and distribution services.

Scientific literature also provided examples of GHG reduction actions in the water sector, as well as data needed to assess the actions.

The interviews’ purpose was to benefit from the experience of the people working at Veolia as a complementary source to literature. They also helped narrow down the focus of literature review, and confront the results of these studies to the reality of the ground. Finally, interviewees could provide site-specific data, which helped assess the actions, as well as new contacts to interview (snowball sampling).

Figure 6. Articulation of the methodology used in this thesis (own figure).

4.2. Literature study

The literature review was conducted on Google and Google Scholar, by searching for combinations of terms such as “LCA”, “carbon accounting”, “GHG”, “water services”, “water production”, “water distribution”, “action plan”, “framework”. More documents were also accessed through the references of other papers.

Three categories of documents were searched for:

- methodological guides: for GHG accounting frameworks, action plans…

- scientific literature: application of LCA and carbon accounting to water production and distribution services, assessment of GHG reduction actions

- framework documents: national and international legislation and regulations on water, French regional plans covering the subject of water management

Many internal reports from Veolia were also consulted, among which:

- GHG accounting reports

- technical studies (energy use, water treatment, pressure management…) - life cycle assessments of some devices (such as water meters)

4.3. Interviews

Along with the literature review, many interviews were conducted within Veolia during the internship.

These interviews took the form of informal discussions, and can be divided in three categories:

- semi-structured interviews ending in open discussion: a few questions were prepared in advance with regards to the interviewee’s job and position, but the interviews remained very open. This choice aims at drawing the best out of the interviewees’ experience.

- unprepared interviews: some of the interviewees were recommended by other people, but it was not possible to know how they would be able to help before the interview. These interviews would typically start by one very general and open question (such as: “What can you tell me about the GHG emissions in your line of work?”), and then continue as an open and informal discussion.

- work meetings: some people were interviewed more than once, and helped develop some of the results that are presented in this thesis. They helped improve the results by iteration from one meeting to another.

The goals of the interviews were to:

- confront ideas (from both literature and previous interviews) to the situation on the ground - validate and improve the methodology by iteration

- get data

- get new people to get in touch with (snowball interviews)

All of the interviewees were people working in many different departments within Veolia. For confidentiality reasons, the names of the interviewees cannot be published here. They will be referred to using letters, according to the following Table 2, which lists all the interviews conducted along with the function of each interviewee.

Table 2. List of the interviews conducted within Veolia.

A Head of Quality, Security and Environment department B In charge of energy-related matters in water production C Head of the Production service

D Technical engineer, in charge of environmental and mobilities studies

E Technical engineer, in charge of environmental studies and carbon accounting F In charge of climate strategy

G Technical engineer, in charge of a study of the pressure within the network H Head of department of network maintenance

I Head of department for the repair of leakages J Data scientist

K In charge of a decarbonation service

4.4. Justification of methodology

Literature review is a staple of scientific research. The choice of searching for the three different categories of documents mentioned previously (section 4.1.) is motivated by the complexity of the problem, and the fact that it spans on different levels, due to its linking nature between the theoretical scale and the operational scale.

However, literature study would not have been enough on its own. The choice of interviewing people at Veolia was motivated by the availability of that interactive source of information and data. Open discussions allowed for a better understanding of the scope and helped narrow it down to the most important elements.

Moreover, previous methodological guidance (ADEME, 2018) was elaborated by people working in water companies and in the regional water authorities, under the supervision of the ADEME and the ASTEE (French technical and scientific association for water and environment), which provided the organizational and methodological skills needed to build up the framework. In this thesis, the organizational and methodological elements were extracted from literature, whereas the interviews served more as a technical basis. Thus, the methodology in this thesis was chosen to match the configuration of the teams that worked on previous guidance.

5. Results

In this section the results will be presented divided in two parts. The first part (section 5.1.) details the framework for the assessment of GHG reduction actions, and the design of an action plan.

Whereas the second part (section 5.2.) contains an application of this framework. In this application,

some actions are assessed according to the framework, and then an example of an action plan is

proposed.

5.1. The framework

The framework is divided in four steps (Fig. 7):

- Data collection - Goal setting

- Assessment of actions - Action plan development

The definition of the system boundaries and the identification of actions are not included in this list, as the framework takes as a starting point that the company is already reporting its GHG emissions, and has already identified a set of actions.

Figure 7. The four components of the framework (own figure)

5.1.1. Data collection

Activity data and emission factors

Two different sorts of data are needed to calculate GHG emissions: activity data (AD) and emission factors (EF).

Activity data: represents, for every specific activity that emits GHG, the extent of that activity.

Examples: number of kilometers travelled by car, amount of electricity used, weight of sand used in the basins…

Emission factors: for every activity, gives the GHG emissions related to 1 unit of activity. Example:

standard passenger cars emit ~0,2 kg CO

/km, electricity production in France emits 0,056 kg CO

/kWh… (ADEME website, n.d. a).

The product of these two numbers gives the total GHG emissions for one specific activity / product / process.

Activity data is to be gathered in the company. How difficult it is to gather depends on the company’s processes, how much they keep track of their activity, the way the data is stored, as well as the permissions needed to access the data.

Figure 8. Important elements of the quantitative data chain (own figure).

Fig. 8 highlights the elements that can make data collection easier or more difficult. First, it is important to know who is in charge of collecting the data. Is the same person in charge of this for different activities? Is data gathering an important part of that person’s job? How is the data collected? The method will impact the uncertainty, or even the ability to use the data at all. This is where the purpose comes into play: the way the data is collected can be perfectly adapted to a specific purpose (e.g. data needed for the operation), but not for estimating GHG emissions. The sample is also important for that reason: is it possible to extrapolate without leading to too much uncertainty?

The sample has to be either big enough, or sufficiently representative of the whole. The sample might be fitting for a specific purpose that the data was originally used for, but not for the estimation of GHG emissions. The data also has to be gathered at a sufficient frequency, so that it matches at least the frequency of the reports (which are usually yearly).

Then comes the question of data storage and access. Which software is used to treat and save the data? Is the data in a form that can be directly used for the exercise? Are some steps automated? How difficult is it to combine data from different sources and/or different activities? Can the data be accessed directly, or does the person in charge of assessing the GHG emissions need to ask someone else? Usually, the easiest case is when all the data is stored in well-organized Excel sheets on one or more shared folders, but I encountered situations where I had to write down some numbers given to me through the phone (which is clearly not the best way of collecting data).

As for the emission factors, they can be obtained through certified databases: in France, the most widely used is probably ADEME’s Base Carbone (which is free to access) (ADEME website, n.d. b).

Other databases are paid, like the widely used Ecoinvent (for LCA) (Ecoinvent website, n.d.). These

databases give average values for different systems (products, activities, processes…) on different

scales (national average, European average, world average…). However, it is always better to rely on

case-specific data that is known to correspond accurately to the actual situation of the company. This work can be done by the company, by assessing in a detailed way some of its processes that already have an existing national average in public databases, using site-specific knowledge. When the product is bought from a supplier, or when the process is the object of subcontracting, then the company can reach out to the supplier/subcontractor and ask them for their specific analysis.

However, doing so raises the question of consistency, as different companies are likely to use different methodologies for their assessments, and they might be making hypotheses that will end up minimizing the environmental impacts of their products/services.

Quantitative & qualitative data

Activity data and emission factors are quantitative data. However, the assessment cannot be completed without some additional information, which will take the form of qualitative data. Most of the criteria described in section 4.4. (Assessment of solutions) can only be assessed using qualitative data. This data can be gathered as feedback from people working on the field, or by discussing with the heads of the relevant departments.

Snowball sampling consists in getting feedback from new people, based on what information the previous interviewees were able to give. Snowball sampling is particularly effective here, as moving from one interviewee to another allows for a better global understanding, by going through different viewpoints and seeing the same system from different angles, with people that might have different knowledge and skills.

Knowledge sharing

Companies in the water sector are often huge bodies operating on contracts in different cities, and even in different countries. This can, in theory, create a lot of opportunities to get feedback from different contexts, like for example the testing of new technologies or new methods. However, the ability to get such feedback depends highly on the focus that is placed by the company on sharing knowledge. The company’s organization makes it more or less easy to reach out to collaborators from other regions and exchange experiences. In order to increase shared knowledge, some companies organize thematic meetings and workshops that will create information flow between teams that do not work together on a daily basis.

Knowledge sharing is important for critical analysis of processes and technologies, especially for sensitivity analysis, as it allows for a better understanding of the influence of the elements of context on the efficiency of an action.

5.1.2. Indicator of global performance & goal setting

In order to monitor GHG emissions, an indicator of global performance is needed. It has to be the ratio between the entity’s GHG emissions over a period of time (generally one year), and an activity data. In the context of drinking water production and distribution, relevant indicators can be:

- yearly GHG emissions divided by the yearly volume of drinking water sold - yearly GHG emissions divided by the yearly number of consumers

- yearly GHG emissions divided by the product of yearly number of consumers and yearly

volume sold

Replacing the volume of water sold by the volume produced would take water distribution (and its related losses) out of the scope, turning it into a good indicator for water production only. Using the volume of water (either produced or sold) makes the indicator insensitive to the variation of water consumption, so it focuses on improving the process of water production and distribution itself.

Keeping it as a second indicator allows for taking into account actions such as campaigns to raise awareness so that people use less water for instance.

The entity wishing to contribute to carbon neutrality has to reduce its GHG emissions first. But to what extent? When does it become “reasonable” to say that the entity has in fact done its best to reduce its GHG emissions? This depends on:

- the company’s motivations

- the physical limits of the company’s activity

According to ADEME (2014), the company’s motivations can be:

- anticipating the evolution of energy prices - reducing its operating costs

- innovation / standing out - adapting its commercial offer - improving its brand image - acting for the climate

- anticipating new externalities - building new partnerships

In line with its ambitions, the entity can define a goal to reach, which is a combination of a value of the indicator and a deadline (in comparison to a reference year), with eventually intermediate goals.

It can also be a yearly decrease (in %) of the indicator over a period of time (e.g. “-2,5% in total CO

emitted per m³ sold every year until 2030”).

5.1.3. Assessment of actions

When designing an action plan, all possible actions and solutions must be listed. In order to choose which actions to implement, and how to implement them, they must be assessed. As criteria are needed for any assessment, here is a set of relevant criteria in this context.

As for any framework, having too many steps and too many criteria in the process can jeopardize its usefulness. Over-categorizing and over-processing can turn out to be counterproductive. It is up to the person in charge of the action plan to limit the framework to a range of criteria that are deemed most relevant, and sufficient for the purpose. However, all the criteria must always be kept in mind, in order not to miss on something important.

Below is a list of criteria that was found to be useful during the internship. Each aspect is detailed after the list:

- chain of consequences (direct, indirect)

- related actors and stakeholders - characteristics

- setup difficulty

- relations with other actions Chain of consequences

Understanding the chain of consequences of an action is the first step of the assessment. A thorough analysis will make it possible to detect potential side-effects, rebound effects, and other negative consequences early on in the process. This will also help identifying synergies and disturbances between actions.

Even though this present thesis work only focuses on GHG emissions, and does not calculate the severity of other types of environmental impacts, that does not mean that these other impacts should be disregarded altogether. Sometimes, it is known that a specific action will have a sizable negative impact, and it should be taken into account in the trade-off.

An action can also have social, ethical, or political consequences. These should be identified as well.

Related actors and stakeholders

While it is not always necessary to undergo a full stakeholder analysis, it is important to know who is involved in every suggested action. After identifying the actors, the next step is to identify their interests and the means in their possession that can be mobilized toward their interests. According to this, actors can either act as allies, or as hindrances that will need to be dealt with, making the action harder to implement. Here are some examples: if two competing companies are involved in the small water cycle, and one action requires to be planned at a scale that extends on the organizational perimeter of these two companies, then it might be difficult to make them work together. Another example would be the case in which the action would be politically unpopular: for example, if it relies on public authorities to change the legislation, then it might just never happen. Thus, the stakeholders can be part of the levers or hindrances that will make an action more or less effective.

Identifying the actors is also the first step of co-construction. The level of involvement of each actor in the making of the action plan needs to be determined to best serve the purpose of the action plan.

Involving too many actors will slow down the process, and will make it more difficult to reach consensus. On the other hand, associating actors to the process early on can also help avoid some issues that would cause some precious time to be lost.

Furthermore, identifying the stakeholders for every action is necessary for the later step of implementing the action plan and its follow-up: any action must have a pilot, in charge of the supervision, and a team to implement it and monitor its effectiveness.

Characteristics

Listing the characteristics of an action allows for a better understanding of it. It will also help go through the assessment of the actions and be more efficient when designing the actual action plan.

Here is an overview of relevant characteristics and their explanations. Please note that the distinction between two opposite terms is not always clear: for instance, an action can be both new and relying on some aspects that were already present; an action can be reversible but with a high reversal cost;

or be a technical solution that requires organizational changes, and so on…

Moreover, some of the following categories might be irrelevant for some specific actions, and these

actions should hence not be forced into categories that do not fit them well.

Table 3. Description of the different criteria for assessing actions.

Effectiveness How much GHG reduction?

New action / continuation Is the action bringing something totally new in the service, or is it just pushing further something that was already in place? In other terms, is there a precedent that can be used as a start, or does the action need to be started from scratch?

Technical / organizational Technical: relies on a new technology

Organizational: relies on the optimization of flows (physical flows, energy, data, knowledge, human labor…)

Doing better / different Doing better: improving an existing process

Doing different: changing the process so it gives better results.

Reversible / irreversible Once the action is implemented, is it possible to return to the previous state? And at what cost?

One-time / continuous Is the action done once, or is it something that needs to be continuously done? Is it periodic?

All-or-none / gradual Does the action need to be implemented at full scale / full intensity at once? Is it possible to have small-scale tests, or increase the intensity gradually?

Setup difficulty

Among the different difficulties that can be encountered when trying to implement an action are:

- financial costs - technical difficulties - aversion to change - cost of consequences - deployment time

All these aspects will make an action more difficult to implement. Some difficulties can be mitigated, but others are inherent to the action. How efficient an action is will depend on its effectiveness, the setup difficulties, and the company’s ability to deal with these difficulties.

Relations with other actions

It is important to detect potential correlations to other actions, were they positive or negative. Some

actions might also have cumulative effects, meaning that the final outcome will be more than the sum

of the outcomes of each action taken separately. Correlations can be a cause of variations in the

predictions if not addressed properly.

Since a company’s budget is not infinite, there is a limited number of actions that can be financed simultaneously. The financing of some actions might hinder others. On the contrary, some business models might be design in order to make some actions finance others.

5.1.4. Action plan development

Once the emission hotspots have been identified and analyzed, and once the possible solutions have been assessed as well, these latter have to be set up into an action plan. The purpose of an action plan is to organize the actions spatially, temporally, logistically, and financially, in order to draw maximal benefits from lowest cost and/or negative externalities. The terms “costs” and “benefits” here include all kinds of costs and benefits, not only the financial ones: environmental, social, organizational…

Temporal, spatial, financial, and logistical factors are often intertwined, and it is hardly possible to consider them separately.

In Table 4 are (some of) the questions that the planner needs to answer before setting up the action plan, for every action.

Table 4. Elements to be considered when designing an action plan.

Temporal planning When does the action need to start?

How much time does it need to take effect?

Does it call for preliminary studies?

Are other actions a prerequisite?

With limited available resources at a given time, which actions are to be prioritized?

How fast will the solution be deployed?

Will there be a trial period?

At what frequency are the indicators to be monitored?

Spatial planning Where will the action be implemented?

Will there be a small-scale test?

What are the characteristics specific to the different locations considered?

Financial planning What is the assigned budget?

How much will it cost?

Preliminary studies?

Deployment cost?

Additional operation costs?

Training costs?

Maintenance costs?

Dismantling costs?

Does it need hiring? Subcontracting?

How much profits are expected (if any)?

Where will the funding come from?

Logistical planning Who will be in charge of implementing the action?

Are construction works needed?

What are the negative externalities?

How can they be minimized? (cf. spatial and temporal planning)

Deployment / operation / maintenance: is hiring and/or subcontracting

needed?

Figure 9. Usual evolution of GHG emissions for a specific hotspot when an effective action plan is set up (own figure).

The influence of a successful action plan for the reduction of GHG emissions (for a single hotspot) can be divided in 4 periods (Fig. 9), as observed during the internship through Veolia internal data:

1: before taking on actions to reduce GHG emissions, they grow due to economic growth, as more human activity with no specific mitigation measures equals more pollution and GHG emissions. 2:

the first actions make the GHG emissions drop drastically, as the plan will try to achieve the most effective enhancements as quickly as possible. 3: the following solutions are less effective than the first ones, as it becomes more and more difficult to further reduce the emissions. 4: the new solutions barely compensate for the increasing emissions due to economic growth. However, please note that the horizontal axis is not to scale, and that the different phases could take relatively more or less time depending on the context, and the investments and other difficulties linked to the implementation of the relevant actions.

Unexpected events may disturb the overall shape of the curve: accidents or breakdowns can cause additional GHG emissions linked to maintenance works or the use of backup solutions that might be more energy intensive.

Moreover, it is not impossible that the major drop from 1 to 2 happens at a different time, with many possible explanations: technology not yet available at the start of the action plan (especially if it started early), high financial costs that are only becoming worth bearing now as the concerns regarding climate crisis are growing…

It is also possible that for some areas of activity, no radical solution exists, only small improvements can be made, and hence no big drop (1/2) can be observed. On the other hand, there might be situations where there is no radical solution, but rather two main solutions, so that two drops may be observed. These drops could be merged in one, if the corresponding actions were implemented at the same time, or be well differentiated.

This leads us to one of the biggest issues in the assessment of actions: when the benefits of different actions take place at the same time, it is very difficult to estimate the contribution of each action to the observed decrease. This is especially complex when the different actions have a cumulative effect, meaning that they interact with one another.

However, knowing the exact contribution of every measure is not always of the greatest relevance. For

instance, if the action is one-shot, non-reversible, and is known to have positive effects, then once the