OCH HUVUDOMRÅDET
MECHANICAL ENGINEERING, AVANCERAD NIVÅ, 30 HP STOCKHOLM SVERIGE 2020 ,
Participatory Modelling for Carbon Footprint Analysis
- A Case Study at DeLaval
EMIL DECKNER CARL MAILER
KTH
SCHOOL OF INDUSTRIAL ENGINEERING AND MANAGEMENT
Participatory Modelling for Carbon Footprint Analysis
- A Case Study at DeLaval
Emil Deckner Carl Mailer
2020-06-17
Master of Science Thesis TRITA-ITM-EX-2020:333
KTH School of Industrial Engineering and Management Mechanical Engineering
SE-100 44 STOCKHOLM
Participativ modellering för analys av koldioxidutsläpp
- En fallstudie på DeLaval
Emil Deckner Carl Mailer
2020-06-17
Examensarbete TRITA-ITM-EX-2020:333
KTH Skolan för industriell teknik och management Maskinteknik
SE-100 44 STOCKHOLM
Abstract
Because of global warming, companies have started to tackle sustainability issues within their operations, but major uncertainties exist on how to establish a quantitative baseline of the current environmental performance of companies. Numerous investigations have been made to assess the carbon footprint of companies with a variety of methods, tools and strategies.
However, the lack of transparency in the methods used and the assumptions made could prevent companies to replicate methods and to analyse the results. Because of this, we will investigate how participatory modelling could be used to create a model of the carbon emissions of a company, but also how this method enables the company to understand the methods used and the results from the investigation. By doing this, we also aim to clarify how the process could be set up, which stakeholders that need to be involved and what data sources that could be used.
The thesis was conducted as a single case study at a manufacturing company named DeLaval.
A participatory modelling process with three major phases was carried out according to proposed methodologies in previous research. In the first phase, a conceptual model of the system accounting for the emissions was created. In the second phase, a quantitative model of the system was developed by gathering data and validating the calculation methodologies with operational stakeholders within the company. In the third phase, the results were verified, and the company could set up sustainability targets based on the findings.
The outcomes of the case study showed that there are major benefits with applying participatory modelling because different perspectives throughout the organisation could be gathered efficiently to create a representable model of the company. The modelling strategy had more benefits in organisational areas characterised by high complexity with numerous of different stakeholders with different roles or by geographical distribution. To create the model, primary data consisting of master product data and transactional data was used, together with secondary data, consisting of carbon emission coefficients and gap filling data created by the modellers.
By basing the calculations on the methodology set up by the GHG protocol and anchoring the root definition of the system with strategic stakeholders, the results were trusted by the organisation.
Keywords: carbon footprint, corporate sustainability, participatory modelling, mediated modelling, system thinking, organisational learning, sustainability strategy
Participatory Modelling for Carbon Footprint Analysis
- A Case Study at DeLaval
Emil Deckner Carl Mailer
Approved
2020-06-17
Examiner
Björn Palm
Supervisor
Per Lundqvist
CommissionerDeLaval International AB
Contact person
Rania Karat
Sammanfattning
På grund av den globala uppvärmningen har företag börjat att hantera hållbarhetsutmaningar inom sin verksamhet, men stora frågetecken kvarstår gällande hur en kvantifierad bild av företagets nuvarande utsläpp ska beräknas. Flertalet studier har genomförts för att undersöka koldioxidavtrycket på företag, med flertalet olika metoder och verktyg. Metoderna och antaganden som gjorts har dock bristande transparens, vilket hindrar andra företag från att replikera beräkningarna och att göra analyser av resultatet. Baserat på detta kommer denna studie att undersöka hur participativ modellering kan användas för att skapa en modell av koldioxidutsläppen från ett företag, men också hur denna metod underlättar för företaget att förstå metoderna som använts och resultaten från undersökningen. Genom detta ämnar vi att bringa klarhet gällande hur processen kan se ut, vilka intressenter som ska vara delaktiga och vilka datakällor som kan vara användbara.
Studien genomfördes som en enkel fallstudie på det producerande företaget DeLaval. En participativ modelleringsprocess med tre faser genomfördes i enlighet med etablerade modelleringsprinciper från tidigare studier. I den första fasen utvecklades en konceptuell modell av systemet för estimering av koldioxidutsläppen. I den andra fasen utvecklades en kvantitativ modell as systemet genom att samla in data och validera beräkningsmetoderna tillsammans med operative intressenter på företaget. I den tredje fasen verifierades resultaten och företaget hade möjlighet att sätta upp hållbarhetsmål baserat på resultatet.
Utfallet av fallstudien visar att det finns stora fördelar med att använda participativ modellering eftersom olika perspektiv i organisationen kunde inhämtas på ett effektivt sätt för att skapa en representativ modell av företaget. Modelleringsstrategin hade större fördelar i delar av företaget som karakteriserades av hög komplexitet, med många olika intressenter med olika roller eller av geografisk utspriddhet. För att skapa modellen krävdes primärdata innehållande produktinformation och transaktionsdata samt sekundärdata, innehållande utsläppsfaktorer och överbryggande data skapad av modellerarna. Genom att basera beräkningarna på metodiken skapad av GHG protocol och förankra syftet med systemet tillsammans med strategiska intressenter, skapades en tillit till resultaten inom organisationen.
Nyckelord: koldioxidutsläpp, hållbart företagande, participativ modellering, medlande modellering, systemtänkande, lärande i organisation, hållbarhetsstrategi
Participativ modellering för analys av koldioxidutsläpp
- En fallstudie på DeLaval
Emil Deckner Carl Mailer
Godkänt
2020-06-17
Examinator
Björn Palm
Handledare
Per Lundqvist
UppdragsgivareDeLaval International AB
Kontaktperson
Rania Karat
Foreword and Acknowledgement
Firstly, we would like to thank our thesis supervisor at the Royal Institute of Technology, Per Lundqvist. He has guided us through a major part of our master program and has headed several course modules within the sustainability area. This inspired us to conduct this thesis and to take the specific angle that we did, trying to apply a modelling technique to a real-world problem.
Throughout the semester Per has provided us with inspiration in how to tackle the problems we encountered and finding ways to solve them.
The thesis proposal was provided by the Global Sourcing Department at DeLaval who we are incredibly thankful to for giving us the opportunity to write our Master Thesis at their company.
More importantly, we would like to thank our supervisors Rania Karat and Anna Lindquist for always being supportive, for challenging us and for providing their immense knowledge about every detail that emerged. We are inspired by their determination of finding ways to make DeLaval sustainable and we would not have reached the same level of details in this work without them.
We would also like to thank all of the informants and employees at DeLaval for their support and for always being accommodating to our questions and discussions. Therefore, we would like to thank them for participating in our interviews and workshops where they surprised us time after time with their interest in investigating DeLaval with a carbon emission lens.
Lastly, we want to direct a lot of thank you to our friends, family and loved ones for always having our backs and pushing us in these tough times with a lot of love and support. Writing a master thesis within a pandemic has been interesting but we are happy to conclude that working remotely can work just as fine as being in place at the office.
Emil Deckner and Carl Mailer
Stockholm, June 2020
Table of Contents
ABSTRACT ... I SAMMANFATTNING ... II FOREWORD AND ACKNOWLEDGEMENT ... III LIST OF FIGURES ... VI LIST OF TABLES ... VII LIST OF ABBREVIATIONS ... VIII
1 INTRODUCTION ... 1
1.1 A
IM... 2
1.2 R
ESEARCHQ
UESTIONS... 3
1.3 S
COPE ANDL
IMITATIONS... 3
1.4 D
ISPOSITION OFR
EPORT... 4
2 THEORETICAL FRAMEWORK ... 5
2.1 S
YSTEMT
HINKING... 5
2.1.1 Orientations and Considerations for System Thinking ... 6
2.1.2 System Thinking for Problem Solving ... 7
2.1.3 Social Constructions and Mental Models ... 8
2.1.4 Applications of System Thinking ... 9
2.2 P
ARTICIPATORYM
ODELLING... 11
2.2.1 Purpose of Participatory Modelling ... 12
2.2.2 Examples of Utilising Participatory Modelling ... 13
2.2.3 General Process of Participatory Modelling ... 13
2.2.4 Variations of Participatory Modelling Methodologies ... 15
2.2.5 Methods in Participatory Modelling ... 18
2.3 S
UPPLYC
HAINM
ANAGEMENT... 19
2.3.1 Green and Sustainable Supply Chains... 22
2.4 C
ARBONA
CCOUNTING... 26
2.4.1 Greenhouse Gases... 26
2.4.2 GHG Protocol Standard ... 27
3 METHODOLOGY ... 29
3.1 R
ESEARCHD
ESIGN... 29
3.2 M
ODELLINGS
TRATEGY... 32
3.2.1 First Phase: Conceptual Model... 34
3.2.2 Second Phase: Quantitative Model... 36
3.2.3 Third Phase: Result Verification ... 38
3.3 R
EPLICABILITY ANDV
ALIDITY... 39
3.4 E
THICALC
ONSIDERATIONS... 40
4 CASE STUDY ...41
4.1 I
NTRODUCTION TO THEC
OMPANY... 41
4.1.1 Simplified Supply Chain ... 41
4.1.2 Stakeholders in the Organization ... 42
4.1.3 Sourcing Categories ... 43
4.2 O
UTCOMES OF THEM
ODELLINGP
ROCESS... 44
4.2.1 Scope Definition ... 44
4.2.2 Social Network Analysis ... 45
4.2.3 Model Structure and System ... 46
4.2.4 Stakeholder Interaction in Modelling Process ... 49
4.2.5 Data Sources... 52
4.2.6 Results of Carbon Footprint ... 54
5 GENERAL FINDINGS ... 55
5.1 M
ODELS
TRUCTURE& S
YSTEM... 55
5.2 D
ATAS
OURCES... 56
5.2.1 Internal (Primary) Data Sources ... 56
5.2.2 External (Secondary) Data Sources... 57
5.3 S
TAKEHOLDERI
NTERACTIONP
ROCESS... 57
5.3.1 First Phase: Conceptual Model... 58
5.3.2 Second Phase: Quantitative Model... 59
5.3.3 Third Phase: Result Verification ... 59
5.4 B
ENEFITS OFP
ARTICIPATORYM
ODELLING... 59
5.4.1 First Phase: Qualitative Model ... 60
5.4.2 Second Phase: Quantitative Model... 60
5.4.3 Third Phase: Result Verification ... 62
6 DISCUSSION ... 63
6.1 S
TAKEHOLDERP
ROCESS ANDD
ATAC
OLLECTIONP
ROCESS... 63
6.2 D
ATAS
OURCES AND THEIRR
ELIABILITY... 65
6.3 M
ODEL ANDD
ATAS
OURCEA
DAPTATION... 67
6.4 M
ODELLINGT
RACEABILITY... 68
6.5 F
UTUREW
ORK... 69
7 CONCLUSIONS ...71
8 BIBLIOGRAPHY ... 72
List of Figures
F
IGURE1, T
HEORETHICAL FRAMEWORK... 5
F
IGURE2, V
ISUALISATION OF ORGANISED AND UNORGANISED COMPLEXITY... 8
F
IGURE3, P
ARTICIPATORY MODELLING PROCESS(
ADAPTED FROMV
OINOV ET AL. (2018)) ... 15
F
IGURE4, M
ATERIAL AND INFORMATION FLOWS IN SUPPLY CHAIN(
ADAPTED FROMS
COTT ANDW
ESTBROOK(1991)... 20
F
IGURE5, P
RESSURE ON SUSTAINABILITY IN SUPPLY CHAIN... 24
F
IGURE6, M
ATERIAL AND INFORMATION PRESSURE IN SUPPLY CHAIN(
ADAPTED FROM(H
ILL, 1997)) ... 25
F
IGURE7, P
ROCESS OF GREEN SUPPLY CHAIN STRATEGY(
ADAPTED FROM(P
INTOT
ABORGA ET AL., 2018)) ... 26
F
IGURE8, V
ISUAL REPRESENTATION OF METHODOLOGY... 29
F
IGURE9, P
ARTICIPATORY MODELLING PROCESS... 33
F
IGURE10, S
IMPLIFIEDS
UPPLYC
HAIN OFD
EL
AVAL... 41
F
IGURE11, O
RGANISATIONALS
CHEMED
EL
AVAL... 42
F
IGURE12, S
TAKEHOLDERS ATD
EL
AVAL... 46
F
IGURE13, M
ETHODOLOGYD
IRECTM
ATERIAL... 47
F
IGURE14, M
ETHODOLOGYP
ACKAGING... 47
F
IGURE15, M
ETHODOLOGYT
RANSPORTS... 48
F
IGURE16, M
ETHODOLOGYF
ACILITIES... 49
F
IGURE17, B
USINESST
RAVEL... 49
List of Tables
T
ABLE1, S
COPEC
ATEGORIES(A
DAPTED FROMGHG P
ROTOCOL) ... 27
T
ABLE2, C
ARBONA
CCOUNTINGM
ETHODS... 28
T
ABLE3, S
UPPLYC
HAINS
TAKEHOLDERS ATD
EL
AVAL... 43
T
ABLE4, D
IRECTM
ATERIALC
ATEGORIES... 43
T
ABLE5, I
NDIRECTS
OURCINGC
ATEGORIES... 44
T
ABLE6, S
COPED
EFINITIONA
CCORDING TOGHG P
ROTOCOL... 45
T
ABLE7, D
ISTRIBUTION OFE
MISSIONS PERS
COPE... 54
List of Abbreviations
ANT Actor-Network Theory
APAC Asia, Pacific-region
BBHV Biases, Beliefs, Heuristics and Values
CATWOE Customers, Actors, Transformation, World View, Owner, Enviro..
CO2e Carbon Dioxide Equivalents
DC Distribution Centre (Internal DeLaval Acronym) EMEA Europe, Middle East, Africa
ERP Enterprise Resource Planning
EU European Union
FCM Fuzzy Cognitive Mapping
GHG Green House Gases
ISO International Standard Organisation
MQAH Milk Quality, Animal Health (Internal DeLaval Acronym)
RQ Research Question
SC Supply Chain
SSM Soft Systems Methodology
PDM Product Development Management
PM Participatory Modelling
SNA Social Network Analysis
UNFCCC United Nations Framework Convention on Climate Change
WRI World Resource Institute
WBCSD World Business Council for Sustainable Development
1 Introduction
The question of environmental sustainability has been increasing in interest the last decades and concrete actions must be initiated to prevent global warming and the resulting consequences of a higher mean temperature on earth (IPCC, 2014). The topic of sustainability has become both controversial and complex regarding the measured that needs to be taken. Initiatives from both individuals, countries and international bodies have arisen to fight the global warming.
This includes the decision to make Sweden climate neutral by year 2045 (Miljödepartementet, 2017) and the global Paris agreement, signed by nearly 200 countries in December 2015, aiming to prevent the global warming from exceeding 2° (UNFCCC, n.d.). In Sweden, industries are responsible for 32% of the CO2 emissions annually, which results in a pressure on companies to increase their sustainability work in order to meet the targets set to combat the global warming (Naturvårdsverket, 2019). Because of this, companies are now directing more focus to sustainability issues to follow the regulations set up and to contribute to the general sustainable development (Tseng & Hung, 2014).
By establishing a carbon-mitigation strategy, the company may also get a competitive advantage against competitors (Pinto Taborga, Lusa, & Coves, 2018). Focusing on the environmental consequences of operations to acquire competitive advantage was argued earlier as well as an add-on to Porter’s work on firm competitiveness (Porter & van der Linde, 1995).
Over 60% of consumers get a better perception of a company if they have reduced the carbon footprint of a product and over, indicating that carbon emission reduction could be an important marketing tool (Carbon Trust, 2020). Damert, Paul, and Baumgartner (2017) also found that companies that are highly committed in mitigating carbon-emissions have higher chance of acquiring investors and environmental-conscious customers which may drive financial performance. In the EU, directives are also requiring large companies to make a non-financial reporting, including information about their work within social and environmental sustainability, which increases the focus on sustainability (European Commission, n.d.).
Therefore, there are multiple reasons for companies to work with their carbon-footprint in terms of both regulation, competitiveness and future survival.
There are guidance and standards available to help companies to design mitigation strategies for carbon emissions. Two major ones are the GHG Protocol (WRI & WBCSD, n.d.) and ISO 14064 ((ISO), 2006). They are compatible with each other and generally ISO 14604 are able to tell “What” and the GHG Protocol explains “How” and “Why”. The ISO 14604 standard focuses mainly on auditing current situations, whereas the GHG Protocol focuses on laying out options for reducing the emissions and guidance to calculate the carbon footprint (Pinto Taborga et al., 2018). In order for companies to make well considered decisions to increase their sustainability, firstly they must know the current situation, how much carbon they emit and where the emissions come from. To get a full overview of the environmental impact of the company, the entire supply chain must be taken into multiple considerations, including sourced materials, transports, production, etcetera.
By restricting the focus only to the companies’ own productions, emissions depending on
different providers of materials and logistics services are disregarded. Accordingly, an
important step to improve the sustainability work of the company is to understand the different
sources of emissions from the entire supply chain. To make an analysis of the carbon footprint
of a company, internal or external competences could be used. However, to establish a correct analysis of the emissions, a wide range of information from the company must be used.
Previous studies have discussed green supply chain strategies and its implication on the environmental footprint on companies (Pinto Taborga et al., 2018) and there are also available more in depth analysis regarding both transportation emissions (Pan, Ballot, & Fontane, 2013) and energy-saving technology (Wang, Lai, & Shi, 2011). Many studies are also investigating the carbon footprint of companies, and many companies report the results in their sustainability reports. However, few articles touch the actual working process of calculating the carbon footprint of a company. There are often gaps in between the high-level reporting and how it is connected with the data sources that builds up the analysis. We have also identified a gap since many companies concern themselves with creating reports of their carbon emissions, yet few show transparencies in the methodology used and the assumptions made, which makes the investigations difficult to recreate. This is agreed upon by Busch, Johnson, and Pioch (2020), who identify a gap between the performance indicators and how they were measured, which further indicates the need of establishing a common process for acquiring data to measure the carbon footprint. Because of this, this study will investigate the process of establishing a carbon footprint model of a company.
Participatory modelling is a modelling technique which involves different stakeholders when creating models of reality (D. M. Hall, Lazarus, & Thompson, 2019), which previously have been used to model complex systems with different stakeholders with diverging apprehensions of reality (W. E. Grant, Pedersen, & Marin, 1997). The technique also brings increased knowledge of the investigated system since multiple perspectives are gathered throughout the process (Voinov et al., 2018), leading to increased credence to the results, which also makes them suitable for decision making (Smajgl & Ward, 2013). Due to the complexity of establishing the carbon footprint of a company, participatory modelling possibly could be a useful method to ease the process and to reach more accurate results. Therefore, the process of participatory modelling will be used when creating a model of the carbon footprint of a company, to gather information about the company, but also to involve employees in the process to reach a result that is as accurate as possible. To evaluate participatory modelling as a way of conducting a carbon footprint analysis, a case study will be made on a Swedish, global, manufacturing company, where the results of the model, as well as the process, will be discussed and evaluated.
1.1 Aim
The aim of this master thesis is to investigate if, and in which way, participatory modelling is beneficial when creating a model of the carbon footprint for a global, manufacturing company.
The results will give new information if participatory modelling can contribute to more accurate
carbon footprint models of companies and if it makes it easier for companies to conduct the
analysis by utilising this method. The results will also showcase a process that could be used
for carbon accounting and how the model could be designed.
1.2 Research Questions
In order to fulfil the aim of this thesis, and to understand the benefits and the outcomes from utilising participatory modelling when creating a model of the carbon emissions of a company, the following research questions need to be answered:
RQ1. What are the benefits of utilising participatory modelling?
RQ2. Is participatory modelling more beneficial for some organisational areas within the company?
RQ3. How does the stakeholder interaction look like in the process and when should different stakeholders be involved?
RQ4. What methods could be applicable when creating a model of the carbon footprint?
RQ5. What data sources could be used to complete the model and make it representative for the business processes?
1.3 Scope and Limitations
The thesis will be conducted as a case study at a Swedish, global, manufacturing company, where parts of the global organisation will be used to investigate the process of using participatory modelling. Only emissions connected to the supply chain and its functions will be included in the scope. Due to the time and resource limitations of the thesis, only one case study will be performed, which limits the possibility to compare the results from applications at different companies. The scope of this thesis is only to investigate the result from using participatory modelling and the resulting model. No further analyses of possible changes in the organisation or strategy to increase the sustainability of the company will be made. Therefore, we will not conduct any recommendations for companies wanting to change their strategy since that can change greatly from company to company.
The model will not be described in detail, since the design will need to be adapted to each
company and the data available. Further, limited data will be disclosed in this thesis because of
confidentiality of this information. Instead, we will focus on the overall structure of the model
to describe what type of data sources that can be used and which calculation steps that could be
applied. Furthermore, we will not discuss organisational psychology and organisational
learning since that would widen the scope substantially and possibly divert too much focus from
the modelling process. However, we do believe that this is a logical next step for evaluating the
effect of the participatory modelling, from an organisational perspective.
1.4 Disposition of Report
In this section, we will shortly describe the content of each chapter to guide the reader through this thesis.
Chapter one presents a short introduction into the investigated area of this thesis, motivating on why the aim of this thesis is interesting and important. Furthermore, the aim, research questions and the boundaries of the thesis are also described.
Chapter two presents the theoretical framework of this study. This chapter describes previous findings which is essential for understanding the basis of this thesis. The theoretical framework consists of four major areas: System Thinking, Modelling Strategies, Carbon Accounting and Supply Chain Management. Together, these areas provide the reader with background information needed to understand the theories that this thesis is based on.
Chapter three presents the methodology of this thesis. The chapter starts with a description of the general research design and continues with describing the modelling process used at the company, to create a model of the carbon emissions which includes three major phases.
Chapter four presents the case study performed at DeLaval. In this chapter the company is described, and the outcomes of the modelling process are presented in detail. The results of the study are based on the outcomes from the case study and specific findings regarding DeLaval is also presented.
Chapter five presents the results of the thesis, where the research questions are answered. The results consist of general findings based on the outcomes of the case study.
Chapter six presents a discussion connected to the findings of this thesis. It also includes recommendations on further research questions that could be interesting to investigate, as well as how companies can utilise the results of this participatory modelling within their organisational development.
Chapter seven presents a brief summary of the findings of this thesis.
2 Theoretical Framework
In this chapter, we will describe the theoretical framework used in this thesis. The chapter provides information and knowledge used further on to create a modelling strategy, in order to understand system dynamics and to account for the carbon emissions. The theoretical framework consists of four major pillars: System Thinking, Participatory Modelling, Carbon Accounting and Supply Chain Management, visualised in Figure 1. Together, these areas provide a fundamental basis for the continued study.
2.1 System Thinking
“Managers are not confronted with problems that are independent of each other, but with dynamic situations that consist of complex systems of changing problems that interact with each other. I call such situations messes… Managers do not solve problems, they manage messes. – Russell Ackoff, operations theorist” (Meadows, 2008). This citation illuminates the reason for applying a system thinking perspective on organisations, because managers do not solve independent problems, they manage the total system that the organisation consists of.
Historically, there have been many different interpretations of system thinking and methodologies on how to apply it in different contexts (Reynolds & Holwell, 2010). A number of researchers have tried to describe the concept of systems and have created definitions of such. An early definition of what a system is originates from Young (1964) who described that
“A system is a set of objects together with relationships between the objects and between their attributes”. Churchman (1978) provides an alternative definition: “A system is a set of parts coordinated to accomplish a set of goals…”. Meadows (2008) provides a third: “A system is an interconnected set of elements that is coherently organized in a way that achieves something.
If you look at that definition closely for a minute, you can see that a system must consist of three kinds of things: elements, interconnections, and a function or purpose”.
In general, models of human-environment dynamics contain a way to communicate and organize complex knowledge (Forrester, 1994; Wierzbicki, 2007). With this model, intricate questions could be understood by looking at the dynamics with a holistic view, to further make sense of interactions within the system (Reynolds & Holwell, 2010).
Connections within a system could be described as a flow of information. The systems have both functions and purposes that may or may not be explicitly expressed, thus may only be
Figure 1, Theorethical framework
deducted while observing system behaviour (Meadows, 2008). When created, the model can describe the relationship between entities within a system, which are hard to describe only in writing (Heemskerk, Wilson, & Pavao-Zuckerman, 2003; van der Leeuw, 2004). Ingelstam (2012) defines three foundational components within system theory:
• Components of the systems and their connections – what is inside the system
• Border of the system – what divides the system and its surroundings
• The system surroundings – what is outside the system
Wiener (1961) further discussed the components and connections within the system and discussed that components in the system could be both machineries and humans, which are held connected by an exchange of information or by communication. Meadows (2008) further discussed this and describes that “many of the interconnections in systems operate through the flow of information”. An effective model should also describe the relationships among the system, with a precision that is beyond what mere words are able to describe (Heemskerk et al., 2003; van der Leeuw, 2004). For example, by looking at the world as one system, small actions in our daily lives could be connected to implications on the orangutans in Indonesia’s rainforests and their life qualities. Accordingly, minor actions in one part of the system could affect other parts at far distance, even though the actions in isolation might be small (Reynolds & Holwell, 2010).
One application of system thinking is the Soft Systems Methodology (SSM), which was developed by Checkland (1999), Checkland and Poulter (2006) and Wilson (2001), among others. With SSM, system thinking approaches could be utilised to solve management or business problems. When studying systems, it could be hard to define the main problem. SSM then is a practical and pragmatic approach of identifying and solving soft issues (Burge, 2015).
It was also created as a tool to understand unstructured problem situations by creating a structured system with definitions, and by identifying actions to tackle the problems (Gasson, 1994).
2.1.1 Orientations and Considerations for System Thinking
In order to clarify the workability of system thinking, Reynolds and Holwell (2010) defined three generalised purposeful orientations of systems thinking, describing the reason to utilise this methodology:
• Orientation 1: Understanding and simplifying the relationship between components in a complex situation. The main idea is not to comprehensively understand situations but rather to find ways to improve them by clarify the relationships between them.
• Orientation 2: Illuminating contrasting perspectives in a complex situation. The value in creating systems is often that it creates different perspectives for users in the system.
• Orientation 3: Exploring and investigating power relations, boundary issues or potential conflicts, among different entities in the system or between the different perspectives.
The system approach is gently disrupting the previous connections and enables change
in the system by enlightening its users.
When studying systems, there are some basic considerations to take into account. S. Gray et al.
(2018) notes that the process of modelling is determined by purpose, process of participation, partnerships and the product. Furthermore, Churchman (1978) discusses a number of useful considerations:
• The system objectives, which encompasses the results of the model, or the problem that the model is intended to solve.
• The system environment and the fixed constraints around it, which includes the boundaries and the scope of the system.
• The resources of the system, which include what the system can access and encompass.
• The management of the system, which includes the different stakeholders in the system.
This could be single or multiple stakeholders with diverging interests in said system, which could increase the complexity in decision making.
• The possibility of interactions in the system, which involves defining interactive and fixed parts of the system, where the fixed parts are not changed by the practitioners.
The listed considerations are important to keep in mind for someone thinking about the meaning of a system, while building the system up around a particular situation.
Clark et al. (2016) also discusses the reasoning and purpose for letting a system depict a situation by acknowledging that “Any abstracted representation of a system must be accurate to technical experts, accessible to stakeholders, and politically acceptable to decision makers if it is to be used to guide management actions.” This can be connected to the management consideration discussed by Churchman (1978), and the orientation about lifting contrasting perspectives discussed by Reynolds and Holwell (2010). Further, policy tools could be introduced to change behaviour, for example by depicting on how a given natural resource should be used. In this case, “system representations must be relevant to those living and working within the system, credible to decision authorities, and legitimate to those within and outside of the decision making process” (Clark et al., 2016).
2.1.2 System Thinking for Problem Solving
One major purpose of utilising system thinking, as discussed, is to solve problems. W. E. Grant et al. (1997) discusses a variety of techniques to address issues in complex systems.
Mathematical methodologies are described as typical problem-solving methods in system analysis, however the dynamics within complex systems are difficult to present statistically with mathematics. Within complex systems, one often starts with an “unorganized complexity”, where the number of components in the system are high and the interrelatedness of the components are low. However, analysing the system means that an “organized complexity”
could be reached, by either decreasing the number of components or by increasing the interrelatedness of the components (Bertalanffy, 1968; W. E. Grant et al., 1997). This is illustrated in Figure 2. A similar framework to Grant’s was proposed by Ingelstam (2012), who also introduced two axes with the number of components and their connections respectively.
W. E. Grant et al. (1997) continues with describing the relationship between the amount of data
in a system, the level of understanding of the system and how the methods for describing the
system could be customized depending on the information available. Systems with a high
amount of data could preferably be simulated and processed with mathematical models, while
systems with little data need to be handled by system analysis and simulations. To do so, an
understanding of the system must be established to form hypothesis concerning the structure and function of the system. After this, simulations can be made to replace the missing data and to draw conclusions about the system. In this way, a system with little data and limited understanding could be treated through system analysis, where important areas could be identified, and work could be done to acquire more data in those areas. Over time, this increases the knowledge of the system and data is accumulated in the process, which makes it possible to solve problems within the system.
Especially for systems characterized by “organized complexity” with limited possibility of acquiring complete data, the system approach is suitable for creating an understanding of the system. However, the system approach only provides an efficient way to obtain knowledge and needs to be combined with methods for problem-solving, such as classification and mathematical or statistical analysis (W. E. Grant et al., 1997).
2.1.3 Social Constructions and Mental Models
D. M. Hall et al. (2019) discusses that “people intuitively construct a practical understanding of how a particular system they live in works, and how to live with it”. This creates mental models of social and ecological systems, which the people refer to as their home. These mental models are constructed without explicit analysis and are not articulated (Meadows, 2008;
Westervelt & Cohen, 2012). The mental models, as well as shared cultural models, can be influenced by ideas from the everyday life and thoughts of how the system works. (Glynn, Voinov, Shapiro, & White, 2017; Paolisso, 2002; Özesmi & Özesmi, 2004). Furthermore, Reynolds and Holwell (2010) also discuss that all systems are social constructions, where the description of the system will be a subjective view of the real world built up by the model creators. The systems then work to create a conceptual construction with which to investigate real world entities (Reynolds & Holwell, 2010), similarly to mental models described by D. M.
Hall et al. (2019) .
All stakeholders of a system could be assumed to have their own mental model of a specific system, including a perception of how the system looks and how parts of the system are related and interact with each other. Senge and Sterman (1992) also describe that the mental models
Figure 2, Visualisation of organised and unorganised complexity
can be an important tool for managers to understand situations and to analyse them systematically. Such as within a company with a complex business-related problem.
Furthermore, by combining the mental models from stakeholders in the system, an objective and holistic system model could be built which represents the system in a complete way, since different views and opinions are combined into a single holistic model.
A system model also enables the model builders, alternatively known as the investigators, to acquire insights into the system, since scenarios and hypotheses can be tested. For model builders, Senge and Sterman (1992) defines three stages of an effective learning process where the mental models of the stakeholders are combined, which increases the system thinking within management teams, and in turn eases the work of the model builders:
• Mapping mental models: initial structuring of assumptions utilising system models.
• Challenging mental models: finding inconsistencies in the initial assumption.
• Improving mental models: continuously extend the model and testing it.
These three stages provide an iterative and flexible process that can be used as a foundation to work with system thinking and models for decision-making.
Furthermore, the mental model of the model builders is discussed, where challenges can arise when contradictions to the mental model is brought up because of new data or knowledge from informants and stakeholders. One example is that experienced managers often have accurate perceptions of areas closely connected to their work but have limited understanding of how their area connect with other parts of the system. Therefore, they risk drawing inaccurate conclusions about the system. This puts high pressure on the model builders to get a holistic view of the system and to connect information provided from different managers in the system, to challenge their understanding and to draw accurate conclusions of the system. Consequently, Senge and Sterman (1992) conclude that mental models are challenging and that the creation of a system is an open-ended process of testing; revising the model based on the assumptions and input from the stakeholders. By doing this thoroughly, awareness could be extended to the managers in order to enlighten them about important feedback dynamics of the system and linkages between them and other functions inside and outside the organization.
Throughout the process of building a model, the stakeholders taking part of the process need to clearly articulate their assumptions concerning the system , much like the modellers need to account for the different perceptions of the stakeholders (Krebs, 2000; Voinov, Seppelt, Reis, Nabel, & Shokravi, 2014). It is also important to consider that the model is a tool which needs to be clarified throughout the modelling process (D. M. Hall et al., 2019). Consequently, clarifications regarding the system are important and the purpose of the model needs to be agreed on among the stakeholders. This include the usage of the model as the limitation throughout the modelling process. By clearly framing the process, it is easier to address differences in opinions and expectations from the start, resulting in a more uniform decision- making process (D. M. Hall, Lazarus, & Swannack, 2014).
2.1.4 Applications of System Thinking
System thinking could be useful regardless if you are working within a company and wants to
improve the business, or if you are a policy maker who wants to intervene in a collective issue
(Meadows, 2008). Especially in situations characterised by complexity and uncertainty of
feedback effects, system thinking, the model creation can be useful. In general, models could fall under two different categories: Descriptive and predictive models (Murray, 2007).
Descriptive models are designed to explain system components and dynamics both qualitatively and quantitatively. Models can also provide a simplified representation of the big picture of an abstract perspective of a system for decision-making. This can improve the ability to evaluate problems, assess management strategies or test hypotheses (W. E. Grant & Swannack, 2008; S.
Gray et al., 2018). Models can also be used as predictive tools, “for stimulating biophysical systems to forecast trends, evaluate social-ecological interactions under varying management scenarios”, allowing comparison of how different changes in the strategy effect the system and the future outcomes of it (D. M. Hall et al., 2019).
The usefulness of system models was also discussed earlier by Schön (1983), who described the important role of models in learning among professionals. For example, models provide possibilities to experiment with decisions without compromising the actual business. Instead, only successful actions and hypotheses from the modelling procedure could be implemented in the real-world business, with fewer decisions effecting the organisation negatively. One example is during a case study performed by Senge and Sterman (1992) at Hanover Insurance, where the objective was to reveal to managers how different decisions could affect the company’s financial results both directly and indirectly. Hanover had a traditional management style with extensive autonomy where the need was extensive to gather a complete picture.
Therefore they implemented a system-based model which allowed managers to see “the full picture”, which in turn enabled them to make better decisions (Senge & Sterman, 1992).
The Role of the Modeller
The modeller, or investigator, plays a central part in creating the model (Senge & Sterman, 1992) and needs to adopt the role as an expert with the role of understanding the system (Gasson, 1994). To understand the system, and to get information in order to create the model, the investigator can interview potential system users and stakeholders in order to gather expert knowledge concerning the system. However, the investigator might not get a complete understanding of the system until the organisational users test the system. To deal with this problem, a system prototype could be used to try the system under the development phase.
Comments and feedback from the users could then be utilised to improve the system and iteratively finish the model, by making changes and retrying it on the users. In complex systems, it can be important to include a large number of users to collect information, since each person might only be familiar to a restricted part of it (Gasson, 1994). However, the investigator should not be regarded as an expert with vast experience within the field of the system. Instead the investigator should be a facilitator who can provide support to the organisations in order for them to define their own system, by combining different perspectives and making compromises.
To secure a reasonable final product, it is important to gather different perspectives from the system to create a representative sample of the organisation. If this is not done properly, the model might end up skewed and only serve parts of the organisation (Gasson, 1994).
System Engineering Methods
As part of the system approach there are a number of methods and approaches that could be
utilised to accommodate modelling process. This includes both the process of understanding
and modelling the system. Burge (2015) defined a set of tools based on the Soft Systems
Methodology to facilitate the process of setting up a system based on this methodology:
• Rich Picture: Visualizes the system in pictures, graphs and diagrams, since they are easier to understand than descriptions in words. In this way, a common ground for the model is created and agreed on by the users (Burge, 2015). Another definition tells that rich pictures utilises parts of pictures, texts, symbols to showcase how a group of people see a certain situation which is analogue to the description (Bell & Morse, 2013).
• Root Definition and CATWOE: Describes the essence of the particular system, what the outcomes of the model should be, as well as the transformation role of the system. The route definition could be evaluated and reworked with the help of a CATWOE (Customers, Actors, Transformation, World View, Owner, Environmental Constraints) to ensure that the input and output parameters of the systems are logically integrated (Checkland & Smyth, 1976). The root definition is important when creating the system, since it becomes clear to all users what the purpose of the model is. This ensures that all participants work toward the same goal in building the system with less discrepancies (Burge, 2015).
• Conceptual Model: Shows the necessary activities in the system and the dependencies between them. By starting with investigating the activities in the conceptual model and comparing them to reality, the modeller can identify potential changes and improvements of the model. This ensures that different activities and processes are logically connected with each other and that the wished for results are presented (Burge, 2015).
2.2 Participatory Modelling
The process of utilising groups of stakeholders and groups of experts to construct integrated and collectively constructed “representations of coupled social-ecological systems” is generally called “participatory modelling” (D. M. Hall et al., 2019). There are also a number of other definitions for participatory modelling, where Stave (2010) defines the process as “… an approach for including a broad group of stakeholders in the process of formal decision analysis”. Furthermore, Schmolke, Thorbek, DeAngelis, and Grimm (2010) all describe participatory modelling as “a tool for understanding, conceptualising and communicating complex systems to evaluate and improve policy”. Cornwall (1995) defines participatory research as “putting an emphasis on a “bottom-up” approach, with a focus on locally defined priorities.” The concept of participatory modelling encompasses several different stakeholder- based modelling approaches. Meanwhile, the different modelling approaches are closely related to each other, utilising the concept of acquiring knowledge through multiple sources instead of a single or pair of researchers, with limited knowledge and experience of the system (Cornwall, 1995). In summary, they are all parts of participatory modelling and utilising this methodology could enhance effectiveness, saving time and money in the long term.
The process also constitutes social learning (Matt Hare, 2011; M. Hare & Deadman, 2004; Pahl-
Wostl & Hare, 2004) in the same way as collaborative learning (D. M. Hall et al., 2019). This
expresses the connection between participatory modelling and the soft system methodology
described in section 2.1. According to D. M. Hall et al. (2019), stakeholders and experts can be
part of the participatory process for different reasons. Stakeholders may care about the specific
system in question, depending on how it functions, and may subsequently want to influence the
policy that regulates the system or as a contrast, making sure that it does not become falsely
modelled (S. Gray et al., 2018; Voinov & Gaddis, 2008).
2.2.1 Purpose of Participatory Modelling
By utilizing participatory modelling, the idea is to bring multiple perspectives, better data and a more complete picture of the situation and the problems in the system. Furthermore, by utilising participatory modelling, knowledge from different stakeholders are combined to create one shared model of reality which promotes system understanding and awareness in the organisation (Voinov et al., 2018). In extension, this could be used as a tool for conflict management and resolution within the organisation since the shared understanding is created (R. Fischer, Ury, & Patton, 2011; D. M. Hall et al., 2019). Because of the shared understanding that is created, participatory modelling could also be used as a decision tool which makes a diverse group of stakeholders work towards the same goal (D. M. Hall et al., 2019).
There have also been multiple studies which demonstrate that participatory processes produce better outcomes than traditional top-down processes (Etienne, Du Toit, & Pollard, 2011;
Lynam, de Jong, Sheil, Kusumanto, & Evans, 2007; Reed, 2008; Voinov & Bousquet, 2010;
Voinov & Gaddis, 2008). In this perspective, participatory modelling brings collaborative learning into the process. The learning process in participatory modelling is referred to as a double loop learning process (Argyris & Schön, 2002). This means that stakeholders could question and learn about the system and modelling themselves. This may change their fundamental hypotheses, values, norms, and beliefs depending on the results of the modelling (Zellner & Campbell, 2015). This is opposed to the single loop learning “where individuals and groups act within a single reference frame and where specific hypotheses, values, norms, beliefs and objectives are assumed to describe the world.” (Argyris & Schön, 2002). By combining multiple mental models, participatory modelling can also modify the perception of the organisation and/or the system (Daré et al., 2014). This is similar to the modelling processes described by Senge and Sterman (1992) and Gasson (1994), which includes major elements of participatory modelling. The participatory modelling process also enables groups to learn through their diverse experiences and perspectives, which exposes data gaps and information needs (Ritzema, Froebrich, Raju, Sreenivas, & Kselik, 2010).
Because of the shared model that is created through the modelling process, the notion that the stakeholders have about the system could be changed and the mental models among the stakeholders could be aligned. In return, this could also help the stakeholders to identify common issues and goals (Daré et al., 2014). Consequently, the actual modelling process in itself acts as knowledge creation for participants at both the “single loop” and the “double loop levels (Argyris & Schön, 2002; Voinov et al., 2018).
As well as benefits with this modelling strategy, there are also risks and limitations and subsequently some considerations to be aware of. One issue is that the representation of the system is only a selection of reality, and may not consider Biases, Beliefs, Heuristics and Values (BBHV) (D. M. Hall et al., 2019). These are cultural differences which are important to take into consideration since good facilitators could help participants to recognize, mitigate and shape their BBHV to improve the participatory modelling (PM) process. In turn, this could reduce the risk of biased and skewed results, as well as increase the understanding of both social and physical contexts, yielding more robust decision-making instruments (Glynn et al., 2017).
Another risk is that simplifications and removal of variables could be needed which could
change the system dynamics and lead to over-simplifications. If not done properly, this could
result in skewed results and decisions based on incorrect data (Allen, Zellmer, & Wuennenberg, 2005; W. E. Grant & Swannack, 2008; D. M. Hall et al., 2019; Murray, 2007).
2.2.2 Examples of Utilising Participatory Modelling
Participatory modelling can be useful in multiple applications, especially when evaluating situations with high complexity, a large number of components, and/or stakeholders with little initial knowledge, as described by W. E. Grant et al. (1997). We have chosen to give some examples of such applications. The different applications range over different subjects and levels such as consulting, discussing, co-designing, co-decision-making and decision-making (Mostert, 2003).
Environmental resources are a prominent theme for participatory modelling studies. This is because this issue often includes a large number of stakeholders with conflicting interests and decisions, potentially causing major effects on the participants, both positively and negatively, directly and indirectly. One example is the study on community water resources which aims to identify new solutions to manage them efficiently and fair (Gaddis, Falk, Ginger, & Voinov, 2010). In another study, participatory modelling was used to improve fishing practices on the coast of Nicaragua (Casillas & Ray, 2018).
Public policymakers have also utilised this method to ensure diverse input and to ensure a balanced policy result towards stakeholders in the model (Smajgl & Ward, 2013). Furthermore, a combination between public policymaking and environmental resources was investigated when assessing the wind energy of a country, utilizing participatory modelling (Höltinger, Salak, Schauppenlehner, Scherhaufer, & Schmidt, 2016). A similar study also used this framework to study the solar energy potential for a village (Krzywoszynska et al., 2016).
A common denominator for the different studies utilising participatory modelling is that the results from the modelling are used for decision-making. By gathering different perspectives, a solution could be found with compromises that are beneficial to all stakeholders, regardless of previous power structures in the system.
2.2.3 General Process of Participatory Modelling
There are several alternative descriptions of typical process methodologies for participatory modelling in literature. We will focus on two major descriptions. The first method was created by D. M. Hall et al. (2019) and is based on an iterative process with three phases:
• In the first phase, the problem definition is made (van den Belt, 2004), and mental and cultural models of the system are gathered from participants (Andersen & Richardson, 1997; Luna-Reyes et al., 2006). This could be done by asking participants to describe and picture their understanding of the system, identify variables or by mapping model components (Jones, Ross, Lynam, Perez, & Leitch, 2011).
• In the second phase, a shared conceptual model of the system is set up and agreed on by the participants, including flows within the model and boundary feedback loops. This is done by facilitated discussions and system modellers working together with the participants to adjust the model and to gather more data to the model if needed. (van den Belt, 2004)
• In the third phase, the model is tested and verified with the participants. Different
scenarios could be tested and the outcomes could be evaluated (van den Belt, 2004).
This modelling process described by Hall et al. could be connected to the process of collecting mental models from participants, described by Senge and Sterman (1992), where mental models are gathered, combined and challenged in three phases to find one shared understanding of the system.
Voinov et al. (2018) provide a second process of participatory modelling, where the different phases also include different tools and methods. A tool is defined as “a modelling technique used to carry out a particular function to achieve a specific goal”, while methods are defined as
“a way of doing something, in particular, a way of using tools”. By setting up the methods and tools correctly, the process gets more efficient and ensures that all participants’ feedback are handled equally without biases of participant opinions (Voinov et al., 2018). The modelling process set up by Voinov et al. (2018) includes five stages which contain a number of methods, each of which will be described generally. Figure 3 shows a visual representation of the process.
• Fact finding: The first stage focuses on “finding, generating, and communicating data, information, and knowledge relevant to the problem considered”. This stage can be reoccurring throughout the process and could consist of a traditional literature study as well as interviews, surveys and crowdsourcing. Albeit that fact finding typically occurs in the beginning of the modelling process (Voinov et al., 2018).
• Process Orchestration: The second stage is informal and depends on the set up of the organisation. The orchestration surrounds all of the different stages in the process and could assume the form of three methods: Facilitation, Role-playing games or Brainstorming. In this stage, it is important to foster a collaborative environment, allowing all stakeholders to express themselves and create mutual learning and understanding (Voinov et al., 2018). To improve facilitation it is proposed to actively document and record the facilitation process and the results of which, increasing transparency, as well as enabling reconstruction and analysis of the outcomes (Voinov et al., 2018). The facilitator should work with the procedure and process of the model, rather than interfere with the content of the modelling (Vennix, 1999).
• Qualitative Modelling: During the third stage a visual representation of the included components of the system is created. This can be done with methods such as Rich Picture or Cognitive/Concept Mapping (Voinov et al., 2018).
• Semi-Quantitative: The fourth stage is a compromise between qualitative and quantitative modelling. The stage is described as a necessity when the data is qualitative or semi-quantitative, which may imply qualitative information, numeric estimates or values agreed upon by the participants. The major methods used at this stage are Fuzzy cognitive mapping (FCM) and Social network analysis (SNA) (Voinov et al., 2018).
• Quantitative Modelling: The fifth and final stage includes the actual quantitative
modelling, which should utilise the already collected and reworked input data. At this
stage, the main idea of the model is already defined, while quantitative tools are used
implement data to deliver results from the model. Some applicable methods at this stage
are Cost-Benefit Analyses (CBA) and Agent Based Modelling (ABM) (Voinov et al.,
2018).
Figure 3, Participatory modelling process (adapted from Voinov et al. (2018))