2. RapportRIVMWaardevGroen.pdf

Amsterdam’s Green Infrastructure Valuing Nature’s Contributions to People

RIVM Letter report 2019-0021 M. Paulin │ R. Remme │ T. de Nijs

Amsterdam’s Green Infrastructure Valuing Nature’s Contributions to People RIVM Letter Report 2019-0021 M. Paulin │ R. Remme │ T. de Nijs

RIVM Letter report 2019-0021

Colophon

© RIVM 2019 Parts of this publication may be reproduced, provided acknowledgement is given to the: National Institute for Public Health and the Environment, and the title and year of publication are cited. DOI 10.21945/RIVM-2019-0021 M. Paulin (author), RIVM R. Remme (author), Stanford University T. de Nijs (author), RIVM Contact: Ton de Nijs Centre for Sustainability, Environment and Health (DMG) ton.de.nijs@rivm.nl This investigation was performed by order of the City of Amsterdam in the project “The Value of Amsterdam’s Green” and the Ministry of Agriculture, Nature and Food Policy (LNV) in the project “Doorontwikkeling TEEB-Stad” (M360013).

This is a publication of: National Institute for Public Health and the Environment P.O. Box 1 | 3720 BA Bilthoven The Netherlands www.rivm.nl/en

Page 2 of 79

RIVM Letter report 2019-0021

Synopsis

Amsterdam’s Green Infrastructure Valuing Nature’s Contributions to People The Municipality of Amsterdam developed a strategy to enable its growing population to better enjoy the benefits provided by nature, while endowing it with a more attractive living environment. The strategies in the Quality Impulse Green (KwaliteitsImpuls Groen) were translated into four scenarios that describe how the city’s green infrastructure will be expanded and improved. For every scenario, RIVM quantified and mapped the benefits that ‘green and blue’ generate for human health and well-being, as well as for mitigating the impacts of climate change, for instance by storing rainwater and excess heat. While the values quantified vary per scenario, the benefits are considerable, between 1 and 5 €/m2 additional green per year. Benefits were found to be most prominent in densely populated areas with relatively low amounts of green space. Urban green, such as parks, canals, and green roofs, can lead to benefits that improve the quality of the living environment. While some of the benefits are not difficult to envision, such as the use of parks as recreation sites, green also contributes to complex processes, such as enhanced air quality, cooling temperatures in cities, more biodiversity, and the storage of water in the soil, which leads to reduced pressure on sewer systems.

Scenarios contain different aspects from the green infrastructure strategy. The scenario ‘Green Neighbourhoods’ entails an increase in the amount of green within the city centre, for instance, by increasing the amount of green roofs and trees. In the ‘Green Networks’ scenario, the tree structure is strengthened and the connections between Amsterdam’s network of recreational areas is improved. In the scenario ‘Urban Parks’, the size and quality of existing parks is enhanced, and new parks are created. Finally, in the scenario ‘Metropolitan Parks’, metropolitan areas in the outskirts of the city are optimized for recreational use by expanding the cycling and walking network, making such areas more accessible. In order to quantify the value generated by green and blue, the Natural Capital Model, developed by RIVM, Wageningen Environmental Research (WEnR), and the Netherlands Environmental Assessment Agency (PBL), was implemented. Keywords: urban ecosystem services, green infrastructure, Natural Capital Model, social cost benefit analysis, SCBA, indicator

Page 3 of 79

RIVM Letter report 2019-0021

Page 4 of 79

RIVM Letter report 2019-0021

Publiekssamenvatting

Amsterdam’s Groen Waardering van de bijdrage van de natuur aan de mens De gemeente Amsterdam heeft een strategie ontworpen om de waarde van natuur en water in de stad beter te benutten en zo het groeiende aantal inwoners een aantrekkelijke leefomgeving te bieden. In deze KwaliteitsImpuls Groen zijn vier scenario’s opgesteld om de groene infrastructuur van Amsterdam uit te breiden en te verbeteren. Het RIVM heeft per scenario in kaart gebracht wat investeringen in ‘groen en blauw’ opleveren voor de gezondheid en het welzijn van de stedelingen en om de gevolgen van klimaatverandering, zoals wateroverlast en hitte, op te vangen. Dat varieert van minder zieken, minder ziekteverzuim en meer lichamelijke activiteit, tot een betere luchtkwaliteit, minder kosten voor de waterzuivering en een hogere vastgoedwaarde. De baten verschillen per scenario maar blijken aanzienlijk, tussen de 1 en 5 €/m2 extra groen per jaar. De baten zijn het hoogst in wijken waar nu weinig groen is en veel mensen wonen. Natuur in de stad, zoals parken, grachten en groene daken, kunnen de kwaliteit van de leefomgeving verbeteren. Sommige vormen zijn duidelijk zichtbaar, zoals parken en recreatiegebieden. Maar al het groen draagt ook bij aan complexere processen, zoals een betere luchtkwaliteit, verkoeling van de stad, biodiversiteit en de berging van water in de bodem waardoor het riool minder wordt belast.

De scenario’s verschillen van opzet. Het scenario Groen Dichtbij betreft meer groen in de binnenstad, bijvoorbeeld met behulp van groene daken en meer bomen. In het scenario Groenblauwe Verbindingen worden de boomstructuur in de straten hersteld en in de hele stad recreatieve fiets- en wandelpaden aangelegd. In het scenario Stadsparken worden bestaande parken verbeterd en nieuwe aangelegd. Ten slotte worden in het scenario Metropolitane Parken de grote parken aan de rand van de stad ontsloten voor recreatief gebruik door de uitbreiding van het fiets- en wandelnetwerk. Om de waarde van groen en blauw te kunnen kwantificeren is in deze studie gebruikgemaakt van het Natuurlijk Kapitaal Model, dat is ontwikkeld door het RIVM, Wageningen Environmental Research (WEnR) en het Planbureau voor de Leefomgeving (PBL). Kernwoorden: stedelijke ecosysteemdiensten, groene infrastructuur, Natuurlijk Kapitaal Model, maatschappelijke kosten-batenanalyse, MKBA, indicator

Page 5 of 79

RIVM Letter report 2019-0021

Page 6 of 79

RIVM Letter report 2019-0021

Contents

Summary — 9 1

Introduction — 11

2 2.1 2.2 2.2.1 2.2.2 2.2.3 2.2.4 2.2.5 2.3 2.4 2.4.1 2.4.2 2.4.3 2.4.4 2.4.5 2.4.6 2.4.7 2.5 2.5.1 2.5.2 2.5.3 2.5.4 2.5.5

Methods — 15 Social cost-benefit analysis — 15 Scenarios — 15 Business-As-Usual — 16 Green Neighbourhoods — 16 Green Network — 21 Urban Parks — 21 Metropolitan Landscape — 21 Green Benefit Planner — 21 Ecosystem services and their benefits — 22 Air quality — 23 Health — 25 Physical activity — 25 Property value — 26 Recreation — 26 Urban cooling — 26 Water storage — 27 Input data — 27 Housing Plans Map — 28 Inhabitants Map — 28 Land Cover Map — 28 Vegetation Maps — 29 Further input data — 30

3 3.1 3.1.1 3.1.2 3.1.3 3.1.4 3.1.5 3.1.6 3.1.7 3.2 3.2.1 3.2.2 3.2.3 3.2.4 3.2.5 3.2.6 3.2.7 3.3

Results — 31 Business-As-Usual — 31 Air quality — 31 Health — 32 Physical activity — 35 Property value — 37 Recreation — 37 Urban cooling — 37 Water storage — 40 Changes in ecosystem service values across scenarios — 42 Air quality — 42 Health — 45 Physical activity — 46 Property value — 49 Recreation — 49 Urban cooling — 51 Water storage — 51 Associations between ecosystem services — 54

4 4.1 4.1.1

Discussion and conclusions — 57 Supply and demand — 57 Supply — 57 Page 7 of 79

RIVM Letter report 2019-0021

4.1.2 4.2 4.2.1 4.2.2 4.2.3 4.2.4 4.3 4.4 4.4.1 4.4.2 4.4.3 4.4.4

Demand — 58 Limitations — 58 Qualitative assessment — 58 Aggregation — 58 Monetization — 59 Model uncertainty — 59 Uses and recommendations — 61 Implications for Amsterdam’s Structural Vision — 61 Quality Impulse Green strategy 1: Urban Parks — 61 Quality Impulse Green strategy 2: Climate and biodiversity — 62 Quality Impulse Green strategy 3: Green Neighbourhoods — 62 Quality Impulse Green strategy 4: Green Networks and accessibility — 63 References — 65 Appendix 1 Social cost-benefit analysis — 67 Appendix 2 Additional models — 68 A2.1 Health — 68 A2.2 Physical activity — 69 A2.3 Recreation — 76 A2.4 Water storage — 78

Page 8 of 79

RIVM Letter report 2019-0021

Summary

Amsterdam’s ‘Structural Vision 2040’ (Structuurvisie 2040) comprises the roadmap for the city’s spatial development until the year 2040 (Gemeente Amsterdam, 2011). Its main objectives include enhancing the attractiveness and quality of public spaces, increasing the availability and use of natural capital, and reducing Amsterdam’s contribution to climate change. Addressing these challenges, Amsterdam’s Green Agenda 2015-2018 (Agenda Groen 2015-2018) laid out strategies for the accomplishment of these objectives, namely by strengthening the city’s green infrastructure (Gemeente Amsterdam, 2015). Its successor, the ‘Quality Impulse Green’ (Kwaliteitsimpuls Groen), will further outline the implementation approach of the city’s green infrastructure strategy, additionally considering the pressures entailed by increasing population growth (Gemeente Amsterdam, 2017). To inform this initiative, this study explores the potential impact of the implementation of green strategies envisioned in the Quality Impulse Green on natural capital and its contributions to human health and well-being. The strategies presented within the Quality Impulse Green were translated into four spatial scenarios, expressed as maps, by urban design firm ‘De Urbanisten’.

These maps were used as input to quantify ecosystem services within a Business-As-Usual situation in 2025, where no change other than the anticipated population and residential growth takes place, and within the four scenarios. By comparing ecosystem service quantities in the Business-As-Usual (reference) scenario against the four scenarios, it is possible to determine the costs and benefits that would result from the implementation of different green infrastructure measures. This information can be adopted for its inclusion in social cost-benefit analyses, conform the ‘General guidelines for social costbenefit analyses’ (Algemene leidraad voor maatschappelijke kostenbatenanalyse) and the ‘Nature Guideline’ (Werkwijzer Natuur). These guidelines describe the steps that must be fulfilled prior to the implementation of policy strategies that affect the quantity and quality of natural capital within the Netherlands (Arcadis & CE-Delft, 2018; Romijn & Renes, 2013). Seven ecosystem services were quantified, expressed in 17 physical, social, and monetary indicators. The Green Benefit Planner, a spatial decision-support tool from RIVM, was used to quantify ecosystem services and evaluate fluctuations in values associated with the implementation of every spatial-strategy developed by De Urbanisten.

The Green Benefit Planner makes use of models from the Natural Capital Model (NK-Model) (Remme et al., 2018), a collaborative modelling platform from RIVM, Wageningen Environmental Research (WEnR), and The Netherlands Environmental Assessment Agency (Planbureau voor de Leefomgeving, PBL), to quantify ecosystem services. Additional models were developed to generate a general overview of the benefits of ecosystem services produced by Amsterdam’s green infrastructure to human health and well-being, and to evaluate the linkages that exist among them. It was found that substantial benefits can be accrued through the enhancement of Amsterdam’s green infrastructure. The values vary per scenario between 1 and 5 €/m2 additional green per year. Two scenarios Page 9 of 79

RIVM Letter report 2019-0021

generally seem to accrue the highest benefits within the Amsterdam region, namely the ‘Green Neighbourhoods’ and the ‘Green Network’ scenarios. The Green Neighbourhoods scenario entails enhancing the amount of green in the near vicinity of households, homogeneously distributed across Amsterdam’s most populated areas, including areas where green is currently scarce. Relative to other scenarios, this one was found to yield high potential improvements in terms of increased physical activity, less visits to general practitioners and increased property value. The ‘Green Network’ scenario entails strengthening Amsterdam’s ecological network, including adding trees in segments connecting existing green areas. This scenario was found to yield substantial benefits in terms of enhanced physical activity, reduced visits to general practitioners, temperature regulation, and water storage within the city. The ‘Urban Parks’ scenario includes expanding and enhancing the quality of existing parks, and creating new. It was found to lead to relatively high potential improvements in air quality and to more opportunities for recreational visits to green areas. The ‘Metropolitan Landscape’ scenario did not entail any alterations in Amsterdam’s green infrastructure but rather an enhancement in the accessibility to green areas through the addition of walking and cycling paths.

This scenario was found to lead to relatively large potential increases in physical activity and greater opportunities for recreational visits to green areas. Regarding ecosystem service demand and supply, it was found that the supply of ecosystem services is mainly determined by the amount, type, and distribution of vegetation, while the demand is mainly determined by the population density. Population density is the main determinant for ecosystem service demand since it is an indicator for the number of ecosystem service beneficiaries in an area and the environmental pressures they exert. Across scenarios, more green generally entails a higher delivery of ecosystem services, yet the distribution of green also plays an important role. For instance, adding green homogeneously across large areal extents, especially in areas where green is currently scarce, leads to a higher total increase in ecosystem service delivery per spatial unit than adding green within specific locations, especially those that already contain substantial amounts of green. In addition, trees often lead to higher ecosystem service benefits compared to other categories of green, namely shrubs and grass. Future research should focus on the further development of models for quantifying and mapping ecosystem services at different scales.

For instance, models for quantifying and mapping water storage by vegetation and recreational opportunities in green and blue areas within cities, were not readily available. Models tailored to the Netherlands and more specifically to different regions or levels of urbanization could also lead to more accurate results and conclusions. This would enhance the explanatory and predictive capability of such models to support spatial planning. Participatory semi-quantitative models could also be adopted to identify stakeholder preferences and their knowledge gaps regarding ecosystem services. This could shed light on niches that require further attention by decision-makers involved in spatial planning, contributing to inclusive and legitimized decision-making.

Page 10 of 79

RIVM Letter report 2019-0021

1

Introduction

“Amsterdam: economically strong and sustainable” (Gemeente Amsterdam, 2011, p.9), is the motto of Amsterdam’s Structural Vision 2040 (Structuurvisie 2040), which lays out the roadmap for the city’s spatial development until the year 2040. The Structural Vision recognizes the challenge of maintaining a robust economy in the face of a rapidly growing population and the environmental pressures it entails. It envisions Amsterdam as the core of an internationally competitive, sustainable, and European metropole (Gemeente Amsterdam, 2011). The main objectives include enhancing the attractiveness and quality of public spaces, increasing the availability and use of natural capital, and reducing Amsterdam’s contribution to climate change. Amsterdam’s Green Agenda 2015-2018 (Agenda Groen 2015-2018) laid out strategies for the accomplishment of these objectives, namely by strengthening the city’s green infrastructure (Gemeente Amsterdam, 2015). Its successor, the ‘Quality Impulse Green’ (Kwaliteitsimpuls Groen), will further outline the implementation approach of the city’s green infrastructure strategy, additionally considering the pressures entailed by a growing population (Gemeente Amsterdam, 2017). Green and blue elements (also ‘green infrastructure’), such as parks, canals, and green roofs, possess more than economic and recreational qualities.

While some of the benefits that green infrastructure generates are not difficult to envision (e.g. parks as recreation sites), ecosystems fulfil a gamma of complex functions within cities, crucial for human survival and well-being (e.g. vegetation that contributes to the retention of air pollution and to heat mitigation). Despite significant advances, knowledge on the benefits generated by green infrastructure remains fragmented and hence tends to be overlooked within decision-making. In order to build a case for the implementation of green infrastructure measures, it is necessary to identify and quantify the environmental costs and benefits that would result from their implementation. Measurements can be integrated within social cost-benefit analyses (SCBA) (Romijn & Renes, 2013; Arcadis & CE-Delft, 2018) to evaluate the contributions of green infrastructure to society. By making nature’s contributions to people tangible, it is possible to build a scientific basis for the implementation of nature-based solutions within cities, and to get policy-makers and investors aboard this mission. A promising concept for measuring nature’s contributions to people is the ecosystem services concept. ‘Ecosystem services’ (MA, 2005), or ‘nature’s contributions to people’ (Diaz et al., 2018), are the benefits that natural capital generates for humans.

Similar to goods and services in an economy, ecosystem services can be tangible or intangible, and are classified into three categories, conform the Common International Classification of Ecosystem Services (CICES): provisioning, regulating, and cultural services. ‘Provisioning services’ are the material benefits nature provides to humans, such as trees for timber or food for feeding the rapidly growing human population. ‘Regulating services’ are natural processes Page 11 of 79

RIVM Letter report 2019-0021

that fulfil different functions that benefit humans, such as the role vegetation plays within carbon sequestration and air quality regulation. While often crucial for human survival, regulating services are often intangible or invisible. This leads to a general lack of awareness of their existence and significance, and often results in their overexploitation. ‘Cultural services’ are non-material benefits provided to humans by natural capital, such as the spiritual or recreational value people assign to different natural elements (TEEB, 2010). Often neglected within markets, these benefits are essential for human well-being and form a core element of everyday life.

Figure 1.1: Forest ecosystem services (source: De Urbanisten) The ecosystem services concept adds to our understanding of fundamental ecological structures and processes by assigning labels to their ‘usefulness’ in maintaining and enhancing human well-being, including our survival needs (Burkhard & Maes, 2017). While natural capital provides us with the preconditions required for human existence, their accrued benefits materialize in various forms, including enhanced health, recreation, and higher property values. The delivery of these final ‘benefits’ often requires additional human or built capital inputs, such as human labour for the extraction of resources and roads for their transportation to the end-user (Burkhard & Maes, 2017). The Green Benefit Planner is a tool developed to deepen decisionmakers’ understanding on the effects of spatial planning on ecosystem services and how this, in turn, affects a city’s inhabitants. It does so by Page 12 of 79

RIVM Letter report 2019-0021

quantifying ecosystem service indicators associated with green and blue elements found in the area under assessment, and identifying potential fluctuations in ecosystem service values given the application of different scenarios. Ecosystem service indicators adopted can reflect physical quantities (e.g. particulate matter retention by plants), social quantities (e.g. the effects of particulate matter retention on human health), and monetary quantities (e.g. the reduced health costs associated with increased particulate matter retention) (Remme et al., 2018). This study explores the potential impact of the implementation of strategies from the Quality Impulse Green on natural capital and its contributions to human health and well-being (ecosystem services). It does so by making use of the Green Benefit Planner, which adopts models from the Natural Capital Model (NK-Model) (Remme et al., 2018). In order to depict a wide-ranging and diverse set of ecosystem services, an additional set of models were designed for this study’s purpose. The aim of the study is to generate insights on the value that green and blue generates for the City of Amsterdam and provide quantitative information that can be integrated into SCBAs, endorsing the scientifically-sound and sustainable management of Amsterdam’s natural capital.

Ultimately, this will contribute to the achievement of the green and blue objectives envisioned within Amsterdam’s Structural Vision. Section 2 presents the methods used within this study to quantify ecosystem services within the Business-As-Usual (no-change) situation in 2025, and within envisioned green infrastructure scenarios. Detailed information is provided on the ecosystem services quantified, and the models, indicators, and input data used for their quantification. Section 3 presents the results of this study, including ecosystem service quantities and maps the year 2025. Section 4 discusses this study’s results and concluding remarks, including the implications of implementing green infrastructure strategies from the Quality Impulse Green on the well-being of Amsterdam’s citizens.

Page 13 of 79

RIVM Letter report 2019-0021

Page 14 of 79

RIVM Letter report 2019-0021

2

Methods

2.1

Social cost-benefit analysis Social cost-benefit analyses (SCBA) weigh the marketed and nonmarketed costs and benefits associated with recommended policies against each other to endorse objective decision-making (Romijn & Renes, 2013). In the Netherlands, policy-related SCBAs are required to follow a set of guidelines, which were published by the Central Planning Office (Centraal Planbureau, CPB) as the ‘General guidelines for social cost-benefit analyses’ (Algemene leidraad voor maatschappelijke kosten-batenanalyse) . The ‘Nature Guideline’ (Werkwijzer Natuur) is a subset of guidelines for conducting SCBAs concerning policies that lead to changes in natural capital and the subsequent effect on human wellbeing (Arcadis & CE-Delft, 2018). The Nature Guideline emphasizes the need to measure the changes in ecosystem services and biodiversity that result from policies that affect natural capital. A reference alternative (nulaternatief), here referred to as the ‘Business-As-Usual’ alternative, is used as a benchmark to compare the costs and benefits that would result from the implementation of a strategy, relative to a situation where no changes other than the anticipated occur.

Within study, four scenarios were developed comprising strategies from Amsterdam’s Quality Impulse Green, to evaluate how changes in natural capital resulting from the implementation of such strategies will affect ecosystem service delivery. Solutions considered within the Nature Guideline to address policyrelated questions come in two forms: investments and policyinstruments. Investments lead to improvements in the size or quality of natural capital, whereas policy-instruments set boundaries or restrictions to endorse the sustainable use of natural capital. This study focuses on how changes in the quantity and quality of natural capital (investment solutions) affect ecosystem service delivery. The study evaluates how changes in Amsterdam’s natural capital due to the implementation of the Structural Vision’s green objectives, affect ecosystem services. Various steps required to conduct a SCBA conform the Nature Guideline, are covered. These steps are described in Appendix 1, along with the location in this report (or otherwise) where different aspects from the guideline are addressed.

2.2

Scenarios The strategies envisioned by the Quality Impulse Green (Gemeente Amsterdam, 2017) were translated into four green infrastructure scenarios by ‘De Urbanisten’, an innovative consultancy firm for urban research, design, and landscape based in Rotterdam, The Netherlands. Scenarios illustrate the potential alterations to Amsterdam’s green infrastructure conform strategies from the Quality Impulse Green. A Business-As-Usual (BAU) scenario was also developed, illustrating ecosystem service values in the year 2025, considering only residential expansion and population growth as the main change. By comparing ecosystem service values resulting from the implementation of green Page 15 of 79

RIVM Letter report 2019-0021

infrastructure strategies, to ecosystem services values generated in a BAU situation, light is shed on the benefits (and costs) generated by green infrastructure. A brief description of each scenario is provided hereunder. 2.2.1

Business-As-Usual The ‘Business-As-Usual’ scenario projects the changes in natural capital that would occur in the City of Amsterdam by 2025 if the city continued to expand but green infrastructure strategies were not implemented. This scenario serves as a benchmark to compare the benefits (or losses) that would result from the implementation of different strategies from the Quality Impulse Green. The main changes considered consist of the expected growth of residential areas in Amsterdam, and the estimated population increase of approximately 70,000 people.

2.2.2

Green Neighbourhoods The ‘Green Neighbourhoods’ scenario (Figure 2.1) entails a general increment in the amount of green within areas where there is little green space, such as the city centre, and improving the quality of green in areas where green is relatively abundant, such as in peripheral neighbourhoods. The set of strategic actions designed to meet these objectives include: • In the historical city centre and along the urban canal, turning 100 % of all parking places into green surfaces, including trees; • Along the first city extensions and extensions made in the 40s, turning 50% of all parking places into green surfaces, including trees; • In the historic centre, turning grey roofs into green roofs, including 50% in the first extensions and extensions made in the 40s, and 15% along urban canals; • Implementing water sensitive design in areas that experience a high flood-risk, according to ‘Amsterdam Rainproof’ (www.rainproof.nl), including creating new green areas that implement water sensitive technologies, such as Bioswales, wadis, and raingardens.

Page 16 of 79

RIVM Letter report 2019-0021

Figure 2.1: Green Neighbourhoods base map, developed and mapped by De Urbanisten

Page 17 of 79

RIVM Letter report 2019-0021

Figure 2.2: Green Network base map, developed and mapped by De Urbanisten

Page 18 of 79

RIVM Letter report 2019-0021

Figure 2.3: Urban Parks base map, developed and mapped by De Urbanisten

Page 19 of 79

RIVM Letter report 2019-0021

Figure 2.4: Metropolitan Landscape base map, developed and mapped by De Urbanisten

Page 20 of 79

RIVM Letter report 2019-0021

2.2.3

Green Network The ‘Green Network’ scenario (Figure 2.2) entails investing in a recreational network of cycling, sports and hiking trails. This includes completing the main tree structure of Amsterdam’s network connecting green areas, thereby strengthening the ecological structure. The set of strategic actions designed to meet these objectives include: • Adding trees to the main tree structure to ensure its continuity; • Adapting the main tree structure to store rainwater in areas that experience a high flood-risk, according to Amsterdam Rainproof (www.rainproof.nl). • Altering the ecological structure by turning grass lawns into richer landscape typologies, such as meadows and different types of shrubs; • Formalizing Amsterdam’s Green Net (Groennet), or Amsterdam’s network of recreational areas, by improving the quality of various segments and adding new walking- and cycling-paths between areas that are not connected.

2.2.4

Urban Parks Strategic actions for the scenario ‘Urban Parks’ (Figure 2.3) were suggested by the Amsterdam Municipality. These actions entail creating new parks, expanding existing parks, and transforming parks of extensive use into intensive use parks. The set of strategic actions designed to meet these objectives include: • Creating new city parks by adding more grass, more shrubs and more trees to existing recreational areas; • Transforming green areas into city parks by (1) opening private green areas (mainly allotment gardens) into public areas, (2) making them accessible by expanding biking- and walking-paths (25km extra), and by (3) raising the maintenance rate from nomaintenance to C-level maintenance; • Expanding walking- and cycling-paths (198km extra) within extensive-use park to attract more people, reducing the pressure in intensive-use parks; • Improving city parks by improving the maintenance level

2.2.5

Metropolitan Landscape Metropolitan parks are large green and blue representative landscapes that surround the city. These areas are already rich in green and blue elements, which is why strategic aspects in the ‘Metropolitan Landscape’ scenario (Figure 2.4) aim to bring these elements closer to Amsterdam’s residents. The strategic actions designed to meet these objectives include expanding the walking- and cycling-paths linking metropolitan areas with the city, and designing programs and routes to access these areas.

2.3

Green Benefit Planner The Green Benefit Planner was used to quantify several ecosystem services within this report. The Green Benefit Planner is a spatial decision-support tool for calculating the effects of spatial changes on natural capital, to support spatial planning. Through the use of the Green Benefit Planner, it is possible to calculate the effect of different scenarios on ecosystem services and the accrued societal benefits. The Page 21 of 79

RIVM Letter report 2019-0021

tool applies models from the Natural Capital Model (NK-Model) (Remme et al., 2018), a collaborative modelling platform from RIVM, Wageningen Environmental Research (WEnR), and The Netherlands Environmental Assessment Agency (Planbureau voor de Leefomgeving, PBL). For detailed descriptions of the first published set of models from the NKModel, see Remme et al. (2018). As input, the Green Benefit Planner makes use of spatial plans developed by, for instance, urban planners or designers. These plans are georeferenced and translated into various input maps required to run the NK-Model. In this report, ecosystem service values for different green infrastructure scenarios were mapped by altering and then modelling input maps on land cover, inhabitants, and vegetation (low vegetation, shrubs, trees). For each scenario, the NK-Model was run, providing output in the form of maps and statistics (see Figure 2.5). The NK-Model can produce results for scenarios and for a BAU situation. The latter serves as a benchmark to illustrate fluctuations in ecosystem service delivery resulting from the implementation of different strategies. Since the Metropolitan Landscape scenario does not entail any alterations to vegetation or water compared to the BAU situation, models from the NK-Model were not applicable to this scenario.

Figure 2.5: Schematic representation of the Green Benefit Planner (source: www.atlasnatuurlijkkapitaal.nl). The physical and monetary value of relevant ecosystem services are estimated by a set of ecosystem service models and a standard set with input maps.

2.4

Ecosystem services and their benefits Table 2.1 presents the ecosystem services quantified in this study, the indicators adopted and their respective units of measurement, and the methods adopted to model them. Ten models were implemented to quantify 17 indicators that reflect seven ecosystem services and their potential impacts. The indicators quantified include air quality, health, physical activity, property value, recreation, urban cooling, and water storage. In some cases, a single model was adopted to quantify more than one indicator. Indicators falling under one model often reflect the same value in a different unit of measurement (e.g. particulate matter retention in kg/yr or in €/yr). Most of the ecosystem services were quantified by using the NK-Model (Remme et al., 2018). A set of additional models were developed for this study to project a diverse set of ecosystem service indicators (indicated as ‘See Appendix 2’ in Table 2.1). This contributes to generating a comprehensive picture of the ecosystem services that natural capital delivers in Amsterdam, and the Page 22 of 79

RIVM Letter report 2019-0021

potential trade-offs and synergies that exist among them. Additional calculations made are described in detail in Appendix 2. 2.4.1

Air quality Air pollution is a common problem and challenge within industrialized countries all over the world. It is caused by sources such as car traffic, industry, and livestock farming. One of the most common forms of pollution found in cities is particulate matter, which is associated with respiratory, cardiovascular diseases, and mortality. Vegetation, especially trees, can behave as a means of particulate matter retention, by affecting its deposition, the airflow, and turbidity. Particulate matter retention and the monetary benefits generated by the ecosystem service, were modelled through the use of the NK-Model (Remme et al., 2018). The model focuses on the retention by vegetation of particulate matter up to 10 micrograms (PM10). Despite the potential air quality regulation benefits that vegetation can generate, recent studies have shown that trees may also have the opposite effect when located along street canyons, streets with significant amounts of buildings, traffic, and human activity (Wesseling et al., 2011). Hence, the ecosystem service was not quantified for the scenario ‘Green Network’, where a significant amount of trees is added to the main tree structure along streets. While the ecosystem service was quantified for the remaining scenarios, the results are to be used with precaution since they constitute a high degree of uncertainty.

These indicators rather serve as a means to reify nature’s contributions to people in a holistic fashion, enabling the consideration and the comparison of the ecosystem service ‘air quality’ with other ecosystem services considered in this study.

Page 23 of 79

RIVM Letter report 2019-0021

Table 2.1: Ecosystem services modelled, indicators used, and methods applied

Ecosystem Service Air quality

Health

Physical activity

Property value Recreation Urban cooling Water storage

Model 1 2 3

4 5 6

7 8 9 10

Indicator

Unit

Method

PM10 retention PM10 retention Reduction in probability of being overweight Reduced number of visits to general practitioner Reduced health costs due to urban green Reduced health-related labour costs due to urban green Additional time spent on outdoor physical activity Cycling Time spent cycling to-from work Avoided premature deaths from cycling to-from work Avoided premature deaths from cycling to-from work Contribution to property value Visits to recreation areas Visitation expenditures Decrease in temperature Reduced rainwater in sewers Reduced water treatment costs

kg/yr €/yr % visits/yr €/yr

NK-Model NK-Model See Appendix 2 NK-Model NK-Model

€/yr

NK-Model

min/yr km/inhabitant/yr min/yr

See Appendix 2 See Appendix 2 See Appendix 2

lives/yr

See Appendix 2

€/yr

See Appendix 2

€ visits/yr €/yr °C m3/yr €/yr

NK-Model See Appendix 2 See Appendix 2 NK-Model See Appendix 2 See Appendix 2

Page 24 of 79

RIVM Letter report 2019-0021

2.4.2

Health A positive relationship exists between urban green spaces and human health. This report uses the NK-Model (Remme et al., 2018) and a Dutch study produced by RIVM (Klompmaker et al., 2017) to translate some of the various health benefits generated by green elements into four indicators describing health impacts and costs. Based on Klompmaker et al. (2017), the effect of the percentage of green space surrounding a household on the general probability of being overweight (BMI ≥ 25 kg/m2), was modelled. Through the use of the NK-Model, the effect of urban green on three human health indicators was modelled. The NKModel extrapolates the effect of urban green on general health by considering the effect of green space on the mitigation of nine diseases (KPMG, 2012; Remme et al., 2018). It also models the effect of urban green on the number of visits paid to general practitioners by patients (Maas, 2008). Finally, the NK-Model quantifies the effect of urban green on health-related labour costs. Health-related labour costs considered include three main factors: absenteeism, labour productivity, and job losses (Steenbeek et al., 2010).

The associations adopted to quantify this ecosystem service were obtained from studies conducted at the national scale, which is why the results obtained in this study are not meant to be used for purposes such as forecasting or analysing the actual cell-by-cell distribution of health benefits in Amsterdam.

2.4.3

Physical activity Evidence on the association between green space and physical activity (e.g. walking, cycling) is mixed (Staatsen et al., 2017). For instance, research conducted within the Netherlands found a positive relationship between the percentage of green space in a household’s vicinity and the amount of time its dwellers spend on outdoor physical activity, for instance cycling to-from work (Klompmaker et al., 2018; Maas et al., 2008). For this study, a model capturing the effect of green space on outdoor physical activity and cycling for commuting purposes, was developed (Appendix 2). The HEAT Tool (Kahlmeier et al., 2017) was applied to determine the effect of cycling on human lifespans (reduced risk of all-cause mortality) and the associated economic benefits (Appendix 2). The HEAT Tool, developed by the World Health Organization, measures the effects of walking and cycling on human health (Kahlmeier et al., 2017). The effect of cycling-network size on cycling is not an ecosystem service in theory, since cycling lanes do not constitute natural capital elements. However, strategies envisioned in the Quality Impulse Green include vast expansions in the cycling-lane network as a means to improve the quality and accessibility of green and blue areas. Hence, the inclusion of ‘kilometres cycled’ as a result of the size of the cycling network, was requested.

Since no readily available scientific research was found providing information on the relationship between the length of cycling lanes and cycling within the Netherlands, a basic extrapolation was made. The extrapolation is based on trends from reported cycling statistics for Amsterdam and projected increases in bike-lanes within every scenario (see Appendix 2). Page 25 of 79

RIVM Letter report 2019-0021

Uncertainties associated with physical activity indicators considered in this study result mainly from (1) the lack of quantitative evidence capturing the hypothesized relationship between explanatory variables, and (2) the failure to integrate socio-cultural elements that play a crucial role on people’s decision to engage in physical activity. For instance, indicators reflecting the relationship between green space and the time individuals engage in general outdoor physical activity or cycling for commuting purposes, only considers the percentage of green space and population growth as explanatory factors. Socio-cultural and other demographic aspects (e.g. age, socio-economic background, neighborhood safety level) were not considered, due to the lack of quantitative evidence reflecting these relationships. Further research should address these relationships in order to enable the integration of additional demographic and socio-cultural factors within the models adopted in this study. 2.4.4

Property value Natural elements, such as trees, gardens, parks, and water, have a positive influence on property values within urban areas (Czembrowski & Kronenberg, 2016; Franco & McDonald, 2017; Luttik & Zijlstra, 1997). To model the influence of green infrastructure on property prices in the city of Amsterdam, the NK-Model was used (Remme et al., 2018; Luttik & Zijlstra, 1997; Ruijgrok, 2006). The model captures the contribution to property prices by vegetation and water, based on their availability and accessibility. Uncertainties within this calculation result mainly from the date of the associations that were used as input for the model (Luttik & Zijlstra, 1997) and from common uncertainties relating to the use of benefit transfer techniques (i.e. transposing quantitative associations from a different site to the site under assessment). These uncertainties should be taken into consideration when adopting the results obtained from this model’s implementation to inform policy-making.

2.4.5

Recreation Aside from playing a crucial role on human survival, green spaces serve an important function as sites for recreation. Due to the lack of readily available tools and scientific research required to model the relationship between green spaces and recreation within the Netherlands, a basic extrapolation was made (see Appendix 2). The extrapolation is based on reported visits to a number of Amsterdam’s green recreational areas of different configurations (e.g. size, vegetation types, cycling/walking network size, access to water), and population growth in the region. Further research on the relationship between recreational area configurations and visitation rates/expenditures, is required to enable the development of a model that reflects these relationships with a lower degree of uncertainty.

2.4.6

Urban cooling Cities around the world increasingly experience higher temperatures than their surrounding rural areas, a phenomenon commonly referred to as the ‘urban heat island effect’. This often results from human activity and infrastructure that enhances heat storage while inhibiting its release. This includes the use of materials such as asphalt and concrete, Page 26 of 79

RIVM Letter report 2019-0021

which increase sunlight absorption while delaying its release, reduced wind speed between buildings, and reduced natural evaporation due to soil sealing (Remme et al., 2018). The effect of vegetation in mitigating the urban heat island effect was estimated through the implementation of the NK-Model (Remme et al., 2018; Lauwaet et al., 2018), which combines information on population, wind speed, vegetation, and soil sealing to model the ecosystem service. 2.4.7

Water storage Soils and vegetation play an influential role in water retention, a largely ignored but crucial function contributing to human well-being. Soils and vegetation behave as a sponge that absorbs water from rainfall, reducing the risk of flooding while also filtering unwanted pollutants from water. Water itself carries nutrients that are essential for plant growth, which in turn contributes to the delivery of other ecosystem services. Due to the absence of a water storage model within the NKModel, the supply of water storage by natural elements was calculated based on a simple model. The model makes use of a reference value (www.sbrcurnet.nl) to estimate the amount of precipitation (rainfall) that can be stored by vegetation in the City of Amsterdam. The reduced costs of sewage treatment are used as a proxy for the avoided costs associated with enhanced water storage. Appendix 2 provides an overview of the method adopted to extrapolate the potential water storage by vegetation and the associated monetary benefits. The main uncertainties from this calculation stem from common issues associated with benefit transfer techniques (i.e. transposing quantitative associations from a different site to the site under assessment).

As such, the results obtained from this model’s implementation can be adopted as a means to inform policy-makers of the benefits vegetation delivers for water retention rather than to analyse the cell-by-cell value of water storage across Amsterdam.

2.5

Input data The models from the NK-Model make use of a standardized set of input maps and tables (Remme et al., 2018). These maps also served as input for the additional models developed for this study. Input maps provide spatial information that, combined with formulas and look-up tables, can be incorporated into ecosystem service maps. Look-up tables provide information necessary for reclassifying the information from input maps. Generally, standard input maps and tables from the NK-Model were used in this study. However, some input maps were adapted to reflect predicted changes (i.e. residential expansion and population growth) based on information provided by the Municipality of Amsterdam, and changes based on the spatial strategies developed by De Urbanisten. The most important input maps adapted were the Land Cover Map (Ecosystem Unit map), Inhabitants Map, and vegetation maps (i.e. Trees Map; Shrubs Map; Low Vegetation Map). The following sub-sections provide information on the input maps that were used to model ecosystem services, and the alterations that were made to these maps to develop scenarios. Page 27 of 79

RIVM Letter report 2019-0021

2.5.1

Housing Plans Map Some of the main adaptations to input maps used in the NK-Model were based on data from the Housing Plans Map from the City of Amsterdam (maps.amsterdam.nl). This spatial dataset shows areas where new residential areas have been planned for the upcoming years, including the number of units that are expected to be built. The version of the dataset here used dates to 20-12-2017. Any subsequent alterations made to the dataset by the City of Amsterdam, have not been considered within this report. Additional information on the average number of inhabitants per housing was provided by the City of Amsterdam.

2.5.2

Inhabitants Map The Inhabitants Map used contains spatial information on the number of inhabitants living in the Municipality of Amsterdam during the year 2016. In 2025, it is expected that approximately 70,000 additional inhabitants will live within the municipality, compared to 2016. To extrapolate the additional number of inhabitants, the Housing Plans data from the Municipality of Amsterdam was used. Plans that fell within the phases ‘investment decision taken’ and ‘in construction’ with planned completion in the period 2018-2025, were included. Gross inhabitant values from the dataset were applied (number of housing units multiplied by the average number of inhabitants for a particular type of neighbourhood). This filtering provided a spatially disaggregated map of the new inhabitants, close to the expected number of additional inhabitants (final count 75,137). The inhabitants of a single building site were allocated evenly over the entire site. For instance, if a new building block of 100 10x10m cells houses 200 people, each cell would house 2 people. The map with the predicted (gross) number of inhabitants per cell was combined with the current Inhabitants Map by overlaying values in the original map with all non-zero values from the map with the predicted additional inhabitants. This resulted in an Inhabitants Map for Amsterdam for 2025.

The new Inhabitants Map was used as input for all scenarios, including the Business-As-Usual scenario.

2.5.3

Land Cover Map To develop a new Land Cover Map for each future scenario, the current Land Cover Map (Ecosystem Unit Map 2013, Statistics Netherlands), was adapted based on the maps for each scenario developed by De Urbanisten. Within every scenario, the current land cover was altered for certain areas into six land cover types: forest, grassland, shrubs, water, housing, and roads. All areas that were indicated to become ‘forest’ within a scenario, were altered to ‘deciduous forest’ in the Ecosystem Unit map, as this is the dominant forest type in the Amsterdam metropolitan region. All areas that were indicated to become ‘low vegetation,’ were altered to ‘natural grassland’ in the Ecosystem Unit Map. All areas indicated to become ‘shrubs’ were, altered to ‘other unpaved terrain’ for the Ecosystem Unit Map, as there is no separate class for shrubs in the Ecosystem Unit map. All additional water bodies and roads were altered to the ‘water’ and ‘road’ land cover classes, respectively. Finally, all areas where additional housing is expected (see 2.5.1) were altered to the ‘housing’ land cover type. Page 28 of 79

RIVM Letter report 2019-0021

2.5.4

Vegetation Maps Three vegetation maps were used: tree cover, shrub cover and low vegetation cover with scale 0 to 1, where 1 represents full cover. See Remme et al. (2018) for a detailed description of the maps. The vegetation cover maps were updated individually for each scenario and then revised to ensure that there were no inconsistencies once different vegetation cover maps were aggregated. Table 2.2 shows the changes in the number of units of various natural elements across different scenarios, with the Business-As-Usual scenario serving as a benchmark. Alterations made are described below. Table 2.2: Difference (+/-) in the number of green elements within green infrastructure scenarios

Green/blue element

Unit

Green Neighb.

Green Netw.

Metro. Lands.

Trees Shrubs Grass Total

units/ha units/ha units/ha units/ha

124 125 249

522 571 -653 440

-

Urba n parks 258 -79 138 316

Tree cover The tree cover map was updated first. All areas that were indicated as forest in a respective scenario, received a tree cover of 1. All areas that were altered to shrubs or low vegetation were set to 0 in the tree cover map. All areas that were altered to new housing areas (see 2.5.1) were set to 0.2, assuming that 20% of new residential areas would be covered by trees. In the Green Network scenario tree cover in the network structure was increased from 10% average cover to 40% average cover (ranging from 0 to 0.82 cover), based on a random distribution within the network. Low vegetation cover Second, the low vegetation cover map was updated. All areas indicated as low vegetation in a respective scenario, received a low vegetation cover of 1. All areas that were altered to shrubs or trees were set to 0 in the low vegetation cover map. All areas that were altered to new residential areas (see 2.5.1) were set to 0.2, assuming that 20% of new residential areas would be covered by low vegetation. Table 2.3: Additional low vegetation at street level and on green roofs in the Green Neighbourhoods scenario for different city zones City zone Additional street-level Additional low low vegetation vegetation on roofs % of total street area % of residential roof area City centre 7 5 Canal belt 5 15 First extensions 10 50 1940s extensions 10 50

In the Green Neighbourhoods scenario additional low vegetation has been added in Amsterdam’s central neighbourhoods, both at street level, and as green roofs (Table 2.3). Page 29 of 79

RIVM Letter report 2019-0021

Shrub cover Finally, the shrub cover map was updated. All areas indicated as shrubs in a respective scenario received a shrub cover of 1. All areas that were altered to low vegetation or trees were set to 0 in the shrub cover map. All areas that were altered to new housing areas (see 2.5.1) were set to 0, assuming that new housing areas would be partially covered only by trees and low vegetation. Total vegetation cover check Total vegetation cover cannot exceed 1 (100% cover). Therefore, after updating the tree separate Vegetation Maps for each scenario, a total vegetation cover map was made as a check, by adding up the tree cover, shrub cover and low Vegetation Maps. 2.5.5

Further input data All other input files (based on the current situation) were left equal in the current situation and all scenarios, even if future changes are likely. Examples of datasets that are likely to change in the future are housing prices, concentrations of particulate matter (PM10) and potentially average wind speed. As the future developments of these variables are unknown, they have been left unchanged, also making the effects of natural capital more visible in the results.

Page 30 of 79

RIVM Letter report 2019-0021

3

Results

This Section presents the predicted ecosystem service values for the City of Amsterdam, for the year 2025. Results for one Business-As-Usual (BAU) and four green infrastructure scenarios, are presented. For every scenario, seven ecosystem services were modelled, expressed in 17 indicators. Indicators describe physical impacts (e.g. water storage, temperature), well-being impacts (e.g. improved health), and the associated monetary benefits. The BAU scenario serves as a benchmark for the value that natural capital would generate in Amsterdam in 2025 if the green infrastructure strategies from the Quality Impulse Green were not implemented. The four green infrastructure scenarios reveal ecosystem service values after alterations in Amsterdam’s green infrastructure take place, due to the implementation of strategies form the Quality Impulse Green. Section 3.1 presents the ecosystem service quantities that would be produced in the year 2025 if Amsterdam’s Structural Vision were not implemented, and Section 3.2 expands on the fluctuations in ecosystem service delivery that would take place provided the implementation of each scenario. Section 3.3 provides a general overview on ecosystem service values within each scenario, also generating insights on ecosystem service trade-offs and interactions. 3.1

Business-As-Usual This Section presents the results showing the total values for the BAU scenario delivered by every ecosystem service. Here, ecosystem service values for the year 2025 are presented. Not all indicators modelled are presented as maps. This is the case since not all models here used are spatial models (i.e. recreation; cycling) and since some of the indicators modelled present similar information within different units of measurement (e.g. the value of PM10 retention in kg/yr and in €/yr). Despite no improvements in Amsterdam’s green infrastructure being considered for the period ranging from 2016-2025, demographic changes (i.e. expected growth of residential areas and population increase) will often lead to fluctuations in ecosystem service values throughout the period (e.g. through more beneficiaries).

3.1.1

Air quality The ecosystem service ‘air quality regulation’ was measured by observing the amount of particulate matter (PM10) retained by vegetation and water in kg/year, and the accrued economic benefits in €/year. The monetary benefits from increase particulate matter retention were calculated based on the assumption that reduced PM10 concentrations in the air lead to a lower incidence of air-pollution related diseases and related costs (CE Delft, 2017). It was estimated that green infrastructure captures more than 340 thousand kg of particulate matter per year in Amsteredam 1, valued at €15 million in avoided health costs from air pollution related diseases. In 2025, the city’s population is 1

This calculation does not consider the inverse effect on PM10 retention that trees have in street canyons

Page 31 of 79

RIVM Letter report 2019-0021

predicted to increase sharply. The larger number of benefactors implies an even higher value generated from reduced health costs. Figure 3.1 presents the map for PM10 retention by vegetation within the BAU 2025 scenario in kg/year, showing the areas where the highest (blue) and lowest amounts (yellow) of PM10 are retained. The capture of PM10 relies on two main factors: the land cover type (i.e. trees, shrubs, grass, water) and the total concentration of PM10 within a specific cell. Trees have the highest capacity for PM10 retention, followed by shrubs, and grass (in descending order). Hence, areas with a high percentage of trees (e.g. Amsterdamse Bos) tend to show high values of PM10 retention. Moreover, the need for PM10 retention is dependent on PM10 concentrations within an area. Densely populated areas with a high degree of human activity tend to have higher particulate matter concentrations and thereby require a higher PM10 uptake. Areas in the northernmost part of Amsterdam (e.g. Waterland) experience lower PM10 uptake compared to green areas in populated areas, which can be seen in dark blue hues. This is the case since the vegetation in the outskirts of the city consist mainly of grass, which retains less PM10 than trees or shrubs, and also since the demand for PM10 retention in such areas is lower due to lower overall concentrations in less populated areas.

Built-up and other paved areas with little to no vegetation, show lower amounts of PM10 retention, visible in yellow hues. 3.1.2

Health The percentage of green space in the near proximity of households is positively associated with improved health. For instance, a relationship exists between exposure to urban green and the number of visits to a general practitioner (GP). It is expected that the number of avoided visits to GPs in Amsterdam will increase from above 31,000 visits in 2016 to above 32,000 visits in 2025. The reduced costs from improved health related to green space are expected to increase from roughly €30.6 billion to €32.4 billion throughout the same period. This projection considers reduced costs based on reductions in the incidence of nine diseases related to urban green (Remme et al., 2018). The reduced labour costs associated with improved health were estimated by calculating the reduced costs of absenteeism, reduced labour productivity, and job losses (KPMG, 2012; Steenbeck et al., 2010). Reduced labour costs are predicted to increase from roughly €150 million in 2016 to €158 million in 2025.

Page 32 of 79

RIVM Letter report 2019-0021

Figure 3.1: Air quality regulation – Business-As-Usual – year: 2025

Page 33 of 79

RIVM Letter report 2019-0021

Figure 3.2: Health – Business-As-Usual – 2025

Page 34 of 79

RIVM Letter report 2019-0021

Figure 3.2 shows the distribution of reduced health costs due to urban green, based on the influence of green on the incidence of nine diseases (KPMG, 2012; RIVM, 2003; Maas, 2008). One of the main determining factors in this calculation is the population density within a 1km radius (buffer) around a cell, which is why dark blue areas cover most inhabited areas in a fairly homogeneous fashion. The percentage of green space within a 1km buffer of every household also plays a determining role in the extrapolation of this ecosystem service’s value, contributing to the fairly homogeneous hues of green, blue, and yellow here seen. Areas in the outskirts of Amsterdam are visible in yellow and white hues primarily, which is mainly attributed to the low population densities in such areas and thereby lower demand for the ecosystem service. 3.1.3

Physical activity Within the Netherlands, a positive relationship exists between green space and the amount of time individuals spend on physical activity (Klompmaker et al., 2017; Maas et al., 2008). Green space has a positive influence on the amount of time people are willing to spend on outdoor physical activity in general, and more specifically cycling tofrom work. In Amsterdam, it is estimated that the contribution of green space to the time people decide to spend on outdoor physical activity will increase by 2.1% throughout the period 2016-2025, while the total amount of time people spend cycling for commuting purposes will increase by 0.1%. Figure 3.3 projects the distribution of the total amount of time people will spend cycling to-from work in 2025, based on the amount of urban green in their surroundings. The projected total amount of time cycled for commuting purposes is dependent both on the amount of green space in an area and on the number on inhabitants within each household. Figure 3.3 shows lower cycling values in the centre of Amsterdam (i.e. de Wallen) compare to the surrounding rings. This is the case since the ratio of households to green in the centre of Amsterdam is relatively low compared to its surrounding neighbourhoods.