2. RapportRIVMWaardevGroen.pdf

Within the outskirts of Amsterdam, the percentage of green increases but population also decreases, so the total amount of cycling for commuting purposes become less visible. Research has shown that cyclists have a lower risk of all-cause mortality compared to non-cyclists, even when considering the risks associated with active transport (e.g. exposure to air pollution or traffic injuries) (Kelly et al., 2014; Staatsen et al., 2017). It was estimated that the number of avoided premature deaths per year associated with cycling for commuting purposes will increase from approximately 192 in 2016 to 193 in 2025. The economic benefit due to increased lifespans was estimated at approximately €555 million for 2016 and €558 million for 2025. These economic benefits were obtained by translating the avoided premature deaths from increased cycling to the value of a statistical life, as suggested by the WHO HEAT Tool (https://www.heatwalkingcycling.org/). Default values for the WHO European region were used, which were set at €2.132 million per individual for 2015 (www.who.int/). Page 35 of 79

RIVM Letter report 2019-0021

Figure 3.3: Physical activity – Business-As-Usual – 2025

Page 36 of 79

RIVM Letter report 2019-0021

3.1.4

Property value In the Netherlands, green infrastructure has a positive influence on property values. The effect of green infrastructure elements (e.g. trees, water, parks) on property values for the City of Amsterdam is expected to increase from €11.5 billion in 2016 and to €11.8 billion in 2025. Figure 3.4 shows the economic contribution of vegetation and water to property values, calculated as a fraction of the total property value. As such, areas which are generally defined by relatively higher prices (e.g. Amsterdam centre and old south) often experience higher net contributions to property value.

3.1.5

Recreation For the ecosystem service ‘recreation’, a non-spatial model was developed (Appendix 2). Based on statistics on visitations to recreational areas in Amsterdam, it is estimated that more than 174 million visits are made to recreational areas in Amsterdam (year=2016). This number is expected to increase to more than 210 million visits in 2025, with associated increased recreational expenditures valued at approximately €200,000 per year.

3.1.6

Urban cooling The effect of vegetation on urban cooling was extrapolated from the difference between the maximum potential urban heat island (UHI) effect and the actual urban heat island effect for a given area. The maximum potential UHI effect represents the total demand for urban cooling and the actual UHI effect as the unmet demand for urban cooling. The resulting value (maximum UHI effect - actual UHI effect) is thus the supply of urban cooling by green and blue. The maximum potential UHI effect is dependent on three main variables: soil sealing (including built-up areas), population density, and wind speed. Different natural (non-sealed) areas will lead to different percentage reductions in the UHI effect. In Amsterdam, it is expected that green infrastructure will increase its contribution to the reduction of the UHI effect by 1%. This estimate assumes an increase in the UHI effect due to an increase in population.

Page 37 of 79

RIVM Letter report 2019-0021

Figure 3.4: Property value – Business-As-Usual – 2025

Page 38 of 79

RIVM Letter report 2019-0021

Figure 3.5: Urban cooling - Business-As-Usual – 2025

Page 39 of 79

RIVM Letter report 2019-0021

Figure 3.5 illustrates the reduction of the UHI effect by green and blue, or the supply of urban cooling. Areas with low supply of the ecosystem service (yellow) consist of areas with very high population densities and vast sealing. Areas in blue are mainly characterized by larger extents of semi-natural and agricultural land, with lower population densities and less sealing, thus requiring a lower demand for urban cooling. 3.1.7

Water storage Vegetated areas play an influential role in storing water from rainfall, avoiding runoff into sewers and reducing the risk of flooding. In Amsterdam, water storage by vegetation leads to around 31 million m3 of reduced rainwater in sewers per year. The value accrued to reduced water treatment costs form increased storage is estimated at more than €24 million per year. Figure 3.6 illustrates the distribution of water retention in the city of Amsterdam, in m3 per year. Here, water storage is directly related to the amount of vegetation in inhabited areas, with highly vegetated areas visible in dark blue hues. The amount of water stored can be viewed as potential runoff stored by vegetation that would otherwise end up on streets and sewers.

Page 40 of 79

RIVM Letter report 2019-0021

Figure 3.6: Water storage - Business-As-Usual – 2025

Page 41 of 79

RIVM Letter report 2019-0021

3.2

Changes in ecosystem service values across scenarios To assess the value generated (or lost) through the implementation of strategies from Amsterdam’s Quality Impulse Green, the difference between ecosystem service values quantities obtained for each green infrastructure scenario were compared to those from the business-asusual (BAU) scenario. Within Table 3.1, differences for seven ecosystem services are expressed in 17 indicators. The values presented consist of the difference between the total value for each indicator within the BAU scenario and the value for each indicator within the four green infrastructure scenarios. Since the Metropolitan Landscape scenario does not entail any alterations to vegetation or water compared to the BAU situation, several ecosystem service indicators do not experience any change. Table 3.2 presents the changes in ecosystem service values accrued to changes in green (green ha). This provides an overview of the average changes in ecosystem services resulting from the implementation of every strategy. The monetary benefits vary, between 1 and 5 €/m2 additional green per year. Aggregation of the different ecosystem services is strongly discouraged as there are overlaps within the elements covered by different indicators, giving way to doublecounting.

The strategies considered within each scenario are not meant to be viewed as alternatives to one another but as different elements of the Quality Impulse Green.

3.2.1

Air quality Two indicators were adopted to reflect fluctuations in particulate matter (PM10) retention and the associated benefits: the reduction of PM10 in kg per year and the reduction in health costs from the reduced incidence of related diseases. The model applied in this study was not applicable for the Green Network scenario, since a significant number of trees were added along streets for this scenario. The large number of trees within street canyons (i.e. narrow streets with significant infrastructure and human activity) often lead to dust getting trapped, which in turn leads to more PM10 in the air and has the opposite effect of the desired. Within the NK-Model, trees have the highest capacity for PM10 retention, which would lead to an overrepresentation of the ecosystem service within the Green Network scenario. While PM10 retention and the associated economic benefits do not differ largely for the Green Neighbourhoods and Urban Parks scenarios, the distribution of the ecosystem service significantly differs. The distribution of the changes in air quality regulation for these scenarios is presented in Figure 3.7, expressed in kg/yr.

For the Green Neighbourhoods scenario, changes in vegetation are distributed all around the core of the city and surrounding rings in a somewhat homogeneous fashion, while in the Urban Parks scenario the main changes include alterations in vegetation within existing parks and expansions of existing green areas. This is clearly visible in Figure 3.7, where increases in PM10 retention are observed mainly in these areas. In the Urban Parks scenario, PM10 retention decreases in some areas (in red). These are areas where trees were replaced by shrubs or grass, which have a lower capacity for PM10 retention.

Page 42 of 79

RIVM Letter report 2019-0021

Ecosystem Service Air quality

Table 3.1: Total changes in ecosystem service values compared to the BAU scenario (scenario-BAU2025). Aggregation of the different ecosystem services is strongly discouraged as there are overlaps within the elements covered by different indicators, giving way to double-counting. The strategies considered within each scenario are not meant to be viewed as alternatives to one another but as different elements of the Quality Impulse Green.

Health

Physical activity

Property value Recreation Urban cooling

Water storage

Green Neighbourhoods 2.4 0.13 0 3 2.4

Green Network 1.0 2 1.9

Metropolitan Landscape 0 0 2.0 0 0

Urban parks 2.7 0.15 3.0 1 1.0

million €/yr

11.8

9.0

0

5.0

million min/yr million km/yr km/person/yr million min/yr

0.3 143 137 0.5

0.3 49 47 0.4

214 205 -

0.2 199 190 0.1

lives/yr

4

3

-

1

million €/yr

12

9

-

3

million € million visits/yr million €/yr degrees C thousand ha thousand ha thousand ha million m3/yr million €/yr

51 0 0 0.0 9 9 0 1.2 0.9

41 0 0 0.0 7 8 6 1.1 1.1

0 10 49 0 0 0 0 0 0

12 19 88 0.0 1 4 3 0.8 0.6

Indicator

Unit

PM10 retention PM10 retention Reduction in probability of being overweight Reduced number of visits to GP Reduced health costs due to urban green Reduced health-related labor costs due to urban green Additional time spent on outdoor physical activity Cycling 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 Area with >1.5 degrees C cooling by green Area with >2 degrees C cooling by green Area with >2.5 degrees C cooling by green Reduced rainwater in sewers Reduced water treatment costs

thousand kg/yr million €/yr % thousand visits/yr million €/yr

Page 43 of 79

RIVM Letter report 2019-0021

Table 3.2: Changes in ecosystem service values compared to the BAU scenario (scenario-BAU2025). Changes are calculated per ha of green added. Changes in ecosystem service values not associated with changes in green space are not considered in this table. Aggregation of the different ecosystem services is strongly discouraged as there are overlaps within the elements covered by different indicators, giving way to double-counting. The strategies considered within each scenario are not meant to be viewed as alternatives to one another but as different elements of the Quality Impulse Green.

Ecosystem Service Air quality Health

Physical activity

Property value Recreation Water storage

kg/green ha/yr thousand €/green ha/yr visits/green ha/yr thousand €/green ha/yr

Green Neighborhoods 9.6 0.54 11 10

Green Network 5 4

Metropolitan Landscape -

Urban parks 8.9 0.47 4 3

thousand €/green ha/yr

47

21

-

16

thousand min/green ha/yr thousand min/green ha/yr

1.1 2.2

0.8 0.9

-

0.6 0.5

lives/green ha/yr

0.016

0.007

-

0.003

thousand €/green ha/yr

46

20

-

9

thousand €/green ha thousand visits/green ha/yr thousand €/green ha/yr thousand m3/green ha/yr thousand €/green ha/yr

202 4.8 3.8

95 2.4 2.6

-

40 2.5 2.0

Indicator

Unit

PM10 retention PM10 retention Reduced number of visits to GP Reduced health costs due to urban green Reduced health-related labor costs due to urban green Time spent on outdoor physical activity Time spent cycling to-from work Avoided premature deaths from cycling tofrom work Avoided premature deaths from cycling tofrom work Contribution to property value Visits to recreation areas Visitation expenditures Reduced rainwater in sewers Reduced water treatment costs

Page 44 of 79

RIVM Letter report 2019-0021

Figure 3.7: Difference between the value for ‘air regulation’ in the Business-AsUsual scenario compared to the green infrastructure scenarios

3.2.2

Health The health benefits generated by green infrastructure within every scenario are expressed as changes in four indicators, as visible in Table 3.2. First, the change in the probability for being overweight, depending on the percentage of green space surrounding a household, was modelled. Here, the Green Network scenario witnesses a higher overall benefit (a reduction of 14.3%) as compared to the Green Neighbourhoods and Urban Parks scenarios. This is possibly due to the modelling approach, which considers the percentage of green space within a 3km buffer as input (see Appendix 2). As such, the overall increase in green has a higher effect than the distribution of green. The Green Network scenario has the highest increase in vegetation, explaining the higher reduction in the probability of being overweight.

Page 45 of 79

RIVM Letter report 2019-0021

To measure changes in the contribution of green space to general health, health-related labour costs, and avoided visits to GPs, the NKModel was implemented. Within the Model, the percentage of green within a buffer of 1km is considered as the main factor affecting urban health. As such, the distribution and distance to green space surrounding a household play an important role in the delivery of these benefits. Increases in green space in the Green Neighbourhoods scenario are homogeneously distributed throughout city’s centre and surrounding rings, leading to the highest increase in health benefits compared to other scenarios. Moreover, the changes in green infrastructure within this scenario take place within Amsterdam’s most densely populated areas, leading to more beneficiaries and a higher increase in the value generated. The distributions of the benefits accrued through the implementation of strategies in each green infrastructure scenario, estimated using the NK-Model, are illustrated in Figure 3.8. 3.2.3

Physical activity For the ecosystem service ‘physical activity’, mixed results were found. Fluctuations in the additional time spent on outdoor physical activity are mainly dependent on changes in the percentage of green space within a 1km buffer and changes in population density. Increases in the percentage of green space in an area lead to a more than proportional increase in outdoor physical activity. As such, the total amount of green is as important as its distribution within a buffer. For this indicator, the highest benefit is projected to occur within the Green Network scenario. When measuring the influence of green space on cycling to-from work, as well as its associated benefits (i.e. avoided premature deaths and reduced health costs), the Green Neighbourhoods scenario seems to accrue the highest benefit. This is likely the case since green space is introduced homogeneously across the city, including areas where there is currently little to no green and are thus areas where high amount of benefits can be accrued. The changes in the time spent cycling for commuting purposes resulting from increases in green space, are visible in Figure 3.9.

Page 46 of 79

RIVM Letter report 2019-0021

Figure 3.8: Difference between the value for ‘health’ in the Business-As-Usual scenario compared to the green infrastructure scenarios

Page 47 of 79

RIVM Letter report 2019-0021

Figure 3.9: Difference between the value for ‘physical activity’ in the Business-As-Usual scenario compared to the green infrastructure scenarios

Page 48 of 79

RIVM Letter report 2019-0021

Although the association between the lengths of cycling lanes and kilometres cycled does not reflect benefits generated by green infrastructure per se, its inclusion was requested since cycling-lane extensions reflect an improvement in the quality and accessibility of green and blue areas. To calculate fluctuations in total cycling due to the extension or creation of bike lanes, a simple estimate was made, whereby the increase in cycling is proportional to the increase in the length of bike lanes. Since the Metropolitan Landscape scenario includes a significant expansion in the length of bike lanes within the city, this scenario also witnesses the highest increase in cycling. 3.2.4

Property value Figure 3.10 shows difference maps for property values for every scenario. These maps reveal very few areas where a noticeable change in property value occurs. Factors influencing the value of property include the proximity of a residence to parks, and its visibility of trees and water. The main change within scenarios that would thus affect residence property values includes their access to trees, as all other factors remain constant. Within the Green Network and Green Neighbourhoods scenarios, trees are added within areas where a few or no trees were found, significantly raising the value of property. As such, these two scenarios show the highest relative increases in property value. Within the Urban Parks scenario, trees are either added to existing parks, having a relatively low or neutral effect on property value, or within new parks. Additionally, the creation of new parks can generate significant contributions to property value, however, for a confined number of residences. As such, the increase in property value is relatively lower within the Urban Parks scenario as compare to the Green Network and Green Neighbourhoods scenarios.

3.2.5

Recreation In Amsterdam, the number of visits to recreational areas is expected to increase sharply throughout the upcoming years, due to the high demand from a growing population. The strategies in Amsterdam’s Quality Impulse Green aim to enhance not just the quantity but also the quality of green spaces throughout the city. The scenarios Green Neighbourhoods and Green Network do not comprise strategies for addressing the management of recreational areas. As such, the visitation rate is not expected to change, compared to the Business-AsUsual situation. Within the Metropolitan Landscape and Urban Parks scenarios, the maintenance and management of a number of recreational sites is enhanced. This in turn leads to certain recreational areas shifting from extensive to intensive use, increasing their attractiveness for recreational visits. In 2025, enhanced management of recreational areas is expected to accommodate 10 million additional visitors provided the implementation of the Metropolitan Landscape scenario, and 19 million provided the implementation of the Urban Parks scenario. The distribution of new visits differs from new visits in urban parks in the Urban Parks scenario, to visits in metropolitan areas in the Metropolitan Landscape scenario.

Page 49 of 79

RIVM Letter report 2019-0021

Figure 3.10: Difference between the value for ‘property’ in the Business-As-Usual scenario compared to the green infrastructure scenarios

Page 50 of 79

RIVM Letter report 2019-0021

3.2.6

Urban cooling Just as with changes in the delivery of other ecosystem services quantified in this study, the distribution of vegetation plays a significantly influential role on the distribution of urban cooling across the city. As such, areas where new vegetation has been added experience the highest reductions in temperature, as can be seen in Figure 3.11. Similar to the ecosystem service ‘air quality regulation’, trees tend to have a higher capacity to reduce temperature, followed by shrubs and grass at last. As such, areas where a high number of trees is added will see the highest reductions in temperature.

3.2.7

Water storage The amount of water stored by green spaces is correlated with the amount of vegetation in an area. As such, areas where more vegetation is added tend to experience the highest increases in water stored. This is clearly visible in Figure 3.12, where increases in water storage are visible particularly in areas where green is added. Within the Green Neighbourhoods scenario, the increase is sparsely distributed throughout Amsterdam’s central areas, while in the Green Network scenario, increases in water storage are mainly seen along areas where the city’s ecological network is expanded. In the Urban Parks scenario, increases are mainly seen in recreational areas, where green spaces were extended or made greener.

Page 51 of 79

RIVM Letter report 2019-0021

Figure 3.11: Difference between the value for ‘urban cooling’ in the Business-As-Usual scenario compared to the green infrastructure scenarios

Page 52 of 79

RIVM Letter report 2019-0021

Figure 3.12: Difference between the value for ‘water storage’ in the Business-As-Usual scenario compared to the green infrastructure scenarios

Page 53 of 79

RIVM Letter report 2019-0021

3.3

Associations between ecosystem services Ecosystem services are often analysed independently from one another within ecosystem service assessments. In reality, ecosystem services do not behave as individual elements in a vacuum but as interconnected parts of a network. They coexist across space and time as bundles of services that are associated with one another, often in a non-linear fashion (Bennett et al., 2009). It is crucial to understand these associations when analysing the implications of ecosystem change (Hattam et al., 2015). In essence, an increase the availability an ecosystem service can be associated with an increase or even a decrease, in the availability of another ecosystem service (Lester et al., 2013; Mouchet et al., 2014). One way to illustrate these trade-offs is by creating amoeba diagrams. Figure 3.13 presents amoeba diagrams depicting relative values for ecosystem services across the four green infrastructure scenarios. Relative values were calculated by dividing the total value for an indicator within a particular scenario against the average value for the same indicator for all four scenarios. Indicators in the amoeba diagrams showing values higher (lower) than 1 suggest that the total quantity measured for that indicator within that scenario is higher (lower) than the average of the same indicator for all scenarios.

In other words, the value of the indicator for a particular scenario is higher (lower) than average. For the Green Network scenario, the indicator for the ecosystem service ‘air quality regulation’ is excluded since, as previously mentioned, the model implemented does not consider the negative effect of street canyons on air quality. Expected values for particular indicators across scenarios do not vary significantly (value≈1) with a few exceptions. Within the Metropolitan Landscape scenario, indicator values are generally lower than average since the scenario comprises no changes in land cover so the values for most indicators are equivalent to the values obtained for the BAU situation. Other scenarios generally witness an increase in ecosystem service delivery compared to the BAU situation, due to additions in vegetation. Moreover, the cycling indicator for physical activity is correlated with the length of cycling lanes, which is why the indicator is significantly higher where bike lanes experience significant expansions (i.e. Metropolitan Landscape and Urban Parks) and lower where bikelane expansion is relatively low (i.e. Green Neighbourhoods and Green Network). Changes in the recreation indicator depend on whether increased management measures are taken in recreational areas or not.

This explains why scenarios where enhanced management of recreational areas takes place (i.e. Metropolitan Landscape and Urban Parks) see a significantly relatively higher value for recreation, compared to scenarios where no enhanced management takes place (i.e. Green Neighbourhoods and Green Networks). Since the ‘cycling tofrom work’ measure for physical activity has a positive relationship to the amount of green, all scenarios see an increase in the indicator, except for the Metropolitan Landscape scenario, where green remains constant. In terms of health benefits, the Green Neighbourhoods and Green Network scenarios seem to yield the highest relative benefits.

Page 54 of 79

RIVM Letter report 2019-0021

Green Network

Green Neighbourhoods

1

1

Metropolitan Landscape

1

Urban Parks

1

Figure 3.13: Relative values of ecosystem services across different scenarios (total value of scenario/average of four green infrastructure scenarios). For every ecosystem service, one indicator is presented per model used (in cases where one model leads to more than one indicator). Values higher (lower) than 1 imply that the value for that ecosystem service indicator for a scenario is higher (lower) than the average for all scenarios. For illustrative purposes, the origin was set at 0.9, accentuating relative changes in indicators across scenarios. Page 55 of 79

RIVM Letter report 2019-0021

Amoebas can also be generated by comparing total values for each scenario against BAU values. Values can be shown as the ratio or the percentage difference between the scenario and the BAU situation. This approach was not adopted to develop the amoeba diagrams since the Metropolitan Landscape scenario experiences significantly large increases for the recreation and one physical activity (i.e. cycling) indicators. As such, relative changes in other indicators, which were relatively smaller, would be difficult to distinguish, defeating the purpose of the illustration. Table 3.3 provides an overview of the relative values of ecosystem service indicators compared to BAU values, as ratios. By analysing the Table, it becomes clear that the Green Neighbourhoods and Green Network scenarios generate the highest relative increases in ecosystem service values for most indicators considered, followed by the Urban Parks and, finally, the Metropolitan Landscape scenarios. It is worth repeating that scenarios are not meant to be viewed as alternatives but rather as different aspects from Amsterdam’s Quality Impulse Green. Hence, the values here presented serve as a means to provide information on how different strategies from the Quality Impulse Green could influence ecosystem services amounts. Table 3.3: Relative values of ecosystem service against the BAU situation.

every ecosystem service, one indicator is presented per model used (in cases where one model leads to more than one indicator). For each indicator, the total value for each scenario is divided by the average of all four scenarios. Values higher (lower) than 1 imply that the value for that ecosystem service indicator for that scenario is higher (lower) than the value within the BAU scenario.

Ecosystem Service Air quality Health

Physical activity

Property value Recreation Urban cooling Water storage

Indicator PM10 retention Reduction in probability of being overweight Reduced health costs due to urban green Additional time spent on outdoor physical activity Cycling Time spent cycling tofrom work Contribution to property value Visits to recreation areas Decrease in temperature Reduced rainwater in sewers

Page 56 of 79

Green Neigh. 1.01

Green Netw.

Metro. Lands. 1.00

Urban parks 1.01

1.00

1.14

1.00

1.14

1.07

1.06

1.00

1.03

1.12

1.15

1.00

1.08

1.14

1.05

1.21

1.20

1.02

1.01

1.00

1.01

1.00

1.00

1.00

1.00

1.00 1.01

1.00 1.01

1.05 1.00

1.09 1.00

1.04

1.03

1.00

1.02

RIVM Letter report 2019-0021

4

Discussion and conclusions

An innovative spatial strategy is required to accommodate the needs of Amsterdam’s growing population, if it wishes to maintain a robust economy while fostering a good quality of life and tackling the critical environmental challenges of our time. The Structural Vision 2040 envisions the solutions that are required to tackle these challenges. The strategies designed to achieve these solutions were envisioned in Amsterdam’s Green Agenda 2015-2018 (Gemeente Amsterdam, 2015) and will be further formulated within the Quality Impulse Green (Gemeente Amsterdam, 2017). To understand the consequences of the implementation of these strategies, a quick scan was requested, evaluating the potential effect of their implementation on the health and well-being of Amsterdam’s citizens. This study explored how the implementation of different green infrastructure measures from the Quality Impulse Green could impact Amsterdam’s natural capital and the ecosystem services it generates. Urban design firm De Urbanisten designed four scenarios capturing strategies from the Quality Impulse Green. Scenarios, expressed as maps, illustrate the potential distribution of new green infrastructure measures in Amsterdam by the year 2025. Input maps on one BusinessAs-Usual (BAU) (i.e.

only population growth and residential expansions considered) and the four scenarios considering changes in green infrastructure, served as input for the quantification and mapping of ecosystem services in the year 2025. Indicators adopted as proxies for ecosystem services consider biophysical, economic, and social values. Described hereunder are this study’s main findings, limitations, and recommendations for future research. The Section discusses the main factors identified as influential for ecosystem service delivery across scenarios. It puts into perspective the limitations and recommendations necessary to make adequate use of this study’s results. Finally, the benefits (and costs) that strategies from the Quality Impulse Green could generate for the city of Amsterdam, are described. 4.1

Supply and demand In most cases, the main factor influencing ecosystem service supply (production) was the vegetation cover, including the amount, type, and distribution of vegetation. In terms of ecosystem service demand (use), the main influential factor was population density, which determines both ecosystem service usage and environmental pressures.

4.1.1

Supply Vegetation characteristics played a deterministic role in the delivery of ecosystem services within this study. Results reveal a general increase in ecosystem service delivery as a result of increases in the amount of green. Vegetation cover was classified intro three main types: trees, shrubs, and grass. Within various models, trees have a higher positive effect on the delivery of ecosystem services (e.g. air quality regulation; urban cooling) than shrubs and grass, at last. Aside from the amount Page 57 of 79

RIVM Letter report 2019-0021

and type of vegetation, its distribution plays a key role. Areas where the relative increase in vegetation is high (e.g. from none or low, to high amounts of green), often experience sharper increases in ecosystem service delivery. This might explain why the Green Neighbourhoods and Green Network scenarios score highest for most ecosystem service indicators, compared to other scenarios. These two scenarios are characterized by not just a higher amount, but also a more homogenous distribution of green, compared to the Urban Parks and Metropolitan Landscape scenarios. This includes adding green in areas where green is currently rather scarce, such as the central and old south areas. 4.1.2

Demand The ecosystem services concept is anthropocentric in nature, since the demand for an ecosystem service only exists if humans can benefit from it. This is why, in most cases, the highest demand for ecosystem services was found within densely populated areas. However, this study also makes reference to ecosystem service bundles, where ecosystem services act as interdependent elements of a system that influence one another. As such, all natural systems provide, to some extent, benefits to humans and their well-being. Demand for ecosystem services also exists when environmental pressures detrimental to human beings need to be mitigated. To a large extent, environmental pressures are determined by population density (e.g. the urban heat island effect; air pollution), another reason why densely populated areas often experience higher ecosystem service demand.

4.2

Limitations

4.2.1

Qualitative assessment A number of relevant ecosystem services were not included within this study. This resulted from restrictions in models, data, and time required to quantify a broader set of ecosystem service indicators. Additional models and indicators are required to quantify ecosystem services not considered in this study, such as biodiversity (often not considered an ecosystem service), and to monetize all ecosystem services not reflected in monetary units. The quantification of ecosystem services that are qualitative in nature (e.g. the spiritual or intrinsic value of nature) is subject to controversy, since defining boundaries and concepts proves a highly subjective task. This often leads to a lack of evidence on the relationship between natural elements and the qualitative values people assign to them. Socio-cultural ecosystem services can be quantified through the use of proxy indicators and through participatory processes. Participatory processes are time and resource intensive. This is why, in this study, only one socio-cultural ecosystem service (i.e. recreation) was quantified, using both physical and monetary proxy-indicators.

4.2.2

Aggregation Aggregating ecosystem service indicators is strongly discouraged. A common way of aggregating ecosystem services is by normalizing quantified indicators and subsequently obtaining their sum or average. However, this practice assumes that all ecosystem services have the same weight, which is unlikely to be the case. One way to solve for this is by assigning different weights to ecosystem services, based on stakeholder preferences obtained through participatory exercises, or Page 58 of 79

RIVM Letter report 2019-0021

based on expert judgment. However, this reduces the objectivity of the results obtained, resulting in a lower degree of accuracy. Moreover, each ecosystem service can be measured in a variety of metrics, quantified through the use of more than one model. This can lead to conflicting changes in different indicators adopted to represent one ecosystem service. For instance, in this study three different indicators were used to quantify the health benefits of urban green, all of them yielding different results (e.g. reduction in the probability of being overweight vs. reduction in visits to GP). Aggregation can also lead to double-counting, which occur when there are overlaps within the elements covered by different indicators. This leads to the inflation of ecosystem service values where double-counting takes place, at the expense of the remaining ecosystem services. Moreover, this study does not discern between intermediate and final benefits (e.g. PM10 retention is an intermediate benefit, while the reduction of diseases associated with PM10 retention is a final benefit), which can also lead to double-counting. 4.2.3

Monetization It is inappropriate and not recommended to aggregate monetized ecosystem service indicators since the value generated can be misleading and may be easily misconstrued. Firstly, not all ecosystem service indicators were monetized. For instance, in addition to reduced sewage costs, water storage leads to a reduction in the risk of flooding and the associated costs, yet the latter was not included in this study. Furthermore, various ecosystem services were not considered in this study, due to restrictions in time, models, and data. One example is biodiversity, which although often not considered an ecosystem service, underpins ecosystem service delivery. The aggregation of monetized ecosystem service indicators may lead to the overstatement of monetized values against non-monetized values. Aggregating monetized ecosystem service indicators may also lead to overstated values due to the potential occurrence of double-counting (i.e. adding two or more indicators that partially overlap).

4.2.4

Model uncertainty For this study, the Natural Capital Model (NK-Model) was used, alongside additional calculations developed for this study’s purpose (Remme et al., 2018). The NK-Model is currently under development and undergoing its first implementation stages. This study made use of published urban subset of models from the NK-Model. Additional models developed for this study have not been published and are not all based on peer-reviewed literature. Table 4.1 presents a qualitative overview on the level of uncertainty associated with the models adopted.

Page 59 of 79

RIVM Letter report 2019-0021

Table 4.1: A qualitative overview on the degree of uncertainty per ecosystem service model. For every ecosystem service, one indicator is presented per model (in cases where one model leas to more than one indicator). The degree of uncertainty is categorized into three levels: moderately uncertain, uncertain, and very uncertain.

Ecosyste m Service

Indicator

Air quality

PM10 retention

Health

Reduction in probability of being overweight

Uncertaint y level

Reduced health costs due to urban green Additional time spent on outdoor physical activity

Physical activity

Cycling Time spent cycling to-from work

Property value

Contribution to real estate value

Recreation

Visits to recreation areas

Urban cooling Water storage

Decrease in temperature Additional storage by green areas Moderately uncertain: Based on published models, based on peer-reviewed literature Uncertain: Based on peer-reviewed literature Very uncertain: Not based on peer-reviewed literature

Identified model-related uncertainties result mainly from the lack of evidence necessary to model ecological functions (e.g. water retention by vegetation and soil in cities) and their benefits (e.g. relationship between green spaces and cycling) at various scales and levels of urbanization. This sometimes led to the adoption of relationships which are either too general (e.g. the application of Dutch relationships to the Amsterdam scale) or somewhat outdated (e.g. modelling visits to GPs based on Maas (2008). In some cases, associations have either only been discussed in theory (e.g. association between cycling-lane length and cycling) or may lead to contradicting conclusions (e.g. studies showing both positive and negative relationships between green space and cycling). Where no relationships were found (i.e. cycling, visits to recreational areas, expenditures made in recreational areas), extrapolations were made based on statistical trends. These are to be used with precaution since they constitute a high degree of uncertainty and were rather produced as a means of comparison among ecosystem service indicators. Provided the substantial amount of readily available Dutch data (input maps) relevant for modelling relationships between green spaces and Page 60 of 79

RIVM Letter report 2019-0021

ecosystem services, it would be useful to place more efforts in conducting new studies that provide insights on current relationships between green space and ecosystem service indicators, and at various scales (city, region, areas with similar demographic structures). 4.3

Uses and recommendations This study’s results serve as an instrument to (1) inform decisionmakers on the value of natural capital, (2) communicate the value of natural capital to citizens, (3) evaluate the consequences of changes in natural capital (e.g. based on trends or policy strategies) on human well-being. This study’s results are not meant to be used for forecasting purposes or for analysing the actual cell-by-cell distribution of ecosystem services. While methods for quantifying and mapping ecosystem services have seen significant advances in the last decades, various models constitute a high degree of uncertainty, due to limitations in data and knowledge or evidence. It is important to bear these points in mind when using ecosystem service assessment results to support decision-making. Future research should focus on the further development of models for quantifying and mapping ecosystem services at different scales. For instance, models tailored to the Netherlands and more specifically to different regions or levels of urbanization could 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.

4.4

Implications for Amsterdam’s Structural Vision Amsterdam’s Structural Vision 2040 laid out an ambitious mission of achieving an economically robust, internationally competitive, and sustainable city by the year 2040 (Gemeente Amsterdam, 2011). The four strategies designed to address this challenge were embodied in Amsterdam’s Green Agenda 2015-2018 and is further developed within the Quality Impulse Green (Gemeente Amsterdam, 2015, 2017). Green infrastructure includes a range of societal benefits, improving health, water storage and recreation. This study showed that the average benefits for the different strategies range between 1 and 5 €/m2/year. An overview on the way in which each green infrastructure scenario can contribute to the objectives addressed by each strategy, is presented below. To avoid confusion, it is worth mentioning that the names from the strategies from the Green Agenda and Quality Impulse Green (Gemeente Amsterdam, 2015) may overlap with the names that were given to green infrastructure scenarios.

4.4.1

Quality Impulse Green strategy 1: Urban Parks Park visitation rates have sharply risen in Amsterdam throughout the past decades, a continuing trend (Gemeente Amsterdam, 2015). The rising demand for green recreational sites is addressed by the strategy ‘Urban Parks’ (Stadsparken) (Gemeente Amsterdam, 2015). The Page 61 of 79

RIVM Letter report 2019-0021

strategy proposes an increase in the quality and availability of green spaces for recreational use, addressing the rising demand for recreational sites while also reducing pressure on sites where use is highly intensive (e.g. Vondelpark). The increase in quality of green recreational sites includes altering the vegetation in parks in order to ensure the appropriate dryness of the grass, while ensuring that the quality of the soil remains constant. This strategy is addressed by the Urban Parks and Metropolitan Landscape scenarios. Within the two scenarios, recreational areas are expanded, their quality is enhanced through an increase in their management level, and accessibility is enhanced through the expansion of cycling and walking paths. This study’s results show that the implementation of these scenarios will lead to a higher use of green spaces, with a redistribution of recreational visits to urban parks where use is currently non-intensive, and to green areas in the outskirts of the city, which will be made more accessible. Moreover, all scenarios entail an increase in outdoor physical activity due to the expansion of green areas, accentuating the need for not just more but also higher quality green. 4.4.2

Quality Impulse Green strategy 2: Climate and biodiversity Climate change leads to more severe weather extremes, such as heatwaves and more intense precipitation, posing a risk to people living in cities. The strategy ‘Climate and Biodiversity’ (Klimaat en biodiversiteit) addresses the challenges posed by climate change by increasing the overall area that green spaces cover, thereby also enhancing biodiversity (Gemeente Amsterdam, 2015). Strategies for expanding green coverage include the implementation of green roofs and areas for rainwater storage where it poses a risk. This strategy is addressed by the Green Neighborhoods, Green Network, and Urban Parks scenarios, which comprise a substantial increase in the amount of green throughout the Amsterdam region. Green space has a positive effect on the amount of water stored from rainfall. It also contributes to lower temperatures during warm days, reducing the heat island effect commonly experienced by cities. As such, it can contribute to tackling the weather extremes that result from climate change, positively affecting people’s health and well-being. Green spaces also contribute to enhancements in biodiversity, playing a crucial role as habitats for various organisms. However, biodiversity was not quantified in this study, due to time, model, and data restrictions.

4.4.3

Quality Impulse Green strategy 3: Green Neighbourhoods Green areas in local neighbourhoods are important sites of social cohesion, while also providing a number of additional benefits, such as enhanced health and improved air quality. The strategy ‘Green Neighbourhoods’ (Groen in de buurt) targets an increase in the quality and availability of green spaces in Amsterdam’s neighbourhoods (Gemeente Amsterdam, 2015). This includes adding more green elements in the near vicinity of households and making green spaces available for local residents (children, elderly, people with disabilities) to engage in green initiatives. Page 62 of 79

RIVM Letter report 2019-0021

The scenario Green Neighbourhoods addresses this strategy directly, making green spaces available in the near vicinity of households. The effects of small, sparsely-distributed green spaces on social cohesion, were not quantified, due to model, data, and time restrictions. However, it is clear that adding green in the near vicinity of households can deliver an array of additional benefits to Amsterdam’s residents. The implementation of this scenario was found to generate substantial benefits in terms of air quality, physical activity, health, property value and heat reduction. This is especially the case in areas where green is currently scarce. 4.4.4

Quality Impulse Green strategy 4: Green Networks and accessibility The rate at which people recreate in green spaces in Amsterdam is sharply rising. A large portion of recreational visits remains within a limited number of recreational spaces (e.g. Vondelpark, Westerpark), despite the increasing rate of visitation to these sites. This results, in part, from the lack of appropriate infrastructure required for people to reach other existing green areas. The strategy ‘Green Networks and Accessibility’ (Verbindingen en toegankelijkheid) entails an improvement in the city’s ecological network and its accessibility (Gemeente Amsterdam, 2015). Enhancing accessibility to green elements includes expanding and interconnecting existing walking and cycling paths. The enhancement of ecological networks is addressed by the scenario Green Network, while the scenario Metropolitan Landscape addresses the need to improve accessibility to green spaces. Green networks are a useful way to enhance biodiversity. This ecosystem service was not quantified in this study due to time, model, and data restrictions. However, ecological networks were also found to generate significant contributions to temperature decrease, water storage, and the amount of time people spend on physical activity.

The implementation of the Metropolitan Landscape scenario could additionally lead to significant increases in cycling and recreation in the outskirts of the city, due to the significant improvement in accessibility to green areas.

Page 63 of 79

RIVM Letter report 2019-0021

Page 64 of 79

RIVM Letter report 2019-0021

References

Arcadis & CE-Delft, 2018. Werkwijzer Natuur: Maatschappelijke KostenBaten Analyses. Bennett E.M., Peterson G.D., Gordon L.J., 2009. Understanding relationships among multiple ecosystem services. Ecol Lett. CE-Delft, 2017. Handboek Milieuprijzen 2017. Methodische onderbouwing van kengetallen gebruikt voor waardering van emissies en milieu-impact. Publicatienummer: 17.7A76.64, Delft. Czembrowski P., & Kronenberg J., 2016. Hedonic pricing and different urban green space types and sizes: Insights into the discussion on valuing ecosystem services. Landscape and Urban Planning. 146. 11-19. Franco S.F., & Macdonald J.L., 2017. Measurement and valuation of urban greenness: Remote sensing and hedonic applications to Lisbon, Portugal. Regional Science and Urban Economics. Gemeente Amsterdam, 2011. Amsterdam 2040 Economisch Sterk en Duurzaam. Amsterdam. Gemeente Amsterdam, 2015. Agenda Groen 2015-2018: Investeren in de Tuin van de Amsterdammer. Amsterdam. Gemeente Amsterdam, 2017. Tussenrapportage Ruimte voor de Stad: naar een metropool op menselijke maat. Amsterdam. Hattam C., Atkins, J.P., Beaumont N, Börger, T., Böhnke-Henrichs, A., Burdon, D., De Groot R., Hoefnagel, E., Nunes, P.A.L.D., Piwowarczyk, J., et al., 2015. Marine ecosystem services: Linking indicators to their classification. Ecol Indic. 49:61–75.

Kahlmeier S., Götschi T., Cavill N., Castro Fernandez A., Brand, David Rojas Rueda C., Woodcock J., Kelly P., Lieb C., Oja P., Foster, Harry Rutter C., Racioppi F., 2017. Health economic assessment tool (HEAT) for walking and for cycling. Methods and user guide on physical activity, air pollution, injuries and carbon impact assessments. World Health Organization. Kelly P., Kahlmeier S., Götschi T., Orsini N., Richards J., Roberts N., ... & Foster C., 2014. Systematic review and meta-analysis of reduction in all-cause mortality from walking and cycling and shape of dose response relationship. International journal of behavioral nutrition and physical activity. 11(1). 132. Klompmaker J.O., Hoek G., Bloemsma L.D., Gehring U., Strak M., Wijga A.H., ... & Janssen N.A., 2018. Green space definition affects associations of green space with overweight and physical activity. Environmental research. 160. 531-540. KPMG, 2012. Groen, gezond en productief. The Economics of Ecosystems & Biodiversity (TEEB NL): natuur en gezondheid. Lauwaet D., Hooyberghs H., Maiheu B., Lefebvre W., Driesen G., Van Looy S., De Ridder K., 2015. Detailed Urban Heat Island projections for cities worldwide: dynamical downscaling CMIP5 global climate models. Climate, 3, 391-415. Lester S.E., Costello C., Halpern B.S., Gaines S.D., White C., Barth J.A., 2013.

Evaluating tradeoffs among ecosystem services to inform marine spatial planning. Mar Policy. 38:80–89.

Page 65 of 79

RIVM Letter report 2019-0021

Luttik J. & Zijlstra M., 1997. Woongenot heeft een prijs; Het waardeverhogend effect van een groene en waterrijke omgeving op huizenprijzen. Wageningen SC-DLO (Rapport 562). Maas J., 2008. Vitamin G: green environments-healthy environments (Doctoral dissertation. NIVEL). Maas J., Verheij. R. A., Spreeuwenberg. P., & Groenewegen. P.P., 2008. Physical activity as a possible mechanism behind the relationship between green space and health: a multilevel analysis. BMC public health. 8(1). 206. Mouchet M.A., Lamarque P., Martín-López B., Crouzat E., Gos P., Byczek C., Lavorel S., 2014. An interdisciplinary methodological guide for quantifying associations between ecosystem services. Glob Environ Chang. 28:298–308. Provincie Noord Holland, 2016. Bezoekersonderzoek natuur- en recreatiegebieden Noord Holland 2016. Remme R., de Nijs T., Paulin M., 2018. Natural Capital Model: Technical documentation of the quantification, mapping, and monetary valuation of urban ecosystem services. National Institute for Public Health and the Environment (2017-0040). RIVM, 2003 onwards. Cijfertool kosten van ziekten met cijfers uit de 'kosten van ziektestudie’. RIVM, Bilthoven. Romijn, G. & Renes, G., 2013. Algemene leidraad voor maatschappelijke kosten-batenanalyse. Central Planbureau voor de Leefomgeving.

Ruijgrok E.C.M., Smale A.J., Zijlstra R., Abma R., Berkers R.F.A., Németh A.A., Asselman N., de Kluiver P.P., de Groot D., Kirchholtes U., Todd P.G., Buter E., Hellegers P.J.G.J. and Rosenberg F.A., 2006. Kentallen Waardering Natuur, Water, Bodem en Landschap, Hulpmiddel voor de MKBA. Witteveen+Bos, commissioned by Ministerie van Landbouw Natuurbeheer en Voedselkwaliteit, The Hague. Staatsen B., van der Vliet N., Kruize H., et al. 2017. INHERIT: Exploring triple-win solutions for living, moving and consuming that encourage behavioural change, protect the environment, promote health and health equity. EuroHealthNet, Brussels. Steenbeek R., Hooftman W., Geuskens G., Wevers C., 2010. Objectiveren van gezondheidsgerelateerde non-participatie en de vermijdbare bijdrage van de gezondheidszorg hieraan. TNO, TNO-rapport 2010.171/13738.01.01. Wesseling J., van der Zee S., van Overveld A., 2011. Het effect van vegetatie op de luchtkwaliteit. National Institute for Public Health and the Environment (680705019/201).

Page 66 of 79

RIVM Letter report 2019-0021

Appendix 1 Social cost-benefit analysis

The Nature Guideline (Arcadis & CE-Delft, 2018) lays out a number of steps that should be fulfilled when conducting SCBAs regarding policies that lead to changes in natural capital in the Netherlands. Hereunder, the steps are presented in brief and insights are provided on the steps that are covered within this report. Table A1: Criteria that need to be covered within a SCBA to meet the guidelines expressed within the ‘General guidelines for social cost-benefit analyses’ (Romijn & Renes, 2013) and the ‘Nature Guideline’ (Arcadis & CE-Delft, 2018). Column 2 presents the location within this report where each step is represented (unless stated otherwise).

Step Report section 1. Structuring the SCBA Problem analysis and scope 1, 4.3 Policy alternatives/scenarios 2.2, 4.4 Reference alternative 2.2 Market demarcation Amsterdam Rule of half Not considered 2. Determining the physical changes in nature Expected and unexpected changes 2.2, 2.5 Uncertainties within predicted changes 4.2 3. From physical changes in nature to effects on human well-being Effects on human well-being from changes on 3.1, 3.2 ecosystem services Quantification of relevant ecosystem services 3.1 Quantification of changes in ecosystem services 3.2 Dealing with biodiversity Not considered 4. Valuation of effects on human well-being Implementation advice regarding valuation methods 4.2, 4.3, 4.4 Overview and description of valuation methods 2.3, 2.4, 2.5 Use of benefit transfers 2.4 5. Determine the costs Relevant costs Not considered Taxes and subsidies Not considered 6. Uncertainties and risks Future, policy, and knowledge uncertainties 2.4, 4.2 How to deal with uncertainties and risks 4.2, 4.3 7. Present results Overview of costs and benefits 3 3.1, 3.2 Distribution effects (partial)

Page 67 of 79

RIVM Letter report 2019-0021

Appendix 2 Additional models

A2.1 Health Indicator: Probability of being overweight Unit: Value higher or lower than 1. where 1 is the reference value. respective to 50% or less green space within a buffer of 3000 meters Source: Klompmaker et al. (2017) A study conducted by Klompmaker et al. (2017) found a negative relationship between green space and being overweight (body mass index (BMI) ≥ 25 kg/m2) within the Netherlands. The study was based on a national health survey (Public Health Monitor 2012. PHM) conducted with 387,195 adults. To measure green space. the Normalized Difference Vegetation Index (NDVI) was used. NDVI captures the density of green vegetation at a spatial resolution of 30m. Surrounding greenness was measured as the average NDVI within a circular buffer of the participant's residential address. The study found a negative relationship between the percentage of green space within a 3000m buffer and the probability of being overweight (Table A2). Percentages of green space were divided into quintiles (column 1). where the first quintile represents the reference category. As Table A2 reveals, as the percentage of green space increases within every quintile, the probability of having a BMI ≥ 25 kg/m2 decreases. Values in bold are significant at a 5% level. Table A2: Relationship between different percentages of green space within a buffer of 3000m and being overweight (BMI ≥ 25 kg/m2)

Percentage of green space (quintiles – NDVI) ≤ 50 51–57 58–63 64–69 > 69

Probability of being overweight (95% CI) ref 1.00 (0.97, 1.02) 0.96 (0.93, 0.99) 0.93 (0.91, 0.96) 0.92 (0.89, 0.95)

Steps 1. For every grid cell. estimate the percentage of green space within a 3000m buffer (X1) 2. For every grid cell. estimate the probability that individuals are overweight (BMI ≥ 25 kg/m2) (X2) Percentage of green space (X1) ≤ 0.50 0.51–0.57 0.58–0.63 0.64–0.69 > 0.69 Page 68 of 79

Probability of being overweight (X2) ref 1.00 0.96 0.93 0.92

RIVM Letter report 2019-0021

A2.2 Physical activity Indicator: Additional time spent on outdoor physical activity Unit: minutes/year Source: Klompmaker et al. (2017) A study conducted by Klompmaker et al. (2017) found a positive relationship between green space and outdoor physical activity within the Netherlands. The study was based on a national health survey (Public Health Monitor 2012, PHM) with 387,195 adults. Outdoor physical activity was defined as all moderate and vigorous physical activities that can be done outdoors (physical activity for commuting purposes, leisure time physical activity (walking, cycling, gardening). and outdoor sports). To measure green space. the Normalized Difference Vegetation Index (NDVI) was used. NDVI captures the density of green vegetation at a spatial resolution of 30m. Surrounding greenness was measured as the average NDVI within a circular buffer of the participant's residential address. The study found a positive relationship between green space within a buffer of 300m and the change in minutes adults engage in outdoor physical activities for every quintile (Table A3). Percentages of green space were divided into quintiles (column 1), where the first quintile represents the reference category. As Table A3 reveals, as the percentage of green space increases within every quintile, the minutes individuals invest in outdoor physical activities increases.

Table A3: Percentage of green space within a buffer of 300m and changes in the amount of min/week that individuals invest in outdoor physical activity. Values in bold are significant at a 5% level.

Percentage of green space (quintiles – NDVI) ≤ 40 45-50 51-55 53–59 > 59

Minutes extra spent on outdoor physical activity (95% CI) ref 3.5 (-3.5. 10.6) 9.4 (2.2. 16.5) 18.6 (11.3. 25.9) 43.8 (36. 51.6)

Page 69 of 79

RIVM Letter report 2019-0021

Steps 1. For every grid cell, estimate the percentage of green space within a 300m buffer (X1) 2. For every grid cell, estimate the population (X2) 3. For every grid cell, estimate the individual increase in in outdoor physical activity (min) (X3) Percentage of green space (X1) ≤ 0.40 0.45-0.50 0.51-0.55 0.53–0.59 > 0.59

Minute increase in outdoor physical activity (X3) 0 3.5 9.4 18.6 43.8

4. For every grid cell, estimate the total increase in outdoor physical activity (min) (X4) X2 * X 3

Page 70 of 79

RIVM Letter report 2019-0021

Indicator: Cycling Unit: km/year Source: Gemeente Amsterdam, CBS To calculate the increase in cycling in Amsterdam, a simple calculation was performed, where the increase in kilometres cycled by the city’s inhabitants is proportional to the increase in the length of Amsterdam’s cycling network within each scenario. The amount of km cycled per resident per year in North Holland was used as a proxy for the amount of km cycled per resident per year in Amsterdam. The growth rate in km cycled per resident per year was used to estimate the increase in km cycled by 2025, within a Business-As-Usual situation.

Page 71 of 79

RIVM Letter report 2019-0021

Indicator: Time spent cycling to-from work Unit: minutes per year Source: Maas et al. (2008) A study conducted by Maas et al. (2008) studied the relationship between green space and cycling within the Netherlands. The study was based on a national health survey (Second Dutch National Survey of General Practice, DNSGP-2) with 4,899 people. To measure green space, the National Land Cover Classification database (LGN4) was used. LGN4 contains the dominant type of land use of each 25×25m grid cell in the Netherlands in 2001. Surrounding greenness was measured as the percentage of green space within a circular buffer of the participant's address. 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 𝑝𝑝𝑝𝑝𝑝𝑝 𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 (𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑) = 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 + (0.83 × 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑥𝑥 100)

Formula 1

The study found a positive relationship between the percentage of green space within a buffer of 1000m and the number of minutes cycled for commuting purposes. It was found that, for every percentage increase in green space, people who cycle to work for commuting purposes will cycle 0.83 additional minutes on average. This information can be used to extrapolate Formula 1. Moreover, people with 20% green space within a 1000m radius around their home cycle 120 minutes per week for commuting purposes, whereas people with 80% green space within a 1000m radius cycle approximately 170 minutes per week for commuting purposes. Based on this information, the intercept is estimated at 103.4 minutes per week per person (Figure A1). Figure A1: Relationship between the percentage of green space and the minutes people engage in cycling for commuting purposes