HEAT VULNERABILITY
IN AMSTERDAM
Urban Climate AdaptationINGE WISSEL
Bachelor project Future Planet Studies
28 May 2021, Amsterdam 12218480
Supervised by Katinka Wijsman and Jannes Willems
1
Abstract
Due to climate change, average temperatures are rising and more extreme weather events are projected to take place. This development is reinforced by the Urban Heat Island effect, which causes urban areas to have an increased temperature up to ten degrees Celsius. Heatwaves have caused great quantities of excess deaths, inducing governments to create adaptation strategies. This research aims to determine where these adaptations are most needed in Amsterdam, by assessing heat vulnerability. The research question is therefore: Which factors explain differences in heat vulnerability between neighbourhoods in Amsterdam?. Vulnerability is divided into two dimensions; heat exposure, which is influenced by spatial factors such as building density and green space, and heat sensitivity, which depends on age and weight, among other indicators. First, a literature review of vulnerability is conducted. Subsequently, spatial data is processed with ArcGIS. Finally, thematic maps are used to identify the most vulnerable parts of Amsterdam. The results show that heat vulnerability is an urgent problem in the centre and West of Amsterdam. In these neighbourhoods there is a deficiency of green and a high level of building density. Furthermore, large numbers of people with an increased vulnerability to heat live in these areas, such as elderly people, people with overweight and people who live alone. This research adds a new understanding and operationalisation of heat vulnerability to the existing literature, which can be applied to other cases, as heat vulnerability is an issue which will only increase in complexity and scope over the coming years.
2
Index
Abstract ... 1 Introduction ... 3 Theoretical framework ... 6 Vulnerability ... 6 Heat exposure ... 6 Heat sensitivity ... 8 Conceptual framework ... 9 Methodology ... 11 Research Design ... 11 Operationalisation ... 12 Data collection ... 15 Data analysis ... 16Justification and limitation of methods ... 18
Data handling ... 18 Ethical considerations ... 18 Results ... 19 Discussion ... 27 Methodological reflection ... 27 Theoretical reflection ... 29 Conclusion ... 33 Bibliography ... 35
3
Introduction
The global climate is changing, due to increasing greenhouse gas emissions since the industrial revolution (Parry et al., 2007). One of the consequences of climate change is an increase in average global and regional temperatures. For example, Europe has experienced an increase of 0.90 °C over the last century (1901–2005) (De Sario, Katsoujanni, & Michelozzi, 2013). Furthermore, more extreme weather events will take place globally, leading to an increase in heat waves in the Netherlands (Meehl & Tebaldi, 2004; Parry et al., 2007). This increase is even more problematic in cities, because of the Urban Heat Island (UHI) effect. This effect refers to the fact that cities can be up to ten degrees Celsius warmer compared to rural areas (Heaviside, Macintyre, & Vardoulakis, 2017). As more than half of the global population currently lives in urban areas, it is important to research the effects of (extreme) heat (Ritchie & Roser, 2018).
In 2003, the population of Europe experienced a severe heatwave, which caused approximately 35.000 excess deaths (Parry et al., 2007). This event induced governments to prepare their countries for future heat waves (Kovats & Hajat, 2008). This was often in the format of a heat warning system, such as the Dutch National Heat plan, but was less translated into preventive spatial implications (Hagens & Van Bruggen, 2014). Furthermore, the Gemeentelijke Gezondheidsdiensten (GGD), the municipal health services of the Netherlands, have developed a list with symptoms that could be caused by extreme heat and how to cope with them (GGD Leefomgeving, 2021). The GGD state that some people are more vulnerable to the effects of a heatwave, such as elderly people, people who live alone and overweight people (ibid.). Furthermore, the organisation of someone’s spatial environment can make people more exposed to heat, such as the building density, amount of asphalt, and a lack of green space (Rizwan, Dennis, & Liu, 2008; Harlan et al., 2006). When the spatial distribution of these two factors – heat sensitivity and heat exposure – are researched and combined, differences in vulnerability and the distribution of environmental disadvantages in society can be distinguished (Lindley et al., 2011).
Contrary to popular belief, deaths caused by extreme weather events are not mere acts of nature (Hamstead et al., 2021). The effects of extreme weather events, such as heatwaves, are not distributed evenly over society (Lindley et al., 2011). Which social groups have, or receive, resources to prepare for or recover from such events is influenced by multiple factors, including politics, power, and finances (Hamstead et al., 2021). To which extend someone is affected
4 does not only depend on their exposure to the event, but also on their sensitivity to it (Lindley et al., 2011). These environmental injustice practices are often deeply imbedded in the predominant institutions, resulting in an unequal distribution of climate burden (Hamstead et al., 2021). The findings of Johnson et al. (2012) indicate that heat vulnerability models should be developed on a local scale, so local variations in vulnerabilities can be taken into account and inequality is diminished. Harlan et al. (2006) state; “Urban heat island reduction policies should specifically target vulnerable residential areas and take into account equitable distribution and preservation of environmental resources” (p. 2847). Given their importance in (urban) climate adaptation issues, this research aims to map local variations in heat vulnerabilities.
This research is conducted in the capital of the Netherlands: Amsterdam. Amsterdam has one of the strongest UHI of Europe, with a maximum increase of 7-9 degrees Celsius (Van der Hoeven & Wandl, 2015). Furthermore, Amsterdam is the largest city of the Netherlands and an important hub with Schiphol airport and Westport harbour (ibid.). At the beginning of 2021, 872.922 people lived in the city, a number that increases by a rate of ten thousand people per year (CBS, 2021a; Couzy, 2021). However, due to the Covid-19 global pandemic, this increase has stagnated (Couzy, 2021). One previous study on heat vulnerability in Amsterdam has been conducted by Van der Hoeven & Wandl, (2015), but a different conceptual framework is constructed in this research.
Figure 1. Districts of Amsterdam (Gemeente Amsterdam, n.d. a)
North Waterland Centre West South East South-East New-West Westpoort
5 Recently, the municipality of Amsterdam published a policy document on their adaptation strategy to increasing urban heat (Gemeente Amsterdam, 2020). However, this policy does not address local differences in vulnerability and exposure. As discussed by Kleerekoper, Van Esch, and Baldiri Salcedo (2012), quite some knowledge on the UHI effect has been accumulated. However, the biggest obstacle now is implementing this knowledge into the urban planning process. Transferring knowledge into practice is difficult, because a lot of different actors are involved in the planning process. Additionally, the scientific literature about the UHI effect is by nature more theoretical than practical. Some tools that could improve the urban planning domain are an easily accessible overview of basic principles and guidelines for design (Kleerekoper, Van Esch, & Baldiri Salcedo, 2012). To create the most effective policy against the UHI effect, it is important to know where the most heat-sensitive people live and in which neighbourhoods the spatial planning situation creates the most heat exposure. To study this distribution, the following research question was constructed: Which factors explain differences in heat vulnerability between neighbourhoods in Amsterdam?. To answer this research question, the following sub-questions were studied;
1. Which spatial factors make people more exposed to urban heat? 2. Which individual factors make people more sensitive to urban heat? 3. Where are these factors present in Amsterdam?
The research results are useful for a wide array of stakeholders. The findings of this paper can guide urban planners from the municipality of Amsterdam, in making the city more resilient to climate change. Besides, NGOs might find the data of this paper interesting. For instance, small initiatives could use this information to motivate people to participate, such as the NGO Grachten van Smaragd which aims to effectuate an increase of green rooftops on the houseboats of Amsterdam (Grachten van Smaragd, n.d.). Furthermore, this paper can alarm citizens about the circumstances in their neighbourhood and motivate them to take action. Additionally, the conceptualization of heat vulnerability can be used by other researchers and applied to other cases. The conceptual framework and operationalisation of this research provide new insights to the academical literature about urban heat vulnerability, which could be applied to locations with similar urban and climate characteristics.
6
Theoretical framework
Vulnerability
In urban policy literature, multiple definitions and operationalisations of vulnerability are used. Lindley et al. (2011) state “Vulnerability of an individual or group is characterised by the degree to which an external event converts into losses in their well-being.” (p. 6). According to Johnson et al. (2012), “It [vulnerability] is typically defined as the inability, of a specific group or population, to appropriately respond or adapt to a specific harmful stressor” (p. 23). This paper builds further on the definition of vulnerability composed by Adger (2006); “Vulnerability is the state of susceptibility to harm from exposure to stresses associated with environmental and social change and from the absence of capacity to adapt” (p. 268). The exposure to increasing urban heat, ‘stresses associated with environmental change’, is measured, as well as the sensitivity to this heat, ‘people who are less capable to change’, in this research. In the following paragraphs these two concepts will be elaborated.
Bao, Li, and Yu (2015) reviewed fifteen articles about heat vulnerability assessments. The thirteen most used indicators for heat vulnerability are; temperature level, population density, age, gender, pre-existing medical conditions, education, income, poverty, minority status, acclimatization, and access to home amenities (e.g. air conditioning and swimming pools). However, for each location different indicators are the most determinative.
Heat exposure
In the future, the inhabitants of The Netherlands have to cope with an increased amount of heat waves (Meehl & Tebaldi, 2004). The National Knowledge and innovation program Water and Climate (NKWK) defines a heat wave as a subsequent period of five summer days (25 °C or above), with at least three tropical days (30 °C or higher) (NKWK, 2019). In just sixty years (from 1950 to 2010), the number of summer days has more than doubled in the Netherlands (KNMI, 2018). Because of the moderate maritime climate (Cfb) of the country, based on the Köppen-system, the relative humidity is often high (Peel, Finlayson, & McMahon, 2007). The average humidity at twelve o’clock UTC in Amsterdam during the summer is 66-68 percentage, which significantly increases the experienced temperature by humans (KNMI, n.d.; Davis et al., 2003). Table 1 shows the heat index created by Robert Steadman, which combines temperature and relative humidity to calculate the apparent temperature (NKWK, 2019). The
7 index indicates a notable temperature rise due to the Dutch climate, resulting in an increased exposure to heat and risk of heat disorders (Harlan et al., 2006).
Table 1. Heat index (Gendler, 2017; Harlan et al., 2006)
The following definition for heat exposure is used in this research; “realized contact between a person and an indoor or outdoor environment that poses a risk of increases in body core temperature and/or perceived discomfort” (Kuras et al., 2017, p. 1). Many factors influence the temperature of an environment, therefore considerable local differences can be experienced (Harlan et al., 2006). A commonly known example of this is the difference between standing in the shadow or in direct sunlight (Van der Hoeven & Wandl, 2013). However, there are also permanent factors which influence heat, such as sky view, albedo, green space, and building density (Harlan et al., 2006; Van der Hoeven & Wandl, 2013). This research aims to find if these characteristics are unevenly present over different neighbourhoods in Amsterdam. The sky view factor indicates how much the earth’s surface is exposed to the sky vault (Van der Hoeven & Wandl, 2013). When there is a low amount of exposure, the build structures and ground cannot release heat, which was locked in during the day, at night (ibid.). For example, the water of a ditch does not freeze underneath a bridge, because there is less release of heat, so the water stays liquid. However, this indicator is primarily of importance during the night. The albedo level indicates to which extend solar radiation is reflected of a certain type of surface (ibid.). When solar radiation is reflected, instead of detained, this has a diminishing effect on
Label Heat disorders Caution Fatigue possible;
discomfort Extreme
caution
Sunstroke, heat cramps, heat exhaustion possible Danger Sunstroke, heat cramps,
heat exhaustion likely and heat stroke possible Extreme
danger
Sunstroke and heat stroke highly likely
8 the temperature. However, due to clouds and air pollution, the radiation can be reverberated back to the earth. Nevertheless, in general, surfaces with a high albedo level have a reducing effect on the UHI effect, compared to surfaces with a low albedo level, such as asphalt. Nature decrease urban heat by releasing water vapor from transpiration and using latent energy for photosynthesis (ibid.). The amount of UHI reduction differs per type of green space; a forest provides lots of shade, but retains more heat at night, compared to a meadow with a high sky view factor.
Buildings attribute to the UHI effect in multiple ways (ibid.). When a building is constructed, the space it occupies becomes sealed from water evaporation and vegetation growth, so solar radiation cannot be converted into latent energy. Furthermore, materials which are often used to build, such as concrete and bricks, have a relative low albedo. This effect can be countered by applying a coating with a high albedo level or painting buildings in light colours, which is often done in areas with hotter climates. The building envelope is the physical separation between the interior and exterior climate, which determines the exposure of the buildings to sunlight, as well as the sky view factor of the building volume. At daytime, a small building envelope has a cooling effect, as little solar radiation is received. However, when a building has a high sky view level, it is more beneficial to have a large building envelope at night.
Heat sensitivity
Heat stress occurs when people are exposed to (extreme) heat and can no longer release this heat, resulting in an increased body temperature (Cheung, McLellan, & Tenaglia, 2000; Jay & Kenny, 2010). This condition has symptoms such as thirst, headaches, heavy sweating, weakness, muscle cramp, fatigue, dizziness, and skin rashes (Carter et al., 2020; Jay & Kenny, 2010). Even more severe symptoms can arise; for example fainting, vomiting, loss of consciousness and irrational behaviour, which might lead to heat strokes and eventually death (Carter et al., 2020; Jay & Kenny, 2010; Cheung, McLellan, & Tenaglia, 2000). The previously discussed heat index is based on many generalized factors, such as height, body mass, clothing, heat tolerance, and amount of physical activity (Epstein & Moran, 2006). Therefore, individual experiences of heat will differ and some people are more likely to suffer from heat stress (ibid.). Research has shown that older people, young children and people with cardio-vascular diseases have an increased sensitivity for heat illness (Bytomski & Squire, 2003). During extremely warm weather, illnesses and deaths from air pollution and infectious diseases also increase, making people with respiratory diseases more vulnerable (Harlan et al., 2006). Furthermore,
9 people who live in a single person household also have an increased risk of heat illness; there diminished social contact can make them less prepared for heatwaves or less checked upon (Lindley et al., 2011; Reid et al., 2009).
In this research, locations with an increased exposure to heat are examined, which could be seen as the distribution of environmental (heat) risk in the city. This approach, looking at the spread of environmental good and bads, is part of environmental justice theory (Schlosberg, 2013). Furthermore, environmental (in)justice is about “the material relationships between human disadvantage and vulnerability and the condition of the environment and natural world in which that experience is immersed” (ibid., p. 51). By mapping the differences in environmental risk, this research aims to effectuate more equity and set the subject on the political agenda (Schlosberg, 2004).
Conceptual framework
For this research, a new conceptual framework was created from the previously explained concepts (see figure 3). The conceptual model of this paper builds further on the work of Lindley et al. (2011), which is shown in figure 2. Lindley et al. (2011) define sensitivity as “personal biophysical characteristics such as age and health which affect the likelihood that a heatwave or flood event will have negative welfare impacts” (p. 9). However, in this research, a broader understanding of sensitivity is used, which also includes social influences, such as socially isolated people. The conceptual framework visualized in figure 2 also contains the concept adaptive capacity, which is divided into three sub-categories; ability to prepare, ability to respond and ability to recover, which consist of personal, environmental and social factors. However in this research adaptive capacity is seen as an inherent part of sensitivity, as biophysical and social factors are often intertwined.
10
Figure 2. Conceptual framework for assessing socio-spatial vulnerability and climate disadvantage (Lindley et al., 2011).
Figure 3 shows the conceptual model of this thesis, which is a simplified version of the framework of Lindley et al. (2011) and specified op UHI. Heat exposure and heat sensitivity both lead to an increase in vulnerability. Some individual's exposure to heat is enhanced because of their local circumstances, which increases their vulnerability. Furthermore, people who are more sensitive to the effects of heat stress or develop heat stress more quickly, are more vulnerable.
11
Methodology
Research Design
The methodology of this paper consists of qualitative and quantitative elements. The operationalisation of vulnerability is based on qualitative research methods, but the Geographical Information System (GIS) analysis is based on numerical data to provide a spatial answer to the research question. Previous research has been conducted on heat exposure in Amsterdam, for example by Van der Hoeven and Wandl (2013), however this paper proposes a different definition for vulnerability. A deductive approach with a case study is used to show the outcome of a different vulnerability understanding.
The research question; Which factors explain differences in heat vulnerability between neighbourhoods in Amsterdam? is answered through a literature review and GIS analysis. The first two sub-questions, Which spatial factors make people more exposed to urban heat? and Which individual factors make people more sensitive to urban heat? have been answered through literature research. This examination has resulted in a thorough operationalisation of these core concepts, as explained in the previous chapter. Subsequently, data about the spatial distribution of these indicators in Amsterdam was searched and processed with a GIS, resulting in multiple thematic maps. Combining these maps has uncovered the most vulnerable parts of Amsterdam, answering the final sub-question, Where are these factors present in Amsterdam?, and the main research question. This thematic mapping method is chosen for a clear representation and comprehension of the data for a broad audience.
Amsterdam was selected as case study for this research for multiple reasons. Heatwaves used to be a relative uncommon phenomenon in the Netherlands, but the number of summer days has doubled over the past sixty years and heatwaves last for longer periods of time (AMS, 2020; Meehl & Tebaldi, 2004; KNMI, 2018). These numbers will continue to increase, but because the issue is relatively new, our existing spatial structures are not build in an efficient way to obstruct heat (ibid.). This issue is mainly present in the centre of Amsterdam where the seventeenth century canal ring is located, which is listed as UNESCO world heritage, so making major changes to the spatial structure is prohibited (United Nations, n.d.). Furthermore, compared to other metropoles in Europe, Amsterdam is subject to a very strong UHI, which can be 7 to 9 degrees Celsius warmer than the surrounding rural areas (Van der Hoeven &
12 Wandl, 2015). As the quantity of inhabitants of the city keeps increasing, with ten thousand people per year, the housing stock of the city will also increase (CBS, 2021a). Therefore, research about the relationship between urban heat and spatial conditions is important. Research into the UHI effect has been conducted in multiple cities, for example Phoenix, Athens, Philadelphia, and Manchester (Bao, Li, & Yu, 2015; De Sario, Katsoujanni, & Michelozzi, 2013; Harlan et al., 2006; Lindley et al., 2011). However, in Amsterdam few studies has been done about the UHI effect, so the aim of this research was to create and test a new operationalisation of vulnerability for Amsterdam and other cities with similar characteristics. Amsterdam is an international city with important economic hubs, such as Schiphol airport, financial business district Zuid as and the Westport harbour (Van der Hoeven & Wandl, 2015). The city is located in the province North-Holland and connected to the North Sea through a channel (ibid.). Currently, the city has around 873.000 inhabitants, with more than 170 nationalities from around the world (CBS, 2021a; Ujcic-Voortman, 2020).
Operationalisation
Epstein and Moran (2006) have analysed different indices for heat exposure and divided them into three categories; rational, empirical and direct indices. The most comprehensive indices are the rational ones which are based on heat balance equations, however, the authors argue for the use of direct indices. As direct indices, based on direct measurements of environmental variables, are more practical and reflective of reality. Therefore, in this paper direct indices are used to measure heat exposure. Based on the definition of heat exposure discussed in the previous paragraph and the research of Harlan et al. (2006) and Van der Hoeven and Wandl (2013), the following indicators are selected; building density, population density, and green space. These indicators are represented on a ratio scale with five classes of equal intervals. Previous research has demonstrated that land-use patterns and landcover are the most important determinators of urban temperatures, which can be measured through these indicators (Harlan et al., 2006; Van der Hoeven & Wandl, 2013). Population density was also chosen as indicator to determine the spread of the population of Amsterdam and where measures are the most effective. Furthermore, the population density provides context for the heat sensitivity indicators; a large quantity of children in the zip code area with the highest number of inhabitants is more apparent than in a different area.
13
Table 2. Operationalization of the core concept heat exposure
Core concept Indicators Values
Heat exposure Building density Surrounding address density, which indicates the average number of addresses which fall into one kilometre radius around a point
Population density Number of people
Green space Square meters of green space
Heat sensitivity can have many underlying causes, which are elaborated in this paragraph. Elderly people have a reduced amount of total body water and sense of thirst, making them more at risk of dehydration and heat illnesses (Rikkert et al., 2009; Bytomski & Squire, 2003; Lundgren et al., 2013). Besides, this group is less active and has a reduced ability to transport heat from the body core to the skin (Lundgren et al., 2013). Following the work of Loughnan, Nicholls, and Tapper (2012), the indicator age above sixty-five is chosen. Secondly, Bytomski and Squire (2003) found children more susceptible to heat illness, due to multiple reasons; such as their lower rate of sweating, decreased thirst response, and greater surface area to body mass ratio. In the literature, different ranges of age are used for the indicator children, so in accordance with the available data, an age from zero to fourteen years was chosen as extend in this research. Furthermore, people with higher percentages of body fat have a lower heat tolerance, because they are less capable of storing heat (Cheung, McLellan, & Tenaglia, 2000). Additionally, people with specific diseases are more at risk for heat illnesses, such as lung, skin, liver and kidney problems (Lundgren et al., 2013). Flensner et al. (2011) found that many people diagnosed with Multiple Sclerosis are sensitive to increased body temperature. Individuals with Sickle cell disease are also hypersensitive to thermal change and experience heat pain sooner (Brandow et al., 2013). Table 3 shows a list of medical conditions which make people more vulnerable to heat, however for most of these conditions there is no spatial data available on a small scale, so they could not be incorporated in this research (Reid et al., 2009).
14
Table 3. Medical conditions linked with heat sensitivity (Brandow et al., 2013; Bytomski & Squire, 2003; Flensner et al., 2011; Lindley et al., 2011; Rikkert et al., 2009)
Excessive fluid loss Febrile state
Gastrointestinal illness Diabetes insipidus Diabetes mellitus Diminished thirst/intake Mental retardation
Young children Elderly people Suboptimal sweating Obesity Spina bifida
Cystic fibrosis
Sweating insufficiency syndrome Excessive sweating Congenital heart defects
Hyperthyroidism Temperature sensitivity Multiple sclerosis
Sickle cell disease Hypothalamic dysfunction Anorexia nervosa
Recent episode of heat illness Decreased adaptation capacity Disabled
People who live alone also have an increased risk of heat illness, as social contacts are important sources for preparing for heat events (Lindley et al., 2011). Researchers of the 1995 Chicago heatwave found that many victims of the heat episode died alone (Reid et al., 2009). Moreover, examinations indicated that people who lived alone or did not leave their homes each day, had a higher risk of death compared to people with a social network, during the same heatwave and the 1999 Chicago heatwave (ibid.). Harlan et al. (2006) found that “social isolation – living alone without regular contact with others – is a significant predictor of who succumbs to heat” (p. 2848). Living alone is not equal to social isolation, but the quantity of single person households in an area does give an indication of the scope of the problem.
15
Table 4. Operationalization of the core concept heat sensitivity
Core concept Indicators Values
Heat sensitivity Elderly people (age > 65) Young children (age < 14) Overweight/obese people Single person households
Number of inhabitants with this characteristic, also visualized on a ratio scale
Data collection
The data for this research is collected from multiple sources. However, a large part of it can be retrieved from the Centraal Bureau Statistiek (Statistics Netherlands, CBS) as shapefiles (CBS, 2020). From this extensive dataset, the necessary data for this research were selected; zip code areas, inhabitants, children, elderly, single person households and surrounding address density. To calculate the surface of green space per zip code area, the Main Green Structure data was used from the municipality of Amsterdam (Gemeente Amsterdam - Ruimte en Duurzaamheid, 2021). Prior to uploading the data in ArcGIS, the file was converted from a GEOjson file to a shapefile with http://mapshaper.org/>. Furthermore, excel data from the GGD was used to acquire information about inhabitants (age > 18) who are overweight or suffer from obesity (GGD Amsterdam, 2016). This data set included many factors, therefore a separate .csv file were created, which contained only the area code, overweight percentages and OID. Because the excel data set included specific local areas and a shapefile was available with these local areas, the data could be used in ArcGIS. To be able to combine the GGD data and the other data, an additional shapefile from the municipality of Amsterdam was used to assign the zip code areas to a division of Amsterdam into twenty-two regions (Gemeente Amsterdam - Onderzoek, Informatie en Statistiek, 2019). This file also had to be converted using http://mapshaper.org/>.
16
Data analysis
The flowchart of figure 4 shows the steps that were conducted to create the spatial maps of heat vulnerability in Amsterdam.
17 To create the CBS_AMS_selection shapefile, zip code areas with a population below 150 inhabitants were removed from the dataset, which were zip code area 1037, 1041, 1042, 1044 to 1047, and 1101 (see figure 5). This decision was made because the small populations made the data less reliable and significant. Furthermore, Amsterdam also has zip code areas with over 26.000 inhabitants, therefore the municipality will be more likely to implement heat measures in these areas.
Figure 5. Zip code areas of Amsterdam. Areas with inhabitants < 150 are excluded (marked in purple).
As previously mentioned, the values of the indicators are divided into five different classes of equal sizes to make the maps more comprehensive. Except for the representation of the green space; the data of this indicator has very large differences, from 0 to over 10,7 km2 of green space. If the equal size method would be used for this data, only four areas would have different classes from the rest, which would not provide useful information for this research. Therefore, the classes are divided into five natural breaks, following the method of Jenks (esri, n.d.). This method creates groups in the data with similar values, to maximize the differences between the classes (ibid.).
18
Justification and limitation of methods
Mapping spatial data is a precise method to investigate where social vulnerability and climate disadvantage come together (Lindley et al., 2011). This research is limited by available data about spatial distribution of characteristics, resulting in varying geographic units. The findings give an indication of vulnerability, but local-neighbourhood knowledge will give even more insights. Additionally, during this research generalizations of population groups are used to make it concise, however personal experience might differ from this study.
Data handling
For this research the software ArcGIS was used, which is a secure program (https://trust.arcgis.com/en/). The data of this research was stored on the laptop of the researcher and Google Drive. This website does not have optimal privacy, however the data used for this research is open data, so data backup is regarded as more important. To ensure the quality of the final data, a GIS teacher from the University of Amsterdam has given advise during the process and raw data was always saved unedited in a separate folder.
Ethical considerations
Resnik (2020) defines research ethics as “the norms for conduct, that distinguish between acceptable and unacceptable behaviour during research” (para. 1). Transparency and social responsibility are considered as key principles during this research. While composing a definition for vulnerability, the research aims to be inclusive, however this can be complicated, therefore it is salient to remain transparent about the made decisions. Furthermore, this research reflects the pragmatic view on epistemology and ontology of the researcher, but also her critical interests.
19
Results
The GIS analysis has resulted in twelve thematic maps, which are elaborated in this chapter. The data on heat exposure was combined with the data on heat vulnerability to determine where heat vulnerability is the greatest in Amsterdam. The indicators of heat exposure are represented in varying shades of a colour and the heat sensitivity indicators with differing circle sizes. First, the heat exposure results are discussed based on figure 7. Subsequently, the heat vulnerability indicators are elaborated.
The upper left map of figure 7 (p. 23) shows the amount of green in m2 per zip code area and indicates a lack of green space in the city centre and West-region. In a national research about green space, 75 m2 green per house was used as a minimum quantity for cities and villages (Bezemer & Visschedijk, 2003). In the research area of this study 447.300 houses are present, and 52.458.066 m2 of green space in total, which results in 117 m2 of green space per house (CBS, 2020). This seems sufficient, but this green space is unevenly distributed in the city, as visible in the upper left map of figure 7. When the five zip code areas with the highest surface of green space are taken out of the equation, only 49 m2 green space per house results (21.231.045 m2 green space divided by 431.605 houses). The limited amount of green space in the centre and West region of Amsterdam is confirmed by the research of Bezemer & Visschedijk (2003). Figure 6 shows the different types of green areas in the city and buffers of 500 meter around these areas. This map signals a lack of access to green in the centre and West of the city, making it inhabitants more susceptible to heat exposure.
20 Higher population densities can be found in New-West, East, and South-East, as shown in the right map in figure 7. The city has an average of 11.944 inhabitants per zip code area, with a range from 535 to 26.345 inhabitants (CBS, 2020). The zip code areas 1069 (Osdorp) in New-West and 1102 (Bijlmer Centrum) in South-East do stand out, with respectively 26.345 and 24.935 inhabitants. In the Netherlands, the average number of citizens per zip code area is 7.607, so compared to the rest of the country, some of the zip code areas which fall into the second lowest category are still densely populated (CBS, 2018).
The surrounding address density is visualized in the bottom left map of figure 7. The surrounding address density is particularly high in the centre of Amsterdam and the region West of the centre, so the habitants of these areas are more exposed to urban heat (CBS, 2020). Based on the surrounding address density, the CBS has created a national index for the level of urbanisation of an area, as shown in table 5 (CBS, n.d.). However, sixty from the seventy-three zip code areas of Amsterdam fall into category one (CBS, 2020). Only the zip code areas which are marked in the lightest shade in the bottom left map of figure 7, and zip code area 1106, fall into the other categories. This suggest that even the areas which are not marked as the highest level of surrounding address density in figure 7, have a high level of urbanization, with an
Figure 6. Accessibility green space in Amsterdam (Bezemer & Visschedijk, 2003).
Agriculture Cemetery Forest
Daytime recreational object Dry natural terrain Wet natural terrain Park and plantation Sports field Allotment Residential area Access to green space Built-up area boundary
21 increased exposure to heat. Furthermore, it shows that the situation is even more pressing in the darker-tinted areas than at first glance.
Table 5. Level of urbanization
Level Surrounding address density per km2
1. Very highly urban 2.500 or more
2. Highly urban 1.500 – 2.500
3. Moderately urban 1.000 – 1.500
4. Little urban 500 – 1.000
5. Non-urban Less than 500
Figure 7. Children 0-14 and spatial factors
Population density
Green space Building density Children 0-14
22 The blue circles in figure 7 indicate the number of children from the age of zero to fourteen who are living in the zip code areas. Remarkably, fewer children live in the city centre and South part of the city. Higher concentrations of children can be found in New-West and South-East. The average number of children per zip code area in Amsterdam is 1.713, but with a standard deviation of 1.085 (CBS, 2020). This high standard deviation indicates substantial differences between the zip code areas. The zip code areas 1069 (Osdorp) and 1087 (IJburg) have by far the largest quantities of children, respectively 5.235 and 5.015. IJburg consist of four artificial islands which are located in the East part of the city (Gemeente Amsterdam, n.d. a). This newly developed neighbourhood is popular by young families and has the highest percentage of children in Amsterdam; almost one third of its population is younger than eighteen years (ibid.). Furthermore, Osdorp also has the largest quantity of inhabitants in Amsterdam and IJburg is also part of the top ten most populated zip code areas, which could be linked to the high number of children.
The relative small group of children in the city centre is a positive finding, as this area has a small amount of green space, which would make the children more exposed to heat. Nonetheless, there also lives a larger group of children in the West area around the neighbourhood the Baarsjes, where the heat exposure is quite high, with few green space, high population density, and a relative high surrounding address density. The distribution of children is less optimal in relation to the population density (upper-right map), but this is not surprising as this research uses exact numbers and not percentages. Considering the surrounding address density, the distribution of children makes them less exposed to heat, as they tend to live in the less built-up areas of Amsterdam.
23
Figure 8. Elderly above 65 years old and spatial factors
Figure 8 shows the distribution of people with an age above sixty-five years, which is rather even. Compared to the distribution of children, there are less extreme values in the data, whit the highest number is 3.345 elderly people in Osdorp. On average, 1.519 elderly people live in a zip code area in Amsterdam, with a standard deviation of 850 people. On a national scale, more elderly people live in the rural, East part of the country and less in the larger cities (CBS, n.d.). 19,5 percentage of the Dutch population is older than 65, but in Amsterdam only 12,7 percentage of the inhabitants are above the age of 65 (ibid.). However, there is often little green space in the zip code areas with more elderly people, which increases their heat exposure. As elderly people are often less able to travel, it is more important to have public services, such as green space, nearby (Böcker, Van Amen, & Helbich, 2017).
Population density
Building density Green space
24
Figure 9. Overweight and spatial factors
The red circles in figure 9 indicate the percentage of the population with overweight or obesity. Overweight is defined as a Body Mass Index (BMI) of above 25 kg/m2, which is based on self-reported weight and length (Volksgezondheidenzorg.info, 2021). The data about overweight was acquired per city district instead of zip code areas, therefore clusters of equal percentages can be seen. Notably, weight issues are more common in the city centre and Amsterdam-North, compared to the outer districts New-West and South-East. Moreover, also in the areas Watergraafsmeer, Stlotervaart, and the Pijp/Rivierenbuurt more than half of the inhabitants are dealing with overweight. However, on a national level, the Amsterdam region has the lowest percentage of adults with overweight: 42,5% (Volksgezondheidenzorg.info, 2021; CBS, 2020). In 2016, on average 48,9% of the population was overweight in the Netherlands, which has risen to 50,0% in 2020 (Volksgezondheidenzorg.info, 2021). These statistics indicate that this is a national issue, which should also be researched in other parts of the country. The top left map of figure 9 shows that the heat exposure due to a lack of green space to the people with
Population density
Building density Green space
25 overweight in the northern part of the city is relatively low. However, the opposite is true for the inhabitants with overweight in the city centre and West. Furthermore, there is not a large group of people with overweight in the areas which have the highest population density, such as Osdrop and Bijlmer Centrum. However, as this indicator is expressed in percentages, the relative number of overweight people still might be high in these zip code areas. The heat exposure of people with overweight in the city centre is enlarged due to the high surrounding address density of this area.
Figure 10. Single person households and spatial factors
Figure 10 shows the total number of single person households per zip code area. Amsterdam has a relatively high percentages of single person households, with a small average household size of 2,05 persons (CBS, 2020). The average household size in the Netherlands has rapidly decreased over the past decades; from 3,56 persons in 1960 to 2,14 persons in 2020 (CBS,
Population density Building density Green space Single person households
26 2021b). This trend is likely to continue in the coming decades, making this factor even more valuable to research.
In the data, the zip code area 1105 Bijlmer Centrum has the greatest quantity of 9.095 single person households. The second highest ranked area has 1.350 less single person households; 7.745 in Centre-East. Furthermore, large numbers of single person households can be found in the West area of the centre. This coincides with a high level of the surrounding address density and a low amount of green space, resulting in a high level of heat vulnerability in this area. The higher numbers of single person households often fall in the more densely populated areas. However, this could be expected as they are both expressed in exact numbers.
In the previous paragraphs, the distribution of people who are more sensitive to heat and spatial factors which increase heat exposure have been discussed. Local temperature was not used as an indicator in this research, due to practical issues, which is elaborated in the discussion. Nonetheless, this is a clear indicator where heat exposure is a pressing issue (Van der Hoeven & Wandl, 2015), therefore the map of Van der Hoeven and Wandl (2015) is used. Figure 11 shows the difference in temperature between the Waterland area outside the city (North-East) and the other parts of the city during the 2006 heatwave, which can be up to 20 degrees Celsius. Remarkably, strong heat islands are found in the areas which were excluded from this research because of their low population density, such as Westport and Amstel III. Furthermore, in Zuid, West and North some heat clusters can be found.
Figure 11. UHI land surface temperature, during day 16-07-2006 heatwave at 12:30 CET (Van der Hoeven & Wandl, 2015)
27
Discussion
Methodological reflection
Data availability
The accessibility of GIS data was a large impediment to this project. Firstly, some data is too personal to be published on a small spatial scale, such as diseases, education level or religion. This data may stigmatise the zip code area or even bring its inhabitants in danger. Secondly, some data is difficult to acquire from a practical perspective, for instance the sky-view factor or local surface temperatures. The sky view factor represents how much the earth’s surface is exposed to the sky vault, so a high sky view level makes it easier for the surface to cooldown. To calculate the sky-view factor, Van der Hoeven and Wandl (2015) created a 3D model of Amsterdam, however the researcher of this study does not have the skillset to execute this. Furthermore, to measure local temperatures, multiple weather stations should be placed throughout the city, as the national weather institute only has data on a larger, regional scale, which is beyond the scope of this research (KNMI, 2020).
Subsequently, some issues occurred with the available data. In the CBS data, a lot of cells had the value of -99997. This code can imply multiple complications; the data is classified, the data is missing, or the cell has a value of 0 to 4 (Van Leeuwen, 2019). To deal with this matter, the zip code areas with less than 150 inhabitants were excluded from this research (see figure 5), as these had the most -99997 cells and are less significant for this research. However, despite the small quantity of inhabitants, these areas are important work-hubs; Amstel III/ArenApoort provides jobs to 50.000 people, and Westpoort is an international port and one of the largest business parks in the Netherlands (Gemeente Amsterdam, 2021; Gemeente Amsterdam Zuidoost, 2018). Besides, figure 11 shows that these areas are relatively hot places in Amsterdam, making further research in these areas on the relationship between urban heat and work circumstances valuable (Van der Hoeven and Wandl, 2015).
28
Figure 12. Harbour Westpoort (Dolman, 2017)
Furthermore, an unreliable aspect of this research is the differences in the size of the zip code areas, which makes the data less suitable to compare. The CBS data was also available on a 500x500 meter grid, however this file contained so many fault codes that it was unusable for this study. Besides, the other data sources which were used in this research were only available on a larger scale. Therefore, the zip code area file was deemed more practical and a more commonly used level in policy making, so the results of this paper are more suitable for application with this scale. It was considered to create percentages of the indicators (e.g. the number of children from the total number of inhabitants of a zip code area), which would make the comparison between different zip code areas more representative. Nonetheless, it was considered more important to indicate the areas where the quantity was the greatest for this research, as an intervention is the most required there.
Manual editing
As shown in figure 4, some of the data had to be manually edited, however this practice could be more susceptible to mistakes. This is mainly the case with the Green PC4 file; green spaces sometimes fell into multiple zip code areas and had to be manually divided between these areas. The data contained 365 green space polygons, from which 72 required manual editing, or were deleted when they fell in the Westport/Bullewijk region. Furthermore, zip code area 1043 also has been manually edited, as it contained the code -99997 for its number of inhabitants between
29 0-14 years old and its inhabitants of 65 years and older. There was no information available online about these inhabitants, therefore the numbers were based on the total number of inhabitants (725), minus the sum of the other age categories (720). The fields were both assigned the quantity 3.
Overweight data
Another uncertainty in this research is inherent to the GGD excel data about obese and overweight people. This data was retrieved from the Amsterdamse Gezondheidsmonitor, (Amsterdam Health Monitor) which is conducted through a questionnaire (GGD Amsterdam, n.d.). This method is less reliable compared to, for example, using confirmed medical information, however this is not possible as this personal data is classified. Furthermore, the GGD data was only available from the year 2016, which still gives a clear indication of the dispersal, however BMI can easily change in five years.
Multiple attributes tool
The initial idea for this thesis was to create one map, which would show the combined results of the different indicators. To do so, the Multiple attributes tool in ArcGIS would be used, however the data of this research was not significant enough to create a clear map with this method. Therefore, the decision was made to create multiple maps, so readers would experience less difficulty in understanding the data and findings.
Theoretical reflection
Operationalisation
In this paper, heat sensitivity was measured through age (young/old), weight (overweight/not overweight) and household size (single person household/larger household). When using such broad groups as indicators, prudence for generalisation is needed, because not all elderly people have the same amount of sensitivity to heat or every person who lives alone. By choosing indicators which have more researched and overarching symptoms (e.g. table 3), generalisation is less likely to happen. However, this can result in less significant data, as these are often smaller groups and less data is available. Nevertheless, further research on these indicators in Amsterdam could be enriching.
30 To measure heat exposure, building density, population density, and green space were used as indicators in this study. In the research of Van der Hoeven and Wandl (2015), sky-view factor, albedo, imperviousness, and local surface temperature were added to this list, which gave even more insights, but were beyond the scope of this thesis. Further research could combine the findings of both these papers in order to distinguish more vulnerable locations in Amsterdam. Furthermore, the role of water could be an interesting addition to the spatial factors. As this is a complex topic and could form a stand-alone research, it was not included in this thesis. The role of water is complicated as it provides cooling, as well as heating (Steeneveld et al., 2014). Because of the density of water, it is slower to heat-up and cool-down compared to, for instance, build structures or green space (ibid.). Resulting in a cooling effect in the morning, but a heating effect during the night, and a cooling effect in spring, but heating effect in autumn (ibid.). Figure 13 shows the difference between day and night temperatures in the Amsterdam region, which are the largest at the water body Ijsselmeer. Moreover, water can be used by people to cool-down through swimming, but this depends on the water quality and is not a possibility for everyone, such as disabled people.
Figure 13. Air temperature difference night and day (18/19-07-2006) (Van der Hoeven & Wandl, 2015).
Furthermore, in a more extensive research about future scenarios for Amsterdam, it could be interesting to incorporate adaptation capacity in the theoretical framework. The Royal Dutch Meteorological Institute predicts an average daytime temperature of 42 degrees Celsius during
31 heatwaves in Amsterdam in their warm/high climate scenario for the year 2085 (Amsterdam Institute for Advanced Metropolitan Solutions, 2020). When Amsterdam’s climate reaches a point where the heat becomes almost unbearable, it is increasingly important to look at the differences in adaptation opportunities between people, for example who has resources to afford a summer home outside of the city, or to instal an air conditioning system.
Fasting and Ramadan
In the operationalisation, multiple factors which influence heat sensitivity are listed, such as diabetes insipidus or anorexia nervosa. This list can be extended with (intermittent) fasting, as fasting can lead to dehydration. Research has shown that the effects of dehydration are more profound in hot than in cold circumstances (Kolasa, Lackey, & Grandjean, 2009). Nevertheless, this indicator was not included in this research, as fasting is often a temporary condition and there is very little data available about this topic. However, an interesting group in this category are Muslims who participate in the Ramadan. During Ramadan, they do not eat or drink between sunrise and sundown and it takes place during the ninth month of the Islamic calendar, therefore it can take place during the winter but also in the middle of the summer (Manjunath, Aravindhakshan, & Varghese, 2019). When this second scenario takes place, research has shown that Ramadan participants are more vulnerable to heat illnesses (ibid.). As there is no data about the distribution of practicing Muslims in Amsterdam, multiple alternative operationalizations of this indicator have been considered. For instance, the distribution of specific nationalities, as Muslims who live in the Netherlands often originate from Morocco or Turkey (De Koning, 2020). Figure 14, shows the distribution of mosques in Amsterdam, which could give a notion of the distribution of Muslims (Gemeente Amsterdam, n.d. b). However, both these indicators were deemed too generalizing, so future research of this topic is needed.
32
Figure 14. Distribution of Mosques in Amsterdam (Gemeente Amsterdam, n.d. b).
Individual/local differences
A final point to consider when reading this paper are the local and individual differences in the city. In this paper generalization is used for clarity and efficiency, but it is important to keep in mind that experiences differ from person to person. It is fundamental to recognize the diversity of affected people and experiences for environmental justice (Schlosberg, 2004). Therefore, future research and policy developments should involve local participants in their work and collaborate with them to counter this.
33
Conclusion
The research question of this paper was: Which factors explain differences in heat vulnerability between neighbourhoods in Amsterdam?. The results from the ArcGIS research show that heat vulnerability is especially high in the city centre and West; where little green space is present, the building density is high, many elderly people and people with overweight live, and single person households are abundant. The research question was divided into three sub questions, which are discussed in the following paragraphs.
The first sub question, Which spatial factors make people more exposed to urban heat?, is elaborated upon in the operationalization. Green space, population density, and building density where found to be the most influential on urban heat exposure. Larger numbers of building and population density increase heat exposure, whereas a lack of green space results in more urban heat.
The second sub question of this thesis was: Which individual factors make people more sensitive to urban heat?. After the literature review, the following four indicators were selected for this research; young children (age < 14), elderly people (age > 65), overweight/obese people, and single person households.
Where are these factors present in Amsterdam?, was the final sub question stated in the introduction. The average amount of green space in Amsterdam is relatively high, however, due to an uneven distribution, there is a lack of green space in the centre and West areas of the city. The population is fairly evenly spread over the city, but the highest population densities can be found in the suburbs. The building density, measured in the surrounding address density, is the highest in the city centre and West, making these areas more exposed to urban heat. However, compared to the rest of the Netherlands, almost all parts of Amsterdam are densely populated and have a high surrounding address density. To conclude, spatial characteristics which increase heat exposure are mainly present in the centre and West.
Many children live in the outskirts of the city, which have relative low levels of heat exposure. However, there is also a large group of children living in the West part of Amsterdam. Elderly people live quite evenly dispersed over Amsterdam; the statistics show a small standard deviation. However, they do often live relatively far away from green space and are less ambulant, which increases their heat exposure. Higher percentages of people with overweight
34 can be distinguished in Amsterdam North and centre. Especially in the centre, people with overweight are more exposed to heat, as well as the zip code areas around the centre. Amsterdam has a large quantity of single person households on a national scale, in the city itself, high numbers of single person households can be found in the centre, West and Bijlmer Centre. The individual indicators related to age did not result in evident problem areas, children are even often living in areas with more green space and a low surrounding address density and elderly people live dispersed over the city. So, people with an increased sensitivity to heat are present in the centre, North and West of Amsterdam.
In conclusion, heat vulnerability is a large problem in the centre and West of Amsterdam, where there is a lack of green space and a high surrounding address density, and many people who are vulnerable to heat live, such as elderly people, people with overweight and people who live alone.
The results of this research indicate a need for measures against urban heat in the centre and West of Amsterdam. Therefore, it is strongly advised to the municipality to assess the needs of the residents of these regions and create a local heat adaptation strategy. Measures against urban heat include; trees and green space, green/cool rooftops, cooler and less pavement, surface water and energy-efficient buildings (Van der Hoeven & Wandl, 2013). Furthermore, the previously explained results confirm the importance of local variations in policies against UHI effect, which might also be the case in different urban policies.
The findings of this research add to the existing literature about heat vulnerability by using a new operationalisation. The theoretical framework of this thesis can be used for cities with similar characteristics and climatological circumstances to Amsterdam, for example other cities in the Netherlands and West-Europe. Furthermore, new data about the case study Amsterdam is acquired in this research.
35
Bibliography
Adger, W. N. (2006). Vulnerability. Global environmental change, 16(3), 268-281.
Ambrose, K. (2019, July 25). Photos: Tourists and residents of Amsterdam try to stay cool in
record heat. Retrieved from
https://www.washingtonpost.com/weather/2019/07/25/photos-tourists-residents-amsterdam-try-stay-cool-record-heat/
AMS. (2020, September 15). 100 years of Dutch summers: The clock is ticking on climate change. Retrieved from https://www.ams-institute.org/news/100-years-of-dutch-summers-the-clock-is-ticking-on-climate-change/
Amsterdam Institute for Advanced Metropolitan Solutions. (2020, February 20). City of Amsterdam launches Climate Adaptation Strategy. Retrieved from https://www.ams-institute.org/news/city-amsterdam-launches-climate-adaptation-strategy/
Bao, J., Li, X., & Yu, C. (2015). The construction and validation of the heat vulnerability index, a review. International journal of environmental research and public health, 12(7), 7220-7234.
Bezemer, V., & Visschedijk, P. A. M. (2003). Groene meters deel II: Analyse van het stedelijk groen in de G30 steden (No. 833). Wageningen: Alterra.
Böcker, L., Van Amen, P., & Helbich, M. (2017). Elderly travel frequencies and transport mode choices in Greater Rotterdam, the Netherlands. Transportation, 44(4), 831-852.
Brandow, A. M., Stucky, C. L., Hillery, C. A., Hoffmann, R. G., & Panepinto, J. A. (2013). Patients with sickle cell disease have increased sensitivity to cold and heat. American journal of hematology, 88(1), 37-43.
Bytomski, J. R., & Squire, D. L. (2003). Heat illness in children. Current sports medicine reports, 2(6), 320-324.
36 Carter, S., Field, E., Oppermann, E., & Brearley, M. (2020).The impact of perceived heat stress symptoms on work-related tasks and social factors: A cross-sectional survey of Australia's Monsoonal North. Applied Ergonomics, 82(102918). https://doi.org/10.1016/j.apergo.2019.102918
CBS. (2018, January 6). Bevolking en huishoudens; 4-cijferige postcode 1-1-2018. Retrieved from https://www.cbs.nl/nl-nl/maatwerk/2018/49/bevolking-en-huishoudens-4-cijferige-postcode-1-1-2018
CBS. (2020). Numeriek deel van de postcode (PC4) 2020-V1 [Data file]. Retrieved from https://www.cbs.nl/nl-nl/dossier/nederland-regionaal/geografische-data/gegevens-per-postcode
CBS. (2021a). Bevolkingsontwikkeling; regio per maand. Retrieved from https://opendata.cbs.nl/#/CBS/nl/dataset/37230ned/table?ts=1621514563336
CBS. (2021b). Huishoudens nu. Retrieved from https://www.cbs.nl/nl-
nl/visualisaties/dashboard-bevolking/woonsituatie/huishoudens-nu#:~:text=Hoeveel%20mensen%20wonen%20alleen%3F,alleenstaande%20ouder%2 0in%20een%20eenoudergezin.
CBS. (n.d.). Stedelijkheid. Retrieved on 16-05-2021, from https://www.cbs.nl/nl-
nl/nieuws/2019/44/meeste-afval-per-inwoner-in-minst-stedelijke-gemeenten/stedelijkheid
Cheung, S. S., McLellan, T. M., & Tenaglia, S. (2000). The thermophysiology of uncompensable heat stress. Sports Medicine, 29(5), 329-359.
Couzy, M. (2021, January 5). CBS: Amsterdamse bevolking groeide nauwelijks vanwege corona. Het Parool. Retrieved from https://www.parool.nl/nederland/cbs-amsterdamse-bevolking-groeide-nauwelijks-vanwege-corona~bdccff66/
37 Davis, R. E., Knappenberger, P. C., Michaeals, P. J., & Novicoff, W. M. (2003). Changing
heat-related mortality in the United States. Environmental Health Perspectives. https://doi.org/10.1289/ehp.6336
De Koning, M. (2020). From Turks and renegades to citizens and radicals: The historical trajectories of ‘good’ and ‘bad’ Muslims in the Netherlands. Trajecta, 29(1), 3-26. De Sario, M., Katsoujanni, K., & Michelozzi, P. (2013). Climate change, extreme weather
events, air pollution and respiratory health in Europe. European Respiratory Journal, 42(3) 826-843. DOI: 10.1183/09031936.00074712
Doman, N. (2017, July 12). Klimaatstresstest Waterbestendig Westpoort. Retrieved from https://www.rainproof.nl/klimaatstresstest-waterbestendig-westpoort
Epstein, Y., & Moran, D. S. (2006). Thermal comfort and the heat stress indices. Industrial health, 44(3), 388-398.
esri. (n.d.). Data classification methods. Retrieved on 30-04-2021, from
https://pro.arcgis.com/en/pro-app/latest/help/mapping/layer-properties/data-classification-methods.htm
Flensner, G., Ek, A. C., Söderhamn, O., & Landtblom, A. M. (2011). Sensitivity to heat in MS patients: a factor strongly influencing symptomology-an explorative survey. BMC neurology, 11(1), 1-8.
Gemeente Amsterdam - Onderzoek, Informatie en Statistiek. (2019). Gebieden22 [Data file]. Retrieved from https://maps.amsterdam.nl/open_geodata/?k=204
Gemeente Amsterdam - Ruimte en Duurzaamheid. (2021). Hoofdgroenstructuur [Data file]. Retrieved from https://maps.amsterdam.nl/open_geodata/?k=55
Gemeente Amsterdam Zuidoost. (2018). Gebiedsplan 2017 Amstel III/ArenApoort. Retrieved
38 https://daf9627eib4jq.cloudfront.net/app/uploads/2018/06/160574_arenapoort_en_ams tel_iii.pdf
Gemeente Amsterdam. (2020, Februari 4). Strategie Klimaatadaptatie Amsterdam. Retrieved from https://www.amsterdam.nl/bestuur-organisatie/volg-beleid/coalitieakkoord-uitvoeringsagenda/gezonde-duurzame-stad/samen-slag-klimaatbestendig-amsterdam/ Gemeente Amsterdam. (2021, March 31). Westpoort Werkt: werk en stage om de hoek.
Retrieved from https://www.amsterdam.nl/westpoort/nieuws-westpoort/westpoort-werkt-werk-stage-hoek/
Gemeente Amsterdam. (n.d. a). Districts and neighbourhoods. Retrieved from https://www.amsterdam.nl/en/districts/
Gemeente Amsterdam. (n.d. a). Gebiedsplan IJburg en Zeeburgereiland 2019. Retrieved on
17-05-2021, from
https://www.amsterdam.nl/bestuur- organisatie/stadsdelen/gebiedsgericht-werken/gebiedsplannen-2019/gebiedsplannen-oost/ijburg/
Gemeente Amsterdam. (n.d. b). Locations for religious gatherings. Retrieved on 10-05-2021, from https://maps.amsterdam.nl/religie/?LANG=en
Gendler, I. (2017). The Heat Index. Retrieved from
https://isaacscienceblog.com/2017/11/18/the-heat-index/
GGD Amsterdam. (2016). Open Data Volwassenen: Amsterdamse Gezondheidsmonitor (AGM) 19 jaar en ouder / 2016 / Gebied (22) [Data file]. Retrieved from https://amsterdam.ggdgezondheidinbeeld.nl/openapi/agm/2016/region2
GGD Amsterdam. (n.d.). Amsterdamse Gezondheidsmonitor. Retrieved on 10-05-2021, from https://www.ggd.amsterdam.nl/beleid-onderzoek/gezondheidsmonitors/amsterdamse-0/
39 GGD Leefomgeving. (2021). Hitte en gezondheid. Retrieved from
https://ggdleefomgeving.nl/omgeving/hitte/hitte-en-gezondheid/
Grachten van Smaragd. (n.d.) Waarom. Retrieved on 18-03-2021, from https://www.grachtenvansmaragd.nl/waarom/
Hagens, W. I. & Van Bruggen, M. (2014). Nationaal Hitteplan: versie 2015. Bilthoven: Rijksinstituut voor Volksgezondheid en Milieu.
Hamstead, Z. A. (2021). How We Got Here: Producing Climate Inequity and Vulnerability to Urban Weather Extremes. In: Hamstead Z.A., Iwaniec D.M., McPhearson T., Berbés-Blázquez M., Cook E.M., Muñoz-Erickson T.A. (Eds.) Resilient Urban Futures: The Urban Book Series (p. 11-28). Cham: Springer. https://doi.org/10.1007/978-3-030-63131-4_2
Harlan, S. L., Brazel, A. J., Prashad, L., Stefanov, W. L., & Larsen, L. (2006). Neighborhood microclimates and vulnerability to heat stress. Social science & medicine, 63(11), 2847-2863.
Heaviside, C., Macintyre, H. & Vardoulakis, S. (2017). The Urban Heat Island: Implications for Health in a Changing Environment. Current Environmental Health Reports, 4, 296– 305. https://doi.org/10.1007/s40572-017-0150-3
Jay, O., & Kenny, G. P. (2010). Heat exposure in the Canadian workplace. American journal of industrial medicine, 53(8), 842-853.
Johnson, D. P., Stanforth, A., Lulla, V., & Luber, G. (2012). Developing an applied extreme heat vulnerability index utilizing socioeconomic and environmental data. Applied Geography, 35(1-2), 23-31.
Kleerekoper, L., Van Esch, M., Baldiri Salcedo, T. (2012). How to make a city climate-proof, addressing the urban heat island effect. Resources, Conservation and Recycling, 64, 30-38. https://doi.org/10.1016/j.resconrec.2011.06.004
40 KNMI. (2018). Ons klimaat verandert. Retrieved from
http://klimaatverandering-mra.vormgeving.com/
KNMI. (n.d.) Klimaatviewer. Retrieved on 13-05-2021, from https://www.knmi.nl/klimaat- viewer/kaarten/vocht/gemiddelde-relatieve-vochtigheid-12UT/zomer/Periode_1991-2020
Kolasa, K., Lackey, C., & Grandjean, A. (2009). Hydration and health promotion. Nutrition Today, 44(5), 190–201. http://dx.doi. org/10.1097/NT.0b013e3181b9c970
Kovats, R. S., & Hajat, S. (2008). Heat stress and public health: a critical review. Annual Review Public Health, 29, 41-55.
Kuras, E. R., Richardson, M. B., Calkins, M. M., Ebi, K. L., Hess, J. J., Kintziger, K. W., ... & Hondula, D. M. (2017). Opportunities and challenges for personal heat exposure research. Environmental health perspectives, 125(8).
Lindley, S., O’Neill, J., Kandeh, J., Lawson, N., Christian, R., & O’Neill, M. (2011). Climate change, justice and vulnerability. York: Joseph Rowntree Foundation.
Loughnan, M, Nicholls, N, Tapper, NJ. (2012). Mapping heat health risks in urban areas. International Journal of Population Research, 2012. DOI:10.1155/2012/518687 Lundgren, K., Kuklane, K., Gao, C., & Holmer, I. (2013). Effects of heat stress on working
populations when facing climate change. Industrial health, 51(1), 3-15.
Manjunath, G., Aravindhakshan, R., & Varghese, S. (2019). Effect of fasting during Ramadan on thermal stress parameters. East Mediterranean Health Journal, 25(1), 34-39.
Meehl, G. A., & Tebaldi, C. (2004). More intense, more frequent, and longer lasting heat waves in the 21st century. Science, 305(5686), 994-997.
NKWK. (2019). Steeds meer hitte records. Retrieved from
https://waterenklimaat.nl/nl/onderzoekslijnen/klimaatbestendige-stad/kenniskrant-voor-een-klimaatbestendige-stad/heet-nieuws/hitte/
