Climate Proof Cities - Final Report

Abstract

Policy summary
All cities in the Netherlands, large and small, are vulnerable to the effects of climate change. The degree of vulnerability varies considerably within urban areas. This means that making cities more climate proof can be done most efficiently by taking many relatively small and local measures. Many of these can be carried out simultaneously with major repairs or renovations. This does require collaboration with many and various parties.

These are the most important findings of the Climate Proof Cities (CPC) research programme. This programme has yielded much insight in making Dutch cities climate proof, with a focus on heat stress and flooding due to heavy rainfall. The programme was carried out by a consortium of ten universities and knowledge institutes that worked together for four years with municipalities, water boards and the national government to provide answers to knowledge questions from practice.

Full text

Climate Proof Cities
Final report

Eindrapport Climate Proof Cities
1
Final report Climate Proof Cities 2010-2014
Date: October 2014
Climate Proof Cities consortium
1
Final editing: Vera Rovers, Peter Bosch, Ronald Albers (TNO)
Translation: Hannah Kousbroek Text & Translation, Amsterdam
KfC report nr: 129/2014
Climate Proof Cities was conducted as part of the Dutch research programme ‘Knowledge for Climate’,
co-financed by the Dutch ministry of Infrastructure and Environment.
1
http://www.knowledgeforclimate.nl/climateproofcities

Eindrapport Climate Proof Cities
2
Colophon
Many researchers (see Appendix A) contributed to this report (and the research programme). The editing
of the content-based chapters was done by the work package leaders:
 Chapter 1 Bert van Hove (Wageningen University)
 Chapter 2 Bert Blocken (Eindhoven University of Technology)
 Chapter 3 Andy van den Dobbelsteen (Delft University of Technology)
 Chapter 4 Tejo Spit, Marjolein Dikmans (Utrecht University)
 Chapter 5 Peter Bosch (TNO)
The following people are responsible for the content of the chapters (texts and research results):
Hoofdstuk
Naam
Instituut
1
Bert van Hove
Cor Jacobs
Bert Heusinkveld
Oscar Hartogensis
Reinder Ronda
Reinder Brolsma
Lisette Klok
Patrick Schrijvers
Ronald Hutjes
Wageningen University
WUR-Alterra
Wageningen University
Wageningen University
Wageningen University
Deltares
TNO
TU Delftt
WUR-Alterra
2
Bert Blocken
Hein Daanen
Twan van Hooff
Karin Stone
Frank van der Hoeven
Alexander Wandl
Jan Hensen
Frans van de Ven
William Veerbeek
TU Eindhoven
TNO
TU Eindhoven
Deltares
TU Delft
TU Delft
TU Delft
Deltares
Unesco-IHE
3
Andy van den Dobbelsteen
Hamid Montazeri
Twan van Hooff
Harry Timmermans
Wiebke Klemm
Toine Vergroesen
Reinder Brolsma
Laura Kleerekoper
Leyre Echevarría Icaza
Jan Hofman
Chris Zevenbergen
Herbert ter Maat
TU Delft
TU Eindhoven
TU Eindhoven
TU Eindhoven
Wageningen University
Deltares
Deltares
TU Delft
TU Delft
KWR
Unesco-IHE
WUR-Alterra
4
Tejo Spit
Caroline Uittenbroek
Anita Kokx †
Liz Root
Beitske Boonstra
Martin Roders
Marjolein Dikmans
Universiteit Utrecht
Universiteit van Amsterdam/Universiteit Utrecht
Universiteit Utrecht
Radboud Universiteit
Universiteit Utrecht/TNO
TU Delft
Universiteit Utrecht
5
Peter Bosch
Annemarie Groot
Cor Jacobs
TNO
Wageningen University
Wageningen University

Eindrapport Climate Proof Cities
3
Policy summary
All cities in the Netherlands, large and small, are vulnerable to the effects of climate change. The degree
of vulnerability varies considerably within urban areas. This means that making cities more climate proof
can be done most efficiently by taking many relatively small and local measures. Many of these can be
carried out simultaneously with major repairs or renovations. This does require collaboration with many
and various parties.
These are the most important findings of the Climate Proof Cities (CPC) research programme. This
programme has yielded much insight in making Dutch cities climate proof, with a focus on heat stress and
flooding due to heavy rainfall. The programme was carried out by a consortium of ten universities and
knowledge institutes that worked together for four years with municipalities, water boards and the
national government to provide answers to knowledge questions from practice.
The urban climate is changing
Climate change leads to more heat waves, more frequent heavy rainfall events, and more periods of
drought. If cities do not prepare for this, it will influence people’s health, quality of life in city districts,
comfort in houses and buildings, and productivity, leading to economic problems.
The high percentage of paved area in the city, combined with the increasing chances of heavy rainfall, can
lead to greater material and financial damage through traffic disruptions, problems with infrastructure
and the expense of calling in emergency services. The thresholds for flooding in the urban environment
have stayed the same or even decreased in recent years, and flooding is a recurring problem in some
districts. More summery and tropical days are also expected in the future. Without an explosive increase
in air conditioning in buildings, this will lead to much higher temperatures in a vast proportion of Dutch
housing. Heat stress can lead to illness and increased mortality among sensitive sections of the
population, such as the elderly and the chronically ill, but also to decreased productivity and sleeping
disorders.
Both large and small cities are vulnerable
During heatwaves, it is warmer in every city in the Netherlands, large or small, than it is in the
surrounding area. This heat island effect is clearly noticeable and can reach a difference of more than 7
˚C, especially in the evening. Because of climate change, the number of days with heat stress in the city
can increase substantially. Heavy rainfall can also hit any city.
Vulnerability varies greatly within the city
A striking conclusion of the CPC research is that within the urban area, there is great spatial variation in
vulnerability, depending on the properties of the district and the building and the distribution of sensitive
persons and objects. Exposure to heat and flooding, for instance, is mainly determined by the amount of
paved area and the density of buildings in an area. Overheating in buildings strongly depends on the
presence of sun blinds and degree of insulation. Information about exposure, combined with the
locations of sensitive groups (for instance, the elderly) and of objects (such as switch boxes and houses
with cellars), forms the basis for identifying areas that need attention.
Adapting to climate change is a matter of the combined effects of relatively small, local measures
Because vulnerability to the effects of the climate is determined locally, the choice of measures is also
dependent on the local context. The input of generic measures for a whole city is less effective. A wide
variety of adjustment measures exists, ranging from influencing the urban climate or the urban water
system (for instance, collecting and storing rain water, creating a cooler layout of streets and squares),
adapting buildings and infrastructure (e.g. installing doorsteps), changing human behavior and increasing

Eindrapport Climate Proof Cities
4
acceptance of discomfort and preventing damage when an extreme event does take place (such as care
for the elderly). Various adjustment measures contribute to easing problems with flooding, heat and
drought at the same time, and an integral approach to these three problems is preferred. Rain water
from wetter periods could for instance be stored underground and used to combat dryness, and, through
evaporation, heat. Many measures have a positive effect on other policy themes, such as migration and
biodiversity, and/or contribute to the improvement of the general living conditions in buildings and in
public spaces. The CPC research has provided a number of new and sometimes startling insights about
the effectivity of measures:
 Traditional green roofs, without restricted discharge measures, are hardly effective for both the
indoor climate, the outdoor climate and the temporary storage of extreme rainfall.
 The cooling effect of the surface water in the city is not unequivocal: bodies of water can even
contribute to the warming of the city; large bodies of water, depending on their orientation in terms
of the direction of the wind, can have a cooling effect.
 Insulating buildings without paying attention to protection against sunshine can lead to more heat
problems in hot summers.
Planting deciduous trees with large crowns, and more generally adding green elements in private and
public spaces leads to better thermal comfort and lessens problems caused by extreme rainfall.
Many measures can easily be integrated into other policy, but require interdisciplinary collaboration
Many measures require collaboration between different parties: the departments within a municipality,
water boards, home owners, sometimes businesses. However, integration of climate adaptation in other
sectors is not self-evident. Institutional entrepreneurs can help to connect different goals and ensure
widely supported solutions for urban development and realising cost savings simultaneously. Making
cities climate proof should be an integral part of decision-making for all sorts of parties interested in the
urban environment. Only when authorities, citizens and private parties realise a climate proof city
requires combined effort, will there be a basis for success.
Now is the time to define the areas for special attention and to develop a strategy, and in the
execution, join in with larger renovation and restructuring projects
The climate is changing slowly but steadily. Because investments that are currently being made in the
urban environment, for instance in renovations or new construction projects, will result in buildings and
infrastructures that will still exist in roughly fifty years, it is important to determine already whether
adjustments to a future climate can be made. More and more studies, both international and national,
show that the costs of adjustments made now are limited compared to the damage that can be caused in
one day due to extreme weather conditions.
Because becoming climate proof requires a long-term plan, it is important to clarify already which
measures should be applied in which areas. In policy terms this is known as a climate stress test and a
climate adaptation strategy. The execution can subsequently take place in phases in the next decennia by
joining in with regular maintenance and renovations, so that costs are limited. Identifying these windows
of opportunity for planning and executing adaptation measures gives a time plan for implementation.
Missing opportunities for including adaptation measures during large transformations can lead to greater
costs later.

Eindrapport Climate Proof Cities
5
Inhoudsopgave
Policy summary ............................................................................................................................................... 3
Introduction .................................................................................................................................................... 8
1 How does the local climate work in Dutch cities, and how does urban design influence the local
climate?.........................................................................................................................................................11
Summary ...................................................................................................................................................11
1.1 Introduction ........................................................................................................................................12
1.2 The interaction between the city and the countryside ......................................................................17
1.3 The influence of anthropogenic heat production ...............................................................................19
1.4 Evaporation in the city ........................................................................................................................20
1.5 Climate variations within the city .......................................................................................................23
1.5.1 The variation in temperature .......................................................................................................23
1.5.2 The influence of neighbourhood characteristics on temperature ...............................................24
1.5.3 The variation in thermal comfort .................................................................................................30
1.5.4 The influence of neighbourhood characteristics on thermal comfort.........................................31
Conclusions ...................................................................................................................................................32
2 How vulnerable are Dutch cities to climate change? ...............................................................................34
Summary ...................................................................................................................................................34
2.1 Introduction ........................................................................................................................................35
2.2 Heat stress ..........................................................................................................................................37
2.2.1 Sensitivity .....................................................................................................................................37
2.2.2 the role of buildings .....................................................................................................................39
2.2.3 Vulnerability maps for heat .........................................................................................................41
2.3 Pluvial flooding....................................................................................................................................44
2.3.1 Vulnerability to damage and tresholds ........................................................................................44
2.3.2 Vulnerability maps for flooding ...................................................................................................46
2.4 Tools for policy makers .......................................................................................................................49
2.4.1 3Di area model for flooding .........................................................................................................49
2.4.2 Heat/Drought Stress Model .........................................................................................................50
2.5 Conclusion ...........................................................................................................................................51

Eindrapport Climate Proof Cities
6
3 Which measures can be taken to better adapt cities to climate change? ...............................................52
Summary ...................................................................................................................................................52
3.1 Introduction ........................................................................................................................................53
3.1.1 Categorising climate adaptation ..................................................................................................53
3.1.2 Description of the measures ........................................................................................................54
3.2 Buildings ..............................................................................................................................................56
3.2.1 Goal ..............................................................................................................................................56
3.2.2 Measures .....................................................................................................................................56
3.2.3 Design principles ..........................................................................................................................62
3.3 From street to neighbourhood ...........................................................................................................64
3.3.1 Goal ..............................................................................................................................................64
3.3.2 Measures .....................................................................................................................................64
3.3.3 Design principles ..........................................................................................................................72
3.4 City and region ....................................................................................................................................78
3.4.1 Goal ..............................................................................................................................................78
3.4.2 Measures .....................................................................................................................................78
3.4.3 Design principles ..........................................................................................................................83
3.5 The Linking Method ............................................................................................................................84
3.6 Tools for adaption planning ................................................................................................................86
3.6.1 Use of calculation models for flooding ........................................................................................87
3.6.2 3Di area model for flooding .........................................................................................................88
3.6.3 Climate Adaptation App ...............................................................................................................89
3.6.4 Adaptation Support Tool..............................................................................................................89
3.6.5 Heat and Drought Stress Model...................................................................................................89
3.6.6 Heallth scan ..................................................................................................................................90
4 Urban governance: the implementation of climate adaptation in urban development .........................91
Summary ...................................................................................................................................................91
4.1 Introduction ........................................................................................................................................92
4.2 Municipalities ......................................................................................................................................92
4.2.1 Organisation .................................................................................................................................92
4.2.2 Financial instruments: the use of TIFs .........................................................................................95
4.3 Housing associations ...........................................................................................................................97
4.4 Citizens ................................................................................................................................................98
4.5 Conclusions .......................................................................................................................................100

Eindrapport Climate Proof Cities
7
5 Integration in Bergpolder Zuid ...............................................................................................................101
Summary .................................................................................................................................................101
5.1 Introduction ......................................................................................................................................101
5.2 Integration within CPC ......................................................................................................................101
5.3 Bergpolder Zuid.................................................................................................................................102
5.4 The stakeholders’desires for the study .............................................................................................102
5.5 The research results ..........................................................................................................................103
5.6 Lessons on integration ......................................................................................................................108
5.7 Conclusions and related research questions ....................................................................................110
ANNEXES .....................................................................................................................................................111
Annex A Researchers CPC ...........................................................................................................................112
Annex B bibliography ..................................................................................................................................115
Bijlage C Thermal comfort and indicators...................................................................................................126
Bijlage D Heat exchange between the urban environment and the atmosphere ......................................128

Eindrapport Climate Proof Cities
8
Introduction
The climate is changing
Worldwide climate change leads to more summery and tropical days and more days with extreme rainfall
in summer, among other things. The KNMI’14 climate scenarios
2
indicate what climate change in the
Netherlands most likely entails (KNMI, 2014). According to these scenarios, in 2050, the average
temperature in summer will be 1.0 to 2.3 ˚C higher than in the reference period of 1981-2010. The
number of summery days (with a max temperature of ≥25 ˚C) will increase with 5 to 15 days in view of
an average of 21 now. Heat waves will occur more often. Depending on the scenario it will become dryer
or wetter in summer. The most extreme rainfall events in summer, however, are influenced by local
processes and cannot be predicted with climate models. In all scenarios, warming leads to more water
vapour in the air, which increases the chances of heavy rainfall.
Cities are vulnerable
The concentration of population and economic capital makes cities important centres for a well-
functioning economy and society. At the same time, it makes cities vulnerable to the effects of climate
change. In the Netherlands, 40% of the population lives in the 36 biggest cities and this number is
growing. These cities generate three quarters of the gross national product (G32, 2011). Extreme weather
conditions, therefore, such as heat waves and extreme rainfall, threaten a large number of people, vital
infrastructures and value chains. The combination of urbanization and climate change demands that
cities take a proactive approach towards increasing their resilience in order to guarantee good quality of
life for citizens and to maintain their competitive position.
Cities are dynamic systems that are continuously in development. Adapting to climate change is only one
aspect of this development. Making cities climate proof must therefore be an integral part of decision-
making on the part of all stakeholders in the urban environment.
Action for climate proofing is urgent
The climate is changing slowly but steadily. Because investments that are currently being made in the
urban environment, for instance in renovations or new construction projects, will result in buildings and
infrastructures that will still exist in roughly fifty years, it is important to determine already whether
adjustments to a future climate can be made (EEA 2010). In addition, more and more studies, both
international (Isoard, 2011; Stern, 2006; Watkiss, 2011) and national (Court of Audit, 2012; PBL, 2011),
show that, when compared to the possible damage caused by climate change in the future, the costs of
adjustments made now are low and the advantages are significant. In a number of cases, adjustments
within the built environment are already necessary now in order to decrease inconvenience and damage
caused by current extreme weather conditions.
Climate Proof Cities
In order to adapt existing structures to a changing climate, it is necessary to make decisions based on
well-founded knowledge and to take an integral approach. From 2010 to 2014, the Climate Proof Cities
(CPC) research programme devoted itself to generating this knowledge for climate proof urban policy.
2
Voor onderzoek binnen CPC is nog vooral gebruik gemaakt van de KNMI’06 scenario’s.

Eindrapport Climate Proof Cities
9
At the beginning of the programme, municipalities and water boards outlined the 5 most important
knowledge questions that formed the point of departure for 5 work packages in CPC (Figure 0.1):
1. How does the local climate work in Dutch cities?
2. How vulnerable are Dutch cities to the effects of climate change?
3. Which measures can be taken to better adapt cities to a future climate?
4. How can these measures be implemented in urban areas?
5. What is the final cost-benefit balance of the adaptation measures?
These five questions also form the skeleton of this report. Each chapter answers one of the questions.
The fifth and final chapter dwells especially on the integration of knowledge about adaptation, both
within science and between science and policy.
The research programme has especially paid attention to heat in the city and the increasing risks of
prolonged periods of warm weather, and to inconvenience caused by more frequent and more intensive
rainfall. Water and heat are connected with each other, for instance because prolonged periods of warm
weather can lead to drought and because water can bring down temperatures through evaporative
cooling. In practice, the research was organized into twenty studies, carried out by 9 PhD students, 3
postdocs and many researchers from 10 different research institutes. In order to increase the usefulness
of the results in practice, the researchers worked together in 5 case studies in different Dutch urban
areas, namely Rotterdam, The Hague Region, Amsterdam, Arnhem/Nijmegen, Utrecht and cities in North
Brabant (Figure 0.2).
Figure 0.1 Main structure of the Climate Proof Cities research programme

Eindrapport Climate Proof Cities
10
Figure 0.2 Case studies within the CPC research programme
The Spatial Adaptation Delta Decision
Around the same time as the Knowledge for Climate programme was set up, awareness of climate
change began to grow among Dutch cities. At the beginning of the CPC research programme there were
only a handful of cities (and individuals) who concerned themselves with climate change. There were
barely any connections between municipalities and water boards when it came to this subject. Partially
stimulated by the New Construction and Restructuring Delta Programme, a broad movement has formed
in the past few years, leading to, for instance, the ‘Manifesto for the Climate Proof City’ and the ‘Guide to
Spatial Adaptation’
3
. Both give advice about making cities climate proof, formulated around the three
steps;
1. Knowledge: analysis of the area (links to CPC research questions 1 and 2);
2. Desire: formulating ambitions (research question 3), and;
3. Practice: implementation in policy and regulations (research questions 4 and 5).
On Prinsjesdag 2014 (Prince’s Day, the opening of the Dutch parliament), the Spatial Adaptation Delta
Decision was presented to the House of Representatives; its aim is to make spatial policy more water-
robust and climate proof. This report and all underlying CPC studies
4
offer support in this.
3
http://www.ruimtelijkeadaptatie.nl/en/
4
All CPC publications can be found at http://www.kennisvoorklimaat.nl/publicaties

Eindrapport Climate Proof Cities
11
1 How does the local climate work in Dutch cities, and how does urban
design influence the local climate?
Summary
Climate change leads to more summery and tropical days and more days with extreme rainfall in
summer. The average amount of precipitation in an urban area does not differ from that in the
surrounding countryside. However, this does not apply to temperatures. It is almost always warmer in
the city than in the surrounding areas, which is known as the urban heat island effect (or UHI). The results
of the Climate Proof Cities (CPC) programme offer more insight into this heat island effect.
The urban heat island effect is caused by the absorption of sunlight by (stony) materials, the lack of
evaporation and the emission of heat caused by human activities (‘anthropogenic heat’). The emission of
heat through industry, houses, buildings, traffic, people and animals contributes substantially to the
development of the UHI: in Rotterdam, it reaches around 10%. In the daytime, the difference in
temperature between the city and the countryside are minimal (< 2 ˚C). The differences are especially
great after sunset because the city cools off more slowly than the surrounding areas do. The maximum
UHI intensities in Dutch cities range from 3 to 7 ˚C. With global warming continuing throughout the next
decades heat stress can become an important issue.
Within an urban area, there are substantial spatial variations in UHI. The properties of the direct
surroundings turn out to be of great influence here. The most influential factors are the proportion of
built surfaces, paved surfaces and the proportion of vegetated surfaces. In addition, the average building
height has a clear effect. The ratio of building height to street width also influences the absorption of
sunlight, thermal emissions from buildings and other surfaces into the atmosphere, and the
transportation of heat within the street. The optimal ratio of height to width seems to be around 1.
Higher or lower ratios both have advantages and disadvantages when it comes to ventilation and shade
effects.
The final effect of open water on temperature is not unequivocal and strongly depends on the
dimensions (surface area, depth), the situation in terms of the direction of the wind and the situation in
terms of buildings and other structures in the surroundings.
Thermal comfort for human beings varies even more than the temperature of the surroundings and is
also dependent on atmospheric radiation, humidity and wind velocity. During the day, the thermal
comfort in the city is largely determined by the differences in wind velocity; the differences in humidity
and radiation are too minimal to have a noticeable effect. After sunset, temperature plays a more
important role, and factors that influence the air temperature are important in determining the thermal
comfort.
Furthermore, because of the changes in the climate we will have deal with long, warm and dry periods in
the future. Understanding the city’s water balance is essential in order to plan the urban area in such a
way that cooling through evaporation is secured with as little water consumption as possible.
Evaporation, however, is an unknown quantity. In CPC first estimates have been made of the evaporation
in Rotterdam and Arnhem.

Eindrapport Climate Proof Cities
12
1.1 Introduction
Following on from the ‘Analysis’ step of the ‘Guide to Spatial Adaptation’
5
, it is important to understand
the workings of the urban climate and the interaction between the city and the regional climate. The
average amount of precipitation in an urban area does not differ from that of the surrounding area
6
.
However, this does not apply to the temperature. In the city it is almost always warmer than in the
surrounding area, which is known as the ‘Urban Heat Island’ (UHI) (see text box on ‘The Urban Heat
Island effect’). This means that cities stand a greater chance of experiencing extremely high temperatures
than the rest of the Netherlands. This chapter therefore focuses on the results of climate change on heat
and drought in the city. Understanding the way in which cities themselves influence the urban climate
offers insight into the choice of measures against extreme heat.
The urban climate and climate change
The UHI effect has already been discussed in the international literature for a century. Maximum
temperature differences between cities and surrounding areas measured and calculated in international
studies (Memon et al., 2009) show values of up to 12 °C, where the greatest differences usually occur at
night. In the Netherlands, Conrads (1975) was the first, in the 70s, to research the urban effects for a
Dutch city. Using temperatures measured in Utrecht in summer it turned out that at night, the
temperature in Utrecht is an average of 2.7 °C higher than outside the city, with peaks of up to 8 °C. The
urban effects in Rotterdam were studied a decade later by Roodenburg (1983). Here, too, maximum
temperature differences of 8 °C were found between the city and its surroundings, especially during
windless nights with few clouds.
After this, the research into the urban climate in Dutch cities was at a standstill for almost 30 years. The
thread was eventually resumed in 2009. In the summer of that year, orientation measurements were
taken in Rotterdam
7
and Arnhem
8
with mobile measuring platforms (meteorological measuring
instruments attached to a cargo bicycle). The results of these measurements also show a substantial heat
island effect. After sunset the differences in temperature between densely constructed areas and the
surrounding areas can reach over 7 °C, especially on clear and windless summer days. In the daytime, the
measured differences in temperature are less noticeable, with a maximum of up to 2 °C (Van Hove et al.,
2010; Van Hove et al., 2011c; Heusinkveld et al., 2010, 2014). Since then, this data has been confirmed by
the results of the CPC’s permanent monitoring network in the Rotterdam area (Van Hove et al., 2011a,b).
A detailed interpretation of the measurements can be found in section 1.5.
In order to form a national impression, surface temperatures from satellite images from the 2006 heat
wave were analyzed (Klok et al., 2012). These images show that each city in the Netherlands, large or
small, experiences a heat island effect. (Figure 1.3). It is important to note that this concerns the surface
UHI that is especially present during the daytime. Discussions about the urban climate almost always
refer to the atmospheric UHI of the ‘Urban Canopy Layer’ because of the effect on living conditions (see
also the text box on the ‘Urban Heat Island effect’). The atmospheric UHI is the difference in air
temperature between the city and the nearby countryside. Unlike the surface UHI, the atmospheric UHI is
minimal during the day; a maximum intensity (UHImax) is reached after sunset because the city cools
down more slowly than the nearby countryside.
5
http://www.ruimtelijkeadaptatie.nl/en/
6
It is noticeable that there is more than the average amount of rainfall near large urban agglomerations, such as the greater Rotterdam
area (see De Bosatlast van het Klimaat; www.klimaatatlas.nl). Possible causes are blocking of the wind by buildings, extra warming and the
presence of more cloud condensation nuclei (fine particles that water drops condense on). These factors are conducive to cloud formation
and the development of precipitation on the lee side of cities.
7
This project was carried out as part of the first section of KvC (Heat stress in Rotterdam project)
8
Part of the EU Future Cities project

Eindrapport Climate Proof Cities
13
The Urban Heat Island effect (UHI)
Cities are generally warmer than their surrounding areas. Because of the high volume of buildings and the properties of the
urban material, warmth is retained better in cities and the so-called heat island effect occurs (Figure 1.1). There are three
types of urban heat island effect (UHI):
 The surface UHI, the difference in surface temperature between the city and the surrounding countryside.
 The atmospheric UHI, the difference in air temperature between the city and the surrounding countryside. The
atmospheric UHI can be subdivided into:
o The UHI at the atmospheric boundary layer above the city (Urban Boundary Layer UHI), of which the
intensity depends on the geographical situation of the city, general configuration and morphology.
o The UHI at living level (Urban Canopy Layer UHI), where the presence of buildings, street surfacing, trees
and water have a direct and noticeable effect on the climate at living level (microclimate). Discussions
about the urban climate generally concern this heat island effect.
The surface UHI exists both during the day and after sunset. A maximum is reached in the daytime when the surfaces absorb
sunlight. After sunset, the differences are smaller, but can still be substantial. In contrast, the atmospheric UHI is minimal or
absent during the day. A maximum is reached after sunset because the countryside cools down faster than the city.
Discussions about the urban climate generally concern the atmospheric UHI of the Urban Canopy Layer, because of the effect
on the living environment.
The local climate and microclimate are influenced by processes that take place at city level (the mesoscale) and vice versa
(Figure 1.2). The spatial planning of an urban area, for instance, has an effect on local wind patterns, and the materials used in
buildings in a neighbourhood (such as the use of materials with high sun reflectivity) directly influence not only the indoor
climate, but also the climate in the area surrounding these buildings. In order to develop effective adaptation strategies and
measures it is important to take all levels of scale into account. Therefore, as part of the CPC programme, research was done
on meteorological processes at all levels of scale.
Figure 1.1 The urban heat island effect:
difference in air and surface temperature
between the city and surrounding
countryside in/during the daytime and at
night. Source:
http://www.epa.gov/heatisland/
about/index.htm (last accessed: July 2014).
Figure 1.2 The horizontal city scales and the related
vertical atmospheric layers. The large arrows
indicate the usual direction of the wind; the small
arrows show the turbulent air movements. We
distinguish three horizontal scales and the related
vertical atmospheric layers: 1. Mesoscale (city and
surroundings) and the ‘Urban Boundary Layer’), 2.
Local scale (city district) and ‘Urban Boundary
Layer’/’Urban Canopy Layer’, and 3. Micro scale
(street, block of housing) and ‘Urban Canopy
Layer’). PBL-planetary boundary layer, (Source:
Oke, 1976).

Eindrapport Climate Proof Cities
14
Figure 1.3 The surface heat island effect of Dutch cities during the day (left) and at night (right). The maps are
based on two NOAA-AVHRR satellite images of surface temperature taken during the heat wave period of 2006
(Source: Klok et al., 2012).
Amateur meteorologist databases show that the UHImax values of Dutch cities range from 3 to more than
7 C (Steeneveld et al., 2011;, Wolters en Brandsma, 2012) (Figure 1.4). These values can be compared to
UHI values determined for other European cities. Oke (1973) found a linear relationship between the
UHImax and a city’s population
9
. This relationship does not exist for Dutch cities (Figure 1.5); the UHI can
also be substantial in smaller cities and villages. This shows that local features are highly important for
the UHI intensity.
Figure 1.4 also shows the effect on thermal comfort. The 95th percentile values calculated for thermal
comfort (based on the ‘Approximated Wet Bulb Globe Temperature’ (AWBGT) , see Appendix C) in
densely built urban areas in the Netherlands are now just below the threshold value for heat stress. This
means that thermal discomfort and heat stress can become an important issue if global warming
continues throughout the next decades.
9
He used the logarithmic value of the number of inhabitants

Eindrapport Climate Proof Cities
15
Figure 1.4 Median and percentile values for UHImax and thermal comfort in Dutch cities, based on AWBGT. The
dotted line is the threshold value for thermal discomfort. Rooftop stations are shaded (source: Steeneveld et al.,
2011).
Figure 1.5 UHImax (95th percentile values, in Kelvin) for cities versus the number of inhabitants of cities
(logarithmic scale) for European cities and Dutch cities. Dotted lines are linear regression lines calculated for
Oke’s results (1973), results published between 1987 and 2006 and for Dutch cities (source: Van Hove et al.
2011c).

Eindrapport Climate Proof Cities
16
Themal comfort and other meteorological variables
Thermal comfort is not only dependent on the air temperature, but also on other meteorological
variables such as atmospheric humidity, radiation and wind speed. These parameters were also analyzed
using the measuring network.
The atmospheric humidity in the city is lower than that of the countryside, which is essentially beneficial
for the thermal comfort during warm summer’s days. However, the differences are minimal: less than 5%
for the absolute atmospheric humidity
10
and 9-15% for the relative atmospheric humidity.
The differences in absolute humidity are present especially during the day; due to evaporation from
vegetation the air above the countryside contains more water vapour, while the amount of water vapour
in the urban air stays more or less the same. However, the differences in relative atmospheric humidity
are present both in the day time and at night. During the day, the lower relative atmospheric humidity in
the city is mostly due to minimal evaporation, and at night, it is due to the higher temperature.
The average global solar radiation (i.e. the amount of solar radiation per surface area unit) in the city is
also lower (12-14%) compared to the reference location. This is mainly due to shadows cast by buildings
and other objects (such as trees) near the weather stations. During summery days, a lower amount of
direct solar radiation is beneficial for thermal comfort. This also applies to diffuse radiation, but this was
not measured separately.
The average wind speed measured in cities is considerably lower (40-65%) than in rural areas. Especially
during summery days, the lack of a breeze is detrimental for thermal comfort. This also applies to the air
quality. Both have detrimental health effects on humans and animals.
Climate change and the future urban climate
In order to have an idea of the urgency of the heat problem, current temperature values from the ‘Zuid’
11
weather station in Rotterdam and the reference location were transformed into temperature values for
2050 and 2100
12
. This took place before the KNMI’06 ‘W+’ climate scenario, which can be seen as a
realistic worst case scenario in terms of heat issues. In this scenario, we can expect a substantial increase
in the number of days with lower thermal comfort in both cities and the countryside (Figure 1.6). We
would like to emphasize that this is a first rough result, where only the difference in temperature has
been examined. For a complete analysis, variables such as discussed above also need to be taken into
account.
10
The absolute atmospheric humidity is the amount of water vapour per volume of air. The relative atmospheric humidity is the amount of
water vapour in the air compared to the maximum amount of water vapour the air can contain. As opposed to the absolute atmospheric
humidity, the relative atmospheric humidity depends on the air temperature; air with a higher temperature can contain more water
vapour. It is not yet known exactly if the absolute atmospheric humidity or the relative atmospheric humidity is the determining factor for
thermal comfort. Both variables occur in the thermal indices.
11
Near Zuidplein
12
http://www.knmi.nl/index_en.html

Eindrapport Climate Proof Cities
17
Figure 1.6. The number of days with moderate to strong heat stress (Effective Temperature, see Appendix C)
for the countryside and ‘Rotterdam Zuid’ locations, calculated for the current situation, 2050 and 2100
according to the KNMI’06 W+ scenario.
1.2 The interaction between the city and the countryside
It is known that the city influences the climate of the surrounding countryside. In reverse, the use of the
land in the countryside could have an effect on the climate in the city. However, it is unclear how large
the footprints of both effects are. In order to gain more insight into this, CPC carried out airplane
measurements and used model simulations.
Six flights were carried out above Rotterdam and its surroundings, where air temperature, surface
temperature, atmospheric humidity and CO2 concentration were measured. The measurements offer
insight into not only the horizontal footprint of the UHI effect, but also the vertical footprint. The
measurements took place during the daytime
13
, usually at a height of around 300 metres. In addition,
vertical profiles (300 – 1700 metres) were measured in order to characterize the build-up of the
atmospheric boundary layer (Figure 1.7).
The UHI effect at 300 metres is minimal and difficult to distinguish from the daily range of variables
measured. On warm days the air at 300 metres above the city is around 1 C warmer. Equal differences
were found during the daytime for the air temperature between the weather stations in Rotterdam and
the reference station (to the north of Rotterdam).
The leeward air temperatures (legs 2-4) are higher (0.2 – 1.0 degrees) than the windward air
temperatures (leg 1). The higher air temperatures above the greenhouse area and above the coastline
are also striking (leg 4). In contrast, the surface temperatures show great disparities, for instance
between surface temperature for water and for built surfaces (asphalt on roads or roofs). The difference
can reach up to 40 C. The diffusion of the urban heat measured downwind from the urban areas is also
found in model simulations, the so-called ‘urban plume’ effect (Figure 1.8) (see also Theeuwes et al.,
2013).
It thus seems as though the city’s vertical footprint is limited (around 300 metres), but that the horizontal
footprint reaches dozens of kilometres into the rural areas downwind from built areas. It is also
interesting to note that the measurements show that the air above the city contains, on average, 4 ppm
more CO2 than the air above the countryside, peaking above the Botlek area (a difference of around 8
ppm).
13
Permission was only granted for flying during the daytime

Eindrapport Climate Proof Cities
18
Figure 1.7 Flight paths and measurement results of Lagrangian flights above the southern Randstad at 300
metres on 26 May 2012 between 10:20 and 13:12 UTC (easterly 25-35 kn, clear Qn~600W m-2, Tmax 26

C). The
boundary layer height (for explanation, see Figure 1.2) was around 1200 metres at that time. Leg 1 is upwind
of the urban area, leg 2 follows a trajectory right across the city (or between the urban areas), leg 3 is
downwind. In addition, measurements were taken along the coast (leg 4). The colour of the trajectory
corresponds to the surface temperature measured.
Figure 1.8 Model simulation of the temperature distribution in the southwestern Randstad at UT 20:00 (22:00
LT). Temperatures are in o Celsius (source: Ronda et al., 2010).

Eindrapport Climate Proof Cities
19
1.3 The influence of anthropogenic heat production
Within CPC, much time has been spent on improving the representation of the urban area in the WRF
(Weather and Research Forecasting) mesoscale model, for instance with regard to anthropogenic heat
sources (Ronda et al., 2012). Important anthropogenic heat sources include industry, individual
households, buildings, traffic, people and animals. Until recently, little data was available about the size
of the anthropogenic heat sources on city and neighbourhood level in Dutch cities on the one hand, and
the locations of these sources on the other (Klok et al., 2010). This is why these emissions are not
generally included in calculations of the UHI effect (or only in a relatively simple way) by mesoscale
models.
Using the LUCY model (Large scale Urban Comsumption of energY; Lindberg en Grimmond, 2013) spatial
variation in anthropogenic emissions in the Randstad was first examined. This took place for an area of 5
x 5 km. On a warm day in the Netherlands, the local differences in anthropogenic heat emitted turned
out to be quite large (Figure 1.9): in the urban areas around The Hague and Rotterdam the emission of
anthropogenic heat reaches values of around 20 W m-2 at night and around 70 W m-2 during the day,
while the emission of heat is much lower in the countryside. These spatial differences in anthropogenic
emissions found with the LUCY model were then implemented into the WRF model (Ronda et al., 2012).
Figure 1.9 Antropogenic emissions of heat (in W m-2) for the Randstad at 2 AM local time (left) and 12 noon
local time (right) as estimated using version 3.1 of the LUCY model (Lindberg en Grimmond, 2013).
The most important conclusions are:
1. in the Netherlands, anthropogenic emissions of heat are an important parameter that determine the
UHI effect in Dutch cities. Incorporating anthropogenic emissions of heat from LUCY leads to
simulated temperatures that are (locally) up to 0.6 ˚C higher or 0.3 ˚C lower than the temperatures
that were calculated without taking into account anthropogenic emissions of heat. These
simulations suggest that anthropogenic emissions in the Randstad are locally responsible for 10%
extra UHI effect;
2. the spatial variations in anthropogenic emissions have an effect on the local climate at city and
neighbourhood level in the Netherlands that cannot be ignored. If temporal and spatial variations in
anthropogenic emissions are not taken into account in the model, the local temperature is
underestimated by up to 0.2 ˚C or overestimated by up to 0.6 ˚C. Traditionally, this spatial variation
is not taken into account in mesoscale models for the atmosphere. This means that weather
predictions based on these models calculate an overestimation of the temperature, while for other
areas, the temperature is underestimated.

Eindrapport Climate Proof Cities
20
1.4 Evaporation in the city
Under the influence of climate change, long warm and possibly dry periods will have to be dealt with
more often in the future. There will be more demand for cooling of the urban area. At the same time it is
important to be careful with water usage, especially in such periods, in order to prevent nature reserves
from drying up and groundwater levels from decreasing. Evaporation is central to this problem:
evaporation can help moderate heat in the city, but by definition, this requires water.
Relatively little is known about the evaporation of water in the city. Data about evaporation can help in
urban water management policy. A good estimate about the evaporation during warm, dry periods can
help in the distribution of the available water across different needs (such as vegetation management,
drinking water) in relation to different policy goals (such as cooling urban areas and preventing the
degradation of wooden pile foundations
14
and salt intrusion (Brolsma et al., 2012). Insight into the
workings of evaporation in the city can aid in designing the urban area in such a way that cooling is
ensured through evaporation, using as little water as possible. This makes it easier, for instance, to assess
how much water the vegetation needs for survival and maintenance of its cooling properties.
Figure 1.10 Left: Scintillometer path between the Sint Franciscus Gasthuis (Lat/Lon 51.56478/4.27747,
elevation 51 metres) and the Erasmus MC (Lat/Lon 51.54632/4.28128, hoogte 77 m) in Rotterdam. The
distance between transmitter and receiver is 3451 metres, orientation

180

. Right: set-up for evaporation
measurements on top of an apartment complex on the Ingenieur J. P. Van Muijlwijkstraat in Arnhem
(51°59'4.97"N, 5°55'5.73"E)
http://www.climatexchange.nl/sites/arnhem/index.htm
. The measuring system consists
of a 3D ultrasonic anemometer (Gill R3-50) in combination with a fast open-path infrared gas analyzer (Li-Cor
LI-7500) attached to the top of a 4-metre-high mast.
In CPC, first estimates of the evaporation in Arnhem and Rotterdam were made (Jacobs et al., 2014). The
results for Arnhem come from Eddy covariance measurements carried out since spring 2012 (Figure 1.10
right). For Rotterdam, the Large Aperture Scintillometer data was used, through which the evaporation
can be calculated indirectly
15
(Figure 1.10 left; Appendix D). In addition, the results of the sap flow
measurements were analyzed (Slingerland, 2012), which also give an indication of the effects of
evaporation caused by trees on the city’s water balance.
14
Fluctuations in ground water level can provoke a rotting process in wooden piles that support older buildings.
15
Recenty a so-called ‘microwave’ scintillometer was developed by WUR-MAQ for an STW project (Hartogensis et al., 2012). Together with
an optical scintillometer, this can determine both the average perceptible heat flux for an area and the evaporation. This development
offers new possibilities (i.e. routine assessments of average evaporation in cities) for the water management in the city.

Eindrapport Climate Proof Cities
21
The results of the scintillometer measurements in Rotterdam show a pattern where high peaks in
evaporation go hand in hand with relatively sunny days. (Figure 1.11 top). Approximately 21% of the
average precipitation in Rotterdam in the summer months (3.2 mm per day
16
) evaporates again (0.67
mm). This evaporation amounts to a cooling rate of 20 W m-2 (approx. 11% of the incoming solar
radiation) (Table 1.1).
In Arnhem, the evaporation correlates strongly with precipitation (Figure 1.11 bottom). Around 60% of
the average precipitation per day (= 24 hours) from April to September (1.44 mm) is used for evaporation
(0.86 mm). This amounts to a cooling rate (E) of 25 Wm-2 per day; this is approx. 14% of the average daily
solar radiation incoming during that period in Arnhem (ca. 180 W m-2).
According to these measurements, the connection with precipitation during the summer months is
clearly stronger in Arnhem than in Rotterdam (Figure 1.11). The evaporation in Arnhem decreases much
less slowly after precipitation. Jacobs et al. (2014) offer a possible explanation for this difference: the flat
roofs around the weather station in Arnhem retain the water better and for longer than in Rotterdam.
This would mean that building styles or other measures that help retain rainwater better and for longer
benefit cooling at the beginning of warm, dry periods.
Sap flow measurements are a very different kind of measurements to the previously mentioned
measurements. However, we find comparable evaporation rates: 0.72 and 0.98 mm per day, amounting
to a cooling rate of 21 and 28 W m-2. Calculations show that the average cooling rate varies from 1.1 kW
to 2.2 kW from tree to tree. On some clear days the hourly average of maximum water consumption
reaches 12 to 16 litres, which amounts to an hourly average of 8.2-10.9 kW per tree.
The average water consumption of the 5 trees studied was 50 litres per day (April – September). Taking
the crown diameter into account, we used this to calculate an average water consumption of 0.64 mm
per day. Extrapolating this result to all trees (600 000) in the Rotterdam area (319 km2) amounts to a total
evaporation of < 4% of the precipitation in this period (386 mm). Although this concerns rough estimates,
it shows that the water consumption of the current number of trees only has a minimal effect on the
city’s water balance. However, this can vary locally, especially on days that the trees’ water consumption
reaches its maximum (approx. 170 litres per day).
Because it is difficult to carry out routine measurements of evaporation in the city, scientists sometimes
try to deduce them from the so-called reference evaporation. That is the evaporation of ‘a healthy and
actively growing field with a good water supply’ which is subsequently corrected for the properties of the
urban surface. However, it turns out that this is not possible: the evaporation in a city reacts differently to
the weather than that in a field. In the city, evaporation decreased on dry days, while it increased in fields
and woods (Jacobs et al., 2014).
There is much less evaporation in a city than in the countryside. As a result, a large part of the incoming
solar energy is transformed into perceptible heat. In order to limit the UHI effect, there should be more
evaporation. More vegetation and more water in the city contribute to this. The water supply of urban
vegetation can also be improved during dry periods, which would keep the evaporation at more or less
the same level.
16
This did concern an extremely wet summer in Rotterdam.

Eindrapport Climate Proof Cities
22
Figure 1.11 Daily evaporation measured (green foreground, mm per day) in Rotterdam in the year 2012 (top)
and in Arnhem from June 2012 – October 2013 (bottom).The grey bars in the background indicate days with
more than 1 mm of precipitation (source: Jacobs et al., 2014). (mrt=March, mei=May, Okt=October).
Table 1.1 Comparison of the average daily evaporation in Arnhem (Eddy covariance measurements) and in
Rotterdam (LAS), and evaporation calculated on the basis of sap flow measurements of trees in Rotterdam. S in
incoming shortwave radiation; Lin incoming longwave radiation; All_in total incoming shortwave and longwave
radiation; E evaporation (source: Jacobs et al., 2014).
Sin
(W m-2)
Lin
(W m-2)
All_in
(W m-2)
Evaporation
(mm day-1)
E
(W m-2)
E/Sin
(%)
E/All_in
(%)
Arnhem EC
183
352
535
0.86
25
14
5
Rotterdam LAS
188
354
542
0.68
20
11
4
Sap flow park
surroundings
190
364
554
0.72
21
11
4
Sap flow street
surroundings
190
364
554
0.98
28
15
5
 All radiation fluxes are own measurements, on the roof in Arnhem or from the reference station from the Rotterdam
monitoring network.
 Arnhem EC period: June-September 2012 and April-September 2013
 Rotterdam LAS period: April-September 2012
 Sap flow measurements: June-September 2012

Eindrapport Climate Proof Cities
23
1.5 Climate variations within the city
The results below of local differences in urban climate are largely based on the data from the CPC
monitoring network in Rotterdam (Van Hove et al., 2011b, 2014). The measuring network makes it
possible to analyze the temporal and spatial variation in the local urban climate in the metropolitan area
in more detail and to relate it to neighbourhood properties. Each weather station measures not only the
usual variables (air temperature, humidity and wind speed and direction), but also the global radiation
and black globe temperature. This can also offer insight into the temporal and spatial variation in thermal
comfort outdoors and the influence of neighbourhood properties on this. Note that this concerns average
values for an area; very locally (on a micro scale), large differences in thermal comfort occur. In addition
we have used results derived from satellite images (Klok et al., 2012), mobile measurements (Heusinkveld
et al., 2014) and model simulations (Schrijvers et al., 2014).
1.5.1 The variation in temperature
The UHI in the urban area of Rotterdam can be substantial: maximum differences in temperature
(UHImax) between the city and the surrounding countryside of 7 degrees and more are not an exception
(Figure 1.12). It turns out that this does not only go for the summer months, but also for a large part of
the year. In the winter months (DJF) the UHI intensities are generally minimal. However, on some
winter’s days the UHI effect can be considerable. The effect is usually of short duration (less than 1 day)
and it occurs when the wind turns to the east and brings in cold air. There is a sharp decrease in
temperature in the rural area, while the temperature in the city remains unchanged for some time.
Figure 1.12 Box-whisker plot of UHImax at the measuring locations in the Rotterdam metropolitan area. NB:
UHImax is defined as the maximum difference in air temperature between city and surrounding area during a
twenty-four hour period. The values have been calculated for the months of June, July and August (JJA) in 2010,
2011 and 2012 and for the months of December, January and February (DJF) in 2009/2010, 2010/2011 and
2011/2012. A distinction was made between rooftop and ground stations (source: Van Hove et al., 2014).

Eindrapport Climate Proof Cities
24
The variation in UHI within Rotterdam is considerable, as Figure 1.13 also shows. The densely built
locations ‘Centrum’, ‘Rijnhaven’, ‘Zuid’, and ‘Spaanse polder’ show the highest UHI intensities. This
applies to all years (2010-2012) and seasons studied. Furthermore, it is striking to note that that the
temperatures measured by the KNMI station at the Rotterdam-The Hague airport are higher on average
than at the reference location in the countryside north of Rotterdam. A possible cause is the city’s urban
plume effect mentioned earlier.
Figure 1.13 Topographic map (left) and spatial variation in UHI (right) in Rotterdam and surroundings (14.9 x
14.3 km). Normalized UHI values are presented (UHI Centre = 1). In the summer (JJA) the average median and
95th percentile values for UHImax in the centre are 4.2 and 7.5 K (Bron: Heusinkveld et al., 2014).
1.5.2 The influence of neighbourhood characteristics on temperature
In order to have an impression of the influence of neighbourhood properties on temperature, the land
use, geometry and ‘urban canyon’ effect were examined.
Urban land use
For both the surface temperature and the air temperature significant
17
correlations (p <0.05) were found
for the fraction of built surface, the fraction of paved surface and the fraction of urban vegetation. This
was not the case for the fraction of surface water (Tables 1.2 and 1.3).
Buildings and surfacing
Urban areas with many buildings and a lot of surfacing have a higher surface temperature and UHI
intensity. The fraction of built surface appears to be of decisive influence. The surface temperature rises
by 1.4 C for each 10% increase in the built fraction. In this case, the median value for air temperature
rises by 0.34 C and the 95th percentile value by 0.63 C. A 10% increase in the fraction of paved surface
gives a 0.7 C higher surface temperature and an increase of the median and 95th percentile UHImax of
respectively 0.25 C and 0.44 C.
17
The p-value is used to assess if the correlation is ‘significant’. A p-value lower than 0.05 (i.e. a 5% chance that the correlation is a
coincidence) shows that a correlation is statistically significant.

Eindrapport Climate Proof Cities
25
Table 1.2 The influence of urban land use and geometry on the variation in surface temperature in the daytime
in neighbourhoods in the Rotterdam metropolitan area. The surface temperatures were calculated using
satellite images. The correlations are significant based on a 95% confidence interval (source: Klok et al. 2012).
1:determined for neighbourhoods
Table 1.3 The influence of the urban land use and geometry on the variation in UHImax () within the Rotterdam
metropolitan area. The UHImax values are based on air temperature data after sunset. The correlations are
significant based on a 95% confidence interval (p>0.05). (Source: Van Hove et al. 2014).
Heat factor
Range of
values
Increase/decrease in
surface temperature (C)
with 0.1 increase (10%)
Pearson
correlation
- r
Comments
Urban land use1
Built fraction
0.00 – 0.39
1.4
0.54
Fully paved fraction
0.00 – 0.96
0.7
0.62
Green fraction
0.02 – 0.66
-1.3
-0.83
Water fraction
0.00 – 0.63
0.2
0.13
Insignificant correlation
Urban geometry1
Sky View Factor (SVF)
0.52-1.00
-1.4
-0.61
Upon increase in SVF
Building height
3 - 38 m
0.3
0.52
Upon increase of 1 m
Albedo
0.06 – 0.16
-0.8
-0.64
Upon increase of 0.01
Emissivity
0.92 – 1.00
-1.7
-0.90
Upon increase of 0.01
Increase/decrease in UHImax (in C)
with increase of 0.1 (10%)
Heat factor
Range of
values2
median
r2
P95
r2
Comments
Urban land use1
Built fraction
0.03-0.38
0.34
0.64
0.63
0.60
Fully paved fraction
0.14-0.74
0.22
0.58
0.44
0.60
Green fraction
0.01-0.64
-0.33
0.65
-0.62
0.48
Water fraction
0.00-0.39
Geen significante relatie
zowel toename als afname
Urban geometry1
Sky View Factor (SVF)
0.44-0.78
Geen significante relatie
Building height
2.3 - 26.6 m
0.08
0.69
0.19
0.80
bij toename van 1 m
Albedo
0.08-0.17
Geen significante relatie
1: determined within a radius of 250m around each weather station; 2excl. Zestienhoven (WMO) and Reference

Eindrapport Climate Proof Cities
26
Vegetation
The fraction of vegetation is often inversely proportional to the fraction of fully paved surface. Indeed, an
increase in the amount of vegetation is often at the cost of the paved surface
18
. If 10% of the paved and
built surface makes way for vegetation, the surface temperature decreased by 1.3 C. In this case the
median value for the UHImax decreases by 0.33 C and the 95th percentile value by 0.62 C. The results of
mobile measurements show a comparable reduction (Heusinkveld et al. 2014). The same applies to the
correlation between UHImax values in different cities and the amount of vegetation in those cities
(Steeneveld et al., 2011). The correlation between UHImax and the amount of vegetation is thus a robust
one (Figure 1.14).
Figure 1.14 Maximum UHI intensity (UHImax, 95th percentile values) as a function of the percentage of
vegetation in an urban area, determined for the Rotterdam metropolitan area and Dutch cities (source:
Steeneveld et al., 2011).
Surface water
In general, it is assumed that surface water in the city has a cooling effect on the surrounding area in
summer. However, this is not always the case. The cooling effect occurs thanks to the fact that part of the
solar energy is absorbed and transformed into evaporation of the water. In addition, solar energy is
stored. Water has a great capacity for heat and can emit the stored energy as heat. The cooling effect of
open water is therefore highly dependent on the water temperature in comparison with the temperature
of the area surrounding it. During the summer, the water heats up gradually, as a result of which the
cooling effect on the surroundings decreases. After sunset the water temperature can even be higher
than the temperature of the surrounding built area, with the result that the last cools down less quickly
(Figure 1.15).
18
Green roofs, façades and trees on streets form an exception to this.

Eindrapport Climate Proof Cities
27
Figure 1.15 Variation of the temperature
of the air and the water, measured in 2010
in the Westersingel in Rotterdam (Brolsma
et al., 2011; Slingerland, 2012).
Steeneveld et al. (2014) even give a weak positive correlation between UHI intensities in Dutch cities and
the fraction of surface water in these cities. However, large bodies of water also offer a surface across
which the wind may blow without obstacles. During the day this natural ventilation can have a beneficial
effect on the thermal comfort during warm days. The ‘Rijnhaven’ neighbourhood can be used as an
example to show the contradictory effects of urban surface water. For this location the highest UHImax
values were found as a result of the warming effect of the surface water after sunset. During the day, the
situation is different, however: the relatively high wind speed ensures that it is more pleasant than other
locations in Rotterdam on summery days.
The eventual effect of open water is thus highly dependent on the dimensions (surface, depth), the
situation in terms of the direction of the wind and in terms of buildings and other structures in the area.
This ‘complex character’ of water also explains the absence of a clear, strong correlation between air
temperature and the fraction of surface water.
As noted earlier, the above analyses provide information about the influence of the properties of an area
on a neighbourhood level. Within that (i.e. on the micro scale) the differences can be considerable. The
results of measurements carried out in a small park in Rotterdam illustrate this. They show that on
summery days (days with a maximum temperature of 25 to 30 C) the average air temperature in a park
can be up to 3 C lower than outside the park (Figure 1.16; from Slingerland, 2012). This makes the air
temperature equal to the temperature outside the city. However, the measurements also indicate that
this ‘Park Cool Island’ effect only has a limited influence on the air temperature in the surrounding built
area. Comparable results were found with mobile measurements (Heusinkveld et al., 2010).

Eindrapport Climate Proof Cities
28
Figure 1.16 Temperature variation at
the edge of a park (sensor 2) and in a
park (sensor 11) in Rotterdam. The
differences in temperature measured
by the sensors and the KNMI are also
shown. KNMI is the weather station at
Zestienhoven airport on the outskirts of
Rotterdam (source: Slingerland, 2012).
Urban geometry
The spatial variation in both surface temperature and air temperature within Rotterdam turns out to be
related to local differences in average heights of buildings and other obstacles. This is a highly influential
factor especially for the UHImax (r2 = 0.69-0.80): if the average height in an area increases by 1m, the
median value increases by approx. 0.1 C and the 95th percentile value increases by approx. 0.2 C.
The spatial variation in surface temperature also turns out to be related to the average ‘Sky View Factor’
(SVF) and surface albedo in an area. Neighbourhoods in Rotterdam with a larger average SVF and a
greater surface albedo have a lower surface temperature. A possible explanation is that a higher SVF and
greater surface albedo mean that less solar radiation is absorbed, so that surfaces heat up less during the
day. However, we did not find a clear correlation between these parameters and the spatial variation in
air temperature within Rotterdam. Apparently, thermal properties of buildings in an area, such as
‘thermal admittance’ (the ability to store heat and emit it) play a greater role after sunset. In addition,
the differences in air temperature between the locations are less substantial than those in surface
temperature, which could be a consequence of advection (the sideways influx of air).
Sep-30 Oct-01 Oct-02 Oct-03 Oct-04 Oct-05
0
5
10
15
20
25
30
Date
Temperature
Sensor 2 against KNMI
Sensor 2
KNMI
difference
Sep-30 Oct-01 Oct-02 Oct-03 Oct-04 Oct-05
-5
0
5
10
15
20
25
30
Date
Temperature
Sensor 11 against KNMI
Sensor 11
KNMI
difference

Eindrapport Climate Proof Cities
29
Urban Canyon Effect
An important phenomenon on the micro scale is the so-called ‘Urban Canyon’ effect. An Urban Canyon
represents a narrow street with tall buildings on both sides. Within CPC a microclimate model was
developed that makes it possible to analyze processes in the Urban Canyon more accurately. The
simulation model combines radiative transfer, conductive heat transfer and convective heat transport by
Computational Fluid Dynamics (CFD) modelling at 1 meter spatial resolution. (Figure 1.17). This makes
the model unique in comparison to other models for the microclimate (Schrijvers et al., 2014).
During the day it is possible to distinguish two opposite effects: tall buildings provide shade, with the
result that street and wall surfaces heat up less in the Urban Canyon. However, the model simulation also
shows that incoming sunlight is very efficiently absorbed between tall building through ‘multi-reflection’.
Then, warming of the street and wall surfaces takes place in the Urban Canyon. After sunset, the high
buildings decrease the thermal emissions from buildings into the atmosphere (‘long-wave trapping’) so
that it stays warm for longer in the Urban Canyon.
In addition, the model simulation shows that ventilation, or rather the transportation of heat through
convection, is of high importance. The ratio of building height to street width (H/W) is significant here. In
model simulations with (‘Weather and Research Forecasting’) WRF, Theeuwes et al. (2014) find an
optimum H/W ratio of around 1 (the buildings are as tall as the street is wide). Higher or lower ratios
both have advantages and disadvantages in terms of ventilation and shade.
The best ventilation in the street is achieved through a H/W ratio of 0.5 or lower (the street is (more
than) twice as wide as the buildings are tall). Up to a height-width ratio of 1.0, the air at street level still
mixes with the canopy layer (the air above the city)
19
. At higher ratios (the buildings are taller than the
street is wide), especially the top part of the Urban Canopy is mixed. In this case a highly stable air
situation occurs in the lower part of the canyon where wind speeds are very low and there is hardly any
mixing of the air. There is, however, more shade, and there is therefore less warming of surfaces in the
Urban Canyon (although it is not the case that a H/W ratio of 1 or higher means that there is no warming
through solar radiation at all) (Kleerekoper 2012).
Figure 1.17 Schematic reproduction of the
micro scale model developed. The input is
on the left, with buildings and
accompanying parameters such as height-
width ratio, and material properties such
as albedo and heat capacity. The various
physical properties can be switched on and
off independently (ventilation, radiative
transfer (short/longwave), etc.). The
output is on the right, with surface
temperature, air temperature and air
currents (Schrijvers et al., 2014).
19
In the Netherlands most streets are wider than the buildings are high (ratio below 1).

Eindrapport Climate Proof Cities
30
1.5.3 The variation in thermal comfort
The variation in thermal comfort in the Rotterdam metropolitan area was determined using the
Physiologically Equivalent Temperature (PET). The PET values calculated were subsequently related to
physiological stress and stress perception (Appendix C).
The number of hours that can be classified as hours with moderate to high heat stress (PET > 23 ) is
greater at the city locations than at the reference location in the countryside (Figure 1.18). Exceeding the
threshold value for thermal discomfort almost always took place during the day (157 hours at the city
locations and 93 hours at the reference location, or 21 and 12.5% of the total number of hours that
month). The months of July in 2011 and 2012 were cooler than usual. Although the number of hours with
PET > 23 C was lower (32 hours in 2011 and 77 hours in 2012), the relative differences found between
the locations is comparable to those of 2011.
Figure 1.18 Frequency distribution for the different thermal comfort classes during the day and at night for July
2010 for the different locations in the city and for the reference location (source: Van Hove et al., 2014).
We can ascribe the greater number of hours with lower thermal comfort in the urban areas to the lower
wind speeds. The differences in air temperature between the city locations and the countryside are
minimal during the day (< 2 C) or even negative (for instance “Rijnhaven”). In addition, we have seen
that the direct radiation from the sun on the urban locations is less on average than in the countryside.
The same applies to the atmospheric humidity. The differences in radiation and humidity, however, do
not have a noticeable effect on PET.
The variation in thermal comfort within the metropolitan area also turns out to be largely related to
differences in wind speed. The wind speed at the ‘Rijnhaven’ location, for instance, is relatively high
(approx. 80% of reference) due to the presence of a large body of water. This also explains why the
number of hours with reduced thermal comfort is relatively low for this location. In reverse, the large
number of hours that exceed the threshold in Ridderkerk can be explained by much lower wind speeds at
this location.
The situation changes after sunset when the UHI effect plays a greater role. The variation in PET in the
urban area is then determined in large part by local differences in temperature. As we have seen earlier,
Rijnhaven has the highest maximum UHI values, while relatively low values are found for the green
location of Ridderkerk. An important conclusion is therefore that a greater UHImax at a certain location
does not automatically mean less thermal comfort during the day.

Eindrapport Climate Proof Cities
31
1.5.4 The influence of neighbourhood characteristics on thermal comfort
The spatial variation in PET during the day is largely determined by differences in average wind speed at
the locations. It is not possible to deduce clear, direct correlations with land use or geometric factors,
such as building height, from the measurements from the measuring network in Rotterdam. When the
situation changes after sunset and the UHI effect begins to play a more prominent role, outdoor thermal
comfort is tied to urban properties that are important for the UHI effect.
However, this result requires further explanation:
 The PET values calculated for the locations in Rotterdam are average values for the areas. Very
locally (on the micro scale) large differences in PET can occur. Figure 1.19 shows this for a street
in the Geitenkamp neighbourhood in Arnhem (Heusinkveld et al. 2012). In the same street there
are 15 degree differences in PET because the south of the street is in the shade (trees and
houses) and the north is in full sun. Wind can have a cooling effect but on this particular day
wind was not significant in lowering the PET.
 PET is one of the many thermal comfort indices that have been developed. The sensitivity of the
different indices for meteorological variables turns out to vary greatly.
 PET is calculated based on physical and physiological factors. The latter factors were only studied
for a standard person. Subsequently the results from German research were used to relate the
values calculated to stress perception. However, this relationship could be different for Dutch
citizens, for Dutch weather conditions. In addition, psychological factors were not taken into
account. According to research done by Klemm et al. (2014) these are highly influential for how
people truly perceive thermal comfort in an environment.
Figure 1.19 Physiologically Equivalent Temperature (PET) and radiation exposure (mean radiant
temperature,Tmrt) in the Doctor Schaepmanlaan and Rozendaalseweg in Arnhem. (source: Heusinkveld et al.,
2012).

Eindrapport Climate Proof Cities
32
Conclusions
In relation to the vulnerability of the built environment, the CPC programme has yielded the following
conclusions:
How does the local climate work in Dutch cities?
Temperature
 Each city or city district in the Netherlands experiences an Urban Heat Island effect (UHI) and
substantial differences can occur in UHI at living level within the city;
 The UHI intensity of Dutch cities is considerable and comparable to that of other European cities;
 The UHI intensity is especially great after sunset because the countryside cools down much faster
than the city, where cooling only takes place at the end of the night;
 The UHI intensity is the greatest in the summer months and in spring and much lower in winter.
However, even on some winter’s days the night-time differences in temperature between the
city and the countryside can be large. The latter is often a short-lived phenomenon (< 1 day);
 The difference in temperature between the city and the countryside is especially large at living
level; the UHI at a greater altitude in the boundary layer above the city is minimal;
 The developments of the micro scale model show that the increased absorption of shortwave
solar radiation through reflection between tall buildings is the driving force behind the UHI
effect; evaporation was not considered in this model for densely constructed high-rise areas.
 Heat production through human activities contributes to the UHI. In and around the large cities
of The Hague and Rotterdam, the anthropogenic emissions of heat are a maximum of 20 W m-2
at night and around 70 W m-2 during the day.
Thermal comfort
 In general, the number of days with heat stress in urban areas is greater than in the countryside.
In the coming decades, thermal discomfort and heat stress can become important issues for
many cities;
 A greater UHImax for a certain location does not always mean less thermal discomfort during the
day. The UHI’s value as a proxy or indicator for thermal comfort is therefore limited;
 The spatial variation in thermal comfort during the day seems to be predominantly determined
by differences in average wind speed at the locations, while the spatial variation during the night
is determined in large part by differences in maximum temperature.
Evaporation
 In the period from April until September, 20-60% of the average precipitation is lost through
evaporation; this creates an average cooling speed of 20-25 W m-2 per day (i.e. 11-15% of the
incoming solar radiation)
 First rough estimates for Rotterdam show that the water consumption of trees only has a minor
effect on the water balance; locally, however, this may vary.
The minimal evaporation – partly caused by paved surfaces in the city, party by lack of moisture for
evapotranspiration – ensures that a city’s temperature increases. It is not yet known how great this effect
is. It is also not yet clear what the effect of evaporation is on the thermal comfort.

Eindrapport Climate Proof Cities
33
What is the influence of spatial planning?
 The correlation between UHI and a city’s population (as proxy for the size of the city) as reported
by Oke (1973) is not confirmed in our research; other factors such as population density and
city/neighbourhood properties are most likely more important for the UHI;
 The properties of a city or neighbourhood also seem to be more important for the UHI than the
geographical situation;
 Both the daytime surface temperature and the maximum UHI intensity during the night show
significant (linear) relationships to factors of urban land use, such as the fraction of build surface,
the fraction of paved surface and the fraction of vegetation (see Table 1.4).
 The surface temperature during the day and maximum UHI intensity during the night also show a
significant correlation to the average building height.
 The longer the warm spell takes, the less quickly a densely built neighbourhood cools off at night.
In a green neighbourhood this accumulating effect is less present.
 Urban vegetation cools the environment through transpiration and shade. This does mean
enough water must be present.
 On summery days it can be 3 C cooler in a small park than in the surrounding built area.
However, the influence of the Park Cool Island effect on the surrounding built environment is
minimal.
 The ratio of building height to street width (H/W) is at its best at H/W = 1. At H/W < 1 there is
good ventilation, but little shade, while at H/W >1 there is more shade, but there is no mixing of
the air near the ground.
 Due to water’s vast capacity for heat, surface water in the city can have both a cooling and
warming effect on the surroundings. The cooling abilities of surface water, for instance, decrease
in the summer months because of an increase in the water temperature. The final effect of open
water is therefore highly dependent on the dimensions (surface, depth), the situation in terms of
the direction of the wind and in terms of buildings and other structures in the surroundings.
Table 1.4 Summarizing overview of the influence of neighbourhood properties on the UHImax (air temperature)
(source: Steeneveld et al., 2011)
Factor
Effect or average UHImax summer months
Antropogenic heat
+0.5 ˚C average across Rotterdam (38 W/m2)
+2.0 ˚C industrial area (200 W/m2)
Population density*
+0,1 ˚C tot +0.3 ˚C per increase of 1000
inhabitants/km2
Built surface
+0,4 ˚C tot +0.6 ˚C per 10% increase
Paved surface
+0,2 ˚C tot +0.4 ˚C per 10% increase
Urban vegetation
-0,3 ˚C tot -0.6 ˚C per 10% increase
Open water
no significant correlation
Sky View Factor
no significant correlation
Albedo
no significant correlation
Building height
+0.08-+0.19 ˚C per increase of 1 m

Eindrapport Climate Proof Cities
34
2 How vulnerable are Dutch cities to climate change?
Summary
The persons or objects affected by a climate effect are the starting point for a vulnerability analysis. In
the case of heat, it is people who suffer from heat stress, while for flooding because of extreme rainfall it
is capital goods that risk damage. As part of the Climate Proof Cities programme, research was done on
the vulnerability of these objects, which is determined by three factors: the sensitivity to a climate threat,
the level of exposure and the adaptive capacity, and which therefore varies greatly by location.
Heat stress
Elderly over the age of 75 are especially sensitive to periods of heat and can become ill or even die.
Research carried out during a heat wave in 2010 shows that many old people suffered and the heat stress
was considerable. At rest, body temperatures were measured that suggest fever. It seems that the elderly
do not acclimatize to heat within a warning period of three days. This means that this vulnerable group
needs extra attention during heat waves. The productivity of workers outdoors or in buildings without air
conditioning also decreases during heat waves, with macro-economic costs as a consequence.
Buildings can reduce the exposure to heat by providing a cool indoor climate, but research based on the
KNMI’06 climate scenarios shows that in the future a large proportion of Dutch houses will regularly
experience indoor temperatures higher than the accepted levels (the temperature at which the
inhabitants find it ‘warm’). Important factors for the heating of buildings is the amount of insulation and
the degree to which the sun can shine directly into the building. The latter depends mainly on the surface
area of windows facing east and west and the presence of solar shading.
By projecting information about the properties of neighbourhoods, buildings and communities on a map,
a vulnerability map can be composed that shows which parts of the city require extra attention. In
Rotterdam, vulnerable neighbourhoods in terms of the elderly are Spangen, Bospolder, parts of the old
north, Feijenoord, Charlois and other parts of Zuid. In Amsterdam, an analysis was carried out into the
vulnerability of workers, where the historical centre stands out with its combination of a high density of
workplaces in badly insulated buildings.
Pluvial flooding
Buildings, especially their interiors, and switch boxes are in particular sensitive to material damage
through flooding. Furthermore, economic damage can take place through interruption of activities, traffic
disturbances and power cuts. In addition, there are the costs of calling in emergency services and there
are social implications if hospitals and so on are less accessible or functional. Risks and damage due to
extreme rainfall are often dependent on a threshold that differs per object, for instance the height at
which switch boxes are installed. Reducing the exposure during extreme rainfall can be achieved locally
by ensuring that the water stays below the threshold (by increasing local storage and infiltration) or by
increasing the threshold (e.g. higher doorsteps or installation of switch boxes). Analyses carried out
within the Climate Proof Cities (CPC) programme show that a large number of vulnerable and
simultaneously vital objects and networks are present in urban areas. Making an inventory and overview
of this is highly important in order to understand the vulnerability of an urban area as a whole.
Building on chapter 1, the vulnerability maps also show that the vulnerability to heat or flooding in a city
shows great spatial variation. Vulnerability analyses or climate stress tests, as they are called by the
coalitions of the New Construction and Restructuring Delta sub-programme , should pay attention to the
factors named and can then offer insight into the vulnerability of an area at a very local level.

Eindrapport Climate Proof Cities
35
2.1 Introduction
In order to gain insight into which buildings require more or less attention in terms of climate change, a
vulnerability analysis can be carried out (the ‘Analysis’ step from the ‘Guide to Spatial Adaptation’
20
).
Vulnerability of cities to climate change is defined as the degree to which the urban system is receptive to
changes in climate parameters and not able to deal with the negative consequences. Within the Climate
Proof Cities (CPC) research programme there was a focus on changes in the climate as expressed in
temperature and rainfall (see chapter 1). Extremely high temperatures and heavy rainfall can form a
threat to the city and its inhabitants, but a climate threat alone does not immediately imply damage. The
true vulnerability is determined by three factors: the sensitivity to the climate threat, the level of
exposure and the adaptive capacity(IPCC, 2007). Together with the (increasing) chance of an extreme
event of a certain intensity, this vulnerability determines the risk for an urban system. The ‘Vulnerability’
text box explains these concepts. Because the exposure has already been discussed in chapter 1 and no
research was done on the adaptive capacity within CPC, this chapter is especially concerned with the
sensitivity and resulting vulnerability of the urban area.
Two types of methods can be distinguished for analyzing vulnerability, each with its advantages and
disadvantages (Veerbeek and Husson, 2013); the contextual vulnerability analysis, an approach that
examines the causes and determining factors of vulnerability and uses this to identifies areas with
relatively higher and lower vulnerability. The outcome vulnerability analysis focuses, as the name
suggests, on the potential effects of climate change at a given moment in the future. Veerbeek and
Husson (2013) argue that insight into the urgency and severity of the climate threat is what is most
needed at the moment and that an outcome analysis is most effective here. This was confirmed by the
input of stakeholders during CPC meetings, who indicated that they would like information about
thresholds or tipping points, especially in terms of time.
However, an outcome analysis requires a lot of data for scenario development, and spatial details, as well
as agreements about norms and goals. For water management this method seems workable and we will
explore it further in section 2.3. This is more difficult for heat stress. How do you determine, for instance,
how much heat stress one can handle or how many heat-related deaths are acceptable? Therefore, the
contextual vulnerability analysis was used for heat (2.2). The chapter closes with an overview of tools for
carrying out vulnerability analyses that are available or in development (2.4) and main conclusions (2.5).
20
http://www.ruimtelijkeadaptatie.nl/en/

Eindrapport Climate Proof Cities
36
Vulnerability
The exposure to climate phenomena simply has to do with the degree to which the system comes in contact with
that threat. The surroundings of a system is key here: cities in the higher parts of the Netherlands, for instance, are
not exposed to flooding from the sea. Precipitation patterns will not differ much within a city, but some parts are
lower than others, which means that the exposure can vary from neighbourhood to neighbourhood. Neighbourhood
characteristics do influence the (comfort) temperature, so that the exposure to high temperatures varies greatly
according to location (see chapter 1).
The sensitivity to a climate threat concerns the degree to which a system is influenced by the changing climate
parameters. Unlike exposure, sensitivity relates to the intrinsic characteristics of a system. The sensitivity of the
urban area is determined by the number and type of sensitive elements in the system, such as people and objects,
and the sensitivity of these elements to impact or damage. Buildings in a lower part of the city with many basements
containing small workplaces, for instance, are more sensitive to flooding than buildings without basements.
The capacity for adaptation is a system’s (i.e. a city’s) capacity to deal with the effects of climate change, realize
possible adaptations and limit the damage (Smit et al., 2001). A good capacity for adaptation can decrease the
overall vulnerability to a climate threat and increase resilience. The capacity for adaptation depends on many (social)
factors that are difficult to quantify. The question is to what extent the dimension of capacity for adaptation is
relevant at neighbourhood level. ‘Access to technology’, for example, is often not organized at neighbourhood level.
Figure 2.1 Impact of climate change on the (urban) system (Pásztor and Bosch, 2011).

Eindrapport Climate Proof Cities
37
2.2 Heat stress
In order to carry out a contextual vulnerability analysis, the sensitive elements affected by climate change
must first be identified. For heat waves this is first and foremost the people themselves who experience
heat stress. A next question is whether there are sections of the population that are more sensitive to
heat stress than others (2.2.1) and where these are located. Are they located in an area where they are
exposed to high temperatures?
In terms of exposure, it is possible to draw a distinction between indoors and outdoors, with the
terminology of “first shell” and “second shell” respectively (Figure 2.2). A building can be seen as a first
shell that can reduce people’s exposure to heat stress, depending on the characteristics of the building
(2.2.2). In terms of the outdoor climate, several neighbourhood properties were named in chapter 1 that
can increase or decrease the exposure. The presence of trees in a neighbourhood, for instance, decreases
the exposure to extreme heat (of buildings and their inhabitants). The elements for exposure and
sensitivity are combined in 2.2.3 to form vulnerability maps that offer insight into both the location of
vulnerable groups and the exposure to climate influences.
Figure 2.2 First and second shell around sensitive people and objects.
2.2.1 Sensitivity
Above a certain limit, high temperatures lead to heat stress. This heat stress can lead to a decreased
thermal comfort, sleep disruption, behavior changes (greater aggression) and decreased productivity.
However, heat stress can also lead to serious heat-related illnesses such as skin rashes, cramps,
exhaustion, strokes, kidney failure and breathing problems. Heat stress can sometimes even lead to
death (Howe and Boden, 2007).
During heat waves both hospital admissions (for emergencies) and death rates increase significantly
(Kovats and Hajat, 2008). In the Netherlands, death rates increase by 12% during heat waves
(approximately 40 deaths more per day) (Huynen et al., 2001). Within CPC, a database of climate data
(KNMI) and death rates (CBS) was made. This charted the extent of the excess mortality in hot and cold
periods. Figure 2.3 shows that during heat waves, there is an excess mortality of 8 extra people for every
degree above 20°C. That is more than was initially thought. In addition, it turns out that the combination
of temperature and humidity, for which combined indices exist, such as the heat index and the humidex,
is a slightly better predictor for mortality than temperature alone.

Eindrapport Climate Proof Cities
38
The people who are the most sensitive to heat-related illnesses and death are the elderly over the age of
75 and the chronically ill, especially if they have heart, breathing and kidney diseases (Kovats and Hajat
2008; Hajat et al., 2010). An investigatory CPC study in Tilburg during a heat wave in 2010 showed that
the elderly suffered many symptoms and the heat stress was considerable. At rest, body temperatures of
over 38 degrees Celsius were measured (Daanen et al., 2011) (Figure 2.4).
Aside from the factors mentioned above, the elderly often exhibit sub-optimal behavior during heat
waves. The elderly are often afraid of draughts, and this means that they keep the windows closed in the
morning, even though that would be a good time to create ventilation. They also do not tend to turn on
the air conditioning, so that the temperature in the house increases needlessly. Because of their impaired
perception of temperature they tend to wear too much clothing.
Based on the research done on the elderly in Tilburg, the most important behavioural factors that
increase heat sensitivity are:
 Not drinking enough
 Preventing ventilation by keeping windows shut
 Wearing inappropriate clothing
 Moving too much
 Not looking for cool spots enough
Within CPC research was also done on the possibility of acclimatization of the elderly, which comes down
to increasing the people’s own capacity for adaptation. Since meteorological services like the KNMI can
predict a heat wave a number of days in advance, this period could be used to prepare the elderly
physically for the coming heat wave. Healthy young people are able to adapt well to heat (Strydom et al.,
1966). There have been studies in which sweat production doubled itself in seven days, increasing these
people’s cooling abilities. The capacity to acclimatize was never studied among the elderly.
Figure 2.3 Excess mortality per day per °C for
heat and cold related to the number of days
after the hot or cold day.
Figure 2.4 Typical temperature course in living room,
bedroom, on the skin and inside the body of an elderly
woman during the heat wave of 2010 (Daanen et al., 2011).

Eindrapport Climate Proof Cities
39
Eight women aged over 75 and eight women aged 20-30 took part in the CPC experiment. Under
controlled circumstances they did moderate exercises at high temperatures for a number of days, to see
if they adapted to the heat. The older women had a significantly lower sweat production than the
younger women, but neither of the groups showed signs of acclimatization. Possibly a period of three
days is too short for acclimatization (Daanen and Herweijer, 2015).
The insights above have already been incorporated into the national heat plan (Ministry of Health,
Welfare and Sport, 2007), which offers guidelines for heat waves to the elderly, their families, carers and
health care institutions. Aside from guidelines derived from the abovementioned vulnerability factors
(such as drinking enough, providing ventilation, or having a bath or shower to cool off) and ensuring that
information about this is available, an important guideline is that help is offered. Many elderly people are
not able to carry out these actions independently.
Heat also leads to decreased productivity. With temperatures above 25 °C productivity decreases by 2%
per degree of temperature increase (Seppanen et al., 2004). The increase of the outdoor temperature will
not have equal consequences for every sector. People who work outdoors, such as in the agrarian sector,
will experience the temperature effects the most directly. People who work in buildings sometimes have
access to active cooling (air conditioning) and can use this to regulate temperature and humidity. The
increased morbidity (illness) and mortality (death) and decreased productivity during a period of extreme
heat have economic consequences.
For the Netherlands a model was developed as part of CPC which made estimates of this (Daanen et al.,
2013). Based on the W+ KNMI’06 scenario (van den Hurk et al., 2006), the economic costs of climate
change were estimated at approximately 100 million Euro per year around 2050. The most important
factor in the costs is decreased productivity as a result of heat. This is partially compensated by the lower
costs of decreased mortality rates and hospital stays due to less cold in winter. Since more people die of
cold than heat in the Netherlands, each of the KNMI’06 scenarios leads to a decrease of the number of
deaths per year. For the “warmer” scenarios, however, the number of heat-related deaths increases with
higher average temperatures (without adaptation). The same correlation exists for temperature-related
hospital stays: it decreases throughout the year, but increases in summer (Stone et al., 2013).
Because cities house both old and young, ill and healthy people, heat sensitivity varies greatly locally.
2.2.2 the role of buildings
Buildings can both increase and decrease the exposure to heat. Research within CPC, based on the
KNMI’06 climate scenarios (van den Hurk et al., 2006), shows that overheating indoors will occur more
often due to climate change and will last for longer in a large proportion of Dutch residential buildings
(Van Hooff et al., 2014). Overheating occurs when indoor temperatures are higher than the thresholds
for thermal comfort. This threshold is variable and depends on the outdoor temperature.
The research focused on detached houses, terraced houses and apartments. The numerical simulations
were carried out with a category of models known as Building Energy Simulation software (Hensen et al.,
2002; Crawley et al., 2001, 2008; Costola et al., 2009). These simulations provide detailed insight into the
temperatures that occur in each room of these houses, during each hour of the year. The simulations
show that detached houses and terraced houses suffer from overheating less often than apartments
(Figure 2.5). This applies both to apartments directly situated under the roof and apartments located
more centrally in the building. The orientation of the building is an important factor here: windows facing
east and west cause an increase in overheating hours because of the lower position of the sun in
summer, which enables more solar radiation to enter the house (Figure 2.6).

Eindrapport Climate Proof Cities
40
Figure 2.5 Annual number of overheating hours for three types of houses and on average for four orientations
N, E, S, W per building type (Building Energy Simulations). Per building type there are also two different
construction years and accompanying insulation values (1970 and 2012). The simulations were carried out for
the climate year 2006 in De Bilt. This heat wave year is a year that can be seen as a typical “future” year, with a
hot summer and various heat consequences (van Hooff et al., 2014).
Figure 2.6 Schematic overview of the sun’s path in summer and winter. The figure on the right shows that the
sun penetrates comes into the rooms furthermore through the windows facing east and west than through
windows facing south in summer.

Eindrapport Climate Proof Cities
41
In addition, modern and therefore more well-insulated houses (often constructed without outdoor
sunblinds) heat up faster than older houses: they have around 3x as many excess hours (Figure 2.5 and
Figure 3.2). Having more insulation and still suffering from heat seems like a contradiction, but this is
caused by the greenhouse effect in combination with the degree of insulation. The greenhouse effect
means that shortwave solar radiation enters the house easily through the glass, is subsequently absorbed
by the materials in the house, and is then emitted as longwave radiation. This longwave radiation cannot
pass through glass, and this form of heat must therefore leave in another way (convection and
conduction). In more well-insulated houses this heat has more trouble leaving than in less well-insulated
houses. Good insulation is key for limiting heat loss in winter, but the same insulation ensures that in the
summer, an overheated space cools down less quickly. Sunblinds (especially outdoor sunblinds) and
ventilation are therefore very important, especially for modern and well-insulated houses (Van Hooff et
al., 2014) (see also chapter 3.2).
The role of buildings can differ greatly according to construction year and typology. Because cities consist
of a broad range of different building types, heat sensitivity is very locally determined here too, and
adaptation is an accumulation of relatively small, local measures.
2.2.3 Vulnerability maps for heat
The vulnerability maps that were drawn up as part of CPC for the cities of Amsterdam and Rotterdam
21
bring together the information about the exposure and sensitivity of man, buildings and environment, as
described in the previous chapters (van der Hoeven and Wandl, 2014). Combining aspects of the built
environment with the people in it offers insight into areas requiring attention (“hotspots”), and thus into
priorities for policy.
This results in different maps: Figure 2.7 focuses on the elderly in Amsterdam and is based on the surface
temperature, the liveability in the neighbourhood, the average energy label of the houses and the
number of people aged 75 or older per hectare. Large social areas requiring attention can especially be
found in the western part of the city, but also in Noord, Oost and Zuidoost. The most vulnerable type is
characterized by an extraordinarily high number of elderly people per hectare while the average energy
label is mediocre. In addition there are parts of the city where, although there are fewer people over the
age of 75, both the liveability and the energy label of the buildings leaves much to be desired.
Based on the surface temperature, the average energy label and the number of workers per hectare, a
vulnerability map for workers was made for Amsterdam (Figure 2.8). The assumption here is that
workplaces with a bad energy label in warm parts of the city rate low on comfort or use a relatively large
amount of energy for cooling. The historical centre of Amsterdam (Burgwallen) deserves special
attention. Here the combination of a high density of workplaces (more than 1000 workplaces per
hectare) and bad energy performance (energy label G) occurs frequently.
Both maps show that the vulnerability to heat in a city shows great spatial variation. Within a
neighbourhood, areas with high vulnerability (local “hotspots”) can be interspersed with less vulnerable
parts.
21
The vulnerability map for Rotterdam can be found in Chapter 5.

Eindrapport Climate Proof Cities
42
Figure 2.7 Vulnerability map of Amsterdam’s inhabitants related to heat stress. (Van der Hoeven and Wandl,
2014).

Eindrapport Climate Proof Cities
43
Figure 2.8 Vulnerability map of the working population of Amsterdam for heat stress (Van der Hoeven and
Wandl, 2014).

Eindrapport Climate Proof Cities
44
2.3 Pluvial flooding
The sensitivity to flooding in streets can be mapped with measuring instruments such as 3Di (see 2.4.1).
Map layers which show potential damage can be added to the results of these model calculations,
according to the contextual vulnerability analysis model. For flooding in streets an outcome vulnerability
analysis is also possible based on the Adaptation Tipping Point method (Kwadijk et al., 2010; Veerbeek
and Husson, 2013). A tipping point is the moment at which the degree of climate change is such that
current strategy or policy no longer reaches its goals. At this point, different policy is needed. The tipping
point method takes into account climate uncertainties, but without being dependent on (more and more)
climate scenarios. This method was used experimentally within CPC for Rotterdam Noord and Nijmegen
(2.3.2). Information about the urban system’s vulnerability to damage and thresholds is important input
for the tipping point method. Within CPC, separate research was carried out on this subject (2.3.1).
2.3.1 Vulnerability to damage and tresholds
Cities have many paved surfaces, with the result that water tends to stay in the street during heavy
rainfall. Indeed, in the summer of 2014 various urban areas in the Netherlands had to deal with flooding,
causing damage to buildings and requiring the fire brigade to clear flooded basements, tunnels and roads
blocked for traffic. Due to the increasing amount of surfacing in urban areas, the increased frequency of
extreme rainfall and overdue maintenance work the design standards
22
are not adhered to more and
more often. Flooding occurs in more than 90% of municipalities, mainly in one location (Luijtelaar, 2008).
The chances of extreme rainfall increase due to climate change and can cause urban drainage systems to
fail more frequently. Perhaps equally important is the increased sensitivity to damage of the urban
system. Because of the more intensive and expensive design of the urban area, flooding can cause more
damage than it could several decades ago. Urban areas have therefore become more vulnerable to
flooding.
Damage functions that describe the consequences of flooding quantitatively are an aid to municipalities
in determining which elements in the city are most sensitive to pluvial flooding. Subsequently one can
look at which measures contribute most to reducing the flooding and the sensitivity to damage. Within
CPC a first version of damage functions was developed that describes the consequences of flooding
(Stone et al., 2013).
Research by Spekkers et al. (2012) about the relation between precipitation intensity and damage to
buildings and interiors shows that the variation in damage cannot only be explained through precipitation
characteristics. There are also other variables, such as building characteristics and properties of the
drainage system, that can play a role in explaining this. The functions developed by CPC describe the
damages in relation to different variables such as the depth of the water, the duration of the flooding or
the presence of basements.
The first version of damage functions have in common a threshold for water depth at which damage
begins to occur. Up to this depth no damage occurs. This damage already offers insights into the
sensitivity of the various urban elements. An urban element with a low threshold will become damaged
sooner than an element with a high threshold. Water in the streets is the first hindrance that occurs, and
a water depth of 30 cm causes serious traffic disruptions. Especially in areas with more relief or
depressions (such as tunnels), this water depth is quickly reached. However, traffic disruptions only occur
when alternative routes are barely available, or not at all.
22
Of old, drains were designed to manage a type of heavy rainfall that is expected to occur once every two years. That is to say that once
every two years, such heavy rainfall can take place that it (almost) causes there to be water on the streets. For surface water in urban
areas, a maximum flooding frequency of once per 100 years is assumed.

Eindrapport Climate Proof Cities
45
Electricity supplies have a threshold of 30 cm or higher, but these supplies are often higher up on the
pavement, against the façades of buildings, and at these locations a water depth of 30 cm is not quickly
reached. However, it is assumed that a building with a basement is already flooded as soon as the water
reaches the façade. Water in buildings causes considerable damage to houses, and in the case of a
commercial building, also brings risks of a temporary disruption of business.
These results show that especially the flooding of buildings (basements, in particular) and traffic
disruptions can cause more serious damage. The costs for the fire brigade are also considerable,
considering they are always called in case of flooding. Damage to electricity supplies does not occur often
but does have far-reaching consequences. A summary of the thresholds and range of damage caused by
precipitation is included in Table 2.1.
Table 2.1 Summary of the thresholds and range of damage caused by precipitation (Stone et al., 2013).
Impact
Urban aspect
Threshold (m)
Average costs per unit (euros
2012) ((min - max)
Unit
Material damage
Houses and interior
0 (basement)
0.1 (w/o basement)
Content : 750- - 1750
House: 400- - 1200
House per event
Electricity supply
0.3 (low tension)
0,35 (street lights)
0,5 (middle tension)
5000
55.000,
Substation
Street light per
event
Economic
damages
Business interruption
0 (basement)
0.1 (w/o basement)
5,- - 2000,- (2010)
Business per hour
Traffic disruption
0.3
Transport: 10 – 40
Personal: 1,50 – 6
Commuter traffic: 2 – 8,5
Business traffic: 7,5 – 30
(2006)
Vehicle per 15 –
60 min
Electricity failure
0.3 (low tension)
0,5 (middle tension)
Households: 0 – 80
Businesses: 80 – 2500
(based on legal compensation)
Per event 1 – 8
hour
Emergency
assistance
Fire brigade
0 (house with
basement)
0.3 (roads)
250 - 1000
Per turn-out
Social disruption
Accessibility health
facilities
0 (basement)
0.1 (w/o basement)
0.3 (roads)
-
Per health facility

Eindrapport Climate Proof Cities
46
The damaging effects of water in the streets on public health were not included in the analysis. Recent
research (De Man and Leenen, 2014; Sales Ortells and Medema, 2014) indicates that the number of
pathogens in ‘water on the streets’ is so high that it forms a threat to public health. In other sectors there
are calculation models to transform the health risks into a damage function (Kemmeren et al., 2006), but
these have not (yet) been applied to water.
The urban area’s sensitivity to damage can be lowered by raising the water depth at which an urban
element experiences damage. Veerbeek and Husson (2013) have shown that a small increase in doorstep
height has a great effect on buildings’ sensitivity to damage. However, lowering water depths by, for
instance, adjusting the street profiles – making them deeper – can also contribute to this. Reducing the
duration of the inconvenience and the time it takes to clean up, especially when it concerns roads and
commercial buildings, also contributes to lowering the costs of the damage. Decreasing the amount of
flooding also always lowers the costs of calling in the fire brigade.
2.3.2 Vulnerability maps for flooding
In order to test whether the tipping point approach is suitable as a vulnerability analysis, it was used
experimentally within CPC at two locations and analyzed with a SWOT4 analysis (Veerbeek and Husson,
2013). These tests must be seen as proof-of-concepts, and not as substantiated vulnerability analyses.
The Rotterdam Noord case is explained below as an example.
The tipping point method attempts to offer insight into the moment when a system or policy no longer
complies with the predetermined aims or standard for a range of future scenarios. The methodology can,
for instance, be used to estimate until when the current flood protection policy is sufficient under a range
of climate change scenarios, but also, on a smaller scale, to research until when a local drainage system
complies with the prescribed standard in scenarios involving a changing number of users or an
intensification of rainfall.
In essence, the methodology is a sensitivity analysis in which uncertainties are classified under scenarios.
First of all, a series of indicators is defined, as well as the accompanying measuring methodology and the
resulting thresholds that operationalize the aim or standard. The indicators and thresholds are based on
the current standard or are formulated in consultation with stakeholders and depend on the local
situation and ambitions. The indicators chosen for Rotterdam Noord are the percentage of flooded
buildings and traffic disruption in the area (see Figure 2.9) which thus gives shape to a broader
interpretation of flooding in terms of simply not accepting water nuisance. Thresholds were defined for a
number of standard showers with different repetition times, where the percentage of flooded houses
must be smaller than 0.1% for a shower that occurs twice a year, for instance (this takes into account the
height of doorsteps at entrances). As such, many indicators can be used based on different criteria,
measuring methodologies and thresholds within the same analysis.

Eindrapport Climate Proof Cities
47
Figure 2.9 Thresholds chosen for Rotterdam Noord (“bui’ = rain shower).
The system’s performance is then calculated using a quantitative model, based on both the design value
(i.e. a heavy shower with a return period of two years) and a series of increments of the most influential
variables (i.e. +5%, 10%). These increments are related to a series of scenarios where the values are
plotted against time: a 5% and 10% increase of river drainage are reached in 2040 and 2070 according to
climate scenario A, for example, and in 2060 and 2100 according to climate scenario B. The next step
examines at which increment the threshold is reached. Considering the increments are related to the
scenario, it is possible to determine for each scenario when the threshold is reached: the so-called tipping
points. For Rotterdam Noord these have been illustrated in Figure 2.10 based on KNMI’06 climate
scenarios G and W. Figure 2.11 shows which buildings will flood in each scenario. This often turns out to
be individual cases and not groups of buildings.
Figure 2.10 shows that for both KNMI’06 scenario G and W, the tipping point is not reached before 2028
(traffic disruption Bergpolder, for a shower of T=50). For many other indicators, however, it looks as
though the tipping points are reached in 2040 for the W scenario. However, for the G scenario, most
tipping points shift to past 2090.
In this example, this information would tell the municipality that urgent measures are not necessary and
that the vulnerability is distributed relatively equally across neighbourhoods. In their report, Veerbeek
and Husson (2013) show that the results for the Provenierswijk, Liskwartier and Bergpolder areas are
characteristic of Rotterdam Noord. Because of the relatively uniform character of the neighbourhood,
there are no hotspots where measures need to be taken with extra urgency. However, this is not the case
everywhere. For the case study area in Nijmegen, which is characterized by a greater difference in height,
the tipping points have already been reached under the current conditions in the Benedenstad area.
In addition to the current policy and accompanying system (i.e. drainage system), the tipping points for
alternatives are also often calculated. Depending on the implementation period and the possibilities of
building on the current system, adaptation paths can be defined. The so-called adaptation tipping point
method thus facilitates flexible adaptation strategies that can be maintained under various scenarios.
An important aspect in the evaluation of flooding in the urban environment is a better understanding of
the vulnerabilities. Aside from the possible flooding of buildings and roads, it is important to gain insight
into the vulnerability of vital objects, networks and locations where large groups of people are present
(i.e. day care centres or care institutions). Insight into the relationship with indirect damage (for instance
disruptions of services in businesses) is essential here, which means that further inventory of critical
infrastructure, especially electricity supplies, is more and more important.

Eindrapport Climate Proof Cities
48
Figure 2.10 The Adaptation Tipping Points (ATPs) calculated for Rotterdam Noord. The top lines for the
Provenierswijk, for instance, show that the tipping point for flooded houses will be reached in 2095 according
to KNMI’06 climate scenario G (blue line) and in 2040 according to scenario W (black line) (Veerbeek and
Husson 2013). NB: this is a Proof-of-Concept, not a substantiated vulnerability analysis.
Figure 2.11 Map of Rotterdam Noord, marking the buildings that will flood at least once every two years
(Veerbeek and Husson, 2013). NB: this is a Proof-of-Concept, not a substantiated vulnerability analysis.

Eindrapport Climate Proof Cities
49
2.4 Tools for policy makers
In order to organize the urban environment in a more climate-robust way and compile effective packages
of measures, a number of tools were developed within and in connection with Climate Proof Cities. The
New Constructions and Restructuring Delta sub-Programme was an important motive for this
development. As part of this, guides, procedures and aids were developed and made publicly available.
23
Municipalities, water boards and other parties concerned are facilitated by these products and tools in
making their areas, objects and networks more climate proof. A number of tools are explained below.
2.4.1 3Di area model for flooding
CPC has contributed to the development of the 3Di area model. 3Di makes it possible to chart quickly and
accurately where, to what extent, and how quickly extreme precipitation leads to water hindrance and
pluvial flooding (Figure 2.12). Using the extremely detailed AHN2 height database it becomes easy to see
how much water there will be in roads, properties and gardens, and where this water will go. If required,
this can be visualized in 3Di; it is possible to fly virtually over the area to see where the problems are
concentrated. This makes it possible to determine if vital or vulnerable objects, networks and groups are
affected by the problems.
A 3Di area model can be used for a desk study to identify the vulnerable places, but also as an interactive
instrument in workshops, to study the how effective certain adaptation measures are; see 3.6.2).
CPC partner involved: Deltares.
More info: http://www.3di.nu/international/
Figure 2.12 Screenshot of a vulnerability analysis with the 3Di area model.
23
http://www.ruimtelijkeadaptatie.nl/en/

Eindrapport Climate Proof Cities
50
2.4.2 Heat/Drought Stress Model
In order to support the first phase of policy development for municipalities and water boards, partners in
CPC are developing a quick-scan tool that quickly offers insight into areas sensitive to heat and drought
stress, both in terms of the exposure and the sensitivity to damage. The model can also determine
indicatively how effective the adaptation measures are. The resulting heat/drought maps give highly
detailed input for prioritising actions for climate proof cities.
CPC partners involved: TNO, Deltares, Wageningen University.

Eindrapport Climate Proof Cities
51
2.5 Conclusion
The CPC programme has yielded the following conclusions in relation to the vulnerability of the built
environment:
 The vulnerability of our built environment depends on the exposure, its sensitivity (to damage)
and its capacity for adaptation. The CPC research has focused on the exposure (chapter 1) and
sensitivity (chapter 2) for heat and flooding.
 The elderly over the age of 75 and the chronically ill are the most sensitive to heat-related illness
and death. The elderly often exhibit sub-optimal behaviour during heat waves: behaviour factors
(drinking, clothing, ventilation, moving) can strongly influence heat sensitivity. Acclimatization to
heat was not observed for periods of up to three days.
 For exposure to heat, a distinction can be made between two shells around the vulnerable
people or objects: the first shell (building, infrastructure) and the second shell (neighbourhood,
city).
 Heat stress and flooding have various serious consequences, for instance for public health.
Flooding is the most obvious one and leads to damage quickest and most directly. Drought and
heat have fewer immediate results, but can also lead to a decreased quality of life and enormous
damage.
 Buildings can reduce the exposure, and therefore also the sensitivity to heat stress. Newly
constructed houses are more sensitive to heat stress than old houses. Because of the improved
insulation, once heat has entered the house, it is retained for longer. The presence of outdoor
sunblinds and ventilation can play a crucial role in limiting the overheating of houses in general
and of newly constructed houses in particular.
 Apartments heat up faster than terraced and detached houses.
 The orientation of buildings/windows is an important factor that can strongly influence the
degree to which heat is a problem. Windows facing east and west will lead to more overheating
than if they face north and south.
 The indirect and subsequent damage through loss of productivity due to heat stress, extra costs
of calling out the fire brigade for flooding and so on are often considerable and definitely not
negligible;
 Up to a certain level – literally and figuratively – water damage can be prevented by increasing
the threshold and reducing the sensitivity of objects; it is essential that measures are also taken
to limit the damage when this threshold is passed.
 Vulnerability maps for heat for Amsterdam and Rotterdam show that there is great spatial
variation in vulnerability to heat in the city.
 The vulnerability is very locally determined because of the great spatial variation in exposure to
heat stress and flooding and because of the location of sensitive objects and groups of the
population. Especially vital and vulnerable objects, networks and groups (hospitals, shopping
centres, transformer buildings, tunnels, cellars, electricity, roads, telephone, the elderly) lead to
areas that are locally very sensitive to damage in the urban area.

Eindrapport Climate Proof Cities
52
3 Which measures can be taken to better adapt cities to climate change?
Summary
There are many and various measures available to achieve a climate proof city. The Climate Proof Cities
(CPC) research has focused on measures that prevent or decrease flooding due to extreme rainfall or heat
stress. Because both themes are connected to each other, an integral approach would be preferred:
water from wet periods, for instance, should not be drained away, but should be stored and used to
combat drought and heat, the latter through evapotranspiration. More vegetation has a positive effect
on both climate aspects. For the implementation of measures, the effect on other policy themes such as
climate mitigation, biodiversity or air quality is often an important argument. Because the vulnerability
for climate effects is determined locally, the choice of measures also depends on the local context. More
generic design guidelines for 3 levels of scale are:
Buildings
As mentioned in chapter 2, a building’s orientation is highly important; south-facing windows offer the
best indoor climate in all seasons. Applying exterior solar shading and extra ventilation of the house when
the outdoor temperature is lower than the indoor temperature are also effective measures and can
almost completely prevent overheating in modern, well-insulated housing.
In addition, homeowners can help to prevent flooding by collecting as much rainwater as possible, for
instance in a tank or water bag, or by reducing the amount of impervious surface around the building so
that the water can infiltrate in the soil. The harvested water can be used for toilet flushing or cooling.
Water storage on a small scale, however, is relatively expensive. Traditional green roofs, without delayed
drainage, only offer a limited capacity for water storage and have little insulating effect on the building
underneath (because of the good insulation value of modern buildings) and little cooling effect on the
surroundings. A rooftop water storage with delayed drainage (green or blue) is more effective from the
point of view of water storage.
From street to neigbourhood
Applying more vegetation in streets is a good measure against the heat island effect. Street trees are the
most effective in reducing the local outdoor temperature through their combination of shade and cooling
through evaporation. The type and placement of the trees regarding orientation towards the sun and the
direction of the wind is important here. If the street is supplied with enough elements providing shade, it
is ideally also twice as wide as the buildings are high to ensure good ventilation. Also at street or
neighbourhood level, flooding can be prevented by collecting and storing rainwater. Because this involves
larger volumes than at building level, the storage capacity must also be greater. Solutions include water
squares or underground storage. The latter makes it possible to reuse water again.
City and region
During heat waves, inhabitants gratefully make use of city parks to cool off. During hot summer days it
can be an average of 2 oC cooler in parks than in the city and 5 oC cooler than the surrounding area
(expressed in thermal comfort), which makes them real cool islands. By laying out more parks with
different microclimates, citizens can go to places where they feel comfortable.

Eindrapport Climate Proof Cities
53
Parks can simultaneously act as a water buffer during extreme rainfall. In order to enlarge the water
storage function, elements such as wadis, infiltration ditches, crates and drains, and surface level
adjustments can be applied. For open water storage, the quality of the water and possible health effects
must, however, be taken into account. In addition, coolspots can be created just outside the city that can
ensure cool air in the city.
Given these design guidelines, adaptation seems to be an accumulation of relatively small, local
measures. By taking these guidelines and available adaptation measures into account already during the
planning stage, steps can be taken at the same time as maintenance and renovation works, and costs can
be kept down.
3.1 Introduction
3.1.1 Categorising climate adaptation
Because vulnerability is very much locally determined, adaptation will also depend on the local context. A
broad range of adaptation measures is available, but which measures against flooding and urban
warming are more effective than others, and which measures are best suited to a certain urban context?
This is the central question in this chapter of the Climate Proof Cities (CPC) programme and forms part of
the ‘Ambition’ step of the ‘Guide to Spatial Adaptation’
24
.
Until now, rising temperatures in and around urban agglomerations have been part of relatively
unexplored and unexplained territory, which is why the studies presented here have focused in large part
on the effects of measures on heat in the city. Water management in urban areas has a much longer
history and much more is known about it. Designing an urban water system that can drain away extreme
rainfall immediately through drains may be technically possible, but is not desirable from a financial point
of view. An assessment framework that also considers alternative options is important here.
In addition, many suitable measures for combating flooding in urban areas can lead to an increase in
drought and heat stress in that same area. Measures to better adapt cities to the climate must therefore
be aimed at draining away the excess water so slowly that there is enough water present to bridge dry
and warm periods without any problems. It is also important to see to what extent adaptation measures
can contribute to the mitigation question or other policy themes. Adapting cities to the (changing)
climate therefore requires an integral approach in most cases.
In this chapter we offer an overview of adaptation measures that have been studied at three scale levels
in the CPC programme: buildings (3.2), from street to neighbourhood (3.3) and city & region (3.4). There
is a short description of the working and effectiveness of a number of measures in terms of flooding
and/or heat stress. The complete analyses can be found in the various reports and publications referred
to in the text. Finally, design guidelines are set out for each scale level. Climate change occurs gradually
during a period of dozens of years. This offers the possibility of introducing measures in stages and
carrying them out at the same time as other activities such as maintenance, renovations, new
constructions, reconstructions and urban development, at lower costs. The last chapter (3.5) then offers
an overview of helpful tools for outlining the adaptation strategy that are available in or in development.
24
http://www.ruimtelijkeadaptatie.nl/en/

Eindrapport Climate Proof Cities
54
3.1.2 Description of the measures
A great many different adaptations to a changing climate are possible in urban areas. These range from
influencing the urban climate (water and temperature) and adjusting buildings and infrastructure to
adapting human behaviour and increasing the acceptance of the inconvenience that occurs. The costs of
the measures named here can also differ greatly, both the costs of implementing these changes and
managing and maintaining them. The parties who carry these costs and enjoy their benefits also vary
greatly per measure. Therefore, a the cost-benefit analysis must be built up from within the local
situation. For an area-specific approach, however, where the emphasis is placed on co-benefits, a
breakdown of the expenses that are incurred specifically for climate adaptation is less essential.
Important aspects that determine the suitability of a measure are the primary goal of the measure, its
effectiveness and its versatility. The overview table at the beginning of each scale level paragraph uses
letters and symbols to show how the measure scores for these properties.
Goal
CPC research has focused on measures that can prevent or reduce heat stress (H) or damage caused by
flooding due to extreme rainfall (W). Some measures, such as vegetation, influence both climate aspects
(H/W).
All heat measures described here aim at decreasing the exposure of people to high (comfort)
temperatures. People’s sensitivity to heat stress is an intrinsic property and adaptation measures
regarding this aspect are not explored. All measures for the prevention of flooding mentioned in this
chapter are also geared at decreasing exposure. There are measures that can decrease an urban
element’s sensitivity, like raising the threshold of a building or placing switch boxes higher up or making
them waterproof, but these measures were not studied within CPC (see also chapter 2 and Van de Ven et
al. (2009)). Within CPC, there was a focus on measures for preventing the accumulation of water in the
streets, such as water storage, infiltration and drainage (delayed or speeded up), and as such decreasing
the exposure of damage-sensitive objects to flooding.
Effectiveness
Pluvial flooding
In order to make a first order estimate of the effectiveness of the measures against flooding, calculations
were made for each measure to determine how much area (per ha) is needed to process an extra 10 mm
of rainfall (Vergroesen, 2013). Per ha, this assumes 0.20 ha of roof surface, 0.10 ha of road surface, 0.2 ha
of other paved surface, 0.45 ha of unpaved surface and 0.05 ha of surface water. This means that an
extra 50 m3 must be stored per hectare.
The figures that follow from this give an impression of the required scope of a measure. However, before
another 10 mm of rainfall can be processed, the storage space in this facility must be fully available again.
This is why the approximate emptying time of the measure is also listed. The longer the emptying time of
a measure is, the lower the effectiveness of the measure is assessed. In the overview table the
effectiveness of a measure is therefore rated ++ when it is available again within a day, and + when the
emptying time is longer than 24 hours.

Eindrapport Climate Proof Cities
55
Heat stress
The effectiveness of measures in lowering the indoor temperature was calculated with computer
simulations and expressed in terms of the number of hours that the temperature in a house exceeds a
certain limit. Determining the effectiveness of measures on the outdoor climate is more complex. On the
one hand, one can track the actual physical cooling, for instance by looking at the effect of measures on
the comfort temperature, calculated in PET or UTCI
25
(see Appendix C), air temperature or radiative
temperature. On the other hand, one can look at the psychological effects of measures on human
perception (of temperature): for instance, how do people perceive warmth in a green street or in a paved
street? Both aspects are important in order the describe the effectiveness of measures for the perception
of heat and thermal comfort comprehensively (Nikolopoulou 2003; Klemm et al., 2014a; Klemm et al.,
2014b in review). However, there is no standard scale with which to read the effectiveness of a measure.
The arrangement in the overview table is therefore an estimate on the part of the researchers and can
vary from negative (-), little to no effect (+/-), to very positive (++).
Type
In addition to the direct effectiveness, a distinction can be made in the type of measure in terms of the
versatility:
The measure is generically (G) applicable, which means that the effects found are generally valid,
irrespective of urban or building typology and also independent of a specific context.
It is typology-linked (T), which means that the effects found are valid for a certain urban or architectonic
typology, independent of a specific context.
It is context-dependent (C), which is to say that the effects of these measures are so dependent on the
local circumstances that they must only be determined for each individual situation.
25
PET: Physical Equivalent Temperature; UTCI: Urban Thermal Comfort Index (see also Appendix C).

Eindrapport Climate Proof Cities
56
3.2 Buildings
3.2.1 Goal
Measures at building level apply to the indoor climate of buildings or on water processing on top of and
directly surrounding buildings:
1. Preventing rising temperatures indoors, or active cooling (H)
A cool indoor climate can decrease the exposure of humans to high temperatures. This can be achieved
by preventing rising temperatures indoors, so-called passive measures. Chapter 2.2.2 gives insight into
the characteristics of a building that cause overheating quickly, such as the type of building, its
orientation and the degree of insulation. A building can also be cooled actively, preferably in a manner
that does not use fossil energy and does not contribute to climate change. Cooling through evaporation,
for instance by keeping roofs and façades wet, is one example of this.
2. Infiltration and rainwater storage in and around buildings (W)
Buildings and the impervious surface around them prevent the infiltration of water. They collect the
water that is often drained away through sewers in cities. Buildings can also be deployed to store water
and/or drain it slowly during heavy showers. A contra-productive trend is the surfacing of gardens, so
that the percentage of paved surface increases and more water flows away faster. Through ‘rainwater
harvesting’, the collecting and utilizing of rainwater on top of, in or near the building, flooding can be
prevented, and water can be stored in order to be employed for something useful.
3.2.2 Measures
Table 3.1 gives an overview of the building measures that were studied within CPC. A number of
measures are briefly explained below. More information can be found in CPC publications such as
Vergroesen (2013), Van Hooff et al. (2014) and Brolsma (2013).
Table 3.1 Overview of ‘Buildings’ measures
Measure
Goal
Effectiveness
Type
Solar shading
H
++
G
Thermal mass
H
+/-
G
Ventilation
H
++
G
Albedo
H
+
G
Building orientation
H
++
G
Green roof (extensive; traditional)
H/W
+/-, +/-
T
Green dak (intensive; delayed drainage)
H/W
+/-, +
T
Blue roof
W
+
T
Rain barrel/storage tank
W
+
G
Water storage in crawl space
W
++
T
Green façades
H/W
+/-
G
Water storage in buildings, gardens and
courtyards
W
++
Desurfacing private gardens
W/H
+/++
G
Active cooling
H
++
H: prevention of heat stress; W: prevention of damage through flooding
G: Generic; T: Typology-linked; C: Context-dependent

Eindrapport Climate Proof Cities
57
In CPC, six climate adaptation measures for heat stress were studied for a typical Dutch detached house,
terraced house and apartment (according to Agentschap NL (2013)) (Figure 3.1). Calculations were
carried out for houses built in two different construction periods: houses built according to guidelines
and practices from the 70s of the last century, and houses built according to the Building Policy of 2012.
The research studied the number of hours the temperature in a house exceeds a certain limit (=
overheating hours) (Van Hooff et al., 2014). This limit is variable and depends on the temperature
outdoors.
Figure 3.1 The three types of houses studied: (a) detached house; (b) terraced house; (c) apartment.
The building performance simulations were carried out for the climate year 2006 (KNMI, 2014). The
summer of 2006 saw an above-average number of warm and tropical days and it can therefore be
assumed that there will be more years like this in the future, due to climate change. The results of the
simulations can be found in Figure 3.2.
In general, it can be concluded that applying an extensive green roof has little effect on the indoor
temperatures in the houses. Using exterior solar shading and additional natural ventilation in the house
when the outdoor temperature is lower than the indoor temperature has the most effect and can also
almost completely prevent overheating in modern, well-insulated houses (see section 2.2.2). The overall
findings can be called generic, although the absolute effects will depend on building type, the
surroundings, the inhabitants’ behavior, etc. The measures of albedo, use of solar shading, green roof
and opening of windows are briefly explained below.
Albedo values (H)
Increasing the albedo values (reflection factor of shortwave solar radiation) of the building’s shell limits
the overheating of the walls and roof through solar radiation and therefore results in fewer overheating
hours. For a terraced house, the decrease in average number of overheating hours is around 50% (house
dating from the 70s) to 14% (house dating from 2012). The absolute extent of the effect depends on the
thermal resistance of the building’s shell and the type of house. The effect becomes much greater as the
insulation rating of the building’s shell is lower. In winter, however, a higher albedo value will lead to an
increased consumption of energy for heating. This increase is at around 2% for a terraced house dating
from 2012 and around 7% for a terraced house built in the 70s.

Eindrapport Climate Proof Cities
58
Figure 3.2 Total number of annual overheating hours in the terraced house for the base case and for the
various adaptation measures. (a) Terraced house from the 70s (low thermal resistance). (b) Terraced house
with thermal insulation according to the 2012 Building Policy (high thermal resistance), where ■ stands for the
average number, ● for the minimum number, and ♦ for the maximum number of overheating hours for the four
calculated orientations (north, east, south, west) (based on van Hooff et al., 2014).
Solar shading (H)
The sun is one of the most important causes of high indoor temperatures in houses. The solar radiation
that enters the house through transparent parts of the façade warms up the space considerably (see also
chapter 2.2.2). This strong overheating can be prevented by applying sunblinds to the windows. This
study has shown that fitting windows with moveable exterior solar shading and lowering these as soon as
the total solar radiation on the window is greater than 150 W/m2 has a substantial effect on the number
of overheating hours per year in the three types of housing. For detached houses and terraced houses,
the number of overheating hours can be reduced to around 0-100, while for apartments the number can
usually be reduced to around 200.

Eindrapport Climate Proof Cities
59
Green roofs (H/W)
A green roof is called extensive when the growing material is a
maximum of 15 cm thick. An intensive green roof is between
15 and 30 cm thick. Anything beyond that is known as a roof
garden.
Extensive green roofs with growing material of a maximum of
15 cm (sedum roof, grass roof, etc.) should reduce the heat
transfer from outdoors to indoors due to (1) a different
reflection factor for shortwave radiation (albedo value); (2) an
increase in the thickness of the insulation layer; (3) an increase
in the convective heat transfer; (4) evapotranspiration. However, in the simulation study, applying an
extensive green roof only had a very limited effect on the number of overheating hours indoors; the
number of overheating hours stays more or less the same. The positive effects are countered by the
adverse effect that extra insulation brings. The effect of applying a green roof is greater the lower the
insulation rating of the building’s shell is.
Vergroesen (2013) examined the impact of green roofs for storage of the water that falls on the roof. The
storage capacity of green roofs is significantly lower than the porosity, and decreases (relatively) as
thickness increases. Experience shows that for extensive green roofs an effective storage of 5 – 20 mm
can be expected. The effectiveness of green roofs depends primarily on the evaporation, which
determines how quickly the storage in the growing material becomes available again. The effectiveness
of a green roof is therefore much higher in summer than in winter.
A green roof only stores the water that falls on the roof itself. Assuming a roof surface of 2000 m2 per
hectare and a storage capacity of 15 mm for an extensive roof, a maximum of 30 m3 can be stored, where
50 m3 is needed (see 3.1.2 for an explanation of the principles behind this). The time needed to be able to
use the storage again depends strongly on the evaporation taking place, and can vary from three days in
summer to almost three weeks in winter. An intensive green roof with a minimal thickness of 15 cm (e.g.
a roof garden) has a greater margin in terms of storage capacity. In addition, the temporary storage of
water that has seeped into the drainage layer under the growing material and the delayed drainage of
water can increase the effectiveness of green roofs as a measure for prevention of flooding.
Opening windows (H)
Opening windows in the night/morning enables additional ventilation on top of the usual ventilation
which takes place throughout the whole day. By ventilating when it is cooler outside than inside, excess
heat can be diverted. The windows only need to be opened when it is cooler outside than in, otherwise
opening windows has a counterproductive effect. The research has shown that applying additional
natural ventilation by opening windows above a certain indoor temperature, and only when the outdoor
temperature is lower than the indoor temperature, significantly lessens the number of overheating hours
(to almost 0) for all types of housing.
Active cooling (H)
Buildings can also be actively cooled during periods of heat to create a pleasant indoor climate, for
instance by using aquifer thermal energy storage. In CPC, research was done to determine whether open
water in a neighbourhood could supply buildings in the Netherlands with heating and cooling. In this
country heating forms the greatest challenge because of the relatively large demand for heating in winter
and limited demand for cooling in summer. The balance will become more advantageous in the future
because of climate change.

Eindrapport Climate Proof Cities
60
A case study from Watergraafsmeer illustrates that it is in fact possible to obtain the amount of heat that
is currently used for heating buildings in this neighbourhood from the water system and the water chain
(Brolsma et al., 2013). Because the heat is mostly collected in summer and used in winter, it must be
stored. The capacity for storage of heat and cold in the subsoil of Watergraafsmeer is also sufficient to
satisfy peak demand.
Water storage in the crawl space (W)
Run-off water can be collected in a tank or bag in the basement
or crawl space of houses and other buildings. This harvested
water can be used for toilet flushing and the irrigation of
gardens and roofs (green, blue or grey) in dry, warm periods
26
.
Depending on what the water is used for it can be necessary to
add a purification step. From the point of view of quality, it is
also advisable not to store the first rainwater after a dry period
(the so-called first flush) in the tank.
Assuming a depth of 50 cm, 100 m2 of crawl space is needed per
hectare to store 50 m3 of water (see 3.1.2 for an explanation of
the principles behind this) (Vergroesen, 2013). This corresponds to around 4 houses. For emptying the
water bag it was assumed that if necessary, the water bag can be emptied more quickly than is needed
for using this water in the house or garden. With a pumping capacity of 4 m3/hour the emptying time is
12 hours.
A study within CPC further shows that storage and use of rainwater at house level is relatively expensive
(Hofman and Paalman, 2014). If the costs for such a system are compared to the use of drinking water, it
is not possible to earn back the investment. If this option is chosen anyway, the economic comparison
will have to be made in a broader context, related to the damage through flooding that is thus avoided. In
addition, deciding to purchase these systems will have to be made attractive to homeowners, for
instance through subsidies or reductions on sewage charges. Applying more large-scale systems does
seem to be economically viable.
Desurfacing private gardens (W/H)
The paving of private gardens is often mentioned as a (joint) cause of climate problems because more
water needs to be drained through the sewage and the water balance in the subsoil is disturbed.
Initiatives like ‘Operatie Steenbreek’
27
are focused on depaving private gardens. Still, further
differentiation of the problem and the solution is relevant here. Many of these paved gardens drain the
rainwater into unsurfaced areas, where it will infiltrate. The surfacing then has few negative
consequences for water management. In addition, much of the surfacing used is gravel or small stones.
These retain a certain level of porosity, so that the natural replenishment of the groundwater is largely
preserved. Surfacing made of asphalt, concrete and old tiles and bricks, however, are not at all or
minimally porous. Assuming that an example neighbourhood consists of 25% private gardens, a total
surfacing of all gardens would yield a loss of subsoil storage of around 25-37 m3. These gardens provide
extra drainage to sewers.
26
The Drinking Water Provision (Drinkwaterbesluit) (2011) currently permits collected rainwater to be used within the household, but
only for flushing the toilet: see http://wetten.overheid.nl/BWBR0030111
27
‘Operatie Steenbreek’ (‘Operation Depave’) is a national initiative where nature organizations and ecologists from universities such as
Groningen and Wageningen declare war on paving in Dutch gardens.

Eindrapport Climate Proof Cities
61
Less storage of rainwater in the ground in private gardens can have negative effects on street trees or
other public vegetation in the direct surroundings of private gardens. When there are more unsurfaced
areas and water storage nearby, street trees are better able to withstand dry periods. In addition, green
(depaved) front gardens contribute to a better perception of thermal comfort for passers-by in the street
(3.3.2) (Klemm et al., 2013a; Klemm et al., 2014b in review). Following on from the depaving of private
gardens, it is important to safeguard unsurfaced areas in cities. The more surfacing, the higher
temperatures become (Table 1.4, Van Hove et al. 2014; Table 3.3, Kleerekoper in review a).
At the same time, improvements can be made in industrial zones and large shopping centres, where large
areas are surfaced for parking and the storage of materials. Desurfacing these areas, or at least providing
them with a porous surface, leads to substantial improvements to local water management. If these
areas become greener, this also leads to a reduction in the possibility of heat stress.

Eindrapport Climate Proof Cities
62
3.2.3 Design principles
Considering that 90% of the real estate of 2030 has already been built, it is important to analyze the weak
spots of existing buildings in detail in terms of heat and flooding and to design specific measures in
response. For heat, consecutive steps can be taken for the prevention of indoor overheating, passive
cooling and finally active cooling. A number of general guidelines for this are given below and are
summarized in Figure 3.3.
General guidelines
 Exterior solar shading is important for the prevention of overheating indoors, particularly in well-
insulated modern buildings, but also in older buildings.
 South-facing windows contribute to passive heating in winter and the prevention of indoor
overheating in summer. This is because having fewer windows facing west or east ensures less solar
radiation (overheating) in summer. A roof surface that faces south is also advantageous for solar
energy production.
 Ensuring good ventilation in houses so that heat (and moisture) can be disposed of through
ventilation when it is colder outside than in.
 Storing as much run-off rainwater as possible (‘harvesting’), in, under or near buildings and reusing it.
 Desurfacing gardens and large parking and storage areas contributes to more balanced water
management, an improved perception of temperature for passers-by, and to the cooling of the urban
environment.
 Consider the pros and cons when deciding whether to use green roofs: traditional green roofs –
without delayed drainage – only offer a limited amount of extra water storage and insulation.
 Determine whether the heating and cooling needs of a neighbourhood can be covered locally by heat
and cold from the local urban surface water, groundwater and water chain.
 Avoid the use of air conditioners when cooling is needed. They work negatively because of the heat
production in the urban environment and use energy so that they contribute, in turn, to climate
change.
 Provide information to inhabitants on how they can prevent overheating in their houses: using
ventilation and sunblinds efficiently and at the right moment.

Eindrapport Climate Proof Cities
63
Figure 3.3 Design guidelines for a climate proof house.

Eindrapport Climate Proof Cities
64
3.3 From street to neighbourhood
3.3.1 Goal
Measures at street and neighbourhood level are geared at limiting rising outdoor temperatures,
improving thermal comfort and storing, infiltrating or draining away water on a local scale:
1. Limiting rising outdoor temperatures and improving the thermal comfort with regard to the
surrounding country side (H)
Chapter 1 dealt with the neighbourhood properties that contribute to the heat island effect. It turned out
that the percentage of built/surfaced/green surface and the height of the buildings are the most
important factors. Measures in this chapter take this as a starting point.
2. Water storage, infiltration and transport (W)
The high percentage of surfacing in the urban environment ensures that rainfall during extreme showers
cannot infiltrate or drain away quickly enough, which can cause local flooding and can lead to damage
and traffic disruption (see also chapter 2). There is a range of measures available for water storage,
infiltration or transport.
3.3.2 Measures
Table 3.2 gives an overview of measures at street and neighbourhood level that have been studied in
CPC. There is a brief explanation of a number of measures below. More information can be found in CPC
publications such as Vergroesen (2013), Kleerekoper (2012) , Klemm (2013a, 2014b in review) and
Montazeri et al. (2015).
Table 3.2 Overview of ‘Street and neighbourhood’ measures
Measure
Goal
Effectiveness
Type
Street trees
H/W
++/+
T/C
Green façades
H
+/-
C
Green roofs
H/W
+/-
C
Green gardens
H/W
+/-
G
Light-coloured roofs
H
+/-
G
Evaporative cooling
H
++
G
Height/width ratio
H
+
Enlarging sewage system
W
++
G/C
Depaving
W
+/++
G
Underground storage tanks
W
+/++*
G
Storage on existing surface water
W
++
G
Sunken roads/raised sidewalks
W
++
G
Storage in/under roads
W
+/++*
G
Water square
W
++
G
Infiltration units/crates
W
+/++*
T
Infiltration wells (deep/shallow)
W
+
T
Infiltration ditches
W
+
T
Infiltration transport sewers
W
+
T
Infiltration drainage sewers
W
+
T
Wadis
W
+/++*
T
Adapting ground level
W
+
C
Disconnecting sewers
W
+/++*
C
H: prevention of heat stress; W: prevention of damage due to flooding
G: Generic; T: Typology-linked; C: Context-dependent
* The effectiveness depends on the size of the storage and/or the local situation.

Eindrapport Climate Proof Cities
65
General heat measures (H)
Many parameters are important for the microclimate, such as the degree of surfacing, the type and
colour of the material, the amount and type of vegetation, the amount and type of water, the height and
width of the streets, and within areas, openness, orientation, type of buildings, the land utilization
(residential, mixed urban functions, industry, city centre, business district, agriculture, sports and
recreation) and density (population per ha, houses/ha, FSI (Floor Space Index), GSI (Gross Space Index),
height/width ratio). In choosing a certain measure all the aspects mentioned above play a role.
Comparing the effectiveness of various heat measures using existing studies with measurements and/or
simulations is barely possible. This is because they all concern different conditions: different climate
zones, weather conditions and other urban contexts. In addition, the measuring methods and simulations
differ in terms of measuring locations (especially the height), and different indicators are used to express
comfort, such as air temperature, radiative temperature, surface temperature, thermal comfort (PET,
UTCI), etc. (see also Appendix C).
By carrying out model calculations in the ENVI-met microclimate model, the difference in effects of the
different measures on thermal comfort was studied under comparable conditions and with comparable
measuring methods in CPC (Kleerekoper et al., in review a and b). The effects were studied in a specific
urban context and in an open field where the complexity was slowly built up. By changing one aspect
each time it is possible to see what effect this measure has (Figure 3.4). In Table 3.3 below it can be seen
which measures were studied and what the calculated effect was. Measurements were taken at four
positions at a height of one metre (Kleerekoper et al., in review b).
This study of the generic effects of measures has led to the conclusion that design measures and
measures that influence solar radiation and wind have a great effect on the local thermal comfort. By
contrast, measures that influence air temperature and atmospheric humidity have a more limited effect
in terms of degrees Celsius, but a larger action radius.
The evaporation from the ground and through vegetation was also considered in the simulations.
Evaporative cooling by, for instance, mist spraying was not examined. Local effects can be expected from
this, but this will be because of direct evaporation and not because of a higher atmospheric humidity.
Figure 3.4 Example of an open field where the effects of grass vs. pavementing are analyzed (left); and the four
measuring positions in this field (right) (Kleerekoper et al., in review b).

Eindrapport Climate Proof Cities
66
Table 3.3 The maximum effects in PET measured at one of the four measuring positions and the average effects
in PET across the four measuring positions (Kleerekoper et al., in review b).
Maximum effects in PET
measured at one of the
four measuring positions
(˚C)
Average effects in PET
across the four measuring
positions (˚C)
A. Grass vs pavement
-8
-5.5
B. Detached building of 20*40*8 m (l*w*h)
with brick paving vs. empty field with brick
paving
-8
-0.6 – 0.7
C. Direction of wind (North, South, East,
West, North-East and South-West)
3
0.0 – 0.9
D. Wind speed from 1 to 6 m/s
-12.4
-11.6
E. Grid size
F. Area rotation
G. Two buildings vs 1 building
10
3.5 – 4.2
H. Two buildings with different heights vs
two buildings of the same height
-3.5 and 4.5
-1.1 – 0.9
I. Half-closed building with courtyard vs two
rectangular buildings
0.2
0.1
J. Detached building with trees vs building
without trees
-20
-5.8 – 0.3
K. Half-closed building with courtyard and
trees vs courtyard without trees
-16
-0.5 – -0.1
L. Detached building with hedges vs building
without hedges
-13
-2.9 – 3.5
M. 20 m high detached building vs 8 m high
detached building
-1.5 and 3.5
0.5
Green elements in the street (H/W)
International research has shown that urban vegetation can improve the urban climate because of the
shade provided by tree crowns and the cooling effect of evapotranspiration (Bowler et al., 2010). The
goal of the research on green infrastructures within CPC is therefore to chart the influence of urban
vegetation on the urban climate in warm summer periods – especially in terms of the thermal comfort of
the city’s inhabitants. Data about this was collected via micrometeorological measurements,
Computational Fluid Dynamics (CFD) simulations and street interviews with inhabitants.
In addition, green infrastructures can play a role in the buffering and infiltration of excess rainwater. The
effects of street vegetation on flooding reduction was charted using the 3Di calculation model. 3Di allows
one to trace the effects of one or more adaptation measures on the scope and degree of the flooding
(see chapter 3.5.2).
Street trees and green gardens
(Cargo bicycle) measurements taken at street level in Utrecht show a limited cooling effect for street
trees in terms of air temperature (Tair)
28
, but a clear reduction of the radiative temperature (mean radiant
temperature, Tmrt), especially because of large tree crowns. The average Tmrt in a street with a 54%
surface of tree crowns was 4.5 oC lower than in a street without trees. 10% more tree crowns in a street
28
Presumably partly because of the fact that the measurements took place in the middle of the street and not directly in the shade
beneath the trees.

Eindrapport Climate Proof Cities
67
leads to a a cooling of 1 K Tmrt (radiative temperature) (Klemm et al., 2013a; Klemm et al., planned for
2014). The measurements show a limited influence of green gardens on the physical climate conditions in
outdoor areas, such as air temperature or radiative temperature, excepting, of course, large trees that
offer shade.
Using CFD simulations for the J.P. van Muijlwijkstraat in the city centre of Arnhem (based on the warm
summer’s day of 16 July 2003) a reduction in the average and maximum temperatures of respectively 0.6
°C and 1.6 °C was measured for street trees in comparison with the situation without trees in the street
(Figure 3.6 and 3.7) (Gromke et al., 2015).
Green gardens show a significant positive influence on how passers-by perceive temperature (Figure 3.5).
Seeing green elements at different heights (low shrubs, hedges, tree crowns) makes heat more bearable
for people, and they also appreciate such streets more from an esthetic point of view (Klemm et al.,
2013a, Klemm et al., planned for 2014).
Figure 3.5 Temperature perception of pedestrians in three types of streets (Klemm, 2013a).
In addition to making gardens greener (depaving), tree pits can also be applied with vegetation All
unsurfaced (or semi-surfaced) parts of the streets thus contribute to the improvement of the perceived
temperature for passers-by, the infiltration of rainwater and the storage of rainwater for irrigation of
vegetation in the street.
Green façades
The CFD simulations show that applying green façades results in relatively low reductions of the air
temperature in the street: an average of 0.1 °C and a maximum of 0.3 °C (Figure 3.5 and Figure 3.6)
(Gromke et al., 2015). The effect of green façades on the outdoor temperature strongly depends on the
type of green façade, but for each façade type the effect is only noticeable very close to the façade. In a
study in Singapore, where different green façades were examined, it turned out that the vegetation gives
a reduced temperature of around 2 °C at around 30 cm away from the façade (Wong et al 2010). The
façades with a good substrate seem to be the most effective for this.

Eindrapport Climate Proof Cities
68
In order to prevent overheating of buildings indoors, a green façade is certainly not the most effective
measure. There is a limited effect for poorly insulated buildings, and the effect is negligible for well-
insulated buildings (van Hooff et al., 2014). Green façades also do not play a role as a measure against
flooding due to rainwater (Vergroesen, 2013). Due to their visual impact, green façades or climbing
constructions for plants do contribute to a better temperature perception outdoors (Klemm et al., 2013a;
Klemm et al., planned 2014).
Green roofs
Applying green roofs in CFD simulations did not result in a noticeable reduction of the air temperature at
pedestrian level in the street (Figure 3.6 and Figure 3.6) (Gromke et al., 2015). In general, the cooling
effects were limited to a distance of a few meters from the vegetation. These results reflect the
differences in temperature measured in earlier studies (e.g. Alexandri & Jones 2008, Errel et al. 2009).
Green roofs can play a role in the storage of water during extreme rainfall (see also chapter 3.2.1 and
Vergroesen (2013)). Furthermore, green roofs have little effect on the indoor temperature (see also
section 3.2.2).
Figure 3.6 Air temperature at 2 metres above the ground in the J.P. van Muijlwijkstraat in Arnhem on paths 1
and 2, as shown in Figure 3.7, both for the current situation and for the three alternative vegetation scenarios
(Gromke et al., 2015).
To sum up, placing trees along the side of the street (due to the added shade) has an especially positive
effect on the physical conditions of thermal comfort, namely air and radiative temperature. In addition,
the evaporation from trees can prevent the translation of a significant part of the incoming shortwave
solar radiation into a rise in the air temperature in the city. Street trees are an effective way of improving
thermal comfort in existing streets that catch a lot of sun. But street trees are not needed everywhere;
depending on the orientation of the street or the street profile (height-width ratio) buildings themselves
can create shade for pedestrians. In streets with busy traffic too many trees can even have a negative
effect, due to the dense canopy: the tree crowns then create a ‘tunnel effect’ so that the air cannot
circulate and the exhaust fumes linger in the street.
In addition, all green elements in streets, such as front gardens, façades, tree pits, have a psychological
effect, namely that pedestrians see vegetation and therefore experience better thermal comfort and find
the heat more bearable.

Eindrapport Climate Proof Cities
69
Figuur 3.7 Air temperature at pedestrian height (2 m above the ground) in the J.P. van Muijlwijkstraat in
Arnhem for the current situation (top) and for the scenario with rows of trees in the whole street (bottom). The
rows of trees are represented by black rectangles (Gromke et al., 2015).
Increasing albedo/light-coloured roofs
Satellite images show that a higher albedo (and therefore a higher reflection value) leads to a lower
surface temperature (section 1.3.2), but the effect on the air temperature at street level is not
unambiguous. The measuring data in Van Hove et al. (2014) (section 1.3.2) does not show a clear relation
between albedo and air temperature in Rotterdam. It is important to note here that the albedo value was
registered as an average value for the area around the measuring equipment.
As part of a simulation study, Kleerekoper (in review a) calculated that the transformation from a black to
a white roof at a height of 9 m does offer a temperature reduction of 0.5-1 °C at a height of 2 metres.
This reduction is obtained in 50% of a radius of around 15 metres in and around the building block.
Various foreign studies also describe the cooling effects of albedo on the air temperature (Taha et al.,
1988; Sailor, 1995).
CPC research showed that high albedo values are in any case undesirable for façades, because the
reflected radiation can in turn be reflected towards people in the street and can influence the thermal
comfort or temperature perception of pedestrians at living level.
Height/width ratio
The ratio between building height and street width (H/W) is important. In model simulations with the
‘Weather and Research Forecasting’ model, Theeuwes et al. (2014) find an optimal H/W ratio of around 1
(the building is just as tall as the street is wide) (see also 1.3.2 – Urban Canyon Effect). A wider street
profile (low H/W) is preferable, however, because of better ventilation; sufficient shade can then be
attained through well-located deciduous trees. In addition, in winter, the street actually benefits from the
solar radiation.
Evaporative cooling (H)
Mist spraying systems are used more and more often as a system for local cooling and the improvement
of thermal comfort in the built environment (Figure 3.8a and b). Because the fine water droplets take in
the heat and evaporate, the air around them cools down. To prevent Legionella, the application of
spraying systems must take place with great care, and only with purified water that has been kept
refrigerated during storage.

Eindrapport Climate Proof Cities
70
Figure 3.8 (a) Direct evaporative cooling (through humidifiers) in an urban area. (b) Schematic overview of a
case study including the nozzles of the humidifying system.
Within CPC, high-resolution numerical models were developed to analyze the performance of
evaporative cooling through humidifying systems in urban areas (Figure 3.8b) (Montazeri et al., 2015).
The evaluation, done using Computational Fluid Dynamics (CFD), is based on grid sensitivity analysis and
on a validation study using wind tunnel measurements by Sureshkumar et al. (2008).
The simulation consists of a water spray system with a hollow-cone nozzle configuration used at
pedestrian level and on balconies of a building of medium height (Figure 3.9). The study was carried out
for various wind speeds and properties of the water vapour, in order to evaluate the performance of the
spraying system in terms of reducing heat stress (based on UTCI, see also Appendix C). One of the
evaluations shows that for lower wind speeds, the system reduces temperatures especially at pedestrian
level, while at higher wind speeds, the façades are cooled more (Figure 3.10). The results show that the
upstream wind speed has an important effect on the cooling performance of the spraying system at
pedestrian level and on balconies. In addition, the first results show that the UTCI can decrease locally by
around 2 °C due to the use of the spraying system, and the effects are noticeable at up to 10-20 metres
away from the system (Montazeri et al., 2015).
Figure 3.9 Image of the hollow cone nozzles for
the generic building with balconies (source:
Montazeri et al., 2015).
Figure 3.10 Contours of the air temperatures around
the building when a humidifying spraying system is
applied for two different wind speeds at a height of 10
metres (V10) (source: Montazeri et al., 2015).

Eindrapport Climate Proof Cities
71
Water square (W)
A water square, like Benthem square in Rotterdam, is a
square that is primarily meant for water storage in shallow
constructions during extreme rainfall. In dry periods a water
square is no different to a normal square. When it is not
raining, the square is dry and accessible and it can be used
for all sorts of recreational purposes.
A surface of 167 m2 per hectare is needed to store 50 m3 at
an average water depth of 30 cm (see 3.1.2 for an
explanation of the principles behind this) (Vergroesen et al.,
2013).
The collected water can be drained into the sewer with a pump, or the water can infiltrate. The latter will
depend on the surface and the design of the square. The emptying time for a pump capacity of 4 m3/hour
is half a day. For infiltration, the emptying time can vary from 0.5-5 days. Infiltration can combat drought
and can locally reduce a lack of moisture by a maximum of 20 mm.
When infiltration is applied in the deep substratum, the water can be stored for a longer time and be
pumped up for reuse. A good example of this is a practical experiment in Westland, where rainwaterfrom
the roofs of greenhouses is collected and infiltrated into the ground. As a result, the above-ground basins
near the greenhouses can be kept at a low level, so that when there is intensive rainfall, there is enough
water storage available to prevent flooding, and in addition, there is always water of sufficient quality
available for agriculture in the underground storage. A similar system could also be realized in the urban
environment, where the underground water store could be used for low-grade water applications in the
city (toilet flushing, water for extinguishing fire, cooling, etc.) (Hofman and Paalman, 2014).
Depaving (W)
Porous surfacing allows rainwater to locally infiltrate into the
substratum below. Water infiltrates through the joints of the
bricks or paving or in the granulation of porous asphalt. Brick and
paved surfacing are especially effective for low intensity
showers, while the high permeability of porous asphalt ensures
that relatively large amounts of water can be stored even for
high intensity. Because more water is retained in the area, the
water balance improves and the chance of drought-related
problems decreases.
Assuming a road surface of 1000 m2 and a sand substratum with an effective porosity of 25%, a depth of
20 cm is required to store 50 m3 of rainfall (see the principles behind this in section 3.1.2) (Vergroesen,
2013). The time needed to use the storage again varies from a few hours to about a day, depending on
the porosity of the surface below.
Figuur 3.11 Benthemplein Rotterdam Figure 3.11 Benthemplein Rotterdam

Eindrapport Climate Proof Cities
72
Sunken roads/raised sidewalks (W)
Water can be temporarily stored in a controlled way on streets and in
verges, especially away from main roads. This is the most effective in
a combination of a small drainage capacity and deeper roads so that
more run-off water is stored temporarily and can drain away slowly.
Compartmentalizing roads, for instance with ramps, prevents water
from flowing (too quickly) to lower areas that are sensitive to flooding
due to rainfall. This measure is especially suited to relatively flat
areas.
To store 50 m3 of water with a depth of 5 cm, 1000 m2 of street surface is needed per hectare. The
emptying time through sewers is half a day (Vergroesen, 2013). If infiltration is possible, for instance
through porous surfacing, the measure can also have an effect on drought.
Roads and pavements are constructed on top of a bed of sand. Infiltration crates and other forms of
water storage under roads – often in combination with drainage to prevent high groundwater levels in
the sand bed – can offer extra water storage underground.
Verges are often very suitable for storing water, better than allowing rainwater to stay in streets. By
lowering verges and making as much space as possible for an infiltration ditch or groove, or putting
infiltration crates under them, a lot of room can be created for water storage. If necessary, some extra
drainage can be put in in order to control the groundwater levels on the spot and to guarantee the
emptying time.
3.3.3 Design principles
When examining a street or a square the local context plays a decisive role: where is it convenient to
have sun/shade, and where should there be shelter/ventilation? Also, the physical circumstances
determine the effect of, for instance vegetation. On this scale, too, various relatively small, local
measures can make the street/neighbourhood less vulnerable to climate change. A number of general
guidelines are summarized below and in Figure 3.13.
General guidelines
 Add green elements (in private and public spaces), preferably at as many different heights as
possible and with different microclimates. (Street) vegetation ensures shade and evaporative
cooling and is appreciated esthetically. More vegetation also ensures better infiltration in the
subsoil. To prevent overheating of the urban environment, spreading out green elements is more
effective; however, for offering cool spaces, recreation and biodiversity, parks are very
important. To allow inhabitants to choose themselves what they find pleasant, creating
differences between microclimates (sun and shade) is important in the layout of the city.
 Add trees with large crowns in streets, parks or squares that catch a lot of sun. Considering that
the shade effect of (large) tree crowns plays an important role (in streets and squares, Figure
3.12), the location of trees, types of tree and maintenance policy should be taken into account
specifically during the design phase, in order to place trees as effectively as possible. Therefore,
pay attention to optimum shade and sufficient maintenance (irrigation). Use deciduous trees in
particular: shade in summer, sun and light in winter.