WMO 2023 - Guidance On Measuring, Modelling and Monitoring The Canopy Layer Urban Heat Island

Guidance on Measuring, Modelling
and Monitoring the Canopy Layer

Urban Heat Island (CL‑UHI)
2023 edition
WEATHER CLIMATE WATER
WMO-No. 1292
Guidance on Measuring, Modelling
and Monitoring the Canopy Layer
Urban Heat Island (CL‑UHI)
2023 edition
WMO-No. 1292
EDITORIAL NOTE
METEOTERM, the WMO terminology database, may be consulted at https://public.wmo.int/en/
meteoterm.
WMO-No. 1292
© World Meteorological Organization, 2023
The right of publication in print, electronic and any other form and in any language is reserved by
WMO. Short extracts from WMO publications may be reproduced without authorization, provided
that the complete source is clearly indicated. Editorial correspondence and requests to publish,
reproduce or translate this publication in part or in whole should be addressed to:
Chair, Publications Board

World Meteorological Organization (WMO)
7 bis, avenue de la Paix Tel.: +41 (0) 22 730 84 03
P.O. Box 2300 Email: publications@wmo.int
CH-1211 Geneva 2, Switzerland
ISBN 978-92-63-11292-2
NOTE
The designations employed in WMO publications and the presentation of material in this publication do
not imply the expression of any opinion whatsoever on the part of WMO concerning the legal status of any
country, territory, city or area, or of its authorities, or concerning the delimitation of its frontiers or boundaries.
The mention of specific companies or products does not imply that they are endorsed or recommended by
WMO in preference to others of a similar nature which are not mentioned or advertised.
CONTENTS
Page
PREFACE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
ACKNOWLEDGEMENTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii
EXECUTIVE SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
CHAPTER 1. INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
CHAPTER 2. GENERAL INFORMATION ON THE CL‑UHI. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2.1 Historical background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Horizontal scales and vertical layers in urban areas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.3 Defining the CL‑UHI. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.4 CL‑UHI metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.5 Development of the CL‑UHI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.5.1 Surface influences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.5.2 Interaction of physical processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.5.3 Ideal weather conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.5.4 Influences of actual weather conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.5.5 Influences of city form and geographic setting . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.6 Distinguishing the CL‑UHI from other urban heat island types. . . . . . . . . . . . . . . . . . . . . 12
2.7 CL‑UHI effects on meteorological and climatological conditions . . . . . . . . . . . . . . . . . . . 13
2.8 The CL‑UHI in brief. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
CHAPTER 3. URBAN SERVICES NEEDING CL‑UHI INFORMATION . . . . . . . . . . . . . . . . . . . . . . 15

3.1 Heat stress information as a service – CL‑UHI impacts on health . . . . . . . . . . . . . . . . . . . 15
3.1.1 Thermal comfort assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.1.2 Heat‑health outcomes and early warnings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.2 Air pollutant concentration information as a service – CL‑UHI impacts on
atmospheric composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.3 Energy provision as a service – CL‑UHI impacts of and on energy use . . . . . . . . . . . . . . . 16
3.4 Urban vegetation as a service – CL‑UHI impacts on and from vegetation . . . . . . . . . . . . 18
3.5 Multi‑hazard early warning system – consideration of the CL‑UHI. . . . . . . . . . . . . . . . . . 18
CHAPTER 4. CHARACTERIZATION OF THE URBAN AREA AND ITS SURROUNDINGS. . . . . 19

4.1 Types of parameters needed. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.2 Detailed parameters at microscale. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.3 Characterization at the local scale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.5 Regional topographic characterization relevant for the CL‑UHI. . . . . . . . . . . . . . . . . . . . . 25
4.6 Main aspects for characterizing urban areas in brief. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
CHAPTER 5. DETERMINING THE CL‑UHI FROM OBSERVATIONS. . . . . . . . . . . . . . . . . . . . . . . 26

5.1 Background on measurements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
5.2 Calculation of the CL‑UHI intensity from urban and rural observations. . . . . . . . . . . . . . 26
5.3 Measurement approaches. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.3.1 Single pair of sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.3.2 Traverse approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.3.3 Meteorological networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5.3.4 Opportunistic sensing – crowdsourcing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5.4 Choosing a site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
5.4.1 Selection of rural reference sites. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
5.4.2 Purpose of the urban station. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
iv GUIDANCE ON MEASURING, MODELLING AND MONITORING THE CANOPY LAYER URBAN HEAT ISLAND
(CL-UHI)
Page
5.4.3 Thermal source areas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

5.4.4 Sensor positioning for representative measurements of neighbourhoods. . . . . 32
5.4.5 Vertical positioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
5.5 Sensors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
5.5.1 Instruments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
5.5.2 Radiation shielding and ventilation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
5.5.3 Mounting and measurement protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
5.5.3.1 Fixed sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
5.5.3.2 Sensors on mobile platforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
5.6 Metadata for observations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
5.7 Data transfer and availability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
5.8 Observational challenges in brief. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
CHAPTER 6. DETERMINING THE CL‑UHI WITH MATHEMATICAL MODELS . . . . . . . . . . . . . 38

6.1 Model types. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
6.1.1 Statistical models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
6.1.2 Obstacle‑resolving models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
6.1.3 Numerical weather prediction and climate models. . . . . . . . . . . . . . . . . . . . . . . . 40
6.2 Calculation of a CL‑UHI metric from models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
6.3 Evaluation of model skill . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
6.4 Application of models to determine the CL‑UHI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
6.4.1 Analyses of current and past CL‑UHI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
6.4.2 CL‑UHI forecasts (several days and shorter). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
6.4.3 Climate predictions (sub‑seasonal and longer) and projections . . . . . . . . . . . . . 46
6.4.4 Urban development and the CL‑UHI. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
6.5 Metadata for modelling results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
6.6 Modelling challenges in brief. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
CHAPTER 7. MONITORING THE CL‑UHI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

7.1 Definition of monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
7.2 Purpose of monitoring. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
7.3 Partnering in the installation of monitoring systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
7.4 Monitoring the CL‑UHI using observation networks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
7.4.1 General considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
7.4.2 Monitoring with reference stations in standard instrument shelters. . . . . . . . . . 53
7.4.3 Monitoring with small automatic weather stations in small
instrument shelters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
7.4.4 Mobile monitoring approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
7.4.5 Opportunistic sensing – crowdsourcing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
7.5 Monitoring the CL‑UHI as part of an IUS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
7.5.1 Monitoring CL‑UHI from the past to today. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
7.5.2 Monitoring CL‑UHI into the future. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
7.5.3 Data transfer, archiving and licensing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
7.6 Monitoring challenges in brief. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
CHAPTER 8. UNDERSTANDING THE IMPACTS OF CL‑UHI MITIGATION AND

ADAPTATION EFFORTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
8.1 Change of materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
8.2 Nature‑based solutions – including green and blue infrastructures . . . . . . . . . . . . . . . . . 57
8.3 Adjusting urban form. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
8.4 Limit energy use. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
8.5 Design and planning considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
8.6 Challenges for finding the right measures to mitigate CL‑UHI. . . . . . . . . . . . . . . . . . . . . . 59
CONTENTS v
Page
APPENDIX 1. A CASE STUDY OF INFLUENCES ON CL‑UHI AND OTHER

TEMPERATURES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
APPENDIX 2. EXAMPLES OF OBSERVATION NETWORKS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
APPENDIX 3. MONITORING EXAMPLES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
ABBREVIATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
REFERENCES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
BIBLIOGRAPHY FOR FURTHER READING. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

PREFACE
Urban areas modify the surface energy exchanges in comparison with non‑urban areas
and generally exhibit higher night‑time air temperatures and similar or lower daytime air
temperatures compared with the surrounding rural areas, a phenomenon called the urban heat
island (UHI) effect. The UHI in the lower part of the atmospheric canopy layer, the canopy layer
UHI (CL‑UHI), is directly experienced by humans. The heat burden in cities induced by the UHI is
an additional signal on top of the climate warming that cities experience. Urbanization continues
in almost all WMO Members, thus the additional urban‑induced warming from the CL‑UHI will
increase, becoming an additional burden for the population in cities.
Despite the importance of the CL‑UHI, WMO has not issued guidance on assessing the CL‑UHI.
At its eighteenth session, the World Meteorological Congress emphasized the importance of
this issue and approved Resolution 32 (Cg‑18) – Advancing Integrated Urban Services and
Resolution 61 (Cg‑18) – Integrated and Coordinated WMO Research to Serve Society. These
request WMO technical commissions, the Research Board and other relevant bodies to develop
science‑based technical guidance on the measuring, modelling and monitoring of the CL‑UHI
effect to support Members’ service delivery needs and planning efforts to mitigate the impacts of
CL‑UHI to support the advancement and development of integrated urban hydrometeorological,
climate and environmental services, also referred to as an integrated urban service (IUS).
Resolution 61 (CG‑18) calls for the long‑term expertise of the WMO GAW (Global Atmosphere
Watch) Urban Research Meteorology and Environment (GURME) project as well as the World
Weather Research Programme (WWRP) to be used in this effort. The Congress also urged:
Members to improve connections among [National Meteorological and

Hydrological Services], research institutions, academia, stakeholders and
end‑users of services on a national level to ensure that research responds to
requirements for the development of new and improved services, and that
advances in research are appropriately included in operations;
This connection between research and operations is an example of the “value chain approach”
promoted by WMO.
Developed in response to the request from the World Meteorological Congress, the present
guidance provides an overview of and recommendations for measuring, modelling and
monitoring the CL‑UHI, which is based on temperature information at about 1.5 m above
ground. Other UHIs, such as the boundary layer, surface or subsurface types, have no direct
influences on human health and are therefore only briefly addressed here. CL‑UHI information is
one component of an IUS. These services may be provided directly by National Meteorological
and Hydrological Service (NMHS) operations, in cooperation with stakeholders or partners, or
indirectly through stakeholders or partners in cities or public and private agencies. Concepts
for these services are explained in Guidance on Integrated Urban Hydrometeorological, Climate and
Environmental Services (WMO-No. 1234), Volume I: Concept and Methodology, and Grimmond
et al. (2020). Detailed examples of integrated services in several demonstration cities are
found in Guidance on Integrated Urban Hydrometeorological, Climate and Environment Services
(WMO-No. 1234), Volume II: Demonstration Cities, and in Baklanov et al. (2020).
ACKNOWLEDGEMENTS
The present guidance was developed by an interprogramme working group of WMO GAW
(Global Atmosphere Watch) Urban Research Meteorology and Environment (GURME) as well
as the World Weather Research Programme (WWRP) and includes many contributions from
different countries and institutions.
Coordinating lead authors: K. Heinke Schlünzen (Universität Hamburg, Germany),

Sue Grimmond (University of Reading, United Kingdom of Great Britain and Northern Ireland),
Alexander Baklanov (WMO).
Lead authors: Alberto Martilli (CIEMAT, Spain), Valéry Masson (Météo France), Shiguang Miao
(IUM CMA, China), Chao Ren (The University of Hong Kong), Matthias Roth (National University
of Singapore), Iain D. Stewart (Global Cities Institute, Canada).
Contributing authors: Felix Ament (Universität Hamburg, Germany), Lee Chapman (University
of Birmingham, United Kingdom), Lesley Choo (NEA Singapore), Andreas Christen (University of
Freiburg, Germany), Anurag Dipankar (NEA Singapore), Evyatar Erell (Ben‑Gurion University of
the Negev, Israel), Clare Heaviside (University College London, United Kingdom), Jorge Gonzalez
(City College of New York, United States of America), Guido Halbig (DWD, Germany), Peter
Hoffmann (GERICS, Germany), Meinolf Kossmann (DWD, Germany), Scott Krayenhoff (University
of Guelph, Canada), Stephen Po‑wing Lau (Hong Kong Observatory), Humphrey Lean (UK Met
Office, United Kingdom), Glenn McGregor (Durham University, United Kingdom), Gerald Mills
(University College Dublin, Ireland), Negin Nazarian (University of New South Wales, Australia),
Paulo Saldiva (University of São Paulo, Brazil), Vivek Shandas (Portland State University, United
States), Stefan Smith (University of Reading, United Kingdom), Jianguo Tan (Shanghai Institute
of Meteorological Science, CAMS, China), Oksana Tarasova (WMO), James Voogt (University of
Western Ontario, Canada).
REVIEWERS
Benjamin Bechtel (Ruhr‑University Bochum, Germany), Edmilson Dias de Freitas (University of

São Paulo, Brazil), Paul Joe (ECCC, Canada), Hunter Jones (NOAA, United States), Feng Liang
(Beijing Meteorological Service, China), Juerg Luterbacher (WMO), Yves‑Alain Roulet (Meteo
Swiss, Switzerland), Marcus Thatcher (CSIRO, Australia), Alberto Troccoli (World Energy &
Meteorology Council (WEMC), United Kingdom), Helen Ward (University of Innsbruck, Austria).
EXECUTIVE SUMMARY
The urban heat island (UHI) is one of the earliest documented effects on air temperature
observed in urban areas. It results from differences in the surface energy balance between
urban and rural areas, caused by the interaction of urban surfaces with atmospheric processes.
The focus of this guidance is the canopy layer UHI (CL‑UHI). The CL‑UHI is concerned with
near‑surface air temperatures differences (~1.5 m above ground) between urban and rural
areas. With more than half of the world’s population living in cities, these temperature effects
are directly experienced by many humans and can be linked to changes in human comfort,
public health and human activities. Exposure to high temperatures can increase morbidity and
mortality, especially during heat waves and at night, when urban CL air temperature can be
elevated compared with the rural surroundings. Other types of UHIs, such as the boundary layer,
surface or subsurface types, have less direct influences on human health and are therefore only
briefly addressed in the present guidance.
The scientific background needed to understand the processes creating and influencing the
CL‑UHI is presented in the present guidance, with examples of different agencies and services
that need information about the CL‑UHI. Critical to this is understanding the role of scale and the
link to urban form. Guidance is provided on the parameters required to characterize urban areas
at microscale and local scale, and the ways in which they influence the CL‑UHI are discussed.
The guidance reviews the “ideal” weather conditions (calm, clear days and nights) for
development of large CL‑UHI intensities, and the influences of meteorological factors,
topography and urban features. It introduces different approaches to measure the CL‑UHI
using sensors placed in different locations, for example a pair of sites, traverses, meteorological
networks and opportunistic sensing, with discussion of site and sensor selection criteria.
Similarly, modelling approaches, including statistical, obstacle‑resolving and numerical weather
prediction models, are compared, along with their advantages and limitations for determining
CL‑UHI values. Measurements can support CL‑UHI assessments and enable robust evaluations
of model outputs. Models performing satisfactorily for their intended applications can be used
for simulating CL‑UHIs for current and future scenarios such as city development and climate
change. Monitoring the CL‑UHI requires both observations and modelling. Key to monitoring
are long‑term maintenance and documentation of changes, for example, of site and instruments
(that is, metadata). Mitigation and adaptation efforts to reduce the effects of CL‑UHIs are
discussed, such as efforts to reduce heat stress, air pollutant concentrations and energy use.
The guidance is written for WMO Members, National Meteorological and Hydrological Services
(NMHSs) and their many potential partner agencies and stakeholders undertaking activities in
cities that are impacted by weather and climate across a wide range of time and space scales.
Information on CL‑UHIs is one part of an IUS and may be part of multi‑hazard early warning
systems and high‑resolution weather prediction systems for urban areas.
CHAPTER 1. INTRODUCTION
The present guidance focuses on urban‑induced heat effects. The guidance uses the term
“urban” for populated areas that are built‑up, with increased density of built structures such
as houses, commercial buildings, roads, industrial facilities and other urban forms. Towns, cities
or suburbs are all referred to as urban areas. Non‑urban areas outside the urban area are called
“rural” in the present guidance.
Urban areas have modified surface energy balances and frequently exhibit higher night‑time air
temperatures and lower daytime near‑surface air temperatures compared with the surrounding
rural area, a phenomenon called the urban heat island (UHI) effect. An UHI with negative
intensities – thus exhibiting urban air temperatures below the rural ones – can be referred to as
an “urban cool island”. In the present guidance, the term urban heat island is used to discuss the
phenomenon whether the air temperatures are elevated or reduced.
The UHI in the lower part of the atmospheric canopy layer, the canopy layer UHI (CL‑UHI),
is directly experienced by humans. In some situations, the CL‑UHI effect poses a threat to
public health: more than half of the world’s population now lives in cities, and exposure to
high temperatures can increase morbidity and mortality, especially during heat waves. With
temperature values elevated at night, the CL‑UHI directly influences human well‑being.
Therefore, one major reason to assess the CL‑UHI effect is to improve forecasts provided by
an IUS (Guidance on Integrated Urban Hydrometeorological, Climate and Environmental Services
(WMO-No. 1234), Volume I: Concept and Methodology and Guidance on Integrated Urban
Hydrometeorological, Climate and Environment Services (WMO-No. 1234), Volume II: Demonstration
Cities). In contrast to the CL‑UHI, the surface UHI (S‑UHI) is based on surface temperature
values. It has its maximum during daytime, when humans can better deal with heat through
behavioural changes than at night when trying to sleep; the focus of the present guidance is thus
on the CL‑UHI.
The CL‑UHI is a type of urban‑induced warming first recognized more than 200 years ago by
Howard (1818). It is one of the most evident direct atmospheric signatures of humans’ alteration
of the Earth’s surface. The characteristics of CL‑UHIs differ between cities, within cities, with rural
surrounding type and with time. In addition to their influence on short‑term weather, health
and human comfort, CL‑UHIs have major implications for energy demand and urban climate
adaptation policies. Urbanization continues in almost all WMO Members, and the heat burden in
cities induced by the CL‑UHI is an additional signal that cities experience especially at night – in
addition to warming caused by climate change. Urban planning measures can help to reduce
CL‑UHI intensities, where beneficial.
This guidance provides general information on the CL‑UHI (Sections 2.1–2.5) and on other
types of UHIs (Section 2.6) as well as CL‑UHI effects on the surroundings (Section 2.7).
The Section 2.8 summary provides a general orientation that can be quickly read. Selected
applications are provided in Chapter 3, including their integration in multi‑hazard early warning
systems (Section 3.5). Chapter 4 explains in detail how urban areas need to be characterized to
determine the CL‑UHI, with the main aspects briefly summarized in Section 4.6. Observational
approaches (Chapter 5), use of models (Chapter 6) and monitoring (Chapter 7) are introduced in
detail with their respective challenges summarized in the last section of each chapter. Chapter 8
discusses mitigation measures typically used and their influence on the CL‑UHI, as well as
challenges (summarized in Section 8.6). In Appendix 1 an example application is provided to
highlight how the CL‑UHI is influenced by the rural surroundings. Typical misunderstandings
of the CL‑UHI are clarified in this example application. Examples of observation set‑ups
(Appendix 2) and monitoring (Appendix 3) are provided. The publications cited in the main text
(References) are supplemented with a Bibliography for further reading organized by the different
aspects addressed in this guidance.
2 GUIDANCE ON MEASURING, MODELLING AND MONITORING THE CANOPY LAYER URBAN HEAT ISLAND
(CL-UHI)
Table 1.1 summarizes the structure of the present publication with respect to processes,
applications needing CL‑UHI information, methods for CL‑UHI determination and mitigation of
the CL‑UHI.
Table 1.1. Main elements of this document
Overview of processes Chapter 2:

relevant to CL‑UHI – Urban scales
– Definition of CL‑UHI
– Metrics for CL‑UHI
– Processes related to the genesis and development of CL‑UHI
– Other types of UHI
Appendix 1:
Application example illustrating information of Chapter 2
Applications needing Chapter 3:

CL‑UHI Where CL‑UHI information is needed
Appendix 1:
Application example illustrating influences on CL‑UHI
Methods to determine Chapter 4:

CL‑UHI Characterization of the urban area and its surroundings
Chapter 5:
Observations
Chapter 6:
Modelling
Chapter 7:
Monitoring
Appendix 2:
Observational networks
Appendix 3:
Examples of monitoring CL‑UHI
Mitigation of CL‑UHI Chapter 8:

How to mitigate CL‑UHI effects
CHAPTER 2. GENERAL INFORMATION ON THE CL‑UHI
2.1 HISTORICAL BACKGROUND
Some of the earliest documented urban effects on air temperature were published by Luke
Howard. From 1806 to 1830, Howard analysed data from a small network of thermometers
across London (United Kingdom) and its surroundings, and surmised that London’s climate
was warmer than the surrounding rural areas. He suggested that this warmth was caused by
radiation trapping and airflow interference by building walls, heat release from domestic and
industrial combustion processes, and lack of evaporation from built surfaces. Howard’s account
of the anomaly is remarkably accurate considering the few instruments that were available at that
time. Studies in other cities soon followed Howard’s ground‑breaking work, each with a similar
outcome. In 1897 this inspired Julius Hann, the pioneer of modern meteorology, to formally
label the recurring warmth in large cities the “city temperature” effect. Several decades later, the
thermal anomaly became known as the “urban heat island” effect, owing to the innovative work
of Austrian and German researchers using motorcars to carry thermometers through city streets
and record air temperatures in hundreds of locations. Their approach uncovered thermal spatial
patterns in cities that resembled “islands” of heat when plotted on a map. The patterns were
clearly relatable to the built form of the urban area and were relevant to local concerns about
air pollution, weather forecasting, settlement design and human health. The combined use of
motorcar surveys and climate observatories eventually enabled researchers to study the spatial
and temporal characteristics of heat islands in individual cities.
Scientific understanding of the CL‑UHI has advanced based on this foundation. The controls,
spatial and temporal patterns, as well as the meteorological and societal implications of heat
islands are now well understood (Oke et al., 2017). Assessments and predictions of CL‑UHI effects
can now be made with reasonable accuracy in most cities.
2.2 HORIZONTAL SCALES AND VERTICAL LAYERS IN URBAN AREAS
In the present guidance, horizontal scales are defined as a function of the horizontal extent of
urban form (Table 2.1, column 1). This follows the nomenclature used in several urban climate
publications. For those more familiar with Orlanski’s (1975) definition of horizontal scales for
atmospheric phenomena, we also ‘translate’ the two scales (Table 2.1): with column 2 listing
urban forms and column 6 containing the corresponding horizontal length of atmospheric
phenomena. Urban forms create atmospheric phenomena that may have the horizontal length
of the urban form or its facets, such as a roof vortex, or larger such as a wake downwind of a
building. Thus, the resulting atmospheric phenomena are multi‑scale. The present guidance
uses horizontal scales derived from urban form. When it deviates from this and uses the
Orlanski (1975) scales, this is explicitly mentioned in the text.
The focus of this publication is the UHI in the lower part of the urban canopy layer (UCL),
defined as the air volume between buildings and trees (Figure 2.1a, b). The vertical extent of the
UCL is from the ground to the top of the buildings and trees. The horizontal extent is across the
urban area (10–100 km, Figure 2.1). Critically, this is the region where people are exposed to the
environment, including the effects of the CL‑UHI, both indoors and outdoors.
The urban form has facets consisting of walls, roofs and ground (Table 2.1). Within the UCL,
microscale heating and cooling processes at timescales of seconds to days dominate. In
compact neighbourhoods with closely spaced buildings and small amounts of vegetation,
the surfaces are predominately impervious with bluff bodies having sharp edges. The facet’s
arrangement results in complex radiative interactions. These include shadowing, multiple
reflections, horizon screening by building walls and thereby lowered sky view factors, as well as
(CL-UHI)
Table 2.1. Characteristic scales for urban form based on surface characteristics
Horizontal Vertical Related Atmospheric

Scale Urban form
length (HL) extent parameters phenomena HL scale
Micro Facet (roof, wall, road) 1–10 m UCL Materials Microscale γ
Building 10+ m UCL H Microscale γ
Street, canyon 30–200 m UCL H, W Microscale β
Local Block (bounded by
canyons, interior 300–500 m RSL λp, Hmax, σH Microscale α
courtyards)
Neighbourhood 1–2 km RSL, ISL λp, Hmax, σH Microscale α
Meso Urban area (city centre
Mesoscale γ,
to low‑density residential 10–100 km UBL λp, Hmax, σH
Mesoscale β
areas that are contiguous)
Regional Region (urban and Mesoscale β,
>100 km PBL λp, Hmax, σH
non‑urban surroundings) Mesoscale α
Note: The vertical extent (Figure 2.1) is influenced by the nature of urban form,
where H: building/urban canyon height; W: street/urban canyon width; λp: plan area fraction of buildings;
Hmax: maximum H; σ H : standard deviation of H; UCL: urban canopy layer, RSL: roughness sublayer, ISL: inertial
sublayer, UBL: urban boundary layer, PBL: planetary boundary layer.
Source: Modified from Cleugh and Grimmond (2012) and Oke et al. (2017). Atmospheric phenomena characteristic
scales based on Orlanski (1975).
(a) Microscale (b) Local scale
ISL
SL
RSL
UCL RSL
UCL
(c) Mesoscale
Urban “plume”
Mixing layer
PBL
UBL
b) Rural BL
SL
Rural Urban Rural
Figure 2.1. Vertical extent of atmospheric layers for (a–c) different horizontal scales
of urban form. UCL: urban canopy layer; RSL: roughness sublayer; SL: surface layer;
ISL: inertial sublayer; PBL: planetary boundary layer; UBL: urban boundary layer.
Source: Adapted from Oke (1997).
CHAPTER 2. GENERAL INFORMATION ON THE CL-UHI 5
airflow effects like channelling, blocking, wakes and flow vortices. Buildings, the basic structural
unit of every urban area, are commonly organized along streets, and in densely‑built areas create
“urban canyons”.
The UCL is part of the roughness sublayer (RSL); this is the volume of air that is directly
influenced by individual surface roughness elements like buildings or trees (Figure 2.1).
This layer may extend to 2–5 times the height of the roughness elements, with the spacing
of roughness elements and individual tall buildings having an influence on its depth. The
depth of the RSL further varies with thermal stratification. Within the UCL and RSL, there is a
three‑dimensional (3D) time‑dependent variation in the atmospheric characteristics over short
distances (1–100 m).
Above the RSL there is an inertial sublayer (ISL), or constant flux layer (Figure 2.1). Within this
layer, the influence of individual roughness elements is blended so that there is a small variability
in atmospheric characteristics in the horizontal direction. For an ISL to form, there needs to be an
area of similar fetch below, for example, neighbourhoods with similar building shape, building
layout, construction materials and anthropogenic heat emissions. Atmospheric properties
in the ISL are referred to as having one‑dimensional (vertical) scaling and are local scale
(on the order of 100 m–1 km) (Table 2.1).
Combined vertically, the RSL and ISL make up the surface layer (SL) of the urban boundary
layer (UBL) (Figure 2.1). The UBL, influenced by the rough and heated urban surfaces, has
dynamic vertical dimensions that vary both diurnally and seasonally. On clear days with strong
solar irradiance, the UBL may reach 1–2 km in the afternoon, before falling to 100 m (or less) by
night as a nocturnal inversion forms. In large cities the UBL is regarded as a mesoscale feature
(Table 2.1) that is affected by the geographic setting, for instance orography or large water
bodies (Subsection 2.5.5), and impacts the region as winds can advect the UBL downwind of the
urban area for tens or hundreds of kilometres as an elevated urban plume (Figure 2.1c). All these
features are within the planetary boundary layer (PBL), where all atmospheric processes are
influenced by the surface.
2.3 DEFINING THE CL‑UHI
The CL‑UHI is a microscale to mesoscale atmospheric warming effect associated with cities.
The CL‑UHI is usually estimated from synchronous differences in near‑surface air temperatures
between urban and non‑urban areas. The typical measurement height is about 1.5 m above
ground level (AGL) for both areas. The present guidance uses the term “urban” for areas that
are built‑up with increased density of structures such as houses, commercial buildings, roads,
industrial facilities and city parks. Towns, cities and suburbs are all referred to as urban areas.
Areas surrounding urban environments are called “rural” in the present guidance. Rural areas
may include both natural areas, where human changes to the landscape are minimal or not
evident, or anthropogenically modified areas, such as agriculture and forestry areas. Rural areas
can also include water bodies.
Ideally, the CL‑UHI would be based on near‑surface temperature differences at the same place
with and without urbanization. However, observed values would be separated in time and
therefore include the climate signal, making the true urban effect difficult to isolate. Modelling
studies can simulate urban and pre‑urban effects; however, it is difficult to determine a pre‑urban
case (Section 6.2). Thus, pragmatically, the urban effect is estimated from the difference in
simultaneously observed or simulated temperatures at carefully selected urban and rural sites
within the region of interest. As the strength of the CL‑UHI varies with the urban area, it is
important to consider this as well as the upwind rural surroundings that are the reference.
In the urban area, the near‑surface canopy layer temperature (CL‑T) is taken between the
roughness elements (for example, buildings and trees), typically at a height of 1.25–2 m AGL
(Subsection 5.4.5). If the rural environment consists of a large grass area, the air temperature is
(CL-UHI)
measured within the ISL and not in the very short rural canopy layer. Thus, the CL‑UHI refers to
the near‑surface air temperature difference representing different sizes and patterns in thermal
source areas (Subsection 5.4.3).
The CL‑UHI intensity is normally defined as a synchronous air temperature difference between
one or more urban and rural measurement sites (selection of sites is discussed in Chapter 5).
Commonly a temperature difference (ΔTu–r) is measured between the urban (u) location with
the slowest cooling (which might be the city centre) and a location in the rural (r) surroundings,
to capture the maximum urban thermal effect relative to a place where urbanization is
absent (“Maximum” metric, Table 2.2, part (a)). This will not account for variations in types of
urbanization or changes and heterogeneities in the rural setting. An approach that identifies
“hot” and “cool” spots will capture intra‑urban thermal differences, providing information
on local‑scale variations (“Spatial pattern” metric, Table 2.2, part (b)) influenced by surface
properties. Similarly, using several rural sites and determining average rural values helps to avoid
biases resulting from microscale effects at a selected rural site (Section 5.2). Note, the CL‑UHI
intensity is a temperature difference and is not indicative of the actual air temperature at either
location (urban or rural).
Table 2.2. CL-UHI metrics based on ΔTu–r, use cases with

(a) conditions of likely occurrence and (b) main influences
(a) CL‑UHI metric Use case Conditions of likely occurrence

Minimum AC, CM, EU, During midday hours in urban centres with shaded deep
HIC, R, WF urban canyons or intense vegetation
Irrigated urban parks with dense vegetation
Cities surrounded by arid or semi‑arid rural areas that

warm rapidly after sunrise
Maximum AC, CM, EU, HIH, At night under ideal weather conditions (Subsection 2.5.3)
R, UDP, WF
Cold climates with high anthropogenic heat emissions any
time of day
(b) CL‑UHI metric Use case Main influences
Daily mean CM, EU, EV, HI, Weather situation, time of year, urban form
R, UDP, WF
Seasonal mean AA, CM, EU, EV, Leaf on/leaf off, anthropogenic heat emissions, sky view
HI, R factor, soil moisture, synoptic conditions, urban form
Annual mean AA, CM, EU, EV, Urban form, surface materials, population, urban
R, UDP areal extent, regional climate, inter‑annual variability
of weather, anthropogenic emissions (depending on
cool/warm winters, quality of building insulation)
Spatial pattern: AA, AC, CM, EU, Urban form, population, urban areal extent, terrain,
use the other metrics EV, HI, HIC, HIH, surface materials, land use
at several sites R, UDP, WF
Note:
AA – agricultural applications such as urban horticulture, city parks and green spaces (Section 3.4);
AC – atmospheric chemistry (Section 3.2);
CM – climate monitoring (Section 3.5);
EU – energy use (Section 3.3);
EV – ecology and vegetation, for example, assessing earlier spring and later autumn onset compared to rural phenology
(Section 3.4);
HI – health impact (Section 3.1);
HIC – health impact cold, such as cold waves and extreme weather (Section 3.1);
HIH – health impact hot: emergency response, for example, heat stress and heat warning due to high temperatures
(Section 3.1);
R – research;
UDP – urban design and planning (smart city applications) (Chapter 8);
WF – weather forecasting (Section 3.5).
2.4 CL‑UHI METRICS
Table 2.2 provides a list of CL‑UHI metrics and corresponding use cases. The maximum CL‑UHI
intensity may be of interest to those concerned with public health and heat stress at night
(Section 3.1). The highest reported value in any city is about 12 K, assuming larger‑than‑normal
anthropogenic heating is not a factor (Oke et al., 2017). Minimum values can be small but
positive, and negative values can also occur, indicating an urban cool island, where the urban
location is cooler than the rural one.
Daily, seasonal or annual means include a time definition (day, season, year), but time averaging
can be done, for example, over several days or over several years. Similarly, the minimum or
maximum metrics might be used to determine average values for specific synoptic situations
in order to better understand the impact of the climate or weather patterns on the CL‑UHI
development (Subsection 2.5.4). Average intensities are always smaller than maximum intensities
measured under ideal weather conditions for individual nights (Subsection 2.5.3). For any metric,
the temporal averaging (or sampling method) needs to be specified to properly interpret the
intensities.
Each metric can be applied to one site or to multiple sites. Using multiple sites provides
information on the spatial pattern. However, averaging the CL‑UHI intensity over several urban
sites will reduce the CL‑UHI intensity for the urban area. Therefore, not only the temporal but also
the spatial averaging needs to be specified and documented in each application. The metric and
averaging to apply will differ depending on the end‑user applications (Chapter 3).
2.5 DEVELOPMENT OF THE CL‑UHI
2.5.1 Surface influences
The physical causes of the CL‑UHI are well known. As urban populations grow, more buildings
are erected, often replacing vegetated landscapes that include meadows, forests or fields with
3D impervious surfaces that store heat due to their thermal properties, emit heat associated
with human activities, and, in general, alter the surface energy exchanges. The CL‑UHI is
fundamentally a thermal anomaly that results from differences in the energy balance in urban
and surrounding rural environments. In the UCL, the relevant energy exchanges occur at
different heights within the canopy volume (Figure 2.1). These urban‑induced changes to
the energy balance compared to rural areas result in differences in cooling and heating rates
between the two areas. This creates the distinct thermal environment of urban areas. The
alterations to the radiative, thermal, moisture and roughness properties in cities, together with
emissions of heat and pollution from human activities, give rise to the four main causes of the
CL‑UHI (Oke, 1982):
– Thermal properties: Building materials often have greater heat capacity than open
vegetation and store heat during daytime that is released at night.
– Surface state: Surface waterproofing by buildings and paving reduces subsurface

moisture storage and directs available energy into surface heating rather than evaporation.
Immediately after rain events, there might be a cooling effect for surfaces and air due to the
evaporation of intercepted water.
– Surface geometry: The total area of the urban surface (horizontal, vertical and sloped
facets) is increased relative to that of the non‑urban surface due to its corrugated form. This
increase results in: (i) reduced average wind speeds, (ii) multiple reflections and greater
absorption of solar radiation, (iii) increased surface area for heat storage, (iv) reduced
(CL-UHI)
net energy loss to the atmosphere via longwave radiation in densely built‑up areas due to
horizon screening during daytime (increased loss at night‑time), and (v) reduced heat loss
by convection and advection in the sheltered UCL.
– Anthropogenic heat: The release of heat from fuel combustion and electricity use is
much greater in urban areas compared to rural areas.
2.5.2 Interaction of physical processes
The 3D structure (geometry) of the urban surface and the high thermal mass of the materials are
the main causes of the nocturnal CL‑UHI. The high level of roughness of urban surfaces reduces
the wind speed in urban areas (Subsection 2.5.1); thus, the sensible heat fluxes are smaller than
in less rough rural areas. In addition, sealed surfaces with less vegetation combined with lower
wind speeds than in the rural surroundings also reduce the latent heat fluxes. Thus, the surface
energy budget is mainly determined by radiative fluxes and heat storage/release by the 3D urban
forms (Table 2.1). Wind speed is a critical influence on heat fluxes and therefore surface–air
exchange. In general, stronger wind speeds result in smaller CL‑UHI intensities.
After sunrise, rural areas warm rapidly because they are fully exposed to the sun when they
have a large sky view factor. The warming is concentrated in a shallow unstable layer of air
near the surface, capped by the remnants of a nocturnal inversion. Warming is slower in urban
areas because the trees and buildings shade large areas of the UCL and have smaller sky view
factors. Thus there is less direct solar radiant energy but more diffuse radiation across the
numerous 3D surfaces. This energy is conducted into the urban mass. The nocturnal cooling
of the 3D surfaces and near‑surface air is comparatively slower in the urban canyons because
of reduced sky view factors and the heat stored in the urban fabric. This results in near‑surface
air temperatures remaining higher in urban areas at night compared to more rapidly cooling
surrounding rural areas. The lower rural temperatures result in a more stable atmosphere at
night in these areas than in urban areas, which can stay unstable. This impacts air pollutant
concentrations (Section 3.2). The CL‑UHI intensity rapidly decreases to a minimum near sunrise
or even becomes negative a few hours after sunrise.
In some cases, especially for urban areas surrounded by arid or semi‑arid landscapes, the
lag period in near‑surface urban air warming is sufficiently long to reverse the thermal
anomaly. In addition, these cities might have more vegetation (or green infrastructure) than
the rural surroundings, which might change the surface energy balance in favour of more
evapotranspiration than in the rural surroundings, such that the urban area is cooler than the
rural surroundings during the middle of the day. This is called an urban cool island (see ”Soil
humidity matters” example in Appendix 1).
As solar day length varies throughout the year, the times that exhibit the greatest differences
in heating/cooling rates shift. Therefore, comparative analysis of heating/cooling rates
must be normalized relative to solar day length (sunset, sunrise). In addition, the synoptic
conditions that favour or suppress CL‑UHIs vary with season. Generally, CL‑UHI intensities
are greatest in summer months for mid‑latitude cities, and in dry months for tropical cities
(Roth, 2007). This pattern results from the important driving forces of solar radiation, soil
moisture and synoptic‑scale weather patterns (wind, cloud, precipitation). Seasonal variations in
anthropogenic heat inputs, most notably space heating in cold seasons or air conditioning in hot
seasons, also increase CL‑UHI intensities. An example of the influence of latitude and temperature
is given in “Latitude matters” example in Appendix 1.
The intensity of the CL‑UHI is often approximated by in situ measurements of air temperature
using sensors positioned at or near standard screen‑level height of 1.25–2 m above the ground
(see Guide to Instruments and Methods of Observation (WMO-No. 8), Volume I: Measurement of
Meteorological Variables, Section 2.1.4.2.1) at multiple sites within and around the urban area
(for observational set‑ups see Section 5.3). When the resulting isotherms are mapped, they
resemble the altitude contours of an island, with variations corresponding to urban land cover,
building morphology, human activity and surface relief characteristics (Figure 2.3).
2.5.3 Ideal weather conditions
These factors (Subsections 2.5.1 and 2.5.2) all contribute to the formation of a CL‑UHI. The
fundamental processes can be most easily examined under “ideal” weather conditions when the
CL‑UHI intensity is maximal. These are calm, clear days with strong daytime insolation followed
by calm, clear nights.
With very low wind speeds, turbulent mixing is weak and thus sensible and latent heat fluxes are
smaller throughout the day than for clear, slightly windy days. To achieve energy balance, heat
loss to the atmosphere must be matched by heat conduction from the surface mass. In these
“ideal” circumstances, the CL‑UHI intensity primarily depends on the comparative radiative
and conductive heat fluxes in urban and rural environments. On calm and clear nights, surface
cooling is driven by radiation loss, with the surface‑emitted longwave radiation exceeding that
received from the atmosphere. The radiative processes are influenced by the sky view factor,
which is reduced by trees and building walls. The conductive heat losses are regulated by the
properties of the surface fabric such as thermal admittance, which is greater for urban materials.
Maximum CL‑UHI intensities develop under ideal weather conditions as defined above.
A corresponding diurnal cycle is illustrated in Figure 2.2. Urban and rural areas might exhibit
similar daytime air temperatures but distinctly different nocturnal temperatures, with higher
values in the urban area (Figure 2.2a). The formation of the nocturnal CL‑UHI is associated
with a strong cooling rate in rural areas around sunset (Figure 2.2b). The cooling rate in urban
areas varies with building density. Differences in evening cooling are magnified when rural
areas are dry (such that thermal admittance is low) and heat withdrawal from the subsurface is
limited. While the cooling rates are distinct at sunset (SS), the rates start to converge as the night
progresses (Figure 2.2b). This results in a characteristic temporal pattern in the CL‑UHI intensity,
with significant growth in the early evening right after sunset, followed by a relatively constant
period through the middle of the night (Figure 2.2c).
2.5.4 Influences of actual weather conditions
The ideal weather conditions resulting in maximum CL‑UHI intensities (Subsection 2.5.3)
rarely occur. The actual intensity and time/space characteristics of the CL‑UHI are controlled by
many interdependent factors related to actual weather, as well as city size, surface properties,
solar zenith angle and geographic setting. The most relevant meteorological variables for the
intensity of the CL‑UHI include near‑surface wind speed and humidity as well as cloud cover and
precipitation.
Wind speed and cloud cover are both inversely related to CL‑UHI intensity. As wind speeds
increase (if skies are clear and surface moisture is available), sensible and latent heat fluxes
will increase because of mixing of the air between the surfaces. Eventually the increasing
wind speed reduces both the near‑surface air temperatures and urban–rural air temperature
gradients. The wind speed at which the CL‑UHI intensity falls to below 1 K differs by city and
the prevailing atmospheric conditions, such as cloudiness and precipitation. Regional wind may
displace warmer isotherms from the city centre downwind. The atmospheric stability within
the UCL, UBL and rural PBL all play a role in vertical heat exchange and thus influence CL‑UHI
development. Cloud type and cloud amount control the CL‑UHI intensity by modifying radiative
receipt and cooling. Clear skies at night permit stronger radiative cooling of rural than urban
air, while a whole day with overcast skies suppresses these differences. Cloud type and cloud
altitude moderate this, with high‑altitude clouds (for example cirrus) impacting the radiative
heat transfer less locally than low‑altitude clouds (for example cumulus).
Since the meteorological conditions influence the CL‑UHI development, synoptic‑scale weather
patterns are also important. Weather situations that quickly change are frequently linked to
higher wind speeds that reduce the intensity of the CL‑UHI due to increased turbulent mixing.
In addition, rainfall and snowfall patterns modify soil moisture and thereby water availability for
evapotranspiration. The spatial and temporal distribution of dry/wet soils across an urban–rural
area influences the diurnal and seasonal expressions of the CL‑UHI.
(CL-UHI)
(a)
Urban
–
TU
Air temperature
–
TR
Rural
Heating/cooling rate
(b)
Urban
0
Rural
Heat island magnitude
ΔTU-R
(c)
SS SR
Middle Middle Middle

of day of night of day
Time (h)
Figure 2.2. Diurnal variation at an urban (brown) and a rural (green) site under favourable
meteorological conditions for a CL‑UHI (clear skies, high solar radiation, light winds),
(SS = sunset; SR = sunrise):
(a) ~2 m air temperatures with daily mean for urban (dashed black, TU ) and rural (black, TR )
site; (b) heating/cooling rates; and (c) CL‑UHI intensity (red, ΔTU–R).
Vertical scale units are approximately (a, c) 2 K and (b) 2 K h –1.
Source: Oke et al. (2017). Reproduced with permission of the licensor through PLSclear.
2.5.5 Influences of city form and geographic setting
A third set of controls on the CL‑UHI relates to the size of a city, defined by its form and
function (Table 4.1) and its geographic setting. Cities with urban areas that have extensive
paving and tightly spaced tall buildings tend to exhibit strong CL‑UHI intensities in those
areas (Figure 2.3a, b). The extent of the urban area influences the CL‑UHI given the potential
sources of heat for advection across the urban area. Heat, moisture and pollution emissions
from human activities vary with socioeconomic and cultural practices both in and out of the
city (for example use of irrigation). Industrial areas with many factories and little or no green
spaces may experience large CL‑UHI intensities. Residential populations may live in tall towers or
low‑rise, low‑density buildings. Population density tells us about the likely urban form, although
it is not helpful for the central business district. Population density alone also does not provide
information about how city form might change as population grows or shrinks (Oke, 1981).
For example, a larger population does not necessarily increase the urban extent, building
density, vertical extent or reduce the pervious surface cover.
(a) (c)
Built-up area B
City core Lago de
Texcoco
Park +6
d
win +4
i g ht A
L
(b)
Day
T0
5K UHISurf ⟫ UHIUCL
Temperature
Ta
Night
Ta UHISurf ∼ UHIUCL
5K
T0
Light wind City core
Park
Rural Built-up Rural 5 km
A B
Figure 2.3. Isothermal maps during clear, calm weather.

(a) CL‑UHI intensity for a “typical” city with flat terrain 2–3 h after sunset,
with (b) cross-section (A → B in (a)) of building density and temperature.
(c) Air isotherms (thick lines, °C) for Mexico City in the early morning (8 Feb. 1972)
at minimum daily temperature. Urban areas shown by shading.
Source: (a), (b) Oke et al. (2017). Reproduced with permission of the licensor through PLSclear. (c) Oke et al. (2017)
based on data from Jáuregui (1973). Reproduced with permission of the licensor through PLSclear.
The topographic setting of a city influences the intensity and spatial characteristics of the CL‑UHI.
Urban areas in valleys are often warmer during day than their more elevated rural surroundings,
since temperatures are generally warmer at lower altitudes above sea level (ASL). This influences
the calculated CL‑UHI intensity, which is based on urban–rural temperate differences.
Therefore, CL‑UHI calculations should preferably be based on values taken at the same altitude
(Section 5.2).
Mesoscale circulations induced by the geographic setting (valley, basin, coastal, mountain)
influence wind fields, causing for instance cold air currents, land–sea breezes or downslope
winds. These pre‑settlement circulations influence the surface energy budget. Their development
is influenced by the urban area, and they may weaken or strengthen the CL‑UHI depending
on their diurnal and seasonal cycles, the thermal characteristics of the winds and possibly
associated cloudiness, as well as the city’s location upwind or downwind of major orographic
and topographic features (Wanner and Filliger, 1989).
Modest terrain can create upslope flows during the day and downslope flows at night. Cold
pools are often formed close to the valley floor during the night by the accumulation of the cool
air flowing down the slopes (katabatic flow), but dense urban structures might block these flows.
These orography effects can be mixed with the urban‑fabric‑induced temperature change and
affect the CL‑UHI intensity. An example is given in Figure 2.3c for Mexico City. During night‑time,
under clear‑sky conditions, cold‑air katabatic flows cool the north and west of the city, as shown
by the spatial pattern of isotherms at the perimeter of the urban area.
Cities located at higher altitudes may experience larger CL‑UHIs as there is more daytime
heating of the urban area from the clearer skies (lower optical thickness). If a nocturnal thermal
inversion occurs in these elevated areas, the CL‑UHI may be enhanced when calculated based
on measurements made at rural stations situated at lower altitudes. Apatity, Russian Federation
(~60 000 inhabitants) provides an example. During winter, a CL‑UHI of ~10 K was observed
(CL-UHI)
in the city, which is located on a small hill with the rural station ~50 m lower at a frozen lake.
About 50% of the CL‑UHI was attributed to anthropogenic heating based on model studies
(Varentsov et al., 2018).
The CL‑UHI intensity is also impacted by water bodies such as lakes and rivers in the urban area.
The daytime CL‑UHI can be reduced close to the water body by onshore advection of cool air, but
the large thermal mass of water keeps nocturnal water temperatures more constant, resulting
in a high CL‑UHI intensity at night close to the water body (Schlünzen et al., 2010). The CL‑UHI
in coastal cities can be strongly modulated by sea/lake breezes. Daytime anthropogenic heat
emissions may be transported downwind, causing the peak CL‑UHI to be displaced from the
densest parts of the city, often historically near the shore, to inland suburban areas. Sea breeze
fronts can travel several tens of kilometres inland. The interaction of the moister sea breeze air
with the urban thermal structure can initiate thunderstorms as in São Paulo, Brazil (Vemado
et al., 2016), or increase rainfall as in Singapore (Doan et al., 2021), which in turn may reduce the
CL‑UHI because of storage heat loss associated with runoff and evapotranspiration. Similarly,
sea breezes can cause clouds (for example in Tokyo or London) over the urban area and increase
night‑time CL‑UHI intensities if the rural area remains less cloudy. However, while coastal areas
remain cooler than the inland urban regions by day, if the inland movement of the sea breeze
front is hindered by the urban roughness, the opposite occurs at night, as found in Mumbai,
India (Maral and Mukhopadhyay, 2015).
The city latitude is an important determinant of the annual and seasonal intensities of solar
radiation received depending on the regional climate (in particular the amount of cloudiness).
This influences the heat that can be stored in the urban fabric, the demand for energy to heat and
cool buildings, and, while also considering socioeconomic conditions, release of anthropogenic
heat. Given the greater frequency of cloud and rain in tropical latitudes (excluding deserts), and
to some extent different city forms in these regions (high albedos, traditionally lower building
heights), it is generally observed that tropical cities experience smaller CL‑UHI intensities
than temperate cities (Urban Climatology and its Applications with Special Regard to Tropical Areas:
Proceedings of the Technical Conference (WMO‑No. 652); Roth, 2007).
2.6 DISTINGUISHING THE CL‑UHI FROM OTHER URBAN HEAT ISLAND TYPES
The CL‑UHI is a measure of the temperature response of the UCL atmosphere, and it can be
attributed mostly to changes in the energy exchanges at the surface due to the characteristics of
the urban environment (Subsection 2.5.2). These characteristics also give rise to other types of
UHI, including the following (Oke, 1995):
– Urban boundary layer UHI (UBL‑UHI): The air above an urban area in the UBL
(Figure 2.1) is warmer compared to air at a similar height above the rural area.
– Surface UHI (S‑UHI): S‑UHI is defined as the heat island resulting from the surface
temperature difference between urban and rural areas. Sunlit urban surfaces (notably
ground, roofs and walls) are mostly warmer during the day than rural surfaces in the
surroundings. At night the urban surfaces are in many cases warmer than rural surfaces
(see “Large CL‑UHI is not necessarily related to high urban temperatures” example in
Appendix 1). In cold climate cities, the S‑UHI increases the number of frost‑free days, and
snow is less likely to remain on the ground. The S‑UHI does not represent the near‑surface
air temperature (Stewart et al., 2021).
– Subsurface UHI (SS‑UHI): SS‑UHI is defined as the heat island resulting from the
subsurface (in the ground) temperature difference between urban and rural areas. Urban
substrate materials (for example soil, rock, groundwater, and city infrastructure such
as sewers, subways and building foundations) are often warmer than those in the rural
surroundings over long time periods. An example of SS‑UHI impacts in Arctic cities is the
change in depth of the permanently frozen soil, causing a mass movement of soil in urban
areas that can weaken the infrastructure foundation.
Each UHI type is closely linked to the vertical structure of the urban area and the atmosphere
above it. These other UHI types have spatial and temporal expressions different from those of the
CL‑UHI. Some are less studied because of the difficulty of obtaining the observations needed to
measure them, for instance above the canopy layer (UBL‑UHI) and below the surface (SS‑UHI).
Observations of the S‑UHI have increased with the availability of satellite and aircraft‑based
infrared sensors, but often using brightness temperatures because of the challenge of obtaining
all the surface emissivities.
The different UHI types share the same fundamental energy balance origins. Differences between
the types reflect the scale and medium (air/solid) under investigation. The S‑UHI responds nearly
immediately to insolation and quickly reacts to changes in cloud cover, for example. If cloudiness
differs at the urban and rural sites, the cloudiness effect is also part of the S‑UHI signal. Apart
from that, the characteristics of each facet (Section 2.2) determine the surface temperatures and
thus the S‑UHI. As it is based on the difference in the temperatures of surfaces (for instance at
top of the trees), the values are not directly relevant for human health or urban planning. The
CL‑UHI response to the surface temperature depends upon molecular diffusivity and convection
to transport heat from surfaces to the near‑surface atmosphere and the atmospheric turbulence
close to the surfaces. The CL‑UHI blends contributions of proximate surfaces and anthropogenic
heating/solar energy removal within the UCL with contributions advected and turbulently mixed
from microscale, local scale and larger scales.
Understanding these distinctions between CL‑UHI and S‑UHI is critical in evaluating policies that
are designed to mitigate the CL‑UHI (Chapter 8). Often, satellite‑derived S‑UHI is confused with
the CL‑UHI based on near‑surface air temperature differences, despite the fact that they represent
different thermal responses at different spatial and temporal scales. The CL‑UHI should be viewed
as one expression of the urban thermal effect relevant for several applications (Chapter 3).
Addressing its causes and mitigating its effects (Chapter 8) will affect the urban‑induced
warming and impact the development of the other UHI types.
2.7 CL‑UHI EFFECTS ON METEOROLOGICAL AND CLIMATOLOGICAL

CONDITIONS
When there are light regional winds, the heat exchange from surfaces including roof and wall
facets (this is the net contribution of wall and ground surfaces to the atmosphere) can initiate a
thermodynamic circulation called the “country breeze”. It is caused by lower air pressure above
the warmer urban area. This leads to updrafts in the warm centre, resulting in a horizontal flow
convergence there, and return flows in higher layers to the rural surroundings with subsidence
there. The circulation is potentially important to the CL‑UHI as a self‑regulating mechanism
moderating its intensity: urban temperatures are limited by the inflow of cooler air from the
surroundings, while rural cooling rates are partly compensated by the subsidence of warmer
urban air from aloft (Haeger‑Eugensson and Holmer, 1999).
The updrafts created by an intense CL‑UHI may increase convection in the overlying air and
contribute to cloud development, lightning strikes and extreme precipitation. The upper UBL
often contains higher concentrations of pollutants, many emitted from within the UCL. These
might provide additional cloud nuclei.
Radiative fog formation occurs under conditions that also favour high CL‑UHI intensities. The
slower urban cooling rate means cities are more likely to remain fog‑free under such conditions,
and with the warmer temperatures the urban relative humidity is likely to be less than in the
rural surroundings. The warmer nocturnal urban temperatures reduce the spatially variable
dew formation on surfaces. Reduced urban dewfall may be a contributor to the observed
nocturnal urban moisture excess, where absolute humidity shows increases at night as in
London (Lee, 1991) or Belgrade (Unkašević et al., 2001), unlike relative humidity. Anthropogenic
emissions of water vapour, from combustion processes, air conditioning, or evaporation after
street cleaning or irrigation, also contribute to the moisture excess in cities.
(CL-UHI)
The CL‑UHI creates a longer growing period and earlier greening of trees in the spring
along with a later leaf fall. The UBL‑UHI contributes to mixing emissions into a larger volume
of air than in the rural areas, especially at night, and the elevated temperatures influence
temperature‑dependent chemical transformation of pollutants. For more details see Chapter 3.
2.8 THE CL‑UHI IN BRIEF
The CL‑UHI is a microscale to mesoscale atmospheric warming effect associated with the
characteristics of the urban environment. It describes near‑surface air temperatures that are
higher in urban areas than in the surrounding rural areas, particularly at night. It is one of four
UHI types that also include the boundary layer UHI (UBL‑UHI), the surface UHI (S‑UHI) and
the subsurface UHI (SS‑UHI). The CL‑UHI is observed in the airspace between buildings and
extends to the height of building roofs and treetops (the urban canopy layer – UCL). The present
guidance focuses on the CL‑UHI at near‑surface levels (1.5 m).
When mapped across the urban landscape, isotherms take the form of an “island” with values
increasing from the outskirts towards the centre of the urban area. The CL‑UHI intensity is
greatest under clear skies and with weak winds following a short dry period; the highest
intensities typically last for several hours after sunset. The energetic basis for its formation
is as follows:
– Urban fabric permits greater storage of heat during the day which is released at night.
– Urban surface geometry obstructs the sky view by its complex 3D form, thus impeding
ground cooling by multiple wall and street facets within the UCL.
– Urban imperviousness reduces subsurface water for evaporation and directs available
heat into sensible rather than latent heat energy exchanges.
– Anthropogenic heat adds an additional source to the CL heat energy balance resulting
from human activities related to the industries, buildings and transportation.
– The meteorological situation influences exchange processes at surfaces (especially heat

and radiative fluxes).
CHAPTER 3. URBAN SERVICES NEEDING CL‑UHI INFORMATION
Information on current and future CL‑UHIs helps to improve forecasts provided by an

IUS (Guidance on Integrated Urban Hydrometeorological, Climate and Environmental Services
(WMO-No. 1234), Volume I: Concept and Methodology and Guidance on Integrated Urban
Hydrometeorological, Climate and Environment Services (WMO-No. 1234), Volume II: Demonstration
Cities). The forecast, either itself or in combination with a statistical model (Section 6.4), can
consider the CL‑UHI in the total temperature, CL‑T, to be provided to citizens. Cities, mostly
in cooperation with NMHSs, provide several services to their citizens, including those linked
to health (Section 3.1), air pollution (Section 3.2) and energy use (Section 3.3). The latter also
impacts the CL‑UHI, for example through anthropogenic heat emissions, while the atmospheric
composition has significant implications for human health (WHO, 2021). The CL‑UHI can
modify vegetation’s phenology (Section 3.4), which influences the health of those sensitive
to pollen, while vegetation in turn impacts the CL‑UHI intensity (Chapter 2). The CL‑UHI may
also be important in early warning systems for extreme weather and emergency management
operations (Section 3.5). Depending on the application, one or many metrics (Table 2.2) may
be relevant.
Newly planned urban forms and designs can intentionally or unintentionally impact the CL‑UHI
intensity. Assessments can be made in advance by using models (Chapter 6). Dedicated urban
planning and design may mitigate CL‑UHI intensities (Chapter 8).
3.1 HEAT STRESS INFORMATION AS A SERVICE – CL‑UHI IMPACTS ON HEALTH
Elevated heat stress is particularly problematic for vulnerable groups, for example children,
adults with certain pre‑existing medical conditions and outdoor workers (Heatwaves and Health:
Guidance on Warning-System Development (WMO-No. 1142)). As noted (Chapter 2), the CL‑UHI
characterizes the urban temperature deviation compared to the rural temperature signal, with
higher temperatures in urban areas, especially at night. Maximum CL‑UHI values are a good
metric to assess the intensity of nocturnal extremes (Table 2.2). This can be used in combination
with forecasts to determine night‑time CL‑T (Section 6.4).
Human thermal comfort is estimated by many thermal indices that combine various
environmental factors (Subsection 3.1.1). The WMO Commission for Weather, Climate, Water
and Related Environmental Services and Applications (SERCOM) Study Group on Integrated
Health Services, together with the World Health Organization (WHO), is considering the impacts
of heat stress on health, and the two organizations have provided guidance on warning‑system
development (Heatwaves and Health: Guidance on Warning-System Development (WMO-No. 1142)).
See also Subsection 3.1.2.
3.1.1 Thermal comfort assessment
Elevated skin and body temperatures, heart rate or sweating are all manifestations of heat strain
arising from heat stress. Heat‑related illnesses, in increasing order of severity, include heat rash,
heat oedema, heat syncope, heat cramps, heat exhaustion and life‑threatening heat stroke. Even
moderately high temperatures are known to be associated with increased all‑cause population
mortality (Basu, 2009). However, although studies have linked mean, minimum and maximum
CL‑T at different timescales with mortality and morbidity in cities, the effects of temperature
extremes in the urban canopy on heat‑ and cold‑related mortality also depend on non‑climate
factors such as the demographic profile of each city. This is considered in thermal indices that
depend on a combination of environmental factors, including temperature, humidity, ventilation
(wind), and solar and longwave radiation, as well as personal factors such as metabolic heat,
age and clothing insulation. Numerous thermal comfort‑related bio‑meteorological indices with
different complexity exist (see reviews such as de Freitas and Grigorieva (2017), and Fischereit
and Schlünzen (2018)).
(CL-UHI)
Daily and monthly CL‑UHI maps vary across a city (use cases HI, HIC and HIH in Table 2.2).
When combined with background temperature data and social and economic characteristics,
spatial heat‑risk maps can identify potentially vulnerable groups. Heat‑related mortality can be
estimated using time‑series and epidemiological methods. Furthermore, recommendations can
be given regarding places where citizens can go to cool off.
As exposure to wind and mean radiant temperature can vary over metres, using the CL‑UHI
(or CL‑T) as a proxy for thermal comfort in an urban canopy might be misleading. However, the
heterogeneity of the CL‑T may influence indoor air temperatures, particularly in poorly insulated
buildings, contributing to indoor overheating and risk.
3.1.2 Heat‑health outcomes and early warnings
The CL‑UHI is only one contributor to heat stress in urban areas. However, as the CL‑UHI is a
nocturnal phenomenon it may result in poor sleep quality, with consecutive hot nights causing
greater health risks than hot days. Therefore, monitoring and providing CL‑UHI information
is important in early warning systems for predicting and informing the public about heat
waves and heat stress events in urban areas (Heatwaves and Health: Guidance on Warning-
System Development (WMO-No. 1142); Multi-hazard Early Warning Systems: A Checklist; Guidance
on Integrated Urban Hydrometeorological, Climate and Environmental Services (WMO-No. 1234),
Volume I: Concept and Methodology).
3.2 AIR POLLUTANT CONCENTRATION INFORMATION AS A SERVICE –

CL‑UHI IMPACTS ON ATMOSPHERIC COMPOSITION
With increasing urbanization, a larger proportion of the world’s population is exposed to

concurrent urban environmental risks such as overheating and poor air quality that may have
independent or synergistic impacts. During heatwaves with high air pollution, mortality rates
increase in cities; therefore, the combined effects need to be considered. Anthropogenic heat
fluxes (Section 3.3) modify the CL‑UHI, and the urban microscale to local‑scale heterogeneity
can increase turbulent mixing within the UCL and above, both increasing the UBL height. With
deeper UBLs the near‑surface concentrations of constituents like aerosols are reduced. The
CL‑UHI may cause modifications of the physical (for example moisture, wind, temperature,
aerosols), biological (for example pollen, viruses, spores, plant diversity) and chemical
(for example air quality, particle formation) state of the urban air. These in turn can impact the
health of the citizens. A higher CL‑T will accelerate air chemical processes such as night‑time
chemical reactions and daytime photochemistry. For example, the rate constants of the
chemical reactions responsible for the formation of ozone increase with air temperature and
solar radiation. These processes are indirectly impacted by the CL‑UHI since the urban fabric
influences the CL‑T and shortwave radiation pattern in a city. The CL‑UHI in turn influences the
circulation (see description of “country breeze”, Section 2.7).
The relations between the CL‑UHI and pollutant concentrations are complex and pollutant
dependent. Therefore, it is important to include information on CL‑UHI values in assessments
of air quality as well as in any mitigation plan. Changes in vertical mixing, horizontal transport,
temperatures, shading as well as 3D urban forms directly influence near‑surface concentrations.
3.3 ENERGY PROVISION AS A SERVICE – CL‑UHI IMPACTS OF AND

ON ENERGY USE
The amount of energy used to keep indoor environments comfortable depends on the energy
balance of the building envelope, ventilation rates, solar gain and the demographics of the
occupants. The energy balance of the building envelope is a function of air temperature,
wind speed, the amount of radiation exchanged with the environment, the building’s shape,
building materials, airtightness, internal sensible and latent heat loads resulting from, for
CHAPTER 3. URBAN SERVICES NEEDING CL-UHI INFORMATION 17
instance, cooking or lighting and the density of people in a room (100–200 W per person).
Cities with different building designs and construction standards, and different socioeconomic
characteristics (for example air‑conditioning usage patterns) will differentially impact or be
impacted by CL‑T. For example, major increases in space cooling can increase energy demand,
and the resulting anthropogenic heat emission can amplify at microscale the night‑time CL‑UHI
by 0.5–1 K (de Munck et al., 2013) and thereby CL‑T. However, the CL‑UHI reduces winter energy
demands in moderate to large cities in cold climate regions (Varentsov et al., 2018).
Energy use associated with buildings and transport is a major anthropogenic heat source in
urban areas that needs to be accounted for in the urban energy balance and can affect the
CL‑UHI. The amounts and spatial distribution of this heat will change with the decarbonization of
the energy sector. This is dependent on political, technical and social changes at different spatial
scales. Within cities, there are moves towards shared/common energy distribution networks
to reduce energy waste, with a shift to greater use of renewable energy sources to overcome
intermittency in supply. The technological changes associated with energy transitions, whether
replacement technology (for instance replacing combustion boilers with electric heat pumps) or
new technology (for example photovoltaic solar panels), have shifted energy considerations from
being predominantly focused on large‑scale supply management to a model where distributed
production and consumption has increased relevance. Balancing energy demand with supply
is requiring more sensitivity to microscale and local‑scale conditions that, when combined with
varying social influences across a city, are giving greater weight to the argument that automated
and “smart” controls are critical.
Climate change (global warming) will increase demand for summer cooling at higher latitudes
than today. Demand profiles will differ vertically and horizontally within the UCL (Hertwig
et al., 2021a). Knowledge of the high‑resolution annual and seasonal CL‑UHI values, as well as
diurnal and extreme values, will help to facilitate the basic engineering tasks of system sizing and
design. Modelling (Chapter 6) and monitoring (Chapter 7) of the CL‑UHI provide an opportunity
for more informed ventilation and heating/cooling control that is particularly relevant to
buildings with mixed‑mode strategies and for maintaining critical infrastructure (for example
cooling demand in data centres).
Shared energy networks, such as those for buildings and transport, face operational constraints
at the local scale and microscale. In urban areas with pronounced CL‑UHIs, for instance, increases
in need for cooling might coincide with higher energy demands for electromobility, requiring
intelligent system management and control of energy by end users. The stress placed on existing
energy networks from an expectation of increased demand at peak times is leading to new
energy services where data and knowledge on the CL‑UHI has a part to play. Such services
include the consolidation of energy end‑ user ability/capacity to shift energy demand (in time)
through demand‑side management programmes and new business models with an emphasis
on “service” rather than “energy” provision, for instance by contracts/payment for guaranteeing
building indoor comfort rather than energy use. Approaches for assessing the viability of these
services are beginning to be explored where differentiation in the CL‑UHI will add value in
assessing the capacity of weather‑sensitive energy infrastructure, for instance through local
zone classification (Section 4.3), CL‑UHI measurements (Chapter 5), modelling (Chapter 6) and
monitoring (Chapter 7).
Temporal and spatial dynamics of energy use are not only influenced by the heterogeneity
in design, built features (Section 2.2, Table 2.1) and societal variance (Sections 4.1, 4.2),
but also by movement and transport characteristics. Incorporating CL UHI data into traffic
management procedures may impact behaviour and thus place and time of energy demand.
These dynamics are beginning to be considered in relation to energy and CL‑UHI modelling
(Capel‑Timms et al., 2020).
(CL-UHI)
3.4 URBAN VEGETATION AS A SERVICE – CL‑UHI IMPACTS ON AND FROM

VEGETATION
Urban vegetation not only provides ecosystem services, but also helps to influence CL‑UHI
patterns (Subsection 2.5.2) and is thus a relevant urban climate service. It provides shade that
reduces radiant temperatures (Section 3.1) and cultural services, as green spaces contribute to
human well‑being (Chang et al., 2017; von Szombathely et al., 2017). The CL‑UHI can extend the
frost‑free season by several weeks each year, affect temperatures during individual freeze events
and provide warmer habitats. Thereby, the CL‑UHI can influence plant composition and diversity,
with those needing higher temperatures found in areas that often have larger CL‑UHI intensities
(Figure A3.2). In addition to plant composition, the growing season and thus the period of active
photosynthesis is influenced by the CL‑UHI: it is longer in mid‑ and high‑latitude urban areas
than in their surroundings, with an earlier flowering time, a later end of the growing season and
an increase of growing degree days. As one result, the total length of the pollen‑allergy season is
longer. Furthermore, the higher temperatures may allow exotic plants to grow, introducing new
allergens or causing ecological issues. There is evidence that pollen in polluted urban areas has a
higher allergenicity, with several reasons for this effect (D’Amato et al. 2007).
Urban vegetation clearly impacts CL‑T and hence the CL‑UHI through increased latent heat fluxes
which reduce the turbulent sensible heat flux. Street trees also provide shade, which reduces
daytime temperatures and can trigger a cool island (Subsection 2.5.2). Urban vegetation can
reduce CL‑UHI intensities by 0.5 °C–4.0 °C at microscale. Urban gardening and green roofs or
façades may also increase latent heat fluxes and thereby reduce the CL‑UHI intensity; however,
sufficient water needs to be available. In addition, some vegetation may improve urban air
quality by absorbing gases and particles (Section 3.2), and vegetation provides water retention
areas for rainfall, and modifies the thermal mass and, therefore, heat storage. Cities may reduce
the CL‑UHI by implementing sustainable urban agriculture policy and zoning interventions such
as promoting the concept of intensive production and creating “edible urban landscapes” by
building green roofs and greenways, and encouraging vertical farming. The CL‑UHI may assist in
colder regions, making it possible to grow vegetables within the urban area when not possible in
the cooler surroundings.
3.5 MULTI‑HAZARD EARLY WARNING SYSTEM –

CONSIDERATION OF THE CL‑UHI
Knowledge about the CL‑UHI is relevant to delivery of an IUS, being included in daily forecasts
and multi‑hazard early warning systems that provide air quality data and thermal comfort
indices. The time‑dependent pattern of CL‑UHI intensity will vary with the meteorological
situation. This information can help to derive city‑wide temperature patterns supported by
measurements (Section 5.2) or by model results (Section 6.2). CL‑UHI information may also
be a valuable input for issuing targeted warnings based on large‑scale weather forecasts
(Subsection 6.4.2) or for assessing thermal comfort using climate projections (Subsection 6.4.3).
It should be remembered that the CL‑UHI intensity indicates the difference in temperature in
an urban area relative to a nearby rural area. Thus, changes in rural surface cover (for instance
wildfire, rural flooding) influence the CL‑UHI intensity without necessarily resulting in higher or
lower temperatures in the urban area.
CHAPTER 4. CHARACTERIZATION OF THE URBAN AREA AND
ITS SURROUNDINGS
Given that the CL‑UHI is defined as the difference between CL‑T values in different areas
(Section 2.3), the characterization of the region of interest (urban and rural areas) is essential
for understanding, measuring, modelling, monitoring (Chapter 5, Chapter 6 and Chapter 7,
respectively) and mitigating (Chapter 8) the CL‑UHI. Several parameters are needed for
characterizing feature types at urban and rural sites in order to provide more complete urban
services (Chapter 3). To determine the intensity, location and timing of the maximum CL‑UHI,
it is advisable to characterize the whole urban area and its surroundings (region of interest, see
Section 2.3). As the methods to obtain these parameters require expertise and regular updating,
there is an associated cost in terms of time and money.
4.1 TYPES OF PARAMETERS NEEDED
The numerous features to characterize the urban area and its surroundings can be subdivided
into four types (Table 4.1) related to: (1) built form, (2) vegetation and other land cover forms,
(3) function (human activities), and (4) geographic setting. The parameters vary between and
within different cities.
The level of detail should ensure that local‑scale variability within the urban area is captured
and that for observation sites the microscale variations are monitored over time to ensure
appropriate interpretation of observations (Chapter 5). For modelling applications (Chapter 6),
this characterization is needed at least at the grid resolution, preferably with more detailed
information within each grid cell to estimate the parameter’s variability and to provide correct
parameters to the model. Without the correct parameters, the model physics cannot provide
reliable output. Given the dynamic nature of cities and nearby rural areas, the characteristics
must be updated regularly.
4.2 DETAILED PARAMETERS AT MICROSCALE
For characterizing the immediate measurement site surroundings (Section 5.6) or for
obstacle‑resolving modelling (Subsection 6.1.2) or similar applications, several parameter
values (Table 4.1) are needed to reliably describe both the 2D land cover and 3D form as well as
the materials. Table 4.2 gives an overview of data sets typically used and possible data sources
Table 4.1. Characteristics to describe urban and rural areas
Vegetation and other

Feature type Built form Function Geographic setting
land cover forms
Related Morphology Morphology Land use Terrain (elevation,
parameters (heights, shape, (e.g. tree structure) slope, orientation)
density, surface Emissions (heat,
covers, roughness) Phenology water, aerosols, Major water bodies
gases)
Materials (heat Soil Latitude
capacity, diffusivity, Human behaviour
albedo, emissivity, Water (e.g. mobility
soil texture) patterns)
Transport (e.g. car,

motorbike, train)
(CL-UHI)
Table 4.2. Data sources typically used to determine parameters for built,
vegetated and other land cover forms at microscale
Built, vegetated and other land

Data sets Data sources
cover form (urban form)
Morphology – of roughness – Building inventories – City administration
elements: – Derived from heights – City planning departments
– Wall area density or frontal and spacing between – City parks agencies
area density roughness elements – Tree advocacy groups
– View factors (vegetation)
– Distance between – Transport departments
roughness elements (street trees)
(e.g. width between – City and regional planning GIS
buildings/trees)
– Street directions
Heights – of roughness – Stereographic images – City administration
elements: – LIDAR – City planning departments
– Buildings – Synthetic aperture radar – City parks agencies
– Trees – Tree advocacy groups
– Other (vegetation)
– Transport departments
(street trees)
– National agencies
Surface cover: – Satellite or aerial – Cartography departments
– Horizontal plan area photography – City administration
fraction – City land use and cover
– Impervious: buildings, inventories
paved, rocks – Google street view
– Pervious: vegetation
(e.g. trees, crops, grass),
water, soil
Materials: – Architectural information – Planning department
– Fraction of wall occupied database
by windows – Spectral libraries
– Radiative characteristics – Planning regulations by
– Thermal characteristics period
– Building information
systems
– Building archetype
– Analysis of street‑level,
aerial or satellite images
(e.g. material properties)
Vegetation: – Satellite or aerial – City parks agencies
– Leaf area index observations – Tree advocacy groups
– Phenology – Ground surveys – Crowdsourcing
– Type
Soil: – Satellite images – City database
– Type – Measured in situ soil – Model databases
– Moisture (dynamic, initial profiles
conditions)
Hydrology: – Modelled sewage system/ – City planning department
– City water drainage and water retention map – Flood planning/protection
sewage systems agencies
– Wastewater disposal service of the
city
Note: GIS – geographical information system, LIDAR – light detection and ranging.
CHAPTER 4. CHARACTERIZATION OF THE URBAN AREA AND ITS SURROUNDINGS 21
providing these parameter values. Built, vegetated and other land cover forms are mixed in
urban areas (for example street trees) and together shape the urban form (Figure 2.1). While
data sets for built form and vegetation are quite different, the data sources are often the same.
In many cities, the planning departments have detailed geographical information systems (GISs)
with much of the 3D information needed, so collaboration is mutually beneficial to ensure the
urban form information used is current and correct. Many of the morphometric parameters
needed can be approximated from existing parameters, using empirical relations for example.
If no detailed information is available, parameters need to be taken from the literature, ideally
taking the local context into account.
More challenging than obtaining morphology or height data (Table 4.2) is obtaining the
material characteristics which influence radiative characteristics (for example albedo, emissivity)
and thermal characteristics (for example thermal conductivity, admittance, density) at microscale
resolution. Information exists for many new buildings today. However, the construction process
may cause, for example, thermal contacts, or maintenance may result in building surfaces being
cleaned or not cleaned, so that the existing information is not fully applicable to the existing
building. Furthermore, surface characteristics, most importantly albedo, may be changed by
deposition of aerosols and dust. At the building scale, large window fractions and their thermal
properties (for example, heat transfer coefficient and emissivity) impact heating, cooling
and ventilation. Several methods can be used to estimate architectural information (Masson
et al., 2020). One example is the use of street‑level, aerial or satellite remote sensing images to
determine materials like thermal and radiative properties, which can benefit from crowdsourcing
(for example, as done in Tunis, Tunisia (Mhedhbi et al., 2019)). Table 4.3 gives an overview of
how people’s activities influence the functioning of cities and the data sets typically used to
capture these.
Despite their local‑scale resolution (Table 4.4), land‑use data provide some information about
when and how an area is likely to be occupied (for instance residential versus commercial).
Microscale data are preferable, but often not available. Other issues to consider for determining
parameters at the microscale include the following:
– Vegetation data tend to be biased by their sources. For example, the city GIS may have a
good record of large street and park trees but lack data for private properties. Required
urban vegetation parameters are similar to those for buildings (for example heights, areal
extent by type), but leaf area index and phenology are additionally required. Similar
techniques are used for the built form (Table 4.2).
– Timing is important (such as of crop rotations, typical work days and times, snow
clearing patterns).
– At middle and high latitudes, anthropogenic heat emissions due to space heating may be
larger than the energy input by net all‑wave radiation in winter.
– In cities with extensive air conditioning, extraction of heat from buildings and emission to
the urban atmosphere can increase the CL‑UHI intensity by ~1 K (Takane et al., 2019).
– Some industries inject large amounts of heat into the atmosphere.
– Heat release from transport can be important near large roads or at local openings of
metro lines.
– Water removal from impervious surfaces after rainfall events may result in large amounts
of heat being moved to the storm water and sewage infrastructure. With nature‑based
solutions (see Section 8.2) being advocated, new rainwater retention areas are created
influencing latent heat fluxes in urban areas. As this adaptation approach for cities to
climate change is increasingly applied, the areal extent and location of the nature‑based
solutions need to be regularly updated, for instance in city GIS databases, and used.
(CL-UHI)
Table 4.3. Data sources typically used to determine parameters to account for human
activities that can characterize urban function. These need to be determined
for the appropriate scale.
Function Data sets Data sources

Land use – City and regional planning – City administration
maps – Agricultural department (rural crops)
– Copernicus, USGS for regional
Anthropogenic emissions – Fuel consumption data – Energy department/companies
– Energy use – Building construction data – Transport department/companies
– Heat emission – Socioeconomic
characteristics
– Gas and particle – Emission inventory – Government/local agencies with
emissions to air or different responsibilities which may
water vary with source type (e.g. industrial,
residential, transport)
– Environmental agencies, which may
mandate activities
– Water – Emission inventory – Companies or local government
agencies that supply water
– Government agencies, which may
limit permitted uses (e.g. hose‑pipe
or irrigation bans or restrictions)
– Building usage – People’s activity patterns – Crowdsourcing
(offices, homes) – Census data (home, work – Planning department
– Heating and air population) at small spatial – Energy regulators
conditioning (usage units – Public transport providers
pattern) – Typical work/home/
– Ventilation recreation patterns
– Building information systems
– Work related activities
Transport – Fuel consumption data – Energy department/companies
– Public transport (bus, – Socioeconomic – Transport department/companies
subway, etc.) characteristics – Crowdsourcing
– Car (types) – People’s activity patterns – Planning department
– Biking
– Walking
Note: CORINE – Coordination of Information on the Environment (European Union programme),

USGS – United States Geological Survey.
Table 4.4. Data sources typically used to determine parameters for functions at local scale
Function Data sets Data sources

Land use – City and regional planning – City administration
maps – Agricultural department (rural crops)
– CORINE land cover data – https://land.copernicus.eu/pan-european/
(Europe) corine-land-cover
– Copernicus Global Land – https://land.copernicus.eu/global/content/
Cover annual-100m-global-land-cover-maps-available
– USGS maps – https://www.usgs.gov/
– LCZ search?keywords=Land%20Cover
– https://www.wudapt.org/lcz-maps/
Anthropogenic Statistical data: – National/county/city administration data sets
heat emission – Energy consumption – National/county statistical offices
– House types
– House insolation status
Further data: – Table 4.3 – Table 4.3
Note: LCZ – Local Climate Zone (Figure 4.1).

– Microscale data should be updated if major changes occur, and should be checked at least
every 5 years, if not rapidly changing, but more frequently in periods of development and
gentrification.
For the rural surroundings, similar data are required as for the urban area (Table 4.2), but the
potential partner agencies may use different techniques to gather these data. The proportion of
individual land cover types will vary throughout the year and from year to year, and with that,
the heights of roughness elements such as crops and trees that are essential to characterize the
microscale. In these areas, the most critical parameters are likely related to vegetation (type,
phenology, leaf area index), soil type, soil moisture condition and water bodies.
Water balance modelling accounting for both precipitation and irrigation can provide soil
moisture conditions. Spatial and seasonal variations of soil moisture impact both urban and
rural areas, and both urban and rural soils may or may not be irrigated depending on human
activities. For example, in very dry regions the urban area may have moist soils, while the rural
region may be more spatially variable in terms of humidity conditions, resulting in very different
CL‑UHIs depending on the rural reference used.
4.3 CHARACTERIZATION AT THE LOCAL SCALE
A useful starting point for characterizing the region of interest at local scale may be to map the
urban area and its surroundings into areas with similar characteristics using one of the many
existing classifications for urban and rural areas. For example, land use, such as residential,
commercial, agriculture, forest, transportation (Table 4.4), is frequently used for this purpose
(land use is one of the functions in Table 4.1). The attributes derived from a classification depend
on the intended use of the data. The data are one input for mesoscale models (Subsection 6.1.3).
Many global maps of land use only have one urban class, whereas others may provide several
classes, for example the Urban Atlas (European Environment Agency, 2021).
To estimate anthropogenic heat emissions at local scale (Table 4.4), two general approaches are
used (Sailor, 2011):
– Top‑down inventories: Large‑scale (mesoscale or larger) energy consumption

information is disaggregated to local scale using ancillary data. These may be building
data or local‑scale annual consumption data. The large‑scale data may be available for the
whole urban area or the country with time intervals such as 30 min allowing for example
assessment of heating/air conditioning responses to regional temperatures.
– Bottom‑up inventories: Components (buildings, vehicles, people) are individually

simulated and then aggregated to the desired scales (Capel‑Timms et al., 2020).
Both approaches require data about the structure of the urban area (urban form) and how
people use the urban area (function). These data can be challenging to obtain. As data may be
different between administrative units, harmonization will likely be required.
One local zone classification scheme designed for UHI studies characterizing urban form and
considering urban function is the Local Climate Zone (LCZ) scheme created by Stewart and Oke
(2012) (Figure 4.1). The scheme divides urban and rural landscapes into 17 classes, of which
seven are land cover types (LCZs A–G) and 10 are built types (LCZs 1–10). Each class describes a
relatively uniform landscape with a distinct land cover and built form, though not necessarily a
distinct thermal climate.
The LCZ classification is widely used by the urban climate community and interpretable by urban
planners and city administration, as it is based on form (built types in Figure 4.1) and function
(land cover types in Figure 4.1). The LCZ scheme serves many purposes, of which the main ones
are: (i) the general classification of neighbourhoods where meteorological stations are installed
(Chapter 5); (ii) the specification of surface parameter values for atmospheric models; and
(iii) the exchange of urban climate knowledge with city planners.
(CL-UHI)
Built types Definition Land cover types Definition
1. Compact high-rise Dense mix of tall buildings to A. Dense trees Heavily wooded landscape of
tens of stories. Few or no trees. deciduous and/or evergreen
Land cover mostly paved. trees. Land cover mostly pervious
Concrete, steel, stone, and (low plants). Zone function is
glass construction materials. natural forest, tree cultivation, or
urban park.
2. Compact midrise Dense mix of midrise buildings B. Scattered trees Lightly wooded landscape of
(3–9 stories). Few or no trees. deciduous and/or evergreen
Land cover mostly paved. trees. Land cover mostly pervious
Stone, brick, tile, and concrete (low plants). Zone function is
construction materials. natural forest, tree cultivation, or
urban park.
3. Compact low-rise Dense mix of low-rise C. Bush, scrub Open arrangement of bushes,
buildings (1–3 stories). Few or shrubs, and short, woody trees.
no trees. Land cover mostly Land cover mostly pervious (bare
paved. Stone, brick, tile, and soil or sand). Zone function is
concrete construction natural scrubland or agriculture.
materials.
4. Open high-rise Open arrangement of tall D. Low plants Featureless landscape of grass
buildings to tens of stories. or herbaceous plants/crops. Few
Abundance of pervious land or no trees. Zone function is
cover (low plants, scattered natural grassland, agriculture,
trees). Concrete, steel, stone, or urban park.
and glass construction
materials.
5. Open midrise Open arrangement of midrise E. Bare rock or paved Featureless landscape of rock or
buildings (3–9 stories). paved cover. Few or no trees or
Abundance of pervious land plants. Zone function is natural
cover (low plants, scattered desert (rock) or urban
trees). Concrete, steel, stone, transportation.
and glass construction
materials.
6. Open low-rise Open arrangement of low-rise F. Bare soil or sand Featureless landscape of soil or
buildings (1–3 stories). sand cover. Few or no trees or
Abundance of pervious land plants. Zone function is natural
cover (low plants, scattered desert or agriculture.
trees). Wood, brick, stone, tile,
and concrete construction
materials.
7. Lightweight low-rise Dense mix of single-story G. Water Large, open water bodies such
buildings. Few or no trees. Land as seas and lakes, or small
cover mostly hard-packed. bodies such as rivers, reservoirs,
Lightweight construction and lagoons.
materials (e.g., wood, thatch,
corrugated metal).
8. Large low-rise Open arrangement of large
low-rise buildings (1–3 stories). VARIABLE LAND COVER PROPERTIES
Few or no trees. Land cover
mostly paved. Steel, concrete, Variable or ephemeral land cover properties that change
metal, and stone construction significantly with synoptic weather patterns, agricultural
materials. practices, and/or seasonal cycles.
9. Sparsely built Sparse arrangement of
small or medium-sized b. bare trees Leafless deciduous trees
buildings in a natural (e.g., winter). Increased sky view factor.
setting. Abundance of Reduced albedo.
pervious land cover (low
plants, scattered trees). s. snow cover Snow cover >10 cm in depth. Low
admittance. High albedo.
10. Heavy industry Low-rise and midrise industrial
structures (towers, tanks, d. dry ground Parched soil. Low admittance. Large
stacks). Few or no trees. Land Bowen ratio. Increased albedo.
cover mostly paved or
hard-packed. Metal, steel, and w. wet ground Waterlogged soil. High admittance. Small
concrete construction materials. Bowen ratio. Reduced albedo.
Figure 4.1. Local Climate Zones (LCZs) to characterize the land cover and surface structure
of urban and rural areas
Source: Stewart and Oke (2012) © American Meteorological Society. Used with permission.
A popular means to implement the scheme in urban climate work is through LCZ mapping.
LCZ maps can be produced in multiple ways but ideally should be comparable when finished.
For example, satellite imagery is used by the World Urban Database and Access Portal Tool
(WUDAPT) (Ching et al., 2018), while Open Street Map (Coast, 2015; www.openstreetmap
.org) is used by other investigators. A global LCZ map produced by Demuzere et al. (2022) is
freely available.
It should be noted that characterizations at local scale (Table 4.4) currently do not include
spatial detail within the pixel. Calculating all urban parameters in detail as needed for the
microscale (4.2), either for characterization of observations (Chapter 5) or for obstacle‑resolving
models (Subsection 6.1.2), may be difficult or impossible from these data, because of their
insufficient resolutions of ~100 m (for example, CORINE Land Cover data for Europe, Copernicus
Global Land Cover). In even coarser resolution (~1 km), several additional data sets, such
as for leaf area index, can be obtained from Copernicus or the USGS. For characterizing the
direct surroundings of a measurement site (Section 5.6) or for obstacle‑resolving modelling
applications (Subsection 6.1.2) or similar applications, these local data sets provide insufficient
spatial accuracy.
4.5 REGIONAL TOPOGRAPHIC CHARACTERIZATION RELEVANT FOR THE

CL‑UHI
The geographic setting of the city is important and influences the CL‑UHI intensity (see
description of orography effects in Subsection 2.5.5). This can include relative terrain differences,
the presence of large water bodies (including the ocean) and the potential impact of ocean
currents. Even small variations in altitude in the region of interest need to be accounted for
(Section 5.2). To perform corrections or consider uncertainties induced by the geographic
setting when interpreting the CL‑UHI, the geographic location of major water bodies and
altitudes have to be determined and documented for the whole region of interest. In addition,
regional influences of water bodies and topography on air flows and temperature distributions
from outside the region of interest should be assessed to ensure that the determined CL‑UHI
intensities are really caused by an urban effect and not by the topographic conditions.
4.6 MAIN ASPECTS FOR CHARACTERIZING URBAN AREAS IN BRIEF
If the CL‑UHI is only determined for a pair of observation sites (urban and rural), only microscale
details are needed and the CL‑UHI is only assessed for this pair of sites (see information on site
selection in Section 5.4). Local‑scale information helps to assess if the CL‑UHI might have similar
values in other parts of the city. On the other hand, when the whole region, with its true spatial
variability, needs to be assessed, local‑scale spatial data are needed for the whole region of
interest. Examples are CL‑UHI forecasts (Subsection 6.4.2), assessment of a future climate CL‑UHI
(Subsection 6.4.3), urban development and the CL‑UHI (Subsection 6.4.4), urban services
(Chapter 3) or mitigation of the CL‑UHI (Chapter 8).
CHAPTER 5. DETERMINING THE CL‑UHI FROM OBSERVATIONS
5.1 BACKGROUND ON MEASUREMENTS
As cities expand and the absolute number of people living in cities grows, meteorological
stations that provide regular observations of the urban thermal conditions become increasingly
important. Given the complexities of the UCL (Section 2.2), it is difficult to make observations
in the UCL with a large spatial and temporal representativeness. Most urban stations do
not conform to the WMO standard guidelines for site selection and instrument exposure
applicable to flat and homogeneous terrain (Guide to Instruments and Methods of Observation
(WMO-No. 8), Volume I: Measurement of Meteorological Variables); instead, the guidelines for
the urban environment need to be considered (Guide to Instruments and Methods of Observation
(WMO-No. 8), Volume III: Observing Systems, Chapter 9). Despite the complexity and
inhomogeneity of urban environments, useful and repeatable observations can be obtained
by careful experimental design that reduces uncertainties and thereby enhances the value of
observations. The challenge of measuring the CL‑UHI and the many aspects to be considered are
the focus of this chapter.
5.2 CALCULATION OF THE CL‑UHI INTENSITY FROM URBAN AND RURAL

OBSERVATIONS
The challenge posed is to isolate the urban effect in the observations. Although CL‑UHI intensity
appears to be the simple temperature difference between an urban and a rural measurement site
(Section 2.3), establishing its intensity accurately requires the removal of several confounding
effects. The air temperature is influenced by processes at multiple scales (Chapter 2), including
the weather conditions (Subsection 2.5.4) and geographic setting Subsection 2.5.5), which
complicate the detection of the urban effect as the CL‑UHI, particularly over long time periods in
which one or more influencing factors are likely changing.
If the regional climate remained ‘constant’, the urban influence on temperature could be most
directly assessed from a continuous set of synchronous observations at reference sites that
respond to urban development. If the measurements began prior to any urban development,
then a ‘true’ urban signal could be obtained. Such opportunities rarely exist and, furthermore,
climate changes; thus, temperature values taken in an urban area and compared with values
at the same place before urbanization include both the urban and the climate change signal.
Thus, pragmatically, the urban effect is estimated from the difference in simultaneously observed
temperatures at carefully selected urban and rural sites.
Details on how to choose the sites are provided in Section 5.4. In particular, the use of a
rural reference (Subsection 5.4.1) poses some conceptual issues (Oke et al., 2017). From the
perspective of the CL‑UHI, these include the following:
– Measurements in the rural area are not equivalent to pre‑urban values.
– As the rural area is not static, the urban–rural difference can change through time without
urban changes.
– Rural stations may be affected by advection of warmer air from the urban area.
– Rural and urban sensors must be exposed to the same regional climate, for example they
should be placed at the same distance from the coast.
– Rural and urban sensors should be placed at the same altitude and height above ground.
CHAPTER 5. DETERMINING THE CL-UHI FROM OBSERVATIONS 27
– At both urban and rural sites, careful consideration of sensor location is required,
recognizing that the surface characteristics and wind variabilities that influence
temperature at both sites will have spatial variability. This variability is particularly marked
in urban areas, resulting in temperatures that generally vary more rapidly in space and time
compared to temperatures in a rural environment.
As air temperature decreases in a well‑mixed boundary layer without phase changes with a dry
adiabatic lapse rate of about 1 K per 100 m, a height‑based altitude correction is needed for all
the urban and rural reference sensors used to estimate the CL‑UHI. This avoids systematic errors
while analysing differences between stations at different altitudes.
5.3 MEASUREMENT APPROACHES
Although there are no clear guidelines, the use case and metric under investigation (Table 2.2)
largely determine how many stations are required. This section outlines four distinct but
complementary approaches. Examples of network layouts are provided in Appendix 2.
5.3.1 Single pair of sites
One way to determine the CL‑UHI is to use ground‑based fixed stations. This allows the
assessment of temporal variations on a continuous and long‑term basis. Careful consideration
should be given to which metrics are needed for the use cases (Table 2.2). If only one pair of sites
is to be selected it is critical to acknowledge that an urban site truly representative of the urban
area does not exist, as microscale and local impacts always occur (Section 5.4). Reference WMO
sites in green spaces, over grass, or at semi‑urban locations will at best only partially represent
urban effects. Although a site may be maintained for decades with the same microscale exposure
(facet‑scale, 10 m), there is often no control over changes in the larger area around the site which
will affect the observed near‑surface temperatures and thus the derived CL‑UHI intensities.
This highlights the essential need to regularly update station metadata (Section 5.6). Sites with
instruments on rooftops exist in some cities, but these locations are too high to monitor the
CL‑UHI (Subsection 5.4.5).
If the goal is to establish the temporal variations of the maximum CL‑UHI in the urban area, one
urban station may suffice after appropriate preparatory observations (for example traverses,
see Subsection 5.3.2). Classifying neighbourhoods, for example using LCZs (Section 4.3), can
provide initial guidance to select possible sites to obtain this metric. Other use cases where a
single site may be appropriate include monitoring the impact of a particular neighbourhood,
such as the LCZ with the greatest extent in an urban area, assessing a site before and after
development or assessing the CL‑UHI of a neighbourhood with a population vulnerable to
potential excessive heat (Section 3.1). However, one urban station will not capture spatial
differences in the CL‑UHI, and it will be extremely challenging to justify the representativeness of
one site without undertaking traverses or considering measurements at other sites.
5.3.2 Traverse approach
Suitable location(s) to install sensors to measure the CL‑UHI can be derived by mobile surveys
to identify the spatial variations of air temperature in response to variations in urban and rural
characteristics. Any mobile platform can be used for such traverse approaches, including
a person, bicycle, automobile, truck or public transport vehicle, as long as care is taken that
the instrument location on the mobile platform will not introduce unnecessary errors into
the measurements (Subsection 5.5.3). Drones are generally not helpful, since a flight level of
1.25–2 m AGL is likely to be difficult and permissions are hard to obtain in cities. However, in rural
areas there may be greater opportunity to use them as it will be easier for the pilot to maintain
eye contact with the drone.
(CL-UHI)
A common approach is to use an out and back traverse that samples the same route in two
directions. Traverse routes can be designed to sample selected neighbourhoods to identify the
major features of the CL‑UHI with sample points also in the rural surroundings. Using multiple
mobile platforms simultaneously allows for greater spatial coverage and better resolution of
spatial patterns. Using a single sensor in the traverse approach has the advantage of avoiding
inter‑instrument comparison issues. It is limited by the need to apply temporal corrections for
the temperature changes that occur in the time required to complete the traverse. These might
be considerable around sunrise or sunset. The design of traverse routes should limit the total
traverse time to approximately 1 hour, otherwise temperature and weather will have changed,
potentially making uncertainties and magnitudes of the temporal corrections large compared to
the spatial differences of temperature.
A common starting point and end point are essential to allow estimation of a linear rate of
change of temperature occurring during the traverse. If the platform speed is approximately
constant and the rate of change in temperature with time is linear, the average of the outbound
and inbound points provides a simple and sufficiently accurate correction. More complex
corrections may use data taken at network stations (Subsection 5.3.3) or be based on repeated
measurements of more points over the traverse to determine the temperature changes
associated with different land cover types (Section 4.3). After correcting for temporal effects, the
data can then be mapped to show the spatial variation of temperature.
When the traverse is carried out using a vehicle, temperature readings must be made with
sensors that have a rapid response time, sufficient to register changes caused by spatial variations
such as a local park. This is quite difficult to achieve in practice: considering that even at a
moderate speed of 36 km hr–1 a vehicle covers a distance of 10 m s–1, there are few temperature
sensors that can record temperature differences that occur over distances of this order
of magnitude.
Each mobile platform type is limited in terms of the locations it can access (for example it may be
limited to the road or footpath network), hence many areas may not be accessible to a particular
mobile platform (for instance private property). To detect maximum CL‑UHI values, traverses
are best conducted at night, typically some hours after sunset or before sunrise, when winds are
calm and skies are clear (Subsection 2.5.3).
5.3.3 Meteorological networks
A network of urban and rural stations can provide information about the spatial character of the
CL‑UHI. Each site in an urban meteorological network needs to adhere to the guidance in the
present publication. It is critical to select sites that are representative (Section 5.4), and that site
metadata (Section 5.6) be gathered regularly to ensure appropriate interpretation. With enough
sites, the spatial form of the CL‑UHI intensity emerges and different metrics can be determined.
Weather station costs have decreased markedly over the last couple of decades, making it now
possible to deploy relatively dense networks of weather stations across cities provided space is
available. There is a growing body of examples of urban meteorological networks now deployed
across the world, utilizing a range of deployment strategies (some examples are provided in
Appendix 2). However, the ongoing costs of maintenance and governance of large networks
have led to opportunistic approaches (Subsection 5.3.4) being explored as an alternative for
monitoring. In addition, university research networks were set up in places such as Birmingham
(Chapman et al., 2015), Hamburg (Wiesner et al., 2014) or Oklahoma City (Hu et al., 2016). Full
weather stations are not required for CL‑UHI estimation, and adding more temperature sensors
to measure just air temperature can provide useful additional information.
5.3.4 Opportunistic sensing – crowdsourcing
Beyond traditional, planned urban meteorological networks, a range of techniques is available

for opportunistic sensing (see Muller et al., 2015 for a summary), including crowdsourcing,
citizen (or public) science weather stations, connected vehicles and mobile phone data. Each
can provide abundant low‑cost (effectively free) data potentially with high density compared
to traditional meteorological networks. However, they need quality control that includes
corrections for instrument and measurement errors (Section 5.5) and attribution of the thermal
source area that they measure (Subsection 5.4.3). It is often difficult or impossible to obtain
relevant metadata (Section 5.6).
Citizen science stations run by weather enthusiasts arguably provide a viable option for
temperature data, though a challenge is often discoverability as there are many companies
that provide interfaces to data from their own sensors but not from others. As these free data
are collected via multiple platforms, some of which are well maintained and include metadata,
they can significantly increase the number of CL‑T measurement points in a city (Chapman
et al., 2017). Likewise, data from sensors installed as standard equipment in vehicles are
increasingly available, but both sensor location on the vehicle and data protection regulations
can make some data unusable or inaccessible. For both citizen science weather stations and
vehicles, the meteorological community has been actively involved in assessing their utility
with efforts to assimilate the data to enhance forecast efforts. There is a potential to estimate
air temperatures from mobile phone battery temperature data, but this is arguably the least
reliable approach.
Despite the dense spatial coverage, data need to be checked (outliers, noise) and preferably
statistically analysed to provide anything meaningful, with the only benefit so far being at
city‑scale. Geostatistical or machine learning techniques may potentially be transformative in the
future and provide usable temperature patterns, but data protection regulations and metadata
issues will remain.
At the time of writing of the present publication, the use of crowdsourcing data is recommended
only with great caution, and only to complement other sources.
5.4 CHOOSING A SITE
5.4.1 Selection of rural reference sites
As rural reference sites used to assess the CL‑UHI also have their own microclimates, their
placement needs careful consideration. A siting classification can be found in Annex 1.D of
Guide to Instruments and Methods of Observation (WMO-No. 8), Volume I: Measurement of
Meteorological Variables. The LCZ scheme (Section 4.3) allows rural areas as well as urban ones
to be locally characterized. Choosing a site that is representative of one LCZ is challenging, as the
spatial variability within one may be as great as within the urban area. Furthermore, agricultural
areas undergo seasonal changes that will affect urban–rural differences, as does the vegetation
around rural reference sites. In addition, agriculture creates some side effects influencing the
CL‑UHI, as there is a drastic difference in rural temperatures caused by the varying degrees of
vegetation cover, irrigation rates and concentrations of atmospheric aerosols. This will change
the apparent urban–rural temperature difference without any changes in the urban area.
Information on vegetation type and phenology helps to interpret the measured temperatures.
Thus, as in the urban area, the characterization of the rural area needs to be both detailed and
regularly updated (see Guide to Instruments and Methods of Observation (WMO-No. 8), Volume I:
Measurement of Meteorological Variables, Annex 1.F), and include phenology information
(Section 5.6).
Rural sites are prone to advection from urban areas. It is impossible to know the exact spatial
extent of the downwind urban thermal anomaly, but a rural station may be affected by
the advection of warmer air from the urban area (see description of thermal source area in
Subsection 5.4.3). Rural sites in the dominant upwind direction are therefore preferred and
can be identified using wind rose data derived from long‑term data from an operational
meteorological station nearby. If selecting multiple sites, locations perpendicular to the
prevailing wind direction can also be considered. For coastal cities, rural reference sites are
ideally located at a similar distance from the coastline as the urban site to avoid different
advective effects arising from sea/land breezes. In more mountainous or hilly areas, reference
(CL-UHI)
sites are ideally situated at similar altitudes to that of the urban site. They should not be sited
in areas where they would be differentially impacted by topographically influenced winds,
exposure to the sun and/or temperature stratifications (Subsection 2.5.5).
Temperature measurements from an urban park may be used to assess intra‑urban temperature
variability, but urban parks do not qualify as true rural reference sites for determining the CL‑UHI.
Similarly, an airport sensor must be carefully assessed for its ability to be a true rural reference, as
airports are characterized by extensive impervious areas, have anthropogenic activity and often
have significant urban‑like development surrounding them.
5.4.2 Purpose of the urban station
Site selection requires an initial definition of the purpose of the station: the use case has to be
clear (Table 2.2). Consideration should be given to the following questions:
– Is the station intended to characterize the greatest impact of the urban area on temperature
and the CL‑UHI (use cases such as weather forecasting or research)?
– Is the station targeting the area in the city with the highest population (use cases such as
energy use or health impact)?
– Is the station intended to represent a block, a neighbourhood or a larger urban area (use
cases such as climate monitoring or urban design)?
– Is the station part of a meteorological network intended to define the intra‑urban

temperature variability for specific neighbourhoods and the associated spatial structure of
the CL‑UHI (use cases such as health impact or agricultural applications)?
– Is the station targeting a particular site (for instance intended for development or
redevelopment) or is it intended to characterize a particular environmental issue or issues
(use cases such as ecology and vegetation or atmospheric chemistry)?
– Is the station intended to be in place for a short experiment or for a long period of time
(longitudinal study), potentially before and after heat mitigation interventions are put in
place (use cases such as urban design and planning or energy use)?
The specific siting of the station should consider the local land cover mix it is intended to
represent and the thermal source area (Subsection 5.4.3). Typically, local‑scale representativeness
is intended (Section 2.2); for this purpose large patches of relatively homogeneous urban
development are required, with minimum microscale influences for instance from facets,
buildings or urban canyons. In addition, if the aim is to monitor the local urban effects, then
mesoscale influences such as cold air drainage, river valleys or shoreline locations must be
carefully evaluated. As local‑scale cold air drainage effects can exist and influence both urban
and rural sites, there is a need to consider whether they should be captured within a site setting,
or not. If a potential location is too heterogeneous to derive locally representative values or
has anomalous microscale features that may distort the signal, it is advised to find a more
homogeneous setting for observations. The station may be classified according to Guide to
Instruments and Methods of Observation (WMO-No. 8), Volume I: Measurement of Meteorological
Variables, Annex 1.D; sites should be preferably restricted to Classes 1–3.
5.4.3 Thermal source areas
The temperature recorded by a sensor depends on the conditions along the back trajectory of
air parcels reaching the sensor; this area is named thermal source area. Interactions with various
different urban surfaces and facets generate the observed temperature characteristics. Turbulent
mixing with ambient air dampens these characteristics and diminishes the effect of distant
surfaces. Accordingly, the measured temperature originates from multiple surfaces at an area
upstream of the sensor location defined as thermal source area, or footprint (Figure 5.1).
0°
270° 90°
Temperature
sensor
100 m Wind direction and source area: daily mean

180°
Wind direction and source area: short-term mean
Figure 5.1. Neighbourhood-scale surface projections of hypothetical thermal source areas for
a standard screen-level-height temperature sensor. Depicted are two source areas
for short‑term (<1 h) intervals (orange, dashed lines) and daily mean considering
a predominant south-easterly wind direction (blue, solid line).
Source: Adapted from Stewart and Oke (2012). Drawn by K. Van Kerkoerle (University of Western Ontario) with input
from J. Voogt and M. Roth.
Conceptually the thermal source area is the total surface area “seen” by the sensor. The size,
shape and orientation of the source areas evolve with time and meteorological conditions
(turbulent mixing, wind speed and direction). Over short periods (<1 h) the area is roughly
elliptical and oriented in the upwind direction from the sensor (orange area in Figure 5.1).
Because of temporal changes in wind direction and stability, these areas oscillate around the
measurement site so, over time, they form a misshapen “circle” (blue area in Figure 5.1). Surfaces
and roughness elements (for instance buildings, trees, street furniture) will also influence
the thermal source area, particularly for the near‑surface measurements that are needed for
determining CL‑UHI intensities.
Analytical solutions to quantify the extent of the thermal source area have been developed for
homogeneous and isotropic turbulent conditions for the ISL – thus well above the roughness
elements (Figure 2.1). These models predict elliptical‑shaped source areas with increasing extent
for higher wind speed, higher measurement height and more stable conditions. The source
areas thus vary with the meteorological conditions, and they are time dependent. The analytical
solutions help to determine thermal source areas above the UCL but are not applicable within
it as the turbulence is non‑homogeneous, and blocking and wakes by buildings substantially
modify back trajectories. These effects can only be resolved by computationally demanding
obstacle‑resolving Reynolds‑averaged Navier–Stokes (RANS) or large‑eddy simulation (LES)
models (Subsection 6.1.2). Their results can be used to calculate more reliable backward
trajectories, or they provide information by using an embedded Lagrangian stochastic particle
model (LES–LS), or Eulerian passive tracer dispersion model. The resulting source area estimates
feature complex structures and will typically differ between nearby positions within an urban
canyon and/or among other roughness elements.
Exposing a sensor optimally to monitor effects within one specific neighbourhood, LCZ, or
situation of interest is a challenging task. Sufficient upstream fetch with homogeneous and
characteristic conditions of the local climate is needed. If no RANS, LES or LES–LS source area
calculations are available, analytical models might give some rough guidance. The appropriate
orientation of the source area can alternatively also be estimated from long‑term wind data for
the location, recognizing that the thermal source area size increases with the height of sensor and
atmospheric stability. Classification of measured temperature data according to wind direction,
wind speed or stability enables a better allocation of the thermal source area. However, the
immediate environment of a station remains most important since microscale effects dominate
(CL-UHI)
the signal and therefore are an essential part of the metadata (Section 5.6). Sensors located in
urban canyons with densely packed buildings have smaller and more poorly defined thermal
source areas and smaller areal representativeness compared to those sited in more open zones.
5.4.4 Sensor positioning for representative measurements of neighbourhoods
If the thermal modification caused by a particular neighbourhood is of interest, then sites

should be surrounded by conditions that are average or typical for the neighbourhood and the
microscale setting should be representative of its surface cover, geometry and human activity.
The site should be in the centre of an open space where the surroundings’ aspect ratio (H/W)
is approximately representative of the neighbourhood. To avoid heterogeneous microscale
influences in the measurements the following general considerations should be noted (Stewart
and Oke, 2012):
– Sensors within the UCL are probably affected by the environment within a radius of up to
500–1 000 m.
– Sensors should be located within a reasonably homogeneous part of the LCZ and away
from borders between LCZs with different surface properties.
– Sites with anomalous structure, surface cover, materials or other properties within the
thermal source area (Subsection 5.4.3) should be avoided. This applies in particular to
unusually moist or dry patches or sites near a concentrated heat source such as a heating
plant or an outlet of a heating, ventilation and air conditioning (HVAC) system.
– Sites along a small urban canyon (large aspect ratio H/W) at the same distance from the
buildings and with buildings of mostly similar materials (for instance LCZ 1) provide a
relatively similar temperature signal, if they are aligned with the street axis and situated on
the same side of the street.
– If the site is within an urban canyon, the dimensions (height, width and length) of that
canyon should ideally be representative of the surrounding neighbourhood.
– For compact built zones (for instance LCZ 1–3) “representative” implies a sheltered urban
canyon with paved ground; for open built zones (for instance LCZ 4–6) it implies an
exposed setting with vegetated ground, scattered trees and nearby buildings.
For continuous monitoring, north–south oriented streets are favoured because there is less
phase distortion in the temperature signal (from solar radiation), although the daytime course
of temperature may be more peaked. However, for some applications, east–west aligned roads
might be a better choice to avoid distortions due to prevailing winds or flow channelling effects.
If possible, north–south and east–west aligned streets should be observed together.
5.4.5 Vertical positioning
For rural stations, the WMO standard screen‑level height (see Guide to Instruments and Methods of
Observation (WMO-No. 8), Volume I: Measurement of Meteorological Variables, Section 2.1.4.2.1)
of 1.25–2 m AGL is recommended. Using the same height AGL for urban and rural stations is
recommended. As adhering to this guideline may be more difficult in urban areas, it can be
relaxed to allow greater heights, considering the vertical character of the temperature gradient in
the urban canopy. With the increased turbulence and vertical mixing in the urban environment,
often the stratification is neutral with narrow canyons having only slight air temperature
gradients (for heights >1 m AGL) through most of the UCL. Therefore, measurement data at
3–5 m AGL are only slightly different from those at the standard screen‑level height. In this height
range they are beyond easy reach (to prevent vandalism) and are out of the path of vehicles,
and they also have a slightly larger thermal source area. The higher measurement heights also
ensure greater dilution of vehicle exhaust heat and reduce contamination from dust (see Guide to
Instruments and Methods of Observation (WMO-No. 8), Volume III: Observing Systems, Chapter 9).
Installing sensors on street poles (such as those holding street lights or road signs) may simplify
access, if permissions can be obtained from a single utility provider or authority, and allow a
consistent installation height. Some structures may provide access to mains power.
Roofs should be avoided for determining the CL‑UHI for most use cases as the sensors installed
there do not provide canopy layer temperatures.
5.5 SENSORS
5.5.1 Instruments
There is a wide range of sensors that can measure air temperature (see Guide to Instruments and
Methods of Observation (WMO-No. 8), Volume I: Measurement of Meteorological Variables,
Chapter 2; Foken, 2021), but not all are appropriate to calculate the CL‑UHI intensity. Some
sensors measure in the wrong part of the vertical structure of the atmosphere (for example
radio‑acoustic sounding systems (RASSs), microwave radiometers) and/or are too noisy to
operate in urban areas without generating complaints from nearby residents (for example sound
detecting and ranging (SODAR), RASSs) or do not measure the air temperature (for example
LIDAR, satellite measurements of infrared temperature, surface temperature measurements
(Section 2.6; Appendix 1)).
Instrument requirements and operation should follow WMO guidelines (Guide to Instruments
and Methods of Observation (WMO-No. 8), Volume I: Measurement of Meteorological Variables,
Chapter 2). Sensors used can include electrical (resistance, semiconductor or thermocouples)
or glass thermometers. The latter, while cheap and not reliant on power, can only be operated
manually and are therefore not suited for long‑term monitoring. Sensors should be calibrated
at regular intervals following the guidelines (Guide to Instruments and Methods of Observation
(WMO-No. 8), Volume I: Measurement of Meteorological Variables, Chapter 2). Calibration in
the field when the atmosphere is well mixed, that is, during strong winds, is also acceptable.
The response time of the instrument should be sufficient to achieve thermal equilibrium in the
environment, but not so long as to mask trends caused by anthropogenic influences.
5.5.2 Radiation shielding and ventilation
Sensors need to be shielded from radiation to permit measurements in the presence of solar
radiation. Radiation reflected off sunlit walls, roads or shiny surfaces in the proximity of the
sensor can distort the measured signal and hence needs to be effectively blocked out. Since there
may not be sufficient space at urban sites to install a Stevenson screen, which is the traditional
standard for a naturally aspirated shield, compact radiation shields made of a series of stacked,
inverted bowl‑shaped plates with an open interior can be used. However, they may introduce
differences of up to 2 °C (–1 °C) relative to the Stevenson screen in sunny conditions with light
wind (on a calm clear night); smaller sensors are less susceptible to radiant effects than otherwise
identical but larger ones (Erell et al., 2005). The shield should also protect the instruments from
precipitation while still allowing the free circulation of air.
Since the shield itself can also be a heat source, forced ventilation (or aspiration) is preferred. The
latter is important to ensure that the sensor is in equilibrium with the adjacent air. Particularly
in sheltered canyon environments when wind speeds may be low, lack of ventilation can result
in erroneous readings being caused for instance by radiative errors from nearby sources such as
traffic heat or walls. Ideally, sensors should be >3 m from an obstacle to allow for less obstructed
airflow. Critically, to avoid sensor‑induced biases, similar sensor housings and measurement
protocols need to be used across all sites that will be compared.
(CL-UHI)
5.5.3 Mounting and measurement protocol
5.5.3.1 Fixed sites
Limited space, numerous regulations and sensor design may restrict mounting options in urban
areas. Traditional approaches to protecting stations such as fences may be neither possible nor
desirable. If a dedicated space is unavailable, existing urban infrastructure such as street poles
can be used to mount small, unobtrusive sensor packages. Care must be taken to avoid mounting
options that distort the measured signal. Solar panels, which tend to be prone to vandalism,
can be avoided by using a fixed power supply. Free Wi‑Fi networks or long‑range wide area
networks, if available, permit data transfer in real time. Real‑time monitoring and automated
post‑processing allow quick detection of measurement issues, which can help to minimize gaps
in the data set.
Measurements should be conducted continuously, if possible. The sampling interval should

be at least once every second, and averages should be calculated and reported for 1–10 min
periods following the WMO automatic weather station (AWS) guidelines (Guide to Instruments
Subsection 2.1.3.3 and Volume V: Quality Assurance and Management of Observing Systems,
Subsection 2.1.2). Clearly, for manual systems, longer sampling and reporting intervals are
more practical.
Quality control of the recorded data should include tests for step changes to identify and remove
erroneous data caused by irregular anthropogenic activities for instance by vehicles stopping at
the site, construction work nearby or sun glints from glass at particular times of year at particular
solar angles. Real‑time or near‑real‑time monitoring is useful to perform online data quality
control and quickly detect issues with the measurements.
Regular maintenance visits should be part of the measurement protocol (Section 5.6), especially
if the data are used for monitoring (Chapter 7). The physical condition of the instruments needs
to be inspected not only to ensure that they are in their regular operational state but also to
check for vandalism. Additional cleaning may be necessary given the prevalence of particulate
matter, soot and other air pollution in urban areas that might be deposited onto the sensors
and/or radiation shields and thereby affect the readings.
5.5.3.2 Sensors on mobile platforms
Sensors mounted on a mobile platform need to avoid heat sources on the scale of facets (10 m)
from both the platform and the surroundings such as vehicles or other canopy layer point sources
of heat. Detailed wind‑tunnel studies, either self‑conducted or based on the literature, may be
needed to assess appropriate locations on mobile platforms. Mounting sensors to the front,
and preferably at some distance out from the platform, should increase the exposure to the
ambient air and reduce impacts from platform heat sources (exhaust or motor). A sensor height
of 1.5 m or higher avoids the lower‑level exhaust emissions from car‑type vehicles. Radiation
shielding and adequate ventilation are required (Subsection 5.5.2), the latter especially at low
vehicle speeds.
The response time of the sensor affects the ability to resolve fine spatial scale temperature
differences and their apparent location. A sampling rate of at least once every 1 s using a sensor
with a matching response time is recommended; spatial sampling controls are needed to avoid
oversampling and a resulting misinterpretation of data when the platform is stopped. Detailed
consideration needs to be given to platform speed, spatial heterogeneity and sensor source
area. Speeds should not exceed 60 km h–1 to avoid dynamic influences on the temperature
measurement or these influences need to be corrected.
Traverse data need careful quality assurance and control to check for effects from heat sources
linked to the mobile platform that are not representative of the near‑surface UCL in general.
Temperature spikes might occur associated with vehicle exhaust emissions at intersections when
the platform is in proximity (<5 m) to other motorized vehicles. Data analysis needs to consider
the sampling frequency in terms of both temporal intervals and the corresponding spatial
locations, to consider the representativeness for example of a constant sampling frequency when
stopped at a traffic signal. Hence, accurate geo‑referencing of the measurements is essential.
5.6 METADATA FOR OBSERVATIONS
Collection and regular reporting of site surroundings metadata is an important part of any
measurement programme. Metadata categories are given in WIGOS Metadata Standard
(WMO-No. 1192) for WIGOS observations. Site metadata describe the location, instrument
systems and sensor exposure. They provide the basis to describe the respective urban and rural
stations selected to define the CL‑UHI as well as for comparisons across different urban and rural
sites or to document changes in time with urban development or land use change. Furthermore,
metadata are needed to properly interpret CL‑UHI intensities across different cities. Metadata are
also important for using the data from other agencies or for other urban services, given that there
is not always the opportunity to set up new stations, and selection from existing stations may
be needed.
Table 5.1 summarizes the recommendations regarding critical quantitative and qualitative
metadata that should be collected and updated over time. These relate to the instrumentation,
sensor mounting platform and site surroundings at both the microscale and local scale. On all
Table 5.1. Metadata needed to support air temperature measurements for determining
the CL‑UHI. For all categories, a time series must be collected.
Observing
Sensor Microscale Local‑scale
platform
Manufacturer Sensor height Height AGL Topographic setting of the site
AGL (including altitude)
Model Relation to local‑scale
Sensor distance area Relation to the larger urban area
Serial number from the
mounting Surface cover (built‑up, Distance to major changes in
Maintenance structure paved, vegetated, bare surface character
history including soil, water)
cleaning Platform location Distance and direction to major
Structure (dimensions geographic features (e.g. water
Calibration history Platform type of buildings and spaces bodies, hills, valleys, swamps,
between them, street deserts, forests)
Aspiration and Platform porosity widths and street
shielding spacing) LCZ class (Section 4.3)
Sensor
Measurement orientation Fabric (construction and Surface cover fractions (built‑up,
uncertainty, natural materials) paved, vegetated, bare soil, water)
precision and Surfaces
response time of underlying Human activity Structure (dimensions of buildings
the sensor individual (emissions of heat, water, and spaces between them, street
instruments pollutants) widths and street spacing)
Sky view factor Vegetation and Fabric (construction and natural

phenology materials)
Ground view
factor Short‑term changes Human activity (emissions of heat,
(e.g. construction water, pollutants)
activity, road works)
Vegetation and phenology
Short‑term changes
(e.g. construction activity, road
works)
(CL-UHI)
site visits, photographs should be taken of the instruments, platform and surroundings so that
changes are documented. It may not be until sometime later that an aberrant measurement is
discovered, for instance related to a small but influential change in sensor exposure, such as new
windows casting a sun glint, vegetation phenology or growth. Photographic records can help to
explain why the change likely occurred; however, permanent photography might not be possible
due to privacy restrictions. All metadata need to be updated over time, especially after significant
changes in the microscale or local‑scale surroundings.
5.7 DATA TRANSFER AND AVAILABILITY
Transfer of raw data from the sensor to the storage will ideally be automated but may need to be
manual if a data network such as the mobile phone network or Wi‑Fi is unavailable. Data stored
locally on a data logger should be downloaded regularly (for instance hourly (automatic), daily
(automatic), monthly (manual)). Frequent downloads are needed for many IUS and multi‑hazard
early warning systems (Section 3.5) to allow use of the data in near real time but may be costly.
Automated downloads reduce the probability of poor or missing data as real‑time checks and
quality control can occur, facilitating rapid repair if necessary. In some cases, advanced “edge
computing” capabilities may enable data storage, quality control and compression to happen at
the station (edge) rather than at a central computer so that processed data can be transmitted
more efficiently.
To ease data use, following the variable naming and format in the Guide to the WMO Information
System (WMO-No. 1061) and the Manual on the WMO Information System (WMO-No. 1060) may
help to ensure standardization in data, information and communication procedures.
5.8 OBSERVATIONAL CHALLENGES IN BRIEF
The wide range of measurement techniques that exist for measuring atmospheric and other
closely linked variables such as those related to soil, hydrology or vegetation are relevant to
the urban environment (Foken, 2021) and the rural comparisons site(s) (see Subsection 5.4.1).
It is important to choose a suitable type of sensor, with the appropriate type of shielding
(see Section 5.5 and Guide to Instruments and Methods of Observation (WMO-No. 8), Volume I:
Measurement of Meteorological Variables) at the appropriate site (Section 5.4). As noted
(Section 5.3), lack of awareness of how a sensor operates and therefore of the thermal source
area the sensor will represent (Subsection 5.4.3), or lack of consideration of the original objective
in siting a sensor in a particular location (Section 5.4), will lead to results for the CL‑UHI values
that are not helpful in the use case considered (Table 2.2), even if the sensor is an excellent
instrument.
The LCZs (Figure 4.1) were originally designed to allow a global, cross‑cultural description of
the neighbourhood area surrounding a measurement site to aid interpretation of the CL‑UHI
observed (Stewart and Oke, 2012). It is essential to consider neighbourhood scale, but the
microscale also needs consideration as poor siting (for example, a site with too much influence
from close‑by facets) will make the LCZ consideration irrelevant. As the atmosphere has larger
spatial variability close to the surface, the central task is to ensure good microscale siting for
investigating the process of interest (Section 5.4). This is challenging in cities because of the wide
range of materials and facets nearby.
There are numerous reasons to undertake observations within urban areas, beyond observing
the CL‑UHI (for instance wind, humidity, radiation). Some of these additional measurements
may be very helpful for purposes related to the CL‑UHI including bio‑meteorological indices
(Subsection 3.1.1), data assimilation (Subsection 6.1.3), model evaluation (Section 6.3),
interpretation of CL‑UHI data, or for supporting other needs for an IUS (Section 3.5). If the
data taken for CL‑UHI assessments are to be used for another purpose than that for which they
were originally intended, consideration needs to be given to whether the new purpose is an
appropriate use of these data. The site selection, sensor, metadata and other details need to be
reviewed in light of this. Therefore, metadata are essential (Section 5.6).
CHAPTER 6. DETERMINING THE CL‑UHI WITH MATHEMATICAL MODELS
There are three mathematical model types to simulate CL‑UHI intensities (Section 6.1):
statistical, obstacle‑resolving and numerical weather prediction (NWP) models. With many
numerical models, numerous processes that cause the CL‑UHI can be modelled through time in
conjunction with the regional‑scale heat, moisture and momentum exchange between surfaces,
UCL, UBL and rural PBL. However, like observed data, numerical models have limitations that
make these tools only applicable for specific purposes and scales. Not every CL‑UHI metric
can be calculated with every model type (Table 6.1). Any attempt to simulate the CL‑UHI with
atmospheric models parameterizing urban effects will need to consider multi‑scale effects. These
can be included either explicitly (model domain corresponding to region of interest), through
nesting, or by employing stretched grids. With models that have been previously evaluated for
the intended application range and the metric to be used (Section 6.3), short‑term assessments
as well as those related to climate or urban development are possible (Section 6.4).
6.1 MODEL TYPES
6.1.1 Statistical models
Statistical relations for the CL‑UHI dependence on both meteorological conditions and
morphological characteristics are widely used. The statistical models construct spatial and
temporal characteristics of the CL‑UHI from environmental parameters and require only
small computational costs. Models may be applied using few observations, however model
development benefits from large, high‑quality, long‑term data sets within both the urban and
rural areas (Hu et al., 2016). Observations can include multi‑scale atmospheric processes that are
not only surface driven. Meteorological data, such as for wind speed and cloudiness, are needed
to derive reliable statistical models for CL‑UHI assessment.
The predictor variables may be determined through a variety of artificial intelligence/machine

learning/statistical techniques, including regression analysis (Wilby, 2003) and neural networks.
The temporal variability of the CL‑UHI may be calculated using local meteorological variables
such as near‑surface wind speed, cloud cover, near‑surface humidity or synoptic‑scale weather
patterns, yielding an explained variance of 50% and above. To refine the spatial pattern of the
CL‑UHI beyond the observations, morphological parameters (for example building height or
building density, land use, land cover, surface roughness), material characteristics (for example
albedo, thermal capacity), and geographical factors (for example terrain height and form,
distance to water bodies) are employed.
Combining statistical CL‑UHI models with NWP (Subsection 6.1.3) can extend the use of the
statistical model, for instance by helping identify situations that increase or weaken the intensity
of the CL‑UHI using local weather types and using this for a forecast.
Limitations of statistical models include the following (summarized in Section 6.2):
– There is a need for long‑term, high‑quality training data for developing the
statistical relation.
– As the CL‑UHI is relatively independent of CL‑T, but depends on meteorological conditions

and location, these other data such as wind speed and radiation (Section 2.5) are needed at
the site of the air temperature observations or a station nearby.
– The applications to scenarios are limited to situations either covered by previous data or
consistent with data used for training.
CHAPTER 6. DETERMINING THE CL-UHI WITH MATHEMATICAL MODELS 39
6.1.2 Obstacle‑resolving models
Obstacle‑resolving models (ORMs)1 aim to explicitly resolve processes in the UCL with buildings
and trees being realistically included at a spatial resolution of ~1 m. Typical model domain sizes
are 1–100 km2. Using Orlanski (1975) scaling, ORMs are also named microscale models.
The 3D time‑dependent temperature, airflow and humidity fields are calculated fulfilling the
conservation laws of mass, momentum and energy. The ORMs employ uniform, stretched
or non‑uniform computational grids in the 3D domain. ORMs parameterize sub‑grid‑scale
turbulent processes typically using either a RANS or LES approach. LES resolves the time
dependency of the larger but still small vortices (size about 5–8 times the grid width) within
the UCL, while RANS provides time‑averaged values (averaging time ~1–10 min) with the same
spatial resolution as LES, if using the same grid. One realization with an LES model represents an
ensemble member; these instantaneous values are not comparable with measured data. Thus,
statistical analysis of LES results such as averaging in time or averaging of multiple realizations
is needed. Direct numerical simulation (DNS) resolves nearly all turbulent processes but needs
even higher resolutions and thus requires larger computational resources; currently it is not used
for operational applications in urban meteorology.
For UHI purposes, ORMs can simulate UCL flow and should calculate temperature fields,
including radiation and surface heating processes. However, if dynamic radiative exchanges are
absent, and heat fluxes are only prescribed at the urban surfaces, the surface heating is crudely
parameterized. Also the lack of other important physical details, for example clouds, convection
or precipitation, limits the applicability of the ORM. Simulating diurnal cycles of the CL‑T and
thus the CL‑UHI is only possible with inclusion of the dynamic radiative and energy balance
exchanges. Active, ongoing research is improving many aspects of the ORMs.
With the high resolution of the ORM, its time step is very small, as it is limited by the Courant–
Friedrichs–Lewy (CFL) criterion for ensuring the stability of numerical methods. The number
of grid points additionally determines computer resources needed for integration. These are
therefore large, and ORM runs are often limited to hours or days. Thus, the horizontal extent of
the model domain often only includes a neighbourhood (1–2 km) and not the region of interest
with the urban and rural areas. Furthermore, the vertical extent of the ORM may be limited to
the ISL (that is, a few building heights), but frequently at least the PBL is included. Some ORMs
extend to/above the UBL/PBL to include cloud effects. As is common in atmospheric models,
most ORMs use a decreasing vertical resolution at higher model levels, typically coarsening the
grid well above building height.
Scenario simulations are possible with ORMs by replacing urban features with natural
surfaces to determine microscale to local‑scale UCL effects. With urban/non‑urban scenarios,
neighbourhood‑scale modifications of the CL‑T and its influence on the CL‑UHI can be evaluated
from differences in model results. However, the whole intensity of the CL‑UHI relative to the
non‑urban situation depends on the prescribed non‑urban surface characteristics. In addition,
advective transport from other neighbourhoods might be missing. This and/or differences
with respect to rural areas can only be calculated if the spatial extent of the simulation domain
includes surrounding rural areas. Ideally, two‑way coupled or nested models are used,
or stretched grids are employed in the ORM. For deriving the resulting CL‑UHI intensity,
atmospheric boundary conditions and boundary values are critical.
Limitations of ORMs include the following (summarized in Section 6.2):
– The horizontal extent often only includes a neighbourhood (1–2 km) and not both the
urban and rural areas in one model domain.
1
ORMs used in CL‑UHI investigations are either extended computational fluid dynamics (CFD) models that include
thermodynamic processes or meteorology models that resolve building, tree and other obstacle effects including
their thermodynamics.
(CL-UHI)
– Care is needed when using an ORM and its results: for instance, are non‑neutral
atmospheric stability conditions simulated, are all appropriate energy exchange processes
simulated, are anthropogenic heat fluxes parameterized?
– The ORMs need to have sufficient resolution to resolve buildings and trees.
Parameterizations for partly‑resolved buildings currently do not exist.
– LES models need to use a sufficient resolution to also resolve small vortices with stable
stratification at night, otherwise the turbulent mixing may be underestimated.
– High‑resolution data on morphology and thermal and radiative properties are required,
but these might not be available for all areas.
– Providing model boundary values is particularly difficult for realistic simulations

with LES models.
6.1.3 Numerical weather prediction and climate models
Regional atmospheric and NWP models with the capability to account for urban energy and
momentum fluxes at the ground may be able to simulate mesoscale variations of CL‑UHI
intensity. They cover domains on the order of 102–103 km and more in the horizontal direction at
a grid resolution on the order of 1 km. At these spatial scales, these models can resolve mesoscale
atmospheric phenomena. To avoid any confusion with the scales used for classifying urban form
(Table 2.1), these models are referred to as regional models (RMs), regional climate models
(RCMs) or, if they cover the globe, global models (GMs) and global climate models (GCMs) in
the present guidance.
Operational weather forecasting with urban characteristics accounted for is still uncommon
in most NMHSs, but such models are already operational in 13 countries in Europe as well as
in Brazil, Canada and northern China, with typical resolutions of 1–4 km, and they are used
by research institutions. Like ORMs, RMs and GMs numerically calculate the time‑dependent
3D temperature, airflow and humidity fields, as well as cloud, rain and – for some models – air
pollutant concentration fields fulfilling the conservation laws of mass, momentum and energy.
As with ORMs, the computational cost is controlled by spatial extent, resolution and related
time step (CFL criterion). All models employ land surface schemes to simulate the heat, moisture
and momentum exchanges for each surface environment. This scheme needs to ensure that a
consistent approach is taken for both the urban and the rural surroundings to allow the CL‑UHI
to be determined from the simulation results as the difference between urban and rural areas.
However, with coarser resolutions than ORMs, individual buildings and other urban structures
(for example, trees) are not resolved, though their effects may be considered in the urban
canopy parameterization (UCP). Some model resolutions may be too coarse, or the models may
not have any UCP (such as the North American Mesoscale Forecast System (NAM), with 12‑km
horizontal grid spacing; or the High‑Resolution Rapid Refresh (HRRR) model), thus making
them inappropriate for direct CL‑UHI calculations. To calculate a CL‑UHI metric (Table 2.2) using
NWP results derived without an UCP requires post‑processing (for example a statistical model
(Subsection 6.1.1)).
In urban areas, the influences of built form and surface cover including vegetation and bodies of
water should be simulated or their effects parameterized. There have been numerous approaches
to try to capture critical features. The UCP approaches include the following:
– Bulk: This may not distinguish building facets, but combines the built land cover in one tile.
– Single‑layer: This separates the built facets such as walls and roofs and may use a
tile approach.
– Multi‑layer: This can have multiple auxiliary model levels within the RSL and UCL. The heat,
moisture and momentum exchanges are calculated by a turbulence scheme. The schemes
also require the radiation and anthropogenic heat fluxes to be vertically resolved. To
directly obtain the CL‑UHI, a high vertical resolution and the lowest model levels below 5 m
AGL are needed, making the models more computationally expensive.
A key difference amongst the various UCPs is the number of facets and other surfaces that
are considered within a grid cell. The UCPs also differ in how the surface characteristics are
aggregated (tile approach) or in the number of auxiliary height levels used (Martilli et al., 2002).
The UCP may provide integrated fluxes for all the surfaces within a grid cell. However, quite
commonly a tile approach is used, where the impacts of vegetated (and other) surfaces are
treated separately from those of built surfaces and combined at first model level. This level
may be the displacement height or higher or some other height above the roughness length,
assuming the displacement height is accounted for (for example, assumed to be included in the
ground level height).
Some models additionally include parameterizations of the urban water budget. The approach
used impacts the turbulent flux (sensible/latent heat) partitioning model skill.
Like ORMs, RMs require initial and boundary values to be provided from a larger‑scale (often
global) model to correctly represent the larger scale meteorology and soil moisture. The model
performance may be optimized by data assimilation in the RM, but care has to be taken to
consider the representativeness of the data used; they might have only microscale or at best
local‑scale representativeness (Subsection 5.4.3). Data assimilation at local scale and microscale
is still a research activity, although forward operators have been developed that help the
evaluation of UCPs (Warren et al., 2018).
The NWP and RMs simulate the mesoscale and regional‑scale motions and boundary layer
turbulent exchanges and take into account the meteorological effects at different scales such
as influences of cities on sea breezes and fronts. The large domain size allows whole cities and
their surrounding rural areas to be simulated, with the smallest resolved phenomenon about
5–8 times the grid width. The CL‑UHI can be simulated for any situation, including day or night,
complex terrain, large water bodies and different seasons, within the limitations of the model.
The models allow assessments of the CL‑UHI for changes in climate and for urban development
scenarios, if the urban features are properly considered in the simulations. CL‑UHI intensity
can be calculated as the difference between urban grid cells and surrounding rural grid cells.
This requires consideration of many of the same issues as when the CL‑UHI is obtained from
observations (Chapter 5). Alternatively, the surface cover can be prescribed as non‑urban in a
scenario simulation so that the temperatures can be simulated without urban influences and the
temperature differences can be calculated and interpreted as the CL‑UHI. As previously noted,
non‑urban surface characteristics are also important.
For calculating the CL‑UHI, values at ~1.5 m AGL should be used (Section 2.3). The lowest
model levels of RMs and GMs are ~10 m AGL, but higher levels (~45 m AGL) are still in use. As
the CL‑T is needed below this lowest atmospheric model level, it can be taken from auxiliary
grid level values (multi‑layer UCP) or needs to be diagnosed. The latter often assumes the use of
Monin–Obukhov similarity theory (MOST) or MOST‑derived approaches, despite MOST being
only applicable in the ISL (Roth, 2000) and not in the UCL (Section 2.2). Hence, care is needed
in using the derived CL‑T and resulting CL‑UHI intensities. Other ways to diagnose CL‑T are to
use an equation linking the heat fluxes between the various surfaces and the atmosphere, or an
offline vertical turbulence scheme with several layers, or to use a multi‑layer UCP with the lowest
auxiliary level at ~1.5 m.
Limitations of NWP and climate models include the following (summarized in Table 6.1):
– Large cities that cover several model grid cells can impact the modelled atmospheric
environment if properly resolved. Note that 5–8 grid lengths is the scale of the atmospheric
process that is captured by a model, thus 1 km grids do not cover the urban effects at
neighbourhood scale. As the next generation of RMs, now being tested, has grid lengths
on the order of 100 m, they will be able to represent surface flux heterogeneities within the
neighbourhood scale.
(CL-UHI)
Table 6.1. Advantages and limitations of different model types and urban canopy
parameterizations
Model type/UCP Advantages Limitations

Statistical model/ Traditional approach with many Completely dependent on urban/rural
none references observations available and sites chosen
Low computational costs High‑quality data needed
Relatively easy to set up Need variables (e.g. wind speed,

radiation) in addition to rural and urban
temperatures
Microscale to local‑scale
representativeness
Robustness of CL‑UHI values depends on

number of sites, length of time series
Statistical analogue needed to use these

models for scenario assessments
Daily mean by synoptic condition:

observations for different synoptic
conditions needed
Spatial pattern: observations at several

sites and morphology data needed
Seasonal or annual mean: at least one year

of data needed
Numerical model/ Traditional model approach (NWP, RM, Does not directly deliver results
bulk approach GM) with many references at ~1.5 m AGL; CL‑UHI metric needs
vertical interpolation/extrapolation
Fast to integrate on a computer (Subsection 6.1.3) and/or statistical model
for post‑processing
Nesting available
Vertical extent of buildings treated by
Non‑urbanized simulation for logarithmic law or not considered
comparison possible
Vertical interpolation has to consider the
Model area includes at least region of sub‑grid surface data
interest (Section 2.3)
Heat storage of urban fabric not always
Temperature differences can be considered
calculated in space and time
Daily mean by synoptic condition:
simulation for different synoptic conditions
needed
Seasonal or annual mean: at least one year

of model results or statistical‑dynamical
downscaling needed
Model type/UCP Advantages Limitations

Numerical model/ Fast to integrate on a computer Same as bulk approach
single layer UCP
Nesting available Requires more computing and data
characterization resources than above
Sub‑grid‑scale surfaces considered to approaches
calculate heat storage and emission
Vertical heat/humidity/ momentum

exchange within urban area calculated
in ISL
Anthropogenic heat may be added
Non‑urbanized simulation for

comparison possible
Numerical model/ Compared to ORM, fast to integrate on High vertical resolution with the lowest
multi‑layer UCP a computer model or auxiliary level below 5 m
is needed for direct CL‑UHI intensity
Nesting available calculation (otherwise interpolation
method is needed, see bulk approach)
Vertical profiles of heat, moisture,
momentum fluxes calculated within the Requires computing resources larger
UCL than for single‑layer approaches due to
time step limitations and more complex
Vertical variability in heat storage, calculations
radiation and anthropogenic heat
emission can be considered
Non‑urbanized simulation for

comparison possible
Temperature differences can be

calculated
Numerical model Each building and tree can be Without rural areas in the model domain
(ORM)/UCL realistically included or nesting, only microscale to local‑scale
resolving urban effects can be simulated; however,
Local flow, humidity, heat and radiation pre‑urban scenarios to determine changes
patterns simulated in time‑dependent in CL‑UHI intensities are possible
3D
Without temperature calculation and
Non‑urbanized simulation for consideration of atmospheric stabilities,
comparison is possible not usable for CL‑UHI assessment
Microscale and local‑scale influences Substantial computing resources needed

can be determined (time and space)
Lowest level below 3 m, model High‑resolution input data required

temperature values can be directly
used Boundary values are difficult to provide for
LES‑type models
Influences of neighbourhood changes
in urban fabric on microscale to Daily mean by synoptic condition:
local‑scale CL‑UHI changes can be simulations for different synoptic
assessed conditions needed
Seasonal or annual mean: at least one

year of simulation or statistical‑dynamical
downscaling needed
(CL-UHI)
– The current UCPs need further testing to ensure they properly account for the effects of
urban form in model simulations with grid widths below 1 km, where large buildings
may partially or completely cover a grid cell (for example train stations, exhibition halls,
industrial plants).
– The CL‑UHI needs near‑surface temperatures at ~1.5 m AGL. With a fine auxiliary grid
(multi‑layer UCP) these might be directly available from model results; alternatively,
they need to be diagnosed, which often assumes the use of MOST or other vertical
interpolation schemes.
– The lowest (auxiliary) model level needs to be much higher than the local roughness
length to determine heat and momentum fluxes in the energy balance at the surfaces using
wall functions.
6.2 CALCULATION OF A CL‑UHI METRIC FROM MODELS
An ideal way for determining the CL‑UHI would be to have data on the pre‑urban setting. This
may be possible in land‑use scenario studies with models. The pre‑urban land use might be the
same as the current rural land use in the surrounding areas, so that the results show the CL‑UHI
based on the current urban area versus the current rural area. However, other pre‑urban land
uses are possible (for instance completely forested, grass, swamp). The selected pre‑urban
land use will influence model results and thus the near‑surface temperature difference for
the scenarios. An alternative, applicable for many applications, is to take model results for an
upwind and preferably undisturbed area in the rural surroundings of the urban area as a rural
(non‑urban) reference, as is done for observations.
The different model types introduced in Section 6.1 may all be used for performing simulations
including pre‑urban land‑use scenarios. However, not all are currently applicable to calculate
the CL‑UHI based on urban and rural CL‑T values. Table 6.1 summarizes model approaches
and their advantages and limitations. In general, all model types can be used for each metric
(minimum, maximum, daily mean, daily mean by synoptic condition, spatial pattern, seasonal
or annual mean). Depending on the model type, the resolution can be very high (some metres).
For statistical models, the limitation is dictated by the training data, which have to have at
least the spatial and time resolution needed for the metric considered. For NWP (RMs and
GMs) and corresponding climate models hourly data are possible, while ORMs can provide
time‑dependent data on a basis of some minutes or even shorter (LES‑type models). In addition,
besides large‑scale meteorology, the quality of the model type‑specific data on the urban area
(Table 4.1) is important for achieving a reliable model result. Please note: the quality of the
model results depends on model skill (Section 6.3) and how well the user is trained in the use
of the model.
6.3 EVALUATION OF MODEL SKILL
Model performance should be assessed with observations and should include more than CL‑T
data (Warren et al., 2018) for both the urban area and the surrounding rural areas. In both
environments the skill to model vegetation and soil moisture is critical. Evaluations should not
only compare CL‑T values or CL‑UHI intensities but should also focus on the performance for the
relevant CL‑UHI metric (Table 2.2) for the use case. Thus, the evaluation needs to assess if the
model is applicable for the intended use.
The needed skill cannot be generally prescribed but depends on (i) the purpose of the
modelling, (ii) the resolution of the model used, and (iii) the uncertainty inherent to the class
of the observational site. For example, the estimated uncertainty including the measurement
uncertainty of 0.2 K is 1.2 K for an urban Class 3 site, while for a typical Class 1 rural site it is 0.2 K
(see Guide to Instruments and Methods of Observation (WMO-No. 8), Volume I: Measurement
of Meteorological Variables, Annexes 1.A, 1.D). Thus, for a single pair of measurements, the
simulated maximum CL‑UHI may differ by about 1.4 K for a model that does not resolve the
microscale effects but uses UCP (Subsection 6.1.3). Note, within this deviation the model is
assumed perfect, while modelling assumptions or constraints such as the vertical resolution
impact simulated CL‑UHI intensities. Knowing the degree of uncertainty in the simulated CL‑UHI
and its spatial distribution is important to determine the model’s applicability for a specific use
case (Table 2.2).
Ensuring consistency between the observations (Section 5.2) and the model output (for example
location, surface cover, thermal source area, spatial representativeness) for a range of different
neighbourhoods and within those neighbourhoods is important for assessing the model
performance and for simulating the CL‑UHI spatially and temporally. If observations are used for
data assimilation, care needs to be taken to account for the instrument siting (Chapter 5). Data
assimilated in a model simulation cannot be re‑used for evaluation; instead, independent data
are needed.
When simulating the current CL‑UHI (analyses, nowcasting, NWP) or past CL‑UHI (reanalyses,
past sub‑seasonal to annual), sufficient observational data may exist (Section 7.4) for thorough
model evaluation. However, for future climate (decadal or longer) CL‑UHI, models can only
be evaluated for the current or past climates and with current or past urban characteristics
such as fabric and morphology. Sufficient model performance is a prerequisite to applying a
model for assessing the CL‑UHI dependent on future climate and/or on urban development.
However, modelling provides the only opportunity to assess the future changes of the CL‑UHI
depending on, for instance, city growth or climate change. Therefore, different evaluation
approaches (operational, diagnostic, dynamic, probabilistic) (Dennis et al., 2010) may be applied
to assess the reliability of a model for scenarios of urban development as well as future climate.
For example, dynamic model evaluation focuses on the ability to predict changes in the response
of the model to changes in meteorological conditions and, more generally, to changes in forcing.
This includes changes in the urban characteristics and thus model application to different types
of urban areas. Diagnostic evaluation might focus on single processes in a model, for example
how well wind speed or heat fluxes are simulated using corresponding measurements. Given the
challenges of assessing future climate CL‑UHI prediction, a variety of evaluation methods need to
be considered.
6.4 APPLICATION OF MODELS TO DETERMINE THE CL‑UHI
6.4.1 Analyses of current and past CL‑UHI
The current CL‑UHI can be analysed with all model types within their type‑specific advantages
and limitations (Table 6.1). For this purpose, observations might be assimilated in the model,
however, care has to be taken to correctly include the representativeness of the data and what
their thermal source area is. For example, an urban observation of microscale representativeness
(very small thermal source area; Class 3–5 as per Guide to Instruments and Methods of Observation
(WMO-No. 8), Volume I: Measurement of Meteorological Variables, Annex 1.D) might be
inconsistent with a model with a 1 km grid, but suitable for use in ORMs with 1–10 m grids or
NWP model with grids well below 100 m.
To assist interpretation of observational data (Chapter 5), models that take into account dynamic
surface energy balance and vegetation phenology may be used to determine the impact of
microscale and local‑scale morphology on the CL‑T and CL‑UHI (Subsections 6.1.2, 6.1.3).
For reanalyses of past or even historic CL‑UHI it is important to know the urban form for the
period that the reanalysis is focusing on. For historic CL‑UHI values it is often difficult to obtain
this needed information about the urban area. However, as long as the urban area is little
changed, the current information on the urban form might be a good first guess.
(CL-UHI)
6.4.2 CL‑UHI forecasts (several days and shorter)
A very dense urban observation network may allow nowcasts of higher spatial resolution of the
CL‑UHI than the grid width of the forecast model (NWP, RM, GM). This may be critical to some
applications (Chapter 3). Weather forecasts can provide variables (temperature, cloud cover,
rain, soil moisture, wind) relevant to CL‑UHI intensity and spatial variability. Depending on
forecast model complexity and the UCPs employed, the forecast itself or in combination with a
statistical model can provide CL‑UHI information. However, mesoscale atmospheric phenomena
and urban processes need to be resolved in the CL‑UHI forecast to be appropriate (Section 6.1).
As noted, this is limited by model resolution and domain size as well as the UCP employed.
With the CL‑UHI considered, the short‑ and medium‑range weather forecasts (1 to 10 days) and
nowcasts (up to 6 hours) allow heat stress and extreme weather warnings for urban areas (Multi-
hazard Early Warning Systems: A Checklist; Guidance on Integrated Urban Hydrometeorological, Climate
and Environmental Services (WMO-No. 1234), Volume I: Concept and Methodology; Heatwaves
and Health: Guidance on Warning-System Development (WMO-No. 1142)).
In general, where features of low predictability are important, forecasts should be run as an
ensemble and the results presented probabilistically to consider uncertainty in the synoptic
evolution on the timescale of several days. It is important to avoid presenting information
to users which is not justified by the predictability of the solution. However, as the CL‑UHI is
strongly influenced by both the local‑scale urban form and the meteorological conditions in
the rural surroundings, it is likely to be predictable for longer periods than other shorter‑lived
phenomena such as urban area‑influenced or ‑initiated thunderstorms.
6.4.3 Climate predictions (sub‑seasonal and longer) and projections
As atmospheric conditions (Section 2.5) influence CL‑UHI intensity, the intensity is modified
by global and regional climate, and the CL‑UHI in turn influences the climate of the urban
surroundings. GCMs and RCMs that employ UCPs (Jacobson and Ten Hoeve, 2012; Katzfey
et al., 2020; Hertwig et al., 2021b) can, in principle, consider both future climate and
urbanization scenarios. The changes in the CL‑UHI may be due to regional changes in cloud and
wind conditions. The challenge lies in obtaining results at the appropriate spatial scales for urban
and rural sites to assess the CL‑UHI. To quantify uncertainties in climate projections, full model
ensembles are needed. Methods to determine the CL‑UHI are given in Section 6.2.
As most GCMs have coarse grid resolutions (~200 to ~50 km) (Haarsma et al., 2016), the
urban extent is insufficient to cover enough grid cells to be meaningful within the urban area.
Furthermore, in many models there is only a simple urban representation (bulk approach,
Table 6.1), or the urban parameters used in the model are poor at the global scale (Hertwig et al.,
2021b). Therefore, model and input data improvements are needed to enhance the model skill
for determining the CL‑UHI. High‑resolution model runs (~10 km) that are now available will
start to overcome some of the current limitations for megacities. The tile approach combines
the UCP with the non‑urban part within the same grid cell allowing the CL‑UHI to be derived
from the CL‑T with appropriate tile weightings to characterize different states of urbanization
(Section 6.1) in the same grid cell (Katzfey et al., 2020). However, this has several constraints.
The appropriate urban/rural CL‑T needs to be combined using appropriate land cover data. Thus,
the resulting CL‑UHI values should be treated as a rough estimate for a region. If a multi‑layer
UCP is employed, the vertical resolution should account for the urban form. Evaluation of GCM
performance under current conditions is essential to understand the nature of the CL‑UHI values
derived (Haarsma et al., 2016). Current GCMs using UCPs show small increases, decreases or
minor changes in the CL‑UHI for different times of year and regions compared to today’s climate
without changing the urban form (Oleson et al., 2011; Katzfey et al., 2020).
RCMs generally have a better grid resolution (~50 to ~1 km) than GCMs. The km‑scale RCMs
(Takane et al., 2020) permit predictions of CL‑UHI for larger cities.
Currently, GCM or RCM projections frequently use bias corrections typically based on results
for current climate (Guide to Climatological Practices (WMO-No. 100); WMO Guidelines on the
Calculation of Climate Normals (WMO-No. 1203)). Several grid points should be used to enhance
the reliability of the climate model results, and thus one would have to include urban and rural
areas with several grid cells each. If the model grid cell includes both urban and rural areas,
the bias corrections need to be weighted by area fraction. Bias‑corrected CL‑UHI will have little
reliability if the whole urban and rural areas are within the same grid cell. The CL‑UHI can be
determined from downscaled, bias‑corrected future climate data or using simulated atmospheric
data driven by the land surface parameterizations (Section 6.1) for the urban and non‑urban tiles
within a grid cell.
High‑resolution climate projections are computationally expensive. Downscaling methods to

estimate the CL‑UHI include using statistical models (Subsection 6.1.1) combined with GCM or
RCM projections (Wilby, 2003). The lower computational demand enables the use of ensemble
techniques. As the statistical models are sensitive to biases in the predictors, a bias correction
to the GCM/RCM input data is necessary. However, statistical models do not generally consider
changes in the urban characteristics (Subsection 6.1.1), and their reliability is highest where
training data are used. Given that changes in urban characteristics are very likely in climate
timescales, both should be considered. However, the prediction of urban development includes
uncertainties and thus scenarios are used (Subsection 6.4.4).
Statistical‑dynamical downscaling uses high‑resolution numerical model results of simulations

performed for selected weather types. These are identified from GCM or RCM results to describe
the temporal CL‑UHI variability or meteorological extremes (cuboid method) (Früh et al., 2011).
The GCM or RCM weather type frequency is used to weight the high‑resolution model results
to compute the projected climatological mean CL‑UHI. Statistical uncertainty is reduced by
increasing the number of simulations conducted for each weather type. For example, instead of
2 simulations per weather type (Hoffmann et al., 2018), 3–6 simulations were used by Schoetter
et al. (2020).
Another statistical‑dynamical downscaling approach uses short periods of high‑resolution RCM

simulations with and without urban surface characteristics. For each day the time‑dependent
spatial pattern of the CL‑UHI is determined, and the determined CL‑UHI signal is added to results
of RCM‑based meteorologically similar days employing an analogue approach.
Temperatures simulated by GCMs (Subsection 6.1.3) are often used to estimate the heat‑related
health impacts without including the intensity of the CL‑UHI. This leads to an underestimation
of the CL‑UHI effects on health. With refinement GCMs might be usable to consider the CL‑UHI
and estimate health impacts (see Takane et al., 2020 for mortality and exposure). This is also done
using regional numerical models and epidemiological analysis. The results show that the CL‑UHI
increases the health risks in warm seasons. However, the CL‑UHI may moderate winter‑related
cold and thereby have positive health effects in that season (Macintyre et al., 2021).
6.4.4 Urban development and the CL‑UHI
All model types introduced in Section 6.1 can be applied to assess the impact of urban
development, if they have sufficient skill (Section 6.3). Simulations may be performed for the
current climate with and without an urban development scenario to study the effect of urban
development in isolation or in combination with climate change (Subsection 6.4.3). These
simulations may help to identify the appropriate choice of urban development scenario. Essential
for scenario studies is the ability to include the urban fabric of the current situation and of the
scenario in the model. The challenge is to provide information on the future urban area such
as urban sprawl versus compact urban development, urban facets including the materials (for
instance wood versus concrete), or anthropogenic heat emissions (for example fewer commuters
due to more work from home in a digitalized world). To set up and assess the probability of these
scenarios, close collaboration with city planners and scientists is essential (Chapter 8). It should
be noted that statistical models may only be used for urban development scenarios if a statistical
analogue exists (Table 6.1).
(CL-UHI)
6.5 METADATA FOR MODELLING RESULTS
Model metadata are an essential part of providing model results. Krayenhoff et al. (2021) define
a useful set of criteria that may be considered by modellers or result users to assess the reliability
and contextualization of model studies. They suggest the following for documenting model
studies on urban heat reduction strategies:
– Provision of site metadata
– Characterization of forcing meteorology
– Provision of heat mitigation implementation metadata
– Provision of air temperature in space and time
They suggest that the model skill for the application be assessed using the following
three criteria:
– Accurate representation of relevant physical processes
– Successful model evaluation for the targeted model application (impact of heat
mitigation measures)
– Model application is sound
For scientific and public information considerations, modellers are encouraged to publish model
results (data sets) for assessing CL‑UHI as open data fulfilling the FAIR principles (findable,
accessible, inter‑operable, reusable). For example, the Coupled Model Intercomparison Project
Phase 6 (CMIP6) (https://pcmdi.llnl.gov/CMIP6/) provides guidance for modellers, data
managers and users including model output specifications for global and regional climate
simulations. Downscaled results for several regions are available (Haarsma et al., 2016; CMIP6:
https://pcmdi.llnl.gov/CMIP6/; World Climate Research Programme (WCRP) Coordinated
Regional Climate Downscaling Experiment (CORDEX): https://cordex.org/). They use
standardized variable names following the Climate and Forecast (CF) metadata conventions
(Andrioni et al., 2022). Standardization and publication of ORM results are also advisable to make
these computationally demanding model results more easily and widely usable, for instance
by using the new Earth System Data Branding (EASYDAB) (http://www.easydab.de; Ganske
et al., 2021). Many countries now have open data requirements which facilitate data use and a
more complete communication of methods.
6.6 MODELLING CHALLENGES IN BRIEF
Given the constraints to modelling the CL‑UHI (Table 6.1), the primary challenge results from
the gap between typical mesoscale model resolution (O(1 km)) and the microscale that needs to
be resolved to explicitly represent the CL effects (~O(1 m)). While advances have been made in
representing the complexities within the UCL, much work is aimed at improving the fluxes from
the UCL to the atmosphere above for urban NWP and climate assessments. Besides the correct
consideration of the horizontal heterogeneity of the urban area, the effects of individual tall
buildings or small groups of tall buildings have not yet been properly parameterized (Hertwig
et al., 2021a). The simulation of the urban vegetation, and its interaction with the other surfaces,
is also important and still a challenge.
To simulate the CL‑UHI microscale variability, further advances in ORM modelling are needed.
A major challenge is to ensure the microscale models are appropriately coupled to a regional
model, for instance by using stretched grids to ensure that the lateral boundaries include the
necessary rural environment, or by nesting within coarser atmospheric models.
A wide range of observations are needed to evaluate the characteristics of the modelled urban
atmosphere in 3D. In addition, detailed and complete information on the 3D city morphology
and building materials, as well as human activities, is needed as input (Chapter 4); this
information is needed regardless of the model scale or type, but in varying degrees of detail.
CHAPTER 7. MONITORING THE CL‑UHI
Monitoring the CL‑UHI allows provision of urban services (Chapter 3) as well as ongoing
assessments of the impact of both urban development and climate change on the
CL‑UHI. Furthermore, CL‑UHI monitoring is helpful for assessing the effects of potential
adaptation measures.
7.1 DEFINITION OF MONITORING
The goal of CL‑UHI monitoring is to provide long‑term information. For this purpose,
observations and corresponding data analysis are used, as well as modelling, or a combination
of both methods (Table 7.1). Monitoring requires observing all the relevant meteorological
variables (Subsection 2.5.4; Chapter 5), obtaining the urban–rural characterization parameters
(Chapter 4) and metadata (Sections 5.6, 6.5), combined with using the appropriate model(s)
(Chapter 6). The metadata are needed to interpret, homogenize and compare intensities of the
CL‑UHI at different times. The CL‑UHI metrics of relevance include the diurnal and annual values,
extremes (minimum, maximum), trends with reference periods, and intra‑ and inter‑urban
values. Databases from CL‑UHI networks for different cities could be useful in regional and/or
global upscaling for global climate solutions (Creutzig et al., 2019). The approaches used to
monitor the CL‑UHI are diverse with complementary attributes (Table 7.1).
7.2 PURPOSE OF MONITORING
Prior to designing a monitoring programme with one or several of the approaches summarized
in Table 7.1, the use cases and corresponding CL‑UHI metric have to be determined (Table 2.2)
with the purpose of identifying the monitoring system needed. The following steps have
to be taken:
– The intended use of the data from a CL‑UHI monitoring network has to be defined.
– Based on the intended use of the data, the necessary scale (spatial coverage and
resolution) of the monitoring network can be decided and for what purpose models are to
be employed (for instance spatial refinement of observational data).
– Given the intended use of the data, the necessary temporal resolution has
to be defined.
– Data accessibility, formats and any post‑processing needed for specific uses should
be assessed.
An existing measurement network may be insufficient for the intended use. However, the
spatial/temporal resolution needed for the intended use might be only achievable using
additional model simulations.
7.3 PARTNERING IN THE INSTALLATION OF MONITORING SYSTEMS
As urban observations and services involve multiple stakeholder groups, they present
opportunities for collaboration between NMHSs, agencies and research institutions (Guidance
on Integrated Urban Hydrometeorological, Climate and Environmental Services (WMO-No. 1234),
Volume I: Concept and Methodology and Guidance on Integrated Urban Hydrometeorological,
Climate and Environment Services (WMO-No. 1234), Volume II: Demonstration Cities). A
coordinated approach between operators of systems ensures that more comprehensive
observational and modelling data sets are available for the community, and that siting and
CHAPTER 7. MONITORING THE CL-UHI 51
Table 7.1. Advantages and disadvantages of CL‑UHI monitoring approaches
Approach Advantages Disadvantages

Observations Extensive temporal Limited spatial coverage
at fixed sites coverage
(Chapter 5) At some sites poor spatial representativeness
Metadata must be regularly updated to monitor changes

in surroundings
Potentially expensive if a large number of sites are used

Observations Good spatial coverage Extensive post‑collection data processing and quality
on mobile along the track used control needed
platforms
(Chapter 5) Detection of hot spots and Each observation’s surroundings differ
spatial variability
Metadata must be regularly updated to monitor changes
in the surroundings
Collecting field data is labour intensive
Measurements not all simultaneous in different parts of

the urban area
Speed of platform determines spatial resolution

Observations Good spatial coverage Extensive post‑collection data processing and quality
by citizens control needed
(opportunistic Detection of hot spots and
approach) spatial variability Observation surroundings often unknown (metadata
missing)
Unknown maintenance of instruments
Metadata may not be reliable or not be updated

Modelling Extensive temporal Spatial resolution might be insufficient
(Chapter 6) coverage
Data assimilation advisable for the assessment of the
Good spatial coverage current state
within model domain
Model parameters to characterize neighbourhoods need
Detection of errors and to be regularly updated
inhomogeneities in
observations Reliability of results depends on model physics
considered and successful evaluation for the intended
purpose
Metadata (e.g. model physics, resolution, input data,

urban form data) must be regularly updated to monitor
changes in the model
Regular model evaluation needed
maintenance of the monitoring system are done in coordination with relevant agencies.
Collaboration should start very early in the planning process so the observing systems and
model set‑ups can be designed to support multiple application areas from the outset. Numerous
agencies, institutions and companies may deploy sensors for temperature, wind, air quality
or hydrological parameters for different purposes (for example hydrology, health, air quality)
and with different goals (for example regulatory, safety, planning) and time frames such as
day‑to‑day management, extreme event preparedness or long‑term planning with a wide range
of scenarios. As NMHSs have a broad overview of requirements, gaps and standards, they can
play a central role in facilitating the evolution of observing systems and modelling possibilities
(CL-UHI)
within cities. The collaborative approach promotes overall cost efficiencies by standardizing
observation and modelling methods and avoiding duplication between agencies and
institutions.
Traditional NMHS meteorological networks are installed to observe synoptic or large mesoscale
features for the country, typically intentionally avoiding heterogeneous urban areas. Some
countries that are small (for instance Singapore) or have NMHSs with an urban focus have
already set up urban networks (Guidance on Integrated Urban Hydrometeorological, Climate and
Environment Services (WMO-No. 1234), Volume II: Demonstration Cities). NMHSs are generally
well positioned to implement the networks for CL‑UHI monitoring, as they deploy and
maintain systems over the long term. Those installing sensors must take into account the siting
considerations in Chapter 5. Given the heterogeneity of urban areas and the increasing need
for high‑resolution data in urban areas, it is likely that sensors in specific areas of interest will be
required and/or may be operated by non‑NMHS entities such as city agencies and companies.
City agencies may implement CL‑UHI monitoring, having monitored air quality already regularly
and for longer periods. Setting up combined networks may save resources, if both purposes can
be addressed, for instance determining exceedances of limit values for particles and maximum
CL‑UHI intensities. City‑funded CL‑UHI monitoring networks are likely to be permanent and
maintained as a public service (IUS: Guidance on Integrated Urban Hydrometeorological, Climate
and Environment Services (WMO-No. 1234), Volume II: Demonstration Cities and Multi-hazard
Early Warning Systems: A Checklist). As with air quality monitoring, expertise in meteorological
observation is important. If expertise is lacking, collaboration with the private sector and
others may be challenging. However, university researchers who have established CL‑UHI
networks from project funding with limited time horizons (Subsection 5.3.3) might become
experienced partners.
Ideally, a jointly installed and maintained measurement network should follow the WMO
Observing Systems Capability Analysis and Review (OSCAR) tool (OSCAR: Observing Systems
Capability Analysis and Review Tool, OSCAR/Surface User Manual). OSCAR targets users interested in
both status and planning of global observing systems and the specifications of instruments for
data users:
– OSCAR/Surface and OSCAR/Space: surface‑ and space‑based observing system capabilities
– OSCAR/Requirements: user requirements
– OSCAR/Analysis: compares requirements with the observing system capabilities (Rolling

Review of Requirements (RRR))
This information pool allows both experts and observing system operators to identify
for example spatial gaps in their observing network and thereby supports their planning efforts.
As for observing networks, the use of models for monitoring will require collaboration with
NMHSs and other groups experienced in model application. These include consultants (air
quality, urban climate assessment) and university researchers with expertise in modelling
urban climate.
7.4 MONITORING THE CL‑UHI USING OBSERVATION NETWORKS
In general, monitoring combines observations, analysis and modelling (Section 7.1), and it is the
combination of all three that is needed. Combining these methods makes it possible to detect
unexpected changes in CL‑UHI that would not have been identified by one method alone.
This section summarizes the aspects of observation networks described in Chapter 5, focusing
on monitoring.
7.4.1 General considerations
Before planning an observational network for monitoring, the user needs must be identified
(Section 7.2). With these identified, the sensor, network design, communication technologies
and power options can be selected (Chapter 5). A choice must be made between a larger
network of lower‑cost sensors, a few high‑quality meteorological stations, the use of existing
measurement sites or a blend of these three, and this choice depends on the use case
(Table 2.2). High‑quality instruments do not compensate for poor or inadequate exposure and
network layout. Any operator, whether of a single station or a massive urban meteorological
network, must provide metadata for the sites, instrumentation, station surroundings and
exposure (Section 5.6). The (spatial) representativeness, time‑dependent thermal source area
(Subsection 5.4.3) and network layouts (Section 5.3) need to be considered. Examples are given
in Appendix 2.
7.4.2 Monitoring with reference stations in standard instrument shelters
Reference WMO meteorological stations in cities provide high‑quality measurements of air

temperature with the same instrumentation as rural WMO stations (see Guide to Instruments
Chapter 1). However, few reference WMO meteorological stations exist in cities. They are sited
at airports or in larger green spaces, and there is rarely more than one in any city. Reference
meteorological stations in cities to monitor the CL‑UHI intensity must be complemented
with reference meteorological stations in the rural surroundings (see Section 5.2 and Guide to
Instruments and Methods of Observation (WMO-No. 8), Volume I: Measurement of Meteorological
Variables, Chapters 1 and 2). With this approach, any city can be monitored but with extremely
limited areal coverage; the intra‑urban variability of the CL‑UHI is not addressed (Section 5.3). If
the urban site is not in a built area (or an urban LCZ, Figure 4.1) the CL‑UHI will be biased.
A standard meteorological station approach, that is an urban and a rural site pair, should
guarantee, if operated following WMO guidelines (Guide to Instruments and Methods of
Observation (WMO-No. 8), Volume I: Measurement of Meteorological Variables, Chapter 1),
that a long‑term record can be used to identify how weather and climate in an urban area
and the rural surroundings are different. In selecting a representative rural reference station
to complement the urban site, it is important to ensure that both sites experience the same
mesoscale climate and have the same topographic setting, while ensuring the rural site is free of
urban influences currently and for the next decades (Section 5.4).
7.4.3 Monitoring with small automatic weather stations in small

instrument shelters
Installing small autonomous thermometers or multi‑variable weather sensors on street poles or

other street furniture is a logistically feasible, cost‑effective option to monitor the intra‑urban
variability of CL‑T with many sites (Subsection 5.4.5). Such networks can provide a detailed
distribution of CL‑T in different LCZs across a city and can be used to map the CL‑UHI across
neighbourhoods. Influence from a variety of features such as green space, cold‑air drainage
and hot spots within cities may be identified. The networks need multiple rural sites (LCZs A–G,
Figure 4.1) with the same instrumentation/site configurations, located in different wind
directions (upwind) and distances from the city to avoid downwind effects of the urban plume
(Lowry, 1977) (Subsection 5.4.1).
Using a larger number of sites requires a much larger effort for communication, maintenance
and quality control. Geostatistical or machine learning techniques can help identify poor data
and guide the placement of new sensors within existing networks or identify redundant sites.
Partnerships with municipal agencies such as transportation and public works departments can
facilitate siting and maintenance of sensors.
(CL-UHI)
7.4.4 Mobile monitoring approaches
For long‑term monitoring, mobile measurements are challenging, with a few promising
exceptions. Sensors (including GPS) mounted on above‑ground public transport can provide
continuous data from transects through cities, covering large areas across urban–rural
transects, though they need to be interpreted with care (Subsection 5.3.2). However, sources
of contamination of measured CL‑T, such as vehicle engine exhaust, need to be considered
(Subsection 5.5.3.2).
7.4.5 Opportunistic sensing – crowdsourcing
As opportunistic sensing (crowdsourcing) is relatively new, the climate record is short and
the technologies are still evolving, with implications for long‑term monitoring campaigns.
Opportunistic sensing (Subsection 5.3.4) can provide data where no permanent measurements
exist. Networks of opportunistic sensors could potentially complement well‑maintained urban
meteorological networks. The largest barriers are associated with being able to undertake
rigorous quality control/data assurance (Chapman et al., 2017) linked to lack of metadata (Fenner
et al., 2021) and sensor drift from unregulated sensor maintenance and calibration regimes.
Rigorous quality control can enhance the data quality. However, deploying well‑documented
sensors and gathering the required metadata after careful site selection remains the
best practice.
7.5 MONITORING THE CL‑UHI AS PART OF AN IUS
CL‑UHI monitoring and information provision is part of an IUS. The CL‑UHI maps produced
from measurements or model results, or their combination, can be used for warnings, in the
planning process or in other use cases for calculating different CL‑UHI metrics (Table 2.2).
However, the CL‑UHI monitoring information needs to be transformed into other measures,
for instance risk maps, to provide services. These services require additional information, for
instance on transportation flows or socioeconomic characteristics (Table 4.3), to support
the IUS concept (Guidance on Integrated Urban Hydrometeorological, Climate and Environmental
Services (WMO-No. 1234), Volume I: Concept and Methodology and Guidance on Integrated
Urban Hydrometeorological, Climate and Environment Services (WMO-No. 1234), Volume II:
Demonstration Cities).
7.5.1 Monitoring CL‑UHI from the past to today
Examples of observational networks are provided in Appendix 2, and examples of monitoring

approaches are provided in Appendix 3. Observed values combined with urban fraction and
altitude data (Appendix 3, Example 1), or CL‑UHI intensities based on observations combined
with plant diversity data (Appendix 3, Example 2), may both provide an enhanced spatial
pattern of CL‑UHI.
Instruments/models are continuously being further developed, and the network/model settings
can change; all this should be documented in the corresponding metadata (Sections 5.6, 6.5).
A changed model needs to be evaluated again for the CL‑UHI metric used (Section 6.3).
For monitoring the CL‑UHI with the use of observations or/and models, past changes in the
observational data availability as well as the thermal source area of the observations and changes
in urban form need to be considered. This is also true for changes in models if the past CL‑UHI
information is stored in a database and used to derive time series or changes in the CL‑UHI.
Care has to be taken in deriving temporal developments from data that might have different
representativeness (observations) or are now influenced by “new” processes for instance via
new model versions. Reanalysing the past CL‑UHI with a frozen model version may solve this
problem, provided that the data on the urban form are available for the past. Nevertheless, the
configuration of models for monitoring requires a compromise between data requirements for
characterization of the urban area (Table 6.1), the intended data use (Table 2.2), computational
cost and model performance (Chapter 6). The model used should be evaluated for the CL‑UHI
metric that will be used for monitoring (Section 6.3).
7.5.2 Monitoring CL‑UHI into the future
Future changes may include the urban area itself – for instance, linked to infrastructure,
technology, people’s behaviour and vegetation – as well as climate change. Information about
future CL‑UHI intensities is likely to be needed for many applications (Chapter 3, Chapter 8).
The combined influences will be observed by sensors in these future times. However, by using
modelling likely future CL‑UHI intensities can be assessed, with the modelling allowing the
different origins of effects to be separated. This may help to identify if changes in CL‑UHI found
for the future are to be attributed to changes in the urban features (Table 4.1) or are a result of
climate change.
If bias corrections are used for the model output, these corrections need to be updated whenever
the model or the thermal source area for observations assimilated in the modelling change.
Therefore, the thermal source area of an observation needs to be assessed from time to time to
determine if the thermal source has changed.
7.5.3 Data transfer, archiving and licensing
For combining measurements and modelling for monitoring purposes, automated downloads
are critical. Specialized meteoalarm (for example, https://www.meteoalarm.org/en/), early
warning and emergency management systems require automated transfer protocols.
The principles of collaboration (Section 7.3) and standardization of provided data are central
to long‑term monitoring. The distribution of CL‑UHI monitoring data, from observations and
modelling, requires data sharing and corresponding licensing options to be established. These
range between open‑data and proprietary (private/commercial) approaches. For scientific and
public information considerations, network operators and modellers are encouraged to publish
CL‑UHI data as open data fulfilling the FAIR principles (findable, accessible, inter‑operable,
reusable; Section 6.5). Metadata are essential for successful use and interpretation of both
observations (Section 5.6) and model results (Section 6.5).
Standardization between operators is challenging but essential as data originates from

sources with different reliability and accuracy. Common databases, platforms and protocols,
if implemented, enable full utilization of all the monitoring data. Data sharing portals may be
operated by NMHSs, agencies or institutions, or by leveraging the WMO Information System
(WIS) (Guide to the WMO Information System (WMO-No. 1061); Manual on the WMO Information
System (WMO-No. 1060). The data are to be stored in a long‑term, actively maintained
data archive.
7.6 MONITORING CHALLENGES IN BRIEF
Monitoring of CL‑UHI intensities based on long‑term measurements, modelling and analyses is

not without challenges:
– All of the challenges outlined for measurements (Section 5.8) and modelling
(Section 6.6) apply.
– Methods to integrate and homogenize CL‑UHI monitoring (measurements, modelling) data

and, in turn, to integrate these between NMHS meteorological services and multi‑agency
IUSs are important.
(CL-UHI)
– Long‑term availability of measurement sites is still rather limited.
– Determination of site characteristics and metadata for historic CL‑UHI measurement data
is demanding.
– Computing resources and expertise are required for microscale and local‑scale
model simulations that assess past (reanalyse), current (analyse) and future (forecast,
scenarios) CL‑UHI.
CHAPTER 8. UNDERSTANDING THE IMPACTS OF CL‑UHI MITIGATION
AND ADAPTATION EFFORTS
A range of adaptation and mitigation strategies applied at the microscale to the mesoscale
could be used to influence CL‑UHI. Examples are modification of urban materials (Section 8.1)
or providing shading and vegetation and water infrastructure (Section 8.2). Through targeted
urban design and planning, the characteristics of the CL‑UHI can be modified. The built form,
land cover and materials (Table 4.1, 4.2) influence the energy exchange processes at many scales
and hence the CL‑UHI. For mitigating the CL‑UHI several approaches are used, including the
examples in the following sections.
8.1 CHANGE OF MATERIALS
At the microscale, materials can be changed through retrofitting or new construction. Careful
selection can mitigate the CL‑UHI. Possible materials include the following:
– Cool materials with high reflectivity values that can reduce absorption of shortwave
radiation and/or the near‑infrared band;
– Thermochromic materials with reflective properties that increase at higher temperatures;
– Insulation and glazing materials reducing anthropogenic heat emissions;
– Permeable paving materials modifying water drainage and increasing soil moisture
and thus enhancing water availability for evapotranspiration, which in turn reduces
CL‑UHI intensity.
To determine probable quantitative impacts of choices on CL‑UHI intensity, modelling studies

are needed, as the near‑surface air temperatures also depend on the meteorological conditions
(Section 2.5). At the microscale, detailed studies are possible with ORMs (Subsection 6.1.2).
For assessing influences of the whole urban area being changed, scenario studies of coarser
resolution may be advisable (Subsection 6.4.4).
8.2 NATURE‑BASED SOLUTIONS – INCLUDING GREEN AND BLUE

INFRASTRUCTURES
With widespread implementation, the microscale changes to mitigate the CL‑UHI (Section 8.1)
may combine to influence the local scale. Nature‑based solutions (including more vegetation
and water features in urban areas (Miles et al., 2021)) are another way to mitigate the CL‑UHI,
with more than a thousand examples around the world (https://una.city/). Green roofs, urban
gardens and water bodies (for instance urban aquaculture) can be effective ways to lower urban
temperatures in some climates. Measures include the following:
– Intensive green roofs can reduce roof temperature (increased evapotranspiration, and
possibly higher albedo than roof shingles), slightly reduce the near‑ground temperatures
and thereby mitigate the CL‑UHI. Green roofs have to have sufficient water supply
(see Table 2 in Norton et al., 2015).
– Urban parks reduce temperature during daytime and thereby daytime CL‑UHI, with
a tendency for larger parks to cool more. How far into the built‑up area the cooling effect
spreads depends on wind speed, building density and the vegetation in the park.
(CL-UHI)
– Water bodies such as lakes and ponds reduce daytime temperature, especially
downwind of the water body, but due to the high heat capacity of water, the night‑time
temperatures and thus night‑time CL‑UHIs are often higher than in the built‑up urban areas
(Schlünzen et al., 2010).
The influences of the nature‑based solutions on the CL‑UHI have to be carefully assessed as
increases in humidity can cause negative influences on heat stress (Section 3.1).
8.3 ADJUSTING URBAN FORM
At the microscale to the local scale, designers and planners can change and regulate urban
structures (Sections 8.1, 8.2). These measures may modify the sky view factor (microscale) and
ventilation pathways (local scale). Given that heat stress is associated with multiple indices
(Subsection 3.1.1), its mitigation may involve modifying other variables beyond air temperature,
such as increasing wind speed and/or decreasing humidity.
Modelling studies can help to determine the consequences of cumulative proposed microscale
and local‑scale changes (for example, whether there are critical thresholds). Examination of
proposed measures (for example increased ventilation in town centres) can identify problems,
investigate the consequences of such measures for the CL‑UHI from timescales of weather cycles
to future climates, and explore potential solutions.
8.4 LIMIT ENERGY USE
At the local scale to the mesoscale, energy use, design and people’s behaviours (Capel‑Timms
et al., 2020) influence the heat emissions into the UCL. A sizeable proportion of building energy
used is linked to HVAC systems. For example, in the European Union, space heating/cooling
amounts to 68%/6% of the total final energy use in households (data for 2012; https://
energy.ec.europa.eu/document/download/59766187-f07a- 40e1- 8097- d5dd09f5341d_en).
For some Mediterranean countries, the space heating contribution is considerably lower
(for example <40% for Malta and Portugal), while for some north European countries it is higher
(for example >90% for Denmark). The energy demand for heating and cooling is closely linked
to the current and future climate of the region and, in urban areas, the additional influence
of the CL‑UHI. It is also linked to the building materials used (Section 8.1), wind speed/cloud
cover (Section 2.5), the indoor temperature desired by residents and anthropogenic heat
emission (Section 3.3). Reducing anthropogenic heat emissions helps to reduce CL‑UHI values.
If the summer CL‑UHI values are reduced and the winter values kept similar, the annual energy
demand may also be reduced. This may be achieved by simple microscale measures such as
the following:
– Intensified microscale urban greening (Norton et al., 2015) with vertical greening
systems. These increase shading and evapotranspiration if sufficiently supplied with water
and slightly reduce the CL‑UHI if they are installed on the sunny façade and begin near
the ground.
– Window shading with a fixed angle (shade in summer, sunshine in winter) or outdoor
blinds/shutters.
– Placing exhaust units on rooftops to avoid anthropogenic heat emission into the
urban canyon.
– Deciduous vegetation that shades façades in summer but not in winter.
For these, and other examples, the quantitative impact depends on the actual situation and how
widely they are employed. They are best assessed in advance by dedicated modelling studies to
increase the likelihood of avoiding unexpected negative impacts.
CHAPTER 8. UNDERSTANDING THE IMPACTS OF CL-UHI MITIGATION AND ADAPTATION EFFORTS 59
8.5 DESIGN AND PLANNING CONSIDERATIONS
Urban planning is a joint effort of all groups involved in urban services and demands expertise in
meteorological influences if the CL‑UHI is to be kept as it is or even mitigated. The following may
be considered:
– The impacts of the CL‑UHI are not always negative; for instance, the CL‑UHI may result in
reduced heating demands in winter and more comfortable outdoor temperatures in cold
climates. Therefore, the heat‑induced health effects may need to be assessed for a full year.
– For mitigating the CL‑UHI, intended urban design and changes in the microscale to
regional‑scale climate have to be considered in the planning phase. To achieve this,
there is a critical need for urban planners, architects, meteorologists, hydrologists and
city administrators (see Recommendations section in Guidance on Integrated Urban
Hydrometeorological, Climate and Environmental Services (WMO-No. 1234), Volume I: Concept
and Methodology) to work together on scenarios.
– The intra‑urban variability in urban form may induce strong spatial variability in the CL‑UHI
across an urban area. Awareness of these intra‑urban differences is important. In rapidly
developing cities, with new neighbourhoods being built and existing neighbourhoods
being rebuilt, it is important to “consider the climate at the early stages of urban
development” (Oke et al., 2017). Consequently, a single CL‑UHI value is not sufficient for
developing urban heat mitigation strategies.
8.6 CHALLENGES FOR FINDING THE RIGHT MEASURES TO MITIGATE CL‑UHI
It should be noted that there is no best design that has positive impacts on all urban services
(Chapter 3). Typically, trade‑offs are needed. The CL‑UHI mitigation strategies may have
secondary negative impacts that have to be assessed in the context of all urban services.
Examples include:
– A reduction of CL‑UHI intensity can have both positive and negative impacts on
concentrations of primary and secondary pollutants. A reduction of the daytime
temperatures reduces the potential for ozone formation. In addition, urban greening can
reduce ozone formation potential through shade, reducing radiation. However, a reduced
night‑time CL‑UHI might lower the UBL height, increasing pollutant concentrations
(Section 3.2).
– CL‑UHI reduction measures are likely to have different effects on air quality in cities of
different sizes and urban forms, in different locations and with different emission and
climate conditions. Therefore, it is recommended that urban design considerations
should not only include impacts on CL‑UHI using the corresponding metrics for extremes
(Table 2.2, part (a)) or seasonal and annual values (Table 2.2, part (b)) but also other
impacts, including those associated with air quality and energy use. Taking an IUS
view allows for unintended consequences to be explored at different scales and from
a multi‑agency perspective (Grimmond et al., 2020; Guidance on Integrated Urban
Hydrometeorological, Climate and Environmental Services (WMO-No. 1234), Volume I: Concept
and Methodology and Guidance on Integrated Urban Hydrometeorological, Climate and
Environment Services (WMO-No. 1234), Volume II: Demonstration Cities).
– Increasing pavement and wall albedo may decrease CL‑T and thus CL‑UHI, but it increases
mean radiant temperature as well as the amount of energy reflected to the buildings
with potential net adverse effects for outdoor thermal comfort and building energy
consumption. In addition, high albedo “smart” surfaces such as cool streets, roofs and walls
(CL-UHI)
can reduce energy demand for cooling but might increase ozone formation (Section 3.2).
Improved insulation can reduce demand for both cooling and heating and also reduces
anthropogenic heat emissions, which also reduces CL‑UHI.
– Roof albedo changes that are beneficial in summer may be detrimental in winter if they
increase heating demand. If the goal of an intervention is to save energy, the net annual
balance should be considered. The intervention should be compared to other practices
such as efficient roof insulation or other measures which may have a smaller impact on
CL‑UHI but result in greater savings relative to investment.
– Increasing evaporation may reduce CL‑UHI but has little impact on thermal comfort and
building energy loads, as both depend on the enthalpy of the air (which includes humidity
effects) and not just temperature.
– The beneficial impacts of vegetation on CL‑UHI by evapotranspiration depend on water

availability. In many climates, increasingly severe water shortages may limit water for
urban irrigation. Thus, even if vegetation may be the most effective temperature reduction
measure, it may not be possible to implement it because of water scarcity.
– Street trees reduce CL‑UHI and daytime mean radiant temperature by providing shade, but
may reduce ventilation in small urban canyons (decreasing comfort) and in consequence
reduce dispersion (increasing pollutant concentrations and exposure). Additionally, trees
emit a range of volatile organic compounds, many of which are precursors for ozone
formation. Increased ozone concentrations impact the health of people (respiratory
and cardiovascular diseases). However, vegetation takes up CO2, and other pollutants
are deposited onto the vegetation, which reduces concentrations of primary pollutants.
Trade‑offs need to be found.
– Reducing the S‑UHI does not automatically lead to a reduction in CL‑UHI, or a reduction in
heat stress and energy consumption (Stewart et al., 2021).
– There are co‑benefits associated with measures lowering the urban area‑induced heat
risk: increased access to green and blue spaces can benefit mental health and well‑being,
reduce heat and air pollution and help reduce atmospheric CO2 concentrations. However,
more shading and trees might reduce the ventilation in the canopy layer and the stagnant
air might lead to increased air pollution and reduced thermal comfort. In some areas, the
additional green and blue spaces may increase relative humidity to levels that might further
reduce well‑being as assessed by thermal indices (Subsection 3.1.2). In addition, more
vegetation increases the pollen load in urban areas which is an additional threat for allergy
sufferers (combined with a longer pollen season in urban areas; Section 3.4).
– Increasing the streets’ ventilation potential by a better street channelling can reduce both
heat stress and air pollution in the UCL. However, a reduced CL‑UHI intensity might reduce
turbulent mixing, thus decreasing the UBL, resulting in higher near‑surface concentrations
of primary pollutants (CO, NOx). Thus, both effects need to be carefully assessed.
– The measure taken, such as a new green area, must be clearly defined before implementing
it, to assess its impacts on and interactions with the CL‑UHI and other features, for instance
air quality. Many of the feedbacks within the urban area are dynamic and therefore need
to be described within the models used for assessment (Chapter 6). Their influences may
be largest when natural radiative fluxes are smallest. For example, wintertime CL‑UHI
may be influenced by anthropogenic heat emission from both traffic and building energy
management, which in turn impacts atmospheric composition (Section 3.2) and health
(Section 3.1). Once the measure has been implemented it is necessary to ensure that the
changes that are occurring in the urban fabric are incorporated into the databases that
describe the city (Chapter 4).
The presence of a city, with its buildings, streets, parks and other urban structures, modifies
the local climate making it different from the surrounding rural climate. While daytime CL‑UHI
intensities are low, the values at night‑time are higher. Through effective management of the
CHAPTER 8. UNDERSTANDING THE IMPACTS OF CL-UHI MITIGATION AND ADAPTATION EFFORTS 61
urban form, it is possible to create microclimates that are more comfortable than those of the
rural surroundings, especially during daytime, as the presence of daytime urban cool islands
indicates (Kuttler et al., 2015; Martilli et al., 2020). Night‑time reductions remain difficult for
two principal reasons: (a) with heat storage in dense urban areas being larger than in the rural
surroundings, the heat stored during the daytime is released during the night; and (b) wind
speed is reduced by the flow resistance of the 3D urban structures, with the result that the
surface energy budget has reduced heat fluxes due to lower wind speeds. Despite these
difficulties, well‑designed urban development can keep CL‑UHI within tolerable limits.
APPENDIX 1. A CASE STUDY OF INFLUENCES ON CL‑UHI AND OTHER
TEMPERATURES
The CL‑UHI is the canopy layer air temperature difference between urban and surrounding rural
areas (Chapter 2). It is important to understand the types of UHI (Section 2.6) and atmospheric
temperature‑related features when talking to stakeholders and end users with different
application needs (Chapter 3). This Appendix can help to ensure that the same feature is being
discussed. For illustration purposes, model simulation results are used.
Set‑up of sensitivity simulations
To illustrate influences contributing to the CL‑UHI, a set of idealized simulations with the
mesoscale Weather Research and Forecasting (WRF) model coupled to the multi‑layer urban
canopy parametrization Building Effect Parameterization – Building Energy Model (BEP–BEM)
(Martilli et al., 2002) are presented (Table A1.1). In all cases, the urban area has the same
characteristics (compact low rise, LCZ 3; Figure 4.1), size (10 × 10 km2), and surroundings (flat
terrain) with case‑specific homogeneous rural land use. The simulations are performed for
21 June (no particular year, northern hemisphere summer solstice) with a geostrophic wind
of 6 m s–1 from the West. The initial specific humidity is the same in all simulations (0.005 kg kg–1),
but the soil moisture, land use in the rural area and latitude differ. Soil moisture is the same
in both urban and rural areas but varies from 0.1 m3 m–3 (dry case) to 0.4 m3 m–3 (wet case).
The urban fraction (building and paved) is assumed to be 90% in the urban area. Simulations
start for 18:00 local solar time (LST). The assessments provided here are for a full day, from
00:00 to 24:00 LST. The CL‑UHI is computed as the difference between the value at the centre
of the city and the value 15 km upwind of the city’s upwind edge, which is considered to be far
enough from the city to reduce the upwind influence of the city.
The simulation results are discussed in the sections that follow. Note that these results are
selected to illustrate the influence factors (first two sections below) and illustrate differences
between CL‑UHI intensity and other temperature‑related phenomena that cannot be described
by the CL‑UHI (remaining sections).
Soil humidity matters
The differences in soil moisture cause differences in the latent heat fluxes between the dry and
wet cases (Table A1.1). In the dry cases, the atmospheric humidity stays almost constant at
0.005 kg kg–1, while for the wet case humidity increased to 0.012–0.014 kg kg–1 in the atmosphere
and the soil moisture decreased due to evapotranspiration by the end of the simulation
(not shown). In the dry_mid_lat case the urban area has a higher maximum CL‑T than in
the wet_mid_lat case (Figure A1.1a), but lower daytime CL‑UHI values (Figure A1.1b).
Table A1.1. Characteristics of idealized mesoscale model case study simulations

with WRF/BEP–BEM
2 m air Soil
Case Land use in rural area Latitude Climate temperature moisture
(°C) (m3 m –3)
dry_mid_lat closed‑canopy shrubland (LCZ C) 45° N Dry 27 0.1
wet_mid_lat deciduous broadleaf forest (LCZ A) 45° N Wet 27 0.4
dry_low_lat closed‑canopy shrubland (LCZ C) 25° N Dry 32 0.1
dry_high_lat closed‑canopy shrubland (LCZ C) 55° N Dry 22 0.1
Note: Temperature and soil moisture values are initial values for the whole domain.
Source: Table provided by Alberto Martilli.
APPENDIX 1. A CASE STUDY OF INFLUENCES ON CL-UHI AND OTHER TEMPERATURES 63
35
(a) dry_mid_lat
30
CL–T (°C)
25
wet_mid_lat
20
5.0 (b) dry_mid_lat
2.5 wet_mid_lat
CL–UHI (°C)
–2.5
10 (c)
wet_mid_lat
5
S–UHI (°C)
–5
dry_mid_lat
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (h)
Figure A1.1. Results of idealized simulations with WRF/BEP–BEM with values for:
(a) urban near-surface air temperature (CL‑T) at the urban centre,
(b) the CL‑UHI as the difference of temperatures upwind of the urban area and at the urban
centre, and (c) S‑UHI for the same places as (b).
Source: Figure provided by Alberto Martilli.
In the dry_mid_lat case, CL‑UHI may be negative during the day, indicating an urban cool
island (Figure A1.1b). For this case CL‑UHI intensity is at its maximum at night (for reasons see
Subsection 2.5.2) and is higher than for the wet_mid_lat case, which has similar values at day and
night (Figure A1.1b).
Latitude matters
The latitude (altitude is not investigated here) determines the amount of solar radiation
received and angles of incidence of direct sunlight, and thus the available energy for the surface
energy budget. For latitudes 25 °N (dry_low_lat case), 45 °N (dry_mid_lat case) and 55 °N
(dry_high_lat case), latitude‑representative air temperatures were chosen to initialize the
simulations (Table A1.1).
In the simulations, CL‑T for dry_high_lat (55° N) is in a range between 14 °C and 27 °C,
for dry_mid_lat (45° N) it is in the range between 18 °C and 32 °C, while for the dry_low_lat
(25° N) case CL‑T reaches almost 40 °C (Figure A1.2a). However, the CL‑UHI values are very
(CL-UHI)
40
(a)
dry_low_lat
dry_mid_lat
30
CL–T (°C)
dry_high_lat
20
5.0 (b)
dry_low_lat
CL–UHI (°C)
2.5 dry_mid_lat
0 dry_high_lat
–2.5
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (h)
Figure A1.2. Modelled (a) CL‑T and (b) CL‑UHI for three cases: dry_low_lat (25° N),
dry_mid_lat (45° N) and dry_high_lat (55° N) (Table A1.1).
Source: Figure provided by Alberto Martilli.
similar for all three cases (Figure A1.2b), especially during the day. Therefore, while cities may
have similarly large CL‑UHI intensities, only those at lower latitudes, which already experience a
warmer climate, might have large urban heat problems.
S‑UHIs versus CL‑UHIs
Depending on humidity, S‑UHIs and CL‑UHIs peak at different times (Figure A1.1b, c). The
two quantities have different magnitudes and different controls (Section 2.6). Note that the
modelled urban surface temperature used to calculate the S‑UHI (Figure A1.1c) is derived from
the upward longwave radiation from roof and roads only, not from the total upwelling longwave
radiation, and therefore is most comparable to a satellite nadir (which would see the plan area of
vegetation) view (a non‑nadir view sees walls, vegetation, roof shading). The differences in the
results for the CL‑UHI and S‑UHI show that it is not possible to design or evaluate an urban heat
mitigation strategy (Chapter 8) aiming to improve (decrease or increase) the CL‑UHI based on
the S‑UHI alone (Martilli et al., 2020).
Large CL‑UHI is not necessarily related to high urban temperatures
A large CL‑UHI intensity (see dry cases in Figure A1.1b, A1.2b) does not necessarily indicate
high temperatures in the urban area in the absolute sense, and therefore it cannot be used to
assess the existence of urban overheating (for instance CL‑T >30 °C). The CL‑UHI is part of the
temperature signal measured within the urban area and depends on several factors, including
the regional climate of the urban area and its surroundings (Chapter 2). Furthermore, the timing
of maximum values for the CL‑UHI and CL‑T differs: in general, daily CL‑UHI intensity is at its
maximum at night (Chapter 2, Figure A1.1b, A1.2b), whereas the maximum of CL‑T is during the
APPENDIX 1. A CASE STUDY OF INFLUENCES ON CL-UHI AND OTHER TEMPERATURES 65
day (Figure A1.1a, A1.2a). This is relevant because the time of maximum CL‑T has implications for
heat exposure in urban areas (Section 3.1) which cannot be addressed exclusively by reducing
the maximum CL‑UHI (for metrics see Table 2.2; for mitigation see Chapter 8).
The CL‑UHI does not create heatwaves
A heatwave is driven by regional synoptic conditions that influence both urban and rural areas,
and is therefore not caused by the presence of an urban area. The heatwave and CL‑UHI will,
however, interact, modifying local weather. The spatial scale of the heat wave is larger than
the urban area itself. A heat wave lasts several days, while the CL‑UHI has a diurnal pattern
in addition to annual changes. Because synoptic conditions associated with heatwaves such
as high pressure, stagnation and low wind speeds are often favourable for CL‑UHI formation
(Subsection 2.5.3), large CL‑UHI values can often be observed during heat waves. Thus,
consecutive “hot nights” are more likely to occur in urban than rural areas.
The CL‑UHI is not global warming
Due to the CL‑UHI, urban areas are warmer than rural ones, especially at night, and thus climate
change will lead to higher values in urban areas than in rural areas. However, the CL‑UHI should
not be mistaken for an indication of global warming or climate change. Greenhouse‑gas induced
global warming occurs at different spatial and temporal scales and has different causes than
the CL‑UHI. The CL‑UHI contributes only slightly to global warming; however, when using
measurements to assess global warming it needs to be considered that the data taken in an
urban area might be influenced by the urban area (Subsection 5.4.3).
There can be feedbacks between global warming and the CL‑UHI. For example, a warmer
climate might lead to more heatwaves, increasing CL‑UHI (see previous subsection). Also,
excessive energy release, for instance from space cooling, increases anthropogenic heat flux
that in turn increases CL‑UHI intensities (de Munck et al., 2013) (surface influences explained in
Subsection 2.5.1). There are indications that CL‑UHI intensities may change with climate change:
no changes to slight increases or decreases are found for cities around the globe in model
studies (Oleson et al., 2011; Katzfey et al., 2020). To assess this effect for a specific urban area,
changes in the synoptic situation, as well as changes in the urban characteristics (Chapter 4)
must be considered for the scenario assessment. Modifications to urban design (Chapter 8)
might change the CL‑UHI but might not influence greenhouse gas emissions which contribute to
global warming.
CL‑UHI intensity is not simply an assessment of heat added by the urban area
As explained in Section 2.3, the CL‑UHI is an atmospheric warming effect associated with cities.
However, there are many influences and factors. Furthermore, CL‑UHI intensities are based
on temperature differences and thus need a non‑urban reference, and difficulties can arise in
finding the proper reference values. The ideal way for determining the CL‑UHI would be to
have data from the pre‑urban setting. This might be possible in model studies, but despite the
possibility of performing scenarios with a pre‑urban land cover the question that needs to be
answered is what the pre‑urban land cover might have been (Section 6.2). More pragmatically,
measurements or model results in the rural surroundings of the urban area are taken as
non‑urban reference. However, the temperature measured in a rural area close to an urban area
might be influenced by it (advection, see Section 5.2), but taking the rural values upstream of
the urban area (as done in the example case study in this Appendix) avoids these problems.
Nonetheless, finding the rural reference can be challenging in the Anthropocene when cities
are becoming larger and more numerous, their boundaries are less clear, and the closest rural
locations to the urban area are frequently agricultural, and therefore also modified by humans
(Lowry, 1977; Martilli et al., 2020). Especially in spring and autumn, when the agricultural areas
quickly change their vegetation cover, a difference in CL‑UHI intensities might be more a result of
land cover changes at the rural reference site than at the urban site.
(CL-UHI)
CL‑UHI intensities cannot be reduced to zero
The presence of an urban area with its different urban structures modifies the local climate
(Section 2.7), making it different from the surrounding rural climate, which may also be modified
by human activities. While daytime CL‑UHI intensities are low, the values at night‑time are
higher. Effective management of the urban form might create microclimates more comfortable
than in the surrounding rural areas, especially during daytime, while night‑time reductions
remain challenging (Section 8.6). The heat storage in dense urban areas might still be larger than
in the rural surroundings, with the result that daytime stored heat is released to the air during
the night. Moreover, wind speeds reduced by the drag from buildings impact the near‑surface
heat fluxes.
APPENDIX 2. EXAMPLES OF OBSERVATION NETWORKS
Four network approaches with examples are shown here. Germany’s Deutscher Wetterdienst
(DWD) uses urban–rural pairs (Figure A2a) in seven cities to monitor long‑term urban effects on
weather and climate. Each urban station is sited in one dominant LCZ and is paired with a more
rural site (often the city’s airport).
(a) Weather station pair (Hanover) (b) LCZ-based sensor network (Bern)
Official station Sensors

5 km 2 km
Official station
(c) Grid-based sensor network (Zurich) (d) Citizen weather stations (Berlin)
Sensors Citizen
on street lights 2 km 10 km
weather station
Built-up areas Agriculture Forests Water bodies Airports
Figure A2. Examples of different observation networks to monitor UCL temperatures used to
determine CL-UHI: (a) standard WMO site pair in Hannover, Germany (operated by DWD);
(b) sensor network designed based on LCZ classification in Bern, Switzerland
(Gubler et al., 2021); (c) grid-based network layout of small automatic weather stations
in Zurich, Switzerland (meteoblue AG); and (d) more than 3 000 citizen weather stations
(Netatmo) in Berlin, Germany (Fenner et al., 2019).
Source: Background map: © OpenStreetMap contributors; land cover: © European Union, Copernicus Land Monitoring
Service 2018 European Environment Agency (EEA). Figure provided by Andreas Christen.
(CL-UHI)
Strategies to plan urban meteorological networks include: mapping the urban area, such as
using LCZs (Section 4.3), and manually selecting representative sites (e.g. Bern, Switzerland
in Figure A2b) to ensure site‑specific thermal source areas are similar independent of wind
direction (Subsection 5.4.3). Using a regular grid to distribute sensors (e.g. Zurich, Switzerland,
Figure A2c) provides an equal density of sensors across an urban environment that matches
numerical model grids. An irregular but dense network is created by citizen weather stations
(e.g. Berlin, Germany, Figure A2d).
APPENDIX 3. MONITORING EXAMPLES
Example 1. Meteorological measurements
Météo‑France and the city of Toulouse (France) have constructed an IUS based on a network of
approximately 70 stations (Figure A3.1c, d) across the urban to rural area. The network is owned
by the city. The IUS provides real‑time data and maps of the CL‑UHI (Figure A3.1c, d).
The ‘rural reference’ uses observations in areas of low vegetation (individual area size >0.1 km2),
including fields and crops around the urban area. The average temperatures for these areas are
normalized to have a zero mean CL‑UHI intensity. For the S‑UHI (70 m resolution, Figure A3.1a,
b) the same rural areas are used to define the base‑level S‑UHI.
The CL‑UHI and S‑UHI intensities can be compared for one day (Figure A3.1a, c) and another
night (Figure A3.1b, d):
– The midday S‑UHI reaches 15 °C west of central Toulouse at the Airbus factory (extensive
sealed surfaces) and at light industrial and commercial areas in the south. At this time the
river Garonne appears cold (Figure A3.1a).
– The CL‑UHI is more homogeneous, with a lower intensity (Figure A3.1c).
Daytime S-UHI from satellite Night-time S-UHI from satellite

07/08/2020 12H LT 15/07/2019 22H LT
(a) (b)
Daytime CL-UHI from weather stations Night-time CL-UHI from weather stations
07/08/2020 12H LT 15/07/2019 22H LT
(c) (d)
Figure A3.1. Toulouse, France (a, b) S‑UHI (ECOSTRESS satellite, 70 m resolution) and
(c, d) CL‑UHI spatially interpolated network (dots) to 250 m using additional information
on urban fraction and altitude (Touati et al., 2020). (a, c) 12:00 local time, 7 August 2020,
(b, d) 22:00 local time 15 July 2019.
Source: Guillaume Dumas, National Centre for Meteorological Research (CNRM) laboratory & Toulouse Métropole
local authority; with contributions from Aurélie Michel, National Office for Aerospace Studies and Research (ONERA)
laboratory.
(CL-UHI)
– At night the CL‑UHI is well pronounced, but the S‑UHI is more intense than the CL‑UHI
(Figure A3.1b, d). The river, visible by higher temperatures, influences both signals.
As satellite data cannot provide the CL‑UHI information, a network of meteorological stations
(Figure A3.1c, d) was installed. The monitoring network data will be used to develop more IUSs,
including supporting future urban planning.
Example 2. Floristic mapping for deriving climate maps of the annual average CL‑UHI
Bechtel and Schmidt (2011) used floristic mapping of species composition in combination with
in situ measurements (15 years of data) (Schlünzen et al., 2010) to determine the spatial pattern
of the climatological average CL‑UHI. The presence and absence of 1 643 vascular wild plant
species was observed at 1 km2 resolution (by volunteers) in areas of no or very little maintenance.
The species composition data were used to determine Ellenberg indicator values for temperature
(EIT) to derive overall temperature preference of the plant species (at 1 km2) (Bechtel and
Schmidt, 2011). EIT values are correlated with CL‑UHI intensities based on five measurement sites
(Figure A3.2).
UHIEIT (K)
>1.1
1.0 – 1.1
0.9 – 1.0
0.8 – 0.9
0.7 – 0.8
0.6 – 0.7
0.5 – 0.6
0.4 – 0.5
0.3 – 0.4
0.2 – 0.3
0.1 – 0.2
< 0.1
Figure A3.2. Climate average (15 years) heat island intensity for Hamburg, Germany, derived
from mean Ellenberg indicator values for temperature (small circles, 1 km2 grid) compared
with decadal average CL‑UHI values derived from measurements.
Source: Bechtel and Schmidt (2011). © Inter‑Research 2011.
APPENDIX 3. MONITORING EXAMPLES 71
Using the EIT pattern, a CL‑UHI with climate‑averaged intensities (Table 4.3 metric: annual
average CL‑UHI (15 years)) was derived (Figure A3.2). Influences from urban morphology
(city centre at St. Pauli) and water bodies (river Elbe, lake and harbour in the centre and south;
indicated with dashed lines in Figure A3.2) are evident, with larger values in the city centre and
near water bodies. The CL‑UHI pattern derived from floristic mapping is much more detailed
than using the five measurement sites, allowing average simulated CL‑UHI pattern to be verified
(Hoffmann et al., 2018).
This approach uses information on plant community diversity for long‑term integrated
monitoring, and requires knowledgeable plant enthusiasts (for example, community
associations, ecology/biogeography students). In addition, the city must allow wild plants
to grow (this is becoming common again for many ecological reasons). The community
involvement has other benefits beyond providing continuous monitoring of the CL‑UHI
distribution and intensity.
ABBREVIATIONS
2D two‑dimensional
3D three‑dimensional
AC atmospheric chemistry
AGL above ground level
AWS automatic weather station
BEP–BEM Building Effect Parameterization – Building Energy Model
CFL Courant–Friedrichs–Lewy
CL canopy layer
CL‑T canopy layer temperature (near‑surface air temperature)
CL‑UHI canopy layer urban heat island
CORINE Coordination of Information on the Environment (European Union programme)
DNS direct numerical simulation
EASYDAB Earth System Data Branding
GAW Global Atmosphere Watch
GCM global climate model
GIS geographical information system
GM global model
GURME GAW Urban Research Meteorology Environment (WMO project)
HVAC heating, ventilation, and air conditioning
ISL inertial sublayer or constant flux layer
IUS integrated urban services (a term comprising integrated urban

hydrometeorological, climate and environmental services)
LCZ Local Climate Zone (classification scheme for land cover and surface form)
LES large eddy simulation
LST local solar time
MOST Monin–Obukhov similarity theory
NMHS National Meteorological and Hydrological Service
NOAA National Oceanic and Atmospheric Administration
NWP numerical weather prediction

ABBREVIATIONS 73
ORM obstacle‑resolving model
OSCAR Observing Systems Capability Analysis and Review
PBL planetary boundary layer
RANS Reynolds‑averaged Navier–Stokes
RCM regional climate model
RM regional model
RRR Rolling Review of Requirements
RSL roughness sublayer
SL surface layer (UCL + RSL + ISL)
S‑UHI surface urban heat island
SS‑UHI subsurface urban heat island
UBL urban boundary layer
UBL‑UHI urban boundary layer urban heat island
UCL urban canopy layer
UHI urban heat island
UCP urban canopy parameterization
USGS United States Geological Survey
WIGOS WMO Integrated Global Observing System
WIS WMO Information System
WRF Weather Research and Forecasting model
WUDAPT World Urban Database and Access Portal Tool

REFERENCES
Andrioni, M.; Austin, J.; Bailey, K. et al. CF Standard Name Table; CF Metadata Conventions, 2022. https://
cfconventions.org/Data/c f-standard-names/current/build/c f-standard-name-table.html.
Baklanov, A.; Cárdenas, B.; Lee, T. et al. Integrated Urban Services: Experience from Four Cities on Different
Continents. Urban Climate 2020, 32. https://doi.org/10.1016/j.uclim. 2020.100610.
Basu, R. High Ambient Temperature and Mortality: A Review of Epidemiologic Studies from 2001 to 2008.
Environmental Health 2009, 8 (1), 40. https://doi.org/10.1186/1476- 069X- 8- 40.
Bechtel, B.; Schmidt, K. J. Floristic Mapping Data as a Proxy for the Mean Urban Heat Island. Climate
Research 2011, 49 (1), 45–58. https://doi.org/10.3354/cr01009.
Capel‑Timms, I.; Smith, S. T.; Sun, T. et al. GMD – Dynamic Anthropogenic Activities Impacting Heat
Emissions (DASH v1.0): Development and Evaluation. Geoscientific Model Development 2020,
13 (10), 4891–4924. https://doi.org/10. 5194/gmd-13- 4891-2020.
Chang, J.; Qu, Z.; Xu, R. et al. Assessing the Ecosystem Services Provided by Urban Green Spaces along
Urban Center‑Edge Gradients. Sci Rep 2017, 7 (1). https://doi.org/10.1038/s41598- 017-11559-5.
Chapman, L.; Bell, C.; Bell, S. Can the Crowdsourcing Data Paradigm Take Atmospheric Science to a New
Level? A Case Study of the Urban Heat Island of London Quantified Using Netatmo Weather
Stations. International Journal of Climatology 2017, 37 (9), 3597–3605. https://doi.org/10
.1002/joc.4940.
Chapman, L.; Muller, C. L.; Young, D. T. et al. The Birmingham Urban Climate Laboratory: An Open
Meteorological Test Bed and Challenges of the Smart City. Bulletin of the American Meteorological
Society 2015, 96 (9), 1545–1560. https://doi.org/10.1175/BAMS -D -13 - 00193.1.
Ching, J.; Mills, G.; Bechtel, B. et al. WUDAPT: An Urban Weather, Climate, and Environmental Modeling
Infrastructure for the Anthropocene. Bulletin of the American Meteorological Society 2018, 99 (9),
1907–1924. https://doi.org/10.1175/BAMS -D -16 - 0236.1.
Cleugh, H.; Grimmond, S. Urban Climates and Global Climate Change. In The Future of the World’s Climate;
2nd ed. Henderson‑Sellers, A., McGuffie, K., Eds.; Elsevier: Boston, 2012; 47–76. https://doi.org/
10.1016/B978- 0-12-386917-3.00003- 8.
Coast, S. The Book of OSM; CreateSpace Independent Publishing Platform, 2015.
Creutzig, F.; Lohrey, S.; Bai, X. et al. Upscaling Urban Data Science for Global Climate Solutions. Global
Sustainability 2019, 2, 1–25. https://doi.org/10.1017/sus. 2018.16.
D’Amato, G.; Cecchi, L.; Bonini, S. et al. Allergenic Pollen and Pollen Allergy in Europe. Allergy 2007, 62 (9),
976–990. https://doi.org/10.1111/j.1398-9995. 2007.01393.x.
de Munck, C.; Pigeon, G.; Masson, V. et al. How Much Can Air Conditioning Increase Air Temperatures for a
City like Paris, France? International Journal of Climatology 2013, 33 (1), 210–227. https://doi.org/
10.1002/joc.3415.
de Freitas, C. R.; Grigorieva, E. A. A Comparison and Appraisal of a Comprehensive Range of Human
Thermal Climate Indices. Int J Biometeorol 2017, 61, 487–512. https://doi.org/10.1007/s00484
-016-1228- 6.
Demuzere, M.; Kittner, J.; Martilli, A. et al. A Global Map of Local Climate Zones to Support Earth System
Modelling and Urban‑Scale Environmental Science. Earth System Science Data 2022, 14 (8),
3835–3873. https://doi.org/10. 5194/essd-14-3835-2022.
Dennis, R.; Fox, T.; Fuentes, M. et al. A Framework for Evaluating Regional‑Scale Numerical Photochemical
Modeling Systems. Environ Fluid Mech 2010, 10 (4), 471–489. https://doi.org/10.1007/s10652
-009-9163-2.
Doan, Q.‑V.; Dipankar, A.; Simón‑Moral, A. et al. Urban‐induced Modifications to the Diurnal Cycle of
Rainfall over a Tropical City. Quarterly Journal of the Royal Meteorological Society 2021, 147 (735),
1189–1201. https://doi.org/10.1002/qj.3966.
Erell, E.; Leal, V.; Maldonado, E. Measurement of Air Temperature in the Presence of a Large Radiant Flux:
An Assessment of Passively Ventilated Thermometer Screens. Boundary‑Layer Meteorol 2005,
114 (1), 205–231. https://doi.org/10.1007/s10546- 004- 8946- 8.
European Environment Agency (EEA). Urban Atlas LCLU 2018, v013; Copernicus, 2021. https://land
.copernicus.eu/local/urban-atlas/urban-atlas-2018.
Fenner, D.; Bechtel, B.; Demuzere, M. et al. CrowdQC+ – A Quality‑Control for Crowdsourced
Air‑Temperature Observations Enabling World‑Wide Urban Climate Applications. Frontiers in
Environmental Science 2021, 9. https://doi.org/10.3389/fenvs. 2021.720747.
Fenner, D.; Holtmann, A.; Meier, F. et al. Contrasting Changes of Urban Heat Island Intensity during Hot
Weather Episodes. Environ Res Lett 2019, 14 (12). https://doi.org/10.1088/1748-9326/ab506b.
REFERENCES 75
Fischereit, J.; Schlünzen, K. H. Evaluation of Thermal Indices for Their Applicability in Obstacle‑Resolving
Meteorology Models. Int J Biometeorol 2018, 62 (10), 1887–1900. https://doi.org/10.1007/
s00484- 018-1591- 6.
Foken, T., Ed. Springer Handbook of Atmospheric Measurements; Springer Nature Switzerland AG: Cham,
Switzerland, 2021. https://doi.org/10.1007/978-3- 030-52171- 4.
Früh, B.; Becker, P.; Deutschländer, T. et al. Estimation of Climate‑Change Impacts on the Urban Heat Load
Using an Urban Climate Model and Regional Climate Projections. Journal of Applied Meteorology
and Climatology 2011, 50 (1), 167–184. https://doi.org/10.1175/2010JAMC2377.1.
Ganske, A.; Kraft, A.; Kaiser, A. et al. ATMODAT Standard; v3.0, 2021. https://doi.org/10.35095/WDCC/
atmodat_ standard_en_v3_ 0.
Grimmond, S.; Bouchet, V.; Molina, L. T. et al. Integrated Urban Hydrometeorological, Climate and
Environmental Services: Concept, Methodology and Key Messages. Urban Climate 2020, 33.
https://doi.org/10.1016/j.uclim. 2020.100623.
Gubler, M.; Christen, A.; Remund, J. et al. Evaluation and Application of a Low‑Cost Measurement Network
to Study Intra‑Urban Temperature Differences during Summer 2018 in Bern, Switzerland.
Urban Climate 2021, 37. https://doi.org/10.1016/j.uclim. 2021.100817.
Haarsma, R. J.; Roberts, M. J.; Vidale, P. L. et al. High Resolution Model Intercomparison Project
(HighResMIP v1.0) for CMIP6. Geoscientific Model Development 2016, 9 (11), 4185–4208. https://
doi.org/10. 5194/gmd-9- 4185-2016.
Haeger‑Eugensson, M.; Holmer, B. Advection Caused by the Urban Heat Island Circulation as a Regulating
Factor on the Nocturnal Urban Heat Island. International Journal of Climatology 1999, 19 (9),
975–988. https://doi.org/10.1002/(SICI)1097- 0088(199907)19:9<975::AID-JOC399>3.0.CO;2-J.
Hertwig, D.; Grimmond, S.; Kotthaus, S. et al. Variability of Physical Meteorology in Urban Areas at
Different Scales: Implications for Air Quality. Faraday Discuss 2021a, 226, 149–172. https://doi
.org/10.1039/D0FD00098A.
Hertwig, D.; Ng, M.; Grimmond, S. et al. High‑Resolution Global Climate Simulations: Representation
of Cities. International Journal of Climatology 2021b, 41 (5), 3266–3285. https://doi.org/10
.1002/joc.7018.
Hoffmann, P.; Schoetter, R.; Schlünzen, K. H. Statistical‑Dynamical Downscaling of the Urban Heat Island in
Hamburg, Germany. Meteorologische Zeitschrift 2018, 27 (2), 89–109. https://doi.org/10.1127/
metz/2016/0773.
Howard, L. The Climate of London: Deduced from Meteorological Observations, Made at Different Places in the
Neighbourhood of the Metropolis, Vol. 1; W. Phillips, George Yard, Lombard Street: London, 1818.
Hu, X.‑M.; Xue, M.; Klein, P. M. et al. Analysis of Urban Effects in Oklahoma City Using a Dense Surface
Observing Network. Journal of Applied Meteorology and Climatology 2016, 55 (3), 723–741.
https://doi.org/10.1175/JAMC-D-15- 0206.1.
Jacobson, M. Z.; Ten Hoeve, J. E. Effects of Urban Surfaces and White Roofs on Global and Regional Climate.
Journal of Climate 2012, 25 (3), 1028–1044. https://doi.org/10.1175/JCLI-D-11- 00032.1.
Jáuregui, E. The Urban Climate of Mexico City. Erdkunde 1973, 27 (4), 298–307. https://doi.org/10.3112/
erdkunde.1973.04.06.
Katzfey, J.; Schlünzen, H.; Hoffmann, P. et al. How an Urban Parameterization Affects a High‐resolution
Global Climate Simulation. Quarterly Journal of the Royal Meteorological Society 2020, 146 (733),
3808–3829. https://doi.org/10.1002/qj.3874.
Krayenhoff, E. S.; Broadbent, A. M.; Zhao, L. et al. Cooling Hot Cities: A Systematic and Critical Review of
the Numerical Modelling Literature. Environ Res Lett 2021, 16 (5). https://doi.org/10.1088/1748
-9326/abdcf1.
Kuttler, W.; Miethke, A.; Dütemeyer, D. et al., Eds. Das Klima von Essen, Vol. 3; Westarp Wiss:
Hohenwarsleben, 2015. https://bibliographie.ub.uni- due.de/servlets/DozBibEntryServlet?id=
ubo_mods_ 00060091&lang=en.
Lee, D. Urban–Rural Humidity Differences in London. International Journal of Climatology 1991, 11, 577–582.
https://doi.org/10.1002/joc.3370110509.
Lowry, W. P. Empirical Estimation of Urban Effects on Climate: A Problem Analysis. Journal of Applied
Meteorology and Climatology 1977, 16 (2), 129–135. https://doi.org/10.1175/1520
-0450(1977)016<0129:EEOUEO>2.0.CO;2.
Macintyre, H. L.; Heaviside, C.; Cai, X. et al. The Winter Urban Heat Island: Impacts on Cold‑Related
Mortality in a Highly Urbanized European Region for Present and Future Climate. Environment
International 2021, 154. https://doi.org/10.1016/j.envint. 2021.106530.
Maral, S. G.; Mukhopadhyay, T. Signal of Urban Heat Island (UHI) Effect: A Case Study of Mumbai
Metropolitan Region. MAUSAM 2015, 66 (4), 729–740. https://doi.org/10. 54302/
mausam.v66i4. 580.
(CL-UHI)
Martilli, A.; Clappier, A.; Rotach, M. W. An Urban Surface Exchange Parameterisation for Mesoscale Models.
Boundary‑Layer Meteorology 2002, 104, 261–304. https://doi.org/10.1023/A:1016099921195.
Martilli, A.; Krayenhoff, E. S.; Nazarian, N. Is the Urban Heat Island Intensity Relevant for Heat Mitigation
Studies? Urban Climate 2020, 31. https://doi.org/10.1016/j.uclim. 2019.100541.
Masson, V.; Heldens, W.; Bocher, E. et al. City‑Descriptive Input Data for Urban Climate Models: Model
Requirements, Data Sources and Challenges. Urban Climate 2020, 31. https://doi.org/10.1016/j
.uclim. 2019.100536.
Mhedhbi, Z.; Masson, V.; Hidalgo, J. et al. Collection of Refined Architectural Parameters by Crowdsourcing
Using Facebook Social Network: Case of Greater Tunis. Urban Climate 2019, 29. https://doi.org/
10.1016/j.uclim. 2019.100499.
Miles, L.; Agra, R.; Sengupta, S. et al. Nature‑Based Solutions for Climate Change Mitigation; United Nations
Environment Programme (UNEP), International Union for Conservation of Nature (IUCN):
Nairobi, Gland, 2021. http://www.unep.org/resources/report/nature-based-solutions- climate
-change-mitigation.
Muller, C. l.; Chapman, L.; Johnston, S. et al. Crowdsourcing for Climate and Atmospheric Sciences: Current
Status and Future Potential. International Journal of Climatology 2015, 35 (11), 3185–3203.
https://doi.org/10.1002/joc.4210.
Norton, B. A.; Coutts, A. M.; Livesley, S. J. et al. Planning for Cooler Cities: A Framework to Prioritise Green
Infrastructure to Mitigate High Temperatures in Urban Landscapes. Landscape and Urban
Planning 2015, 134, 127–138. https://doi.org/10.1016/j.landurbplan. 2014.10.018.
Oke, T. R. Canyon Geometry and the Nocturnal Urban Heat Island: Comparison of Scale Model and
Field Observations. Journal of Climatology 1981, 1 (3), 237–254. https://doi.org/10.1002/joc
.3370010304.
Oke, T. R. The Energetic Basis of the Urban Heat Island. Quarterly Journal of the Royal Meteorological Society
1982, 108 (455), 1–24. https://doi.org/10.1002/qj.49710845502.
Oke, T. R. The Heat Island of the Urban Boundary Layer: Characteristics, Causes and Effects. In Wind
Climate in Cities; NATO ASI Series, Vol. 277; Cermak, J. E., Davenport, A. G., Plate, E. J. et al. Eds.;
Springer: Dordrecht, Netherlands, 1995; 81–107. https://doi.org/10.1007/978-94- 017-3686-2_ 5.
Oke, T. R. Urban Environment. In The Surface Climates of Canada; Bailey, W. G., Oke, T. R., Rouse, W. R., Eds.;
McGill‑Queen’s University Press: Montreal, 1997; 303–327. https://books.google.fr/books?id=
oxNMhw-rRrQC&source= gbs _ book _other_versions.
Oke, T. R.; Mills, G.; Christen, A. et al. Urban Climates; Cambridge University Press: Cambridge, 2017. https://
doi.org/10.1017/9781139016476.
Oleson, K. W.; Bonan, G. B.; Feddema, J. et al. An Examination of Urban Heat Island Characteristics in a
Global Climate Model. International Journal of Climatology 2011, 31 (12), 1848–1865. https://doi
.org/10.1002/joc. 2201.
Orlanski, I. A Rational Subdivision of Scales for Atmospheric Processes. Bulletin of the American Meteorological
Society 1975, 56 (5), 527–530. http://www.jstor.org/stable/26216020.
Roth, M. Review of Atmospheric Turbulence over Cities. Quarterly Journal of the Royal Meteorological Society
2000, 126 (564), 941–990. https://doi.org/10.1002/qj.49712656409.
Roth, M. Review of Urban Climate Research in (Sub)tropical Regions. International Journal of Climatology
2007, 27 (14), 1859–1873. https://doi.org/10.1002/joc.1591.
Sailor, D. J. A Review of Methods for Estimating Anthropogenic Heat and Moisture Emissions in the Urban
Environment. International Journal of Climatology 2011, 31 (2), 189–199. https://doi.org/10
.1002/joc. 2106.
Schlünzen, K. H.; Hoffmann, P.; Rosenhagen, G. et al. Long-term changes and regional differences in
temperature and precipitation in the metropolitan area of Hamburg. International Journal of
Climatology 2010, 30 (8), 1121–1136. https://doi.org/10.1002/joc.1968.
Schoetter, R.; Hidalgo, J.; Jougla, R. et al. A Statistical–Dynamical Downscaling for the Urban Heat Island
and Building Energy Consumption – Analysis of Its Uncertainties. Journal of Applied Meteorology
and Climatology 2020, 59 (5), 859–883. https://doi.org/10.1175/JAMC-D-19- 0182.1.
Stewart, I. D.; Krayenhoff, E. S.; Voogt, J. A. et al. Time Evolution of the Surface Urban Heat Island. Earth’s
Future 2021, 9 (10). https://doi.org/10.1029/2021EF002178.
Stewart, I. D.; Oke, T. R. Local Climate Zones for Urban Temperature Studies. Bulletin of the American
Meteorological Society 2012, 93 (12), 1879–1900. https://doi.org/10.1175/BAMS-D-11- 00019.1.
Takane, Y.; Kikegawa, Y.; Hara, M. et al. Urban Warming and Future Air-Conditioning Use in an Asian
Megacity: Importance of Positive Feedback. npj Clim Atmos Sci 2019, 2 (1), 1–11. https://doi
.org/10.1038/s41612- 019- 0096-2.
REFERENCES 77
Takane, Y.; Ohashi, Y.; Grimmond, C. S. B. et al. Asian Megacity Heat Stress under Future Climate Scenarios:
Impact of Air-Conditioning Feedback. Environ Res Commun 2020, 2 (1). https://doi.org/10
.1088/2515-7620/ab6933.
Touati, N.; Gardes, T.; Hidalgo, J. A GIS Plugin to Model the Near Surface Air Temperature from Urban
Meteorological Networks. Urban Climate 2020, 34. https://doi.org/10.1016/j.uclim
.2020.100692.
Unkašević, M.; Jovanović, O.; Popović, T. Urban-Suburban/Rural Vapour Pressure and Relative Humidity
Differences at Fixed Hours over the Area of Belgrade City. Theor Appl Climatol 2001, 68 (1),
67–73. https://doi.org/10.1007/s007040170054.
Varentsov, M.; Konstantinov, P.; Baklanov, A. et al. Anthropogenic and Natural Drivers of a Strong Winter
Urban Heat Island in a Typical Arctic City. Atmospheric Chemistry and Physics 2018, 18 (23),
17573–17587. https://doi.org/10. 5194/acp-18-17573-2018.
Vemado, F.; Pereira Filho, A. J. Severe Weather Caused by Heat Island and Sea Breeze Effects in the
Metropolitan Area of São Paulo, Brazil. Advances in Meteorology 2016. https://doi.org/10.1155/
2016/8364134.
von Szombathely, M.; Albrecht, M.; Antanaskovic, D. et al. A Conceptual Modeling Approach to
Health‑Related Urban Well‑Being. Urban Science 2017, 1 (2), 17. https://doi.org/10.3390/
urbansci1020017.
Wanner, H.; Filliger, P. Orographic Influence on Urban Climate. Weather and Climate 1989, 9 (1), 22–28.
https://doi.org/10. 2307/4 4279768.
Warren, E.; Charlton‑Perez, C.; Kotthaus, S. et al. Evaluation of Forward‑Modelled Attenuated Backscatter
Using an Urban Ceilometer Network in London under Clear‑Sky Conditions. Atmospheric
Environment 2018, 191, 532–547. https://doi.org/10.1016/j.atmosenv. 2018.04.045.
Wiesner, S.; Eschenbach, A.; Ament, F. Urban Air Temperature Anomalies and Their Relation to Soil
Moisture Observed in the City of Hamburg. Meteorologische Zeitschrift 2014, 23 (2), 143–157.
https://doi.org/10.1127/0941-2948/2014/0571.
Wilby, R. L. Past and Projected Trends in London’s Urban Heat Island. Weather 2003, 58 (7), 251–260.
https://doi.org/10.1256/wea.183.02.
World Health Organization (WHO). WHO Global Air Quality Guidelines: Particulate Matter (PM2.5 and PM10),
Ozone, Nitrogen Dioxide, Sulfur Dioxide and Carbon Monoxide; WHO, 2021. https://apps.who.int/
iris/handle/10665/3 45329.
World Meteorological Organization (WMO). Guidance on Integrated Urban Hydrometeorological, Climate and
Environmental Services ‑ Volume I: Concept and Methodology (WMO-No. 1234). Geneva, 2019.
World Meteorological Organization (WMO). Guidance on Integrated Urban Hydrometeorological, Climate and
Environment Services ‑ Volume II: Demonstration Cities (WMO‑No. 1234). Geneva, 2021.
World Meteorological Organization (WMO). Guide to Climatological Practices (WMO‑No. 100).
Geneva, 2018.
World Meteorological Organization (WMO). Guide to Instruments and Methods of Observation (WMO‑No. 8),
Volume I: Measurement of Meteorological Variables. Geneva, 2018.
Volume III: Observing Systems. Geneva, 2018.
Volume V: Quality Assurance and Management of Observing Systems. Geneva, 2018.
World Meteorological Organization (WMO). Guide to the WMO Information System (WMO‑No. 1061).
Geneva, 2021.
World Meteorological Organization (WMO). Initial Guidance to Obtain Representative Meteorological
Observations at Urban Sites (WMO/TD-No. 1250). Geneva, 2006.
World Meteorological Organization (WMO). Manual on the WMO Information System: Annex VII to the WMO
Technical Regulations (WMO‑No. 1060). Geneva, 2021.
World Meteorological Organization (WMO). Multi‑hazard Early Warning Systems: A Checklist: Outcome of the
First Multi‑hazard Early Warning Conference; WMO: Geneva, 2018.
World Meteorological Organization (WMO). OSCAR/Surface User Manual; WMO: Geneva, 2022.
World Meteorological Organization (WMO). WIGOS Metadata Standard (WMO‑No. 1192). Geneva, 2019.
World Meteorological Organization (WMO). WMO Guidelines on the Calculation of Climate Normals
(WMO‑No. 1203). Geneva, 2017.
World Meteorological Organization (WMO)/MeteoSwiss/Swiss Confederation. OSCAR: Observing Systems
Capability Analysis and Review Tool; WMO: Geneva, 2015.
(CL-UHI)
World Meteorological Organization (WMO)/World Health Organization (WHO). Urban Climatology

and its Applications with Special Regard to Tropical Areas: Proceedings of the Technical Conference
(WMO‑No. 652). Geneva, 1986.
World Meteorological Organization (WMO)/World Health Organization (WHO). Heatwaves and Health:
Guidance on Warning‑System Development (WMO‑No. 1142). Geneva, 2015.
BIBLIOGRAPHY FOR FURTHER READING
The references included here provide both broad and/or detailed information to supplement the
chapters of this guidance. Other examples could have been chosen.
Section 2.1. Historical background
Chandler, T. J. The Climate of London; Hutchinson & Co: London, 1965.

Mills, G. Luke Howard and The Climate of London. Weather 2008, 63 (6), 153–157. https://doi.org/10
.1002/wea.195.
Peppler, A. Das Auto als Hilfsmittel der meteorologischen Forschung. Zeitschrift für angewandte Meteorologie
1929, 46, 305–308.
Stewart, I. D. Why Should Urban Heat Island Researchers Study History? Urban Climate 2019, 30. https://doi
.org/10.1016/j.uclim. 2019.100484.
von Hann, J. Handbuch Der Klimatologie, 2nd ed.; Vol. 2; Stuttgart: J. Engelhorn, 1897.
Section 2.2. Horizontal scales and vertical layers in urban areas
Feddersen, B. Wind Tunnel Modelling of Turbulence and Dispersion above Tall and Highly Dense Urban
Roughness. Dr. Thesis, Swiss Federal University of Technology, Zurich, 2005. https://doi.org/10
.3929/ethz-a- 004941441.
Grimmond, C. S. B.; Oke, T. R. Aerodynamic Properties of Urban Areas Derived from Analysis of Surface
Form. Journal of Applied Meteorology and Climatology 1999, 38 (9), 1262–1292. https://doi.org/10
.1175/1520 - 0450(1999)038<1262:APOUAD>2.0.CO;2.
Kanda, M.; Inagaki, A.; Miyamoto, T. et al. A New Aerodynamic Parametrization for Real Urban Surfaces.
Boundary‑Layer Meteorol 2013, 148 (2), 357–377. https://doi.org/10.1007/s10546- 013-9818-x.
Kastner‑Klein, P.; Rotach, M. W. Mean Flow and Turbulence Characteristics in an Urban Roughness
Sublayer. Boundary‑Layer Meteorology 2004, 111 (1), 55–84. https://doi.org/10.1023/B:BOUN
.0000010994.32240.b1.
Oke, T. R. The Distinction between Canopy and Boundary‐layer Urban Heat Islands. Atmosphere 1976,
14 (4), 268–277. https://doi.org/10.1080/0 0046973.1976.9648422.
Zilitinkevich, S. S.; Mammarella, I.; Baklanov, A. A. et al. The Effect of Stratification on the Aerodynamic
Roughness Length and Displacement Height. Boundary‑Layer Meteorol 2008, 129 (2), 179–190.
https://doi.org/10.1007/s10546- 008-9307-9.
Section 2.5. Development of the CL‑UHI
Ketterer, C.; Matzarakis, A. Human‑Biometeorological Assessment of the Urban Heat Island in a City with
Complex Topography – The Case of Stuttgart, Germany. Urban Climate 2014, 10, 573–584.
https://doi.org/10.1016/j.uclim. 2014.01.003.
Masson, V.; Lemonsu, A.; Hidalgo, J. et al. Urban Climates and Climate Change. Annual Review of
Environment and Resources 2020, 45 (1), 411–444. https://doi.org/10.1146/annurev- environ
-012320- 083623.
Runnalls, K. E.; Oke, T. R. Dynamics and Controls of the Near‑Surface Heat Island of Vancouver, British
Columbia. Physical Geography 2000, 21 (4), 283–304. https://doi.org/10.1080/02723646
.2000.10642711.
Schluenzen, K. H. Numerical Studies on the Inland Penetration of Sea Breeze Fronts at a Coastline with
Tidally Flooded Mudflats. Contributions to Atmospheric Physics 1990, 63 (3/4), 243–256.
von Glasow, R.; Jickells, T. D.; Baklanov, A. et al. Megacities and Large Urban Agglomerations in the Coastal
Zone: Interactions Between Atmosphere, Land, and Marine Ecosystems. AMBIO 2013, 42 (1),
13–28. https://doi.org/10.1007/s13280- 012- 0343-9.
Wiesner, S.; Bechtel, B.; Fischereit, J. et al. Is It Possible to Distinguish Global and Regional Climate Change
from Urban Land Cover Induced Signals? A Mid‑Latitude City Example. Urban Science 2018,
2 (1), 12. https://doi.org/10.3390/urbansci2010012.
(CL-UHI)
Section 3.1. Heat stress information as a service – CL‑UHI impacts on health
Gasparrini, A.; Guo, Y.; Hashizume, M. et al. Mortality Risk Attributable to High and Low Ambient
Temperature: A Multicountry Observational Study. The Lancet 2015, 386 (9991), 369–375.
https://doi.org/10.1016/S0140- 6736(14)62114- 0.
Gasparrini, A.; Guo, Y.; Sera, F. et al. Projections of Temperature‑Related Excess Mortality under Climate
Change Scenarios. The Lancet Planetary Health 2017, 1 (9), e360–e367. https://doi.org/10.1016/
S2542-5196(17)30156 - 0.
Ho, H. C.; Lau, K. K.‑L.; Ren, C. et al. Characterizing Prolonged Heat Effects on Mortality in a Sub‑Tropical
High‑Density City, Hong Kong. Int J Biometeorol 2017, 61 (11), 1935–1944. https://doi.org/10
.1007/s00484- 017-1383- 4.
Hua, J.; Zhang, X.; Ren, C. et al. Spatiotemporal Assessment of Extreme Heat Risk for High‑Density Cities:
A Case Study of Hong Kong from 2006 to 2016. Sustainable Cities and Society 2021, 64. https://
doi.org/10.1016/j.scs. 2020.102507.
Katavoutas, G.; Founda, D. Response of Urban Heat Stress to Heat Waves in Athens (1960–2017).
Atmosphere 2019, 10 (9). https://doi.org/10.3390/atmos10090483.
McGregor, G. R.; Vanos, J. K. Heat: A Primer for Public Health Researchers. Public Health 2018, 161, 138–146.
https://doi.org/10.1016/j.puhe. 2017.11.005.
McMichael, A. J.; Wilkinson, P.; Kovats, R. S. et al. International Study of Temperature, Heat and Urban
Mortality: The ‘ISOTHURM’ Project. International Journal of Epidemiology 2008, 37 (5),
1121–1131. https://doi.org/10.1093/ije/dyn086.
Wang, D.; Lau, K. K.‑L.; Ren, C. et al. The Impact of Extremely Hot Weather Events on All‑Cause Mortality
in a Highly Urbanized and Densely Populated Subtropical City: A 10‑Year Time‑Series Study
(2006–2015). Science of The Total Environment 2019, 690, 923–931. https://doi.org/10.1016/j
.scitotenv. 2019.07.039.
Wolf, T.; McGregor, G. The Development of a Heat Wave Vulnerability Index for London, United Kingdom.
Weather and Climate Extremes 2013, 1, 59–68. https://doi.org/10.1016/j.wace. 2013.07.004.
Section 3.2. Air pollutant concentration information as a service – CL‑UHI impacts on

atmospheric composition
Baklanov, A.; Sørensen, J. H.; Hoe, S. C. et al. Urban Meteorological Modelling for Nuclear Emergency
Preparedness. Journal of Environmental Radioactivity 2006, 85 (2), 154–170. https://doi.org/10
.1016/j.jenvrad. 2005.01.018.
Bohnenstengel, S. I.; Hamilton, I.; Davies, M. et al. Impact of Anthropogenic Heat Emissions on London’s
Temperatures. Quarterly Journal of the Royal Meteorological Society 2014, 140 (679), 687–698.
https://doi.org/10.1002/qj. 2144.
Le Tertre, A.; Lefranc, A.; Eilstein, D. et al. Impact of the 2003 Heatwave on All‑Cause Mortality in 9 French
Cities. Epidemiology 2006, 17 (1), 75–79. https://doi.org/10.1097/01.ede.0000187650.36636.1f.
Section 3.3. Energy provision as a service – CL‑UHI impacts of and on energy use
Boßmann, T.; Staffell, I. The Shape of Future Electricity Demand: Exploring Load Curves in 2050s Germany
and Britain. Energy 2015, 90, 1317–1333. https://doi.org/10.1016/j.energy. 2015.06.082.
Curtis, M.; Torriti, J.; Smith, S. T. Demand Side Flexibility and Responsiveness: Moving Demand in Time
Through Technology. In Demanding Energy: Space, Time and Change; Hui, A., Day, R., Walker, G.,
Eds.; Springer International Publishing: Cham, 2018; 283–312. https://doi.org/10.1007/978-3
-319- 61991- 0_13.
Dodoo, A.; Gustavsson, L. Energy Use and Overheating Risk of Swedish Multi‑Storey Residential Buildings
under Different Climate Scenarios. Energy 2016, 97, 534–548. https://doi.org/10.1016/j.energy
.2015.12.086.
Essletzbichler, J. Renewable Energy Technology and Path Creation: A Multi‑Scalar Approach to Energy
Transition in the UK. European Planning Studies 2012, 20 (5), 791–816. https://doi.org/10.1080/
09654313. 2012.667926.
Habitzreuter, L.; Smith, S. T.; Keeling, T. Modelling the Overheating Risk in an Uniform High‑Rise Building
Design with a Consideration of Urban Context and Heatwaves. Indoor and Built Environment
2020, 29 (5), 671–688. https://doi.org/10.1177/1420326X19856400.
Hodson, M.; Marvin, S. Cities Mediating Technological Transitions: Understanding Visions, Intermediation
and Consequences. Technology Analysis & Strategic Management 2009, 21 (4), 515–534. https://
doi.org/10.1080/09537320902819213.
BIBLIOGRAPHY FOR FURTHER READING 81
Kamga, C.; Yazıcı, M. A. Temporal and Weather Related Variation Patterns of Urban Travel Time:
Considerations and Caveats for Value of Travel Time, Value of Variability, and Mode Choice
Studies. Transportation Research Part C: Emerging Technologies 2014, 45, 4–16. https://doi.org/10
.1016/j.trc. 2014.02.020.
Lindberg, F.; Grimmond, C. S. B.; Yogeswaran, N. et al. Impact of City Changes and Weather on
Anthropogenic Heat Flux in Europe 1995–2015. Urban Climate 2013, 4, 1–15. https://doi.org/10
.1016/j.uclim. 2013.03.002.
Santamouris, M.; Cartalis, C.; Synnefa, A. et al. On the Impact of Urban Heat Island and Global Warming on
the Power Demand and Electricity Consumption of Buildings – A Review. Energy and Buildings
2015, 98, 119–124. https://doi.org/10.1016/j.enbuild. 2014.09.052.
Strbac, G. Demand Side Management: Benefits and Challenges. Energy Policy 2008, 36 (12), 4419–4426.
https://doi.org/10.1016/j.enpol. 2008.09.030.
Tsapakis, I.; Cheng, T.; Bolbol, A. Impact of Weather Conditions on Macroscopic Urban Travel Times. Journal
of Transport Geography 2013, 28, 204–211. https://doi.org/10.1016/j.jtrangeo. 2012.11.003.
van Leeuwen, R. P.; de Wit, J. B.; Smit, G. J. M. Review of Urban Energy Transition in the Netherlands and
the Role of Smart Energy Management. Energy Conversion and Management 2017, 150, 941–948.
https://doi.org/10.1016/j.enconman. 2017.05.081.
Section 3.4. Urban vegetation as a service – CL‑UHI impacts on and from vegetation
Cariñanos, P.; Adinolfi, C.; Díaz de la Guardia, C. et al. Characterization of Allergen Emission Sources in
Urban Areas. Journal of Environmental Quality 2016, 45 (1), 244–252. https://doi.org/10. 2134/
jeq2015.02.0075.
Cariñanos, P.; Casares‑Porcel, M. Urban Green Zones and Related Pollen Allergy: A Review. Some
Guidelines for Designing Spaces with Low Allergy Impact. Landscape and Urban Planning 2011,
101 (3), 205–214. https://doi.org/10.1016/j.landurbplan. 2011.03.006.
Schatz, J.; Kucharik, C. J. Urban Heat Island Effects on Growing Seasons and Heating and Cooling Degree
Days in Madison, Wisconsin USA. International Journal of Climatology 2016, 36 (15), 4873–4884.
Zipper, S. C.; Schatz, J.; Singh, A. et al. Urban Heat Island Impacts on Plant Phenology: Intra‑Urban
Variability and Response to Land Cover. Environ Res Lett 2016, 11 (5). https://doi.org/10.1088/
1748-9326/11/5/054023.
Section 4.2. Detailed parameters at microscale
Li, M.; de Beurs, K. M.; Stein, A. et al. Incorporating Open Source Data for Bayesian Classification of Urban
Land Use From VHR Stereo Images. IEEE J Sel Top Appl Earth Observations Remote Sensing 2017,
10 (11), 4930–4943. https://doi.org/10.1109/JSTARS. 2017. 2737702.
Tremeac, B.; Bousquet, P.; de Munck, C. et al. Influence of Air Conditioning Management on Heat Island
in Paris Air Street Temperatures. Applied Energy 2012, 95, 102–110. https://doi.org/10.1016/j
.apenergy. 2012.02.015.
Section 4.3. Characterization at the local scale
Bechtel, B.; Alexander, P. J.; Böhner, J. et al. Mapping Local Climate Zones for a Worldwide Database of the
Form and Function of Cities. ISPRS International Journal of Geo‑Information 2015, 4 (1), 199–219.
https://doi.org/10.3390/ijgi4010199.
Bocher, E.; Petit, G.; Bernard, J. et al. A Geoprocessing Framework to Compute Urban Indicators: The
MApUCE Tools Chain. Urban Climate 2018, 24, 153–174. https://doi.org/10.1016/j.uclim
.2018.01.008.
Ching, J.; Aliaga, D.; Mills, G. et al. Pathway Using WUDAPT’s Digital Synthetic City Tool towards
Generating Urban Canopy Parameters for Multi‑Scale Urban Atmospheric Modeling. Urban
Climate 2019, 28. https://doi.org/10.1016/j.uclim. 2019.100459.
Demuzere, M.; Bechtel, B.; Middel, A. et al. Mapping Europe into Local Climate Zones. PLOS ONE 2019,
14 (4). https://doi.org/10.1371/journal.pone.0214474.
Demuzere, M.; Kittner, J.; Bechtel, B. LCZ Generator: A Web Application to Create Local Climate Zone
Maps. Frontiers in Environmental Science 2021, 9. https://doi.org/10.3389/fenvs. 2021.637455.
Gabey, A. M.; Grimmond, C. S. B.; Capel‑Timms, I. Anthropogenic Heat Flux: Advisable Spatial Resolutions
When Input Data Are Scarce. Theor Appl Climatol 2019, 135 (1), 791–807. https://doi.org/10
.1007/s00704- 018-2367-y.
(CL-UHI)
Haklay, M. How Good Is Volunteered Geographical Information? A Comparative Study of OpenStreetMap

and Ordnance Survey Datasets. Environ Plann B Plann Des 2010, 37 (4), 682–703. https://doi
.org/10.1068/b35097.
Hidalgo, J.; Dumas, G.; Masson, V. et al. Comparison between Local Climate Zones Maps Derived from
Administrative Datasets and Satellite Observations. Urban Climate 2019, 27, 64–89. https://doi
.org/10.1016/j.uclim. 2018.10.004.
Stewart, I. D.; Oke, T. R.; Krayenhoff, E. S. Evaluation of the ‘Local Climate Zone’ Scheme Using Temperature
Observations and Model Simulations. International Journal of Climatology 2014, 34 (4),
1062–1080. https://doi.org/10.1002/joc.3746.
Section 5.3. Measurement approaches
Basara, J. B.; Illston, B. G.; Fiebrich, C. A. et al. The Oklahoma City Micronet. Meteorological Applications
2011, 18 (3), 252–261. https://doi.org/10.1002/met.189.
Bell, S.; Cornford, D.; Bastin, L. How Good Are Citizen Weather Stations? Addressing a Biased Opinion.
Weather 2015, 70 (3), 75–84. https://doi.org/10.1002/wea. 2316.
Droste, A. M.; Pape, J. J.; Overeem, A. et al. Crowdsourcing Urban Air Temperatures through Smartphone
Battery Temperatures in São Paulo, Brazil. Journal of Atmospheric and Oceanic Technology 2017,
34 (9), 1853–1866. https://doi.org/10.1175/JTECH-D-16- 0150.1.
Kirk, P. J.; Clark, M. R.; Creed, E. Weather Observations Website. Weather 2021, 76 (2), 47–49. https://doi
.org/10.1002/wea.3856.
Mahoney, W. P.; O’Sullivan, J. M. Realizing the Potential of Vehicle‑Based Observations. Bulletin of the
American Meteorological Society 2013, 94 (7), 1007–1018. https://doi.org/10.1175/BAMS-D
-12- 00044.1.
Napoly, A.; Grassmann, T.; Meier, F. et al. Development and Application of a Statistically‑Based Quality
Control for Crowdsourced Air Temperature Data. Frontiers in Earth Science 2018, 6. https://doi
.org/10.3389/feart. 2018.00118.
Oke, T. R.; East, C. The Urban Boundary Layer in Montreal. Boundary‑Layer Meteorol 1971, 1 (4), 411–437.
https://doi.org/10.1007/BF00184781.
Overeem, A.; R. Robinson, J. C.; Leijnse, H. et al. Crowdsourcing Urban Air Temperatures from Smartphone
Battery Temperatures. Geophysical Research Letters 2013, 40 (15), 4081–4085. https://doi.org/
10.1002/grl. 50786.
Young, D. T.; Chapman, L.; Muller, C. L. et al. A Low‑Cost Wireless Temperature Sensor: Evaluation for Use
in Environmental Applications. Journal of Atmospheric and Oceanic Technology 2014, 31 (4),
938–944. https://doi.org/10.1175/JTECH-D-13- 00217.1.
Section 5.4. Choosing a site
Hellsten, A.; Luukkonen, S.‑M.; Steinfeld, G. et al. Footprint Evaluation for Flux and Concentration
Measurements for an Urban‑Like Canopy with Coupled Lagrangian Stochastic and Large‑Eddy
Simulation Models. Boundary‑Layer Meteorol 2015, 157 (2), 191–217. https://doi.org/10.1007/
s10546- 015- 0062- 4.
Kormann, R.; Meixner, F. X. An Analytical Footprint Model For Non‑Neutral Stratification. Boundary‑Layer
Meteorology 2001, 99 (2), 207–224. https://doi.org/10.1023/A:1018991015119.
Kumar, R.; Mishra, V.; Buzan, J. et al. Dominant Control of Agriculture and Irrigation on Urban Heat Island
in India. Sci Rep 2017, 7 (1). https://doi.org/10.1038/s41598- 017-14213-2.
Nakamura, Y.; Oke, T. R. Wind, Temperature and Stability Conditions in an East‑West Oriented Urban
Canyon. Atmospheric Environment (1967) 1988, 22 (12), 2691–2700. https://doi.org/10.1016/
0004- 6981(88)90437- 4.
Núñez Peiró, M.; Sánchez‑Guevara Sánchez, C.; Neila González, F. J. Source Area Definition for Local
Climate Zones Studies. A Systematic Review. Building and Environment 2019, 148, 258–285.
https://doi.org/10.1016/j.buildenv. 2018.10.050.
Section 5.6. Metadata for observations
Muller, C. L.; Chapman, L.; Grimmond, C. S. B. et al. Toward a Standardized Metadata Protocol for
Urban Meteorological Networks. Bulletin of the American Meteorological Society 2013, 94 (8),
1161–1185. https://doi.org/10.1175/BAMS-D-12- 00096.1.
Section 5.8. Observational challenges in brief
Foken, T., Ed. Springer Handbook of Atmospheric Measurements, 1st ed., 2021. https://doi.org/10.1007/978-3
-030-52171- 4.
World Meteorological Organization (WMO). WIGOS Metadata Standard (WMO‑No. 1192). Geneva, 2019.
Subsection 6.1.1. Statistical models
Arnds, D.; Böhner, J.; Bechtel, B. Spatio‑Temporal Variance and Meteorological Drivers of the Urban Heat
Island in a European City. Theor Appl Climatol 2017, 128 (1), 43–61. https://doi.org/10.1007/
s00704- 015-1687- 4.
Fortuniak, K. An Application of the Urban Energy Balance Scheme for a Statistical Modelling of the UHI
Intensity. Proc 5th Int Conf on Urban Climate, 2003, 1, 59‑62.
Gardes, T.; Schoetter, R.; Hidalgo, J. et al. Statistical Prediction of the Nocturnal Urban Heat Island Intensity
Based on Urban Morphology and Geographical Factors – An Investigation Based on Numerical
Model Results for a Large Ensemble of French Cities. Science of The Total Environment 2020, 737.
https://doi.org/10.1016/j.scitotenv. 2020.139253.
Hidalgo, J.; Jougla, R. On the Use of Local Weather Types Classification to Improve Climate Understanding:
An Application on the Urban Climate of Toulouse. PLOS ONE 2018, 13 (12). https://doi.org/10
.1371/journal.pone.0208138.
Schatz, J.; Kucharik, C. J. Seasonality of the Urban Heat Island Effect in Madison, Wisconsin. Journal of
Applied Meteorology and Climatology 2014, 53 (10), 2371–2386. https://doi.org/10.1175/JAMC
-D-14- 0107.1.
Straub, A.; Berger, K.; Breitner, S. et al. Statistical Modelling of Spatial Patterns of the Urban Heat Island
Intensity in the Urban Environment of Augsburg, Germany. Urban Climate 2019, 29. https://doi
.org/10.1016/j.uclim. 2019.100491.
Theeuwes, N. E.; Steeneveld, G.‑J.; Ronda, R. J. et al. A Diagnostic Equation for the Daily Maximum Urban
Heat Island Effect for Cities in Northwestern Europe. International Journal of Climatology 2017,
37 (1), 443–454. https://doi.org/10.1002/joc.4717.
Subsection 6.1.2. Obstacle‑resolving models
Baklanov, A.; Nuterman, R. Multi‑Scale Atmospheric Environment Modelling for Urban Areas. Advances in
Science and Research 2009, 3 (1), 53–57. https://doi.org/10. 5194/asr-3-53-2009.
Courant, R.; Friedrichs, K.; Lewy, H. On the Partial Difference Equations of Mathematical Physics. IBM
Journal of Research and Development 1967, 11 (2), 215–234. https://doi.org/10.1147/rd.112.0215.
Giometto, M. G.; Christen, A.; Meneveau, C. et al. Spatial Characteristics of Roughness Sublayer Mean
Flow and Turbulence Over a Realistic Urban Surface. Boundary‑Layer Meteorol 2016, 160 (3),
425–452. https://doi.org/10.1007/s10546- 016- 0157- 6.
Hertwig, D.; Gough, H. L.; Grimmond, S. et al. Wake Characteristics of Tall Buildings in a Realistic Urban
Canopy. Boundary‑Layer Meteorol 2019, 172 (2), 239–270. https://doi.org/10.1007/s10546
-019- 00450-7.
Kanda, M.; Inagaki, A.; Miyamoto, T. et al. A New Aerodynamic Parametrization for Real Urban Surfaces.
Boundary‑Layer Meteorol 2013, 148 (2), 357–377. https://doi.org/10.1007/s10546- 013-9818-x.
Maronga, B.; Gross, G.; Raasch, S. et al. Development of a New Urban Climate Model Based on the Model
PALM – Project Overview, Planned Work, and First Achievements. Meteorologische Zeitschrift
2019, 105–119. https://doi.org/10.1127/metz/2019/0909.
Resler, J.; Krč, P.; Belda, M. et al. PALM‑USM v1.0: A New Urban Surface Model Integrated into the PALM
Large‑Eddy Simulation Model. Geoscientific Model Development 2017, 10 (10), 3635–3659.
https://doi.org/10. 5194/gmd-10-3635-2017.
Salim, M. H.; Schlünzen, K. H.; Grawe, D. Including Trees in the Numerical Simulations of the Wind Flow
in Urban Areas: Should We Care? Journal of Wind Engineering and Industrial Aerodynamics 2015,
144, 84–95. https://doi.org/10.1016/j.jweia. 2015.05.004.
Salim, M. H.; Schlünzen, K. H.; Grawe, D. et al. The Microscale Obstacle‑Resolving Meteorological Model
MITRAS v2.0: Model Theory. Geoscientific Model Development 2018, 11 (8), 3427–3445. https://
doi.org/10. 5194/gmd-11-3427-2018.
Schlünzen, K. H.; Grawe, D.; Bohnenstengel, S. I. et al. Joint Modelling of Obstacle Induced and
Mesoscale Changes – Current Limits and Challenges. Journal of Wind Engineering and Industrial
Aerodynamics 2011, 99 (4), 217–225. https://doi.org/10.1016/j.jweia. 2011.01.009.
(CL-UHI)
Subsection 6.1.3. Numerical weather prediction and climate models
Baklanov, A.; Mestayer, P. G.; Clappier, A. et al. Towards Improving the Simulation of Meteorological Fields
in Urban Areas through Updated/Advanced Surface Fluxes Description. Atmospheric Chemistry
and Physics 2008, 8 (3), 523–543. https://doi.org/10. 5194/acp- 8-523-2008.
Best, M. J. Representing Urban Areas within Operational Numerical Weather Prediction Models.
Boundary‑Layer Meteorol 2005, 114 (1), 91–109. https://doi.org/10.1007/s10546- 004- 4834-5.
Best, M. J.; Pryor, M.; Clark, D. B. et al. The Joint UK Land Environment Simulator (JULES), Model
Description – Part 1: Energy and Water Fluxes. Geoscientific Model Development 2011, 4 (3),
677–699. https://doi.org/10. 5194/gmd- 4- 677-2011.
Dupont, S.; Mestayer, P. G. Parameterization of the Urban Energy Budget with the Submesoscale Soil
Model. Journal of Applied Meteorology and Climatology 2006, 45 (12), 1744–1765. https://doi
.org/10.1175/JAM2417.1.
Dupont, S.; Mestayer, P. G.; Guilloteau, E. et al. Parameterization of the Urban Water Budget with the
Submesoscale Soil Model. Journal of Applied Meteorology and Climatology 2006, 45 (4), 624–648.
https://doi.org/10.1175/JAM2363.1.
Fisher, B.; Kukkonen, J.; Piringer, M. et al. Meteorology Applied to Urban Air Pollution Problems: Concepts
from COST 715. Atmos Chem Phys Discuss 2005, 5, 7903–7937. https://doi.org/10. 5194/acpd
-5-7903-2005.
Grawe, D.; Thompson, H. L.; Salmond, J. A. et al. Modelling the Impact of Urbanisation on Regional Climate
in the Greater London Area. International Journal of Climatology 2013, 33 (10), 2388–2401.
Grimmond, C. S. B.; Best, M. J.; Barlow, J. et al. Urban Surface Energy Balance Models: Model
Characteristics and Methodology for a Comparison Study. In Meteorological and Air Quality
Models for Urban Areas; Baklanov, A., Sue, G., Alexander, M., Athanassiadou, M., Eds.; Springer:
Berlin, Heidelberg, 2009; 97–123. https://doi.org/10.1007/978-3- 642- 00298- 4_11.
Grimmond, C. S. B.; Blackett, M.; Best, M. J. et al. Initial Results from Phase 2 of the International Urban
Energy Balance Model Comparison. International Journal of Climatology 2011, 31 (2), 244–272.
https://doi.org/10.1002/joc. 2227.
Grimmond, C. S. B.; Blackett, M.; Best, M. J. et al. The International Urban Energy Balance Models
Comparison Project: First Results from Phase 1. Journal of Applied Meteorology and Climatology
2010, 49 (6), 1268–1292. https://doi.org/10.1175/2010JAMC2354.1.
Hamdi, R.; Masson, V. Inclusion of a Drag Approach in the Town Energy Balance (TEB) Scheme: Offline 1D
Evaluation in a Street Canyon. Journal of Applied Meteorology and Climatology 2008, 47 (10),
2627–2644. https://doi.org/10.1175/2008JAMC1865.1.
Hertwig, D.; Grimmond, S.; Hendry, M. A. et al. Urban Signals in High‑Resolution Weather and Climate
Simulations: Role of Urban Land‑Surface Characterisation. Theor Appl Climatol 2020, 142 (1),
701–728. https://doi.org/10.1007/s00704- 020- 03294-1.
Järvi, L.; Grimmond, C. S. B.; Christen, A. The Surface Urban Energy and Water Balance Scheme (SUEWS):
Evaluation in Los Angeles and Vancouver. Journal of Hydrology 2011, 411 (3), 219–237. https://
doi.org/10.1016/j.jhydrol. 2011.10.001.
Kusaka, H.; Kondo, H.; Kikegawa, Y. et al. A Simple Single‑Layer Urban Canopy Model For Atmospheric
Models: Comparison With Multi‑Layer And Slab Models. Boundary‑Layer Meteorology 2001,
101 (3), 329–358. https://doi.org/10.1023/A:1019207923078.
Lean, H. W.; Barlow, J. F.; Halios, C. H. The Impact of Spin‑up and Resolution on the Representation of a
Clear Convective Boundary Layer over London in Order 100 m Grid‑Length Versions of the
Met Office Unified Model. Quarterly Journal of the Royal Meteorological Society 2019, 145 (721),
1674–1689. https://doi.org/10.1002/qj.3519.
Lemonsu, A.; Masson, V.; Shashua‑Bar, L. et al. Inclusion of Vegetation in the Town Energy Balance Model
for Modelling Urban Green Areas. Geoscientific Model Development 2012, 5 (6), 1377–1393.
https://doi.org/10. 5194/gmd-5-1377-2012.
Loridan, T.; Grimmond, C. S. B. Characterization of Energy Flux Partitioning in Urban Environments: Links
with Surface Seasonal Properties. Journal of Applied Meteorology and Climatology 2012, 51 (2),
219–241. https://doi.org/10.1175/JAMC-D-11- 038.1.
Loridan, T.; Grimmond, C. S. B. Multi‑Site Evaluation of an Urban Land‑Surface Model: Intra‑Urban
Heterogeneity, Seasonality and Parameter Complexity Requirements. Quarterly Journal of the
Royal Meteorological Society 2012, 138 (665), 1094–1113. https://doi.org/10.1002/qj.963.
Masson, V. A Physically‑Based Scheme For The Urban Energy Budget In Atmospheric Models.
Boundary‑Layer Meteorology 2000, 94 (3), 357–397. https://doi.org/10.1023/A:1002463829265.
Oleson, K. W.; Bonan, G. B.; Feddema, J. et al. An Urban Parameterization for a Global Climate Model.
Part I: Formulation and Evaluation for Two Cities. Journal of Applied Meteorology and Climatology
2008, 47 (4), 1038–1060. https://doi.org/10.1175/2007JAMC1597.1.
Porson, A.; Clark, P. A.; Harman, I. N. et al. Implementation of a New Urban Energy Budget Scheme
in the MetUM. Part I: Description and Idealized Simulations. Quarterly Journal of the Royal
Meteorological Society 2010, 136 (651), 1514–1529. https://doi.org/10.1002/qj.668.
Schlünzen, K. H.; Katzfey, J. J. Relevance of Sub‑Grid‑Scale Land‑Use Effects for Mesoscale Models. Tellus A:
Dynamic Meteorology and Oceanography 2003, 55 (3), 232–246. https://doi.org/10.3402/tellusa
.v55i3.12095.
Schoetter, R.; Kwok, Y. T.; de Munck, C. et al. Multi‑Layer Coupling between SURFEX‑TEB‑v9.0 and
Meso‑NH‑v5.3 for Modelling the Urban Climate of High‑Rise Cities. Geoscientific Model
Development 2020, 13 (11), 5609–5643. https://doi.org/10. 5194/gmd-13-5609-2020.
Seity, Y.; Brousseau, P.; Malardel, S. et al. The AROME‑France Convective‑Scale Operational Model. Monthly
Weather Review 2011, 139 (3), 976–991. https://doi.org/10.1175/2010MWR3425.1.
Stavropulos‑Laffaille, X.; Chancibault, K.; Andrieu, H. et al. Coupling Detailed Urban Energy and Water
Budgets with TEB‑Hydro Model: Towards an Assessment Tool for Nature Based Solution
Performances. Urban Climate 2021, 39. https://doi.org/10.1016/j.uclim. 2021.100925.
Section 6.3. Evaluation of model skill
Best, M. J.; Grimmond, C. S. B. Importance of Initial State and Atmospheric Conditions for Urban Land
Surface Models’ Performance. Urban Climate 2014, 10, 387–406. https://doi.org/10.1016/j
.uclim. 2013.10.006.
Best, M. J.; Grimmond, C. S. B. Modeling the Partitioning of Turbulent Fluxes at Urban Sites with Varying
Vegetation Cover. Journal of Hydrometeorology 2016, 17 (10), 2537–2553. https://doi.org/10
.1175/JHM-D-15- 0126.1.
Schlünzen, K. H. Standards for Evaluation of Atmospheric Models in Environmental Meteorology.
In Computer Simulation Validation: Fundamental Concepts, Methodological Frameworks, and
Philosophical Perspectives; Beisbart, C., Saam, N. J., Eds.; Springer International Publishing:
Cham, 2019; 563–586. https://doi.org/10.1007/978-3-319-70766-2_ 23.
Schlünzen, K. H.; Builtjes, M. R.; Deserti, J. et al. Evaluating the performance of mesoscale meteorology
models used for air quality simulations. In Mesoscale Modelling for Meteorological and Air
Pollution Applications; Sokhi, R. A.; Baklanov, A.; Schlünzen, K. H. et al., Eds. Anthem, London,
2018. https://anthempress.com/mesoscale-modelling-for-meteorological-and-air-pollution
-applications-hb.
Verein Deutscher Ingenieure. VDI 3783 Blatt 7 – Environmental Meteorology – Prognostic Microscale Wind Field
Models – Evaluation for dynamically and thermally induced flow fields; Düsseldorf, 2017. https://
www.beuth.de/de/technische-regel/vdi-3783 -blatt-7/267500583?websource=vdin.
Verein Deutscher Ingenieure. VDI 3783 Blatt 9 – Environmental Meteorology – Prognostic Microscale Wind Field
Models – Evaluation for Flow around Buildings and Obstacles; Düsseldorf, 2017. https://www.beuth
.de/en/technical-rule/vdi-3783-blatt-9/267500591.
Subsection 6.4.3. Climate predictions (sub‑seasonal and longer) and projections
Amorim, J. H.; Segersson, D.; Körnich, H. et al. High Resolution Simulation of Stockholm’s Air Temperature
and Its Interactions with Urban Development. Urban Climate 2020, 32. https://doi.org/10
.1016/j.uclim. 2020.100632.
Bruyère, C. L.; Done, J. M.; Holland, G. J. et al. Bias Corrections of Global Models for Regional Climate
Simulations of High‑Impact Weather. Clim Dyn 2014, 43 (7), 1847–1856. https://doi.org/10
.1007/s00382- 013-2011- 6.
Bruyere, L.; Monaghan, J.; Steinhoff, F. et al. Bias‑Corrected CMIP5 CESM Data in WRF/MPAS Intermediate
File Format. 2015. https://doi.org/10. 5065/D6445JJ7.
Chan, S. C.; Kendon, E. J.; Berthou, S. et al. Europe‑Wide Precipitation Projections at Convection Permitting
Scale with the Unified Model. Clim Dyn 2020, 55 (3), 409–428. https://doi.org/10.1007/s00382
-020- 05192- 8.
Chapman, S.; Thatcher, M.; Salazar, A. et al. The Impact of Climate Change and Urban Growth on Urban
Climate and Heat Stress in a Subtropical City. International Journal of Climatology 2019, 39 (6),
3013–3030. https://doi.org/10.1002/joc. 5998.
(CL-UHI)
Daniel, M.; Lemonsu, A.; Déqué, M. et al. Benefits of Explicit Urban Parameterization in Regional Climate
Modeling to Study Climate and City Interactions. Clim Dyn 2019, 52 (5), 2745–2764. https://doi
.org/10.1007/s00382- 018- 4289-x.
Duchêne, F.; Schaeybroeck, B. V.; Caluwaerts, S. et al. A Statistical–Dynamical Methodology to Downscale
Regional Climate Projections to Urban Scale. Journal of Applied Meteorology and Climatology
2020, 59 (6), 1109–1123. https://doi.org/10.1175/JAMC-D-19- 0104.1.
Goggins, W. B.; Chan, E. Y. Y.; Ng, E. et al. Effect Modification of the Association between Short‑Term
Meteorological Factors and Mortality by Urban Heat Islands in Hong Kong. PLOS ONE 2012,
7 (6), e38551. https://doi.org/10.1371/journal.pone.0038551.
González‑Aparicio, I.; Baklanov, A.; Hidalgo, J. et al. Impact of City Expansion and Increased Heat Fluxes
Scenarios on the Urban Boundary Layer of Bilbao Using Enviro‑HIRLAM. Urban Climate 2014,
10, 831–845. https://doi.org/10.1016/j.uclim. 2014.07.010.
Hamdi, R.; Van de Vyver, H.; De Troch, R. et al. Assessment of Three Dynamical Urban Climate Downscaling
Methods: Brussels’s Future Urban Heat Island under an A1B Emission Scenario. International
Journal of Climatology 2014, 34 (4), 978–999. https://doi.org/10.1002/joc.3734.
Hawkins, E.; Osborne, T. M.; Ho, C. K. et al. Calibration and Bias Correction of Climate Projections for Crop
Modelling: An Idealised Case Study over Europe. Agricultural and Forest Meteorology 2013, 170,
19–31. https://doi.org/10.1016/j.agrformet. 2012.04.007.
Heaviside, C.; Vardoulakis, S.; Cai, X.‑M. Attribution of Mortality to the Urban Heat Island during
Heatwaves in the West Midlands, UK. Environmental Health 2016, 15 (1). https://doi.org/10
.1186/s12940- 016- 0100-9.
Hoffmann, P.; Krueger, O.; Schlünzen, K. H. A Statistical Model for the Urban Heat Island and Its Application
to a Climate Change Scenario. International Journal of Climatology 2012, 32 (8), 1238–1248.
https://doi.org/10.1002/joc. 2348.
Holland, G.; Done, J.; Bruyère, C. L. et al. Model Investigations of the Effects of Climate Variability and
Change on Future Gulf of Mexico Tropical Cyclone Activity. 2010, 2. https://doi.org/10
.4043/20690 - MS.
Hondula, D. M.; Georgescu, M.; Balling, R. C. Challenges Associated with Projecting Urbanization‑Induced
Heat‑Related Mortality. Science of The Total Environment 2014, 490, 538–544. https://doi.org/10
.1016/j.scitotenv. 2014.04.130.
Kendon, E. J.; Roberts, N. M.; Senior, C. A. et al. Realism of Rainfall in a Very High‑Resolution Regional
Climate Model. Journal of Climate 2012, 25 (17), 5791–5806. https://doi.org/10.1175/JCLI-D
-11- 00562.1.
Laaidi, K.; Zeghnoun, A.; Dousset, B. et al. The Impact of Heat Islands on Mortality in Paris during the
August 2003 Heat Wave. Environ Health Perspect 2012, 120 (2), 254–259. https://doi.org/10
.1289/ehp.1103532.
Le Roy, B.; Lemonsu, A.; Schoetter, R. A Statistical–Dynamical Downscaling Methodology for the Urban
Heat Island Applied to the EURO‑CORDEX Ensemble. Clim Dyn 2021, 56 (7), 2487–2508.
https://doi.org/10.1007/s00382- 020- 05600-z.
Milojevic, A.; Armstrong, B. G.; Gasparrini, A. et al. Methods to Estimate Acclimatization to Urban Heat
Island Effects on Heat‑ and Cold‑Related Mortality. Environ Health Perspect 2016, 124 (7),
1016–1022. https://doi.org/10.1289/ehp.1510109.
Piani, C.; Weedon, G. P.; Best, M. et al. Statistical Bias Correction of Global Simulated Daily Precipitation
and Temperature for the Application of Hydrological Models. Journal of Hydrology 2010,
395 (3), 199–215. https://doi.org/10.1016/j.jhydrol. 2010.10.024.
Taylor, J.; Wilkinson, P.; Davies, M. et al. Mapping the Effects of Urban Heat Island, Housing, and Age on
Excess Heat‑Related Mortality in London. Urban Climate 2015, 14, 517–528. https://doi.org/10
.1016/j.uclim. 2015.08.001.
Wuebbles, D. Ed., Report on the Workshop on Urban Scale Processes and their Representation in High Spatial
Resolution Earth System Models; Argonne National Laboratory: Lemont, USA, 2019.
Section 6.5. Metadata for modelling results
Eyring, V.; Bony, S.; Meehl, G. A. et al. Overview of the Coupled Model Intercomparison Project Phase 6
(CMIP6) Experimental Design and Organization. Geoscientific Model Development 2016, 9 (5),
1937–1958. https://doi.org/10. 5194/gmd-9-1937-2016.
Taylor, K. E.; Stouffer, R. J.; Meehl, G. A. An Overview of CMIP5 and the Experiment Design. Bulletin of
the American Meteorological Society 2012, 93 (4), 485–498. https://doi.org/10.1175/BAMS-D
-11- 00094.1.
Chapter 7. Monitoring the CL‑UHI
Tan, J.; Yang, L.; Grimmond, C. S. B. et al. Urban Integrated Meteorological Observations: Practice and
Experience in Shanghai, China. Bulletin of the American Meteorological Society 2015, 96 (1),
85–102. https://doi.org/10.1175/BAMS-D-13- 00216.1.
Chapter 8. Understanding the impacts of CL‑UHI mitigation and adaptation efforts
Akbari, H.; Kolokotsa, D. Three Decades of Urban Heat Islands and Mitigation Technologies Research.
Energy and Buildings 2016, 133, 834–842. https://doi.org/10.1016/j.enbuild. 2016.09.067.
Bowler, D. E.; Buyung‑Ali, L.; Knight, T. M. et al. Urban Greening to Cool Towns and Cities: A Systematic
Review of the Empirical Evidence. Landscape and Urban Planning 2010, 97 (3), 147–155. https://
doi.org/10.1016/j.landurbplan. 2010.05.006.
Cui, F.; Hamdi, R.; Yuan, X. et al. Quantifying the Response of Surface Urban Heat Island to Urban Greening
in Global North Megacities. Science of The Total Environment 2021, 801. https://doi.org/10.1016/
j.scitotenv. 2021.149553.
Erell, E. Is Urban Heat Island Mitigation Necessarily a Worthy Objective? Proceedings of 33rd PLEA
International Conference 2017, 2, 1693–1700. http://www.scopus.com/inward/record.url?scp =
85043994421&partnerID = 8YFLogxK.
Erell, E. The Application of Urban Climate Research in the Design of Cities. Advances in Building Energy
Research 2008, 2 (1), 95–121. https://doi.org/10.3763/aber. 2008.0204.
Erell, E.; Pearlmutter, D.; Boneh, D. Effect of High‑Albedo Materials on Pedestrian Heat Stress in Urban
Street Canyons. Urban Climate 2013, 10, 367–386. https://doi.org/10.1016/j.uclim. 2013.10.005.
Erell, E.; Pearlmutter, D.; Williamson, T. Urban Microclimate: Designing the Spaces Between Buildings;
Routledge: London, 2010. https://doi.org/10.4324/9781849775397.
Gromke, C.; Ruck, B. Pollutant Concentrations in Street Canyons of Different Aspect Ratio with Avenues of
Trees for Various Wind Directions. Boundary‑Layer Meteorol 2012, 144 (1), 41–64. https://doi
.org/10.1007/s10546- 012-9703-z.
He, J. F.; Liu, J. Y.; Zhuang, D. F. et al. Assessing the Effect of Land Use/Land Cover Change on the Change of
Urban Heat Island Intensity. Theor Appl Climatol 2007, 90 (3), 217–226. https://doi.org/10.1007/
s00704- 006- 0273-1.
Hoffmann, P.; Fischereit, J.; Heitmann, S. et al. Modeling Exposure to Heat Stress with a Simple Urban
Model. Urban Science 2018, 2 (1), 9. https://doi.org/10.3390/urbansci2010009.
Kokkonen, T. V.; Grimmond, C. S. B.; Christen, A. et al. Changes to the Water Balance Over a Century of
Urban Development in Two Neighborhoods: Vancouver, Canada. Water Resources Research
2018, 54 (9), 6625–6642. https://doi.org/10.1029/2017WR022445.
Li, D.; Bou‑Zeid, E.; Oppenheimer, M. The Effectiveness of Cool and Green Roofs as Urban Heat
Island Mitigation Strategies. Environ Res Lett 2014, 9 (5). https://doi.org/10.1088/1748
-9326/9/5/055002.
Martilli, A.; Roth, M.; Chow, W. T. L. et al. Summer Average Urban‑Rural Surface Temperature Differences Do Not
Indicate the Need for Urban Heat Reduction; OSF, 2020. https://doi.org/10.31219/osf.io/8gnbf.
Ng, E.; Yuan, C.; Chen, L. et al. Improving the Wind Environment in High‑Density Cities by Understanding
Urban Morphology and Surface Roughness: A Study in Hong Kong. Landscape and Urban
Planning 2011, 101 (1), 59–74. https://doi.org/10.1016/j.landurbplan. 2011.01.004.
Ren, C.; Ng, E. Y.; Katzschner, L. Urban Climatic Map Studies: A Review. International Journal of Climatology
2011, 31 (15), 2213–2233. https://doi.org/10.1002/joc. 2237.
Ren, C.; Yang, R.; Cheng, C. et al. Creating Breathing Cities by Adopting Urban Ventilation Assessment and
Wind Corridor Plan – The Implementation in Chinese Cities. Journal of Wind Engineering and
Industrial Aerodynamics 2018, 182, 170–188. https://doi.org/10.1016/j.jweia. 2018.09.023.
Santamouris, M.; Synnefa, A.; Karlessi, T. Using Advanced Cool Materials in the Urban Built Environment
to Mitigate Heat Islands and Improve Thermal Comfort Conditions. Solar Energy 2011, 85 (12),
3085–3102. https://doi.org/10.1016/j.solener. 2010.12.023.
Santiago, J.‑L.; Martilli, A.; Martin, F. On Dry Deposition Modelling of Atmospheric Pollutants on
Vegetation at the Microscale: Application to the Impact of Street Vegetation on Air Quality.
Boundary‑Layer Meteorology 2017, 162, 451–474. https://doi.org/10.1007/s10546- 016- 0210-5.
Shashua‑Bar, L.; Pearlmutter, D.; Erell, E. The Influence of Trees and Grass on Outdoor Thermal Comfort in a
Hot‑Arid Environment. International Journal of Climatology 2011, 31 (10), 1498–1506. https://doi
.org/10.1002/joc. 2177.
(CL-UHI)
Sproul, J.; Wan, M. P.; Mandel, B. H. et al. Economic Comparison of White, Green, and Black Flat Roofs
in the United States. Energy and Buildings 2014, 71, 20–27. https://doi.org/10.1016/j.enbuild
.2013.11.058.
van den Bosch, M.; Nieuwenhuijsen, M. No Time to Lose – Green the Cities Now. Environment International
2017, 99, 343–350. https://doi.org/10.1016/j.envint. 2016.11.025.
Wheeler, B. W.; Lovell, R.; Higgins, S. L. et al. Beyond Greenspace: An Ecological Study of Population
General Health and Indicators of Natural Environment Type and Quality. International Journal of
Health Geographics 2015, 14 (1). https://doi.org/10.1186/s12942- 015- 0009-5.
Appendix 1. Example case study for influences on CL‑UHI and other temperatures
Lauwaet, D.; Hooyberghs, H.; Maiheu, B. et al. Detailed Urban Heat Island Projections for Cities Worldwide:
Dynamical Downscaling CMIP5 Global Climate Models. Climate 2015, 3 (2), 391–415. https://
doi.org/10.3390/cli3020391.
Rasilla, D.; Allende, F.; Martilli, A. et al. Heat Waves and Human Well‑Being in Madrid (Spain). Atmosphere
2019, 10 (5). https://doi.org/10.3390/atmos10050288.
Salamanca, F.; Krpo, A.; Martilli, A. et al. A New Building Energy Model Coupled with an Urban Canopy
Parameterization for Urban Climate Simulations – Part I. Formulation, Verification, and
Sensitivity Analysis of the Model. Theoretical and Applied Climatology 2009, 99, 331–344. https://
doi.org/10.1007/s00704- 009- 0142-9.
Skamarock, C.; Klemp, B.; Dudhia, J. et al. A Description of the Advanced Research WRF Version 3. 2008.
https://doi.org/10. 5065/D68S4MVH.
For more information, please contact:
World Meteorological Organization

7 bis, avenue de la Paix – P.O. Box 2300 – CH 1211 Geneva 2 – Switzerland
Strategic Communications Office

Tel.: +41 (0) 22 730 83 14 – Fax: +41 (0) 22 730 80 27
Email: cpa@wmo.int
public.wmo.int
JN 22330

WMO 2023 - Guidance On Measuring, Modelling and Monitoring The Canopy Layer Urban Heat Island

Uploaded by

Copyright:

Available Formats

WMO 2023 - Guidance On Measuring, Modelling and Monitoring The Canopy Layer Urban Heat Island

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

WMO 2023 - Guidance On Measuring, Modelling and Monitoring The Canopy Layer Urban Heat Island

Uploaded by

Copyright:

Available Formats

Guidance on Measuring, Modelling

and Monitoring the Canopy Layer

© World Meteorological Organization, 2023

Chair, Publications Board

CHAPTER 2. GENERAL INFORMATION ON THE CL‑UHI. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  3

CHAPTER 3. URBAN SERVICES NEEDING CL‑UHI INFORMATION . . . . . . . . . . . . . . . . . . . . . .  15

CHAPTER 4. CHARACTERIZATION OF THE URBAN AREA AND ITS SURROUNDINGS. . . . .  19

CHAPTER 5. DETERMINING THE CL‑UHI FROM OBSERVATIONS. . . . . . . . . . . . . . . . . . . . . . .  26

5.4.3 Thermal source areas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

CHAPTER 6. DETERMINING THE CL‑UHI WITH MATHEMATICAL MODELS . . . . . . . . . . . . .  38

CHAPTER 7. MONITORING THE CL‑UHI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  50

CHAPTER 8. UNDERSTANDING THE IMPACTS OF CL‑UHI MITIGATION AND

APPENDIX 1. A CASE STUDY OF INFLUENCES ON CL‑UHI AND OTHER

APPENDIX 2. EXAMPLES OF OBSERVATION NETWORKS. . . . . . . . . . . . . . . . . . . . . . . . . . . . .  67

APPENDIX 3. MONITORING EXAMPLES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  69

BIBLIOGRAPHY FOR FURTHER READING. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  79

Members to improve connections among [National Meteorological and

Coordinating lead authors: K. Heinke Schlünzen (Universität Hamburg, Germany),

Benjamin Bechtel (Ruhr‑University Bochum, Germany), Edmilson Dias de Freitas (University of

Table 1.1. Main elements of this document

Overview of processes Chapter 2:

Applications needing Chapter 3:

Methods to determine Chapter 4:

Mitigation of CL‑UHI Chapter 8:

2.1 HISTORICAL BACKGROUND

2.2 HORIZONTAL SCALES AND VERTICAL LAYERS IN URBAN AREAS

Horizontal Vertical Related Atmo­spheric

(a) Microscale (b) Local scale

Rural Urban Rural

2.3 DEFINING THE CL‑UHI

Table 2.2. CL-UHI metrics based on ΔTu–r, use cases with

(a) CL‑UHI metric Use case Conditions of likely occurrence

Irrigated urban parks with dense vegetation

Cities surrounded by arid or semi‑arid rural areas that

2.4 CL‑UHI METRICS

2.5 DEVELOPMENT OF THE CL‑UHI

2.5.1 Surface influences

– Surface state: Surface waterproofing by buildings and paving reduces subsurface

2.5.2 Interaction of physical processes

2.5.3 Ideal weather conditions

2.5.4 Influences of actual weather conditions

Middle Middle Middle

2.5.5 Influences of city form and geographic setting

Figure 2.3. Isothermal maps during clear, calm weather.

2.7 CL‑UHI EFFECTS ON METEOROLOGICAL AND CLIMATOLOGICAL

2.8 THE CL‑UHI IN BRIEF

– The meteorological situation influences exchange processes at surfaces (especially heat

Information on current and future CL‑UHIs helps to improve forecasts provided by an

3.1 HEAT STRESS INFORMATION AS A SERVICE – CL‑UHI IMPACTS ON HEALTH

3.1.1 Thermal comfort assessment

3.1.2 Heat‑health outcomes and early warnings

3.2 AIR POLLUTANT CONCENTRATION INFORMATION AS A SERVICE –

With increasing urbanization, a larger proportion of the world’s population is exposed to

3.3 ENERGY PROVISION AS A SERVICE – CL‑UHI IMPACTS OF AND

3.4 URBAN VEGETATION AS A SERVICE – CL‑UHI IMPACTS ON AND FROM

3.5 MULTI‑HAZARD EARLY WARNING SYSTEM –

4.1 TYPES OF PARAMETERS NEEDED

4.2 DETAILED PARAMETERS AT MICROSCALE

Table 4.1. Characteristics to describe urban and rural areas

CHAPTER 2. GENERAL INFORMATION ON THE CL‑UHI. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

CHAPTER 3. URBAN SERVICES NEEDING CL‑UHI INFORMATION . . . . . . . . . . . . . . . . . . . . . . 15

CHAPTER 4. CHARACTERIZATION OF THE URBAN AREA AND ITS SURROUNDINGS. . . . . 19

CHAPTER 5. DETERMINING THE CL‑UHI FROM OBSERVATIONS. . . . . . . . . . . . . . . . . . . . . . . 26

5.4.3 Thermal source areas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

CHAPTER 6. DETERMINING THE CL‑UHI WITH MATHEMATICAL MODELS . . . . . . . . . . . . . 38

CHAPTER 7. MONITORING THE CL‑UHI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

APPENDIX 2. EXAMPLES OF OBSERVATION NETWORKS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

APPENDIX 3. MONITORING EXAMPLES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

BIBLIOGRAPHY FOR FURTHER READING. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

Horizontal Vertical Related Atmospheric