Water mist spray for outdoor cooling: A systematic review of technologies, methods and impacts

Abstract

For the first time, a systematic review was conducted on mist spraying systems used for outdoor cooling by perusing twenty years of publications from 12 countries and 7 climatic zones. The twofold aim was to emphasize both the potential against local overheating in a variety of climatic contexts and the extreme heterogeneity in terms of investigation techniques and performance metrics that hinder the construction of a cohesive body of knowledge. In addition to statistics and patterns, data were screened to outline theoretical and methodological trends and gaps and to detect geographic biases and climate dependencies. Indeed, each study was thoroughly described and comparatively discussed according to (i) the investigational method (purely experimental studies, purely numerical studies and those combining field tests with simulations), (ii) the results in terms of cooling, humidification and comfort, also in relation to the adopted performance metrics (iii) the design novelty. Most relevant approaches and findings were discussed and compared to identify governing variables, optimized configurations, unchartered solutions and criticalities. Overall, the collected data qualify water spraying as a cost-effective, versatile and high-impact blue mitigator. Opportunities and challenges towards an informed use emerged and will help delineating appropriate guidelines for practitioners involved in town development, to deliver strategies and precautions.

Full text

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
Water mist spray for outdoor cooling: A systematic review of technologies,
methods and impacts
Giulia Ulpiani1
Published in
Applied Energy
https://doi.org/10.1016/j.apenergy.2019.113647
©2019. This manuscript version is made available under the Creative Commons Attribution-
ShareAlike 4.0 International License. To view a copy of this license, visit
http://creativecommons.org/licenses/by-sa/4.0/
or send a letter to
Creative Commons, PO Box 1866, Mountain View, CA 94042, USA.
1Dipartimento di Ingegneria Industriale e Scienze Matematiche (DIISM), Facoltà di Ingegneria, Università Politecnica delle Marche,
Via Brecce Bianche, 60131 Ancona, Italy.

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
Title
Water mist spray for outdoor cooling: A systematic review of technologies, methods and impacts.
Author
Giulia Ulpiani
E-mail: g.ulpiani@pm.univpm.it
Mobile phone: +393290615752
Polytechnic University of Marche
Department of Industrial Engineering and Mathematical Sciences (DIISM)
12, Via Brecce Bianche, Ancona 60131, Italy
Abstract
For the first time, a systematic review was conducted on mist spraying systems used for outdoor cooling by
perusing twenty years of publications from 12 countries and 7 climatic zones. The twofold aim was to
emphasize both the potential against local overheating in a variety of climatic contexts and the extreme
heterogeneity in terms of investigation techniques and performance metrics that hinder the construction of a
cohesive body of knowledge. In addition to statistics and patterns, data were screened to outline theoretical
and methodological trends and gaps and to detect geographic biases and climate dependencies. Indeed, each
study was thoroughly described and comparatively discussed according to i) the investigational method
(purely experimental studies, purely numerical studies and those combining field tests with simulations), ii)
the results in terms of cooling, humidification and comfort, also in relation to the adopted performance metrics
iii) the design novelty. Most relevant approaches and findings were discussed and compared to identify
governing variables, optimized configurations, unchartered solutions and criticalities. Overall, the collected
data qualify water spraying as a cost-effective, versatile and high-impact blue mitigator. Opportunities and
challenges towards an informed use emerged and will help delineating appropriate guidelines for practitioners
involved in town development, to deliver strategies and precautions.
Keywords: water mist spray; evaporative cooling; energy efficiency; outdoor comfort; urban climate; heat
mitigation.
Nomenclature (alphabetic order)
A.I. Artificial Intelligence H Humidex
ACEC Air Conditioning Energy Consumption IPCC Intergovernmental Panel on Climate
Change
Af Tropical rainforest climate n.s. Non specified
ANOVA Analysis of variance PA Parametric Analysis
ASHRAE American Society of Heating Refrigerating
and Air-conditioning Engineers PE Cooling Power Index [mcal/cm
2
s]
Aw Tropical wet and dry or savanna climate PET Physiological Equivalent Temperature [°C]
Bwh Hot desert climate PV Photovoltaic
CE Cooling Efficacy, also η or ε [%] RH relative humidity [%]
Cfa Humid subtropical climate ROI Return On Investment [years]
Cfb Temperate oceanic climate SET* Standard Effective Temperature [°C]
CFD Computational Fluid Dynamics T Ambient temperature [°C]
COP Coefficient of Performance Tdrop Temperature drop [°C]
Csa Hot-summer Mediterranean climate Twb Wet bulb temperature [°C]
d32 Mean Sauter Diameter [µm] UHI Urban Heat Island
Dv0.99 Droplet diameter, below which is the 99%
of the cumulative volume [µm] UTCI Universal Thermal Climate Index [°C]
Dwa Monsoon-influenced hot-summer humid
continental climate WBGT Wet-Bulb Globe Temperature
EER Energy Efficiency Rating ws wind speed [m/s]

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
ET* New Effective Temperature Δ Difference between homogeneous
parameters
FP-growth Frequent Pattern growth algorithm
1. Introduction
The global climatic trends portrayed by the latest IPCC’s statistics [1] leave no doubts on the urge for bold
interventions to ensure heat-safe and liveable cities. Global warming and Urban Heat Island (UHI) effect are
major stressors on economy, energy, comfort and health, especially under extreme weather events and where
social disparity and energy poverty thrive [2]. The interaction between urban microclimate and electric air-
conditioning energy consumption has been investigated not just in hot arid regions [3], but also in sub-tropical
contexts [4], oceanic climates [5] and Mediterranean climates [6] with daunting results.
Science has been moving fast to propose solutions and new criteria for urban planning, boosting a massive
body of sector-specific knowledge and technological progress that demands for periodical wrap-ups.
Consequently, several review papers tackle the development and assessment of UHI countermeasures (see
Table 1 in [7]). Most of these collections compare information on measurement systems and
simulation/forecast models, on facilitators and constraints to UHI escalation, on geographical and historical
patterns, on strategic and remedial planning, on energy-saving and comfort-enhancing metrics. A substantial
quota [8–19] focuses on specific technologies used to increase the albedo of cities (e.g. vegetation, cool/green
roof and pavements, …).
A much less appraised category is that of water-based technologies, namely those that improve the local
climate by means of evaporative processes. Most papers look into the impacts of open water bodies which,
however, are prerogative of few conurbations in the world. A review on the temperature-mitigating effects of
urban wetlands, dating back to 2015 [20], and another systematic collection [21] on various types of urban
blue features (ponds, lakes or rivers) agree on a mean cooling effect of about 2.5 °C during the warmest
months.
Nevertheless, the meta-analysis by Gunawardena et al. [22] suggest that multiple small-scale interventions,
devised in view of dominant wind patterns and synergistic cooling tend to impact more than a solitary larger
feature. Indeed, in a 2017 review paper by Santamouris et al. [23], the authors compared 5 studies on pools,
ponds and open water bodies, three on evaporative wind towers, four on water sprinklers, two on fountains,
four on combinations (both monitored real scale applications and simulation studies included). Water spraying
was found to exert the highest local impact. In the same vein, Teleghani and Berardi [24] recently ran high-
resolution simulations in ENVI-met to check the impacts of different materials and features in a main urban
square in Toronto under summer 2015 heat wave conditions. Water ponds and water sprays emerged as
especially promising mitigation strategies (Physiological Equivalent Temperature - PET - reduced by 3.6 °C).
In view of its envisaged potential, this paper aims at depicting a comprehensive literature review on the use
of water mist spraying for outdoor cooling. The arena of existing research trends and evolutions is here
perused, largely expanding the collection of projects reviewed by Nouri et al. [25].
These solutions are being implemented widely in the world yet mostly by private parties (in pubs and
restaurants pergolas, balconies, gardens …) and respecting no well-grounded procedures. The lack of cohesive
design guidelines on how to maximise the cooling potential in view of the local specificities and climatic
frames is hindering any effective and long-standing intervention. It is not a trivial point, as, despite the dim
spotlight, water spraying multifariously contributes to the enhancement of outdoor well-being and health: it
cools the air, attenuates solar radiation, scavenges dust and repels insects. Most of these positive impacts on
the environment are not fully unveiled [23,25], partly because of the early integration of misting systems into
urban planning, partly because most of the existing information is simulation-based (and given the complexity
of the heat transfer processes, the accuracy of simulation algorithms often calls for further verification) and
partly because monitoring campaigns on real scale projects are challenging, expensive and unstandardized.
The main objective of this review paper is to organically synthetize the major findings in the field, the main
areas of debate and outstanding research questions to unify all acquired knowledge, gather attention on
unsolved issues and pave the way for a truly informed use.
Firstly, an overview of spray cooling is provided along with a general discussion on the governing physical
processes. Secondly, inclusion criteria for the review process are presented. Thirdly, geographical and climatic
patterns are discussed, as well as the technological options. Finally, each study is thoroughly described and

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
comparatively discussed according to 1) the investigational method (purely experimental studies, purely
numerical studies and those combining field tests with simulations) to look for methodological trends and
potential discordances, 2) the results in terms of cooling, humidification and comfort, also in relation to the
adopted performance metrics 3) the design novelty. General remarks are conclusively drawn.
2. Spray cooling
Spray cooling applies to a variety of fields such as metallurgy [26], livestock science [27], agriculture [28],
safety sciences [29]… It is actually from these sectors that most of the insight on the underlying physics was
collected over the years.
The implementation of water spray features in the built environment, to create pedestrian cool spots, was
boosted during international exhibitions in cities like Osaka (1970), Seville (1992), Aichi (2005), and
Shanghai (2010). Despite customarily utilized for aesthetic and sculptural purposes, water spraying
progressively drew attention as an effective thermal stress mitigator, as blatantly testified by projects like the
TVK urban installation in Place de la Republique [30], the revitalization project for the city of Athens [31]
and the microclimatic study in Khan Antoun Bey Square [32].
Direct evaporative cooling stems from the 17th century Japanese practice of “Uchimizu”, which consists in
splashing rainwater/grey water on the heated asphalt to cool the surrounding air down. Anyway, fine water
spraying (termed micro-mist or dry-mist), in place of surface watering or sprinkling, guarantees total or almost
total evaporation immediately after the injection, with minimum water usage and wetting risk at pedestrian
level. Indeed, smaller droplets i) spend longer time being airborne due to the frictional forces, and thus
evaporate faster [33], ii) tend to rebound rather than to break, without adhering [34]. Furthermore, sprayed
mist expels dust and pollen, repels insects and attenuates solar radiation in the long wave ranges of near
infrared and mid infrared [35]: fine droplets (tens of microns or less) are especially effective against harmful
UV radiation. Such a fine pulverization is typically attained through atomization nozzles at high water
pressures or ultrasonic methods [36].
These systems have an edge over water bodies or surfaces, too, since evaporation can still efficiently cool the
air down when air relative humidity (henceforth named RH) is relatively high [37] which explains the interest
and the application of such a technology in humid countries as well.
The process of spray cooling is similar to that of adiabatic humidification (ideally isenthalpic), thus the wet
bulb depression, ΔTwb (difference between dry-bulb and wet-bulb temperature) represents the theoretical
limit [38]. Indeed, heat removal by misting systems largely relies on phase change and only marginally, yet
additionally, on the convective heat transfer of the fluid in motion (unless mated with a fan). Heat and mass
transfer mechanisms are well defined in the studies by Kachhwaha [39,40] and Sureshkumar [41–43]. With a
latent heat of evaporation of 226 kJ/kg, and considering the range of water specific heat and density values as
the temperature (henceforth named T) varies from 26 °C (cooling setpoint according to ASHRAE [44]) to 45
°C (representative of heat wave conditions) under atmospheric pressure, we can assume that 1 kg of
evaporating water is capable of cooling 200 m3of air by 1 °C. Always to give a measure of the orders of
magnitude at stake, a 20 µm radius droplet would take about 2 s to evaporate and about 4 s to produce a-5 °C
decrement in a hot and dry environment (e.g. Twb of 10 °C) [45].
Indeed, the efficiency of this technology not only depends on its design, but also on the climatic context all
around: the water vapour pressure in the air determines the capacity to include additional moisture, whereas
wind speed (henceforth named ws) and gusts dictate the severity of dilution.
Generally speaking, the cooling effect is more apparent and easily measurable in still air conditions and close
to leeward zones, especially in case of micro-nebulization, given the ease with which pulverized water gets
entrained into the airflow. Nozzle’s typology and injection layout should be accurately selected to generate
and preserve droplets at a size that significantly contributes to the temperature drop (henceforth named Tdrop)
with no unpleasant wetting [46].
Against this backdrop, many questions are currently subject of international research. In this review a synthesis
of the most representative studies is presented to disclose which countries and climatic zones are most active
on the topic, what are the typical experimental and numerical approaches, what are the reported performances
and how do they relate to the meteorological context.
3. Review methods

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
To systematically review the existing knowledge on spray cooling, the approach of Pickering and Byrne [47]
was adopted: thereby, in addition to statistics and patterns, data were screened not just to outline theoretical
and methodological trends and gaps, but also to detect geographic biases and climate dependencies. Four main
databases were consulted (ScienceDirect, Scopus, Google Scholar and Web of Science), according to the
taxonomy portrayed in Fig. 1.
Publications were selected according to the following inclusion criteria:
written in English, to retain only manuscripts of international impact;
covering the last 20 years (1999-2019) to account for the most recent developments. Antecedent
studies were mentioned when disclosing theoretical or applicative turning points, but were not included
in the analysis;
specifically addressing outdoor air cooling by mist spray systems. Results borrowed from other
applicative fields (such as fire suppression or irrigation), or limited to indoor effects (e.g. roof spray
cooling studies) or to a general characterization of the spray process with no mention of the achieved
Tdrop, were excluded as hardly transferrable to climatic mitigation;
non peer-reviewed manuscripts were included when rich in innovative contents, to acknowledge
emerging trends.
Affiliations of main authors were used to determine the origin of publications, when no specific locations were
provided for on-site monitoring and/or numerical simulation.
Fig.1. Lemmas used to scan and skim the databases.
4. Synthesis of evidence
In Fig. 2 the number of on-topic publications in the last 20 years is presented. Colour differentiates numerical
from experimental studies to verify whether any methodological dominance occurred as a result of either more
efficient components/constructive techniques or computational power. Relevant events (e.g. Expos) were
additionally reported to catch causal links. Apparently, the interest for outdoor spray cooling over the years
fluctuated with no repetitive/monotonic patterns, but stably persisted since 2007. Experimentation started with
a 2-year delay. Before 2007, the interest was merely architectural, despite solid theoretical basis on evaporative
cooling had been laid during the previous decade. The inauguration of the Drymist technology (finely
pulverized water), during Aichi Expo in 2005, was crucial for ushering in a new wave of applicative domains.
In 2015, much more on-topic literature was published, possibly to gather data and evidence to support urban
development plans.

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
Fig.2. Number and typology of studies related to evaporative cooling by water spraying in the last 20 years.
82 % of the papers quantified the Tdrop and 75 % determined or discussed on the RH gain. The efficacy of
the cooling process was computed in 50 % of the selected manuscripts adopting a variety of indicators,
whereas the comfort benefit was estimated in 46 % of them via questionnaires and/or specific metrics. Cost
analysis was reported in the 18 % of the publications, usually to support a proposed technological variation.
Only four papers mentioned the inertial effects due to secondary evaporation, by attesting persistent and even
higher cooling after injections’ shut-off. More than 60 % of the reviewed literature included parametric
analyses, notably: 7 papers investigated the impact of different initial conditions (in terms of temperature T,
RH or Twb), 4 papers that of droplets size and/or height of the injection and/or airflow rate (representative of
wind speed in outdoor environments), 3 studies looked at pressure and/or water flow rate, 2 at water
temperature and/or pump regulation, while number/typology of the nozzles or solar radiation were only
considered by 1 paper each. Some studies considered specific parameterizations relevant to the
technology/methodology in object, such as solar panel loads and tilting [48], fan oscillation and tilt angle [34],
numerical settings (turbulence model, number of streams, spray angle) [49].
The investigated settings could be open, semi-enclosed and indoor spaces. Semi-enclosed environments were
investigated as they represent the potential best fitting with mist cooling systems (sheltered from scattering
winds and receiving fresh air income through the openings to the outdoors with minimum risk of over-
moistening and degraded performance). Open areas, with wind breaking provisions or exposed to a
dominantly light summer breeze, may still measurably benefit from misting systems even in relatively humid
climates. Indoor environments, especially large air volume workplaces, are also well suited, since their
conditioning via standard devices is extremely expensive. In this case, the permissible spray duration should
be accurately set in function of the desirable RH. Overall, 10 papers focused on semi-enclosed spaces (large
atria, courtyards…), 19 on open spaces either unobstructed or shaded (under pergola, arbor…), 3 on indoor-
like environments (rooms or climatic chambers) and 2 left it unspecified.
The considered papers were published in 23 journals: Building and Environment (5 papers), Applied Energy
(2), Applied Thermal Engineering (2) and Journal of Heat Island Institute International (2) were the most
common editorials. This demonstrates that urban planning, environmental sciences and energy were the
dominant disciplines. Frequent taxonomy denotes that the most recurrent lemmas were “mist”, “cooling”,
“water”, “spray”, “urban”, “evaporation”, “outdoor” and “environment” in the order. See Table A1 in
Appendix for all the details.
4.1 Geographical patterns
28 manuscripts from 12 countries were included. Based on authors’ affiliations and/or experimental sites,
papers were predominantly produced in Japan (39 %). It is clearly no coincidence that most of the research
effort on this technology still comes from the cradle of Uchimizu. Especially overhead hydraulic misting

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
nozzles, producing droplets with average diameter of below 100 µm, are becoming increasingly popular in
Japan to cool urban spaces [46], yielding Tdrops in the order of 2–3 °C on hot summer days [45]. The leading
city is Osaka (6 publications), also due to the Expo installations, followed by Tokyo (3). Research and
application of evaporative cooling have been very fruitful in China too (14 %), notably in Shanghai (2).
Nevertheless, little was and is disclosed worldwide because most of published works maintains the original
language (Japanese and Chinese) and requires authorizations.
Other than that, the interest for spray cooling was not prerogative of Asian countries, since 1) the Netherlands
(11 %) laid the criteria for a solid numerical methodology thanks to the work of Montazeri et al., 2) Italy (7
%) pioneered the introduction of advanced control logic, A.I. techniques and statistical procedures thanks to
the experimental campaigns by Ulpiani et al., 3) Spain and UK (4 % each) contributed to data collection and
mathematical modelling. Middle Asia countries like Saudi Arabia, India, Turkey and Iraq mostly published
on technological upgrades and self-sustained configurations. In these countries, especially India, mist spraying
has been used since 1950 to cool residential complexes, restaurants, amusement parks, playgrounds, sports
stadium, hotels ... [50].
Conversely, the application of spray systems as microclimate mitigator remained unattended by American,
African and Oceania countries. No study was conducted in the Southern Hemisphere.
In a climate-based analysis, the geographical bias is further stressed. Sticking to the updated Köppen-Geiger
classification by Peel et al. [51], spray cooling finds a fertile ground for investigation and implementation in
warm, temperate and humid climates, notably the Humid subtropical (Cfa) and the Hot-summer Mediterranean
(Csa), that together cover the 75 % of the reviewed literature. Notably, more than half of the manuscripts
originated in Cfa climatic contexts.
Most of the major cities in Japan, eastern China, south-eastern Asia, south-eastern Australia and America have
similar subtropical climates. As Cfa areas experience no annual dry seasons, possibly and promisingly, when
transferring this technology to Csa contexts, the cooling could be pushed further without exacerbating
acceptable humidity levels. Indeed, the highest Tdrops are reported in Csa-georeferenced studies (see Fig. 3).
The interest, from Dutch or English cities, featuring Cfb climate (14 %), with mild winters and moderately
warm summers, derives from the increasing frequency of weather extremes, like heat waves: countries less
accustomed to thermal stress are also less armoured against overheating, which explains why some research
lines originated there, to support novel urban strategies.
Few instances arose from semi-arid and tropical climates, namely Bwh (hot desert climate), Af (tropical
rainforest climate), Aw (tropical wet and dry orsavanna climate) and Dwa (monsoon-influenced hot-summer
humid continental climate).

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
Fig.3. Geographic distribution of spray cooling studies: in the upper map, colour denotes the climate and size
the number of instances. In the lower close-ups, shape defines the type of investigation and size the maximum
reported Tdrop.
4.2 Injection characteristics
Nozzle design determines the average size of sprayed droplets. Instant flash-evaporation is sought-after feature
in mist cooling, thus smaller orifices mated with high pressure pumps are generally preferred. Nevertheless,
cost and effectiveness should be balanced. For instance, in [52] the authors stress that over-40 µm diameters
strongly weaken the cooling potential, which becomes negligible around 100 µm. On the other hand, below
this threshold, the cost premium could be inconvenient.
In the reviewed literature, droplets size scattered from 5 to 420 µm, expressed as Sauter mean diameter (d32),
unless unspecified. In numerical studies, the Rosin-Rammler particle-size distribution was the most common
metric.
The pump input power, when declared, ranged between 30 to 1500 W, while the developed pressure varied
between 0.3 and 7 MPa. High-pressure systems were generally preferred (50 % more studies). In terms of
nozzle typology, few papers reported the geometrical details. Hollow-cone, pressure swirl atomizers were the
most popular. The discharge diameter ranged between 0.16 and 4 mm. Pneumatic nozzles featured more rarely
compared to hydraulic nozzles, since the relatively smaller amount of sprayed water limited the available
evaporative cooling [45]. Also, hydraulic spray nozzles usually consume less than air-driven solutions, but at
the expense of larger droplets. The outlet velocity, in the few (5) papers where it is mentioned (usually
measured), varied between 7 and 22m/s. Commonly, few litres per hour were processed by a single nozzle.
Water temperature was investigated in almost the entire liquid range by Farnham et al. [53] proving to have a
minor impact on the cooling performance. On average, water stayed slightly above 30 °C.

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
In terms of configuration, 7 studies focused on single-nozzle systems to characterize the process. The
remainder considered up to 136 nozzles. Their arrangement was predominantly vertical (misting lines,
matrices or cross-shaped assemblies), pointing downwards (about 50 % of the studies). Other than that, the
setup could be horizontal (6) or rotational (4). The latter was attained by mounting the nozzles around the
circular header of a fan. Less frequent configurations included spirals of nozzles wrapped around vertical
columns (2010 Shanghai Expo) and nozzle lines over the roof of a moving bus [54]. Limited to vertical
arrangements, the height above the ground varied between 2 and 25 m. On average, overhead systems were
mounted at 2.7 m height. Spacing between nozzles varied between 0.4 and 3 m, but its role was never
scrutinized.
Most of the proposed systems worked continuously, 6 papers mention on/off modes (timing variable from few
seconds to 7 minutes, some emphasizing the off phase some others the on phase) and only one [6] fuzzy logic.
Temporizations allowed to better handle the RH gain and keep it under a ceiling, whereas more advanced
algorithms could track specific comfort targets.
All details are carefully collected in Tables A2 and A3 (Appendix). Statistics were conducted on declared or
computable data (i.e. not considering total flow rates when the number of nozzles was ungiven). Apparent
typos were excluded.
5 Contents overview
In this section, each reviewed publication is accurately discussed according to the main target: parametric
analysis, comfort analysis and proposal of innovative solutions.
Experimental investigations took place in summertime (but for [48]), usually covering the entire day. In few
cases (4) the observation was limited to a specific time window (night time in [46], 11am-3pm in [45], 10am-
4pm in [34] and 1:30-2:40pm in [52]). The duration of the measurements varied between few hours to months,
mostly depending on the typology and temporality of the installation. In nearly 60 % of the monitoring
campaigns, T and RH probes were variously arranged to determine the Tdrop and/or map horizontal and
vertical distributions in and around the misted area. In 4 studies, RH was disregarded and thus only T probes
were in place. Other commonly measured environmental parameters were wind speed (42 %), solar radiation
(26 %) and mean radiant temperature (21 %, prerogative of comfort-oriented studies). In 3 studies, thermal
imagery was used to verify surface temperatures, including skin temperature, whereas Phased Doppler Particle
analysis was conducted in 2 experiments to determine the size distribution. Finally, only 2 papers report
measured water temperature. The sampling rate was highly variable, from seconds to tens of minutes. The
experimental settings were predominantly shaded (58 %), either totally or partly, while 26 % were vegetated.
No study, to date, quantitatively differentiates water mist performance with and without solar shadings or local
greenery, at equal boundary conditions. In [55], Ulpiani et al. monitored the same misting system in two urban
settings, one with no vegetation or shading, the other with massive arboreal shelter, measuring up to 2.3 °C
extra cooling in the first location. Anyway, direct comparison cannot be fully trusted, as experiments took
place over different days and geographic areas.
Along with experimental experience, computational fluid dynamic (CFD) models have been extensively
adopted to simulate the droplet-air interaction during an injection. About half of the 14 papers including a
numerical part, considered wind and solar influences, the other half disregarded any atmospheric boundary
layer.
ANSYS Fluent was, by far, the most popular software to assess the thermal regimes of misted areas at fine
spatial resolutions and short time steps (70 % of the reviewed studies, considering those explicitly mentioning
the name of the software). Particles, including water droplets, were handled as Lagrangian points (discrete
phase) traveling through the Eulerian grid elements representing the airflow (continuous phase). The control
volume varied between less than 1m3to over a billion m3[56]. Nevertheless, ENVI-met is expected to gain
popularity, since the introduction of the blue mitigators simulator (water mist cooling, wet surfaces), although
the spatial resolution is much reduced.
Only a limited number of papers (5) combined measurements and simulations.
Fig. 4 reports the max Tdrop and RH gain in experimental and numerical studies, comparatively. See the
following paragraphs and the Appendix for a comprehensive description of the methodologies (Tables A4,
A5) and of the main results in terms of cooling, humidification and comfort (Tables A6, A7). Note that only
explicit information was reported.

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
The variety of approaches and metrics sounds a warning bell on the need for a higher level of standardization
to get comparable results.
Fig.4. Tdrop and RH gain according to experimental and numerical studies (absolute maxima).
5.1 Parametric identification of key players
Most of the reviewed literature (57 %) aims at understanding and weighting the role of a variety of influential
parameters: initial thermohygrometric conditions (23 %) and wind/airflow (13 %) were primarily investigated
to delimit climatic suitability, droplets size (13 %) and pump pressure (10 %) were explored to delimit the
technologic arena, while the distance between injection and target object (13 %) was examined to delineate
substantiated design guidelines. Other factors, such as water temperature (7 %), were found to play a minor
role and thus examined to a lesser extent.
In 2007, Barrow and Pope [33] focused on water mist as a counter to overheating in railway tunnels.
Persistence time of airborne droplets affected the severity and frequency of impingement on the walls, to the
detriment of the cooling effectiveness. Thus, the authors investigated the life-time of freely falling droplets of
different sizes by applying the transient form of the First Law of Thermodynamics to the moving spherical
control volume of the droplet, while travelling in moist air (28 °C, 40 %, no ventilation). Relevant mass,
momentum and energy-conservation equations were coded in Q-basic, using a time-marching numerical
procedure. The results revealed that droplets rapidly reached a terminal T very close to the thermodynamic
Twb. Indeed, the evaporation time ranged between 0.66 s and 28 s, while the travelled distance increased from
0.006 m to 15.7 m as the droplet diameter was varied between 25 µm and 200 µm. Life-time was expected to
decline in presence of ventilation.
In 2008, simulations by Yoon and Yamada [57] again on droplets size, suggested that in case of high-pressure
nozzles, larger dimensions provoke longer persistence and greater penetration, but negligible impacts on
Tdrop. Three different d32 were examined (16.9 µm, 20.8 µm, and 32.6 µm): the Tdrop was close to 1.5 °C
either case. The same authors, and again by numerical approach, tackled the role of outdoor air conditions and
potential control logics [58]. They simulated the performance of a misting system in a semi-outdoor space,
dominated by light breeze (0.1 m/s) and weak solar radiation (363 W/m2). Three injections at 3.5m above the
ground were tested in as much T/RH scenarios (30℃/80 %, 30℃/60 % and 34℃/60 %). The maximum Tdrop
at a height of 1.5 m was 2.5 °C, recorded in the 30℃/60 % scenario. Regardless of T, higher RH depreciated
the evaporation rate and increased the mass of unevaporated droplets in the computational domain. The mist
worked best with RH <70 % and T in the 30-34 °C range. Accordingly, the authors proposed the following
control strategies, based on RH thresholds: i) RH <75 % to avoid wetting, ii) RH <70 % to avoid mist
persistence within 1.5 m of the ground. The occurrence of favourable conditions was computed for seven

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
Japanese cities, in consideration of the 2 control strategies demonstrating that, in very humid regions such as
Naha, control schemes should be mandatory to avert discomfort.
In 2009, an experimental and numerical investigation in open outdoors [59] was conducted to check the role
of wind direction. The maximum measured Tdrop was of 6.7 ℃, whereas numerical simulation suggested a 4
°C extra decrement in the downwind case compared to the 3-5 °C Tdrop when injection and wind were counter
to each other. The following year, Wang et al. [60] numerically simulated a two-phase flow injection within
a 600m3control volume. They varied both the inlet velocity and the typology of space (open versus semi-
enclosed) and found that, by increasing the relative Reynolds number, the cooled area expanded due to the
increase of the spray penetration length, and this positively affected the level of comfort in the pedestrian
zone. The Tdrop in the semi-outdoor environment was very evenly distributed in the 3-5 ℃ range, with
humidity gain of about 1-6 %. Conversely, in the open space, the Tdrop ranged between 1 and 2 ℃. The
authors concluded that the presence and the height of the roof strongly influenced magnitude and homogeneity
of the cooling.
In 2011, Huang et al. experimentally investigated the effectiveness of high-pressure spray cooling in a
residential pavilion first and then in an outdoor waiting area during 2010 Shanghai Expo [61]. In the first
location, when T surpassed 30 °C and RH stayed below 70 %, a 1-2 °C Tdrop was observed, whereas with T
and RH around 35 °C and 45 % respectively, the reduction rose up to 5-7 °C. In the EXPO site, with T in the
34-40 °C range and RH oscillating between 32 and 55 %, the cooling was in the order of 6-12 °C within 1 m
of the spray column, of 1.5-4 °C in a 3 m radius, and of 0.5-1.5 °C in a 7 m radius. The authors also computed
the cooling efficiency (defined as the ratio of the actual Tdrop to the ideal ΔTwb). It exceeded 65 % in the
first run and 90 % in the second. The COP (ratio of total cooling power over the pump power input) was over
30 when water was pushed at more than 5.5 MPa. Looking at comfort enhancement, the WBGT (calculated
from the regression curve by L. Zhang et al. [15]) in the sprayed area could be reduced by up to 5.2 °C.
In the following years, the numerical approach was largely pursued by Montazeri et al. [49,56,62]. In their
latest publication [56], they applied high-resolution Computational Fluid Dynamics based on the 3D unsteady
Reynolds-Averaged Navier-Stokes equations to systematically characterize spray cooling in a Dutch courtyard
under heat wave conditions. A row of 15, 0.5 m equally spaced, hollow-cone nozzles was simulated and
parameterized in terms of water flow rate (total 9 l/min, progressively halved) and height above the ground
(3, 4 and 5 m). Validation was based on wind-tunnel measurements and satellite imagery data (90 sampling
points for 42 specific times). The maximum T and UTCI reductions at pedestrian height (7 °C and 5 °C
respectively) corresponded to the highest flow rate, with Tdrop of -2 °C up to a distance of 8 m. No residual
cooling was observed when the flow rate was halved twice. Two years earlier [62], the same authors ran
another systematic parametric analysis focusing on the impact of physical parameters (inlet air T, humidity
ratio, velocity, inlet water T and size distribution). For each, five discrete values were simulated: the impacts
on air T [°C], humidity ratio [g/kg], sensible cooling [kW] and UTCI [°C] were parametrically displayed.
ASHRAE cooling efficiency (CE) was additionally computed. It was observed that 1) injecting droplets with
a water T sensibly lower than the air T boosted the cooling (+40 % for a 8 °C difference), 2) lower vapor in
the inlet air and lower air-droplet relative velocity amplified both T and UTCI reductions, 3) small droplets
with a wide spread triggered complete evaporation with no wetting. The CFD model was fully optimized [49]
via grid-sensitivity analysis and validated on the wind-tunnel measurements by Sureshkumar [41]. No
significant sensitivity to mesh grids or different turbulence model was observed. A minimum of 100 streams
proved to be sufficient for stable CFD results.
In the same period, Farnham et al. [46] monitored three hydraulic nozzles (d32 from 41 to 79 µm) working at
two operating pressures (0.7 and 5.5 MPa) and flow rates ranging from 38 to 660 ml/min inside a large atrium.
They tweaked the height from 6 to 25 m above the ground and mapped the vertical gradient to determine the
altitude at which evaporative cooling caused no wetting in still air conditions. The authors reported maximum
T and ET* drops of about 2 °C and signalled the difficulties of measuring the cooling effect inside the mist
spray, because of sensors wetting: probes tended to reach the wet bulb T in 1-10 minutes.
A similar parameterization was carried out by Mahmoud [63] in 2014 on a low pressure system. Measurements
at pedestrian height showed that 1) at the highest tested water flow rate (1.2 l/min), the maximum dry bulb
Tdrop was 9.4˚C, 2) at the lowest tested height above the ground (2.25 m), keeping the above flow rate, the
Tdrop peaked at 11 °C, 3) the inlet water T had a negligible impact (only 15.9 % difference for a 10 °C

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
variation). The effectiveness hit its maximum (62.5 %, with respect to 21 °C Twb) at 2 m from the nozzles,
assuming the highest flow rate, the lowest height and the coldest water inlet.
In 2015 the minor impact of water T was confirmed by Farnham et al. [53], who studied the effect of water T
on flow rate, droplet diameter, evaporation rate and cooling effect of water mist by both developing a CFD
model and performing measurements on a test bed. The water temperature was varied between 8 and 92 °C
by defining the saturation vapour pressure and binary diffusivity of the air-water interface. The water flow
rate decreased with higher T according to a simple linear curve. Also, a smaller d32 was observed, similar to
the trend predicted by Lefebvre [64]. The authors concluded that a pedestrian in an evaporative mist cool
would not perceive sprayed hot water unless by touching the nozzles or headers. Thus, hot water may be
deliberately pumped through the system, to avert wetting of the target in tight spaces and as a legionella
countermeasure.
The latest study on water temperature dates back to 2018 [65]. The authors sought to optimize the cooling
impact of dry mist systems by tweaking ΔTwb, pressure and water T in a climatic chamber to simulate
different climates, from temperate (Tokio-inspired) to tropical ones (Singapore-inspired). Double-flow
pneumatic spray with the same pressure for air and water was adopted. The authors found that 1) ΔTwb was
a good estimator of the cooling performance with Tdrop and ΔTwb linked by an inverse “S”-shaped curve 2)
pressure could be tuned to achieve optimal cooling at various heights on a fixed distance downstream, 3) water
T had negligible impact (in line with [53]). Indeed, the Tdrop ranged between 2.9 and 7.5 °C by varying
ΔTwb, between 4.1 and 4.3 °C by changing pressure and levelled out at 4.9 °C by tweaking the water T. The
authors also computed the cooling efficacy CE according to the ASHRAE handbook [66] and the sensible
heat ratio (between sensible and total heat load) demonstrating that dry mist systems would be a favorable
choice in tropical climates. Results need further confirmation and should be trusted for large indoor spaces,
not prone to oversaturation, or semi-enclosed environments where solar radiation and wind are not significant.
On another subject and again in 2015, Kim et al. [67] analyzed the cooling provided in open outdoors by a
cross-shaped system of overhead blast sprayers, at three different water flow rates. They investigated the
mutual link between solar radiation and evaporation by performing Duncan multiple range test and one-way
ANOVA, concluding that statistically higher Tdrop occurred at higher ambient T and irradiation. Furthermore,
by gradually doubling the flow rate from 0.16 to 0.48 l/min the recorded Tdrop increased from 3.1 °C to 3.5
°C and finally to 4.4 °C in terms of daily average. The absolute maximum (around 6 °C) occurred at peak
hours (12-2 pm). In terms of spatial spread, Tdrops of over 1℃ could be still measured at 7 m from the misted
area with the highest water flow rate. The authors suggested to adjust the cooling by increasing mist exposure
time or the number of active nozzles.
Later in 2016, Farnham et al. extended their research on mist cooling to large indoor workspaces (such as
factories) [34]. A mist fan (32 nozzles, 6 MPa pressure, 86 l/h) was characterized by mapping T and RH in
more than 10 points along the centreline and the action edges. Furthermore, the system was tested both
oscillating (60 ° in a 50 s period) and fixed, both at 0 ° and -4 ° tilt angle. The best performance was achieved
when the spray was fixed, with maximum average of 1.7-3.7 °C and peak of 3.1-4.8 °C in the horizontal and
tilted cases respectively. Additionally, by developing a silicone skin-analogue (heated to near-body T and fit
with heat flux and T sensors) the authors demonstrated that the misted fan removed 17 W/m2more heat
compared to the fan alone on time-averaged basis (peak close to +26 W/m2). By adapting the ASHRAE 55-
2013 comfort model to include the misting term, it could be concluded that a 2 m/s air jet combined with a 20
W/m2mist cooling doubles the effect of the no-mist case, with negligible humidification penalty.
Finally, in 2017, Huang et al. [52] developed a mathematical model to calculate T and RH in the misted area
as a function of nozzle pressure, droplet diameters, environment airflow rate, initial T and RH, and distance
from the nozzle. The results were validated through a day-long experimental campaign during Shanghai Expo.
The cooling incremented with: 1) higher pressure, but not proportionally, 2) smaller droplets, 3) lower ws
(Tdrop increased by 2.5 °C for a 2 m/s difference), 4) lower RH (despite accentuated humidification). Overall,
the authors suggested to 1) use a 3 MPa pressure to pulverize the water down to 40 µm diameter, 2) turn off
the system in case of T <30 °C, RH >70 % and ws >3 m/s. The provision of windscreen features was
recommended.
5.2 Characterization of comfort implications

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
In addition to parameterization, some manuscripts focused on how and how much mist cooling could improve
pedestrian’s comfort.
In 2009, Ishii et al. [68] tested the dry-mist system. Considering a quasi-isenthalpic process, they projected i)
a 5 % humidity premium for each 1 °C Tdrop, ii) a maximum cooling of about 2 °C in the Japanese climate
and iii) a 10 % ACEC reduction. These projections were experimentally confirmed in a railway station in
Yokohama, where 30 nozzles were installed. In the day-long summertime experiment, the cooling reached
1.63 °C between 9 am and 1 pm and 1.90 °C over peak hours (1–3 pm). The on off temporization allowed 2-
min lasting injections: authors affirmed that longer operation could have led to over 2 °C mean cooling
throughout the day. Comfort questionnaires were additionally distributed: 80 % of respondents reported
comfortable conditions. The authors considered automatic control too: the injection was started if T surpassed
28 °C, RH didn’t exceed 70 % and ws stayed below 3 m/s with no rain.
The comfort benefit of on/off operation was later (2016) remarked by Domínguez et al.’s [69]. They measured
the environmental impact of 128 nozzles installed in the open areas of the Seville Expo (1992). The Tdrop
reached 5 °C at peak hours and touched 10 °C on a 42 °C day. RH was kept under a 65 % ceiling thanks to
the intermittent control. The comfort premium was quantified by computing the sweating ratio, according to
Fanger’s theory [70]: it stayed below 90 g/h.
In 2010, Wong and Chong [71] implemented misting fans in two food centres and a misting line in a coffee
outlet. Over 4 days, 40 men and 40 women responded to the comfort questionnaires (on thermal sensation,
overall comfort, humidity sensation and preference, airflow sensation and preference) repeatedly in specific
time windows, while sitting at the dining areas. Air T, RH, ws and globe T were measured at sitting height (1
m), whereas meteorological data were collected outside the influenced areas. Additionally, air sampling was
conducted to check biological proliferation harbored by the mist. The T drop reached 1.6 °C, yet at the expense
of over 10 % RH gain and consistently greater biological proliferation (up to 10 times larger bacteria colony
units, minor impact on mould and yeast growth). Questionnaires substantiated that: 1) the mist features were
capable of producing a 1-vote shift to fresher feedbacks, both in terms of thermal sensation and overall
comfort, 2) humidity perception was fairly unaltered, although potential discomfort could have occurred if the
interviewees were to undertake higher activity levels, 3) systematic SET/PET reductions occurred at equal
ET* in the misted area (up to nearly 3.5 °C), 4) thermal neutrality could be guaranteed at a higher outdoor
ET* with the misting fan.
Five year later, in Osaka, Farnham et al. [45] replicated the experience of combining misting lines and fan in
an outdoor space. 4 T and RH loggers, 1 hot-sphere anemometer and a globe thermometer were distributed in
the misted area to compute the PMV and the thermal load. Fast, non-intrusive skin T measurements were
performed by means of an infrared thermometer, designed for medical use, while surface heat flux
measurements were conducted using the same skin-analogue as in [34]. Also, comfort questionnaire were
distributed to 141 participants on hot summer days, to verify thermal sensation, overall comfort, skin
wittedness sensation and wettedness impression. The average thermal sensation changed from +2.7 (hot) to -
1.41 (slightly cool), while comfort sensation changed from -1.41 (uncomfortable) to 1.82 (comfortable)
moving from outside to inside the mist. The humidity increase was perceived, but pleasantly, by 75 % of the
respondents. Skin T sizeably and swiftly decremented when exposed to the misting fan (averaged drop of 1.18
°C at the forearm and 1.08 °C at the head in 10 s). Conversely, no significant change occurred with only the
fan and no mist. The heat flux averaged at 60 W/m2for natural convection (no fan, no mist), 81 W/m2for
forced convection (fan and no mist) and was substantially higher (201 W/m2) for the fan+mist case, at 4 m
distance. No correlation was identified between residence time in the mist and degree of effect and no
verification of the impacts on the core T was included. In a previous study [72], the same authors measured
skin T reductions up to 3 times the expected air T reductions, which explains why people’s satisfaction
drastically improves in a misted environment, even if the cooling is limited (1-2 °C).
The latest comfort-oriented study, dated 2019, is by Ulpiani et al. [55].The authors collected on-field
measurements and comfort questionnaires, for a total of 332 feedbacks in about 11 days. Thestudy took place
in two Italian urban settings featuring Cfa and Csa climates. To uncover frequent patterns, the authors
combined statistical tests and data mining algorithms (FP-growth). Also, the vertical thermohygrometric
profile beneath the mist was analyzed by multiple regression, denoting a Lorentzian distribution, peaking at
about 0.7 m from the nozzles. Reportedly, the maximum T and UTCI drops were 8.2 °C and 7.9 °C
respectively, at the expense of a 7 % average RH gain. While under the mist, 67 % to 90.6 % of respondents

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
reported almost perfect neutrality, and accentuated appreciation for solar radiation, RH and ws. Some design
guidelines were derived, pretty much in accordance with Huang et al. [52]: misting was suggested with
dominant and steady light airflows (below 2 m/s), in significantly sunlit sites and suspended 1.2-1.5 m over
users’ average height.
5.3 Innovative setups and renewable solutions
Along with investigational studies, some publications proposed novel configurations.
In 2012, Narumi et al. compared three types of evaporative cooling techniques (rooftop spraying, veranda
spraying, and outdoor air-conditioning spraying) via both on-site measurements and numerical simulations
[73]. The systems were installed in an apartment house in Osaka. Firstly, the ACEC saving was experimentally
monitored for a single room with respect to the two adjacent ones (not fitted with the water features); secondly,
the results were extended to the simulation model of the entire building. Overall, the cooling decremented by
more than 80 %, notably that of the top floor. The daily average outdoor T drop was 16.4 °C and the emitted
thermal load (urban heat flux) decremented by 60 %, especially due to the outdoor AC spraying since the
subtracted latent heat exceeded the exhaust heat.
An innovative and scarcely investigated implementation of spray cooling was envisaged by Kojima and
Nakashima in [54]. They suggested micro-mist spraying systems mounted on urban means of transport (buses)
to bring thermal relief into dense urban areas, where buildings are clustered together, with no need for large-
scale installations nor water pipelining along the roads: since vehicle exhaust emission discharge is a major
escalator to urban overheating and pollution, on-board mist cooling could alleviate the discomfort directly
from the source. A very preliminary numerical analysis on the surroundings of Sakae Station in Nagoya City,
was proposed, yet with no quantitative estimation of the cooling potential.
Later on, self-sustained configurations started being scientifically investigated, notably solar-assisted
solutions. A first demonstratation by Atieh and Al Shariff [48] took place in Saudi Arabia (Medina) in 2013.
The authors monitored the ambient T according to different timing signals applied to the pump control. The
prototype reached Tdrops of up to 10 °C. Approximately 5 °C extra cooling was provided when the on/off
timing was switched from 20 s/20 s to 60 s/20 s. Nevertheless, the humidity gain rose from 5 % to about 26
%. The energy efficiency rating (EER) touched 128.4, corresponding to a COP of 37.7, but, as the campaign
was conducted under relatively cold conditions (December-January), much higher efficiencies could be
expected in hotter seasons. The return on investment (ROI) for the self-sustained prototype (10 m long misting
system+ solar panel+DC/AC inverter+charge controller, rated at about $ 1500) was nearly two years and half.
It was calculated by comparing the cost of operating an air condition (AC) to cool the equivalent of the misted
area.
Solar-assisted mist cooling was investigated in Turkey too [74]. In this study, dated 2015, nebulization was
used to cool down a 24 m2arbor, adopting PV technology to pressurize the water flow. Two 150 W
monocrystalline solar panels and three 100 Ah 12 V batteries powered up a misting system composed by up
to 28 hexagonal nozzles. T mapping was performed via thermal imagery, according to randomly selected X
and Y coordinates. The arbor T reportedly dropped from 35 °C to 15 °C, with RH gains of 20–25 %. Following
the procedure by Atieh and Al Shariff [48], the authors computed max COP and EER (18.8 and 64.15,
respectively) and ROI, established at 3.7 years. These values were calculated for an average measured T of 35
°C and RH of 40 %. The authors alerted about potential inrush currents induced by on-off operation on the
inverter. For this reason, the system was operated continuously during the experiment.
Later in 2016, Joshi et al. [50] re-proposed to harness solar radiation to power a mist cooling system in India.
The authors provided a very detailed description of connection schemes and dimensioning of each prototype
component, but no data about the actual cooling performance. They concluded that six 60 W standard PV
modules and four 100 Ah, 12 V batteries could suffice for a 140 psi system, regardless of the number of
connected nozzles and considering a maximum lifting height of the submersible pump equal to 2.75 m.
Self-sustained solutions become a more viable path when the energy absorption of the misting system is
rationalized. This is advocated by Ulpiani et al. in [6], where the authors focused on advanced control logics
to calibrate the cooling intensity and mitigate spray cooling’s strong dependency on site’s topography, ground
surfaces, vegetation, built volumes and shade regimes [75]. The authors programmed and field tested a smart
overhead misting system based on approximate reasoning (fuzzy inference) in two Italian urban locations. The
controller managed the pump activation on the basis of i) the offset from the neutral air T, ii) solar radiation,

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
iii) Humidex (H) and iv) cooling power index (PE), lumped indices that quantify the aggregated thermal strain
of T-RH and T-ws respectively. The prototype could reduce the T up to 7.5 °C, in function of PE and ws, as
demonstrated by the R2=0.98 equation the authors extracted via evolutionary algorithms. The logic kept the T
within ±2 °C of perfect neutrality while guaranteeing a mean energy saving of 51.2 %, on average, compared
to intermittent temporization, touching a high of −67.5 %. Ultimately, the authors proposed a preliminary
design estimation for solar-powered assemblies: an endowment of two to three PV panels, plus three 200 Ah,
12 V batteries could power the smart system, essentially half the provision needed for a temporized control.
Also, they declared that a small quota of water, retrieved from public fountains, sufficed to sustain their high-
pressure system, with no need to interrupt the flow of potable water to the citizens.
The schematic representation of all mentioned solar-powered prototypes is proposed in Fig. 5.
Fig.5. Solar-powered misting technologies, proposed in a) Saudi Arabia [48], b) Turkey [74], c) India [50]
and d) Italy [6]. Adapted from sources.

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
6. Discussion
This review paper collects all internationally relevant knowledge on mist cooling, applied or applicable to
open, semi-enclosed and closed urban spaces. Asian countries (Japan and China first in line) dominate the
scientific panorama, even without considering the extensive untranslated literature (beyond the scope of this
document). The interest in spray cooling shows no stable trend over the years, but a strong geographic
dominance in warm, temperate and humid climates. The most successful applications are reported in fairly
dry climates. Nonetheless, recent studies, such as [65], demonstrated dry-mist suitability to tropical climates:
countries like Singapore, that heavily rely on mechanical AC systems to achieve thermal comfort, would
especially benefit from less energy-intensive and refrigerant-free solutions [71]. No strong disparity between
developed and underdeveloped countries could be established, since nebulization is a fairly economic blue
mitigator.
Overall, given the extremely high climate sensitivity, it is hard to guarantee perfect performance replicability
in a different context. Another general criticism is that reported information is often vague and descriptive,
and even when performance parameters are introduced, literature still lacks a cohesive approach. Anyway,
what most hinders the rigorous assessment of spray cooling is that data collection and reporting protocols are
not established yet. To reduce bias and inaccuracy of results, but also to ensure comparability, probes typology
and positioning should be homologated. Besides, adequate countermeasures against probes trend to the wet
bulb condition should be encouraged to avoid overestimating the temperature reduction [45]. High temporal
resolution and fine-grained spatial mapping are advisable, but require extensive sensor networks, which are
unlikely to be the standard. Furthermore, verifying different exposure times and configurational variations
over a wide range of environmental conditions, mist characteristics and human samples is challenging.
Bespoke equipment might be necessary. The paraffin human-skin analogue, proposed by Farnham et al.
[45,72] could overcome most limits, but simulation of physiological processes like sweating should be
included. Thermal imaging should be very carefully deployed: although an attractive solution, it is sensitive
to emissivity settings, in addition to mist, irradiative and on-surface-adhering water interferences, especially
in open settings. In [72] its use led to systematic overestimation of the skin cooling compared to thermocouple
readings. To sum up, Table A4 demonstrates the variety of monitoring equipment, layouts, observation periods
and sampling rates that hampers any accurate benchmarking.
Notwithstanding the need for field-collected evidence, many studies outline that, as fine water mist scatter
within a turbulent airflow, strongly perturbed by natural wind (velocity, direction, gusts, buoyancy…), the T
distribution would unlikely be axial symmetric and could profoundly differ along horizontal and vertical
planes [76]. Therefore, simulations by CFD code should be developed alongside. This review (see Table A5)
also provides a list of inputs that should be declared in numerical studies for the sake of comparison. Indeed,
the arena of technological alternatives in terms of nozzle typology, discharge diameters and cone angles,
configurations and layouts is extensive, but rarely set out in detail (see Tables A2 and A3).
In terms of parametric analysis, the roles of nozzle density, configuration, temporization and other control
logics are still poorly investigated. For instance, on/off timing varied from seconds to tens of minutes, with
no justification for a specific setting, nor a tuning procedure. This would be an interesting knot to untangle to
effectively lower the ambient T without imperilling hygrometric comfort. Appropriate water pressure, nozzle
type, and exposure time have been identified as pivotal players, notably when approaching thermal comfort
in areas that witness high humidity levels [77], but no all-purpose and all-climate control strategy has been
proposed yet. Ulpiani et al. demonstrated that using cooling power indices (Humidex) among the governing
variables of a smart logic [6] is a promising path, yet only verified in Mediterranean climates at present.
Wind control is another grey area. Misting systems in urban open spaces are especially sensitive to the airflow
rate: the droplets may be blown sideways, evaporating before reaching the intended target. Forced convection
(misting fans) could help overcoming the scattering effect of moderate breezes by directing the injections.
Nevertheless forced mists favour much localized cooling on the body, which may be detrimental in terms of
comfort.
In the reviewed literature, the average maximum Tdrop was 8 °C, which qualify mist cooling as a tremendous
asset against urban overheating at local scale (see Table A6). Anyhow, clear statement of where and when the
maximum cooling was recorded is omitted in some cases. Although different metrics were used, the benefit
is apparent in terms of comfort, too (see Table A7), particularly when skin temperature is analysed [34,45].
SET* [58,73] and UTCI [55,56,62] were the most frequent indices to quantify the improvement. The

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
maximum SET* reduction is reported by Ulpiani et al. in Italy (Ancona, Cfa conditions) with an average of -
3.5 °C [55]. The maximum UTCI reduction is reported by Montazeri for the city of Rotterdam [62] (peak of -
8.7 °C). Many secondary impacts (dust scavenging and pollutants removal [78], erythema reduction [35] and
solar attenuation [79]) were out of the scope of the present review. Anyway, they have been scarcely
investigated by experimental means, hence they remain underrepresented in research as well as in policy
making. On the other side, the study by Wong and Chong [71], here discussed, demonstrates that nebulization
may be responsible for a more conducive environment for fungi and bacteria proliferation (in addition to the
renowned legionella risk, possibly causing fatal pneumonia when inhaled). Indeed, no paper investigates water
quality as a key-player for the long term effectiveness and the risk of bacterial growth. Besides, as legionella
risk makes water storage an ethically unviable solution, recovery solutions for discharged water, notably with
on/off systems in place, should be considered. With high-pressure systems the pump may recirculate
unsprayed water if the waste heat is sufficient to raise the water T in the tank to over 65 °C and instantly kill
legionella [80]. Additional disinfection measures should be still provided.
In terms of self-sustained solutions, solar-powered systems might be suitable, notably for those countries in
the equatorial sun belt, receiving abundant radiant energy from the sun and experiencing hot and dry weather.
Cost estimation of the potential upgrade would be especially useful to developing countries with limited
resources, where demographic growth, urbanization and/or severe weather extremes (heat waves, bushfires,
droughts …) are expected to climb higher in the next future.
Countries like Japan [81], Germany [82], France [83], Greece [84] and the Netherlands [85] are already
working on heat countermeasure guidelines for practitioners involved in town development, merging
academic and industrial results. Mist spraying is included, thus design guidelines are urgently needed. Also
suitable and shareable indicators should be indicated/developed to describe the efficacy of the cooling process,
univocally. In this review, half of the studies reported the cooling efficacy, but adopting a variety of indicators
(CE, COP, EER, wet bulb depression…).
All things considered, future studies may address several non-fully-unveiled aspects: i) potential applications
in tropical, AC-intensive conurbations, ii) potential and need for methodological standardization, iii) potential
for comparison and design of monitoring equipment and protocols, iv) potential investigation of unchartered
parameterizations, v) potential for comparison and design of wind control provisions, vi) potential
investigation of secondary impacts, notably health repercussions, vii) potential for conceptualization of water
management systems and practices, viii) potential for alternative renewable-powered solutions.
7. Conclusions
This manuscript is a compendium of 28 studies on outdoor cooling by water mist spraying, statistically
analysed and perused to delineate trends, gaps and drivers.
The revised studies demonstrate how spray cooling efficiently and easily integrates in existing city
infrastructures or renovation projects, as it achieves high local impacts in a cost-effective (potential
evaporative cooling effect over 200 times the consumed power [45]) and extremely versatile way.
Furthermore, the effect is controllable, unlike most urban heat island mitigation/adaptation approaches, thus
encouraging climate-adaptive and renewable-powered solutions.
Nevertheless, by amalgamating all data, three main criticalities emerged:
standardized protocols for collecting and reporting on the cooling potential of spray systems are
imperative to build a consistent and ample casuistry. Climate-sensitive metrics and classification
should be introduced to help selecting the most appropriate components. Besides, specific quantitative
performance descriptors should be developed, based on general empirical laws;
much more research is required to master the functional and geometrical alternatives for targeted
applications. Notably, wind speed, exposure time, injection density should be included in the
parameterizations;
the interplay and cumulative effects of competing evaporative processes (evapotranspiration,
condensed moisture evaporation, secondary and inertial effects of the spray and of precipitation …)
should be integrated in a holistic approach.
The results collected in this review allow to draw some preliminary optimization rules: i) high-pressure
systems around 3 MPa, should be preferred to generate droplet diameter below 40 µm, ii) the injection should
be stopped whenever ambient temperature is below 30 °C, relative humidity over 70 % and wind speed over

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
3 m/s, iii) windscreens should be provided in open spaces to maintain a dominant light breeze of 1 m/s, iv)
control logics should be mandatory in humid climates, v) the cooling should be adjusted by changing flow
rate or the number of active nozzles as strongly affecting intensity and spread of the action.
Further investigation is needed, but aligning the experiences of different countries constitutes a first step to
avoid a cloud of incompatible results. Against the current backdrop, fully depicted in this paper, fertile is the
opportunity for future studies to dilute segregation and move towards a targeted use of this technology in
different contexts, taking advantage of water spraying multiple impacts on outdoor environmental quality.
Appendix
Table A1 List of referenced works and general information. Field M stands for Methodology (N=Numerical,
E=Experimental, B=Both).
Title and reference Year Editorial Space M Parametric analyses
Droplet evaporation with reference to the
effectiveness of water-mist cooling [33] 2007 Applied Energy semi-enclosed
(railway tunnels) N droplet size
Study of cooling system with water mist
sprayers: Fundamental examination of
particle size distribution and cooling
effects [57]
2008 Building
Simulation semi-enclosed N 1) droplet size (by varying the sheet constant under
the same spray conditions)
2) number of nozzles
Study on a cooling system using water
mist sprayers; system control considering
outdoor environment [58]
2008 Korea-Japan Joint
Symposium on
Human-
Environment
System, Cheju,
Korea
open space
(under-canopy
waiting area)
N initial T and RH
Experimental study and numerical
simulation on evaporative cooling of fine
water mist in outdoor environment [59]
2009 2009 International
Conference on
Energy and
Environment
Technology
open space B airflow: downwind and upwind
Cooling system with water mist sprayers
for mitigation of heat-island drymist
system: Uchimizu [68]
2009 Seventh
International
Conference on
Urban Climate
semi-enclosed
(railway station
staircase)
E none
Quantification of the effect of cooling
mists on individual thermal comfort [72] 2009 The seventh
International
Conference on
Urban Climate
indoors (3.5m x
4m x 3.5m
concrete room)
E height above the target (heat-controlled, instrumented
arm)
Application and numerical simulation on
water mist cooling for urban environment
regulation [60]
2010 Lecture Notes in
Computer Science open and semi-
enclosed spaces N 1) airflow rate
2) initial conditions
Performance evaluation of misting fans in
hot and humid climate [71] 2010 Building and
Environment open spaces (2
food centres and
1 coffee outlet)
E none
The research and application of spray
cooling technology in Shanghai Expo
[61]
2011 Applied Thermal
Engineering First run: semi-
enclosed
(pavilion)
Second run:
open space
(EXPO waiting
area)
E none
Study of mist-cooling for semi-enclosed
spaces in Osaka, Japan [46] 2011 Procedia
Environmental
Sciences
semi-enclosed
(large atrium
20mx8mx30m)
B 1) nozzle typology (3 cases)
2) pressure (2 cases)
3) height above the ground
A study of mist spraying system by urban
transportation [54] 2012 Design for
Innovative Value
Towards a
Sustainable Society
open space N none
Effect of the evaporative cooling
techniques by spraying mist water on
reducing urban heat flux and saving
energy in apartment house [73]
2012 Journal of Heat
Island Institute
International
open space
(residential
building
exteriors)
B none
Solar energy powering up aerial misting
systems for cooling surroundings in Saudi
Arabia [48]
2013 Energy Conversion
and Management open space (at
Taibah
University)
E 1) initial T
2) control (on/off timing 20s/20s vs 60s/20s)
3) solar panel load (3 cases, low=10Ω, medium=20Ω,
high=40Ω)
4) solar panel tilting (0° flat, 25° and 45°)
Experimental study to evaluate mist
system performance [63] 2014 International
Journal of
Innovative
Research in
n.s. (possibly
indoors from
pictures)
E 1) water flow rate
2) height above the ground
3) inlet water temperature

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
Advanced
Engineering
Effect of water temperature on
evaporation of mist sprayed from a nozzle
[53]
2015 Journal of Heat
Island Institute
International
n.s B 1) water temperature (by defining the saturation
vapor pressure and binary diffusivity of the air-water
interface)
2) distance from injection and initial T, RH
conditions (7 cases)
CFD analysis of the impact of physical
parameters on evaporative cooling by a
mist spray system [62]
2015 Applied Thermal
Engineering semi-enclosed or
open space
(wind tunnel
type)
N 1) inlet air temperature (5 cases)
2) inlet air humidity ratio ω (5 cases)
3) inlet airflow rate (5 cases)
4) inlet water temperature (5 cases)
5) inleTdroplet size distribution (d32 and spread, 5
cases each)
Evaporative cooling by water spray
systems: CFD simulation, experimental
validation and sensitivity analysis [49]
2015 Building and
Environment semi-enclosed or
open space
(wind tunnel
type)
N 1) finer and coarser mesh grid
2) turbulence model (5 cases)
3) number of streams (2 declared cases)
4) spray angle (3 cases)
Investigation of photovoltaic assisted
misting system application for arbor
refreshment [74]
2015 International
Journal of
Photoenergy
open space
(24m2arbor) E none
Evaluation of cooling effects: Outdoor
water mist fan [45] 2015 Building Research
and Information 24m
2
open space
for public events E none
A verification study on thetemperature
reduction effect of water mist injection
[67]
2015 The International
Journal of The
Korea Institute of
Ecological
Architecture and
Environment
open space (at
Kunkuk
university
campus)
E 1) water flow rate
2) solar radiation
3) horizontal distance
Design of solar powered mist cooling
system for a typical semi-outdoor area in
Nagpur [50]
2016 Researchgate semi-enclosed
(9m2)N none
Cooling effect of a mist fan for large
indoor spaces [34] 2016 Proceedings of 9th
Windsor
Conference.
Windsor, UK
large indoor
workspace
(40,000m2
factory space)
E 1) fan oscillation (60°, 50s period for 12 trials, fixed
for 4 trials)
2) fan tilt angle (0˚ for 12 trials, -4˚ downward for 4
trials)
The effect of evaporative cooling
techniques on reducing urban heat [69] 2016 Urban Climate
Mitigation
Techniques
open space E none
Simulating the cooling effects of water
spray systems in urban landscapes: A
computational fluid dynamics study in
Rotterdam, The Netherlands [56]
2017 Landscape and
Urban Planning semi-enclosed
(courtyard) N 1) total mass flow (3 cases)
2) height above the ground (3 cases)
Solving model of temperature and
humidity profiles in spray cooling zone
[52]
2017 Building and
Environment open space
(Expo security-
check area)
B 1) initial T and RH
2) pressure
3) droplet size
4) airflow rate
Parametric study on the cooling effects
from dry mists in a controlled
environment [65]
2018 Building and
Environment indoors (climatic
chamber) E 1) wet bulb depression (4 cases, variable air
temperature, constant pressure and water
temperature)
2) pressure (3 cases, constant air temperature at 20
°C, humidity at 50 % and water temperature)
4) water temperature (3 cases, constant air
temperature at 30 °C, humidity at 60 % and pressure)
Water nebulization to counteract urban
overheating : Development and
experimental test of a smart logic to
maximize energy efficiency and outdoor
environmental quality [6]
2019 Applied Energy Ancona: open
space (terrace)
Rome: open
space (50m ×
16m park)
E control logic (on-off, fuzzy, no-stop)
Outdoor comfort in urban open spaces:
Transversal field surveyand experimental
investigation of the benefits of water mist
cooling [55]
2019 Building and
Environment Ancona: open
space (terrace)
Rome: open
space (50m ×
16m park)
E none
Table A2 Injection characteristics (part 1). Vertical configuration=nozzle pointing downward, Ø=diameter
[mm]
Ref Nozzle type Cone
angle n° nozzles Arrangement
& operating mode Height above
ground
[33] n.s. n.s. 1 (single-
droplet study) not applicable n.s. (freely falling
droplets)
[57] pressure-swirl atomizer,
diameter=0.16mm 50°, 6°
spread 1-3 vertical configuration, multiple nozzles at 2.5m intervals 3.5m
[58] pressure-swirl atomizer 50° 3 Vertical configuration, 3 nozzles aligned in a single row at
approximately 2.5m intervals 3.5m
[59] n.s. n.s. n.s. rotational configuration (mated with fan) 2m

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
[68] dry-mist nozzle n.s. 30 vertical configuration
on-off control (2min on, 3min off) n.s.
[72] hydraulic nozzle n.s. 1 vertical configuration n.s. Positioned at
1.0, 1.5 and 2.0m
above the arm
[60] n.s. n.s. n.s. rotational configuration (mated with fan) 2m
[71] n.s. n.s. n.s. n.s. n.s.
[61] n.s. variable First run: 20
Second run:
136
First run: transversal arrangement (400mm spacing)
Second run: four spray columns, 34 nozzles around each spray
column pointing in four directions
First run ≈2.5m
Second run:
variable
[46] 3 types A, B and C (no
further specs) n.s. 1 vertical configuration from 6 to 25m
[54] n.s. n.s. n.s. along the longitudinal border of buses roof ≈2.5m
[73] single-flow, two types (A
for rooftop spraying and
outdoor AC spraying, B
for veranda spraying). No
further specs
n.s. rooftop
spraying: 12
veranda
spraying: 10
outdoor AC
spraying: 1
rooftop spraying: 4 rowsof 3 nozzles operating from 9 a.m. to
5 p.m. two pairs of rows sprayed simultaneously for 15
seconds, then stopped for 7minutes
veranda spraying: 8 nozzles on the south-facing veranda + 2
nozzles on the north window railing, no-stop operation
outdoor AC spraying: over 24 hours, 1s/29s on/off cycle
not applicable
[48] n.s. n.s. 4 vertical configuration, 4 nozzles aligned in a single row at 3m
intervals, on-off operation (2tested timings) 2.5m
[63] solid cone 45˚ n.s. horizontal configuration, 0.5m intervals 2.25m, but for PA 2
(2.25, 2.5 and
2.75m)
[53] hollow-cone, high-
pressure swirl atomizer 40° 1 horizontal configuration, U-shaped nozzle header wrapped by a
400W wire heater n.s.
[62] hollow-cone, 4mm
diameter 36° 1 horizontal configuration, downwind. 300 particle streams,
evenly distributed along the orifice perimeter 0.2925m
[49] hollow-cone nozzle, 4mm
diameter from 32°
to 48° 1 horizontal configuration, downwind. 300 particle streams,
evenly distributed along the orifice perimeter (set, after PA 3) 0.2925m
[74] hexagonal, 0.2mm
diameter n.s. n.s. 16- 28
provision vertical configuration, in a line n.s. (≈2.5m from
thermal images)
[45] n.s. n.s. 16 8 nozzles in rotational configuration 60° span, 30s period),
mounted around the circular header of a Ø350 fan
+ 4 nozzles, from stationary header 1 (near sitting users,
spraying diagonally upward) + 4 nozzles from stationary header
2 (on building wall, spraying diagonally downward)
≈1.5m fan misting,
≈50cm header 1,
≈3m for header 2
(from photos)
[67] blast spray n.s. 20 vertical configuration, 4 pipes with 5 nozzles each on a cross
shaped support. Operated from10am to 4pm 3m
[50] solid cone type turbulence
nozzle 15-120° n.s. vertical configuration up to 2.75m
[34] hydraulic nozzle n.s. 32 rotational configuration, mounted around the circular header of
a Ø600 fan. 2.5m (fan axis
center)
[69] micronizer (no further
specs) n.s. 128 vertical configuration, intermittent operation n.s. >4m (from
pictures)
[56] hollow-cone, 4mm
diameter n.s. 15 vertical configuration, 0.5 m intervals on a single horizontal
line 3m, but for PA 3
(3,4 and 5m)
[52] n.s. 30° 136 horizontal configuration, 4 spray columns, 34 nozzles each,
pointing in four directions from different heights (vertical
spacing=0.4m). Active from 1pm
not applicable
[65] double-flow pneumatics
spray nozzle with the
same pressure for air and
water
n.s. 1 horizontal configuration 1.8m
[6] hollow-cone, 0.20mm
orifice n.s. 24 vertical configuration. 6 nozzles x 4 branches, symmetrically
designed to balance the pressure drops. Customizable spacing,
set to 1m between nozzles and 1.25m between branches.
Covered area≈3m x 6m area.
Operating mode: no-stop vs on/off (20s/10s) vs fuzzy
≈3m
[55] hollow-cone, 0.20mm
orifice n.s. 24 vertical configuration, at 1m intervals, arranged in 4 parallel
strings. Operating mode: no-stop Ancona: 3.3m
Rome: 2.8m
Table A3 Injection characteristics (part 2). d32=Sauter mean diameter
Ref Droplet size Pump
power Outlet speed Pressure Nozzle flow rate Water T
[33] 7 diameters in the 25-200 µm range n.s. n.s. n.s. not applicable 288.5K
[57] Rosin-Rammler distribution
Sheet constant=24 (case 1), 12 (case 2), 6
(case 3). Volumetric mean diameter=9.0
µm (case 1), 12.7 µm (case 2), 17.9 µm
(case 3)
d32=16.9 µm (case 1), 20.8 µm (case 2),
32.6 µm (case 3)
n.s. n.s. 6MPa 0.83g/s 33.4 °C
[58] n.s. n.s. n.s. 6MPa 0.83g/s 28 °C
[59] Rosin-Rammler distribution
d30=72.9μm (measured)
d
32
=106μm (calculated)
n.s. 8m/s 0.5-0.7Mpa 80ml/min (measured)
0.012kg/s (modelled,
total)
n.s.

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
[68] n.s. n.s. n.s. 6MPa n.s. n.s.
[72] nearly log-normal droplet
distribution with d
32
=20 µm n.s. n.s. n.s. 36ml/min n.s.
[60] mean diameter=60 µm (imposed) n.s. 6m/s and 10m/s
for PA 1. 10m/s
for PA 2
n.s. 0.012 kg/s (total) n.s.
[71] n.s. n.s. n.s. n.s. n.s. n.s.
[61] n.s. 500W n.s. 1.5-7MPa ≈ 0.12-0.55m
3
/s (from
chart, probable typo) n.s.
[46] d
32
=41-79 µm (measured) 750W n.s. 0.7and 5.5MPa 3-40l/h n.s.
[54] n.s. n.s. n.s. n.s. n.s. n.s.
[73] average diameter=300μm (Nozzle A),
40μm (Nozzle B)
60W
and
30W
n.s. main water
pressure
(Nozzle A),
1.8MPa
(Nozzle B)
Rooftop spraying:
261l/day. Veranda
spraying: 322l/day.
Outdoor AC spraying:
66l/day
n.s.
[48] 5–10 µm n.s. n.s. 0.276MPa n.s. n.s.
[63] n.s. n.s. n.s. 0.4MPa 1.2l/min, but for PA 1
(0.42, 0.84 and 1.2l/min)
- expressed as total flow
rate
31 °C for PA 1
32.5 °C for PA 2
34.8, 39.7 and 44˚C for
PA 3
[53] Measured d32=21 µm (supply water
temperature=27 °C)
Numerical distribution: Rosin-Rammler
.Max diameter=25-100 µm; mean
diameter=23.4-25.5 µm; min diameter=1-
25 µm; Spread=1.75-3.53 µm;
#diameters=20; 30 particle streams
n.s. Measured average
at 2 cm from the
nozzle: 10-15m/s.
Reynolds <100.
Numerical input:
15m/s
5.5MPa 2.1l/h 8-92 °C (measured)
300-363K (simulated)
[62] Rosin-Rammler distribution
d32=369 µm, min diameter=74 µm, max
diameter=518 µm. 20 discrete droplet
diameters. Spread=3.67.
PA 5.1: d32=310, 340, 369, 400 and 430
µm. PA 5.2: spread=3.02, 3.35, 3.67, 4.00
and 4.33
n.s. 22m/s 0.3MPa 12.5l/min 35.2 °C for all PAs, but
PA 4 (31.2-47.2 °C)
[49] Rosin-Rammler distribution
d32=369 µm, min diameter=74 µm, max
diameter=518 µm. 20 discrete droplet
diameters. Spread=3.67.
n.s. n.s. 0.3MPa 12.5l/min 35.0-35.2 °C
[74] n.s. 550W 7MPa 0.075l/min n.s.
[45] d32=25 µm 750W n.s.
Airspeed=7m/s at
inlet, 5.5m/s at
2m, 3.3m/s at 3m,
2.7m/s at 4m
distance
6MPa 2.375l/min (19l/min in
total) n.s.
[67] <100 µm n.s. n.s. 5-6MPa,
regulated by
inverter
0.04l/min. PA by closing
nozzles (total rate of
0.16, 0.32 and 0.48l/min)
n.s.
[50] n.s. 184W
(total) n.s. 0.965MPa (140
psi) n.s. n.s.
[34] d
32
=25 µm 1500W n.s. 6MPa 86l/h (total) n.s.
[69] mean diameter=20mm (probable typo) n.s. n.s. n.s. high-
pressure 7.2l/h n.s.
[56] Rosin-Rammler distribution
d32=20 µm, min diameter=10 µm, max
diameter=60 µm. Spread=3.5
n.s. n.s. 0.3MPa 0.6l/min (total of
9.0l/min). For PA 1: total
flow rates=2.25, 4.5 and
9.0l/min
25 °C
[52] Experimental: diameter=16 µm (as per
manufacturer's specification)
Numerical: as for experiment, but for PA 3
(16, 26, 40, 60 µm)
n.s. n.s. Droplet and
airflow at same
velocity
Measured:
5MPa
Modelled: 1.5,
2, 3, 5 and
7MPa
Measured: 0.44g/s
Modelled: ≈0.15-
0.55m3/s for increasing
pressure
n.s.
[65] d32=7-9 µm (as per manufacturer's
specification) n.s. n.s. 0.3MPa for
PAs 1 and 3
0.25-0.35MPa
for PA 2
from 0.9l/h (at 0.25 MPa)
to 1.1l/h (at 0.34 MPa) 19 °C for PA 1
17 °C for PA 2
18.5-29 °C for PA 3
[6] droplet size distribution centred below 10
µm 919W n.s. 7MPa 1.5l/min (total) Ancona: 13-30 °C
Rome: 19±1 °C (daily
swings)
[55] droplet size distribution centred below 10
µm (as per manufacturer’s specification) 919W n.s. 7MPa 1.5l/min (total) n.s.
Table A4 Experimental setups

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
Ref Shadings/Ve
getation Duration Probes and loggers Points of measurement Sampling rate
[59] n.s. multiple 30-
min lasting
runs at
different times
of the day with
ambient
temperature at
about 35.5 °C
thermometers
evaporative dish 1m inter-spaced thermometers, 1.7m above the
ground (average stature)
diameters of droplets observed in evaporativedish
n.s.
[68] Shadings
n.s., no
vegetation
1 summer day
of 2007 temperature sensors T between inside and outside of the spray zone n.s.
[72] Shadings all
around 1-min tests
(30s misting,
after which the
arm is dried
and heated
again)
3 T-type thermocouples
Thermal imaging
Digital scale (accuracy 0.1g)
2 T points in the forearm mold to monitor the deep
core temperature, 1 T point 1mm below the
paraffin surface to monitor the skin temperature
n.s.
[71] n.s. Monitoring: 8
days in Aug
and Sept
Questionnaires:
Sept. 21-24
2009, three
time slots:
9:30-11am
(morning),
12:30-2pm
(afternoon) and
3:30-5pm (late
afternoon)
Babuc A data logger + Testo 445 (T,
RH ws probes, standard 15cm globe
thermometer)
HOBO weather station (T, RH ws
probes)
Single stage N6 Andersen air samplers
(Tryptic Soya Agar plates for bacteria,
Potato Dextrose Agar plate for yeast
and moulds)
Environmental measurements at 1.0 m above the
ground
Meteorological measurements outside the dining
areas
3/4 air sampling points, selected based on the
misting and seating locations (3 samples
consecutively taken per point)
Climatic
measurements
every 15min
intervals from
9am to 5pm
(3min time-
average)
Air sampling
over 4min
period
[61] First run:
semi-shaded
pergola +
greenery
Second run:
n.s.
First run: May
1 to Oct 31,
2008
Second run: 2
days, July 18
and Aug 13,
2010
HUMLOG10 hygro-thermograph
(accuracy ±0.3 °C), FYF-1 portable
wind velocity indicator (accuracy
±(0.3+0.03 v)) and Delta-HD2107.1
globe thermometer (accuracy ±2 °C)
First run: 5 points for T and RH: 1 at 1m from the
injection, 3 at 1.5m from the injection (0.3m and
0.6m relative distance), 1 at 2m from the injection
and 1 point for undisturbed ambient parameters at
1.5m above the ground
Second run: 5 to 10 points for T and RH diagonally
arranged between the spray columns (0.1 to 0.25m
relative distance) and in undisturbed location under
the awnings 40m from the columns
1min for T and
RH, 5min for
other
parameters
[46] atrium roof
shading 5 nights PDPA (Phased Doppler Particle
Analyzer)
Electronic temperature/humidity
loggers (± 0.5 °C and ± 5 %
accuracies, some provided with
ventilated weather shelters)
T-type thermocouples
PDPA measurements along the centreline of the
spray, 50cm below the nozzle
Flow rates measured by placing a beaker over the
nozzle for 120 seconds and measuring the weight of
accumulated water
Evaporation rate measured using a timer and
weighting the unevaporated potion of the mist,
captured on a waterproof vinyl tarp
4 points for T and RH via electronic loggers: 3
sheltered probes 3m distant from the centreline of
the mist spray cone at 1m, 13m and 25m height + 1
unsheltered probe placed at20cm horizontal
distance from the nozzle
16 points for T and RH via thermocouples: 4X4
pattern, intervals of 50cm, suspended 1m above the
floor centred below the nozzle for horizontal
temperature distributions, or propped up
perpendicularly in contact with the floor for vertical
temperature distributions
10s for T and
RH via
electronic
loggers
1s for T and
RH via
thermocouples
[73] No greenery Aug 10 -Sept
27,2007 n.s. n.s. n.s.
[48] No shadings Dec and early
Jan n.s. n.s. n.s.
[63] No greenery n.s. AR847 digital T and RH meter, sensor
type k
Digital anemometer model 8901
T and RH measured at pedestrian height, 6
horizontal points reported in graphs
Air velocity measurement point n.s.
n.s.
[53] n.s. n.s. PDPA (Phased Doppler Particle
Analyzer)
T-type uncoated thermocouples with
solder beads under 1 mm
8 PDPA measurements taken 10cm from the nozzle
along the centreline every 2 degrees (repeated at
30cm)
2 water temperature measurements: 1) 5cm
upstream of the nozzle; 2) closely over the nozzle
embedded in sponge inside a bottle
Same bottle+sponge to measure flow rate
Evaporation rate and cooling measured in a
45cmx45cm forced-draft air duct (2-3m/s).
n.s.

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
Measurements taken for 7 different cases of
distance from nozzle and inlet air temperature by a
grid of 25 evenlyspaced thermocouples
Direct measurement of temperature in mist cone by
6 thermocouples, 3cm spaced in a row along the
nozzle spray centreline, starting 1.5cm from the
nozzle first, and then from closer (5mm)
[74] arbor
shading n.s. Solarimeter
Thermal camera Solar radiation measured for 10 days
arbor temperature mapped by thermal imagery
Humidity measurements n.s.
n.s.
[45] area well
shaded by
trees and
buildings
10-11 Aug
2013 (11am-
3pm)
IR thermometer (accuracy 0.2 °C)
hot-wire anemometer
Data recorders with temperature
(thermistor)/humidity (capacitive
polymer) sensors (±0.5 °C and ±3 %
accuracies)
Globe thermometer (Ø150, T-type
thermocouple, 20min response time)
Stopwatches
3 K-type thermocouples, 5cm square,
0.5mm thick
Heat flux sensor
IR skin measurements at head and forearm level
Injection velocity measurement along the fan
centreline, repeated in front of the fan, at 2, 3 and
4m distance.
4 points for T and RH via data recorders: at the fan
inlet, at a shaded brick wall 6m from the fan along
the fan centreline, 4m from the misting area but
within the same shaded area, 4m from the misting
area in an unshaded area with the same brick
ground cover (shielded from sunlight)
Globe temperature measured in the centre of the
misted area, but after the experiment, under similar
sun and air conditions. Correlation for the misting
days (R2=0.74).
All-sky insolation measured on a rooftop 50m from
the site (sensor n.s.)
Stopwatches used to record the time each person
spent in the mist
1 K-type themocouple embedded in a 60 W silicone
rubber heater with automated temperature
controller, 2 K-type themocouples and heat flux
sensor on top of the silicone rubber sheets. The
skin-analogue system was placed along the
centreline of the misting fan at 2–5m distance.
5s for IR
measurements
1s for heat flux
measurements
(intervals of 3–
10 min)
[67] highly
vegetated
University
campus +
artificial
shielding
features
Approximately
5 months (Jun-
Nov 2014)
Watchdog 2000 Series weather station
(measuring solar radiation,
temperature, wind direction, wind
velocity, relative humidity)
LI-19 datatalogger with secondclass
solarimeter
VelociCalc Multi-Function Ventilation
Meter (to measure air velocity and
temperature)
5 Temperature sensors (n.s.)
Solar radiation measured inundisturbed location
4 T points inside the mist at 1m from the centre,
corresponding to cardinal points + 1 T point at 10m
distance from the mist system perimeter. Over peak
hours the 4 probes in the mist were moved at 1,3,5
and 7m horizontal distance from the centre (height
n.s.)
1min
[34] shadings all
around, No
greenery
Sept. 21-22,
2014 between
10am and 4pm
16 x 30-min
testing cycles
hot-wire anemometer
10 T and RH data loggers (thermistors
with capacitive film humidity sensors,
response time=7min)
10 T-type thermocouples (0.4mm
single-core wire with solder beads)
Globe thermometer (Ø150, T-type
thermocouple)
2 heat flux meters (5cm-square)
The air speed profile of the fan was measured along
the centreline (see Figure 3) at 1.2m height at 5m
intervals
10 T and RH points, 8 within the sprayed area at
1.2m height from 10m distance, 5m spaced (3 on
the centreline, 3 on the 60° oscillation edge line, 2
on the outer edge 1m farther) + 2 outside the
sprayed area (one near the fan inlet and one at 15m
distance perpendicular to the fan airflow). Both
built-in loggers and thermocouples in the same
locations.
Globe temperature measured at 1.2m height
perpendicular to the air flow at a distance of 5m
from the fan, outside the misted area.
Heat flux measured between the sheets of the
heated skin analogue
20s for T and
RH loggers, 1s
for
themocouples
[69] pergola
shading Apr.20 -
Oct.12, 1992 n.s. n.s. n.s.
[52] visible
shadings all
over the area
1 day, from
1:30pm to
2:40pm
Temperature and humidity sensors n.s. 10 T and RH points, located downwind of the spray
column and set 1m apart from each other at a height
of 1.5m along the diagonal line between opposite
columns
10min
[65] none 10 cycles: 1
hour for
chamber
operation and
10 minutes of
spray operation
per cycle
4mx2.7mx2.5m temperature and
humidity climatic chamber in the
range range of -10 to +80 °C (accuracy
±0.5 °C), and a humidity range of 20
to 95 % RH (accuracy ±4 %)
HOBO temperature/relative humidity
dataloggers, type #U12-011 (±0.35 °C
and ±2.5 % accuracies)
Water thermometer
10 points for T and RH: 1 behind the fogger,
upstream of the nozzle to measure inlet dry bulb
temperature + 9 (3 by 3 arrangement) 2m from the
fogger at heights of 60, 120 and 180cm height from
the ground
1 point for water temperature in the tank
1s for T and
RH
[6] Ancona:
none Rome:
partial tree
canopy
Ancona: 9 days
(July and Aug
2018), 10am-
8pm
Weather station connected to LSI
Lastem Elog: thermohygrometer (0.2
°C and 1.5 % accuracies, response
time=10s), cup anemometer (1.5 %
and 1° accuracies, 2.5s and 0.7s time
Meteorological station placed 50m (Ancona) and
16m (Rome) from the spraying system
(thermohygrometer, anemometer and radiometer at
1.7, 2.0 and 2.5m above the ground, respectively)
5 points for T and RH, distributed at the centre and
1min for meteo
measurements,
100 samples at
1 kHz for T
and RH, 10

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
Rome: 6 days
(Aug 2018),
11am-7pm
constants), radiometer (<5 % accuracy,
response time<30s)
5 PCMINI52 miniaturized
thermohygrometers by Michell
Instruments, connected to NI 9205
acquisition module (0.3 °C and ±2 %
accuracies, response time<10s, anti-
wetting plastic protection, solar shield)
Rain detector Kemo M152, connected
to NI 9219
Mechanic flowmeter
Manometer
RTD (Pt100) for water temperature
Digital I/O NI 9401 module to pilot
the solenoid valve
+ only in Rome and for one day: 5
ATMOS 14 thermohygrometers
connected to DECAGON EM60Gs
(±0.2 °C and ±2 % accuracies)
mid-points of the ground-projected perimeter of the
spraying rack, aligned with the cardinal directions,
at 1.1m above the ground
Flow, pressure and water temperature spot
measurements performed alongthe pump line
Only in Rome for oneday: 4 T and RH points in the
middle of the mist at 1.1 m, 1.7 m, 2.1 m e 2.5 m
above the ground + 1 T and RH point 100m away,
at about 2.5 m above the ground
samples at 1.95
Hz for rain
detection,
sampling rate
of 10s for all
the other
probes. All
records
averaged and
databased with
a 1min time
step
[55] Ancona:
none Rome:
partial tree
canopy
Ancona: 4 days
(July 2018),
10am-8pm
Rome: 7 days
(Aug 2018),
11am-7pm
As in [6] + a Ø150mm, Pt100 globe
thermometer (accurate to ±0.2 °C). No
water temperature measurements.
As in [6]
Globe temperature measured in the middle of the
mist.
As in [6]
Included globe
temperature
Table A5 Numerical setups
Ref Software Control
volume Mesh features Models & Boundary Conditions
[33] Q-basic Railway tunnel
(dimensions
n.s.)
none T=301K, RH=40 %, initial droplet temperature=288.5K
Freely-falling droplets, 0m/s initial velocity
[57] ANSYS Fluent 50m x 15m x
4m 11,200 cells Transient regime (convergence criterion=T fluctuations<10
-
6
K)
Discrete phase model
Standard k-ε turbulence model
No-slip condition (at horizontal surface), free-slip (at side surface)
Ground=concrete, 5m thick (T=29.8 °C, thermal conductivity=1.4W/(m·K), solar
absorptance=60 %)
Roof=PVC-coated glass-fiber, 0.53mm thick (external T=38.2 °C, external heat
transfer coefficient=23W/(m2·K), thermal conductivity=0.11W/(m·K), solar
absorptance=10.8 %, solar transmittance=13.7 %)
T=33.4 °C, RH=58 % (Tokyo typical summer day), wind=0.1m/s
[58] ANSYS Fluent
6.3 50m x 15m x
4m n.s. Discrete phase model
Free-slip condition (at side surface, adiabatic)
Ground=concrete (heat transfer coefficient=23W/(m2·K), thermal
conductivity=1.4W/(m·K), solar absorptance=60 %)
Roof=PVC-coated glass-fibre plain-weave (heat transfer coefficient=23W/(m2·K),
thermal conductivity=0.11W/(m·K), solar absorptance=10.8 %, solar
transmittance=13.7 %)
T and RH scenarios: 30 °C/80 %, 30 °C/60 % and 34 °C/60 %; wind=0.1m/s, solar
radiation=363W/m2(average June to September 2008, 9am-7pm)
[59] n.s. 15m x 10m x
6m 720000
tetrahedral cells Realizable k-ε turbulence model
Standard wall function (near-wall), No-slip condition (at wall)
Thermal condition: combined external radiation and convection heat transfer
Ground=concrete (T=50 °C, external emissivity=0.95)
Boundaries=outflow, fan=inlet velocity
T=35 °C, RH=60 %, wind=1.5m/s
[60] ANSYS Fluent 15m × 10m x
4m non-structured
and non-uniform
mesh, generated
with hexahedral
scheme. 570,000
cells
Transient regime
Discrete phase model
SIMPLE algorithm for coupling pressure and velocity
Second-order upwind differencing scheme for convection
Realizable k-ε turbulence model
Standard wall function (near-wall), No-slip condition (at wall)
T=35 °C, RH=60 %, solar radiation=600 W/m2
Ground=concrete (T=50 °C for parameterization 1, 55 °C in open space 39 °C in
semi-enclosed space for parameterization 2, external emissivity=0.71)
Roof (PA 2, semi-enclosed space)=PVC
[46] Fire Dynamics
Simulator (FDS
Version 5.5, Build
6004)
n.s. n.s. Turbulence model: Smagorinsky form of Large Eddy Simulation
[54] n.s. Open urban
areas around
Sakae Station.
Dimensions n.s.
n.s. n.s.

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
[73] SMASH
(Simplified
Analysis System
for Housing Air
Conditioning
Energy) +
Assessment Tool
for the Thermal
Load on Urban
Atmosphere
5 story building,
20 dwelling
units
none Occupancy and consumption schedules: 20 units, each occupied by a family of four
Rooftop spraying (and sidewall cooling) simulated by varying solar absorption for
evaporation at rooftop from 93 % to 30 %
Veranda spraying simulated by applying a -2 °C Tdrop
Outdoor AC spraying simulated by considering 36 % less cooling demand
[53] ANSYS Fluent
13.0 400cm x 45cm
x 45cm duct 1cm tetrahedral
mesh near the
nozzle, ranging
up to 4cm mesh at
the duct inlet and
outlet, total of 1.2
million elements
Steady state calculation (by imposing duct air speed at 0.3m/s)
Discrete phase model (injection), Spherical drag law, Gravity
Realizable k-ε turbulence model
Water density, thermal conductivity and latent heat of evaporation set constant at their
values for 27 °C.
Duct inlet air set at 27˚C and 75 % humidity as in experiments
[62] ANSYS Fluent
12.1 0.585m x
0.585m x 1.9m
wind tunnel
1,018,725
hexahedral cells.
Min cell
volume≈1.9·10-
8m3, max cell
volume≈5.9·10-
6m3, y*=35 and
135 (for max inlet
velocity of 3 m/s)
Discrete phase model
3D steady RANS equations
SIMPLE algorithm for coupling pressure and velocity
Automated Tracking Scheme
Realizable k-ε turbulence model
Spherical drag law
Standard wall functions, no-slip condition, adiabatic,
"reflect" discrete phase at walls, "escape" discrete phase at outlet plane
T=39.2 °C for all parameterizations, but parameterization 1 (27.2-43.2 °C), wet bulb
temperature=18.7 °C but for parameterization 2 (RH=5.9 %, 11.8 %, 17.6 %, 22.4 %
and 29.1 %), mean radiant temperature=35 °C
wind=3-15m/s, turbulence intensity=10 %
[49] ANSYS Fluent
12.1 0.585m x
0.585m x 1.9m
wind tunnel
Coarse
grid=360,000
cells, fine
grid=2,880,000
cells for
parameterization
1.
After sensitivity
analysis:
1,018,725
hexahedral cells.
Stretching
ratio=1.05 around
the nozzle. Min
cell
volume≈1.9·10-
8m3, max cell
volume≈5.9·10-
6m3, y*=35 and
135 (for max inlet
velocity of 3m/s)
Discrete phase model
3D steady RANS equations
SIMPLE algorithm for coupling pressure and velocity
Droplet deformation, secondary break-up or collisions are neglected
Automated Tracking Scheme
Turbulence model: Realizable k-ε for all parameterizations, but parameterization 2
(standard k-ε , realizable k-ε, renormalization Group k-ε, standard k-u, Reynolds
Stress Model)
Spherical drag law
Standard wall functions, no-slip condition, adiabatic,
"reflected" discrete phase at walls, "escaped" discrete phase at outlet plane
T=39.2 °C-41.4 °C, wet bulb temperature=18.5-18.7 °C
wind=1-3m/s, turbulence intensity=10 %
[56] ANSYS Fluent
12.1 Circular inner
subdomain
(diameter=1200
m) with
explicitly
modelled
buildings
(resolution=1-3-
8m,
progressively
reduced) +
surrounding
outer hexagonal
subdomain
(side=1200m)
with implicitly
modelled
obstacles
(roughness).
Height=400m
6,610,456
hexahedral and
prismatic cells
(surface-grid
extrusion
technique by van
Hooff and
Blocken). Min
cell size=4.1·10−2-
24m3
Min 10 cells
along building
edges and
between
neighbouring
buildings and wall
adjacent cell faces
Discrete phase model
3D Unsteady Reynolds-Averaged Navier-Stokes (URANS) + energy equation (time
step=900s, 60 iterations per time step)
SIMPLE algorithm for coupling pressure and velocity
Automated Tracking Scheme
Realizable k-ε turbulence model
Conduction, convection (Boussinesqapproximation) and radiation (P-1 radiation
model + solar calculator) fully coupled with the wind flow
One-directional conduction at solid boundaries.
Boundary conditions according to 17 July 2006 heat wave 12:00, hourly
meteorological data from 4km distant station=Inlet mean ws=3 m/s, wind
direction=80°, inlet airT=29.7 °C, RH=33 %, mean radiant temperature=45 °C
Aerodynamic roughness length (z0) of the surroundings=0.5-1m
Ground= 10m thick earth layer (density=1150kg/m3, specific heat=650J/kgK, thermal
conductivity=1.5W/mK, absorptivity=0.60, emissivity=0.90, imposed T=10 °C at10m
depth, imposed evapotranspiration power=80W/m2 (6am–11am, 3pm-6pm), 80W/m2
(11am–3pm)
Building walls= 0.4 m thick brick layer (density=1400kg/m3, specific heat=900J/kgK,
thermal conductivity=1.7W/mK, absorptivity=0.75, emissivity=0.88, imposed indoor
T=24 °C, internal convective heat transfer coefficient and emissivity=8W/m2K and
0.95)
Sky= free-slip wall, outflow
[52] n.s. same as
experimental not applicable Spherically symmetric model with a saturated boundary layer
Stable diffusion theory of one-dimensional spherical coordinates
Stephen's radial diffusion law
Gilliland liquid-phase diffusion formulation
Droplet velocity assumed similar to the airflow velocity
Gravity field, border influence (infinite space), condensation on the droplet=neglected
T=39.4 °C, RH=35 %, ws=2m/s but for PA 4 (1, 2, 3, 5m/s)
Outdoor environment not affected by wind

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
Table A6 Cooling results. Cooling effectiveness (achieved Tdrop over theoretical wet bulb depression) is
indicated by either η or ε or CE (ASHRAE version). COP is the heat removed by evaporation over the input
energy for spraying. ACEC is the air-conditioning energy consumption.
Ref Temperature drop Cooling efficacy
Min/Max/Avg, Spatial spread Inertial
effects?
[33] Spatial spread: 0.006-15.7m for increasing diameter n.s. n.s.
[57] Tdrop=1.5 °C at the spray location and 0.5 °C about 2.5m behind
Same results with multiple nozzles n.s. n.s.
[58] 30℃/80 % scenario: max Tdrop=2.0 °C under the nozzles, max
Tdrop=0.5 °C on 1-m adjacent plane (at 1.5m)
30℃/60 % and 34℃/60 % scenarios: max Tdrop=2.5 °C under the
nozzles, max Tdrop=1.0 °C on 1-m adjacent plane (at 1.5m)
n.s. n.s.
[59] Experimental max Tdrop=6.7 °C, cooling distance along the axis
direction=4.5-5m
Numerical max Tdrop=9 ℃ in the central plane along the axial direction,
max Tdrop=6.5 ℃ at 4-9 m distance (downwind flow case), Tdrop=3-5
℃ in the vertical cross-section along the axial direction (upwind flow
case)
n.s. n.s.
[68] max Tdrop=1.63 °C between 9am and 1pm, 1.9 °C between 1pm and
3pm. n.s. COP=300 (down to 50 for miniaturized systems)
[72] For initial T=20.7-21.2 °C, RH=42-49 %, wet bulb temperature=13.4-
14.5 °C: skin temperature drop=5.3, 3.3 and 3.4 °C at 1,1.5 and 2m
distance, respectively. Negligiblecore temperature drop
n.s. n.s.
[60] 3 °C Tdrop in 4-10m for 6m/s inlet velocity
3 °C Tdrop in 3-13m for 10m/s inlet velocity
3-5 °C Tdrop in the semi-enclosed space along the injection
1-2 °C Tdrop in the open space along the injection
n.s. n.s.
[71] Average Tdrop (4-day mean, 95 % confidence)=1.5±0.1 °C n.s. n.s.
[61] First run: For T≈30 °C and RH<70 % Max Tdrop=1-2 °C, for T≈ 35 °C
and RH≈45 %, Max Tdrop=5-7 °C
Second run: Max Tdrop=6-12 °C (on colder and hotter days
respectively); Avg Tdrop=1.5-4 °C at a distance of 3 m from spray
column, 0.5-1.5 °C at a distance of 7 m.
Yes, cooling
persisted
after tens of
minutes
ηmax=65 % (first run) and 90 % (second run)
COPmax=30 for inlet pressure≥5MPa
[46] mean Tdrop=0.7 °C horizontally and 0.5 °C vertically for Nozzle A
mean Tdrop=1.8 °C horizontally and 2.1 °C vertically for Nozzle B
NB: the climate control system in the atrium kept T in the 25-28 °C range
and RH in the 44-69 % range
n.s. Evaporation rates: near-complete or complete (95-100 %)
at 15m height with Nozzle A (for both pressure levels) or
with Nozzle B at 25m (0.7MPa)
∆Twb=4.4-8.2 °C
[54] n.s.. A weak global effect is denounced n.s. n.s.
[73] Rooftop spraying: day average outdoor Tdrop=16.4 °C, indoor
Tdrop=1.2 °C at 1.2 m above floor level
no effect below the third floor (simulated)
Veranda spraying: day average , indoor Tdrop=1.9 °C at 1.2m
Outdoor AC spraying: n.s
rooftop
spraying:
cooling
observed
even after
cessation of
spraying
(5pm)
Experimental: rooftop spraying: 9.7 % ACEC reduction;
veranda spraying: n.s., outdoor AC spraying: 36 % average
ACEC
Simulation: : rooftop spraying generated -1 % and -22 %
ACEC on 4th and 5th floors respectively; veranda spraying
generated -65 % ACEC at each floor, but the 5th (-51 %);
outdoor AC spraying generated -36 % ACEC; sidewall
cooling contributed for extra - 4/5 %. Total=79/80 %. Max
total energy saving of each dwelling units=15 % (August).
max AC using time reduction=90 % (September). Water
consumption increased by a factor of 1.3
[48] max Tdrop≈10 °C
PA 2: max Tdrop≈3-7.8 °C for increasing ON time.
n.s. max EER=energy efficiency ratio calculated for the best
cooling performance=ratio of the output cooling [Btu] to
the input electrical energy consumed by the system [Wh] at
a given point in time=128.4
max COP=37.7
[63] PA 1: max Tdrop=9.4˚C (water flow rate=1.2l/min, inlet air conditions
T=35 °C RH=21 %, measured at pedestrian height)
PA 2: max Tdrop=6.6-11 °C for decreasing height above the ground
(water flow rate=1.2l/min, inlet air conditions T=35 °C RH=21 %,
measured at pedestrian height)
PA 3: max Tdrop≈13-15 °C (water flow rate=1.2l/min, inlet air
conditions T=44 °C RH=17 % at 1.1 m/s, measured at pedestrian height)
n.s. PA 1: max ε=cooling effectiveness=59.1 at 2m horizontal
distance and max flow rate (with respect to 19.1 °C wet bulb
temperature)
PA 2: max ε=cooling effectiveness=59.1 at 2m horizontal
distance and min height above the ground (with respect to
19.1 °C wet bulb temperature)
PA 3: max ε=cooling effectiveness=62.5 at 2m horizontal
distance and min water inlet temperature (with respect to 21
°C wet bulb temperature)
[53] Tdrop≈0.5-2.5 °C. Average cooling loss=0.26 °C (when water T is
increased from 20 °C to 60 °C). Min cooling loss=0.04 °C for T=19 °C,
RH=45 % at 0.75m distance. Max cooling loss=0.88 °C for T=23 °C,
RH=38 % at 1.25m distance
n.s. Min average evaporation rate=37 % for T=20 °C, RH=69
% at 0.5m distance, Max average evaporation rate=95 % for
T=18 °C, RH=68 % at 2.0m distance
Average evaporation rate gain=7.3 % (when water T is
increased from 20 °C to 60 °C). Min gain=5.8 % for T=19
°C, RH=45 % at 0.75m distance. Max gain=12.9 % for T=7
°C, RH=48 % at 1.5m distance.
[62] PA 1: max Tdrop=3.3-9.3 °C for increasing air temperature. PA 2: max
Tdrop≈3-9 °C for decreasing humidity ratio. PA 3: max Tdrop≈0.8-7 °C
for decreasing relative velocity. PA 4: max Tdrop≈2-9 °C for decreasing
n.s. PA 1: Max sensible cooling=4.1-9.8kW for increasing air
temperature. CE=25-40 % for increasing air temperature.

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
water temperature. PA 5.1: max Tdrop≈4-10 °C for decreasing mean
diameter. PA 5.2: max Tdrop≈5.8 °C for increasing spread
PA 2: Max sensible cooling≈4-10kW, CE=21-37 % for
decreasing humidity ratio.
PA 3: Max sensible cooling≈8.5kW for the lowest relative
velocity. CE=3-35 % for decreasing relative velocity.
PA 4: Max sensible cooling=3.6-10.2kW for decreasing
water temperature. CE n.s.
PA 5.1: Max sensible cooling≈5.8-12kW for decreasing
mean diameter. CE improves by >110 %
PA 5.2: Max sensible cooling≈6.5-9.8kW for increasing
spread. CE n.s.
[49] Tdrop=4-13 °C for smaller droplets n.s. Max sensible cooling=8kW
[74] max surface Tdrop=20 °C n.s. max EER=64.15, max COP=18.8 (average measured T=35
°C, RH=40 %)
[45] Complete evaporation within about 5m of the fan.
Skin temperature average drop=1.18 °C at the arm and 1.08 °C at the face.
1–3 °C within 10 s vs 0.5 °C with only fan (and no mist)
Skin Tdrop
=1 °C after
leaving the
mist
Heat flux average=60 W/m
2 (
no fan, no mist), 81 W/m
2
(fan
and no mist) and 201 W/m2(fan and mist) at 4m distance.
Over 300W/m2for misted convection at 2m distance. With
thermal loads around 70 W/m2, the cooling efficacy was
guaranteed. No specific indicator.
[67] max Tdrop≈6 °C
For PA 1:
Lower flow rate (0.16l/min): daily average Tdrop=3.1 °C (2.6 °C between
10am-12pm, 3.5 °C between 12-2pm, 2.9 °C between 2-4pm). Mid-range
flow rate (0.32l/min): daily average Tdrop=3.5 °C (2.8 °C between 10am-
12pm, 4.0 °C between 12-2pm, 3.4 °C between 2-4pm). Higher flow rate
(0.48l/min): daily average Tdrop=4.4 °C (4.1 °C between 10am-12pm,
4.7 °C between 12-2pm, 4.3°C between 2-4pm)
For PA 2:
Lower solar radiation (700-750 W/m2): average Tdrop=2.0, 2.4 and 3.1℃
for increasing water flow rate. Mid-range solar radiation (750-800
W/m2): average Tdrop=3.1, 4.2 and 4.5℃ for increasing water flow rate.
Higher solar radiation (800-850 W/m2): average Tdrop=3.7, 4.8 and
5.4℃ for increasing water flow rate
For PA 3 (at daily maximum temperature):
1m distance: average Tdrop=3.0, 3.5 and 4.5℃ for increasing water flow
rate. 3m distance: average Tdrop=2.0, 2.8 and 3.2℃ for increasing water
flow rate. 5m distance: average Tdrop=0.5, 1.6 and 2.1℃ for increasing
water flow rate. 7m distance: average Tdrop=0.2, 0.5 and 1.2℃ for
increasing water flow rate
n.s. n.s.
[50] n.s. n.s. n.s.
[34] For initial T=6-29˚C and RH=30-45 %. Max Tdrop in a 10m radius.
PA 1 (tilt=0°): max Tdrop=2.6-3.1 °C (oscillating vs fixed fan). PA 2
(tilt=-4°): max Tdrop=3.0-4.8 °C (oscillating vs fixed fan), complete
evaporation within 10-15m radius
n.s. n.s.
[69] max Tdrop=5 °C at peak hours, 10 °C (absolute maximum, for 42 °C
initial temperature) n.s. n.s.
[56] PA 1: Max Tdrop≈1-7 ◦C for increasing flow rate, underneath the
injection in the middle of the spray line, at pedestrian height (1.75m from
ground), spatial spread: 0-2 °C up to a distance of 8m from the spray line
for increasing flow rate
PA 2: Max Tdrop≈5-7 ◦C for decreasing height
n.s. n.s.
[52] Experimental results (for average ambient temperature=39.4 °C, relative
humidity=35 % and ws=2m/s):
Tdrop>9 °C at 1m distance, ≈6 °C at 2m, ≈3.5 °C at 3m, ≈2 °C to 1 °C
between 4 and 10m
Numerical results: max Tdrop (close to the column)≈14 °C
38 °C/40 % vs 30 °C/70 % scenarios: Tdrop (3m distance)≈1.2-3.0 °C vs
0.3-1.2 °C for increasing pressure, Radius of the cooled area=2.2-5.0m
for increasing pressure, Tdrop (16 µm diameter, 3m distance)=3.1 °C vs
1.2 °C, Tdrop (26 µm diameter, 3m distance)=1.4 °C vs 0.5 °C, Tdrop
(40 µm diameter, 3m distance)=negligible, radius of the cooled area=2.4-
5.7 vs 1.4-3.3 m for decreasing diameter, Tdrop (3m/s airflow, 3m
distance)=2.3 °C vs 0.9 °C
38 °C/40 %: Tdrop (3m distance)=4.8 °C and 3.1 °C for 2m/s and 1m/s
airflow rate. Radius of the cooled area=4-7.8 m for decreasing airflow
rate. Similar patterns are achieved in 30 °C/70 % scenario.
38 °C/40 % vs 38 °C/70 % scenarios: Tdrop (3m distance)=3.1 °C vs 1.3
°C. 30 °C/40 % vs 30 °C/70 % scenarios:Tdrop (3m distance)=2.7 °C vs
1.2 °C
n.s. n.s.
[65] Tdrop range=2.85-7.52 °C for PA 1
Tdrop range=4.09-4.33 °C for PA 2
Tdrop=4.9 °C for PA 3
n.s. ∆Twb=3.67-8.96 °C for PA 1. ∆Twb=5.8 to 6 °C for PA 2
and 3
CE=50.32-77.66 %. SHR=sensible heat ratio=ratio
between sensible and total heat loads=0.4-0.75
[6] Max, min, average, 99th, 90th, 50th and 1th percentiles, interquartile
range and standard deviation given for each monitoringday.
Ancona: average conditions: T=30.7 °C, RH=59.5 % and ws=2.6m/s
Fuzzy control: max Tdrop=5-7.4 °C, on the coldest (29.0 °C) and hottest
(32.6 °C) days. Max Tdrop rate=1 °C/min
on/off control: max Tdrop=5-7.5 °C, on the coldest (30.3 °C) and hottest
(33.3 °C) days. Max Tdrop rate=1 °C/min
Increased
cooling after
spraying
at times,
notably
under fuzzy
control
Fuzzy vs on/off: energy saving= 67.5 % (max, on coldest
and cloudiest day), 51.2 % (average), 25 % (min), 58.8 %
(for comparable boundary conditions)

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
Fuzzy vs on-off: horizontal non-uniformity under the mist (among the 5
monitored points): 4 °C vs 5 °C range
Rome: average conditions: T=29.9 °C, RH=47.9 % and ws=0.8m/s
Max Tdrop=4.4 °C (fuzzy control), 6.1 °C (on/off), 6.2 °C (no-stop
control). Max Tdrop rate=0.8 °C/min (fuzzy), 1.0 °C/min (on/off and no-
stop)
Fuzzy vs on-off: horizontal non-uniformity under the mist (among the 5
monitored points): 8 °C vs 13 °C range. Max Tdrop at 2.1m above the
ground (-1 °C vs -2.3 °C compared to 1.1m)
[55] Max, min, average, 99th percentile, interquartile range and standard
deviation reported.
Ancona: average conditions: T=30.9 °C, RH=44.9 %, ws=2.0m/s and
558W/m2 solar radiation. Max Tdrop=8.2 °C, average Tdrop=2.8 °C
Rome: average conditions: T=29.0 °C, RH=50.2 %, ws=0.7m/s and
394W/m2 solar radiation. Max Tdrop=5.9 °C, average Tdrop=1.6 °C
Lorentzian vertical distribution for cooling and humidification, peaking
at about 0.5m of the injection
n.s. n.s.
Table A7 Humidification and comfort results.
Ref Humidity gain Comfort premium
[33] n.s. n.s.
[57] absolute humidity increased from 19.2 to 20.0g/kg at the spray
location, and to 19.5g/kg about 2.5m behind,
0.2g/kg further increase for the farthest windward nozzle in the
multiple nozzle case
n.s.
[58] When RH=80 % incomplete evaporation and surface wetting at
a height of 0 to 0.25m above the ground. Max SET* reduction >0.5° C under the nozzle (30 °C/60 % scenario). Min SET*
reduction on 1-m adjacent plane at 1.5m (30 °C/80 % scenario)
[59] Only numerical estimation: max ΔRH=20 % at the height of 2m,
2-3.5 m behind the spray (vertically), max ΔRH=11 %
(horizontally), moistening on the radial direction extends for 3.5-
4 m (downwind flow case)
ΔRH=9.5 % and 6.5 % in the vertical and horizontal cross-
section along the axial direction respectively, no significant
effect at 7- 15 m. (upwind flow case)
n.s.
[68] n.s. comfort questionnaires (200): 80 % of answers among “very comfortable”,
“comfortable” or “a little comfortable”
[72] There was no detectable difference in weight after misting. n.s.
[60] RH over 70 % within 2-6m for 6m/s inlet velocity
RH=64-67 % within 2-6m for 10m/s inlet velocity
negligible ΔRH even in the semi-enclosed case (1-6 %)
n.s.
[71] Average ΔRH (4-day mean, 95 % confidence)=9.497±0.882 % Comfort questionnaires: 80 participants (20/day), no gender bias, college-age, 1 met
(sitting), average clo≈0.3, residency in the space>15min. ASHRAE scale (thermal
sensation) and Bedford scale (comfort sensation), both 7-point scales.
Majority of votes (under misting fan vs under non misting fan): 28 % slightly cool
vs 32 % neutral (thermal sensation), 69 % between comfortable and comfortably
cool vs 69 % between comfortable and comfortably warm (overall comfort), 83 %
vs 81 % just right RH or slightly humid (humidity sensation), 63 % vs 66 %
unaltered RH levels (humidity preference),
Outdoor ET*(UC BerkeleyThermal Comfort Program Version 1.03), semi-outdoor
PET and SET (RayMan version 1.2) computed for each respondent. Max
statistically significant (R2>0.78) SET and SET difference≈3.5 °C under the fan mist
for equal ET*. Thermal neutrality=31.8 °C (mist fan) and 31.0 °C (non-mist fan),
28.9 °C (with mist line) and 28.6 °C (without mist line)
Average number of Colony-forming units (mist vs no mist) in the three locations:
110-674-1060 CFU m-3 vs 71-400-150 CFU m-3 (for bacteria), 598-271-277 CFU
m-3 vs 550-233-206 CFU m-3 (for fungi)
[61] Saturation within 1m from the spray even in the hot-dry
condition in the EXPO site Max WBGT reduction=5.2 °C
[46] max ΔRH=1 % both horizontally and vertically for Nozzle A, 6
% horizontally and 2 % vertically for Nozzle B Effective Temperature (ET*) reduction=0.7 °C horizontally and 0.5 °C vertically
for Nozzle A, 1.6 °C horizontallyand 2.1 °C vertically for Nozzle B (almost equal
to Tdrop)
[54] n.s.
[73] absolute humidity increased to 1.7 g/kg (veranda spraying) day average SET* reduction of 0.9° C and a maximum of 1.4° C
[48] PA 2: ΔRH=5-26 % for increasing ON time. n.s.
[63] PA 1: ΔRH=9-36 % for increasing water flow rate, at 2m
horizontal distance
PA 2: ΔRH=7-28 % for decreasing height above the ground, at
2m horizontal distance
PA 3: ΔRH=34-45 % for decreasing water temperature, at 2m
horizontal distance
n.s.

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
[53] n.s. n.s.
[62] the change in the humidity ratio is approximately independent of
the inlet air-water temperature difference (<3 % at the outlet
plane), but heavily enhanced by lower relative air-water velocity
(+90 % in PA 3) and smaller mean diameter (+50 % in PA5.1).
Δw≈15 % for different spread.
No considerations for surface wetting
PA 1: max UTCI reduction≈1.5-7.5 °C for increasing temperature difference; PA 2:
max UTCI reduction≈3-7 °C for decreasing humidity ratio; PA 3: max UTCI
reduction≈1-7.5 °C for decreasing relative velocity; PA 4: max UTCI reduction≈2-
7.5 °C for decreasing water temperature; PA 5.1: max UTCI reduction=3.7-8.7 °C
for decreasing mean diameter; PA 5.2: max UTCI reduction=4-7 °C for increasing
spread
[49] n.s. n.s.
[74] ΔRH=20-25 % n.s.
[45] Skin resulted slightly wet after 10s spray PMV (1.2met, 0.35clo)=3.48-3.96, Thermal load (W/m
2
)=66.7-72.8. For both
indicators: max around 2pm, min at the beginning and end of the experiment. 141
Comfort questionnaires: average age=20; gender bias:70 % female; average
clo=0.37
Average thermal sensation (9-point scale) =+2.7 (hot) before misting, -0.35 (neutral)
after misting; average comfort sensation (7-point scale)=-1.41 (uncomfortable)
before misting, 1.82 (comfortable) after misting
Average wettedness sensation (7-point scale)=0.82 (slightly humid) before misting,
1.56 (humid) after misting
[67] n.s. n.s.
[50] n.s. n.s.
[34] ΔRH=5-10 % over 8 hours Skin heat flux increased by 17-26W/m
2
with mist+fan vs only fan
[69] RH<65 % Sweating ratio<90g/h
[56] PA 1: vapor mass fraction gain=0.8-3·10-3 for increasing flow
rate
PA 2: vapor mass fraction gain=2.3-3·10-3 for decreasing height
PA 1: max UTCI reduction≈1-5 °C for increasing flow rate
PA 2: max UTCI reduction≈3-5 °C for decreasing height
[52] Experimental results: ΔRH>40 % at 1m distance, >20 % at 2m,
≈15 % at 3m, ≈10 to 0 % between 4 and 10m
Numerical results: max ΔRH (close to the column)≈65 %
38 °C/40 % vs 30 °C/70 % scenarios: ΔRH (3MPa pressure, 3m
distance)=8.5 % vs 5.3 %, ΔRH (26 µm diameter, 3m
distance)=4.5 % vs 2.7 %, radius of +10 % humidified area=1.3-
3.1 m vs 1.0-2.3m for decreasing diameter, ΔRH (3m/s airflow,
3m distance)=7.8 % vs 4.8 %
ΔRH rate>5, 10 and 20 %/m for increasing pressure
ΔRH patterns similar to Tdrop's as for PA 4
n.s.
[65] ΔRH=19.85-47.2 % for PA 1
ΔRH=41 % for PA 2
ΔRH=28 % for PA 3
n.s.
[6] Max, min, average, 99th, 90th, 50th and 1th percentiles,
interquartile range and standard deviation given for each
monitoring day
Ancona: 50th percentile<70 % (always), <65 % (for77.8 % of
the time, only with fuzzy control). Max ΔRH rate=+13.4 %/min,
average ΔRH rate=+2.4 %/min.
Rome: 50th percentile<65 % (always). Maxima above 65 % inly
with fuzzy and no-stop. Max ΔRH rate=+5.4 %/min, average
ΔRH rate=+3.9 %/min.
Fuzzy vs on-off: Offset from neutrality<2 °C for 98 % vs 80 % of the time
[55] Ancona: max ΔRH=24.7 %, average ΔRH=7.7 %
Rome: max ΔRH=22.1 %, average ΔRH=6.8 %
8 comfort indicators: thermal sensation (7-point scale), overall comfort,
humidity/wind/sun sensation and preference vote, all 3-point scale. Bouden and
Ghrab met formulation, ASHRAE pumping effect formulation, min 15min
occupancy
332 questionnaires (211 in Ancona + 121 in Rome), slight male bias (max 20 %),
dominant age range: 21-35 in Ancona ( 64.5 %), 36-50 in Rome (34.5 %)
Thermal sensation vote=uncomfortably hot (81 % in Ancona, 50 % in Rome) out
the misted area, uncomfortably hot (7 % in Ancona, ≈0 % in Rome) in the misted
area. Overall comfort=uncomfortable (63 % in Ancona, 32 % in Rome) out the
misted area, uncomfortable (5 % in Ancona, 0 % in Rome) in the misted area
Significant greater satisfaction with sun, wind and humidity while under the mist.
PET, SET* and UTCI difference computed via Rayman Pro. Max, min, average,
99th percentile, interquartile range and standard deviation reported.
in Ancona (SVF=0.87): ΔPET (max, average)=11.8 °C, 4.3 °C; ΔSET* (max,
average)=10.7 °C, 3.5 °C; ΔUTCI (max, average)=7.9 °C, 2.8 °C
in Rome (SVF=0.48): ΔPET (max, average)=4.2 °C, 2.0 °C; ΔSET* (max,
average)=3.9 °C, 1.7 °C; ΔUTCI (max, average)=3.3 °C, 0.9 °C
References
[1] IPCC, Fifth Assessment Report (AR5), 2018.
[2] M. Dijst, E. Worrell, L. Böcker, P. Brunner, S. Davoudi, S. Geertman, R. Harmsen, M. Helbich,
A.A.M. Holtslag, M.P. Kwan, B. Lenz, G. Lyons, P.L. Mokhtarian, P. Newman, A. Perrels, A.P.
Ribeiro, J. Rosales Carreón, G. Thomson, D. Urge-Vorsatz, M. Zeyringer, Exploring urban
metabolism—Towards an interdisciplinary perspective, Resour. Conserv. Recycl. 132 (2018) 190–
203. doi:10.1016/j.resconrec.2017.09.014.

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
[3] H. Radhi, S. Sharples, Quantifying the domestic electricity consumption for air-conditioning due to
urban heat islands in hot arid regions, Appl. Energy. 112 (2013) 371–380.
doi:10.1016/j.apenergy.2013.06.013.
[4] T.N.T. Lam, K.K.W. Wan, S.L. Wong, J.C. Lam, Impact of climate change on commercial sector air
conditioning energy consumption in subtropical Hong Kong, Appl. Energy. 87 (2010) 2321–2327.
doi:10.1016/j.apenergy.2009.11.003.
[5] Y. Toparlar, B. Blocken, B. Maiheu, G.J.F. van Heijst, Impact of urban microclimate on summertime
building cooling demand: A parametric analysis for Antwerp, Belgium, Appl. Energy. 228 (2018)
852–872. doi:10.1016/j.apenergy.2018.06.110.
[6] G. Ulpiani, C. Di Perna, M. Zinzi, Water nebulization to counteract urban overheating : Development
and experimental test of a smart logic to maximize energy efficiency and outdoor environmental
quality, Appl. Energy. 239 (2019) 1091–1113. doi:10.1016/j.apenergy.2019.01.231.
[7] K. Deilami, M. Kamruzzaman, Y. Liu, Urban heat island effect: A systematic review of spatio-
temporal factors, data, methods, and mitigation measures, Int. J. Appl. Earth Obs. Geoinf. 67 (2018)
30–42. doi:10.1016/j.jag.2017.12.009.
[8] M. Santamouris, Using cool pavements as a mitigation strategy to fight urban heat island - A review
of the actual developments, Renew. Sustain. Energy Rev. 26 (2013) 224–240.
doi:10.1016/j.rser.2013.05.047.
[9] D.E. Bowler, L. Buyung-Ali, T.M. Knight, A.S. Pullin, Urban greening to cool towns and cities: A
systematic review of the empirical evidence, Landsc. Urban Plan. (2010).
doi:10.1016/j.landurbplan.2010.05.006.
[10] A.B. Besir, E. Cuce, Green roofs and facades: A comprehensive review, Renew. Sustain. Energy Rev.
(2018). doi:10.1016/j.rser.2017.09.106.
[11] B. Amor, D.E.Z. Lacheheb, Y. Bouchahm, Improvement of Thermal Comfort Conditions in an Urban
Space (Case Study: The Square of Independence, Sétif, Algeria), Eur. J. Sustain. Dev. 4 (2015) 407–
416. doi:10.14207/ejsd.2015.v4n2p407.
[12] C. Bartesaghi Koc, P. Osmond, A. Peters, Evaluating the cooling effects of green infrastructure: A
systematic review of methods, indicators and data sources, Sol. Energy. (2018).
doi:10.1016/j.solener.2018.03.008.
[13] Y. Qin, A review on the development of cool pavements to mitigate urban heat island effect, Renew.
Sustain. Energy Rev. (2015). doi:10.1016/j.rser.2015.07.177.
[14] M. Santamouris, A. Synnefa, T. Karlessi, Using advanced cool materials in the urban built
environment to mitigate heat islands and improve thermal comfort conditions, Sol. Energy. (2011).
doi:10.1016/j.solener.2010.12.023.
[15] H. Akbari, H.D. Matthews, Global cooling updates: Reflective roofs and pavements, in: Energy
Build., 2012. doi:10.1016/j.enbuild.2012.02.055.
[16] M. Santamouris, Cooling the cities - A review of reflective and green roof mitigation technologies to
fight heat island and improve comfort in urban environments, Sol. Energy. (2014).
doi:10.1016/j.solener.2012.07.003.
[17] A.M. Hunter, N.S.G. Williams, J.P. Rayner, L. Aye, D. Hes, S.J. Livesley, Quantifying the thermal
performance of green façades: A critical review, Ecol. Eng. (2014).
doi:10.1016/j.ecoleng.2013.12.021.
[18] H. Akbari, D. Kolokotsa, Three decades of urban heat islands and mitigation technologies research,
Energy Build. (2016). doi:10.1016/j.enbuild.2016.09.067.
[19] J.S. Lee, J.T. Kim, M.G. Lee, Mitigation of urban heat island effect and greenroofs, Indoor Built
Environ. (2014). doi:10.1177/1420326X12474483.
[20] G. Manteghi, H. Bin Limit, D. Remaz, Water Bodies an Urban Microclimate: A Review, Mod. Appl.

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
Sci. 9 (2015) 1. doi:10.5539/mas.v9n6p1.
[21] H. Baumeister, L.G. Nordrhein-westfalen, C. Hornberg, Evidence for the temperature-mitigating
capacity of urban blue space - A health geographic perspective, (2013).
doi:10.3112/erdkunde.2013.04.05.
[22] K.R. Gunawardena, M.J. Wells, T. Kershaw, Utilising green and bluespace to mitigate urban heat
island intensity, Sci. Total Environ. 584–585 (2017) 1040–1055. doi:10.1016/j.scitotenv.2017.01.158.
[23] M. Santamouris, L. Ding, F. Fiorito, P. Oldfield, P. Osmond, R. Paolini, D. Prasad, A. Synnefa,
Passive and active cooling for the outdoor built environment – Analysis and assessment of the cooling
potential of mitigation technologies using performance data from 220 large scale projects, Sol.
Energy. 154 (2017) 14–33. doi:10.1016/j.solener.2016.12.006.
[24] M. Taleghani, U. Berardi, The effect of pavement characteristics on pedestrians’ thermal comfort in
Toronto, Urban Clim. 24 (2018) 449–459. doi:10.1016/j.uclim.2017.05.007.
[25] A.S. Nouri, J.P. Costa, M. Santamouris, A. Matzarakis, Approaches to outdoor thermal comfort
thresholds through public space design: A review, Atmosphere (Basel). 9 (2018).
doi:10.3390/atmos9030108.
[26] J. Kim, Spray cooling heat transfer : The state of the art, 28 (2007) 753–767.
doi:10.1016/j.ijheatfluidflow.2006.09.003.
[27] A. Berman, Extending the Potential of Evaporative Cooling for Heat-Stress Relief, J. Dairy Sci. 89
(2006) 3817–3825. doi:10.3168/jds.S0022-0302(06)72423-7.
[28] I. Sanchez, J.M. Faci, N. Zapata, The effects of pressure , nozzle diameter and meteorological
conditions on the performance of agricultural impact sprinklers, Agric. Water Manag. 102 (2011) 13–
24. doi:10.1016/j.agwat.2011.10.002.
[29] J.R. Mawhinney, J.K. Richardson, A Review of Water Mist Fire Suppression Research and
Development, 1996, Fire Technol. 33 (1997) 54–90. doi:10.1023/A:1015322428719.
[30] Architectes Trevelo & Viger-Kohler, Urbanistes, TVK. Place de la Republique. In TVK/Press Kit—
Place de la Republique, Paris, France, 2013.
[31] M. Knuijt, One Step Beyond. Open Space., Topos. 85 (2013) 60–67.
[32] J. Nunes, I. Zolio, N. Jacinto, A. Nunes, T. Campos, M. Pacheco, D. Fonseca, Misting-cooling
systems for microclimatic control in public space, Nunes, J., Ed. (2013) 1–16.
[33] H. Barrow, C.W. Pope, Droplet evaporation with reference to the effectiveness of water-mist cooling,
Appl. Energy. 84 (2007) 404–412. doi:10.1016/j.apenergy.2006.09.007.
[34] C. Farnham, L. Zhang, J. Yuan, T. Mizuno, K. Emura, A. Ashraful, Cooling Effect of a Mist Fan for
Large Indoor Spaces, in: Proc. 9th Wind. Conf. Wind. UK, 2016.
[35] L.A. Dombrovsky, V.P. Solovjov, B.W. Webb, Attenuation of solar radiation by a water mist from
the ultraviolet to the infrared range, J. Quant. Spectrosc. Radiat. Transf. 112 (2011) 1182–1190.
doi:10.1016/j.jqsrt.2010.08.018.
[36] J.J. Nijdam, S.H. Stårner, T.A.G. Langrish, An experimental investigation of droplet evaporation and
coalescence in a simple jet flow, Exp. Fluids. 37 (2004) 504–517.
[37] A. Haeussermann, E. Hartung, T. Jungbluth, E. Vranken, J.-M. Aerts, D. Berckmans, Cooling effects
and evaporation characteristics of fogging systems in an experimental piggery, Biosyst. Eng. 97
(2007) 395–405.
[38] H.R. Pruppacher, J.D. Klett, Microphysics of Clouds and Precipitation, 1997.
[39] S.S. Kachhwaha, P.L. Dhar, S.R. Kales, Experimental studies and numerical simulation of
evaporative cooling of air with a Horizontal parallel flow, Int. J. Heat Mass Transf. (1998).
[40] S.S. Kachhwaha, P.L. Dhar, S.R. Kale, Experimental studies and numerical simulation of evaporative
cooling of air with a water spray - II. Horizontal counter flow, Int. J. Heat Mass Transf. 41 (1998)

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
465–474. doi:10.1016/S0017-9310(97)00131-2.
[41] R. Sureshkumar, S.R. Kale, P.L. Dhar, Heat and mass transfer processes between a water spray and
ambient air – I . Experimental data, 28 (2008) 349–360. doi:10.1016/j.applthermaleng.2007.09.010.
[42] R. Sureshkumar, S.R. Kale, P.L. Dhar, Heat and mass transfer processes between a water spray and
ambient air - II. Simulations, Appl. Therm. Eng. 28 (2008) 361–371.
doi:10.1016/j.applthermaleng.2007.09.015.
[43] R. Sureshkumar, P.L. Dhar, S.R. Kale, Effects of spray modeling on heat and mass transfer in air-
water spray systems in parallel flow, Int. Commun. Heat Mass Transf. 34 (2007) 878–886.
doi:10.1016/j.icheatmasstransfer.2007.03.001.
[44] G. ASHRAE, Atlanta, ASHRAE 55 Thermal Environmental Conditions for Human Occupancy,
(2004).
[45] C. Farnham, K. Emura, T. Mizuno, Evaluation of cooling effects: Outdoor water mist fan, Build. Res.
Inf. 43 (2015) 334–345. doi:10.1080/09613218.2015.1004844.
[46] C. Farnham, M. Nakao, M. Nishioka, M. Nabeshima, T. Mizuno, Study of mist-cooling for semi-
enclosed spaces in Osaka, Japan, Procedia Environ. Sci. 4 (2011) 228–238.
doi:10.1016/j.proenv.2011.03.027.
[47] C. Pickering, J. Byrne, The benefits of publishing systematic quantitative literature reviews for PhD
candidates and other early-career researchers, 4360 (2014). doi:10.1080/07294360.2013.841651.
[48] A. Atieh, S. Al Shariff, Solar energy powering up aerial misting systems for cooling surroundings in
Saudi Arabia, Energy Convers. Manag. 65 (2013) 670–674. doi:10.1016/j.enconman.2011.10.031.
[49] H. Montazeri, B. Blocken, J.L.M. Hensen, Evaporative cooling by water spray systems: CFD
simulation, experimental validation and sensitivity analysis, Build. Environ. 83 (2015) 129–141.
doi:10.1016/j.buildenv.2014.03.022.
[50] S. Joshi, S. Gomekar, S. Pokle, Design of solar powered mist cooling system for a typical semi-
outdoor area in Nagpur, (2016).
[51] M.C. Peel, B.L. Finlayson, T.A. McMahon, Updated world map of the Köppen-Geiger climate
classification, Hydrol. Earth Syst. Sci. Discuss. 4 (2007) 439–473.
[52] C. Huang, J. Cai, Z. Lin, Q. Zhang, Y. Cui, Solving model of temperature and humidity profiles in
spray cooling zone, Build. Environ. 123 (2017) 189–199. doi:10.1016/j.buildenv.2017.06.043.
[53] C. Farnham, M. Nakao, M. Nishioka, M. Nabeshima, T. Mizuno, Effect of Water Temperature on
Evaporation of Mist Sprayed From a Nozzle, J. Heat Isl. Inst. Int. 10 (2015) 35–44.
[54] M. Kojima, K. Nakashima, S. Management, A Study of Mist Spraying System by Urban
Transportation, Des. Innov. Value Towar. a Sustain. Soc. Springer, Dordr. (2012) 720–723.
[55] G. Ulpiani, E. Di Giuseppe, C. Di Perna, M. D’Orazio, M. Zinzi, Thermal comfort improvement in
urban spaces with water spray systems: Field measurements and survey, Build. Environ. 156 (2019)
46–61. doi:https://doi.org/10.1016/j.buildenv.2019.04.007.
[56] H. Montazeri, Y. Toparlar, B. Blocken, J.L.M. Hensen, Simulating the cooling effects of water spray
systems in urban landscapes: A computational fluid dynamics study in Rotterdam, The Netherlands,
Landsc. Urban Plan. 159 (2017) 85–100. doi:10.1016/j.landurbplan.2016.10.001.
[57] H. Yamada, G. Yoon, M. Okumiya, H. Okuyama, Study of cooling system with water mist sprayers:
Fundamental examination of particle size distribution and cooling effects, Build. Simul. 1 (2008)
214–222. doi:10.1007/s12273-008-8115-y.
[58] G. Yoon, H. Yamada, M. Okumiya, Study on a cooling system using water mist sprayers; system
control considering outdoor environment., Korea-Japan Jt. Symp. Human-Environment Syst. Cheju,
Korea, p. 4. (2008).
[59] W. Jun-feng, T. Xin-cheng, Experimental Study and Numerical Simulation on Evaporative Cooling of

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
Fine Water Mist in Outdoor Environment, 2009 Int. Conf. Energy Environ. Technol. (2009) 156–159.
doi:10.1109/ICEET.2009.44.
[60] J. Wang, X. Tu, Z. Wang, J. Huang, Application and Numerical Simulation on Water Mist Cooling
for Urban Environment Regulation, in: Lect. Notes Comput. Sci., 2010. doi:10.1007/978-3-642-
15859-9.
[61] C. Huang, D. Ye, H. Zhao, T. Liang, Z. Lin, H. Yin, Y. Yang, The research and application of spray
cooling technology in Shanghai Expo, Appl. Therm. Eng. 31 (2011) 3726–3735.
doi:10.1016/j.applthermaleng.2011.03.039.
[62] H. Montazeri, B. Blocken, J.L.M. Hensen, CFD analysis of the impact of physical parameters on
evaporative cooling by a mist spray system, Appl. Therm. Eng. 75 (2015) 608–622.
doi:10.1016/j.applthermaleng.2014.09.078.
[63] M.S. Mahmoud, Experimental Study to Evaluate Mist System Performance, Int. J. Innov. Res. Adv.
Eng. 2 (2014) 41–48.
[64] A. Lefebvre, Atomization and Sprays, Boca Raton, n.d. doi:https://doi.org/10.1201/9781315120911.
[65] K. Zheng, M. Ichinose, N.H. Wong, Parametric study on the cooling effects from dry mists in a
controlled environment, Build. Environ. 141 (2018) 61–70. doi:10.1016/j.buildenv.2018.05.053.
[66] ASHRAE, Ashrae handbook: Fundamentals (Inch-pound ed.), Atlanta, 2013.
[67] J.-H. Kim, M.-H. Lee, Y.-H. Yoon, A Verification Study on the Temperature reduction Effect of
Water Mist Injection, Int. J. Korea Inst. Ecol. Archit. Environ. 15 (2015) 5–14.
[68] T. Ishii, M. Tsujimoto, G. Yoon, M. Okumiya, Cooling System with Water Mist Sprayers for
Mitigation of Heat-island Drymist system : Uchimizu, Seventh Int. Conf. Urban Clim. (2009) 2–3.
[69] S.A. Domínguez, F.J. Sánchez, The effect of evaporative cooling techniques on reducing urban heat,
in: Urban Clim. Mitig. Tech., Routledge, 2016: pp. 127–144.
[70] P.O. Fanger, Thermal comfort, analysis and application in environmental engineering, New York,
McGrew-Hill. (1972).
[71] N.H. Wong, A.Z.M. Chong, Performance evaluation of misting fans in hot and humid climate, Build.
Environ. 45 (2010) 2666–2678. doi:10.1016/j.buildenv.2010.05.026.
[72] C. Farnham, M. Nakao, M. Nishioka, M. Nabeshima, Quantification of the Effect of Cooling Mists on
Individual Thermal Comfort, in: Seventh Int. Conf. Urban Clim., 2009.
[73] D. Narumi, K. Shigematsu, Y. Shimoda, Effect of the Evaporative Cooling Techniques by Spraying
Mist Water on Reducing Urban Heat Flux and Saving Energy in Apartment House, J. Heat Isl. Inst.
Int. 7 (2012) 175–181. http://heatisland2009.lbl.gov/docs/221000-narumi-doc.pdf.
[74] H. Esen, O. Tuna, Investigation of photovoltaic assisted misting system application for arbor
refreshment, Int. J. Photoenergy. 2015 (2015). doi:10.1155/2015/748219.
[75] J. Nunes, I. Zoilo, N. Jacinto, A. Nunes, T. Campos, M. Pacheco, D. Fonseca, Misting-cooling
systems for microclimatic control in public space., Available through
<http//Www.Proap.Pt/847/Misting-Cooling-Systemsfor- Microclim. (2016).
[76] G.R. Hunt, P.P. Linden, The fluid mechanics of natural ventilation-displacement ventilation by
buoyancy-driven flows assisted by wind, Build. Environ. 34 (1999) 707–720.
[77] A.S. Nouri, J.P. Costa, Addressing thermophysiological thresholds and psychological aspects during
hot and dry mediterranean summers through public space design: The case of Rossio, Build. Environ.
118 (2017) 67–90. doi:10.1016/j.buildenv.2017.03.027.
[78] S. Yu, Water spray geoengineering to clean air pollution for mitigating haze in China’s cities,
Environ. Chem. Lett. 12 (2014) 109–116. doi:10.1007/s10311-013-0444-0.
[79] Q. Zhao, W. Yao, C. Zhang, X. Wang, Y. Wang, Study on the influence of fog and haze on solar
radiation based on scattering-weakening effect, Renew. Energy. (2018).

Ulpiani (2019). Applied Energy - https://doi.org/10.1016/j.apenergy.2019.113647
doi:10.1016/j.renene.2018.11.027.
[80] Y. Chartier, J. V Lee, K. Pond, S. Surman-Lee, Legionella and the prevention of legionellosis, World
Health Organization, 2007.
[81] H. Takebayashi, A Simple Method to Evaluate Adaptation Measures for Urban Heat Island,
Environments. 5 (2018) 70.
[82] M. Beermann, B.; Berchtold, M.; Baumüller, J.; Gross, G.; Kratz, Städtebaulicher Rahmenplan
Klimaanpassung für Die Stadt Karlsruhe (Teil II), Karlsruhe, Germany, 2014.
https://fachdokumente.lubw.baden-wuerttemberg.de/servlet/is/109723/U83-W03-
N11.pdf?command=downloadContent&filename=U83-W03-N11.pdf.
[83] T.A.L. Martins, L. Adolphe, M. Bonhomme, F. Bonneaud, S. Faraut, S. Ginestet, C. Michel, W.
Guyard, Impact of Urban Cool Island measures on outdoor climate and pedestrian comfort:
Simulations for a new district of Toulouse, France, Sustain. Cities Soc. 26 (2016) 9–26.
doi:10.1016/J.SCS.2016.05.003.
[84] A. Chatzidimitriou, P. Liveris, M. Bruse, L. Topli, Urban Redevelopment and Microclimate
Improvement : A Design Project in Thessaloniki , Greece, PLEA 2013 Sustain. Archit. a Renew.
Futur. (2013).
[85] R.A.W. Albers, P.R. Bosch, B. Blocken, A. Van Den Dobbelsteen, L.W.A. Van Hove, T.J.M. Spit, F.
de Ven, T. Van Hooff, V. Rovers, Overview of challenges and achievements in the climate adaptation
of cities and in the Climate Proof Cities program, (2015).