Analyzing the Impact of Urban Planning and Building Typologies in Urban Heat Island Mitigation

Abstract

Urban and building typologies have a serious impact on the urban climate and determine at large the magnitude of the urban overheating and urban heat island intensity. The present study aims to analyze the impact of various city typologies and urban planning characteristics on the mitigation of the urban heat island. The effect of the building height, street width, aspect ratio, built area ratio, orientation, and dimensions of open spaces on the distribution of the ambient and surface temperature in open spaces is analyzed using the Sydney Metropolitan Area as a case study for both unmitigated and mitigated scenarios. Fourteen precincts are developed and simulated using ENVI-met the simulation tool. The ambient temperature, surface temperature, and wind speed are extracted. The parameter ‘Gradient of the Temperature Decrease along the Precinct Axis’ (GTD) is introduced to study the cooling potential of the various precincts. In the mitigated precincts, the GTD ranges between 0.01 K/m to 0.004 K/m. In the non-mitigated precincts, the GTD ranges between 0.0093 K/m to 0.0024 K/m. A strong correlation is observed between the GTD of all the precincts, with and without mitigation, and their corresponding average aspect ratio, (Height of buildings to Width of streets). The higher the aspect ratio of the precinct, the lower the cooling potential. It is also observed that the higher the Built Area Ratio of the precincts, the lower the cooling contribution of the mitigation measures.

1. Introduction

The heat island effect is the most documented phenomenon of climate change. The phenomenon has been known for almost a century and is related to higher urban temperatures compared to the adjacent suburban and rural areas. Higher urban temperatures are due to the positive thermal balance of urban areas caused by several factors such as [1,2,3,4,5]:

• The release of anthropogenic heat;
• The excess storage of solar radiation by the city structures;
• The lack of green spaces and cool sinks;
• The non-circulation of air in urban canyons;
• The reduced ability of the emitted infrared radiation to escape in the atmosphere.

Summer urban heat islands with daytime average air temperatures 4 °C higher than the surrounding rural areas are present in many cities around the world.

Urban overheating has a serious impact on the energy consumption of buildings and the peak electricity demand, while it increases the concentration of urban pollutants and seriously affects the levels of heat-related mortality and morbidity, [6,7,8,9].

In Sydney, the phenomenon has been studied in depth [10,11,12,13] with a maximum recorded gradient of peak temperature of 9 °C. Such a high magnitude of the urban heat island is due to the existence of two synoptic meteorological systems affecting the local climate. In particular, sea breeze contributes to lower ambient temperature in the eastern coastal part of the city, while warm western winds from the desert heat the western part of the city, [14,15].

To counterbalance the impact of urban overheating, important heat mitigation technologies and techniques have been developed [16,17,18]. Mitigation technologies aim to decrease the maximum possible heat gains and maximize the heat losses in a city to decrease the magnitude of the urban overheating. The increase of the urban albedo as well as the use of additional greenery (green roofs, walls, green parks, pocket parks, etc.) [19,20,21], evaporative sources, blue infrastructure [22,23] low-temperatureperature heat sinks seems to be among the most efficient ones [24,25,26,27].

Urban and building typologies have a serious impact on the urban climate and determine at large the magnitude of the urban overheating [13,28,29]. The urban density, height of buildings, size of streets, aspect ratio, and size of open spaces affect the heat and solar gain while determining the heat losses through radiation and convection [30,31]. While numerous studies are available on the impact of the various mitigation technologies, i.e., an increase of urban albedo and greenery, evaporative sources, etc., on urban climate, very little is known about the heat impact of the main landscape parameters as well as their impact on the cooling potential of the main mitigation technologies [32,33]. Therefore, although the impact of the various mitigation techniques on the urban heat island has been studied, the role of the spatial distribution, typologies, and urban forms play a significant role in the cooling potential of the various technologies. The present study aims study is to analyze the impact of building height, street width, aspect ratio, built area ratio, orientation, and dimensions of open spaces on ton distribution of the ambient and surface temperature as well as on thermal comfort in open spaces. Moreover, the study analyses the impact of cooling potential of the most commonly used urban heat island mitigation technologies. The overall study is based on real urban neighborhoods with significant overheating problems.

The paper is structured in six more sections. Section 2 includes the methodology and approach of the research while Section 3 includes the description of the urban context in Sydney. The modeling procedures and simulation results are included in Section 4 and Section 5 respectively while the analysis of results and conclusions are incorporated in Section 6 and Section 7.

2. Materials and Methods

The present study aims to analyze the impact of various city typologies and urban planning characteristics on the mitigation of the urban heat island. For that purpose, Syndey Metropolitan Area is selected as the case study area for the following reasons:

• The neighborhoods represent real case areas.
• There is a good representation of open and compact typologies defined officially by the Australian Association of Planners.

There is a 100% representation of the specific urban area characteristics.

Sydney’s climatic conditions are humid subtropical (Köppen: Cfa) in Eastern Australia [34]. In the Sydney Metropolitan Area (SMA), seven urban areas with diverse urban fabric characteristics are selected (see

Table 1. Building Types.

Type	Description	Figure	No of Storeys	People per Hectare	Location	Amenities
T1: Single Dwellings	Single dwellings areas include houses, terrace houses, dual occupancies, and semi-detached dwellings.		1–2	30–100	Suburban areas	Local parks, distant shops
T2: Low Rise	Low rise housing typically involves townhouses/terrace housing or small-scale buildings with street-level retail shops and cafes with residential apartments above.		3–4	70–200	Close to village centers, along transport corridors	Parks, shops
T3: Low/Medium Rise	Low/medium-rise housing involves apartment buildings sometimes with cafes or small shops at the ground level.		5–6	150–300	Close to town centers and urban renewal areas	Park, shops, swimming pools
T4: Medium Rise	Medium rise hosing involves apartment buildings, sometimes with cafes or medium shops at the ground level.		7–8	250–400	Urban corridors, urban renewal areas, city centers	Parks, shops, gyms, child cares, swimming pools, buses, trains
T5: Medium/High Rise	Medium /high-rise housing involves apartment buildings, sometimes with retail, medium, and large shops at the ground level.		9–12	300–400	Urban corridors, urban renewal areas, near railway stations	Parks, supermarket, gyms, child cares, swimming pools, buses, trains, theatres/cinemas
T6: High Rise 1	High rise housing 1 comprises standalone apartment buildings and mixed-use buildings that incorporate retail shops and/or commercial uses on the lower levels.		13–25	400–800	Near transport nodes, urban renewal areas, city centers	Parks, jobs, supermarkets, gyms, child cares, swimming pools, buses, trains, theatres/cinemas
T7: High Rise 2	High rise housing 1 comprises standalone residential and mixed-use towers that incorporate retail shops and/or commercial uses on the lower levels.		25+	600–1200	City centers, near railway stations	Parks, jobs, supermarkets, gyms, clubs, swimming pools, buses, theatres/cinemas, major railway st.

and

Table 2. District Precincts.

Open Arrangements
Type	OT1: Open Single Dwellings	OT2: Open Low Rise	OT3: Open Low/Medium Rise	OT4: Open Medium Rise	OT5: Open Medium/High Rise	OT6: Open High Rise 1	OT7: Open High Rise 2
Figure
Region in Sydney
Location	Normanhurst	Kooloora	Rosebery	Raleigh Park	Parramatta	Waterloo	Syndey Olympic Park
No. Storeys	1	3	5–6	8	9–12	18	30–35
Building height	4–8	5–12	9–18	25–30	21–42	60	100–130
Street width	25–35	25–30	15–20	35–45	20–30	45–55	30–70
Building size	200–350	250–500	1000–2000	1000–1500	1000–1500	1000	1000–1500
Compact arrangements
Type	CT1: Compact single dwellings	CT2: Compact Low rise	CT3: Compact Low/Medium Rise	CT4: Compact Medium rise	CT5: Compact Medium/high rise	CT6: Compact High rise 1	CT7: Compact High rise 2
Figure
Region in Sydney
Location	Kellyville	Epping	Meadowback	Harold Park	Mascot	Wentworth Point	Chatswood
No. Storeys	1–2	3–4	6	7–8	10–12	15–25	35–40
Building height	4–12	8–12	12–18	21–30	28–42	60–75	130–145
Street width	25–30	15–30	15–20	20–25	25–30	25–30	25–40
Building size	150–300	650–1000	1000–2000	1000–1500	4000–6000	1500–2000	1000–1500

). The residential areas have been categorized into 14 residential precincts to support microclimatic analysis. The Local Climates Zones (LCZs) approach has been followed for selecting the typologies. The LCZs is a standardized approach, widely implemented for the analysis of urban areas’ overheating [35].

Seven residential building types (T1-T7, Table 1) have also been selected, representative of the area under investigation in this study. The building types are categorized according to height based on the typologies defined by the Department of Planning and Environment in New South Wales (NSW).

The seven building types are arranged under two urban design scenarios including one low density-open scenario (open represented by the letter ‘O’) and one high density-compact scenario (compact represented by the letter ‘C’). Therefore 14 precincts are investigated corresponding to OT1-OT7 and CT1-CT7.

Current climate and land use as well as future climate and land use in 2050 have been simulated with mesoscale models. In addition, microscale climatic simulations for all the climatic scenarios are performed with hourly climatic files [36].

Microscale simulations are performed to predict the distribution of ambient temperature, wind speed, surface temperature, and outdoor thermal comfort, in the 14 precincts. The simulations are run for representative summer days based on the 2050 climate scenario. Two scenarios are simulated: (a) Full mitigation scenario, including implementation of greenery, evaporation, and cool materials in all precincts, and (b) Non-mitigation scenario where no mitigation measures are implemented.

The simulation results of the peak daytime temperature conditions are analyzed for assessing the distribution of the main climatic parameters and the outdoor thermal comfort. Moreover, the urban typologies characteristics are analyzed versus their cooling potential

The methodological steps are the following:

• The building typologies and urban precincts are selected to fully represent the urban characteristics and neighborhoods.
• The precincts are modeled using ENVI-met for the mitigated and unmitigated scenarios.
• The ambient temperature, surface temperature, outdoor comfort indices, and wind flow regimes for both mitigated and unmitigated scenarios are extracted and compared.
• The cooling potential is then analyzed by introducing a specific parameter called ‘Gradient of the Temperature Decrease along the Precinct Axis’ (GTD).
• The GTD is evaluated versus the flow through open areas, the aspect ration (H/W), and the Built Area Ratio.

3. Buildings and Urban Context

Most urban precincts are not completely homogenous, and the different regions are composed of buildings with varying heights arranged in distinct patterns and densities thus providing various open spaces’ typologies. Accordingly, the seven building typologies (Table 1) are further categorized into two arrangement types (open and compact) as mentioned in the previous section. The residential precincts are presented in Table 2. The further separation is based on four basic characteristics (no. stories, building height, street width, and building size).

4. Modeling Procedures

The simulations of the unmitigated and mitigated scenarios based on future climate (2050) were performed with the software ENVI-met V4.4.2 [37,38]. This program is a reliable tool for the simulation of the main climatic parameters’ distribution in the urban environment. The program uses a three-dimensional microclimate model with a resolution that allows for simulating the surface, plant, and air interactions in urban environments. It supports microscale modeling with the following characteristics:

• A typical horizontal resolution from 0.5 to 5 m
• A typical time frame of 24 to 48 h
• A time step of 1 to 5 s.

ENVI-met solves the Reynolds-averaged non-hydrostatic Navier-Stokes equations for each grid in space and for each time step.

The spatial resolution used in the simulations of the present study is 1.5 m horizontally. The area has been represented with 150 × 150 × 30 (x − y − z) cells. Each cell size is dx = 1.5 m, dy = 1.5 m, and base dz = 0.5 m. The grid at the z-axis is telescopic with a thicker cell near the ground, allowing a better accuracy for edge effects.

Future climate conditions for 2050 are predicted with mesoscale models using the Weather Research and Forecasting (WRF) Model. The outputs of the mesoscale models are used as inputs to the ENVI-met microscale models.

For the ENVI-met simulations, two approaches are used: full forcing and simple forcing. Full forcing is used for the simulation of air temperature, relative humidity, and solar radiation. This means that the diurnal variation of the atmospheric boundary conditions and the incoming radiation is defined in each simulation step. The initial conditions for wind speed and direction are:

• Wind speed: 2.5 m/s;
• Wind direction: 250°;
• The start time and date of simulation: 18:00 21/2/2050;
• The end time and date of simulation: 00:00 23/2/2050 correspond to summer conditions in Sydney.

The buildings’ characteristics are tabulated in

Table 3. The buildings’ construction characteristics.

Code	Name	Construction
		Outside Layer	1st Layer	2nd Layer
000000	Default wall-moderate insulation	0100PL (1cm)	0100IN (11 cm)	0100CO (6 cm)
0100Q2	CoolRoof-moderate insulation	0100Q1 (1 cm)	0100FE (11 cm)	0100F3 (6 cm)

while the buildings’ materials properties are tabulated in

Table 4. The building materials’ properties are used for all typologies.

Code	Name	Absorption	Reflection	Emissivity	Specific Heat (J/(kgK))	Thermal Conductivity (w/(mK)	Density (kg/m3)
0100PL	Default Plaster	0.50	0.50	0.90	850	0.60	1500
0100Q1	CoolPaint	0.30	0.70	0.90	830	0.84	1856
0100IN	Default Insulation	0.50	0.50	0.90	1500	0.07	400
0100CO	Default Concrete	0.50	0.50	0.90	850	1.60	2220
0100F3	Moderate insulation	0.42	0.45	0.90	1033	1.00	1687

and

Table 5. The urban surfaces albedo and emissivity.

Code	Name	Albedo	Emissivity	Used in OT	Used in CT
0100ST	Asphalt Road	0.2	0.9		2
0100PD	Concrete Pavement Dark	0.20	0.90	1-6-7	1-3
0100PG	Concrete Pavement Gray	0.50	0.90	1-3-4-5-6-7	1-2-3-7
0100PL	Concrete Pavement Light	0.80	0.90	1-3-4-5-7	1-2-3-5-7
0100Q3	Cool Pavement	0.50	0.90	1-2-3-4-5-6-7	1-2-3-4-5-6-7
0100Q5	Cool Asphalt Road	0.55	0.90	1-2-3-4-5-6-7	1-2-3-4-5-6-7
0100KK	Brick road (red stones)	0.3	0.9		2-7
0100GG	Dark Granit Pavement	0.3	0.9		3
0100WW	Deepwater (swimming pools)	0.00	0.96	1-4	1-

. The vegetation types are included in

Table 6. The vegetation characteristics.

Code	Name	OT	CT
0100XX	Grass 25 cm aver. Dense	1-2-3-4-5-6	1-2-3-4-5-6-7
0100H2	Hedge dense, 2 m	1-3-4-5-6	7
0100H4	Hedge dense, 4 m	1
01ALDM	Conic, large trunk, dense, medium (15 m)		4
01ALDL	Conic, large trunk, dense, large (25 m)		4
01ALDS	Conic, large trunk, dense, small (5 m)		4
01CMSS	Cylindric, medium trunk, sparse, small (5 m)		5
01CSSS	Cylindric, small trunk, sparse, small (5 m)		5
01CLSS	Cylindric, large trunk, sparse, small (5 m)		5
01CLDM	Cylindric, large trunk, dense, medium (15 m)	1-3-4-5-6	1-2-3-5-6-7
01CLDS	Cylindric, large trunk, dense, small (5 m)	1-3-4	1-2-3-5-6-7
01CSDS	Cylindric, small trunk, dense, small (5 m)	3-4-6	1
01CLDL	Cylindric, large trunk, dense, large (25 m)	3-4	1-2-3-7
01CSDM	Cylindric, small trunk, dense, medium (15 m)	3-4-6	1
01CMDM	Cylindric, medium trunk, dense, medium (15 m)	1-4-5-6
01CLDL	Cylindric, large trunk, dense, large (25 m)	1-2-6	7
01HLDL	Heart-shaped, large trunk, dense, large (25 m)	1	2
01PSDS	Palm, small trunk, dense, small (5 m)		7
01CMDS	Cylindric, medium trunk, dense, small (5 m)	1-4-5	7
01PSDS	Palm, small trunk, dense, small (5 m)	1
01PLDS	Palm, large trunk, dense, small (5 m)	2-4-6	3-7
01PLDM	Palm, large trunk, dense, medium (15 m)	1-4	6-7
01PLDL	Palm, large trunk, dense, large (25 m)	2	6
01OMDS	Cylindric, medium trunk, dense, small (5 m)		6
01OLDM	Cylindric, large trunk, dense, medium (15 m)	2-4	6
01CMDL	Cylindric, medium trunk, dense, large (25 m)	1-3-4-6
01OLDS	Cylindric, large trunk, dense, small (5 m)	2	6
01OLDL	Cylindric, large trunk, dense, large (25 m)	2	6
01SLDS	Spherical, large trunk, dense, small (5 m)	6
01SMSL	Spherical, medium trunk, sparse, large (25 m)	6
01SMDS	Spherical, medium trunk, dense, small (5 m)	6
01CLSM	Cylindric, large trunk, sparse, medium (15 m)	6
01SMDM	Spherical, medium trunk, dense, medium (15 m)	6

and the ENV-met models in

Table 7. The models developed in ENVI-met for all typologies and the mitigated scenarios.

Open Precincts
Type	OT1: Open Single Dwellings	OT2: Open Low Rise	OT3: Open Low/Medium Rise
Envimet Model
	OT4: Open Medium rise	OT5: Open Medium/high rise	OT6: Open High rise 1	OT7: Open High rise 2
ENVI-met Model
Compact precincts
Type	CT1: Compact single dwellings	CT2: Compact Low rise	CT3: Compact Low/Medium Rise	CT4: Compact Medium rise
ENVI-met Model
	CT5: Compact Medium/high rise	CT6: Compact High rise 1	CT7: Compact High rise 2
ENVI-met Model

.

5. Simulation Results

Simulations are performed for both the unmitigated and mitigated cases. The basic values of the urban configuration and mitigated scenarios are tabulated in

Table 8. Basic Values of the urban configuration used for the base run and mitigation simulations.

Urban Canopy Parameters				Base Run and Unmitigated Scenario		Mitigated Scenario Values
Urban Categories	Cat	Building Height	Urban Fraction	Roof Albedo	Road Albedo	Roof Albedo	Roof Albedo
Commercial Business Dist.	CBT	28	0.95	0.15	0.08	0.6	0.6
High Density	HD	13	0.66	0.15	0.08	0.6	0.6
Medium Density	MD	6	0.62	0.15	0.08	0.6	0.6
Low Density	LD	4	0.55	0.15	0.08	0.6	0.6
Industrial	IN	6	0.60	0.6	0.08	0.6	0.6

. The simulation results are included in the following subsections.

5.1. Simulation Results for the Unmitigated Cases

Simulations have been performed using an unmitigated weather file for the year 2050. Furthermore, the various ENVI-met model’s settings are set to the base case with no mitigation techniques. The urban canopy settings for all unmitigated simulations are: Road Albedo = 0.08, Roof Albedo = 0.15, a limited number of trees, and no water sprinklers in the precincts for each precinct, the air temperature, surface temperature, wind speed, and thermal comfort indices are extracted.

For all precincts, the ambient temperature ranges are tabulated in

Table 9. The results of the unmitigated cases for 14:00.

Precinct	Air Temperature Range (°C)	Maximum Wind Speed (m/s)	Surface Temperature Range (°C)	UTCI Range (°C)
CT1	31.9–35.3	2	26.6–57.8	35.0–43.6
CT2	31.8–34.9	2.8	25.8–55.9	34.6–43.2
CT3	31.7–35.3	2.3	27.4–57.0	34.5–43.7
CT4	32.2–35.6	3.2	27.7–57.0	34.5–43.8
CT5	32.0–35.0	4	26.1–56.3	33.8–43.4
CT6	31.0–34.6	3.4	24.4–55.6	31.9–41.5
CT7	31.5–34.4	5.1	24.5–55.4	30.9–41.5
OT1	32.1–34.7	1.8	25.5–56.2	35.2–43.4
OT2	31.9– 35.2	2.1	26.1–57.0	34.4–43.5
OT3	31.9–35.2	2.4	26.0–58.4	34.1–44.0
OT4	32.2–34.6	2.3	24.3–56.6	34.9–43.5
OT5	31.6–34.8	3.7	25.9–57.1	34.0–43.0
OT6	32.3–34.7	3.4	25.4–56.6	32.9–42.9
OT7	31.9–34.9	4.1	25.5–56.8	32.9–42.6

. For all CT cases, the thermal comfort is mostly improved in the shaded areas while there is moderate to strong heat stress. The wind speed is significantly reduced between the buildings.

Indicatively the air temperature distribution for CT1 and OT7 are depicted in Figure 1 and Figure 2 respectively. The air movement around buildings for CT5 is illustrated in Figure 3 and the surface temperature in Figure 4.

5.2. Simulation Results for the Mitigated Cases

The mitigated scenario is based on the 2050 land use and climate. It assumes the implementation of mitigation technologies in the whole Sydney area. The RCP4.5 Future Scenario [39,40] dataset has been used for the Urban Plan simulation. It includes (a) increased building density based on future projections (b) new Urban Growth Areas (c) increased Anthropogenic Heat Flux (d) 2 million irrigated trees planted in the “Third City” area and 3 million irrigated trees planted in the rest of Sydney (e) increased albedo (0.6) of urban impervious surfaces and (f) Water in the landscape. The mitigated scenario values are tabulated in Table 8.

Table 10. Statistical summary of the ambient temperature results at 14:00 for all precincts.

Ambient Temperature Statistical Results (°C)								Percentiles
Layout	TMax 1	TMin 1	Mean	Median	Std.	TMax 2	TMin 2	25	50	75	90	95
OT1	33	27.9	31	30.6	0.9	33	26.3	30.3	30.6	31.8	32.4	32.5
OT2	32.2	28.5	30.4	30.3	0.8	32.5	25.1	29.9	30.3	30.8	31.4	31.8
OT3	32.7	27.7	30.8	30.8	1	32.7	24.8	30	30.8	31.5	32.1	32.4
OT4	32.7	29.8	31.3	31.3	0.9	32.9	22.1	30.9	31.3	31.6	32.4	32.6
OT5	32.9	28.8	31.1	31	0.8	32.9	26.6	30.5	31	31.7	32.3	32.6
OT6	32.7	27.7	30.2	30.5	1.7	32.7	18.6	29.7	30.5	31.1	31.8	32
OT7	33	29.2	31.2	31.1	0.8	33.1	26.3	30.7	31.1	31.6	32.2	32.5
CT1	33.4	28.4	31.1	30.8	1	33.4	25.8	30.4	30.8	31.8	32.7	32.8
CT2	33	28	31.2	31.2	1	33	25.7	30.5	31.2	32.1	32.5	32.6
CT3	32.9	27.9	31	30.7	0.9	32.9	27.1	30.2	30.7	31.8	32.4	32.7
CT4	33.2	29.3	31.4	31.3	0.8	33.2	26.3	30.9	31.3	32	32.5	32.7
CT5	33.2	28.3	31.5	31.3	1	33.2	27.2	30.7	31.3	32.3	32.8	33
CT6	32.7	27.5	30.4	30.2	1	32.7	24.5	29.7	30.2	31.1	31.8	32
CT7	32.4	28.2	30.4	30.2	0.8	32.4	27.1	29.8	30.2	31	31.6	31.9

¹: The minimum and maximum excluding outliers. ²: The absolute minimum and maximum temperature data.

shows the statistical summary of the ambient temperature at 14:00. CT1 has the maximum ambient temperature followed by CT4 and CT5. The minimum ambient temperature is 27.5 °C in CT6. The lowest 25th percentile of temperature data is 29.7 °C, observed in CT6 and OT6. The 95th percentile of data is 33.0 °C in CT4. The average ambient temperature in CT layouts is higher than that in OT layouts.

The ambient temperature difference of the mitigated versus the unmitigated simulations in each precinct is reported in

Table 11. Ambient air temperature differences of the unmitigated versus the mitigated scenarios for all precincts.

	CT1	CT2	CT3	CT4	CT5	CT6	CT7	OT1	OT2	OT3	OT4	OT5	OT6	OT7
Mean	2.20	2.22	2.32	2.14	2.13	2.13	2.09	2.33	3.04	2.78	2.39	2.27	3.26	2.16
Std.	0.48	0.69	0.63	0.70	0.61	0.52	0.37	0.51	0.68	0.73	0.90	0.89	1.49	0.56
Minimum	1.42	1.41	1.64	1.10	1.57	1.37	1.54	1.48	1.88	1.62	1.21	0.93	1.69	0.81
Maximum	8.18	8.17	7.59	6.84	7.25	7.74	5.20	7.33	8.76	9.59	11.98	7.70	14.50	7.35

. The average reduction of ambient temperature ranges between 2.1 °C ± 0.4 and 3.3 °C ± 1.5°C. The minimum reduction of ambient temperature is observed in CT7 and the maximum in OT6.

The charts of the ambient temperature differences are depicted in Figure 5. The maximum air temperature differences occur close to the water sprays and these high-temperature differences extend beyond the point of water spray position depending on the wind direction and the typology examined.

The charts of the surface temperature differences are depicted in Figure 6. The areas shaded by buildings and trees present lower surface temperatures.

The ambient temperature, surface temperature, and thermal comfort results for the mitigated versus the unmitigated scenarios are similar to those extracted from other research published and used the ENVI-met model [41,42,43,44].

6. Analysis of Results and Discussion

Based on the simulation results a more detailed analysis is performed. The cooling potential of each of the 14 studied precincts is evaluated using the ‘Gradient of the Temperature Decrease along the Precinct Axis’ (GTD) parameter. The GTD parameter measures the average temperature decrease along the axis of the canyon that is closer to the wind direction.

In the present study, the angle of the wind speed direction (250°) with the X-axis is smaller than the angle with the Y-axis, thus the GTD parameter is calculated along the X-axis.

The climatic parameters in each precinct are calculated for a 224 × 224 cells grid. For each parameter, W, its average value is defined as W(X_i, Y_1-224), corresponding to one X cell and all the 224 Y cells of the same X value. For example, for the cell X = 3, the sum of the W values corresponding to cells with X = 3 and Y = 1 to 224, is calculated and divided by 224. In case cells do not include numerical values of the parameter W (if the area is covered by buildings), then the corresponding cells are not considered.

The GTD for the average distribution of ambient temperature, and wind speed, is calculated along the axis X. The GTD(x), is calculated as the average difference between the initial and the final value of the ambient temperature along the X-axis:

$$GTD\left(x\right)=\frac{T_{\left(averagex=1\right)}-T_{\left(averagex=224\right)}}{224}$$

(1)

GTD(x) counts for the temperature decrease along the X-axis per meter of length of the precinct and expresses the potential of the precinct to mitigate the ambient overheating along the axis closer to the wind direction. The procedure is followed for all precincts with mitigation and without mitigation.

Indicative results are shown for CT1 and OT3 in Figure 7 and Figure 8 respectively while all results are tabulated in

Table 12. The Gradient Temperature Decrease at the wind flow direction for all precincts.

Mitigated GTD across the Precincts (K/100 m)
1.1	0.95	0.90	0.87	0.84	0.84	0.83
CT1	OT6	OT1	OT3	CT6	OT2	CT2
0.80	0.77	0.70	0.68	0.58	0.42	0.4
OT4	CT3	OT5	CT4	CT5	CT7	OT7
Unmitigated GTD across the Precincts (K/100 m)
0.93	0.82	0.79	0.75	0.73	0.73	0.70
CT1	OT3	CT2	CT6	CT3	OT1	OT6
0.69	0.66	0.60	0.53	0.50	0.30	0.24
OT2	CT4	OT4	OT5	CT5	CT7	OT7

. For the mitigated scenarios the GTD varies between 0.01 K/m to 0.004 K/m. In the unmitigated scenarios, the GTD varies between 0.0093 K/m to 0.0024 K/m. The maximum expected temperature difference between precincts of about 40,000 m² with different layouts, building typologies, and open spaces types, where the same mitigation measures are implemented, may be close to 0.9 K, for a reference ambient temperature of 32 °C and a wind speed of about 2 m/s. Moreover, the maximum expected temperature difference between precincts of about 40,000 m², without mitigation, is close to 1.5 K, for reference ambient temperature of 33 °C, and wind speed of about 2 m/s.

The wind flow regimes of each precinct are depicted in Figure 9. For each precinct, the areas are divided into different zones that are either parallel or perpendicular to the wind speed. In Figure 9 the GTD is also included.

For CT1 the ambient temperature is strongly related to the wind speed. As the advection heat increases with higher wind speeds, the ambient temperature increases too. The sections of zone 1, in both CT2 and CT3, provide low advection and good solar protection thus highly contributing to increasing the cooling gradient. Zones 2 and 3 of CT2 are parallel to the wind direction; they present however a significant cooling rate that is attributed to the intensive planting mitigation. The CT3 zone 2 and CT4 zone 2 sections, despite being perpendicular to the wind direction, present a low average wind speed and a substantial cooling rate because of the intensive planting considered. In CT4 the sections of zone 1, specifically canyon 1a and 1c, provide high advection air flows, contributing highly to ambient temperature increase. CT5 precinct is densely built with high-rise buildings. In Zone 1 of CT5, the high advection air flows contribute highly to increasing the ambient temperature, while the impact of zone 2 is negligible. Similarly, in canyon 2a in zone 2 of CT6 as well as in CT7 high advection flows are observed that result in increased ambient temperature. In all precincts, except CT5, the combination of mitigation measures, namely cool pavements, cool roofs, and vegetation, results in a substantially ambient temperature decrease. Especially significant is the mitigation impact of tree shading. In CT5, due to its density, fewer trees are used, and the high-rise buildings minimize the cool roof impact.

In the OT1 precinct, the zone 1 sections provide high advection air flows and contribute highly to increasing the temperature in the precinct. In OT2 the building arrangement provides wind shading resulting in lower wind speeds and lower heat advection. Similar conditions are observed in OT3 zones 2 and 3. In both OT2 and OT3, solar shading is not optimized resulting in high surface temperatures in the non-shaded zones. The combination of cool pavements, cool roofs, and vegetation results in a substantial decrease in the ambient temperature. Especially significant is the mitigation impact of tree shading. No wind or solar shading is present in theOT4 and OT7 configurations. In these two precincts, the mitigation measures contribute to the ambient temperature decrease along the canyon axis. Zone 3 in OT5 configuration, s is well protected from the wind. As a result, the wind speed and ambient temperature are reduced. Solar shading, however, is not optimized, and the temperature decrease along the canyon axis is owed to the mitigation measures. The OT6 configuration partly benefits from wind shading, while trees offer insignificant solar shading. In this case, too, the mitigation measures result in a temperature decrease along the precinct axis.

Analysis of the results shows that advection heat transfer is prominent in the precincts. The heat flux is strongly related to the wind and the GTD values.

The advected heat is a function of the open spaces across the precinct where wind can flow and of the corresponding wind speed.

Given that the analyzed precincts are of square form, an average cross-section, S_average, is calculated as:

$$S_{average}={\left[\left(1-BAR\right)\cdot A\right]}^{0.5}$$

(2)

where: $BAR$ is the Built Area Ratio of each precinct and $A$ is the total area of the precinct.

The average wind speed is calculated for the whole area of the precinct, V_average.

The average advected heat, (Q_average), for each precinct can be approximated with the following equation:

$$Q_{average}=S_{average}\cdot V_{average}$$

(3)

The GTD value of each precinct is observed to have a strong correlation with the corresponding $Q_{average}$ for both the mitigated and non-mitigated precincts (Figure 10).

The correlation has the form:

$$GTD\left(i\right)=a\cdot Q_{average}{}^b,$$

(4)

The $a$, $b$ values are extracted from Figure 10. The corresponding R2 values for both the mitigation and non-mitigation scenarios are close to 0.9 (see Figure 10).

The correlation clearly shows that when the heat advection in the precinct is lower, the $GTD$ is higher, meaning that the protection against overheating is higher.

A strong correlation is also found between the ratio of the average wind speed in the precinct $V_{average}$, and the incident wind speed in the limits of the precinct, $V_{inc}$ with the average aspect ratio of the precinct, $H/W$.

Where: H is the average height of the buildings and W is the average width of the streets

The correlation is expressed with the equation:

$$\frac{V_{average}}{V_{inc}}=a_1+b_1\cdot H/W$$

(5)

The parameters $a_1$ and $b_1$ can be taken from Figure 11.

A comparison of the GTD values for the mitigated and non-mitigated scenarios, shows clearly that the mitigation measures, (vegetation, evaporation, and cool materials), increase significantly the GTD of the precincts as well as their cooling capacity.

In the mitigated scenarios, the total GTD_{(total, mit)} can be attributed to (a) The mitigation measures and (b) The layout of the precinct. In this case, it can be written that:

$$GT{D}_{total,mit}=GT{D}_{mitigation}+GT{D}_{layout}$$

(6)

For a given advection rate $Q_{average}$, the GTD_layout can be calculated from the corresponding expression of Equation (4), and the specific contribution of the mitigation can be calculated from Equation (6).

Based on the analysis of the role of $Q_{average}$, it can be concluded that the cooling potential of the precincts based on their layout is increased when advection is low and decreases as advection is rising.

Moreover, a quite strong correlation is observed between the $GT{D}_{mitigation}$ and the Built Area Ratio, (Figure 12). As the Built Area Ratio increases, the contribution of the mitigation measures to the cooling rate of the precincts decreases, since less space is available for implementing the mitigation measures.

A clear and strong correlation is found between the GTD of all the precincts with and without mitigation, and their corresponding average aspect ratio, (H/W), (Figure 13). The relation is:

$$GTD\left(i\right)=a_2+b_{2}HW$$

(7)

where $a_2$ and $b_2$ are coefficients provided in Figure 13.

As shown in Figure 13, the higher the aspect ratio of a precinct the lower the cooling potential and GTD. This is expected, as (a) the application of cool roofs in high-rise buildings has a lower mitigation impact and (b) in canyons of high aspect ratios the wind speed is high and results in a much higher advection rate.

Both prediction methods proposed to calculate the GTD of the mitigated and non-mitigated precincts are of sufficient accuracy. Figure 14 and Figure 15 compare the predicted GTD values, calculated with the two methods, against the original data for the mitigated and the non-mitigated precincts. The average relative prediction error of both methods for the mitigated precincts is close to 10%. For the non-mitigated precincts, the average prediction error of the method based on the aspect ratio is close to 17%, while the corresponding error of the method based on the estimation of the advection rate is close to 12%.

The advection rate in the precincts is highly dependent on the canyon orientation and aspect ratio. The wind speed in canyons with an axis vertical or oblique to the wind direction presents a lower wind speed is lower compared to wind speed in canyons with a parallel axis to the wind direction. This depends highly on the aspect ratio, (H/W), of the canyons, as derived by Oke [45]. For canyons vertical or oblique to the wind direction, a high H/W value > 0.8 signifies that the flow is under a skimming regime and corresponds to a local vortex inside the canyon, and a bypass of the flowing air above the height of the buildings. For H/W values between 0.8 and 0.3, the flow is the wake interface, and for lower values is isolated roughness [45].

For all canyons of the precincts that have their axis vertical or slightly oblique to the wind direction, a strong correlation between the average wind speed in the canyon, V_average, and the aspect ratio (H/W, is found (Figure 16). The relation has the form:

$$V_{average}=0.2046\cdot {\left(\frac{H}{W}\right)}^{-1.745}$$

(8)

For all canyons that have their axis parallel to the wind direction, a strong correlation is observed between the length of the canyon and the product of the entry wind speed and width of the canyon as well as with the product of the exit wind speed with the width of the canyon (Figure 17).

7. Conclusions

Urban overheating causes energy, environmental, and health problems while also impacting the overall economic and cultural life of cities. The implementation of a series of mitigation measures can compensate for the negative impact of the high urban temperatures.

The mitigation measures considered in this study include the increase of greenery in the various areas, the introduction of water sprays, and the change of urban materials’ albedo. The change of ambient temperature and surface temperature of the mitigated versus the unmitigated areas is significantly higher close to water sprays. The cooling potential is then studied using the GTD values along the precinct and is strongly influenced by the wind speed and direction as well as the urban form

The effect of building height, street width, aspect ratio, built area ratio, orientation, and dimensions of open spaces on the distribution of the ambient and surface temperature is further analyzed. The cooling potential of different district arrangements and building typologies is analyzed using the parameter ‘Gradient of the Temperature Decrease along the Precinct Axis’, GTD. The GTD measures the average temperature decrease along the X or the Y-axis of the canyon. In the mitigated precincts the GTD ranges between 0.01 K/m to 0.004 K/m. In precincts without mitigation, GTD ranges between 0.0093 K/m to 0.0024 K/m. Considering the layout of the precincts only, the cooling potential decreases when mitigation measures are implemented, compared to the cooling potential of the same precinct without mitigation.

Moreover, the main heat transfer mechanism in the precincts is advection and there is a strong relationship between the wind-caused heat flux and the GTD values. There is also a strong correlation between the GTD of all the precincts, with and without mitigation, and their corresponding average aspect ratio, (Height of buildings to Width of streets). The higher the aspect ratio of the precinct the lower the cooling potential. Finally, it is worth mentioning that a high Built Area Ratio means that less space is available for the installation of mitigation measures. Therefore, the cooling contribution of mitigation measures is lower in precincts with a higher Built Area Ratio.