Assessment of Outdoor Thermal Comfort in Urban Public Space, during the Hottest Period in Annaba City, Algeria

Abstract

In this study; the outdoor thermal comfort of the users of two urban plazas with different morphologies in Annaba city, Algeria, have been evaluated. First, field measurements of the microclimatic parameters took place; namely the air temperature and the relative humidity in the two urban plazas, during hot days. Then, these measurements were compared with the results of the numerical simulations carried out by ENVI-met software in order to validate the model. The outdoor thermal comfort was evaluated by microclimatic measurements as well as a questionnaire survey consisting of interviews during the measurement days. The main objective was to determine the neutral Physiological Equivalent Temperature (PET) and to examine the influence of urban microclimatic conditions on the subjective thermal perception of people, as well as to compare it with different indexes of thermal comfort. Hence, the aim was to compare the microclimatic parameters of the two plazas, with and without vegetation and to see their impact on the thermal comfort indices. The results show that there is a difference between the two morphologies. Moreover, vegetation and urban morphology influence the neutral values of the thermal comfort indices.

1. Introduction

According to the 6th report of the Intergovernmental Panel on Climate Change, an increase in the global average temperature of 1.8 °C to 4.0 °C by 2100 is assumed, due to global warming [1]. The consequences of this climate change will be significant in urban areas, where more than 50% of the world’s population lives, thus posing particular challenges to urban areas [2]. According to the United Nations, urban areas are increasing, and it is expected that more than 70% of the world’s population will be located in urban centres by 2050. This growth leads to an increase in the urban density of buildings, especially in city centres, thus influencing the urban microclimate and the thermal comfort at the level of urban areas [3]. Urban settings present a higher threat of heat stress than rural settings, because of the phenomenon called the “Urban Heat Island (UHI)” effect [4,5]. This phenomenon occurs because of the change in the natural landscape, introducing buildings and infrastructure that replace natural terrain and vegetation cover. These changes make urban areas warmer than their rural surroundings, forming an “island” of higher temperatures [1,6,7,8].

In this context of climate change and urban growth, the issue of outdoor thermal comfort in open urban spaces is attracting more and more attention from researchers. As a result, many studies have been conducted on outdoor thermal comfort around the world [9,10,11,12,13,14]. Some studies have focused on evaluating the level of thermal comfort in outdoor urban spaces [15] and others have tried to understand the impact of urban morphology parameters (such as building height, density, h/l ratio) on the urban microclimate and on user comfort [16,17,18]. This particular attention paid to the topic of outdoor thermal comfort at the level of the outdoor public space is explained by the important role of the urban public space in improving the quality of life in urban centres. According to Kiven Lynch, the urban space is the most important part of the city, where the greatest number of human contacts and interactions occur [19]. Thermally comfortable urban outdoor spaces, including streets, squares, parks, playgrounds and other common areas, provide residents with quality public places, and more people would be drawn outdoors, thereby contributing to a more active social life and to the habitability and urban vitality of cities [20,21]. Thus, outdoor thermal comfort is a key aspect of urban public spaces for sustainable urban development [22].

Outdoor thermal comfort is a complex topic because it is influenced by various factors [23]. It is difficult to understand all of its interrelated aspects [3]. Its definition remains ambiguous and difficult to define. There are several definitions of thermal comfort in the literature. It is defined as “the absence of any discomfort and dissatisfaction or when individuals feel neither cold nor hot” [24,25,26]. It has been defined by Hensen as “a state in which there are no motor impulses to correct the Environment through behavior” [27,28]. ASHRAE defined outdoor thermal comfort as “the state of mind that expresses satisfaction with the outdoor thermal Environment” [29].

Thermal indices play a crucial role in evaluating thermal perception and are widely used for design purposes [30]. Over the past century, numerous studies have been conducted to determine thermal conditions and assess the thermal sensation of individuals, leading to the proposal of numerous thermal indices [31]. In recent decades, there has also been the emergence of many indices aiming to evaluate bioclimatic conditions for humans. Some of these indices are based on general measurements, while others rely on empirical reactions of the human body to thermal stress [32]. Several studies have been conducted to evaluate outdoor thermal comfort indices and their use in different climatic contexts.

A study conducted in a subtropical climate in China compared four indices, revealing that the Universal Thermal Climate Index (UTCI) performed the best, followed by Physiologically Equivalent Temperature (PET) and Standard Effective Temperature (SET*). UTCI was recommended for evaluating outdoor thermal comfort in the humid subtropical zones of China [33]. In Mexico, another study examined changes in the degree of thermal comfort and compared it to physiologically equivalent temperature (PET). This study used microclimatic measurements and surveys on the thermal perception of space users. It emphasised the need to adjust thermal stress categories for a better assessment of residents’ thermal comfort [34].

Many studies have been conducted to examine the relationship between thermal comfort indices. For example, a study in Cyprus evaluated the outdoor thermal environment using data on actual thermal sensation. The results showed that all studied indices had low predictive capacity for actual thermal sensation when compared to data collected from pedestrians [35]. Another study adjusted the comfort/stress ranges of PET for three different climatic contexts (Curitiba in Brazil, Rio de Janeiro in Brazil, and Glasgow in the United Kingdom), based on pedestrian thermal sensation data and on-site recorded microclimatic data [36]. These studies emphasise the importance of carefully choosing thermal comfort indices based on specific climatic contexts and considering relevant microclimatic variables to obtain more accurate assessments of outdoor thermal comfort.

Furthermore, several studies have been conducted to investigate the impact of morphological parameters on microclimate and outdoor thermal comfort [37,38,39,40]. Lai et al. conducted a literature review on “mitigation strategies to enhance thermal environment and outdoor thermal comfort in urban outdoor spaces.” They addressed studies conducted on the impact of urban geometry on microclimatic parameters and outdoor thermal comfort. They summarised studies conducted to evaluate the impact of various temperature mitigation strategies, including changes in building geometry, vegetation, surface, and water [41]. Several studies around the world have indicated that densely built urban areas have an impact on the urban microclimate and the formation of urban climatic conditions different from those of the countryside, forming what is called the UHI effect [42]. The expression “urban heat island” means the difference in temperature observed between urban environments and the surrounding rural areas. Observations have shown that temperatures in urban centres can be up to 12 °C warmer than neighbouring regions [43].

The urban open space, with its various morphologies, can significantly impact the microclimate and thermal comfort. To understand this impact, various morphological parameters have been studied, including aspect ratio (H/W), street orientation, sky view factor (SVF), density, height, compactness, sinuosity, and others [44,45]. A good understanding of the impact of urban morphology on microclimatic parameters provides important information to improve outdoor thermal comfort. Appropriate modifications to the geometry of the urban area can then be made to improve the outdoor thermal conditions [46,47].

In this study, the outdoor thermal comfort of the users of two urban plazas of different morphology has been evaluated. First, the impact of morphological parameters on the microclimate in the two spaces; then, the thermal comfort of users using a questionnaire-based survey, and a series of numerical simulations have been undertaken. This case study is situated in the city of Annaba, a Mediterranean coastal city located in the northeast of Algeria (at 36°, 52′ N, 6.57 E) (Figure 1). Annaba is the fourth largest city in Algeria in terms of population. It has recently recognised significant urban growth.

The city of Annaba is characterised by a Mediterranean climate, with seasonal variations marked by a hot summer, and a cold and humid winter with a rainy and temperate character. The main wind directions are from north to north-east. The maximum temperatures are recorded during the month of August, with a maximum average of 33 °C, and minimum during the month of January, with a value of 7 °C (Figure 2).

2. Materials and Methods

This study is based on two complementary approaches: a quantitative approach, carried out using field measurements and numerical simulations to calculate physical parameters of the microclimate, namely Ta, RH, Tmrt; and the thermal comfort index, namely PET, SET, UTCI. The qualitative approach is based on a subjective survey by questionnaire. Figure 3 summarises the method adopted in this study.

The outdoor thermal comfort was evaluated by microclimatic measurements as well as a questionnaire survey consisting of 336 interviews conducted in two public places of the city of Annaba during the summer season. The main objective is to determine the neutral PET and to examine the influence of urban microclimatic conditions on the subjective thermal perception of people and to compare it with different indices of thermal comfort. The objective is also to compare two types of public spaces in urban areas characterised by different microclimates and their relationship with morphological parameters.

2.1. Case Study

This study was carried out in two public spaces (plazas) with different morphologies; they were selected particularly because they comprised a variety of morphological factors, including the built density, H/W ratio, and SVF (

Table 1. The morphological parameters for the two chosen spaces.

The Public Space	Aspect Ratio H/W	Sky View Factor	Surface Density of the Building
Plazas of Revolution	0.28	0.43	0.5
Plazas of El Bouni	0.15	0.49	0.32

). Urban surface density measures the use of built structures in urban areas. It is determined by the proportion of built surface area compared to the total surface area of the urban zone. From a morphological standpoint, urban surface density varies depending on different modes of spatial occupation. In this study, the surface density of the building has been calculated by determining the ratio between the building footprint area and the site plan area.

The investigation was carried out during the hottest days of the year (July and August 2019), the field measurements were taken for every 2 h of the selected days, starting from 8:00 a.m. to 6:00 p.m.

The first public space is the plazas of Revolution (Figure 4); it was built from 1850 on an orderly Haussmannian layout. The volumes built are implanted in a continuous manner, forming closed islands. It is planted with several Ficus trees in its southern part, unlike the northern part where there is a lower plant density. The second public space is the square of El Bouni, which has been built in recent years, with a totally different urban form that is characterised by wide streets and large open spaces (Figure 5).

The qualitative approach was carried out in two phases; field measurements considering the physical parameters of microclimate, and numerical simulations.

2.2. Field Measurements

Field measurements were carried out over four days and these measures were taken on 29 July and 2 August 2019, at the level of the plot of El Bouni, and others were taken on 28 July and 1 August 2019, at the level of the plot of the Revolution. Four measurement points were set for each public space (Figure 6). The field measurements of the climatic parameters (temperature, humidity, wind speed) were taken every two hours from 8:00 a.m. to 6:00 p.m. using a mobile measurement station: a Testo 480 device (Figure 6) positioned at 1.50 m above the ground.

2.3. Numerical Simulations

ENVI-met is a Computational Fluid Dynamics (CFD) model that relies on RANS equations (Reynolds-averaged Navier–Stokes equations) to solve for atmospheric flow and heat transfer in urban settings. This model was initially developed by Bruse during his dissertation work in Germany in the late 1990s (Bruse and Fleer, 1998). The software is based on a holistic, three-dimensional microclimate model. It generates microclimate calculations and analyses the effects of planning measures on the urban climate (ENVI-met, 2023). In this study, the numerical simulation was carried out using the ENVI-met software to calculate the microclimatic parameters (temperature, humidity, mean radiant temperature), and the most appropriate thermal comfort index among PET, SET, and UTCI.

The objective of the numerical simulations, initially, was to compare the field measurement data with the results of the simulation to see the accuracy of the software and validate the numerical model. Then, to compare between the two public spaces during the hottest day, which is 1 August 2019. A simulation with and without vegetation was carried out with the aim of evaluating the impact of urban geometry and vegetation at the level of the two spaces, on the microclimate and thermal comfort. So, the microclimatic parameters (Tair, RH, Tmrt) and the thermal comfort index (PET, SET, UTCI) were calculated using numerical simulations. The simulation process is summarised in a synthetic diagram (Figure 7). The input data are also represented in

Table 2. ENVI-met input data for the current state and scenarios.

Input File	Settings	El Bouni Plazas		Revolution Plazas
		Current State	Scenarios	Curent State		Scenarios
ENVI-guide	simulation date	29 July 2019 and 2 August 2019	1 August 2019	28 July 2019 and 1 August 2019		1 August 2019
	Simulation time	6:00–18:00
	Meteorological boundary conditions	Simple forcing by integrating data from the weather station of the simulation date
Spaces	Grid	Dimension: (117 × 227 × 40) m Resolution: (1 × 1 × 2) m		Dimension: (108 × 319 × 40) m Resolution: (1 × 1 × 2) m
	Paving material	Road: Asphalt (0100 ST), Curb: Pavement concrete used/dirty (0100 PP), Ground: loamy soil (000000)
	Build material	Wall: default brick aerated (0100B1) Roof: concrete wall (0100C1)	Wall: brick reinforced (0100B3) Roof: roofing terracotta (0100R2)	Wall: default brick aerated (0100B1) Roof: concrete wall (0100C1)		Wall: (0100B1) Roof: concrete wall (0100C1)
	Vegetation	Ficus, deciduous tree and palm	with végétation: Ficus, deciduous tree and palm	without végétation	Ficus and palm	with végétation (Figus, and Palm)	without végétation

.

2.4. Questionnaire Survey

While the field measurements were being carried out, the questionnaire survey was performed, consisting of 336 participants (47% males and 53% females). This questionnaire was conducted in both public spaces during the summer of 2019. The questionnaire contains two parts (

Table 3. 7-point scale of TSV defined by ASHRAE Standard 55.

Part 01	Part 02
Sex	At the moment, how do you feel about the thermal environment in this place?
Age
Residence	Hot
How long have you been in this place?	Warm
Less than 10 min	Slightly warm
A half hour	Comfortable, neutral
More than an hour	Slightly cool
Your location before coming here (last hour)	Cool
In the street	Cold
In a building	Do you feel a thermal variation from your previous feeling?
At home	cooler
Own car	No variation
Taxi	Hotter
Bus	How are you feeling overall?
Other (Please specify)	Very uncomfortable
What have you consumed recently?	Uncomfortable
Ice cream	Comfortable
A cold drink	Very comfortable
A cigarette	Do you feel the thermal environment comfortable or not?
A hot drink	Yes
Cold meal	No
Hot meal	In your opinion, at what temperature we feel Comfortable in summer
Other (Please specify)

). In the first part, participants answer personal questions (age, gender) and other questions related to factors influencing outdoor thermal comfort, such as thermal history and behavioural factors. The second part deals with outdoor thermal comfort, applying a 7-point scale of Thermal Sensation Votes (TSV) defined by ASHRAE Standard 55.

3. Results

The results are presented in two parts; the results of the quantitative part, which includes the field measurements, and the simulation of the microclimatic parameters of the outdoor thermal comfort indexes. The second part concerns the results of the evaluation of the outdoor thermal comfort of the users of the two chosen plazas by questionnaire survey.

3.1. Field Measurements and Numerical Simulation

In order to validate the ENVI-met model, measured values at 4 different locations within both plazas have been compared with the simulated air temperature and relative humidity. The instruments were placed 1.5 m above the ground of the study site. The measured and simulated Ta and RH of this area were analysed at different times (in July and August 2019 from 8:00 a.m. to 6:00 p.m. in local time).

These analyses are based on the measurements taken on 28 July, 29 July, 1 August, and 2 August 2016, from 8:00 a.m. to 6:00 p.m. The field measurements of 29 July and 2 August 2016 were carried out at the level of El Bouni plazas (Figure 8). And the in situ measurements of 28 July and 1 August 2016 at the level of the Placette de la Revolution (the Revolution plazas) (Figure 9). The results are represented in graph form for each measurement point associated with current state simulation maps and have been chosen at 2:00 p.m. as a representative time. The graphs allow a comparison between the in situ measurements data, the simulation data, and the meteorological station data for each measurement point in both plazas.

3.2. Comparison between the Two Plazas

On the other hand, the objective of the numerical simulation is, first, to compare between the microclimatic parameters (air temperature, relative humidity, and average radiant temperature) and the thermal comfort indexes of the two studied plazas, with and without vegetation. Then, to see the impact of microclimatic parameters on the comfort indexes and the impact of the morphology of the two plazas on thermal comfort. The day chosen for the simulation is 1 August 2019. The climatic data during this day were retrieved from the weather station and integrated into the ENVI-met model. Four receptors were defined to establish graphs in order to make a comparison over time between the two spaces in the two scenarios (without and with vegetation). Maps are also established in order to make a comparison on a spatial scale; the time chosen to extract the maps is 2:00 p.m. (Figure 10, Figure 11 and Figure 12).

The thermal comfort indices were calculated for the two plazas and in the two scenarios using Biomet. Maps and graphs were established in order to make the comparisons where each comfort index has been presented separately (Figure 13, Figure 14 and Figure 15).

3.3. Evaluation of the Outdoor Thermal Comfort Using Questionnaire Survey

While field measurements were being carried out, the questionnaire survey was performed. Participants answer about their thermal sensation applying a 7-point scale of TSV defined by ASHRAE Standard 55. Two measuring points for each plaza were fixed to carry out the questionnaires: measuring points 2 and 3 (Figure 6).

To identify the acceptable range of the PET index that corresponds to TSV in the studied areas, the neutral temperatures of each measuring point were calculated. The neutral temperature can be calculated by regression equations in each of the four points, with the Mean Thermal Sensation Vote (MTSV = 0). The correlations between MTSV and PET are shown in Figure 16.

4. Discussion

4.1. Comparison between the In Situ Measurements Data, the Simulation Data and the Meteorological Station

After comparing the air temperature value of the ENVI-met model and the air temperature value of the in situ measurements, it was found that for measurements before 1:00 p.m. during the day of 29 July 2019, the Tair of the in situ measurements, at the level of the measurement point 3, was generally higher with an average of 1.3 °C, than the Tair of the ENVI-met model; but, after 1:00 p.m. the ENVI-met model overestimates the Tair with an average of 0.5 °C. At the measurement points 2 and 4 before 2:00 p.m., it was noted that the recorded temperature is higher than that of the simulation with an average of 1.6 °C; this difference becomes almost zero after 2:00 p.m. (Figure 8). Concerning the first measurement point, the Tair of the simulation is always lower than the measured Tair. For the relative humidity during the same day, it was noted that after 10:00 a.m., the ENVI-met model underestimates the relative humidity on the four measurement points with an average of 2.2% for measurement points 1 and 2, and an average of 3.3% for measurement points 3 and 4. During the second day of measurement at the El Bouni plazas (2 August 2019), which is a relatively warmer day than that of 29 July 2019 (10 °C difference), it was noted that the air temperatures recorded at all the measurement points were higher than those of the ENVI-met model with an average difference of 2 °C. This difference becomes almost zero at 2:00 p.m. and reaches its peaks at 8:00 a.m. and 4:00 p.m. On the other hand, the results of the simulation concerning the relative humidity are closer to the measurement data during this day on all the measurement points except for the 3rd point (Figure 8). Note that the difference between the measured data and the simulation data decreases as the simulation time progresses.

At the level of the Revolution plazas, during the day of 28 July 2019, it was noted that the ENVI-met model overestimated the air temperature after 12:00 p.m., where an average difference of 1.5 °C was noticed, except for the 3rd point where it was noted that simulated air temperatures are lower than those measured with an average of 1.4 °C (Figure 9) For humidity, it was noted that the measured values are higher than those of the ENVI-met model on all of the measurement points and throughout the day (Figure 9); this difference is an average of 4.2%. This can be explained by the close proximity to the sea. During 1 August 2019, which is a warmer day than the previous one, the simulated air temperatures are higher than those measured in the measuring points 1, 2, and 4 from 10:00 a.m.; the difference is 2 °C on average. At receptor 3, from 2:00 p.m. the temperature measured is almost the same temperature predicted by ENVI-met. On the other hand, it was noted that the measured relative humidity was higher than that of the simulation on all the receptors and throughout the day. It was also noted at the level of the Revolution plazas that the difference between the measured data and the simulation data decreased as the simulation time progressed.

Overall, the model represented very similar results between the measured and simulated data of Ta and RH for the El Bouni plazas (Figure 8). This similarity is less important for the Revolution plazas, precisely for the relative humidity (Figure 9); this difference can be explained by the proximity of the plazas to the sea, therefore the input data concerning the weather conditions should be reviewed.

As regards the comparison with the data from the meteorological station, in El Bouni plazas, the temperatures measured are higher than those recorded in the meteorological station, and this is at all of the measurement points and throughout the day, where there is a difference of up to 3.6 °C at 8:00 a.m. during the cool day (29 July 2019), and a difference of 3.9 °C during the hot day at 8:00 a.m. during the hot day (1 August 2019). Except for at 2:00 p.m., it was noticed that the temperature of the meteorological station was higher than the measured temperature with an average difference of 1.6 °C. On the other hand, it was noted that the temperature measured at the level of the Revolution plazas on almost all the receptors were lower than the temperature recorded in the meteorological station, and this was for almost the whole day. During the cool day (28 July 2019), the temperature recorded in receptors 1 and 2 in the Revolution plazas is lower than the temperature of the meteorological station after 12:00 p.m. with an average of 2.5 °C, and higher before 12:00 p.m. 0.9 °C. During the same day and in receptor 3, the temperature of the meteorological station is lower before 3:00 p.m. with an average of 1.3 °C and higher after 3:00 p.m. with an average of 0.6 °C, while the temperature of receptor 4 is lower than that of the meteorological station from 9:00 a.m. with an average difference of 1.9 °C. During the hot day (1 August 2019), the temperature recorded in the four receptors is always lower than the temperature of the meteorological station with an average difference of 1.8 °C except at 8:00 a.m., when it was noticed that the temperature of the meteorological station is less than that of the plazas with a difference of 1.4 °C. Concerning the relative humidity, at the level of the plazas of El Bouni, the measured humidity is lower than that of the meteorological station during the two measurement days. But at the level of the Revolution plazas, the measured relative humidity is always higher than that of the meteorological station during the hottest day and lower than that recorded at the level of the meteorological station during the cool day.

The coefficient of determination R² was calculated for the air temperature and the relative humidity in the two plots, to see the precision between the measured data and the simulated data (Figure 17 and Figure 18).

4.2. Comparison between the Two Studied Plazas, with and without Vegetation

4.2.1. Air Temperature

It was noticed from the maps (Figure 10) that the Revolution plazas represent lower air temperature levels than the El Bouni plazas in both scenarios (without and with vegetation). Also, the northern part of the Revolution plazas is cooler than its southern part, whether with or without vegetation. The presence of vegetation has strongly contributed to the decrease in air temperature. This decrease is more significant at the level of the Revolution plazas because the vegetation in this space is denser than El Bouni plazas. What can also be noticed from the maps that there is a big difference between the northern and southern part of the Revolution plazas (without vegetation), where it was noted that the northern part is cooler than the south (Figure 14); this can be explained by the difference in the morphology of the building, where there is more open space, in which are found higher temperatures. It was also noticed that the east-west oriented streets were the hottest.

In order to compare the difference in air temperature of the two plazas considering the two scenarios and their variation over time during the day (from 8:00 a.m. to 6:00 p.m.), the graphs are presented in Figure 19 using four receptors.

Receptor 1 (Figure 19a): The highest air temperatures are located at the level of El Bouni plazas with a maximum of 36.2 °C. Then, almost the same temperatures were recorded at the El Bouni plazas with vegetation and the Revolution plazas without vegetation, with a maximum of 36 °C and 35.9 °C, respectively. The lowest air temperatures are at the level of the Revolution plazas with vegetation with a maximum of 35.5 °C.

Receptor 2 (Figure 19b): The highest air temperatures are located at the level of the Revolution plazas without vegetation with a maximum of 36.5 °C. Then, during the morning period (before 1:00 p.m.), the same temperatures were recorded at the El Bouni plazas without vegetation and the Revolution plazas with vegetation; after 1:00 p.m., it was clearly seen that the temperatures in the El Bouni plazas without vegetation were higher than those of the Revolution plazas with vegetation, with an average difference of 0.7 °C. The air temperatures at the El Bouni plazas with vegetation were the lowest before 1:00 p.m., then they became a little higher than that of the Revolution plazas with vegetation, with a difference of 0.3 °C.

Receptor 3 (Figure 19c): From 1:00 p.m., there is a difference of 0.6 °C between the plot of the Revolution and El Bouni plazas. The temperatures of El Bouni are higher in both scenarios (without and with vegetation). There is a slight difference between the two scenarios in the two plazas (a difference of 0.3 °C at the level of the Revolution plazas and of 0.2 °C at the level of El Bouni plazas).

Receptor 4 (Figure 19d): It was noted that the same was found as in Receptor 2. The highest air temperature is located at the level of the Revolution plazas without vegetation, with a maximum of 36.7 °C. Then during the morning period (before 1:00 p.m.), it was seen that the temperatures at the plot of El Bouni without vegetation were lower than in the Revolution plazas with vegetation. After 1:00 p.m., it is clearly seen that the temperatures in El Bouni plazas are higher than those of the Revolution plazas in the second scenario (with vegetation). While the lowest temperatures of the four cases are those of the El Bouni plazas with vegetation before 2:00 p.m., and the Revolution plazas after 2:00 p.m.

4.2.2. Relative Humidity

It was noted from the maps (Figure 11) that the Revolution plazas represent overall higher humidity levels than El Bouni plazas in both scenarios (without and with vegetation). Also, it is found that the northern part of the two plazas is wetter than the southern part, in the first scenario (without vegetation). The presence of vegetation has strongly contributed to the increase in the level of humidity and especially at the level of the southern part of the Revolution plazas (an increase, which can reach up to 5%).

To be more precise in the comparison, the graphs represented in Figure 20 are referred to. From the graphs, it was found that at receptor 3 (Figure 20c), the relative humidity is almost the same in both spaces (without and with vegetation). While in the other three receptors, a difference was noted. At receptor 1 (Figure 20a), the humidity is higher in the Revolution plazas with vegetation compared to the three other cases (Revolution plazas without vegetation, and El Bouni plazas without vegetation and with vegetation) where the same level of humidity was recorded. At receptor 2, the humidity in the El Bouni plazas and the Revolution plazas without vegetation is the lowest. Concerning the second scenario (with vegetation), before 1:00 p.m. the level of humidity at the level of El Bouni is higher. The same concerning receptor 4 (Figure 20d), the Revolution plazas present lower humidity rates in the two scenarios and especially before 1:00 p.m.

4.2.3. Mean Radiant Temperature

According to the maps (Figure 12), it was noted that, in the first scenario (without vegetation), the mean radiant temperature is almost the same at the level of the two plazas. There is a very high temperature that exceeds 64 °C, which decreases to 52 °C in shaded areas (the shadow of buildings). For the second scenario (with vegetation), the impact of vegetation on the reduction of Tmrt was clearly seen, especially at the level of the plot of the Revolution. In the unshaded areas, the temperature at the level of the plot of the Revolution with vegetation is around 61 °C, while in El Bouni plazas is 63 °C. Hence, there is a decrease in the unshaded areas, of 1 °C in the El Bouni plazas and 3 °C in the Revolution plazas, by integrating the vegetation. The drop in temperature is much greater where the vegetation is located, where there is a drop of 16 °C in the Revolution plazas, and a drop of 12 °C in El Bouni plazas; this difference between the two plazas is due to the difference in plant density.

The graphs shown in Figure 18 better explain the difference in Tmrt over time, between the two plazas, in the two scenarios.

Receptor 1 (Figure 21a): It was noted that the highest Tmrt is that of the El Bouni plazas without vegetation with a maximum of 66.3 °C, then the Revolution plazas without vegetation, with a maximum of 64.8 °C. Regarding the second scenario, it was noted that there is not a big difference at the level of the El Bouni plazas, except at 3:00 p.m., when the temperature drops with 10 °C. On the other hand, at the level of the Revolution plazas, there is a difference of 16.5 °C between the two scenarios.

Receptor 2 (Figure 21b): There is a big difference between the two scenarios. In the first scenario, the temperatures recorded at the level of the Revolution plazas and El Bouni plazas are almost the same throughout the day with a slight difference, where it was noticed that the Tmrt at the level of El Bouni is higher than that of the Revolution plazas with 3 °C. This difference becomes more significant at 6 p.m. (16.3 °C). By integrating the vegetation, the temperature at the level of the Revolution plazas decreases with an average of 15.2 °C and 13.9 °C at the level of El Bouni.

Receptor 3 (Figure 21c): The difference in Tmrt is minimal (an average between 0.7 °C and 1.2 °C). There are lower temperatures at the Revolution plazas with vegetation, especially at 5:00 p.m., where the Revolution plazas are cooler, with a difference of 21 °C.

Receptor 4 (Figure 21d): In the first scenario (without vegetation), the temperature of the two plazas is the highest compared to the second scenario. The Revolution plazas without vegetation present lower temperatures compared to the El Bouni plazas, with a flagrant decrease from 4:00 p.m., and which reaches the same temperatures as recorded in the second scenario. Concerning the second scenario, the Tmrt is much lower compared to the first scenario in the two plazas. Nonetheless, a significant increase in Tmrt was noticed, which reaches its maximum at 11:00 p.m. (approximately 60 °C), for the two plazas. The temperature gradually decreases in both plazas. However, in the Revolution plazas, there is a re-increase at 4:00 p.m. and it reaches up to 62.5 °C, and then it drops to 33.7 °C at 6:00 p.m.

4.2.4. The Physiological Equivalent Temperature (PET)

According to the maps (Figure 13), the PET is very high in the eastern part of the two plazas and it can reach up to 58 °C (Figure 20a,c). This part is located next to the west facade of the buildings which limit the plazas. These facades are exposed to the sun in the afternoon and contribute to the increase in thermal discomfort by indirect solar radiation. Unlike the east facade, where the PET is lower thanks to the presence of shade. The presence of vegetation also contributes to the local decrease in PET; this local decrease can reach up to 10 °C in the Revolution plazas, and 6 °C in the El Bouni plazas, and a decrease of about 2 °C in non-vegetated areas.

From the graphs (Figure 22), it was noted that in the second scenario (with vegetation), the curves of the PET are almost the same as the curves of the mean radiant temperature in all the receptors. But in the first scenario (without vegetation), the PET of the Revolution plazas is the highest compared to the other cases, while the mean radiant temperature of the plot of El Bouni was the highest at the level of receptor 1 and 3 (Figure 22a,c).

Receptor 1 (Figure 22a): The highest PET is that of the Revolution plazas without vegetation with a maximum of 52.6 °C, then El Bouni plazas without vegetation with a maximum of 49.4 °C, and this after 9:00 a.m.. Concerning the second scenario, it can be seen that there is not a big difference at the level of the El Bouni plazas, except at 3:00 p.m., when the PET drops with 4.2 °C. On the other hand, at the level of the Revolution plazas, there is a difference of 9.1 °C between the two scenarios.

Receptor 2 (Figure 22b): There is a big difference between the two scenarios. In the first scenario, the PET recorded at the level of the Revolution plazas and El Bouni plazas are almost the same throughout the day with a slight difference. By integrating the vegetation, the PET at the level of the Revolution plazas decreases with an average of 9.5 °C and 7.8 °C at the level of El Bouni plazas.

Receptor 3 (Figure 22c): The difference in PET at receptor 3 is very small (between 0.5 °C and 0.7 °C). With a lower PET at the level of the Revolution plazas with vegetation, especially at 5:00 p.m. (a decrease of 11 °C) and at the level of the Revolution plazas without vegetation at 6:00 p.m. (a decrease of 8 °C).

Receptor 4 (Figure 22d): In the first scenario (without vegetation), the PET of the two plazas is almost the same, except at 4:00 p.m. where a significant drop at the level of the Revolution plazas were noticed. Concerning the second scenario, the PET is much lower compared to the first scenario in the two plazas. Nonetheless, a significant increase in PET was noticed, which reaches its maximum at 11:00 a.m. (about 45 °C), for the two plazas. The temperature gradually decreases in both plazas. However, in the Revolution plazas, there is a re-increase at 4:00 p.m. and reaches up to 47.4 °C, then it drops to 32.3 °C at 6:00 p.m.

4.2.5. Standard Effective Temperature (SET)

There is a similar scenario concerning the SET index; it can be seen that the distribution of the SET is similar to the distribution of the PET. According to the maps (Figure 14), it was noticed that the SET is very high in the eastern part of the two plazas and it can reach up to 43.4 °C (Figure 23a,c). This increase is at the level of the plazas adjacent to the western facades of the buildings. Unlike the east facade, where the PET is lower thanks to the presence of shade. The presence of vegetation also participates in the local decrease in the SET; this local decrease can reach up to 4 °C in the Revolution plazas, and 3 °C in the plot of El Bouni, and a decrease of about 1 °C in non-vegetated areas.

It is noted that the difference in the SET between the two plazas and in the two scenarios is less important compared to the PET and Tmrt.

Receptor 1 (Figure 23a): the highest SET is that of the Revolution plazas without vegetation with a maximum of 42.4 °C; then, the El Bouni plazas without vegetation with a maximum of 39.9 °C and this after 9:00 a.m.. Concerning the second scenario, it can be seen that there is not a big difference at the level of the El Bouni plazas, except at 3:00 p.m. when the SET drops with 2 °C; on the other hand, at the level of the Revolution plazas an average difference of 4.5 °C between the two scenarios. In the first scenario, the difference between the SET of the Revolution plazas and El Bouni plazas is 3 °C; this difference is higher compared to the PET which is 2 °C.

Receptor 2 (Figure 23b): there is a noticeable difference between the two scenarios. In the first scenario, the SETs recorded at the level of the Revolution plazas and El Bouni plazas are almost the same throughout the day with a slight difference, where it was noticed that the SET at the level of the Revolution plazas is greater than that of the El Bouni plazas, with 0.5 °C. By integrating the vegetation, the PET at the level of the Revolution plazas decreases with an average of 3.6 °C and 2.4 °C at the level of the El Bouni plazas. It was noted that in the second scenario that the SETs in the Revolution plazas are lower than in the El Bouni plazas, especially between 12:00 p.m. and 3:00 p.m.

Receptor 3 (Figure 23c): The difference in SET at receptor 3 is insignificant. With a lower SET at the level of the Revolution plazas with vegetation, at 5:00 p.m., and at the level of the Revolution plazas without vegetation at 6:00 p.m.

Receptor 4 (Figure 23d): In the first scenario (without vegetation), the SET of the two plazas is almost the same, except from 4:00 p.m., it was noticed that there was a significant drop at the level of the Revolution plazas. Regarding the second scenario, there is a significant drop in the Revolution plazas (3.2 °C); this difference decreases at 11:00 a.m. and at 4:00 p.m. to become 1 °C. The SET of the El Bouni plazas in the second scenario drops with an average of 1.6 °C between 7:00 a.m. and 9:00 a.m., and an average of 1.2 °C between 12:00 p.m. and 6:00 p.m. Between 10:00 a.m. and 12:00 p.m., it can be seen that the SET of the El Bouni plazas without vegetation is lower than the SET of the El Bouni plazas with vegetation, with an average of 1.8 °C.

4.2.6. UTCI

In both scenarios, it can be noted that the values of the UTCI at the level of the Revolution plazas are generally lower than the El Bouni plazas, where it was recorded in the first scenario: a maximum UTCI equal to 45.5 at the level of the Revolution plazas and 45.7 °C at the level of the El Bouni plazas. The addition of vegetation contributes to a reduction of 1 °C in the non-vegetated areas of the Revolution plazas, and 0.7 °C in the non-vegetated areas of the El Bouni plazas (Figure 15).

Receptor 1 (Figure 24a): It was noted that the highest UTCI is that of the El Bouni plazas without vegetation with a maximum of 44 °C, then the Revolution plazas without vegetation, with a maximum of 43.3 °C. Regarding the second scenario, it can be seen that there is not a big difference at the level of the El Bouni plazas, except at 3:00 p.m., when the temperature drops with 3 °C. On the other hand, at the level of the Revolution plazas a difference of 4 °C is recorded from 9:00 a.m. to 5:00 p.m., between the two scenarios.

Receptor 2 (Figure 24b): At receptor 2, there is a big difference between the two scenarios. In the first scenario, the UTCI values recorded at the level of the Revolution plazas and El Bouni plazas are almost the same throughout the day except at 6:00 p.m., when the UTCI at the level of the Revolution plazas is lower than that of the El Bouni plazas, with 4 °C of difference. By integrating the vegetation, the UTCI at the level of the Revolution plazas decreases with an average of 5.2 °C and 4.3 °C at the level of the El Bouni plazas.

Receptor 3 (Figure 24c): The difference in UTCI at receptor 3 is minimal, with lower temperatures at the level of the Revolution plazas with vegetation, especially at 5:00 p.m. (a drop of 5.5 °C). Also, it can be seen with the difference between the two plazas that the Revolution plazas, are cooler, with a difference of 1 °C in the second scenario and 0.7 °C in the first scenario.

Receptor 4 (Figure 24d): In the first scenario (without vegetation), the UTCI values of the two plazas are higher compared to the second scenario. In the first scenario, the plazas of El Bouni present lower temperatures (difference of 0.5 °C) compared to the Revolution plazas until 4:00 p.m.; after 4:00 p.m. there is a flagrant decrease in the UTCI of the Revolution plazas (a difference of 5.3 °C between the two plazas). Concerning the second scenario, the UTCI is much lower compared to the first scenario in the two plazas. There is a significant difference in the Revolution plazas (about 5.3 °C) between 7:00 a.m. and 11:00 a.m. and 4.4 °C between 11:00 a.m. and 4:00 p.m.; this difference decreases at 11:00 a.m. (1.1 °C) and at 16:00 p.m. (1.4 °C). The UTCI of the El Bouni plazas in the second scenario drops with an average of 4.4 °C between 7:00 a.m. and 10:00 a.m. and an average of 4 °C between 12:00 p.m. and 6:00 p.m. Between 10:00 a.m. and 12:00 p.m., it can be seen that the UTCI of the El Bouni plazas without vegetation is equal to the UTCI of the El Bouni plazas with vegetation.

The coefficient of determination for each comfort index was calculated in order to grasp the correlation between each of these indices with the air temperature on one side (Figure 25), and the average radiant temperature on the other side (Figure 26). The results show that the three comfort indices, namely PET, SET, and UTCI, have a strong relationship with the mean radiant temperature, especially PET with the coefficient of determination R² = 0.965. This correlation is less significant for air temperature: R² = 0.804 for PET. The relation of the UTCI with the air temperature is more important compared to the other indexes (R² = 0.868). The disparity in the correlation between thermal indices and climatic parameters can be attributed to variations in their calculation approaches and the range of variables they incorporate. This explains the significant relationship between UTCI and air temperature, as compared to the other indices. Furthermore, these results demonstrate that investigating the differences between thermal comfort indices can be an intriguing research topic, examining the relationship between various comfort indices and climatic parameters such as air temperature, mean radiant temperature, and relative humidity.

4.3. Questionnaire

The calculation of neutral temperatures for each measuring point, using equations where MTSV = 0, shows that in the Revolution plazas, the neutral PET is lower than that of the El Bouni plazas (22.30 °C for measuring point 2 and 24.16 °C measuring point 3). The neutral PET in the El Bouni plazas is 25.11 °C in measuring point 2 and 25.95 °C in measuring point 3. The reason for this variation could be explained by the variation in the microclimatic parameters, but also that the adaptation and short-term thermal history of users.

Table 4. Neutral values of Tmrt, Tair and thermal comfort indexes.

	Neutral PET	Neutral SET	Neutral UTCI	Neutral Tair	Neutral Tmrt
Receptor 2 The Revolution	22.3 °C	28.58 °C	25.80 °C	26.87 °C	20.93 °C
Receptor 3 The Revolution	24.16 °C	29.33 °C	27.01 °C	27.51 °C	23.76 °C
Receptor 2 El Bouni	25.11 °C	29.72 °C	27.63 °C	27.84 °C	25.44 °C
Receptor 3 El Bouni	25.95 °C	30.06 °C	28.18 °C	28.13 °C	26.79 °C

shows the neutral values of the Tmrt and Tair, as well as the comfort indexes.

It is important to carefully choose thermal comfort indices based on specific climatic conditions and consider relevant microclimatic variables to obtain more accurate assessments of outdoor thermal comfort. Several studies have been conducted to evaluate outdoor thermal comfort indices, including UTCI, PET, and SET, and their application in different climatic contexts [31,32,33,34,35,36]. These studies have highlighted the need to adjust thermal stress categories for a better evaluation of residents’ thermal comfort, based on pedestrian thermal sensation data and on-site microclimatic data.

In a study conducted at 14 sites in five European cities, Nikolopoulou found that recent experience and expectations play a major role and account for a variation of more than 10 °C in neutral temperatures [21]. Another study in Brazil aimed to calibrate the thermal perception categories of the Universal Thermal Climate Index (UTCI). The results revealed that when UTCI ranges from 19.3 °C to 26.7 °C, individuals perceive a “neutral” thermal comfort sensation [48], whereas in our study during the summer period, the recorded UTCI values ranged from 25.80 °C to 28.18 °C.

In the context of Algeria, two studies were conducted in two cities with different climatic contexts: A Mediterranean climate (Djijel city) [49] and a semi-arid climate (Biskra city) [50]. The results indicated that in the arid climate, the neutral PET had an average value of 28.4 °C, while for Djijel city, characterised by a Mediterranean climate similar to our study case, the neutral PET value was 21.2 °C. The neutral PET recorded in our study ranged from 22.3 °C to 25.95 °C.

These differences in neutral values of thermal comfort indices are due to variations in subjective and objective parameters considered in each studied climatic context. Indeed, it is possible to predict urban thermal comfort using a calibrated evaluation scale of thermal indices such as UTCI, PET, and SET, providing relevant information for decision making in urban planning strategies.

5. Conclusions

In this study, the outdoor thermal comfort of the users of two urban plazas of different morphology were evaluated. First, studies of the impact of morphological parameters on the microclimate in both plazas and then the thermal comfort of users by using questionnaires and numerical simulations have been undertaken. The outdoor thermal comfort was evaluated by microclimatic measurements as well as a questionnaire survey consisting of 336 interviews conducted in two public spaces (plazas) of the city of Annaba during the summer season. The main objective was to determine the neutral PET and to examine the influence of urban microclimatic conditions on the subjective thermal perception of people, and to compare it with different indices of thermal comfort. Hence, another objective was to compare two types of public spaces in urban areas characterised by different microclimates and their relationship with morphological parameters.

The results of this study can be summarised by the following points:

-

The ENVI-met software is a reliable tool for studying the urban microclimate. The results of the numerical simulations have proven to be close to the data measured in situ, knowing that the input data concerning the meteorological boundary conditions used in this study is simple forcing. The coefficient of determination R² was calculated for the air temperature and the relative humidity in the two plazas, to find the precision between the measured data and the simulated data. Results showed that the ENVI-met model provided values close to the measured data concerning the air temperature with the coefficient of determination R² = 0.87 in the El Bouni plazas, and R² = 0.70 in the Revolution plazas. This accuracy is less concerning humidity (R² = 0.82 in the El Bouni plazas and R² = 0.63 in the Revolution plazas). The proximity to the sea could explain this difference recorded between the measured and simulated humidity values. In addition, some more information about the surrounding meteorology could further improve the results of the simulations and best results can be expected by using the full forcing option.

-

The less dense urban fabric (the El Bouni plazas) has higher air temperatures and mean radiant temperatures compared to the denser urban fabric (the Revolution plazas). The temperature difference between the two plazas can vary by 0.4 °C for Tair and 1.1 °C for Tmrt. Thus, the density contributes to the improvement of thermal comfort during the day and especially between 4:00 p.m. and 6:00 p.m.

-

The results of the questionnaire survey allowed us to calculate the neutral PET for each studied plaza, and in vegetated and non-vegetated areas. The neutral PET in the Revolution plazas is 22.3 °C in the vegetated area and 24.16 °C in the non-vegetated area. Neutral PET in the El Bouni plazas is 25.11 °C in the vegetated space and 25.95 °C in the non-vegetated space. It can therefore be seen that the urban morphology and the vegetation influenced the neutral values of the comfort indexes (Table 4).

-

The vegetation participates in the reduction of the air temperature and the mean radiant temperature. This decrease can reach 1 °C in the Tair and 20 °C in the Tmrt. The improvement of thermal comfort through vegetation can also indirectly affect non-vegetated areas. The improvement of thermal comfort by vegetation decreases at the end of the day. Thus, a comparative study between the impact of vegetation during the daytime and nighttime could be an interesting study subject for further future studies.