1 Introduction

Cities have well known effects on climate parameters like temperature, humidity, precipitation, wind, sunshine duration and total solar radiation, at micro, local, meso and regional scales depending on various urban features including the geography, forms, surface characteristics and anthropogenic activities (Oke et al., 2017). Cities create their own unique climate conditions as microclimate with the urban heat island effect (Cleugh & Grimmond, 2012). Urban areas affect the balance of heat exchange processes by influencing the uptake and transmission of solar radiation (WMO, 2023). Thermal comfort can be defined as a state in which individuals feel comfortable in their environment or a state in which they are not exposed to heat or cold stress (Parsons, 2003). City dwellers are exposed to the combined effect of more than one climate element (Matzarakis, 2020) and percept this effect in their surrounding as thermal comfort condition together with the activity performed, clothing insulation and physiological characteristics of human body (ISO, 1994). Harmful effects of uncomfortable conditions on humans are stated in previous studies like health problems (weakness, chronic diseases, fatigue, headaches, etc.), the increase in death rates, decrease in work efficiency, and mental and psychological depressions (Błażejczyk et al., 2018; Nastos & Matzarakis, 2011). Today, urban areas are the problematic points on the earth as they cause likely the largest negative impacts on natural environment (Azeri et al., 2015; Boukhabla & Alkama, 2012). Cities represent higher temperature averages, lower mean relative humidity, calmer (windless) air conditions, higher precipitation rate, larger amount of polluting gas and particulate matter and faster surface runoff waters (from storms) (Cohen et al., 2012; Grimmond, 2007; Landsberg, 1981; Oke, 1973). Urban surfaces are transformed from natural lands as the result of human activities and such a transformation intervenes heating mechanisms and the heat balance in urban atmosphere (Murshed & Saadat, 2018). In spite of all these negative impacts, the cities house and provide economic welfare with over 4 billion people (nearly half of the world's population) while producing 80% of the global gross product. Although cities account only for 2–3% of earth's land surface, they consume more than 60% of global energy and produce 70% of greenhouse gas emissions and global waste (C40 Cities, 2012; Liu et al., 2014; United Nations, 2016) due to the increase in the number, size and population of (mega)cities (Demircan & Toy, 2019; Naboni et al., 2019; United Nations, 2018). If urbanization is managed well, it can contribute to sustainable growth and development, increase productivity, contribute to innovation and the development of new ideas (Sayad et al., 2021; United Nations, 2019).

It is possible through new spatial planning and design strategies and technologies to develop and transform cities with favourable characteristics towards environment which can also provide human quality of life and economic growth, environmental awareness and create cleaner, safer and aesthetically more attractive sites (Karimi et al., 2023; United Nations, 2019). In this respect, the use of efficient open green spaces in cities is at the center of all these strategies. Open green spaces in urban areas can reduce surface temperatures as they provide shade (Narimani et al., 2022; Taleghani, 2018) moisture to the environment (Yuan & Bauer, 2007) while urban impervious and brown surfaces without vegetation (like sidewalks, roofs, asphalt roads, parking lots etc.) and shading effects cause overheating (Boukhabla & Alkama, 2012; Chapman et al., 2017). Urban green spaces cover a wide variety land-use from urban parks, green walls, roof gardens and urban forests to allotment gardens and all vegetated areas along with water surfaces like lakes and rivers in the urban environment (Omar, 2019). The construction of low albedo sidewalks and the use of trees with wide crowns in urban parks contribute to thermal comfort conditions (Karimi et al., 2020).

It was not until the nineteenth century (Swanwick et al., 2003) that the urban dwellers did not perceive the deficiency and the need of their cities for urban green spaces throughout the history since the cities before modern times were smaller-sized and less populated as well as showing the countryside characteristics. In today’s cities, urban green spaces are among the vital elements to make cities more livable and functional (Gómez et al., 1998; Sodoudi et al., 2018; Spiller, 1993).

Urban green spaces offer recreational possibilities in daily life, contribute to the conservation of biodiversity, the formation and preservation of the cultural identity in a city, the protection and maintenance of the environmental quality of a city, brings natural solutions to technical problems (Akbarı et al., 2001; Azeri et al., 2015; Sandstrom, 2002) and reduce average air temperature by 5 °C (Spronken-Smith & Oke, 1998) and mean radiant temperature (Tmrt) by 20.04 °C (Klemm et al., 2015). Mean radiant temperature (TMRT) has an important place in the evaluation of thermal comfort conditions of cities (Leroyer et al., 2018).

The aim of this study is to evaluate the spatial distribution of thermal comfort conditions in the sample of a cold city centre, Yakutiye district, Erzurum, Turkey based on long term meteorological values (from 2004 to 2020) in the hottest 2 months, July and August by considering the effects of the factors including both meteorological parameters and surface characteristics like the deficiency in green spaces on the distribution of thermal comfort.

Thanks to the findings obtained, it is aimed to determine the factors that cause adverse thermal comfort conditions in cities. In addition, it is thought to contribute to ecological urban designs and landscape studies that take into account the physical geography conditions.

2 Materials and Methods

2.1 Study Area

Erzurum city center, which was established at an altitude varying between 1758 and 2100 m, in the Northeast Anatolia (NUTS II) sub-region of Turkey (39° 55′ N and 41° 16′ E), is far from the maritime effect. The city is exposed to the seasonal fluctuations in meteorological parameters very strongly and harsh continental-mountainous climate characteristics. The city can be categorised among winter cities with humid–semi–humid, continental–mountain climate with extremely severe winters, no dry season, cool-temperate summers and seasonal effects. The study area is selected to be a densely built area of Yakutiye central district, which is the historical core of Erzurum city center (Fig. 1) and represents the characteristics of urban core dominated by the built environment and offers the variety of land uses such as green areas and impervious surfaces.

Figure 1
figure 1

Location map of the study area

Full size image

According to the long-term averages obtained from the meteorology station, which has been operating in the city center since 1929 but moved to the airport in 1988, the annual average temperature is 5.7 °C with a range from 36.5 °C in summer and − 37.2 °C in winter (record values). The city receives 403.3-mm annual rainfall which is higher in the spring, early summer and late autumn periods, while the lowest in August. In the city center, where snow cover can be seen on the ground from mid-November to late April, average snow depth does not fall below 20 cm in the majority of the mentioned period. Depending on its elevation and topographic structure (bowl-shape) the city is exposed to an intense inversion event from the late autumn to mid-winter, when the city is under the influence of stable air masses in which long-term dry, frosty and low-quality air prevails. In summer time, highest temperature values are measured in July and August. Some average and extreme values of the city center are given in Table 1.

Table 1 Average and extreme values for Erzurum
Full size table

2.2 Data

In the study, data were obtained from the meteorological station located in the yard of the Regional Directorate of Meteorology in the city center (39° 54″ N; 41° 17″ E; 1869 m; Fig. 2). The surrounding of the station is densely built and exhibits urban features. Hourly air temperature (°C), relative humidity (%), wind (m/s) and cloudiness (octa) values measured in a period from 2004 to 2020 were used to determine the effect of the urban environment on human thermal comfort.

Figure 2
figure 2

Study area

Full size image

2.3 Methods

Thermal comfort conditions assessments were conducted for the hottest 2 months, July and August (62 days in total). PET (Physiological Equivalent Temperature; Höppe, 1999; Matzarakis et al., 1999; Verein Deutscher Ingenieure, 1998) index was used to calculate thermal comfort values. This index is frequently used all around the world because it takes into account many factors including those related to both human body and environment. Calculations were performed using the RayMan radiation model (Matzarakis et al., 2007). The RayMan model was preferred in this study because it is a radiation model that takes into account many climatic parameters (air temperature, relative humidity, wind speed, cloudiness). In the calculation, a 35-year-old 175 cm tall 75 kg male with 0.9 clo clothing load and 80 W workload was taken into account (Matzarakis & Mayer, 1996; Matzarakis et al., 1999). The comfort ranges given in Table 2 were taken into account in the classification of the data obtained.

Table 2 Human thermal sensation and stress ranges for PET (Höppe, 1999; Matzarakis et al., 1999)
Full size table

ArcGIS 10.5 software program was used to determine the spatial distribution of the PET values and monthly distribution maps are presented individually. In addition to spatial distribution, temporal ones are given in percentages. Elevation values and land use characteristics were taken into account in the distribution of values. The values were converted into raster maps, and then the spatial distribution of PET was obtained by overlapping land use raster maps, elevation maps and solar radiation maps (Fig. 3).

Figure 3
figure 3

Work flow chart showing the acquisition of thermal comfort maps

Full size image

In the spatial distribution of PET values, a newly developed model approach, which has been tested for its validity in different climate zones, was used (Çağlak, 2024).

3 Results

In the study where real-time measured meteorological values are used to calculate thermal comfort values for a densely-built city centre, distribution of thermal comfort values was determined considering land-use types to determine also their effects on thermal comfort in the hottest period of a winter city. Distribution ratios of land use types are given in Fig. 4 and their distribution map is given in Fig. 5. According to the distribution map and the figure, the study area covers at least 11 land-use types including dense, moderate and sparse housing (built-up) areas (most dominant), public area where public-uses like schools, public buildings and mosques are located, green areas including parks and gardens of public institutions, impervious road surface (asphalt) and a sport facility. Ratio of total green areas is only 6,6% compared to built-up surfaces.

Figure 4
figure 4

Ratios of land-use types

Full size image
Figure 5
figure 5

Space usage distributions

Full size image

3.1 Human Thermal Comfort Conditions in the Area

Distribution maps were prepared for both study months considering the means of hourly PET values calculated for the study period. For each month two types of distribution intervals were adopted. For the first interval minimum and maximum PET values were taken and 2 °C is added to the baseline value, which is 18 °C, and 20 °C for July and August respectively. The second interval is based on the standard thermal sensation intervals given in Table 3 above. In addition, the surface area exposed to heat stress was also calculated on the maps.

Table 3 Percentage rates of predetermined thermal comfort intervals
Full size table

Figure 6 represents the distribution of mean thermal comfort values in July for both categorisation intervals. It is seen from the maps that heat stress (red in the above; 24.1–26.2 °C and yellow in the below; 23.1–29.0 °C) prevails dominantly in the study area, nearly entire built-up area. Green areas i.e. parks and gardens of public institutions, which are generally sparsely structured represent the lowest and light heat stress in the map on the left while in the map on the below include only two standard thermal sensation categories, “comfortable” and “slightly warm”. The map on the below reflects nearly the same characteristics with that on the above where dominant interval is slightly warm i.e. heat stress. The ratio of the area exposed to slightly warm interval is 67.5% of the whole study area and the rest (32.5%) is in the comfortable range.

Figure 6
figure 6

Mean thermal comfort conditions in July (distribution of PET values above, distribution of comfort ranges below)

Full size image

Figure 7 represents the distribution of mean thermal comfort values in August for also both categorisation intervals. It is seen from the maps that heat stress (red in the above; 26.1–28.9 °C and yellow in the below; 23.1–29.0 °C) prevails more dominantly in the study area than that in July. The area exposed to heat stress covers nearly whole study area except only for parks small parts of public institutions’ gardens. Entire of the built-up area no matter how sparsely it is structured is under heat stress. Green areas i.e. parks and gardens of public institutions are also under heat stress in both maps in August. The map on the below includes again only two standard thermal sensation categories, “comfortable” and “slightly warm”. Dominant interval is slightly warm i.e. heat stress again. The ratio of the area exposed to slightly warm interval in August is 91.9% of the whole study area and the rest (8.1%) is in the comfortable range.

Figure 7
figure 7

Mean thermal comfort conditions in August (distribution of PET values above, distribution of comfort ranges below)

Full size image

Figure 8 represents the distribution of maximum thermal comfort values in July for both categorisation intervals. Maximum thermal comfort values are calculated by taking the daily maximum temperature together with other parameters given above in the daytime. It is seen from the maps that heat stress (all the colours in both maps; 23.8–32.6 °C) prevails in entire study area. Green areas i.e. parks and gardens of public institutions and sparsely structured areas correspond lighter heat stress in both maps. The map on the below which is the result of standard thermal sensation categories include again two standard thermal sensation categories but this time with higher PET values and higher heat stress, so-called “slightly warm” and “warm”. The ratio of the area exposed to slightly warm interval (light heat stress) is 60.1% of the whole study area and the rest (39.9%) is in the warm range (moderate heat stress).

Fig. 8
figure 8

Maximum thermal comfort conditions in July (distribution of PET values above, distribution of comfort ranges below)

Full size image

Figure 9 represents the distribution of maximum thermal comfort values in August for both categorisation intervals. It is seen from the maps that heat stress (PET values ranging from 28.3 to 36.4 °C) is prevalent in entire study area. Heat stress is lighter with lower PET values in green areas and sparsely structured areas in both maps as in July but this time with higher values. The map on the below based on standard thermal sensation categories gives three standard thermal sensation categories with higher PET values and higher heat stress, so-called “slightly warm” and “warm” and “hot”. The ratio of the area exposed to slightly warm interval (light heat stress) is 1.1%, warm (moderate heat stress) 71.6% and hot (strong heat stress) 27.3% of the study area.

Figure 9
figure 9

Maximum thermal comfort conditions in August (distribution of PET values above, distribution of comfort ranges below)

Full size image

Figure 10 represents the distribution of minimum thermal comfort values in July for both categorisation intervals. Minimum thermal comfort values are calculated by taking the daily minimum temperature together with other parameters given above in a day. It is seen from the maps that rather than heat stress, light cold stress (blue and light greens in both maps; 12.1–19.3 °C) prevails in the study area. Parks and gardens of public institutions and sparsely structured areas correspond light cold stress in both maps while built-up areas are in comfortable range. The map on the below prepared for standard thermal sensation categories include again three standard thermal sensation categories with lower PET values and higher cold stress, so-called “cool” and “slightly cool” and “comfortable”. The ratio of the area exposed to “cool”, “slightly cool” and “comfortable” intervals is 1.8%, 47.6% and 50.6%, respectively.

Figure 10
figure 10

Minimum thermal comfort conditions in July (distribution of PET values above, distribution of comfort ranges below)

Full size image

Figure 11 represents the distribution of minimum thermal comfort values in August for both categorisation intervals. It is seen from the maps that no heat stress is present but slight cold stress and neutral conditions (blue and light greens in the map on the left and light green on the right; 14.3–22.4 °C) are prevalent in the study area. Parks and gardens of public institutions and sparsely structured areas correspond light cold stress in both maps while built-up areas are in comfortable range as in July. The map on the below covering the standard thermal sensation categories shows two standard thermal sensation categories with lower PET values and higher cold stress again, so-called “slightly cool” and “comfortable”. The ratio of the area exposed to “slightly cool” and “comfortable” intervals is 19.6% and 80.4%, respectively.

Figure 11
figure 11

Minimum thermal comfort conditions in August (distribution of PET values above, distribution of comfort ranges below)

Full size image

4 Discussion

In the scope of the study an approach was developed to reveal the effects of land use types on the degree and distribution of human thermal comfort conditions in the hottest 2 months of a historical urban core in a high-elevated winter city over a period from 2004 to 2020. In this approach real time monitored meteorological values were used to calculate thermal conditions and the spatial distribution of these conditions throughout the study area was detected using ArcGIS software considering also elevation. From the results of the analysis, spatial percentage rates of the land use types and the spatial distribution of predetermined thermal comfort intervals (i.e. thermal sensation cool, slightly cool, comfortable, slightly warm, warm and hot) were detected to monitor the land use impact on thermal comfort conditions in urban environment (Table 3). It is seen from the table that totally six thermal comfort intervals can be seen in the study area revealing cold stress, neutral conditions and heat stress in the hottest months of the study area when mean, maximum and minimum PET values were evaluated. The values obtained for the study area are in the range of those found in previous studies (Toy & Kantor, 2017).

Cold stress in the area is experienced majorly at night-time and early morning hours when minimum temperature is measured, while heat stress is effective during daytime and comfortable conditions are generally prevalent at the same time with minimum temperatures i.e. morning hours. Such extreme values even in summer months can be explained by the high density of solar radiation depending on city’s elevation and diversity in urban morphology i.e. sky view factor (SVF; Dirksen et al., 2019). Impact of urban structure on climate elements especially temperature related to local solar conditions (direct and reflected) has long been shown in literature (Chen et al., 2012; Dirksen et al., 2019; Oke, 1973, 1981). Moisture content of the urban atmosphere coming especially from vegetated and wet surfaces can help balance temperature fluctuations. In this respect, the city is far from marine effect and relative humidity in summer time is generally below 60% (yearly 64.6%; Table 1). As the result of the analysis, it can be stated that in the daytime, especially from noon till sun set heat stress is possible especially under the sun but at night time soon after sunset cold stress begins with the effect of north-easterly chilly winds due to local conditions (Toy et al., 2007).

In addition to meteorological and topographical factors to describe the urban bioclimatological conditions and their temporal and spatial distribution, manmade land-use types also take key parts as they first impact shading (canopy) effect due to the sky view factor (SVF; Oke, 1973), which corresponds the urban geometry and the density of structured areas (Dirksen et al., 2019; Dokhanien et al., 2023; Gal et al., 2007) then the formation of UHI effect.

In the scope of the study SVF observations and imaging were also performed but not taken into account since the effects of land-use types mainly in two groups (impervious and green areas) tried to be investigated. As the result of the study given especially in the distribution maps green areas (parks and institutions’ gardens) show the conditions closer to neutral (comfortable) due to well-known reasons e.g. optimum heat storage and moisture content of vegetated surfaces (Lindberg & Grimmond, 2011; Theeuwes et al., 2017).

5 Conclusion

The results of the present study show that in a medium-sized and high-elevated city like Erzurum, where solar radiation is strong and not heavy industry and human activity take place, thermal comfort conditions exhibit unfavourable characteristics during the hottest summer months in mean and maximum conditions especially in built-up areas.

This study also shows that thermal comfort (sensation) intervals fall short in representing the values obtained to define real feelings of city dwellers since the categorisation based on maximum and minimum values showed well diversified intervals. Therefore, a new classification study should be performed for thermal comfort intervals in high-elevated winter cities with extreme heat stress in summer.

This study is also important to show that real time measured, long term and standardised values by authorised institution are important to give accurate and comparable results rather than short term and individual measurement. Therefore, in order for studies using data analysis to give more concrete and accurate results, their data should be long term and well-standardized. Short term data series fail to give and represent perfect results.

As a conclusion, adverse thermal comfort conditions are expected to be more prevalent in future cities (as stated in IPCC Assessment Reports) due to the climate hazards like heat waves, extreme events etc. therefore cities must have high adaptation capacity to climate change to reduce climate risks. In this respect, environmentally respectful spatial design applications should be employed for well-balanced distribution of land-use types to reduce the size and impact of impervious surfaces and increase green areas to improve thermal comfort conditions. It is important to increase the resilience of the city to climate risks, density of built-up areas should be relieved, empty and suitable spaces for green infrastructure should be employed to increase the size of green areas per capita thus preventing excessive absorption of solar radiation and balancing other climatic elements like providing humidity to urban environment. High branching tall trees should be preferred on the sidewalks in order to benefit from their shading effects against the sun rays.