Analyzing the Contribution of Urban Land Uses to the Formation of Urban Heat Islands in Urmia City

Abstract

An Urban Heat Island (UHI) is an important variable in climate and environmental studies. Nowadays, population growth and urbanization development are the most important factors affecting the temperature increase in urban areas, which cause the creation of heat islands in urban areas. (1) Background: This study explores the influence of major land uses on the creation of Urban Heat Islands in Urmia city, Iran. (2) Methods: To achieve the aim of this study, Landsat satellite data including Landsat 5 and 8 imageries from the time periods of 1990 and 2023 were used. With the series of data processing and analyses on vegetation cover and land surface temperature, the impact of main land uses on the creation of Urban Heat Islands and the intensification of their effects have been investigated. (3) Results: The analysis reveals that barren lands consistently exhibit the highest temperature, while garden lands show the lowest temperature across both periods. In addition, the spatial distribution of Urban Heat Islands demonstrates a clustered pattern throughout the study period, with hot spots mainly located in the northwestern and southwestern areas. (4) Conclusions: This study’s findings can be helpful for urban policymakers and planners to develop practical strategies to mitigate UHIs and improve climate resilience in cities.

1. Introduction

Rapid urbanization leads to UHI creation which promotes energy consumption and causes lower thermal comfort, and increased risks to human health [1,2,3]. For example, urbanization growth in Shanghai has raised the average summer temperature by approximately 1.5 °C, leading to a 16% growth in cooling energy consumption [4]. The increase in UHIs in Tokyo has caused a 20% rise in cooling energy demand over the last decades [5].

The development of cities and ongoing urban constructions have impacted the climate locally and globally [6]. Nowadays, climate change has become one of the most important issues in the world due to its environmental, economic, social, and even political consequences [7,8]. Increasing greenhouse gases due to human activities and the industrialization of countries are among the top drivers of climate change, leading to an increase in the Earth’s temperature and more frequent natural disasters [9]. Therefore, human activities affect the atmospheric condition of the Earth, causing destructive impacts on the growth of plants, biodiversity, soil, forests, and underground water, ultimately resulting in climate change [10]. The phenomenon of Urban Heat Islands (UHIs) is one of the clearest evidence of human-made climate change in urban areas [11,12]. When a heat island forms, isothermal curves are drawn in the city and its surrounding areas, depicting the city as an island surrounded by water [13]. Urban Heat Islands are urban areas that are significantly warmer than their surroundings. The intensity of heat islands can vary between 0 and 12 degrees Celsius globally, depending on the season, sunlight, and urban characteristics [14]. In moderate UHIs, urban areas can experience temperature increases of 1–3 °C [15], while in extreme UHIs, the heat island effect can reach 8–12 °C, particularly in dense urban centers [16].

UHIs often occur during both day and night, but the sun’s radiation intensifies them during the day. Depending on the season, as vegetation and weather change, the intensity of surface UHIs also varies [17]. In most cases, this phenomenon forms when a large percentage of natural and green covers are wiped out [18], and the natural surface of the Earth is replaced with paving, buildings, concrete, asphalt, and other urban constructions which eliminate the cooling effects of these natural surfaces. In addition, the heat produced by vehicles, factories, and air conditioners increases ambient heat. Tall buildings and narrow streets also reduce airflow and contribute to high temperatures. These changes create Urban Heat Islands [8,19]. The phenomenon whereby urban areas experience warmer temperatures than the surrounding rural areas occurs when the sun’s rays are trapped by the urban structure during the day and are reflected at night [20,21,22]. These processes cause cities to be surrounded by a hot air mass, which is about 120 m high during the day and more than doubles during the night [23]. As a result, the natural cooling process of the Earth’s surface occurs at a slower rate during the night; therefore, the air temperature of cities is usually higher than that of surrounding areas [24]. This temperature difference sometimes reaches 5 to 6 degrees Celsius, and even at night in large cities, it can reach 6 to 8 degrees Celsius [25].

In general, an UHI is revealed in two ways: a. Surface Urban Heat Island (SUHI), which is the temperature change between the surfaces of urban areas and nearby rural areas [26], and b. Atmospheric Urban Heat Island (AUHI) which includes the difference in air temperature pattern between urban and rural areas [27].

There are three main factors involved in the emergence of Urban Heat Islands, which can generally be classified into three types: variables with a temporary effect, such as wind speed and cloud cover; variables with a constant and stable effect, such as green spaces, building materials, and height; finally, variables with a periodic or cyclical effect, such as solar radiation and heat sources caused by human activity [20,28]. In general, the heat produced on the surface comes from the sun in the form of solar radiation, large industries and factories, cars, air conditioning systems, and other sources related to human activities [29]. Since the first observations of Urban Heat Islands, many policies have been implemented to cool cities. This issue has gained attention due to the increasing urban population and the desire to create more comfortable and high-quality urban environments, as well as concerns related to energy consumption and minimizing energy usage to achieve thermal comfort. Additionally, with predictions of global warming and increasing urban temperatures worldwide, urban cooling techniques may become even more important in the coming decades [30,31].

With the introduction of thermal remote sensing technology, it became possible to investigate UHIs indirectly using satellite and aircraft platforms on continental and global scales. Therefore, new avenues for observing heat islands and analyzing their causes and factors were provided [17,32]. The study of the UHI phenomenon by measuring the temperature of the Earth’s surface was initially conducted using the NOAA AVHRR satellite, with a spatial resolution of 1.1 km, suitable for small-scale city-level temperature mapping. Subsequently, thermal infrared data from Landsat and Landsat ETM+ with resolutions of 120 and 60 m, respectively, have been used to extract the surface temperature.

In recent years, the population of Urmia city has increased from 300,746 people in 1983 to 794,300 people in 2021 [33]. According to a previous study, population growth leads to changes in land cover, increased building construction, and higher traffic volumes, which hinder the flow of wind streams and cause the retention of pollutants and hot air masses, contributing to the formation of Urban Heat Islands [34]. Therefore, along with the physical development of the city, significant changes have occurred in the land cover of the city due to massive constructions. This situation has led to the destruction of the vast vegetation cover and, consequently, the development of UHIs in the city, which have adverse effects on the lives of the city dwellers. Therefore, this study aims to investigate the effects of urban land use on the creation of UHIs in Urmia city. The results of this study would be beneficial for urban planners and policymakers to consider in their land use planning programs to reduce the possibilities of creating UHIs and prevent their harmful effects.

2. Materials and Methods

2.1. Case Study

Urmia is one of the largest cities in the northwest of the country. It is located between 37 degrees and 28 min to 37 degrees and 35 min north latitude and 44 degrees and 58 min to 45 degrees and 8 min east longitude [35]. According to the 2021 census, the city has a population of 794,300 [33]. Urmia city spreads across a plain next to Urmia Lake, extending 70 km in length and 30 km in width. It is bordered by the city of Salmas to the north, Ashnoviyeh and Naqadeh to the south, Lake Urmia to the west, and Turkey to the east. The elevation of Urmia city is 1332 m above sea level [36]. The location of Urmia city is shown in Figure 1.

2.2. Methods

This research aims to identify the effective factors for the creation of UHIs in Urmia city, extract their thermal patterns, and simultaneously monitor urban land use changes from 1990 to 2023. To achieve this goal, remote sensing technology has been adapted, as it offers valuable information and facilitates the study of UHI phenomena from a thermal perspective. The primary component of this method is electromagnetic waves. By analyzing the reflection of these waves from various objects, diverse information, including surface temperature distribution patterns in Urmia have been collected. Therefore, Landsat satellite images, including Landsat 5 (TM) and Landsat 8 (TIRS/OLI) sensors were used. These images consist of two groups of spectral bands: reflective and thermal bands (see

Table 1. Characteristics of satellite images.

Satellite	Sensor	The Number of Bands	Imaging Date	Row	Path
Landsat5	TM	7	02 July 1990	34	169
Landsat8	TIRS/OLI	11	13 July 2023	34	169

). The thermal bands were employed to identify the temperature of the ground surface and heat island, while the reflective bands were utilized to apply indicators for image processing. The TM data from Landsat 5 satellite comprise 6 bands, whereas the OLI/TIRS data from Landsat 8 comprise 11. Thermal band 6 data from Landsat 5 and band 10 data from Landsat 8, with a wavelength of 10.60 μm, were used to calculate the surface temperature distribution patterns of the city.

• Converting DNs (Digital Numbers) to Radiance

DNs are unitless integers used by satellites to record the relative radiance values, typically ranges from 0 to 255 for 8-bit images [37]. This step involves converting DNs (raw images) to radiance, which is calculated using the following equation:

• Relationship 2–1

Lλ = MLQcal + AL − Oi

(1)

In the above relation, Lλ represents spectral radiation, ML is the multiple coefficient specific for Radiance Mult Band corresponding to the selected band, which is found in the metadata of the studied image. Qcal is a digital number at the desired pixel location (the raw image corresponding to the selected band). AL is the cumulative coefficient for the Radiance Add Band, and Oi is the offsets issued by USGS for the calibration of the TIRS bands [38].

• Computation of Brightness Temperature

“Brightness temperature is the temperature required by a blackbody to emit the same amount of radiation per unit of surface area as the body being observed” [39,40]. Spectral radiation is converted to brightness temperature using the following relationship:

• Relationship 2–2

$$\mathrm{BT}={\displaystyle \frac{K2}{Ln\left(\frac{k1}{l\lambda }+1\right)}}$$

(2)

In this relationship, BT represents the at-sensor brightness temperature in degrees Kelvin, where K1 and K2 are the band-specific thermal conversion constants obtained from the metadata. $\mathrm{l}\mathsf{\lambda }$ is the corrected thermal band radiance, which represents the actual radiance emitted from the Earth’s surface after atmospheric correction.

The Earth’s surface temperature was calculated using ENVI5.1 software, employing Planck’s method. Planck’s algorithm is one of the most basic methods for calculating the Earth’s surface temperature in remote sensing [41,42].

• Relationship 2–3

$$\mathrm{LST}={\displaystyle \frac{\mathrm{BT}}{1+\left(\frac{\mathsf{\lambda }\ast Tb}{\rho .}\right)ln\mathsf{\epsilon }}}$$

(3)

LST is the land surface temperature (K); BT is the at-sensor brightness temperature (K); λ is the wavelength of the emitted radiance; ρ is the constant (h * c/σ), where h is Planck’s constant, c is the speed of light, and σ is the Stephan-Boltzmann consent, equal to 1.438 × 10⁻² mK; and ε is the spectral emissivity.

• NDVI Calculation

This index is sensitive to the amount and condition of vegetation cover and is calculated using Equations (2)–(4). Its values range between 1 and −1.

• Relationship 2–4

$$\mathrm{NDVI}={\displaystyle \frac{\left(p4-p3\right)}{\left(p4-p3\right)}}$$

(4)

In the above relation, p3 represents the ability to emit in red band, and p4 represents the ability to emit in the near-infrared band [43].

The processes of preparing the maps for evaluating UHIs in Urmia city is presented in Figure 2.

3. Results

In this study, radiometric and atmospheric corrections were first performed on the images in the studied area. Then, based on the distribution of land use in terms of land cover and land use within Urmia city, four major land use classes were selected as follows: barren lands, constructed lands, agricultural lands, and garden lands. Then, the final images were processed using the supervised method of neural network and a maximum 3 × 3 filter, and subsequently, the land cover maps were extracted. To assess the accuracy of the classification, an errors matrix was first constructed, followed using two indicators: overall accuracy and kappa coefficient. The error matrix is shown in

Table 2. Kappa coefficient and total accuracy of images.

Images	Kappa Coefficient	Total Accuracy
1990	0.9014	0.9117
2023	0.9382	0.9571

.

By conducting analyses on land use changes in GIS, a comparison of land use and land cover changes during the studied period was obtained using classified images from consecutive years (Figure 3).

The changes in land use within Urmia city indicated that the area of constructed lands, which was approximately 2900.79 hectares in 1990, expanded to 7636.89 hectares at the end of the period. Conversely, garden lands, spanning 1749.63 hectares in 1990, have decreased to 284.97 hectares in 2023. Additionally, agricultural land, covering 1166.78 hectares in 1990, decreased to 153.98 hectares in 2023 (

Table 3. Land use changes in Urmia from 1990 to 2023 (in hectares).

Land Use	(1990)	(2023)
constructed lands	2900.79	7636.89
garden lands	1749.63	284.97
agricultural land	1166.78	274.98
barren lands	2533.62	153.98
Total	8350.82	8350.82

).

The surface temperature map of Urmia city for 1990 is depicted in Figure 4 and

Table 4. Thermal classes in Urmia city in 1990.

Thermal Class	Area (Hectares)	Area Percentage
20.57–25	509.70	6.10
25–30	1447.89	17.34
30–35	5370.51	64.31
35–42.46	1022.72	12.25

. The map illustrates very hot temperature ranges represented by dark brown coloration. The findings reveal that the temperature range of 20.57–25 degrees Celsius is about 6.10 percent of the area, while the range of 25–30 degrees Celsius covers about 17.34 percent. Additionally, the temperature range of 30–35 degrees Celsius encompasses approximately 31.64%, and the range of 35–42.46 degrees Celsius is around 12.25%. Consequently, the results indicate evaluated temperature in the center, west, and southwest of Urmia city, attributed to the presence of constructed and barren lands. On the other hand, the lowest temperatures are observed in garden lands.

The results of the UHI extraction conducted in Urmia city in 2023, depicted in Figure 5 and

Table 5. Thermal classes in Urmia city in 2023.

Thermal Class	Area (Hectares)	Area Percentage
16.23–25	72.53	0.87
25–30	1957.23	23.43
30–35	3046.90	36.49
35–41.22	3274.16	39.21

, reveal that the lowest recorded temperature in the city is 23.16 degrees Celsius, while the highest recorded temperature is 47.22 degrees Celsius. Therefore, UHI formation is observed throughout most areas of Urmia, primarily attributed to the expansion of urban construction.

The Relationship Between Land Use and Surface Temperature

Surface temperature is influenced by various surface conditions. Areas with more vegetation tend to exhibit lower surface temperatures compared to those without vegetation. Vegetation acts as a natural air conditioner by absorbing solar energy and releasing water vapor through transpiration. However, according to

Table 6. The relationship between temperature and land use in 1990 and 2023.

Year	Class	Range	Min	Max	Mean	Std
1990	Constructed lands	18.41	22.79	20.41	22.79	1.82
	Garden lands	15.64	20.57	36.21	20.57	2.48
	Agricultural lands	16.74	24.90	41.64	24.90	2.06
	Barren lands	18.00	24.46	42.46	24.46	2.68
2023	Constructed lands	22.66	26.15	45.82	31.37	2.78
	Garden lands	16.23	23.16	44.66	33.80	2.79
	Agricultural lands	17.08	30.14	45.85	38.68	2.33
	Barren lands	18.51	29.62	47.22	39.52	2.49

, the ongoing construction process in Urmia city over the past decades has resulted in significant physical development, expanding from 2900.79 hectares in 1990 to 7636.89 hectares in 2023, consequently posing numerous environmental challenges. The conversion of farmlands, agricultural areas, and green spaces on the city’s outskirts into urban development has led to the replacement of natural surfaces with impermeable ones such as buildings and roads. This has resulted in a decline in vegetation cover, the disruption of natural surface cooling mechanisms, and an overall increase in surface temperature. In 1990, the highest average temperature recorded for barren lands was 42.46 degrees Celsius, while the lowest temperature was observed in garden lands at 20.57 degrees Celsius. By 2023, the highest average temperature for barren lands had risen to 47.22 degrees Celsius, with the lowest temperature for garden lands recorded at 23.16 degrees Celsius.

Moran’s index was employed to investigate the spatial correlation of UHIs. This index not only yields global values but also interprets the local areas [44]. Moran’s index provides two types of outputs: numerical and graphical. Generally, if the value of Moran’s index is close to +1, it indicates spatial autocorrelation and a cluster pattern in the data. Conversely, if the value of Moran’s index is close to −1, it suggests that the data are dispersed and separated [45]. To implement Moran’s index, first, the Zonal Statistics As Table algorithm was conducted to analyze the relationship between temperature and land use. Then, the temperature layer was converted into a point layer using the Raster to Point algorithm. Next, the Spatial Autocorrelation algorithm was applied to perform Moran’s correlation. After that, the Cluster and Outlier Analysis algorithm was utilized to create the Moran image layer. Finally, the Hot Spot Analysis algorithm was employed to determine the spatial distribution of the indicators.

The graphical output, illustrated in Figure 5 and Figure 6, demonstrates whether the data are scattered or clustered. In this method, the null hypothesis posits that there is no spatial clustering between the element values associated with the considered geographic features. When the significance level is very small and the calculated Z value exceeds the confidence limit, the null hypothesis can be rejected. A positive value of Moran’s index (greater than zero) indicates spatial clustering, while a negative value of (less than zero) suggests a scattered pattern.

Table 7. Moran’s index for the studied periods.

Moran	1990	2023
Moran’s Index	0.8049	0.8897
Expected Index	−0.000011	−0.000011
Variance	0.000001	0.000001
z-score	945.0113	1044.5373
p-value	0.0000	0.0000

presents Moran’s index in the studied periods.

According to the Moran index table, the UHIs in Urmia city exhibit a cluster pattern in all periods. The calculated values of z for the studied periods are 945.01 and 1044.53, respectively. Thus, the high z value and the low significance levels confirm the hypothesis of spatial autocorrelation between the UIH data in Urmia city. The graphical results of the Moran model, presented in Figure 6 and Figure 7, depict output clusters with abbreviations such as HH, indicating clusters of high values or positive spatial autocorrelation at the 99% confidence level. Conversely, LL indicates clusters of low values or negative spatial autocorrelation at the 99% confidence level. HL indicates scatter, wherein high values are surrounded by low values, while LH represents a single cell with a low value surrounded by high values, distinguishing them from each other, and are statistically significant at the 5% level.

According to the findings, a part of the city shows no discernible pattern; in other words, there is a lack of spatial autocorrelation. This situation is evident throughout the city during the studied periods, but it is more pronounced in the city center. In terms of single cells, HL and LH indicate non-clustering within the city. Regarding the LL spots’ condition in 1990, this index is observed in areas with dense vegetation and coincides with points of the lowest temperature in the city. In 2023, it is observed in the peripheral areas of the city. The hotspot index has been used to identify the spatial distribution of clusters in the study area. This method employs a type of z-score. For a positive and statistically significant z-score, the larger the z-score, the more clustered the values become, forming a hot spot. For a negative and statistically significant z-score, the smaller the z-score, indicating a more severe clustering of low values, which represent cold spots.

The hot and cold spots index results of the UHIs in Urmia city reveal that in 2023, hot spots are concentrated in the central, northeastern and northern regions. These areas are significantly impacted by land use, contributing to the city’s elevated temperature. Conversely, cold spots are identified in the southwestern and northwestern parts of the city with a confidence level of 99%. This suggests that both natural conditions and human activities play a role in the formation of these cold spots (see Figure 8 and Figure 9).

4. Discussion

Rapid urbanization and development have led to the substantial replacement of green spaces, agricultural lands, and forests by urban areas [46,47]. Particularly in large cities, urban growth accelerates land cover changes at a remarkable pace. Surfaces with high heat capacity, constructions materials, reduced albedo, and the presence of heat sources from human activities collectively contribute to elevating temperature in urban centers compared to surrounding areas [48,49]. Therefore, due to the harmful effects of UHIs, researchers are seeking solutions to reduce the intensity of their effects. Some researchers have suggested increasing surfaces with the capability of returning solar energy and urban greenery to mitigate UHI impacts [50].

Zhang et al. (2012) found a negative correlation between Normalized Difference Vegetation Index (NDVI) and surface radiant temperature, indicating higher temperature in commercial and industrial areas [51]. Conversely, areas with abundant green and blue spaces play a vital role in mitigating the UHI effect. The findings of this study corroborate similar research findings.

This research specifically examines the impact of land use on UHI formation in Urmia city, employing Landsat satellite data for analysis.

The results of this study reveal a significant rise in temperature across constructed lands in 2023, with an increase of 4.62 degrees Celsius compared to 1990. Concurrently, there is a noticeable upward trend in the area covered by constructed lands. Garden lands also experienced a substantial temperature hike, registering a 13.23 degrees Celsius in 2023 compared to the average temperature recorded in 1990. However, despite this temperature surge, the area covered by garden lands decreased by 1464.66 hectares in 2023 compared to 1990. Moreover, the spatial autocorrelation index results indicated a cluster pattern of UHI effects in Urmia city across all periods, with a confidence level of 99%. The hotspot analysis results also reveal that hot spots are concentrated in the central, northeast, and northern sectors of the city, while cold spots are prevalent in the southwest and northwest areas, with a confidence level of 99%. Through the examination of Urmia city’s thermal islands, it becomes evident that the distribution pattern of surface temperatures has undergone notable transformations during the study period. This shift favors the expansion of the areas characterized by very high temperatures while concurrently diminishing the extent of regions with very cool temperatures. Consequently, while the extent of moderate and very cold temperatures has decreased significantly, very hot areas have expanded considerably. Thus, it can be inferred that the increase in constructed land, residential areas development and urban expansion, has led to a reduction in garden lands. This phenomenon has contributed to the formation of Urban Heat Islands in Urmia city.

In this study, Planck’s algorithm was employed to extract land surface temperature in Urmia using Landsat data. Analysis of the retrieved temperature data revealed an escalating trend in the distribution of UHIs across Urmia city in 2023. Particularly in the central area, UHIs have intensified into a vast regional heat island, attributed to the interconnected urban constructions. In light of these findings, it is reasonable that city growth plans and programs consider the establishment of satellite cities or towns surrounding the central city, independent of the commercial center. This approach aims to mitigate the formation of a large-scale regional Urban Heat Islands by distributing urban development across multiple areas rather than concentrating it solely within the central city.

Furthermore, the correlation analysis conducted between the retrieved Land Surface Temperature (LST) and Normalized Difference Vegetation Index (NDVI), revealed that green lands mitigate the UHI effect, whereas constructed lands exacerbate it. The findings align with the results of Grover and Singh’s study (2015) [52], which demonstrated that areas with vegetation and water exhibit lower temperatures compared to constructed and barren lands. This research underscores the direct relationship between land surface temperature and the normalized index of vegetation cover, indicating that urban greening should receive an increased emphasis in future urban planning and development initiatives. Enhancing urban vegetation cover emerges as a potent strategy to mitigate the creation and impact of UHIs by providing a cooler urban environment. This assertion is further corroborated by the findings of Harun (2020) [53], who highlighted the significant role of land cover in influencing temperature fluctuations within cities. Therefore, increasing urban greenery, as one of the nature-based solutions, is commonly used to improve urban thermal environment and address UHI issues [54,55].

Furthermore, the calculation of Moran’s index shows that during the studied periods, the heat islands in Urmia city demonstrate a clustered pattern, representing the strongest Urban Heat Island phenomenon. Therefore, to mitigate the creation and effects of UHIs, urban planning and development programs should prioritize increasing greenery and reducing the concentration of built-up areas in Urmia city.

5. Conclusions

Due to extensive urbanization, increased human activities, and extensive urban constructions, the phenomenon of Urban Heat Islands has emerged as one of the most significant urban challenges worldwide. Identifying the most influential factors contributing to the creation of UHIs across urban areas is crucial for preventing their formation and mitigating their harmful effects. The aim of the current research was to investigate the role of land use in UHI formation using satellite data and to identify areas with elevated temperatures in Urmia city. In addition, by analyzing urban land use and identifying critical environmental areas in the Urmia, this study tried to provide guidelines for urban spatial and environmental planning programs to prevent UHI formation and its effects. However, in this regard, it is imperative to provide a comprehensive policy and approach for urban planners and designers to incorporate into land use planning programs. Policies such as increasing public green spaces along streets and roads, expanding tree canopies, and green roofs can improve shading and natural cooling. In addition, monitoring and enforcing environmental standards in the industries, particularly the automobile industry, and promoting optimal public transportation usage can serve as valuable initiatives in addressing UHI effects. Applying permeable or reflective materials to pavements and roofs minimize heat absorption, while incorporating water features helps cool areas through evaporation. Moreover, strategic zoning can allocate green areas in densely built-up zones. Public awareness can also encourage community involvement and the adoption of these cooling techniques on provided properties and within their neighborhoods.

Therefore, the results of this study, which model the UHI effect in Urmia city, provide essential guidance for urban planners and decision makers seeking the most effective methods to control temperatures and prevent the adverse effects of UHIs on the local environment.

This study focused on the role of major urban land uses in heat emission and the formation of UHIs. While urban surface temperature primarily depends on land cover, it is important to recognize that some three-dimensional elements such as urban trees, building height, wind direction, and solar radiation intensity also significantly contribute to the UHI effect. Therefore, future studies should comprehensively consider the impacts of various multidimensional factors, including both artificial and natural elements, on Urban Heat Islands.