Urban Microclimates in a Warming World: Land Surface Temperature (LST) Trends Across Ten Major Cities on Seven Continents

Abstract

Understanding microclimatic changes driven by urbanization is critical in the context of global warming and climate change. This study investigates the land surface temperature (LST), the normalized difference vegetation index (NDVI), and changes in land use types for 10 major cities across seven continents between 2001 and 2021. Utilizing MODIS satellite data processed on the Google Earth Engine (GEE) platform, the analysis focused on yearly median values to examine variations in LST during the day and night, as well as temperature dynamics across different land types, including vegetation and bare land. The global mean LST trend from 2001 to 2021, derived from Terra MODIS MOD11A2 data, was found to be 0.025 °C/year. The analysis of daytime and nighttime (nocturnal) land surface temperature (LST) trends across the ten cities examined in this study reveals notable variations, with most cities exhibiting an increasing trend in LST within urban mosaics. Airports exhibited a mean daytime land surface temperature (LST) that was 2.5 °C higher than surrounding areas, while industrial zones demonstrated an even greater temperature disparity, with an average increase of 2.81 °C. In contrast, cold spots characterized by dense vegetation showed a notable cooling effect, with LST differences reaching −3.7 °C. Similarly, proximity to water bodies contributed to temperature mitigation, as areas near significant water sources recorded lower daytime LST differences, averaging −4.09 °C. A strong negative correlation was found between NDVI and LST, underscoring the cooling effect of vegetation through evapotranspiration and shading. This study provides a comprehensive global perspective on the commonalities of urban temperature dynamics in cities across diverse geographical regions and climates, contributing to a deeper understanding of how urbanization and land use changes influence surface temperatures and climate change.

1. Introduction

Global climate change, influenced by urbanization and land use changes, has significant effects on land surface temperature (LST) worldwide [1,2,3]. Between 2001 and 2020, the global average land surface temperature increased at a rate of 0.26 °C to 0.34 °C per decade, with the Arctic permafrost regions experiencing the most rapid warming at the start of the 21st century [4]. While urbanization contributes minimally to overall global temperature trends, it significantly intensifies land surface warming at local and regional scales, particularly in densely urbanized areas such as eastern China, the eastern United States, and Europe [5]. This localized warming effect is expected to become more pronounced as urban land expansion accelerates, with projections indicating a 78% to 171% increase in urbanized areas by 2050. Such expansion is anticipated to exacerbate urban heat island effects, disproportionately increasing extreme heat risks. These risks will be particularly severe in the tropical regions of the Global South, where nearly half of the future urban population will be exposed to heightened thermal stress [6].

Urbanization, particularly through the urban heat island (UHI) effect, contributes to microclimatic changes, especially in surface temperatures [7,8]. The UHI effect becomes more pronounced during the day and in summer [9], exacerbating heatwaves [10] and increasing heat stress in urban populations [11], which are expected to intensify with future climate change [12]. It has been reported that urban areas can be up to 5 °C warmer than rural areas depending on local climate and urban density [7,13]. In Tianjin, China, urbanization caused a ground temperature increase of 13 °C over 15 years as the area transitioned from agricultural land to a built-up environment [14]. Research in North America has shown that urbanization has a more pronounced effect on nighttime temperatures [15,16]. Generally, nighttime temperatures in urban areas are higher compared to rural areas due to reduced vegetation, human activities, and impervious surfaces [17,18,19]. Additionally, urban architecture and geometry [20] and haze pollution [21] have been significant factors influencing nighttime UHI.

Changes in land types are another important factor influencing LST. Urban and built-up areas increase LST due to the UHI effect, with built-up surfaces (e.g., concrete, asphalt, etc.) absorbing and retaining heat [22,23,24]. Increased vegetation cover (e.g., forests and grasslands) lowers LST through cooling effects such as transpiration and shading [25,26,27]. Water bodies also act as cooling zones, reducing LST [28]. Spatiotemporal variations show that LST varies significantly across regions and seasons. For example, arid regions and barren lands tend to have higher LST values compared to vegetated areas [22,25,26]. Seasonally, LST is more influenced in summer and lower in winter [28,29]. Human activities and urbanization, which involve the replacement of natural landscapes with impervious surfaces [23,24], and land use changes, such as the conversion of grasslands to cultivated lands, increase LST [22]. The composition and distribution of landscape patterns are key factors in determining LST. Large built-up lands tend to increase LST, while mixed areas with significant vegetation reduce it [25,28]. In European cities, a 5% reduction in bare land has led to a 1–2 °C drop in nighttime LST [30].

With the help of remote sensing technologies, long-term changes in geographical areas (terrain type, surface features, LST, and NDVI) can be easily analyzed. Remote sensing and geographic information systems (GISs) have been used for spatiotemporal pattern analysis over many years, particularly for vegetation and land use changes [31,32]. In environmental and agricultural studies, MODIS datasets are widely used, particularly for LST and NDVI analyses. The high temporal and spatial resolution of MODIS data makes it an essential tool for thermal analysis [33,34,35,36]. Many environmental studies conducted with this dataset have shown a strong negative correlation between LST and NDVI [33,37]. In another study, the vegetation temperature condition index (VTCI), an integrated index of LST and NDVI, was used for drought monitoring, and strong correlations with past rainfall events were observed [38]. In agricultural studies, MODIS data are often preferred for yield forecasting. In a study conducted in the Corn Belt region of the United States, yield forecasts for corn and soybean were made using LST and NDVI derived from MODIS. The study found a positive correlation between mid-summer NDVI and yields, and a negative correlation with daytime LST. High coefficients of determination (R²) were achieved for both crops [39]. In another urban and environmental study, a negative correlation was found between LST and NDVI, while a positive correlation was identified with the normalized difference built index (NDBI) [36].

Although there are many studies in the literature examining spatial and temporal trends of LST and its correlation with land use types [40,41,42,43], there is a lack of intercontinental comparative analyses. Moreover, there is a notable gap in research that investigates the global impacts of urbanization on different land types. The cloud-based infrastructure provided by the GEE platform enables the efficient analysis of large datasets, playing a significant role in global analyses over long periods.

This study uses GEE and MODIS datasets to conduct a global-scale analysis of LST and NDVI, offering a comprehensive view of thermal and ecological dynamics. Through satellite-based data, the changes in LST, NDVI, and land types were analyzed in detail for 10 major cities, providing valuable global insights into the microclimatic impacts of urbanization. The findings of this study have the potential to guide urban planning and climate change adaptation policies.

2. Study Area, Data and Methodology

2.1. Study Area

This study focuses on ten major cities representing the largest urban centers across different continents (Figure 1): Cairo (Africa), Chongqing (Asia), Delhi (Asia), Istanbul (Europe/Asia), Melbourne (Oceania), Mexico City (North America), Moscow (Europe), Nuuk (North America), São Paulo (South America), and Tokyo (Asia). These cities were selected to capture diverse climatic zones, including humid subtropical (Tokyo and Chongqing), semi-arid/tropical (Delhi), tropical monsoon (São Paulo), subtropical highland (Mexico City), hot desert (Cairo), Mediterranean (Istanbul), humid continental (Moscow), oceanic (Melbourne), and Arctic (Nuuk) (

Table 1. Selected cities in different continents with their populations.

Continent	City	Country	Population	Climate
Asia	Tokyo	Japan	37,115,000	Humid Subtropical
Asia	Delhi	India	33,807,400	Semi-Arid/Tropical
Asia	Chongqing	China	32,100,000	Humid Subtropical
South America	São Paulo	Brazil	22,806,700	Tropical Monsoon
North America	Mexico City	Mexico	22,505,300	Subtropical Highland
Africa	Cairo	Egypt	22,623,900	Hot Desert (Arid)
Europe/Asia	Istanbul	Turkey	16,047,400	Mediterranean
Europe	Moscow	Russia	12,712,300	Humid Continental
Oceania	Melbourne	Australia	5,315,600	Oceanic
Greenland	Nuuk	Greenland	19,783	Arctic

Footnote: Population data have been gathered from World Population Review (https://worldpopulationreview.com/cities).

). The cities also reflect varied urbanization patterns and socio-economic contexts, making the analysis globally representative [44]. To comprehensively analyze vegetation areas and their interaction with urban environments, this study includes not only each city’s metropolitan region but also its surrounding green areas, enabling a holistic evaluation of urban and peri-urban environmental changes.

2.2. Data and Method

Three MODIS datasets were utilized in this study to evaluate the land surface temperature (LST), vegetation cover (NDVI), and true color reflectance across the selected cities (Figure 2). These datasets were selected due to their high temporal resolution, global coverage, and proven reliability in capturing key environmental variables over large spatial extents. The temporal scope of the analysis focused on the years between 2001 and 2021 to capture long-term changes.

The LST data were sourced from the MOD11A2 product [45], which provides 8-day composite thermal readings at a 1 km spatial resolution. This dataset includes both daytime (LST_Day_1 km) and nighttime (LST_Night_1 km) surface temperature measurements, crucial for assessing urban heat island effects. Each pixel value in MOD11A2 represents the average of all corresponding MOD11A1 LST pixels collected during that 8-day period. LST values were converted to Celsius using a scaling factor provided by the dataset’s documentation with the following equations (Equations (1) and (2)):

$$LSTday(°C)=0.02\ast LST\_Day\_1km-273.15$$

(1)

$$LSTnight(°C)=0.02\ast LST\_Day\_1km-273.15$$

(2)

Vegetation cover was analyzed using the NDVI band from the MOD13A2 product [46], which provides bi-weekly vegetation indices at a 1 km spatial resolution. This dataset captures vegetation health and density, enabling an assessment of urban green spaces and surrounding areas. NDVI values were scaled to their standard range of 0–1 using Equation (3):

$$NDVI(0-1)=0.0001\ast NDVI$$

(3)

Additionally, true color imagery was derived from the MOD09A1 product [47], which includes 8-day composite surface reflectance data for the red, green, and blue bands. This dataset, provided at a 500 m spatial resolution, was normalized to a 0–1 range for an accurate visualization of urban and vegetated areas. The composite images, combining all variables, were clipped to the boundaries of each city and visualized for analysis. Separate layers for LST (day and night) and NDVI were also generated for a detailed assessment of urban environmental conditions.

The MCD12Q1.061 MODIS Land Cover Type Yearly Global dataset [48] at a 500 m resolution was used to assess changes across various land types in relation to daytime LST and nighttime LST across years. This dataset provides yearly classifications based on the International Geosphere-Biosphere Programme (IGBP), which were reclassified for focused analysis. These categories were further isolated using specific masks: Urban and Built-up Lands, Vegetation, Bareland, Permanent Snow and Ice, Sea and Water sources (

Table 2. Original and re-classified land usage classes.

Re-classified Land Usage Classes	Original LC_Type1 Classes
Urban and Built-up Lands	13
Vegetation	1 + 2 + 3 + 4+5 + 6+7 + 10 + 12 + 14
Bareland	8 + 9 + 16
Permanent Snow and Ice	15
Sea and Water sources	11 + 17

Footnote: “Original LC_Type1 Classe” refers to the IGBP classification scheme used in the LC_Type1 layer of the MCD12Q1.061 MODIS Land Cover dataset.

).

2.3. Data Handling and Quality Control

2.3.1. Resampling

The datasets used in this study—MOD11A2 (land surface temperature), MOD09A1 (surface reflectance), and MCD12Q1.061 (land cover)—have different spatial resolutions, which necessitates resampling to ensure consistency in spatial analysis. MOD11A2 has a native resolution of 1 km; MOD09A1 is available at 500 m, and MCD12Q1.061 is provided at 500 m. To facilitate direct comparisons and minimize spatial inconsistencies, all datasets were resampled to a common resolution of 500 m using the bilinear interpolation method.

2.3.2. Quality Assurance (QA) Flag Detection

Each MODIS product comes with a quality assurance (QA) flag, which is used to indicate the reliability of each observation [49]. These flags provide information about the conditions under which data were collected, including cloud cover, data validity, and sensor performance.

QA Flag Value	Action
For a value of 00, data are valid:	Proceed with using the LST data.
For a value of 01, data are low quality:	Proceed with caution, potentially applying additional quality control (temporal smoothing).
For a value of 10 or 11, LST is not produced:	Do not use the data for LST calculations. The pixel is either affected by clouds or sensor errors, and the LST values are unreliable. All remaining QA bits and SDS layers are undefined in these cases.

2.3.3. Temporal Smoothing (1-Year Median Average)

The median smoothing method is a robust statistical way that helps reduce noise and smooth the data, particularly when working with time-series data from remote sensing products [50]. The median filter [51] is especially useful in removing spurious outliers caused by temporary data issues, such as sensor calibration errors or atmospheric interference, because the median is less sensitive to extreme values than the mean. The 1-year median (annual smoothing) is a simple method but helps reduce the influence of outliers and noise while preserving the key long-term trends. It is applied as follows (Equation (4)):

$$X_{median}\left(Y\right)=median(X\left(Y_1\right),X\left(Y_2\right),\dots ,X(Y_n\left)\right)$$

(4)

where

$X_{median}\left(Y\right)$ is the median value for year Y;

$(X\left(Y_1\right), X\left(Y_2\right),\dots ,X(Y_n\left)\right)$ represent all the data values for year Y.

2.3.4. Linear Regression Model

To calculate the trend, this study applied a linear regression model to each LST dataset (both daytime and nighttime). The method involved first adding a time band to each image, representing the number of years since the start of the study period (1 January 2001). This time band was then used as an independent variable in the linear regression, with LST as the dependent variable. The reduce(ee.Reducer.linearFit()) function in Google Earth Engine was used to calculate the slope of the linear trend (α) for each pixel (Equation (5)), providing a measure of how LST has changed over time.

$${LST}_{Day,Night}(°C)=Slope\left(\alpha \right)\times Time\left(Year\right)+Offset\left(\beta \right)$$

(5)

3. Results

3.1. Land Surface Temperature Trend Evaluation on a Global Scale

The global mean land surface temperature (LST) trend from 2001 to 2021, as derived from Terra MODIS MOD11A2 data, was found to be 0.025 °C per year, indicating a gradual increase in land surface temperatures over the two decades. This warming trend highlights significant regional variations, with noticeable hotspots in northern latitudes, such as Siberia and parts of North America, where warming is particularly pronounced. Conversely, some regions, such as central Africa, parts of India, and South America, exhibit cooling trends during the same period. Changes in surface temperatures can be influenced by various factors studied in the previous literature, including global events such as El Niño and La Niña, teleconnection patterns, solar variability, radiation feedbacks, as well as human activities like land use changes. But, in this study, the focus is on the surface temperature trends in specific cities, including Tokyo, Delhi, Chongqing, São Paulo, Mexico City, Cairo, Istanbul, Moscow, Melbourne, and Nuuk, with a particular emphasis on the impacts of land use changes and urban mosaics (Figure 3).

3.2. City-Based Evaluation

3.2.1. Cairo

Figure 4 illustrates changes in Cairo’s land cover, vegetation, and surface temperatures between 2001 and 2021, emphasizing the impacts of urbanization and climate trends over two decades. The true color images in the top row (2001 and 2021) highlight the spatial extent of urban with vegetation development in Cairo. Between 2001 and 2021, Cairo experienced significant changes in land cover and surface temperatures, which were shaped by urbanization and targeted greening initiatives. Areas such as New Cairo, New Heliopolis City, and El Obour City underwent extensive vegetation projects that integrated green spaces into urban designs. These efforts resulted in a notable reduction in daytime land surface temperatures (LSTs) in vegetated regions, dropping from 37.94 °C to 35.87 °C. The cooling effect underscores the role of increased vegetation in moderating urban heat. However, central urban zones, characterized by dense impervious surfaces, showed minimal changes in daytime LST, remaining around 33.35 °C. Nighttime trends revealed a contrasting dynamic, with LST rising in vegetated regions from 16.95 °C to 20.68 °C, likely due to heat retention in surrounding built-up areas. In central Cairo, nighttime LST also increased, from 20.48 °C to 22.43 °C, reflecting intensified urban heat island effects. Long-term analysis indicated trends of −0.13 °C/year for daytime and +0.19 °C/year for nighttime temperatures in greened areas.

From 2001 to 2021 in Cairo, urban areas showed no significant daytime surface temperature trend but experienced a yearly nighttime increase of 0.1 °C, highlighting urban heat island effects. Vegetation areas saw a slight cooling trend during the day (−0.04°/year) and warming at night (0.12 °C/year) (Figure 5). Bareland exhibited a modest warming trend both during the day (0.02 °C /year) and at night (0.1 °C/year).

3.2.2. Chongqing

Chongqing, characterized by its mountainous terrain and dispersed urban layout, exhibits a highly heterogeneous land surface temperature (LST) profile. From 2001 to 2021, significant changes in LST were observed due to urbanization and varied land use patterns (Figure 6). The Sahikiaopu residential district, located in the southwestern part of Chongqing, experienced an increase in daytime LST from 28.83 °C to 30.85 °C, while nighttime LST saw a dramatic rise from 14.07 °C to 25.70 °C. However, in the city center, the daytime LST changed only slightly from 27.02 °C to 27.03 °C. This minimal change is due to the cooling effect of the Yangtze River passing through the city center. In forested northeastern areas, daytime temperatures slightly decreased from 20.39 °C to 19.80 °C, likely due to consistent vegetation cover. However, nighttime LST in these areas remained nearly stable, dropping marginally from 10.59 °C to 10.35 °C. In contrast, southern regions dominated by cropland showed increases in both daytime and nighttime LST. Daytime temperatures rose from 27.10 °C to 28.56 °C, while nighttime LST increased from 13.09 °C to 16.07 °C. A similar study states that Chongqing, as a mountainous city, features a complex terrain with intervalley denudation platforms and sloping landforms. These significant topographical variations play a crucial role in shaping its urban thermal environment, influencing how urbanization impacts land surface temperatures and heat distribution [52]. The analysis reveals that in Chongqing’s city center, the annual land surface temperature (LST) trends are +0.18 °C/year during the day and +0.06 °C/year at night. In the scattered settlements along the urban periphery, these trends are slightly lower, with daytime increases of +0.12 °C/year and nighttime increases of +0.01 °C/year. This disparity underscores the variability in the urban heat island effect, which is influenced by the size and density of the metropolitan area. In the northeastern forested regions, the trends differ, with a decrease in daytime LST by −0.065 °C/year, contrasting with a slight nighttime increase of +0.036 °C/year. These results reflect the cooling effect of vegetation during the day and minor heat retention at night. The broader Yangtze River region’s industrial distribution, with factories, power plants, and storage facilities scattered along its banks, contributes to localized thermal variations.

From 2001 to 2021, Chongqing exhibited a surface temperature increase of 0.04 °C/year during the day and a decrease of −0.05 °C/year at night (Figure 7). Vegetation experienced a notable cooling trend of −0.07 °C/year during the day and −0.64 °C/year at night. Bareland showed a slight cooling trend of −0.03 °C/year during the day, with no nighttime trend. Water body surface temperatures decreased by −0.04 °C/year during the day and −0.03 °C/year at night.

3.2.3. Delhi

Delhi’s urban thermal landscape exhibits significant changes between 2001 and 2021, reflecting the impact of urbanization and land use dynamics. In consistently urbanized areas, the daytime LST decreased slightly from 32.48 °C in 2001 to 31.78 °C in 2021, while nighttime LST increased from 21.76 °C to 24.73 °C (Figure 8). Hotspots, particularly in 2021, were concentrated around the airport and nearby industrial zones such as Bawana, Rohini, and Narela, where daytime LST reached 34.52 °C and nighttime LST stood at 20.02 °C, indicating intensified urban heat island effects in these high-activity areas [53]. Notably, the northern “Restricted Forest” area demonstrated cooling trends during the day, with LST dropping from 30.87 °C in 2001 to 28.71 °C in 2021, while nighttime LST increased slightly from 20.24 °C to 22.35 °C. The Yamuna River played a crucial role in moderating temperatures, providing a cooling effect in its vicinity [54]. For instance, in Seelampur, areas closer to the river in 2021 recorded lower average LST values (daytime: 30.71 °C; nighttime: 21.7 °C) compared to more inland locations (daytime: 31.93 °C; nighttime: 22.23 °C). In the city center, the daytime trend is negligible, at −0.004 °C/year, while nighttime LST exhibits a modest increase of +0.047 °C/year. This minimal change reflects the stabilization of thermal conditions in the urban core, possibly due to long-term urban saturation and limited new development. In contrast, newly urbanized and industrialized areas, such as the NTPC water treatment plant, Sector 17 industrial zones, Kundli, and the DSIIDC industrial hub, demonstrate more significant trends, with daytime LST rising by +0.10 °C/year and nighttime LST by +0.065 °C/year. Adjacent forested regions near the city center show a cooling trend during the day, at −0.037 °C/year, attributed to the preservation of vegetation, which aids in moderating daytime heat. However, nighttime LST in these areas is increasing, with a trend of +0.061 °C/year, likely due to thermal retention in nearby urban areas that influences nocturnal cooling.

From 2001 to 2021 in Delhi, urban areas showed a slight cooling trend of −0.01 °C/year during the day and a significant warming trend of 0.12 °C/year at night, highlighting urban heat island effects (Figure 9). Vegetation experienced a cooling of −0.05 °C/year during the day and a warming of 0.09 °C/year at night, reflecting changes in land use and evapotranspiration. Bareland exhibited no daytime temperature trend but warmed significantly at night by 0.13 °C/year.

3.2.4. Istanbul

Istanbul, a city bridging two continents, displays distinct variations in land surface temperatures (LSTs) across its regions due to diverse land use and urbanization patterns. In the densely populated southern areas, the daytime LST increased from 23.41 °C in 2001 to 25.31 °C in 2021, while nighttime LST decreased slightly from 12.09 °C to 11.75 °C (Figure 10). Conversely, in the northern forested regions, the daytime LST showed a minor decrease, dropping from 18.91 °C in 2001 to 18.64 °C in 2021, while the nighttime LST rose from 9.23 °C to 11.07 °C. These shifts underline the contrasting effects of urban expansion and vegetation on LST dynamics. Hotspots on the European side, such as Bahçelievler, Esenyurt, and Fatih, and on the Asian side, like Ataşehir and Samandıra, correspond to densely urbanized zones. A particularly noteworthy development is the newly constructed airport (Istanbul Airport), where significant land use changes drove dramatic increases in LST. At the airport site, the daytime LST soared from 18.89 °C in 2001 to 28.51 °C in 2021, with nighttime LST rising from 10.01 °C to 10.87 °C. This translates to a linear trend of +0.38 °C/year during the day and +0.10 °C/year at night, driven by the replacement of natural surfaces with impervious infrastructure. Meanwhile, in the northern Çatalca forests, trends were less pronounced, with daytime LST increasing by +0.031 °C/year and nighttime LST by +0.046 °C/year, reflecting the stabilizing influence of forest cover amidst regional changes.

From 2001 to 2021, in Istanbul, urban areas showed a significant surface temperature increase of 0.13 °C/year during the day and 0.03 °C/year at night, indicating urban heat island effects (Figure 11). Vegetation warmed by 0.03 °C/year during the day and 0.04 °C/year at night, while bareland exhibited similar trends, with increases of 0.03 °C/year during the day and 0.04 °C/year at night. Water body surface temperatures rose by 0.06 °C/year both during the day and at night, reflecting regional warming trends.

3.2.5. Melbourne

In Melbourne, land surface temperatures (LSTs) around the city center have shown moderate increases over the past two decades. In 2001, the average daytime LST was 23.74 °C, and nighttime LST was 8.17 °C (Figure 12). By 2021, these values had risen to 24.13 °C during the day and 8.73 °C at night, reflecting a slight warming trend associated with urbanization and infrastructure development. Hotspot regions primarily include industrial and commercial areas such as Laverton North, Melbourne Airport, Somerton, Campbellfield, Essendon, and Dandenong South. Among these, industrial facilities (recycling, logistics, and warehouses) on Boundary Road in Laverton North recorded the most dramatic rise, with daytime LST increasing from 22.49 °C in 2001 to 30.82 °C in 2021 and nighttime LST rising from 7.29 °C to 8.79 °C. This sharp increase underscores the thermal impact of industrial activities and impervious surface expansion. In contrast, forested northern regions showed minimal temperature changes, maintaining their cooling effect. In 2001, daytime and nighttime LSTs in these areas were 13.13 °C and 7.49 °C, respectively, compared to 13.17 °C and 7.87 °C in 2021. The analysis of temperature trends in Melbourne indicates gradual warming in the city center, with a daytime LST trend of +0.030 °C/year and a nighttime trend of +0.033 °C/year. The most significant increases were observed in hotspot areas such as Dandenong South, where daytime LST rose by +0.17 °C/year and nighttime LST by +0.067 °C/year, reflecting intense industrial activities and urban development. An intriguing contrast emerges in the western Quandong region, which is primarily composed of agricultural lands. Here, a cooling trend of −0.12 °C/year during the day is observed, alongside a minimal nighttime increase of +0.007 °C/year. This suggests that the retention of vegetation and agricultural practices may be mitigating daytime heat, though slight thermal retention is evident during the night.

From 2001 to 2021, in Melbourne, urban areas showed surface temperature increases of 0.02 °C/year during the day and 0.03 °C/year at night, reflecting urban warming trends (Figure 13). Vegetation experienced a slight cooling of −0.02 °C/year during the day and a warming of 0.02 °C/year at night, while bareland cooled during the day by −0.03 °C/year and warmed at night by 0.03 °C/year. Water body surface temperatures showed no daytime trend but increased significantly at night by 0.05 °C/year, indicating nighttime oceanic warming near urban areas.

3.2.6. Mexico

In Mexico, surface temperature patterns across different urban areas have shown notable changes between 2001 and 2021, reflecting the impacts of urban growth and industrial activities (Figure 14). In neighborhoods such as Santa Cruz, Alfredo Del Mazo, and Benito Juarez, the average daytime LST increased from 31.12 °C in 2001 to 32.21 °C in 2021, while nighttime LST rose from 11.27 °C to 12.65 °C, highlighting the intensification of urban heat island effects. The hotspot analysis identified the tultitlán de mariano Escobedo, Industrial Cuamatla, and the Toluca industrial region as significant contributors to rising temperatures. In this area, daytime LST increased from 31.51 °C in 2001 to 32.61 °C in 2021, while nighttime LST rose from 10.15 °C to 11.83 °C, driven by industrial expansion and reduced vegetation cover. In contrast, forested areas near the city center, such as Sierra Guadalupe and Cañada del Otatal, showed a slight decrease in daytime LST, dropping from 20.64 °C to 19.92 °C over the same period. However, nighttime LST in these regions increased from 8.27 °C to 9.59 °C, indicating that while forests mitigate daytime heat, nearby urban developments may be influencing nighttime warming. The analysis of temperature trends in the Mexico City area highlights notable shifts in thermal dynamics due to rapid urbanization, particularly in previously rural regions like Los Heroes Tecámac. This area transitioned from a rural mosaic with agricultural fields in 2001 to an urbanized landscape by 2021, resulting in high temperature trends of +0.10 °C/year during the day and +0.21 °C/year at night. In contrast, a nearby forested area, located just 6 km away from the urbanized zones, demonstrated cooling trends during the day (−0.054 °C/year), while nighttime temperatures showed a slight warming trend (+0.0597 °C/year). This stark contrast within a relatively small spatial range illustrates the variability in land surface temperature (LST) changes, driven by differences in land use and vegetation cover.

From 2001 to 2021, in Mexico, urban areas experienced surface temperature increases of 0.05 °C/year during the day and 0.07 °C/year at night, highlighting urban heat island effects (Figure 15). Vegetation showed a slight cooling trend of −0.02 °C/year during the day and a warming of 0.04 °C/year at night. Bareland warmed by 0.02 °C/year during the day and 0.07 °C/year at night. Water body surface temperatures showed notable increases of 0.11 °C/year during the day and 0.12 °C/year at night.

3.2.7. Moscow

Moscow’s urban thermal dynamics have undergone notable changes between 2001 and 2021, reflecting the combined effects of urban infrastructure and climate shifts (Figure 16). In the city center, daytime LST decreased significantly from 14.75 °C in 2001 to 11.99 °C in 2021, while nighttime LST rose from −1.34 °C to 2.09 °C over the same period. This pattern suggests a cooling trend during the day, but a warming trend at night, which may result from the thermal retention of urban surfaces. Hotspot analysis for 2021 highlights areas of significant thermal activity, such as the Yaroslavskiy Railway Terminal, with a daytime LST of 17.31 °C and a nighttime LST of 2.79 °C, and Aviapark shopping mall, with 16.74 °C during the day and 2.64 °C at night. In contrast, Sokolniki Park, one of the largest green spaces in the city center, exhibited a marked cooling trend. Daytime LST in the park dropped from 15.33 °C in 2001 to 11.28 °C in 2021, while nighttime LST decreased from 5.07 °C to 2.43 °C, showcasing the cooling benefits of urban green spaces. In Moscow, the analysis of temperature trends reveals significant spatial variations driven by land use and redevelopment activities. In residential and commercial areas such as Lefortovo, Meshchansky, Presnensky, and Tagansky districts, daytime LST has been increasing at a rate of +0.11 °C/year, while nighttime LST shows a more moderate rise of +0.05 °C/year. An intriguing exception is the Danilovsky District, which was historically an industrial zone with aging and abandoned buildings. Recent redevelopment initiatives in this area, emphasizing eco-friendly urban design and green space integration, have reversed the typical warming trends seen elsewhere. The daytime LST trend in Danilovsky has decreased by −0.039 °C/year, while the nighttime trend, although positive, remains relatively low at +0.048 °C/year.

From 2001 to 2021, in Moscow, urban areas exhibited a cooling trend of −0.07 °C/year during the day and a slight warming of 0.03 °C/year at night (Figure 17). Vegetation showed a cooling of −0.04 °C/year during the day and a warming of 0.04 °C/year at night. Bareland temperatures decreased by −0.04 °C/year during the day and increased slightly by 0.01 °C/year at night. Water body surface temperatures displayed a cooling trend of −0.08 °C/year during the day and a significant warming of 0.09 °C/year at night.

3.2.8. Nuuk

Nuuk, the capital of western Greenland, demonstrates significant changes in land surface temperature (LST) from 2001 to 2021, highlighting the impact of Arctic warming and polar amplification (Figure 18). In Nuuk, the LST increased from −2.71 °C (day) and −7.58 °C (night) in 2001 to 1.49 °C (day) and −4.00 °C (night) in 2021. The western part of Greenland, in particular, showed even more pronounced warming. In 2001, the LST for this region was 2.55 °C (day) and −5.90 °C (night), increasing to 8.31 °C (day) and −3.15 °C (night) in 2021. This marked rise underscores the pronounced vulnerability of Greenland’s coastal areas, which have less snow and ice cover compared to inland regions. In contrast, Greenland’s interior exhibits a different pattern. In 2001, LST in the interior was recorded at −30.62 °C (day) and −36.64 °C (night), declining further to −33.04 °C (day) and −39.19 °C (night) by 2021. This cooling trend is attributed to differences in atmospheric and oceanic circulation systems affecting inland areas, compared to the warming observed along the coasts. Polar amplification is especially prominent in coastal regions due to feedback mechanisms like reduced snow and ice cover, which enhance heat absorption and accelerate warming. The most significant warming has been observed in the easternmost region, particularly in the Vandrebrok area. Here, the LST trends show an increase of +0.11 °C/year during the day and +0.053 °C/year at night, highlighting the impact of reduced ice cover and enhanced polar amplification in these coastal regions. In contrast, interior Greenland displays cooling or stable trends, with daytime LST decreasing by approximately −0.11 °C/year and nighttime trends showing a minimal change, around −0.01 °C/year.

From 2001 to 2021, in Nuuk, vegetation showed a consistent warming trend of 0.03 °C/year both during the day and at night, while bareland warmed more significantly, with increases of 0.05 °C/year during the day and 0.06 °C/year at night (Figure 19). Water body surface temperatures decreased by −0.03 °C/year during the day but increased slightly by 0.02 °C/year at night. Snow and ice areas experienced cooling trends of −0.05 °C/year during the day and −0.06 °C/year at night, reflecting ongoing cryospheric changes in the region.

3.2.9. Sao Paulo

In São Paulo, areas such as Vila Mariana, Tatuapé, and Guarulhos experienced slight changes in land surface temperature (LST) between 2001 and 2021 (Figure 20). In these urban regions, the average LST decreased from 33.16 °C (day) and 17.77 °C (night) in 2001 to 32.66 °C (day) and 16.35 °C (night) in 2021, reflecting localized cooling likely influenced by urban planning and increased vegetation in certain pockets. In the green corridor of Parque Ecológico do Tietê, which is intersected by the Tietê River, the LST values also declined. The park recorded 29.49 °C (day) and 16.84 °C (night) in 2001, compared to 29.37 °C (day) and 15.13 °C (night) in 2021. This cooling trend emphasizes the role of protected natural areas in mitigating urban heat island effects. Hotspot analysis across São Paulo’s urban and semi-urban sprawl highlights industrial and densely populated zones such as Ribeirão Preto, Barretos, São José do Rio Preto, Franca, Campinas, and Cruzeiro. These regions show elevated LSTs relative to their surroundings, largely attributed to industrial and commercial developments. Examining temperature trends in São Paulo reveals distinct variations between the city center and its surrounding regions. In the urban core, daytime LST trends show a decline of −0.05 °C/year, while nighttime temperatures remain nearly stable at −0.001 °C/year. This suggests some success in mitigating daytime urban heat island effects, potentially through urban planning and vegetation efforts. However, the surrounding municipalities, such as Guarulhos, Arujá, Osasco, Taboão da Serra, and São Bernardo do Campo, located 20–30 km from the center, exhibit rising LSTs, with trends of +0.08 °C/year (day) and +0.002 °C/year (night). These suburban and peri-urban areas are likely experiencing increased urbanization and industrial development. In the northern agricultural zones, the trends are even more pronounced. Daytime LST shows a significant rise of +0.13 °C/year, while nighttime LST decreases slightly by −0.02 °C/year. This divergence highlights the influence of changing land use, such as the intensification of agricultural activities and reduced vegetative cover, which exacerbates temperature increases during the day while allowing more heat dissipation at night.

From 2001 to 2021 in São Paulo, urban areas experienced a warming trend of 0.05 °C/year during the day and a slight increase of 0.02 °C/year at night (Figure 21). Vegetation showed a slight warming of 0.02 °C/year during the day, with no significant trend at night. Bareland exhibited a modest increase of 0.02 °C/year both during the day and at night. Water body surface temperatures warmed by 0.02 °C/year during both day and night.

3.2.10. Tokyo

In Tokyo, regions such as Nakano, Ōta, Nishitōkyō, and Meguro stand out due to their relatively high land surface temperature (LST) values (Figure 22). In 2001, the average LST was 27.15 °C (day) and 9.08 °C (night), which increased slightly to 27.23 °C (day) and 10.66 °C (night) by 2021. These trends reflect urban heat island effects in the densely populated metropolitan areas. The Imperial Palace National Gardens, Tokyo’s largest central green space, exhibited a cooling influence, although some warming was observed over two decades. In 2001, the LST values were 21.01 °C (day) and 12.03 °C (night), rising to 22.53 °C (day) and 12.53 °C (night) in 2021. Examining temperature trends in Tokyo highlights significant warming, particularly in the Nakano district. Here, land surface temperature (LST) trends show increases of +0.10 °C/year (day) and +0.062 °C/year (night), reflecting urban densification and the intensifying urban heat island effect in this central area. In contrast, the Imperial Palace National Gardens, despite being a large green area in the heart of the city, also exhibit warming trends, albeit at a slower pace. LST trends here are +0.04 °C/year (day) and +0.073 °C/year (night). The steady increase indicates that even expansive urban greenery is not fully immune to broader climatic influences and rising temperatures in the surrounding urban environment.

From 2001 to 2021 in Tokyo, urban areas showed a warming trend of 0.03 °C/year during the day and 0.04 °C/year at night, indicating urban heat island effects (Figure 23). Vegetation experienced a slight warming of 0.01 °C/year during the day and 0.03 °C/year at night. Bareland showed a modest increase of 0.02 °C/year during the day, but a cooling trend of −0.01 °C/year at night. Water body surface temperatures rose by 0.06 °C/year during the day and 0.04 °C/year at night, reflecting significant oceanic warming.

3.3. Land Cover Change and Transformations

When analyzing the trends presented in the previous section alongside the Land Use Change Table (

Table 3. Land use change by percent (2001–2021 difference).

City	Urban Change (%)	Vegetation (%)	Bareland (%)	Waterbody (%)
Cairo	6.1	4990.8	−3.0	0.0
Chongqing	48.7	18.4	−16.0	19.0
Delhi	6.4	−4.8	29.4	−18.6
Istanbul	10.4	9.8	−15.0	−20.3
Melbourne	10.8	0.1	−5.0	−33.9
Mexico	7.5	−0.6	1.5	−51.9
Moscow	5.6	−13.7	21.2	0.0
Nuuk	0.0	5.6	−11.7	5.9
Sao Paulo	0.9	1.5	−4.4	−16.4
Tokyo	0.1	5.5	14.4	−21.4

), a discernible pattern emerges regarding the relationship between land cover transformations and the daytime land surface temperature (LST Day). The findings indicate that reductions in vegetation and water bodies are consistently correlated with an increase in LST Day, whereas expansions in urban areas and bareland exhibit a more complex and context-dependent influence, potentially increasing or, in some cases, even reducing LST Day. A particularly notable case is Cairo, where an extraordinary increase in vegetation cover (4990.8%) is primarily attributed to large-scale greening initiatives rather than a naturally occurring trend. In contrast, Mexico exhibits the most pronounced decline in water bodies, with a reduction of 51.9%, which corresponds to an observable increase in LST Day within water body areas (r = 0.11). This trend aligns with the well-documented thermal regulation function of water surfaces, whose loss contributes to localized warming effects. Similarly, cities such as Chongqing (−16.0%) and Istanbul (−15.0%) have experienced reductions in bareland, largely due to urban expansion or afforestation. However, the impact on LST Day differs between these two cases, reflecting variations in urban form and climate. Notably, the correlation between urban expansion and LST Day also varies significantly, with Istanbul exhibiting a trend (r = 0.13) and Chongqing demonstrating a lower trend (r = 0.4).

3.4. Effect of the NDVI Value on Both Daytime and Nighttime LSTs

Table 4. Relationship between NDVI classes and daytime land surface temperatures (LSTs).

Cities	NDVI ≤ 0	0 < X ≤ 0.2	0.2 < NDVI ≤ 0.6	0.6 < NDVI ≤ 1
Cairo.	-	35.61 ± 1.4	33.4 ± 1.9	-
Chongqing	20.62 ± 1.8	25.42 ± 2.5	24.26 ± 3.8	22.07 ± 3.4
Delhi	-	31.81 ± 0.7	30.20 ± 1.4	29.61 ± 0.4
Istanbul	15.70 ± 0.0	24.29 ± 2.2	21.45 ± 1.6	19.33 ± 1.0
Melbourne	-	23.44 ± 1.1	22.71 ± 1.3	-
Mexico	21.6 ± 0.8	29.81 ± 3.3	25.7 ± 3.1	21.9 ± 4.5
Moscow	-	12.7 ± 1.6	11.03 ± 1.5	9.43 ± 1.0
Nuuk	−20.19 ± 8.3	−1.44 ± 4.2	0.46 ± 2.13	-
Sao Paulo	24.9 ± 1.5	26.61 ± 3.0	31.92 ± 1.8	25.62 ± 3.2
Tokyo	-	24.17 ± 1.48	23.78 ± 2.8	16.13 ± 3.5

presents the relationship between NDVI (normalized difference vegetation index) classes and land surface temperatures (LSTs) across different cities. The general trend suggests that higher NDVI values, which indicate greater vegetation cover, correspond to lower LST values. For example, in cities like Istanbul and Chongqing, LST decreases as NDVI increases, with Istanbul showing a drop from 24.29 °C (NDVI 0 < X ≤ 0.2) to 19.33 °C (NDVI 0.6 < X ≤ 1). Similarly, in Moscow, LST decreases from 12.7 °C to 9.43 °C as NDVI increases. Unlike daytime LST trends, nighttime temperatures do not always show a clear linear relationship with NDVI (

Table 5. Relationship between NDVI classes and nighttime land surface temperatures (LSTs).

Cities	NDVI ≤ 0	0 < X ≤ 0.2	0.2 < NDVI ≤ 0.6	0.6 < NDVI ≤ 1
Cairo	-	19.47 ± 1.6	21.08 ± 0.7
Chongqing	14.66 ± 0.71	14.91 ± 1.0	13.65 ± 2.8	−12.56 ± 2.3
Delhi	-	23.90 ± 0.4	21.02 ± 2.0	20.88 ± 1.5
Istanbul	13.47 ± 0.19	11.60 ± 0.9	10.38 ± 1.0	10.10 ± 0.8
Melbourne	-	9.18 ± 0.4	9.32 ± 0.6	-
Mexico	15.36 ± 1.99	11.63 ± 3.1	9.12 ± 4.2	11.59 ± 5.3
Moscow	-	1.48 ± 1.1	1.39 ± 1.0	1.78 ± 0.7
Nuuk	−24.93 ± 9.1	−6.56 ± 3.3	−5.58 ± 1.7	−5.0 ± 1.5
Sao Paulo	22.71 ± 1.5	21.47 ± 2.5	18.00 ± 1.50	15.95 ± 1.7
Tokyo	-	12.27 ± 0.9	10.38 ± 1.8	9.74 ± 3.6

). In many cities, such as Istanbul, Delhi, and São Paulo, higher NDVI values still correspond to lower nighttime LST, reinforcing the cooling effect of vegetation. For instance, in Istanbul, LST decreases from 13.47 °C (NDVI ≤ 0) to 10.10 °C (NDVI 0.6 < X ≤ 1), and in São Paulo, it drops from 22.71 °C (NDVI ≤ 0) to 15.95 °C (NDVI 0.6 < X ≤ 1). This suggests that vegetation helps retain cooler temperatures at night, reducing heat retention in urban areas. However, some cities exhibit anomalies. In Cairo, nighttime LST is slightly higher (21.08 °C) for moderate NDVI values (0.2 < NDVI ≤ 0.6) compared to low NDVI (19.47 °C). In Nuuk, an extremely cold region, LST remains negative across all NDVI classes, though it rises slightly as NDVI increases. The most unusual case is Chongqing, where the highest NDVI category shows a negative LST (−12.56 °C), which is due to local climatic influences, such as dense vegetation in mountainous regions and seasonal variations.

3.5. Common Hotspots and Cold Spots in City Landscapes

Table 6. Land surface temperature (LST) variations across different urban zones in selected megacities.

City	City Center	Airport	Industrial Zone	Green Space	Water Bodies
Cairo	32.19	35.47 (Cairo International Airport)	36.52 (Industrial Zone Badr City)	31.50 (Gharb el-Golf)	30.58 (Nil River)
Chongqing	26.61	35.05 (Chongqing Jiangbei Airport)	35.10 (Gecaoba)	23.11 (Wumaguicao)	22.35 (Yangtze River)
Delhi	32.39	34.51 (Indira Gandhi Airport)	33.05 (Mundka Industrial Area)	29.24 (Central Ridge Reserve Forest)	27.55 (Yamuna River)
Istanbul	24.19	28.07 (Istanbul Airport)	25.89 (Ataşehir Industrial Zone)	18.77 (Çatalca Forest)	15.63 (Bosphorus)
Melbourne	24.63	25.70 (Melbourne Airport)	27.70 (Truganina)	20.27 (North Warrandyte)	20.52 (Yarra River)
Mexico	32.16	33.19 (Mexico Airport)	30.77 San Luis Tlatilco Industrial Zone	20.03 (Sierra de Guadalupe State Park)	22.30 (Laguna de Zumpango)
Moscow	12.44	Vnukovo Airport (12.85)	15.55 (Podolsk Industrial Area)	10.77 (Sokoliki Park)	11.15 (Moscow River)
Sao Paulo	31.65	32.39 (Sao Paulo Airport)	32.80 (Viela Sabesp)	29.17 (Parque Ibirapuera)	28.05 (Jurubatuba River)
Tokyo	22.31	23.89 (Tokyo International Airport)	26.51 (Toshibacho)	21.81 (Tokyo Imperial Palace)	23.62 (Arakawa River)

highlights land surface temperature (LST) variations across different urban zones in selected megacities, revealing distinct thermal patterns. Industrial zones and airports consistently exhibit the highest temperatures, underscoring their role in intensifying the urban heat island effect. For instance, Cairo’s Industrial Zone Badr City (36.52 °C) and Chongqing’s Gecaoba (35.10 °C) record the highest LSTs within their respective cities, exceeding temperatures in city centers and even airports. Airports, with their vast concrete surfaces and high energy consumption, also contribute significantly to urban heating, as seen in Cairo International Airport (35.47 °C) and Chongqing Jiangbei Airport (35.05 °C). In contrast, Moscow, due to its colder climate, registers much lower LSTs across all zones, with its industrial area (15.55 °C) and airport (12.85 °C) showing significantly lower values than other cities. Green spaces and water bodies provide substantial cooling effects, with temperature reductions of up to 12 °C compared to surrounding urban areas. Istanbul’s Çatalca Forest (18.77 °C) and Mexico City’s Sierra de Guadalupe State Park (20.03 °C) demonstrate the strong cooling impact of vegetation, significantly lowering LSTs relative to their city centers. Similarly, water bodies like the Bosphorus (15.63 °C) in Istanbul and the Yangtze River (22.35 °C) in Chongqing act as natural temperature regulators. However, their cooling effectiveness varies. Tokyo’s Arakawa River (23.62 °C) and Cairo’s Nile River (30.58 °C) exhibit higher temperatures, likely due to heat absorption from surrounding infrastructure.

The analysis of land surface temperature (LST) differences between urban mosaics and different urban zones reveals notable variations across cities (

Table 7. Land surface temperature (LST) difference between urban mosaic and different urban zones.

	Difference Value ( °C)				Percentage (%)
City	Airport	Industrial Zone	Green Space	Water Bodies	Airport	Industrial Zone	Green Space	Water Bodies
Cairo	3.28	4.33	−0.69	−1.61	10%	13%	−2%	−5%
Chongqing	8.44	8.49	−3.5	−4.26	32%	32%	−13%	−16%
Delhi	2.12	0.66	−3.15	−4.84	7%	2%	−10%	−15%
Istanbul	3.88	1.7	−5.42	−8.56	16%	7%	−22%	−35%
Melbourne	1.07	3.07	−4.36	−4.11	4%	12%	−18%	−17%
Mexico	1.03	−1.39	−12.13	−9.86	3%	−4%	−38%	−31%
Moscow	0.41	3.11	−1.67	−1.29	3%	25%	−13%	−10%
São Paulo	0.74	1.15	−2.48	−3.6	2%	4%	−8%	−11%
Tokyo	1.58	4.2	−0.5	1.31	7%	19%	−2%	6%

and Figure 24). Airports consistently exhibit higher temperatures than their respective city centers, with differences ranging from 0.41 °C in Moscow to 8.44 °C in Chongqing. This suggests that large expanses of impervious surfaces, such as runways and terminals, contribute significantly to the urban heat island (UHI) effect. The highest percentage increase is observed in Chongqing (32%), followed by Istanbul (16%) and Cairo (10%), indicating that climatic conditions and local land cover types play a crucial role in the extent of temperature disparities. Industrial zones also tend to be hotter than the urban mosaic, with differences peaking at 8.49 °C in Chongqing and 4.33 °C in Cairo. This trend can be attributed to the high concentration of factories, warehouses, and paved surfaces, which absorb and retain heat. However, in some cases, such as Mexico (−1.39 °C), industrial zones exhibit a cooling effect due to the presence of green buffers and different green land use policies. The highest percentage increases occur in Moscow (25%) and Tokyo (19%), reinforcing the idea that industrial activities, combined with urban density and materials used, significantly impact local LST variations. Conversely, green spaces and water bodies generally exhibit lower temperatures than urban mosaics, highlighting their crucial role in mitigating the UHI effect. Green spaces in Istanbul (−5.42 °C), Mexico (−12.13 °C), and Melbourne (−4.36 °C) show substantial cooling, with percentage reductions as high as −38% in Mexico. Water bodies further amplify this effect, with the most significant reductions in Istanbul (−8.56 °C) and Mexico (−9.86 °C). Interestingly, Tokyo shows a slight warming effect (+6%) for water bodies, which is influenced by specific urban hydrology dynamics and heat retention in the nearby urban fabric.

4. Discussion

4.1. General Characteristics of Urban Thermal Dynamics

In this study, a comparative analysis of urban thermal dynamics in 10 major cities across seven continents was conducted. Within this context, shared characteristics and distinct differences shaped by urbanization, land use, vegetation cover, and climate change were identified. Across all cities examined, it was observed that UHIs intensified particularly during nighttime. This phenomenon is attributed to the retention of heat due to impervious surfaces and reduced vegetation in urbanized areas. The literature has extensively documented the similar microclimatic effects of urbanization on UHIs, particularly its association with elevated surface temperatures [2,3,7,8,15,16].

The analysis of daytime land surface temperature (LST) trends across the ten cities examined in this study reveals notable variations in urban thermal dynamics. While some cities, such as Cairo (0 °C/year) and Moscow (−0.07 °C/year), exhibit minimal or negative trends, others, including Chongqing (0.4 °C/year) and Istanbul (0.13 °C/year), show a more pronounced increase in daytime LST within urban mosaics. Cities like Delhi (−0.1 °C/year) demonstrate a slight cooling trend, whereas Melbourne (0.02 °C/year), Mexico (0.05 °C/year), Nuuk (0.05 °C/year), São Paulo (0.05 °C/year), and Tokyo (0.03 °C/year) indicate modest warming patterns. In cities like Mexico City and Melbourne, newly urbanized areas significantly reflected the thermal impacts of replacing natural surfaces with impervious materials [55], leading to notable warming trends in these regions. Variations were identified in the patterns and magnitudes of LST changes, attributed to geographical and associated climatic differences. For instance, Arctic warming observed in Nuuk showed significant increases in both daytime and nighttime LSTs, consistent with the effects of polar amplification [56]. In contrast, tropical cities like Delhi and Mexico City exhibited more stable daytime trends due to the broader coverage and continuous presence of vegetation and the associated evapotranspiration. Similarly, the literature underscores that spatiotemporal variations induce significant regional and seasonal modifications in LST [25,26,57]. In cities with complex terrains, such as Chongqing, LST changes were found to be heterogeneous, with cooler daytime LSTs identified particularly in forested and mountainous areas, highlighting the critical influence of elevation and vegetation on LST. In Tokyo and SaoPaulo, despite the absence of a spatial urban expansion (%1 and %0.9), both daytime and nighttime urban heat island (UHI) effects have increased over the past 20 years. This suggests that even without an expansion of the urban mosaic, the intensification of the UHI effect can be attributed to global warming and climate change.

4.2. Climatic Conditions

Indisputably, regional climatic conditions constitute a fundamental determinant of LST variations [58]. Moscow, with its colder climate, has the lowest overall temperatures, whereas warmer cities like Cairo, Delhi, Mexico City, and São Paulo see city center temperatures exceeding 30 °C. Melbourne and Tokyo fall into a mid-range category, showing moderate temperature variations across different land use zones. The findings suggest that industrial and airport zones require targeted mitigation strategies such as increased vegetation cover and reflective materials to reduce heat buildup. In addition, the integration of water bodies and vegetation generates a powerful synergistic cooling effect, helping to regulate temperatures in urban environments [59]. Interestingly, in Delhi, the daytime land surface temperature (LST) exhibits a decreasing trend, while the nighttime LST is showing an upward trajectory, indicating a shift in the diurnal thermal pattern. This trend is consistent with findings from previous studies, [28], which documented similar variations in LST over the region. Additionally, [60] reported the formation of a cold island effect over Delhi during the daytime, reinforcing the observed cooling trend in daytime LSTs. By strategically planning and optimizing the layout of green spaces, cities can significantly enhance cooling efficiency within residential areas, creating more comfortable and sustainable living conditions.

4.3. Ecological Assessment of Urban Microclimates

4.3.1. The Relationship Between Vegetation and Urban Surface Temperatures

The analysis highlights a significant negative correlation between NDVI values and daytime/ nighttime LSTs, illustrating the critical role of vegetation in influencing urban surface temperatures [61]. Areas with low NDVI, typically barren land or sparsely vegetated regions, exhibit the highest LST values, indicating the heat-retaining nature of these surfaces. As NDVI increases, representing areas with denser vegetation, LST decreases, reflecting the cooling effect of vegetation through evapotranspiration and shading [62]. Water bodies consistently display the lowest LST values when present, further emphasizing their ability to regulate temperatures. This relationship is evident across diverse climatic and geographical contexts, from arid regions like Cairo to temperate and densely populated cities like Tokyo and Moscow. Dense vegetation areas, when present, exhibit the strongest cooling effect, demonstrating their importance in mitigating urban heat islands and enhancing climate resilience [63].

4.3.2. The Role of Green and Water Spaces in Urban Heat Regulation

The urban heat island (UHI) effect is evident in all cities, with industrial zones [64] and airports [65] consistently exhibiting the highest land surface temperatures (Figure 24). These areas contribute significantly to urban heat stress due to dense infrastructure, a lack of vegetation, and high energy consumption. Cairo (36.52 °C) and Chongqing (35.10 °C) stand out with particularly high industrial zone temperatures. While city centers generally have elevated temperatures, they are often surpassed by industrial zones and airports, as seen in cities like Cairo, Chongqing, Delhi, and São Paulo. However, Moscow presents an exception, with a significantly lower city center LST (12.44 °C), reflecting its colder climate and different urban composition. Green spaces and water bodies play a crucial role in mitigating heat, though their effectiveness varies across cities [66,67,68]. Green spaces consistently show lower LSTs, with Istanbul’s Çatalca Forest (18.77 °C) and Delhi’s Central Ridge Reserve Forest (29.24 °C) offering noticeable cooling effects. In some cases, the difference is striking; for example, Mexico City’s green spaces are up to 12.13 °C cooler than its urban core. Similarly, water bodies such as the Bosphorus (15.63 °C) in Istanbul and the Moscow River (11.15 °C) demonstrate significant temperature reductions. However, in cities like Tokyo (23.62 °C) and Cairo (30.58 °C), water bodies do not always exhibit strong cooling effects due to factors such heat absorption from surrounding dense urban infrastructure [69].

The findings of this study also highlight the dual role of vegetation in urban environments, specifically its ability to provide cooling during the day and retain heat at night. The intensity of urban heat island (UHI) effects across cities was primarily influenced by factors such as land use policies, urban design, and geographical characteristics, all of which shape the thermal dynamics of urban spaces [70]. In cities like Cairo and Moscow, the expansion of green infrastructure, such as increasing tree cover and green spaces, demonstrated clear and measurable cooling benefits [71]. These measures will not only mitigate daytime heat but also improve the overall thermal comfort of urban areas. To address significant nighttime and daytime UHI effects observed in cities like Delhi, Sao Paulo, and Istanbul, it is essential to develop strategies that increase the albedo (reflectivity) of urban surfaces. This can be achieved through the use of reflective materials for roads, roofs, and pavements, which will help reduce heat retention during the night and mitigate the intensity of nighttime UHI [72]. Tailoring urban interventions to the specific thermal dynamics of each city is crucial for effective mitigation. It is vital that cities adopt localized adaptation strategies that consider the unique climatic, geographical, and socio-economic contexts of each urban area. In this regard, climate-sensitive urban expansion will be key for maintaining desirable thermal comfort and minimizing UHI effects [73]. Furthermore, as seen in the case of Mexico City’s rural–urban transformations, proactive measures must be taken to preserve agricultural land and promote sustainable development practices. Ensuring the conservation of open green spaces, forests, and agricultural land will not only mitigate the effects of UHI but also promote more sustainable and resilient urban growth [74].

5. Conclusions

In conclusion, this study observed an overall increase in land surface temperature (LST) both during the day and at night across major cities on all continents. Even in the absence of urban expansion, daytime and nighttime temperatures can increase due to global warming and climate change. Given the compounded effects of this, the rise in temperature exacerbates the impact of heatwaves, posing significant challenges for urban populations.

5.1. Contribution to the Scientific Community

The present study makes a significant contribution to the scientific community by offering a comprehensive, comparative analysis of urban thermal dynamics across ten major cities representing seven continents. By integrating multi-temporal MODIS datasets and applying consistent spatial analyses, the research bridges global and local perspectives on land surface temperature (LST) trends and urban heat island (UHI) effects. It highlights the interplay between urbanization, vegetation cover (NDVI), land use, and climatic zones, offering novel insights into how these factors jointly influence urban thermal behavior both during the day and at night. Unlike many previous studies focused on individual cities or regions, this study systematically reveals cross-continental patterns and discrepancies, such as the unexpected rise in nighttime UHI intensity even in cities with minimal urban expansion. The inclusion of ecological assessments, such as the cooling roles of vegetation and water bodies, further deepens our understanding of urban microclimates. By identifying key thermal hotspots and evaluating the impacts of industrial zones, airports, and green infrastructure, this study provides a robust scientific basis for climate-responsive urban planning and sustainable development. The findings contribute valuable empirical evidence to the ongoing discourse on global urban climate adaptation and resilience strategies in the face of accelerating climate change.

Therefore, integrating city-specific assessments into urban planning strategies is critical for addressing the challenges brought about by these changing climate conditions. A proactive approach that incorporates climate resilience into urban planning will be essential to safeguarding the well-being of urban populations and ensuring the long-term sustainability of urban areas. This will help foster resilience, reduce the severity of heat-related impacts, and improve the quality of life for future generations.

5.2. Guidelines for Future Research Directions

• Future research should delve deeper into understanding the specific drivers of urban heat island (UHI) effects across diverse regions, particularly in cities with distinct climatic, geographical, and socio-economic conditions. This would involve comparing UHI intensities across different climates (e.g., tropical, arid, and temperate) and topographies (e.g., coastal, mountainous, and flat) and understanding how these factors interact with land use and urban design. More detailed, city-specific case studies will help identify context-sensitive UHI mitigation strategies.
• While the benefits of green infrastructure are well documented in this study, there is a need for more longitudinal studies that track the long-term effects of implementing vegetation-based solutions, such as green roofs, urban parks, and tree cover, on UHI reduction. Future research should assess how these interventions evolve over time, considering factors such as growth cycles, maintenance, and the impact of climate change on their effectiveness. This would provide valuable insights into the sustainability and durability of green infrastructure.
• While water bodies have been shown to provide significant cooling benefits, future research should explore in more detail how different types of water features (e.g., ponds, lakes, and rivers) interact with urban environments to mitigate UHI effects. Research could examine factors such as the size, depth, and location of water bodies and their effectiveness in cooling surrounding urban areas, particularly in the context of water scarcity and sustainability.
• Lastly, as cities continue to grow under the pressures of climate change, there is a need for climate-sensitive urban planning models that can predict UHI intensity and guide mitigation efforts in future developments. Research should focus on creating urban planning frameworks that incorporate climate projections, spatial modeling, and thermal dynamics to inform urban expansion.