Evaluating the Impact of Green Spaces on Urban Heat Reduction in Rajshahi, Bangladesh Using the InVEST Model

Abstract

Urban heat poses significant challenges in rapidly developing cities, particularly in countries like Bangladesh. This study investigates the cooling effects of urban green spaces in Rajshahi city, addressing a critical research gap in developing urban contexts. We examined the relationships among urban vegetation, heat mitigation, and temperature variables using the InVEST Urban Cooling Model and spatial analysis techniques. This study focused on three key relationships: Normalized Difference Vegetation Index (NDVI) and Heat Mitigation Index (HMI), HMI and Land Sur face Temperature (LST), and HMI and Air Temperature (AT). Analysis revealed a strong positive correlation between NDVI and HMI, indicating the effectiveness of vegetation in enhancing urban cooling. A robust inverse relationship between HMI and LST was observed (R² = 0.78, r = −0.88), with every 0.1 unit increase in HMI corresponding to a 0.53 °C decrease in LST. The HMI−AT relationship showed an even stronger correlation (R² = 0.84, r = −0.87), with each unit increase in HMI associated with a 2.80 °C decrease in air temperature. These findings quantify the significant role of urban green spaces in mitigating heat and provide valuable insights for urban planning in developing cities, underscoring the importance of integrating green infrastructure into urban-development strategies to combat urban heat and improve livability.

1. Introduction

Urbanization has led to significant environmental changes worldwide, particularly in the transformation of natural landscapes into built environments. One of the most pressing issues associated with urbanization is the urban heat island (UHI) effect, in which urban areas experience higher temperatures than their rural counterparts. This phenomenon is mainly caused by the substitution of vegetation for impervious surfaces such as roads and buildings, which absorb and retain heat [1,2,3]. The urban heat effect not only increases energy consumption and greenhouse gas emissions but also poses serious health risks, particularly during heatwaves [4]. Given the growing concerns over climate change and urban sustainability, the need for effective heat-mitigation strategies has become more urgent than ever [5]. Urban green spaces (UGSs), such as parks, gardens, green roofs, and vertical greening (green walls and green façades), are increasingly recognized as crucial elements in plans for mitigating urban heat. These green spaces can lower urban temperatures through processes like shading, evapotranspiration, and the cooling effect of vegetation [6]. Vertical greening can contribute significantly to urban heat mitigation by providing additional surfaces for vegetation, improving air quality, and enhancing the aesthetic appeal of urban environments. This approach is suitable for compact cities where traditional UGSs are difficult to implement due to the lack of space. Consequently, the integration of green spaces into urban planning is vital for improving urban living conditions, enhancing public health, and contributing to climate resilience.

Understanding the role of UGSs in mitigating heat is crucial for urban planning and sustainability. As cities continue to grow, incorporating green spaces effectively can improve urban living conditions, enhance public health, and contribute to climate resilience [7]. In Bangladesh, rapid urbanization and population growth have intensified urban heat, particularly in cities like Rajshahi, which is known for its extreme temperatures. As urban areas expand, it is becoming crucial to understand and quantify the heat-mitigation services provided by UGSs to inform urban planning and policy decisions. This study focuses on evaluating the heat-reduction benefits of UGSs in Rajshahi City using the InVEST Urban Cooling Model (UCM), a tool designed to assess ecosystem services.

Numerous studies have demonstrated the significant role of UGSs in mitigating urban heat. Bowler et al. [8] conducted a meta-analysis of the cooling effects of UGSs and found that the typical temperature in parks during daylight hours was about 0.94 °C lower than that in surrounding areas. Similarly, Zhang et al. [9] examined the cooling effects of green roofs and walls, highlighting their substantial contribution to urban temperature reduction. Norton et al. [10] examined the role of urban green infra structure in mitigating the UHI effect in Melbourne, Australia, and found that green roofs and walls significantly reduce urban temperatures. Hu et al. [11] applied the InVEST UCM along with scenario analysis to examine urban heat-reduction patterns. They accomplished this by developing different scenarios for the arrangement of blue and green spaces in Wuhan, China. Their findings indicated that the InVEST UCM successfully identified some variations in Wuhan’s surface-temperature responses. Sodoudi et al. [12] tested the HMI they developed for three urban areas with distinct structural layouts in Tehran, Iran. They achieved this by comparing linear regression analysis with land surface temperature (LST) data. Their results showed that the HMI could account for varying amounts of LST variation. The coefficient of determination (R²) ranged from 0.48 to 0.64, differing by town. Notably, stronger correlations were found in towns with a wider LST value range. Gunawardena et al. [13] analysed how green and blue spaces cool cities at different atmospheric levels. They found that evapotranspiration mainly cools the urban canopy layer, with tree-dominated areas providing the most heat relief when it was needed most. Additionally, a study by Demuzere et al. [14] utilized advanced modelling techniques to assess the cooling impacts of green spaces in different climatic contexts, emphasizing the versatility and effectiveness of these methods. These studies underscore the potential of UGSs to alleviate urban heat.

Several methods have been employed to assess the cooling benefits of UGSs. Traditional approaches often involve in situ temperature measurements and remote sensing techniques. For instance, Potchter et al. [15] utilized fixed meteorological stations to compare temperatures in different urban environments, while Chen et al. [2] employed remote sensing data to analyse the spatial distribution of LSTs and their relationship with vegetation cover. These methods provide valuable insights but can be limited by their spatial and temporal coverage. In recent years, advanced modelling tools have been developed to provide more comprehensive assessments of urban cooling services. The InVEST UCM is one such tool; it integrates various factors, including vegetation cover, albedo, and evapotranspiration, to estimate the cooling benefits of UGSs. The model has been applied in various contexts to quantify the cooling effects of green infrastructure. For instance, Zawadzka et al. [16] applied the InVEST UCM to develop an HMI. This index was designed to quantify the cooling potential of vegetation across different spatial areas. By applying this model, the researchers aimed to create a tool that could estimate how effectively various types and distributions of urban greenery could reduce heat in specific locations within a city. This approach allows for a more nuanced understanding of how vegetation contributes to cooling in urban environments, considering the spatial context of greenspaces. Hu et al. [11] applied the InVEST UCM to examine urban heat-reduction patterns. They created multiple scenarios featuring different arrangements of blue (water) and green spaces in urban environments. This approach allowed them to simulate and analyse how various configurations of these natural elements could potentially mitigate urban heat issues.

Despite extensive research on UGSs and their cooling effects, empirical studies focusing on developing countries like Bangladesh remain limited. While numerous studies have explored the broader environmental benefits of UGSs, few have specifically examined their role in mitigating urban heat, particularly in cities other than Dhaka [17,18,19]. Previous studies have primarily considered the NDVI and LST but have not included the HMI and AT in their analyses. This study addresses this gap by incorporating HMI and AT alongside NDVI and LST, using the InVEST UCM and comprehensive spatial analysis in Rajshahi. Our research provides novel insights into how different types of green space can mitigate urban heat in Rajshahi by adding HMI and AT dimensions, offering valuable data to inform urban planning and policymaking in similar developing cities.

Furthermore, the application of advanced modelling tools like the InVEST UCM to quantify the cooling benefits of UGSs is limited in Bangladesh. While studies in developed countries have increasingly utilized such models to provide comprehensive assessments of urban cooling services, similar applications in Bangladesh are rare. This lack of advanced modelling applications hampers the ability of urban planners and policymakers to make informed decisions regarding the integration of green spaces into urban environments for heat mitigation.

Moreover, there is a need for studies that investigate the relationship between urban green space (measured as NDVI) and HMI, as well as the relationship between HMI and both LST and AT. LST and AT are crucial yet distinct parameters in environmental and climatic studies. LST measures the radiative temperature of the Earth’s surface, including soil, vegetation, and built structures, using thermal infrared sensors on satellites, providing data at high spatial resolution but with limited temporal frequency. In contrast, AT is the measure of the temperature of the air at a specific height above the ground, typically between 1.25 m and 2 m, as standardized by meteorological practices. It is a crucial parameter for weather forecasting, climate monitoring, and various environmental studies. AT is measured using thermometers or electronic sensors housed in meteorological stations, which provide continuous data with high temporal resolution, though spatial coverage is limited to the locations of these stations. Accurate AT measurements are vital for understanding atmospheric processes, predicting weather patterns, and assessing climate variability and change. While these relationships have been explored in other contexts, they remain understudied in Bangladesh [20,21,22]. Understanding these relationships is crucial for developing effective strategies to combat urban heat and improve urban livability in Bangladeshi cities.

Addressing these research gaps is crucial for informing urban planning and policy decisions aimed at enhancing urban sustainability and resilience in Bangladesh. This study aims to fill these gaps by providing a comprehensive analysis of the heat-mitigation services of UGSs in Rajshahi City, utilizing the InVEST UCM. Specifically, this study attempted to (a) investigate the relationship between urban green space, measured as NDVI, and HMI (b) examine how HMI varies with LST and (c) explore the relationship between HMI and AT.

This study focuses on the broader concept of mitigating urban heat rather than solely on the UHI effect. While the UHI effect specifically refers to the temperature difference between urban areas and their rural surroundings due to human activities and alterations to the natural environment, our research aims to understand how UGSs can reduce overall urban temperatures. By analysing the relationships among NDVI, HMI, LST, and AT, we provide insights into the general benefits of vegetation in cooling urban areas. This comprehensive approach helps in developing strategies for urban planning and climate adaptation that are applicable to various urban settings, irrespective of their rural counterparts.

While this study focuses on outdoor urban heat, it is important to acknowledge that people spend a significant portion of their time indoors, in environments such as schools, workplaces, and homes. Indoor air temperature is influenced by factors like thermal insulation, cooling devices, heating sources, and indoor vegetation, as well as vertical and roof vegetation. Although outdoor temperatures can affect indoor conditions and thermal comfort, they do not always directly reflect indoor temperatures due to these moderating factors. Therefore, comprehensive urban heat-mitigation strategies should consider both outdoor and indoor environments to effectively enhance overall thermal comfort and public health. Future research should explore the specific effects of outdoor heat-mitigation measures on indoor thermal environments to provide a more holistic understanding of urban heat dynamics and their implications for public health and comfort.

The paper is structured as follows: Section 2 details the methodology, including data collection, processing, and the application of the InVEST UCM. Section 3 presents the results of the study followed by a discussion in Section 4 that interprets the findings in the context of existing literature and the specific conditions of Rajshahi. Finally, Section 5 concludes the paper with key insights and recommendations for policy and future research.

2. Materials and Methods

2.1. Study Area

The study focused on the Rajshahi Metropolitan Area (RMA) in Bangladesh, one of the country’s eight metropolitan cities and administrative divisions. Situated in north-western Bangladesh, Rajshahi is bordered by the Padma River to the south and the Jamuna River to the east. The city lies within the Barind Tract, 23 m above sea level, at coordinates 24°22′26′′ N 88°36′04′′ E.

Rajshahi spans 365.55 km2 (Figure 1) and houses 1.3 million residents. Located 243 km from Dhaka, the capital, it is near the India-Bangladesh border. The city is a key centre for administration, education, culture, and commerce. Its high density of educational institutions has earned it the nickname “Bangladesh’s educational city”.

Rajshahi’s climate is categorized as tropical wet and dry, a climate classified as Aw according to the Köppen classification system. The weather is typically characterized by monsoons, high temperatures, significant humidity, and moderate rainfall.

The warm season begins in early March and lasts until mid-July. During the peak months of April through July, average high temperatures range from 32 to 36 °C. January sees the coldest temperatures, with lows between 7 and 16 °C.

Rainfall is the heaviest during the monsoon period. On average, the district receives about 1448 millimetres (57.0 inches) of rain annually [23].

2.2. Dataset for InVEST UCM

A variety of datasets were required to run InVEST UCM in this study. The primary data required in InVEST UCM included land cover, LST, evapotranspiration, NDVI and biophysical table. Descriptions of the datasets follow. ArcGIS 10.8 software was used to produce all maps shown in this manuscript.

2.2.1. Land Cover

Landsat 8 imagery was utilized for land-cover classification in the study area. The specific image dated 25 April 2023 was selected for this purpose. The Collection 2 Level 2 dataset was obtained and processed. The image corresponded to path 138 and row 43 of the Landsat orbital system. The Landsat image dated 25 April 2023 was specifically selected for this study due to several key factors. April falls within the pre-monsoon season in Rajshahi, offering optimal conditions for clear satellite imagery, with rising temperatures and relatively low cloud cover. April typically is associated with lower chances of cloud cover compared to the monsoon months, ensuring better image quality. Additionally, the 2023 image offers up-to-date information, reflecting recent land-use and land-cover changes in the study area.

To derive the land-cover classes, an unsupervised-classification approach was employed. The imagery was processed and analysed using appropriate remote sensing software. Multiple spectral bands from the Landsat 8 sensor were incorporated into the classification algorithm. The unsupervised classification method was applied to group pixels with similar spectral characteristics into distinct clusters.

After the initial clustering, the resulting classes were interpreted and labelled based on known land-cover types in the study area. The classification results were then refined and validated to ensure accuracy. Post-classification processing techniques may have been applied to improve the final land-cover map. The classification accuracy of the un-supervised classification of land-cover classes is 87.86%.

The resulting land-cover classification was used as a basis for further analysis in the study, providing crucial information about the distribution of different land-cover types within the Rajshahi Metropolitan Area. The derived land-cover map of Rajshahi is presented in Figure 2.

2.2.2. Land Surface Temperature

LST for the study area was derived using Landsat 8 imagery. The thermal infrared sensor (TIRS) Band 10 from the surface reflectance product, dated 25 April 2023, was utilised for this purpose.

The methodology for LST calculation followed established procedures from the literature, specifically drawing upon the methods developed by Avdan et al. [24] and Zhao et al. [25]. These methods are thoroughly documented and widely recognized within the remote sensing community for extracting LST from Landsat thermal data.

As the temperature-retrieval process is a well-established procedure in remote sensing and the methods used are thoroughly documented in the cited literature, detailed step-by-step calculations are not given in this study. The resulting LST map provided crucial information about the spatial distribution of surface temperatures across the study area, which was then used for further analysis in the research.

2.2.3. Evapotranspiration

Evapotranspiration (ET) quantifies the water-vapour transfer from terrestrial surfaces to the atmosphere over a specific time frame. This measurement combines two processes: direct evaporation from surfaces like soil and water bodies and transpiration through vegetation [26]. ET₀ is typically expressed as a water depth, usually in millimetres, per unit of time.

The MOD16 Global Evapotranspiration (ET) product, derived from MODIS (Moderate Resolution Imaging Spectroradiometer) data, was utilized in this research. This dataset was obtained from NASA’s Application for Extracting and Exploring Analysis Ready Samples (AppEEARS) platform (https://appeears.earthdatacloud.nasa.gov/) (accessed on 12 May 2024).

MOD16 is a widely used global ET product that provides estimates of evapotranspiration at regular intervals. The specific dataset used in this study was the 8-day composite product, which has a spatial resolution of 500 m. So, the ET image was resampled at 30 m spatial resolution to match with the land-cover map of 30 m spatial resolution. The ET was also divided by 8 to obtain daily measurements. The ET map of the study area is presented in Figure 3.

2.2.4. Normalized Difference Vegetation Index

The NDVI for the study area was calculated using Landsat 8 imagery from April 24, 2023. Band 4 (Red) and Band 5 (Near-Infrared) were extracted and pre-processed to surface-reflectance values. The NDVI was then computed using the standard formula, as follows: NDVI = (NIR-Red)/(NIR + Red) [27]. This calculation was conducted on a pixel-by-pixel basis across the entire study area, yielding values between −1 and +1, with more positive values representing denser vegetation. The resulting NDVI map provided a spatial representation of vegetation health and density, which was subsequently used for further analysis in the research.

2.2.5. Biophysical Table

The biophysical table presents parameters for each land-cover type, including shade, crop coefficient (Kc), and albedo. These values typically range from 0 to 1, though Kc can extend up to 1.5 in practice. Shade is categorized as either 1 (above 2 m) or 0 (below 2 m) [28]. However, the shade value was modified based on experience and local conditions for this study. Kc and albedo values were determined based on recommendations from Zawadzka et al. [16] and the InVEST model guidelines. Green spaces, including water bodies, were assigned a value of 1. For this study, we used the values described in Hu et al. [11].

Table 1. Key parameters of the biophysical table assigned to each land-cover type.

Land Cover	Shade 1	Kc	Albedo	Green Space
Built up land	0.05	0.33	0.21	0
Agricultural Land	0.2	0.72	0.16	1
Vegetation	0.5	0.97	0.19	1
Bare Land	0	0.61	0.23	0
Water	0.5	1.00	0.06	1

¹ Shade value has been modified based on experience and local condition.

provides a detailed breakdown of these parameter values for each land-cover type.

2.2.6. Additional Parameters

In addition to the core data described above, InVEST UCM requires some additional parameters. These parameters include reference air temperature (T_ref), magnitude of the UHI effect (UHI_max), air blending distance (d_air) and maximum cooling distance (d_cool) [29]. These values can vary significantly based on the specific urban context, climate zone, and scale of the study area. It’s important to calibrate these parameters using local data, when possible, for more accurate results.

Reference air temperature is the baseline temperature and is typically measured in rural or less-urbanized areas. It represents the temperature without UHI effects. The reference value for reference air temperature often ranges from 20 °C to 35 °C, depending on the local climate and time of year [11]. In this study, 32 °C was considered as the reference air temperature.

The magnitude of the UHI effect represents the maximum temperature difference between urban and rural areas due to the UHI effect. Reference values typically range from 2 °C to 6 °C but can be higher in large cities or extreme conditions [16]. In this study, 3.5 °C was considered as the magnitude of the UHI effect based on the data at https://yceo.users.earthengine.app/view/uhimap (accessed on 13 June 2024).

Air blending distance is the distance over which the cooling effect of green spaces blends with the surrounding air temperature. Reference values are often set between 500 to 600 m, depending on local conditions and the study scale [30]. An air blending distance of 550 was used here.

Maximum Cooling Distance is the maximum distance at which the cooling effect of a green space larger than 2 hectares can be detected. The recommended value of 450 m has been considered.

2.3. Working Principles of InVEST UCM

2.3.1. Cooling Capacity Index

The InVEST UCM model first calculates the cooling capacity (CC) index for each pixel by considering local shade, evapotranspiration, and albedo. This method is based on indices from Zardo et al. [31] and Kunapo et al. [32], with the addition of albedo to enhance heat reduction. The shade factor represents the proportion of tree canopy (≥2 m) per land use/land-cover (LULC) category, ranging from 0 to 1. The evapotranspiration index (ETI) is a normalized value of potential evapotranspiration, computed by multiplying the reference evapotranspiration (ETo) and the crop coefficient (Kc) for each LULC type, then dividing by the maximum ETo value in the area (ETmax) [33]:

ETI = (Kc · ETo)/ETmax

(1)

This equation assumes adequate irrigation for vegetated areas, though Kc values can be adjusted for water-limited conditions. The albedo factor, ranging from 0 to 1, represents the proportion of solar radiation reflected by the LULC type. The CC index is computed as follows [34]:

CCi = 0.6 · shade + 0.2 · albedo + 0.2 · ETI

(2)

The recommended weights (0.6; 0.2; 0.2) reflect the higher impact of shading. In smaller areas, evapotranspiration is weighted less compared to shading. The model was adjusted to give equal weight to albedo and ETI, based on findings by Phelan et al. [34]. Option ally, the model can include a building-intensity factor, which influences nighttime temperatures due to heat released by buildings. This factor needs to be provided in the Biophysical Table for each land-use class, modifying the following equation accordingly [34]:

CCCi = 1 − building. intensity

(3)

2.3.2. Urban Heat-Mitigation Index

To consider the cooling effect of large green spaces (>2 ha) on surrounding areas (as discussed in Zardo et al. [31], the model computes the urban HM index. If a pixel is not influenced by any large green spaces, HM equals CC; otherwise, it is a distance-weighted average of the CC values from the large green spaces and the pixel of interest.

To do this, the model first calculates the area of green spaces within a search distance d_cool around each pixel (GA_i) and the CC provided by each park (CC_parki), as follows [31]:

$$G{A}_i=cel{l}_{area}{\displaystyle \sum }_{\in d radius from i}gj$$

(4)

$$C{C}_{par{k}_i}={\displaystyle \sum }_{\in d radius from i}gj .C{C}_j .e^{\left({\displaystyle \frac{d\left(i,j\right)}{d_{cool}}}\right)}$$

(5)

where cell_area is the area of a cell in hectares, g_j is 1 if the pixel is green space or 0 if not, d(i, j) is the distance between pixels i and i, d_cool is the distance over which a green space has a cooling effect, and CC_parki is the distance-weighted average of the CC values from green spaces. Next, the HM index is calculated as follows [31]:

$$H{M}_i=\left\{\begin{array}{c}C{C}_i if C{C}_i\ge C{C}_{par{k}_i} or G{A}_i<2ha\\ C{C}_{par{k}_i} otherwise\end{array}\right\}$$

2.3.3. Estimation of Air Temperature

To estimate heat reduction across the city, the model uses the city-scale UHI magnitude. Users can obtain UHI values from local studies or global resources, such as the Global Surface UHI Explorer by Yale University, which provides estimates of annual, seasonal, daytime, and nighttime UHI (https://yceo.users.earthengine.app/view/uhimap) (accessed on 18 May 2024). It is important to note that UHI magnitude is specific to a particular period (e.g., current or future climate) and time (e.g., nighttime or daytime temperatures) and that this selection will influence the service quantification and valuation.

Air temperature without air mixing (T_{no_mix}T _{no_mix}) is calculated for each pixel as follows [32]:

T_{air nomix,i} = T_{air, ref} + (1 − HMi) ⋅ UHI_max

where T_{air, ref} is the rural reference temperature and UHI_max is the maximum magnitude of the UHI effect for the city, or more precisely, the difference between T_{air, ref} and the highest temperature observed in the city.

Due to air mixing, these temperatures average spatially. The actual air temperature (with mixing), actual T_{air, ref}, is derived from T_{air nomix,i} using a Gaussian function with a kernel radius, r, defined by the user.

For each area of interest (provided by the user as a vector GIS layer), we calculated the average temperature and temperature anomaly (T_air,i − T_{air, ref}).

3. Results

This section presents the key findings of our analysis, focusing on the relationships between the NDVI, HMI, LST, and air temperature. We examine the correlations between NDVI and HMI, HMI and LST, and HMI and air temperature to provide insights into the interconnections between vegetation, heat, moisture, and temperature in the study area.

3.1. Normalized Difference Vegetation Index vs. Heat Mitigation Index

One of the critical analyses performed was to understand the relationship between the NDVI and the HMI. NDVI is a widely used indicator of green space, while HMI quantifies the cooling benefits provided by green spaces. A linear regression analysis was con ducted between NDVI and HMI. HMI was calculated using InVEST UCM as described in Section 2.3.2. The result of regression analysis has been presented in Figure 4 while the NDVI and HMI map have been presented in Figure 5.

The regression analysis revealed a significant relationship between NDVI and HMI, with an R² value of 0.76 and a correlation coefficient of 0.88. The slope of the regression line indicates that for a 0.1 unit increase in NDVI, the HMI increases by approximately 0.066 units. This positive slope underscores the significant contribution of vegetation density and health to heat mitigation in urban areas. The higher the NDVI, the greater the amount of green space, which enhances the cooling effect. The intercept represents the expected value of HMI when NDVI is zero. Although an NDVI of zero is theoretical and implies a lack of vegetation, the intercept value of 0.46 provides a baseline HMI level in the absence of vegetation. This baseline might be due to other nonvegetative cooling factors present in the urban environment. The coefficient of determination, R², is 0.76, indicating that approximately 77.54% of the variability in HMI can be explained by the variability in NDVI. This high R² value suggests a strong fit of the regression model to the observed data. It signifies that NDVI is a robust predictor of HMI, highlighting the importance of vegetation in urban heat mitigation.

3.2. Heat-Mitigation Index vs. Land Surface Temperature (LST)

Examining the connection between LST and the HMI was the second main objective of the study. The HMI measures the amount of cooling that UGSs provide, and the LST shows the earth’s surface temperature in urban settings. This association was investigated in depth using a linear regression analysis. The Figure 6 and the LST map shown in Figure 7, present the findings of the regression analysis. Regression analysis, with an R² value of 0.78 and a correlation coefficient of −0.88, provides important insights into this relationship between HMI and LST.

The regression line’s negative slope shows that the LST drops by about 0.53 °C for every 0.1 unit rise in HMI. The cooling benefit offered by higher HMI values—which are indicative of improved heat mitigation through UGSs—is shown by this considerable inverse relationship. When the HMI is zero, the intercept shows the expected value of LST. This figure, indicating the potential intensity of urban heat in the absence of effective green space interventions, is 35.99 °C. It serves as a baseline temperature in the absence of heat-mitigation measures. With a coefficient of determination R² = 0.78, it can be inferred that the variability in HMI accounts for roughly 78.22% of the variability in LST. This high number indicates that the regression model and the observed data have a good fit. It shows that HMI is a reliable indicator of LST and emphasizes the significance of heat-mitigation strategies.

3.3. HMI vs. AT

Understanding the connection between air temperature and the HMI was the third goal of the research. While air temperature is an important indicator of urban heat levels, the HMI measures the cooling impacts of UGSs. The association was examined using a linear regression analysis to gain a better understanding. Figure 8 displays the outcome of the regression study, and Figure 9 displays the air-temperature map. HMI and air temperature have a strong association, as indicated by the regression analysis, which produced a correlation coefficient of −0.87 and an R² value of 0.84.

The regression line’s negative slope shows that there is a about 2.80 degree Celsius drop in air temperature for every unit increase in HMI. Higher HMI values, which indicate superior heat mitigation through UGSs, have a cooling effect, as this substantial inverse relationship highlights. When the HMI is zero, the intercept shows the predicted air temperature. Without heat-reduction strategies, this figure of 35.05 °C serves as a baseline, illustrating the potential intensity of urban heat in the absence of effective green-space interventions. The coefficient of determination (R²) is 0.84, meaning that the variation in HMI can account for roughly 84.03% of the variation in air temperature. An excep tional match between the regression model and the observed data is indicated by this ex ceptionally high R² score. It shows that HMI is a reliable indicator of air temperature, emphasizing the role that heat-mitigation strategies have in regulating air temperatures in cities.

4. Discussion

4.1. Influence of NDVI on HMI

The findings of the study revealed a strong positive relationship between the NDVI and the HMI. This finding provides valuable insights into the role of urban green space in mitigating heat and improving thermal comfort in urban environments.

The linear regression analysis of the relationship between NDVI and HMI yielded a high coefficient of determination (R² = 0.76) and a strong positive correlation (r = 0.88). These strong statistical measures suggest that NDVI is a dependable indicator of heat-mitigation potential in urban areas. The findings align with and build upon previous research in the field. A study by Kong et al. [35] in Nanjing, China, found a strong negative correlation between NDVI and LST, with R² values ranging from 0.727 to 0.934. This aligns with our findings, as HMI is inversely related to LST. A study by Solecki et al. [36] in Newark, New Jersey, found a strong positive correlation between NDVI and surface temperature reduction, supporting our findings on the relationship between vegetation indices and heat mitigation. The consistency between these previous studies and our findings in Rajshahi City, Bangladesh, strengthens the generalizability of the relationship between vegetation indices and urban-cooling potential across different geographical and climatic contexts. It also validates the use of NDVI as a reliable indicator for estimating the heat-mitigation capabilities of UGSs.

According to Figure 5, the northern region shows high HMI values, corresponding with the high NDVI values. This indicates that areas with dense vegetation also have strong heat-mitigation capabilities. The high NDVI values reflect a significant presence of healthy vegetation, which contributes to cooling through processes such as evapotranspiration and shading. Evapotranspiration involves the transfer of water from soil and vegetation into the atmosphere, which absorbs heat and cools the surrounding air [8]. Shading from trees and other vegetation reduces the amount of solar radiation reaching the ground, thereby lowering surface temperatures [37].

In contrast, the southern part exhibits low HMI values, aligning with the low NDVI values. The lack of vegetation here results in poor heat mitigation, making this region more susceptible to higher temperatures. The absence of trees and other vegetation means there is minimal evapotranspiration and shading, leading to increased surface and air temperatures. This phenomenon is often observed in UHI, where built-up areas with limited vegetation experience higher temperatures compared to their rural counterparts [38].

In the central part of the study area, both NDVI and HMI values are moderate, indicating a mix of vegetation density and heat-mitigation capabilities. This region benefits from its vegetation cover but also faces challenges due to urbanization. Urban areas often have less vegetation and more impervious surfaces, which absorb and retain heat, contributing to higher temperatures. However, the presence of moderate vegetation in this region helps to mitigate some of the heat through the mechanisms mentioned earlier.

Strategic planning to increase the number of green spaces and vegetation can improve heat mitigation and contribute to a healthier urban environment. Implementing urban-greening initiatives, such as planting trees, creating parks, and creating green roofs, can significantly reduce urban temperatures and improve overall thermal comfort [8,39]. By understanding the relationship between NDVI, HMI, and urban heat, city planners and policymakers can develop targeted strategies to enhance urban resilience to heat stress.

4.2. Impact of Heat-Mitigation Index on Land Surface Temperature

The study revealed a strong inverse relationship between the HMI and LST in urban environments. This relationship was quantified through linear regression analysis, which yielded significant results that highlight the cooling effect of UGSs. The strong negative correlation (r = −0.88) between HMI and LST indicates that as the heat-mitigation potential of an area increases, the surface temperature decreases substantially. This relationship is further supported by the high coefficient of determination (R² = 0.78), suggesting that 78.22% of the variability in LST can be explained by changes in HMI. These robust statistical measures emphasize the reliability of HMI in predicting urban surface temperatures. The regression model showed that for every 0.1 unit increase in HMI, there was a corresponding decrease of 0.53 °C in LST. This quantifiable cooling effect can be attributed to several factors, including evapotranspiration [40], shading [35], albedo effect [41], and air circulation [36]. These findings align with and build upon the results of previous studies in the field.

A study by Kong et al. [35] in Nanjing, China, found that UGSs could lower LST by 0.4 °C to 2.2 °C, depending on the type and configuration of vegetation. This aligns with our finding of a 0.53 °C decrease per 0.1 HMI unit increase. Research by Feyisa et al. [42] in Addis Ababa, Ethiopia, reported that the cooling effect of parks extended up to 240 m beyond their boundaries, with an average temperature reduction of 6.72 °C. This supports the broader implications of our HMI−LST relationship for urban planning.

4.3. Impact of Heat-Mitigation Index on Air Temperature

The strong inverse relationship between AT and the HMI revealed in our study underscores the significant role of UGSs in mitigating urban heat in Rajshahi City. With a correlation coefficient of −0.87 and an R² value of 0.84, our findings demonstrate that HMI is a robust predictor of air-temperature variations in urban environments. The observed 2.80 °C decrease in air temperature for every unit increase in HMI highlights the substantial cooling effect of UGSs. This cooling effect can be attributed to several mechanisms:

• Evapotranspiration: Plants release water vapour through their leaves, absorbing heat from the surrounding air [43].
• Shading: Tree canopies intercept solar radiation, reducing surface and air temperatures [8].
• Albedo modification: Green spaces generally have higher albedo than built surfaces, reflecting more solar radiation [37].
• Air circulation: Vegetation can influence local air movement, potentially enhancing cooling effects (Guo et al., 2014) [44].

These findings align with and expand upon those of several previous studies. For instance, Oliveira et al. [45] observed temperature reductions of up to 6.9 °C in small green spaces in Lisbon, Portugal.

Our results are particularly consistent with those of Zölch et al. [46], who used the InVEST UCM in Munich, Germany. They found that increasing the vegetation cover by 20% could reduce surface temperatures by up to 1.7 °C, demonstrating the model’s efficacy in quantifying urban cooling effects.

The baseline AT of 35.05 °C, predicted when HMI is zero, serves as a critical reference point. This value illustrates the potential severity of urban heat in the absence of effective greenspace interventions, emphasizing the importance of incorporating heat-mitigation strategies in urban-development plans. This aligns with findings from Heaviside et al. [47], who projected that UHI effects could increase heat-related mortality by up to 50% in some cities by the 2050s.

The exceptionally high R² value (0.84) in our study suggests that the InVEST UCM, coupled with the HMI, may be particularly effective in quantifying the thermal benefits of green spaces in tropical urban contexts like Rajshahi City. This builds upon the work of Deilami et al. [1], who successfully used the model to assess urban cooling services in Brisbane, Australia, demonstrating its applicability across diverse climatic zones.

These findings have important implications for urban heat-adaptation strategies in Bangladesh and similar climatic regions. The strong predictive power of HMI−AT variations suggests that it could be a valuable tool for urban heat mapping and for assessing the potential impact of green infrastructure projects. This is particularly relevant in the context of rapid urbanization in developing countries, as highlighted by Ramaiah et al. [41] in their study of UGSs in Chennai, India.

4.4. Similarities in the Heat-Mitigation Index’s Influences on Land Surface Temperature and Air Temperature

We examined the similarities in the influences of the HMI on LST and AT. We observed a robust inverse relationship between HMI and LST, with an R² value of 0.78 and a correlation coefficient of −0.88. Specifically, each 0.1 unit increase in HMI was associated with a 0.53 °C decrease in LST. Similarly, the HMI−AT relationship demonstrated a strong correlation (R² = 0.84, r = −0.87), where each unit increase in HMI corresponded to a 2.80 °C decrease in AT.

These results indicate a notable commonality in the effects of HMI on LST and AT, showing that HMI consistently mitigates heat across both temperature metrics. The stronger correlation with AT compared to LST suggests that while the HMI’s influence on the two metrics is broadly similar, its impact on air temperature is more pronounced.

4.5. Implications of the Study

The results of this study have significant implications for urban planning in Bangladesh and other developing countries. The findings underscore the importance of integrating green infrastructure into urban-development plans to mitigate urban heat and enhance thermal comfort. Strategic placement and maintenance of UGSs, such as parks, green roofs, and street trees, can effectively reduce surface and air temperatures in densely populated urban areas. This is particularly crucial in rapidly urbanizing cities like Rajshahi, where the expansion of built-up areas often comes at the expense of natural vegetation.

Urban planners and policymakers in Bangladesh should prioritize the creation and preservation of green spaces in urban areas. Incorporating green infrastructure into city planning not only helps in mitigating urban heat but also provides multiple co-benefits, including improved air quality, enhanced biodiversity, and increased recreational opportunities for residents [48]. Additionally, green spaces can play a critical role in climate-adaptation strategies by reducing the vulnerability of urban populations to heat waves, which are expected to become more frequent and intense due to climate change [49].

For other developing countries with similar urban and climatic conditions, the study’s findings highlight the need for context-specific approaches to urban heat mitigation. In cities where informal settlements and rapid urban growth are prevalent, integrating green infrastructure can be challenging but essential. Policymakers must consider local socioeconomic factors, land-use patterns, and governance structures when designing and implementing green infrastructure projects [50].

Moreover, effective urban heat mitigation requires a participatory approach, involving community engagement and collaboration with various stakeholders. This ensures that green infrastructure projects are sustainable, socially inclusive, and responsive to the needs of local populations [51]. Future research should explore the socioeconomic and health impacts of UGSs, as well as the cost-effectiveness of different green infrastructure interventions in developing country contexts.

4.6. Limitation of the Study

While the InVEST UCM provides valuable insights into the cooling effects of UGSs, it is important to acknowledge its limitations and potential sources of uncertainty. The model relies on several assumptions, such as the homogeneity of vegetation types and the uniformity of urban structures, which may not fully capture the complexities of urban environments. Additionally, the model’s resolution and the quality of input data, primarily derived from remote sensing, can introduce errors.

Another limitation is the lack of in situ temperature measurements within the study area, which hinders the validation of model outputs. Ground-truthing data from localized air temperature sensors would enhance the accuracy and reliability of the findings. Future studies should incorporate such ground-truthing efforts to validate and refine the model predictions.

Furthermore, while our study demonstrates the cooling potential of green spaces, it does not address the practical challenges of implementing and maintaining urban green infrastructure in a rapidly developing city. Future studies could investigate the cost-effectiveness of various green-space interventions and their long-term sustainability in the local context, building on work such as that by Guo et al. [44] on the economic valuation of urban green infrastructure.

5. Conclusions

This study highlights the crucial role of UGSs in mitigating urban heat, with a focus on Rajshahi, Bangladesh. Using the InVEST UCM and spatial analysis, we found strong correlations between the NDVI, HMI, LST, and air temperature. Our results show a significant inverse relationship between HMI and both AT and LST and a positive association between NDVI and HMI, indicating that areas with dense vegetation have better heat-mitigation capabilities. These findings align with those of previous studies from various regions and underscore the cooling benefits of vegetation.

In addition to mitigating urban heat, UGSs provide a range of other environmental benefits that are critical for improving urban living conditions, especially in developing cities. Vegetation helps to remove air pollutants and improve air quality, which is vital in areas with high levels of pollution [52,53]. Moreover, UGSs can also reduce noise pollution, creating more pleasant and healthier urban environments. These complementary benefits underscore the multifaceted value of UGSs in addressing various environmental challenges in urban areas.

The study’s results suggest that the cooling effects of UGSs are substantial and predictable, which is vital for urban planning and climate adaptation in rapidly urbanizing and warming regions. Despite the limitation of lacking in situ air-temperature data for model validation, the findings support the inclusion of green infrastructure in urban development plans to reduce the effects of heat islands and enhance urban livability. Future research should focus on collecting field data for model validation, optimizing green space configurations, and exploring socioeconomic and public-health impacts.

To sum up, this study offers strong support for the inclusion of green infrastructure in urban-development plans as a practical way to reduce the effects of urban heat islands and enhance livability. Policymakers, urban planners, and environmental managers can all benefit from the quantitative correlations among NDVI, HMI, LST, and air temperature when constructing heat-resilient cities. Subsequent investigations may go deeper into the ideal arrangement and distribution of urban green areas to optimize their thermal capabilities and extrapolate these findings to various urban scenarios. Addressing environmental-justice concerns and measuring the public-health benefits through socioeconomic studies and health impact assessments would also be valuable. Furthermore, cost−benefit evaluations and research on integrating green spaces with other urban systems could provide a more comprehensive understanding of the value of green infrastructure, contributing to the development of more resilient, equitable, and climate-adaptive urban environments.