Introduction

Urban temperatures are rising gradually all over the globe1. In recent years, Pakistan has experienced a significant rise in the population of urban areas2. The growth in urban development has led to changes in Land Use Land Cover (LULC) as building areas expand to meet the needs of the growing population3. This shift is crucial in modern times4. Every change in LULC has a global impact, affecting fragile ecological balances and climate change5. Urbanization is a key driver of LULC change, crucial in altering Land Surface Temperature (LST)6. With population growth, expanding cities, and shifting agricultural boundaries, the impact of these changes goes beyond a single region7. Rapid Urbanization disrupts the energy balance of the local environment, giving rise to the Urban Heat Islands (UHI) effect8. This effect refers to a situation where the surface temperature in the urban areas surpasses that of the surrounding areas9. This transformation in LULC leads to more impervious surfaces and increases heat storage capacity, a primary cause of UHI. UHI intensification has adverse effects, such as sudden temperature rises and changes in urban climate, leading to unpredictable rainfall, health issues, disasters like floods, and worsened air quality. Moreover, LULC changes may impact the variation in LST and the urban microclimate10.

LULC variations in regional and local environments significantly affect the area’s radiation and LST11. Studies have consistently demonstrated that LST increases correlate with changes in LULC, particularly in urban areas12. Several studies have pointed out significant health consequences of increased area radiation and LST13,14,15. These health impacts manifest in diverse forms, including thermal comfort index, discomfort index, and temperature relative index16. Globally, LST has experienced an upward trend, rising from 0.6 to 0.9 °C between 1960 and 2015. Thus, continuous monitoring of LST changes, aligned with global solar radiation, is crucial, especially in urban environments4. This monitoring is advantageous for planning effective adaptation and mitigation policies. Urban expansions have impacted the climate and LULC changes and the development of rural regions and societies17. Analyzing LULC changes is crucial for addressing environmental issues such as heat islands, flooding, biodiversity loss, arable land depletion, local warming, ecosystem degradation, and habitat loss18. Rapid changes in LULC and their impact on LST, if not addressed, pose substantial threats to the overall health of our planet. Exploring the relationship between LST and LULC aims to deepen our understanding of the environmental implications of LST and address specific challenges in local ecosystems19.

The photosynthetic process, indicated by the Normalized Difference Vegetation Index (NDVI), is linked to factors such as plant biomass, water stress, and biodiversity20. In recent years, the NDVI has become a widely utilized tool for explaining the spatiotemporal characteristics of LULC. Remote Sensing (RS) studies have consistently employed NDVI values to detect and interpret vegetation patterns21. The NDVI tool has been instrumental in deriving vegetation areas, while the Normalized Difference Built-up Index (NDBI) has been crucial for identifying the concentration and expansion of urban areas. Destruction of vegetation areas leads to a decline in NDVI, while the expansion of the built-up regions increases NDBI22. These factors, because of LULC changes, play a crucial role in the fluctuations of LST. Integrating GIS and RS is a valuable approach for urban modeling, LULC mapping, and assessing the environmental impacts of urban growth over specific time intervals23 Spatial data derived from RS and GIS provides accurate and timely information about degraded areas during particular periods and is a cost-effective solution24.

Recent LULC studies have emphasized the importance of focusing on specific regions25,26. In the study of LULC, certain places are frequently overlooked27. The geographical uniqueness of the study area, with its varied terrain and ecosystems, makes it an important place to study the localized impacts of LULC changes. Urban development in this region reflects the global trend of urban sprawl but with a unique touch specific to the study area. Examining the characteristics and dynamics of the study area not only reveals a regional case study but also offers insights into the broader implications of human-environment interactions globally. Over the past few years, Pakistan, including the Punjab region, has undergone considerable changes in LULC patterns, mainly attributed to insufficient management practices28. The study area is home to a population of over three million and is a focal point for immensely fertile agricultural land27. Unfortunately, despite its geopolitical and socio-economic significance, the study area lacks comprehensive research studies on its LULC dynamics.

This study is designed to fill the gap by systematically analyzing the changes in LULC and LST in the study area. Our primary objectives include: (i)To analyze the impact of specific LULC changes, particularly urban expansion and vegetation dynamics, on LST in the study area over the years 1991, 2001, 2011, and 2023. (ii) To examine the changes in NDVI as indicators of vegetation cover and urban growth, respectively, and their relationship with LST over time. (iii) To investigate correlations between LST and selected LULC changes, with a focus on understanding the temperature effects of urbanization and vegetation shifts, which are critical for sustainable land use planning. (iv) To provide insights into the broader implications of LULC and temperature changes for sustainable urban planning and climate adaptation strategies.

Our findings aim to inform discussions on the Sustainable Development Goals (SDGs), particularly SDG 11 (Sustainable Cities and Communities) and SDG 13 (Climate Action), which relate to responsible land use, climate action, biodiversity conservation, and the development of sustainable cities and communities.

Materials and methods

Study area

The district of Okara (Fig. 1) lies between approximately 73.273° E to 74.162° E longitude and 30.254° N to 31.139° N latitude. Okara shares its eastern boundary with Kasur, its western boundary with Sahiwal, its southern boundary with Bahawalnagar, its southwestern boundary with Pakpattan, its southeastern boundary with India, and its northern boundary with the districts of Nankana Sahib and Faisalabad. The three tehsils of Okara, Renala Khurd, and Depalpur comprise the approximately 4377 km³ research area. The research area has a combination of planned and unintentional LULC patterns. The cow and buffalo breeds found in the Okara district are also renowned29. The research area experiences a dry, scorching summer and a frigid winter. The winter season is shortened; it begins in December and finishes in February. Summertime brings regular hot, dusty storms. The region receives 200 mm of rain on average every year. With a high temperature of 44 °C in May and June and a low temperature of 2 °C in January, these months are regarded as the hottest27,30. The study area is famous for its natural environment, fertile lands, and green fields of different crops, with a population of more than 3 million. The land used for vegetation has now changed into Punjab’s urban areas, including Okara. Beyond its geographical significance, the study area emerges as a microcosm for studying the delicate balance between development and environmental sustainability. The agricultural heritage of the region, coupled with the expanding urban escape, necessitates a detailed exploration of how these changes impact local ecosystems, climate conditions, and overall ecological resilience.

Fig. 1
figure 1

Location of study district in the maps of Pakistan and Punjab Province (this figure was generated by the author using ArcGIS version 10.8 (Esri Inc., Redlands, CA, USA; https://www.esri.com/en-us/arcgis/about-arcgis/overview).

Full size image

Materials

This study used Landsat satellite imagery to investigate LULC changes in the study area. The datasets utilized encompassed images from 1991, 2001, and 2011, acquired from Landsat 5, while the 2023 imagery was sourced from the latest Landsat 9 satellite downloaded from the website https://earthexplorer.usgs.gov/ (accessed on 20 January 2024) of the United States Geological Survey (USGS). Landsat satellites are renowned for providing comprehensive and consistent Earth observation data over the past few decades. Landsat 5, a keystone in remote sensing, operated from 1984 to 2013 and featured the Thematic Mapper (TM) sensor, offering multispectral data for detailed land cover analysis. Landsat 9, the latest addition to the program, represents a significant advancement, equipped with the Operational Land Imager 2 (OLI-2) and Thermal Infrared Sensor 2 (TIRS-2). This cutting-edge sensor technology ensures improved spectral resolution and data continuity, enhancing the precision and reliability of land cover assessments. The combined use of Landsat 5 and Landsat 9 imagery enabled a comprehensive temporal analysis, capturing the evolving land dynamics in the Okara district with increased accuracy and detail. We chose the years 1991, 2001, 2011, and 2023 to examine changes in land use and land cover (LULC) and how they affect land surface temperature (LST). First, these years indicate noteworthy socioeconomic shifts and infrastructure developments, and they align with important stages of Okara District’s urban development and environmental changes. In particular, the region began to become heavily urbanized in 1991, and the years 2001 and 2011 were selected to evaluate the mid-term effects of urbanization and policy changes on LULC. To assess current trends and their implications for sustainable urban development in light of current climatic issues, 2023 was included. All maps (Figs. 1, 2, 3 and 4, 6, 7, and 8) generated by the author using ArcGIS version 10.8 (Esri Inc., Redlands, CA, USA; https://www.esri.com/en-us/arcgis/about-arcgis/overview).

Fig. 2
figure 2

Flowchart diagram of the materials and methods used in this study (this figure was prepared in Microsoft Office 365, https://www.office.com/).

Full size image
Fig. 3
figure 3

The classified maps of the study district for the period (a) 1991, (b) 2001, (c) 2011 and (d) 2023 (this figure generated by the author using ArcGIS version 10.8 (Esri Inc., Redlands, CA, USA; https://www.esri.com/en-us/arcgis/about-arcgis/overview).

Full size image
Fig. 4
figure 4

LULC change detection (1991–2023) map (this figure was generated by the author using ArcGIS version 10.8 (Esri Inc., Redlands, CA, USA; https://www.esri.com/en-us/arcgis/about-arcgis/overview).

Full size image

To ensure data quality and minimize atmospheric interference, we selected images with less than 5% cloud cover, prioritizing cloud-free dates when available each time. We excluded any scenes with significant cloud cover or missing data to maintain consistency and reliability across datasets. The geographic information system (GIS) environment was used for pre-processing, which included geometric and atmospheric corrections, to guarantee the dataset’s temporal and spatial comparability31. Data were then pre-processed using the software ERDAS Imagine 14 and ArcMap 10.8 for layer stacking and sub-setting of the image based on Area of Interest (AOI) (Fig. 2). Further details are shown in (Table 1).

Table 1 Landsat metadata for the study area.
Full size table

LULC classification and analysis

A series of samples were generated using training samples to represent the intended land use types effectively32,33These samples played a pivotal role in the image classification process. Selecting training samples entailed outlining polygons around the distinctive locations representing each LULC class34. Researchers leaned on their expertise and understanding of the geographic features of the research area to lay the groundwork for these training samples35. Utilizing color composite images (543 for Landsat 9 OLI and 432 for Landsat TM), the team prepared visual interpretation aids and delineated training areas, focusing on vegetation (1). The process of digitizing polygons encompassing each training sample for comparable land cover types was executed utilizing the selected color combination, allocating a distinct label to every class of LULC. Subsequently, spectral signatures for each LULC class were developed35, and the classification process adopted a supervised approach. The Maximum Likelihood (MLH) algorithm, implemented in ERDAS Imagine 2016, was chosen for supervised classification into different LULC categories36. Recognized for its accuracy and precision in generating precise training data, the MLH algorithm is extensively utilized. It considers the average vector and the diversity of pixels within a multispectral feature space37. To further refine classification results, applying the NDVI and NDBI was crucial, particularly for enhancing the classification of Landsat MSS images at a 30 m resolution for 1991, 2001, and 2011, alongside the Landsat OLI image 2023. The study centers on four identified land cover types: Vegetation, Urban, Water, and Bare Land. Additional details for the LULC classes are presented in (Table 2).

Table 2 Classes identified based on supervised classification.
Full size table

Accuracy assessment

The classification accuracy reflects the extent of concordance between remote sensing data and reference data38. Assessing the precision of every classification is pivotal in determining the usability of the classified data for change detection analysis39. For the accuracy evaluation, high-resolution satellite imagery from Google Earth and visual interpretation were utilized35. Error matrices were used to statistically compare the reference data and the outcomes of the classification process. In this study, the accuracy of images was evaluated through the error matrix. These matrices were created to assess user accuracy, producer accuracy, and the overall accuracy of the classification. Furthermore, a nonparametric Kappa test was carried out to comprehensively measure the classification accuracy, considering all elements in the confusion matrix, not just the diagonal ones40. The Kappa coefficient was computed using the formula presented in Eq. 1:

$$\:\text{I}\text{K}=\frac{\text{N}\:{\sum\:}_{\text{i}=1}^{\text{r}}{\text{X}}_{\text{i}\text{i}}-\sum\:_{\text{i}=1}^{\text{r}}{\text{X}}_{+\text{i}}}{{\text{N}}^{2}\sum\:_{\text{i}=1}^{\text{r}}{(\text{X}}_{\text{i}+}{\times\:\text{X}}_{+\text{i}})}\:$$
(1)

where \(\:K\) = kappa coefficient, \(\:N\:\)=t the total number of values, \(\:\pi\:\sum\:_{i=1}^{r}{X}_{ii}\) = observed accuracy and \(\:\sum\:_{i=1}^{r}{(X}_{i+}{\times\:X}_{+i})\) = chance accuracy.

LULC change detection and analysis

This study detected changes in land use/cover using ArcGIS 10.8 to pinpoint modifications. The aim of employing change detection was to compare and analyze the LULC maps derived from visual interpretation and subsequent supervised classification, covering the period from 1991 to 202134. The change detection technique involves describing and quantifying images captured of the same scene at different points in time. Classified images from various temporal instances were employed to ascertain the area covered by different land types, measured in km2 and percentages. Additionally, this approach facilitated the observation and identification of alterations in various LULC classes, such as the expansion of urban built-up areas or the reduction of forested regions, among other changes41. The calculation of percentage change within the identical LULC type between two time intervals was executed using Eq. 2:

$$\:\text{P}\text{e}\text{r}\text{c}\text{e}\text{n}\text{t}\left(\text{\%}\right)\text{C}\text{h}\text{a}\text{n}\text{g}\text{e}=\frac{{\text{A}}_{\text{x}}-\:{\text{A}}_{\text{x}-1}}{{\text{A}}_{\text{x}-1}}$$
(2)

where \(\:{A}_{x}\)= area of specific LULC type at time x, \(\:{A}_{x-1}\)= area of specific LULC at the time \(\:x-1\).

Normalized difference vegetation index estimation

The computation of the NDVI, a metric indicating the abundance of vegetation, was performed for each image. NDVI relies on contrasting the peak reflectance of radiation in the near-infrared spectral range with the absorption of radiation in the red spectral range. Mathematically, it is calculated using Eq. 342:

$$\:\text{N}\text{D}\text{V}\text{I}=\frac{\:\text{N}\text{I}\text{R}-\text{R}\text{e}\text{d}}{\text{N}\text{I}\text{R}+\text{R}\text{e}\text{d}}\:$$
(3)

The range of NDVI values is from − 1.0 to + 1.0. Universally recognized as the most commonly utilized vegetation index for monitoring plant status and abundance, NDVI is chosen for its positive correlation with these characteristics43.

Normalized difference built-up index estimation

The NDBI, an indicator of built-up areas, was computed for each available image. NDBI is derived from the difference between the highest reflection of radiation in the shortwave infrared spectral bands (1.55–1.75 μm) and the highest absorption of radiation in the visible and near-infrared spectral bands (0.63–0.69 μm)44. Mathematically, it is calculated using Eq. 4:

$$\:\text{N}\text{D}\text{B}\text{I}=\frac{\:\text{S}\text{W}\text{I}\text{R}-\text{N}\text{I}\text{R}}{\text{S}\text{W}\text{I}\text{R}+\text{N}\text{I}\text{R}}$$
(4)

NDBI values range from − 1.0 to + 1.044 NDBI is recognized for its effectiveness and is commonly employed as an index for monitoring built-up areas due to its positive correlation with urbanization and impervious surfaces45.

Retrieving land surface temperature

Obtaining the LST involved utilizing data from Band 6 of the Landsat TM, which operates in the thermal infrared spectrum with a spatial resolution of 60 m. Subsequently, the TIRS bands, initially acquired at a resolution of 100 m, underwent resampling to achieve a 30-meter spatial resolution in the delivered data product.

The calculation of LST was performed utilizing Eq. 5

$$\:\text{L}\text{S}\text{T}\:=\frac{\text{T}}{1+\text{w}\times\:\left[\left(\frac{\text{T}}{\text{p}}\right)\times\:\text{ln}\left(\text{e}\right)\right]}\:\:$$
(5)

Where (\(\:T\)) = At-satellite brightness temperature, (\(\:w\)) = wavelength of the emitted radiance from the thermal infrared band, and (\(\:e\)) = land surface emissivity, typically ranging between 0.99 and 1.01.

$$\:p\:=\frac{h\:\times\:\:c}{s\:\left(1.438\:\times\:\:{10}^{-2}\:m\:K\right)}\:\:\:\:$$
(6)

where \(\:h\) =Planck’s constant (6.626 × 10−34 Js), \(\:c\) =velocity of light (2.998 × 108 m/s), \(\:s\) = Boltzmann constant (1.38 × 10−23 J/K).

Conversion of the digital number (DN) to TOA spectral radiance (L λ)

This process of the transformation of digital number (DN) values into spectral radiance was carried out using Eq. 7:

$$\:{L}_{\lambda\:}=\:{M}_{L}{Q}_{cal}+\:{A}_{L}$$
(7)

Where (\(\:{L}_{\lambda\:}\)) represents spectral radiance in (\(\:\text{W}/{(\text{m}}^{2}\times\:\:\text{s}\text{r}\text{a}\text{d}\:\times\:\:{\upmu\:}\text{m})\)), (\(\:{M}_{L}\)) and (\(\:{A}_{L}\)) denote band-specific multiplication and additive rescaling factors obtainable from the header file of the images, and (\(\:{Q}_{cal}\)) signifies the digital number (DN), representing quantized and calibrated standard product pixel values.

Transformation of spectral radiance (L λ) into satellite brightness temperature (T)

The conversion of spectral radiance to At-satellite brightness temperature involved utilizing thermal constants specified in the metadata as detailed in Eq. 8:

$$\:T=\frac{{K}_{2}}{\text{ln}\left(\frac{{K}_{1}}{{L}_{\lambda\:}}\right)}\:\:$$
(8)

Where (T) = At-satellite brightness temperature in Kelvin, (\(\:{K}_{1}\)) and (\(\:{K}_{2}\)) = band-specific thermal conversion constants (calibration constants) (\(\:{L}_{\lambda\:}\)) = spectral radiance in (\(\:W/{(m}^{2}\times\:\:srad\:\times\:\:\mu\:m)\)). Specifically, for Landsat 5 Band 6, (\(\:{K}_{1}\)= \(\:607.76/666.09)\:\)and (\(\:{K}_{2}\)= \(\:1260.56/1282.71\)) (\(\:\text{W}{\text{m}}^{-2}\:{\text{s}\text{r}}^{-1}\:{\upmu\:}{\:\text{m}}^{-1}\)), while for Landsat 9 Bands 10 (\(\:{K}_{1}\)= \(\:174.85/480.89)\:\)and (\(\:{K}_{2}\)= \(\:1321.08/1210.14\)) (\(\:\text{W}{\text{m}}^{-2}\:{\text{s}\text{r}}^{-1}\:{\upmu\:}{\:\text{m}}^{-1}\)).

Derivation of land surface emissivity (LSE)

The determination of LSE plays a vital role in obtaining LST. The calculation of LSE (\(\:e\)) is based on Eq. 9. To initiate the process, the calculation involves the derivation of the vegetation proportion (\(\:{P}_{v}\)) using Eqs. 10,

$$\:e\:=\:\left(0.004\:\times\:{P}^{v}\right)+\:0.986\:\:\:$$
(9)
$$\:{{P}_{v}=\left(\frac{NDVI-\:{NDVI}_{min}}{{NDVI}_{max}-\:{NDVI}_{min}}\right)}^{2}\:$$
(10)

Conversion of LST from Kelvin to degree celsius

Upon computing the emissivity-corrected land surface temperatures in Kelvin, for clarity, the LSTs were converted to degrees Celsius using the conversion factor, where 0 °C corresponds to 273.15 K. Following the relation given below, the temperature values were subsequently converted to degrees Celsius by subtracting 273.15:

$$\:LST\:\left(℃\right)=\:LST\:\left(K\right)\--\:273.15\:\:\:$$
(11)

Results

LULC classification

The comprehensive examination of LULC dynamics in the study area, spanning the years 1991 to 2023, reveals substantial changes in various key categories (Fig. 3). The findings from the analysis of multi-temporal satellite images are visually depicted in a table (Table 3). In 1991, the study area had a considerable vegetation cover of 1607.6 km2, constituting 63.6% of the total area covered by the land. In the following years, this vegetation area experienced a noticeable decrease., reaching 1424.8 km2 by 2023, marking a substantial reduction of 415.1 km2 or 22.6%. The percentage of land covered by vegetation decreased from 63.6 to 56.35% during this period, emphasizing a concerning trend in diminishing green cover. This decline may be attributed to increased agricultural expansion and urban development, leading to habitat loss and potential impacts on biodiversity. Concurrently, urbanization in the study area underwent significant growth. In 1991, metropolitan areas covered 172.7 km2, accounting for 6.8% of the district’s land. By 2023, urban areas expanded remarkably to 283.4 km2, marking an impressive growth of 110.7 km2.

Table 3 LULC classes of the study district (km2).
Full size table

Or 64.1%. The percentage of land occupied by urban areas rose from 6.8 to 11.21%, reflecting a substantial shift in the district’s landscape towards increased urban development. This urban sprawl raises concerns about land use conflicts, increased heat islands, and strain on local resources. Water bodies in the district displayed dynamic fluctuations. In 1991, the area covered by water bodies was 42.5 km2, constituting 1.7% of the total area covered by the land. However, by 2001, there was a deterioration of 28.5 km2 (1.1%), followed by a succeeding increase to 71.4 km2 (2.82%) in 2023. The rise in water-covered regions in the research area results from the recent flooding that impacted the locality https://www.dawn.com/news/1771059. Further, the fluctuations underscore the dynamic nature of water bodies, mainly the Satluj River, which has passed through the study area over the years, necessitating a closer examination of local hydrological processes and potential impacts on water resources. Bare land, a significant component of the district’s LULC, experienced substantial changes. In 1991, the area covered by bare land was 705.7 km2, representing 27.9% of the district. By 2023, there was a notable increase to 748.9 km2, accounting for 29.62% of the total land area. This shift indicates changes in land use practices, potentially driven by agricultural intensification or inadequate soil conservation measures, raising considerations for sustainable land management and the need for strategies to mitigate soil degradation.

An accuracy evaluation was conducted to appraise the applicability of the classified data in change analysis. Confusion Matrices for the classified images of the study district are shown in (Supplementary Fig. 1). The comprehensive classification accuracy assessment and Kappa statistics of the LULC information derived from Landsat scenes reached approximately 90%, by the established accuracy criterion of 90% for LULC mapping studies, as suggested by Lea & Curtis (2010). These are distribution values for LULC (Table 3) for the years 1991, 2001, 2003, and 2023 provide a quantitative foundation for assessing the impact of anthropogenic activities on land cover, informing future research directions, and facilitating evidence-based decision-making in the realms of sustainable land management and environmental conservation.

LULC dynamics (1991–2023)

The thorough examination of LULC changes (Fig. 4) in the study area from 1991 to 2023 reveals complex patterns in land cover classes, including vegetation, urban areas, water bodies, and bare land (Tables 3 and 4). In the vegetation domain, the period from 1991 to 2001.

Table 4 Changes in LULC for 1991–2001, 2001–2011, 2011–2023 and 1991–2023.
Full size table

Witnessed a notable increase of 138.4 km2, constituting 8.6%. This growth was followed by a subsequent decrease of 93.8 km2 or 5.4% from 2001 to 2011. However, the most substantial shift occurred from 2011 to 2023, revealing a significant decline of 415.1 km2 or 22.6%. This cumulative reduction over the entire period amounted to 182.8 km2 or 11.4%, indicative of a consistent negative trend. The calculated rate of change stands at -4.35 km2 per year, emphasizing the urgency for conservation efforts and the implementation of reforestation initiatives to counteract this declining trend. Urban areas in the study area experienced noteworthy growth. From 1991 to 2001, there was an increase of 22.8 km2, representing a growth of 13.2%. From 2001 to 2011, the subsequent decade, we witnessed a further rise of 13.7 km2 or 7.0%. The most significant urban expansion was observed from 2011 to 2023, resulting in a substantial change of 110.7 km2 or 64.1%. The overall growth from 1991 to 2023 is an impressive 74.2 km2 or 35.5%. The calculated rate of change is 2.64 km2 per year, underscoring the need for careful urban planning to effectively manage associated challenges and sustain the district’s developmental trajectory. While exhibiting dynamic changes, water bodies declined − 14.1 km2 or -33.1% from 1991 to 2001. Subsequently, the period from 2001 to 2011 displayed a positive change of 11.1 km2 or 39.0%. The most substantial increase occurred from 2011 to 2023, resulting in a change of 31.9 km2 or 80.5%. The overall growth from 1991 to 2023 is 28.9 km2 or 67.9%. The calculated rate of change is 0.69 km2 per year, highlighting the need for a detailed examination of local hydrological processes and sustainable water resource management techniques to maintain the delicate balance of water bodies. Bare land, a critical component of the study area’s landscape, showed obscure changes over the study period. There was a substantial decrease of -147.1 km2 or -20.8% from 1991 to 2001, followed by a further decline of -118.6 km2 or -21.2% from 2001 to 2011. However, from 2011 to 2023, there was a significant increase of 308.9 km2 or 70.2%. The overall change from 1991 to 2023 is an addition of 43.2 km2 or 6.1%. The calculated rate of change is 1.03 km2 per year, indicating a consistent upward trajectory in bare land cover (Table 4). Effectively addressing land degradation concerns and implementing soil conservation measures centers on thoroughly understanding this expansion’s underlying drivers.

Trends in transition between LULCs

The examination of LULC changes in the study area from 1991 to 2023 reveals subtle trends in the transitions between various land use categories, providing valuable insights into the evolving landscape (Supplementary Fig. 2). During the initial decade (1991–2001), a noteworthy positive trend was observed in the transition from vegetation to urban areas, signified by a substantial increase of 138.4 km2 in urban coverage (Fig. 5). This shift shows a growing emphasis on urban development, potentially driven by population growth and economic activities. Simultaneously, there is a noticeable decrease in water bodies by 14.1 km2 and bare land by -147.1 km2. This reduction in water bodies and raw land signifies a transition towards more developed and built-up areas. Interestingly, vegetation cover shows significant expansion during this period, reflecting a positive trend in green cover. In the subsequent decade (2001–2011), urban areas continue to grow, albeit slightly slower, with an additional 93.8 km2 (Fig. 5). This sustained urbanization suggests ongoing developmental activities and infrastructure expansion. The growth in water bodies by 11.1 km2 during this phase indicates efforts towards sustainable development and environmental conservation. However, there is a substantial decrease in bare land by 118.6 km2, signaling a reduction in unused or undeveloped areas. This shift aligns with the broader trend of converting open spaces into more urbanized and developed zones. The period from 2011 to 2023 unfolds significant transitions, characterized by a substantial decrease of 415.1 km2 in vegetation (Fig. 5).

Fig. 5
figure 5

Change detection between 1991 and 2001, 2001 and 2011, 2011 and 2023, and 1991 and 2023.

Full size image

This decline raises concerns about the loss of green cover, potentially due to urban expansion or other anthropogenic factors. Notably, urban areas experience the most significant expansion during this phase, growing by 74.2 km2. This emphasizes an ongoing trend of urbanization, indicating increased urban infrastructure and residential development. The positive change in water bodies by 31.9 km2 and the increase in bare land by 308.9 km2 during this period highlight efforts towards balancing development with environmental preservation. The growth in water bodies suggests a conscious effort to maintain or enhance natural features. At the same time, expanding bare land may indicate strategic planning to conserve certain areas from intensive development. The overall trend from 1991 to 2023 reveals a net loss of 182.8 km2 in vegetation, signaling a concerning decline in green spaces (Fig. 5). Urban areas, on the other hand, exhibit consistent growth, increasing by 110.7 km2. The positive change in water bodies by 28.9 km2 and the addition of 43.2 km2 in bare land underscores efforts towards sustainable development practices. Highlighting these trends emphasizes the imperative of well-informed and equitable decision-making to secure the region’s long-term ecological sustainability and resilience.

Normalized difference vegetation index (NDVI)

The analysis of NDVI data from 1991 to 2023 provides a comprehensive overview of vegetative health and dynamics in the study area over the years (Fig. 6). The NDVI values, calculated based on satellite imagery, serve as a crucial indicator of the density and vitality of vegetation in the region46. In 1991, the maximum NDVI value was 0.5, indicating a relatively dense and healthy vegetation cover, while the minimum reached − 0.21, representing areas with sparse or no vegetation (Table 5). The average NDVI for the year 1991 is 0.23, reflecting a moderate overall vegetative condition. Moving to 2001, the NDVI values displayed subtle changes. The maximum NDVI slightly decreased to 0.48, indicating a slight reduction in the density of the healthiest vegetation. However, the minimum NDVI improved to -0.05, representing a reduction in areas with sparse or no vegetation. The average NDVI increased to 0.26, indicating a general improvement in the overall vegetative health and density compared to 1991. In 2011, the maximum NDVI was reduced to 0.47, implying a potential decrease in the healthiest vegetation cover. However, the minimum NDVI further improved to -0.02, pointing towards a reduction in areas with sparse or no vegetation. The average NDVI remained at 0.23, like the value in 1991, suggesting a maintenance of the overall moderate vegetative condition. The most recent data from 2023 indicates a slight recovery in vegetative health. The maximum NDVI increased to 0.48, potentially signaling an enhancement of the healthiest vegetation cover (Table 5). The minimum NDVI, at -0.09, shows a slight decrease in areas with sparse or no vegetation compared to 2011. The average NDVI reached 0.27, reflecting an overall improvement in vegetative health and density compared to the previous years. These variations are influenced by climate, land use changes, or other environmental dynamics47. The consistent monitoring of NDVI provides valuable insights for assessing the impact of natural and anthropogenic factors on the region’s vegetation, guiding informed environmental management and conservation efforts.

Table 5 Statistics describing the NDVI, NDBI, and LST across the four periods.
Full size table
Fig. 6
figure 6

NDVI maps of the study district for the period (a) 1991, (b) 2001, (c) 2011 and (d) 2023 (this figure was generated by the author using ArcGIS version 10.8 (Esri Inc., Redlands, CA, USA; https://www.esri.com/en-us/arcgis/about-arcgis/overview).

Full size image

Normalized difference built-up index (NDBI)

The NDBI data from 1991 to 2023 shows essential results regarding the built-up and urbanization trends in the study area over the years (Fig. 7). NDBI values, derived from satellite imagery, serve as a critical indicator of the extent of built-up areas within the region48. In 1991, the NDBI data indicated a maximum value of 0.28, reflecting the presence of built-up structures and urban surfaces. The minimum value recorded was − 0.34, representing areas with little to no built-up structures. The average NDBI for the year was − 0.1, indicating a moderate level of built-up coverage (Table 5). Moving to 2001, the NDBI values displayed some noteworthy changes. The maximum NDBI was reduced to 0.09, and the minimum NDBI was − 0.34. The average NDBI decreased slightly to -0.12, indicating a subtle decrease compared to 1991. In 2011, the NDBI values exhibited further adjustments. The maximum NDBI increased to 0.41, marking a significant expansion or modification of built-up areas. The minimum NDBI remained at -0.34, and the average NDBI stayed at -0.11, comparable to 2001. The most recent data from 2023 reveals some interesting trends. The maximum NDBI decreased to 0.13, and the minimum NDBI also saw a slight decrease of -0.31. The average NDBI remained consistent.

Fig. 7
figure 7

NDBI maps of the study district for the period (a) 1991, (b) 2001, (c) 2011 and (d) 2023 (this figure was generated by the author using ArcGIS version 10.8 (Esri Inc., Redlands, CA, USA; https://www.esri.com/en-us/arcgis/about-arcgis/overview).

Full size image

They reflect a stable built-up intensity and impervious surface coverage compared to the preceding years. The fluctuations in maximum, minimum, and average values suggest changes in the extent and intensity of built-up areas over the years. These variations are influenced by urban expansion, infrastructure development, or land use changes49. Continually monitoring NDBI is crucial for understanding the evolving urban landscape, informing urban planning strategies, and ensuring sustainable development practices in the region.

Land surface temperature (LST)

The examination of LST data from 1991 to 2023 provides significant insights into the thermal characteristics of the study area over the years (Fig. 8). LST, derived from satellite observations, is a crucial indicator of the landscape’s heat distribution and temperature patterns50. In 1991, the LST data showed a maximum temperature of 30.49 °C, indicating the warmest recorded surface temperature in the district. The minimum temperature was 2.59 °C, reflecting the most excellent recorded surface temperature (Table 5). The average LST for the year was 13.56 °C, representing the overall thermal condition of the landscape. Moving to 2001, the maximum temperature decreased to 28.44 °C, suggesting a reduction in the highest surface temperatures. The minimum temperature increased to 5.26 °C, indicating a rise in the coolest recorded temperatures. The average LST remained relatively stable at 13.82 °C, indicating a consistent thermal condition compared to 1991. In 2011, the maximum temperature increased significantly to 35.13 °C, indicating a notable rise in the warmest surface temperatures. The minimum temperature also increased to 4.33 °C, indicating a slight rise in the coolest recorded temperatures. The average LST saw a notable increase to 14.71 °C, reflecting an overall increase in thermal conditions compared to the preceding years. The most recent data from 2023 indicates some exciting trends. The maximum temperature decreased substantially to 24.45 °C, indicating a considerable reduction in the warmest surface temperatures. The minimum temperature dropped to -3.16 °C, indicating a shift toward more excellent conditions or potential changes in local climate patterns. The average LST further decreased to 8.97 °C, representing a significant decrease in the overall thermal conditions compared to 2011. There is notable spatial diversity in Okara District, according to the analysis of LST trends. Temperature increases are not consistent; in contrast to more vegetated or rural areas, certain places—especially metropolitan centers and areas with less vegetation cover—experience greater temperature increases. The variations in maximum, minimum, and moderate temperatures suggest changes in the local climate or environmental conditions. Land use changes, vegetation dynamics, or broader climatic trends influence these fluctuations51,52. Continuous monitoring of LST is crucial for understanding temperature variations, guiding climate studies, and informing environmental management practices in the region.

Fig. 8
figure 8

LST maps of the study district for the period (a) 1991, (b) 2001, (c) 2011 and (d) 2023 (this figure was generated by the author using ArcGIS version 10.8 (Esri Inc., Redlands, CA, USA; https://www.esri.com/en-us/arcgis/about-arcgis/overview).

Full size image

LULC classification and changes analysis

The analysis of LULC classification and changes reveals a detailed dataset presenting the dynamic transformation of the study area’s landscape between 1991 and 2023 (Supplementary Fig. 3). The categorization into land cover classes, including Vegetation, Urban areas, Water bodies, and Bare Land, facilitates a comprehensive examination of the spatial dynamics in the study area. This detailed breakdown forms a solid foundation for understanding the varied changes.

They were observed in the study area. In 1991, the landscape had a substantial Vegetation cover, spanning 1607.6 km2 or 63.6% of the total area. Urban areas accounted for 172.7 km2 (6.8%), Water bodies covered 42.5 km2 (1.7%), and Bare Land constituted 705.7 km2 (27.9%) (Table 4). These baseline figures serve as pivotal reference points for analyzing the subsequent shifts in land use patterns. The period from 1991 to 2001 witnessed essential transitions. Vegetation exhibited a subsequent increase of 138.4 km2 (8.6%), indicating growth (Supplementary Fig. 3). Urban areas expanded by 22.8 km2 (13.2%), underscoring an expanding urbanization trend. Water bodies experienced a decrease of -14.1 km2 (-33.1%), suggesting alterations in hydrological patterns. Simultaneously, Bare Land saw a reduction of -147.1 km2 (-20.8%), representing dynamic changes in land utilization practices. Advancing to the years between 2001 and 2011, Vegetation sustained its positive trajectory, adding 93.8 km2 (5.4%). The expansion of Urban areas continued, amounting to 13.7 km2 (7.0%), emphasizing the persistence of urbanization trends. Water bodies displayed a positive change of 11.1 km2 (39.0%), and Bare Land decreased. by 118.6 km2 (21.2%). The most recent phase, from 2011 to 2023, shows distinctive trends. Vegetation encountered a substantial decrease of 415.1 km2 (22.6%), emphasizing a concerning loss of green cover. While Urban areas expanded significantly by 110.7 km2 (64.1%), indicating the study area’s expanding urban footprint. Water bodies witnessed a positive change of 31.9 km2 (80.5%). Simultaneously, Bare Land saw a substantial increase of 308.9 km2 (70.2%), raising questions about the factors influencing this expansion. The cumulative changes from 1991 to 2023 reveal a net decrease of 182.8 km2 in Vegetation along with a significant growth of 110.7 km2 in Urban areas. This signifies strategic urban planning to balance development with environmental sustainability. The positive change of 28.9 km2 in Water bodies means successful conservation measures, while the addition of 43.2 km2 in Bare Land indicates an exploration of the drivers behind this expansion. Undoubtedly, these statistics represent more than numerical variations; they encapsulate the narrative of a changing landscape. They call upon decision-makers, urban planners, and environmentalists to explore the complexities of these changes more profoundly, aiming for sustainable solutions that guarantee the resilience and equilibrium of the environment in the study area amid continuing urbanization and environmental challenges53,54.

Correlation between LST and NDVI

The examination of the correlation between LST and NDVI across the study area over the years 1991, 2001, 2011, and 2023 consistently reveals negative associations, underscoring the nuanced interplay between surface temperature and vegetation abundance and density (Fig. 9). 1991, the correlation coefficient (R2) of 0.55 indicates a moderate correlation. As NDVI values rise, representing denser vegetation, LST tends to decrease. For every 1 unit decrease in NDVI, the regression model indicates a predicted increase of approximately 30.12 units in LST. About 55% of the variability in LST can be explained by changes in NDVI, underscoring the significant influence of vegetation on temperature regulation. Moving to 2001, the R2 value 0.52 signifies a similar moderate correlation. As NDVI increases, LST tends to decrease, with approximately 52% of the variability in LST attributed to changes in NDVI. This consistent negative correlation emphasizes the persistent impact of vegetation on surface temperature dynamics. In 2011, the R2 value of 0.40 indicates a moderate correlation, though slightly weaker than in previous years. The negative.

Fig. 9
figure 9

Correlation plots between NDVI and LST, NDVI and NDBI, and NDBI and LST in the study district.

Full size image

The relationship persists, suggesting that about 40% of the variability in LST is explained by changes in NDVI. This underscores the ongoing influence of vegetation on temperature patterns. The relationship in 2023 stands out with an R2 value of 0.65, indicating a relatively strong correlation. As NDVI increases, LST tends to decrease, with approximately 65% of the variability in LST attributed to changes in NDVI. This suggests a more noticeable influence of vegetation on temperature dynamics this year. With each unit decrease in NDVI, the regression model indicates an estimated rise of around 17.57 units in LST. While the strength of the correlation varies across the years, the overall trend highlights the significant and persistent influence of vegetation dynamics on the thermal characteristics of the study area55. These findings imply that maintaining and enhancing vegetation cover is crucial for regulating surface temperatures. Land management practices should prioritize the preservation and expansion of green spaces to mitigate the effects of rising temperatures and improve urban climate resilience.

Correlation between NDVI and NDBI

The correlation between NDBI and NDVI in the 1991, 2001, 2011, and 2023 study areas provides significant insights into the intricate dynamics between urbanization and vegetation density (Fig. 9). In 1991, the negative correlation with an R2 value of 0.82 indicates a strong relationship. As the NDBI, representing built-up areas, increases, there is a corresponding decrease in NDVI, signifying a reduction in vegetation density. Approximately 82% of the variability in NDVI can be attributed to changes in NDBI. For every unit increase in NDBI, the regression model suggests an expected decrease of approximately 0.75 units in NDVI. This highlights a significant pattern, indicating that the considerable reduction of vegetation cover is linked to urban expansion in 1991. This emphasizes the transformative influence of urbanization on the landscape. In 2001, the R2 value 0.84 represents a continued strong negative correlation. As NDBI increases, indicating further urban development, NDVI experiences a decrease. Around 84% of the variability in NDVI is explained by changes in NDBI. This restates the continued influence of urbanization on.

Vegetation reduction in 2001, emphasizing the sustained impact of built-up areas on the surrounding green cover. In 2011, the R2 value 0.68 signifies a moderately strong negative correlation. The negative relationship persists, indicating that approximately 68% of the NDVI variability is explained by NDBI changes. Although slightly weaker than in previous years, this correlation highlights the continued impact of urbanization on vegetation dynamics in 2011. The relationship in 2023, with an R2 value of 0.80, maintains a strong negative correlation. As NDBI increases, suggesting further urban expansion, NDVI tends to decrease. Approximately 80% of the variability in NDVI can be attributed to changes in NDBI. As per the regression model, a unit increase in NDBI is anticipated to decrease by approximately 0.66 units in NDVI. The variations in the strength of these correlations over the years emphasize the evolving nature of this interaction, offering valuable insights for sustainable urban planning and environmental conservation strategies in the face of urban expansion56. These results highlight the need for integrated urban planning that balances development with the preservation of green spaces. Strategies such as green roofs, urban forests, and parks can help mitigate the negative impacts of urbanization on vegetation density and enhance the overall environmental quality.

Correlation between LST and NDBI

The persistent moderate positive correlations observed between LST and NDBI in the study area for 1991, 2001, 2011, and 2023 highlight the thermal dynamics influenced by urban development (Fig. 9). In 1991, the moderate correlation with an R2 value of 0.65 indicates that as the NDBI, representing built-up areas, increases, there is a corresponding rise in LST. Approximately 65% of the variability in LST can be explained by changes in NDBI. According to the regression model, for each unit increase in NDBI, there is an expected decrease of around 38.81 units in LST. This implies that increased urban areas and reduced vegetative cover associated with urbanization contributed to elevated surface temperatures. Moving to 2001, the R2 value of 0.63 signifies a continued moderate correlation. As NDBI increases, LST tends to rise, with around 63% of the variability in LST attributed to changes in NDBI. This persistent positive correlation suggests the ongoing influence of the UHI effect in 2001. In 2011, the R2 value of 0.62 indicates a moderate correlation.

The positive relationship continues, suggesting that about 62% of the variability in LST is explained by changes in NDBI. This highlights the continued impact of urban development on local thermal conditions. The correlation in 2023, with an R2 value of 0.69, maintains a moderate correlation. As NDBI increases, LST tends to rise, with approximately 69% of the variability in LST attributed to changes in NDBI. By the regression model, a unit increase in NDBI is associated with an estimated decrease of approximately 20.21 units in LST. The variations in the strength of these correlations over the years provide subtle details of the evolving thermal dynamics influenced by urbanization, emphasizing the need for sustainable urban planning to mitigate the heat island effect and ensure the well-being of the local environment57.These findings underscore the importance of implementing urban heat island mitigation strategies, such as increasing urban greenery, using reflective building materials, and enhancing natural ventilation in urban areas. Such measures can help reduce surface temperatures and improve the livability of urban environments.

Discussion

The increasing trend of global urbanization, particularly in the burgeoning urban areas of Pakistan, necessitates a thorough examination of LULC changes and their extensive environmental repercussions. This urbanization is closely linked to climate change, as it contributes to increased greenhouse gas emissions and alters local climate conditions. This rise in urbanization poses significant challenges to sustainable development, natural resource management, and ecosystem integrity. As cities expand and populations grow, the demand for land for infrastructure, housing, and commercial activities intensifies, leading to widespread alterations in land cover patterns. Consequently, these changes can have profound implications for biodiversity, water resources, air quality, and climate stability. Extensive research has been conducted globally on LULC change, focusing on developed regions like China, Europe, and the United States58. However, Pakistan stands out as one of the least-studied Asian countries at local and national levels28. This knowledge gap impacts global efforts to manage Earth’s resources effectively.

Okara, a critical agricultural region in Punjab, Pakistan, has undergone significant changes in LULC, as demonstrated by59. The region’s susceptibility to natural disasters further emphasizes the importance of monitoring these changes60. For instance, our findings reveal a correlation between rapid LULC changes and increased vulnerability to climate-related hazards, necessitating integrated disaster risk management strategies. The research conducted by Hadeel et al., in Northern, Australia highlighted the utilization of remote sensing data and GIS techniques in LULC change detection61, echoing findings from studies in other regions such as India and Egypt62,63.

Building upon this research, our current study classified LULC into four categories: Vegetation, Built-up, Water, and Bare Land. The analysis revealed changes from 1991 to 2023 for each category, with percentages of -11.4%, 64.1%, 67.9%, and 6.1%, respectively. These patterns reflect broader trends of urbanization and land degradation documented in similar studies, underscoring the urgent need for targeted conservation strategies. Most importantly, the detected patterns of LULC change closely resembled those observed in the research of Hussain65, even though slight variation is ascribed to periodic differences. The results highlight the rapid pace of LULC changes occurring nationally and locally throughout Pakistan. Rapid population growth and the shrinking per capita agricultural land are pressing concerns for food security. The scientific documentation of LULC changes over multiple decades is crucial for understanding the implications of these transformations on human well-being66. Significantly, a multitude of factors, including poverty, overcrowding, illegal deforestation, agricultural expansion, and insufficient legislation and policy enforcement, have collectively contributed to the decline in forest cover67,68,69. For instance, our study highlights a reduction in the Vegetation area in Okara by 11.4% between 1991 and 2023, coinciding with a significant increase in barren land by 64.1% during the same period.

Butt et al., studied the LULC changes in Islamabad’s watershed using supervised classification from 1992 to 2012, revealing substantial alterations in agricultural land, vegetative area, urban area, bare soil, and aquatic bodies34. Moreover, in their study, Hussain et al., documented the conversion of cropland and forest areas to roads and human settlements in the Lodhran district, reducing bare soil and vegetation by 5.2% between 1977 and 201770. Rapid urbanization in districts like Lodhran has led to haphazard expansion, straining available resources, and degrading infrastructure. Concurrently, Manzoor et al., underscored the substantial urbanization occurring in metropolitan areas, which has resulted in a gap of information concerning LULC development at both local and regional scales, including within Pakistan71. This gap poses challenges for comprehensive understanding and effective management of LULC dynamics in rapidly urbanizing areas. Nevertheless, local and regional LULC development gaps persist due to a need for more technical expertise and efficient investigation methods65.

Numerous socioeconomic factors have an impact on the LULC alterations that have been noticed in Okara District. Urban expansion is being driven by the increased demand for housing and infrastructure because of rapid population growth. Services, industry, and agriculture are examples of economic activity that are important. Economic incentives and the prospect of greater returns are frequently the driving forces behind the conversion of agricultural land to urban and industrial usage. These changes are also influenced by patterns of land ownership, such as the consolidation of land for extensive industrial or agricultural enterprises. LULC dynamics are also influenced by national and local policy decisions, including those pertaining to development incentives, land use rules, and zoning laws. Policies that encourage urbanization or industrialization, for example, might hasten the transformation of natural or agricultural regions into populated places. Comprehending these socioeconomic factors is essential to creating all-encompassing land management plans that strike a balance between environmental sustainability and development requirements.

The examination revealed significant correlations among land surface variables over the study period. Specifically, there was a discernible negative correlation between NDVI and LST, with R2 values ranging from 0.55 to 0.69, indicating that as NDVI increased, LST tended to decrease. This negative relationship suggests that areas with higher vegetation cover experience lower surface temperatures. The analysis revealed notable insights into the relationship between vegetation fraction and NDVI, showing a strong positive correlation. At the same time, urban areas and barren land areas exhibited a negative correlation with NDVI over the three-decade period. This observation aligns with previous studies72,73, which reported similar negative correlations between NDVI and LST in different geographical regions. Additionally, a strong negative correlation was observed between NDVI and NDBI, having R2 values from 0.62 to 0.69, implying that as NDVI values rose, NDBI values tended to decline. Conversely, a positive correlation emerged between NDBI and LST, with R2 values from 0.62 to 0.69, indicating that higher built-up densities corresponded to higher surface temperatures. This positive relationship is consistent with previous research74, which similarly identified positive correlations between NDBI and LST in urban environments. These findings underscore the complex interplay among vegetation cover, urban development, and surface temperature dynamics within the study area.

Changes in LULC have substantial effects on human health and agriculture. Reduced vegetation cover and urbanization have a significant impact on agricultural operations, affecting soil quality, crop yields, and water availability. While changes in vegetation cover can cause soil erosion and degradation, which negatively affects soil fertility, urban sprawl-induced loss of arable land can result in lower agricultural production. Increased temperatures and changed precipitation patterns can also stress crops, requiring more thorough irrigation techniques. Urbanization-related increases in temperature pose health dangers to people, including the escalation of long-term illnesses. Mitigation techniques are crucial to addressing these issues. These include improving urban green spaces, putting heat-health action plans into place, and raising public awareness of the dangers of heat-related illnesses. In addition to lowering surface temperatures, these actions enhance urban people’s general resilience and well-being.

Our findings highlight the need for informed urban planning and management strategies to minimize and mitigate the adverse impacts of LULC changes. Expanding infrastructure and commercial sectors have profound implications for vegetated regions, necessitating informed policymaking and sustainable management practices. Furthermore, integrating RS data and GIS techniques in LULC management can enhance decision-making processes and facilitate sustainable development initiatives. Governmental authorities must prioritize education and training in cutting-edge techniques to effectively deal with the challenges caused by the ever-changing LULC dynamics.

Conclusion

This study presents a comprehensive and systematic analysis of the changes in Land Use and Land Cover (LULC) and Land Surface Temperature (LST), along with an in-depth examination of the Normalized Difference Vegetation Index (NDVI) and Normalized Difference Built-up Index (NDBI) over the years 1991, 2001, 2011, and 2023. The data reveals significant transformations in the district’s landscape, providing critical insights into the dynamics of urbanization, vegetation cover, and their environmental implications. Key findings include a substantial 22.6% decline in vegetation cover (415.1 km2) and a 64.1% increase in urban areas (110.7 km2) from 1991 to 2023. Changes in water bodies and bare land further contribute to the evolving landscape. The exploration of correlations between NDVI and LST highlights a consistent negative relationship, with R2 values ranging from 0.55 to 0.69, emphasizing the cooling effect of vegetation on surface temperatures. Correlation analyses underline the influence of urbanization on surface temperature, with a statistically significant positive correlation between LST and NDBI, featuring R2 values from 0.62 to 0.69, indicating the persistent urban heat island effect. These findings emphasize the need for sustainable urban planning to balance developmental needs and ecological preservation. As we navigate the practical implications of these findings, it becomes evident that informed decision-making in urban planning is paramount. Acknowledging urbanization’s impact on temperature dynamics and vegetation cover, authorities can adopt strategic measures to mitigate the UHI effect in the study area.

Based on these findings, we urge future studies to examine the long-term ecological implications of urbanization on local biodiversity and the efficiency of green spaces in moderating the urban heat island effect. We recommend policymakers implement sustainable urban planning models that incorporate green infrastructure and community participation. Specific recommendations include developing green roofs, urban forests, and parks; implementing zoning laws to ensure a balance between built-up areas and natural landscapes; fostering community involvement in urban planning; promoting the use of reflective and sustainable building materials; and establishing continuous monitoring systems for LST and vegetation cover. These steps aim to balance urban development with environmental sustainability, ensuring the well-being of both the environment and the community.