Spatio-Temporal Assessment of Land Surface Temperature, Vegetation Cover, and Built-Up Areas Using LST, NDVI, and NDBI in Balıkesir, Türkiye (1985–2025)

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript uses multi-temporal Landsat imagery, combined with LST, NDVI, and NDBI indices, to analyze land surface temperature changes in Balıkesir, Türkiye, from 1985 to 2025, as well as their relationships with vegetation cover and built-up area expansion. After reading the manuscript, I have the following comments and suggestions:

1. Analyses combining LST, NDVI, and NDBI have already been extensively reported in the literature. The study lacks originality and should clearly articulate its added value compared with previous work.
2. I suggest you should update the literature and define the novelty. Some references are interesting for you to refer: Contributions of natural and anthropogenic factors to summertime thermal environments across different urban scales: An investigation in Chengdu-Chongqing agglomeration, China. Environmental Impact Assessment Review, 115, 107981. Capability of LCZ scheme to differentiate urban thermal environments in five megacities of China: Implications for integrating LCZ system into heat–resilient planning and design. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.
3. Although the formulas for calculating LST, NDVI, and NDBI are presented, there is no validation using ground-based observations or comparisons with existing studies.
4. The analysis relies solely on Pearson correlation, and the conclusions overemphasize “strong correlations.” Nonlinear relationships, multiple regression, or partial correlation should be considered to avoid overstating the results.
5. Correlation coefficients are reported without p-values or significance levels.
6. The study lacks concrete, actionable planning or management recommendations.
7. Some sentences are overly long and contain grammatical errors; the English writing requires professional editing.

In summary, the study suffers from insufficient originality, inadequate methodological details, vague presentation of results, and a lack of in-depth discussion. It is not suitable for acceptance in its current form.

Author Response

Comments 1: [We would like to thank the reviewers for their constructive comments on the article. In consideration of these comments, the article has been improved with the following remarks]

Response 1: [Analyses combining LST, NDVI, and NDBI have already been extensively reported in the literature. The study lacks originality and should clearly articulate its added value compared with previous work.

The requested changes have been made as follows.

This study, based on a continuous 40-year time series derived from Landsat (30 m) data for the period 1985–2025, revealed in detail the micro-scale heterogeneity of urban surface temperatures. In addition, MODIS (1 km) data provided a comparative perspective on macro-scale trends for the post-2005 period, while Sentinel-3 (1 km) data offered such a comparison only for the year 2025. Unlike most urban heat island studies in the literature, which are limited to a single satellite or short-term analyses, this study systematically compared micro (30 m) and macro (1 km) scales within the same warm-season windows, thereby developing both a long-term and multi-scale perspective. Furthermore, the increasingly strong positive correlation between LST and built-up index (NDBI), and the intensifying negative correlation between LST and vegetation index (NDVI), were statistically validated through Pearson and partial correlation analyses, thereby isolating the reinforcing role of urbanization on heat island intensity and the cooling effect of vegetation.

Building on these findings, a detailed planning output was developed for 2025 based on high-resolution Landsat LST trends, providing concrete recommendations for strengthening cooling corridors, prioritizing interventions in heat-intensive zones, and advancing green infrastructure strategies. Unlike previous studies, this approach not only described the spatial distribution of surface temperatures but also generated actionable strategies that can be directly integrated into urban planning processes. Moreover, it clearly demonstrated through comparative analysis the insufficiency of MODIS and Sentinel-3 data in capturing cool islands and micro-climatic variations at the urban texture scale, while highlighting the success of Landsat (30 m) data in accurately identifying urban hot spots and green cool islands, thus contributing a methodological framework on product–scale compatibility.

The issue raised in Reviewer Comment 1 has already been elaborated in detail in the manuscript, specifically between lines 158 and 189.]

Comments 2: [I suggest you should update the literature and define the novelty. Some references are interesting for you to refer: Contributions of natural and anthropogenic factors to summertime thermal environments across different urban scales: An investigation in Chengdu-Chongqing agglomeration, China. Environmental Impact Assessment Review, 115, 107981. Capability of LCZ scheme to differentiate urban thermal environments in five megacities of China: Implications for integrating LCZ system into heat–resilient planning and design. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.]

Response 2: [We consider the reviewers' feedback and add the following justifications to the introduction section in order to broaden the study:

The recommended articles have been thoroughly reviewed and incorporated into the study where relevant. In addition, to strengthen the scientific depth of the research, numerous other scholarly publications were also examined. Accordingly, the Introduction section has been substantially expanded, with the newly integrated content presented between lines 31–55 and 78–122.

The temperature increases caused by climate change, extreme weather events, and the urban heat island (UHI) effect are creating increasing pressures on the ecological balance and social living conditions of cities [1-3]. The Mediterranean Basin is one of the regions most intensely affected by climate change and stands out as a critical ‘hot spot’ due to the increase in extreme heat events [4]. The Intergovernmental Panel on Climate Change (IPCC) Sixth Assessment Report (AR6) reveals that temperature in-creases in the Mediterranean region are above the global average, with significant in-creases expected in the frequency, duration, and intensity of heatwaves during the summer months.

In addition, according to AR6, the UHI effect and increasing humidity deficit in Mediterranean cities are exacerbating thermal stress conditions in cities, thereby re-ducing quality of life. These trends reveal the increasing vulnerability of cities in southern Europe and western Turkey to climate-induced extreme temperatures and the UHI effect. In this context, the UHI, one of the most critical consequences of the climate crisis, seriously threatens the liveability and sustainable development capacity of cities worldwide, particularly in the Mediterranean Basin [5] Managing this threat requires not only technical interventions but also the adoption of heat-resilient urban planning and design approaches that comprehensively integrate mitigation, adapta-tion, and governance strategies [6,7].

The global impacts of climate change and the environmental pressures arising from rapid urbanisation processes are among the primary interlinked dynamics today [8].In particular, increasing energy demand, intensive construction trends, and the covering of natural surfaces with impermeable materials are intensifying both the UHI effect and the frequency of extreme heat events [9].Consequently, the global impacts of climate change and the local consequences of urban growth are intertwined, creating a critical area of vulnerability in terms of ecological sustainability [10]. (Lines 31-55).

The role of Remote Sensing (RS) and Geographic Information Systems (GIS) is sig-nificant in the examination and modelling of LST and LULC changes. In particular, the integration of LST and LULC data covering different periods from the 1980s to the present day allows for a detailed understanding of the ecological impacts of urbanisa-tion processes [25].When UA and GIS-based studies are integrated with spatial corre-lation, regression, and temporal change models, they provide a scientific basis for de-termining UHI dynamics and developing nature-based solutions to these effects [26]. Furthermore, the fact that these methods encompass not only current trends but also future scenarios supports the development of resilient and sustainable planning strat-egies against thermal stresses arising from rapid urbanisation and climate change [27].

UA technologies enable the long-term monitoring of LST and LULC classes by uti-lising multi-temporal data obtained from various satellite sensors such as Landsat, Sentinel, and MODIS (Moderate Resolution Imaging Spectroradiometer). Geographic Information Systems (GIS) provide a complementary infrastructure for analysing, classifying, and interpreting these data through spatial outputs [28]. Landsat has been providing long-term, reliable data archives since 1972, offering multispectral optical images with a spatial resolution of 30 metres. Sentinel-2 provides high-resolution data for LULC classifications and vegetation indices, while Sentinel-3 has been used for large-scale LST studies since 2016 with its 1 km SLSTR thermal measurements. Fur-thermore, Sentinel-3 stands out as a more suitable data source for monitoring regional heat waves, assessing land–sea interactions, and examining large-scale climatic tem-perature trends such as those in the Mediterranean Basin [29,30].

MODIS sensors are located on the Terra (1999) and Aqua (2002) satellites and are used to monitor LST and LULC changes over large areas with a spatial resolution of 250 m–1 km and a daily revisit capability [31]. Thanks to their high temporal resolu-tion, they offer significant advantages in revealing short-term dynamics such as heat waves and seasonal land cover changes. Therefore, when Landsat's long-term high spatial resolution, Sentinel's current and medium-high resolution data, and MODIS's large-scale temporal observations are used together, it is possible to comprehensively analyse LST–LULC relationships both spatially and temporally and to more accurately predict pressures caused by climate change and urbanisation [32,33].

Scientific studies reveal that different satellite systems are used both inde-pendently and in an integrated manner in the analysis of LST changes in urban areas. For example, Zhan et al. [34] examined the spatial differentiation of LST distribution by using MODIS and Landsat data together, demonstrating that MODIS's high tem-poral resolution is effective in capturing short-term temperature changes, while Landsat's high spatial resolution is effective in capturing micro-scale urban differ-ences. Peng et al. [35] analysed urban heat island (UHI) dynamics by integrating MODIS and Landsat data, noting that increasing urbanisation significantly raised LST values, particularly during summer months. Similarly, Odindi et al. [36] used Landsat and MODIS data in long-term (1984–2018) comparative analyses and showed that LST trends increased in parallel with the rate of urbanisation, but that green infrastructure applications partially offset this increase. Sentinel-3 data have also been used in the literature in different contexts. Jiménez-Muñoz et al. [37] conducted a comparative analysis of Sentinel-3, MODIS, and Landsat data and emphasised that LST data at dif-ferent resolutions are complementary. (Line 78-122)].

Commens 3: [Although the formulas for calculating LST, NDVI, and NDBI are presented, there is no validation using ground-based observations or comparisons with existing studies.]

Response 3: [Although ground-based measurements from earlier periods are not available, the Landsat-derived LST maps at 30 m resolution were indirectly validated through consistency with existing field studies. The cool islands detected in the LST outputs were found to correspond to green spaces such as Başçeşme Cemetery, Atatürk Park, and Avlu, whereas the hot spots coincided with industrial centers, major transport hubs, and newly developed high-rise residential zones. Furthermore, an assessment of the historical urban fabric indicated a comparable spatial pattern, with cool islands aligning with parks and cemeteries, and hot spots associated with industrial and transportation areas. Collectively, these results substantiate the reliability of the 30 m resolution Landsat-based LST datasets employed in this study.]

Commens 4: [The analysis relies solely on Pearson correlation, and the conclusions overemphasize “strong correlations.” Nonlinear relationships, multiple regression, or partial correlation should be considered to avoid overstating the results.]

Response 4: [Following the reviewer’s observation, the correlation analyses were further refined through the application of partial correlation. Within this framework, p-values were rigorously calculated and confidence intervals established, thereby demonstrating that the findings are statistically significant and exhibit a high degree of reliability.]

Commens 5: [Correlation coefficients are reported without p-values or significance levels.]

Response 5: [Within the methodological framework of the study, p-values were calculated and confidence intervals established, thereby demonstrating that the results are both statistically significant and robust.]

Commens 6: [The study does not provide concrete and applicable planning or management recommendations.]

Response 6: [As a result of the reviewer’s comments, the planning recommendations presented in our study have been elaborated in greater detail. The application/planning map produced from the 2025 LST data at 30 m resolution highlights strategic interventions aimed at mitigating the urban heat island effect. In this context, directly applicable spatial planning strategies have been developed, including the establishment of cooling corridors, the design of 250 m-wide green belts, the promotion of green roofs and vertical greening, the use of high-albedo and permeable surfaces, the integration of water features, and the redevelopment of heat-intensive transport hubs as city squares. In particular, the 250 m green belt proposal is applicable around industrial areas where there are currently no residential structures, such as vacant lots, partial agricultural lands, and bare open spaces surrounding newly developed high-rise buildings. Thus, the study not only provides analytical results but also delivers directly applicable outputs that can be utilized at the scale of spatial planning and urban management.]

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript presents a study on the relationship between LST, NDVI and NDBI in a city in Turkey. The article has shortcomings that need to be addressed.

1. The introduction seems like a catalogue of previous studies without any critical evaluation of previous research and justification for the manuscript. Which gap is the manuscript filling? What is the contribution of the manuscript?
2. The methodology includes mainly basic analysis. The study of LST has gone beyond the analysis of the relationship between LST and NDBI and NDVI. This is known already. The methods can be improved by examining the relationship during different seasons, adding other variables such as building heights, nighttime light data, using machine learning, or using Google Earth Engine to analyze several images and not just five epochs. 
3. The use of English needs improvement. Moreover, the similarity is 21%. Journal article should not have similarity more than 15%, even some journals accept only similarity that is less than 10%. 

Comments on the Quality of English Language

The English could be improved to more clearly express the research.

Author Response

Comments 1: [The introduction seems like a catalogue of previous studies without any critical evaluation of previous research and justification for the manuscript. Which gap is the manuscript filling? What is the contribution of the manuscript?]

Response 1: [This study, based on a continuous 40-year time series derived from Landsat (30 m) data for the period 1985–2025, revealed in detail the micro-scale heterogeneity of urban surface temperatures. In addition, MODIS (1 km) data provided a comparative perspective on macro-scale trends for the post-2005 period, while Sentinel-3 (1 km) data offered such a comparison only for the year 2025. Unlike most urban heat island studies in the literature, which are limited to a single satellite or short-term analyses, this study systematically compared micro (30 m) and macro (1 km) scales within the same warm-season windows, thereby developing both a long-term and multi-scale perspective. Furthermore, the increasingly strong positive correlation between LST and built-up index (NDBI), and the intensifying negative correlation between LST and vegetation index (NDVI), were statistically validated through Pearson and partial correlation analyses, thereby isolating the reinforcing role of urbanization on heat island intensity and the cooling effect of vegetation.

Building on these findings, a detailed planning output was developed for 2025 based on high-resolution Landsat LST trends, providing concrete recommendations for strengthening cooling corridors, prioritizing interventions in heat-intensive zones, and advancing green infrastructure strategies. Unlike previous studies, this approach not only described the spatial distribution of surface temperatures but also generated actionable strategies that can be directly integrated into urban planning processes. Moreover, it clearly demonstrated through comparative analysis the insufficiency of MODIS and Sentinel-3 data in capturing cool islands and micro-climatic variations at the urban texture scale, while highlighting the success of Landsat (30 m) data in accurately identifying urban hot spots and green cool islands, thus contributing a methodological framework on product–scale compatibility.]

Comments 2: [The methodology includes mainly basic analysis. The study of LST has gone beyond the analysis of the relationship between LST and NDBI and NDVI. This is known already. The methods can be improved by examining the relationship during different seasons, adding other variables such as building heights, nighttime light data, using machine learning, or using Google Earth Engine to analyze several images and not just five epochs. ]

Response 2: [As a result of the reviewer’s suggestion, the study has been further developed. Based on a 40-year continuous time series derived from Landsat (30 m) data covering the 1985–2025 period, the research has revealed in detail the micro-scale heterogeneity of urban surface temperature. In addition, MODIS (1 km) data provided comparative insights at the macro scale for the post-2005 period, while Sentinel-3 (1 km) data contributed only for the year 2025. Unlike most previous urban heat island studies that rely on a single satellite or short-term analyses, this study systematically compared micro (30 m) and macro (1 km) scales within the same warm-season windows, thereby establishing both a long-term and multi-scale perspective. Furthermore, the strengthening positive correlation between LST and built-up areas (NDBI), and the increasing negative correlation between LST and vegetation (NDVI), were statistically validated through both Pearson and partial correlation analyses, thus isolating the independent role of urbanization in intensifying the heat island effect and the cooling function of vegetation.

Building upon these findings, a high-resolution planning map for 2025 was developed based on Landsat LST trends, through which concrete recommendations were made, including the strengthening of cooling corridors, prioritized interventions in heat-intensive areas, and the implementation of green infrastructure strategies. Unlike previous studies that were limited to describing the spatial distribution of surface temperatures, this approach delivers actionable strategies that can be directly integrated into urban planning processes. Moreover, the study clearly demonstrated, through comparative analyses, the insufficiency of MODIS and Sentinel-3 data in capturing cool islands and micro-climatic differences at the urban fabric scale, while highlighting the superior capability of Landsat (30 m) in delineating urban hot spots and green cool islands with high accuracy, thereby developing a methodological framework for product–scale compatibility.]

Comments 3: [The use of English needs improvement. Moreover, the similarity is 21%. Journal article should not have similarity more than 15%, even some journals accept only similarity that is less than 10%]. 

Response 3: [The similarity ratio has been reduced to 15% again.]

Reviewer 3 Report

Comments and Suggestions for Authors

The research was interesting, with a significant approach to data analysis by remotely sensing data. However, some improvements are needed to improve the impact of the research.

In terms of the introduction, I suggest the authors should integrate references to the IPCC AR6, which highlights the increased vulnerability of Mediterranean regions, such as western Turkey, to extreme heat and the urban heat island effect. This would provide an important climate science context and align the study with global research priorities.

Also, the paper shows limited engagement with the main literature on climate and remote sensing, citing only one or two pieces from this field. Consequently, the manuscript seems repetitive regarding the relationships between NDVI and LST and between NDBI and LST. For instance, the inverse NDVI–LST relationship is mentioned in the abstract and reiterated in the introduction and methods sections without further developing the argument.

Expand the data context. For example, acknowledge the study's Landsat strength, but add a paragraph on Sentinel-3 SLSTR (daily and thermal) and MODIS LST (high temporal coverage) as complementary sources for heat wave and seasonal dynamics.

For the Materials and Methods, although Landsat provides excellent data for long-term analysis, it would be good to acknowledge the Sentinel-3 SLSTR data as an alternative or complementary source for land surface temperature (LST). Sentinel-3 provides a higher temporal resolution (daily), which could improve future analyses, especially in identifying seasonal anomalies or heat wave events.

For the results. The authors might consider adding a time-series anomaly chart (for example, LST anomalies relative to the 1985–2005 baseline), which would highlight the long-term warming trend.

The authors could also compare and discuss further contexts, such as the magnitude of LST increases (from 41 °C in 1985 to 52 °C in 2025), alongside broader regional warming trends reported by AR6 for southern Europe and the Mediterranean.

The authors could also briefly discuss the potential for integrating adaptation strategies in the context of the IPCC AR6 Working Group II. Linking the local Balıkesir case to these global assessments would greatly enhance its policy relevance.

 

 

 

Author Response

Comments 1: [The research was interesting, with a significant approach to data analysis by remotely sensing data. However, some improvements are needed to improve the impact of the research.]

Response 1: [In light of the reviewer’s valuable comments, both the methodology and the scope of the study have been further expanded.

This study, based on a continuous 40-year time series derived from Landsat (30 m) data for the period 1985–2025, revealed in detail the micro-scale heterogeneity of urban surface temperatures. In addition, MODIS (1 km) data provided a comparative perspective on macro-scale trends for the post-2005 period, while Sentinel-3 (1 km) data offered such a comparison only for the year 2025. Unlike most urban heat island studies in the literature, which are limited to a single satellite or short-term analyses, this study systematically compared micro (30 m) and macro (1 km) scales within the same warm-season windows, thereby developing both a long-term and multi-scale perspective. Furthermore, the increasingly strong positive correlation between LST and built-up index (NDBI), and the intensifying negative correlation between LST and vegetation index (NDVI), were statistically validated through Pearson and partial correlation analyses, thereby isolating the reinforcing role of urbanization on heat island intensity and the cooling effect of vegetation.

Building on these findings, a detailed planning output was developed for 2025 based on high-resolution Landsat LST trends, providing concrete recommendations for strengthening cooling corridors, prioritizing interventions in heat-intensive zones, and advancing green infrastructure strategies. Unlike previous studies, this approach not only described the spatial distribution of surface temperatures but also generated actionable strategies that can be directly integrated into urban planning processes. Moreover, it clearly demonstrated through comparative analysis the insufficiency of MODIS and Sentinel-3 data in capturing cool islands and micro-climatic variations at the urban texture scale, while highlighting the success of Landsat (30 m) data in accurately identifying urban hot spots and green cool islands, thus contributing a methodological framework on product–scale compatibility.

The issue raised in Reviewer Comment 1 has already been elaborated in detail in the manuscript, specifically between lines 158 and 189. ]

Comments 2: [ In terms of the introduction, I suggest the authors should integrate references to the IPCC AR6, which highlights the increased vulnerability of Mediterranean regions, such as western Turkey, to extreme heat and the urban heat island effect. This would provide an important climate science context and align the study with global research priorities. Also, the paper shows limited engagement with the main literature on climate .]

Response 2: [We sincerely thank the reviewer for this valuable suggestion. In response, the Introduction section has been expanded to incorporate references to the IPCC AR6, which emphasizes the heightened vulnerability of Mediterranean regions, including western Turkey, to extreme heat and the urban heat island effect. This integration provides a stronger climate science context and aligns the study with global research priorities. Furthermore, to address the reviewer’s observation regarding limited engagement with the broader climate literature, additional relevant references have been incorporated into the Introduction (lines 31–55), thereby strengthening the theoretical and contextual foundation of the study.

The temperature increases caused by climate change, extreme weather events, and the urban heat island (UHI) effect are creating increasing pressures on the ecological balance and social living conditions of cities [1-3]. The Mediterranean Basin is one of the regions most intensely affected by climate change and stands out as a critical ‘hot spot’ due to the increase in extreme heat events [4]. The Intergovernmental Panel on Climate Change (IPCC) Sixth Assessment Report (AR6) reveals that temperature in-creases in the Mediterranean region are above the global average, with significant in-creases expected in the frequency, duration, and intensity of heatwaves during the summer months.

In addition, according to AR6, the UHI effect and increasing humidity deficit in Mediterranean cities are exacerbating thermal stress conditions in cities, thereby re-ducing quality of life. These trends reveal the increasing vulnerability of cities in southern Europe and western Turkey to climate-induced extreme temperatures and the UHI effect. In this context, the UHI, one of the most critical consequences of the climate crisis, seriously threatens the liveability and sustainable development capacity of cities worldwide, particularly in the Mediterranean Basin [5] Managing this threat requires not only technical interventions but also the adoption of heat-resilient urban planning and design approaches that comprehensively integrate mitigation, adapta-tion, and governance strategies [6,7].

The global impacts of climate change and the environmental pressures arising from rapid urbanisation processes are among the primary interlinked dynamics today [8].In particular, increasing energy demand, intensive construction trends, and the covering of natural surfaces with impermeable materials are intensifying both the UHI effect and the frequency of extreme heat events [9].Consequently, the global impacts of climate change and the local consequences of urban growth are intertwined, creating a critical area of vulnerability in terms of ecological sustainability [10].]

Response 3: [In line with the reviewers’ suggestions, the role of remote sensing and geographic information systems has been elaborated in greater detail within the introduction section of the manuscript (Lines 78-99).

The role of Remote Sensing (RS) and Geographic Information Systems (GIS) is sig-nificant in the examination and modelling of LST and LULC changes. In particular, the integration of LST and LULC data covering different periods from the 1980s to the present day allows for a detailed understanding of the ecological impacts of urbanisa-tion processes [25].When UA and GIS-based studies are integrated with spatial corre-lation, regression, and temporal change models, they provide a scientific basis for de-termining UHI dynamics and developing nature-based solutions to these effects [26]. Furthermore, the fact that these methods encompass not only current trends but also future scenarios supports the development of resilient and sustainable planning strat-egies against thermal stresses arising from rapid urbanisation and climate change [27].

UA technologies enable the long-term monitoring of LST and LULC classes by uti-lising multi-temporal data obtained from various satellite sensors such as Landsat, Sentinel, and MODIS (Moderate Resolution Imaging Spectroradiometer). Geographic Information Systems (GIS) provide a complementary infrastructure for analysing, classifying, and interpreting these data through spatial outputs [28]. Landsat has been providing long-term, reliable data archives since 1972, offering multispectral optical images with a spatial resolution of 30 metres. Sentinel-2 provides high-resolution data for LULC classifications and vegetation indices, while Sentinel-3 has been used for large-scale LST studies since 2016 with its 1 km SLSTR thermal measurements. Fur-thermore, Sentinel-3 stands out as a more suitable data source for monitoring regional heat waves, assessing land–sea interactions, and examining large-scale climatic tem-perature trends such as those in the Mediterranean Basin [29,30].]

Comments 4: [Consequently, the manuscript seems repetitive regarding the relationships between NDVI and LST and between NDBI and LST. For instance, the inverse NDVI–LST relationship is mentioned in the abstract and reiterated in the introduction and methods sections without further developing the argument.]

Response 4: [The necessary adjustments have been made in the abstract section of the article.]

Response 5: [While the initial version of this study was conducted solely using Landsat datasets, MODIS and Sentinel-3 products were subsequently incorporated. Accordingly, these additions have been explicitly addressed in the introduction, as well as in the methodology and results sections of the revised manuscript.

INTODUCTION

UA technologies enable the long-term monitoring of LST and LULC classes by uti-lising multi-temporal data obtained from various satellite sensors such as Landsat, Sentinel, and MODIS (Moderate Resolution Imaging Spectroradiometer). Geographic Information Systems (GIS) provide a complementary infrastructure for analysing, classifying, and interpreting these data through spatial outputs [28]. Landsat has been providing long-term, reliable data archives since 1972, offering multispectral optical images with a spatial resolution of 30 metres. Sentinel-2 provides high-resolution data for LULC classifications and vegetation indices, while Sentinel-3 has been used for large-scale LST studies since 2016 with its 1 km SLSTR thermal measurements. Fur-thermore, Sentinel-3 stands out as a more suitable data source for monitoring regional heat waves, assessing land–sea interactions, and examining large-scale climatic tem-perature trends such as those in the Mediterranean Basin [29,30].

MODIS sensors are located on the Terra (1999) and Aqua (2002) satellites and are used to monitor LST and LULC changes over large areas with a spatial resolution of 250 m–1 km and a daily revisit capability [31]. Thanks to their high temporal resolu-tion, they offer significant advantages in revealing short-term dynamics such as heat waves and seasonal land cover changes. Therefore, when Landsat's long-term high spatial resolution, Sentinel's current and medium-high resolution data, and MODIS's large-scale temporal observations are used together, it is possible to comprehensively analyse LST–LULC relationships both spatially and temporally and to more accurately predict pressures caused by climate change and urbanisation [32,33].

Scientific studies reveal that different satellite systems are used both inde-pendently and in an integrated manner in the analysis of LST changes in urban areas. For example, Zhan et al. [34] examined the spatial differentiation of LST distribution by using MODIS and Landsat data together, demonstrating that MODIS's high tem-poral resolution is effective in capturing short-term temperature changes, while Landsat's high spatial resolution is effective in capturing micro-scale urban differ-ences. Peng et al. [35] analysed urban heat island (UHI) dynamics by integrating MODIS and Landsat data, noting that increasing urbanisation significantly raised LST values, particularly during summer months. Similarly, Odindi et al. [36] used Landsat and MODIS data in long-term (1984–2018) comparative analyses and showed that LST trends increased in parallel with the rate of urbanisation, but that green infrastructure applications partially offset this increase. Sentinel-3 data have also been used in the literature in different contexts. Jiménez-Muñoz et al. [37] conducted a comparative analysis of Sentinel-3, MODIS, and Landsat data and emphasised that LST data at dif-ferent resolutions are complementary (Lines 88-122).]

Comments 6: [For the Materials and Methods, although Landsat provides excellent data for long-term analysis, it would be good to acknowledge the Sentinel-3 SLSTR data as an alternative or complementary source for land surface temperature (LST). Sentinel-3 provides a higher temporal resolution (daily), which could improve future analyses, especially in identifying seasonal anomalies or heat wave events.]

Response 6: [Thank you for the suggestion regarding the use of Sentinel-3 SLSTR data as a complementary source. In line with the reviewer’s suggestion, MODIS (2005, 2015, 2025) and Sentinel-3 (2025) datasets were incorporated as complementary sources; however, the 1 km spatial resolution prevented the results from being consistent at the urban scale. Moreover, the comparative analyses clearly demonstrated the insufficiency of MODIS and Sentinel-3 data in capturing cool islands and micro-climatic variations at the urban fabric scale. In contrast, Landsat (30 m) data proved to be highly effective in accurately delineating urban hot spots and green cool islands, thereby establishing a methodological framework for product–scale compatibility.

The issue raised in Reviewer Comment 6 has already been elaborated in detail in the manuscript, specifically between lines 158 and 189.]

Comments 7: [For the results. The authors might consider adding a time-series anomaly chart (for example, LST anomalies relative to the 1985–2005 baseline), which would highlight the long-term warming trend.]

Response 7: [As a result of the referee evaluation, the ‘Long-Term Trends of LST: Temporal Change Graphs’ for the period 1985-2025 are presented in Figure 6.]

Reviewer 4 Report

Comments and Suggestions for Authors

Major Reviews

1. The flow chart in Figure 2 should be revised so that readers can fully understand the detailed procedure by looking at the figure alone. Specifically, it is recommended that the figure include more concrete details corresponding to the descriptions provided in the main text.
2. In the current manuscript, the map legends for LST (Figure 3), NDVI (Figure 4), and NDBI (Figure 5) are presented with varying interval ranges for each year. This approach makes it difficult for readers to intuitively and consistently compare changes across years. I therefore recommend applying a fixed and common classification scheme, based on the overall minimum and maximum values of LST, NDVI, and NDBI observed during the study period. In other words, the same interval ranges and color scales should be used for all years. This would allow readers to more clearly and coherently assess the magnitude and spatial patterns of these changes. In addition, it is suggested that Table 2 be revised accordingly, using the same consistent classification method. Based on these changes and numerical adjustments, corresponding revisions in the main text are also necessary.

 

Minor Reviews

1. Line 35: Please revise “This in turn. often~” to “This, in turn, often~”.
2. Lines 98–99: Please revise “The hill which has an elevation of 181-275 metres above sea level. was included ~” to “The hill, which has an elevation of 181-275 metres above sea level, was included ~”.
3. Line 185: As TB has been defined as brightness temperature in line 144, please revise “Blackbody temperature” to “Brightness temperature”.
4. The commas in the numerical values in Figures 3, 4, and 5 should be corrected to periods (decimal points).
5. Please express the unit of LST in Figure 3 in degrees Celsius (°C).
6. Please remove the “Direction” column in Table 3. Since the correlation values are already indicated with + and − signs, this information can be inferred without the additional column.

Author Response

Comments 1: [The flow chart in Figure 2 should be revised so that readers can fully understand the detailed procedure by looking at the figure alone. Specifically, it is recommended that the figure include more concrete details corresponding to the descriptions provided in the main text.]

Response 1: [As suggested by the reviewer, the methodological flowchart has been substantially revised and elaborated to provide a more detailed representation of the research process (Figure 2).]

Comments 2: [ In the current manuscript, the map legends for LST (Figure 3), NDVI (Figure 4), and NDBI (Figure 5) are presented with varying interval ranges for each year. This approach makes it difficult for readers to intuitively and consistently compare changes across years. I therefore recommend applying a fixed and common classification scheme, based on the overall minimum and maximum values of LST, NDVI, and NDBI observed during the study period. In other words, the same interval ranges and color scales should be used for all years. This would allow readers to more clearly and coherently assess the magnitude and spatial patterns of these changes. In addition, it is suggested that Table 2 be revised accordingly, using the same consistent classification method. Based on these changes and numerical adjustments, corresponding revisions in the main text are also necessary.]

Response 2: [In the present study, the thematic maps of LST (Figure 3), NDVI (Figure 4), and NDBI (Figure 5) were generated using different classification ranges for each year. In line with the reviewer’s suggestion, an initial attempt was made to apply a fixed classification scheme by considering the overall minimum and maximum values obtained from all Landsat 30 m resolution images throughout the study period. However, this approach proved insufficient in capturing the micro-climatic variations that emerged annually in urban areas, particularly in reflecting the spatial patterns of cool (green) and hot (red) islands. The use of fixed intervals largely masked the inter-annual thermal and structural differences and restricted the realistic representation of spatial patterns.

Therefore, in order to achieve more precise and representative results, dynamic classification ranges were established based on the minimum and maximum values specific to each year. This method, by taking into account the annual temperature distributions, allowed for a more accurate delineation of urban heat island (UHI) cool and hot zones. Following the same principle, NDVI and NDBI classifications were also reorganized according to their annual value ranges, thereby enabling a more reliable and comparable assessment of changes in vegetation density and built-up intensity across years. Examination of the thematic maps clearly indicates that the spatial distribution of cool (green) and hot (red) urban islands has been meaningfully and distinctly reflected over time. This finding demonstrates that the year-specific classification approach is effective in revealing urban micro-climatic variations.]

Comments 3: [Minor Reviews]

Response 3: [All minor revisions indicated have been revised.]

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Authors have improved the quality of this paper significantly.

Author Response

Comments 1: [Authors have improved the quality of this paper significantly]

Response 1:[We sincerely thank you for your valuable comments.]

[When the study was examined in detail, the acronyms appearing in lines 78 and 88 were corrected.For example, the section referred to as AU (UA) has been corrected to RS. Additionally, the inverted English sentences in lines 45–48, 204–209, 451–458, and 636–640 have been revised to improve clarity and fluency. Furthermore, all minor typographical and punctuation errors throughout the manuscript have been carefully reviewed and corrected.]

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have corrected the manuscript. However, they need to check and correct minor error and improve language style. For example, what is the meaning of the acronym "AU"?

Comments on the Quality of English Language

The English could be improved to more clearly express the research.

Author Response

Comments 1: [The authors have corrected the manuscript. However, they need to check and correct minor error and improve language style. For example, what is the meaning of the acronym "AU"?]

Response 1: [In accordance with the reviewer’s comment, minor errors in the manuscript have been corrected and the overall language style has been improved. In addition, the meaning of the acronym “AU” has been clearly defined in the revised version of the manuscript] Thank you for pointing this out. We agree with this comment. Therefore, we have carefully revised the related sections and made the necessary corrections to improve the manuscrip.[In accordance with the comment, the abbreviations mentioned in lines 78 and 88 have been corrected. For example, the section referred to as AU (UA) has been corrected to RS. Additionally, the inverted English sentences in lines 45–48, 204–209, 451–458, and 636–640 have been revised to improve clarity and fluency. Furthermore, all minor typographical and punctuation errors throughout the manuscript have been carefully reviewed and corrected.]

Author Response File: Author Response.docx

Reviewer 4 Report

Comments and Suggestions for Authors

The revised manuscript has faithfully reflected the earlier review comments and has been substantially improved. The authors have made appropriate efforts to revise the methodological flow chart, enhance the clarity of figures, and address the technical concerns raised. As a result, the manuscript is clearer, more detailed, and scientifically sound. I believe the paper has now reached a publishable level.

Author Response

Comments 1: [The revised manuscript has faithfully reflected the earlier review comments and has been substantially improved. The authors have made appropriate efforts to revise the methodological flow chart, enhance the clarity of figures, and address the technical concerns raised. As a result, the manuscript is clearer, more detailed, and scientifically sound. I believe the paper has now reached a publishable level.]

Response 1: [We sincerely thank you for your valuable comments]

[When the study was examined in detail, the acronyms appearing in lines 78 and 88 were corrected.For example, the section referred to as AU (UA) has been corrected to RS. Additionally, the inverted English sentences in lines 45–48, 204–209, 451–458, and 636–640 have been revised to improve clarity and fluency. Furthermore, all minor typographical and punctuation errors throughout the manuscript have been carefully reviewed and corrected.]

Author Response File: Author Response.docx