Characterizing Spatiotemporal Variations in the Urban Thermal Environment Related to Land Cover Changes in Karachi, Pakistan, from 2000 to 2020

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors：

Review Report on "Characterizing spatiotemporal variations in the urban thermal environment related to land-cover changes in Karachi, Pakistan from 2000 to 2020"

Overall description and shortcoming

The study explores the LUCC that took place in Karachi, Pakistan in recent decades, and connects with UHI effects, as well as SUHII and other spatial transformation. The idea is good, and the flow of the entire manuscript is smooth and easy to follow, especially the application of remote sensing datasets and technologies in less developed cities, for land use retrieval and LST estimation etc..

Nevertheless, the research methodologies have to be explained in a clearer manner, and proper citations and acknowledgments should be included for different parts and explanations provided in the manuscript, for example recent land use retrieval work conducted in Pakistan or other developing countries.

Further, the categorization of land use types is not correct, with reference to many related literature, there should be 5 land use types (instead of 4) after handling and processing remotely sensed imageries. In this research, the authors have missed out "agricultural land / areas", which is very important, and is different from "greenery areas" or "bare land", thus the authors should repeat the study again based on the proposed approach / mechanism, then update the corresponding LUCC results. Definitely, the statistical correlation with temperature, SUHII and other environmental quantities should also be updated accordingly, and the statistical figures within different tables should be revised too.

Some specific problems and proposed grammatical changes are stated as follows:

Major Problems

Lines 46-73 (Introduction): Please refer to the Introduction of
https://www.mdpi.com/2072-4292/13/16/3337/htm  

https://springerplus.springeropen.com/articles/10.1186/s40064-016-2414-z
for (i) devastating environmental impacts; (ii) the use of remote sensing techniques, machine learning approaches and geospatial techniques in investigating LUCC within Pakistan

Line 58: What do you mean by local UHI effect? To what spatial level, say street level, or by region, or by village etc.?

Lines 63-66: It's better to provide a summary of previous studies, especially within less developed cities / developing countries, as well as the spatial techniques applied.

Lines 83-84: What is the main difference of your study when compared to Rizvi et al.? It would be better to point out the research gaps at the end of this paragraph, thus laying down the foundation of your current manuscript.

Lines 90-95: Again, it would be better to add in some "technical methods" or "approaches" of handling satellite imageries / data within the Introduction, otherwise the research gap(s) may not be effectively revealed.

Lines 126-127: What kind of atmospheric and seasonal effects could take place? Also, how much could it affect the eventual satellite retrieval result. Provision of these information can highlight the importance of your selection criteria.

Line 132: The authors should take reference of the following articles, which have made use of SVM for land use land cover retrieval as well - please also include some more details of the kind of SVM used in current study.

https://www.sciencedirect.com/science/article/abs/pii/S0924271606001122

https://www.mdpi.com/2072-4292/13/16/3337/htm

https://www.sciencedirect.com/science/article/abs/pii/S2352938520301075?via%3Dihub

Section 3.1 (in overall and the rest of this manuscript): How about agricultural land / areas? They are different from vegetation - the current study should divide into 5 different land use types, and the corresponding results of the whole manuscript should be updated and revised as well. The 5 land use types include built-up areas, bare land, waterbodies, vegetation, and agricultural areas / land.

Lines 140-141: Can you provide some details of these ancillary land cover data? It will provide more information to your study.

Lines 163-165: How many pixels have you considered here to get the "averaged LSTs" of built up area and green space respectively? How many days have you considered as well?

Line 180: How about using Monte Carlo approach here? It will be more scientific to select these random samples.

Line 190: "distance decay function" - do you mean inverse-distance weighting approach? If so, please cite relevant usage of this approach in related studies, for example for estimating pollutant emission etc.

Lines 200-206: The LULC categorization here should include agricultural land as well, 5 categories of land use types should be included. The description of results and statistical analysis in Table 3 should be updated accordingly.

Lines 255-258: Are there any relevant studies that "proves" this phenomenon in a scientific manner? If so, please include proper citations as well.

Lines 273-279: Some numerical values / comparisons should be included in the descriptions of Figure 5.

Figure 6: The authors should also include a box plot for "agricultural areas/lands" after re-performing the satellite retrieval and classification.

Lines 293-310: The revised manuscript should include the discussion of 5 land-cover or land-use categories

Lines 372-374: How could the uncertainty of NDVI estimates be remedied in future, especially using more advanced remote sensing technologies?

Lines 383-387: The observation was actually observed in Hyderabad, Pakistan, as verified and validated in https://www.mdpi.com/2072-4292/13/16/3337/htm 

Is this a common observation within the entire Pakistan? Or just these several cities of Pakistan? The authors can include a bit of discussion here.

Line 401: Are there any environmental factors that cause such large-scale conversion? The authors should also link with the local policies with Karachi, Pakistan within recent decades - this may also be connected with implementation of environmental policies, urban planning strategies etc. All these are also important.

Lines 403-405: Have you compared the statistical performance of GWR model with other existing models based on previous or recent studies? In particular, whether the combination of remote sensing strategies, image processing techniques and machine learning approaches will work better? Any automatic correction will be conducted by your GWR model?

Lines 402-417: Has modeling work been conducted to investigate the thermal conditions within Karachi, Pakistan? If so, it's good to include the results and findings of these literature in environmental modeling.

Lines 418-422: The authors have pointed out an important point here, but how could the influence of spatial patterns of LUCC towards SUHII been monitored or traced? Any technical approaches can be suggested?

Section 5 (Discussion): The authors also pointed out at the Introduction of this manuscript - that data availability (i.e., lack of open data) is a big issue in Pakistan and other developing cities, the authors should mention something about "open data initiatives that enhance sustainability of Karachi or other cities of Pakistan". Data openness of land use, air quality and environmental data should be highlighted, either in the Discussion, or in the Conclusion of this manuscript. This addresses the authors' concerns and comments in the Introduction.

Lines 428-432: The authors should also mention something about "agricultural land" changes within the recent decades after the study has been repeated, where "agricultural land" is included as one of the land use types.

Line 439: It is still a bit unclear why GWR technique is preferred in related studies, when compared with other available statistical algorithms. Perhaps the authors can include some discussion of this in the Methodology section.

Lines 443-445: Based on what the authors have concluded here, what can be done in those green spaces? Any political / environmental suggestions can be outlined at the end of this manuscript?

Minor Problems / Further clarifications needed

Line 22: Climate warming is another issue and can be caused by many factors, so the choice of word here is somehow inaccurate. Line 60 has the same problem as well.

Lines 23-24: Which sustainable development goals based on United Nation standard?

Line 35: About correlation statistics - some statistical figures in correlation will be needed here.

Lines 35-36: About the importance of NDBI - how much in average? It will be better to state the number here in Abstract.

Line 40: What do you mean / define as a "meaningful relationship"? In terms of statistics?

Line 99: It's a bit unclear, need some further details of "statistical analysis".

Line 126: What's the percentage of cloud cover here?

Table 1: Path / Row is somehow consistent among all years, thus there is not much need to open a column for this. The details can be provided at the footnote of Table 1. Meanwhile, some formatting problem appears in Table 1 - caption should be placed on top of the table. Further, the format of date in Table 1 is not presented in a reader-friendly manner, please use standard way of presentation.

Line 142: "previous" - what do you mean here? It sounds a bit vague here

Lines 159-160: How would you distinguish / classify urban areas and rural areas usually, say in literature?

Lines 192-193: Can perhaps provide some more details on how to adjust the regression coefficients based on "locations".

Line 223: any proper reference in city level regarding the "high demand for residential and business activities", say within Karachi?

Line 239: Are there any proper official documents from Karachi that emphasized "construction is prohibited in these regions"?

Figure 3: The resolution of the legend of Figure 3 on Page 9 should be updated, so that readers can visualize the changes easily when referring to the figures directly.

Lines 244-245: The authors need to restate the 5 classes of NLST here in the text.

Table 5: The number of decimal places / significant figures within data of Table 5 should be consistent, currently, some has only 1 decimal place, while most have 2.

Lines 278-279: What do you mean in Lines 278-279? How can these factors "produce" SUHI that are more intense? Need to rephrase the sentence too.

Line 297: It's a bit unclear - what's the "background of climate change"?

Lines 305-306: should include the label of the "river" in the map of this study, back in Section 2.

Figure 7 labelling: should include the label of Kiamari and Malir in all those figures on the first 4 panels of Figure 7, for clear visualization.

Figure 8(a) and (b) can be put in Appendix, as the implications and results are basically the same as that inferred from previous figures.

Lines 363-364: How strong is the positive correlation in average?

Grammatical changes / English errors

Line 22: help to alleviate

Line 26: due to the lack of consistent data

Line 28: Land-cover maps were "retrieved"?

Line 46: Humans have witnessed a significant increase in urbanization across the globe in the last several decades.

Line 60: mitigating

Line 151: are comparable

Line 159: calculate the temperature difference

Line 168: calculated by

Lines 171-172: some English problem here, please kindly rewrite the sentence again

Line 198: individual maps of land cover changes...

Line 199: expansion

Line 206: more dense 

Caption of Table 3: Accuracy

Lines 212-213: during the period of 2000-2020

Line 214: by nearly

Line 238: only been

Caption of Figure 4 (Line 260): map in... and 2020 respectively

Line 264: Variations "of"

Line 266: of SUHI

Line 268: SUHII growth

Line 273: of SUHII

Lines 276-278: too long, should break into two sentences

Line 284: Variations of LST

Line 299: cultivated and became built-up areas

Line 307: Due to the inclusion of mixed pixels,

Line 308: for example bare land.

Line 309: were observed in pixels that represent waterbodies.

Line 311: It is observed that...

Lines 312-313: For example,

Lines 314-317: in Figure 7, where the change from vegetation to open bare land has taken place, and the NLST has increased by up to 0.2, as comparable to between 0.3 to 0.4 for pixels that have become built-up class.

Line 324: shown in Figure 8

Line 326: NDVI values were

Line 347: reflects

Lines 347-348: the higher the vegetation density within a region, the lower the land surface temperature being detected.

Line 360: variations of 

Line 362: for different land-cover classes

Line 386: "the"

Line 387: being re-built or re-structured.

Line 392: proportion of areas where...

Lines 397-398: In contrast, our study has found out that...

Line 409: variations of NDVI

Lines 413-414: English error, please rewrite this sentence again.

Line 433: These changes have significantly altered the patterns of SUHII within Karachi, Pakistan.

Line 440: in the districts situated in the East and the West, including Kamari and Malir.

Lines 442-443: at regional scale, and fine-tune planning practice at local scales, with the aim of improving thermal conditions and promoting sustainable development of a community or less developed regions.

Lines 444-445: will be the most effective for mitigating UHI effects.

Finally, I believe a proper round of English editing will help improving the effectiveness of presentation of this manuscript as a whole. Further, it would be even better and nice to have some connections with other recent studies, thus better indicate the importance of land use retrieval through satellite means, and for promoting the future development of urban remote sensing.

Overall, I believe there are some scientific impacts within this manuscript, and it will really help developing countries a lot, however the authors should go through all aforementioned points and suggestions, and address these major / minor points one by one, so that the manuscript will look more professional, and of higher impact within the entire community. The data openness within developing nations can be promoted / enhanced for the purpose of sustainable development and enhancing thermal comfort etc. This could be mentioned in the revised manuscript too.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The topic of the paper CHARACTERIZING SPATIOTEMPORAL VARIATIONS IN THE URBAN THERMAL ENVIRONMENT RELATED TO LAND-COVER CHANGES IN KARACHI, PAKISTAN FROM 2000 TO 2020 is relevant, timely, and of interest to the audience of this journal. In this paper land-use land-cover changes and their impact on the urban thermal environment in a tropical megacity - Karachi, Pakistan was presented. Land-cover maps were prepared, and the land surface temperature (LST) was estimated using Landsat images from five different years in the period 2000 to 2020. The surface urban heat island intensity was then quantified based on the LST data. Statistical analysis, geographically weighted regression including correlation analysis, was performed to analyse the relationship between the land-cover composition and LST. These findings can inform adaptive land-use planning that aims to mitigate the effects of the UHI and aid efforts towards sustainable urban growth. This research area is great importance not only on for the Pakistan but in the many places of the world. The paper is clearly written and is very easy to read. The paper is based on academic standards. The abstract is concise and sufficient. The introduction provides the necessary background information. References are complete and appropriate. The manuscript would benefit from a broader treatment of the Conclusions section e.g. what is the unique contribution of this research? The manuscript should be published after a minor revision.

The following comments are suggested to be addressed in the revised manuscript:

Some of the References items:  5, 17, 27, 28, 46 and 51 has incomplete description or DOI identifiers are missing.

Invalid DOI identifier of the References item 30.

Incorrect description of the Table 1.

Error of column designations in the  Table 4.

Use the same number of decimal places in the Tables 4 and 5.

Lines 293-297 - The increases of the NLST shown in the Figure 6 are not observed until the 2020.

The Figure 7 is not readable enough.

Use the same number of decimal places in the charts presented in the Figure 9.

There is no description of 2020 in the charts presented in the Figure 9.

The Figure 10 is not readable enough.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Dear authors, the manuscript is well conceived and informative but it has some inadequacies.  I shall highlight them and the authors might improve the quality and readability of this research paper accordingly. 

1. The cross-cutting explanation of the methodology should be described by a schematic and conceptual flowchart.  The methodology should be described clarifying the definition of the variables and their modelling.   
2. Please include all the Software codes that have been implemented in this particular research work in the appendix section for independent simulation, testing, validation and integration.  
3. The introduction section does not provide a succinct theoretical basis for the study.  I would like to ask the authors to expand and highlight the advantages of their approaches that bring benefits in the solved issues.  
4. The English in several parts of the text is not good enough and does not allow the understanding of the authors' sentences.  Please have the manuscript proofread by competent authorities.  Please make it coherent dealing with each section at a time, deleting all the repetitions and linking one paragraph to the next having focus, scale, novelty, importance, clarity and rigour.  I suggest using Journal's assistance with proof reading services.  
5. If possible, it will be good if the authors could add a graphical representation summarizing their results which compares controls, simulation results, all the parameters and variables directly related to SUHI, SUHII and LST.  
6. Please, briefly add future perspectives and further applied applications of this specific research work in the discussion section before the conclusion.  It is not well interpreted to show how and why this work was undertaken relative to what is already known.  
7. The techniques and/or models presented and mentioned in the manuscript require sufficient details (including calibration, sensitivity analysis and validation) to allow other researchers to develop and test the applications later on.  Please include the parameters that I have mentioned here.    
8. Please, highlight the outliers in all the tables and graphs, where relevant.  
9. The most relevant data-results should be summarized and demonstrated by a graph and a corresponding table.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Review Report (Round 2)

Dear authors,

The revised manuscript looks much clearer than before. There are still several more problems as follows:

Problems / Improvements needed:

Line 58 (previous manuscript): Actually we want the authors to be more specific here, rather than removing "local" solely.

Response of Lines 63-66 (previous manuscript): "Similar studies have been conducted in numerous cities in developing nations, such as Abuja, Nigeria [13], Cairo, Egypt [14], Kandy, Sri Lanka [15], Tehran, Iran [16], Bangkok, Thailand [17], Rajshahi, Bangladesh [18], Bogor, Indonesia [19], and Shanghai, China [20].” - the authors only listed out these studies, but didn't go into deep. Therefore, the effects is not obvious, and the authors can remove some of these unnecessary information. Instead, the authors can focus more on the study of Pakistan, because the authors are interested in Karachi in this manuscript.

Line 132 (of previous manuscript): the authors mentioned that "radial basis function kernel is adopted in SVM", but what kind of radial basis function, any further descriptions can be provided in the manuscript as well?

New Problems found / Suggestions

Line 111 (current manuscript): Please remove "Rizvi et al. . Ranagalage et al."

Lines 271-273: "The weighting of an observation is not constant, but is a  function of the geographical location. Larger weights are assigned to  observations closer to point i [44,56]" - inverse distance weighting has been adopted in different environmental studies / investigation too, please include the scientific strength of this approach for estimation and relevant environmental studies, including the followings:

https://www.sciencedirect.com/science/article/pii/S187802961100586X

https://www.mdpi.com/1660-4601/18/12/6532

https://www.tandfonline.com/doi/full/10.1080/23311843.2015.1038944 

After including more relevant references and sources, the approach used here (inverse distance weighting) for estimation will become much more scientific and persuasive.

Line 474 (revised manuscript) - "optimized characteristics", what do the authors mean by optimized? Some more explanations may be better.

Lines 528-531 (revised manuscript) - "Furthermore, in Pakistan, geospatial data are not freely shared due to various institutional and technical problems. The openness of geospatial data collected by different  government departments can facilitate the understanding of the relationship between urbanization and UHI." The authors should cite some relevant work that discusses "open data initiatives that enhance sustainability of Karachi or other cities of Pakistan, or Data openness of land use, air quality and environmental datasets". Then, it can bring more positive impacts to the whole research community, as well as city development, as these are hot issues in developing cities / these are concepts that steer smart cities forward nowadays.

Section 6 (Conclusions) - the concept of "modelling work" for land use and land use change detection can also be mentioned as a future goal of this study.

Further, the concepts or approaches of distinguishing urban areas from rural areas can also be included - the authors can simply include some very basic information based on their response letter (Line 159-160 on Page 11 of the response file), and cite relevant references that the authors have provided in the Response Letter.

Other than the problems as aforementioned, I think the manuscript is very meaningful and worth publishing. Thanks for giving me an opportunity for me to read this manuscript. Good luck!

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Dear Authors, I am okay with the changes made. Thanks. 

Author Response

Thank you very much.