Remote Sensing-Based Urban Sprawl Modeling Using Multilayer Perceptron Neural Network Markov Chain in Baghdad, Iraq

Round 1

Reviewer 1 Report

The work is characterized by a good methodological approach related to urban sprawl, but the choice of models for predicting the increase of the phenomenon  as it shows us is not sufficiently justified.

Explanations are needed based on bibliographic human-geographical and urban research on internal and external migration policies to and from Baghdad (for example from the North to the South, or from other war-torn countries), internal demographic development, and existing urban planning forecasts, even if they are incomplete.

As there is insufficient urban geographical analysis of Baghdad, as well as seismic projection, no conclusions can be drawn regarding the proposal to build towers in the area. Personally, I would prefer the polycentric development of the country, but this should also be documented. Finally, the different typology of urban sprawl in relation to its urban social and morphological characteristics could be explored through an eco-landscape approach, if this is of interest to the authors.

Author Response

Dear Reviewer,

We have revised the comments and suggestion accordingly. Please kindly check the file attached bellow,

Thank you very much

Wafaa Majeed,

Author Response File: Author Response.pdf

Reviewer 2 Report

REVIEW REPORT - Remote Sensing based Urban Sprawl Modelling Using Multi-Layer Perception Neural Network Markov Chain in Baghdad, Iraq

In this article, the authors carry out a comprehensive evaluation of urban sprawl in the city of Bagdad. They applied remote sensing data with the aided of multiple softwares to explore, classification and analysis. The Future land use/ cover change (FLUCC) results from 2030 to 2050 recorded that agriculture was the only land use type that has observed a massive decreasing trend from 1985 to 2050 compared to other categories. Moreover, the results also suggest that, the rapid urbanization along with unplanned urban growth and rising populations including immigration to rural cities are the main drivers of all transformation in land use.

This research is a case study about LUCC in the city of Bagdad (Irak). The study is quite interesting and valuable. Results are of great value, especially for authorities in the city. I can recommend it for publication after some few suggestions are taken into account. These are the following ones:

• In the abstract the authors introduce the case study area and after that, they argue “Comprehensive Evaluation and understanding the effect of urban sprawl and its LUCC is therefore of great significance to manage land surface resources and sustainable development at the regional and global level.” My point is why to mix the global and local levels, once you are presenting the local study area already.
• Why some terms are in capital letters and others not such “Comprehensive Evaluation” or “Future land use/ cover change (FLUCC)”
• Also in the abstract the authors argue “Moreover, the results also suggest that, the rapid urbanization along with unplanned urban growth and rising populations including immigration to rural cities are the main drivers of all transformation in land use.” The authors refer to phenomena of false urbanization or pseudo-urbanization, right? But what do you mean with “including immigration to rural cities”? Do you mean to suburban areas close to the cities or remote rural areas?
• The authors introduce the study area such “Baghdad city, the oldest city in the region”. What do you mean with region here: the province of the city, the whole Irak, middle East, southern Asia?
• The authors introduce the paper like this: “Human socioeconomic development activity is the connection between human population increasing, residential expansion, and the general tendency of global urbanization were the sources of excessive remarkable variations in the urban land, intimidating the land use/cover change (LUCC) sustainability of the regional area [1–3].” It is always difficult to introduce a paper with very general terms. In any case, I do not know if “intimidating” is the right word here. I would recommend “pressuring” or something similar. After that, they argue “To date, urban population growth has been immensely impacted almost the entire world, lead to huge changes in land use/cover which have become a major concern in both natural resource management and sustainable development in the big city [8].” Referring to a global presentation of the work, I strongly recommend you to refer how some human dynamics are shifting at a global scale. Thus, the wealth is moving to the east whereas the global population is moving more to the South. This can be measured by using the center of gravity. Thus, it is more clear a broad perspective of the topic here presented and also to understand how the pseudo-urbanization processes are a huge problem in Africa and some Asian megalopolis. You can check these results in Globalization and the shifting centers of gravity of world's human dynamics: Implications for sustainability.
• After that, the authors present “About 60% of the 45 world’s city population will increase by 2050, which predicts by the United Nations [4–7].” Why do you need 3 references for presenting a simple population forecast?
• Just few lines later, they argue “From the problems existing in urbanization, it is necessary to identify the factors responsible to these changes by modeling future LUCC [16,17]. In order to comprehend the process of future land use/cover change (FLUCC) and its variations in the upcoming years of urbanization, LUCC modeling is very critical [10,17,18].” I think you could simplify this paragraph easily. We already know that is a crucial topic.
• Some lines later, they introduce the methodological tools “Currently, the remote sensing (RS) and geographic information systems (GIS) have great techniques for understanding landscape dynamics and provide valuable information [19]”. I would suggest that you rephrase this like “The combination between RS data and GIS tools offer a new valuable perspective for monitoring LUCC changes”.
• Previously in the abstract the authors argue “The present research has applied remote sensing data such as Landsat5 Thematic Mapper (TM), and Landsat8 Operation Land Imager (OLI) from 1985 to 2020 with the aided of multiple softwares namely ERDAS Imagine, GIS, QGIS and TerrSet to explored, classification and analysis. The result analysis found a significant increase in land surface cover in water bodies 2.39%, urban construction land 8.72%, and other land use types were decreased from -6.38%, -3.47%, -1.18% in agriculture, natural vegetation, and bare land, respectively. ”I would suggest you remove the name of the softwares and detailed results of the abstract. You should present these in amore simple and clear way.
• Also I could recommend you to incorporate to the list of references some relevant studies in the topic in other international cities/regions where they raise some similar approaches under different circumstances. For example, the study of Hermosilla et al. Land-use mapping of Valencia city area from aerial images and LiDAR data present a novel method for land use classification in a consolidated city of the south of Europe and very accurate results.
• Finally, I would suggest to expand (doing bigger and clear) the maps (Figures 5, 9, 11, A1 and A2) and arranging all the maps scales for doing them easier to read and to interpret.
• I have observed some typos in the text such “Figure 12. Overal change of urban CL from 1985 to 2050” instead of “Overall”, and other ones.

Author Response

Dear Reviewer, 

We have revised all the comments and suggestions which mentioned by you in the attached file bellow. Please kindly see it.

Thanks very much

Wafaa Majeed,

Author Response File: Author Response.pdf

Reviewer 3 Report

Introduction section provide a lot of general aspects of urban land use change and it can focus on remote sensing approaches of this issue. I appreciate the explanation about the employment of the Markov Chain prediction modelling of urban area change.

Objectives must be written to the end of the introduction as a separate paragraph. We can extract them from text but is important to list and argue them in the context of your approach.

Methodology

If geographical names are used it is good to display them on the maps. They occur in figure 1 but please check if all names from text are on the map.

Figure 1. There is no source of the map of urban area footprint. Graphical scales can be simplified with few exact values like tens or hundred kilometers. The legend of the map of Bagdad area needs an explanation of the acronyms.

I recommend to explain a little more the study area sub-section because it is important to know why it was selected for the current approach. Also the statistical demographic data need a little more explanation.

Row 159 - phenological is correct; I recommend an English language detailed check of the paper by a specialist.

The tool used from Google Earth Pro is not suitable explained. Please provide some details.

2.2. Data sources - What is the processing level of these Landsat scenes? Did you used atmospherically corrected data?

2.3.1. Image processing - row 179 please use ...classification accuracy... ( again, English language editing is necessary...).

Usually the software used can be mentioned only if there are particular or specific issues in image processing, and not to the beginning of the sub-section.

Figure 3. The scheme is limited to a visual analysis and does not provide the advanced processing of change detection analysis. This might be better included together with figure 4 scheme in order to better extract the original contribution.

Kappa coefficient is well known from literature and the formula can be mentioned only as a reference. 

For the entire methodological section - It is important to explain and argue the use of the methodology from literature and software manual to the specific issues of your approach than to explain again theory items.  Please extract the most representative aspects from the methodology (examples explained) because the journal scope focused on remote sensing data processing methodologies/techniques.

Row 284 - correct is Determinants...

Results section can be separated from analysis (discussion section or interpretation of results).

Statistics are difficult to be followed without a graphical or even cartographical presentation (partly done). It is useful in order to understand the sense and the magnitude of changes before the predictive modelling is applied. Avoid general statements and extract the most particular aspects to build the findings of your approach. Readers search for new, particular aspects.

Figure 10 is difficult to be understood (probably the change of scale or legend can solve the problem).

3.5.1. Scenarios are important in the context of the current approach. This is the reason we recommend to explain them in detail and not only display maps and statistics. Extract the most specific issues from methodological approach to the realism of the data produced. It might be interesting to connect them with terrain and historical reality and especially to the regional or National urban planning and regional development strategies (ex. population dynamics and migration, effects of wars, economy, transport infrastructure etc.). They can be a helping tool for future planning strategies.

I suggest to focus on a detailed case study from these scenarios. 

Figure 5 - maps are displayed in a small format and are difficult to be red and understood. It is necessary to explain a little more the classification algorithm selection and even some specific parameters. Maps are very general. Probably you might focus on a detail showing the most spectacular change and provide an explanation too.

Figure 6 - legends need to be explained.

Page 11 it is not necessary to explain the cardinal points acronyms. These occurs on figure 7, too.

Figure 8 is useful but you can connect the demographic data with the LULC classes and their dynamics.

Map interpretation can be better linked with examples from the region and probably validated with other information from literature. Some images like orthophoto samples could be used to explain the phenomena.

Section 3.4 is not clearly explained. Why searching for the relationship between building and the altitude or slope? I think an explanation can ease the understanding of this aspect. Explain if the topography is an essential factor of urban area dynamics...

Figure 9 is difficult to be red and understood. 

3.4.1 Section is important but the maps are displayed at small scales and difficult to be red.  Please focus on some details in order to explain the predictions.

Discussion section - the validation of the prediction can be done by explaining the results in the regional context with specific data (general explanation provided). There are prediction maps to be interpreted in connection with the regional background, historical aspects, resources, population and migration pattern, transport infrastructure and sustainability issues (if planning strategies mention them).

The evaluation of the methodology needs to be done in comparison with similar approaches. It is necessary to extract the positive parts and the limitations of the modelling. 

Conclusion section - it is important to explain the degree the objectives are covered by the current approach. It is also useful to extract data from the analysis to be used in regional development strategies (partly done). What is planned and what is not in LULC dynamics? 

Author Response

Dear Reviewer, 

We have revised all the comments and suggestions with professional language editing which mentioned by you in the attached file bellow, the professional language editing is attached in the last page of the response version. Please kindly see it.

Thanks very much

Wafaa Majeed,

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The work is characterized by a good methodological approach related to urban sprawl

But I suggest that you remove from the conclusions the phrase that suggests building office and residential towers in the area because you do not have the necessary elements of urban planning and urban landscape needed for such a proposal. Remote Sensing is not an Urban Planning Journal. 

Author Response

Dear Reviewer,

First, we hope you are doing good with healthy. By the way, we thank you very much for the all comments and suggestions, it was critical for our manuscript for publication. All the comments and suggestion has been revised as attached in the file below. please kindly check it. We hopefully this version is ready for publication.

Best regard,

Waffa Majeed,

Author Response File: Author Response.docx

Reviewer 3 Report

It is recommended to read again carefully the manuscript and check some English language issues. For example row 610.

Avoid details with no importance for remote sensing data analysis.

Recommendations can be the last part of discussions section.

Figures 9 and 10 legends need some changes. Red color of built-up area can be difficultly noticed  from the red color of elevation and slope gradient areas.  

Author Response

Dear Reviewer,

First, we hope you are doing good with healthy. By the way, we thank you very much for the all comments and suggestions, it was critical for our manuscript for publication. All the comments and suggestion has been revised as attached in the file below. please kindly check it. We hopefully this version is ready for publication.

Best regard,

Waffa Majeed,

Author Response File: Author Response.docx