Urban Growth Forecast Using Machine Learning Algorithms and GIS-Based Novel Techniques: A Case Study Focusing on Nasiriyah City, Southern Iraq

Round 1

Reviewer 1 Report

This manuscript (ijgi-2176184) tries to present a machine learning technique that can simulate and predict urban sprawl for a long time and can be integrated with optimisation land use techniques to make suitable decisions in Nasiriyah City, Iraq. Although it fits the aim and scope of this journal and the amount of the work is enough, its contribution to land use and land cover change modeling is not significant. Another serious concern is that some related latest studies have been neglected. Also, the current results of this study can hardly be reviewed because of those problems about data and methodology. Therefore, a Major Revision is required. More detailed comments and suggestions are presented as follows:

-1. The logic of the Introduction Section needs to be improved because it is currently a bit confusing. The authors mentioned "land use optimisation" at the very first part, but land use optimisation has not been significantly involved in this current study.

-2. The scientific question or research gap is missing in the Abstract. Similarly, the Introduction Section is a bit weak because the authors failed to raise an important scientific question or gap related to this study and beyond this study area. Therefore, potential readers can hardly identify the need that the authors should have to provide a new solution. What I learn from the introduction is that the authors apply some previous established models to a specific study area (Nasiriyah City, Iraq). Note that the multi-criteria decision, Markov Chain, and ANN-CA models are not even new methods for future land use and land cover change prediction and urban growth boundaries optimisation.

-3. The authors need to explain why the linear regression (LR), K-nearest neighbour (KNN), AdaBoost (AB) and random forest (RF) were selected in this study as there are many other novel methods, such as genetic algorithm.

-4. The authors have mentioned that: "The main objective of the present research is to innovate a simple novel technique based on ML and GIS to predict urban expansion for a long time and detect changes in land use land cover that have happened and may take place in the future for any plot that may be assigned to construct industrial projects or other activities". However, a large number of previous studies have already done so (see below as examples, and many more).

Simulating Urban Growth Using a Random Forest-Cellular Automata (RF-CA) Model. ISPRS Int. J. Geo-Inf. 2015, 4(2), 447-470

A spatial error-based cellular automata approach to reproducing and projecting dynamic urban expansion. 2022, 37: 560-580

-5. The authors have mentioned that: "Given that BAUS was adopted in this study, the effect of socioeconomic factors on urban growth was assumed to be constant and unchanging", which is not convincing. It is because the socioeconomic factors, such as population density and regional GDP, are very important to urban land use changes and urban growth boundaries.

-6. Section 2.1. Neighbourhood factors and Section 2.2 Natural driving factors: The authors need to explain why just these driving factors have been selected.

-7. It seems that this manuscript did not consider the serious problem of the multicollinearity of different driving factors.

-8. The authors devote too much space to describing the background of the study area, which should be briefly mentioned in several simple sentences. In addition, the authors also need to look further into the latest research in this field. In fact, the literature review is far from enough. Especially, the more advanced patch-based CA model has been widely utilized by many cutting-edge research (see below as examples). However, this well-accepted technique is not even mentioned in the manuscript.

Modeling urban land-use changes using a landscape-driven patch-based cellular automaton (LP-CA). Cities, 2023, 132: 103906.

The Literature Review section is meant to set the context for your research work and highlight how it contributes to the knowledge in this field and builds on previous similar studies. In particular, the authors have mentioned that: "However, for modelling of urban expansion, the CA simulation technique applies similar transition rules for all cells in the model space. Thus, the model ignores spatial heterogeneity change, which indicates that such systems are prone to over or under simulation". Actually, this disadvantage can be effectively tackled by the patch-based CA models.

-9. At the beginning of the Section 3, I suggest the authors to provide an overall flowchart of the methodology part.

-10. Section 3.2, Table 1: the data description section failed to provide the specific details of the input data (in addition to land use data), such as the dates in acquiring them, and accuracies. In particular, what are the years of the driving factors? Are they consistent with the years of the land use data?

-11. Section 3.3.1 GIS-based classification Landsat images using RF: what are the validation data for land use classification? In addition, what are the classification accuracies?

-12. To prepare for the simulation process, the following ANN algorithm settings were used: 1000 iterations, a neighbourhood value of 3 pixels, a learning rate of 0.001, hidden layer of 10 and 0.05 momentum. The authors need to explain clearly the determination of these different parameters.

-13. Table 2: how to determine these radius of buffer zones for different types?

-14. The authors also need to improve the Conclusion Section by mentioning the main shortages of your work.

Author Response

The response is submitted 

Author Response File: Author Response.pdf

Reviewer 2 Report

This paper analyzed urban land use changes in a city in Iraq during 1992-2022, and then predicted its urban growth in the following three decades (2022-2052). Urban expansion analysis is a classical theme in Geography. Overall, the novelty of this paper adds to this field is limited. My specific comments are as follows for improving the quality of this paper.

1. The severe flaw of this paper is its lengthy writing. Where there could have been a brief explanation, the author wrote so much text that it drowned out the most critical information. (1) Please concise the Introduction section and propose your scientific questions distinctly. (2) Please concise the discussion and conclusion sections. Please write your conclusions directly in the conclusion section without repeating the research background. The first four sentences of the conclusion section could be removed. (3) Please remove unnecessary and repetitive references. I suggest containing only about 45 references.

2. This paper only reported a tiny story but was written on 24 pages. I strongly suggest concise the whole paper by 50%.

3. Regarding driving forces of urban expansion (Section 2), authors reviewed neighborhood and physic-geographical factors but missed socio-economic factors.

4. Please direct introduce your methods without too many reviewing statements.

5. This paper has four flowcharts (Figure 2-5), which must be combined into one figure. Please remove unnecessary details in the four flowcharts. There are too many acronyms in Figure 3 to understand clearly. Please write their full names.

6. There are too many keywords. Some keywords should be combined, for instance, “urban growth prediction”. GIS and RS are common acronyms that should be removed.

Author Response

The response is submitted.

Author Response File: Author Response.pdf

Reviewer 3 Report

Dear Authors

Many thanks for submitting your paper for possible review.

I have carefully go through the paper that you have studied for urban sprawl growth mapping in the Nasiriyah City, South Iraq was critically investigated and well organized. However, I am requesting the authors to incorporate my comments and suggestions during the revision stage.

My decision and comments and suggestions are specified below:

Decision: Paper can be accept after minor correction.

Comments and Suggestions:

(1) Title should add the Nasiriyah City, South Iraq 

revised suggested title: "Urban Growth Forecast Using Machine learning Algorithms and GIS based Novel Techniques: A case study focuses on the Nasiriyah City, Southern Iraq"  

(2) Keywords should be limit to 5-6 with important one only your key index. please add Iraq, as keyword.

3. section 3. Methodology should be heading as Materials and Methods

4. In Figure-1, study area map, left key maps should be use scale and coordinate information. In Right map, insert important locations and show river name in the map. Without the location including surrounding name,  map has no sense. 

5. DEM, S, PMR, PCC, PWP, PSN and PR, the meaning of these should be declare first and use this acronym. Next throughput the text, use use only DEM, S, PMR etc.

6. In Figure 6, all maps must be prepared with scale, north line, coordinates, specific year in title. You could separately prepared layout and place in same sizes.

7. 4.2. Insert of : Prediction of urban growth result

8. Figure 9, zoom view should be separately identified and showing each zooming view of different parts of the areas to be highlighted in different corners with marking in the existing map.

9. Figure 10, map shows at the bottom marked as 2092, why? you have mentioned that predicted time period is 2052. Is it mistake? Insert location names of the city and surrounding important locations to identify the details of it. Coordinates should be displayed in deg.min.sec.

10. Figure 11, both figure caption, and within figure R2 should use in superscript.

11. Conclusion part should be specific and it should include in heading as Conclusion and future implications. Major findings and specific outcome of this study should be presented in a point-by-point basis. so readers will see in a overview.

Overall, the manuscript need revision. 

Final decision: Manuscript can be accept after minor correction.

Thanking you

Good luck

Reviewer

Dated: 31 January, 2023, Time: 12;15 PM

Author Response

The response is submitted 

Author Response File: Author Response.pdf

Reviewer 4 Report

The authors propose a composite ML model to predict city expansion and its environmental impact.

The idea of inferring the update rules of a CA by means of different ML techniques seems to be quite efficient.

The work is quite clear and detailed, and this makes it useful for anyone who wants to reuse the same technique for the study of another city.

Perhaps the text as a whole could be written in a more concise way.

It would be useful to have a list of all the acronyms used (there is a special section in the template for this purpose).

Author Response

The response is submitted 

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Thank you for incorporating my comments and suggestions.

Author Response

Thank you very much for your efforts. We are grateful for your contribution to the success of this work.

Reviewer 2 Report

My concerns have been addressed. Thank you for your great work. I again suggest cutting down the length of the paper and reducing unnecessary sentences. The publication of the paper is only the first step, but more important is the spread of the paper and its influence on others. The concise writing and clear logic can make readers like your paper.

Author Response

Thank you very much for your efforts. We are grateful for your suggestions; thus, they were considered carefully. 
1-    Section 2 (driving forces of urban growth) was summarized ;
2-     Section 3.2 (RS and data collection) was summarized;  
3-    The number of references decreased from 88 to 85.