Examining the Density and Diversity of Human Activity in the Built Environment: The Case of the Pearl River Delta, China

Round 1

Reviewer 1 Report

Examining the Density and Diversity of Human Activity in the Built Environment: The Case of the Pearl River Delta, China

Reviewer’s assessment

This paper has been developed specifically to analyse spatial patterns in the Pearl River delta and includes a review of relevant previous literature of spatial analysis in China. As presented, however, it is only of local rather than international interest and does not seen relevant at all to a journal dealing with sustainability. The paper does not discuss sustainability or sustainable development and should not be accepted for that reason alone. Sustainability is only mentioned once in the body of the text. The wording is as follows.

“Based on active data mining and spatial orientation, this article explores the spatial characteristics of human activity within the built environment. ArcGIS, serving as an information platform for sustainable land management …, is extensively used.”

It is not enough to generate spatial patterns from novel sources to justify publication in a journal dealing with sustainability. The implications of the data also need to be discussed. ArcGIS is, in any case, a multipurpose mapping platform, not one that is only concerned with sustainability.

The lack of discussion of sustainability is a reflection of poorly articulated aims and research questions. These are specified as:

• How can we determine the density and diversity of human activity in built-up areas?
• Which factors, besides urban scale, indicate higher efficiency for a city?

Question 1 is appropriate. However, the authors need to relate their findings to the wider literature about urban densities. The study is based on the overall densities of activity in uniform sized cells in a spatial grid. For spatial planning, activity density is an abstract and largely untried concept. It needs to be related to city wide population densities, to neighbourhood and to local design densities. The concept of diversity of activities is likewise unconventional. The authors mention Jane Jacobs as a source. Jane Jacobs argued for mixed land uses, for example neighbourhoods with a mixture of housing, retail and employment, rather than single use zoning. Land use, other than in relation to density (intensity of use)  does not figure at all in the paper. Activities are related to land use, but a firmer and clearer relationship needs to be established. One of the components of diversity as measured here- retail sales- is an indicator of local purchasing power and therefore of income inequality. It is therefore very unclear as to the meaning of the data as aggregated.

Question 2 is far too ambitious. Sustainability implies a particular type of efficiency- ecological efficiency. However, the definition and implications of efficiency are not discussed. The development of ghost towns is obviously inefficient, whether looked at from an ecological or economic viewpoint. However this paper does not provide insights into why ghost towns emerged or how they can be avoided in the future. The factors that determine urban spatial patterns and the use of land are numerous  and cannot be determined by a mapping exercise alone. 

In addition to all this, the methodology needs attention. The authors should explain why they have chosen the range of data, how they obtained the data and whether any commercial or other constraints stood in their way. Data from online shopping or from social media is not readily available for public researchers in Europe and the US. The authors state that they geocoded online/ TV shopping data. Does this mean that they had access to the addresses of purchasers? It would be of interest to know what exactly could be learnt be learn from each of these sources, with very much more detailed and qualitative discussion of the implications, including the implications for personal privacy. There are potential ethical concerns with the data itself. 

For an international journal in spatial planning or sustainable development, the paper needs to display very much more critical, theoretical and methodological awareness. The paper might nevertheless be better suited to a journal specialising in GIS, smart cities or urban modelling.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Good paper overall, needing some revisions before publication. See some comments below:

• please clarify (intro + disc) the INTERNATIONAL relevance of your study
• please provide further details on the methology, especially how to reproduce the maps you have presented and clearly illustrated in the paper
• Literature review is focused on Chinese references, but you should improve the international perspective of the paper.
• What is the originality and novelty of this paper? Please add some comments in the conclusion section.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Generally speaking, this work has an appropriate scientific level and I would judge its study objective to be appropriate and in line with the title and subject of the publication. I similarly think that one of the strong points of this proposal is its originality, both in terms of the subject treated and the hypotheses analysed.

We appreciate the measured use of the bibliography cited and how this largely focuses n works that are relevant to the subject and well-contrasted. The structure used is in line with accepted practice and follows the prescribed order for a scientific article. This work should have an appropriate writing style and correct syntax and grammar.

Questions relating to the approach and content:

This study covers a matter of current relevance, rapid urbanization in China has been accompanied by spatial inefficiency in patterns of human activity, of which ‘ghost towns’ are the most visible result. Using the Pearl River Delta (PRD) as an empirical case.

The results presented are consistent with the analysis carried out. This empirical study indicates that cities with higher population are more likely to avoid becoming ‘ghost cities’ and smaller cities face the challenge of inefficient construction land. This means that most small cities in the PRD are faced with the difficulty of attracting a sufficiently large population. While smaller ones have also greater marginal benefits in improving land use efficiency.

 

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Examining the Density and Diversity of Human Activity in the Built Environment: The Case of the Pearl River Delta, China

Revised

Assessor’s review

I have provided a lengthy review as assessment of the paper is very much more negative than that of the other reviewers and it is important to understand why my view is negative.

The authors have sought to respond to earlier criticisms. The issue of ghost towns is now a central theme and the sources of data are better explained. It is possible to see what the research is intended to achieve. Sime small ethical concerns have been resolved. The structure of the paper has not been changed, however and is confusing, especially for an international audience. Material has been to the discussion and the conclusions without undertaking concomitant changes to other aspects of the paper. Basic problems remain moreover about the validity, meaning and implications of the data, as interpreted. Having read the paper again, it is impossible to see how, in their present form, the results can be used to promote effective or sustainable spatial planning.

A fundamental flaw is that the authors ignore related economic and social research concerning urban form and urban patterns and ignore in addition research in the field of urban design. They state: “Since there is little research directly examining which factors have impacts on human activity in the built environment, this paper proposes three potential factors as pivotal elements in making land use policies, namely urban scale, morphology and administration.” A brief review of Google scholar will reveal a huge amount of literature covering the factors of urban scale, morphology and administration and its impact on spatial planning. There are full blown academic disciplines concerned with the factors that influence or ought to influence spatial planning. It is not new, for example, to suggest that the density and diversity of human activities is related to city size and that larger cities have higher densities and possess a wider range of human activities. Urban geographers have been arguing this in Europe and the US for many years. The paper might therefore be better presented as a confirmation in a new setting, southern China and with new methods of older observations and theories of growth. The paper should not, however, be presented as novel at an international level.

The literature review:
The relation between diversity of activity and density and ghost towns is, in any case, more problematic than the authors suppose. To say that larger cities have higher densities and more diversity of activity does not necessarily mean that they grow more quickly. The pattern in England and some other European countries is, for example, for most cities, whether large or small, to grow. Equally, there are some large cities, such as Liverpool or those in Eastern Europe that have experienced population decline and have had extensive areas of unused or underused land. The authors cite the definition of Shepard [24] of a ghost town as “a new development that is running at significantly under capacity, a place with drastically fewer people and businesses there is an available space for.” The reference to new development is significant as ghost towns can also be caused by gradual, long term population loss or by the collapse of local industry or by the closure of mines. Ghost towns and ghost cities in China are, in contrast, as I understand it, the product of excessive local building rates outstripping demand. However, the paper fails to follow Shepard’s definition. The paper maps all urban activities rather than those linked specifically to new development, it does not define urban capacity, another aspect of Shepard’s definition  and it does not discuss what is ‘significant under capacity’. The analysis lacks sensitivity in relation to the phenomenon being studied. Some variation in both the diversity and density of activities is inevitable in a large metropolitan area. Demonstration of such a variation is simply not important in itself.

Methodology:
The methodology section brings in some good material, especially about the evolution of relevant studies in China. More detail is nevertheless required. The authors state ‘Chi et al. [4], Jin et al. [92 11] and Zheng et al. 93 [12] explored human activities to detect possible ‘ghost towns’ in China’. The authors must discuss the strengths and weaknesses of the methods used in these previous studies, at greater length. The authors are dealing with proxy data for ghost towns. The availability of real ‘ghost town’ data, for example, the number of property vacancies and the amount of unused land should also be discussed. Presumably data on vacancies and unused land is not available, but this is not stated. Finally, the authors are using data sources that are only available at the discretion of commercial operators. Panoramio is reported to have closed down after being taken over by Google. Other data would probably not be available from equivalent blogging or online shopping services in Europe or North America. The unusual or unique character of the dataset needs to be mentioned for an international audience, especially if the authors claim, as they do in the conclusions, that their study is an international contribution to open data.

Discussion:
Though there is a section called discussion, it does not go back to the literature on ghost towns, other than in the most general manner. In addition, and most fundamentally, the discussion either repeats reasons why the paper is important (more appropriate for the introduction) or brings in new material connected to urban compactness and open data. Compactness of urban form is potentially important, but there is no discussion anywhere in the paper about how it has been measured in previous studies in China and elsewhere. The authors need to reflect on the contribution of knowledge above to all ghost towns but also to compactness if they wish, first to equivalent previous Chinese studies and then to previous international studies and the previous studies need to be discussed first in the literature review. They need to say what is new, original and interesting about their study. The suggestion that governments should promote the availability of open data is particularly problematic. The promotion of open data is not a recommendation that flows from the study as presented, the usual criterion of whether a conclusion is justified in a scientific study. The recommendation is referenced instead to an Australian government report written with a different purpose and in a completely different political context. If the authors want to discuss open data, they would have to discuss the relevant literature first. Most European countries have been pursuing open data polices with varied degrees of success over the past decade. If there is an equivalent policy for open data in China, this should be stated near the beginning of the paper. Finally, the authors do not reflect on a point I raised in my previous review. One of the components of diversity- retail sales- is an indicator of local purchasing power and therefore of income inequality, rather than urban form.

In its present form, the paper lacks credibility for an international audience in spatial planning. It may, however, be suitable for a journal written specifically for a Chinese audience and may also, with significant variations, be suitable for a journal dealing with e-planning and technology.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 3

Reviewer 1 Report

The revised draft is better, but serious flaws of logic and evidence remain. The revision locates the study in the context of debates in China, explains how the study has sought an innovative approach in the face of data limitations, explains the technical aspects in a convincing manner and has clarified the issue of ghost towns to an extent, but not sufficiently so. The main weakness remains, as it was in the first and second drafts, a failure to be absolutely clear about the meaning and implications of the heat maps and a tendency to overstate the implications. Clarification has raised further, deeper issues and questions. The authors need to be very much more aware of the relation between the data and the phenomenon- ghost towns- they are examining, including the limitations of that data.

The paper contains the very odd statement  ‘Thus far, we have not seen any widely recognized definition of ghost towns.’ The conclusions then state ‘This (ie the paper) helps further the understanding of ghost towns because Woodworth and Wallace [4], Shepard [9] and Sorace and Hurst [10] are in agreement that there is not yet any consensus in the definition and factors leading to this phenomenon. Furthermore, our analysis provides a valuable reference for cities worldwide considering how to avoid ghost towns development and ensure urban sustainability.’

The paper does not explain how and why the cited authors differ in their definition. From the limited extracts cited, the difference between the cited authors mostly arises because they are concerned with different aspects of ghost towns and does not involve any fundamental definitional dispute. Moreover, a good quality scientific paper should not simply admit the existence of disagreement. It should either attempt to resolve any differences or to determine which author is correct. The test of originality in a scientific paper is whether it adds to a body of knowledge as defined in the literature review. The international literature on ghost towns is larger than the authors cited, but at least the paper makes a start through  identifying a literature. Then in a departure from standard practice in research, the paper states that it does not seek to develop the ideas of the cited authors. That failure means, as a point of logic, that contrary to the statement in the conclusions, the paper fails to add to knowledge or understanding of ghost towns. Why cite authors in the literature review, if the research cannot reflect on what they have said?

It is, in any case, unclear why the definition of a ghost town has caused so many problems. The Oxford English Dictionary defines a ghost town as ‘a town partially or completely devoid of its inhabitants’. Shepard’s definition is consistent with that of the OED. A ghost town can be defined in principle, therefore. The unanswered questions are partly those of quantification. At what point, does a lack of population or a lack of economic activity become so severe as to count as a ghost town? The paper as presented here does not examine this issue at all. The variations in the heat maps are variations in intensity and no attempt has been made to indicate a critical level beyond which real problems are likely to exist.

There is more than one cause and more than one type of ghost town, as the paper correctly recognises. Following Sorace and Hurst, the authors might wish to distinguish between a ghost town that has involved massive recent development and a ghost town with sufficient existing built environment but a fairly small amount of new construction. Again the authors make no attempt either assess the usefulness of the distinction by Sorace and Hurst or to specify which type of ghost town exists in the case study area. Surely construction data is available to apply to the case study area, so testing the definition by Sorace and Hurst. The authors recognise that construction data has been used in previous studies but then seem to dismiss its relevance, because ‘these topics have already been addressed’. The authors do not, however, show how construction data issues have been addressed in the case study region, however. In the context of ghost towns, it would be more logical to examine human activity in addition to (rather than instead of) construction data and then to examine how the construction data and the human activity data relate to one another. Again, the authors seem to be dismissing rather than building on previous research.

A related problem is the way the authors repeatedly talk about construction land and its distribution without explaining the precise characteristics of the data set. Figure 1 shows, for example, the distribution of construction land in the Pearl River area, though it does not give the time period and it does not say whether this is vacant construction land (land allocated for development, but not developed), construction land in the process of development or land that has been developed. Of crucial importance in defining the type of ghost town is the amount of vacant construction land, the time the land has remained vacant and the relation between construction land availability and construction rates. The maps are meaningless without the relevant detail.

The paper hints at it origin. It states that ‘Data collection was completed in July, 2015, contributing to the Master Plan for the region of Pearl River Delta (2015-2020)’. Could the authors explain the form in which the data was presented to those preparing the Pearl River Delta Plan and how the data was used in plan preparation? Inclusion of this information will be helpful in assessing the usefulness of the analysis and its local impact. If the data was presented in a different format, for example without the heat maps or was not used in plan preparation, this should also be stated. At the moment, the authors assert that their study helps promote sustainability, but they provide no evidence that planners and policy makers in the region have conceived of sustainability in a similar way.

Further, in terms of policy implications, the paper states that the ‘optimum population size for megacities such as Guangzhou and Shenzhen should also be considered’. Again, is the optimum population size for Guangzhou and Shenzhen seriously being considered at the moment? If yes, please provide evidence. The question of optimum city size is very old and also very contested. Many would argue that there is no single population optimum. In any case, there is no obvious link between the data analysis as presented and any specific population or other city size optimum. An optimum city size implies limits to growth and the study indicates no limits. The analysis is consistent with established research showing that the centres of larger cities are denser and more diverse than those of smaller cities, so supporting the development of ever larger cities, whilst at the same time not postulating an upper limit to city size. The observation that, in the case study, the subregions of low human activity and density are located away from the main urban concentrations has a similar implication of favouring still further concentration.

My reservations about the paper remain. The technical content of the paper, once fully explained, looks sound. The paper could therefore be acceptable to a specialist GIS journal or a specialist journal on e-planning or data analytics. However, the paper would not be published in an international scientific journal in spatial planning, urban studies or sustainability, without further major changes. The construction land data still needs clarification. However, it is unlikely further clarifications will be sufficient. At the very least, the authors will have to include construction data so that it is possible to see trends over time as well as in space and determine whether ghost towns have emerged as a result of massive new development proceeding ahead of demand. Ghost towns of this type emerged in Europe, for example in Spain an Ireland just after the financial crash of 2008. I have also seen examples in the UK and Malaysia. The authors might also pay more attention to local and regional policy and planning debates. The technical content is presented in a way that is poorly related to a practical planning context, even though it was apparently prepared for a regional spatial planning exercise. In some ways the paper is a demonstration of the limitations as well as the strengths of data analytics as applied to spatial planning, but those limitations have not been grasped.