[JCR] Urban Wind Environments on the Distribution of COVID-19

The Impact of High-Density Urban Wind Environments on the Distribution of COVID-19 Based on Machine Learning: A Case Study of Macau

Abstract: The COVID-19 epidemic poses a global challenge. Urban wind environments, as integral parts of cities, may contribute to virus spread. Understanding their impact is vital for effective prevention strategies. This study employs a conditional generative confrontation network (CGAN) to train a model using simulated urban wind and COVID-19 distribution data. The model predicts COVID-19 distribution probability under various wind environments, revealing their relationship.

Key findings: (1) Different wind environments show significant variations in COVID-19 distribution, with high building density areas being more susceptible. (2) COVID-19 hotspots in building complexes correlate with building characteristics. (3) Building area influences COVID-19 spread. Post-epidemic planning in high-density cities can consider building houses on the northeast side of mountains, adopting “strip” or “rectangular” layouts, and ensuring the long side faces southeast. (4) Overall wind speed around buildings should exceed 2.91 m/s, with an optimal range of 4.85-8.73 m/s. These findings inform urban planning and public health strategies for more effective prevention and control. This study utilizes machine learning to uncover the impact of urban wind environments on COVID-19 distribution, providing crucial insights for planning and public health. [SCITIP]

DOI:  https://doi.org/10.3390/buildings13071711

CITE: Zheng, Liang, Yile Chen, Lina Yan, and Jianyi Zheng. 2023. “The Impact of High-Density Urban Wind Environments on the Distribution of COVID-19 Based on Machine Learning: A Case Study of Macau” Buildings 13, no. 7: 1711. https://doi.org/10.3390/buildings13071711

Read more

作者的话

1. 心得分享

今日分享之论文，是本月见刊上线的第一篇 SCIE 期刊论文，按照目前的学习计划安排，这一篇应该是这个月唯一的一篇的。

特别强调：这一篇论文有“Feature Paper”标识。如官方解释所示：

Feature papers represent the most advanced research with significant potential for high impact in the field. A Feature Paper should be a substantial original Article that involves several techniques or approaches, provides an outlook for future research directions and describes possible research applications.

Feature papers are submitted upon individual invitation or recommendation by the scientific editors and must receive positive feedback from the reviewers.

专题论文代表了最先进的研究，在该领域具有巨大影响力的巨大潜力。专题论文应该是一篇实质性的原创文章，涉及多种技术或方法，提供对未来研究方向的展望并描述可能的研究应用。专题论文是根据科学编辑的个人邀请或推荐提交的，并且必须得到审稿人的积极反馈。

上期预告也有提及，《The Impact of High-density Urban Wind Environments on the Distribution of COVID-19 Based on Machine Learning: A Case Study of Macau》（基于机器学习的高密度城市风环境对 COVID-19 分布的影响：以澳门为例）一文于本月在线发表在期刊 Buildings（ISSN: 2075-5309）（SCIE、Scopus）之 Special Issue “The Adaptability of Residential Planning and Design to World-Changing Events”（住宅规划和设计对世界变化事件的适应性）。本期重点介绍一下。

本人为共一作者，在本次论文中主要是从写作、全文翻译（含改写）、写 coverletter、返修改动了 1 轮、投稿、校稿 1 轮等。返修确实是一个很头大的事情，这一次活生生返修多了 5 -6p 的版面，我还要想审稿人这个疑问，系统回应了，但是对应的文字插入哪里，怎么把原来的逻辑顺下来。综合来看，我个人感受，返修比写第一稿更费时间。返修完了在第二轮那会，审稿人 2 给了看上去非常高的评价：“The paper has been revised correctly.  I can recommend it to be published now, as it presents very high scientific level.”

[JCR] Urban Wind Environments on the Distribution of COVID-19

中间很长一段时间是整个澳门特区的风环境的图卡住了，为了出图，中间辗转了很多。这个重大功劳要归功于我们家娜娜，最后跑了一张全澳门的风环境超级叼图。

原图可 152MB 呢老铁~！打晕我绑架我我都不会给的一张图。

[JCR] Urban Wind Environments on the Distribution of COVID-19

2. 作者解读

[JCR] Urban Wind Environments on the Distribution of COVID-19

线上阅读地址：https://www.mdpi.com/2075-5309/13/7/1711

引用格式：

Zheng, Liang, Yile Chen, Lina Yan, and Jianyi Zheng. 2023. “The Impact of High-Density Urban Wind Environments on the Distribution of COVID-19 Based on Machine Learning: A Case Study of Macau” Buildings 13, no. 7: 1711. https://doi.org/10.3390/buildings13071711

大白话解释一下这篇论文：说来话长，这个征稿其实是去年就收到了，后面做了别的事情以后，在 DDL 第一生产力的促进之下，终于给投了出去。熟悉我的小伙伴都知道，一年前澳门是大爆发了 COVID-19 当时我也很久不出门，年底 + 今年年头发了两篇关于 COVID-19 和城市形态的论文，然后那 500 个病例的数据还在，刚好，延续了之前的研究，叠加了一个风环境的元素。然而，时过境迁，现在讲 COVID-19 显得有点过时，但我们更想强调，这是一种规划设计前期的方法，澳门这个 case 只是我们讲解这项方法的一个例题。understand？在本文中，探讨了以下六个问题：（参考正文的 1.3. Problem Statement and Objectives）

  1. 以中国澳门为例，高密度的城市空间及其塑造的城市风环境对 COVID-19 的分布有何影响？
  2. 机器学习技术如何协助分析 COVID-19 的分布？
  3. 进一步，根据 2022 年 6 月中国澳门 500 例 COVID-19 爆发的足迹数据，COVID-19 与城市风环境之间的相关性如何？
  4. 不同形态布局下城市风环境如何促进或抑制 COVID-19 的分布？
  5. 采用哪种形式的可持续居住区布局规划设计，更有利于适应流行病环境？
  6. 在后疫情时代，这项研究能为其他类似疫情提供哪些思考？

Keywords: machine learning; COVID-19; urban wind environment; high density city; urban planning; urban public health

Institution:Faculty of Humanities and Arts, Macau University of Science and Technology

Funding: This research was funded by the National Social Science Foundation’s special academic team project for unpopular research (21VJXT011).

论文标题框架：

1. Introduction

1.1. Research Background

1.2. Literature Review

1.2.1 The Pandemic and the Urban Wind Environment

1.2.2 Application Areas of Machine Learning in COVID-19

1.2.3 COVID-19 and Housing Conditions

1.3. Problem Statement and Objectives

2. Materials and Methods

2.1 Data collection

2.2 Data processing

2.3 CGAN method

3. Model Training

3.1 Model training process and verification

3.2 Correlation Analysis of the Wind Environment and COVID-19 in Different Building Layout Types

3.3 Robustness test of the model

4. Discussion: Residential Planning under Long COVID

4.1 Typical Residential Building Types in Low-Epidemic-Risk Areas

4.2 Wind environment simulation and epidemic situation analysis and verification

4.3 Design principles

5. Conclusions

Appendix A

Appendix B

References

文章插图

1. 研究背景

[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure 1. Macau’s location. The three small pictures are, respectively, the geographic location of Macau in China, the location of Macau in the Pearl River Delta, and the Macau Special Administrative Region.
[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure 2. Analysis of prevailing winds and annual wind speed in Macau.

2. 材料准备

[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure 3.Macau wind environment simulation map and COVID-19 heat map. (1) Urban wind environment data simulation; (2) COVID-19 data collection and aggregation.
[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure 4.Wind environment simulation map of Macau Special Administrative Region (including Macau Peninsula, Taipa, and Coloane) and its distribution map of COVID-19 hotspots. Among them, Taipa and Coloane belong to the outlying islands of the Macau Special Administrative Region.
[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure 5.A slice of the Macau Peninsula wind environment simulation map and COVID-19 hotspot distribution map. (1) Partial slice of wind simulation on the Macau Peninsula; (2) Partial slice of COVID-19 hot spot on the Macau Peninsula.

3. 模型训练

[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure 6. CGAN principle flow chart.
[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure 7. Line chart of model training log.
[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure 8. Comparison of the images generated during the training process of Model 1 with the original images. “Epoch” indicates the number of training iterations. This model has been trained for 200 epochs in total, so the test images of the 50th, 100th, 150th, and 200th epochs during the model training process are selected for analysis. “Input” represents the input wind environment simulation. The slice material of the result “Real” represents the actual COVID-19 hotspot image corresponding to the slice. “Generated” represents the image generated by the model through the material of “input”.
[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure 9. Comparison of the images generated during the training process of Model 2 with the original images. “Epoch” indicates the number of training iterations. This model has been trained for 200 epochs in total, so the test images of the 50th, 100th, 150th, and 200th epochs during the model training process are selected for analysis. “Input” represents the input wind environment simulation. The slice material of the result “Real” represents the actual COVID-19 hotspot image corresponding to the slice. “Generated” represents the image generated by the model through the material of “input”.

4. 模型 测试

[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure 10. Different wind environments compared to COVID-19 hotspots. “Input” represents the input wind environment simulation. “Generated” represents the image generated by the model through the material of “input”.
[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure 11. Different COVID-19 hotspots compared to wind environments. “Input” represents the input wind environment simulation. “Generated” represents the image generated by the model through the material of “input”.
[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure 12. Results of model testing using wind simulation data from Macau’s outlying islands. “Input” represents the input wind environment simulation. “Generated” represents the image generated by the model through the material of “input”. The slice material of the result “Real” represents the actual COVID-19 hotspot image corresponding to the slice. “Difference” represents an overlay of “Generated” and “Input” materials. The same parts are black, and the different parts are white to better analyze the accuracy difference between the results generated by the model and the actual results.
[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure 13. Results of model testing using COVID-19 data from Macau’s outlying islands. “Input” represents the input wind environment simulation. “Generated” represents the image generated by the model through the material of “input”. The slice material of the result “Real” represents the actual COVID-19 hotspot image corresponding to the slice. “Difference” represents an overlay of “Gen-erated” and “Input” materials. The same parts are black, and the different parts are white so as to better analyze the accuracy difference between the results generated by the model and the actual results.

5. 统计、分析、归纳、总结

[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure14. Wind speed distribution map of the affected area and the unaffected area
[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure15. A statistical map of wind speeds in areas affected by the epidemic and those unaffected.
[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure16. High-density residential buildings without mountains. (a) Wind environment map; (b) Epidemic distribution map; (c) “C” layout of typical residential buildings; (d) Building space model.
[JCR] Urban Wind Environments on the Distribution of COVID-19
[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure17. High-density residential buildings without mountains. (a)(e) Wind environment map; (b)(f) Epidemic distribution map; (c) “long strip” layout of typical residential buildings; (g) “rec-tangular” layout of typical residential buildings; (d)(h) Building space model.
[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure18. Low-density residential buildings without mountains. (a) Wind environment map; (b) Epidemic distribution map; (c) “L” or “+” layout of typical residential buildings; (d) Building space model.
[JCR] Urban Wind Environments on the Distribution of COVID-19
[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure19. High-density residential buildings without mountains. (a)(e) Wind environment map; (b)(f) Epidemic distribution map; (c) “C” shape layout of typical residential buildings; (g) “rectan-gular” layout of typical residential buildings; (d)(h) Building space model.
[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure20. Analysis of typical residential building types in areas with low incidence of COVID-19.
[JCR] Urban Wind Environments on the Distribution of COVID-19
Figure21. Wind environment simulation and COVID-19 prediction of typical layout of residential buildings.
正文完
更多资讯，敬请关注「走访建筑」微信公众号😘
post-qrcode
SCIE SCOPUS
发表至： 论文成果
2023-07-07