How Human Mobility Models Can Help to Deal with COVID-19
Round 1
Reviewer 1 Report
The manuscript presented a work on studying how Opportunistic Networks mobility models can be applied to fight against COVID pandemic by measuring the exposure risks from spatial and temporal perspectives. They want to reuse OppNets to evaluate the opportunity of transmitting coronavirus when two individuals are close to each other. They claims the results obtained using models equipped with pedestrian simulator are realistic. The topic is within the scope, the idea is interesting, and the work is important. Here are some suggestions for improvements:
- My major concern is its feasibility in real-world scenarios. Comparing to the simple practical statistics-based risk estimation such as COSRE score (Sun et al 2020), what is the advantage and disadvantage of the proposed approach?
Sun, Ziheng, Liping Di, William Sprigg, Daniel Tong, and Mariana Casal. "Community venue exposure risk estimator for the COVID-19 pandemic." Health & place 66 (2020): 102450.
- In the case of United States, surveillance and control of all the traveling people is unrealistic. If participation in contact tracing programs is not mandatory, we can only get partial and very likely biased datasets. Based on such dataset, will this approach yield the risk estimation with a similar accuracy and frequency?
- There are two possible fallacies within the approach and authors need rethink and clarify. If I didn’t comprehend it wrong, the proposed risk estimation need know the real-time density of emitters (COVID-infection sources) in the plaza and subway. That is impossible to do in real world, right? If we know who is COVID-positive and infectious, they will never be allowed to be in plaza or subway station. Another potential fallacy is the authors mention “based on these results, health authorities could recommend pedestrians to avoid these risky spots in order to reduce the opportunity of contacts..” If everyone avoid the spots, the spots would be the safest COVID-free place, and some other places will become new hot spots. The statement about how the results can be used should be rewritten.
- The equation of exposure risk was put in the background section, but has no references. Is this equation originally invented? If no, references are needed, and if yes, clarification is needed.
- The paper is too long and spent many paragraphs talking about the basic synthetic models which are impossible to use to fight COVID. Those sections need be simplified or pruned.
- Although the work aims at fighting COVID pandemic, no COVID data is used in the experiments. The real measurements were taken on Jan 23. COVID has significant impacts on the mobility pattern of people and the authors should discuss more on the pattern in the pandemic instead of pre-pandemic.
Minor:
“the irruption of the COVID-19 pandemic” -> “the eruption of the COVID-19 pandemic”
Author Response
Please, see attached file.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Paper is very extensive and detaily describes 3 pure synthetic and 2 pedestrian models generated for real urban scenarios with considering very simplified qualities of the environment (medium).
However, its basis should not be considered as relevant. The authors used quite obsolete observations from January 2020, when the social mobility and distancing behaviour were still not modified by later constantly repeated advices to keep a distance. In addition, current researches about Covid-19 dissemination confirms that this pandemic unlike the common contagious diseases is disseminated in clusters not linearly. More appropriate approach should not simulate quick occasional contacts but massive social events with super disseminators.
The simulation scenario of metro in Valencia is also not very appropriate. As the authors mention its frequency is from 5 to 10 minutes, that means it does not fulfil the condition of exposure duration of at least 15 minutes (according to the classification of ECDC and CDC).
As the authors also mentioned on page 13 (lines 455-456) the proposed approach take into consideration the only linear spread of the virus with no additional factors that have an important impact on the spreading (no super disseminators, no social relations between pedestrians, individual biological characteristics etc.).
Mayor part of conclusions is very general. Authors try to identify risky spots as if the risk level would be only an attribute of the space itself. But distancing is a highly individual, personal and spatial-independent approach. As we all have recently seen it can be significantly changed during a few days.
The applied approach is not very contributive from the point of view of understanding the spatial and temporal spreading of Covid-19. There were too many factors omitted (it is also recognized by authors).
On the other hand, the authors precisely compared 5 different human mobility models, calculated individual interpersonal exposures, density, usual pedestrian paths and local attractors.
Author Response
Please, see attached file.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The authors address all my concerns. Congrats!
Author Response
Thank you again for your kind revision.
Reviewer 2 Report
A significant part of the reviewer´s comments was taken into consideration by the authors and to a greater or lesser extent was included in the new version of the paper.
My main remark is that the research is based on obsolete data that are not valid anymore. Measurements were realized in January 2020.
In the new version of paper, authors stated that there is no significant change in the number of passengers transported by metro in the city of Valencia (lines 339: „Although this data was taken before the339 COVID-19pandemic, the current use of the subway has not significantly changed.“). That sounds strange as several studies demonstrated that Covid-19 pandemic had an enormous influence on the use of public transport. In the majority of European countries, there was identified about 70% reduction in the use of public transport. In my opinion, this argument was not reflected in the paper.
According to the information that metro valencia is providing on its webpage there were also significant decreases (p. ej.46 % decrease in the number of people using the metro in November 2020 compared to the same month of the previous year - https://www.metrovalencia.es/page.php?page=200&id=4412). I would like to ask the authors for clarification and also for providing sources of the information in the case of such important data that are directly influencing the results of their research.
I strongly recommend to update the research and use the data from the pandemic period. For the publication of the paper, it is crucial to solve this issue.
If there is no possibility to use new real measurements, I recommend that the authors provide at least basic comparison between normal evolution and COVID-19 period data with an objective to create some variable/s that will make possible adjustment of inputs related to the correct use of the proposed model.
In the introduction section, there is used almost the same text two times (paragraphs in lines 13-17 and 18-22).
The topic of the paper is actual and important and its title is tentative, so there is a huge probability that in the case of the paper is published it will have many reads. This is very compromising and it means that special attention should be paid to provide really high-quality paper. For this reason, the major revision of the manuscript is still needed.
Author Response
Please, see the attached document.
Author Response File: Author Response.pdf
Round 3
Reviewer 2 Report
I appreciate that the authors have retaken in site measures and repeated the experiments. With updated information paper provides a COVID-19 related scenario and is fulling the expectations which arouse from its title. In present form, the paper had big potential to be of high readers’ interest.
