City-Scale Air Quality Network of Low-Cost Sensors

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The present study proposed an air quality monitoring network composed of low-cost sensors for enhancing the existing ones employed by the governmental institutions. The author set up such a network in Sarajevo, Bosnia and Herzegovina and measured PM2.5 concentration and relevant meteorological parameters. The system (MAQS) can post-process the bias in the measurement, and visualize the air quality on a map in real-time.

 

General comments: This manuscript has several missing elements:

A.    How many sensors were there at the beginning and at the time the manuscript is written?

B.    How are the location of the sensors selected?

C.   What are the quality control procedures taken to detect and eliminate any possible erroneous measurement?

D.   Evaluation of sensor performance is insufficient: Fig. 3 shows how good the system can be, but it is important to know the situations that it does not work well. Although it is briefly mentioned in L118-120, it is better to specify such cases in Fig. 3 or another figure as well.

E.    Do the measurements from the MAQS comparable to those from the governmental network? Although it is clear that the sensors of two networks are not installed at the same location, it is necessary to know whether the measurements from the two networks agree to a certain degree as I assume the measuring equipment from the governmental institutions should have a better fidelity than low-cost sensors. I recommend the author using a couple of pollution episodes to illustrate this with the PM2.5 time series of a governmental monitoring station and a (or a few) MAQS sensors in a neighboring area. The current manuscript shows too less example.

F.    The author should write a paragraph to discuss the current limitation of the MAQS in the conclusion.

 

Considering all the problems I observed, I recommend a major revision. The author needs to address the above issues as well as the major and minor questions below during the revision.

 

Major comments/questions

1.   The title is not specific enough. In the study, the measurement of air pollutant is limited to PM2.5, and the sensor deployed only works for PM2.5 as well. Therefore, I suggest something like “City-scale Air Quality Network of Low-cost Sensors: an example of PM2.5 measurement in Sarajevo, Bosnia and Herzegovina”.

2.   L1: Does “possibly global” mean you expect the low-cost sensors that you developed can be potentially applied across the world? If true, it is better to mention it at the end of the abstract because your study shows that your system has only been applied to a city. In addition, if the author thinks the MAQS has the potential to be applied globally, it should be discussed (e.g., what steps are needed to make it happen) in the conclusion section.

3.   L6-7: In this study, the measured PM2.5 concentrations are not compared to those from the governmental institution. I recommend the author to add this part to the manuscript.

4.   L10: Remove “(cable car)”. It can give the illusion to readers that the sensors are installed to the body of the cable cars.

5.   L19: Do you want to say that other air pollutants are at a relatively “acceptable” level compared to PM? For me, it is hard to accept that air pollution is acceptable. I would suggest phrases like “within safe exposure level”, “below a certain limit”.

6.   L24 and Table 1: PM2.5 in this location probably does not follow a normal distribution, so the annual average is not informative for showing the air quality of the location. Other indicators like exceedances, percentiles (or boxplot) can be better choices in this context.

7.   Fig. 1: Is your faculty, where the data in Table 1 was measured, in the figure? If yes, it is better to indicate the location.

8.   L34-35: This sentence makes a comparison between PM2.5 concentration and AQI, which is an inappropriate and unfair comparison. The author should compare measured PM2.5 with a PM2.5 standard. Moreover, it is more appropriate to quote local/national/WHO guidelines instead of that of the USA.

9.   Fig. 2: I am not sure whether the author should draw the double-headed arrow between the database and internet. Is the database inside the server in the actual setting? From this figure, one may infer that the clients can access the database directly without communicating with the server, which does not make sense.

10.    L61-66: Are both generations still in use? Or is Generation 1 phased out?

11.    L127-128: How is the calibration coefficient of each sensor determined? Is it computed based on the data collected during the campaign mentioned in L115? Is the coefficient the same over time without modification? The author should describe the computation of the calibration coefficient more clearly in the manuscript.

12.    Figs. 5 and 6: Due to the small number of sensors installed across the city, most of the grid points in the pollution map are far from any sensor, so the associated values in the map are not representative of the PM2.5 concentration at their locations. More sensors are required to give a better picture of air pollution in the city. Alternatively, numerical model output can be used for estimating the PM2.5 distribution over the area where observation is missing. These points can be added to the conclusion as possible improvement of the pollution map.

13.    Fig. 6: How is AQI calculated? AQI is calculated by considering multiple pollutants concurrently, but your sensors only measure PM2.5. Does it mean that AQI is calculated based on PM2.5 only?

 

Minor comments:

14.    L5: Remove “a city of”.

15.    L9: Remove “many more”. You only need to highlight the major advantages in the abstract.

16.    L17: Remove “concentrations in the air”.

17.    L161: “as fast as possible” should be written as “as frequent as possible”.

18.    L163-164: It is more appropriate to move “We have collected more than 60 million database records so far. This database is very valuable asset for research of air pollution” to the conclusion.

19.    L197: “Periferal” should be written as “peripheral”.

20.    The date format is inconsistent throughout the manuscript. Please write the date in the same way.

Comments on the Quality of English Language

See comments and suggestions for authors.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The paper presented an air quality measurement system based on low-cost PM2.5 sensors in the city of Sarajevo. Co-located measurement from the Low cost PM2.5 sensor PMS5003 was found to match the filter-based gravimetric method very well except under high humidity conditions during which hygroscopic growth correction is needed. The paper also presented drone based measurement of temperature, humidity and PM2.5 concentrations, which is very interesting results. In addition, the paper also generated interpolated concentration field for the city based on the low-cost sensor measurement. The paper also described how data from measurement is transmitted and stored. A very well written paper, I recommend its publication. There is only one comment. In the abstract, it says "particulate metal concentrations in real-time". It should be "particulate matter concentrations", correct.    

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This manuscript addresses the problem of air pollution monitoring in urban environments. This topic is very relevant for monitoring air pollutions in large cities. The author presents a new system called MAQS for high-resolution spatial and temporal monitoring - a city-scale (possibly global) air pollution network consisting of low-cost sensors for particulate matter concentration in the air.

The followings are some comments.

1. The article describes the MAQS system in detail, but does not provide a review of similar systems. The only mention is on line 156: “MAQS is not the only project of this type. For example, PurpleAir network is often discussed”. The author needs to insert a “Related works” section, because in fact there are many similar systems. It is necessary to make the analytical review to show the advantages of the MAQS system.
2. Line 53-55: In order to test the scalability of MAQS server code, a simulator was created to test the performance of MAQS server under the concurrent activity of many thousand sensors, and the results were outstanding: MAQS server worked well even on Raspberry Pi 3 single board computer (i.e. with very limited hardware resources). Please explain what the phrase "server worked well" means? What criteria were used here to evaluate the work - data transfer speed, reliability?
3. There is no description of the algorithm for placing sensors - by what principle they are placed. Is the algorithm for placing sensors related to the location of fixed (governmental) stations? The author writes (Line 250): “This system proved its capabilities to augment sparse governmental network of monitoring stations”. How do they interact –governmental network and sensor network?
4. Line 67-68: Communication between sensors and server is implemented using our own protocol and TCP/IP connection.  It seems to me that at this point it is necessary to describe in more detail the algorithm of own protocol.
5. The author writes that the MAQS system works with a large number of sensors. Are node clustering techniques used to optimize data collection from a large number of sensors? Please write about the methods used for collecting data from sensors and transmitting them to the server.

Despite these comments, I believe that the article describes an interesting system for high resolution spatial and temporal monitoring of air pollution in urban environments.  I think that after making corrections, the article can be accepted into the journal.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Please see the attachment.

Comments for author File: Comments.pdf

Comments on the Quality of English Language

Please see the attachment.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf