Pre-monsoon air quality over Lumbini, a world heritage site along the Himalayan foothills

Rupakheti, Dipesh; Adhikary, Bhupesh; Praveen, Puppala Siva; Rupakheti, Maheswar; Kang, Shichang; Mahata, Khadak Singh; Naja, Manish; Zhang, Qianggong; Panday, Arnico Kumar; Lawrence, Mark G.

doi:https://doi.org/10.5194/acp-17-11041-2017

Articles | Volume 17, issue 18

Atmos. Chem. Phys., 17, 11041–11063, 2017

https://doi.org/10.5194/acp-17-11041-2017

© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 3.0 License.

Special issue: Atmospheric pollution in the Himalayan foothills: The SusKat-ABC...

Atmos. Chem. Phys., 17, 11041–11063, 2017

https://doi.org/10.5194/acp-17-11041-2017

© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 3.0 License.

Articles | Volume 17, issue 18

Research article 18 Sep 2017

Research article | 18 Sep 2017

Pre-monsoon air quality over Lumbini, a world heritage site along the Himalayan foothills

Dipesh Rupakheti et al.

Download

Final revised paper (published on 18 Sep 2017)
Supplement to the final revised paper
Preprint (discussion started on 17 Jun 2016)
Supplement to the preprint

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Review of Rupakheti et al.', Anonymous Referee #1, 21 Jul 2016
RC2: 'Review', Anonymous Referee #2, 07 Nov 2016
RC3: 'Review', Anonymous Referee #3, 09 Dec 2016
AC1: 'Compiled Response to the reviewers comments', D. Rupakheti, 10 Feb 2017

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Dipesh Rupakheti on behalf of the Authors (10 Feb 2017) Author's response Manuscript

ED: Referee Nomination & Report Request started (09 Mar 2017) by Kim Oanh Nguyen Thi

RR by Anonymous Referee #4 (02 Apr 2017)

Suggestions for revision or reasons for rejection

Review of Rupakheti et al.

The authors present PM, BC, CO and O3 concentrations measured at Lumbini during April-June 2013 and explained meteorology, pollutant concentrations by conducting WRF-Stem model simulation. They also estimated the regional contributions of CO and aerosol composition to local air quality based only on the model simulation results. This reviewer full agree that the presented observational data set in this study are unique and very useful to understating the level of air pollution in the study area. However, in this revised manuscript, there are several important issues on model simulation and scientific discussion. Therefore, this revised manuscript cannot be accepted in its current form. Before publishing in ACP, several points should be clarified. Specific comments and suggestions are below:

L68, Fig 1: Information given in Figs 1, 2 and 5 are overlapped. This reviewer recommends to merge Fig 1 and Fig 5. That is, plot both monthly mean AOD and winds for separately in April, May and June. Those figures will give more direct insight on aerosol distribution and regional-scale circulation during the intensive measurement period. Fig 2 is not necessary in main body text, so please move it to the supplement.

L68, Fig 1: Which version of MODIS TERRA data have used? Why the authors are not used MODIS Aqua data?

L54-139: This reviewer recommends to reduce the length of INTRODUCTION section with deleting sentences are not closely related the topic of this paper. For example, the authors emphasize many times in the paper that Lumbini is a UNESCO world heritage. This is interested to the authors, but not to all readers. So please minimize the statements on this.
L 99 ~ : As the authors mentioned, sulfuric acid is more critically important in historical heritages. Please more carefully and clearly explain why the air pollutants presented in this study is important should added in the INTRODUCTION.

L128: remove “Aerosol optical depth – not discussed on the present study”. This is not necessary here.

L 165: Table 1 - The sampling period should be more clearly clarified. As shown in Fig. 6, all instruments had not properly operated during the study period.

L169: What’s the uncertainty of PM concentration measured with GRIMM EDM164? Especially the quantitative uncertainty of EDM164 for such a high PM concentration level should be discussed, because PM concentration in here is estimated from light scattering measurements.
Figures 3 & 6 : There is large differences between observations and model simulations. First, more specific explanation and discussion on why the model results were not well agreed with the observations must be addressed. Why the WRF-STEM simulation cannot well simulate the precipitation events and why there is big difference in RH, WD and WS. Why BC is too underestimated compare to the aethalometer data? This should be made for all parameters. This is very important to convincing the results given in Section 3.3 and Section 3.4, as the authors mentioned in L261-262.
Figure 3: WD should NOT be plotted with solid line, because, for example, WD at 355 and 5 degree is almost the same direction. So make a plot with dots.
L242-243: It is hard to agree to this argument. As shown in Fig. 3, there is large difference between modeled and observed wind speed.

L258-259: This reviewer also cannot agree to this conclusive sentence. Apparently, the observed RH is two times higher than the modeled one. There is no evidence that RH by model is how well captured the regional variation. Is there a reference data to back this up?

L265- Figure 5: How about the winds at 850 hPa or 700 hPa pressure level?

L266-267: Here, what “calm winds” means? This discussion in Figure 6 is conflict the winds discussed in Figures 3 and 4. Please clarify.

Figure 6: As commented above, more explanations are needed why there is a large discrepancy between modeled and observed values. Without clarifying this, the results given in the next sections (sections 3.3.2. and 3.4) are not truly reliable.

L384-385: This reviewer understand the PBL height observation was not available during the measurement period. However, the modelled PBL height has also large uncertainty and not believe it. The authors cited several previous works, but need to add some information on the PBL height, not general seasonal characteristics.

L407-408: The authors mentioned ‘Global Monthly Fire Location Products’ were used. However, daily data were used in Figure 9. Please clarify this.

L416: Clearly present how much higher? This is very vague sentence.

L421: Quality of Figure 10 is very bad. It’s hard to read.

Figure 10: How the authors get the modelled biomass CO concentration? Generally the modelled biomass CO concentration is not capture the observed CO concentration. What is the major reason that the author gusse? This reviewer cannot agree to the sentence given in Li435-436.

Figure 10 (I) and (J): Instead of wind roses, regional-scale wind patterns will be more helpful to understand the transport in the interested region.

L441-450 and Figure 11: The two ozone peaks were possible contributed by local pollution, induced by NO2, but also by the transport from the fire plume. However, satellite NO2 data shown in Figure 11 is not direct evidence of the effects of fires on high ozone concentration. Clarify this.

L455 and Figure 12: It should be provided how the authors calculated the contributions from different conturies? Can you provide PSCF results to back this up? In addition, have the authors estimated other pollutants (i.e., PM2.5 and PM10) for their contribution like CO?

Figure 14: The authors only showed wavelength dependency of normalized light-absorption coefficients for two time periods. First question is why the authors do not show all times? The authors can present with time (x-axis), wavelength (y-axis) and normalized light-absorption coefficient with different color. This figure will be more helpful to understand the difference of light-absorption coefficient in around 380 nm wavelength. Second, the difference at 380 nm in current Figure 14 is statistically significant? And how many data points were used? Last question for Figure 12 is that data during the fire periods discussed section 3.2 were included or excluded here?

Hide

RR by Anonymous Referee #3 (01 May 2017)

Suggestions for revision or reasons for rejection

Overall comment: The authors have improved the MS significantly, especially the monitoring data analysis and interpretation. The major uncertainty however still remains with the modeling part and it was, as pointed out by the authors, mainly due to the emission input data.
Major comments:
• It is not clear how authors extracted/projected the emission provided by EDGAR-HTAP_v2 for the simulation period and how the emissions were segregated temporally (hourly, daily etc.) for the simulation.
• One of the reason of the discrepancy between the modeled and monitored levels is the point-based monitoring as compared to model produced grid average values. However, the model significantly underestimated all species, especially PM. Even CO levels were not reasonably produced as shown in Figure 6, the two events were not reproduced that well (as stated in line 509) as the modeling results appear to be fluctuating continuously during the period.
• Why modeled PM levels are not presented in Figure 6? It is suggested that authors include scatter plots to show the relationship between the monitoring and modeling results for each species in Figure 6. Due to the uncertainty in the model output, all the results and discussion based on the model results may be questionable, i.e. those discussed in Section 3.3.2 (line 524).
• Section 2.3: provide the reasons why this particular model was selected, i.e. if it performs better than other models for the region etc., and how the emission input data was prepared for the modeling period.
• Section 3.1: WRF overestimated temperature and wind speed, underestimated precipitation and RH. Common statistical measures should be used to assess the model performance. For wind direction, because of the circular scale of the measurements (near 0 and near 360 degrees are almost the same) the interpretation of the time series should be made with caution or should be avoided. The comparison should be made for different wind sectors or simply by comparing the modeled windroses with the observed windroses to be presented along in Figure 4.

Suggestion: Since the purpose of using WRF-STEM model was “...to understand pollution source region as well as the contribution of open biomass burning to air quality in Lumbini...” as stated in lines 126-127, it is suggested that authors can use alternative ways of the data analysis to achieve the same aims. For example, analysis of wind field (such as Figure 5) or HYSPLIT trajectories and source locations (hotspots, urban etc.) to show the potential of regional transport of the biomass smoke to the site.

Referee Report: PDF

Hide

ED: Reconsider after major revisions (01 May 2017) by Kim Oanh Nguyen Thi

AR by Dipesh Rupakheti on behalf of the Authors (03 Jul 2017) Author's response Manuscript

ED: Reconsider after minor revisions (Editor review) (25 Jul 2017) by Kim Oanh Nguyen Thi

Dear Authors,

I have carefully consider the last version of the MS and responses to all reviewers comments. There are some minor revisions needed before the paper can be accepted for publications. Please see the comments below and also those directly inserted in the PDF (the comments are included in the box of "replace text" and the parts need authors' attention are highlighted in yellow color.

- The MS would benefit much from the improvement of written language to provide sharper discussion of the findings
- The authors’ efforts to compare the results with other studies are appreciable. However, in some places too much comparison without highlighting implications of similarity/difference would dilute the findings.
- Please elaborate the emission input data for the modeling task, i.e. how the annual emissions (and which year) were interpreted into necessary temporal distributions for the modeling purpose. Please note that the comparison between observed and modelled in this MS is to test the model performance and it is not for the model VALIDATION. Hence, it is the model “performance” for this case to be discussed and not the model itself (i.e. not the validity of physico-chemical processes incorporated in the model development).
- The altitude of the model layer, which provided the simulated results for comparison with the observation data, was not given in the MS.
- Other minor comments are directly provided in the PDF of the MS. Please note that the comments are inserted in the boxes of the “replace text” function and the MS content needs attention is yellow-marked in the text, references and figure caption.

Comments to the Author: PDF

Hide

AR by Dipesh Rupakheti on behalf of the Authors (04 Aug 2017) Author's response Manuscript

ED: Publish as is (11 Aug 2017) by Kim Oanh Nguyen Thi

Short summary

For the first time, atmospheric composition was monitored during pre-monsoon season of 2013 at Lumbini (UNESCO world heritage site as birthplace of the Buddha). PM and O₃ frequently exceeded WHO guidelines. Pollution concentration, diurnal characteristics and influence of open burning on air quality in Lumbini were investigated. Potential source regions were also identified. Results show that air pollution at this site is of a great concern, requiring prompt attention for mitigation.