the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 05 Apr 2024
An Air Quality and Boundary Layer Dynamics Analysis of the Los Angeles Basin Area During the Southwest Urban NOx and VOCs Experiment (SUNVEx)
Abstract. The NOAA Chemical Sciences Laboratory (CSL) conducted the Southwest Urban NOx and VOCs Experiment (SUNVEx) to study emissions and the role of boundary layer (BL) dynamics and seabreeze (SB) transitions on the time evolution of coastal air quality. The study presented focuses utilizes remote sensing and in situ observations in Pasadena, California.
Investigations of the synoptic conditions during days when O3 was greater than 70 ppb led to the identification of high pressure conditions and an overall reduction in BL height throughout the day as being primary dynamical factors responsible for enhanced ozone. Enhanced trapping of pollutants at night resulted in reduced O3 and increased NOx (titration), while trapping during the day coincided with a simultaneous decrease in NOx and increase in VOCs that promoted favorable O3 conditions. To evaluate micrometeorological impacts, we selected a day when O3 exceeded 70 ppb during a SB, used empirical mode decomposition to isolate higher frequency variations (micrometeorology), and developed a multivariate spectral coherence mapping (MSCM) technique using the Ricker wavelet to study the role of BL growth and SB transitions on the evolution of air quality measurements. The extraction of time-scales from chemistry and dynamics scaleograms led to a quantitative evaluation of dynamical contributions from BL growth and a SB transition. A statistical evaluation of chemistry data during August 2021 during BL growth supported findings from the case study, but with the caveat of SB interactions with land BL and complex chemical reactions contributing to the scatter distribution caused by day-to-day variation.
- Preprint
(2482 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-447', Ian Faloona, 26 Apr 2024
This work strives to link time scales of variability in the boundary layer to those of the chemical constituents measured (VOCs, NOx, NOy) and develops some interesting and unique methods for doing so (EMD and wavelet scaleograms). Furthermore, the presentation of the air quality data is unique and informative (diurnal and monthly times mapped in 2 dimensions, and conditional sampling based on an O3 peak value threshold.) However, this is a very complex system with a sprawling heterogeneous air pollutant source field and a mountain-plain dynamical system linked to a sea breeze one. The dynamics of precursor emissions and boundary layer dynamics are both extremely complicated in this area. So while there is bound to be much ambiguity about what exactly is controlling what, as there most often is, the manuscript tends to overstate the causes of fluctuating boundary layer depth on concentrations and leaves the exact arrival times of the sea breeze quite unspecified in the context of a continually veering wind throughout each day.
For example, the abstract states, “Enhanced trapping of pollutants at night resulted in reduced O3 and increased NOx (titration), while trapping during the day coincided with a simultaneous decrease in NOx and increase in VOCs that promoted favorable O3 conditions.” However, It is by no means proven in this work that ABL height differences are what caused the differences in overnight NOx and VOCs. In fact, Fig. 4 indicates that the source air overnight for the two conditions (O3_max <, > 70 ppb, Fig. 4) are almost diametrically opposed (~190 vs. ~50 degrees), which in a complex urban environment like SoCal will likely bring in an entirely different chemical composition to the region. Moreover, the difference in overnight ABL depth is quite small and not always easy to measure (it would be helpful to see examples of the DL output over the course of a day and where it identifies the ABL height relative to w-variance signal.) Similarly, the boundary layer height differences in the afternoon could have much more to do with the timing of the arrival of the sea breeze rather than “trapping” emissions. This is even more so true in the case of veering winds consistently changing the advection of these pollutants. Perhaps my overall point can be made by referring to the fact that just as temperature correlates with just about everything in an air quality time series, yet is not the direct causal agent for them all, advection is likely at play influencing the diurnal evolution of the pollutants *and* the boundary layer height, yet it is not the ABL depth, per se, that is causing higher or lower concentrations. Overall, advection is very likely to be a big player with the arrival of a sea breeze front and its lateral mixing with the continental air masses being crucial to understanding the variability.
I think the work is novel and worthy of publication in ACP with the large caveat of many concerns about the analysis outlined below.
General Comments:
Be specific about which BLH data is being shown and used in the analysis (HRRR is hourly, Doppler Lidar is 4/hr). If you are using the higher rate w-variance technique from the DL, then perhaps you could compare it with the HRRR data set to see if the model is doing a decent job of estimating ABLH (I would guess that it is not).
The machinations to develop a vertical velocity scale seems unnecessary: you can use the w-variance measurements to estimate w* from convective similarity theory.
Throughout solar noon is identified as 19 UTC, but solar noon is much closer to PST, not PDT, and is therefore more like 20 UTC. Also note that the length of the day decreases by about 1 hour across the month of August in SoCal.
The exact arrival time of the sea breeze should be better established. I would suggest plotting dew point temperature (or water mixing ratio) which are not directly influenced by air temperature to most clearly indicate the arrival of the marine layer into Pasadena. See Mayor (2011) and Wang & Ullrich (2018). This should be included in Figure 6.
There seems to be a confusion between angular frequency and linear frequency which makes the interpretation of observed time scales slightly occluded. The periodicity in Figure 1b is clearly 2 hours, yet the scalogram is reporting it in Figure 1c as 0.3 hrs., which is probably a factor of 2*pi shorter. However, the physical process that is affecting the “ripple” in ozone has an obvious time scale of 2 hrs. This is important because later on the authors divide by 2*pi to make a derivative of the timescales with BLH match what is expected to be the entrainment velocity.
Specific Comments line by line:
Fig. 1: It is unclear how the periodicity that is so obvious in (b) as ~2 hours, is reported as 0.3 hrs., unless you are reporting the inverse of angular frequency, (Period)/2*pi. This seems important because later on the authors divide by a factor of 2*pi to interpret the time scales changes in a growing ABL as corresponding to an entrainment rate (Eq. 13).
Fig. 1c: Typically wavelet coherence figures include a cone of influence to direct the eye away from the extremes of the figures where the uncertainty in the method is greatest. This might assist the reader in interpreting this and other figures in the manuscript (7-9).
Fig. 2: Solar noon is more closely tied to PST, not PDT, therefore more like 20 UTC (not 19 UTC as specified in the caption.) Also, while Figs. 2 & 3 are interesting ways to present data, it is very difficult to eyeball a correlation with them. For example, to my eye the (anti) correlation between O3 and NOx/NOy seems stronger than with VOC/NOx. Maybe a simple table with the daytime correlation coefficients for all of these parameters would be helpful to the reader to “calibrate” their eyes. It would help to give some quantified sense of proportion when making statements such as, ““…increased NOx during nights that preceded O3 exceedances.” (Line 587)
Fig. 3: Because there is no ABLH data before 8/10, this work presents an anti-correlation dominated by 2 elevated O3 episodes which occur overlapping the presence of ~3 ABLH minima in the same 20 day time period. This is not a very solid correlation. In fact, what seems more interesting is that the O3 events seem to occur on the “falling edge” of a high ABLH period, that is as the BLH_max is decreasing on the synoptic scale. Nevertheless, the limited time series makes such ideas very limited conjectures.
Figs. 2&3: I think it might be more clear and accurate to run the hour along a slight diagonal as the orthogonal date axis increases (upward to the right). When comparing patterns from left to right (the way we read) it really should be done slightly obliquely in time. But that is just a thought, not a strong recommendation.
Line 292: “A large increase in NOx leads to a lowering of VOC:NOx ratio” seems more like a tautology than an interesting point. When the value of the denominator goes up, there typically exists a substantial reduction in the ratio.
Fig. 4: It seems like the main determinant of the high afternoon O3 could easily boil down to which direction the overnight winds are coming from: high O3 is preceded by NE-erly flow that has a lot more VOCs (and potentially many more biogenics from the San Gabriel Mountains) and early a.m. NOx (and PM2.5). This chemical preconditioning gives rise to much greater O3 production throughout the daylight hours.
Fig. 4l: The small difference in afternoon wind direction may be quite significant. The 10-15 degrees greater WDR on low O3 days shows that the Sea Breeze is developing earlier, which is why the T is lower and potentially the advection of precursors has different timing. The longer southerly air is brought to Pasadena during the peak photochemical production hours, the higher the overall O3 will be. Also, just a reminder that simply averaging the numerical wind direction in these plots can be misleading. I am assuming the “average” wind directions are vector averages. Please confirm that is so.
Also, because RH is so strongly dependent on T, I would recommend trying to look at specific humidity or dew point temperature instead of RH. I suspect it would be the best indicator of the sea breeze that there (aside from lower T).
Fig. 4: It would help to put down the N, number of data points, for each, to get a sense of the statistical power of these comparisons when sampled conditionally against the O3 peak threshold.
Line 322: “Increased temperature” should be changed to “increased afternoon temperatures” because the high O3 subsample actually has lower overnight lows.
Line 324: When is wind speed shear reduced? They seem to vary out of phase quite a bit. Also, it would be a lot better of a variable, if you are trying to indicate turbulent production, to calculate vector shear (not the wind speed shear): sqrt[(du/dz)^2 + (dv/dz)^2]
Lines 365-371: The discussion of synoptic details that exist downstream (e.g., tropical cyclone Fred) does not seem all that relevant. On the other hand, the inverted trough is a common pattern in the warm season across California. This is relevant because there appears to be a lot of wildfire smoke from the north all throughout the region. In fact, the Suomi NPP/VIRS Deep Blue Aerosol Type product shows considerable wildfire smoke in the vicinity of Pasadena.
(https://worldview.earthdata.nasa.gov/?v=-126.41006362210211,30.280133300830197,-109.61712583773692,38.709529522001745&l=Reference_Features_15m,Reference_Labels_15m(hidden),Coastlines_15m,VIIRS_SNPP_Aerosol_Type_Deep_Blue_Best_Estimate,MODIS_Combined_MAIAC_L2G_AerosolOpticalDepth(hidden),VIIRS_SNPP_CorrectedReflectance_TrueColor(hidden),MODIS_Aqua_CorrectedReflectance_TrueColor,MODIS_Terra_CorrectedReflectance_TrueColor(hidden)&lg=false&t=2021-08-16-T05%3A53%3A12Z ).
Line 372: I believe that it is very important to be sure that the HRRR output for ABL height is in altitude above ground surface, as opposed to above mean sea level. This should be made clear in the units (m-agl) throughout the manuscript.
Line 387: You should probably define what a convergence line is exactly. The convergence around Pasadena tends to exist most days with onshore flow because of the San Gabriel Mountains on its northern flank which naturally forces horizontal flow convergence in the presence of southerly wind.
Fig. 5: It might be more instructive to not fill in the marker identifying Pasadena, so the color scale can be read within the region. How well does your DL estimate of ABL depth compare with the HRRR output in general?
Line 409-410: Can you explain this suggested mechanism? Is there any evidence of a strengthening inversion in this time? In my opinion, there is no sound evidnece of strong subsidence on this day (despite the qualitative arrows annotated in Fig. 6) and there is no reason to believe that increasing static stability increased the winds at that elevation.
Lines 424-426: I do not understand this argument. What “oscillations” exactly are you referring to? Second, how/why do they appear related? Which characteristics of the ABL are you referring to and how would those influence the “oscillations”?
Fig. 6a: The annotated gray arrows indicate a vertical velocity of -200 m/hr = - 5.5 cm/s (day and night). This is a very strong subsidence rate in the lower troposphere, and it is not clear how they were inferred from profiles of wind speed and direction. What makes you confident that the wind is transported downward in such a manner? Momentum is definitely *not* a conservative tracer in the atmosphere. Regardless, the arrow runs from very low wind speed (blue) to strong wind speed (red). Why are these two locations related as indicated by the gray arrow? Furthermore, any observation of a descending scalar from a fixed point measurement is always subject to the potential aliasing by the horizontal advection of a slightly slanted layer. There is no reason to believe those gray arrows, in my opinion.
Fig. 6c: This is a very unusual Doppler lidar diurnal vertical velocity plot. Typically there are domains of updrafts and downdrafts intercalated every 5-10 minutes throughout the ABL (e.g., Lothon et al., 2009; Maurer et al., 2016). There appears to be no downdrafts observed anywhere in this plot, which violates mass continuity. Furthermore, since the DL measurement is an average fo 11.5 minutes every 15 minutes, there is a chance of aliasing higher frequency components into this dataset. This brings up the fact that it would be reassuring to see some of the data from the DL prior to getting handed over to the more complex mathematical treatments. For example, how do the w-variance profiles compare with the literature, and the boundary layer heights compare to HRRR, etc.
Line 447: However, unlike BLH and VOCs, the NOx periodicity drops down again to shorter values matching O3, not VOCs.
Line 464: “…provided that the velocity of overturning eddies does not change appreciably.” But quite the contrary: it absolutely does! The convective velocity scale is going to increase with increasing surface buoyancy fluxes throughout the day, and while the BLH will also, the former increases at a power of 1/3. Thus the large eddy turnover time scale will be proportional to BLH^(2/3). Using a simple slab model convective boundary layer model (e.g. CLASS, https://classmodel.github.io/) one finds that this time scale increases monotonically over the course of the daytime heating (from ~5 to ~20 minutes).
The Doppler Lidar data should allow for an estimate of w* by convective similarity, and BLH, therefore the large eddy turnover time should be able to be estimated: tau = BLH/w* from the measurements directly.
Line 466: There is so much turning of the winds throughout this day that you are observing many different things affecting your time series via simple advection differences. This is the crux of the problem with the interpretation of all these “covariances”. Differential advection is likely very dominant in this system.
Lines 521-523: Again, this discussion entirely ignores the time scales of horizontal advection and the veering wind which brings in different concentrations which is likely contributing to the variations in the chemical species significantly. Furthermore, when you bring up processes like mixing changing concentrations which change chemical reaction rates, you are blending the transport and chemical reaction terms (all of which are going to have a wide range of time scales: from 10 minute for a reactive VOC and NO2, to half a day for less reactive VOCs).
Line 545: On Aug 16 it looks like the SB did not really fully influence the sampling site until 20-21 UTC (nearly outside of the subdomain you are studying here: 14-20 UTC). I believe it is critical to mark the arrival of the marine layer at Pasadena, and it will likely be most apparent when looking at dew point temperature or specific humidity in conjunction with the other variables (e.g. Fig. 6).
Line 549: Where does this 17 UTC time come from? The August 16 case study indicates the SB arrival time is more like 19-21 UTC (Fig. 6, based on wind direction veering to southwesterly, the direction of the nearest coastline, and the premature fall of the air temperature).
Line 551-553: This discussion seems very speculative. For instance, to eliminate the other variable pairings is to assume that n=2 is a decent account of how they ‘normally’ behave. Further, the subjective grouping of “high frequency” and SB arrival is extremely fuzzy. When does the SB arrive on each day (Aug 16 it looked more like 20 UTC), and what is high frequency? There are plenty of scatter points that are below 1 hr in period.
Lines 554-556: This exercise surrounding Figure 12 seems fraught. BLH and time are going to be strongly correlated in this time interval (in fact, monotonically linked). So these figures (Fig 12) look a lot like the previous set (Fig. 11) just rotated around the x=y line. And the selection of the subset in the red circles seems arbitrary as they do not visually cluster in any noticeable way.
Line 563-566: It seems unlikely that the SB arrives in Pasadena by 8-10 a.m. You can look for yourself with the wind direction, specific humidity, etc. shifts daily. But even so, you are saying you recognize that everything that can affect high frequency changes in a reactive scalar like O3 could be happening. That is true in the most general sense. What type of “dynamical interactions” and “precursor reactions” are being referred to? It might be instructive to inspect the scalar budget equation of these reactive compounds.
Lines 569-570: Why present this data if any associations that a reader is inclined to infer from the figures is always going to be statistically insignificant? I would recommend leaving these small event counts out of your analysis altogether. They are misleading, in my opinion.
Line 574: There are plenty of studies in the literature that reported on these dynamical influences on ozone chemistry in the continental ABL (Wolfe et al., 2015; Trousdell et al., 2016; Kaser et al., 2017; Trousdell et al., 2019). These can at the very least give a sense of proportion when attempting to apportion causes.
Line 578: I would avoid the use of the word “stable” because of its preeminence in buoyancy/mixing. If what you mean is “stationary” (i.e., not time dependent) I would use that term instead.
Line 587: “…increased NOx during nights that preceded O3 exceedances.” The statement raises the question of how much of the high O3 episodes are related to high a.m. NOx, relative to other influences (BLH differences, VOCs, SB arrival time, etc.) It might be instructive to look at the 8 hours when O3 is high (~18 - 1 UTC), and scattering the average O3 with the average NOx and average VOC. What fraction of the day to day O3 variability can be explained by traditional chemistry. It looks like the correlation with VOC will be large. Then also do it with T and BLH. Regardless this gives you a baseline of what is controlling the ozone peaks day to day. And then you can get into the minute to minute to hour variations afterwards. This is just a suggestion.
Line 594: I believe it is very unlikely that the difference of 0.5 m/s at the surface is going to influence BLH. August in SoCal will not typically produce neutral ABLs, they tend to be strongly convective. Surface shear production is not the dominant source of turbulent kinetic energy. Without knowing what the subsidence difference is between low and high O3 days, you cannot suggest that this is a reason the ABL top is lower. The differences in the strength and timing of the sea breeze, which brings lower T air into the region (and lower BLH), is much more likely to be the cause of these modest differences.
Lines 597-599: The greater winds and deeper nocturnal ABLs would lead to increased dry deposition of O3 and NO2 in a thicker layer overnight. This does not necessarily reduce the role of titration, but reduces the next day’s Ox levels. Also the daytime BL heights have very little to do with the ~50 m difference in their initial morning values (Driedonks, 1982).
Line 603: At what elevation is the “observed” wind shift you are referring to? One can find a wind shift in that figure at some elevation just about any time of day.
Line 608-612: The winds are not advected around by the flow in a conservative manner like a non-reactive scalar is. They are strongly influenced by several thermal circulations in this region (varying pressure gradients with height) that are all changing strength throughout the day (upslope southerly flow in the a.m. and southwesterly sea breeze flow later in the day.) “Patterns of descent” suggested by a wind pattern is highly speculative without interrogating the entire Navier-Stokes equation (and also importantly the thermal wind.)
Line 613: The region of higher wind speeds (>5 m/s) above the ABL (~600-1500 m) is more or less continuous throughout the day from the southeast. There is only one period of ~ 1hr near solar noon when the winds accelerate to 8-9 m/s.
Line 614: Bear in mind that one does not need to hypothesize a temporary, thin wind jet atop the ABL to “initiate” entrainment. Entrainment is sure to be occurring vigorously throughout the day because of strong surface heat fluxes in SoCal in August.
Line 628: The NOx and VOC concentrations do not increase, but rather their variance does.
Line 628: I think it is better to be more specific with the wording here: it is not any other “structure” than the ABL height, correct? If not, then specify what “structure” parameters you are referring to.
References
Driedonks, A. G. M. (1982). Sensitivity analysis of the equations for a convective mixed layer. Boundary-Layer Meteorology, 22, 475-480.
Kaser, L., E. G. Patton, G. G. Pfister, A. J. Weinheimer, D. D. Montzka, F. Flocke, A. M. Thompson, R. M. Stauffer, and H. S. Halliday (2017), The effect of entrainment through atmospheric boundary layer growth on observed and modeled surface ozone in the Colorado Front Range, J. Geophys. Res. Atmos., 122, 6075–6093, doi:10.1002/ 2016JD026245.
Lothon, M., Lenschow, D. H., & Mayor, S. D. (2009). Doppler lidar measurements of vertical velocity spectra in the convective planetary boundary layer. Boundary-layer meteorology, 132, 205-226.
Maurer, V., Kalthoff, N., Wieser, A., Kohler, M., Mauder, M., and Gantner, L.: Observed spatiotemporal variability of boundary-layer turbulence over flat, heterogeneous terrain, Atmos. Chem. Phys., 16, 1377–1400, https://doi.org/10.5194/acp-16-1377-2016, 2016.
Mayor, S. D. (2011). Observations of seven atmospheric density current fronts in Dixon, California. Monthly weather review, 139(5), 1338-1351.
Trousdell, J. F., Conley, S. A., Post, A., and Faloona, I. C.: Observing entrainment mixing, photochemical ozone production, and regional methane emissions by aircraft using a simple mixed-layer framework, Atmos. Chem. Phys., 16, 15433–15450, https://doi.org/10.5194/acp-16-15433-2016, 2016.
Trousdell, J. F., Caputi, D., Smoot, J., Conley, S. A., and Faloona, I. C.: Photochemical production of ozone and emissions of NOx and CH4 in the San Joaquin Valley, Atmos. Chem. Phys., 19, 10697–10716, https://doi.org/10.5194/acp-19-10697-2019, 2019.
Wang, M., & Ullrich, P. (2018). Marine air penetration in California’s Central Valley: Meteorological drivers and the impact of climate change. Journal of Applied Meteorology and Climatology, 57(1), 137-154.
Wolfe, G. M., et al. (2015), Quantifying sources and sinks of reactive gases in the lower atmosphere using airborne flux observations, Geophys. Res. Lett., 42, 8231–8240, doi:10.1002/2015GL065839.
Citation: https://doi.org/10.5194/egusphere-2024-447-RC1 -
AC1: 'Reply on RC1', Edward Strobach, 11 May 2024
Hello Ian,
Your review is greatly appreciated. I'm attaching the response document below. I'm unable to submit the revised manuscript at this time, but I do mention where I have made changes as well as arguments in support of key points made in the initial manuscript with supporting analysis.
-
AC1: 'Reply on RC1', Edward Strobach, 11 May 2024
-
RC2: 'Comment on egusphere-2024-447', Anonymous Referee #2, 24 May 2024
This manuscript aims to understand meteorological and chemical variability associated with boundary layer growth and sea breezes. The authors use a novel statistical approach to understand relationships between the observed variability. The topic is suitable for publication to ACP. Two major comments and a few minor comments are below.
Major comments:
1) The paper does not mention nighttime land breezes, but Figure 4l suggests that nighttime land breezes occurred prior to high ozone days, which probably played a role in poor air quality. A discussion on these land breezes on the observed air quality should be worked into the manuscript. See more pertaining to this issue below.
2) While the scaleogram technique is a novel way of investigating the variability of meteorological and chemical variability within the atmosphere (Figures 7-9), it is not clear if combining the maxima of the spectral peaks from the scaleograms during these two different meteorological processes (boundary layer growth and sea breezes) results in statistically significant relationships on variability and boundary layer height (Figures 10-12). The analysis and results associated with Figures 10, 11, and 12 might show more statistically significant results if you only look at the impact of boundary layer growth by not including data points once a SB arrives. At least that is my hypothesis.
Other comments:
Page 1, line 14: Briefly state what the findings from the cast study are.
Page 4, line 4: Capitalize the first letter in this sentence.
Page 5, line 23: Re-word “accessible online at S. (2021).”
Page 15, line 327, Figure 4l, and part of the discussion in the conclusion: The nighttime northerly winds is likely a sign that a land breeze has formed. Including a discussion on the land breeze in the paper would be beneficial. This probably plays a large role in the high ozone days. It results in air pollution to recirculate and stick around in the LA Basin until synoptic scale winds are strong enough to push them over the mountains. The nighttime southwesterly nighttime winds preceding low ozone days suggest pollution is being transported over the mountains. Does the HRRR simulate a land breeze at night? Figure 5 does not cover nighttime hours.
Figure 5: Consider showing a nighttime hour plot. Also add observed surface wind barbs to the figures.
Page 18, line 407: change “BL height, and” to “BL height growth, and”
Page 18, line 417: In addition to entrainment, also detrainment. Consider also mentioning detrainment or BL-free troposphere exchange.
Page 18, lines 418-420: Why do some pulses cause an increase in NOx and decrease in O3 and some cause the opposite?
Page 20, line 456. Change “NOx, or VOCs-Ox did” to “NOx, or VOCs. Ox pairings did”
Figure 10: The NOx lines may look similar to the others if you don’t include data points dealing with the SB. Consider only using data points before the SB moves over the measurement location.
Page 26, lines 541-553: Quantify. Hard to see a trend when looking at any of these figures.
Page 26, line 550: In addition to RH and VOCs, it looks like O3, NOx, and Ox feature high frequency variability spanning this time.
Figures 11 and 12: Removing data points when the SB arrives might improve the analysis and make it easier to see a relationship / trend lines between these variables.
Citation: https://doi.org/10.5194/egusphere-2024-447-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 277 | 111 | 23 | 411 | 15 | 21 |
- HTML: 277
- PDF: 111
- XML: 23
- Total: 411
- BibTeX: 15
- EndNote: 21
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1