the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 27 Jun 2024
Global carbon emission accounting: national-level assessment of wildfire CO2 emission – a case study of China
Abstract. Wildfires release large amounts of greenhouse gases into the atmosphere, exacerbating climate change and causing severe impacts on air quality and human health. Including carbon dioxide (CO2) emissions from wildfires in international assessments and national emission reduction responsibilities is crucial for global carbon reduction and environmental governance. In this study, based on a bottom-up approach and using satellite data, combined with emission factor and aboveground biomass data for different vegetation cover types (forest, shrub, grassland, cropland), the dynamic changes in CO2 emissions from wildfires in China from 2001 to 2022 were analyzed. The results showed that between 2001 and 2022, the total CO2 emissions from wildfires in China were 693.7 Tg (1 Tg = 1012 g), with an annual average of 31.5 Tg. The CO2 emissions from cropland and forest fires were relatively high, accounting for 46 % and 32 %, respectively. The yearly variation in CO2 emissions from forest and shrub fires showed a significant downward trend, while emissions from grassland fires remained relatively stable. In contrast, the CO2 emissions from cropland fires showed a clear upward trend. High CO2 emissions from wildfires were mainly concentrated in the eastern regions of Heilongjiang and Inner Mongolia Provinces in China, accounting for 44 % of the total annual emissions. Various factors such as daily cumulative sunshine hours (Spearman’s correlation coefficient, forest: -0.41, shrub:0.25; p < 0.001) and the normalized difference vegetation index (NDVI; Spearman’s correlation coefficient, forest: -0.35, shrub: 0.37; p < 0.001), influenced CO2 emissions from forest and shrub fires. Moreover, temperature (Spearman’s correlation coefficient, -0.45, p < 0.001) primarily affected CO2 emissions from grassland fires. The CO2 emissions from cropland fires negatively correlated with the gross domestic product (GDP) (Spearman’s correlation coefficient, -0.52, p < 0.001) and population density (Spearman’s correlation coefficient, -0.51, p < 0.001). China's policy management has been crucial in reducing CO2 emissions from wildfires. By accurately assessing CO2 emissions from wildfires, governments worldwide can better set CO2 reduction targets, take corresponding measures, and contribute to the global response to climate change.
- Preprint
(2649 KB) - Metadata XML
-
Supplement
(1665 KB) - BibTeX
- EndNote
Status: open (until 25 Apr 2025)
-
RC1: 'Comment on egusphere-2024-1684', Anonymous Referee #1, 03 Apr 2025
reply
Review of Global carbon emission accounting: national-level assessment of wildfire CO2 emission—a case study of China by Gong et al., 2024
General comments: Wildfire emissions are challenging to quantify but often impact air quality and climate. The study by Gong et al. 2024 estimates China's CO2 emissions caused by wildfires from 2001-2022, finding high emission potential from cropland and forests. The authors use the bottom-up method using MODIS-MCD64A1 burned area product, Chinese LULC (CNLUCC) product, and NDVI product in conjunction with ancillary datasets such as harvesting area map, meteorological variables and a set of literature-based emission factors to assess changes in fire emissions across China for nearly two decades. The study highlights a declining trend of forest emissions while cropland fires have risen over the years. Further, statistical correlations of fire emissions with sunshine hours, NDVI, and temperature are analysed to determine the dominant influencing factors of the forest and shrub fires.
From a research perspective, however, the current version of the manuscript lacks certain major aspects (outlined in detail below), which require careful evaluation of results, such as the comparison with existing inventory and estimations of the emission uncertainty that may arise from both data and methods. The result interpretation and discussion are found weak on a number of occasions, limiting the scientific scope of the article - the study may need to provide more evidence (direct or indirect) to back up the interpretations of the results. The overall structure of the paper could be improved by adding the novelty of the present study and avoiding disjoints (detailed below). Only after adequately addressing these major issues can I recommend this study for a potential publication in ACP.
Major Comments:
As the authors point out, there exists substantial uncertainty in wildfire emissions in existing emission inventories. In the present study, there are no attempts to quantify or report the emission uncertainties. For instance, MODIS products usually miss the small fires (typical for croplands) due to spatial and temporal resolutions, and the classification algorithm can misclassify burnt areas/vegetation types. Without the use of additional information (e.g. ground-truth data) to calibrate and validate the burnt area product/classification algorithm, the estimated emissions raise concerns, and the robustness of the results cannot be examined. I find this to be a serious limitation of the study. As an example, how do those results vary if we use other satellite products (e.g. burnt area from FIRECCI, other NDVI) or ancillary data? What are the associated errors when using derived/literature-based combusting efficiency or emission factors? Combusting efficiency can be influenced by a number of factors, such as the intensity and type of fires, moisture content and load of combustibles, and meteorological conditions, as noted. In that case, how are those estimations impacted when eq. 2 considers only the fraction of vegetation in the empirical relationships? These issues have to be addressed before the manuscript is considered for publication.
Further, the novelty of this study and the value-addition to the existing emission estimate approaches need to be clarified. There are different global fire emission inventories available (e.g. GFED, GFAS). Also, the authors cite a set of national emission inventories in China or the sub-national level. How do the present estimations compare to those? It is unclear how the present approach is advantageous (or even different) from already existing inventory-based methods for estimating fire emissions. Moreover, the current version of the manuscript lacks a comprehensive and comparative assessment of available inventory data, especially when reporting annual/monthly emissions and their trends, as well as relating them to environmental or socioeconomic factors.
Sect.3.4 raises some concerns about reported correlations (detailed below), which need to be addressed/clarified.
Additionally, I have a concern over the title and the use of the term: ‘Global carbon emission accounting’. I do believe that the study is more towards the bottom-up based wildfire emissions from China rather than describing it in the context of global carbon emissions - although it can be linked further as a value addition. Hence, I suggest the authors reconsider the title to represent the present content of the manuscript appropriately.
Comments Specific to Introduction:
An overview of previous/available methods for estimating wildfire emissions and the comparison (or shortcomings) of those methods can be included in the introduction section.
L.37: “The global annual CO2 emissions from wildfires are approximately 6.5 to 11 billion tons” - please check the range (max. value); it seems to be quite large as per current records.
L.62-64: However, CO2 emissions from wildfires are not included in international assessments..” Doesn’t seem to be true. Please rewrite or justify.
Comments Specific to Data + Methods
Currently, Sections 2.2 and 2.3 are disjoined, providing cluttered details of the approach utilized. I suggest combining the Data and Methods sections, which helps the readers to follow the methodology easily.
Also, the resolution of the final product (Ei or Ej) is not clear. From eq. 3, it might be emissions per vegetation cover or per crop type. In that case, how do you address differences in spatial resolutions, given that the vegetation cover is at 250 m, the burnt area is at 500 m, and the harvesting area dataset is at 1 km? Please clarify.
What is Tc in eq. 2?
Comments Specific to Results and Discussion
L.155: “The total CO2 emissions from wildfires in China from 2001 to 2022 were 693.7 Tg, with an average annual value of 31.5 Tg, accounting for 0.46% of the total global emission of All fire types (GFED4)...” I assume that 693.7 Tg is the estimate for this study, and the percentage ( 0.46%) is calculated by considering the total global emission from GFED. How do your estimates compare with existing inventories? What is the contribution percentage of Chinese wildfire emissions according to GFED only?
L.157: Does 0.52% mean net forest fire emissions to fossil fuel emissions? Biomes have sequestration potential, and vegetation regrowth in later years can help offset the impact of wildfires. So, it is not straightforward to compare annual biomass emissions (only) with fossil fuel emissions to assess the biomass contribution to CO2 emissions. How does the number or extent of forest fires compare with the forest regrowth patterns across China? Please clarify.
L.159: “accounting for 46% and 32%”. Please correct it to: accounting for 46% and 32% of the total wildfire emissions in China.”
Fig. 2: Forest fires show a declining trend, but I am not sure how this can be attributed solely to effective forestry management strategies. How about the delays in forest regrowth after the extreme fires? For example, the extreme forest fires occurred in 2003 (around Heilongjiang and Inner Mongolia Provinces) and 2008 (Inner Mongolia), as per Fig. 8. Did this study consider the effects of forest regrowth time?
L.183: “.. policy management.” What about the influence of other environmental conditions, such as crop types, harvesting cycles, crop-specific-suitable temperature, precipitation, water availability, etc?
Table 1: Not clear what additional information is conveyed through spatial autocorrelations. Please clarify.
L211: The hotspot analysis method is not clear, though the results are nicely outlined. Please include additional details on the method.
L245: Please rewrite clearly - there is no downward trend (in Fig. 8d and Fig.3) for Heilongjiang and Inner Mongolia regions after 2012 (after the implementation of China's strict ban on open-air biomass burning) - is there any reason for this?
Sect. 3.4: The use of vegetarian primary productivity and the NDVI in factor analyses is inappropriate. They are already part of the wildfire emission calculation; hence, they must be highly correlated with emissions. Please clarify.
Sect. 3.4: “… daily cumulative sunshine hours (forest:-0.41, shrub:0.25; p < 0.001) …, while the main factor affecting CO2 emissions from grassland fires was temperature (-0.45, p < 0.001)”
How have these correlation coefficients become negative?
L275: “An increase in GDP and population density was often accompanied by better agricultural technology and management practices, including more effective management alternatives to straw burning.” Please backup.
L280: “China's policies have also significantly reduced CO2 emissions from opening biomass burning fires.” This is quite contradictory to Fig. 3
L311: “However, the current international assessment and national emission reduction responsibilities do not include wildfire carbon emissions or consider measures such as reducing wildfire frequency and intensity through wildfire management.” I’m not sure if this is entirely correct.
Comments Specific to Conclusion
L329: “Human activities significantly influence CO2 emissions from cropland fires. Emissions negatively correlated with GDP (-0.52) and population density (-0.51).” These two sentences seem to be contradictory.
L330: “Various factors, such as accumulated sunshine hours (-0.41, p < 0.001) …), mainly influenced emissions from forest and shrub fires, while temperature (-0.45, p < 0.001) primarily affected emissions from grassland fires.” Please see my comments above - why are they negatively correlated?
Citation: https://doi.org/10.5194/egusphere-2024-1684-RC1 -
RC2: 'Comment on egusphere-2024-1684', Anonymous Referee #2, 04 Apr 2025
reply
Comments on Gong et al ‘Global carbon emission accounting: national-level assessment of wildfire CO2 emission – a case study of China’
The study presents a well-structured and timely study on wildfire CO₂ emissions in China, using satellite data and a bottom-up approach to analyze spatiotemporal trends from 2001 to 2022. The topic is highly relevant to climate change mitigation and environmental governance, and the methodology appears robust. However, several aspects could be clarified or expanded to strengthen the manuscript’s impact and scientific rigor.
(1) The authors should reconsider the title of the manuscript. While this study focuses exclusively on wildfire CO₂ emissions in China, the current title, “Global carbon emission accounting: national-level assessment of wildfire CO₂ emission—a case study of China”, suggests a broader global scope. This could be misleading to readers, as the study does not provide data or analysis on global carbon emissions or wildfire CO₂ emissions beyond China. If this is part of a series of publications on global carbon emission accounting with a focus on individual countries, it would be helpful to clarify this in the title. A more precise and region-specific title would better reflect the study's content and avoid confusion.
(2) This study only presents the average CO₂ emissions from wildfires but does not estimate the uncertainties propagating from emission factors and activity data. Without such an uncertainty assessment, readers cannot evaluate the reliability and accuracy of the wildfire CO₂ emission estimates. Consequently, the claim on page 18, lines 306–307, that “wildfire CO₂ emissions provide accurate input data for simulating the effects of wildfires on air quality, climate, and human health” warrants reconsideration. The reviewer suggests to include an uncertainty assessment, which would significantly strengthen the study’s conclusions and enhance its credibility for use in simulation models. In addition, the uncertainty of CO2 emissions can be included in figures, for example, Figure 2, Figure 3.
(3) The reviewer suggests that the authors consider reorganizing method Section 2.3 to improve the logical flow. Specifically, moving Section 2.3.4 (CO₂ emission estimation) to the beginning of Section 2.3, before Section 2.3.1 (Emission Factors), would create a more coherent structure. This reorganization would allow the authors to first present the overall approach for estimating CO₂ emissions, followed by the details of the input parameters, such as emission factors (Section 2.3.1) and activity data (Sections 2.3.2 and 2.3.3). This adjustment would enhance the clarity and readability of the Methods section, making it easier for readers to follow.
(4) In the in-text citations and reference list, the surnames “van der Werf” and “van Wees” are incorrectly formatted, with “van der”, “van”, capitalized. According to academic citation conventions, prefixes like “van der”, “van”, and similar are not capitalized unless they appear at the beginning of a sentence. These prefixes are considered part of the Dutch surname but are not treated as independent proper nouns. Correcting this will ensure the
citations follow established formatting standards and provide proper attribution to the authors.
(5) Page 3, Line 79: Replace “The study results can provide…” with “The results of this study can provide”
(6) Page 3, Line87: Add “geographical” before the “distribution” in the sentence “There are differences in the distribution of cropland, grassland, shrubs, and forests in China.”
(7) Page 4, Line 101: Replace “spatial” with “spatiotemporal” in the sentence “Vegetation cover data were combined with fire area data to extract spatial data, including the time and geographic coordinates of fire occurrence, burned area,”
(8) Page 5, Line 126: The citation for the China Statistical Yearbook should be included both as an in-text citation and in the reference list.
(9) Page 6, Line 137: “Combustion efficiency” should more appropriately be abbreviated as “CE” rather than “CF”.
(10) Page 9, Line 183–185: The manuscript states that “The high emissions of cropland fires in March and April mainly originated from Heilongjiang and Jilin Provinces. The high emissions of cropland fires in May and June mainly came from the Anhui, Henan, and Jiangsu Provinces.” Could the authors explain why the cropland fire emissions peak at different months between the two regions (“Heilongjiang and Jilin Provinces” vs. “Anhui, Henan, and Jiangsu Provinces”)?
(11) Throughout the manuscript, the “p” in p-values should be italicized to align with standard formatting conventions in scientific writing.
(12) Page 11, Line 202: do the authors intent to say “national wide” instead of “global”? Again, the CO2 emission in China is clearly not global. In addition, the spatial autocorrelation analysis (now in the supplement) and Table 1 should be moved to the method section in the main text. Further, the authors state that “the p values were all less than 0.01, with a confidence level of 99%; the Moran's I values were all positive, with a Z score greater than 2.58, indicating a significant positive spatial autocorrelation of CO₂ emissions from wildfires, exhibiting an aggregation pattern in spatial distribution.” However, the terms “positive spatial autocorrelation” and “aggregation pattern in spatial distribution” are not clearly explained in the manuscript. Could the authors elaborate on what these terms specifically mean in the context of wildfire CO₂ emissions?
(13) Page 13, Line 244-245: The manuscript states that “After the implementation of China's strict ban on open-air biomass burning in 2012, emissions decreased, showing an overall downward trend.” However, the Fig. 8d shows the opposite for Inner Mongolia that the cropland CO2 emissions in Inner Mongolia increased strongly from 2012. Could the authors explain why is it?
(14) Page 14, Line 253: The statement “High emissions still existed in Heilongjiang and Inner Mongolia in the east” is somewhat unclear. It is not immediately apparent whether “in the east” refers to the eastern parts of Heilongjiang and Inner Mongolia, or whether it refers to these regions being in the eastern part of a larger geographical context (e.g., China).
(15) Page 16, Line 280: The statement “China's policies have also significantly reduced CO₂ emissions from opening biomass burning fires” requires clarification. Are these policies specifically targeting cropland fires, or do they apply more broadly? If the focus is on cropland fires, the sentence should be revised to clarify. Additionally, it should be noted that, in principle, all wildfires are open biomass burning.
(16) Inconsistency between text and figure. line 295-206 states “Since 2012, following the implementation of policies for air pollution prevention and control, CO2 emissions from cropland fires have decreased (Fig. 7d)”, however, Figure 3d appears to show the opposite trend. Additionally, Figure 7d does not show the temporal variations of CO2 emissions.
Citation: https://doi.org/10.5194/egusphere-2024-1684-RC2 -
RC3: 'Comment on egusphere-2024-1684', Anonymous Referee #3, 08 Apr 2025
reply
Xuehong Gong et al. “Global carbon emission accounting: national-level assessment of wildfire CO2 emission—a case study of China”
GENERAL:
This study analyzes the spatiotemporal variation of carbon dioxide emissions from wildfires in China from 2001 to 2022 based on high-resolution satellite-derived data and emission factors of various vegetation cover types, then further examine the correlation between wildfire emissions and multiple meteorological and anthropogenic factors. Overall, this study makes a valuable contribution by providing a comprehensive, long-term dataset with high temporal and spatial resolution, which enhances our understanding of the evolution of wildfire emissions across different ecosystems in China. Compared with previous studies, the differentiation of emission contributions from forest, shrubland, grassland, and cropland can suggest finely the contribution of different vegetation types. Furthermore, the use of spatial autocorrelation analysis (Moran’s I) is relatively uncommon in existing wildfire emission studies. However, a few minor issues should be addressed before the manuscript can be considered for publication in ACP.
First, the relationship between CO2 emissions and air quality assessment is unclear. While the study focuses exclusively on CO₂ emissions, it repeatedly mentions potential implications for air quality assessment. However, this relationship is not clearly defined in the current version of the manuscript. Since CO₂ is not a traditional air pollutant in the context of air quality standards, the relevance of this link should either be clarified or reconsidered to avoid conceptual ambiguity.
Second, there are some insufficient in interpretation of spatial correlation analysis. The current discussion is limited to high-value clusters. It may also be important to consider spatial outliers, such as high-emission areas surrounded by low-emission regions. These areas could represent localized fire hotspots or regions vulnerable to extreme events, and further discussion would enrich the interpretation of spatial dynamics.
Finally, the mechanism analysis in this study is insufficient. For example, this work attributes a reduction in agricultural fire emissions to China’s 2012 policy banning open-air biomass burning. While this explanation is reasonable for cropland-related fires, the simultaneous decline in emissions from other land types (forest, shrub, and grassland) around the same period is not addressed. Exploring additional factors, such as changes in land management, fire suppression efforts, or climatic influences could strengthen the mechanism analysis.
SPECIFIC
Line 81: Consider revising “air pollution control strategies” to “global warming control strategies”.
Line 275: Lower emissions in high-GDP regions may be due to a smaller proportion of cropland in these areas. To assess the impact of farmland management practices, it is suggested to analyze emissions per unit area in different area.
Table 2: The bottom border line of this table is missing and should be corrected.
Citation: https://doi.org/10.5194/egusphere-2024-1684-RC3 -
RC4: 'Comment on egusphere-2024-1684', Anonymous Referee #4, 10 Apr 2025
reply
General Comments
This manuscript presents a national-scale inventory of wildfire CO2 emissions in China using a bottom-up approach based on multiple observed datasets. In specific, they used MODIS burned area data, vegetation and NDVI data, land cover data, data on farms and harvesting, meteorological data and empirical emission factors. In no cases were direct observations of fires or species emitted from the fires observed, even though there is extensive literature looking at the problem from this perspective. Overall, the topic is both of interest and relevant to the larger community. The authors also did a great job gathering and applying datasets that spanned over two decades.
However, the manuscript suffers from multiple critical limitations that currently preclude publication. In specific, the study lacks methodological novelty and primarily replicates well-established techniques without introducing new approaches or refinements. Key variables—such as emission factors, biomass, and combustion efficiency—are treated as fixed, without uncertainty quantification or regional calibration. The analysis remains largely descriptive, and the use of bivariate correlations (e.g., Spearman) to explore emission drivers is insufficient, especially given the presence of spatial autocorrelation. Moreover, important limitations of the MODIS burned area product are not addressed, particularly its tendency to underestimate small fires in cropland regions, obscured fires, fires occurring under cloudy conditions or wet surfaces both of which are commonly found throughout areas in eastern and southern China. The policy implications presented are not supported by causal analysis or empirical evaluation. In addition, the integration of multi-resolution datasets lacks clarity, raising concerns about data consistency. I have seen photographs published which demonstrate fires in an area of Southern China which does not show up on your map. Uncertainties of this nature need to at least be written about and limitations discussed.
Overall, the manuscript requires substantial revisions in terms of methods, statistical analysis, and interpretation to reach the level of scientific rigor expected for publication in ACP. I am happy to look at a future revision if all of these points are addressed.
Major Comments:
Lines 15–80 (Introduction): The manuscript adopts a conventional bottom-up emissions estimation framework without methodological innovation. The research question is descriptive in nature and lacks originality. Similar approaches using MODIS and empirical EF models have been extensively applied in both national and global contexts.
Lines 110–150 (Methods): The emission estimation follows a static multiplicative model (BA × AGB × EF × CF), which does not account for critical factors such as fire behavior, combustion completeness, fuel moisture, or fire weather variability. The combustion efficiency model, based solely on fractional vegetation cover (FVC), is overly simplistic.
Lines 110–330: The manuscript fails to provide uncertainty ranges or confidence intervals for any of the key input variables (BA, AGB, EF, CF). No Monte Carlo simulation, error propagation, or even basic upper/lower bounds are included.
Lines 270–290 (Figure 10): The authors only apply Spearman correlation to assess drivers of emissions, without controlling for multicollinearity or confounding factors. No multivariate regression, GAM, or causal modeling is attempted, leading to misleading interpretations.
Lines 95–100 (Data section): MODIS MCD64A1 is known to underestimate small fires, fires occurring under cloudy conditions, and fires occurring on wet ground, all of which are commonly found in different places in eastern and southern China. The authors do not validate this data or compare it to higher-resolution alternatives (e.g., VIIRS, Sentinel-2).
Lines 155–265 (Figures 2–9): Results are largely descriptive summaries of emission patterns over time and space. Figures are repetitive and lack analytical depth. There is little attempt to explain spatial heterogeneity beyond surface-level summaries.
Lines 295–310 (Implications): The manuscript claims that policies reduced emissions (e.g., crop burning bans) but offers no empirical evaluation, no event-based analysis, and no causal testing (e.g., Difference-in-Differences or time series break analysis).
Minor Comments:
L120-L125 (Methods): The AGB data for forests/shrubs are sourced from Su et al. (2016) and Yan et al. (2023), which likely differ in methodology and spatial resolution. No effort is made to harmonize these datasets, risking inconsistencies in long-term trends.
L120-L125 (Methods): The NDVI-based exponential model (Gao et al., 2012) may saturate in high-biomass regions, leading to underestimation. The authors do not address this limitation or compare their AGB estimates with field measurements or independent datasets (e.g., LiDAR), casting doubt on the reliability of the method.
L135-L40 (Methods): The combustion efficiency (CF) equations rely on FVC thresholds derived from Hély et al. (2003), which were developed for African savannas. The authors do not validate whether these thresholds are appropriate for Chinese ecosystems (e.g., boreal forests in Heilongjiang vs. arid grasslands in Inner Mongolia and tropical forests in Hainan and Yunnan). This raises concerns about model transferability and regional accuracy.
L220-225 The term "high-confidence hotspot" is used without defining confidence thresholds (e.g., 95% confidence). Clarify how hotspots were statistically determined to be of high-confidence.
L235-240 "Human activities and fire management may affect cropland fire emissions more significantly, resulting in more significant variability..." Replace the second "significant" with "pronounced" for clarity.
L205 (Table 1): While Moran’s I confirms clustering, the authors do not explore why spatial clusters exist (e.g., links to regional policies, economic activities, climate zones, fire spread, etc.). This limits the utility of spatial analysis for informing targeted mitigation strategies.
L285-290 (Table 2): Citations for "Chen et al. (2022)" (GDP data) and "LandScan" (population density) are missing from the reference list. Ensure all data sources are fully cited.
L290 (Figure 10): Spearman’s correlations assume independence of observations, but spatial autocorrelation violates this assumption. The reported significance levels (p-values) may be inflated, leading to spurious correlations. Spatial regression or geographically weighted regression (GWR) should be used instead.
Lines 315–330 (Conclusion): The conclusion restates earlier findings without elevating the discussion. Consider framing in terms of implications for global carbon inventories, LULUCF reporting, or wildfire mitigation strategies.
Lines 360 (References): Several references (e.g., Cao et al., 2004) lack English translations or full journal details. Ensure uniform citation style. Furthermore, while some people may be able to read the article, I am not sure if all of the reviewers can understand it fully.
Citation: https://doi.org/10.5194/egusphere-2024-1684-RC4
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 516 | 162 | 26 | 704 | 45 | 29 | 25 |
- HTML: 516
- PDF: 162
- XML: 26
- Total: 704
- Supplement: 45
- BibTeX: 29
- EndNote: 25
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 195 | 29 |
| China | 2 | 188 | 28 |
| India | 3 | 53 | 8 |
| Germany | 4 | 24 | 3 |
| undefined | 5 | 18 | 2 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 195