Significant contrasts in aerosol acidity between China and the Unites States

Aerosol acidity governs several key processes in aerosol physics and chemistry, thus affecting aerosol mass and composition, and ultimately the climate and human health. Previous studies have reported the aerosol pH separately in China and the United States, implying a different aerosol acidity between these two countries. However, underlying mechanisms responsible for the pH difference are not fully understood, limited by the scarcity of simultaneous measurements of aerosol 15 composition and gas species, especially in China. Here we conduct a comprehensive assessment of the aerosol acidity in China and the United States, using extended ground-level measurements and regional chemical transport model simulations. We show aerosol in China is significantly less acidic than that in the United States, with pH values 1–2 units higher. Based on a multivariable Taylor Series method and a series of sensitivity tests, we identify several major factors leading to the pH difference. Compared to the United States, aerosols in China are generally in total ammonia (TNH3=NH4+NH3) rich 20 conditions where particle phase ammonium (NH4) concentrations are adequate enough to nearly neutralize major acidic inorganic anions such as sulfate, nitrate, and chloride, leading to a higher aerosol pH. Higher relative availability of the stronger acidic component, sulfate, compared with the weaker acidic component, total nitrate (TNO3=NO3+HNO3), also contributes to the lower aerosol pH in the United States. As a response to higher aerosol pH, the higher nitrate to sulfate molar ratios in China indicates a nitrate-rich condition, further leading to higher aerosol water uptake which will continually promote nitrate aerosol 25 formation. Considering the historical emissions trends, the difference in aerosol acidity between these two countries is expected to continue as SO2 and NOx emissions are further controlled. The differences in aerosol acidity highlight in the present study imply potential differences in formation mechanisms, physicochemical properties, and toxicity of aerosol particles between China and the United States. 30 https://doi.org/10.5194/acp-2020-879 Preprint. Discussion started: 10 September 2020 c © Author(s) 2020. CC BY 4.0 License.

and emissions is consistent with the year of the measurements in each country (i.e., 2017 for China and 2011 for the United States).
In order to evaluate model performance against observations, we calculate normalized mean bias (NMB) and normalized rootmean-square errors (NRMSE) to evaluate the spatial variation of pH, species concentrations and partitioning ratios with 125 following equations, where Cm is the CMAQ-modeled value, Co is the observational value, N is the number of simulation-observation pairs used in NMB and NRMSE calculations.

Aerosol pH calculation
In this study, we use the ISORROPIA-II thermodynamic model (Fountoukis and Nenes, 2007) to determine the composition in a K + -Ca 2+ -Mg 2+ -NH4 + -Na + -SO4 2--NO3 --Cl --H2O aerosol system under equilibrium condition with gas phase precursors. 135 Aerosol pH is calculated based on H + air and liquid water content uptake by inorganic species ( where γH + is the activity coefficient of hydronium ion and is assumed to be 1 in this study (note that, the binary activity 140 coefficients of ionic pairs, including H + , are calculated in ISORROPIA-II), H + aq (mol•L -1 ) is the hydronium ion concentration in the aerosol liquid water, H + air (µg•m -3 ) is the equilibrium particle hydronium ion concentration per volume air. LWC includes LWCi and LWCo, which means the water uptake by inorganic species and organic species (µg•m -3 ) are modeled separately because both organic and inorganic species are hygroscopic. In this study, we only consider the effect of LWCi since the effect of LWCo on aerosol pH has been found to be minor (Guo et al., 2015). 145 There are two modes in ISORROPIA-II's calculation, i.e. forward mode and reverse mode. In the forward mode, the inputs include total concentrations (i.e. gas+aerosol) of TNH3, TNO3, TCl (HCl+Cl -), SO4 and NVCs and meteorological parameters (temperature, relative humidity); in the reverse mode, only the aerosol phase of compounds and meteorological parameters are needed (Fountoukis and Nenes, 2007). In this study, the ISORROPIA-II model is run in the forward mode for aerosol in https://doi.org/10.5194/acp-2020-879 Preprint. Discussion started: 10 September 2020 c Author(s) 2020. CC BY 4.0 License. metastable state because reverse mode has been reported to be more sensitive to measurement errors (Hennigan et al., 150 2015;Song et al., 2018).
We also find that there are measurements with unrealistically high Ca 2+ concentrations (such that Ca 2+ is more than LWC×0.002, i.e., the solubility of Ca 2+ in aerosol liquid water). This may be due to the measurement method of Ca 2+ which needs to use large amount of water to dissolve filter-based particles. This process will likely dissolve the water-insoluble part of Ca 2+ in aerosols which may cause higher bias of aerosol Ca 2+ concentration. In the existence of aerosol SO4 2-, Ca 2+ precipitates along 155 with SO4 2as CaSO4 because of the low solubility (Seinfeld and Pandis, 2006 We evaluate the model performance by comparing the gas-particle partitioning of semi-volatile compounds between measured and simulated values such as ε(NO3 -) and ε(NH4 + ). This method is effective when the species have substantial fractions in both gas and particle phases (Guo et al., 2017a). The comparison results of ε(NH4 + ) and ε(NO3 -) are shown in Fig. S3. The correlation coefficients and the slopes of linear regression are all close to 1, suggesting good agreement between the simulations and observations. In terms of these partitioning ratios, the model performs better in the United States than in China, which may 165 attribute partly to the more even partitioning of the species between gas and particle phase in the United States.

Multivariable Taylor Series Method (MSTM)
In order to separate the contribution of each component (8 species in total, include Na + , SO4, TNO3, TNH3, TCl, Ca 2+ , K + , and Mg 2+ ) to the pH difference between China and the United States, we use a multivariable Taylor Series Method (MTSM). First, we derive the average conditions (i.e., species concentrations and meteorological conditions) across all the sites in the United 170 States and China. We then use the United States as the starting point and China as the end point and decompose the contributions of individual compounds to the pH difference based on the following equations:  (7)  180 where subscript i denotes a specific species; ci,China and ci,US represent the concentration of compound i in China and the United States, respectively; ∆ci is the difference in ci between China and the United States; ci,λ is an intervening ci between ci,China and ci,US defined by λ∈ [0, 1]; ci,λ is ci,US when λ is 0; ci, ,λ is ci,China when λ is 1. In this study we assume negligible interaction between species, therefore the increasing concentration of species i will not have the effect of changing the concentration of other species. The pH difference between China and the United States (i.e., ∆pH) can be expressed as the sum of the partial 185 derivatives of pH with respect to ci,λ which is then integrated from ci,US to ci,China , as described by Eq. (6). In this study, we take 100 steps with equal intervals to gradually change λ from 0 to 1 (Eq. (6)) and record the partial derivatives of pH with respect to individual ci,λ, and derive the contributions of all the species and meteorological variables to the pH change at every step. By summing up the contributions at all the steps, we characterize the contributions of individual components to the overall pH difference (Eq. (7)). 190

The pH difference based on observation
The aerosol pH values calculated based on observational data show a significant difference between China (most observation sites in NCP) and the United States. In China, the 2017 annual average pH level is 4.3 and ranges from 3.3 to 5.4 by monitoring 195 sites with an interquartile range of 3.9-4.6. In the United States, the 2011 annual average pH level is 2.6, ranging from 1.9 to 3.9 with an inter-quartile range of 2.2-3.0 (Fig. 1). The t-test shows a significant difference between the two groups (p<0.0001), suggesting that the aerosols are on average more acidic at the monitoring sites in the United States than in China.
The pH difference can also be illustrated by the cumulative distribution function (CDF) curves (Fig. 2, solid lines). The shapes of the CDF curves are similar in these two countries with a slightly steeper slope in the United States (Fig. 2a). The pH values, 200 however, are 1-2 units higher in China than in the United States at across levels of cumulative frequencies. In some cases, the aerosol acidities could be completely neutral in China (the frequency is 2% for pH ≥7), while in the United States, the pH values in all the cases are below 6.
Spatially, 14 out of the 16 sampling sites in China are in the NCP (Fig. S2c) which is one of the most populous and polluted regions in China (Hu et al., 2014;Cui et al., 2020). Our pH results in this region are consistent with other studies (ranging from

The pH difference based on model simulation 210
To solve the uneven spatial coverage of observational data in China, we conduct simulations using CMAQ, in company with the observational data, to further address the pH difference on a nation-wide scale. We evaluate the model performance by comparing the modeled and observed aerosol pH values (Fig. 3), major aerosol and gaseous species including SO4 2-, NO3 -, NH4 + and HNO3, NH3, and the partitioning ratios including ε(NH4 + ) and ε(NO3 -), at monitoring sites ( Fig. S4-S5). Spatially, the model simulations generally capture the observed variations in pH, species concentrations, and partitioning ratios, 215 although some biases occurred. For SO4 2-, the model captures the high concentration in the NCP and the eastern US, but it shows low biases in some sites in the southern NCP. This leads to a more negative NMB of the modelled SO4 2in China than in the United States, which can also be seen from Fig  With respect to the temporal variation, the model captures the seasonal trends of pH, ε(NH4 + ), and ε(NO3 -) in both countries, all of which are lower in summer and higher in winter (Fig. 4). The lower temperature in wintertime favors the particle-phase 230 for semi-volatile species. Comparison of the seasonal trends of the individual aerosol components shows a better agreement in the United States than in China. For example, the simulation in China misses the peaks of SO4 2in winter and NH3 in summer, and biases for HNO3 in summer (Fig. S6a, i, e). On the other hand, the simulation in the United States captures the trends of almost all the components though is biased low for SO4 2and NH4 + in summer (Fig. S6b, h). These results indicate the need for better quantification of the monthly emission trends in China which are currently subject to high uncertainty. Overall, the 235 spatial and temporal evaluation suggests generally good agreement between the model simulations and observations in both countries.
In line with the pH comparison based on observational data (Sect.3.1.1), the nationwide model simulations show significant differences in aerosol acidity between the two countries. Almost all the areas in the United States have aerosol pH values lower than 3 according to the CDF plot (Fig. 2b). Higher pH values are found in the middle and eastern United States, while in the 240 western United States except California, the pH values are lower (Fig. 3). In China, a large portion of areas (87%) have aerosol pH values above 3 according to the CDF plot, which is especially true in the eastern China with the largest population (Fig.   3). Aerosol pH values in the western and southeastern China are generally lower than in the east. The nationwide annual average pH values in China and the United States are 2.7±0.6 and 0.8±0.8, respectively, lower than the observation-based values because most of the monitoring sites are in the high pH areas (Fig. 3) and the bias in model simulation (Fig. 4a, Fig 4b). 245 Given the adverse health impacts of ambient aerosols (Burnett et al., 2014;Freedman et al., 2019) and the potential linkage of aerosol acidity with aerosol toxicity through the solubility of redox-active metals (Oakes et al., 2012;Fang et al., 2015;Ye et al., 2018), we further calculate and compare the population-weighted averages of the aerosol pH in the two countries in order to highlight the pH levels in densely populated areas. The calculation shows the weighted pH values of 3.3±0.4 and 2.2±0.5 in China and the United States, respectively, both of which are higher than non-weighted averages, which means that aerosols in 250 more populous areas tend to be less acidic (Fig. 2b). This finding is further confirmed by the significant positive correlation, within each country, between the aerosol pH and population density (China: r=0.42, p<0.0001; the United States: r=0.28, p<0.0001). Consistent with the observation-based results, the t-test for the model simulations shows a significant difference in either the population-weighted or non-weighted aerosol pH values between the two countries (p< 0.001).

Gaseous and aerosol compound profiles between China and the United States
We further investigate the factors leading to the pH difference. Although both observations and simulations are subject to uncertainty, we expect that the observational data should provide more direct and reliable evidence for this investigation, when available. Table 1  States. Similar to other studies in China (Yao et al., 2002;Pathak et al., 2009;Liu et al., 2016)

and the United
States (Guo et al., 2015;Feng et al., 2020), NH4 + , NO3and SO4 2-, contribute more than 80% of the total WSI concentrations 265 in both countries. The mass fractions of individual WSIs, however, differ between the two countries (Fig. 5). In China, the dominant WSI was NO3 -(34.6%), followed by SO4 2-(26.3%) and NH4 + (25.5%). In the United States in 2011, SO4 2contributed nearly half of the total WSI concentration (49.4%), and the contributions of NO3and NH4 + are comparable (NO3 -17.6%, NH4 + https://doi.org/10.5194/acp-2020-879 Preprint. Discussion started: 10 September 2020 c Author(s) 2020. CC BY 4.0 License. pH increase per tenfold increase in TNH3 at current TNH3 level, is higher in the United States (3.0) than in China (0.4), indicated a higher sensitivity of aerosol pH to TNH3 in the United States than in China. Besides, we find that the responses of 330 pH to TNH3 are nonlinear and anisotropic. With all others equal, pH in the United States could be closer to the level in China if the TNH3 increases to the level in China. On the other hand, the pH in China would be lower than the United States if the TNH3 decreases to the United States level because of the relative higher abundances of acidic components (SO4, TNO3, TCl) than basic ions (TNH3, NVCs) (Fig. 7a). In both countries, the sensitivities would quickly diverge from the original values toward higher values as TNH3 decreases, with the sensitivities in China changing at a faster pace. As TNH3 increases, however, 335 the sensitivities in these two countries would gradually become constant, stabilizing at comparable levels (0.002 pH unit per TNH3 increase in both two countries).
The effects of TNH3 on the gas-particle partitioning of NH3-NH4 + and HNO3-NO3are illustrated in Fig. 7b and 7c, showing a decreasing trend of ε(NH4 + ) and an increasing trend of ε(NO3 -) as TNH3 increases. In the range of observation cases the value of ε(NH4 + ) in China is smaller than in the United States, suggesting excess presence of TNH3 compared to other aerosol 340 components (e.g., TNO3 and SO4). ε(NO3 -) increased with increased TNH3, due to higher aerosol pH which promote TNO3 shifting to the particle phase as well as increased NH4 + promote the condensation of HNO3 to form NH4NO3. Higher ε(NO3 -) in China than in the United States with an average ε(NO3 -) in China being close to 1 confirmed the excess presence of TNH3.
Both the lower ε(NH4 + ) and higher ε(NO3 -) in China estimated by the sensitivity curves are consistent with observations. The gas to particle partitioning of NH3 produces inorganic ammonium salt of ammonium bisulfate (NH4HSO4) and ammonium 345 sulfate ((NH4)2SO4) first because the affinity of sulfuric acid for NH3 is much larger than that of nitric and hydrochloric acid for NH3, especially when TNH3 concentration is relatively low (Behera et al., 2013). The excess TNH3 may also react with nitric acid and hydrochloric acid to form salt of NH4NO3 and NH4Cl which will dissolve in the aerosol liquid water (Zhao et al., 2016 ], where available NH4 + is enough to balance particle phase anions. We then investigate 355 the sensitivities of pH to TNH3 in these three groups for China and the United States separately by changing the input TNH3 from a median variation range (i.e. 55% to 150%) in each group in the two countries, respectively, and keeping all other components (i.e., concentrations and meteorological conditions) unchanged. Note that no data in the United States fall in Group https://doi.org/10.5194/acp-2020-879 Preprint. Discussion started: 10 September 2020 c Author(s) 2020. CC BY 4.0 License. C, making up only two groups in the United States (i.e., Groups A and B). The results with average values of each group are shown in Fig. 8. 360 The aerosol pH increases with the increases in TNH3 in all groups, which consist with the result of the sensitivity test in Fig.   7, but the increasing rates (i.e., the sensitivities of pH to TNH3) and the pH levels vary among different groups (Fig. 8a). In China, Group C that represents aerosol systems with largest amount of excess NH4 + shows the highest pH levels and the flattest slopes of pH with TNH3, suggesting a relatively low sensitivity of pH to the change in TNH3 when TNH3 is abundant. Group A that represents aerosol systems with insufficient NH4 + , shows the lowest pH levels with the slopes slightly steeper than in 365 Group C. As TNH3 decreases to 55%, the average pH in China in Group A can be as low as 2.3, closer to the pH level in the United States, consist with the conclusion in sensitivity test using average value only (Fig. 7a). Group B can be regarded as an intermediate group between Groups A and C. But the sensitivities of pH to TNH3 changes in group B are the highest among the three groups when reducing TNH3, which could be due to the rapid increase in ε(NH4 + ) in this group as TNH3 decreases (Fig. 8b), that leads to a faster loss of NH4 + . Note that although the relative abundance of NH4 + in group 370 B is smaller than in group C, the transition from group B to group C due to TNH3 increase does not always happen. Because

if TNH3 increase in an aerosol system with 2×[SO4 2-] < [NH4 + ] < 2×[SO4 2-]+[NO3 -]+[Cl -], [NH4 + ] would increase, and more
TNO3 and TCl would shift into the particle phase, leading to the increase of WSI concentration. However, the average WSI concentration in group B is 55.03±46.79 µg•m -3 in China, significantly higher than that in group C in China (31.60±20.29 µg•m -3 ). LWC in group B (22.90±7.38 µg•m -3 ) is also higher than that in group C (14.37±16.85 µg•m -3 ). We find that most of 375 the cases in group B could be identified as highly polluted cases where large amount of NH4NO3 is formed and dissolves in the aerosol water. As a result, despite the higher abundance of NH4 + in group B than group A, ε(NH4 + ) in group B is the highest among all the groups (Fig. 8b).
Throughout the observed cases, 85% in China are in Group C (i.e., aerosol systems with excess NH4 + ), and 55% in the United States are in Group A (i.e., aerosol systems with insufficient NH4 + ). The higher sensitivity of pH to TNH3 in group A than in 380 group C explains why the pH sensitivity to TNH3 increases more significantly in the United States than in China as TNH3 decreases (Fig. 7a). Overall, the positive sensitivity of pH to TNH3 and the different dominant groups in these two countries (Group C in China, Group A in the United States) suggest that the high abundance of TNH3 in China increases the aerosol pH and is one of the major reasons for the pH difference between the two countries.

The relationship between sulfate/nitrate and aerosol pH 385
Besides the effect of TNH3 on aerosol pH discussed in Sec 3.2.3, other species, especially the acidic species which mainly include SO4 and TNO3, could also affect aerosol pH because of their effects on H + air concentration as well as on LWC (Ding et al., 2019). This effect is investigated in a sensitivity test by changing TNO3 or SO4 concentration while keeping other inputs constant as the average levels ( Fig. 9). Similar to the MSTM results as shown in Fig. 6, elevated SO4 significantly increases https://doi.org/10.5194/acp-2020-879 Preprint. Discussion started: 10 September 2020 c Author(s) 2020. CC BY 4.0 License. aerosol pH by increasing H + air. On the other hand, elevated TNO3 only slightly increases H + air, indicating a weaker acidity than 390 that of TSO4, in line with the result in a previous study (Guo et al., 2017b). This is partially due to the semi-volatile property of TNO3 (Ding et al., 2019). Notably, even in China where ε(NO3 -) are mostly close to 1, the variation of aerosol pH with TNO3 (roughly equals to NO3in this case) is also subtle. Therefore, for two systems with different moles of SO4 2and NO3neutralized by same moles of NH4 + , the system with more SO4 2will likely have a lower pH. This result indicates that higher aerosol acidity is associated with higher availability of TSO4 rather than TNO3, which can be confirmed by observed data in 395 Fig. 10.
Compared to the difference in TNO3/TSO4, the difference in NO3 -/SO4 2molar ratio is more significant due to higher aerosol pH and ammonium in China promotes TNO3 shift in particle phase as NH4NO3, leading to a higher NO3 -/SO4 2molar ratio, while low pH in the United States promotes TNO3 stay in gas phase, leading to a lower NO3 -/SO4 2ratio. Based on observation data, 74.5% of the cases in China have NO3 -/SO4 2molar ratio larger than one, while only 22.3% in the United States. The 400 different NO3 -/SO4 2ratios, as a result of the pH difference as well as TNO3/SO4 difference in two countries, could subsequently affect other aerosol properties, such as aerosol water uptake ability, which is one of the important reasons causing haze events in China during winter time Wang et al., 2020b). Although nitrate aerosol and sulfate aerosol absorbs similar amounts of water per mass (Fig. S7), heavy haze events in China are usually associated with increased LWC with enhanced RH levels under nitrate-dominate condition (Wang et al., 2020b). In order to study this effect, we categorize the observation 405 data into a nitrate-rich group (Group N, where [NO3 -]/[SO4 2-] > 3) and a sulfate-rich group (Group S, where [NO3 -]/[SO4 2-] < 1) and compare these two groups under different RH conditions. The ratio 3 in group N is mentioned in lab studies and is a more typical value of nitrate-rich conditions in field observations (Ge et al., 1998;Xie et al., 2019).
The results in Fig. 11 show that aerosol pH values in the same groups in China and the United States have similar responses to the changes in RH. In both countries, as RH increases, the pH in group N decreases, and the pH in group S increases (Fig.  410   11a). Both the values and the increasing rate of LWC in group N is larger than in group S, suggesting a higher water uptake ability in nitrate-rich condition, which is likely due to higher aerosol mass compared with group S as shown in Fig. 11f. The nearly two times aerosol mass in group N as in group S indicates the co-condensation effect of nitrate aerosol and LWC (Guo et al., 2017a), which suggests that NO3formed in aerosol leads to a higher LWC due to the increase in aerosol mass, while higher LWC dilutes H + air and increases pH, which is favorable for more HNO3 shifting from gas phase to particle phase and 415 thus continually increases particle NO3concentration. This effect will reach a balance when most of the gas phase HNO3 is in the particle phase with enough NH4 + , and, therefore, ε(NO3 -) is close to 100% in group N in the two countries (Fig. 11e).
The condition in group N usually has a higher LWC and aerosol mass, due to the mutual promotion between LWC and particle 420 nitrate. And such a condition in group N occurs more often in China than in the United States, which is probably one of the reasons leading to high particle concentrations on hazy days in China.

Discussions and implications
Based on extended ground-level measurements and regional air quality model simulations, we find significant differences in aerosol pH between China and the United States. Aerosols in the United States are on average more acidic with pH generally 425 1-2 units lower than in China. We use two independent methods, i.e., the MTSM method and sensitivity tests, to identify the key factors leading to the pH difference. These two methods consistently reveal the important role of TNH3 in causing the pH difference. The MTSM method further shows a significant contribution of NVCs on the pH difference, and the sensitivity tests highlight the high nitrate/sulfate ratios as one of the important responses to the pH difference, and high nitrate aerosol in China will further lead to higher aerosol water uptake, which may have other effects to aerosol conditions. 430 The nitrate/sulfate ratio depends on the emission ratio of NOx/SO2, the availability of cations due to the dependency of ε(NO3 -) on TNH3 (Fig. 8c, Fig. 9c), and other factors such as the atmospheric oxidizing capacity. Further investigation into the total emissions shows that the emission molar ratios of [NOx]/[SO2] are close to 3:1 in both countries (2.92 In China in 2017 and 3.12 in the United States in 2011 when assuming the emission NOx is in the form of NO2), indicating that the emission difference is not the major factor leading to the nitrate/sulfate ratio difference. On the other hand, the emission molar ratio of 435 [NH3]/([NOx]+2×[SO2]) in China (0.75) is 1.6 times higher than that in the United States (0.46), which is consistent with the measured high relative abundance of TNH3 in China and confirms that high availability of cations (mainly NH4 + caused by high NH3 emission ) is one of the causes for the high nitrate/sulfate ratio in China.
Will the aerosols in China be as acidic as in the United States as emissions are further controlled without significant reductions in TNH3? Unlikely. Although both countries have been taking actions to cut down pollutant emissions (Pinder et al., 2007;Hand 440 et al., 2012;Zhang et al., 2019), the reduction rates of NOx and SO2 emissions are quite different between the two countries ( Fig. 12). In the United States, the reduction rates of NOx and SO2 emissions (mainly from mobile and power sectors) were similar during the past two decades, while the emission of NH3 (mainly from the agricultural sector) kept constant. The data in the monitoring sites in the United States showed a decreasing SO4 2concentration over the years due to the SO2 emission reduction, but the reduction of NO3is not obvious compared with SO4 2- (Fig. S8). Lower SO4 2concentration could lead to a 445 higher aerosol pH in the United States, but this effect could be buffered by partitioning of TNH3, leading to a lower aerosol pH than expected (Weber et al., 2016). Overall, significant higher SO4 2concentration compared with relative stable NO3concentration still led to nitrate to sulfate ratio smaller than one. This ratio, however, reached a value higher than 1 in 2015, four years after the period of this study (2011). In China, on the other hand, SO2 emission reduction rate has been higher than https://doi.org/10.5194/acp-2020-879 Preprint. Discussion started: 10 September 2020 c Author(s) 2020. CC BY 4.0 License.   https://doi.org/10.5194/acp-2020-879 Preprint. Discussion started: 10 September 2020 c Author(s) 2020. CC BY 4.0 License.  https://doi.org/10.5194/acp-2020-879 Preprint. Discussion started: 10 September 2020 c Author(s) 2020. CC BY 4.0 License.