[Establishment and Effective Evaluation of Haikou Ozone Concentration Statistical Prediction Model].

This study selected 15 key predictors of the maximum of 8-hour averaged ozone (O3) concentration (O3-8h), using the O3 concentration of Haikou and ERA5 reanalysis data from 2015 to 2020, and constructed a multiple linear regression (MLR) model, support vector machine (SVM) model, and BP neural network (BPNN) model, to predict and test the O3-8h concentration of Haikou in 2021. The results showed that the absolute value of correlation coefficients between the O3-8h and related key prediction factors was mainly among 0.2 and 0.507. The 1 000 hPa relative humidity (RH1000), wind direction (WD1000), and 875 hPa meridional wind (v875) showed a good indicative effect on the O3-8h, with the absolute correlation value exceeding 0.4. The three prediction models could predict the seasonal variation in the O3-8h in Haikou, which was larger in the winter half year and smaller in the summer half year. The root mean square error(RMSE) was the smallest (22.29 µg·m-3) in the BPNN model. The correlation coefficients between the predicted values of three statistical models and observations were ranked as 0.733 (BPNN) > 0.724 (SVM) > 0.591 (MLR), all passing the 99.9% significance test. For the prediction of the O3-8h level, we found that TS scores of these three prediction models decreased with the increase in O3-8h concentration level. Relatively, the point over rate and not hit rate increased with the rise in O3-8h concentration level. TS scores of the SVM and BPNN model were relatively larger than those of MLR, especially in the light pollution level with TS scores remaining above 70%, indicating a better prediction capability.

[Spatial and Temporal Variations in Ozone Pollution and Sensitivity Characteristics in Hainan Island].

Based on ambient air quality data, meteorological observation data, and satellite remote sensing data, the temporal and spatial variations in ozone (O3) pollution, the sensitivity of O3, and its relationship with meteorological factors in Hainan Island were analyzed in this study. The results showed that the maximum daily 8-h moving mean (O3-8h) in western and northern cities in Hainan Island was higher than that in the central, eastern, and southern cities. O3-8h was the highest in 2015, and O3-8h exceeding the standard proportion was the largest in 2019. In addition, O3-8h was positively correlated with average temperature (P<0.1), sunshine duration (P<0.01), total solar radiation (P<0.01), atmospheric pressure, and average wind speed and was negatively correlated with precipitation (P<0.05) and relative humidity. The satellite remote sensing data showed that the tropospheric NO2 column concentration (NO2-OMI) and HCHO column concentration (HCHO-OMI) displayed opposite trends in Hainan Island from 2015 to 2020. Compared with those in 2015, NO2-OMI increased by 7.74% and HCHO-OMI decreased by 10.2% in 2020. Moreover, Hainan Island belongs to the NOx control area, and the FNR value exhibited a fluctuating downward trend in the past 6 years, with a trend coefficient and climatic trend rate of -0.514 and -0.123 a-1, respectively. A strong correlation was observed between meteorological factors and the FNR value of Hainan Island.

[Effects of Tropical Cyclones on Ozone Pollution in Hainan Island].

Based on the 181 tropical cyclones data in the western North Pacific Ocean from 2015 to 2020, hourly ozone (O3) concentration data, and meteorological observation data of 18 cities and counties in Hainan Island, this study analyzed the impacts of tropical cyclones on ozone pollution in Hainan Island. We found that 40 (22.1%) tropical cyclones experienced O3 pollution in Hainan Island during the lifetime of tropical cyclones in the past six years. During the years with more tropical cyclones, more O3- polluted days occurred in Hainan Island. Highly polluted days, which were defined as more than or equal to three cities and counties exceeding the standard, were the most serious in 2019 with 39 (54.9%) polluted days. The tropical cyclones related to high pollution (HP) showed an increasing trend, with the trend coefficient and climatic trend rate of 0.725 (exceeding the 95% significance level) and 0.667 a-1, respectively. Tropical cyclone intensity was positively correlated with the maximum 8 h moving average (O3-8h) concentration in Hainan Island. Among them, HP-type tropical cyclones accounted for 35.4% of all samples in the typhoon (TY) intensity level. Cluster analysis of tropical cyclone paths showed that tropical cyclones from the South China Sea (type A) were the most common of the 67 (37%) and were the most likely to cause large-scale and high-concentration O3 pollution events in Hainan Island. The average number of HP tropical cyclones and ρ(O3-8h) of Hainan Island in type A were 7 and 121.90 µg·m-3, respectively. In addition, the tropical cyclone centers were located generally in the middle part of the South China Sea and the western Pacific Ocean near the Bashi Strait during the HP period. The change in meteorological conditions in Hainan Island under the influence of HP tropical cyclones was conducive to the increase in O3 concentration.

[Influence of Typhoon Nangka Process on Ozone Pollution in Hainan Island].

In order to explore the influence of the 2016 typhoon Nangka on ozone (O3) concentration in Hainan Island, this study employed correlation analysis and backward trajectory simulated methods, using the hourly air quality monitoring data, meteorological data, EAR5 reanalysis data, and bright temperature (TBB) infrared from the Himawari-8 satellite from October 10 to 14, 2020. The results showed that, during the typhoon process from October 11 to 13, the O3-8h (maximum of 8 hours sliding average) concentration exceeded the standard. The averaged ρ(O3-8h) of Hainan Island reached its peak value (130.5 µg·m-3) in October 12, 2020, with the O3-8h concentration exceeding the standard in four cities and counties, in which Lingao County reached the highest concentration of 198.44 µg·m-3. During the typhoon process, the hourly O3 concentration was negatively correlated with precipitation, relative humidity, and wind speed and positively correlated with air pressure and air temperature. Moreover, the correlation coefficients between O3 with precipitation, air pressure, and relative humidity exceeded the 99.9% confidence level. The southeastern coastal provinces of China were the main contribution source area for this ozone pollution process on Hainan Island. Affected by the downdraft inside the western Pacific subtropical high (WPSH) and the downdraft outside the circumference of typhoon Nangka, vertical transport in the atmosphere appeared over the contribution source area. The meteorological conditions were favorable for atmospheric photochemical reactions. Air pollutants were transported to Hainan Island along the northeast airflow around the typhoon, resulting in the occurrence of an O3 pollution event. In addition, this study built a weather conceptual model of O3 pollution in Hainan Island affected by typhoon Nangka, which can be used as a reference for the joint pollution prevention and control by air pollution prediction, early warning, and environmental management departments.

[Characteristics and Potential Sources of Four Ozone Pollution Processes in Hainan Province in Autumn of 2019].

Based on the air quality and meteorological monitoring data of Hainan province in autumn of 2019, this study analyzed the characteristics and potential sources for the four O3 polluted processes in Hainan province, using the methods of correlation analysis, HYSPLIT backward trajectory modeling, PSCF (potential source contribution function), and CWT (concentration weighted-trajectory). The results showed that ① the average concentrations of the maximum 8h average (O3-8h) for process 1 and process 3, which occurred from September 21st to 30th and November 3rd to 11th with the durations of 10 d and 9 d, were 145.52 µg·m-3 and 143.55 µg·m-3, respectively. Process 2 and process 4 occurred from October 18th to 21st and November 20th to 25th, with the durations of 4 d and 6 d, and the average concentrations of O3-8h were 130.79 µg·m-3 and 115.46 µg·m-3, respectively. ② High air pressure, low precipitation and relative humidity, long sunshine duration, and strong solar radiation favored the occurrence of O3-polluted weather in Hainan province. Northerly wind was conducive to the increase in O3-8h concentration, and wind speeds affected the regional distribution of high-value areas of O3-8h concentration in Hainan province. ③ Furthermore, process 1 and process 3 with more serious pollution had a larger air flow divergence, and there were two airflows originating from the inland area and the southeast coastal area, respectively. Air flow of process 2 and process 4 was relatively more concentrated with less O3 pollution and was classified as southeast coastal air flow. 4 The analysis of potential contribution sources showed that transport from Zhejiang, Jiangxi, Fujian, and Guangdong provinces were the main sources of O3 pollution in Hainan province in autumn 2019. Among them, the weight potential source contribution function (WPSCF) and weight concentration weighted-trajectory (WCWT) values were larger than 0.36 and 90 µg·m-3 in the PRD (Pearl River Delta) and western Guangdong province regions, respectively.

[Temporal and Spatial Variations in Ozone and Its Causes over Hainan Province from 2015 to 2020].

This study investigated temporal and spatial variations in O3-8h (defined as the maximum 8 h average result) in Hainan Province from 2015 to 2020 and further analyzed its relationships with precursors and meteorological factors based on a dataset of observations from 32 environmental monitoring stations in Hainan. Basic statistical methods, including the empirical orthogonal function (EOF), climatic tendency rate, and climatic trend coefficient analysis, were used here. The results showed that ρ(O3-8h) was higher in northern and western Hainan than that in other regions, with the maximum value occurring in Dongfang City (91.5 µg·m-3). Twelve cities and counties experienced a downward trend from 2015 to 2020, and six cities and counties reached a 95% confidence level. The variation in ρ(O3-8h) in Hainan Province demonstrated remarkable seasonal changes, which were the largest in the autumn, spring, and winter followed by the smallest in the summer, exhibiting a clear declining trend in all seasons except autumn. In addition, the cumulative variance of the first two eigenvector fields decomposed by EOF was 72.58%, which could well describe the distributed characteristics of ρ(O3-8h) in Hainan Province. The first mode reflected the consistency of ρ(O3-8h) variation, and the second mode reflected regional differences. Meanwhile, the change in ρ(O3-8h) had a good correlation with the precursors and meteorological factors. Among them, the correlation coefficients between ρ(O3-8h) and ρ(NO2), precipitation, sunshine duration, average temperature, average wind speed, atmospheric pressure, and total radiation passed the 99% confidence test. The results of multiple linear regression showed that the variation in regressed ρ(O3-8h) was consistent with the observed ρ(O3-8h), and the correlation coefficient between them was 0.853, which passed the 99.9% confidence test. The regression value explained 0.72 variance of the observed value.

[Temporal and Spatial Variation of Haze Pollution over China from 1960 to 2013].

Based on haze observation and related meteorological data from the surface weather stations in China during 1960-2013, we analyzed the variation of haze days and its relationship to meteorological elements using climatic linear-trend estimation, cluster analysis, cumulative departure and the Mann-Kendall test. The results showed that the haze events occurred mainly in mid-eastern and southern China, especially in Beijing, central and southern areas of Shanxi, Henan, the Yangtze River Delta and Pearl River Delta areas. At the same time, the haze days were less frequently observed in western and northeastern China. The occurrence of haze days was increasing in the past 54 years, which was consistent with the total energy consumption in China. While the emission of air pollutant could be an important factor of haze increase, the adverse weather conditions also played a role. The correlation coefficients of precipitation, average wind speed, sunshine duration and relative humidity with haze days were -0.653, -0.635, -0.462 and -0.699, respectively, which all exceeded the 99.9% significance. Cluster analysis showed that haze days of stations with very significant increase, significant increase and obvious increase have accelerated in recent years, and the trend of cumulative variance type was fall-flat-rise. Haze days of slight rise station increased from 1960s to the late1970s and after 2000, of which cumulative variance type was multi wobbled. Moreover, Haze days of slight fall and obvious fall stations increased from 1960s to 1970s, and their cumulative variance types were rise-gently-fall. Furthermore, haze days showed mutation from more to less during 1992 to 1993.

[Temporal and spatial characteristics of atmospheric NO2 over Hainan Island and the pollutant sources in recent 10 years].

The temporal-spatial characteristics of the tropospheric column NO2 (TroNO2) and total column NO2 (TotNO2) over Hainan Island are analyzed using remote sensing data derived from OMI sensor, and also combining surface wind, SO2, HYSPLIT model to research the source of atmospheric pollutants over Hainan Island. The results show that: The value of NO2 in northern area is higher than that in southern area, and the value of NO, in central mountainous area is lower than those other places. In addition, the seasonal variation of NO2 indicates that NO2 is higher in winter and lower in summer, which can be attributed to precipitation in summer and external transport of atmospheric pollutants in winter. Long-term changes of NO2 in Hainan Island appear opposite trends during winter and summer, which is declining in winter and has a weak increase in summer. The reasonable explanation is that local emissions of pollutants play an important role in summer, but external transport is the main resource of pollutants over Hainan Island. The TroNO2 in Haikou City has a good relationship with favorable delivered days in PRD, the correlation coefficient is 0.84 with 99% confidence level. Moreover, there are 3 transport paths in Dec. 2013 which can impact Haikou City from backward trajectory analysis, but all of them pass through the PRD, which can further prove that atmospheric pollutants of Hainan Island in winter are mainly delivery from PRD region.