Using a Pollution-to-Risk Method to Evaluate the Impact of a Cold Front: A Case Study in a Downstream Region in Southeastern China

: Cold fronts frequently intrude China in winter, causing air pollution episodes in downwind regions. Fine particulate matter (PM 2.5 ) has been used as a major proxy of air pollution to examine the impacts of cold fronts. Compared to particles, gaseous pollutants can cause comparable or even higher levels of short-term health risks. In this study, a pollution-to-risk model was used to systematically evaluate the impacts of cold fronts on the combined health risks of air pollution mixtures, including PM 2.5 , nitrogen dioxide (NO 2 ), ozone (O 3 ), and sulfur dioxide (SO 2 ). Dominant pollutants that caused signiﬁcant mortality risks during a cold frontal passage in December 2019 over Fuzhou, a downwind city in southeastern China, were then examined. Under northerly frontal airﬂows, a pollution belt propagated southwards. In Fuzhou, two pollution peaks existed during the cold frontal passage. At the ﬁrst peak, convergence and stagnant air in the frontal zone rapidly accumulated local air pollutants. The dominant pollutants that caused the mortality risk were identiﬁed as NO 2 and PM 2.5 , both of which contributed 45% to the total risk. At the second peak, advection transported a signiﬁcant amount of secondary pollutants from the upwind regions. Although PM 2.5 was the dominant pollutant at this peak, gaseous pollutants still accounted for 34% of the total risk. Our risk analyses underscore the signiﬁcant health impacts of gaseous pollutants during cold frontal passages in winter. The results generated from this study will help guide environmental policy makers in forming and improving air pollution control strategies during pollution episodes.


Introduction
Severe air pollution episodes are frequently reported in China [1]. Because pollutant emissions are relatively steady from day to day, the evolution and occurrence of pollution episodes are largely governed by synoptic patterns and meteorological conditions [2][3][4]. Cold fronts are common synoptic patterns that frequently intrude China in winter [5,6]. They play a significant role in the earth's heat regulation system, transferring heat between polar and tropical regions. As the cold front passes a location, temperature falls, wind speed increases, and wind direction changes. In pollution source regions (e.g., northern China), the windy conditions in the immediate aftermath of a cold front often help to dilute and disperse air pollution [7].
However, downwind regions of a cold front are often affected by pollutants transported from source regions through the advection process [8,9]. Hu et al. [10] investigated a severe air pollution event in eastern China and found that a belt of fine particulate matter (PM 2.5 ) pollution was transported at the leading edge of a cold front. Similar transport of PM 2.5 pollution was found during another cold frontal passage that occurred in central China [5]. Kang et al. [11] examined the pollutant transport in the Yangtze River Delta (YRD) during a cold frontal passage, and showed a southward propagation of PM 2.5 peaks

Materials and Methods
The research design of our study is summarized as follows: (1) Because cold fronts can significantly worsen air quality in downwind regions, this study targets a downwind city in southeastern China. (2) Both transport and convergence can play a significant role in the occurrence of a pollution episode that is associated with a cold front. The equation of continuity is used to understand the different impacts of transport and convergence on pollution variations during a cold frontal passage. (3) Given the considerable short-term health risks of both particles and gaseous pollutants, a pollution-to-risk model is used to systematically evaluate the impacts of cold fronts on the health risks of air pollution mixtures. The dominant pollutants that cause significant health risks during a cold frontal passage are then examined.

Study Region
As shown in Figure 1, Fuzhou is one of the major cities located in the WTS economic zone. As a new economic zone, the WTS was developed by the Chinese government in 2009. A sustainable ecological environment is one of the major requirements of economic development [23]. The growth of the regional economy, however, rapidly worsens air quality in this region. The conflict between rapid economic development and ecological sustainability has become increasingly prominent [24]. In the winter, this region is frequently affected by cold fronts, which may bring a substantial amount of air pollutants from northern China and the YRD. Accordingly, the primary study region of our analyses is Fuzhou, the capital of Fujian Province, China.

Materials and Methods
The research design of our study is summarized as follows: (1) Because cold front can significantly worsen air quality in downwind regions, this study targets a downwin city in southeastern China. (2) Both transport and convergence can play a significant rol in the occurrence of a pollution episode that is associated with a cold front. The equatio of continuity is used to understand the different impacts of transport and convergence o pollution variations during a cold frontal passage. (3) Given the considerable short-term health risks of both particles and gaseous pollutants, a pollution-to-risk model is used t systematically evaluate the impacts of cold fronts on the health risks of air pollution mix tures. The dominant pollutants that cause significant health risks during a cold fronta passage are then examined.

Study Region
As shown in Figure 1, Fuzhou is one of the major cities located in the WTS economi zone. As a new economic zone, the WTS was developed by the Chinese government i 2009. A sustainable ecological environment is one of the major requirements of economi development [23]. The growth of the regional economy, however, rapidly worsens ai quality in this region. The conflict between rapid economic development and ecologica sustainability has become increasingly prominent [24]. In the winter, this region is fre quently affected by cold fronts, which may bring a substantial amount of air pollutant from northern China and the YRD. Accordingly, the primary study region of our analyse is Fuzhou, the capital of Fujian Province, China.

Air Quality and Meteorological Data
Hourly air pollution (PM2.5, NO2, O3, and SO2) mass concentration data for the fou cities (shown in Figure 1) were obtained from the Chinese national air quality monitorin network (http://www.cnemc.cn/, accessed on 20 December 2019). To evaluate the effect of meteorological variations, hourly wind data over East Asia were acquired from th global telecommunications system of the World Meteorological Organization (WMO). T

Air Quality and Meteorological Data
Hourly air pollution (PM 2.5 , NO 2 , O 3 , and SO 2 ) mass concentration data for the four cities (shown in Figure 1) were obtained from the Chinese national air quality monitoring network (http://www.cnemc.cn/, accessed on 20 December 2019). To evaluate the effects of meteorological variations, hourly wind data over East Asia were acquired from the global telecommunications system of the World Meteorological Organization (WMO). To assist the analyses, weather charts over East Asia during the pollution episode were obtained from the Hong Kong Observatory (https://www.hko.gov.hk/, accessed on 20 December 2019).

Air Quality Model Configuration
To better understand pollutant transport, air quality model simulations were also used in this study. Our air quality modelling system comprises the Sparse Matrix Operator Kernel Emissions Processing System (SMOKE) as an emission module, the Weather Research and Forecast model (WRF) as a meteorological module, and the Community Multiscale Air Quality (CMAQ) model as a chemical transport module. The WRF model provides a time-varying meteorological field, which powers the CMAQ model. The modelling system is set up on a nested domain with a grid size of 27 km over East Asia. More details of the model setup can be found in Zhang et al. [25].

Equation of Continuity
Assuming local emissions remain steady, severe air pollution can be caused by the advection and convergence processes. The continuity equation in fluid dynamics states that the rate of change in mass within a volume is equal to the net inflow rate of mass. The differential form of the continuity equation can be expressed as: Here, c represents fluid density, which is pollutant concentration (unit: µg/m 3 ) in this study; t represents time (unit: s); and V represents the field of velocity vector (unit for wind speed: m/s). The equation can be decomposed by separating pollutant concentration and the velocity vector: The first term is the effect of pollutant transport, resulting from a decreasing concentration gradient from the upwind region to the study region. Both V and ∇c are vectors. The dot product of these two vectors produces a scalar. The second term is the effect of convergence, resulting from the change in wind velocity. With a southward-moving cold front, surface wind in the frontal zone is governed by the convergence of cold air from the north and warm air from the south. The equation of continuity will be used to explain and understand the occurrence and evolution of the pollution episode during the cold frontal passage.

Pollution-to-Risk Model
The Air Pollution Index (API) is a common value reported by governments in a lot of countries, including China and the United States. AQHI, which directly calculates the combined health risk of air pollutants, is used in Canada and Hong Kong. Compared with the API system, the AQHI system can help better inform the public of the short-term air quality-related health risks in a more timely and quantitative manner. In this study, the quantification of the short-term health risks of air pollutants relies on the pollution-to-risk model used in the AQHI systems of Canada and Hong Kong [16,17]. The model estimates the percentage of added health risk (%AR, or so-called excess risk) of all-cause mortality or cause-specific hospital admission for different air pollutants. For specific pollutants, i, the %AR can be calculated as: where C i represents the concentration level of a specific pollutant and β i represents the corresponding risk coefficient. The sum of the %AR for individual pollutants represents the combined health risk of air pollution mixtures. Table 1 summarizes the β coefficients for different air pollutants from a global literature review and the AQHI systems of Canada and Hong Kong. The β coefficients from the global review and the AQHI of Canada are associated with all-cause mortality. SO 2 was dropped from the Canadian AQHI because of its low additional health risk. The β coefficients from the AQHI of Hong Kong were associated with hospital admissions for respiratory and cardiovascular diseases. The primary β coefficients used in this study were those obtained from the global review. Results based on the β coefficients from the two AQHI systems were used for comparisons.

Impact of the Cold Front on Meteorology
Weather charts provide basic information on synoptic patterns. Panels (a) and (b) in Figure 2 display the weather charts at 8:00 a.m. local standard time (LST) on 17 and 18 December 2019, respectively. In the winter, low temperatures continuously created high surface air pressures over the Asian continent. After the cold air mass intensified over its source region at high latitudes, it began to move southwards. As the cold air mass advanced towards a warm region, a cold front was generated at the boundary surface that separated the cold and warm air masses. The cold front arrived at Shanghai in the early morning of 17 December, and at Fuzhou overnight between 17 and 18 December.
Behind the cold front, wind speed increased and the wind direction turned northerly due to the change in pressure gradient force. Figure 2c,d show the spatial distributions of wind conditions (wind barbs) from ground weather stations over East Asia at 2:00 a.m. LST on 17 and 18 December 2019, respectively. Black straight lines represent the locations of the cold front. Behind the cold front, the dominant winds were either northwesterly (orange barbs) or northeasterly (red barbs). These northerly winds could transport air pollutants from northern China and the YRD region into southeastern China.
Atmosphere 2022, 13, 1944 5 of 15 for different air pollutants from a global literature review and the AQHI systems of Canada and Hong Kong. The β coefficients from the global review and the AQHI of Canada are associated with all-cause mortality. SO2 was dropped from the Canadian AQHI because of its low additional health risk. The β coefficients from the AQHI of Hong Kong were associated with hospital admissions for respiratory and cardiovascular diseases. The primary β coefficients used in this study were those obtained from the global review. Results based on the β coefficients from the two AQHI systems were used for comparisons.

Impact of the Cold Front on Meteorology
Weather charts provide basic information on synoptic patterns. Panels (a) and (b) in Figure 2 display the weather charts at 8:00 a.m. local standard time (LST) on 17 and 18 December 2019, respectively. In the winter, low temperatures continuously created high surface air pressures over the Asian continent. After the cold air mass intensified over its source region at high latitudes, it began to move southwards. As the cold air mass advanced towards a warm region, a cold front was generated at the boundary surface that separated the cold and warm air masses. The cold front arrived at Shanghai in the early morning of 17 December, and at Fuzhou overnight between 17 and 18 December.
Behind the cold front, wind speed increased and the wind direction turned northerly due to the change in pressure gradient force. Figure 2c,d show the spatial distributions of wind conditions (wind barbs) from ground weather stations over East Asia at 2:00 a.m. LST on 17 and 18 December 2019, respectively. Black straight lines represent the locations of the cold front. Behind the cold front, the dominant winds were either northwesterly (orange barbs) or northeasterly (red barbs). These northerly winds could transport air pollutants from northern China and the YRD region into southeastern China.  To better understand the impacts of the cold front on wind conditions, the time series of wind speed and direction are analysed. Figure 3a,b display the time series of wind speed and direction at Shanghai and Fuzhou, respectively, from 16 to 19 December 2019. Before the cold front arrived at Shanghai, the dominant winds in Shanghai were southerly. the cold front arrived at Shanghai, the dominant winds in Shanghai were southerly. As the cold front arrived at Shanghai in the early morning of 17 December, the wind speed rapidly increased and wind direction turned northerly. The changes in wind conditions were found in Fuzhou overnight between December 17 and 18, when the cold front arrived at the city. With this southward-moving cold front, the wind in the frontal zone was governed by significant air convergence, causing stagnant air with low wind speeds.
Atmosphere 2022, 13, 1944 7 of 15 As the cold front arrived at Shanghai in the early morning of 17 December, the wind speed rapidly increased and wind direction turned northerly. The changes in wind conditions were found in Fuzhou overnight between December 17 and 18, when the cold front arrived at the city. With this southward-moving cold front, the wind in the frontal zone was governed by significant air convergence, causing stagnant air with low wind speeds.

Impact of the Cold Front on Air Pollution
Cold fronts can transport pollutants from upstream polluted regions. Figure 4

Impact of the Cold Front on Air Pollution
Cold fronts can transport pollutants from upstream polluted regions. Figure 4   We now focus on the impacts of the cold front on PM2.5 variation in Fuzhou, a downwind city in southeastern China. Figure 6   We now focus on the impacts of the cold front on PM 2.5 variation in Fuzhou, a downwind city in southeastern China. Figure 6   These results underscore the significant effects of both convergence and advection on the occurrence of the pollution episode during the cold frontal passage. Convergence plays a dominant role in the frontal region because wind patterns suddenly change during a short period of time. The advection effect plays a significant role when a concentration gradient exits in the upwind region. The effects of convergence and advection sometimes mix. As shown in Figure 4, the time series of PM2.5 in Shanghai and Ningbo exhibited only one significant peak, suggesting that the two effects mixed during the cold frontal passage in these regions.   These results underscore the significant effects of both convergence and advection on the occurrence of the pollution episode during the cold frontal passage. Convergence plays a dominant role in the frontal region because wind patterns suddenly change during a short period of time. The advection effect plays a significant role when a concentration gradient exits in the upwind region. The effects of convergence and advection sometimes mix. As shown in Figure 4, the time series of PM2.5 in Shanghai and Ningbo exhibited only one significant peak, suggesting that the two effects mixed during the cold frontal passage in these regions.  These results underscore the significant effects of both convergence and advection on the occurrence of the pollution episode during the cold frontal passage. Convergence plays a dominant role in the frontal region because wind patterns suddenly change during a short period of time. The advection effect plays a significant role when a concentration gradient exits in the upwind region. The effects of convergence and advection sometimes mix. As shown in Figure 4, the time series of PM 2.5 in Shanghai and Ningbo exhibited only one significant peak, suggesting that the two effects mixed during the cold frontal passage in these regions.

Impact of the Cold Front on Health Risk
The pollution-to-risk model, together with the β coefficients of the global literature review, are now applied to evaluate the variations in the %AR of all-cause mortality that are associated with air pollutants. Figure 7a shows the time series of the %AR of all-cause mortality for PM 2.5 , NO 2 , O 3 , and SO 2 in Fuzhou from 16 to 19 December 2019. The time series of the combined %AR for the four pollutants is shown in Figure 7b. Before the cold frontal passage (i.e., 16 December), the concentration levels of air pollutants exhibited significant diurnal cycles. O 3 was the dominant pollutant that caused a significant level of short-term mortality risk during the daytime, with enhanced photochemical formations. During the nighttime with a shallow boundary layer, the dominant pollutant with shortterm health risks was NO 2 , which was mainly produced in local combustion sources (e.g., local traffic).

Impact of the Cold Front on Health Risk
The pollution-to-risk model, together with the coefficients of the global literature review, are now applied to evaluate the variations in the %AR of all-cause mortality that are associated with air pollutants. Figure 7a shows the time series of the %AR of all-cause mortality for PM2.5, NO2, O3, and SO2 in Fuzhou from 16 to 19 December 2019. The time series of the combined %AR for the four pollutants is shown in Figure 7b. Before the cold frontal passage (i.e., 16 December), the concentration levels of air pollutants exhibited significant diurnal cycles. O3 was the dominant pollutant that caused a significant level of short-term mortality risk during the daytime, with enhanced photochemical formations. During the nighttime with a shallow boundary layer, the dominant pollutant with shortterm health risks was NO2, which was mainly produced in local combustion sources (e.g., local traffic). At pollution peak T1, the convergence and stagnant air rapidly accumulated local air pollutants, such as NO2 and PM2.5. As a result, the %AR of mortality for NO2 and PM2.5 rapidly increased. The combined %AR for the four pollutants rapidly increased from 7.2 at 17:00 p.m. to 15.8 at 21:00 p.m. on 17 December. Within these four hours, the %AR for PM2.5, NO2, O3, and SO2 changed by 5.9, 5.0, −2.4, and 0.1, respectively. These results indicate that the increase in the health risk at T1 was mainly caused by PM2.5 and NO2. Fractions of the %AR of mortality for the four air pollutants at the pollution peak T1 are illustrated in Figure 8a. The dominant pollutants that caused the health risks were NO2 and PM2.5, both of which contributed around 45% to the total short-term health risks. The three gaseous pollutants together contributed 55% to the total risk, a level higher than PM2.5.
At pollution peak T2, the advection process transported a large amount of secondary pollutants, such as particles and O3, from northern China and the YRD region into Fuzhou. As a result, the %AR of mortality for PM2.5 and O3 rapidly increased. The combined %AR for the four pollutants rapidly increased from 7.3 at 0:00 a.m. to 13.1 at 5:00 a.m. on December 18. Within these five hours, the %AR for PM2.5, NO2, O3, and SO2 changed by 6.4, −2.3, 1.8, and −0.1, respectively. These results indicate that the increase in health risk at T2 was mainly caused by PM2.5 and O3. Fractions of the %AR of mortality for the four air pollutants at the pollution peak T2 are illustrated in Figure 8b. The dominant pollutant At pollution peak T 1 , the convergence and stagnant air rapidly accumulated local air pollutants, such as NO 2 and PM 2.5 . As a result, the %AR of mortality for NO 2 and PM 2.5 rapidly increased. The combined %AR for the four pollutants rapidly increased from 7.2 at 17:00 p.m. to 15.8 at 21:00 p.m. on 17 December. Within these four hours, the %AR for PM 2.5 , NO 2 , O 3 , and SO 2 changed by 5.9, 5.0, −2.4, and 0.1, respectively. These results indicate that the increase in the health risk at T 1 was mainly caused by PM 2.5 and NO 2 . Fractions of the %AR of mortality for the four air pollutants at the pollution peak T 1 are illustrated in Figure 8a. The dominant pollutants that caused the health risks were NO 2 and PM 2.5 , both of which contributed around 45% to the total short-term health risks. The three gaseous pollutants together contributed 55% to the total risk, a level higher than PM 2.5 . overwhelm the impacts of particulate matter. The second pollution peak was caused by the transport of secondary pollutants. Particulate matter and O3 are common secondary pollutants that can be transported between cities. Because most cold fronts occur in the cold season, concentration levels of O3 are usually not that high. As a result, the health impacts of particulate matter are often dominant during such a pollutant transport episode in winter. The impacts of gaseous pollutants may increase if a large amount of O3 is produced and transported.

Uncertainty Analyses
To better understand the uncertainties of the estimations, we compared the fractions of the %AR for the four air pollutants at two pollution peaks (i.e., T1 and T2) using the three sets of coefficients (i.e., from the global review and the AQHI systems of Canada and Hong Kong). The results are shown in Table 2. Although differences between the three sets of β coefficients exist, major results from the risk analyses using the β coefficients of the global review are consistent with those using the two AQHI systems. At the pollution peak T1, the dominant pollutants were still local pollutants, such as NO2 and PM2.5. At the pollution peak T2, the dominant pollutant was still PM2.5. The fraction of the risk of gaseous pollutants increased because the relative significance of the coefficients for gases in these two AQHI systems was considerably higher than the global review. Table 2. A comparison of the fractions of the %AR for the four air pollutants at two pollution peaks (i.e., T1 and T2) using the three sets of coefficients (i.e., from global review and the AQHI systems of Canada and Hong Kong).

Pollutants
Global  At pollution peak T 2 , the advection process transported a large amount of secondary pollutants, such as particles and O 3 , from northern China and the YRD region into Fuzhou. As a result, the %AR of mortality for PM 2.5 and O 3 rapidly increased. The combined %AR for the four pollutants rapidly increased from 7.3 at 0:00 a.m. to 13.1 at 5:00 a.m. on December 18. Within these five hours, the %AR for PM 2.5 , NO 2 , O 3 , and SO 2 changed by 6.4, −2.3, 1.8, and −0.1, respectively. These results indicate that the increase in health risk at T 2 was mainly caused by PM 2.5 and O 3 . Fractions of the %AR of mortality for the four air pollutants at the pollution peak T 2 are illustrated in Figure 8b. The dominant pollutant that caused health risks was PM 2.5 , contributing 66% to total short-term mortality risk. After excluding the effects of PM 2.5 , the gaseous pollutants still significantly contributed to the total risk (e.g., O 3 : 17.2%, NO 2 : 14.9%, and SO 2 : 1.8%).
The causes of the two pollution peaks were different. The first pollution peak was mainly caused by the accumulation of local pollutants. NO 2 is a typical local pollutant that is generated from combustion sources, such as traffic. Concentration levels of NO 2 in urban areas, with a large amount of vehicles, are usually much higher than rural areas. As NO 2 accumulated in the city, the combined health impacts of gaseous pollutants could overwhelm the impacts of particulate matter. The second pollution peak was caused by the transport of secondary pollutants. Particulate matter and O 3 are common secondary pollutants that can be transported between cities. Because most cold fronts occur in the cold season, concentration levels of O 3 are usually not that high. As a result, the health impacts of particulate matter are often dominant during such a pollutant transport episode in winter. The impacts of gaseous pollutants may increase if a large amount of O 3 is produced and transported.

Uncertainty Analyses
To better understand the uncertainties of the estimations, we compared the fractions of the %AR for the four air pollutants at two pollution peaks (i.e., T 1 and T 2 ) using the three sets of β coefficients (i.e., from the global review and the AQHI systems of Canada and Hong Kong). The results are shown in Table 2. Although differences between the three sets of β coefficients exist, major results from the risk analyses using the β coefficients of the global review are consistent with those using the two AQHI systems. At the pollution peak T 1 , the dominant pollutants were still local pollutants, such as NO 2 and PM 2.5 . At the pollution peak T 2 , the dominant pollutant was still PM 2.5 . The fraction of the risk of gaseous pollutants increased because the relative significance of the β coefficients for gases in these two AQHI systems was considerably higher than the global review.

Discussion
Assessments of the impacts of cold fronts in winter mainly use particulate matter as the proxy of air pollution. Although particulate matter is an important pollutant in the cold season, gaseous pollutants can still pose a significant level of short-term health risks on the population. Given the considerable health risk of gaseous pollutants, this study used the pollution-to-risk model from the AQHI systems of Canada and Hong Kong to systematically evaluate the impact of a cold front on the health risks of air pollution mixtures. In our risk analyses, the gaseous pollutants played a dominant role in increasing the short-term health risk. Our results indicate that the effects of gaseous pollutants should not be ignored in the assessment of pollution episodes in the winter.
A propagation of the pollution peak was detected in southeastern China during the cold frontal passage in December 2019. Pollution propagation is a common feature that is associated with cold fronts [11]. The strengthened airflows transport a large amount of pollutants from pollution source regions into downstream regions. Given the significant impacts of cold fronts on worsening air quality in downwind regions, this study targeted Fuzhou, a key downwind city in southeastern China and the WTS economic zone.
The impacts of synoptic patterns and meteorological conditions on the occurrence of pollution episodes are clear. Meteorological factors play significant roles in air pollution formation, transport, and dispersion [26,27]. Before the cold front arrived at Fuzhou, concentration levels of air pollutants exhibited significant diurnal cycles. For instance, O 3 concentration reached its peak in the early afternoon because intense solar radiation enhanced photochemical reactions. In the frontal zone, convergence and stagnant atmosphere hindered the dispersion of air pollutants and thus rapidly accumulated local air pollutants near the ground. Behind the cold front, the steady northerly winds transported a significant amount of secondary pollutants from the surrounding polluted regions.
The effects of convergence and advection on the occurrence of air pollution episodes can be explained very well by the equation of continuity. After the local emission is assumed to be steady, the equation of continuity can be decomposed into two terms, representing the convergence and advection effects. The effect of convergence is caused by the change in wind velocity, while the effect of transport results from the concentration gradient. The convergence usually occurs in the frontal zone. The effect only lasts for a short period of time, which is likely to be within a few hours. As the frontal zone moves forward, the strengthened winds behind the cold front rapidly dilute air pollution. The duration of pollutant transport depends on the amount of pollutants in the upwind regions. Compared to the effect of convergence, the effect of advection can last for a much longer time. It is worth noting that the impacts of local emission change should be taken into account when there is a significant change in local emission (e.g., fire accident, biomass burning, or explosion).
Air pollutants pose a range of adverse health effects on humans. PM 2.5 , NO 2 , and O 3 pollution issues are particularly severe in China, having considerably higher levels than in developed countries. Our risk analyses underscore the great health impacts of