The magnitude of the effect of air pollution on sunshine hours in China

Authors

  • Yawen Wang,

    1. Key Laboratory of Agricultural Water Resources, Hebei Laboratory of Agricultural Water-Saving, Center for Agricultural Resources Research, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Shijiazhuang, China
    2. Graduate University of Chinese Academy of Sciences, Beijing, China
    Search for more papers by this author
  • Yonghui Yang,

    Corresponding author
    1. Key Laboratory of Agricultural Water Resources, Hebei Laboratory of Agricultural Water-Saving, Center for Agricultural Resources Research, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Shijiazhuang, China
    • Corresponding author: Y. Yang, Key Laboratory of Agricultural Water Resources, Hebei Laboratory of Agricultural Water-Saving, Center for Agricultural Resources Research, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, 286 Huaizhong Rd., Shijiazhuang 050021, China. (yonghui.yang@sjziam.ac.cn)

    Search for more papers by this author
  • Na Zhao,

    1. Key Laboratory of Agricultural Water Resources, Hebei Laboratory of Agricultural Water-Saving, Center for Agricultural Resources Research, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Shijiazhuang, China
    2. School of Hydropower and Information Engineering, Huazhong University of Science and Technology, Wuhan, China
    Search for more papers by this author
  • Chen Liu,

    1. Graduate School of Environmental Studies, Nagoya University, Nagoya, Japan
    Search for more papers by this author
  • Qinxue Wang

    1. National Institute for Environmental Studies, Tsukuba, Japan
    Search for more papers by this author

Abstract

[1] This study investigates the changes in sunshine hours in relation to API (Air Pollution Index) across China. Data were collected from a total of 38 cities over the period of 1960–2009. Sunshine hours in over 84% of the cities significantly (p < 0.05) decline with an average of 16.7% for the 1960s–2000s. This decline is mainly prevalent over Sichuan Basin (22.4%), North China Plain (18.8%), and Yangtze River Delta (18.2%). While the sunshine hour decline is largely in the 20th century (with the strongest drop in the 1980s and the least in the 1990s), it rebounds by 0.3% after 2000. For especially in winter seasons and the North China region, API is negatively related with sunshine hours. For days with API > 80, sunshine hours are on the average 0.7 h d−1 (8.4%) shorter than for days with API ≤ 80 under clear-sky condition for 2001–2005. In cities with average daily API ≤ 80 and >80 for the 2000s, sunshine hour decline for the 1960s–2000s is 0.8 h d−1 (13.4%) and 1.0 h d−1 (15.9%) respectively. Winter seasons with high API (90) exhibit the highest sunshine hour decline (21.5%). The study shows that spatiotemporal changes in sunshine hours in China could largely be explained in terms of API.

1. Introduction

[2] Solar radiation reaching the Earth's surface is the primary energy source for all forms of life. Variations in solar radiation profoundly affect land surface climate, hydrologic/carbon cycle, and human activity. Global decline in surface solar radiation, the so-called global dimming, was first noted between the 1950s and 1980s [Wild et al., 2005; Wild, 2009]. Along with variations in solar radiation were reports of sunshine hour decline in different regions of the world since the 1950s. For instance, Beniston et al. [1994] observed monthly, seasonal and yearly declines in sunshine hours in Switzerland until the late 1980s. Aksoy [1999] examined changes in hourly sunshine duration over Turkey for the period 1955–1996 and detected a declining trend for each successive day. Sanchez-Lorenzo et al. [2008] reported an overall decline in annual sunshine hours over Western Europe for the 1950s through 1980s. With the exception of the months of June, Jaswal [2009] verified a significant decline in sunshine hours over India for 1970–2006. Like in other countries/regions of the world, downward trends in sunshine hours have been found in China for the latter half of the 20th century [Kaiser and Qian, 2002; Che et al., 2005]. Southeast China [Liu et al., 2002], Yangtze River Delta [Chen et al., 2006], Southwest China [Zhang et al., 2002; Zheng et al., 2008] and North China [Yang et al., 2009a, 2009b] have all experienced strong declines in sunshine hours. Following global dimming was a reversal in the form of global brightening since the late 1980s [Wild et al., 2005]. In fact, sunshine hour recovery has been reported prior to the 1980s in certain developed countries such as the U.S. [Angell and Korshover, 1978], Western Europe [Sanchez-Lorenzo et al., 2008] and Japan [Stanhill and Cohen, 2008].

[3] Aerosols, particles smaller than 1 μm, have been identified directly or indirectly modify surface solar radiation [Charlson et al., 1992; Ramanathan et al., 2001; Wild, 2009]. Due to their composition, aerosols can directly scatter and/or absorb surface solar radiation (direct radiative forcing). Also as cloud condensation nuclei, aerosols can indirectly affect surface solar radiation by altering the number of droplet condensation and cloudiness (the first and second indirect radiative forcings). Moreover, in heavily polluted area, absorbing aerosols may inhibit cloud formation or dissolve existing clouds for heating and stabilizing the atmosphere (semi-direct radiative forcing) [Wild, 2009]. All the radiative forcings of aerosols tend to reduce surface solar radiation in the atmosphere. Human activities are releasing anthropogenic particles into the atmosphere, thereby altering aerosol density [Coe, 2011]. It is apparent from previous studies that global dimming and brightening are related with anthropogenic alteration of aerosol and pollutant load in the atmosphere [Alpert et al., 2005; Streets et al., 2006; Wild, 2009]. The effects of the changes in aerosols on solar radiation trends are adequately explored. Similarly, aerosols related with air pollution affect sunshine hour duration. There is compelling evidence that increased air pollution led to declines in sunshine hours and solar radiation over China for the last half of the 20th century [Kaiser and Qian, 2002; Che et al., 2005]. So far, however, the magnitude of the effect of air pollution on sunshine hours in China largely remains unclear.

[4] The main objective of this study was to quantify the effect of air pollution on sunshine hours in China. This was achieved by analyzing, in space and time, the variations in sunshine hours under different API (Air Pollution Index) conditions.

2. Data and Methods

[5] Data were collected in 38 cities across China (Figure 1), mainly provincial and medium- to large-size cities for which SEPA (State Environmental Protection Agency) monitored pollution data were available.

Figure 1.

Location map of China showing the studied meteorological stations, provinces and typical regions.

[6] Daily meteorological data for sunshine hours, wind speed, total cloud cover and precipitation were collected from standard weather stations of China Meteorological Data Sharing Service System (CMDS, http://cdc.cma.gov.cn/) in the 38 cities across the country. The instrumentation and routine monitoring methods of the data have been discussed by Tao et al. [1997]. Sunshine hours for 1960–2009 were observed using the Campbell-Stokes Sunshine Recorder. Wind speed was measured at 10 m height above the ground surface via either the EL electric wind direction and speed device or Dines wind direction and speed recorder. Total cloud cover was subjectively estimated by experienced observers based on recommended procedures of World Meteorological Organization (WMO). Precipitation was recorded by the Dines tilting-siphon rain gauge and siphon rainfall recorder. All the meteorological observations were in accordance with the specifications of authorized surface meteorological observations by China Meteorological Administration (CMA). Besides, strict quality control and assessment were maintained in the meteorological data collection by CMA. Quality, errors and time consistency checks were applied and all flawed records adjusted or omitted (see interpretation of documented data sets).

[7] Daily API (Air Pollution Index) data were collected for 2001–2009 (except for November 2006) from China National Environmental Monitoring Center (CNEMC, http://www.cnemc.cn/). Based on the principal pollutants of SO2, NO2 and inhalable particulates (PM10), API was calculated on the 0–500 sub-index scale for air pollutants at the monitoring stations on 24-h average concentrations. The maximum IPM10, ISO2 and INO2 values formed the upper limit of the API. The factor I was the API 24-h score for the pollutant subscripts of PM10, SO2 and NO2. Note, however, that PM10 was the most dominant air pollutant in China.

[8] Spatial trends in seasonal (March to May for spring, June to August for summer, September to November for autumn, and December to February for winter), decadal (1960s–1970s, 1970s–1980s, 1980s–1990s, and 1990s–2000s) and total percent changes of sunshine hour for the 1960s–2000s were computed and displayed in GIS using data from the 38 cities. Stepwise linear regression analysis was used to determine whether the trend in sunshine hours was significant at the 95% confidence level. Temporal trends were also illustrated for the whole of China by plotting time series of annual and seasonal averages of daily sunshine hours for 1960–2009.

[9] Stepwise linear regression was also used to analyze the effects of total cloud cover, API, precipitation and wind speed on sunshine hours, and total cloud cover, precipitation and wind speed on API using daily data for the period from January 01 2001 to December 31 2005, averaged for the 38 cities across the country. This period was used because total cloud cover data were available only up to 2005, while API measurements only started in 2001. To avoid the effect of seasonal fluctuation in sunshine hours, daily data were weighted by multiplying a ratio of annual average to monthly average.

[10] The effect of air pollution on sunshine hours was qualified under clear-sky condition (i.e., days with no precipitation and total cloud cover ≤10%) for 2001–2005 by grouping sunshine hours into days with API ≤ 80 (DAPI80) and API > 80 (DAPI80+). Daily sunshine hours in the DAPI80 and DAPI80+ categories were then separately averaged and compared at yearly and seasonal cycles. To minimize random effects on the analysis, only cities with >15 clear-sky days in a season or with >60 clear-sky days in a year for 2001–2005 in both categories were considered for comparison. Furthermore, spatial and temporal (seasonal and annual) trends of API were calculated for the study area and for 2001–2009, and compared with those of sunshine hours.

3. Results

3.1. Spatial and Temporal Trends in Sunshine Hours

[11] Differences in sunshine hours between the 1960s and 2000s for the 38 city sites are presented in Figure 2a. A drastic decline, on the average of 14.6%, is observed in all the 38 cities across China. Nanchang of Jiangxi province is the only city with no decline in sunshine hours. Significant (p < 0.05) sunshine hour declines (on the average of 16.7%) are noted in 32 cities, of which 27 cities are with >10% sunshine hour decline. The strongest declines in sunshine hours are in Sichuan Basin (22.4%), North China Plain (18.8%) and Yangtze River Delta (18.2%).

Figure 2.

Spatial patterns of sunshine hour change (%) for the 1960s–2000s and average daily API (Air Pollution Index) for 2001–2009 over the 38 cities across China at seasonal and annual time scales. The grey circles denote an upward trend and the black circles indicate a downward trend. The open circles denote that average annual trend in sunshine hours for 1960–2009 is significant at the 95% confidence level.

[12] Seasonal changes in sunshine hours between the 1960s and 2000s are illustrated in Figures 2b–2e. The strongest sunshine hour decline is in winter (21.5%), followed by summer (17.7%), autumn (12.5%), and then spring (6.9%). Note that the percent changes in sunshine hours in winter could still be somewhat higher than in summer due to the shorter day-driven smaller winter denominator. In winter (Figure 2e), sunshine hour declines occur in all 38 cities, with the strongest (>10% and significant at p < 0.05) distributions around Central China. In summer (Figure 2c), sunshine hours decline in 35 cities and are strongest/significant in regions around Sichuan Basin, North China Plain, and Yangtze River Delta. Although sunshine hour decline is noted in 36 cities in autumn (Figure 2d), the number of cities with strong/significant declines is much less than that in winter and summer. Regions around Yangtze River Delta and North China Plain still represent the most obvious decline in sunshine hours during autumn. In spring (Figure 2b), sunshine hours decrease in 27 cities, and are especially strong/significant around Southeast and Southwest China; while increase (on the average of 4.6%) in the other 11 cities, but are only significant for Nanchang city of Jiangxi province.

[13] Figure 3 depicts a time series of annually averaged daily sunshine hours over the 50-year period. There is noticeably a steady decline in sunshine hours in the 20th century. The sunshine hour decline is highest in the 1980s (5.2%), followed by the 1970s (4.3%), and then the 1990s (3.9%). Sunshine hour decline hits the lowest point in the 1990s (12.8% from the 1960s). Since the 2000s, there is a slight recovery (0.3% from the 1990s) in sunshine hours.

Figure 3.

Annual time series of average daily sunshine hours for the 38 investigated cities in China for 1960–2009. Values with a single asterisk denote percent changes in sunshine hours from the preceding to the succeeding decade, and those with two asterisks denote percent changes in sunshine hours from the 1960s to the respective decade.

[14] Figure 4 shows the seasonal trends in average daily sunshine hours across China for 1960–2009. While there is a mild decline in sunshine hours for 1960–2000 in spring, it rebounds by 6.7% after 2000. In summers of the 50-year period, there is a relatively steady decline in sunshine hours. While sunshine hour decline in autumn is strong for the early 1980s, it stabilizes after the late 1980s. Although the annual fluctuation in sunshine hours is highest in winter, a strong decline is still noticeable (especially prior to 1990).

Figure 4.

Seasonal time series of average daily sunshine hours for the 38 investigated cities in China for 1960–2009.

3.2. Effect of Air Pollution on Sunshine Hours

[15] The effects of air pollution index (API) and the meteorological variables of total cloud cover (TCC, %), precipitation (P, mm) and wind speed (WS, m s−1) on sunshine hours (SH, h d−1) was determined using stepwise linear regression analysis. A multiple regression relationship was established as follows:

display math

[16] From Table 1, the measure of goodness-of-fit R2 for equation (1) is of 0.726. F test and T-test respectively show that the equation and coefficients are significant (p < 0.0001). Sunshine hours are strongly and significantly related to the four potential driving factors in China. Of course, total cloud cover negatively affects sunshine hours by reflecting solar radiation. The critical effects of cloud amount on sunshine hours in China have also been verified for 1954–2005 by Xia [2010]. Increased aerosol load could enhance cloud formation, suppress precipitation, expand cloud lifetime, and hence reduce surface solar radiation for further increasing the reflection effect of clouds on solar radiation [Charlson et al., 1992; Ramanathan et al., 2001; Wild, 2009]. On the basis of the negative relationship between aerosols and surface solar radiation, API also exhibits negative correlation with sunshine hours.

display math
Table 1. Summary Statistics for Stepwise Liner Regression Equations and Coefficients of Total Cloud Cover (TCC, %), Air Pollution Index (API), Precipitation (P, mm) and Wind Speed (WS, m s−1)a
 R2FFactorMin.Max.MeanStd. Dev.T
  • a

    R2, coefficient of determination; F, F test statistic; Min., minimum; Max., maximum; Std. Dev., standard deviation; T, T-test statistic.

  • b

    At the significance level p < 0.0001.

  • c

    At the significance level p < 0.05.

Equation (1)0.7261206bTCC6925916.217−63.465b
   API451988119.596−12.321b
   P01833.1885.166b
   WS1.363.642.230.3778.247b
Equation (2)0.273342bTCC6925916.217−8.222b
   P01833.188−16.184b
Equation (3)0.2186cAPI7912910415.007−2.533c

[17] Equation (2) shows that total cloud cover and precipitation are significantly (p < 0.0001) and negatively correlated with API (Table 1). Precipitation and cloudiness could directly or indirectly affect the amount of pollutants in the air. Although precipitation washes out aerosols, interactions with cloud could lead to convective mixing of air pollutants in the deeper boundary layer that eventually reduces measured API. Assuming that weakening the influences of precipitation and total cloud cover helps determining any direct correlation between sunshine hours and API, a stepwise linear regression analysis was performed for days with average precipitation = 0 mm and total cloud cover ≤25% for the total 38 cities. With this assumption, the effect of API on sunshine hours was quantified as follows:

display math

[18] API plays a significantly (p < 0.05) negative role in driving the changes in sunshine hours (Table 1). Furthermore, it is further deducible that precipitation and wind speed positively influence sunshine hours by reducing air pollutants and aerosols. Precipitation and wind speed have been identified as the main drivers of sunshine hour duration respectively in Tibet [Du et al., 2007] and North China [Yang et al., 2009a, 2009b].

[19] Therefore, to isolate the impact of air pollution on sunshine hours, clear-sky days (i.e., days with 0 mm precipitation and ≤10% total cloud cover) were selected and grouped into DAPI80 (API ≤ 80) and DAPI80+ (API > 80). Figure 5 illustrates the effects of the two categories of air pollution on sunshine hours at seasonal and annual cycles under clear-sky condition for 2001–2005. Except for Lhasa in winter, sunshine hours are longer under DAPI80 than DAPI80+ conditions for all the cities and seasons. It is discernible that irrespective of precipitation and cloudiness, air pollution has a negative effect on sunshine hours. Over a year, average daily sunshine hours are 0.7 h d−1 (8.4%) longer under DAPI80 than DAPI80+ conditions. Large differences (>1.0 h d−1) in DAPI80 and DAPI80+ average daily sunshine hours exist for the cities of Xi'an (1.6 h d−1), Taiyuan (1.4 h d−1), Urumqi (1.4 h d−1) and Zhengzhou (1.1 h d−1). This implies that sunshine hours in these cities are more sensitive to air pollution condition. Moreover, sunshine hours under DAPI80 are on the average 0.6 h d−1 (5.9%), 0.6 h d−1 (5.8%), 0.9 h d−1 (10.8%), and 0.8 h d−1 (9.9%) longer than under DAPI80+ condition in spring, summer, autumn, and winter, respectively. Correspondingly, the effect of air pollution on sunshine hours is relatively strong in autumn and winter. Large differences (>1.5 h d−1) in DAPI80 and DAPI80+ average daily sunshine hours exist for Urumqi (1.9 h d−1) in spring; for Xi'an (1.8 h d−1), Zhengzhou (1.6 h d−1), and Urumqi (1.6 h d−1) in autumn; and for Xi'an (1.9 h d−1) and Taiyuan (1.6 h d−1) in winter.

Figure 5.

Comparisons of average daily sunshine hours under DAPI80 (API ≤ 80) and DAPI80+ (API > 80) categories across China for 2001–2005. Note that the plots are only for clear-sky condition (i.e., days with 0 mm precipitation and ≤10% total cloud cover).

[20] Figures 2g–2j represents the average (for the 2000s) seasonal spatial distributions of daily API over the 38 cities across China. The highest API is in winter (90), followed by spring (82), autumn (75), and summer (63). In winter (Figure 2j), cities with high API (>80) mainly occur in North China, the main coal consumption area with strong (>10%) and significant (p < 0.05) sunshine hour declines (Figure 2e). The extent and distribution of cities with high API in spring (Figure 2g) are generally similar to those in winter (Figure 2j). This is because severe dust storms generate dust particles that aggravate air pollution in the North China region during spring [Qu et al., 2010; Xin et al., 2010]. Comparing Figure 2g with Figure 2b, 12 of the 16 cities with low API (≤80), which are mainly in South China, show sunshine hour decline; while, 6 of the 22 cities with high API (>80), which are mainly in North China, show sunshine hour rise. This is apparently inconsistent with the verified negative relationship between sunshine hours and API, because of differences in humidity and water vapor between North and South China in spring. Mixing of pollutants with high water vapor could easily develop cloudy and rainy conditions in South China. This explains the sunshine hour decline under relatively low API. In autumn (Figure 2i), cities with high API largely take place in regions around North China Plain, Yangtze River Delta, and Loess Plateau; with relatively heavy declines in sunshine hours (Figure 2d). In summer (Figure 2h), which is the monsoon season in China, mainly regions round North China Plain have high API. The vast of China otherwise has low API, due probably to the frequent rain-wash of air pollutants. In contrast to Figure 2c, strong and significant sunshine hour declines occur not only in cities with high API, but also in some cities with low API. In addition, it is also conclusive from Table 2 that with the exception of spring, sunshine hour decline for the 1960s–2000s is stronger in cities with average API > 80 than in cities with average API ≤ 80 for the 2000s in all the seasons. This necessitates further analysis of the effects of other meteorological variables on sunshine hours and API during the spring season.

Table 2. Comparisons of Average Seasonal and Annual Sunshine Hour Declines (h d−1, %) for the 1960s–2000s Between Cities With Average Daily API ≤ 80 and >80 for 2001–2009 in China
 Entire YearSpringSummerAutumnWinter
API ≤ 800.8 (12.6%)0.6(11.4%)1.3(17.3%)0.7(12.2%)0.7(15.5%)
API > 801.0 (15.6%)0.3(3.7%)1.7(20.0%)0.8(12.8%)1.3(25.2%)

[21] There is almost a linear decline in average annual daily API across China for 2001–2009 (Figure 6), which corresponds with sunshine hour recovery in Figure 3. API also declines for all the seasons, with the strongest decline (from 99 to 69) in spring. In fact, spring is the only season with increasing (6.7%) sunshine hours during the 2000s (Figure 4).

Figure 6.

Seasonal and annual time series of average daily API (Air Pollution Index) for the 38 investigated cities in China for 2001–2009.

4. Discussions and Conclusions

[22] Sunshine hours quantify the length of time in each day when direct solar radiation is greater than a certain threshold (often set at 120 Wm−2). Under clear-sky condition, direct solar radiation largely occurs during the hours of the day with sunshine above this threshold. Given that sunshine records span much longer than solar radiation records, sunshine hour is widely used as a surrogate for global radiation in studies of recent trends and evolution of solar radiation, particularly in the so-called global dimming and brightening [Sanchez-Lorenzo et al., 2007, 2008; Stanhill and Cohen, 2008; Yang et al., 2009a].

[23] For the 1960s–2000s, sunshine hours in over 84% of the investigated 38 cities across China significantly (p < 0.05) decline with an average of 16.7% (Figure 2a). The strongest declines mainly distribute in regions around Sichuan Basin (22.4%), North China Plain (18.8%) and Yangtze River Delta (18.2%). Over the 50-year period, sunshine hours of China largely decrease in the last 40 years of the 20th century, and especially in the 1980s—consistent with the findings of Kaiser and Qian [2002]. After reaching the lowest point in the 1990s (−12.8% from the 1960s), a slight recovery (0.3% from the 1990s) in sunshine hours in the 2000s is noted in the 38 cities (Figure 3). In China, a similar recovery trend was observed in surface solar radiation after 1990 by Wild et al. [2005], though his update work [Wild et al., 2009] has reported a renewed solar dimming beyond 2000. Among the seasons, sunshine hour decline is strongest in winter (21.5%), followed by summer (17.7%), autumn (12.5%), and spring (6.9%). In India, a drastic decline of sunshine hours in winter was also reported by Jaswal [2009]. In China, Xia [2010] pointed out a low rate of sunshine hour decline in the spring of 1954–2005. Regions around Sichuan Basin, North China Plain, and Yangtze River Delta all exhibit great declines in sunshine hours across all the seasons of the year. Severe trends of decline in sunshine hours in these regions has also been referenced in scientific works of Kaiser and Qian [2002], Chen et al. [2006], and Yang et al. [2009a, 2009b].

[24] An important finding in this study is that in China, API negatively relates with sunshine hours. This is especially clear when the effects of precipitation and cloudiness on air pollutants and sunshine hours are excluded. The effect of API on sunshine hours is visible in three ways: 1. contingency with API categories of trends in annual and seasonal sunshine hours in the mean across all cities under clear-sky condition; 2. contingency of a city's mean API category with its trend in annual and seasonal sunshine hour declines; and 3. concurrence of the season (winter) of greatest decline in sunshine hours with that of highest API. In the first way, the averages across all cities, both for the whole year and for the separate seasons, show progressively shorter sunshine hours at times of higher API under clear-sky condition (Figure 5). This negative relationship is more obvious in North China during autumn and winter, under which condition API is relatively high (Figure 2) and clean-up effect of wind and precipitation on air pollutants is comparatively weak. In the second way, subdividing the cities by average API shows that higher API cities (mainly in North China) have stronger sunshine hour declines for the entire year and for all seasons except spring (Figure 2 and Table 2). This is due to frequent sandstorms and dust in the driest season of spring [Qu et al., 2010; Xin et al., 2010]. Dust, and not human air pollution, hardly enhances aerosols under low water vapor conditions to influence sunshine hour decline [Yang et al., 2009b]. In the third way, in winter, the season with the highest API due largely to fossil (coal) fuel combustion [Yang, 2009; Wang et al., 2010], sunshine hour decline in the 1960s–2000s is strongest. This is especially noticeable in North China, the main coal combustion area (Figure 2). It then becomes clear that air pollution has a negative effect on sunshine hours. This is consistent with the findings of Charlson et al. [1992] and Ramanathan et al. [2001]; where direct (radiation scattering and absorption) and indirect (cloud condensation nuclei formation) effects of aerosols on surface solar radiation are noted.

[25] Aerosol load, linked to air pollution, has been recognized as the most plausible cause of decadal changes in solar radiation, particularly under cloudless condition [Che et al., 2005; Streets et al., 2006]. Similarly, Kaiser and Qian [2002] noted the effect of aerosol-driven extinction in solar radiation (by enhanced backscattering and absorption of incoming solar radiation) reduces in sunshine hours (by weakening direct solar radiation needed to activate the Campbell-Stokes Sunshine Recorder). This especially occurs just after sunrise and before sunset when aerosols could affect direct beam effect on sunshine recorder because of the long distances the sunshine travels to reach the land surface [Sanchez-Lorenzo et al., 2007].

[26] In fact, API and AOD (Aerosol Optical Depth) have been widely used as representative indices of aerosols in studies on air pollution [Khanna, 2000; He et al., 2002; Li et al., 2003, 2005; Pour-Biazar et al., 2011]. The 24-h ground-based monitoring of API is performed and piecewise values derived from mass concentration measurements for dry condition. AOD measurements, on the other hand, are based on remote sensing (spatial and temporal) with background influences such as relative humidity, vertical structures and of atmospheric mixing layer heights [Li et al., 2005; Qu et al., 2010]. Various studies have shown significantly strong correlations between API and remotely sensed AOD after corrections for seasonal vertical distributions of aerosols and relative humidity real-time [Li et al., 2003, 2005; He et al., 2010]. In China, Guo et al. [2011] reported an upward tendency in AOD since 1980, and which has weakened since 2000. This is consistent with the declining trend in API (Figure 6). API, as mass concentrations of aerosols/pollutants, was proved to be negatively related with sunshine hours in our study. This phenomenon was further verified by the effects of meteorological factors such as total cloud cover, precipitation, and wind speed, which directly or indirectly affect sunshine hours and pollutant volume in the air. In their studies, Du et al. [2007], Yang et al. [2009a, 2009b], and Xia [2010], reached similar conclusions.

[27] Although API, a routinely measured air quality index, has been ascertained strongly linked with aerosols, which directly or indirectly affect solar radiation and sunshine hours, negative relationships between API and sunshine hours have hardly been addressed. This study, for the first time, proved the value of API as a separate indicator in studying the influence of air pollution on sunshine hours, even though the lack of long-term data of API and total cloud cover limited our ability in confirming the relation between API and sunshine hours for entire 50 years.

Acknowledgments

[28] This study is funded by International Collaborative Project (2009DFA21690) and National 973 Project (2010CB951002) of the Ministry of Science and Technology. Meteorological data are provided by China National Meteorological Information Centre (NMIC). The authors also acknowledge Sara C. Pryor and the anonymous reviewers for their patient work and useful comments.

Ancillary