In this study, we focus on Tongling city and the area surrounding it. The study region (117.5E-118E, 30.8–31.1N) is shown in Figure 1. In this domain, the Yangtze River passes through from south to north and is almost perpendicular to the latitude in all lines. Therefore, the river breeze caused by the difference of land and surface contrast is mainly in east-west direction. Observation data from all meteorological stations (total 22 stations) and air quality monitoring stations (total four stations) in study region are collected for analyzing river breeze and its influences on pollutant particles distribution. These four air quality monitoring stations are located at the ninth middle school (NS), the wastewater factory (WF), the fourth middle school (FS) and the road bureau (RB).

The data from air pollutant observation stations (national stations of Ministry of Ecology and Environment, PRC) and meteorological observation stations (regional stations of China Meteorological Administration) in the whole year of 2011 (observation experiment period in Tongling) are used in this study, include concentration of PM₁₀, the wind speed and direction. The time frequency of these observation data is 1 h. Those concentration of PM₁₀ were observed by the LGH-01B PM₁₀ air particulate matter monitoring instrument (produced by Anhui Landun Photoelectron CO., LTD., Anhui, PRC). The wind speed was observed by the three-cup wind speed sensor (EL15-1A) and the wind direction was observed by the sensor of EL15-2A (produced by Zhonghuan TIG, Tianjin, PRC).

The data of PM₁₀ are disposed according to the technical specifications of Ministry of Ecology and Environment, PRC (Technical Specifications for Installation and Acceptance of Continuous Automatic Monitoring System for Ambient Air Particulate Matter (PM₁₀ and PM_2.5), HJ 655–2013). And the supplemental quality control of all these datasets had been carried out before analysis. The primary purpose of quality control is to eliminate the outliers. However, it is also assumed 10% of air pollutant data are unreliable. The quality control aims to eliminate not only the outliers, but also those unreliable data. So, first step is to calculate its probability distribution, then, a threshold is obtained to eliminate those 10% data.

According to formulas (1), it is clear that total wind ( V ⇀ )can be decomposed into zonal ( u i ⇀ ), meridional ( v j ⇀ ) and vertical component ( w k ⇀ ). Because only horizontal wind data are collected in meteorological stations, and only zonal and meridional wind data are available. And for our case, the river-land breeze mainly affects the winds in east-west direction; therefore, only zonal winds (positive for westerly wind and negative for easterly wind) are analyzed. In addition, since river-land breeze is local deviated wind from environmental background wind, we further decompose zonal wind into environmental background wind ( u ¯ ) and local deviated wind ( ∆ u ) as shown in formulas (2). V ⇀ = u i ⇀ + v j ⇀ + w k ⇀ (1) u = u ¯ + ∆ u (2) u ¯ = 1 n ∑ i = 1 n u i (3) n is the number of weather stations in the study region, u i is the zonal wind of the weather station i for the per hour observation data.

As the environment background wind ( u ¯ ) is calculated with domain average method as formula (3) shown, the local deviated wind ( ∆ u ) is subtracted from the total zonal as formula (2). As the local deviated wind ( ∆ u ) is caused by the local geomorphic characteristics (river and land), and it is used to express river-land breeze in the following articles.

Environment background wind ( u ¯ ) is calculated with domain average method, and the formula is showed in formula (3). The difference (Diff) between the local deviated wind in the east ( ∆ u e a s t ) and west ( ∆ u w e s t ) bank of the Yangtze River is also calculated as formula (4). D i f f = ∆ u w e s t − ∆ u e a s t (4)

Diurnal variation of annual mean local deviated wind ( ∆ u ) is shown in Figure 2. Generally, A “change inversely” relationship can be noticed between the west and east side of the river. During 8:00–16:00 (local time, and same hereafter), It is easterly wind in the west of the Yangtze River and westerly wind in the east, exhibiting a divergence pattern. In contrast, during 16:00–1:00, the local winds reverse the direction, which forms a convergence pattern. It can also be noticed that the local deviated zonal wind speed ∆ u in both east and west sides of the Yangtze River increases from 8:00a.m., reaching the maximum between 12:00–14:00, and gradually decreases afterwards. According to the difference between ∆ u in the east and west (Diff), there are two maximums in a diurnal cycle. The first one occurs between 12:00–14:00 when both u' and Diff reach their maximums (the largest ∆ u is 0.28 m/s and occurs in the west side of the Yangtze River, while the largest Diff is 0.37 m/s). The second maximum occurs between 16:00–22:00, and both ∆ u and Diff (0.2 m/s and 0.29 m/s respectively) are smaller in magnitude compared with the first maximum.

The diurnal variation of local deviated wind ( ∆ u ) in the east and west side of Yangtze River and their difference for different seasons are shown in Figure 3. It shows the following information. 1) For the pairs of ∆ u maximum at any season, in the daytime, it is always negative in the west side and positive in the east side of the Yangtze River. The only difference about this aspect for different seasons is the strength of ∆ u and the duration time of strong wind reversal. 2) In the daytime, the difference between the winds in either side of the Yangtze River is most significant in spring, followed by autumn, winter and summer. 3) In the nighttime, the difference between the winds in either side of the Yangtze River is the most obvious in winter, followed by spring and autumn, and the smallest difference occurs in summer too. 4) “Diff” is larger in the nighttime than daytime in winter, while vice versa for other seasons. For the Diff represents the difference of wind between the east and west sides of the Yangtze River, as formula (4), it can be used to represent the speed of river-land breeze. As Figure 3 shown, the speed of river-land breeze at night is greater than that in daytime in winter, and it is just the opposite in other seasons.

The mean local deviated wind speed and standard deviations are calculated during daytime (9:00–16:00) and nighttime (19:00–04:00). As shown in Table 1, the land breeze in the west side of Yangtze River is stronger than that in the east. The mean ∆ u in the west and east are −0.18 m/s (negative sign indicates easterly wind) and .07 m/s (positive sign indicates westerly wind) respectively. In the west side, the river-land breeze is strongest in spring. And the average easterly/westerly wind speed is 0.29/0.27 m/s. However, in the east side, the river breeze in summer is the most prominent (0.11 m/s), so is the land breeze in winter. For the wind different between the west and east side, the maximum occurs during daytime in spring, and the minimum occurs during nighttime in summer.

As shown in Figure 4, the results from this analysis support that easterly winds (land breezes) are dominant during nighttime, but westerly winds (river breezes) appear during 10:00–16:00, which has crucial impacts on dominant winds over Tongling. During the diurnal cycle, when the maximum occurs with westerly winds in the east and easterly winds in the west during 10:00–16:00, the river breezes are dominant winds; when the maximum occurs with westerly winds in the west and easterly winds in the east during 16:00–22:00, the land breezes are dominant winds. The difference is more significant in the daytime than nighttime.

The main cause of river-land breeze is the different heat capacity between river and land. Suppose Q (J/m²) is solar radiation flux, C is specific heat capacity of surface (land and water), dT represents the change of temperature due to radiation, and m indicates unit mass of atmosphere. Then the formula of solar radiation flux is as following: Q = C ⋅ d T ⋅ m (5)

Because specific heat capacity of land (denoted as CL) is smaller than that of water (denoted as CW), for same solar radiation, temperature change of water (denoted as dTW) is less than that of land (denoted as dTL). In the daytime, temperatures of both land and the Yangtze River rise because of solar radiation, but the land is heated much quicker than the river, resulting in warm land and cold water. In contrast, during the night, temperatures of both land and the Yangtze River decrease due to radiation cooling, but the land cools much quicker than the river, resulting in cold land and warm water. A pressure gradient is generated due to the temperature difference between the Yangtze River, which leads to airflow movement.

According to atmospheric state equation p = ρ R T , if there is a difference between the temperatures in two regions, their corresponding atmospheric pressure (p) will respond to the change of temperature gradient by increasing their gradient accordingly. Based on equation of atmospheric motion [formula (6)], wind speed as well as the change rate of the wind speed with time will follow the changes of the pressure gradient, resulting in an increase of the local deviated wind speed. In another word, the river-land breeze increases the amplitude of the perturbed wind component, which further aggravate the inhomogeneity of the wind fields. Therefore, river-land breeze increases the change rate of the wind with time, amplifies the perturbations of the total wind and finally effects the diffusion of the atmospheric polluted matters. d u d t − f v = − 1 ρ ∂ p ∂ x + F x d v d t + f u = − 1 ρ ∂ p ∂ y + F y d w d t = − 1 ρ ∂ p ∂ z − g + F z (6)

Diurnal variation of the speed of ∆ u and the regional average concentration of PM₁₀ are shown in Figure 5. We can find that the ∆ u is generally low speed (<0.35 m/s) from 0:00 to 7:00. A quick increase can be noticed after 7:00, and it reaches maximum around 16:00 and then decreases abruptly afterwards. Also, it is obvious that the speed of ∆ u is small during the night, and less than 0.4 m/s after 20:00.