**3. Results and Discussion**

#### *3.1. Relationship between Precipitation, Particle Mass Concentration, and SE*

First, we investigated potential size effects on the scavenging efficiencies. The figure shows the relationship between the *RI* and *SE*. In the distribution of the precipitation intensity, most precipitation intensities are lower than 5 mm/h. A precipitation intensity above 5 mm/h takes a relatively low proportion in the samples. Therefore, for the segment with rainfall less than 1 mm, a 0.2 mm interval is adopted, while for the segment with rainfall greater than 1 mm, a 2 mm interval is adopted. From Figure 2, we can see that when the rain intensity (*RI*) is less than 0.4 mm h<sup>−</sup>1, the *SE* of PM2.5 is almost zero, but the *SE* of PM10 can reach ~2 μg m−<sup>3</sup> h<sup>−</sup>1. The concentration of PM2.5 often rises during weak precipitation (*RI* lower than 0.5 mm h<sup>−</sup>1); when the *RI* is 7 mm h<sup>−</sup>1, the *SEs* of PM2.5 and PM10 are 2.7 and 6.3 μg m−<sup>3</sup> h<sup>−</sup>1, respectively. The *SE* is positively correlated with the *RI*: the greater the *RI*, the higher the *SE*. When the *RI* is 10 mm h<sup>−</sup>1, the *SE* of PM2.5 reaches 5.1 μg m−<sup>3</sup> h<sup>−</sup>1, and the *SE* of PM10 reaches 15.8 μg m−<sup>3</sup> h<sup>−</sup>1.

**Figure 2.** Relationship between precipitation intensity and scavenging efficiency (*SE*) (the dashed lines are the fitted curves, the green area is the interquartile span of PM10, the red area is the interquartile span of PM2.5, and the gray area is the overlapping region).

By using functions to fit the rainfall intensity and scavenging efficiency, where *SEpm*2.5 and *SEpm*<sup>10</sup> are the *SEs* of precipitation on PM2.5 and PM10, respectively, *RI* is precipitation, and *t* is the precipitation duration, we relate the *RI* and *SE* as follows:

$$SE\_{pm2.5} = 0.275 \times \exp\left(\frac{RI}{6.468 \times t}\right) - 0.068\tag{2}$$

$$SE\_{pm10} = 0.218 \times \exp\left(\frac{RI}{5.019 \times t}\right) + 1.653\tag{3}$$

The changes in PM2.5 and PM10 caused by precipitation under a stable *RI* (the precipitation intensity remains unchanged) are as follows:

$$\text{CON}\_{pm2.5} = \left[ 0.275 \times \exp\left(\frac{RI}{6.468 \times t}\right) - 0.068 \right] \times t \tag{4}$$

$$\text{CON}\_{pm10} = \left[ 0.218 \times \exp\left(\frac{RI}{5.019 \times t}\right) + 1.653 \right] \times t \tag{5}$$

The derivatives of Equations (3) and (4), respectively, are

$$\text{CON}'\_{pm2.5} = 0.275 \times \exp\left(\frac{RI}{6.468 \times t}\right) \times \left(1 - \frac{RI}{6.468 \times t}\right) - 0.068 \tag{6}$$

$$\text{CON}'\_{pm10} = 0.218 \times \exp\left(\frac{RI}{5.019 \times t}\right) \times \left(1 - \frac{RI}{5.019 \times t}\right) + 1.653\tag{7}$$

Therefore, when *CON pm*2.5 or *CON pm*<sup>10</sup> is equal to zero, changes in *CONpm*2.5 and *CONpm*<sup>10</sup> caused by precipitation reach the maximum values; the corresponding *RI* reaches 5.8 mm/h and 10.1 mm h<sup>−</sup>1, respectively.

Based on the analysis of the 27,219 precipitation processes from 2013 to 2017 in Jiangsu Province, we find that the effect of precipitation is greater on coarse particles than on smaller particles.

Figure 3 shows the effect of the particle mass concentration on the *SE* under different *RI*s when precipitation processes are classified according to the *RI*. Higher particle mass concentrations under the same *RI* and heavy rain under the same particle mass concentration all have a higher *SE*. The precipitation *SE* on PM10 is higher than that of PM2.5 for the same *RI* and the same particle mass concentration. The *SE* is an increasing function of both the *RI* and the initial concentration. Precipitation has a limited effect on particulate matter and even has no effective clearance when the particle mass concentration is below the thresholds (PM2.5 below 40 μg m−<sup>3</sup> and PM10 below 60 μg m<sup>−</sup>3). The precipitation *SE* on the particle is significantly enhanced with the increase in the particle mass concentration. The precipitation *SE* with an intensity below 0.5 mm h−<sup>1</sup> can reach more than 15 μg m−<sup>3</sup> h−<sup>1</sup> when the concentration of PM2.5 or PM10 is above 140 μg m<sup>−</sup>3. This also explains why sometimes the particle mass concentration rises after strong precipitation and why sometimes the particle mass concentration decreases after weak precipitation. This is because the *SE* is not only determined by accumulative precipitation or the *RI* but also by the two combined. Strong precipitation with a low particle mass concentration may result in a negative clearance effect, and a high particle mass concentration with weak precipitation may lead to a positive *SE*. This result is similar to the numerical model simulation results, in that the same amount of precipitation may lead to different removal efficiencies of atmospheric aerosols [2]. Jose Nicolás et al. [34] and Yoo et al. [35] also found a higher atmospheric removal efficiency for coarse particles than for fine particles.
