Effects of wintertime polluted aerosol on cloud over the Yangtze 1 River Delta : case study

31 The effects of polluted aerosol on cloud are examined over the Yangtze 32 River Delta (YRD) using three-month satellite data during wintertime from 33 December 2013 to January 2014. The relationships between aerosol 34 properties and cloud parameters are analyzed in detail to clarify the 35 differences of cloud development under varying aerosol and meteorology 36 conditions. Complex relationships between aerosol optical depth (AOD) 37 and cloud droplet radius (CDR), liquid water path (LWP) and cloud optical 38 thickness (COT) exist in four sub-regions. High aerosol loading does not 39 obviously affect the distributions of cloud LWP and COT. In fact, an 40 inhibiting effect of aerosol occurs in coastal area for low-and medium-low 41 clouds, more pronounced in low clouds (<5km) than high clouds. Low 42 aerosol loading plays a positive role in promoting COTs of highand low43 clouds in areas dominated by marine aerosol. The most significant effect 44 presents in valley and coal industry districts for clouds except high-cloud. 45 The smallest values and variations of cloud parameters are observed in dry46 polluted area, which suggests that dust aerosol makes little difference on 47 clouds properties. Synoptic conditions also cast strong impacts on cloud 48 distribution, particularly the unstable synoptic condition leads to cloud 49 development at larger horizontal and vertical scales. The ground pollution 50 enhances the amount of low-level cloud coverage even under stable 51 condition. Aerosol plays an important role in cloud evolution for the low 52 layers of troposphere(below 5km) in case of the stable atmosphere in 53 wintertime. 54

COTs of all height-type clouds are affected equally by AODs at low-level.

266
As for high-clouds, the inhibiting effect of aerosol on COTs is more 267 outstanding (R 2 =0.47) than the promoting effect. In the sub-region C, 268 except for high-clouds, the influence of low-level AODs on COTs of other 269 type clouds is relatively stronger than that in the sub-region B, while high-270 level AODs are less influential in the sub-region B than the sub-region C 271 and cast no evident impacts on high-clouds. In the sub-region D, COTs and 272 AODs show a significant positive correlativity at low-level AODs, for 273 example, a steep slope (3.58) appears in high-clouds. Generally, COT links 274 closely with AOD, in particular of low-and medium-low clouds in the sub-region A, low-and high-clouds in the sub-region B and D, and other types except high-clouds in the sub-region C.  Although aerosol plays equal roles in all height-type clouds in the sub-303 region B, the best-fit slopes at high-level AODs are twice as large as those 304 at low-level AODs, and correlation coefficients for the clouds below 4.6km 305 are larger than clouds in higher layers (Table S1). In other words, for each 306 level of clouds, LWPs increase slowly (AOD<0.6) but decrease sharply 307 (AOD>0.6) with AODs growing. Opposite to the sub-region B, the 308 promoting effect of AOD on LWP in the sub-region C at low AODs is 309 marked, while the inhibiting effect is not significant at high AODs (Table   310   S1). In addition, the promoting effect of low clouds in the sub-region C is 311 most outstanding. In the sub-region D, the pronounced effect of AOD on 312 LWP mainly works on low-and high-clouds at low-level AODs (Table S1).

313
Particularly, the best-fit slope of high-clouds, such as 2.53 at low-level 314 AODs and -3.46 at high-level AODs, is much higher than that of other 315 height-type clouds.    (Table S1).

345
When AOD grows larger than 0.6, the cloud development is reduced, 346 probably due to that aerosols shade the surface. The reducing surface  (Table S1). Firstly, it is notable that a considerable portion of relatively high correlation that clouds with low LWP (<500 g/m 2 ) generate little rain and are not 452 strongly susceptible due to aerosol.    Regarding the problem of aerosol and cloud data matching, we add a 516 case study that attempts to match the observed aerosols by satellite to the same source influencing clouds over a series of days. Figure. (Fig.14b), will take air horizontal and vertical movements into consideration. With 562 sharp decline of low-cloud fraction and unremarkable variation of high-563 cloud parameters (Fig.16), it can be inferred that the enhanced high-cloud 564 fraction is mainly caused by transmission. In other words, the occurrence 565 of high aerosol layer on 4 th February is mainly caused by vertical elevation 566 of air mass from polluted ground and long-distance horizontal 567 transportation from west.

568
Air mass transportation has great influences on aerosol micro-properties 569 (e.g. particle size, shape, composition) and then clouds development. For consequently, the relatively stable atmosphere appears at low altitude. With 585 the low values of SLI and SLP (Fig.13), large concentrations of PM 2.5 are 586 left on the ground in these two days. Atmosphere suddenly becomes 587 unstable from 3 rd February (Fig.13) as AODs and aerosol layer fractions 588 decrease on 2 nd February. Also, as shown in Fig.16, from 4 th to 7 th February,

589
LWPs of low-and mid-low clouds increase systemically from noon to     The whole dataset is sorted as low to high polluted atmospheres by AOD at interval of

972
The whole dataset is sorted as low to high polluted atmospheres by AOD at interval of 973 0.2.