Object-based Tracking Analyses

To examine the impact of CCN on diurnally developed deep convection based on

the properties of each convective system, object-based tracking analyses are conducted,

combining cloud object connecting and rain cell tracking algorithms. The six-connected segmentation method, proposed by Tsai and Wu (2017), connects horizontally and vertically adjacent cloudy grid boxes as the same cloud object, and its details are given in Appendix C. In this study, only the convective clouds are analyzed, whose cloud base height is lower than 0.5 km, cloud depth is thicker than 1.0 km, and center of cloud mass is higher than 0.5 km. The iterative rain cell tracking (IRT; Moseley et al., 2019) links rain cells at each time step and forms the rain tracks of precipitating systems. IRT provides a Lagrangian framework focusing on the life cycle of precipitating systems, and its details are provided in Appendix D. By co-locating the rain cells with the cloud objects overhead, the evolution of the diurnal precipitating systems would be completed (Appendix E).

Figure 8 provides the numbers of rain tracks identified under different CCN

concentration scenarios and for different types of precipitation patterns. Under the clean scenario, the total number of rain tracks for the WEAK type is almost 1.5 times of that for the STRONG type. However, the total number of rain tracks for the two types are about the same amount under the normal scenario. Increasing CCN leads to a decline in the rain tracks for both types. For the WEAK type, the reduction is nearly 35%, which is much higher than that for the STRONG type (about 10%). The rain tracks that can

only produce light rain rates are the majorities of the reduction. Especially for the WEAK type, the rain tracks that can only produce rain rates lighter than 20 mm·hr^-1 decreases for about 50% when CCN concentration increases. Since the influence of CCN is mainly on the rain tracks with light rain rates, it might explain why CCN concentration has less effect on the composite precipitation for the WEAK type (Figure 6c and Figure 6d).

Figure 9 illustrates the probability density function of the clouds with various maximum rain rates. For the STRONG type, the probability of heavy rain rates is notably higher under the normal scenario. On the other hand, the increase in the probability of heavy rain rates due to rising CCN is less noticeable for the WEAK type.

Furthermore, the decrease in the probability of light rain rates (less than 20 mm·hr^-1) resulting from increasing CCN is more symbolic for the WEAK type. The size of the circles in Figure 9 represents the critical cloud size (CCS), defined as the minimum

cloud size that can produce the corresponding rain rate (Tsai and Wu, 2017). For the STRONG type, increasing CCN generally leads to larger CCS for heavy rain rates, representing that the clouds have to grow larger to produce the same rain rate under the normal scenario.

Figure 10 shows the rain contribution of the rain tracks with different maximum

rain rates. For the STRONG type, more than half of total precipitation is furnished by

the rain tracks with maximum rain rate heavier than 100 mm·hr^-1. Furthermore, the rain

tracks with the ability to precipitate heavier than 130 mm·hr^-1 provide about 15% of total precipitation under the clean scenario and increases to more than 40% when CCN concentration rises. On the other hand, the rain contribution from the rain tracks with

maximum rain rate heavier than 130 mm·hr^-1 is lower than 10% for the WEAK type, regardless of CCN concentration. This result reveals that the rain contribution from the

extreme rain rates increases with CCN concentration for the STRONG type, while it is

less affected by CCN concentration for the WEAK type.

We have learned that the initiation time of precipitation can be influenced by CCN from the previous composite results (Figure 7), and thus we further examine the timing of the rain tracks, displayed in Figure 11. The initiation time is about the same under the clean scenario for the two types and it generally lies between 13:00 and 17:00.

Increasing CCN leads to a significant delay in both the initiation time and the ending time. The delay for the STRONG type is about an hour, which is longer than that for the WEAK type (about 20 minutes). The duration of the rain tracks usually keeps within an hour and is less affected by CCN. Although only less than 5% of the rain tracks could survive for more than 3 hours, the extreme rain tracks could last for almost 12 hours.

Thus, the mean of the duration is about the same as the 3^rd quartile.

Figure 12 demonstrates the statistics of the maximum rain rate and the maximum

rain area during the lifetime of the rain tracks, representing the strength of precipitation in the mature stage of the diurnal precipitating systems. Increasing CCN leads to a significantly larger mean maximum rain rate for the WEAK type. On the other hand, for the STRONG type, although increasing CCN leads to a less significant increment in the mean maximum rain rate, the enhancement in the 99^th percentile (P99) of the maximum rain rate is about 22.1 mm·hr^-1, which is 2.5 times higher than that for the WEAK type (8.8 mm·hr^-1). The means of the maximum rain area are around 25 km²

under different CCN concentration scenarios and for different types of precipitation

patterns. Nevertheless, the P99 of the maximum rain area for the STRONG type expands from 876 km² to 1073 km² when CCN concentration increases, while that for the WEAK type remains around 275 km². The extreme values of the maximum rain rate and the maximum rain area increase with increasing CCN for the STRONG type. This result reveals that the rise of CCN concentration strengthens the extreme precipitation for the diurnal precipitating systems which are already more intense than the others.

By co-locating the rain cells with the cloud objects overhead (Appendix E), the statistics of the maximum cloud depth and the maximum cloud size during the lifetime of the rain tracks are presented in Figure 13, representing the strength of cloud development in the mature stage of the diurnal precipitating systems. Increasing CCN leads to a significantly larger mean maximum cloud depth and cloud size for both the

STRONG and the WEAK type. However, it is more evident for the STRONG type that the majority of the clouds become thicker and larger when CCN concentration rises.

The interquartile range (IQR) of the maximum cloud depth decreases by 7.0 km, and the 1^st quartile (Q1) of the maximum cloud depth surges from 6.9 km to 14.9 km under the normal scenario. Also, the IQR of the maximum cloud size reduces by 5.5×10³ km³,

and the Q1 of the maximum cloud size rises from 6.7×10³ km³ to 11.8×10³ km³ under the normal scenario. On the other hand, for the WEAK type, CCN concentration has less effect on the IQR and the Q1 of the maximum cloud depth and the maximum cloud size. The result indicates that the rise of CCN also enhances the growth of clouds being more well-developed than the others, viewed from the macro-scale cloud properties of the diurnal precipitating systems.

Figure 14 gives the statistics of the maximum in-cloud vertical velocity and the maximum core ratio during the lifetime of the rain tracks, representing the cloud dynamical features in the mature stage of the precipitating systems. The core ratio is defined as the proportion of the clouds with a vertical velocity larger than 0.5 m·s^-1, representing the updraft region. Increasing CCN leads to a narrower distribution of vertical velocity, a stronger upward motion, and a relatively smaller core ratio. This phenomenon is more compelling for the STRONG type: the IQR of the maximum vertical velocity declines from 18.3 m·s^-1 to 14.0 m·s^-1, the mean maximum vertical

velocity rises significantly from 21.1 m·s^-1 to 24.6 m·s^-1, and the mean maximum core ratio decreases from 44% to 37%. As for the WEAK type, the IQR of the maximum vertical velocity declines from 10.6 m·s^-1 to 9.0 m·s^-1, the mean maximum vertical velocity rises slightly from 15.3 m·s^-1 to 16.7 m·s^-1, and the mean maximum core ratio decreases marginally from 46% to 43%. The result reveals that the updraft regions of the more severe diurnal precipitating systems are more concentrated and more vigorous under the normal scenario.

In conclusion, when CCN concentration is higher, the diurnal precipitating systems with greater ability to produce heavy rain rates have a higher probability to occur, tend to develop into larger sizes, and are more responsible for total precipitation. Also, from

the rain properties, cloud macro-scale characteristics, and cloud dynamical features of the diurnal precipitating systems, a “strong get stronger” response is identified with

increasing CCN concentration.