Abstract

Cloud microphysics and cloud condensation nuclei (CCN) measurements from two marine stratus cloud projects are presented and analyzed. Results show that the increase of cloud droplet concentrations Nc with CCN concentrations NCCN rolls off for NCCN at 1% supersaturation (S)N1% above 400 cm−3. Moreover, at such high concentrations Nc was not so well correlated with NCCN but tended to be more closely related to vertical velocity W or variations of Ww). This changeover from predominate Nc dependence on NCCN to Nc dependence on W or σw is due to the higher slope k of CCN spectra at lower S, which is made more relevant by the lower cloud S that is forced by higher NCCN. Higher k makes greater influence of W or σw variations than NCCN variations on Nc. This changeover at high NCCN thus seems to limit the indirect aerosol effect (IAE).

On the other hand, in clean-air stratus cloud S often exceeded 1% and decreased to slightly less than 0.1% in polluted conditions. This means that smaller CCN [those with higher critical S (Sc)], which are generally more numerous than larger CCN (lower Sc), are capable of producing stratus cloud droplets, especially when they are advected into clean marine air masses where they can induce IAE. Positive correlations between turbulence σw and NCCN are attributed to greater differential latent heat exchange of smaller more numerous cloud droplets that evaporate more readily. Such apparent CCN influences on cloud dynamics tend to support trends that oppose conventional IAE, that is, less rather than greater cloudiness in polluted environments.

1. Introduction

Low stratus clouds that predominate the east sides of oceans provide a majority of the indirect aerosol effect (IAE; Warren et al. 1988; Platnick and Twomey 1994; Kogan et al. 1996), which remains the largest climate uncertainty (Alley et al. 2007). The high radiative temperatures and large albedo contrasts with the underlying ocean provide substantial global cooling that can be increased by the advection and injection of anthropogenic cloud condensation nuclei (CCN), which increase cloud droplet concentrations Nc (albedo; first IAE; Twomey 1977a) and decrease droplet sizes (lifetime; second IAE; Albrecht 1989). CCN concentrations (NCCN) or more specifically CCN spectra (NCCN active at various levels of supersaturation S) and vertical velocity W at cloud base are the main determinants of Nc and the subsequent spectra of cloud droplets. Hudson et al. (2010a, hereafter H10) broke the conventional wisdom of maximum S < 0.3% (100.3% RH) in stratus clouds (Hudson 1983; Hoppel et al. 1986; Leaitch et al. 1996; Roberts et al. 2006; Hegg et al. 2009) by comparing below-cloud CCN spectra with nearby Nc during the Physics of Stratocumulus Tops (POST) project (Carman et al. 2012).

Here we present further analysis of POST along with measurements and analysis of a complementary field project, the Marine Stratus/Stratocumulus Experiment (MASE) of July 2005 (Wang et al. 2009). Although both of these aircraft research projects were done in the same area off the central California (CA) coast (Tables 12) in summertime stratus clouds, there were several important contrasts between them (Table 3). The differences were greatly due to the shorter period of the MASE data that are presented here from 9 flights on 9 days over 11 days (15–25 July 2005) whereas POST data are presented from 13 flights on 13 days over 29 days (18 July–15 August 2008). The foremost differences were higher NCCN (Table 3, columns 2–7, Fig. 1a) and Nc (Table 3, columns 8–10, Fig. 2) and smaller droplets (Table 3, columns 11–13 and Fig. 2) in MASE. The longer time period and greater number of flights in POST contributed to the greater variety of aerosol (Table 3, columns 3–4 and 6–7), Nc (Table 3, columns 9–10) and droplet mean diameters (MD; Table 3, columns 12–13) during POST. For all of these variables the maximum values of the two projects are similar but the minimal values of the two projects differ by a factor of 2 to an order of magnitude. One important aspect that could have been related to these microphysical differences was the lower cloud tops and especially lower cloud bases during MASE (Table 2, column 6), which made it difficult to obtain below-cloud CCN measurements. Higher cloud bases and a slower airplane (Twin Otter compared to Gulfstream 1; Table 2, column 2) allowed more extensive below-cloud measurements during POST (Table 2, columns 9–10). The persistent aerosol layer above CA stratus (e.g., Hudson and Frisbie 1991) was often absent during POST (H10) but always present during the MASE flights considered here. Here we take the opportunity to advance understanding of IAE by comparing and contrasting aerosol–cloud relationships between these two field projects done in clouds that are the most relevant to IAE.

Table 1.

Project (Proj), number of flights (Flts), number of horizontal cloud penetrations (Clds), total duration of cloud data (Cloud dur), range of number of penetrations per flight (Clds/flt), mean, median (Med), minimum, and maximum duration of cloud penetrations, number of penetrations after splitting 6 (POST) and 12 (MASE) (Clds split), range of number of penetrations per flight after splitting (Clds/flt), and mean, median, maximum, and minimum durations after splitting.

Project (Proj), number of flights (Flts), number of horizontal cloud penetrations (Clds), total duration of cloud data (Cloud dur), range of number of penetrations per flight (Clds/flt), mean, median (Med), minimum, and maximum duration of cloud penetrations, number of penetrations after splitting 6 (POST) and 12 (MASE) (Clds split), range of number of penetrations per flight after splitting (Clds/flt), and mean, median, maximum, and minimum durations after splitting.
Project (Proj), number of flights (Flts), number of horizontal cloud penetrations (Clds), total duration of cloud data (Cloud dur), range of number of penetrations per flight (Clds/flt), mean, median (Med), minimum, and maximum duration of cloud penetrations, number of penetrations after splitting 6 (POST) and 12 (MASE) (Clds split), range of number of penetrations per flight after splitting (Clds/flt), and mean, median, maximum, and minimum durations after splitting.
Table 2.

Project, aircraft speed, mean, standard deviation, minimum, and maximum altitude of cloud measurements, mean cloud-base altitude, mean cloud liquid water content (LWC) from CAS, mean drizzle LWC from CIP, number of CCN measurements, duration of CCN measurements, range of CCN measurements per flight, and mean, median, minimum, and maximum duration.

Project, aircraft speed, mean, standard deviation, minimum, and maximum altitude of cloud measurements, mean cloud-base altitude, mean cloud liquid water content (LWC) from CAS, mean drizzle LWC from CIP, number of CCN measurements, duration of CCN measurements, range of CCN measurements per flight, and mean, median, minimum, and maximum duration.
Project, aircraft speed, mean, standard deviation, minimum, and maximum altitude of cloud measurements, mean cloud-base altitude, mean cloud liquid water content (LWC) from CAS, mean drizzle LWC from CIP, number of CCN measurements, duration of CCN measurements, range of CCN measurements per flight, and mean, median, minimum, and maximum duration.
Table 3.

Project, mean and std dev of N1%, minimum N1%, maximum N1%, mean and std dev of N0.1%, minimum N0.1%, maximum N0.1%, mean and std dev of Nc, minimum Nc, maximum Nc, mean and std dev of droplet mean diameter (MD), minimum MD, maximum MD, activation ratio, and mean and std dev of slope of CCN spectra at NCCN(S) = mean Nc.

Project, mean and std dev of N1%, minimum N1%, maximum N1%, mean and std dev of N0.1%, minimum N0.1%, maximum N0.1%, mean and std dev of Nc, minimum Nc, maximum Nc, mean and std dev of droplet mean diameter (MD), minimum MD, maximum MD, activation ratio, and mean and std dev of slope of CCN spectra at NCCN(S) = mean Nc.
Project, mean and std dev of N1%, minimum N1%, maximum N1%, mean and std dev of N0.1%, minimum N0.1%, maximum N0.1%, mean and std dev of Nc, minimum Nc, maximum Nc, mean and std dev of droplet mean diameter (MD), minimum MD, maximum MD, activation ratio, and mean and std dev of slope of CCN spectra at NCCN(S) = mean Nc.
Fig. 1.

(a) Mean POST and MASE CCN spectra. The most polluted POST flight (18 Jul) is plotted separately. (b) The k of (a) are shown.

Fig. 1.

(a) Mean POST and MASE CCN spectra. The most polluted POST flight (18 Jul) is plotted separately. (b) The k of (a) are shown.

Fig. 2.

Mean differential droplet spectra of all of the horizontal cloud passes analyzed in each project.

Fig. 2.

Mean differential droplet spectra of all of the horizontal cloud passes analyzed in each project.

2. Measurements

As with H10 all POST measurements were made on board the Center for Interdisciplinary Remotely-Piloted Aircraft Studies (CIRPAS) Twin Otter airplane based at Marina, California, just north of Monterey. Cloud and drizzle droplet probes and W measurements from this airplane have been presented by Conant et al. (2004), Lu et al. (2007, 2008, 2009, 2012), Small et al. (2009), and H10. Cloud measurements were made by the cloud, aerosol, and precipitation spectrometer (CAPS) of Droplet Measurements Technologies of Boulder, Colorado (Baumgardner et al. 2001). The two components of CAPS used in this study are the cloud and aerosol spectrometer (CAS) for cloud droplets between 0.58 and 51-μm diameter and the cloud imaging probe (CIP) for drizzle drops between 50- and 1500-μm diameter. A five-hole gust probe on the radome of the airplane and a pitot-static pressure tube with a GPS-corrected C-MIGITS III using the technique of Lenschow (1986) (Lenschow and Spyers-Duran 1989) as described by Khelif et al. (1999, 2005) determined W. The W measurements from this aircraft have been presented by Lu et al. (2007, 2008, 2009, 2012).

All measurements during MASE were made on the Department of Energy Gulfstream 1 (G1) airplane, which was based in Sacramento, California, during MASE. A similar CAPS probe and a slightly different five-hole gust probe (Brown et al. 1983; Chan et al. 1998) were used on the G1. During one of the MASE flights the W measurements were favorably compared with model simulations (Guo et al. 2008).

CCN spectra in both projects were measured with the Desert Research Institute (DRI) CCN spectrometers (Hudson 1989), which have been used and reported from numerous aircraft field experiments (Hudson and Yum 2002, 2001; Yum and Hudson 2004; Hudson and Mishra 2007; Hudson and Noble 2009; Hudson et al. 2009, 2010b, 2012). In both the POST and MASE projects this instrument was calibrated at least once during each flight. During MASE there were two DRI CCN spectrometers operating over slightly different S ranges and there was good agreement between the two instruments in the overlapping S range (not shown).

Flight durations in each project were ~4 h. With the exception of Fig. 3a, all cloud data are from 1-s averages during horizontal flight at constant altitudes. As in previous publications by the authors, cloud data were restricted to only those seconds with cloud probe (here the CAS) liquid water content (LWC) greater than 0.1 g m−3. Mean values of all of these 1-s values in each cloud pass are used in this analysis. As described in section 6, 6 of the 28 POST clouds and 12 of the 38 MASE clouds were divided so that 34 POST clouds and 50 MASE clouds were considered throughout this analysis; Table 1 displays differences before and after these splits of some of the cloud passes.

Fig. 3.

(a) Revision of Fig. 2 of H10, and (b) revision of Fig. 4 of H10. These show Seff against N1% for 69 vertical penetrations in (a) and 34 horizontal penetrations of POST clouds. The Seff is the S for which nearby below-cloud NCCN(S) equals mean Nc. Major corrected data for 18 Jul are denoted in green. Linear regression lines and correlation coefficients R are shown.

Fig. 3.

(a) Revision of Fig. 2 of H10, and (b) revision of Fig. 4 of H10. These show Seff against N1% for 69 vertical penetrations in (a) and 34 horizontal penetrations of POST clouds. The Seff is the S for which nearby below-cloud NCCN(S) equals mean Nc. Major corrected data for 18 Jul are denoted in green. Linear regression lines and correlation coefficients R are shown.

All CCN measurements were made in horizontal flight legs below cloud as outlined in Tables 2 and 4. Mean concentrations of the data obtained in each of these legs are used in this analysis. The CCN measurements were made continuously at a rate of 1–3 s−1 with less than a tenth of a second of dead time between each measurement. In POST there was one below-cloud CCN measurement for each of the original 28 cloud passes, thus the same CCN measurement was used for both sections of the 6 divided passes. The POST CCN measurements were made well below cloud base at either 30- or 100-m altitude before or after each of the cloud passes (Tables 2 and 4). Because of the lower cloud bases and greater speed of the MASE airplane, CCN measurement legs were usually very close to cloud base at ~100-m altitude. Since cloud-free air at these altitudes was limited in MASE, there were only 26 separate CCN measurements that had to be used for the 38 original cloud passes, and as in POST, there was only one CCN measurement for both sides of the 12 divided cloud passes. Therefore, many more of the MASE cloud passes used the same CCN measurements, especially within the flights with the largest numbers of cloud passes.

Table 4.

Project, mean, std dev, minimum, and maximum lengths of the cloud penetrations (columns 2–4), and of the CCN below-cloud measurements (columns 5–7); mean, std dev, minimum, and maximum distances between the nearest edges of each of the clouds to their corresponding CCN measurements (columns 8–10); and distances between the farthest edges of the clouds from the corresponding CCN measurements (columns 11–13). All distances in km.

Project, mean, std dev, minimum, and maximum lengths of the cloud penetrations (columns 2–4), and of the CCN below-cloud measurements (columns 5–7); mean, std dev, minimum, and maximum distances between the nearest edges of each of the clouds to their corresponding CCN measurements (columns 8–10); and distances between the farthest edges of the clouds from the corresponding CCN measurements (columns 11–13). All distances in km.
Project, mean, std dev, minimum, and maximum lengths of the cloud penetrations (columns 2–4), and of the CCN below-cloud measurements (columns 5–7); mean, std dev, minimum, and maximum distances between the nearest edges of each of the clouds to their corresponding CCN measurements (columns 8–10); and distances between the farthest edges of the clouds from the corresponding CCN measurements (columns 11–13). All distances in km.

Table 4 outlines the lengths of the cloud passes and CCN measurement paths used here along with characterizations of the separations between each of the clouds from their corresponding CCN measurements. The separations between cloud and corresponding CCN measurements usually consisted of slant path ascents or descents, data from which are not considered here. These separations shown in Table 4 are maximum possible distances because, for simplicity, they assume that the airplane continued to fly in straight lines for each cloud and corresponding CCN measurement and separation between them. Since this was not always the case, especially for MASE where the airplane usually flew back through the same clouds at different altitudes in opposite directions, the distance separations are often overestimated. However, this does not reduce the corresponding time differences between the CCN and cloud measurements, which can be computed from the airspeed. During those separation periods the air in the regions of the measurements could have changed anyway. On the other hand, Table 5 indicates that there were relatively small differences in Nc and NCCN within the flights. Thus, Table 6 shows that the correlation coefficients (R; Pearson product moment) based on flight averages were similar to corresponding R based on all of the individual clouds in each project. Furthermore, H10 showed the same N1%Nc R values for vertical cloud penetrations that had much smaller distance and time separations than the separations between horizontal cloud penetrations and their corresponding below-cloud CCN measurements.

Table 5.

Variations in terms of the standard deviations/means of Nc for the 11 POST flights and 8 MASE flights with multiple cloud passes (columns 2–4) and of the CCN concentrations at 1% S for the 10 POST flights and 8 MASE flights with multiple CCN measurements.

Variations in terms of the standard deviations/means of Nc for the 11 POST flights and 8 MASE flights with multiple cloud passes (columns 2–4) and of the CCN concentrations at 1% S for the 10 POST flights and 8 MASE flights with multiple CCN measurements.
Variations in terms of the standard deviations/means of Nc for the 11 POST flights and 8 MASE flights with multiple cloud passes (columns 2–4) and of the CCN concentrations at 1% S for the 10 POST flights and 8 MASE flights with multiple CCN measurements.
Table 6.

The R for all of the cloud penetrations and for the flight averages of these penetrations.

The R for all of the cloud penetrations and for the flight averages of these penetrations.
The R for all of the cloud penetrations and for the flight averages of these penetrations.

Cloud passes during POST were made at only one altitude for each cloud. Multiple altitude cloud passes were done in MASE during six of the nine flights. The Nc was roughly constant with altitude for three of these flights and increased by ~25% with altitude for two flights and increased by ~50% with altitude in the highest-altitude MASE flight. Cloud pass altitudes were confined to 150–440 m for all but one MASE flight where they ranged from 560 to 700 m. In MASE divisions of data according to altitude showed little difference from correlations that considered all MASE cloud altitudes. Nevertheless, the altitude differences of the MASE clouds are a source of uncertainty.

Since W measurements are less reliable during altitude changes, data during slant soundings are not considered in this analysis. The W was recorded at 10 Hz in both projects. The W considered here are means of all of these measurements where LWC of the CAS exceeded 0.1 g m−3 within each of the 34 and 50 cloud passes. There is also consideration of the standard deviations of Ww), which was computed from the same 10-Hz W measurements.

3. Corrections of H10

Estimates of stratus cloud supersaturations S during POST (H10) demonstrated the reduction of cloud S [effective supersaturation; Seff; S for which NCCN(S) = Nc] by the competition among cloud droplets brought about by high NCCN as had been predicted by Twomey (1959, hereafter T59). However, it has since been discovered that an incorrect calibration was applied to the CCN spectra for the POST flight with the highest NCCN and Nc. This produced erroneously low Seff estimates for 18 July, which had provided 8 of the 69 vertical cloud penetrations shown in Fig. 2 of H10 and 3 of the 28 horizontal penetrations shown in Fig. 4 of H10. Figure 3 here shows corrected versions of those two H10 figures. There was also a minor correction for the 4 August flight that slightly adjusted 7 vertical (Fig. 2 of H10, Fig. 3a here) and 3 horizontal (Fig. 4 of H10, Fig. 3b here) data points and there were a few other tiny revisions. Also the Seff here are more precise because all CCN channels were used to determine Seff rather than interpolations between 10 NCCN channels in H10.

The horizontal data (Fig. 3b) also add 6 data points because 6 of the original horizontal cloud penetrations have been divided because of abrupt differences in Nc and W during these penetrations (Fig. 4a); thus 34 instead of 28 data points. Nonetheless, these revised figures still exhibit the decrease of Seff with NCCN, although the decrease is not as sharp as in H10 because Seff remained above 0.1% even for the most polluted POST flight of 18 July. Figures 1 and 3 of H10 [Nc versus NCCN at 1% S (N1%)] are imperceptibly changed because the calibration corrections affected only lower S NCCN and not N1%. Revisions of those two figures still show the expected high positive correlation between N1% and Nc—for example, Fig. 5a, which is an update of Fig. 1 of H10.

Fig. 4.

(a),(b) 1-s Nc and (a) W and (b) σw variation with time during divided 18 Jul horizontal cloud penetrations for POST in (a) (Table 9, row 2) and for MASE in (b) (Table 10, row 2). Mean Nc and mean W in (a) and mean σw in (b) for each division are shown.

Fig. 4.

(a),(b) 1-s Nc and (a) W and (b) σw variation with time during divided 18 Jul horizontal cloud penetrations for POST in (a) (Table 9, row 2) and for MASE in (b) (Table 10, row 2). Mean Nc and mean W in (a) and mean σw in (b) for each division are shown.

Fig. 5.

Mean Nc vs below-cloud N1% for (a) 34 POST horizontal cloud passes (as part of Fig. 3 of H10) and (b) 50 horizontal passes in MASE. (c) Data from (a),(b) together are shown. Linear regressions are shown in (a),(b), and second-order regression is shown in (c).

Fig. 5.

Mean Nc vs below-cloud N1% for (a) 34 POST horizontal cloud passes (as part of Fig. 3 of H10) and (b) 50 horizontal passes in MASE. (c) Data from (a),(b) together are shown. Linear regressions are shown in (a),(b), and second-order regression is shown in (c).

4. Influences on cloud droplet concentrations Nc

Figure 6a shows the R patterns between N1% and cumulative droplet concentrations larger than various threshold sizes Nt—that is, the R of Fig. 5a (black circles) and Fig. 5b (red squares) are displayed on the far left sides of Fig. 6a, which are the R values for the smallest Nt, which is Nc, thus the R of N1%Nc. Patterns similar to those of POST (black circles in Fig. 6a) were also found in the Rain in Cumulus over the Ocean (RICO) (Hudson et al. 2009), the Pacific Atmospheric Sulfur Experiment (PASE) (Hudson and Noble 2009), and the Ice in Clouds Experiment—Tropical (ICE-T) (Hudson et al. 2010b) where the strong positive R for Nc (all cloud droplets) and Nt larger than small size thresholds (<~10 μm) gives way to negative R for Nt larger than bigger size thresholds (>~10 μm). As in the other projects, this change from positive to negative R of N1%Nt occurs at approximately the peak of the mean droplet spectrum (Fig. 2, black data) where competition among droplets for condensate is greatest. This is what causes higher NCCN to produce fewer droplets larger than these larger size thresholds, which makes the negative R between N1% and Nt. This is the principle behind pluvial constipation, second IAE (Hudson 1993)—that is, higher NCCN inhibits production of larger cloud droplets that can precipitate. Black circles in Fig. 7 show that the same strong positive R for Nc (smallest droplet size threshold of 0.7 μm) with NCCN extends down to 0.15% S (N0.15%) and continues positive down to N0.02%.

Fig. 6.

(a) The R between N1% and cumulative Nc larger than various size thresholds Nt vs the size thresholds. (b),(c) As in (a), but R of W and σw measured within cloud with Nt, R of W and σw measured where CCN were measured with Nt, and R of W and σw measured over more extensive low altitudes with Nt.

Fig. 6.

(a) The R between N1% and cumulative Nc larger than various size thresholds Nt vs the size thresholds. (b),(c) As in (a), but R of W and σw measured within cloud with Nt, R of W and σw measured where CCN were measured with Nt, and R of W and σw measured over more extensive low altitudes with Nt.

Fig. 7.

The R for NCCN at each S with the various mean droplet concentrations in both projects and with W in POST and σw in MASE.

Fig. 7.

The R for NCCN at each S with the various mean droplet concentrations in both projects and with W in POST and σw in MASE.

In complete contrast with POST and probably every other comparison between NCCN and Nc (e.g., Hudson and Yum 2001, 2002; Conant et al. 2004; Yum et al. 1998; Yum and Hudson 2002, 2004; Hudson and Mishra 2007; Rosenfeld et al. 2008; Hudson et al. 2009; Hudson and Noble 2009; Hegg et al. 2012; Antilla et al. 2012) or any aerosol measurement with Nc (Leaitch et al. 1992, 1996; Twohy et al. 2005; Peng at al. 2005; Lu et al. 2007) Figs. 5b and 6a (red squares) and Fig. 7 (red squares) show inverse NCCNNc relationships for 50 horizontal cloud penetrations during MASE. Figure 5c puts data from the two projects together (Figs. 5a,b) to show that they overlap within the middle N1% range of 350–800 cm−3. Figure 5c then displays the decrease in the slope of NcN1%, which is referred to as roll off, which occurs because competition among droplets reduces cloud S as shown in Fig. 3. Furthermore, Fig. 5c shows more than a reduction of the slope, but here there is even a reversal of the NcN1% slope to negative values for the high N1% of MASE.

On the other hand, NCCN is not the only factor that determines Nc; the vertical velocity (W) is the second major influence on initial Nc. Consistent with this H10 showed positive relationships between W and Nc in POST though with lower R than NCCNNc—that is, Figs. 8a and 6b (black). Such expected positive relationships between W and Nc have been previously reported (Leaitch et al. 1996; Snider and Brenguier 2000; Conant et al. 2004; Peng et al. 2005; Hudson et al. 2010b). Since both NCCN and W should and indeed do display positive influences on Nc in POST a multiple regression analysis ought to show a higher R than R of either NCCNNc or WNc. However, such a multiple regression instead shows an R that only equals the R of NCCNNc. The reason for this curious result is the unexpected positive R between N1% and W (H10) that is of similar magnitude (0.54, Fig. 9a) to R of WNc (0.60, Fig. 8a). Gray diamonds in Fig. 7 show multiple significantly positive R of NCCNW at all measured S in POST. Thus, NCCN and W are not independent of each other in POST and this is why multiple regression does not make a higher R than the higher R of WNc or NCCNNc, which is the latter.

Fig. 8.

(a) Mean Nc vs W for 34 POST horizontal cloud passes and (b) 50 MASE horizontal cloud passes. Linear regressions are shown. Note the different units on the abscissas.

Fig. 8.

(a) Mean Nc vs W for 34 POST horizontal cloud passes and (b) 50 MASE horizontal cloud passes. Linear regressions are shown. Note the different units on the abscissas.

Fig. 9.

(a) Mean W vs mean N1% below the corresponding clouds during 34 POST horizontal cloud passes. (b) The σw vs below-cloud N1% during 50 MASE horizontal cloud passes.

Fig. 9.

(a) Mean W vs mean N1% below the corresponding clouds during 34 POST horizontal cloud passes. (b) The σw vs below-cloud N1% during 50 MASE horizontal cloud passes.

It is also somewhat unexpected that there are finite and somewhat large values of W within each POST cloud pass (e.g., Fig. 4a, blue line; maximum mean W range −40 to +40 cm s−1 in Figs. 8a and 9a) because stratus cloud W is usually thought to average zero over sufficient distances (e.g., Leaitch et al. 1996; Peng et al. 2005). Possibly the POST cloud passes were not long enough to produce near-zero mean W (Table 4). However, this did not seem to be the case for the MASE cloud penetrations, which consistently showed mean W so close to zero (maximum range −4 to +2 cm s−1; Fig. 8b) during all MASE horizontal cloud passes that the tiny mean W of the MASE clouds was not correlated with cloud microphysical variables, especially mean Nc (R = 0.04 in Fig. 8b and Table 6, 0.07 for flight averages—or with any Nt, Fig. 6c, black line). Although the MASE cloud passes were on average a factor of 3 longer than those of POST (Table 4, column 2) there was considerable overlap of the durations of the cloud passes between the two projects (Table 4, column 2, std dev). Although, absolute values of mean W are larger for the shorter MASE durations, R of NcW for the 10 MASE passes with durations less than the median duration of the POST passes (17 of 34) is 0.15. Furthermore, R for the 21 MASE passes with durations less than the mean of the POST passes (21 of 34) is 0.06. Thus, the durations of the cloud passes do not seem to be the reasons for the differences of the mean W between the two projects or the fact that mean W is related to Nc in POST but not MASE. It is unlikely that there were such fundamental differences in cloud dynamics between these two projects in the same area, season, and cloud type. More to the point, Twin Otter measurements obtained on the three MASE days when both airplanes flew (15–17 July) also showed the same magnitude differences in mean W during scores of horizontal cloud penetrations. Lu et al. (2009) also found similar large mean W in MASE-II. It is thus much more likely that the different methods of measuring and interpreting W between the two different aircraft were probably the reasons for the differences in apparent W behavior between the two projects. The slower speed and smaller size of the Twin Otter probably made it more difficult to discount the airplane motions in response to air motions. It may also be possible that the smaller, slower airplane has greater sensitivity to W or that its W measurement system is more sensitive.

It is often thought that in stratus clouds the variability of W, that is, the standard deviation of W (σw), can be a surrogate for W (Leaitch et al. 1996; Peng et al. 2005). Indeed Figs. 10b and 6c (red line) show that mean Nc and Nt seemed to respond to σw during MASE (σwNc, R = 0.51) and not to mean W (Fig. 8b; R = 0.04; Fig. 6c black line). An even higher R of 0.58 is found in MASE for σw with 90th percentiles of Nc (top decile, Ntd), which will be shown later to approximate adiabatic Nc (Na). In POST, σwNc was only 0.28 (Fig. 10a and green line in Fig. 6b where R for σwNt is lower than all R for WNt). Based on all of this information W is predominantly considered in POST and σw is predominantly considered in MASE. Moreover, in both projects W and σw are only considered in relative terms, that is, precision rather than accuracy; e.g., that mean W or σw was larger or smaller than mean W or σw during other situations—that is, Fig. 4. Indeed the negative mean W frequently observed during POST cannot produce clouds and the σw of MASE cannot be easily converted into mean W that could be used as input for adiabatic model predictions of Nc. Mean W and σw here serve only as indicators of the possible dynamic effects on cloud microphysics, that is, relative dynamic history of the air parcels.

Fig. 10.

As in Fig. 8, but plotted against standard deviation of W σw. Note the different abscissa scales.

Fig. 10.

As in Fig. 8, but plotted against standard deviation of W σw. Note the different abscissa scales.

Unlike the positive 0.54 R of WN1% in POST (Fig. 9a), in MASE Fig. 9b shows no correlation between σw and N1% (R = 0.005) or W and N1% or any NCCN with σw (Fig. 7, dark pink). Thus, unlike POST where the positive influences of NCCN and W on Nc are linked, the influences of NCCN and σw on Nc were not linked in MASE; thus they cannot reinforce each other in MASE as they do in POST.

Since NCCN and W should both have positive influences on Nc, the fact that all three of these relationships (NCCNNc, WNc, and WNCCN) are positive in POST suggests that the R of NCCNNc and WNc are both enhanced by the influence of the other factor. To illustrate the effect of the interactions of the two major influences on Nc we chose a subset of the POST data with R of WN1% that is neutral (uncorrelated). This is demonstrated by the green triangles in Fig. 11. The 23 green points in Fig. 11a have a neutral R of 0.01, which is quite different from the positive R of 0.54 for all 34 POST data points in Fig. 9a. Figure 11b then shows the lower 0.74 R for NcN1% for this data subset than the 0.86 R for the entire POST dataset (Fig. 5a). Then Fig. 11c shows that this same data subset has a neutral (R = −0.08) WNc relationship that thus does not enhance the NcN1% relationship as is the case for the entire POST data set where NcW is positively related (R = 0.60, Fig. 8a).

Fig. 11.

(a) As in Fig. 9a, but highlighting data with a neutral relationship for POST. (b) As in Fig. 5a, but with same highlighted data as in (a). (c) As in Fig. 8a, but with same highlighted data in (a),(b).

Fig. 11.

(a) As in Fig. 9a, but highlighting data with a neutral relationship for POST. (b) As in Fig. 5a, but with same highlighted data as in (a). (c) As in Fig. 8a, but with same highlighted data in (a),(b).

So far W and σw within the clouds have been considered. For the purposes of determining the dynamic influences on initial cloud microphysics (i.e., Nc) W or σw measurements should have been made just below or just above cloud base. But these are stratus clouds, which have longer lifetimes that do not fit the cloud base formation processes of cumulus clouds. At any rate, Fig. 6b (POST) shows that W and σw measured below cloud also showed positive R with small size threshold Nt (<~10 μm), albeit with lower R than within-cloud W and σw, except that the more extensive below-cloud σw (low) has a higher R than the within-cloud σw with Nt. Furthermore, Fig. 6c (MASE) shows that both below-cloud σw measurements have nearly as high R with small size threshold Nt as the within-cloud σw measurements have with small threshold Nt.

5. Effects on cloud supersaturation Seff

H10 noted that estimates of Seff by comparing CCN spectra with mean Nc are underestimates of the S at cloud base that determines initial Nc, which is the most closely related to NCCN (e.g., Hegg et al. 2012). This is because mean Nc has often been reduced by entrainment, which is not expected to depend on aerosol characteristics. Entrainment also inherently reduces LWC below adiabatic values. Figure 12 demonstrates estimates of adiabatic Nc in POST by using the ratio of measured LWC to adiabatic LWC (LWCadia) (Hudson and Yum 2001, 2002). LWCadia is calculated from cloud-base temperature, pressure, and altitude for the altitude of each Nc measurement. The assumption of identical cloud-base altitude, pressure, and temperature throughout each horizontal cloud penetration is an uncertainty of this technique and so is the difficulty of ascertaining cloud base, which was done during vertical slant ascents or descents prior or after each horizontal cloud pass. During POST 31 of the 34 horizontal cloud passes displayed patterns similar to the examples in Fig. 12 where there are linear relationships between measured Nc and measured LWC/LWCadia for which extrapolations of the regressions provide Nc at LWC/LWCadia = 1.0. These adiabatic Nc are Na, which should be the Nc that were produced at cloud base, which should be the Nc most closely related to the input CCN. The Na are then compared with below-cloud NCCN(S) to produce alternate Seff estimates that are displayed in Fig. 13a, which also shows Seff based on mean Nc that were displayed in Fig. 3b. Figure 13a also uses another estimate of Na, the 90th percentile of Nc measurements Ntd within each cloud pass. The Ntd was also employed by H10. Figure 13b shows similar Seff estimates for MASE, absent Na because Na could not always be estimated because of the difficulty of determining cloud base in MASE. The agreement between the Na and Ntd estimates of Seff in Fig. 13a (POST) gives credence to the utility of the Ntd estimates of Seff in MASE (Fig. 13b). Figure 13c shows Seff consistency between POST and MASE within the overlapping N1% range (380–800 cm−3) and the overall decrease of Seff with N1% that was shown in Fig. 3. Table 7 shows the average Seff within three N1% bins for each project. Because of the vast difference in N1% between the two projects it is not practical to use the same bins in each project because they would not have a sufficient number of cases. Nevertheless, Table 7 does show rather good agreement between the two projects, both of which show the decrease of Seff with N1%, and the higher Seff by using mean Ntd rather than mean Nc as well as the slightly higher but agreeable Seff by using mean Na for POST that is also displayed in Fig. 13a.

Fig. 12.

1-s Nc against ratio of measured LWC to calculated LWCadia for two different POST cloud passes out of the 34 POST horizontal cloud passes.

Fig. 12.

1-s Nc against ratio of measured LWC to calculated LWCadia for two different POST cloud passes out of the 34 POST horizontal cloud passes.

Fig. 13.

(a) As in Fig. 3b, but with the addition of estimates based on Na and Ntd. (b) As in (a), but for MASE. (c) Combination of (a),(b) for NtdSeff.

Fig. 13.

(a) As in Fig. 3b, but with the addition of estimates based on Na and Ntd. (b) As in (a), but for MASE. (c) Combination of (a),(b) for NtdSeff.

Table 7.

The Seff within various N1% bins for each project. Columns 2 and 6 are based on mean Nc, columns 3 and 7 are based on Ntd, and columns 4 and 8 are based on Na.

The Seff within various N1% bins for each project. Columns 2 and 6 are based on mean Nc, columns 3 and 7 are based on Ntd, and columns 4 and 8 are based on Na.
The Seff within various N1% bins for each project. Columns 2 and 6 are based on mean Nc, columns 3 and 7 are based on Ntd, and columns 4 and 8 are based on Na.

Low Seff so frequently displayed in Figs. 3 and 13, especially for MASE, indicate that N1% might not be the most appropriate NCCN. Figure 14, however, shows that R for Seff with NCCN at various S is consistent over wide ranges of S in both projects. Although the slopes and R of the regressions are smaller for SeffNCCN at lower S they are not positive for any SeffNCCN in either project. In POST, NCCNNc relationships are positive at all S, albeit the slopes are higher and the R values are progressively smaller for lower S (Fig. 7; black, blue and green data points).

Fig. 14.

The R between Seff and NCCN at various S as a function of S for (a) POST and (b) MASE.

Fig. 14.

The R between Seff and NCCN at various S as a function of S for (a) POST and (b) MASE.

Though higher W should produce higher Seff, Fig. 15a shows negative R for SeffW. This occurs because of the coupling of W with N1% in POST (Fig. 9a). The gray diamonds in Fig. 7 show that in POST this WNCCN coupling exists for NCCN at all measured S. Therefore, higher NCCN comes along with the higher W and this makes even higher Nc than that due to just the higher W. The greater droplet surface area of the higher Nc reduces Seff because of the competition among the droplets for condensate as demonstrated in Figs. 3 and 13. Thus, in POST the negative NCCN effect on Seff prevails over the positive W influence on Seff. This is demonstrated by the black, blue, and green data points in Fig. 7 where R for NCCNNc for S > 0.08% is greater than the 0.60 R for WNc shown in Fig. 8a. Figure 15b, on the other hand, does show positive R for Seff–σw because the lack of coupling between σw and NCCN in MASE (Fig. 9b and dark pink in Fig. 7) does not bring higher NCCN along with higher W. This allows the positive influence of σw on Nc (Fig. 10b) to go unhindered in MASE and thus produce the positive R of Seffσw shown in Fig. 15b. Moreover, the strong positive influence of NCCN on Nc in POST shown in Fig. 5a and further in Fig. 7 (black, blue, and green data points for nearly all S) does not exist in MASE (Figs. 5b) where this R is even negative for NCCN at all S as shown by the red and pink data points in Fig. 7. The Seff calculated from mean Nc, Na, or Ntd show the same trends with W in POST and with σw in MASE.

Fig. 15.

The Seff based on Ntd for (a) POST vs W and (b) Seff vs σw for MASE.

Fig. 15.

The Seff based on Ntd for (a) POST vs W and (b) Seff vs σw for MASE.

Although it might seem that Nc should be positively related to cloud S, Fig. 16 displays a negative SeffNc relationship for POST. This is because the lower Seff estimates occur when NCCN is higher, which then produce higher Nc—that is, even though a smaller proportion of CCN activate to droplets when Seff is lower the number of activated droplets is still larger than is the case for cleaner conditions when greater percentages of CCN activate out of the lower NCCN. This is elucidated in Table 3, column 13, by the higher mean activation ratio (Nc/N1%) of POST. Thus, in POST Nc is still higher when NCCN is higher because of the greater influence of NCCN than W on Nc and the positive coupling of NCCN with W. In MASE the lack of apparent influence of NCCN on Nc allows the σw influence on Nc to prevail. Thus, higher σw force a higher percentage of CCN to activate to droplets, which generally produces higher Nc and Seff. The lower Seff for higher NCCN in MASE (N1% in Figs. 13b,c) results from the lack of coupling between σw and NCCN, that is, within each NCCN (N1%) band a similar mix of σw will activate a smaller percentage of CCN to droplets when NCCN is higher and a larger percentage of NCCN to droplets when NCCN is lower thus making the descending SeffN1% relationships shown in Figs. 13b and 13c. The fact that there are different reasons for the similar relationships of the two projects in Fig. 13 is reflected in the very different relationships shown in Fig. 16—that is, the predominant NCCN influence in POST and the predominant σw influence in MASE.

Fig. 16.

The Seff based on Ntd vs mean Nc.

Fig. 16.

The Seff based on Ntd vs mean Nc.

The difference between the two major influences on Nc exhibited by POST and MASE was predicted by T59. That explanation lies in the differences of the slopes (k) of the cumulative CCN spectra at and below Seff shown in Fig. 1b and Table 3, last column. At the lower values of Seff of MASE k is higher than the lower k at the higher Seff of most of POST. Table 8 displays the exponents of N1% and W of the T59 equation

 
formula

for various k. Equation (1) and Table 8 demonstrate the increasing influence of W and decreasing influence of NCCN as k increases. Twomey (1977b, p. 104) presented the same equation and pointed out that for k ≪ 1.0, Nc is approximately proportional to N1% but for k ≫ 1.0, Nc is approximately proportional to W, “Had the slope of natural supersaturation spectra proved to be large the drop concentration in cloud would have been determined almost exclusively by the dynamics of the environment in which the cloud formed rather than the aerosol content.” The same differences of W or σw produce greater differences in Nc at higher k than at lower k. For MASE the transition from predominant NCCN influence to predominant σw influence on Nc occurred at a lower k (<0.8) than predicted by T59 (k = 1.33).

Table 8.

Exponents of the two main factors that determine Nc by the equation of Twomey (1959).

Exponents of the two main factors that determine Nc by the equation of Twomey (1959).
Exponents of the two main factors that determine Nc by the equation of Twomey (1959).

6. Divided cloud passes

Table 9 details the six divided POST horizontal cloud passes, an example of which is displayed in Fig. 4a. The second and third columns show the mean W of the higher and lower W portions of each divided cloud leg. The fourth, fifth, and sixth columns show the mean Nc, Na, and Ntd of each of the respective portions of each flight leg. Mean Nc is higher for the leg portions with higher mean W in four of the six cases while mean Na and Ntd are higher for five of the six leg portions with higher mean W. Mean Na and Ntd are more pertinent to initial NCCN and W because mean Nc are more likely to be reduced by entrainment. The two Nc exceptions, 28 July and 1 August, have the smallest and third smallest mean W differences and the smallest mean Nc differences between the two divisions of their cloud passes out of the six complete cloud passes considered in Table 9. The 1 August exception for all three droplet concentrations (lower for the higher W portion) has nearly the lowest Ntd difference; 28 July has only 1 cm−3 smaller Ntd difference in the opposite direction, higher for higher W. The 1 August exception has by far the lowest LWC and average ratio of measured to adiabatic LWC of any of the passes (column 7), which makes a much greater and more uncertain extrapolation of the linear regression of the data points to LWC/LWCadia = 1 that is needed to estimate Na (see Fig. 12). Furthermore, the distribution of data points (e.g., Fig. 12) for the high mean W portion of the 1 August distribution is flat whereas this distribution for the lower W portion is very steep. These are the reasons for the lower Na estimate for the high mean W portion of this leg and the higher Na estimate for the lower mean W portion of the 1 August divided cloud.

Table 9.

Six pairs of POST adjacent cloud segments divided according to abrupt major differences in mean W and Nc; example in Fig. 4a. Column 2 is the mean W of the cloud portion with the higher mean W, column 3 is the mean W of the lower mean W portion, column 4 shows the mean Nc of the higher and lower W portions of each cloud leg, column 5 shows the Na of each cloud portion, column 6 is the Ntd of each cloud division, column 7 is the mean cloud (CAS) LWC of each cloud section, column 8 is mean drizzle (CIP) LWC of each section, column 9 is σw of each section, and column 10 shows the durations of each cloud leg. The bottom rows labeled x1A are the mean and std dev of the five rows exclusive of 1 Aug.

Six pairs of POST adjacent cloud segments divided according to abrupt major differences in mean W and Nc; example in Fig. 4a. Column 2 is the mean W of the cloud portion with the higher mean W, column 3 is the mean W of the lower mean W portion, column 4 shows the mean Nc of the higher and lower W portions of each cloud leg, column 5 shows the Na of each cloud portion, column 6 is the Ntd of each cloud division, column 7 is the mean cloud (CAS) LWC of each cloud section, column 8 is mean drizzle (CIP) LWC of each section, column 9 is σw of each section, and column 10 shows the durations of each cloud leg. The bottom rows labeled x1A are the mean and std dev of the five rows exclusive of 1 Aug.
Six pairs of POST adjacent cloud segments divided according to abrupt major differences in mean W and Nc; example in Fig. 4a. Column 2 is the mean W of the cloud portion with the higher mean W, column 3 is the mean W of the lower mean W portion, column 4 shows the mean Nc of the higher and lower W portions of each cloud leg, column 5 shows the Na of each cloud portion, column 6 is the Ntd of each cloud division, column 7 is the mean cloud (CAS) LWC of each cloud section, column 8 is mean drizzle (CIP) LWC of each section, column 9 is σw of each section, and column 10 shows the durations of each cloud leg. The bottom rows labeled x1A are the mean and std dev of the five rows exclusive of 1 Aug.

It must be assumed that NCCN is the same for both sections of each of the divided cloud passes. The NCCN differences on scales as small as the two divisions of each of the cloud passes are unlikely (Table 5, columns 5–7) and thus unlikely to be the cause of the Nc differences (Table 5, columns 2–4) between the divisions of the cloud passes (Table 6), especially since the W differences are generally commensurate with the Nc, Na, and Ntd differences that are shown in Table 9. The R of the linear regressions N1%Nc, N1%Ntd, WNc, and WNtd are all higher for the 12 data points in Table 9 than the corresponding R values for the entire 34 case dataset.

Table 10 shows the analogous set of divided cloud passes from MASE; Fig. 4b is an example. The running σw in Fig. 4b do not correspond with the overall σw because they are with respect to the 50 record (5 s) running-mean W whereas the overall σw is with respect to the mean W over the entirety of each cloud section. In all 12 cases there were substantially higher Nc (column 4) and Ntd (column 5) for the legs with higher σw. Moreover, the σw differences between each of the adjacent legs were usually substantial. Column 8 shows that mean W was lower for 5 of the 12 legs with higher σw, Nc, and Ntd, thus again mean W does not correlate in MASE. As in POST the same NCCN is assumed for both divisions of each of the 12 cloud legs and it seems unlikely that NCCN would display differences on such a small scale (i.e., Table 5, columns 5–7) that are as drastic as the Nc differences (Table 5, columns 2–4). Since the Nc and Ntd differences between each of the adjacent cloud legs correspond to the σw differences, it is likely that σw is the cause of the Nc and Ntd differences in each of the 12 adjacent cloud legs. This is especially so since 10 of the 12 Nc and Ntd values are more than 39% and 21% higher in the higher σw than the lower σw sections of each leg. Furthermore, 4 of 12 cases have more than a factor of 2 higher Nc and more than 61% higher Ntd in the higher σw sections of each leg. The R values of N1%Nc, N1%Ntd, σwNc, and σwNtd for these 24 cases are all similar to the corresponding R values of the entire 50-case MASE dataset. The huge Nc and Ntd differences for the same aerosol inputs and the negative R for NCCNNc and NCCNNtd and the consistently higher Nc and Ntd for the higher σw divisions of each of the 12 divided cloud legs indicate the impossibility that any aerosol parameter can positively correlate with Nc or Ntd in MASE.

Table 10.

As in Table 9, but for MASE. The 12 pairs of adjacent divided cloud passes based on abrupt differences in σw and Nc (Fig. 4b is an example). Column 2 is σw of the higher σw portion, column 3 is σw of the lower σw portion, column 4 is the mean Nc of the higher and lower σw portions, column 5 is Ntd of the higher and lower σw portion, column 6 is the cloud (CAS) LWC of the higher and lower σw portions, column 7 is the drizzle LWC (CIP) of the higher and lower portions, column 8 is the mean W for the higher and lower σw portions, and column 9 is the durations of the higher and lower σw portions of each adjacent cloud pass.

As in Table 9, but for MASE. The 12 pairs of adjacent divided cloud passes based on abrupt differences in σw and Nc (Fig. 4b is an example). Column 2 is σw of the higher σw portion, column 3 is σw of the lower σw portion, column 4 is the mean Nc of the higher and lower σw portions, column 5 is Ntd of the higher and lower σw portion, column 6 is the cloud (CAS) LWC of the higher and lower σw portions, column 7 is the drizzle LWC (CIP) of the higher and lower portions, column 8 is the mean W for the higher and lower σw portions, and column 9 is the durations of the higher and lower σw portions of each adjacent cloud pass.
As in Table 9, but for MASE. The 12 pairs of adjacent divided cloud passes based on abrupt differences in σw and Nc (Fig. 4b is an example). Column 2 is σw of the higher σw portion, column 3 is σw of the lower σw portion, column 4 is the mean Nc of the higher and lower σw portions, column 5 is Ntd of the higher and lower σw portion, column 6 is the cloud (CAS) LWC of the higher and lower σw portions, column 7 is the drizzle LWC (CIP) of the higher and lower portions, column 8 is the mean W for the higher and lower σw portions, and column 9 is the durations of the higher and lower σw portions of each adjacent cloud pass.

Questions about the viability of the W and σw measurements are answered in Tables 9 and 10 where adjacent pairs of cloud passes display distinct contrasts in Nc and W (POST) or σw (MASE) (Fig. 4). Another possibility for the cause of the droplet concentration differences would be the injection of higher NCCN from above the clouds, which might be more likely with higher W or σw. However, H10 showed that in POST there was no difference in the N1%Nc R for situations with higher total particle (condensation nuclei; NCN) concentrations above the clouds than below the clouds than for situations with the same or lower NCN above the clouds compared to below the clouds. The NCN was used for this purpose because the faster response of the CN counter provided better separation from the undesired splashing artifacts of in-cloud measurements and since NCCN are generally proportional to NCN. Above-cloud NCCN did not seem to be the cause of the differences in three of the six POST cases because there were lower or the same NCCN above the clouds for these cases. The large differences of mean W (POST) and σw (MASE) between the divisions of each of the divided cloud passes are more than adequate to produce the observed mean Nc differences between the divisions of each pass without any help of injections of above-cloud aerosol.

7. CCN–vertical velocity relationships

The correlation between NCCN and mean W in POST may just be coincidental—that is, the airplane happened to go through high W cloud regions when NCCN was higher. Nonetheless, the enhancement of Nc by higher W was a fact to deal with regardless of the cause. Moreover, if stratus clouds have near-zero mean W over sufficient distances (Leaitch et al. 1996; Peng et al. 2005) then WNCCN coupling does seem merely coincidental. However, this does not preclude σwNCCN coupling in stratus. Although σw in POST does not display as high R as W (Fig. 8a) with either Nc (Fig. 10a) or Nt (Fig. 6b) or NCCN (Fig. 17) all Rs for σw in POST are consistently positive, except those with Nt for thresholds >10 μm; this is similar to other R values for these large sizes in Fig. 6b.

Fig. 17.

The R for WNCCN in POST and σwNCCN for POST and MASE. POST W cloud (gray diamonds) is same as gray diamonds in Fig. 7, MASE cloud (red squares) is same as dark pink diamonds in Fig. 7. CCN refers to σw measurements at the same subcloud locations as the CCN measurements, and low refers to σw measured over longer below-cloud distances.

Fig. 17.

The R for WNCCN in POST and σwNCCN for POST and MASE. POST W cloud (gray diamonds) is same as gray diamonds in Fig. 7, MASE cloud (red squares) is same as dark pink diamonds in Fig. 7. CCN refers to σw measurements at the same subcloud locations as the CCN measurements, and low refers to σw measured over longer below-cloud distances.

There are two theories that suggest σwNCCN coupling, both of which involve differential latent heat exchange. Jiang et al. (2002, 2010) and Ackerman et al. (2004) suggested that since evaporation of drizzle below cloud tends to stabilize the boundary layer, precipitation suppression by higher NCCN should result in reduced stabilization of the boundary layer. Such a reduction of evaporative cooling below cloud would then increase TKE and buoyancy, which would result in an increase of σw. This should then result in a negative R between drizzle (LWC of the CIP, Ld) and σw below cloud base where evaporation of drizzle occurs. Table 11, rows 1–3, shows that in both projects there is less precipitation when Nc is higher, which tends to be associated with smaller mean diameter (MD; row 2) and narrower droplet spectra (σc; row 3), which inhibit Ld. Row 4 shows how Ld is associated with lower mean W in POST, but again mean W has no associations in MASE. Row 5 shows that lower within-cloud σw may promote Ld in both projects. Row 6 shows that below-cloud Ld (measured at the same locations and times as the CCN measurements) is well correlated with within-cloud Ld in both projects. Row 7 shows that σw measured at the CCN measurement locations and times is related to σw measured within cloud in POST but not in MASE. Row 8 shows that in both projects Ld below cloud is nearly as negatively correlated with cloud droplet concentrations as is Ld within cloud (row 1). Row 9 shows that below-cloud σw is negatively correlated with below-cloud Ld in MASE but not in POST. This indicates the differential latent heat release effect on dynamics suggested by Jiang et al. (2002) and Ackerman et al. (2004) in MASE but not in POST.

Table 11.

The R between drizzle LWC (Ld) within cloud (Ldcld) and mean Nc, mean diameter of droplet spectra (MD), std dev of droplet spectra (σc), within-cloud W, within-cloud σw, and Ld below cloud where CCN were measured. Then row 7 is the R between within-cloud and below-cloud (where CCN were measured) σw. Then row 8 is the R between below-cloud Ld and Nc. Then row 9 is the R between below-cloud Ld and σw below cloud. Then row 10 is the below-cloud Ld with N1% and then within-cloud Ld with N1% (row 11) and then within-cloud σw (row 12) with N1%, and finally below-cloud σw with N1%.

The R between drizzle LWC (Ld) within cloud (Ldcld) and mean Nc, mean diameter of droplet spectra (MD), std dev of droplet spectra (σc), within-cloud W, within-cloud σw, and Ld below cloud where CCN were measured. Then row 7 is the R between within-cloud and below-cloud (where CCN were measured) σw. Then row 8 is the R between below-cloud Ld and Nc. Then row 9 is the R between below-cloud Ld and σw below cloud. Then row 10 is the below-cloud Ld with N1% and then within-cloud Ld with N1% (row 11) and then within-cloud σw (row 12) with N1%, and finally below-cloud σw with N1%.
The R between drizzle LWC (Ld) within cloud (Ldcld) and mean Nc, mean diameter of droplet spectra (MD), std dev of droplet spectra (σc), within-cloud W, within-cloud σw, and Ld below cloud where CCN were measured. Then row 7 is the R between within-cloud and below-cloud (where CCN were measured) σw. Then row 8 is the R between below-cloud Ld and Nc. Then row 9 is the R between below-cloud Ld and σw below cloud. Then row 10 is the below-cloud Ld with N1% and then within-cloud Ld with N1% (row 11) and then within-cloud σw (row 12) with N1%, and finally below-cloud σw with N1%.

On the other hand Fig. 18 verifies this below-cloud negative Ldσw relationship in POST if only the 6 cases with substantial drizzle are considered (Ld > 5 mg m−3) or the 10 cases with relative humidity (RH) less than 88.5% where the CCN were measured. There should be more significant evaporation where RH is lower. Furthermore, if only the 19 cases with σw greater than 40 cm s−1are considered R = −0.33 for LdCCNσwCCN. Rows 10 and 11 of Table 11 show that drizzle is negatively related to NCCN in POST but positively related to NCCN in MASE. Rows 12 and 13 of Table 11 show that σw within and below cloud are related to N1% in POST but not in MASE. Figure 17 shows that similar R values are found for σw with NCCN at all S in both projects. Evidence of this differential drizzle evaporation effect as a cause of the coupling between σw and NCCN in POST is not substantial. Although there is significant evidence that differential drizzle evaporation causes differential below-cloud σw in MASE, it is not due to NCCN but rather to within-cloud σw that inversely promotes Ld, that is, low σwcld promotes lower Nc, greater MD, and σc, which promote Ld. Consistent with this, Table 10, column 7, shows greater drizzle in 11 of the 12 divided cloud penetrations for the divisions with lower σw. The one exception (20 July middle) has the second-smallest σw difference and the smallest Ld percentage difference. There is a factor of 2 greater mean drizzle for the lower σw divisions of MASE (Table 10, column 7) as well as for the lower W divisions of POST (Table 9, column 8) even though half of the six POST cases show greater drizzle in the higher W divisions (by small margins).

Fig. 18.

POST drizzle drop LWC (Ld) measured below cloud where CCN were measured against σw at the same location.

Fig. 18.

POST drizzle drop LWC (Ld) measured below cloud where CCN were measured against σw at the same location.

Another theory that would connect σw with NCCN stems from the realization that smaller cloud droplets evaporate more readily (Xue and Feingold 2006; Jiang et al. 2002; Zuidema et al. 2008; Xue et al. 2008); some evidence of this in the RICO experiment was presented by Hudson et al. (2009). The resulting greater latent heat exchange of the greater evaporation in more polluted clouds adds TKE and buoyancy gradients that can enhance W, mixing, and entrainment (Blyth et al. 1988; Zhao and Austin 2005). The stronger W in the more polluted clouds can also lead to more horizontal motions, and this can further enhance evaporation of droplets, which thus further enhances latent heat exchange and vertical motions, thus, positive feedback. However, these cloud modeling simulations were applied to clouds of a more cumulus nature than the stratocumuli of POST and MASE, especially MASE, which was more stratus than stratocumulus, that is, more solid. At any rate, it might be possible that the WNCCN and σwNCCN coupling in POST is due to greater latent heat exchanges in the more polluted stratus. This could also include latent heat released during condensation (Lee and Feingold 2010; Storer and van den Heever 2013), which is more rapid for the greater surface areas of the smaller more numerous droplets of more polluted clouds. Especially for stratus clouds, these differential latent heat theories imply a positive relationship between cloud σw and NCCN rather than W and NCCN. This is a suggestion of greater turbulence in polluted stratus that could result from the greater latent heat exchanges when there is greater NCCN and Nc.

Figure 17 shows positive R for all σwNCCN in POST. This includes not only σw measured within the cloud passes but also σw measurements at two different sets of below-cloud locations, the same places and times where and when the presented CCN measurements were made and over longer distances and time periods that include those of the CCN measurements. It might be notable that the R values for within-cloud σw measurements with NCCN exceed R of the below-cloud σw measurements with NCCN at low S that are more characteristic of the Seff of the more polluted clouds, which should have the greater latent heat exchanges. Higher R for within-cloud σw with NCCN than below-cloud σw with NCCN might be expected for this process, although the increased turbulence because of higher NCCN might be efficiently imparted to below-cloud air in such shallow boundary layers. As should be expected, mean σw of POST was 29% higher in-cloud than below cloud at the CCN locations and 12% higher than the more extensive below-cloud σw measurements—that is, greater turbulence within clouds. Also to be expected the standard deviation (std dev) and relative std dev of the within-cloud σw was greater than the below-cloud σw measurements, that is, greater turbulence within clouds. Low R values such as shown in Fig. 17 are sometimes scoffed, but the significance level (sl) for the 34 POST data points is ~98% for the low S σwNCCN and > 85% at high S.

The only positive σwNCCN Rs in MASE are at low S where the sl of the 50 data points does exceed 90%. The uniformly high NCCN and Nc of MASE may not produce as much differential latent heat exchange as occurs within the lower NCCN and Nc ranges of POST; that is, this latent heat effect on dynamics may not continue to increase at high NCCN just as Nc does not seem to continue to increase at high NCCN (i.e., Fig. 5c; roll off).

8. Conclusions

Results of two contrasting stratus cloud airborne field experiments (POST and MASE) reveal the following characteristics that may or may not be universal. Stratus cloud supersaturations Seff decrease with higher CCN concentrations NCCN as predicted by T59 (Fig. 13). In clean marine air (i.e., N1% < 200 cm−3) Seff often exceeded 1%. In polluted stratus (N1% > 700 cm−3) average Seff was ~0.1%. This indicates that much smaller particles (i.e., 20 nm) than conventional wisdom (>60 nm) can nucleate stratus cloud droplets (H10), especially in cleaner air masses, which are more susceptible to IAE (Platnick and Twomey 1994). Thus, more numerous smaller particles are capable of inflicting IAE.

Stratus cloud droplet concentrations Nc increase with NCCN up to at least N1% = 400 cm−3 above which there is a roll off of Nc with further NCCN increase (Fig. 5, Table 6, column 3). Roll off of Nc with various aerosol measurements other than CCN has been observed by Raga and Jonas (1993), Martin et al. (1994), Twohy et al. (2005), and Lu et al. (2007; 2008) but not by Bowers et al. (2000). This roll off of Nc with NCCN then tends to limit IAE. Vertical velocity W or the standard deviation of W (σw) have a secondary positive effect on Nc that became the primary effect on Nc at high NCCN. T59 showed that differences in the slope k of CCN spectra might account for the greater–overwhelming W or σw influence on Nc at the high NCCN of MASE. Although Fig. 1b shows no significant differences in mean k between POST and MASE, Fig. 13 shows large differences in cloud effective S (Seff) between the two projects. At the lower Seff of MASE there is higher k than k at the higher Seff of most of POST. Although k is just as high at low S in POST, Seff in POST was seldom so low. Thus, in POST the low k values at high S are the most relevant for determining cloud microphysics (i.e., Nc). The low Seff forced by the higher NCCN of MASE makes irrelevant the lower k values observed at high S that were seldom achieved in MASE. Thus, the high k values most relevant to MASE reduce the relationship between NCCN and Nc and enhance the relationship between W or σw and Nc (T59). This reduction of NCCN and enhancement of W or σw influences seems to occur at smaller k than predicted by T59. However, T59 did not consider variations in k over S, which are observed to increase to even higher values at S lower than Seff (Fig. 1). The NCCNs at all S < Seff are important for producing Nc. Furthermore, although k consistently increased with decreasing S, the CCN spectra were different for each flight and each cloud penetration. This makes a more complicated situation than considered by T59.

A coupling of W (and σw) with NCCN enhanced the correlations of each with Nc. The NCCNσw coupling could be a result of two different responses to different levels of latent heat exchange between clean and more polluted clouds. One effect is through below-cloud differential drizzle evaporation due to drizzle suppression in more polluted clouds. The other process is greater cloud droplet evaporation of smaller polluted cloud droplets. Both of these processes are predicted to preferentially enhance turbulence in more polluted clouds, which can be expressed by σw. A limited amount of evidence of these effects has been presented for the POST project, which has a greater range of NCCN and Nc and displays the WNCCN and σwNCCN coupling. These results are consistent with Chen et al. (2012), who found greater W variance in ship trail clouds than in adjacent natural clouds. Such coupling between CCN and cloud dynamics tends to support effects opposite of conventional IAE (Xue and Feingold 2006); less cloudiness with higher NCCN.

The predominating effect of vertical velocity variations (σw) in the MASE polluted clouds precluded positive correlations of any aerosol parameter with cloud microphysics in MASE. However, this does not mean the irrelevance of the aerosol for cloud microphysics because it is the high NCCN and high k that produce the conditions where σw has so much influence on stratus cloud microphysics. The Seff also decreased with Nc in POST because of the suppression of Seff with NCCN. But in the more polluted MASE clouds, Seff increased with Nc because of the lack of correlation between NCCN and Nc and the high k at the relevant S of the CCN spectra.

Acknowledgments

MASE was supported by U.S. DOE Grant DE-FG02-05ER63999 and DE-SC0009162. POST was supported by U.S. NSF ATM-0734441. For POST Haflidi Jonsson of CIRPAS, Naval Postgraduate School provided the cloud data while Djamal Khelif of University of California, Irvine, provided the W measurements. Gunnar Senum of Brookhaven National Laboratory provided the cloud and W data for MASE.

REFERENCES

REFERENCES
Ackerman
,
A. A.
,
M. P.
Kirkpatrick
,
D. E.
Stevens
, and
O. B.
Toon
,
2004
:
The impact of humidity above stratiform clouds on indirect aerosol climate forcing
.
Nature
,
432
,
1014
1017
.
Albrecht
,
B. A.
,
1989
:
Aerosols, cloud microphysics, and fractional cloudiness
.
Science
,
245
,
1227
1230
.
Alley
,
R. B.
, and
Coauthors
,
2007
: Summary for policymakers. Climate Change 2007: The Physical Science Basis, S. Solomon et al., Eds., Cambridge University Press, 1–18.
Antilla
,
T.
,
D.
Brus
,
A.
Jaatinen
,
A.-P.
Hyvarinen
,
N.
Kivekas
,
S.
Romakkaniemi
,
M.
Komppula
, and
H.
Lihavainen
,
2012
:
Relationships between particles, cloud condensation nuclei and cloud droplet activation during the third Pallas Cloud Experiment
.
Atmos. Chem. Phys.
,
12
,
11 435
11 450
,
doi:10.5194/acp-12-11435-2012
.
Baumgardner
,
D.
,
H.
Jonsson
,
W.
Dawson
,
D.
O’Connor
, and
R.
Newton
,
2001
:
The cloud, aerosol and precipitation spectrometer: A new instrument for cloud investigations
.
Atmos. Res.
,
59–60
,
251
264
.
Blyth
,
A. M.
,
W. A.
Cooper
, and
J. B.
Jensen
,
1988
:
A study of the source of entrained air in Montana cumuli
.
J. Atmos. Sci.
,
45
,
3944
3964
.
Bowers
,
K. N.
, and
Coauthors
,
2000
:
ACE-2 HILLCLOUID, An overview of the ACE-2 ground-based cloud experiment
.
Tellus
,
52B
,
750
778
.
Brown
,
E. N.
,
C. A.
Friehe
, and
D. H.
Lenschow
,
1983
:
The use of pressure fluctuations on the nose of an aircraft for measuring air motion
.
J. Climate Appl. Meteor.
,
22
,
171
180
.
Carman
,
J. K.
,
D. L.
Rossiter
,
D.
Khelif
,
H. H.
Jonsson
,
I. C.
Faloona
, and
P. Y.
Chuang
,
2012
:
Observational constraints on entrainment and the entrainment interface layer in stratocumulus
.
Atmos. Chem. Phys.
,
12
,
11 135
11 152
.
Chan
,
K. R.
,
J.
Dean-Day
,
S. W.
Bowen
, and
T. P.
Bui
,
1998
:
Turbulence measurements by the DC-8 meteorological measurement system
.
Geophys. Res. Lett.
,
25
,
1355
1358
.
Chen
,
Y.-C.
,
M. W.
Christensen
,
L.
Xue
,
A.
Sorooshian
,
G. L.
Stephens
,
R. M.
Rassmussen
, and
J. H.
Seinfeld
,
2012
:
Occurrence of lower albedo in ship tracks
.
Atmos. Chem. Phys.
,
12
,
8223
8235
.
Conant
,
W. C.
, and
Coauthors
,
2004
:
Aerosol-cloud drop concentration closure in warm cumulus
.
J. Geophys. Res.
,
109
,
D13204
,
doi:10.1029/2003JD004324
.
Guo
,
H.
,
Y.
Liu
,
P. H.
Daum
,
G. I.
Senum
, and
W.-K.
Tao
,
2008
:
Characteristics of vertical velocity in marine stratocumulus: Comparison of large eddy simulations with observations
.
Environ. Res. Lett.
,
3
, 045020,
doi:10.1088/1748-9326/3/4/045020
.
Hegg
,
D. A.
,
D. S.
Covert
,
H. H.
Jonsson
, and
R.
Woods
,
2009
:
Differentiating natural and anthropogenic cloud condensation nuclei in the California coastal zone
.
Tellus
,
61B
,
669
676
.
Hegg
,
D. A.
,
D. S.
Covert
,
H. H.
Jonsson
, and
R.
Woods
,
2012
:
A simple relationship between cloud drop number concentration and precursor aerosol concentration for the regions of the earth’s large marine stratocumulus decks
.
Atmos. Chem. Phys.
,
12
,
1229
1238
,
doi:10.5194/acp-12-1229-2012
.
Hoppel
,
W. A.
,
G. M.
Frick
, and
R. E.
Larson
,
1986
:
Effect of nonprecipitating clouds on the aerosol size distribution in the marine boundary layer
.
Geophys. Res. Lett.
,
13
,
125
128
.
Hudson
,
J. G.
,
1983
:
Effects of CCN on stratus clouds
.
J. Atmos. Sci.
,
40
,
480
486
.
Hudson
,
J. G.
,
1989
:
An instantaneous CCN spectrometer
.
J. Atmos. Oceanic Technol.
,
6
,
1055
1065
.
Hudson
,
J. G.
,
1993
:
Cloud condensation nuclei
.
J. Appl. Meteor.
,
32
,
596
607
.
Hudson
,
J. G.
, and
P. R.
Frisbie
,
1991
:
Cloud condensation nuclei near marine stratus
.
J. Geophys. Res.
,
96
(
D11
),
20 795
20 808
.
Hudson
,
J. G.
, and
S. S.
Yum
,
2001
:
Maritime–continental drizzle contrasts in small cumuli
.
J. Atmos. Sci.
,
58
,
915
926
.
Hudson
,
J. G.
, and
S. S.
Yum
,
2002
:
Cloud condensation nuclei spectra and polluted and clean clouds over the Indian Ocean
.
J. Geophys. Res.
,
107
,
8022
,
doi:10.1029/2001JD000829
.
Hudson
,
J. G.
, and
S.
Mishra
,
2007
:
Relationships between CCN and cloud microphysics variations in clean maritime air
.
Geophys. Res. Lett.
,
34
,
L16804
,
doi:10.1029/2007GL030044
.
Hudson
,
J. G.
, and
S.
Noble
,
2009
:
CCN and cloud droplet concentrations at a remote ocean site
.
Geophys. Res. Lett.
,
36
,
L13812
,
doi:10.1029/2009GL038465
.
Hudson
,
J. G.
,
S.
Noble
,
V.
Jha
, and
S.
Mishra
,
2009
:
Correlations of small cumuli droplet and drizzle drop concentrations with cloud condensation nuclei concentrations
.
J. Geophys. Res.
,
114
,
D05201
,
doi:10.1029/2008JD010581
.
Hudson
,
J. G.
,
S.
Noble
, and
V.
Jha
,
2010a
:
Stratus cloud supersaturations
.
Geophys. Res. Lett.
,
37
,
L21813
,
doi:10.1029/2010GL045197
.
Hudson
,
J. G.
,
S.
Noble
, and
V.
Jha
,
2010b
:
Comparisons of CCN with supercooled clouds
.
J. Atmos. Sci.
,
67
,
3006
3018
.
Hudson
,
J. G.
,
S.
Noble
, and
V.
Jha
,
2012
:
Cloud droplet spectral width relationship to CCN spectra and vertical velocity
.
J. Geophys. Res.
,
117
,
D11211
,
doi:10.1029/2012JD017546
.
Jiang
,
H.
,
G.
Feingold
, and
W. R.
Cotton
,
2002
: Simulations of aerosol–cloud–dynamical feedbacks resulting from entrainment of aerosol into the marine boundary layer during the Atlantic Stratocumulus Transition Experiment. J. Geophys. Res.,107, 4813,
doi:10.1029/2001JD001502
.
Jiang
,
H.
,
G.
Feingold
, and
A.
Sorooshian
,
2010
:
Effect of aerosol on the susceptibility and efficiency of precipitation in warm trade cumulus clouds
.
J. Atmos. Sci.
,
67
,
3525
3540
.
Khelif
,
D.
,
S. P.
Burns
, and
C. A.
Friehe
,
1999
:
Improved wind measurements on research aircraft
.
J. Atmos. Oceanic Technol.
,
16
,
860
875
.
Khelif
,
D.
,
C. A.
Friehe
,
H.
Jonsson
,
Q.
Wang
, and
J.
Kalogiros
,
2005
:
Wintertime boundary layer structure and air-sea interaction over the Japan/East Sea
.
Deep-Sea Res. II
,
52
,
1525
1546
.
Kogan
,
Z. N.
,
Y. L.
Kogan
, and
D. K.
Lilly
,
1996
:
Evaluation of sulfate aerosols indirect effect in marine stratocumulus clouds using observation-derived cloud climatology
.
Geophys. Res. Lett.
,
23
,
1937
1940
.
Leaitch
,
W. R.
,
G. A.
Isaac
,
J. W.
Strapp
,
C. M.
Banic
, and
H. A.
Wiebe
,
1992
:
The relationship between cloud droplet number concentrations and anthropogenic pollution: Observations and climatic implications
.
J. Geophys. Res.
,
97
(
D2
),
2463
2474
.
Leaitch
,
W. R.
,
C. M.
Banic
,
G. A.
Isaac
,
M. D.
Couture
,
P. S. K.
Liu
,
I.
Gultepe
, and
S.-M.
Li
,
1996
:
Physical and chemical observations in marine stratus during the 1993 North Atlantic Regional Experiment: Factors controlling cloud droplet number concentrations
.
J. Geophys. Res.
,
101
(
D22
),
29 123
29 135
.
Lee
,
S.-S.
, and
G.
Feingold
,
2010
:
Precipitating cloud-system response to aerosol perturbations
.
Geophys. Res. Lett.
,
37
,
L23806
,
doi:10.1029/2010GL045596
.
Lenschow
,
D. H.
,
1986
: Aircraft measurements in the boundary layer. Probing the Atmospheric Boundary Layer, D. H. Lenschow, Ed., American Meteorological Society, 39–55.
Lenschow
,
D. H.
, and
P.
Spyers-Duran
, cited
1989
: Measurement techniques: Air motion sensing. NCAR RAF Bull. 23. [Available online at http://www.eol.ucar.edu/raf/Bulletins/bulletin23.html.]
Lu
,
M.-L.
,
W. C.
Conant
,
H. H.
Jonsson
,
V.
Varutbangkul
,
R. C.
Flagan
, and
J. H.
Seinfeld
,
2007
:
The Marine Stratus/Stratocumulus Experiment (MASE): Aerosol-cloud relationships in marine stratocumulus
.
J. Geophys. Res.
,
112
,
D10209
,
doi:10.1029/2006JD007985
.
Lu
,
M.-L.
,
G.
Feingold
,
H. H.
Jonsson
,
P. Y.
Chuang
,
H.
Gates
,
R. C.
Flagan
, and
J. H.
Seinfeld
,
2008
:
Aerosol–cloud relationship in continental shallow cumulus
.
J. Geophys. Res.
,
113
,
D15201
,
doi:10.1029/2007JD009254
.
Lu
,
M.-L.
,
A.
Sorooshian
,
H. H.
Jonsson
,
G.
Feingold
,
R. C.
Flagan
, and
J. H.
Seinfeld
,
2009
:
Marine stratocumulus aerosol-cloud relationships in the MASE-II experiment: Precipitation susceptibility in eastern Pacific marine stratocumulus
.
J. Geophys. Res.
,
114
,
D24203
,
doi:10.1029/2009JD012774
.
Lu
,
M.-L.
,
Y.
Liu
,
S.
Niu
, and
A. M.
Vogelmann
,
2012
:
Observed impacts of vertical velocity on cloud microphysics and implications for aerosol indirect effects
.
Geophys. Res. Lett.
,
39
,
L21808
,
doi:10.1029/2012GL053599
.
Martin
,
G. M.
,
D. W.
Johnson
, and
A.
Spice
,
1994
:
The measurement and parameterization of effective radius of droplets in warm stratocumulus clouds
.
J. Atmos. Sci.
,
51
,
1823
1842
.
Peng
,
Y.
,
U.
Lohmann
, and
W. R.
Leaitch
,
2005
: Importance of vertical velocity variations in the cloud droplet nucleation process of marine stratus clouds. J. Geophys. Res.,110, D21213, doi:10.1029/2004JD004922.
Platnick
,
S.
, and
S.
Twomey
,
1994
:
Determining the susceptibility of cloud albedo to changes in droplet concentration with the Advanced Very High Resolution Radiometer
.
J. Appl. Meteor.
,
33
,
334
347
.
Raga
,
G. B.
, and
P. R.
Jonas
,
1993
:
On the link between cloud-top radiative properties and sub-cloud aerosol concentrations
.
Quart. J. Roy. Meteor. Soc.
,
119
,
1419
1425
.
Roberts
,
G.
,
G.
Mauger
,
O.
Hadley
, and
V.
Ramanathan
,
2006
:
North American and Asian aerosols over the eastern Pacific Ocean and their role in regulating cloud condensation nuclei
.
J. Geophys. Res.
,
111
,
D13205
,
doi:10.1029/2005JD006661
.
Rosenfeld
,
D.
,
W. L.
Woodley
,
D.
Axisa
,
E.
Freud
,
J. G.
Hudson
, and
A.
Givati
,
2008
:
Aircraft measurements of the impacts of pollution aerosols on clouds and precipitation over the Sierra Nevada
.
J. Geophys. Res.
,
113
,
D15203
,
doi:10.1029/2007JD009544
.
Small
,
J. D.
,
P. Y.
Chuang
,
G.
Feingold
, and
H.
Jiang
,
2009
:
Can aerosol decrease cloud lifetime?
Geophys. Res. Lett.
,
36
,
L16806
,
doi:10.1029/2009GL038888
.
Snider
,
J. R.
, and
J.-L.
Brenguier
,
2000
:
Cloud condensation nuclei and cloud droplet measurements during ACE-2
.
Tellus
,
52B
,
823
842
.
Storer
,
R. L.
, and
S. C.
van den Heever
,
2013
:
Microphysical processes evident in aerosol forcing of tropical deep convective clouds
.
J. Atmos. Sci.
,
70
,
430
446
.
Twohy
,
C. H.
,
M. D.
Petters
,
J. R.
Snider
,
B.
Stevens
,
W.
Tahnk
,
M.
Wetzel
,
L.
Russell
, and
F.
Burnet
,
2005
:
Evaluation of the aerosol indirect effect in marine stratocumulus clouds: Droplet number, size, liquid water path, and radiative impact
.
J. Geophys. Res.
,
110
,
D08203
, doi:10.1029/2004JD005116.
Twomey
,
S.
,
1959
:
The nuclei of natural cloud formation. II: The supersaturation in natural clouds and the variation of cloud droplet concentration
.
Geophys. Pure. Appl.
,
43
,
243
249
.
Twomey
,
S.
,
1977a
:
The influence of pollution on the shortwave albedo of clouds
.
J. Atmos. Sci.
,
34
,
1149
1152
.
Twomey
,
S.
,
1977b
: Atmospheric Aerosols. Elsevier Scientific Publishing Company, 302 pp.
Wang
,
J.
,
P. H.
Daum
,
S. S.
Yum
,
Y.
Liu
,
G. I.
Senum
,
M.-L.
Lu
,
J. H.
Seinfeld
, and
H.
Jonsson
,
2009
:
Observations of marine stratocumulus microphysics and implications for processes controlling droplet spectra: Results from marine stratus/stratocumulus experiment
.
J. Geophys. Res.
,
114
,
D18210
,
doi:10.1029/2008JD011035
.
Warren
,
S. G.
,
C. J.
Hahn
,
J.
London
,
R. M.
Chervin
, and
R. L.
Jenne
,
1988
: Global distribution of total cloud cover and cloud type amounts over the ocean. NCAR Tech. Note NCAR/TN-317+STR, doi:10.5065/D6QC01D1.
Xue
,
H.
, and
G.
Feingold
,
2006
:
Large-eddy simulations of trade wind cumuli: Investigation of aerosol indirect effects
.
J. Atmos. Sci.
,
63
,
1605
1622
.
Xue
,
H.
,
G.
Feingold
, and
B.
Stevens
,
2008
:
Aerosol effects on clouds, precipitation and the organization of shallow cumulus convection
.
J. Atmos. Sci.
,
65
,
392
406
.
Yum
,
S. S.
, and
J. G.
Hudson
,
2002
:
Maritime/continental microphysical contrasts in stratus
.
Tellus
,
54B
,
61
73
.
Yum
,
S. S.
, and
J. G.
Hudson
,
2004
:
Wintertime/summertime contrasts of cloud condensation nuclei and cloud microphysics over the Southern Ocean
.
J. Geophys. Res.
,
109
,
D06204
,
doi:10.1029/2003JD003864
.
Yum
,
S. S.
,
J. G.
Hudson
, and
Y.
Xie
,
1998
:
Comparisons of cloud microphysics with cloud condensation nuclei spectra over the summertime Southern Ocean
.
J. Geophys. Res.
,
103
,
16 625
16 636
.
Zhao
,
M.
, and
P. H.
Austin
,
2005
:
Life cycle of numerically simulated shallow cumulus clouds. Part II: Mixing dynamics
.
J. Atmos. Sci.
,
62
,
1291
1310
.
Zuidema
,
P.
,
H.
Xue
, and
G.
Feingold
,
2008
:
Shortwave radiative impacts from aerosol effect in marine shallow cumuli
.
J. Atmos. Sci.
,
65
,
1979
1990
.