Abstract

A climatology of the structure of the low-altitude cloud field (tops below 4 km) over the Southern Ocean (40°–65°S) in the vicinity of Australia (100°–160°E) has been constructed with CloudSat products for liquid water and ice water clouds. Averaging over longitude and time, CloudSat produces a roughly uniform cloud field between heights of approximately 750 and 2250 m across the extent of the domain for both winter and summer. This cloud field makes a transition from consisting primarily of liquid water at the lower latitudes to ice water at the higher latitudes. This transition is primarily driven by the gradient in the temperature, which is commonly between 0° and −20°C, rather than by direct physical observation.

The uniform lower boundary is a consequence of the CloudSat cloud detection algorithm being unable to reliably separate radar returns because of the bright surface versus returns due to clouds, in the lowest four range bins above the surface. This is potentially very problematic over the Southern Ocean where the depth of the boundary layer has been observed to be as shallow as 500 m. Cloud fields inferred from upper-air soundings at Macquarie Island (54.62°S, 158.85°E) similarly suggest that the peak frequency lies between 260 and 500 m for both summer and winter. No immediate explanation is available for the uniformity of the cloud-top boundary. This lack of a strong seasonal cycle is, perhaps, remarkable given the large seasonal cycles in both the shortwave (SW) radiative forcing experienced and the cloud condensation nuclei (CCN) concentration over the Southern Ocean.

1. Introduction

The present generation of satellite imagery offers an unprecedented view into the radiative and microphysical properties of clouds over the Southern Ocean (e.g., Mace 2010; Bennartz 2007; Stephens et al. 2002). Simply using CloudSat to observe that the fractional cloud cover of the low-altitude clouds (tops less than 3 km) commonly exceeded 0.80, Mace et al. (2007) underscored the importance of these clouds to the energy and water fluxes across the surface of the Southern Ocean. Haynes et al. (2011) demonstrated the prevalence of these low clouds over the Southern Hemisphere middle latitudes, and the significant shortwave (SW) cloud radiative effect associated with them. These clouds and their corresponding thermodynamic fluxes are of importance to diverse climate processes such as the transport of water and energy to the Antarctic (Fitzpatrick and Warren 2007; Yin 2005), the uptake of carbon dioxide into the Southern Ocean (Caldeira and Duffy 2000), and the aerosol budget over the Southern Ocean (Korhonen et al. 2010).

Active satellite observations (i.e., CloudSat and CALIPSO) are able to build upon the observations of passive instrumentation to provide depth and fresh insight. O’Dell et al. (2008) constructed an 18-yr climatology of the cloud liquid water path (LWP) using a host of satellite-based passive microwave observations. This extended climatology is of particular interest over the Southern Ocean where it is observed that the LWP weakly peaks along the storm track and experiences a relatively weak annual cycle. Morrison et al. (2011) used Moderate Resolution Imaging Spectroradiometer (MODIS) observations to highlight the prevalence of supercooled liquid water and the absence of clear glaciation in the low-altitude clouds over the Southern Ocean. Greenwald (2009) constructed a 2-yr climatology of LWP for both MODIS and Advanced Microwave Scanning Radiometer for Earth Observing System (EOS) (AMSR-E) and highlighted the differences arising at high latitudes during the winter months when the satellite zenith angle remains large. Seethala and Horvath (2010) constructed a similar comparison of LWP for 1 yr and noted that the MODIS climatology overestimated LWP at high latitudes.

In spite of these new capabilities, Trenberth and Fasullo (2010) conclude that disproportionately large biases “exist in both reanalysis and global coupled models in the energy budget of the Southern Hemisphere (SH) that is directly linked to simulation of clouds.” Moving beyond the fractional cloud cover and LWP, an immediate question arises over the physical boundaries of these low-altitude clouds. Few in situ observations exist at higher latitudes, primarily because of their remote location and the difficulty in operating at low altitudes in an environment of high wind shear and heavy icing. Notable field experiments in this region like the Aerosol Characterization Experiment (ACE-1; Bates et al. 1998) and the Southern Ocean Cloud Experiments (SOCEX I and II; Boers et al. 1998) are now over 15 yr old. A more recent field study off the coast of Tasmania (Morrison et al. 2010) identified the difficulty in modeling supercooled liquid water due primarily to the inability of the reanalyses to capture the wind shear and temperature inversion across the cloud top.

The aim of this paper is to explore the physical structure of the low-altitude clouds over the Southern Ocean as observed by CloudSat. While it is common to employ an integration of a set of A-Train observations (e.g., Mace et al. 2009; Delanoë and Hogan 2010), it is important to appreciate the behavior of individual components particularly the CloudSat observations, which are the backbone of many of these integrated products. The simple climatology of low-altitude cloud boundaries derived by CloudSat is then evaluated against inferred observations from upper-air soundings from Macquarie Island and Hobart.

2. CloudSat profiles

An example of a single wintertime granule, or overpass, of the CloudSat radar reflectivity, the cloud radar-only (RO) liquid water content (CloudSat RO LWC) and ice water content (IWC; CloudSat RO IWC) is presented in Fig. 1, along with the thermodynamic phase from Cloud–Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) and the visible imagery from MODIS. This particular wintertime example has been selected as it did not intersect any midlatitude cyclones, which are commonly present in the Southern Ocean storm track, nor any high-altitude cirrus, which may complicate the observation of low-altitude clouds. The granule runs from 65° to 40°S at a longitude of roughly 122°–134°E placing it to the south of Australia.

Fig. 1.

An example of a CloudSat granule over the Southern Ocean for 7 Aug 2008 along with the CALIPSO water phase partition and MODIS visible image. (a) CloudSat radar reflectivity, (b) CloudSat RO liquid water content, (c) CloudSat RO ice water content, (d) CALIPSO water phase partition, and (e) MODIS 11-μm imagery are shown.

Fig. 1.

An example of a CloudSat granule over the Southern Ocean for 7 Aug 2008 along with the CALIPSO water phase partition and MODIS visible image. (a) CloudSat radar reflectivity, (b) CloudSat RO liquid water content, (c) CloudSat RO ice water content, (d) CALIPSO water phase partition, and (e) MODIS 11-μm imagery are shown.

Over an extent of more than 2500 km, the reflectivity predominantly emanates from within the lowest 2 km above the ocean surface. The strongest returns, not surprisingly, are generated in the lowest 500 m and highlight the limitation of CloudSat to identify clouds within the boundary layer. Specifically, the CloudSat cloud detection algorithm is unable to reliably separate radar returns because of the bright surface versus returns due to clouds in the lowest four range bins above the ocean surface (0–980 m). As such, this product is unable to detect clouds with confidence within the lowest kilometer over the Southern Ocean as illustrated in the images of the cloud LWC and IWC. For this specific example, the CloudSat LWC and IWC algorithms produce fields with tops primarily below 2 km and a base around 0.75 km. Two peaks in the concentration of LWC occur around 47° and 52°S. Across the domain, these fields exhibit intermittency on the horizontal scale of 5 km and do not suggest uniform stratocumulus decks exist on the scale of hundreds of kilometers.

The IWC is sparse in comparison to the LWC with a more uniform coverage only from 55° to 58°S. The CloudSat radar-only products assign hydrometers as ice or liquid as a function of the temperature as defined by the CloudSat European Centre for Medium-Range Weather Forecasts (ECMWF) auxiliary (AUX) data product, which is overlain on this example. The absence of liquid water at temperatures colder than −20°C at high latitudes is a consequence of this algorithm, as is the ratio between IWC and LWC at temperatures between 0° and −20°C. As the LWC and IWC are not direct, independent physical observations in this range, quantitative values should be interpreted with great caution. It is not the intent of this paper to produce a quantitative climatology of the IWC and LWC, rather the primary focus is on the physical boundaries. Mace (2010) details the error analysis in the retrievals of the ice water path (IWP) and LWP for an algorithm employed not only the CloudSat reflectivity and the ECMWF temperature but also employing the CALIPSO, AMSR-E, and MODIS observations. The error analysis for the IWP was detailed to be around 40%, while the LWP could only be constrained up “to a factor of 2 or more.” It was further noted that the integrated LWP product agreed rather poorly with the relatively simple RO product, but the IWP product showed a stronger agreement.

To contrast the cloud location generated from CloudSat, the same granule is also observed with CALIPSO (Hu et al. 2009) and the MODIS 11-μm channel (Platnick et al. 2003), which is used to define the cloud-top temperature. At this resolution, the MODIS imagery is roughly consistent with the CloudSat LWC profiles: the cloud cover is largely solid from 55° to 60°S and scattered at lower latitudes. The lidar image, by comparison, suggests a near-solid cloud deck with intermittency primarily limited to regions north of 44°S, around 47°S, and the region between 51° and 53°S. CALIPSO continues to detect clouds to the south of 58°S up to 61°S, although at cloud tops well below 1 km in altitude. It is also interesting to note that CALIPSO observes only liquid water, as opposed to any ice. The difference in the vertical and horizontal resolution between CloudSat and CALIPSO is also apparent. CloudSat is recorded at a relatively coarse 240-m vertical resolution and 1.4-km horizontal resolution; CALIPSO is recorded at a 30-m vertical resolution and 333-m horizontal resolution at this altitude.

This single example highlights potential limitations to CloudSat over the remote Southern Ocean, where CloudSat cannot make observations within the lowest kilometer and is limited to a relatively coarse resolution. Further for this example, most of the cloud field exists in the critical temperature range between 0° and −20°C making it difficult to directly define the cloud thermodynamic phase and thus fully interpret the reflectivity. It remains to be seen how frequent these low-altitude clouds exist over the Southern Ocean.

The CloudSat observations of ice water and liquid water can readily be compiled into a simple climatology. The climatology of the austral wintertime (June, July, and August) average cloud LWC and IWC is presented for the three years 2006–08 for all granules that reside over the domain of 100°–160°E and 40°–65°S (Fig. 2). In total 64 tracks are contained in this domain. Repeating every 16 days, there are on average 15–18 granules for each track in 3 yr. Thus for any given season there are over 1100 granules for the radar-only cloud mask (CM) (RO_CM) calculation [level 2 Geometric Profiling Product (2B-GEOPROF)], more than 900 granules for I/LWC [level 2 Radar-Only Cloud Water Content Product (2B-CWC-RO)], and more than 850 granules for radar–lidar cloud mask (RL–CM) (2B-GEOPROF-lidar) calculations. For an individual granule, there are approximately 2668 footprints across the selected domain, which have been reassigned to the nearest of 2500 bins with 0.01° width.

Fig. 2.

The austral wintertime (June, July, and August) climatology of the CloudSat liquid/ice water frequency for the three years, 2006–08, along with the liquid/ice water content. (top left) Liquid water frequency, (top right) ice water frequency, (middle left) average liquid water content when cloud is present, (middle right) average ice water content when cloud is present, (bottom left) average liquid water content over all observations, and (bottom right) average ice water content over all observations are shown.

Fig. 2.

The austral wintertime (June, July, and August) climatology of the CloudSat liquid/ice water frequency for the three years, 2006–08, along with the liquid/ice water content. (top left) Liquid water frequency, (top right) ice water frequency, (middle left) average liquid water content when cloud is present, (middle right) average ice water content when cloud is present, (bottom left) average liquid water content over all observations, and (bottom right) average ice water content over all observations are shown.

Looking first at the frequency of observing liquid water (LW; Fig. 2 left), a rather uniform layer is observed at heights between 750 and 2000 m for all latitudes. The magnitude of the frequency lies below 0.35 with the peak occurring at the higher latitudes of 55°–62°S. These values are for layers of 240-m thickness and, not surprisingly, are smaller than the integrated value of 0.80 commonly observed in Mace et al. (2007). The cloud frequency quickly drops to zero at heights below 750 m where the CloudSat RO algorithm cannot assign the radar return. Above 2000 m the frequency quickly drops below a magnitude of 0.05. It is interesting to note that this upper boundary is largely independent of the underlying latitude over the extent of the 25° of latitude.

The average LWC when cloud is present (cloud only) is next compiled. Here a strong gradient is observed across the domain with LWC increasing beyond 300 mg m−3 at the lower latitudes. It is again highlighted that the quantitative values should be viewed with great caution; the shape of the contours is the primary interest. This field is strongly correlated with the average atmospheric temperature computed from the ECMWF-AUX temperature product, highlighting again the importance of this product in defining the thermodynamic cloud phase as seen by CloudSat over the Southern Ocean.

The LWC can also be averaged over all observations (cloudy and clear) to create a broader climatological average that provides insight into the energy budget over the Southern Ocean. At first glance, one might readily interpret the average LWC (all) as that of a roughly uniform boundary layer cloud from 40° to 55°S with a maximum near 46°S. South of 55°S, the average liquid water concentration drops off primarily in response to colder temperatures and there is a weak slope in the contours. North of 50°S, there is little ice, and the average LWC (all) is largely independent of the latitude.

Such a uniform picture of the average liquid water content (all) of clouds over the Southern Ocean is remarkable especially when considering the gradients in the synoptic-scale forcing over the region. For example, from 55° to 40°S the sea surface temperature (SST) roughly increases from 5° to 15°C. In the subtropical belts in the world, it is common to find a correlation between the height of boundary layer clouds and the underlying SST (e.g., Stevens et al. 2007; Bretherton et al. 1995). This lack of variation with latitude also arises in spite of the presence of the Southern Ocean storm track (Simmonds and Keay 2000) and a large meridional gradient in the mean sea level pressure (MSLP) field across the domain. The MSLP reflects, in part, the synoptic-scale subsidence, which helps define the height and thickness of boundary layer clouds. In this light, the level contours in the liquid water content at heights from 1.5 to 2.5 km are a remarkable feature, too.

Turning next to the IWC, a similar climatology of frequency may be constructed (Fig. 2, right). Unlike the frequency of the LW, the frequency of the IW is strongly a function of the underlying latitude for both the magnitude and the height of the peak. Ice is rarely observed north of 45°S but builds quickly at the higher latitudes. At these higher latitudes, the 800-m cut off in the frequency of IWC is most striking; again, this is simply a reflection of the lack of reliability in CloudSat retrievals near the surface. Still focusing on the higher latitudes between 55° and 65°S, the frequency in IW is constantly above 0.3 at heights roughly between 1000 and 2000 m. This is somewhat in contrast to Morrison et al. (2011), where MODIS commonly observed cloud tops consisting of supercooled liquid water and uncertain in this region but rarely of clear glaciation. As the CloudSat observations are not limited to cloud top, it is not a direct comparison.

The average IWC (cloud only) complements that of the average LWC (cloud only). It is interesting to observe the peak in the average IWC (cloud only) steadily decreases from a height of 5 km at 40°S to a height 3 km at 65°S. Unlike for liquid water, the average IWC (all) is quite similar in shape to that for the average IWC (cloud only), although the magnitude of the concentrations are substantially lower at the lower latitudes where ice is rarely observed. No constant threshold or barrier is evident at elevations below 3 km, in contrast to the average LWC and the climatology of Mace et al. (2007).

If no distinction was made between the ice and liquid, the CloudSat climatology produces a uniform cloud layer across the whole of the Southern Ocean. The distinction between ice and liquid is a function of the CloudSat ECMWF-AUX temperature product. This dependence on the temperature may be explicitly calculated (Fig. 3) and highlights the dominance of clouds in the temperature range of 0° to −20°C. Including the uncertain class, the CloudSat observes clouds in this range nearly 60% of the time during the winter months (June, July and August, Fig. 3 top). The uncertain class comes from either the bright surface or the difficulty in partitioning ice and liquid in clouds. A climatology of the concentration of ice and water versus temperature (cloud only) may readily be calculated to illustrate the dependence CloudSat places on these products between freezing and −20°C. This climatology has also been produced for all pixels, cloudy and clear.

Fig. 3.

Climatology of the CloudSat liquid/ice water frequency vs temperature and average liquid/ice water content vs temperature for the three years, 2006–08. (top left) Wintertime liquid/ice water frequency, (top middle) wintertime average liquid/ice water content when cloud is present, (top right) wintertime average liquid/ice water content over all observations, (bottom left) summertime liquid/ice water frequency, (bottom middle) summertime average liquid/ice water content when cloud is present, and (bottom right) summertime average liquid/ice water content over all observations are shown.

Fig. 3.

Climatology of the CloudSat liquid/ice water frequency vs temperature and average liquid/ice water content vs temperature for the three years, 2006–08. (top left) Wintertime liquid/ice water frequency, (top middle) wintertime average liquid/ice water content when cloud is present, (top right) wintertime average liquid/ice water content over all observations, (bottom left) summertime liquid/ice water frequency, (bottom middle) summertime average liquid/ice water content when cloud is present, and (bottom right) summertime average liquid/ice water content over all observations are shown.

The seasonality of these clouds can be examined by repeating these calculations for the austral summertime (December, January, and February) (Fig. 4). Starting again with the frequency of observing any LW, one sees that the location of the cloud field remains relatively stable between seasons with the magnitude of the frequency marginally decreasing in summer. South of 60°S, however, liquid water is more frequently observed and at greater elevations. The summertime average LWC (cloud only) is similar to that of the wintertime only shifted slightly poleward. The summertime average LWC (all) is again quite similar to that of the wintertime, shifting slightly poleward. On average, CloudSat observes a uniform blanket of liquid water across the whole of the Southern Ocean over the summer months. The peak values in the average LWC (all) decreases from values above 110 mg m−3 during winter to 90 mg m−3 during summer, yet the upper boundary near 2000 m remains largely unchanged across the entire domain. This decrease in concentration over the summer is not immediately intuitive.

Fig. 4.

As in Fig. 2, but for the austral summer months (December, January, and February).

Fig. 4.

As in Fig. 2, but for the austral summer months (December, January, and February).

Overall this weak seasonal cycle is quite remarkable when compared to that of stratocumulus clouds in subtropical belts. One might expect to observe a much larger cycle simply because of the larger cycle in the incoming SW radiative forcing over the Southern Ocean. Further, Boers et al. (1998) observed a seasonal cycle in the effective radius arising from a dramatic increase in biogenic activity in the Southern Ocean that underpins the Charlson–Lovelock–Andreae–Warren (CLAW) hypothesis (Charlson et al. 1987; Ayers and Cainey 2007). Both ground-based (Gras 1995) and in situ (Yum and Hudson 2004) observations have revealed an order of magnitude increase in cloud condensation nuclei (CCN) concentrations over the summer months.

Turning to the ice water content, a strong seasonality is observed across the domain with ice much less prevalent over the summer and located at slightly higher altitudes. This is consistent with the climatology of Morrison et al. (2011). This decrease in average ice water content (all) is most distinct over the higher latitudes. Overall there appears to be a poleward shift in the height of the peak values during the summer, which would be expected with the migration of the Southern Ocean storm track (Simmonds and Keay 2000).

Finally, the climatology of the thermodynamic phase versus temperature may be calculated for the summer months, as well (Fig. 3, bottom). Here, 68% of the clouds are located in the temperature range of freezing to −20°C. This increase is, presumably, in direct response to the decrease in colder cirrus clouds over the summer. Overall there is an increase in the LWC of warm clouds between freezing and 15°C during the summer months.

3. Discussion

The CloudSat climatology of the average liquid water content (all) over the Southern Ocean displays what could roughly be described as a uniform “boundary layer” liquid phase cloud with a peak at just over 1 km and top near 2 km that ranges from 40° to 60°S in the winter and from 40° to 65°S in the summer. The climatology produces a crisp cloud base around 750 m, which is simply an artifact of the inability of CloudSat to make observations in the lowest kilometer with confidence (Sassen and Wang 2008). We find the average liquid water content climatology to be remarkable in three ways. First, the average LWC (all) is roughly constant with latitude and independent of large gradients to both the SST and the MSLP. Second, the vertical cap in the frequency of LWC is fixed around 2000 m across the domain (except at high latitudes during summer) and does not reflect the frequent presence of midlatitude cyclones along the Southern Ocean storm track. Finally, the average LWC (all) displays only a weak seasonal cycle with the peak concentration decreasing during the summer. The magnitude of the frequency of liquid water content undergoes a slightly greater seasonal cycle with greater values during the winter.

The CloudSat climatology of the average IWC (all), on the other hand, displays a strong meridional gradient with glaciated clouds increasing in concentration and decreasing in elevation when moving to higher latitudes. The climatology also portrays a strong seasonal cycle with the average IWC (all) decreasing over the summer, consistent with a warming atmosphere and greater incident solar radiation. The boundaries of the average ice water field shifts poleward during the summer, consistent with the migration of the Southern Ocean storm track. This portrait is consistent with the MODIS cloud phase climatology of Morrison et al. (2011), which highlights the decrease in upper-level cirrus clouds over the domain during the summer.

CloudSat frequently encounters clouds at temperatures between 0° and −20°C, at heights below 3 km. At these temperatures CloudSat assigns hydrometeors to both LWC and IWC as a function of the temperature. The average IWC (all) increases at the expense of the LWC (all) when moving from warmer temperature to colder. The Morrison et al. (2011) climatology, on the other hand, observed little evidence of clear glaciations in the tops of such clouds with MODIS, even during the winter months.

There are little means of directly confirming the CloudSat portrait of the structure of the liquid and ice water clouds over the Southern Ocean through independent observations. The Southern Ocean has simply proven to be too remote for routine in situ observations, and field observations in this region are scarce. Perhaps the most comprehensive field experiment was the ACE-1 (Bates et al. 1998) in 1995. Unfortunately, the National Center for Atmospheric Research (NCAR) C-130 aircraft flown during ACE-1 largely attempted to avoid these low elevation clouds because of an operational constraint to avoid the wetting of certain instruments. It is interesting to note that during ACE-1 some of these low-elevation clouds were observed not to reside in the boundary layer but rather in a “buffer layer” (Russell et al. 1998; Wang et al. 1999) that actually resided above a shallow boundary layer of 500–700-m depth. A distinct cloud layer has also been identified in a case study during the SOCEX experiment as an “intermediate layer” (Jensen et al. 2000) and again recently in the case studies of Morrison et al. (2010). Interestingly, large concentrations of supercooled liquid water have commonly been observed in the vicinity of Tasmania dating back to the 1960s (Mossop et al. 1970) with the discovery of the Hallett–Mossop ice-multiplication process. Many of these observations were made at heights above 2000 m. Morrison et al. (2010) encountered supercooled liquid water (SLW) at heights in excess of 3000 m.

One independent source of observations that can be employed to infer the structure of the cloud field is the routine upper-air soundings taken by Australian Bureau of Meteorology on Macquarie Island (54.62°S, 158.85°E). There were 23 896 soundings available for the period from 1973 to 2009. As the resolution of these data at low elevations is of importance, the soundings were further required to have least 15 data points up to 600 hPa, which reduced the number of soundings to 10 296. Introducing this resolution constraint meant that, due to improved instrumentation over time, the last decade was much more heavily represented in this subset. Prior to 1997, only 21% of the soundings met this resolution control constraint, compared to 77% for the last 13 yr. The soundings in the high-resolution subset are evenly distributed throughout the months of the year; it is not anticipated that this further constraint introduced any significant biases; especially as the CloudSat observations are limited to the three years, 2006–08.

Macquarie Island is ideally located within the domain; moreover, the site is largely free of any orographic complications. The launch site is at an elevation of 8 m above mean sea level and has direct exposure to the prevailing westerly winds. While these routine soundings do not directly observe the presence of clouds, let alone the phase or concentration of any such clouds, the relative humidity over water from such soundings may be employed as a proxy for the presence of liquid water cloud. If a sounding records a relative humidity (RH) above some threshold (say 90%) then the presence of liquid water is assumed. These proxy cloud layers may then be compiled from all of the soundings to create and a profile of frequency (Fig. 5) of LWC. Note that RH thresholds of 80% and 95% were similarly calculated, and while the magnitude of the frequency profiles changed with the threshold, the shape of the profile did not. The 90% threshold was chosen simply as it best matched the CloudSat frequency profile at higher elevations.

Fig. 5.

Vertical profiles of the frequencies of observing clouds over the Macquarie Island from both CloudSat (CALIPSO) and upper-air soundings. Ice water content (IWC) from CloudSat, liquid water content (LWC) from CloudSat, radar-only cloud mask [CM (RO)] from CloudSat, and the merged cloud mask [CM (R+L)] from CloudSat and CALIPSO. “Sounding” refers to the observed clouds from a sounding using a proxy of 90% RH. (left) Wintertime and (right) summertime climatologies are shown.

Fig. 5.

Vertical profiles of the frequencies of observing clouds over the Macquarie Island from both CloudSat (CALIPSO) and upper-air soundings. Ice water content (IWC) from CloudSat, liquid water content (LWC) from CloudSat, radar-only cloud mask [CM (RO)] from CloudSat, and the merged cloud mask [CM (R+L)] from CloudSat and CALIPSO. “Sounding” refers to the observed clouds from a sounding using a proxy of 90% RH. (left) Wintertime and (right) summertime climatologies are shown.

The frequency profiles for both the Macquarie proxy data (MAC) and the CloudSat data were calculated from a height of 20–4000 m in bins of 240 m (Fig. 5). These bins are labeled by the midpoint elevation. Focusing on the winter profiles first (Fig. 5 left), the MAC frequency profile peaks with a value of 0.60 in the second bin (380 m) and then steadily decreases with height up to 1580 m. Above this height the frequency decreases asymptotically toward 0.1 at 4000 m, although there is some noise above 2500 m. This profile is directly compared against the CloudSat LW frequency profile (threshold of 0) constructed for a narrow slice of latitude from 53° to 55°S. Above 1500 m these two frequency profiles behave similarly, although the MAC frequency profile is consistently of slightly greater magnitude. Below 1340 m, however, the CloudSat LW frequency profile increases marginally to a peak of 0.24 in the 1100-m bin and then drops off dramatically in the lowest bins where the CloudSat retrievals become more uncertain.

Profiles of the frequency of the CloudSat IW (IWC threshold of 0) and CM (RO_CM, threshold of 10, Marchand et al. 2007) have similarly been constructed (Fig. 5). The CloudSat IW frequency profile has a peak value of approximately 0.35 in the 1340 m bin, while the CloudSat CM frequency profile has a peak value of approximately 0.53 in the 1100-m bin. Again, the decline in the frequency of any of the CloudSat profiles over the lowest 1000 m is simply a reflection of the limited ability of CloudSat to make observations across this layer. The frequency profiles of the CloudSat IW and CM behave significantly different than that of the LW at elevations above 2000 m; instead of monotonically decaying, the frequency remains largely constant near 0.3, which is roughly consistent with Fig. 2.

A combined CloudSat and CALIPSO (radar + lidar) frequency profile of the CM is also considered in Fig. 5. This is the radar–lidar CM in which the radar threshold was again set at 10 while the lidar (2B-GEOPROF-lidar) threshold was set to require at least 50% of the bin to contain lidar-detected cloud (Mace et al. 2009). The CALIPSO lidar has the advantage of being able to make observations below 1000 m with greater confidence. The immediate drawback is that the lidar signal has often been attenuated at some height below 1000 m. The combined radar–lidar profile comes from a collection of 850 granules instead of 1100 granules for the CloudSat only profile, which potentially introduces a bias. Nevertheless, it is interesting to look at the radar–lidar frequency profile of the CM. The radar–lidar frequency profile of CM again has a peak of approximately 0.6 in the 1100-m bin; however, the combined radar–lidar frequency profile increases most in the lowest 1000 m, where it is roughly 0.2 greater than the CloudSat frequency of the CM.

The primary conclusion to be reached from Fig. 5 is that a substantial portion of the low-altitude clouds over the Southern Ocean could exist in the lowest 1000 m immediately above the ocean surface, and that CloudSat has a limited ability to view such clouds. Even a combined radar–lidar profile may have difficulty in picking out clouds in the lowest 500 m over the Southern Ocean. There are very few in situ observations that can be used to either support or counter this conclusion. Detailed observations of the boundary layer observed during ACE-1 (Russell et al. 1998) found its depth to be on the order of 500–700 m, which is more closely aligned with the peak in the MAC frequency profile in the 380-m bin.

The same profiles can be made for the summer (December, January, and February) to more clearly look for any seasonal cycle (Fig. 5 right). Focusing first on the MAC and the CloudSat LW frequency profiles, only a weak seasonal cycle is evident, especially for the MAC profile. The MAC frequency peak drops in magnitude from 0.6 (winter) to 0.56 (summer) while the CloudSat LW frequency peak correspondingly drops from 0.24 to 0.19. The heights of these peaks, however, remain unchanged between seasons. Again for summer, a conclusion can be reached that there is a potential for CloudSat to miss a substantial fraction of the cloud cover in the lowest 1000 m.

The CloudSat IW frequency profile displays the most striking seasonal cycle where the wintertime peak in the 1340-m bin no longer exists. Above 2000 m, the summer profile is similar to that of the winter profile. Next the CloudSat CM frequency profile can be examined. During the summer the magnitude in the frequency decreases from the surface to a height of about 1600 m with the peak frequency dropping from 0.53 to 0.51. Above 2000 m, however, the magnitude of the CloudSat CM frequency is greater during summer than during winter. The radar–lidar CM frequency profiles behave similarly to the CloudSat CM frequency profiles.

The second conclusion to reach from Fig. 5 is that there is indeed only a weak seasonal cycle in these low-altitude clouds according to the MAC frequency profiles and that this weak cycle is roughly consistent with the CloudSat LW frequency profile. Over the lowest 1600 m, both the MAC and the CloudSat LW frequency are marginally greater in the winter than during the summer. The CloudSat IW frequency profile displays the greatest seasonal cycle, presumably during the summer the atmospheric temperature increases in the lowest 2000 m, which requires CloudSat to classify less of these hydrometeors as ice. The two CM frequency profiles display a weak seasonal cycle similar to those for the MAC and the CloudSat LW frequency profiles.

The only other sounding site in the domain is that of Hobart, Tasmania (42.83°S, 147.50°E). Frequency profiles may readily be constructed for this site as before for both the proxy cloud data (HOB) and the CloudSat products (Fig. 6). Unfortunately, there is little that can be concluded from the HOB frequency profiles at heights below 2000 m, as the clouds and precipitation at this site are strongly affected by the local orography. Immediately upwind (to the west) of Hobart lies Mt. Wellington at an elevation of 1271 m, which creates a strong rain shadow effect for Hobart. From 2000 to 4000 m the HOB frequency profile looks qualitatively and quantitatively similar to the MAC frequency profile for both summer and winter.

Fig. 6.

As in Fig. 5, but for the profiles from Hobart, Tasmania.

Fig. 6.

As in Fig. 5, but for the profiles from Hobart, Tasmania.

Looking at the CloudSat LW frequency profile in the latitude band from 42° to 44°S, a stronger seasonal cycle is evident as not only has the peak frequency value decreased from 0.2 (winter) to 0.17 (summer), but also the elevation of this peak has changed from the 1580-m bin to the 1820-m bin. Once again CloudSat picks up very little cloud cover in the lowest 1000 m over the Southern Ocean. The radar–lidar CM frequency profile again has its peak in the 1100-m bin for both seasons with the peak frequency dropping from 0.50 (winter) to 0.45 (summer). Qualitatively, the radar–lidar CM frequency profile is similar between the 42°–44°S and the 53°–55°S latitude bands.

The CloudSat IW frequency profile demonstrates the greatest change when moving from the higher latitudes of Macquarie Island to the lower latitudes of Hobart. No local maximum is observed in the frequency during either summer or winter. During summer the frequency does not exceed 0.1 at any altitude; during winter the frequency approaches 0.2 at 4000 m.

4. Summary

A climatology of the structure of the low-altitude cloud field (tops below 4 km) over the Southern Ocean (40°–65°S) in the vicinity of Australia (100°–160°E) has been constructed with CloudSat products for liquid water (RO_liquid_water_content) and ice water (RO_ice_water_content) clouds. Averaging over the longitudes, the primary observation is that CloudSat produces a roughly uniform liquid phase cloud field with boundaries of 750 and 2250 m across much of the extent of the domain for both winter and summer. This cloud field makes a transition from primarily being composed of liquid water at the lower latitudes to ice water at the higher latitudes. This transition is driven simply by the gradient in the temperature field. Given the large uncertainties in the liquid water product, the focus is on the structure of these clouds rather than the quantitative value.

Upon further analysis, the climatology highlights the potential for severe limitations in the CloudSat observations over the Southern Ocean. First, CloudSat cannot make reliable observations in the lowest kilometer above the surface, yet the few direct observations in the region suggest that the boundary layer depth is often below 1-km altitude and cloudy. Second, for such low elevation clouds, the 240-m vertical resolution of CloudSat is relatively coarse and may bias the structure of these clouds. Third, the vast majority of these clouds reside in the temperature range of freezing to −20°C, precisely where CloudSat is unable to directly determine the thermodynamic phase. The algorithms that ultimately interpret the radar reflectivity remain to be explicitly evaluated over the unique, pristine environment of the Southern Ocean. For example, an assumption that the cloud droplet concentration is invariant in time and space is almost certainly invalid on the seasonal time scale over the Southern Ocean (Boers et al. 1998). Morrison et al. (2011) suggests that even at high latitudes (50°–60°S) during winter MODIS observes little evidence of clear glaciation in the low-altitude clouds.

With these caveats in mind, it is still interesting to examine the climatology of the liquid water field and ice water field, as seen by CloudSat. It is remarkable that the boundaries of the low-altitude liquid water cloud field are so stable given the extent of the domain and the large meridional gradients in the air temperature, the sea surface temperature, and the mean sea level pressure across the domain. The lower boundary is almost certainly defined by the limitation in CloudSat to reliably make observations across the lowest 1000 m immediately above the surface. With regard to the rigid upper boundary for these low-altitude liquid phase clouds, there is no evidence of any immediate bias in the CloudSat algorithm, although this may be consequence of the relatively coarse resolution (240 m) for such shallow clouds.

The climatology is also remarkable in that structure and amount undergoes only a very weak seasonal cycle. This is in contrast to the vast maritime stratocumulus fields in the subtropics (e.g., the eastern Pacific off of California), which effectively disappear over the course of a year. This weak seasonal cycle is largely consistent with the climatology of the LWP derived from passive satellite imagery (O’Dell et al. 2008).

Employing the routine upper-air soundings from Macquarie Island (54.62°S, 158.85°E), a proxy for liquid water cloud layers was constructed from observations that have a high relative humidity. Profiles of the frequency of observing proxy cloud were constructed for altitude bins matching CloudSat. The maximum frequency peaks at 380 m for this proxy cloud field for both summer and winter and is substantially lower than the altitude of the peak for the CloudSat cloud mask or a combined CloudSat–CALIPSO cloud mask. This further highlights potential limitations in employing CloudSat over the Southern Ocean. The frequency of the Macquarie Island proxy cloud field displays only a weak seasonal cycle with a slight maximum during the winter, which is consistent with the CloudSat climatology.

Ultimately, this research highlights the continued uncertainty in the structure of the cloud field over the Southern Ocean and, thus, the energy budget. Integrated products that employ CloudSat may potentially assume this uncertainty and are the focus of future work. It supports the call for a long-term ground-based radar–lidar observation program in this region coupled with in situ microphysics observations. Such observations would then be able to address the relatively large errors in the energy budget present in the global reanalyses and the Coupled Model Intercomparison Project (CMIP) simulations as discussed in Trenberth and Fasullo (2010).

Acknowledgments

We are happy to acknowledge the assistance of Larry Oolman from the University of Wyoming in procuring the upper-air soundings. We have also benefited from discussions with John Gras, Alain Protat, and Christian Jakob.

REFERENCES

REFERENCES
Ayers
,
G. P.
, and
J. M.
Cainey
,
2007
:
The CLAW hypothesis: A review of the major developments
.
Environ. Chem.
,
4
,
doi:10.1071/EN07080
.
Bates
,
T. S.
,
B. J.
Huebert
,
J. L.
Gras
,
F. B.
Griffiths
, and
P. A.
Durkee
,
1998
:
International Global Atmospheric Chemistry (IGAC) project’s first Aerosol Characterization Experiment (ACE1): Overview
.
J. Geophys. Res.
,
103
(
D13
),
16 297
16 318
.
Bennartz
,
R.
,
2007
:
Global assessment of marine boundary layer cloud droplet number concentration from satellite
.
Geophys. Res. Lett.
,
113
,
D20107
,
doi:10.1029/2006JD007547
.
Boers
,
R.
,
J. B.
Jensen
, and
P. B.
Krummel
,
1998
:
Microphysical and radiative structure of marine stratocumulus clouds over the Southern Ocean: Summer results and seasonal differences
.
Quart. J. Roy. Meteor. Soc.
,
124
,
151
168
.
Bretherton
,
C. S.
,
P.
Austin
, and
S. T.
Siems
,
1995
:
Cloudiness and marine boundary layer dynamics in the ASTEX Lagrangian experiments. Part II: Cloudiness, drizzle, surface fluxes, and entrainment
.
J. Atmos. Sci.
,
52
,
2724
2735
.
Caldeira
,
K.
, and
P. B.
Duffy
,
2000
:
The role of the Southern Ocean in uptake and storage of anthropogenic carbon dioxide
.
Science
,
287
,
620
622
.
Charlson
,
R. J.
,
J. E.
Lovelock
,
M. O.
Andreae
, and
S. G.
Warren
,
1987
:
Oceanic phytoplankton, atmospheric sulphur, cloud albedo and climate
.
Nature
,
326
(
6114
),
655
661
.
Delanoë
,
J.
, and
R. J.
Hogan
,
2010
:
Combined CloudSat–CALIPSO–MODIS retrievals of the properties of ice clouds
.
J. Geophys. Res.
,
115
,
D00H29
,
doi:10.1029/2009JD012346
.
Fitzpatrick
,
M. F.
, and
S. G.
Warren
,
2007
:
The relative importance of clouds and sea ice for the solar energy budget of the Southern Ocean
.
J. Climate
,
20
,
941
954
.
Gras
,
J. L.
,
1995
:
CN, CCN and particle size in Southern Ocean air at Cape Grim
.
J. Atmos. Res.
,
35
,
233
251
.
Greenwald
,
T. J.
,
2009
:
A 2-year comparison of AMSR-E and MODIS cloud liquid water path observations
.
Geophys. Res. Lett.
,
36
,
L20805
,
doi:10.1029/2009GL040394
.
Haynes
,
J. M.
,
C.
Jacob
,
W. B.
Rossow
,
G.
Tselioudis
, and
J.
Brown
,
2011
:
Major characteristics of Southern Ocean cloud regimes and their effects on the energy budget
.
J. Climate
,
24
,
5061
5080
.
Hu
,
Y.
, and
Coauthors
,
2009
:
CALIPSO/CALIOP cloud phase discrimination algorithm
.
J. Atmos. Oceanic Technol.
,
26
,
2293
2309
.
Jensen
,
J. B.
,
S.
Lee
,
P. B.
Krummel
,
J.
Katzfey
, and
D.
Gogassa
,
2000
:
Precipitation in marine cumulus and stratocumulus. Part I: Thermodynamic and dynamic observations of closed cell circulations and cumulus bands
.
Atmos. Res.
,
54
,
117
155
.
Korhonen
,
H.
,
K. S.
Karslaw
,
P. M.
Forster
,
S.
Mikkonen
,
N. D.
Gordon
, and
H.
Kokkola
,
2010
:
Aerosol climate feedback due to decadal increases in Southern Hemisphere wind speeds
.
Geophys. Res. Lett.
,
37
,
L02805
,
doi:10.1029/2009GL041320
.
Mace
,
G. G.
,
2010
:
Cloud properties and radiative forcing over the maritime storm tracks of the Southern Ocean and North Atlantic derived from A-Train
.
J. Geophys. Res.
,
115
,
D10201
,
doi:10.1029/2009JD012517
.
Mace
,
G. G.
,
R.
Marchand
,
Q.
Zhang
, and
G.
Stephens
,
2007
:
Global hydrometeor occurrence as observed by CloudSat: Initial observations from summer 2006
.
Geophys. Res. Lett.
,
34
,
L09808
,
doi:10.1029/2006GL029017
.
Mace
,
G. G.
,
Q.
Zhang
,
M.
Vaughan
,
R.
Marchand
,
G.
Stephens
,
C.
Trepte
, and
D.
Winker
,
2009
:
A description of hydrometeor layer occurrence statistics derived from the first year of merged CloudSat and CALIPSO data
.
J. Geophys. Res.
,
114
,
D00A26
,
doi:10.1029/2007JD009755
.
Marchand
,
R.
,
G. G.
Mace
,
T.
Ackerman
, and
G.
Stephens
,
2007
:
Hydrometeor detection using CloudSat—An earth-orbiting 94-GHz cloud radar
.
J. Atmos. Oceanic Technol.
,
25
,
519
533
.
Morrison
,
A. E.
,
S. T.
Siems
,
M. J.
Manton
, and
A.
Nazarov
,
2010
:
A modeling case study of mixed-phase clouds over the Southern Ocean and Tasmania
.
Mon. Wea. Rev.
,
138
,
839
862
.
Morrison
,
A. E.
,
S. T.
Siems
, and
M. J.
Manton
,
2011
:
A cloud-top phase climatology of Southern Ocean clouds
.
J. Climate
,
24
,
2405
2418
.
Mossop
,
S. C.
,
A.
Ono
, and
E. R.
Wishart
,
1970
:
Ice particles in maritime clouds near Tasmania
.
Quart. J. Roy. Meteor. Soc.
,
96
,
487
508
.
O’Dell
,
C. W.
,
F. J.
Wentz
, and
R.
Bennartz
,
2008
:
Cloud liquid water path from satellite-based passive microwave observations: A new climatology over the global oceans
.
J. Climate
,
21
,
1721
1739
.
Platnick
,
S.
,
M. D.
King
,
S. A.
Ackerman
,
W. P.
Menzel
,
B. A.
Baum
,
J. C.
Riedi
, and
R. A.
Frey
,
2003
:
The MODIS cloud products: Algorithms and examples from Terra
.
IEEE Trans. Geosci. Remote Sens.
,
41
,
459
473
.
Russell
,
L. M.
,
D. H.
Lenschow
,
K. K.
Laursen
,
P. B.
Krummel
,
S. T.
Siems
,
A. R.
Bandy
,
D. C.
Thornton
, and
T. S.
Bates
,
1998
:
Bidirectional mixing in an ACE 1 marine boundary layer overlain by a second turbulent layer
.
J. Geophys. Res.
,
103
(
D13
),
16 411
16 432
.
Sassen
,
K.
, and
Z.
Wang
,
2008
:
Classifying clouds around the globe with the CloudSat radar: 1-year of results
.
Geophys. Res. Lett.
,
35
,
L04805
,
doi:10.1029/2007GL032591
.
Seethala
,
C.
, and
A.
Horvath
,
2010
:
Global assessment of AMSR-E and MODIS cloud liquid water path retrievals in warm oceanic clouds
.
J. Geophys. Res.
,
115
,
D13202
,
doi:10.1029/2009JD012662
.
Simmonds
,
I.
, and
K.
Keay
,
2000
:
Mean Southern Hemisphere extratropical cyclone behavior in the 40-yr NCEP–NCAR reanalysis
.
J. Climate
,
13
,
873
885
.
Stephens
,
G. L.
, and
Coauthors
,
2002
:
The CloudSat Mission and the A-Train: A new dimension of space-based observations of clouds and precipitation
.
Bull. Amer. Meteor. Soc.
,
83
,
1771
1790
.
Stevens
,
B.
,
A.
Beljaars
,
S.
Bordoni
,
C.
Holloway
,
M.
Köhler
,
S.
Krueger
,
V.
Savic-Jovcic
, and
Y.
Zhang
,
2007
:
On the structure of the lower troposphere in the summertime stratocumulus regime of the northeast Pacific
.
Mon. Wea. Rev.
,
135
,
985
1005
.
Trenberth
,
K. E.
, and
J. T.
Fasullo
,
2010
:
Simulation of present-day and twenty-first-century energy budgets of the Southern Oceans
.
J. Climate
,
23
,
440
454
.
Wang
,
Q.
, and
Coauthors
,
1999
:
Characteristics of the marine boundary layers during two Lagrangian measurement periods. 2. Turbulence structure
.
J. Geophys. Res.
,
104
,
21 767
21 784
.
Yin
,
J. H.
,
2005
:
A consistent poleward shift of the storm tracks in simulations of 21st century climate
.
Geophys. Res. Lett.
,
32
,
L18701
,
doi:10.1029/2005GL023684
.
Yum
,
S. S.
, and
J. G.
Hudson
,
2004
:
Wintertime and summertime contrasts of cloud condensation nuclei and cloud microphysics over the Southern Ocean
.
J. Geophys. Res.
,
109
,
D06204
,
doi:10.1029/2003JD003864
.