1. Introduction and background
During the last two decades, marine stratocumulus clouds have been the subject of many theoretical/modeling studies (e.g., Garreaud and Muñoz 2004; Bretherton and Wyant 1997) and field experiments (e.g., Albrecht et al. 1988, 1995b). This type of cloud is mainly observed at low levels over the eastern side of the subtropical oceans. The conditions there (cool surface waters and warm, dry air subsiding aloft) favor the creation of a sharp temperature inversion and a sharp decrease in moisture that caps the marine atmospheric boundary layer (MABL) and leads to extensive stratocumulus with tops at the inversion base. Both surface-based cloud climatologies (Klein and Hartmann 1993) and satellite studies (Ramanathan et al. 1989) have clearly indicated the impact of marine boundary layer clouds on the global radiation budget; their high albedo results in a substantial decrease of the amount of solar radiation reaching the ocean’s surface, while their low altitude results in a relatively small temperature difference between cloud top and the ocean’s surface, which does not significantly affect the amounts of thermal radiation emitted to space. Although the role of stratocumulus clouds in affecting the radiation balance by cooling the ocean was recognized through early studies (e.g., Randall et al. 1984; Philander et al. 1996), the growing need of a more accurate representation in the global climate models (GCMs) has engaged many scientists in the pursuit of gaining a better understanding of their radiative, microphysical, and dynamical properties, the thermodynamic structure of the MABL, and the climatological variability of the respective areas (e.g., Stevens et al. 2003).
One of the most prevalent stratocumulus cloud decks in the world is located over the subtropical southeast (SE) Pacific, extending about 1500 km offshore from the equator to the latitude of central Chile (25°–30°S; Klein and Hartmann 1993). In addition to the large latitudinal extent, the interaction with El Niño–Southern Oscillation (ENSO) and the special morphology of the western South American continent (e.g., the presence of Andes) also contribute to the unique character and great importance of the SE Pacific stratocumulus regime (Li and Philander 1996). Despite their importance and special feedbacks on the global climate, the SE Pacific stratocumulus areas were largely unexplored in the past; almost all associated research until the end of the previous century focused on the northeast Pacific (e.g., Albrecht et al. 1988; Stevens et al. 2003) and Atlantic (e.g., Albrecht et al. 1995b) stratocumulus, while the SE Pacific stratocumulus regime received far less attention, and its characteristics were considered a priori to be similar to the other stratocumulus regimes. This lack of in situ data lead scientists to the organization of the East Pacific Investigation of Climate (EPIC) processes in the coupled ocean–atmosphere system (1999–2004), and the subsequent EPIC 2001 field experiment (Weller 1999). The second leg of the EPIC field campaign was an extensive stratocumulus study (Bretherton et al. 2004), which took place in October 2001 and revealed the complex structure of the stratocumulus-topped boundary layer in the subtropical SE Pacific. An important component of EPIC long-term monitoring is the Stratus Ocean Reference Station (Stratus ORS), which was launched in October 2000 at the geographical location of 20°S, 85°W by the Woods Hole Oceanographic Institution (WHOI) Upper Ocean Processes (UOP) Group. The recovery and replacement of the Stratus ORS buoy was one of the primary objectives of the EPIC 2001 stratocumulus cruise (hereafter EPIC 2001). Thereafter (with the exception of 2002), the annual ship campaigns to maintain and replace the buoy are further exploited as a means to deploy remote sensors and other instrumentation, and to conduct observations that will eventually advance our knowledge and understanding of the various processes associated with the SE Pacific stratus deck. The Pan-American Climate Studies (PACS) Stratus 2003 (Kollias et al. 2004) and Stratus 2004 (Serpetzoglou et al. 2005) research cruises provided, in combination with EPIC 2001, a unique dataset by capturing most of the properties that are fundamental for studying and analyzing the complex features of stratocumulus clouds and MABL in the subtropical SE Pacific. These measurements also allow stratocumulus in this region to be compared with the better-studied stratocumulus of the northeast Pacific, and to those sampled in a less-instrumented Chilean cruise off of central Chile in October 1999 (Garreaud et al. 2001).
In this study, data collected during the EPIC/PACS Stratus research cruises form the basis for exploring clouds and boundary layer structures in this climate sensitive area. The research cruises and datasets are briefly described in section 2. In section 3 we describe the spatiotemporal variability of the various boundary layer and cloud properties observed on the three cruises and highlight the differences between the three observational periods. Mean profiles of the MABL thermodynamic structure for each cruise for the period for which the research vessels remained stationed at the ORS location are also constructed and compared in section 3. Section 4 discusses the physical properties of clouds and drizzle in an attempt to illuminate the various processes that modulate cloud life cycle and drizzle occurrence. A summary and discussion on the analysis results are provided in section 5. Composite sounding data and buoy-period statistical values are provided in the appendix.
2. Datasets and analysis procedures
a. Domain setup
The ship track for each of the three cruises under consideration is shown in Fig. 1. Table 1 provides the significant dates and times of each cruise. Although the cruise paths followed by the Research Vessels (R/Vs) Ronald H. Brown in 2001 and 2004 and Roger Revelle in 2003 differ, a sufficient overlap in their domains allows crucial comparisons to be made among the three field experiments. All three cruises feature a focus on the Stratus ORS location at 20°S, 85°W, where 5–6 days are spent recovering the old buoy and placing the new buoy, and with the ship holding station for intercomparisons. Thus, we focus on observations from this location to extract statistical characteristics of the basic cloud and drizzle properties. The transect along 20°S from 85° to 72°W is also common among the three cruises, and captures the evolution of the MABL in the transition from the deeper-ocean waters to the coastal regime.
b. Instrumentation and data
The EPIC 2001, PACS Stratus 2003, and PACS Stratus 2004 research cruises were collaborative efforts among various institutions and universities. An extensive suite of instruments was deployed onboard the research vessels for making measurements of boundary layer clouds, thermodynamic structure, surface fluxes, and near-surface meteorology (Fairall et al. 1997). The remote sensing systems on each cruise and their respective products are briefly described in Table 2. All three cruises included a laser ceilometer, a three-channel microwave radiometer, and an 8.6-mm Doppler cloud radar (MMCR)—although the latter suffered a component failure early in the PACS Stratus 2004 cruise. The EPIC 2001 and PACS Stratus 2004 cruises also included the operation of the C-band radar on board the Ronald H. Brown and a 915-MHz wind profiler, while a new, very high-resolution but low-sensitivity 3.2-mm Doppler cloud radar was only used during PACS Stratus 2004 [a frequency-modulated continuous-wave (FMCW) radar].
The investigation of the relationship between cloud morphology, fractional cloudiness, and drizzle occurrence requires detailed analysis of measurements from the ship-based active remote sensors. Because of their short wavelength, millimeter radars are capable of detecting very small droplets with diameters of 5–10 μm. Furthermore, millimeter radars have narrow beams that result in small sampling volumes. As a result, these radars provide excellent resolution in space and time (e.g., Clothiaux et al. 1995; Kollias and Albrecht 2000). Radar reflectivity is used in this study to extract drizzle occurrence during each cruise. The threshold value used to identify drizzle is a maximum radar reflectivity in the column greater than −10 dBZ (Frisch et al. 1995). The classification allows the calculation of the hourly drizzle fractional coverage, which is defined as percent of profiles that contain drizzle (according to the reflectivity threshold) within each hour. Radar reflectivity data also provided observations of cloud-top height. Drizzle occurrence and cloud-top height were calculated from MMCR data for EPIC 2001 and PACS Stratus 2003. For PACS Stratus 2004, however, the lack of MMCR data forced the use of wind profiler reflectivity data for the estimation of cloud-top height and FMCW reflectivity data for the extraction of hourly drizzle occurrence. The ceilometer provided estimates of cloud-base height with a temporal resolution of 15–30 s and a vertical resolution of 15 m. The cloud-base height retrievals were then used to calculate hourly zenith cloud fraction, which is defined as the number of cloud-base height samples observed over the total number of samples within each hour. Because of a problem with portions of the ceilometer data from PACS Stratus 2004, measurements of incoming longwave radiation were used as a surrogate of fractional cloudiness.
Surface meteorology and turbulent and radiative flux measurements provided a near-surface complement to the remote sensing instruments during each cruise. Continuous measurements of surface meteorological data and surface incoming solar and IR fluxes were obtained from the National Oceanic and Atmospheric Administration (NOAA)/Earth System Research Laboratory (ESRL)/Physical Sciences Division (PSD) air–sea flux system, while surface latent heat (LH), sensible heat (SH), and virtual heat (VH) fluxes were calculated from the surface meteorological data by applying a bulk flux parameterization (Fairall et al. 1997, 2003). The lifting condensation levels (LCLs) were also calculated from the surface meteorological data (air temperature and specific humidity time series data) using the equation described in Bolton (1980). Further, rawinsondes launched during the three field experiments provided high-resolution vertical profiles of the MABL thermodynamic structure. The frequency of the sounding launches in EPIC 2001 was relatively high (eight per day), compared with that in PACS Stratus 2003 (four per day) and PACS Stratus 2004 (four per day, with the exception of six per day while at the ORS location).
The Vaisala sounding systems (RS-80 sondes in EPIC 2001, RS-90 sondes in PACS Stratus 2003, and RS-92 sondes in PACS Stratus 2004) provided profiles of temperature (T), pressure (p), relative humidity (RH), and horizontal wind speed and direction. Based on the methods described by Bolton (1980), these data were then used to calculate profiles of potential temperature (θ), virtual potential temperature (θυ), equivalent and saturation equivalent potential temperature (θe and θes, respectively), and mixing ratio (r). To obtain mean thermodynamic profiles for each of the datasets we used two different approaches. The first included linear interpolation of the initial (raw) sounding data, obtained at variable height levels, to new vertical bins with a height increment of 10 m. This approach allows for an objective quantitative comparison of the vertical MABL profiles sampled during the three cruises, but limits the detailed analysis of the inversion characteristics. To maintain the structure of the inversion in the composite soundings, a nondimensional height scale was used, following Albrecht et al. (1995a); using this approach, the height (z) is normalized with the inversion-base height (zi) of each sounding to give a nondimensional vertical coordinate z/zi. The estimation of the inversion-base height for each sounding was performed objectively, using the μ parameter described in Yin and Albrecht (2000). Average soundings were then obtained by using vertical bins with a nondimensional height increment of 0.01. Using a subjectively selected threshold value of μ (0.25), the upper and lower boundaries of the capping inversion were retrieved (as the nearest heights that corresponded to a value of μ = 0.25 above and below its maximum value, respectively), along with the corresponding values of potential temperature and mixing ratio.
3. Boundary layer structures and cloudiness
During all three cruises, a wide range of cloud conditions were encountered that included extensive periods of complete cloud cover, broken cloud, and clear sky. The data reveal qualitative differences in the MABL structure and cloud conditions from year to year.
a. Moisture structure and cloud boundaries
The MABL mixing ratio structures from the rawinsondes launched during the three cruises are shown in Fig. 2. The cloud boundaries and the LCLs are also depicted. The three panels of Fig. 2 clearly demonstrate the differences between the boundary layer and cloud structures observed during the three cruises and provide a point of reference for the complexity and variability of the SE Pacific stratocumulus regime. A well-mixed stratocumulus-capped boundary layer was observed throughout the entire EPIC 2001 cruise (Bretherton et al. 2004). The fact that few broken-cloud and nearly no clear-sky periods were reported is confirmed by the very high value of the cruise-averaged ceilometer-derived zenith cloud fraction (92%). Conditions differed, however, during the PACS Stratus 2003 cruise (Kollias et al. 2004). The MABL structure was occasionally characterized by the strong capping inversion and the well-mixed vertical thermodynamic structure often observed in 2001, but there were also days, especially at the ORS location, with moderate vertical gradients of potential temperature and mixing ratio. This was reflected in the cloud coverage, with a reduced average cloud fraction (about 80%) compared to EPIC 2001, and the occasional presence of decoupled layers with shallow cumuli clouds. PACS Stratus 2004 included some of the features observed in 2003, but revealed further differences and interesting new features with respect to the previous field experiments. The boundary layer was relatively well mixed in the beginning of the cruise (westerly route toward the ORS location), with rather thin clouds and a good correspondence between LCL and cloud base. Conditions changed drastically, however, while the ship was stationed at the buoy location; the boundary layer deepened with time, and was characterized by very high and relatively thinner stratocumulus clouds and the formation of a second cloud base associated with small cumuli rising into the stratocumulus. The cruise-averaged zenith cloud fraction value was similar to that of PACS Stratus 2003 (about 78%).
A closer look at the lower panel of Fig. 2 reveals that the inversion height was about 1.2 km at the beginning of the 2004 buoy period (as in EPIC 2001 and somewhat lower than that of PACS Stratus 2003), but its gradual increase resulted in a maximum inversion height of about 1.7 km about 3 days later that persisted for the remaining 2 days. After the ship left the ORS station and headed southeast, the height of the inversion increased even more to 1.8–1.9 km, and then decreased to a minimum of about 500 m near the coast. During EPIC 2001 and PACS Stratus 2003, the maximum inversion heights observed did not exceed 1.5 km. A smaller-scale deepening of the boundary layer was observed at the beginning of these two cruises along the southerly tracks toward the mooring location and into the stratus deck. The upper panel of Fig. 2 further shows that the mean daily inversion height is slightly lowering with time during the EPIC 2001 buoy period, while the diurnal variability of the inversion height is prominent, as described in Bretherton et al. (2004). Some signs of a similar diurnal cycle in the inversion height can be seen in the moisture structure and cloud-top height recorded in 2003 and 2004 (middle and lower panels of Fig. 2), although the cycle is weaker and much more irregular. The cloud-base height, however, does show a strong diurnal variability during PACS Stratus 2004, in contrast to the EPIC 2001 observations that showed a pronounced diurnal cycle of inversion height/cloud top and almost no cloud-base diurnal variability. Bretherton et al. (2004) indicate, however, that a strong diurnal cycle of cloud-top height is at odds with the expectation based on previous observational (e.g., Minnis et al. 1992) and modeling (e.g., Bougeault 1985) studies, indicating the importance of cloud-base variations for diurnal variability in cloud thickness, in addition to inversion height variations.
Some of the gaps in the observed cloud-base heights of PACS Stratus 2004 (especially after 13 December) are due to the malfunctioning of the ceilometer during the daytime, and are not necessarily associated with the nonexistence of clouds. Fortunately, the ceilometer malfunction did not affect the representation of the cloud-base height increase during the boundary layer deepening observed after 11 December. The observed deepening is also highly correlated with the onset and gradual intensification of strong vertical gradients of the boundary layer moisture and significant divergence between LCL and cloud-base height, indicating that the subcloud layer remains “decoupled” for several days. During this time the stratus clouds are partially disconnected from the surface temperature and moisture fluxes (Bretherton and Wyant 1997; Wood and Bretherton 2004). This decoupling during the PACS Stratus 2004 cruise appears to begin the third day that the Ronald H. Brown is stationed at the buoy location and is actually enhanced during the southeasterly route that followed. The decoupling was associated with a small decrease in cloud thickness and the intermittent presence of shallow cumuli clouds below the higher stratocumulus cloud base, as indicated by the ceilometer cloud-base estimates (black dots near 600–800 m in the lower panel of Fig. 2). The daily ceilometer backscatter intensity and cloud-base height were compared with FMCW reflectivity data (graphs not shown here), revealing that some of the low-level cloud-base returns correspond to drizzle, while the rest are associated with low cumulus clouds.
Another spatial domain of interest is along the easterly route that the Ronald H. Brown and the Roger Revelle followed after leaving the WHOI buoy during EPIC 2001 and PACS Stratus 2003, respectively. This transect along the 20°S parallel from the ORS location (85°W) to Arica, Chile (∼70°W), is repeated in PACS Stratus 2004 (hereafter the 20°S transect), but in the opposite direction, because Arica was then the departure and not the ending point for the Ronald H. Brown. During the EPIC 2001 transect, the boundary layer becomes somewhat shallower than at the buoy with an average height of 1 km. During this period, the boundary layer was even closer to being well mixed than along the southerly transect and the buoy period, as indicated by the close correspondence between the LCL and the cloud-base height, but the clouds were thinner along this leg mostly because of the lower inversion heights. The respective transect in PACS Stratus 2003 was characterized by a constant inversion-base height of approximately 1.3 km. The mixing ratios during the PACS Stratus 2003 transect are much higher (∼8–10 g kg−1) than the respective EPIC 2001 period (∼6–8 g kg−1). The same applies for the PACS Stratus 2004 transect (beginning of PACS Stratus 2004); the boundary layer remains well mixed throughout, but with a very high moisture content (∼9–12 g kg−1). After a gradual decrease during the first 2 days of the cruise, the inversion base rises again to reach 1.2 km at the beginning of the buoy period.
b. Boundary layer structures (buoy periods)
Boundary layer structures observed at the WHOI buoy further illustrate the different mean states observed on the three cruises. The mean vertical profiles for the MABL thermodynamic and dynamical variables were constructed from the soundings launched during the WHOI buoy periods, following the analysis techniques described in section 2. The geometric height technique is used for the profiles shown in Fig. 3. Profiles with height scales normalized by the inversion-base heights are not shown here, but numerical values for these nondimensional soundings are provided in the appendix (see Table A1).
The normalized vertical profiles indicate that the inversion top lies between 1.1 and 1.2 z/zi for all three buoy periods, and these nondimensional heights correspond to about 1450 m for EPIC 2001, 1400 m for PACS Stratus 2003, and 1700 m for PACS Stratus 2004, as seen in Fig. 3. These mean inversion-top heights are slightly overestimated with respect to the values objectively calculated with the use of the μ parameter and presented in the appendix (see Table A2), and this can be attributed to the unconditional averaging technique that was applied to provide the profiles of Fig. 3. Further, the highest inversion thickness (i.e., difference between inversion-top and inversion-base heights) is observed in EPIC 2001 and the lowest is in PACS Stratus 2003, as indicated by the nondimensional soundings and the mean values (see Table A2). The inversion strength (as indicated by Δθ and Δr) is similar for the EPIC 2001 and PACS Stratus 2004 buoy periods, while PACS Stratus 2003 maintained a weaker inversion.
The potential temperature structure is similar for the three composite soundings and shows the typical characteristics of a stratocumulus-capped marine boundary layer: nearly well mixed in the subcloud layer, moist adiabatic in the cloud layer, and with a strong capping inversion with an exponential θ profile above the inversion. The higher inversions observed on the PACS Stratus 2004 cruise are well indicated by the composite sounding. The differences in the mean temperatures of the boundary layer are consistent with the changes in sea surface temperature (SST) on the cruises. The mean SSTs are 18.6°, 19.3°, and 19.5°C for the EPIC 2001, PACS Stratus 2003 and PACS Stratus 2004 buoy periods, respectively (see the appendix, Table A2), and these differences are reflected in the boundary layer temperatures; the PACS Stratus 2003 buoy-period boundary layer is about 1°–2°C warmer than that of EPIC 2001, and the boundary layer during the PACS Stratus 2004 buoy period is even warmer by about 0.5°–1°C (Fig. 3, θ profile).
The moisture structures shown in the composite sounding (Fig. 3, r profile) differ substantially among the three cruises, and this is further reflected in the θe profiles. The EPIC 2001 composite sounding is fairly well mixed, showing only a weak decrease in mixing ratio from the surface to the inversion-base height. The PACS Stratus 2003 buoy-period sounding is moister than that of EPIC 2001, especially in the lower boundary layer. From the surface to about 500 m, mixing ratio decreases slightly with height similarly to the EPIC 2001 sounding, but above 500 m it decreases at a greater rate, which is consistent with the partially decoupled conditions observed intermittently during the PACS Stratus 2003 buoy period. The same structure is observed in the PACS Stratus 2004 composite sounding as well, although this sounding is even moister in the lower layers and more decoupled than PACS Stratus 2003 (Fig. 3, r profile). Both PACS Stratus 2003 and PACS Stratus 2004 soundings are characteristic of the existence of a second cloud base (shallow cumuli clouds) below the stratocumulus. The base of the cumuli clouds is marked by the transition layer in the two composite soundings. The height of this layer is 0.35 z/zi, and this value corresponds to a geometric height of about 450–500 m. This is consistent with the respective heights measured for the northeast Pacific stratocumulus regime during the First International Satellite Cloud Climatology Project (ISCCP) Regional Experiment (FIRE; in 1987), and the Atlantic stratocumulus-to-cumulus transition regime that was the focus of the Atlantic Stratocumulus Transition Experiment (ASTEX; in 1992; see Albrecht et al. 1995a). The middle and lower panels of Fig. 2 indicate that the LCL, calculated by surface values of temperature and mixing ratio, closely matches the height of the transition layer as indicated by the moisture gradient. In addition, a surface layer of about 50-m depth is indicated in all composite soundings. The θe profiles indicate nearly well-mixed conditions in the top 25%–30% of the boundary layer, which is consistent with the decoupled boundary layer structures observed by Nicholls (1984).
The variations in mixing ratio profiles are also consistent with the changes in SST, because the relative humidity near the surface remains relatively constant (Fig. 3, RH profile). Thus, the mixing ratio at the lowest levels increases with the boundary layer temperature. The relative humidity profiles below 500 m (height of the transition layer for the PACS Stratus 2003 and PACS Stratus 2004 buoy periods) are similar for the three boundary layers, with relative humidity increasing with height from a minimum of about 73% near the surface. Above 500 m, however, relative humidity for the PACS Stratus 2003 buoy period is substantially lower than that of the EPIC 2001 sounding, and this difference is about 10% (88% and 98%, respectively) at the height of the stratocumulus cloud layer (1000–1100 m). Although the PACS Stratus 2004 composite sounding shows more enhanced decoupling than PACS Stratus 2003, the mean relative humidity of the stratocumulus cloud layer is higher (∼93%), indicating a fairly solid cloud layer despite the persistent decoupling and the higher cloud bases.
c. Above-inversion moisture
One factor important to the boundary layer variability is the moisture content above the boundary layer. Time–height sections of relative humidity above the boundary layer indicate synoptic-scale variability in the moisture field for all three cruises, as shown in Fig. 4. Layers of dry and moist air can be seen descending with time during all three cruises. These features, which seem to be more pronounced in EPIC 2001 and PACS Stratus 2004, may be attributed to the persistent subsidence over the region of the SE Pacific stratocumulus (Bretherton et al. 2004; Zuidema et al. 2006). This area lies beneath the descending branch of the local Hadley cell, which may result in the descent of layers with high moisture content that may originate from the deep convection in surrounding areas. The intertropical convergence zone (ITCZ) is one possible source. Another potential source is from deep convection over the Amazon and the surrounding areas of South America, rising high above the Andes, and being transferred over the SE Pacific area through westward-propagating upper- or midtropospheric Rossby waves (Bretherton et al. 2004; Garreaud and Muñoz 2004). A sign of this circulation pattern is the high moisture content of the upper-level air masses located close to the South American coast as indicated in Fig. 4 between 4 and 10 km during the 20°S transect in all three cruises. The midtropospheric humidity boundary in the PACS Stratus 2004 cruise is quite striking, and its origin should be further illuminated in the future. The moist air above the MABL during the initial days of the EPIC 2001 cruise could be either a manifestation of the deep convection over the equatorial areas (the EPIC 2001 cruise was initiated at the Galapagos Islands and the first sounding shown in the upper panels of Fig. 4 was released at approximately 2°S, 95°W) or a sign of the shallow meridional circulation (Zhang et al. 2004).
d. Near-surface conditions
Temperature and moisture conditions at and near the surface were obtained continuously for the three cruises. The cruise-track SSTs shown in Fig. 5 are consistent with the temporal and spatial climatology of the area. Large sea–air temperature differences of 1°–3°C were observed throughout the EPIC 2001 buoy period, and were primarily modulated by fluctuations of surface air temperature (Tair), because SST varied slightly between 18.5° and 19°C. Relatively low values of Tair (∼15°C) were recorded from 1400 UTC 18 October to 1400 UTC 19 October and again from 0800 to 2200 UTC 21 October. Both events are associated with a moistening and cooling of the lower 500 m of the boundary layer, a significant decrease in the LCL, and a partial decoupling from the stratocumulus cloud base (see upper panel of Fig. 2). Comstock et al. (2005) noted that the air–sea temperature difference was typically enhanced by about 1°C during drizzling periods of EPIC 2001. Indeed, the prescribed events correlate with extensive drizzle periods (see Fig. 12 and section 4b for further discussion).
Although the sea–air measurements collected during the PACS Stratus 2003 and PACS Stratus 2004 cruise experiments were similar to those during EPIC 2001, some differences were noted. The initial SST and Tair values during PACS Stratus 2003 were close to 24°C, thus much higher than the respective values at the departure of the Ronald H. Brown from the Galapagos Islands in 2001. Although the latitude was the same, the absence of a pronounced manifestation of the cold tongue in 2003 seems to be associated with longitudinal differences (Pyatt et al. 2005). Both SST and Tair decreased rapidly as the Roger Revelle steamed away from the warm equatorial waters to enter the cool stratus region of the subtropical SE Pacific, with their difference remaining at quite low levels (0°–1°C) compared with the sea–air temperature difference recorded during the respective EPIC 2001 route. The SSTs during the PACS Stratus 2003 WHOI buoy period varied between 19° and 20°C, and showed enhanced diurnal variability compared with the respective periods in EPIC 2001 and PACS Stratus 2004. This should be attributed to the broken-cloud or clear-sky periods observed at the buoy location in PACS Stratus 2003, especially just after the solar flux maximum (Kollias et al. 2004). Events like those during the EPIC 2001 buoy period associated with low values of Tair that result in a large sea–air temperature difference were observed during the PACS Stratus 2003 buoy period as well (16, 19, and 20 November), and correlated with higher values of temperature and relative humidity in the lower boundary layer. The sea–air temperature difference was, on average, much smaller (0°–1°C) than that during EPIC 2001.
The PACS Stratus 2004 buoy period was characterized by nearly constant SSTs (∼19.5°C), with surface air temperatures very close to this value and at times slightly larger. One event of a sudden rise of Tair at the end of 13 December resulted in the minimum sea–air temperature difference observed on all cruises (−1.5°C). An expected decrease in SST and Tair, with their difference rising gradually, marked the southeastward and eastward routes of PACS Stratus 2004, while the path along the Chilean coast that concluded the cruise was characterized by even lower SSTs but higher surface air temperatures, which is indicative of the coastal upwelling and the land effects influencing the ocean and boundary layer temperatures.
Along the 20°S transect, common with all three cruises, SST and Tair increased in an eastward direction toward the coast. This is evident both on the ending part of PACS Stratus 2003 (22–24 November 2003) and the initial part of PACS Stratus 2004 (6–8 December 2004), while the concluding days of EPIC 2001 (22–25 October 2001) are characterized by an extremely pronounced SST and Tair variability.
e. Winds
The vertical structure of the wind field from the rawinsondes launched during the three field experiments is shown in Fig. 6. In all three cruises, the winds are generally consistent with climatology, with quite strong southeasterlies prevailing in the lower 1–2 km of the troposphere. Some signs of large-scale variability are further evident at these levels (i.e., within the boundary layer), while substantial mesoscale and/or synoptic-scale variability is observed in the winds above the boundary layer.
Another feature of the cruise-track wind profiles is the weak northerly flow in the layer between 1 and 2 km on 11 October for the EPIC 2001 cruise from 2° to 8°S along 95°W, and 13 November for the PACS Stratus 2003 cruise at about 10°–11°S, 85°W. Similar events, although at a more northern location along 95°W, are documented in de Szoeke and Bretherton (2005), and are suggestive of the shallow meridional circulation proposed by Zhang et al. (2004): southerly trades in the MABL and a low-level return flow from the ITCZ atop the MABL. This circulation may be responsible for the increase in moisture content just above the boundary layer during this time (see upper and middle panels of Fig. 4, respectively). Further, in the PACS Stratus 2004 field experiment there is a pronounced southerly/southwesterly flow in the MABL during the eastward route toward and the southward route along the Chilean coast (20–23 December). This flow reflects the low-level jet off the west coast of subtropical South America (Rutllant 1993; Garreaud and Muñoz 2005).
Figure 6 also indicates a distinct difference among the three cruises regarding the strength of the southeasterlies that prevailed in the lower troposphere. The sharp pressure gradients, forming as a result of the enhanced anticyclonic circulation over the SE Pacific during PACS Stratus 2004 (graphs not shown here), seem to be primarily responsible for the strong trade winds observed in and above the MABL throughout the cruise. These stronger winds during the PACS Stratus 2004 buoy period are also shown in the composite boundary layer soundings (Fig. 3). The EPIC 2001 and PACS Stratus 2003 buoy periods are characterized by wind speed profiles with winds of about 7–8 m s−1 in the boundary layer and decreasing above the inversion. In contrast, during the PACS Stratus 2004 buoy period, the mean winds were consistently 3–4 m s−1 stronger for the entire lower-tropospheric profile. Although these conditions are considered favorable for enhanced turbulent mixing within the boundary layer that would normally result in well-mixed temperature and moisture soundings, the extensive period of decoupled conditions does not show such influence. Another contrasting feature is that the strong winds characterizing PACS Stratus 2004 are accompanied by smaller sea–air temperature differences than the other two cruises.
f. Wind and SST fields
To put the point measurements from the ship in perspective with the larger-scale SST and wind fields, we use SST data from the Tropical Rainfall Measuring Mission (TRMM) Microwave Imager (TMI) and wind retrievals from the Quick Scatterometer (QuikSCAT). Both satellite datasets include gridded data in very high resolution (0.25° or approximately 27.5 km). The prescribed fields, averaged over the buoy period for each of the three cruises, are shown in Fig. 7. They show how the variations in the position of the SE Pacific anticyclone appear to affect the wind and SST variability observed at the buoy for the three different observing periods. The SSTs to the southeast of the ORS location are lower during the EPIC 2001 buoy period than the respective periods in PACS Stratus 2003 and PACS Stratus 2004. The lower temperatures during this time are consistent with the climatology; EPIC 2001 took place in midspring for the Southern Hemisphere compared with late spring for the PACS Stratus 2003 cruise and early summer for PACS Stratus 2004.
These fields can also be used to estimate the magnitude of temperature advection at the buoy; a factor that can affect the sea–air temperature differences. Temperature advection near the surface was estimated using the mean buoy-period zonal and meridional wind components at 1000 mb (see Table A2) with the zonal and meridional temperature gradients estimated from the gridded satellite SST fields. The SSTs at the four grid points surrounding the WHOI buoy location were averaged to provide the SST at 20°S, 85°W. Because the calculations are sensitive to the choice of the location used to estimate the SST gradient upstream, we select a location and average the four surrounding grid points about 700 km upstream (24.5°S, 80.5°W), which is approximately the advection distance for 1 day, given a mean southeasterly wind of 8 m s−1. Using this methodology we obtained temperature advection values of −0.86°, 0.18°, and 0.08°C day−1 for the EPIC 2001, PACS Stratus 2003, and PACS Stratus 2004 buoy periods, respectively. Thus, the EPIC 2001 buoy period was characterized by significant cold advection compared with the other two field experiments. Moreover, the lack of cold-air advection during the PACS Stratus 2003 and PACS Stratus 2004 buoy periods is consistent with the low sea–air temperature difference (see Table A2). Although the wind speeds during the 2004 buoy period were substantially higher than those during EPIC 2001, there is little or no SST contrast upstream of the buoy.
g. Radiative and turbulent fluxes
Surface radiative and turbulent fluxes are coupled to the structure of stratocumulus-capped boundary layers. Both incoming shortwave and longwave radiation demonstrated a rather expected variability throughout each cruise (graph not shown here), because they were primarily modulated by fractional cloudiness. During periods with overcast skies, the IR flux ranged from 390 to 410 W m−2, while the maximum (noontime) solar flux varied between 600 and 800 W m−2. As expected, clear-sky periods are associated with reduced incoming longwave radiation (310–320 W m−2) and much higher noontime solar fluxes (1100–1200 W m−2). Values between these extremes correspond to broken-sky periods. This span of values for both longwave and shortwave radiation provides a rough estimate of the intensity of the radiative forcing associated with the SE Pacific stratocumulus cloud deck, and highlights the importance of an accurate representation of these clouds in the radiative transfer schemes of regional and global climate models.
The surface turbulent fluxes (Fig. 8) are primarily modulated by the sea–air temperature difference variations and the winds. The sensible heat flux is tied to the SST–Tair evolution; thus, it is characterized by relatively high values during the EPIC 2001 cruise and significantly lower values during PACS Stratus 2003 and PACS Stratus 2004. Actually, intermittent periods in the later cruises are associated with negative values of SH flux. Latent heat fluxes exhibit much higher values (measured in W m−2) than SH fluxes in general, although their contribution to the virtual heat flux (or buoyancy flux) is limited. Figure 8 clearly shows that the VH flux closely follows the SH flux evolution. There is a pronounced diurnal cycle observed in the LH fluxes during the first half of EPIC 2001 that is not as prevalent on the other cruises. The LH fluxes exhibit large variability (20–150 W m−2) during PACS Stratus 2003, especially during the WHOI buoy period. During the PACS Stratus 2004 cruise LH flux values range between 50 and 150 W m−2.
4. Physical properties of clouds, drizzle, and boundary layer
The preceding discussion and analysis of boundary layer properties provided a full and in-depth description of the stratocumulus boundary layer variability during the three-cruise sampling. Cloud morphology and life cycle and drizzle occurrence are highly coupled to this variability, and the systematic analysis and comparison of the physical properties of clouds and drizzle can further provide a wealth of information for additional modeling and field studies.
a. Cloud morphology
The macroscopic properties of the clouds observed during the three cruises are summarized in Fig. 9. These frequency distributions reflect the characteristics shown in the time–height sections of Fig. 2. The average cloud base is lower during EPIC 2001 than the other two cruises where more decoupled boundary layers are observed. Similar differences in the cloud-top heights are apparent in the middle panel of Fig. 9 with lower cloud tops observed during EPIC 2001 and higher cloud tops during the other two cruises. The striking feature here is the bimodal distribution of cloud-top height during PACS Stratus 2004, which reflects the boundary layer deepening that occurred while at the WHOI buoy (see section 3a). A similar but less intense bimodality is observed in the PACS Stratus 2004 cloud-base height distribution. Cloud thickness distribution is similar for EPIC 2001 and PACS Stratus 2003, despite the different cloud regimes encountered in the two cruises, while PACS Stratus 2004 shows a more confined distribution with less occurrence of cloud decks thinner than 200 m and higher occurrence of clouds with depths between 250 and 400 m than the preceding cruises.
b. Fractional cloudiness and drizzle occurrence
Figure 10 shows the time–height mapping of MMCR reflectivity during EPIC 2001 and PACS Stratus 2003. As noted previously, there are almost no available radar data from the NOAA/ESRL/PSD MMCR for the PACS Stratus 2004 cruise. Complementary observations from the 94-GHz FMCW radar, along with ceilometer observations, provide estimates of cloud fraction and drizzle occurrence during this cruise. The observations shown in Fig. 10 illustrate the variability of marine stratus occurrence over the length of the two cruises. There are periods of continuous cloud coverage, especially during the EPIC 2001 cruise, and clear-sky periods, especially during the PACS Stratus 2003 cruise. The presence of radar returns below the ceilometer cloud base is a general indicator of drizzling periods.
To further elaborate on the cloud and drizzle occurrence on all three cruises, time series of hourly estimates of cloud and drizzle fraction are shown in Figs. 11 and 12, respectively. The cruise-averaged cloud fraction is about 92% for EPIC 2001, 80% for PACS Stratus 2003, and 78% for PACS Stratus 2004, as mentioned earlier in section 3a. Comparing all three cruises, overcast conditions were observed during EPIC 2001, with a large fraction of drizzle occurrence during nighttime and early morning hours, when the cloud layer was thick and contained high values of liquid water. Although drizzle is thought to have a stabilizing effect on the MABL, a well-mixed state was maintained and the clouds persisted throughout the EPIC 2001 cruise. During PACS Stratus 2003, there were periods before and after the buoy stationing with fairly high cloud and drizzle fractions that resembled the persistent cloud decks observed during EPIC 2001, but also extensive clear-sky periods that were mainly documented by the ceilometer while the research vessel was stationed at the WHOI buoy. A reduction in cloud cover during or just after the characteristic EPIC 2001 (18–19 and 21 October 2001) and PACS Stratus 2003 (16, 19, and 20 November 2003) buoy-period events (addressed earlier in section 3d) can be further seen in Fig. 11. In addition, extensive drizzle occurs during or before these periods (Fig. 12), providing evidence of a mechanism of marine stratiform cloud dissipation that has been proposed by many studies in the past (e.g., Albrecht 1989); drizzle evaporates below cloud base, and the resulting evaporative cooling stabilizes the boundary layer and inhibits surface turbulent fluxes from reaching the cloud layer. As a result, cloud base rises and clouds get thinner or even dissipate, leading to broken-sky areas and reduced cloud cover. In PACS Stratus 2004, the cloud fraction oscillated from 100% at night to much lower values during daytime, particularly near the solar maximum period. During the same cruise the lowest drizzle fraction was observed.
The differences in fractional cloudiness among the three cruises may be related to the stability factor as defined by the Klein and Hartmann (1993) relationship. For the three cruise-composite soundings the average stability factor calculated as the difference between mean potential temperature at 700 and 1000 hPa was found to be about 24°C. The cloud fraction observed on the PACS Stratus 2003 and PACS Stratus 2004 cruises fit well with the Klein line representation (graph not shown here). However, the nearly constant value of the stability factor for the three cruises does not capture the higher fractional cloudiness observed on the EPIC 2001 cruise.
Various local factors may affect the drizzle variability observed within marine stratocumulus. Some that have been explored include cloud thickness, cloud microphysical properties, and above-inversion moisture. The effects of cloud thickness on drizzle occurrence are shown in Fig. 13. As shown in the figure, clouds need to be about 200 m thick for drizzle to occur and the occurrence increases sharply as the thickness increases to 500 m. This simple analysis does not attempt to separate the various effects that can cause drizzle and merely illustrates a thickness dependence that is consistent with that found in previous studies. During decoupled conditions, the cloud thickness may be a less reliable predictor of the drizzle occurrence, because some drizzle events may be coupled to the subcloud layer through convective elements with lower cloud bases than decoupled stratocumulus layers at the top of the boundary layer.
The possible effects of the above-inversion moisture on drizzle occurrence are examined by comparing the occurrence with the magnitude of the moisture jump at the inversion. The dataset for this comparison is smaller than the previous one, because it is limited to times when soundings are available. The results shown in Fig. 14 indicate little dependence of the frequency of drizzle occurrence on the moisture jump. Thus, for the area sampled, the controls on the drizzle frequency appear to be closely coupled to cloud thickness changes, although cloud microphysical observations were either not available or very limited to try to infer what impact these properties might have on drizzle occurrence.
c. Diurnal variability
Strong surface fluxes and cloud-top IR cooling are the primary mechanisms that maintain a well-mixed MABL and the marine stratus deck near the top of the boundary layer during nighttime. During daytime, the absorption of solar radiation near the cloud top partially offsets the IR cooling and thus reduces the turbulence kinetic energy that promotes vertical mixing and supplies the stratus deck with moisture. As a result, the cloud layer can partially thin or completely evaporate, leading to clear-sky periods (e.g., Miller and Albrecht 1995; Wood et al. 2002). The diurnal cycle signature is often disturbed by synoptic- and large-scale features, such as inertia–gravity waves (Bretherton et al. 2004) and fluctuations in the subsidence rate at the top of the MABL (Garreaud and Muñoz 2004).
Using the cloud and drizzle fraction hourly estimates reported in the previous section we construct the diurnal cycle of cloud fraction and drizzle occurrence for the three cruises (Fig. 15). In general, drizzle occurrence seems to vary diurnally in accordance with fractional cloudiness in all three cruises. The EPIC 2001 diurnal cycle of cloud fraction is relatively weak compared with the subsequent cruises, although the diurnal variation in drizzle occurrence during EPIC 2001 is pronounced and similar to that during PACS Stratus 2003. The highest values of cloud and drizzle fraction are observed during the night and early morning hours. In EPIC 2001, cloud fraction values remain remarkably high (above 90%) almost for the entire day, that is, from early evening [1700 local time (LT)] to late morning (1000 LT). Even at local noon cloud fraction does not drop below 80% on the EPIC 2001 cruise. Drizzle occurrence shows higher diurnal variability than cloud fraction during the same cruise, with a distinct maximum at 0500 LT (44%) and a minimum at local noon (4%). During PACS Stratus 2003, the cloud and drizzle fraction demonstrate higher diurnal variability. The maximum values of cloud and drizzle fraction are observed at 0600 LT (97% and 48%, respectively) and the minimum values are observed right after local noon (56% and 1%, respectively). PACS Stratus 2004 is also characterized by pronounced diurnal variability in cloud fraction, with higher values during nighttime and lower values during daytime, compared with PACS Stratus 2003. The maximum cloud fraction value was recorded at 0300 LT (93%), whereas the maximum in drizzle occurrence was recorded a few hours earlier (30% at local midnight). The lowest cloud fraction values occurred at 1200 and 1300 LT (55%), when the clouds had no drizzle in or below the cloud layer. These differences on the diurnal signature of cloud and drizzle occurrence between the three cruises may be attributed to the variations observed in the boundary layer structures and cloud regimes discussed in earlier sections.
d. Boundary layer decoupling
Although the EPIC 2001 cruise was characterized by boundary layers that were generally well mixed, the soundings and cloud properties during the PACS Stratus cruises were characterized by the frequent occurrence of decoupled boundary layers. To illuminate the possible mechanisms responsible for the decoupling, individual soundings launched during PACS Stratus 2004 were subjectively classified as either coupled or decoupled, based on the mixing ratio and relative humidity profiles. The soundings successfully launched during PACS Stratus 2004 totaled 79, from which 41 were characterized as coupled and 33 as decoupled (indicated by blue solid lines in Fig. 2), while 5 were disregarded. Composite soundings were then constructed for these two classifications. These soundings are shown in Fig. 16. The depth of the decoupled boundary layer is about 300 m higher than that of the coupled soundings. Further, the boundary layer wind speeds are about 2 m s−1 lower and the wind directions are 15° more to the south for the coupled soundings than for the decoupled soundings. The potential temperature and mixing ratio differences shown in Fig. 16 illustrate how the boundary layer differs for these two cases, with higher stability in the decoupled cases associated with higher moisture levels in the subcloud layer and lower in the cloud layer.
Table 3 summarizes the mean values of basic boundary layer and cloud properties for the two cases of coupled and decoupled soundings. Hourly estimates of each property were considered and averaged for those hours that matched the rawinsondes launch times. The differences in mixing ratio and potential temperature profiles described above are consistent with the changes in the moisture transports between these two cases, as seen in Table 3. Mean surface turbulent fluxes are significantly lower for the decoupled soundings (virtual heat flux is reduced by almost 2.5 W m−2), which is consistent with the lower differences between sea surface and surface air temperatures and specific humidities (about 0.2°C and 0.5 g kg−1, respectively), compared with the coupled case. Cloud and drizzle occurrence are lower by 16% and 8.5%, respectively, for the decoupled composite sounding, whereas cloud thickness only varies by 25 m. The high value of surface incoming solar flux associated with the decoupled case indicates that decoupling occurs mostly during daytime in contrast to well-mixed boundary layers that occur during the night. Nonetheless, the 140 W m−2 difference in solar radiation reaching the surface between the coupled and decoupled cases and the respective 8 W m−2 difference in surface incoming longwave radiation are noteworthy and should be further examined in terms of associating decoupling with variations on the ocean’s surface radiation budget.
5. Summary and discussion
Ship-based observations of boundary layer, cloud, and drizzle properties during the EPIC 2001 and the PACS Stratus 2003 and PACS Stratus 2004 cruises have been used to study the variability of these properties in the SE Pacific. The EPIC 2001 field experiment was the first attempt to study marine stratus clouds in this regime using ship-based instrumentation (Bretherton et al. 2004). During the PACS Stratus 2003 (Kollias et al. 2004) and PACS Stratus 2004 (Serpetzoglou et al. 2005) cruises new observational datasets of marine stratus clouds were collected. Here, the observations from these three cruises are used to document the structure and variability of the MABL, clouds, and drizzle, and provide a cohesive description of their differences and similarities. Some of the main features observed during the three cruises are summarized below.
During EPIC 2001, the MABL was well mixed, resulting in small LCL variability. The cloud fraction was very high, nearly no clear-sky periods were observed, and high nighttime drizzle occurrence and drizzle rates were recorded (Comstock et al. 2005). Stratus clouds with cloud thickness greater than 250 m had drizzle below the cloud base. Despite the stabilization of the boundary layer induced by the evaporation of drizzle, the MABL maintained a well-mixed vertical structure that helped maintain the cloud layer. A strong diurnal cycle of marine stratocumulus cloud-top height was documented throughout the cruise. Overall, the EPIC 2001 observations of marine stratus revealed an omnipresent stratus deck, with little or no transition to other MABL regimes, such as broken clouds and decoupled conditions. These conditions contrast with those observed on the subsequent cruises.
During the PACS Stratus 2003 cruise, moderate vertical gradients of potential temperature and mixing ratio that overlap with periods of low cloud fractional coverage, decoupled layers, and shallow cumuli clouds were observed. Furthermore, during PACS Stratus 2003 the LCL varied substantially with time in conjunction with MABL variability. Large periods of clear skies were observed at the WHOI buoy location, especially during midday. The stratus observed at the buoy during PACS Stratus 2003 differed from that during EPIC 2001, with sharp transitions from a solid cloud deck to broken cumuli, and moderate vertical gradients of thermodynamic properties in the MABL.
During the PACS Stratus 2004 cruise, the observed MABL, cloud, and drizzle structures showed similar features with those observed in PACS Stratus 2003. However, the presence of decoupled conditions in the MABL was more pronounced. The decoupled MABL conditions resulted in strong vertical temperature and moisture gradients and the intermittent presence of shallow cumuli clouds below the higher stratocumulus cloud bases. The inversion height during PACS Stratus 2004 was greater than the other cruises. Although the PACS Stratus 2004 composite sounding showed more enhanced decoupling than PACS Stratus 2003, the mean relative humidity in the stratocumulus cloud layer was higher, indicating a fairly solid cloud layer at the top of the boundary layer despite the persistent decoupling and the higher stratocumulus cloud bases.
The EPIC 2001 buoy period was characterized by large sea–air temperature differences (∼1.5°C), in contrast to PACS Stratus 2003 (∼0.5°C) and PACS Stratus 2004 (0°C). This was due to the lack of cold-air advection in the proximity of the WHOI buoy location during the later cruises. The wind direction during all three cruises was relatively persistent from the southeast (120°–130°) and the wind speed at the surface (1000 hPa) was 7–10 m s−1 on average for the three cruises. PACS Stratus 2004 was characterized by stronger winds that resulted from the enhanced anticyclonic circulation during the same period.
The typical depth of the MABL capping inversion was about 100 m, with an increase of potential temperature across the inversion of about 6–9 K, and a decrease in mixing ratio of about 4–5.5 g kg−1. Midtroposphere moisture features were observed by the soundings in all three cruises. These features were propagating downward, reaching the layer above the capping inversion. Typical cloud thickness was about 300–350 m for all three cruises and the thickest clouds were observed during EPIC 2001 and PACS Stratus 2004, despite the presence of decoupled boundary layers in the later cruise. Drizzle occurrence was substantially reduced during PACS Stratus 2004. Cloud fraction and drizzle occurrence typically exhibit strong diurnal cycles with maximum values during nighttime and minimum near the local noon time. The examination of the dependence of drizzle occurrence on cloud thickness and above-inversion moisture reveals that drizzle is closely coupled to cloud thickness changes and is far less dependent on the inversion moisture jump for the area sampled during the cruises. A study on the microphysical characteristics of cloud and drizzle would further illuminate the features and processes that influence drizzle formation and evaporation.
The classification of soundings launched on the PACS Stratus 2004 cruise for coupled and decoupled cases provided an in-depth analysis on the differences between coupled and decoupled boundary layer structures. Compared with the coupled case, decoupled boundary layers are shown to be associated with much higher boundary layer depths, higher moisture levels in the subcloud layer and lower levels in the cloud layer, lower surface turbulent fluxes, much higher surface incoming shortwave radiation and somewhat lower surface incoming IR flux, and reduced occurrence of clouds and drizzle. While all differences described above are consistent with the physical processes affecting boundary layer structures and with each other, two contrasting features are observed: cloud thickness does not vary significantly for the two cases, although the presence of thinner clouds would be expected for the decoupled case; and wind speeds are higher in the case of decoupled boundary layers, although stronger winds may induce stronger surface fluxes in the presence of nonnegligible sea–air temperature and moisture differences and may thus favor a stronger forcing (e.g., greater turbulence) toward coupled boundary layers.
The documentation of the temporal and spatial variability of the MABL and clouds in the SE Pacific is an important step in understanding the physical processes that contribute to the formation, maintenance, and dissipation of marine stratocumulus. We anticipate that the findings presented will help in the design and implementation of future field programs [e.g., Variability of the American Monsoon Systems (VAMOS) Ocean–Atmosphere–Land Study (VOCALS) 2008]. Further, the systematical comparison among the three cruises will provide a benchmark for the modeling community [e.g., large eddy simulations (LESs)], where modelers can test their parameterization schemes and representation of marine stratus clouds for a variety of MABL, surface, and large-scale forcing conditions.
Acknowledgments
This work was supported by NOAA Grant NA17RJ1226. As mentioned before, the EPIC and PACS Stratus cruises were collaborative efforts among many scientists, students, staff, and the crew of the research vessels Ronald H. Brown and Roger Revelle. The authors especially thank Dan Wolfe and Dr. Paquita Zuidema for their work and dedication on the ocean field. The cooperation of Dr. Robert Weller and the Woods Hole Oceanographic Institution made these projects possible.
REFERENCES
Albrecht, B. A., 1989: Aerosols, cloud microphysics and fractional cloudiness. Science, 245 , 1227–1230.
Albrecht, B. A., D. A. Randall, and S. Nicholls, 1988: Observations of marine stratocumulus during FIRE. Bull. Amer. Meteor. Soc., 69 , 618–626.
Albrecht, B. A., M. P. Jensen, and W. J. Syrett, 1995a: Marine boundary layer structure and fractional cloudiness. J. Geophys. Res., 100 , 14209–14222.
Albrecht, B. A., C. S. Bretherton, D. Johnson, W. H. Schubert, and A. S. Frisch, 1995b: The Atlantic stratocumulus transition experiment—ASTEX. Bull. Amer. Meteor. Soc., 76 , 889–904.
Bolton, D., 1980: The computation of equivalent potential temperature. Mon. Wea. Rev., 108 , 1046–1053.
Bougeault, P., 1985: The diurnal cycle of the marine stratocumulus layer: A higher-order model study. J. Atmos. Sci., 42 , 2826–2843.
Bretherton, C. S., and M. C. Wyant, 1997: Moisture transport, lower-tropospheric stability, and decoupling of cloud-topped marine boundary layers. J. Atmos. Sci., 54 , 148–167.
Bretherton, C. S., and Coauthors, 2004: The EPIC 2001 stratocumulus study. Bull. Amer. Meteor. Soc., 85 , 967–977.
Clothiaux, E. E., M. A. Miller, B. A. Albrecht, T. P. Ackerman, J. Verlinde, D. M. Babb, R. M. Peters, and W. J. Syrett, 1995: An evaluation of a 94-GHz radar for remote sensing of cloud properties. J. Atmos. Oceanic Technol., 12 , 201–229.
Comstock, K. K., C. S. Bretherton, and S. E. Yuter, 2005: Mesoscale variability and drizzle in southeast Pacific stratocumulus. J. Atmos. Sci., 62 , 3792–3807.
de Szoeke, S. P., and C. S. Bretherton, 2005: Variability in the southerly flow into the eastern Pacific ITCZ. J. Atmos. Sci., 62 , 4400–4411.
Fairall, C. W., A. B. White, J. B. Edson, and J. E. Hare, 1997: Integrated shipboard measurements of the marine boundary layer. J. Atmos. Oceanic Technol., 14 , 338–359.
Fairall, C. W., E. F. Bradley, J. E. Hare, A. A. Grachev, and J. B. Edson, 2003: Bulk parameterization of air–sea fluxes: Updates and verification for the COARE algorithm. J. Climate, 16 , 571–591.
Frisch, A. S., C. W. Fairall, and J. B. Snider, 1995: Measurement of stratus cloud and drizzle parameters in ASTEX with a Kα-band Doppler radar and microwave radiometer. J. Atmos. Sci., 52 , 2788–2799.
Garreaud, R. D., and R. Muñoz, 2004: The diurnal cycle in circulation and cloudiness over the subtropical southeast Pacific: A modeling study. J. Climate, 17 , 1699–1710.
Garreaud, R. D., and R. Muñoz, 2005: The low-level jet off the west coast of subtropical South America: Structure and variability. Mon. Wea. Rev., 133 , 2246–2261.
Garreaud, R. D., J. Rutllant, J. Quintana, J. Carrasco, and P. Minnis, 2001: CIMAR-5: A snapshot of the lower troposphere over the subtropical southeast Pacific. Bull. Amer. Meteor. Soc., 82 , 2193–2207.
Klein, S. A., and D. L. Hartmann, 1993: The seasonal cycle of low stratiform clouds. J. Climate, 6 , 1587–1606.
Kollias, P., and B. A. Albrecht, 2000: The turbulence structure in a continental stratocumulus cloud from millimeter-wavelength radar observations. J. Atmos. Sci., 57 , 2417–2434.
Kollias, P., C. W. Fairall, P. Zuidema, J. Tomlinson, and G. A. Wick, 2004: Observations of marine stratocumulus in SE Pacific during the PACS 2003 cruise. Geophys. Res. Lett., 31 .L22110, doi:10.1029/2004GL020751.
Li, T., and S. G. H. Philander, 1996: On the annual cycle of the eastern equatorial Pacific. J. Climate, 9 , 2986–2998.
Miller, M. A., and B. A. Albrecht, 1995: Surface-based observations of mesoscale cumulus–stratocumulus interaction during ASTEX. J. Atmos. Sci., 52 , 2809–2826.
Minnis, P., P. W. Heck, D. F. Young, C. W. Fairall, and J. B. Snider, 1992: Stratocumulus cloud properties derived from simultaneous satellite and island-based instrumentation during FIRE. J. Appl. Meteor., 31 , 317–339.
Nicholls, S., 1984: The dynamics of stratocumulus: Aircraft observations and comparisons with a mixed-layer model. Quart. J. Roy. Meteor. Soc., 110 , 783–820.
Philander, S. G. H., D. Gu, D. Halpern, G. Lambert, N. C. Lau, T. Li, and R. C. Pacanowski, 1996: Why the ITCZ is mostly north of the equator. J. Climate, 9 , 2958–2972.
Pyatt, H. E., B. A. Albrecht, C. Fairall, J. E. Hare, N. Bond, P. Minnis, and J. K. Ayers, 2005: Evolution of marine atmospheric boundary layer structure across the cold tongue–ITCZ complex. J. Climate, 18 , 737–753.
Ramanathan, V., R. D. Cess, E. F. Harrison, P. Minnis, B. R. Barkstrom, E. Ahmad, and D. Hartmann, 1989: Cloud-radiative forcing and climate: Results from the Earth Radiation Budget Experiment. Science, 243 , 57–63.
Randall, D. A., J. A. Coakley Jr., D. H. Lenschow, C. W. Fairall, and R. A. Kropfli, 1984: Outlook for research on subtropical marine stratiform clouds. Bull. Amer. Meteor. Soc., 65 , 1290–1301.
Rutllant, J., 1993: Coastal lows and associated southerly winds in north-central Chile. Preprints, Fourth Int. Conf. on Southern Hemisphere Meteorology, Hobart, Australia, Amer. Meteor. Soc., 268–269.
Serpetzoglou, E., C. W. Fairall, D. E. Wolfe, V. P. Ghate, I. Jo, B. A. Albrecht, and P. Kollias, 2005: Properties of the SE Pacific stratocumulus using mm-wave radars and other remote sensors. Extended Abstracts, 32nd Conf. on Radar Meteorology, Albuquerque, NM, Amer. Meteor. Soc., 2R.2. [Available online at http://ams.confex.com/ams/pdfpapers/96599.pdf.].
Stevens, B., and Coauthors, 2003: Dynamics and chemistry of marine stratocumulus—DYCOMS-II. Bull. Amer. Meteor. Soc., 84 , 579–593.
Weller, R., 1999: A science and implementation plan for EPIC: An eastern Pacific investigation of climate processes in the coupled ocean–atmosphere system. University of Washington, 105 pp. [Available online at http://www.atmos.washington.edu/gcg/EPIC/EPIC_rev.pdf.].
Wood, R., and C. S. Bretherton, 2004: Boundary layer depth, entrainment, and decoupling in the cloud-capped subtropical and tropical marine boundary layer. J. Climate, 17 , 3576–3588.
Wood, R., C. S. Bretherton, and D. L. Hartmann, 2002: Diurnal cycle of liquid water path over the subtropical and tropical oceans. Geophys. Res. Lett., 29 .2092, doi:10.1029/2002GL015371.
Yin, B., and B. A. Albrecht, 2000: Spatial variability of atmospheric boundary layer structure over the eastern equatorial Pacific. J. Climate, 13 , 1574–1592.
Zhang, C., M. McGauley, and N. A. Bond, 2004: Shallow meridional circulation in the tropical eastern Pacific. J. Climate, 17 , 133–139.
Zuidema, P., B. E. Mapes, J. L. Lin, and C. Fairall, 2006: The interaction of clouds and dry air in the eastern tropical Pacific. J. Climate, 19 , 4531–4544.
APPENDIX
Buoy-Period Composite Soundings and Statistics
Table A1 includes the nondimensional buoy-period composite soundings used in this study at intervals of 0.1 z/zi. Table A2 includes buoy-period mean and standard deviation values for the MABL, cloud, and drizzle properties explored in this study. The buoy-period soundings are used for the calculation of the means and standard deviations of all temperature, moisture, and wind parameters for the surface (1000 hPa), inversion, and above-inversion (700 hPa) levels. Hourly estimates/averages of the ceilometer and radar data, and 5-min averages of the air–sea flux system data were used before extracting the buoy-period means and standard deviations for the respective properties.
The routes that the research vessels followed during EPIC 2001 (blue, solid), PACS Stratus 2003 (red, dashed), and PACS Stratus 2004 (black, dashed–dotted). The respective triangles indicate the ship trajectory. The gray square denotes the location of the Stratus ORS buoy (20°S, 85°W). Topography (m) is also shown in colored contours.
Citation: Journal of Climate 21, 23; 10.1175/2008JCLI2186.1
Time–height section of mixing ratio r (g kg−1) from the soundings launched during (top) EPIC 2001, (middle) PACS Stratus 2003, and (bottom) PACS Stratus 2004. The cloud boundaries and the LCLs are also displayed. The cloud top (red) is retrieved from the MMCR for EPIC 2001 and PACS Stratus 2003, while for PACS Stratus 2004 it is approximated by the inversion-base height derived from the wind profiler reflectivity. The cloud base (black) is derived from the ceilometer and the LCL (blue) from surface meteorological data. All estimates are 10-min averaged or linearly interpolated from a higher resolution, with the exception of the hourly averaged cloud-top heights. The periods when the vessels were stationed at the WHOI buoy (20°S, 85°W) are bounded by black vertical lines, while white segments indicate missing or bad sounding values. The blue solid lines on the time axis of PACS Stratus 2004 panel indicate decoupled soundings.
Citation: Journal of Climate 21, 23; 10.1175/2008JCLI2186.1
Mean profiles derived from the soundings launched during the three WHOI buoy periods: EPIC 2001 (solid; 6 days: 16–22 Oct 2001), PACS Stratus 2003 (dashed; 5 days: 16–21 Nov 2003), and PACS Stratus 2004 (dashed–dotted; 5 days: 11–16 Dec 2004). Each variable is noted at the top of each subplot.
Citation: Journal of Climate 21, 23; 10.1175/2008JCLI2186.1
Time–height mapping of RH (%) from the soundings launched during (top) EPIC 2001, (middle) PACS Stratus 2003, and (bottom) PACS Stratus 2004. Dashed lines indicate the period when the ship was stationed at the WHOI buoy; white segments indicate missing or bad sounding values.
Citation: Journal of Climate 21, 23; 10.1175/2008JCLI2186.1
SST (black) and surface air temperature Tair (gray) during (top) EPIC 2001, (middle) PACS Stratus 2003, and (bottom) PACS Stratus 2004, as recorded from the NOAA/ESRL/PSD air–sea flux system (5-min temporal resolution). Dashed lines indicate the period when the ship was stationed at the WHOI buoy.
Citation: Journal of Climate 21, 23; 10.1175/2008JCLI2186.1
Time–height mapping of wind speed (m s−1, colors) and wind direction (° relative to the north clockwise, arrows) from the soundings launched during (top) EPIC 2001, (middle) PACS Stratus 2003, and (bottom) PACS Stratus 2004. Dashed lines indicate the period when the ship was stationed at the WHOI buoy; white segments indicate missing or bad sounding values.
Citation: Journal of Climate 21, 23; 10.1175/2008JCLI2186.1
SST (colors) from TMI data and surface winds (arrows) from QuickSCAT data for the (a) 6-day EPIC 2001, (b) 5-day PACS Stratus 2003, and (c) 5-day PACS Stratus 2004 buoy periods. The black square indicates the Stratus ORS location.
Citation: Journal of Climate 21, 23; 10.1175/2008JCLI2186.1
Surface sensible heat (black), latent heat (light gray), and virtual heat (dark gray) fluxes during (top) EPIC 2001, (middle) PACS Stratus 2003, and (bottom) PACS Stratus 2004. The temporal resolution is 5 min. Dashed lines indicate the period when the ship was stationed at the WHOI buoy.
Citation: Journal of Climate 21, 23; 10.1175/2008JCLI2186.1
Frequency distributions of cloud properties during EPIC 2001 (stars, solid line), PACS Stratus 2003 (squares, dashed line), and PACS Stratus 2004 (circles, dashed–dotted line). (top) Cloud-base height distribution in bins of 100 m. (middle) Cloud-top height distribution in bins of 100 m. (bottom) Cloud thickness distribution in bins of 50 m.
Citation: Journal of Climate 21, 23; 10.1175/2008JCLI2186.1
Reflectivity from the MMCR during (top) EPIC 2001 and (bottom) PACS Stratus 2003. The ceilometer cloud-base height is shown with the black stars. Dashed lines indicate the period when the ship was stationed at the WHOI buoy.
Citation: Journal of Climate 21, 23; 10.1175/2008JCLI2186.1
Hourly estimates of zenith-point fractional cloudiness from the ceilometer for (top) EPIC 2001, (middle) PACS Stratus 2003, and (bottom) PACS Stratus 2004. During PACS Stratus 2004, the daytime cloud fraction values were adjusted using the observed downward longwave radiation. Dashed lines indicate the period when the ship was stationed at the WHOI buoy.
Citation: Journal of Climate 21, 23; 10.1175/2008JCLI2186.1
Hourly fractional drizzle occurrence for (top) EPIC 2001, (middle) PACS Stratus 2003, and (bottom) PACS Stratus 2004. Drizzle is defined as MMCR (for EPIC 2001 and PACS Stratus 2003) or FMCW (for PACS Stratus 2004) radar profiles having maximum (column integrated) reflectivity greater than −10 dBZ. Dashed lines indicate the period when the ship was stationed at the WHOI buoy.
Citation: Journal of Climate 21, 23; 10.1175/2008JCLI2186.1
Drizzle occurrence as a function of cloud depth. The linear fit is drawn subjectively to avoid the bias on the mathematical calculation imposed by “extreme” domains (e.g., drizzle occurrence equal to zero). The overall correlation coefficient is R = 0.61.
Citation: Journal of Climate 21, 23; 10.1175/2008JCLI2186.1
Drizzle occurrence as function of mixing ratio jump across the inversion. The Δr values were initially estimated for each sounding (every 4 or 6 h, depending on the cruise), then linearly interpolated to the hourly estimates of drizzle occurrence.
Citation: Journal of Climate 21, 23; 10.1175/2008JCLI2186.1
Diurnal cycle of cloud fraction (black-colored lines) and drizzle occurrence (gray-colored lines) during EPIC 2001 (stars, solid line), PACS Stratus 2003 (squares, dashed line), and PACS Stratus 2004 (circles, dashed–dotted line). A −10-dBZ reflectivity threshold is used in the MMCR/FMCW data for the retrieval of the drizzle fraction. Ceilometer data complemented by longwave radiation data are used for extracting the PACS Stratus 2004 cloud fraction diurnal cycle.
Citation: Journal of Climate 21, 23; 10.1175/2008JCLI2186.1
Coupled (solid) and decoupled (dashed) composite soundings for PACS Stratus 2004. Each variable is noted at the top of each subplot.
Citation: Journal of Climate 21, 23; 10.1175/2008JCLI2186.1
Time schedule for the three stratus cruises.
A list of the remote sensing instruments on board the Ronald H. Brown and Roger Revelle, and their respective products.
Mean values of boundary layer and cloud properties for the coupled/decoupled sounding classification.
Table A1. Composite (buoy period) soundings used in this study at intervals of 0.1 z/zi.
Table A2. Buoy-period statistics.
