1. Introduction
Low cloud cover over eastern subtropical oceans, such as develops off the Atacama Desert along the subtropical west coast of South America, features large and persistent stratocumulus decks. As elsewhere, this extensive cloud sheet develops mainly in connection with large-scale subsidence brought about by the descending branch of the Hadley–Walker circulation cells in connection with the southeast (SE) Pacific subtropical anticyclone. Large-scale subsidence results in a low-level temperature inversion separating a moist and cool boundary layer (BL), frequently capped by stratocumulus clouds, from warm and dry air aloft. Large-scale subsidence is regionally enhanced by atmosphere–land–ocean interactions including cold water upwelling and diurnal circulation flows forced by the solar heating of the regional alongshore orography: the coastal and the Andes mountain ranges. While the offshore SE Pacific marine boundary layer (MBL) has been long recognized as very important for global climate due to the radiative forcing of its persistent stratocumulus layer, its eastern coastal margin directly affects local weather and has large impacts on fisheries, wind power production, and aeronautical transportation.
Adequate representation of this MBL has proven a significant challenge for global and mesocale numerical models (Bony and Dufresne 2005; Leidner et al. 2001). Boundary layer height underestimation and fractional cloudiness low biases are frequently reported in the literature (Garreaud and Muñoz 2005; Zeng et al. 2004; Stevens et al. 2007; Ahlgrimm et al. 2009; Hannay et al. 2009). This difficulty in modeling arises in part due to the many interacting processes that affect this BL (Moeng 1998). Closer to the coast these modeling problems are aggravated as the real BL is shallower, its variability is larger, and additional processes like the interaction of the BL flow with the coastal topography enter the problem (Burk and Thompson 1996; Koracin and Dorman 2001; Strom and Tjernstrom 2004).
Advances in modeling this BL require observations as a basis to check and constrain the various physical parameterizations affecting model results. This is particularly important for the SE Pacific region, an area with historically sparse meteorological observations. Recognition of this fact prompted the execution of several research cruises in the tropical–subtropical SE Pacific during the last 10 years (e.g., Serpetzoglou et al. 2008) and, more recently, the carrying out in 2008 of the comprehensive Variability of the American Monsoon Systems (VAMOS) Ocean–Cloud–Atmosphere–Land Study Regional Experiment (VOCALS-Rex) field campaign aimed at gathering a large suite of meteorological and physical–chemical observations in the region (Wood et al. 2010). The principal focus of these efforts has been, however, on the offshore marine boundary layer, and meteorological data on the coastal band is still limited.
The general goal of the present work is to provide a basic observational reference of the main characteristics of the coastal boundary layer along northern Chile, based on a long record of surface and upper-air observations available at 23.4°S, 70.4°W. Besides surface meteorological data, these observations provide information on two of the main elements affecting coastal weather in this region: low-level cloudiness and the state of the subsidence temperature inversion. The objectives of the paper are, thus, to present the basic climatological features of these elements and to document the differences in the structure of this coastal boundary layer associated with the presence or absence of low-level clouds.
Several previous studies have documented differences in the vertical thermodynamic structure of the tropical and subtropical marine low-level troposphere depending upon cloudiness. Betts et al. (1995) presented 7-day averages of thermodynamic profiles at 28°N, 24°W measured in June 1992 during the Atlantic Stratocumulus Transition Experiment (ASTEX). Capped by an inversion near 850 hPa, the boundary layer shows a structure characterized by a conditionally stable cloud layer above 960 hPa, and a more well-mixed subcloud layer. When partitioned by cloud cover, the high-cloud-cover group (mean cloud fraction of 80%) shows a more unstable cloud layer and relative humidity increasing linearly with height from ∼75% near surface to ∼96% near inversion base. The low-cloud-cover group (mean cloud fraction of 10%), on the other hand, shows a more stable cloud layer, a more unstable subcloud layer, and relative humidity remaining between 75% and 85% in the whole BL.
Albrecht et al. (1995) presented mean thermodynamic soundings obtained from three different observational experiments—representative of four subtropical and tropical MBLs differing in terms of inversion base height, sea surface temperature, and cloud fraction. The profiles representing a subtropical stratocumulus-capped MBL were obtained from averaging 65 soundings performed in the summer of 1987 at San Nicolas Island (33.43°N, 119.57°W) 100 km southwest of Los Angeles, California. The inversion base in this case was ∼700 m, the SST was ∼16°C, and the mean cloud fraction was 83%. The BL showed a coupled structure, with a moist adiabatic temperature profile in the cloud layer and relative humidity increasing from ∼75% near surface to little less than 100% near inversion base. The other three climatological regimes represented in the analysis show larger SSTs, higher inversion bases, lower cloud fractions, a decoupled boundary layer structure, and a conditionally stable cloud layer. Cloud-layer relative humidity was shown to be useful in the diagnosis of fractional cloudiness.
Klein (1997) described the subtropical trade wind boundary layer based on 25 years of surface and upper-air observations at Ocean Weather Station November (OWS-N): 30°N, 140°W. A total of 2374 50-mb vertical resolution soundings for the months of June–September were used to produce averaged thermodynamic profiles composited upon three classes of nighttime low-cloud amount. Before averaging, the vertical scale of individual soundings was first normalized by the trade inversion base height. Results show that lower tropospheric stability increases with cloud fraction, associated with a cooler boundary layer and a warmer trade inversion when cloud fraction is high. All composited relative humidity profiles have values between 75% and 80% near the surface. Near the MBL top, however, the nearly overcast composite reaches ∼93% relative humidity, while the less cloudy composite shows a more vertically uniform profile, not surpassing 85%. Mixing increases with cloud fraction, although even the nearly overcast composite is not fully well mixed.
Norris (1998) presented cloud-conditioned upper-air and surface data averages over five ocean weather stations, including OWS-N analyzed by Klein (1997). In this case, however, composites were built based on cloud type rather than cloud fraction, and the radiosonde database used was much more limited (generally less than 70 profiles in each group) because only those with significant levels available were considered. In cases of low clouds under capping inversions (most prevalent in the eastern subtropical ocean OWS-N and summer midlatitude OWS-C and OWS-B) the composite vertical thermodynamic profiles show a systematic deepening and decoupling or stabilization of the MBL as cloud types vary from stratocumulus to cumulus-under-stratocumulus to cumulus.
More recently, Serpetzoglou et al. (2008) summarized observational studies of the MBL structure in the SE Pacific stratocumulus region based on three research vessel cruises, between 2001 and 2004, that conducted intensive observations at 20°S, 85°W for 5–6-day periods in each case. Their Fig. 16 shows mean thermodynamic profiles at this site conditioned subjectively upon the degree of coupling of the MBL. The coupled MBL results in shallower, more well mixed potential temperature and water vapor with somewhat weaker and more well-mixed winds, as compared to the uncoupled composite. The relative humidity profile in the latter is close to 80% in all the MBL, while in the coupled case relative humidity increases linearly up to ∼95% near inversion base.
As in the works reviewed above, the present study also documents differences in the MBL structure effects on low-level cloudiness in a subtropical environment. The site considered, however, is unique in the sense that it is located near the southern edge of a wide peninsula along the northern Chilean coast, at the eastern limit of the extensive and semipermanent stratocumulus region of the SE Pacific and at the western border of the extremely arid Atacama Desert (details of the site are described in section 2). On the one hand, its coastal character makes the site subject to more ample diurnal cycles as compared to open ocean locations, both in the boundary layer as well as in the free troposphere above (Muñoz 2008; Rutllant et al. 2003). Its peninsular character, on the other hand, makes the site very much influenced by the nearby MBL.
The peculiarities of the site compound to produce a complex BL with relatively large variability at seasonal, synoptic, and diurnal scales. For example, while the stratocumulus offshore are extremely persistent (e.g., Ghate et al. 2009), low-level cloudiness at the site varies much more in time and space (Berríos 2008; González 2009; Garreaud and Rutllant 2006). The potential problem with the variability and complex conditions at the site is compensated for by one factor: it has routine daily upper-air observations since 1957, together with an even longer record of hourly and synoptic weather observations. This large database makes it possible to investigate variability of the structure of the BL over this site from several perspectives and time scales. In the present work we focus on the differences in the BL mean structure associated with nighttime cloudiness, as the presence or absence of low-level clouds has first-order effects on radiative and turbulent processes in the MBL. Ceiling observations are used to distinguish between nearly overcast and nearly clear nights, and, in contrast to the studies referred to above, the impact of the inversion height is explicitly used as an additional discriminating variable. Furthermore, not only thermodynamic profiles are composited, but also wind profiles, making a more complete analysis of mechanisms determining the resulting mean structures possible.
The paper is organized as follows. Section 2 describes the site and the surface, upper-air, and regional datasets used, including the definition of parameters used to characterize the subsidence inversion and the presence of clouds in the BL. Section 3 provides a climatological description of subsidence inversion properties and low-level cloudiness at the site. Section 4 explores the covariability of low-level cloudiness and subsidence inversion properties and shows the cloudiness-conditioned climatological averages of surface meteorology and of the vertical structure of the lower troposphere. Section 5 presents additional local diagnostics aimed at better understanding mechanisms behind the cloud-conditioned climatology of this BL. Section 6 provides a regional view of the results, addressing the representativity of coastal cloudiness and the regional circulation anomalies associated to it. The last section summarizes the results and presents the conclusions of this work.
2. Site and data description
a. Site and regional climatology
Data analyzed herein correspond to routine meteorological measurements performed at Cerro Moreno Airport (23.4°S, 70.4°W) ∼30 km north of Antofagasta, one of the major cities in northern Chile. The airport is ∼7 km north of the southern edge of the Mejillones Peninsula (Fig. 1a), a conspicuous singularity in the otherwise regular and meridionally oriented northern Chile coastline. Mejillones Peninsula is located on the sea side of the Atacama Desert, almost permanently under the influence of the subtropical anticyclone of the SE Pacific, with persistent low-level southerly winds and a strong subsidence inversion. Upwind of the southern edge of the peninsula, the Coloso coastal upwelling center fosters cool air advection in spite of the fact that sea surface temperatures in the Antofagasta Bay are relatively warmer (Piñones et al. 2007). The meridionally oriented orography at the eastern and western flanks of Mejillones Peninsula encloses the gently sloping Mejillones pampa where the airport is located. In the northern and southern edges of the Mejillones Peninsula opposite sea breeze circulations develop during daytime, although regional southerly winds usually dominate over the whole Mejillones Peninsula late in the afternoon and evening (Flores et al. 2009).
The Andes Mountains, with elevations exceeding 4000 m, are located about 250 km east from Mejillones Peninsula, but their imprint is still noticeable in the regional circulation. Their barrier effect on the prevailing westerlies aloft is likely responsible for the northerly wind component above the subsidence inversion (Kalthoff et al. 2002). Strong diurnal cycles in temperature and wind exist above the marine boundary layer, associated with the diurnal cycle of heating and cooling of the western Andes slopes (Muñoz 2008). During daytime, a mean upslope flow divergence is induced in the lower troposphere over the desert producing a regionally enhanced subsidence along the coast in the afternoon, and concomitant low-cloud clearing (Fig. 1d). Conversely, during nighttime coastal convergence frequently brings the stratocumulus cloud cover onshore, and early morning satellite visible images commonly reveal a continuum between the coastal and offshore cloud fields (Figs. 1b,c; Rutllant et al. 2003).
Synoptic-scale forcing from midlatitudes modulates also the depth of the BL, producing variability in daily cycles of coastal cloud cover. This synoptic forcing induces the regular development of coastal troughs and lows that significantly modify the near-shore marine boundary layer characteristics, with alternating quasi-weekly ups and downs in the subsidence inversion base (Garreaud et al. 2002).
b. Surface meteorological observations
Meteorological surface data used in this study are based on conventional synoptic surface observations (SYNOP) and aviation routine weather reports (METAR) carried out by the Chilean Weather Service (DMC) at the Cerro Moreno Airport and compiled in the integrated surface dataset (ISD) by the National Climatic Data Center (NCDC; dataset ID DS3505, available at http://www7.ncdc.noaa.gov/CDO/cdo). The period considered in the analysis is 1979–2007, for which more than 80% of hourly observations of temperature, relative humidity, wind direction, and wind speed exist. Low-level cloudiness will be characterized based on the ceiling height (Zc) variable included in the ISD. As documented in National Climate Data Center (2008), Zc corresponds to “the height above ground level of the lowest cloud or obscuring phenomena layer aloft with ⅝ or more summation total sky cover.” Cloud observations at this site are visual estimates performed by professional meteorological observers. The vast majority of ceilings reported at this station is associated with stratus and stratocumulus cloud types, and have values below 1500 m AGL (see section 3). We shall use ceilings mainly as an indicator that at the corresponding hour low-level cloud fraction was large and not put too much emphasis on the actual value of Zc, except for checking that they in fact correspond to low-level clouds. Moreover, the conditional climatology, shown in section 4, is based upon a cloudy nighttime fraction (CNF) index defined as the fraction of the five SYNOP reports between 0000 and 1200 UTC on each day with Zc < 1500 m. High values of CNF will be considered to represent nights with persistent large cloud cover (hereafter referred to as cloudy nights), while low values of CNF will represent nights with relatively low cloud fraction (hereafter referred to as clear nights). Nighttime cloud observations, in principle, could be questioned because of the inherent difficulty in evaluating cloud properties at night. We argue, however, that at this site nocturnal low-level cloud fraction is generally not difficult to estimate by trained observers because of the extreme clearness of the skies above the MBL in this region.
c. Upper-air observations
Antofagasta station (WMO Station 85442, located also at Cerro Moreno airport) is one of five radiosonde stations operated in Chile by DMC and is a member of the Global Climate Observing System (GCOS) Upper-Air Network (World Meteorological Organization 2002). Radiosonde observations in Antofagasta started in 1957 in a 1200–0000 UTC mode, although 0000 UTC observations were discontinued in 1986. The period of analysis in the present work is 1979–2007 and only 1200 UTC soundings are considered. The radiosonde database was constructed based on three main sources: the integrated global radiosonde archive (IGRA; Durre et al. 2006), the University of Wyoming radiosonde archive available online (http://weather.uwyo.edu/), and a 1957–86 database acquired in 1987 by the Department of Geophysics of the University of Chile directly to NCDC. Temperature vertical profiles were quality-checked focusing on characteristics relevant for the appropriate determination of boundary layer height and subsidence inversion parameters. In particular, vertical resolution indices were defined in order to determine inversion base and top heights adequately, and physical plausibility indices were defined in order to detect and eliminate soundings with erroneous highly unstable or stable layers (Zamora 2010). Based upon these indices a revised database was constructed that has valid 1200 UTC soundings for 89% of the days in the 1979–2007 period.
Vertical heights of radiosonde variables are described herein using standard pressure altitudes (Glickman 2000, p. 721). Vertical profiles are truncated at 137 m, taken as the nominal elevation of Antofagasta station. Based upon the radiosonde temperature vertical profiles, all layers with thermal inversions up to 3000 m were identified. Surface-based inversions were separated from the analysis (see section 3). When more than one inversion layer existed aloft, they were merged if the intermediate layer was isothermal or shallower than 100 m. In cases when more than one inversion remained, the one with a larger temperature increase was selected. As a final filter, only inversions with temperature increments larger than 2°C and a relative humidity decrease larger than 20% were considered as representative of the subsidence inversion over the site.
For analyses requiring averaging of vertical profiles, they were first interpolated to a common vertical grid with 25-m resolution. Only wind profiles with more than eight significant and mandatory levels below 4000 m were considered in the analysis involving winds. A final remark about this upper-air dataset is to note some inhomogeneities apparent in the relative humidity data, probably due to changes in sensors over the years. Because of this problem, absolute values of humidity parameters must be taken with caution and are not heavily weighted in the analysis.
d. Regional fields
To put the climatological results for Antofagasta into a regional context (section 6), we used the National Center for Atmospheric Research–National Centers for Environmental Prediction (NCEP–NCAR) 40-Year Reanalysis (Kalnay et al. 1996) to describe wind and mass spatial structures, especially in the lower troposphere above the MBL. For near surface winds over the ocean, we have used the Quick Scatterometer (QuikSCAT) database (Gille et al. 2003) for the period 2000–07. Finally, for an exploratory assessment of the regional nocturnal cloud field, we have used 0845 UTC infrared images of the Geostationary Operational Environmental Satellite (GOES-12) for the period 2004–08. In particular, we have used the difference of channels 2 and 4 brightness temperatures to diagnose the presence of low-clouds as will be described in section 6.
3. Ceiling and inversion climatology
Figure 2 provides a basic characterization of ceiling observations at Antofagasta. Ceiling heights were grouped in four categories: Zc < 500 m, 500 m ≤ Zc < 1500 m, 1500 m (all AGL) ≤ Zc < unlimited, and unlimited. Figure 2a shows that during the night (0000 to 1200 UTC) the most frequent categories are those associated with low clouds, for values of Zc concentrated mainly between 300 and 900 m AGL (not shown). During daytime (1200 to 0000 UTC) the most frequent class corresponds to unlimited ceiling, presumably associated to the coastal “burning off” of low clouds. Figure 2b shows the seasonal variation of ceiling classes at 0900 UTC [0500 local time (LT)], considered as representative of end-of-the-night conditions. During austral summer unlimited ceiling at this hour is more frequent, while during winter and spring low clouds are most prevalent. Occurrence of very low ceilings (less than 500 m AGL) peaks more in austral winter from June to August. Finally, Fig. 2c describes the seasonal variation of the cloudy nighttime fraction (see section 2) index. Clear nights are more prevalent from January to March, while from July to October a large fraction of nights are persistently cloud covered.
Figure 3a shows histograms of inversion base heights determined from radiosonde temperature profiles, as described in the previous section. Solid bars correspond to the height distribution of the bases of all inversions found in the profiles. Surface-based inversions show a distinct peak in the distribution, separated from a wider peak around 900 m. After discarding surface-based inversions and the merging and selection procedure described in section 2, the resulting distribution of base heights is shown by the white bars in Fig. 3a. About 90% of subsidence inversion bases occur in the 500–1300-m range, with maximum frequency between 800 and 1100 m.
Figure 3b describes the seasonal variation in the occurrence frequency of surface and elevated inversions, as defined in section 2. The latter are extremely frequent at this site all year round. Surface inversions, on the other hand, show a clear annual cycle with maximum (40%) frequency in April and minimum frequency (10%) in October. This annual cycle appears related to the nighttime cloudiness cycle described earlier, with surface inversions less frequent in the most cloudy period of the year. The opposite is not completely true since surface inversions are found most frequently in April, although nighttime cloudiness is minimal in February. A plausible reason for this mismatch is that the soundings correspond to 1200 UTC (0800 LT), a time when in summer the sun is already above the horizon, and hence clear-night surface-based inversions may have become already mixed when the soundings were performed. In this respect, it is also worth noticing that the surface-based inversions are generally weak (median temperature jump of 1°C) and shallow (median depth of 34 m). A final remark about surface inversions is that radiosonde profiles show a steady increase in their annual prevalence from about 10% in 1979 up to 35% in 2007. Durre and Yin (2008) suggest that tendencies in shallow surface inversions could be explained by an increase in the number of significant levels available in the radiosonde databases, although it could as well be related to trends in cloudiness or sea surface temperature. Analysis of trends in this BL has been left out of the scope of the present paper (Zamora 2010).
Annual variation of properties of the subsidence inversion is shown in Fig. 4. Figure 4a shows the monthly distributions of inversion base heights, Zb. Its median values in winter are ∼900 m, increasing to ∼1100 m in summer. Variability does not change much throughout the year with monthly interquartile ranges of about 300 m. Temperature jumps across the inversion, on the other hand, show a strong annual cycle (Fig. 4b), with median values larger than 11°C from June to October and about 5°C from January to March. Interquartile ranges of ∼5°C do not change much along the year. Figure 4c shows monthly distributions of relative humidity jumps across the inversion. Humidity contrasts across the inversions are generally large over this site, especially during winter and spring when median values of 80% in relative humidity decrements are observed. In summer, together with the smaller values of the temperature jumps, the humidity contrasts are smaller as well.
The mean annual variation of the vertical thermodynamic profiles and inversion properties is illustrated further in Fig. 5a. Monthly median values of temperature and dewpoint at the inversion base and top are plotted together with the interquartile ranges of temperature and dewpoint measured at each level. The high recurrence of the subsidence inversion in this region leaves a clear mark in the statistical distribution of the temperature profiles. The annual cycle of the inversion top is stronger in height than in temperature, the latter closely following the free troposphere temperature profile. This suggests that a subsidence intensity annual cycle at this level may be important. The inversion base height, on the other hand, has an annual cycle similar to that of the inversion top, but it shows also a strong autumn cooling, which is responsible for the large annual cycle in inversion strength observed in Fig. 4b. This low-level cooling follows closely the sea surface temperature annual cycle illustrated at the bottom of Fig. 5a.
Variability of the 1200 UTC (0800 LT) wind profiles at the site is illustrated in Fig. 5b. At this morning transition time, near surface winds are weak with a typical southerly component. Around 1200 m the zonal component is generally easterly and the meridional component shows a transition to the northerly components that prevail at around 2000 m. Farther aloft the wind profiles transition to a decidedly westerly regime in the free troposphere above 500 hPa (not shown).
4. Cloudiness impact on the coastal boundary layer
a. Cloudiness and temperature inversions
Figure 6 explores the association between nighttime cloudiness, as described by CNF classes, and temperature inversions. Figure 6a shows the prevalence of CNF classes conditioned by inversion base height ranges. With Zb values less than 900 m, cloudy nights are more probable as Zb increases. With larger values of Zb this relationship stops or even reverts. In any case these results show that, although Zb and low-level clouds are somewhat related in the Mejillones Peninsula (Cerro Moreno airport), the relationship is not close, and both variables provide independent information on BL structure and dynamics.
Figure 6b shows the relationship between CNF and the occurrence of surface inversions. As nighttime cloudiness increases, the probability of finding a surface inversion in the 1200 UTC temperature profile greatly diminishes. For the clear-night class, on the other hand, surface inversions may or may not be present with equal probability, which shows that a clear night is not a sufficient condition for having a surface inversion at 1200 UTC (0800 LT) in the site.
b. Cloudiness and surface meteorology
The relationship between nighttime cloudiness and surface meteorology is illustrated in Fig. 7. Diurnal cycles of temperature, water vapor mixing ratio, and zonal and meridional winds are shown for low (≤30%) and high (≥70%) values of CNF. To reduce the effect of the annual cycle of the variables, the diurnal cycles have been computed for their anomalies with respect to their monthly means. Suppression of nocturnal surface cooling is the clearest effect of low-level cloudiness in air temperature (Fig. 7a), with mean dawn temperatures after a clear night being ∼2°C lower than after a cloudy night. The diurnal phase of the mean temperature cycle is much less affected by nighttime cloudiness, indicative of the efficient morning dissipation of cloudiness above the Mejillones Peninsula due to surface heating.
Mean water-vapor mixing ratio diurnal cycles (Fig. 7b) are generally weak but still show a different shape in clear and cloudy nights. The clear-night group shows a larger decrease during the night and a conspicuous peak during early morning. Both features suggest the nighttime development of a vertical gradient in moisture during clear nights and probably dew formation at the surface, which at sunrise mixes back into the BL producing the morning peak [a similar behavior is described for a prairie site by Betts and Ball (1995)]. After 1500 UTC the clear-night composite has slightly larger mixing ratios than the cloudy-night composite. The cooler and dryer surface conditions prevalent during clear nights have a net effect of a typical reduction of the lifting condensation level for surface air parcels as compared to cloudy nights (not shown), which emphasizes that this BL is not well mixed during clear nights.
Contrasts in surface winds (Figs. 7c,d) related to nighttime cloudiness are small. After a clear night mean southerly winds (onshore winds for the airport location) appear to be slightly larger (∼1 m s−1) than after a cloudy night, suggesting perhaps a minor intensification of the sea breeze during clear conditions. This larger southerly winds could also explain the slightly larger daytime mixing ratios observed in the clear-night composites of Fig. 7b.
c. Cloudiness and vertical structure
Figure 8 shows the mean vertical structure of temperature, humidity, and wind at 1200 UTC, conditioned by inversion base height and nocturnal cloudiness. To reduce the effect of the annual cycle, we have considered in this figure only data for the transitional months of April–May when the frequency of cloudy and clear nights is more similar. The largest differences in vertical structure are associated to the clear/cloudy contrast. Consistent with previous studies for offshore sites, cloudy nights show a boundary layer that is more well mixed in potential temperature and water vapor mixing ratio, and relative humidity that increases with height, reaching maximum values near 90% at its top. The mean temperature profiles associated with clear nights, on the other hand, are more stable and, although water vapor mixing ratios are larger than in the cloudy cases, the mean relative humidity in the boundary layer is remarkably uniform in the vertical, with values between 75% and 85%.
Some of the differences noted below the inversion base can, at least qualitatively, be explained by the mere presence or absence of a cloud layer at the top of the boundary layer, for example, the more well-mixing of the cloudy BL. Differences in profiles above the inversion base, however, may be more indirectly related to the presence of clouds. The most noticeable differences among them are the larger temperatures at the inversion top associated with cloudy nights and the larger northerly winds at the 1000–3000-m layer associated with clear nights. Larger inversion-top temperatures on cloudy nights may indicate relatively stronger subsidence and horizontal divergence in the lower troposphere associated to synoptic-scale forcings and/or to regional effects.
5. Local analysis
The cloudiness-conditioned mean vertical structures observed over Antofagasta are interpreted in this section with the aim of providing physical insights on the factors underlying the differences found for clear and cloudy night cases.
a. Cloud-topped boundary layer
The thermodynamic vertical structure of cloudy cases in Fig. 8 is similar to what is expected for a coupled stratocumulus-capped marine boundary layer. It is approximately well mixed in potential temperature and a little less in water mixing ratio with its relative humidity increasing steadily to maximum values at its top. To put these results in context, we use the model of Schubert et al. (1979) for a horizontally homogeneous stratus-capped marine boundary layer. In particular, they compute stationary solutions of the model and describe the sensitivity of cloud-top (or inversion base) heights upon sea surface temperature (SST) and lower-tropospheric horizontal wind divergence. For lack of observations of the latter in our case, we simply compute stationary solutions of the model for an ensemble of forcing conditions and then average the results for the same inversion height classes as those used in Fig. 8. The background potential temperature and mixing ratio profiles have been defined so as to approximate the mean profiles of Fig. 8 above the boundary layer. The other two differences with the Schubert et al. (1979) computations are a 3 m s−1 surface wind and a zero shortwave radiative flux.
Bold lines in Fig. 9 show results obtained with the model for a forcing ensemble constructed with a fixed SST = 15°C and varying divergence values (in the range associated to vertical velocities at 2000 m between 0.4 and 0.9 cm s−1). These profiles share many of the features of the cloudy case mean profiles in Fig. 8. An alternative forcing ensemble can be constructed by fixing the divergence value and varying the SST. Results for this ensemble are shown with fine lines in Fig. 9a (obtained with a divergence corresponding to a vertical velocity of 0.5 cm s−1 at 2000 m and SSTs in the 10.7°–16.5°C range). An interesting qualitative difference between the results obtained with the two forcing ensembles is the change of BL top temperature with inversion base height. In the divergence ensemble, deeper boundary layers have a much cooler top, while in the SST ensemble the BL top temperature is similar in both shallow and deep BL cases, indicating that in these cases the warming of the BL top by the sea surface is offset by adiabatic cooling associated with the deeper BL. Figure 8 shows a cooler BL top for a deeper BL. We interpret this result as suggesting that at this location the variability of the cloudy BL height is more controlled by variability in lower-tropospheric divergence than by SST variability.
b. Temperature advection and vertical velocity
Under the geostrophic and hydrostatic assumptions the wind vertical profile permits the estimation of horizontal advection terms in the temperature budget equation by using the thermal wind relationship. Figures 10a–c show the zonal, meridional, and total horizontal temperature advection computed using smoothed versions of the mean wind vertical profiles of Figs. 8e,f. Warm zonal advection appears more important in the inversion region right above the BL and has larger values in the clear than in the cloudy cases. This zonal advection is associated with easterly winds prevalent in the region around 1200 m and with the negative meridional wind shear existing below 1500 m (see Fig. 5b and Figs. 8e,f). The latter shear is indicative of the persistent west–east temperature gradient associated with the cold sea − warm desert contrast. The meridional thermal advection, on the other hand, is small in the inversion layer, but is larger and positive above it. This positive meridional advection results from winds with a generally northerly component and a positive zonal wind shear (see again Fig. 5b and Figs. 8e,f). This latter shear is indicative of a south–north temperature gradient of climatological and/or regional origin. Combining both advections, Fig. 10c shows that the column above the BL is subject to generally warm advection enhanced in the inversion layer, especially in the clear-night cases.
In the long term the advective tendencies above the BL have to be balanced by the remaining terms in the budget, namely, vertical advection, radiative effects, and the mean temperature tendency at 1200 UTC. Just to get a rough estimate of the vertical advection term, we shall consider that the mean temperature tendency at 1200 UTC is negligible in the balance and that the free tropospheric radiative cooling rate is 1.5°C day−1, in the middle of the range of typical clear-air cooling rates as mentioned by Moeng (1998). With these assumptions we can estimate the vertical advection term of the temperature budget, which results from the composition of the vertical velocity and the vertical gradient of potential temperature. The latter gradient can be estimated from Fig. 8b so that we can estimate the vertical profile of vertical velocity, shown in Fig. 10d. Values estimated with this simple methodology have signs and magnitudes that are reasonable for the region. Subsidence appears to predominate above the BL, and larger subsidence magnitudes in the lower troposphere are associated to the cloudy and shallower MBL cases, as compared to the clear and deeper MBLs. The large values of vertical velocity in the BL that appear in Fig. 10d have no physical meaning and arise because in this layer the assumptions of the estimates are clearly not valid. In particular, turbulence is an important term of the temperature and the momentum budget equations in the BL, but it was ignored in the estimate.
c. Turbulence-related parameters
Figure 11 shows profiles of derived parameters associated to the turbulence and stability state in the column. In this case the parameters have been computed for the individual soundings, and the resulting median profiles are shown (computation of wind shear profiles includes additional smoothing in order to condition this noisier variable). The same transitional months as in the previous section have been considered in the analysis. Figures 11a,b show wind shear and stability profiles. Both parameters show peaks in the region above the MBL. Shear and stability of the individual profiles were used to compute gradient Richardson numbers (Ri), whose median profiles are shown in Fig. 11c. Like shear and stability, Ri also maximizes in the inversion layer above the MBL. Maximum values of Ri at the top of the MBL are generally smaller for the deeper and clear MBLs, as compared to the shallower and cloudy cases. A physical interpretation of this feature will be undertaken in the last section of the paper.
A last parameter computed is a so-called Bowen ratio (Bo), calculated as the ratio of the potential temperature and water vapor mixing ratio vertical gradients multiplied by the specific heat of air at constant pressure and latent heat of evaporation of water, respectively. In turbulent conditions with similar eddy diffusivities for water vapor and temperature, this ratio should approach the ratio of sensible and latent heat fluxes. Figure 11d shows the median Bo profiles for the four composites considered. As turbulence is more prevalent below the inversion base, the significance of these Bo profiles is restricted to the lower layers. Negative values of Bo indicate the prevalence of stable conditions during these clear-night cases, as has been found in other similar coastal upwelling regions (e.g., Strom and Tjernstrom 2004). The Bo values in the MBL appear controlled more by the clear/cloudy difference than by the MBL depth.
6. Regional analysis
The previous sections have considered data only in the atmospheric column above Antofagasta, including surface meteorology, cloud observations, and radiosonde thermodynamic and wind vertical profiles. In this section we expand the analysis into a regional scale, trying to answer the questions 1) how representative are the Antofagasta clouds of the regional low-cloud field? and 2) are the differences observed in the vertical atmospheric structure (especially above the MBL) for clear and cloudy nights associated to any regional circulation pattern?
a. Representativeness of Antofagasta cloudiness
To address the first question above we compared Antofagasta cloud observations with a satellite-derived low-cloud status. The latter was based on the brightness temperature difference of channels 2 (3.9-μm wavelength) and 4 (10.7-μm wavelength), DT2m4, of the 0845 UTC GOES-12 infrared images for the period 2004–08. Values of DT2m4 in this region have a clear bimodal distribution (not shown) associated with low and high low-level cloudiness (presence of high clouds are not very frequent and can be clearly distinguished and left out of the analysis by detecting very large values of DT2m4). Presence (absence) of low clouds in the satellite pixel closest to Antofagasta was diagnosed if DT2m4 was smaller (larger) than a threshold temperature, Tc (Lee et al. 1997). An optimal value of Tc was chosen by maximizing the consensus between satellite-derived and ceiling-derived low-cloud status. Using Tc = −1.5°C, the consensus between both cloudiness estimates is between 80% and 95% for all months in spite of the large seasonal variation in nighttime cloudiness (section 3).
After establishing an appropriate value of Tc, we use it to diagnose a low-cloudiness status over all of the region covered by the available infrared images. Then, we compute a consensus fraction between the cloudiness status of the Antofagasta pixel as compared with all pixels in the region. The resulting consensus fraction maps are shown in Figs. 12a,b for the periods March–May and September–November. The regional representativeness of low clouds in Antofagasta changes considerably over the seasons. In austral spring, Antofagasta nocturnal cloudiness is clearly part of the regional stratocumulus field. In the fall, however, its representation is more constrained to a coastal band. According to these maps, low-cloudiness status for Antofagasta at 0845 UTC agrees with the cloud status in a coastal band about 100 km wide in more than 60% of the days in the period from March to May. Note that this representation applies to instantaneous 0845 UTC cloud status, and it may be larger if we consider cases in which Antofagasta was almost all-night clear or cloudy, as the low or high CNF values represent. Figures 12c,d further validate ceiling observations at Antofagasta. They show the differences in satellite-derived mean cloud frequencies for low-CNF and high-CNF days, for the same 3-month periods shown in Figs. 12a,b. During austral fall (Fig. 12c), clear (cloudy) nights observed at Antofagasta correspond to low (high) cloudiness along the same coastal band to the north of Antofagasta described in Fig. 12a. In austral spring, on the other hand, Antofagasta low-level cloudiness is more related to the cloudiness status of a coastal band located to the south of Antofagasta.
b. Regional circulation anomalies
To address the second question we analyzed NCEP–NCAR 40-Year Reanalysis fields of winds, vertical velocity, and geopotential heights for the months of April–May in the 1979–2007 period. As in the previous sections, these transitional months were selected so as to have a similar number of cloudy and clear nights in the composites. We calculated the anomalies of each variable with respect to their averages and then computed the mean anomalies of cases with low CNF values (clear nights) minus the mean anomalies of cases with high CNF values (cloudy nights). These differences can be compared with the differences in the vertical structures of Fig. 8, and should point to the regional anomaly patterns to which the latter can be related.
The vector field in Fig. 13 shows the clear − cloudy anomaly differences of 1200 UTC winds at 700 hPa. The northerly anomaly above Antofagasta is comparable with what was found for radiosonde profiles (Fig. 8). The figure indicates that this anomaly is associated with a cyclonic anomaly over the SE Pacific basin off the Peru and Chile coast. In turn, this anomaly is related to a relatively colder lower troposphere in the same region, manifested for example in the anomaly difference of 700-hPa geopotential height shown in Fig. 14. Along the coast of northern Chile, the relatively colder lower troposphere over the ocean is associated, by thermal wind arguments, to enhanced northerly winds at 700 hPa and enhanced southerlies near the surface, as evinced by the shaded contours in Fig. 13, which show anomaly differences of near-surface meridional winds derived from morning-pass QuikSCAT data (2000–07 period). Finally, the interaction of the cyclonic wind anomaly with the topography induces positive vertical velocity anomalies along the coastal regions of Peru and northern Chile, as illustrated by the shaded contours in Fig. 14. These differences in vertical velocity for clear and cloudy cases agree in sign and order of magnitude with the estimates shown in Fig. 10d.
7. Summary and concluding remarks
Based on 29 years (1979–2007) of surface and upper-air observations at Cerro Moreno Airport (23.4°S, 70.4°W) we have presented 1) a basic climatology of low-level cloudiness and subsidence inversion properties and 2) an analysis of the association between nighttime cloudiness and the mean structure of the boundary layer and lower troposphere. We consider the site as generally representative of the coastal marine boundary layer along northern Chile and of the eastern coastal margin of the SE Pacific stratocumulus region.
The main results of the climatological characterization are the following. Low-level clouds (ceilings less than 1500 m AGL) occur at the site mostly during the night, as daytime surface warming rapidly dissipates them. Seasonal frequency of nighttime low clouds peaks during austral winter and spring, although very low ceilings (less than 500 m AGL) are more frequent in winter. Vertical profiles of temperature at 1200 UTC (0800 LT) show a very large prevalence of elevated subsidence inversions with base heights typically between 800 and 1100 m. Their seasonal cycle shows ∼300 m amplitude at inversion base and top. As the inversion base cools significantly in austral winter, median inversion strengths larger than 10°C are found between May and October. Surface-based inversions at 1200 UTC are present in about 15% of the soundings. They are generally shallow and weak and are more frequent in the fall season, as nighttime cloudiness is less frequent than in winter and spring and nights are longer than in the summer.
Nighttime cloudiness was shown to have a large association with BL dynamics and structure. Although the amplitude of the diurnal cycle of temperature is relatively small at the coast, nocturnal cooling at this site is strongly conditioned by nighttime clouds, with end-of-the-night temperatures after a clear night being typically 2°C less than after cloudy nights, for which average nocturnal cooling is very little. Clear nights show also a larger decrease of near-surface water vapor mixing ratio probably due to dew formation. As the sun rises, dew at the surface rapidly evaporates producing a conspicuous morning peak in the surface mixing ratio diurnal cycle for clear nights. Diurnal cycle of surface winds show little association with nighttime cloudiness. For clear and cloudy cases, mean winds at the coast diminish steadily along the night, reaching near calm conditions at dawn.
Similar to what has been found for open ocean sites, like those described in the introduction, the thermodynamic structure of this coastal BL differs markedly between cloudy and clear conditions. In the former the BL is more well mixed in temperature and mixing ratio, with mean relative humidity profiles increasing from ∼80% near surface up to maximum values ∼90% at the top. In contrast, the clear-night BL shows a stable temperature profile and larger vertical gradients in mixing ratio, although mean relative humidity is more uniform with height as compared to the cloudy case.
Differences between the cloudy and clear composites are also noticeable above the BL. Inversion top temperatures are on average ∼1°C larger in cloudy cases, which, together with their ∼2°C cooler inversion bases, produce a MBL capping inversion that is more stable compared to the clear night cases. Another difference aloft is observed in the vertical profiles of meridional winds. Clear nights appear associated to more intensified northerly winds in the 1000–3000-m layer as compared to the cloudy cases. Mean winds in the BL at this time (1200 UTC, 0800 LT) are weak in all cases, consistent with the analysis of the diurnal cycles of surface winds.
Insights into physical mechanisms underlying the findings above have been sought by comparing the cloudy composites to results of the simple model of Schubert et al. (1979) and by computing derived parameters pertaining to the temperature budget and the turbulent state of the lower troposphere. The results show physically significant differences in subsidence, horizontal temperature advection, and winds in the lower troposphere associated with the mean clear and cloudy coastal BL, as explained below.
A simplified thermal balance above the BL suggests that cloudy cases are associated with smaller warm advection and larger subsidence values at the 1000–3000-m layer, as compared to the clear cases. The importance of subsidence in the cloudy cases is also suggested by the simple model, which reproduces well the observed change of BL structure with inversion base height when forced by an ensemble of divergence values at fixed SST, more than by an ensemble of SSTs at fixed divergence. Finally, it was shown that deeper cloudy BLs are generally capped by a less stable inversion layer, probably related to weaker subsidence. By contrast, the clear BLs are capped by less stable inversion layers, although the median stability values do not change much with inversion base height. Richardson number profiles summarize well the relationship between BL height and cloudiness described above. Typical capping Ri values at the top of the MBL are smaller for clear and deeper BLs, as compared to the clear or shallower cases. We interpret these results considering that they describe an equilibrium condition of the BL. Comparing the cloudy versus clear conditions, for example, the cloudy BL is more well mixed, presumably more turbulent, and, therefore, it requires a larger subsidence rate and a larger capping Ri in order to maintain an equilibrium in its height and turbulence intensity. Comparing the shallow versus deeper cases, on the other hand, a shallower mean MBL may be associated to a larger subsidence rate, which, in equilibrium, requires a more turbulent BL and a larger capping Ri value at its top.
The smaller subsidence values, enhanced northerly winds above the BL, and cooler lower troposphere inferred over Antofagasta for clear nights, as compared to cloudy nights, have been examined in a regional context by analysis of NCEP–NCAR 40-Year Reanalysis fields and near-surface QuikSCAT satellite-derived winds. The anomaly differences shown by the reanalysis are generally consistent with the local analysis. They suggest that coastal clear nights are associated with a cold anomaly in the lower troposphere over the SE Pacific basin off Peru and Chile, which by thermal wind argument induces, along the Chilean coast, anomalies of southerly winds near the surface and northerly winds above the BL, while at the same time reducing the coastal subsidence in the lower troposphere.
The description above is based on averages of a large number of cases and for rather idealized (mostly) clear and (mostly) cloudy nighttime conditions. Therefore, we have strived not to discuss about cause–effect relationships between the variables, but more about association between them. What happens on individual days, when, for example, nighttime coastal cloud cover may be more variable, has not been addressed here nor have we addressed the covariability of cloudiness and inversion properties at synoptic or long-term scales. Other recent papers have addressed the synoptic variability of the MBL and low-level clouds in the SE Pacific (Painemal and Zuidema 2010; George and Wood 2010; Rahn and Garreaud 2010). While synoptic variability of this MBL is considered to be reasonably well represented in numerical models, the persistent biases in height and cloudiness mentioned in the introduction suggest that the factors controlling long-term averages and equilibriums of this MBL are still not well understood. To this purpose, we considered it relevant to present this conditional climatology, as the idealized average structures shown here may represent states toward which this coastal BL is attracted depending on its forcings. In this regard it is interesting to note that for the same BL height, we have shown clear and cloudy mean states that are significantly different in vertical structure and forcing mechanisms. This is reminiscent of the multiequilibrium states of MBLs suggested by Randall and Suarez (1984) based upon modeling results, and it is probably one more reason why this MBL is so difficult to model correctly.
Acknowledgments
Partial funding by projects Conicyt ACT-19 and Fondecyt 1090412 is acknowledged. Data analyzed in this work originates in long-term operational observational efforts by the Dirección Meteorológica de Chile (DMC). Comments by Drs. Patricio Aceituno, René Garreaud, and two anonymous reviewers improved the manuscript. QuikSCAT data are produced by Remote Sensing Systems and sponsored by the NASA Ocean Vector Winds Science Team. Data are available at www.remss.com.
REFERENCES
Ahlgrimm, M., D. A. Randall, and M. Kohler, 2009: Evaluating cloud frequency of occurrence and cloud-top height using spaceborne lidar observations. Mon. Wea. Rev., 137 , 4225–4237.
Albrecht, B. A., M. P. Jensen, and W. J. Syrett, 1995: Marine boundary layer structure and fractional cloudiness. J. Geophys. Res., 100D , 14209–14222.
Berríos, P., 2008: Stratocumulus along the northern coast of Chile: Variability and tendencies (in Spanish). B.S. thesis, Dept. of Meteorology, University of Valparaíso, Chile, 118 pp.
Betts, A. K., and J. H. Ball, 1995: The FIFE surface diurnal cycle climate. J. Geophys. Res., 100D , 25679–25693.
Betts, A. K., C. S. Bretherton, and E. Klinker, 1995: Relation between mean boundary-layer structure and cloudiness at the RV Valdivia during ASTEX. J. Atmos. Sci., 52 , 2752–2762.
Bony, S., and J.-L. Dufresne, 2005: Marine boundary layer clouds at the heart of tropical cloud feedback uncertainties in climate models. Geophys. Res. Lett., 32 , L20806. doi:10.1029/2005GL023851.
Burk, S. D., and W. T. Thompson, 1996: The summertime low-level jet and marine boundary layer structure along the California coast. Mon. Wea. Rev., 124 , 668–686.
Durre, I., and X. Yin, 2008: Enhanced radiosonde data for studies of vertical structure. Bull. Amer. Meteor. Soc., 89 , 1257–1262.
Durre, I., R. S. Vose, and D. B. Wuertz, 2006: Overview of the integrated global radiosonde archive. J. Climate, 19 , 53–68.
Flores, V., G. Vargas, J. Rutllant, and J. L. Roux, 2009: Estacionalidad de la erosión y el transporte eólico de partículas en el desierto costero de Atacama, Chile (23 s). Andean Geol., 36 , 288–310.
Garreaud, R., and R. C. 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., and J. Rutllant, 2006: High frequency atmospheric variability in the eastern rim of the southeast Pacific anticyclone. Cienc. Tecnol. Mar, 29 , 5–13. [Available online at www.cona.cl/ctmol2/revista.htm].
Garreaud, R., J. Rutllant, and H. Fuenzalida, 2002: Coastal lows along the subtropical west coast of South America: Mean structure and evolution. Mon. Wea. Rev., 130 , 75–88.
George, R. C., and R. Wood, 2010: Subseasonal variability of low cloud radiative properties over the southeast Pacific Ocean. Atmos. Chem. Phys., 10 , 4047–4063.
Ghate, V. P., B. Albrecht, C. W. Fairall, and R. A. Weller, 2009: Climatology of surface meteorology, surface fluxes, cloud fraction, and radiative forcing over the southeast Pacific from buoy observations. J. Climate, 22 , 5527–5540.
Gille, S. T., S. G. L. Smith, and S. M. Lee, 2003: Measuring the sea breeze from QuikSCAT scatterometry. Geophys. Res. Lett., 30 , 1114. doi:10.1029/2002GL016230.
Glickman, T. E., 2000: Glossary of Meteorology. 2nd ed. Amer. Meteor. Soc., 855 pp.
González, C., 2009: Frequency of coastal stratocumulus between Paposo (24.5°S) and Papudo (32.5°S) in the 1998–2002 period (in Spanish). B.S. thesis, Dept. of Meteorology, University of Valparaiso, Chile, 46 pp.
Hannay, C., D. L. Williamson, J. J. Hack, J. T. Kiehl, J. G. Olson, S. A. Klein, C. S. Bretherton, and M. Kohler, 2009: Evaluation of forecasted southeast Pacific stratocumulus in the NCAR, GFDL, and ECMWF models. J. Climate, 22 , 2871–2889.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437–472.
Kalthoff, N., and Coauthors, 2002: Mesoscale wind regimes in Chile at 30°S. J. Appl. Meteor., 41 , 953–970.
Klein, S. A., 1997: Synoptic variability of low-cloud properties and meteorological parameters in the subtropical trade wind boundary layer. J. Climate, 10 , 2018–2039.
Koracin, D., and C. E. Dorman, 2001: Marine atmospheric boundary layer divergence and clouds along California in June 1996. Mon. Wea. Rev., 129 , 2040–2056.
Lee, T. F., F. J. Turk, and K. Richardson, 1997: Stratus and fog products using GOES-8–9 3.9-μm data. Wea. Forecasting, 12 , 664–677.
Leidner, S. M., D. R. Stauffer, and N. L. Seaman, 2001: Improving short-term numerical weather prediction in the California coastal zone by dynamic initialization of the marine boundary layer. Mon. Wea. Rev., 129 , 275–294.
Moeng, C.-H., 1998: Stratocumulus-topped atmospheric planetary boundary layer. Buoyant Convection in Geophysical Flows, E. J. Plate, Ed., Kluwer, 421–440.
Muñoz, R. C., 2008: Diurnal cycle of surface winds over the subtropical south east Pacific. J. Geophys. Res., 113 , D13107. doi:10.1029/2008JD009957.
National Climate Data Center 2008: Federal climate complex data documentation for integrated surface data. Tech. Rep., National Climate Data Center, 121 pp. [Available online at www1.ncdc.noaa.gov/pub/data/ish/ish-format-document.pdf].
Norris, J. R., 1998: Low cloud type over the ocean from surface observations. Part I: Relationship to surface meteorology and the vertical distribution of temperature and moisture. J. Climate, 11 , 369–382.
Painemal, D., and P. Zuidema, 2010: Microphysical variability in southeast Pacific stratocumulus clouds: Synoptic conditions and radiative response. Atmos. Chem. Phys., 10 , 6255–6269.
Piñones, A., J. C. Castilla, R. Guiñez, and H. Largier, 2007: Nearshore surface temperatures in Antofagasta bay (Chile) and adjacent upwelling centers. Cienc. Mar., 33 , 37–48.
Rahn, D. A., and R. Garreaud, 2010: Marine boundary layer over the subtropical southeast Pacific during VOCALS-REx. Part 2: Synoptic variability. Atmos. Chem. Phys., 10 , 4507–4519.
Randall, D. A., and M. J. Suarez, 1984: On the dynamics of stratocumulus formation and dissipation. J. Atmos. Sci., 41 , 3052–3057.
Rutllant, J. A., H. Fuenzalida, and P. Aceituno, 2003: Climate dynamics along the arid northern coast of Chile: The 1997–1998 Dinámica del Clima de la Región de Antofagasta (DICLIMA) experiment. J. Geophys. Res., 108 , 4538. doi:10.1029/2002JD003357.
Schubert, W. H., J. S. Wakefield, E. J. Steiner, and S. K. Cox, 1979: Marine stratocumulus convection. Part I: Governing equations and horizontally homogeneous solutions. J. Atmos. Sci., 36 , 1286–1307.
Serpetzoglou, E., B. A. Albrecht, P. Kollias, and C. W. Fairall, 2008: Boundary layer, cloud, and drizzle variability in the southeast Pacific stratocumulus regime. J. Climate, 21 , 6191–6214.
Stevens, B., A. Beljaars, S. Bordoni, C. Holloway, M. Kohler, 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.
Strom, L., and M. Tjernstrom, 2004: Variability in the summertime coastal marine atmospheric boundary-layer off California, USA. Quart. J. Roy. Meteor. Soc., 130 , 423–448.
World Meteorological Organization 2002: Guide to the GCOS surface and upper-air networks: GSN and GUAN. version 1.1. Tech. Rep., World Meteorological Organization, 37 pp. [Available online at www.wmo.int/pages/prog/gcos/Publications/gcos-73.pdf].
Wood, R., and Coauthors, 2010: The VAMOS Ocean–Cloud–Atmosphere–Land Study Regional Experiment (VOCALS-REx): Goals, platforms, and field operations. Atmos. Chem. Phys. Discuss., 10 , 20769–20822. doi:10.5194/acpd-10-20769-2010.
Zamora, R., 2010: Observational characterization of the marine boundary layer in Antofagasta (in Spanish). M.S. thesis, Dept. of Geophysics, University of Chile, 122 pp.
Zeng, X., M. A. Brunke, M. Zhou, C. Fairall, N. A. Bond, and D. H. Lenshow, 2004: Marine atmospheric boundary layer height over the eastern Pacific: Data analysis and model evaluation. J. Climate, 17 , 4159–4170.
(a) Topography around Mejillones Peninsula. Line contours from 0 to 900 m MSL every 100 m. Shaded contours from 1000 to 2000 m MSL every 250 m. Black circle marks location of Antofagasta radiosonde station. Rectangle in the inset shows location of Mejillones Peninsula in South America. Visible GOES image at (b) 1039, (c) 1139, and (d) 1639 UTC 11 Nov 2006.
Citation: Journal of Climate 24, 4; 10.1175/2010JCLI3714.1
(a) Hourly variation (for all months) and (b) monthly variation (for 0900 UTC) of the occurrence frequency of four ceiling categories. From darker to lighter shading the categories are ceiling < 500 m, 500 m ≤ ceiling < 1500 m, 1500 m (all AGL) ≤ ceiling < limited, and unlimited ceiling. (c) Monthly variation of the occurrence frequency of three CNF categories. From darker to lighter shading the categories are CNF ≥ 70%, 30% < CNF < 70%, and CNF ≤ 30%.
Citation: Journal of Climate 24, 4; 10.1175/2010JCLI3714.1
Frequency distribution of (a) all inversion bases in radiosonde temperature profiles (black bars) at 1200 UTC in the 1979–2007 period and of elevated inversion bases after application of selection definition algorithm (white bars) (see text). (b) Monthly frequency of surface (black bars) and elevated (white bars) inversions.
Citation: Journal of Climate 24, 4; 10.1175/2010JCLI3714.1
Boxplots of the (a) monthly distributions of inversion base heights, (b) inversion temperature jump, and (c) inversion-relative humidity jump. Rectangular boxes extend between the lower and upper quartiles of each distribution, and horizontal lines inside the boxes mark the location of the median. Vertical segmented lines outside the boxes show the extent of the data. Data points lying more than 1.5 times the interquartile range from the upper or lower quartile are considered outliers and plotted individually with a cross.
Citation: Journal of Climate 24, 4; 10.1175/2010JCLI3714.1
(a) Shading marks the interquartile ranges of temperature (continuous border line) and dewpoint temperature (dashed border line) profiles measured at Antofagasta at 1200 UTC in the 1979–2007 period. Lines with symbols show the monthly evolution of the medians of inversion properties: inversion base and top temperatures (solid lines); inversion base and top dewpoint temperatures (dashed lines). Symbols mark each month; big circle marks January and big triangle marks February. For reference, near-surface points show the mean annual cycle in the period 1979–2006 of minimum sea surface temperature measured in Antofagasta bay (data available online at www.shoa.cl). (b) As in (a), but for zonal (continuous lines) and meridional wind (dashed lines).
Citation: Journal of Climate 24, 4; 10.1175/2010JCLI3714.1
(a) Occurrence frequency of three CNF classes categorized by inversion base height (Zb) ranges: CNF classes and shading code as in Fig. 2c. (b) Frequency of surface inversions categorized by CNF ranges. Gray shading indicates no surface inversion; white shading indicates existence of surface inversion.
Citation: Journal of Climate 24, 4; 10.1175/2010JCLI3714.1
Diurnal cycles of surface meteorology conditioned by CNF: (a) air temperature, (b) water vapor mixing ratio, (c) zonal wind, and (d) meridional wind. Values correspond to diurnal cycles of anomalies with respect to the monthly averages of each variable, to which the global mean (indicated by the fine horizontal line) has been added. Solid lines are for cases with CNF ≤ 30% and dashed lines are for cases with CNF ≥ 70%.
Citation: Journal of Climate 24, 4; 10.1175/2010JCLI3714.1
Mean April–May vertical profiles at 1200 UTC conditioned by CNF and inversion base height:(a) temperature profiles, (b) potential temperature, (c) relative humidity, (d) water vapor mixing ratio, (e) zonal wind speed, and (f) meridional wind speed for CNF ≤ 30% (bold lines) and CNF ≥ 70% (fine lines) and 500 m < Zb < 800 m (dashed) and 900 m < Zb < 1200 m (continuous).
Citation: Journal of Climate 24, 4; 10.1175/2010JCLI3714.1
Vertical profiles of (a) temperature, (b) potential temperature, (c) relative humidity, and (d) total and liquid water mixing ratios, as computed with stationary model of Schubert et al. (1979). Bold continuous (dashed) lines are averages for cases in which inversion base is in the range 900–1200 m (500–800 m) for fixed SST = 15°C, and varying divergence values. In (a) fine lines are results for fixed divergence of 2.5 × 10−6 s−1 and varying SST values. See text for details.
Citation: Journal of Climate 24, 4; 10.1175/2010JCLI3714.1
Components of the temperature budget derived from smoothed mean profiles in Fig. 8 using thermal wind assumption (see text): (a) zonal temperature advection, (b) meridional temperature advection, (c) total horizontal temperature advection, and (d) vertical velocity required to close the temperature budget assuming a 1.5°C day−1 radiative cooling rate. Line codes as in Fig. 8.
Citation: Journal of Climate 24, 4; 10.1175/2010JCLI3714.1
Median of turbulent indices computed for soundings composited upon CNF and inversion base height: line codes as in Fig. 8. Resulting profiles have been smoothed. (a) Shear squared, (b) Brunt–Väisälä frequency squared, (c) gradient Richardson number, and (d) Bowen ratio estimate (see text).
Citation: Journal of Climate 24, 4; 10.1175/2010JCLI3714.1
Consensus fraction between low clouds at the Antofagasta pixel and the regional cloud field as derived from GOES-12 0845 UTC infrared images for (a) March–May and (b) September–November 2004–08. Differences in satellite-derived mean cloud frequencies for low-CNF and high-CNF days for (c) March–May and (d) September–November 2004–07. A value of −100% in (c) and (d) would correspond to perfect coincidence between low-cloud frequency derived from satellite and that derived from Antofagasta ceiling observations.
Citation: Journal of Climate 24, 4; 10.1175/2010JCLI3714.1
Mean differences of anomalies for almost-clear (CNF < 30%) minus almost-cloudy (CNF > 70%) night conditions at Antofagasta: anomalies of NCEP–NCAR 40-Year Reanalysis winds (vectors) at 700 hPa and 1200 UTC (April–May 1979–2007). A cross at the coast marks the Antofagasta location. The anomaly vector offshore from Antofagasta has a magnitude of 1.2 m s−1. Anomalies of QuikSCAT meridional component of surface morning winds (contours) for the April–May 2000–07 period: shaded contours plotted at 0.5, 1, and 2 m s−1; fine-line contour plotted at −0.5 m s−1.
Citation: Journal of Climate 24, 4; 10.1175/2010JCLI3714.1
Mean differences of anomalies for almost-clear (CNF < 30%) minus almost-cloudy (CNF > 70%) night conditions at Antofagasta: anomalies of 700-hPa NCEP–NCAR Reanalysis geopotential heights (gpm) in bold contours and anomalies of NCEP–NCAR Reanalysis vertical velocity (thin contours) averaged between 925 and 700 hPa: shaded contours plotted at 0.05, 0.10, and 0.15 cm s−1; fine-line contours plotted at −0.05 and −0.10 cm s−1. Analysis period is April–May 1979–2007. Basic variables used in both cases correspond to daily averages. A cross at the coast marks Antofagasta location.
Citation: Journal of Climate 24, 4; 10.1175/2010JCLI3714.1
