1. Introduction
Mesoscale convective systems (MCSs) are the major source of rain and cold cloudiness (Houze 2004; Nesbitt et al. 2000; Del Genio and Kovari 2002; Fiolleau and Roca 2013a; Roca et al. 2014) in the tropics. They inject at middle-to-high altitudes of the troposphere large quantities of ice that may persist for several hours after the rain has ceased. The tropical dynamics is also strongly coupled with the diabatic processes, in which MCSs are key factors, through latent heat release or modulation by the radiative cooling. By synthesizing former studies, Houze (1982) proposed a schematic life cycle of the MCS with three main stages: initiation, maturity, and dissipation. Embedded within these stages is the evolution of internal subregions of the MCS. In the initiation stage, the MCS may be seen as isolated convective towers leading to intense precipitation at ground level. These towers persist in the mature stage, but there is also a region of lighter precipitation, the stratiform region. Both the convective cores and the stratiform regions are surrounded by nonprecipitating high clouds: the nonprecipitating anvil (or cirriform anvil). In the dissipation stage, only this cloud shield remains. Throughout the life cycle, each of these MCS subregions interacts with its environment through latent and radiative processes (Sherwood and Wahrlich 1999) with an intensity that is not necessarily accurately quantified at the scale of the tropics. However, Schumacher et al. (2004) show that an improved representation of the vertical shape and geographical distribution of the diabatic heating profile leads to a more realistic upper-level circulation.
Because of their horizontal scale, convective processes remain parameterized within general circulation models. Some of these parameterizations, in particular that of Donner et al. (2001), include a degree of mesoscale organization by including a representation of the internal circulation. However, in spite of this example, MCSs and their associated effects (particularly in terms of diabatic heating) are partly represented through the convective and large-scale cloud schemes that in turn feed microphysical and radiative schemes. In this way, MCSs can have impact on a larger scale as a result of the combined effects of convective, large-scale cloud, and radiative schemes. However, all these parameterized mesoscale processes have a major contribution on the radiative budget of the tropics. Although earlier studies have established that the effects of deep convective clouds in the longwave and shortwave domains nearly balance each other out at the top of the atmosphere (Hartmann et al. 2001; Ramanathan et al. 1989; Harrison et al. 1990; Thampi and Roca 2014), this balance may be disturbed if physical properties (size, optical properties, depth, and cloud-top altitude) are modified (Behrangi et al. 2012). This mean equilibrium is the result of a subtle combination of various processes that evolve throughout the life cycle, and better documentation and understanding of this complex combination is needed.
Physical processes involved in the life cycle differ according to the MCS subregion (i.e., convective, stratiform, or cirriform anvil). Several observational studies have focused on convective and stratiform regions, in particular because they are the regions that lead to precipitation (Schumacher and Houze 2006; Futyan and Del Genio 2007; Schumacher et al. 2007; Cetrone and Houze 2009). Most use the TRMM Precipitation Radar data, which, with 13.8-GHz frequency, appear very well suited to document differences that exist within these two regions. First of all, they found that oceanic systems have a higher stratiform rain fraction than continental systems in conjunction with a greater sustainability over oceans (Schumacher and Houze 2003, 2006; Yuter and Houze 1998). Because of differences in the physical processes of the two MCS subregions (Houze 2004), different hydrometeors are expected between them, leading to different rain intensity at the surface, as well as different radiative properties. Ultimately, these regions have very different latent heating profiles, with top-heavy profiles for stratiform regions and bottom-heavy profiles within convective parts that arise from differences in evaporation/condensation processes (Schumacher et al. 2004).
The nonprecipitating anvil properties have been widely investigated using satellite (Webster and Stephens 1980; Cetrone and Houze 2009), ground-based (Cetrone and Houze 2011; Powell et al. 2012), and aircraft measurements (Heymsfield and McFarqhar 1996; McFarquhar and Heymsfield 1996; Bouniol et al. 2010). Webster and Stephens (1980) infer differential radiative heating between the cloud base and cloud top that can reach 35 K day−1, demonstrating that these clouds and associated radiative effects must be taken into account for accurate heating calculations in the tropics. The anvil appears to be fed by detrainment from convective and stratiform regions, and lower ice contents are found in this region. As the age increases and convection intensity decreases, additional nucleation processes are at play and permit the anvil to remain. Therefore, an evolution of microphysical and radiative processes is expected over the life cycle.
Since it is difficult to access the life cycle using orbiting satellites (even with relatively low-orbiting satellites such as TRMM) or ground-based measurements (especially for propagating systems), earlier studies (see, e.g., Powell et al. 2012) tended to separate anvil clouds with respect to their physical depth or to the distance to convective cores (see Yuan and Houze 2010). This type of separation underlies the assumption that the decrease in depth of the anvil and/or the distance to convective cores measures the age of the anvil. In contrast, geostationary satellites are a relevant data source for following the evolution of convective systems (Futyan and Del Genio 2007; Fiolleau and Roca 2013a), in particular the infrared channel, which is not very sensitive to the diurnal cycle. However, it is well known that internal regions within an MCS cannot be accurately identified at these wavelengths alone and thus the physical properties (size and radiation) are generally only documented at the scale of the MCS. By combining NEXRAD and GOES images, Feng et al. (2011) discussed the opportunity to split the cirriform anvils into transitional (with echo base below 3 km and echo top above 6 km) thick and thin anvils thanks to the different dynamical characteristics among them. However, since their radiative characteristics were not different, these MCS subregions were ultimately grouped together; in the end, only three MCS subregions were documented. Using the combination of Feng et al. (2011) and the life stage definition of Futyan and Del Genio (2007), Feng et al. (2012) document the evolution of MCS and the factors affecting the anvil production over the United States. They found that the short lifetime of MCSs (less 6 h) is strongly correlated with the intensity of convection at the initial stage. However, for a longer lifetime, more convectively favorable environmental characteristics (such as midtropospheric humidity and upper-tropospheric wind shear) and convection organization are needed. The nonprecipitating anvil cloud area is strongly correlated with the size of the intense convective core, updraft strength, and stratiform region area.
In this paper, we make use of another composite approach that takes advantage of the ability of the geostationary satellite to follow the overall morphological evolution of MCS and of orbiting satellites, such as those of the A-Train constellation, which are able to sample the macrophysical, microphysical, and radiative properties of individual subregions of MCSs. The use of satellites ensures homogeneity in data sampling all over the tropics. A further aim of this paper is to document the evolution of the nonprecipitating anvil throughout the life cycle in conjunction with the evolution of the convective and stratiform regions, as it dominates the radiative budget (Feng et al. 2012) at the scale of the MCS. However, the active instrumentation of the A-Train used in this paper does not have scanning capability. Therefore the morphological evolution and the contribution of each MCS subregions (e.g., in terms of fractional coverage of the MCS area) over the life cycle will not be addressed in this paper. The evolution of physical properties throughout the MCS life cycle is contrasted among three geographical areas: one continental area (West Africa) and two oceanic areas (Atlantic Ocean and Indian Ocean). The first section describes the data and methodology used in this composite approach. This paper concentrates on the radiation aspects, and related MCS properties are documented and compared among these different geographical areas in order to contrast among environments with different sustainability (Houze 2004). The TRMM mission has demonstrated the variable nature of convection across the tropics (Houze et al. 2015), and our approach may be useful in understanding how different environments of convective processes impact the microphysical and radiative processes. Concluding remarks are given in the final section.
2. Data and method
The proposed approach for this study is Lagrangian, as opposed to the Eulerian approach, which derives the cloud properties at a given place. The MCSs are followed throughout their life cycle using an automatic algorithm called tracking of organized convection algorithm through a 3D segmentation (TOOCAN; Fiolleau and Roca 2013b). In TOOCAN, and following Houze and Betts (1981), an MCS is defined as a convective core with a cold brightness temperature associated with a stratiform part with a warmer temperature and lighter precipitation, and a cirriform part where no rain occurs in the spatiotemporal domain. In contrast, at an instantaneous time scale, we refer to a cloud cluster. A multiple detection threshold method is applied in order to identify convective cores. A clustering technique is then applied simultaneously in the space and time domain, allowing us to identify and characterize the MCSs with consistency throughout their life cycle [see Fiolleau and Roca (2013b) for details]. In the present study, the cold cloud shield is defined by a 235-K threshold. A segmented infrared image is then produced at the 30-min time scale, allowing us to locate the cloud shield associated with a convective core, and then to delineate a cloud cluster.
The choice of the 235-K level for segmentation of the infrared images has direct consequences on the area of cloud (and in particular cirriform cloud) that are considered to belong to a given cluster. Different infrared thresholds have been used in former studies: Futyan and Del Genio (2007) used a 235 W m−2 threshold (corresponding grossly to 254 K); Feng et al. (2011) identified the cold cloud shields for temperatures lower than 270 K; Tobin et al. (2012) used a 240-K threshold; and Yuan and Houze (2010) used a 260-K threshold in order to enclose the anvil out to its thinner edges. Mathon and Laurent (2001) estimated that the 253-K level is the highest temperature associated with convection. A threshold value of about 235 K is maintained in this paper because it has been commonly used (e.g., Kondo et al. 2006; Yuter and Houze 1998; Pope et al. 2008), it represents a good compromise for attributing nonprecipitating clouds to a cluster, and it excludes mid- or low-level clouds that would be in the vicinity of the main convective system. As an indirect validation, Table 1 and Fig. 1 demonstrate that a large fraction of the cloud shield with a brightness temperature lower than 235 K corresponds to nonconvective and nonstratiform regions of the MCS and that the mean cloud-top altitude (as observed by radar and lidar) is in the expected range for tropical cirrus.
Number of MCS subregions [convective (conv), stratiform (strat), and cirriform (cirri)] and the corresponding number of profiles sampled at each normalized life step (between 1 and 10) for the three geographical areas. Boldface numbers in the table highlight the subregions that were sampled in less than 20 independent MCSs.
Mean cloud-top height evolution over the MCS life cycle retrieved from radar data (solid line) and radar and lidar data (dashed line) for the three MCS subregions and the three included geographical areas.
Citation: Journal of Climate 29, 9; 10.1175/JCLI-D-15-0551.1
Mean cloud-top height evolution over the MCS life cycle retrieved from radar data (solid line) and radar and lidar data (dashed line) for the three MCS subregions and the three included geographical areas.
Citation: Journal of Climate 29, 9; 10.1175/JCLI-D-15-0551.1
Mean cloud-top height evolution over the MCS life cycle retrieved from radar data (solid line) and radar and lidar data (dashed line) for the three MCS subregions and the three included geographical areas.
Citation: Journal of Climate 29, 9; 10.1175/JCLI-D-15-0551.1
The objective of this paper is to monitor the evolution of moist physical properties during each phase of the life cycle. The main advantage of the TOOCAN algorithm is that it avoids the splitting and merging drawbacks commonly encountered in traditional tracking algorithms (Fiolleau and Roca 2013b). The MCS life cycle is thus well documented from the early initiation to the dissipation for all segmented MCSs. No selection needs to be applied to the MCSs according to the split/merge artifacts, and all detected MCSs can be included in our study, increasing the statistical representativeness of our results. In the following, the simple classification proposed by Fiolleau and Roca (2013b) is adopted. A 5-h threshold is applied to the life duration. The evolution of the size of the MCS is then examined, and only MCSs that have a unique well-defined maximum size evolution are kept. This simple classification ensures that the life cycles of the aggregated individual MCSs in our composites are as similar as possible [class 2a of Fiolleau and Roca (2013a)] and that systems with complex life cycles are omitted. The application of both these constraints means that about 35% of MCSs (Fiolleau and Roca 2013a) cannot be included in the present study. However, as already stressed for the tropics by Futyan and Del Genio (2007) and Fiolleau and Roca (2013b), if it is true that small short-lived systems dominate the population over both land and ocean, the less frequent larger and longer-lived systems provide most of the cloud cover and rainfall (Williams and Houze 1987; Mapes and Houze 1993; Machado et al. 1998; Roca and Ramanathan 2000; Mathon and Laurent 2001; Wilcox and Ramanathan 2001). The life cycle of each MCS is then normalized between 0% and 100% and discretized into 10 life steps. This normalization process differs substantially from the one used by Futyan and Del Genio (2007) or Feng et al. (2012), where a sixth-order polynomial is fitted to the radius and temperature data according to time of detection in order to define five life steps for Futyan and Del Genio (2007), and three for Feng et al. (2012). In this study, each step of the life cycle has a fixed duration that will therefore be different for systems with different life duration. This means that data from MCSs of different ages can be mixed in the same step. Pope et al. (2008) show the mean evolution of the MCS equivalent radius and the minimum temperature according to life duration. They found that the mean maximum equivalent radius and the MCS life time are related, and showed that the minimum temperature occurs approximately halfway through the MCS life span. Feng et al. (2012) look at the same relationship according to the MCS subregions and found the convective core size peaking before those of stratiform and anvil cloud. They also show that this lag increases linearly with system life time. Computing mean physical properties as a function of the normalized life stage rather than the system age allows information from MCSs with different life times, but with similarities in their evolution, to be combined.
To compare the evolution of physical properties over the life cycle in different environmental conditions (i.e., land or ocean), three main geographical areas are explored via two different geostationary satellites [Meteosat Second Generation-2 (MSG-2) and Meteosat-7]. MSG-2 acquires an image of the African continent and the adjacent Atlantic Ocean every 15 min. Meteosat-7 samples the Indian Ocean every 30 min. To remain homogeneous over the whole tropical belt, the TOOCAN algorithm is run at a 30-min resolution for both satellites. The periods of interest are in phase with the MCS peak of occurrence over the African continent and adjacent ocean (i.e., boreal summer) and in the ITCZ region of the Indian Ocean (i.e., boreal winter). In the following, we mainly focus on MCS properties in the three following domains:
continental West Africa (AF) between −20° and 25°E and 2° and 25°N and land profiles only,
Atlantic Ocean (ATL) between −40° and −10°E and 0° and 25°N and sea profiles only, and
open Indian Ocean (OIO) between 60° and 90°E and −25° and 5°N and sea profiles only.
Each A-Train sample that matches a TOOCAN trajectory is associated with a life step (between 1 and 10), belongs to a MCS subregion, and can be composited accordingly. The following properties are documented:
Macrophysical properties (number of cloud layers, cloud-top and cloud-base altitude, cloud frequency of occurrence, etc.), combining CloudSat radar and CALIPSO lidar detection strength, are obtained through the Level-2B Radar–Lidar Cloud Geometrical Profile (2B-GEOPROF-LIDAR; Mace and Zhang 2014). Both instruments are reprojected at the CloudSat radar resolution: 1.3 km (across track) × 1.7 km (along track) × 240 m vertically.
Microphysical properties are inferred from the CloudSat radar reflectivity provided in the Level-2B Cloud Geometrical Profile product (2B-GEOPROF; Marchand et al. 2008).
Radiative properties [outgoing longwave radiation (OLR), albedo, cloud radiative effect, etc.] at the top or bottom of the atmosphere using CloudSat Level-2B Radar–Lidar Fluxes and Heating Rates (2B-FLXHR-LIDAR; L’Ecuyer et al. 2008; Henderson et al. 2013) and the various estimates (CERES, MODIS, and MODIS enhanced) are provided along the CloudSat tracks in the CALIPSO, CloudSat, CERES, and MODIS (CCCM) dataset (Kato et al. 2011). However, to keep the figures readable, only the 2B-FLXHR-LIDAR and CERES products are shown. The CERES product, denoted in the following as CERES_standard, is an extraction along the CloudSat track of the single satellite footprint (SSF) retrieval for the top-of-atmosphere (TOA) fluxes and the clouds and radiative swath (CRS) retrieval for the clear-sky fluxes (Wielicki et al. 1996). The CERES footprint is about 20 km in diameter; therefore, several 2B-FLXHR-LIDAR profiles may correspond to the same CERES observation. In this case the CERES value is considered several times in the analysis.
Radiative heating profiles from the CloudSat 2B-FLXHR-LIDAR are used. This product is an upgraded version of the 2B-FLXHR (L’Ecuyer et al. 2008), and the improvements are fully described in Henderson et al. (2013). The heating rates are derived from the vertical distribution of liquid and ice cloud effective radii and water content retrieved from the CloudSat measurements. Additional thermodynamic profiles and surface properties are taken from the European Centre for Medium-Range Weather Forecasts analyses and the Geosphere–Biosphere Program, respectively. These parameters are then used as inputs to a broadband radiative flux model. The 2B-FLXHR-LIDAR takes advantage of improvements in cloud and precipitation products. Indeed, cloud layers (and associated properties) not detected by the radar are obtained from CALIPSO and MODIS data. Special processing is also applied when precipitation is detected or even when a total attenuation of the CloudSat profile is encountered. When a rain rate is obtained a partition is made between rainwater and cloud water content according to Lebsock and L’Ecuyer (2011). In circumstances where the rain rate is not retrieved (e.g., over land or when the CloudSat radar is totally attenuated), microphysical properties are specified for the precipitation layers using climatological values: 0.1 g m−3 for cloudy pixels and 0.15 g m−3 for pixels containing rain. In case of heavy precipitation (total attenuation of the radar) the pixels containing rain are set to 0.6 g m−3. Henderson et al. (2013) demonstrated that the microphysical assumptions made in the precipitation layers mainly impact the shortwave (SW) fluxes at the surfaces.
Because of the polar orbit of the A-Train, each detected and followed MCS of class 2a (Fiolleau and Roca 2013a) is only sampled once and therefore a substantial amount of data is needed to document the whole life cycle of the three subregions of the MCS. This is why the longest A-Train dataset (between 2006 and 2011) is considered to lead to the statistics summarized in Table 1. The first number in each cell corresponds to the number of individual MCSs for which at least one given subregion was sampled at a given time of its normalized life time. The second number is the corresponding number of profiles that can be used to build the composites. Since individual profiles obtained within a given MCS are not necessarily independent, in order to ensure a minimum representativeness of the composites the results are considered only when at least 20 MCSs have been sampled. Therefore, results for cells with bold numbers in Table 1 are not shown. For comparison, Cetrone and Houze (2009) document convective and stratiform/cirriform regions using the sampling of 120 and 82 MCSs over West Africa, 190 and 78 over the Maritime Continent, and 141 and 42 over the Bay of Bengal. The number of MCSs shown in Table 1 is of the same order of magnitude, and therefore our composites can be expected to be fully representative of the sampled geographical areas.
Another interesting question is the potential bias that may occur as the result of the fixed sampling time of the sun-synchronous orbit of the A-Train. Convection triggering is affected by a strong diurnal cycle, particularly over land. The fixed sampling time of the polar-orbiting satellite may therefore focus on a particular type of system, or all systems in a given stage of their life cycle may originate from a particular area. These points were investigated (not shown) and the main results can be summarized as follow:
From the geostationary satellite, more than 63%, 64%, and 56% of the class 2a (Fiolleau and Roca 2013a) tracked MCSs for AF, ATL, and OIO, respectively, have a life duration between 5 and 10 h. If the life duration distribution for class 2a MCSs sampled by the A-Train is built, they only represent 30%, 21%, and 33% respectively. This means that our composites are biased toward systems longer than 10 h.
A strong correlation exists between maximum area and life duration (Feng et al. 2012; Pope et al. 2008). Therefore, the bias on “long” systems translates into a relative bias on maximum area.
Over land, convection triggers in the afternoon between (1200 and 1800 UTC); therefore the beginning of the life cycle is sampled largely during the daytime; in contrast, the end of the life cycle is largely sampled during night time. If the diurnal cycle of convection exists over the ocean, there is a larger spread of initiation time. Therefore, the oceanic MCSs included in our statistics birth at different times throughout the day.
The geographical distribution of sampling of each life step was also examined, and it appears that they are homogeneously distributed over the considered geographical regions (not shown).
In summary, although the polar-orbiting satellites provide only a twice-daily sample of the tropics, the data accumulation over several years and the buildup of composites over the life cycle as tracked in geostationary images can provide useful information on the physical processes involved in each subregion of the system over the whole life cycle.
3. Evolution of the cloudiness vertical structure over MCS life cycle
The vertical structure of cloudiness associated with MCSs, and by extension of individual subregions, is a first-order parameter for shaping the vertical profile of latent heating (bottom heavy or top heavy) and its modulation by radiative heating. The radar–lidar combination (from CloudSat and CALIPSO) is particularly well suited for this purpose and the differences in macrophysical properties derived from the combination of both instruments with respect to the radar alone provide useful insights into variations in hydrometeor size and is a good proxy for the detection of vertical variation in the water content.
Vertical convective motion in the tropics injects large quantities of ice into the upper troposphere. Therefore, variations in convective intensity impact the amount of ice that can be found in the upper level as well as the physical and optical thickness of clouds, which in turn affects the radiative heating profile (Ackerman et al. 1988; Li and Schumacher 2011). This basic relationship between vertical velocity intensity and macrophysical cloud properties also needs to be documented with regard to the life cycle of the MCS subregions.
Yuan et al. (2011) investigated the vertical layering of MCS from the CloudSat radar and found that a cloud layer with a top higher than 10 km exists even in situations of multilayering. The distributions of cloud-top height displayed in their study show similar values for weakly raining and nonraining regions of MCSs and a distribution that peaks slightly higher for raining regions. Further, this study does not take the MCS life cycle into account. Figure 1 shows the evolution of the mean cloud-top height throughout the life cycle for the three MCS subregions and the three geographical areas. This diagnostic was compared to the probability density function of cloud top as in Yuan et al. (2011, their Fig. 2), and the mean value was found to correspond well to the peak value of the distribution. For instance, if the mean cloud-top height of the convective subregion deduced from the radar data (Fig. 1, solid line with squares) is compared to the raining probability density function (PDF) of Yuan et al. (2011) (at least at the beginning of the MCS life), the same variations are observed from one geographic area to the other: a mean value of about 15 km over Africa and about 14 km for oceanic regions. Cetrone and Houze (2009) found the same kind of behavior for continental and oceanic MCSs and attributed it to a greater potential buoyancy over the continent. However, in particular for AF, the mean cloud-top value strongly decreases after the middle of the life cycle (with a mean cloud top at 12 km at step 10), as it remains nearly constant for the other geographical areas. For stratiform and cirriform parts, the cloud-top height is similar from one geographical area to another and they all exhibit a loss of about 1 km in the second half of the life cycle. These results are not in line with Cetrone and Houze (2009), who found a lower radar cloud-top altitude for the anvils for their West Africa region compared to two other oceanic monsoon regions. They invoke tropopause height as an explanation of lower cloud top in West Africa. One possibility is that their signal arises rather from a mixing of systems at different life stages, since this dimension is not included in their study and because they accumulate a limited number of MCSs.
The operational sensitivity of CloudSat radar is about −30 dBZ (Stephens et al. 2002) and it is insufficient to sample clouds with low water content and low optical thickness. For a more accurate detection of cloud top, a mean cloud-top height was computed using a radar–lidar combination (dashed line in Fig. 1). The difference between the solid and dashed lines tends to increase faster in AF than in ATL or OIO. This suggests that large amounts of ice are lifted due to intense motion over land, but that the convective intensity decreases faster in the life cycle, leaving a deeper layer of smaller ice water content at the top of the anvil within this region. However, this deeper layer at the end of the life cycle may have some consequences on the radiation properties, and this point is investigated in section 5. Jensen and Del Genio (2003) already stressed that this region between the radar cloud top and the infrared equivalent top height is likely to be occupied by a population of small ice crystals, identified as an important contributor to the storm albedo. Even if the radar–lidar cloud-top height remains relatively constant over the life cycle, the ATL MCSs have a lower vertical (by about 500 m) extent with respect to OIO for all MCS subregions, in spite of similar cloud-top altitudes, as observed by radar between the two geographical areas.
In summary, the examination of the cloud-top height evolution highlights different intensities in the convective processes according to the various geographical areas, with a specific behavior over land. Some differences also exist among the oceanic areas in term of cloud-top altitude. The following section investigates the associated microphysical processes in more detail.
4. Evolution of MCS microphysics over the life cycle
The best way to characterize and identify microphysical processes at play in the MCS life cycle is through in situ measurements. Some experiments have been designed in this way: see, for instance, McFarquhar and Heymsfield (1996), Heymsfield et al. (2002), Bouniol et al. (2010), and Fontaine et al. (2014). However, for aircraft safety, the MCS sampling is generally confined to stratiform and cirriform parts. In addition, because of flight time limitations, it is very difficult to gather information about the entire life cycle. Ground-based radar, using classification algorithms and dynamical retrievals, is also powerful tool for gaining insight into microphysical processes (see, e.g., Hauser et al. 1988; Evaristo et al. 2010; Xu and Zipser 2015). However, this documentation does not provide a total view of the life cycle, and full documentation of the MCS (including nonprecipitating anvil) is in general beyond the scope of such experiments due to the radar frequency. The use of satellite data and of the aforementioned composite strategy offers valuable, albeit indirect, insights into microphysical processes, since radar reflectivity is an integrated measurement of microphysical properties.
At each altitude, a distribution of reflectivity values is built for each part of the system and step of the life cycle. Then, statistical parameters—mean and interquartile range—are computed. The interquartile range is preferred to the standard deviation because of the Mie effect, which tends to saturate the reflectivity values by about 20 dBZ at 94 GHz. For the same reason, the median value may be expected to differ from the mean. However, a systematic comparison of both parameters shows very close values (not shown).
Former studies have investigated the vertical distribution of reflectivity. Yuan et al. (2011) built joint altitude–reflectivity probability distribution functions normalized by the maximum frequency in any height–reflectivity bin (according to raining, weakly raining, and nonraining systems). Cetrone and Houze (2009) used the same kind of diagnostics, separating the distributions built from the TRMM profiles according to convective and stratiform precipitation, and from CloudSat with respect to the depth of the anvil. Based on the cloud-top height shown in Fig. 1, our results appear to be more comparable to their statistics for thick anvils (deeper than 6 km). None of these studies decomposed the statistics according to the life cycle; we can learn from Table 1 that the beginning and the end of the life cycle are undersampled compared to the middle of the life cycle. Therefore, convective cloud statistics built from A-Train data and the identification of related processes, without decomposition according to the step of the life cycle, are more representative of the occurrences at the middle of the MCS life cycle.
a. Convective regions
The mean reflectivity evolution (Fig. 2) for the convective part has different features according to geographical area. The largest and deepest values (larger than 10 dBZ) are found for AF MCSs (in the first third of the life cycle) and extend between heights of 6 and 13 km. In contrast, high values over a layer of that depth are not found for oceanic regions (even if they can be even more intense in the lower part of the cloud). For those systems, the reflectivity remains high throughout the life cycle over a relatively constant depth. These reflectivity values cannot be directly interpreted in terms of ice water content or liquid water content due to the Mie scattering at the CloudSat frequency, and the multiple scattering or attenuation. Several algorithms that take such processes into account exist in order to retrieve the water content. However, high reflectivity values can be viewed as an index of high water and/or ice content where attenuation is likely to occur even in the ice phase. By definition of the convective stratiform flag in CloudSat products, these profiles are precipitating, but the reflectivity value decreases strongly below the 0°C isotherm altitude toward the surface due to attenuation in rain at 94 GHz. This signature cannot be unequivocally attributed to evaporation of precipitation. These high values of reflectivity are also accompanied by a low value of interquartile range, which means that the reflectivity values are really confined around the mean value. Within this region, the precipitation-growth process is a collection of cloud water by precipitation particles (like riming) in the strong updraft cores (Yuter and Houze 1998). The associated ice particles are then dense hydrometeors, explaining the high value of reflectivity. Updrafts of oceanic storms are expected to be less vigorous than continental ones, leading to deeper systems over continent. Lower values of reflectivity can be attributed to smaller ice crystals for ATL and OIO (Jensen and Del Genio 2003). The evolution of the mean reflectivity profile (as well as the mean radar cloud top; left panel of Fig. 1 for continental MCS) thus suggests strong vertical updrafts at the beginning of the life cycle, with a rapid weakening in intensity.
Evolution of the mean profile of reflectivity (first row) and of the reflectivity interquartile range (second row) for (a) AF, (b) ATL, and (c) OIO. Solid upper line with squares and dashed line are the mean cloud-top height obtained from radar and radar–lidar, respectively. The lower line with squares is the mean cloud base.
Citation: Journal of Climate 29, 9; 10.1175/JCLI-D-15-0551.1
Evolution of the mean profile of reflectivity (first row) and of the reflectivity interquartile range (second row) for (a) AF, (b) ATL, and (c) OIO. Solid upper line with squares and dashed line are the mean cloud-top height obtained from radar and radar–lidar, respectively. The lower line with squares is the mean cloud base.
Citation: Journal of Climate 29, 9; 10.1175/JCLI-D-15-0551.1
Evolution of the mean profile of reflectivity (first row) and of the reflectivity interquartile range (second row) for (a) AF, (b) ATL, and (c) OIO. Solid upper line with squares and dashed line are the mean cloud-top height obtained from radar and radar–lidar, respectively. The lower line with squares is the mean cloud base.
Citation: Journal of Climate 29, 9; 10.1175/JCLI-D-15-0551.1
The value of the interquartile range increases toward the cloud top, corresponding to a larger range of reflectivity values (and therefore includes lower values) in agreement with the abovementioned existence of a layer of smaller size hydrometeors at the cloud top, but also smaller particles that are buoyant enough to be lifted to such altitudes.
b. Stratiform regions
Reflectivity in the stratiform subregions is about 5 dBZ smaller than in convective regions. A faster decrease in reflectivity can also be observed as the altitude increases. This decrease in reflectivity likely reflects the settling of larger particles (Mather et al. 2007) that are able to aggregate and hence form low-density but large hydrometeors (Yuan et al. 2011). A particular signature exists, however, in the interquartile range (particularly marked for AF) at the beginning of the life cycle, with very high values of this parameter. Examination of the individual distribution for these early stages of the life cycle (not shown) reveals a bimodal distribution in reflectivity values (persistence of a high value mode) at altitudes above 8 km. These high values in the stratiform anvil can be attributed to detrainment from the convective towers and hence advection toward the anvil. Indeed, larger numbers of ice particles generated by dynamical processes in the convective updrafts are detrained laterally into the anvil (raining and nonraining) (Yuan et al. 2011). Existence of this double mode of reflectivity distribution is in agreement with the in situ observations of Bouniol et al. (2010), who found rimed particles in the stratiform anvil of West Africa MCSs.
c. Cirriform regions
Within the cirriform anvil, the mesoscale updrafts are an order of magnitude smaller than that observed in the convective core. The major growth process is then through vapor deposition on ice particles (Yuter and Houze 1998), which is a slow process. The decrease of reflectivity with altitude (Fig. 2) also appears within this part; however, reflectivity also tends to decrease toward the cloud base (with an increase in the interquartile range), suggesting that sublimation or evaporation of the larger particle in the subsaturated environment occurs near the anvil base. However, this evaporation process, combined with radiative processes, may feed the mesoscale circulation in the anvil through stabilization or destabilization processes that contribute to its maintenance (Webster and Stephens 1980; Houze 1982).
In summary, the major differences in microphysical properties are linked to the land–ocean contrast with deeper systems over land due to intense convective motions. Dense hydrometeors are detrained toward the stratiform region, and a signature is found up to the cirriform region. However, this signature exists mainly in the first half of the life cycle. In contrast, oceanic MCSs have a more constant behavior throughout their life cycle. As emphasized by Yuan et al. (2011) these differences in microphysical properties of clouds may have direct consequences on the radiative properties of the MCS, which are investigated in the next section.
5. Radiative property evolution over the life cycle
As stressed early on by Leary and Houze (1979), convective processes have a direct influence on the microphysical processes that can take place in the life cycle. These processes determine which types of hydrometeors are present in each part of the MCS, as well as their sizes and concentrations. These microphysical properties are then crucial in determining the radiative properties of a MCS.
a. Top of the atmosphere
Figure 3 compares the evolution of the albedo and the OLR at the TOA as a function of life step. For simplicity, only the 2B-FLXHR-LIDAR and CERES_standard products are shown. The other products MODIS or MODIS enhanced, also present in the CCCM dataset, give results very close to one another and generally between CERES_standard and 2B-FLXHR-LIDAR. The 2B-FLXHR-LIDAR seems to systematically underestimate the OLR and the albedo of the cirriform part with respect to CERES_standard and overestimates it for convective part. Henderson et al. (2013) quantified the bias of 2B-FLXHR-LIDAR with respect to CERES data at global scale and in all-sky conditions and found a −4.9 W m−2 for OLR and 4.1 W m−2 for the outgoing shortwave radiation. These biases are explained mainly by the errors in surface temperature and changes in specific humidity below the 500-hPa level for OLR and the CloudSat liquid water content estimate for shortwave radiation.
Evolution of albedo and OLR at TOA for (a) AF, (b) ATL, and (c) OIO. Each color represents part of the system; different symbols are for different products (see legend).
Citation: Journal of Climate 29, 9; 10.1175/JCLI-D-15-0551.1
Evolution of albedo and OLR at TOA for (a) AF, (b) ATL, and (c) OIO. Each color represents part of the system; different symbols are for different products (see legend).
Citation: Journal of Climate 29, 9; 10.1175/JCLI-D-15-0551.1
Evolution of albedo and OLR at TOA for (a) AF, (b) ATL, and (c) OIO. Each color represents part of the system; different symbols are for different products (see legend).
Citation: Journal of Climate 29, 9; 10.1175/JCLI-D-15-0551.1
In terms of albedo, convective and stratiform regions have values of the same order of magnitude (between 0.6 and 0.65) for all geographical areas (convective values being slightly larger). Larger values can be observed at the beginning of the life cycle, along with a decrease in magnitude after step 5, which is more pronounced for the AF area.
The evolution of the albedo of the cirriform part differs according to geographical area. A substantial decrease, from about 0.55 to about 0.42, is observed in AF, while a weak decrease, or nearly constant values, is found for the oceanic areas (between 0.45 and 0.5). This decrease supports the idea of a thinning (at least in the optical sense) of the AF nonprecipitating anvils, while at the end of the life cycle the cloud is no longer fed by detrainment, and it is mainly the anvil’s own dynamical processes that maintain the cloud, involving less dense hydrometeors.
In terms of OLR, convective and stratiform regions have values of the same order of magnitude, but values for cirriform regions are about 10 W m−2 higher (particularly over the ocean). The variations over the life cycle are of the order of 30 W m−2 for AF and 20 W m−2 for OIO and even weaker for ATL, and the minimum value is observed around step 3 for ATL and step 5 for the two others.
Futyan and DelGenio (2007) proposed life cycle composites of the same parameters, but without discriminating the internal parts of the MCS. Their results (which can be interpreted as a weighted combination of our composites) give different orders of magnitude, with a minimum in MCS averaged OLR of 170 W m−2 for continental MCS and 182 W m−2 for oceanic MCS and the amplitudes of the life cycle are 45 and 28 W m−2, respectively. Our disagreement on absolute value arises from the way in which the infrared images (converted to synthetic GERB data in their cases) are segmented up to an equivalent infrared temperature of 254 K, as segmentation stopped at 235 K in our case. This higher value of segmentation likely includes higher OLR values. However, we have in common the amplitude of variation of the OLR throughout the life cycle. The separation within the internal part is of importance in guiding parameterization development. Indeed, clouds associated with MCS can, depending on the model, be represented partly by the convective scheme or the large-scale scheme. The combination of cloud properties resulting from both schemes should be in the observed range in order to represent the impact of MCS at larger scale correctly. For the same thresholding reason, their albedo magnitudes are smaller and a variation in amplitude of about 0.1 throughout the life cycle is observed in their composites in agreement with what is shown in Fig. 3. It would be interesting to go further in the comparison of our results with their study by recombining our estimates with the fractional size of each MCS subpart. Futyan and Del Genio (2007) found a lower mean albedo and larger OLR for oceanic systems compared to land systems. These results do not seem obvious from our results and the recombination of the subregion values may help in understanding the origin of these differences. They propose that the more stratiform nature of oceanic convection explains the land–ocean differences. Such a recombination is beyond the scope of the present paper and should be considered in future work.
b. Cloud radiative forcing at the top of the atmosphere
Previous studies by Ramanathan et al. (1989), Harrison et al. (1990), and Cess et al. (2001) showed from observations that in the convective areas, CRFLW and CRFSW tend to balance within ±10 W m−2. This near cancellation between CRFLW and CRFSW results from an average of forcing from different cloud types ranging from high thick clouds to thinner anvils. Kiehl and Ramanathan (1990) and Hartmann et al. (2001) proposed a feedback mechanism based on the energy balance between convective and neighboring nonconvective regions that implies that if the CRFnet in the convective region is positive, the corresponding net fluxes increase up to be balanced by that of the environment. This is posited to result in changes in the strength of the overturning circulation between these regions, which in turn alters the cloud properties driving the fluxes associated with the convective area to that of the nonconvective area. Kiehl (1994) proposed that the cancellation of CRFLW and CRFSW is a generic property of all tropical convective regions. However, Rajeevan and Srinivasan (2000) and Thampi and Roca (2014) showed that in particular times and places, like in the Bay of Bengal, or more generally in the Asian monsoon region, this near-cancellation no longer occurs. They link this imbalance to a combination of a large amount of high clouds and high clouds with large optical depth found only in the Asian monsoon, and also to the significant water vapor loading of the atmosphere.
More generally, previous studies have examined this forcing balance at the regional and monthly scales. Futyan et al. (2004) emphasized that the Hartmann et al. (2001) process is expected to act at the time and space scales at which the individual convective systems develop and that this near-cancellation should occur for the spatially and temporally averaged radiative effects of individual convective systems over their lifetime. The diurnal cycle is a strong contributor to this balance (since CRFSW vanishes during night time but CRFLW remains positive) and therefore a cancellation at the scale of the lifetime requires a balance of the forcing between day and night times in addition to convection occurrence within the diurnal cycle.
To investigate the possibility of cancellation between CRFSW and CRFLW, these parameters are shown in Fig. 4, along with their sum. One should keep in mind that this data is systematically collected at 0130 and 1330 LT, and therefore the CRFSW corresponds to an instantaneous value. Because of the small differences between daytime and nighttime values of CRFLW, its values have been averaged and may be interpreted as a daily mean for each individual subpart of the MCS.
Evolution of the CRF at the TOA in (left) the SW (daytime only), (center) the LW, and (right) total for (a) AF, (b) ATL, and (c) OIO. CRFnet is split according to day (solid) and night (dashed). Each color represents part of the system; different symbols are for different products (see legend).
Citation: Journal of Climate 29, 9; 10.1175/JCLI-D-15-0551.1
Evolution of the CRF at the TOA in (left) the SW (daytime only), (center) the LW, and (right) total for (a) AF, (b) ATL, and (c) OIO. CRFnet is split according to day (solid) and night (dashed). Each color represents part of the system; different symbols are for different products (see legend).
Citation: Journal of Climate 29, 9; 10.1175/JCLI-D-15-0551.1
Evolution of the CRF at the TOA in (left) the SW (daytime only), (center) the LW, and (right) total for (a) AF, (b) ATL, and (c) OIO. CRFnet is split according to day (solid) and night (dashed). Each color represents part of the system; different symbols are for different products (see legend).
Citation: Journal of Climate 29, 9; 10.1175/JCLI-D-15-0551.1
CRFSW is highly negative and is strongly controlled by the cloud albedo. Therefore, as in the previous section, values for convective and stratiform regions are of the same order of magnitude. The life cycle appears more pronounced for continental MCSs, and the minimum values are larger for these MCSs. This point may be attributed to the larger surface albedo of the continental surface compared to the ocean (Nowicki and Merchant 2004). Higher values are also observed for the cirriform part in AF, and the CRFSW evolution correlates well with the evolution of the albedo. For the oceanic areas, in contrast, a nearly constant CRFSW of −450 W m−2 is obtained.
The value of CRFLW rarely exceeds 170 W m−2 and larger values are found in convective subregions of AF. The other subregions have smaller forcing values (about 130 W m−2 for the cirriform subregion). Based on these values, a given number of hours with low insolation are needed in order to compensate for the high amount of energy accumulated during daytime. This point is illustrated by the third panel, which shows the CRFnet separately for day and night and a ratio of about 3:1 exists between day and night time. This suggests that the SW–longwave (LW) cancellation is very sensitive to the time at which the different life steps occur as well as the fractional area of each MCS subregion for a particular step (Nowicki and Merchant 2004). An accurate estimation of the CRFSW and CRFLW balance over the MCS life cycle should be considered in future work; however, it is worth noting the strong dependency on the diurnal cycle implied by such a balance, which points toward the difficulty of its representation within models. Our results show that, because of the amount of energy involved, the convective cloud radiative feedback would be hard to represent without a good timing of the convection triggering within the model.
c. Radiative heating profiles
Clouds interact with the dynamical circulation through their associated heating profile. Li et al. (2013) show that the radiative heating is an order of magnitude smaller than the latent heating at the regional scale. However, nonprecipitating anvil clouds associated with MCS have a lifespan far longer than the intense convective processes and therefore cannot be neglected even at the regional scale. Radiative heating within the clouds is also a key factor in explaining the longevity of the clouds through stabilizing or destabilizing processes within the anvil (Webster and Stephens 1980; Jensen and Del Genio 2003) that models should be able to simulate in order to incorporate a realistic feedback of the MCS anvils. Muller and Bony (2015) also demonstrate how the contrast of the MCS radiative heating with respect to its dry environment explains the self-aggregation of convection. In addition, MCSs at different stages of their life cycles may coexist at the regional scale, and a nonhomogenous distribution of this heating may be expected at this scale.
Figure 5 shows the evolution of the radiative heating profiles in the shortwave and longwave for our three areas of interest, as well as the anomaly with respect to clear sky in superimposed contours. This anomaly corresponds to radiative forcing as defined in Li et al. (2013) or Haynes et al. (2013). The shortwave heating is always positive due to the absorption of SW radiation and leads to warming throughout the cloud. The impact of optically thick clouds is to reduce the insolation reaching the lower levels, explaining the negative anomaly (dotted line in Fig. 5) below heights of 6 km (Hartmann et al. 2001; Haynes et al. 2013). In the SW, the more noticeable pattern of the convective subregion is the strong heating just below the cloud top of the MCS, as seen by the radar (solid line with squares in Fig. 5). This heating decreases in intensity over the life cycle. The detrainment of dense hydrometeor in the stratiform anvil of AF MCS observed up to the middle of the life cycle implies a heating that exceeds the value of clear sky in the upper portion of the anvil by more than 10 K day−1. In the cirriform anvil, the heating remains higher at the beginning of the life cycle and then decreases. Higher values are observed for AF MCSs. However, while different magnitudes are observed for a given MCS subregion from one geographical area to another, the tendency throughout the life cycle remains identical. The SW radiative heating at the scale of the whole MCS is a combination of these three heating profiles, with more weight given to the cirriform subregion due its size. In addition, the triggering time also has a strong influence on the budget at the MCS scale since no contributions are expected during night time. Figure 5 also shows that it is the uppermost part (the upper third) of the high ice content clouds that contributes the most.
Evolution of the radiative heating (W m−2) in the SW (first row) and in the LW (second row) for (a) AF, (b) ATL, and (c) OIO over the MCS life cycle in color with superimposed the anomaly with respect to clear-sky radiative heating [dashed (dotted) contours for positive (negative) values]. The lines are as in Fig. 2.
Citation: Journal of Climate 29, 9; 10.1175/JCLI-D-15-0551.1
Evolution of the radiative heating (W m−2) in the SW (first row) and in the LW (second row) for (a) AF, (b) ATL, and (c) OIO over the MCS life cycle in color with superimposed the anomaly with respect to clear-sky radiative heating [dashed (dotted) contours for positive (negative) values]. The lines are as in Fig. 2.
Citation: Journal of Climate 29, 9; 10.1175/JCLI-D-15-0551.1
Evolution of the radiative heating (W m−2) in the SW (first row) and in the LW (second row) for (a) AF, (b) ATL, and (c) OIO over the MCS life cycle in color with superimposed the anomaly with respect to clear-sky radiative heating [dashed (dotted) contours for positive (negative) values]. The lines are as in Fig. 2.
Citation: Journal of Climate 29, 9; 10.1175/JCLI-D-15-0551.1
The LW heating does not vary significantly over the life cycle, and the heating anomaly is moderate (maximum anomaly of about ±3–4 W m−2) compared to the values observed in the SW. A dipole of cooling near the cloud top and heating at the base of the ice cloud is found in the composites for all subregions (just above the melting level in convective and stratiform subregions). This dipole corresponds well to the signature of optically thick clouds, as highlighted by Mather et al. (2007) and Webster and Stephens (1980). The heating anomaly with respect to clear sky is systematically thicker for cirriform regions. Muller and Bony (2015) show that the combination of this midlevel warming with radiative cooling in the low-level cooling of clear-sky regions is necessary to maintain aggregation.
It is the cooling at the cloud top that is the most influenced by the life cycle. Indeed, its magnitude and thickness grow with the life step to reach lower values at the end of the life cycle in the cirriform region. The minimum value is located at the mean cloud-top height as observed by the radar, and becomes null at the cloud top height as seen by the lidar (dashed line). As noted by Webster and Stephens (1980), Mather et al. (2007), Powell et al. (2012), and Jensen and Del Genio (2003), heating at the bottom of the cloud and cooling in the upper portion may fuel the internal circulations of the anvil and mixing with surrounding clear air, thus destabilizing the anvil cloud top and base while stabilizing the interior of the anvil.
Powell et al. (2012) stated that during daytime, the SW component dominates the radiative budget, leading to warming throughout the cloud. Taking the life cycle into account leads to a more complicated view. Indeed, if the SW and LW components shown in Fig. 5 are summed, the convective and stratiform subregions remain dominated by the SW heating. However, the LW contribution dominates at the cloud top (between 11 and 15.5 km height for AF and between 12.5 and 15.5 km height for the oceanic region) for the second half of the life cycle. 1300 LT is close to the maximum of insolation, and the cooling at the cloud top is therefore even larger at other times of day.
6. Conclusions
The evolution of physical properties over the MCS life cycle was investigated over three different geographical areas, one continental (AF) and two oceanic (ATL and OIO), using the complementarity of geostationary and orbiting satellites. Because of intrinsic differences in physical processes, the MCSs are divided into three internal subregions: deep convective, stratiform anvil, and cirriform anvil. The life steps (normalized between 1 and 10) are derived from the output of a tracking algorithm applied to infrared images. The physical properties are obtained individually for each subregion of the MCS when the A-Train orbits intersect a MCS trajectory. The evolution of each individual subregion is then followed throughout the entire MCS life cycle. Several physical properties are documented in the present study, from macrophysical properties to radiative heating profiles.
Macrophysical properties are expected to directly influence the radiative properties (Mather et al. 2007). Differences in cloud-top altitude as seen by the radar and the lidar indicate the highest altitude where larger hydrometeors are found within the MCSs. The cloud top as observed by the lidar does not have such a marked life cycle and does not vary according to the MCS subregions. The cloud top, as seen by the radar, has a different altitude according to the MCS subregions, which implies a deeper layer of small crystals at the top of the nonprecipitating anvil than in the stratiform part. For stratiform and cirriform subregions, the evolution is similar for all geographical areas; however, in the case of the convective part over land, the cloud top is higher but decreases strongly (loss of about 3 km) after the middle of the life cycle. This evolution of the convective subpart is in accordance with more intense convective updrafts over the continent with respect to the ocean (Zipser et al. 2006).
Examination of the radar reflectivity composites highlights some microphysical properties and underlying processes. First, a decrease in the magnitude of reflectivity is observed from the convective subregion up to the cirriform anvil. This observation supports the idea of a less dense and smaller hydrometeor in the anvil and corroborates in situ observations (Bouniol et al. 2010). The more striking feature derives from the differences in life cycle when comparing the continental MCSs with the oceanic. Strong updrafts at the beginning of the life cycle are able to generate large hydrometeors that are lifted up to a high altitude and even detrained to the stratiform subregion, and a signature is also found in the nonprecipitating anvil. However, the convective intensity weakens after the first half of the life cycle. In contrast, for oceanic MCSs, reflectivity is not as high in altitude, but its magnitude remains in the same range throughout the life cycle. The signature of detrainment from the convective towers is found only in the stratiform region.
Associated radiative properties and cloud forcing are well documented. Convective and stratiform subregions have relatively similar albedo values that do not evolve markedly over the life cycle. The main differences are seen in the cirriform subregion for the continental MCSs with an albedo that is higher (with respect to oceanic MCS) at the beginning of the life cycle and then reaches the value observed over the ocean. The question of the respective roles of small and large hydrometeors in shaping the albedo remains open (Jensen and Del Genio 2003). Our results suggest that the large and dense hydrometeors tend to control the evolution of this parameter in agreement with McFarquhar and Heymsfield (1997). Indeed, the deepening of the small crystal layer in the cirriform subregion combined with the decrease in the intensity of reflectivity is accompanied by a decrease in albedo. The OLR evolution for continental MCSs is also strongly shaped by the evolution of the large hydrometeors. Lower values are found in continental, versus oceanic, MCSs in the first half of the life cycle. This marked life cycle in the LW for continental MCSs transfers to the associated forcing. In contrast, oceanic MCSs have a LW contribution of smaller amplitude. The importance of separating the nonprecipitating anvil from stratiform and convective subregions appears crucial in the SW CRF. Indeed, the contribution of the cirriform anvil can be twice as small as the others.
Finally, the radiative heating is examined and the impact of dense hydrometeors detrained from the convective towers leads to more intense heating in the SW, particularly at the beginning of the life cycle for continental MCS. A dipole of cooling near the cloud top and heating near the cloud base is found in the LW. This cooling intensifies for all MCS subregions and all geographical areas near the end of the life cycle, but with different intensity.
It appears from the preceding analysis that the convective intensity strongly constrains the evolution of the cloud parameters. By analyzing the thicker part of the anvils using TRMM data, Li and Schumacher (2011) reach a similar conclusion, namely that the diversity of anvil properties is linked to factors that affect the parent convection. The time of triggering is another important feature for understanding how the balances are built at the scale of the MCS. While some physical parameters, in particular in the LW, are relatively insensitive to the diurnal cycle, this is not the case for SW parameters.
Acknowledgments
Jérémy Aublanc and Bastien Sauvage are acknowledged for their help in data processing. We also thank Catherine Rio for our scientific discussions. CloudSat data are obtained from CIRA of Colorado State University. ICARE and NASA provided access to the CALIOP data. We also acknowledge the Atmospheric Sciences Data Center of Langley Research Center for access to the CCCM dataset and S. Kato for his assistance in its processing. This work was partially supported by CNES.
REFERENCES
Ackerman, T. P., K.-N. Liou, F. P. J. Valero, and L. Pfister, 1988: Heating rate in tropical anvils. J. Atmos. Sci., 45, 1606–1623, doi:10.1175/1520-0469(1988)045<1606:HRITA>2.0.CO;2.
Awaka, J., 1998: Algorithm 2A23-rain type classification. Proc. Symp. on the Precipitation Observation from Non-Sun Synchronous Orbit, Nagoya, Japan, 215–220.
Behrangi, A., T. Kubar, and B. Lambrigtsen, 2012: Phenomenological description of tropical clouds using CloudSat cloud classification. Mon. Wea. Rev., 140, 3235–3249, doi:10.1175/MWR-D-11-00247.1.
Bouniol, D., J. Delanoë, C. Duroure, A. Protat, V. Giraud, and G. Penide, 2010: Microphysical characterisation of West African MCS anvils. Quart. J. Roy. Meteor. Soc., 136 (Suppl.), 323–344, doi:10.1002/qj.557.
Cess, R. D., M. Zhang, B. A. Wielicki, D. F. Young, X.-L. Zhou, and Y. Nikitenko, 2001: The influence of the 1998 El Niño upon cloud-radiative forcing over the Pacific warm pool. J. Climate, 14, 2129–2137, doi:10.1175/1520-0442(2001)014<2129:TIOTEN>2.0.CO;2.
Cetrone, J., and R. A. Houze Jr., 2009: Anvil clouds of tropical mesoscale convective systems in monsoon regions. Quart. J. Roy. Meteor. Soc., 135, 305–317, doi:10.1002/qj.389.
Cetrone, J., and R. A. Houze Jr., 2011: Leading and trailing anvil clouds of West African squall lines. J. Atmos. Sci., 68, 1114–1123, doi:10.1175/2011JAS3580.1.
Del Genio, A. D., and W. Kovari, 2002: Climatic properties of tropical precipitating convection under varying environmental conditions. J. Climate, 15, 2597–2615, doi:10.1175/1520-0442(2002)015<2597:CPOTPC>2.0.CO;2.
Donner, L. J., C. J. Seman, R. S. Hemler, and S. Fan, 2001: A cumulus parameterization including mass fluxes, vertical momentum dynamics, and mesoscale effects: Thermodynamic and hydrological aspects in a general circulation model. J. Climate, 14, 3444–3463, doi:10.1175/1520-0442(2001)014<3444:ACPIMF>2.0.CO;2.
Evaristo, R., G. Scialom, N. Viltard, and Y. Lemaître, 2010: Polarimetric signature and hydrometeor classification of West Africa squall lines. Quart. J. Roy. Meteor. Soc., 136 (Suppl.), 272–288, doi:10.1002/qj.561.
Feng, Z., X. Dong, B. Xi, C. Schumacher, P. Minnis, and M. Khaiyer, 2011: Top-of-atmosphere radiation budget of convective core/stratiform rain and anvil clouds from deep convective systems. J. Geophys. Res., 116, D23202, doi:10.1029/2011JD016451.
Feng, Z., X. Dong, B. Xi, S. A. McFarlane, A. Kennedy, B. Lin, and P. Minnis, 2012: Life cycle of midlatitude deep convective systems in a Lagrangian framework. J. Geophys. Res., 117, D23201, doi:10.1029/2012JD018362.
Fiolleau, T., and R. Roca, 2013a: Composite life cycle of tropical mesoscale convective systems from geostationary and low Earth orbit satellite observations: Method and sampling considerations. Quart. J. Roy. Meteor. Soc., 139, 941–953, doi:10.1002/qj.2174.
Fiolleau, T., and R. Roca, 2013b: An algorithm for the detection and tracking of tropical mesoscale convective systems using infrared images from geostationary satellite. IEEE Trans. Geosci. Remote Sens., 51, 4302–4315, doi:10.1109/TGRS.2012.2227762.
Fontaine, E., A. Schwarzenboeck, J. Delanoë, W. Wobrock, D. Leroy, R. Dupuy, and A. Protat, 2014: Constraining mass–diameter relations from hydrometeor images and cloud radar reflectivities in tropical continental and oceanic convective anvils. Atmos. Chem. Phys., 14, 11 367–11 392, doi:10.5194/acp-14-11367-2014.
Futyan, J. M., and A. D. Del Genio, 2007: Deep convective system evolution over Africa and the tropical Atlantic. J. Climate, 20, 5041–5060, doi:10.1175/JCLI4297.1.
Futyan, J. M., J. E. Russell, and J. E. Harries, 2004: Cloud radiative forcing in Pacific, African, and Atlantic tropical convective regions. J. Climate, 17, 3192–3202, doi:10.1175/1520-0442(2004)017<3192:CRFIPA>2.0.CO;2.
Harrison, E. F., P. Minnis, B. R. Barkstrom, V. Ramanathan, E. D. Cess, and G. G. Gibson, 1990: Seasonal variation of cloud radiative forcing derived from the Earth Radiation Budget Experiment. J. Geophys. Res., 95, 18 687–18 703, doi:10.1029/JD095iD11p18687.
Hartmann, D., L. A. Moy, and Q. Fu, 2001: Tropical convection and the energy balance at the top of the atmosphere. J. Climate, 14, 4495–4511, doi:10.1175/1520-0442(2001)014<4495:TCATEB>2.0.CO;2.
Hauser, D., F. Roux, and P. Amayenc, 1988: Comparison of two methods for the retrieval of thermodynamic and microphysical variables from Doppler radar measurements: Application to the case of a tropical squall line. J. Atmos. Sci., 45, 1285–1303, doi:10.1175/1520-0469(1988)045<1285:COTMFT>2.0.CO;2.
Haynes, J. M., T. S. L’Ecuyer, G. L. Stephens, S. D. Miller, C. Mitrescu, N. B. Wood, and S. Tanelli, 2009: Rainfall retrieval over the ocean with spaceborne W-band radar. J. Geophys. Res., 114, D00A22, doi:10.1029/2008JD009973.
Haynes, J. M., T. H. Vonder Haar, T. L’Ecuyer, and D. Henderson, 2013: Radiative heating characteristics of Earth’s cloudy atmosphere from vertically resolved active sensors. Geophys. Res. Lett., 40, 624–630, doi:10.1002/grl.50145.
Henderson, D. S., T. L’Ecuyer, G. Stephens, P. Partain, and M. Sekiguchi, 2013: A multisensor perspective on the radiative impacts of clouds and aerosols. J. Appl. Meteor. Climatol., 52, 853–871, doi:10.1175/JAMC-D-12-025.1.
Heymsfield, A. J., and G. M. McFarqhar, 1996: High albedos of cirrus in the tropical Pacific warm pool: Microphysical interpretations from CEPEX and from Kwajalein, Marshall Islands. J. Atmos. Sci., 53, 2424–2451, doi:10.1175/1520-0469(1996)053<2424:HAOCIT>2.0.CO;2.
Heymsfield, A. J., A. Bansemer, P. R. Field, S. L. Durden, J. L. Stith, J. E. Dye, W. Hall, and C. A. Grainger, 2002: Observations and parameterizations of particle size distributions in deep tropical cirrus and stratiform precipitating clouds: Results from in situ observations in TRMM field campaigns. J. Atmos. Sci., 59, 3457–3491, doi:10.1175/1520-0469(2002)059<3457:OAPOPS>2.0.CO;2.
Houze, R. A., Jr., 1982: Cloud clusters and large-scale vertical motions in the tropics. J. Meteor. Soc. Japan, 60, 396–410.
Houze, R. A., Jr., 2004: Mesoscale convective systems. Rev. Geophys., 42, RG4003, doi:10.1029/2004RG000150.
Houze, R. A., Jr., and A. K. Betts, 1981: Convection in GATE. Rev. Geophys., 19, 541–576, doi:10.1029/RG019i004p00541.
Houze, R. A., Jr., K. L. Rasmussen, M. D. Zuluaga, and S. R. Brodzik, 2015: The variable nature of convection in the tropics and subtropics. A legacy of 16 years of the Tropical Rainfall Measuring Mission satellite. Rev. Geophys., 53, 994–1021, doi:10.1002/2015RG000488.
Jensen, M. P., and A. D. Del Genio, 2003: Radiative and microphysical characteristics of deep convective systems in the tropical western Pacific. J. Appl. Meteor., 42, 1234–1254, doi:10.1175/1520-0450(2003)042<1234:RAMCOD>2.0.CO;2.
Kato, S., and Coauthors, 2011: Improvements of top‐of‐atmosphere and surface irradiance computations with CALIPSO‐, CloudSat‐, and MODIS‐derived cloud and aerosol properties. J. Geophys. Res., 116, D19209, doi:10.1029/2011JD016050.
Kiehl, J. T., 1994: On the observed near cancellation between longwave and shortwave cloud forcing in tropical regions. J. Climate, 7, 559–565, doi:10.1175/1520-0442(1994)007<0559:OTONCB>2.0.CO;2.
Kiehl, J. T., and V. Ramanathan, 1990: Comparison of cloud forcing derived from the Earth Radiation Budget Experiment with that simulated by the NCAR Community Climate Model. J. Geophys. Res., 95, 11 679–11 698, doi:10.1029/JD095iD08p11679.
Kondo, Y., A. Higuchi, and K. Nakamura, 2006: Small-scale cloud activity over the Maritime Continent and the western Pacific as revealed by satellite data. Mon. Wea. Rev., 134, 1581–1599, doi:10.1175/MWR3132.1.
Leary, C. A., and R. A. Houze, Jr., 1979: Melting and evaporation of hydrometeors in precipitation from the anvil clouds of deep tropical convection. J. Atmos. Sci., 36, 669–679, doi:10.1175/1520-0469(1979)036<0669:MAEOHI>2.0.CO;2.
Lebsock, M. D., and T. S. L’Ecuyer, 2011: The retrieval of warm rain from CloudSat. J. Geophys. Res., 116, D20209, doi:10.1029/2011JD016076.
L’Ecuyer, T. S., N. B. Wood, T. Haladay, G. L. Stephens, and P. W. Stackhouse, Jr., 2008: Impact of clouds on atmospheric heating based on the R04 CloudSat fluxes and heating rates data set. J. Geophys. Res., 113, D00A15, doi:10.1029/2008JD009951.
Li, W., and C. Schumacher, 2011: Thick anvils as viewed by the TRMM Precipitation Radar. J. Climate, 24, 1718–1735, doi:10.1175/2010JCLI3793.1.
Li, W., C. Schumacher, and S. A. McFarlane, 2013: Radiative heating of the ISCCP upper level cloud regimes and its impact on the large-scale tropical circulation. J. Geophys. Res. Atmos., 118, 592–604, doi:10.1002/jgrd.50114.
Mace, G. G., and Q. Zhang, 2014: The CloudSat radar–lidar geometrical profile product (RL-GeoProf): Updates, improvements, and selected results. J. Geophys. Res. Atmos., 119, 9441–9462, doi:10.1002/2013JD021374.
Machado, L. A. T., W. B. Rossow, R. L. Guedes, and A. W. Walker, 1998: Life cycle variations of mesoscale convective systems over the Americas. Mon. Wea. Rev., 126, 1630–1654, doi:10.1175/1520-0493(1998)126<1630:LCVOMC>2.0.CO;2.
Mapes, B. E., and R. A. Houze Jr., 1993: Cloud clusters and superclusters over the oceanic warm pool. Mon. Wea. Rev., 121, 1398–1415, doi:10.1175/1520-0493(1993)121<1398:CCASOT>2.0.CO;2.
Marchand, R. G., G. Mace, T. Ackerman, and G. Stephens, 2008: Hydrometeor detection using CloudSat—An Earth orbiting 94-GHz cloud radar. J. Atmos. Oceanic Technol., 25, 519–533, doi:10.1175/2007JTECHA1006.1.
Mather, J. H., S. A. McFarlane, M. A. Miller, and K. L. Johnson, 2007: Cloud properties and associated radiative heating rates in the tropical western Pacific. J. Geophys. Res., 112, D05201, doi:10.1029/2006JD007555.
Mathon, V., and H. Laurent, 2001: The life cycle of Sahelian mesoscale convective cloud systems. Quart. J. Roy. Meteor. Soc., 127, 377–406, doi:10.1002/qj.49712757208.
McFarquhar, G., and A. Heymsfield, 1996: Microphysical characteristics of three anvils sampled during the Central Equatorial Pacific Experiment. J. Atmos. Sci., 53, 2401–2423, doi:10.1175/1520-0469(1996)053<2401:MCOTAS>2.0.CO;2.
McFarquhar, G., and A. Heymsfield, 1997: Parameterization of tropical cirrus ice crystal size distributions and implications for radiative transfer: Results from CEPEX. J. Atmos. Sci., 54, 2187–2200, doi:10.1175/1520-0469(1997)054<2187:POTCIC>2.0.CO;2.
Mohr, K. I., and E. J. Zipser, 1996: Defining mesoscale convective system by their 85-GHz ice-scattering signatures. Bull. Amer. Soc., 77, 1179–1189, doi:10.1175/1520-0477(1996)077<1179:DMCSBT>2.0.CO;2.
Muller, C., and S. Bony, 2015: What favors convective aggregation, and why? Geophys. Res. Lett., 42, 5626–5634, doi:10.1002/2015GL064260.
Nesbitt, S. W., E. J. Zipser, and D. J. Cecil, 2000: A census of precipitation features in the tropics using TRMM: Radar, ice scattering, and lightning observations. J. Climate, 13, 4087–4106, doi:10.1175/1520-0442(2000)013<4087:ACOPFI>2.0.CO;2.
Nowicki, S. M., and C. J. Merchant, 2004: Observations of diurnal and spatial variability of radiative forcing by equatorial deep convective clouds. J. Geophys. Res., 109, D11202, doi:10.1029/2003JD004176.
Pope, M., C. Jakob, and M. J. Reeder, 2008: Convective systems of the north Australian monsoon. J. Climate, 21, 5091–5112, doi:10.1175/2008JCLI2304.1.
Powell, S. W., R. A. Houze Jr., A. Kumar, and S. A. McFarlane, 2012: Comparison of simulated and observed continental tropical anvil clouds and their radiative heating profiles. J. Atmos. Sci., 69, 2662–2681, doi:10.1175/JAS-D-11-0251.1.
Rajeevan, M., and J. Srinivasan, 2000: Net cloud radiative forcing at the top of the atmosphere in the Asian monsoon region. J. Climate, 13, 650–657, doi:10.1175/1520-0442(2000)013<0650:NCRFAT>2.0.CO;2.
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, doi:10.1126/science.243.4887.57.
Roca, R., and V. Ramanathan, 2000: Scale dependence of monsoonal convective systems over the Indian Ocean. J. Climate, 13, 1286–1298, doi:10.1175/1520-0442(2000)013<1286:SDOMCS>2.0.CO;2.
Roca, R., J. Aublanc, P. Chambon, T. Fiolleau, and N. Viltard, 2014: Robust observational quantification of the contribution of mesoscale convective systems to rainfall in the tropics. J. Climate, 27, 4952–4958, doi:10.1175/JCLI-D-13-00628.1.
Schumacher, C., and R. A. Houze Jr., 2003: Stratiform rain in the tropics as seen by the TRMM Precipitation Radar. J. Climate, 16, 1739–1756, doi:10.1175/1520-0442(2003)016<1739:SRITTA>2.0.CO;2.
Schumacher, C., and R. A. Houze Jr., 2006: Stratiform precipitation production over sub-Saharan Africa and the tropical east Atlantic as observed by TRMM. Quart. J. Roy. Meteor. Soc., 132, 2235–2255, doi:10.1256/qj.05.121.
Schumacher, C., R. A. Houze Jr., and I. Kraucunas, 2004: The tropical dynamical response to latent heating estimates derived from the TRMM Precipitation Radar. J. Atmos. Sci., 61, 1341–1358, doi:10.1175/1520-0469(2004)061<1341:TTDRTL>2.0.CO;2.
Schumacher, C., M. H. Zhang, and P. E. Ciesielski, 2007: Heating structures of the TRMM field campaigns. J. Atmos. Sci., 64, 2593–2610, doi:10.1175/JAS3938.1.
Sherwood, S. C., and R. Wahrlich, 1999: Observed evolution of tropical deep convective events and their environment. Mon. Wea. Rev., 127, 1777–1795, doi:10.1175/1520-0493(1999)127<1777:OEOTDC>2.0.CO;2.
Sohn, B.-J., J. Schmetz, R. Stuhlmann, and J.-Y. Lee, 2006: Dry bias in satellite-derived clear-sky water vapor and its contribution to longwave cloud radiative forcing. J. Climate, 19, 5570–5580, doi:10.1175/JCLI3948.1.
Stephens, G. L., and Coauthors, 2002: The CloudSat mission and the A-Train: A new dimension of space-based observations of clouds and precipitation. Bull. Amer. Meteor. Soc., 83, 1771–1790, doi:10.1175/BAMS-83-12-1771.
Thampi, B. V., and R. Roca, 2014: Investigation of negative cloud radiative forcing over the Indian subcontinent and adjacent oceans during the summer monsoon season. Atmos. Chem. Phys., 14, 6739–6758, doi:10.5194/acp-14-6739-2014.
Tobin, I., S. Bony, and R. Roca, 2012: Observational evidence for relationships between the degree of aggregation of deep convection, water vapor, surface fluxes, and radiation. J. Climate, 25, 6885–6904, doi:10.1175/JCLI-D-11-00258.1.
Webster, P. J., and G. L. Stephens, 1980: Tropical upper-tropospheric extended clouds: Inferences from winter MONEX. J. Atmos. Sci., 37, 1521–1541, doi:10.1175/1520-0469-37.7.1521.
Wielicki, B. A., B. R. Barkstrom, E. F. Harrison, R. B. Lee III, G. L. Smith, and J. E. Cooper, 1996: Clouds and the Earth’s Radiant Energy System (CERES): An Earth observing system experiment. Bull. Amer. Meteor. Soc., 77, 853–868, doi:10.1175/1520-0477(1996)077<0853:CATERE>2.0.CO;2.
Wilcox, E. M., and V. Ramanathan, 2001: Scale dependence of the thermodynamic forcing of tropical monsoon clouds: Results from TRMM observations. J. Climate, 14, 1511–1524, doi:10.1175/1520-0442(2001)014<1511:SDOTTF>2.0.CO;2.
Williams, M., and R. A. Houze Jr., 1987: Satellite-observed characteristics of winter monsoon cloud clusters. Mon. Wea. Rev., 115, 505–519, doi:10.1175/1520-0493(1987)115<0505:SOCOWM>2.0.CO;2.
Xu, W., and E. J. Zipser, 2015: Convective intensity, vertical precipitation structures, and microphysics of two contrasting convective regimes during the 2008 TiMREX. J. Geophys. Res. Atmos., 120, 4000–4016, doi:10.1002/2014JD022927.
Yuan, J., and R. A. Houze Jr., 2010: Global variability of mesoscale convective system anvil structure from A-Train satellite data. J. Climate, 23, 5864–5888, doi:10.1175/2010JCLI3671.1.
Yuan, J., R. A. Houze Jr., and A. Heymsfield, 2011: Vertical structures of anvil clouds of tropical mesoscale convective systems observed by CloudSat. J. Atmos. Sci., 68, 1653–1674, doi:10.1175/2011JAS3687.1.
Yuter, S. E., and R. A. Houze Jr., 1998: The natural variability of precipitating clouds over the western Pacific warm pool. Quart. J. Roy. Meteor. Soc., 124, 53–99, doi:10.1002/qj.49712454504.
Zipser, E. J., D. J. Cecil, C. Liu, S. W. Nesbitt, and D. P. Yorty, 2006: Where are the most intense thunderstorms on Earth? Bull. Amer. Meteor. Soc., 87, 1057–1071, doi:10.1175/BAMS-87-8-1057.
