An Observational Comparison of Level of Neutral Buoyancy and Level of Maximum Detrainment in Tropical Deep Convective Clouds

Abstract

Tropical deep convective clouds are important drivers of large‐scale atmospheric circulation representing the main vertical transport pathway through the depth of the troposphere for heat, momentum, water, and chemical species. The strength and depth of this transport are impacted by the convective updraft size and intensity that are driven by buoyancy, dynamical forcing, and mixing of environmental air, that is, entrainment. In this study, we identify tropical deep convective systems with well‐defined forward anvils using Atmospheric Radiation Measurement (ARM) ground‐based profiling radars, at three ARM fixed sites in the Tropical Western Pacific (TWP; i.e., Manus, Nauru, and Darwin) and three ARM Mobile Facility deployments in Niamey, Niger; Gan Island, Maldives; and Manacapuru, Brazil. We use the difference between the level of neutral buoyancy (LNB) and the level of maximum detrainment (LMD) as a proxy for the effective bulk convective entrainment (εproxy). The LNB, the theoretical height that a parcel raised above the level of free convection would reach with no mixing, is calculated based on preconvection radiosonde measurements using parcel theory. The LMD is the height of the maximum reflectivity observed in forward anvil clouds by profiling radars. Deep convective systems over the TWP show higher LNBs that extend to 16.3 km on average and larger εproxy (median value of LNB minus LMD up to 6.5 km) compared to their continental counterparts in the Amazon and West Africa. Oceanic conditions show larger convective available potential energy (CAPE) coupled with higher moisture at low levels, which favors larger εproxy. In contrast, continental cases initiate and develop, under high convective inhibition, steeper environmental lapse rate, and high wind shear conditions, which show smaller offset between LNB and LMD. Deep convective cases that promote significant cold pools at the surface experience less εproxy. Using a Random Forest regression algorithm, CAPE is associated with the highest feature importance score for predicting convective εproxy, followed by low‐level relative humidity. For continental cases, the low‐level wind shear also indicates higher importance.

Improved Convective Ice Microphysics Parameterization in the NCAR CAM Model

“…One striking feature is the marked increase in the detrained mass above 10 km relative to the XReICE_EL17_rime simulation. The range of mass detrainment heights (from 8 to 15 km) for deep convection is comparable to previous findings informed by analysis of satellite and ground‐based observations over the ARM TWP Darwin site (Deng et al., 2016; Takahashi et al., 2017; Wang et al., 2020). Though quantifying and deriving detrained ice requires further novel observational techniques and modeling studies, inclusion of snow detrainment does alleviate a suspected underestimation of detrained FWC (Figure 8a).…”

Factors Governing Cloud Growth and Entrainment Rates in Shallow Cumulus and Cumulus Congestus During GoAmazon2014/5

“…Differences between our study and JD06 could be attributed to the geographic differences between the maritime tropical congestus observed by JD06 and the continental tropical clouds of the Amazon, a difference also recently found by Wang et al. (2020b). Sea breezes and other mesoscale circulations (Burleyson et al., 2016) unique to the Amazon may contribute to the differences, along with possible aerosol‐mediated feedbacks on the local humidity (Abbot & Cronin, 2021).…”

Vertical Structure of Tropical Deep Convective Systems at Different Life Stages From CloudSat Observations

“…All the DCSs in the following text are one‐convective‐pillar DCSs, as shown in Figure 2a. Furthermore, for the one‐convective‐pillar DCS samples, we separate the anvil clouds on both sides of the convective pillar as the “major anvil” (longer anvil) and the “minor anvil” (shorter anvil) by their anvil lengths (Cetrone & Houze, 2011; Wang et al., 2020), as shown in Figure 2a4. To construct the composite cloud images (Section 2.4) of the total DCS samples, we also align all the cloudy pixels according to the distance from the DCS center, which would maintain the mushroom‐like structure of the DCS.…”