Tropical deep convective cloud morphology

Abstract

A cloud-object partitioning algorithm is developed. It takes contiguous CloudSat cloudy regions and identifies various length scales of deep convective clouds from a tropical, oceanic subset of data. The methodology identifies a level above which anvil characteristics become important by analyzing the cloud object shape. Below this level in what is termed the pedestal region, convective cores are identified based on reflectivity maxima. Identifying these regions allows for the assessment of length scales of the anvil and pedestal of the deep convective clouds. Cloud objects are also appended with certain environmental quantities from the ECMWF reanalysis. Simple geospatial and temporal assessments show that the cloud object technique agrees with standard observations of local frequency of deep-convective cloudiness. Additionally, the nature of cloud volume scale populations is investigated. Deep convection is seen to exhibit power-law scaling. It is suggested that this scaling has implications for the continuous, scale invariant, and random nature of the physics controlling tropical deep convection and therefore on the potentially unphysical nature of contemporary convective parameterizations. Deep-convective clouds over tropical oceans play important roles in Earth's climate system. The response of tropical, deep convective clouds to sea surface temperatures (SSTs) is investigated using this new data set. Several previously proposed feedbacks are examined: the FAT hypothesis, the Iris hypothesis, and the Thermostat hypothesis. When the data are analyzed per cloud object, each hypothesis is broadly found to correctly predict cloud behavior in nature, although it appears that the FAT hypothesis needs a slight modification to allow for cooling cloud top temperatures with increasing SSTs. A new response that shows that the base temperature of deep convective anvils remains approximately constant with increasing SSTs is introduced. These cloud-climate feedbacks are integrated to form a more comprehensive theory for deep convective anvil responses to SST. An investigation into the physical shape and size of mature, oceanic, tropical, deep convective clouds is conducted. Mean cloud objects are discussed. For single-core clouds, the mean cloud has an anvil width of 95 km, a pedestal width of 11 km, and an anvil thickness of 6.4 km. The number of identified convective cores within pedestal correlates well with certain length scales and morphological attributes of cloud objects. As the number of cores increases, so does the size of the mean cloud object. Pedestal width is shown to regress linearly to anvil width when a 2/3rd power scaling is applied to pedestal width. This result implies continuous but retarded growth of anvils with growing pedestals and equivalence in the mass flux convecting through the pedestal and into the anvil. Trends in cloud scales with cloud base and top heights are investigated to shed light on related convective parameterization assumptions and on convective transport, respectively. Many of the results obtained using the CloudSat methodology are also examined with a large-domain radiative-convective equilibrium numerical simulation and are found to exhibit similar trends when modeled. Finally, various CloudSat sampling issues are discussed in several appendices. Utilizing the CloudSat cloud object database, an examination of the sensitivity of oceanic, mature, deep convective cloud morphology to environmental characteristics is conducted. Convective available potential energy (CAPE), aerosol optical depth, mid-level vertical velocity, and troposphere deep shear are all included as meteorological measures. The sensitivity of various aspects of convective morphology to each one of these environmental characteristics is assessed individually. The results demonstrate that clouds tend to be invigorated by higher CAPE, aerosol amount, and upward mid-level vertical velocity. Stronger shear tends to make clouds wider but also shallower. The relative importance of each of these, and some additional, environmental measures to trends in cloud morphology are compared. It is found that aerosol, mid-level vertical velocity, and sea surface temperature tend to be the most influential environmental characteristics to convective morphology. The results are shown to be insensitive to the manner in which the environment is measured. The potentially surprising insensitivity of cloud morphology to CAPE is discussed in detail.