Cloud clustering and

microphysical processes

Lau et al. (1991) analyzed infrared radiance measurements at the cloud top from the Japanese GMS to study the structure and propagation of tropical cloud clusters over the equatorial western Pacific. The observed cloud clusters display a hierarchy of collective motions at time scales of 1 day, 2-3 days, and 10-15 days, respectively. The 1-15-day time scale is closely related to the intraseasonal oscillation, and their super cloud clusters propagate eastward along the equator from the Indian Ocean to the western Pacific all around the global tropics. The cloud clusters embedded in the super cloud clusters have the 2-3-day time scale, and propagate in the opposite direction to the super cloud clusters. The diurnal time scale is most significant in the cloud clusters, in which the signals are more pronounced over the continent than over the open ocean.

Sui and Lau (1992) analyzed the First GARP Global Experiment Illb circulation data along with the Japanese GMS-1 IR data to study the multiscale variability in the atmosphere over the tropical western Pacific during the 1979 Northern Hemisphere. Two intraseasonal oscillations propagate eastward from the Indian Ocean to the western Pacific. Over the western Pacific warm pool, the intraseasonal oscillations develop into quasi-stationary systems with the enhanced rotational circulations. The intrasea-sonal oscillations interact with regional and synoptic-scale systems such as monsoon circulations. The intraseasonal oscillations also excite the 2-4-day disturbances. Sui and Lau also found that the diurnal signal becomes strong when the intraseasonal oscillation loses the intensity whereas the opposite is true.

To investigate the relevant cloud clustering processes (formation and evolution of cloud ensembles), Peng et al. (2001) performed a numerical experiment using a two-dimensional cloud ensemble model covering a basin-scale domain (15 360 km) with prescribed warm SST surrounded by cold SST, to mimic the equatorial western Pacific. The model used an open lateral boundary. Under the condition of no prescribed basic zonal flow and no initial perturbation, deep convective clouds develop in hierarchical clustered patterns, which are limited to the area of warm SST above 28° C. The most fundamental cloud cluster in the model has a horizontal scale of a few hundred kilometers, in which new cumulus clouds are generated at the leading edge of a propagating surface cold pool, the "gust front." It may last for days and propagate for a long distance if the background flow is broad and persistent, as is the case in the low-level convergence zone of the SST-induced background flow.

The largest hierarchical propagating cloud systems in the model have horizontal scales of up to 3000 km and consist of up to four cloud clusters that are generally of the gust-front type. The constituent cloud clusters are generated intermittently and have life spans of 12-36 h. The internal heating of the constituent clusters collectively induces an overall troposphere-deep gravity wave (Mapes and Houze, 1993; Mapes, 1993). The overall wave travels in the direction of the tropospheric deep shear at a speed determined by the thermodynamic asymmetry in the wave created by the transition from warm and moist incoming air in the front to drier and cooler air in the rear.

The development of new cumulus clusters in the gust-front region of the hierarchical system is due to the combined effect of the overall wave and the gravity waves excited by the constituent clusters on the lower-tropospheric stability. When there are no interruptions from outside the cloud system, new cloud clusters develop intermittently from shallow disturbances hundreds of kilometers ahead of the existing deep convection. The resulting hierarchical cloud pattern resembles the observed equatorial super cloud cluster (SCC) in the time-longitude diagram. However, the life spans of the constituent clusters of the simulated system are shorter than that in the observed SCC.

The dynamic processes for cloud clustering are intimately coupled to microphysical processes. This coupling may be revealed by water- and ice-cloud contents and their corresponding microphysics. This is investigated in several companion papers. Li et al. (2002c) simulated the cloud clusters using the two-dimensional cloud-resolving model with the imposed forcing from the TOGA COARE data during the disturbed period. The cloud clusters move westward, while the embedded individual clouds propagate eastward. Along with the westward propagation and during the development of tropical convection, the area-mean vertical velocity profiles exhibit the major ascending motion below 500 mb in the western half of the cloud, and the maximum ascending motion between 300 and 500 mb in the eastern half, indicating that the western half of the cloud undergoes the deep convective development whereas the anvil cloud grows in the eastern half. The surface rainfall is much larger in the western half than in the eastern half. The amount of water hydrometeors is much larger than that of ice hydrometeors in the western half, whereas ice and water hydrometeors have similar amounts in the eastern half. The analysis of the rainwater budget reveals that the collection of cloud water by raindrops is a major process for the surface rainfall, and thus the water hydrometeor processes are dominant in the deep convective clouds in the western half, whereas both the collection of cloud water by raindrops and the melting of precipitation ice into raindrops are responsible for the surface rainfall in the anvil clouds in the eastern half.

The simulations show that the performance of cloud-microphysical parametrization schemes has the direct, crucial impacts on the simulations of cloud clusters in the genesis, evolution, propagation, and amplitudes. However, the computation of the full set of cloud-microphysical equations is time-consuming. Li et al. (2002c) found from their analysis of cloud-microphysical budgets that in the deep tropical convective regime, the magnitudes of 12 terms out of the total of 29 cloud-microphysical processes are negligibly small. Thus, they proposed a simplified set of cloud-microphysical equations, which saves 30-40% of CPU time. The neglected terms in the simplified set include the accretion of cloud ice and snow by raindrops, the evaporation of melting snow, the accretion of cloud water and raindrops by snow, the accretion of raindrops and the homogeneous freezing of cloud water by cloud ice, the accretion and freezing of raindrops by graupel, the growth of cloud water by the melting of cloud ice, and the growth of cloud ice and snow by the deposition of cloud water. An experiment with the simplified set of cloud-microphysical equations was conducted and compared to an experiment with the original set of cloud-microphysical equations. The two experiments show similar time evolution and magnitudes of temperature and moisture profiles, surface rain rates including stratiform percentage and fractional coverage of convective, raining and nonraining stratiform clouds. This suggests that the original set of cloud-microphysical equations could be replaced by the simplified set in simulations of tropical oceanic convection.

Sui and Li (2005) analyzed the same TOGA COARE experiment performed by Li et al. (2002c) to show that interaction between ice and water clouds is crucial in determining the life span of deep convective and stratiform clouds and the evolution of cloud clusters. They defined a cloud ratio that is the ratio of the vertically integrated content of ice clouds (ice water path, IWP) to the liquid water path (LWP) to study the ice-water-hydrometeor interaction processes and their impacts in the development of convec-tive and stratiform clouds. Clouds become more stratiform when the tendency of the cloud ratio is positive whereas they become more convective when the tendency of the cloud ratio is negative. The advantage of the definition of the cloud ratio is to mathematically derive a tendency equation of the cloud ratio based on the prognostic cloud equations in the cloud-resolving framework. The derived cloud ratio budget indicates that the tendency of the cloud ratio is determined by the vapor condensation and deposition (cloud sources), rainfall and evaporation (cloud sinks), and conversion between ice and water hydrometeors including melting of precipitation ice to raindrops and accretion of cloud water by precipitation ice. The analysis reveals that the tendency of the cloud ratio is mainly determined by the vapor condensation and deposition during the genesis and decay of the tropical convection when the system is relatively weak, whereas the tendency is controlled by the convection process during the development of tropical convection, when the system is relatively strong.

Sui et al. (2007a) proposed using threshold values of the cloud ratio to separate the convec-tive component of the precipitation from the remainder. The cloud variables (IWP, LWP, and their ratio) are physically linked to the cloud microphysics, as demonstrated by an analysis of simulated cloud microphysics budgets in the same two-dimensional cloud-resolving model experiment subject to the imposed forcing from the TOGA COARE. Their analysis suggests that rainfall can be designated convective when the corresponding value of the cloud ratio is smaller than 0.2, or the value of IWP is larger than 2.55 mm. The remaining grids are classified as mixed and stratiform when the corresponding range of the cloud ratio is 0.2-1.0, and greater than 1, respectively. The new partition method is evaluated by the vertical velocity (w) data. The frequency distribution of w shows that w in the convective region has a wide distribution, with maximum values exceeding 15 ms-1. In the designated stratiform region, the distribution is narrow, with absolute values of w confined within 5ms-1. The statistics of w and the budgets of cloud microphysics are consistent with corresponding physical characteristics of the convective and stratiform regions of precipitation. The w distribution in the mixed region exhibits features more convective than stratiform, indicating a transition stage of convective development. But the consideration of features like fractional cloud covers, rain rates, vertical velocity profiles, and the corresponding wave response leads us to regard the mixed and stratiform regions as the nonconvective region.

Was this article helpful?

0 0

Post a comment