Tropical deep convection, entrainment, and dilution during the DYNAMO field campaign
Date
2014
Authors
Hannah, Walter, author
Maloney, Eric, advisor
Randall, David, committee member
Johnson, Richard, committee member
Venayagamoorthy, Karan, committee member
Journal Title
Journal ISSN
Volume Title
Abstract
This dissertation presents a study of outstanding questions in tropical meteorology relating to tropical deep convection, entrainment, and dilution. Much of the discussion in this study will focus on an important convectively-coupled phenomenon in the tropical atmosphere known as the Madden-Julian Oscillation (MJO), which is an eastward propagating atmospheric disturbance over the Indian and West Pacific Oceans that dominates the tropical variability on intraseasonal timescales (30-90 days). The MJO is most active during the Northern Hemisphere winter season and is characterized by alternating periods of enhanced and suppressed convective activity. A field campaign known as the "Dynamics of the MJO" (DYNAMO) was conducted in the boreal winter months from October 2011 through February 2012 to study the initialization of the MJO with in-situ observations. The first part of this study examines hindcast simulations of the first two MJO events during DYNAMO in a general circulation model (GCM). The model used for this is the National Center for Atmospheric Research (NCAR) Community Atmosphere Model (CAM5) version 5, which uses parameterized convection. In these simulations, an entrainment rate parameter is varied to test its effects on the representation of the MJO, following previous studies. Hindcast simulations with CAM5 reveal that the entrainment parameter can improve the representation of the MJO. However, analysis of the column integrated moist static energy (MSE) budget reveals that this improvement is the right answer for the wrong reason. CAM5 incorrectly enhances vertical MSE advection, which compensates for cloud radiative feedbacks that are too weak. A promising theory for the MJOs fundamental dynamics is that of a moisture mode. The gross moist stability (GMS) describes the ratio of advective MSE import to a measure of convective activity. Negative GMS, and specifically the vertical component of GMS (VGMS), is thought to be a necessary condition for the destabilization of a moisture mode. In CAM5, VGMS becomes negative when the entrainment parameter is increased, indicating that the model can more easily destabilize moisture modes. However, this is inconsistent with re-analysis data, which exhibits positive VGMS. The second part of the study examines hindcasts using the super-parameterized version of CAM5 (SP-CAM) that uses embedded cloud-resolving models (CRM) to explicitly simulate convection on the sub-grid scale. SP-CAM was used for these hindcast simulations because previous studies have shown this type of model can reproduce the MJO much better than conventional GCMs. SP-CAM hindcasts yield a more robust MJO representation than CAM5, as expected. SP-CAM has an interesting systematic drift away from the initial conditions that projects well on the Real-time Multivariate MJO index (RMM), which negatively impacts the RMM skill scores. Analysis of the column MSE budget shows that SP-CAM has more realistic cloud-radiative feedbacks when compared to CAM5. SP-CAM also has a bias towards stronger import by vertical MSE advection that is similar CAM5 and inconsistent with re-analysis data. VGMS in SP-CAM is also found to be negative, which is similar to CAM5 and inconsistent with re-analysis data. The results from the first part of this study highlight a paradox surrounding entrainment. Although, previous studies have shown that entrainment rates should be larger than typical values used in parameterizations, increasing the entrainment rate does not make global model simulations more realistic. This prompted a detailed investigation into entrainment processes in high-resolution CRM simulations. A series of simulations are conducted where deep convection is initiated with a warm humid bubble. The bubble simulations are compared to a more realistic field of deep convection driven by forcing derived from the DYNAMO northern sounding array data. Entrainment and detrainment are found to be associated with toroidal circulations, consistent with recent studies. Analysis of the directly measured fractional entrainment rates does support an inverse relationship between entrainment and cloud radius, as is often assumed in simple models of deep convection. A method for quantifying the dilution by entrainment is developed and tested. Dilution and entrainment are generally not synonymous, but dilution is found to have a weak inverse relationship to cloud core radius. Sensitivity experiments show that entrainment and total water dilution are enhanced with environmental humidity is increased, contrary to the assumptions of some parameterizations. More vigorous convection in a more humid environment is better explained by a reduction of buoyancy dilution. An additional sensitivity experiment shows that entrainment and dilution are enhanced when convection is organized by the presence of vertical wind shear. The enhanced dilution is associated with entrainment of drier air on average.
Description
Rights Access
Subject
parameterization
intraseasonal variability
tropical meteorology
convection
entrainment