Date of Award
Doctor of Philosophy (PhD)
Convection can rapidly and efficiently transport polluted boundary layer air to the upper troposphere and lower stratosphere, thereby influencing the chemical composition and distribution of greenhouse gases in the atmosphere. Whether mass detrains into the upper troposphere or lower stratosphere has differing impacts on the radiative budget and hence, on climate. Currently, there have been only a few observing platforms capable of studying convective mass transport, which have significant limitations and are frequently restricted to field campaigns resulting in a small number of case studies. Outside of these case studies, little is known about the actual heights that convection detrains mass to or how much dilution a parcel rising in the updraft experiences due to processes such as entrainment. Entrainment not only reduces updraft buoyancy resulting in lower mass detrainment altitudes, but also dilutes updrafts that may be vertically transporting polluted boundary layer air, changing the chemistry of the detrained air aloft. To account for many of the limitations in observations, model simulations are commonly utilized; however, these models are unconstrained and need to correctly depict both the chemistry and dynamics. To improve our understanding of convective mass transport and help constrain model simulations, this study focuses on 1) identifying whether convection-allowing models can accurately depict the dynamics of mass transport, 2) building a large database of observed convective detrainment heights to determine the heights that convection detrains mass to, and 3) developing a methodology to retrieve observed fractional entrainment rates for deep convection that can be used to determine how much dilution is experienced by rising parcels.
These three objectives were researched as follows. First, biases within high-resolution convection-allowing model forecasts were identified with focus on the vertical structure and depth of deep moist convection. The object-based validation revealed that while the models performed well near the surface, there were large biases aloft. Overall, model forecasts generated too many convective elements that were individually too large and contained convection that reached the mid-troposphere twice as often as observations, leading to an over-estimation of the amount of mass being transported. Second, to determine the heights that convection actually detrains mass to, a large observational database of convective detrainment heights for the midlatitudes was built using ground-based radar observations. A newly developed radar echo stratification scheme was combined with high-resolution radar composites and an anvil-proxy methodology to retrieve the level of maximum detrainment (LMD) for convection across seven years for the months of May and July. Results showed that on average the LMD height was around 4.3 km below the tropopause, but can be as high as 2 km above the tropopause, with at least some mass transport occurring up to 6 km above the tropopause. May storms had a slightly higher mean tropopause-relative LMD height but July contained storms with the deepest transport. An analysis focusing on morphology found that quasi-isolated strong convection had higher LMD heights than mesoscale convective systems, with the highest LMD heights belonging to supercells. When subset by region, the southern regions were found to have lower mean LMD heights due to a large amount of diurnally-driven convection. Third and finally, to better understand why storms detrain mass to certain altitudes and to investigate the dilution of parcels with updrafts, a buoyancy-based methodology was developed that builds upon and constrains plume theory with observations. The methodology works on the principles of comparing the buoyancy of an ideal parcel to that of a mixed parcel with attributes derived from observations of vertical velocity and environmental temperature and moisture. The method was applied to a case of weaker, mid-level convection and a case of a deep convective cluster. The deep convective cluster was found to have mean fractional entrainment rates of around 0.26 km-1, which was about half of the mean rate found for the weaker, mid-level convective cell. The entrainment results also illustrated the importance of accounting for processes such as hydrometeor drag and the ice phase within the rising plume.
Overall, this study demonstrates the importance of including vertical information in analysis of both models and observations. The identified model biases in convective structure showcase where the convection-allowing models still need improvement and can be used to investigate where biases in precipitation fields originate. The LMD height retrievals depict the heights of mass detrainment and can be used to constrain chemical transport models in order to get more accurate approximations of transport heights for radiative and climate models. The statistical distribution of detrainment heights can also be used to estimate the amount of mass transport that occurs into the troposphere and stratosphere. The entrainment retrieval methodology can be applied to several observational datasets to retrieve fractional entrainment rates for convection of various morphologies and depths as long as vertical velocity and environmental temperature and moisture information is present. By incorporating observations, the entrainment rate retrievals can be used to constrain cumulus parameterization and idealized parcel models. Furthermore, the LMD and detrainment envelope retrievals can be merged with the entrainment retrieval methodology to determine how much dilution parcels experience before being detrained. Lastly, further study is required to investigate why supercells detrain mass to higher altitudes than other forms of convection.
Starzec, Mariusz, "Modeled And Observed Dynamical Characteristics Of Convective Mass Transport" (2018). Theses and Dissertations. 2431.