3.4 Exploring the Microphysical Structure of Mesoscale Convective Systems through Theory, Observations, and Numerical Simulations

Tuesday, 24 January 2017: 11:15 AM
2AB (Washington State Convention Center )
Hannah C. Barnes, PNNL, Richland, WA

Professor Houze’s work to understand the microphysical processes in MCSs over the last thirty years has come full circle. The theories he proposed early in his career are now being confirmed through observations and used to validate numerical simulations.

In the 1970s and 1980s Professor Houze used a combination of ground observations, aircraft data, and theory to suggest how microphysical processes are spatially organized within MCSs. For example, within the stratiform region,Leary and Houze [1979b] indirectly inferred that rimed particles may exist in a layer just above the brightband using drop-size distributions measured below the brightband.Houze[1989] developed a comprehensive conceptual model that described how microphysical processes are spatially distributed within MCSs. Within the stratiform region, this conceptual model suggested that deposition occurs whenever upward motion is found above the 0°C level but aggregation and riming are restricted to temperatures between -12°C and 0°C.

While these conceptual models have been referenced in numerous papers, obtaining data capable of validating these holistic conceptual models has been challenging since traditional aircraft and ground observations only obtain data within a small region of the storm. Additionally, the ability for numerical simulations to represent these microphysical patterns has not been extensively investigated. However, recent work published by Professor Houze that uses dual-polarimetric radar observations and the Weather Research and Forecasting (WRF) model starts to provide insight into both of these gaps in the literature.

Dual-polarimetric radar data is a powerful tool for investigating the microphysical structure of MCSs. These radars emit and receive both horizontally and vertically polarized pulses. By comparing the power and phase shift of these pulses, dual-polarimetric radars obtain variables that indicate the physical characteristics of the particles. Particle identification algorithms (PIDs) combine these dual-polarimetric radar variables with a temperature profile in order to provide an indication of the dominant hydrometeor and/or microphysical process at a given location within a storm.

Using dual-polarimetric and Doppler radar data obtained by the NCAR S-PolKa radar during the Dynamics of the Madden-Julian Oscillation / ARM MJO Initiation Experiment (DYNAMO/AMIE) field campaign,Barnes and Houze[2014] spatially composited PID data around the airflow patterns observed in MCSs and showed that hydrometeors and microphysical processes are systematically organized around the kinematic structure of MCSs. Specifically, the frozen hydrometeors in the stratiform portion of MCSs were characterized by a layered pattern: small ice crystals were concentrated near echo top; dry aggregates existed in a relatively deep layer above the brightband, but did not reach echo top height; rimed particles occurred in isolated, shallow pockets just above the brightband; wet aggregates were concentrated in a thin layer at and just below the brightband. If these hydrometeor types are interpreted with respect to the microphysical process involved in their generation, these composites suggest that deposition occurs everywhere above the brightband, aggregation occurs near the brightband band and extends upward, riming is concentrated in isolated, shallow regions near the brightband, and melting occurs near the brightband. Thus, radar observations presented inBarnes and Houze[2014] provides observational evidence that supports the presence of rimed particles above the brightband that was inferred inLeary and Houze[1979] and the layered microphysical structure theorized inHouze[1989].

One of the most challenging aspects of validating microphysical processes in numerical simulations is the fact that microphysical processes are dependent on the dynamical structure of convection. If the airflow in the observed and simulated convection significantly differs it is impossible to determine if microphysical differences between the simulations and observations are the result of the dynamical or microphysical differences. Dual-polarimetric radars again provide a means to overcome this challenge.

Barnes and Houze[2016] first assimilated Doppler velocity data into WRF to force simulated MCSs to have midlevel inflows that were as similar to DYNAMO/AMIE observations as possible. Then the spatial distribution of deposition, aggregation, riming, and melting from three microphysical parameterizations was composited around the simulated midlevel inflow. It was found that each parameterization had a broad layered pattern that was consistent with observations presented inBarnes and Houze[2014] and conceptual model presented inHouze[1989]. Additionally, each parameterized supported the inference made byLeary and Houze[1979] that rimed particles exist in the stratiform region above the brightband. However, each parameterization tended to allow riming to occur too frequently and over too deep of a layer. Additionally, simulated aggregation was found to extend over too great of a depth. This new knowledge of the spatial organization of simulated microphysical processes provides motivation to continue ongoing efforts to improve microphysical parameterizations.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner