Degree Type


Date of Award


Degree Name

Doctor of Philosophy


Geological and Atmospheric Sciences

First Advisor

William A. Gallus Jr.


As computer power increases and model grid spacing decreases, more emphasis will be put on model microphysics to produce accurate forecasts of rainfall including that from warm-season mesoscale convective systems (MCSs). Some believe bin microphysical schemes are far superior to the commonly used bulk microphysical schemes because of their ability to more accurately depict certain processes like sedimentation. However, bin schemes are computationally inefficient and there are no plans in the near future to implement such schemes operationally. Instead, this study proposes to use a technique in Weather Research and Forecast (WRF) Advanced Research WRF (ARW) simulations that attempts to improve bulk microphysical forecasts of warm-season MCSs by harnessing the intrinsic characteristics of bin fall speed distributions that are important for the sedimentation process provided the fall speed characteristics in bin schemes differ from those in commonly used bulk schemes.

Fall speed distributions of rain, snow, graupel and cloud ice were compared between a bin scheme and three bulk schemes, and were found to be different between the different schemes. The microphysical processes that contributed the largest to the microphysical budget in the bin scheme often occurred with the slower fall speeds, but the opposite was true for the bulk schemes. There was evidence of size-sorting in the bin and Thompson bulk schemes, a naturally occurring phenomenon. This feature was not simulated in the WSM6 and Lin schemes and can be attributed to those schemes being single moment and the Thompson scheme being double moment in ice and rain.

Since the characteristics of the bin fall speeds were different from those in the bulk schemes, bin fall speeds were used to modify bulk scheme fall speeds using a probability matching technique that was developed to improve the prediction of warm-season MCSs. The sensitivity of different convective morphologies to the fall speed modifications was also evaluated. First, various tests were performed with various microphysical schemes and cases in order to find the vertical grid configuration that provides the best rainfall forecast. Rainfall forecasts worsened when the number of vertical levels was doubled from a control configuration of 31 levels and an over prediction of rainfall occurred. The largest improvement in skill occurred when the levels above the melting level were doubled and this was attributed to better resolved cold-cloud microphysical processes. As such, simulations using the probability matching technique employed the vertical configuration with refined vertical grid resolution above the melting layer.

The different convective morphologies responded similarly when the fall speed modifications were made with all systems simulating a narrower stratiform region, less stratiform rainfall and a larger anvil. Rainfall forecasts generally improved with the use of the probability matching technique with improvements in the lightest and heaviest rainfall. The reduced stratiform rainfall occurred as a result of slower falling snow and a reduction in downward fluxes of snow, while forecasts of convective rainfall intensity improved as a result of faster falling graupel. Sensitivity tests were performed by computing bulk-like fall speeds in the bin scheme, which resulted in a modification of the particle size distribution of snow, which led to faster falling snow, larger downward fluxes of snow and a larger stratiform rain region.


Copyright Owner

Eric Anthony Aligo



Date Available


File Format


File Size

200 pages