RELATIVE OCCURRENCE OF LIQUID, ICE AND MIXED-PHASE CONDITIONS WITHIN CLOUD AND PRECIPITATION REGIMES: LONG TERM GROUND-BASED OBSERVATIONS FOR GCM MODEL EVALUATION
Open Access
- Author:
- Lamer, Katia
- Graduate Program:
- Meteorology and Atmospheric Science
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- December 04, 2018
- Committee Members:
- Eugene E. Clothiaux, Dissertation Advisor/Co-Advisor
Johannes Verlinde, Committee Chair/Co-Chair
Andrew Mark Carleton, Committee Member
Chris Eliot Forest, Committee Member
Ann M. Fridlind, Outside Member - Keywords:
- model evaluation
mixed-phase clouds
rain rate
observational benchmark
forward-simulators
subcolumn generator
ground-based observations
general circulation models - Abstract:
- The representation of clouds and precipitation in the NASA general circulation model ModelE3 is evaluated using observations collected by ground-based millimeter-wavelength radars and near-infrared lidars operated by the Atmospheric Radiation Measurement (ARM) Program. Using a retrieval, I produced a climatological benchmark to evaluate ModelE3 representation of warm precipitation intensity and occurrence in the Eastern North Atlantic. To objectively evaluate the occurrence of liquid water and ice in single and multi-layer hydrometeor systems produced by ModelE3, I addressed the many factors that differentiate ModelE3 and ARM observations. I constructed a forward-simulator to transform simulated hydrometeor mass-weighted properties to sensor backscattering-based observables. To assess uncertainty in this process, I generated an ensemble of 576 forward simulations using different water-content-to-backscattering relationships which propagated to no more than 3.7 % uncertainty in hydrometeor frequency of occurrence. I also showed that, by considering instrument detection capabilities in forward simulations, it is possible to objectively determine which hydrometeor-containing model grid cells are comparable to observations, bypassing the need to arbitrarily filter trace amounts that may be numerical noise. Then, I applied the same threshold-based water phase retrieval to both forward-simulated and observed depolarization lidar and Doppler radar observables. Using output from a ModelE3 simulation in the forward simulator framework, I evaluated that using fixed empirical thresholds in this retrieval leads to only 7.6 % unphysical water-phase classifications. I addressed the scale difference between ARM observations and ModelE3 by degrading the vertical resolution of observations to match ModelE3 using a backscatter-weighting technique and by increasing ModelE3 horizontal resolution through the distribution of each of its simulated stratiform and convective clouds and precipitation fractions across 5000 horizontally-microphysically-homogenous subcolumns following assumptions consistent with ModelE3’s microphysical scheme. Finally, I created a display that summarizes statistics on the vertical distribution of liquid water, ice water and mixed-phase conditions within the context of thirteen different hydrometeor vertical layering regimes, which I defined based on three altitude-defined atmospheric layers. A statistical comparison of ModelE3 and observations for the month of December at the North Slope of Alaska suggested that ModelE3 produced insufficient amounts of cirrus and too much liquid water in low-level clouds.