A GCM-Oriented Ground-Observation and Phase Retrieval Simulator for a definition aware model evaluation of cloud and precipitation phase
Poster PDF
Authors
Katia Lamer — Brookhaven National Laboratory
Ann M. Fridlind — NASA - Goddard Institute for Space Studies
Andrew Ackerman — NASA - Goddard Institute for Space Studies
Pavlos Kollias — Stony Brook University
Eugene E. Clothiaux — Pennsylvania State University
Category
High-latitude clouds and aerosols
Description
With general circulation model (GCM) treatment of phase increasing in complexity comes a demand for more robust phase evaluation. Observations remain incomplete benchmarks but their value is substantially enhanced when model outputs are transformed in such a way as to mimic their limitations. Here, a GCM-oriented ground-observation and phase-retrieval forward-simulator ((GO)2-SIM) is presented. This forward-simulator currently interfaces with the ModelE3 GCM’s grid-average hydrometeor area fractions, mixing ratios, mass-weighted fall speeds and effective radii. Water content-based empirical relationships are used to emulate zenith observations from polarized micropulse lidar and Ka-band Doppler radar. For uncertainty quantification, 18 different relationships are implemented creating an ensemble of 576 forward-realizations.
Current results reinforce the idea that hydrometeor phase characteristics lead to distinct signatures in radar-lidar space. Modeled mixing ratios are used to determine thresholds in the forward space to spatially isolate observable ice, liquid and mixed-phase conditions in single and multi-layer clouds. An equivalent retrieval algorithm can be applied to the observed fields, thereby producing forward-retrieved and observationally-retrieved hydrometeor phase maps with similar limitations and uncertainties. Results over the North Slope of Alaska extracted from a 1-year global ModelE3 simulation show that (GO)2-SIM phase frequency of occurence assignment is insensitive to the choice of empirical relationships (IQR < 6.4%). Over all realizations, the technique leads to phase misclassification in no more than 7% of the cloudy grid cells; its main limitations lying above the first liquid layer where it may misinterpret ice-snow mixtures as mixed-phase and cannot identify pure liquid layers.
Due to its ability to capture effects such as gaps in observability of high clouds and supercooled liquid, (GO)2-SIM is expected to help assist in choosing between several possible diagnostic schemes for GCMs such as ModelE.
