Development of an upper-level cloud parameterization for large scale models
Date
1995-08-03
Authors
Nebuda, Sharon E., author
Journal Title
Journal ISSN
Volume Title
Abstract
The interaction of clouds with the general circulation is generally agreed upon to be the most important physical process requiring improvement in today's climate models. Due to the limited spatial and temperal resolution of most large-scale models, representing the detailed physical properties of clouds has been difficult. To overcome this limitation, a one-dimensional, upper-level cloud model has been developed which can be nested in time and space in a. localized area with limited frequency. The adaptive cloud model will provide microphysical and radiative information for the large-scale model. The cloud parameterization was developed using the existing physics in the Regional Atmospheric Modeling Systems (RAMS) developed at Colorado State University (CSU). The microphysical routine requires the large-scale model to maintain the variables of liquid and ice water as well as total ice number concentration. The cloud model includes six water species of total water, water vapor, rain, small ice (pristine ice), large ice (snow), and aggregates. Cloud water is computed as a residual of the other water categories. A subgrid turbulence scheme which predicts the vertical velocity variance provides the mixing created by radiative destabilization. The radiation routine distinguishes between liquid and ice water and computes heating rates which can significantly influence the large-scale circulation. To evaluate the upper-level cloud model several initializations were used by both a psuedo-lD format of RAMS and a psuedo-GCM format of RAMS in conjunction with the cloud model to simulate cirrus. The results from the cirrus cases created by the pseudo-lD format of RAMS provides an understanding of the response of RAMS physics to the lD dynamics. These simulations also provide a control run for judging the cloud model when called by the psuedo-GCM RAMS. The cirrus initializations range in altitude from 125 mb to 450 mb at both tropical and middle latitudes. The cloud was initiated by either elevated relative humidity (large-scale weather disturbance) or addition of cloud water (detrainment from convection). The results indicate the upper-level cloud model is successful at reproducing the same features of the cloud as modeled by the psuedo-lD version of RAMS. Among the features of the cirrus is the deepening of the cloud layer through gravitational settling of larger ice particles which can significantly alter the radiative heating rates in the cloud layer. Clouds created by large-scale weather disturbances warmer than -50°C or initiated by convection with significant amounts of ice (0.5 g/kg) a.re more likely to generate radiative forcing that would noticeably impact the large-scale circulation. The performance of the upper-level cloud model was highly dependent on the quality of initialization by the large-scale model. Preliminary results suggest the cost of the upper-level cloud model can be reduced by limiting its frequency and diagnosing the precipitation of large ice crystals from the cloud layer similar to the technique of Ghan and Easter (1992).
Description
August 3, 1995.
Also issued as author's thesis (M.S.) -- Colorado State University, 1995.
Also issued as author's thesis (M.S.) -- Colorado State University, 1995.
Rights Access
Subject
Cloud physics -- Mathematical models
Atmospheric circulation