1.5
Sensitivities of SCM models to improved parameterizations of cloud-radiative interactions for tropical cirrus
Greg M. McFarquhar, NCAR, Boulder, CO; and R. C. J. Somerville, S. F. Iacobellis, and P. Yang
Parameterizations of the mean single-scattering properties (e.g., single-scatter albedo, extinction coefficient, asymmetry parameter) for distributions of ice crystals in tropical cirrus clouds are developed using measurements of ice crystal sizes and habits derived from two-dimensional crystal images, together with an improved geometric ray-tracing method. This effort represents an improvement over past studies in that observed mixtures of crystal shapes and sizes are used rather than single crystal habits to develop the parameterization, the numbers of ice crystals smaller than 100 micrometers are included in the observations and hence in the parameterization, and uncertainty estimates based on a Monte Carlo approach are incorporated in the derived relations. A new simplified parameterization of effective radius in terms of IWC and temperature is also introduced based on the observations.
Using the Scripps single-column model (SCM) and observational data collected during TOGA COARE that have been previously used to simulate cloud properties and radiative fluxes in the tropical environment, the sensitivity of model-produced cloud properties and radiative fluxes to the new parameterization is examined. Comparing the mean modeled fields with those obtained from prior simulations allows a determination of which cloud and radiative properties are most affected by the cirrus parameterization, and comparing the temporal evolution from different simulations allows a determination of the conditions under which these cirrus parameterizations are most important. Further, the accuracy of the modeled cloud and radiative fields is estimated by using the uncertainty estimates of the parameterization coefficients. Simulations with the extreme outliers of the coefficients allow a determination of the maximum variation of the cloud and radiative properties from their mean values. Monte Carlo simulations indicate how large the deviations in the radiative fluxes can be. Implications for climate studies are discussed.
Session 1, Climate Modeling: Climate Change (parallel with Session 2)
Monday, 15 January 2001, 9:00 AM-12:00 PM
Previous paper Next paper