Approaches for Better Use of Observations

Analysis of Time History in Satellite Retrievals

Observations and analyses from a combination of various satellite instruments (e.g., from the A-Train constellation) provide a comprehensive dataset for studying global cloud and aerosol properties, their radiative effects, and their interrelationship. It is critical to determine whether the correlations imply causation (e.g., whether correlations between cloud albedo and aerosol optical depth from adjacent clear-sky regions constitute evidence for aerosol indirect effects). Correct construction and interpretation of the correlations require methods for identifying measurements with similar physical and chemical tendencies over the lifetime of the clouds prior to measurement. These tendencies are not available from the instantaneous observations collected by polar-orbiting satellites. However, the tendencies could be constructed in a chemical and meteorological assimilation of satellite observations. Analyses of satellite data might be enhanced to include the tendencies of physical and chemical environmental factors that have been shown to govern cloud-aerosol interactions.

Model-Data Comparisons to Infer Causalities

To explore the extent to which observed correlations between cloud and aerosol properties represent causality, similar correlations can be computed in models. Consistency between a process model and observations and sensitivity studies with the model, with and without relevant processes, provide confidence that causality can be inferred from observed correlations (Feingold and Siebert, this volume).

Careful Choice of Analyzed Situations

The use of natural meteorological variability is a means of testing the ability of a model to represent correctly the key physics of the aerosol indirect effects. Careful choice needs to be made regarding possible locations where this might be beneficial. The system should have some simple and effective proxies for meteorological control, the meteorological variability must not greatly swamp the essential aerosol signal, the albedo of the clouds in the system need to be potentially susceptible to perturbations in aerosols, and the clouds in the system should have sufficient horizontal homogeneity such that their properties (microphysical and macrophysical) are detectable from space. To some degree, marine stratocumulus satisfies these criteria. Spaceborne retrievals of CDNC, LWP, and cloud cover, and possibly other variables (e.g., drizzle from Cloudsat), can be used for model evaluation. CDNC could serve as a proxy for the aerosol variability. In principle, models could be used to calculate the radiances actually measured by satellites.

Ensembles of LES/CRMModel Simulations

The understanding of aerosol-cloud interactions at a large scale might be advanced by simulating broad ranges of meteorological situations with process models. LESs are highly idealized, limited domain (order of 10 km) simulations. When embedding LES within a regional model, mesoscale forcing is provided to the finer LES grid. Two-way nesting allows for communication between grids. In a different approach, meteorological parameters, such as SST

and large-scale subsidence, may be varied along with aerosol perturbations to explore the relative importance of aerosol and meteorology.

A multiscale modeling framework, which replaces the conventional cloud parameterizations with a CRM in each grid-column of a general circulation model, or global CRMs are ideal tools to apply CRMs to a large variety of meteorological situations for all regions of the globe. These could be used to investigate the effects of indirect aerosols on a large scale (Collins and Satoh; Grabowski and Petch, both this volume).

Improving Large-scale Model Parameterizations

It has been a long-standing goal to use field programs, process models, and satellite observations effectively to improve GCM parameterizations, which are of critical importance to represent the effects of indirect aerosols accurately. Still, this remains a daunting challenge. A basic strategy to achieve this goal is first to develop confidence in process models (e.g., LES, CRMs). Field campaigns can play a central role in doing this. Next, process models must be used in the parameterization process. Finally, satellite observations, including statistical analysis of covariances, provide a means to evaluate the GCM with its parameterization.

Ice and Mixed-phase Clouds

Much of the previous discussion centered on the link from aerosols to the CDNC and albedo of warm clouds. Whereas warm clouds (extensive stratus or stratocumulus sheets especially over the oceans) are important for the radiation balance, mixed-phase and ice clouds are more important in terms of the possible effect of aerosols on precipitation. However, there are much larger uncertainties associated with the processes acting in these clouds.

0 0

Post a comment