AMBIENT CLOUD SAMPLING STRATEGY
The strategy to sample ambient (unseeded) cloud systems draws on the "closure study" paradigm that was shown to be very successful during ACE-2 and CRYSTAL-FACE. It consists of four sampling strategies conducted in succession. The first leg is flown below cloud to fully characterize the aerosol feeding the clouds at their bases. Self-consistency among aerosol size, chemistry, hygroscopic growth, and CCN behavior will be tested in an "aerosol-CCN" closure. The second leg samples the lowest 100 meters or so of the cloud base to measure cloud microphysics and cloud turbulence in the region where activation takes place. These data will be used to test cloud parcel models that predict cloud drop number concentration based on sub-cloud aerosol properties and updraft velocity. The third leg is a spiral up through the cloud to characterize the vertical profile of cloud microphysics and precipitation. The following questions will be addressed using the microphysical profile: Is it possible to link the vertical and horizontal variation of cloud drop dispersion to turbulent processes? to mixing with detrained air? to the chemistry of the aerosol on which the cloud forms? to giant nuclei? Does entrainment mixing appear to be homogeneous or inhomogeneous or a conditional combination of the two? Is it possible to predict which clouds are precipitating based on the updraft strength, cloud thickness, and sub-cloud aerosol characteristics? The last leg is an extended run above the cloud system to see how well the radiometric retrievals of cloud properties compare to the observed microphysical profile, and to characterize the properties of the free troposphere aerosol that mixes with the cloud from above. These radiometric runs provide evidence for indirect effects (if they are apparent) as well as validation of satellite remote sensing techniques for these cloud systems.

CLOUD SEEDING EXERCISES
Two cloud seeding sources will be used: ship tracks (if and when they can be identified using satellite remote sensing) and the cloud seeding flares. When a ship track is found, sampling will proceed as described above, but in a race-track pattern transverse to the track, so that at each altitude leg, both in- and out-of-track samples will be characterized to maximize the contrast. Use of the cloud seeding flares is more exploratory. The first step will be to characterize the particle emission by the flares. Cloud activation models will be used in the field to determine the predicted effect these flares will have on the cloud as a function of measured updraft velocity distribution and sub-cloud ambient aerosol characteristics. It has been shown that different cloud systems have different susceptibilities based on the local meteorology and ambient aerosol characteristics. The activation model will predict the susceptibility of these different cloud systems to various types of cloud forcing: Twomey effects; precipitation suppression (Albrecht effect); and precipitation enhancement (precipitation seeding effects). It is anticipated that in some meteorological conditions the perturbation may induce drizzle (i.e. the cases where there are few giant nuclei in the upper reaches of the cloud), and in other cases it may suppress drizzle (cases in which the mean droplet radius is suppress to below 14 microns the nominal threshold for the onset of collision-coalescence drizzle formation). Coupled with a Gaussian dispersion model, the effects will be predicted as a function of time and space as the perturbation slowly disperses.

IMPORTANCE OF FORECASTING

The Naval Postgraduate School yields access to a wealth of satellite images and forecast modeling data that can be used to identify ship tracks and those meteorological variables (cloud cover, inversion strength, sea surface temperature) that are most important for predicting cloud susceptibility. These observations and predictions will thus guide the choice of where to fly, which types of experiments to conduct, and even whether or not to have a "no-fly" day.

PRE-EXPERIMENT MODELING
Key to the strategy will be modeling the experiment well before reaching the field. A 3-D cloud resolving model (RAMS) that was successfully used to model sub-tropical trade cumulus in HALO and tropical cumulus congestus in CRYSTAL-FACE is presently being modified to describe marine stratocumulus systems. The planned sampling strategies can be tested against the predicted cloud fields by asking: What is the evolution of the cloud over the time it takes to make a single sampling run? What is the timescale for the dispersion of the cloud seeding flare? Can Lagrangian parcels be coupled to the cloud activation model to describe the penetration of a perturbation to upper cloud regions where remote sensing strategies can detect it? Under what background conditions a strong signal be most likely? Do dynamical responses to aerosol forcing have a signature that can be identified by either the in situ measurements or the highly resolved remote sensing measurements? By answering these questions before going to the field, it is possible to design sampling strategies that provide the best opportunity for success.

In summary, the following elements of the above strategy give this proposed experiment a high probability for success. (1) Use of a full hierarchy of models both before and during the field campaign to anticipate the meteorological conditions and to shape the sampling strategies; (2) use of closure study strategies using state-of-the-art instrumentation that proved very successful in the CRYSTAL-FACE campaign last July; and (3) use of local contrasts in aerosol conditions under identical meteorological conditions by cloud seeding and pursuit of ship tracks to isolate the aerosol indirect effect. The data from these experiments should validate an array of process models that are central to the development of fundamentally based algorithms for the prediction of aerosol indirect effects in local, mesoscale, and global models.

| top |



©
Caltech Last Update: April 9, 2003