The SCMS was a field program observing small cumulus with aircraft and dual-wavelength radar in Florida in 1995. Charles Knight and L. Jay Miller completed a study of the very early radar echo histories of small cumulus, at 3- and 10-cm wavelengths. As a first step toward using the radar data to help understand the first stages of droplet coalescence, the Bragg and Rayleigh components of the echo must be unraveled so as to determine the echo intensities that depend upon droplet sizes. It appears now that there may be two distinct, important Bragg scattering sources with very different wavelength dependencies: one from humidity fluctuations within regions of mixing of cloud and dry, surrounding air, and the other from liquid water content variations within continuous cloud. The latter is a conjecture based upon correlations within the dual-wavelength data and the surprising but probable presence of this Bragg-scattering source within unmixed updraft. The in-cloud source appears not to obey the -5/3 scale-dependence law, while the former roughly does. The origin of these hypothetical liquid-water-content fluctuations is obscure, and at present efforts are focused upon aircraft data to test the hypothesis.

William Cooper (joint appointment with ASP) and Roelof Bruintjes (joint appointment with RAP), in collaboration with Graeme Mather (South African Weather Bureau) and Jean-Louis Brenguier (Meteo-France), used calculations and field observations to investigate the roles of large particles in the formation of rain by coalescence. The field observations from the Small Cumulus Microphysics Study (SCMS) indicate that the rate of the initial coalescence process is reasonably consistent with that calculated from the particle size distribution and CCN concentration measured at cloud base. The calculations also supported the importance of particles larger than about 1 µm, because the calculated coalescence rates were significantly higher when the effects of this size range were included. A related study of the effects of hygroscopic seeding, completed and published, also pointed to the importance of large CCN in rain formation.

Knight continued his collaborations with several university groups on the nonequilibrium antifreeze compounds from insects and fish, and has continued his collaboration with a group at Colorado School of Mines (under the leadership of Prof. E. Dendy Sloan) on the crystallization of clathrate hydrates. Knight and Richard Sommerfeld (U.S. Forest Service, Fort Collins), completed a study of diffusion of nitric acid within ice crystals, concluding that the diffusion constant and/or solubility of nitric acid within ice is much less than has been reported, and is in fact undetectably small in the experiments performed.

Charles and Nancy Knight completed a survey of the state of knowledge of hail formation within storms, for a projected new AMS Severe Storms Monograph.

Andrew Heymsfield, Larry Miloshevich and Steve Aulenbach, along with Glen Sachse (NASA Langley Research Center), continued their work on determining the ambient conditions required for ice crystals to form in cirrus clouds. These researchers used a combination of measurements acquired with airborne sensors during several NASA sponsored field campaigns, and with balloon-borne cryogenic hygrometer operated by Sam Oltmans (NOAA) over a six-year period. Consistent with earlier studies, they found that relative humidities with respect to water that are required for crystals to originate decline from almost 100 percent near -40°C to 75 or 80 percent from -55 to -65°C. The data they collected at temperatures below -50°C, however, added considerable new information, and were noteworthy in several respects. High relative humidities, approaching 90 percent, were measured in clear air at -52°C off the coast of California. Furthermore, RHs approaching 100% are noted in orographic wave clouds between -62 and -65°C, resulting from high vertical velocities and low ice crystal growth rates. These results indicate that very high RHs can build up at low temperatures in certain instances, prior to the formation of ice crystals.

Results from several balloon-borne replicator flights were also used to determine how ice particles develop in the vertical in mid-latitude cirrus clouds. The "ice crystal replicator is a simple, cost-effective, cloud sampling instrument capable of collecting particles as small as 10 microns while preserving the detailed morphological characteristics of the ice crystals. Detailed analysis of data from several replicator launches during the FIRE II cirrus experiment and an experiment in Arizona by Heymsfield, Miloshevich, and Aulenbach over the past year supported the view that cirrus crystals originate near cloud top, then grow in highly (ice) supersaturated air to near cloud base, and sublimate as they fall to cloud base.

Accurate profiles of atmospheric relative humidity (RH) from radiosonde measurements are important to the research community for initializing numerical models, determining the RH at which ice clouds form, operationally predicting the formation of aircraft contrails, developing and validating radiative transfer algorithms, and validating remotely-sensed retrievals of RH profiles. Because ice crystals replicated in cirrus clouds by the NCAR balloon-borne cloud particle replicators consistently showed characteristics of active crystal growth in an ice-supersaturated environment while simultaneous RH measurements from RS80-A radiosondes on the same balloon packages showed substantial ice-subsaturation, Miloshevich and Heymsfield in collaboration with Sam Oltmans (NOAA CMDL), investigated the accuracy of RH measurements from the widely-used Vaisala RS80-A radiosondes. An extensive set of simultaneous RH measurements from RS80-A radiosondes and the reference-quality NOAA/CMDL balloon-borne cryogenic frost-point hygrometer (see figure) show that: 1) ice-supersaturation is not uncommon in the atmosphere (Panel A), 2) the RS80-A rarely measures ice-saturation, and 3) the RS80-A increasingly under measures the RH as the temperature decreases, as judged by an imaginary envelope that overlies the bulk of the data in Panel B. A correction was developed from statistical analysis of this dataset based on the assumption that the cryogenic hygrometer measurements are equal to the true ambient RH. The correction under ambient conditions of ice-saturation is about 15 percent RH at temperatures of -30°C or warmer, increasing to 38 percent RH at -60°C. The correction is also strongly dependent on the ambient RH, increasing at -40°C from 6 percent RH when the ambient RH is 28 percent, to 28 percent RH when the ambient RH is 78 percent. The large magnitude of this measurement error has potentially significant implications for the research community.

Heymsfield and Aulenbach, along with Glen Sachse (NASA Langley Research Center) and Paul Lawson (SPEC, Inc.) extended their analysis of data from the SUCCESS project (Subsonic aircraft: Contrail & Cloud Effects Special Study), sponsored by NASA to examine the relationship of persistent contrail development on the ambient relative humidity field. In one instance, the NASA DC-8, which collected the microphysical data, produced a contrail at a temperature of -52°C, generated a contrail and then sampled the resulting contrail for over an hour. In this instance, the contrail developed ice streamers (precipitation trails) which contained ice particles hundreds of microns long that descended up to 1 km during the observational period.

These researchers found that inside the contrail core, ice particles remained relatively small due to high crystal concentrations (~ 10 cm ^-3) which reduced the vapor density to ice saturation. Mixing of moist environmental air and vapor-depleted contrail air produced localized regions of ice supersaturations along the contrail periphery where contrail crystals grew large enough to fall from the contrail into the vapor-rich environment below. As the heavier crystals left the contrail, a stochastic selection process caused others to move into the regions of ice supersaturation. As this process continued over time, it produced precipitation trails. In effect, the contrail core acted as a source of ice crystals, the contrail periphery acted as a growth region, and the environment provided a continuous source of vapor for particle growth.

Knowledge of the dependence of radiative transfer on tropical cirrus microphysical properties is needed to better understand the effects of clouds on climate. Many radiative parameterizations for climate models are formulated in terms of an effective radius, r_e, of a population of ice crystals. Although r_e is well defined for water clouds, a review of the different definitions for ice clouds showed substantial discrepancies, making direct comparison of different parameterization schemes difficult. Further, remote sensing retrievals that are frequently used to derive cloud properties for evaluation of climate models also generate a r_e value. An examination of the relative importance of different cloud layers for the retrieval of a single r_e representing the entire cloud showed that only the uppermost regions of thick cirrus would be detected using current visible and near-infrared satellite retrieval methods. Heymsfield and McFarquhar are currently working with James Spinhirne (NASA Goddard) for testing their r_e definitions against his LIDAR-based retrievals of r_e and extinction coefficient using data obtained during the Central Equatorial Pacific Experiment (CEPEX).

Another study, in collaboration with David Mitchell (Desert Research Institute) and Andreas Macke (University of Kiel), which aims to determine how well cirrus radiative properties can be modeled, is underway. Scattering phase functions are being computed from ice crystal size and shape distributions measured during CEPEX, and the extinction coefficients are similarly estimated from anomalous diffraction theory. Multi-channel measured reflectances are available from the MODIS airborne simulator for comparison with predicted radiative properties and for prediction of retrieved particle shapes and sizes (which can be directly compared with the microphysical measurements).

Modeling of global sulfur distributions (aerosol sulfate, sulfur dioxide, and dimethyl sulfide) was continued by Mary Barth (joint appointment with ACD), Philip Rasch (CGD) and Jeffrey Kiehl (CGD). Control simulations of present day emissions and pre-industrial emissions were performed. In these simulations, the parameterization depicting the aqueous-phase oxidation of sulfur dioxide was done in detail using a small timestep of two minutes and a diagnostic pH that varied as sulfate was produced because this aqueous oxidation process is the dominant pathway of producing aerosol sulfate (60-70 percent in the annual average). A sensitivity study showed that there was no difference in aerosol sulfate loadings between the present-day-emissions control simulation and a simulation that prescribed the pH of the cloud and raindrops to 4.5. From the present-day emissions control simulation, sulfur emitted in Europe and Asia contribute significantly to sulfate loadings in the Arctic. Comparison of model results to observations continues. For both remote and industrial regions, the comparison looks good for all four species prognosticated in the model. Future work includes using the global sulfur model for sensitivity tests such as incorporating a three-dimensional ammonia field to better characterize the sulfate aerosol, examining the effect of the dimethyl sulfide emissions, and investigating the nucleation rates of sulfate aerosols.

Cooper, Larry Radke (ATD), Alan Hills (joint appointment with ACD), and Charles Brock (ASP), in collaboration with David Rogers (Colorado State University) studied a new instrument in laboratory tests for the measurement of cloud condensation nuclei (CCN) and ice nuclei (IN), and the results were used to evaluate design modifications that will increase its temperature range and its reliability. As part of this design, some data collected in its first field deployment were examined. The instrument uses a controlled expansion to produce supersaturated conditions, then counts droplets that form under these conditions as they pass through an optical counter. After air is humidified to near 100 percent relative humidity, small pressure changes are produced by flow through capillaries, and capillaries of different sizes are switched into the line to produce a sequence of supersaturations. The instrument can also be operated at low temperature, where it can detect nuclei active as condensation-freezing or deposition nuclei. The objective is to extend its operation to temperatures below -40°C in order to detect the homogeneous freezing of solution droplets, thought to occur in the formation of cirrus and perhaps of contrails.

Processing and analysis of data collected during the STERAO/Deep Convection experiment are being coordinated by James Dye (joint appointment with RAP) in close collaboration with Steve Rutledge (Colorado State University), Jeff Stith (University of North Dakota), Brian Ridley (ACD), Pierre Laroche (ONERA, France), Gerd Hubler (NOAA Aeronomy Laboratory), Thomas Matejka (NSSL), and others. The STERAO-A experiment (Stratospheric Tropospheric Experiment: Radiation, Aerosols and Ozone, Part A. Deep Convection and the Chemical Composition of the Upper Troposphere and Lower Stratosphere) was conducted in northeastern Colorado during June through August 1996. The project had major goals of investigating the transport of chemical constituents by deep convection and the production of NO_x (comprised of NO and NO₂) by lightning and was unique in combining and coordinating extensive chemistry, air motion and electrical measurements in and around thunderstorms. It included participants from several universities and national laboratories. The major research facilities in the experiment included:

Many of the early analysis efforts focused on the storms of 10 and 12 July. The 10 July storm was a multicell storm, which evolved late in its life to become a quasi-supercell storm. Although the areal extent was small it had cloud tops above 16 km, which is very high for northeast Colorado. During the quasi-supercell stage work by Tim Lang (Colorado State University), Laroche and Dye showed that the lightning was almost all intra-cloud with very few cloud-to-ground flashes. Early Doppler synthesis by Matejka suggests that there was not an organized downdraft in this storm; thus most transport would have been from the boundary layer to the upper troposphere. In contrast, the 12 July storm covered a much larger area with tops to roughly 12 km and much CG and IC lightning. The Doppler analysis suggests that this storm did have an organized downdraft and might be expected to transport constituents from midlevels down into the boundary layer.

Dye worked with and guided Preston Heard (SOARS) on the small storm, which occurred on 27 June. This storm produced only nine CG flashes and about 55 IC flashes, but did have elevated levels of NO_x in the anvil. This small case provided a nice contrast to the much larger and more lightning-prolific storms of 10 and 12 July.

Analysis of NO_x production by lightning is in an early stage. Measurements obtained by Ridley from the Citation show a clear signature of significant enhancements of NO_x in the upper parts of the storm anvils, much like Ridley and Dye showed in an exploratory experiment in New Mexico in 1989. However, unlike New Mexico there is a substantial anthropogenic contribution to NO_x in the boundary layer, and the effects of transport of boundary layer NO_x compared to that of lightning will need to be separated. A somewhat surprising result is there was no striking evidence of enhancements of NO_x from lightning found in the boundary layer, unlike the large enhancements found in the anvil. Substantially more effort will be needed to understand the roles of transport and lightning in these storms.

NCAR | UCAR | NSF | NCAR FY96 ASR | NCAR FY97 ASR