This section emphasizes microphysical aspects of severe thunderstorms, squall lines, convective complexes, rainbands, and hurricanes. In-depth reviews of these topics are considered in separate reviews of mesoscale phenomena---warm season, (Smull [1995]), and cold season (Businger [1995]).
The North Dakota Thunderstorm Project was a field program designed to study entrainment, cloud ice initiation and evolution, storm structure, atmospheric chemistry, and electrification in thunderstorms and was conducted in the Bismark, North Dakota area from 12 June through 22 July 1989 (Boe et al. [1992]). Reinking et al. [1992] studied a small, strongly sheared storm during this project using an airborne Doppler radar and other in-situ measurements and found a highly inefficient storm that transported a large portion of the water substance out through the anvil, in agreement with previous theories and observations of highly sheared storms. The storm inflow was initially near the surface in response to surface heating; during the mature storm stage the surface inflow was diminished and replaced by mid-level inflow from a south-flank seeder cell field.
Brooks and Wilhelmson [1992] performed a three-dimensional model simulation of a low precipitation supercell thunderstorm. They showed that low precipitation storms can form in environments supportive of supercell storms when the initial temperature pulse is small and suggest that these storms produce low precipitation due to the enhanced turbulent entrainment in the highly sheared environment.
Chilson et al. [1993] studied a tropical thunderstorm near Puerto Rico using collinear dual wavelength Doppler radars and found structures similar to thunderstorms observed during the GARP Atlantic Tropical Experiment (GATE). They also showed that the use of reflectivity to estimate the fall speed of particles as part of determining the vertical component of wind during Doppler wind field analysis is unreliable when turbulence is present.
Raymond et al. [1991] studied vertical and detrained mass fluxes throughout the life cycle of four thunderstorms in New Mexico using measurements from four Doppler radars and an instrumented aircraft. Updrafts were found to entrain a certain amount of radar invisible air from neighboring towering cumulus clouds. Entrained air did not immediately mix with the cloud base air, but tended to maintain its own identity in ascent. Levels with decreasing parcel buoyancy induced significant detrainment from the updraft. These results support the buoyancy sorting arguments discussed by a number of authors.
Biggerstaff and Houze [1991] analyzed the mature phase of the 10-11 June 1985 squall line during the Preliminary Regional Experiment--STORM (PRE-STORM) and found that the region of heaviest stratiform precipitation was immediately downwind of the most intense portion of the convective line, and that the width of the trailing stratiform region was controlled by a combination of the wind velocity and the fall speed of the hydrometeors.
Gamache [1990] presented in-situ aircraft measurements of the
microphysical structure of Summer MONEX (Monsoon Experiment) convective
and stratiform clouds and showed the stratiform regions to be dominated
by branched crystals and aggregates. Ice crystal concentrations in the
convective clouds were found to be very high (average of 230 l
), while
the concentrations of particles in the stratiform regions were found to
be an order of magnitude lower (average of 20 l
). The convective
regions were characterized by a near absence of liquid water in the
temperature range between -10 and -22
C, suggesting that the
convective updrafts were nearly completely glaciated. They also suggest
that the convective regions may be the source of ice for the stratiform
region due to the high concentrations of ice in the former region.
McCumber et al. [1991] performed model simulations of a tropical squall line and a non-squall type convective line using three different microphysical options. The inclusion of ice processes enhanced the agreement of the simulated convection with some features of the observed convection, including the proportion of rainfall in the anvil region, and the intensity and structure of the radar brightband near the melting level in the anvil. The use of three ice classes produced better results than two ice classes or ice-free conditions.
Rasmussen and Smolarkiewicz [1993] studied the airflow and accompanying cloud bands forming upstream of the island of Hawaii and showed that the initiation of these cloud bands under typical trade wind conditions could be understood by the dynamics of strongly stratified flow past three-dimensional obstacles. Under the low Froude number conditions pertinent to the island of Hawaii, a low-level convergence zone forms upstream of the island over which clouds are initiated and then advected onshore by the cloud layer winds. Nocturnal surface cooling was shown to double the depth and maximum speed of the downslope flow, but had little effect on the location of the offshore convergence zone. Analysis of upstream soundings collected during the Hawaiian Rainband Project (HaRP) showed that the trade wind environment upstream of the island is favorable to the formation of cloud bands consisting of isolated cells advected onshore by the mean cloud layer winds.
Dual-aircraft investigations of Hurricane Norbert by Houze et al. [1992] and Gamache et al.[1993] determined the mesoscale distribution of ice particles and the water budget of the storm. Houze et al. [1992] show that large graupel particles were found in the eyewall, with evidence for rime-splintering on the outside edges of the eyewall convective updrafts. In the outer stratiform region, the predominant particle type was aggregates.
Microburst studies continued to show the importance of evaporative cooling of rain, cooling due to melting, and water loading to the production of microbursts (Straka and Anderson [1993], Proctor and Bowles [1992], Lin and Coover [1990], McNulty [1991], Lee et al. [1992], Kingsmill and Wakimoto [1991], and Mahoney and Elmore [1991]). Of particular interest are microbursts created by light precipitation (Proctor and Bowles [1992]). The light precipitation case studied by Proctor and Bowles occurred at Denver on 11 July 1988 and produced its most intense microburst downwind of the main precipitation shaft. The main precipitation shaft produced a relatively weaker microburst by comparison. Modeling studies suggest that the intense microburst was driven primarily by sublimating snow.