The Arizona Monsoon Boundary is defined as the boundary separating two distinctly different air masses over Mexico, the southwestern United States, and the adjacent Pacific during the summer. The structure and dynamics of this boundary are examined by cross-sectional analysis using three different data sources: (1) a time-height cross section, constructed using radiosonde observations, at the time the boundary initially passed through Tucson in 1984; (2) a composite cross section through the boundary, constructed from the Fleet Numerical Oceanography Center analysis; and (3) a cross section through the boundary using high-resolution fields of temperature, moisture, and geopotential height obtained from the VISSR Atmospheric Sounder (VAS). All three cross sections showed similar structure. In some respects, the Arizona monsoon boundary resembles a mid-latitude front with a distinct and relatively sharp air mass change across the boundary, forced almost entirely by confluence. A direct ageostrophic circulation is produced by this forcing, giving weak ascent on the warm, moist side of the boundary. The gradients and flow associated with the composite boundary are weaker, by a factor of four, than those associated with strong mid-latitude fronts. However, the VAS cross section suggests that, at times, the strength of the boundary approaches that of middle-latitude fronts. The wind shear suggested by the composite boundary ought to be unstable to baroclinic or barotropic processes. Disturbances developing along the boundary have been observed. One example of such a disturbance is examined using GOES imagery, lightning strike data, cloud track winds, and VAS data. Satellite images show the disturbance resembling a mid-latitude occluded cyclone, with an apparent low pressure center over northern Baja California and front-like cloud features extending eastward and southward from the low. Lightning strike data show convective activity occurring along the front-like features. Wind data indicate the presence of a cyclonic circulation south of San Diego along the Baja California coast. Cross sections using VAS data suggest that barotropic and baroclinic energy sources are present and suggest the front-like nature of the cloud feature extending southward from the low pressure center. Additionally, a second disturbance that eventually interacted with the monsoon boundary is briefly examined using satellite imagery.

Squall lines possessing nearly all the characteristics of tropical squall lines occasionally develop during the summer monsoon over southern Arizona and northwestern Mexico. Initial thunderstorm formation is over the mountains along the Continental Divide in the late afternoon. Satellite imagery, cloud-to-ground lightning strike data, and surface observations indicate the squall lines move from east to west or northeast to southwest by discrete propagation faster than all the winds below 20 kPa so that most of the anvil clouds lag behind. The synoptic-scale circulation is anomalous with a strong ridge located over the western United States and a deep trough located over the eastern United States. West to northwest winds are found in the boundary layer over southern Arizona and northwest Mexico while a deep layer of east winds are observed above. As a result, most of the environmental wind shear is confined to the lowest 2.5 km above the ground. The low-level wind shear seems to be required for the westward propagation of thunderstorms and the formation of the squall lines. Extremely dry midtropospheric air develops in the easterly flow through some combination of advection and subsidence and also appears to be an important factor in the development of the squall lines. A two-dimensional, nonhydrostatic, numerical model was able to simulate many of the features observed in these squall lines. Solar heating of the elevated terrain in the model caused the initial thunderstorm to develop over the Continental Divide. Continued development of new thunderstorms to the west of the Divide produced a squall line that travelled westward by translation of cells and discrete propagation, wherein new cells would develop 10-25 km ahead of the old ones, at a speed greater than all the winds below 30 kPa. Upward motion produced by westward propagating gravity waves and by the strong low-level convergence found just ahead of the gust front appeared to cause several episodes of discrete propagation. The creation of horizontal potential temperature gradients and the vertical and horizontal advection of preexisting vorticity gradients combined to produce the vorticity field associated with the rear inflow jet that developed beneath the simulated squall line.

A ground-based near-infrared spectroradiometer was built and used to measure relative spectral reflectance from cumulus congestus and cumulonimbus clouds during the 1985 and 1986 Arizona summer monsoon seasons. Thermodynamic phase was inferred from spectral features in the regions between 1.55-1.75μm and 2.1-2.3μm where there are distinct differences between absorption in liquid water and ice and absorption by water vapor is very weak. Although liquid water and ice are nearly transparent in the visible, they absorb weakly in the near-infrared and that absorption is amplified by multiple scattering in clouds. Reflectance measurements are simple to make, requiring neither high spectral resolution nor absolute detector response. Three distinct aspects of differences between absorption in liquid water and ice were used to infer phase: (a) Ratio of the signal at 1.65 μm to that at 2.2 μm; (b) Wavelength of peak signal in the 1.65 μm water vapor transmission window; (c) Half-bandwidth of the 2.1-2.3 μm feature. Representative spectra are presented and analyzed on the basis of the predicted behavior of liquid water and ice cloud absorption. The results are consistent with young cumuli rapidly glaciating as they reach cooler levels, well before evidence of anvil formation or fibrous structure, contrary to the notion that phase can be inferred from visible cloud features.