This paper summarizes the results of a detailed study from the Coopera
tive Convective Precipitation Experiment (CCOPE) of the vertical struc
ture of mesoscale gravity waves that disturbed a sizable part of the t
roposphere and that played a significant role in the generation of a m
esoscale convective complex. These bimodal waves displayed periods of
148 (50) min, wavelengths of 135 (60) km, and phase speeds of 15.2 (19
.8) m s-1. A comparison is made between wave-induced pressure perturba
tion fields derived from triple-Doppler wind fields within regions of
essentially nonconvective precipitation, pressure perturbation fields
obtained by bandpass filtering of surface mesonetwork data, and the ve
rtical structure of the pressure eigenfunctions as predicted from a li
near stability analysis. It is believed that this represents the first
such application of the Doppler radar pressure retrieval technique to
the study of gravity waves. In addition, an analysis of the potential
for shear instability was performed on all of the special CCOPE sound
ings taken on this day to determine the representativeness of the chos
en soundings for the theoretical analysis and the likelihood that a wa
ve maintenance mechanism endured throughout the 33-h wave event. The a
nalysis of the potential for shear instability and the eigenfunctions
both indicate that the bimodal waves were able to efficiently extract
energy from the mean flow near several closely spaced critical levels
in the 4.0-6.5-km layer to maintain their coherence for many wave cycl
es. This result serves as the explanation for the observed ability of
the waves to organize precipitation into long convective bands whose a
xes were along and just ahead of the wave crests. The eigenvalue analy
sis predicts unstable modes that are hydrostatic, nondispersive, ducte
d gravity waves characterized by half of a vertical wavelength contain
ed between the ground and the lowest critical level (at z = 4 km). Eig
enfunctions of pressure and other variables all display negligible til
t below 2.3-3.3 km, above which a sudden reversal in phase occurs. The
vertical structure of the Doppler-derived fields associated with one
of these gravity waves agrees in terms of the following respects with
the eigenfunction predictions and/or the surface mesoanalyses: (a) the
vertical wavelength, horizontal structure, and amplitude of the pertu
rbation horizontal wind and pressure fields, and (b) the in-phase cova
riance between the pressure and horizontal wind fields at levels below
2.5 km. On the other hand, the theory predicted a much more abrupt ve
rtical transition in phase in the pressure fields and weaker amplitude
s aloft than were evident in the Doppler analyses. In addition, the si
ze of the multiple-Doppler analysis domain was too small to capture an
entire horizontal wavelength of the 135-km-scale gravity wave, which
made direct comparisons difficult. Furthermore, the linear theory pred
icts much smaller amplitudes and somewhat longer horizontal wavelength
s for the vertical motions characterizing both wave modes than those s
een in the Doppler winds, which likely also contain nonwave effects. T
hese discrepancies are largely due to the combined effects of weak con
vection, turbulence, and data sampling problems. Despite these drawbac
ks, the findings from this and other recent studies using Doppler rada
rs and ground-based radiometers suggest that remote sensing of mesosca
le gravity waves that occupy a significant fraction of the troposphere
should be exploited further.