Abstract
We study two microwave bursts (which occurred about 12 minutes apart in the same active region) that
show contrasting characteristics in morphology and spectral variation in microwave data from the Owens
Valley Radio Observatory (OVRO) Solar Array and the Nobeyama Radioheliograph (NoRH). The first
flare (X-ray class C2.8, 23:22 UT) shows radio morphology that changes from a single source at a low
frequency (5 GHz) to a double source at high frequencies (10-17 GHz), and shows a total power spectrum
that hardens in the decay phase. In contrast, the second flare (X-ray class M1.0, 23:35 UT) shows a simple,
single-source morphology at 7-17 GHz. This source is located at one of the two footpoints of a small
loop identified in Yohkoh/HXT maps, and its spectrum evolves in a typical soft-hard-soft pattern. We
infer the trap properties of these two loops from the microwave spectral behaviors and a coronal field
extrapolation from the vector magnetogram of the active region obtained from Mees Solar Observatory,
along with X-ray maps obtained from Yohkoh/SXT. It is shown that the radio characteristics of the first
event are consistent with the hypothesis in which Coulomb collisions dominate in a highly anisotropic
loop with low ambient electron density ~5x109 cm-3. The second flare involves a small loop in which a
more uniform field strength allows rapid loss of electrons by precipitation, resulting in a spectral
variation mainly controlled by the energy dependence and time profile of the injection. In spite of the
significantly differing decay times (~8 and ~1 minutes, respectively), very similar injection times of ~30 s
are inferred from the analysis of time profiles under the above hypotheses. We therefore present a view
that the largely different morphologies and the corresponding spectral variations are consequences of the
differing trap properties rather than dissimilar injection properties. The low trap density inferred for the
first event (~5x109 cm-3) as compared with the second event (~8x1010 cm-3) also explains why this event
showed richer microwave output in spite of weaker GOES activity.