Solar flares are generally identified as an important source of disturbances that affect space weather, while many aspects of the basic flare process are still not well understood. The purpose of the present work has been to investigate the small and large-scale magnetic structures and their evolution associated with flares in the context of magnetic reconnection, based on which the goal of this dissertation is to further the understanding of the various properties of flares and related phenomena, including their origin, precursors, and evolution of morphology in solar atmosphere.

The research presented in this dissertation relied upon multiwavelength observations of flares from hard X-rays to radio wavelengths obtained from several ground-and space-based instruments. The studied topics in the flare core regions include microflares, flare-induced evolution of the photospheric magnetic field, and sigmoids. The large scale study includes remote brightenings, Moreton waves, type II and III radio bursts, and coronal mass ejections (CMEs). Statistical studies were carried out for microflares and the evolution of flare core fields. Large-scale activity was analyzed by examining two major eruptions, the 2003 October 29 X10 flare and the 2005 May 13 M8.0 flare.

The main findings in this dissertation are as follows: (1) results of X-ray spectral fitting indicate the nonthermal origin of X-ray emission at over ~10 keV during the impulsive phase of microfiares and the photon spectra of the microflares associated with type III bursts are generally harder than those without type III bursts; (2) white-light and magnetogram observations of 6 sunspots reveal the rapid penumbral decay and central umbral/penumbral darkening associated with flares and suggest that the flaring magnetic fields change from a highly inclined to a more vertical configuration. Moreover, these changes are irreversible; (3) in the 2003 October 29 X10 event, remote brightenings more than 2 x 10⁵ km away from the main flare were ignited by hot particles transported along closed magnetic fields; meanwhile, this event was strong enough to indicate the common origin of the Moreton waves, the type II radio bursts and the CME; and (4) multiwavelength signatures were found to support the argument that the 2005 May 13 M8.0 flare originated from a characteristic sigmoidal active region following the tether-cutting flare model.

The major contribution of this dissertation is the discovery of new observational evidence to enlighten the future flare/CME modeling: